Untitled

ORIGINAL ARTICLE
A Prospective, Multicenter Study of Developmental

Dysplasia of the Hip: What Can Patients Expect After
OOIXlWaXUZnczw/iMdshREa5aLNbRoI8k4/i3nEue4+M9SuoRpOlsks8q++lgMau5as6ohRU05i11Ve4Ujf1jppCIQP47X7YIo=
Downloaded from http://journals.lww.com/pedorthopaedics by VA2NP/EQXy7tVERrPFBmNAUmP6w2i5xIZXinZP55F
Open Reduction?
Sara N. Kiani, MPH,* Alex L. Gornitzky, MD,* Travis H. Matheney, MD, MLA,†
Emily K. Schaeffer, PhD,‡ Kishore Mulpuri, MD,‡ Hitesh H. Shah, MD,§ Ge Yihua, MD,∥
Vidyadhar Upasani, MD,¶ Alaric Aroojis, MD,# Venkatadass Krishnamoorthy, MD,**
Global Hip Dysplasia Registry, and Wudbhav N. Sankar, MD*
The majority (81%; n = 13/16) occurred in the first year after

Background: Although there are several predominantly initial OR. Excluding patients with repeat dislocation, 94.5% of
single-center case series in the literature, relatively little pro- hips were IHDI 1 at most recent follow-up. On the basis of strict
on 04/15/2023
spectively collected data exist regarding the outcomes of open hip radiographic review, some degree of PFGD was present in 44%
reduction (OR) for infantile developmental dysplasia of the hip of hips (n = 101/230) at most recent follow-up. Seventy-eight hips
(DDH). The purpose of this prospective, multi-center study was to (55%) demonstrated residual dysplasia compared with estab-
determine the outcomes after OR in a diverse patient population. lished normative data. Hips that had a pelvic osteotomy at index
Methods: The prospectively collected database of an interna- surgery had about half the rate of residual dysplasia (39%;
tional multicenter study group was queried for all patients n = 32/82) versus those without a pelvic osteotomy with at least
treated with OR for DDH. Minimum follow-up was 1 year. 2 years follow-up (78%; n = 46/59).
Proximal femoral growth disturbance (PFGD) was defined by Conclusions: In the largest prospective, multicenter study to date,
consensus review using Salter’s criteria. Persistent acetabular OR for infantile DDH was associated with a 7% risk of
dysplasia was defined as an acetabular index > 90th percentile re-dislocation, 44% risk of PFGD, and 55% risk of residual
for age. Statistical analyses were performed to compare pre- acetabular dysplasia at short term follow-up. The incidence of
operative and operative characteristics that predicted re-dis- these adverse outcomes is higher than previous reports. Patients
location, PFGD, and residual acetabular dysplasia. treated with concomitant pelvic osteotomy had lower rates of
Results: A cohort of 232 hips (195 patients) was identified; me- residual dysplasia. These prospectively collected, multicenter
dian age at OR was 19 months (interquartile range 13 to 28) and data provide better generalizable information to improve family
median follow-up length was 21 months (interquartile range education and appropriately set expectations.
16 to 32). Re-dislocation occurred in 7% of hips (n = 16/228). Level of Evidence: Level II, prospective comparative study.
From the *Department of Orthopaedic Surgery, Children’s Hospital of Key Words: Developmental Dysplasia of the Hip (DDH), Open
Philadelphia, Philadelphia, PA; †Department of Orthopaedics, Bos- Reduction, Re-dislocation, Avascular Necrosis, Proximal
ton Children’s Hospital, Boston, MA; ‡Department of Orthopaedics, Femoral Growth Disturbance (6)
University of British Columbia, Vancouver, BC, Canada; §Depart-
ment of Orthopaedics, Kasturba Medical College, Manipal, Karna- (J Pediatr Orthop 2023;43:279–285)
taka; #Department of Orthopaedics, Bai Jerbai Wadia Hospital for
Children, Mumbai, Maharashtra; **Department of Orthopaedics,
Ganga Hospital, Coimbatore, Tamil Nadu, India; ∥Department of
Orthopaedics, Shanghai Children’s Medical Center, Shanghai, China;
and ¶Department of Orthopaedics, Rady Children’s Hospital, San
Diego, CA.
This research received funding support from the Pediatric Orthopaedic
D evelopmental dysplasia of the hip (DDH) is the most
common developmental deformity of the lower
extremity.1 Many therapeutic options exist, with the
Society of North America (POSNA) the Canadian Orthopaedic
Foundation, the I’m a HIPpy Foundation, the Peterson Fund for overall goal of treatment being a stable, concentric
Global Hip Health, BC Children’s Hospital Foundation, Divi’s reduction that will facilitate continued femoral head
Foundation for Gifted Children and the Munday Family. development and acetabular growth and remodeling.2
This authors declare no conflicts of interest.
Reprints: Wudbhav N. Sankar MD, Department of Orthopaedics,
Depending upon the patient’s age at the time of initial
Children’s Hospital of Philadelphia, Philadelphia, PA 19146. E-mail: presentation, typical first-line treatment for infantile DDH
sankarw@chop.edu. includes bracing and potentially closed reduction/spica
Supplemental Digital Content is available for this article. Direct URL casting. In turn, open reductions are typically reserved for
citations appear in the printed text and are provided in the HTML patients who present at an older age and/or those who
and PDF versions of this article on the journal’s website, www.
pedorthopaedics.com. have failed previous orthotic treatment.
Copyright © 2023 Wolters Kluwer Health, Inc. All rights reserved. There has been a decrease in the rate of both closed
DOI: 10.1097/BPO.0000000000002383 and open hip reductions for the treatment of DDH, with
J Pediatr Orthop Volume 43, Number 5, May/June 2023 www.pedorthopaedics.com | 279
Copyright r 2023 Wolters Kluwer Health, Inc. All rights reserved.

Kiani et al J Pediatr Orthop Volume 43, Number 5, May/June 2023
the rate of closed reductions declining more rapidly than

that of open procedures.3 Although Sankar et al4 pre-
viously explored early results after closed reductions
using a prospective multicenter cohort, research on open
reductions has remained limited by small sample sizes
that are primarily retrospective in nature and usually

single-center based. Multiple studies have demonstrated

that clinical outcomes after open hip reduction are better
for younger patients and those treated with concomitant
pelvic osteotomy.5–7 Additional data also suggest that
patients who undergo open reduction alone are at in-
creased risk for proximal femoral growth disturbance
(PFGD) (often termed avascular necrosis) and
reoperation.5,8 Nevertheless, assessment of preoperative
and intraoperative risk factors for complications after
open reductions has been limited to smaller, single-center
cohorts.
The purpose of this study was to assess early out-
comes of open reduction for DDH using a prospectively
on 04/15/2023
collected database of an international multicenter study

group. Our primary goal was to report the overall results
of this procedure with specific attention to three major
adverse outcomes: re-dislocation, PFGD, and residual
acetabular dysplasia (RAD).
FIGURE 1. Consolidated Standards of Reporting Trials (CON-
METHODS SORT) diagram showing patients evaluated, excluded, and
enrolled by stage, along with primary and secondary outcomes
The data for this study were queried from the pro- at latest radiographic follow-up.
spective registry of a multicenter, international study
group - the Global Hip Dysplasia Registry - that collects
data on infants and children treated for DDH by any different institutions; group consensus was reached for
means. Patients with neuromuscular and syndromic di- any discrepancies.4 Each reviewer based their desig-
agnoses were excluded from the database. Institutional nations on the Salter criteria, but for the purposes of this
Review Board approval was obtained by all sites before manuscript, each hip was only classified with a simple
patient enrollment and written consent was obtained from “yes” versus “no” for radiographic evidence of PFGD to
all participating families. At each encounter, contributing reduce subtype variability.4,9,10
surgeons collected and uploaded all relevant patient in- RAD was assessed using Novais et al.’s normal-
formation and pertinent clinical details to a central data- ization formulas to calculate a z-score and corresponding
base, including hip radiographs. percentile for a patient’s acetabular index by age.11 RAD
Eligible patients were enrolled in the database con- was defined as an acetabular index at or above the 90th
secutively. All patients treated from 2010 to 2020 with an percentile at last follow-up. In addition, any child who
open hip reduction were queried. Bilateral hips were underwent a follow-up pelvic osteotomy after the index
counted separately. Exclusion criteria were lack of base- procedure was considered to have had RAD. For patients
line imaging before treatment (n = 43 patients) and <1-year with a pelvic osteotomy at index surgery, RAD was ana-
follow-up after open reduction (n = 194 patients; Fig. 1). lyzed for patients with at least 1 year of follow-up,
The database was designed to collect patient in- whereas for patients without a pelvic osteotomy at index
formation and clinical details for each included hip from surgery, RAD was analyzed for patients with at least
the time of initial consultation through most recent fol- 2 years of follow-up. This distinction was made as the
low-up. A standardization manual was used to provide a acetabulum remodels more slowly in the absence of a
uniform definition of the variables and outcomes col- pelvic osteotomy but tends to achieve most of its remod-
lected. Apart from PFGD, all radiographic measure- eling within the first 2 years.12
ments were utilized as originally entered by the treating Comparative analyses were performed for all out-
surgeon including follow-up IHDI grades and acetabular comes, with 2-tailed Fisher exact and χ2 tests for catego-
indices. A patient was classified as having a re-dis- rical variables and student t tests and Mann-Whitney U
location if their surgeon recorded their femoral head as tests for continuous variables. Categorical variables are
“dislocated” at a point after their open reduction. The reported as frequency and percentage. Continuous varia-
presence or absence of PFGD by most recent follow-up bles are reported as median and interquartile range, as all
was determined by independent, blinded assessment of continuous variables had non-normal distributions. A
follow-up radiographs by 3 study group members from 3 multivariate logistic regression was run to identify factors
280 | www.pedorthopaedics.com Copyright © 2023 Wolters Kluwer Health, Inc. All rights reserved.

J Pediatr Orthop Volume 43, Number 5, May/June 2023 Outcomes After Open Reduction
associated with the primary outcomes of the study

(re-dislocation, PFGD, and RAD). The following varia-
bles were used in the model: age at open reduction, lat-
erality (bilateral patients included in both), clinical
reducibility before open reduction, IHDI grade, ossific
center status and location, previous treatment (harness or

closed reduction), and open reduction approach. IHDI

grade was excluded from the re-dislocation variable due to
collinearity. If the presence of an ossific center was noted,
it was considered normally located if it was central within
the cartilaginous femoral head and thus appeared in the FIGURE 2. Summary findings for incidence of adverse clinical
middle of the hypoechoic femoral head region on ultra- outcomes after open hip reduction; “Missing hips” refer to hips
sound. It was considered eccentrically located if it was missing outcome data for a specific outcome measure at final
located closer to the periphery of the hypoechoic region of follow-up.
the femoral head on ultrasound. For RAD, pelvic
osteotomy was also included in the model. These factors factors or intra-operative characteristics that were
were chosen as they are commonly assumed to be risk associated with increased rates of re-dislocation in the
factors for outcomes after open reduction. The sig- multivariable analysis (Table 1; Fig. 3). Of those hips that
nificance level was set at p-value <0.05. Statistical analysis re-dislocated, 62.5% (n = 10) underwent further surgery,
on 04/15/2023
was performed using Stata BE/17.0 (StataCorp, College including eight hips treated with repeat open reduction
Station, TX). and 3 hips treated with pelvic osteotomy.
Radiographic evidence of PFGD was present in
43.9% of hips (101/230) by the time of final follow-up
RESULTS according to consensus review (Fig. 4A). Baseline
The study cohort consisted of 232 hips (195 characteristics associated with an increased risk of PFGD
patients) across 16 institutions and 6 countries. Median (Supplementary Table 2, Supplemental Digital Content 2,
age at the time of index open reduction was 19 months http://links.lww.com/BPO/A590; Fig. 3) included older age
(interquartile range 13 to 28 mo). For those with pre- at first encounter (P < 0.001), unilateral hip dysplasia
operative radiographs (n = 204; remaining had only (vs. bilateral; P = 0.004) and an irreducible hip on clinical
preoperative ultrasounds), the proportion of infants with examination at any time before open reduction (P = 0.007).
a prereduction IHDI grade of 2, 3, or 4 was 4.9%, 26.5%, Patients with a history of previous harness use had
and 68.6%, respectively. An ossific center was present in decreased rates of PFGD compared with those who did
81.0% of hips and 26.3% of all hips had an eccentric not (P = 0.007). Intraoperative details associated with
ossific nucleus. increased risk for PFGD were an older age at time of
In the overall cohort, 24.2% of patients had pre- open reduction (P < 0.001), adductor tenotomy (P = 0.022),
viously been treated with a harness and 11.4% had a prior a concomitant pelvic osteotomy (p-value = 0.007), or a
attempt at closed reduction. The anterior approach was femoral shortening osteotomy (P < 0.001). In the multi-
used for 83.0% of hips. Soft tissue releases were performed variable regression model (Table 1), factors protective
in 62.1% of all hips, including 56.5% with adductor tenot- against PFGD included bilateral DDH [odds ratio (OR)
omy and 52.6% with psoas release. A concomitant pelvic 0.37, 95% CI 0.17-0.82; P = 0.014], having a small or
osteotomy was performed in 39.2% of hips at the time of the ischemic ossific center as compared with none (OR 0.24,
index surgery and a femoral osteotomy was performed in 95% CI 0.09-0.68; P = 0.007), and prior treatment (OR
32.8% of patients. Fifty-two hips (22.5%) had both a pelvic 0.39, 95% CI 0.15-0.99; P = 0.049). In contrast, older age at
and femoral osteotomy. After the index surgery, further the time of open reduction (OR 1.04, 95% CI 1.01-1.07;
corrective surgery was required in 30.6% of hips at some P = 0.003) was predictive of PFGD.
point during the follow-up period. The overall incidence of residual radiographic dys-
At final follow-up, the proportion of children with a plasia after open reduction was 55.3%. (Fig. 4B) A lower
final IHDI grade of 1, 2, 3, or 4 was 92.4%, 6.7%, 0.5%, rate of RAD (39.0%, n = 32/82) was seen in patients treated
and 0.5%. Re-dislocation during the follow-up period with a concomitant pelvic osteotomy at index surgery,
occurred in 7.0% of all hips (n = 16 hips; Fig. 2). The rather than an open reduction alone (Supplemental Table 3,
median time to re-dislocation was 100.5 days (interquartile Supplemental Digital Content 3, http://links.lww.com/BPO/
range 45.5 to 139 days), and most re-dislocations (n = 13 A591). Of note, patients who underwent a concomitant
hips, 81.3%) occurred within the first year after open pelvic osteotomy were more likely to have a prereduction
reduction. In the univariable analysis there was a lower IHDI grade of 4 (73.6%, n = 64/87) than those who did not
rate of post open-reduction immobilization among (65.0%, n = 76/117). Patient characteristics associated with
patients who experienced re-dislocation (75.0%; n = 6/16) RAD in this osteotomy subcohort included older age at
than who did not (94.3%, n = 200/212) (Supplementary baseline (P < 0.001), older age at the time of surgery
Table 1, Supplemental Digital Content 1, http://links.lww. (P < 0.001), and an eccentric ossific nucleus before surgery
com/BPO/A589; Fig. 3). There were no preoperative risk (P = 0.038). The rate of RAD was double in those not
Copyright © 2023 Wolters Kluwer Health, Inc. All rights reserved. www.pedorthopaedics.com | 281

FIGURE 3. Associated risk factors for the primary clinical outcomes after open hip reduction. DDH indicates developmental
dysplasia of the hip; OR, open reduction.
treated with a pelvic osteotomy at the time of open DISCUSSION

reduction (78.0%; n = 46/59; Supplemental Table 4, Open reduction of a dislocated hip is the next step
Supplemental Digital Content 4, http://links.lww.com/ in the treatment algorithm for infants with DDH who
on 04/15/2023
BPO/A592). Age at open reduction did not significantly either present late or have failed previous attempts at
differ between those who did and those who did not develop closed management. Although typically successful, this
RAD among those who did not have a pelvic osteotomy procedure may lead to several adverse outcomes in-
(P = 0.178). Furthermore, 35.6% (n = 21/59) of those cluding re-dislocation, PFGD, and RAD. Using data
without a pelvic osteotomy at the time of index surgery from a prospective, multicenter study group, the purpose
eventually required one during the study period. Finally, of this study was to report on general outcomes with
combining these 2 groups, the multivariable analysis specific attention to the rates and risk factors for each of
confirmed that a pelvic osteotomy was protective against these main complications after open reduction of an in-
RAD (OR 0.09, 95% CI 0.03-0.24; P < 0.001), whereas older fant’s dislocated hip. In this cohort of 232 hips, 92% of
age at the time of open reduction (OR 1.05, 95% CI 1.02- hips were IHDI grade 1 at final follow-up, indicating a
1.08; P = 0.001) increased the risk of dysplasia (Table 1). high rate of successful maintenance of reduction.
TABLE 1. Multivariate Analysis of Factors Associated With Re-dislocation, Proximal Femoral Growth Disturbance, and Residual
Dysplasia After Open Reduction for Developmental Dysplasia of the Hip
OR (95% CI)
Re-dislocation Proximal femoral growth disturbance Residual dysplasia
Older age at open reduction 0.99 (0.94-1.03) 1.04 (1.01-1.07)* 1.05 (1.02-1.08)*
Laterality
Right Reference Reference Reference
Left 0.52 (0.14-1.97) 1.38 (0.68-2.80) 1.34 (0.62-2.87)
Bilateral 0.48 (0.10-2.26) 0.37 (0.17-0.82)* 0.91 (0.41-2.02)
Clinical reducibility at any time before open reduction 1.11 (0.20-6.06) 0.49 (0.17-1.39) 0.65 (0.22-1.86)
IHDI Grade†
2 — Reference Reference
3 — 0.52 (0.09-3.08) 0.35 (0.05-2.24)
4 — 0.56 (0.09-3.40) 0.15 (0.02-1.03)
Ossific center
None Reference Reference Reference
Yes—normal 2.46 (0.18-33.16) 1.02 (0.33-3.11) 1.60 (0.50-5.14)
Yes—small or ischemic 2.68 (0.26-27.24) 0.24 (0.09-0.68)* 1.05 (0.37-2.97)
Previous treatment 0.79 (0.12-5.19) 0.39 (0.15-0.99) 1.55 (0.61-3.97)
Open reduction approach
Anterior Reference Reference Reference
Medial 1.04 (0.09-11.48) 2.34 (0.64-8.51) 0.72 (0.19-2.79)
Pelvic osteotomy — — 0.09 (0.03-0.24)*
Values <1 indicate possible protective factors. Values ˃1 indicate possible risk factors.
*Bolded values indicate statistically significant findings, with P < 0.05.
†IHDI Grade was unable to be included in the re-dislocation model, as there were no events of redislocation in hips with a preoperative IHDI Grade of 2.
OR indicates odds ratio.

FIGURE 4. Anteroposterior view of the left hip demonstrating a (A) 28-month-old male with proximal femoral growth disturbance
13 months after open reduction and (B) 33-month-old female with residual dysplasia (acetabular index = 34 degrees) 26 months
after open reduction without pelvic osteotomy.
on 04/15/2023
Obtaining and maintaining a stable and concentric reduction.22 It should be noted that our methodology was
joint is the primary goal of open hip reduction. In this a rigorous one that defined PFGD as a yes/no entity, but
study, the rate of re-dislocation after open reduction was did not distinguish clinically significant PFGD from other
7.0%, and the majority of dislocations occurred within the subtypes. As such, this likely represents the most stringent
first year after surgery. This is consistent with prior view of the incidence of this complication. However, it
reported rates of re-dislocation ranging from 3% to should also be noted that the incidence in the present study
12%.4,13–20 These rates include those from previous large is higher than the 25% rate found after closed reductions
studies by Pospschill and colleagues, Luhmann and which used the same methodology for determining
colleagues, and Wang and colleagues, with sample sizes of PFGD.4
78, 153, and 268 hips, respectively. On the basis of this Identifying risk factors for the future development
study—the largest to date using prospective data—our of PFGD has important prognostic value for clinicians
study suggests that the rate of re-dislocation likely falls in and families. In this cohort, the multivariable analysis
the middle of the range found from prior research. Only identified bilateral DDH, the presence of a small or is-
postoperative immobilization was associated with a lower chemic ossific nucleus (vs. no ossific nucleus), and pre-
rate of this adverse complication in the univariable anal- vious orthotic treatment or closed reduction as protective
ysis, and no variables were associated with a lower risk in against PFGD, whereas older age (at the time of surgery)
the multivariable analysis. Future research should aim was a risk factors for growth abnormalities. Previous
to further characterize the relationship between posttreatment may in part be a proxy for earlier access to
operative immobilization and re-dislocation. As the treatment, which could contribute to the protective
majority of these patients (60% in this cohort) went on to nature of this variable. Risk factors for the later devel-
have additional surgery, developing strategies to mitigate opment of PFGD in the univariable analysis included
this risk should remain one of the primary goals of older age at open reduction, and performance of an
ongoing investigations. adductor tenotomy, a pelvic osteotomy, or a femoral
Given the potential deleterious impact PFGD on shortening osteotomy; these operations are likely proxies
long-term patient outcomes and physical function,21 it for increasing disease severity (and thus age as well).
remains one of the most potentially significant complica- Previous research supports the association between older
tions after open hip reduction. The accuracy of previous age and PFGD.20 The association between the presence
reports on incidence has been limited by varying defi- of an ossific nucleus and PFGD is more controversial,
nitions and subjective interpretation of this complication. with many studies identifying a protective effect but
To reduce the risk of methodological bias inherent to several other studies, including a recent meta-analysis,
many previous single-center, retrospective series,22 this finding no protective effect.14,22,25,29,30 A better under-
study classified PFGD using a group consensus of 3 in- standing of these factors is necessary to improve our
dependent and blinded senior pediatric orthopaedic sur- abilities to predict and prevent PFGD after open
geons from 3 different institutions. Our findings showed a reduction for DDH.
rate of PFGD of 43.9%, which falls in the range of Pelvic osteotomies directly improve acetabular dys-
other studies that report rates of 9% to 58%,13,15,17–20,23–26 plasia, but there is limited data on how they affect sub-
using the 3 most commonly utilized criteria to assess sequent acetabular growth and remodeling throughout
PFGD,9,27,28 but is significantly higher than a recent follow-up. In this study, those treated without a pelvic
meta-analysis documenting a rate of 19% after open osteotomy at the time of index reduction had double the

risk of developing RAD (78.0% at 2 years) as those treated dependent on the data reported by individual centers.
with an osteotomy (39.0% at 1 year). These rates are much Because of the large number of surgeons contributing to
higher than previous reports, which more often have the database, there is considerable variation in treatment
shown rates of RAD around 20% with a follow-up period practices, but the authors view this as a strength in that it
of 1-2 years.20,31 However, these analyses did not use age- more accurately reflects the diversity of real-world prac-
adjusted acetabular index values for defining RAD and tice. To improve reliability of the data, all contributors to
often relied on rates of further corrective surgery, the in- the database were provided with a Standardization Man-
dications for which are largely subjective. Notably, among ual with definitions and descriptions of the variables and
patients without an osteotomy, the average age at open outcomes. In addition, the database has periodic reliability
reduction did not differ between those who did and did not studies performed using collected x-ray images to ensure
develop RAD. Yet, when looking at the full sample in the the accuracy and consistency of the data. In addition, we
multivariate analysis, the only risk factor identified for addressed quality for the PFGD variable by retro-
RAD was older age at time of surgery, whereas pelvic spectively characterizing imaging through blinded expert
osteotomy was protective against RAD. This was true consensus and establishing study group protocols that
even though patients who underwent pelvic osteotomy require the use of universal data collection forms and
were more likely to have a prereduction IHDI grade of 4. periodical data review to improve consistency.
Taken together, these findings suggest that there may be In conclusion, this prospective, multicenter cohort of
benefits to performing single stage hip reconstructions infants undergoing open reduction for DDH found high
(open reduction with pelvic osteotomy) at an earlier age, success rates in terms of femoral head reduction but higher
on 04/15/2023
although additional information is needed to characterize rates of PFGD and RAD then previous smaller, retro-
which patients would benefit most from this approach.32 If spective reports. Late diagnosis and treatment were asso-
appropriately indicated, performing concomitant pelvic ciated with an increased risk of PFGD and RAD, whereas
osteotomies could reduce the risk of future surgery, ex- the absence of an initial pelvic osteotomy also predicted
posure to anesthesia, and additional time in a spica cast.33 future dysplasia. Moving forward, surgeons can utilize this
However, more work is needed to further substantiate this information to improve the shared decision-making
finding and to assess the effects of different types of pelvic process with families and potentially consider performing
osteotomies on long term rates of RAD. pelvic osteotomies more often at the index procedure.
This study had a number of limitations. Although Further long-term follow-up of this cohort will allow us to
we utilized a large, prospectively collected data set, we ascertain more accurate long-term success and complica-
were nevertheless limited by the available length of the tion rates after open reduction for infantile DDH.
follow-up included within the database, and longer follow-
up certainly could provide a clearer picture of outcomes.
However, this study, which had a median follow-up of REFERENCES
1. Shipman SA, Helfand M, Moyer VA, et al. Screening for
21 months, was done to assess early outcomes after open developmental dysplasia of the hip: a systematic literature review
reduction. Previous research has shown that the majority for the US Preventive Services Task Force. Pediatrics. 2006;117:
of pelvic remodeling after open reduction occurs within e557–e576.
the first 2 years and that follow-up beyond two years is 2. Cooper AP, Doddabasappa SN, Mulpuri K. Evidence-based
unlikely to identify additional cases of redislocation or management of developmental dysplasia of the hip. Orthop Clin
North Am. 2014;45:341–354.
PFGD.10,12,34 This suggests that this study likely captures 3. Nelson SE, DeFrancesco CJ, Sankar WN. Operative reduction for
a large portion of these outcomes, but as part of this study developmental dysplasia of the hip: epidemiology over 16 years.
group’s ongoing efforts, future studies will further report J Pediatr Orthop. 2019;39:e272–e277.
on this cohort after longer-term follow-up. A second 4. Sankar WN, Gornitzky AL, Clarke NMP, et al. Closed reduction for
developmental dysplasia of the hip: early-term results from a
limitation was the small sample sizes for the RAD anal- prospective, multicenter cohort. J Pediatr Orthop. 2019;39:111–118.
yses. This was due to a requisite 2-year minimum follow- 5. Castañeda P, Masrouha KZ, Ruiz CV, et al. Outcomes following
up for anyone treated without an initial pelvic osteotomy, open reduction for late-presenting developmental dysplasia of the
as the authors felt that this best accounted for the bulk of hip. J Child Orthop. 2018;12:323–330.
acetabular remodeling that would occur after femoral 6. Subasi M, Arslan H, Cebesoy O, et al. Outcome in unilateral or
bilateral DDH treated with one-stage combined procedure. Clin
head relocation.12 Because of the small sample sizes here, Orthop. 2008;466:830–836.
we could not further stratify by type of pelvic osteotomy 7. Ning B, Yuan Y, Yao J, et al. Analyses of outcomes of one-stage
and results may vary by subgroup. In addition, indications operation for treatment of late-diagnosed developmental dislocation
for further corrective surgery during the follow-up period of the hip: 864 hips followed for 3.2 to 8.9 years. BMC Musculoskelet
Disord. 2014;15:401.
were at the discretion of the treating surgeon, and there- 8. Qiu M, Chen M, Sun H, et al. Avascular necrosis under different
fore subjective. We believed it was necessary to include treatment in children with developmental dysplasia of the hip: A
postoperative pelvic osteotomy in the RAD group, as network meta-analysis. J Pediatr Orthop B. 2022;31:319–326.
RAD is a common surgical indication for this procedure 9. Salter RB, Kostuik J, Dallas S. Avascular necrosis of the femoral
head as a complication of treatment for congenital dislocation of the
and using acetabular index alone would not fully capture hip in young children: a clinical and experimental investigation. Can
the outcome of interest. Finally, as is true with any pro- J Surg J Can Chir. 1969;12:44–61.
spective multicenter study, there is expected heterogeneity 10. Tiderius C, Jaramillo D, Connolly S, et al. Post-closed reduction
in treatment across centers and the quality of results is perfusion magnetic resonance imaging as a predictor of avascular

necrosis in developmental hip dysplasia: a preliminary report. 23. Gulman B, Tuncay IC, Dabak N, et al. Salter’s innominate
J Pediatr Orthop. 2009;29:14–20. osteotomy in the treatment of congenital hip dislocation: a long-
11. Novais EN, Pan Z, Autruong PT, et al. Normal Percentile Reference term review. J Pediatr Orthop. 1994;14:662–666.
Curves and Correlation of Acetabular Index and Acetabular Depth 24. Mardam-Bey TH, MacEwen GD. Congenital hip dislocation after
Ratio in Children. J Pediatr Orthop. 2018;38:163–169. walking age. J Pediatr Orthop. 1982;2:478–486.
12. Albinana J, Dolan LA, Spratt KF, et al. Acetabular dysplasia after 25. Schoenecker PL, Strecker WB. Congenital dislocation of the hip in
treatment for developmental dysplasia of the hip. Implications for children. Comparison of the effects of femoral shortening and of
secondary procedures. J Bone Joint Surg Br. 2004;86:876–886. skeletal traction in treatment. J Bone Joint Surg Am. 1984;66:21–27.
13. Pospischill R, Weninger J, Ganger R, et al. Does open reduction of 26. Clarke NMP, Jowett AJL, Parker L. The surgical treatment of
the developmental dislocated hip increase the risk of osteonecrosis? established congenital dislocation of the hip: results of surgery after
Clin Orthop. 2012;470:250–260. planned delayed intervention following the appearance of the capital
14. Luhmann SJ, Schoenecker PL, Anderson AM, et al. The prognostic femoral ossific nucleus. J Pediatr Orthop. 2005;25:434–439.
importance of the ossific nucleus in the treatment of congenital 27. Kalamchi A, MacEwen GD. Avascular necrosis following treatment
dysplasia of the hip. J Bone Joint Surg Am. 1998;80:1719–1727. of congenital dislocation of the hip. J Bone Joint Surg Am. 1980;62:
15. Wang TM, Wu KW, Shih SF, et al. Outcomes of open reduction for 876–888.
developmental dysplasia of the hip: does bilateral dysplasia have a 28. Bucholz RW, Ogden JA. Patterns of ischemic necrosis of the
poorer outcome? J Bone Joint Surg Am. 2013;95:1081–1086. proximal femur in nonoperatively treated congenital hip disease. The
16. Badawy MY, Afifi A, Shaker ES, et al. Is capsulorrhaphy a necessary Hip: Proceedings of the Sixth Open Scientific Meeting of the Hip
step during open reduction of developmental dysplasia of the hip? A Society. Mosby; 1978:43–63.
randomized controlled trial. Int Orthop. 2022;46:589–596. 29. Segal LS, Boal DK, Borthwick L, et al. Avascular necrosis after
17. Klisic P, Jankovic L. Combined procedure of open reduction and treatment of DDH: the protective influence of the ossific nucleus.
shortening of the femur in treatment of congenital dislocation of the J Pediatr Orthop. 1999;19:177–184.
hips in older children. Clin Orthop. 1976;119:60–69. 30. Apostolides M, Gowda SR, Roslee C, et al. The presence of the
on 04/15/2023
18. Barrett WP, Staheli LT, Chew DE. The effectiveness of the Salter ossific nucleus and its relation to avascular necrosis rates and the
innominate osteotomy in the treatment of congenital dislocation of number of secondary procedures in late-presenting developmental
the hip. J Bone Joint Surg Am. 1986;68:79–87. dysplasia of the hip. J Pediatr Orthop Part B. 2021;30:139–145.
19. Galpin RD, Roach JW, Wenger DR, et al. One-stage treatment of 31. Abousamra O, Deliberato D, Singh S, et al. Closed vs open reduction
congenital dislocation of the hip in older children, including femoral in developmental dysplasia of the hip: The short-term effect on
shortening. J Bone Joint Surg Am. 1989;71:734–741. acetabular remodeling. J Clin Orthop Trauma. 2020;11:213–216.
20. Tennant SJ, Hashemi-Nejad A, Calder P, et al. Bilateral devel- 32. Salter RB. The classic. Innominate osteotomy in the treatment of
opmental dysplasia of the hip: does closed reduction have a role in congenital dislocation and subluxation of the hip by Robert B. Salter,
management? outcome of closed and open reduction in 92 hips. J. Bone Joint Surg. (Brit) 43B:3:518, 1961. Clin Orthop. 1978;178:2–14.
J Pediatr Orthop. 2019;39:e264–e271. 33. Yilar S, Topal M, Zencirli K, et al. Comparison of total cost and
21. Marks A, Cortina-Borja M, Maor D, et al. Patient-reported outcomes between single-stage open reduction and Pemberton
outcomes in young adults with osteonecrosis secondary to devel- periacetabular osteotomy operation and two separate consecutive
opmental dysplasia of the hip - a longitudinal and cross-sectional operations in treatment of bilateral developmental hip dysplasia in
evaluation. BMC Musculoskelet Disord. 2021;22:42. children at walking age. J Pediatr Orthop Part B. 2020;29:256–260.
22. Novais EN, Hill MK, Carry PM, et al. Is age or surgical approach 34. Buchanan JR, Greer RB, Cotler JM. Management strategy for
associated with osteonecrosis in patients with developmental dysplasia prevention of avascular necrosis during treatment of congenital
of the hip? A meta-analysis. Clin Orthop. 2016;474:1166–1177. dislocation of the hip. J Bone Joint Surg Am. 1981;63:140–146.

ORIGINAL ARTICLE
Age Influence Upon Glenohumeral Remodeling After

Shoulder Axial Rebalancing Surgery in Brachial Plexus
OOIXlWaXUZnczw/iMdshREa5aLNbRoI8k4/i3nEue4+M9SuoRpOlsks8q++lgMau5as6ohRU05i1xIfaj6mOhBR5kiITxc4+W0=
Birth Injury
Malo Le Hanneur, MD, MSc,*† Lilia Brahim, MD,* Tristan Langlais, MD, MSc,*
Pierre-Alban Bouché, MD, PhD,* and Franck Fitoussi, MD, PhD*
[CI = (−0.04; −0.01), P = 0.002], 0.12% [CI = (−0.21; −0.04),

Background: Shoulder rebalancing procedures have been proven P = 0.0076], and 0.01 grade [CI = (−0.02; −0.01), P = 0.0078] per ad-
to provide satisfactory functional improvements in patients with ditional month of age at the time of surgery, respectively. The threshold
shoulder external rotation (ER) deficit due to brachial plexus of 5 years was identified as the age at the time of surgery after which
birth injury (BPBI). However, the influence of age at the time of significant remodeling no longer occurred. No significant postoperative
on 04/15/2023
surgery on osteoarticular remodeling remains uncertain. The changes were observed in patients without glenohumeral dysplasia on
purposes of this retrospective case series were (1) to assess the age preoperative magnetic resonance imaging.
impact on glenohumeral remodeling and (2) to determine an age Conclusion: In the setting of BPBI-related glenohumeral dys-
limit after which significant changes can no longer be expected. plasia, the younger the surgical axial rebalancing of the shoulder,
Methods: We reviewed preoperative and postoperative magnetic the greater the glenohumeral remodeling seems to be. Such
resonance imaging data of 49 children with BPBI who underwent procedure seems to be safe in patients without significant joint
a tendon transfer to reanimate active shoulder ER, with (n = 41) deformity on preoperative imaging.
or without (n = 8) concomitant anterior shoulder release to re- Level of Evidence: Therapeutic—Level IV.
store passive shoulder ER, at a mean age of 72 ± 40 months
(19;172). Mean radiographic follow-up was 35 ± 20 months Key Words: glenohumeral remodeling, brachial plexus, birth
(12;95). Univariate linear regressions were used to assess the in- palsy, shoulder, tendon transfer, glenohumeral dysplasia
fluence of age at the time of surgery upon changes of glenoid (J Pediatr Orthop 2023;43:e389–e395)
version, glenoid shape, percentage of the humeral head anterior
to the glenoid midline, and glenohumeral deformity. Beta
coefficients with 95% CI were calculated.
Results: Improvements of glenoid version, glenoid shape, percentage of
the humeral head anterior and glenohumeral deformity significantly
decreased by 0.19 degrees [CI = (−0.31; −0.06), P = 0.0046], 0.02 grade
B rachial plexus birth injury (BPBI) is still a rather
common condition whose recovery can vary sub-
stantially between patients.1 At the shoulder, external ro-
tators are commonly more weakened than internal
From the *Department of Pediatric Orthopedics and Reconstruction, rotators, inducing a lack of active shoulder external ro-
Armand Trousseau Hospital—Sorbonne University, Paris; and tation (ER).2 In addition, impaired growth of paralyzed
†Hand to Shoulder Mediterranean Center, ELSAN, Clinique Bou- internal rotator muscles can cause capsulo-ligamentous
chard, France. contractures on the joint anterior aspect, leading to pas-
The review was conducted by an independent observer according to the sive shoulder ER limitations.3 Subsequently, as the ante-
ethical standards of the 1964 Declaration of Helsinki and to the
Methodology of Reference MR-003; as data were obtained by re- rior aspect of the humeral head is relatively fixed on the
viewing charts, no Institutional Review Board (IRB) consultation was posterior aspect of the glenoid process, glenohumeral
required, as stated in the Official Journal of the French Republic*. dysplasia progressively appears.4 The advent of magnetic
The study was registered in the National Committee of Computer resonance imaging (MRI) systems has allowed to describe
Science and Liberties register (CNIL – No 2223039 version 0), and
legal guardians were individually informed and consented to data
and quantify these deformities, to follow their natural
collection and analysis. history over time, and to assess the impact of various
The authors received no financial support. therapeutic strategies.4–6
The authors declare no conflicts of interest. To axially rebalance the glenohumeral joint, the
Reprints: Malo Le Hanneur, MD, MSc, Department of Pediatric
Orthopedics, Armand Trousseau Hospital—Sorbonne University, 26
most widely adopted surgical approach consists in rean-
avenue du Dr Arnold Netter, 75012 Paris, France. E-mail: malo. imating active shoulder ER, using nerve and/or tendon
lehanneur@gmail.com. transfers, and restoring passive shoulder ER by releasing
Supplemental Digital Content is available for this article. Direct URL the joint anterior aspect.5,7–9 Numerous release techniques
citations appear in the printed text and are provided in the HTML have been described, most commonly involving the sub-
pedorthopaedics.com. scapularis muscle, other internal rotators (e.g., pectoralis
Copyright © 2023 Wolters Kluwer Health, Inc. All rights reserved. major), and capsulo-ligamentous elements anterior to the
DOI: 10.1097/BPO.0000000000002380 joint. Regarding the muscles used as donors to reanimate
J Pediatr Orthop Volume 43, Number 5, May/June 2023 www.pedorthopaedics.com | e389

Le Hanneur et al J Pediatr Orthop Volume 43, Number 5, May/June 2023
TABLE 1. Cohort’s Characteristics*

Demographics
No. patients (N) 49
Sex (F/M, N) 28 / 21
Birth characteristics
Term (wk) 39 ± 2 (31;41)

Weight (g) 3844 ± 775 (1100;5400)

Delivery (V/C, N) 48 / 1
Palsy characteristics
Affected side (R/L, N) 30 / 19
Narakas’ grade (I/II/III/IV, N) 15 / 5 / 17 / 12
Received treatments in infancy
Nerve surgeries (N) 27
Nerve grafts 25
Nerve transfers 9
Other surgeries (N) 7
Wrist extension tendon transfer 4
Elbow flexion tendon transfer 1
Forearm derotational osteotomy 2
*Results are presented as Mean ± SD (Range), unless otherwise stated.
C indicates cesarean; F, female; g, grams; L, left; M, male; N, number of
patients/cases in absolute values; R, right; V, vaginal.
on 04/15/2023
shoulder release to restore passive shoulder ER. Patients

with additional procedures that could modify the shoulder
axial balance were excluded from the study, and patients
with <12 months of radiographic follow-up.
Cohort Characteristics
Among the 69 patients who underwent a tendon
transfer to reanimate shoulder ER within the inclusion
FIGURE 1. Flow chart of the cohort. BPBI indicates brachial period, 20 were excluded (Fig. 1). Forty-nine patient’s
plexus birth injury; ER: external rotation. charts were analyzed after an average radiographic follow-
up of 35 ± 20 months (12;95) (Table 1), including 41 who
active shoulder ER, latissimus dorsi (LD) and lower tra- underwent a tendon transfer with a concomitant release
pezius (LT) are the most transferred muscles to the rotator and 8 who received an isolated tendon transfer, at a mean
cuff posterior aspect (i.e., infraspinatus muscle). The ef- age of 72 ± 40 months (19;172) at the time of surgery
fectiveness of such strategy has been demonstrated over (Table 2).
the years, both in terms of clinical and radiologic
outcomes.5–13 However, the influence of age at the time of Preoperative Care and Surgical Indications
surgery on postoperative osteoarticular remodeling re- The type of palsy was clinically determined in the
mains unclear, especially regarding the age limit after perinatal period according to Narakas’ classification.14
which significant changes can no longer be expected. Depending on the spontaneous recovery that was observed
The objectives of this study were, based on the during follow-up, reconstructive nerve surgeries were
comparisons of preoperative and postoperative MRI elected in some patients, within the first 6 months of life
measurements, to assess the impact of patient’s age at the (i.e., no recovery of active elbow flexion—brachial plexus
time of surgery on postoperative glenohumeral remodel- exploration and reconstruction using nerve grafts and/or
ing, and to identify a cutoff age after which such remod- transfers) and/or between 6 and 18 months of life (i.e.,
eling would no longer be significant. On the basis of our incomplete recovery—elective nerve transfers).
clinical experience and the existing literature,5,8,9,11,13 our
hypothesis was that younger children would have the
TABLE 2. Surgical Data*
greatest postoperative joint modifications.
Age at surgery (mo) 72 ± 40 (19;172)
Tendon transfer (N) 49
METHODS Latissimus dorsi 43
We conducted a retrospective chart review of all Lower trapezius 6
Shoulder anterior release (N) 41
patients who underwent surgical management of active Open 24
shoulder ER deficit due to BPBI from January 2012 to Arthroscopic 17
January 2020 at our institution. The inclusion criterion
was having received a tendon transfer to reanimate active N indicates number of cases in absolute values.
shoulder ER, with or without concomitant anterior
e390 | www.pedorthopaedics.com Copyright © 2023 Wolters Kluwer Health, Inc. All rights reserved.

J Pediatr Orthop Volume 43, Number 5, May/June 2023 Age Influence on GH Remodeling in BPBI
A tendon transfer to the infraspinatus muscle was

elected in patients with <0 degree of active shoulder ER
with the arm in adduction (ER1) after 18 months of life. LD
was our first choice of donor. LT was selected in case of LD
weakness (i.e., grade-3 strength or less according to the
British Medical Council Research (BMRC) grading system

in patients older than 3; grade-2 strength or less according

to the modified BMRC scale in younger patients 15). In
addition, an anterior shoulder release was performed con-
comitantly in patients with <30 degrees of passive shoulder
ER1, in an open manner or arthroscopically depending on
the preoperative clinical assessment.10,16
Surgical treatment was undertaken at the age at
which the patients were referred to us. Preoperatively, a
shoulder MRI scan was systematically performed to
evaluate the glenohumeral joint morphology.
Surgical Techniques and Postoperative Care

All patients were operated on by the senior author.
on 04/15/2023
Tendon transfers were performed in an open manner,

without graft augmentation. A single axillary incision was
used for LD transfers, whereas 2 posterior incisions were
made for LT transfers, according to previously described
techniques.7,17,18 When indicated, a release of the joint
anterior aspect was performed concomitantly, involving
the subscapularis muscle (i.e., released from the sub-
scapular fossa but without tenotomy), the rotator interval
including the coracohumeral ligament, the upper and
middle glenohumeral ligaments, and the anterior gleno-
humeral capsule; it was considered satisfactory when at
least 45 degrees of passive shoulder ER1 was obtained
intraoperatively.
Postoperatively, the patient was immobilized in an
upper extremity spica cast for 6 weeks, with the shoulder
positioned in abduction and ER. Once the cast was re-
moved, physiotherapy was resumed to maintain passive
ER and strengthen the transferred muscle. Systematic
shoulder MRI scans were performed 12 months after the
tendon transfer and then repeated yearly.
Data Collection
The review was conducted by an independent sur-
geon experienced in BPBI, according to the ethical
standards of the 1964 Declaration of Helsinki and the
Methodology of Reference MR-003; as data were ob-
tained by reviewing charts, no Institutional Review Board
(IRB) consultation was required.19 The study was regis- FIGURE 2. A and B, Satisfactory remodeling of the right gle-
tered in the National Committee of Computer Science and nohumeral joint in a 38 months-old boy at the time of surgery.
Liberties register (CNIL – No 2223039 version 0), and Preoperative (A) axial MR image (proton density-fast spin echo
legal guardians were individually informed and consented sequence) showing a grade-V glenohumeral deformity, with
to data collection and analysis. severe flattening and posterior dislocation of the humeral head
Patients’ demographics, medical histories and surgical (black arrow) and moderate pseudoglenoid dysplasia (i.e.,
information were collected, as well as preoperative and grade-VI glenoid shape) (white arrow). Fifty-nine months after
a latissimus dorsi transfer to the infraspinatus (white arrow-
postoperative (i.e., at last follow-up) MRI data. Glenoid
heads) and an arthroscopic anterior glenohumeral release, a
morphology was assessed using the glenoid version, as de- grade-I residual glenohumeral deformity was observed on
fined by Kozin et al,5 and the glenoid shape, as classified by postoperative (B) magnetic resonance images (proton density-
Pearl et al. in seven grades of increasing severity.6 Humeral fast spin echo sequence), with a centered and spheric humeral
head position was assessed using the percentage of the head and a concentric glenoid process (i.e., grade-I glenoid
humeral head anterior (PHHA) to the glenoid midline.5 shape).4,6
Copyright © 2023 Wolters Kluwer Health, Inc. All rights reserved. www.pedorthopaedics.com | e391

TABLE 3. Preoperative and postoperative MRI outcomes*

RESULTS
In a large majority of cases, glenohumeral dysplasia was
Preoperative Postoperative p
improved by the surgery (Fig. 2a and Fig. 2b), with significant
Glenoid version (deg) −24 ± 16 (−62;3) −11 ± 16 (−68;24) 0.0001 improvements observed between mean preoperative and
Glenoid shape (Grade)† 3.5 ± 2.2 (1;7) 2.2 ± 1.8 (1;7) 0.0016 postoperative values of all MRI criteria (Table 3).
PHHA (%) 31 ± 15 (0;55) 39 ± 13 (0;63) 0.0016

In 3 cases, glenohumeral dysplasia worsened after
Glenohumeral deformity 3.3 ± 1.6 (1;6) 2.1 ± 1.5 (1;7) 0.0002

(Grade) ‡ surgery. A 9-year-old girl with a grade-III glenoid process
developed a moderate pseudoglenoid postoperatively. A
†According to Pearl et al.6 6-year-old boy with a preoperative grade-VI glenoid
‡According to Waters et al.4 process developed a severe pseudoglenoid postoperatively,
MRI indicates magnetic resonance imaging; PHHA, percentage of humeral with a glenoid version decrease of 24 degrees. Lastly, a
head anterior to the midline of the glenoid.
growth arrest of the humerus proximal aspect (i.e., grade-
VII glenohumeral deformity) was noted in a 6.5-year-old
Finally, glenohumeral deformity was classified in 7 grades girl who presented a severe flattening of the humeral head
of increasing severity, according to Waters et al.4 MR im- on preoperative images.
ages were analyzed using the PACS software (Picture Ar- Nine patients who had normal preoperative MRI
chiving and Communication Systems – Carestream Health, scans (i.e., grade-I glenohumeral deformity4) did not de-
Rochester, NY). teriorate their joints postoperatively.
Influence of Age on Glenohumeral Remodeling
on 04/15/2023
Regarding linear regression analyses, improvements

Statistical Analysis between preoperative and postoperative values of glenoid
Continuous quantitative variables were described by version, glenoid shape grade, PHHA and glenohumeral
mean, SD, and range. Dichotomous variables were de- deformity grade significantly decreased by 0.19 degrees
scribed by their number of events. Wilcoxon tests were used [CI = (−0.31;−0.06), P = 0.0046], 0.02 grade [CI = (−0.04;
to compare continuous variables. To assess whether the age −0.01), P = 0.002], 0.12% [CI = (−0.21;−0.04], P = 0.0076],
at the time of surgery affected joint remodeling, univariate and 0.01 grade [CI = (−0.02;−0.01), P = 0.0078) per addi-
linear regressions were used and Beta coefficients with 95% tional month of age at the time of surgery, respectively.
CIs were calculated. For all linear regressions, the nor-
mality assumption and the homoskedasticity of the data Age Limit for Significant Remodeling
were checked. The threshold of significance retained was On the basis of the graphical reading of the re-
5% for a power of 80% and a risk of the first species at 5%. gression curves, the threshold of 5 years was identified to
All tests were 2-sided. The R software (version 3.5.0) was be the age at the time of surgery after which differences
used to perform statistical analyzes. between preoperative and postoperative MRI values
FIGURE 3. Influence of the age at surgery upon modifications of the glenohumeral deformity grades (abscissa: age at surgery in
months; ordinate: differences between preoperative and postoperative grades of glenohumeral deformity). As age increases,
glenohumeral deformity changes decrease, with a break in the regression curve (blue line) observed at the age of 5 years (dotted
red line).4

This 5-years age limit seems consistent with the ex-

TABLE 4. Comparisons of MRI Improvements Between
Patients Operated Before and After 5 Years of Age* isting literature. In 2008, Waters and colleagues studied
the effects of latissimus dorsi and teres major (LD+TM)
< 5 years old ≥ 5 years old
tendon transfers and joint open reduction in a cohort of 23
(N = 23) (N = 26) p
patients of a mean age of 2.3 years.13 After a mean ra-
Follow-up (mo) 40 ± 24 (12;95) 31 ± 15 (13;60) 0.2176 diographic follow-up of 25 months, the average glenoid
Glenoid version 20 ± 19 (−3;62) 6 ± 15 (−24;42) 0.0089

(deg)
version, PHHA and glenohumeral deformity grade sig-
Glenoid shape −2.3 ± 2.3 (−5;1) −0.4 ± 1.4 (−4;3) 0.0087 nificantly improved. The same team reported on the re-
(Grade) † sults of LD+TM tendon transfers associated with
PHHA (%) 15 ± 14 (−5;42) 3 ± 10 (−12;26) 0.0017 musculotendinous lengthening on the anterior aspect of
Glenohumeral −1.9 ± 1.6 (-5;0) −0.5 ± 0.9 (−2;2) 0.0033 the shoulder in a cohort of older children (i.e., mean age of
deformity
(Grade) ‡ 3.5 years).9 The authors noted that glenoid version and
PHHA improved, but that the improvements were less
*Results are presented as Mean ± SD (Range).
†According to Pearl et al.6
important than in their younger patients.9,13 Similarly,
‡According to Waters et al.4 Kozin and colleagues published two studies in 2010, re-
MRI indicates magnetic resonance imaging; N, number of patients; PHHA, porting on axial rebalancing procedures.5,11 In the first
percentage of the humeral head anterior to the midline of the glenoid.
one, 44 children underwent glenohumeral joint arthro-
scopic release, including 16 of them who underwent a
would decrease (Fig. 3; SDC 1-3, Supplemental Digital concomitant LD+TM tendon transfer to the infra-
on 04/15/2023
Content 1, http://links.lww.com/BPO/A593, Supplemental spinatus, with an average age of 2.7 years at the time of
Digital Content 2, http://links.lww.com/BPO/A594, surgery.11 After 1 year of follow-up, significant glenoid
Supplemental Digital Content 3, http://links.lww.com/ version and PHHA improvements were observed. In their
BPO/A595). Subsequently, we conducted an age-based second study, the authors reported on 24 children who
subgroup analysis of the differences between preoperative underwent LD+TM tendon transfers at an average age of
and postoperative MRI values, and we observed that 5 ± 1.8 years, without any significant improvement re-
those differences were significantly higher in younger garding glenoid version, PHHA and/or glenohumeral
patients, for all MRI criteria (Table 4). Furthermore, deformity.5
those improvements were significant in the younger In their cohort of 109 BPBI-patients treated with
subgroup whereas they were not significant in older TM tendon transfers to the infraspinatus and sub-
patients, for all MRI criteria (Table 5). scapularis release, El-Gammal et al8 found a highly sig-
nificant positive correlation between the age at surgery
and the residual glenoid retroversion. Significantly higher
DISCUSSION residual glenoid retroversion was observed in patients
In this study, we reviewed the MRI data of 49 patients operated after 4 years of age, and significantly higher re-
who underwent a tendon transfer to reanimate active sidual humeral head subluxation was observed in patients
shoulder ER in the setting of BPBI. We highlighted the in- operated on after 2 years of age. The differences regarding
fluence of age at the time of surgery upon glenohumeral the remodeling capacities and the cutoff ages between
remodeling capacities, with significant decreases of the their study and the present series may be explained by the
differences between preoperative and postoperative values fact that El-Gammal and colleagues used computerized
observed in all MRI criteria as the patients’ age at the time of tomographic scans to assess glenohumeral remodeling,
surgery increased. In addition, the age limit of 5 years which seem to be less accurate in evaluating osteoarticular
was identified as the age at the time of surgery after which remodeling in children.3,4 Nonetheless, the authors’ con-
significant remodeling no longer occurred. clusions are in the same line as ours, namely that the
TABLE 5. Age-based Subgroup Analysis of MRI Improvements*

Patients operated before 5 years of age Preoperative Postoperative p
Glenoid version (deg) −32 ± 15 (−62;−3) −13 ± 15 (−42;19) 0.0002
Glenoid shape (Grade)† 4.6 ± 1.9 (1;7) 2.3 ± 1.8 (1;6) 0.0003
PHHA (%) 24 ± 13 (0;52) 38 ± 14 (0;63) 0.0004
Glenohumeral deformity (Grade)‡ 3.9 ± 1.4 (1;6) 1.9 ± 1.1 (1;4) < 0.0001
Patients operated at or after 5 years of age
Glenoid version (deg) −17 ± 14 (−44;3) −10 ± 18 (−68;24) 0.0509
Glenoid shape (Grade)† 2.6 ± 1.9 (1;6) 2.1 ± 1.8 (1;7) 0.3194
PHHA (%) 37 ± 13 (0;55) 41 ± 13 (0;63) 0.1930
Glenohumeral deformity (Grade)‡ 2.7 ± 1.6 (1;5) 2.2 ± 1.7 (1.7) 0.1189
*Results are presented as Mean ± SD (Range).
†According to Pearl et al.6
‡According to Waters et al.4
MRI indicates magnetic resonance imaging; PHHA, percentage of the humeral head anterior to the midline of the glenoid.

earlier the rebalancing surgery is performed, the better the between the 2 groups was not significant (Table 4).
remodeling will be. Finally, we did not investigate the functional
We observed 3 aggravations of the preexisting gle- improvements that such procedures can provide,
nohumeral dysplasia. These 3 patients were over 5 years of considering that was not the purpose of the current
age at the time of surgery, which seems to be consistent study and that such data have been extensively reported
with the existing literature,5,9 even though such compli- by previous authors.5–9 Accordingly, no conclusions can
cations have already been reported in younger patients.13 be drawn from this study regarding the influence of the age
Such observations stress out the fact that in patients with at surgery upon clinical outcomes after these procedures.
limited potentials of osteoarticular remodeling and/or
advanced glenohumeral deformations, especially in cases CONCLUSIONS
of severe flattening of the humeral head, alternative ex- In patients with glenohumeral dysplasia secondary
traarticular surgical options should be considered, such as to BPBI, axial rebalancing procedures seem to allow for
humeral external derotational or glenoid anteversion significant remodeling of the glenohumeral joint, provided
osteotomies.20–22 Considering the 9 patients without pre- that the age at the time of surgery is <5 years. As the
operative glenohumeral dysplasia, no significant remod- analysis was based solely on MRI criteria and not clinical
eling was observed on postoperative imaging. Such criteria (e.g., shoulder ranges of motion, functional
findings seem to be consistent with the available scores), our results do not allow us to comment on the
literature,12 and we believe they are important to highlight clinical implications of such management. However, early
as they support the harmlessness of these rebalancing and close MRI monitoring seems to us to be essential in
on 04/15/2023
surgeries on intact joints. patients with BPBI-related active shoulder ER deficit; in

The results of this study should be considered in light case of dysplasia, axial rebalancing surgery should be
of its limitations, including the relatively small sample size performed as early as possible to achieve the best possible
which prevented us from highlighting the potential impact remodeling.
of other factors on remodeling. We conducted a multi-
variate analysis that included sex (Male/Female), release REFERENCES
(Yes/No), type of transfer (LD/LT), preoperative gleno- 1. Hale HB, Bae DS, Waters PM. Current concepts in the management
humeral deformity according to Waters et al., in addition of brachial plexus birth palsy. J Hand Surg Am. 2010;35:322–331.
to the age at surgery (months); with the exception of the 2. Waters PM, Monica JT, Earp BE, et al. Correlation of radiographic
latter, no variable was responsible for significant changes muscle cross-sectional area with glenohumeral deformity in children
with brachial plexus birth palsy. J Bone Joint Surg Am. 2009;91:
between preoperative and postoperative MRI data. Such 2367–2375.
lack of significant differences regarding these potential 3. Nikolaou S, Peterson E, Kim A, et al. Impaired growth of denervated
confounding variables may be due to type II statistical muscle contributes to contracture formation following neonatal
errors, related to the above-mentioned small sample size. brachial plexus injury. J Bone Joint Surg Am. 2011;93:461–470.
However, our cohort seems consistent with previously 4. Waters PM, Smith GR, Jaramillo D. Glenohumeral deformity
secondary to brachial plexus birth palsy. J Bone Joint Surg Am.
published series on this relatively rare subject, and the fact 1998;80:668–677.
that we could observe significant differences related to the 5. Kozin SH, Chafetz RS, Shaffer A, et al. Magnetic resonance imaging
age at surgery despite such a small sample highlight the and clinical findings before and after tendon transfers about the
importance of this variable for glenohumeral remodeling. shoulder in children with residual brachial plexus birth palsy: a 3-
year follow-up study. J Pediatr Orthop. 2010;30:154–160.
Another limit of the study regarding patients’ selection 6. Pearl ML, Edgerton BW, Kon DS, et al. Comparison of arthroscopic
may lie in the fact that we included patients in whom findings with magnetic resonance imaging and arthrography in
different muscles, with subsequently different lines of ac- children with glenohumeral deformities secondary to brachial plexus
tion, were used as donors, and different release procedures birth palsy. J Bone Joint Surg Am. 2003;85:890–898.
7. Allard R, Fitoussi F, Azarpira MR, et al. Shoulder internal rotation
were performed (i.e., open or arthroscopic). When taking
contracture in brachial plexus birth injury: proximal or distal
into consideration the flourishing literature on shoulder subscapularis release? J Shoulder Elbow Surg. 2021;30:1117–1127.
rebalancing procedures in patients with BPBI, with nu- 8. El-Gammal TA, Saleh WR, El-Sayed A, et al. Tendon transfer
merous surgical teams performing different procedures around the shoulder in obstetric brachial plexus paralysis: clinical
and reporting relatively similar outcomes,5–13 we believe and computed tomographic study. J Pediatr Orthop. 2006;26:
641–646.
that such selection bias may be insignificant. An evalua- 9. Waters PM, Bae DS. Effect of tendon transfers and extra-articular
tion bias may be highlighted, considering that radio- soft-tissue balancing on glenohumeral development in brachial
graphic measurements were conducted by a single plexus birth palsy. J Bone Joint Surg Am. 2005;87:320–325.
observer who was not a board-certified radiologist; 10. Kany J, Kumar HA, Amaravathi RS, et al. A subscapularis-
nonetheless, the interobserver reproducibility of the used preserving arthroscopic release of capsule in the treatment of internal
rotation contracture of shoulder in Erb’s palsy (SPARC procedure).
measurements have been shown to be satisfactory.4,23 In J Pediatr Orthop B. 2012;21:469–473.
addition, patients had not reached bone maturity at the 11. Kozin SH, Boardman MJ, Chafetz RS, et al. Arthroscopic treatment of
time of evaluation which may limit our conclusions as internal rotation contracture and glenohumeral dysplasia in children with
remodeling could still have occurred. However, significant brachial plexus birth palsy. J Shoulder Elbow Surg. 2010;19:102–110.
12. Pearl ML, Edgerton BW, Kazimiroff PA, et al. Arthroscopic release
changes between pre- and postoperative values were out- and latissimus dorsi transfer for shoulder internal rotation contrac-
lined for all MRI criteria despite this limited period of tures and glenohumeral deformity secondary to brachial plexus birth
follow-up (Table 3), and the difference of follow-up palsy. J Bone Joint Surg Am. 2006;88:564–574.

13. Waters PM, Bae DS. The early effects of tendon transfers and open 19. Journal Officiel de la République Française n°0189 du 14 août 2016.
capsulorrhaphy on glenohumeral deformity in brachial plexus birth Texte N°77. legifrance.gouv.fr. Accessed January 1, 2022. Available
palsy. J Bone Joint Surg Am. 2008;90:2171–2179. at: https://www.legifrance.gouv.fr/affichTexte.do?cidTexte=JORFTE
14. Narakas AO. [Injuries of the brachial plexus and neighboring XT000033028290&dateTexte=&categorieLien=id
peripheral nerves in vertebral fractures and other trauma of the 20. Abzug JM, Chafetz RS, Gaughan JP, et al. Shoulder function after
cervical spine]. Orthopade. 1987;16:81–86; German. medial approach and derotational humeral osteotomy in patients
15. Gilbert A, Tassin JL. [Surgical repair of the brachial plexus in with brachial plexus birth palsy. J Pediatr Orthop. 2010;30:469–474.
obstetric paralysis]. Chirurgie. 1984;110:70–75; French. 21. Dodwell E, O’Callaghan J, Anthony A, et al. Combined glenoid
16. Carlioz H, Brahimi L. [Place of internal disinsertion of the anteversion osteotomy and tendon transfers for brachial plexus birth
subscapularis muscle in the treatment of obstetric paralysis of the palsy: Early outcomes. J Bone Joint Surg Am. 2012;94:2145–2152.
upper limb in children]. Ann Chir Infant. 1971;12:159–167; French. 22. Waters PM, Bae DS. The effect of derotational humeral osteotomy
17. Elhassan B, Bishop AT, Hartzler RU, et al. Tendon transfer options on global shoulder function in brachial plexus birth palsy. J Bone
about the shoulder in patients with brachial plexus injury. J Bone Joint Surg Am. 2006;88:1035–1042.
Joint Surg Am. 2012;94:1391–1398. 23. van der Sluijs JA, van der Meij M, Verbeke J, et al. Measuring
18. Pagnotta A, Haerle M, Gilbert A. Long-term results on abduction and secondary deformities of the shoulder in children with obstetric
external rotation of the shoulder after latissimus dorsi transfer for brachial plexus lesion: reliability of three methods. J Pediatr Orthop
sequelae of obstetric palsy. Clin Orthop Relat Res. 2004;426:199–205. B. 2003;12:211–214.
on 04/15/2023

ORIGINAL ARTICLE
Clinical Characteristics and Treatment Patterns of Open

Hand Fractures in the Pediatric Population
OOIXlWaXUZnczw/iMdshREa5aLNbRoI8k4/i3nEue4+M9SuoRpOlsks8q++lgMau5as6ohRU05i1wneHACYVztkiv2UtI1CQT0=
Raphael H. Parrado, MD,* Lukas Foster, MD,* Megan Gilbert, CPNP-AC,*

Nellie Movtchan, MD,† Lois Sayrs, PhD,* Emily Khoury, RN, MNA,* Wassim Ballan, MD,‡
and Timothy Schaub, MD§
Key Words: antibiotic treatment, children, open hand fracture,

Background: Open hand fractures are one of the most common postoperative infection
injuries in the pediatric population. These injuries are at higher
risk of infection, especially in cases of frank contamination. (J Pediatr Orthop 2023;43:e358–e362)
Several studies on adult hand fractures are available in the lit-
erature; however, pediatric open hand fractures have yet to be
INTRODUCTION
on 04/15/2023
extensively studied. This study aimed to define pediatric open

hand fracture’s demographics, clinical characteristics, and
treatment patterns. H and fractures are among the most common pediatric
injuries encountered in the United States, with pha-
lanx fractures most frequent in the 10 to 19 years old age
Methods: Using the Protected Health Information database,
pediatric patients (< 18 y old) with the diagnosis of open hand group.1,2 When considering all fractures, the proximal
fracture from June 2016 to June 2018 were extracted. Demo- phalanx and the fifth digit are the most affected sites.3
graphic, treatment, and follow-up data were collected. Clinical In toddlers, distal phalangeal fractures are the most
outcomes included readmission and postoperative infection rates. common. As age increases, the distribution becomes more
Results: There were a total of 4516 patients who met the in- proximal fractures into the metacarpals and the wrist due
clusion criteria; the median age was 7 years (interquartile range: to increasing body weight and sporting activities.4 Most
3 to 11); 60% males; 60% white. Displaced fractures occurred in pediatric hand fractures are closed, and up to 90% do not
74% of patients, with the right hand (52%) and middle finger require any operative interventions.5 However, open hand
(27%) predominance. The most common mechanism of injury fractures pose a special consideration as there is a risk or
was a crushing injury in-between objects (56%). Associated nerve degree of contamination. Differences in vascular supply
injury occurred in 78 patients (4%) and vascular injury in 43 and soft tissue make them more prone to infection and
patients (2%). Open reduction and internal fixation were per- often necessitate surgical intervention.6
formed in 30% of patients. Cephalosporins were the most com- In terms of infection, there is extensive research on
monly prescribed antibiotics (73%), followed by aminopenicillins the adult population, but data are scarce on the pediatric
(7%). Nine patients had complications related to surgical inter- population. As a result, often adult recommendations are
vention (0.2%), and postoperative infection occurred in 44 pa- applied to pediatric patients. Adult open hand fracture
tients (1%). infection rates are ~5%1, with reduced postoperative in-
Conclusions: Pediatric open hand fractures most often occur fection correlated with rapid administration of antibiotics
during childhood and more frequently in males. These fractures upon admission.7 In the adult population, multiple factors
tend to be more distal and displaced; reduction and fixation are have been shown to decrease open hand fracture infection
required in one-third of the cases. Despite the absence of treat- rates, -including the amount of irrigation used for washout
ment guidelines and variability, this injury exhibits low in the emergency department and the oral antibiotic type.8
complication rates. Recent findings also show that open hand fractures have
Level of Evidence: Level III, retrospective study. lower infection rates than lower extremity open fractures,
believed to be secondary to an increased blood supply to
the hand in the pediatric population.9 Current infection
treatment protocols recommend debridement/admin-
From the *Division of Trauma Surgery, Department of Surgery; ‡Di- istration of antibiotics before operative intervention.
vision of Infectious Diseases, Department of Pediatrics; §Division of
Plastic Surgery, Department of Surgery, Phoenix Children’s Hospital;
However, to date, there have been no studies to evaluate
and †Department of Plastic Surgery, Mayo Clinic, Phoenix, AZ. the efficacy of a perioperative antibiotic protocol, and
None of the authors received financial support for this study. there is considerable empirical variation in clinical
The authors declare no conflicts of interest. practice.10
Reprints: Raphael H. Parrado, MD, Department of Plastic Surgery, 1919 Pediatric hand fractures differ from adults in
E Thomas Road, Phoenix, AZ 85016. E-mail: raphaelpar393@gmail.
com. the physiological characteristics of the hand, environ-
Copyright © 2023 Wolters Kluwer Health, Inc. All rights reserved. mental conditions influencing treatment, and mechanism
DOI: 10.1097/BPO.0000000000002379 of injury.
e358 | www.pedorthopaedics.com J Pediatr Orthop Volume 43, Number 5, May/June 2023

J Pediatr Orthop Volume 43, Number 5, May/June 2023 Open Hand Fractures in Children
Although hand fractures are considered a high risk The most common race was white, with 60%. Crush
for infection in children, research for pediatric hand injury between two objects was the most common mech-
fracture infection rates and associated clinical outcomes anism of injury (n = 2186, 56%), followed by being struck
has not yet been studied. by an object (n = 550, 14%) and a fall injury (n = 243,
The ambiguity in open hand fracture wound char- 6.2%). The mechanism of injury was not identified in 191
acterization and management, especially in the pediatric (4.5%) cases (Table 1). Sports was the most common
population, has led to variation in practice and a void in activity reported as associated with an injury: basketball
specific guidelines for infection prevention. One recent accounted for 57 cases (18.4%), football, 54 cases (17.5%),
study aimed to fill this void by proposing a new classi- and baseball, 44 (14.2%) (Table 2).
fication scheme for managing open hand fractures in
adults. The classification is based on fracture location Injury Location and Characteristics
within the hand, the integrity of soft tissue remaining There were 2271 (52%) patients, in which the right
available for coverage, the contamination level, and the hand was involved. The middle finger was the most involved
affected area’s vascular integrity. The application of this digit (29.2%). When the metacarpals were involved, the fifth
classification system has resulted in promising clinical was the most commonly injured (28.7%). Distal phalanx
outcomes in adult infection prevention but has not yet open fractures comprised 88% of all cases (Table 3).
been evaluated in the pediatric population.11,12 This study Displaced fractures occurred in 3364 of the patients (74%).
aimed to define pediatric open hand fracture’s demo- Nerve injury occurred in 78 cases (1.7%), with the radial
graphics, clinical characteristics, and treatment patterns. nerve being most common in 24 cases (31%). Forty-three
(0.9%) fractures had an associated vascular injury, with the
on 04/15/2023
METHODS ulnar artery being most involved in 7 cases (16%). Partial

amputation occurred in 98 patients (2.2%) and complete
Data Source amputation occurred in 21 patients (0.5%). Regarding
Data for the present study were collected from the surgical treatment, reduction, and internal fixation were
Protected Health Information database. The Protected performed in 1355 patients (30%). One hundred ninety-eight
Health Information database is a comprehensive admin- (4.4%) patients received a surgical digit amputation and 3
istrative database managed by the Children’s Hospital As- received a surgical hand amputation (0.07%).
sociation that collects data from 49 Children’s hospitals and
includes demographic and clinical information.13 Pediatric Return Visits, Complications, and Length of Stay
patients (< 18 y old) with an ICD10 diagnosis of open hand In this study, the median hospital days for a patient
fracture (including digits, metacarpal bones, and wrist) who with an open hand fracture was 1 day.1,2 Return emergency
presented from June 2016 to June 2018 were extracted. Ex- room visit rates at 15 and 30 days were 3% (141) and 4.3%
tracted data on injury demographics included: patient age, (195), respectively. The most common chief complaint for
race, and mechanism of injury. In addition, using the diag- repeat visits was hand pain (19.4%). The average length of
nosis codes, the following variables were extracted: digit in- stay for emergency room revisits was <1 day.
volved, fracture displacement, presence of nerve injury,
presence of arterial injury, and partial/complete amputation. TABLE 1. Mechanism of Injury for Open Hand Fractures
CPT codes for fixation, debridement, nerve, and vascular Mechanism of injury Count (N = 4516); n (%)
repair were extracted for treatment. Outcome-related varia-
bles extracted included any antibiotic received, postoperative Crush 2,186 (56)
Struck by an object 550 (14)
infection presence, hospital days, readmission at 15 days, Fall 243 (6.2)
readmission at 30 days, and reason for readmission. Sharp object 230 (5.9)
Bite 74 (1.9)
Statistical Analyses Accidental discharge from gun 68 (1.7)
Descriptive statistics (mean and median) were used Accidental hit by another person 66 (1.7)
Bicycle accident 52 (1.3)
to compare pediatric patients across demographic and ATV accident 49 (1.3)
injury characteristics. The Kruskal-Wallis test was used to Assault with firearm/fight 43 (1.1)
assess statistical significance for continuous covariates and Struck by patient 33 (0.8)
χ2 or Fisher exact test for categorical variables. All data MVC 27 (0.7)
were analyzed using R version 3.3.1 (R core Team 2015, Motorcycle accident 26 (0.7)
Boarding a car 22 (0.6)
Vienna, Austria). Statistical significance was set at P <0.05 Fireworks 22 (0.6)
for all statistical tests. Auto vs pedestrian 16 (0.4)
Overexertion of hand 14 (0.4)
Explosion 4 (0.1)
RESULTS Self-harm 1 (< 0.1)
SNAT 1 (< 0.1)
Demographics and Mechanism Unspecified 191 (4.9)
A total of 4516 patients were included based on our
ATV, All Terrain Vehicle MVC, Motor Vehicle Collision SNAT, Suspected
criteria. There were 2712 (60%) male patients; the median Non-Accidental Trauma
age of the patients was 7 years.

Parrado et al J Pediatr Orthop Volume 43, Number 5, May/June 2023
TABLE 2. Activity During Injury for Open Hand Fractures TABLE 4. Risk for Infection in Children With Open Hand
Most common activities Count (N = 317); n (%) Fractures
YES (n = 30); No (n = 2500);
Basketball 57 (18)
American football 54 (17) n (%) n (%) P
Baseball 44 (14) Age; median (IQR) 7 (1-10) 7 (3-11) 0.34

Bike riding 40 (13) Gender (M) 23 (76) 1550 (62) 0.09

Free weights 29 (9.4) Any comorbidity (yes) 2 (6) 31 (1.2) < 0.05
Bowling 23 (7.4) Displaced fracture (yes) 20 (66) 1946 (78) 0.09
Walking, hiking 19 (6.1) Multiple fractures (yes) 4 (13) 285 (11) 0.74
Running 18 (5.8) Vascular injury (yes) 0 41 (1) 0.47
Roller skating 13 (4.2) Nerve injury (yes) 0 69 (3) 0.35
Wrestling 12 (3.9) LOS; mean (SD) 8.6 (30.7) 1.3 (2.6) < 0.05
Amputation (yes) 2 (7) 95 (4) 0.41
Readmission in 30 days 2 (7) 119 (5) 0.62
Infection Characteristics and Treatment (yes)
There was a total of 44 patients who developed a IQR indicates interquartile range; LOS, length of stay.
hand infection, making the overall infection rate ~1%. Of
44 patients with infections, 35 (0.8%) developed cellulitis,
mainly in the middle finger (n = 12 or 34%). Abscess vs 1.3%, P < 0.05), but there was no difference in ampu-
requiring drainage occurred in 9 patients (0.2%) and was tation (7% vs 4%, P = 0.4) or readmission rates (7% vs
most commonly on the index finger (n = 3 or 33%). For
on 04/15/2023
5%, P = 0.6) in all children (Table 4).

children with any comorbidity (cardiac, respiratory, neu- Antibiotic treatment was documented in 2623 of the
rological, hematologic, or neurological), infection rates 4516 patients (58%). Cephalosporins were the most com-
were significantly higher than in children without mon antibiotic class administered, comprising 1921 (73%)
comorbidities (6% vs 1.2%, P < 0.05). Patients with documented cases. A first-generation cephalosporin was
postoperative infection had increased hospital days (8.6% used in 1910 (99.4%) cases and third-generation cepha-
losporins comprised the other 11 (0.6%) cases. Other an-
tibiotic classes used included: aminopenicillins (177, 6.7%)
TABLE 3. Open Hand Fracture Location Distribution by and clindamycin (170, 6.5%) (Table 5). Cephalosporins
Number of Cases were more commonly used in displaced fractures. Mean
Open Hand Fracture Location No. cases; n (%) hospital days and postoperative infection rates were both
First finger 553 (13) lower in the cephalosporin group than in other antibiotics
Distal phalanx 480 (11) used (1.2 vs 2 d, P < 0.05 and 0.3% vs 4%, P < 0.05,
Proximal phalanx 49 (1.0) respectively) (Supplemental Data, Table 6). There was
Second finger 731 (17) no significant difference in readmit rates between patients
Distal phalanx 615 (14)
Middle phalanx 39 (0.9)
where cephalosporins were utilized (5%) and patients
Proximal phalanx 48 (1.1) where other antibiotics (5%) were used to manage
Third finger 1220 (28) infection (P = 1).
Proximal phalanx 42 (1.0) DISCUSSION
Fourth finger 994 (23) The principal aims of this study were to characterize
Distal phalanx 901 (21) the demographic, treatment characteristics, and short-
Proximal phalanx 40 (0.9)
term outcomes for pediatric open hand fractures. We ex-
Fifth finger 684 (16) amined 4516 injuries in 2 years using an administrative
Proximal phalanx 53 (1.2) TABLE 5. Antibiotic Treatment of Open Hand Fractures in
Metacarpals 157 (3.6) Children
First 30 (0.7) Antibiotic class Count (N = 2631); n (%)
Second 23 (0.5)
Third 31 (0.7) Cephalosporin 1921 (73)
Fourth 28 (0.6) Aminopenicillin 177 (6.7)
Fifth 45 (1.0) Clindamycin 170 (6.5)
Carpals 31 (0.7) Bacitracin 159 (6.1)
Scaphoid 9 (0.2) Combination topical 93 (3.5)
Lunate 0 Aminoglycoside 69 (2.6)
Triquetrum 2 (< 0.1) Ureidopenicillins 18 (0.7)
Pisiform 1 (< 0.1) Folate inhibitors 6 (0.2)
Hamate 3 (< 0.1) Fluoroquinolones 4 (0.2)
Capitate 4 (< 0.1) Vancomycin 3 (0.1)
Trapezoid 0 Penicillin 2 (0.1)
Trapezium 3 (< 0.1) Albendazole 1 (< 0.1)

J Pediatr Orthop Volume 43, Number 5, May/June 2023 Open Hand Fractures in Children
results have specific implications for injury prevention in

TABLE 6. Cephalosporin use and Outcomes in Open Hand
Fractures very young children. There could be improvements to be
performed in parent education primarily followed
Yes (n = 1921) No (n = 609) P
by potential changes in chair design once specific
Age; median (IQR) 6 (3-11) 7 (3-12) 0.1 mechanisms have been elucidated.
Any comorbidity (yes) 24 (1.2) 9 (1.5) 0.7 Antibiotics are considered a standard of care for
Displaced fracture (yes) 1511 (78) 117 (19) < 0.05

Multiple fractures (yes) 209 (11) 80 (13) 0.12 open pediatric fractures, and their timely administration
Vascular injury (yes) 31 (1.6) 10 (1.6) 1 has been demonstrated to reduce infection rates7. In this
Nerve injury (yes) 53 (2.8) 16 (2.6) 1 study, 73% of patients with documented antibiotic treat-
LOS; mean (SD) 1.2 (2.4) 2 (7.4) < 0.05 ment were given cephalosporin; however, we have no data
Postoperative infection (yes) 6 (0.3) 24 (4) < 0.05
Amputation (yes) 75 (4) 22 (4) 0.1
regarding timing.
Readmission in 30 days 92 (5) 29 (5) 1 Contamination for open hand fractures in children is
(yes) often associated with a gram-positive environmental
IQR indicates interquartile range; LOS, length of stay.
source, which may help explain why cephalosporins have
demonstrated exceptional coverage in open pediatric
fractures.19 Aminopenicillins were the next most
database. Of the patients, 60% in this study were males, commonly prescribed antibiotic, likely reflecting added
which is consistent with the 60% to 73% reported in the anaerobic and Clostridium as a source of frank
literature concerning all pediatric hand fractures.14,15 In contamination.20 Clindamycin was the third most popular
contrast, Kreutz-Rodrigues et al14 report a mean age of
on 04/15/2023
choice, providing gram-positive (including Methicillin-

any pediatric hand fractures to be 12.2 years in 4356 resistant Staphylococcus aureus) and anaerobic coverage;
fractures over 27 years. In this study, the median patient it is also a popular alternative when a patient is allergic to
age of pediatric open hand fracture presentation was cephalosporins or penicillins.20 The addition of an ami-
7 years, with 67.2% of injuries occurring at 8 or younger. noglycoside to a cephalosporin is recommended for Gus-
Although Young et al16 reported a similar mean age of tilo-Anderson class II and III open fractures, likely
12.1 years in 303 open and closed pediatric hand fractures explaining why they were used in 69 cases in this study.
in 283 patients in a single institution, the authors also Unfortunately, we do not have a way to classify the
reported that when open hand fractures were isolated, fractures comparatively, and the addition of an amino-
over half of the cases involved patients younger than glycoside may not be indicated for these particular
7 years old. fractures.11,20
Our findings mirror those of Young and colleagues, A 2005 retrospective study of 554 pediatric open
suggesting pediatric open hand fractures occur, on aver- fractures (including but not limited to open hand frac-
age, at a younger age than pediatric patients with closed- tures) reported an overall infection rate of 3%20. This
hand fractures. study of 4516 patients from 2016 to 2018 suggests an
Another notable finding of this study was that most overall infection rate for open hand fractures of 1%, sug-
open pediatric hand fractures (88%) occurred at the gesting that open hand fracture rates are lower than gen-
distal phalanges. In contrast, closed pediatric hand eral open orthopaedic fractures. Despite the overall rate
fractures most commonly occur in the proximal pha- for open hand fractures being lower than the rate for all
langes and metacarpals.14–17 This difference may be at- open fractures, age and mechanism of injury may alter
tributable to tuft fractures accounting for up to 80% of that risk. In the absence of specific antibiotic treatment
hand fractures in toddlers and preschool-age children.18 guidelines for open pediatric hand fractures, age, and
This fact also supports the younger median patient age of mechanism should also be considered.
injury found in this study. Regarding the mechanism of
injury, it has been documented that crush injuries are the
most common cause of hand fractures in toddlers, Limitations
whereas sports-related injuries are the most common in The retrospective nature of this study is an inherent
adolescence. Our study of 4516 patients found that a limitation. In addition, we based our data collection on
crushing injury caused 56% of the open pediatric hand diagnosis and procedure coding, which might be variable
fractures, and only 7.9% were sustained during sports- and not wholly reflective of the injury. We had no imaging
related activity. We also found that 30% of cases utilized or other data to classify or determine the severity and to
open reduction and internal fixation, compared with a classify the injury. As discussed, we also did not have in-
recent study reporting ~5% of pediatric hand fractures formation regarding patient allergies or comorbidities that
needing surgical fixation.16 Our results for open hand would have changed the antibiotic choice. The database
fractures compared with all hand fractures suggest that also does not provide specific dates and geographical lo-
open pediatric hand fractures are likely sustained at a cations of the events, for which it would be difficult to
higher energy level than closed fractures. It is interesting provide a geographical or seasonal pattern for these in-
to note that crush injuries were most often associated juries. Finally, unfortunately, we have no data on long-
with car doors. Because younger pediatric patients are term complications as well as the functional impact of
more likely to sustain an open hand fracture, these these injuries.

Parrado et al J Pediatr Orthop Volume 43, Number 5, May/June 2023
CONCLUSION 7. Patzakis MJ, Wilkins J. Factors influencing infection rate in open

Pediatric open hand fractures occur at a median age fracture wounds. Clin Orthop Relat Res. 1989;243:36–40.
8. Ketonis C, Dwyer J, Ilyas AM. Timing of debridement and infection
of 7 and are slightly more predominant in males. These rates in open fractures of the hand: a systematic review. Hand (N Y).
injuries are most seen in the distal phalanges. A crushing 2017;12:119–126.
mechanism most frequently causes open pediatric hand 9. Naranje SM, Erali RA, Warner WC Jr, et al. Epidemiology of
fractures. Concerning the treatment of open pediatric pediatric fractures presenting to emergency departments in the
hand fractures, open reduction, and internal fixation is United States. J Pediatr Orthop. 2016;36:e45–e48.
10. Basat NB, Allon R, Nagmi A, et al. Treatment of open fractures of the
needed in approximately one-third of the cases. Vascular hand in the emergency department. Eur J Orthop Surg Traumatol.
or nerve repairs are rare; first-generation cephalosporins 2017;27:415–419.
are the most used antibiotic. The overall infection rate is 11. Gustilo RB, Anderson JT. Prevention of infection in the treatment of
<1%. We are currently working on a more specific and one thousand and twenty-five open fractures of long bones:
retrospective and prospective analyses. J Bone Joint Surg Am. 1976;58:
detailed assessment of this fracture in terms of the injury 453–458.
itself, timing of antibiotics, and futural outcomes. We 12. Shah SS, Rochette LM, Smith GA. Epidemiology of pediatric hand
encourage future prospective studies after the treatment of injuries presenting to United States emergency departments, 1990 to
pediatric open hand fractures as these are common and 2009. J Trauma Acute Care Surg. 2012;72:1688–1694.
potentially functionally limiting injuries. 13. Narus SP, Srivastava R, Gouripeddi R, et al. Federating clinical data
from six pediatric hospitals: process and initial results from the PHIS
+ Consortium. AMIA Annu Symp Proc. 2011;2011:994–1003.
REFERENCES 14. Kreutz-Rodrigues L, Gibreel W, Moran SL, et al. Frequency,
1. Warrender WJ, Lucasti CJ, Chapman TR, et al. Antibiotic Pattern, and Treatment of Hand Fractures in Children and
on 04/15/2023
management and operative debridement in open fractures of the Adolescents: A 27-Year Review of 4356 Pediatric Hand Fractures.
hand and upper extremity: a systematic review. Hand Clin. 2018;34: Hand (N Y). 2020;17:92–97.
9–16. 15. Liu EH, Alqahtani S, Alsaaran RN, et al. A prospective study of
2. Zalavras CG. Prevention of infection in open fractures. Infect Dis pediatric hand fractures and review of the literature. Pediatr Emerg
Clin North Am. 2017;31:339–352. Care. 2014;30:299–304.
3. Chew EM, Chong AK. Hand fractures in children: epidemiology and 16. Young K, Greenwood A, MacQuillan A, et al. Paediatric hand
misdiagnosis in a tertiary referral hospital. J Hand Surg Am. 2012;37: fractures. J Hand Surg Eur Vol. 2013;38:898–902.
1684–1688. 17. Sivit AP, Dupont EP, Sivit CJ. Pediatric hand injuries: essentials you
4. Liao JCY, Chong AKS. Pediatric hand and wrist fractures. Clin need to know. Emerg Radiol. 2014;21:197–206.
Plast Surg. 2019;46:425–436. 18. Nellans KW, Chung KC. Pediatric hand fractures. Hand Clin.
5. Hartley RL, Lam J, Kinlin C, et al. Surgical and nonsurgical 2013;29:569–578.
pediatric hand fractures: a cohort study. Plast Reconstr Surg Glob 19. Trionfo A, Cavanaugh PK, Herman MJ. Pediatric Open Fractures.
Open. 2020;8:e2703. Orthop Clin North Am. 2016;47:565–578.
6. Tulipan JE, Ilyas AM. Open fractures of the hand: review of 20. Skaggs DL, Friend L, Alman B, et al. The effect of surgical delay on
pathogenesis and introduction of a new classification system. Orthop acute infection following 554 open fractures in children. J Bone Joint
Clin North Am. 2016;47:245–251. Surg Am. 2005;87:8–12.

ORIGINAL ARTICLE
Comparison of Different Surgical Techniques in Correction

of Congenital Vertical Talus Deformity: A Systematic
OOIXlWaXUZnczw/iMdshREa5aLNbRoI8k4/i3nEue4+M9SuoRpOlsks8q++lgMau5as6ohRU05i184jGgQK4fdlkjixHt2Jp1Q=
Review and Meta-Analysis of the Literature

Jason L. Cummings, BS, Asdrubal E. Rivera, MD, Abhishek Tippabhatla, BS,
and Pooya Hosseinzadeh, MD
Approach (7.81). The Dobbs Method resulted in the largest

Background: Congenital Vertical Talus (CVT) is a rare form of ankle arc of motion.
congenital rigid flatfoot. Numerous surgical techniques have Conclusion: We found the lowest radiographic recurrence and
been developed over the years in an attempt to definitively cor- reoperation rates in the Single-Stage Dorsal Approach cohort,
rect this deformity. We performed a systematic review and meta- while the highest rate of radiographic recurrence was seen in
on 04/15/2023
analysis of the existing literature to compare the outcomes of those treated with the Direct Medial Approach. The Dobbs
children with CVT treated with different methods. Method results in higher clinical scores and ankle arc of motion.
Methods: A detailed systematic search was conducted in ac- Future long-term studies focusing on patient-reported outcomes
cordance with PRISMA guidelines. Radiographic recurrence of are needed.
the deformity, reoperation rate, ankle arc of motion, and clinical Level of Evidence: Level III.
scoring was compared between the following 5 methods: Two-
Stage Coleman-Stelling Technique, Direct Medial Approach, Key Words: congenital vertical talus, congenital foot, dobbs
Single-Stage Dorsal (Seimon) Approach, Cincinnati Incision, method, two-stage correction, single-stage correction, minimally
and Dobbs Method. Meta-analyses of proportions were per- invasive correction, treatment outcomes, meta-analysis
formed, and data were pooled through a random effects model (J Pediatr Orthop 2023;43:317–325)
using the DerSimonian and Laird approach. Heterogeneity was
assessed using I^2 statistics. The authors used a modified version
of the Adelaar scoring system to assess clinical outcomes. An
alpha of 0.05 was used for all statistical analysis.
Results: Thirty-one studies (580 feet) met the inclusion criteria.
The reported incidence of radiographic recurrence of talona-
C ongenital Vertical Talus (CVT) is a rare form of
congenital rigid flatfoot with an incidence of roughly
1 in 10,000 births.1,2 This condition is defined by irredu-
vicular subluxation was 19.3%, with 7.8% requiring reoperation. cible dorsolateral dislocation of the navicular on the talar
Radiographic recurrence of the deformity was highest in the head and neck.3
children treated with the direct medial approach (29.3%) and Numerous surgical techniques have been developed
lowest in the Single-Stage Dorsal Approach cohort (11%) over the years in an attempt to definitively correct this
(P < 0.05). The reoperation rate was significantly lower in the deformity.4 In 1970, Coleman, Stelling, and Jarrett de-
Single-Stage Dorsal Approach cohort (2%) compared with all scribed a 2-staged approach for treating CVT.3 In 1987,
other methods (P < 0.05). There was no significant difference in Seimon described a single-staged approach involving a
the reoperation rates between the other methods. The highest single straight longitudinal incision over the mid-dorsum
clinical score was seen in the Dobbs Method cohort (8.36), of the foot.5 Most recently, Dobbs and colleagues de-
followed by the group treated with the Single-Stage Dorsal scribed a minimally invasive approach for treating CVT.6
None of these techniques come without complica-
tions, and varying degrees of success have been reported
From the Department of Orthopedic Surgery, Washington University in
St. Louis, Saint Louis, MO.
among different studies. Short-term complications include
Investigation performed at Washington University School of Medicine- wound necrosis, talar avascular necrosis (AVN), and both
St. Louis, MO undercorrection and overcorrection of the deformity.7–9
The authors declare no conflicts of interest. To the best of our knowledge, there has not yet been a
Reprints: Pooya Hosseinzadeh, MD, Associate Professor, Department of
Orthopaedic Surgery, 660 S. Euclid, Campus Box 8233, St. Louis,
systematic literature review and meta-analysis performed
MO 63110, Washington University School of Medicine. E-mail: that compares the outcomes of each of the treatment
hosseinzadehp@wustl.edu. methods listed above. This manuscript thus seeks to con-
Supplemental Digital Content is available for this article. Direct URL tribute to the current body of CVT literature by providing
citations appear in the printed text and are provided in the HTML the first systematic literature review on the topic. We hy-
pedorthopaedics.com. pothesize that the Dobbs Method will result in a lower
Copyright © 2023 Wolters Kluwer Health, Inc. All rights reserved. recurrence and reoperation rate than the other methods
DOI: 10.1097/BPO.0000000000002369 studied.

Cummings et al J Pediatr Orthop Volume 43, Number 5, May/June 2023
METHODS Given the variations in reporting across the different

studies, the authors decided to use a modified version of
Search Strategy the Adelaar scoring system for evaluating clinical out-
A detailed systematic search was conducted in comes. When individual Adelaar scores were given, these
PubMed/MEDLINE (1946-), Cumulative Index of Nurs- were the scores used for calculating clinical outcomes.
ing and Allied Health Literature (CINAHL, 1937-), However, in cases where the authors did not provide nu-
EMBASE (1947-), SCOPUS (1823-), Web of Science merical scores but rather reported outcomes as “ex-
(WOS, 1900-), The Cochrane Library, and Clinicaltrials. cellent,” “good,” “fair,” or “poor,” scores of 10, 8, 5, and
gov. The search was executed using the standardized in- 3 were assigned, respectively. An alpha level of 0.05 was
dexing terms and keywords: Foot deformities, congenital, used for all analyses.
foot deformities, congenital vertical talus, acquired, talus,
treatment outcome, treatment, clubfoot, and clubfeet. The RESULTS
final search results were limited to the English language
In the initial literature search, 13,948 articles were
using the database-supplied limits.
identified from the 7 databases: PubMed (5371), EM-
Study Selection BASE (3604), CINAHL (2638), SCOPUS (988), Web of
Science (896), clinicaltrials.gov (311), and Cochrane Li-
Title and Abstract Screening brary (140). In the first stages of the screening, 13,736
The title and abstract of the articles related to articles were excluded through screening the title and ab-
Congenital Vertical Talus Treatments and Outcomes were stract based on inclusion and exclusion criteria, and we
on 04/15/2023
reviewed and extracted by the investigators based on the included 212 articles for further screening. At the second
inclusion and exclusion criteria of this study. Inclusion stage of the screening process, an additional 172 articles
criteria included studies that mention “congenital vertical were excluded. After both stages of screening, 31 studies
talus;” “vertical talus,” “congenital convex pes valgus,” or were included in our final literature review (Fig. 1). Fifteen
“rocker-bottom foot.” Exclusion criteria included articles (50%) of these studies were of high quality based on the
that lack the keywords noted in the inclusion criteria or MINORS criteria (Table 1).
that focus on “oblique talus;” “calcaneovalgus;” “tarsal A total of 580 feet were included in this systematic
coalition;” “paralytic pes valgus;” “pes planovalgus;” review. There were 4 studies that included the Coleman-
“congenital talipes equinovarus;” or “clubfoot.” Stelling technique (22 patients, 34 feet); 8 studies (109
patients, 157 feet) that included the Direct Medial Ap-
Full-Text Screening proach, 5 studies (57 patients, 91 feet) that included the
Once selected for full-text screening, the selected Single-Stage Dorsal Approach, 5 studies (58 patients, 79
articles then underwent a second round of screening by the feet) that included the Cincinnati incision and 9 studies
same 2 investigators based on the inclusion and exclusion (145 patients, 219 feet) that included the Dobbs Method.
criteria of this study. Inclusion criteria included patients A pooling meta-analysis of proportions revealed a Tau^2
under 18 years of age, a minimum follow-up time of of 0.019 and I^2 of 59.42% when comparing reoperation
1 year, and studies reporting outcomes of congenital ver- rates across the studies and a Tau ^2 of 0.035 and I^2 of
tical talus treatment. Exclusion criteria included articles 73.49% when comparing rates of recurrence of talona-
that included the aforementioned key words in the ab- vicular subluxation. The forest plots for these comparisons
stract but were not primarily focused on the treatments of are shown in Supplemental Fig. 2, Digital Content 1,
interest and/or outcomes. Discrepancies in inclusion status http://links.lww.com/BPO/A585 and Supplemental Fig. 3,
were discussed between 2 reviewers until a consensus was Digital Content 2, http://links.lww.com/BPO/A586, re-
reached. A quality assessment was then performed on all spectively.
studies included in the final analysis using the Method- The reported incidence of radiographic recurrence of
ological Index for Non-randomized Studies (MINORS), talonavicular subluxation was 19.3% among all patients
similar to previous meta-analyses on pediatric foot included in this study. There was a total of 45 (7.8%) cases
deformities.10,11 of talonavicular subluxation that required a repeat oper-
ation. Only 2 cases of talar avascular necrosis were re-
Data Extraction and Statistical Analysis ported, both of which occurred in patients treated with the
Two investigators extracted data based on the search Direct Medial Approach. The results of each individual
criteria. After tabulating all relevant data, statistical treatment cohort are listed below.
analysis was performed on IBM SPSS Statistics Version
28.0 (IBM Corporation; Armonk, NY, USA) and RStu- The Coleman-Stelling Technique
dio 2022.02.0. χ2 analysis was used to compare recurrence In 1970, Coleman, Stelling, and Jarrett described a
and reoperation rates among the different treatment co- 2-staged approach for treating CVT.3 The first stage
horts. Meta-analyses of proportions were performed, and consists of an obliquely placed incision centered over the
data were pooled through a random effects model using sinus tarsi that is used for the lengthening of the extensor
the DerSimonian and Laird approach. Heterogeneity was digitorum longus, extensor hallucis longus, and anterior
assessed using I^2 statistics, as is routine for meta-analyses tibial tendons. In addition, a complete talonavicular and
of orthopaedic literature.10,12 calcaneocuboid capsulotomy is done. A long leg cast is

J Pediatr Orthop Volume 43, Number 5, May/June 2023 Systematic Review on Congenital Vertical Talus
on 04/15/2023
FIGURE 1. PRISMA flowchart.
then applied for 6 to 8 weeks with the knee bent and the Single-Stage Dorsal (Seimon) Approach
foot and ankle in the equinus position. Following cast In 1987, Leonard P. Seimon described a single-
removal, the second stage is then performed, which con- staged approach involving a single straight incision over
sists of heel cord lengthening, posterior capsulotomy, and the dorsum of the foot. This is followed by extensive soft-
advancement of the posterior tibial tendon to the plantar tissue release and K-wire fixation of the reduced talona-
surface of the navicular. vicular joint. The Achilles contracture is then corrected by
Our search yielded a total of 4 studies (22 patients, percutaneous lengthening or open lengthening plus pos-
34 feet) meeting our inclusion criteria in which the Cole- terior capsulotomy.5
man-Stelling technique was used.3,13–15 The results are There was a total of 5 studies (57 patients, 91 feet)
displayed in Table 2. Overall, no cases of talar AVN were that used the Single-stage Dorsal Approach and met our
reported in this cohort, with a 21.1% reoperation rate in inclusion criteria.5,24–27 The results are shown in Table 4.
those treated with this technique. Radiographic recurrence of the deformity was noted in
11.0% of the patients, with 2% requiring reoperation. No
cases of talar AVN were reported.
The Direct Medial Approach and its Variants
In 1978, Ogata et al39 described a 1-stage open me-
dial reduction of the talonavicular joint, with imbrication
of the joint capsule combined with posterior ankle and The Cincinnati Approach
subtalar joint capsulotomy. In 1982, Kumar and col- In 1982, Crawford et al41 described a transverse
leagues described a similar approach that involved 3 in- incision that extends from the anteromedial to the an-
cisions. The first incision is made on the lateral side of the terolateral aspect of the foot over the back of the ankle
foot, centered over the sinus tarsi. A second incision is at the level of the tibiotalar joint. Raab and Krauspe
made on the medial side of the foot, over the talus. A third were the first to utilize the Cincinnati incision
incision is made on the medial side of the Achilles tendon on CVT patients in 1997, and this technique was fur-
to be used for Z-lengthening.40 ther popularized after a 1999 paper by Kodros and
There were 8 studies (157 feet) in which the direct Dias.28,29
medial approach was used, and our inclusion criteria were There was a total of 5 studies (58 patients, 79 feet) in
met.16–23 The results are displayed in Table 3. which the Cincinnati incision was used.7,8,28–30 The results
Radiographic recurrence of the deformity was noted in are shown in Table 5. No cases of talar AVN were
29.3% of these patients, with 9.1% requiring reoperation. reported, with 16.1% of feet requiring reoperation for
There were 2 cases (1%) of talar AVN seen in this cohort. recurrence of the deformity.

on 04/15/2023
Cummings et al
320 | www.pedorthopaedics.com
TABLE 1. Quality Assessment of Included Studies

6. Follow-up
period
2. Inclusion of 4. Endpoint 5. Unbiased appropriate to 7. Loss to 8. Prospective
1. Stated consecutive 3. Prospective appropriate to evaluation of the major follow-up not calculation of Total Study
References Technique aim patients data collection the study aim endpoints endpoint exceeding 5% sample size score* quality†
Coleman et al3 Coleman- 2 1 2 2 1 1 1 0 10 Low

Stelling
Zimbler and Craig13 Coleman- 2 0 2 2 1 2 1 0 10 Low
Stelling
Jacobsen and Coleman- 0 2 2 2 1 1 1 0 9 Low
Crawford14 Stelling
Walker et al15 Coleman- 2 2 2 2 2 2 2 0 14 High
Stelling
Zhu et al16 Direct Medial 2 0 2 2 2 2 2 0 12 High
Sanzarello et al17 Direct Medial 2 2 2 2 1 2 2 0 13 High
Ramanoudjame Direct Medial 2 2 2 2 2 2 2 0 14 High
et al18
Ayadi et al19 Direct Medial 2 1 2 2 1 2 1 0 11 Low
Schwering 20 Direct Medial 2 2 2 2 1 2 1 0 12 High
Wirth et al21 Direct Medial 2 0 2 2 1 1 1 0 9 Low
Oppenheim et al22 Direct Medial 2 0 2 2 1 2 0 0 9 Low
Clark et al23 Direct Medial 2 0 0 2 1 1 1 0 7 Low
Seimon5 Single-Stage 1 1 2 2 1 2 1 0 10 Low
Copyright © 2023 Wolters Kluwer Health, Inc. All rights reserved.
Dorsal
Stricker and Rosen24 Single-Stage 2 2 2 2 1 2 2 0 13 High
Dorsal
J Pediatr Orthop
Mazzocca et al25 Single-Stage 2 0 0 2 2 2 0 0 8 Low
Dorsal
Saini et al26 Single-Stage 1 1 2 2 1 2 0 0 9 Low
Dorsal
Malhotra and Single-Stage 2 2 2 2 1 2 1 0 12 High
Shah27 Dorsal
Raab and Krauspe28 Cincinnati 0 1 2 2 2 2 2 0 11 Low

Kodros and Dias29 Cincinnati 2 2 2 2 2 2 0 0 12 High
Volume 43, Number 5, May/June 2023

Zorer et al7 Cincinnati 1 1 0 2 1 2 1 0 8 Low
Mathew et al8 Cincinnati 0 2 2 2 2 2 2 0 12 High
Bray et al30 Cincinnati 1 0 2 2 0 1 0 0 6 Low
Dobbs et al9 Dobbs 2 2 0 2 2 2 2 0 12 High
Aslani et al31 Dobbs 2 0 0 2 2 2 0 0 8 Low
Chalayon et al32 Dobbs 2 2 0 2 2 2 2 0 12 High
Yang and Dobbs33 Dobbs 2 2 0 2 2 2 2 0 12 High
Hafez and Davis34 Dobbs 2 2 0 2 2 2 2 0 10 Low
Chan et al35 Dobbs 2 2 0 2 2 2 2 0 12 High
Wright et al36 Dobbs 2 2 2 2 2 2 2 0 14 High
Eberhardt et al37 Dobbs 1 1 2 2 2 2 1 0 11 Low
Cummings and Dobbs 2 2 0 2 2 2 2 0 12 High
Hosseinzadeh38
*recorded as 0 (non-reported), 1 (reported but inadequate), or 2 (reported and adequate).
†a score of 12 or higher was considered high methodological quality.
TABLE 2. Coleman-Stelling Technique

Decreased
Patients Average age of Average follow-up subtalar Talar
References Year (feet) patient (range) (range) Recurrence Reoperation motion AVN
Coleman et al3 1970 4 (6) 44.07 mo (36-48 mo) 4.63 y (6 mo-10 y) n=1 n=1 — n=0
Zimbler and Craig13 1983 2 (4) — 10 y n=0 n=0 n=2 n=0

Jacobsen and 1983 6 (9) 5.17 mo (birth-17 mo) 2.1 y n=3 n=3 — n=0
Crawford14
Walker et al15 1985 10 (15) 8.5 mo (1 wk-30.02 10.2 y (15 mo-21 y) n=1 — n=6 n=0
mo)
Totals, n (%) 22 (34) 19.25 mo 6.73 y n = 5 (14.7) n = 4 (21.1) n = 8 (42.1) n=0
The Dobbs Method though the sample size of this data was too small to
Treatment for CVT was transformed in 2006 when confirm this difference statistically.
Dobbs introduced a new, minimally invasive approach to
treating this condition.6 This novel treatment method be-
gins with weekly manipulations and castings followed by a
minimally invasive surgery that includes percutaneous
DISCUSSION
fixation of the talonavicular joint (closed or open) and a While the evolution of surgical correction of con-
on 04/15/2023
percutaneous Achilles tendon tenotomy. Reduction of the genital vertical talus (CVT) over the past several decades
talonavicular joint is maintained with a Kirschner wire. has been well documented, there has not been a meta-
There was a total of 9 studies (145 patients, 219 feet) analysis performed that compares the outcomes of these
meeting our inclusion criteria in which the Dobbs Method different techniques. The minimally invasive Dobbs
was utilized.6,31–37,42 Results are shown in Table 6. Method, which was first published in 2006, provides nu-
Radiographic recurrence of the deformity was noted in merous theoretical advantages, including fewer surgical
21.5% of the patients, with 14% requiring reoperation. No complications and improved long-term ankle range of
cases of talar AVN were reported. motion, but the clinical outcomes in these patients have
never been compared with those who underwent treatment
Comparison Between the Methods with the other techniques. We present the first meta-
analysis on the outcomes of 5 different surgical techniques
Radiographic Recurrence of the Deformity used for CVT correction. We found the Seimon Approach
χ2 analysis revealed that radiographic recurrence of to result in a lower rate of radiographic recurrence of the
the deformity was significantly higher in the children deformity and reoperation compared with other methods,
treated with Direct Medial Approach (29.3%) compared while the Direct Medial Approach resulted in the highest
with all other methods. Children in the Single-Stage rate of radiographic recurrence. In the area of clinical
Dorsal (Seimon) Approach (11%) cohort had significantly scoring, we found the Dobbs Method resulted in higher
lower recurrence rates than those treated with the other clinical scoring and ankle arc of motion compared with
methods, including the Dobbs Method. other methods.
All of these techniques appear to be successful at
Reoperation Rate correcting the initial talonavicular complex deformity, as
χ2 analysis revealed that the reoperation rate was evidenced by normal radiographic TAMBA parameters in
significantly lower in the Single-Stage Dorsal (Seimon) the immediate post-op period. In addition, the majority of
Approach cohort (2%) compared with all other methods cases do seem to maintain long-term correction of the
and significantly higher in the Cincinnati incision cohort talonavicular reduction regardless of what treatment
(16.1%). method is used. However, time reveals differences in both
complications and long-term clinical outcomes among
Clinical Outcomes these different methods.
Clinical scores were compared among the different Although the few articles that cover the Coleman-
treatment cohorts using the modified Adelaar scoring Stelling technique do not mention many complications,
system described in the Methods section above. Patients this 2-stage approach poses the obvious deleterious con-
treated with the Dobbs Method had a significantly higher sequences associated with more time in the operative
average clinical score (8.36) than patients treated with all room, including more anesthesia time and increased risk
other methods, while those treated with the Direct Medial of infection.43 Furthermore, the 2 studies on the Coleman-
Approach had a significantly lower average clinical score Stelling technique that do report ankle range of motion
(6.96) than that seen across the other methods. values report a lower average ankle arch of motion (dor-
siflexion+plantar flexion) than all the other techniques
Ankle arc of Motion that were analyzed.13,15 It is therefore not surprising
The Dobbs method resulted in a larger average ankle that this 2-staged technique has largely been replaced by
range of motion compared with all other methods, al- 1-stage techniques.

on 04/15/2023
Cummings et al
322 | www.pedorthopaedics.com
TABLE 3. Direct Medial Approach
Development
Average age Average Loss of of
Patients of patient follow-up Loss of Hindfoot medial Pes Overcorrection Talar
References Year (feet) (range) (range) reduction Reoperation valgus arch planovalgus to pes cavus AVN
Zhu et al16 2010 7 (9) 18 mo 2 y (1.67-2.42 y) n=0 n=0 n=2 n=5 n=0 n=0 n=0
(10-42 mo)
Sanzarello et al17 2019 5 (9) 38.43 mo 2.60 y (2-4 y) — n=0 — n=3 n=0 n=0 n=0
(24-48 mo)
Ramanoudjame 2014 22 (31) 33 mo 11 y (2-21 y) n=2 n=5 — n = 31 n=3 n=0 n=0
et al18 (10 mo-10 y)
Ayadi et al19 2008 7 (10) 36.14 mo 9.29 y (5-9 y) — — n=3 — n=0 n=0 n=1
(16 mo-6 y)
Schwering20 2005 35 (59) 54.04 mo 7.25 y n = 26 — n = 24 — n=4 n=5 n=0
(6 mo-25.5 y) (7 mo-15.4 y)
Wirth et al21 1994 10 (13) 8.35 mo 3.5 y n=2 n=1 — n=3 — n=1 n=0
(3 mo-6 y) (10 mo-8.5 y)
Oppenheim et al22 1985 11 38 mo 8y n=8 — — — — — n=1
Clark et al23 1977 12 (15) 19.27 mo 3.9 (2-15 y) n=8 n=1 n=6 n = 15 — — n=0
(5 mo-42.03 mo)
Totals, n (%) 157 30.65 mo 5.94 y n = 46 n=7 n = 35 n = 57 n=7 n=6 n=2
(29.30) (4.46) (22.29) (36.31) (4.46) (3.82) (1.27)
J Pediatr Orthop

TABLE 4. Single-Stage Dorsal Approach
Residual
Patients Average age of Average Loss of Hindfoot Loss of forefoot
References Year (feet) patient (range) follow-up (range) reduction Reoperation valgus medial arch abduction
Seimon5 1987 7 (10) 9.57 mo (5-13 mo) 5.17 y (1.25-17 y) n=2 n=1 n=2 n=5 —
Stricker and 1997 13 (20) 11.8 mo (5-27 mo) 3.42 y (1.5-7.33 y) n=5 n=0 n=4 n = 11 n=2
Rosen24
Mazzocca et al25 2001 6 (8) 26.54 mo (6 mo-5.42 y) 3.2 y (3-3.5 y) n=0 n=0 — — —
Saini et al26 2009 12 (20) 16 mo (12-23 mo) 4y n=0 n=0 — n=0 —
Malhotra and 2020 19 (33) 27.5 mo (12 mo-5.33 y) 5.5 y n=3 n=1 see comments see comments n=4
Shah27
Totals, n (%) 57 (91) 18.28 4.26 n = 10 (10.99) n = 2 (2.20) n = 6 (6.59) n = 16 (17.58) n = 6 (6.59)
on 04/15/2023
J Pediatr Orthop
TABLE 5. Cincinnati Incision
Decreased
Average age Average Average Average subtalar
Patients of patient follow-up Loss of Hindfoot plantarflexion dorsiflexion Average arc of range of
References Year (feet) (range) (range) reduction Reoperation valgus (range)(deg) (range)(deg) motion (deg) motion

Raab and 1997 10 (14) 14 mo 3.5 y (1.5-6 y) n=0 n=0 n=2 see comments see comments see comments n=7

Krauspe28 (7-30.02 mo)
Kodros and 1999 32 (42) 24 mo 7 y (2-12 y) n=9 n = 10 n=6 17 16 33 n = 41
Dias29
Zorer et al7 2002 12 (17) 28.4 mo 3.56 y (19 — — n=9 8.1 (5-39) 32.9 (4-48) 41 n = 12
(12 mo-5.08 y) mo-6.75 y)
Mathew et al8 2009 3 (5) 32 mo 7.5 y(7-8 y) n=0 n=0 n=4 25 (15-50) 14 (10-15) 39 n=5
Bray et al30 2017 1 (1) 10 mo 3y n=1 — n=1 — — — —
Totals, n (%) 58 (79) 21.68 4.91 y n = 10 n = 10 n = 22 16.7 (5-50) 20.97 37.67 n = 65
(16.1) (16.1) (27.85) (4-48) (82.28)
TABLE 6. Dobbs Method

Average Average Average Average arc Initial
Patients Average age of follow-up Loss of Hindfoot plantarflexion dorsiflexion of motion correction
References Year (feet) patient (range) (range) reduction Reoperation valgus (range)(deg) (range)(deg) (deg) achieved
Dobbs et al9 2006 11 (19) 8 mo ≥ 2y n=6 n=4 n=4 33 (15-35) 25 (18-35) 58 n = 19
(2-18 mo)
Systematic Review on Congenital Vertical Talus

Aslani et al31 2012 10 (15) 77 mo 2y n=0 n=0 — 16 (12-18) 26 (14-28) 42 n = 15
(1 mo-9 y)
Chalayon et al32 2012 15 (25) 6 mo ≥ 2y n=5 n=5 n=5 25 (15-32) 22 (15-30) 47 n = 25
(1-11 mo)
Yang and Dobbs33 n=4 n=4 — n = 24
www.pedorthopaedics.com
2015 16 (24) 7 mo 7y 23.9 (0-35) 18.5 (5 -30) 42.4

(1-28 mo) (5-11.3 y)
Hafez and Davis34 2021 21 (30) 6 mo 6.5 y n=5 n=5 — 20 10 30 n = 30
(1-17 mo) (1-11 y)
Chan et al35 2016 10 (18) 5 mo 4.42 y n=6 — — — — — —
(2-8 mo) (2-6.67 y)
Wright et al36 2014 13 (21) 5 mo 3y n = 10 n=6 — — — — n = 21
(1-23 mo) (8 mo-4.75 y)
Eberhardt et al37 2012 14 (20) (20 days-14 mo) 2 y (3 mo-5 y) n=1 — — 35.5 (20-40) 23.3 (20-30) 58.8 n = 14
Cummings and 2022 35 (47) 13 mo 4 y (1-9 y) n=4 n=0 n = 10 31 (20-40) 25 (20-35) 56 n = 47
Hosseinzadeh38 (1 wk-5 y)
Totals, n (%) 145 (219) 16 mo 3.7 y n = 41 n = 24 n = 19 26.34 21.4 47.74 n = 195
(18.72) (13.26) (20.88) (97.01)
| 323
Cummings et al J Pediatr Orthop Volume 43, Number 5, May/June 2023
The other 4 treatment techniques each were asso- radiographic recurrence was seen in the Direct Medial Ap-
ciated with unique complications. For example, a rela- proach cohort. Clinical scores were highest in children
tively common complication in patients who underwent a treated with the Dobbs Method. Future long-term studies
direct medial approach was the development of pes pla- focusing on patient-reported outcomes are needed.
novalgus (n = 7,4.46%) and overcorrection to pes cavus
(n = 6,3.82%). In addition, this was the only treatment REFERENCES

group in which there were reports of talar osteonecrosis 1. Miller M, Dobbs MB. Congenital vertical talus: Etiology and
(n = 2).19,22 management. J Am Acad Orthop Surg. 2015;23:604–611.
When compared with the other 4 treatment tech- 2. Lamy L, Weissman L. “CONGENITAL CONVEX PES VALGUS.”.
niques, the Dobbs Method does indeed seem to offer Journal of Bone and Joint Surgery, American Volume. 1939;21:79–91.
3. Coleman SS, Stelling FH 3rd, Jarrett J. Pathomechanics and
certain advantages, as confirmed by the clinical scores treatment of congenital vertical talus. Clin Orthop Relat Res.
reported in our results section. This finding is consistent 1970;70:62–72.
with the 2015 study by Yang and Dobbs that directly 4. Osmond-Clarke H. Congenital vertical talus. J Bone Joint Surg Br.
compared outcomes of the Dobbs Method with those who 1956;38-B:334–341.
underwent more extensive surgery in 42 feet. After a mean 5. Seimon LP. Surgical correction of congenital vertical talus under the
age of 2 years. J Pediatr Orthop. 1987;7:405–411.
follow-up of 7 years, the average total ankle range of 6. Alaee F, Boehm S, Dobbs MB. A new approach to the treatment of
motion in the minimally invasive cohort was 42 ± 18 (5-60) congenital vertical talus. J Child Orthop. 2007;1:165–174.
degrees compared with 12.7 ± 6.8 (5-25) degrees in the 7. Zorer G, Bagatur AE, Dogan A. Single stage surgical correction of
extensive surgery cohort.33 Both these findings and ours congenital vertical talus by complete subtalar release and peritalar
reduction by using the Cincinnati incision. J Pediatr Orthop B.
on 04/15/2023
underscore the potential superiority of the Dobbs Method 2002;11:60–67.

with regard to long-term functional outcomes of the ankle, 8. Mathew PG, Sponer P, Karpas K, et al. Mid-term results of one-
which could ultimately translate to an improved quality of stage surgical correction of congenital vertical talus. Bratisl Lek
life. The potentially improved quality of life afforded by Listy. 2009;110:390–393.
the Dobbs technique was further highlighted in a 2022 9. Dobbs MB, Nunley R, Schoenecker PL. Long-term follow-up of
patients with clubfeet treated with extensive soft-tissue release.
study on patient-reported outcomes from this patient J Bone Joint Surg Am. 2006;88:986–996.
population which found that, on average, these patients 10. Ferreira GF, Stéfani KC, Haje DP, et al. The Ponseti method in
were within 1 SD of the reference population in the do- children with clubfoot after walking age - Systematic review and
mains of pain interference, mobility, and peer relations.38 metanalysis of observational studies. PLoS ONE. 2018;13:e0207153.
11. Slim K, Nini E, Forestier D, et al. Methodological index for non-
The main limitation of our analysis was the lack of randomized studies (MINORS): development and validation of a
standardization in how outcomes were reported among new instrument. ANZ J Surg. 2003;73:712–716.
different studies, including differences in the number of 12. Evans JT, Evans JP, Walker RW, et al. How long does a hip
patients and length of follow-up among the different replacement last? A systematic review and meta-analysis of case
studies. Furthermore, only a few studies reported detailed series and national registry reports with more than 15 years of follow-
up. Lancet. 2019;393:647–654.
ankle range of motion findings, which prevented us from
13. Zimbler S, Craig CL. The arthrogrypotic foot plan of management
performing meaningful statistical comparisons on this and results of treatment. Foot & Ankle. 1983;3:211–219.
topic, and there was only 1 study that described patient- 14. Jacobsen ST, Crawford AH. Congenital vertical talus. J Pediatr
reported outcomes of this population. As many of the Orthop. 1983;3:306–310.
historical studies did not separate idiopathic patients from 15. Walker AP, Ghali NN, Silk FF. Congenital vertical talus. The results
of staged operative reduction. J Bone Joint Surg Br. 1985;67:117–121.
non-idiopathic patients, we similarly did not make this 16. Zhu ZX, Lei W, Huang LY. The short-term effect of an anchor in
distinction during our analysis. In addition, there were treatment of congenital vertical talus in infants. Orthop Surg. 2010;2:
substantially fewer cases reported in the literature that 218–222.
underwent treatment with the Coleman-Stelling Techni- 17. Sanzarello I, Nanni M, Perna F, et al. One-stage release by double
surgical approach for neglected congenital vertical talus: results in a
que and Cincinnati Incision than with the other 3 tech- series of walking children in Tanzania. J Pediatr Orthop B. 2019;28:
niques. Another limitation of this review is the fact that 586–590.
many of the studies included were single-institution studies 18. Ramanoudjame M, Loriaut P, Seringe R, et al. The surgical
that involved procedures performed by the study authors. treatment of children with congenital convex foot (vertical talus):
Some of these authors published multiple studies that were evaluation of midtarsal surgical release and open reduction. Bone
Joint J. 2014;96-B:837–844.
included in this meta-analysis, and it is therefore possible 19. Ayadi K, Trigui M, Gdoura F, et al. Results of surgical treatment of
that a small subset of patients from these studies were congenital convex pes valgus (ten nonidiopathic feet). Rev Chir
duplicates. Lastly, the rarity of this condition is reflected Orthop Reparatrice Appar Mot. 2008;94:e28-34.
by a relative sparsity of overall literature on this topic 20. Schwering L. Surgical correction of the true vertical talus deformity.
compared with more common congenital abnormalities. Oper Orthop Traumatol. 2005;17:211–231; English, German.
21. Wirth T, Schuler P, Griss P. Early surgical treatment for congenital
In conclusion, we have conducted the first meta-anal- vertical talus. Arch Orthop Trauma Surg. 1994;113:248–253.
ysis that compares the results of the 5 most commonly used 22. Oppenheim W, Smith C, Christie W. Congenital vertical talus. Foot
surgical techniques for the correction of CVT in 580 feet. We Ankle. 1985;5:198–204.
found radiographic recurrence of the deformity to occur in 23. Clark MW, D’Ambrosia RD, Ferguson AB. Congenital vertical
talus: treatment by open reduction and navicular excision. J Bone
around 20% of patients, while 8% need reoperation. We Joint Surg Am. 1977;59:816–824.
found the lowest radiographic recurrence and reoperation 24. Stricker SJ, Rosen E. Early one-stage reconstruction of congenital
rate in the Seimon Approach cohort, while the highest rate of vertical talus. Foot Ankle Int. 1997;18:535–543.

25. Mazzocca AD, Thomson JD, Deluca PA, et al. Comparison of the 35. Chan Y, Selvaratnam V, Garg N. A comparison of the Dobbs
posterior approach versus the dorsal approach in the treatment of method for correction of idiopathic and teratological congenital
congenital vertical talus. J Pediatr Orthop. 2001;21:212–217. vertical talus. J Child Orthop. 2016;10:93–99.
26. Saini R, Gill SS, Dhillon MS, et al. Results of dorsal approach in 36. Wright J, Coggings D, Maizen C, et al. Reverse Ponseti-type
surgical correction of congenital vertical talus: an Indian experience. treatment for children with congenital vertical talus: comparison
J Pediatr Orthop B. 2009;18:63–68. between idiopathic and teratological patients. Bone Joint J. 2014;96-
27. Malhotra M, Shah H. Comparison of outcome between idiopathic B:274–278.

and non-idiopathic congenital vertical talus treated with soft tissue 37. Eberhardt O, Fernandez FF, Wirth T. The talar axis-first metatarsal
release. J Pediatr Orthop B. 2020;30:458–466. base angle in CVT treatment: a comparison of idiopathic and non-
28. Raab P, Krauspe R. One-stage procedure for surgical correction of idiopathic cases treated with the Dobbs method. J Child Orthop.
congenital vertical talus. Foot Ankle Surg. 1997;3:71–76. 2012;6:491–496.
29. Kodros SA, Dias LS. Single-stage surgical correction of congenital 38. Cummings JL, Hosseinzadeh P. Patient-reported pain and func-
vertical talus. J Pediatr Orthop. 1999;19:42–48. tion outcomes in children with congenital vertical talus treated
30. Bray JJH, Crosswell S, Brown R. Congenital talipes equinovarus and with the minimally invasive method. J Pediatr Orthop. 2022;42:
congenital vertical talus secondary to sacral agenesis. BMJ Case Rep. e981–e984.
2017;2017:bcr2017219786. 39. Ogata K, Schoenecker PL, Sheridan J. Congenital vertical talus and
31. Aslani H, Sadigi A, Tabrizi A, et al. Primary outcomes of the its familial occurrence: an analysis of 36 patients. Clin Orthop Relat
congenital vertical talus correction using the Dobbs method of serial Res. 1979;139:128–132.
casting and limited surgery. J Child Orthop. 2012;6:307–311. 40. Kumar SJ, Cowell HR, Ramsey PL. Vertical and oblique talus. Instr
32. Chalayon O, Adams A, Dobbs MB. Minimally invasive approach Course Lect. 1982;31:235–251.
for the treatment of non-isolated congenital vertical talus. J Bone 41. Crawford AH, Marxen JL, Osterfeld DL. The Cincinnati incision:
Joint Surg Am. 2012;94:e73. a comprehensive approach for surgical procedures of the foot
33. Yang JS, Dobbs MB. Treatment of congenital vertical talus: and ankle in childhood. J Bone Joint Surg Am. 1982;64:
on 04/15/2023
comparison of minimally invasive and extensive soft-tissue release 1355–1358.

procedures at minimum five-year follow-up. J Bone Joint Surg Am. 42. Cummings JL, Hosseinzadeh P. Minimally invasive method in
2015;97:1354–1365. treatment of idiopathic congenital vertical talus: recurrence is
34. Hafez M, Davis N. Outcomes of a minimally invasive approach for uncommon. J Pediatr Orthop. 2022;42:503–508.
congenital vertical talus with a comparison between the idiopathic 43. Linnaus ME, Ostlie DJ. Complications in common general pediatric
and syndromic feet. J Pediatr Orthop. 2021;41:249–254. surgery procedures. Semin Pediatr Surg. 2016;25:404–411.

ORIGINAL ARTICLE
Comparison of Distal Spine Anchors and Distal Pelvic

Anchors in Children With Hypotonic Neuromuscular
Scoliosis Treated With Growth-friendly Instrumentation

Arya Ahmady, MD,* Lindsay Rosenthal, MS,* Adam C. Abraham, PhD,* Bianca Parker, MS,†
Jaysson T. Brooks, MD,‡ Patrick J. Cahill, MD,§ John T. Smith, MD,∥ Paul D. Sponseller, MD,¶
Peter F. Sturm, MD,# and Ying Li, MD,* on behalf of the Pediatric Spine Study Group
returns to the operating room rates. Subanalysis of the DSA

Background: Lower preoperative pelvic obliquity (PO) and L5 group based on ambulatory status showed similar radiographic
tilt have been associated with good radiographic outcomes when measures except the ambulatory patients had a lower PO at all
the fusion ended short of the pelvis in children with neuro- time points (preindex: 5 vs. 16 degrees, P = 0.011; postindex: 6 vs.
on 04/15/2023
muscular scoliosis (NMS). Our purpose was to identify in- 10 degrees, P = 0.045; most recent follow-up: 5 vs. 14 degrees,
dications to exclude the pelvis in children with hypotonic NMS P = 0.028). Only 1 ambulatory DSA patient had a PO ≥ 10 de-
treated with growth-friendly instrumentation. grees at most recent follow-up compared with 6 nonambulatory
Methods: This was a multicenter retrospective review. Children DSA patients. Three (8%) DSA patients, all nonambulatory,
with spinal muscular atrophy and muscular dystrophy treated with underwent extension of their instrumentation to the pelvis.
dual traditional growing rod, magnetically controlled growing rod, Conclusions: Pelvic fixation should be strongly considered in
or vertical expandable prosthetic titanium rib with minimum nonambulatory children with hypotonic NMS treated with
2-year follow-up after the index surgery were identified. growth-friendly instrumentation. At intermediate-term follow-
Results: A total of 125 patients met the inclusion criteria. Thirty- up, revision surgery to include the pelvis was rare but DSAs do
eight patients had distal spine anchors (DSAs) and 87 patients not seem effective at maintaining control of PO in non-
had distal pelvic anchors (DPAs) placed at the index surgery. ambulatory patients. DSA and DPA were equally effective at
Demographics and length of follow-up were similar between the maintaining major curve control, and complication and un-
groups but there was a greater percentage of DPA patients who planned returns to the operating room rates were similar.
were nonambulatory [79 patients (91%) vs. 18 patients (47%), Level of Evidence: Level III—therapeutic.
P < 0.0001]. Preindex radiographic measures were similar except
the DSA patients had a lower PO (11 vs. 19 degrees, P = 0.0001) Key Words: spinal muscular atrophy, muscular dystrophy, early
and L5 tilt (8 vs. 12 degrees, P = 0.001). Postindex and most onset scoliosis, pelvic fixation, growing rods, VEPTR, pelvic obliquity
recent radiographic data were comparable between the groups.
There was no difference in the complication and unplanned (J Pediatr Orthop 2023;43:e319–e325)
From the *Department of Orthopaedic Surgery, C.S. Mott Children’s Hospital, Michigan Medicine, Ann Arbor, MI; †Wayne State University School
of Medicine, Detroit, MI; ‡Department of Orthopaedic Surgery, Texas Scottish Rite Hospital for Children, University of Texas Southwestern,
Dallas, TX; §Department of Orthopaedic Surgery, Children’s Hospital of Philadelphia, Philadelphia, PA; ∥Department of Orthopaedic Surgery,
Primary Children’s Hospital, University of Utah, Salt Lake City, UT; ¶Department of Orthopaedic Surgery, Johns Hopkins Medicine, Baltimore,
MD; and #Department of Orthopaedic Surgery, Cincinnati Children’s Hospital Medical Center, Cincinnati, OH.
The Pediatric Spine Study Group collaborators are listed in the Appendix.
None of the authors received financial support for this study.
J.T.B. has received consulting fees from Depuy Synthes, Medtronic, and OrthoPediatrics. P.J.C. has received research grants from the Setting Scoliosis
Straight Foundation and the Children’s Spine Study Group, has a patent for dynamic lung magnetic resonance imaging, and is a board or committee
member of the Scoliosis Research Society, Journal of Bone and Joint Surgery American, Spine Deformity, and the Pediatric Orthopaedic Society of
North America. J.T.S. has received royalties from Globus, has a patent for a hydraulic growing rod, and is a board member of the Pediatric Spine
Study Group. P.D.S. has received a grant from Ipsen, has received royalties from Globus, has received consulting fees from Depuy Synthes, and has
received payment for lectures from NuVasive. P.F.S. has received consulting fees from NuVasive, is a board or committee member of the Pediatric
Orthopaedic Society of North America, the Scoliosis Research Society, and Journal of Children’s Orthopaedics, and has stock or stock options from
Green Sun Medical. Y.L. has received a research grant from the Scoliosis Research Society, has received consulting fees from Medtronic, has
received support for attending an educational event from Zimmer, and is a committee member of the Pediatric Orthopaedic Society of North
America and the Scoliosis Research Society. The Pediatric Spine Study Group has received research grants from Boston Orthotics and Prosthetics,
Depuy Synthes Spine, Globus Medical, Medtronic, NuVasive, nView Medical, OrthoPediatrics, Pacira, the Pediatric Spine Foundation, Stryker, and
Zimmer Biomet. The remaining authors declare no conflicts of interest.
Reprints: Ying Li, MD, Department of Orthopaedic Surgery, C.S. Mott Children’s Hospital, 1540 E. Hospital Drive, SPC 4241, Ann Arbor, MI 48109-
4241. E-mail: yingyuli@med.umich.edu.
DOI: 10.1097/BPO.0000000000002376

Ahmady et al J Pediatr Orthop Volume 43, Number 5, May/June 2023
N euromuscular scoliosis (NMS) is defined as a spinal

curvature caused by disorders of the brain, spinal cord,
and muscular system. Examples of such diagnoses include
base of patients with early onset scoliosis. Children with
SMA and muscular dystrophy treated with dual tradi-
tional growing rods (TGRs), magnetically controlled
cerebral palsy, muscular dystrophy, myelomeningocele, and growing rods (MCGRs), or vertical expandable prosthetic
spinal muscular atrophy (SMA).1 The onset of deformity can titanium rib (VEPTR) with minimum 2-year follow-up
occur at any age, and the deformity is more likely to progress after the index surgery were identified. Patients with as-
with increased systemic involvement of the disease.2,3 In sociated congenital scoliosis, unknown ambulatory status,
contrast to adolescent idiopathic scoliosis, NMS is associated and prior surgical treatment for scoliosis were excluded.
with more rapid curve progression, higher degrees of pelvic Demographic data collected included age at index
obliquity (PO), involvement of more vertebrae, and higher surgery, sex, and preindex body mass index and ambula-
rates of pulmonary complications.4 tory status. Radiographic measures, including major curve
The magnitude of the spinal curvature in these patients magnitude, L5 tilt, PO, and maximum sagittal deformity,
can significantly affect mobility and gait in ambulatory pa- were collected before the index surgery, immediately after
tients and sitting balance in nonambulatory patients. In ad- the index surgery, and at most recent follow-up. The
dition, more severe deformities may lead to increased pain and measurement technique for L5 tilt and PO are shown in
skin compromise, which can significantly increase the burden Figure 1. L5 tilt and PO were both measured on an
on caregivers. Bracing has little effect on curve progression anteroposterior radiograph. L5 tilt was determined by
and mainly acts as an external support for those with flexible measuring the angle between a line extending along the
curves.5 In patients with SMA, bracing is rarely used due to upper endplate of L5 and the intercristal line.10 PO was
on 04/15/2023
the potential constrictive effects of the brace on an already determined by measuring the angle between the
compromised ability to achieve adequate lung volumes.6 perpendicular of a line extending from the center of the
Surgical intervention for NMS is indicated when T1 vertebral body to the center of S1 and the intercristal
there is progressive deformity with an unacceptable trun- line.10 All radiographic measurements were performed
cal shift or PO that affects standing or sitting balance and using Surgimap software (Nemaris Inc., New York, NY).
positioning.1 Because of the high incidence of PO in NMS, The type of growth-friendly device, type and loca-
pelvic instrumentation is often required to achieve better tion of the distal anchors (spine or pelvis), number of
truncal balance.7 Pelvic fixation is generally obtained with patients with DSAs who underwent extension of their
iliac screws, which can be prominent and lead to skin construct to the pelvis, number of patients with compli-
breakdown and infection.1,7 In addition, instrumenting to cations, type of complications, and number of patients
the pelvis carries the risk of increased blood loss during with unplanned returns to the operating room (UPROR)
surgery, longer operative times, and limited mobility.8 were also recorded. Complications were categorized as
Rates of pelvic implant failure have also been reported to implant-related, wound-related, neurologic, medical/oth-
be as high as 29% in patients with NMS.9 er, and death. The 24-Item Early Onset Scoliosis Ques-
Because of the risks associated with pelvic fixation, tionnaire (EOSQ-24)12 was used to assess quality of life.
several authors have attempted to identify indications to Statistical analyses were performed using GraphPad
exclude the pelvis when performing a spinal fusion for Prism 9.3.1 (GraphPad Software, San Diego, CA). Con-
NMS. A prior study showed that exclusion of the pelvis at tinuous variables were checked for normality using a
the time of posterior spinal fusion in patients with flaccid D’Agostino-Pearson test. Normally distributed data were
NMS with preoperative L5 tilt <15 degrees and curve evaluated using an independent t test, and non-normally
apex at L2 or higher resulted in good radiographic out- distributed data were evaluated using a Mann-Whitney U
comes with a low complication rate.10 L5 tilt ≤ 10 degrees test. Categorical variables were evaluated using a χ2 test.
and PO <10 degrees have also been reported as possible All significance tests were 2 tailed and a P-value of <0.05
indications to exclude the pelvis in children with cerebral was considered significant. For the multivariable re-
palsy and scoliosis treated with distraction-based growth- gression analyses, the dependent variables were postindex
friendly instrumentation.11 Indications to exclude the and most recent PO and the independent variables were
pelvis in children with hypotonic NMS undergoing treat- preindex apex of major curve (L2 and proximal vs. distal
ment with growth-friendly implants has not been studied. to L2), lowest instrumented vertebra [(LIV), proximal to
The purpose of this study was to identify indications to L4 vs. L4 and distal], and type of pelvic anchor (screws vs.
exclude the pelvis in children with hypotonic NMS treated S hooks). The cutoff for the significance for the regression
with distraction-based growth-friendly instrumentation. We analyses was set at P = 0.10.
hypothesized that the pelvis can be successfully excluded in
properly selected patients, and radiographic measures and
outcomes will be similar between patients with distal spine RESULTS
anchors (DSA) and distal pelvic anchors (DPAs). One hundred and twenty-five patients met the in-
clusion criteria. Thirty-eight patients had DSA and 87 had
DPA placed at the index surgery. There were no sig-
METHODS nificant differences in mean preindex age, sex, preindex
This was a retrospective review of an Institution BMI, or length of follow-up between the groups, but there
Review Board–approved multicenter international data- was a greater percentage of DPA patients who were

J Pediatr Orthop Volume 43, Number 5, May/June 2023 Spine vs Pelvic Anchors in Neuromuscular Scoliosis
nonambulatory before the index surgery [79 patients (47%) vs. 54 patients (62%), P = 0.233) and UPRORs
(91%) vs. 18 patients (47%), P < 0.0001]. Demographic [7 patients (18%) vs. 23 patients (26%), P = 0.493]
and surgical data for the DSA and DPA groups are shown (Table 4). Implant-related, wound-related, and medical
in Table 1. complications were equally common in the DSA group. In
Radiographic data are shown in Table 2. Location the DPA group, medical complications were the most
of the apex of the major curve, mean preindex major curve common, followed by a similar rate of implant and wound
magnitude, and mean preindex maximum sagittal complications.

deformity were similar between the DSA and DPA The analysis of EOSQ-24 scores at most recent
groups. However, the DSA patients had significantly follow-up showed that the DSA patients had significantly
lower mean preindex PO (11 vs. 19 degrees, P = 0.0001) better mean physical function (49 vs. 24, P = 0.0002),
and L5 tilt (8 vs. 12 degrees, P = 0.001) compared with the daily living (44 vs. 21, P < 0.0001), and satisfaction
DPA patients. There were no significant differences in any (61 vs. 49, P = 0.033) compared with the DPA patients
of the radiographic parameters at the postindex and most (Table 5).
recent time points between the DSA and DPA groups.
Both the DSA and DPA patients maintained correction of
their PO during the follow-up period, and the most recent
A
PO was 9 degrees in both groups (P = 0.229). Eight (27%)
of the DSA and 29 (39%) of the DPA patients had a PO
≥ 10 degrees at most recent follow-up (P = 0.245).
on 04/15/2023
Multivariable regression analysis did not demonstrate an

association between apex of major curve or LIV and
postindex or most recent PO in the DSA group. Type of
pelvic anchor also did not have an association with
postindex or most recent PO in the DPA group.
As there was a similar number of ambulatory and
nonambulatory patients in the DSA group, a subanalysis
was performed to compare radiographic parameters
based on ambulatory status (Table 3). There was no
difference in the radiographic measures except for PO, L5 tilt
which was significantly lower in the ambulatory DSA
patients at every timepoint (preindex: 5 vs. 16 degrees,
P = 0.011; postindex: 6 vs. 10 degrees, P = 0.045; most
recent follow-up: 5 vs. 14 degrees, P = 0.028). Only 1
ambulatory DSA patient had a PO ≥ 10 degrees at most
recent follow-up compared with 6 nonambulatory DSA
patients.
Three (8%) DSA patients, all nonambulatory, B
required later extension of their construct to the pelvis.
One patient was a male with muscular dystrophy who had
a PO of 26 degrees and a L5 tilt of 7 degrees before TGR
insertion with a LIV of L5. His postindex PO was 12
degrees. The PO increased to 20 degrees during the course
of growing rod treatment and he underwent pelvic fixation
with screws at the time of final fusion. The second patient
was a female with spinal muscular atrophy who had a PO
of 27 degrees and L5 tilt of 28 degrees before TGR in-
sertion with a LIV of L5. Fixation to the pelvis was unable
to be performed initially because of significant malnu-
trition. Her postindex PO was 6 degrees. She underwent
planned extension of the TGR construct to the pelvis Pelvic obliquity
1.2 years after her index surgery. The third patient was a
male with spinal muscular atrophy who had a PO of 1
degrees and L5 tilt of 2 degrees before MCGR insertion
with a LIV of L5. His postindex PO was 2 degrees. He
underwent extension of the MCGR construct to the pelvis
at a planned rod exchange 3.6 years after the index
surgery.
There was a similar percentage of patients in the FIGURE 1. Demonstration of the measurement technique for
DSA and DPA groups with complications [18 patients L5 tilt (A) and pelvic obliquity (B).

TABLE 1. Demographic and Surgical Data for Entire Cohort TABLE 3. Radiographic Data for Ambulatory and Nonambulatory
Distal spine Distal pelvic Distal Spine Anchor Patients
anchors anchors Ambulatory Nonambulatory
(N = 38) (N = 87) P (N = 16) (N = 18) P
Age at index surgery 7.4 ± 2.3 7.6 ± 2.2 0.698 Apex of major curve [n (%)]
(years; mean ± SD) L2 or higher 15 (94) 17 (94) —

Sex [n (%) female] 21 (55) 48 (55) 0.955 Below L2 1 (6) 1 (6) 0.932
BMI at index surgery (kg/ 17.4 ± 5.2 16.5 ± 4.7 0.231 Major curve (degree; mean ± SD)
2
m ; mean ± SD) Preindex 67 ± 25 71 ± 18 0.529
Nonambulatory status at 18 (47) 79 (91) < 0.0001 Postindex 38 ± 16 47 ± 20 0.182
index surgery [n (%)] Most recent 45 ± 16 44 ± 21 0.911
Length of follow-up 5.1 ± 2.2 5.4 ± 2.9 0.937 Pelvic obliquity (degree; mean ± SD)
(years; mean ± SD) Preindex 5±4 16 ± 15 0.011
Type of growth-friendly device (n) Postindex 6±4 10 ± 7 0.045
TGR 20 15 — Most recent 5±4 14 ± 15 0.028
MCGR 15 27 — L5 tilt (degree; mean ± SD)
VEPTR 3 44 — Preindex 9±7 8±7 0.777
Hybrid construct — 1 — Postindex 6±4 3±3 0.044
LIV — NA — Most recent 4±4 9 ± 10 0.317
L1 1 — — Maximum sagittal deformity (degree; mean ± SD)
L2 1 — — Preindex 46 ± 30 55 ± 28 0.447
L3 11 — — Postindex 40 ± 19 43 ± 19 0.608
on 04/15/2023
L4 14 — — Most recent 51 ± 28 39 ± 20 0.205

L5 11 — —
Type of pelvic anchor N/A — — Values in bold are significant.
Screws — 38 —
S hooks — 49 —
Undergone final fusion 10 (26) 23 (26) — distraction-based growth-friendly treatment, with similar
[n (%)]
complication and UPROR rates. However, we found that
Value in bold is significant. DSAs were unable to maintain permanent correction of PO
BMI indicates body mass index; LIV, lowest instrumented vertebra; MCGR,
magnetically controlled growing rod; TGR, traditional growing rod; VEPTR,
in nonambulatory patients. Although mean PO improved
vertical expandable prosthetic titanium rib. from 16 to 10 degrees after the index surgery in our non-
ambulatory DSA group, PO increased to 14 degrees at a
mean follow-up of 5.1 years. We are unaware of any pub-
DISCUSSION lished measurement error for PO but the reported meas-
DSAs and DPAs seem to be equally effective at long- urement error for Cobb angle is ≤ 5 degrees.13,14 If we
term control of the coronal and sagittal spinal deformity presume that PO has a similar measurement error, then the
in children with hypotonic NM scoliosis undergoing nonambulatory and ambulatory DSA patients had a com-
parable (although statistically significant) PO at the post-
index timepoint (10 vs. 6 degrees). However, PO increased
TABLE 2. Radiographic Data for Entire Cohort in the nonambulatory group during the follow-up period
Distal spine Distal pelvic and they had a PO that was 9 degrees higher than the
anchors (N = 38) anchors (N = 87) P ambulatory DSA group at the most recent timepoint. Only
3 DSA patients in our series had extension of their construct
Apex of major curve — — 0.537
[n (%)] to the pelvis but not all patients have undergone final
L2 or higher 36 (95) 76 (87) — fusion. As such, we recommend pelvic fixation in non-
Below L2 2 (5) 11 (13) — ambulatory children with hypotonic NMS treated with
Major curve (degree; mean ± SD) TGR, MCGR, and VEPTR.
Preindex 70 ± 22 71 ± 25 0.812
Postindex 43 ± 18 43 ± 18 0.884
Children with NMS can experience rapid curve
Most recent 43 ± 18 48 ± 21 0.184 progression and require early operative treatment4 so it is
Pelvic obliquity (degree; mean ± SD) critical to assess which surgical interventions will be most
Preindex 11 ± 12 19 ± 12 0.0001 effective long term with regard to curve correction, im-
Postindex 8±6 8±6 0.716 provement in PO, and mobility while also maintaining
Most recent 9 ± 12 9±7 0.229
L5 tilt (degree; mean ± SD) high caregiver satisfaction and low complication rates.
Preindex 8±7 12 ± 9 0.001 Although we did not find a difference in the complication
Postindex 5±4 5±4 0.318 rate between patients with distal spine or DPAs, increased
Most recent 7±8 6±5 0.969 rates of skin breakdown and infection, implant failure,
Maximum sagittal deformity (degree; mean ± SD)
Preindex 52 ± 28 54 ± 29 0.950
intraoperative blood loss, longer operative times, and
Postindex 43 ± 20 36 ± 18 0.071 limited mobility have been reported with pelvic
Most recent 43 ± 24 41 ± 24 0.704 instrumentation.1,7,8,15 Thus, it seems pertinent to de-
Values in bold are significant.
termine whether exclusion of the pelvis in spinal deformity
procedures is warranted in certain subsets of pediatric

Although none of our ambulatory DSA patients have

TABLE 4. Complications Data for Entire Cohort
necessitated revision surgery to include the pelvis, not all
Distal spine patients have undergone final fusion. These children are
anchors Distal pelvic
likely returning to the operating room for multiple sur-
(N = 38) anchors (N = 87) P
geries throughout the course of treatment, whether for rod
Total complications (n) 32 144 — lengthening, rod exchange, or final fusion, and if sig-
Type of complication [n (%)]

Implant-related 11 (35) 35 (24) —
nificant PO develops over time, then the instrumentation
Wound-related 10 (31) 34 (24) — can be extended to the pelvis during a planned procedure.
Neurological 1 (3) 8 (6) — One reported benefit of excluding the pelvis is maintained
Medical/other 10 (31) 64 (44) — mobility at the L5-S1 joint, which may be advantageous
Death 0 3 (2) — for transfers, weight shifts, and rotational mobility in
No. patients with 18 (47) 54 (62) 0.233
complications [n (%)] ambulatory NMS patients.16 In addition, many surgeons
Mean complications per 0.86 1.66 0.120 are concerned that fixation to the pelvis will result in de-
patients (n) creased function, including limiting ambulatory potential
Total UPRORs (n) 8 32 — in a patient population who already has challenges with
No. patients with 7 (18) 23 (26) 0.493
UPRORs [n (%)]
independent ambulation. However, this is controversial as
Tsirikos et al17 found the maintenance of ambulatory
UPROR indicates unplanned return to operating room. ability in 23 of 24 cerebral palsy patients at a mean follow-
up of nearly 3 years after posterior spinal fusion with
on 04/15/2023
pelvic instrumentation. The benefits of increased mobility

NMS patients. Takaso et al10 retrospectively evaluated at the lumbosacral junction with DSAs and possible
patients with flaccid NMS who underwent posterior spinal preservation of ambulatory potential likely outweigh the
fusion to L5 using segmental pedicle screw in- small risk of future PO progression and potential need for
strumentation. These authors found that patients with pelvic instrumentation in this subset of hypotonic NMS
preoperative L5 tilt <15 degrees and a curve apex at L2 or patients.
higher had excellent coronal curve correction and im- A significantly greater percentage of patients who
provement in PO to <10 degrees that was maintained at received DPAs at the index surgery in our study were
minimum 2-year follow-up. Neither the location of the nonambulatory compared with the patients who received
apex of the major curve nor the LIV was associated with DSAs. The DPA group also had significantly higher mean
postindex or most recent PO in our DSA patients. preindex PO and L5 tilt, suggesting that nonambulatory
In our study, the ambulatory DSA patients had a hypotonic NMS patients are more likely to have an un-
mean PO that remained stable at most recent follow-up. Li balanced pelvis that requires correction at the time of in-
et al11 previously showed that preoperative PO <10 de- sertion of growth-friendly implants. This is supported by
grees and L5 tilt ≤ 10 degrees may be the indications to the results of our subanalysis of the DSA group, which
exclude the pelvis in children with cerebral palsy and EOS showed a significantly higher mean preindex PO in the
treated with distraction-based growth-friendly implants. nonambulatory patients compared with the ambulatory
Similar criteria could be followed for ambulatory hypo- patients. Although the mean postindex PO improved in
tonic NMS patients, as our ambulatory DSA patients had the nonambulatory DSA patients, this correction could
a mean preindex PO of 5 degrees and L5 tilt of 9 degrees. not be maintained and 6 patients had a PO ≥ 10 degrees at
most recent follow-up. Only 3 nonambulatory DSA pa-
TABLE 5. EOSQ-24 Data for Entire Cohort at Most Recent tients underwent extension of their construct to the pelvis
Follow-up but since the majority of patients in our series are still
Distal spine Distal pelvic undergoing growth-friendly treatment, it is yet to be de-
anchors (N = 38) anchors (N = 87) P termined how many patients may require pelvic fixation in
the future. DPAs were effective at maintaining long-term
General health 71 ± 16 70 ± 18 0.696
Pain/discomfort 69 ± 22 62 ± 26 0.243 control of PO in our DPA patients so we recommend
Pulmonary 85 ± 21 81 ± 27 0.585 including the pelvis in nonambulatory hypotonic NMS
function patients. Nielsen and colleagues presented similar con-
Transfer 64 ± 31 55 ± 32 0.185 clusions in their review of 285 patients with a variety of
Physical function 49 ± 32 24 ± 28 0.0002 neuromuscular diagnoses who had undergone posterior
Daily living 44 ± 27 21 ± 27 < 0.0001
Fatigue/energy 65 ± 24 59 ± 28 0.293 spinal fusion with pelvic fixation either at the index sur-
level gery or during revision surgery. Mean preoperative PO
Emotion 69 ± 20 63 ± 24 0.303 (22.1 vs. 18.3 degrees) and percent improvement in PO
Parental impact 63 ± 22 55 ± 22 0.089 (50.7% vs. 46.3%) were comparable between groups
Financial impact 68 ± 25 66 ± 33 0.957
Family burden 65 ± 21 61 ± 24 0.351
postoperatively. However, a significantly greater percent-
Satisfaction 61 ± 23 49 ± 27 0.033 age of patients in the revision pelvic fusion group under-
went reoperations at a later time than in the index pelvic
Value in bold is significant.
Presented as the mean ± SD. fusion group (50.0% vs. 22.9%) and implant failure was a
significantly more common reason for reoperation in the

revision pelvic fusion patients. As such, these authors Erickson, Jorge Fabregas, Frances Farley, David Far-
suggest including the pelvis at the index posterior spinal rington, Graham Fedorak, Ryan Fitzgerald, Lorena
fusion in this patient population.8 Floccari, Jack Flynn, Peter Gabos, Adrian Gardner, Su-
Although the DSA patients in our study had better meet Garg, Frank Gerow, Michael Glotzbecker, Jaime
scores in several EOSQ-24 domains compared with the Gomez, David Gonda, Tenner Guillaume, Purnendu
DPA patients at their most recent follow-up, it is difficult Gupta, Kyle Halvorson, Kim Hammerberg, Christina
to know how to interpret these data, as we had incomplete Hardesty, Daniel Hedequist, Michael Heffernan, John
preindex EOSQ-24 scores in the database. Nearly 50% of Heflin, Ilkka Helenius, Jose Herrera, Grant Hogue, Josh
the complications in the DPA group were medical com- Holt, Jason Howard, Michael Timothy Hresko, Steven
pared with one third of the complications in the DSA Hwang, Stephanie Ihnow, Brice Ilharreborde, Kenneth
group. As such, the DPA patients may have had more Illingworth, Viral Jain, Andrew Jea, Megan Johnson,
medical comorbidities that would have affected their Charles Johnston, Morgan Jones, Judson Karlen, Law-
EOSQ-24 scores. In addition, a greater percentage of our rence Karlin, Danielle Katz, Noriaki Kawakami, Brian
DPA group was nonambulatory, and a negative correla- Kelly, Derek Kelly, Raymond Knapp, Paul Koljonen,
tion between EOSQ-24 scores and nonambulatory status Kenny Kwan, Hubert Labelle, Robert Lark, A. Noelle
and NMS has been reported.18 Larson, William Lavelle, Lawrence Lenke, Sean Lew,
This study has several limitations. There was almost Ying Li, Craig Louer, Scott Luhmann, Jean-Marc Mac-
certainly treatment bias between the DSA and DPA Thiong, Stuart Mackenzie, Erin MacKintosh, Francesco
groups, given a significant difference in ambulatory status, Mangano, David Marks, Sanchez Marquez, Jonathan
indicating surgeons’ preference to avoid pelvic fixation in
on 04/15/2023
Martin, Jeffrey Martus, Antonia Matamalas, Oscar

ambulatory NM patients. As this is a multicenter retro- Mayer, Richard McCarthy, Amy McIntosh, Jessica
spective study, it is not possible to determine the criteria McQuerry, Jwalant Mehta, Lionel Metz, Daniel Miller,
that surgeons used to select the LIV for each patient at the Firoz Miyanji, Greg Mundis, Josh Murphy, Robert
index surgery. Similarly, we do not know the indications Murphy, Karen Myung, Susan Nelson, Peter Newton,
to extend the construct to the pelvis in the 3 DSA patients. Matthew Newton Ede, Cynthia Nguyen, Susana Nunez,
End vertebrae and traction radiographs are not collected Matthew Oetgen, Timothy Oswald, Jean Ouellet, Josh
in the database so we were unable to evaluate whether Pahys, Kathryn Palomino, Stefan Parent, Alejandro Peiro
these factors could assist with determining the need for Garcia, Ferran Pellise, Joseph Perra, Jonathan Phillips,
pelvic fixation. In addition, the majority of our patients Javier Pizones, Selina Poon, Nigel Price, Norman Ram-
have not completed growth-friendly treatment or under- irez-Lluch, Brandon Ramo, Gregory Redding, Todd
gone final fusion. As such, additional patients with DSAs Ritzman, Luis Rodriguez, Juan Carlos Rodriguez-Ola-
may undergo extension of their construct to the pelvis in verri, David Roye, Benjamin Roye, Lisa Saiman, Amer
the future. Samdani, Francisco Sanchez Perez-Grueso, James Sand-
In conclusion, pelvic fixation should be strongly con- ers, Jeffrey Sawyer, Christina Sayama, Michael Schmitz,
sidered in nonambulatory children with hypotonic NMS Jacob Schulz, Richard Schwend, Suken Shah, Jay Sha-
treated with distraction-based growth-friendly instrumen- piro, Harry Shufflebarger, David Skaggs, Kevin Smit,
tation. At intermediate-term follow-up, revision surgery to John Smith, Brian Snyder, Paul Sponseller, George Ste-
include the pelvis was rare but DSAs do not seem effective at phen, Joe Stone, Peter Sturm, Hamdi Sukkarieh, Ishaan
maintaining the control of PO in nonambulatory patients. Swarup, Michal Szczodry, John Thometz, George
PO remained stable at most recent follow-up in our ambu- Thompson, Tanaka Tomoko, Walter Truong, Raphael
latory DSA patients and DPA patients. DSA and DPA were Vialle, Michael Vitale, John Vorhies, Eric Wall, Shengru
equally effective at long-term major curve control, and Wang, Bill Warner, Stuart Weinstein, Michelle Welborn,
complication and UPROR rates were similar. Klane White, David Wrubel, Nan Wu, Kwadwo Yankey,
Burt Yaszay, Muharrem Yazici, and Terry Jianguo Zhang
Appendix
The Pediatric Spine Study Group: Abdullah Saad REFERENCES
1. Murphy RF, Mooney JF. Current concepts in neuromuscular
Abdulfattah Abdullah, Edward Ahn, Behrooz Akbarnia, scoliosis. Curr Rev Musculoskelet Med. 2019;12:220–227.
Harry Akoto, Stephen Albanese, Jason Anari, John An- 2. Persson-Bunke M, Hägglund G, Lauge-Pedersen H, et al. Scoliosis in
derson, Richard Anderson, Lindsay Andras, Jennifer Ba- a total population of children with cerebral palsy. Spine. 2012;37:
uer, Laura Bellaire, Randy Betz, Craig Birch, Laurel E708–E713.
3. Hägglund G, Pettersson K, Czuba T, et al. Incidence of scoliosis in
Blakemore, Oheneba Boachie-Adjei, Chris Bonfield, cerebral palsy: a population-based study of 962 young individuals.
Daniel Bouton, Felix Brassard, Douglas Brockmeyer, Acta Orthop. 2018;89:443–447.
Jaysson Brooks, David Bumpass, Pat Cahill, Olivier 4. Allam AM, Schwabe AL. Neuromuscular scoliosis. PM&R. 2013;5:
Chemaly, Jason Cheung, Kenneth Cheung, Robert Cho, 957–963.
5. Olafsson Y, Saraste H, Al-Dabbagh Z. Brace treatment in neuro-
Tyler Christman, Eduardo C. Beauchamp, Daniel Cou-
muscular spine deformity. J Pediatr Orthop. 1999;19:376–379.
ture, Haemish Crawford, Alvin Crawford, Benny Dahl, 6. Chandran S, McCarthy J, Noonan K, et al. Early treatment of
Gokhan Demirkiran, Dennis Devito, Mohammad Diab, scoliosis with growing rods in children with severe spinal muscular
Hazem El Sebaie, Ron El-Hawary, John Emans, Mark atrophy: a preliminary report. J Pediatr Orthop. 2011;31:450–454.

7. Rumalla K, Yarbrough CK, Pugely AJ, et al. Spinal fusion for 13. Morrissy RT, Goldsdmith GS, Hall EC, et al. Measurement of the
pediatric neuromuscular scoliosis: national trends, complications, Cobb angle on radiographs of patients who have scoliosis. Evaluation
and in-hospital outcomes. J Neurosurg Spine. 2016;25:500–508. of intrinsic error. J Bone Joint Surg Am. 1990;72:320–327.
8. Nielsen E, Andras LM, Bellaire LL, et al. Don’t you wish you had 14. Carman DL, Browne RH, Birch JG. Measurement of scoliosis and
fused to the pelvis the first time: a comparison of reopera- kyphosis radiographs. Intraobserver and interobserver variation. J
tion rate and correction of pelvic obliquity. Spine. 2019;44: Bone Joint Surg Am. 1990;72:328–333.
E465–E469. 15. Ramo BA, Roberts DW, Tuason D, et al. Surgical site infections
9. Myung KS, Lee C, Skaggs DL. Early pelvic fixation failure in after posterior spinal fusion for neuromuscular scoliosis: a thirty-year
neuromuscular scoliosis. J Pediatr Orthop. 2015;35:258–265. experience at a single institution. J Bone Joint Surg Am. 2014;96:
10. Takaso M, Nakazawa T, Imura T, et al. Segmental pedicle screw 2038–2048.
instrumentation and fusion only to L5 in the surgical treatment of 16. McCall RE, Hayes B. Long-term outcome in neuromuscular
flaccid neuromuscular scoliosis. Spine. 2018;43:331–338. scoliosis fused only to lumbar 5. Spine. 2005;30:2056–2060.
11. Li Y, Swallow J, Gagnier J, et al. Pediatric Spine Study Group. 17. Tsirikos AI, Chang WN, Shah SA, et al. Preserving ambulatory
Pelvic fixation is not always necessary in children with cerebral palsy potential in pediatric patients with cerebral palsy who undergo spinal
scoliosis treated with growth-friendly instrumentation. Spine Deform. fusion using unit rod instrumentation. Spine. 2003;28:480–483.
2022;10:925–932. 18. Hell AK, Braunschweig L, Behrend J, et al. Health-related quality of
12. Matsumoto H, Williams B, Park HY, et al. The final 24-item early life in early-onset-scoliosis patients treated with growth-friendly
onset scoliosis questionnaires (EOSQ-24): validity, reliability and implants is influenced by etiology, complication rate and ambulatory
responsiveness. J Pediatr Orthop. 2018;38:144–151. ability. BMC Musculoskelet Disord. 2019;20:1–6.
on 04/15/2023

ORIGINAL ARTICLE
Do Osteochondroplasty Alone, Intertrochanteric

Derotation Osteotomy, and Flexion-Derotation Osteotomy
Improve Hip Flexion and Internal Rotation to Normal

Range in Hips With Severe SCFE? - A 3D-CT
Simulation Study
Till D. Lerch, MD, PhD,*† Young-Jo Kim, MD, PhD,† Ata Kiapour, PhD,†
Simon D. Steppacher, MD,‡ Adam Boschung, MD,‡§ Moritz Tannast, MD,‡§
Klaus A. Siebenrock, MD,‡ and Eduardo N. Novais, MD†
on 04/15/2023
treatment was performed before CT. Specific collision detection

Background: Severe slipped capital femoral epiphysis (SCFE) software was used for the calculation of impingement-free
leads to femoroacetabular impingement and restricted hip flexion and IR in 90 degrees of flexion and simulation of os-
motion. We investigated the improvement of impingement-free teochondroplasty, derotation osteotomy, and combined flex-
flexion and internal rotation (IR) in 90 degrees of flexion ion-derotation osteotomy.
following a simulated osteochondroplasty, a derotation
Results: Osteochondroplasty alone improved impingement-free
osteotomy, and a combined flexion-derotation osteotomy in
motion but compared with the uninvolved contralateral control
severe SCFE patients using 3D-CT-based collision detection
group, severe SCFE hips had persistently significantly
software.
decreased motion (mean flexion 59 ± 32 degrees vs. 122 ± 9
Methods: Preoperative pelvic CT of 18 untreated patients degrees, P < 0.001; mean IR in 90 degrees of flexion −5 ± 14
(21 hips) with severe SCFE (slip-angle > 60 degrees) was used degrees vs. 36 ± 11 degrees, P < 0.001). Similarly, the impinge-
to generate patient-specific 3D models. The contralateral hips ment-free motion was improved after derotation osteotomy,
of the 15 patients with unilateral SCFE served as the control and impingement-free flexion after a 30 degrees derotation was
group. There were 14 male hips (mean age 13 ± 2 y). No equivalent to the control group (113 ± 42 degrees vs. 122 ± 9
degrees, P = 0.052). However, even after the 30 degrees der-
From the *Department of Diagnostic, Interventional and Pediatric otation, the impingement-free IR in 90 degrees of flexion
Radiology, Inselspital, Bern University Hospital; ‡Department of
Orthopedic Surgery, Inselspital, University of Bern, Bern; §Depart- persisted lower (13 ± 15 degrees vs. 36 ± 11 degrees, P < 0.001).
ment of Orthopaedic Surgery, HFR Fribourg, University of Fri- Following the simulation of flexion-derotation osteotomy,
bourg, Fribourg, Switzerland; and †Department of Orthopedic mean impingement-free flexion and IR in 90 degrees of flexion
Surgery, Child and Young Adult Hip Preservation Program at Bos- increased for combined correction of 20 degrees (20 degrees
ton Children’s Hospital, Harvard Medical School, Boston, MA.
Till Lerch has received funding from the Swiss National Science Foun- flexion and 20 degrees derotation) and 30 degrees (30 degrees
dation (Early Postdoc.Mobility P2BEP3_195241). flexion and 30 degrees derotation). Although mean flexion was
Each author certifies that his or her institution approved the human equivalent to the control group for both (20 degrees and 30
protocol for this investigation, that all investigations were conducted degrees) combined correction, the mean IR in 90 degrees of
in conformity with ethical principles of research, and that informed flexion persisted decreased, even after the 30 degrees combined
consent for participation in the study was obtained.
The authors declare no conflicts of interest. flexion-derotation (22 ± 22 degrees vs. 36 degrees ± 11,
Reprints: Till Lerch, MD, Department of Diagnostic, Interventional and P = 0.009).
Pediatric Radiology, University of Bern, Inselspital, Bern University Conclusions: Simulation of derotation-osteotomy (30 degrees
Hospital, Bern, Switzerland, Freiburgstrasse, 3010 Bern, Switzerland.
correction) and flexion-derotation-osteotomy (20 degrees cor-
E-mails: till.lerch@insel.ch; till.lerch@childrens.harvard.edu.
Supplemental Digital Content is available for this article. Direct URL rection) normalized hip flexion for severe SCFE patients, but IR
citations appear in the printed text and are provided in the HTML in 90 degrees of flexion persisted slightly lower despite significant
and PDF versions of this article on the journal’s website, www. improvement. Not all SCFE patients had improved hip motion
pedorthopaedics.com. with the performed simulations; therefore, some patients may
Copyright © 2023 The Author(s). Published by Wolters Kluwer Health,
Inc. This is an open access article distributed under the terms of the need a higher degree of correction or combined treatment with
Creative Commons Attribution-Non Commercial-No Derivatives osteotomy and cam-resection, although not directly investigated
License 4.0 (CCBY-NC-ND), where it is permissible to download and in this study. Patient-specific 3D-models could help individual
share the work provided it is properly cited. The work cannot be preoperative planning for severe SCFE patients to normalize the
changed in any way or used commercially without permission from
hip motion.
the journal.
DOI: 10.1097/BPO.0000000000002371 Level of Evidence: III, case-control study.
286 | www.pedorthopaedics.com J Pediatr Orthop Volume 43, Number 5, May/June 2023

J Pediatr Orthop Volume 43, Number 5, May/June 2023 Simulation of treatment for Severe SCFE
Key Words: hip, SCFE, slipped capital femoral epiphysis a derotation-osteotomy, and a combined flexion-derotation
femoral derotation osteotomy, femoroacetabular impingement, osteotomy in hips with untreated severe SCFE using 3D-CT
hip preservation surgery and collision detection software.
(J Pediatr Orthop 2023;43:286–293)
METHODS
Study Population
I n slipped capital femoral epiphysis (SCFE), shearing
forces surpass the stability of the proximal femoral
physis, causing the metaphysis to rotate on the epiphyseal
After obtaining institutional review board approval,
we identified 123 patients treated for SCFE who underwent
a pelvic CT between 1998 and 2016. Out of 123 patients
tubercle as a fulcrum, resulting in an extension and a with CT scans, we excluded 105 patients due to a mild or
retroversion deformity.1,2 The resulting deformity is clas- moderate SCFE, postoperative CT, and those with CT that
sified based on the head-shaft angle as mild (< 30 degrees), did not involve the femoral condyles. The study cohort
moderate (30 degrees to 60 degrees), and severe SCFE comprised 18 patients with severe SCFE (slip angle > 60
( > 60 degrees).3 Previous studies suggested that treatment degrees), and preoperative CT was used to generate patient-
with in-situ fixation is associated with good Iowa hip specific 3D models. Of the 18 patients, no previous treat-
scores in the long term.4,5 However, the severity of SCFE ment was noted. Three patients (15%) had bilateral SCFE
deformity is directly related to the development of symp- yielding a total of 21 hips evaluated. The contralateral hips
toms and early hip osteoarthritis.6 SCFE deformity leads of the 15 patients with unilateral SCFE were used as a
on 04/15/2023
to femoroacetabular impingement (FAI), resulting in control group. The mean age was 13 years (SD, ± 2 y,
articular cartilage damage and posing a risk for long-term Table 1). There were 14 male hips, and the mean body mass
osteoarthritis.7–11 Hence, the goals of surgical treatment of index (BMI) was 28 ± 5. Seventeen (81%) hips were further
SCFE have expanded from the short-term perspective, classified as chronic ( > 3 wk of symptoms), while the
focusing on stabilizing the physis and preventing further remaining were considered acute on chronic ( > 3 wk of
slip, towards a long-term goal to prevent secondary FAI prodrome symptoms but with acute exacerbation in the last
and further osteoarthritis.12 3 wk). Eighteen out of the 21 hips (86%) were stable, and 3
Hip arthroscopy is the mainstay of surgical treatment hips were unstable slips, according to the Loder et al34
of cam-type FAI, and the role of arthroscopic cam resection classification system (Table 1).
of the femoral head-neck junction has been expanded to the
treatment of symptomatic residual deformity associated Imaging, Bone Segmentation, and 3D Modeling
with SCFE.13–16 However, arthroscopic treatment may be All patients underwent standardized AP, lateral
limited to restoring the normal range of motion in hips with radiographs, and CT scans, including the bilateral hip
severe SCFE.17 Several femoral osteotomies have been joint and the distal femoral condyles. CT scan was
described to improve the alignment of the proximal femur performed to assess SCFE severity, to measure the
in hips with residual deformity due to severe SCFE, femoral version, and for surgical planning. Following
including the Imhauser flexion-derotational intertrochan-
teric osteotomy,18 the Southwick triplane proximal
osteotomy,3 an osteotomy at the base of the femoral TABLE 1. Demographic Information of the Patient Series
neck,19,20 the femoral neck closing wedge osteotomy is Shown
according to Dunn21 and Fish,22 a simple diaphyseal Parameter Value
derotation osteotomy with intramedullary fixation.23 More Total hips (patients) 36 (18)
recently, the modified Dunn procedure using the surgical Total hips with severe SCFE (patients) 21 (18)
hip dislocation approach has gained popularity because of Total hips of asymptomatic controls (patients) 15 (15)
the possible complete restoration of the proximal femoral Age (years) 13 ± 2 (10–16)
anatomy and improved outcomes.24–27 However, this pro- Sex (% male of all hips) 10 hips, 48
Side (% left of all hips) 12 hips, 57
cedure is not widely available, and indications are discussed Height (cm) 166 ± 9 (152–179)
controversially. Therefore, proximal femoral osteotomies Weight (kg) 80 ± 12 (53–97)
still play an important role in treatment. Nevertheless, Body mass index (kg/m2) 28 ± 5 (22–36)
independent of the treatment used, deformity correction of BMI percentile 93
Unstable hips according to Loder classification 3 hips, 14
severe SCFE is challenging due to its 3-dimensional (% unstable of all hips)34
complexity.1,28–31 Although 2-dimensional radiographs are Severity based on slip Angle3 (% of all hips)
used to diagnose SCFE, the head-neck angle measurement, Mild <30 (deg) 0
and SCFE classification are affected by the patient’s posi- Moderate 30–60 (deg) 0
tion, and radiographs do not allow for a comprehensive Severe > 60 (deg) 21 hips, 100
Classification based on the duration of symptoms (% of all hips), 52
deformity analysis.32 Previous studies have described the Acute 0
application of 3D-CT for surgical planning.33 However, its Acute on chronic 4 hips, 19
application remains restricted due to radiation exposure. Chronic 17 hips, 81
This study investigates the improvement of impinge- Continuous values are expressed as mean ± SD, and range in parenthesis
ment-free motion following a simulated osteochondroplasty,
Copyright © 2023 The Author(s). Published by Wolters Kluwer Health, Inc. www.pedorthopaedics.com | 287
Lerch et al J Pediatr Orthop Volume 43, Number 5, May/June 2023
the segmentation of CT images, we built a 3D virtual

bone model (Fig. 1) of the pelvis and the femur for each
hip using the Amira Visualization Toolkit (Visage
Imaging Inc, Carlsbad, CA). Computer-assisted bone
segmentation was performed by 2 independent observers
not involved in the surgical care of the patients (TL and

AB).. All 3D models were available from a previous

study.35 Limited Hip Flexion and Internal Rotation
Resulting From Early Hip Impingement Conflict on
Anterior Metaphysis of Patients With Untreated Severe
SCFE Using 3D Modelling (TL and AB). The reference
coordinate system for the acetabulum was the anterior
pelvic plane (APP), defined by both anteroinferior iliac
spines and the pubic tubercles, and for the femur, it was
defined by the center of the femoral head, the knee
center, and both femoral condyles as previously
described.36
Simulation of Hip Impingement and Surgery

on 04/15/2023
The virtual 3D bone models (Fig. 1) were exported to

specific software for collision detection and quantification
of impingement-free hip range of motion. This software was
previously validated and allowed to simulate human hip
motion using CT scans of plastic and cadaveric hips.36 The
software uses the so-called equidistant method, a hip
simulation algorithm that accounts for a dynamic hip
joint center and allows to calculate the hip range of motion
with a higher linear and angular accuracy compared with
other methods.37 FIGURE 2. A femoral 3D model of a severe slipped capital
Hip flexion and internal rotation (IR) in 90 degrees femoral epiphysis patient after virtual cam resection is shown
of flexion were simulated using the CT-based 3D models, above. Below the resected bone volume is shown red trans-
and the impingement-free flexion and impingement-free parent.
IR in 90 degrees of flexion were recorded for the untreated
severe SCFE hips. Impingement-free motion was assessed The femoral 3D models were used to simulate
by the patient-specific initial point of impingement, osteochondroplasty of the femoral head-neck junction
defined as the amount of flexion and internal rotation (cam resection, Fig. 2 and figures in supplementary
recorded at initial detection of collision between the ace- material, Supplemental Digital Content 1, http://links.
tabular rim and the proximal femur. The effect of each lww.com/BPO/A583) as previously described by Ecker
specific surgical intervention on the impingement-free et al.38 computer-assisted femoral head-neck osteochon-
flexion and IR in 90 degrees of flexion was calculated and droplasty using a surgical milling device an in vitro ac-
compared with the impingement hip motion before the curacy study. Briefly, the software was used to plan and
intervention for the hips with severe SCFE and to the perform repeated computer-assisted osteochondroplasty
contralateral uninvolved hips as a control group. (cam resection, Fig. 2) using a virtual burr (surgical
reaming device, similar to the electric pen drive of Synthes
AG, Switzerland). The virtual sphere of the burr was used
in a stepwise fashion while the extent and depth of the
resection was visualized (red transparent in Fig. 2 and in
figures in supplementary material, Supplemental Digital
Content 1, http://links.lww.com/BPO/A583) with the goal
to improve the femoral head-neck offset. Simulation of
virtual surgeries was performed by a resident with 5 y
of experience in hip impingement simulation and 5 y of
experience in musculoskeletal imaging.
For the simulation of a derotation osteotomy
(Fig. 3A), a virtual intertrochanteric osteotomy was
performed using the software (figures in supplementary
material, Supplemental Digital Content 1, http://links.
FIGURE 1. A virtual CT-based 3D model of a severe slipped lww.com/BPO/A583). The landmarks and the reference
capital femoral epiphysis patient is shown. system used are the same as for the equidistant method.
288 | www.pedorthopaedics.com Copyright © 2023 The Author(s). Published by Wolters Kluwer Health, Inc.
of flexion, respectively. The flexion component of the

osteotomy was created by manipulating the distal femoral
segment in an anterior direction according to the desired
flexion correction, simulating an anterior wedge resection
(Fig. 3). We then compared the impingement-free range of
motion to the baseline measurements for each patient and to

the contralateral uninvolved hip (Fig. 5).

Statistical Analysis
Statistical analysis was performed using Winstat soft-
ware (R. Fitch software, Bad Krozingen, Germany). Normal
distribution was tested using the Kolmogorov-Smirnov test.
Because the data were not normally distributed, we only used
nonparametric tests. Continuous variables were compared
FIGURE 3. A-C, Schematic views of femoral derotation (A) and using the Friedman test and the Wilcoxon test because the
flexion-derotation-osteotomy (B and C) are shown. data were not normally distributed. For the range of motion
testing, the baseline (preoperative) range of motion was
After performing the virtual osteotomy perpendicular to compared with (1) cam resection, (2) derotation osteotomy
the femoral shaft axis, the distal femur is rotated inwards for 10 degrees correction, (3) 20 degrees correction, (4) der-
otation 30 degrees correction, (5) derotation, and flexion
on 04/15/2023
(medial, in the anterior direction of the contralateral

limb) to increase the femoral version. Three virtual 10 degrees, 6) derotation and flexion 20 degrees, and (7)
femoral derotation osteotomies were created for each derotation and flexion 30 degrees. Because of the 7 sub-
patient (1 model for 10 degrees 1 model for 20 degrees groups, the level of significance was adjusted with the Bon-
and 1 model for 30 degrees of correction, figures in ferroni correction after counselling a statistician. This is a
supplementary material, Supplemental Digital Content 1, simple method for adjustment. The level of significance was
http://links.lww.com/BPO/A583) as described by Stevens 0.05/7 = 0.0071. This means a P value below 0.0071 was
et al.23 Then, the postoperative 3D models were compared considered significant.
with the preoperative 3D models to calculate the improve-
ment in hip motion and to the control group (Fig. 4). RESULTS
Finally, we used the software to simulate a flexion The impingement-free motion was limited in the
derotation osteotomy as described by Imhauser.18 Using the hips with severe SCFE at baseline (Table 2). The mean
virtual femoral 3D model of the severe SCFE hips, the impingement-free flexion (46 ± 32 degrees vs. 122 ± 9
flexion derotation intertrochanteric osteotomy (Fig. 3) was degrees, P < 0.001) and mean impingement-free IR in 90
simulated with 10 degrees of derotation correction and 10 degrees of flexion (−17 ± 18 degrees vs. 36 ± 11 degrees,
degrees of flexion correction. A second and a third simulation P < 0.001) were significantly decreased in the severe SCFE
was performed with 20 degrees of derotation and 20 degrees hips compared with the contralateral control hips
of flexion and with 30 degrees of derotation and 30 degrees (Table 2).
FIGURE 4. Simulation of femoral derotation osteotomy to increase femoral version is shown for a severe slipped capital femoral epiphysis
patient. The red zone represents the impingement zone. No impingement was noted after 30 degrees of derotation.
(P < 0.001) with the 10 degrees combined correction but

improved after the 20 and 30 degrees combined correction.
(Table 3) When compared with the contralateral uninvolved
control hips, mean flexion was no different for the 20 degrees
combined correction (119 ± 45 degrees vs. 122 ± 9 degrees;
P = 0.052). However, the mean impingement-free IR in 90

degrees of flexion persisted decreased after the 30 degrees

combined flexion-derotation when compared with the
contralateral uninvolved control hips (22 ± 22 degrees vs.
36 degrees ± 11, P = 0.009).
DISCUSSION
We compared the improvement of impingement-free
FIGURE 5. Simulation of combined flexion and derotation flexion and IR in 90 degrees of flexion following simulated
osteotomy is shown for internal rotation in 90 degrees of osteochondroplasty of the femoral head-neck junction,
flexion.
femoral derotation-osteotomy, and combined flexion-
derotation-osteotomy in hips with severe SCFE (Fig. 3) to
Simulation of motion after osteochondroplasty the contralateral uninvolved hips using specific software for
compared with the baseline motion of the severe SCFE collision detection. Although osteochondroplasty improved
on 04/15/2023
hips showed improvement in impingement-free flexion the limited hip motion, the achieved range of impingement-
(59 ± 32 degrees, range 0 to 121 degrees vs. 46 ± 32 de- free motion was far from the normal hip motion simulated
grees; P < 0.001) and impingement-free IR in 90 degrees for the contralateral uninvolved hips. Similar findings were
of flexion (–5 ± 14 degrees, range −50 to 19 degrees vs. noted for improving impingement-free motion following a
−17 ± 18 degrees, P = 0.002). However, when compared derotation osteotomy of 10 and 20 degrees and even after a
with the contralateral control group, the hip motion of flexion-derotation osteotomy with 10 degrees combined
severe SCFE hips after osteochondroplasty (Fig. 2) was correction. A 30 degrees derotation osteotomy (Fig. 4)
significantly decreased (mean flexion 59 ± 32 degrees vs. improved hip flexion to no difference compared with the
122 ± 9 degrees, P < 0.001 and mean IR in 90 degrees of contralateral uninvolved hip, but with persistent limited IR in
flexion (−5 ± 14 degrees vs. 36 ± 11 degrees, P < 0.001). 90 degrees of flexion (Table 2).
After derotation-osteotomy (Fig. 4), mean impinge- We found that simulated osteochondroplasty of the
ment-free flexion increased significantly with 10, 20 and 30 femoral head and neck junction did not normalize
degrees of correction (P < 0.001, Table 2). However, mean impingement-free motion as compared with the contra-
impingement-free IRF-90 degrees was not improved after lateral uninvolved control hips. Along the same lines,
10 degrees correction but significantly (P < 0.001) improved Wylie et al17 compared the functional results of arthro-
to 3 ± 15 degrees and 13 ± 15 degrees (Table 2) with 20 and scopic treatment with osteochondroplasty versus open
30 degrees of simulated derotation (P < 0.001), respectively. surgical treatment through a surgical hip dislocation
The mean impingement-free flexion after a 30 degrees with or without a femoral osteotomy. They suggested
derotation was not different compared with the control arthroscopic osteochondroplasty to treat hips with mild
group (113 ± 42 degrees vs. 122 ± 9 degrees; P = 0.052). SCFE deformity and only slightly limited hip IR in 90
However, after the 30 degrees derotation, the impingement- degrees of flexion. Patients with severe SCFE deformity
free IRF-90 degrees persisted lower compared with with obligatory external rotation in flexion benefit from a
the control group (13 ± 15 degrees vs. 36 ± 11 degrees, flexion-derotation osteotomy to improve the range of
P < 0.001). impingement-free motion. Besomi et al14 reported on
Following the simulation of the flexion-derotation os- the results of hip arthroscopy treatment, including
teotomy (Fig. 5), mean impingement-free flexion increased osteochondroplasty in 17 patients with residual SCFE
significantly (P < 0.001) for the 10, 20, and 30 degrees of deformity, and found only 6 degrees of hip flexion and 14
combined correction (eg, to 138 ± 47 degrees for the 30 degrees of IR improvement. Balakumar et al37 reported a
degrees combined correction, Table 3). Mean impingement- less significant improvement of hip IR for patients with
free IR in 90 degrees of flexion did not improve significantly severe and moderate SCFE deformity treated with ar-
TABLE 2. Virtual Treatment Simulation of 21 Hips with Untreated Severe SCFE Using Preoperative 3D-CT
Parameter No treatment Derotation 10 (deg) Derotation 20 (deg) Derotation 30 (deg)
Flexion 46 ± 32 (0 to 113) 79 ± 44 (10 to 130)* 103 ± 43 (14 to 150)* 113 ± 42 (19 to 160)*
Internal rotation in 90 degrees of flexion (deg) −17 ± 18 (−60 to 10) −7 ± 15 (−53 to 15) 3 ± 15 (−43 to 25)* 13 ± 15 (−33 to 35)*
*signifies significant difference compared with no treatment, P values below 0.0071 were considered significant.
level of significance was adjusted with the Bonferroni correction to 0.05/7 = 0.0071.
Derotation osteotomy was simulated to increase femoral version.
TABLE 3. Virtual Treatment Simulation of 21 Hips with Untreated Severe SCFE Using Preoperative 3D-CT
Derotation 10 and Flexion Derotation 20 and Flexion Derotation 30 and Flexion
Parameter No treatment 10 (deg) 20 (deg) 30 (deg)
Flexion 46 ± 32 (0 to 113) 86 ± 45 (10 to 139)* 119 ± 45 (14 to 160)* 138 ± 47 (39 to 180)*
Internal rotation in 90 −17 ± 18 (−60 to 10) -5 ± 15 (−50 to 29) 11 ± 21 (−47 to 37)* 22 ± 22 (−40 to 49)*
degrees of flexion (deg)

*signifies significant difference compared with no treatment, P values below 0.0071 were considered significant.
level of significance was adjusted with the Bonferroni correction to 0.05/7 = 0.0071.
Flexion and derotation osteotomy was simulated to increase femoral version.
throscopic osteochondroplasty (mean improvement 20 analyzed 19 patients with moderate or severe SCFE and
degrees, range 0 to 20 degrees) compared with open reported lower values for hip flexion and IR (61 and 66
femoral neck osteotomy (mean improvement 50 degrees, degrees) after the simulation of a multiplanar femoral
range 30 to 70 degrees). Based on our findings and intertrochanteric osteotomy and the simulation of
previous studies, isolated osteochondroplasty improves uniplanar flexion osteotomy (63 and 54 degrees). Inter-
hip motion in hips with severe SCFE; the improvement, estingly, they found similar improvements in hip motion
however, is modest and insufficient to restore impinge- (apart from abduction) for both multiplanar and uni-
ment-free motion. planar femoral osteotomy. One advantage of the collision
on 04/15/2023
In this study, we observed significant improvement in detection software used in the current study is the ability
impingement-free flexion and IR in 90 degrees of flexion to determine the patient-specific impingement-free range
following an isolated derotation osteotomy of the femur of motion of the hip joint.
when we simulated a 30 degrees correction. Although cor- This study has several limitations. First, the software
recting femoral retroversion is a component of most fem- for collision detection calculates the osseous range of
oral osteotomies independent of the correction level, motion without considering soft tissue (labrum, ligaments,
applying a pure derotation osteotomy to the treatment of or cartilage). Therefore, we believe the clinical hip motion
FAI secondary to SCFE is relatively novel. Stevens et al23 may be even lower in these hips. Assessment of the soft
described a femoral midshaft 45 degrees derotation tissues limiting motion may be unavoidable using pelvic
osteotomy fixed with an intramedullary nail in 4 patients CT imaging, although it could be integrated using mag-
with severe residual SCFE deformity. They reported sig- netic resonance imaging in the future. However, previous
nificant improvement in hip flexion, IR, and gait analysis. collision detection studies used the same specific software
However, 2 patients required secondary arthroscopic to analyze hips with complex morphology, including
osteochondroplasty to alleviate residual FAI, and 1 patient post-Perthes disease, underlining the software’s validity
underwent a total hip replacement 62 months after the for collision detection in hips with severe deformity.
osteotomy. In addition, 2 patients had delayed union and Second, we only tested 3 types of surgical intervention
some loss of correction, secondary to broken interlocking (osteochondroplasty, femoral derotation osteotomy, and
screws, and required revision surgery with reaming and nail flexion-derotation osteotomy, figures in supplementary
exchange. Although our study confirms that a 30 degrees material, Supplemental Digital Content 1, http://links.
simple derotation osteotomy may normalize hip flexion, lww.com/BPO/A583). Other types of proximal femoral
future studies are necessary to determine the clinical and osteotomies could result in more significant improvement
functional outcomes of a pure derotation osteotomy and to in impingement-free motion. No translation was per-
define the specific preoperative criteria for the indication of formed for the flexion osteotomy. The rationale for
this procedure. Applying collision detection software may selecting the 3 specific surgical interventions was based on
enhance the ability of the treating surgeon to determine the the increased popularity of arthroscopic osteochon-
amount of retroversion correction during surgery. Further droplasty for treating hips with SCFE and the fact that
steps are needed to apply this software in clinical practice; femoral derotation osteotomy is a universally available
so far the software has been used for research purposes only and relatively low-demanding technique. Similarly, the
(figures in supplementary material, Supplemental Digital flexion-derotation osteotomy is a well-accepted procedure
Content 1, http://links.lww.com/BPO/A583). for treating SCFE deformity that is technically less
Simulation of hip range of motion and evaluation of demanding than some of the procedures we did not test
different techniques of femoral osteotomy to assess the (eg, the modified Dunn procedure). Third, our study
improvement in range of hip motion and the hip-joint focused on the simulated surgical procedure without as-
geometry is not a novel concept, but the literature is sessing the actual results of such procedures. No post-
scarce.31,40,41 One previous study evaluated 3D models of operative alpha angles after cam resection were calculated
11 hips with severe SCFE and reported lower values for because of the severe displacement of the epiphysis and the
flexion and IR compared with the current study.31 femoral head center. Therefore, we lack information
Mamisch et al41 simulated the effect of a multiplanar about those procedures’ complications and clinical
Southwick3 intertrochanteric osteotomy with flexion, val- improvement. Although comparing patient-specific inter-
gus, and IR to a uniplanar purely flexion osteotomy. They ventions and outcomes was not the goal of our study, all
patients were symptomatic at the time of image acquis- 4. Boyer DW, Mickelson MR. Ponseti IV. Slipped capital femoral
ition, and most of them underwent surgical treatment. epiphysis. Long-term follow-up study of one hundred and twenty-one
patients. J Bone Joint Surg Am. 1981;63:85–95.
Therefore, future studies should investigate the simulation 5. Carney BT, Weinstein SL, Noble J. Long-term follow-up of slipped
of additional isolated or combined hip preservation capital femoral epiphysis. J Bone Joint Surg Am. 1991;73:667–674.
procedures to investigate clinical and functional 6. Castaneda P, Ponce C, Villareal G, et al. The natural history of
improvements of such procedures to establish the clinical osteoarthritis after a slipped capital femoral epiphysis/the pistol grip
application of the software in the clinical setting. A com- deformity. J Pediatr Orthop. 2013;33(Suppl 1):S76–S82.
7. Fraitzl CR, Kafer W, Nelitz M, et al. Radiological evidence of
bination of femoral osteotomy and cam resection was not femoroacetabular impingement in mild slipped capital femoral
studied and could be an additional treatment approach for epiphysis: a mean follow-up of 14.4 years after pinning in situ. J
severe SCFE patients. Finally, our study was limited to Bone Joint Surg Br. 2007;89:1592–1596.
hips with severe SCFE, and we cannot make any inference 8. Leunig M, Casillas MM, Hamlet M, et al. Slipped capital femoral
about the effects of the investigated procedures on the epiphysis: early mechanical damage to the acetabular cartilage by a
prominent femoral metaphysis. Acta Orthop Scand. 2000;71:
improvement of hip motion for hips with mild and 370–375.
moderate SCFE. 9. Sink EL, Zaltz I, Heare T, et al. Acetabular cartilage and labral
Preoperative planning is crucial before surgical cor- damage observed during surgical hip dislocation for stable slipped
rections in patients with severe SCFE. The severity of the capital femoral epiphysis. J Pediatr Orthop. 2010;30:26–30.
femoral deformity and the morphology of the acetabulum 10. Ziebarth K, Leunig M, Slongo T, et al. Slipped capital femoral
epiphysis: relevant pathophysiological findings with open surgery.
vary between patients resulting in complex FAI. Although Clin Orthop Relat Res. 2013;471:2156–2162.
femoral derotation osteotomy (30 degrees correction) or 11. Lieberman EG, Pascual-Garrido C, Abu-Amer W, et al. Patients
combined flexion and derotation osteotomy (20 degrees
on 04/15/2023
with symptomatic sequelae of slipped capital femoral epiphysis have

correction) enabled the restoration of impingement-free advanced cartilage wear at the time of surgical intervention. J Pediatr
flexion, similar to the hip flexion of the control group, IR Orthop. 2021;41:e398–e403.
12. Millis MB, Novais EN. In situ fixation for slipped capital femoral epiphysis:
in 90 degrees of flexion remained lower than the control perspectives in 2011. J Bone Joint Surg Am. 2011;93(Suppl 2):46–51.
group, despite significant improvement. Furthermore, not 13. Saito M, Kuroda Y, Sunil Kumar KH, et al. Outcomes after
all severe SCFE hips achieved normal motion after the arthroscopic osteochondroplasty for femoroacetabular impingement
simulation of derotational and flexion-derotation osteot- secondary to slipped capital femoral epiphysis: A systematic review.
Arthroscopy. 2021;37:1973–1982.
omy (Figs. 3 and 4). Patient-specific 3D models and
14. Besomi J, Escobar V, Alvarez S, et al. Hip arthroscopy following
virtual surgical simulation provide a unique opportunity slipped capital femoral epiphysis fixation: chondral damage and labral
to understand the deformity better and determine the tears findings. J Child Orthop. 2021;15:24–34.
personalized bony correction needed to optimize 15. Basheer SZ, Cooper AP, Maheshwari R, et al. Arthroscopic
impingement-free motion in patients with severe SCFE, treatment of femoroacetabular impingement following slipped
capital femoral epiphysis. . Bone Joint J. 2016;98-B:21–27.
but future studies should determine whether preoperative 16. Wylie JD, Beckmann JT, Maak TG, et al. Arthroscopic treatment of
simulation and planning impact patient-specific symptoms mild to moderate deformity after slipped capital femoral epiphysis:
and hip function. intra-operative findings and functional outcomes. Arthroscopy.
2015;31:247–253.
17. Wylie JD, McClincy MP, Uppal N, et al. Surgical treatment of
CONCLUSION symptomatic post-slipped capital femoral epiphysis deformity: a
Simulation of derotation-osteotomy (30 degrees comparative study between hip arthroscopy and surgical hip
correction) and flexion-derotation-osteotomy (20 degrees dislocation with or without intertrochanteric osteotomy. J Child
combined correction) normalized hip flexion for severe Orthop. 2020;14:98–105.
18. Imhauser G. [Late results of Imhauser’s osteotomy for slipped capital
SCFE patients, but IR in 90 degrees of flexion persisted femoral epiphysis (author’s transl)]. Z Orthop Ihre Grenzgeb.
slightly lower despite significant improvement. Not all 1977;115:716–725.
SCFE patients had improved hip motion with the per- 19. Barmada R, Bruch RF, Gimbel JS, et al. Base of the neck
formed simulations; therefore, some patients may need a extracapsular osteotomy for correction of deformity in slipped
higher degree of correction or combined treatment with capital femoral epiphysis. Clin Orthop Relat Res. 1978;132:98–101.
20. Kramer WG, Craig WA, Noel S. Compensating osteotomy at the
osteotomy and cam-resection, although not directly in- base of the femoral neck for slipped capital femoral epiphysis. J Bone
vestigated in this study. Patient-specific 3D models could Joint Surg Am. 1976;58:796–800.
help individual preoperative planning for severe SCFE 21. Dunn DM. The treatment of adolescent slipping of the upper femoral
patients to normalize hip motion. epiphysis. J Bone Joint Surg Br. 1964;46:621–629.
22. Fish JB. Cuneiform osteotomy of the femoral neck in the treatment
of slipped capital femoral epiphysis. J Bone Joint Surg Am. 1984;66:
REFERENCES 1153–1168.
1. Novais EN, Hosseinzadeh S, Emami SA, et al. What is the association 23. Stevens PM, Anderson L, MacWilliams BA. Femoral shaft
among epiphyseal rotation, translation, and the morphology of the osteotomy for obligate outward rotation due to SCFE. Strategies
epiphysis and metaphysis in slipped capital femoral epiphysis? Clin Trauma Limb Reconstr. 2017;12:27–33.
Orthop Relat Res. 2021;479:935–944. 24. Ganz R, Gill TJ, Gautier E, et al. Surgical dislocation of the adult
2. Maranho DA, Bixby S, Miller PE, et al. A novel classification system hip a technique with full access to the femoral head and acetabulum
for slipped capital femoral epiphysis based on the radiographic without the risk of avascular necrosis. J Bone Joint Surg Br. 2001;83:
relationship of the epiphyseal tubercle and the metaphyseal socket. 1119–1124.
JB JS Open Access. 2019;4:e0033. 25. Ganz R, Huff TW, Leunig M. Extended retinacular soft-tissue flap
3. Southwick WO. Osteotomy through the lesser trochanter for slipped for intra-articular hip surgery: surgical technique, indications, and
capital femoral epiphysis. J Bone Joint Surg Am. 1967;49:807–835. results of application. Instr Course Lect. 2009;58:241–255.
26. Leunig M, Slongo T, Kleinschmidt M, et al. Subcapital correction 35. Lerch TD, Kim YJ, Kiapour AM, et al. Limited Hip Flexion and
osteotomy in slipped capital femoral epiphysis by means of surgical Internal Rotation Resulting From Early Hip Impingement Conflict
hip dislocation. Oper Orthop Traumatol. 2007;19:389–410. on Anterior Metaphysis of Patients With Untreated Severe SCFE
27. Novais EN, Hill MK, Carry PM, et al. Modified Dunn Procedure is Using 3D Modelling. J Pediatr Orthop. 2022;42:e963–e970.
superior to in situ pinning for short-term clinical and radiographic 36. Tannast M, Kubiak-Langer M, Langlotz F, et al. Noninvasive three-
improvement in severe stable SCFE. Clin Orthop Relat Res. dimensional assessment of femoroacetabular impingement. J Orthop
2015;473:2108–2117. Res. 2007;25:122–131.

28. Allen MM, Rosenfeld SB. Treatment for post-slipped capital femoral 37. Puls M, Ecker TM, Tannast M, et al. The Equidistant Method - a
epiphysis deformity. Orthop Clin North Am. 2020;51:37–53. novel hip joint simulation algorithm for detection of femoroacetabular
29. Bland DC, Valdovino AG, Jeffords ME, et al. Evaluation of the impingement. Comput Aided Surg. 2010;15:75–82.
three-dimensional translational and angular deformity in slipped 38. Ecker TM, Puls M, Steppacher SD, et al. Computer-assisted femoral
capital femoral epiphysis. J Orthop Res. 2020;38:1081–1088. head-neck osteochondroplasty using a surgical milling device an in
30. Kordelle J, Millis M, Jolesz FA, et al. Three-dimensional analysis of vitro accuracy study. J Arthroplasty. 2012;27:310–6.
the proximal femur in patients with slipped capital femoral epiphysis 39. Balakumar B, Flatt E, Madan S. Moderate and severe SCFE
based on computed tomography. J Pediatr Orthop. 2001;21:179–182. (Slipped Capital Femoral Epiphysis) arthroscopic osteoplasty vs
31. Mamisch TC, Kim YJ, Richolt JA, et al. Femoral morphology due open neck osteotomy-a retrospective analysis of results. Int Orthop.
to impingement influences the range of motion in slipped capital 2019;43:2375–2382.
femoral epiphysis. Clin Orthop Relat Res. 2009;467:692–698. 40. Kordelle J, Mamisch C, Kikinis R, et al. Anatomical analysis and
32. Jones CE, Cooper AP, Doucette J, et al. Southwick angle measure- preoperative planning of correctional osteotomies: slipped capital
ments and SCFE slip severity classifications are affected by frog- femoral epiphysis (SCFE). Minim Invasive Ther Allied Technol.
lateral positioning. Skeletal Radiol. 2018;47:79–84. 2000;9:269–276.
33. Cherkasskiy L, Caffrey JP, Szewczyk AF, et al. Patient-specific 3D 41. Mamisch TC, Kim YJ, Richolt J, et al. Range of motion after
models aid planning for triplane proximal femoral osteotomy in computed tomography-based simulation of intertrochanteric correc-
on 04/15/2023
slipped capital femoral epiphysis. J Child Orthop. 2017;11:147–153. tive osteotomy in cases of slipped capital femoral epiphysis:
34. Loder RT, Richards BS, Shapiro PS, et al. Acute slipped capital comparison of uniplanar flexion osteotomy and multiplanar flexion,
femoral epiphysis: the importance of physeal stability. J Bone Joint valgisation, and rotational osteotomies. J Pediatr Orthop. 2009;29:
Surg Am. 1993;75:1134–1140. 336–340.
ORIGINAL ARTICLE
Early Complications After Posterior Spinal Fusion

in Patients With Rett Syndrome
Tristen N. Taylor, BS, Callie S. Bridges, BS, Luke A. Nordstrom, BS, Darrell S. Hanson, MD,
Frank T. Gerow, MD, and Brian G. Smith, MD
Level of Evidence: Level IV—therapeutic study.

Background: Neuromuscular scoliosis in Rett syndrome (RS) is
common, progressive, and often requires posterior spinal fusion Key Words: Rett syndrome, scoliosis, posterior spinal fusion,
(PSF). While PSF is associated with improved overall outcomes, complications
there is a paucity of information describing complications. We (J Pediatr Orthop 2023;43:e326–e330)
aimed to report the postoperative complications, readmissions,
and reoperations for patients with RS undergoing PSF.
on 04/15/2023
Methods: Female pediatric patients with RS treated by PSF with

segmental instrumentation, with or without concurrent pelvis
fixation, during January 2012 to August 2022 were included.
Preoperative patient characteristics, intraoperative data (esti-
R ett syndrome (RS) is an X-linked dominant disorder
associated with mutations in the methyl-CpG-binding
protein 2 (MECP2).1 This disorder is classically present in
mated blood loss, cell saver, packed red blood cells transfused), females and is characterized by normal early development
postoperative complications according to the Modified Clavien- followed by sudden and progressive neurodevelopmental
Dindo-Sink classification within 90 days, unplanned read- regression, stereotypic hand movements, cognitive and
missions within 30 days, and unplanned reoperations within communication impairment, epilepsy, and neuromuscular
90 days were recorded. scoliosis.2
Results: A total of 25 females were included. The mean (SD) age By age 13, 80% of RS patients have measurable
at surgery was 12.9 (1.8) years and the mean follow-up of 38.6 scoliosis, and 70% of those patients undergo surgical
(24.9) months. The mean preoperative major coronal curve was correction.2 Kerr et al3 found that correction is associated
79 degrees (23 degrees) which decreased to 32 degrees (15 de- with improved overall outcomes in at least 84% of RS
grees) by the last follow-up (P < 0.001). The median estimated patients, and Downs et al4 found a strong protective effect
blood loss was 600 mL and length of stay was 7 days. There were of posterior spinal fusion (PSF) on overall mortality in RS
81 total postoperative complications (3.2 complications/patient). patients. However, this surgery has also been associated
Eight (32%) had grade IVa complications (disseminated intra- with a high risk of pulmonary complications, prolonged
vascular coagulopathy, hypotensive shock, respiratory failure, length of stay (LOS), and wound infection.5–7 Despite the
chronic urosepsis). Five (20%) patients experienced seizures, 48% prominence of scoliosis in RS and the need for surgical
had pulmonary complications, and 56% had gastrointestinal correction, the literature is limited by the few studies
complications. There were 3 readmissions (12%) within 30 days evaluating postoperative complications, small sample
for pneumonia and 2 (8%) reoperations (an incision and drainage sizes, or nonstandard reporting methods.6–8 A lack of
and C2-T2 fusion for significant kyphosis) within 90 days. One detailed and consistent information on the complications
patient also had their fusion extended to the pelvis 1 year later. experienced by these patients hinders the potential for
There were more nonambulatory patients in the group fused to surgeons to adequately counsel families on risks and
the pelvis, but otherwise no differences between those fused and communicate with other members of the perioperative
unfused to the pelvis. care team.
Conclusions: This is the largest review of early postoperative In 1992, Clavien et al9 described complications as
complications for patients with RS who underwent PSF. PSF unexpected events not intrinsic to the procedure and
effectively reduced the major coronal curve, but surgeons and proposed a classification system that did not rely on
families should be aware of a high postoperative seizure and “major” or “minor” as outcome measurements, thereby
respiratory complication rate, as well as 8% having reoperations reducing subjectivity. The modified Clavien-Dindo-Sink
within 90 days and 12% being readmitted within 30 days. (CDS) classification has been validated to describe com-
plications after spinal fusion for children with adolescent
From the Texas Children’s Hospital, Houston, TX. idiopathic scoliosis and cerebral palsy (CP).10–13 Given the
This study was internally funded. sparse and inconsistent data on early RS complications
The authors declare no conflicts of interest. after PSF, our primary objective was to describe post-
Reprints: Tristen N. Taylor, BS, 1 Hermann Museum Circle Drive, APT
4037, Houston, TX 77004. E-mail: tristen.taylor@bcm.edu. operative complications, readmissions, and reoperations
Copyright © 2023 Wolters Kluwer Health, Inc. All rights reserved. using the modified CDS classification system. Secondarily,
DOI: 10.1097/BPO.0000000000002384 we also described preoperative and intraoperative

J Pediatr Orthop Volume 43, Number 5, May/June 2023 Early Complications After PSF in Patients With RS
characteristics, and compared patients fused and not fused with a mean (SD) current age of 17.97 (3.8) years, mean
to the pelvis. age at surgery of 12.9 (1.8) years, and mean follow-up of
38.6 (24.9) months (range: 7 to 94 mo) (Table 1). Twenty
METHODS patients (80%) were fused to the pelvis from T2 to T4.
Only 3 (12%) patients were ambulatory without assistance
After institutional review board approval, we con-
preoperatively. The mean preoperative major angle was 79

ducted a retrospective review of female patients with a

degrees (23 degrees) which decreased to 32 degrees (15
diagnosis of RS who were 18 years old or younger and
degrees) by the last follow-up (P < 0.001). Median EBL
underwent PSF for scoliosis correction between January
was 600 mL (200 to 2500 mL) and the mean pRBC
2012 and August 2022. Patients were excluded if they had
transfused was 234 (334) mL [ie, 0.7 (0.96) U of pRBC].
a follow-up of <6 months, prior spinal fusion surgeries,
The median LOS was 7 days (range: 3 to 20 d).
insufficient data, or received a MAGnetic expansion
control rod for fixation.
The following data were collected for each patient: Complications Overview
preoperative demographic characteristics (current age, age Despite no intraoperative complications, this cohort
at surgery, race, ambulatory status, follow-up time), experienced 81 postoperative complications (3.2 per pa-
imaging data (major coronal angles preoperatively and by tient) within 90 days (Table 2). Eight (32%) patients had
last follow-up), intraoperative parameters [estimated CDS grade IVa complications, 2 patients (8%) had grade
blood loss (EBL), complications, surgical time, cell saver III complications, 16 patients (64%) had grade II
and packed red blood cells (pRBC) volume transfused, complications, and 21 patients (84%) had grade I
on 04/15/2023
highest and lowest vertebrae fused], LOS, readmissions complications (Table 3). There were no grade IVb (grade
within 30 days of discharge, total reoperations, and re- IVa, but with long-term disability) or grade V (death)
operations within 90 days. complications. There was no significant difference between
Complications that required treatment in some form each surgeon and complications per patient by CDS grade
were compiled from every patient note after surgery severity (P = 0.84). Overall, there were 7 different
90 days postoperation and then categorized according to nosocomial infections among 4 patients (16%), 3
the modified CDS classification. Complications were readmissions within 30 days, and 2 reoperations within
categorized by a consensus of authors and then reviewed 90 days of discharge
for accuracy by the senior author. All data analysis was
performed using RStudio, version 2022.07.0 (R Core Complication Descriptions
Team: Vienna, Austria, 2014). Descriptive statistics were Two patients developed disseminated intravascular
reported as counts (%), mean with SD, or median with coagulopathy and were treated with fresh-frozen plasma.
range. Continuous data was compared with the Student t Two patients developed postoperative hypotensive shock
test, and categorical data was compared with χ2 or Fisher and were treated with vasopressors. Two patients had
exact test. P-values <0.05 were considered significant. respiratory failure requiring prolonged ventilation
( > 48 h) or reintubation. Patient #7 (Table 3) developed
Surgical Technique catheter-associated Pseudomonas urosepsis requiring long-
Five fellowship-trained surgeons with an average term antibiotic therapy.
27 years in practice performed standard PSF with seg- Five (20%) patients had postoperative seizures.
mental instrumentation (Fig. 1) on 32 patients.14 The Thirteen (52%) had postoperative transfusions, 2 patients
operating surgeon determined the starting and end (8%) had superficial wound dehiscence, and patient #25
ing vertebrae for the surgery and whether pelvic developed a sacral pressure ulcer. Twelve (48%) patients
instrumentation was necessary15 based on preoperative had respiratory complications: 8 patients (32%) developed
pelvic obliquity, kyphosis, and ambulatory status. atelectasis or hypoxia, which required noninvasive mech-
Fluoroscopy or intraoperative computed tomography anical ventilation in 4 patients (16%). Gastrointestinal
scans were used to ensure proper placement of hardware. complications were all grade I, present in 14 patients
Sacropelvic fixation was performed after exposing the (56%), and managed with antiemetics, nasogastric tubes,
posterior spine and sacral foramen, followed by inferior or stool softeners. Patient #5 had a significant loss in
facetectomies and bone grafting. The rods were then fixed ambulation from baseline and required inpatient re-
to the pedicle screws. All patients received antibiotic- habilitation. Patient #21 developed aspiration pneumonia
infused auto/allografts after decortication. The patients and was treated with intravenous antibiotics.
were admitted to the pediatric intensive care unit (ICU) and
closely monitored for complications postoperatively. Readmissions and Reoperations
There were 3 readmissions (12%) within 30 days of
RESULTS discharge. Patient #19 was readmitted to the pediatric
Of the 32 patients identified, 6 were excluded due to ICU 1 day after discharge for parainfluenza pneumonia
lack of follow-up or improper diagnosis. One patient was and pleural effusion that required a chest tube placement.
fixed with a MAGnetic expansion control rod and was Patient #11 was readmitted for systemic inflammatory
excluded. Thus, this study included 25 female patients response syndrome that required broad-spectrum IV

Taylor et al J Pediatr Orthop Volume 43, Number 5, May/June 2023
TABLE 2. Postoperative Complications*

Complication No. patients [n (%)]
Neurological (ie, seizure) 5 (20)
Cardiovascular/hematologic
Postoperative transfusions 13 (52)

Abnormal vitals (eg, fever, tachycardia, 10 (40)

hypotension)
Disseminated intravascular coagulopathy 2 (8)
Hypotensive shock 2 (8)
Pulmonary
Atelectasis/hypoxia 8 (32)
Respiratory failure requiring mechanical 2 (8)
ventilation
Gastrointestinal (eg, ileus, dysphagia, emesis) 14 (56)
Renal/GU (ie, urinary retention) 1 (4)
Infection
Pulmonary infections 4 (16)
Urinary tract infections 2 (8)
Wound infections 1 (4)
MSK
FIGURE 1. A, A 12-year-old female patient with Rett syndrome Spasm, gait disturbance, chronic pain 4 (16)
with a 115-degree major coronal curve and significant pelvic Need for reoperation 3 (12)
on 04/15/2023
Wound complications (eg, dehiscence, ulcer) 3 (12)

obliquity. B, The same patient 2 years after posterior spinal
Reoperation within 90 d 2 (8)
fusion with segmental instrumentation and pelvis fixation. Readmission within 30 d 3 (12)
Mortality 0 (0)
antibiotics. Patient #15 was readmitted for Escherichia *n = 25.
coli pneumonia. GU indicates genitourinary; MSK, musculoskeletal.
There were 2 reoperations (8%) within 90 days of
discharge. Patient #15 received an incision and drainage
for distal surgical site infection during their readmission, fused (P = 0.35), procedure time (P = 0.84), LOS (P = 0.27),
and patient #9 received a C2-T2 fusion for significant number of complications and their severity (P = 0.07), read-
proximal junctional kyphosis. Patient #5 who was am- missions (P = 0.50), or reoperations (P = 0.10). There were
bulatory preoperatively, had an L4-pelvis fusion for per- significantly more nonambulatory patients that were fused to
sistent sagittal imbalance 1-year postoperation. the pelvis than not (80% vs. 20%; P = 0.02).
Fused to Pelvis Versus Unfused
DISCUSSION
There were no significant differences between patients
RS is a rare X-linked dominant disorder that fre-
fused to the pelvis and not in EBL (P = 0.29), pRBC trans-
quently causes severe scoliosis in females. We found 81
complications in 25 RS patients (3.2 complications per
TABLE 1. Patient Characteristics patient) after PSF, which was significantly higher com-
Characteristics Values [n (%)] pared with the CP patients (0.4) reported by Yaszay
et al,16 and slightly less than Gabos et al,7 (4) though both
No. patients 25
Age at surgery [mean (SD)] (y) 12.9 (1.8)
authors used a “major-minor” complication system. Due
Length of stay [median (range)] (d) 7 (3-20) to the rarity of this condition, there is a dearth of studies
Ethnicity describing the complications with varying reporting
White 11 (42) methods.3,5–8,17 In 2002, Kerr et al3 described the early
Hispanic 11 (42) complications of 24 patients, though complications were
Black 3 (11)
Preoperative ambulatory status retrospectively reported by patient surveys. More recently
Ambulatory 3 (12) Rocos and Zeller8 found no complications in their 8 pa-
Ambulatory with assistance 5 (20) tients, and studies by Cohen et al5 and Rumbak et al6 with
Nonambulatory 17 (68) 21 and 8 patients, respectively, focused on respiratory
Imaging [mean (SD)]
Preoperative major curve 79 (23) complications. This lack of information and reliability of
Postoperative major curve 32 (15)* reporting makes it difficult for surgeons to evaluate and
Intraoperative characteristics relay the risks and benefits of spinal fusion to patients and
Fused to pelvis 20 (80) their families.
No. vertebrae fused [median (range)] 15 (11-16) Postoperative seizures were present in 5 (20%) of our
Procedure time [median (range)] (min) 305 (227-487)
Estimated blood loss [median (range)] (mL) 600 (200-2500) patients and is a uniquely prevalent postoperative com-
Cell saver transfused [median (range)] (mL) 200 (0-1015) plication for this population. Epilepsy has been reported
Packed red blood cells transfused [mean (SD)] (mL) 243.4 (337.2) in 60% to 80% of patients with RS and has been associated
*P < 0.001; n = 25. with more severe disability.18 Furthermore, postoperative
seizures have occurred in 16% to 31% of cases after PSF in

J Pediatr Orthop Volume 43, Number 5, May/June 2023 Early Complications After PSF in Patients With RS
TABLE 3. CDS Classification of Postoperative Complications Within 90 Days*

Classification Grade I Grade II Grade III Grade IVa
No deviation from normal care Deviation from normal care Treatable but requires repeat Life-threatening condition or
and requires minimal requiring prolonged stay or surgery, intervention, or requires PICU admission, no
treatment increased postoperative follow-up readmission long-term disability

Patient #
1 Constipation; hypothermia; Transfusion; NIMV
tachycardia
2 Constipation; fever; required Transfusion
NG tube
3 Transfusion
4 Transfusion
5 Dysphagia; gait disturbance Seizure; superficial wound
requiring inpatient dehiscence
rehabilitation; tachycardia
6 Constipation; emesis; urinary Transfusion
retention
7 Emesis Chronic urosepsis
8 Emesis; required NG tube
9 Fever C2-T2 fusion reoperation Respiratory failure
10 Transfusion; seizure Respiratory failure
11 Atelectasis Transfusion; NIMV Readmission for SIRS
on 04/15/2023
12 Emesis Hypotensive shock

13 Fever; hypotension; emesis
14 Atelectasis; constipation; Transfusion
required NG tube
15 Muscle spasms Pneumonia and I&D for SSI
16 Emesis; fever
17 Transfusion; seizure; superficial DIC
wound dehiscence
18 Constipation; emesis; fever
19 Atelectasis; muscle spasms; Transfusion DIC; PICU readmission for
fungal UTI pneumonia and pleural
effusion
20 Tachycardia Seizure; chronic hip pain
21 Fever Aspiration pneumonia
22 Diarrhea; tachycardia Transfusion; NIMV
23 Atelectasis
24 Atelectasis; constipation Transfusion; seizure
25 Emesis Transfusion; NIMV; sacral Hypotensive shock
pressure ulcer
CDS grades IVb and V were excluded from the table due to lack of events.
CDS indicates Clavien-Dindo-Sink; DIC, disseminated intravascular coagulopathy; I&D, incision and drainage; NG, nasogastric; NIMV, noninvasive mechanical
ventilation; PICU, pediatric intensive care unit; SIRS, systemic inflammatory response syndrome; UTI, urinary tract infection.
the literature,7,19 perhaps due to medication resistance in Other studies have also found high rates of respiratory
46% of patients,19,20 hemodilution, preoperative medi- complications in RS patients, such as 63% major respiratory
cation noncompliance, anesthetic exposure, or changes in complications,7 and 38% mortality due to lower respiratory
gastrointestinal motility and emesis (56%). Seizures pose a tract infection or respiratory failure.4 In addition, respiratory
theoretical risk of postoperative implant displacement, complications may be uniquely more significant in RS pa-
though this was not corroborated by Abousamra and tients, as Cohen et al5 and Rumbak et al6 both found higher
colleagues in patients with CP and epilepsy. Rather, an rates of respiratory failure, longer ventilation times, and in-
increased risk of respiratory complications and longer creased noninvasive mechanical ventilation compared with
ICU stay was observed.21 Although the relationship be- patients with NMS. Whereas in-utero hypoxic injury may
tween postoperative seizures and outcomes in children lead to respiratory compromise and vulnerability in patients
with RS has yet to be studied, neurology consultation with CP, pontine noradrenergic deficits due to loss of the
perioperatively for all RS patients with epilepsy is neces- MECP2 gene leads to progressive central dysautonomia,
sary to avoid seizure-related respiratory complications and ventilation and perfusion mismatches, and dysphagia which
prolonged ICU stays. may ultimately increase the risk infection and respiratory
Families should be made aware of the potential for failure in patients with RS.5,6,22,23 Given the high frequency
respiratory complications in patients with RS. In our study, and severity of respiratory complications in our cohort, we
32% of patients developed postoperative atelectasis, 12% re- support recommendations made by Cohen et al5 for diligent
quired postoperative noninvasive mechanical ventilation, 8% multidisciplinary perioperative respiratory care that empha-
had respiratory failure and 16% had nosocomial pneumonia. sizes appropriate extubation timing, pain control, early

Taylor et al J Pediatr Orthop Volume 43, Number 5, May/June 2023
mobilization and feeding, and ultimately the development of 3. Kerr AM, Webb P, Prescott RJ, et al. Results of surgery for scoliosis
a systematic perioperative care guideline. in Rett syndrome. J Child Neurol. 2003;18:703–708.
4. Downs J, Torode I, Wong K, et al. Surgical fusion of early onset
Our study found no significant differences in com- severe scoliosis increases survival in Rett syndrome: a cohort study.
plications between patients who were fused to the pelvis Dev Med Child Neurol. 2016;58:632–638.
and those who were not. However, 1 patient did require 5. Cohen JL, Klyce W, Kudchadkar SR, et al. Respiratory complica-
pelvic fixation 1 year later due to sacropelvic instability. tions after posterior spinal fusion for neuromuscular scoliosis:
Our 2 spinal reoperations were similar to Larsson et al,17 children with Rett syndrome at greater risk than those with cerebral
palsy. Spine (Phila Pa 1976). 2019;44:1396–1402.
who also reported a complimentary stabilization to the 6. Rumbak DM, Mowrey W, Schwartz SW, et al. Spinal fusion for
pelvis in 1 patient and cervical spine in another. The scoliosis in Rett syndrome with an emphasis on respiratory failure
majority of our patients fused to the pelvis were non- and opioid usage. J Child Neurol. 2016;31:153–158.
ambulatory (80%) as this allowed for improved seated 7. Gabos PG, Inan M, Thacker M, et al. Spinal fusion for scoliosis in
balance and stability. Similar to Rocos and Zeller, we Rett syndrome with an emphasis on early postoperative complica-
tions. Spine. 2012;37:E90.
avoided pelvic fixation for patients without significant 8. Rocos B, Zeller R. Correcting scoliosis in Rett syndrome. Cureus.
(< 15 degrees) pelvic obliquity and if they were capable of 2021;13:e15411.
controlling the lumbosacral junction.8,24 However, be- 9. Clavien PA, Sanabria JR, Strasberg SM. Proposed classification of
cause RS is a progressive condition, it may be appropriate complications of surgery with examples of utility in cholecystectomy.
Surgery. 1992;111:518–526.
to perform pelvic fixation in equivocal cases to avoid re- 10. Keil LG, Himmelberg SM, Guissé NF, et al. Complications
operations. Four of our patients also had superficial following posterior spinal fusion for adolescent idiopathic scoliosis:
wound complications, and 1 required an incision and a retrospective cohort study using the modified Clavien-Dindo-Sink
on 04/15/2023
drainage reoperation. To reduce wound complications, system. Spine Deform. 2022;10:607–614.

collaboration with plastic surgery for multilayered clo- 11. Roye BD, Fano AN, Quan T, et al. Modified Clavien-Dindo-
Sink system is reliable for classifying complications following
sures for patients with pelvic instrumentation has become surgical treatment of early-onset scoliosis. Spine Deform. 2023;11:
more frequent at our institution, as Imahiyerobo et al25 205–212.
found significantly decreased wound complications, and 12. Zhou L, Willoughby K, Strobel N, et al. Classifying adverse events
Ward et al26 found a 0% return to the operating room. following lower limb orthopaedic surgery in children with cerebral
palsy: reliability of the modified Clavien-Dindo System. J Pediatr
This study is not without limitations. Complications are Orthop. 2018;38:e604–e609.
often subjectively classified; as most grade I complications 13. Guissé NF, Stone JD, Keil LG, et al. Modified Clavien–Dindo–Sink
would not necessarily be considered a “complication” de- classification system for adolescent idiopathic scoliosis. Spine
pending on the treating physician, patient population, surgery, Deform. 2022;10:87–95.
14. Helenius IJ. Standard and magnetically controlled growing rods for
or severity. Our study is also a small retrospective review of only
the treatment of early onset scoliosis. Ann Transl Med. 2020;8:26.
25 patients, which precluded a meaningful subanalysis of risk 15. Dayer R, Ouellet J, Saran N. Pelvic fixation for neuromuscular scoliosis
factors. However, because RS is a rare condition, most studies deformity correction. Curr Rev Musculoskelet Med. 2012;5:91–101.
and institutions experience few cases and are thus logistically 16. Yaszay B, Bartley CE, Sponseller PD, et al. Major complications
limited to descriptive reporting and low-powered comparisons. following surgical correction of spine deformity in 257 patients with
cerebral palsy. Spine Deform. 2020;8:1305–1312.
Comparatively, our study is higher powered in both sample size 17. Larsson EL, Aaro S, Ahlinder P, et al. Long-term follow-up of
and description. Due to the retrospective nature of this project functioning after spinal surgery in patients with Rett syndrome. Eur
over a large timespan, patient-reported outcome measures were Spine J. 2009;18:506–511.
unable to be uniformly obtained. Last, our study included pa- 18. Operto FF, Mazza R, Pastorino GMG, et al. Epilepsy and genetic in
Rett syndrome: a review. Brain Behav. 2019;9:e01250.
tients from 5 different surgeons which may introduce variation 19. Karmaniolou I, Krishnan R, Galtrey E, et al. Perioperative manage-
in surgical technique and thereby outcomes. However, each ment and outcome of patients with Rett syndrome undergoing scoliosis
surgeon is well experienced with this surgical procedure and had surgery: a retrospective review. J Anesth. 2015;29:492–498.
similar complication rates. 20. Haas RH, Dixon SD, Sartoris DJ, et al. Osteopenia in Rett
Despite these limitations, we have carefully identi- syndrome. J Pediatr. 1997;131:771–774.
21. Abousamra O, Sullivan BT, Shah SA, et al. Do seizures compromise
fied and reported the complications of children with RS correction maintenance after spinal fusion in cerebral palsy scoliosis?
who underwent PSF according to the CDS classification J Pediatr Orthop B. 2020;29:538.
system. This data represents the largest series to report 22. MacKay J, Leonard H, Wong K, et al. Respiratory morbidity in Rett
early postoperative complications. PSF can effectively syndrome: an observational study. Dev Med Child Neurol. 2018;60:
correct spinal deformity, but surgeons and families should 951–957.
23. Taneja P, Ogier M, Brooks-Harris G, et al. Pathophysiology of locus
be aware of a 3.2 complications/child rate with 32% hav- ceruleus neurons in a mouse model of Rett syndrome. J Neurosci.
ing grade IVa complications, a 20% seizure rate, 48% 2009;29:12187–12195.
respiratory complication rate, 8% having reoperations 24. Modi HN, Woo Suh S, Song HR, et al. Evaluation of pelvic fixation
within 90 days, and 12% being readmitted within 30 days. in neuromuscular scoliosis: a retrospective study in 55 patients. Int
Orthop. 2010;34:89–96.
25. Imahiyerobo T, Minkara AA, Matsumoto H, et al. Plastic multi-
REFERENCES layered closure in pediatric nonidiopathic scoliosis is associated with
1. Rett A. On a unusual brain atrophy syndrome in hyperammonemia a lower than expected incidence of wound complications and surgical
in childhood. Wien Med Wochenschr. 1966;116:723–726. site infections. Spine Deform. 2018;6:454–459.
2. Killian JT, Lane JB, Lee HS, et al. Scoliosis in Rett syndrome: progression, 26. Ward JP, Feldman DS, Paul J, et al. Wound closure in nonidiopathic
comorbidities, and predictors. Pediatr Neurol. 2017;70:20–25. scoliosis: does closure matter? J Pediatr Orthop. 2017;37:166–170.

ORIGINAL ARTICLE
Femoral Deformity in Tibia Vara and Its Response

to Growth Modulation
Janet L. Walker, MD,* David M. Dueber, PhD,† Allison C. Scott, MD,‡

Lindsay P. Stephenson, MD,‡ Joel A. Lerman, MD,§ Sarah B. Nossov, MD,∥
Corinna C. Franklin, MD,∥ Kenneth P. Powell, MD,¶ David E. Westberry, MD,#
and Jeffery D. Ackman, MD**
varus, had no change in their mean mLDFA of 87 degrees.

Background: While tibia vara is a disorder of the proximal tibial However, 4 femurs (10%) ended with posttreatment varus.
physis, femoral deformity frequently contributes to the overall Conclusions: Femoral LTBP is effective in correcting femoral
limb malalignment. Our purpose was to determine how femoral varus deformity in the tibia vara. For femoral varus associated
varus deformity in tibia vara responds to growth modulation, with late-onset tibia vara, femoral LTBP should be considered.
on 04/15/2023
with/without lateral tension band plating (LTBP) to the femur. Those that had femoral LTBP had statistically more successful
Methods: One-hundred twenty-seven limbs undergoing LTBP for femoral and overall limb varus correction. However, in early-
tibia vara were reviewed. All had tibial LTBP and 35 limbs also onset tibia vara, with associated femoral varus, observation is
had femoral LTBP for varus. Radiographs were measured for warranted because 73% of femurs are corrected without femoral
correction of the mechanical lateral distal femoral angle intervention. This study was underpowered to show additional
(mLDFA) and mechanical axis deviation (MAD). Preoperative- improvement with femoral LTBP in the early-onset group. Even
femoral varus was defined with an age-adjusted guide: mLDFA limbs with normal femoral alignment, should be observed closely
> 95 degrees for 2 to below 4 years and mLDFA > 90 degrees for for the development of femoral varus, during tibial LTBP
4 to 18 years. The 35 limbs having femoral LTBP were compared treatment for tibia vara.
with 50 limbs with femoral varus and no femoral LTBP. In ad- Level of Evidence: Level III.
dition, 42 limbs that did not have preoperative-femoral varus
were followed. Patients with early-onset (below 7 y) tibia vara Key Words: Blount disease, guided growth, knee, children, lower
were compared with those with late-onset ( ≥ 8 y). Outcome limb deformity
success was based on published age-adjusted mLDFA and (J Pediatr Orthop 2023;43:303–310)
MAD norms.
Results: Following femoral LTBP, the mean mLDFA decreased
from 98.0 to 87.1 degrees. All femurs had some improvement,
with 28/35 femurs (80%) achieving complete correction. One
limb, with late follow-up, overcorrected, requiring reverse (me-
dial) femoral tension band plating. For the 50 limbs with femoral
varus and only tibial LTBP, 16/22 limbs (73%) with early-onset
T ibia vara, also known as Blount disease or os-
teochondrosis deformans tibiae, is a developmental
disorder that affects growth at the medial proximal tibial
and 11/28 limbs (39%) with late-onset completely corrected their
physis. The resulting tibial varus deformity has been di-
femoral deformities. If the limb had preoperative-femoral varus,
rectly correlated with body mass index (BMI) suggesting a
femoral LTBP statistically correlated with successful mLDFA
mechanical etiology.1,2 In young children, excess weight
correction and improvement of MAD, only in the late-
may compress the medial physis in a limb physiologically
onset group. Forty-two limbs, without preoperative-femoral
aligned in varus. In older children, who may not be
physiologically in varus, abnormal compressive forces on
the medial physis3 could be related to gait deviations re-
From the *Department of Orthopaedic Surgery and Sports Medicine, sulting from wider thighs in children with obesity.4
Shriners Children’s and University of Kentucky; †Department of Accompanying the tibial varus, femoral deformity,
Statistics, University of Kentucky, Lexington, KY; ‡Shriners Child- especially varus, has been reported.5–8 It is seen more
ren’s, Galveston, TX; §Shriners Children’s, Sacramento, CA;
∥Shriners Children’s, Philadelphia and Erie, PA; ¶Shriners Children’s,
frequently in those with adolescent-onset tibia vara. The
Shreveport, LA; #Shriners Children’s, Greenville, SC; and **Shriners abnormal compressive forces on the medial knee, due to
Children’s, Chicago, IL. excessive weight, are also implicated in the femoral de-
No funding sources. formity, and occasional widening of the lateral distal
The authors declare no conflicts of interest. femoral physis has been reported.3 However, the radio-
Reprints: Janet L. Walker, MD, Shriners Hospital for Children, 110
Conn Terrace, Lexington, KY 40508. E-mail: jwalker@shrinenet.org. graphic changes of medial physeal widening, metaphyseal/
Copyright © 2023 Wolters Kluwer Health, Inc. All rights reserved. epiphyseal fragmentation, and physeal bar formation seen
DOI: 10.1097/BPO.0000000000002358 in the proximal tibia, are not typically seen in the femur.

Walker et al J Pediatr Orthop Volume 43, Number 5, May/June 2023
on 04/15/2023
FIGURE 1. Flow chart of study patients with tibia vara and tension band plating (TBP) groups in bold are analyzed in this study.
LTBP indicates lateral tension band plating.
Temporary hemiepiphyseodesis has gained popu- patients were reviewed. All had tibial LTBP for tibia
larity in the management of varus deformity in vara. Those with prior surgery were excluded. Patients
tibia vara.9–14 Tension band plates have been introduced were stratified by age at tibia vara diagnosis into early-
as a successful method of growth modulation.15 onset (74 with below 7 y) and late-onset (70 with 8 y and
While their use in tibia vara deformity appears less above), based on the bimodal age distribution of our
successful than in other disorders,16,17 their technical entire population of 147 patients. As there were only 2
ease and low complication rates make them attractive patients aged 7 years, this we defined this the nadir
as tibia vara patients frequently have surgical comorbid- frequency age to separate our age at diagnosis groups.
ities. These 2 patients, along with one other, where the age at
The purpose of our study is to determine the re- onset was not recorded, were excluded to create the early
sponse of femoral deformity in tibia vara to temporary (below 7 y) and late (8 y and above) onset groups.
hemiepiphyseodesis by lateral tension band plating Preoperative radiographs were measured to determine
(LTBP) to the proximal tibia, with and without tension the prevalence and severity of deformities. To assess the
band procedures to the distal femur. changes due to surgery, patients with femoral LTBP were
followed for a minimum of 2 years after their femoral
procedures unless they corrected the deformity, had
METHODS subsequent surgery, or became skeletally mature.
We performed a retrospective review of 185 pa- The patients not having femoral procedures were
tients identified by coding query of tibia vara and hem- followed through their growth modulation series
iepiphyseodesis at 7 centers between 2008 and 2018. The (GMS) for a minimum of 2 years unless they had
study was reviewed by an Institutional Review Board for subsequent osteotomy/physeal bar surgery, corrected
all centers. Using the selection tree in Figure 1, 98 their tibial deformity, or became skeletally mature.

J Pediatr Orthop Volume 43, Number 5, May/June 2023 Femoral Deformity in Tibia Vara
GMS occurring within the same limb during the study

follow-up period could encompass the effects of initial
tibial LTBP, any tibial implant revision or removal,
rebound growth after implant removal, and/or
reimplantation to treat tibial recurrence.
Radiographic outcomes were determined by com-

paring preoperative to follow-up limb alignment on digital

standing anteroposterior full-length lower limb radio-
graphs performed by the technologists with the patellae
facing forward. Due to the time frame of this study, both
conventional and slit-beam digital radiography systems
were in use. Conventional films were obtained with a
single exposure to films aligned in a single cassette. Elec-
tronic stitching of the conventional films was performed
automatically by the system and not modified by the
technologists. All films included the ankle and femoral
head. Inaccurately aligned films were excluded by the
senior author measuring the films if the patella was not
centered within the femoral condyles. A single senior in-
on 04/15/2023
vestigator (JLW) with 36 years of pediatric orthopaedic

experience measured the films to minimize variability us-
ing McKesson Radiology Station 12.3 64 bit 1989-2019
(Irvine, TX). The measurements (Fig. 2) included medial
proximal tibial angle (MPTA), mechanical lateral distal
femoral angle (mLDFA), mechanical tibio-femoral angle
(mTFA), and mechanical axis deviation (MAD).18 The
investigator’s reliability to measure these films was
assessed while blinded to outcome from radiographic
measurements of children with tibia vara, 38 early-onset,
and 29 late-onset limbs, at 2 time points. The intraclass
coefficients ranged from 0.907 to 0.998, indicating FIGURE 2. The radiographic measurements used in this study
excellent agreement. are shown here. MAD indicates mechanical axis deviation is the
The measures for mLDFA and MAD were com- distance from the middle of the knee joint to the overall limb
pared with reported age-adjusted normal values for mechanical axis; mLDFA, mechanical lateral distal femoral an-
children.19 We noted that remembering this age-adjusted gle is the lateral angle formed between the mechanical femoral
table in a clinical setting is not practical. Therefore, we axis and the joint line of the femoral condyles; MPTA, medial
propose a short-cut guide for determining age-adjusted proximal tibial angle is the medial angle formed by the tibial
mechanical axis and the joint line of the tibial plateaus; mTFA,
varus by using the mLDFA > 95 degrees for children
mechanical tibio-femoral angle is the angle formed by the
below 4 years and mLDFA > 90 degrees for children mechanical axes of the tibia and femur.
4 years and above as a definition of preoperative femoral
varus. The comparison of this preoperative guide for RESULTS
femoral varus and the age-adjusted mLDFA norms are
shown in Figures 3A and B. Femoral LTBP
Logistic regression models were estimated to assess Three patients with 5 limbs were excluded for in-
whether femoral LTBP influenced the odds of success of sufficient follow-up leaving 40 femoral tension band
correcting deformity in the femur (mLDFA) and in the plate procedures for review, 38 for initial varus de-
overall limb (MAD), controlling for preoperative deformity, and 2 for initial valgus deformity. The 2 limbs
formity severity. Analyses were performed separately for having tension band plating for valgus deformity and 3
groups with early-onset and late-onset types of tibia additional limbs having LTBP’s with femoral alignment
vara. While our preoperative guide for mLDFA was within the age-adjusted norm for mLDFA were excluded
used for preoperative femoral deformity grouping, the from our comparative analyses. This left 35 limbs with
outcome success was determined by the age-adjusted femoral LTBP for varus and their population summary,
norms for mLDFA and MAD. Because of the small with clinical outcomes, are shown in Table 1. There
sample size, the power for detecting an effect using lo- was one infection and another wound had persistent
gistic regression is expected to be small; accordingly, drainage. Both healed without further surgery. One
linear regression models were also estimated to assess metaphyseal screw broke but did not need revision and
whether femoral LTBP influenced the postoperative 1 patient, with late follow-up ( > 6-mo interval),
values of mLDFA and MAD, controlling for pre- required medial femoral tension band plating for
operative severity. overcorrection. Twenty-nine limbs had their femoral

TABLE 1. Population Summary for Femoral Lateral Tension Band Plate Procedures Used in Analysis
All femoral lateral tension band plate Age <7 y at Age ≥ 8 y at
procedures diagnosis diagnosis
Patients N = 25 N=5 N = 20
Sex (at birth, males:females) 15:10 1:4 14:6

Race/ethnicity (self-reported)
Black/not Hispanic 15 3 12
White/Hispanic 2 1 1
White/not Hispanic 6 6
Other 2 1 1
Limbs N = 35 N=8 N = 27
Side (right:left) 18:17 5:3 13:14
Mean preoperative age (y) 11.2 6.4 12.6
Mean preoperative years of growth remaining* (y) 5.0 8.9 3.8
Mean reoperative body weight (kg) 87.2 50.4 98.2
Mean preoperative body mass index (kg/m2) 36.0 29.9 37.8
Clinical outcome (N)
Number with age-adjusted femoral correction/ 28 6 22
overcorrection
Femurs not completely corrected 7 2 5
Femoral implants removed 23 6 17
on 04/15/2023
Femoral epiphyseodesis 2 1 1
Femoral lateral tension band plate reimplanted for 2 2
recurrence
Femoral tension band plate reversed for 1 1
overcorrection
Femurs mature without implant removal 7 7
Corrected 4 4
Undercorrected 3 4
Femurs immature with implants 5 2 3

Corrected 3 1 2
Overcorrected 2 1 1
*Assumes maturity at age 15 years for girls and 17 years for boys.20
procedures with the tibial LTBP and 6 had them at a late-onset group (P = 0.002). The overall femoral correction
mean of 2.0 years following their tibial LTBP. Their rate for the early-onset group was not affected by femoral
radiographic findings are shown in Table 2, along with LTBP, as many corrected on their own. Limbs without in-
the femoral deformity groups to be compared. Those itial femoral varus were more likely to achieve MAD cor-
having femoral procedures for varus had more varus rection (P = 0.01). Using 87 degrees as a norm for mLDFA
than the population as a whole (P < 0.001). At follow-up, and 0 degree as the norm for mTFA, we calculated the
28 of 35 (80%) of femurs were corrected but only 16 percentage of the mTFA attributed to the femoral de-
(46%) had physeal closure. formity = (mLDFA−87 degrees)/(mTFA)×100%. For those
Tibial LTBP Only limbs with femoral varus, 52% of the overall deformity was
The 92 limbs with tibial only LTBP were used for in the femur for those having femoral LTBP and 46%
femoral outcome comparisons. Of these, 44 limbs pre- femoral deformity for those not having femoral LTBP. The
operatively had age-adjusted femoral varus, 47 had nor- percentage of limb deformity due to the femur was not
mal femoral alignment, and 1 femur was in valgus. Using statistically different between the groups (P = 0.266).
our preoperative mLDFA guide, 50 femurs were classified The logistic and regression model estimates are shown
as varus and 42 were not. Their radiographic summary is in Table 3. Femoral LTBP improved the likelihood of
shown in Table 2. This table also includes tibial (MPTA) femoral alignment (mLDFA) success in the late-onset group
and overall limb alignment changes (mTFA and MAD) with an odds ratio of 21.8, controlling for preoperative
with GMS. The rate of successful correction of age- severity. In the early onset group, femoral LTBP was not
adjusted mLDFA and MAD was higher for those in the found to significantly improve the odds of femoral alignment
early-onset group compared with the late-onset group. success. Femoral LTBP was not found to significantly
improve the odds of overall limb alignment (MAD age-
Comparisons adjusted) success in either the early onset or the late onset
For those limbs with femoral varus based on our groups. Femoral LTBP was found to have a significant
preoperative guide, having femoral LTBP statistically im- impact on postoperative mLDFA in both the early onset
proved their success of femoral deformity correction for the (−6.5 degrees compared with no surgery) and late onset

(−6.8 degrees compared with no surgery), controlling for
modulation final
preoperative severity. Femoral LTBP was also found to have
mLDFA ≤ 90 degrees, except
≤ 95 degrees for under 4 y
Growth
−30, 51
LTBP indicates lateral tension band plate; MAD, mechanical axis deviation; mLDFA, mechanical lateral distal femoral angle; MPTA, medial proximal tibial angle; mTFA, mechanical tibiofemoral angle.
−9, 14
82-94
a significant reduction on postoperative MAD in the late
7499
90%
71%
87.7
86.4
1.4
6.2
onset group (20.4 mm compared with no surgery) but not the
(n = 42 limbs)
early onset group, controlling for preoperative severity.

DISCUSSION
Preoperative
−3.8, 105
While several authors have reported on femoral varus
−2, 33
83-95
57-87
88.8
77.3
14.2
34.0
in tibia vara,5–8 none have specifically studied the effect of
tension band plate surgeries. We demonstrated that, as ex-
pected, femoral LTBP does correct femoral varus associated
with tibia vara in 80% of limbs, using pediatric normative
modulation final
values. We had a 6% wound complication rate and one

mLDFA > 90 degrees without
overcorrection, requiring treatment. Bushnell et al21 found

Growth
−6, 131
85-104
−2, 36
femoral LTBP late onset
71-98
39%
35%
93.0
83.9
11.2
43.0
femoral physeal staples corrected the femoral deformities in
( ≥ 8 y, n = 28 limbs)
tibia vara from mean = 8 degrees to mean = 0 degree.

However, the amount of correction ranged from 0 to 19
degrees. Our mean correction was 11 degrees with a range of
on 04/15/2023
2 to 34 degrees. Like Bushnell and colleagues, we could not

Preoperative
correlate amount of correction with deformity severity or

17-115
90-99
58-85
5-33
93.6
76.8
18.1
63.3
other predictors in our study.

In addition, we showed that femoral varus deformity
is capable of significant remodeling in tibia vara. Danino
et al20 reported that 64% of their patients had mLDFA
modulation final
LTBP early onset (< 7 y, n = 22 limbs)
correction to adult norm in the first 24 months following

degrees for under 4 y without femoral
mLDFA > 90 degrees except > 95
TBP in tibia vara of all ages. We found if the child had

Growth
−28, 16
−94, 55
81-104
81-97
73%
55%
89.8
88.5
2.3
7.2
late-onset tibia vara (onset 8 y and above) and femoral

varus, their femoral deformities only achieved successful
correction in 39%, with tibial LTBP alone. This group did
have statistically better femoral deformity success by
adding femoral LTBP.
However, for children with early-onset tibia vara
Preoperative
(onset below 7 y) and femoral varus, we found 73% of

90-108
66-86
10-88
4-48
97.0
77.4
23.6
51.4
the femoral deformities corrected during LTBP treat-

ment of the tibia, without femoral LTBP. We found no
statistical outcome predictors for this group. Those with
additional femoral LTBP did not perform better than
those with tibial LTBP surgery alone. This may be re-
Final or At femoral
implant removal
lated to the high rate of remodeling, but this study is

−43, 107
−13, 35
limited by the small number of limbs having femoral

70-100
70-103
(n = 35 age-adjusted
80%
55%
87.1
84.1
20.1
5.3
Femoral LTBP
LTBP in the early-onset group. While the regression

varus limbs)
analyses regarding femoral LTBP in early-onset limbs

was not significant, the odds ratios and beta value were
substantial, suggesting a more powered study might find
TABLE 2. Femoral Group Comparisons
Preoperative
significance.
91-115
11-117
61-94
The goal of GMS is correction of the overall limb

5-51
98.0
78.5
23.5
72.1
alignment. Those limbs with initial femoral varus were less

likely to correct their MAD with GMS than those with-
out. In the late-onset group, femoral LTBP had a sig-
% age-adjusted correction/
% age-adjusted correction/
nificant reduction on the final MAD, 20.4 mm compared

with no femoral LTBP. As femoral varus was shown to
Measures of deformity
improve with femoral LTBP, the improvement in MAD

overcorrection
overcorrection
would be expected. It is possible that this reduction in

mLDFA (deg.)
forces by femoral improvement would reduce the stresses

MPTA (deg.)
mTFA (deg.)
MAD (mm)
on the tibia, improving the chances of tibial correction.

Range
Range
Range
Range
However, the change in mean MPTA was lowest in those

Mean
Mean
Mean
Mean
limbs having both tibial and femoral LTBP. Finally, we

found femoral varus can develop during treatment with

TABLE 3. Regression Analyses Predicting Femoral and Overall Limb Alignment Correction Success With Growth Modulation
Using LTBP
Early-onset Late-onset
Logistic regression Femoral alignment success (age adjusted mLDFA) Femoral alignment success (age adjusted mLDFA)
Model 1 Model 1
Predictor Odds ratio 95% CI P Predictor Odds ratio 95% CI P

Pre-mLDFA 0.85 0.71-0.99 0.066 Pre-mLDFA 0.76 0.58-0.93 0.016*
Femoral LTBP yes/no 4.39 0.41-10.48 0.283 Femoral LTBP yes/no 21.76 4.41-177.03 < 0.001*
Limb alignment success (age adjusted MAD) Limb alignment success (age adjusted MAD)
Model 1 Model 1
Predictor Odds ratio 95% CI P Predictor Odds ratio 95% CI P
Pre-MAD 0.96 0.91-1.00 0.079 Pre-MAD 0.97 0.95-1.00 0.055
Femoral LTBP yes/no 7.09 0.89-94.20 0.089 Femoral LTBP yes/no 2.25 0.70-7.75 0.182
Linear regression Predicting final mLDFA Predicting final mLDFA

Model 1 Model 1
Predictor b P Predictor b P
Pre-mLDFA 0.35 0.015* Pre-mLDFA 0.48 < 0.001*
Femoral LTBP yes/no −6.51 0.016* Femoral LTBP yes/no −6.83 < 0.001*
Predicting final MAD Predicting final MAD
Model 1 Model 1
on 04/15/2023
Predictor b P Predictor b P
Pre-MAD 0.37 0.008* Pre-MAD 0.39 0.032*
Femoral LTBP yes/no −14.8 0.098 Femoral LTBP yes/no −20.42 0.011*
*Denotes statistical significance.
CI, confidence interval; LTBP, lateral tension band plate surgery; MAD, mechanical axis deviation; mLDFA, mechanical lateral distal femoral angle.
tibial LTBPs. If it does, there should be a consideration for the commonly cited adult norm of ≤ 90 degrees.18 We
adding a femoral LTBP later. proposed a short-cut guide using this adult norm from 4 to
A number of studies have looked at femoral de- 18 years and expanding it to ≤ 95 degrees for those age 2
formity in tibia vara at single time-points. They showed to below 4 years (Figure 3). When we analyzed our results
findings suggestive of femoral varus correction with age in using the strict age-adjusted definition of femoral varus,
early-onset but not the late-onset tibia vara. These findings we had the same statistical predictions as when the
could be attributed to the natural history of physiological proposed guide was used.
femoral varus as shown by the dashed lines in Figure 3A. Response of femoral deformity to tibial osteotomy
Gordon et al8 noted that the mLDFA in those patients was described in 2 prior studies. Abraham et al22 found
with onset below 4 years decreased as their age at aLDFA did not change in children after tibial osteotomy
evaluation increased. They found by age 7 years, the and all of their subjects were above 7 years at osteotomy.
mLDFA was in the normal adult range. In those with LaMont et al23 found the preoperative LDFA was ac-
onset 4 years and above, there was no change in their tually lower in those limbs that recurred after osteotomy
mLDFA with increasing ages. Myers et al7 found the compared to those that did not. Those patients that re-
anatomic LDFA (aLDFA) in children with tibia vara, curred were also significantly older at surgery and may
onset age 4 to 9 years, to be within the normal adult range, have already had the expected physiological decline in
as well. In those with onset 10 years and above, femurs their LDFA with age, before surgery. The findings of these
were statistically more varus than the adult norms. Kline studies are consistent with those of our patients following
et al5 found that in limbs with age of onset below 6 years, tibial LTBP.
all of their deformity was coming from the tibia, based on This retrospective study is limited as there was no
ideal adult alignment. Those with age at onset above formal protocol for patient selection or management. We
6 years had 34% to 76% of their limb varum coming from did include patients who had prior bracing but their
the femur. Firth et al6 were the only authors to use a deformities were still severe enough to have surgical
pediatric control group, consisting of normal or mid- treatment. The decision to perform femoral LTBP was
proximal femur fracture radiographs, for their aLDFA that of the patients’ individual physicians. We had no
measures in tibia vara. They found statistically more control group of our own and deferred to the published
femoral varus in tibia vara for those with onset below reference population radiographic data.19 It is possible
4 years and above 10 years but not in the 4 to 9 years that this group of typical children is not representative of
onset group. our population of children with tibia vara or the way
We took advantage of the published normative obese children have to stand for radiographs. Some
values of pediatric lower limb alignment for our study studies have used the contralateral limb as a control but
comparisons.19 The upper limit of mLDFA varies it is unclear what effect abnormal gait and stance due to
throughout childhood, making it harder to remember than the tibia vara has on that limb. The literature supports

on 04/15/2023
FIGURE 3. These graphs demonstrate the mechanical lateral distal femoral angle (mLDFA) changes that occurred in response to
tibial only lateral tension band plating (LTBP) in limbs with tibia vara. This is shown in relation to the published age-adjusted norms
as indicated by the dashed lines.19 (A) Shows the mLDFA changes for those with early-onset (below 7 y) and (B) shows those with
late-onset (8 y and above). The solid gray areas show our proposed short-cut guide for age-adjustment of femoral varus to facilitate
use in the clinical setting.
tibia vara having different characteristics based on age of growth after physeal untethering by LTBP removal, we
onset. They are frequently grouped into infantile-onset were unable to make a recommendation about the ap-
(below 4 y) and adolescent-onset (above 10 y) with the propriateness of overcorrection in growing children.
intermediate ages, so called juvenile-onset (4 to 10 y).6,8 Lastly, most of our patients were immature at last as-
However, the clinical distinctions are not well-defined sessment and continued remodeling and correction/
and studies frequently combine these arbitrary age overcorrection may occur.
groups, shifting the intermediate ages in either direction In conclusion, femoral LTBP is effective in correct-
into various groups called early and late-onset.2,5,7,8,11,12 ing femoral varus deformity in tibia vara. For femoral
Instead, we based our groups on our entire tibia vara varus associated with tibia vara, onset 8 years and above,
population, which had a bimodal age distribution. We femoral LTBP should be considered. Few femurs
removed those few patients in the nadir frequency age of corrected with tibial LTBP alone and those who had
7 years to clearly define our early-onset (below 7 y) and femoral LTBP had statistically more successful femoral
late-onset (8 y and above) groups. In doing so, we div- and overall limb varus correction. However, in tibia vara,
ided the juvenile-onset group which may affect our onset below 7 years, with associated femoral varus, ob-
comparisons to the literature. Due to the variability of servation is warranted because 73% of femurs corrected

without femoral intervention. This study was under- 11. Funk SS, Mignemi ME, Schoenecker JG, et al. Hemiepiphyseodesis
powered to show additional improvement from adding implants for late-onset tibia vara: a comparison of cost, surgical
success, and implant failure. J Pediatr Orthop. 2016;36:29–35.
femoral LTBP in the early-onset group. Even limbs with 12. Jain MJ, Inneh IA, Zhu H, et al. Tension band plate (TBP)-guided
tibia vara and normal femoral alignment, should be hemiepiphysiodesis in Blount disease: 10-year single-center experience with
observed closely for the development of femoral varus, a systematic review of literature. J Pediatr Orthop. 2020;40:e138–e143.
during tibial LTBP treatment. 13. Danino B, Rödl R, Herzenberg JE, et al. The efficacy of guided
growth as an initial strategy for Blount disease treatment. J Child

Orthop. 2020;14:312–317.
REFERENCES 14. Griswold BG, Shaw KA, Houston H, et al. Guided growth for the
1. Scott AC, Kelly CH, Sullivan E. Body mass index as a prognostic treatment of infantile Blount’s disease: Is it a viable option? J Orthop.
factor in development of infantile Blount disease. J Pediatr Orthop. 2020;20:41–45.
2007;27:921–925. 15. Stevens PM. Guided growth for angular correction: a preliminary
2. Sabharwal S, Zhao C, McClemens E. Correlation of body mass series using a tension band plate. J Pediatr Orthop. 2007;27:253–259.
index and radiographic deformities in children with Blount disease. J 16. Fan B, Zhao C, Sabharwal S. Risk factors for failure of temporary
Bone Joint Surg Am. 2007;89:1275–1283. hemiepiphyseodesis in Blount disease: a systematic review. J Pediatr
3. Beskin JL, Burke SW, Johnston CE, et al. Clinical basis for a Orthop B. 2020;29:65–72.
mechanical etiology in adolescent Blount’s disease. Orthopedics. 17. Burhardt RD, Herzenberg JE, Strahl A, et al. Treatment failures and
1986;9:365–370. complications in patients with Blount disease treated with temporary
4. Davids JR, Huskamp M, Bagley AM. A dynamic biomechanical hemiepiphyseodesis: a critical systematic literature review. J Pediatr
analysis of the etiology of adolescent tibia vara. J Pediat Orthop. Orthop B. 2018;27:522–529.
1996;17:461–468. 18. Paley D. Normal lower limb alignment and joint orientation. In:
5. Kline SC, Bostrum M, Griffen PP. Femoral varus: an important Paley D, Herzenburg JE, eds. Principles of Deformity Correction. 1st
component in late-onset Blount’s disease. J Pediat Orthop. 1992;12:
on 04/15/2023
edn. New York, NY: Springer-Verlag; 2005;Chapter 1:1–18.

197–206. 19. Sabharwal S, Zhao C, Edgar M. Lower limb alignment in children.
6. Firth GB, Ngcakani A, Ramguthy Y, et al. The femoral deformity in J Pediatr Orthop. 2008;28:740–746.
Blount’s disease: a comparative study of infantile, juvenile and 20. Danino B, Rödl R, Herzenberg JE, et al. Guided growth:
adolescent Blount’s disease. J Pediat Orthop B. 2020;29:317–322. preliminary results of a multinational study of 967 physes in 537
7. Myers TG, Fishman MK, McCarthy JJ, et al. Incidence of distal patients. J Child Orthop. 2018;12:91–96.
femoral and distal tibial deformities in infantile and adolscent Blount 21. Bushnell BD, May R, Campion ER, et al. Hemiepiphyseodesis for
disease. J Pediat Orthop. 2005;25:215–218. late-onset tibia vara. J Pediatr Orthop. 2009;29:285–289.
8. Gordon JE, King DJ, Luhmann SJ, et al. Femoral deformity in tibia 22. Abraham E, Toby D, Welborn MC, et al. New single-stage double
vara. J Bone Joint Surg Am. 2006;88:380–386. osteotomy for late-presenting infantile tibia vara: a comprehensive
9. Westberry DE, Davids JR, Pugh LI, et al. Tibia Vara: results of approach. J Pediatr Orthop. 2019;39:247–256.
hemiepiphyseodesis. J Pediatr Orthop B. 2004;13:374–378. 23. LaMont LE, McIntosh AL, Jo CH, et al. Recurrence after surgical
10. Scott AC. Treatment of infantile Blount disease with lateral tension intervention for infantile tibia vara: assessment of a new modified
band plating. J Pediatr Orthop. 2012;32:29–34. classification. J Pediatr Orthop. 2019;39:65–70.

ORIGINAL ARTICLE
Fibrin Glue is a Viable Alternative to Fat Graft

for Interposition After Tarsal Coalition Resection
Ronald M. Swonger, BA,* Jessica M. Bernstein, MD,† Olivia F. Perez, BS,* Alina Syros, MPH,*
Kevin S. Horowitz, MD,‡ and Verena M. Schreiber, MD‡
Level of Evidence: Level III, a retrospective comparative study

Background: Tarsal coalition is one of the most common foot between treatment groups.
and ankle pathologies in children, yet there is no consensus re-
garding what to interpose after resection. Fibrin glue could be Key Words: fibrin glue, tarsal coalition, fat graft, interposition,
considered, but the literature comparing fibrin glue to other in- calcaneonavicular coalition, talocalcaneal coalition, coalition
terposition types is sparse. The purpose of this study was to recurrence, wound complications, foot, Pediatrics
evaluate the effectiveness of fibrin glue for interposition com- (J Pediatr Orthop 2023;43:e370–e373)
on 04/15/2023
pared with fat graft by analyzing the rate of coalition recurrence

and wound complications. We hypothesized that fibrin glue
would have similar rates of coalition recurrence and fewer wound
complications compared with fat graft interposition.
Methods: A retrospective cohort study was performed examining
all patients who underwent a tarsal coalition resection at a free-
standing children’s hospital in the United States from 2000 to
T arsal coalition is an abnormal connection between two
or more tarsal bones caused by a failure in mesen-
chymal segmentation during foot development.1,2 Patients
2021. Only patients undergoing isolated primary tarsal coalition typically present between the ages of 8 years and 16 years
resection with interposition of fibrin glue or fat graft were in- with painful flatfoot and recurrent ankle sprains.3 The true
cluded. Wound complications were defined as any concern for an incidence of tarsal coalition is unknown, but estimates
incision site that prompted the use of antibiotics. Comparative vary between 1% and 13% making it one of the most
analyses were conducted using χ2 and Fisher exact test to ex- common foot and ankle pathologies in children.1 In pa-
amine relationships among interposition type, coalition re- tients with a symptomatic tarsal coalition, initial treat-
currence, and wound complications. ment should focus on minimizing pain.4,5 Nonoperative
Results: One hundred twenty-two tarsal coalition resections met management commonly fails to leave surgical intervention
our inclusion criteria. Fibrin glue was used for interposition in 29 as the only recourse.1 However, the optimal surgical
cases and fat graft was used in 93 cases. The difference in the treatment has not yet been determined.6
coalition recurrence rate between fibrin glue and fat graft inter- Resection of tarsal coalitions was first reported by
position was not statistically significant (6.9% vs. 4.3%, Badgley7 in 1927.8 Since that time, surgical resection of
P = 0.627). The difference in wound complication rate between tarsal coalitions without soft tissue interposition has
fibrin glue and fat graft interposition was not statistically sig- shown to have a high recurrence rate.8–10 In 1982,
nificant (3.4% vs 7.5%, P = 0.679). Cowell11 was the first to describe using the origin of the
Conclusion: Fibrin glue interposition after tarsal coalition re- extensor digitorum brevis (EDB) muscle as an inter-
section is a viable alternative to fat graft interposition. Fibrin position after tarsal coalition resection.8 In 2009,
glue has similar rates of coalition recurrence and wound com- Mubarak et al12 described using gluteal fat as an alter-
plications when compared with fat grafts. Given our results and native to EDB interposition.12 In 2017, Masquijo et al8
the lack of tissue harvesting required with fibrin glue, fibrin glue compared resections of calcaneonavicular coalitions with
may be superior to fat grafts for interposition after tarsal coali- the interposition of bone wax, fat graft, and EDB.8 The
tion resection. authors found that fat graft and bone wax interposition
provided better pain relief, gave better functional scores,
and avoided coalition reossification more effectively than
EDB interposition.8 Another possible interposition is fi-
From the *Miller School of Medicine; †Department of Orthopaedic
Surgery, University of Miami; and ‡Department of Orthopaedic
brin glue.5 Weatherall and Price5 reported that fibrin glue
Surgery, Nicklaus Children’s Hospital, Miami, FL. is a safe and reliable alternative to tissue grafts for inter-
None of the authors received financial support for this study. position after tarsal coalition resection.
The authors declare no conflicts of interest. To the author’s knowledge, there are no studies that
Reprints: Verena M. Schreiber, MD, Department of Orthopaedic Sur- compare fibrin glue and fat graft as interpositions after
gery, Nicklaus Children’s Hospital, 3100 SW 62nd Avenue, Miami,
FL 33155. E-mail: verena.schreiber@nicklaushealth.org. tarsal coalition resection. We sought to make this com-
Copyright © 2023 Wolters Kluwer Health, Inc. All rights reserved. parison and hypothesized that fibrin glue would have
DOI: 10.1097/BPO.0000000000002386 similar rates of coalition recurrence and fewer wound

J Pediatr Orthop Volume 43, Number 5, May/June 2023 Fibrin Glue vs Fat Graft Interposition
complications compared with fat graft interposition while relatively uncommon occurrences such as recurrence of
eliminating the need for tissue harvesting. coalition and wound complications, a sample size of N =
5 per treatment group provides 80% power (alpha = 0.05)
METHODS to detect a difference of 0.5 SDs.
Descriptive statistics were calculated for all variables.
Retrospective Chart Review Comparative analyses were conducted using χ2 and Fisher
After Institutional Review Board approval, a retro- exact test when appropriate to examine relationships among
spective cohort study was performed using a billing query interposition type, coalition recurrence, and wound compli-
to identify all patients who underwent a tarsal coalition cations. All statistics were performed using SAS Studio 3.81.
resection at a free-standing children’s hospital in the
United States from 2000 to 2021. Records were reviewed
for demographic data, procedures performed, coalition RESULTS
recurrence, and wound complications. Inclusion criteria Over the 22-year study period, tarsal coalition re-
were primary tarsal coalition resections with interposition section was performed on 140 patients (173 feet). Further, a
of fibrin glue or fat graft (gluteal or local fat). Exclusion review of charts yielded 97 patients (122 feet) that met the
criteria were surgeries performed before 2000, those per- inclusion criteria. There were 61 females and 61 males in-
formed at the same time as another deformity correction cluded in the final analysis, with a mean age of 12 ± 2 years
surgery, revision tarsal coalition resections, coalitions old. Of the 122 coalition resections, the most common
other than calcaneonavicular or talocalcaneal, inter- anatomic classification was calcaneonavicular (n = 85)
position other than fibrin glue or fat graft, and those followed by talocalcaneal (n = 37). Fibrin glue was used as
on 04/15/2023
without follow-up. Evicel (OMRIX Biopharmaceuticals an interposition in 29 cases and a fat graft (gluteal, n = 56
Ltd., Tel Aviv, Israel), Vistaseal (Instituto Grifols, S.A., or local fat, n = 37) was used as an interposition in 93 cases.
Barcelona, Spain), and Tisseel (Baxter Healthcare Corp, There was a total of 10 surgeons in the study, and the
Deerfield, IL) were the 3 brands of fibrin glue used for average length of follow-up was 14 months with a range of 2
interposition. Follow-up examinations were performed by weeks to 7 years. There was no statistical difference in pa-
the primary surgeon. Coalition recurrence was defined as tient demographics between interposition types (Table 1).
symptomatic reossification confirmed by a radiograph that Six patients had a symptomatic recurrence of the
subsequently underwent reoperation. Wound complica- coalition and underwent reoperation (Table 2). The
tions were defined as any concern for an incision site that difference in the coalition recurrence rate between fibrin
prompted the use of antibiotics. When gluteal fat was glue (n = 2) and fat graft (n = 4) interposition was not
used, it was noted whether the wound complication was statistically significant (6.9% vs 4.3%, P = 0.627). Eight
from the foot incision or the gluteal incision. Microsoft patients developed postoperative wound complications.
Excel was used to organize and document patient data. The difference in wound complication rate between fibrin
glue (n = 1) and fat graft (n = 7) interposition was not
Statistical Analyses statistically significant (3.4% vs 7.5%, P = 0.679).
A power analysis was performed between the two It is worth noting that calcaneonavicular and talo-
interposition groups (fibrin glue and fat graft). A sample calcaneal coalitions were grouped together for analysis.
size of N = 14 per treatment group provides 80% power However, when comparing the anatomic classification, the
(alpha = 0.05) to detect a difference of 0.5 SDs. A power rate of coalition recurrence between calcaneonavicular (n
analysis was also performed for complications. For = 5) and talocalcaneal (n = 1) coalitions was not statis-
tically significant (P > 0.5), and the rate of wound com-
TABLE 1. Patient Demographics plications between calcaneonavicular (n = 5) and
Fibrin Glue Fat Graft
talocalcaneal (n = 3) coalitions was not statistically sig-
Characteristics (n = 29) (n = 93) P nificant (P > 0.5).
It is also worth noting that we considered both local
Age 11.7 12.2 0.451*
fat and gluteal fat under the same fat graft classification.
Sex 0.288†
Female 17 (58.6) 44 (47.3) — In terms of wound complications, we felt it may be worth
Male 12 (41.4) 49 (52.7) — considering gluteal fat and local fat as separate categories
Ethnicity 0.154† because harvesting gluteal fat necessitates a separate in-
Hispanic 21 (72.4) 51 (54.8) — cision along the gluteal fold. We speculated that this sec-
White 4 (13.8) 22 (23.7) —
Black 3 (10.3) 4 (4.3) —
ond incision could be a conduit to increased infections.
Other 0 4 (4.3) — When comparing wound complication rates for fibrin
Unknown 1 (3.5) 12 (12.9) — glue, gluteal fat, and local fat [3.4% (n = 1), 8.9% (n = 5),
Classification 0.890† 5.4% (n = 2), P = 0.621], gluteal fat did demonstrate the
Calcaneonavicular 21 (72.4) 64 (68.8) — highest complication rate with 8.9%, although this was not
Talocalcaneal 8 (27.6) 29 (31.2) —
statistically significant. Of the 5 patients with a gluteal fat
Values are reported as averages and absolute frequency (percentage). interposition that developed a wound complication, 4 out
*Analysis of variance test.
†χ2 test. of 5 (80%) developed the complication at the gluteal in-
cision.

Swonger et al J Pediatr Orthop Volume 43, Number 5, May/June 2023
with fat graft and 6% of their patients who received in-

TABLE 2. Association Between Interposition Type and
Complications terposition with bone wax.8 We found no statistically
significant difference in the coalition recurrence rate be-
Fibrin glue Fat graft
tween fibrin glue and fat graft interpositions, and the re-
Characteristics (n = 29) (n = 93) P*
currence rates in our study between both interposition
Coalition recurrence 2 (6.9) 4 (4.3) 0.627 types were similar to other published literature.8,12
Wound complications 1 (3.4) 7 (7.5) 0.679

In terms of wound complications, 3.4% of our pa-
Values are reported as absolute frequency (percentage). tients with a fibrin glue interposition had a postoperative
*Fisher exact test. wound complication. Comparing this to published liter-
ature, Weatherall and Price5 reported that none of their 9
patients acquired a wound complication when using fibrin
DISCUSSION glue as the interposition.5 Of our patients, 7.5% treated
Fibrin has been used as a biological agent for over with a fat graft (gluteal or local fat) interposition devel-
100 years.13,14 Its use to establish surgical hemostasis was oped a wound complication. Comparing this to published
first reported by Bergel15 in the early 1900s.13 Fibrin glue literature, Masquijo et al8 reported that 13.0% of their
was first commercially available in 1982 with subsequent patients treated with a fat graft interposition developed a
Food and Drug Administration approval in 1998.13 Since wound complication.8 Although our patients with a fibrin
that time, fibrin glue has been successfully used as a bio- glue interposition developed fewer wound complications
material in a variety of specialties including orthopaedic compared with our patients with fat graft interposition, it
surgery.13 The biocompatibility, biodegradability, and
on 04/15/2023
did not rise to the level of statistical significance.

injectability of fibrin glue are traits that make it an at- This study is not without its limitations. Firstly, the
tractive option in a variety of procedures.16–19 It is the retrospective nature of the study comes with its own in-
only agent approved by the Food and Drug Admin- herent limitations. The small cohort size and data from a
istration for use as a hemostat, sealant, and adhesive.20 single institution limit the validity and generalizations of
Fibrin glue is composed of clotting proteins, mainly the conclusions. The study did not have sufficient power to
fibrinogen and Factor XIII, and freeze-dried thrombin, analyze the uncommon occurrences of coalition re-
which functions as a catalyst.21 The native biological currence and wound complications, which opens our re-
origin of these components contributes to the success and sults to a possible type II error. Postoperative imaging was
longevity of fibrin glue in most settings.22 The glue es- not obtained in every patient, and we were unable to
sentially mimics the final common pathway of coagulation comment on radiographic evidence of coalition re-
in vivo to induce tissue adhesion.23 In the setting of tarsal currence. We were only able to comment on symptomatic
coalition resections, Weatherall and Price5 suggested that recurrence that required revision surgery. Another limi-
fibrin glue works within the resected coalition by forming tation to consider is that we were unable to document
a barrier between the bony edges.5 This barrier then patient-reported outcome measures. The Oxford Ankle
functions to prevent coalition recurrence and scar tissue Foot Questionnaire for Children, the Foot and Ankle
from building up in the area.5 They also argued that the Outcome Instrument, and the 36-item Short Form Health
cost of fibrin glue is offset by the longer operating time Survey are all validated measures that could be used to
and higher morbidity associated with fat graft harvesting.5 gauge successful outcomes.1,5,24,25 We were unable to re-
In their retrospective study of 9 patients (12 feet) with port this data due to the lack of this documentation in the
calcaneonavicular coalitions, Weatherall and Price5 found reviewed charts. We were also unable to quantify the pre
that all 9 patients reported excellent outcomes, were able to and postoperative range of motion due to the lack of this
return to sports and had no postoperative complications documentation in the reviewed charts. In addition, be-
when using fibrin glue as the interposition.5 Likewise, we cause this study involved 10 surgeons and each surgeon
report positive outcomes when using fibrin glue as an in- had a preferred interposition and surgical technique, the
terposition in our study. In our cohort of 23 patients (29 surgeon himself or herself is a confounding variable that
feet) that underwent resection with interposition of fibrin limits the validity of the results. Lastly, we were unable to
glue, 20 patients (26 feet) reported no postoperative com- access information about whether patients may have
plications. The coalition recurrence rate was 6.9% (n = 2) sought reoperation at another institution.
and the wound complication rate was 3.4% (n = 1). The lack of literature leaves the orthopaedic com-
The coalition recurrence rate of fibrin glue inter- munity with no consensus on what to interpose after tarsal
position in our study (6.9%) is similar to the recurrence coalition resection. Although fat graft interposition is of-
rates of other interposition types in published ten a preferred technique, other materials such as bone
literature.5,8,12 Mubarak et al12 reported radiographic wax and fibrin glue have been used with promising
coalition recurrence in 13% and symptomatic coalition results.1,5,8 In our study, we found that interposition with
recurrence in 5% of their patients who underwent tarsal fibrin glue after tarsal coalition resection is a viable al-
coalition resection with interposition of fat graft.12 In the ternative to interposition with fat grafts. Fibrin glue has
only comparative study of different interpositions in the similar rates of coalition recurrence and similar rates of
literature, Masquijo et al8 reported radiographic re- wound complications when compared with fat graft in-
currence in 4% of their patients who received interposition terposition. To our knowledge, this study is the first to

J Pediatr Orthop Volume 43, Number 5, May/June 2023 Fibrin Glue vs Fat Graft Interposition
compare fibrin glue interposition to fat graft interposition 12. Mubarak SJ, Patel PN, Upasani VV, et al. Calcaneonavicular
after tarsal coalition resection. coalition: treatment by excision and fat graft. J Pediatr Orthop.
2009;29:418–426.
The authors encourage further studies directly
13. Patel S, Rodriguez-Merchan EC, Haddad FS. The use of fibrin glue
comparing fat graft, bone wax, and fibrin glue for re- in surgery of the knee. J Bone Joint Surg Br. 2010;92:1325–1331.
currence rates, wound complications, operative time, and
14. Carless PA, Henry DA, Anthony DM. Fibrin sealant use for
patient-reported outcome measures. Robust, randomized, minimising peri-operative allogeneic blood transfusion. Cochrane
prospective, multicenter studies comparing these inter- Database Syst Rev. 2003;2003:Cd004171.
position techniques are needed to further guide surgical 15. Bergel S. About the effects of fibrin [Ueber wirkungen des fibrins].
techniques and improve patient outcomes. DMW-German Med Week J. 1909;35:663–665.
16. Kim I, Lee SK, Yoon JI, et al. Fibrin glue improves the therapeutic
effect of MSCs by sustaining survival and paracrine function. Tissue
Eng Part A. 2013;19:2373–2381.
REFERENCES 17. Farra J, Zhuge Y, Neville HL, et al. Submucosal fibrin glue injection
1. Kothari A, Masquijo J. Surgical treatment of tarsal coalitions in for closure of recurrent tracheoesophageal fistula. Pediatr Surg Int.
children and adolescents. EFORT Open Rev. 2020;5:80–89. 2010;26:237–240.
2. Soni JF, Valenza W, Matsunaga C. Tarsal coalition. Curr Opin 18. Othman S, Messa CAIV, Elfanagely O, et al. Sticking to what
Pediatr. 2020;32:93–99. matters: a matched comparative study of fibrin glue and mechanical
3. Mosca VS. Subtalar coalition in pediatrics. Foot Ankle Clin. 2015;20:
fixation for split-thickness skin grafts in the lower extremity. Int J
265–281. Low Extrem Wounds. 2021:15347346211047748. doi:10.1177/
4. Docquier PL, Maldaque P, Bouchard M. Tarsal coalition in 15347346211047748.. [Online ahead of print].
paediatric patients. Orthop Traumatol Surg Res. 2019;105(suppl 1): 19. Kim YS, Choi YJ, Suh DS, et al. Mesenchymal stem cell
S123–s131. implantation in osteoarthritic knees: is fibrin glue effective as a
on 04/15/2023
5. Weatherall JM, Price AE. Fibrin glue as interposition graft for tarsal scaffold? Am J Sports Med. 2015;43:176–185.
coalition. Am J Orthop (Belle Mead NJ). 2013;42:26–29.
6. Krief E, Ferraz L, Appy-Fedida B, et al. Tarsal coalitions: 20. Spotnitz WD. Fibrin Sealant: the only approved hemostat, sealant,
and adhesive-a laboratory and clinical perspective. ISRN Surg.
preliminary results after operative excision and silicone sheet
2014;2014:203943.
interposition in children. J Foot Ankle Surg. 2016;55:1264–1270.
7. Badgley CE. Coalition of the calcaneus and the navicular. Arch Surg. 21. Canonico S. The use of human fibrin glue in the surgical operations.
1927;15:75–88. Acta Biomed. 2003;74(suppl 2):21–25.
8. Masquijo J, Allende V, Torres-Gomez A, et al. Fat graft and bone 22. Chen FM, Liu X. Advancing biomaterials of human origin for tissue
wax interposition provides better functional outcomes and lower engineering. Prog Polym Sci. 2016;53:86–168.
reossification rates than extensor digitorum brevis after calcaneona- 23. Panda A, Kumar S, Kumar A, et al. Fibrin glue in ophthalmology.
vicular coalition resection. J Pediatr Orthop. 2017;37:e427–e431. Indian J Ophthalmol. 2009;57:371–379.
9. Inglis G, Buxton RA, Macnicol MF. Symptomatic calcaneonavic- 24. Morris C, Doll HA, Wainwright A, et al. The Oxford ankle foot
ular bars. The results 20 years after surgical excision. J Bone Joint questionnaire for children: scaling, reliability and validity. J Bone
Surg Br. 1986;68:128–131. Joint Surg Br. 2008;90:1451–1456.
10. Swiontkowski MF, Scranton PE, Hansen S. Tarsal coalitions: long- 25. Johanson NA, Liang MH, Daltroy L, et al. American Academy of
term results of surgical treatment. J Pediatr Orthop. 1983;3:287–292. Orthopaedic Surgeons lower limb outcomes assessment instruments.
11. Cowell HR. Tarsal coalition–review and update. Instr Course Lect. Reliability, validity, and sensitivity to change. J Bone Joint Surg Am.
1982;31:264–271. 2004;86:902–909.

ORIGINAL ARTICLE
Is There a Relationship Between the Functional Level of

Juvenile and Adolescent Patients With Down Syndrome
and Hip Dysplasia?

María Galán-Olleros, MD,*† Ángel Palazón-Quevedo, MD,† Rosa M. Egea-Gámez, MD, PhD,‡
Ana Ramírez-Barragán, MD, PhD,*† J. Ignacio Serrano, Eng,§
and Ignacio Martínez-Caballero, MD, PhD*†
(P < 0.0005). A significant receiver operating characteristic curve

Background: The prevalence of hip dysplasia among patients was obtained for W-CEA with a cutt-off point at 26.4 degrees for
with Down syndrome (DS) is higher than in the general pop- level I (area under the curve = 0.763; P < 0.005; sensitivity =
ulation. We hypothesize that a relationship may exist between 0.800 and specificity = 0.644). There was a fairly high correla-
on 04/15/2023
functional level and hip dysplasia in DS, but this has not been tion between EI and TA (0.749; P < 0.0005), EI and W-CEA
studied to date. The aim of this study is to evaluate whether there (−0.817; P < 0.0005), and TA and W-CEA (−0.748; P < 0.0005).
is a relationship between functional level and radiographic pa- Numerous hips showed signs of acetabular retroversion, with no
rameters of hip dysplasia or other measures. significant differences found between functional levels or asso-
Methods: Retrospective cross-sectional comparative study of 652 ciation with hip dysplasia measures.
patients with DS from a pediatric referral center database. Patients Conclusions: The present study reveals a relationship between an
over 8 years of age with an anteroposterior pelvis radiograph and increased risk of hip dysplasia and reduced functional levels in
with no exclusion criteria were selected, totaling 132 patients (264 DS children older than 8 years. These findings may guide in-
hips; 54.55% females; mean age 12.96 ± 2.87 y). Several radio- dividualized clinical follow-up of hip development in DS children
graphic parameters of the acetabulum [Sharp angle (SA), Tönnis considering their functional level.
angle (TA), Wiberg center-edge angle (W-CEA), extrusion index Level of Evidence: Level III, retrospective comparative study.
(EI), and acetabular retroversion signs], the proximal femur [neck
shaft angle (NSA)], and joint congruence [Shenton line (SL)] were Key Words: down syndrome, hip dysplasia, gross motor func-
assessed. Patients were classified into 2 levels based on functional tion, functional level, motor skills, radiograph
skills. A multivariate association analysis was performed between (J Pediatr Orthop 2023;43:e311–e318)
radiographic parameters and functional level.
Results: Sixty-one patients were compatible with a functional
level I and 71 with a level II. Forty-six hips were dysplastic and
60 were borderline according to the W-CEA. A statistically sig-
nificant relationship was found between the categorical dis- T he prevalence of hip dysplasia in patients with Down
syndrome (DS) is estimated to be between 1% and 7%,
which exceeds that of the general population.1,2 As with
tribution of certain radiographic measurements of hip dysplasia
(EI, SA, TA, W-CEA, SL, and classification by functional level most musculoskeletal manifestations of DS, this condition
seems to be related to generalized ligamentous laxity, joint
hypermobility, and hypotonia.3 Structural abnormalities
From the *Neuro-Orthopaedic Unit; †Pediatric Orthopaedics; ‡Spine
Unit, Orthopaedic Surgery and Traumatology Department. Hospital in tissues seem to be caused by an increase in type VI
Infantil Universitario Niño Jesús; and §Neural and Cognitive En- collagen, which is partially encoded by genes located on
gineering group, Center for Automation and Robotics, CAR CSIC- chromosome 21.4 Patients with DS have been found to
UPM, Arganda del Rey, Madrid, Spain. present osseous abnormalities of the proximal femur and
Institutional Review Board approval was obtained for this study
(n°. R-0018/22). This study was performed in accordance with the
acetabulum, consisting of a retroverted,5–7 dysplastic, and
ethical standards in the 1964 Declaration of Helsinki. shallower acetabulum, with global insufficiency, compared
No funding was received for conducting this study. with controls8 as well as femoral anteversion and coxa
The authors declare no conflicts of interest. valga.2,9
Reprints: María Galán-Olleros, MD, Neuro-Orthopaedic Unit, Ortho-
paedic Surgery and Traumatology Department, Hospital Infantil
In addition to impaired mental ability, motor de-
Universitario Niño Jesús, Avenue Menéndez Pelayo 65, Madrid velopment is delayed in children with DS. Decreased
28009, Spain. E-mail: mgalanolleros@gmail.com. motor development seems closely related to the degree of
Supplemental Digital Content is available for this article. Direct URL hypotonia observed.10 Hypotonia in young children with
citations appear in the printed text and are provided in the HTML DS negatively influences muscle contraction and balance,
pedorthopaedics.com. leading to problems in postural control.11,12 This, com-
Copyright © 2023 Wolters Kluwer Health, Inc. All rights reserved. bined with generalized ligamentous laxity and joint
DOI: 10.1097/BPO.0000000000002370 hypermobility, has a significant functional impact, and

Galán-Olleros et al J Pediatr Orthop Volume 43, Number 5, May/June 2023
contributes to reduced levels of physical activity.13,14 Radiographic Evaluation

Motor ability varies widely in children with DS.15,16 Fine At our institution, anteroposterior radiographs of
and gross motor skills are both impaired in DS children, the pelvis are performed supine, with the legs in internal
though to differing degrees. Some authors have found that rotation, a film-focus distance of 120 cm, and the central
fine motor skills, in particular bimanual coordination, beam directed at the midpoint. Radiographs were con-
were more impaired than gross motor skills;15 however, sidered adequate if the sacrococcygeal joint was 1 to 3 cm
the gross motor ability seems to be a better predictor of from the superior aspect of the symphysis pubis, the
limitations in functional skills.16 Likewise, these in- greater trochanter laterally, and the lesser trochanter
dividuals present a wide spectrum of hip dysplasia, and partially superimposed on the femoral neck, the obturator
several factors have been proposed to influence its occur- rings, and acetabular teardrops symmetrical, and the
rence and degree.2,5,6,8,9,17 midsacral line aligned with the symphysis pubis.
The functional level has been related to hip dys- Plain radiographs were digitally transferred, and
plasia in other neuromuscular disorders, such as cerebral different parameters of the acetabulum and proximal fe-
palsy, where the risk and severity of the hip subluxation mur were measured by 2 experienced pediatric orthopae-
and dislocation are closely related to the patient’s gross dic surgeons (M.G.O. and A.P.Q.) using the Syngo.plaza
motor function.18 The pediatric hip is particularly PACS software (Siemens Healthcare 2022). For greater
susceptible to change during development,19 and func- reliability, each investigator performed the assessment
tion determine the bone shape and structure of the hip. twice, with a 4 weeks minimum interval between each, and
The morphology and stability of the hip joint during the mean of these measurements was used. Investigators
on 04/15/2023
development are influenced by loading, movement, and were blinded to the patient’s functional status. Interob-
the action of muscle forces.20,21 Greater motor impair- server reliability for the radiologic evaluation was calcu-
ment leads to less bearing and movement, resulting in lated with the Cohen k coefficient, showing a high degree
reduced reaction forces at the hip during development, of consistency between the 2 observers, ranging from 0.84
thereby predisposing patients to morphologic alterations to 0.89.
of the proximal femur and acetabulum. Furthermore, Acetabular coverage was analyzed quantitatively
hip instability can result in progressive loss of mobility using Sharp angle (SA), the acetabular roof angle of
and function.22 Tönnis (TA) or acetabular inclination, the center-edge
The relationship between functional level and hip angle of Wiberg (W-CEA), and the extrusion index (EI),
dysplasia in DS has not been studied thus far. However, whereas the presence or absence of the crossover sign
we hypothesize that there may be a relationship between
these 2 variables. The aim of this study is to evaluate
whether a relationship exists between the functional level
of DS patients and the radiographic parameters of hip
dysplasia or other radiographic measures.
METHODS
Study Design and Population
After the approval from our Institutional Review
Board (n°. R-0018/22), a retrospective cross-sectional
comparative study was conducted. A search of the data-
base of our pediatric referral center revealed 652 patients
diagnosed with DS from 1999 to 2021, and of these, we
identified 169 patients with an anteroposterior pelvis ra-
diograph performed over age 8 years. Several radiographic
measures of the acetabulum were assessed, and medical
records were reviewed to obtain information on functional
skills. Exclusion criteria were inadequate radiographs,
incomplete medical records preventing proper functional
assessment, and the presence of coexisting diseases that
could affect hip morphology. Five were excluded for
having Perthes disease, one because of spastic tetraplegia,
and one for slipped capital femoral epiphysis. Inadequate
radiographs led to the exclusion of 26 patients and in-
complete medical records to exclusion of 4 more subjects FIGURE 1. Study flow chart. AP indicates anteroposterior; CP,
(Fig. 1). The final study group consisted of 132 patients cerebral palsy; DS, Down syndrome; LCPD, Legg-Calvé-Perthes
(264 hips). disease; SCFE, slipped capital femoral epiphysis.

J Pediatr Orthop Volume 43, Number 5, May/June 2023 Function and Hip Dysplasia in Down Syndrome
(COS), posterior wall sign (PWS), and ischial spine sign Therefore, the financial level of the families had no
(ISS) were used as qualitative indicators of reduced ace- influence on this classification.
tabular version, acetabular retroversion, or posterior ace- Although some specialists often use adapted scales
tabular wall deficiency. The retroversion index (RI) was originally created for other disorders, such as the Gross
calculated only in cases with findings suggestive of ace- Motor Function Classification System,23–25 doubts have
tabular retroversion. The anatomy of the femoral neck in been raised concerning the use of this classification system
the coronal plane was described by the femoral neck-shaft in children with DS.26 Also, as reported,27 we found that
angle (NSA), and the Shenton line (SL) was used to assess the gait skills of patients with DS are unique in certain
joint congruence and was classified as continuous or dis- ways, as severe functional limitations are rare in these
rupted. A description of the radiographic parameters and patients, though they are more likely to require caregiver
the reference values used can be seen in Table 1. Figure 2 assistance and supervision due to syndrome-related in-
is a graphical representation of all the measurements tellectual disability and hypotonia. Moreover, DS children
performed on each patient. scarcely use walking aids such as crutches or a walker and
generally do not use ankle-foot orthoses. In addition,
younger DS children display more functional variability,
Functional Classification while function in older children is relatively more
Patients were classified into 2 levels according to consistent.28 For these reasons, this simple classification
their functional skills, with level I broadly representing was developed, which allowed patients to be divided into 2
patients who participated and were actively engaged in distinct groups according to their functional level.
on 04/15/2023
community-based sports and recreational activities and

level II comprising those who did not. Other details used Statistical Analyses
to classify patients by functional level can be found in All data were analyzed using SPSS Statistics soft-
Figure 3, which was developed by the authors. It should ware, version 23 (IBM Inc., Chicago, IL). Continuous
also be noted that in our environment there are no major quantitative data were expressed as mean and SD,
differences in access to education and care for children whereas categorical variables were reported as frequency
with DS and that all of them can practice adapted sports and percentage values. A multivariate analysis was per-
and actively participate in community recreational formed to compare numerical radiographic parameters
activities if they have the motor skills to do so. between functional levels using the multivariate analysis of
Table 1. Description of the Radiographic Parameters Measured in Each Patient on Anteroposterior Pelvic Radiographs and
Reference Values Used in the Study
Radiographic parameter Measurement method Reference values
Sharp’s angle (SA) Angle between an HRL and a line through the caudal tip of the teardrop < 32º, overcoverage
and the lateral edge of the acetabulum. 33-38º, normal
39-42, borderline
> 42º, dysplasia
Acetabular roof angle of Angle formed by an HRL and a line through the most medial point of > 10º, acetabular dysplasia
Tönnis (TA) the sclerotic zone of the acetabular roof and the lateral edge of the -10º to 10º, normal
acetabulum. < -10º, overcoverage
Wiberg center-edge angle Angle formed by a VRL and a line connecting the center of the FH with < 15º, severe dysplasia
(W-CEA) the lateral edge of the acetabular sourcil (the weightbearing area of the < 20º, mild dysplasia
acetabulum). 20-25º, borderline
dysplasia
> 25º, normal
> 42º, overcoverage
Extrusion index (EI) Percentage of uncovered FH (B) in comparison to the total horizontal < 30%, normal
head diameter (A). 30-60%, hip at risk
> 60%, partial extrusion
> 100%, complete
extrusion
Crossover sign (COS) Positive if the projected anterior wall crosses the PW laterally. Positive / Negative
Posterior wall sign (PWS) Positive if the PW runs medially to the center of the FH. Positive / Negative
Ischial spine sign (ISS) Positive if IS is projected medially to the pelvic brim. Positive / Negative
Retroversion index (RI) Percentage of retroverted acetabular opening (C) divided by the entire %
opening (D).
Neck-shaft angle (NSA) Angle formed by the femoral neck axis and femoral shaft axis. < 120º, coxa vara
120-140º, normal
> 140º, coxa valga
Shenton’s line (SL) Harmonic arc formed by a line between the caudal femoral head and Continuous / Disrupted
neck contour and superior border of the obturator foramen.
FH indicates femoral head; HRL, horizontal reference line; IS, ischial spine; PW, posterior wall; VRL, vertical reference line.
A, B, C and C are represented in figure 1.

SL (P < 0.0005), but not for NSA (P = 0.186), COS (P =

0.098), PWS (P = 0.074), ISS (P = 0.193), or RI (P =
0.106) (Table 2, Fig. 5).
In addition, the receiver operating characteristic
curve for W-CEA displayed as a binary classification
system (level I: yes/no), had a discrimination threshold at

26.4 degrees for level I (area under the curve = 0.763; P <
0.005), with a sensitivity of 0.800 and specificity of 0.644
(Fig. 6).
An analysis of the relationship between the category
of both hips from a patient and the functional level for the
same patient yielded similar results, with statistically sig-
nificant differences for EI (P < 0.0005), SA (P = 0.012),
TA (P < 0.0005), and W-CEA (P < 0.0005) (Electronic
Supplemental Table E-1, Supplemental Digital Content 1,
http://links.lww.com/BPO/A587 and Electronic Supple-
mental Fig. E-1, Supplemental Digital Content 2, http://
FIGURE 2. Different measurements performed in each patient links.lww.com/BPO/A588).
using AP pelvic radiographs. A represents total horizontal head
Correlation Between Radiographic Measures
on 04/15/2023
diameter; B, percentage of femoral head uncovered by the

acetabular roof; C, percentage of retroverted acetabular By analyzing the correlation between the different
opening; D, entire acetabular posterior wall. AP indicates an- radiographic measures of hip dysplasia, a fairly high
teroposterior; COS, crossover sign; EI, extrusion index; HRL, correlation was found between EI and TA (0.749; P <
horizontal reference line; ISS, ischial spine sign; NSA, neck-shaft 0.0005), EI and W-CEA (−0.817; P < 0.0005), and TA
angle; PWS, posterior wall sign; RI, retroversion index; SA, and W-CEA (−0.748; P < 0.0005), but not for other pa-
Sharp angle; SL, Shenton line; TA, Tönnis acetabular roof an-
rameters such as NSA or RI. Likewise, a statistically
gle; W-CEA, Wiberg center-edge angle.
significant difference was obtained when comparing
quantitative radiographic measures of hip dysplasia and
variance test, with Tukey correction for pairwise post hoc the continuous or disrupted SL as a qualitative indicator:
comparisons. The χ2 test was used to compare the dis- EI [18.6 (10.1) vs 44.3 (21); P < 0.005], SA [41.9 (5.9) vs 47
tributions of categorical variables between functional (4.4); P < 0.005], TA [7.8 (5.6) vs 18.7 (5.8); P < 0.005),
levels. The Pearson correlation coefficient was used to find W-CEA [28.9 (9) vs 8.1 (14.2); P < 0.005), and NSA [142.3
the association between the different radiographic mea- (7.7) vs 153.9 (8.3); P < 0.005]. However, no significant
sures of the hip. All tests were 2-tailed, and statistical differences were found between radiographic values of hip
significance was set at P values <0.05. For statistically dysplasia and the presence of signs of acetabular retro-
significant differences, we also reported the effect size version (COS, PWS, and ISS).
(partial eta squared, denoted as η2p) and the observed
power (1-ß).
DISCUSSION
Due to the higher prevalence of hip dysplasia in patients
RESULTS with DS and given our hypothesis of a possible relationship
Demographic Data
Sixty-one patients were compatible with functional
level I (46.21%) and 71 with level II (53.59%). The mean
age was 12.96 ± 2.87 years and 54.55% (72) were females,
with a similar distribution between functional groups for
age and sex (P = 0.061 and 0.914, respectively), showing
that both groups were comparable in this regard.
Comparison Between Radiographic Parameters

and Functional Groups
Statistically significant differences were found be-
tween functional groups for EI, SA, TA, W-CEA
(P < 0.0005), and NSA (P = 0.002) (Table 2; Fig. 4).
When comparing the distribution of certain radio-
graphic parameters of hip dysplasia by category between
the different groups of functional level, a statistically sig-
nificant relationship was also found for EI (P < 0.0005), FIGURE 3. Classification of Down syndrome patient’s func-
SA (0.002), TA (P < 0.0005), W-CEA (P < 0.0005), and tional skills developed by the authors.

Table 2. Mean ± standard Deviation And Distribution by Category—Percentage (Number of Participants)—of Radiographic
Variables for the Cohort Studied per Functional Group and Statistical Differences Between Groups
Level I Level II Statistics
EI (%) 15.3 (7.8) 26 (15.9) F(1,262) = 45.685, P < .0005, η2p = .148, 1-β = 1
χ2(2) = 31.923, P < .0005
At risk 2.46% (3) 24.65% (35)

Subluxated 0.00% (0) 3.52% (5)

Normal 97.54% (119) 71.83% (102)
SA (degrees) 40.5 (5.8) 44 (5.6) F(1,262) = 24.625, P < .0005, η2p = .086, 1-β = .999
Dysplasia 37.70% (46) 64.79% (92) χ2(3) = 20.007, P < .0005
Borderline 35.25% (43) 22.54% (32)
Normal 24.59% (30) 11.27% (16)
Overcoverage 2.46% (3) 1.41% (2)
TA (degrees) 6.2 (4.1) 11.1 (7.4) F(1,262) = 42.753, P < .0005, η2p = .140, 1-β = 1
Acetabular dysplasia 13.93% (17) 59.15% (84) χ2(1) = 56.809, P < .0005
Normal 86.07% (105) 40.85% (58)
W-CEA (degrees) 32.2 (6.7) 22.8 (12.1) F(2,261) = 58.673, P < .0005, η2p = .183, 1-β = 1
Severe dysplasia 0.82% (1) 19.01% (27) χ2(4) = 63.072, P < .0005
Mild dysplasia 0.82% (1) 11.97% (17)
Borderline dysplasia 13.93% (17) 30.28% (43)
Normal 77.87% (95) 35.92% (51)
Overcoverage 6.56% (8) 2.82% (4)
NSA (degrees) 141.6 (7.2) 144.9 (9.2) F(1,240) = 9.594, P = .002, η2p = .038, 1-β = .870
on 04/15/2023
Coxa valga 56.36% (62) 66.66% (88) χ2(2) = 3.367, P = .186

Normal 43.64% (48) 33.33% (44)
SL disrupted 0.82% (1) 14.49% (20) F(1,258) = 17.783, P < .0005, η2p = .064, 1-β = .987
Acetabular retroversion signs
Positive COS 33.61% (41) 38.73% (55) F(1,226) = 2.641, P = .098
Positive PWS 23.77% (29) 34.51% (49) F(1,224) = 9.239, P = .074
Positive ISS 20.49% (25) 26.06% (37) F(1,231) = 3.325, P = .193
RI (%) 31.55 (13.43) 34.73 (17.43) F(1,95) = 2.664, P = .106
Numbers in the same row not sharing subscript letters present statistically significant post hoc differences.
1-β indicates observed statistical power; COS, Crossover sign; EI, extrusion index; ISS, Ischial spine sign; NSA, Neck-shaft angle of the femur; PWS, Posterior wall sign;
RI, Retroversion index; SA, Sharp’s angle; SL, Shenton’s line; TA, Acetabular roof angle of Tönnis; W-CEA, Wiberg center-edge angle; η2p, effect size.
Bold data represent statistically significant differences.
with functional level, in the present study, we have analyzed a symptoms had a more retroverted and shallower acetab-
total of 132 patients and 264 hips, measuring different radio- ulum, with an overall reduction in femoral head coverage
graphic parameters of acetabular dysplasia as well as signs of compared with healthy subjects.8 Aggregating the radio-
acetabular retroversion. Our aim was to establish whether graphic values of both groups, our results suggest that 46
these variables were related to functional skills in DS patients hips (17.42%) showed signs of acetabular dysplasia and 60
divided into 2 levels (61 level I and 71 level II). (22.73%) were borderline hips according to the W-CEA,
Traditionally, it has been assumed that patients with DS which is the most sensitive parameter for assessing ace-
have less mobility than the general population, which causes tabular dysplasia, whereas 101 (38.26%) were dysplastic
lifestyle limitations and a low life expectancy.1 Nowadays, based on the TA. According to the EI, there were 38
however, this concept has changed, and these patients cur- (14.39%) at risk and 5 (1.89%) subluxated hips, whereas
rently enjoy a longer life expectancy, with greater mobility and 150 (56.82%) showed valgus morphology.
participation in sports activities as key factors,29,30 although Although previous studies have mentioned ace-
motor skills vary widely in children with DS. tabular retroversion as a cause of hip instability in patients
Different studies have evaluated acetabular mor- with DS6 due to a deficiency of the posterior wall of the
phology in patients with DS to identify the presence of acetabulum, other studies have found no such relation-
acetabular dysplasia or the existence of acetabular retro- ship, suggesting that hip instability is associated with dif-
version. The study by Abousamra et al7 included 23 pa- ferent degrees of acetabular retroversion; furthermore,
tients with available computed tomography (CT) or normal acetabular anteversion and even moderately in-
magnetic resonance imaging scans out of a total of 308 creased anteversion are possible findings.7 Also, Bulat
patients with DS; among the patients who had undergone et al8 found that hip instability among patients with DS
these imaging tests, the authors identified 13 unstable hips was associated with worse overall acetabular insufficiency
and 26 stable hips, with no differences in acetabular ver- and greater femoral anteversion, though not with more
sion between them, although there was an overall decrease severe acetabular retroversion. We found numerous hips
in acetabular anteversion. Another study evaluated 42 in our study with signs suggestive of retroversion [96 hips
patients with DS and hip symptoms with a CT scan versus (36.36%) with COS with a mean RI of 32.4 ± 16%], as
42 healthy patients who had a CT scan requested due to mentioned in previous studies.6,17 However, we found no
abdominal pain, finding that patients with DS and hip association between the presence of signs of acetabular

retroversion and the radiographic values of hip dysplasia and W-CEA, meaning that the correlation with func-
and found no relationship with their functional level. tional level was higher as both hips of a patient had a
Our research is novel in that it explores the rela- dysplastic morphology.
tionship between the radiographic morphology of hip The finding of a cutoff point of 26.4 degrees for
dysplasia and functional level in DS children over 8 years W-CEA, as indicated by the receiver operating characteristic
of age, regardless of the presence of symptoms or not. curve, which enables patients to be classified as functional
Our results suggest that there are significant differences level 1 or 2, has clinical relevance for patient follow-up.
between active and sedentary DS patients in terms of EI, Thus, patients compatible with a functional level I, who
SA, TA, W-CEA, and NSA, both quantitatively and by have few limitations and can practice adapted sports and
category (except for the category of coxa valga). How- actively participate in community recreational activities,
ever, no significant differences were found between have a very low risk of presenting radiographic signs of hip
functional levels with respect to signs of acetabular ret- dysplasia, whereas patients with functional abilities com-
roversion. When analyzing both hips from the same patible with level II, who are more sedentary, neither prac-
patient according to a functional group, statistically tice adapted sports nor actively participate in community
significant differences were also observed for EI, SA, TA, recreational activities, and have some limitations in walking,
on 04/15/2023
FIGURE 4. Boxplots representing the statistically significant differences between functional levels for the parameters studied. EI
indicates extrusion index; NSA, neck-shaft angle; SA, Sharp angle; TA, Tönnis acetabular roof angle; W-CEA, Wiberg center-
edge angle.

on 04/15/2023
FIGURE 5. Bar plots representing the distribution of each radiographic parameter by category. EI indicates extrusion index; SA,
Sharp angle; SL, Shenton line; TA, Tönnis acetabular roof angle; W-CEA, Wiberg center-edge angle.
have a higher risk of radiographic signs of hip dysplasia. patients with DS. In fact, to our knowledge, this is the
Therefore, follow-up of DS patients older than 8 years can largest series published to date analyzing plain hip ra-
be further individualized considering their functional level, diographs of patients with DS to assess morphology. In
and attention should be paid to the increased likelihood of addition, to assess acetabular retroversion, only signs
hip dysplasia in those with worse gross motor functional suggestive of acetabular retroversion were used, and this
abilities. suspicion was not confirmed by further imaging tests.
There are some limitations to this study. First, Also, EI, SL, and NSA vary with leg rotation and with
radiographs were evaluated based on availability, re- femoral anteversion, which cannot be measured on plain
gardless of the reason, for which they were requested, the radiographs. Third, as this was a retrospective cross-
vast majority being requested by social pediatricians to sectional study, the functional classification of the pa-
evaluate the hips of patients with DS. Therefore, there tients was based on the interpretation of medical records.
could be a selection bias, although we consider the Although our analysis simplified the classification into 2
sample to be fairly representative as most of the patients functional levels, we consider that these generally define
did not show symptoms. Second, our results are based patients with DS, and we were able to obtain significant
on plain radiographic measurements, which are less ac- intergroup differences in relation to the radiographic
curate than those performed by CT or magnetic reso- parameters of dysplasia. This approach has also taken
nance imaging scan, although radiography is the most into account that functional assessment in DS patients is
widely available imaging test in the clinic and, therefore, challenging, and this classification, which is not vali-
the one used for radiologic follow-up of the hips of dated, is simple and can be easily used in pediatric

5. Woolf SK, Gross RH. Posterior acetabular wall deficiency in Down

syndrome. J Pediatr Orthop. 2003;23:708–713.
6. Sankar WN, Schoenecker JG, Mayfield ME, et al. Acetabular
retroversion in Down syndrome. J Pediatr Orthop. 2012;32:277–281.
7. Abousamra O, Bayhan IA, Rogers KJ, et al. Hip instability in Down
syndrome: a focus on acetabular retroversion. J Pediatr Orthop.
2016;36:499–504.
8. Bulat E, Maranho DA, Kalish LA, et al. Acetabular global insufficiency

in patients with Down syndrome and hip-related symptoms: a matched-
cohort study. J Bone Joint Surg Am. 2017;99:1760–1768.
9. Knight DM, Alves C, Wedge JH. Femoral varus derotation
osteotomy for the treatment of habitual subluxation and dislocation
of the pediatric hip in trisomy 21: a 10-year experience. J Pediatr
Orthop. 2011;31:638–643.
FIGURE 6. Plotting of the ROC curves for all radiographic 10. Harris SR. Relationship of mental and motor development in Down’s
parameters to discriminate functional level I. Best discrim- syndrome infants. Phys Occupat Ther Pediatr. 1981;1:13–18.
ination at a threshold of 26.35 according to the W-CEA grad- 11. Block ME. Motor development in children with Down syndrome: a
ing system. AUC indicates area under the curve; EI, extrusion review of the literature. Adapted Phys Activity Q. 1991;8:179–209.
index, FL, functional level; NSA, neck-shaft angle, ROC, re- 12. Lauteslager PEM, Vermeer A, Helders PJM. Disturbances in the
ceiver operating characteristic; SA, Sharp angle, TA, Tönnis motor behaviour of children with Down’s syndrome: the need for a
acetabular roof angle, W-CEA, Wiberg center-edge angle. theoretical framework. Physiotherapy. 1998;84:5–14.
13. Caird MS, Wills BP, Dormans JP. Down syndrome in children: the role
of the orthopaedic surgeon. J Am Acad Orthop Surg. 2006;14:610–619.
on 04/15/2023
14. Chang CL, Kubo M, Ulrich BD. Emergence of neuromuscular

neuro-orthopaedic clinical practice, making these find- patterns during walking in toddlers with typical development and
ings easily applicable to routine settings. with Down syndrome. Hum Mov Sci. 2009;28:283–296.
As correlation does not imply causation, the associa- 15. Spano M, Mercuri E, Rando T, et al. Motor and perceptual-motor
tions observed in our study do not indicate whether a func- competence in children with Down syndrome: variation in perfor-
mance with age. Eur J Paediatric Neurol. 1999;3:7–13.
tional impairment is the cause of hip dysplasia in DS patients 16. Volman MJ, Visser JJ, Lensvelt-Mulders GJ. Functional status in 5 to
or vice versa. It seems more logical to view these relation- 7-year-old children with Down syndrome in relation to motor ability
ships as being both bidirectional and multifactorial in nature, and performance mental ability. Disabil Rehabil. 2007;29:25–31.
such that the hypotonia, hypermobility, and hyperlaxity that 17. Eshuis R, Boonzaaijer M, van Wieringen H, et al. Assessment of the
characterize patients with DS potentially contribute to both relationship between joint laxity and migration of the hip in children
with Down syndrome. J Child Orthop. 2012;6:373–377.
hip dysplasia and lower functionality, and these 2 influence 18. Hägglund G, Pettersson K, Czuba T, et al. Incidence of scoliosis in
each other reciprocally, as reduced mobility influences hip cerebral palsy. Acta Orthop. 2018;89:443–447.
dysplasia and increased hip dysplasia has an influence on 19. Giorgi M, Carriero A, Shefelbine S. Effects of normal and abnormal
functionality. Despite the inability of our study to demon- loading conditions on morphogenesis of the prenatal hip joint:
application to hip dysplasia. J Biomech. 2015;48:3390–3397.
strate a causal relationship, our findings do show a clear 20. Sharir A, Stern T, Rot C. Muscle force regulates bone shaping for
relationship between hip dysplasia and functional level, optimal load-bearing capacity during embryogenesis. Development.
which has not been previously reported. The relevance of this 2011;138:3247–3259.
finding lies precisely in the fact that this association may raise 21. Ford CA, Nowlan NC, Thomopoulos S, et al. Effects of imbalanced
clinician awareness of the increased likelihood of hip dys- muscle loading on hip joint development and maturation. J Orthop
Res. 2017;35:1128–36.
plasia in DS children older than 8 years with reduced am- 22. Rako K, Ranade S, Allen A. Orthopaedic Management in Down
bulatory function and lead to more individualized Syndrome: Current Concept Review. JPOSNA [Internet]. 2021 May 1
monitoring of the hip based on their functional level. [cited 2023 Feb. 11];3(2). Available from: https://www.jposna.org/
index.php/jposna/article/view/283
ACKNOWLEDGMENTS 23. Palisano R, Rosenbaum P, Walter S, et al. Development and
reliability of a system to classify gross motor function in children
The authors thank Dr Julian Lirio Casero, with cerebral palsy. Dev Med Child Neurol. 1997;39:214–223.
Dr Azucena Retuerta Oliva, and Dr Enrique Villalobos- 24. Russell DJ, Rosenbaum PL, Wright M, et al. Gross Motor Function
Pinto (Social Pediatrics Unit, Pediatrics Department) for Measure (GMFM-66 & GMFM 88) User’s Manual, 2nd edn.
their thorough evaluation and continuous follow-up of pa- London: Mac Keith Press,; 2013.
25. Moriyama CH, Massetti T, Crocetta TB, et al. Systematic review of
tients with Down syndrome over the years at our institution. the main motor scales for clinical assessment of individuals with
Down syndrome. Dev Neurorehabil. 2020;23:39–49.
REFERENCES 26. Towns M, Rosenbaum P, Palisano R, et al. Should the gross motor
1. Bennet GC, Rang M, Roye DP, et al. Dislocation of the hip in function classification system be used for children who do not have
trisomy 21. J Bone Joint Surg Br. 1982;64:289–294. cerebral palsy? Dev Med Child Neurol. 2018;60:147–154.
2. Shaw ED, Beals RK. The hip joint in Down’s syndrome. A study of 27. Leonard S, Msall M, Bower C, et al. Functional status of school-aged
its structure and associated disease. Clin Orthop Relat Res. 1992;278: children with Down syndrome. J Paediatr Child Health. 2002;38:160–165.
101–107. 28. Bodkin AW, Robinson C, Perales FP. Reliability and validity of the
3. Foley C, Killeen OG. Musculoskeletal anomalies in children with Down gross motor function classification system for cerebral palsy. Pediatr
syndrome: an observational study. Arch Dis Child. 2019;104:482–487. Phys Ther. 2003;15:247–252.
4. Karousou E, Stachtea X, Moretto P, et al. New insights into the 29. Eyman RK, Call TL. Life expectancy of persons with Down
pathobiology of Down syndrome–hyaluronan synthase-2 overex- syndrome. Am J Ment Retard. 1991;95:603–612.
pression is regulated by collagen VI α2 chain. FEBS J. 2013;280: 30. Winell J, Burke SW. Sports participation of children with Down
2418–2430. syndrome. Orthop Clin North Am. 2003;34:439–443.

ORIGINAL ARTICLE
Magnetic Resonance Imaging in the Management

of Significantly Displaced Adolescent Posterior
Sternoclavicular Joint Injuries

Graham Tytherleigh-Strong, MBBS, Rory Cuthbert, MBBS,
Freideriki Poutoglidou, MD, PhD, and Quen Tang, MBBS
Level of Evidence: Level IV—case series.

Background: Computed tomography (CT) scans are the standard
imaging modality for the diagnosis and treatment guide for Key Words: posterior sternoclavicular joint injury, physeal fracture,
adolescent posterior sternoclavicular joint (SCJ) injuries. How- CT scan, MRI scan
ever, the medial clavicular physis is not visualized and it is not (J Pediatr Orthop 2023;43:e374–e382)
on 04/15/2023
possible to differentiate between a true SCJ dislocation and a

physeal injury (PI). An magnetic resonance imaging (MRI) scan
can visualize the bone and the physis.
Methods: We treated a series of patients with adolescent poste-
rior SCJ injuries diagnosed on CT scan. Patients underwent an
P osterior sternoclavicular joint (SCJ) injuries are very
rare in all age groups, accounting for <5% of all
shoulder girdle injuries. They are particularly challenging
MRI scan to differentiate between a true SCJ dislocation and a in adolescents where the medial clavicular physis is the last
PI and to further differentiate between a PI with or without growth plate in the body to begin ossification (18 to 20 y)
residual medial end clavicular bone contact. Patients with a true and the injury may consist of either a true SCJ dislocation
SCJ dislocation and a PI with no contact underwent an open or a medial clavicular physeal injury (PI) with displace-
reduction and fixation. Patients with a PI with contact were ment. As a result, the optimal management of a posterior
treated nonoperatively with repeat CT scans at 1 and 3 months. SCJ injury in an adolescent can vary significantly.
At final follow-up SCJ clinical function was assessed Traditionally the imaging modality of choice for an
using Quick-DASH, Rockwood, modified Constant, and single adolescent posterior SCJ injury has been either a computed
assessment numeric evaluation (SANE) scores. tomography (CT) or a CT arteriogram. Whilst this can ac-
Results: Thirteen patients (2 female and 11 male) with an average curately assess the position of the medial end of the clavicle
age of 14.9 years (12 to 17) were included in the study. Twelve in relation to the sternum it is rarely able to differentiate
patients were available at final follow-up (mean 50 mo, 26 to 84). between a true dislocation, a PI and, when a PI has occurred,
One patient had a true SCJ dislocation and 3 had an off-ended PI any residual contact with the medial end of the clavicle. The
and were treated with an open reduction and fixation. Eight difficulty in differentiating posterior SCJ dislocations from
patients had a PI with residual bone contact and were treated physeal fractures on preoperative imaging studies has pre-
nonoperatively. For these patients serial CT scans showed that viously been noted by several authors.1,2 Generally, a sig-
the position was maintained, with a serial increase in callus nificantly displaced posterior SCJ is treated with an initial
formation and bone remodeling. The average follow-up was closed reduction and, if not successful, an open reduction. It
42.9 months (24 to 62). At final follow-up the mean Quick- is only when an open procedure is undertaken that diagnosis
disabilities of the arm, shoulder and hand (DASH) was 0.4 (0 to between a true dislocation and a PI can be made.
2.3), Rockwood was 15, modified Constant was 98.8 (89 to 100) Magnetic resonance imaging (MRI) scans of sig-
and SANE was 99.5% (95 to 100). nificantly displaced adolescent SCJ injuries can clearly dif-
Conclusion: In this case series of significantly displaced adoles- ferentiate between a true dislocation and a PI and, when
cent posterior SCJ injuries MRI scans allowed identification of there is a PI, whether there is any residual contact with the
true SCJ dislocations and off-ended PIs, which were successfully medial end of the clavicle and the presence of an intact
treated by open reduction, and PIs with residual physeal contact periosteal sleeve.3,4 We have used MRI scans to guide the
which were successfully treated nonoperatively. treatment of a series of adolescent patients with significantly
displaced posterior SCJ injuries. Patients where there was a
From the Division of Orthopaedics, Addenbrooke’s Hospital, Cambridge true SCJ dislocation or a PI where there was no contact with
University Hospitals Trust, Hills Road, Cambridge, UK. the medial clavicle were treated operatively with an open
The authors declare no conflicts of interest. reduction and fixation and patients where there was any
Reprints: Graham Tytherleigh-Strong, MBBS, Addenbrooke’s Hospital, residual contact between the physes and the medial end of
Cambridge CB2 2QQ, United Kingdom. E-mail: graham.tytherleigh-
strong@addenbrookes.nhs.uk. the clavicle were managed nonoperatively.
Copyright © 2023 Wolters Kluwer Health, Inc. All rights reserved. The purpose of this study was to assess the clinical
DOI: 10.1097/BPO.0000000000002378 outcomes in a series of patients in which the diagnosis and

J Pediatr Orthop Volume 43, Number 5, May/June 2023 Adolescent Posterior SCJ Injuries
initial management of significantly displaced posterior anterior-posterior length of the physis compared with the
SCJ injuries was based on the MRI findings. We anterior-posterior length of the residual contact between the
hypothesized that using an MRI scan to assess the extent posterior end of the physis and the anterior edge of the
of the physeal component of the SCJ injury as a guide to clavicle.
treatment would result in a decrease in the requirement for Patients where there was a true SCJ dislocation or
surgical intervention to obtain a good clinical outcome. where there was no contact between the medial end of the
clavicle and the physis, even if the periosteal sleeve was

METHODS intact, were treated with an open reduction and fixation.
Patients that had sustained a significantly displaced For patients with a true SCJ dislocation the joint was
posterior SCJ injury between January 2016 and February reduced and the anterior capsule repaired with sutures. We
2020 were included in this study. We considered the dis- protected the repair with an internal brace (Arthrex, Na-
placement to be significant when there was <50% contact ples, FL) using the technique previously described for
between the anterior edge of the medial end of the clavicle and adult first-time anterior SCJ dislocations.5 For patients
the posterior edge of the physis. Institutional review board with an off-ended PI the fracture was reduced and stabi-
approval was obtained for this study and informed consent lized with transosseous sutures passed through the medial
was obtained from all of the patients and their parents. end of the clavicle and transphyseal sutures passed
Inclusion criteria included patients that had sus- through the physis. Postoperatively, for both groups, a CT
tained an acute first-time significantly displaced posterior scan was undertaken and they remained in a sling for 2
SCJ injury without an associated retrosternal vascular weeks. They were reviewed at 1 month and 3 months.
on 04/15/2023
injury diagnosed on CT arteriogram. Exclusion criteria Patients with residual contact between the medial
included patients with evidence of a retrosternal vascular end of the clavicle and the physis were treated non-
injury, patients with a less significant posterior SCJ injury, operatively. They remained in a sling and at 4 weeks un-
patients who had had a previous SCJ injury or fracture derwent a repeat CT scan to assess the position of the
and patients with a history of SCJ instability. medial end of the clavicle for further displacement and any
All the patients were referred from outside our in- evidence of initial healing callus. If the patients were pain
stitution and had a proven posterior displacement of the free and callus was present, they were then allowed to
medial end of the clavicle diagnosed on CT scan. They begin to mobilize out of their sling. They were then re-
had also undergone, at our request, a CT arteriogram to viewed at 3 months and a further CT scan was undertaken
ascertain and rule out a retrosternal vascular injury (in- to assess healing and evidence of remodeling.
timal damage, hematoma or active bleeding). They all Final follow-up was undertaken by a specialist pe-
underwent an ambulatory transfer. diatric practice nurse independent of the treating surgeons.
On admission all the patients were clinically assessed SCJ function was assessed using the short version of the
with specific focus on their SCJ injury and any evidence of Disabilities of the Arm, Shoulder and Hand (Quick-DASH)
on-going mediastinal or vascular compromise. They all score, Rockwood SCJ Score, the modified Constant
then underwent an MRI. The MRI scans were reviewed Shoulder Score, and the Single Assessment Numerical
by a consultant radiologist and the senior author. Evaluation score.6–9 Because of ethical reasons regarding
From the MRI T1 axial view it was possible to clearly exposure to radiation no further imaging was undertaken.
differentiate between a true SCJ dislocation and a medial
end of clavicle PI. For PIs with residual contact the position RESULTS
of the medial end of the clavicle in relation to the physis Over the study period 13 patients with a significant
was assessed. This was expressed as a percentage of the posterior SCJ injury were treated. One patient with a PI injury
TABLE 1. Patient Demographics

Mechanism of Presenting Days to CTA—Vessel MRI physeal
Patient Age Sex Side injury symptoms referral compression contact (Yes/No) Treatment Follow-Up
1 13 M L Soccer None 3 Y Y37% Nonop 84
2 15 M L Judo None 5 Y Y—27% Nonop 75
3 12 M L Rugby None 3 Y Y—33% Nonop 68
4 14 M R Fall None 4 Y Y—15% Nonop 59
5 14 M R Rugby Stridor 6 Y N Surgery 57
6 16 F R Hockey None 11 Y Y—31% Nonop 49
7 15 M L Bicycle None 2 Y Y—31% Nonop 44
8 13 F R Hit by car None 1 Y Y—42% Nonop 38
9 16 M L Toboggan None 6 Y SCJ dislocation Surgery 37
10 17 M R Rugby None 4 Y N Surgery 33
11 15 M R Rugby None 3 Y Y—23% Nonop 30
12 15 M R Fall None 5 Y N Surgery 26
14.6 4.4 29.1% 50
CTA indicates CT arteriogram; F, female; L, left; M, male; MRI, magnetic resonance imaging; N, no; Nonop, nonoperatively; R, right; SCJ, sternoclavicular joint; Y, yes.

Tytherleigh-Strong et al J Pediatr Orthop Volume 43, Number 5, May/June 2023
with residual contact was lost to follow-up. The remaining 12 On the MRI scan T1 axial view of the patients with
patients were included in the study. Details of these patients’ PIs with residual contact the mean amount of residual
demographics are shown in Table 1. There were 2 female and contact between the physis and the medial end of the
10 male patients with a mean age of 14.6 years (12 to 17) at the clavicle was 29.1% (15 to 42) (Fig. 1). The MRI scan T1
time of injury. The mean time to referral was 4.4 days (1 to 11) axial view of the patients with an off-ended PI clearly
and the mean follow-up was 50 months (26 to 84). demonstrated the physis located in the sternal facet and
On arrival at our institution one patient had symptoms the posteriorly and medially dislocated medial clavicle
of mild stridor, which settled within 12 hours. None of the (Fig. 2). The MRI scan T1 axial view of the patient with
patients had any clinical signs suggestive of a significant as- the true SCJ dislocation showed the intact physis on the
sociated mediastinal or intrathoracic injury. Most of the pa- medial end of the clavicle which was dislocated posteriorly
tients were very comfortable with the affected arm in a sling. out of the SCJ (Fig. 3).
The plain CT scans demonstrated a posteriorly and For the 4 patients that underwent surgery there were
medially displaced clavicle in all of the patients. There did no intraoperative complications and specifically no prob-
not appear to be any obvious difference on the CT scans to lems with mobilizing and reducing either the PI or the SCJ
differentiate between patients with a true SCJ dislocation dislocation. The postoperative CT scans showed that the
and a PI. There was some evidence of posterior vessel reduction was maintained and that the injured SCJ was
compression, predominantly the brachiocephalic vein, on symmetrical compared with the uninjured side (Fig. 4).
all of the CT arteriograms. However, there was no evi- For the 8 patients with PIs with residual contact the CT
dence of any associated intimal or vascular injury or of a scans at 1 month all demonstrated no further displacement of
on 04/15/2023
retrosternal hematoma. the medial end of the clavicle and evidence of healing callus
FIGURE 1. Images of a 15-year-old male with a left posteriorly displaced medial end of clavicle physeal injury with residual contact.
A, Axial view computed tomography scan demonstrating posterior displacement of the medial end of the left clavicle. B, Re-
formatted computed tomography arteriogram demonstrating compression of the brachiocephalic artery without evidence of an
intimal injury or associated hematoma (arrow). C, Axial magnetic resonance imaging scan (T1) demonstrating the physis situated
in joint (arrow) with the anterior edge of posteriorly displaced medial end of the clavicle impacted and in contact with the posterior
edge of the physis. D, Enlarged image of axial magnetic resonance imaging scan (T1). The length of the physis is 15 mm and the
length of the overlap between the physis and medial end of the clavicle is 4 mm. The amount of contact is 27%.

on 04/15/2023
FIGURE 2. Images of a 17-year-old male with a right posteriorly displaced medial end of clavicle with an off-ended physeal injury.
A, Axial view computed tomography scan demonstrating posterior displacement of the medial end of the clavicle. B, Reformatted
computed tomography arteriogram demonstrating compression of the brachiocephalic artery without evidence of an intimal
injury or associated hematoma (arrow). C, Axial magnetic resonance imaging (MRI) scan (T1) demonstrating the physis situated in
joint (arrow) with the medial end of the clavicle displaced posteriorly and medially. D, Enlarged image of axial MRI scan (T1). The
length of the physis is 14 mm and there is no contact with the medial end of the clavicle. E, Coronal MRI scan (T1) demonstrating
the right physis situated within the joint (arrow) and the medial end of the clavicle absent. F, Coronal MRI scan (T1) 2 cuts more
posterior to (E) demonstrating the medial end of the clavicle with the physis absent.
(Fig. 5). The CT scans at 3 months demonstrated significant chest CT scan, for a separate clinical problem, at 7 months.
further osseous healing, remodeling at the fracture site and This showed that the physeal fracture had healed and
some resorption of the prominent posterior edge of the remodeled back into alignment and that the clavicular
medial end of the clavicle. One patient underwent a further prominence had completely resorbed.10

on 04/15/2023
FIGURE 3. Images of a 16-year-old male with a right posterior sternoclavicular joint dislocation. A, Axial view computed
tomography scan demonstrating posterior displacement of the medial end of the clavicle. B, Reformatted computed tomography
arteriogram demonstrating compression of the brachiocephalic artery without evidence of an intimal injury or associated hema-
toma (arrow). C, Axial magnetic resonance imaging scan (T1) demonstrating a true posterior sternoclavicular joint dislocation. The
physis is situated normally on the medial end of the clavicle (arrow) and the sternal notch is empty. D, Enlarged image of axial
magnetic resonance imaging scan (T1).
At final follow-up the mean QUICK-Dash score was reduction has been undertaken. As a result, the general
0.4 (0 to 2.3), the mean Rockwood score was 15 (15), the recommended management for these injuries is essentially
mean modified Constant score was 98.8 (89 to 100) and empirical and aimed at reducing the medial end of the
the mean SANE score was 99.6 (95 to 100) (Table 2). clavicle back to its anatomical position.
None of the patients who underwent surgery had any More recently the use of MRI scans to diagnose and
intraoperative or postoperative complications. treat posteriorly displaced PIs has been described in 2 case
reports. Ozer et al4 used an MRI scan to diagnose and
DISCUSSION then treat a displaced medial end clavicle PI with a closed
The main findings of this study are that, following an reduction in a 16-year-old boy. They confirmed the re-
adolescent posterior SCJ injury, an MRI scan is able to clearly duction with a postoperative MRI. An MRI scan was also
differentiate between a true SCJ dislocation and a medial end used to diagnose a left displaced medial end clavicle PI,
clavicle PI and whether there is any residual contact between the where impression of the brachiocephalic vein was noted, in
physis and medial end of the clavicle. Treating patients with a an 18-year-old boy by Wagner et al.3 They treated the
true SCJ dislocation or an off-ended PI with an open reduction patient nonoperatively, in a similar way to the patients in
and fixation and treating patients where there was residual our study with displaced injuries were treated. Serial ser-
contact nonoperatively provided a satisfactory clinical outcome. endipity plain radiographs were used to confirm the po-
Adolescent posterior SCJ injuries have previously sition and the patient returned to contact sports after 8
been diagnosed by CT scan, where it has not been possible weeks. In neither of the studies did the authors attempt to
to differentiate between a true SCJ dislocation and a quantify the extent of the PI displacement.
posteriorly displaced medial end clavicle PI. The only way In our study all of the patients, having a confirmed
the difference has been established is after an open posterior SCJ injury on a CT scan, underwent an MRI

it to be a complication.2,11,12 However, in a recent study

by Tytherleigh-Strong et al13 of 17 adult patients with an
acute posteriorly dislocated SCJ they noted that all of the
patients had vessel compression on the CT arteriogram but
none had evidence of an intimal injury or a hematoma.
They considered the compression, without intimal injury, to

be an expected finding with its clinical relevance only re-

lated to other associated clinical signs. The experience in
our region with all posterior SCJ injuries/dislocations are
that most of these injuries have evidence of vessel com-
pression without evidence of hematoma or intimal injury on
CT arteriogram. WE do not feel that this finding on its own,
would routinely require reduction.
For patients without any evidence of mediastinal
compromise a closed reduction is recommended as the in-
FIGURE 4. Reformatted computed tomography images fol- itial treatment. However, the success of a closed reduction
lowing open reduction and fixation of the patient in Figure 2. has been shown to diminish over time, with a cutoff of
A, Postoperative reformatted computed tomography scan of 48 hours being the usual convention.11,14,15 In a metanalysis
the patient in Figure 2 after open reduction and transosseous – by Tepolt et al1 of 140 patients with adolescent posterior
on 04/15/2023
transphyseal suture fixation. The transosseous drill holes can be SCJ injuries, extracted from 79 studies, 90 (64%) of the
seen in the medial end of the clavicle (arrow). patients were initially treated with a closed reduction. Of
those treated within 48 hours 43 of 77 (58%) were successful
scan. The axial T1 images were used to assess the physeal whilst only 4 of 13 (30.8%) were successful when treated
component of the medial clavicle and could clearly and after 48 hours. In the largest single case series of 48 patients
reliably differentiate between a true SCJ dislocation with adolescent posterior SCJ injuries by Lee et al,2 22
without injury to the physis, a posteriorly displaced PI (46%) of the patients had an attempted closed reduction. Of
where there was residual contact with the metaphyseal these 11 (50%) were initially successful but 3 of these had a
bone and where the metaphyseal bone was completely off- recurrence and required a subsequent open procedure.
ended. For the 3 patients that had completely off-ended The indications for an open reduction as the initial
PIs evidence of residual periosteal sleeve contact was vi- procedure are variable and surgeon dependent. They may
sualized on certain axial cuts. The coronal images were be considered when a closed reduction has been un-
generally less useful for interpreting PIs with residual successful, if the injury is more than 48 hours old or if there
contact but could nicely demonstrate off-ended injuries. is concern with regards to a retrosternal injury.15,16 Surgery
Currently there is a lack of consensus in the literature is undertaken through an incision in line with the medial
as to the optimal treatment modality for an adolescent end of the clavicle and SCJ. It is only at this point that the
posteriorly displaced SCJ injury. This, in part, may be due distinction between a PI and a true dislocation can be made.
to the fact that it has not previously been possible to dif- For a PI the medial end of the clavicle is reduced
ferentiate between a PI, and its extent, and a true SCJ dis- back onto the physis and stabilized with transphyseal su-
location at the time of injury. There is also some confusion tures. For a true dislocation the joint is reduced and the
with regards to the significance of posterior vessel com- anterior capsular tissues repaired. In the metanalysis of
pression on the CT arteriogram. Some authors have con- Tepolt et al1 42 (30%) of the patients underwent a primary
sidered this to always be of significance and even considered open reduction and 47 (33.6%) of patients who had a
failed closed reduction, then underwent an open proce-
dure. In Lee et al2 case series 26 (54%) of the patients
TABLE 2. Patient Clinical Outcomes underwent a primary open repair and 14 (35%) had an
Quick-DASH Rockwood Modified SANE
open procedure after a failed closed reduction. At surgery
Patient (0–100) (0–15) constant (0–100) (0%–100%) 20 patients were found to have had a true SCJ dislocation
and 20 had had a PI.
1 0 15 100 100 Interestingly the functional outcomes following
2 0 15 100 100
3 0 15 100 100 treatment for a posterior SCJ injury have rarely been re-
4 0 15 100 100 ported in the literature. In the metanalysis by Tepolt et al1
5 0 15 100 100 only 61 of the 140 patients had a follow-up of more than
6 0 15 100 100 1 year and only 29 patients from 5 case series/reports were
7 0 15 100 100
8 0 15 100 100
evaluated using functional scores. In this group only 1
9 2.3 15 97 100 patient had been treated by closed reduction and the rest
10 0 15 100 100 by open surgery. The largest case series is the study by
11 0 15 100 100 Waters and colleagues of 12 patients, all treated with open
12 2.3 15 89 95 surgery, who, at a mean follow-up of 23 months (6 to 50)
0.4 15 98.8 99.6
had “perfect” functional outcome scores.

on 04/15/2023
FIGURE 5. Serial computed tomography (CT) images of the patient from Figure 1 with a left posteriorly displaced medial end of
clavicle physeal injury with residual contact treated nonoperatively. A, Axial view CT scan at 4 weeks postinjury. The position of the
fracture has remained the same and there is evidence of healing callus (arrow). B, Reformatted CT scan at 4 weeks. The position of
the fracture has remained the same and there is evidence of healing callus between the physis and the clavicle shaft (arrow). C,
Axial CT scan at 3 months postinjury. The physeal fracture has healed and the displaced medial end of the clavicle is beginning to
remodel. D, Reformatted CT scan at 3 months. The physeal fracture has healed and the displaced medial end of the clavicle has
begun to resorb. E, Reformatted CT scan at 7 months postinjury. The physeal fracture has healed and fully remodeled with near
complete resorption of the medial end of the clavicle.

There is very little in the literature describing the 15 years of age and the fracture was able to remodel ad-
nonoperative management of posterior SCJ injuries. In equately. However, we have no experience of older pa-
general, it is only described in patients that have had a tients where the ability to remodel is diminished.
delayed presentation and usually not considered as a pri- The functional outcome scores at final follow-up for
mary treatment. In the metanalysis by Tepolt et al1 only 2 all patients treated in this series were excellent. This is in
of the 140 patients were treated nonoperatively and, in keeping with previous studies where essentially all of the
both cases, this was due to a delayed presentation of over 4 patients, regardless of the type of posterior SCJ injury,
weeks. However, the patients made a full functional re- were treated with open reduction and fixation.
covery. Carbone et al12 described a case report of a patient In this study the benefit of undertaking an MRI scan
that had sustained a fracture to the medial clavicle physis at the time of the patients’ posterior SCJ injury provided
that had abnormally remodeled but then had sustained a useful diagnostic information and enabled us to differ-
further injury resulting in posterior displacement with entiate between a true SCJ dislocation and a PI. For the
superior vena cava impingement. This was successfully patients with a PI with residual contact, which accounted
treated by surgical reduction and fixation. They postulated for two-thirds of the total, we were able to successfully
that patients presenting 4–6 weeks after a medial clavicle treat their injury nonoperatively, relying on the inherent
PI may have a significant deformity due to osseous over- remodeling abilities of youth. We would recommend the
growth and may require a resection arthroplasty or a more routine additional use of an MRI scan for patients that
complex reconstruction. In a case report by Oykenami sustain a posterior SCJ injury and to initially treat patients
et al17 a 16-year-old male patient who had presented with with a PI with residual contact without surgery.
on 04/15/2023
an acute left clavicle midshaft fracture was noted to have One of the advantages of an MRI scan over a CT
an incidental asymptomatic fibrous nonunion at the me- scan is that it does not require the use of any radiation.
dial end of the clavicle. The patient recounted an injury at Patients in our series underwent as many as 4 CT scans
the medial end of his clavicle 2 years previously from during their treatment, which represents a significant
which he had made a full recovery. However, this diag- amount of radiation over a short period. On the basis of
nosis was based on a plain x-ray and no further imaging in the findings of this study, following a CT scan/arteriogram
the form of a CT or MRI scan was undertaken. and MRI diagnosis of a PI, we have altered our protocol.
In this study, our approach to the management of For patients with residual physeal contact that we do not
adolescent posterior SCJ injuries was significantly different routinely undertake a follow-up CT scan unless there is
and based on the initial MRI scan findings. We treated pa- clinical concern but undertake an MRI scan 6 months
tients with a PI and residual contact nonoperatively and relied after injury. For patients that undergo open reduction we
on the inherent remodeling process present in adolescence to undertake a postoperative CT scan only.
eliminate the bony deformity and restore the medial physis There are a several limitations to this study. This is a
within the periosteal sleeve.18–20 In all of the cases a repeat CT retrospective study and a single institution series and could
scan at 4 weeks showed that the medial end of the clavicle had be susceptible to observer bias. Also, this is a rare injury
not displaced and evidence of plentiful healing callus. A fur- and, as a result, there were a small number of patients
ther scan at 3 months showed further healing and remodeling available for analysis. In addition, 2 of the 4 outcome
of the physis and evidence of resorption of the clavicular measures have not been validated for use in SCJ disorders,
posterior medial prominence. In the patient that had an ad- however, they have been used in other SCJ studies.
ditional CT scan at 7 months the physis and clavicle had
completed remodeled, and the clavicular prominence had CONCLUSION
completely resorbed. As none of the patients had any clinical In this case series of significantly displaced adoles-
evidence of vascular compromise at follow-up and there was cent posterior SCJ injuries MRI scans allowed identi-
either remodeling with no evidence of fracture displacement or fication of true SCJ dislocations and off-ended PIs, which
the fracture had been reduced, we did not undertake repeat were successfully treated by open reduction, and PIs with
CT arteriograms. residual physeal contact which were successfully treated
The patient with a true SCJ dislocation and all of the nonoperatively.
patients with an off-ended PI in our series had presented
more than 48 hours after injury. On the basis of the senior REFERENCES
author’s previous experience and evidence from the literature 1. Tepolt F, Carry PM, Heyn PC, et al. Posterior sternoclavicular joint
regarding the diminishing success of attempting a closed re- injuries in the adolescent population: a meta-analysis. Am J Sports
duction more than 48 hours after injury, they were all treated Med. 2014;42:2517–2524.
by open reduction and fixation. We chose to treat patients 2. Lee JT, Nasreddine AY, Black EM, et al. Posterior sternoclavicular
joint injuries in skeletally immature patients. J Pediatr Orthop.
with an off-ended PI by open reduction as we were concerned 2014;34:369–375.
that, although the fracture might heal, the medial end of the 3. Wagner RJ, Symanski JS, Raasch WG, et al. Successful nonsurgical
clavicle was displaced more medially and posteriorly and management of a posteriorly displaced medial clavicular physeal
might be less stable and may not remodel adequately. fracture in an adolescent athlete: a case report. Clin J Sport Med.
2022;32:e319–e321.
The physis at the medial end of the clavicle does not 4. Ozer UE, Yalcin MB, Kanberoglu K, et al. Retrosternal displace-
begin to ossify until the age of 18 to 20 years. In our series ment of the clavicle after medial physeal fracture in an adolescent:
the oldest patient that was treated nonoperatively was MRI. J Pediatr Orthop B. 2014;23:375–378.

5. Tytherleigh-Strong G, Pecheva M, Titchener A. Treatment of first- 13. Tytherleigh-Strong G, Sabharwal S, Peryt A. Clinical outcomes and
time traumatic anterior dislocation of the sternoclavicular joint with return to sports after open reduction and hamstring tendon autograft
surgical repair of the anterior capsule augmented with internal reconstruction in patients with acute traumatic first-time posterior
bracing. Orthop J Sports Med. 2018;6:2325967118783717. dislocation of the sternoclavicular joint. Am J Sports Med. 2022;50:
6. Constant CR, Murley AH. A clinical method of functional assess- 3635–3642.
ment of the shoulder. Clin Orthop Relat Res. 1987:160–164. 14. Selesnick FH, Jablon M, Frank C, et al. Retrosternal dislocation of the
7. Gummesson C, Ward MM, Atroshi I. The shortened disabilities of clavicle. Report of four cases. J Bone Joint Surg Am. 1984;66:287–291.
the arm, shoulder and hand questionnaire (QuickDASH): validity 15. Laffosse JM, Espie A, Bonnevialle N, et al. Posterior dislocation of
and reliability based on responses within the full-length DASH. the sternoclavicular joint and epiphyseal disruption of the medial
BMC Musculoskelet Disord. 2006;7:44–51. clavicle with posterior displacement in sports participants. J Bone
8. Rockwood CA Jr, Groh GI, Wirth MA, et al. Resection arthroplasty Joint Surg Br. 2010;92:103–109.
of the sternoclavicular joint. J Bone Joint Surg Am. 1997;79:387–393. 16. Leighton RK, Buhr AJ, Sinclair AM. Posterior sternoclavicular
9. Williams GN, Gangel TJ, Arciero RA, et al. Comparison of the dislocations. Can J Surg. 1986;29:104–106.
single assessment numeric evaluation method and two shoulder 17. Oyekanmii CHG, Nwachukwu I. Asymptomatic fibrous non-union
rating scales. Outcomes measures after shoulder surgery. Am J Sports of displaced sternoclavicular epiphyseal fracture: ase report and
Med. 1999;27:214–221. literature review. Injury Extra. 2005;36:297–299.
10. Virk MS, Apostolakos J, Cote MP, et al. Operative and nonoperative 18. Sewell MD, Al-Hadithy N, Le Leu A, et al. Instability of the
treatment of acromioclavicular dislocation: a critical analysis review. sternoclavicular joint: current concepts in classification, treatment
JBJS Rev. 2015;3:e5. and outcomes. Bone Joint J. 2013;95-B:721–731.
11. Tepolt F, Carry PM, Taylor M, et al. Posterior sternoclavicular joint 19. Tennent TD, Pearse EO, Eastwood DM. A new technique for
injuries in skeletally immature patients. Orthopedics. 2014;37: stabilizing adolescent posteriorly displaced physeal medial clavicular
e174–e181. fractures. J Shoulder Elbow Surg. 2012;21:1734–1739.
12. Carbone P, Rose M, O’Daniel JA, et al. Posterior remodeling of 20. Zaslav KR, Ray S, Neer CS 2nd. Conservative management of a
on 04/15/2023
medial clavicle causing superior vena cava impingement. Am J displaced medial clavicular physeal injury in an adolescent athlete. A
Orthop (Belle Mead NJ). 2011;40:297–300. case report and literature review. Am J Sports Med. 1989;17:833–836.

ORIGINAL ARTICLE
Novel Machine Vision Image Guidance System

Significantly Reduces Procedural Time and Radiation
Exposure Compared With 2-dimensional Fluoroscopy-

based Guidance in Pediatric Deformity Surgery
Christopher P. Comstock, MD and Eric Wait, MD
(3.4 ± 3.2 vs 9.9 ± 10.5 mGy, P < 0.001) respectively. The

Background: Intraoperative 2-dimensional (2D) fluoroscopy length of stay displayed a decreasing trend with MVIGS, and the
imaging has been commonly adopted for guidance during com- operative time was significantly reduced in MvIGS compared
plex pediatric spinal deformity correction. Despite the benefits, with 2D fluoroscopy for an average of 63.6 minutes (294.5 ±
fluoroscopy imaging emits harmful ionizing radiation, which has 15.5 vs 358.1 ± 60.6 min, P < 0.001).
on 04/15/2023
been well-established to have deleterious effects on the surgeon Conclusion: In pediatric spinal deformity correction surgery,
and operating room staff. This study investigated the difference MvIGS was able to significantly reduce intraoperative fluoro-
in intraoperative fluoroscopy time and radiation exposure during scopy time, intraoperative radiation exposure, and total surgical
pediatric spine surgery between 2D fluoroscopy-based navigation time, compared with traditional fluoroscopy methods. MvIGS
and a novel machine vision navigation system [machine vision reduced the operative time by 63.6 minutes and reduced intra-
image guidance system (MvIGS)]. operative radiation exposure by 66%, which may play an im-
Methods: This retrospective chart review was conducted at a portant role in reducing the risks to the surgeon and operating
pediatric hospital with patients who underwent posterior spinal room staff associated with radiation in spinal surgery procedures.
fusion for spinal deformity correction from 2018 to 2021. Patient Level of Evidence: Level III; retrospective comparative study.
allocation to the navigation modality was determined by the date
of their surgery and the date of implementation of the MvIGS. Key Words: 2D fluoroscopy, image guidance, machine vision,
Both modalities were the standard of care. Intraoperative radi- deformity correction, scoliosis, pediatric
ation exposure was collected from the fluoroscopy system re- (J Pediatr Orthop 2023;43:e331–e336)
ports.
Results: A total of 1442 pedicle screws were placed in 77 children:
714 using MvIGS and 728 using 2D fluoroscopy. There were no
significant differences in the male-to-female ratio, age range,
body mass index, distribution of spinal pathologies, number of
levels operated on, types of levels operated on, and the number of
P osterior segmental spinal instrumentation and fusion
are commonly performed for scoliosis deformity
correction.1,2 There are inherent risks while implanting
pedicle screws implanted. Total intraoperative fluoroscopy time pedicle screws due to the adjacent critical neurovascular
was significantly reduced in cases utilizing MvIGS (18.6 ± 6.3 s) anatomy. This has led to continuous advancements in
compared with 2D fluoroscopy (58.5 ± 19.0 s) (P < 0.001). This intraoperative technology, which have been heavily relied
represents a relative reduction of 68%. Intraoperative radiation on for guidance in complex cases3Although 2-dimensional
dose area product and cumulative air kerma were reduced by (2D) fluoroscopy has historically been the most commonly
66% (0.69 ± 0.62 vs 2.0 ± 2.1 Gycm2, P < 0.001) and 66% adopted intraoperative imaging modality, improvements
from 2D plain radiographs to 3-dimension (3D) real-time
imaging have given surgeons access to multiple planes of
From the Driscoll Children’s Hospital, Corpus Christi, TX.
No funding was received to assist with the preparation of this manuscript.
vision. Despite the advancements in imaging and navi-
The authors declare no conflicts of interest. gation, intraoperative radiation exposure remains a con-
Reprints: Christopher P. Comstock, MD, Department of Pediatric Ortho- cern to surgeons and operating room (OR) staff.
paedics, Driscoll Children’s Hospital, 3533 South Alameda, P.O. Box The effects of radiation are especially concerning to
6530, Corpus Christi, TX 78466. E-mail: christopher.comstock@dchstx.
org.
the surgeon and staff who are frequently exposed to
Copyright © 2023 The Author(s). Published by Wolters Kluwer Health, multiple doses of radiation. This accumulation is also
Inc. This is an open access article distributed under the terms of the exacerbated as the surgeon must remain in the surgical
Creative Commons Attribution-Non Commercial-No Derivatives field, directly adjacent to the image intensifier. Spine sur-
License 4.0 (CCBY-NC-ND), where it is permissible to download and geons are exposed to 4 to 12 times higher doses of radia-
share the work provided it is properly cited. The work cannot be
changed in any way or used commercially without permission from tion compared with other orthopaedic specialties.4,5
the journal. Numerous studies quantify the amount of occupatio-
DOI: 10.1097/BPO.0000000000002377 nal radiation exposure to health care professionals and

Comstock and Wait J Pediatr Orthop Volume 43, Number 5, May/June 2023
provide evidence that occupational exposure is well above Institutional Review Board approval was obtained
the recommended values for annual allowable occupa- per hospital guidelines to access data from patient’s
tional radiation exposure.5–9 With increased intra- medical and operative records and discharge summaries,
operative radiation exposure, surgeons are potentially which included: patient demographics, number of verte-
subjected to deterministic effects (such as hair loss, skin brae treated and location (thoracic, lumbar, and pelvis),
erythema, skin burns, and cataract formation)10 and sto- number of pedicle screws implanted, operative time (de-
chastic effects (carcinogenesis and teratogenesis).11–13 In a fined as the time from the first incision to closure), length
study of female orthopaedic, urology, and plastic surgeons of stay, and intraoperative radiation exposure. All intra-
who were exposed to fluoroscopy, it was found that or- operative and postoperative complications were recorded.
thopaedic surgeons had twice the expected rate of total
cancers and 2.9 times the rate of expected breast cancer.14 Surgical Technique
Thus, there is a need to develop intraoperative systems A standard posterior midline subperiosteal spine
that limit radiation to ensure the safety of the surgeon. exposure was utilized for both imaging modalities. In the
A new technology in spinal navigation that utilizes 2D fluoroscopy group, the entry point was defined by
machine vision has been developed to reduce intra- determining the anatomic landmarks such as the trans-
operative radiation to surgeons and OR staff. This system verse process, and lateral facet border after facetectomy at
uses advanced optics combining a light projector with 2 each level. For thoracic levels, fluoroscopy was used to
stereoscopic video cameras to create a 3D map of the confirm the entry point. The majority of lumbar pedicle
patient’s anatomy and correlates this information with screws were implanted using the freehand technique and
preoperative computerized tomography (CT).15 This sys- without the aid of fluoroscopy. If pedicle probing was
on 04/15/2023
tem has previously been studied in the adult population difficult due to dense pedicles or abnormal anatomy, flu-
undergoing posterior spinal fusion in up to 4 levels and oroscopy would be used to confirm trajectory. After de-
was found to reduce intraoperative radiation compared termining the cancellous part within the pedicle by using a
with both 2D-fluoroscopy-based navigation (2D fluoro- pedicle sounder, sequentially, tapping and screw place-
scopy) and a different 3D navigation system.16,17 The ment were performed. Anteroposterior and lateral fluo-
present study is the first study to compare intraoperative roscopic images were obtained to check whether the
radiation emissions between machine vision navigation to screws were positioned properly. All screws were stimu-
2D-fluoroscopy in a surgeon population performing spinal lated with a threshold of up to 12 mA for acceptance of
deformity correction surgery in a pediatric hospital. The the screw.
objectives were to compare intraoperative radiation ex- In the machine vision navigation group, a low-dose
posure and procedural operative time. preoperative CT scan was obtained and uploaded to the
system for each patient. Preoperative CT scans were used
for registration of the vertebral anatomy. Facetectomy was
METHODS done on each level before registration. The surgeon per-
This was a single-center retrospective study com- formed segmental registration (registering to a single verte-
paring surgical parameters of pediatric ( ≤ 18 y of age) bra) for regions of the spine that were flexible. The reference
posterior spinal deformity correction surgery with the use frame was clamped to the spinous process of the nearest
of conventional 2D-fluoroscopy-based navigation, OEC vertebra to place the pedicle screws. By using a navigated
Elite (GE Healthcare) to FLASH navigation (SeaSpine, pedicle probe, the ideal entry point and trajectory of pedicle
CA). All procedures were performed by a single surgeon screws were defined. Preparing the entry point was done
with over 20 years of experience in orthopaedic surgery. with a navigated awl and probe. The pedicle screws were
Surgeries were performed between 2018 and 2021 implanted with the guidance of a virtual trajectory feature
and were conducted using either FLASH navigation fea- (Fig. 1A), allowing the surgeon to follow the trajectory
turing machine vision or conventional 2D fluoroscopy planned using a navigated pedicle probe. The Reslicer
assistance. Patients were assessed, and treatment was de- (Fig. 1B) was also used during navigation for screw
termined by their treating physician as part of their implantation in spinal levels with severe rotation, which
standard of care. Analyzed surgeries were an open poste- allows the surgeon to align the axis marker with the
rior fusion of a long construct (6+ vertebras) during a transverse process of the level and create an accurate axial
spinal deformity correction surgery. Other procedures and sagittal view. Anteroposterior and lateral fluoroscopic
(such as transforaminal lumbar interbody fusion) or in- images were obtained to check whether the screws were
complete data sets were excluded. Navigation modality positioned properly, and all screws were stimulated similarly
allocation was determined by the date of the surgery and to the 2D fluoroscopy group.
the date of implementation of FLASH navigation at the
participating hospital (Driscoll’s Children’s Hospital, TX). Radiation Exposure
All procedures performed before September 2020 were For each case, intraoperative radiation exposure was
conducted using 2D fluoroscopy, which was the standard collected from the radiology department, which was pro-
of care at that time. All procedures performed after this vided on reports from the fluoroscopy system for the du-
date were performed with FLASH navigation, which be- ration of the procedure (incision to close). Intraoperative
came the new standard of care. radiation was evaluated by collecting the fluoroscopy
e332 | www.pedorthopaedics.com Copyright © 2023 The Author(s). Published by Wolters Kluwer Health, Inc.
J Pediatr Orthop Volume 43, Number 5, May/June 2023 MvIGS vs 2D Fluoroscopy
on 04/15/2023
FIGURE 1. A, Live video of the saved augmented reality trajectory, as shown through FLASH navigation. B, The Reslicer feature was
used throughout the surgical procedure. It was used to navigate L2 as the axial image resembles an irregular sagittal image (left
image) due to severe spinal rotation. The Reslicer is implemented (right image) correcting for the misaligned plane and creating an
axial image.
system output in dose area product (DAP) in units of gray This data set is not available as it includes patient-identi-
centimeters squared (Gycm2) and as cumulative air kerma fiable information.
(CAK) in units of milligray (mGy). DAP is a measure of
the total energy delivered throughout the entire radiation
field and is the product of CAK and the cross-sectional RESULTS
area of the radiation field. CAK is a measure of the energy A total of 77 children were enrolled in this study: 40
delivered per unit mass of air to a reference point in the in the 2D fluoroscopy group and 37 in the machine vision
radiation field and is also considered as the scatter radia- navigation group. A total of 1442 pedicle screws were
tion as radiation from the source that is deflected off of a placed: 714 using machine vision navigation and 728 using
surface (such as the patient in an operative setting).18 2D fluoroscopy. There were no statistically significant
Scatter radiation exposure is the primary form of exposure differences in the sex, body mass index, or spinal pathol-
to operative staff who stand further away from the sur- ogies distributions between groups (P = 0.81, P = 0.81,
gical table.19 and P = 0.48, respectively). The mean age was also
similar (P = 0.21). No patients received osteotomies or
vertebrectomies. All demographic information is pre-
Statistical Methods sented in Table 1.
Means and SDs are reported for continuous varia- The overall operative time was significantly reduced
bles; counts and proportions are reported for discrete with machine vision navigation compared with 2D fluo-
variables. For continuous variables following approx- roscopy (294.5 ± 15.5 vs 358.1 ± 60.6 min, respectively,
imately normal distributions, the 2 groups were compared P < 0.001). This is an average reduction of 63.6 minutes or
using a 2-tailed t test. For discrete variables, a Fisher exact 1.06 hours. The average hospital length of stay trended
test was used to compare the 2 groups as appropriate. lower with the machine vision navigation (3.4 ± 0.70 vs
A P value of <0.05 was considered statistically significant. 3.8 ± 0.97 d, P = 0.05). Other characteristics of the
Copyright © 2023 The Author(s). Published by Wolters Kluwer Health, Inc. www.pedorthopaedics.com | e333
its advantages, fluoroscopy-based imaging emits harmful

TABLE 1. Patient Demographics and Distribution of
Pathologies Underlying Spinal Deformity ionizing radiation, to which the risks of iatrogenic radia-
tion exposure have been well-established.20–24 Spine sur-
Navigation 2D
geons operating on complex anatomy are particularly
featuring Fluoroscopy-based
machine vision navigation susceptible to these effects because of repeated doses
(N = 37); (N = 40); throughout a procedure and throughout their career re-

n (%) n (%) P sulting in a significantly higher incidence and mortality

from cancer compared with that of the general
Sex
Female 27 (73) 28 (70) 0.81 population.13,14 Intraoperative imaging and navigation
Age (y) have remained a primary source of radiation for spine
Mean 13.8 13.2 0.21 surgeons and OR staff and warrant a reduced radiation
Range 9-18 10-17 — navigation alternative.
BMI
Underweight 1 (3) 3 (8) 0.81
Based on this retrospective study involving 77 chil-
Healthy 19 (51) 22 (55) — dren and 1441 pedicle screws, machine vision navigation
Overweight 6 (16) 5 (13) — was able to provide 3D intraoperative navigation with
Obese 11 (30) 10 (25) — reduced intraoperative radiation and operative time. Both
Pathologies modality groups were well matched with no statistically
Idiopathic scoliosis 28 (76) 25 (63) 0.48
Neuromuscular 7 (19) 12 (33) — significant difference in age, sex, body mass index, primary
scoliosis pathology, number of operative levels, or number of
on 04/15/2023
Congenital scoliosis 2 (5) 2 (5) — pedicle screws implanted. Our study found that the in-
BMI indicates body mass index; 2D, 2-dimension.
traoperative radiation during spinal deformity correction
surgery with machine vision navigation compared with 2D
fluoroscopy was exposed to significantly less radiation
deformity correction were similar between groups (such as (DAP and CAK), less total fluoroscopy time, reduced
the number of levels operated on, types of levels operated operative time, and a trend of reduced length of stay. This
on, and the number of pedicle screws implanted). Char- represents intraoperative radiation reductions of 68% and
acteristics of the procedure are presented in Table 2. No 66%, respectively, and an average time savings of
intraoperative complications were reported. Two screws in 63.6 minutes. These findings show the substantial advan-
the 2D fluoroscopy group were repositioned and 1 screw tages and the potential for intraoperative reduced radia-
was abandoned and replaced with a pedicle hook based on tion with the implementation of FLASH navigation.
the neuromonitoring data. No neurological complications Intraoperative radiation exposure of both modalities
were reported postoperatively. used in our study has been previously studied. One study
All intraoperative radiation parameters were sig- found that 2D fluoroscopy time ranged from 46 to 69
nificantly reduced with the machine vision navigation seconds, the DAP ranged from 1.5 versus 2.5 Gycm2 and
group compared with 2D fluoroscopy. The total intra- the CAK ranged from 5.1 to 8.8 mGy.25 This range was
operative exposure time was significantly reduced with dependent on the use of a dedicated spine radiology
machine vision navigation compared with 2D fluoroscopy technologist. These values are comparative and validate
(18.6 ± 6.3 vs 58.5 ± 19.0 s, P ≤ 0.001). This represents a our findings in the 2D fluoroscopy group (fluoroscopy
relative reduction of 68%. The total intraoperative DAP time; 58.5 s; DAP, 2.0 Gycm2; CAK, 9.9 mGy). A study
and CAK were also significantly reduced (0.69 ± 0.62 vs looking at machine vision navigation in adults undergoing
2.0 ± 2.1 Gycm2, P < 0.00001 and 3.4 ± 3.2 vs posterior spinal fusion of 1 to 4 levels, found that the
9.9 ± 10.5 mGy, P < 0.001, respectively). This represents a fluoroscopy time was 4.51 seconds and the DAP was
relative reduction of 66% in DAP and CAK. Intra- 0.80 Gycm2.17 This represents less fluoroscopy time com-
operative radiation parameters are presented in Table 3. pared with our study (18.6 s), however, this is likely due to
fewer levels fused and less complicated anatomy. The
DAP from this previous study is comparable to our find-
DISCUSSION ings in the machine vision navigation group (0.69 Gycm2)
This is the first study to compare 2D fluoroscopy- and is likely because the only potential source of intra-
based navigation to a novel machine vision navigation operative radiation when using machine vision navigation
system [machine vision image guidance system (MvIGS)] is confirmatory fluoroscopic images. Although this is the
during corrective surgery for a pediatric spinal deformity first study to demonstrate that machine vision navigation
on the intraoperative radiation of the surgeon and OR reduced radiation exposure in pediatric spinal deformity
staff. These results indicate that the utilization of FLASH corrections, our findings are comparable to the existing
navigation featuring machine vision resulted in a sig- literature in the adult population.
nificant intraoperative reduction in fluoroscopy time, ra- Furthermore, there was a significant reduction in the
diation exposure, and total surgical time. total operative time in the machine vision navigation
Historically, 2D fluoroscopy has been a widely uti- group compared with the 2D fluoroscopy group. The total
lized method for the guidance of placement of pedicle operative time was reduced from 358.4 to 294.5 minutes
screws, especially in spines with variable anatomy. Despite or an average reduction of 63.6 minutes. The decreased
J Pediatr Orthop Volume 43, Number 5, May/June 2023 MvIGS vs 2D Fluoroscopy
TABLE 2. Characteristics of Spinal Correction Procedures the accuracy, no patients had symptomatic indications of
complications relating to pedicle screw breaches, such
Navigation featuring 2D Fluoroscopy-
machine vision based navigation
as neurological deficits. Another limitation of the present
(N = 37); n (%) (N = 40); n (%) P study is the impact of coronavirus disease 2019
(COVID-19). COVID-19 has led to numerous cancella-
Pedicle screws 19.3 ( ± 4.1) 18.2 ( ± 4.4) 0.26 tions and rescheduled surgeries. Studies suggest that a wait
No. levels 13.0 ( ± 2.1) 13.0 ( ± 2.5) 0.64

Levels included in the procedure time of 6 months puts patients with idiopathic scoliosis at
Thoracic 37 (100) 40 (100) 1.0 risk of deformity progression and delaying intervention is
Lumbar 36 (97) 39 (98) 1.0 associated with prolonged operative times and un-
Pelvis 6 (16) 10 (25) 0.40 favorable outcomes.28,29 Factors related to COVID-19
Operative 294.5 ( ± 15.5) 358.4 ( ± 60.6) < 0.001
time (min)
were not collected, however, the demographic character-
Length of 3.4 ( ± 0.70) 3.8 ( ± 0.97) 0.05 istics suggest that there was no difference between the
stay (d) populations despite machine vision navigation being im-
2D indicates 2-dimension.
plemented during COVID-19 in September 2020. There
are slightly more neuromuscular patients in the 2D fluo-
roscopy cohort, though not significant, that may indirectly
surgical time in the machine vision navigation group may influence the studied outcomes. In addition, 1 author has
have been due to increased confidence in the ability to an active consultation agreement with SeaSpine at the
visualize and correct the angle of severely rotated spines time of this study, which may have contributed to bias.
However, the authors received no financial support for the
on 04/15/2023
leading to a true axial view. It was also due to more

confidence in identifying appropriate entry points, and research, authorship, or publication of this article.
trajectories and understanding the anatomy of dense The scope of this study focused on scatter intra-
pedicles. Shorter operative times can deliver advantages operative radiation that affects the surgeon and OR staff and
including decreased health care costs, decreased potential thus radiation to the patient preoperatively or postoperative
for surgical and anesthesia complications, and decreased has not been directly captured. Scatter radiation is known to
length of stay.26,27 It is suspected that the trend of de- be affected by anatomy and the distance of the surgeon and
creased length of stay in the MvIGS cohort in this study, staff to the radiation; thus the actual exposure may vary.
was related to the significant reduction in total operative Radiation to the patient from the preoperative CT is out of
time. However, further studies should include patient the scope of this study, however, the authors recognize that
safety data and other factors that may be correlated to the further investigation is required for patient safety. Pre-
length of stay. Although limitations exist in quantifying operative CT has the potential to increase the radiation
operative time and may include confounding factors such exposure to the patient depending on the protocol used,30
as the effects of rotating staff, machine vision navigation however, additional radiation exposure to the patient must
had a significant impact on operative time. be weighed with the potential safety advantages provided by
Several limitations exist in the present study. This the additional 3D data of navigation systems. Low-dose CT
study was limited by its retrospective nature; thus, it is scan protocols continue to be experimented with and are
susceptible to selection bias and group heterogeneity. In- crucial in future navigation and radiation studies.
herent to studies spanning multiyear, differences in prac- The cost of MvIGS is $500,000 with the addition of
tice patterns and surgical staff may have resulted in disposable passive spheres to navigate instruments (< $200
confounding factors, which are not accounted for in the per case). This is compared with other navigation systems
present study, such as different expertise levels, evolving that can cost anywhere from $250,000 to $1,500,000.31
techniques, etc. However, the placement of pedicle screw The 2D fluoroscopic system used costs ~$195,000 and does
instrumentation remained reasonably standardized. Pedi- not require preoperative CT imaging. A carbon fiber table
cle screw placement accuracy was not captured due to or module is required: upgrades to an existing surgical
ethical reasons. To assess pedicle screw accuracy, the pa- table may cost $30,000 to 60,000 whereas a full carbon
tient would be required to receive additional radiation fiber table costs ~$100,000 to $200,000. The 2D fluoro-
from a postoperative CT scan, which is not the current scopy system also requires a radiology technician, which
standard of care at the institution. Despite not assessing has a typical salary range between $56,867 and $66,220.
TABLE 3. Radiation Exposure During Spinal Correction Procedures

Navigation featuring machine vision (N = 37) 2D Fluoroscopy-based navigation (N = 40) P
Events 33.6 ( ± 11.6) 104.4 ( ± 32.8) < 0.001
Pedal time (s) 21.6 ( ± 7.8) 71.7 ( ± 25.1) < 0.001
Total exposure time (s) 18.6 ( ± 6.3) 58.5 ( ± 19.0) < 0.001
Total DAP (Gycm2) 0.69 ( ± 0.62) 2.0 ( ± 2.1) < 0.001
CAK (mGy) 3.4 ( ± 3.2) 9.9 ( ± 10.5) < 0.001
CAK indicates cumulative air kerma; 2D, 2-dimension; DAP, dose area product.
Copyright © 2023 The Author(s). Published by Wolters Kluwer Health, Inc. www.pedorthopaedics.com | e335
In this study, the implementation of FLASH 15. Kalfas IH. Machine vision navigation in spine surgery. Front Surg.
navigation yielded considerable benefits for the surgeon 2021;8:41.
16. Dorilio J, Utah N, Dowe C, et al. Comparing the efficacy of
performing pediatric spinal deformity corrections. Com- radiation free machine-vision image-guided surgery with traditional
pared with 2D fluoroscopy-based imaging, total intra- 2-dimensional fluoroscopy: a randomized, single-center study. HSS
operative fluoroscopy time and intraoperative radiation J. 2021;17:274–280.
exposure were all significantly reduced by a relative 17. Malham GM, Munday NR. Comparison of novel machine vision
reduction of 68% and 66%, respectively. Operative time was spinal image guidance system with existing 3D fluoroscopy-based
navigation system: a randomized prospective study. Spine J. 2022;22:
also significantly reduced by an average of 63.6 minutes. 561–569.
This novel FLASH navigation featuring a machine vision 18. Frush D, Huda W. Categorical Course in Diagnostic Radiology
system has reduced intraoperative radiation exposure, thus Physics: From Invisible to Visible –The Science and Practice of X-Ray
potentially decreasing the proven harmful effects associated Imaging and Radiation Dose Optimization. Radiological Society of
with ionizing radiation toward the surgeon and OR staff. North America 2006 Scientific Assembly and Annual Meeting,
November 26 - December 1, 2006. Accessed January 1, 2023. http://
archive.rsna.org/2006/4425855.html
REFERENCES 19. Narain AS, Hijji FY, Yom KH, et al. Radiation exposure and
1. Weinstein SL, Dolan LA, Cheng JC, et al. Adolescent idiopathic reduction in the operating room: Perspectives and future directions in
scoliosis. Lancet. 2008;371:1527–1537. spine surgery. World J Orthop. 2017;8:524–530.
2. Hicks JM, Singla A, Shen FH, et al. Complications of pedicle screw 20. National Research Council. Health Risks from Exposure to Low
fixation in scoliosis surgery: a systematic review. Spine. 2010;35: Levels of Ionizing Radiation: BEIR VII Phase 2. Washington, D.C:
E465–E470. The National Academies Press; 2006.
3. Rawicki N, Dowdell JE, Sandhu HS. Current state of navigation in 21. Kleinerman RA. Cancer risks following diagnostic and therapeutic
spine surgery. Ann Transl Med. 2021;9:85–85.
on 04/15/2023
radiation exposure in children. Pediatr Radiol. 2006;36(suppl 2):

4. Theocharopoulos N, Perisinakis K, Damilakis J, et al. Occupational 121–125.
exposure from common fluoroscopic projections used in orthopaedic 22. Ronckers CM, Land CE, Miller JS, et al. Cancer mortality among
surgery. J Bone Joint Surg Am. 2003;85:1698–1703. women frequently exposed to radiographic examinations for spinal
5. Rampersaud YR, Foley KT, Shen AC, et al. Radiation exposure to disorders. Radiat Res. 2010;174:83–90.
the spine surgeon during fluoroscopically assisted pedicle screw 23. Morin Doody M, Lonstein JE, Stovall M, et al. Breast cancer
insertion. Spine. 2000;25:2637–2645. mortality after diagnostic radiography: findings from the U.S
6. ICRP. 1990 Recommendations of the International Commission on scoliosis cohort study. Spine. 2000;25:2052–2063.
Radiological Protection. Ann ICRP. 1991;21:1–201. 24. Levy AR, Goldberg MS, Mayo NE, et al. Reducing the lifetime risk
7. Goldstone KE, Wright IH, Cohen B. Radiation exposure to the of cancer from spinal radiographs among people with adolescent
hands of orthopaedic surgeons during procedures under fluoroscopic idiopathic scoliosis. Spine. 1996;21:1540–1547.
x-ray control. Br J Radiol. 1993;66:899–901. 25. Siddiqui AA, Andras LM, Obana KK, et al. Using a dedicated spine
8. Mehlman CT, DiPasquale TG. Radiation exposure to the orthopae- radiology technologist is associated with reduced fluoroscopy time,
dic surgical team during fluoroscopy: “how far away is far enough?” radiation dose, and surgical time in pediatric spinal deformity
J Orthop Trauma. 1997;11:392–398. surgery. Spine Deform. 2021;9:85–89.
9. Oddy MJ, Aldam CH. Ionising radiation exposure to orthopaedic 26. Carreon LY, Puno RM, Lenke LG, et al. Non-neurologic
trainees: the effect of sub-specialty training. Ann R Coll Surg Engl. complications following surgery for adolescent idiopathic scoliosis.
2006;88:297–301. J Bone Joint Surg Am. 2007;89:2427–2432.
10. Fujibayashi S, Murata K, Shimizu T, et al. An observational 27. Diefenbach C, Ialenti MN, Lonner BS, et al. Hospital cost analysis
survey of nail and skin of spine surgeons-possible damage by of neuromuscular scoliosis surgery. Bull Hosp Jt Dis (2013). 2013;71:
occupational ionizing radiation exposure. Spine Surg Relat Res. 272–277.
2021;5:359–64. 28. Miyanji F, Newton PO, Samdani AF, et al. Impact of surgical
11. Mendelsohn D, Strelzow J, Dea N, et al. Patient and surgeon radiation waiting-list times on scoliosis surgery: the surgeon’s perspective.
exposure during spinal instrumentation using intraoperative computed Spine. 2015;40:823–828.
tomography-based navigation. Spine J. 2016;16:343–354. 29. Anari JB, Baldwin KD, Flynn JM, et al. What’s important:
12. Perisinakis K, Theocharopoulos N, Damilakis J, et al. Estimation of managing the impact of coronavirus on pediatric spine surgery.
patient dose and associated radiogenic risks from fluoroscopically J Bone Joint Surg Am. 2020;102:e94.
guided pedicle screw insertion. Spine. 2004;29:1555–1560. 30. Larson AN, Schueler BA, Dubousset J. Radiation in spine
13. Mastrangelo G, Fedeli U, Fadda E, et al. Increased cancer risk deformity: state-of-the-art reviews. Spine Deform. 2019;7:
among surgeons in an orthopaedic hospital. Occup Med (Lond). 386–394.
2005;55:498–500. 31. Malham GM, Wells-Quinn T. What should my hospital buy next?-
14. Chou LB, Lerner LB, Harris AHS, et al. Cancer prevalence among a Guidelines for the acquisition and application of imaging,
cross-sectional survey of female orthopedic, urology, and plastic navigation, and robotics for spine surgery. J Spine Surg. 2019;5:
surgeons in the United States. Womens Health Issues. 2015;25:476–481. 155–165.
ORIGINAL ARTICLE
Operative Fractures of the Phalangeal Head and Neck in

Children—Does Open Reduction Affect Outcomes?
Aaron S. Vaslow, MD,* James Banks Deal Jr, MD,† and Christine A. Ho, MD‡§∥
Conclusions: Open phalangeal head and neck fractures have

Background: Closed reduction percutaneous pinning of displaced more concomitant digital injuries and postoperative complica-
pediatric phalangeal head and neck fractures is preferred to tions compared with injuries closed on presentation, regardless of
prevent malunion and loss of motion and function. However, whether the fracture underwent open or closed reduction. Al-
open reduction is required for irreducible fractures and open though osteonecrosis occurred in all 3 cohorts, it was most fre-
injuries. We hypothesize that osteonecrosis is more common in quent in open injuries. This study allows surgeons to discuss rates
open injuries than closed injuries that require either open re- of osteonecrosis and resultant complications with families whose
duction or closed reduction percutaneous pinning. child presents with phalangeal head and neck fractures that are
Methods: Retrospective chart review of 165 phalangeal head and
on 04/15/2023
indicated for surgical treatment.

neck fractures treated surgically with pin fixation at a single Level of Evidence: Therapeutic, Level III.
tertiary pediatric trauma center from 2007 to 2017. Fractures
were stratified as open injuries (OI), closed injuries undergoing Key Words: phalangeal neck, fracture, children, necrosis,
open reduction (COR), or closed injuries treated with closed phalangeal head, avascular necrosis
reduction (CCR). The groups were compared using Pearson χ2 (J Pediatr Orthop 2023;43:311–316)
tests and ANOVA. Two group comparisons were made with
Student t test.
Results: There were 17 OI fractures, 14 COR fractures, and 136
CCR fractures. Crush injury was the predominant mechanism in
OI versus COR and CCR groups. The average time from injury
to surgery was 1.6 days for OI, 20.4 days for COR, and 10.4 days
H and fractures are common in the pediatric pop-
ulation, comprising an estimated 15% of fractures
seen in the emergency department, and of these fractures,
for CCR. The average follow-up was 86.5 days (range, 0 to most occur in the phalanges.1–5 10 percent to 15% of pe-
1204). The osteonecrosis rate differed between the OI versus diatric phalangeal fractures are phalangeal head or neck
COR and OI versus CCR groups (71% for OI, 7.1% for COR, sub-types.2,5–9
and 1.5% for CCR). Rates of coronal malangulation > 15 de- Pediatric phalangeal neck fractures are described as
grees differed between the OI and COR or CCR groups, but the inherently unstable with poor remodeling potential be-
2 closed groups did not differ. Outcomes were defined using Al- cause they occur at the distal end of the phalanges, away
Qattan’s system; CCR had the most excellent and fewest poor from the physis.10,11 Al-Qattan originally described and
outcomes. One OI patient underwent partial finger amputation. developed a classification and treatment algorithm for
One CCR patient had rotational malunion but declined derota- phalangeal neck fractures, with Type I being non-
tional osteotomy. displaced, Type II displaced with bony contact, and Type
III displaced without any bony contact (Fig. 1). He
recommended pin fixation for unstable, displaced Type II
From the *T Brian D. Allgood Army Community Hospital, Camp and all Type III fractures, with a strong preference for
Humpreys, Pyeongtaek, South Korea; †Department of Orthopedic open reduction.10 Most displaced phalangeal head and
Surgery, Walter Reed National Military Medical Center, Bethesda, neck fractures in children are treated with closed reduction
MD; ‡Department of Orthopaedic Surgery, Children’s Health Dallas; and percutaneous pinning (CRPP), but if healing and
§Scottish Rite for Children; and ∥Department of Orthopaedic Sur-
gery, University of Texas Southwestern, Dallas, TX.
early callus render closed reduction impossible, these
The authors declare no conflicts of interest. fractures require anatomic reduction with either open or
Reprints: Christine A. Ho, MD, Scottish Rite for Children; Children’s percutaneous osteoclasis and percutaneous pinning.12,13
Medical Center Dallas; University of Texas Southwestern Medical Additional indications for open reduction include failure
School, 1935 Medical Center Dr., E2300-E2.01, Dallas, TX 75235.
E-mail: christine.ho@childrens.com.
of conservative treatment, unsuccessful closed reduction,
Copyright © 2023 The Author(s). Published by Wolters Kluwer Health, and delayed presentation with impending malunion.2,8
Inc. This is an open access article distributed under the terms of the Phalangeal head and neck fractures occur in a vas-
Creative Commons Attribution-Non Commercial-No Derivatives cular watershed area of the bone, and disruptions of the
License 4.0 (CCBY-NC-ND), where it is permissible to download and blood supply are a concern for both closed versus open
share the work provided it is properly cited. The work cannot be
changed in any way or used commercially without permission from pinning procedures. Fortunately, avascular necrosis
the journal. (AVN) is a rare complication following a phalangeal neck
DOI: 10.1097/BPO.0000000000002364 fracture.9,14 Risk factors for AVN include open fractures,

Vaslow et al J Pediatr Orthop Volume 43, Number 5, May/June 2023
on 04/15/2023
FIGURE 1. Illustration of Al-Qattan’s classification of types of phalangeal neck fractures. Reprinted from Injury, Vol 46, Al-Qattan,
and Al-Qattan, “A review of phalangeal neck fractures in children, 935-944, Copyright (2015), with permission from Elsevier.
Copyright [Al-Qattan and Al-Qattan]. All permission requests for this image should be made to the copyright holder.
Al-Qattan type III fractures, open reduction of closed database. Subjects presenting with open injuries (OI) were
fractures, and multiple unskilled attempts at closed compared with patients who presented with closed injuries
reduction.9,10,14,15 The rate of AVN following either open that required open reduction (COR) and closed injuries
or closed treatment of these pediatric fractures is currently with closed reduction (CCR).
unknown. We hypothesize that subjects with open injuries For the purposes of this study, AVN was defined as
would display higher rates of osteonecrosis and poorer radiographic sclerosis and fragmentation of the phalangeal
postsurgical outcomes than subjects with closed injuries head. In addition, the fracture outcomes were measured
requiring open reduction or those treated with CRPP. with the Al-Qattan outcome scores, stiffness, range of
motion (ROM), and coronal malangulation. Stiffness was
MATERIALS AND METHODS defined as loss of motion, extensor lag, or contracture of at
We performed a retrospective study of subjects who least 10 degrees. The ROM was categorized as excellent,
underwent open or closed surgical treatment of phalangeal good, fair, or poor according to the Al-Qattan classi-
head and neck fractures at a single pediatric tertiary re- fication. Coronal malangulation was defined as radio-
ferral center from 2007 to 2017. After institutional review graphic angulation of ≥ 10 degrees. The 3 groups were
board approval, we performed a manual chart review of compared using Pearson χ2 tests for categorical variables
all phalangeal fractures treated surgically. Cases were and ANOVA for continuous data. Two group comparisons
identified by ICD-9 and ICD-10 codes for phalangeal were made with the Student t test with significance set to
fractures. Inclusion criteria were subject age less than P < 0.05. Additional analysis of the surgical outcomes was
eighteen years, fracture pattern involving the phalangeal performed, grouping fractures by intra-articular versus ex-
head or neck of the proximal or middle phalanx of the tra-articular fractures using Pearson χ2 tests.
fingers or the proximal phalanx of the thumb. Patients
were not excluded if there were other fractures or injuries. RESULTS
Subjects’ demographic data, mechanism of injury, oper- One hundred sixty-five patients were identified.
ative reports, radiographs, associated injuries, and clinical There were 17 OI in 16 patients, 14 COR fractures, and
course were recorded in a de-identified electronic 136 CCR fractures. The mean age for the OI was
J Pediatr Orthop Volume 43, Number 5, May/June 2023 Operative Fractures of the Phalangeal Head and Neck
TABLE 1. Demographic, Presurgical and Follow-up Data

CCR COR OI (Fractures, n = 17.
(n = 136) (n = 14) Patients, n = 16) OI vs. CCR OI vs. COR COR vs. CCR
Age, years (Range, SD) 9.8 (1-17, 4.21) 11.4 (4-16, 4.08) 5.9 (1-16, 4.40) P < 0.005 P < 0.005 P = 0.08
P = 0.723 P = 0.919 P = 0.842
Male Sex, (%) 66.9 64.3 62.5

Mechanism, n (%)
Crush 25(19.2) 2 (15.4) 9 (56.3) P < 0.0005 P < 0.005 P = 0.704
Sport 103 (79.2) 11 (84.6) 1 (6.25) P < 0.0005 P < 0.0005 P = 0.813
Machines 2 (1.54) 0 (0) 3 (18.8) P < 0.0005 P = 0.0720 P = 0.648
Sharp 0 (0) 0 (0) 3 (18.8) P < 0.0005 P = 0.0810 P > 0.999
Unknown 6 (4.41) 1 (7.14) 0 (0) P = 0.387 P = 0.256 P = 0.644
Type of fracture, n (%)
Extra-articular 123 (90.4) 9 (64.3) 11 (64.7) P < 0.005 P = 0.980 P < 0.005
(Subcondylar)
Intra-articular 12 (8.8) 5 (35.7) 6 (35.3)
Bi-condylar 4 (2.9) 4 (28.5) 3 (17.6) P < 0.05 P = 0.469 P < 0.0005
Unicondylar 8 (5.9) 1 (7.1) 3 (17.6) P = 0.078 P = 0.385 P = 0.856
Unknown 1 (0.7) 0 (0) 0 (0) P < 0.05 P < 0.05
Laterality right, % Right 77 (56.6) 8 (57.1) 8 (47.1) P = 0.455 P = 0.576 P = 0.970
Finger distribution, n (%)
Thumb 6 (4.4) 0 (0) 2 (11.8) P = 0.200 P = 0.185 P = 0.423
Index 13 (9.6) 2 (14.2) 1 (5.9) P = 0.620 P = 0.431 P = 0.575
on 04/15/2023
Middle 23 (16.9) 2 (14.2) 0 (0) P = 0.066 P = 0.156 P = 0.802

Ring 29 (21.3) 1 (7.1) 4 (23.5) P = 0.835 P = 0.217 P = 0.207
Small 65 (47.8) 9 (64.3) 10 (58.8) P = 0.391 P = 0.756 P = 0.240
Phalanx distribution, n (%)
P1 90 (66.2) 10 (71.4) 6 (35.3) P < 0.05 P = .0108 P = 0.887
P2 46 (33.8) 4 (28.6) 11 (64.7)
Associated Injuries 1 3 15 P < 0.0005 P < 0.0005 P < 0.05
Injury to surgery time, 10.41 (1-30, 5.78) 22.42 (3-76, 18.9) 1.625 (0-11, 2.66) P < 0.05 P < 0.05 P < 0.05
days (Range, SD)
Follow-up length, days 61.9 (23-229, 35.0) 147.6 (42-515, 122.2) 250.1 (49-1203, 318.7) P < 0.0001 P = 0.133 P < 0.0001
(Range, SD)
AVN indicates Avascular Necrosis; CCR, closed injury treated with closed reduction; COR, closed injury undergoing open reduction; OI, open injury.
5.9 years; this was significantly younger than CCR (9.8 y) Concomitant injuries were present in 15 of 16 OI
and COR groups (P < 0.005, Table 1). The patients were subjects, consisting of 12 extensor tendon ruptures, 6 ip-
predominantly male, with no statistically significant sidigital fractures, 3 digital nerve transections, 2 flexor
difference between groups (P > 0.5). tendon ruptures, and 1 vascular disruption requiring mi-
Mechanisms of injury were classified as crushing, crosurgical repair. The COR group had 3 patients with
sports (low-energy blunt injuries including falls), ma- concomitant injuries, 1 with a 90% partial rupture of the
chines, and sharp. Open injuries were more likely to be terminal extensor tendon and 2 patients with 2 P2 shaft
caused by a crush injury, while in the CCR and COR fractures on other digits. The OI group was significantly
groups, sports injuries predominated. Intra-articular in- more likely to have associated injuries than the COR or
juries, particularly bi-condylar fractures, were seen more CCR group (P < 0.0005).
frequently in the COR and OI groups (P < 0.05) Postoperatively, AVN was noted in 70.6% in the OI
(Table 1). group, 7.1% in COR, and 1.5% in CCR (Table 2). The 1
OI subjects were operated on earlier than those in fracture in the COR group with AVN was hit by a
the COR or CCR group, even with 1 outlier in the OI baseball. The 2 AVN patients in the CCR group sustained
group; this patient’s procedure was delayed until 11 days a crush injury and the other a fall, with the latter patient
post-injury, following irrigation and primary closure at an having complete radiographic remodeling of the digit over
outlying facility. In the COR group, 3 patients required a period of 6 years. AVN was diagnosed at a mean time of
open reduction of fractures that were irreducible by closed 132 days from injury with a median time of 71 days
means (injury-surgery interval 3-5 d), while the remainder (range, 28 to 608 d). However, there were 4 patients who
were able to be reduced by percutaneous osteoclasis (in- did not return to the clinic after the pin pull and
terval 12-76 d). Each group showed a statistically sig- subsequently re-presented from 4 to 20 months after the
nificant difference in the time to operation, with the OI initial injury due to deformity and stiffness. Without the 4
group undergoing the soonest operation and the COR outliers, the mean time to radiographic AVN diagnosis
undergoing the latest (Table 1). Follow-up for all fractures was 60.5 days from injury with a median time of 61 days
averaged 86.5 days (range, 0 to 1203), with the OI group (range, 28 to 101).
having the longest follow-up (250.1 d) and the CCR group No significant coronal malangulation on the final
the shortest (61.9 d). follow-up x-rays for each patient was found in the COR
TABLE 2. Postoperative Outcomes

CCR COR OI OI vs CCR OI vs COR COR vs CCR
AVN, n (%) 2 (1.5) 1 (7.1) 12 (70.6) P < 0.0005 P < 0.0005 P > 0.05
Coronal malangulation, n (%) 3 (2.2) 0 (0) 5 (29.4) P < 0.0005 P = 0.396 P < 0.005
P < 0.0005 P < 0.0005 P = 0.400
Stiffness, n (%) 24 (17.6) 8 (57.1) 11 (64.7)

— — —
Range of motion, n (%) (120) (12) (15)

Excellent 111 (92.5) 3 (25.0) 4 (23.5) P < 0.0005 P = 0.921 P < 0.0005
Good 4 (3.3) 2 (16.7) 2 (11.7) P = 0.076 P = 0.809 P < 0.05
Fair 5 (4.2) 6 (50.0) 6 (35.3) P < 0.0005 P = 0.414 P < 0.0005
Poor 0 (0) 1 (8.3) 3 (17.6) P < 0.0005 P < 0.05 P < 0.0005
Mean total active motion (deg) 279 261 258 P < 0.05 P = 0.394 P < 0.05
Al-Qattan Outcome, n (%)
Excellent 111 (81.6) 3 (21.4) 1 (5.88) P < 0.0005 P = 0.141 P < 0.0005
Good 15 (11.0) 3 (21.4) 1 (5.88) P = 0.513 P = 0.198 P = 0.254
Fair 7 (5.1) 4 (28.6) 2 (11.8) P < 0.005 P = 0.842 P < 0.005
Poor 3 (2.20) 4 (28.6) 13 (76.4) P < 0.0005 P < 0.05 P < 0.0005
AVN indicates Avascular Necrosis; CCR, closed injuries undergoing closed reduction; COR, closed injuries undergoing open reduction; OI, open injuries.
group, but 3 (2.2%) patients in the CCR group and 5 patient developed a nonunion that was asymptomatic. For
(29.4%) patients in the OI group (P < 0.05) had significant
on 04/15/2023
the 2 poor results in the COR group, 1 was stiff at the PIP
coronal malangulation. joint, and 1 had AVN. Of the CCR group’s poor results
Postoperative stiffness was noted in 64.7% and (3 patients) 2 had AVN, and the other had cascade-
57.1% of patients in OI and COR groups, respectively, altering rotational deformity but ultimately declined
versus only 17.1% in the CCR group (P < 0.05, Table 2). corrective surgery. Three infections were reported; 2 in
Detailed ROM data was available for 15 fractures in OI the OI group and 1 in CCR which all resolved with a
group, 12 in COR group, and 120 fractures in the CCR 14-day course of Clindamycin.
group. Two patients in the OI group were excluded. One There were a total of 142 extra-articular fractures, 23
patient went on to amputation, and the other was noted to intra-articular fractures, and 1 unknown without available
have limited ROM, but no measurements were available. injury radiographs. Intra-articular fractures showed a
Two fractures in the COR group were excluded due to a significantly higher rate of AVN (30.4% vs. 5.6%,
lack of measurements, but both were noted to have limited P < 0.0005) (Table 3). Stiffness was present in 19.7% of
ROM. The exclusions in the CCR group included 9 extra-articular versus 39.1% of intra-articular fractures
patients who were referred to OT or started on a home (P < 0.05). ROM data was available for 128 extra-
exercise program due to concern for stiffness, but no articular and 19 intra-articular fractures. ROM was
measurements of ROM were made, 6 patients due to lack excellent in 85.2% of extra-articular and 47.4% of intra-
of data, but no complications were noted, 1 excluded due articular fractures and fair in 7.0% of extra-articular and
to reinjury, and another was referred for surgery for 42.1% of intra-articular fractures (P < 0.05). No difference
cascade-altering deformity. The total active motion was was seen between a good and poor range of motion
279 degrees in the CCR group, 261 degrees in COR, and outcomes for the 2 groups. Based on Al-Qattan’s
257 degrees in OI. The CCR group had higher total active
motion than the OI and COR groups, and there was no
difference between the COR and OI group. Using the Al- TABLE 3. POSTOPERATIVE OUTCOMES Intra-articular Versus
Qattan ROM classification, the CCR group had more Extra-articular fractures
excellent and fewer good, fair, and poor ROM outcomes Extra-articular Intra-articular
than both the OI and COR groups (Table 2). (n = 142) (n = 23)
Based on Al-Qattan’s classification, the CCR group AVN, n (%) 8 (5.6) 7 (30.4) P < 0.0005
had significantly more excellent results than both COR Stiffness, n (%) 28 (19.7) 9 (39.1) P < 0.05
and OI groups (P < 0.00050), with OI having significantly Range of motion, n (%) 128 19 —
more poor results compared with COR and CCR Excellent 109 (85.2) 9 (47.4) P < 0.0005
Good 7 (5.5) 1 (5.3) P = 0.971
(P < 0.05) (Table 2). Among 13 patients with poor results Fair 9 (7.0) 8 (42.1) P < 0.0005
in the OI group, 12 had AVN, 8 had significant stiffness, Poor 3 (2.3) 1 (5.3) P = 0.465
and 3 had residual deformity great enough to compromise Coronal 5 (3.5) 3 (13.0) P = 0.054
function. Several patients had multiple issues that caused malangulation, n (%)
poor function. One patient with a proximal phalanx neck Al-Qattan outcome, n (%)
Excellent 106 (81.6) 9 (21.4) P < 0.005
fracture and associated digital nerve, artery, and extensor Good 16 (11.0) 3 (21.4) P = 0.786
mechanism injuries went on to undergo a partial finger Fair 10 (5.1) 3 (28.6) P = 0.311
amputation at the P1 level of the fracture after the failure Poor 12 (2.20) 8 (28.6) P < 0.0005
of arterial repair 11 days postoperatively. No AVN of the AVN indicates Avascular Necrosis.
bone was noted at the time of amputation. Another
J Pediatr Orthop Volume 43, Number 5, May/June 2023 Operative Fractures of the Phalangeal Head and Neck
classification of results, the extra-articular fractures had reduction. Given the short duration of follow-up for the
significantly more excellent results, and the intra-articular CCR group compared with the COR and OI groups,
group had a greater proportion of poor results than the leading to follow-up bias, we may have been able to detect
extra-articular group (P < 0.005). more AVN in the CCR group with a longer duration of
follow-up, while the increased duration of follow-up and
increased injury severity may have led to the increased rate

DISCUSSION of detection in the OI group. Time to the identification

Pediatric phalangeal neck fractures can be chal- of AVN was also subject to follow-up bias due to the
lenging to treat. Displaced fractures are often unstable, 4 patients who failed to return for routine follow-up
and the potential exists for nonunion and poor remodel- until months to years after injury, but for those who re-
ing. Many authors recommend percutaneous pin fixation turned to the clinic as scheduled, both mean and median
of these fractures with either open or closed reduction to time for radiographic identification of AVN were 61 days
reduce those risks.10,12,13 Delayed presentation of these from injury, leading the authors to believe that AVN
fractures provides additional challenges as a closed re- may be identified much earlier than we had previously
duction may not be sufficient to achieve adequate align- assumed.
ment, and open reduction increases the risk for It remains unclear how much remodeling potential
postoperative stiffness and other complications. phalangeal neck fractures in patients with residual ma-
Two previous authors have described AVN com- langulation or AVN possess. We had 1 patient with AVN
plications as a result of open reduction.10,14 To mitigate who demonstrated excellent remodeling after 6 years.
In addition, Cornwall et al17 presented a case report of a
on 04/15/2023
this risk, some surgeons avoid open treatment until no

other alternative is viable. Waters et al16 described their 5-year-old child with near complete remodeling of a
treatment of 8 patients with percutaneous osteoclasis proximal phalangeal neck fracture malunion with an ex-
through the dorsal callus for malunions. They described cellent final ROM and function as did Hennrikus and
good results with no cases of AVN. Nearly all reports Cohen18 in 3 late presenting phalangeal neck fractures in
discuss AVN in the setting of closed injuries with open 2 children. Puckett and colleagues retrospectively reviewed
treatment, and few reports exist regarding the rate after 8 patients with displaced and angulated pediatric pha-
open injuries. Our results show that opening the fracture langeal neck fracture malunions treated without surgery.
either surgically or at the time of injury increases the All patients went on to union with significant remodeling
postoperative risk of AVN compared with closed treat- and no functional loss of motion.19 Although these reports
ment. The OI and COR study groups had higher pro- have small numbers, they support the potential for
portions of high energy mechanisms, including crush continued remodeling in skeletally immature, displaced
injuries and concomitant injuries, than the CCR group. pediatric phalangeal neck fractures with advanced healing
This supports that the differences in outcomes are related rather than risk AVN and stiffness after open reduction.
to the higher amount of initial trauma that each finger
sustains. Soft tissue injury from the initial mechanism as CONCLUSION
well as the surgical procedure, increases the risk for stiff- Pediatric phalangeal head and neck fractures pre-
ness and less than optimal range of motion postprocedure.
senting as open injuries or requiring open reduction have
In addition, open treatment was more likely to be required
increased postoperative risks of stiffness, AVN, and
for fractures that were intra-articular, as they are often overall less favorable outcomes compared with fractures
more difficult to treat and involve smaller bone fragments
with closed reduction and pin fixation. Parents should be
which may be more vulnerable to vascular compromise.
counselled appropriately for these risks.
Our data supports this, as intra-articular fractures dem-
onstrated a higher rate of AVN than extra-articular frac- REFERENCES
tures. Fractures of the joint surface may also comprise the 1. Landin LA. Epidemiology of children’s fractures. J Pediatr Orthop
motion of the joint. Parents of patients with open injuries, Part B. 1997;6:79–83.
intra-articular fractures, and fractures that may require 2. Nellans KW, Chung KC. Pediatric hand fractures. Hand Clin.
open reduction should be counselled appropriately to the 2013;29:569–578.
3. Rajesh A, Basu AK, Vaidhyanath R, et al. Hand fractures: a study of
increased risk of AVN of their child’s fracture and in- their site and type in childhood. Clin Radiol. 2001;56:667–669.
creased likelihood of a less than the excellent outcome. 4. Mahabir RC, Kazemi AR, Cannon WG, et al. Pediatric hand
Limitations of this study include its retrospective fractures: a review. Pediatr Emerg Care. 2001;17:153–156.
nature, short follow-up period (as is typical for post- 5. Cornwall R, Ricchetti ET. Pediatric phalanx fractures: unique
challenges and pitfalls. Clin Orthop Relat Res. 2006;445:146–156.
operative uncomplicated CRPP of finger fractures), lack 6. Abzug JM, Dua K, Sesko Bauer A, et al. Pediatric phalanx fractures.
of detailed OT notes, and small group sizes of the OI and Instr Course Lect. 2017;66:417–427.
COR cohorts. With longer follow-up and/or OT, many 7. Leonard MH, Dubravcik P. Management of fractured fingers in the
patients with stiffness may have achieved a full ROM. In child. Clin Orthop Relat Res. 1970;73:160–168.
addition, disruption to the blood supply to the phalangeal 8. Kang HJ, Sung SY, Ha JW, et al. Operative treatment for proximal
phalangeal neck fractures of the finger in children. Yonsei Med J.
head may occur after aggressive open reduction or open 2005;46:491–495.
trauma, but we were not able to prove conclusively, which 9. Al-Qattan MM, Al-Qattan AM. A review of phalangeal neck
was the culprit for open fractures that had an open fractures in children. Injury. 2015;46:935–944.
10. Al-Qattan MM. Phalangeal neck fractures in children: classification 15. Al-Qattan MM, Al-Munif DS, AlHammad AK, et al. The outcome
and outcome in 66 cases. J Hand Surg (Edinburgh, Scotland). of management of “troublesome” vs “non-troublesome” phalangeal
2001;26:112–121. neck fractures in children less than 2 years of age. J Plastic Surg
11. Barton NJ. Fractures of the hand. J Bone Joint Surg. 1984;66B: Hand Surg. 2016;50:93–101.
159–167. 16. Waters PM, Taylor BA, Kuo AY. Percutaneous reduction of
12. Paksima N, Johnson J, Brown A, et al. Percutaneous pinning of incipient malunion of phalangeal neck fractures in children. J Hand
middle phalangeal neck fractures: surgical technique. J Hand Surg. Surg. 2004;29A:707–711.
2012;37A:1913–1916. 17. Cornwall R, Waters P. Remodeling of phalangeal neck fracture

13. Karl JW, White NJ, Strauch RJ. Percutaneous Reduction and malunions in children: case report. J Hand Surg. 2004;29A:458–461.
fixation of displaced phalangeal neck fractures in children. J Pediatr 18. Hennrikus WL, Cohen MR. Complete remodeling of displaced fractures
Orthop. 2012;32:156–161. of the neck of the phalanx. J Bone Joint Surg Br. 2003;85:273–274.
14. Topouchian V, Fitoussi F, Jehanno P, et al. [Treatment of 19. Pucket BN, Gaston RG, Peljovich AE, et al. Remodeling potential
phalangeal neck fractures in children: technical suggestion]. Chir of phalangeal distal condylar malunions in children. J Hand Surg.
Main. 2003;22:299–304. 2012;37A:34–41.
on 04/15/2023
ORIGINAL ARTICLE
Pediatric Supracondylar Humerus Fracture

AAOS Appropriate Use Criteria: Does Treatment
at a Pediatric Level I Trauma Center Comply?

Diana G. Douleh, MD,*† Patrick Carry, MS,† Julia S. Sanders, MD,† and Jason T. Rhodes, MD†
Background: The primary purpose of this analysis was to com-

pare supracondylar humerus fracture (SCHF) treatment patterns
S upracondylar humerus fractures (SCHFs) are among the
most commonly treated fractures among the pediatric
population encountered by orthopaedic surgeons.1,2 Tradi-
at a single quaternary pediatric hospital relative to the American
tionally, displaced fractures have been managed with prompt
Academy of Orthopedic Surgeons (AAOS) appropriate use
surgical treatment secondary to the risk of complications,
treatment recommendation(s).
including vascular or neurological injury, compartment
Methods: Among all fractures included in the cohort (n = 571),
on 04/15/2023
syndrome, and malunion.2–5 In 2012, the American Acad-

the observed treatment approach was evaluated relative to the
emy of Orthopedic Surgeons (AAOS) aimed to provide
AAOS “Appropriate” treatment recommendation(s). The pro-
guidance for the treatment of SCHFs based on best available
portion, and corresponding 95% confidence interval, of cases
evidence, and published these recommendations in the form
that agreed with the “Appropriate” treatment recommendation
of the Clinical Guideline Practice Summary in 2012.6 In
was estimated. Demographics and clinical characteristics
2014, the AAOS formulated these recommendations into the
among cases that were managed in accordance with the “Ap-
appropriate use criteria (AUC).7 Indications for manage-
propriate,” “May be Appropriate,” or “Rarely Appropriate”
ment were determined by systematic review of the literature
were compared.
with consideration of fracture type, vascular status, asso-
Results: All fractures were treated according to the “Appro-
ciated nerve injuries, soft tissue envelope, ipsilateral radius
priate,” “May be Appropriate,” or “Rarely Appropriate” AAOS
and/or ulna fracture, and degree of swelling.7
treatment guidelines. The observed treatment among fractures
The AUC document lists 220 scenarios which were
included in the cohort agreed with AAOS “Appropriate” recom-
reviewed by a multidisciplinary panel, and appropriateness
mendations in 92.1% [95% confidence interval (CI): 89.6%-94.2%]
of treatment was determined to be “Appropriate,” “May be
of the cases. Fracture type differed significantly between patients
Appropriate,” or “Rarely Appropriate.”7 The salient rec-
treated according to AAOS “Appropriate” recommendations
ommendations of the AUC on the basis of Gartland fracture
compared to those treated according to “May be Appropriate,” or
type8 advocate for nonoperative treatment of type I fractures
“Rarely Appropriate” recommendation.
with immobilization and splinting or casting, and closed
Conclusions: The treatment approach implemented at a single
reduction and percutaneous pinning of type II, type III, and
level 1 trauma center was in concordance with the appropriate
flexion-type fractures.7 However, recent evidence has sug-
use criteria treatment recommendations in a significant majority
gested deviation from these recommendations in clinical
of cases. Fractures not treated according to “Appopriate” rec-
practice may be acceptable.9–11
ommendations were primarily type IIA injuries, and were treated
Wang et al12 characterized management of SCHFs at
with closed reduction and casting instead of the recommended
their institution in comparison to the AUC recommendations,
closed reduction and percutaneous pinning.
and found that the AUC recommended operative manage-
Level of Evidence: Level III.
ment of type II fractures, and urgent/emergent operative in-
Key Words: supracondylar fractures, humerus, pediatrics, appropriate tervention more frequently than was performed. Prompted by
use criteria these findings, the purpose of our study was to augment ex-
isting evidence for the applicability of the AUC in clinical
(J Pediatr Orthop 2022;42:e470–e473) practice and potentially support modifications to the criteria to
improve utility in clinical decision making. We sought to
compare SCHF treatment patterns at a single quaternary pe-
From the *Department of Orthopaedics, University of Colorado School
of Medicine, Aurora; and †Department of Orthopedic Surgery, diatric hospital relative to the AAOS AUC, and describe areas
Children’s Hospital Colorado, Anschutz, CO. of treatment agreement and disagreement.
No funding was required for the currently submitted project.
The authors declare no conflicts of interest.
Reprints: Diana G. Douleh, MD, Department of Orthopedic Surgery, METHODS
University of Colorado, 12631 E. 17th Avenue, Mail Stop B202,
Aurora, CO 80045. E-mail: diana.douleh@cuanschutz.edu. After obtaining institutional IRB approval, we retro-
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved. spectively reviewed 571 cases of SCHFs treated at our
DOI: 10.1097/BPO.0000000000002115 institution between April 2012 and March 2015. Eleven

J Pediatr Orthop Volume 42, Number 5, May/June 2022 Pediatric Supracondylar Humerus Fracture
TABLE 1. Demographics and Fracture Characteristics Among Fractures That Were Versus Were Not Treated According to AAOS
Appropriate Use Criteria Recommendations
Agrees With Appropriate (n = 526) Agrees With May Be/Rarely Appropriate (n = 45)
Mean |Median|, N SD |IQR| Mean |Median|, N SD |IQR| P
Age at Treatment, mean (SD) 5.5 2.4 5 2.7 0.1662

Not Hispanic or Latino, n (%) 213 40.5 17 37.8 0.7214

Female, n (%) 271 51.5 25 55.6 0.6031
Ipsilateral Fracture, n (%) 503 95.6 44 97.8 0.7119
White, n (%) 336 63.9 30 66.7 0.7082
Right side, n (%) 319 60.6 29 64.4 0.6268
Open fracture, n (%) 10 1.9 0 0.0 > 0.9999
Severe swelling, n (%) 32 6.1 0 0.0 0.1643
Fracture type, n (%) < 0.0001
Type 1 108 20.5 0 0.0
Type 2 a 12 2.3 42 93.3
Type 2 b 121 23.0 1 2.2
Type 3 271 51.5 2 4.4
Flexion 14 2.7 0 0.0
Vascular injury, n (%) 17 3.2 0 0.0 0.3854
Nerve injury, n (%) 63 12.0 0 0.0 0.0059
on 04/15/2023
Statistical significance values are in bold.
pediatric-fellowship trained orthopaedic surgeons were in- 42 of these cases were treated with immobilization and casting.
volved in the management of all fractures. Patient data were AAOS recommended treatment for these fractures included
obtained from the electronic medical record and entered into emergent or urgent closed reduction and pinning in all cases.
a REDCap database. Patient age, sex, date/time of injury, Only 2 IIA patients eventually resulted in operative treatment
laterality of fracture, hand dominance, Gartland fracture after initially being treated with immobilization and casting;
type,8 vascular and neurological status, condition of soft therefore, the failure rate of this treatment method was 4.8%.
tissue envelope, concomitant ipsilateral upper extremity The decision to convert to closed reduction and percutaneous
fracture, injury treatment modality, and date/time of surgery pinning was made at the time of follow-up, if a loss of
were reviewed from the medical record. reduction was noted with radiographic parameters (anterior
humeral line intersecting the capitellum).
Statistical Methods
Among all fractures included in the cohort (n = 571), the
observed treatment approach was evaluated relative to the DISCUSSION
AAOS “Appropriate” treatment recommendation(s). The In the management of SCHFs at our institution,
proportion, and corresponding 95% confidence interval (CI), fractures that were managed with a treatment approach
of cases that agreed with the “Appropriate” treatment rec-
ommendation was estimated. Demographics and clinical
characteristics among cases that were managed in accordance TABLE 2. Treatment and Fracture Characteristics among Type
IIa Fractures Versus All Other Fractures
with the “Appropriate,” versus the “May be Appropriate,” or
“Rarely Appropriate” recommendation were compared. Type IIA Fractures All Other Fractures
(n = 54) (n = 517)
Frequency % Frequency %
RESULTS
Agreement category
The observed treatment among fractures included in Appropriate 12 22.2 515 99.6
the cohort agreed with AAOS “Appropriate” treatment in May Be 1 1.9 0 0.0
92.1% (95% CI: 89.6%-94.2%) of the cases (Table 1). All Appropriate
fractures agreed with either the “Appropriate” treatment Rarely Appropriate 41 75.9 2 0.4
recommendation, “May be Appropriate,” or “Rarely Vascular Injury 0 0.0 17 3.3
Nerve injury 0 0.0 63 12.2
Appropriate” treatment recommendation. Fractures that Neurovascular injury 0 0.0 69 13.3
did not agree with the “Appropriate” treatment approach Treatment
tended to be less severe in nature. Casting 42 77.8 110 21.3
Wilkins modified the Gartland classification, subdividing CRPP 12 22.2 397 76.8
ORIF 0 0.0 10 1.9
type II fractures into type IIA fractures, which have no rota- Unanticipated surgical 2 3.7 1 0.2
tional deformity or translation, and IIB fractures, which do intervention
have these characteristics of instability.13 Among type IIA
CRPP indicates closed reduction percutaneous fixation; ORIF, open reduction
fractures, treatment at our institution did not agree with internal fixation.
AAOS recommendations in 42/56 cases, or 75% (Table 2). All

Douleh et al J Pediatr Orthop Volume 42, Number 5, May/June 2022
TABLE 3. Treatment Characteristics Among Type III Fractures that Did Not Agree With AAOS Appropriate Use Criteria Appropriate
Treatment Recommendation
Fracture Characteristics Actual Treatment AAOS Appropriate Recommendation
Open Vascular Nerve Pin Pin
ID Type Fracture Injury Injury Treatment Configuration Timing Treatment Configuration Timing
278 Type III Absent Absent Absent Closed Lateral OP Closed Lateral OR crossed Urgent/emergent
240 Type III Absent Absent Absent Open NA Urgent/emergent Closed Lateral OR crossed Urgent/emergent
AAOS indicates American Academy of Orthopedic Surgeons; NA, not applicable; OP, outpatient.
that did not agree with the “Appropriate” treatment rec- treated operatively, and based on our findings has a <5%
ommendation tended to be less severe than fractures that chance of failure and need for subsequent surgical fixation.
were treated in a manner that was consistent with the This cohort of patients constituted the majority of patients
AAOS AUC. All treating surgeons were familiar with that were treated in a manner that did not comply with
AAOS AUC guidelines throughout the study period. AUC’s “Appropriate” recommendations.
Treatment was in accordance with AUC’s recom- To our knowledge, 2 prior studies have evaluated
mendations in 92.1% of cases. Of the 44 cases in which clinical management of SCHFs as it compares to AUC’s
treatment was not in compliance with the “Appropriate” recommendations.12,14 Ibrahim et al14 evaluated operative
on 04/15/2023
treatment recommendation, 42 occurred in the manage- cases of SCHFs and determined that practice at their in-
ment of type IIA fractures. Two instances occurred in the stitution complied with AUC’s recommendations in 89.4%
management of type III fractures (Table 3). These injuries of cases. Wang et al12 found further discordance with
included a type III injury which required open reduction, AUC’s recommendations, with disagreement in 31% of
and a type III injury that was splinted at an outside urgent cases, in their evaluation of management of both non-
care, referred to our institution for outpatient follow-up, operative and operative fractures at their institution. The
and indicated for surgical management. majority of instances of disagreement stemmed from type
The AUC does not differentiate between type IIA II fractures treated nonoperatively despite AUC’s recom-
and type IIB subtypes in outlining treatment recommendation for closed reduction and percutaneous pinning,
mendations, although the presence or absence of angula- and from type III fractures not treated in an urgent/
tion is included. The authors of this study have included a emergent manner.12
differentiation between IIA and IIB fracture types, as the The AUC recommended operative management of
distinction is routinely among our providers in clinical type II SCHFs more frequently than was practiced, and the
decision making. The AUC itemizes recommendations results from our institution add to the growing literature
categorially as “Appropriate,” “May be Appropriate,” supporting practice deviation from AUC’s “Appropriate”
and “Rarely Appropriate.” For type II fractures without recommendations may be acceptable in the management of
angulation, and a perfused hand, no nerve injury, no ip- type IIA injuries. Treatment decisions for type IIA fractures
silateral fracture, and typical swelling, the “Appropriate” may have been influenced by surgeon preference, history of
recommendation is urgent or outpatient operative reduc- fracture, as well as timing and location of patient’s initial
tion with percutaneous pinning. Reduction with sub- presentation at our institution. However, treatment was in
sequent casting at 70 to 90 degrees “May be Appropriate.” agreement with the AAOS “Appropriate” use recom-
Several recent studies have offered validation for mendation in instances of more severe injury (type IIB and
nonoperative management of certain type II fractures.9–11 above).
Spencer et al established guidelines for the management of Although our study is limited by its retrospective design
type II SCHFs; in the absence excessive swelling, varus, without availability of patient outcome data, our study is
valgus, or rotational deformity, shaft-condylar axis <15 strengthened by the inclusion all SCHFs treated a single
degrees, or loss of reduction, the authors recommend quaternary care hospital over a 3-year period in a large co-
closed reduction and casting.9 At our institution, there is hort of 571 patients. Our results provide further justification
no standardized protocol guiding management of type II for an evaluation of the practical application of the AAOS
fractures, and pediatric orthopaedic surgeons individually AUC in future iterations. Future directions will include
decide on course of treatment. Our results indicate the evaluation of clinical outcomes of type II fractures treated
majority clinicians elected for nonoperative management non-operatively. Such data will aim to provide possible sup-
of type IIA fractures. When nonoperative management is port for an “Appropriate” treatment recommendation for
chosen, close follow-up is instituted. If the fracture is in- nonoperative management of certain type II fractures.
adequately reduced or loses reduction at follow up (the We hope the results from our institution will provide
anterior humeral line does not intersect the capitellum), data from the current practice of pediatric orthopaedic
then the patient is indicated for closed reduction and surgeons that may be used to guide clinical outcomes
percutaneous pinning. Treatment of type IIA fractures studies to support future modifications to the AUC,
with closed reduction and casting results in more rapid and create a tool that can guide decision making to better
treatment of these fractures as compared with fractures reflect current practice.

J Pediatr Orthop Volume 42, Number 5, May/June 2022 Pediatric Supracondylar Humerus Fracture
REFERENCES 8. Gartland JJ. Management of supracondylar fractures of the humerus

1. Kumar V, Singh A. Fracture supracondylar humerus: a review. in children. Surg Gynecol Obstet. 1959;109:145–154.
J Clin Diagn Res. 2016;10:RE01–RE06. 9. Spencer HT, Dorey FJ, Zionts LE, et al. Type II supracondylar
2. Tomaszewski R, Wozowicz A, Wysocka-Wojakiewicz P. Analysis of humerus fractures: can some be treated nonoperatively? J Pediatr
early neurovascular complications of pediatric supracondylar humerus Orthop. 2012;32:675–681.
fractures: a long-term observation. BioMed Res Int. 2017;2017:2803790. 10. Silva M, Delfosse EM, Park H, et al. Is the “Appropriate Use
3. Otsuka NY, Kasser JR. Supracondylar fractures of the humerus in Criteria” for type II supracondylar humerus fractures really
children. J Am Acad Orthop Surg. 1997;5:19–26. appropriate? J Pediatr Orthop. 2019;39:1–7.

4. Omid R, Choi PD, Skaggs DL. Supracondylar humeral fractures in 11. Kazimoglu C, Turgut A, Reisoglu A, et al. Are the Appropriate Use
children. J Bone Jt Surg. 2008;90:1121–1132. Criteria for the management of pediatric supracondylar humerus
5. Carrazzone OL, Belloti JC, Matsunaga FT, et al. Surgical interventions fractures useful in clinical practice? J Pediatr Orthop B. 2017;26:395–399.
for the treatment of supracondylar humerus fractures in children: 12. Wang JH, Morris WZ, Bafus BT, et al. Pediatric supracondylar humerus
protocol of a systematic review. JMIR Res Protoc. 2017;6:e232. fractures: AAOS appropriate use criteria versus actual management at a
6. Howard A, Mulpuri K, Abel MF, et al. The treatment of pediatric pediatric level 1 trauma center. J Pediatr Orthop. 2019;39:e578–e585.
supracondylar humerus fractures. J Am Acad Orthop Surg. 2012;20: 13. Alton TB, Werner SE, Gee AO. Classifications in brief: the Gartland
320–327. classification of supracondylar humerus fractures. Clin Orthop Relat
7. Appropriate Use Criteria for the Management of Pediatric Supra- Res. 2015;473:738–741.
condylar Humerus Fractures. American Academy of Orthopedic 14. Ibrahim T, Hegazy A, Abulhail SI, et al. Utility of the AAOS
Surgeons. 2014. Available at: https://www.aaos.org/research/ Appropriate Use Criteria (AUC) for pediatric supracondylar humerus
Appropriate_Use/PSHF_AUC.pdf. Accessed January 4, 2020. fractures in clinical practice. J Pediatr Orthop. 2017;37:14–19.
on 04/15/2023

LETTER TO THE EDITOR
data and indicate that 40 of their multiple authors having shown that this
Pediatric Supracondylar included children with a type IIa frac- is unnecessary for the vast majority of
Humerus Fracture AAOS ture would have had unnecessary sur- type IIa fractures.1,3,4 There is possibly
gery, had the surgeons followed the also a need for these complex fractures
Appropriate Use Criteria: AUC and that 12 of their included to be operated on only by a limited
children with a type IIa fracture most number of experienced surgeons (non-
Does Treatment at a probably had unnecessary surgery be- pediatric orthopaedic surgeons or pe-
Pediatric Level I Trauma cause the surgeons had followed the diatric orthopaedic surgeons) to increase
AUC, exposing the latter children and cases and experience per surgeon, which
Center Comply? their parents to unnecessary stresses and could be facilitated by trauma care cen-
costs (psychological stress; potential loss tralization and by delaying surgery for
of earnings to attend hospital) and sur- those fractures without neurovascular
gical risks and created unnecessary costs compromise until a surgeon with the
To the Editor: for health insurances. appropriate skills is available.6
We read with interest the recent pub- We would like to ask Douleh
lication by Douleh et al.1 Wilkins2 et al1 why 12 of their IIa fractures were Andreas Rehm, MD*
described the modification of the primarily treated with CRPP and about Azeem Thahir, MD†
on 04/15/2023
Gartland classification in 1984, divid- their opinion why malrotation is not Elizabeth Ashby, MD*
ing type II supracondylar humerus considered in the AUC? Could it be Tamás Kobezda, PhD, MD‡
fractures (SCHFs) into type IIa (pos- that a lot of these fractures are Pinelopi Linardatou Novak, MBBCh*
terior angulation without malrotation) managed by nonpediatric orthopaedic Departments of *Paediatric Orthopaedics
and IIb (posterior angulation with surgeons5 in other hospitals, with the †Trauma and Orthopaedics
malrotation). In 2000, O’Hara et al3 AUC creators possibly being concerned ‡Department of Paediatric Orthopaedics
reported a series of 29 children with about some of these surgeons not being Paediatric Division, Cambridge University
type IIa fractures which were treated able to judge malrotation and not Hospitals NHS Trust, Cambridge, UK
with cast immobilization without having the experience to differentiate
K-wire fixation (7 underwent a closed IIa from IIb fractures, potentially re- All authors contributed in literature review and
manuscript preparation.
reduction), with all having achieved sulting in under-treatment of IIb frac-
full range of movement and a normal tures, which require reduction and The authors did not receive any funding for
carrying angle at last follow-up. K-wire fixation? The nonappreciation this work.
Spencer et al4 reported in 2012 that of malrotation is reflected in Sullivan DOI: 10.1097/BPO.0000000000002324
type II fractures without varus/valgus et al’s5 paper where the authors did not
malalignment, without initial malro- differentiate between type IIa and IIb
tation and a shaft-condylar angle ≥ 15 fractures, did not consider malrotation REFERENCES
degrees can be treated successfully as part of their assessment of radio- 1. Douleh DG, Carry P, Sanders JS, et al.
with cast immobilization without Pediatric supracondylar humerus fracture
graphic outcomes of surgically treated AAOS appropriate use criteria: does treat-
K-wires. The latter authors4 stated SCHFs and their included intra- ment at a Pediatric Level I Trauma Center
that their selective approach spared operative lateral radiographs of 1 case, comply? J Pediatr Orthop. 2022;42:e470–e473.
58% of type II fractures from surgery which shows that the distal fragment 2. Wilkins KE. Fractures and dislocations of
the elbow region. In: Rockwood CA, Wilkins
(150 of 259 fractures). was fixed in an unacceptable malro- KE, King RE, eds. Fractures in Children,
Despite Wilkins modification tated position, which was not recog- Vol. 3Philadelphia, PA: JB Lippincott Co.;
and the latter evidence, the American nized as abnormal by the surgeons and 1984:363–575.
Academy of Orthopaedic Surgeons had to be revised.5 Sullivan and col- 3. O’Hara LJ, Barlow JW, Clarke NMP.
(AAOS) (https://www.aaos.org/quality/ leagues also reported that SCHFs were Displaced supracondylar fractures of the
quality-ograms/pediatric-supracondyl- humerus in children. J Bone Joint Surg Br.
operated on by a large number of sur- 2000;82-B:204–210.
ar-humerus-fractures) did not consider geons with the vast majority (21 of 24) 4. Spencer HT, Dorey FJ, Zionts LE, et al. Type
malrotation when it formulated the having performed only very small II supracondylar humerus fractures: can some
“Appropriate use criteria” (AUC) in numbers of surgeries, ranging from 1 be treated nonoperatively? J Pediatr Orthop.
2014. Instead of including malrotation fracture in 77 months to about 3 frac- 2022;32:675–681.
5. Sullivan MH, Stillwagon MR, Nash AB, et al.
as a criterium to decide between non- tures per year. Complications with surgical treatment of pedia-
operative and operative treatment of We conclude, the new evidence tric supracondylar humerus fractures: does
type IIa and IIb fractures respectively, provided by Douleh et al1 and previous surgeon training matter. J Pediatr Orthop.
the AUC recommends closed reduc- 2022;42:e8–e14.
evidence from the literature suggest that 6. Mehlman CT, Strub WM, Roy DR, et al.
tion and percutaneous pinning (CRPP)
there is an urgent need for a change of The effect of surgical timing on the perioper-
for all type II fractures. ative complications of treatment of supra-
Douleh et al’s1 data complement the current AUC, since it recommends condylar humeral fractures in children.
O’Hara et al’s3 and Spencer et al’s4 CRPP for all type II SCHFs, with J Bone Joint Surg. 2001;83:323–327.

ORIGINAL ARTICLE
Policies, Practices, and Attitudes Related to Parental Leave

for Practicing Pediatric Orthopaedic Surgeons
Candice S. Legister, BS,* Sara J. Morgan, PhD,*†‡ Julie B. Samora, MD, PhD,§
Jennifer M. Weiss, MD,∥ Michelle S. Caird, MD,¶ and Daniel J. Miller, MD#**
believed 7 to 12 weeks should be offered for maternity leave

Background: Parental leave impacts family engagement, bonding, (66.2%), 1 to 6 weeks for paternity leave (54.6%), and 7 to 12
stress, and happiness. Because parental leave benefits are im- weeks for adoption leave (46.8%). Many respondents reported
portant to all surgeons regardless of sex, understanding parental taking 1 to 6 weeks of parental leave as a surgeon (53.3%) and
leave practices in pediatric orthopaedic surgery is critical to pro- that their colleagues were supportive of their parental leave
mote equity within the profession and supporting balance in work (40.3%).
and family life. The aim of this study was to survey pediatric Conclusions: Most pediatric orthopaedic surgeons were un-
on 04/15/2023
orthopaedic surgeons about their knowledge of parental leave familiar with parental leave benefits provided by employers.
policies, attitudes towards parental leave, and their individual Respondents who were familiar with these policies believed that
experiences taking leave. more parental leave should be provided, especially for men who
Methods: A 34-question anonymous survey was distributed to may feel social pressure to take less time for leave. Although
the Pediatric Orthopaedic Society of North America member- respondents reported that their work environments were sup-
ship. Eligible respondents were attending pediatric orthopaedic portive, this study identified opportunities for improvement to
surgeons practicing in the United States or Canada. The survey support surgeons who wish to balance parental experiences with
gathered information about employer parental leave policies, work responsibilities.
perceptions about and experiences with parental leave while Level of Evidence: Level V.
practicing as a surgeon, and demographic information about
respondents. Key Words: pediatric orthopaedics, parental leave, paternity
Results: A total of 77 responses were completed and used for leave, maternity leave
analysis. Most respondents were men (59.7%), <50 years old (J Pediatr Orthop 2023;43:337–342)
(67.5%), married (90.9%), and in urban communities (75.3%). A
large majority were practicing in the United States (97.4%). Most
respondents were unfamiliar with employer parental leave poli-
cies (maternity: 53.3%; paternity: 67.5%; and adoption: 85.7%).
Those familiar with policies reported that employers offered 7 to
12 weeks for maternity leave (45.7%) and <1 week for paternity
T he years during medical training and early practice
generally coincide with family planning and child-
bearing pursuits. Medical trainees are often forced to
leave (50%) and adoption leave (45.5%). Most respondents make difficult decisions to juggle personal and pro-
fessional endeavors. Traditionally, this burden is felt
heaviest by women navigating pregnancy and their duties
From the *Research Department; #Department of Orthopaedics, as trainees. Reports in the literature outline professional
Gillette Children’s, St. Paul; †Department of Rehabilitation difficulties, including unsupportive peers or attendings,
Medicine; **Department of Orthopaedic Surgery, University
of Minnesota, Minneapolis, MN; ‡Department of Rehabilitation feelings of guilt, and the risk of graduation delay im-
Medicine, University of Washington, Seattle, WA; §Orthopaedics, pacting board certification and career advancement.1,2 To
Nationwide Children’s Hospital, Columbus, OH; ∥Orthopaedics mitigate these professional challenges, many residents
Department, Southern California Permanente Medical Group, Los delay the start of family expansion until after residency, a
Angeles, CA; and ¶Department of Orthopaedic Surgery, University
of Michigan, Ann Arbor, MI.
trend more common among women than men (73% vs
Financial support for this project was provided by the Pediatric Ortho- 59%, respectively).3,4
paedic Society of North America (POSNA) and the Gillette Children’s There has been a recent push to recruit and retain
Spine Fund. women in orthopaedics, and thus discussions pertaining to
Reprints: Daniel J. Miller, MD, Department of Orthopaedics—Spine,
parental leave benefits and policies have risen to the
Gillette Children’s, 200 University Avenue East, St. Paul, MN 55101. forefront.4–6 However, increased interest in parental leave
E-mail: danmiller@gillettechildrens.com. policies may also reflect a generational shift toward the
Supplemental Digital Content is available for this article. Direct URL desire for better work-life integration by providers in
citations appear in the printed text and are provided in the HTML medicine, regardless of sex.7 Beyond the desire for work-life
pedorthopaedics.com. balance, parental leave offers other tangible benefits to
Copyright © 2023 Wolters Kluwer Health, Inc. All rights reserved. parents and their children. Adequate parental leave im-
DOI: 10.1097/BPO.0000000000002360 proves family engagement, provides more time for bonding,

Legister et al J Pediatr Orthop Volume 43, Number 5, May/June 2023
and decreases parental stress while caring for new the survey. Potential participants who did not fulfill the
children.8–12 Paid maternity leave is associated with lower inclusion criteria based on the screening questions were
rates of neonatal and infant mortality, higher rates of vac- not able to complete the survey. Survey completion was
cination for children, and improved maternal mental voluntary, and participants had the ability to skip any
health.9,10,13 Similarly, men who take paternity leave are question they did not want to answer.
more likely to be involved in the care of their children well The study design and the content of the survey were
after the postpartum period.11 reviewed and approved by POSNA.

Within pediatric orthopaedics, there is a perceived
greater level of commitment to making parental policies Data Analysis
reasonable and accessible, when compared with other or- SAS v.9.4 (SAS Institute Inc., Cary, NC) was used
thopaedic subspecialties.14 The objectives of this study are for analysis. Descriptive statistics were calculated for all
to survey practicing pediatric orthopaedic surgeons about quantitative variables collected, including counts and
parental benefits and policies at their places of employ- percentages for categorical variables and means and SD
ment and to ascertain their level of familiarity with these (or median and ranges) for continuous variables.
policies. In addition, we aim to report on the general at-
titudes of pediatric orthopaedic surgeons toward parental RESULTS
leave. To our knowledge, there are no prior studies on A total of 102 respondents completed the screening
parental leave policies available to pediatric orthopaedic questions preceding the survey (response rate: 6.5%). Five
surgeons, and how they compare to other orthopaedic respondents were not eligible for the study: 3 were not at-
on 04/15/2023
subspecialties. Exploring the current policies may illumi- tending pediatric orthopaedic surgeons, 1 did not practice in
nate potential areas of improvement and lay the ground- the United States or Canada, and 1 opened the survey but
work for establishing uniform guidelines across the field of opted out of participation. Of the remaining 97 eligible re-
orthopaedics. spondents, 20 were excluded due to incomplete responses.
The final data set included 77 survey participants.
METHODS
This cross-sectional study used an online, anony- Participant Characteristics
mous survey to evaluate parental leave policies, attitudes, Of 77 participants, 75 practiced in the United States
and experiences among pediatric orthopaedic surgeons in (Table 1). Survey respondents were predominantly men
the United States and Canada. This study was deemed (59.7% vs 38.9% women), which represents a higher
exempt and approved by an Institutional Review Board. proportion of women compared with the general POSNA
membership (69.4% men and 21.7% women). A larger
Survey Design proportion of respondents in our study were between ages
We designed a 34-question online survey that con- 30 and 39 (33.8% of our sample compared with 15.5% of
sisted of 4 major sections (Supplement 1, Supplemental POSNA members). Most respondents were married or
Digital Content 1, http://links.lww.com/BPO/A579). The partnered (94.8%). Of participants who were married or
first section asked about employer policies regarding pa- partnered, 64.4% had a partner that was employed. Most
rental leave. Survey participants had the option of re- of these employed partners worked full-time (80.9%).
sponding “I do not know” if they were unfamiliar with Most (66.2%) survey participants practiced in an
leave policies. The second section asked about the re- academic setting, 12.9% practiced in a mixed academic and
spondent’s opinions regarding parental leave. Section 3 private environment, 7.8% were in private practice, 6.5% in
sought information from participants who had children a specialty care center, and 6.5% in a community hospital.
within the past 10 years. The fourth section gathered Many of these practices (75.3%) were in urban areas.
demographic data, including age, sex, and marital status.
Survey data were collected and managed with REDCap Employer Parental Leave Policies
electronic data capture tools.15,16 More than half of the survey, participants were un-
familiar with employer maternity and paternity leave poli-
Survey Participants cies (53.3% and 67.5%, respectively) (Table 2). Most
We piloted the survey with 5 attending pediatric participants were unfamiliar with policies surrounding
orthopaedic surgeons to assess content and clarity. The parental leave for adoption (85.7%). Of participants
final survey was initially disseminated to the membership aware of their institutional policies, 7 to 12 weeks were
body of the Pediatric Orthopaedic Society of North offered for maternity leave for 45.7% of respondents, 1 to 6
America (POSNA) in March 2022, and a reminder was weeks were offered to 37.1% of respondents, and <1 week
sent to the membership 2 weeks later. Inclusion criteria was offered to 17.1%. In contrast, <1 week was offered for
required survey participants to (1) be attending pediatric paternity and adoption leave to many participants aware of
orthopaedic surgeons who completed all levels of training employer policies (50% and 45.5%, respectively). Many
and (2) practice in the United States or Canada. Surgeons survey participants (48.1%) were unaware if parental leave
practicing in other orthopaedic subspecialties and trainees was fully or partially paid. Of those aware of compensation
(eg, residents and fellows) were excluded. Eligibility for during parental leave, 42.5% reported that it was fully paid
the study was determined by screening questions preceding and 27.5% reported partial payment.

J Pediatr Orthop Volume 43, Number 5, May/June 2023 Parental Leave Policies in Pediatric Orthopaedics
TABLE 1. Characteristics of Pediatric Orthopaedic Surgeon TABLE 2. Reported Employer Leave Policies and Desired Leave
Respondents (n = 77) Time and Pay for Pediatric Orthopaedic Surgeons
Respondent Characteristic n (%) Current Leave Time Desired Leave Time
and Pay Policies and Pay Policies
Age
30-39 26 (33.8) n (%) n (%)

40-49 26 (33.8)
50-59 13 (16.9) Maternal leave
> 60 12 (15.6) < 1 wk 6 (7.8) 1(1.3)
Mean (SD); years 46.2 (10.3) 1-6 wk 13 (16.9) 10(13.0)
Sex 7-12 wk 16 (20.8) 51(66.2)
Female 30 (39.0) 13-18 wk 0 5(6.5)
Male 46 (59.8) > 19 wk 0 8(10.4)
No response 1 (1.3) I do not know 41 (53.2) —
Race/ethnicity No response 1 (1.3) 2 (2.6)
Alaska Native or American Indian 0 Mean (SD); 7.4 (4.6) 12.1 (5.2)
Asian 9 (11.7) weeks
Black or African American 1 (1.3) Paternal leave
Hispanic or Latinx 0 < 1 wk 12 (15.6) 1 (1.3)
Native Hawaiian or Pacific Islander 0 1-6 wk 7 (9.1) 42 (54.6)
White 66 (85.7) 7-12 wk 5 (6.5) 26 (33.8)
Other 1 (1.3) 13-18 wk 0 3 (3.9)
Marital status > 19 wk 0 3 (3.9)
on 04/15/2023
Married 70 (90.9) I do not know 52 (67.5) —

Partnered/not married 3 (3.9) Missing 1 (1.3) 2 (2.6)
Single 4 (5.2) Mean (SD); 3.3 (4.5) 7.7 (5.3)
Partner employed (n = 73) weeks
Yes 47 (64.4) Adoption leave
No 24 (32.9) < 1 wk 5 (6.5) 2 (2.6)
No response 2 (2.7) 1-6 wk 4 (5.2) 27 (35.6)
Partner full or part time employed (n = 47) 7-12 wk 2 (2.6) 36 (46.8)
Full time 38 (80.9) 13-18 wk 0 4 (5.2)
Part time 9 (19.1) > 19 wk 0 5 (6.5)
No. surgeons in practice I do not know 66 (85.7) —
0-3 11 (14.3) Missing 0 3 (3.9)
4-6 16 (20.8) Mean (SD); 4.2 (4.7) 9.6 (5.8)
7-15 40 (51.9) weeks
> 16 9 (11.7) Payment for leave
No response 1 (1.3) Fully paid 17 (22.1) 47 (61.0)
Years in practice Partially paid 11 (14.3) 24 (31.2)
<3 13 (16.88) Unpaid 12 (15.6) 5 (6.5)
3-5 13 (16.88) I do not know 37 (48.1) —
6-10 16 (20.78) Missing 0 1 (1.3)
11-20 16 (20.78)
> 21 19 (24.68)
Mean (SD); years 13.4 (10.9) participants, 37.7% believed that surgeons should still pay
Trauma call overhead costs while on leave. Most participants (75.3%)
Yes 70 (90.9) did not agree that call responsibilities had to be paid back
No 7 (9.1)
Compensation model
or paid forward when taking parental leave.
Salary 24 (31.17)
Salary + bonus based on productivity 37 (48.05) Surgeon Experiences With Parental Leave
Compensation solely based on productivity 16 (20.78) Of 30 participants who had children within the past
10 years, 16 took 1 to 6 weeks of parental leave (Table 4).
Perceptions About Parental Leave Fourteen of those 16 surgeons were men. Most women
Most believed that at least 7 to 12 weeks should be took 7 to 12 weeks for parental leave. Of the surgeons,
offered for maternity leave (66.2%), that 1 to 6 weeks 53.3% who had children within the last 10 years took more
should be offered for paternity leave (54.6%), and that 7 to calls before and/or after their parental leave to compensate
12 weeks should be offered for adoption leave (46.8%) for missed call responsibilities. Most participants who had
(Table 2). More surgeons believed parental leave should children within the past 10 years reported no negative
be fully paid (61.0%), compared with 31.2% who experiences related to taking parental leave.
supported partial compensation. Almost half of the
survey, participants disagreed that colleagues should DISCUSSION
minimize their leave to decrease the burden on other The lack of established and publicized parental
surgeons (Table 3). Most participants reported that their leave policies within orthopaedics perpetuates a culture
work environments were supportive or somewhat that penalizes parents who may request time off from
supportive to surgeons taking parental leave. Of the work.17,18 Many practicing surgeons may be unaware of

The results of this study revealed that most pediatric

TABLE 3. Perceptions of Pediatric Orthopaedic Surgeon
Respondents About Parental Leave orthopaedic surgeons who completed the survey were not
familiar with the parental leave policies available to them.
n (%)
It is difficult to determine whether this finding is due to a
Surgeons should minimize parental leave to reduce colleague burden lack of transparency of the policies that employers pro-
Strongly agree 6 (7.8) vide, or if surgeons are not interested in knowing existing
Agree 8 (10.4)
Neither agree nor disagree 23 (29.8) policies unless they need to use them. Approximately half
Disagree 22 (28.6) of the participants who were unaware of the policies have
Strongly disagree 18 (23.4) been in practice for more than 11 years, which may in-
How would you describe your work environment as it pertains to dicate that their family-building efforts had been com-
parental leave?
Supportive 31 (40.3)
pleted at the time of this survey. Furthermore, most of the
Somewhat supportive 27 (35.1) participants who were unaware of policies were men,
Somewhat unsupportive 5 (6.5) which could be due to parental leave policies having been
Unsupportive 10 (13.0) historically perceived as pertaining to women, with little
I do not know 3 (3.9) impact on the career of men surgeons.1,19,20
Missing 1 (1.3)
Parents taking leave should still pay overhead costs while on leave
Strongly agree 7 (9.1) Maternity Leave
Agree 22 (28.6) Less than half of pediatric orthopaedic surgeons
Neither agree nor disagree 32 (41.6) familiar with leave policies report 7 to 12 weeks were of-
Disagree 12 (15.6)
on 04/15/2023
Strongly disagree 4 (5.2)

fered for maternity leave by their employers. This was
On-call responsibilities covered by colleagues should be “paid forward” more than the average 4.6 weeks offered to women or-
or “paid back” thopaedic surgeons from all subspecialties, reported by
Yes 19 (24.7) Nguyen et al.4 Nguyen et al4 also reported that women
No 58 (75.3) orthopaedic surgeons from all subspecialties took an
average of 8 weeks for maternity leave, which is com-
parable to the 7 to 12 weeks taken by more than 80% of
parental leave policies at their places of employment. women pediatric orthopaedic surgeons in this study. These
Some surgeons also hesitate to take leave due to per- data suggest that pediatric orthopaedic surgeons may be
ceptions of decreased commitment to their careers or offered longer maternity leave than women in other or-
feelings of guilt about increased colleague workload.18 It thopaedic subspecialties. Although the number of weeks
is evident that more work needs to be done to decrease of parental leave taken by women pediatric orthopaedic
stress on new parents, decrease burnout, and improve job surgeons are comparable to what is offered, most of the
satisfaction.7,17,18 women in our study believe more maternity leave should
be offered by employers. This finding suggests that per-
haps surgeons returned to work sooner than they should
TABLE 4. Experiences With Parental Leave for Pediatric have, either due to self-imposed or collegial pressures to
Orthopaedic Surgeon Respondents Who Have Had a Child do so. However, most surgeons reported that their work
While Working as an Orthopaedic Surgeon Within the Past 10 environment was supportive of the time taken for parental
Years (n = 30) leave and that the amount of time they took depended on
n (%) their own conscious efforts not to burden their colleagues.
Weeks of parental leave taken
< 1 wk 3 (10.0) Paternity Leave
1-6 wk 16 (53.3) Most men pediatric orthopaedic surgeons reported
7-12 wk 10 (33.3) taking 1 to 6 weeks for paternity leave, which was more
13-18 wk 0
> 19 wk 0
than the average <1 week offered by employers. This
Missing 1 (3.3) finding is comparable to the average number of weeks
Mean (SD); weeks 4.6 (4.4) taken by fathers in the general population.21 Many men
Call experience pediatric orthopaedic surgeons, some women surgeons,
Fully compensated for missed call 16 (53.3) men within other surgical specialties, and nonmedical
Partially compensated for missed call 4 (13.3)
Did not take on extra call to compensate 10 (33.3) professions believe that more time should be offered to
Experiences on parental leave (respondents could choose > 1) men who request paternity leave.11,17,21 Adequate parental
Stalled or delayed professional advancement 4 (13.3) leave is not just a concern for women surgeons. Though
Lower compensation or slower increase in compensation 2 (6.7) most surgeons describe supportive colleagues, some men
compared with peers
Pressure to minimize or shorten leave time 6 (20.0)
surgeons report feeling pressure from peers to minimize
Concerns about employment security 2 (6.7) leave time because they are men. It is imperative that men
Negative questions/comments about your profession 1 (3.3) surgeons feel comfortable taking the time needed to sup-
commitment port their families. The ability to bond with their child and
Negative questions/comments about your parenthood 0 assist with childcare will allow them to be mentally present
No, none of these apply 18 (60.0)
when returning to work, improve physician wellness, and

J Pediatr Orthop Volume 43, Number 5, May/June 2023 Parental Leave Policies in Pediatric Orthopaedics
ultimately improve patient outcomes.7,14,22 In addition, if more should be done to support men who desire parental
men are encouraged to take more parental leave, this leave.17,18
would promote equity within the workforce by normaliz-
ing the practice of parental leave by all providers. In our Limitations
opinion, if parental leave practices by men were more This study is limited by the number of survey re-
expansive and/or uniform, there would be a decreased sponses we received. The survey was distributed to the
“motherhood penalty” faced by women face when taking POSNA membership, but the 6.5% response rate is sig-
parental leave.23 nificantly lower than the 42% average response rate for a
POSNA survey.27 There has been a trend toward de-
Adoption Leave clining survey response rates from the POSNA mem-
Most women surgeons who completed this survey bership in recent years,27 with recent publications
were 49 years old or younger. However, surgeons 50 years reporting response rates between 9.7% and 21.3%.28–31 In
and older with more than 10 years in practice believe that addition, surgeons who opted to respond to the survey
more parental leave should be provided for surgeons who may have increased interest in parental leave policies due
would like to adopt children. The decision to adopt chil- to personal positive or negative experiences. Thus, re-
dren is not age-specific, but surgeons who delayed child- sponses are likely skewed to represent the perceptions of
bearing or family expansion to advance their careers these respondents and may not accurately represent the
experience increased complications and infertility with attitudes of most pediatric orthopaedic surgeons. In ad-
time.3,6 Thus, sufficient adoption leave is also a necessity dition, approximately one-third of respondents in the
on 04/15/2023
for surgeons who may pursue this alternate route of family sample were in the childbearing age range and were
building. partnered, which may suggest increased interest in family
planning and more knowledge of employer parental
leave policies than would be typical for pediatric ortho-
Parental Leave Pay paedic surgeons. Although the survey was anonymous,
More than 90% of pediatric orthopaedic surgeons respondents may have answered questions based on
who responded believe that parental leave should be fully perceived acceptable societal norms and not personal
or partially paid. Most of these were surgeons 49 years old sentiments. It is difficult to say how much this affected
or younger, and more than half were men. Surgeons, like the results, despite efforts to reduce response bias. Of the
other new parents, have many financial responsibilities 20 participants with incomplete data, 11 only answered
related to childbirth, childcare, adoption, or assisted re- the eligibility questions and 8 answered the first 3 survey
productive treatments when expanding their families. questions on employer leave policies. We hypothesize
These responsibilities are compounded by overhead costs that these participants may have been discouraged by the
related to maintaining a practice, and the loss of income length of the IRB-required information statement at the
proportionate to the length of time taken for parental start of the survey, or decided not to finish the survey due
leave.4,22 It is a challenge to achieve standardization to their lack of knowledge of employer leave policies.
of policies that promote adequate pay for surgeons on One strength of this study is that a high proportion
parental leave due to the varying structures of practice of respondents were men (59.7%). This strengthens the
environments, but one worth addressing to improve the argument that both men and women are concerned about
well-being of the pediatric orthopaedic workforce. policies surrounding parental leave.
The Federal Family and Medical Leave Act entitle
eligible employees to 12 weeks of unpaid leave, but there
are no clearly defined policies in medicine regarding paid CONCLUSION
benefits.24–26 Itum et al24 reported only 53% of top aca- With a generational shift toward improved work-life
demic medical institutions have any defined policies about balance, new parents desire adequate parental leave time
paid parental leave within their surgical subspecialties. and sufficient pay to support their families, regardless of
Parental leave can be expensive for academic and private their sex. Pediatric orthopaedic surgeons report receiving
practice physicians who still have professional and per- more time for parental leave than orthopaedic surgeons
sonal financial responsibilities. A survey by Nguyen et al4 from other subspecialties, but they still believe more can
found that practicing women orthopaedic surgeons were be done to improve parental leave policies. Established
offered 4.6 ± 4.2 weeks of paid leave but took an average and adequate parental leave policies will be an integral
of 8.2 ± 7.4 weeks of total leave time. The discrepancy part of ensuring the well-being of the physician workforce
between the paid time offered and the amount of leave and the patients they treat.
time taken highlights the need for more compensation due
to the increased financial burden during this critical time. ACKNOWLEDGMENTS
It also suggests that 4 to 6 weeks generally offered for The authors thank the Pediatric Orthopaedic Society
parental leave is inadequate for postpartum recovery and of North America (POSNA) for distributing our survey to
parent-child bonding.4,24 Some employers provide little to the POSNA membership, survey respondents for partic-
no unpaid or paid time off for paternity leave, and there is ipating and making this study possible, and PDA statistics
evidence to support that both men and women believe that for statistical consultation.

REFERENCES 17. MacVane CZ, Fix ML, Strout TD, et al. Congratulations, you’re
1. Blair JE, Mayer AP, Caubet SL, et al. Pregnancy and parental leave pregnant! now about your shifts: the state of maternity leave attitudes
during graduate medical education. Acad Med. 2016;91:972–978. and culture in EM. West J Emerg Med. 2017;18:800–810.
2. Stack SW, Jagsi R, Biermann JS, et al. Childbearing decisions in 18. Gaines T, Harkhani N, Chen H, et al. Current policies and
residency: a multicenter survey of female residents. Acad Med. practicing surgeon perspectives on parental leave. Am J Surg.
2020;95:1550–1557. 2019;218:772–779.
3. Willett LL, Wellons MF, Hartig JR, et al. Do women residents delay 19. Association of American Medical Colleges. 2020 Physician Specialty
childbearing due to perceived career threats? Acad Med. 2010;85: Report Data Highlights; 2020. Accessed: 6 October 2022. https://
640–646. www.aamc.org/data-reports/workforce/interactive-data/2020-physicia
4. Nguyen CV, Luong M, Weiss JM, et al. The cost of maternity leave n-specialty-report-data-highlights.
for the orthopaedic surgeon. J Am Acad Orthop Surg. 2020;28: 20. United States Department of Labor. DOL POLICY BRIEF
e1001–e1005. Paternity Leave: Why Parental Leave For Fathers Is So Important
5. Poon S, Kiridly D, Mutawakkil M, et al. Current trends in sex, race, For Working Families; 2015. Accessed: 6 October 2022. https://www.
and ethnic diversity in orthopaedic surgery residency. J Am Acad dol.gov/sites/dolgov/files/OASP/Paternity-Leave.pdf.
Orthop Surg. 2019;27:e725–e733. 21. Nepomnyaschy LWJ. Paternity leave and fathers’ involvement with
6. Hamilton AR, Tyson MD, Braga JA, et al. Childbearing and their young children: evidence from the American Ecls-B. Commun-
pregnancy characteristics of female orthopaedic surgeons. J Bone ity Work Fam. 2007;10:427–453.
Joint Surg Am. 2012;94:e77. 22. Scully RE, Davids JS, Melnitchouk N. Impact of procedural
7. Howell LP, Beckett LA, Nettiksimmons J, et al. Generational and specialty on maternity leave and career satisfaction among female
gender perspectives on career flexibility: ensuring the faculty work- physicians. Ann Surg. 2017;266:210–217.
force of the future. Am J Med. 2012;125:719–728. 23. Zalis S. Men Should Take Parental Leave- Here’s Why. Forbes
8. Jou J, Kozhimannil KB, Abraham JM, et al. Paid maternity leave in Magazine; 2018. Accessed: 6 October 2022. https://www.forbes.com/
the United States: associations with maternal and infant health. sites/shelleyzalis/2018/05/03/why-mandatory-parental-leave-is-good-
on 04/15/2023
Matern Child Health J. 2018;22:216–225. for-business/

9. Dagher RK, McGovern PM, Dowd BE. Maternity leave duration 24. Itum DS, Oltmann SC, Choti MA, et al. Access to paid parental
and postpartum mental and physical health: implications for leave leave for academic surgeons. J Surg Res. 2019;233:144–148.
policies. J Health Polit Policy Law. 2014;39:369–416. 25. Gupta M, Reichl A, Diaz-Aguilar LD, et al. Pregnancy and parental
10. Heymann J, Raub A, Earle A. Creating and using new data sources leave among neurosurgeons and neurosurgical trainees. J Neurosurg.
to analyze the relationship between social policy and global health: 2020;134:1325–1333.
the case of maternal leave. Public Health Rep. 2011;126(suppl 3): 26. United States Department of Labor WaHD. Family and Medical
127–134. Leave Act. Accessed: 6 October 2022. https://www.dol.gov/agencies/
11. Huerta MC, Adema W, Baxter J, et al. Fathers’ leave and fathers’ whd/fmla
involvement: evidence from four OECD countries. Eur J Soc Secur. 27. Williams BA, Guerrero A, Blakemore LC, et al. Surveying the
2014;16:308–346. POSNA landscape: what can we learn from society survey studies? J
12. Van Niel MS, Bhatia R, Riano NS, et al. The impact of paid Pediatr Orthop. 2020;40:e63–e67.
maternity leave on the mental and physical health of mothers and 28. Beck JJ, Carpenter CM, West N, et al. Treatment variability and
children: a review of the literature and policy Implications. Harv Rev complications associated with pediatric lateral ankle injuries: a
Psychiatry. 2020;28:113–126. POSNA quality, safety, and value initiative survey. Orthop J Sports
13. Hajizadeh M, Heymann J, Strumpf E, et al. Paid maternity leave and Med. 2022;10:23259671221100223.
childhood vaccination uptake: longitudinal evidence from 20 low- 29. Johnson MA, Ganley TJ, Crawford L, et al. Pediatric orthopedic
and-middle-income countries. Soc Sci Med. 2015;140:104–117. trauma care during the COVID-19 pandemic: a survey of the
14. Goldberg MJ, Shea KG, Weiss JM, et al. The Pediatric Orthopaedic Pediatric Orthopedic Society of North America. HSS J. 2022;18:
Society of North America (POSNA) adopts a member health and 205–211.
wellness charter. J Pediatr Orthop. 2019;39:e241–e244. 30. Lindsay SE, Holmes S, Swarup I, et al. Trends in management of
15. Harris PA, Taylor R, Thielke R, et al. Research electronic data pediatric distal radius buckle fractures. J Pediatr Orthop. 2022;42:
capture (REDCap)–a metadata-driven methodology and workflow 367–371.
process for providing translational research informatics support. 31. Ho D, Du JY, Erkilinc M, et al. Getting them back in the game:
J Biomed Inform. 2009;42:377–381. when can athletes with adolescent idiopathic scoliosis safely return
16. Harris PA, Taylor R, Minor BL, et al. The REDCap consortium: to sports? A mixed-effects study of the Pediatric Orthopaedic
building an international community of software platform partners. Association of North America. J Pediatr Orthop. 2021;41:
J Biomed Inform. 2019;95:103208. e717–e721.

ORIGINAL ARTICLE
Predicting Deformity Correction of Growth Modulation

in Late-onset Tibia Vara
Janet L. Walker, MD,* David M. Dueber, PhD,† Kenneth P. Powell, MD,‡

Lindsay P. Stephenson, MD,§ Allison C. Scott, MD,§ Joel A. Lerman, MD,∥
Sarah B. Nossov, MD,¶ Corinna C. Franklin, MD,¶ and David E. Westberry, MD#
success for final-mTFA with GMS by 82%, controlling for pre-

Background: Growth modulation in late-onset tibia vara (LOTV) operative mTFA. Age, sex, race/ethnicity, type of implant, and
has been reported to yield variable results. We hypothesized that knee center peak value adjusted age (a method for bone age)
parameters of deformity severity, skeletal maturity, and body were not predictive of outcome.
weight could predict the odds of a successful outcome. Conclusions: Resolution of varus alignment in LOTV using first
Methods: A retrospective review of tension band growth modu- LTTBP and GMS, as quantified by MPTA and mTFA, re-
on 04/15/2023
lation for LOTV (onset ≥ 8 y) was performed at 7 centers. Tibial/ spectively, is negatively impacted by deformity magnitude, hip
overall limb deformity and hip/knee physeal maturity were as- physeal closure, and/or body weight ≥ 100 kg. The presented
sessed on preoperative anteroposterior standing lower-extremity table, utilizing these variables, is helpful in the prediction of the
digital radiographs. Tibial deformity change with first-time lat- outcome of the first LTTBP and GMS. Even if complete cor-
eral tibial tension band plating (first LTTBP) was assessed by rection is not predicted, growth modulation may still be appro-
medial proximal tibia angle (MPTA). Effects of a growth mod- priate to reduce deformity in high-risk patients.
ulation series (GMS) on overall limb alignment were assessed by Level of Evidence: Level III.
mechanical tibiofemoral angle (mTFA) and included changes
from implant removal, revision, reimplantation, subsequent Key Words: adolescent tibia vara, Blount disease, genu varum,
growth, and femoral procedures during the study period. The guided growth, children
successful outcome was defined as radiographic resolution of (J Pediatr Orthop 2023;43:e343–e349)
varus deformity or valgus overcorrection. Patient demographics,
characteristics, maturity, deformity, and implant selections were
assessed as outcome predictors using multiple logistic regression.
Results: Fifty-four patients (76 limbs) had 84 LTTBP procedures
and 29 femoral tension band procedures. For each 1-degree de- L ate-onset tibia vara (LOTV) results from abnormal
growth of the medial proximal tibial physis. It occurs
with greater frequency in males and black/African
crease in preoperative MPTA or 1-degree increase in pre-
operative mTFA the odds of their successful correction decreased American individuals. The deformity may be related to
by 26% in the first LTTBP and 6% by GMS, respectively, con- increased mechanical stress, as it is often associated with
trolling for maturity. The change in odds of success for GMS obesity.1 In addition to proximal tibial varus, there is often
assessed by mTFA was similar when controlling for weight. deformity of the distal femur, further complicating the
Closure of a proximal femoral physis decreased the odds of overall limb alignment.2 If not corrected, lower limb
success for postoperative-MPTA by 91% with first LTTBP and mal-alignment results in abnormal forces across the knee.
for final-mTFA by 90% with GMS, controlling for preoperative Hemiepiphyseodesis is commonly used to treat lower
deformity. Preoperative weight ≥ 100 kg decreased the odds of limb malalignment.3 The reported success of hemi-
epiphyseodesis for the management of tibia vara through
“growth modulation” has been disappointing when com-
From the *Shriners Children’s and University of Kentucky Department pared with other diagnoses.4–6 The literature cites age,
of Orthopaedic Surgery and Sports Medicine; †University of Ken-
tucky Department of Statistics, Lexington KY; ‡Shriners Children’s,
severity of deformity, and body mass index (BMI) as
Shreveport LA; §Shriners Children’s, Galveston TX; ∥Shriners predictors of poor results of hemiepiphyseodesis in
Children’s, Sacramento CA; ¶Shriners Children’s, Philadelphia and LOTV.6–10 However, these conclusions are predicated
Erie PA; and #Shriners Children’s, Greenville SC. on studies, which frequently have small sample sizes,
The authors declare no funding for this work.
heterogeneous groups of patients (early and late-onset)
Reprints: Janet L. Walker, MD, Shriners Hospital for Children, 110 managed with varying techniques, or results from meta-
Conn Terrace, Lexington, KY 40508. E-mail: jwalker@shrinenet.org analyses.6,9 As such, they may be underpowered to
Supplemental Digital Content is available for this article. Direct URL determine the relative risk of these variables.
citations appear in the printed text and are provided in the HTML Tension band plating, as a method for temporary
pedorthopaedics.com. hemiepiphyseodesis11 has gained wide popularity due to
Copyright © 2023 Wolters Kluwer Health, Inc. All rights reserved. its technical ease, potential reversibility, and low compli-
DOI: 10.1097/BPO.0000000000002373 cation rate. The technique of tension band plating in

LOTV has been used at our 7 centers. The purpose of this some children classified as “juvenile” onset by others.
study was to report our results and determine, which pa- After applying exclusionary criteria, 70 patients with
rameters could predict the ability of growth modulation tibia vara, onset at age 8 years or older, with 94 affected
to correct deformity in LOTV. We assessed outcomes of limbs were included for short-term complications.
first-time lateral tibial tension band plating (first LTTBP), Patient characteristics recorded included sex (as
using the medial proximal tibial angle (MPTA). As com- assigned at birth), age at diagnosis and surgery, self-
prehensive limb growth modulation is often a sequence reported race/ethnicity, preoperative body weight (kg),
of procedures, we, in addition, assessed the outcome BMI, and BMI percentile. Surgical parameters recorded
of the overall growth modulation series (GMS) using included location, implant characteristics, complications,
measures of overall limb alignment with the mechanical and follow-up. Minimum follow-up for deformity
tibiofemoral angle (mTFA), mechanical axis deviation correction prediction analysis was accepted at 2 years
(MAD), and mechanical axis zone (MAZone). unless deformity correction, subsequent surgery, or
skeletal maturity occurred. As such, an additional 16
METHODS patients were excluded from this portion of the study.
A retrospective review was performed at 7 tertiary All implants were manufactured by Orthofix
pediatric orthopaedic centers, and was reviewed by an Medical, Inc. (Guided Growth System, Lewisville TX) or
Institutional Review Board for all participating in- OrthoPediatrics (PediPlates, Warsaw IN). The plates were
stitutions. One hundred forty-eight patients with tibia placed extraperiosteally, spanning the lateral tibial physis,
vara having tension band plating from June 2008 to June without other intervention. There was no specified treat-
on 04/15/2023
2018 were identified through a review of the hospital ment protocol during the study period. All patients were
coding systems. Figure 1 depicts the study flow diagram managed at the discretion of their individual surgeons.
of the patient selection identified by coding queries. Treatment outcome was defined by radiographic
Based on the bimodal age distribution of our patient change. Pre and postsurgical digital standing full-length
population with tibia vara, we defined those children anteroposterior lower-extremity radiographs, with patella
with tibia vara, age at diagnosis 8 years or older, as late- forward, were analyzed by a single investigator (JLW)
onset. This also excluded limbs with the medial with 36 years of pediatric orthopaedic experience to
epiphyseal/plateau depression complication common in minimize measurement variability. Measurements were
those of earlier onset. Our LOTV includes “adolescent” performed using McKesson Radiology Station 12.3 64-bit
tibia vara described in some studies but also includes 1989-2019 (Irvine, TX). Radiographic measurements in-
cluded: MPTA, mechanical lateral distal femoral angle
(mLDFA), mTFA, MAD,12 and MAZone.3 The ob-
server’s reliability was tested while blinded to treatment
outcome with repeated assessment of 29 limb radiographs
of patients with LOTV at 2 time points. The intraclass
coefficients were all > 0.95 and kappa = 0.897, indicating
excellent agreement.
Skeletal maturity assigned on each radiograph was
determined through the closure of the hip physes. Using
the anteroposterior films of the knee, a previously reported
age-adjustment method was applied, using the height of
the distal femoral physeal peak compared with its width.13
This has been shown to correlate with skeletal maturity.
The observer’s repeatability was excellent for hip physeal
closure and knee age-adjustment method with a kappa
value = 0.935 and an intraclass correlation = 0.982, re-
spectively. Years of remaining growth were also calcu-
lated, assuming growth cessation at 17 years for boys and
15 years for girls.5
Age, sex, race/ethnicity, weight parameters, severity
of deformity, bone maturity measures, and implant co-
variates were assessed as predictors of outcome. To dem-
onstrate the effectiveness of growth modulation, success
was defined as the correction of radiographic parameters to
the normal adult range or overcorrection into valgus. The
success of the first LTTBP (MPTA) and the success of GMS
(mTFA, MAD, and MAZone) were analyzed separately.
The normal adult values accepted were: MPTA = 85 de-
grees to 90 degrees, MAD = 3 to 17 mm medial, mTFA =
FIGURE 1. Study flow chart for subject selection. 0 degrees to 3 degrees varus,12 and MAZone = I (+/−).3

J Pediatr Orthop Volume 43, Number 5, May/June 2023 Deformity Correction of Growth Modulation in LOTV
The outcome of the first LTTBP was assessed from the last RESULTS
radiograph, before implant removal/revision or other sur- Seventy patients, with 94 limbs, satisfied the inclusion
gery, or at the last visit, for those not having further surgery. criteria and had 103 LTTBPs. Short-term perioperative
The final end-point for GMS outcome was defined as the complications included 3 with wound dehiscence/delayed
last radiograph, before any bony surgery, or at the last visit, healing, 1 prolonged drainage, and 1 incisional site in-
if no further surgery was performed. A GMS could en- fection. One patient required a 3-day hospital stay for
compass the effects of primary and revision tibial implants, postoperative nausea. In the first postoperative year, 1 pa-
femoral implants, implant removal, rebound growth after tient died and 1 patient had a stroke, unrelated to LOTV.
physeal untethering, and reimplantation to treat recurrence, An additional 14 patients were excluded from further
all occurring within the same limb throughout the study deformity correction analysis due to insufficient follow-up
period. or radiographic documentation.
The association of patient characteristics, pre- Fifty-four patients with 76 limbs having 84 LTTBPs
operative deformity, maturity measures, and surgical and 29 femoral tension band procedures were reviewed for
covariates with the successful outcome of the first deformity correction analysis. These are summarized in
LTTBP and overall GMS were assessed using univariate Table 1. All were included in the GMS/overall limb
logistic regression. Several patients had 2 limbs included analysis. To study the effect of only the first LTTBP on
in the analysis, as such, their limbs are not independent MPTA we excluded 7 tibias that had had LTTBPs before
observations. This failure of independence can bias es- the study, 6 revision LTTBPs, and one LTTBP
timated standard errors and, thereby, significance tests reimplanted after recurrence following removal.
For the first LTTBP, follow-up duration (minimum
on 04/15/2023
for parameter estimates. The robust sandwich estimator

accounts for correlations within subjects, resulting in 2 y unless deformity correction, additional surgery, or
unbiased estimated standard errors.13 Logistic regression skeletal maturity) was mean = 2.2 years (range: 0.46 to 7.0
models were estimated using the “glm” function in the R y). For GMS the follow-up was mean = 2.7 years (range:
statistical computing environment and the robust sand- 0.56 to 7.0 y). Follow-up for GMS includes changes related
wich estimator was applied using the sandwich to femoral implants (n = 29, 28 for varus, and 1 for valgus),
package.14 Variables showing significance (P < 0.05) in tibial implant revisions (n = 7), removals (n = 50, 18
the univariate analysis were assessed through multi- femoral, and 32 tibial), reimplantation (n = 1) and any
variate logistic regression models. Stepwise model se- subsequent rebound growth during the study period.
lection procedures were used to assist in constructing the Seven broken tibial implants in 4 patients were revised.
final predictive models. In this stepwise procedure, mul- Six revisions were secondary to a broken metaphyseal screw.
tivariate models were assessed based on sensitivity, One revision was for a broken plate along with one of 2
specificity, overall accuracy, positive predictive value, metaphyseal screws. Eight additional broken implants were
and negative predictive value. found. Three were retained and 5 were routinely removed.
TABLE 1. Deformity Correction Analysis Population Summary

Patients (n = 54)
Sex (at birth) 40 males, 14 females
Race/ethnicity (self-reported) 31 black/non-Hispanic, 13 white/non-Hispanic, 9 white/Hispanic, 1 other
Age at diagnosis mean = 11 y, range = 8-15 y
Late follow-up 11 with > 6 mo follow-up intervals
Limbs (n = 76)
Side 43 left, 33 right
Preoperative age mean = 12.0 y, range = 8.3-15.5 y
Preoperative body weight mean = 89.5 kg, range = 39.6-149 kg
Preoperative BMI mean = 37 kg/m2, range = 17-58 kg/m2
Preoperative BMI percentile mean = 99.2 percentile, 75% over the 99th percentile
Preoperative age by femoral center peak adjustment mean = 12.0 y, range = 8.6-14.7 y
Preoperative years of growth remaining mean = 4.5 y, range = 1.0-8.7 y
Hip physes 23 triradiate open
33 triradiate closed, proximal femur open
19 any closure in proximal femur
First-time, single-event tibial implants (n = 70)

Implant manufacturer 49 Orthopediatrics, 19 Orthofix, 2 not recorded
Metal composition 52 stainless steel, 16 titanium, 2 not recorded
Cannulated vs solid screws 48 cannulated, 20 solid, 2 unclear
Screw diameter 60: 4.5 mm, 7: 3.5 mm, 1 not recorded
No. screws (4 vs 2) 35 = 4-screws, 35 = 2-screws
BMI indicates body mass index.

Of the initial 70 patients who had LTTBP, 22 pa- The results of univariate analyses are shown in Sup-
tients with open perigenu physes (including 11 originally plemental Table 1 (Supplemental Digital Content 1, http://
excluded for insufficient follow-up) were lost to follow- links.lww.com/BPO/A582). Deformity severity and closure
up, 15 with implants in place, and 7 with implants re- of hip physes were predictive of outcome for both first
moved. One limb required LTTBP reimplantation LTTBP and GMS. In addition, variables of weight were
for recurrence after implant removal. Another limb predictive of GMS outcome. BMI percentile was too
had progressive valgus despite implant removal after skewed in this population and, therefore, was not used in
overcorrection. analyses. Treatment success was not predicted by sex, race/
The radiographic outcomes are presented in Table 2. ethnicity, age at diagnosis or surgery, years of growth re-
The mean MPTA at the treatment end was 84.9 degrees after maining, the knee center peak value adjusted age,13 implant
the first LTTBP. This resulted in mean change in the MPTA characteristics, or additional femoral hemiepiphyseodesis.
of +7.5 degrees with a correction rate of +0.33 degrees/ Table 3 presents the multivariate analyses. Preopera-
month of implantation. Subsequent LTTBPs (n = 14) tive mTFA was selected as the measure of overall limb
resulted in a mean correction of +6.0 at a rate of +0.13 alignment because it was most predictive of successful
degrees/month. The overall GMS outcomes are shown in GMS outcome, outperforming preoperative-MAD or
Table 2. Forty-one implants in 30 patients were removed or preoperative-MAZone. Preoperative body weight was the
revised at a mean = 2.0 years postinsertion and at mean age more predictive weight variable compared to BMI.
= 14.0 years. Limbs that had femoral tension band Interpreting the odds ratios using the models, controlling
procedures (n = 29) had a mean preoperative-mLDFA = for maturity, each 1-degree decrease in preoperative MPTA
94.7 degrees correcting to a mean = 88.6 degrees at the study
on 04/15/2023
or 1-degree increase in preoperative mTFA, the odds of

end. Those limbs that did not have femoral implants (n = their successful correction decreased by 26% for the first
47) had a mean preoperative-mLDFA = 91.7 degrees LTTBP and 6% with GMS, respectively. The odds of
correcting to mean = 91.2 degrees at the study end. One success for mTFA with GMS were similar when controlling
lateral femoral tension band plate required medial for the weight (Model 3). Controlling for preoperative
femoral tension band plating for overcorrection after a late MPTA or preoperative mTFA (first models), any proximal
follow-up. femoral physeal closure decreased the odds of achieving
With GMS, MAD was successfully corrected, ≤ 17 success with MPTA and mTFA by 91% by first LTTBP and
mm medial, in 47% of limbs whereas mTFA was only 90% with GMS, respectively. Controlling for preoperative
corrected in 40%. At the final GMS follow-up, 48 limbs MPTA, closure of the triradiate physis alone was associated
had no remaining proximal tibial growth, 43 with physeal with a negligible decrease in odds of success for the
maturity, and 5 with proximal tibial epiphysiodesis. Four first LTTBP. In our mTFA analysis for GMS, triradiate
limbs underwent valgus tibial osteotomy and 2 limbs re- physeal closure was both nonsignificant and theoretically
quired derotational osteotomy. All 3 limbs in zone II nonsensical, therefore, a separate analysis of triradiate
valgus at the GMS end, still had open physes without physeal closure was not included. Nineteen limbs, with
implants and the potential for rebound correction. The full some physeal closure in the proximal femur, had mean
data set is available from the corresponding author with deformities of MPTA = 74 degrees and mTFA = 21 degrees,
an approved data share agreement. and successful correction rates of 15% (first LTTBP) and
TABLE 2. Radiographic Outcome of Growth Modulation in LOTV (54 Patients, 76 Limbs)

Mean Mean Mean Median Number
Preoperative Treatment end Correction Correction Achieving success* n (%)
First-time, Single-event lateral tibial tension band plates (n = 70)
Medial proximal tibial angle (degrees) 77.4 85.0 +7.5 +8.0 38 (54)
Overall GMS (n = 76 Limbs) Preoperative Study end

Medial proximal tibial angle (degrees) 76.1 83.5 +6.9 +8.2 40 (53)
mTFA (degrees varus) 18.4 8.0 −10.4 −12.0 30 (40)
MAD (millimeters varus) 63.8 29.9 −34.5 −40.5 36 (47)
MAZone (n) Preoperative Study end 37 (49)
II valgus 3
I valgus 9
I varus 25
II varus 14 16
III varus 62 23
Mature/tibial epiphyseodesis (n) 48
*Success = correction or overcorrection, relative to normal adult limb parameters.
GMS indicates growth modulation series; LOTV, late-onset tibia vara; MAD, mechanical axis deviation; MAZone, mechanical axis zone; mTFA, mechanical
tibio-femoral angle.

TABLE 3. Results of Multivariate Logistic Regression Analysis

Variable Odds ratio 95% CI P
*Tibial alignment success, as determined by MPTA with first-time single-event first LTTBP
Model 1
Preoperative MPTA 1.26 1.12, 1.45 < 0.001

Preoperative physeal maturity

Triradiate closed 0.57 0.14, 2.13 0.409
Any proximal femur closed 0.09 0.01, 0.48 0.009
Overall accuracy Sensitivity Specificity Positive predictive value Negative predictive value
Model 1 0.77 0.82 0.71 0.78 0.76
*Overall limb alignment success, as determined by mTFA with GMS

Model 1
Preoperative mTFA 0.94 0.87, 0.99 0.048
Preoperative physeal maturity
Any proximal femur closed 0.1 0.01, 0.41 0.004
Model 2
Weight, per 1 kg increase 0.63 0.36, 1.05 0.082
Model 3
on 04/15/2023
Weight ( > 100 kg) 0.18 0.04, 0.58 0.007
Overall accuracy Sensitivity Specificity Positive predictive value Negative predictive value
Model 1 0.71 0.73 0.69 0.65 0.76
Model 2 0.69 0.67 0.70 0.68 0.71
Model 3 0.68 0.67 0.70 0.63 0.73
*As the models for MPTA and mTFA success refer to different procedures (first LTTBP and GMS), they should not be compared with each other.
CI indicates confidence interval; GMS, growth modulation series; LTTBP, lateral tibial tension band plating; MPTA, medial proximal tibia angle; mTFA, mechanical
tibio-femoral angle.
5% (GMS), respectively. Weight ≥ 100 kg also decreased found preoperative mTFA, was the most predictive limb
the odds for successful correction of mTFA by 82% with alignment parameter for GMS, MAD is frequently
GMS, controlling for preoperative mTFA. reported for the outcome of LOTV studies.4,7,10 These
Prediction accuracy for the multivariate models is statistical analyses, using MAD as an outcome do not
shown in Table 3, whereas Table 4 indicates the threshold change our conclusions.
values of preoperative MPTA for the first LTTBP and
preoperative mTFA in GMS required for specified
predicted probabilities of success. These threshold values DISCUSSION
indicate that, with any proximal femoral physeal closure, Growth modulation in our study of LOTV patients
there is only a 40% chance of correcting the MPTA from had a 6% complication rate in the perioperative period.
82.2 degrees to ≥ 85 degrees with a first LTTBP procedure. First-time LTTBP achieved successful correction of MPTA
There were insufficient numbers of more mature limbs with in 54% of tibias. The rate of tibial deformity correction was
MPTA > 82.2 degrees (< 3 degrees away from success) lower for subsequent revisions for implant failure or re-
having first LTTBP to calculate thresholds for higher odds implantation. Complete correction of the mTFA, MAD,
of success. In the case of overall limb alignment and and MAZone by GMS, was only successful in 40%, 47%,
complete GMS, the odds are only 20% that a varus and 49% of limbs, respectively. The definition of success for
deformity of 6.9 degrees mTFA can be corrected to ≤ 3 hemiepiphyseodesis and GMS in LOTV has no clear con-
degrees varus, if a proximal femoral physis is closed. sensus among investigators.6–10,15–17 Reported success rates
Similarly, body weight ≥ 100 kg substantially reduces the of achieving normal adult limb parameters range from 33%
odds of correcting mTFA with GMS. In our models to 67%, comparable to our study.7,10,15,16 Deformity sta-
evaluating GMS, we had insufficient numbers of mild bilization, in this progressive condition, is also a viable
deformities to allow the calculation of thresholds for 80% definition of success for a patient at high risk for
odds of success for mTFA. This population would be osteotomy.17 Even partial correction of deformity may re-
expected to have a more unpredictable outcome with the duce the risks of a subsequent osteotomy, though our study
multiple procedural variations of GMS and an additional does not address this hypothesis. For the limbs that did not
contribution of the femur and knee joint in overall limb achieve complete correction, there was still a mean im-
alignment. All 3 predictors (severity, maturity, and weight) provement in the mTFA of 6.5 degrees and MAD of 18 mm
are moderately to highly correlated so that the model with 3 with GMS.
predictors did not perform any better than the models with The rate of tibial deformity correction of the first
2 (severity + maturity, and severity + weight). Although we LTTBP in our study was 0.33 degrees/month and slower

growth remaining5 and knee age adjustment, based on the

TABLE 4. Predicting Tibial and Overall Limb Alignment
Correction/Overcorrection With Growth Modulation center peak of the distal femur,19 were not. We acknowl-
edge these are not completely independent variables as
Odds of success
larger deformity and greater body weight would be ex-
Preoperative predictor variable 20% 40% 60% 80% pected with more skeletally mature patients. However,
Preoperative MPTA angle (degrees) needed for MPTA success with they are preoperative parameters that can be measured
first-time LTTBP and are predictive of the outcome as demonstrated in

Hip physes Table 4. The success of treatment also was not predicted
All open 67.2 71.5 75.0 79.3
Triradiate closed, femur open 69.7 74.0 77.5 81.8
by sex, race/ethnicity, age at surgery, age at diagnosis, or
Any proximal femoral closure 77.9 82.2 NA NA implant characteristics.
This study is limited by its retrospective design. The
Preoperative mTFA angle (degrees) needed for final-mTFA success from patients were selected and treated by a variety of surgeons
entire GMS using implants of their choice. Surgeons might be expected
Hip physes
Proximal femur open 42.9 27.5 14.8 NA to choose stronger implants for larger or older patients.
Any proximal femoral closure 6.9 NA NA NA The addition of femoral growth modulation may have
Weight (kg) also been chosen preferentially for those limbs with more
< 100 43.2 17.4 14.3 NA severe deformity or advanced maturity. The distal femoral
≥ 100 15.9 NA NA NA
center peak value adjustment of age for maturity13 was not
NA indicates that there were insufficient number of limbs with small enough originally described in standing lower-extremity radio-
on 04/15/2023
deformities having these procedures to calculate the threshold for these higher odds graphs. This may have affected its usefulness with our
of success.
GMS indicates growth modulation series; LTTBP, lateral tibial tension band patients. At the final follow-up, completion of tibial
plating; MPTA, medial proximal tibia angle; mTFA, mechanical tibio-femoral angle. growth was present in 48 of 76 limbs (63%). Therefore, the
final outcomes are also not complete.
This study presents the first prediction tables (Table 4)
than the 0.79 degrees/month reported for tibial tension for the success of the first LTTBP and GMS in LOTV as
band plating, in general.5 This slower rate of correction is quantified by MPTA and mTFA, respectively. Using
in agreement with other studies in tibia vara.4,5 The suc- MPTA, mTFA, body weight, and maturity of the hip
cessful correction of overall limb alignment (mTFA, physes, informed decisions can be made regarding the
MAD, and MAZone) with our GMS has multiple com- expected outcome of the first LTTBP and GMS for an
ponents such as implant failure, revision, rebound, and individual patient. Even if complete correction is not possible,
reimplantation. In addition, the effect of the femoral growth modulation may still be appropriate to reduce
component confounds these parameters. The mean fem- deformity in high-risk patients with few complications.
oral deformity did improve with those femurs having
femoral tension band plating. In contrast, those femurs REFERENCES
that did not have tension band plating had no change in 1. Beskin JL, Burke SW, Johnston CE, et al. Clinical basis for a
their mean mLDFA. Due to the small number of these mechanical etiology in adolescent Blount’s disease. Orthopedics.
adjunctive procedures, we did not have statistical power to 1986;9:365–370.
discern their effects separately. 2. Kline SC, Bostrum M, Griffin PP. Femoral varus: an important
component in late-onset Blount’s disease. J Pediatr Orthop. 1992;13:
We found correction/overcorrection of MPTA with 197–206.
first LTTBP and mTFA by GMS are negatively impacted 3. Mielke CH, Stevens PM. Hemiepiphyseal stapling for knee
and best predicted by deformity magnitude, weight ≥ 100 deformities in children younger than 10 years: a preliminary report.
kg, and hip physeal closure. Others have reported success J Pediatr Orthop. 1996;16:423–429.
4. Castañada P, Urquhart B, Sullivan E, et al. Hemiepiphysiodesis for
in hemiepiphyseodesis in LOTV to correlate with de- correction of angular deformity about the knee. J Pediatr Orthop.
formity severity in the tibia8 and overall limb.7,10,15,16 In 2008;28:188–191.
our study, weight and BMI were correlated with overall 5. Danino B, Rödl R, Herzenberg JE, et al. Guided growth:
limb alignment outcome with GMS, with weight being a preliminary results of a multinational study of 967 physes in 537
better predictor. Two studies have reported overall limb patients. J Child Orthop. 2018;12:91–96.
6. Burghardt RD, Herzenberg JE, Strahl A, et al. Treatment failures
alignment after hemiepiphyseodesis in tibia vara correlates and complications in patients with Blount disease treated with
with BMI7,15 and 1 found no association with weight on temporary hemiepiphyseodesis: a critical systematic literature review.
the tibial outcome.8 J Pediatr Orthop B. 2018;27:522–529.
We did not find an association with age at surgery 7. McIntosh AL, Hanson CM, Rathjen KW. Treatment of adolescent
tibia vara with hemiepiphyseodesis: risk factors for failure. J Bone
but other studies have reported a correlation.7,8,16 Bone Joint Surg. 2009;91:2873–2879.
age may be advanced in children with tibia vara.18 De- 8. Bushnell BD, May R, Campion ER, et al. Hemiepiphysodesis for
termination of skeletal maturity is important in planning late-onset tibia vara. J Pediatr Orthop. 2009;29:285–289.
for growth modulation in this patient population. Hips are 9. Fan B, Zhao C, Sabharwal S. Risk factors for failure of temporary
conveniently present on lower-extremity radiographs and hemiepiphyseodesis in Blount disease: a systematic review. J Pediatr
Orthop B. 2020;29:65–72.
we found that their physeal closure was highly correlated 10. Funk SS, Mignemi ME, Schoenecker JG, et al. Hemiepiphyseodesis
with the success of correction for both the tibia, first implants for late-onset tibia vara: a comparison of cost, surgical success,
LTTBP, and the overall limb, by GMS, whereas years of and implant failure. J Pediatr Orthop. 2016;36:29–35.

11. Stevens PM. Guided growth for angular correction: a preliminary series experience with a systematic review of literature. J Pediatr Orthop.
using a tension band plate. J Pediatr Orthop. 2007;27:253–259. 2020;40:e138–e143.
12. Paley D. Normal lower limb alignment and joint orientation. In: 16. Park SS, Gordon JE, Luhmann SJ, et al. Outcome of hemiepiphyseal
Paley D, Herzenberg JE, eds. Principles of Deformity Correction, stapling for late-onset tibia vara. J Bone Joint Surg. 2005;87-A:
1st ed. New York, NY: Springer-Verlag; 2005;Chapter 1:1–18. 2259–2266.
13. Cameron AC, Miller DL. Robust inference with clustered data. In: 17. Westberry DE, Davids JR, Pugh LI, et al. Tibia vara: results of
Ullah A, Giles DE, eds. Handbook of Empirical Economics and hemiepiphyseodesis. J Pediatr Orthop B. 2004;13:374–378.
Finance, 1st ed. Boca Raton, FL: CRC Press; 2011;Chapter 1:1–28. 18. Sabharwal S, Sakamoto SM, Zhao C. Advanced bone age in children
14. Zeileis A, Köll S, Graham N. Various versatile variances: an object- with Blount disease: a case-control study. J Pediatr Orthop. 2013;33:
oriented implementation of clustered covariances in R. J Stat Softw. 555–557.
2020;95:1–36. 19. Knapik DM, Sanders JO, Gilmore A, et al. A quantitative method
15. Jain MJ, Inneh IA, Zhu H, et al. Tension band plate (TBP)- for the radiological assessment of skeletal maturity using the distal
guided hemiepiphysiodesis in Blount disease: 10-year single-center femur. Bone Joint J. 2018;100B:1106–1111.
on 04/15/2023

ORIGINAL ARTICLE
Predicting Success of Deformity Correction With Tension

Band Plating in Early-Onset Tibia Vara
Janet L. Walker, MD,*† David M. Dueber, PhD,‡ Lindsay P. Stephenson, MD,§

Allison C. Scott, MD,§ Joel A. Lerman, MD,∥ Kenneth P. Powell, MD,¶ Jeffery D. Ackman, MD,#
David E. Westberry, MD,** Sarah B. Nossov, MD,†† and Corinna C. Franklin, MD††
Conclusion: Successful correction of MPTA to age-adjusted

Background: Angular deformity correction with tension band norms following a single-event LTTBP occurred in 53% of tibias
plating has not been as successful in early-onset tibia vara and was best predicted by preoperative-MPTA and preoperative
(EOTV) as it has been in other conditions. Our hypothesis is that body weight <70 kg. Comprehensive growth modulation cor-
perioperative factors can predict the success of lateral tibial rected limbs in 54%. The probability of correction to age-
tension band plating (LTTBP) in patients with EOTV. adjusted MAD is best estimated by preoperative-MAZone 1 or 2
on 04/15/2023
Methods: A retrospective review was performed at 7 centers (MAD ≤ 40 mm). Limbs with preoperative-MAD > 80 mm im-
evaluating radiographic outcomes of LTTBP in patients with proved, but ultimately all failed to correct completely with
EOTV (onset <7 y of age). Single-event tibial LTTBP outcome CLGM. Osteotomy may need to be considered with these severe
was assessed through medial proximal tibial angle (MPTA). The deformities. While modified Langenskiöld classification and
final limb alignment following comprehensive limb growth medial physeal slope have been shown to predict the outcome of
modulation (CLGM), which could include multiple procedures, osteotomy, they were not predictive for LTTBP. Change in
was assessed by mechanical axis zone (MAZone), mechanical MPTA was common after physeal untethering.
tibio-femoral angle (mTFA), and mechanical axis deviation Level of Evidence: Level—III.
(MAD). Preoperative age, weight, deformity severity, medial
physeal slope, and Langenskiöld classification +/- modification Key Words: Blount disease, osteochondritis deformans
were investigated as predictors of outcome. Success was defined tibiae, hemiepiphyseodesis, infantile, growth modulation, guided
as the correction or overcorrection to normal age-adjusted growth, genu varum
alignment. The minimum follow-up was 2 years except when (J Pediatr Orthop 2023;43:e350–e357)
deformity correction, skeletal maturity, or additional surgery
occurred.
Results: Fifty-two patients with 80 limbs underwent 115 tibial
LTTBP procedures at a mean age of 5.3 y, including 78 primary,
21 implant revisions, and 15 reimplantations for recurrence. Ti-
bial LTTBP resulted in a mean change of +8.6o in MPTA and E arly-onset tibia vara (EOTV) results from disordered
growth at the proximal tibial physis. While etiology is
unknown, developmental causes are theorized, as pre-
corrected 53% of tibias. CLGM resulted in MAD correction for
54% of limbs. Univariate analysis showed that success was best sentation before age 2 years is rare. Unlike children with
predicted by preoperative age, weight, MPTA, and MAD. physiological bowlegs, those with EOTV tend to be obese
Multivariate analysis identified that preoperative-MPTA/MAD and walk with a lateral thrust.1 Deformity magnitude
and preoperative-weight < 70 kg were predictive of MPTA and is directly proportional to body mass index (BMI).2
MAD correction, respectively. The probability of success tables Compared to children with late-onset tibia vara, those with
are presented for reference. EOTV are less likely to be male or of Black race, and their
deformities are more frequently bilateral and symmetrical.3
Traditional management for EOTV has been cor-
From the *Shriners Children’s and the University of Kentucky Depart-
ment of Orthopaedic Surgery and Sports Medicine, Lexington KY;
rective tibial osteotomy before age 4 years to minimize
†Shriners Hospital for Children, Lexington, KY; ‡University of recurrence.4,5 Hemiepiphyseodesis has also been reported
Kentucky Department of Statistics, Lexington, KY; §Shriners to manage the varus deformities in the tibia vara.6–13
Children’s, Galveston, TX; ∥Shriners Children’s, Sacramento CA; Compared with osteotomy, hemiepiphyseodesis has a low
¶Shriners Children’s, Shreveport LA; #Shriners Children’s, Chicago
IL; **Shriners Children’s, Greenville SC; and ††Shriners Children’s,
rate of surgical morbidity, though studies report high rates
Philadelphia and Erie PA. of correction failure. To date, preoperative factors pre-
The authors declare no conflicts of interest. dictive of successful outcomes with hemiepiphyseodesis
Reprints: Janet L. Walker, MD, Shriners Children’s and University of are absent. Available studies are difficult to compare due
Kentucky Department of Orthopaedic Surgery and Sports Medicine, to differences in methodology. They are limited by small
Lexington KY, Shriners Hospital for Children, 110 Conn Terrace,
Lexington KY 40508. E-mail: jwalker@shrinenet.org. patient series, heterogeneous populations (late combined
Copyright © 2023 Wolters Kluwer Health, Inc. All rights reserved. with early-onset tibia vara), varying surgical techniques,
DOI: 10.1097/BPO.0000000000002375 or requiring meta-analyses.13

J Pediatr Orthop Volume 43, Number 5, May/June 2023 Deformity Correction of Growth Modulation in EOTV
The purpose of this large, multicenter study is to length, anteroposterior, and lower-extremity radiographs,
evaluate preoperative factors, which may predict out- with patella forward, were measured by a single inves-
comes following hemiephyseodesis, using lateral tibial tigator (JLW) with 36 years of pediatric orthopedic ex-
tension band plating (LTTBP).14 perience to control for variability using McKesson
Radiology Station 12.3 64 bit 1989-2019 (Irvine, TX).
Measurements of deformity severity included: medial

METHODS
proximal tibial angle (MPTA); mechanical lateral distal

We retrospectively reviewed all LTTBP performed at 7 femoral angle (mLDFA); mechanical tibio-femoral angle
tertiary pediatric orthopedic centers from June 2008 to June (mTFA); mechanical axis deviation (MAD)15; meta-
2018. An Institutional Review Board reviewed the study, and physeal-diaphyseal angle (MDA)16; angle of depression of
patients were identified by coding queries at each center. the medial tibial plateau (ADMTP)17; and medial physeal
(Fig. 1) For perioperative review, we report on 58 patients slope (MPS).18,19 As recommended by Sabharwal et al,20
with tibia vara, with 125 LTTBP’s. We defined early-onset as whose normative values we used for comparison, the
age below 7 years at diagnosis, based on the bimodal age MPTA measurement was modified when there was
distribution of patients with tibia vara we found at our rounded or insufficient ossification in the proximal tibial
centers. Surgical procedures were frequently performed epiphysis to define a straight line at the joint surface. In-
above age 7 years and included in this study as long as the stead, a transverse line at the level of the tibial plateau was
patient was below 7 years at diagnosis. For deformity drawn parallel to the proximal tibial physis and used to
outcome analysis, we included only patients with a minimum measure the MPTA.20 In addition, MAD was grouped
follow-up of 2 years except when deformity correction, into 4 zones (MAZone)21: [1] 0 to 20 mm, [2] 21 to 40 mm,
on 04/15/2023
skeletal maturity, or additional surgery occurred. These [3] 41 to 80 mm, and [4] > 80 mm.
criteria resulted in the exclusion of 4 patients due to The severity of the physeal changes was classified
insufficient follow-up and 2 without digital radiographs. using the Langenskiöld22 classification and modifications
Sex, age at diagnosis and surgery, race/ethnicity, body described by LaMont et al.23 Skeletal maturity was as-
weight, BMI and BMI percentile, implant characteristics, sessed by preoperative closure of the triradiate physis,
complications, and follow-up were recorded. which was readily available on the radiographs. Years of
All plates were placed extraperiosteally, spanning remaining growth were calculated assuming maturity at
the lateral proximal tibial physis, without other inter- age 15 years for girls and 17 years for boys and have been
vention. There was no standard protocol during the study reported to correlate with guided growth outcomes.24
period, and patients were managed by their surgeons. Patient demographics, weight parameters, deformity
Implants were manufactured by Orthofix Medical Inc. severity, physeal classification, maturity, and implant
(Guided Growth System, Lewisville, TX) or OrthoPedi- characteristics were assessed as predictors of outcome.
atrics (PediPlates, Warsaw, IN). Eleven limbs also had Complete deformity correction or overcorrection into val-
femoral tension band procedures. gus were considered successful outcomes for the tibia, in
Radiographic change defined deformity outcomes. isolation, and the limb, in its entirety. Previously reported
Preoperative and postoperative digital, standing, full- age-adjusted normative values20 were referenced for MPTA
and MAD success. Success values for mTFA and MAZone
were ≤ 3o varus and Zone 1 or valgus, respectively.
The treatment outcome of single-event LTTBP was
quantified by the MPTA on the radiograph at the final
follow-up or immediately preceding additional surgery.
Overall alignment, after comprehensive limb growth
modulation (CLGM), was described by MAD, mTFA,
and MAZone measured from the last radiograph or before
any bony surgery. Limb alignment following CLGM ac-
counts for the impacts of primary and revision LTTBP,
femoral hemiepiphyseodesis, rebound deformity after
physeal untethering, and reimplantation for management
of recurrence; all common scenarios with CLGM.
The measurement reliability of the investigator was
determined from 38 limb radiographs from children with
EOTV measured at 2-time points while blinded to treat-
ment outcome. Intra-class coefficients for MDA, MPTA,
mLDFA, MAD, and mTFA measurements were all ex-
cellent (0.907 to 0.998). Intra-class coefficients for MPS
and ADMTP were 0.700 (moderate) and 0.869 (good),
respectively.
Univariate logistic regression was employed to assess
FIGURE 1. Selection Tree for Study Subjects. patient characteristics, preoperative deformity, and surgical

Deformity Correction of Growth Modulation in EOTV J Pediatr Orthop Volume 43, Number 5, May/June 2023
covariates with the success of single-event LTTBP and Clinical outcomes are shown in Table 3. Two
CLGM. For patients with bilateral or initial and sub- unplanned surgeries for overcorrection occurred in patients
sequent procedures, their limbs are not independent ob- with greater than 6-month follow-up intervals. At the final
servations in the analysis. This failure of independence can follow-up, there were 21 limbs with closed proximal tibial
bias estimated SE and tests of significance. Therefore, the physes: 8 after epiphyseodeses (corrected), 1 premature
robust sandwich was employed to account for correlations physeal closure (corrected), and 12 mature (Table 3). Thirty-
within subjects, yielding unbiased estimated SE.25 Logistic three immature limbs had implants out, and 12 still had
regression models were estimated using the glm function in implants in place. Fourteen immature and 2 mature limbs
the R statistical computing environment, and the robust had osteotomies. Twenty-four percent of the original 58
sandwich estimator was applied using the sandwich patients were ultimately considered lost to follow-up
package.26 If a variable showed significance (P < 0.05) in (Table 3). Data is available from the corresponding author
the univariate analysis, further evaluation through multi- with an approved data share agreement.
variate logistic regression models followed. Stepwise model- The effect of body size was evaluated by preoperative-
selection procedures were employed in constructing the weight, BMI, and BMI percentile. BMI percentile was
final prediction models. Sensitivity, specificity, positive skewed (skewness = -2.6, kurtosis = 9.9) with 60% of the
predictive value, negative predictive value, and overall population above the 99th percentile and was therefore
classification accuracy were used to assess the performance excluded from further analyses. Univariate analysis
of the multivariate models in the stepwise procedure. (Table 4) indicates the odds of success correlated with
preoperative-weight, BMI, and all measures of deformity
on 04/15/2023
severity. Using logistic regression, preoperative body

weight ≤ 70 kg, was associated with greater success in
RESULTS MPTA and MAD than if preoperative weight was > 70 kg.
The 58 patients had 125 LTTBP for EOTV for (66% vs. 28%, P = 0.004 and 58% vs. 18%, P = 0.026,
perioperative review. Prior treatment included: bracing (20 respectively) Single-event tibial correction, of age-adjusted
limbs), 1 prior broken LTTBP (revision is included), and MPTA, was 67% if preoperative-BMI was <30 kg/m2,
1 prior correction with LTTBP (reimplantation for re- compared with 46% if preoperative-BMI was ≥ 30 kg/m2.
currence is included). Perioperative complications in- (P = 0.04) In addition, when considering limb alignment,
cluded 2 tibial wound dehiscences and 3 tibial plate age/years of growth remaining and additional surgery on
infections, 1 requiring implant removal and 1 arthrotomy the femur had an impact on the correction. The rate of age-
for septic knee arthritis. adjusted MAD success with CLGM was 60% for patients
Fifty-two patients with 80 limbs, encompassing 115 below 8 years at surgery compared with 28% for those
LTTBPs, were available for radiographic outcome anal- who were above 8 years of age. However, this was not
ysis. Their descriptors are presented in Table 1. No significant.(P = 0.138)
statistical differences were identified in MPTA age- Multivariate regressions and model accuracies are
adjusted outcome success between initial LTTBP (45 of shown in Table 5. Each 1-degree increase in the
78 = 58%, 19 required revision) and subsequent LTTBP preoperative-MPTA increased the odds of success by
(24 of 37 = 64%, 1 required revision). (P = 0.54) Therefore, 16%, controlling for preoperative weight. Preoperative
all 115 LTTBPs were included in the univariate analysis weight > 70 kg decreases the odds of success by 81%
for predictors of MPTA outcome. relative to preoperative weight ≤ 70 kg, controlling for
Radiographic outcomes are shown in Table 2. Tibial preoperative-MPTA. Preoperative-MAZone was found to
LTTBP resulted in a mean change of +8.6o in the MPTA be most predictive of successful limb alignment following
(mean rate = +0.46o/month). A single-event LTTBP CLGM. Additional covariates did not improve the
corrected the tibial varus to age-adjusted and adult discriminatory ability in the multivariate model. The
normal or valgus in 53% and 65% of procedures, model-estimated probabilities of success are presented in
respectively. CLGM outcomes were successful in 53% to Table 6. Fisher exact tests indicated the probability of
65%, depending on the measure and are summarized in success was lower for preoperative-MAZone 3 than
Table 2. These included any deformity change after the preoperative-MAZones 1 and 2 (P = 0.008), and was
removal of 48 implants. Forty-one tibias followed for ≥ lower for preoperative-MAZone 4 than for preoperative-
1 year to 6.8 years after implant removal, demonstrated a MAZone 3 (P = 0.021).
mean loss of 6.7o MPTA correction (range = 17 degrees
loss to 6.6 degrees increased valgus), over a mean
duration of 2.6years (mean rate = 0.28 degrees/month). DISCUSSION
Final assessment of these 41 tibias revealed: 9 with There is no consensus definition of success for
MPTA > 90 degrees, 8 with MPTA < 85 degrees, and 24 growth modulation in tibia vara, making comparison to
with MPTA 85degrees-90degrees. Other deformity the literature difficult. We are the first to use age-adjusted
parameters at final evaluation include the mean mLDFA norms and expect the deformities to be corrected or
of 89.1 degrees (range = 80.6 degrees to 104 degrees) and overcorrected for a successful outcome. A LTTBP should
the mean ADMTP of 26.9 degrees (range = 0.7 degrees to most directly affect the tibial alignment as measured by
54 degrees). MPTA. A 65% rate of MPTA correction for a single-event

limb alignment was predicted by preoperative-MAD

TABLE 1. Deformity Correction Analysis Population Summary
≤ 40 mm (MAZone 1+2) and preoperative-MAD ≤ 80
Patients (n = 52) mm (MAZone 1-3). While this MAZone, described by an
Sex (at birth) 38 females, 14 males
Race/Ethnicity 20 Black/not Hispanic, 14 White/Hispanic, absolute MAD distance instead of a percentage, has been
(self-reported) 13 White/not Hispanic, 5 other used in studies of late-onset tibia vara,21 it also appears
Age at Diagnosis mean = 3.1 y, range, 1–6 y relevant for EOTV. The use of MAZone simplifies clinical
Limbs (n = 80) use for prediction as it does not require knowledge of age-
Side 42 left, 38 right
Preoperative age mean = 5.3 y, range, 1.9–13.4 y
adjusted normative values of limb alignment. We present
Preoperative years of mean = 10.2 y, range, 3.6–14.3 y the prediction tables (Table 6) to allow surgeons to
growth remaining consider the appropriateness of LTTBP for an individual
Preoperative body mean = 37.5 kg, range , 13.3–150 kg patient. Since limbs with preoperative-MAD > 80 mm did
weight not fully correct during the study period, consideration
Preoperative body mass mean = 26 kg/m2, range, 16–49 kg/m2
index (BMI) for osteotomy rather than growth modulation may be
Preoperative-BMI mean = 95th percentile, 73% over the 95th appropriate. Adding femoral procedures may be indicated
percentile percentile in those limbs that have femoral varus. This investigation
Preoperative 2-I, 21-II, 32-III, 19-IV, 2-V did not have a sufficient volume of femoral growth
Langenskiöld Class
Femoral tension band 11 (10 varus, 1 valgus)
modulation procedures, and future studies would be
plates necessary to understand the impact of the combined
Tibial Implants (n = 115) tibial and femoral intervention.
We found that age at first surgery below 8 years was
on 04/15/2023
Procedure type
Primary 78 predictive of success in overall MAD correction. Power
Revisions 7 loss of bone fixation
7 broken implants (1 broken before the analysis suggests that a sample size of 275 may be suffi-
study) ciently large to detect the effect of age at the surgery on
4 migration with bone growth overall MAD correction, controlling for MAZone. The
4 maximal screw divergence bone age in patients with early-onset Blount disease has
Reimplantations for 15 (1 recurrence before the study)
recurrence
been reported to be advanced by a mean of 26 months.27
Implant manufacturer 58 Orthofix©, 55 Orthopediatrics©, and 2 Proposed knee bone age methods target older children
not recorded using knee films. We used hip physes seen on our radio-
Metal composition 61 stainless steel, 46 titanium, and 8 not graphs. Only 8 limbs had preoperative-triradiate physeal
recorded closure, and its status could not be correlated with success.
Cannulated vs. solid 87 cannulated, 28 solid
screws Despite this, skeletal age is likely related to outcome.
Screw diameter 107 4.5 mm, 8 3.5 mm Future studies to evaluate bone age as a predictor appear
No. screws (4 vs. 2) 84 2-screws, 31 4-screws warranted.
While not frequently reported in studies on EOTV,
preoperative-weight < 70 kg was also predictive of suc-
LTTBP has been reported in an EOTV subset (< 10 y at cessful correction in single-event LTTBP. The odds of that
onset)11 and 70% tibial correction in an infantile tibia vara success can be predicted by preoperative-MPTA (Table 6).
cohort (< 4 y at onset).12 Using the adult norms for Eighteen tibias had LTTBP in patients weighing ≥ 70 kg.
MPTA (Table 2), our 65% success rate compares with Ten were first-time implants, and 8 were subsequent
these studies. We achieved this with 4% wound procedures. All the patients were older than 6 years at
complications but had 18% implant revisions. surgery, and 12 had triradiate physeal closure. In our
The goal of growth modulation is to correct the univariate analysis, we found preoperative-BMI correlated
entire limb. In our patients with EOTV, CLGM yielded with tibial and overall limb deformity outcomes. BMI has
overall limb deformity correction in 53% to 65%, de- been shown to be correlated with the outcome of
pending on the measure. This is less than the 89% success hemiepiphyseodesis in late-onset tibia vara21,28 and in a
rate reported by Scott et al.8 The main difference between mixed tibia vara population.10 Forces across the physis are
this study and ours is deformity severity. The limb with the implicated in the development and the severity of
most severe deformity in the Scott et al8 study had pre- EOTV.1,2 Weight plays a role in implant breakage29 and
operative-MAD = 50 mm. In our study, the 44 limbs with might be expected to play a role in the outcome of
preoperative-MAD ≤ 50 mm had a success rate of 75% LTTBP. Unfortunately, age, weight, and deformity are
measured by age-adjusted MAD, but 86% as measured by highly correlated, and the statistical models could not
MAZone. Deformity severity impacts outcome as dem- reliably distinguish the most important variable.
onstrated by our finding that limbs with preoperative- The literature reports MPS,19 Langenskiöld
MAD > 80 mm improved, but none were corrected com- classification,22 and its LaMont et al23 modification cor-
pletely. relate with recurrence in early-onset Blount disease after
Our multivariate models demonstrated that the treatment with early tibial osteotomy.19,23 We did not find
probability of success in the tibia was predicted by them predictive of success with LTTBP and theorize that
increasing preoperative-mMPTA and preoperative- this is due to the low inter-rater reliability of these
weight < 70 kg, while the probability of successful overall measures.30,31 Of the 2 limbs retrospectively assigned by

TABLE 2. Radiographic Outcome of Lateral Tibial Tension Band Plating (LTTBP) for Early-Onset Tibia Vara (52 Patients, 80 Limbs,
and 115 LLTBP)
Mean Mean Mean Median Number, % Number, %
Achieving age- Achieving adult norm
Single-Event Tibial LTTBP (n = 115) Preoperative Treatment end Correction Correction adjusted success* success†
Medial Proximal Tibial Angle (degrees) 77.0 85.6 +8.6 +8 61 (53) 72 (65)
Mean Mean Mean Median Number Number

Comprehensive Limb Growth Modulation Preoperative Study End Correction Correction Achieving Achieving adult norm
(n = 80) age-adjusted success* success†
Medial Proximal Tibial Angle 75.3 83.9 +8.5 +8 47 (59) 50 (63)
(degrees)
Mechanical Tibio-Femoral Angle 22.0 6.5 –15.5 –14 — 42 (53)
(degrees varus)
Mechanical Axis Deviation (millimeters 49.6 19.3 –30.3 –29 43 (54) 49 (61)
varus)
Mechanical Axis Zone (n)21 — — — 52 (65)
4 Valgus — 1 — — — —
2 Valgus — 1 — — — —
1 Valgus — 21 — — — —
1 Varus 17 29 — — — —
on 04/15/2023
2 Varus 16 16 — — — —
3 Varus 36 6 — — — —
4 Varus 11 6 — — — —
*Age-adjusted success = correction or overcorrection to age-adjusted limb parameters.
†Adult norm success = medial proximal tibial angle ≥ 85o, mechanical tibio-femoral angle ≤ 3o varus, mechanical axis deviation <17 mm varus and MAZone valgus or
1 varus.
the principal investigator to preoperative-Langenskiöld V, those that corrected or overcorrected was 1.7 years, but
1 corrected to age-adjusted MAD and the other did not. there was a wide range so we cannot determine a time after
Some have recommended growth modulation im- which correction will no longer occur. Change in de-
plant removal after a certain duration of physeal tethering formity after implant removal has been reported to occur
for concerns of bony overgrowth or physeal arrest. In our less frequently in tibia vara.32 We found its occurrence to
study, there were no implant removals without complete be quite variable, with MPTA decreasing, remaining
correction, skeletal maturity, or other bony procedures. static, or increasing after implant removal. Therefore, this
This suggests the implants were not removed because of data does not support routine overcorrection. Girls under
implant duration alone. The mean implant duration for 10 years and boys under 12 years are reported to be at the
TABLE 3. Clinical Outcomes of Lateral Tibial Tension Band Plating for Early-Onset Tibia Vara
Correction Analysis Patients (n = 52)
Late follow-up 16 with > 6 mo follow-up intervals between visits at some point
1 required varus osteotomy and 1 required reverse plating of the femoral side
Original Patients (n = 58)
Lost to follow-up by study end 12 immature patients/18 limbs: 17 without implants, 1 with implants, and 1 referred outside
2 immature patients were originally excluded due to insufficient follow-up
Limbs (n = 80)
Epiphyseodesis 7 tibial, 1 tibial, and femoral
Osteotomies 14 valgus (1 with epiphysiolysis), 1 rotational only, and 1 varus (for overcorrection)
Growth arrest 1 complete after correction and 1 progressive valgus after implant removal
Skeletally mature 12: 2 had osteotomies, 3 corrected their MAD, and 7 remain uncorrected
Proximal Tibial Physis Open 33 implants out and 12 implants in
Tibial Implants (n = 115)*
Implant Removals 53 removals for correction/overcorrection
mean implant duration = 1.7 y, range = 0.4–5.2 y
mean MPTA at removal = 93.7o, range = 85o-106o
1 for infection
1 for pain at maturity with MPTA = 84o
1 at physeal bar resection
8 at epiphyseodesis
9 at osteotomy
Broken Screws 6 revised and 4 removed at osteotomy
*for all revisions and reimplantations, see Table 1.

greatest risk for rebound phenomenon.32 Ninety-five per-

TABLE 4. Results of Univariate Logistic Regression Analysis for cent of our population with implant removal was in these
Success high-risk groups. Continued assessment until maturity
Odds ratio appears warranted. Reimplantation with LTTBP was
Preoperative predictor [95% CI] P successful in the correction of the recurrent deformity.
Tibial Alignment, as determined by the age-adjusted medial proximal tibial angle (MPTA) Our retrospective design limits this study. Patients
(n = 115 tibial procedures) were selected and procedures were performed by a vari-
Demographic covariates
Age at surgery 0.92 [0.79, 1.06] 0.253
ety of surgeons. This is also a strength of the study as it is
Female vs. male 1.3 [0.56, 3.07] 0.542 a more accurate representation of clinical practice. The
Race/Ethnicity — 0.469 use of 4 rather than 2-hole plate or additional femoral
Black vs. White 1.11 [0.38, 3.20] —
Hispanic vs. White 0.78 [0.25, 2.36] —
procedures varies among surgeons. The literature is in-
Weight covariates consistent in the stratification of patients with tibia vara
Weight, per 1-kg increase 0.98 [0.97, 1.00] 0.012* based on their age at onset. The most common is a di-
Body mass index 0.92 [0.88, 0.96] < 0.001* vision of infantile <4 years, juvenile 4 to 10 years and
Tibial deformity severity covariates
MPTA, per 1o increase 1.16 [1.10, 1.24] < 0.001* adolescent ≥ 11 years.33 However, studies may then re-
Metaphyseal-diaphyseal angle 0.91 [0.87, 0.95] < 0.001* port results combining groups or blurring the age
Medial physeal slope 0.97 [0.94, 1.01] 0.087 ranges.7,9,10,13,19,21–23,28 There are distinct clinical dif-
Physeal changes covariates
Langenskiöld classification — 0.098 ferences between the younger and older onset types of
III vs. II 1.1 [0.39, 3.08] — tibia vara, but those of intermediate age onset are more
—
on 04/15/2023
IV vs. II 0.87 [0.30, 2.49] variable. We looked at the total tibia vara population of
V vs. II N/A† N/A†
N/A† 148 patients at our centers. Finding a nadir frequency of
Lamont et al Modified Classification — 0.311 2 patients at age 7 to 8 years, we used <7 years as the
B vs. A 0.88 [0.27, 2.83] — breakpoint of our EOTV study. Comparing our patients
C vs. A 0.52 [0.18, 1.44] —
Maturity covariates with age at diagnosis < 4 years versus 4 to 6 years showed
Years of growth remaining 1.16 [1.01, 1.34] 0.04* expected differences in body weight but no statistical
Triradiate physis open 2.63 [0.95, 7.69] 0.062 difference in the outcome. Power analysis suggested that
Surgical covariates
Cannulated vs. solid screws 2.06 [0.88, 4.81] 0.096
a sample size of 764 would be necessary to detect dif-
Screw diameter (4.5 mm vs. 3.5 mm) 1.12 [0.25, 4.93] 0.881 ferences in outcome. However, our comparison to the
Number of screws (4 vs. 2) 0.66 [0.43, 1.01] 0.052 literature is affected because we chose a different
Metal (Titanium vs. Steel) 1.70 [0.77, 3.75] 0.189
classification point.
Overall Limb alignment, as determined by age-adjusted mechanical axis deviation (MAD) In our study, we combined initial and subsequent
(n = 80 limbs)
Preoperative predictor LTTBP procedures for the MPTA outcome analysis. The
Demographic covariates rate of success was slightly higher for subsequent pro-
Age at surgery 0.78 [0.65, 0.94] 0.009*
Female vs. male 0.78 [0.28, 2.13] 0.622
cedures, despite an older and heavier population, but
Race/Ethnicity — 0.083 was not statistically different. Power analysis suggests
Black vs. White 0.25 [0.06, 1.04] — that to have an 80% power to detect a significant effect,
Hispanic vs. White 0.75 [0.20, 2.80] —
Weight covariates given group population proportions equal to observed
Weight, per 1-kg increase vs. Pre-WT 0.97 [0.95, 0.99] 0.004* proportions, a total sample size of 388 (half initial and
Body mass index 0.93 [0.88, 0.98] 0.005 half subsequent procedures) or 433 using a sample ratio
Overall limb deformity severity covariates
Mechanical tibio-femoral angle 0.91 [0.86, 0.96] 0.001* of 2 initial:1 subsequent (similar to ours) would be re-
MAD 0.94 [0.92, 0.97 < 0.001* quired. We selected the radiographic parameters used in
Mechanical axis zone (MAZone) ‡ 21 — < 0.001*
MAZone 2 vs. 1 0.4 [0.06, 2.60] —
previous studies on EOTV but acknowledge that those
MAZone 3 vs. 1 0.1 [0.02, 0.48] — measuring overall limb deformity (MAD, mTFA, and
Physeal changes covariates MAZone) are affected by the femoral deformity. The
Langenskiöld — 0.07
III vs. II 0.56 [0.18, 1.74] —
femoral deformity may change in EOTV with remodel-
IV vs. II 0.23 [0.06, 0.87] — ing. EOTV results in 3-dimensional changes in the tibia.2
V vs. II 0.5 [0.03, 9.24] — but records were limited regarding torsional or sagittal
Lamont modified Classification — 0.167
B vs. A 0.47 [0.13, 1.65] — deformity. Many limbs had more than 1 LTTBP, and
C vs. A 0.35 [0.11, 1.06] — prior LTTBPs distort the shape of the lateral tibial
Maturity covariates
Years of growth remaining 1.28 [1.08, 1.53] 0.005*
physis. This may have affected those variables dependent
Triradiate physis open 3.7 [0.70, 18.62] 0.122 on the lateral physeal-metaphyseal landmarks such as
Surgical covariates the MDA, MPS, and the MPTA. Closure of the prox-
Addition of a femoral implant 0.07 [0.01, 0.57] 0.013*
imal tibial physis was only present in 21 patients at the
*Significance p < 0.05. final follow-up. Therefore, this study cannot assess final
†Of the 3 cases with Langenskiöld V, none were successful in correcting
mMPTA (age-adjusted). Accordingly, an odds ratio for Langenskiöld V vs. II could
outcomes.
not be computed. In conclusion, single-event LTTBP can correct varus
‡Of the 11 cases with MAZone = 4, none were successful in correcting in 53% of tibias. This correction to age-adjusted norms is
mechanical axis deviation (age-adjusted). Accordingly, an odds ratio for MAZone
= 4 vs. MAZone = 2 could not be computed. best predicted by preoperative-MPTA and preoperative
body weight < 70 kg (Table 6). Comprehensive limb growth

TABLE 5. Results of Multivariate Logistic Regression Analysis and Prediction Accuracy

Odds Ratio 95% Confidence Overall Sensitivity/ Predictive ValuePositive/
Variable Interval P Accuracy Specificity Negative
Tibial alignment success, as determined by medial proximal tibial angle (mMPTA; age-adjusted)
Model 1
Preoperative-MPTA 1.16 [1.09, 1.25] < 0.001 — —

Weight, per 1 kg increase 0.98 [0.97, 1.00] 0.07 — —
Model 2
Preoperative-MPTA 1.16 [1.09, 1.25] < 0.001 — —
Weight (> 70 kg) 0.19 [0.04. 0.73] 0.02 — —
Model 2 (Preoperative-MPTA and preoperative- 0.77 0.86/0.65 0.79/0.75
weight > 70 kg)
Overall Limb alignment success, as determined by mechanical axis deviation (MAD; age-adjusted)
Model 1
Preoperative-mechanical axis zone (MAZone (1)
MAZone 2 vs. MAZone1 0.40 [0.05, 2.42] 0.334 — —
MAZone 3 vs. MAZone 1 0.1 [0.01, 0.40] 0.004 — —
MAZone 4 vs. MAZone 1 N/A N/A — —
Model 1 (Preoperative-MAZone)* 0.74 0.64/0.84 0.82/0.69
*None of the 11 cases with MAZone = 4 were successfully corrected according to MAD (age-adjusted)
on 04/15/2023
modulation was successful in 54%. The probability of ACKNOWLEDGMENTS

correction to age-adjusted MAD is best estimated by the The authors thank the administrative assistance of the
preoperative-MAZone 1 or 2 (MAD ≤ 40 mm). Limbs with research staff at each of the centers for their efforts in this
preoperative-MAD > 80 mm improved, but all failed to study: Rebecca Davis, Judy Linton, Kory Bettencourt,
correct completely with CLGM. Osteotomy may need to be Marnita Thomas, Sahar Hassani, Ashley Carpenter, and
considered with these severe deformities. While modified Solomon Praveen Samuel.
Langenskiöld classification and medial physeal slope have
been shown to predict the outcome of osteotomy, they were
not predictive for LTTBP. Change in deformity was REFERENCES
common after physeal untethering, so patients must be 1. Scott AC, Kelly CH, Sullivan E. Body mass index as a prognostic
followed until physeal closure. factor in development of infantile Blount disease. J Pediatr Orthop.
2007;27:921–925.
2. Sabharwal S, Zhao C, McClemens E. Correlation of body mass
index and radiographic deformities in children with Blount disease.
J Bone Joint Surg Am. 2007;89:1275–1283.
TABLE 6. Predicting Tibial and Overall Limb Alignment 3. Rivero SM, Zhao C, Sabharwal S. Are patient demographics
Correction/Overcorrection With Lateral Tibial Tension Band different for early-onset and late-onset Blount disease? Results based
Plating (LTTBP) and Comprehensive Limb Growth on meta-analysis. J Pediatr Orthop B. 2015;24:515–520.
Modulation (CLGM) 4. Loder RT, Johnston CE. Infantile tibia vara. J Pediat Orthop.
Probability of success 1987;7:639–646.
5. Chotigavanichaya C, Salinas G, Green T, et al. Recurrence of varus
Preoperative predictor deformity after proximal tibial osteotomy in Blount disease: long
variable 20% 40% 60% 80% term follow-up. J Pediatr Orthop. 2002;22:638–641.
Preoperative-MPTA* angle (degrees) needed for age-adjusted MPTA 6. Henderson RC, Kemp GJ Jr, Greene WB. Adolescent tibia vara:
success with single-event LTTBP alternatives for operative treatment. J Bone Joint Surg Am. 1992;74:
Weight 342–350.
< 70 kg 63.6 70.1 75.4 81.9 7. Westberry DE, Davids JR, Pugh LI, et al. Tibia vara: results of
≥ 70 kg 74.5 81.0 N/A N/A hemiepiphyseodesis. J Pediatr Orthop B. 2004;13:374–378.
8. Scott AC. Treatment of infantile Blount disease with lateral tension
Estimated probability of achieving age-adjusted normal mechanical axis band plating. J Pediatr Orthop. 2012;32:29–34.
deviation (MAD) with CLGM 9. Burghardt RD, Herzenberg JE, Strahl A, et al. Treatment failures
MAZone† and complications in patients with Blount disease treated with
1 (MAD 0–20 mm varus) 88 temporary hemiepiphyseodesis: a critical systematic literature review.
2 (MAD 21–40 mm varus) 75 J Pediatr Orthop B. 2018;27:522–529.
3 (MAD 41–80 mm varus) 42 10. Jain MJ, Inneh IA, Zhu H, et al. Tension band plate (TBP)-guided
4‡ (MAD > 80 mm varus) 0 hemiepiphyseodeis in Blount disease: 10-year single-center experience with
a systematic review of literature. J Pediatr Orthop. 2020;40:e138–e143.
*medial proximal tibial angle. 11. Danino B, Rödl R, Herzenberg JE, et al. The efficacy of guided
†mechanical axis zone.21
growth as an initial strategy for Blount disease treatment. J Child
‡In our sample, none of the surgeries on the 11 limbs with preoperative-MA-
Zone = 4 were successful. Orthop. 2020;14:312–317.
MAD indicates mechanical axis deviation; MPTA, medial proximal 12. Griswold BG, Shaw KA, Houston H, et al. Guided growth for the
tibial angle. treatment of infantile Blount’s disease: Is it a viable option? J Orthop.
2020;20:41–45.

13. Fan B, Zhao C, Sabharwal S. Risk factors for failure of temporary 24. Danino B, Rödl R, Herzenberg JE, et al. Guided growth:
hemiepiphyseodesis in Blount disease: a systematic review. J Pediatr preliminary results of a multinational study of 967 physes in 537
Orthop B. 2020;29:65–72. patients. J Child Orthop. 2018;12:91–96.
14. Stevens PM. Guided growth for angular correction: a preliminary 25. Cameron AC, Miller DL. Robust inference with clustered data. In:
series using a tension band plate. J Pediatr Orthop. 2007;27:253–259. Ullah A, Giles DE, eds. Handbook Empir Econ Fin. Boca Raton, FL:
15. Paley D. Normal lower limb alignment and joint orientation. In: CRC Press; 2011:1–28.
Paley D, Herzenberg JE, eds. Principles of Deformity Correction. 26. Zeileis A, Köll S, Graham N. Various versatile variances: an object-
New York, NY: Springer-Verlag; 2005:1–18. oriented implementation of clustered covariances in R. J Stat Softw.
16. Levine AM, Drennan JC. Physiological bowing and tibia vara. The 2020;95:1–36.
metaphyseal-diaphyseal angle in the measurement of bowleg 27. Sabharwal S, Sakamoto SM, Zhao C. Advanced bone age in children
deformities. J Bone Joint Surg Am. 1982;64:1158–1163. with Blount disease: a case-control study. J Pediatr Orthop. 2013;33:
17. Janoyer M, Jabbari H, Rouvillian JL, et al. Infantile Blount’s disease 551–557.
treated by hemiplateau elevation and epiphyseal distraction using a 28. McIntosh AL, Hanson CM, Rathjen KW. Treatment of adolescent
specific external fixator: preliminary report. J Pediatr Orthop B. tibia vara with hemiepiphyseodesis: Risk factors for failure. J Bone
2007;16:273–280. Joint Surg. 2009;91:2873–2879.
18. Kling TF, Volk AG, Dias L, et al. Infantile Blount’s treated with 29. Chapman P, Dueber DM, Stephenson LP, et al. Guided growth
osteotomy: follow-up to maturity. Orthop Transac. 1990;14:634–635. procedures: broken tension band implants in patients with Blount
19. Maré PH, Thompson DM, Marais LC. Predictive factors for disease. J Pediatr Orthop. 2022;42:e435–e440.
recurrence in infantile Blount disease treated by osteotomy. J Pediatr 30. du Plessis J, Firth GB, Robertson A. Assessment of the reliability
Orthop. 2021;41:e36–e43. and reproducibility of the Langenskiold classification in Blount’s
20. Sabharwal S, Zhao C, Edgar M. Lower limb alignment in children. disease. J Pediatr Orthop B. 2020;29:311–316.
J Pediatr Orthop. 2008;28:740–746. 31. Vosoughi F, Nabian MH, Simon AL, et al. Langenskiold
21. Funk SS, Mignemi ME, Schoenecker JG, et al. Hemiepiphyseodesis classification of tibia vara: a multicenter study on interrater
on 04/15/2023
implants for late-onset tibia vara: A comparison of cost, surgical reliability. J Pediatr Orthop B. 2022;31:114–119.
success, and implant failure. J Pediatr Orthop. 2016;36:29–35. 32. Leveille LA, Razi O, Johnston CE. Rebound deformity after
22. Langenskiöld A. Tibia vara: (osteochondrosis deformans tibiae): growth modulation in patients with coronal plane angular deformities
a survey of 23 cases. Acta Chir Scand. 1952;103:1–22. about the knee: Who gets it and how much? J Pediatr Orthop. 2012;32:
23. LaMont LE, McIntosh AL, Jo CH, et al. Recurrence after surgical 29–34.
intervention for infantile tibia vara: assessment of a new modified 33. Thompson GH, Carter JR, Smith CW. Late-onset tibia vara:
classification. J Pediatr Orthop. 2019;39:65–70. a comparative analysis. J Pediatr Orthop. 1864;4:186–194.

ORIGINAL ARTICLE
Robotics Coupled With Navigation for Pediatric Spine

Surgery: Initial Intraoperative Experience With 162 Cases
Nicole Welch, BA,* Frank Mota, MD,* Craig Birch, MD,*† Lauren Hutchinson, MPH,*
and Daniel Hedequist, MD*†
registration, as no loss of registration occurred after adopting

Background: This study assesses intraoperative efficacy, accuracy, this technique.
and complications of pedicle screw placement using robotic- Conclusion: This study highlights the safety and screw accuracy
assisted navigation (RAN) in pediatric spine surgery. associated with the use of RAN in pediatric patients.
Methods: A retrospective review of patients who underwent spine Level of Evidence: Level III.
deformity surgery using RAN at a single pediatric institution
from 2019 to 2021 was conducted. Patient demographics, peri- Key Words: robotics coupled with navigation, pediatric, spine
surgery, high-speed navigated drill, adolescent idiopathic
on 04/15/2023
operative metrics, screw execution and accuracy, technical diffi-

culties, and other outcomes were summarized. In cases with scoliosis
postoperative computed tomography scans, screws were classi- (J Pediatr Orthop 2023;43:e337–e342)
fied using the Gertzbein and Robbins classification scale. Fisher
exact tests were used to assess the relationship between proce-
dural changes and lateral screw malposition.
Results: One hundred sixty-two cases with an average patient age
of 15.1 years (range, 4 to 31 y) were reviewed. The most common
diagnosis was adolescent idiopathic scoliosis (n = 80) with an
E nabling technologies in surgery are by definition de-
veloped to improve operational outcomes. The out-
comes may be a variety of measurements but usually
average major curve of 65 degrees. Of 1467 screws attempted, revolve around surgical safety and efficacy. The use of
1461 were executed successfully (99.6%). All failures were in spinal instrumentation for the treatment of scoliosis has
Type D pedicles and were lateral deviations recognized with progressed from Harrington instrumentation consisting of
routine intraoperative fluoroscopy. In cases with postoperative simple hook-rod constructs to modern day segmental in-
computed tomography imaging, 100% of screws (n = 197) were strumentation using pedicle screws. Pedicle screws for
placed with complete containment (Grade A). Remaining screws pediatric spine deformity surgery were designed to enable
were graded as accurate by mirroring fluoroscopy and planned surgical correction of the spine while simultaneously al-
computer software positions. In 4% of cases, loss of registration lowing for a stable mode of fixation. Pedicle screw fixation
was detected by a safety check before drilling at the planned has since evolved to include routine use of thoracic pedicle
level. There were no neurological deficits or returns to the op- screws in pediatric deformity cases. The use of screw fix-
erating room. Two changes occurred as part of the learning curve ation for scoliosis cases became more commonplace after
associated with this technique. (1) Adoption of a high-speed literature showed a high rate of successful screw placement
navigated drill: Change 1 (last 74 cases). (2) Drilling all pilot in pediatric patients using freehand techniques.1 However,
holes robotically first, then placing screws within the robotically complications related to spinal instrumentation remain the
established tracts to avoid motion and subsequent registration leading cause of reoperation in pediatric deformity cases.2
disruption: Change 2 (last 39 cases). Change 1 was less likely to The majority of these reoperations in the immediate
result in screw malposition as no screws skived lateral with the postoperative period are related to malpositioned pedicle
technique (P = 0.03). Change 2 trended toward statistical screws, which place the patient at risk of neurological or
significance for avoidance of screw malposition and loss of visceral injuries.2,3
To avoid potential screw malposition, surgical
From the *Department of Orthopaedic Surgery, Boston Children’s navigation techniques were developed to aid in pedicle
Hospital; and †Harvard Medical School, Boston, MA. screw placement. Computer-assisted navigation (CAN)
This study was approved by the Boston Children’s Hospital Institutional of pedicle screw placement in pediatric patients has been
Review Board.
This research was not funded.
shown to be associated with a decreased rate of reoper-
D.H. is a consultant for Medtronic. The remaining authors declare no ation and improved accuracy of screw placement.4,5 The
conflicts of interest. ability to place screws in an “acceptable” manner,
Reprints: Daniel Hedequist, MD, Boston Children’s Hospital, Chief of meaning down the expected trajectory, is upwards of
Spine Division, Associate Professor of Orthopaedic Surgery, 300 98% in a recent study by Baldwin, et al.6 There remains
Longwood Avenue, Boston, MA 02115. E-mail: Daniel.Hedequist@
childrens.harvard.edu. some pitfalls to standard CAN, including inaccurate
Copyright © 2023 Wolters Kluwer Health, Inc. All rights reserved. registration, line of site workflow problems, and manual
DOI: 10.1097/BPO.0000000000002381 dexterity errors by the surgeon. Robotics has evolved

Welch et al J Pediatr Orthop Volume 43, Number 5, May/June 2023
on 04/15/2023
IMAGE 1. The robotic-assisted navigation planning software allows for planning of the size and trajectory of pedicle screws before
placement.
from original systems that relied on accurate placement with navigation were identified. Standard demographic
of Kirschner wires through the robotic arm, to current data were collected for these patients through chart re-
systems that have improved surgical planning software view. Patient medical records, operative reports, and ra-
and robotic arm technology (Image 1).7,8 These systems diographic studies were reviewed. Standard robotic
rely on patient registration and planning of screw variables were obtained for each case from the computer
placement with the robotic arm guiding the instrument software component of RAN. Surgical variables obtained
path down the pedicle and theoretically minimizing any included the procedure performed, the total surgical and
manual dexterity errors by the surgeon.7,9 total robotic times, the total number of pedicle screws
In the spring of 2019, the Food and Drug Admin- placed, and the number of pedicle screws placed using
istration (FDA) cleared the first platform that coupled the RAN. In addition, the number of malpositioned screws
technologies of robotics with CAN for clinical use in the and any complications related to screw placement were
United States. This platform, robotic-assisted navigation recorded. Regarding RAN, the mode of registration was
(RAN), is multidimensional: it allows for registration of collected for each patient and any technical difficulties
patient anatomy with the computer software, planning of associated with the platform, specifically loss of registra-
the size and trajectory of pedicle screws, sending of the tion. There are currently 2 methods to register patient
robotic arm to a predetermined trajectory at each level, anatomy to the RAN system. One method involves
and use of navigated instruments to confirm in real time merging a preoperative computed tomography (CT) scan
accurate instrument and screw placement.10 Only sparse with intraoperative fluoroscopy images. This technique
literature is available on the combination of these tech- was used for patients who had a preoperative CT scan as
nologies mainly due to the recent clearance by the FDA standard of care (i.e., patients with congenital scoliosis,
for their use.9,11 The initial reports suggest up to a 98.9% spondylolisthesis). The other registration method uses an
accuracy rate using RAN but are limited to a small intraoperative O-arm scan. The protocol for this scan is
number of cases.9,11 The purpose of this study was to
present the intraoperative experience with a large number
of pediatric patients treated using this platform at a single TABLE 1. Summary of Cohort Diagnoses (N = 162)
tertiary, urban institution. Characteristic Freq. (%)
Adolescent idiopathic scoliosis 80 49
METHODS Spondylolisthesis 26 16
Congenital scoliosis 10 6
Institutional Review Board approval was obtained Neuromuscular scoliosis 20 12
before searching a single-center surgical spine database for Tumor 2 1
patients who were operated on between June 2019 and Other, including kyphosis 24 15
September 2021. Consecutive patients who underwent Freq. indicates frequency.
spinal instrumentation with the use of robotics coupled

J Pediatr Orthop Volume 43, Number 5, May/June 2023 Use of RAN in Pediatric Spine Surgery
CT scans with the pediatric dose protocol applied. Intra-

operative connection of the robot to the patient was done
by a Schanz pin to the posterior superior iliac spine in 48
(30%) patients and spinous process clamping in 114 (70%)
patients (Image 2). Mean RAN time, measured from the
start of robotic registration to the last screw trajectory,

was 52.3 minutes (range, 11 to 131 min).

In 162 patients, 1467 pedicle screws were attempted
with RAN—a mean of 9.1 screws per patient (range, 1 to
29 screws). Of the 1467 screws attempted using RAN,
1461 were executed successfully (99.6%). The 6 screws that
failed were in morphologic Type D pedicles and were
lateral deviations recognized with routine intraoperative
IMAGE 2. The robot clamped to the spinous process before fluoroscopy. There was one medial perforation related to
registration to the patient anatomy. drilling, recognized by probing allowing a deficit of the
medial wall to be appreciated. No screw was placed at that
one that is published in the pediatric literature and widely level and upon registration check, it was realized that there
used at pediatric institutions.12 In cases with three- had been loss of registration accuracy at that level. Oth-
dimensional intraoperative imaging or postoperative CT erwise, there were no noted medial screw deviations and
on 04/15/2023
scans, pedicle screws were graded using the Gertzbein and no noted neurological deficits or returns to the operating
Robbins classification scale. The scale includes five clas- room. All screws (other than the 6 lateral deviations) were
sifications: Grade A: accurate screws with no cortical considered accurate using the Lenke criteria (medial bor-
breach, Grade B: screws with 0 to 2 mm cortical breach der of the pedicle should not be completely visualized,
without requirement for surgical revision, Grade C: screws screw tips should not cross the midline, cascade of screws
with 2 to 4 mm cortical breach, Grade D: screws with 4 to should be harmonious) and comparing the computer
6 mm cortical breach, Grade E: screws with more than software plan with the post-screw placement fluoroscopic
6 mm cortical breach (Grade A and Grade B were con- views. In cases with post-screw 3-dimensional imaging,
sidered accurate). Fisher exact tests were used to assess the 100% of screws (n = 197) graded with complete contain-
relationship between procedural techniques and lateral ment (Grade A).
screw malposition. Registration failures occurred in seven cases (4%), all
Study data were collected and managed using of which were due to loss of registration noted on safety
REDCap electronic data capture tools hosted at Boston checks done before drilling at each planned vertebral level
Children’s Hospital.13,14 REDCap (Research Electronic (Table 2).
Data Capture) is a secure, web-based software platform Two major workflow changes took place throughout
designed to support data capture for research studies, this study and were recorded. The first workflow change
providing (1) an intuitive interface for validated data (Change 1, last 74 cases) was the transition to use of a
capture; (2) audit trails for tracking data manipulation high-speed navigated drill, which is a high-speed and low
and export procedures; (3) automated export procedures torque drill with more reliable technique, especially with
for seamless data downloads to common statistical pack- respect to the in-out-in pedicles with Type D morphology.
ages; and (4) procedures for data integration and inter- Adoption of Change 1 was less likely to result in screw
operability with external sources. malposition, as no screws skived lateral with the technique
(P = 0.03, Fisher exact test) (Fig. 1).
The second workflow technique (Change 2, last 39
RESULTS cases) involved a change from the original practice of drilling
One hundred sixty-two patients operated on with
RAN technology with a mean age at surgery of 15.1 years
(range, 4 to 31 y) were identified. The most common TABLE 2. Summary of Cohort Outcomes (N = 162)
diagnosis was adolescent idiopathic scoliosis (n = 80) with Characteristic Freq. (%)
an average major curve of 65 degrees (range, 40 to 128
degrees). The remaining diagnoses included spondylolis- Total surgical time [min; mean (IQR)] 294.79 (226-348)
Total RAN time [min; mean (IQR)] 52.30 (38-64)
thesis, congenital scoliosis, neuromuscular scoliosis, tu- Successfully executed screws 1461 99.6
mor, and others, including kyphosis (Table 1). Return to operating room for screw malposition 0 0
The mean number of levels fused was 9.1 (range, 1 to Technical difficulties 2 1
18) with a mean surgical time of 294.8 minutes (range, 105 Loss of registration 7 4
Intraoperative complications 1 1
to 636 min). Methods of RAN registration included in- Neurological deficits 0 0
traoperative O-arm scan (Scan & Plan) in 131 (81%) pa- EBL [ccs; mean (IQR)] 326.64 (100-450)
tients and preoperative CT to intraoperative fluoroscopic
EBL indicates estimated blood loss; Freq., frequency; IQR, interquartile range;
registration in 31 (19%) patients. The patients with the RAN, robotic-assisted navigation.
latter RAN registration method received standard of care

FIGURE 1. Presence of screw malposition before and after the incorporation of the high-speed navigated drill. Before the use of the
high-speed drill, 6 screws were observed to be malpositioned. With the use of the high-speed drill, no screw malposition occurred.
each pilot hole and then tapping the hole if needed, followed increased accuracy of screw placement compared with
immediately by placing the screw. This practice was changed standard freehand techniques.5,6
to drilling the pilot hole (and tapping if needed) at each level The coupling of robotics and CAN follows the natural
through the robotic arm and moving to the next level without evolution of enabling technology and the promise of enhanced
screw placement. Once the pilot holes were prepared, the ro- safety, accuracy, and improved ergonomics. These platforms
botic arm was removed from the field and screws were se- became available in the United States in the spring of 2019
on 04/15/2023
quentially placed. This workflow change was introduced to and include sophisticated preoperative planning software to
prevent the significant vertebral and spine motion occurring plan screw placement at each level, robotic arm technology to
with the placement of ideal screws in healthy bone, sub- guide accurate trajectory with minimal potential for dexterity
sequently disrupting registration. Change 2 trended toward errors, and navigation to confirm accurate instrument and
statistical significance for avoidance of both screw malposition implant placement (Image 3). Although there have been few
and loss of RAN registration, as no loss of registration oc- reports of this technology in the literature over recent years,
curred after adopting this technique (Fig. 2). In addition, it is those published yielded excellent screw placement accuracy.
now the practice of the senior author to probe all drilled and One of these reports was published by the study institution,
tapped screw tracts before placing the screw using the where the first 40 patients who underwent spine surgery with
navigated passive planar probe to confer correct trajectory adoption of RAN technology demonstrated a screw execution
has been drilled. rate of 98.7% and no complications.9 The experience reported
here is one of execution. Postoperative O-arm and CT scans
are not routinely done at this institution, so only a small subset
(100%) of the total screws placed robotically could be
DISCUSSION analyzed for true accuracy.
Accurate placement of pedicle screws is necessary to The intraoperative experience in a larger cohort of
avoid neurological or visceral injuries and to avoid po- patients encompassing a variety of diagnoses and repre-
tential unplanned return to the operating room. Initially, senting a heterogeneous patient population, which is more
thoracic pedicle screws in pediatric deformity were in- illustrative of a pediatric spine deformity practice, is de-
serted using freehand techniques, yielding an accuracy rate scribed by this study. This report also allows for expansion
of 90%.1 Enabling technologies have been described in of workflow description, which has aided in improvement in
detail for placement of screws in the pediatric population. screw placement and maintaining accurate registration
The original robotic platforms have been reported in the throughout cases. A high rate of screw execution, especially
literature to be associated with a 98% accuracy.8 CAN has after adoption of the use of a high-speed navigated drill,
been studied extensively in the pediatric literature with the was experienced. This execution is a proxy for
advantage of diminished return to the operating room and accuracy in the majority of screws placed, as intraoperative
FIGURE 2. Presence of loss of registration before and after the change in screw insertion workflow. Although the screws were
placed through the robotic arm, 6 screws were observed to be malpositioned. When the screws were inserted after the robotic arm
was removed from the field, no loss of registration occurred.

J Pediatr Orthop Volume 43, Number 5, May/June 2023 Use of RAN in Pediatric Spine Surgery
use as is the case with many enabling technologies.

Training of surgeons and surgical staff at the study in-
stitution was done by industry with the initial launch of
the platform, and thus required no additional costs. As
previously mentioned, the senior author’s preferred
method of registration is with an O-arm, which the study

institution had before adoption of robotics. Institutions

that do not have 3-dimensional intraoperative registra-
tion technology may require this as an additional pur-
chase. There are also additional costs with
instrumentation compatible with the system as well as
the disposable costs related to arrays for the navigation
and draping the robotic arm. The cost savings with ro-
IMAGE 3. Insertion of pedicle screw through the robotic arm. botics has yet to be determined. As these platforms be-
come more accurate and surgical efficiency increases,
O-arm scans or CT scans in patients after screw placement then the potential costs savings will be tied to avoiding
is not routinely obtained due to radiation. The literature on unplanned return to the operating room, diminished
CAN suggests that this execution is correlated with accu- surgical time, and efficient use of optimally placed screws
racy. Only 6 screws were noted lateral, all of which were which may in turn decrease implant density.
placed in hypoplastic pedicles and all with an in-out-in
on 04/15/2023
Enabling technologies continue to expand in medi-

technique. The ideal indication for RAN is the ability to cine, with RAN being the most recent technique aimed at
place in-out-in pedicle screws in difficult anatomy, as the improving accuracy and efficiency with screw placement.
software allows optimal planning of these screws to assure In our report, RAN was an efficient and safe means of
the drill and screw path maintain medial cortical integrity pedicle screw placement in a representative cohort of pe-
and allow for the endpoint to be in the vertebral body. diatric patients. With increasing experience, workflow
These lateral screws were easily seen on fluoroscopic check changes were adopted to diminish the potential for skiving
of screw placement. It is believed that the lateral position of and registration loss, 2 drawbacks which have plagued
these screws was due to a failure to maintain the trajectory robotics in the past. Successful adoption of this technology
as the screw tip re-engaged bone in the lateral portion of the is possible and should be considered for spinal deformity
vertebral body and was pushed lateral. The use of the high- surgeons.
speed navigated drill allows for a clean drill hole and for the
drill bit to stay down its true path when re-entering the
vertebral body. In these type of pedicles, it is also routine to REFERENCES
1. Lehman RA Jr, Lenke LG, Keeler KA, et al. Computed
tap before placing the screws to help maintain trajectory. tomography evaluation of pedicle screws placed in the pediatric
Workflow changes also diminished any chance of deformed spine over an 8-year period. Spine (Phila Pa 1976).
skiving and any chance of registration loss, 2 inherent 2007;32:2679–2684.
challenges with robotics.10,15 The addition of the high- 2. Shillingford JN, Laratta JL, Sarpong NO, et al. Instrumentation
speed drill allows for a direct cutting path through the complication rates following spine surgery: a report from the
Scoliosis Research Society (SRS) morbidity and mortality database.
bone at entry given the high speed (75,000 revolutions per J Spine Surg. 2019;5:110–115.
minute) compared with the previous navigated drill bit 3. Sarwahi V, Suggs W, Wollowick AL, et al. Pedicle screws adjacent to
that was run at 200 rpms. The speed and cutting path at the great vessels or viscera: a study of 2132 pedicle screws in pediatric
entry lessens the chance for the drill bit to skive as long as spine deformity. J Spinal Disord Tech. 2014;27:64–69.
4. Larson AN, Santos ERG, Polly DW, et al. Pediatric pedicle screw
the revolutions are started before hitting the bone, which placement using intraoperative computed tomography and 3-dimen-
in essence is “hitting the ground running.” sional image-guided navigation. Spine (Phila Pa 1976). 2012;37:
The second workflow technique changed was to drill E188–E194.
(and tap if necessary) the pilot holes at each trajectory 5. Baky FJ, Milbrandt T, Echternacht S, et al. Intraoperative computed
before placing the screws. The robotic arm was then re- tomography-guided navigation for pediatric spine patients reduced
return to operating room for screw malposition compared with
moved from the field and screws were placed at each level. freehand/fluoroscopic techniques. Spine Deform. 2019;7:577–581.
This change was made to counteract the potentially great 6. Baldwin KD, Kadiyala M, Talwar D, et al. Does intraoperative CT
insertional torque observed when placing the screw, re- navigation increase the accuracy of pedicle screw placement in
sulting in inadvertent motion of the surrounding segments. pediatric spinal deformity surgery? A systematic review and meta-
analysis. Spine Deform. 2021;10:19–29.
Such scenarios may result in loss of registration of the 7. Devito DP, Woo R. History and evolution of spinal robotics
robotic system and subsequent inefficiency. in pediatric spinal deformity. Int J Spine Surg. 2021;15(s2):
The adoption of robotics coupled with navigation S65–S73.
is associated with some costs to the institution, and 8. Devito DP, Kaplan L, Dietl R, et al. Clinical acceptance and
whereas a thorough discussion is out of the realm of this accuracy assessment of spinal implants guided with SpineAssist
surgical robot: retrospective study. Spine (Phila Pa 1976). 2010;35:
paper, it certainly warrants mention. The initial costs 2109–2115.
include the purchase of the robotic platform, which can 9. Gonzalez D, Ghessese S, Cook D, et al. Initial intraoperative experience
be done either as a capital purchase or tied to implant with robotic-assisted pedicle screw placement with stealth navigation in

pediatric spine deformity: an evaluation of the first 40 cases. J Robot Surg. 13. Harris PA, Taylor R, Thielke R, et al. Research electronic data
2021;15:687–693. capture (REDCap)—A metadata-driven methodology and workflow
10. Lieberman IH, Kisinde S, Hesselbacher S. Robotic-assisted pedicle screw process for providing translational research informatics support. J
placement during spine surgery. JBJS Essent Surg Tech. 2020;10:e0020. Biomed Inform. 2009;42:377–381.
11. Sawires AN, Birch CM, Hedequist D. The use of robotics coupled 14. Harris PA, Taylor R, Minor BL, et al. REDCap Consortium, The
with navigation for pediatric congenital spine deformity. HSS J. REDCap consortium: Building an international community of
2021;17:289–293. software partners. J Biomed Inform. 2019;95:103208.

12. Su AW, Luo TD, McIntosh AL, et al. Switching to a pediatric Dose 15. Hu X, Ohnmeiss DD, Lieberman IH. Robotic-assisted pedicle screw
O-Arm protocol in spine surgery significantly reduced patient placement: lessons learned from the first 102 patients. Eur Spine J.
radiation exposure. J Pediatr Orthop. 2016;36:621–626. 2013;22:661–666.
on 04/15/2023

ORIGINAL ARTICLE
Scoliosis and Kyphosis Prevalence in Turner Syndrome:

A Retrospective Review at a Pediatric Tertiary Care
Medical Center
Jeremy S. Marx, MD,*† Manasa Pagadala, BA,*† John Carney, MD,*† Erik Gerlach, MD,*†
Peter Swiatek, MD,*† Jennifer A. Zabinsky, MD,‡§‖ John Sarwark, MD,*
Wendy J. Brickman, MD,‡§ and Reema L. Habiby, MD‡§
Level of Evidence: Level III.

Background: The prevalence of major coronal and sagittal spinal Clinical Relevance: This retrospective case series serves to review
curves (scoliosis and kyphosis) in Turner syndrome (TS) is not and address the prevalence of spinal deformity in TS patients and
well established due to limited reporting. The relationship be- whether GH impacts worsening deformity.
on 04/15/2023
tween growth hormone (GH) therapy and its effect on TS spinal

curve incidence is also not well established. Key Words: turner syndrome, spinal deformity, scoliosis, ky-
Methods: A retrospective chart review of 306 TS patients from phosis, growth hormone
2007 to 2021 evaluated major coronal and sagittal spinal curves, (J Pediatr Orthop 2023;43:299–302)
progression of the curve, and treatment with GH. Statistical
significance (defined as P <0.05) between curvature rates and
curve progression was compared between GH-treated patients
and non–GH-treated patients using a χ2 or Fisher exact test
when appropriate. T urner syndrome (TS) occurs in ~1 in 2500 live births
and is due to the complete or partial absence of 1 sex
chromosome.1 Skeletal deformity in TS has been ex-
Results: Thirty-seven of 306 (12%) TS patients had a radio-
graphically relevant spinal deformity. Twenty-seven of 37 (73%) tensively described. These patients present to orthopaedic
had mild; 4 of 37 (11%) had moderate, and 6 of 37 (16%) had specialists for evaluation and management of spine
severe curves. Of those with severe, 4 underwent spinal fusion, 1 and musculoskeletal concerns (scoliosis, thoracic hyper-
was treated with bracing, and 1 was braced before a car- kyphosis, slipped capital femoral epiphysis, Madelung
diovascular-related death. Regarding GH use among TS pa- deformity, and fragility fractures later in life.1) A major
tients, 190 of 306 (62%) used GH versus 116 of 306 (38%) who coronal curve (scoliosis) is defined by age of onset and
did not. Of those with a spinal curve, 24 of 37 (65%) used GH etiology of the curve, and it is recognized that curves in TS
compared with 13 of 37 (35%) who did not. On univariate patients align most closely with idiopathic scoliosis.1,2
analysis, GH therapy was not a risk factor for the diagnosis of a Spinal curves seem to occur more frequently in pa-
major spinal curve, a more severe degree of the curve at the time tients with TS compared with the general population. In
of diagnosis, or spinal curve progression (P > 0.05 for all). the general population, 1% to 4% of children aged 10 to
Conclusions: This is the largest single institution retrospective 16 years will have adolescent idiopathic scoliosis (AIS),
review of a TS cohort known to the authors assessing spinal whereas in TS populations the literature widely varies with
curve prevalence and relation to GH treatment and demonstrates a range of 10% to 59%.2–8 Data pooled from 3 case series
a TS spinal curve rate of 12% (37/306). Four of six (11%) TS and an Australian cohort study that totaled 205 TS pa-
patients with a severe curve underwent corrective spine fusion. tients demonstrated 28% with spinal curves.4–6,8
There was no relationship between the use of GH and the The relationship between growth hormone (GH)
presence of a spinal curve or curve progression. Further study is therapy and the incidence and clinical course of spinal
warranted to determine risk factors for curve progression. curves in TS is inconclusive. GH is indicated for TS pa-
tients at early ages (around 4 to 6 y of age, preferably
before 12 to 13 y) for short stature and growth failure.9
From the *Division of Orthopaedics; ‡Division of Endocrinology, Ann GH efficacy is monitored by height measurement every
and Robert H. Lurie Children’s Hospital of Chicago; Departments of
†Orthopaedics; §Pediatrics, Northwestern University Feinberg School
6 months to verify the appropriate response.9 Other pos-
of Medicine, Chicago, IL; and ‖Division of Endocrinology, UCSF itive impacts of GH include a lower body mass index, a
Department of Pediatrics, San Francisco, CA. more favorable lipid profile, and lower rates of
The authors declare no conflicts of interest. Hypertension.10 Negative GH effects can include elevated
Reprints: Jeremy S. Marx, MD, Arkes Family Pavilion, 676 N. Saint triglycerides, retrognathism, nail anomalies, and higher
Clair, Suite 1350, Chicago, IL 60611. E-mail: jeremy.marx@north
western.edu. Follicle-Stimulating Hormone when compared with those
Copyright © 2023 Wolters Kluwer Health, Inc. All rights reserved. not exposed to GH therapy.10 Reports have also suggested
DOI: 10.1097/BPO.0000000000002367 that GH use may either cause or accelerate preexisting

Marx et al J Pediatr Orthop Volume 43, Number 5, May/June 2023
curvatures in idiopathic scoliosis, yet extensive data Moderate coronal curves were defined as 26 to 45 de-
specific to GH’s effect on TS patients is limited.11–13 Day grees. Severe coronal curves were defined as > 46 de-
et al,14 in a study of Australian TS patients treated with grees. Hyperkyphosis describes curvatures in the sagittal
GH, reported an increased spinal curve prevalence of 29% plane. Mild hyperkyphosis was defined as 46 to 59 de-
compared with prior reports of ~10%. The authors con- grees, moderate as 60 to 74 degrees, and severe as 75
cluded that either spinal curves were more common in TS degrees or greater. Patients who had a clinical diagnosis
than previously believed or that GH may lead to increased of a spinal curve, without appropriate radiographic
development. imaging, were not considered as having a relevant di-
The purpose of this study was to evaluate the prev- agnosis and were therefore excluded.
alence of coronal and sagittal curves in TS patients at a The identified charts with spinal curves were
large academic institution, with limited reporting in the searched for GH prescriptions or clinical notes that veri-
literature, and report outcomes of the orthopaedic man- fied GH use. If GH use was not documented, then patients
agement of these patients based on their degree of a curve were considered as not having received GH therapy. Age
at presentation. We also evaluated whether there was an at curve diagnosis was reviewed and annotated for each
association between GH treatment of TS individuals and patient.
the presence and progression of spinal curves. Statistical significance between curve rates and curve
progression was compared between GH-treated patients
and non–GH-treated patients using a χ2 or Fisher exact
METHODS test for expected values of <5. Statistical significance was
defined as P <0.05.
on 04/15/2023
The study was approved by the Institutional Review

Board for our large, single academic pediatric institution.
The requirement for informed consent was waived for this RESULTS
retrospective study. We queried the hospital’s electronic
medical record (EPIC) for patients diagnosed with TS Coronal or Sagittal Plane Curves
from 2007 to 2021, for ICD9 and ICD10 codes of 758.6 The initial query resulted in 555 patients. After in-
and q96.9, representing gonadal dysgenesis and TS clusion and exclusion criteria were applied, the final TS
diagnoses. cohort consisted of 306 patients. Of this cohort, 37 of 306
Exclusion criteria included patients with male phe- (12%) had a coronal or sagittal spinal curve confirmed by
notype, nonbinary genitalia, mixed gonadal dysgenesis, radiographs. Of the 37 TS patients, 27 (73%) had mild
those that did not have a confirmed TS karyotype or those curves, 1 with mild kyphosis and 26 with mild scoliosis; 4
who were not adequately evaluated for spinal curves. (11%) had moderate curves, 1 with moderate kyphosis and
The diagnosis of a coronal curve (scoliosis) was 3 with moderate scoliosis; and 6 (16%) had severe
determined radiographically by using the Cobb angle scoliosis. The 27 patients with mild curves were observed.
technique at ≥ 10 degrees in the coronal plane. Mild Of the 4 with moderate curves, 1 received the bracing
coronal curves were defined as 10 to 25 degrees. intervention, 1 was advised bracing but did not comply,
FIGURE 1. Turner syndrome patient with severe thoracic right coronal curve (64 degrees, T4 to T10); pre and postfusion X-ray.

J Pediatr Orthop Volume 43, Number 5, May/June 2023 Scoliosis and Kyphosis Prevalence in TS
and 2 were observed. Of the 6 with severe coronal curves, Pooling the data from 3 case series and an Australian
4 underwent corrective spine fusion surgery, 1 was advised cohort study, 28% of 205 TS patients had a major coronal
spinal fusion but did not consent and was treated curve.4–6,8 In our retrospective review of 306 patients with
with bracing, and the final patient was braced before TS, we found a 12% spinal curve prevalence, which falls
cardiovascular-related death. Figure 1 demonstrates a pre within the range of prior reported rates, though previously
and postoperative radiograph of a TS patient who had a reported rates were widely varied and based on small
severe coronal curve that required subsequent posterior numbers.1 In addition, in the general population, about
spinal fusion with instrumentation. 10% of adolescents with AIS will progress to a severity
The average age of diagnosis for the cohort of TS that requires consideration of surgery.2 In our study of TS
patients with a spinal curve was 10.6 ± 4.4 years with a patients with spinal curves, 16% had progression deemed
median of 12 years, and a mode of 14 years. The average eligible for surgery, and 11% underwent fusion. This is in
age for TS patients with mild curves was 11.7 ± 3 years, contrast to pooled TS studies that reported a 7% rate of
median of 13 years, and mode of 14 years. The average fusion.4–6,8
age for moderate curves was 10.0 ± 6.8 years, median of During chart review, we identified some patients
12, and mode of 13 years. The average age of severe curves diagnosed with spinal curves clinically, without radio-
was 6.2 ± 5.7 years, median of 7 years, and mode of graphic confirmation. The rate of spinal curves in our
11 years. cohort would have increased to 20.38% if those diagnosed
only clinically were included. This is relevant as prior
Growth Hormone Treatment studies did not always define their inclusion criteria with
Regarding GH therapy, 190 of 306 (62%) TS pa-
on 04/15/2023
strict radiographic examination as inclusion criteria for

tients received GH. The cohort consisted of 269 of 306 diagnosis.6,8 This also calls into question the specificity of
(88%) TS patients without spinal curves, and of these, 166 a clinical examination versus a radiographic evaluation.
of 269 (62%) received GH therapy and 103 of 269 (38%) Table 2 summarizes the pooled studies compared with
did not. For TS patients with a curve, 24 of 37 (65%) data from our retrospective review.
received GH and 13 of 37 (35%) did not. Regarding the
severity of the curve and GH, 17 of 27 (63%) with mild
curves, 3 of 4 (75%) with moderate curves, and 4 of 6 Growth Hormone
(67%) with severe curves received GH. Rates of curves The relationship between GH therapy and the in-
were not found to be significantly different between GH cidence of spinal curves in TS has been inconclusive. Re-
and non-GH users (Table 1, P = 0.71). In addition, GH ports have suggested that the use of GH may either cause
use was not associated with a higher grade of curve or accelerate preexisting scoliosis in AIS,11–13 yet extensive
severity (Table 1, P = 0.99) or with increased rates of data specific to GH’s effect on TS patients has been lim-
curve progression (Table 1, P = 1.00). ited. Day et al14 studied a group of Australian patients
with TS being treated with GH. Their cohort was found to
DISCUSSION have increased rates of spinal curves at 29% compared
with prior reports of ~10%.14 The authors concluded
Coronal and Sagittal Curves that the curves were more common in TS than previously
To our knowledge, this is the largest retrospective believed or that GH may lead to an increased
case review of the prevalence of spinal curvature in TS development.14 However, as the reporting of the preva-
patients. In our study, curves occurred more frequently in lence of curves in TS patients is so widely variable, no
patients with TS compared with the general population. In conclusions have been made from this data.
the general population, 1% to 4% of children aged 10 to Our findings demonstrated that GH use was not a
16 years will have AIS, whereas in TS populations the risk factor for developing curves (P = 0.71), being diag-
literature widely varies with a range of 10% to 59%.2–8 nosed with a higher grade of curvature (P = 0.99), or
worsening of curvature compared with non-GH users
(P = 1.00).
TABLE 1. TS Patient GH Use, Spinal Curvature, and
Progression
TS Patients n = 306 GH Use No GH P Limitations
As in any retrospective chart review, there are limi-
190/306 (62.1) 116/306 (38.0) —
Normal (n = 269) 166/269 (61.7) 103/269 (38.3) 0.71 tations to our study. First, not all patients with a clinical
Spinal curves (n = 37) 24/37 (64.9) 13/37 (35.1) — diagnosis of a spinal curve received appropriate imaging
TS curves n = 37 and were therefore excluded from our study, likely
24/37 (65.0) 13/37 (35.1) — underestimating the prevalence in our TS cohort. In ad-
Mild curve (n = 27) 17/27 (63.0) 10/27 (37.0) 0.99
Moderate curve (n = 4) 3/4 (75.0) 1/4 (25.0) —
dition, not all TS patients were seen in the Endocrine or
Severe curve (n = 6) 4/6 (66.7) 2/6 (33.3) — Orthopaedic clinics, and therefore a diagnosis of a mild
No progression (n = 27) 17/27 (63.0) 10/27 (37.0) 1.0 curve may have been missed. In addition, estrogen
Curve progression (n = 10) 7/10 (70.0) 3/10 (30.0) — replacement was not assessed among the cohort, which
GH indicates growth hormone; TS, Turner syndrome. could be an associated risk factor for spinal curvature
progression or worsening.

Marx et al J Pediatr Orthop Volume 43, Number 5, May/June 2023
There was no statistical significance in GH use in either the

TABLE 2. Published TS Patient Pooled Data vs. Single Institution
Retrospective Review development or worsening of curves. Further study is war-
ranted to determine the risk factors for curve progression.
TS Patients Curves Brace Fusion
Cohort retrospective 306 37/306 3/37 4/37 REFERENCES
review (12.1) (8.1) (10.8) 1. Acosta AM, Steinman SE, White KK. Orthopaedic manifesta-
Ricotti 20116 49 29 0 0 tions in Turner syndrome. J Am Acad Orthop Surg. 2019;27:

Day 20048 88 18 1 1 e1021–e1028.
Kim 20014 43 5 2 2 2. Altaf F, Gibson A, Dannawi Z, et al. Adolescent idiopathic scoliosis.
Elder 20025 25 5 2 1 BMJ. 2013;346:1–7.
Pooled total 205 57/205 5/57 4/57 3. Reamy BV, Slakey JB. Adolescent idiopathic scoliosis: review and
(27.8) (8.8) (7.0) current concepts. Am Fam Physician. 2001;64:111–116.
4. Kim JY, Rosenfeld SR, Keyak JH. Increased prevalence of scoliosis
TS indicates Turner syndrome.
in Turner syndrome. J Pediatr Orthop. 2001;21:765–766.
5. Elder DA, Roper MG, Henderson RC, et al. Kyphosis in a Turner
syndrome population. Pediatrics. 2002;109:e93–e97.
6. Ricotti S, Petrucci L, Carenzio G, et al. Prevalence and incidence of
Moreover, patients with inadequate follow-up were scoliosis in Turner Syndrome: a study in 49 girls followed-up for
4 years. Eur J Phys Rehabil Med. 2011;47:447–453.
excluded from the cohort analysis and may have even- 7. Bolar K, Hoffman AR, Maneatis T, et al. Long-term safety of
tually developed a spinal curvature. This effect on the final recombinant human growth hormone in Turner syndrome. J Clin
prevalence rate is therefore uncertain. Some patients who Endocrinol Metab. 2008;93:344–351.
8. Day GA, McPhee IB, Batch J. The incidence of idiopathic scoliosis
on 04/15/2023
were diagnosed with mild or moderate curves during their

in Turner syndrome—growth hormone treated and non-treated.
juvenile years were lost to follow-up, and it is unknown if Orthop Proc. 2004;86-B:455.
there was a progression. Also, a small number of TS pa- 9. Gravholt CH, Andersen NH, Conway GS, et al. Clinical practice
tients passed away during their childhood and adolescent guidelines for the care of girls and women with Turner syndrome:
years, some of whom had spinal curves; therefore it is proceedings from the 2016 Cincinnati International Turner
unclear if they would have had curve progression requir- Syndrome Meeting. Eur J Endocrinol. 2017;177:G1–G70.
10. Irzyniec T, Jeż W, Lepska K, et al. Childhood growth hormone
ing spinal fusion. Finally, many TS patients are still in- treatment in women with Turner syndrome—benefits and adverse
fants, juveniles, or adolescents and their curve magnitudes effects. Sci Rep. 2019;9:15951.
are not yet fully realized. 11. Ahn UM, Ahn NU, Nallamshetty L, et al. The etiology of adolescent
idiopathic scoliosis. Am J Orthop. 2002;31:387–395.
12. Burwell RG. Aetiology of idiopathic scoliosis: current concepts.
CONCLUSIONS Pediatr Rehabil. 2003;6:137–170.
This is the largest retrospective review of a TS cohort 13. Wang ED, Drummond DS, Dormans JP, et al. Scoliosis in
(including 306 patients) assessing for the prevalence of coronal patients treated with growth hormone. J Pediatr Orthop. 1997;17:
and sagittal spinal curves and concomitant GH therapy 708–711.
14. Day G, Szvetko A, Griffiths L, et al. SHOX gene is expressed in
known by the authors. Our study demonstrated a spinal curve vertebral body growth plates in idiopathic and congenital scoliosis:
prevalence rate of 12%, which is consistent with prior Implications for the etiology of scoliosis in Turner syndrome.
reporting.1 Our TS cohort demonstrated a fusion rate of 11%. J Orthop Res. 2009;27:807–813.

ORIGINAL ARTICLE
Skeletal Maturity in Legg-Calve-Perthes Disease: Significant

Discrepancy Present Between the Hand and the Hip
K. Aaron Shaw, DO and John A. Herring, MD
Oxford bone age more closely mirrors the patient’s CA and does
Background: The concept of delayed skeletal maturity in Legg- not correlate with the GP bone age, which reveals a delayed
Calve-Perthes Disease (LCPD) has been well identified with the maturation.
Greulich and Pyle (GP) atlas showing 1 to 2 years delay. Re-
cently the optimized Oxford hip skeletal age (Optimized Oxford) Key Words: Legg-Calve-Perthes, skeletal maturity, bone age,
system has been developed and shown to have similar accuracy optimized Oxford
as the GP atlas for assessing skeletal maturity. However, this (J Pediatr Orthop 2023;43:294–298)
system has not been used to assess skeletal maturity in LCPD.
Methods: A retrospective review of a prospective, multicenter
on 04/15/2023
study of patients with LCPD treated from 1984 to 1991 and

followed to skeletal maturity was performed. We identified all
patients who had a left-hand radiograph at the time of pre-
sentation with an accompanying anteroposterior pelvis radio-
L egg-Calve-Perthes Disease (LCPD) remains an elusive
condition for the practicing pediatric orthopaedic sur-
geon. Although the etiology remains largely unknown,
graph including the contralateral hip. Patients were excluded if previous studies have demonstrated that immaturity at the
their age at presentation fell outside the validated range for the time of disease onset is one factor associated with im-
Optimized Oxford system. GP atlas was used to determine bone proved femoral head shape and functional outcomes at
age using left-hand radiographs and the nonaffected hip radio- skeletal maturity.1–3 Herring et al1 reported that children
graphs were used to calculate the Optimized Oxford bone age. 8 years and younger had superior outcomes compared
Skeletal maturity indices were compared with chronological age with children older than 8 years at the time of pre-
(CA) to determine the discrepancy between methodologies. sentation. Subsequent studies have suggested that the op-
Results: A total of 71 patients met inclusion criteria (mean 9.5 ± timal age for a favorable outcome at maturity is 6 years or
1.2 y at presentation, 42.2% females). The mean GP bone age younger at presentation.4,5
was 1.4 years younger than CA (95% CI: 1.01-1.76 y), with the However, chronologic age does not necessarily
discrepancy being greater for boys than girls (1.8 vs 0.86 y, P = equate to skeletal age, particularly in LCPD where there is
0.02). The mean Optimized Oxford bone age was 0.31 years older a known association with delayed skeletal maturity of up
than CA (95% CI: 0.24-0.38 y) and correlated significantly with to 1.9 years in affected children.1,6 These data have been
CA (R = 0.97, P < 0.001). There were no sex differences in the based on maturity assessment with the Greulich and Pyle
Optimized Oxford bone age relative to CA (P = 0.32). The GP (GP) maturity atlas. Although the GP atlas remains the
bone age was a mean of 1.7 years younger than the Optimized accepted standard for assessing bone age, this requires
Oxford bone age (95% CI: 1.35-2.05 y). obtaining a separate radiograph of the hand. Recently, the
Conclusion: Skeletal maturity assessment in children with LCPD optimized Oxford hip skeletal age system (Optimized
varies according to the utilized maturity system. The Optimized Oxford) has been introduced and shown to be comparable
for predicting 90% of adult standing height in comparison
to the GP maturity atlas.7 To date, however, this
From the Department of Pediatric Orthopaedic Surgery, Scottish Rite for
Children Hospital, Dallas, TX.
methodology has not been used to assess skeletal maturity
The authors declare no funding for this work. in children with LCPD.
Approved by IRB: STU 2021-0603. In this study, we sought to investigate the use of the
K.A.S is a committee member for NASS and AAOS; J.A.H reports Optimized Oxford system in the determination of skeletal
receiving publishing royalties from Elsevier, IP royalties from Med- maturity in children with LCPD. We hypothesized that
tronic, is a board/committee member for GSSG, POSNA, and SRS,
and editorial board member for JBJS, Spine, and JPO, and is an the Optimized Oxford system would provide a comparable
unpaid consultant for OrthoPediatrics. maturity assessment to the GP maturity atlas.
Reprints: K. Aaron Shaw, DO, Department of Pediatric Orthopaedic
Surgery, Scottish Rite for Children Hospital, 2222 Welborn St,
Dallas, TX 75219. E-mail: aaron.shaw@tsrh.org. METHODS
Supplemental Digital Content is available for this article. Direct URL A retrospective review of a prospective, multicenter
citations appear in the printed text and are provided in the HTML study of patients with LCPD treated from 1984 to 1991
pedorthopaedics.com. and followed to skeletal maturity was performed. The
Copyright © 2023 Wolters Kluwer Health, Inc. All rights reserved. study was approved by the Institutional Review Board.
DOI: 10.1097/BPO.0000000000002368 Inclusionary criteria consisted of patients with LCPD

J Pediatr Orthop Volume 43, Number 5, May/June 2023 Skeletal Maturity in Legg-Calve-Perthes Disease
diagnosed on anteroposterior pelvis radiographs that in-

cluded the contralateral hip who had bone age radio-
graphs obtained at the time of presentation. Patients were
excluded from participation if they presented with
bilateral LCPD, the contralateral hip was not visualized
on their presenting radiographs, or they lacked bone age

radiographs at the time of presentation. In addition, we

excluded girls who presented before age 7 years and boys
presenting before age 9 as the Optimized Oxford bone age
has only been validated for children above these age
thresholds.8
Skeletal Maturity Assessment
Skeletal maturity was assessed with both the GP and
the Optimized Oxford maturity systems. The GP bone age
evaluation was performed by the senior author, a pediatric
orthopaedic surgeon with extensive familiarity with the FIGURE 1. Histogram depicting Greulich and Pyle (GP) and
system. The GP atlas was set as the standard for skeletal optimized Oxford bone ages relative to chronologic age for
maturity assessment given its previous utilization in children with Legg-Calve-Perthes Disease (LCPD).
LCPD for characterizing delayed skeletal maturation.1,6
on 04/15/2023
The Optimized Oxford system was performed using the

methodology described by Furdock et al.7 The system Optimized Oxford assessment leaving a total of 71 patients
consists of an assessment of radiographic parameters for for inclusion (mean 9.5 ± 1.2 y at presentation, 42.2%
the femoral head, greater trochanter, ilium, triradiate females). The majority of children were classified as
cartilage, and the height of the greater trochanter relative modified lateral pillar B (N = 47, 66%), followed by B/C
to the femoral head diameter based upon the modified (N = 14), and C (N = 10).
Oxford system (Supplemental Digital Content 1, http:// Skeletal maturity assessment by the GP bone age
links.lww.com/BPO/A584). The results of these parame- method demonstrated a mean discrepancy of 1.4 years
ters are then combined into validated equations with the younger than CA (95% CI: 1.01-1.76 y) whereas the Op-
patient’s chronologic age to determine their skeletal ma- timized Oxford bone age demonstrated a mean discrep-
turity. Assessment of the hip radiographic parameters was ancy of 0.31 years older than CA (95% CI: 0.24-0.38 y).
performed on the contralateral hip with the use of PACS The GP bone age was a mean of 1.7 years younger than
software (IDS 7 Sectra, Sweden). Intra and interrater re- the Optimized Oxford bone age (95% CI: 1.35-2.05 y),
liability was assessed with 6 weeks in between measure- Figure 1. The Optimized Oxford bone age correlated
ments indicating excellent intrarater (ICC 0.93) and significantly with CA, (R = 0.97, P < 0.001). Male
interrater reliability (ICC 0.84). patients had a significantly younger GP bone age relative
to CA compared with female patients (1.8 y vs 0.86, P =
Statistical Analysis
0.02), however, there were no sex differences in the
Statistical analysis was performed using SPSS Optimized Oxford bone age relative to CA (Male, 0.34 y
software (SPSS version 24, IBM, Chicago, IL). De- older vs 0.26 y older; P = 0.32), Figure 2.
scriptive statistics were generated. GP bone age was used
as the accepted standard. The GP and Optimized Oxford
bone ages were compared with the child’s chronologic DISCUSSION
age to determine mean discrepancy. The Optimized LCPD is a challenging condition, which continues to
Oxford score was compared with the chronological age remain an elusive entity for the treating pediatric ortho-
(CA) and GP bone age using a 2-way Student t Test. paedic surgeon. Maturity, using chronologic age at pre-
Pearson correlation coefficients were also performed by sentation as a proxy, is one factor, which has been shown
comparing the Optimized Oxford score to GP bone to influence patient outcomes with older patients demon-
age and CA. Statistical significance was predetermined as strating worse outcomes at the time of skeletal
P = 0.05. maturity.1–3 In this study, we investigated the utility of the
optimized Oxford system as a means of assessing current
RESULTS skeletal age at the time of presentation in a known cohort
A total of 440 total patients were identified in the of children with LCPD. Our results indicated significant
LCPD repository (80.4% males, mean 7.9 ± 1.4 y). Of variability between the optimized Oxford system and the
these 202 were excluded due to a lack of presenting hand GP maturity atlas with the optimized Oxford closely
radiographs for GP bone age assessment. An additional mirroring chronologic age, demonstrating a 1.7-year older
163 were excluded for not meeting the age threshold for bone age than the GP atlas, Figure 3.
the Optimized Oxford assessment with 4 additional pa- The Optimized Oxford maturity system, similar to
tients excluded due to inadequate radiographs for the the GP bone age, was generated from the Bolton-Brush

Shaw and Herring J Pediatr Orthop Volume 43, Number 5, May/June 2023
on 04/15/2023
FIGURE 2. Histogram depicting skeletal maturity bone age predictions according to sex [(A) female and (B) male], relative to
chronologic age.
Study database based upon serial radiographs of healthy LCPD has a known delay in skeletal maturity rel-
middle to upper-class, primarily white, children in Cleve- ative to chronologic age, ranging from a 1 to 1.9-year
land, OH from 1929 to 1942.9 This score was developed by delay in skeletal maturity as assessed with the GP ma-
Furdock et al7 as an accurate methodology to predict 90% turity atlas.1,6 Herring et al1 found that this delay in
final height in comparison to the GP bone age. The Op- skeletal maturity varied by sex with boys having a
timized Oxford system is a derivation of select parameters greater delayed maturity relative to girls. Similarly, Ki-
from the Achesons10 Oxford method of assessing skeletal toh et al11 reported a delay in the ossification in the
maturity, including the femoral head, greater trochanter, unaffected proximal femoral epiphysis in a series of 125
triradiate cartilage, and ilium components.8 In addition to Japanese children. The ability to utilize radiographic
these oxford parameters, greater trochanteric height ratio parameters from the preexisting pelvis imaging studies is
and patient chronologic age make up the final parameters, advantageous as it mitigates the need for additional
both of which were shown to heighten the accuracy of 90% imaging studies and minimizes resource utilization.
height prediction.8 Nelson et al9 first reported on the However, no study to date has applied a quantitative
utility of greater trochanteric height as an accurate and hip-specific maturity assessment system to assess skeletal
efficient parameter for predicting 90% adult height with maturity in children with LCPD.
Castillo Tafur et al8 showing it to be a resilient parameter In this study of 71 children with LCPD with GP
even in the setting of imperfect lower-extremity rotation. bone ages obtained at the time of diagnosis, we found

J Pediatr Orthop Volume 43, Number 5, May/June 2023 Skeletal Maturity in Legg-Calve-Perthes Disease
FIGURE 3. A 9 + 11-year-old boy with onset of left hip pain and findings consistent with LCPD. Radiographs were consistent with a
bone age of 5 years according to the GP atlas (A) and an optimized Oxford bone age of 9.97 (B).
on 04/15/2023
significant discordance between the optimized Oxford and shown the GP atlas to be valid in estimating final
GP bone ages. The Optimized Oxford system was, on adult height.7 Interestingly, previous studies have
average, 1.7 years older than the GP bone age and was also shown that lower-extremity-based maturity
strongly correlated with patient chronologic age (R = methodologies tend to report further delayed maturation
0.97, P < 0.001). This finding suggests that the inclusion of relative to the GP atlas in cases of delayed skeletal
the patient’s chronologic age in the Optimized Oxford maturity,12 which given the inverse findings in the
maturity system skews the maturity data toward chrono- current study raises further concern regarding the
logic age and in conditions with a known delay in skeletal clinical application of the Optimized Oxford system
maturities, such as LCPD, may not provide an accurate in LCPD.
assessment of the true bone age. As such, we recommend
against the use of the optimized Oxford system as a stand- CONCLUSION
alone assessment of skeletal maturity in children with This study found significant discrepancies with the
LCPD and suggest caution with its application in other application of the Optimized Oxford skeletal maturity
conditions with a known delay or advancement in skeletal assessment system in children with LCPD when compared
maturity. with the GP bone age system. The Optimized Oxford
The results of this study cannot be viewed without system includes the patient’s chronologic age in its deter-
recognition of its limitation. As a retrospective review, mination of maturity and as such, is significantly corre-
there are certain inherent limitations. The data for the lated with chronologic age. In its current form, we
current study were based on a multicenter prospective recommend against the use of the Optimized Oxford sys-
study of children with LCPD treated in the continental tem for the stand-alone assessment of skeletal maturity in
United States and followed to skeletal maturity. The conditions with a known aberration in skeletal maturity.
heterogeneity of the population minimizes potential se- Further studies are needed to assess the utility of hip and
lection bias. However, a significant portion of the pelvic maturity parameters for the assessment of skeletal
available study population had to be excluded due to the maturity in LCPD.
lack of GP bone ages at the time of presentation (45.9%)
with an additional 37% excluded due to their age at
diagnosis being outside the validated age range for ap- REFERENCES
1. Herring JA, Kim HT, Browne R. Legg-Calve-Perthes disease. Part
plication of the Optimized Oxford system (girls <7 y and II: prospective multicenter study of the effect of treatment on
boys <9 y). As such, the current data are based on an outcome. J Bone Joint Surg Am. 2004;86:2121–2134.
older CA cohort of children than the typical LCPD 2. Herring JA, Neustadt JB, Williams JJ, et al. The lateral pillar
population and may introduce selection bias. In addi- classification of Legg-Calvé-Perthes disease. J Pediatr Orthop.
1992;12:143–150.
tion, these patients were all treated between 1984 and 3. Stančák A, Kautzner J, Chládek P, et al. Predictors of radiographic
1991 and may not be directly comparable to a con- outcomes of conservative and surgical treatment of Legg-Calvé-Perthes
temporary patient population as has been reported disease. Int Orthop. 2022;46:2869–2875.
in the skeletal maturity assessment.9 To date, no study 4. Rosenfeld SB, Herring JA, Chao JC. Legg-calve-perthes disease:
has applied a hip-based skeletal maturity system for a review of cases with onset before six years of age. J Bone Joint Surg
Am. 2007;89:2712–2722.
the evaluation of delayed maturation in LCPD, nor 5. Canavese F, Dimeglio A. Perthes’ disease: prognosis in
validated the application of the GP atlas relative to children under six years of age. J Bone Joint Surg Br. 2008;90:
final skeletal maturity. However, previous studies have 940–945.

Shaw and Herring J Pediatr Orthop Volume 43, Number 5, May/June 2023
6. Kristmundsdottir F, Burwell RG, Hall DJ, et al. A longitudinal 9. Nelson G, Knapik DM, Janes JL, et al. Greater trochanter height:
study of carpal bone development in Perthes’ disease: its significance a quantitative predictor of skeletal maturity. J Pediatr Orthop.
for both radiologic standstill and bilateral disease. Clin Orthop Relat 2021;41:99–104.
Res. 1986;209:115–123. 10. Acheson RM. The Oxford method of assessing skeletal maturity.
7. Furdock RJ, Benedick AJ, Nelson G, et al. Systematic Clin Orthop. 1957;10:19–39.
isolation of key parameters for estimating skeletal maturity 11. Kitoh H, Kitakoji T, Katoh M, et al. Delayed ossification of the
on AP hip radiographs. J Pediatr Orthop. 2021;41: proximal capital femoral epiphysis in Legg-Calvé-Perthes’ disease. J
483–489. Bone Joint Surg Br. 2003;85:121–124.

8. Castillo Tafur JC, Furdock RJ, Sattar A, et al. The optimized 12. Aicardi G, Vignolo M, Milani S, et al. Assessment of skeletal
Oxford hip skeletal maturity system proves resilient to rotational maturity of the hand-wrist and knee: a comparison among methods.
variation. J Pediatr Orthop. 2022;42:186–189. Am J Hum Biol. 2000;12:610–615.
on 04/15/2023

ORIGINAL ARTICLE
Subgroups of Idiopathic Clubfoot Can Predict

Short-term Outcomes
Rachal Quinlan, MRes,* Verity Pacey, PhD,* Emre Ilhan, PhD,*† Paul Gibbons, MD,†‡
and Kelly Gray, PhD*
highlighting the clinical benefit of subgrouping to predict out-

Background: The Pirani scale is used for the assessment of comes in Ponseti-managed idiopathic clubfoot.
Ponseti-managed clubfoot. Predicting outcomes using the total Level of Evidence: Level II, prognostic.
Pirani scale score has varied results, however, the prognostic
value of midfoot and hindfoot components remains unknown. Key Words: Clubfoot, Ponseti, Pirani, group- based trajectory
The purpose was to (1) determine the existence of subgroups of modelling
Ponseti-managed idiopathic clubfoot based on the trajectory of (J Pediatr Orthop 2023;43:326–331)
on 04/15/2023
change in midfoot and hindfoot Pirani scale scores, (2) identify

time points, at which subgroups can be distinguished, and (3)
determine whether subgroups are associated with the number of
casts required for correction and need for Achilles tenotomy.
Methods: Medical records of 226 children with 335 idiopathic
clubfeet, over a 12-year period, were reviewed. Group-based
I diopathic clubfoot is an isolated condition in a typically
developing infant, where the foot presents in severe
equinus, adduction, and supination.1,2 Ponseti manage-
trajectory modeling of the Pirani scale midfoot score and hind- ment uses gentle manipulation and serial casting to correct
foot score identified subgroups of clubfoot that followed statis- clubfoot deformity.1 Cavus is initially reduced followed by
tically distinct patterns of change during initial Ponseti correction of the adductus and heel varus deformity.3
management. Generalized estimating equations determined the After casting, an Achilles tenotomy is commonly required
time point, at which subgroups could be distinguished. Com- to correct residual hindfoot deformity.4
parisons between groups were determined using the Kruskal- The Pirani scale is commonly used for initial as-
Wallis test for the number of casts required for correction and sessment and monitoring correction.5,6 Six signs of club-
binary logistic regression analysis for the need for tenotomy. foot contracture are scored on a 3-part scale (0 = no
Results: Four subgroups were identified based on the rate of abnormality, 0.5 = moderate abnormality, 1 = severe
midfoot-hindfoot change: (1) fast-steady (61%), (2) steady-steady abnormality).7 Scores for the medial crease, curvature of
(19%), (3) fast-nil (7%), and (4) steady-nil (14%). The fast-steady the lateral border, and reducibility of the lateral head of
subgroup can be distinguished at the removal of the second cast talus are combined for a midfoot score (MFS), and scores
and all other subgroups can be distinguished at the removal of for the posterior crease, emptiness of heel, and rigidity of
the fourth cast [H (3) = 228.76, P < 0.001]. There was a sig- equinus combined for a hindfoot score (HFS).6,7 MFS and
nificant statistical, not clinical, difference in the total number of HFS are combined for a Pirani scale total foot score (TFS)
casts required for correction across the 4 subgroups [median out of 6 with a higher score demonstrating a more severe
number of casts 5 to 6 in all groups, H (3) = 43.82, P < 0.001]. foot.7 The progression of correction during Ponseti man-
Need for tenotomy was significantly less in the fast-steady (51%) agement has been reported by plotting mean Pirani scale
subgroup compared with the steady-steady (80%) subgroup [H scores in subgroups requiring a different number of casts
(1) = 16.23, P < 0.001]; tenotomy rates did not differ between for correction.6 Change in Pirani scale scores during
the fast-nil (91%) and steady-nil (100%) subgroups [H (1) = 4.13, Ponseti management was described as gradual improve-
P = 0.04]. ment in MFS, minimal improvement in HFS, progressive
Conclusions: Four distinct subgroups of idiopathic clubfoot were correction of TFS, and rapid improvement in HFS and
identified. Tenotomy rate differs between the subgroups TFS with tenotomy.6 Similar to anecdotal clinical ob-
servation, not all clubfoot followed the same pattern of
correction. TFS did not progressively correct week to
From the *Macquarie University; †The Children’s Hospital at West-
mead; and ‡University of Sydney, NSW, Australia.
week during casting in all clubfoot, with mean TFS in
No funding or support was received for this work. 18.4% of feet remaining unchanged at some stage of
The authors declare no conflicts of interest. Ponseti management.6
Reprints: Rachal Quinlan, MRes, Department of Health Sciences, Using TFS at initial assessment to predict short-term
Health and Human Sciences, Ground floor, 75 Talavera Road, outcomes has produced conflicting results. Although a
Macquarie University, NSW 2109, Australia. E-mail: rachal.
quinlan@hdr.mq.edu.au. positive correlation is reported between TFS at initial as-
Copyright © 2023 Wolters Kluwer Health, Inc. All rights reserved. sessment and the number of casts required for correction,
DOI: 10.1097/BPO.0000000000002382 the correlation ranges from weak to strong (r = 0.12 to

J Pediatr Orthop Volume 43, Number 5, May/June 2023 Subgroups of Idiopathic Clubfoot
0.72).8–14 Although the need for tenotomy is also bracing (23 hours a day for 3 months) was prescribed.
positively correlated with TFS at initial assessment, there Part-time bracing (nighttime and naps) was then advised
is a lack of consensus on the cutoff score used for until 4 to 5 years of age.
prediction.9,14–16 Furthermore a lower TFS at initial as- The following data were extracted from electronic
sessment does not eliminate the need for a tenotomy.14 medical records: age at initial assessment, assigned sex,
When examining the relationship between Pirani scale comorbidities, record of prenatal diagnosis, family history
TFS and the number of casts to correction and the need of clubfoot, total number of casts required for correction,
for tenotomy the strongest correlation was after the sec- and need for tenotomy. The total number of casts required
ond cast and not at the initial assessment.17 for correction was calculated from the application of the
Few studies have explored the MFS and HFS as first cast at the initial assessment to the decision on the
independent predictors of outcomes. MFS and HFS at need for tenotomy or progression to abduction bracing.
initial assessment are moderately correlated with the Anycast required after either decision was not included.
number of casts required for correction.9,14 Furthermore, Pirani scale MFS, HFS, and TFS were extracted from
the MFS has been hypothesized to more objectively reflect initial assessment records and, when available, from sub-
the effect of Ponseti casting on deformity correction as the sequent cast change records. Interrater reliability of the
hindfoot deformity is mainly corrected by Achilles Pirani scale was determined when scores had been re-
tenotomy.10 At initial assessment, HFS significantly pre- corded by both the clinic orthopaedic surgeon and a
dicts the need for tenotomy, with a higher score indicating physiotherapist on the same visit. A range of Pirani scale
an increased need for tenotomy.9,14,15 scores was included in the examination of reliability (TFS
on 04/15/2023
Result variability in studies using Pirani scale scores 1.5 to 6).

to predict treatment outcomes could indicate the existence All data, except group-based trajectory modeling
of other factors. The rate of response to treatment in MFS (GBTM), were analyzed with SPSS version 25.0 (IBM
and HFS, and their interaction, may be contributing to the Corp., Armonk, NY). Data were assessed for normality
variability of predicting outcomes seen in the literature. and were not always normally distributed so descriptive
Differences in treatment trajectories of Pirani MFS and data are reported as medians and ranges. Interrater reli-
HFS during Ponseti management have not been inves- ability using the intraclass correlation coefficient 2,1
tigated. (ICC), was determined for Pirani scale MFS, HFS, and
Therefore, the aims of this study are to (1) determine TFS. ICC values above 0.75 were considered excellent,
whether subgroups of Ponseti-managed idiopathic club- 0.40 to 0.75 fair to good, and <0.40 poor.18
foot exist based on the trajectory of change in Pirani scale Clubfoot subgroups were identified statistically using
MFS and HFS, (2) identify the time points, at which GBTM through the “proc traj” plugin in SAS 9.4 (SAS
subgroups can be distinguished, and (3) determine whether Institute).19,20 Trajectories were derived using a censored
subgroups are associated with the number of casts re- normal model based on observed changes in MFS and
quired for correction and need for Achilles tenotomy. HFS from initial assessment to assessment at each cast
change, up to the time when the foot was considered
corrected or needing tenotomy. Censored normal model-
METHODS ing allowed for the clustering of data on the Pirani scale’s
This study was a retrospective medical record review minimum and maximum scores. In turn, the curves of
of infants with clubfoot attending a clubfoot clinic in an each trajectory were calculated using these parameter es-
Australian tertiary children’s hospital. Ethics approval timates. Thus, the curve for each trajectory represented the
(#2019/ETH12935) was gained before the study com- predicted means in the change in MFS and HFS at each
mencement. Infants referred to the clubfoot clinic who had timepoint and its 95% CI. Descriptive statistics were re-
structural idiopathic clubfoot treated with Ponseti man- ported based on statistically derived subgroups. Model
agement between January 2008 and March 2020 were selection was based on widely used criteria including;
included. The minimum follow-up time was the com- obtaining the most parsimonious and clinically interpret-
pletion of initial correction with either Ponseti casting or able model, adequate average posterior probability of
Ponseti casting and Achilles tenotomy. Exclusion criteria group membership, tight confidence intervals around
included (1) nonidiopathic clubfoot, (2) gestational age at predicted curves, and obtaining a model, which accounted
birth of <36 weeks, (3) prior treatment, (4) shared care for the largest variance in observations according to
with another service, (5) Pirani scale MFS and HFS not Bayesian information criteria.20,21 GBTM has been used
recorded at initial assessment. widely in pediatric research to describe changes in a var-
Infants attended the clinic weekly for cast removal, iable across time in underlying subgroups.22
assessment of the clubfoot, and reapplication of the Pon- Validation of the subgroups was performed using
seti cast. Clubfoot severity was assessed using the Pirani generalized estimating equations to enable the analysis of
scale by the clinic orthopaedic surgeon or physiotherapist. pairs of limbs while minimizing the correlation between
The need for Achilles tenotomy was determined by pairs.23 Changes in MFS and HFS at each time point from
the orthopaedic surgeon. Tenotomy was performed if the initial assessment were compared between subgroups
ankle dorsiflexion was <10 degrees after the correction of to determine, at which time point subgroups were dis-
the cavus, adductus, and varus. Post correction, full-time tinguishable. Six post hoc comparisons resulted in a

Quinlan et al J Pediatr Orthop Volume 43, Number 5, May/June 2023
Bonferroni correction to the 0.05 significance level to (fast-steady), the second subgroup (n = 64, 19%) had
0.008. A difference of 0.5 points in the 95% CI between steady midfoot and steady hindfoot change (steady-
assessments was considered as significantly different. steady), the third group (n = 22, 7%) had fast midfoot
As the dependent variable was not normally dis- and nil hindfoot change (fast-nil), and the final subgroup
tributed, a Kruskal-Wallis H test, with a Bonferroni- (n = 44, 13%) had steady midfoot and nil hindfoot change
corrected significance level of 0.008, was used to determine (steady-nil). The baseline demographic and treatment
whether the total number of casts required for correction profiles of each subgroup are presented in Table 2.
differed according to subgroups. The association between MFS, HFS, and TFS were similar at the initial as-
the rate of tenotomy and subgroup was determined using sessment for feet from all subgroups. Infants with bilateral
binary logistic regression. clubfoot did not always have both feet classified into the
same subgroup (n = 54 clubfeet, 25%).
RESULTS The validity of the trajectory subgroups was tested
Five hundred forty-nine infants with clubfoot were using a generalized estimating equation based on changes
referred to the clinic and 226 infants with 335 clubfeet in the MFS and HFS from the initial assessment to each
were eligible for inclusion. The baseline and treatment subsequent cast change. The earliest time point, at which
profile of the study population is presented in Table 1. any subgroup was statistically and clinically distinguish-
Reliability of TFS, MFS, and HFS was good to able according to MFS and HFS was at the removal of the
excellent between the orthopaedic surgeon and physi- second cast, with all subgroups distinguishable at the re-
otherapists in the clinic who assessed 131 clubfeet, for moval of the fourth cast [H (3) = 228.76, P < 0.001]. The
on 04/15/2023
MFS (ICC: 0.88, 95% CI: 0.84-0.91), HFS (ICC: 0.70, clinical significance of this analysis is presented in Table 3.
95% CI: 0.60-0.78) and TFS (ICC: 0.86, 95% CI: 0.80- While all groups demonstrated a median of 5 or 6 casts
0.90). (Table 2) required for initial correction, an omnibus test
Two midfoot and 2 hindfoot trajectories were iden- revealed that there was a statistically significant difference in
tified using GBTM (Fig. 1) and named according to the the total number of casts for correction between at least 2
relative speed of correction compared with other groups. subgroups [Kruskal-Wallis H (3) = 43.82, P < 0.001].
During initial Ponseti management, the rate of change in Adjusted for multiple comparisons, the fast-steady subgroup
the midfoot (midfoot-steady or midfoot-fast) and hindfoot achieved correction with fewer casts than the steady-steady
(hindfoot-nil and hindfoot-steady) differed. These midfoot and steady-nil subgroups (mean differences: 0.5 to 0.8, P <
and hindfoot trajectories were combined to identify 4 0.008). No other statistically significant differences were
subgroups. The first subgroup (n = 205, 61%) found (mean differences: 0.2 to 0.6, all P > 0.02).
demonstrated fast midfoot and steady hindfoot change The need for tenotomy differed significantly between
subgroups [H (1) = 16.23, P < 0.001]. Subgroup fast-steady
had significantly fewer tenotomies than subgroup steady-
TABLE 1. Baseline Demographic and Treatment Profile of the steady [odds ratio (OR): 0.27, 95% CI: 0.14-0.52, P <
Study Population 0.001]. The need did not differ between the fast-nil and
N = 226 steady-nil subgroups [H (1) = 4.13, P = 0.042], therefore,
infants, 335 feet; they were combined for binary logistic regression analysis.
Demographic n (%) Accordingly, the fast-steady (OR: 0.03, 95% CI: 0.01-0.14,
Sex P < 0.001) and steady-steady (OR: 0.12, 95% CI: 0.03-0.57,
Male 157 (69)
Laterality
Unilateral 117 (52)
Right 175 (52)
Family history of clubfoot (N = 220)* —
Yes 75 (34)
Prenatal diagnosis (N = 219)* —
Yes 171 (76)
Age at initial assessment (d); median(range) 16 (5-64)
Pirani scale scores at initial assessment; median(range)
MFS/3 2.5 (0.5-3)
HFS/3 3 (1-3)
TFS/6 5.5 (2-6)
No. casts for correction (N = 333)*; median(range) 5 (2-9)
Time between cast changes (d); median(range) 7 (3-28)
Rate of correction
Casting and tenotomy 220 (66)
Casting only 115 (34)
Bilateral cases (N = 218)
Casting and tenotomy 150 (69)
Casting only 68 (31)
*Reduced N due to missing data.
HFS indicates hindfoot score; MFS, midfoot score; TFS, total foot score.
FIGURE 1. Trajectory groups for idiopathic clubfoot.

TABLE 2. Baseline Demographic and Treatment Profiles of Each Subgroup

Demographic N = 205; n (%) N = 64; n (%) N = 22; n (%) N = 44; n (%)
Subgroup name (midfoot-hindfoot) Fast-steady Steady-steady Fast-nil Steady-nil
Age at initial assessment (d); median(range) 15 (5-61) 17 (5-57) 20 (12-39) 20 (5-64)
Laterality
Unilateral 82 (40) 17 (27) 7 (32) 11 (25)

Bilateral both feet in subgroup 106 (52) 28 (44) 8 (36) 22 (50)
Bilateral 1 foot in a single subgroup 17 (8) 19 (29) 7 (32) 11 (25)
Pirani scale scores at initial assessment; median(range)
MFS/3 2.5 (0.5-3) 3 (2-3) 2.5 (1.5-3) 3 (2-3)
HFS/3 3 (1-3) 3 (1.5-3) 3 (2.5-3) 3 (2.5-3)
TFS/6 5 (2-6) 6 (4.5-6) 5.5 (4.5-6) 6 (4.5-6)
No. casts for correction (N = 333)*; median(range) 5 (2-9) 5 (3-9) 5 (3-7) 6 (4-9)
Rate of correction
Casting and tenotomy 105 (51) 51 (80) 20 (91) 44 (100)
Casting only 100 (49) 13 (20) 2 (9) 0
*Reduced N due to missing data.
HFS indicates hindfoot score; MFS, midfoot score; TFS, total foot score.
P < 0.007) subgroups had significantly fewer tenomoties for each clubfoot were examined and changes between
on 04/15/2023
than the combined subgroup. casts were analyzed. The relationship between MFS and
The fast-steady subgroup was 4 times less likely to HFS and their combined contribution towards patterns of
need tenotomy than the steady-steady subgroup. Com- correction was also examined. The existence of subgroups
pared with the combined group (fast-nil and steady-nil) could account for the TFS of all clubfeet in previous lit-
the fast-steady subgroup was 33 times less likely, and the erature not following the same pattern of change.
steady-steady group was 8 times less likely to require te- This study identified time points in Ponseti man-
notomy. The hindfoot trajectory of an infant’s clubfoot agement, at which clubfoot can be distinguished into
significantly predicted the need for tenotomy. A steady subgroups. To our knowledge, this is the first report in the
hindfoot trajectory was less likely to require tenotomy literature. All 4 subgroups in this study presented similarly
than a nil hindfoot trajectory in all subgroups, irrespective at the initial assessment in terms of median TFS, MFS,
of midfoot trajectory. and HFS. All feet could be distinguished into subgroups at
the removal of 2 to 4 casts. After a clubfoot is dis-
DISCUSSION tinguished into a subgroup then outcomes may be pre-
This study identified 4 subgroups in a cohort of in- dicted. This study’s findings suggest that the initial
fants with idiopathic clubfoot. Subgroups based on the assessment is too early to predict outcomes without al-
change in MFS and HFS during Ponseti management lowing for analysis of the change in scores during Ponseti
demonstrated 2 distinct midfoot trajectories (steady and management.
fast) and 2 distinct hindfoot trajectories (steady and nil). Regardless of subgroup, all infants with clubfoot
These trajectories combined to form 4 distinct subgroups, required a similar number of casts for correction. Al-
which were all clinically distinguishable at the removal of though there was a statistically significant difference be-
the fourth cast. To our knowledge, this is the first study to tween the number of casts to correction between the
identify subgroups of idiopathic clubfoot based on the subgroups, this was <1 cast (0.5 to 0.8) and not considered
change in Pirani scale MFS and HFS in individual feet clinically significant. Clinicians could still consider this
during Ponseti management. Previous literature reported useful information regarding the expected duration of
the progression of correction during Ponseti management Ponseti management.
by plotting mean Pirani scale scores in subgroups requir- Subgroups were associated with the need for tenot-
ing a different number of casts for correction, with TFS of omy, with both the midfoot and hindfoot trajectory pre-
all clubfeet not following the same pattern of change.6 dicting the need for tenotomy. The nil hindfoot trajectory
This current study differed as individual MFS and HFS predicted the need for tenotomy in 2 subgroups. Previous
TABLE 3. Timepoint and Minimum Reduction in Pirani Scale Score Distinguishing Subgroups
Timepoint From Initial Assessment Reduction in MFS* (points) Reduction in HFS* (points) Subgroup
At removal of second cast ≥ 1.5 AND ≥ 1.0 Fast-steady
At removal of third cast ≥ 1.0 AND ≥ 1.0 Steady-steady
≥ 1.0 AND ≥ 0.0 Steady-nil
At removal of fourth cast ≥ 2.0 AND ≥ 0.5 Fast-nil
*Change in score from initial assessment.
HFS indicates hindfoot score; MFS, midfoot score.

Quinlan et al J Pediatr Orthop Volume 43, Number 5, May/June 2023
studies suggested that Pirani scale HFS at initial assessment recurrence, and the need for further treatment such as
is associated with the need for tenotomy and have reported tenotomy and tibialis anterior tendon transfer. Finally,
differing baseline TFS cut-offs to predict the need for this study has produced a statistically-derived model of
tenotomy.9,14–16 Existence of subgroups in these study co- subgroups based on the change in MFS and HFS. Further
horts may have contributed to the variability of baseline research is required to externally validate this model in
scores predicting tenotomy. Furthermore, while there was terms of its performance, calibration, and discrimination
no difference between the rate of tenotomy in feet with a nil in a prospective inception cohort.26 It is possible that other
improvement in hindfoot trajectory improvement, a dif- variables, unable to be considered in this study, may have
ference in the rate of tenotomy in the 2 subgroups with a influenced findings. Future prospective studies may be
steady hindfoot trajectory was identified, suggestive of an able to determine such factors.
effect of midfoot improvement on the rate of tenotomy in Documentation of Pirani scale MFS and HFS dur-
some clubfoot. ing Ponseti management may assist clinicians to identify
This study utilizes the change in MFS and HFS, subgroups of clubfoot. This may provide clinicians and
early in Ponseti management, to distinguish idiopathic families of infants with clubfoot clearer expectations of the
clubfoot from subgroups and predict the need for tenot- short-term outcomes of Ponseti management.
omy. A previous study suggested using percentage change
in Pirani scale TFS between initial assessment and point of REFERENCES
correction, to predict the probability of needing 1. Ponseti IV. The treatment of congenital clubfoot. J Orthop Sports
tenotomy.16 However, this approach only provides prog- Phys Ther. 1994;20:1.
2. Ponseti VI, Zhivkov BM, Davis AN, et al. Treatment of the complex
on 04/15/2023
nostic information at the point of correction. The ex-

idiopathic clubfoot. Clin Orthop Relat Res. 2006;451:171–176.
istence of subgroups and their ability to predict outcomes, 3. Cooper DM, Dietz FR. Treatment of idiopathic clubfoot. a thirty-
as early as at removal of the second cast, could optimize year follow-up note. J Bone Joint Surg Am. 1995;77:1477–1489.
the timing of tenotomy by helping clinicians make a de- 4. Jowett CR, Morcuende JA, Ramachandran M. Management of
cision earlier in Ponseti management. Families of infants congenital talipes equinovarus using the Ponseti method: a system-
with clubfoot can also be provided with more timely in- atic review. J Bone Joint Surg Br. 2011;93:1160.
5. Gelfer Y, Weintroub S, Hughes K, et al. Congential Talipes
formation regarding the need for tenotomy. equinovarus: a systematic review of relapse as a primary outcome
This study found that some feet of infants with bi- of the Ponseti method. Bone Joint J. 2019;101:639–645.
lateral clubfoot can follow different trajectories during 6. Lampasi M, Abati CN, Stilli S, et al. Use of the Pirani score in
Ponseti management. In infants with bilateral clubfoot, monitoring progression of correction and in guiding indications for
tenotomy in the Ponseti method: are we coming to the same
both feet were not always distinguished into the same decisions? J Orthop Surg (Hong Kong). 2017;25:1–8.
subgroup. In 3 of the 4 subgroups in this study, both feet 7. Pirani S, Hodges D, Sekeramayi F. A reliable and valid method of
of bilateral cases in the same subgroup occurred at least assessing the amount of deformity in the congenital clubfoot
1.5 times the rate of those with bilateral feet in different deformity. Orthop Proc. 2018;90:53.
subgroups (fast-steady 52% vs 8%, steady-steady 44% vs 8. Tahririan MA, Ardakani MP, Kheiri S. Can clubfoot scoring
systems predict the number of casts and future recurrences in patients
29%, steady-nil 50% vs 25%). Previous studies have undergoing Ponseti method? J Orthop Surg Res. 2021;16:238.
demonstrated that baseline Pirani scale scores, number of 9. Lampasi M, Abati CN, Bettuzzi C, et al. Comparison of Dimeglio and
casts required for correction, and the rate of tenotomy in Pirani score in predicting number of casts and need for tenotomy in clubfoot
left and right feet are highly positively correlated in chil- correction using the Ponseti method. Int Orthop. 2018;42:2429–2436.
10. Fan H, Liu Y, Zhao L, et al. The correlation of Pirani and Dimeglio
dren with bilateral clubfoot.24,25 Despite similar Pirani scoring systems for Ponseti management at different levels of
scores at baseline a proportion of bilateral feet from single deformity severity. Sci Rep. 2017;7:14578.
infants in this cohort differed in the number of casts for 11. Zhao D, Li H, Zhao L, et al. Prognosticating factors of relapse in clubfoot
correction, rate of tenotomy, and trajectories of change in management by Ponseti method. J Pediatr Orthop. 2018;38:514–520.
12. Gao R. Correlation of Pirani and Dimeglio Scores with number of
MFS and HFS during Ponseti management. Therefore,
Ponseti casts required for clubfoot correction. J Pediatr Orthop.
analysis including both feet from bilateral cases, and 2014;34:639–642.
considering correlations between feet, is required to ensure 13. Chu A, Labar AS, Sala AD, et al. Clubfoot classification: correlation
appropriate representation of varied bilateral cases. with Ponseti cast treatment. J Pediatr Orthop. 2010;30:695–699.
There are limitations to this study. This study is 14. Dyer PJ, Davis N. The role of the Pirani scoring system in the
management of club foot by the Ponseti method. J Bone Joint Surg
based on retrospective data from medical records with Br. 2006;88:1082–1084.
prospective research required. Secondly, multiple clini- 15. Scher MD, Feldman SD, Van Bosse JPH, et al. Predicting the need
cians completed the Pirani scale scoring and Ponseti for tenotomy in the Ponseti method for correction of clubfeet. J
casting for infants included in this study. However, the Pediatr Orthop. 2004;24:349–352.
dedicated clubfoot clinic was located in a tertiary hospital 16. Aydin BK, Senaran H, Yilmaz G, et al. The need for Achilles
tenotomy in the Ponseti method: is it predictable at the initiation or
with experienced clinicians. Furthermore, the reliability during the treatment? J Pediatr Orthop B. 2015;24:341–344.
for use of the Pirani scale in the clinic was demonstrated as 17. Jochymek J, Peterkova T. Are scoring systems useful for predicting
good to excellent. Thirdly, with the completion of initial results of treatment for clubfoot using the Ponseti method? Acta
correction as the minimum follow-up time, we were un- Ortop Bras. 2019;27:8–11.
18. Fleiss JL. Design and Analysis of Clinical Experiments. Canada:
able to determine whether there was an association be- Wiley-Interscience; 2011.
tween subgroups and long-term outcomes including gross 19. Jones BL, Nagin DS. A note on a stata plugin for estimating group-
motor development of the infant with clubfoot, based trajectory models. Sociol Methods Res. 2013;42:608–613.

20. Jones BL, Nagin DS, Roeder KA. SAS procedure based on mixture statistical efficiency in musculoskeletal research. Gait Posture.
models for estimating developmental trajectories. Sociol Methods 2018;59:182–187.
Res. 2001;29:374–393. 24. Gray K, Gibbons P, Little D, et al. Bilateral clubfeet are highly
21. Nagin DS, Odgers CL. Group-based trajectory modeling in clinical correlated: a cautionary tale for researchers. Clin Orthop Relat Res.
research. Annu Rev Clin Psychol. 2010;6:109–138. 2014;472:3517–3522.
22. Mattsson M, Maher GM, Boland F, et al. Group-based trajectory 25. Gray K, Barnes E, Gibbons P, et al. Unilateral versus bilateral
modelling for BMI trajectories in childhood: a systematic review. clubfoot: an analysis of severity and correlation. J Pediatr Orthop B.
Obes Rev. 2019;20:998–1015. 2014;23:397–399.

23. Stewart S, Pearson J, Rome K, et al. Analysis of data collected from 26. Altman DG, Vergouwe Y, Royston P, et al. Prognosis and prognostic
right and left limbs: accounting for dependence and improving research: validating a prognostic model. BMJ. 2009;338:605.
on 04/15/2023

ORIGINAL ARTICLE
Subsequent Forearm Fractures Following Initial Surgical

Fixation
Amelia M. Lindgren, MD,*† Gabriela Sendek, BS,*† Claire E. Manhard, MPH,*

Tracey P. Bastrom, MA,* and Andrew T. Pennock, MD*†
ESINs are both less invasive at the time of initial surgery and can
Introduction: Forearm fractures are a common pediatric injury. often be treated nonoperatively if there is a subsequent fracture,
Currently, there is no consensus on treatment for fractures that while plate refractures are more likely to be treated with a second
recur following initial surgical fixation. The objective of this surgery and have a longer average surgery time.
study was to investigate the subsequent fracture rate and patterns Level of Evidence: Level IV—retrospective case series.
and describe the treatment of these forearm fractures.
Methods: We retrospectively identified patients who underwent Key Words: forearm fracture, refracture, plate, elastic stable in-
tramedullary nail
on 04/15/2023
surgical treatment for an initial forearm fracture at our in-

stitution between 2011 and 2019. Patients were included if they (J Pediatr Orthop 2023;43:e383–e388)
sustained a diaphyseal or metadiaphyseal forearm fracture that
was initially treated surgically with a plate and screw construct
(plate) or elastic stable intramedullary nail (ESIN), and if they
subsequently sustained another fracture that was treated at our
institution.
Results: A total of 349 forearm fractures were treated surgically
F orearm fractures are among the most common pedia-
tric injuries. Typically, uncomplicated fractures are
treated with closed reduction and casting.1,2 However,
with ESIN or a plate fixation. Of these, 24 sustained another some fractures fail conservative management or are not
fracture, yielding a subsequent fracture rate of 10.9% for the amenable to nonoperative treatment and require surgical
plate cohort and 5.1% for the ESIN cohort (P = 0.056). The fixation.2,3 Current treatment options include elastic stable
majority of plate refractures (90%) occurred at the proximal or intramedullary nailing (ESIN), plate and screw constructs,
distal plate edge, while 79% of the fractures treated previously k-wire fixation, and external fixation. Prior studies have
with ESINs occurred at the initial fracture site (P < 0.001). demonstrated no significant difference in functional out-
Ninety percent of plate refractures required revision surgery, comes between patients treated with nailing or plating;
with 50% underwent plate removal and conversion to ESIN, and however, nailing may be preferable given the shorter op-
40% underwent revision plating. Within the ESIN cohort, 64% erative time, cosmetic appeal, and ease of hardware
were treated nonsurgically, 21% underwent revision ESINs, and removal.4
14% underwent revision plating. Tourniquet time for revision Regardless of treatment type, there is a risk of re-
surgeries were shorter for the ESIN cohort (46 vs. 92 min; fracture. The estimated rate of refracture following ESIN
P = 0.012). In both cohorts, all revision surgeries had no com- is ∼5% (range: 1.2% to 10.4%).5–17 Similarly, the reported
plications and healed with evidence of radiographic union. refracture rate for forearm fractures treated with plate and
However, 9 patients (37.5%) underwent implant removal (3 screw constructs ranges from 6.9% to 8.5%.13,18–20 Fol-
plates and 6 ESINs) after subsequent fracture healing. lowing elective plate removal, this rate has been docu-
Conclusions: This is the first study to characterize subsequent mented to be as high as 40% (2/5 patients).20 The ability to
forearm fractures following both ESIN and plate fixation and to determine risk factors associated with refracture is limited
describe and compare treatment options. Consistent with the by fairly small sample size. However, male sex, younger
literature, refractures following surgical fixation of pediatric age, higher weight, bone size, and plate design have all
forearm fractures may occur at a rate ranging from 5% to 11%. been associated with this complication.10,15–18
The treatment of these subsequent fractures has
been anecdotally described in the literature. For exam-
From the *Rady Children’s Hospital-San Diego; and †University of ple, refracture with retained ESIN has been treated with
California San Diego, San Diego, CA.
This study was supported by Rady Children’s Orthopaedic Research and
in situ ESIN closed reduction and casting,15,21–23 ex-
Education. change nailing,16,24 as well as ESIN removal and
A.T.P. is a paid consultant for the Pediatric Orthopaedic Society of plating.16 Alternatively, refracture with retained plate
North America and a board or committee member. The remaining has been treated with closed reduction and casting, plate
authors declare no conflicts of interest. removal and ESIN, or revision plating.20,25
Reprints: Andrew T. Pennock, MD, 3020 Children’s Way, MC 5062, San
Diego, CA 92123. E-mail: apennock@rchsd.org. Currently, there is no consensus on treatment of
Copyright © 2023 Wolters Kluwer Health, Inc. All rights reserved. forearm fractures that have previously undergone surgical
DOI: 10.1097/BPO.0000000000002374 fixation in the pediatric population. The purpose of this

Lindgren et al J Pediatr Orthop Volume 43, Number 5, May/June 2023
study was to investigate forearm fracture rates and pat- lations were done by patient. Statistical analysis was per-
terns that were previously treated with ESIN or plate formed per patient with nonparametric Mann-Whitney U
fixation, and to describe a single institution’s approach to test to compare continuous variables and χ2 or the Fisher
these injuries. exact test to compare categorical variables. Alpha was set
at P < 0.05 to declare significance. SPSS v. 27 was utilized
METHODS for the analyses (IBM Corp. Released 2020. IBM SPSS
An IRB approved retrospective study was conducted Statistics for Windows, Version 27.0; IBM Corp., Ar-
and identified patients treated surgically for an initial monk, New York, USA).
forearm fracture at a single level one pediatric trauma
center between October 2011 and September 2019. In- RESULTS
clusion criteria were diaphyseal or metadiaphyseal fore- During the study period, 349 patients that under-
arm fractures that were surgically treated with a plate and went surgical fixation of a forearm fracture were identi-
screws or ESIN, subsequent fracture treated at our in- fied. Of these, 92 were treated with at least one plate, 272
stitution, and below 18 years of age at the time of re-injury were treated with at least one ESIN, and 15 were treated
(Fig. 1). Exclusion criteria were signs of pathologic with a hybrid construct. During the study period, 24 pa-
fracture or an underlying neuromuscular condition. tients that sustained a subsequent fracture were identified
Patients with subsequent fractures were categorized (7 patients initially treated with plates, 13 with ESINs, and
by their initial surgical treatment: ESIN group, a plate and 4 with hybrid constructs) (Table 1). Within the hybrid
screw construct (plate group), or a plate/ESIN hybrid cohort, 3 fractures involved only 1 bone (2 initially treated
on 04/15/2023
construct (hybrid group) where one bone was treated with a with plates and 1 initially treated with an ESIN). Further,
plate and the other was treated with an ESIN. Patient charts the final hybrid fracture involved both bones, but the nail
and radiographs were reviewed. Patient age, sex, mecha- had previously been removed from the ulna and the radius
nism of initial and subsequent injuries, fracture types, fracture occurred at the end of the plate. For subsequent
fracture and subsequent fracture locations, initial surgical analysis, the 3 hybrid fractures involving plate treatment
treatment, time to re-injury, interval removal of implants, were grouped with the plate subsequent fracture cohort
and subsequent treatment technique were analyzed. Fol- and the one hybrid fracture involving the ESIN was
low-up documentation was reviewed for complications and placed into the ESIN cohort. Overall, the subsequent
radiographic evidence of union. Management of the initial fracture group consisted of 20 males and 4 females with an
fracture as well as the subsequent fracture were at the dis- average age of 10.6 ± 3.2 years (range: 6.1 to 16.2 y) at
cretion of the surgeons at our institution. The various time of initial surgical fixation. The mechanism of re-
board-certified orthopaedic surgeons assessed each fracture injury involved the following: device with wheels (bicycle,
and made the decision to operate or not to operate based skateboard, hoverboard) (37.5%), injuries while playing
upon radiographic evidence of injury, expertise, and sports (29.2%), falls from standing heights (20.8%), falls
practice within the field. At the time of initial treatment, from objects (play structure, trampolines, etc.) (12.5%).
patients that were older, fractures that were more length The average time from the initial fracture to the
unstable (comminuted or long oblique fractures), or those subsequent fracture was 1.6 ± 1.6 years (range: 0.2 to
fractures requiring an open approach (irreducible fractures 6.2 y). When comparing the demographic and injury data
or open fractures) were more likely to be considered can- between the ESIN and plate cohorts no differences were
didates for plate fixation versus ESIN. noted between groups with respect to age, sex, or
Descriptive statistics were calculated using Excel mechanism of injury (P > 0.05).
(Microsoft Corporation, Redmond, Washington, USA). The subsequent fracture rate did not reach statistical
Calculations relating to fracture location and fracture significance between the ESIN and plate cohorts, but it
treatment were done by bone and the remaining calcu- trended towards a lower rate in the ESIN group [14/272
FIGURE 1. Flow diagram showing the treatment approach for the 24 subsequent fractures. ESIN indicates elastic stable
intramedullary nail.

J Pediatr Orthop Volume 43, Number 5, May/June 2023 Forearm Refractures After Prior Plating or ESIN
TABLE 1. Subsequent Fracture Demographics and

revision surgery (90% vs. 35.7%; P = 0.025). In the plate
Treatment Data cohort, 9/10 fractures required a revision operation and 1/
10 were successfully managed with closed reduction and
Plate, N = 10, ESIN,
n (%) N = 14, n (%) P
casting for an isolated radius periprosthetic fracture. Of
the 9 patients requiring a revision operation, 6 were
Age at initial surgery (years) 11.7 ± 2.5 9.9 ± 3.4 0.096 treated with plate removal and ESIN insertion, 1 was
Sex (male) 8 (80) 12 (86) 0.71

Time to subsequent fracture 2.1 ± 1.5 1.3 ± 1.6 0.074
treated with adjunctive perpendicular plate placement,
(y) and 2 were treated with re-plating (Fig. 2). In the ESIN
Implant removal before 1 (10) 8 (57) 0.033 cohort, 5/14 required a revision operation and 9/14 were
subsequent fracture successfully managed with closed reduction and casting,
Implant removal following 3 (30) 6 (43) 0.68 including 6/7 patients with retained ESINs. The average
subsequent fracture
healing angulation of retained ESIN after re-injury, but before
Tourniquet time (min) 92 ± 28 45 ± 28 0.012 reduction, was 26.8 ± 25.5 degrees (range: 0 to 76.1
Subsequent fracture type degrees). Of the 5 patients requiring a revision
Both bone 7(70) 8 (57) 0.45 operation, 2 were treated with plating, 2 had ESINs
Radius only 3(30) 4(29)
Ulna only — 2(14)
placed after having been previously removed, and 1 had a
Subsequent fracture location bent ESIN removed and a new ESIN placed.
Plate edge 9 (90) — < 0.001 The average tourniquet time during revision surgery
Distal ESIN — 1 (7) was significantly less in the ESIN cohort as compared with
Original location 1 (10) 11 (79) the plate cohort (46 ± 28 vs. 92 ± 28 min; P = 0.012). All
on 04/15/2023
New fracture — 2 (14)

Treatment patients in both cohorts subsequently healed. In the plate
Cast 1 (10) 9 (64) 0.025 cohort, no complications were noted, other than 1 patient
Plate 4 (40) 2 (14) who underwent implant removal before refracture and
ESIN 5 (50) 3 (22) was found to have a distal radial sensory neuropathy and
ESIN indicates elastic stable intramedullary nail. ulnar nerve palsy on EMG before revision surgery. This
patient’s complication resolved by the 8-month post-
operative visit. Three of the 10 fracture patients in the
(5.1%) vs. 10/92 (10.9%); P = 0.056]. Further, the average plate cohort had their implants removed at the time of
time to re-injury did not differ significantly between the final follow-up. In the ESIN cohort, 6/14 fracture patients
ESIN and plate cohorts [1.3 ± 1.6 y (range: 0.2 to 6.2 y) vs. ultimately had their implants removed. Of note, 1 patient
2.3 ± 1.5 y (range: 0.3 to 4.5 y), respectively; P = 0.074]. In in the ESIN cohort sustained a second refracture, which
the plate cohort, 6/10 (60%) subsequent fracture involved was successfully treated with closed reduction and casting.
both the radius and ulna and 4/10 (40%) were isolated to
the radius. Half of the plate fractures were previously fixed
with 3.5 mm plates while half were fixed with 2.7 mm DISCUSSION
plates. The majority of plates were dynamic compression The subsequent fracture rate following initial surgi-
plates (8/10) and 2 were 1/3 tubular plates. In the ESIN cal intervention for a forearm fracture was 10.9% for the
cohort, 8/14 (57%) subsequent fractures involved the ra- plate cohort and 5.1% for the ESIN cohort. This is con-
dius and ulna, 4/14 (29%) and 2/14 (14%) were isolated to sistent with previously published results and reinforces the
the radius and ulna, respectively. In the plate cohort, 9/10 notion that refractures do occur after prior fixation and
(90%) fractures occurred with the plate in situ, while 1 developing treatment strategies for these injuries is
radius refracture occurred 1 month after plate removal. In important.
the ESIN cohort, 7/14 (50%) of subsequent fractures oc- Published literature supports a wide range in average
curred following ESIN removal. The mean time from time to refracture, however, the time to a subsequent
ESIN removal to the next fracture was 1.5 ± 2.0 years fracture in our study was similar between the ESIN and
(range: 0.1 to 5.5 y). Of note, the majority (77%) of frac- plate cohorts (2.1 vs. 1.3 y, respectively). Studies reporting
tures in the ESIN group had an open reduction to pass the refractures after ESIN demonstrate that the majority of
nail with most being opened to irrigate an open fracture. refractures occur within 1 year of the initial surgery, with
The location of the subsequent fractures was found time to refracture ranging from 4 to 13 months,5–7,14,15,17
to differ significantly between the 2 groups, with 9/10 and as far out as 34 months.18 Furthermore, secondary
(90%) and 1/10 (10%) plate cohort fractures occurring at fractures were more likely to occur in the year following
the distal or proximal edge of the plate and at the original ESIN removal.5,9,12,17 Makki et al13 found a statistically
fracture site following plate removal, and 11/14 (79%), 1/ significant difference in re-injury rate in patients who un-
14 (7%), and 2/14 (14%) ESIN cohort fractures occurring derwent ESIN removal within 6 months of initial treat-
at the original fracture site, distal to the ESIN tip, and at a ment versus after 6 months (5/20 refractures within 6 mo
different location, respectively (P < 0.001). and 0/18 after 6 mo). Interestingly, 7/14 (50%) in our
Treatment approach differed significantly between ESIN cohort had ESIN removal before their subsequent
the 2 cohorts, with a higher percentage of patients in the fracture (4 with ESIN removal within 6 mo of initial
plate cohort as compared with the ESIN cohort requiring treatment and 3 with removal after 6 mo). This suggests

on 04/15/2023
FIGURE 2. Subsequent forearm fractures following initial surgical fixation. A, Initial injury films of patient from plate cohort. B, Five
weeks following revision to ESIN. C, Demonstrating well-healed fractures at 6 months postoperatively. D, Initial injury films from the
ESIN fracture cohort. E, Following closed reduction and casting. F, Well-healed 5 months out from injury.
that the timing of ESIN removal may not be as crucial as within 12 months of the initial surgery, as compared with
previously reported. plate removal 12 months after the initial surgery [11/62
At our institution, most plates tend to be retained plates (17.7%) vs. 0/50 plates (0%), respectively]. In con-
unless symptomatic given the more invasive surgical pro- trast to these findings, only 1 patient in our plate cohort
cedure required for removal. Most published literature had plate removal 17 months after the initial surgery and
reports refractures occurring after plate removal, with had subsequent refracture within 1 month of removal.
time to refracture ranging from 12 days after plate re- Whether these subsequent fractures with longer follow-up
moval to 3 months.19,20 Published literature on refractures and complete remodeling represent refractures or just new
without plate removal report a time to refracture of up to injuries can be debated, but it should be recognized that
4 years after injury.20 Makki and colleagues found a sig- subsequent fractures with or without retained implants
nificantly higher refracture rate following plate removal can occur years after the initial surgical treatment.

J Pediatr Orthop Volume 43, Number 5, May/June 2023 Forearm Refractures After Prior Plating or ESIN
The current study revealed that the subsequent fatigue after 1 cycle of an ESIN being reverse bent to 21
fracture location varied based on the initial surgical fix- degrees, however, the average force required for permanent
ation technique, with 90% of plate refractures occurring at deformation of a previously bent nail was reduced by 37%
the plate edge and 79% of the ESIN refractures occurring for both titanium and stainless steel nails (21 to 13.2 N and
at the original fracture site. Little has been published that 25 to 15.7 N, respectively).21 Interestingly, half of our pa-
describes the subsequent fracture location. Van Egmond tients with retained ESIN at time of re-injury had over 20
et al23 described a case where a refracture occurred 9 degrees of angulation and all were successfully treated with
weeks after an ESIN at the original fracture location, closed reduction and casting. We, therefore, do not believe
while Kelly et al7 describes a refracture distal to the ESIN that a bent nail > 20 degrees is a definite indication for a
implant. Alternatively, McLean et al25 described 4 cases of revision surgery and strong consideration should be given to
plate refractures that occurred at either the proximal or an attempt at closed reduction.
distal screw hole and Vopat et al20 described a refracture In the literature, forearm fractures with retained
at the initial site 2 weeks after plate removal. plates have been managed with closed reduction and
A variety of treatment options are supported for pe- casting,20 revision plating13,18 or conversion to ESIN.25 Our
diatric forearm fractures that have retained implants from study consolidates and highlights the majority of techniques
previous surgical intervention. Within our plate cohort, 50% that have been previously described. However, the decision
of fractures were treated with plate removal and conversion to remove or retain surgical plates in pediatric surgeries
to ESIN, 20% with new plate placement, 10% with perpen- remains a question because of the varying risks and benefits
dicular plate augmentation, and 10% with closed reduction proposed within the literature. Prior work has suggested
on 04/15/2023
and casting. While differing strategies exist for managing that removal of forearm plates in children may be asso-
these plate refractures, many of the surgeons at our institution ciated with a complication rate ranging from 9% to 42%
prefer placing an intramedullary device bypassing all of the with subsequent fractures representing a large percentage of
screw holes as opposed to placing a longer plate which re- these complications.19,26 This has prompted some authors
quires a longer incision and theoretically creates a new stress to advocate the retention of forearm plates, which is the
riser at the end of the longer plate. In these situations, where a standard practice at our institution. Clement and colleagues
flexible nail is used, a Kirschner wire, curette, or drill is often found a refracture rate of 7.3% with retention of the plates,
necessary to recreate the intramedullary canal. Other factors of which 83% required a second surgery to re-plate the
that potentially influenced surgical decision making was the fractured bone. Our findings report a similar rate of sub-
age of the patient, the location of the secondary fracture, as sequent fracture after plate fixation (10.9%) with a similar
well as the difficulty of removing the retained implants. The number requiring a second surgery (90%).
patients that were subsequently plated were averaged 1-year There are limitations to this study. The cohort of pa-
older than those treated with a revision to an ESIN. Patients tients who had subsequent fractures following initial surgical
with a subsequent fracture distal to the original plate that fixation at our institution was relatively small, thus limiting
potentially precluded the placement of a flexible nail were our ability to perform extensive subanalysis including the
treated with revision plates. Furthermore, patients with effect of plate size and type. Given the small sample size, it
stripped screws and or significant bone growth over the plate was not possible to determine significant differences between
were considered for perpendicular plates. Within our ESIN cohorts. As this was a retrospective study reliant on clinic
cohort, the subsequent fractures were managed quite differ- notes and radiographs, there are no specific patient-reported
ently, with the majority being treated with closed reduction outcomes. In addition, some patients with subsequent frac-
and casting. In the ESIN cohort, patients who underwent a tures may have been lost to follow-up. Surgical intervention
second procedure were either re-nailed or converted to a plate was at the surgeon’s discretion and it is not always clear in a
and screw construct, however, regardless of approach, their retrospective study why a specific technique or treatment
revision surgery was nearly half as long as those in the plate approach was used for each specific patient.
cohort. In the ESIN cohort, the preference of the surgeons This is the first study to characterize pediatric fore-
was typically to reinsert ESINs when possible, but if the arm fractures following both ESIN and plate fixation and
fracture site had to be opened to cut and remove the bent nail, to describe and compare treatment options for this chal-
a plate was then considered. Overall, these findings demon- lenging clinical problem. While there were no significant
strate potential advantages of initial treatment with ESIN differences in rates of subsequent fractures between the
over plate fixation. ESIN cohort and the plate cohort, fractures after previous
Similar to the present study, there are many descrip- ESIN treatment were more often managed nonoperatively
tions of forearm fractures with retained ESINs being whereas fractures that had undergone previous plate fix-
treated with closed reduction and immobilization in the ation were more likely to require a second surgery.
literature, sometimes under general anesthesia.7,13,15,17,22,24
Han et al10 recommends fractures with retained ESIN with
<20 degrees angulation undergo closed reduction and REFERENCES
1. Price CT. Acceptable alignment of forearm fractures in children:
casting, while those with over 20 degrees angulation un- open reduction indications. J Pediatr Orthop. 2010;30:S82–S84.
dergo ESIN removal and reinsertion of new ESINs after 2. Zionts LE, Zalavras CG, Gerhardt MB. Closed treatment of
reduction. The biomechanical rationale for this approach is displaced diaphyseal both-bone forearm fractures in older children
that there is no macroscopic evidence of metal fracture or and adolescents. J Pediatr Orthop. 2005;25:507–512.

3. Franklin CC, Robinson J, Noonan K, et al. Evidence-based 15. Rousset M, Mansour M, Samba A, et al. Risk factors for re-fracture
medicine: management of pediatric forearm fractures. J Pediatr in children with diaphyseal fracture of the forearm treated with
Orthop. 2012;32:S131–S134. elastic stable intramedullary nailing. Eur J Orthop Surg Traumatol.
4. Westacott DJ, Jordan RW, Cooke SJ. Functional outcome following 2016;26:145–152.
intramedullary nailing or plate and screw fixation of paediatric 16. Salonen A, Salonen H, Pajulo O. A critical analysis of postoperative
diaphyseal forearm fractures: a systematic review. J Child Orthop. complications of antebrachium TEN-nailing in 35 children. Scand J
2012;6:75–80. Surg. 2012;101:216–221.

5. Fernandez FF, Langendörfer M, Wirth T, et al. Failures and 17. Tsukamoto N, Mae T, Yamashita A, et al. Refracture of pediatric
complications in intramedullary nailing of children’s forearm both-bone diaphyseal forearm fracture following intramedullary
fractures. J Child Orthop. 2010;4:159–167. fixation with Kirschner wires is likely to occur in the presence of
6. Kelly BA, Miller P, Shore BJ, et al. Exposed versus buried immature radiographic healing. Eur J Orthop Surg Traumatol.
intramedullary implants for pediatric forearm fractures: a compar- 2020;30:1231–1241.
ison of complications. J Pediatr Orthop. 2014;34:749–755. 18. Clement ND, Yousif F, Duckworth AD, et al. Retention of forearm
7. Kelly BA, Shore BJ, Bae DS, et al. Pediatric forearm fractures with plates: risks and benefits in a paediatric population. J Bone Joint
in situ intramedullary implants. J Child Orthop. 2016;10:321–327. Surg Br. 2012;94:134–137.
8. Antabak A, Luetic T, Ivo S, et al. Treatment outcomes of both-bone 19. Kim WY, Zenios M, Kumar A, et al. The removal of forearm plates
diaphyseal paediatric forearm fractures. Injury. 2013;44(suppl 3):S11–S15. in children. Injury. 2005;36:1427–1430.
9. Dinçer R, Köse A, Topal M, et al. Surgical treatment of pediatric 20. Vopat BG, Kane PM, Fitzgibbons PG, et al. Complications
forearm fractures with intramedullary nails: is it a disadvantage to associated with retained implants after plate fixation of the pediatric
leave the tip exposed? J Pediatr Orthop B. 2020;29:158–163. forearm. J Orthop Trauma. 2014;28:360–364.
10. Han B, Wang Z, Li Y, et al. Risk factors for refracture of the forearm 21. Muensterer OJ, Regauer MP. Closed reduction of forearm refrac-
in children treated with elastic stable intramedullary nailing. Int tures with flexible intramedullary nails in situ. J Bone Joint Surg Am.
Orthop. 2019;43:2093–2097. 2003;85:2152–2155.
on 04/15/2023
11. Korhonen L, Lutz N, Sinikumpu JJ. The Association of Metal 22. O’Neill CJ, Fitzgerald E, Kaar K, et al. Refracture of the pediatric
Frame Construct of ESIN and radiographic bone healing of pediatric forearm with intramedullary nails in situ. J Orthop Case Rep. 2019;9:
forearm fractures. Injury. 2020;51:856–862. 15–18.
12. Kruppa C, Bunge P, Schildhauer TA, et al. Low complication rate of 23. Van Egmond PW, van der Sluijs HA, Royen BJ, et al. Refractures of
elastic stable intramedullary nailing (ESIN) of pediatric forearm the paediatric forearm with the intramedullary nail in situ. Case Rep.
fractures: a retrospective study of 202 cases. Medicine (Baltimore). 2013;2013:bcr2013200840.
2017;96:e6669. 24. Zhamilov V, Reisoglu A, Basa CD, et al. Pediatric forearm
13. Makki D, Kheiran A, Gadiyar R, et al. Refractures following refracture with intramedullary nail bending in situ: options for
removal of plates and elastic nails from paediatric forearms. treatment. Cureus. 2020;12:e6744.
J Pediatr Orthop B. 2014;23:221–226. 25. McLean C, Adlington H, Houshian S. Paediatric forearm refractures
14. Pogorelić Z, Gulin M, Jukić M, et al. Elastic stable intramedullary with retained plates managed with flexible intramedullary nails.
nailing for treatment of pediatric forearm fractures: a 15-year single Injury. 2007;38:926–930.
centre retrospective study of 173 cases. Acta Orthop Traumatol Turc. 26. Sanderson PL, Ryan W, Turner PG. Complications of metalwork
2020;54:378–384. removal. Injury. 1992;23:29–30.

ORIGINAL ARTICLE
Ten-year Review of Acute Pediatric Hematogenous

Osteomyelitis at a New Zealand Tertiary Referral Center
Sarah Hunter, MBChB,* Haemish Crawford, MBChB, FRACS,†

and Joseph F. Baker, MCh, FRCSI*
childhood AHO within New Zealand and Australia was

Background: Acute hematogenous osteomyelitis (AHO) remains first explored by Gillespie and colleagues in the 1980s.
a cause of severe illness among children with the possibility of This research separated rates of AHO by ethnicity. For
long-term consequences for growth and development. Recent New Zealand Māori and Pacific the highest reported rate
research has highlighted an unusually high burden of disease in was 110/100, 000—more than twice the incidence seen in
the New Zealand population compared with other Western re- New Zealand European children.2 Australian Aboriginal
gions. We have sought to identify trends in presentation, diag- children are also more severely affected.3 Importantly, a
on 04/15/2023
nosis, and management of AHO, with added focus on ethnicity large case series of children treated in New Zealand be-
and access to health care. tween 1997 and 2008 suggests overall incidence is
Methods: A 10-year retrospective review of all patients <16 years declining.1
with presumed AHO presenting to a tertiary referral center be- Although repeated studies confirm an extremely high
tween 2008 and 2018 was performed. risk of disease within New Zealand Māori and Pacific
Results: One hundred fifty-one cases met inclusion criteria. The ethnicities, reasons for this are not fully understood.4,5
median age was 8 years with a male predominance (69.5%). Despite residing in generally well-resourced regions, these
Staphylococcus aureus was the most common pathogen using tra- communities experience poorer health outcomes driven by
ditional laboratory culture method (84%). The number of cases per inequitable access to health care and economic
year decreased from 2008 to 2018. Assessment using New Zealand disparity.6,7 Socioeconomic factors in New Zealand were
deprivation scores showed Māori children were most likely to ex- explored by Hunter et al8 looking specifically at rates of
perience socioeconomic hardship (P ≤ 0.01). Median distance septic arthritis (SA) with regard to deprivation index and
traveled by families to first hospital consult was 26 km (range 1 to distance from tertiary care. A higher deprivation index
178 km). Delayed presentation was associated with need for pro- was associated with development of SA, and the majority
longed antibiotic therapy. Incidence of disease varied by ethnicity of New Zealand Māori and Pacific patients were from the
with 1:9000 cases per year for New Zealand European, 1:6500 for most deprived deciles. Some of these patients also expe-
Pacific, and 1:4000 for Māori. Overall recurrence rate was 11%. rience more frequent skin and soft-tissue infection.9,10
Conclusions: The incidence of AHO in New Zealand is con- The spectrum of illness experienced by children with
cerningly high within Māori and Pacific populations. Future AHO is variable. Children with complicated disease may
health interventions should consider environmental, socio- require multiple surgeries and experience long-term se-
economic, and microbiological trends in the burden of disease. quelae of such as growth arrest, chronic infection, or
Level of Evidence: Retrospective study, Level III. avascular necrosis.11,12 Inequitable access to health care
Key Words: Acute hematogenous osteomyelitis, epidemiology, may exacerbate these outcomes. Studies to date suggest
pediatric bone and joint infection, Complications that the incidence of AHO has changed over time.13
Therefore, the aim of this study was to determine the in-
(J Pediatr Orthop 2023;43:e396–e401) cidence of the AHO in our population over the last
10 years. We have also sought to identify trends in pre-
sentation, diagnosis, and management of AHO, with
added focus on ethnicity and access to health care.
T he incidence of childhood acute hematogenous osteo-
myelitis (AHO) within New Zealand has not been
formally assessed in the last decade.1 The epidemiology of METHODS
From the *University of Auckland, Department of Orthopaedic Surgery,
Health and Disability Ethics Committee (HDEC)
Waikato Hospital, Hamilton, New Zealand; and †Starship Children’s approval was obtained for this study (reference: 21/STH/
Hospital, Auckland District Health Board, Auckland, New Zealand. 89). As all data were collected in retrospect patient consent
No sources of financial support to declare. was not required.
The authors declare no conflicts of interest. This cohort study is a retrospective review of all
Reprints: Sarah Hunter, MBChB, Pembroke St, Waikato Hamilton.
PH:02102520645. E-mail: shun472@aucklanduni.ac.nz. cases of suspected osteomyelitis presenting to our in-
Copyright © 2023 Wolters Kluwer Health, Inc. All rights reserved. stitution between 2008 and 2018. Children from newborn
DOI: 10.1097/BPO.0000000000002385 to age 15 years were included. Osteomyelitis was defined

J Pediatr Orthop Volume 43, Number 5, May/June 2023 Epidemiology of AHO in New Zealand
based on radiographic investigation through magnetic infection of the skull or maxillofacial bones (n = 3),
resonance imaging (MRI) or computed tomography (CT) insufficient data (n = 7), or postoperative infection (n = 7).
and/or positive intraoperative culture or bone biopsy. The remaining 151 cases were considered suitable.
Cases of post-viral or reactive arthritis, postoperative in- The median age was 8 years with a range from neonatal to
fection, or patients with insufficient clinical data for 15 years. The majority were male (69.5%) and the primary
analysis have been excluded. Children with a primary di- ethnicity was Māori (46.4%) followed by New Zealand
agnosis of osteomyelitis with contiguous infection of the European (45%) (Table 1). There were only 9 Pacific
adjacent joint or muscle were included. Comprehensive patients. Electronic record characterized these patients as
review of electronic clinical records was conducted. All Tongan, Samoan, or Polynesian.
patients were followed for a minimum of 2 years. Incidence of disease varied by ethnicity with 1:9000
Data were collected on patient demographics (age, cases per year for New Zealand European, 1:6500 for
sex, ethnicity), presenting complaint, time to presentation, Pacific, and 1:4000 for Māori (Table 1). Considering
and use of local primary care facilities. Deprivation index, incidence over time, the number of cases per year
a surrogate marker for socioeconomic hardship, was also decreased from 2008 to 2018 (Fig. 1). There was a
collected. This is an area-based score from 1 to 10 used in median number of 16 cases/year over the first 3 years of
New Zealand population statistics, where a score of 10 the study and a median of 8 cases/year in the last 3 years.
correlates to significant hardship. It has been calculated MRI use remained relatively constant over the time
here using patient address at the time of admission. period, utilized in 50% to 80% of cases each year.
Demographic analysis was undertaken using the 2013 Evaluating patient presentations, the primary com-
on 04/15/2023
New Zealand Census. Incidence of disease by ethnicity plaint was painful limb or joint (85%) and approximately
was also calculated using data extracted from the 2013 half were febrile (52%). A history of trauma was docu-
census.7 Duration of symptoms before admission classed mented in 35 cases (23%). The most common comorbidity
as <24 hours, 24 to 72 hours, > 72 hours to <2 weeks, or was asthma (9%) followed by eczema or impetigo (7%).
> 2 weeks. Laboratory results included C-reactive protein Anatomic location of osteomyelitis in our cohort is shown
(CRP mg/L), white cell count (L), Hemoglobin (g/dl), al- in Fig. 2.
bumin (g/dL), and microbiological samples from either Patients traveled a median of 26km ( ± 17.2km,
joint aspirate, blood culture, or intraoperative specimen. range 1 to 178 km) to reach an Emergency Department for
Standard agar plate culture was used for synovial fluid assessment. Just over a third of patients waited <72 hours
aspirate, positive result defined by positive gram stain, cell before presenting to the hospital (35%); nearly a quarter
count > 50,000 mm,3 or growth of pathogen on culture. attended primary care consultations before seeking as-
No samples have undergone 16sPCR at our institution sessment at an Emergency Department (22.5%). Waiting
during the study period. Synovial fluid culture with BD > 72 hours before presentation was associated with in-
Bactec bottles is not routinely undertaken at our increased length of antibiotic treatment but was not asso-
stitution. ciated with other markers of disease severity (Table 2).
Information regarding diagnostic method was classi- The median CRP on arrival was 64mg/L ( ± 93) and
fied as surgical, clinical, or radiographic. For the purposes white cell count was 11.3 ( ± 4.8). Median CRP for tended
of this study, a surgical diagnosis required positive intra- to fall over the first 96 hours of admission, from 64 mg/L
operative findings and/or growth of organism on cultured on admission to 29 mg/L ( ± 82) at day 4 to 5. An or-
operative samples. Radiographic diagnosis was made if the ganism was isolated in just over half of cases (56%) from
patient had findings of osteoarticular infection on MRI, either positive blood cultures (44%), positive aspirate
CT, or bone scan. A positive clinical diagnosis is defined as result and/or a positive intraoperative sample (33%).
diagnosis made using history, examination findings, and Staphylococcus aureus was the most frequent
laboratory results but in the absence of advanced imaging pathogen isolated (84% of positive samples) followed by
or operative samples. MRSA (13%). Kingella kingae was only identified in 1
sample (Fig. 3).
Statistics The majority of patients were from low socio-
The incidence of AHO by year for our catchment economic regions with a combined median deprivation
area was analyzed and calculated using publicly available index of 7. Subgroup analysis comparing Māori patients
census record.7 Univariate analysis was undertaken to to New Zealand European/other ethnicities confirmed
assess for differences in disease severity for children pre-
senting later than 72 hours after symptom onset. Finally,
an ethnicity subgroup analysis was performed to evaluate TABLE 1. Incidence of Disease by Ethnicity, Patients With AHO
for differences in presentation, management, and out- Ethnicity Number in cohort Population/census year Incidence
comes of AHO seen in Māori and Pacific patients.
NZE: 68 61,374 1:9000
Maori: 70 28,659 1:4000
RESULTS Pacific: 9 5850 1:6500
An initial 204 cases were identified by clinical cod- Total 151 87,483 1:5800
ing. Patients were excluded if they had chronic or subacute AHO indicates acute hematogenous osteomyelitis; NZE, New Zealand European.
infection (n = 8), an incorrect diagnosis of OM (n = 28),

Hunter et al J Pediatr Orthop Volume 43, Number 5, May/June 2023
25
TABLE 2. Association Between Delayed Presentation and
Worse Disease for Children With AHO
Number of Cases
20
15 Delay to presentation and Delay Delay CHI2/

markers of worse disease > 72 h < 72 h TTEST
10
Median deprivation index 6.8 7.4 0.9
5 Median distance from hospital 48 42 0.39

in Km
0 Median antibiotic treatment 42 36 0.01
2006 2008 2010 2012 2014 2016 2018 2020 in days
Years Median length of stay in days 10 13 0.3
Repeat surgical procedure 14 (68) 9 (74) 0.17
Recurrent disease 9 6 0.32
FIGURE 1. Number of cases of acute hematogenous osteo- AHO indicates acute hematogenous osteomyelitis.
myelitis per year, 2008 to 2018.
response to prolonged fevers, failure to resolve in-
generally worse socioeconomic conditions for these
flammatory markers, or identification of a drainable col-
children (P < 0.001) (Table 3). There were no differences
lection on subsequent advanced imaging. The median
with regard to presentation characteristics, severity
duration of IV antibiotics was 7 days ( ± 5.6) with a
of illness, or complication rate. The duration of
combined oral and IV median treatment period of 42 days
on 04/15/2023
intravenous (IV) antibiotic treatment was shorter for

( ± 12). Of the children treated surgically, 31% required
Māori (2.5 vs.7 d, P = 0.3) but this did not reach statistical
multiple surgeries. The median delay to surgical treatment
significance.
was 2 days ( ± 2.9). Median length of hospital stay was
A final diagnosis of AHO was made radio-
8 days ( ± 21.3).
graphically (through CT or MRI) in 79% of cases, surgi-
After treatment, 24 children were re-admitted within
cally in 5%, and clinically in 17%. Use of advanced
30 days, and a further 18 within 1 year. Recurrence of
imaging revealed subperiosteal abscess in 29%, concurrent
infection was seen in 16 cases (11%) of which 8 children
SA in 3%, and pyomyositis in 5% of cases. Co-existing
required surgical treatment. There were a total of 32
infections were present in 17 (11%) with respiratory tract
documented chronic complications of infection including
infection occurring the most frequently (5%) (Table 3).
chronic osteomyelitis (5%), growth disturbance (4%),
All children were treated with a minimum of 4 to 6
pathologic fracture (3%), or femoral head avascular ne-
weeks of IV antibiotics. Together with antibiotic treat-
crosis secondary to adjacent SA (1%) (Table 4).
ment, 34 (22.5%) children had surgical management at the
time of admission. In addition to this, some children
converted to surgical management after an unsuccessful DISCUSSION
attempt to control infection with IV antibiotics alone We have defined the incidence and presenting fea-
(22.5%). This decision was made by the treating team in tures of pediatric AHO over a 10-year period at our in-
stitution. Although similar to previous reports there was a
OSTEOMYELITIS LOCATION trend of decreasing incidence across the study period, the
incidence of AHO in our cohort is among the highest in
Pelvis
clavicle 8% the world, with disparate burden of disease seen in more
2% deprived and indigenous communities.1 Importantly, the
Radius majority of children in our study reside in the most de-
Femur
3%
27% prived 3 deciles (51%). The highest rates of AHO are for
Humerus
5% Māori children, sitting at 1:4000 cases. From a recently
global review of epidemiology, it seems that the incidence
of AHO is now decreasing in New Zealand and the in-
cidence of SA is rising.13 The decreasing incidence of
Fibula
10% AHO in our population is reflective of trends in other
developed countries.13 Potential reasons for this include
Patella
1% better health literacy, effective national health care sys-
Phalanges
tems, and efficient primary care.
3% S. aureus was the most frequent pathogen and
Talus MRSA was rarely isolated in AHO cases. This relative
1%
absence of MRSA differs strongly from disease patterns
Calcaneus seen within the United States and Europe.13 Importantly,
7%
Tibia
the low rates of MRSA have not resulted in a reduced
Metatarsals complication rate or reduced severity of disease. There are
28%
5%
also no significant differences in pathogen type by eth-
FIGURE 2. Anatomic location of osteomyelitis. nicity. Māori and Pacific children did not experience

FIGURE 3. Organisms isolated from laboratory or intraoperative samples. Dark blue: S. Aureus 84%, Light blue: MRSA (13%),
Orange: S. pyogenes 1%, Gray: Coagulase negative staphylococcus (1%), Yellow: K. kingae (1%)
on 04/15/2023
higher rates of MRSA; other factors should be considered The cohort of patients treated at our institution
to explain the disproportionate burden of disease in these traveled between 1 and 178 km to reach an ED for
populations.8 assessment. Delayed presentation was assessed for a
K. kingae, although an emerging pathogen in Europe, relationship with worse disease showing that children who
was only found in 1 case of AHO.14 This is very likely to be wait > 72 hours before medical assessment require longer
a result of the lack of PCR testing at our institution rather courses of antibiotic treatment. The association between
than a true observation.15 Use of PCR analysis will enable distance traveled to the ED and length of stay/IV anti-
an examination of S. aureus virulence factors such as biotic duration for children with AHO demonstrates a key
panton-valentine leukocidin and give a more accurate es- aspect of practical access to health. Our region, although
timate of infection caused by K. kingae.16 Both of these covering a relatively small catchment of 400,000 patients,
are crucial to understanding the patterns of disease in our traverses more than 21,000 km2. Considering the range of
region. distances families may be required to travel, some more
Māori children within the cohort tend to reside in the than 150 km, care must be taken in prompt diagnosis and
most deprived areas (median NZDep = 9). This is also a treatment. With recent interest in telemedicine due to the
reflection of population trends, whereby 60% of Māori in global pandemic, virtual options could be considered to
New Zealand live below the poverty threshold.6 Although help reduce this inequity.
late presentation is sometimes cited as a reason for worse MRI use in the cohort of children with AHO re-
health outcomes, the data from our research confirms that mained constant over the time period. This may be due to
Māori families did not present significantly later to hos- a combination of factors. Osteomyelitis may only be de-
pital or seek less primary care frequently. Previously, tectable in plain radiographs 2 weeks after the onset of the
cultural factors in particular household overcrowding has disease—those with delayed presentation may have ob-
been used as an explanation for higher rates of S. aureus vious radiographic findings.20 Where the diagnosis is un-
carriage in both Māori and Pacific patients.17 High rates clear, advanced imaging is necessary.20 Alternatively, if a
of S. aureus colonization in New Zealand increase the child presented early with classical features of AHO (in the
likelihood of community-onset skin and soft-tissue absence of any advanced imaging or operative inter-
infection.10 The relationship between skin infection and vention), initiation of IV antibiotics could lead to clinical
socioeconomic factors has been demonstrated in previous improvement. For these children, a presumptive “clinical”
literature from New Zealand. Māori and Pacific children diagnosis of AHO was made. This method of diagnosis
were 2.9 and 4.5 times, respectively, more likely to be was used in 16.5% of cases. This creates a conundrum.
admitted to hospital for impetigo between 2000 and 2007. Although MRI is generally preferred for early identi-
Rates of infection are linked to greater deprivation with fication of AHO, it remains a limited resource in some
almost 3 times the rates of serious skin infections in centers.21 For children with a milder course of illness, in-
NZdep index 9 to 10 versus children without significant itiation of IV antibiotics may be sufficient for diagnosis.
deprivation.18 However, foregoing advanced imaging creates risk in
Socioeconomic determinants of health also have a cases where subperiosteal abscess or pyomyositis require a
clear association with invasive community-onset S. aureus, different treatment plan. Delayed surgical management
further supporting a correlation between deprivation and could also lead to worsening clinical picture and risk of
infection.19 complication.22 Utilization of MRI has been shown to

Hunter et al J Pediatr Orthop Volume 43, Number 5, May/June 2023
advanced imaging for children with AHO will be

TABLE 3. Subgroup analysis Māori patients, Children
With AHO extremely useful.
Identification of those requiring surgical procedures
Māori NZ European/other
remains a challenge. Primary surgical treatment was used
Variables (n = 70) (n = 72) Significance
in 34 AHO cases. However, a further 34 patients sub-
Median age 8 8 0.78 sequently converted to surgical management after an un-

Male 47 49 0.9
Median deprivation 9 7 < 0.01
successful course of IV antibiotics. This decision was made
index by the surgical team, usually due to worsening clinical
Median distance 40 25 0.32 picture or identification of drainable collection on ad-
traveled (km) vanced imaging. There remains no validated criteria for
Deprivation index 1-4 1 25 < 0.01 identifying those likely to experience severe course of
Deprivation index 5-7 21 24 0.54
Deprivation index 8-10 48 23 < 0.01 illness and require surgery in cases of AHO.24,25
> 72 h delay to 34 30 0.4 The outcomes after treatment for AHO suggest a
presentation relatively high recurrence rate of 10.6% with 8.6% needing
Sought primary care 28 27 0.75 re-admission for surgery—this is key data to allow suitable
Contiguous disease 23 22 0.08
Median CRP on arrival 68 59.5 0.63
patient and family counseling at the time of diagnosis and
Multifocal infection 7 7 0.96 the outset of treatment. Development of chronic osteo-
Received MRI 51 46 0.25 myelitis occurred in 5.3% of cases, which is in keeping with
Surgical treatment 30 31 0.98 trends in Western regions identified in systematic review.13
Median days until 2 1 0.72
on 04/15/2023
Follow-up remains a challenge in our region with 26.% of

surgery
Non-op to surgical 9 16 0.14 children not brought to outpatient clinics despite multiple
Median total Abx 42 42 0.55 attempts at contact.
duration We acknowledge this study has limitations. As with
Median IV abx 2.5 10 0.3 all retrospective analyses, the findings are only as strong
duration
> 1 clinic non- 11 6 0.17
as the documentation by the clinicians in the medical
attendance records. Ethnicity in our health care system relies on self-
Recurrence 7 8 0.83 reporting by the individual rather than any confirmation
Acute complication 26 24 0.63 by genetic typing. We have used census data to calculate
Chronic complication 7 11 0.34 the incidences of disease—this is as accurate as we can
Median LOS 9 8 0.43
Abx resistance 8 5 0.34 possibly be with our calculations but we do note that the
MRSA 6 3 0.28 New Zealand population is somewhat itinerant and
Re-admitted within 30 d 11 11 0.94 that the population number may be slightly different in
Abx indicates antibiotics; AHO, acute hematogenous osteomyelitis; CRP,
reality.
C-reactive protein, LOS, length of stay; MRI, magnetic resonance imaging; In summary, although there was a decrease in the
Non-op, non-operative; NZ, New Zealand. incidence of pediatric AHO over the 10-year study pe-
riod, the incidence remains concerningly high, partic-
potentially reduce unnecessary return trips to the operat- ularly in Maori and Pacifica cohorts. A significant
ing room and shorter length of stay in some number of patients require re-admission to hospital
publications.21–23 Further research on the impact of over the subsequent year. Public health measures are
urgently required to improve our understanding of risk
factors for the development of pediatric bone and joint
TABLE 4. Treatment and Outcomes for Children With AHO infections.
Median %/SD
Outcomes after treatment for osteomyelitis
REFERENCES
Median length of stay in days 8 21.3 1. Street M, Puna R, Huang M, et al. Pediatric acute hematogenous
Re-admission within 30 d 24 15.9 osteomyelitis. J Pediatr Orthop. 2015;35:634–639
Re-admission within 1 y 18 11.9 2. Gillespie WJ. The epidemiology of acute haematogenous osteomye-
Recurrence of infection 16 10.6 litis of childhood. Int J Epidemiol. 1985;14:600–606.
Re-admission for surgical 13 8.6 3. Brischetto A, Leung G, Marshall CS, et al. A Retrospective case-
debridement series of children with bone and joint infection from northern
Was not brought to OPC 40 26.5 Australia. Medicine (Baltimore). 2016;95:e2885.
Complications 4. Gillespie WJ. Racial and environmental factors in acute haematog-
Chronic osteomyelitis 8 5.3 enous osteomyelitis in New Zealand. N Zeal Med J. 1979;90:93–95.
Pathologic fracture 5 3.3 5. Williamson DA, Coombs GW, Nimmo GR. Staphylococcus aureus
Growth disturbance 6 4.0 “Down Under”: Contemporary epidemiology of S. aureus in
Lytic lesion 5 3.3 Australia, New Zealand, and the South West Pacific. Clin Microbiol
Chronic pain 5 3.3 Infect. 2014;20:597–604.
Femoral head collapse 1 0.7 6. Child Poverty Statistics. Stats NZ Tatauranga Aotearoa. Published
Leg length discrepancy 2 1.3 2020. Accessed January 19, 2022. https://www.stats.govt.nz/news/
latest-release-of-child-poverty-statistics
AHO indicates acute hematogenous osteomyelitis; OPC, outpatient clinic. 7. NZ S. 2018 Census and Dwelling Counts. Stats NZ Tatauranga
Aotearoa. Published 2018. Accessed November 26, 2019. https://

www.stats.govt.nz/information-releases/2018-census-population-and- 17. Williamson DA, Zhang J, Ritchie SR, et al. Staphylococcus aureus
dwelling-counts-nz-stat-tables infections in New Zealand, 2000-2011. Emerg Infect Dis. 2014;20:
8. Hunter S, Baker JF. Ten-year retrospective review of paediatric 1156–1161.
septic arthritis in a New Zealand centre. Int Orthop. 2021;45: 18. O’SULLIVAN CE, BAKER MG, ZHANG J. Increasing hospital-
147–154. izations for serious skin infections in New Zealand children,
9. Hobbs MR, Grant CC, Thomas MG, et al. Staphylococcus aureus 1990–2007. Epidemiol Infect. 2011;139:1794–1804.
colonisation and its relationship with skin and soft tissue infection in 19. Williamson DA, Lim A, Thomas MG, et al. Incidence, trends and
New Zealand children. Eur J Clin Microbiol Infect Dis. 2018;37: demographics of Staphylococcus aureus infections in Auckland, New
2001–2010. Zealand, 2001-2011. BMC Infect Dis. 2013;13:569.
10. Williamson DA, Ritchie SR, Lennon D, et al. Increasing incidence 20. Manz N, Krieg AH, Heininger U, et al. Evaluation of the current use
and sociodemographic variation in community-onset staphylococcus of imaging modalities and pathogen detection in children with
aureus skin and soft tissue infections in new zealand children. Pediatr acute osteomyelitis and septic arthritis. Eur J Pediatr. 2018;177:
Infect Dis J. 2013;32:923–925. 1071–1080.
11. Martin AC, Anderson D, Lucey J, et al. Predictors of outcome in 21. Schallert EK, Herman Kan J, Monsalve J, et al. Metaphyseal osteomyelitis
pediatric osteomyelitis: five years experience in a single tertiary in children: how often does MRI-documented joint effusion or epiphyseal
center. Pediatr Infect Dis J. 2016;35:387–391. extension of edema indicate coexisting septic arthritis? Pediatr Radiol.
12. Sukswai P, Kovitvanitcha D, Thumkunanon V, et al. Acute 2015;45:1174–1181.
hematogenous osteomyelitis and septic arthritis in children: clinical 22. Griswold BG, Sheppard E, Pitts C, et al. The Introduction of a
characteristics and outcomes study. J Med Assoc Thai. 2011; Preoperative MRI Protocol Significantly Reduces Unplanned Return
94(suppl 3):S209–S216. to the Operating Room in the Treatment of Pediatric Osteoarticular
13. Hunter S, Chan H, Baker JF. Global epidemiology of childhood Infections. Journal of Pediatric Orthopaedics. 2020;40:97–102.
bone and joint infection: a systematic review. Infection. 2022;50: 23. Connolly SA, Connolly LP, Drubach LA, et al. MRI for detection of
329–341. abscess in acute osteomyelitis of the pelvis in children. Am J
on 04/15/2023
14. Samara E, Spyropoulou V, Tabard-Fougère A, et al. Kingella kingae Roentgenol. 2007;18910.2214/AJR.07.2416

and osteoarticular infections. Pediatrics. 2019;144:e20191509. 24. Hunter S, Kioa G, Baker JF Predictive algorithms in the diagnosis
15. van den Boom M, Lennon DR, Crawford H, et al. Microbiology of and management of pediatric hip and periarticular infection.
septic arthritis in young Auckland children. J Paediatr Child Health. doi:10.2106/JBJS.21.01040
2022;58:326–331. 25. Alhinai Z, Elahi M, Park S, et al. Prediction of adverse outcomes in
16. Yagupsky P. Kingella kingae: carriage, transmission, and disease. pediatric acute hematogenous osteomyelitis. Clin Infect Dis. 2020;71:
Clin Microbiol Rev. 2015;28:54–79. e454–e464.

ORIGINAL ARTICLE
The Effect of Epiphysiodesis on the Longitudinal Bone

Growth of Hands or Feet in Children With Macrodactyly
Based on Long-term Quantitative Analysis

Soo Jin Woo, MD,* Jae Hoon Jung, MD,† Jun Ho Choi, MD,‡ Yumin Kim, MD,*
Sung Tack Kwon, MD, PhD,* and Byung Jun Kim, MD, PhD*
Background: Treatment protocols for macrodactyly have not

been elucidated due to its rarity and variety of clinical manifes-
tations. This study aims to share our long-term clinical results of
M acrodactyly is a rare congenital overgrowth dis-
order, which is characterized by enlargement of all
the elements of the affected digits. Isolated macrodactyly
epiphysiodesis in children with macrodactyly.
is distinguished from other conditions with secondary
on 04/15/2023
Methods: A retrospective chart review was performed for 17

hypertrophy of digits such as overgrowth syndromes or
patients with isolated macrodactyly treated with epiphysiodesis
neoplastic lesions.1,2 Recent evidence suggests a correla-
over 20 years. Length and width of each phalanx in both the
tion of PIK3CA gene mutations with isolated macro-
affected finger and the corresponding unaffected finger in the
dactyly [31].
contralateral hand were measured. Results were presented in
Macrodactyly comprises a wide array of phenotypes
ratios of the affected to unaffected side for each phalanx. Mea-
in which the extent of hypertrophy differs significantly
suring of length and width of phalanx was performed pre-
between patients. Many attempts have been made to
operatively and postoperatively at 6, 12, and 24 months, and the
classify this haphazard disease,3,4 and the most commonly
last follow-up session. Postoperative satisfaction scoring was
applied one divides the condition into 2 types based on
done with visual analogue scale.
natural history.5 The progressive type shows a rapid grow
Results: The mean follow-up period was 7 years and 2 months. In
of bone and adipose tissue, whereas the static type displays
the proximal phalanx, length ratio significantly decreased com-
proportionate growth as the child grows.5,6
pared with preoperative state at after more than 24 months, in
Treatment for macrodactyly is categorized according
the middle phalanx after 6 months, in the distal phalanx after
to the patients’ growth potential and individualized accord-
12 months. When classified by the growth patterns, the pro-
ing to their clinical state.7,8 In fully developed macrodactyly
gressive type showed significant decrease in length ratio at after
without growth potential, the surgery aims for size reduction
6 months, and the static type after 12 months. Patients were
by operations such as ostectomy, mid-phalanx shortening
overall satisfied with the results.
(Barsky method), or amputation. On the contrary, for chil-
Conclusion: Epiphysiodesis effectively regulated longitudinal
dren with remnant growth capability, epiphysiodesis is a
growth with different degree of control for different phalanges in
surgical option for growth inhibition.9–11
the long-term follow-up.
As various treatment methods are applied sequen-
Key Words: macrodactyly, growth plate, bone development, tially and simultaneously, it is not easy to distinguish the
long-term care, quantitative evaluation effect of any one procedure. Epiphysiodesis has shown
mixed results with a limited number of patients and in-
(J Pediatr Orthop 2023;43:e363–e369) sufficient quantitative analysis in previous studies.12,13 To
the best of our knowledge, this is the first study to evaluate
From the *Department of Plastic and Reconstructive Surgery, Seoul long-term results of sole effect of epiphysiodesis in patients
National University Hospital, Seoul National University College of with macrodactyly. The purpose of this study is not to
Medicine; †Rex Plastic Surgery Clinic; and ‡345 Plastic Surgery recommend one treatment method but to determine
Clinic, Seoul, Republic of Korea. quantitively how effective epiphysiodesis is. The in-
S.J.W. and J.H.J. equally contributed to this paper as first authors.
S.T.K. and B.J.K. equally contributed to this paper as corresponding
formation will assist hand surgeons in determining
authors. treatment plans and providing counseling to patients or
None of the authors received financial support for this study their guardians regarding the postoperative course.
Reprints: Sung Tack Kwon, MD, PhD (E-mail: stk59@snu.ac.kr); and
Byung Jun Kim, MD, PhD (E-mail: bjkim79@gmail.com), Depart-
ment of Plastic and Reconstructive Surgery, Seoul National Uni-
METHODS
versity Hospital, Seoul National University College of Medicine, 101
Daehak-ro, Jongno-gu, Seoul 110-744, Republic of Korea.
Patients and Data Collection
Copyright © 2023 Wolters Kluwer Health, Inc. All rights reserved. This study was approved by the Institutional Review
DOI: 10.1097/BPO.0000000000002387 Board of Seoul National University College of Medicine

Woo et al J Pediatr Orthop Volume 43, Number 5, May/June 2023
(H-1507-103-689). Written informed consent was obtained surgical treatment, the most severely involved phalanges
for publication. Between January 2000 and December were selected for epiphysiodesis based on X-ray. If nec-
2019, 45 patients with isolated macrodactyly in the hand essary, surgery was performed on another phalanx after
or foot underwent surgical treatment in our department by observing the progress. In the serial procedure, we could
a single surgeon (S.T.K.). Patients who underwent epi- minimize unnecessary problems of performing epi-
physiodesis without any other bone procedures such as physiodesis simultaneously on all phalanxes. It also re-
ostectomy were included in the study. Patients who un- duces overall operation time and facilitates postoperative
derwent soft-tissue debulking involving nerve resection, management.
angulation correction, or repeated epiphysiodesis on the
same phalanx due to failure were excluded. Bilateral in- Surgical Methods
volved cases and patients with postoperative period of A longitudinal incision was made on the dorsal side
<2 years were excluded (Fig. 1). Medical records including of the target epiphyseal plate. After retraction of the ex-
sex, age at operation, growth pattern (progressive or tensor tendon, the periosteum was elevated to expose the
static type), treatment detail, postoperative result, and epiphysis under direct vision or under intraoperative flu-
complication were retrospectively reviewed. oroscopy if necessary. The physis was ablated using a
The length and width of each phalangeal bone in the surgical blade or electrical saw. Two adjacent phalanges
involved digits were measured on anterior-posterior hand were fixated with Kirschner wires inserted crossing the
or foot X-rays (Fig. 2). Anterior-posterior X-rays were epiphysis and joint. After repairing the periosteum with
taken at preoperative, postoperative 6, 12, and 24 months, 4-0 Ethibond (Ethicon, Somerville, NJ), the skin was re-
paired with 5-0 Rapid Vicryl (Ethicon, Somerville, NJ). A
on 04/15/2023
and every outpatient session after 2 years. In each patient,

the affected side was compared with the same digit of the splint was applied for immobilization for 6 weeks and
unaffected hand in ratio. Length (or width) ratio was Kirschner wires were removed with confirmation of bone
calculated as length (or width) of an affected phalangeal union on X-ray.
bone to unaffected phalangeal bone. The closer the ratio
was to 1.0, the greater the similarity between the affected Statistical Methods
and unaffected sides. Statistical analysis was performed using the SAS pro-
As there are no validated scoring systems to eval- gram (version 9.2; SAS Institute Inc., Cary, NC). A linear
uate the satisfaction after treatment of macrodactyly, a mixed-model method was adopted to investigate post-
simple and practical scoring system in a clinical setting operative changes at 5-time points in terms of 2 independent
was applied.14 The visual analogue scale for satisfaction variables: the treated phalanges (proximal vs. middle vs.
is composed of a 100 mm long horizontal line with two distal), growth types (progressive vs. static). P-values <0.05
descriptors at each end, representing 0 (no satisfaction) were considered statistically significant. Statistical analyses
and 100 (extreme satisfaction). The patient or the were supported by the Medical Research Collaborating
guardian rated satisfaction by making a vertical mark Center of Seoul National University Hospital.
based on the functional and aesthetic results of the sur-
gery. RESULTS
Patients were indicated for surgery when the in- Detailed information about the clinical and surgical
volved digit’s total length reached 75% or more of the characteristics of patients is summarized in Table 1. The
same digit in the same-sex parent considering further mean age at first epiphysiodesis procedure was 4.5 years
growth potential. After the patient decided to undergo and second was 5.7 years. In 4 patients, more than 1 digit
FIGURE 1. Flowchart of patient inclusion and exclusion.

J Pediatr Orthop Volume 43, Number 5, May/June 2023 Long-term Clinical Results of Epiphysiodesis
on 04/15/2023
FIGURE 2. Method of calculating the length ratio and width ratio on X-ray.
was involved. Six patients underwent additional DISCUSSION

epiphysiodesis on the remnant physis. There were no Macrodactyly, an isolated congenital overgrowth
complications such as wound infection, dehiscence, and of the hand or foot, is a condition devastating for pa-
bone nonunion (Fig. 3). tients and challenging for surgeons. It has been identi-
After epiphysiodesis, the length ratio showed a ten- fied as a part of the PIK3CA Related Overgrowth
dency to decrease in all phalanges. The length ratio of each Spectrum which is due to a somatic gain-in-function
phalanx was compared before and after surgery at 4 specific mutation in PIK3CA in mTOR pathway.15 It is classi-
time points for significant differences (Fig. 4A). Results fied according to its growth pattern; the progressive
showed that in the proximal phalanx, length ratio type shows a rapid growth of bone and adipose tissue,
significantly decreased at after more than 24 months whereas the static type displays proportionate growth as
nths (P < 0.001), in the middle phalanx after 6 months the child ages.5,6 The treatment plan considers the se-
(P < 0.01), in the distal phalanx after 12 months (P < 0.05). verity and remnant growth potential of the affected
Next, the length ratio of each phalanx was analyzed for the digits. Common surgical options include epi-
difference from 1.0 at each time point to confirm the physiodeses, osteotomies, soft-tissue debulking, nerve
similarity to the contralateral phalanx length. The proximal resection, carpal tunnel releases, toe transfers and
phalanx was not significantly different from 1.0 after amputations.16 Recently, sirolimus, an mTOR inhibitor
24 months (P < 0.05), middle phalanx after 12 months is considered as an adjuvant treatment for isolated limb
(P < 0.05), whereas distal phalanx was constantly overgrowth conditions.17 However, due to rare in-
significantly different from 1.0 postoperatively (P < 0.05). cidence and variability in clinical features of macro-
Regarding the width ratio, there was no common dactyly, there are limited long-term quantitative data
tendency between different phalanges (Fig. 4B). and standard treatment protocols. In the study, we
Statistically significant results were shown in the performed epiphysiodesis in patients with growth po-
proximal phalanx, the width ratio significantly increased tential, and investigated postoperative growth in terms
after more than 24 months (P < 0.05). The width ratio of of the longitudinal and appositional growth. To the best
the middle phalanx was significantly increased at 12 and of our knowledge, no other article has analyzed the sole
24 months (P < 0.05). effect of epiphysiodesis in macrodactyly with quantita-
Cases were also classified by growth patterns tive data over a long-term follow-up period.
(Fig. 5). The length ratio was compared before and after True macrodactyly presents with hyperplasia of all
surgery at 4 specific time points. Both patterns showed a the tissue elements of the affected digits, including the
tendency to decrease. In the progressive type, length ratio phalanges, tendons, skin, subcutaneous fat, nerves, and
significantly decreased after 6 months compared with vessels. Hypertrophy is not always equal in all elements.
preoperative state, and in the static type at 12 months. Previous studies show a tendency for more severe hyper-
Patients and guardians were satisfied with the overall trophic involvement in the distal phalanges.5,7,9 We also
appearance and function with average score of 74.4 in observed that phalanges became more severely affected
static type and 61.7 in progressive type, respectively. when they were closer to the distal end (Fig. 4A).

on 04/15/2023
e366 | www.pedorthopaedics.com
Woo et al
TABLE 1. Demographic and Clinical Data of the 17 Patients

First epiphysiodesis Second epiphysiodesis Third epiphysiodesis
Age at Age at Age at Follow-up
Involved Affected Affected Growth Operated operation Operated operation Operated operation period Satisfaction
Case Sex limb side digit pattern phalanx* conducted (y) phalanx* conducted (y) phalanx* conducted (y) (mo) score (VAS)
1 F Hand Right LF S P1, P2, P3 11.3 — — — — 34 90
2 F Hand Left T, IF S IF P1, IF P2 5.0 IF P3 7.6 — — 115 70
3 M Hand Left IF S P2, P3 6.5 — — — — 70 80
4 F Hand Left LF S P1, P2, P3 11.2 — — — — 63 60
5 F Hand Left LF S P2 0.5 — — — — 23 70
6 F Hand Left T, IF P T P1, T P3 6.2 — — — — 36 60
7 F Foot Right 2T S P1, P2 1.8 — — — — 93 80
8 F Foot Right 2T S P1, P2, P3 4.3 — — — — 201 90
9 M Foot Right 2T S P3 6.3 — — — — 31 80
10 F Foot Left 2T S P2 7.3 P3 8.8 P1 13.2 72 50
11 F Foot Left 2T S P1, P2 5.9 — — — — 44 —
12 F Foot Right 2T P P2, P3 1.5 P1 2.6 — — 96 50

13 F Foot Right BT, 2T P BT P3 1.8 BT P1 3.8 2T P1 7.8 127 40
— — — —
J Pediatr Orthop
14 F Foot Right 2T, 3T P 2T P2, 3T P2 1.4 178 70
15 F Foot Right 2T P P2, P3 1.9 — — — — 156 80
16 M Foot Left 3T P P1, P2 1.2 P3 6.6 — — 78 —
17 F Foot Left BT, 2T P BT P1, 2T P1 2.6 BT P3 4.9 — — 42 70
*Operated digit was mentined in the front when multiple digits were involved
2T indicates second toe; 3T, third toe; BT, big toe; F, female; IF, index finger; LF, long finger; M, male; MT, metatarsal; P1, proximal phalanx; P2, middle phalanx; P3, distal phalanx; T, thumb; VAS, visual analogue
scale.

on 04/15/2023
FIGURE 3. (Case 1) A 11-year-old female patient with macrodactyly of her right long finger. Preoperative (A) and postoperative (C)
medical photographs after 34 months showing good postoperative results. The patient underwent epiphysiodesis on the proximal,
middle, and distal phalanges. Radiologic images also showing good postoperative results (D) compared with preoperative state (B)
without surgery-related complications, impairment in joint movement or evidence of arthritis until the last follow-up.
FIGURE 4. Serial change of the length and width ratio of affected to unaffected in proximal, middle, and distal phalanges at 5
different time points. A, Serial change of the length ratio. B, Serial change of the width ratio. ‘a)’ Statistical different when compared
with preoperative versus postoperative time point with a P-value < 0.05. ‘b)’ Statistical different when compared for difference from
1.0 with a P-value < 0.05. (‘b)’ mark was omitted in in Figure 4B, as all values were statistically different when compared for difference
from 1). M indicates month; Preop, preoperatively; Long-term, long-term follow-up more than 24 months.

Woo et al J Pediatr Orthop Volume 43, Number 5, May/June 2023
static types without statistically significant differences.

Previous studies have reported that the progressive type
take more effort to treat than the static type. Patients with
the progressive type undergo surgery earlier18 and more
frequently,4,9 and they even undergo aggressive methods
such as amputation.5,18 Our results were compatible with

previous findings. The average age at the first epi-

physiodesis was earlier in patients with progressive growth
type than in patients with the static type: 2.4 years (range,
1.2 to 6.2 y), and 6 years (range, 0.5 to 11.3 y), re-
spectively. Four patients (66.7%) had progressive type
macrodactyly out of 6 patients who received secondary
epiphysiodesis. However, there was no patient who un-
derwent amputation in either type. Our data showed that
FIGURE 5. Serial change of the length ratio of affected to proper epiphysiodesis can effectively control the longi-
unaffected in static and progressive growth patterns at 5 dif-
ferent time points. ‘a)’ Statistical different when compared tudinal growth of macrodactyly in both progressive and
with preoperative versus postoperative time point with a static types without the need for amputation.
P-value < 0.05. M indicates month; Preop, preoperatively; As anticipated, the width of the affected digit was not
Long-term, long-term follow-up more than 24 months. influenced by epiphyseal plate arrest. Rather, the width ratio
on 04/15/2023
increased in some points in the postoperative course (Fig. 4B).

Interestingly, this tendency was maintained throughout Topoleski et al11 also reported widening of the metaphysis
the postoperative course as the length ratio of distal after performing epiphysiodesis in toe macrodactyly. They
phalanx was larger than the ratio of other phalanges in all assumed it was due to iatrogenic periosteal stripping during
time points. open epiphysiodesis. Hernandez et al19 described similar
There are mixed opinions on the effectiveness of results from an animal study, and they attributed it to
epiphysiodesis in macrodactyly. Ishida et al10 performed decreased osteoclastic activity. The detailed mechanism
epiphysiodesis in four cases of macrodactyly of the hand. should be investigated in future study.
Results showed prevention of longitudinal overgrowth The limitations of this study are its inclusion of a
after 4 years of follow-up without additional surgery. small number of patients in a retrospective manner. Al-
Topoleski et al11 reported satisfactory results in 2 patients so, even though the decision to operate was based on
with macrodactyly of the foot after proximal phalangeal appearance, postoperative evaluation was done with
epiphysiodesis. In contrast, Tsuge et al7 doubted that X-rays. The study includes selection bias as we included
epiphyseal plate destruction alone could sufficiently con- patients who underwent epiphysiodesis as the first step of
trol longitudinal growth. They suggested that hypertrophy treatment, which can mean that disease was relatively
of bone is not caused by the epiphyseal plate but due to less severe. As shown in the 2 patients who were excluded
hypertrophy of all tissues. Although cases that underwent due to repeated epiphysiodesis on the same phalanx
epiphysiodesis for the same phalanx were excluded, there (Fig. 1), insufficient epiphysiodesis leads to recurrence.
were sections when the length ratio slightly increased due Despite the shortcomings, this is the first study to
to insufficient epiphysiodesis (Fig. 4A). However, perform quantitative analysis of the surgical results
longitudinal growth was overall well regulated with with long-term follow-up. Our observation clearly
epiphysiodesis and showed significant differences in showed that epiphysiodesis can effectively regulate
length ratios before and after the surgery. longitudinal growth of the macrodactyly in both types
The average age of patients who underwent epi- of growth patterns in macrodactyly.
physiodesis for proximal phalanx was 63.4 months, Epiphysiodesis is a promising, satisfactory, and
52.7 months for middle phalanx, and 76.6 months for minimally invasive approach for regulating the longi-
distal phalanx. Results showed that the prevention effect tudinal growth of macrodactyly when growth potential
of length growth seemed the fastest in the middle phalanx. remains. Objective functional analysis and contributing
As effectiveness of epiphysiodesis depends largely on the factors that can influence the growth of macrodactyly,
age at the time of surgery due to different remnant growth other than epiphysis, should be investigated in the future
potential, statistical analysis was done with stand- in a large prospective study.
ardization of age. Still after standardization, the timing of
the significant effect of epiphysiodesis for each phalanx REFERENCES
was measured differently. Other researchers have not 1. Flatt AE. The Care of Congenital Hand Anomalies. Quality Medical
presented such specific results. It was considered that the Publishing; 1994:292–316.
growth of phalanx at a younger age would be faster for 2. Leung P, Chan K, Cheng J. Congenital anomalies of the upper limb
cases operated at a younger age, when growth potential among the Chinese population in Hong Kong. J Hand Surg. 1982;7:
563–565.
remained high. 3. Southerland, Joe T, et al. McGlamry’s comprehensive textbook of foot
Our data showed that epiphysiodesis effectively and ankle surgery. Volume 2. Macrodactyly: Lippincott Williams &
controlled longitudinal growth in both progressive and Wilkins; 2012.

4. Hardwicke J, Khan M, Richards H, et al. Macrodactyly—options 13. Dennyson W, Bear J, Bhoola K. Macrodactyly in the foot. J Bone Jt
and outcomes. J Hand Surg (Eur Vol). 2013;38:297–303. Surg Br Vol. 1977;59:355–359.
5. Barsky AJ. Macrodactyly. JBJS. 1967;49:1255–1266. 14. Voutilainen A, Pitkäaho T, Kvist T, et al. How to ask about patient
6. De Laurenzi V. Macrodactyly of the middle finger. Giornale di satisfaction? The visual analogue scale is less vulnerable to
medicina militare. 1962;112:401–405. confounding factors and ceiling effect than a symmetric Likert scale.
7. Tsuge K. Treatment of macrodactyly. J Hand Surg Am. 1985;10 J Adv Nurs. 2016;72:946–957.
(6 Pt 2):968–969. 15. Rios JJ, Paria N, Burns DK, et al. Somatic gain-of-function
8. Wolfe SW, Pederson WC, Kozin SH, et al. Green’s operative hand mutations in PIK3CA in patients with macrodactyly. Hum Mol
surgery. 8th Edn, Volume 2. Deformities of the Hand and Fingers, Genet. 2013;22:444–451.
1438–1444 Elsevier Health Sciences; 2021. 16. Ezaki M, Beckwith T, Oishi SN. Macrodactyly: decision-making and
9. Cerrato F, Eberlin KR, Waters P, et al. Presentation and treatment surgery timing. J Hand Surg Eur Vol). 2019;44:32–42.
of macrodactyly in children. J Hand Surg Am. 2013;38:2112–2123. 17. Badawy M, Ma Y, Baldrighi C, et al. Efficacy of mTOR inhibitors
10. Ishida O, Ikuta Y. Long-term results of surgical treatment for (sirolimus) in isolated limb overgrowth: a systematic review. J Hand
macrodactyly of the hand. Plast Reconstr Surg. 1998;102:1586–1590. Surg (Eur Vol). 2022;47:698–704.
11. Topoleski TA, Ganel A, Grogan DP. Effect of proximal phalangeal 18. Gluck JS, Ezaki M. Surgical treatment of macrodactyly. J Hand
epiphysiodesis in the treatment of macrodactyly. Foot Ankle Int. Surg Am. 2015;40:1461–1468.
1997;18:500–503. 19. Hernandez JA, Serrano S, Marinoso ML, et al. Bone growth and
12. Jones KG. Megalodactylism: case report of a child treated by modeling changes induced by periosteal stripping in the rat. Clin
epiphyseal resection. JBJS. 1963;45:1704–1708. Orthop Relat Res. 1995;320:211–219.
on 04/15/2023

ORIGINAL ARTICLE
Use of Pediatric Outcomes Data Collection Instrument

to Evaluate Functional Outcomes in Multiple Hereditary
Exostoses
Nathalia Sundin Palmeira de Oliveira, MD,* Marcela Rocha Dias da Silva, MD,†
and Camila Bedeschi Rego de Mattos, MD‡
Results: Children with MHE have significantly lower scores

Background: The Pediatric Outcomes Data Collection Instrument (P < 0,05) in comparison with unaffected children in all domains
(PODCI) is a validated quality-of-life questionnaire with 6 domains using the Student and Welch t test. Parents score differs from
designed to provide a standardized method of measuring outcomes children score with statistically relevance in pain and comfort
in pediatric musculoskeletal conditions. To our knowledge there domain (P < 0,5). The Spearman test showed a negative corre-
on 04/15/2023
are no reports on its use in children with multiple hereditary lation between physical examination and PODCI score with
exostosis (MHE). statistical significance.
Questions/Purposes: Most published studies on MHE patients Conclusions: These results point towards PODCI’s capacity in
have described the efficacy of specific surgical techniques or the evaluating functional outcomes of pediatric patients with MHE.
specification of deformities. Little is known about the general Level of Evidence: Diagnostic Study, Level III.
health status of pediatric patients, the severity of pain, loss of
function, and how MHE influences the activities of daily life. We Key Words: multiple hereditary exostosis, osteochondromatosis,
aim to assess the functional levels of MHE pediatric patients with Health-Related quality of life, patient health questionnaire,
PODCI questionnaire. patient reported outcome measures, PODCI
Patients and Methods: As a cross-sectional study, we pro- (J Pediatr Orthop 2023;43:332–336)
spectively administered PODCI to 34 pediatric patients diag-
nosed with MHE and their families. The score distributions were
compared with values published earlier for children and adoles-
cents without musculoskeletal disorders using the Student and
Welch t tests. Parents and adolescents’ reports were compared
using Wilcoxon signed rank test. Physical examination and
T he main goal of orthopaedic interventions is to im-
prove health and physical function of patients, and
thus their quality of life. It is difficult to measure and
PODCI score relation were evaluated by Spearman test. quantify functional health in children as functional per-
formance increases with normal neurological maturation.1
The patient is the most important source of information
From the *Oncology Orthopedist, Orthopaedic department, Hospital regarding its health status.2 With young or intellectually
Universitário Pedro Ernesto (HUPE-UERJ), Rio de Janeiro, RJ; impaired children, the parent is the proxy, and in many
†Oncology Orthopedist, Instituto de Ortopedia e Traumatologia instances, they do not accurately reflect the child or
(IOT-USP), São Paulo, SP—Brazil; and ‡Oncology and Pediatric adolescent perceptions.1
Orthopedist, Orthopaedic department, Skåne’s University Hospital, The Pediatric Outcomes Data Collection Instrument
Lund, Sweden.
C.B.R.d.M.: Interpretation of data; critical revising for intellectual con- (PODCI) was developed in 1994 by the American
tent; approval and agreement of the version to be published. N.S.P.d. Academy of Orthopaedic Surgeons and the Pediatric Or-
O.: Conception and design of the work; drafting the work, inter- thopaedic Society of North America (POSNA).1 It is a
pretation of data, and critical revising for intellectual content; self-report scale designed to provide a standardized
approval and agreement of the version to be published. M.R.D.d.S.:
Drafting the work, interpretation of data, and critical revising for
method of evaluation of musculoskeletal conditions in
intellectual content; approval and agreement of the version to be pediatric population. It consists of 6 domains: (1) upper
published. extremity function, (2) transfers and mobility, (3) sports
The authors state that the views expressed in the submitted article are participation, (4) pain and comfort, (5) global function,
their own and not an official position of the institutions where the
study took place.
and (6) happiness with physical condition. Each domain is
No financial support was given to the conduction of this study nor the given a numerical score from 0 to 100, which 100 repre-
authors. sents the highest level of function.1 The advantages of
The authors declare no conflicts of interest. using PODCI includes an easy administration to patients,
Reprints: Nathalia Sundin Palmeira de Oliveira, MD, Oncology Orthopedist, parents and caretakers, and, relevance to their interest.
Orthopaedic department, Hospital Universitário Pedro Ernesto (HUPE-
UERJ), Rio de Janeiro, RJ, Brazil. E-mail: nathsundin@gmail.com. The PODCI has been validated and shown to be reliable
Copyright © 2023 Wolters Kluwer Health, Inc. All rights reserved. and sensitive to changes.1 Because of its properties, the
DOI: 10.1097/BPO.0000000000002372 PODCI was translated and validated in multiple

J Pediatr Orthop Volume 43, Number 5, May/June 2023 Use of PODCI
languages.3–7 It has earlier been evaluated in children with or first evaluation between October 2017 and August 2018
many conditions, such as cerebral palsy,8,9 scoliosis and were candidates for inclusion. Patients under completion
kyphosis,10,11 arthrogryposis,12 obstetric brachial palsy,3,13–15 of 2 years old or aged over 18 years old were excluded.
type 1 neurofibromatosis,16 mucopolysaccharidosis,17 idio- The diagnosis of solitary osteochondroma was an ex-
pathic juvenile arthritis, and obesity.6,7,18 clusion criterion regardless of musculoskeletal secondary
The Brazilian version was translated and validated disorders. The non-agreement or consent by the subject or
in 2013 by Do Monte FA et al7 with 57 patients. To our its legal guardians was also an exclusion criterion at any
knowledge, PODCI was never used to evaluate children time of the study.
with multiple hereditary exostoses (MHE). After PODCI’s terminology, were considered chil-
MHE is a rare disorder with an autosomal dominant dren patients between 2 and 10 years old, meanwhile pa-
inheritance though to arise in around 1 in 50.000 in- tients between 11 and 18 years old were considered
dividuals. Tumor suppressor genes EXT1 and EXT2 are adolescents. Children and adolescents will be referred
involved with the etiology of the characteristic non- collectively as patients in this paper.
malignant bone tumors covered by cartilage that appear in The PODCI questionnaire was administered to pa-
the metaphyseal region, close to the growth zone, of long rents of patients between 2 and 18 years old and to ado-
bones, ribs, hips, and vertebraes.19,20 The symptoms are lescents (11 to 18 y old). All patients were evaluated by
frequently related to the size and location of the exostoses, physical examination which assessed lower extremities
varying from totally asymptomatic to severe pain and coronal alignment, leg length discrepancy, pain complain,
deformities, potentially interfering in daily activities and and upper extremities mobility based on Leung and
psychosocial well-being.19–22 Peterson Classification System.23
on 04/15/2023
The main objective of this study was to evaluate the

METHODS results of PODCI for children and adolescents with MHE.
Cross-sectional prospective research was performed Secondary objectives were to correlate physical examina-
in 2 Institutions by the rules and authorization of the tion with PODCI results, and to assess differences between
Brazilian Ethical Research Committee (Comitê de Ética parents’ report and adolescents’ self-report outcomes.
em Pesquisa – CEP). Recommendations and approval of The Student t test and Welch t test were used to
both institutions’ Ethical Research Committees were compare mean scores between children and adolescents
respected before the beginning of the study. without musculoskeletal disorders and patients with
Children and adolescents with MHE diagnosed by MHE.1 Wilcoxon signed rank test was used to compare
the oncological service in our Institutions with follow-up parent-reports and adolescents self-reports. The assessment
TABLE 1. Descriptive Statistical Analysis of MHE Patients Data

N Min Max Mean Std* CI 95 Skewness z-score t test P-value Welch test P-value
Global functioning scale
A† 23 41 100 83,83 13,90 (77,8-89,8) −2,64 < 0,05 < 0,05
PA‡ 23 49 100 78,96 15,70 (72,2-85,7) −1,00 < 0,05 < 0,05
PC§ 11 50 98 80,55 15,88 (69,9-91,1) −1,33 < 0,05 < 0,05
Happiness scale
A† 23 30 100 84,35 15,54 (77,6-91,1) −4,60 0,49 0,45
PA‡ 23 5 100 72,61 29,19 (60,0-85,2) −2,24 < 0,05 0,16
PC§ 11 40 100 77,73 20,54 (63,9-91,5) −1,51 < 0,05 0,08
Pain/comfort scale
A† 23 8 100 68,65 27,35 (56,8-80,5) −1,11 < 0,05 < 0,05
PA‡ 23 0 100 58,09 29,59 (45,3-70,9) −0,14 < 0,05 < 0,05
PC§ 11 35 100 70,27 21,91 (55,6-85,0) −0,26 < 0,05 < 0,05
Sports and physical functioning scale
A† 23 21 100 80,65 20,41 (71,8-89,5) −2,98 < 0,05 < 0,05
PA‡ 23 17 100 74,78 22,86 (64,9-84,7) −2,12 < 0,05 < 0,05
PC§ 11 40 95 78,73 19,85 (65,4-92,1) −2,16 < 0,05 0,08
Transfer and basic mobility scale
A† 23 60 100 92,65 12,06 (87,4-97,9) −3,92 < 0,05 < 0,05
PA‡ 23 73 100 92,09 7,88 (88,7-95,5) −2,02 < 0,05 < 0,05
PC§ 11 74 100 91,36 10,77 (84,1-98,6) −1,03 < 0,05 0,06
Upper extremity scale
A† 23 67 100 93,22 9,35 (89,2-97,3) −3,37 < 0,05 < 0,05
PA‡ 23 38 100 91,26 14,70 (84,9-97,6) −5,43 < 0,05 < 0,05
PC§ 11 48 100 81,64 20,61 (67,8-95,5) −1,46 < 0,05 0,13
*Standard deviation.
†Adolescents self-report.
‡Parents of Adolescents report.
§Parents of Children report.
Max indicates maximum; Min, minimum; N, number.

Bedeschi Rego de Mattos et al J Pediatr Orthop Volume 43, Number 5, May/June 2023
of physical examination was related to PODCI scores using Pain, range of motion of upper extremities, and co-
the Spearman test. The skewness z-score was used to eval- ronal alignment and dysmetria of lower extremities were
uate the discriminative property between individuals assessed by physical examination and compared with
(“ceiling/floor effect”). PODCI score using Spearman test. On the basis of Leung
Statistical significance was defined as P-value <0.05 and Peterson Classification System,23 alterations on
in a 2-tailed test. Data analysis was performed using SciPy physical examination in the upper extremities were in-
1.1.0 open-source software. dicated on a binary score that ranged from 0 to 15 (from
less impairment to great impairment). It was shown a
distribution range of physical examination from 0 to 10.
The comparison between physical examination and parent
RESULTS based reports (N = 33) showed a negative correlation
A total of 34 patients met the criteria to be included (−0,43) with statistical significance and confidence
in this study. There were 21 male and 13 female, of those, (P < 0,05 ; CI 95%) which reflects that disabilities ob-
11 children and 23 adolescents. All questionnaires were jectively measured by physicians have a negative impact
sufficiently complete to calculate a score for all domains. on HRQL perception with lower PODCI scores as a re-
Thirty-three patients had full assessment of physical ex- sult. Physical examination data were also compared with
amination. adolescents self-report (N = 23) which maintained the
The descriptive statistics of PODCI responses for same tendency (correlation −0,42 ; P < 0,05 , CI 95%).
children and adolescents with MHE is shown in Table 1.
on 04/15/2023
The mean scores for patients with MHE were significantly

lower than the mean aggregated scores for children and
adolescents without musculoskeletal disorders (Table 2) in DISCUSSION
all domains using Student and Welch t tests (Fig. 1). MHE is a genetic driven musculoskeletal disorder
Wilcoxon test shows no significant difference among with a wide variety of clinical presentation, from totally
all pairs of comparison between pediatric self-reports and asymptomatic to great disability.19,20,22,24 Decision mak-
parent proxy-reports in all domains, except Pain, and ing in orthopaedic practice can be challenging as patient,
Comfort Scale. (Table 3). In all domains, the parent mean parents, and physician perception about functional con-
score (PA) was lower than the adolescent mean score (A). dition and complains may differ and impact on children
The skewness z-score for children with MHE was below and adolescents’ treatment.1,2,25
zero (Table 1). Patient-based outcomes are used to evaluate a pa-
tient’s subjective health experience and the consequences
of the patient’s disease and medical interventions. The
TABLE 2. Posna Normative Data self-assessment with self-administered questionnaire is
N Mean Std* considered the main standard method of quality-of-life
Global functioning scale evaluation, removing influence or interpretation from a
A† 1834 95.88 5.38 clinical or investigator.2,25 This approach to outcome as-
PA‡ 1834 95.15 7.24 sessment has particularities as functional performance and
PC§ 1791 93.31 7.77
Happiness scale
neurological maturation changes over time in a child/
A† 1834 81.83 17.59 adolescent under either typical or atypical development.25
PA‡ 1834 81.47 18.01 The PODCI was developed to assess musculoskel-
PC§ 1791 89.80 14.10 etal functional outcomes in children and adolescents and
Pain/comfort scale has not been replicated in all areas of pediatric
A† 1834 89.31 14.79
PA‡ 1834 88.96 16.67 orthopaedics.1,25 The purpose of this study was to evaluate
PC§ 1791 92.43 13.75 its use in MHE patients.
Sports and physical functioning scale When PODCI was evaluated in children and ado-
A† 1834 95.51 9.74 lescents without musculoskeletal disorders, a “ceiling ef-
PA‡ 1835 93.66 10.99
PC§ 1791 90.22 12.32
fect” was noted, where the distribution was heavily skewed
Transfer and basic mobility scale with scores clustered near the maximum score (no mean
A† 1834 99.05 4.70 score under 85), resulting in poor ability to differentiate
PA‡ 1834 99.22 4.56 between individuals.26 As MHE can be presented as to-
PC§ 1791 98.35 5.68 tally asymptomatic condition we were concerned about
Upper extremity scale
A† 1834 98.71 4.73 the reproduction of this “ceiling effect”. Although PODCI
PA‡ 1834 98.82 5.08 scores in MHE patients were left skewed, it was shown a
PC§ 1791 91.97 11.49 tendency of distribution proximate to normative values yet
*Standard deviation.
with an absence of “ceiling effect” (means range: 70-93). It
†Adolescents self-report. reflects a good discriminative property on our sample al-
‡Parents of Adolescents report. though we believe selection bias may interfere on this re-
§Parents of Children report.
N indicates number. sult, as pauci or asymptomatic patients often do not have
diagnosis or clinical follow-up on Institutions.19

J Pediatr Orthop Volume 43, Number 5, May/June 2023 Use of PODCI
FIGURE 1. Comparison between normative and MHE scores. MHE indicates multiple hereditary exostosis.
Physical examination evaluation showed a statistical conducted a HRQL study about MHE patients but,
on 04/15/2023
relevant correlation with PODCI score assessing its ca- as highlighted by the authors, no validated pediatric
pacity of revealing quality-of-life impairment related to quality-of-life assessment tool was used.
physical disabilities. The negative correlation between Caino and colleagues were the first authors to con-
physical examination and reports (parents and adolescents duct a MHE evaluation study with a specific HRQL pe-
self-reports) points toward a negative correlation between diatric instrument. The population assessment with the
alteration on physical examinations and quality-of-life, Pediatric Quality of Life Inventory (PedsQL) was made by
where higher scores in the Leung and Peterson classi- self-administration by children aged 8 to 18 years and by
fication, and thus greater impairment in the upper ex- the parents of children aged 2 to 18 years. The PedsQL is a
tremities, in addition to lower extremity deformities are 23-item instrument with 4 domains: (1) physical func-
correlated to lower PODCI scores, as expected. tioning (8 items); (2) emotional functioning (5 items); (3)
The perception of pain and comfort by the parents social functioning (5 items); and (4) school functioning (5
was worse than the patient’s own perception in our data. items). Although no statistically significant difference was
Daltry et al1 reported that the correlation among parents observed between children and caregivers responses, chil-
and adolescents in that scale was also low at follow-up. dren reported school domain as the most affected, whereas
We agree with those authors’ speculation were the health parents referred to the emotional aspect. Caino et al22
and well-being of the parents might influence their per- concluded that pain and disease severity had a negative
ceptions on how well the child is doing, as well as the impact on HRQL.
expectations of 2 different individuals can similarly vary PODCI is a validated HRQL instrument specific to
greatly.1 musculoskeletal conditions in pediatric population. Con-
Our literature review reveled only a few studies sidering the importance of self-administered ques-
dedicated specifically to evaluate quality of life of pe- tionnaires and its properties, adolescent’s responses were
diatric MHE patients. Chhina H and colleagues assessed evaluated in separate and compared with its parents’
health-related quality of life (HRQL) of MHE patients counterparts on this study. In all domains, the parent
using The Child Health Questionnaire (CHQ PF-50). mean score was lower than the adolescent mean score.
The CHQ PF-50 is a generic pediatric HRQL ques- Although Wilcoxon test showed statistically significant
tionnaire that was applied to patients (n = 43; 5 to 18 y difference only in Pain and Comfort domain (P < 0,05), a
old) in the mentioned study by parent proxy. Chhina H tendency of over estimating limitation by parents can be
et al21 concluded that MHE population had lower seen. A bigger sample would be more reliable evaluating
HRQL than the general population. Goud et al24 also this behavior in all domains.
TABLE 3. diference Significance between Self (A) and Parent CONCLUSION

Report (PA) This study was designed to evaluate the results of
Wilcoxon P-value PODCI for children and adolescents with MHE as a
Global functioning scale 0,067
sufficiently sensitive HRQL instrument. Hence, analysis of
Happiness scale 0,081 parent-reported scores and adolescents self-report ach-
Pain/comfort scale 0,004 ieved its purpose and it is in concordance with other
Sports and physical functioning scale 0,191 studies found in literature.21,22 PODCI adds to orthopae-
Transfer and basic mobility scale 0,534 dic practice the patient perspective, in a pediatric scenario,
Upper extremity scale 0,511
about its own quality-of-life related specifically to

Bedeschi Rego de Mattos et al J Pediatr Orthop Volume 43, Number 5, May/June 2023
musculoskeletal conditions. We believe PODCI ques- 10. Lerman JA, Sullivan E, Haynes RJ. The Pediatric Outcomes Data
tionnaire can be used as an assistant managing tool for Collection Instrument (PODCI) and functional assessment in
patients with adolescent or juvenile idiopathic scoliosis and
physicians concerning MHE follow-up and treatment de- congenital scoliosis or kyphosis. Spine (Phila Pa 1976). 2002;27:
cision making. Prospective studies with bigger samples 2052–2057.
focused on assessment of its psychometric properties spe- 11. Ugwonali OF, Lomas G, Choe JC, et al. Effect of bracing on the
cifically for MHE pediatric population may be of quality of life of adolescents with idiopathic scoliosis. Spine J. 2004;4:
greater value. 254–260.

12. Amor CJ, Spaeth MC, Chafey DH, et al. Use of the pediatric
outcomes data collection instrument to evaluate functional outcomes
ACKNOWLEDGMENTS
in arthrogryposis. J Pediatr Orthop. 2011;31:293–296.
The authors thank Lucas Pierezan Magalhães, MSs 13. Dedini RD, Bagley AM, Molitor F, et al. Comparison of pediatric
Computer Science and Bsc Mathematics, for its great as- outcomes data collection instrument scores and range of motion
sistance with statistical methodology and application. This before and after shoulder tendon transfers for children with brachial
plexus birth palsy. J Pediatr Orthop. 2008;28:259–264.
paper would not be possible without his efforts and patience. 14. Huffman GR, Bagley AM, James MA, et al. Assessment of children
The authors also acknowledge their former institutions, In- with brachial plexus birth palsy using the pediatric outcomes data
stituto Nacional de Traumatologia e Ortopedia (INTO), collection instrument. J Pediatr Orthop. 2005;25:400–404.
Rio de Janeiro, RJ—Brazil and Hospital Estadual da 15. Lerman J, Sullivan E, Barnes D, et al. The Pediatric Outcomes Data
Criança (HEC), Rio de Janeiro, RJ—Brazil that made this Collection Instrument (PODCI) and functional assessment of
patients with unilateral upper extremity deficiencies. J Pediatr
research project possible. Orthop. 2005;25:405–407.
16. Johnson BA, Sheng X, Perry AS, et al. Activity and participation in
REFERENCES
on 04/15/2023
children with neurofibromatosis type 1. Res Dev Disabil [Internet].

1. Dantroy L, Liang M, Fossel A, et al. The POSNA Pediatric 2015;36:213–221.
Musculoskeletal Functional Health Questionnaire: Report on Reliability, 17. O’Brien A, Bompadre V, Hale S, et al. Musculoskeletal function
Validity, and Sensitivity to Change. J Pediatr Orthop. 1998;18:561–571. in patients with mucopolysaccharidosis using the pediatric out-
2. Meadows KA. Patient-reported outcome measures: an overview. comes data collection instrument. J Pediatr Orthop. 2014;34:
Br J Community Nurs. 2011;16:146–151. 650–654.
3. Holst M, Vlieland T, Sande M, et al. Translation and adaptation of 18. Podeszwa D, Stanko K, Mooney J, et al. An analysis of the
the Pediatric Outcome Data Collecting Instrument (PODCI) into the functional health of obese children and adolescents utilizing the
Dutch language and preliminary validation in children with Neo- PODC instrument. J Pediatr Orthop. 2006;26:140–143.
natal Brachial Plexus Palsy. J Pediatr Rehabil Med. 2015;8:219–226. 19. WHO Classification of Tumours Editorial Board. Soft Tissue and
4. Kwon DG, Chung CY, Lee KM, et al. Transcultural adaptation and Bone Tumours. International Agency for Research on Cancer
validation of the Korean version of the pediatric outcomes data (IARC); 2020:437–439.
collection instrument (PODCI) in children and adolescents. J Pediatr 20. Ryckx A, Somers J, Allaert L. Hereditary multiple exostosis. Acta
Orthop. 2011;31:102–106. Orthop Belg. 2013;79:597–607.
5. Merder-Coskun D, Kenis-Coskun O, Celenlioğlu AE, et al. Reli- 21. Chhina H, Davis JC, Alvarez CM. Health-related quality of life in
ability of cross-cultural adapted Turkish version of the Pediatric people with hereditary multiple exostoses. J Pediatr Orthop. 2012;32:
Outcomes Data Collection Instrument (PODCI). J Pediatr Rehabil 210–214.
Med. 2016;9:101–105. 22. Caino S, Alba R, Bevilacqua S, et al. Osteochondromatosis: clinical
6. Misterska E, Adamczak K, Kaminiarczyk-Pyzałka D, et al. The variability and factors related to quality of life in children and adults.
Pediatric Outcomes Data Collection Instrument for a Polish sample Arch Argent Pediatr. 2022;120:180–186.
with juvenile idiopathic arthritis: psychometric properties of proxy 23. Leung AG, Peterson HA. Fractures of the proximal radial head and
version. Int J Rheum Dis. 2017;20:2077–2085. neck in children with emphasis on those that involve the articular
7. Do Monte FA, Ferreira MNL, Petribu KCL, et al. Validation of the cartilage. J Pediatr Orthop. 2000;20:7.
Brazilian version of the pediatric outcomes data collection instru- 24. Goud AL, De Lange J, Scholtes VAB, et al. Pain, physical and social
ment: a cross-sectional evaluation in children and adolescents with functioning, and quality of life in individuals with multiple hereditary
juvenile idiopathic arthritis. BMC Pediatr. 2013;13:1–6. exostoses in the Netherlands: a national cohort study. J Bone Jt Surg
8. McMulkin ML, Baird GO, Gordon AB, et al. The pediatric - Ser A. 2012;94:1013–1020.
outcomes data collection instrument detects improvements for 25. Phillips L, Carsen S, Vasireddi A, et al. Use of patient-reported
children with ambulatory cerebral palsy after orthopaedic interven- outcome measures in pediatric orthopaedic literature. J Pediatr
tion. J Pediatr Orthop. 2007;27:1–6. Orthop. 2018;38:393–397.
9. Seok Park M, Youb Chung C, Min Lee K, et al. Rasch analysis of 26. Haynes RJ, Sullivan E. The Pediatric Orthopaedic Society of North
the pediatric outcomes data collection instrument in 720 patients America pediatric orthopaedic functional health questionnaire: an
with cerebral palsy. J Pediatr Orthop. 2012;32:423–431. analysis of normals. J Pediatr Orthop. 2001;21:619–621.

Untitled

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Untitled

Uploaded by

Copyright:

Available Formats

ORIGINAL ARTICLE

A Prospective, Multicenter Study of Developmental

The majority (81%; n = 13/16) occurred in the ﬁrst year after

J Pediatr Orthop  Volume 43, Number 5, May/June 2023 www.pedorthopaedics.com | 279

Copyright r 2023 Wolters Kluwer Health, Inc. All rights reserved.

the rate of closed reductions declining more rapidly than

that are primarily retrospective in nature and usually

single-center based. Multiple studies have demonstrated

collected database of an international multicenter study

Copyright r 2023 Wolters Kluwer Health, Inc. All rights reserved.

associated with the primary outcomes of the study

center status and location, previous treatment (harness or

closed reduction), and open reduction approach. IHDI

Copyright r 2023 Wolters Kluwer Health, Inc. All rights reserved.

treated with a pelvic osteotomy at the time of open DISCUSSION

Copyright r 2023 Wolters Kluwer Health, Inc. All rights reserved.

Copyright r 2023 Wolters Kluwer Health, Inc. All rights reserved.

Copyright r 2023 Wolters Kluwer Health, Inc. All rights reserved.

Copyright r 2023 Wolters Kluwer Health, Inc. All rights reserved.

Age Influence Upon Glenohumeral Remodeling After

[CI = (−0.04; −0.01), P = 0.002], 0.12% [CI = (−0.21; −0.04),

J Pediatr Orthop  Volume 43, Number 5, May/June 2023 www.pedorthopaedics.com | e389

Copyright r 2023 Wolters Kluwer Health, Inc. All rights reserved.

TABLE 1. Cohort’s Characteristics*

Term (wk) 39 ± 2 (31;41)

Weight (g) 3844 ± 775 (1100;5400)

shoulder release to restore passive shoulder ER. Patients

Copyright r 2023 Wolters Kluwer Health, Inc. All rights reserved.

A tendon transfer to the infraspinatus muscle was

British Medical Council Research (BMRC) grading system

in patients older than 3; grade-2 strength or less according

Surgical Techniques and Postoperative Care

Tendon transfers were performed in an open manner,

Copyright r 2023 Wolters Kluwer Health, Inc. All rights reserved.

TABLE 3. Preoperative and postoperative MRI outcomes*

PHHA (%) 31 ± 15 (0;55) 39 ± 13 (0;63) 0.0016

Glenohumeral deformity 3.3 ± 1.6 (1;6) 2.1 ± 1.5 (1;7) 0.0002

Regarding linear regression analyses, improvements

Copyright r 2023 Wolters Kluwer Health, Inc. All rights reserved.

This 5-years age limit seems consistent with the ex-

Glenoid version 20 ± 19 (−3;62) 6 ± 15 (−24;42) 0.0089

TABLE 5. Age-based Subgroup Analysis of MRI Improvements*

Copyright r 2023 Wolters Kluwer Health, Inc. All rights reserved.

surgeries on intact joints. patients with BPBI-related active shoulder ER deﬁcit; in

Copyright r 2023 Wolters Kluwer Health, Inc. All rights reserved.

Copyright r 2023 Wolters Kluwer Health, Inc. All rights reserved.

Clinical Characteristics and Treatment Patterns of Open

Raphael H. Parrado, MD,* Lukas Foster, MD,* Megan Gilbert, CPNP-AC,*

Key Words: antibiotic treatment, children, open hand fracture,

extensively studied. This study aimed to deﬁne pediatric open

e358 | www.pedorthopaedics.com J Pediatr Orthop  Volume 43, Number 5, May/June 2023

Copyright r 2023 Wolters Kluwer Health, Inc. All rights reserved.

METHODS ulnar artery being most involved in 7 cases (16%). Partial

Copyright r 2023 Wolters Kluwer Health, Inc. All rights reserved.

Baseball 44 (14) Age; median (IQR) 7 (1-10) 7 (3-11) 0.34

Bike riding 40 (13) Gender (M) 23 (76) 1550 (62) 0.09

5%, P = 0.6) in all children (Table 4).

Copyright r 2023 Wolters Kluwer Health, Inc. All rights reserved.

results have speciﬁc implications for injury prevention in

Displaced fracture (yes) 1511 (78) 117 (19) < 0.05

choice, providing gram-positive (including Methicillin-

Copyright r 2023 Wolters Kluwer Health, Inc. All rights reserved.

CONCLUSION 7. Patzakis MJ, Wilkins J. Factors influencing infection rate in open

Copyright r 2023 Wolters Kluwer Health, Inc. All rights reserved.

J Pediatr Orthop Volume 43, Number 5, May/June 2023 www.pedorthopaedics.com | 279

J Pediatr Orthop Volume 43, Number 5, May/June 2023 www.pedorthopaedics.com | e389

e358 | www.pedorthopaedics.com J Pediatr Orthop Volume 43, Number 5, May/June 2023

J Pediatr Orthop Volume 43, Number 5, May/June 2023 www.pedorthopaedics.com | 317

J Pediatr Orthop Volume 43, Number 5, May/June 2023 www.pedorthopaedics.com | e319

286 | www.pedorthopaedics.com J Pediatr Orthop Volume 43, Number 5, May/June 2023

e326 | www.pedorthopaedics.com J Pediatr Orthop Volume 43, Number 5, May/June 2023