Professional Documents
Culture Documents
Definingnoptimal Calcar Screw Positioning in Proximal Humerus Fracture Fixation
Definingnoptimal Calcar Screw Positioning in Proximal Humerus Fracture Fixation
www.elsevier.com/locate/ymse
a
Department of Orthopaedic Surgery, Thomas Jefferson University Hospitals, Philadelphia, PA, USA
b
Sidney Kimmel Medical College, Thomas Jefferson University Hospitals, Philadelphia, PA, USA
c
Drexel University College of Medicine, Philadelphia, PA, USA
d
The Rothman Institute of Orthopaedics, Thomas Jefferson University Hospitals, Philadelphia, PA, USA
Background: Anatomic reduction and placement of an inferior calcar screw are strategies to prevent fix-
ation failure in proximal humerus factures. Optimal position of the calcar screw remains unknown.
Methods: There were 168 shoulders (68.5% female; average age, 63.6 ± 11.5 years) that underwent open
reduction and internal fixation of a displaced proximal humerus fracture involving the surgical or ana-
tomic neck. Univariate and multivariate analyses were performed on preoperative clinical, preoperative
radiographic, and postoperative radiographic variables to determine association with fixation failure. A
receiver operating characteristic curve was performed to determine a maximum distance from the inferi-
or screw to the calcar (“calcar distance”) as well as a maximum ratio of this distance and the head diameter
(“calcar ratio”).
Results: There were 26 of 168 (15.5%) patients with radiographic failures (19 related to fixation failure).
Univariate analysis and multivariate analyses found quality of reduction (P < .001), calcar distance (P < .001),
and calcar ratio (P < .001) to be significantly associated with radiographic success. In all patients, receiv-
er operating characteristic analysis found quantifiable thresholds of 12 mm or within the bottom 25% of
the humeral head as measures to prevent fixation failure.
Conclusions: Quality of reduction, calcar distance, and calcar ratio independently correlated with fixa-
tion failure. This study provides optimal distances and ratios for calcar screw placement that can be used
clinically.
Level of evidence: Level III; Retrospective Cohort Design; Treatment Study
© 2017 Journal of Shoulder and Elbow Surgery Board of Trustees. All rights reserved.
Keywords: Proximal humerus fracture; proximal humerus open reduction; internal fixation; calcar screw;
nonunion; malunion; varus collapse
1058-2746/$ - see front matter © 2017 Journal of Shoulder and Elbow Surgery Board of Trustees. All rights reserved.
https://doi.org/10.1016/j.jse.2017.05.003
1932 E.M. Padegimas et al.
humerus fractures increases, the burden of complications, in- films. Direct chart review of all patients and radiographs was per-
cluding fixation failure, will become increasingly significant. formed to determine if they had in fact sustained a proximal humerus
A recent analysis of proximal humerus fracture ORIF found fracture that underwent ORIF, if the fixation was with a locking plate
an 18.4% rate of fixation failure and a reoperation rate of and screws construct, and if the patient did not meet any exclusion
criteria.
27.6%.13 Interestingly, the quality of reduction at the time of
surgery was predictive of both radiographic and clinical out-
comes. Earlier analyses found age, local cortical density, quality Independent variables
of reduction, and restoration of the medial cortical support
to be predictive of radiographic failure.8 Retrospective anal- Preoperative clinical variables were identified for all patients.
ysis of patients undergoing proximal humerus ORIF found These variables included age, gender, race, individual medical
that placement of a medial support (calcar) screw was asso- comorbidities, and age-adjusted Charlson Comorbidity Index.4,14
ciated with a decreased risk of loss of reduction.12,13 Preoperative radiographic variables were identified for all patients
The inter-relatedness of fracture reduction quality and the on true anterior-posterior and Y-view radiographs. These variables
position of the medial support (calcar) screw has not been were cortical density measurement,15 Neer classification,11 medial
hinge displacement,7 medial comminution,7 and presence of varus
quantified. The purpose of this study was to identify radio-
or valgus angulation (Fig. 1). A neck-shaft angle of <120° was con-
graphic variables that were independently associated with sidered varus (Fig. 1, A), whereas a neck-shaft angle >150° was
failure of proximal humerus fracture ORIF and to generate considered valgus (Fig. 1, B).13
thresholds for optimal calcar screw placement based on the Early postoperative radiographs (true anterior-posterior and
quality of the reduction. Y-view radiographs) taken at 2 weeks postoperatively were re-
viewed for a number of variables. All preoperative and postoperative
radiographs are standardized at our institution. Based on the pa-
Methods rameters for reduction quality described by Schnetzke et al, medial
head-shaft displacement of <5 mm, greater tuberosity cranialization
Study population of <5 mm, and varus or valgus angulation of 120°-150° were
considered.13 If all 3 criteria were met, the fracture was considered
All operatively treated proximal humerus fractures were identified adequately reduced. If 2 of the 3 criteria were met, the fracture was
by querying an institutional database by International Classifica- considered partially reduced. If only 1 or 0 of the 3 criteria was met,
tion of Diseases, Ninth Revision and Tenth Revision, Clinical the fracture was considered malreduced. In addition to the reduc-
Modification and Current Procedural Terminology codes. This da- tion quality, the integrity of the calcar, distance of inferior screw
tabase included patients from January 2008 through March 2016. to the calcar (referred to as calcar distance), ratio of calcar dis-
The codes used were 79.31 (open reduction of fracture with inter- tance and the head diameter (referred to as calcar ratio), distance
nal fixation, humerus), S42.2__ (all codes for fracture of upper end of the inferior screw to the humeral head articular surface (re-
of humerus), and 23615 (open reduction with internal fixation ferred to as tip distance), number of diaphyseal screws, number of
of proximal humerus). Exclusion criteria were revision surgery, proximal screws, and number of proximal screws in the inferior half
associated diaphyseal fracture, fixation with an intramedullary of the head were also considered.1,17 The calcar distance was mea-
device, no anatomic or surgical neck component of the fracture sured as the perpendicular from the threads of the calcar screw to
(eg, isolated greater tuberosity fractures were excluded), inadequate the apex of the arch of the calcar. A well-reduced and well-fixed
preoperative or postoperative films, incomplete clinical variables, proximal humerus fracture that went on to successful radiographic
previous shoulder fracture or nonunion, and <12-week postoperative healing is shown in Figure 2 (all measurements shown).
Figure 1 (A) Representative anterior-posterior radiograph of a Neer 2-part11 proximal humerus fracture in varus (neck-shaft angle of <120°).
(B) Representative anterior-posterior radiograph of a Neer 4-part11 proximal humerus fracture in valgus (neck-shaft angle of >150°).
Proximal humerus fixation 1933
collapse with screw penetration (21.1%; 4/19). These failures humeral head vs. 2.9; P = .13), or the distance between the
were identified at a mean of 24.9 weeks (range, 12.0-46.6 tip of the screw and the articular surface (12.6 mm vs.
weeks). None of the preoperative patient or fracture charac- 10.9 mm; P = .21). With regard to the medial support screw,
teristics considered were statistically significant predictors of both the shorter distance from the calcar to the screw (calcar
failure on univariate analysis (Table I). The quality of distance; P < .01) and the ratio of this distance to the humeral
reduction13 was predictive of failure (1.8 average of achieved head diameter (calcar ratio; P < .01) were strongly predic-
parameters in those that failed vs. 2.5 of 3 in those that did tive of failure. Isolating adequately reduced fractures (3/3
not; P < .01). Individually, reduction of angulation (achieved reduction parameters achieved), the average distance from the
in only 52.6% of failures compared with 83.1% in success- screw to the calcar was 19.2 mm in failures and 9.5 mm in
ful outcomes; P = .01) and correction of head-shaft successes (P < .01; Table II), with a calcar ratio of 38.4% vs.
displacement (achieved in 42.1% of failures and 71.1% of 18.6% (P < .01), respectively. Isolating partially reduced frac-
success cases; P = .02) were statistically significant predic- tures (2/3 reduction parameters achieved), the average calcar
tors of failure. distance was 18.8 mm in failures and 10.3 mm in successes
From the perspective of fixation, the fixation failures were (P = .06; Table II), with a calcar ratio of 34.9% vs. 19.5%
similar to the healed group with respect to the number of distal (P = .02), respectively. Isolating malreduced fractures (0 or
screws (mean, 2.9 distal screws vs. 3.0; P = .72), the number 1/3 reduction parameters achieved), the average calcar dis-
of proximal fragment screws (mean, 6.2 proximal screws vs. tance was 16.2 mm in failures and 13.3 mm in successes
5.8; P = .28), the number of screws in the inferior half of the (P = .10; Table II), with a calcar ratio of 33.5% vs. 25.8%
humeral head (mean, 2.4 screws in the inferior half of the (P = .05), respectively.
Proximal humerus fixation 1935
Table III Three models of multivariate analysis of predictors of fixation failure after proximal humerus surgical neck fixation
Univariate correlation Model 1 Model 2 Model 3
coefficient OR P value OR P value OR P value
Neer classification (compared to 2-part)
3-part −0.16 0.13 .09 0.14 .09 0.11 .08
4-part 0.009 0.65 .52 0.69 .56 0.68 .56
Reduction quality −0.27 N/A 0.42 .01 0.42 .01
Head-shaft displacement −0.20 0.38 .16 N/A N/A
Angulation −0.24 0.42 .20
Calcar ratio 0.43 1.14 <.01 N/A 1.14 <.01
Distance to inferior calcar 0.40 N/A 1.20 <.01 N/A
OR, odds ratio; N/A, not applicable.
In multivariate analysis, the quality of reduction was an as the reduction quality improved. For adequately reduced
independent predictor of success (Table III). On isolation of fractures, the optimal threshold was a calcar distance of
the distance to the calcar from the inferior screw or the ratio 13.1 mm or less and a calcar ratio of 26.1% or less. For par-
of this distance to the humeral head diameter, both param- tially reduced fractures, the optimal threshold was a calcar
eters were predictors of failure independent of reduction quality distance of 12.1 mm or less and a calcar ratio of 21.0% or
(odds ratio, 1.2 and 1.1, respectively). In addition, single-factor less. Finally, for malreduced fractures, the optimal thresh-
analysis of variance was used to assess any potential corre- old was a calcar distance of 11.4 mm or less and a calcar ratio
lation between Neer classification and the quality of reduction. of 23.4% or less.
There was no significant difference in the quality of reduc- Ten of the 26 (38.5%) shoulders that went on to radio-
tion based on Neer classification, with 2-part fractures having graphic failure underwent 12 future operations at a mean of
an average of 2.43 ± 0.72 reduction components achieved, 47.3 weeks (range, 2.1-166.3 weeks) after the index proce-
3-part fractures 2.50 ± 0.86, and 4-part fractures 2.38 ± 0.88 dure. These were 4 hardware removals, 1 hematoma washout,
(F = 0.25; P = .78).11,13 1 rotator cuff repair, 2 arthroscopic débridements (1 with
Analysis of ideal medial support screw placement with re- concurrent removal of hardware), 3 revision ORIF, and 1 an-
ceiver operating characteristic curve confirmed the predictive atomic total shoulder arthroplasty. In comparison, 16 of
value of the calcar distance and the calcar ratio (Table II). the 142 (11.3%) shoulders that healed radiographically
The AUC was 0.84 and 0.86 for calcar distance and calcar underwent 16 future operations at a mean of 48.6 weeks
ratio, respectively. The optimal threshold for all patients to (range, 14.0-153.9 weeks) after the index procedure (P < .01).
minimize radiographic failures was a calcar distance of 12 mm Thirteen reoperations were hardware removal (5 of which
or less and a calcar ratio of 25% or less (Table II). On strati- underwent an arthroscopic débridement), 2 were arthro-
fying patients by reduction quality, the identified threshold scopic débridements, and 1 was a reverse total shoulder
for both the calcar distance and calcar ratio became less strict arthroplasty for rotator cuff insufficiency.
1936 E.M. Padegimas et al.