ORIGINAL ARTICLE
Recurrence After Femoral Derotational Osteotomy
in Cerebral Palsy
Hayong Kim, MD, MS,* Michael Aiona, MD, PhD,† and Michael Sussman, MD, PhD†
Abstract: The authors studied the long-term results of femoral
derotational osteotomy (FDO) for medial femoral torsion in ambu-
latory children with cerebral palsy. Thirty children with 45 femurs
that underwent distal FDO were followed for a mean of 6.5 years.
Although correction was achieved after surgery, recurrence occurred
during follow-up in 15 femurs. Preoperative mean external hip rota-
tion of 10.7 6 7.2 degrees increased to 41.3 6 16.6 degrees 1 year
after surgery and decreased to 28.2 6 14.7 degrees 5 years after sur-
gery. On kinematic data, maximum hip rotation in stance of 30 degrees
before surgery decreased to 8.7 degrees 1 year after surgery and in-
creased to 16.1 degrees 5 years after surgery. Minimum hip rotation
of 10.4 degrees before surgery was corrected to –4.3 degrees 1 year
after surgery and was 0.8 degrees 5 years after surgery. Passive hip
external rotation and kinematic hip rotation showed progressive dete-
rioration of the initial correction. Patients having surgery prior to age
10 were more likely to show deterioration.
Key Words: medial femoral torsion, cerebral palsy, femoral
osteotomy
(J Pediatr Orthop 2005;25:739–743)
R otational malalignment of the lower extremities has a
deleterious functional as well as cosmetic influence on
gait.1,2 Medial femoral torsion (MFT) is the most common
bony deformity in children with diplegic cerebral palsy (CP).
Associated gait problems with MFT include excessive internal
hip rotation in all phases of gait, leading to internal foot
progression angle and poor foot clearance with excessive shoe
wear, which causes increased tripping and falling, as well as
a cosmetically poor gait pattern.
MFT does not respond to nonsurgical treatment.1,2
Torsional deformities are, however, treated successfully with
corrective derotational osteotomies of the femur, done at either
the proximal or distal femur.2–8 Distal femoral derotational
osteotomy (distal FDO) is less invasive, provides quicker re-
covery for rehabilitation, and is known to be as effective in
functional and cosmetic terms as proximal osteotomy.7 It has
From the *Eulji University, Daejeon City, Korea; and †Shriners Hospital for
Children, Portland, OR.
None of the authors received financial support for this study.
Study conducted at Shriners Hospital for Children, Portland, OR.
Reprints: Hayong Kim, MD, Eulji University School of Medicine,
Tunsan-dong, Seo-gu 1306, Daejeon City, 302-799, Korea (e-mail:
hayong_kim@yahoo.com or hykim@eulji.ac.kr).
Copyright Ó 2005 by Lippincott Williams & Wilkins
J Pediatr Orthop  Volume 25, Number 6, November/December 2005
an additional advantage of performing an osteotomy with the
opportunity to correct fixed flexion contracture at the knee, if
present. The short-term results of this procedure are generally
reported to be excellent for the correction of hip rotation and
foot progression angles in children with spastic diplegia.3,7
There are few studies in which the static and kinematic
outcome of FDO has been serially followed for the long term.
The purpose of this study was to see whether the distal FDO
may correct the deformity and whether the correction achieved
by FDO is well maintained on long-term follow-up of more
than 5 years in children with spastic diplegia.
METHODS
Patient Selection
The medical records and gait laboratory data for patients
who underwent distal FDO between 1987 and 1998 were re-
viewed retrospectively. Included in the study were ambulatory
children with spastic CP who underwent distal FDO with in-
ternal fixation with plate and screws, and who had been fol-
lowed for at least 5 years or revised during the follow-up with
the purpose of operation to correct MFT. Thirty children
(14 male, 16 female) with 45 femurs met our criteria for
inclusion in this study. The anatomic distribution of the patients
was diplegia in 15, asymmetric diplegia in 2, triplegia in 8,
quadriplegia in 3, and familial spastic diplegia in 2 (Table 1).
Average age at the time of surgery was 9 years 2 months (range
4 years 10 months to 17 years 2 months). Mean follow-up for
the entire group was 6.5 years. Only patients with clinical
examination and follow-up gait laboratory data were included.
Our indications for distal FDO were a clinical history of
functional problem, limited external rotation of the hip of less
than 15 degrees on physical examination, and hip internal rota-
tion and internal foot progression angle more than 10 degrees
of difference from the norm on kinematic data. For the limited
number of patients, we did CT scan torsional studies for an-
other guideline for surgery. The data were acquired before sur-
gery (P0), 1 year after surgery (P1), and usually every 2 years
thereafter for clinical follow-up, up to P9.
Patient Treatment
Two of the authors (M.S., M.A.) performed all of the
surgeries. The amount of derotation was decided with the con-
sideration of examination under anesthesia findings and preop-
erative static and kinematic data. The surgical goal was set to
make the range of hip motion equal to external and internal
rotation or provide a little more external rotation than internal
rotation by about 10 degrees. Distal FDO was carried out in the
739
Kim et al J Pediatr Orthop  Volume 25, Number 6, November/December 2005

Oxford Metrics, Oxford, UK). Helen Hayes reflective markers

TABLE 1. Characteristics of Patients Undergoing Distal
were placed on the pelvis and lower extremities in accordance
Femoral Osteotomy
with the model outlined by Oxford Metrics.10
Pt. No. Diagnosis* Age Sex Derotation Site† Last Gait Lab‡
A few studies performed during the early stage of our
1 A-diplegia 6+7 F 35 L 5 gait laboratory were processed with the Helen Hayes program
2 Diplegia 8+9 M 50 B 4 (Helen Hayes Hospital, West Haverstraw, NY). For these data,
3 Diplegia 8+5 F 40/40 B 5 we could not get the data on an Excel spreadsheet. Those data
4 F-diplegia 4 + 10 F 40/30 B 6 were recorded only by kinematic graphs, and the data values
5 A-diplegia 12 + 5 M 40/40 B 5 for kinematics were calculated manually by one of the authors
6 Diplegia 8+3 F 30/30 R 3 (H.K.) using conventional orthopaedic rulers. The accuracy
7 Triplegia 9+3 M 35/45 B 7 was tested on the graphs acquired with VCM software. The
8 Triplegia 8+6 M 30/35 B 7 range of error for this manual measurement was less than
9 Diplegia 6+3 F 40 R 7 5 degrees.
10 Quadriplegia 12 + 7 M 40 L 5
11 Triplegia 8+1 M NA B 9 Statistics
12 Diplegia 7+6 M NA B 5 The patients were classified according to the age at oper-
13 Diplegia 6+2 M 30/45 L 8 ation (,10 and .10 years of age). The differences produced
14 Triplegia 6 + 11 F 30 L 4 by corrective surgery (P0 vs. P1) were tested for statistical
15 Diplegia 6+1 M NA B 9 significance using the paired t test. The comparison between
16 Quadriplegia 12 + 5 F 35/30 B 6 P1 and P5 was used to assess postoperative loss of correction.
17 Triplegia 9 + 11 F 35 L 7 An analysis of variance (ANOVA) was carried out to deter-
18 Diplegia 10 + 2 M 60/30 B 6 mine the tendency of this recurrence (P1 vs. P5). The dif-
19 Diplegia 17 + 2 M NA R 5 ference of the rate of recurrence according to the age group
20 Triplegia 15 + 1 F NA L 5 was tested with the chi-square test (Fisher exact test).
21 Diplegia 8+2 F NA R 9
22 F-diplegia 6+4 F 40/40 B 7
23 Diplegia 8+8 M NA B 4
RESULTS
24 Diplegia 10 + 4 M 30/45 B 10 Other Associated Operations
25 Diplegia 5+1 F 40 L 5 A variety of soft tissue procedures and bony procedures
26 Diplegia 6 + 11 F 30 R 11 were performed at the time of distal FDO or subsequently. The
27 Diplegia 11 + 1 F 45 L 5 most commonly performed procedures were distal hamstring
28 Triplegia 14 + 4 F 50/50 B 5 release and heel cord lengthening, which were each performed
29 Triplegia 11 + 7 F 30 R 9 on 18 patients (Table 2).
30 Quadriplegia 7+7 M 35 L 12
*A-diplegia, asymmetric diplegia; F-diplegia, familial spastic diplegia. Postoperative Complications
†B, bilateral; L, left; R, right. There were no acute postoperative complications such as
‡Time (years) between surgery and the last gait lab study.
wound problems, infection, and loss of correction for the index
procedure. Only one patient had a complication of metal screw
failure with revision surgery for recurrence of internal femoral
supracondylar area using the lateral approach to the distal torsion; it was corrected with a change of plate and screws.
femur.9 Derotation was done using two guide pins inserted prox-
imally and distally to the osteotomy site to guide angular cor-
rection. Internal fixation was achieved with either small-fragment TABLE 2. Other Associated Operations
or narrow LCDCP AO plates. Location Procedure No. Pts.
Hip Adductor release 15
Methods Psoas release 4
We reviewed the medical records for the age at the time Knee Distal hamstring release 18
of operation, other associated operations, and postoperative com- Rectus transfer 11
plications. We reviewed the gait laboratory data for passive hip Ankle Heel cord lengthening 18
internal and external rotation on static examination and pelvic Post tib lengthening 4
rotation, hip rotation, and foot progression angle on kinematic Split T.P. transfer 2
data. Split T.A. transfer 4
Gait laboratory staff performed the physical examina- Tibia Tibia derotation osteotomy 2
tions using a goniometer for static measurements. Hip rotation Tibia internal
was measured with the patient prone and the knees flexed to rotation osteotomy 5
90 degrees. Others Upper extremity 7
Gait data were collected using 6- to 8-camera VICON SDR 2
370 systems with two AMTI force plates, and subsequently Distal FDO and hardware removal are excluded from this table.
processed using VICON Clinical Manager software (VCM,

740 q 2005 Lippincott Williams & Wilkins

J Pediatr Orthop  Volume 25, Number 6, November/December 2005 Recurrence After Femoral Derotational Osteotomy in CP

Recurrence
Fifteen of the 45 femurs were classified as having re-
currence of deformity requiring revision surgery. Thirteen
femurs underwent a second FDO to treat recurrent MFT. Two
other femurs have met the indication for repeat distal FDO,
which has been recommended but not yet performed.
Fourteen of the 15 recurrences occurred in children
younger than 10 years old at the time of the index surgery.
The timing for revision surgery was P4 for 3 children, P5 for 4,
P8 for 5, and P10 for 1. The timing of the recommendation for
surgery that has not yet been done in two patients was P7 and
P10.
Hip External Rotation on Static Examination
Mean external hip rotation on static examination at
P0 of 10.8 6 7.2 degrees (n = 45) increased to 41.3 6 16.6
degrees at P1 (n = 45) and decreased to 28.2 6 14.7 degrees at
P5 (n = 41) (Fig. 1, Table 3).
FIGURE 1. Follow-up of external rotation in prone position
Hip Internal Rotation on Static Examination (mean values).
Mean internal hip rotation at P0 of 77.7 6 9.8 degrees de-
creased to 45.6 6 12.3 degrees at P1 and 54.0 6 14.4 degrees Statistical Analysis
at P5.
Postoperatively, static examination data (internal and
Hip Rotation on Kinematic Data external rotation) and kinematic data (internal hip rotation,
During gait, maximum hip rotation in stance at P0 of external hip rotation, mean hip rotation) were significantly
27.0 degrees internal decreased to 8.7 degrees at P1 (n = 45) changed compared with preoperative data (paired t test, P ,
and increased to 16.1 degrees at P5 (n = 41) (Fig. 2). Minimum 0.001; Table 4). However, the data on pelvic rotation showed
hip rotation in stance at P0 of 10.0 degrees was corrected no significant changes (P = 0.306 for internal pelvic rotation,
to –4.3 degrees at P1 and 0.8 degrees at P5. P = 0.632 for external pelvic rotation).
Similar results were observed for the comparison of P1
Pelvic Rotation and Foot Progression Angle on and P5. Static examination data (external rotation) and kine-
Kinematic Data matic data (internal hip rotation) of P5 reverted significantly
On the kinematic data, maximum pelvic rotation in back to preoperative status compared with P1 (ANOVA, P ,
stance at P0 of 8.9 degrees was 7.6 degrees at P1 (n = 45) and 0.05). However, the data on pelvic rotation showed no sig-
6.8 degrees at P5 (n = 41) (Fig. 3). Minimum pelvic rotation in nificant changes (P = 0.987 for internal pelvic rotation,
stance at P0 of –12.0 degrees was 212.6 degrees at P1 (n = 45) P = 0.648 for external pelvic rotation). There was a statistically
and –10.0 degrees at P5 (n = 41). Foot progression angles significant tendency of recurrence for hip rotation in static and
changed from an internal progression mean of 19.7 degrees kinematic data, even though P5 still had a corrected value
at P0 (n = 41) to external progression of 3.8 degrees at P1
(n = 44) and external progression of 0.6 degrees at P5 (n = 41).

TABLE 3. Mean Values of Rotational Profiles in Static and

Dynamic Status
P0 P1 P3 P5 P7
ER 10.8 (7.2) 41.3 (16.6) 35.1 (14.3) 28.2 (14.7) 27.4 (16.3)
IR 77.7 (9.8) 45.6 (12.3) 49.9 (16.7) 54.0 (14.4) 49.5 (17.5)
MaxHR 27.0 (11.7) 8.7 (10.0) 11.4 (9.8) 16.1 (12.0) 17.1 (11.1)
MinHR 10.0 (12.6) 24.3 (12.1) 20.1 (12.0) 0.8 (12.6) 1.1 (13.8)
MaxPR 8.9 (10.1) 7.6 (12.2) 8.4 (12.0) 6.8 (10.3) 8.3 (11.1)
MinPR 212.0 (9.1) 212.6 (12.1) 210.0 (9.6) 210.0 (9.7) 211.8 (10.8)
FPA 19.7 (21.4) 23.8 (14.1) 22.0 (17.5) 20.6 (20.6) 21.7 (13.8)
Data are given as mean (SD).
ER, external rotation of hip on static examination; IR, internal rotation of hip on
static examination; MaxHR, maximum hip rotation on kinematics (minus means external
rotation); MinHR, minimum hip rotation on kinematics (minus means external rotation);
MaxPR, maximum pelvic rotation on kinematics (minus means retraction); MinPR,
minimum pelvic rotation on kinematics (minus means retraction); FPA, foot progression
angle (minus means external foot progression angle).
FIGURE 2. Follow-up of maximum hip rotation in stance phase
(mean values).

q 2005 Lippincott Williams & Wilkins 741

Kim et al J Pediatr Orthop  Volume 25, Number 6, November/December 2005
FIGURE 3. Follow-up of maximum pelvic rotation in stance
phase (mean values).
compared with P0 (post hoc multiple comparison test by the
Duncan multiple range test).
The revision rate for children under 10 years of age at
the time of surgery was greater than for those over 10 years of
age (P = 0.007, Fisher exact test).
DISCUSSION
In-toeing gait due to MFT is the most commonly seen
transverse plane deformity in children with CP. When this
deformity is marked, FDO is indicated. Staheli et al11 reported
that distal FDO was associated with complications such
as delayed union, stiffness of the knees, valgus deformity,
undercorrection, and persistent in-toeing compared with
proximal FDO. However, the method of internal fixation in his
paper was pins in plaster. In our study group, osteotomy was
solidly internally fixated with AO plates and screws using AO
techniques. No acute complications were observed in our
study. Recent papers using AO techniques have reported sim-
ilar results, that stable internal fixation of distal FDO yields
satisfactory clinical and radiologic results.3,9,12
Our patients had undergone multilevel procedures
including not only distal FDO but also other bony and soft
tissue procedures. The effects of soft tissue procedures on the
transverse plane kinematics in general are minimal.13,14 Our
results showed that distal FDO could normalize rotational
profiles at the 1-year follow-up. All the data except pelvic
rotation shifted to the side of external hip rotation. Saraph
et al15 also found that patients in the diplegia group showed
significant improvements in hip rotation with no significant
change in pelvic rotation after multilevel surgery.
Most long-term studies of patients undergoing femoral
rotational osteotomy included children with non-neuromus-
cular disorders, such as developmental dysplasia of the hip.16–18
Those studies found variable results in terms of the long-term
maintenance of surgical correction. Kleine16 and Mittel-
meier and Jäger17 reported gross recurrence (up to 40 degrees)
of the anteversion angle after derotational osteotomy. In
contrast, Svenningsen et al18 found in a prospective study that
the mean increase of femoral anteversion after subtrochanteric
742
TABLE 4. Comparison Between Preoperative and
Postoperative Values
P0 vs. P1 P1 vs. P3 P3 vs. P5 P1 vs. P5
ER ,0.001 0.192 0.125 ,0.001
IR ,0.001 0.487 0.531 0.190
MaxHR ,0.001 0.697 0.271 0.013
MinHR ,0.001 0.451 0.989 0.233
MaxPR 0.306 0.989 0.924 0.987
MinPR 0.632 0.677 1.000 0.648
FPA ,0.001 0.974 0.989 0.862
Data are given as P values. P values for P0 vs. P1 were obtained from paired t test.
P values for others were obtained from ANOVA test.
ER, external rotation of hip on static examination; IR, internal rotation of hip on
static examination; MaxHR, maximum hip rotation on kinematics (minus means external
rotation); MinHR, minimum hip rotation on kinematics (minus means external rotation);
MaxPR, maximum pelvic rotation on kinematics (minus means retraction); MinPR,
minimum pelvic rotation on kinematics (minus means retraction); FPA, foot progression
angle (minus means external foot progression angle).
derotational osteotomy in children without neuromuscular
disease was minimal: only 6 degrees (0.7 degrees per year).
Ounpuu et al7 found that correction achieved with FDO was
well maintained on 5-year follow-up for children with CP
(average age at the time of surgery was 8.1 6 3 years), and
concluded that FDO is a lasting treatment for the correction of
anteversion and associated internal hip rotation during gait in
children with CP. However, we found that much of the cor-
rection achieved with surgery was diminished on follow-up
between P1 and P5. Our results on hip external rotation and
kinematic hip rotation showed progressive deterioration of the
initial correction. Only hip internal rotation showed in-
significant change on follow-up. This may secondarily suggest
a decreasing arc of hip rotation, as well as loss of correction
with growth of the children.
Bruner and Baumann5 reported that the anteversion
angle progressed slowly but continuously during the growth
period, and this tendency toward recurrence could not be found
in studies with shorter follow-up times. Their study was done
on children with CP whose hips were subluxated or dislocated,
and with more severe involvement than in our study. As in our
study, most of the recurrence was observed in children younger
than 10 years of age. About 3 degrees of loss of hip external
rotation each year was observed in our patients between
P0 and P5.
FDO may be performed either proximally or distally. It
is unclear whether the difference in osteotomy level might
result in a difference in the recurrence rate.
We found a statistically significant relationship between
deterioration of correction and age at operation. The correction
obtained at operation becomes deteriorated due to the patho-
logic forces that originally led to the deformity of the growing
femur. Therefore, the results of our study on FDO to correct
the in-toeing gait in children with CP suggest the following:
(1) parents and children should be counseled that children
undergoing surgery at younger than 10 years of age may re-
quire repeat surgery; (2) FDO should be done to overcorrect
slightly for patients younger than 10 years; and (3) when de-
ciding on the amount of derotation, the surgeon should take
q 2005 Lippincott Williams & Wilkins
J Pediatr Orthop  Volume 25, Number 6, November/December 2005
into account that the preoperative pelvic rotation might not be
corrected postoperatively for the spastic diplegia.
ACKNOWLEDGMENTS
The authors thank Rosemary Pierce, PT, and Robin
Dorociak, biomechanical engineer, Portland Shriners Hospi-
tal for Children Gait Analysis Laboratory, for their work in
support of this study. We also thank Dr. Hwajung Kim, at the
Seoul National University School of Medicine, Korea, for her
work on statistical analysis.
REFERENCES
1. Bleck EE. Orthopaedic management in cerebral plasy. London: MacKeith
Press, 1987.
2. Gage JR. Gait analysis in cerebral palsy. London: MacKeith Press, 1991.
3. Aiona MD, Spens AK, Orendurff MS, et al. Kinematic changes in gait
after femoral derotational osteotomies in children with cerebral palsy. Dev
Med Child Neurol. 1996;38(Suppl 74):s1.
4. Bleck EE. Mechanism and treatment of internal rotation of the hip
deformity in cerebral palsy. Acta Orthop Belg. 1984;50:273–274.
5. Bruner R, Baumann JU. Long-term effects of intertrochanteric varus-
derotation osteotomy on femur and acetabulum in spastic cerebral palsy:
an 11- to 18-year follow-up study. J Pediatr Orthop. 1997;17:585.
6. Hoffer MM. Current concepts review: management of the hip in cerebral
palsy. J Bone Joint Surg [Am]. 1986;68:629–631.
q 2005 Lippincott Williams & Wilkins
Recurrence After Femoral Derotational Osteotomy in CP
7. Ounpuu S, Deluca P, Davis R, et al. Long-term effects of femoral dero-
tation osteotomies: an evaluation using three-dimensional gait analysis.
J Pediatr Orthop. 2002;22:139–145.
8. Root L. Treatment of hip problems in cerebral palsy. AAOS Instr Course
Lect. 1987;36:237.
9. Cooke PH, Carey RPL, Williams PF. Lower femoral osteotomy in cerebral
palsy: brief report. J Bone Joint Surg [Br]. 1989;71:146–147.
10. Aktas S, Aiona MD, Orendurff M. Evaluation of rotational gait abnor-
malities in the patients with cerebral palsy. J Pediatr Orthop. 2000;20:
217–220.
11. Staheli LT, Clawson DK, Hubbard DD. Medial femoral torsion:
experience with operative treatment. Clin Orthop. 1980;146:222–225.
12. Pirpiris M, Trivett A, Baker R, et al. Femoral derotation osteotomy in
spastic diplegia. J Bone Joint Surg [Br]. 2003;85:265–272.
13. Arnold AS, Delp SL. Rotational moment arms of the medial hamstrings
and adductors vary with femoral geometry and limb position: implication
for the treatment of internally rotated gait. J Biomech. 2001;34:437–447.
14. Schmidt DJ, Arnold AS, Caroll NC, et al. Length changes of the
hamstrings and adductors resulting from derotation osteotomies of the
femur. J Orthop Res. 1999;17:279–285.
15. Saraph V, Zwick EB, Zwick G, et al. Effect of derotation osteotomy of the
femur on hip and pelvis rotations in hemiplegic and diplegic children.
J Pediatr Orthop B. 2002;11:159–166.
16. Kleine L. Nachuntersuchungen von kongenitalen Hüftdyplasien nach in-
tertrochanterer Varisations und Derotationsoteotomi. Z Orthop. 1961;94:
266–286.
17. Mittelmeier H, Jäger M. Ergebnisse der Antetorsionkorrektur bei der
intertrochanteren Femur osteotomy dysplastischer Hüften im Kinderalter.
Arch Orthop Unfallchir. 1969;65:1–12.
18. Svenningsen S, Terjwsen T, Apalset K, et al. Osteotomy for femoral
anteversion. Acta Orthop Scand. 1990;61:360.
743