ORIGINAL ARTICLE

Gait Abnormalities Following Slipped Capital

Femoral Epiphysis
Kit M. Song, MD, Suzanne Halliday, MS, Chris Reilly, MD, and William Keezel, MEngr

recommendations for the use of redirectional osteotomies of

Summary: The authors evaluated 30 subjects with treated unilateral the proximal femur. To our knowledge, the impact of proximal
slipped capital femoral epiphysis and a range of severity from mild to
femoral deformity following SCFE on gait has not been previ-
severe to characterize gait and strength abnormalities using instru-
ously quantified. Additionally, the mechanism by which de-
mented three-dimensional gait analysis and isokinetic muscle testing.
For slip angles less than 30°, kinematic, kinetic, and strength vari- formity leads to long-term development of osteoarthritis is not
ables were not significantly different from age- and weight-matched well established. The purpose of this retrospective review was
controls. For moderate to severe slips, as slip angle increased, passive to identify trends and patterns between the degree of SCFE and
hip flexion, hip abduction, and internal rotation in the flexed and ex- observed biomechanical and gait abnormalities in short-term
tended positions decreased significantly. Persistent pelvic obliquity, follow-up.
medial lateral trunk sway, and trunk obliquity in stance increased, as
did extension, adduction, and external rotation during gait. Gait ve- MATERIALS AND METHODS
locity and step length decreased with increased amount of time spent With institutional review board approval, a search of
in double limb stance. Hip abductor moment, hip extension moment,
computerized records revealed 301 children who had been
knee flexion moment, and ankle dorsiflexion moment were all de-
creased on the involved side. Hip and knee strength also decreased
treated for acute and chronic SCFE at our institution from 1980
with increasing slip severity. All of these changes were present on the to 1995. Eighty-four met our inclusion criteria:
affected and to a lesser degree the unaffected side. Body center of 1. A unilateral SCFE with no clinical or radiographic evi-
mass translation or pelvic obliquity in mid-stance greater than one dence of a contralateral slip by the time of evaluation for
standard deviation above normal correlated well with the impression this study
of compensated or uncompensated Trendelenburg gait. 2. Follow-up at our institution for at least 1 year following in
Key Words: hip, slipped capital femoral epiphysis, gait analysis, out-
situ pinning of the slip
comes 3. No endocrinopathy or associated medical condition that
contributed to the SCFE
(J Pediatr Orthop 2004;24:148–155) 4. No avascular necrosis or chondrolysis following treatment
of the SCFE
5. No revision surgery
C hildren who have a slipped capital femoral epiphysis
(SCFE) are commonly reported to have gait and func-
tional disturbances after in situ pinning. Remodeling of the
Thirty subjects agreed to return for evaluation. Eight
control subjects with no history of any medical problems and
proximal femoral anatomy can lead to improvement of hip mo- with a comparable body mass index were recruited for com-
tion, but concerns for the possible development of degenera- parison.
tive arthritis in the middle to late adult years have led to many All patients were examined and were administered a
functional questionnaire and the Harris Hip Score18 for pain
Study conducted at the Texas Scottish Rite Hospital for Children, Dallas, and function. Body mass index (BMI)35 was calculated as
Texas. mass (kg)/height2 (m). Knee extension and flexion, hip abduc-
From Children’s Hospital and Regional Medical Center of Seattle, Seattle, tion and adduction, and hip flexion and extension were mea-
Washington (Dr Song); Oxford Orthopaedic Engineering Centre, Oxford,
England (Dr Halliday); Children’s Hospital of British Columbia, Vancou-
sured using a goniometer by one of us (C.R.), who was un-
ver, British Columbia, Canada (Dr Reilly); and Texas Scottish Rite Hos- aware of the side of the slip. Measurements were made in the
pital for Children, Dallas, Texas (Mr Keezel). supine position. Hip extension was measured by flexing the
None of the authors received financial support for this study. opposite side until the lumbar lordosis was flattened and then
Reprints: Kit M. Song, MD, Department of Orthopedic Surgery, Children’s checking hip extension of the side to be measured. Hip flexion
Hospital and Regional Medical Center of Seattle, 4800 Sand Point Way
N.E., P.O. Box 5371/CH-59, Seattle, WA 98104 (e-mail: kit.song@
was checked with the opposite hip extended fully by flexing
seattlechildrens.org). the hip to be measured with the leg in neutral ab/adduction and
Copyright © 2004 by Lippincott Williams & Wilkins. rotation until the hip moved into obligatory external rotation or

148 J Pediatr Orthop • Volume 24, Number 2, March/April 2004

J Pediatr Orthop • Volume 24, Number 2, March/April 2004 Gait Abnormalities After SCFE

until static restraints or the abdomen prevented further flexion. tion for the total mechanical work for each limb using the
Hip rotation was assessed in the supine position with the hip method of Eng10 but found that the variability of velocities was
flexed to 90° and in the prone position. All measurements were great enough to invalidate the calculations.
rounded to the nearest 5°. All patients had an instrumented gait A position of the total body center of mass (COM) was
analysis and isokinetic muscle testing of their lower extremi- calculated segmentally in the x, y, and z directions for each
ties. An orthoroentgenogram and an anteroposterior and frog- trial. We used the following assumptions for body segment
leg lateral x-ray of both hips were obtained on the day of test- mass. The trunk was 35.5% of total body mass, each arm was
ing. 5% of total body mass, the head and neck were 8.1%, each
Lower leg length discrepancy was measured from the thigh was 10%, each shank was 4.65%, and each foot was
orthoroentgenogram. Sticky lead balls were placed on the an- 1.45%. The trajectory of the COM during the gait cycle was
terior superior iliac spines to determine the accuracy of marker plotted and the maximum and minimum displacement for each
placement for subsequent gait analysis. Slip severity was cal- trial was determined.
culated using the method of Southwick33 on the lateral radio- We subjectively defined Trendelenburg gait as in-
graph taken at the 6-week postoperative visit. The difference in creased unilateral trunk sway toward (compensated) or unilat-
the Southwick angle between the slip side and the non-slip side eral drop of the hemipelvis away from (uncompensated) the
was used to divide the patients into five groups: slip angle less affected side in mid-stance.36 We objectively defined Tren-
than 20°, slip angle 20° to 29°, slip angle 30° to 39°, slip angle delenburg gait as loss of normal abductor moment peak of
40° to 49°, and slip angle greater than 50°. The articulo- greater than one standard deviation of normal variability or
trochanteric distance was calculated by drawing a transverse reversal to an adductor moment. We then correlated these find-
line between the inferior border of the teardrop of each hip, ings to trunk obliquity and medial and lateral displacement of
drawing a perpendicular vertical line to this, and measuring the the body COM in midstance and to visual impressions of Tren-
distance between a horizontal perpendicular line at the tip of delenburg gait by one of us (K.M.S.).
the greater trochanter and the roof of the acetabulum. We used a Cybex II machine to isokinetically test the
The patients and control subjects walked at a self- muscle strength of the hip flexors and extensors in the supine
selected speed along a 15-m runway into a calibrated volume position, hip abduction and adduction while side-lying, and
of 2.5 m. Simultaneous frontal and sagittal plane videotapes of knee flexion and extension while seated for each patient. The
each walk were made. Thirty-two reflective markers were hips were tested at 30°/s and the knees at 60°/s. The peak
placed as follows: bilaterally at the level of the ear canal, pos- torque per body weight was selected from one of five trials. We
terior glenohumeral joint, radial head, radial styloid; lateral did not normalize data to lean body mass.
iliac crest; anterior superior iliac spine; greater trochanter; lat- Pearson correlations were calculated between the inde-
eral epicondylar ridge of the femur, on a mid-thigh wand pendent variable of the Southwick angle and the dependent
aligned to the long axis of the femur; on the anterior aspect of variables of range of motion, kinematic, kinetic, and strength
the thigh; on a mid-shank wand aligned to the long axis of the variables for the individual patients. The slip severity and all of
tibia, on the anterior aspect of the shank; lateral malleolus; cal- the dependent variables were averaged for the patients within
caneus; on the midfoot between the second and third metatar- the five groups. Pearson correlations were calculated again be-
sal heads, and one on the spinous process midway between the tween all averaged dependent variables and averaged slip se-
posterior superior iliac spines and the spinous process of C7. verity. Correlations were interpreted at a level of significance
Three-dimensional kinematic data were recorded with a of P < 0.05. The use of P values, where reported, is intended to
six-camera 60-Hz Vicon system (Vicon, Oxford, United King- reflect a possible trend or effect, but they do not represent de-
dom). Cadence, velocity, and step length were recorded. Ki- finitive findings in the usual sense.
netic data were recorded with two AMTI strain-gauged force
plates. Four force plate strikes of each limb were recorded for RESULTS
each subject. Lead markers were placed on the anterior supe- There were 20 boys and 10 girls. The average age at sur-
rior iliac spines during orthoroentgenograms to check for ac- gery was 12 years and 8 months (girls 12 years and 2 months,
curacy of position. The obesity of our patients led to inaccurate boys 12 years and 11 months). No child had an intraoperative
placement in many cases. For this reason joint angle, internal or perioperative complication. Twenty-five patients under-
moments, and powers in sagittal, coronal, and transverse went pinning of their slip with a single cannulated screw. Five
planes were calculated using Move3D software from the Na- patients underwent pinning with two screws. Four patients had
tional Institutes of Health.23 This software uses a body seg- elective removal of asymptomatic hardware. No patient was
ment model that does not rely upon specific anatomic land- symptomatic at the operative site at the time of evaluation. The
marks for marker placement. Maximum and minimum values average age at evaluation was 16 years and 10 months. Eigh-
for joint angles, moments, and powers were identified for each teen patients were clinically obese at evaluation, as defined by
phase of the gait cycle. We considered performing a calcula- a BMI greater than 27.8 for boys and greater than 27.3 for girls.

© 2004 Lippincott Williams & Wilkins 149

Song et al J Pediatr Orthop • Volume 24, Number 2, March/April 2004

FIGURE 1. Slip severity as measured by Southwick angle difference between involved and noninvolved side.

Four patients had slip angles less than 20°, nine a slip
angle between 20° and 29°, eight a slip angle between 30° and
39°, six a slip angle between 40° and 49°, and three a slip angle TABLE 1. Harris Hip Scores
greater than 50° (Fig. 1). The epidemiologic characteristics
Harris Hip Score
and body mass index distribution for the five groups of patients Slip Angle
were not significantly different. There was a trend toward Pt. No./Sex Difference Part I Part II Part III Part IV
greater BMI with greater slip severity (r = 0.65).
1/M 7 40 47 4 5
The mean Harris hip score was 96 (range 67–100). The
2/M 14 44 47 4 5
mean scores for pain and function showed a trend for increas-
3/F 16 40 44 4 5
ing pain and diminished function with increasing severity of
4/M 18 44 47 4 5
slip (Table 1) but were not significantly different between the
5/M 20 44 47 4 5
five groups. No other component of the hip score showed a
6/M 20 44 47 4 5
correlation to severity of slip.
7/F 21 44 47 4 5
The mean passive joint range of motion for the affected
8/F 22 30 47 4 5
and unaffected sides of the five groups is shown in Table 2.
9/M 24 44 47 4 5
Measurement of hip adduction was found to be inaccurate due
10/M 25 44 47 4 5
to the obesity of the subjects and is not included. No significant
11/M 26 44 47 4 5
differences were found between the joint range of motion in
12/M 26 44 47 4 5
the normal literature and the unaffected side of our study pa-
13/M 28 44 47 4 5
tients. As slip angle increased, hip flexion on the slip side de-
14/F 33 44 47 4 5
creased (r = 0.85), hip abduction decreased (r = 0.96), and in-
15/M 34 44 47 4 5
ternal hip rotation decreased in both the flexed and prone po-
16/F 34 40 44 4 5
sitions (r = 0.9). Hip abduction also decreased on the non-slip
17/M 35 44 47 4 5
side (r = 0.88), as did hip internal rotation in the flexed (r =
18/F 35 44 44 4 5
0.96) and extended (r = 0.95) positions.
19/M 35 44 44 4 5
Gait velocity decreased with increasing slip severity (r =
20/F 36 44 47 4 5
0.8), and there was increased time spent in double limb support
21/F 37 44 47 4 5
for both the slip (r = 0.81) and normal (r = 0.85) sides. There
22/M 40 30 42 4 5
was a trend toward decreasing step length with increasing slip
23/M 41 44 47 4 5
severity (r = 0.7). Significant differences in velocity compared
24/M 42 44 47 4 5
with our normal population were seen once the slip angle ex-
25/M 45 40 47 4 5
ceeded 50°. No other range of motion or cadence variables
26/M 46 40 44 4 5
showed any relationship to slip severity or BMI.
27/M 46 44 44 4 5
The following significant kinematic correlations were
28/F 50 44 47 4 5
found with increasing slip severity. Persistent pelvic obliquity
29/M 54 20 38 4 5
(down on affected side) increased (r = 0.85), as did trunk obliq-
30/M 74 44 44 4 5
uity (r = 0.81). The slip-side hip became more extended (r =

150 © 2004 Lippincott Williams & Wilkins

J Pediatr Orthop • Volume 24, Number 2, March/April 2004 Gait Abnormalities After SCFE

TABLE 2. Passive Joint Range of Motion, Slip Side and Normal Side

Abduction ER at 90 IR at 90 ER Prone IR Prone Hip Flexion

Slip Angle
Pt. No./Sex Difference Slip Norm Slip Norm Slip Norm Slip Norm Slip Norm Slip Norm
1/M 7 50 50 70 60 5 15 55 55 10 20 130 130
2/M 14 60 60 60 60 20 40 55 60 20 35 130 130
3/F 16 55 55 80 80 25 25 80 80 40 40 130 130
4/M 18 50 50 90 90 20 40 90 90 35 40 135 135
5/M 20 30 40 50 45 0 20 60 40 20 45 130 130
6/M 20 45 45 55 50 20 40 45 55 70 50 130 130
7/F 21 50 50 60 65 30 25 85 50 45 45 130 130
8/F 22 45 55 50 55 0 30 55 50 35 65 130 130
9/M 24 35 40 55 40 25 40 45 40 40 55 135 135
10/M 25 45 50 60 65 40 40 75 60 40 40 130 130
11/M 26 30 40 80 70 −20 15 85 60 −15 30 110 110
12/M 26 40 40 60 50 15 30 80 55 30 60 120 120
13/M 28 60 50 60 60 15 10 60 60 40 45 135 135
14/F 33 40 40 70 50 35 50 80 55 35 50 130 130
15/M 34 30 30 70 60 −5 10 90 80 0 25 130 130
16/F 34 40 50 65 50 −25 0 90 50 −25 30 110 110
17/M 35 40 35 75 50 −20 10 90 65 −20 50 120 120
18/F 35 45 45 70 55 5 30 75 40 15 40 130 130
19/M 35 25 30 75 50 −30 20 75 50 0 30 90 90
20/F 36 45 45 45 30 0 25 90 90 5 45 130 130
21/F 37 45 45 70 30 5 40 75 40 10 40 110 110
22/M 40 50 50 70 50 −20 20 90 70 20 20 130 130
23/M 41 25 30 45 35 −30 25 55 30 −15 45 110 110
24/M 42 30 60 75 80 −15 10 80 55 −15 40 130 130
25/M 45 20 35 85 55 −20 15 50 40 −5 20 125 125
26/M 46 30 45 50 55 10 35 50 55 10 35 130 130
27/F 46 20 30 80 80 −50 20 90 80 −40 20 130 130
28/F 50 40 45 80 35 5 35 75 40 0 50 130 130
29/M 54 10 35 55 45 −35 15 90 55 −30 40 105 125
30/M 74 30 35 70 50 −20 10 90 90 −20 10 110 110

0.9), more adducted (r = 0.99), and more externally rotated (r = for limb advancement at the ankle (r = 0.82) and the hip (r =
0.99) throughout the gait cycle. Knee flexion decreased (r = 0.83). Calculations of total limb work were considered invalid
0.93) and foot progression angle became more external (r = due to variable velocities.
0.97). Nine patients were believed to have a compensated and
The following kinetic correlations were found with in- one an uncompensated Trendelenburg gait upon review of vid-
creasing slip severity. On the slip side, hip extension moment eotapes. Nine had abnormal hip moments and all 10 had ab-
(r = 0.91), knee flexion moment (r = 0.88), and ankle dorsi- normal medial-lateral translation of the body COM and trunk
flexion moment (r = 0.97) all decreased. The changes in hip obliquity or pelvic obliquity. In 5 of the 10, however, the ab-
abductor moments are shown in Figure 2. Hip abductor mo- normal medial-lateral translation of the body COM was away
ment decreased (r = 0.83) with increasing slip severity and be- from the affected side. There was not a good correlation be-
came an adductor moment in the most severe slips. There was tween severity of slip or changes in the articulo-trochanteric
good correlation between medial/lateral displacement of the distance and the magnitude of COM medial-lateral displace-
calculated body COM (r = 0.87), medial lateral trunk sway (r = ment, magnitude of trunk sway, or magnitude of pelvic obliquity.
0.92), and hip ab/adductor moment changes. Power absorption Isokinetic muscle testing revealed a progressive loss of
at the knee decreased (r = 0.94), as did power generation bursts maximal strength for knee extension, knee flexion, hip exten-

© 2004 Lippincott Williams & Wilkins 151

Song et al J Pediatr Orthop • Volume 24, Number 2, March/April 2004

FIGURE 2. Hip abductor and adductor moments.

sion, hip flexion, hip abduction strength, and hip adductor is believed to lead to long-term disability due to degenerative
strength on both the slipped side and the uninvolved side with arthritis and short-term functional deficits due to altered mo-
increasing slip severity. All correlations had an r value greater tion of the hip.2,4,5,13,16,19,29,32,37 Reduction of slipped epiph-
than 0.80 (Table 3). syses3,8,9 and osteotomy of the proximal femur for moderate to
As slip severity increased, the articulo-trochanteric dis- severe slips has been recommended to decrease the risk of de-
tance decreased on the involved side (r = 0.83) and on the un- generative arthritis.1,6,11,14,17,24,27,31,33,34 High complication
involved side (r = 0.91) (Table 4). We did not find a significant rates and inability to prove that corrective osteotomy decreases
relationship between slip severity and the radiographic or the risk of long-term arthritis have led some authors to recom-
clinical lower limb length discrepancy. mend such surgeries only for improvement of function.12,20,25
When comparing passive range of motion, strength, ki- Little information is available, however, on the relationship
nematic, and kinetic variables to our normals, we found that between the severity of proximal femoral displacement and
there were not significant differences for patients with slip functional or gait disturbances.
angles less than 30°. All patients with a slip angle greater than Rab recently defined the geometry of non-remodeled
40° had significant side-to-side differences and deviation from SCFE using a three-dimensional volume/surface computer
normals for all measured variables. We did not find a consis- model.30 He predicted that progressive alteration in gait and
tent pattern of abnormalities for these variables as slip angles range of motion would occur to avoid impaction of the metaph-
increased. yseal portion of the femoral neck. Such changes were pre-
dicted to be linear. We found that Rab’s geometric predictions
DISCUSSION held true in our subjects with moderate to severe unilateral
Children with an SCFE will have residual deformity of SCFE. We found linearly decreased hip flexion, hip abduction
the proximal femur. The deformity in moderate to severe slips and internal hip rotation, decreased velocity, and more exten-

TABLE 3. Mean Cybex Data: Peak Torque/Body Weight

Knee Ext Knee Flex Hip Ext Hip Flex Hip Abd Hip Add

Slip Slip Norm Slip Norm Slip Norm Slip Norm Slip Norm Slip Norm
Angle Cybex Cybex Cybex Cybex Cybex Cybex Cybex Cybex Cybex Cybex Cybex Cybex

<20° 55.3 (10.6) 61 (8.0) 34 (6.5) 38.8 (9.6) 78.3 (42.3) 78.5 (31.5) 34 (15.4) 36 (10.1) 38.8 (11.0) 34.3 (12.7) 57.5 (19.7) 64.8 (22.9)
20–29° 61.8 (13.3) 61.7 (13.0) 41 (9.3) 41.2 (10.0) 73.1 (21.1) 74.2 (21.3) 36.2 (8.8) 37.7 (9.7) 35.6 (9.5) 36.4 (10.5) 56.9 (19.0) 59.2 (13.7)
30–39° 56.1 (9.95) 57.1 (11.5) 35.1 (10.0) 37.1 (11.7) 58 (15.6) 66.1 (18.8) 30.3 (10) 32.4 (6.0) 32.5 (7.5) 32.1 (7.3) 49.6 (15.6) 57.9 (12.9)
40–49° 50.7 (14.1) 60.3 (13.2) 32.7 (9.0) 39.8 (8.7) 62.2 (19.4) 66.3 (19.9) 27.5 (10.2) 31.8 (6.9) 29.5 (10.9) 29.7 (10.2) 50.7 (14.4) 53.8 (7.2)
>50° 40.7 (13.2) 39.7 (14.0) 24 (3.6) 26.3 (5.5) 40.7 (3.79) 44.3 (5.5) 19.7 (0.58) 20.7 (3.2) 18.3 (2.52) 17 (6.1) 29.3 (5.86) 30.7 (6.8)

152 © 2004 Lippincott Williams & Wilkins

J Pediatr Orthop • Volume 24, Number 2, March/April 2004 Gait Abnormalities After SCFE

TABLE 4. Leg Length Discrepancy and shortened step length predicted by Rab30 and confirmed in this
Articulo-Trochanteric Distance study can account for some of the changes in sagittal plane
moments.
Slip Side Non-Slip Diff. The reason for the decrease in hip abductor moments is
Pt. No/Sex ATD CM ATD CM ATD LLD CM not clear. We had initially assumed that a decrease in the ar-
ticulo-trochanteric distance would explain the progressive de-
1/M 2.5 3.8 1.3 −2.5 crease in hip abductor moments and reversal to an adductor
2/M 1.7 2.6 0.9 0 moment in a severe slip due to altered muscle length and
3/F 2.4 3 0.6 0.5 strength. We did find a decrease in isokinetic muscle strength
4/M 3 3.8 0.8 1 with increasing slip severity for all muscle groups about the hip
5/M 2.5 2.7 0.2 0 and knee, but noted this for both the involved and uninvolved
6/M 1.6 2 0.4 1 sides. The presence of bilateral changes can be partly ex-
7/F 1.7 2.4 0.7 1 plained by the increasing BMI with more severe slips. Our Cy-
8/F 2.6 3 0.4 0 bex data were normalized to body weight and not lean body
9/M 1.5 1.8 0.3 0.5 mass, and this may skew the results.
10/M 2.4 3.2 0.8 0 Changes in hip abductor, adductor, extensor, and flexor
11/M 2.4 2.6 0.2 0 moments due to movement of the femoral head were predicted
12/M 2 3.1 0.9 −1 by Delp and Maloney.7 They modeled the effects of displace-
13/M 1.8 2.4 0.6 1 ment of the femoral head in three directions upon muscle mo-
ments of the hip and predicted that superior-inferior displace-
14/F 2 3.1 0.9 −1
ment of the femoral head would affect the abductor and adduc-
15/M 1.2 2.6 1.4 1
tor moments of the hip, and that these effects would be greater
16/F 1.9 3.3 1.4 1
than displacement in the anterior-posterior or medial-lateral
17/M 2.1 2.9 0.8 0.5
directions. Combining displacements resulted in little net
18/F 1.6 2.5 0.9 0
change in hip abductor moments. They postulated that offset in
19/M 1.8 2.3 0.5 0
the anteroposterior direction helped to re-establish muscle ten-
20/F 2 2.8 0.8 1
sion, diminishing the effects of the frontal plane deformity. We
21/F 1.7 1.9 0.2 0
could not verify these predictions in our population of children
22/M 1.8 2.7 0.9 0.5 with unilateral SCFE. Although the articulo-trochanteric dis-
23/M 1.9 2.2 0.3 1 tance decreased with increasing slip severity, the changes in
24/M 1.6 2.6 1 1 distance were comparable between the involved and unin-
25/M 1 2 1 1 volved side and did not correlate well with the presence or
26/M 1.3 2.7 1.4 1 absence of clinically apparent “abductor lurch” or changes in
27/F 2 3.2 1.2 1 abductor moments. This suggests to us that the development of
28/F 0.4 1.6 1.2 −1 an “abductor lurch” is not just related to the frontal plane rela-
29/M 1.2 2.1 0.9 1.5 tionship between the greater trochanter and the femoral head.
30/M −0.6 2 2.4 3 We believe that extension and external rotation at the slip site
contribute significantly to the clinical impression of Tren-
Positive descrepancy = slip leg shorter than normal side. delenburg gait and that this plane of deformity needs to be con-
sidered if alteration of gait is the desired goal of osteotomy.
The long-term consequence of observed gait abnormali-
ties for individuals who have SCFE is unknown. Unlike the
sion and external rotation of the hip during stance with increas- isokinetic muscle strength data, calculated changes in power
ing slip severity. For our subjects with slip angles less than 30°, and overall changes in limb work are greatly influenced by
however, we did not find any significant differences compared changes in the velocity of walking. This varied considerably in
with our normal population. It is known that remodeling of the our study population, and we cannot speculate on the effect of
metaphysis occurs,21,22,26,28,39 and we believe this may ex- progressively more severe deformity on limb work or joint
plain some of our differences from the predictions of Rab. forces. Rab30 noted that both impaction and inclusion were
We found increased slip severity to be highly correlated predicted for severe slips, and he believed that both of these
with a decrease in hip extension and abduction, knee flexion, factors might contribute significantly to the development of
and ankle dorsiflexion moments. Decreased hip extension mo- degenerative changes within the hip. While we could not dem-
ments are explained by the posterior displacement of the femo- onstrate a correlation between the severity of proximal femoral
ral head with relative extension of the proximal femur.7,30 The displacement and disability as measured by the Harris hip

© 2004 Lippincott Williams & Wilkins 153

Song et al J Pediatr Orthop • Volume 24, Number 2, March/April 2004

scores, we did find a trend toward increasing pain and dimin- Future studies will need to determine the optimal
ished function with more severe slips even in this very short amount of correction to strive for when performing surgeries to
follow-up. correct gait disturbances due to SCFE. In this series, children
Although Trendelenburg36 elegantly describes move- with mild slip angles were not found to have significant alter-
ment of the pelvis and of the trunk when walking with congen- ations in gait. Unfortunately, operative procedures done to cor-
ital hip dislocation, he does not offer any criteria for determin- rect deformity may create secondary bony deformities or
ing when gait changes can be considered abnormal. Many au- muscle weakness due to the surgical approach that can also
thors have used the term “Trendelenburg gait” in their lead to gait alterations.
description of undesirable gait abnormalities, but we could not In summary, we found that children with slip angles less
find any objective definition of this gait abnormality. We had than 30° did not exhibit a significant alteration in any definable
hoped to identify criteria that could be used to define “Tren- gait parameter as compared with age- and weight-matched
delenburg gait,” but we could not find a strong correlation be- normals. Progressively greater alteration in hip motion, hip
tween the clinical impression of either a compensated or un- and knee moments, hip strength, and functional limitations can
compensated abductor lurch and any kinetic variables. We did be expected for larger deformities. More severe slips had a
find that hip abductor and extensor moment changes correlated higher correlation with the clinical impression of an “abductor
well with increased medial-lateral sway of the trunk and body lurch” during gait. This was best defined objectively by an ex-
COM translation. cessive medial-lateral trunk sway and pelvic obliquity during
All subjects thought to have a Trendelenburg gait had single leg stance. The articulo-trochanteric distance did not
clinically abnormal medial-lateral translation of their body predict well which patients would develop this gait abnormal-
COM or increased pelvic obliquity in single leg stance. In 5 of ity. It appears that the extension and external rotation deformi-
ty of the proximal femur need to be considered in the manage-
the 10, however, the blinded clinical impression was that the
ment of gait abnormalities related to SCFE.
sway was away from the affected side. Work by Halliday et
al15 and Winter38 using dual force plates showed that initiation
REFERENCES
of gait involved the shifting of 90% of weight onto the stance-
phase limb for able-bodied young people as they threw them- 1. Abraham E, Garst J, Barmada R. Treatment of moderate to severe slipped
capital femoral epiphysis with extracapsular base-of-neck osteotomy. J
selves off balance to begin single leg stance. For elderly or Pediatr Orthop. 1993;13:294–302.
disabled individuals, the transference of weight decreases, as 2. Boyer DW, Mickelson MR, Ponseti IV. Slipped capital femoral epiphysis.
do velocity of gait and the length of the first step. Most clini- Long-term follow-up of one hundred and twenty-one patients. J Bone
Joint Surg [Am]. 1981;63:85–95.
cians have believed that the abductor lurch is done to shift 3. Broughton NS, Todd RC, Dunn DM, et al. Open reduction of the severely
weight more over the stance-phase limb to offset weakened slipped upper femoral epiphysis. J. Bone Joint Surg [Br]. 1988;70:435–
439.
abductors. Based on our findings and the work by Halliday et 4. Carney BT, Weinstein SL. Natural history of untreated chronic slipped
al and Winter, we speculate that the clinical appearance of the capital femoral epiphysis. Clin Orthop. 1996;322:43–47.
abductor lurch is sometimes a result of increased sway over the 5. Cooperman DR, Charles LM, Pathria M, et al. Post-mortem description of
slipped capital femoral epiphysis. J Bone Joint Surg [Br]. 1992;74:595–
unaffected limb during swing phase of gait to provide for better 599.
ability to “launch” oneself up onto the affected limb. In these 6. Crawford AH. The role of osteotomy in the treatment of slipped capital
situations, there may actually not be increased weight transfer femoral epiphysis. AAOS Instr Course Lect. 1989;38:273–279.
7. Delp SL, Maloney W. Effects of hip center location on the moment-
onto the affected limb, but perhaps a more rapid acceleration of generating capacity of the muscles. J Biomech. 1993;26:485–499.
weight onto that limb. We believe that our findings may be a 8. Dietz FR. Traction reduction of acute and acute-on-chronic slipped capital
useful start toward defining Trendelenburg gait abnormalities femoral epiphysis. Clin Orthop. 1994;302:101–110.
9. Dunn DM, Angel JC. Replacement of the femoral head by open operation
and evaluating the impact of treatment in future studies. in severe adolescent slipping of the upper femoral epiphysis. J Bone Joint
Pearson correlations using individual patient values did Surg [Br]. 1978;60:394–403.
10. Eng JJ, Winter DA. Kinetic analysis of the lower limbs during walking:
not reflect the relationships seen when grouping was per- what information can be gained from a three-dimensional model? J Bio-
formed. The grouping of data may have artificially increased mech. 1995;28:753–758.
the correlation coefficients compared with the use of indi- 11. Fish JB. Cuneiform osteotomy of the femoral neck in the treatment of
slipped capital femoral epiphysis. A follow-up note. J Bone Joint Surg
vidual data points and will result in less power when doing a [Am]. 1994;76:46–59.
comparison of means. Despite this, we do not believe it invali- 12. Gage JR, Sundberg BA, Nolan DR, et al. Complications after cuneiform
dates the trends observed in this population of patients with a osteotomy for moderately or severely slipped capital femoral epiphysis. J
Bone Joint Surg [Am]. 1978;60:157–165.
unilateral SCFE. We do acknowledge, however, that applica- 13. Hagglund G, Hannson LE, Sandstrom S. Slipped capital femoral epiphy-
tion of these results to individual patients must be done with sis in southern Sweden. Long-term results after nailing/pinning. Clin Or-
caution. However, the trends that were observed may help us to thop. 1987;217:190–200.
14. Hagglund G, Hansson LE, Ordeberg G, et al. Slipped capital femoral
better understand the underlying mechanisms that drive the pa- epiphysis in southern Sweden. Long-term results after femoral neck oste-
thologies seen. otomy. Clin Orthop. 1986;210:152–159.

154 © 2004 Lippincott Williams & Wilkins

J Pediatr Orthop • Volume 24, Number 2, March/April 2004 Gait Abnormalities After SCFE

15. Halliday E, Winter DA, Frank JS, et al. The initiation of gait in young, wedge osteotomy for severe chronic slipped capital femoral epiphysis. J
elderly, and Parkinson’s disease subjects. Gait Posture. 1998;1:8–14. Pediatr Orthop. 1989;9:412–416.
16. Hansson G, Billing L, Hogstedt B, et al. Long-term results after nailing in 28. O’Brien ET, Fahey JJ. Remodeling of the femoral neck after in-situ pin-
situ of slipped upper femoral epiphysis. J Bone Joint Surg [Br]. 1998;80: ning for slipped capital femoral epiphysis. J Bone Joint Surg [Am]. 1977;
70–77. 59:62–68.
17. Hansson G, Jerre R, Sanders SM, et al. Radiographic assessment of cox- 29. Ordeberg G, Hansson LE, Sandstrom S. Slipped capital femoral epiphysis
arthrosis following slipped capital femoral epiphysis. A 32-year follow- in southern Sweden. Long-term result with no treatment or symptomatic
up study of 151 hips. Acta Radiol. 1993;34:117–123. primary treatment. Clin Orthop. 1984;191:95–104.
18. Harris WH. Traumatic arthritis of the hip after dislocation and acetabular 30. Rab GT. The geometry of slipped capital femoral epiphysis: Implications
fractures: treatment by mold arthorplasty. J Bone Joint Surg [Am]. 1969; for movement, impingement, and corrective osteotomy. J Pediatr Orthop.
51:737–755. 1999;19:419–424.
19. Harris WH. Etiology of osteoarthritis of the hip. Clin Orthop. 1986;213: 31. Salvati EA, Robinson JHJ, O’Down TJ. Southwick osteotomy for severe
20–33. chronic slipped capital femoral epiphysis: results and complications. J
20. Jerre R, Hansson G, Wallin J, et al. Long-term results after realignment Bone Joint Surg [Am]. 1980;62:561–570.
operations for slipped upper femoral epiphysis. J Bone Joint Surg [Br]. 32. Siegel DB, Kasser JR, Sponseller P, et al. Slipped capital femoral epiph-
1996;78:745–750. ysis. A quantitative analysis of motion, gait, and femoral remodeling after
21. Jones JR, Paterson DC, Hillier TM, et al. Remodeling after pinning for in-situ fixation. J Bone Joint Surg [Am]. 1991;73:659–666.
slipped capital femoral epiphysis. J Bone Joint Surg [Br]. 1990;72:568– 33. Southwick WO. Osteotomy through the lesser trochanter for slipped capi-
573. tal femoral epiphysis. J Bone Joint Surg [Am]. 1967;49:807–835.
22. Kallio PE, Foster BK, Le Quesne GW, et al. Remodeling in slipped capital 34. Sugioka Y. Transtrochanteric rotational osteotomy in the treatment of id-
femoral epiphysis: sonographic assessment after pinning. J Pediatr Or- iopathic and steroid-induced femoral head necrosis, Perthes’ disease,
thop. 1992;12:438–443. slipped capital femoral epiphysis, and osteoarthritis of the hip. Indications
23. Kepple T. Move3D software for analyzing human motion. IEEE Future and results. Clin Orthop. 1984;184:12–23.
Directions Workshop, 1990. 35. Thomas CL. Taber’s Cyclopedic Medical Dictionary. Philadelphia: F.A.
24. Kramer WG, Craig WA, Noel S. Compensating osteotomy at the base of Davis, 1989.
the femoral neck for slipped capital femoral epiphysis. J Bone Joint Surg 36. Trendelenburg F. Uber den gang bei angeborener huftgelenksluxation.
[Am]. 1976;58:796–800. Dtsch Med Wchenschr. 1895;21:21.
25. Maussen JP, Rozing PM, Obermann WR. Intertrochanteric corrective os- 37. Weinstein SL. Natural history and treatment outcomes of childhood hip
teotomy in slipped capital femoral epiphysis. A long-term follow-up disorders. Clin Orthop. 1997;344:227–242.
study of 26 patients. Clin Orthop. 1990;259:100–110. 38. Winter DA. Anatomy, Biomechanic and Control of Balance During
26. Moreau MJ. Remodeling in slipped capital femoral epiphysis. Can J Surg. Standing and Walking. Waterloo, Canada: Waterloo Biomechanics, 1995.
1987;30:440–442. 39. Wong-Chung J, Strong ML. Physeal remodeling after internal fixation of
27. Nishiyama K, Sakamaki T, Ishii Y. Follow-up study of the subcapital slipped capital femoral epiphyses. J Pediatr Orthop. 1991;11:2–5.

© 2004 Lippincott Williams & Wilkins 155