Between December 1996 and October 2001, eleven female patients with
a history of Crowe
type-IV7 congenital
dislocation of the hip were treated at our institution with simultaneous
pelvic support osteotomy and distal femoral lengthening with a hybrid Ilizarov
external fixator6
(Table I). The primary
indications for the procedure were severe pain caused by degenerative changes
in the hip, limping with a positive Trendelenburg sign, and a unilateral hip
dislocation. The secondary indications were back and knee pain resulting from
a dislocated hip.
Clinical Evaluations
The Harris hip score, which represents pain, walking function, activities
of daily living, and range of motion of the hip joint, was calculated
preoperatively and at the time of the last follow-up. The Trendelenburg test,
described by Hardcastle and
Nade8, was performed
preoperatively and at the time of the last follow-up by two independent
orthopaedic surgeons. The patient was asked to stand on the affected limb with
the hip on the non-stance side flexed to about 30° and the knee flexed
enough to allow the foot to be clear of the ground. A positive Trendelenburg
test was recorded if the iliac crest was high on the affected side and low on
the unaffected side. A delayed positive Trendelenburg test was recorded if the
patient had an initially negative test, but, after standing on one leg for a
short time, the pelvis gradually began to fall toward the unsupported side and
the patient was not able to maintain the initial pelvic stability. All
patients had a positive Trendelenburg test preoperatively. The postoperative
recovery time needed to achieve a negative Trendelenburg test was recorded for
each patient.
Magnetic Resonance Imaging
The patients were placed supine on the magnetic resonance imaging table.
Imaging of the gluteus medius muscle was performed in the coronal plane with a
1.5-T magnet (Gyroscan Intera Master; Philips, Best, The Netherlands) and a Q
body coil. The coronal T2-weighted sequence was performed with a repetition
time of 3800 msec, an echo time of 100 msec, a field of view of 35 cm, a slice
thickness of 5 mm, an interslice gap of 0 mm, a matrix of 400 × 512
pixels, and three acquisitions with a total imaging time of three minutes and
forty-five seconds.
The length of the gluteus medius muscle from its origin to its insertion
was measured as a straight line preoperatively and postoperatively
(Fig. 1), and the difference
between the measurements was considered to be the change in the gluteus medius
length.
The volume of the gluteus medius muscle was determined with
three-dimensional image-processing on an independent workstation. The image
data were transferred to the workstation, and an isotropic voxel size was then
obtained by a trilinear interpolation routine. The volume of the muscle was
isolated from the total volume by manually drawing contours around the muscle
boundaries on a section-by-section basis
(Fig. 2). All contouring was
done independently by one radiologist and one orthopaedic surgeon. These data
were then resampled by means of bilinear and cubic interpolation for the final
three-dimensional rendering. The volume of the muscle was then determined by
summing all of the pertinent voxels within the resultant binary volume.
Radiographic Evaluation
Limb-length discrepancy was measured with computerized tomography
preoperatively and at the time of the last follow-up. Valgus angulation of the
proximal part of the femur was measured on the anteroposterior supine
radiographs of the pelvis at the time of the last follow-up
(Fig. 3). The radio-graphic
measurement was done by one of us.
Surgical Technique
The proximal osteotomy level and the valgus angle were determined on an
anteroposterior radiograph of the pelvis with the involved limb held in
maximum adduction.
After induction of general anesthesia, the patient was placed supine on a
fracture-table. The arthritic femoral head was resected in all patients
through an anterolateral incision with fluoroscopic guidance. With use of an
image intensifier, a 5 or 6-mm pin was inserted laterally from the greater
trochanter at a predetermined angle. A second pin was inserted 15 cm distally
from the first pin and perpendicular to the distal part of the femoral shaft.
An arch was attached to each pin, and one or two pins were added to each arch.
The proximal osteotomy was performed between the two arches. Valgus angulation
was achieved by bringing the arches parallel. For distal femoral fixation, a
Kirschner wire-and-pin combination was used, and a distal femoral osteotomy
was performed between the second pin and the distal femoral fixation to
equalize limb-length discrepancies and to restore the mechanical axis of the
lower extremity (Fig. 4).
Postoperative Management
Following a recovery period of seven to ten days postoperatively, gradual
lengthening was started from the distal osteotomy site at a rate of 0.25 mm
four times daily (total maximum daily lengthening of 1 mm per day) to
eliminate any limb-length discrepancy. The mechanical axis was corrected at
the end of the distraction period by making frontal plane adjustments at the
distal osteotomy site.
There was no standardized rehabilitation program for the patients. However,
passive knee motion was initiated on the day following the surgery. On the
third postoperative day, partial weight-bearing with crutches and active
flexion-extension of the knee were started. Weight-bearing was increased
gradually, as tolerated by the patients.
Isometric hip exercises were started on the third postoperative day, and
the patient was encouraged to lie on the unaffected side to achieve maximum
passive adduction of the affected hip to obtain bone contact between the
pelvis and the adducted proximal part of the femur. Passive range-of-motion
exercises of the hip in all directions within the limits of pain were
initiated four weeks postoperatively. Active-assisted and active
range-of-motion exercises to strengthen the hip abductor, extensor, and flexor
muscles were commenced as tolerated by the patient. After removal of the
fixator, intensive abductor muscle-strengthening exercises were initiated,
guided by physical therapists, until a negative Trendelenburg test was
achieved or for a maximum of six months. After six months, patients with a
positive Trendelenburg test were usually responsible for their own physical
therapy program at home.
Statistical Analysis
The preoperative and postoperative Harris hip scores and measurements of
the length and volume of the gluteus medius muscle were analyzed with the
Student t test with use of SPSS software (version 12.0; Chicago, Illinois).
The Spearman correlation coefficient test was used to calculate the
correlation between the Trendelenburg sign and proximal femoral valgus
angulation, lengthening of the gluteus medius, change in the gluteus medius
volume, and age; p < 0.05 was considered significant.
The mean interval between the surgical procedure and the magnetic
resonance imaging was forty-two months (range, thirty to sixty-four months).
The mean follow-up time was three years (range, twenty-three to fifty-nine
months) after removal of the fixator. The average age was 25.2 years (range,
thirteen to thirty-nine years) at the time of the operation.
Clinical Results
The average preoperative and last follow-up Harris hip scores were 52
points (range, 32 to 73 points) and 92 points (range, 77 to 98 points),
respectively (p < 0.01). All patients had a painless range of motion and
function of the hip and knee by the last follow-up visit. Two patients (Cases
5 and 9) continued to experience recurring back pain with activity at the time
of the last follow-up.
Of the eleven patients, three (Cases 7, 8, and 11) had a persistently
positive Trendelenburg test and two (Cases 9 and 10) had a delayed positive
Trendelenburg test at the time of the last follow-up. For the six patients
with a negative Trendelenburg test, the average recovery time from the
operation to the negative test was 12.5 months (range, nine to sixteen
months). There was a significant association between the patient's age at the
time of the operation and a positive Trendelenburg test at the time of final
follow-up (p = 0.01): four of the five patients who had a positive test were
at least thirty-one years old at the time of the operation. The average age
was 32.4 years for the patients who had a positive Trendelenburg test and 19.2
years for those who had a negative Trendelenburg test.
Magnetic Resonance Imaging
There was a significant improvement in muscle length following the pelvic
support osteotomy (p < 0.001). The average change in the gluteus medius
length was 19.3 mm (range, 14 to 28 mm); however, with the numbers available,
there was no association between the change in muscle length and the result of
the Trendelenburg test (p = 0.86): the average change was 19.2 mm in those
with a positive test and 19.3 mm in those with a negative test.
The muscle volume increased from a mean (and standard deviation) of 82
± 12.7 cm3 preoperatively to a mean of 87 ± 14.7
cm3 at the time of the last follow-up (p < 0.01)
(Fig. 5). There was a
significant negative correlation between the change in muscle volume and the
result of the Trendelenburg test (r = -0.63; p = 0.03), with an average
increase in the gluteus medius volume of 10% (range, 4.9% to 25.6%) in the
patients with a negative Trendelenburg test and 1% (range, -4.2% to 4.3%) in
those with a positive test. In two patients (Cases 7 and 11) who had a
positive Trendelenburg test, atrophy of the gluteus medius was seen on the
magnetic resonance images acquired thirty-six and fifty-two months
postoperatively.
The increased muscle volumes were still less than the volumes on the normal
contralateral side (p < 0.001): the average muscle volume on the affected
side was restored to only 66% (range, 43% to 89%) of that on the normal side.
The volume was restored to 74% (range, 58% to 89%) of the normal volume in the
patients who had a negative Trendelenburg test and to 61% (range, 43% to 82%)
of the normal volume in the patients who had a positive test.
Radiographic Results
The average limb-length discrepancy (and standard deviation) was 5 ±
0.6 cm preoperatively. At the time of follow-up, only one patient had a
persistent discrepancy, with 3 cm of shortening on the affected side secondary
to loosening of the fixator clamp during the lengthening procedure.
The mean valgus angulation of the proximal part of the femur was 48.7°
(range, 37° to 61°) at the time of the last follow-up. One
thirteen-year-old patient (Case 5) had some loss of angular correction in the
proximal part of the femur caused by bone-remodeling; however, it did not lead
to a positive Trendelenburg gait. With the numbers available, there was no
correlation between the valgus angulation of the proximal part of the femur
and the result of the Trendelenburg test (r = 0.14; p = 0.66).
This study demonstrates that pelvic support osteotomy is an
effective method for restoring abductor muscle length and volume even if a
Trendelenburg gait persists, as it did in five of our eleven patients.
Critical factors influencing the resolution of a Trendelenburg gait include
restoration of the volume of the gluteus medius muscle and a younger age at
the time of the operation. In contrast, with the numbers available, we found
no correlation between a positive Trendelenburg test and the change in the
gluteus medius length or the valgus angle of the proximal part of the
femur.
The high percentage of patients with a positive Trendelenburg sign in the
present study is not consistent with the findings in previous reports.
Manzotti et al.9
reported that nine of eleven patients who had late sequelae of septic
arthritis and were treated with a pelvic support osteotomy had improvement in
lower-limb function with resolution of the Trendelenburg gait. However, the
interval between the onset of symptoms and the operation, which might have
affected the amount of muscular atrophy, was not recorded. Also, there were no
patients with a high dislocated hip in the study by Manzotti et al., which may
explain the high success rate in that series. In another study, Kocaoglu et
al.10 performed a
pelvic support osteotomy in fourteen patients, eleven of whom had a congenital
dislocation. All patients with a congenital dislocation had a negative
Trendelenburg gait after the operation. The authors attributed the high
success rate to increased tensioning of the abductor muscles, accomplished by
moving the greater trochanter distally. In contrast, we found that the amount
of distal transfer of the greater trochanter (as measured by increased gluteus
medius length) alone was not adequate to prevent a Trendelenburg gait.
Age appears to be an important factor for retaining hip function after
pelvic support
osteotomy11. In the
present study, four of the five patients who had a persistently positive
Trendelenburg test were at least thirty-one years of age. On the basis of
these results, we speculated that an atrophied muscle might not be restored by
a pelvic support osteotomy in patients older than this age.
Restoration of the abductor muscle volume was closely correlated with
elimination of a positive Trendelenburg test. Atrophy of the muscle can cause
insufficiency of muscle
contraction11. In
the present study, atrophy of the gluteus medius occurred in two patients and
resulted in a Trendelenburg gait. Even though the cause of muscle atrophy is
unknown, it may explain why one twenty-two-year-old patient had a positive
Trendelenburg test.
In addition to abductor weakness, an unstable fulcrum can be a causative
factor in a Trendelenburg gait. The pelvic support osteotomy produces a stable
fulcrum with the altered weight-bearing surfaces and diminishes the abductor
torque required to achieve pelvic equilibrium by moving the fulcrum point
medially2,4,5.
However, our study showed that, despite the creation of a stable fulcrum in
all patients, insufficient restoration of the gluteus medius led to a positive
Trendelenburg test in five of them.
The disadvantage of the pelvic support osteotomy is limitation of hip
motion, particularly adduction and flexion, because of the valgus and
extension angulation of the proximal part of the
femur3,11.
This limitation is especially important in the sitting position because an
individual can compensate for it only by tilting the pelvis in the standing
position11. Our
study did demonstrate that a pelvic support osteotomy can result in a painless
and functional hip (a good or excellent Harris hip score) despite limitations
of hip adduction and flexion.
Several authors have recommended a total hip arthroplasty with or without
femoral shortening to achieve a painless hip and a negative Trendelenburg sign
in patients with degenerative arthritis as a result of congenital
dislocation7,12-23.
Even though this procedure is commonly successful, complications such as
peroneal nerve palsy, femoral nerve palsy, early postoperative dislocation,
late infection, femoral shaft fracture, nonunion or delayed union of the
femoral osteotomy site, and aseptic loosening have been reported in the
literature7,13,14,16,18,21-24.
Patients retain an active lifestyle following a pelvic support osteotomy,
which may be converted to a total hip arthroplasty later in life. However,
conversion of a pelvic support osteotomy has a high risk of intraoperative
complications15,16,25,26.
To our knowledge, there have been no studies comparing the outcomes and
complication rates of total hip arthroplasty between dislocated hips with and
without a previous pelvic support osteotomy.
A weakness of this study was that we investigated only the effect of
gluteus medius geometry on abductor torque. Other hip abductor muscles may
also generate important abductor torque that serves as a static stabilizer of
the pelvis during unilateral
stance27,28.
However, magnetic resonance imaging measurements of the other muscles,
especially on the affected site, were difficult technically.
In conclusion, pelvic support osteotomy achieved a functional, painless
hip; however, five of eleven patients still had a positive Trendelenburg test.
This study strongly suggests that restoration of abductor muscle volume after
pelvic support osteotomy is not sufficient to prevent a Trendelenburg gait in
older patients with congenital dislocation of the hip. ?