Between January 1995 and June 2001, the senior author (M.N.) performed a
rotational acetabular osteotomy for twenty-one consecutive patients
(twenty-five hips) with extensive osteonecrosis and collapse of the femoral
head who were subsequently followed for a minimum of two years
postoperatively. The mean age of the five male and sixteen female patients was
29.0 years (range, sixteen to thirty-nine years) at the time of surgery. The
inclusion criteria were an age of less than forty years at the time of
surgery, an osteonecrotic lesion that affected >50% of the weight-bearing
surface of the femoral head, collapse of the femoral head as seen on both
anteroposterior and lateral radiographs, and the presence of an intact lateral
cortex of the femoral head.
Rotational acetabular osteotomy was performed according to the method
described by Ninomiya and
Tagawa8. All of the
patients were followed clinically and radiographically at one-year intervals.
The mean duration of follow-up was forty-nine months (range, twenty-four to
101 months), and no patient was lost to follow-up. All but two of the patients
had a history of steroid therapy as part of a treatment for various diseases:
fourteen patients had systemic lupus erythematosus, three had a history of
dermatomyositis, one had acute lymphocytic leukemia, and one had uveitis. The
other two patients had alcohol-associated osteonecrosis.
The radiographic diagnosis of osteonecrosis of the femoral head was made on
the basis of anteroposterior and lateral plain radiographs as well as magnetic
resonance images. The lesions were classified according to the staging system
described by Steinberg et
al.9 (see Appendix).
The extent of osteonecrosis was quantified on plain radiographs according to
the method of Kerboul et
al.10, in which the
arcs of necrotic involvement of the femoral head are measured on
anteroposterior and lateral radiographs and total necrosis is expressed as the
sum of the two angles. Necrosis was considered to be extensive when the sum of
the necrotic angles was =200° and moderate when the sum of the necrotic
angles was 155° to 199° (Figs. 1-A
and 1-B). To assess the extent of collapse of the femoral head, we
measured the distance between the superior subchondral border of the femoral
head and the head-neck junction and compared this value with that for the
contralateral femoral head (Fig.
2).
Operative Technique
Rotational acetabular osteotomy was performed according to the method
described by Ninomiya and
Tagawa8. The
osteotomy line was located 1 to 1.5 cm proximal to the acetabular rim, where a
straight osteotome was inserted into the bone to a depth of approximately 1.5
cm. Then, a curved osteotome was used to complete the osteotomy through the
inner wall of the ilium. A capsulotomy was not performed. After the osteotomy,
adequate coverage of the femoral head could be achieved by moving the
acetabulum inferiorly and laterally. Two or three cortical bone struts,
approximately 5 mm thick, were then placed over the osteotomized bone surfaces
and were supplemented with cancellous bone chips. Two Kirschner wires (2 mm in
diameter) or polylactic acid absorbable screws (4.2 mm in diameter) were then
inserted to fix the rotated acetabulum and the cortical bone grafts to the
ilium.
Postoperative Management
With the patient at bed rest, isometric exercises of the quadriceps were
performed during the first two weeks after surgery. Transfer to a wheelchair
began two weeks postoperatively. Non-weight-bearing walking on two crutches
began at three weeks. Partial weight-bearing was permitted at two months, and
full weight-bearing was resumed at four to six months.
Assessment
Clinical and radiographic follow-up examinations were performed at yearly
intervals. Clinical information was obtained by means of an interview and
physical examination by two of the authors (M.N. and K. Maezawa). The scoring
system of Merle d'Aubigné and
Postel11 was used
to grade the functional level preoperatively and at the time of the final
follow-up (see Appendix). The hip-scoring system of the Japanese Orthopaedic
Association was used for the evaluation of activities of daily living (maximum
score, 20
points)12.
All follow-up anteroposterior and lateral radiographs were made with the
patient in the supine position. To compare femoral head coverage before and
after surgery, we devised an index that was calculated as the distance between
the medial margin of the femoral head and the lateral border of the necrotic
segment divided by the distance between the medial margin of the femoral head
and the acetabular border (Fig.
3). To quantify containment of the lateral femoral head cortex
after surgery, we analyzed the lateral head index as described by Masuda et
al.13
(Fig. 4). To compare the
coverage of the femoral head before and after surgery, we evaluated the
center-edge angle of
Wiberg14 and the
acetabular femoral head index of Heyman and
Herndon15. To
assess arthritic changes, we measured the width of the joint space at the
weight-bearing surface before surgery and at the time of the final follow-up
and recorded the mean joint space at three different points (the lateral edge,
the medial edge, and the center of the weight-bearing surface). To assess the
medialization of the femoral head, we measured the change in the distance
between the symphysis pubis and the medial border of the femoral head before
and after
surgery16. To
reduce the variability of these measurements, we used the shape of the
obturator foramen as a control in every anteroposterior radiograph. To measure
the distal movement of the femoral head, the distance between the inferior
margin of the lateral part of the pubis and the head-neck junction was
compared before and after
surgery16.
Statistical Analysis
The mean clinical scores and radiographic indices were compared with use of
the paired-sample t test. The level of significance was p < 0.05.
Preoperatively, four hips were classified as Steinberg stage IV-B,
seventeen were classified as stage IV-C, and four were classified as stage
V-C. The mean Kerboul necrotic angle was 238° (range, 171° to
464°). There was extensive necrosis in twenty hips and moderate necrosis
in five; no hip had minor necrosis. The extent of preoperative collapse was 2
mm in four hips, 3 mm in eight, 4 mm in six, 5 mm in two, 6 mm in three, 8 mm
in one, and 11 mm in one.
Clinical Results (see Appendix)
At the time of the final follow-up, there was marked improvement in
symptoms (particularly pain). The mean Merle d'Aubigné and Postel score
improved significantly from 11.3 points (range, 8 to 15 points) preoperatively
to 14.9 points (range, 10 to 18 points) at the time of the final follow-up (p
< 0.001). The mean pain score improved significantly from 3.0 points
(range, 0 to 5 points) to 5.6 points (range, 5 to 6 points) (p < 0.001),
and the mean gait score improved significantly from 3.0 points (range, 1 to 5
points) to 4.4 points (range, 2 to 6 points) (p < 0.005). However, the mean
mobility score was 5.3 points (range, 4 to 6 points) preoperatively but only
4.9 points (range, 3 to 6 points) postoperatively (p = 0.1). In eleven hips,
the mobility score at the time of the final follow-up was decreased compared
with the preoperative score. The mean flexion decreased from 96.4° (range,
60° to 135°) to 87.4° (range, 45° to 120°) (p = 0.08), and
the mean abduction decreased significantly from 30.4° (range, 10° to
60°) to 22.4° (range, 10° to 35°) (p < 0.005). The mean
activities of daily living score improved significantly, from 14.0 points
(range, 4 to 20 points) to 17.6 points (range, 14 to 20 points) (p <
0.0005).
Radiographic Results (see Appendix)
All of the osteotomy sites showed radiographic evidence of osseous union at
the time of the latest follow-up. Collapse of the femoral head progressed in
seven hips, with the change being <2 mm in six hips and 4 mm in one hip at
the time of the final follow-up. Marginal osteophytes of the femoral head
appeared or grew larger in twenty hips. The mean width of the joint space was
3.7 mm (range, 2.2 to 6.0 mm) preoperatively and 3.3 mm (range, 1.0 to 5.6 mm)
at the time of the final follow-up (p = 0.1). The joint space was narrower
postoperatively in eleven hips. Two hips showed >2 mm of narrowing of the
joint space; the extent of collapse before surgery had been >8 mm in both
hips. The femoral head coverage index improved significantly in all hips, with
the mean score improving from 1.0 (range, 0.7 to 1.9) to 0.8 (range, 0.5 to
1.0) (p < 0.0002). The mean correction of the femoral head index was 21%,
with a mean value of 78% (range, 47% to 91%) preoperatively and 99% (range,
84% to 112%) postoperatively (p < 0.0002). The mean lateral head index
after surgery was 33.7% (range, 11.9% to 64.5%). The lateral head index was
>20% in all hips but two. The mean correction of the center-edge angle was
29°, with a mean value of 26° (range, 0° to 39°)
preoperatively and 55° (range, 36° to 78°) postoperatively (p <
0.0002). Medialization of the femoral head, as indicated by the mean change in
the distance between the symphysis pubis and the medial border of the femoral
head after surgery, was —1.1 mm (range, —10 to 9 mm). Eight
femoral heads had moved medially and twelve had moved laterally. The mean
change in the distance between the inferior margin of the lateral part of the
pubis and the head-neck junction after surgery was 3.8 mm (range, —10 to
13 mm), with the mean distance being 1.0 mm (range, —5 to 8 mm)
preoperatively and —2.8 mm (range, —11 to 5 mm) postoperatively (p
< 0.005). In nineteen hips, the shift was in a caudal direction. Magnetic
resonance imaging showed no enlargement of the necrotic area, and the
low-intensity region was partly changed to a high-intensity region in three
hips.
Complications
There were no intraoperative complications. We did not observe
postoperative osteonecrosis of the isolated acetabular fragment, permanent
neurovascular complications, or heterotopic ossification. One patient had
dysesthesias in the distribution of the lateral femoral cutaneous nerve, which
resolved with time.
We previously observed that some patients with osteonecrosis of the femoral
head in whom acetabular roof osteophytes developed reported relief of groin
pain17. We
speculated that improved acetabular coverage of the lateral portion of the
femoral head might lead to clinical improvement in patients affected by
extensive necrosis with collapse. On the basis of our success with the use of
rotational acetabular osteotomy for the treatment of osteoarthritis and
acetabular
dysplasia16, we
hypothesized that this procedure might prevent superolateral migration of the
femoral head and thus improve coverage.
Mont and
Hungerford18 stated
that the purpose of osteotomy in the treatment of osteonecrosis of the femoral
head is to move the necrotic segment away from the major load-transmitting
area of the acetabulum and to redistribute weight-bearing forces to articular
cartilage that is still supported by healthy bone. Moreover, Plenk et
al.19 suggested
that procedures that are performed to change the loading of the affected
portion of the femoral head may preserve the blood supply to the main
load-bearing area.
When osteotomy is considered for the treatment of osteonecrosis, it is
essential to evaluate the preoperative extent of collapse, the extent of
osteonecrosis, and the magnitude of osteoarthritic changes in order to choose
the appropriate operative procedure. Gallinaro and
Massé7
reported that the best results after flexion osteotomy of the proximal part of
the femur can be obtained in patients in whom the necrotic angle of Kerboul is
<200°. Pavlovcic and
Dolinar20 also
reported that the necrotic angle of Kerboul is an important indicator of the
final outcome of intertrochanteric osteotomy. Those investigators reported
good results for all of their patients in whom the preoperative necrotic angle
was approximately 180°. While the transtrochanteric rotational osteotomy
described by Sugioka et
al.6 is indicated
for the treatment of osteonecrosis when the lesion is localized to the
anterior aspect of the femoral head and occupies less than two-thirds of the
entire head, the extent of necrosis and collapse of the femoral head in our
patients was too great for this osteotomy to be performed.
In the present series, the joint space became narrower in eleven hips after
rotational acetabular osteotomy and arthritic changes appeared in five of
these eleven hips. In all five hips, the preoperative necrotic angle of
Kerboul had been >250°. However, collapse of the femoral head
progressed by =4 mm in only one patient (a thirty-two-year-old man with a
necrotic angle of 298°). Moreover, that patient had a postoperative
femoral coverage index of 1.0 and a lateral head index of 11.9, which suggests
that there was inadequate acetabular coverage for the lateral portion of the
femoral head. We believe that containment of the lateral femoral head cortex
is an essential component of the success of this operation.
Merle d'Aubigné et
al.21 reported a
good clinical result for forty-seven (80%) of fifty-nine stage-II or stage-III
hips that had been treated with a Pauwels varus or varus-rotation
intertrochanteric osteotomy and had been followed for as long as eight years.
Ito et al.22
reported that varus osteotomy alone can produce sufficient coverage of the
intact portion of the femoral head in hips with necrosis of the femoral head.
However, Ito et
al.22 also reported
that three of four patients who had >25° of varus correction and three
of fifteen patients who had <25° of correction had a permanent limp.
Thus, the value of the varus intertrochanteric osteotomy may be limited when
the extent of correction that is necessary to cover the femoral head results
in a limp and/or excessive shortening of the limb. Additionally, varus
osteotomy is limited in terms of the amount of lateral bone that can be
delivered into the acetabular cavity when the necrotic area is extensive.
We believe that achieving adequate acetabular coverage of the viable
lateral portion of the femoral head was the most important factor in achieving
a good clinical result (Figs. 5-A and
5-B). In comparison with varus osteotomy, the rotational
acetabular osteotomy is not associated with shortening of the involved limb,
and full coverage of the femoral head (as shown by a mean postoperative
center-edge angle of 55°) can be achieved without abduction insufficiency.
Moreover, healing of the osteotomy site was quite rapid and all of the
patients could walk with full weight-bearing three to four months after
surgery.
We believe that patients with Steinberg stage-III, IV, or V lesions who are
younger than forty years of age and in whom the lateral femoral head cortex is
intact can be successfully managed with rotational acetabular osteotomy. This
procedure led to a reduction of symptoms and prevented progression of collapse
(even in hips with stage-IV and V lesions) by providing containment of the
lateral femoral head cortex. Although five hips had progression of arthritic
changes and eleven hips had joint-space narrowing after early to
intermediate-term follow-up, no hip had needed an arthroplasty by the time of
the most recent examination.
We have performed total hip arthroplasty in eight patients who had
undergone a previous rotational acetabular osteotomy. We found that the
osteotomy did not adversely change the acetabular anatomy or present technical
difficulties that would have compromised the results of total hip
arthroplasty. The osteotomy preserves excellent bone stock. After early to
intermediate-term follow-up, rotational acetabular osteotomy for the treatment
of extensive osteonecrosis with collapse of the femoral head in young patients
was associated with improvement in symptoms and a delay in the need for total
hip arthroplasty. Longer follow-up will be necessary to document the
persistence of these results over time.
Tables presenting the Steinberg classification, the scoring system of Merle
d'Aubigné and Postel, and the clinical and radiographic data on all
twenty-one patients are available with the electronic versions of this
article, on our web site at
(go to
the article citation and click on "Supplementary Material") and on
our quarterly CD-ROM (call our subscription department, at 781-449-9780, to
order the CD-ROM). ?