In order to clarify how intra-articular lesions influence the survival of a periacetabular osteotomy in patients with dysplasia of the hip, we performed an observational study of 121 patients (121 hips) who underwent a transposition osteotomy of the acetabulum combined with an arthroscopy. Their mean age was 40.2 years (13 to 64) and the mean follow-up was 9.9 years (2 to 18). Labral and cartilage degeneration tended to originate from the anterosuperior part of the acetabulum, followed by the femoral side. In all, eight hips (6.6%) had post-operative progression to Kellgren–Lawrence grade 4 changes, and these hips were associated with the following factors: moderate osteoarthritis, decreased width of the joint space, joint incongruity, and advanced intra-articular lesions (subchondral bone exposure in the cartilage and a full-thickness labral tear). Multivariate analysis indicated subchondral bone exposure on the femoral head as an independent risk factor for progression of osteoarthritis (p = 0.003). In hips with early stage osteoarthritis, femoral subchondral bone exposure was a risk factor for progression of the grade of osteoarthritis.
Although the outcome of transposition osteotomy of the acetabulum was satisfactory, post-operative progression of osteoarthritis occurred at a high rate in hips with advanced intra-articular lesions, particularly in those where the degenerative process had reached the point of femoral subchondral bone exposure.
Developmental dysplasia of the hip (DDH) is a common cause of secondary osteoarthritis (OA).1 Many morphological features, including a shallow acetabulum and reduced acetabular cover of the femoral head, result in abnormal loading of the articular cartilage and instability.2,3 This can lead to subsequent labral tears and degeneration of the articular cartilage leading to OA.4-6 Periacetabular osteotomies have been developed to manage such abnormalities to delay or prevent the development of OA.7-10 The outcomes of these osteotomies can be affected by various factors,11-16 with a consensus that their benefit is greatest in younger patients with no or mild pre-operative OA.10,17-19
However, previous studies have reported that the post-operative progression of OA can vary between 6% and 35% of the operated hips, even if the osteotomy was performed in the early stages of OA.14,15,19,20 Studies on intra-articular lesions in DDH have shown a high prevalence of articular cartilage degeneration prior to the appearance of osteo-arthritic changes on a plain radiograph, and the degree of these lesions varied widely in hips with similar radiological stages of OA.4,21 This variation in cartilage degeneration is suggested to affect the results of the osteotomy.22 However, the mechanism by which the intra-articular lesions influences the outcome remains unclear.
Transposition osteotomy of the acetabulum (TOA),23,24 first described in 1956, is a type of spherical periacetabular osteotomy which may be used for the treatment of DDH. It is undertaken through a lateral trochanteric approach. The concept has much in common with other periacetabular osteotomies (Fig. 1). In the present study, we sought to determine how the pre-operative state of an intra-articular lesion affects the outcome of an osteotomy, by analysing the results of TOA using an intra-operative arthroscopic examination for DDH.
Patients and Methods
The institutional review board approved this study. Between August 1990 and January 2001, the senior author (YN) performed a TOA, combined with arthroscopy of the hip on 126 consecutive patients (139 hips) with symptomatic DDH. The indication for the osteotomy was radiological evidence of dysplasia with a lateral centre-edge angle25 < 20°, in patients with pain in the hip that interfered with daily activities. No patient had any other orthopaedic conditions, such as skeletal dysplasia or neuromuscular disease. We excluded five hips in five patients, who lacked more than two years of follow-up. In 13 patients who had bilateral osteotomies, only one hip was randomly selected to fulfil the statistical assumption of independent observation.
A total of 121 patients (121 hips) were included in the study. There were 11 men and 110 women. Their mean age at operation was 40.2 years (13 to 64). The mean follow-up was 9.9 years (2 to 18). A total of 48 patients had undergone treatment for congenital dislocation of the hip and an osteotomy had been performed in 13 other patients. The pre-operative stage of OA was graded radiologically according to the Kellgren–Lawrence classification system (Fig. 2).26,27 In all, 25 hips (20.6%) were classified as grade 1; 68 (56%) as grade 2; and 28 (23%) as grade 3. We identified the early stages of OA as grade 1 or 2 hips. Clinical evaluation was performed pre-operatively and one year post-operatively using the Merle d’Aubigné and Postel clinical scores.28
The TOA was performed based on the method introduced by Nishio (Fig. 1).23 Briefly, the patient is placed in the lateral decubitus position. A lateral trochanteric approach with an osteotomy of the greater trochanter through a single posterosuperiorly curved incision is used. After circumferential exposure of the articular capsule, aspherical osteotomy is performed using a curved osteotome, starting 20 mm proximal to the superior acetabular edge and passing through the midpoint between the greater sacral fossa and the posterior edge of the acetabulum and the innominate sulcus of the ischium. A pubic osteotomy is performed lateral to the iliopubic tubercle. The acetabular fragment is then adducted to achieve a horizontal weight-bearing acetabulum and medialisation of the femoral head. The acetabular fragment is rotated anteriorly as is needed to correct anterior deficiency. The acetabular fragment is fixed to the pelvis using two or three 4.5 mm cortical screws. An arthroscopic examination is performed through a small incision in the articular capsule. A concomitant transtrochanteric valgus osteotomy of the femur29 was performed in seven patients with severe femoral head deformity to minimise the effect of joint incongruency.
An active range of hip exercises and non-weight-bearing mobilisation begins one week post-operatively with partial weight-bearing on two crutches two weeks post-operatively progressing to full weight-bearing at five to eight weeks post-operatively.
We reviewed standard supine anteroposterior radiographs of the pelvis that were taken pre- and post-operatively and at the time of annual follow-up. Additional pre-operative radiographs were taken with the hip abducted to simulate joint congruency after the osteotomy. The lateral centre-edge angle, the Sharp angle30 and the acetabular head index31 were recorded. The minimum joint space width was measured as the minimum radial distance between the acetabulum and the femoral head in the weight-bearing area. The severity of the deformity of the femoral head was evaluated using the roundness index.32 Joint congruency was classified into three grades: good, poor and narrowed, based on the classification of Okano et al.16 The presence of a cross-over sign and a posterior wall sign was also recorded as indices of acetabular retroversion and posterior acetabular deficiency, respectively.33 We determined the change in Kellgren–Lawrence grade on annual radiographs taken after the osteotomy.
One author (MF) performed all the radiological evaluations in a blinded manner during two reviewing sessions held one month apart. Intra-observer reliabilities for numerical measurements, which were evaluated using an intraclass correlation coefficient, were excellent (0.9117 to 0.9899). Intra-observer reliabilities for categorical evaluations, which were evaluated using kappa statistics, were in substantial agreement (0.8064 to 0.8630).
Articular cartilage degeneration was classified into five grades according to the classification of Outerbridge (Table I).34 Labral lesions were classified into four grades, based on the severity of the tear, as follows: grade 0, normal; grade I, superficial tear; grade II, substantial tear; grade III, full-thickness tear. The location of intra-articular lesions was evaluated by dividing the articular weight-bearing area of the acetabulum and femoral head into three regions (Fig. 3).4 We adopted evidence for subchondral bone exposure in the cartilage and a full-thickness labral tear as possible risk factors for the progression of OA in subsequent analysis.
Using these data, we conducted the following analysis: 1) incidence and distribution of the intra-articular lesions; 2) clinical and radiological results of the TOA; 3) risk factors affecting the survival of the TOA with progression to Kellgren–Lawrence grade 4 or conversion to total hip replacement (THR) as the endpoint.
The chi-squared test and Wilcoxon rank-sum test were used to compare the categorical and continuous parameters between the two groups, respectively. A Wilcoxon signed-rank test was used to compare the pre- and post-operative parameters. The cumulative probability of the progression of OA was estimated using the Kaplan-Meier product-limited method with 95% confidence intervals (CI). The Cox proportional hazard model was used to analyse the independent risk factors affecting the progression of OA. The univariate Cox model was applied to each possible factor and the significant factors were examined together using the multivariate Cox model to determine their contribution to the progression of OA. A significant difference was defined as a p-value < 0.05. Statistical analyses were performed using JMP Software (Version 8.0; SAS Institute, Cary, North Carolina).
A high incidence of grade I or higher intra-articular lesions was observed, even in Kellgren–Lawrence grade 1 hips; labral lesions were found in 24 hips (96%), acetabular cartilage lesions in 22 hips (88%), and femoral cartilage lesions in 11 hips (44%). The lesions were common in the anterosuperior area of the acetabulum (Fig. 4a). Full-thickness labral tears were found in four hips (16%) and subchondral bone exposure was observed in two hips (8%).
In Kellgren–Lawrence grade 2 hips, the lesions were more advanced and varied more widely (Fig. 4b). A full-thickness labral tear was found in 50 hips (73.5%), acetabular subchondral bone exposure in 43 hips (63.2%), and femoral subchondral bone exposure in 14 hips (20.6%).
In Kellgren–Lawrence grade 3 hips, a full-thickness labral tear was found in 25 hips (89.3%), acetabular subchondral bone exposure in 26 hips (92.9%), and femoral subchondral bone exposure in 19 hips (67.9%) (Fig. 4c).
The mean Merle d’Aubigné and Postel score significantly improved from 13.1 (9 to 17) points pre-operatively to 16.4 (10 to 18) at final follow-up (p < 0.001, Wilcoxon signed-rank test). The radiological indices also showed significant improvement (Table II). The cross-over sign and posterior wall sign were corrected in 15 of 25 hips (60.0%) and 57 of 99 hips (57.6%), respectively. In ten of 93 hips (10.8%) with early stage OA progression to Kellgren–Lawrence grade 3 occurred. Among these, two hips (2.2%) deteriorated to Kellgren–Lawrence grade 4, and one subsequently required a THR. In addition six (21.4%) of 28 Kellgren–Lawrence grade 3 hips deteriorated to 4. A total of eight hips (6.6%) showed progression to grade 4, and these hips had a significantly lower mean clinical score than those without progression at the latest follow-up (12.9 (10 to 15) versus 16.7 (13 to 18); p < 0.001, Wilcoxon rank-sum test). A Kaplan-Meier analysis identified a ten-year survival rate of 94.1% (95% CI 86.6 to 97.5; 70 hips at risk) with progression to Kellgren–Lawrence grade 4 as endpoint (Fig. 5).
Complications related to the surgery were experienced by five patients; three had a superficial infection, which resolved unremarkably with antibiotics. One patient had asymptomatic nonunion of the superior pubic ramus. One patient had a pulmonary embolus following a deep vein thrombosis on the sixteenth post-operative day. This was successfully treated by thrombolysis.
When we compared hips with and without progression with Kellgren–Lawrence grade 4, hips with progressive OA were associated with low pre-operative clinical scores, Kellgren–Lawrence grade 3, decreased minimum joint space width, and joint incongruity (Table III). The incidence of intra-articular lesions was significantly higher in hips with progressive OA, and the location of the lesions was similar between the two groups (Table IV). The incidence of intra-articular lesions was also significantly higher in hips with an unsatisfactory clinical score at the latest follow-up (Table V). A univariate Cox model indicated that a Kellgren–Lawrence grade 3 (p = 0.002), a pre- and post-operative minimum joint space width (p < 0.001 and p = 0.002, respectively), a pre- and post-operative joint incongruity (p = 0.007 and p = 0.049, respectively), subchondral bone exposure in the acetabulum and femur (p = 0.004 and p < 0.001, respectively), and a full-thickness labral tear (p = 0.008) were significant risk factors for progressive OA. When these factors were analysed together, only subchondral bone exposure on the femoral head was an independent risk factor (p = 0.003). While no hip without femoral subchondral bone exposure deteriorated to Kellgren–Lawrence grade 4, eight (23.5%) of 34 hips with femoral subchondral bone exposure deteriorated.
In the subgroup of hips with early stage OA, the incidence of intra-articular lesions was significantly higher in hips with progression to Kellgren–Lawrence grade 3 (Table VI). Hips with femoral subchondral bone exposure showed progression of OA at a high rate (8 of 15 hips, 53.3%). Multivariate analysis showed femoral subchondral bone exposure was the only independent risk factor of progression to Kellgren–Lawrence grade 3 (p = 0.006).
Previous studies have shown favourable intermediate to long-term results of periacetabular osteotomies.11-13,15,19,22,35 A recent study by the Bern group reported a survival rate of 87.6% at ten years, and 60.5% at 20 years after Bernese periacetabular osteotomy.17 We demonstrated that TOA provides significant correction of deformity and symptomatic relief for patients with DDH. The outcome of TOA, with a survival rate of 94% at ten years, is comparable with the outcome of other osteotomies.
In this study, however, eight hips (6.6%) showed deterioration to severe OA and these hips had an unsatisfactory clinical score. According to univariate analysis, the following factors seemed to affect the progression of OA; decreased pre-operative Merle d’Aubigné and Postel score, Kellgren–Lawrence grade 3, decreased minimum joint space width, joint incongruity, and advanced intra-articular lesions. The first three factors are associated with advanced joint degeneration at the time of surgery, and previous studies also identified these factors to be predictors of poor outcomes.12,13,17,19,36 Joint incongruity has been identified as a predictor of the failure in studies involving other osteotomies.11,16 We believe good pre-operative joint congruity in abduction is essential in order to ensure a good outcome after TOA. Several studies have reported an adverse effect of femoral head deformity on the progression of OA,14,32 but we did not find this. We performed transtrochanteric valgus osteotomy in seven hips with deformity of the femoral head to minimise the effect of joint incongruence and this may have had an impact on the results.
Previous studies identified an effect of suboptimal correction.13-15,17 While we found no difference between hips with and without progression of OA with regard to the post-operative radiological indices of DDH, we believe adequate acetabular correction is critical for a successful outcome. Correction of abnormal acetabular retroversion has also been an integral part of a periacetabular osteotomy.35,37,38 The absence of this concept during the study period resulted in ten of 25 hips (40.0%) remaining retroverted, post-operatively. No significant difference was noted in the prevalence of a cross-over sign between hips with and without progression of OA. Garras, Crowder and Olson39 and Yasunaga et al20 also reported no significant effects associated with post-operative acetabular retroversion. Since proximal femoral morphology also affects post-operative femoroacetabular impingement, further study is needed to appreciate fully the effect of acetabular retroversion on the clinical results.
Advanced intra-articular lesions such as subchondral bone exposure and full-thickness labral tears were strongly correlated with post-operative progression of OA. Siebenrock et al13 also reported that the presence of a labral tear negatively influenced the outcome of periacetabular osteotomy. Yasunaga et al22 reported that cartilage degeneration at the time of surgery significantly affected the short-term results of a rotational acetabular osteotomy. Additionally, subchondral bone exposure on the femoral head was the most potent predictor of post-operative progression of OA in this study.
Our results and previous studies4,21 suggest a pathomechanism for intra-articular lesions whereby the degenerative process in DDH begins before the changes of OA appear on a plain radiograph and tends to originate on the acetabular side, from the anterosuperior part of the labrum and the weight-bearing area, followed by the femoral side. When the results of these studies are taken together, a possible mechanism for the post-operative progression of OA emerges. Although the osteotomy redistributes the contact stresses across the joint and significantly decreases the peak pressure,40 the pre-existing femoral cartilage lesions remain in the same weight-bearing area and are exposed to substantial post-operative stress. This may compromise the regeneration of articular cartilage lesions on the femoral head following surgery. If the degenerative process has reached the point where the femoral subchondral bone is exposed at the time of osteotomy, the hip is less likely to benefit from the improved biomechanical environment offered by the osteotomy; thus, OA will continue to progress.
The results of this study indicate limitations associated with the use of pre-operative radiographs for assessment of joint degeneration and for predicting the prognosis after an osteotomy. Even in hips with early stage OA, 15 hips (16.1%) had femoral subchondral bone exposure and these hips showed a high rate of progression of OA. On the other hand, in hips with moderate OA, nine hips (32.1%) without femoral subchondral bone exposure did not deteriorate. Thus, a detailed pre-operative evaluation of intra-articular lesions, with particular attention to the existence of advanced femoral cartilage degeneration, is recommended for identifying hips with a poor prognosis among hips at a similar osteo-arthritic stage and for selecting appropriate candidates for osteotomy. Cunningham et al41 suggest that a delayed gadolinium-enhanced MRI to demonstrate the articular cartilage is useful for predicting the clinical outcome following a periacetabular osteotomy. We believe these imaging modalities enable accurate prediction of the outcome by allowing greater understanding of the correlation between the imaging findings and the actual state of the intra-articular lesions.
In conclusion, TOA provides good correction of acetabular dysplasia and symptomatic relief for patients with DDH at a mean follow-up of ten years. However, care should be taken with hips where intra-articular changes have already advanced to the point of femoral subchondral bone exposure, as these hips may develop progressive OA following osteotomy.
No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article.
- Received April 14, 2011.
- Accepted July 5, 2011.
- © 2011 British Editorial Society of Bone and Joint Surgery