[Outcome of hip shelf arthroplasty in adults after a minimum of 15 years of follow-up. Long term results and analysis of failures of 56 dysplastic hips].

PURPOSE OF THE STUDY: The goal of this study was to evaluate late results of hip shelf arthroplasty in adults after a minimum of 15 years follow-up.

MATERIAL: 65 hip shelf arthroplasty performed for painful hip dysplasia between 1964 and 1977 were studied retrospectively in 1992. These 65 procedures were performed in 57 patients mean aged 32 +/- 14 years [17-56]. Nine patients (9 hips) were excluded (2 deceaded, 5 lost for follow-up, and two reoperated because of severe infection). Consequently, the functional results were evaluated for 56 hips (48 patients). Before surgery, according to Merle d'Aubigné's hip rating system, all the hips were painfull (mean pain score was 2.6 +/- 1.7 [0-5]). On radiography, all the hips had a dysplastic acetabulum and arthritic changes. Arthritic changes were severe in 32 hips (57.1 per cent).

METHODS: The hip shelf arthroplasty was carried out according to Roy-Camille. 10 hips had additional varus femoral osteotomy. The 48 patients (56 hips) included were evaluated by means of Merle d'Aubigné's hip rating system and AP and false lateral weight-bearing Xrays. In 1992, 24 procedures were changed for total hip replacement (THR) (17 before 15 years (early failure) and 7 after 15 years of follow-up (late failure)). These 24 hips were included with their last hip rating observed just before THR. Survival analysis was performed according to Kaplan-Meier using date of revision for THR as end-point.

RESULTS: After 16.1 +/- 5.6 years of follow-up the functional score for 56 hips was: excellent in 4 hips, very good in 7 hips, good in 10, satisfactory in 14, poor in 17, and bad in 4. The survival rate established for 65 hips was 60 per cent at 15 years and 40 per cent at 21 years. Only 39 hips shelf arthroplasties were still functional after 15 years (mean follow-up 19.1 +/- 3 years [16-28], but 18 hips (46.1 per cent) were painfree or slightly painful (pain score to 5 or 6). Among these 39 hips, the results were excellent in 4 hips, very good in 7 hips, good in 10, satisfactory in 10, poor in 6, and bad in 2. Arthritic change was the main reason for failures: the Kaplan-Meier survival rate at 21 years was 87 per cent when arthrosis was slight and only 15 per cent and 42 per cent when arthrosis was moderate to severe (p = 0.0001). The adverse effect of arthrosis was promoted by lack of congruency for early failures, and by severity of dysplasia and hip subluxation for late failures. The additional femoral varus osteotomies had no influence on functional or radiographic outcome.

DISCUSSION: Our study indicated that hip shelf arthroplasty performed for painful acetabular dysplasia in adult has a 40 per cent probability survival rate at 21 years. The high rate of revision (42.8 per cent) could be related to the prevalence of severe pre-operative arthrosis. The low rate of lost for follow-up (8.7 per cent) and the long follow-up period (16.1 years) made our conclusion reliable.

CONCLUSION: In spite of a high revision rate we recommend shelf athroplasty to treat acetabular dysplasia in adults. This procedure, very reliable in cases of moderate arthrosis, could be performed in cases of severe arthrosis to delay and make easier THR, but a low survival rate could be expected unless dysplasia, lack of congruency and subluxation were mild.