Les résultats cliniques et les difficultés techniques rencontrées lors de l'implantation de prothèses totales de hanches, après ostéotomie de l'extrémité supérieure du fémur, ont été rétrospectivement étudiés. Une série de 75 arthroplasties mises en place entre 1978 et 1995 a été revue. La population était composée de 64 sujets, pour la plupart jeunes, actifs, souvent multi-opérés. Les étiologies étaient : la luxation congénitale dans 32 % des cas, la dysplasie dans 23 %, la coxarthrose primitive ne représentant que 34 % des cas. Deux tiers des hanches étaient raides et plus de 3/4 instables. L'intervention initiale était une ostéotomie inter-trochantérienne dans 68 cas et une ostéotomie sous-trochantériennne dans 7 cas.

Il s'est produit 8 fractures du fémur lors de la mise en place de la prothèse. La position médiale de la fossette digitale par rapport à l'axe de la diaphyse a conditionné les difficultés opératoires et les complications fémorales.

Les résultats fonctionnels au dernier recul étaient très bons et bons dans 57 cas (76 %). Les 18 résultats non satisfaisants correspondaient aux 7 échecs repris chirurgicalement, à 5 descellements potentiels et à 6 luxations congénitales raides. Il y a eu 7 reprises chirurgicales (9,3 %) : 3 pour sepsis (dont 2 tardifs hématogènes), 1 luxation précoce, 2 descellements acétabulaires aseptiques et 1 descellement fémoral aseptique.

La courbe de survie des implants, établie selon la méthode actuarielle, retrouvait à 10 ans 94,9 % de probabilité de conserver une prothèse encore en place. L'ostéotomie initiale, hormis les ostéotomies sous-trochantériennes n'a pas entraîné de « perte de chance » pour le fémur, mais le nombre de complications et le risque de malposition de la tige fémorale doivent inciter à recourir à la trochantérotomie lorsque la fossette digitale se projette en dedans de l'axe de la diaphyse.

Clinical outcome and technical difficulties observed after total hip arthroplasty subsequent to osteotomy of the proximal femur were studied in 75 total hip arthroplasties.

Sixty-four patients underwent 75 total hip arthroplasty procedures at the Caen University Hospital between 1978 and 1995. These patients were reviewed at least two years after implantation. The Postel Merle d'Aubigné (PMA) score was used to assess clinical outcome and the Lequesne criteria to determine acetabular and femoral parameters on the weight-bearing AP radiograph. Off-set of the femoral epiphysis was determined by measuring the relation between the femoral shaft axis and the digital fossa. The population was composed of young active subjects who had had several operations. The main underlying diseases were: congenital dislocation of the hip (32%), dysplasia (23%), and primary degenerative hip disease (34%). Two-thirds of the hips were stiff and more than three-quarters were unstable. There were 28 osteotomies for varisation, 19 medial translations, 20 osteotomies for valgisation, and 7 subtrochanteric osteotomies. All femoral pieces except one were cemented. Implantation required 39 trochanterotomies, 39 Hardinge approaches, and 5 deosteotomies.

The medial offset of the digital fossa compared with the diaphysis led to operative difficulties and femoral complications: Their were eight femoral fractures: 7 of these occurred with a digital fossa situated medially to the diaphysis. The Postel-Merle-d'Aubigné score at last follow-up was satisfactory in 57 cases (76%). Among the 18 non-satisfactory results, there was 7 failures requiring surgical revision, 5 potential loosenings, and 6 stiff hips in patients with congenital dislocation. For the 7 revisions (9.3%), 3 were for infection (2 late hematogenous), 1 for early dislocation, 2 for aseptic acetabular loosening, and 1 for aseptic femoral loosening. Medial translation of the digital fossa led to a larger number of varus positions (44%). For femoral cementing, 10 hips showed a non-progressive secondary lucent line; these prostheses were considered as potentially loose. Among them, varisation osteotomies has been performed in 5 cases, and in 8 the digital fossa was medial to the diaphyseal axis. There was only one aseptic femoral loosening requiring surgical revision. Implant survival, established according to the actuarial method, was 94.9% at 10 years.

The almost 95% implant survival at 10 years observed in our patients is comparable with most reports in the literature and slightly better than those reported at the 1997 SOFCOT symposium on total hip arthroplasty after 50 years (84%). The most important factor determining surgical difficulty and operative complications at the femoral level was the position of the digital fossa, more than the type of initial osteotomy.

Indications for osteotomy must account for subsequent total hip arthoplasty. Implantation is difficult after subtrochanteric osteotomy which can cause important technical problems. The rate of failure is very high. The risk of femoral complications and misalignment of the femoral stem, irrespective of the type of initial osteotomy, is greater when the digital fossa lies medially to the diaphyseal axis. We prefer trochanterotomy or desosteotomy for such cases. Osteotomy did not rule out arthroplasty for any of the femurs. There were however important operative difficulties and the frequency of complications suggest we should be most prudent about careful preoperative planning in the frontal plane.