La raideur est un motif relativement fréquent d'échec et de reprise des prothèses du genou. Nous avons étudié les résultats de 34 mobilisations sous anesthésie générale, de 18 arthrolyses sous contrôle arthroscopique et de 10 arthrolyses chirurgicales pratiquées pour la raideur postopératoire d'une prothèse totale du genou. Ces 3 groupes étaient comparables pour tous les paramètres étudiés sauf le délai entre le geste et l'arthroplastie (17 semaines pour les mobilisations, 46 semaines pour les arthrolyses sous contrôle arthroscopique, 97 semaines pour les arthrolyses chirurgicales). Les résultats de ces 62 gestes mobilisateurs ont été analysés rétrospectivement.

Les amplitudes de mobilité articulaire acquises après le geste de libération n'évoluaient plus à partir de 6 mois. La flexion moyenne passait de 58,4 à 94,6° en postopératoire et le flessum de 7,6 à 2,5°. Si les résultats des 3 techniques ne différaient pas globalement, il apparaissait toutefois de façon significative des échecs de la mobilisation sous anesthésie lorsqu'elle était pratiquée au-delà de 8 semaines après la pose de la prothèse et après 6 mois pour l'arthrolyse sous contrôle arthroscopique.

Nous recommandons donc de ne pratiquer une mobilisation sous anesthésie générale que jusqu'à la huitième semaine, puis entre 8 semaines et 6 mois, une arthrolyse sous contrôle arthroscopique. Au-delà de 6 mois, nous recommandons la réalisation d'une arthrolyse chirurgicale. Ce protocole s'adresse à des prothèses non infectées et sans défaut technique de pose.

Stiffness of the knee is a common reason for revision of total knee arthroplasty. Three methods are currently used to mobilize the knee: manipulation under general anesthesia, arthroscopic release, open surgical release. The purpose of the present work was to determine the respective indications of these three procedures in a large single-center study.

We retrospectively assessed all revision procedures without component exchange in patients with a stiff total knee prosthesis. Sixty-two procedures were performed in our institution between 1989 and 2001. All patients were followed for at least one year. There were 34 manipulations under general anesthesia, 18 arthroscopic release procedures, and 10 open surgical release procedures. The three groups were not different for all parameters studied except time interval between implantation of the prosthesis and the mobilization procedure: 17 weeks for manipulation under general anesthesia, 46 weeks for arthroscopic release, 97 weeks for surgical release. A comparable postoperative analgesia and rehabilitation program was instituted for all patients.

Range of flexion improved after all 62 procedures: mean 58.4° before the procedure, mean 94.6° at one-year follow-up. Flexion deformity also improved from 7.6° to 2.5° at one year (p = 0.001). From surgery to one-year follow-up, there was a decrease in flexion (104.6° to 94.6°) and an increase in flexion deformity (1.3° to 2.5°) (NS). The worst postoperative ranges of motion were observed at six weeks after the procedure. Improvement was then observed up to six months but was not significant. There was no improvement in flexion beyond six months after the mobilization procedure. The results of the three techniques were not significantly different. Failures were however more frequent when manipulation under anesthesia was performed more than eight weeks after prosthetic insertion, and when arthroscopic release was performed more than six months after prosthetic insertion (p < 0.01).

We recommend treatment of stiff total knee prosthesis by manipulation under general anesthesis if the procedure is performed less than eight weeks after implantation; a delay of six weeks is even better because intraoperative complications were observed for patients treated between six and eight weeks. Between eight weeks and six months, arthroscopic release should be advised, surgical release thereafter. Whatever the delay, this protocol is appropriate for stiff knee prostheses without infection and without component malposition. Whatever procedure is applied, the definite range of motion is reached six months after the intervention.