Comparison of three methods of gluteal muscle attachment to an allograft/endoprosthetic composite in a canine model

This study used radiography, gait analysis, gluteal muscle mass, mechanical testing. and qualitative histology to compare three methods of gluteal muscle attachment to an allograftiendoprosthetic composite of the proximal 25% of the femur in an in vivo canine model. The three methods of gluteal muscle attachment were identical to those used clinically in human patients lor hip revision and proximal femoral limb salvage: the host gluteal tendon sutured to the allograft tendon (tendon group), the host greater trochanter with intact gluteal tendons secured to the allograft with a cable-grip system (grip group), and periosteally vascularized proximal femoral bone onlay with intact tendons wrapped around the allograft (wrap group). On the basis of radiographs taken every 2 months, the tendon group had more graft fractures than did the grip or wrap group. Radiographic union ol the graft-host bone junction occurred more rapidly and therc was less graft resorption in the wrap group than in the other two groups. In all dogs, peak vertical ground-reaction forces in the treated limb decreased immediately after surgery and then slowly increased over the length of the study. The dogs in the wrap group regained nornial weight-bearing on the treated limb more quickly than did those in the other groups. The constructs in the tendon group were weaker and less stiff immediately after surgery ihan were those in the other groups or in intact cuntrols. Histologic analysis confirmed that the wrap technique rcsulted in complete union of the host bone-allograft junction more often than did the other techniques. The wrap method had the best functional outcome after 9 months when an allograftiendoprosthetic composite was used during total hip arthroplasly in this canine model.