The use of ilioischial cage reconstruction for pelvic discontinuity has been replaced by the Trabecular Metal (Zimmer, Warsaw, Indiana) cup-cage technique in our institution, due to the unsatisfactory outcome of using a cage alone in this situation. We report the outcome of 26 pelvic discontinuities in 24 patients (20 women and four men, mean age 65 years (44 to 84)) treated by the cup-cage technique at a mean follow-up of 82 months (12 to 113) and compared them with a series of 19 pelvic discontinuities in 19 patients (18 women and one man, mean age 70 years (42 to 86)) treated with a cage at a mean follow-up of 69 months (1 to 170). The clinical and radiological outcomes as well as the survivorship of the groups were compared. In all, four of the cup-cage group (15%) and 13 (68%) of the cage group failed due to septic or aseptic loosening. The seven-year survivorship was 87.2% (95% confidence interval (CI) 71 to 103) for the cup-cage group and 49.9% (95% CI 15 to 84) for the cage-alone group (p = 0.009). There were four major complications in the cup-cage group and nine in the cage group. Radiological union of the discontinuity was found in all successful cases in the cup-cage group and three of the successful cage cases. Three hips in the cup-cage group developed early radiological migration of the components, which stabilised with a successful outcome.

Cup-cage reconstruction is a reliable technique for treating pelvic discontinuity in mid-term follow-up and is preferred to ilioischial cage reconstruction. If the continuity of the bone graft at the discontinuity site is not disrupted, early migration of the components does not necessarily result in failure.

Cite this article: Bone Joint J 2014;96-B:195–200.

We report 11 patients having revision of total hip arthroplasty using massive structural allografts for failure due to sepsis and associated bone loss. All patients had a two-stage reconstruction and the mean follow-up was 47.8 months (24 to 72). Positive cultures were obtained at the first stage in nine of the 11 patients, with Staphylococcus epidermidis being the most common organism. The other two patients had draining sinuses with negative cultures.

There was no recurrence of infection in any patient. The mean increase in the modified Harris hip score was 45 and all the grafts appeared to have united to host bone. Two patients required additional procedures, but only one was related to the allograft. Complications included an incomplete sciatic nerve palsy and one case of graft resorption. Our results support the use of massive allografts in failed septic hip arthroplasty in which there is associated bone loss.