Acetabular reconstruction with bulk structural autogenous grafts and allografts from the femoral head in complex total hip arthroplasty was highly successful at an average of five years postoperatively but was much less so by 11.8 years. To assess the longer-term fate of such grafts, we reviewed the results of eighty-one consecutive total hip arthroplasties performed by the senior one of us with use of these grafts. Nine hips in nine patients who had died and two hips that were infected in two patients were excluded. Therefore seventy hips (sixty-two patients) were included in this study. The average duration of follow-up was 16.5 years (range, 14.1 to 21.4 years).Sixty-one arthroplasties were performed to treat various forms of congenital dysplasia. Fifteen arthroplasties, ten of which were revision operations, were performed with allograft, and fifty-five were performed with autogenous graft. The average age of the patients at the time of the index operation was 45.2 years (range, sixteen to sixty-nine years). All of the sockets, which had an average outer diameter of forty millimeters (range, thirty-four to fifty millimeters), were inserted with cement. The average coverage of the acetabular component by the bulk graft was 49 per cent (range, 15 to 100 per cent).All of the grafts united. At the latest follow-up examination, twenty-five acetabular components (36 per cent) had been revised for aseptic loosening, eighteen (26 per cent) had radiographic evidence of loosening, and twenty-seven (39 per cent) were rigidly fixed and in place. The average Harris hip score for the hips in which the implant remained rigidly fixed was 74 points, while that for the hips in which the implant was loose but had not been revised was 69 points.Nine of the fifteen acetabular components supported by allograft and sixteen (29 per cent) of the fifty-five supported by autogenous graft were revised (p = 0.03). However, the total rate of acetabular components that were either loose or revised was ten of fifteen and thirty-three (60 per cent) of fifty-five, respectively. This difference was not significant (p = 0.4), with the numbers available.Regression analysis revealed that a younger age at the time of the operation and the extent of coverage of the acetabular component by the graft were associated with the need for revision. Twenty-one (78 per cent) of the twenty-seven acetabular components that remained rigidly fixed were supported by graft over less than 50 per cent of the contact area, while only nine (36 per cent) of the twenty-five that were revised were so supported (p < 0.05). None of the nine acetabular components with 30 per cent of the contact area or less covered by graft were revised.In nineteen of the twenty-two revisions of the acetabular component performed after the index operation, the socket was inserted without cement; the average outer diameter of the socket was fifty-three millimeters (range, forty to fifty-eight millimeters).Both the structural autogenous grafts and the structural allografts used in acetabular reconstruction in total hip replacement functioned well for the initial five to ten years. By an average of 16.5 years, nine of the fifteen hips treated with allograft and sixteen (29 per cent) of the fifty-five treated with autogenous graft had been revised. The greater the extent of the coverage of the acetabular component by the graft, the greater the rate of late failure.