We present the results for 4762 revision total hip arthroplasties with no previous infection in the hip, which were reported to the Norwegian Arthroplasty Register between 1987 and 2003. The ten-year failure rate for revised prostheses was 26% (95% CI 25 to 26). Cox regression analyses were undertaken separately for acetabular and femoral revision components. Cemented revision components without allograft was the reference category. For acetabular components, we found a significantly reduced risk of failure for uncemented revisions both with (relative risk (RR) = 0.66; 95% CI 0.43 to 0.99) and without (RR = 0.37; 95% CI 0.22 to 0.61) allograft. For femoral components, we found a significantly reduced risk of failure for uncemented revisions, both with (RR = 0.27; 95% CI 0.16 to 0.46) and without (RR = 0.22; 95% CI 0.11 to 0.46) unimpacted allograft. This reduced risk of failure also applied to cemented revision components with allograft (RR = 0.53; 95% CI 0.33 to 0.84) and with impaction bone grafting (RR = 0.34; 95% CI 0.19 to 0.62). Revision prostheses have generally inferior results when compared with primary prostheses. Recementation without allograft, and uncemented revision with bone impaction, were associated with worse results than the other revision techniques which we studied.