A randomized controlled trial was performed to compare the cemented Stanmore metal-on-metal (Biomet, Warsaw, Indiana) total hip arthroplasty (THA; 102 hips) to the cemented Stanmore metal-on-polyethylene (Biomet) THA (98 hips). The primary outcome was clinical performance. Radiological performance, serum cobalt analysis, and prosthetic survival were secondary outcome measures. At a mean follow-up of 5.6 years, 5 patients were lost to follow-up, 18 died, and 4 were revised (3 metal-on- metal, 1 metal-on-polyethylene). Harris Hip Scores improved from 48 to 90 in the metal-on-metal patients (P<.001) and from 46 to 87 in the metal-on-polyethylene patients (P<.001). Oxford Hip Scores changed from 40 to 19 in the metal-on-metal group (P<.001) and from 40 to 18 in the metal-on-polyethylene group (P<.001). For both Harris and Oxford Hip Scores, there was no significant difference between the 2 groups. Five-year survival with revision for any reason was 97% (95% CI 93%-100%) in the metal-on-metal group and 99% (95% CI 97%-100%) in the metal-on-polyethylene group. All revisions were indicated for aseptic loosening (metal-on-metal: 3 cup revisions; metal-on-polyethylene: 1 total revision). At 5-year follow-up, cemented metal-on-metal THA showed no clinical superiority over metal-on-polyethylene THA.