A consecutive series of patients with a hydroxyapatite-coated uncemented total knee replacement (TKR) performed by a single surgeon between 1992 and 1995 was analysed. All patients were invited for clinical review and radiological assessment. Revision for aseptic loosening was the primary outcome. Assessment was based on the Knee Society clinical score (KSS) and an independent radiological analysis. Of 471 TKRs performed in 356 patients, 432 TKRs in 325 patients were followed for a mean of 16.4 years (15 to 18). The 39 TKRs in 31 patients lost to follow-up had a mean KSS of 176 (148 to 198) at a mean of ten years. There were revisions in 26 TKRs (5.5%), of which 11 (2.3%) were for aseptic loosening. Other further surgery was carried out on 49 TKRs (10.4%) including patellar resurfacing in 20, arthrolysis in 19, manipulation under anaesthetic in nine and extensor mechanism reconstruction in one. Survivorship at up to 18 years without aseptic loosening was 96% (95% confidence interval 91.9 to 98.1), at which point the mean KSS was 176 (134 to 200). Of 110 knees that underwent radiological evaluation, osteolysis was observed in five (4.5%), one of which was revised. These data indicate that uncemented hydroxyapatite-coated TKR can achieve favourable long-term survivorship, at least as good as that of cemented designs.