Corrosion and wear of total hip (THA) and knee (TKA) prostheses extricate metallic particles and soluble metallic compounds. The oncogenic risk of these products should be known. Material and methods: Three Nordic cohorts of total hip (THA) and total knee arthroplasty (TKA) patients operated on for primary osteoarthrosis during 1967–1995 were combined for meta-analysis. The number of THA patients was 49,000 and TKA patients 24,000 totaling 497,000 person years. The mean follow-up time was 6.8 years. Standardized incidence ratios (SIRs) with 95% conþdence intervals (95% CI) were calculated for the observed and expected number of cancers. The expected numbers were based on national incidence rates. Results: The allover SIRs as well as the site-speciþc cancer incidences were similar for the THA and TKA patients. The observed number of all site cancers was 7639 and 8202 expected (SIR 0.93, 95% CI 0.91–0.95). The SIR for lung cancer (0.69, 0.64–0.75) was reduced. The incidence was also low for cancers of the stomach (SIR 0.76, 0.67–0.84), colon (SIR 0.86, 0.79–0.93) and rectum (SIR 0.89, 0.80–0.98). Slightly elevated SIRs were seen among TKA patients in cancer of the endometrium and prostate and among both THA and TKA patients in skin melanoma. Conclusions: Total cancer risk was signiþcantly reduced among THA and TKA patients due to decrease of respiratory and gastrointestinal cancers. The present results do not suggest any oncogenic risk of the components of hip and knee prostheses and their degradation products.