A nationwide cohort study of hip and knee replacement patients in Denmark was undertaken to assess any carcinogenic potential of these implants. A cohort of 22,997 osteoarthritis patients who received hip replacements and of 4,771 osteoarthritis patients who received knee replacements during the period 1977 through 1989 were identified using the nationwide Danish Hospital Discharge Registry. These patients were followed for cancer occurrence through 1993, using the Danish Cancer Registry. There was no overall excess of cancer in either the hip implant cohort [standardized incidence ratio (SIR) = 0.94; 95% confidence interval (CI)= 0.91-0.98] or the knee implant cohort (SIR = 0.97; 95% CI = 0.89-1.06). The risk reduction in both groups of patients reflected for the most part reduced risks for cancers of the respiratory system and the digestive tract, particularly stomach cancer (SIR = 0.69; 95% CI = 0.50-0.81 for hip replacement patients; SIR = 0.46; 95% CI = 0.20-0.91 for knee replacement patients). Elevated risks were observed for melanoma of the skin in both groups of patients. There was no clear excess risk for lymphohematopoietic cancers or malignant neoplasms of the bone or connective tissue among implant patients in either implant group. Contrary to an earlier study in Sweden, we did not find an excess risk for kidney or prostate cancers. In summary, these nationwide results indicate no overall cancer hazard among hip and knee implant patients, but limited follow-up warrants continued surveillance of individuals undergoing these increasingly common surgical procedures.