Risk of knee osteoarthritis (OA) was assessed in a population-based case-control study of Japanese men. The study covered three health districts in Wakayama and Osaka prefectures, Japan. Subjects were male individuals >or=45 years old diagnosed radiographically with knee OA, and who did not display any established causes of secondary OA. Controls selected randomly from the general population were individually matched to cases for age, sex, and residential district. Subjects were interviewed using structured questionnaires to determine medical history, physical activity, socio-economic factors, and occupation. Interviews were obtained from 37 cases and 37 controls. In univariate analysis, heaviest weight in the past and physical work such as factory, construction, agricultural, or fishery work as the principal occupation significantly raised the risk of male knee OA (P<0.05). Odds ratios (OR) were determined using conditional logistic regression analysis mutually adjusted for potential risk factors using the results of univariate analysis. Heaviest weight in the past (OR 6.01, 95% confidence interval (CI) 1.18-30.5, P<0.05), past knee injury (OR 6.25, 95% CI 1.13-34.5, P<0.05), and physical work as the principal occupation (OR 6.20, 95% CI 1.40-27.5, P<0.05) represented independent factors associated with knee OA after controlling for other risk factors. Physical work is associated with knee OA, demonstrating the influence of working activity on the development of OA. The present study suggests that risk factors for knee OA in men resemble those in women.