OBJECTIVES--To test the hypothesis that specific occupational physical activities are risk factors for knee osteoarthritis (OA). METHODS--A population-based case-control study of knee osteoarthritis was carried out in which 109 men and women with painful, radiographically confirmed knee OA were compared with 218 age and sex matched controls who had not suffered knee pain and had normal radiographs. Information collected included a lifetime occupational history and details of specific workplace physical activities. RESULTS--After adjustment for obesity and Heberden's nodes, the risk of knee OA was significantly elevated in subjects whose main job entailed more than 30 minutes per day squatting (OR 6.9, 95% CI 1.8-26.4) or kneeling (OR 3.4, 95% CI 1.3-9.1), or climbing more than ten flights of stairs per day (OR 2.7, 95% CI 1.2-6.1). The increase in risk associated with kneeling or squatting appeared to be more marked in subjects whose jobs entailed heavy lifting, but the size of the study did not permit precise delineation of any such interaction. CONCLUSIONS--These data suggest that prolonged or repeated knee bending is a risk factor for knee OA, and that risk may be higher in jobs which entail both knee bending and mechanical loading.