Objective. To estimate the risk of osteoarthritis (OA) of the hip and knee due to long-term weight-bearing sports activity in ex-elite athletes and the general population. Methods. A retrospective cohort study was conducted of 81 female ex-elite athletes (67 middle- and long-distance runners, and 14 tennis players), currently ages 40-65, recruited from original playing records, and 977 age-matched female controls, taken from the age-sex register of the offices of a group general practice in Chingford, Northeast London, England. The definition of OA included radiologic changes (joint space narrowing and osteophytosis) in the hip joints, patellofemoral (PF) joints, and tibiofemoral (TF) joints. Results. Compared with controls of similar age, the ex-athletes had greater rates of radiologic OA at all sites. This association increased further after adjustment for height and weight differences, and was strongest for the presence of osteophytes at the TF joints (odds ratio [OR] 3.57, 95% confidence interval [95% CI] 1.89-6.71), at the PF joints (OR 3.50, 95% CI 1.80-6.81), narrowing at the PF joints (OR 2.97, 95% CI 1.15-7.67), femoral osteophytes (OR 2.52, 95% CI 1.01-6.26), and hip joint narrowing (OR 1.60, 95% CI 0.73-3.48), and was weakest for narrowing at the TF joints (OR 1.17, 95% CI 0.71-1.94). No clear risk factors were seen within the ex- athlete groups, although the tennis players tended to have more osteophytes at the TF joints and hip, but the runners had more PF joint disease. Within the control group, a small subgroup of 22 women who reported long-term vigorous weight-bearing exercise had risks of OA similar to those of the ex- athletes. Ex-athletes had similar rates of symptom reporting but higher pain thresholds than controls, as measured by calibrated dolorimeter. Conclusion. Weight-bearing sports activity in women is associated with a 2-3-fold increased risk of radiologic OA (particularly the presence of osteophytes) of the knees and hips. The risk was similar in ex-elite athletes and in a subgroup from the general population who reported long-term sports activity, suggesting that duration rather than frequency of training is important.