Female participants of the Framingham Osteoarthritis Study (n = 831, mean age 73, age range 63-93) were evaluated for osteoarthritis with weight-bearing radiographs of the knee during 1983-1985. At each biennial examination (1963-1981), the women were asked about their use of estrogen during the previous 2 years. We categorized estrogen use as no use reported, use reported at 1 examination, or use reported at 2 or more examinations. We found no positive association of estrogen use with radiographic knee osteoarthritis after controlling for age, body mass index, age at menopause, physical activity, history of knee injury, and smoking. In fact, a modest but nonsignificant protective effect for both radiographic osteoarthritis (odds ratio 0.71, 95% confidence interval 0.42, 1.20) and severe radiographic osteoarthritis (odds ratio 0.66, 95% confidence interval 0.33, 1.32) was seen in women who reported estrogen use at 2 or more examinations. Subgroup analyses also showed no association between estrogen use and radiographic knee osteoarthritis. We conclude that estrogen use in women is not associated with an increased risk of radiographic knee osteoarthritis.