Incidence and risk factors for radiographic knee osteoarthritis in middle-aged women - The Chingford Study

Objective. To examine the natural history, role of risk factors, and incide nce of knee osteoarthritis (OA) in a prospective study of women from a popu lation cohort. Methods. Women from the Chingford Study who had been recruited in 1989 were followed up with knee radiographs 4 years later, A total of 715 paired rad iographs (71% of the original sample) were graded for osteophytes and 644 f or joint space narrowing (JSN). Women whose radiographs had been graded as 0 in 1989 and as greater than or equal to 14 years later were classified as having incident disease. Incident cases were compared with controls for as sociations,vith a number of risk factors. Results, Eighty-one women (12.6%) developed JSN of the knee, equating to an incidence of 3.1% per year. No clear risk factors for JSN were identified. Reproducibility of measures of joint space is poor, however, leading to in accuracy of definition. Incident knee osteophytes developed in 95 women (13 .3%), equating to an incidence of 3.3% per year. Compared with controls, wo men with incident knee osteophytes were older, heavier, and had more hand O A and knee symptoms. Women in the top tertile of obesity (body mass index > 26.4) had a significantly increased risk of incident knee osteophytes (odds ratio [OR] 2.38, 95% confidence interval [95% CI] 1.29-4.39), Incident kne e osteophytes increased by 20% per 5-year age increase. A nonsignificant pr otective effect for incident knee osteophytes was seen with current estroge n replacement therapy (ERT) (OR 0.41, 95% CI 0.12-1.42), No effect was asso ciated with smoking, physical activity, hysterectomy, or previous knee inju ry, Conclusion. Obesity and aging are associated with a high risk of new knee O A developing in women. Evidence of a protective effect of ERT was seen. No clear association,vas found for incident JSN, suggesting that different eti ologic mechanisms are operating or that standard radiographs are an inaccur ate measure of incident narrowing.