Long-term prediction of incident hip fracture risk in elderly white women: Study of osteoporotic fractures

Objectives: To identify independent risk factors for first hip fracture over 10 years of follow-up. Design: Prospective cohort study. Setting: Four U.S. clinical centers. Participants: A total of 6,787 women aged 66 and older in the Study of Osteoporotic Fractures. Measurements: Total hip bone mineral density (BMD) using dual-energy x-ray absorptiometry and a comprehensive set of potential risk factors were collected. Incident hip fractures were identified prospectively and confirmed using radiographic report. Results: Six hundred two women (8.9%) had a hip fracture during a mean +/-standard deviation (SD) follow-up of 10.1+/-3.2 years. Older age, previous self-reported fracture after age 50, maternal history of hip fracture after age 50, greater height at age 25, impaired cognition, slower walking speed, nulliparity, type II diabetes mellitus, Parkinson's disease, and depth perception each independently predicted a 1.17- to 1.83-fold increase in hip fracture risk, whereas each SD (0.13 g/cm(2)) decrease in hip BMD was independently associated with a 1.84-fold increase in risk. Lower body mass index also was associated with an increased risk of hip fracture, although lower hip BMD largely explained this association. Conclusion: Although hip BMD is strongly related to hip fracture risk in elderly white women, other clinical risk factors also are independent predictors of long-term risk and provide additional insight into the prevention of fracture in high-risk women. Clinicians should be alert to factors other than BMD that place older women at a high risk of hip fracture.