UNLABELLED: In this large prospective cohort study of elderly women, the relationships between prior wrist fracture and incident hip and radiographic vertebral fractures were significantly attenuated when adjusted for BMD. This study suggests that BMD thresholds for drug therapy to prevent osteoporotic fracture should be only modestly adjusted in those with prior wrist fracture compared with those without prior wrist fracture. Validation of such an approach would require intervention trials in patients with prior wrist fracture. INTRODUCTION: Prior wrist fracture has been identified as a risk factor for incident hip and vertebral fractures and proposed as a criterion for determining who should be offered drug therapy to prevent osteoporotic fracture, even if their hip BMD T score is > -2.5. Previously published studies of the relationships between prior wrist fracture and incident hip and vertebral fractures did not adjust for BMD. MATERIALS AND METHODS: We ascertained prior history of wrist fracture since age 50, measured calcaneal and hip BMD, and performed lateral spine films in a cohort of 9704 elderly community-dwelling women, and then followed them prospectively for incident vertebral and hip fractures. Incident vertebral fractures were defined by morphometry using lateral spine radiography at the first examination and an average of 3.7 years later. Incident hip fractures were confirmed with radiographic reports over a mean follow-up period of 10.1 years. RESULTS: Prior wrist fracture was associated with an age-adjusted 72% increased odds of incident radiographic vertebral fracture (odds ratio [OR], 1.72; 95% CI, 1.31-2.25). After adjustment for calcaneal BMD, the association of prior wrist fracture with incident radiographic vertebral fracture was attenuated (OR, 1.39; 95% CI, 1.05-1.83). Prior wrist fracture was also associated with an age-adjusted 43% excess rate of incident hip fracture (hazards ratio [HR], 1.43; 95% CI, 1.17-1.74). After adjustment for hip BMD, the association of prior wrist fracture with rate of incident hip fracture was no longer statistically significant (HR, 1.12; 95% CI, 0.92-1.38). CONCLUSION: In elderly women, prior wrist fracture is a risk factor for radiographic vertebral fracture independent of BMD. The association between prior wrist fracture and incident hip fracture is largely explained by hip BMD. Modest adjustment of BMD drug treatment thresholds for prevention of osteoporotic fractures in those with prior wrist fracture compared with those without prior wrist fracture may be reasonable, but validation of such an approach would require intervention trials in patients with prior wrist fracture.