Low bone density as assessed by calcaneal ultrasound has been associated with mortality in elderly men and women. We examined the relationship between bone density measured at the hip and all cause and cardiovascular mortality in elderly men. Men aged 65–76 years from the general community were recruited from general practices in Cambridge between 1991 and 1995. At baseline survey, data collection included health questionnaires, measures of anthropometry and cardiovascular risk factors, as well as bone mineral density (BMD) measured using dual energy X-ray absorptiometry. All men have been followed up for vital status up to December 1999. BMD was significantly inversely related to mortality from all causes and cardiovascular disease, with decreasing rates with increasing bone density quartile, and an approximate halving of risk between the bottom and top quartile (p <0.002, test for trend all causes and p <0.025, test for trend for cardiovascular deaths). In multivariate analyses using the Cox proportional hazards model, an increase of 1 standard deviation (0.144 g/cm2) in total hip bone density was significantly associated with an age-adjusted 0.77 relative risk (95% CI 0.66–0.91) for all-cause mortality and 0.76 relative risk (95% CI 0.62–0.93) for cardiovascular disease mortality. The association remained significant after adjusting for age, body mass index, cigarette smoking status, serum cholesterol, systolic blood pressure, past history of heart attack, stroke or cancer and other lifestyle factors which included use of alcohol, physical activity and general health status. Low bone density at the hip is thus a strong and independent predictor of all-cause and cardiovascular mortality in older men.