Results of cross-sectional analyses of the association of kidney function with bone mineral density (BMD) have been conflicting. We examined the association of cystatin-C, a new marker of kidney function that is unrelated to lean mass, with initial and follow-up BMD, in an ancillary study of the Cardiovascular Health Study, a population-based cohort of individuals > or = 65 years old. Two years after measurement of cystatin-C and other covariates, the first BMD was measured in Pittsburgh, Pennsylvania and Davis, California, by using dual energy x-ray absorptiometry. Follow-up BMD was measured in Pittsburgh 4 years later. Associations of cystatin-C with initial BMD and the change in BMD (%/y) at the hip were examined with linear regression. Analyses were conducted separately for men and women. In 1519 participants who had cystatin-C and initial BMD assessed, 614 had follow-up BMD. The percent annual change in BMD at the total hip by cystatin-C quartiles was -0.24, -0.13, -0.40, and -0.66%/y (first to fourth quartile) in women and -0.02, -0.30, -0.18, and -0.94%/y in men. After adjusting for potential confounders, cystatin-C was marginally associated with initial BMD in men but not women. Cystatin-C was associated with bone loss in men; after adjustment for weight loss, cystatin-C was not associated with bone loss in women. Kidney dysfunction, as assessed by cystatin-C, is associated with a more rapid loss of BMD at the hip, especially in men. Further studies are needed to confirm these findings and to determine whether this loss leads to an elevated risk of fracture.