To assess whether leptin is an independent predictor of bone mineral density (BMD) in postmenopausal women, we studied the relationships of BMD to serum leptin, 25(OH)D, 1,25(OH)(2)D, PTH, E2, dehydroepiandrosterone sulfate, GH, IGF-I, creatinine clearance, calcium intake, fat mass, and lean mass in 107 women aged 50-90 yr. We also related serum leptin to markers of bone formation [serum bone alkaline phosphatase and osteocalcin (OC)] and resorption (urine C-telopeptide of type I collagen). In stepwise multiple linear regression, lean mass explained 28.5%, age 10.3%, and leptin 7.2% of the whole body BMD variance. Age explained 21.1%, lean mass 12.8%, and leptin 3.7% of the femoral neck BMD variance. After adjustment for fat mass and creatinine clearance, correlations between leptin and bone alkaline phosphatase (positive) and OC (negative) disappeared but, remained significant with urine C-telopeptide of type I collagen (r = -0.27, P < 0.01). Markers of bone formation and resorption were strongly intercorrelated. These data demonstrate that leptin is an independent predictor of whole body and femoral neck BMD in postmenopausal women. Although the relationships between leptin and markers of bone formation appear complex, leptin may exert a protective effect on bone by limiting the excessive bone resorption coupled to bone formation associated with bone loss after menopause.