Whether the higher serum leptin levels in women are due to gender differences in fat mass or to other factors such as sex steroids remains unclear. In addition to sex steroids, serum insulin levels also appear to be related to leptin levels, although whether this effect is independent of the effects of body composition is unclear. The purpose of this study was to identify the major determinants of circulating serum leptin levels. We studied a large, population-based cohort of 345 men (23 to 90 years), 137 premenopausal women (21 to 54 years), and 212 postmenopausal women (34 to 94 years), including 47 women on hormone replacement therapy (HRT). Serum leptin levels were related to body composition as assessed by dual-energy x-ray absorptiometry (DEXA) and to circulating sex steroid and insulin levels. Serum leptin levels remained significantly higher in women versus men even after adjustment for fat mass, and leptin levels were significantly correlated with fat mass independently of age. By univariate analyses, logarithmically transformed serum leptin levels correlated positively with bioavailable estrogen ([E] estradiol plus estrone) in postmenopausal women not on HRT, and negatively with total and bioavailable testosterone (T) levels in men. Serum insulin levels were directly related to leptin levels regardless of gender and age. By multivariate analyses, fat mass, lean mass, and insulin levels were the strongest predictors of leptin levels in all groups. In addition, bioavailable E entered the model in the postmenopausal women not on HRT. These studies indicate that the fat mass, lean mass, and insulin level are the major determinants of the serum leptin level in adults. Moreover, after adjusting for these variables, bioavailable E also explains a significant proportion of the variance in leptin levels among postmenopausal women not on HRT.