OBJECTIVE:In both humans and rodents, males have lower levels of leptin than females at any level of adiposity. Experimental data support the idea that testosterone exerts a negative in¯uence on leptin levels. There are, however, major inconsistencies in available data concerning the possible association between androgenicity and leptin in humans. Reasons could be the in¯uence of androgenicity on leptin production being dependent on body composition, and incomplete measures of biologically active testosterone levels. In the present study we have characterized the relationship between biologically active testosterone and leptin after careful strati®cation for gender and adiposity. DESIGN AND SUBJECTS: Healthy subjects (n 158; 85 men and 73 pre-and postmenopausal women) from the Northern Sweden MONICA (Monitoring of Trends and Determinants in Cardiovascular Disease) population were studied with a crosssectional design. MEASUREMENTS: Anthropometric measurements (body mass index (BMI) and waist circumference) and oral glucose tolerance tests were performed. Circulating levels of leptin, insulin, testosterone, androstenedione, sex hormone-binding globulin (SHBG) and insulin-like growth factor-1 (IGF-1) were measured by radioimmunoassays or microparticle enzyme immunoassays. Apparent concentrations of free testosterone and non-SHBG-bound testosterone were calculated. RESULTS: After adjustments for age, BMI and insulin, leptin levels were inversely correlated to testosterone levels in non-obese men (r À0.56, P`0.01) and obese women (r À0.48, P`0.05). In contrast, leptin and testosterone correlated in a positive manner in non-obese women (r 0.59, P`0.01). Levels of SHBG were negatively associated with leptin in men with low waist circumference (r À0.59, P`0.01). The following factors were associated with leptin in a multivariate model: low levels of biologically active testosterone and SHBG in men with low and medium waist circumference, insulin in men with high waist circumference, high levels of testosterone and insulin in non-obese women, and BMI in obese women. CONCLUSION: We conclude that low leptin levels are associated with androgenicity in non-obese men and women and that the direction of this association is dependent on gender and body fat distribution. Based on these results we suggest that the relation between testosterone and leptin contributes to the gender difference in circulating leptin levels. International Journal of Obesity (2001) 25, 98 ± 105