Objective: To study relationships between leptin and factors regulating body composition as well as metabolic risk factors. Furthermore, to study the effects of GH on leptin., Design: Cross-sectional and population-based. Regarding the effects of GH, prospective and interventional., Patients: One hundred and eleven women and 107 men, 20-70 years old, randomly selected from the population registry in the community of Linköping, Sweden. Ten GH-deficient subjects were given GH until normalization of IGF-I levels., Measurements: Venous blood was drawn in the fasting state. Serum leptin and hormones were analysed by immunoassay., Results: In the population sample the natural logarithm of leptin (in(leptin)) correlated with body mass index (BMI) (men, r = 0.67), P < 0.0001; women, r = 0.71, P < 0.0001). The median value of leptin was 4.6 micrograms/l in men and 12.3 micrograms/l in women (P < 0.0001). Levels of in(leptin) did not correlate with plasma neuropeptide Y (men, P = 0.13; women, P = 0.35). In men only there was an inverse relationship between in(leptin) and testosterone (r = -0.46, P < 0.0001, after correction for BMI standardized r = -0.26, P = 0.03) as well as IGF-I (r = -0.20, P = 0.048). Although BMI was similar, smoking men had higher leptin levels than non-smoking men (median, 6.6 and 4.2 micrograms/l, respectively; Mann-Whitney; P = 0.006). In the GH-deficient subjects leptin levels were elevated and, although GH treatment did not change BMI, leptin levels decreased (median before GH, 21 micrograms/l and after 15 micrograms/l, respectively; P = 0.017)., Conclusion: Serum leptin concentration is closely associated with BMI in the population with a gender difference in absolute levels and a strong negative correlation with testosterone in men. Serum leptin is elevated in GH deficiency and lowered by GH substitution.