AIM: To evaluate whether fat distribution plays a role in determining serum leptin concentrations. PATIENTS AND METHODS: One-hundred and forty-seven obese patients, 77 males and 70 females, aged 45.1 AE 13.2 y (mean AE s.d.; range 21 ± 73 y), with body mass index (BMI) ranging from 30 to 55 kgam 2 (mean 42.3 AE 5.9). Ultrasound assessment of the thickness of subcutaneous and preperitoneal fat was carried out and calculation of their ratio as abdominal fat index (AFI), waist ± hip ratio (WHR), body composition by bioelectrical impedance to evaluate the percentage of fat mass (FM%) and total amount of fat (FMKg) were also determined. Plasma leptin was measured by radio immuno assay (RIA). RESULTS: In the whole group of patients, serum leptin concentrations were 37.2 AE 18.4 ngaml (range 6 ± 101.3 ngaml); in spite of BMI values not being signi®cantly different, women had leptin values signi®cantly higher (47.4 AE 17.4 ngaml) (P`0.01) than males (28.1 AE 15.1 ngaml), also after correction for fat mass. The mean thickness of abdominal subcutaneous fat was 33.7 AE 12.9 mm and it was signi®cantly (P`0.001) higher in female (40.9 AE 10.6 mm) than in male (27.1 AE 11.2 mm) patients; preperitoneal thickness was 22.9 AE 7.1 mm, with signi®cantly (P`0.05) higher values in males (24.2 AE 6.8 mm) than in females (21.7 AE 7.3 mm). Accordingly, AFI (in all patients 0.84 AE 0.6) was signi®cantly higher in males (1.09 AE 0.6) than in females (0.56 AE 0.2). In the overall population, leptin concentrations were directly and signi®cantly related to subcutaneous but not preperitoneal fat; they showed a strong inverse relationship with AFI and WHR. When the results were evaluated dividing the patients according to gender, subcutaneous fat thickness showed a stronger association with leptin levels in males than in females, whereas no association was found with preperitoneal fat thickness. Leptin and AFI values were signi®cantly related only in men. WHR values were not correlated with leptin concentrations in either sex. When fat mass was added to the model, subcutaneous fat thickness, AFI and WHR remained independently associated with leptin concentrations. Age and diabetes did not in¯uence these measures. CONCLUSIONS: Fat distribution contributes to the variability in serum leptin in obese patients. In particular, subcutaneous abdominal fat is a determinant of leptin concentration, also independently of the amount of fat mass, whereas the contribution of preperitoneal visceral fat is not signi®cant.