Objective: In recent studies serum leptin levels were signi®cantly higher in the luteal phase than in the follicular phase, but the mechanism of changing leptin levels are unknown. Several research lines indicate a potential role for tumor necrosis factor (TNF-a) in ovulation and reproductive events. As TNF-a appears to regulate leptin secretion, we speculated that TNF-a might be involved in leptin variations during the menstrual cycle. Design and methods: Nine healthy never obese and ten overweight normally cycling women were studied. TNF-a action ± through the plasma levels of the soluble fraction of the tumor necrosis factor receptors 1 and 2 (sTNFR1 and sTNFR2) ± and leptin concentrations were measured in the follicular (F), peri-ovulatory (PO) and luteal phases (L) of their menstrual cycles. Results: Circulating leptin levels were signi®cantly associated with the stage of the menstrual cycle (P<0.001), being higher in PO and L phases. However, only three of ten overweight subjects vs eight of nine lean women (Chi square P 0.014 after Fisher's exact test) showed signi®cantly higher leptin levels in the PO and L than in the F phase (95% con®dence interval (95% CI) of the differences, 3.7 to 10.2 ng/ml, paired t-test P 0.001). In these women (group 1), the changes in leptin levels parallelled the variations observed in plasma sTNFR1 (2.50 6 0.1 vs 2.11 6 0.05 ng/ml, P < 0.0001, 95% CI, 0.21 to 0.56) and sTNFR2 levels (5.19 6 0.28 vs 4.55 6 0.25 ng/ml, P < 0.0001, 95% CI, 0.47 to 0.81). In the remaining women (group 2), leptin (95% CI, ±1 to 9.2 ng/ml, P not signi®cant (NS)), sTNFR1 (95% CI, ±0.3 to 0.14 ng/ml, P NS) and sTNFR2 levels (95% CI, ±0.95 to 0.39 ng/ml, P NS) were essentially unaltered throughout the menstrual cycle. Group 2 women were similar in age (36.1 6 2.9 vs 37.3 6 1.4 years) and signi®cantly overweight (body mass index 31 6 2.9 vs 23.9 6 1.2 kg/m 2 ) compared with group 1 women. A negative correlation was observed between leptin levels in the follicular phase and the change in plasma leptin from F to L phase in all subjects (r À 0.67, P 0.002). Conclusions: Circulating leptin and sTNFRs levels change signi®cantly during the menstrual cycle of most lean women. In contrast, the levels of these molecules remain essentially unaltered during the F, PO and L phases in the majority of overweight women. Obesity might be associated not only with blunted diurnal excursions and dampened pulsatility, but also with blunted excursions during the menstrual cycle.