Obesity is a cardiovascular risk factor associated with endothelial dysfunction, but the effect of different weight loss strategies on endothelial function is not known. The effect of diet on endothelial function in two hypocaloric diets, a very-low-carbohydrate diet (A) and a Mediterranean diet (M), was measured by brachial artery flow-mediated dilation (FMD). Design Using a longitudinal, randomized, open study design, subjects were engaged in a 2-month weight loss diet. FMD, inflammatory cytokines [interleukin-6 (IL-6) and tumour necrosis factor-alpha] and a marker of oxidative stress [8-iso-prostaglandin F2alpha (8-iso-PGF2alpha)] were measured in subjects on three occasions: before initiating the diet (T0), after 5-7 days of dieting (T5) and after 2 months of dieting (T60). The very short- and medium-term time points were established to discriminate respectively the effect of the diet itself (T5) from that of weight loss (T60). Twenty overweight/obese but otherwise healthy women (BMI: 27-34.9 kg m(-2); age 30-50 years) completed the study. Results Group A lost more weight (mean +/- SEM; -7.6 +/- 0.8 kg) than group M (-4.9 +/- 0.6 kg, P = 0.014) at T60. The FMD was not significantly different between the two groups at T0 (group A: 12.2 +/- 2.9% vs. group B: 10.3 +/- 2.3%, P = ns). In group A, FMD was significantly reduced at T5 and returned to baseline at T60; in group M, FMD increased at T5 and returned to baseline at T60 (P = 0.007 for diet x time interaction). Serum concentrations of IL-6 and 8-iso-PGF2alpha were not significantly different between the two groups at T0 and increased significantly at T5 only in group A (P < 0.001 and P < 0.005 respectively). Conclusion As endothelial dysfunction is known to be associated with acute cardiovascular events, this study suggests that the cardiovascular risk might be increased in the first days of a very-low-carbohydrate diet.