Objective: (1) To compare the effect of an alcohol-free Mediterranean-type diet (MD) and a high-fat diet (HFD) on variables of primary haemostasis (bleeding time, plasma von Willebrand factor and platelet aggregation=secretion). (2) To test whether red wine supplementation modified these variables, independently of the diet. Design, subjects and intervention: Controlled prospective intervention study. Two groups, each consisting of 21 healthy male university students (22 AE 3.4 y), received either MD or HFD during 90 days. Between days 30 and 60, both diets were supplemented with 240 ml=day of red wine. Baseline (T0) and T30, T60 and T90-day samples were drawn. Bleeding time was measured before (day 30) and after (day 60) wine supplementation. No drop out from the study was experienced. Setting: University campus and outpatient nutrition clinic. Results: All baseline (day 0) variables did not differ significantly between study groups. On day 30, individuals on MD had significantly higher levels of plasma b-carotene, folate, ascorbate, and eicosapentaenoic acid in plasma lipid fractions, than those on HFD. Total plasma cholesterol, HDL and LDL did not change significantly in either study group at any time point. After 30 days on each diet, individuals on MD had longer bleeding time (BT) than those on HFD (7.6 AE 2.8 vs 5.8 AE 1.7 min; P ¼ 0.017). BT did not change significantly after I month of wine supplementation (7.1 AE 2.0 vs 5.5 AE 2.0 min, respectively). Plasma von Willebrand factor (vWF : Ag) on day 0 was 89 AE 40 and 111 AE 70% in MD and HFD groups, respectively (P ¼ 0.21). These values did not change significantly at 30, 60 or 90 days. MD intake was associated with an increase in platelet serotonin secretion (P ¼ 0.02) and a marginal increase in platelet aggregation after stimulation with epinephrine (P ¼ 0.07). Wine intake resulted in a marginal decrease in platelet 14 C-5-HT secretion with 4 mM ADP (P ¼ 0.07). However, both platelet aggregation and secretion were consistently increased when using collagen as agonist (1 and 2 mg=ml, P ¼ 0.01). Conclusion: The longer BT in individuals on MD, obtained independently of red wine, denotes less interaction of platelets with the vascular wall, which could be beneficial from the point of view of cardiovascular (CV) risk. This effect is not explained by changes in the measured haemostatic determinants of BT (plasma vWF, ex vivo platelet function), and might be attributed to other as yet unknown vascular factors. Moderate consumption of red wine results in a significant increase in ex vivo platelet aggregation and secretion after stimulation with collagen. This observation contradicts previous reports, although further studies are required to elucidate the influence of this finding on CV risk.