Epidemiological studies have suggested that cardiovascular disease can be decreased by moderate wine consumption, but an overall quantitative estimation of the relationship between wine intake and vascular risk is lacking. A meta-analysis was therefore performed on 19 studies selected on the basis of the availability of specific information on the cardiovascular relative risk (RR) associated with wine consumption. A significant risk reduction (RR: 0.66, 95% CI 0.57-0.75) was associated with moderate (1-2 drinks or 150-300 mL/d) versus no wine consumption. In five studies which excluded ex-drinkers as reference group, the overall RR associated with wine consumption was 0.61 (95% CI 0.57-0.75). A dose-response relation between wine intake and vascular risk resulted in a J-shaped curve, with a significant risk reduction at about 300 mL/d (trend analysis p = 0.032). Two studies were also performed to investigate the effects of wine polyphenols on experimental thrombosis in rats. Supplementation for 10 days with alcohol-free red wine--but not white wine or alcohol--induced a significant reduction of stasis-induced venous thrombosis, an effect blunted by NO synthase inhibitor L-NAME. In rats with diet-induced hyperlipidemia, alcohol-free red wine supplementation significantly delayed the thrombotic occlusion of an artificial prosthesis inserted into the abdominal aorta, but did not affect the increased cholesterol and triglyceride levels. TRAP values were significantly higher in animals receiving alcohol-free wine. Altogether these experimental data support an antithrombotic role of polyphenols in the reduced vascular risk associated with moderate wine consumption in man, as shown by our epidemiological studies.