A placebo-controlled crossover study was conducted to evaluate whether lipid-lowering with gemfibrozil (10 to 12 weeks) affects platelet function in vivo at rest and during mental stress in 21 men with combined hyperlipoproteinemia. Gemfibrozil lowered plasma triglycerides and total and VLDL cholesterol (P < .001 for all), whereas HDL cholesterol increased (P < .001). Gemfibrozil increased platelet counts by 18% (P < .001) and, according to filtragometry measurements, enhanced overall (means of rest and stress values) platelet aggregability in vivo (P = .014); this effect was more evident during mental stress (P = .027) than at rest (P = .18). Moreover, the urinary excretion of 11-dehydro-thromboxane-B2 was 54% higher during treatment with gemfibrozil (P < .001), whereas the excretion of beta-thromboglobulin in urine was unaltered. Plasma beta-thromboglobulin tended to be slightly elevated during active treatment (P = .15). Mental stress increased heart rate and catecholamine levels and elevated all cholesterol fractions (P < .01) and plasma beta-thromboglobulin (during placebo, P = .02), but platelet aggregability did not increase significantly. A positive correlation was found between 11-dehydrothromboxane-B2 excretion and LDL cholesterol levels during placebo (r = .62, P = .0057). In conclusion, gemfibrozil has beneficial effects on plasma lipoprotein levels but enhances several aspects of platelet activity in vivo and increases platelet counts. These changes might be prothrombotic and thus adverse for the hyperlipidemic patient. Primary preventive effects of gemfibrozil might be enhanced if a platelet inhibitor such as aspirin is administered with gemfibrozil.