The effect of gemfibrozil on postprandial lipoprotein metabolism was investigated in a 12-week, randomized, double-blind, placebo-controlled trial in 20 non-insulin-dependent diabetic patients with moderate hypertriglyceridemia. The patients were given a meal containing 78 g of fat and 345,000 units of vitamin A to label chylomicrons and their remnants. Plasma obtained at various times during the fat-load test was separated into six fractions by gradient-density ultracentrifugation. Gemfibrozil reduced the postprandial triglyceride response, measured as the area under the time-dependent concentration curve, on average by 32% in whole plasma, by 38% in the Svedberg flotation unit (Sf) 1,100-3,200 chylomicron fraction, by 36% in Sf 400-1,100 chylomicrons, and by 38% in the Sf 60-400 lipoproteins. Retinyl palmitate, a measure of intestinally derived particles, was reduced in plasma by 34%, in Sf 1,100-3,200 by 46%, in Sf 400-1,100 by 44%, and in Sf 60-400 by 37%. All these reductions were significant in comparison with the placebo group. Particles with Sf < 60 were not significantly affected. In contrast to earlier observations in healthy subjects, no significant negative correlations existed between postprandial lipemia and high density lipoprotein cholesterol or the postheparin lipoprotein lipase activity. The reduction of the potentially atherogenic chylomicron remnants may decrease the risk of atherosclerosis in non-insulin-dependent diabetes mellitus, a hypothesis that awaits testing in prospective studies.