We studied the determinants of postprandial lipemia in 49 post-coronary-bypass men with low HDL cholesterol (< or = 1.1 mmol/l at screening). The subjects were given a mixed meal containing 63 g fat and 150,000 IU vitamin A. Serum was obtained before and 3, 4, 5, 6, and 8 h after the meal. S(f) > 400 and S(f) 12-400 lipoproteins, LDL, and HDL were separated by ultracentrifugation; and triglyceride (TG), retinyl ester (RE), and apolipoprotein (apo)E concentrations were measured. The associations of 15 potential predictor variables with measures of postprandial lipemia were evaluated in univariate and multivariate models. Fasting TG concentration was the most important determinant of postprandial lipid and apoE concentrations. In univariate analyses, neither apoE phenotype nor common genetic polymorphisms in the apoB gene (XbaI and apoB signal peptide length polymorphisms), lipoprotein lipase gene (Hind III polymorphism), or apoC-III gene (C[1100] to T sequence change) significantly predicted the magnitude of postprandial lipemia. In multivariate linear regression analyses, fasting TG concentration (P< 0.001) and postheparin plasma hepatic lipase activity (P = 0.023) were directly, and body mass index (P = 0.007) and the presence of apoE2 (P = 0.029) allele inversely related to the TG increment in S(f) >400 lipoproteins. Fasting TG was associated with a high (P < 0.001) and presence of the SP24 allele of the apoB signal peptide gene with a low (P = 0.014) S(f) 12-400 TG response. Fasting TG concentrations alone predicted 35%, 10%, and 34% of the variability in postprandial S(f) >400 responses of TG, RE, and apoE; multivariate models improved this predictive power to 40-50%. Even multivariate models were poor predictors of postprandial responses in S(f) 12-400 lipoproteins (0-26%). Much of the interindividual variation in the magnitude of postprandial lipemia remained unexplained in the present study.