Postprandial lipemia test (PPLT) results are predictive of cardiovascular disease risk. However, their reproducibility must be established before they can be clinically useful. Therefore, we investigated PPLT reproducibility by testing 9 men and women (body mass index, 20-41 kg/m(2); age, 21-40 years) on 4 separate occasions (n = 36 PPLTs total) separated by 1 week. Furthermore, because PPLTs are time consuming, we assessed the validity of an abbreviated PPLT. During the PPLT, venous blood was obtained before and every hour for 8 hours after a high-fat meal, which consisted of ice cream and heavy cream (approximately 800 kcal, 71% fat calories). Total and triglyceride-rich lipoprotein (TRL) triglyceride concentrations were measured in plasma. Total area under the curve (AUC) for total triglycerides was highly reproducible (within-subject coefficient of variation, 8%; intraclass correlation coefficient, 0.82); however, reproducibility was low for total triglyceride incremental AUC and both total and incremental TRL triglyceride AUCs (within-subject coefficients of variation, 20%-31%; intraclass correlation coefficients, 0.28-0.54). Four-hour lipemic responses were highly predictive of 8-hour responses (R(2) = 0.89-0.96, P <or= .0001). In conclusion, PPLTs are highly reproducible when lipemic responses are determined as the total AUC for total triglycerides. However, large variability in incremental AUC and TRL triglyceride responses may preclude their clinical utility. Furthermore, abbreviated 4-hour PPLTs are a valid surrogate for longer tests and may make PPLTs more feasible in a clinical setting.