Elevated plasma triglyceride has been found to be a univariate risk factor for coronary heart disease in many epidemiologic studies, but this increased risk is not independent of other lipid and lipoprotein measures. Rather than dismiss triglyceride as a risk factor, we can see this disparity between univariate and multivariate results as a clue to better understanding of the potential role of triglyceride in atherosclerosis. From a statistical point of view, triglyceride often is not a significant predictor of coronary heart disease in multivariate statistical models because of the large variation in triglyceride measurements and the strong inverse relation between HDL cholesterol and triglyceride. Heterogeneity in both the biologic mechanisms producing hypertriglyceridemia and the manifestations of cardiovascular disease complicates the study of triglyceride as a risk factor. recently, subclasses of both HDL and LDL have been shown to be related to triglyceride metabolism and may provide more precise variables for predicting risk. However, none of this evidence precludes the involvement of triglyceride in the development of atherosclerosis. The final answer as to whether triglyceride is a causal risk factor for coronary heart disease must come from the biologic sciences, and the complexity of the epidemiologic results likely reflects the intricate metabolic processes involved.