Numerous studies have shown lipoprotein (a) [Lp(a)] serum levels above 0.3 gL-1 to be a genetically determined and independent risk factor for atherosclerotic vascular disease. In this study of sera from 1009 patients attending our lipid clinics, multivariate regression analysis revealed an inverse correlation between the serum concentrations of triglycerides (TG) and Lp(a) (r = -0.31; P < 0.001) as determined by electroimmunodiffusion. This was not observed in 1237 controls from a random population. Detailed analysis of the frequency distribution of Lp(a) levels at different degrees of hypertriglyceridaemia (HTG) revealed a decreasing dosage effect of HTG on Lp(a) serum levels. In 60% of patients with TG > 9.12 mmol L-1, this effect led to undetectable serum Lp(a) levels. Dilution of hypertriglyceridaemic samples with normotriglyceridaemic sera containing high levels of Lp(a) revealed that analytical interference in part accounts for the decreasing effect of TG-rich lipoproteins on Lp(a). Re-evaluation of 45 hypertriglyceridaemic samples by enzyme immunoassay and immunoblotting revealed the presence of small amounts of Lp(a) in several samples which were considered to be free of Lp(a) upon electroimmunodiffusion. We conclude that TG-rich lipoproteins interfere with the quantification of Lp(a), at least by electroimmunodiffusion. However, HTG may also decrease Lp(a) plasma concentrations in vivo, possibly by increased clearance of TG-rich Lp(a).