The relationship between plasma lipoprotein(a) (Lp(a)) levels and other clinical/biochemical variables was investigated in 1200 consecutive hyperlipidemic patients. Plasma Lp(a) concentrations were measured by a sandwich-ELISA method, while the patients were either on diet or diet plus lipid-lowering drugs; 38% of them had a plasma Lp(a) level > 30 mg/dl. The median plasma Lp(a) concentration and the frequency of Lp(a) > 30 mg/dl were significantly lower in individuals with severe hypertriglyceridemia vs. hypercholesterolemics (HC) or mixed hyperlipidemics (M-HLP), but similar to normolipidemic healthy controls. Patients with isolated moderate hypertriglyceridemia had Lp(a) levels intermediate between HC and M-HLP subjects. The in vitro addition of triglyceride-rich lipoproteins to normotriglyceridemic plasma did not affect the Lp(a) measurement. Plasma Lp(a) concentrations in the whole hyperlipidemic population correlated negatively with triglycerides and positively with total cholesterol, HDL-cholesterol and age, being unrelated to either body mass index or lipid-lowering treatment. In HC patients, the presence of tendon xanthomas was associated with twofold higher Lp(a) levels. These findings argue for a regulatory role of triglycerides on plasma Lp(a) levels in hyperlipidemic patients.