BACKGROUND AND AIM: The relationship between hyperlipidemia and blood and plasma viscosity is not completely clear. While increasing viscosity is often reported with increasing blood lipids, lipid-lowering treatments are often unable to normalize the viscosity values. Aim of this study is to try to clarify the relationship between blood lipids and viscosity. METHODS AND RESULTS: Apparently healthy subjects were enrolled (n = 410). Smokers, diabetics, obese, and hypertriglyceridemic (above 400 mg/dl) were excluded. Blood (at shear rate 225/s) and plasma viscosity were measured at 37°C. Erythrocyte rigidity (Tk) was calculated according to Dintenfass. Blood lipids and glucose were measured by routine methods. Hyperlipidemic subjects (n = 315) had higher values of plasma viscosity (1.44 ± 0.13 vs. 1.40 ± 0.12 cP, p = 0.007), and blood viscosity (4.51 ± 0.54 vs. 4.35 ± 0.55 cP, p = 0.013), compared to normolipidemic subjects (n = 95). In simple correlation analysis, plasma viscosity was directly associated with LDL cholesterol, and inversely with Tk and HDL cholesterol. In multiple regression analysis the association with LDL and HDL was strengthened, though these two variables as a whole accounted for only 5% (adjusted R2) of the variability of plasma viscosity. Blood viscosity was significantly associated with haematocrit, plasma viscosity, Tk and all considered variables but age in simple correlation analysis, but only with haematocrit, plasma viscosity and Tk in multiple regression analysis. CONCLUSIONS: LDL cholesterol and HDL cholesterol influence plasma viscosity, but not blood viscosity. Triglycerides up to values of 400 mg/dl do not seem to have important effects, at least in apparently healthy subjects and at the shear rates used in the present study. The contribution of LDL and HDL cholesterol to plasma viscosity seems however quite limited.