To examine the hypothesis that hyperinsulinaemia promotes atherosclerosis, cholesterol-fed rabbits were injected subcutaneously with 6 IU of human insulin (n = 16) or placebo (n = 20) daily for 24 weeks; injection of insulin resulted in hyperinsulinaemia for up to 16 h after injection. Compared to placebo rabbits, insulin-treated rabbits had higher levels of insulin antibodies in plasma, similar levels of intermediate density, low density and high density lipoprotein cholesterol and similar activities of hepatic and lipoprotein lipase in post-heparin plasma, but lower levels of plasma C-peptide, blood glucose, postprandial plasma triglycerides, plasma cholesterol and very low density lipoprotein cholesterol. On univariate analysis, with and without adjustment for differences in plasma cholesterol levels between the two groups, there were no significant differences in extent or severity of atherosclerosis between insulin and placebo rabbits. Furthermore, after combining the results from all the rabbits to examine plasma insulin levels and the other variables mentioned above as predictors of atherosclerosis severity, plasma insulin level was not a predictor, on univariate or multiple linear regression analysis; the first ranked independent predictors were postprandial intermediate density lipoprotein cholesterol in the arch, and postprandial plasma triglyceride in both the thoracic and abdominal aorta. These results suggest that exogenous hyperinsulinaemia does not promote atherogenesis in cholesterol-fed rabbits, but that postprandial levels of intermediate density lipoprotein cholesterol or plasma triglycerides may be involved in atherogenesis. [Diabetologia (1997) 40: 512–520]