Background Atherosclerosis is the most common complication of diabetes. Correction of hyperglycaemia helps to prevent microvascular complications but has little effect on macrovascular disease. Post-hoc analyses of diabetic subpopulations in lipid intervention trials suggest that correction of lipoprotein abnormalities will lead to a decrease in coronary-artery disease. The Diabetes Atherosclerosis Intervention Study (DAIS) was specifically designed to assess the effects of correcting lipoprotein abnormalities on coronary atherosclerosis in type 2 diabetes. Methods 731 men and women with type 2 diabetes were screened by metabolic and angiographic criteria. 418 were randomly assigned micronised fenofibrate (200 mg/day) or placebo for at least 3 years. They were in good glycaemic control (mean haemoglobin A(10) 7.5%), had mild lipoprotein abnormalities, typical of type 2 diabetes, and at least one visible coronary lesion. Half had no previous clinical coronary disease. Initial and final angiograms followed a standard protocol and were analysed by a computer-assisted quantitative approach. Missing data for the primary endpoints (minimum lumen diameter, mean segment diameter, and mean percentage stenosis) were imputed. Analyses were by intention to treat. Findings Total plasma cholesterol, HDL-cholesterol. LDL-cholesterol, and triglyceride concentrations all changed significantly more from baseline in the fenofibrate group (n=207) than in the placebo group (n=211). The fenofibrate group showed a significantly smaller increase in percentage diameter stenosis than the placebo group (mean 2.11 [SE 0.594] vs 3.65 [0.608]%, p=0.02), a significantly smaller decrease in minimum lumen diameter (-0.06 [0.016] vs -0.10 [0.016] mm, p=0.029), and a non-significantly smaller decrease in mean segment diameter (-0.06 [0.017] vs -0.08 [0.018] mm, p=0.171). The trial was not powered to examine clinical endpoints, but there were fewer in the fenofibrate group than the placebo group (38 vs 50). Interpretation DAIS suggests that treatment with fenofibrate reduces the angiographic progression of coronary-artery disease in type 2 diabetes. This effect is related, at least partly, to the correction of lipoprotein abnormalities, even those previously judged not to need treatment.