The incidence of coronary artery disease is greatly increased in those with diabetes mellitus. The largest number of those who have coronary artery disease have non-insulin-dependent diabetes (NIDDM). Lipoprotein abnormalities have been identified among the several risk factors that could account for this increase in atherosclerosis. There have been many studies demonstrating that correction of dyslipoproteinaemias will reduce the risk of coronary disease in non-diabetic populations. Current advice to those with diabetes is based on extrapolations from such studies. However, the justification for this, and the treatment targets are unclear as there has been no direct test of the lipid hypothesis in diabetes. This paper describes the protocol of the first intervention trial designed to examine directly whether correcting dyslipoproteinaemia in men and women with NIDDM will reduce their coronary artery disease. The Diabetes Atherosclerosis Intervention Study (DAIS), is a multinational angiographic study using the 200 mg micronized form of fenofibrate in a double-blind, placebo-controlled protocol. [Diabetologia (1996) 39: 1655–1661]