14 national groups have collaborated under WHO auspices to select, from local defined populations of individuals with clinical diabetes, groups of approximately 500 within the age range 35--55 yr stratified by age, sex, and known duration of diabetes. In each center, the selected patients were submitted to a standardized study protocol, which included systematic inquiry (WHO questionnaire) for the presence of symptoms of angina pectoris, history of myocardial infarction, presence of intermittent claudication, and cigarette smoking history. Examination included standard biometry, blood pressure measurement, 12-lead (centrally Minnesota coded) electrocardiography, and central laboratory measurement of serum cholesterol and creatine. Ophthalmoscopic and urinary examinations were also included. The prevalence of arterial disease symptoms and electrocardiographic abnormalities show very large variation between countries, the lowest rates generally being found in the Oriental samples and the highest in the European. "Risk factors" for arterial disease (blood pressure, serum cholesterol, and cigarette smoking) also vary widely between diabetic groups. Although data are not yet complete, these differences appear unlikely to explain the variation in the atherosclerotic morbidity observed. Diabetic women were at least as vulnerable to arterial disease as diabetic men. A high prevalence of nonspecific abnormalities of the repolarization phase of the ECG was found, even in groups where ischemic abnormalities were rare. The origin of these is uncertain; they may represent variable local changes or possibly diabetic cardiomyopathy. This preliminary report confirms and quantifies previous indications that the impact of atherosclerotic disease on persons with diabetes varies considerably between national groups, in broad terms, running parallel with the variations in prevalence in the populations in general and suggesting that cultural and/or ethnic factors are more important determinants of atherosclerosis in diabetic individuals than is the diabetic state per se.