A cross-sectional study of 1084 Caucasoid diabetic subjects in rural Western Australia revealed a high rate of clinical macrovascular disease (46%), including coronary heart disease (13%), stroke (8%), and peripheral vascular disease (38%). Age was the major time-related variable for total macrovascular disease and for peripheral vascular disease, with identical prevalence rates in Type 1 (insulin-dependent) and Type 2 (non-insulin-dependent) diabetes when age was taken into account. In 179 Type 1 diabetic subjects, logistic regression analysis showed no associated risk factors other than age. In 905 Type 2 diabetic subjects the independent risk factors for total macrovascular disease, identified by a forward step-wise selection procedure, were age as the major contributor, with plasma creatinine levels and plasma glucose levels (all p less than 0.001), high-density lipoprotein cholesterol levels, serum total cholesterol levels, and the (supine-erect) systolic blood pressure difference (all p less than 0.05). There were no direct associations with percentage desirable weight, cigarette smoking or male sex. Type 2 diabetic subjects demonstrated a very strong negative association between high-density lipoprotein cholesterol levels and coronary heart disease, and significant associations were found also between plasma glucose levels and coronary heart disease (p less than 0.01), and glycosylated haemoglobin levels and peripheral vascular disease (p less than 0.001).