Rates of elevated urinary albumin concentration, defined as microalbuminuria (30-299 μg/ml) and macroalbuminuria (≥ 300 μg/ml), were determined on random morning urine specimens in the population of Nauru, which has a high prevalence of non-insulin-dependent diabetes mellitus. The prevalence of elevated urinary albumin levels in the total Nauruan population was very high: 26 and 30% of men and women, respectively, had microalbuminuria, whereas 13% of both sexes had macroalbuminuria. Of the subjects with macroalbuminuria, 66% had diabetes. The prevalence increased with worsening glucose tolerance; 26% of subjects with normal glucose tolerance had either micro- or macroalbuminuria, increasing to 43% of subjects with impaired glucose tolerance, 63% of newly diagnosed diabetic subjects and 75% of previously diagnosed diabetic subjects. Associations between elevated urinary albumin concentration and putative risk factors were assessed for both the total population (n = 1184) and the diabetic subgroup alone (n = 318). Fasting plasma glucose and hypertension were the most important independent correlates for the whole population, whereas plasma creatinine was also important in diabetic subjects. Age at onset and duration of diabetes were not found to be significantly associated with elevated albumin concentration. In subjects with normal glucose tolerance, hypertension and hyperuricemia were the most important associated factors. These results suggest that blood glucose, blood pressure, and possibly obesity and plasma uric acid are important modifiable risk factors for both micro- and macroalbuminuria in this population.