We evaluated the effect of dipyridamole (5 mg/kg/day) for 12 months on renal and platelet function in 53 children with insulin dependent diabetes mellitus (IDDM) in a prospective double blind placebo controlled trial. Urine albumin excretion (expressed as the geometric mean albumin to creatinine concentration ratio (UA/UC) was measured every three months throughout the study. At 12 months, the geometric mean UA/UC was no different in diabetic children receiving dipyridamole, 0.60 mg/mmol, when compared with those receiving placebo, 0.87 mg/mmol. Glomerular filtration rate, urinary excretion of retinol binding protein, and N-acetyl-beta-D-glucosaminidase (NAG), blood pressure, and spontaneous platelet aggregation in response to stirring whole blood did not differ between the two groups at 12 months. Subgroup analysis to include only those children with high UA/UC before entry into the study also failed to show an effect of the drug on UA/UC. Eleven children had either persistently high UA/UC (n = 8: four on dipyridamole, four on placebo) or progression to high UA/UC (n = 3: two on dipyridamole, one on placebo). These children had significantly higher urinary excretion of retinol binding protein and NAG, bigger kidneys, and higher diastolic blood pressure both before and after treatment than the remaining 42 children, whereas there was no difference in spontaneous platelet aggregation between the two groups. These observations on the associations between UA/UC and other parameters of renal function suggest that measurement of 'tubular' proteins and diastolic blood pressure as well as UA/UC may contribute to the identification of those at risk of developing nephropathy.