The aim of this study was to clarify whether early glomerular hyperfiltration, a characteristic feature of insulin-dependent diabetes, is associated with late diabetic nephropathy. In 1984 we re-examined 12 patients previously investigated in our laboratory around 1970; entrance criteria were as follows: male diabetics, clearly insulin-dependent, and age at onset of diabetes less than or equal to 20 years; glomerular filtration rate (GFR) and renal plasma flow (RPF) measured at least 7 years before follow-up study; duration of diabetes at initial examination 3-7 years. All patients fulfilling these criteria accepted a re-examination. The end-point at follow-up (final criterium) was the level of urinary albumin excretion (UAE), either elevated (greater than or equal to 15 micrograms/min) or normal (less than 15 micrograms/min). A clear discrimination was seen, patients being either grossly abnormal (95-4117 micrograms/min) or normal (2.6-7.4 micrograms/min). A marked difference in initial GFR was seen: 166 ml/min +/- 15.4 in those with high UAE at follow-up versus 138 +/- 8.6 in patients with normal UAE at follow-up (2p = 0.2%). The GFR at follow-up was significantly decreased in diabetics with high follow-up UAE (mean values 166----80 ml/min) but stable in patients with low UAE (138----132 ml/min). Initial blood pressure, plasma glucose and RPF were not different between groups. Marked glomerular hyperfiltration, whatever its cause, may contribute to late glomerular damage in diabetic nephropathy. Early measurements of GFR and UAE can be used to identify patients at risk of subsequently developing nephropathy.