Nephropathy is the major life-threatening complication of insulin-dependent diabetes mellitus (IDDM). The clinical syndrome is characterized by persistent albuminuria (greater than 300 mg day), a rise in arterial blood pressure, and a relentless decline in glomerular filtration rate leading to end-stage renal failure. The availability of a radioimmunoassay for detecting albumin in low concentrations in urine has allowed the study of urinary albumin excretion rates in diabetics well before clinically persistent proteinuria develops. An albumin excretion rate greater than that in normal subjects and lower than that in macroalbuminuric subjects is called microalbuminuria (range 20-200 microg/min or 30-300 mg/24 h). Although recent studies have challenged the predictive value of microalbuminuria for later development of overt diabetic nephropathy, albumin excretion rate in the microalbuminuric range and its tracking (i.e. annual increase) are still considered reliable markers for prediction of later overt diabetic kidney disease. Overnight urinary collection is preferred for calculation of the rate of albumin excretion, but may be difficult to perform precisely. The albumin:creatinine ratio of the first morning urine sample is a reliable screening method: the microalbuminuric range is considered to be 2.5-25 mg/mmol or 30-300 mg/g (3.5 mg/mol has been proposed as lower limit in females because of their lower creatinine excretion). Irrespective of the procedure used, at least two samples over a 3-6-month period should test positive before microalbuminuria is confirmed and 'persistent microalbuminuria' defined. If the albumin excretion rate is persistently in the microalbuminuric range it is of crucial importance to define strategies and carry out interventions for prevention of decline in kidney function. The goal of achieving the best glycaemic control as early as possible in as many IDDM patients as is safely possible is particularly important in microalbuminuric patients. Although it is unsafe to reduce dietary protein intake drastically, particularly in children and adolescents, moderate decrease of protein intake (i.e. 0.9-1.1/g/kg day) is advisable in diabetic patients from the very beginning of the disease. Timely treatment with an angiotensin-converting enzyme inhibitor, independently of rise in arterial blood pressure, should be considered if improvement of glycaemic control and moderate decrease of dietary protein intake for 6-12 months have failed to reduce the albumin excretion rate. Screening programmes for microalbuminuria and early intervention can substantially modify the natural history of diabetic renal involvement and disease and possibly reduce the incidence of end-stage renal failure.