We prospectively followed a cohort of 278 non-insulin-dependent (NIDDM) patients for a 6-year period, intending to estimate the rate of increase of albuminuria and to identify clinical variables that influence this increase. At baseline, normo-albuminuria (N) was seen in 74%, microalbuminuria (M) in 19% and 7% presented with proteinuria (P). A total of 80 patients died; they were older (p < 0.001) and had higher albumin excretion both at baseline and as an average during follow-up (p < 0.01). At baseline, patients with proteinuria had higher blood pressures (systolic and diastolic), whereas there was no difference between patients with normo-and microalbuminuria. Glycaemic control was increasingly poor throughout the three groups. At follow-up, an average relative rate of increase of albuminuria (slope) of 17% per year was seen both for patients with complete 6 years, follow-up (n - 135) and patients with at least 4 years follow-up (n = 178). Slope correlated significantly with systolic blood pressure (r = 0.26 and 0.29) in both groups, diastolic blood pressure only in the 4-year group (r = 0.22) and average albuminuria in both (r = 0.31 and 0.24). By multiple regression analyses systolic blood pressure and average albuminuria remained with significant influence on slope. Progression was defined as an increase in the category (e.g. normo- to microalbuminuria) as well as an increase of more than 20% in albumin excretion, and was seen in 46 patients with at least 4 years' follow-up.(ABSTRACT TRUNCATED AT 250 WORDS)