To estimate the frequency of remission/regression of microalbuminuria and to identify factors affecting such outcomes in Japanese patients with type 2 diabetes, we observed 216 patients with type 2 diabetes and microalbuminuria enrolled during an initial 2-year evaluation period for the next 6 years. Remission was defined as shift to normoalbuminuria and regression as a 50% reduction in urinary albumin excretion rate (AER) from one 2-year period to the next. Reduction of urinary AER was frequent, with a 6-year cumulative incidence of 51% (95% CI 42-60) for remission and 54% (45-63) for regression, whereas the frequency of progression to overt proteinuria was 28% (19-37). Microalbuminuria of short duration, the use of renin-angiotensin system-blocking drugs, and lower tertiles for HbA(1c) (<6.95%) and systolic blood pressure (<129 mmHg) were independently associated with remission or regression in the pooled logistic regression analysis. The results indicate that reduction in urinary AER occurs frequently in Japanese patients with type 2 diabetes. Early detection of microalbuminuria and a multifactorial control may result in improved outcomes for diabetic nephropathy and cardiovascular events.