The introduction of more efficacious treatments for diabetic kidney disease may slow its progression, but evidence for their effectiveness in populations is sparse. We examined trends in the incidence of clinical proteinuria, defined as a urinary protein-to-creatinine ratio >0.5 g/g, and diabetic end-stage renal disease (ESRD), defined as death from diabetic nephropathy or onset of dialysis, in Pima Indians with type 2 diabetes between 1967 and 2002. The study included 2189 diabetic subjects >/=25 years old. During follow-up, 366 incident cases of proteinuria occurred in the subset of 1715 subjects without proteinuria at baseline. The age-sex-adjusted incidence rate of proteinuria increased from 24.3 cases/1000 person-years (pyrs) (95% confidence interval (CI) 18.7-30.0) in 1967-1978 to 35.4 cases/1000 pyrs (95% CI 28.1-42.8) in 1979-1990 and 38.9 cases/1000 pyrs (95% CI 31.2-46.5) in 1991-2002 (P(trend)<0.0002). In each period, the age-sex-adjusted incidence of proteinuria increased with diabetes duration, but diabetes duration-specific incidence was stable throughout the study period (P=0.8). The age-sex-adjusted incidence of ESRD increased between 1967 and 1990 and declined thereafter. The incidence of proteinuria increased over 36 years in Pima Indians as the proportion of people with diabetes of long duration increased. On the other hand, the incidence of ESRD declined after 1990, coinciding with improved control of blood pressure, hyperglycemia, and perhaps other risk factors.