This study aimed to establish the time of initiation and the determinants of renal function decline in type 1 diabetes. Until now, such decline has been assumed to be a late-occurring event associated with proteinuria. A total of 267 patients with normoalbuminuria and 301 patients with microalbuminuria were followed for 8 to 12 yr. Linear trends (slopes) in GFR were estimated by serial measurement of serum cystatin C. Cases of early renal function decline were defined by loss in cystatin C GFR that exceeded -3.3%/yr, a threshold that corresponds to the 2.5th percentile of the distribution of GFR slopes in an independent nondiabetic normotensive population. Cases of early renal function decline occurred in 9% (mean slope -4.4; range -5.9 to -3.3%/yr) of the normoalbuminuria group and 31% (mean slope -7.1; range -23.8 to -3.3%/yr) of the microalbuminuria group (P < 0.001). Risk for early renal function decline depended on whether microalbuminuria regressed, remained stable, or progressed, rising from 16 to 32 and 68%, respectively (P < 0.001). In multivariate analysis, risk for decline was higher after age 35 yr or when glycosylated hemoglobin exceeded 9% but did not vary with diabetes duration, smoking, BP, or angiotensin-converting enzyme inhibitor treatment. Contrary to the existing paradigm of diabetic nephropathy, progressive renal function decline in type 1 diabetes is an early event that occurs in a large proportion of patients with microalbuminuria. Together with testing for microalbuminuria, clinical protocols using cystatin C to diagnose early renal function decline and track response to therapeutic interventions should be developed.