We designed a prospective, double-blind controlled trial to determine predictors of loss of renal function in patients with insulin dependent diabetes and established nephropathy. A total of 409 insulin-dependent diabetic patients with established nephropathy enrolled in a trial on the effect of Captopril on the rate of progression of renal disease. Baseline demographic, clinical (history and physical) and laboratory parameters were analyzed as risk factors for time to progression. Dichotomous characteristics were compared by Fisher's exact test and continuous characteristics with the Wilcoxon rank-sum test. Univariate proportional hazards regression analysis was used to estimate relative risk of nephropathy progression, and bivariate proportional hazard regression to identify interactions with the treatment group assignment. Multivariate proportional hazard regression was employed to determine which characteristics were independent risk factors. We found that a number of demographic and clinical characteristics were significantly associated with nephropathy progression even after adjustment for treatment group. However, after multivariate analysis, the risk factors that independently predicted progression were onset of IDDM later in life, parental diagnosis of IDDM, the presence of edema, increased mean arterial pressure, and an abnormal electrocardiogram. Likewise, a number of laboratory characteristics were also predictive of nephropathy progression. A low hematocrit, high blood sugar, and higher protein excretion predicted nephropathy progression as did a higher serum creatinine, particularly in the face of a normal serum albumin. In conclusion, this study identifies a number of clinical and laboratory risk factors that can predict which patients with insulin-dependent diabetes with established nephropathy are more likely to sustain a clinically important decrease in renal function over a median follow-up of three years.