To assess differences in the patterns of age-specific black/white comparisons between end-stage renal disease (ESRD) attributed to diabetes (ESRD-DM) and not attributed to diabetes (ESRD-NON-DM), data for subjects 20 to 79 years of age reported by the US Renal Data System as incident cases of ESRD during 1988 to 1991 were analyzed. While the black to white incidence rate ratio (B/W RR) for ESRD-NON-DM peaked in patients before the age of 40 years, the most striking B/W RRs for ESRD-DM were observed in patients older than 40 years. This study also explored evidence supporting the hypothesis that an increased risk of premature death attributed to cardiovascular disease (CVD death) in black patients, alone or in combination with black/white differences in prevalence of diabetes, influences the pattern of age-specific black/white ESRD-DM comparisons. By using estimates of the diabetic population as denominators for the rates, the incidence or ESRD-DM remained much higher in black patients than in white patients for those aged 45 years or above. However, the incidence of ESRD-DM for patients aged below 45 years was found to be significantly (P < 0.05) lower (B/W RR = 0.6) for black male diabetic patients and slightly, yet significantly, higher (P < 0.05; B/W RR = 1.1) for black female diabetic patients than for their white counterparts. Therefore, prevalence of diabetes could not fully explain the pattern of age-specific B/W RR for ESRD-DM.(ABSTRACT TRUNCATED AT 250 WORDS)