We compared the changes in serum creatinine levels over time after a mean follow-up of 9.8 years in essential hypertensive (EH, n = 56) and control (n = 59) male veteran subjects. All subjects had normal serum creatinine levels (62 to 124 mumol/L) and "normal" urinalysis results on entry into the study. Subjects with comorbid renal diagnoses and diabetes mellitus were eliminated from the analysis. Although not statistically significant, the rate of change in the serum creatinine concentration over time was greater in the EH cohort compared with the control cohort (1.08 +/- 4.8 vs 0.027 +/- 3.5 mumol/L per year). The difference was especially marked in black EH subjects vs black control subjects (1.60 +/- 6.2 mumol/L per year vs -0.21 +/- 3.3 mumol/L per year). When age, race, body mass index, and a diagnosis of EH were entered into a logistic regression analysis, EH subjects had a statistically significantly greater rate of decline in renal function than did control subjects (1.5 +/- 8.3 mumol/L per year). When mean time-averaged systolic blood pressure for each subject was also included in the logistic regression analysis, only systolic time-averaged blood pressure was statistically significant (0.063 +/- 0.029 mumol/L per year). We conclude that in the absence of clinically detected parenchymal renal disease, EH subjects have a greater rate of decline in renal function than do nonhypertensive subjects. Time-averaged blood pressure is predictive of the change in serum creatinine concentration not only in EH subjects but also in nonhypertensive subjects. Thus, preservation of renal function may require a blood pressure lower than the currently accepted normotensive range.