Hypertension-associated renal disease is a major cause of end-stage renal disease (ESRD) in the United States, but its risk factors remain incompletely defined. Identification and correction of amendable ESRD risk factors among patients with essential hypertension could reduce ESRD prevalence. Patients referred by their primary care physician for hypertension management to an academic nephrology clinic during calendar year 1995 were followed up prospectively. Studied patients had no evidence of secondary hypertension, diabetes, or primary renal disease. All were treated pharmacologically toward a target mean blood pressure (MBP) of 100 mm Hg or less. The course of renal function during follow-up was assessed as the slope of the reciprocal of plasma creatinine concentration (1/P(cr)) plotted against months of follow-up and as the change in calculated glomerular filtration rate (GFR) in milliliters per minute per month. The following patient characteristics were prospectively examined as possible predictive factors for altered renal function: age, sex, ethnicity, initial MBP, initial P(cr) level, initial level of urine protein excretion, and smoking status. Fifty-three patients were enrolled, and follow-up data were available for 51 patients after a mean follow-up of 35.5 months. Despite MBP reduction from 126.8 +/- 1.3 to 96.5 +/- 1.1 mm Hg (P < 0.0001), P(cr) level increased from 1.5 +/- 0.1 to 1.9 +/- 0. 2 mg/dL (P < 0.01). Multivariate regression analysis showed that smoking, greater initial P(cr) level, and black ethnicity were the only examined parameters that independently predicted both a decrease in the 1/P(cr) slope and calculated GFR with at least 95% confidence. Smoking was by far the most powerful of the examined factors, with initial P(cr) and ethnicity being much less predictive. These studies show for the first time that smoking is an independent risk factor for renal function decline in patients with severe essential hypertension.