Patients with a stable progression of chronic renal failure with a creatinine clearance of 15-45 mL/min were randomly assigned to two groups of antihypertensive therapy: 1--nisoldipine as the only antihypertensive agent and 2--antihypertensive drugs without calcium channel blockers and a placebo tablet instead of nisoldipine. The patients were already on a low-protein diet and some form of antihypertensive therapy but without calcium channel blockers. There were 18 patients in the placebo group and 20 patients in the nisoldipine group. The follow-up period averaged 23.7 +/- 10.6 (SD) months in the placebo group and 23 +/- 11.3 months in the nisoldipine group. The slopes of the reciprocal of serum creatinine were calculated for the period prior to and following our intervention. The number of patients whose slopes improved following intervention was 6/18 in the placebo group and 15/20 in the nisoldipine group (p less than .02). The patients whose slopes improved had a significant fall in systolic and diastolic BP, as well as in the MAP. Those whose slopes did not improve had a significant decrease in systolic BP, but no change in diastolic BP and no significant difference in the MAP. When all 38 patients are analyzed together, regardless of their grouping, the correlation between the difference percent in the slope, and the difference percent in the MAP, was significant. Furthermore, punch biopsies of the skin showed a markedly different calcium content in the two groups, which was significantly less in the nisoldipine-treated patients as compared with the patients not receiving calcium blockers.(ABSTRACT TRUNCATED AT 250 WORDS)