A comparative, double-blind, placebo-controlled, double-dummy, randomized, crossover trial was performed in 32 hypertensive patients (initial blood pressure [BP] 165/105 +/- 3/2 mm Hg), most of them obese women, to evaluate the antihypertensive efficacy and tolerability of the new calcium channel blocking drug nisoldipine, 10 mg once or twice daily, over 6 weeks compared with the diuretic hydrochlorothiazide-amiloride (hydrochlorothiazide, 25 mg, in combination with amiloride, 2.5 mg once or twice daily). Adequate 24-hour control of BP (supine diastolic BP less than 90 mm Hg) was achieved in 15 of 32 patients (47%) with monotherapy by 1 agent and in 23 of 32 patients (72%) when including those who responded to either agent. Both nisoldipine and the diuretic had a flat dose-response curve. The larger dose of diuretic yielded lower systolic and exercise BP values (erect and supine) than high-dose nisoldipine (p less than 0.05), with less headache (p less than 0.05). In the other 9 patients who did not respond, the combination of nisoldipine 10 mg plus hydrochlorothiazide (25 mg)-amiloride (2.5 mg) was administered and yielded a response in 8 patients (overall response including monotherapy 96%). Ten patients were given nisoldipine for an additional period of 6 months, and they required an average dose of 33 mg/day for BP control. In both the 6-week and 6-month studies, nisoldipine monotherapy caused frequent subjective adverse effects. In contrast, in the 6-week study the combination of low-dose nisoldipine and low-dose diuretic gave good BP control with no adverse effects.