Several reports suggest that when diuretics are added to nifedipine (N), they do not exert any additional hypotensive effect to that produced by N alone. We present the first double-blind, crossover trial to investigate this interaction. Twenty-four black patients with moderate to severe essential hypertension entered the trial. After an initial "open" 4 weeks of therapy with N slow release (SR) 20 mg. b.i.d., those 17 patients whose blood pressures (BP) were not controlled (greater than 160 mm Hg systolic and/or greater than 90 mm Hg diastolic) were randomly allocated (double-blind) to 4 weeks treatment with N SR 20 mg. b.i.d., plus either mefruside 1 q.d. (a thiazide-like diuretic) or matching placebo. Patients then crossed over for a second 4 week treatment period. Blood pressures were measured at 2 weekly intervals under the same conditions using Hawksley random zero sphgymomanometers by one of two standardized observers after patients had been lying for 5 min and standing for 2 min. Analysis (taking account of period effect) of the mean results for the 16 patients completing the trial confirms that, contrary to what previous uncontrolled data suggest, lying and standing systolic and diastolic BPs are significantly lower (8.5/4.5 mm Hg: 2p less than 0.01 and 7.9/5.0 mm Hg: 2p less than 0.05, respectively) with nifedipine plus diuretic than with nifedipine plus placebo.