The relationship between the hypotensive effect of nitrendipine (N), 20 mg/day (n = 17), or atenolol (A), 100 mg/day (n = 17), and the erythrocyte sodium [( Na]i) and potassium [( K]i) concentrations was investigated in hypertensive African blacks during a randomized double-blind study. After 6 weeks, both treatments significantly reduced supine and standing blood pressures; however, the magnitude of the decrease in supine systolic (-22.0 +/- 2.0 vs -12.1 +/- 3.4 mm Hg) and diastolic (-14.1 +/- 1.3 vs -7.6 +/- 2.1 mm Hg) pressures and in standing diastolic pressure (-16.0 +/- 1.7 vs -9.2 +/- 2.0 mm Hg) was more pronounced (p less than 0.05) in the N-treated than in the A-treated group. Pulse rate, plasma aldosterone, and plasma renin activity were lower (p less than 0.05) in the A-treated patients. Neither treatment had significant influence on [Na]i, [K]i, or ouabain-sensitive sodium efflux. The N-induced changes in supine systolic and diastolic pressure correlated (p less than 0.05) with age (r = -0.65 and r = -0.58, respectively) and pretreatment plasma renin activity (r = 0.71). Multiple regression analysis demonstrated a negative association between pretrial [Na]i and the change in systolic pressure during N treatment that was independent of age, pretreatment blood pressure, and change in pulse rate. Age and the change in supine pulse rate were also independently correlated with the change in diastolic pressure during N treatment. The results show a greater antihypertensive efficacy of N than A in the patients entered in this study and suggest that a higher intracellular sodium concentration could predict a better hypotensive response to N.