The benefit of antihypertensive therapy in reducing cardiovascular morbidity and mortality associated with isolated systolic hypertension has now been established by the Systolic Hypertension in the Elderly Program. However, there is little information about the relative effectiveness of different drug regimens in this condition. This study compared the efficacy and tolerability of 50 mg of atenolol, 10 mg of enalapril, 25 mg of hydrochlorothiazide and 2.5 mg of isradipine in the treatment of isolated systolic hypertension. After a 3-week placebo run-in phase, 24 subjects were randomized into a 4-period double-blind crossover study by use of an orthogonal latin square design. Treatment periods were of 6 weeks' duration with titration to a higher dose after 4 weeks in those not reaching goal blood pressure (BP). Each active treatment was followed by a 3-week placebo washout. Casual clinic and 24-hour ambulatory BP (Accutracker II) were measured at the end of each treatment phase. Routine biochemistry was also performed after the placebo run-in, at the end of each active treatment phase, and after the placebo run-out. Of the 24 subjects entered (mean age 72.3 years, 38% men) 20 completed the whole study. Mean ± standard deviation of supine clinic and daytime ambulatory BP on entry were 181/79 ± 21/9 mm Hg and 165/82 ± 23/15 mm Hg, respectively. All drugs reduced mean casual and ambulatory BP significantly relative to placebo but only hydrochlorothiazide and enalapril produced a consistent hypotensive effect throughout the entire 24-hour period. Isradipine and enalapril exhibited a relatively greater effect on reducing systolic BP than either hydrochlorothiazide or atenolol. The frequency of adverse effects was similar except ankle edema was more frequent with isradipine, cough with enalapril, and breathlessness with atenolol. It is concluded that in the dosages used, both hydrochlorothiazide and enalapril provide suitable treatment for isolated systolic hypertension. The place of dihydropyridine like calcium antagonists requires further assessment.