The effect of atenolol and reserpine on incidence of strokes, coronary heart disease (CHD), cardiovascular disease (CVD), and mortality was assessed in 4736 persons aged 60 years and older with isolated systolic hypertension. Participants were randomized to either chlorthalidone (2371), with step-up to atenolol, or reserpine if needed, or placebo (2365). The average baseline SBP/DBP was 170/77 mm Hg. In the active treatment group, step 1, dose 1 was chlorthalidone, 12.5 mg/day; dose 2 was 25 mg/day. For step 2, dose 1 was atenolol 25 mg/day (or reserpine 0.05 mg/day if atenolol was contraindicated); dose 2 was 50 mg/day (reserpine, 0.10 mg/day).
During 4.5 years average follow-up, 32% (757) of the active treatment group were on atenolol, with an average exposure of two years and 8% (193) were on reserpine with an average exposure of 1.7 years. Overall there were 96 strokes, 140 CHD events and 289 CVD events among the 2365 active group participants. Using time-dependent lifetable regression with adjustment for several variables, the addition of either atenolol or reserpine to chlorthalidone did not substantially alter the risk ratios for chlorthalidone alone. The relative risk for CHD events for atenolol versus no atenolol was 1.04 (95% confidence interval: 0.58, 1.86) and for reserpine versus no reserpine was 0.93 (95% confidence interval: 0.29, 2.96). The relative risks for atenolol were 0.84 (95% confidence interval: 0.54, 1.30) for death, 1.34 (95% confidence interval: 0.80, 2.28) for stroke, and 1.07 (95% confidence interval: 0.71, 1.61) for CVD. For reserpine, the corresponding relative risks and confidence intervals were 0.65 (0.26, 1.59) for death, 0.27 (0.04, 2.26) for stroke, and 0.55 (0.20, 1.49) for CVD.
Thus, the beneficial effects in several outcomes in Systolic Hypertension in the Elderly Program (SHEP) were due to the treatment regimen of lowering blood pressure based on low-dose chlorthalidone (plus atenolol or reserpine as required to meet blood pressure criteria). Additional (independent) benefits attributable to atenolol or to reserpine were not identified. However, a greater number of patients might have been necessary to adequately evaluate potential differential effects of these drugs, especially for reserpine.
Am J Hypertens 1995;8:1147–1153