Background: The apparent shortfall in prevention of coronary heart disease (CHD) noted in early hypertension trials has been attributed to disadvantages of the diuretics and beta blockers used. For a given reduction in blood pressure, some suggested that newer agents would confer advantages over diuretics and beta blockers. Our aim, therefore, was to compare the effect on non-fatal myocardial infarction and fatal CHD of combinations of atenolol with a thiazide versus amlodipine with perindopril.
Methods: We did a multicentre, prospective, randomised controlled trial in 19 257 patients with hypertension who were aged 40-79 years and had at least three other cardiovascular risk factors. Patients were assigned either amlodipine 5-10 mg adding perindopril 4-8 mg as required (amlodipine-based regimen; n=9639) or atenolol 50-100 mg adding bendroflumethiazide 1.25-2.5 mg and potassium as required (atenolol-based regimen; n=9618). Our primary endpoint was non-fatal myocardial infarction (including silent myocardial infarction) and fatal CHD. Analysis was by intention to treat.
Findings: The study was stopped prematurely after 5.5 years' median follow-up and accumulated in total 106 153 patient-years of observation. Though not significant, compared with the atenolol-based regimen, fewer individuals on the amlodipine-based regimen had a primary endpoint (429 vs 474; unadjusted HR 0.90, 95% CI 0.79-1.02, p=0.1052), fatal and non-fatal stroke (327 vs 422; 0.77, 0.66-0.89, p=0.0003), total cardiovascular events and procedures (1362 vs 1602; 0.84, 0.78-0.90, p<0.0001), and all-cause mortality (738 vs 820; 0.89, 0.81-0.99, p=0.025). The incidence of developing diabetes was less on the amlodipine-based regimen (567 vs 799; 0.70, 0.63-0.78, p<0.0001).
Interpretation: The amlodipine-based regimen prevented more major cardiovascular events and induced less diabetes than the atenolol-based regimen. On the basis of previous trial evidence, these effects might not be entirely explained by better control of blood pressure, and this issue is addressed in the accompanying article. Nevertheless, the results have implications with respect to optimum combinations of antihypertensive agents.