Cardiovascular protection and blood pressure reduction: a meta-analysis

Background Whether antihypertensive drugs offer cardiovascular protection b eyond blood pressure lowering has not been established. We aimed to investi gate whether pharmacological properties of anti hypertensive drugs or reduc tion of systolic pressure accounted for cardiovascular outcome in hypertens ive or high-risk patients. Methods In a meta-analysis we extracted summary statistics from published r eports, and calculated pooled odds ratios for experimental versus reference treatment. We correlated across-trials odd ratios for differences in systo lic pressure between groups. Findings We analysed nine randomised trials comparing treatments in 62 605 hypertensive patients. Compared with old drugs (diuretics and beta -blocker s), calcium-channel blockers and angiotensin converting-enzyme inhibitors o ffered similar overall cardiovascular protection, but calcium-channel block ers provided more reduction in the risk of stroke (13.5%, 95% Cl 1.3-24.2, p=0.03) and less reduction in the risk of myocardial infarction (19.2%, 3.5 -37.3, p=0.01). Heterogeneity was significant between trials because of hig h risk of cardiovascular events on doxazosin in one trial, and high risk of stroke on captopril in another; but systolic pressure differed between gro ups in these two trials by 2-3 mm Hg. Similar systolic differences occurred in a trial of diltiazem versus old drugs, and in three trials of convertin g-enzyme inhibitor against placebo in high-risk patients. Meta-regression a cross 27 trials (136 124 patients) showed that odds ratios could be explain ed by achieved differences in systolic pressure. Interpretation Our findings emphasise that blood pressure control is import ant. All antihypertensive drugs have similar long-term efficacy and safety. Calcium-channel blockers might be especially effective in stroke preventio n. We did not find that converting-enzyme inhibitors or a-blockers affect c ardiovascular prognosis beyond their antihypertensive effects.