Meta-analyses of antihypertensive therapy: Are some of them misleading?

Meta-analysis has become a very popular tool to compare the efficacy of different antihypertensive regimens. Combining results from various outcome studies may provide evidence to guide the therapeutic approach even before results from large prospective studies are available. However, meta-analysis may be misleading if it is not done meticulously. Some meta-analyses that received broad news media coverage in the recent years were misleading. One analysis suggested that the use of short-acting nifedipine in moderate to high doses in patients with coronary disease increased mortality. This claim was refuted later by observational studies. Based on another meta-analysis, it was claimed that diuretics and β-blockers are equally effective in reducing cardiovascular morbidity and mortality. Another more careful meta-analysis, omitting one study in which most patients were on combination therapy and not on β-blocker monotherapy, showed the superiority of diuretic versus β-blocker treatment in the elderly. Calcium antagonists were recently blamed for increasing the rate of myocardial infarction and congestive heart failure in hypertensive patients, and therefore their use was not recommended as first-line therapy in hypertension. This recommendation was based on a meta-analysis subject to major drawbacks and was misleading. Another notion based on meta-analysis was that angiotensin converting enzyme inhibitors reduce left ventricular mass more than diuretics. This notion was refuted by three large randomized studies. A recent meta-analysis, which showed a similar blood pressure lowering effect for all angiotensin receptor blockers, was refuted by head-to-head studies. Thus, when performed correctly, meta-analysis can be an important tool, but when uncritically employed, it is prone to be misleading.