James Wright and colleagues reported results of a meta-analysis of data on the effects of various antihypertensive drugs.1 Their stated purpose was to assist physicians in choosing an initial antihypertensive drug by systematically quantifying the available evidence on efficacy, defined as lowering blood pressure and preventing adverse outcomes. They did not achieve this goal, however, because they focused exclusively on clinical trial data. Although they mentioned the importance of exercising treatment decisions on the basis of the best available evidence, they failed to remind physicians that the real-world effectiveness of antihypertensive therapies is also largely a function of patient compliance. Unfortunately, although Wright and colleagues included data for study withdrawals, they did not consider that real-world compliance cannot be studied under the conditions imposed by trials.2
If they had deemed results from studies that investigated compliance with antihypertensive therapies in actual practice as additional evidence worthy of consideration, physicians would also have been informed that class-specific patterns of persistence with initial antihypertensive drug therapy have emerged.[3–6] Persistence with antihypertensive therapy, for example, is generally poor, particularly for initial therapy with older agents such as diuretics and β-blockers. Therefore, to conclude, as the authors have, that physicians should select a diuretic in the absence of contraindications ignores the best available evidence. If the ultimate goal of antihypertensive therapy is to control hypertension and to avoid cardiovascular events, then physicians must consider all available evidence. An anti-hypertensive medication is only efficacious if a patient remains on therapy, and initial choice of antihypertensive therapy appears to be a significant factor in achieving this outcome.