Like many of us, when I was completing my graduate medical education, my mentors emphasized that quality clinical trial evidence indicated that β-blockers were among the initial agents of choice in the treatment of hypertension but were contraindicated in patients with chronic heart failure due to systolic dysfunction. Obviously, in the past decade the pendulum has swung; select β-blockers are now seen as a cornerstone in the management of chronic heart failure but are falling out of favor as an initial agent in the treatment of hypertension.
In the United States in 1997, the sixth report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC VI) stated that, in the absence of compelling indications for the use of another antihypertensive class, β-blockers or thiazide-type diuretics were recommended agents for the initial treatment of hypertension, with diuretics preferred in older patients.2 With the publication of JNC 7 in 2003, based on new clinical data, the list of recommended agents for initial hypertensive therapy was expanded to include angiotensin-converting enzyme inhibitors (ACEIs), angiotensin receptor blockers (ARBs), and calcium channel blockers (CCBs) in addition to thiazide-type diuretics and β-blockers. Due to their low cost and unsurpassed clinical efficacy, thiazide-type diuretics were labeled as “preferred” among these agents for most patients.3 Similarly, in 2003 the European Society of Hypertension-European Society of Cardiology (ESH-ESC) included β-blockers among their list of drugs recommended for initial therapy in hypertension.4 Previous guidelines from the British Hypertensive Society (BHS) also included β-blockers as a recommended choice for initial therapy, most recently in 2004. In 2006, the BHS along with the National Collaborating Centre for Chronic Conditions (United Kingdom), published an update of their guidelines for treatment of hypertension that explicitly excludes β-blockers from the list of recommended initial antihypertensive medications in the absence of concomitant compelling indications for their use (such as post-myocardial infarction, chronic angina, or systolic dysfunction).1
In their review of the clinical data concerning this question, the BHS stated: “In head-to-head trials, β-blockers were usually less effective than a comparator drug at reducing major cardiovascular events, particularly stroke.” In the absence of compelling indications for other antihypertensive medications, for patients at least 55 years of age or black patients of any age, BHS recommends initial therapy with either a CCB or a thiazide-type diuretic. For patients younger than 55, BHS recommends initial therapy with an ACEI (or an ARB if the patient cannot tolerate ACEIs). In this new treatment algorithm, β-blockers are relegated to one of a number of fourth-step drugs for patients whose blood pressure (BP) is not controlled despite adequate doses of ACEIs/ARBs, CCBs, or thiazide-type diuretics.1 This fundamental change in the perceived value of β-blockers in the treatment of hypertension occurred without the publication of any single major new study of initial therapy of hypertension. The remainder of this commentary will examine the new thinking that led to this recommendation and attempt to determine whether the pendulum has in fact swung too far.
To examine this question it is helpful to go back and briefly review data that originally led to the inclusion of β-blockers as one of the agents for initial therapy. Essentially, a number of randomized prospective clinical trials were performed, mostly in the 1970s and 1980s, that demonstrated the efficacy of antihypertensive therapy in patients with essential (mostly diastolic) hypertension. The majority of these studies used thiazide-type diuretics or β-blockers. As discussed extensively in JNC VI, there were limited data on the effects of other antihypertensive agents on cardiovascular events at the time of its publication.
The Medical Research Council (MRC) trial5 of antihypertensive therapy in older people was probably the first major clinical trial to cast doubt on the relative efficacy of β-blocker therapy, at least in an older population. In this trial, thiazide-type diuretics were more effective at reducing cardiovascular events than both placebo and the β-blocker atenolol; in fact, atenolol was no better than placebo in reducing the incidence of myocardial infarction in this population. Subsequent analyses of these early clinical trials, especially those using atenolol, have suggested that the majority, if not the totality, of the benefits seen were attributable to thiazide-type diuretic treatment, with no evidence of significant benefit in cardiovascular event rates in patients treated with β-blockers alone.6 Similarly, in the meta-analysis performed as part of the BHS report, in the 3 studies that directly compared β-blocker with thiazide-type diuretic therapy, β-blocker use was associated with a trend toward increased incidence of mortality, myocardial infarction, and stroke. Studies comparing β-blockers with other classes of antihypertensive agents were also analyzed, and in no case did β-blockers appear to provide better results than the comparator.1 Almost all of the data analyzed were in older patients.
Of course, such meta-analyses should only be considered hypothesis generating; however, there have also been several prospective randomized clinical trials that have not demonstrated favorable results with β-blockers compared with other classes of antihypertensive agents. In the Losarten Intervention for Endpoint Reduction in Hypertension (LIFE) study7 an ARB (losartan)-based regimen was demonstrated to be more effective in preventing a composite of cardiovascular events than a β-blocker (atenolol)–based treatment program in subjects with hypertension and electrocardiographic evidence of left ventricular hypertrophy, a finding that was mostly driven by a difference in stroke rate. In LIFE, most subjects in both groups received add-on therapy with hydrochlorothiazide (plus other medications as needed to get BP to goal). In the more recently published Anglo-Scandinavian Cardiac Outcomes Trial (ASCOT),8 subjects with hypertension and multiple other cardiovascular risk factors were randomized in openlabel fashion to initial therapy with either the CCB amlodipine or the β-blocker atenolol; an ACEI was added to the CCB after titration and a diuretic added to the β-blocker. Despite similar peripheral BP at the end of the study, the CCB+ACEI group had significantly fewer cardiovascular events, mostly driven by a difference in stroke and incident heart failure. BP differences were, however, noted during the first few months of the study when the CCB and β-blocker were being titrated; BPs were lower with the CCB. One must question whether these BP differences may have accounted for the differences in outcome.9
Like all clinical trials, LIFE, ASCOT, and other trials that have compared β-blockers with other antihypertensive drug classes have well-publicized limitations. One of the most frequently cited is the use and potential under-dosing of atenolol in the majority of these studies. It has been suggested that atenolol may have specific attributes, including its inability to efficiently cross the blood-brain barrier and its relatively short half-life, that make it less effective than other β-blockers at reducing cardiovascular events, especially when used once daily. Whether the use of different β-blockers, higher doses of atenolol, or atenolol dosed twice daily would have led to different outcomes remains unknown. Determining the validity of the ASCOT results is also limited by the lack of blinding, the reliance on secondary end points (there was no difference in primary outcome between the treated groups), and the inability to separate out the incremental effect of each drug used.
Even taking into account these limitations, the totality of the data available to us seems to show a trend toward less effective reduction in cardiovascular events, particularly stroke, with β-blockers compared with thiazide-type diuretics and other agents. In fact, the Conduit Artery Function Evaluation (CAFE) substudy10 of ASCOT may shed light on a potential physiologic rationale for the suboptimal findings found in many of the β-blocker studies. In CAFE, 2073 patients underwent repeat noninvasive central BP measurement using a well-validated commercially available device. While the peripheral BPs were the same in both treatment groups, the central BP was found to be significantly lower in patients randomized to the CCB+ACEI group (systolic BP, –4.3 mm Hg). For many years it has been well established, but often forgotten, that unlike most other major classes of antihypertensive medications, β-blockers have a greater effect on peripheral BP than central BP. This may be secondary to their effects on pulse wave velocity. Since cardiovascular events, particularly stroke and heart failure, are more likely related to central BP elevation, the difference in central BP seen in CAFE may explain the difference in event rates seen in these β-blocker trials. Perhaps if ASCOT and other clinical trials had been designed to achieve similar central rather than peripheral BP, better results would have been seen in patients randomized to the β-blocker.
Many β-blockers, including atenolol, appear also to increase insulin resistance, although given the relatively short duration of the clinical trials discussed here, that is an unlikely explanation for the clinical results. The risk of worsening insulin resistance is another reason why many β-blockers may not be ideal early therapy. This may be especially true for patients already on a thiazide-type diuretic, which may have similar untoward metabolic effects, but in that case, cardiovascular outcome does not appear to be adversely effected.
This conclusion is not definitive; we certainly don't know that β-blockers, and especially β-blockers other than atenolol, are less effective than other antihypertensive medications. More importantly, however, we lack any solid data demonstrating that β-blockers are as effective as other classes of antihypertensive medications with proven benefits for cardiovascular protection. Based on the results of the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT) and other trials, we have excellent quality data demonstrating relatively equivalent reduction in cardiovascular events with thiazide-type diuretics, ACEIs, CCBs, and ARBs. For initial monotherapy, we cannot say for certain that β-blockers are worse, but we cannot say that they are as good either. Ironically, β-blockers now find themselves in a position similar to that of ACEIs, ARBs, and CCBs at the time JNC VI was published; definitive data showing that they provide equivalent reduction in cardiovascular events to that demonstrated by other available antihypertensive agents are lacking. Future clinical trials will need to be designed to explicitly answer that question; however, given the expense that would be associated with such a trial and the number of available proven options for antihypertensive therapy, it seems unlikely that such a trial will be undertaken. The fact that many of the most popular β-blockers are either currently or soon to be generically available also makes it unlikely that funding from the pharmaceutical industry would be readily available for such a trial.
Of course, this is not equivalent to saying that β-blockers have no role in the management of hypertension. As the BHS guidelines point out, β-blockers should continue to be used in a number of important clinical situations where they have demonstrated clinical efficacy, including in patients with heart failure, angina, and after a myocardial infarction (compelling indications). In addition, since the majority of patients with hypertension will require multiple medications to reach goal, β-blockers will continue to have an important role as add-on medications. The recommendation from BHS that β-blockers not be considered until the fourth drug in most cases appears to be based more on conjecture than solid clinical trial evidence, and it remains unclear exactly when they should be included as part of combination therapy and in which patients. β-Blockers should also be considered as early, if not initial, therapy in younger patients with evidence of increased sympathetic drive or with contraindications to or intolerance of both ACEIs and ARBs.11 If patients are currently on a β-blocker, are tolerating it well, and have well-controlled BP, there is no reason to change their therapy.1 Undoubtedly, given the nature of the evidence available to us, this contentious issue will continue to be hotly debated within the hypertension community. Given the number of well-tolerated classes of antihypertensive medications with proven results as initial therapy and the questions surrounding the relative magnitude of the benefit seen with β-blockers, β-blockers should not at present be recommended as initial therapy in most patients with hypertension. For now, the pendulum appears to have swung just far enough.