Do β-blockers have a role in treating hypertension?

When your heart’s on fire, you must realize, smoke gets in your eyes.

The Platters, Smoke Gets in Your Eyes

In 1996, a colleague and I wrote in the hypertension chapter of a textbook, “[β-blockers] have proven efficacy in reducing morbidity and mortality”1 (M&M). In hindsight, the only way we could have thought this was true was to ignore the results of almost all the large β-blocker hypertension trials.

Table 1 summarizes the available evidence from placebo-controlled trials looking at the effect of β-blockers on M&M when used for hypertension.2 Five of 7 trials were unable to demonstrate that β-blockers had any effect on individual end points of stroke, myocardial infarction (MI), or overall mortality when compared with placebo. In the 2 trials that did show significant differences, approximately 60% of subjects were taking thiazides, which might have contributed to the observed benefit. Lindholm et al recently published these results as a meta-analysis.2 Compared with placebo, β-blockers produced no statistically significant effect on coronary artery disease or mortality, but there was a 19% relative reduction (<0.5% absolute reduction) in stroke. This is half of what is typically seen with other antihypertensives. Looking only at the trials that compared atenolol with placebo, no statistically significant reduction in any end point was found.

These findings are not new. Earlier meta-analyses3,4 questioned the use of β-blockers for hypertension, especially among the elderly. Given this uncertainty, one would expect the use of β-blockers (especially atenolol) in the elderly to be minimal. Nothing could be further from the truth. British Columbia pharmacare data showed the number of atenolol prescriptions in British Columbia for patients older than 65 had increased each year from 2003 to 2005. It was the most prescribed βblocker in both 2004 and 2005, and now sits at ¼ million prescriptions a year.

In a more recent meta-analysis,5 the authors split the β-blocker studies into those that looked at either younger (<60) or older (≥60) patients. In patients younger than 60, using a composite outcome of MI, stroke, and mortality, the authors calculated a risk ratio of 0.86, with the upper limit of the 95% confidence interval just sneaking (0.99) under the magical 1. Using their estimates, β-blockers produced a whopping 0.5% absolute reduction (over 5 years) in this composite outcome. This led the authors to state that “in younger patients ßblockers are associated with a significant reduction in cardiovascular morbidity and mortality.”5 Interestingly, the main study used to support this statement was the Medical Research Council trial of mild hypertension,6 which reported that “cardiovascular events [were] not reduced by … smokers taking propranolol” and that “all cause mortality was reduced in men on active treatment [thiazide or β-blockers] but increased in women.” Space precludes a detailed critique of this meta-analysis, but one of the problems with retrospective analysis of data is that, if one looks at enough end points, something statistically significant can show up, including the things the MRC mentioned.

So why, in the mid-1990s, did my colleague and I state that β-blockers had been shown to reduce M&M associated with hypertension? In hindsight, I believe we erred in the following ways:

• by talking about thiazides and β-blockers as a group rather than looking at individual agents,

• by not questioning the link between surrogate markers (blood pressure) and M&M,

• by not requiring solid evidence from randomized controlled trials to make recommendations, and

• by selecting β-blockers for post-MI therapy or for patients who also had angina, migraine, or essential tremour and thinking we were getting “2-fers.”

We were not the only ones making these statements. For example, the 1993 report from the Canadian Hypertension Society stated, “since 1988 three studies … have confirmed the efficacy … of diuretics and βblockers and have shown a marked and significant reduction in the risk of [cerebrovascular accident].”7 The 3 studies referenced don’t support that conclusion with regard to β-blockers.

The 2004 Canadian Hypertension Education Program guidelines recommended that β-blockers be among 5 first-line drugs for initial therapy. The 2005 and 2006 guidelines were similar, but said β-blockers should be used only for those younger than 60.

So after 42 years of using these agents, what do we know about the effects of β-blockers on the M&M associated with hypertension?

• Studies of patients 60 and older have not shown a benefit.

• Studies using atenolol have not shown a benefit.

• Studies of patients younger than 60 have not shown a benefit in individual cardiovascular end points.

• If one retrospectively searches for statistically significant differences, a 0.5% absolute reduction in a composite cardiovascular end point is possible in patients younger than 60.

• A recent meta-analysis2 showed that β-blockers had less effect on some end points (particularly stroke) than other antihypertensive agents did.

In my opinion, these data do not justify a first-line recommendation for β-blockers in hypertension. But so as to not be roundly criticized for throwing out the baby with the bath water, I strongly believe that for younger, male, non-smoking, hypertensive patients, who wish to reduce their cardiovascular risk by upwards of 0.5% and who have been shown to be intolerant (when appropriately dosed) of every other available first-line class of antihypertensives, β-blockers would clearly be a solid first-line choice (except for atenolol, that is).

• Copyright© the College of Family Physicians of Canada