Beta-Blockade After MI: No Practice Should Be Set in Stone

Benefit Limited in Reperfusion Era

John M. Mandrola, MD


September 13, 2018

A patient presented soon after having chest pain. The electrocardiogram (ECG) showed an ST-elevation myocardial infarction (STEMI). The day after acute percutaneous coronary intervention (PCI), cardiac enzymes met the definition of myocardial infarction (MI) but the echocardiogram showed a normal left ventricular ejection fraction (LVEF).

Should this person be treated with a β-blocker?

The fast-thinking answer is…yes, of course. High-level medical evidence has shown that β-blockers reduce the rate of death after MI. European and North American guidelines give β-blockers a class I indication. Regulators consider use of these drugs after MI as a marker of quality care.

Here is the problem: The "E" in "EBM," or evidence-based medicine, changes over time.

The trials showing benefits from β-blockers[1,2] were performed in an era when we treated patients who had MI with bed rest, morphine, and aspirin. In other words, before there was reperfusion, we let people have heart attacks.

I'm old enough to remember this era; people died of complications of MI—myocardial rupture, ventricular septal defects, refractory ventricular tachycardia—that were mediated by catecholamine excess. You can see why blunting the effects of catecholamines with β-blockade led to better outcomes.

Now that reperfusion therapy prevents or severely limits heart muscle damage in most patients with acute MI, it's possible that one of cardiology's most accepted quality measures—β-blocker use after MI—is no longer valid.

New Meta-analysis

Indeed, that is what a recent systematic review and meta-analysis suggests.[3] A Norwegian-led research team looked at the effect of oral β-blockers on mortality in contemporary patients with MI and found little evidence of an association with reduced all-cause mortality.

This was a meta-analysis of 16 observational studies that were published between 2000 and 2017 and included more than 164,000 patients. The authors included studies of patients with both STEMI and non-ST-elevation MI (NSTEMI), and the majority of patients had preserved LVEF and no signs of heart failure.

The pooled analysis of all studies did in fact show a 26% relative reduction in all-cause death with use of β-blockers, but there was moderate heterogeneity between the studies. This means that some studies showed benefit while others did not. Such variation requires further analysis.

When the researchers looked further, they found that the beneficial association was driven by small studies, which are more susceptible to bias. When they controlled for small-study bias, the association no longer met statistical significance (adjusted RR, 0.90; 95% confidence interval [CI], 0.77 - 1.04). They concluded that this evidence suggested no association between β-blockers and reduced all-cause mortality, and they called for new clinical trials.

The authors made clear the limitations of this analysis. Greg Fonarow, MD, from University of California Los Angeles, also emphasized the limits of this study on Twitter and cautioned that "highly confounded observational data and a meta-analysis of observational data are not a valid basis for revising care/guidelines based on RCT evidence." He did agree with the authors that it is reasonable for new randomized controlled trials (RCTs) to retest β-blockers for this indication.

I concur with Fonarow that one observational study should not upend a major quality measure. But it is not just one study. Others show the same thing.

Two Additional Analyses

Chinese authors reviewed 10 observational studies of oral β-blocker use and outcomes after acute MI in patients who had undergone PCI.[4] The included studies were mostly different from those in the Norwegian analyses. Although the adjusted pooled estimate of all studies revealed a 24% reduced risk for death in patients who took β-blockers, the association was not seen in small studies (sample size ≤ 1000) or in studies with patients who had preserved LVEF. Additional subanalyses found that the benefit of β-blockers did not meet statistical significance in studies that included only patients with STEMI, or those in which the prescription rates for angiotensin-converting enzyme inhibitors/angiotensin-receptor blockers was above 50% or over 90% for antiplatelets.

Even stronger evidence questioning the benefit of β-blockers after MI comes from a meta-analysis of 60 randomized clinical trials with more than 100,000 patients. Sripal Bangalore, MD, from New York University and a group of colleagues reported that β-blockers reduced mortality in the pre-reperfusion era  (incident rate ratio [IRR], 0.86; 95% CI, 0.79 - 0.94) but not in the reperfusion era (IRR, 0.98; 95% CI, 0.92 - 1.05).[5] Importantly, for this meta-analysis, trials enrolling post-MI patients with LV dysfunction or heart failure, such as CAPRICORN,[6] were excluded because β-blockers have been shown beneficial in this group.


I concede: This revelation seems hard to fathom because benefit from β-blockers in patients with coronary disease has been ingrained in us.

I chose this topic because it has specific and wide-ranging implications.

On the matter of treating patients with MI who have been fortunate enough to have reperfusion and preserved heart function, we now have observational and RCT-level[5] evidence suggesting no benefit from use of β-blockers. Not that it is needed, but plausibility of β-blocker nonbenefit also aligns with the contemporary data.

This is good to know because not taking β-blockers means our patients can avoid potential drug side effects and incur less pill burden.[7]

The wide-ranging message relates to the use of quality measures for an established practice in a rapidly changing medical environment. I found it shocking that the cited evidence for use of post-MI β-blockers came from a long-forgotten era of cardiology.

While it is true that some quality practices remain timeless—taking a careful history, doing a good exam, communicating clearly, and interpreting ECG patterns—therapeutics clearly have to be revisited over time.

Two Other Examples of Dated Evidence

The trials showing benefit of implantable cardioverter defibrillators (ICDs) enrolled patients two decades ago.[8,9] The more recent DANISH trial[10] found no mortality benefit in a group of patients with severe LV dysfunction and heart failure. In an accompanying editorial, John J.V. McMurray, MD, from the University of Glasgow, explained the lack of ICD benefit in the overall trial population by highlighting the declining rate of death from heart disease, which is likely due to improvements in heart failure care.[11] His message was that contemporary care of heart failure has diminished the absolute benefit of the ICD. That fact, he wrote, calls us to better target ICD use in patients most likely to benefit.

In the United States, many asymptomatic patients with carotid artery stenosis undergo carotid endarterectomy surgery to prevent stroke. Evidence for this practice stems from two RCTs that found 3% to 6% absolute risk reductions in ipsilateral stroke with surgery over medical therapy.[12,13]  But these trials largely enrolled patients in the 1980s and 1990s. Medical therapy has improved. Contemporary analyses have questioned the benefit of preventive carotid surgery because of the declining rates of stroke with better medical therapy.[14,15,16]

I am not saying medical therapy has progressed to the point where β-blockers, ICDs, or carotid surgery are useless. β-Blockers help patients with LV dysfunction, ICDs improve mortality in selected patients, and carotid surgery is indicated for symptomatic disease.

But clinicians cannot be complacent. Medical science progresses; these gains call for us to question established practices and keep gathering new evidence.

Finally, clinicians are not the only healthcare professionals who should worry about the rapid pace of medical progress. Policymakers bent on measuring quality care should also pay attention to the changing "E" in "EBM." I hope these revelations  infuse much-needed humility into the quest to measure quality care.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: