After a Heart Attack With Normal EF, Beta-Blockers Don't Prevent the Next One

Beta-blockers after myocardial infarction have been a cornerstone of cardiology for more than 40 years. The rationale has always been sound: slow the heart, reduce oxygen demand, prevent dangerous arrhythmias, and stabilize the ventricle during the vulnerable weeks after an ischemic event. The drugs do save lives — but that evidence comes from trials conducted before the era of routine percutaneous coronary intervention (PCI), high-potency statins, dual antiplatelet therapy, and modern antithrombotic regimens. The question REDUCE-AMI asked is whether the benefit still holds when every other element of post-MI care has been modernized. The answer, for patients with preserved ejection fraction, is no[1].

The trial enrolled 5,020 adults across Swedish hospitals between 2017 and 2023[1]. All participants had acute MI, underwent successful PCI, and had a left ventricular ejection fraction (LVEF) of 50% or higher — the "preserved" range. They were randomized to metoprolol succinate at a target dose of 100 mg daily or no beta-blocker. The primary endpoint was a composite of all-cause death or new MI followed for a median of 3.5 years. This was not a tiny pilot — it was powered to detect a 25% relative reduction in the primary outcome if beta-blockers provided benefit.

The event rates were nearly identical: 7.9% in the beta-blocker group versus 8.3% in the control group — a hazard ratio of 0.96 (95% CI 0.79–1.16; p=0.64[1]). That is a null result with confidence intervals tight enough to exclude any clinically meaningful benefit. All-cause death alone: 4.0% vs 4.1%. New MI alone: 4.5% vs 4.7%. No difference anywhere. The only notable findings were that beta-blocker patients had slightly more shortness of breath (12.9% vs 11.4%) and dizziness than controls. They derived no benefit and had marginally more symptoms.

The critical distinction here is ejection fraction. This trial enrolled patients with EF ≥50% — meaning the heart was not weakened by the infarction. Patients with reduced EF (below 40%) still have strong evidence supporting beta-blockers, primarily from older trials and the separate mechanism of preventing ventricular remodeling. What REDUCE-AMI shows is that among the substantial proportion of MI patients who leave the hospital with a normal squeeze — perhaps a third to a half of all MI patients — the long-standing reflex to prescribe a beta-blocker may add side effects without preventing the next cardiac event[1].

Clinical Context

The beta-blocker question matters beyond post-MI care. These drugs are used across cardiology and internal medicine — for heart failure with reduced EF, for rate control in atrial fibrillation, for hypertension, and for anxiety-related symptoms. In the post-MI setting, their use has been driven partly by inertia: once shown to work in the pre-PCI era, they became embedded in guidelines and discharge checklists. REDUCE-AMI[1] is one of several recent trials (including DANAMI-3 and parts of the SWEDEHEART registry) that suggest preserved-EF MI patients may not need them.

What clinicians should carry forward: for patients with MI and reduced EF, beta-blockers remain indicated. For those with EF ≥50% and no other indication (no heart failure, no AFib, no symptomatic hypertension), the evidence for routine post-MI beta-blocker use is now genuinely questionable. This does not mean stopping beta-blockers in patients who are tolerating them well — that evidence doesn't exist either. But it does mean a rethink before reflexively initiating them at discharge in every MI patient.

Limitations

REDUCE-AMI enrolled only patients who underwent successful PCI — it tells us nothing about MI managed medically or with CABG. The trial was conducted in Sweden, where baseline care quality is high and follow-up robust; results may differ in settings with less rigorous medication adherence or follow-up. The open-label design means both patients and physicians knew the assignment, which could affect symptom reporting. Patients with reduced EF were excluded, so this result must not be extrapolated to HFrEF.