Can you stop beta blockers after a heart attack?
For one narrow group of people, yes. Stopping was not worse than staying on the drug. Among 6,238 stable adults more than six months past a heart attack, a major event over the next three years happened to 17.2% of those who stopped their beta blocker and 15.9% of those who continued.
Beta blockers slow the heart and lower the force of each beat. After a heart attack they became standard, and many people simply stay on them for years. The trials that made them standard were run before modern stents, clot-preventing drugs, and cholesterol treatment changed what recovery looks like. So the question is whether the pill is still doing work, or just sitting in the weekly organizer.
This analysis pooled the raw patient-level data from two randomised trials, ABYSS in France and SMART-DECISION in South Korea. Everyone included was stable, at least six months out from a heart attack, had a left ventricular ejection fraction of 40% or higher, meaning the heart’s main pumping chamber was squeezing normally or close to it, had no heart failure, and had no other reason to be on a beta blocker.
What the data show
The main outcome bundled four things together: death from any cause, another heart attack, stroke, or a hospital stay for a heart-related reason. It happened to 532 of the 3,092 people who stopped and 500 of the 3,146 who continued. That is 17.2% against 15.9%, a gap of 1.3 percentage points, or about 13 more people per 1,000 over three years. In relative terms that is about a 9% higher risk (hazard ratio 1.09), but the plausible range ran from 3% lower to 24% higher (95% CI 0.97 to 1.24), which includes no difference at all.
The second outcome is the one I care about more, because it strips out the softer hospital admissions. Death, another heart attack, or a hospital stay for heart failure happened to 201 people in each group: 6.5% of those who stopped and 6.4% of those who continued. Relative risk was about 1% higher (hazard ratio 1.01), with a plausible range from 17% lower to 23% higher (95% CI 0.83 to 1.23). Those curves are on top of each other. The result held no matter where a person’s ejection fraction sat.
Dr. Kumar’s Take
The hard outcomes are what convince me. When you look at death, reinfarction, and heart failure admissions, the two groups produced the identical number of events, 201 and 201. That is not a subtle statistical rescue, that is two lines lying flat against each other.
The primary endpoint is weaker, and the authors say so themselves. It leaned heavily on hospitalisations, and the mix of events differed between the two trials. Hospital admission for a “cardiovascular reason” is a soft outcome. It moves with local practice, with how anxious a patient is, and with how a cardiologist reads a symptom. A drug that lowers heart rate and blood pressure will reduce some of those visits without preventing a single death.
The other thing worth naming is how the non-inferiority test was set up. Researchers decided in advance that stopping would count as “not worse” if the true increase in risk stayed under 25% for the primary endpoint and under 40% for the key secondary one. Those are generous allowances. The primary endpoint passed with less than a 2 in 100 chance of a false pass (p for non-inferiority = 0.016), and the key secondary passed with about a 6 in 10,000 chance (p = 0.0006). So the statistics are sound, but “not worse by less than 25%” is a lower bar than most people hear in the word “safe.”
How the studies were done
Pooling individual patient data is stronger than pooling published summaries. Instead of averaging two sets of conclusions, the researchers reanalysed every patient record from both trials in one model, with trial included as a random effect so that differences between the French and Korean populations did not distort the estimate. Of the 6,238 patients, 3,698 came from ABYSS and 2,540 from SMART-DECISION. Randomisation split them into 3,092 who stopped and 3,146 who continued.
These were not recent heart attack patients. The median time from the index heart attack to randomisation was 3.6 years, and half the group fell between 1.6 and 7.5 years out. Median follow-up ran 3.0 years. The review protocol was registered with PROSPERO in advance, which matters because it locks in the outcomes before anyone sees the data.
Safety, limits, and caveats
Three years is enough to catch a rebound in events but not enough to rule out a slow drift over a decade. The enrollment rules also carve out most of the people who take these drugs. Anyone with heart failure, an ejection fraction under 40%, a recent heart attack, or another reason for a beta blocker, such as an arrhythmia, angina, or blood pressure that needs the drug to stay controlled, was not in this analysis. None of this applies to them.
Stopping a beta blocker is also not a clean subtraction. Heart rate and blood pressure both tend to rise when the drug comes off, and the abstract does not report those measurements here. That alone is a reason nobody should stop one on their own.
Practical Takeaways
- If you are years past a heart attack, have a normal pumping heart, and have no heart failure, it is reasonable to ask your cardiologist whether the beta blocker is still earning its place.
- Never stop a beta blocker on your own or taper it without guidance, because the heart rate and blood pressure rebound can be significant.
- If you take a beta blocker for high blood pressure, angina, an irregular rhythm, or heart failure, this analysis does not apply to you and the drug is doing a job.
- Ask what your most recent ejection fraction was, since the entire question in this study depended on that number being 40% or higher.
Related Studies and Research
- Muscle quality, not size, predicts heart attack and death
- Shingles vaccine cuts heart attack risk nearly in half for heart disease patients
- Treating gout properly may cut heart attack and stroke risk by up to 23%
- Does rosuvastatin prevent heart attacks? A critical review of the JUPITER trial
FAQs
What does “non-inferiority” actually mean in a drug trial?
A regular trial asks whether a treatment is better than the alternative. A non-inferiority trial asks something weaker: whether it is not meaningfully worse. Researchers pick a margin in advance, and here they chose 1.25 for the primary endpoint and 1.40 for the key secondary one. Passing means the data ruled out a risk increase bigger than that margin. It does not mean the two options performed identically, and it is worth knowing the margin before you accept the word “safe.”
Who is left out of this finding?
The entry rules were strict, and that is the point. Everyone had an ejection fraction of 40% or higher, no heart failure, at least six months since the heart attack, and no other indication for a beta blocker. If you are on one for an abnormal heart rhythm, ongoing chest pain, blood pressure control, or a weakened heart muscle, you were not represented in these 6,238 patients. The drug in those settings is treating something specific, and this analysis says nothing about removing it.
Why did the authors flag their own main result as limited?
The primary endpoint counted hospital admissions for cardiovascular reasons alongside death, heart attack, and stroke. Admissions are far more common than deaths, so they dominate a combined count and can drive the whole result. The authors also noted the mix of events differed between the two trials, which means the endpoint was not measuring quite the same thing in France and in South Korea. That is why the secondary endpoint, built only from death, heart attack, and heart failure admission, carries more weight.
Bottom Line
In stable people who are years past a heart attack, whose heart still pumps normally, and who have no heart failure and no other reason to be on the drug, stopping a beta blocker met the trial’s bar for not being worse than continuing it. The hard outcomes were identical, 201 events in each group. The broader endpoint drifted 1.3 percentage points in the wrong direction but sat well inside the range of chance. This is a real argument for reviewing a long-standing prescription with a cardiologist, and it is not a reason to open the cabinet tonight and stop.
