Almost everyone who has felt their heart flutter after a strong coffee has drawn the obvious conclusion. Doctors have drawn it too: advising patients with atrial fibrillation or palpitations to cut out caffeine has been routine for decades, on the reasonable grounds that caffeine is a stimulant and stimulants and heart rhythm seem like a bad combination. It is a good example of how a mechanism that sounds obvious can survive for years without anyone testing whether it is true. When it was finally tested properly, the result went the other way.
What the words mean
A palpitation is the sensation of your own heartbeat, felt as a thump, flutter or skipped beat. Most palpitations are not dangerous. Many are extra beats, known as premature atrial or ventricular complexes, which almost everyone has and most people never notice.
Atrial fibrillation is a different thing: a sustained, disorganised electrical rhythm in the upper chambers of the heart, which makes the pulse irregular, reduces the heart's efficiency and raises the risk of stroke. It is the most common sustained arrhythmia, it becomes more common with age, and it often recurs after being corrected. Cardioversion, a controlled electrical shock, is one way of restoring normal rhythm, and the question of what makes it come back afterwards is where the most interesting coffee research sits.
The trial that tested the advice
In January 2026, Wong and colleagues[1] published the DECAF randomized clinical trial in JAMA, the first randomised test of caffeinated coffee in patients with atrial fibrillation. They enrolled 200 current or recent coffee drinkers with persistent atrial fibrillation, or atrial flutter with a history of atrial fibrillation, all scheduled for electrical cardioversion, across five hospitals in the United States, Canada and Australia between November 2021 and December 2024. Participants had a mean age of 69 and 71% were male.
Half were randomised to drink at least one cup of caffeinated coffee daily for 6 months. The other half were asked to abstain completely from caffeinated and decaffeinated coffee and from other caffeine-containing products. Baseline intake in both groups was 7 cups a week. During follow-up the coffee group drank a median of 7 cups a week and the abstinence group 0, so the two arms really did differ.
The primary outcome was clinically detected recurrence of atrial fibrillation or flutter over 6 months. It occurred in 47% of the coffee group and 64% of the abstinence group, a hazard ratio of 0.61, meaning a 39% lower hazard of recurrence in the people who kept drinking coffee. The benefit was comparable when only atrial fibrillation was counted. Adverse events did not differ significantly between groups.
How to read that result honestly
This is a randomised trial, which puts it above everything else in this article, and it is the only one of its kind. It should still be read with its limits in view. It was open-label, meaning everyone knew which group they were in, and in a trial whose primary outcome is clinically detected recurrence, awareness can influence when someone seeks care. It enrolled 200 people, which is small for a cardiology trial. Everyone in it was already a coffee drinker, so it does not tell you what happens if a non-drinker takes up coffee. And it ran for 6 months in a specific clinical situation, after cardioversion, not across the general population.
What it does do is directly test the advice that was being given. Telling coffee-drinking patients with atrial fibrillation to give up caffeine did not reduce recurrence, and in this trial the group given that instruction did worse. The authors were careful in how they framed it, and so is this article: this is one trial, and it is a reason to stop repeating the old advice reflexively rather than a reason to prescribe coffee.
What the large observational data showed first
The trial did not come out of nowhere. Kim and colleagues[2] analysed 386,258 UK Biobank participants with a mean age of 56, followed for a mean of 4.5 years, during which 16,979 developed an incident arrhythmia. After adjustment for demographics, comorbid conditions and lifestyle, each additional cup of habitual coffee was associated with a 3% lower risk of any arrhythmia. The same size of association appeared for atrial fibrillation or flutter specifically and for supraventricular tachycardia.
That study then did something most cannot. It tested whether genetic differences in caffeine metabolism changed the picture, using both a polygenic score of 7 variants and the CYP1A2 variant rs762551 alone, and found no evidence that they did. A Mendelian randomisation analysis using those same variants found no association between genetically predicted caffeine metabolism and arrhythmia risk at all. In other words, the observational data show coffee drinkers having slightly fewer arrhythmias, and the genetic analysis suggests caffeine is not causing that, in either direction.
A separate Mendelian randomisation study by Yuan and Larsson[3] reached the same conclusion specifically for atrial fibrillation: no association between genetically predicted coffee consumption and risk.
The dose-response picture
Cao and colleagues[4] pooled prospective studies in an updated dose-response meta-analysis of coffee and atrial fibrillation risk. Across the literature, higher coffee consumption was not associated with an increased risk of atrial fibrillation. This matters because it covers ordinary populations over long periods, which neither the DECAF trial nor a 4.5 year cohort can.
The pattern across all of it is consistent, which is unusual for coffee research: a randomised trial, a very large cohort, two Mendelian randomisation analyses and a dose-response meta-analysis all failing to find the harm that the standard advice assumed.
What the cardiology bodies now say
Marcus and colleagues[5], in a scientific statement from the American Heart Association on caffeine and cardiovascular disease, synthesise this literature for clinicians. Statements like this matter because they are what filters into practice, and the older reflex to advise caffeine abstinence in arrhythmia patients was never based on trial evidence in the first place.
It is also worth separating two claims that often get mixed. Nothing here says caffeine has no acute cardiac effects. The randomised crossover work in the CRAVE trial, covered in its own article, found that drinking coffee was associated with more premature ventricular contractions, the extra beats many people feel, while being associated with more daily steps and less sleep. Feeling extra beats after coffee is a real phenomenon. The question this article answers is a different one: whether that translates into sustained arrhythmia or worse outcomes, and the evidence says it does not appear to.
So why do palpitations feel worse after coffee?
Several explanations fit without requiring coffee to cause arrhythmia. Caffeine raises alertness and can raise anxiety, and heightened body awareness makes ordinary extra beats noticeable, an effect explored in our article on caffeine and anxiety. Caffeine produces a short-term rise in blood pressure and heart rate in people who are not tolerant to it. Coffee is often drunk in situations, stress, deadlines, poor sleep, short nights, that independently provoke palpitations, and poor sleep is itself associated with arrhythmia.
None of that makes the sensation imaginary. It means the sensation and the disease are not the same question, and the studies above only answer the second one.
How much caffeine is actually in the cup
Part of why people reach different conclusions about their own palpitations is that a cup of coffee is not a fixed dose. McCusker and colleagues[6] measured specialty coffees and found 58 to 259 mg of caffeine per serving, with a 16-ounce brewed coffee averaging 188 mg. That is a fourfold range between one shop's cup and another's, before you count how many you have had.
This matters for anyone trying to work out their own threshold. If a coffee one morning is 60 mg and the next is 250 mg, the difference in how you feel is not mysterious, and it is not evidence that coffee is unpredictable so much as evidence that the dose was. Someone tracking whether coffee triggers their palpitations is better off noting what they drank and where than counting cups, and better off testing a smaller or single-origin filter coffee against a double-shot drink than giving up entirely on the first bad morning.
The sleep confounder nobody controls for
There is one more reason to be careful attributing palpitations to caffeine directly. Coffee affects sleep, as covered in our article on caffeine and sleep, and short or fragmented sleep is itself associated with arrhythmia. A late afternoon coffee that costs you an hour of sleep and a morning of flutters has produced its effect through two steps, not one, and cutting the coffee and improving the sleep are not the same intervention even when they look identical from the outside.
None of the studies here were designed to separate those paths, and the DECAF trial in particular asked its abstinence group to give up all caffeine, which would have changed their sleep as well as their stimulant intake. That is a limitation of the evidence and also a hint about where to look if coffee seems to be the problem for you.
Who this does not apply to
The DECAF trial enrolled adults with persistent atrial fibrillation undergoing cardioversion. It did not include people with other arrhythmias, people with severe heart failure, or people who do not drink coffee. Energy drinks are also a different exposure: they deliver caffeine in larger, faster doses alongside other ingredients, and none of the studies here tested them. Anyone on a medicine whose handling is affected by caffeine, or whose condition has led a cardiologist to give specific advice, has a reason to follow that advice that does not come from population data. Medication interactions are covered separately in our article on caffeine and medicines.
Reading this from the Philippines
Atrial fibrillation is diagnosed less often in the Philippines than in the European and North American populations these studies drew on, which reflects differences in age structure and in how often people are screened as much as anything about hearts. That has one practical consequence for a reader here: an irregular pulse that is never checked is not a reassuring absence of disease. The coffee question in this article is a small one next to that. Nothing in this research suggests Filipino coffee drinkers should treat their daily cup as a cardiac risk, and nothing in it substitutes for having a persistently irregular pulse looked at.
The limits of this evidence
One open-label randomised trial of 200 people, however well conducted, is one trial, and its finding of benefit from coffee should be treated as a strong argument against the old advice rather than as proof that coffee prevents recurrence. The cohort evidence is observational and subject to the usual problem that people who feel palpitations may already have cut their coffee, which would make coffee drinkers look healthier. Mendelian randomisation avoids that but rests on assumptions about the genetic variants used. Most of these participants were middle-aged or older adults of European ancestry.
The practical answer
On the current evidence, coffee does not appear to cause atrial fibrillation, and in the one randomised trial to test it, coffee-drinking patients told to keep drinking coffee had fewer recurrences than those told to abstain. Habitual coffee drinking is associated with slightly fewer arrhythmias in very large cohorts, and genetic analyses find no causal link. Feeling palpitations after coffee is common and real, and if a particular drink reliably does that to you, drinking less of it is a sensible response to your own experience. What the evidence no longer supports is a blanket instruction to every patient with a rhythm problem to give up coffee. That is a conversation to have with your cardiologist, not a decision to take from an article. See also the CRAVE trial article and our article on coffee and blood pressure. Nothing here is medical advice.