Flu vaccination tracked 20% less new atrial fibrillation across 276,888 matched adults

That a bad case of flu can push a heart into trouble is established. Whether the annual shot does anything about the commonest disturbance of heart rhythm was not, and its role in preventing incident atrial fibrillation and cardiovascular outcomes in clinical practice remains uncertain.

Writing in the European Heart Journal, a team led from Tours worked through the records of 276,888 adults, half vaccinated and half not, matched against each other on 65 separate characteristics. Influenza vaccination was associated with a significantly lower risk of incident AF, by about a fifth.

The same analysis also found the vaccinated group less likely to die of anything at all. That second result is the reason to read the first one carefully.

The rhythm problem in question

Atrial fibrillation, usually shortened to AFib, is one of the most common types of arrhythmias, the family of faults in the rate or rhythm of the heartbeat.

In AFib your heart beats irregularly and sometimes much faster than normal, and the chambers fall out of step with one another. Blood that should be moved along instead pools, and clots that form in a pooling chamber travel. That is why the condition matters: it is a leading cause of stroke, and it becomes steadily more common with age.

Anything that reduced how often it starts would be worth having.

Why influenza is a plausible culprit

The link runs through the illness rather than the vaccine.

A serious respiratory infection is a whole-body event. It brings fever, inflammation, dehydration, low oxygen and a heart working harder for days at a stretch, which is a reasonable set of conditions for an unstable rhythm to establish itself. Influenza infection triggers cardiovascular events often enough that the association is not in dispute.

So the hypothesis is not exotic. Prevent the infection, prevent some share of the events it sets off, including some arrhythmias.

What the records showed

Everyone in the analysis had a routine healthcare visit in 2022. Vaccination in the previous year defined the exposed group; the comparison group had no record of one. After matching, the two groups were followed for a mean of about 2.7 years.

New atrial fibrillation appeared in 0.9% of vaccinated patients each year, against 1.1% of the unvaccinated. That is the whole effect, stated in the units a person can hold: nine cases a year in every thousand vaccinated people, eleven in every thousand who were not.

The relationship held across ages and regardless of whether someone already had heart disease, without significant effect modification.

The result that undermines the result

Vaccinated patients also had lower risks of all-cause mortality, by around 15%.

That figure should stop the reader, because no flu vaccine does that. Influenza kills a small number of people in any given year, and preventing those deaths cannot move the total death rate by a seventh across a general adult population followed for under three years.

What produces a number like that is not a vaccine. It is the difference between the sort of person who turns up for one and the sort who does not. People who get vaccinated are, on average, more engaged with medical care, more mobile, less frail, better housed and better off. Every one of those things predicts survival on its own.

Epidemiologists call it the healthy vaccinee effect, and it is one of the best-documented biases in the field. It inflates every benefit that vaccination appears to confer in observational data, and it inflates them roughly in proportion to how strongly the outcome tracks general health.

Matching on 65 variables is a serious attempt to handle it. It cannot succeed completely, because the variables that matter most here are the ones medical records do not hold: whether someone is frail, whether they have anyone to drive them to an appointment, whether they have given up on their own health.

What survives the objection

Not nothing, and this is where the article gets harder than a simple debunk.

The mortality result is almost certainly confounded. But confounding of that kind should smear itself across every outcome roughly evenly, and it did not. The reduction in atrial fibrillation is larger than the reduction in death, and considerably larger than the reduction in cardiovascular hospitalization, which barely moved.

If healthy-vaccinee bias were the whole story, the rhythm finding should have been the same size as the others. Its being the biggest of them is at least consistent with something real sitting underneath.

That is an argument for taking the hypothesis seriously, not for believing the number. A 20% reduction is the ceiling of what this design can claim, and the floor could be zero.

What it does not change

The case for a flu shot did not need this study, and does not rest on it. Getting vaccinated against the flu each year is best way to lower your risk of the flu and its complications, and that has been the reason all along.

What a result like this adds is a possible extra, in an area where nobody was looking. A randomized trial could settle it, several have already tested flu vaccination against heart outcomes in people with existing disease, and the rhythm question would be a reasonable thing to add to the next one.

Until then, this is a large, careful, thoroughly confounded piece of evidence that the annual shot may do slightly more than advertised, reported by a design that cannot prove it and has the honesty to publish the mortality figure that gives the game away.