Weight loss helped older atrial-fibrillation patients lose weight, but not their symptoms

Guidelines tell doctors to help patients with atrial fibrillation and obesity lose weight. The advice rests on studies in younger patients, where shedding weight reduced the burden of the arrhythmia. LOSE-AF, published in JAMA, asked whether the same holds in older adults — the group most likely to have AF, and the group in whom aggressive weight loss could do harm by precipitating frailty [s1]. The answer is a clean, useful null.

What LOSE-AF did

The trial was a parallel-group, unblinded, randomised study run at two UK hospitals between 14 November 2018 and 25 April 2025 [s1]. Roughly 1,500 people undergoing electrical cardioversion for AF were assessed, and about 500 aged 60 to 85 with a body-mass index of 27 or higher were invited; 119 consented and 118 were randomised [s1]. They were assigned to an eight-month low-calorie diet and behavioural support programme (n=59) or to usual care (n=59) [s1]. The trial is registered as NCT03713775 [s2].

The primary outcome was the change at eight months in the Atrial Fibrillation Severity Scale (AFSS) symptom-severity score, analysed by intention to treat [s1]. Participants had a mean age of 68 years (SD 6), and 33% were female [s1].

The programme worked — on weight

The diet did what a diet is supposed to do. Baseline-adjusted mean weight at eight months was 92.6 kg (SE 0.85) in the intervention group against 99.4 kg (SE 0.85) in the control group (P<.001), an estimated difference of −6.9 kg (95% CI −9.2 to −4.5) [s1]. That corresponds to a weight reduction of 9.7% versus 3.1% (P<.001) [s1]. A near-10% loss, sustained to eight months in people in their late sixties, is a real result for a behavioural programme.

But not on the arrhythmia

The point of losing the weight, though, was to ease the AF. It did not. The baseline-adjusted mean AFSS symptom-severity score at eight months was 7.9 (SE 0.84) in the intervention group and 8.9 (SE 0.84) in the control group — a between-group difference of −0.9 (95% CI −3.3 to 1.4; P=.43) [s1]. The confidence interval straddles zero, and the point estimate is small against a scale where both groups sat near the same value.

Nor was the null confined to symptoms. There were no significant treatment effects on physical performance, AF burden, cardiac imaging parameters, blood pressure, lipid profile, or the incidence of repeat cardioversion or AF ablation during follow-up [s1]. When the intervention that separated the groups so clearly on weight fails to move any of the downstream measures, the burden shifts to explaining why the earlier, younger-patient evidence does not carry over.

One plausible explanation is timing. Much of the earlier evidence tied weight loss to reversal of the structural and electrical changes that let AF establish itself. In an older patient whose arrhythmia is already persistent, and whose atrium has remodelled over years, losing weight over eight months may simply arrive too late to undo what is fixed. That would make LOSE-AF less a contradiction of the earlier work than a boundary marker on it — showing where, and in whom, the strategy stops paying off.

Reading a null honestly

The first thing to say is what the trial does not show. It does not show that weight loss is useless in AF, and it does not overturn the younger-patient trials. It is a single trial of 118 people at two centres, and it was unblinded — participants knew which arm they were in, which can colour a symptom score. With this sample size, a modest true benefit could hide inside that confidence interval.

What it does show is worth stating plainly. In older adults with overweight and persistent AF, an eight-month low-calorie programme produced significant, safe weight loss but did not affect AF symptoms, AF burden, cardiac remodelling, or the need for further rhythm-control procedures [s1]. No serious adverse events related to trial participation were reported in either group, which answers the safety half of the question the trial was built to ask — weight loss did not obviously precipitate frailty here [s1].

That the trial was investigator-led and sponsored by the University of Oxford, rather than by a company with a product to sell, is part of why the null is credible [s2]. Negative results from independent trials are the ones least likely to be buried, and the ones most likely to correct an over-generalised guideline. Extrapolating a benefit seen in 45-year-olds to 75-year-olds was always an assumption; LOSE-AF is the first randomised test of it in the older group, and the assumption did not hold.

The practical takeaway is narrow and real. Weight loss in an older patient with AF may still be worth pursuing for the many other reasons to lose excess weight — but easing the arrhythmia itself is not, on this evidence, one of the things it reliably does.

This article is informational and does not constitute medical advice.

Sources

• [s1] Weight Loss in Older Patients With Persistent Atrial Fibrillation: The LOSE-AF Randomized Clinical Trial. JAMA, 1 July 2026. https://doi.org/10.1001/jama.2026.5787
• [s2] LOSE-AF: Can Weight Loss Help Patients With Atrial Fibrillation? ClinicalTrials.gov identifier NCT03713775, US National Library of Medicine. https://clinicaltrials.gov/study/NCT03713775