Semaglutide improved measures of heart filling and movement in a small MRI study of people with obesity and heart failure with preserved ejection fraction, or HFpEF. The November 8 Nature Medicine report offers a closer look at how the heart may change during treatment. Its most important qualification is who made it into the analysis: 49 people who completed study treatment and had usable scans at both visits.
That is useful evidence about a possible mechanism. The study does not establish that those scan changes prevent hospital admissions or extend life. Nor does it show that every person starting semaglutide would experience the same changes.
The 49-person study inside a larger trial
CAMEO-SEMA enrolled adults with obesity and diagnosed HFpEF. The trial registry required BMI of at least 30 and a left ventricular ejection fraction of at least 50%, alongside specific evidence of heart failure. This was not a study of otherwise healthy adults seeking weight loss. Type 1 diabetes was excluded; type 2 diabetes was not a blanket exclusion.
The parent trial report describes 81 participants randomized to semaglutide or placebo for 52 weeks. Weekly injections were titrated toward 2.4 mg, and both groups received lifestyle counseling. That describes the research regimen, not a dosing recommendation. The trial’s main endpoint was pressure measured invasively during low-intensity exercise, rather than an MRI result.
The MRI substudy included 33 semaglutide recipients and 16 placebo recipients, with 28 women overall. Scans were performed at baseline and 52 weeks, within a day of the invasive exercise assessment. Keeping those measurements close together let investigators examine whether changes on the scans tracked changes in pressure.
What the scans add
The prespecified main MRI endpoint was left atrial reservoir strain. In plain language, this measures how the left upper chamber stretches as it receives blood. It is a particular aspect of heart mechanics, not a score for overall heart health. The published protocol and statistical analysis plan identify this as the MRI substudy’s primary endpoint, separate from the parent trial’s pressure endpoint.
The MRI abstract reports a 5.7-percentage-point greater improvement in reservoir strain with semaglutide than placebo, with a 95% confidence interval from 0.3 to 9.7 points. That interval leaves considerable uncertainty about the size of the effect. It should not be translated into “the heart worked 5.7% better.”
Another prespecified measure assessed how quickly the left ventricle filled early in its relaxation phase, relative to its end-diastolic volume. This also improved compared with placebo. Improvements in both measures were associated with reductions in exercise pulmonary artery wedge pressure, a measure of cardiac congestion.
The larger trial separately found an estimated 10.6 mmHg greater reduction in exercise wedge pressure with semaglutide than placebo. The MRI findings help connect that pressure result to changes in chamber behavior. They do not prove which change caused the other, or isolate a direct drug effect from the consequences of losing weight.
Why the missing scans matter
The study’s reporting summary says people who discontinued study medication were excluded from the MRI analysis. So were people without paired scans because of MRI contraindications, pacemakers or other study or logistical reasons. These are different reasons for missing data; an absent scan does not automatically mean treatment failed.
Nevertheless, the analysis describes a selected group who could remain on study treatment and complete imaging. The original randomization is valuable, but selecting participants afterward can change how comparable the groups are. It also limits how well the result represents everyone who starts treatment. Our guide to trial estimands explains why the people and events included in an analysis matter as much as its headline number.
The summary also states that no separate sample-size calculation was performed for the MRI substudy. It came from one center and had no independent replication cohort. Imaging analysis was blinded to treatment assignment, which is a strength, but these safeguards cannot make a 49-person analysis answer every clinical question. Our trial-reading guide covers the distinction between a measured endpoint and a broader claim.
What this changes for a care discussion
For readers with obesity-related HFpEF, the study adds a plausible explanation for benefit: heart filling and chamber mechanics may improve alongside other treatment effects. It does not establish a cure, a benefit in people without HFpEF, or superiority over another medicine. The comparison was with placebo.
Approval and safety remain separate questions. The current US Wegovy prescribing information includes weight-management uses and cardiovascular risk reduction for specified populations; it does not list HFpEF as a standalone indication. Nausea, vomiting and diarrhea are among its common adverse reactions, and its contraindications and other warnings still apply. A favorable imaging result cannot replace that assessment. The semaglutide profile provides product context.
A useful question at an appointment is which benefits are established for your diagnosed condition and which remain mechanistic research. Our medication discussion guide can help organize that conversation. This report checked the public abstracts, registry, protocol, reporting summary and label. The full journal results tables were unavailable, so it does not present a complete safety analysis or all imaging outcomes.


