Met-HeFT trial finds no cardiovascular benefit from metformin

Metformin did not significantly affect the risk of cardiovascular outcomes in patients with heart failure with reduced ejection fraction and diabetes or at risk of diabetes. These were the main findings of the Met-HeFT trial and a meta-analysis presented in a Hot Line session today at ESC Congress 2026.

Metformin is the most commonly prescribed oral antidiabetic medication worldwide. Type 2 diabetes and heart failure often co-exist and many patients with both conditions are prescribed metformin. However, as Principal Investigator of the Met-HeFT trial, Professor Henrik Wiggers from Aarhus University Hospital, Denmark, explained, there is a lack of randomized trial evidence on the use of metformin in patients with heart failure: "Observational studies and small, short-term trials suggest that metformin may have cardioprotective effects beyond glucose lowering in patients with heart failure but whether metformin improves cardiovascular outcomes has not been tested in a large long-term randomized trial," he noted.

The investigator-initiated, double-blind Met-HeFT trial was conducted in 23 centers in Denmark as part of DANHEART. DANHEART had a factorial design, studying metformin in patients with chronic heart failure and diabetes or prediabetes (Met-HeFT trial) and hydralazine–isosorbide dinitrate in patients with chronic heart failure (H-HeFT trial; also presented at ESC Congress 2026).

In Met-HeFT, eligible patients had symptomatic chronic heart failure, a left ventricular ejection fraction of up to 40% and were required to have type 2 diabetes, prediabetes or be at increased risk of developing type 2 diabetes. A total of 940 participants were randomised (1:1) to metformin or placebo. The mean age was 70 years and 16% were women.

Over a mean follow-up of 3.7 years, there was no significant difference between metformin and placebo for the primary endpoint of death, worsening heart failure, acute myocardial infarction or stroke (25.1% vs. 23.4%; hazard ratio 1.10; 95% confidence interval 0.84 to 1.42; p=0.48).

Metformin treatment did not significantly affect secondary endpoints such as all-cause death, unplanned heart failure events, new-onset diabetes or change in the heart strain marker, NT-proBNP.

As a limitation of the trial, Professor Wiggers noted that only around 10% of participants had type 2 diabetes. "Our lack of enrolment of patients with type 2 diabetes likely reflected the existing widespread use of metformin and the reluctance of patients and clinicians to discontinue ongoing metformin treatment, which was a requirement for inclusion," he said. However, he pointed out that many of the patients had prediabetes and insulin resistance.

Also at the congress, Associate Professor Anders Hostrup Larsen from Goedstrup Hospital, Herning, Denmark, presented a meta-analysis of randomized placebo-controlled trials of metformin in patients with established heart failure, which included data from Met-HeFT, and in patients with ischaemic heart disease.2 Three heart failure trials (involving 1,077 patients) and four ischaemic heart disease trials (involving 1,123 patients) were analyzed.

The meta-analysis came to the same conclusion as the Met-HeFT trial. There was no significant difference in cardiovascular outcomes with metformin in patients with heart failure and a reduced ejection fraction and we also showed no significant effect in established ischaemic heart disease."

Associate Professor Anders Hostrup Larsen, Goedstrup Hospital, Herning, Denmark

Summing up the evidence, Professor Wiggers said: "Metformin was not associated with any harm and it still remains beneficial in terms of glucose lowering, but our analyses indicate no significant evidence for any independent cardioprotective benefits."