GLP-1 drugs and heart failure: what the SOUL trial actually found

Why the heart and GLP-1 receptors are connected

GLP-1 receptors are not limited to the gut and pancreas. Cardiomyocytes have them. So do the coronary arteries. Activating those receptors reduces inflammation, lowers oxidative stress, and improves how cardiac cells handle energy. None of this was understood when the first GLP-1 drugs reached the market for blood sugar control.

The cardiac angle emerged from a regulatory requirement. In 2008, the FDA began requiring cardiovascular outcomes trials for all new diabetes drugs, after one compound in a different drug class caused unexpected cardiac harm. GLP-1 developers ran those trials expecting to show safety. They showed benefit. Semaglutide, liraglutide, and tirzepatide all reduced the rate of major cardiac events compared to placebo in high-risk diabetic populations.

Heart failure was the next question. Cardiovascular outcomes trials track heart attack and stroke well. Heart failure is harder, because there are two distinct types with different underlying mechanisms, and drugs that help one may not help the other.

What the SOUL trial tested

The SOUL trial enrolled 9,647 adults with type 2 diabetes and either established atherosclerotic cardiovascular disease, chronic kidney disease, or both. These were not average patients. They had already survived a cardiac event or were showing measurable organ damage. Half received once-daily oral semaglutide; half received placebo. The trial ran across 33 countries and followed patients for a median of 47.5 months -- close to four years.

Heart failure data came from a pre-specified subgroup analysis: patients who had existing heart failure at enrollment. Pre-specified means the analysis was committed to before the trial data was unblinded. That matters. It is a more trustworthy form of subgroup analysis than findings that emerge after researchers have already looked at the data.

MACE down 14%, heart failure events down 22%

The main result: oral semaglutide reduced MACE by 14%. MACE is the standard composite endpoint for cardiovascular trials -- non-fatal heart attack, non-fatal stroke, and cardiovascular death combined. 12.0% of the semaglutide group hit a MACE event versus 13.8% in the placebo group (HR 0.86, 95% CI 0.77 to 0.96, P=0.0028). Non-fatal heart attack drove most of the benefit, reduced by 26%.

Those numbers do not read like marginal effects. A 14% MACE reduction over four years in a very high-risk population is clinically meaningful. It is in the range of what statins achieve in comparable groups. The heart failure number is more striking: a 22% reduction in a pre-specified subgroup with an upper confidence interval that stays below 1.0.

The critical split: HFpEF benefits, HFrEF does not

The heart failure benefit is not uniform. The trial found something important when it separated patients by ejection fraction type.

Heart failure with preserved ejection fraction, or HFpEF, is the type where the heart pumps at a normal rate but the muscle has become stiff and does not fill properly. Fat infiltration, chronic inflammation, and fibrosis in and around the cardiac muscle drive it. Obesity is the biggest modifiable risk factor. About half of all heart failure cases in wealthy countries are HFpEF, and prevalence is rising with obesity rates.

Heart failure with reduced ejection fraction, or HFrEF, is different. The pump is weak. Ejection fraction falls below 40%. This typically follows a heart attack that damages heart muscle, viral myocarditis, or a genetic cardiomyopathy. It involves structural damage to the pump itself.

In the SOUL HFpEF subgroup, the hazard ratio was 0.59. That is a 41% relative risk reduction. In the HFrEF subgroup, it was 0.98, essentially flat.

This split makes biological sense. GLP-1s reduce visceral fat and systemic inflammation. Those are the primary drivers of HFpEF. The HFrEF pump failure comes from structural damage that anti-inflammatory and metabolic effects do not repair. You cannot un-scar infarcted myocardium with a GLP-1. You can reduce the fat and inflammation that is stiffening cardiac tissue in someone with obesity-driven HFpEF. The data shows exactly that pattern.

STEP-HFpEF: the dedicated heart failure trial

The SOUL subgroup pointed at HFpEF. STEP-HFpEF was built specifically to test it. The trial enrolled 529 adults with obesity and HFpEF -- patients without T2D as a required qualifier, just obesity and the HFpEF diagnosis. That design matters: it tests whether the benefit comes from GLP-1 activity directly, not just from improving glycemic control in diabetics.

Results: semaglutide improved Kansas City Cardiomyopathy Questionnaire (KCCQ) scores by nearly 8 points versus placebo. The KCCQ measures heart failure symptoms and quality of life across domains like physical limitation, social limitation, and symptom frequency. An 8-point improvement is clinically meaningful -- patients reported being able to do more and feeling better.

Body weight fell by 13.3% on average. Six-minute walk distance improved by 21.5 meters more than placebo. In a population limited by cardiac stiffness and breathlessness, walking 21 meters further in six minutes is real functional change.

In HFpEF, reducing the fat load that stiffens and inflames cardiac tissue is not incidental. It is the mechanism of treatment. The KCCQ and walk improvements followed directly from it.

Real-world data from 90,000 patients

Trials show what happens under controlled conditions. A 2025 analysis of real-world records from over 90,000 HFpEF patients with obesity and T2D showed what happens in clinical practice. Patients treated between 2018 and 2024 were compared against those on sitagliptin, a diabetes drug class with no meaningful cardiac remodeling effect.

Semaglutide users had 42% lower risk of hospitalization for heart failure or all-cause mortality. Tirzepatide users had 58% lower risk. Real-world numbers tend to run larger than trial data, because patients who stay on GLP-1s tend to be adherent and responding. The direction is consistent with the trial data, and the effect sizes suggest the trial results were not an overestimate.

The tirzepatide advantage over semaglutide in the real-world HFpEF data mirrors what appears elsewhere in head-to-head comparisons. Tirzepatide hits both GLP-1 and GIP receptors, producing greater weight loss on average. More weight loss from cardiac tissue means more decompression of a stiff heart.

What GLP-1s now cover across organ systems

Heart failure joins a growing list of organ-level indications for this drug class. The FLOW trial found semaglutide cut kidney failure risk by 24% in CKD patients -- notably, the SOUL trial enrolled many of those same patients. The ESSENCE trial found semaglutide resolved MASH in 62.9% of patients versus 34.3% on placebo. The SURMOUNT-OSA trial found tirzepatide cut sleep apnea severity by up to 63%.

The mechanism connecting all of these is not simply weight loss. GLP-1 receptors exist in the liver, kidney, heart, and lung. Activating them reduces local inflammation and shifts metabolic behavior in ways that differ by tissue. The cardiac benefit in HFpEF comes from the fat-inflammation pathway. The kidney benefit comes partly from hemodynamic changes and reduced inflammatory signaling in nephrons. Same drug class, different organs, different mechanisms, real outcomes in each.

What this means if you have heart failure now

If you have HFpEF and obesity or T2D, you are the population the SOUL and STEP-HFpEF data applies to most directly. The evidence is strong enough to bring to your cardiologist. Semaglutide is already FDA-approved for cardiovascular risk reduction in people with obesity and established CVD, which overlaps with this population. The HFpEF-specific data adds depth to that conversation.

If you have HFrEF, the picture is different. The SOUL data showed no meaningful benefit in that group. HFrEF has its own robust evidence base -- ACE inhibitors, beta-blockers, sacubitril/valsartan, SGLT2 inhibitors -- and GLP-1s are not currently indicated as treatments for it.

If you are already on a GLP-1 for obesity, T2D, or kidney disease, the SOUL data is not a reason to change anything. It is a reason to ask your doctor whether a cardiac workup or formal HFpEF assessment makes sense, particularly if you have exertional shortness of breath or swelling that has not been fully evaluated.

For context on where the drug pipeline is heading, including the next-generation triple agonist retatrutide and the amylin-semaglutide combination CagriSema, the compounds being developed to succeed current GLP-1s all produce greater weight loss -- which, if the HFpEF data holds, should translate into stronger cardiac effects. That research is underway.