GLP-1 and Heart Health: What the SELECT Trial Showed About Cardiovascular Risk

For most of the past decade, GLP-1 medications were discussed almost entirely in terms of the number on the scale. That changed in November 2023 when the SELECT trial was published in the New England Journal of Medicine. SELECT was the first large trial to test whether a GLP-1 medication could prevent heart attacks, strokes and cardiovascular deaths in people who had heart disease but did not have diabetes. The answer was yes, by a margin that surprised many cardiologists, and it arrived faster than weight loss alone could explain.

The result reshaped how regulators, insurers and clinicians think about this drug class. In March 2024 the FDA added a cardiovascular risk reduction indication to semaglutide 2.4 mg, the first time a weight management medication had been approved to prevent heart events. Since then, two more trials have extended the story to kidney disease and to a form of heart failure that had almost no effective treatments.

This article explains what SELECT measured, what the numbers actually mean for an individual, why the benefit appeared so early, and where the evidence is still thin. It is educational content about published research and not medical advice. If you have heart disease and are considering a GLP-1 medication, the conversation belongs with the clinician managing your cardiovascular care.

SELECT tested whether semaglutide 2.4 mg once weekly could prevent major cardiovascular events in people who already had heart disease and carried excess weight but did not have diabetes. That last condition was the point. Every previous cardiovascular outcome trial for a GLP-1 medication had been run in people with type 2 diabetes, so nobody knew whether the heart benefit depended on lowering blood sugar or was a property of the drug itself.

The trial, led by A. Michael Lincoff and published in the New England Journal of Medicine on November 11, 2023, enrolled 17,604 participants across 41 countries. To qualify, a person had to be at least 45 years old, have a BMI of 27 or higher, and have a documented history of heart attack, stroke or symptomatic peripheral artery disease. Anyone with an HbA1c of 6.5 percent or higher was excluded. The average participant was 61.6 years old with a BMI of 33.3, and about 72 percent were men.

Participants were randomized to semaglutide or placebo on top of whatever cardiovascular medications they were already taking, which for most meant statins, antiplatelet drugs and blood pressure medication. Semaglutide was titrated up over 16 weeks to the 2.4 mg maintenance dose, the same schedule used in the STEP weight loss trials. Mean exposure to the study drug was 34.2 months and mean follow up was 39.8 months.

The primary endpoint was a composite called three point MACE: death from cardiovascular causes, nonfatal myocardial infarction, or nonfatal stroke. The trial was designed as an event driven study, meaning it ran until 1,225 primary events had occurred, which gave it enough statistical power to detect a 17 percent difference between arms.

Semaglutide reduced the combined rate of cardiovascular death, heart attack and stroke by 20 percent in relative terms and by 1.5 percentage points in absolute terms over roughly three years. A primary endpoint event occurred in 569 of 8,803 participants on semaglutide (6.5 percent) compared with 701 of 8,801 on placebo (8.0 percent). The hazard ratio was 0.80 with a 95 percent confidence interval of 0.72 to 0.90, and the result was highly statistically significant (P less than 0.001).

Each component of the composite moved in the same direction. Cardiovascular death occurred in 2.5 percent of the semaglutide group and 3.0 percent of the placebo group (hazard ratio 0.85). Nonfatal heart attack occurred in 2.7 percent versus 3.7 percent (hazard ratio 0.72), the largest single contributor to the overall benefit. Nonfatal stroke occurred in 1.7 percent versus 1.9 percent (hazard ratio 0.93), a smaller and not individually significant difference. Death from any cause was 4.3 percent on semaglutide and 5.2 percent on placebo (hazard ratio 0.81).

The absolute numbers translate into a number needed to treat of about 67. In plain terms, treating 67 people like the SELECT participants with semaglutide for about three years would be expected to prevent one heart attack, stroke or cardiovascular death. For comparison, the number needed to treat for a moderate intensity statin in secondary prevention over five years is in a similar range, which is part of why cardiologists took notice.

A secondary composite of heart failure outcomes (cardiovascular death or hospitalization or urgent visit for heart failure) was also reduced, with a hazard ratio of 0.82. This was notable because SELECT was not designed as a heart failure trial and did not select for heart failure patients.

The cardiovascular benefit in SELECT emerged within the first few months of treatment, at a point when participants had lost only a fraction of their eventual weight, which strongly suggests the mechanism is not simply a consequence of being lighter. The Kaplan Meier curves for the primary endpoint began to separate almost immediately after randomization and continued to diverge steadily throughout the trial.

Weight loss in SELECT was substantial but slower than the event curves. Participants on semaglutide lost an average of 9.4 percent of body weight by week 104, compared with 0.9 percent on placebo. Yet at month 6, when only a portion of that weight had come off, the event rate difference was already visible. If weight loss were the sole driver, the curves would be expected to run together for a year or more before separating.

A 2024 analysis presented by John Deanfield and colleagues at the European Congress on Obesity examined this directly. The researchers divided SELECT participants by baseline BMI and by the amount of weight they lost in the first 20 weeks. The cardiovascular benefit was consistent across all baseline BMI categories, including participants with a BMI between 27 and 30 who would not be classified as having obesity. It was also similar in people who lost more than 5 percent of their body weight early and in those who lost less than 5 percent. In statistical language, weight change mediated only a small fraction of the treatment effect.

Several weight independent mechanisms are plausible and were measured in the trial. Semaglutide reduced high sensitivity C reactive protein, a circulating marker of vascular inflammation, by roughly 38 percent relative to placebo, an effect that began within weeks. Systolic blood pressure fell by about 3 to 4 mmHg more than placebo. Triglycerides, HbA1c and waist circumference all improved. Preclinical research has also shown that GLP-1 receptors are expressed in the heart, blood vessels and immune cells, and that GLP-1 receptor agonists reduce plaque inflammation and improve endothelial function in animal models. The likeliest explanation is that several of these effects act together rather than any single one dominating.

Earlier cardiovascular outcome trials in people with type 2 diabetes had already shown that some GLP-1 receptor agonists reduce major cardiovascular events, which is why SELECT was designed as a superiority trial rather than a safety study. The two most cited are LEADER and SUSTAIN-6, both published in the New England Journal of Medicine in 2016.

LEADER enrolled 9,340 people with type 2 diabetes at high cardiovascular risk and randomized them to liraglutide 1.8 mg daily or placebo for a median of 3.8 years. Three point MACE occurred in 13.0 percent of the liraglutide group and 14.9 percent of the placebo group, a hazard ratio of 0.87. Cardiovascular death fell by 22 percent and all cause death by 15 percent. LEADER was the first trial to show that a GLP-1 medication could prevent cardiovascular events rather than merely not cause them.

SUSTAIN-6 tested semaglutide at 0.5 mg or 1.0 mg weekly, the diabetes doses, in 3,297 participants over about two years. MACE occurred in 6.6 percent of the semaglutide group and 8.9 percent of the placebo group, a hazard ratio of 0.74, driven mostly by a 39 percent reduction in nonfatal stroke. The trial was smaller and originally designed to rule out harm, so the superiority finding was treated as hypothesis generating until SELECT confirmed it in a much larger population.

Other agents in the class have produced mixed results. REWIND (dulaglutide, 2019) showed a 12 percent MACE reduction. EXSCEL (exenatide once weekly, 2017) and ELIXA (lixisenatide, 2015) showed neutral results, meaning no benefit but no harm. Meta analyses of the class as a whole estimate a roughly 14 percent reduction in MACE, with the long acting agents driving most of the effect.

For tirzepatide, the dual GIP and GLP-1 agonist, the dedicated cardiovascular outcome trial is SURPASS-CVOT, which compared tirzepatide head to head against dulaglutide rather than against placebo in more than 13,000 people with type 2 diabetes and established heart disease. Results reported in 2025 showed tirzepatide met the noninferiority threshold for MACE with a hazard ratio of 0.92, and all cause mortality was lower on tirzepatide. Because dulaglutide itself reduces cardiovascular events, a placebo controlled estimate for tirzepatide is not yet available.

The FLOW trial showed that semaglutide 1.0 mg weekly reduced the risk of major kidney outcomes by 24 percent in people with type 2 diabetes and chronic kidney disease, and it was stopped early because the benefit was so clear. Published in the New England Journal of Medicine in May 2024 by Vlado Perkovic and colleagues, FLOW enrolled 3,533 participants with an estimated glomerular filtration rate between 25 and 75 and elevated urinary albumin.

The primary outcome was a composite of kidney failure requiring dialysis or transplant, a sustained 50 percent or greater decline in eGFR, or death from kidney or cardiovascular causes. Over a median of 3.4 years, the primary outcome occurred in 18.7 percent of the semaglutide group and 23.2 percent of the placebo group, a hazard ratio of 0.76. The annual rate of eGFR decline was slower by 1.16 mL per minute per 1.73 square meters on semaglutide. Cardiovascular death fell by 29 percent and all cause death by 20 percent.

Kidney disease and heart disease share much of their biology, including inflammation, endothelial dysfunction and high blood pressure, so a drug that improves one would be expected to help the other. FLOW confirmed that expectation in a population where kidney function was already impaired, and it led to an FDA indication for semaglutide 1.0 mg in chronic kidney disease with type 2 diabetes in January 2025.

STEP-HFpEF showed that semaglutide 2.4 mg substantially improved symptoms, physical function and quality of life in people with obesity related heart failure with preserved ejection fraction, a condition that previously had very few effective treatments. The trial, led by Mikhail Kosiborod and published in the New England Journal of Medicine in August 2023, enrolled 529 participants with a BMI of 30 or higher and HFpEF but no diabetes.

The primary endpoints were change in the Kansas City Cardiomyopathy Questionnaire Clinical Summary Score (KCCQ-CSS), a 0 to 100 measure of heart failure symptoms and limitations, and change in body weight over 52 weeks. KCCQ-CSS improved by 16.6 points on semaglutide versus 8.7 points on placebo, a difference of 7.8 points, which is roughly three times the threshold clinicians consider meaningful for an individual. Body weight fell by 13.3 percent versus 2.6 percent. Six minute walk distance improved by 21.5 meters more than placebo, and C reactive protein fell by 43 percent relative to placebo.

A companion trial, STEP-HFpEF DM, repeated the design in 616 participants who also had type 2 diabetes and found a similar KCCQ improvement of 7.3 points. Together with the heart failure signal in SELECT, these trials have moved GLP-1 medications into heart failure treatment guidelines for the first time.

The cardiovascular safety profile in SELECT was reassuring on the outcomes that matter most, with fewer serious adverse events on semaglutide than on placebo, but the trial did confirm the known gastrointestinal burden and a small increase in resting heart rate. Serious adverse events occurred in 33.4 percent of the semaglutide group and 36.4 percent of the placebo group, a difference driven partly by fewer cardiac events.

Adverse events leading to permanent discontinuation were more common on semaglutide: 16.6 percent versus 8.2 percent. Gastrointestinal disorders accounted for the bulk of that gap, at 10.0 percent versus 2.0 percent. This is consistent with the STEP weight loss trials and follows the same pattern, with most events concentrated during dose escalation.

Resting heart rate increased by about 3 beats per minute on semaglutide, an effect seen across the GLP-1 class. In LEADER and SELECT this increase did not translate into more arrhythmias or worse outcomes, but it is monitored. Gallbladder related disorders were slightly more frequent on semaglutide (2.8 percent versus 2.3 percent), as they are in any intervention that produces rapid weight loss. Rates of pancreatitis, malignancy and acute kidney injury did not differ between groups.

The trial did not enroll people with heart failure with reduced ejection fraction as a target population, people under 45, or people with a BMI below 27, so the findings should not be extrapolated to those groups. It also does not tell us whether the benefit persists after stopping the medication, a question that matters given how common discontinuation is in practice.

SELECT answered a question that had hung over GLP-1 medications for a decade: the heart benefit is real, it does not depend on having diabetes, and it does not depend on how much weight you lose. In 17,604 people with established heart disease, semaglutide 2.4 mg reduced heart attacks, strokes and cardiovascular deaths by 20 percent over about three years, with the curves separating within months. FLOW then showed a 24 percent reduction in serious kidney outcomes, and STEP-HFpEF showed meaningful relief in a form of heart failure that had few options.

The evidence has limits worth keeping in view. SELECT enrolled people who already had cardiovascular disease and a BMI of 27 or higher, so it says nothing directly about primary prevention in people with a healthy heart, and it cannot tell you what happens after stopping. Tirzepatide has not yet been tested against placebo for cardiovascular outcomes. Gastrointestinal side effects still drove twice as many discontinuations as placebo. If you have heart disease and carry extra weight, though, the conversation with your cardiologist about a GLP-1 medication now rests on outcome data rather than on the hope that a lighter body will translate into a safer heart.

This article is educational and reflects published trial data, not medical advice. Decisions about starting, stopping or changing any cardiovascular or weight management medication belong with your prescribing clinician.