Trial evidence

SELECT Trial Semaglutide Explained: Heart Attack and Stroke Results

Q: Does semaglutide actually prevent heart attacks, or is it just weight loss?

In SELECT, the composite of cardiovascular death, nonfatal heart attack, and nonfatal stroke was 20% lower with semaglutide (HR 0.80, 95% CI 0.72–0.90; P<0.001) [2, 5]. The benefit appeared before substantial weight loss occurred. The presentation's discussant, Ania Jastreboff, said the gains are likely a combination of direct drug effects and indirect effects of weight loss on blood pressure, lipids, and glycemia; lead author A. Michael Lincoff said the effect is probably not simply weight loss, and the mechanism is not settled.

Q: What is the difference between the 20% figure and the 1.5 percentage point figure?

They describe the same result. Major cardiovascular events occurred in 8.0% of placebo participants and 6.5% of semaglutide participants over a mean follow-up of 39.8 months [2, 4]. The 20% is the relative reduction between those rates; the 1.5 percentage points is the absolute difference.

Q: How many people have to take it to prevent one event?

The number needed to treat was 67, based on the 1.5 percentage point absolute reduction over the trial's mean follow-up of 39.8 months [2, 4]. One pharmacist commentator called that a "pretty good number".

Q: Does semaglutide help if you don't have diabetes?

SELECT is the trial that tested this directly. It enrolled 17,604 people with overweight or obesity and established cardiovascular disease but no diabetes, and it showed a 20% relative reduction in major cardiovascular events with semaglutide 2.4 mg weekly versus placebo [2, 5]. It is the only one of the major semaglutide heart trials conducted in people without diabetes.

Q: Does it work if you've never had a heart attack or stroke?

SELECT cannot answer that. Everyone enrolled had established cardiovascular disease — a prior heart attack, prior stroke, or symptomatic peripheral arterial disease — so there are no data from this trial on preventing a first event in people without existing disease.

Q: How long does the weight loss last on treatment?

In SELECT, weight loss continued for about 65 weeks and was sustained for up to four years. At 208 weeks, the semaglutide group had lost an average of 10.2% of body weight versus 1.5% with placebo (P<0.0001).

Q: How many people quit because of side effects?

Adverse events leading to permanent discontinuation of the trial product occurred in 16.6% of the semaglutide group versus 8.2% of the placebo group (P<0.001), driven mainly by gastrointestinal disorders (10.0% vs 2.0%) [2, 4]. Serious adverse events overall were less frequent with semaglutide (33.4% vs 36.4%).

Q: Was there a reduction in deaths?

Deaths from any cause were 4.3% with semaglutide versus 5.2% with placebo (HR 0.81, 95% CI 0.71–0.93). All-cause death was a hierarchically tested secondary endpoint, not the primary one, so SELECT was not designed primarily to prove a mortality benefit [2, 4].

SELECT tested 2.4 mg weekly semaglutide against placebo in 17,604 adults with overweight or obesity and existing heart disease but no diabetes. It cut major cardiovascular events 20% relative — 6.5% vs 8.0% — over about 3.3 years.

Last checked September 14, 2026·13 sources cited·Wegovy

Isometric clay scene of a heart shape, an injection pen, and a downward-trending bar graph representing a cardiovascular trial

Semaglutides.org provides information only, not medical advice. Talk to a licensed provider about your own care before making changes. Some telehealth links are affiliate links and are labeled where they appear.

The SELECT trial (Semaglutide Effects on Cardiovascular Outcomes in People with Overweight or Obesity, ClinicalTrials.gov ID NCT03574597) tested a question that had not been settled before: whether semaglutide can lower the cardiovascular risk associated with overweight and obesity in people who do not have diabetes ^[2]. It enrolled 17,604 adults at 804 sites in 41 countries, randomly assigning them to once-weekly subcutaneous semaglutide at a target dose of 2.4 mg or to placebo, with both groups staying on their usual cardiovascular medicines [2, 4]. The trial was funded by Novo Nordisk, with the protocol designed by the sponsor together with an academic steering committee ^[2].

The headline result: major cardiovascular events — death from cardiovascular causes, nonfatal heart attack, or nonfatal stroke — occurred in 6.5% of the semaglutide group (569 of 8,803 people) versus 8.0% of the placebo group (701 of 8,801), a hazard ratio of 0.80 (95% CI 0.72–0.90; P<0.001) [2, 5]. That is a 20% relative reduction, or a 1.5 percentage point absolute reduction, measured over a mean follow-up of 39.8 months [2, 4]. The full results were presented at the American Heart Association 2023 Scientific Sessions on November 11, 2023, and published in the New England Journal of Medicine on December 14, 2023 (Lincoff AM et al., 2023;389(24):2221–2232) [3, 4].

How was the SELECT trial designed?

SELECT was a multicenter, double-blind, randomized, placebo-controlled, event-driven superiority trial ^[2]. “Event-driven” means the trial was set to continue until enough heart events had accumulated to answer the question, rather than stopping at a fixed calendar date.

Participants were randomized 1:1 with no stratification, and neither the patients nor the investigators knew who received semaglutide and who received placebo ^[2]. The drug and the placebo were both added on top of standard cardiovascular care — investigators were encouraged to follow evidence-based recommendations for managing underlying heart disease [2, 4].

The design gave the trial 90% power to detect a 17% relative risk reduction (hazard ratio 0.83) at a one-sided significance level of 0.025, which required at least 1,225 primary endpoint events ^[2]. Assuming a placebo event rate of 2.2% per year and a planned duration of 59 months, the researchers estimated they would need to enroll about 17,500 people ^[2]. One interim analysis for superiority was prespecified to occur once two-thirds of the planned events had accrued ^[2]. Efficacy was analyzed on an intention-to-treat basis — everyone randomized was counted in their assigned group regardless of whether they kept taking the drug — using Cox proportional hazards models ^[2].

Who was in the trial, and who was left out?

To join SELECT, a person had to be 45 or older, have a body mass index (BMI) of 27 or higher, and have established cardiovascular disease, defined as a prior heart attack, prior stroke, or symptomatic peripheral arterial disease ^[2].

The exclusion list is just as important as the inclusion list. People could not join if they had a prior diabetes diagnosis, a screening HbA1c of 6.5% (48 mmol/mol) or higher, or treatment with any glucose-lowering medication or GLP-1 receptor agonist within the previous 90 days ^[2]. Also excluded were people with New York Heart Association class IV heart failure, end-stage kidney disease or dialysis, anyone enrolled within 60 days of a cardiovascular or neurologic event, and anyone planning coronary, carotid, or peripheral revascularization ^[2].

The 17,604 participants had a mean age of 61.6, were 72.3% men, and had a mean BMI of 33.3 kg/m², with 71.5% having obesity (BMI 30 or higher) ^[4]. Mean HbA1c at baseline was 5.8%, and about two-thirds fell in the prediabetes range of 5.7% to 6.4% ^[4]. They were already well treated for heart disease: 90.1% were on lipid-lowering medication, 86.2% on antiplatelets, 70.2% on beta-blockers, 45.0% on ACE inhibitors, and 29.5% on ARBs ^[4].

What dose did participants take?

The trial used once-weekly subcutaneous semaglutide with a target dose of 2.4 mg ^[2]. Treatment started at 0.24 mg weekly and increased every four weeks through 0.5, 1.0, and 1.7 mg, reaching 2.4 mg by week 16 ^[2]. If dose escalation caused unacceptable side effects, the escalation intervals could be extended, treatment could be paused, or a maintenance dose below 2.4 mg could be used ^[2].

The protocol required stopping the trial product for pregnancy or planned pregnancy, pancreatitis, or a calcitonin level of 100 ng/L or higher ^[2]. By two years, 77% of the semaglutide group was still on the full 2.4 mg target dose ^[4].

What were the results for heart attacks, strokes, and deaths?

The primary endpoint was the first occurrence of death from cardiovascular causes, nonfatal heart attack (myocardial infarction), or nonfatal stroke — the standard “three-point MACE” composite ^[2]. Over a mean exposure of 34.2 ± 13.7 months and mean follow-up of 39.8 ± 9.4 months, the composite favored semaglutide: 6.5% versus 8.0%, hazard ratio 0.80 (95% CI 0.72–0.90; P<0.001) ^[2].

Each piece of the composite pointed in the same direction, but only the heart attack component on its own reached statistical significance ^[4]:

Endpoint	Semaglutide	Placebo	Hazard ratio (95% CI)
Primary composite (CV death, nonfatal MI, nonfatal stroke)	6.5%	8.0%	0.80 (0.72–0.90) ^[2]
Nonfatal heart attack	2.7%	3.7%	0.72 (0.61–0.85) ^[4]
Cardiovascular death	2.5%	3.0%	0.85 (0.71–1.01) ^[4]
Nonfatal stroke	1.7%	1.9%	0.93 (0.74–1.15) ^[4]
Heart failure composite	3.4%	4.1%	0.82 (0.71–0.96) ^[4]
All-cause death	4.3%	5.2%	0.81 (0.71–0.93) ^[4]
Coronary revascularization	5.4%	6.9%	0.77 (0.68–0.87) ^[4]
Progression to diabetes (HbA1c ≥6.5%)	3.5%	12.0%	0.27 (0.24–0.31) ^[4]

The confirmatory secondary endpoints — cardiovascular death, the heart failure composite, and death from any cause — were tested in a fixed hierarchical order ^[2]. All-cause death was a secondary endpoint, not the primary one, so the trial did not set out to prove a mortality benefit; the 4.3% versus 5.2% figure is how deaths from all causes compared over the follow-up period [2, 4].

What does “20% lower risk” actually mean?

The 20% figure is a relative risk reduction. In plain terms, 8.0% of placebo-treated participants had a major cardiovascular event during the trial and 6.5% of semaglutide-treated participants did — a difference of 1.5 percentage points ^[4]. Both numbers describe the same result; they just frame it differently.

That absolute difference translates to a number needed to treat of 67, meaning about 67 people would need to be treated over the trial’s mean follow-up of 39.8 months for one fewer person to have a primary event, on average [2, 4]. One outside pharmacist commentator, Cynthia Jackevicius, PharmD, described 67 as a “pretty good number” ^[4].

How much weight did people lose, and did weight loss explain the benefit?

Weight loss continued for roughly the first 65 weeks and then was sustained for up to four years ^[5]. At 208 weeks, the semaglutide group had lost an average of 10.2% of body weight versus 1.5% with placebo (P<0.0001), with waist circumference down 7.7 cm versus 1.3 cm and waist-to-height ratio down 6.9% versus 1.0% ^[5]. Clinically meaningful weight loss occurred in both sexes and across races, body sizes, and regions ^[5]. At the AHA presentation, weight loss was reported as 9.4% versus 0.9% ^[4]. For reference, an earlier trial found semaglutide 2.4 mg for 104 weeks cut weight by 15.2% in people with overweight or obesity without diabetes ^[2].

Whether the heart benefit comes from weight loss is not settled. The reduction in cardiovascular events appeared early, before substantial weight loss had occurred ^[4]. Ania Jastreboff, MD, PhD, the discussant for the presentation, said the finding could be related to effects on inflammation, endothelial function, or other factors, and that the gains are likely due to a combination of the indirect effects of weight loss — improvements in hypertension, hyperlipidemia, and glycemic control — and direct effects of the drug ^[4]. Lead author A. Michael Lincoff said the effect is unlikely to be weight loss alone but that the mechanism is not fully worked out ^[4]. GLP-1 receptor agonists act on pathways involved in glucose metabolism, energy homeostasis, and inflammation ^[2]. One commentator also raised a caveat: because semaglutide’s gastrointestinal side effects are noticeable, some participants may have guessed their assignment and made extra lifestyle changes ^[4].

Beyond weight, semaglutide was associated with improvements in waist circumference, HbA1c, systolic and diastolic blood pressure, high-sensitivity C-reactive protein, and lipids ^[4].

What were the side effects and dropout rates?

Adverse events leading to permanent discontinuation of the trial product occurred in 16.6% of the semaglutide group (1,461 people) versus 8.2% of the placebo group (718 people), P<0.001 ^[2]. The gap was driven mainly by gastrointestinal disorders, which caused discontinuation in 10.0% versus 2.0%, P<0.001 ^[4]. Counting all reasons, premature discontinuation of the study drug happened in 26.7% of the semaglutide group and 23.6% of the placebo group ^[4].

Serious adverse events were actually less frequent with semaglutide: 33.4% versus 36.4%, P<0.001 ^[4]. A later analysis found lower rates of serious adverse events per 100 years of observation in every BMI band — 43.23, 43.54, 51.07, and 47.06 with semaglutide versus 50.48, 49.66, 52.73, and 60.85 with placebo for BMI under 30, 30 to under 35, 35 to under 40, and 40 or higher — though discontinuation rates rose as BMI class fell ^[5].

Gallbladder-related disorders were slightly more frequent with semaglutide (2.8% vs 2.3%, P=0.04), while cancer, acute kidney failure, and acute pancreatitis occurred at similar rates in both groups ^[4]. Psychiatric disorders were not higher with semaglutide (0.7% vs 0.6%), though no further breakdown of those events was provided ^[4]. No unexpected safety findings were reported ^[4].

How does SELECT compare with the other semaglutide heart trials?

SELECT is the only one of the major semaglutide cardiovascular outcome trials conducted in people without diabetes ^[2]. The others enrolled people with type 2 diabetes:

Trial	Population	Treatment	Primary result
SELECT [2, 4]	17,604 with overweight/obesity and established CVD, no diabetes	Semaglutide 2.4 mg weekly	MACE 6.5% vs 8.0%; HR 0.80 (0.72–0.90)
SUSTAIN 6 ^[3]	3,297 with type 2 diabetes at high CV risk	Semaglutide 0.5 or 1.0 mg weekly, 104 weeks	MACE 6.6% vs 8.9%; HR 0.74 (0.58–0.95)
FLOW ^[6]	3,533 with type 2 diabetes and chronic kidney disease	Semaglutide 1.0 mg weekly	Major kidney disease events 24% lower; HR 0.76, P=0.0003; stopped early at interim; median follow-up 3.4 years
SOUL ^[7]	9,650 with type 2 diabetes and ASCVD, CKD, or both	Oral semaglutide up to 14 mg daily	MACE 12.0% vs 13.8%; HR 0.86 (0.77–0.96), P=0.006; mean follow-up 47.5 months

A few details worth knowing. In SUSTAIN 6, retinopathy complications were significantly higher with semaglutide (HR 1.76, 95% CI 1.11–2.78, P=0.02) ^[3]. In FLOW, cardiovascular death was 29% lower (HR 0.71, 95% CI 0.56–0.89), MACE was 18% lower (HR 0.82), and all-cause death was 20% lower (HR 0.80) ^[6]. In SOUL, the confirmatory secondary outcomes did not differ significantly between groups, and serious adverse events were 47.9% versus 50.3% ^[7].

What questions did SELECT leave open?

SELECT does not tell you what semaglutide does for people who have never had a heart attack, stroke, or symptomatic peripheral arterial disease. Everyone enrolled already had established cardiovascular disease, so there are no data from this trial on preventing a first event in people without it ^[2]. The results also do not directly apply to people under 45 or to people with diabetes, who were excluded ^[2].

The comparator was placebo added to standard care, not another GLP-1 or incretin drug, so SELECT cannot rank semaglutide against similar medications ^[2]. The longest randomized readout reported from the trial is 208 weeks (four years) ^[5]. And because participants were already well treated for cardiovascular disease — more than 90% on lipid-lowering therapy, for example — clinicians noted this shapes how broadly the results transfer to other settings ^[4].

What did clinicians say about cost and access?

At the time of the November 2023 presentation, one commentator cited an out-of-pocket cost of around $1,300 a month for patients whose insurance did not cover the drug ^[4]. That figure was a clinician’s estimate at the time, not an official or current list price. Named barriers included variable payer coverage, high copays when insurance does cover it, and drug shortages driven by weight-loss demand ^[4].

Clinician commentary framed SELECT as establishing overweight and obesity as a modifiable cardiovascular risk factor — Lincoff called semaglutide “the first weight management therapy which is proven in a rigorous randomized controlled trial to reduce the risk of cardiovascular events” ^[4]. Contemporaneous reporting also noted that tirzepatide (Zepbound) received an obesity approval the same week as the SELECT presentation ^[4].

Sources

Frequently asked questions

Does semaglutide actually prevent heart attacks, or is it just weight loss?

In SELECT, the composite of cardiovascular death, nonfatal heart attack, and nonfatal stroke was 20% lower with semaglutide (HR 0.80, 95% CI 0.72–0.90; P<0.001) [2, 5]. The benefit appeared before substantial weight loss occurred. The presentation's discussant, Ania Jastreboff, said the gains are likely a combination of direct drug effects and indirect effects of weight loss on blood pressure, lipids, and glycemia; lead author A. Michael Lincoff said the effect is probably not simply weight loss, and the mechanism is not settled.

What is the difference between the 20% figure and the 1.5 percentage point figure?

They describe the same result. Major cardiovascular events occurred in 8.0% of placebo participants and 6.5% of semaglutide participants over a mean follow-up of 39.8 months [2, 4]. The 20% is the relative reduction between those rates; the 1.5 percentage points is the absolute difference.

How many people have to take it to prevent one event?

The number needed to treat was 67, based on the 1.5 percentage point absolute reduction over the trial's mean follow-up of 39.8 months [2, 4]. One pharmacist commentator called that a "pretty good number".

Does semaglutide help if you don't have diabetes?

SELECT is the trial that tested this directly. It enrolled 17,604 people with overweight or obesity and established cardiovascular disease but no diabetes, and it showed a 20% relative reduction in major cardiovascular events with semaglutide 2.4 mg weekly versus placebo [2, 5]. It is the only one of the major semaglutide heart trials conducted in people without diabetes.

Does it work if you've never had a heart attack or stroke?

SELECT cannot answer that. Everyone enrolled had established cardiovascular disease — a prior heart attack, prior stroke, or symptomatic peripheral arterial disease — so there are no data from this trial on preventing a first event in people without existing disease.

How long does the weight loss last on treatment?

In SELECT, weight loss continued for about 65 weeks and was sustained for up to four years. At 208 weeks, the semaglutide group had lost an average of 10.2% of body weight versus 1.5% with placebo (P<0.0001).

How many people quit because of side effects?

Adverse events leading to permanent discontinuation of the trial product occurred in 16.6% of the semaglutide group versus 8.2% of the placebo group (P<0.001), driven mainly by gastrointestinal disorders (10.0% vs 2.0%) [2, 4]. Serious adverse events overall were less frequent with semaglutide (33.4% vs 36.4%).

Was there a reduction in deaths?

Deaths from any cause were 4.3% with semaglutide versus 5.2% with placebo (HR 0.81, 95% CI 0.71–0.93). All-cause death was a hierarchically tested secondary endpoint, not the primary one, so SELECT was not designed primarily to prove a mortality benefit [2, 4].

This article summarizes FDA labeling, published research and company information current as of September 14, 2026. It is not medical advice and does not replace a conversation with your own healthcare provider. How we research and verify.