Semaglutide and Heart Risk: What SELECT Proved, and for Whom

Semaglutide reduced serious cardiovascular events in SELECT, a trial in adults with established cardiovascular disease and overweight or obesity. The result supports a real clinical benefit in that population. It does not mean every person taking any semaglutide formulation receives the same protection, or that weight loss alone proves a heart benefit¹.

The distinction is worth making before looking at a “20% reduction” headline. A relative reduction describes the difference between groups. Your starting risk, the outcome being counted and the length of follow-up determine what that difference means in practice.

For different treatment goals, read the separate evidence on kidney outcomes in diabetes and CKD and MASH liver-biopsy outcomes. These studies have different populations and endpoints.

The finding that changed the indication

FDA approved Wegovy injection for reducing cardiovascular death, heart attack and stroke in adults with cardiovascular disease and either obesity or overweight in March 2024. Its announcement reports these events in 6.5% of treated participants, compared with 8% receiving placebo. Both groups also received usual cardiovascular care and lifestyle counseling².

The drug was therefore added to care. It was not tested as a replacement for blood-pressure treatment, cholesterol management or the other measures the participants already needed. A claim that the medication makes those measures unnecessary would turn an add-on trial into a replacement trial that was never conducted.

This is also a more meaningful endpoint than a change in a laboratory marker alone. People had fewer events that directly matter to their health. Our guide to surrogate endpoints explains why that distinction makes SELECT important.

Who was actually in SELECT

The study enrolled 17,604 adults at least 45 years old, with existing cardiovascular disease, a BMI of at least 27 and no history of diabetes. Participants received weekly semaglutide injection or placebo. Average follow-up was 39.8 months; the reported hazard ratio for the main outcome was 0.80. Novo Nordisk funded the trial¹.

That population is the boundary of the direct trial claim. A younger adult with no cardiovascular disease may still have a reason to discuss weight management with a clinician. But SELECT alone does not tell that person how much their risk of a first cardiovascular event would fall.

Likewise, the trial's exclusion of diabetes does not make its findings a universal answer for diabetes care. Studies in different populations answer related questions. They should be identified separately rather than combined into a sentence that implies every person has been studied in the same way.

The absolute difference behind the headline

Using the rounded event percentages in FDA's summary, the absolute difference is 1.5 percentage points: 8.0 minus 6.5. Expressed per 1,000 similar participants over the trial's follow-up, that is about 80 people with an event in the placebo group versus 65 in the semaglutide group, a difference of about 15². These are calculations from the reported percentages, not additional trial outcomes.

The hazard ratio compares the rate of a first event over follow-up; the event percentages summarize how many participants had an event. They are related but are not interchangeable calculations. Keep both the relative time-to-event result and the absolute event proportions visible.

Do not multiply the difference by the number of years you expect to take treatment. A several-year event proportion is not a fixed annual discount. Risk changes over time, people discontinue treatment, and the statistical analysis accounts for when events occur.

A composite endpoint is not three separate promises

SELECT's principal outcome combined cardiovascular death, nonfatal heart attack and nonfatal stroke. Combining events helps a trial assess an important overall question. It does not mean the headline relative reduction necessarily applies identically to each component¹.

When a seller writes “20% fewer heart attacks” using the overall SELECT figure, ask for the heart-attack-specific result. The wording has narrowed the outcome while keeping the larger study's headline. The same check applies to a claim about longer life: a composite including nonfatal events is not identical to all-cause mortality.

This is not a reason to dismiss the trial. It is a reason to describe its success with the same outcome it actually measured. A useful review should preserve the endpoints rather than make the strongest possible promotional sentence from them.

Tolerability belongs beside benefit

Adverse events caused permanent treatment discontinuation in 16.6% of the semaglutide group and 8.2% of the placebo group in the published report¹. Those figures do not mean that all discontinuations were severe events. They show that staying on treatment was a meaningful issue even in a closely monitored trial.

Someone comparing options needs both sides of that record. A drug can reduce important long-term risks while being difficult for some people to tolerate. The benefits do not erase the need for a clinician to consider contraindications, other medicines and the current prescribing information.

A trial average cannot tell you which side of that experience you will have. It can support an informed discussion about what is likely, what requires monitoring and what would lead to a change in treatment. That is more useful than presenting the favorable percentage alone.

Why the product matters

SELECT is evidence for the studied injection and regimen. A compounded formulation, a different delivery route or a different ingredient mixture does not inherit a completed outcomes trial merely because its advertising uses the word semaglutide. FDA does not review compounded drugs for safety, effectiveness and quality before marketing in the way it reviews approved products³.

The distinction is not semantic. The claim being made is about clinical outcomes over years. To say a different preparation has the same demonstrated outcome requires evidence for that preparation or an appropriate regulatory basis, not just a shared ingredient name. See why compounded is not the trial drug.

The same caution applies to oral semaglutide. SELECT itself studied an injection. Evidence for a tablet should be assessed from its own trials and current approved indication. One route's trial does not become another route's trial because the mechanism is related.

When comparing a provider's results with SELECT, check who was counted. The Ivim Health review explains its selected-completer results; the WeightWatchers Clinic review examines an engagement comparison. These program weight-loss figures answer different questions from a randomized cardiovascular-outcomes trial.

Questions that make the evidence useful

For a clinical discussion, start with whether you have the kind of established cardiovascular disease represented in the evidence. Then ask which approved product and indication are relevant, how treatment fits with existing care, and what monitoring or side effects could change the plan.

For evaluating an article or advertisement, ask whether it names SELECT, states the studied population, gives the time horizon and distinguishes relative from absolute benefit. If all four are missing, a correct percentage may still be doing misleading work.

Our semaglutide evidence page places this finding beside the weight-loss literature. The cardiovascular result is stronger than a promise about a biomarker, but its boundaries are part of the finding. Keep the population and product attached to the benefit whenever the number is repeated.