For two years the argument about these drugs was conducted almost entirely on aesthetic terrain. Vanity drugs, Hollywood drugs, the cost of thinness. Then a trial reported, and the conversation had to move.
SELECT is the most consequential study in this field, and not primarily for clinical reasons.
What the trial did
Cardiovascular outcome trials exist because regulators, having been burned, now require obesity and diabetes drugs to demonstrate they do not cause heart attacks. Most of them are designed to prove an absence of harm.
SELECT was designed to look for benefit, and it made two deliberate choices that matter.
It excluded people with diabetes. Semaglutide already had cardiovascular data in diabetes from SUSTAIN-6, and the obvious objection was that the benefit belonged to glycaemic control. Removing diabetes removed the objection.
It enrolled by cardiovascular disease, not by weight. Every participant had prior heart attack, stroke or symptomatic peripheral arterial disease. This is a secondary prevention population — people who have already had the event you are trying to prevent the next of.
Follow-up ran a little over three years.
17,604
Participants randomised in SELECT
NEJM, 2023
20%
Relative reduction in major adverse cardiovascular events
HR 0.80
6.5% vs 8.0%
Absolute event rates over roughly three years
The result, and its size
A hazard ratio of 0.80 for the primary composite — cardiovascular death, non-fatal myocardial infarction, non-fatal stroke.
In absolute terms: 1.5 fewer events per hundred people over about three years. Roughly 67 people treated for three years to prevent one major cardiovascular event.
That is a genuinely useful effect size. It is comparable to what statins deliver in similar populations, and statins reshaped cardiology.
It is also not a miracle, and the framing in some coverage — “cuts heart attack risk by twenty percent” — invites people to hear something larger than 1.5 percentage points. Both numbers are true; only one of them tells you what to expect.
Is it the weight loss?
The interesting question, and not fully answered.
Participants lost around nine percent of body weight on average — meaningful, but well short of the fifteen percent seen in STEP 1, because this was an older, sicker population.
Two observations argue against weight loss being the whole story. The event curves began to separate early, within the first several months, before much weight had been lost. And the benefit did not track cleanly with individual weight change — people who lost little still appeared to benefit.
The leading candidate explanations are anti-inflammatory effects, direct action on vascular endothelium, blood pressure reduction, and improved lipid handling. High-sensitivity CRP fell substantially in the semaglutide arm, which is at least consistent with an inflammatory mechanism.
Why the label change mattered more than the trial
Here is the part that has affected more lives than the clinical result.
American insurance treats obesity drugs badly. Medicare Part D is statutorily prohibited from covering drugs for weight loss — a provision written in 2003, when the available products were poor and the political appetite was nil. Many commercial plans copied the exclusion.
In March 2024 the FDA approved Wegovy for reducing cardiovascular risk in adults with established cardiovascular disease and obesity or overweight. That is not a weight-loss indication. CMS subsequently confirmed that Part D plans may cover semaglutide when prescribed for it.
The drug did not change. The molecule, the dose and the pen are identical. What changed was the sentence on the label, and with it the answer to whether a plan is permitted to pay.
The same injection, in the same dose, became coverable — because it acquired a second reason to exist.
This is now a routine part of appeals. A patient with a documented prior myocardial infarction and a BMI over 27 has a coverage pathway that a patient with a BMI of 34 and no cardiac history does not, and the difference has nothing to do with need. If you are fighting a denial, the indication you apply under is frequently the whole game — the mechanics are in appealing a prior authorisation.
What it does not show
It is a secondary prevention trial. Everyone enrolled had established cardiovascular disease. It does not tell you what semaglutide does for a healthy 40-year-old with a BMI of 31 and no cardiac history, and it should not be quoted as though it does.
It is semaglutide. Tirzepatide’s dedicated cardiovascular outcome programme is a separate body of evidence, and the two molecules should not be assumed interchangeable on this endpoint.
It ran three years. These are drugs people may take for decades. Three years is a good trial and a short lifetime.
What it changed
Before SELECT, defending these drugs meant arguing that obesity is a disease — a true claim that has lost every political fight it has ever been in.
After SELECT, the argument is that a specific injection prevents a specific number of heart attacks in a specific population over three years. That is an argument insurers, formulary committees and Medicare are structurally obliged to engage with.
It was always going to take a hard endpoint. This is the one it took.