Beyond glucose and weight: what the GLP-1 hard-endpoint trials actually showed

Reporting the SELECT, LEADER, and FLOW randomized trials on heart and kidney outcomes, with the numbers and the caveats.

Weight loss and lower HbA1c are what most people associate with semaglutide and liraglutide. The clinical question that moved regulators and guidelines, though, was different: do these drugs prevent the events that actually kill people with metabolic disease, namely heart attacks, strokes, cardiovascular death, and kidney failure? Three large randomized trials are the ones usually cited on that question. Here is what they reported.

This is a summary of published evidence, not medical advice. Drug names, doses, and indications differ by country and by product; the figures below come from the trials as published.

SELECT: semaglutide in obesity without diabetes

SELECT enrolled 17,604 adults aged 45 or older who had overweight or obesity and established cardiovascular disease (prior heart attack, stroke, or peripheral artery disease) but no diabetes. Participants were randomized to once-weekly semaglutide 2.4 mg or placebo, both on top of standard care. Mean follow-up was about 40 months.

The primary endpoint was a three-part composite: death from cardiovascular causes, nonfatal myocardial infarction, or nonfatal stroke. It occurred in 6.5% of the semaglutide group versus 8.0% on placebo (hazard ratio 0.80; 95% CI 0.72–0.90; p<0.001) — a 20% relative reduction (SELECT, NEJM 2023).

Safety is part of the picture. Serious adverse events were actually less frequent on semaglutide (33.4% vs 36.4%), but discontinuation because of adverse events ran the other way: 16.6% on semaglutide versus 8.2% on placebo, driven largely by gastrointestinal effects.

This trial is why the U.S. FDA expanded the Wegovy label on March 8, 2024 to include reducing the risk of cardiovascular death, heart attack, and stroke in adults with cardiovascular disease and overweight or obesity. It was the first weight-management drug to get such an indication.

LEADER: liraglutide in type 2 diabetes at high cardiovascular risk

LEADER ran earlier and in a different population: 9,340 adults with type 2 diabetes and high cardiovascular risk, randomized to liraglutide (titrated up to 1.8 mg daily; median dose 1.78 mg) or placebo. Median follow-up was 3.8 years.

The primary composite (cardiovascular death, nonfatal MI, or nonfatal stroke) occurred in 13.0% on liraglutide versus 14.9% on placebo (hazard ratio 0.87; 95% CI 0.78–0.97; p<0.001 for noninferiority, p=0.01 for superiority) (LEADER, NEJM 2016). The mortality signal was notable: death from cardiovascular causes was lower with liraglutide (HR 0.78; 95% CI 0.66–0.93), as was death from any cause (HR 0.85; 95% CI 0.74–0.97).

On safety, discontinuation due to adverse events was 9.5% with liraglutide versus 7.3% with placebo. Acute pancreatitis cases were numerically fewer on liraglutide (18 vs 23), but acute gallstone disease was more common (3.1% vs 1.9%).

FLOW: semaglutide and the kidney

FLOW tested a kidney question. It enrolled 3,533 adults with type 2 diabetes and chronic kidney disease (defined by eGFR and urine albumin thresholds), on a renin-angiotensin-system inhibitor, randomized to once-weekly semaglutide 1.0 mg or placebo. Median follow-up was 3.4 years, and the trial was stopped early for efficacy.

The primary composite was time to first major kidney event: a persistent ≥50% drop in eGFR, kidney failure (eGFR <15 or start of chronic kidney-replacement therapy), or death from kidney or cardiovascular causes. It occurred at 5.8 versus 7.5 events per 100 patient-years, a 24% relative reduction (hazard ratio 0.76; 95% CI 0.66–0.88; p=0.0003) (FLOW, NEJM 2024).

Secondary endpoints pointed the same direction: the kidney-specific composite (HR 0.79; 95% CI 0.66–0.94), a slower annual eGFR decline (difference 1.16 mL/min/1.73 m² per year), major cardiovascular events (HR 0.82; 95% CI 0.68–0.98), and death from any cause (HR 0.80; 95% CI 0.67–0.95). Serious adverse events were less frequent on semaglutide (49.6% vs 53.8%).

Side by side

Trial	Drug, dose	Population	Primary endpoint	Result
SELECT (NEJM 2023)	Semaglutide 2.4 mg weekly	Overweight/obesity + CVD, no diabetes (n=17,604)	CV death, nonfatal MI, nonfatal stroke	6.5% vs 8.0%; HR 0.80 (0.72–0.90)
LEADER (NEJM 2016)	Liraglutide up to 1.8 mg daily	T2D + high CV risk (n=9,340)	CV death, nonfatal MI, nonfatal stroke	13.0% vs 14.9%; HR 0.87 (0.78–0.97)
FLOW (NEJM 2024)	Semaglutide 1.0 mg weekly	T2D + CKD (n=3,533)	Major kidney event composite	HR 0.76 (0.66–0.88); stopped early

What these trials do and don't establish

A few limits are worth keeping straight. The populations were specific. SELECT studied people with prior cardiovascular disease and no diabetes; its result should not be read as a claim about cardiovascular prevention in lower-risk people or in those without established disease. LEADER and FLOW both required type 2 diabetes. None of these trials enrolled healthy adults taking a GLP-1 drug only to lose weight.

The drugs, doses, and routes also differ. SELECT used the 2.4 mg weight-management dose of semaglutide; FLOW used the 1.0 mg dose more typical of diabetes care; LEADER used daily liraglutide. Findings don't automatically transfer between products or doses.

Gastrointestinal effects (nausea, vomiting, diarrhea) drove most of the higher discontinuation in the active arms, and the trials were of finite length — SELECT's roughly three-and-a-half-year mean follow-up says nothing direct about a decade of use. The gallbladder signal in LEADER and the malignancy and retinopathy questions raised in earlier semaglutide work are reasons these drugs carry monitoring and labeling cautions in their approved indications.

A closing point on scope. The trials above studied FDA- and EMA-approved, pharmaceutical-grade products under medical supervision. They say nothing about the safety or content of research-grade or gray-market peptides, which are not approved for human consumption and are sold for laboratory use only. Outcome data from a regulated trial does not carry over to an unverified vial.