Orforglipron and the Heart: The Lancet Cardiovascular Safety Trial

Orforglipron met its cardiovascular safety test in ACHIEVE-4: major cardiovascular events occurred in 4.2% of participants on the pill and 5.0% on insulin glargine over a median of 2 years (hazard ratio 0.84; 95% CI 0.59 to 1.20).^[1] That establishes non-inferiority to insulin in adults with type 2 diabetes and established cardiovascular or chronic kidney disease. It does not establish that orforglipron prevents heart attacks or strokes, and the trial did not study people with obesity who do not have diabetes.

What ACHIEVE-4 tested

ACHIEVE-4 was the first look at cardiovascular safety for a non-peptide GLP-1 receptor agonist, according to its authors, who note that this question “has not been studied” for the class.^[1] Eli Lilly funded the trial, and 7 of the 11 authors are Lilly employees.^[1] The design details below come from the Lancet abstract and the trial registry.^[1][6]

• Design: phase 3, event-driven, randomized, open-label, at 317 sites in 16 countries and territories.^[1] Neither participants nor investigators were masked, per the registry.^[6]
• People: 2,749 adults with type 2 diabetes, HbA1c of 7.0% to 10.5% and BMI of 25 or higher. All had established cardiovascular or chronic kidney disease: 85.9% had cardiovascular disease and 37.6% had chronic kidney disease. Mean age was 63.1 years, mean HbA1c 8.2% and mean BMI 33.^[1]
• Treatments: once-daily orforglipron at the maximum tolerated dose, up to a 36 mg capsule (equivalent to the 17.2 mg tablet), or once-daily titrated insulin glargine injections.^[1] The 17.2 mg tablet is the top dose in the Foundayo label.^[5]
• Primary outcome: time to first MACE-4 event, meaning cardiovascular death, non-fatal heart attack, non-fatal stroke or hospitalization for unstable angina.^[1]
• Analysis: everyone who took at least one dose, using all data points from baseline until withdrawal or study completion, regardless of whether they stayed on treatment.^[1] The result therefore includes time after some people had stopped taking their assigned drug.

How to read a hazard ratio of 0.84

The primary outcome occurred in 57 of 1,358 orforglipron participants and 67 of 1,343 insulin glargine participants.^[1] The trial declared non-inferiority if the upper limit of the 95% confidence interval stayed below 1.8, and the upper limit was 1.20 (p<0.0001 for non-inferiority).^[1] The 1.8 margin is the same one PIONEER 6 used for oral semaglutide, which its authors described as ruling out “80% excess cardiovascular risk.”^[3]

Two limits follow from those numbers. The interval runs from 0.59 to 1.20, so the data fit anything from a 41% lower to a 20% higher rate of events compared with glargine. And the p-value applies to non-inferiority: since the interval crosses 1.0, the point estimate of 0.84 should not be quoted as a 16% reduction. With 124 primary events in total, the estimate is imprecise. The registry lists a three-component MACE outcome, HbA1c and body weight as secondary measures, but the abstract does not report them.^[1][6] Our overview of the ACHIEVE diabetes program covers the glucose side of these trials.

Why insulin glargine as the comparator matters

A placebo-controlled trial asks whether a drug changes risk against background care. ACHIEVE-4 asks something narrower: how does orforglipron compare with an active diabetes drug? Glargine is a defensible yardstick because the ORIGIN trial found it neutral for cardiovascular outcomes over a median of 6.2 years, with a hazard ratio of 1.02 (95% CI 0.94 to 1.11) against standard care in people with cardiovascular risk factors and dysglycemia.^[2] Sanofi funded ORIGIN.^[2]

ORIGIN also found that glargine increased severe hypoglycemia (1.00 versus 0.31 per 100 person-years) and modestly increased weight (median gain of 1.6 kg versus a 0.5 kg loss).^[2] ACHIEVE-4 cannot tell us how much of any gap between its arms reflects a property of orforglipron and how much reflects those features of the comparator. ORIGIN also enrolled a different population with a different glucose target, so its result is context and does not transfer directly.

The deaths: a striking number with caveats

The abstract reports 62 deaths: 19 of 1,371 participants (1.4%) on orforglipron and 43 of 1,355 (3.2%) on insulin glargine. All except one, in the glargine group, were judged unrelated to treatment.^[1] The abstract gives counts without a hazard ratio or significance test, and its conclusions rest on non-inferiority for MACE-4 and do not claim a survival benefit.^[1] An unmasked trial built around a safety margin is a weak basis for concluding that a drug extends life, so we would treat this imbalance as a finding that needs replication.

Oral semaglutide offers a useful parallel. PIONEER 6 also showed lower death from any cause than placebo (1.4% versus 2.8%; hazard ratio 0.51; 95% CI 0.31 to 0.84).^[3] The much larger SOUL trial then tested oral semaglutide for superiority and reported a hazard ratio of 0.86 (95% CI 0.77 to 0.96) for three-component MACE.^[4] The SOUL abstract gives no all-cause death result, so it cannot say whether that early signal held. Still, the sequence shows the kind of larger test a mortality imbalance from a safety trial calls for.

Oral GLP-1 cardiovascular trials side by side

The table below draws on ACHIEVE-4, PIONEER 6 and SOUL.^[1][3][4] These trials used different comparators, endpoints and populations, so any comparison across rows is indirect. ACHIEVE-4 tested orforglipron against an active drug, while the semaglutide trials used placebo. Hazard ratios from these trials cannot rank the two pills.

Gastrointestinal effects and low blood sugar

Gastrointestinal adverse events affected 851 of 1,371 participants (62.1%) on orforglipron and 193 of 1,355 (14.2%) on glargine, and they were the most common reason for stopping orforglipron.^[1] Clinically significant or severe hypoglycemia, with glucose below 54 mg/dL, went the other way: 93 participants (6.8%) on orforglipron and 260 (19.2%) on glargine.^[1] The broader tolerability picture from the obesity trials is in our review of Foundayo side effects.

Liver safety was examined separately. A pooled review of seven phase 3 orforglipron trials, with sponsor employees among its authors and sponsor-provided data analysis, found ALT or AST at least 3 times the upper limit with bilirubin at least 2 times the limit in 6 orforglipron participants (0.1%) and 6 comparator participants (0.1%). The authors attributed all the orforglipron cases to other causes.^[10]

What this means if you take Foundayo for weight

Readers on Foundayo for obesity should be careful about applying ACHIEVE-4 to themselves. In its July 2026 label, Foundayo is indicated “to reduce excess body weight and maintain weight reduction long term” in adults with obesity or adults with overweight and at least one weight-related condition, alongside diet and physical activity.^[5] Every ACHIEVE-4 participant had type 2 diabetes plus established cardiovascular or chronic kidney disease.^[1] The trial offers useful reassurance about a higher-risk group, though it does not measure risk in a healthier person using the drug for weight. Injectable semaglutide has outcome evidence in people without diabetes from the SELECT trial, and no equivalent orforglipron result exists in that group yet.

The label also describes a heart-rate effect. In a pool of its two weight-management trials, tachycardia was reported in 3% of Foundayo patients and 0.9% on placebo, with “a mean increase in heart rate of 4 to 5 beats per minute from baseline compared to 0.5 beat per minute with placebo.”^[5] The ACHIEVE-4 abstract does not report heart rate.^[1]

Risk models are not outcomes

Two recent papers estimate cardiovascular benefit from risk scores, and neither counted actual heart attacks or strokes. A post hoc ATTAIN-1 analysis led by Lilly employees, in people with overweight or obesity and no type 2 diabetes, found that orforglipron 5.5 to 17.2 mg lowered predicted 10-year total cardiovascular risk on the Framingham score at 72 weeks (absolute change of -0.6% to -1.00% versus +0.6% with placebo).^[8] A separate modeling paper from Jacobi Medical Center applied pooled trial effects to hypothetical patients. For a representative 50-year-old man with type 2 diabetes on 36 mg, it projected a 10-year relative risk reduction of 16.8% with the older Pooled Cohort Equations and 31.4% with an HbA1c-enhanced PREVENT model.^[9] Its authors state that these projections “remain model-derived” and need validation in outcomes trials.^[9]

The trial that comes next

The registry lists ATTAIN-Outcomes, a double-blind, placebo-controlled, event-driven trial in people with established atherosclerotic cardiovascular disease, chronic kidney disease or both. Its estimated enrollment is 7,140 and its estimated primary completion is August 2031.^[7] The primary composite is broader than MACE-4: non-fatal heart attack, non-fatal stroke, hospitalization or urgent visit for heart failure, coronary revascularization, or death from any cause.^[7] That design differs from ACHIEVE-4 in comparator, masking and endpoint, so its result will not be a simple repeat. Its outcome is unknown.

What to raise with a prescriber

• Whether you match the ACHIEVE-4 population, which means type 2 diabetes with established cardiovascular or chronic kidney disease, or whether the evidence applies to you only indirectly.
• Why an oral or injectable GLP-1 with a demonstrated reduction in cardiovascular events might or might not suit you better than one with a non-inferiority result, if heart protection is your main goal.
• Whether your insulin or sulfonylurea dose needs adjusting. The label says hypoglycemia risk may be lowered by reducing those doses.^[5]
• What resting heart rate or palpitations should prompt a call, especially if you have a heart rhythm problem or heart failure.