Semaglutide vs tirzepatide: what the head-to-head data shows

The comparison finally has a direct answer. In SURMOUNT-5, the first head-to-head obesity trial of the two drugs, tirzepatide produced a mean body-weight reduction of 20.2% at 72 weeks against 13.7% for semaglutide 2.4 mg (NEJM 2025). That is the headline. The rest of this article is what the headline compresses: two different mechanisms, the landmark trials with their actual numbers, why the cross-trial comparisons everyone made for three years were always shaky, and what the record still leaves open.

Both compounds are FDA-approved and, by trial enrollment alone, among the most thoroughly studied peptides ever made. Full compound profiles — mechanism, half-life, regulatory status, safety signals — live at semaglutide and tirzepatide.

One receptor versus two

Semaglutide is a long-acting analog of GLP-1, an incretin hormone the gut releases after eating. A fatty-acid side chain lets the molecule bind albumin, stretching its half-life to about 7 days and making once-weekly injection practical. At the receptor it does what native GLP-1 does, but continuously: boosts insulin secretion only when glucose is elevated, suppresses glucagon, slows gastric emptying, and dials down appetite through receptors in the brain.

Tirzepatide is a single 39-amino-acid peptide engineered on a GIP backbone to activate two receptors at once — GIP and GLP-1. Its half-life is about 5 days, also dosed weekly. The GLP-1 activity covers the same territory as semaglutide. The GIP activity adds a second incretin signal: amplified glucose-dependent insulin secretion, plus effects on adipose tissue that appear to improve lipid handling.

Why the dual agonist outperforms on weight is genuinely unsettled. GIP agonism was long assumed to be redundant for body weight — tirzepatide's results forced a rethink the pharmacology literature is still working through. The trials below established that the difference exists; they did not establish exactly why.

The landmark trials, with their actual numbers

Behind the means: in STEP 1, 86% of semaglutide participants lost at least 5% of body weight and roughly half lost 15% or more. In SURMOUNT-1, more than half of the 15 mg arm lost at least 20%. Placebo arms in both trials received the same lifestyle intervention as the treatment arms — which is why they lost weight too.

Why the cross-trial comparison misled

For three years the standard move was to read STEP 1's −14.9% against SURMOUNT-1's −20.9% and call it a comparison. It never really was one. The trials enrolled different populations at different times, ran different durations (68 versus 72 weeks), used different statistical estimands, and produced different placebo responses (−2.4% versus −3.1%) — a tell that the trial contexts themselves differed. Cross-trial deltas of a few percentage points can be manufactured by any one of those factors. That is exactly why SURMOUNT-5 had to be run.

SURPASS-2: the first head-to-head, with an asterisk

SURPASS-2 (NEJM 2021) compared the two drugs directly for type 2 diabetes: 1,879 adults on metformin, 40 weeks. On the primary endpoint, every tirzepatide dose reduced A1c more than semaglutide — 2.01, 2.24, and 2.30 percentage points against 1.86. On body weight, a secondary endpoint, the tirzepatide arms lost 7.6, 9.3, and 11.2 kg against 5.7 kg for semaglutide.

The asterisk: the semaglutide comparator was 1 mg — the approved diabetes dose at the time — not the 2 mg diabetes dose approved later, and not the 2.4 mg dose used for weight management. SURPASS-2 answered "which drug at these doses controls glucose better in diabetes," not "which drug at full dose produces more weight loss." Carrying its weight numbers into obesity arguments always overreached the design.

SURMOUNT-5: the direct answer for obesity

SURMOUNT-5 (NEJM 2025) removed the asterisk. It enrolled 751 adults with obesity, or overweight plus a weight-related condition, without diabetes, and randomized them to tirzepatide or semaglutide — each titrated to its maximum tolerated dose (10 or 15 mg for tirzepatide; 1.7 or 2.4 mg for semaglutide) for 72 weeks. Mean weight reduction: 20.2% for tirzepatide, 13.7% for semaglutide.

Two caveats are worth keeping attached. The trial was open-label — participants and clinicians knew the assignment — which matters at the margin even when the endpoint is a scale reading. And a mean is not a prediction: the response distributions overlap, and some semaglutide participants lost more than the tirzepatide average. SURMOUNT-5 settles the population-level question, not the individual one.

Side effects: a shared class signature

Both drugs are GI-dominant in their adverse-event profiles, and both concentrate those events during dose escalation. In STEP 1, nausea was reported by 44% of the semaglutide arm versus 17% on placebo, with diarrhea, vomiting, and constipation following; 7% discontinued for adverse events versus 3.1% on placebo. In SURMOUNT-1, nausea led the adverse-event table at every dose — reported by roughly a quarter to a third of participants — and adverse-event discontinuations ran between 4% and 7% versus 2.6% for placebo. Most events in both programs were mild to moderate and clustered around titration steps.

The labels carry the same class warnings for both compounds: pancreatitis, gallbladder disease, hypoglycemia when combined with insulin or sulfonylureas, acute kidney injury secondary to dehydration, and the boxed rodent thyroid C-cell finding. What those warnings mean for monitoring is its own topic — covered in the GLP-1 bloodwork article.

Cardiovascular outcomes: semaglutide holds the completed superiority trial

SELECT (NEJM 2023) tested semaglutide 2.4 mg against placebo in 17,604 adults with established cardiovascular disease and overweight or obesity, without diabetes. Major adverse cardiovascular events occurred in 6.5% of the semaglutide arm versus 8.0% on placebo — a 20% relative risk reduction, and the first demonstration that a drug in this class reduces cardiovascular events in a non-diabetic population.

Tirzepatide's cardiovascular outcomes trial, SURPASS-CVOT, reported in 2025: in adults with type 2 diabetes, tirzepatide was noninferior to dulaglutide — a comparator that had itself demonstrated cardiovascular benefit against placebo in REWIND (The Lancet, 2019). Noninferiority to an active, cardioprotective drug is a meaningful result, but it is a different kind of evidence than SELECT's placebo-controlled superiority in an obesity population. Tirzepatide's dedicated obesity morbidity-and-mortality trial, SURMOUNT-MMO, is still running.

Titration, as the labels write it

These are label facts, not suggestions — the schedules exist because GI tolerance is built gradually, and every trial above embedded one.

• Wegovy (semaglutide 2.4 mg): 0.25 mg weekly for four weeks, then 0.5, 1.0, and 1.7 mg in successive four-week steps, reaching the 2.4 mg maintenance dose in week 17. The label allows holding maintenance at 1.7 mg when 2.4 mg isn't tolerated.
• Zepbound (tirzepatide): 2.5 mg weekly for four weeks, then 5 mg; further increases in 2.5 mg increments after at least four weeks at a given dose, with maintenance at 5, 10, or 15 mg and a 15 mg maximum.

What the record still leaves open

Durability off-drug. In the STEP 1 extension (Diabetes, Obesity and Metabolism, 2022), participants regained about two-thirds of their lost weight within a year of stopping. SURMOUNT-4 (JAMA, 2024) found the same shape for tirzepatide: after a 36-week lead-in, participants switched to placebo regained roughly 14% of body weight over the following year while those who continued lost about 5.5% more. Neither drug has evidence that its results persist after discontinuation.

Body composition. Both programs measured lean mass only in sub-studies. How much of the loss is fat versus lean tissue — and what that ratio means over years of use — remains an active research question.

Who responds. Neither program produced a usable baseline predictor of response. The overlapping distributions in SURMOUNT-5 mean the individual answer still requires the individual experiment.

The next comparison. The pipeline behind these two is already in phase 3 — retatrutide, a triple agonist, and cagrilintide paired with semaglutide as CagriSema. The head-to-heads of the next cycle are being designed now.

Sources

• STEP 1 — Wilding et al., New England Journal of Medicine, 2021.
• STEP 1 extension — Wilding et al., Diabetes, Obesity and Metabolism, 2022.
• SURMOUNT-1 — Jastreboff et al., New England Journal of Medicine, 2022.
• SURPASS-2 — Frías et al., New England Journal of Medicine, 2021.
• SURMOUNT-5 — Aronne et al., New England Journal of Medicine, 2025.
• SURMOUNT-4 — Aronne et al., JAMA, 2024.
• SELECT — Lincoff et al., New England Journal of Medicine, 2023.
• REWIND — Gerstein et al., The Lancet, 2019.
• SURPASS-CVOT — cardiovascular outcomes trial of tirzepatide versus dulaglutide in type 2 diabetes; results reported 2025.
• Wegovy (semaglutide) prescribing information — U.S. FDA label, Novo Nordisk.
• Zepbound (tirzepatide) prescribing information — U.S. FDA label, Eli Lilly.