Key facts
- Comparison: tirzepatide (dual GIP/GLP-1 agonist) 5, 10 or 15 mg vs semaglutide (GLP-1 agonist) 1 mg, all once weekly, subcutaneous.
- Sponsor: Eli Lilly and Company.
- Design: randomized, open-label, active-controlled, parallel-group phase 3 trial, 40 weeks.
- Population: 1,879 adults with type 2 diabetes inadequately controlled on metformin; mean baseline A1c 8.28%, mean age 56.6 years, mean weight 93.7 kg.
- Randomization: 1:1:1:1 across the three tirzepatide doses and semaglutide.
- Primary endpoint: change in glycated hemoglobin (A1c) from baseline to 40 weeks.
- Headline result: tirzepatide noninferior and superior to semaglutide at all doses on A1c, with greater weight reduction.
- Registration: NCT03987919 — ClinicalTrials.gov.
- Publication: New England Journal of Medicine, 2021 — PMID 34170647.
What was the SURPASS-2 trial?
SURPASS-2 was an open-label, 40-week, phase 3 randomized trial that compared once-weekly tirzepatide with once-weekly semaglutide in adults with type 2 diabetes. It was sponsored by Eli Lilly and published in the New England Journal of Medicine in 2021 (PMID 34170647).
Its value comes from being a direct, head-to-head comparison. Tirzepatide is a dual glucose-dependent insulinotropic polypeptide (GIP) and glucagon-like peptide-1 (GLP-1) receptor agonist, while semaglutide is a selective GLP-1 receptor agonist. Before SURPASS-2, the efficacy and safety of tirzepatide relative to semaglutide were, in the authors' own words, unknown. The trial is registered as NCT03987919 on ClinicalTrials.gov. For the broader match-up between these two molecules, see our dedicated comparison of tirzepatide vs semaglutide.
Who was in it?
SURPASS-2 enrolled 1,879 adults with type 2 diabetes that was inadequately controlled on metformin alone. At baseline, the mean glycated hemoglobin (A1c) level was 8.28%, the mean age was 56.6 years, and the mean body weight was 93.7 kg. Everyone in the trial was already taking metformin, so the study tested each injectable as an add-on to that background therapy rather than as a standalone treatment.
Because the comparison was randomized, the four groups started from a similar place — the differences seen at 40 weeks are attributable to the assigned drug and dose rather than to who happened to be sicker at the outset.
What did participants receive?
Participants were randomly assigned, in a 1:1:1:1 ratio, to one of four once-weekly subcutaneous regimens: tirzepatide at 5 mg, 10 mg or 15 mg, or semaglutide at 1 mg. Each was added on top of ongoing metformin. Splitting tirzepatide across three doses let the trial map how the drug's effect changed as the dose rose, while the single semaglutide arm served as the active comparator.
Describing the assigned regimens is reporting what the trial did — it is not a dosing instruction. Both tirzepatide and semaglutide are prescription-only, and the dose any individual might use is a clinical decision, not something to infer from a trial protocol. The molecules themselves are covered in our tirzepatide and semaglutide research summaries.
Primary results: A1c
The primary endpoint was the change in glycated hemoglobin (A1c) from baseline to 40 weeks. The estimated mean reduction was -2.01 percentage points with tirzepatide 5 mg, -2.24 with 10 mg and -2.30 with 15 mg, compared with -1.86 percentage points for semaglutide 1 mg.
The estimated differences between each tirzepatide dose and semaglutide were -0.15 percentage points (95% CI -0.28 to -0.03; P=0.02) for 5 mg, -0.39 (95% CI -0.51 to -0.26; P<0.001) for 10 mg, and -0.45 (95% CI -0.57 to -0.32; P<0.001) for 15 mg. Tirzepatide at all three doses was both noninferior and superior to semaglutide for A1c lowering.
| A1c change from baseline (40 weeks) | Estimated mean change | Difference vs semaglutide 1 mg |
|---|---|---|
| Tirzepatide 5 mg | -2.01 points | -0.15 (95% CI -0.28 to -0.03); P=0.02 |
| Tirzepatide 10 mg | -2.24 points | -0.39 (95% CI -0.51 to -0.26); P<0.001 |
| Tirzepatide 15 mg | -2.30 points | -0.45 (95% CI -0.57 to -0.32); P<0.001 |
| Semaglutide 1 mg (comparator) | -1.86 points | — |
The gap between the two drugs widened as the tirzepatide dose rose. All four arms produced a large absolute A1c reduction from a high baseline of 8.28%; the trial's contribution is showing that tirzepatide edged out an already-effective comparator.
Weight results
Reductions in body weight were greater with tirzepatide than with semaglutide. Expressed as least-squares mean estimated treatment differences, tirzepatide beat semaglutide by -1.9 kg at 5 mg, -3.6 kg at 10 mg and -5.5 kg at 15 mg, with P<0.001 for all comparisons. As with A1c, the advantage grew with dose.
Weight was a secondary outcome here — the primary question was blood-sugar control — but the direction is consistent with what the wider evidence base shows for these drugs. For how tirzepatide performs when weight loss itself is the goal, see our summaries of the SURMOUNT programme, including SURMOUNT-5, and the broader landscape in peptides for weight loss.
Safety and side effects reported
The most common adverse events were gastrointestinal and were primarily mild to moderate in severity in both the tirzepatide and semaglutide groups. Reported rates were nausea in 17 to 22% of tirzepatide-treated participants versus 18% with semaglutide; diarrhea in 13 to 16% versus 12%; and vomiting in 6 to 10% versus 8%.
Hypoglycemia with a blood-glucose level below 54 mg per deciliter was uncommon: it was reported in 0.6% of the tirzepatide 5-mg group, 0.2% of the 10-mg group and 1.7% of the 15-mg group, versus 0.4% of the semaglutide group. Serious adverse events were reported in 5 to 7% of participants receiving tirzepatide and in 3% of those receiving semaglutide. These figures come directly from the published record; this page does not extrapolate beyond them.
Limitations and what it does not prove
SURPASS-2 is an important head-to-head trial, but it has boundaries that matter when the headline gets quoted:
- It was open-label. Participants and investigators knew which drug was being used. The primary endpoint, A1c, is a laboratory value that is hard to bias, but open-label designs can still colour how subjective side effects are reported.
- It tested semaglutide 1 mg, not 2.4 mg. The comparator was the diabetes dose of semaglutide, not the higher 2.4 mg dose approved for weight management. That is the right comparison for a diabetes trial, but it means the result should not be read as tirzepatide versus semaglutide at its maximum weight-loss dose.
- It measured surrogate outcomes over 40 weeks. A1c and weight are established markers, but SURPASS-2 was not a cardiovascular outcomes trial and did not run long enough to measure hard events such as heart attacks or strokes.
- It studied a specific population. Everyone was an adult with type 2 diabetes on metformin. The result applies most directly to people like those enrolled and does not automatically extend to other backgrounds or to people without diabetes.
None of this diminishes the finding. It keeps it the right size: strong evidence that tirzepatide out-performed semaglutide 1 mg on A1c and weight over 40 weeks in a defined population.
Why the SURPASS-2 trial matters
Most trials compare a new drug against placebo. SURPASS-2 is more informative because it put tirzepatide directly against semaglutide, a drug already regarded as one of the most effective GLP-1 agonists. Showing noninferiority and then superiority against that benchmark is a stronger claim than beating an inert control.
The result helped establish tirzepatide's dual GIP/GLP-1 mechanism as more than a theoretical upgrade over selective GLP-1 receptor agonists, at least on the metabolic markers measured here. It shaped how clinicians and regulators positioned the drug for type 2 diabetes and set expectations that later carried into the dedicated obesity trials. For the full side-by-side, see tirzepatide vs semaglutide; for the weight-loss evidence map, see peptides for weight loss; and to keep exploring the science, browse the full PepMate research library.
Frequently asked questions
What did the SURPASS-2 trial show?
SURPASS-2 was a head-to-head trial in adults with type 2 diabetes on metformin. Once-weekly tirzepatide at 5, 10 and 15 mg was noninferior and superior to semaglutide 1 mg for lowering glycated hemoglobin (A1c) over 40 weeks, and it produced greater reductions in body weight. It was one of the first trials to directly compare tirzepatide, a dual GIP/GLP-1 receptor agonist, against a selective GLP-1 receptor agonist.
What doses were compared in SURPASS-2?
Participants were randomly assigned in a 1:1:1:1 ratio to once-weekly subcutaneous tirzepatide at 5 mg, 10 mg or 15 mg, or to once-weekly semaglutide at 1 mg. Reporting these assigned regimens describes what the trial administered — it is not a dosing recommendation. Both drugs are prescription-only, and dose selection is a clinical decision made with a clinician.
How much more did tirzepatide lower A1c than semaglutide?
The estimated mean A1c change from baseline was -2.01, -2.24 and -2.30 percentage points for tirzepatide 5, 10 and 15 mg, versus -1.86 percentage points for semaglutide 1 mg. The differences between each tirzepatide dose and semaglutide were -0.15 (95% CI -0.28 to -0.03; P=0.02), -0.39 (95% CI -0.51 to -0.26; P<0.001) and -0.45 percentage points (95% CI -0.57 to -0.32; P<0.001). Tirzepatide was noninferior and superior at all doses.
Did tirzepatide cause more weight loss than semaglutide in SURPASS-2?
Yes. Reductions in body weight were greater with tirzepatide than with semaglutide. The least-squares mean estimated treatment differences favouring tirzepatide were -1.9 kg, -3.6 kg and -5.5 kg for the 5, 10 and 15 mg doses, respectively, with P<0.001 for all comparisons. Weight change was a secondary outcome; the primary endpoint was A1c.
Was SURPASS-2 a blinded trial?
No. SURPASS-2 was an open-label trial, meaning participants and investigators knew which drug was being given. That is a real limitation for subjective outcomes, though the primary endpoint — glycated hemoglobin — is a laboratory measurement that is hard to influence by knowing the treatment. It is worth keeping in mind when interpreting side-effect reporting.
Does SURPASS-2 mean tirzepatide is better than semaglutide for everyone?
No. SURPASS-2 compared specific doses over 40 weeks in adults with type 2 diabetes on metformin, and it tested semaglutide 1 mg rather than the higher 2.4 mg dose used for weight management. Its results apply most directly to people like those enrolled. Whether either drug is appropriate for any individual is a clinical decision that weighs personal history, other conditions, tolerability and cost. This page is educational and not medical advice.
Sources
This summary traces to the peer-reviewed publication and trial registration:
- Tirzepatide versus Semaglutide Once Weekly in Patients with Type 2 Diabetes (SURPASS-2) — New England Journal of Medicine, 2021;385(6):503-515; PubMed 34170647.
- SURPASS-2 (tirzepatide vs semaglutide add-on to metformin) — ClinicalTrials.gov NCT03987919.
- Drug reference record: tirzepatide.