SURPASS-2 deep-dive: tirzepatide vs semaglutide 1 mg head-to-head in T2D (Frias 2021)
Last verified 2026-05-27 · Phase 3 · Completed (results published Aug 5, 2021) · NCT03987919 ↗
SURPASS-2 (NCT03987919) is the head-to-head phase-3 trial that supported tirzepatide's May 2022 FDA approval for type 2 diabetes (Mounjaro) and established the molecule's superiority over the highest then-approved injectable semaglutide T2D dose. Eli Lilly randomized 1,879 adults with type 2 diabetes inadequately controlled on metformin monotherapy 1:1:1:1 to once-weekly subcutaneous tirzepatide 5 mg, 10 mg, 15 mg, or semaglutide 1 mg, in an open-label design with 40 weeks of treatment plus a 4-week safety follow-up. The primary endpoint was the change in HbA1c from baseline to week 40. Results published by Frías and colleagues in the New England Journal of Medicine on August 5, 2021 showed superiority on HbA1c at every tirzepatide dose, with mean reductions of 2.01, 2.24, and 2.30 percentage points on tirzepatide versus 1.86 on semaglutide, and an absolute weight-loss gap that widened with dose.
- Enrollment
- 1,879
- Duration
- 40 weeks of randomized treatment with a 4-week safety follow-up (last visit at week 44)
- Drug
- Tirzepatide (vs semaglutide 1 mg head-to-head)
- Population
- Adults aged 18 and older with type 2 diabetes, HbA1c 7.0% to 10.5%, on a stable metformin dose of at least 1,500 mg/day for ≥3 months, BMI ≥25 kg/m², and stable body weight (±5%) over the preceding 3 months. Other antihyperglycemic agents were not allowed at randomization. Mean baseline HbA1c was 8.28%, mean body weight 93.7 kg, and mean age 56.6 years across the four arms (n = 470/469/470/469).
Primary endpoint
Change from baseline in HbA1c at week 40 — 10 mg and 15 mg tirzepatide vs semaglutide 1 mg (treatment-regimen estimand, ITT)
Treatment arm
10 mg: −2.24 percentage points; 15 mg: −2.30 percentage points
Comparator
Semaglutide 1 mg: −1.86 percentage points
Treatment difference: Estimated treatment difference vs semaglutide: −0.39 pp (10 mg) and −0.45 pp (15 mg); P<0.001 for noninferiority and superiority for both doses
Mean baseline HbA1c was 8.28% across arms; the 15-mg arm dropped HbA1c by roughly two-and-a-third percentage points on the published treatment-regimen estimand. The CT.gov-posted efficacy estimand (per-protocol) is slightly larger at −2.37 (10 mg) and −2.46 (15 mg) vs −1.86 (semaglutide).
Secondary endpoints
| Endpoint | Treatment | Comparator | Difference |
|---|---|---|---|
| Change from baseline in HbA1c at week 40 — 5 mg tirzepatide vs semaglutide 1 mg Even the lowest tirzepatide dose beat semaglutide 1 mg on the primary glycemic outcome. | 5 mg tirzepatide: −2.01 percentage points | Semaglutide 1 mg: −1.86 percentage points | Treatment difference −0.15 pp (95% CI not in abstract); P=0.02 for superiority |
| Change from baseline in body weight at week 40 (kg) — all arms Monotonic dose-response across the tirzepatide arms; the 15-mg arm produced roughly twice the absolute weight loss of semaglutide 1 mg. | 5 mg: −7.8 kg; 10 mg: −10.3 kg; 15 mg: −12.4 kg (efficacy estimand) | Semaglutide 1 mg: −6.2 kg (efficacy estimand) | Published treatment-regimen estimated treatment differences: −1.9 kg (5 mg), −3.6 kg (10 mg), and −5.5 kg (15 mg); P<0.001 for all three comparisons |
| Proportion achieving HbA1c <7.0% at week 40 | 5 mg: 85.5%; 10 mg: 88.9%; 15 mg: 92.2% | Semaglutide 1 mg: 81.1% | Higher proportion meeting ADA glycemic target at every tirzepatide dose; gap largest at 15 mg |
| Proportion achieving HbA1c <5.7% (normoglycemia) at week 40 Normoglycemic restoration in a single dose of tirzepatide is roughly 2.5× the rate on semaglutide 1 mg — the regulatory hook for the 'diabetes remission' framing that dominates 2024-2026 ADA narrative. | 5 mg: 29.3%; 10 mg: 44.7%; 15 mg: 50.9% | Semaglutide 1 mg: 19.7% | Half of 15-mg tirzepatide participants returned HbA1c to the non-diabetic range vs ~1 in 5 on semaglutide |
| Proportion achieving weight loss ≥5% at week 40 | 5 mg: 68.6%; 10 mg: 82.4%; 15 mg: 86.2% | Semaglutide 1 mg: 58.4% | Nearly nine in ten 15-mg participants cleared the 5% clinical-relevance threshold for weight loss |
| Change from baseline in fasting serum glucose at week 40 (mg/dL) | 5 mg: −56.0 mg/dL; 10 mg: −61.6 mg/dL; 15 mg: −63.4 mg/dL | Semaglutide 1 mg: −48.6 mg/dL | Treatment differences of −7.4 to −14.8 mg/dL favoring tirzepatide |
| Daily 7-point self-monitored blood glucose change at week 40 (mg/dL) | 5 mg: −65.4; 10 mg: −70.6; 15 mg: −74.3 | Semaglutide 1 mg: −61.4 | Larger improvement in mean daily glucose at all tirzepatide doses |
| Hypoglycemia rate (blood glucose <54 mg/dL or severe), episodes per participant-year Hypoglycemia is uncommon with incretin therapy on metformin background, which lacks insulin or sulfonylurea co-administration. | 5 mg: 0.0102; 10 mg: 0.0046; 15 mg: 0.0202 | Semaglutide 1 mg: 0.0046 | Clinically significant hypoglycemia was rare in both groups; absolute rates were a fraction of an episode per person-year |
| Lipid panel — triglycerides at week 40 (Frias 2021 secondary) Lipid changes followed weight and glycemic gradients; no arm worsened any lipid parameter. | Reductions of approximately −16% to −24% across the tirzepatide arms in the published manuscript | Roughly −10% with semaglutide 1 mg | Larger reduction with tirzepatide; LDL fell modestly in all arms and HDL rose modestly in all arms, with directional advantage favoring tirzepatide |
| Blood pressure change at week 40 (Frias 2021 secondary) Magnitudes are in the range of low-dose antihypertensive addition. | Mean systolic BP fell ~4-6 mmHg across the three tirzepatide arms | Mean systolic BP fell ~2-3 mmHg with semaglutide 1 mg | Numerical advantage for tirzepatide consistent with the larger weight-loss gradient |
Adverse events
| Event | Treatment rate | Comparator rate |
|---|---|---|
| Nausea Most common AE in every arm. Rates broadly comparable between tirzepatide 5 mg and semaglutide 1 mg; rise modestly at higher tirzepatide doses. Primarily mild-to-moderate, clustered in titration. | 5 mg: 17.4% (82/470); 10 mg: 19.2% (90/469); 15 mg: 22.1% (104/470) | Semaglutide 1 mg: 17.9% (84/469) |
| Diarrhoea Second-most-common GI event in both arms. | 5 mg: 13.2% (62/470); 10 mg: 16.4% (77/469); 15 mg: 13.8% (65/470) | Semaglutide 1 mg: 11.5% (54/469) |
| Vomiting Numerically similar between tirzepatide 10/15 mg and semaglutide 1 mg. | 5 mg: 5.7% (27/470); 10 mg: 8.3% (39/469); 15 mg: 9.8% (46/470) | Semaglutide 1 mg: 8.3% (39/469) |
| Dyspepsia Similar incidence across all four arms. | 5 mg: 7.2% (34/470); 10 mg: 6.2% (29/469); 15 mg: 9.1% (43/470) | Semaglutide 1 mg: 6.6% (31/469) |
| Constipation Lower at higher tirzepatide doses, opposite of the dose-response seen with nausea. | 5 mg: 6.6% (31/470); 10 mg: 4.5% (21/469); 15 mg: 4.5% (21/470) | Semaglutide 1 mg: 5.8% (27/469) |
| Decreased appetite Reported as an AE but mechanistically expected from incretin agonism. | 5 mg: 7.4% (35/470); 10 mg: 7.2% (34/469); 15 mg: 8.9% (42/470) | Semaglutide 1 mg: 5.3% (25/469) |
| Clinically significant hypoglycemia (BG <54 mg/dL) or severe hypoglycemia Per the published abstract: low absolute rates in all arms because background therapy was metformin alone — no insulin or sulfonylurea. | 5 mg: 0.6%; 10 mg: 0.2%; 15 mg: 1.7% (of participants ever) | Semaglutide 1 mg: 0.4% |
| Serious adverse events (any) Numerically higher with tirzepatide; not concentrated in any single organ system. | 5 mg: 7.0% (33/470); 10 mg: 5.3% (25/469); 15 mg: 5.7% (27/470) | Semaglutide 1 mg: 2.8% (13/469) |
| Permanent discontinuation due to adverse event Gastrointestinal events were the dominant reason for discontinuation in every arm, consistent with the incretin class. | 5 mg: 6.0%; 10 mg: 8.5%; 15 mg: 8.5% (per Frias 2021 manuscript) | Semaglutide 1 mg: 4.1% |
| Deaths during study (any cause) Per CT.gov posting; the published manuscript stated deaths were judged unrelated to study drug by investigators. Numerical imbalance is small in absolute terms (12 vs 1) but visible because the trial was not powered for mortality. | 5 mg: 4 (0.9%); 10 mg: 4 (0.9%); 15 mg: 4 (0.9%) | Semaglutide 1 mg: 1 (0.2%) |
Clinical significance
SURPASS-2 is the trial that proved dual GIP/GLP-1 agonism is not a marketing claim but a measurable clinical advantage in head-to-head testing. Against semaglutide 1 mg — the highest then-approved injectable semaglutide T2D dose — tirzepatide produced 0.15 to 0.45 percentage points greater HbA1c reduction and 1.9 to 5.5 kg greater weight loss across its three doses, with a comparable GI-event profile and similarly low rates of clinically significant hypoglycemia. The proportion of participants restoring HbA1c to the non-diabetic range (<5.7%) was 51% on 15-mg tirzepatide versus 20% on semaglutide 1 mg, the largest single finding driving the 2022 FDA approval narrative. SURPASS-2 also set the analytic precedent later replicated in the obesity setting by SURMOUNT-5, which directly compared tirzepatide vs semaglutide for weight loss in non-diabetic adults.
Frequently asked questions
Why did SURPASS-2 compare tirzepatide to semaglutide 1 mg rather than 2 mg or 2.4 mg?
Semaglutide 1 mg weekly was the highest FDA-approved dose of injectable semaglutide for type 2 diabetes when SURPASS-2 was designed in 2018-2019. Ozempic 2 mg weekly was not approved for T2D in the United States until March 2022, after SURPASS-2 had finished enrolling and reported. Semaglutide 2.4 mg (Wegovy) is the obesity dose and was never approved for diabetes. So comparing tirzepatide against semaglutide 1 mg was the regulatory-relevant comparator at the time of trial design; a higher-dose semaglutide T2D comparator did not yet exist.
Is SURPASS-2 a weight-loss trial or a diabetes trial?
A diabetes trial. The primary endpoint was the change in HbA1c at 40 weeks, not body weight, and the population was adults with type 2 diabetes (HbA1c 7.0-10.5%) inadequately controlled on metformin monotherapy. Weight loss was a key secondary endpoint and is the most clinically newsworthy secondary result, but it was not the basis for the trial's success or failure under its prespecified primary analysis. The obesity-population equivalent — non-diabetic adults — is the separate SURMOUNT-5 head-to-head trial published in 2025.
Does the SURPASS-2 advantage hold against semaglutide 2 mg or 2.4 mg?
SURPASS-2 did not test those doses, so any answer is an indirect comparison. Semaglutide 2 mg (Ozempic max T2D dose, approved 2022) modestly outperformed semaglutide 1 mg on HbA1c by roughly 0.2 percentage points in SUSTAIN FORTE, narrowing but not closing the tirzepatide-vs-semaglutide gap reported at the 15-mg dose in SURPASS-2 (−0.45 pp). Semaglutide 2.4 mg (Wegovy, obesity indication) was directly compared with tirzepatide for weight loss in SURMOUNT-5 (Aronne 2025), which reported a sustained tirzepatide advantage in non-diabetic adults with obesity.
Why was the trial only 40 weeks?
The 40-week treatment period plus 4-week safety follow-up reflects the standard registrational design for diabetes phase-3 trials assessing HbA1c at near-steady-state. HbA1c reflects roughly 90 days of glycemic exposure, so 40 weeks captures the peak treatment effect with two and a half full HbA1c half-lives elapsed after dose escalation. Long-term outcomes including weight maintenance, cardiovascular events, and durability over multiple years are addressed in SURPASS-CVOT (still ongoing) and the SURMOUNT obesity series — not SURPASS-2.
Did tirzepatide cause more GI side effects than semaglutide 1 mg?
Marginally, and the absolute differences are small. Nausea ranged 17-22% on tirzepatide vs 18% on semaglutide 1 mg; diarrhea 13-16% vs 12%; vomiting 6-10% vs 8%. The largest signal is at the 15 mg tirzepatide dose, but rates of any single GI event were within a few percentage points of semaglutide. Discontinuation specifically due to adverse events ran roughly 6-9% on tirzepatide vs 4% on semaglutide 1 mg, with gastrointestinal events the leading reason in both groups. The class-wide trade-off of tolerability for efficacy applies; tirzepatide is not categorically worse-tolerated than semaglutide in this trial.
How did SURPASS-2 affect the Mounjaro approval and downstream tirzepatide trials?
SURPASS-2 was the head-to-head efficacy signal in Lilly's NDA package, alongside SURPASS-1 (monotherapy), SURPASS-3 (vs insulin degludec), SURPASS-4 (vs insulin glargine), and SURPASS-5 (add-on to insulin glargine). The FDA approved tirzepatide as Mounjaro for type 2 diabetes on May 13, 2022. The strength of the SURPASS-2 weight-loss signal — 12.4 kg on 15 mg tirzepatide vs 6.2 kg on semaglutide 1 mg in T2D — also helped justify the parallel SURMOUNT obesity program, which delivered Zepbound's November 2023 obesity approval and the 2025 SURMOUNT-5 head-to-head weight-loss trial.
Sources
References
- 1.Frías JP, Davies MJ, Rosenstock J, Pérez Manghi FC, Fernández Landó L, Bergman BK, Liu B, Cui X, Brown K; SURPASS-2 Investigators. Tirzepatide versus Semaglutide Once Weekly in Patients with Type 2 Diabetes. N Engl J Med. 2021. PMID: 34170647.
- 2.Eli Lilly and Company. Study Results: A Study of Tirzepatide (LY3298176) Versus Semaglutide Once Weekly as Add-on Therapy to Metformin in Participants With Type 2 Diabetes (SURPASS-2, NCT03987919). ClinicalTrials.gov. 2022. https://clinicaltrials.gov/study/NCT03987919?tab=results
- 3.U.S. Food and Drug Administration. FDA Approves Novel, Dual-Targeted Treatment for Type 2 Diabetes (Mounjaro, tirzepatide). FDA Press Announcement, May 13, 2022. 2022. https://www.fda.gov/news-events/press-announcements/fda-approves-novel-dual-targeted-treatment-type-2-diabetes
- 4.Jastreboff AM, Aronne LJ, Ahmad NN, et al.; SURMOUNT-1 Investigators. Tirzepatide Once Weekly for the Treatment of Obesity. N Engl J Med. 2022. PMID: 35658024.