Tirzepatide versus Semaglutide Once Weekly in Patients with Type 2 Diabetes
Frías JP, Davies MJ, Rosenstock J, Pérez Manghi FC, Fernández Landó L, Bergman BK, Liu B, Cui X, +2 more
New England Journal of Medicine (2021) · n=1879
In the only major head-to-head trial of the modern GLP-1 class, tirzepatide at all three doses produced superior glycemic control to semaglutide 1 mg, with the 15-mg dose producing nearly twice the weight loss — the trial that established the dual-incretin advantage as more than a marginal difference.
SURPASS-2 is the head-to-head trial that quantified what the dual GIP/GLP-1 receptor agonism in tirzepatide adds beyond GLP-1 monoagonism. The 40-week open-label Phase 3 trial randomized 1,879 patients with type 2 diabetes inadequately controlled on metformin in a 1:1:1:1 ratio to subcutaneous tirzepatide 5, 10, or 15 mg or to subcutaneous semaglutide 1 mg, all once weekly. The primary endpoint was change in glycated hemoglobin (HbA1c) from baseline. Tirzepatide met both noninferiority and superiority across all three doses: estimated mean HbA1c change was -2.01, -2.24, and -2.30 percentage points on tirzepatide 5/10/15 mg versus -1.86 percentage points on semaglutide 1 mg. The weight-loss difference was even larger — patients on tirzepatide 15 mg lost roughly twice the body weight of those on semaglutide 1 mg, an effect-size separation that supported the FDA approval of tirzepatide as Mounjaro for type 2 diabetes (May 2022) and informed the subsequent SURMOUNT obesity program. Gastrointestinal adverse events were most common across all four arms, mostly mild-to-moderate, and concentrated during dose escalation; hypoglycemia rates were low across both drugs, consistent with the glucose-dependent insulin-secretion mechanism.
The semaglutide comparator dose was 1 mg — the maximum approved T2D dose at the time of trial design — which is below the 2.4 mg dose used for chronic weight management (Wegovy) and approved subsequently. A head-to-head trial against semaglutide 2.4 mg would test the comparison at the higher semaglutide dose where the obesity-grade effect-size envelope sits, and that trial design has been called for repeatedly in the field; as of early 2026 it has not been published. The trial population was T2D patients on metformin background, not the obesity-without-diabetes population that drives most lay-press comparison between the two drugs. Open-label design (rather than double-blind) is appropriate for a study where injectable appearance differs and where patient-reported outcomes were not the primary endpoint, but it does affect the interpretation of secondary endpoints sensitive to expectancy. Industry sponsorship by Eli Lilly is disclosed; the primary endpoint is regulatory-grade and the manuscript is consistent with the FDA review.