Summary: In a manufacturer-funded post hoc exploratory analysis of a 28-week double-blind randomized trial in adults with type 2 diabetes on metformin, tirzepatide 15 mg was associated with a greater improvement in insulin sensitivity (clamp M value) than semaglutide 1 mg (p < 0.001), and the gain per unit weight loss was steeper (regression slope difference p = 0.0461). The benefit was not fully explained by greater weight or fat loss.
PICO Summary
| Element | Detail |
|---|---|
| Population | Adults with type 2 diabetes treated with metformin; post hoc exploratory analysis of a 28-week, double-blind randomized controlled trial. |
| Intervention | Once-weekly tirzepatide 15 mg. |
| Comparison | Once-weekly semaglutide 1 mg (a placebo arm was also present in the parent trial). |
| Outcome | Insulin sensitivity by hyperinsulinemic euglycemic clamp (M value) improved more with tirzepatide than semaglutide (p < 0.001). The correlation between insulin-sensitivity change and percent weight change was stronger for tirzepatide (R = -0.656, p < 0.0001) than semaglutide (R = -0.268, p = 0.0820); the between-drug correlation difference was p = 0.0242. In regression, the slope relating change in M value to change in weight differed between agents (p = 0.0461). Findings were echoed using the Matsuda index and change in fat mass. No 95% CIs, absolute risk reduction, or NNT were reported for these continuous, exploratory endpoints. |
Expert Commentary
This is an exploratory, post hoc analysis, and it should be read as hypothesis-generating rather than as confirmatory evidence of superiority. Within those limits, the signal is internally consistent: tirzepatide was associated with a greater clamp-measured improvement in insulin sensitivity than semaglutide, the relationship with weight loss was steeper, and the authors argue that the benefit was not fully accounted for by greater weight or fat loss. That residual effect is mechanistically plausible given dual GIP and GLP-1 receptor agonism, but it is inferred from regression behaviour, not from a prespecified head-to-head endpoint. Two cautions temper enthusiasm. First, the analysis was funded by the manufacturer of tirzepatide, and several authors are company employees, so the framing favours the sponsor’s product and warrants independent replication. Second, the comparison rests on continuous correlations and slopes in a modest clamp substudy, with no confidence intervals or effect sizes that translate into clinical thresholds, and the semaglutide weight correlation itself was not statistically significant. Can I use this with my patients? Not yet as a basis for switching therapy; it does not establish that tirzepatide produces better glycaemic or cardiometabolic outcomes than semaglutide, only that a mechanistic insulin-sensitivity signal differs. I would await prospective, independently funded studies with predefined metabolic endpoints before letting these data drive prescribing.
References
Mather KJ, Mari A, Weerakkody G, Heise T, DeVries JH, Urva S, et al. Greater improvement in insulin sensitivity per unit weight loss associated with tirzepatide versus semaglutide: An exploratory analysis. Diabetes Obes Metab. 2025;27(3):1507-1514. doi:10.1111/dom.16159