A cardiovascular comparison against an active treatment asks a different question from a placebo comparison. SURPASS-CVOT also illustrates why the statistical claim attached to a result matters. Primary trial
The trial result
The modified intention-to-treat population included 13,165 people with type 2 diabetes and established atherosclerotic cardiovascular disease. Tirzepatide was compared with dulaglutide for cardiovascular death, myocardial infarction, or stroke. The hazard ratio was 0.92, with a 95.3% confidence interval of 0.83–1.01. The prespecified noninferiority criterion was met; superiority was not. Gastrointestinal adverse events were more common with tirzepatide. Primary report
How to read the margin
The noninferiority rule used an upper confidence bound below 1.05. The superiority rule required the upper bound below 1. Those thresholds answer different questions. A result can satisfy the first without satisfying the second, exactly as happened here.
Noninferiority does not mean that the treatments are identical. Nor does a numerically favorable point estimate prove that one prevents more events. The interval communicates uncertainty around that estimate, and the prespecified testing framework determines the justified claim.
The active comparator is essential context. Replacing its name with “control” can invite an incorrect placebo interpretation. Conversely, evidence against this comparator does not automatically establish the same comparative result against every other drug in its class.
Participant flow also matters: randomized enrollment and the primary analysis population are not interchangeable numbers. A complete review should explain exclusions and assess whether they could influence a noninferiority conclusion.
For clinical interpretation, relative effects should eventually be accompanied by verified absolute event risks, follow-up duration, and tolerability. This record is an appraisal of the trial's principal cardiovascular claim, not a recommendation to select or switch treatment. The studied regulated products and protocol do not validate unverified research preparations.
Frequently asked questions
Was the comparator placebo?
No; it was dulaglutide.
Was noninferiority established?
Yes, under the trial's prespecified rule.
Was superiority established?
No.
Does noninferiority mean identical effects?
No.
Was this a diabetes population?
Yes, with established cardiovascular disease.
Were gastrointestinal effects relevant?
Yes.
What should accompany the hazard ratio?
The comparator, interval, endpoint, absolute risks, and observation period.