Did tirzepatide beat dulaglutide in the cardiovascular outcomes trial?
A clinical trial can succeed without demonstrating that one treatment beats another.
A clinical trial can succeed without demonstrating that one treatment beats another. SURPASS-CVOT compared tirzepatide with an active treatment, dulaglutide, in people with type 2 diabetes and established atherosclerotic cardiovascular disease. Its result makes sense only if you know which statistical question came first.
What margin did the design use?
A noninferiority trial sets a boundary for how much worse a new treatment could be than the comparator on a specific outcome while still meeting the study's criterion. Investigators compare the uncertainty interval around the observed effect with that prespecified boundary. Failure to find a statistically significant difference in an ordinary superiority trial would not prove noninferiority. The new trial had to be designed and analysed for this purpose.
Read the result the authors actually reported
The cardiovascular outcomes investigators found that tirzepatide met their noninferiority criterion against dulaglutide on the composite of cardiovascular death, myocardial infarction or stroke. Superiority was not demonstrated. A favorable-looking point estimate can fall short of the separate superiority test because of its uncertainty. A headline saying that tirzepatide prevented more cardiovascular events with established certainty would therefore be stronger than the reported primary result.
| Read the two statistical questions | What to look for |
|---|---|
| Population | Type 2 diabetes and atherosclerotic cardiovascular disease |
| Primary event | Cardiovascular death, myocardial infarction or stroke |
| Result | Noninferiority demonstrated; superiority was not established |
What this comparison cannot answer
The trial did not compare either medicine against no treatment in an unselected Bali population. Its participants had both diabetes and established vascular disease. A subgroup analysis or a separate study of weight change answers another question. Clinical decisions also include tolerability and other outcomes. The investigators observed more gastrointestinal adverse events with tirzepatide, a relevant detail that cannot be read from the cardiovascular point estimate.
Make a question for your own appointment
Ask the clinician whether the population and cardiovascular endpoint resemble your situation, and which separate outcomes matter for the decision. Write “noninferior on the composite in the studied population” on the research card. Keep product details, clinical eligibility and evidence about other outcomes on separate lines. That preserves a useful long-term outcome without turning the name on a Bali catalogue listing into a promise.
