SURPASS-4 enrolled people with type 2 diabetes and elevated cardiovascular risk. That setting does not make every favorable result proof of cardiovascular protection. Primary report
Study snapshot
This open-label trial randomized 2,002 participants to tirzepatide or insulin glargine; 1,995 received treatment. The primary comparison concerned HbA1c at 52 weeks, while treatment could extend to 104 weeks. Glucose reductions favored tirzepatide. The four-component cardiovascular-event estimate was a hazard ratio of 0.74, with a 95% confidence interval of 0.51–1.08. Gastrointestinal events were more common with tirzepatide, while the specified clinically significant or severe hypoglycemia outcome was less common. Primary report
Our appraisal: do not promote an uncertain secondary signal
The cardiovascular interval includes the no-difference value of one. A headline claiming cardiovascular superiority from this estimate would convey more certainty than the result supports. Conversely, failure to establish superiority is not proof that the treatments have identical effects.
The follow-up period matters as much as the endpoint name. A one-year glucose comparison and a longer event-surveillance analysis should not be placed in one table under a single unlabeled time horizon.
A balanced reading also keeps different harms separate. Fewer hypoglycemia events cannot be substituted for an overall assertion of fewer adverse effects when other adverse-event categories move in the opposite direction. Before release, the complete report should be checked for event adjudication, analysis hierarchy, treatment exposure, and the precise cardiovascular composite definition.
Frequently asked questions
Was cardiovascular superiority the primary endpoint?
No. The primary comparison concerned HbA1c.
Did the cardiovascular confidence interval exclude one?
No. Its upper limit was 1.08.
Does that prove identical cardiovascular effects?
No. An uncertain difference is not an equivalence finding.
Were all endpoints measured over the same period?
No. The report includes different assessment windows.
Were gastrointestinal events absent?
No. They were more frequent with tirzepatide.
Does lower hypoglycemia mean lower risk in every category?
No. Each harm needs its own assessment.
How complete is this review?
It is bounded to the indexed primary abstract, with full-report review pending.