SURPASS-3 addresses treatment intensification in a defined type 2 diabetes population. It is not a study of stopping established insulin treatment. Primary report
Study snapshot
The open-label, randomized trial compared tirzepatide with titrated insulin degludec over 52 weeks. Participants were insulin-naive and taking metformin, with or without an SGLT2 inhibitor. Of 1,444 randomized participants, 1,437 entered the modified intention-to-treat population. HbA1c reductions favored tirzepatide, and the prespecified noninferiority criterion was met. Weight decreased in the tirzepatide groups and increased in the degludec group. These findings describe trial treatment groups, not guaranteed individual outcomes. Primary report
Our appraisal: define the decision before comparing drugs
A comparison is only useful when the reader knows what decision was being tested. Starting an additional treatment in somebody not previously using insulin differs from replacing insulin in somebody already dependent on it. The eligibility criteria determine which of those questions a trial can answer.
An active comparator also requires scrutiny of how it was managed. A medicine's name alone does not describe its titration targets, adherence, or rescue-treatment rules. Those details belong beside the outcome estimates when a complete comparative assessment is prepared.
The difference between randomized and analyzed counts deserves a participant-flow check. It should not be silently erased by quoting one sample size throughout the page. Full release review must also inspect adverse events and missing outcomes; a favorable glucose result is not a complete benefit–harm evaluation.
Frequently asked questions
Was this a placebo comparison?
No. The comparator was insulin degludec.
Were participants already using insulin?
They were insulin-naive at enrollment.
Was tirzepatide studied without background medicines?
No. Participants continued specified oral diabetes treatment.
Did everyone randomized enter the reported analysis population?
No. The two reported counts differ.
Does HbA1c improvement establish fewer heart attacks?
No. That requires an appropriate clinical-outcome assessment.
Is this an instruction to replace insulin?
No. Treatment changes require individual clinical care.
What remains to be reviewed?
Complete methods, safety tables, missing-data handling, and publication updates.