SURMOUNT-3 randomized people after they had already achieved a qualifying weight reduction. That sequence defines what the drug comparison can establish. Primary report
Study snapshot
After a 12-week intensive lifestyle period, 579 adults without diabetes who had lost at least 5% of their weight were randomized to tirzepatide or placebo. Over the subsequent 72 weeks, additional weight change favored tirzepatide. Gastrointestinal events were the most common adverse events. A later correction repaired swapped treatment labels in two figure panels. Primary report, correction
Our appraisal: keep the two baselines separate
The beginning of lifestyle treatment and the point of randomization are different reference points. A percentage change from one should not be presented as if calculated from the other. Simply adding percentages based on different weights can also produce an incorrect total.
Selection after the run-in limits applicability. The randomized result concerns participants who met the initial response criterion, not everyone who entered lifestyle treatment. People who did not qualify are outside that randomized comparison.
This does not diminish the question the study answered: whether an additional intervention helped after an initial successful phase. It does prevent an overly broad interpretation that the same sequence has been demonstrated for all patients regardless of early response or tolerability.
The figure correction matters for visual interpretation, but it should not be exaggerated into a claim that the entire study was invalidated. Use the corrected version and describe the actual scope of the notice. Full review should inspect enrollment flow, estimands, missing outcomes, and harms across both study phases.
Frequently asked questions
Did randomization occur before lifestyle treatment?
No. It occurred after the qualifying run-in.
Could everyone in the run-in enter the randomized phase?
No. A weight-response criterion applied.
Are the two baseline weights interchangeable?
No.
Can percentage changes with different denominators simply be added?
No.
Was a correction published?
Yes. It concerned swapped figure-key labels.
Did the correction retract the trial?
No. The notice describes a specific labeling repair.
What is the central applicability limit?
The randomized findings concern people who first achieved the qualifying lifestyle-associated reduction.