SURMOUNT-4 asks a maintenance question after an initial treatment period. It should not be read as though every participant was randomized at their first treatment visit. Primary trial record
Two phases, different starting points
The study enrolled adults with obesity or overweight and a weight-related complication, without diabetes. After a 36-week open-label tirzepatide lead-in, 670 participants were randomized to continued tirzepatide or placebo for 52 weeks. Mean weight change from randomization to week 88 was −5.5% with continuation and +14.0% with placebo. The reported between-group difference was −19.4 percentage points. Gastrointestinal adverse events were common. Primary trial record
Our interpretation: label the baseline
A percentage is incomplete without its denominator and starting time. Weight change from entry into the lead-in is not the same calculation as weight change from randomization. Combining those estimates in one unlabeled chart could make a correct number communicate the wrong result.
Consider an invented example: a person starts at 100 units, reaches 80 units, and later reaches 88 units. The final value is 12% below the initial value but 10% above the intermediate value. Both statements are arithmetically correct. They answer different questions. These numbers are illustrative, not trial results.
Why the lead-in affects interpretation
When assessing generalizability, inspect who entered the study, who reached randomization, and why others did not. The randomized phase describes the population that reached that phase under the study's rules. A useful evidence table preserves that distinction instead of treating enrollment and randomization counts as interchangeable.
The full report should also be checked for treatment adherence, handling of missing weights, prespecified analyses, and participant flow. Those checks help interpret the estimate; they are not an assertion that the investigators omitted them.
Frequently asked questions
What kind of study was this?
A randomized withdrawal trial following an open-label lead-in.
Does the +14.0% figure start at original enrollment?
No. It describes the randomized period beginning at week 36.
Is a percentage-point difference the same as percent change?
No. A difference between percentages must retain its units.
Did everyone entering the lead-in reach randomization?
No. Participant flow needs to be reported separately for the two phases.
Does the mean predict an individual's trajectory?
No. A group estimate is not an individual forecast.
Should someone change treatment based on this page?
No. Medication changes belong with the prescribing clinician, considering the person's circumstances.
What is this record's scope?
The primary indexed report, not every later analysis or a personalized maintenance plan.