Weight regain after tirzepatide: the two starting points in SURMOUNT-4

The question

People switched from tirzepatide to placebo in SURMOUNT-4 regained substantial weight on average, while those continuing treatment lost more. But the often-quoted “14% regain” needs its starting point attached: it describes weight change after randomization, not the percentage of previously lost weight that returned.

The trial had a 36-week open-label treatment period followed by 52 weeks of randomized continuation or withdrawal. Of 783 people entering the lead-in, 670 were randomized. Adults had obesity or overweight with a weight-related complication; diabetes was excluded. Mean weight reduction during the lead-in among randomized participants was 20.9%. From week 36 to week 88, mean change was −5.5% with continued tirzepatide and +14.0% after switching to placebo. Original SURMOUNT-4 report.

Start the timeline before reading the percentage

There are two reference dates: entry into the study and the later randomization point. A sentence can be numerically correct yet misleading if it quietly moves between them.

Here is a hypothetical illustration, not a trial participant. Someone starts at 100 kilograms and reaches 80 kilograms. A subsequent 14% increase from 80 kilograms is 11.2 kilograms, producing a weight of 91.2 kilograms. The person has regained 56% of the earlier 20-kilogram loss, while remaining 8.8% below the original 100 kilograms.

Three different percentages describe the same example because each divides by a different starting quantity. “14% weight gain,” “56% of the loss regained” and “8.8% below baseline” are not interchangeable. None of those illustrative calculations predicts an individual outcome or reproduces the trial's statistical analysis.

This distinction also explains why subtracting an initial loss percentage from a later gain percentage can give the wrong answer. The later percentage applies to a changed weight. Before combining figures, write down the denominator for each one.

Who reached the randomized comparison?

The randomized phase compared people who had already progressed through the lead-in. It was not a fresh randomization of all treatment starters. Lifestyle intervention continued in both groups. Lilly funded the study. Trial design and funding.

That design is useful for asking about continuation after an initial treatment period. It is less suited to answering what every new patient will experience from the first prescription onward. Readers should resist treating a selected follow-up group as if it were an unselected sample of everyone who might consider treatment.

There is a practical way to recognize this distinction in other studies: look for the event that happens immediately before randomization. Sometimes it is enrollment. Sometimes it is completion of earlier treatment. Those studies can both be rigorous while answering different questions.

Regain and metabolic changes are related questions

A later post-hoc analysis examined participants switched to placebo who had first lost at least 10% of body weight. Greater regain was associated with larger reversals in several cardiometabolic improvements, including waist circumference and blood-pressure measures. It grouped participants by the fraction of their prior weight loss regained. Post-hoc SURMOUNT-4 analysis.

That is an association within a subgroup, not a second randomization assigning people to regain different amounts. It adds useful context but cannot show that a particular regain threshold guarantees a particular change in someone's health. A category boundary in a research table is not automatically a clinical decision threshold.

For an individual, weight, blood pressure, glucose measures and daily functioning remain separate observations. A single scale reading cannot stand in for all of them. Nor does a group association justify assuming that every earlier benefit disappears completely or immediately after treatment stops.

What the trial cannot prescribe

SURMOUNT-4 does not provide a personalized tapering schedule, prove that everyone must remain on one medicine indefinitely, or establish the best alternative for someone who cannot continue it. Those are additional questions requiring their own evidence and clinical assessment.

For someone discussing a treatment change with a clinician, useful questions include which outcomes will be monitored, what follow-up is planned and how changes will be interpreted. This article does not recommend stopping, restarting or adjusting medication. It explains why a maintenance conversation needs more than a percentage copied from a headline.

FDA warns that unapproved GLP-1 products have not undergone its review for safety, effectiveness and quality. A research-material listing is not the finished medicine used in this trial. FDA product context.

For a different withdrawal design, read our semaglutide weight-regain comparison. For another denominator problem, see tirzepatide and lean mass.