The SURMOUNT-4 results, in short: people who stayed on tirzepatide kept their weight off and lost a little more, and people switched by chance to a placebo regained a lot of it. After 36 weeks on the drug, 670 adults had lost an average of 20.9% of their body weight. Over the next year, weight rose 14.0% in the placebo group and fell another 5.5% in the group that stayed on tirzepatide. Both groups had diet and exercise counseling the whole way through.
For the practical side of stopping, see what happens when you stop taking tirzepatide. Here, the trial itself.
What question did the SURMOUNT-4 trial ask?
Tirzepatide, the medicine in Zepbound and Mounjaro, was already known to lower weight. SURMOUNT-4 asked what comes next: once the weight is off, does the medicine have to stay on to keep it off?
The design is called randomized withdrawal. Everyone started on the drug. At week 36, a computer split them by chance: half kept getting tirzepatide, half got a placebo, a look-alike shot with no medicine in it. That phase was double-blind, meaning neither participants nor staff knew who got which.
Louis Aronne and colleagues published the results in JAMA (online December 2023, print January 2024); the trial is registered as NCT04660643. Eli Lilly, which makes tirzepatide, paid for it and took part in its design and analysis.
Who was in it, and who made it to the switch
The trial ran from 2021 to 2023 at 70 sites in four countries. Adults qualified with a BMI (a height-to-weight screening number) of 30 or more, or 27 or more plus a condition such as high blood pressure or sleep apnea. People with diabetes were excluded.
783 people started. Doses began at 2.5 mg and stepped up every four weeks toward each person's maximum tolerated dose, the highest step they could handle, of 10 or 15 mg. During those 36 weeks, 113 people stopped the drug, most often because of side effects or because they left the study.
Only people who reached 10 or 15 mg were randomized: 670 of them, 92.7% on 15 mg. They averaged 48 years old, 71% were women, and they began at 107.3 kg (about 237 pounds), down to 85.2 kg (about 188 pounds) by week 36.
That filter matters more than any other detail. The randomized group was, by design, people who had stuck with the drug and handled its top doses. The authors say so themselves: those people "may represent a subgroup of the general population."
How the trial measured weight maintenance
The primary endpoint, the main result chosen in advance, was the percentage change in weight over the year after the switch, from week 36 to week 88. A key second measure asked how many people kept at least 80% of the weight they had lost.
The headline figures count everyone in their assigned group, including people who stopped their shots early. Counting only time on treatment gave slightly larger numbers, −6.7% versus +14.8%. Counseling continued throughout, and nobody's dose could change after the switch.
SURMOUNT-4 results by the numbers
| After the switch (weeks 36 to 88) | Stayed on tirzepatide | Switched to placebo |
|---|---|---|
| Weight change | −5.5% (about −10 lb) | +14.0% (about +24 lb) |
| Waist change | −4.3 cm | +7.8 cm |
| Kept at least 80% of earlier loss | 89.5% | 16.6% |
| Total change from the very start | −25.3% | −9.9% |
| At least 20% below starting weight | 69.5% | 12.6% |
The gap between the groups was 19.4 percentage points.
Read the 14.0% carefully. It is 14% of the weight people had at the switch, not 14% of the pounds they had lost. The placebo group finished about 10% below its starting weight after being about 21% below at week 36, so on average roughly half the loss came back within a year. Most came early: that group was already up 10.0% by week 64, 28 weeks after the switch.
The continuing group kept drifting down, just more slowly, the flattening described in why a tirzepatide plateau happens.
Health markers followed the scale. In the Zepbound label's summary of the trial, systolic blood pressure (the top number) fell about 11 points in the first 36 weeks, then rose about 8 points over the next year on placebo, against 2 points on tirzepatide. The authors wrote that much of the early improvement in blood pressure, blood sugar and cholesterol "had been reversed."
Side effects in SURMOUNT-4
An adverse event is any health problem reported during a trial, whether or not the drug caused it. In the first 36 weeks, when everyone was on tirzepatide, the most common were nausea (35.5%), diarrhea (21.1%), constipation (20.7%) and vomiting (16.3%), and 7.0% of those enrolled stopped because of side effects.
After the switch, stomach symptoms eased but didn't vanish:
| Weeks 36 to 88 | Tirzepatide | Placebo |
|---|---|---|
| Any adverse event | 60.3% | 55.8% |
| Diarrhea | 10.7% | 4.8% |
| Nausea | 8.1% | 2.7% |
| Vomiting | 5.7% | 1.2% |
| Serious adverse events | 3.0% | 3.0% |
| Stopped because of side effects | 1.8% | 0.9% |
No cases of pancreatitis were confirmed. Three people on placebo had gallbladder inflammation, against none on tirzepatide. Three on tirzepatide had a major cardiovascular event confirmed by an independent committee, against none on placebo; the trial was built to measure weight, and counts this small can't settle a question like that. One person in each group died, and investigators judged neither death related to the study drug.
What critics and later studies said
The trial's own limits. Beyond the filter on who got randomized, doses were fixed after the switch and intensive behavioral therapy wasn't tested. Weight at week 88 was missing for 13.7% of the placebo group and 7.5% of the tirzepatide group and was filled in statistically, according to the current Zepbound label. And with one year of follow-up, the trial can't say where regain levels off.
Regain and health markers moved together. A post hoc analysis (one done afterward, not planned in advance) followed 308 placebo-group participants who had lost at least 10%. Within the year, 82.5% regained at least a quarter of what they had lost. Waist size rose 0.8 cm in those who regained under a quarter and 14.7 cm in those who regained three-quarters or more. Systolic pressure rose 6.8 to 10.4 points across the groups, even among people who kept most of the weight off (Horn et al., JAMA Internal Medicine 2026).
Stepping down instead of stopping. A later Lilly trial, SURMOUNT-MAINTAIN, tested cutting to 5 mg rather than switching to placebo (Horn et al., Lancet 2026); the guide to stopping tirzepatide walks through what it found.
The newest pooled look. As of September 23, 2026, the day this article was written, SURMOUNT-4 had just been folded into a new meta-analysis (a study that pools other studies), published online August 31, 2026. Across six studies and 1,776 people who stopped semaglutide or tirzepatide, it estimated regain at about 1 kg (2.3 pounds) a month, half the lost weight back by about 7.5 months. The authors caution that anything past 52 weeks is a projection, not an observation (Kow et al., Endocrinology, Diabetes & Metabolism 2026).
What SURMOUNT-4 means for a person
The authors' reading: tirzepatide worked while it was being taken, and for most people the effect faded after it stopped. They compared obesity to type 2 diabetes and high blood pressure, conditions that usually need long-term treatment.
As of September 23, 2026, the day this article was written, the current Zepbound label (Revised 08/2026) describes SURMOUNT-4, by its registry number, as "Study 4," and Zepbound's approved use reads "to reduce excess body weight and maintain weight reduction long term."
None of that makes regain a personal failing: counseling ran through the whole trial, and averages hide a wide spread. Tirzepatide weight-loss results shows the on-treatment curve; the clinical trials overview places SURMOUNT-4 among the rest. Whether to continue, step down or stop is a decision for a patient and prescriber together.
What SURMOUNT-4 can't say about compounded tirzepatide
The trial used Lilly's branded tirzepatide under a controlled protocol. Through Promise, tirzepatide is dispensed as a compounded medication, which is different from an FDA-approved product: the formulation offered here is not FDA-approved. The FDA says it does not verify compounded drugs' safety, effectiveness or quality before they are marketed, so SURMOUNT-4's averages can't show that a compounded preparation will perform the same way.
A licensed provider may still prescribe a compounded formulation; that decision is between the patient and the doctor. At Promise, a licensed provider reviews every request, and not everyone qualifies.
