Research

SURMOUNT-MAINTAIN appears in print in The Lancet

The SURMOUNT-MAINTAIN trial of tirzepatide as a weight-loss maintenance drug appeared in print in The Lancet, spelling out the statistical rules behind its widely cited -9.9% placebo figure.

By the Semaglutides news desk·

The SURMOUNT-MAINTAIN trial, a US study testing whether tirzepatide (the drug in Mounjaro and Zepbound) helps people keep weight off after they have already lost it, was published in print in The Lancet on June 6, 2026 [1]. The paper is described as a multicentre, double-blind, randomised, placebo-controlled trial conducted in people with obesity in the USA [1].

The print version lays out the statistical approach used to judge the trial's main result: a "modified treatment-regimen estimand" [1]. In plain terms, an estimand is the specific question a trial is built to answer. A modified treatment-regimen estimand estimates what would happen if people generally stayed on their assigned treatment as planned, while still accounting for people who stopped taking the drug or switched to other treatment during the study. This choice matters because it shapes how the topline numbers, including the placebo group's result, should be read.

The publication also details rescue-therapy rules that took effect starting at week 84 [1]. Under these rules, participants in the trial who were not maintaining an adequate response could be given additional treatment partway through the study. Because some placebo-group participants may have received rescue therapy after week 84, the placebo arm's reported change of -9.9% is not necessarily a clean measure of what happens to weight with no treatment at all over the full study period [1]. Readers trying to compare the drug arm and placebo arm need to know these rules were in place, since they can affect how much weight the placebo group appears to regain or keep off.

The source available for this article is the journal's abstract landing page, which confirms the trial's title, its estimand definition, and its rescue-therapy timing, but does not show the full numeric results for the tirzepatide arm, the size of the study population, or the exact proportion of placebo patients who received rescue therapy [1]. Those figures are not yet known from what is available here.

Why it matters for patients

For people who have lost weight on tirzepatide and are wondering what happens if they stop or reduce the dose, this trial is meant to give a direct answer. But the fine print about the estimand and the rescue rules is not a technicality — it changes what the -9.9% placebo number actually represents. A trial that lets some placebo patients get extra treatment partway through will likely show less weight regain in that group than a trial that does not, simply because some of those patients were no longer receiving nothing at all [1].

That means patients and clinicians comparing tirzepatide's maintenance effect against a true "no treatment" scenario should be cautious about reading -9.9% as a simple measure of natural weight regain after stopping medication. It reflects a mixed group, some of whom may have restarted treatment after week 84 [1].

What happens next

The print appearance in The Lancet on June 6, 2026, follows what appears to be an earlier release of results, since the article already reports a specific placebo-arm figure [1]. What is not yet known from the material reviewed here is the full breakdown of results for the tirzepatide arm, the number of participants who required rescue therapy, and whether the modified treatment-regimen estimand or a alternative approach was used as the trial's primary analysis in the version originally released. Readers looking for those details will need the full text of the published study.

Sources

  1. https://www.thelancet.com/article/S0140-6736(26)00656-2/abstract

Semaglutides.org is for information only and is not medical advice. Always talk to a licensed healthcare provider about your own care. Some links to telehealth services are affiliate links, labeled where they appear.