DailyDoctor Archive Specialties Get app
Back to the 5 September 2026 edition

Clinical update · 02 of 06

Half the weight comes back within eight months of stopping semaglutide or tirzepatide

Tell patients stopping semaglutide or tirzepatide that regain begins straight away at about 1 kg a month, with half the weight lost back within eight months, and arrange a maintenance plan before treatment ends.

This was a Bayesian hierarchical re-analysis of published timepoint-level data from six studies, ten intervention arms and 1,776 participants, modelling what happens to weight after semaglutide or tirzepatide is stopped. Observed follow-up ran from 4 to 52 weeks.

Weight loss at cessation averaged 15.35 kg (95% credible interval 11.18-19.60). Regain then ran at 1.04 kg per month (95% CrI 0.80-1.29). On a linear model, half the loss was back by 7.5 months and baseline weight by 15 months. Tirzepatide looked numerically faster than semaglutide before adjustment, but no independent drug difference survived it. Greater initial loss was directionally associated with faster regain, though that interval included zero.

The practical use is in the conversation before stopping, not after. Patients who stop for cost, supply or side effects should know the slope is steep and starts immediately, and should have a maintenance plan and a review date rather than an open-ended gap. Behavioural support was directionally associated with slower regain, but imprecisely, so offer it as sensible rather than proven.

  • About 15 kg lost at cessation, regained at roughly 1 kg per month
  • Half of the loss back by 7.5 months, baseline by 15 months
  • No clear difference between semaglutide and tirzepatide after adjustment
  • Book a review before stopping, not after the weight has returned

The statistics, in plain English

A credible interval is the Bayesian counterpart of a confidence interval, and the one on the regain rate, 0.80 to 1.29 kg per month, is narrow enough to plan around. The 15-month figure for full return is far less certain: observed data ran to 52 weeks at most, so anything beyond that is the model extending a straight line. Treat the early slope as measured and the endpoint as an estimate.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

QR code to install Daily Doctor
Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.

insulintype1pediatricglp1obesitysglt2type2cgm

Tomorrow morning, before your first patient

One edition a day for diabetes & endocrinology, written by the desk, every claim tied to its paper. Six minutes.

Get the app — free
Daily Doctor All 27 specialties, every morning. Free.
Get the app