PeerJ meta-analysis confirms regain after withdrawal and improvement in side effects

A systematic review and meta-analysis published in PeerJ has confirmed what earlier clinical trials suggested: people who stop taking GLP-1 receptor agonist medications for weight loss tend to regain a substantial share of the weight they lost, while gastrointestinal side effects tend to improve after they quit the drug ^[1][2].

The researchers pooled data from six randomized controlled trials involving 8,993 patients, comparing 5,553 people who stopped treatment against 3,440 who continued it ^[1]. They searched PubMed, Embase, Cochrane Library, and Scopus for studies through December 2025, and included trials of tirzepatide, semaglutide, liraglutide, and dulaglutide ^[1].

The weight difference between those who stopped and those who kept taking the drugs was large. The mean difference was 17.90%, with a 95% confidence interval of 14.11 to 21.69, a result the authors describe as statistically significant ^[1]. Looking only at people who stopped treatment, the percentage difference in body weight before and after discontinuation was 9.11%, with a 95% confidence interval of 7.91 to 10.30 ^[1]. Both findings point the same direction: stopping the drug is followed by meaningful weight regain ^[1][2].

The analysis also found that the size of the rebound was not the same across drugs. Subgroup analysis showed that weight regain after stopping tirzepatide (the active ingredient in Mounjaro and Zepbound) was significantly larger than after stopping semaglutide (the active ingredient in Ozempic, Wegovy, and Rybelsus) ^[1]. The authors suggest this may reflect differences in how the two classes of drugs work in the body, since tirzepatide acts on both GLP-1 and GIP receptor pathways while semaglutide acts on GLP-1 alone ^[1].

A less-discussed part of the analysis concerns side effects. The pooled data showed that after people stopped their medication, their overall rate of adverse reactions went down, and gastrointestinal adverse reactions specifically decreased ^[1]. The analysis found no evidence that the rate of cardiovascular events was affected one way or another by stopping the drug ^[1]. The authors note this side-effect finding is something patients frequently ask about but that the original withdrawal trials often did not report as a headline result ^[1].

The study's authors, writing under the name Qi et al., disclosed no competing interests ^[1]. The paper places its findings in the context of a debate over whether cycling on and off these drugs is safe long-term, noting that some earlier research suggested weight cycling itself could carry health risks, while newer evidence suggests the benefits of weight loss generally outweigh the risks tied to weight fluctuation, and that the harms of sustained obesity are greater than those from cycling ^[2].

Why it matters for patients

For people currently taking a GLP-1 drug for weight loss, this analysis adds numbers to a pattern many have already heard about: stopping the medication is likely to be followed by weight regain, and the amount lost tends to come back substantially within roughly a year, consistent with prior estimates of 55% to 65% regain in that timeframe cited in the underlying research ^[2]. The finding that tirzepatide-based regain was larger than semaglutide-based regain in this pooled analysis may be relevant for people weighing which drug class they are on, though the analysis does not explain why the two differ ^[1].

The side-effect finding may be reassuring to people who experience gastrointestinal problems like nausea or diarrhea while on these drugs. The data show those symptoms tend to ease once the drug is stopped, and stopping did not appear to raise the rate of cardiovascular problems in the pooled trials ^[1]. This does not mean stopping is free of consequences, since the weight regain finding shows the opposite is true for body weight.

What happens next

The review's authors note that follow-up data after GLP-1 discontinuation remain limited and highly variable across existing studies, meaning future trials with longer follow-up periods could refine these estimates ^[1][2]. No specific dates for new trials were given in the sources reviewed here.