Summary
Background
Glucagon-like peptide 1 receptor agonists (GLP-1RAs) have emerged as breakthrough agents for weight loss. However, discontinuation is common, and clinical trials have demonstrated significant weight regain following cessation. In this systematic review, we aimed to characterise the trajectory of weight regain after GLP-1RA cessation.
Methods
This systematic review and meta-regression analysis followed Cochrane and PRISMA guidelines. We searched MEDLINE, Embase, Cochrane Library, Scopus and Web of Science from inception to August 28, 2025 for randomised controlled trials (RCTs), non-randomised interventional studies and observational studies reporting weight outcomes after cessation of GLP-1RAs in adults with overweight or obesity. Weight regain was the primary outcome and was modelled using nonlinear regression. Secondary, exploratory outcomes included HbA1c and systolic blood pressure. The study protocol is registered with PROSPERO (CRD420250631751).
Findings
A total of 48 relevant studies were included. Weight consistently rebounded after cessation of GLP-1RAs. Six RCTs with 3236 participants were included in the nonlinear meta-regression, which utilised a mixed-effect exponential recovery model with random effects applied to the rate constant. At 1 year post-cessation, 60% of the weight lost during treatment was regained. Beyond 52 weeks, weight trajectories were extrapolated, with weight regain estimated to plateau at 75.3% (95% CI 68.9–81.6) of the weight lost on treatment. The rate constant was 0.0302 per week (95% CI 0.0202–0.0401), corresponding to a half-life of 23.0 weeks (95% CI 17.3–34.3). Most studies were assessed to have a moderate risk of bias.
Interpretation
GLP-1RA cessation is associated with a predictable and decelerating pattern of weight regain, which appears to plateau below pre-treatment levels, suggesting that partial weight-loss benefit may persist long-term but is substantially attenuated.
Keywords
GLP-1RAs, Weight regain, Obesity, Treatment discontinuation, Anti-obesity pharmacotherapy, Meta-regression.
Introduction
The increasing prevalence of obesity represents a significant public health concern. Worldwide, more than one billion individuals are living with obesity. At the individual level, obesity has been linked to poor health outcomes and is associated with comorbidities including type 2 diabetes, cardiovascular disease, and metabolic dysfunction-associated steatotic liver disease.
Weight loss of 5–10% can mitigate obesity-related health complications. For patients with a body mass index (BMI) of 35 kg/m2 or higher, evidence-based guidance recommends a greater weight loss of 15–20% for sustained improvement in comorbidity. Lifestyle modifications, primarily diet and exercise, have historically been the cornerstone of weight management. However, despite being widely recommended, lifestyle modification alone is generally ineffective.
Until recently, pharmacological options for weight loss were limited, with either poor efficacy or tolerability. In the past few years, however, glucagon-like peptide 1 receptor agonists (GLP-1RAs) have emerged as highly effective weight loss drugs; clinical trials have demonstrated substantial weight losses of 15–20% with agents such as semaglutide and tirzepatide. In addition, patients treated with GLP-1RAs experience improvements in cardiometabolic parameters, including blood pressure, HbA1c, LDL, and triglycerides.
Despite the successes of GLP-1RAs, approximately half of patients who initiate them discontinue within the first year, likely due to their potential gastrointestinal side effects as well as limited access under insurance coverage policies and national prescribing guidelines.
Outcomes after discontinuing GLP-1RAs have been underexplored in the literature. Two reviews reported weight regain after GLP-1RA cessation but neither modelled the trajectory of weight over time. While clinical trials such as STEP 4 and SURMOUNT-4 have provided insights into post-cessation outcomes, additional relevant data exist across a wider body of trials with greater diversity in drugs and populations. Understanding the nature of weight regain could aid the development of strategies to maintain weight loss after cessation. Therefore, this review aimed to synthesise existing evidence and characterise the typical trajectory of weight regain following cessation of GLP-1RA treatment.
Methods
Search strategy and selection criteria
This systematic review and meta-regression followed the Cochrane Handbook for Systematic Reviews. Findings were reported according to PRISMA guidelines. The protocol is registered with PROSPERO (CRD420250631751).
SL conducted a systematic search in MEDLINE, Embase, Cochrane Library, Scopus and Web of Science from their inception to Aug 28, 2025. We aimed to identify studies which tracked weight outcomes in patients following discontinuation of GLP-1RA therapy. A broad search strategy was used because post-cessation outcomes were often absent from abstracts despite being reported in full texts or supplements.
The search queries comprised variations of the term ‘glucagon-like peptide 1 receptor agonist’ in combination with keywords related to weight regain and treatment cessation. In addition, supplementary searches were conducted by SL in Google Search and Google Scholar to identify further relevant studies not included in the primary databases. Conference abstracts and other grey literature were included if they met the inclusion criteria. Protocol papers and conference abstracts were followed up for full-text publications which, if eligible, were subsequently included in the review. Non-English studies and reviews were excluded from the search results. The complete search strategy is provided in the Appendix (p. 2).
Citations were imported into EndNote, deduplicated and then uploaded to Rayyan for screening. Six authors conducted title and abstract screening as well as full-text screening. Two reviewers independently screened each record and conflicts were resolved through discussion.
Studies were eligible for inclusion if they were randomised controlled trials (RCTs), non-randomised interventional studies or observational studies that reported weight outcomes following cessation of GLP-1RA treatment in adults with overweight or obesity, defined as a BMI of ≥25 kg/m2 or ≥23 kg/m2 for specific populations. Studies were required to have a treatment period of at least 8 weeks and a post-cessation follow-up period of at least 4 weeks, and could not include co-treatment with other FDA-approved weight loss drugs. Case reports, protocol-only publications, reviews, editorials and non-English publications were excluded.
Discussion
The findings of this review indicate that there is significant weight regain following cessation of GLP-1RAs. The meta-regression shows that after 1 year of treatment withdrawal, participants regain 60% of the weight they lost during treatment. The trajectory of regain is nonlinear and decelerating in nature, with an initial rapid recovery followed by a gradual tapering off and eventual plateauing. Notably, the estimated plateau is below the pre-treatment baseline weight, indicating that some beneficial effects may persist at the population level beyond 1 year after cessation, potentially indefinitely based on model extrapolation.
Weight regain poses a major challenge to the long-term efficacy of GLP-1RAs after treatment cessation, which is particularly relevant as around half of patients discontinue treatment within the first year. Discontinuation is often due to gastrointestinal side effects, financial constraints, and limited access through insurance or prescribing guidelines. Despite this, public health bodies have yet to provide clear and coherent guidance on the use of GLP-1RAs in long-term weight management. In the UK, the National Institute for Health and Care Excellence (NICE) recommends that semaglutide should be prescribed for weight loss for a maximum of 2 years, but no strict limit is set for tirzepatide. This inconsistency highlights a blind spot in long-term obesity management as the benefits of GLP-1RAs risk being negated by subsequent weight regain.
While our model predicts considerable weight regain, it also suggests that 25% of the initial weight loss may be sustained in the long term. This residual weight loss corresponds approximately to a 4–5% reduction in body weight relative to the pre-treatment baseline, assuming an initial weight loss of 15–20% as seen in trials of semaglutide and tirzepatide. Notably, we observed that the residual weight loss was consistently greater than that seen in placebo groups. Clinically significant weight loss for individuals with a BMI of less than 35 kg/m2 has been defined as a 5–10% reduction in body weight, with the predicted residual weight loss of 4–5% falling just below this threshold. Nonetheless, GLP-1RAs may confer modest long-term metabolic benefits for some patients even after cessation.
If weight regain does indeed persist, this may be explained by sustained behavioural or physiological adaptations. By reducing appetite, GLP-1RAs may help facilitate the development of healthier eating habits, such as reduced portion sizes or improved nutritional quality, which may persist even after treatment is discontinued. GLP-1RAs may also facilitate long-term physiologic adaptations, such as altering hormone levels and hypothalamic resetting.
When deprescribing GLP-1RAs, physicians and patients should be aware of the potential for weight regain and consider strategies to mitigate this risk. One possible option is individualised dose tapering. This approach was explored by Seier et al., who treated patients with semaglutide until they reached their target weight loss, then gradually reduced the dosage to the minimum dose that maintained their weight. This strategy may improve long-term tolerability, however patients may be unable to stop treatment entirely. Another study is currently investigating whether weight regain after total discontinuation can be mitigated by interventions that include tailored meals and a mobile application tracking food intake and exercise.
Our results are subject to certain limitations and should be interpreted with caution. Most importantly, the trial data used to fit our model only extend to 52 weeks after cessation. No large-scale trials have reported weight regain data beyond this point; thus predictions made outside of this window are extrapolations. Longer-term trials are necessary to confirm that our predictions hold beyond one year.
We restricted our meta-regression analysis to studies reporting at least 3 kg on-treatment mean weight loss. This was a pragmatic threshold chosen to enable reliable estimation of post-cessation weight regain trajectories whilst also reflecting the typical weight loss seen in most GLP-1RA users. We acknowledge, however, that patients achieving more modest weight loss are a clinically important subgroup and their post-cessation weight trajectories may not mirror those presented in our analysis.
We were unable to run sensitivity analyses to determine the extent to which other variables such as concomitant medications, weight maintenance strategies or comorbidities affected the trajectory of weight regain. This was because data were typically not reported in sufficient detail, often only with broad summary statistics or not at all. In particular, reporting and implementation of post-cessation weight maintenance strategies was inconsistent amongst studies included in the meta-regression.
Assessment of publication bias was not feasible as the meta-regression included only six studies; funnel plot methods typically require at least ten studies for reliable interpretation.
We initially aimed to model trajectories of HbA1c and SBP following cessation, as exploratory outcomes, but this proved infeasible due to inconsistent reporting and the lack of a clearly defined trajectory. As a result, insights into these parameters were limited, mainly confirming expected post-cessation rebound trends. We note, however, that study inclusion was primarily based on the availability of post-cessation weight data. Consequently, the HbA1c and SBP results in this review may not capture a fully representative sample of post-cessation outcomes in these variables.
The topic of weight regain after GLP-1RA cessation has been largely underexplored in the literature. There are relatively few major trials investigating weight regain after cessation of GLP-1RAs; we identified only two large-scale RCTs that were designed to measure weight regain as a primary outcome: STEP 4 and SURMOUNT-4. Most of our included trials, including STEP 1and SURMOUNT-1, measured weight regain as a secondary, often exploratory outcome; even these represent a small subset of all GLP-1RA trials, which generally do not report any post-cessation outcomes.
Beyond weight regain, there are significant concerns about the long-term consequences of GLP-1RAs on body composition. Studies indicate that 40–60% of the weight lost during treatment is lean mass, with newer agents such as semaglutide and tirzepatide being less effective in preserving lean mass. In patients who discontinue treatment, it is unknown whether the same proportion of lean mass is recovered as no studies have specifically assessed post-treatment changes in body composition. We recommend that future trials investigate further the effect of GLP-1RAs on the composition of weight loss during and after treatment cessation and how these changes compare to other weight loss modalities.
Our findings paint a nuanced picture regarding long-term post-GLP-1RA outcomes. Although a significant proportion of the weight is regained after cessation, the results of our modelling suggest that some benefit is maintained for at least 1 year; however there is likely significant variation at the individual level. As Seier et al.demonstrated, while some patients may require long-term therapy at the full dose for weight maintenance, others can taper their dose to a reduced maintenance level, and some may be able to discontinue therapy entirely while maintaining their desired weight. These considerations underscore the importance of a flexible, individualised approach to weight management when using GLP-1RAs. Blanket guidelines such as NICE's 2-year limit for semaglutide may fail to accommodate patients' individual needs and risk undermining long-term effectiveness.
The landscape of weight loss medications is rapidly evolving as new agents are brought to market. Although included here as a GLP-1RA, tirzepatide is a dual agonist that also targets the glucose-dependent insulinotropic polypeptide (GIP) receptor, likely contributing to its superior efficacy. Even more potent agents are on the horizon. Currently undergoing clinical trials is retatrutide, a triple agonist of the GLP-1, GIP and glucagon receptors, which has produced weight loss of 24% after 48 weeks of treatment. If the weight dynamics in this review hold for newer agents, greater on-treatment efficacy may translate to sizeable residual weight loss even after discontinuation.
In the face of high discontinuation rates and policy-imposed treatment duration limits, our findings underscore the need for an urgent re-evaluation of prescribing guidelines. Individualised cessation strategies and long-term support are essential to sustain treatment benefits and prevent the reversal of clinical progress. As newer and more potent agents emerge, understanding the trajectory of discontinuation outcomes will be pivotal in shaping future obesity care.
Acknowledgements
ChatGPT was used to verify grammar, refine the English phrasing in our text and for coding assistance in data analysis. All output was manually reviewed and the authors take full responsibility for the accuracy and quality of the publication.