Glucagon‐like peptide‐1 receptor agonists discontinuation is higher in individuals with overweight and obesity without type 2 diabetes

New-generation glucagon-like peptide-1 receptor agonists (GLP-1 RAs), including mono, dual and triple agonists, are transforming obesity care. Although the prevalence of obesity in the United States remains high at 40.3%, 2023 marked the first time in over a decade that the obesity rate did not rise.1 While it is unclear if GLP-1 RAs contribute to slowing obesity rates, studies show greater reductions in body weight with these medications compared to lifestyle intervention and previous medications.2, 3 However, despite many researchers and clinicians calling for their long-term use, emerging data reveal relatively high discontinuation rates. This is concerning, as rapid weight regain and an attenuation of cardiometabolic benefits (e.g., blood pressure, glycaemic status, blood lipids) are often observed when patients discontinue GLP-1 RAs.4 Understanding why patients discontinue GLP-1 RAs is critical for improving persistence and promoting long-term weight loss maintenance and associated health benefits. Although GLP-1 RAs were originally developed for the treatment of type 2 diabetes (T2D), their approval for chronic weight management has led to widespread use in populations with and without T2D. Initial clinical trials of new-generation GLP-1 RAs for weight management showed promising persistence rates, with nearly 90% of participants completing a 72-week tirzepatide trial.2 However, due to the highly controlled nature of clinical trials, these findings may not fully reflect real-world practices, particularly among individuals with obesity but without T2D.5-8 Real-world studies report one-year discontinuation rates as high as 74%,9 though substantial variability exists between studies. To better understand these discrepancies and the potential role of T2D in influencing discontinuation rates, we conducted a systematic review and meta-analysis of real-world studies published on or after the U.S. Food and Drug Administration's 2021 approval of new-generation GLP-1 RAs for chronic weight management. On 26 November 2024, we conducted a PubMed search to identify observational real-world studies reporting one-year discontinuation rates for GLP-1 RAs. The following search terms were used: (‘discontinuation’ [Mesh] or ‘persistence’ [Mesh] or ‘adherence’ [Mesh] or ‘discontinuation’ [tiab] or ‘persistence’ [tiab] or ‘adherence’ [tiab]) and (‘glucagon-like peptide 1’ [Mesh] or ‘glp1’ [tiab] or ‘glp-1ra’ [tiab] or ‘glucagon-like-peptide-1’ [tiab]) and (‘diabetes’ [tiab] or ‘diabetes mellitus, type 2’ [Mesh] or obesity’ [Mesh] or ‘overweight’ [tiab]). We also reviewed reference lists of relevant papers to ensure inclusion of all available studies. During this process, a preprint with a large sample size and a focus on discontinuation among individuals both with and without type 2 diabetes was also identified. Studies were excluded if they were published before 2021, or did not examine records extending to at least 2021, as this was the year that new-generation GLP-1 RAs, such as semaglutide, were approved for weight management in the United States. Articles reporting clinical trial data or only oral medications were excluded to maintain a focus on real-world data for injectable GLP-1 RAs. We specifically focused on one-year discontinuation rates, as this was the most commonly reported time frame. The literature search was conducted by GJG and independently reviewed by SSJ and a third party not involved as an author. A random-effects meta-analysis was performed to estimate the pooled one-year GLP-1 RA discontinuation rate across studies. The random-effects model accounted for heterogeneity between studies and was used to calculate the pooled estimate, 95% confidence interval (CI) and 95% prediction interval (PI). Heterogeneity (τ2) was assessed using established methods.10 We then conducted a meta-regression to explore the potential impact of T2D proportion on GLP-1 RA discontinuation rate. Analyses were performed using Python v3.11. Reasons for discontinuation and associated factors were qualitatively summarized. A PRIMSA flow diagram depicting the process of record identification, screening, eligibility and inclusion is provided in Figure S1. Our initial PubMed search revealed 628 studies, 323 of which remained after filtering for studies published on or after 2021. A total of 214 studies were excluded for improper type (e.g., systematic review, clinical trial, etc.). After evaluation of remaining full-text articles (n = 111), nine articles were deemed eligible for inclusion (Table S1).5, 7, 8, 11-16 Exceptions were made for two international studies examining earlier records that ended in 20199 and 2020,17 due to their specific focus on the use of GLP-1 RAs for weight loss. We also included a preprint describing one-year GLP-1 RA discontinuation patterns among a large cohort of adults with overweight or obesity, including individuals with and without T2D.6 Injectable GLP-1 RA specific patient characteristics and discontinuation numbers were extracted from studies comparing different medications.11-13, 16 The median sample size of included studies was 5103 (range:147–195 915; total:445897), and the one-year median discontinuation rate was 50.1% (range:15.5–74.0%). The pooled one-year discontinuation rate from the random-effects meta-analysis was 48.7% (95% CI:39.8% to 57.5%; 95% PI: 17.0% to 80.3%), as shown in Figure 1. Significant heterogeneity was observed across studies (τ2 = 0.02, I2 = 99.97%, Q = 36 787, p < 0.001), indicating substantial inter-study variation in discontinuation rates. Four of the 11 studies exclusively focused on patients with T2D,11, 13, 15, 16 while three studies excluded individuals with T2D.9, 14, 17 Four studies included a range in the proportion of individuals with T2D, from 31.3% to 89.7%,5-8 and one study did not provide T2D proportions specific to the GLP-1 RA user group.12 A meta-regression exploring the association between the proportion of patients with T2D and the one-year discontinuation rate revealed a significant negative association (Figure 2). The adjusted R2 was 61.6%. For each 1% increase in the proportion of patients with T2D, the one-year discontinuation rate decreased by (β = −0.37 [95% CI: −0.57 to −0.17, p = 0.003]). The y-intercept was 69.4% (95% CI: 54.4% to 84.5%, p < 0.001). One of 12 studies directly reported reasons for discontinuation17 while three provided assumed reasons,7, 14, 15 citing adverse gastrointestinal (GI) events, high out-of-pocket costs, medication shortages and a perceived lack of benefit. Six studies provided prognostic factors such as younger age, lower income and male sex, which were associated with higher discontinuation.5, 6, 8, 11, 13, 15 Three studies did not report on reasons or factors associated with discontinuation.9, 12, 16 We performed a time-restrained systematic review and meta-analysis of observational real-world studies to better understand GLP-1 RA discontinuation rates and factors influencing discontinuation, with a particular focus on T2D within obesity care. Our findings indicate that nearly half of patients who begin a GLP-1 RA will discontinue within a year, though substantial inter-study variability was observed. The proportion of patients with T2D explained a significant proportion of the variation between studies. While the pooled estimate of approximately 49% of patients discontinuing GLP-1 RA medications before 1 year is incongruent with their intended long-term use, it aligns with adherence rates observed for other chronic disease medications,18 and represents an improvement compared to the 60%–65% one-year discontinuation rates reported for earlier GLP-1 RAs used in the treatment of T2D.19 Consistent with this trend, a comparison of GLP-1 RAs for weight management showed substantially lower one-year discontinuation rates with the newer semaglutide (60%) compared to the older liraglutide (87%),12 which may be partially attributed to semaglutide's greater weight loss efficacy. We anticipate that reductions in discontinuation will continue with introduction of lower-cost medications and more convenient injectable formulations that require less frequent dosing. Using meta-regression, we quantified the contribution of the proportion of patients with T2D to GLP-1 RA discontinuation rates. Given recent approval of GLP-1 RAs for chronic weight management, the number of individuals using these medications in the absence of T2D is expected to rise. Our findings indicate that approximately 70% of these individuals will discontinue medications before 1 year. Considering the limited number of studies providing reasons, it is difficult to draw definitive conclusions about primary drivers of discontinuation. However, reported and suspected reasons such as GI events, costs and perceived lack of benefit likely contribute to the higher discontinuation rates seen in those without T2D.20 To support long-term weight loss maintenance in this growing user-group, barriers to continued GLP-1 RA treatment should be addressed through testing strategies such as patient and health insurer education on the chronicity of obesity, the long-term cost-effectiveness of sustained treatment and the importance of lifelong transformation through the integration of long-term medication use and healthy habits.21 While our analysis provides valuable insights, certain limitations should be considered when interpreting these findings. Given the evolving landscape of GLP-1 RAs, and their relatively recent approval for the treatment of excess body mass, our study required a strategic and justifiable focus on more recent studies. However, it is important to recognize that our findings do not offer a comprehensive overview of GLP-1 RA discontinuation. Additional limitations include conducting the search in a single bibliographic database, the absence of a systematic grey literature search and not performing a formal risk of bias assessment. A notable strength of our analysis is that we compiled real-world data from nearly half a million individuals, offering unique and timely insights into discontinuation patterns with new-generation GLP-1 RA medications in the context of weight loss and T2D. We would like to thank Ms. Madelyn Shepherd for assistance with the literature review. Gregory Grosicki and Satya Jonnalagadda are employees of Medifast Inc. Nikhil Dhurandhar, J. Graham Thomas, Holly Lofton and Steven Heymsfield serve on the Medifast Inc. Scientific Advisory Board and received financial support from Medifast during the preparation of this manuscript. The peer review history for this article is available at https://www.webofscience.com/api/gateway/wos/peer-review/10.1111/dom.16151. The contents of this manuscript are based on publicly available data from previously published studies. All data supporting the findings of this review are included within the manuscript and/or its supplementary materials. No additional data are available. Supplementary Table 1. Studies included in meta-analysis and relevant participant characteristics. Supplementary Figure 1. PRISMA flow diagram. We searched the PubMed database using the following search terms: (‘discontinuation’ [Mesh] or ‘persistence’ [Mesh] or ‘adherence’ [Mesh] or ‘discontinuation’ [tiab] or ‘persistence’ [tiab] or ‘adherence’ [tiab]) and (‘glucagon-like peptide 1’ [Mesh] or ‘glp1’ [tiab] or ‘glp-1ra’ [tiab] or ‘glucagon-like-peptide-1’ [tiab]) and (‘diabetes’ [tiab] or ‘diabetes mellitus, type 2’ [Mesh] or ‘obesity’ [Mesh] or ‘overweight’ [tiab]). Filtered records (by publication year, 2021) were complemented with relevant papers from reference lists and a preprint identified during the search process. The screening process excluded articles identified as not being observational real-world studies. The eligibility assessment excluded articles not reporting one-year discontinuation, records not extending to at least 2021 or those reporting only on oral medications. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.