Metabolic rebound after GLP-1 receptor agonist discontinuation: a systematic review and meta-analysis

To quantify the extent of metabolic and cardiovascular rebound following the discontinuation of GLP-1 receptor agonist (GLP-1RA) therapy in adults with obesity, type 2 diabetes, or type 1 diabetes, and identify potential modifiers of these outcomes.

We conducted a systematic review and meta-analysis of randomized controlled trials including adults and adolescents treated with GLP-1RAs, followed by a post-discontinuation follow-up of at least 12 weeks. Databases searched included MEDLINE, Embase, Cochrane CENTRAL, Web of Science, and ClinicalTrials.gov, with no language restrictions, and published up to October 17, 2025. We assessed changes in anthropometric, glycemic, cardiovascular, and lipid parameters between the end of treatment and the post-cessation period. Random-effects models were used to calculate pooled mean differences. The risk of bias and certainty of evidence were evaluated using the Cochrane ROB 2.0 and GRADE frameworks. The study was registered with PROSPERO under CRD42025646185.

Eighteen RCTs (3771 participants) were included. Among individuals with obesity, discontinuation of GLP-1RA resulted in significant metabolic rebound, characterized by a body weight gain of 5.63 kg (95% CI: 3.52-7.73, I2 = 99.57%, moderate) and an increase in HbA1c of 0.25% (95% CI: 0.18-0.32, I² = 98.45%, moderate). Waist circumference, BMI (Body Mass Index), SBP (Systolic Blood Pressure), and FPG (fasting plasma glucose) also showed significant deterioration. In the type 2 diabetes setting, weight gain was 2.03 kg (95% CI: 1.63-2.42, I² = 42.28%, moderate), and HbA1c rose by 0.65% (95% CI: 0.22-1.08, I² = 96.83%, moderate), while FPG remained stable (0.90 mmol/L; 95% CI: -0.36 to 2.17, I² = 98.81%, moderate). Subgroup analyses revealed greater weight regain with longer follow-up (>26 weeks: 7.31 kg vs. 2.51 kg) and with semaglutide compared to liraglutide (8.21 kg vs. 4.29 kg). Semaglutide also led to greater increases in waist circumference (3.80 cm vs. 2.69 cm) and SBP (7.09 mmHg vs. 1.56 mmHg). Most outcomes showed no evidence of significant publication bias, with minor asymmetry detected for VLDL (very-low-density lipoprotein) levels in individuals with obesity and for weight change (kg) in patients with type 2 diabetes. The risk of bias assessment using the ROB 2.0 tool rated all included studies as having a low risk.

The magnitude and consistency of these effects underscore the physiological consequences of GLP-1RA discontinuation. As the clinical use of GLP-1RAs continues to grow in the management of obesity and diabetes, it is imperative that treatment guidelines address not only initiation and titration but also discontinuation and long-term maintenance strategies to sustain therapeutic gains.

This study received no funding.