Glucagon-like peptide-1 receptor agonists (GLP-1RAs) and related dual agonists have emerged as key pharmacologic options for the treatment of overweight and obesity. Despite overall efficacy, there is pronounced interindividual and interstudy heterogeneity in weight-loss outcomes. The authors conducted a meta-regression analysis to identify factors that influence the variability in weight-reduction efficacy of GLP-1RAs and dual agonists.
The investigators performed systematic literature searches of Medline, Embase, The Cochrane Library and Web of Science through February 1, 2026. The review adhered to PRISMA guidelines. Thirty-two randomized trials meeting eligibility criteria were included, representing 40,408 participants in total. Meta-regression and subgroup analyses were conducted using R software. Risk of bias across included trials was assessed with the RoB-2 tool.
Detailed trial-level characteristics are reported in the full article. The pooled dataset included randomized controlled trials of GLP-1RAs and dual agonists used for weight reduction in adults with overweight or obesity. The publication and indexing details are Clin Obes 2026 Oct;16(5):e70112, PMID 42733128, PMCID PMC13572569, DOI 10.1111/cob.70112.
Across the included trials, the pooled weighted mean percentage weight reduction was 9.36% (95% CI 7.92%–10.80%). However, there was very high between-study heterogeneity (I2 = 99.6%, p < 0.001), indicating substantial variability in treatment effects across trials.
Subgroup analyses reported by the authors identified two notable modifiers of treatment magnitude:
GLP-1/GIP dual agonists (specifically tirzepatide) produced superior weight-lowering effects compared with other agonist classes included in the analysis.
Longer treatment duration (defined in the study as ≥ 72 weeks) was associated with greater weight reduction compared with shorter durations.
Additional subgroup displays in the article compare effects across GLP-1RA agents such as semaglutide and liraglutide, and across different dual-agonist classes; figures support these comparisons.
The meta-regression explored potential mediators of the observed heterogeneity. Key findings included:
Baseline anxiety/depression was positively associated with weight-loss efficacy (β = 2.44, p < 0.001). The model including this factor had an R2 of 81.34%, indicating that it accounted for a large proportion of between-study variance in weight loss.
Certain metabolic comorbidities were negatively correlated with treatment response. Specifically, comorbid type 2 diabetes mellitus (T2DM), obstructive sleep apnea (OSA) and metabolic dysfunction–associated fatty liver disease (MAFLD) were each associated with reduced weight-loss efficacy (all p < 0.05).
Figure 8 in the paper visualizes treatment effects for weight reduction by mean differences in potential mediators and presents meta-regression slopes with confidence bands.
Risk of bias across the 32 included trials was evaluated using RoB-2; a dedicated figure summarizes that assessment. The authors also performed sensitivity analyses, including subgroup sensitivity based on the source and imputation of standard deviation (SD) data (for example, comparing studies using an imputed coefficient of variation for missing SDs versus studies reporting SDs). These sensitivity checks are presented in the figures and supplemental analyses.
The authors conclude that both emotional state and specific metabolic comorbidities are important mediators of interindividual differences in weight-loss response to GLP-1RAs and dual agonists. Clinically, their findings suggest that agent selection (for example, tirzepatide) and treatment duration influence outcomes, and that baseline patient factors such as anxiety/depression and presence of T2DM, OSA or MAFLD may guide expectations and personalization of obesity pharmacotherapy.
These results are intended to inform more individualized obesity management by identifying patient characteristics that modulate pharmacologic response.
The abstract, figures and main text provide pooled estimates, subgroup and meta-regression results, risk-of-bias assessments and sensitivity analyses. Specific trial-level details, full numeric results for every subgroup comparison and additional methodological particulars are presented in the full text. Where the abstract or PubMed entry did not report further granular data, those specifics are available in the full article (PMCID PMC13572569).