Bayesian Network Meta-Analysis of Weight Loss Efficacy for GLP-1 Receptor Agonists and Tirzepatide
Author(s)
Rishabh D. Pandey, PhD1, Prabhakar Pandey, M. Pharm2.
1Head, Strategy, Growth and Solutions, SEREXIA CONSULTANCY PVT LTD, BENGALURU, India, 2Head, Delivery & Operations, SEREXIA CONSULTANCY PVT LTD, Bengaluru, India.
1Head, Strategy, Growth and Solutions, SEREXIA CONSULTANCY PVT LTD, BENGALURU, India, 2Head, Delivery & Operations, SEREXIA CONSULTANCY PVT LTD, Bengaluru, India.
OBJECTIVES: To estimate comparative effectiveness of glucagon-like peptide-1 (GLP-1) receptor agonists (liraglutide, semaglutide) and the dual GLP-1/GIP receptor agonist tirzepatide on weight loss by conducting a Bayesian Network Meta-Analysis (NMA).
METHODS: A systematic literature review was conducted to identify relevant randomized controlled trials (RCTs) evaluating liraglutide, semaglutide, and tirzepatide for weight management. Mean weight loss from baseline was the primary endpoint. For multi-arm tirzepatide trials, effects from different dosages were pooled within each study using a fixed-effect meta-analysis to derive a single study-level estimate for Tirzepatide, disregarding dose-dependent efficacy. A Bayesian random-effects NMA was conducted using JAGS in R. Non-informative priors were specified for treatment effects and heterogeneity. Three Markov Chain Monte Carlo Simulations (MCMC) chains were run for 100,000 iterations (50,000 burn-in, 10 thinning). Convergence was assessed via trace plots, Gelman-Rubin diagnostics, and effective sample size. Pairwise mean differences (kg) with 95% credible intervals (CrI), league tables, forest plots, treatment ranks, and Surface Under the Cumulative Ranking Curve (SUCRA) values were generated.
RESULTS: The NMA included 21 comparisons from 21 studies. Model diagnostics indicated excellent convergence and mixing. The estimated between-study heterogeneity was 0.29 (95% CrI: 0.03, 1.25) kg. Tirzepatide (pooled) demonstrated the greatest mean weight loss from baseline relative to placebo: -22.10 (95% CrI: -23.01, -21.19) kg, followed by Semaglutide: -12.35 (95% CrI: -13.04, -11.66) kg, and Liraglutide: -5.35 (95% CrI: -5.95, -4.77) kg. Pairwise comparisons consistently favored Tirzepatide (pooled) over all other active comparators. SUCRA values confirmed Tirzepatide (pooled) as the highest-ranked treatment for weight loss.
CONCLUSIONS: Based on this Bayesian NMA, Tirzepatide appears to be the most effective pharmacotherapy for mean weight loss compared to liraglutide, semaglutide, dulaglutide, insulin degludec, and placebo. However, the methodology disregarded dose-dependent efficacy and not adjusted for patient-level variabilities, which should be explored further.
METHODS: A systematic literature review was conducted to identify relevant randomized controlled trials (RCTs) evaluating liraglutide, semaglutide, and tirzepatide for weight management. Mean weight loss from baseline was the primary endpoint. For multi-arm tirzepatide trials, effects from different dosages were pooled within each study using a fixed-effect meta-analysis to derive a single study-level estimate for Tirzepatide, disregarding dose-dependent efficacy. A Bayesian random-effects NMA was conducted using JAGS in R. Non-informative priors were specified for treatment effects and heterogeneity. Three Markov Chain Monte Carlo Simulations (MCMC) chains were run for 100,000 iterations (50,000 burn-in, 10 thinning). Convergence was assessed via trace plots, Gelman-Rubin diagnostics, and effective sample size. Pairwise mean differences (kg) with 95% credible intervals (CrI), league tables, forest plots, treatment ranks, and Surface Under the Cumulative Ranking Curve (SUCRA) values were generated.
RESULTS: The NMA included 21 comparisons from 21 studies. Model diagnostics indicated excellent convergence and mixing. The estimated between-study heterogeneity was 0.29 (95% CrI: 0.03, 1.25) kg. Tirzepatide (pooled) demonstrated the greatest mean weight loss from baseline relative to placebo: -22.10 (95% CrI: -23.01, -21.19) kg, followed by Semaglutide: -12.35 (95% CrI: -13.04, -11.66) kg, and Liraglutide: -5.35 (95% CrI: -5.95, -4.77) kg. Pairwise comparisons consistently favored Tirzepatide (pooled) over all other active comparators. SUCRA values confirmed Tirzepatide (pooled) as the highest-ranked treatment for weight loss.
CONCLUSIONS: Based on this Bayesian NMA, Tirzepatide appears to be the most effective pharmacotherapy for mean weight loss compared to liraglutide, semaglutide, dulaglutide, insulin degludec, and placebo. However, the methodology disregarded dose-dependent efficacy and not adjusted for patient-level variabilities, which should be explored further.
Conference/Value in Health Info
2025-11, ISPOR Europe 2025, Glasgow, Scotland
Value in Health, Volume 28, Issue S2
Code
CO27
Topic
Clinical Outcomes
Topic Subcategory
Comparative Effectiveness or Efficacy
Disease
Diabetes/Endocrine/Metabolic Disorders (including obesity)