Comparative Exposure-Adjusted Adverse Reaction Reporting Rates for GLP-1-Based Therapies: An Analysis of UK MHRA Yellow Card Data and NHS Prescribing Data for England

Liraglutide had an ADR reporting rate of 1,326.64 per million items, significantly lower than newer therapies like tirzepatide, which reported 27,511.52 per million items.

Summary and findings

This study estimated exposure-adjusted adverse drug reaction (ADR) reporting rates for six GLP-1-based therapies, including liraglutide, using UK MHRA Yellow Card data and NHS prescribing data. The analysis covered 140,462 ADRs against 17,005,924 prescribed items from February 2021 to January 2026. Liraglutide had an ADR reporting rate of 1,326.64 per million items.

Abstract

<h4>Background: </h4> /Objectives: Glucagon-like peptide-1 (GLP-1) receptor agonists and the dual glucose-dependent insulinotropic polypeptide/GLP-1 receptor agonist tirzepatide are increasingly prescribed for type 2 diabetes and obesity, but post-marketing comparisons usually rest on unadjusted report counts, which reflect prescribing volume as much as risk. We estimated exposure-adjusted adverse drug reaction (ADR) reporting rates for six GLP-1-based therapies. <h4>Methods:</h4> This retrospective observational pharmacovigilance study covered February 2021 to January 2026. Suspected ADRs reported UK-wide to the Medicines and Healthcare products Regulatory Agency (MHRA) Yellow Card Scheme were divided by English National Health Service (NHS) general practice (GP) and hospital prescribing volumes, combined at medicine and time-period level; hospital quantities were converted to items using GP quantity-per-item ratios. Rates per million items were pooled in a generalised linear mixed model; temporal trends used Poisson regression with a log-exposure offset. Sensitivity analyses used GP-only and hospital-only denominators and re-fitted the temporal models on reports. <h4>Results:</h4> We analysed 140,462 UK Yellow Card ADRs against 17,005,924 English NHS prescribed items. The denominator omits private prescribing, which supplies most UK weight-management use, so these rates are differentially inflated, most for tirzepatide and semaglutide. Rates per million items varied more than 200-fold: tirzepatide 27,511.52 (95% confidence interval [CI] 27,339.39–27,684.73), semaglutide 5,905.83 (5,847.98–5,964.25), liraglutide 1,326.64 (1,268.16–1,387.82), exenatide 443.36, dulaglutide 256.42 and lixisenatide 129.68; heterogeneity was extreme (I² = 99.995%) and the ordering was unchanged under both sensitivity denominators. Rates rose for semaglutide (incidence rate ratio [IRR] 1.675, 1.657–1.693) and tirzepatide (IRR 1.139, 1.118–1.160) and fell for dulaglutide (IRR 0.810, 0.774–0.848); report-based models reproduced these trends. <h4>Conclusions:</h4> Exposure-adjusted reporting rates differ markedly across GLP-1-based therapies, the newest agents reporting most. These are reporting rates under an incomplete NHS denominator, not incidence estimates; differential exposure capture and reporting behaviour prevent interpretation as comparative medicine risk.

Limitations

Not reported in abstract.

Elsewhere in the Liraglutide corpus