Use of semaglutide and risk of non-arteritic anterior ischemic optic neuropathy: A Danish-Norwegian cohort study

Emma Simonsen,Lars Christian Lund,Martin Thomsen Ernst,Vidar Hjellvik,Laszlo Hegedus,Steffen Hamann,Oystein Kalsnes Jorstad,Hanne Lovdal Gulseth,Oystein Karlstad,Anton Pottegard

DOI: https://doi.org/10.1101/2024.12.09.24318574

2024-12-11

Abstract:Importance Recent findings have raised concerns about an association between semaglutide, a glucagon-like peptide-1 receptor agonist (GLP-1RA), and non-arteritic anterior ischemic optic neuropathy (NAION). Objective To further investigate the putative association between semaglutide use and NAION. Design Bi-national active comparator new-user cohort study with propensity score weighing to adjust for confounding, fixed-effect meta-analysis, and a supplementary self-controlled analysis (symmetry analysis). Setting Data were obtained from national health registries in Denmark and Norway, including prescription registries, patient registries, and civil registration systems. Participants New users of semaglutide and sodium-glucose co-transporter 2 inhibitors (SGLT-2is) during 2018-2022 (Norway) and 2018-2024 (Denmark). Exposures Initiation of semaglutide and SGLT-2is for the management of type 2 diabetes was assessed in the main analysis. In the self-controlled design, exposure to semaglutide, regardless of indication, was investigated. Main Outcomes and Measures Incidence rates of NAION among semaglutide and SGLT-2is users, along with hazard ratios (HRs) and incidence rate differences (IRDs) with robust 95% confidence intervals. Estimates from Denmark and Norway were pooled using a fixed-effects model (inverse variance weighting). The supplementary self-controlled study returned sequence symmetry ratios (SRs). Results We identified 44,517 eligible users of semaglutide for the management of type 2 diabetes in Denmark and 16,860 in Norway. In total, we observed 32 NAION events among semaglutide users (24 in Denmark and 8 in Norway). This yielded an unadjusted incidence rate of NAION of 2.19/10,000 person-years among Danish semaglutide initiators, compared to 1.18 among SGLT-2i initiators. The corresponding unadjusted incidence rates in Norway were 2.90 and 0.92. After adjustment, we obtained a pooled HR of 2.81 (95% CI 1.67 to 4.75) and IRD of +1.41 (95% CI +0.53 to +2.29). Estimates were compatible in the individual countries, although with higher and less precise estimates in Norway (HR 7.25; 95% CI 2.34 to 22.4) compared to Denmark (HR 2.17; 95% CI 1.20 to 3.92). Results were consistent across sensitivity and supplementary analysis, with a notably stronger association observed in a post hoc per-protocol analysis (HR 6.35; 95% CI 2.88 to 14.0). In the supplementary self-controlled study, SRs for NAION were 1.14 (95% CI 0.55 to 2.36) in Denmark and 2.67 (95% CI 0.91 to 8.99) in Norway. Conclusions and Relevance Our findings show an association between use of semaglutide for type 2 diabetes and risk of NAION. However, the absolute risk remains low.