Adverse pregnancy and birth outcomes in women exposed to glucagon-like peptide-1 receptor agonists – A nationwide register-based study

Abstract

Prescription and use of glucagon-like peptide-1 receptor agonist (GLP-1RA) medicines for type 2 diabetes (T2D) and weight management have been rising notably in Denmark in recent years, including among women of fertile age. To date, limited information about the safety of GLP-1RA in pregnancy is available. Animal studies have shown reproductive toxicity, and women are therefore advised to stop treatment if they plan to become pregnant or in the case a pregnancy occurs. Despite clinical recommendations, some women may still become pregnant during treatment. To be able to counsel these women, we investigated the association between periconceptional GLP-1RA use and potential risks of adverse pregnancy and birth outcomes. Nationwide Danish health registry data was utilized to investigate the risks of miscarriage, major congenital malformations (MCM) and small for gestational age (SGA) among pregnancies exposed to GLP-1RA during pregnancy versus no exposure to GLP-1RA during pregnancy. Exposure was defined as prescription between 3 months prior pregnancy and an outcome-specific end date. Confounding was adjusted for by propensity score inverse probability of treatment weighting. Prevalence odds ratios (OR) of MCM and SGA were estimated by log-binomial regression and the hazard ratio (HR) of miscarriage by cox proportional hazards modelling. A total of 1 195 415 pregnancies were eligible for the study of which 900 733 (74.3%) ended as live or still births, 170 331 (14.3%) as terminations, and 124 351 (10.4%) as miscarriage. Of all eligible pregnancies, 1087 (0.09%) had been exposed to GLP-1RA. No statistically significant association was found between GLP-1RA exposure and risk of MCM and SGA (adjusted prevalence OR for MCM, 1.36; 95% CI, 0.88-2.16; adjusted prevalence OR for SGA, 0.86, 95% CI, 0.55-1.34. Although the estimated HR of miscarriage was statistically significant (adjusted HR, 1.24; 95% CI, 1.05 -1.47), pre-specified sensitivity analyses did not support this finding, suggesting possible unmeasured confounding of the association. In conclusion, the findings of this study did not suggest an increased risk of adverse pregnancy and birth outcomes related to GLP-1RA exposure during pregnancy. More research is needed before a pregnancy or birth related risk can be reassuringly excluded.