To assess whether exposure to anti-CGRP monoclonal antibodies or propranolol was associated with spontaneous abortion among pregnancies in patients with pregestational migraine.
Approach:
Design and population: Retrospective cohort study of 7,579 singleton pregnancies among 7,119 patients aged 15–45 years, recorded from November 2018 through December 2021 in a US commercial insurance database.
Exposure and outcome: Claims identified time-varying exposure to anti-CGRP monoclonal antibodies or propranolol and clinically detected spontaneous abortion at or before 20 weeks.
Analysis: Cox models adjusted for maternal age, comorbidities, concomitant medications, and health care utilization; sensitivity and bias analyses assessed exposure duration and potential confounding by migraine severity.
Key Findings:
There were 1,862 clinically detected spontaneous abortions (24.6%) among 7,579 pregnancies.
Anti-CGRP-exposed person-time included 350 pregnancies and 100 spontaneous abortions (28.6%); the adjusted hazard was 42% higher than during unexposed person-time.
The adjusted estimated risk by 20 weeks was 28.7% under a hypothetical all-exposed scenario, compared with 24.5% without exposure.
Propranolol-exposed person-time included 283 pregnancies and 37 spontaneous abortions (12.8%); propranolol was not associated with an increased hazard compared with unexposed person-time.
Adjustment for potential confounding by migraine severity modestly attenuated the anti-CGRP association, which remained elevated.
Interpretation:
The observational findings indicate an association between anti-CGRP monoclonal antibody exposure around conception or early pregnancy and spontaneous abortion, but do not establish causality. The authors called for further research on pregnancy safety and longer-term offspring outcomes.
Limitations:
Claims did not directly measure migraine frequency or severity, leaving potential residual confounding by indication.
The number of anti-CGRP-exposed pregnancies was relatively small, and prescription claims could not confirm medication administration.
Very early unrecognized losses may have been missed, and gestational age was imputed for many losses.
The commercially insured US population may limit generalizability; pregnancy intention and treatment history could also have influenced findings.
Conclusion:
The authors advised that clinicians consider the limited pregnancy-safety evidence when discussing anti-CGRP monoclonal antibody treatment with people who may become pregnant.
Single-nucleus sequencing of menstrual fluid identified epithelial gene signatures and prioritized five candidate biomarkers that warrant validation in larger studies.
Claims-based target trial emulation found no clear association between continued GLP-1 receptor agonist use in early pregnancy and nonlive birth, fetal growth abnormalities, or major congenital malformations.
