Migraine Antibody Drugs Were Tied to 42 Percent Higher Miscarriage Risk. Propranolol Users Showed No Rise.

What is actually known about anti-CGRP migraine antibodies in pregnancy is thin. The injections have been sold since 2018. CGRP helps build the placenta and its blood supply. The safety record has rested mostly on case reports and drug-safety reporting databases. A paper published October 12 in JAMA Neurology adds a larger dataset. Its result is a higher rate of miscarriage in exposed pregnancies than in unexposed ones.

Leah Flatman, Anick Bérard and colleagues at the Azrieli Research Centre of CHU Sainte-Justine in Montreal pulled 7,579 singleton pregnancies in 7,119 women from MarketScan commercial insurance claims, 2018 through 2021. Each woman was 15 to 45. Each had a migraine diagnosis, or a triptan or ergot prescription, in the six months before her last menstrual period. By 20 weeks, 1,862 of those pregnancies had ended in spontaneous abortion, the clinical term for miscarriage.

What the numbers show

Compared with unexposed pregnancies, exposure to erenumab, fremanezumab, galcanezumab or eptinezumab carried an adjusted hazard ratio of 1.42. The interval is wide: the data fit anything from a 12 percent to an 80 percent higher hazard. A separate bias-adjusted analysis gave 1.33. Propranolol, a beta-blocker also used to prevent migraine, showed no association at 1.02.

A hazard ratio compares rates over time. It is not the share of pregnancies that end in loss, and the abstract gives no absolute difference to apply to one pregnancy. Medscape covered the June 5 presentation of this work at the American Headache Society meeting. In its account the comparison rested on 318 anti-CGRP pregnancies, 201 on propranolol alone, and 7,060 on neither. The finding comes from roughly 300 exposed pregnancies.

Base rates complicate the picture. The American College of Obstetricians and Gynecologists puts early pregnancy loss at about 10 in 100 known pregnancies. This cohort logged 1,862 losses in 7,579 pregnancies, about 25 in 100. The abstract does not explain the gap. ScienceAlert notes that insurance records are not collected with rigorous research in mind.

Timing, and a five-month clock

Medscape reported that no losses were recorded in any group before week 8. Between weeks 8 and 12, losses rose more among the anti-CGRP group than the others, and from week 12 to 20 the groups looked similar. The conference estimate was 1.45 (interval 1.14 to 1.85), slightly different from the published 1.42, so the number moved a little between presentation and paper.

Vincent Martin directs the Headache and Facial Pain Center at the University of Cincinnati and was not part of the study. He told Medscape that injected CGRP therapies "stay in the body for 5 months." That matters, he said, for an unplanned pregnancy, or for a patient who does not realize she should be off the drug for five months before trying to conceive. Martin called the finding noteworthy but early, and said it needs to be replicated.

Why migraine severity is not ruled out

The strongest objection is confounding by indication. ScienceAlert points out that these antibodies are typically prescribed for severe or treatment-resistant migraine, so severe migraine itself could be what raises miscarriage risk. The Medscape account shows the exposed group differed on paper: more depression, more use of triptans, anti-nausea drugs and antiepileptics, and heavier health care use. The authors adjusted for those. The claims data hold no measure of migraine severity, and no record of whether prescribed drugs were taken or pregnancies were planned.

Flatman's response at the conference was that the result held across multiple sensitivity analyses, which suggests severity is unlikely to explain all of it. "Unlikely to fully explain" is her wording, and it is not the same as ruled out. The propranolol arm is a partial check, since those women also took a migraine preventive. But the paper gives no severity score for either group, so the check cannot be graded.

Brinder Vij is division director of headache medicine at the University of Cincinnati College of Medicine. He told Medscape the FDA label leaves it unclear whether the antibodies belong in pregnancy, and that he personally does not recommend them to pregnant patients. Medscape's disclosure notes list fees for Martin from AbbVie, Eli Lilly, Pfizer, Tonix and Teva, and for Vij from AbbVie, Amgen and Biohaven, all drugmakers. Flatman reported none, and no outside funding was noted. This piece draws on the published abstract and secondary coverage. The full JAMA Neurology text would not load.