Registry-based safety studies of migraine drugs in pregnancy: Acknowledging our failures

I would have been happy to commend Mitter et al. (1) for developing algorithms to identify and characterize prevalence of migraine in pregnancy. However, their claim that “registry-based safety studies of migraine drugs in pregnancy face multiple challenges” (1) is hardly an acceptable excuse for regulators’ enduring lack of concern over the need to prescribe treatments with evidence for a positive benefit-to-harm ratio.

First, why are registries for monitoring quality of care deemed such a challenge in our data-driven world? The dearth of pregnancy registries is shameful given that clinical trials routinely and systematically exclude pregnant and lactating women from participation. Even where such registries exist, too few data are collected to enable researchers to ascertain benefit-to-harm ratios or guide dosing. Even when available, they are frequently overlooked, and timely and proportionate actions are not taken (2).

Second, with regard to migraine, the case of topiramate illustrates a devastating level of regulatory inertia. Despite the accumulation of confirmatory evidence of neurodevelopmental harms following prenatal exposure from reliable cohort studies since the earliest warning in 2012 (3), the European Medicines Agency waited for a decade before recommending the following Risk Minimization Measures: (i) no use during pregnancy unless there is no other suitable treatment available; and (ii) implementation of a related pregnancy prevention program. Indeed, the issue is also women of child bearing age as 40% of pregnancies are unintended (4).

Worse, on the other side of the Atlantic Ocean, authorities seem to be in complete denial of the dangers posed to the unborn child from topiramate, even though the drug was ranked 66th for prescriptions in 2021, its popularity having soared from 99th in 2013. Indeed, one leading journal went so far as to publish a study's conclusion that confidence interval of the hazard ratio for autism after prenatal exposure to topiramate ranged from 0.56 to 1.65 (5), at the same time failing to warn about the study's serious limitations. For example, it was a retrospective analysis from databases aimed at monitoring reimbursements – Medicaid and commercial health insurance – and used sophisticated ad hoc analyses that had not been described in a pre-registration. These omissions should be considered in the context of globally accumulating evidence of neurodevelopmental drug harms, being generated by reliable cohort studies (3).

This lack of concern for minimizing exposure of pregnant women, even when harms are documented, is a rule not the exception. It represents an unconscionable failure to protect the most vulnerable of our citizens: the developing child. Consider that the US Food and Drug Administration has only ever issued a dozen Risk Evaluation and Mitigation Strategy programs for teratogenic drugs, of which ten remain active, despite the fact that the Teratology Information Service and Clinical Pharmacology databases identify 141 drugs as having a definite teratogenic risk and 65 a potential risk (6). The issue is a systemic one, from the industry to the prescribers, taking in regulatory agencies, professional organizations and academics. As prescribers, should we not provide women with sufficient data and support to enable them to make informed decisions on treatment choices that could seriously impact their offspring?

Acknowledging failure is surely a mandatory prerequisite for improvement.