Introduction

The relationship between migraine and pregnancy is highly variable [1, 2], with the majority of women experiencing an improvement in their symptoms during pregnancy (55 to 90% of cases) [3], while others reporting an unchanged, or rarely, a worsened migraine course during pregnancy especially in the first trimester (8% of cases) [4].

Since 2018, two drug classes inhibiting the signalling of calcitonin gene-related peptide (CGRP) have become available as acute or preventive migraine treatments [5, 6]: monoclonal antibodies targeting CGRP or its receptor and the small-molecule CGRP receptor antagonists (gepants), collectively referred to as CGRP antagonists (CGRP-A).

CGRP-A are effective in both episodic and chronic migraine according to randomized controlled trials and several observational studies [7]. However, safety data on their use in pregnancy are limited to anti-CGRP monoclonal antibodies and stem predominantly from single clinical cases [8,9,10] and pharmacovigilance studies on spontaneous safety reports [11]. Conversely, no safety data in humans on the use of gepants in pregnancy are currently available.

Regardless of the two different therapeutic modalities by which CGRP-A act and the different timing of crossing the placental barrier, it seems reasonable to expect similar consequences from their use in human pregnancy when considering the role of CGRP in the development and regulation of the utero-placental blood flow [12].

To gain further information on the safety of CGRP-A when used in pregnancy, we queried VigiBase®, the World Health Organization global database of spontaneous safety reports, and performed disproportionality analyses to detect differences in the reporting frequency between CGRP-A and the migraine-specific acute treatment with triptans [13] in relation to pregnancy.

Methods

Disproportionality analyses were performed on de-duplicated spontaneous safety reports collected in VigiBase® as of 31.05.2023. Drugs of interest, selected as active ingredients, included suspected monoclonal antibodies erenumab (targeting CGRP receptor), galcanezumab, fremanezumab and eptinezumab (targeting CGRP-ligand), and gepants ubrogepant, rimegepant, and atogepant (all targeting CGRP receptor). Events of interest were captured by using the Standardized Medical Dictionary for Regulatory Activities (MedDRA®) Query (SMQ) “pregnancy and neonatal topics” (version 26.0). Safety reports reporting as suspected drug(s) sumatriptan, naratriptan, zolmitriptan, rizatriptan, almotriptan, eletriptan and/or frovatriptan were used for the single comparator group (Supplementary Fig. 1), to control for confounding by indication [14], and because, in the absence to date of migraine-specific preventive drugs proven safe in pregnancy, use of triptans in pregnancy appears safe [15]. Safety reports with additional suspected/interacting drugs beyond those of interest and safety reports lacking specific terms referring to drug exposure in pregnancy (including “maternal exposure before pregnancy”, “foetal exposure during pregnancy”, “maternal exposure during pregnancy”, “maternal exposure during breastfeeding”, “paternal exposure during pregnancy”, “maternal exposure time unspecified”) were excluded from the study cohort.

Reporting Odds Ratio (ROR) was used as disproportionality measurement along with its 95% confidence interval (CI) and computed when a minimum number of 5 safety reports of interest was present to reduce the likelihood of false positives [16]. Threshold for the detection of a signal of disproportionate reporting was set with 95% CI lower limit > 1 [16]. The primary outcome was to detect signals of disproportionate reporting for pregnancy exposures to CGRP-A regardless of the reporting of pregnancy outcomes in addition to drug exposure.

The secondary outcomes were i) to detect signals of disproportionate reporting for pregnancy exposures to CGRP-A reporting any pregnancy outcomes (maternal and/or foetal/neonatal); and ii) to detect signals of disproportionate reporting for pregnancy exposures to CGRP-A reporting foetal/neonatal outcomes.

The following sensitivity analyses were performed to control for confounding: i) temporal restriction, starting from 01.01.2018 (when the first in class erenumab received marketing authorization); and ii) temporal restriction and subgroup disproportionality analyses by therapeutic modality (i.e. monoclonal antibodies versus gepants). Data management and analyses were performed with Statistical Analysis System Software (version 9.4; SA Institute, Cary, NC).

According to the Human Research Act (810.30, of 30 September 2011 - status as of 1 September 2023, Art. 2), from the Federal Assembly of the Swiss Confederation, ethical approval and written informed consents were not required.

Results

Safety reports’ characteristics

As of 31.05.2023, there were 83′587 de-duplicated safety reports with CGRP-A in VigiBase®. Of these, 81′108 (97%) fulfilled the pre-defined inclusion/exclusion criteria, including 467 (0.6%) safety reports reporting exposures to CGRP-A in pregnancy (with or without pregnancy outcomes) (Supplementary Fig. 2). Most of the safety reports related to CGRP-A and pregnancy originated from the United States of America (360/467, 77%), were more frequently reported by patients (225/467, 48%), who were mainly females (431/467, 92%), and more frequently reported exposure to CGRP-A during pregnancy (400/467, 86%) (Table 1).

Table 1 Baseline characteristics of the safety reports included in the study
Full size table

Disproportionality analyses

By comparing safety reports associated with CGRP-A against triptans, no signals of disproportionate reporting were detected either for the overall reporting in relation to pregnancy, for the reporting of pregnancy outcomes (maternal and/or foetal/neonatal), or for the reporting of foetal/neonatal outcomes (Fig. 1 and Supplementary Table 1).

Fig. 1
figure 1

Forest plot representation of the results of disproportionality analyses. Abbreviations: CGRP, calcitonin gene-related peptide; ROR, reporting odds ratio; CI, confidence interval; mAbs, monoclonal antibodies 

Full size image

Sensitivity analyses assessing the overall reporting of CGRP-A exposure in pregnancy, the reporting of any pregnancy outcomes, and the reporting of foetal/neonatal outcomes since 2018 confirmed these results (Fig. 1 and Supplementary Table 1). Lastly, temporal restriction since 2018 and subgroup disproportionality analyses by therapeutic modality showed that safety reports with anti-CGRP monoclonal antibodies and with gepants, when separately assessed against triptans, were not associated with any signals of increased reporting (overall reporting in relation to pregnancy, reporting of any maternal and/or foetal/neonatal pregnancy outcomes, reporting of foetal/neonatal outcomes) (Fig. 1 and Supplementary Table 1).

Discussion

We did not detect differences in the reporting frequency between CGRP-A and triptans in relation to pregnancy. This result extended to pregnancy-related safety reports with CGRP-A as a whole (i.e. reporting drug exposure in pregnancy with or without pregnancy outcomes), and reporting pregnancy outcomes, both in general and specifically referring to the foetus/neonate.

To our knowledge, current evidence on CGRP-A safety in human pregnancy does not yet provide definitive information. A handful of single clinical cases [8,9,10] and a series of 286 safety reports retrieved from VigiBase® by the end of 2021 [11], showed no patterns of maternal, foetal or neonatal toxicity with anti-CGRP monoclonal antibodies, whereas there is still a lack of data on gepants’ safety when used in human pregnancy. Albeit differences between anti-CGRP monoclonal antibodies and gepants in timing of crossing the placental barrier and differences in half-lives [7], we considered CGRP-A as a single group due to the key role played by CGRP in pregnancy, which increases utero-placental blood flow and decreases uterine vascular resistance [12]. Separate disproportionality analyses by therapeutic modality (i.e. monoclonal antibodies versus gepants) confirmed the absence of signals of disproportionate reporting with anti-CGRP monoclonal antibodies in relation to pregnancy, as previously assessed by our group [11]. Also for gepants, no signals of increased reporting in relation to pregnancy were detected. Interestingly, our study also identified five pregnancy-related safety reports that were associated with the concomitant use of an anti-CGRP monoclonal antibody and a gepant, a combination treatment that still remains debated in the general population [17,18,19].

Our study has several limitations. Firstly, disproportionality analysis is a hypothesis-generating method that does not allow definitive conclusions to be drawn on drug safety. Secondly, the lack of exposure data and the unquantified under-reporting prevent from calculating the incidence of any drug toxicity. Lastly, due to the voluntary nature of spontaneous reporting, clinical details of VigiBase® safety reports are not available, therefore it is not possible to know the exact time of drug exposure in pregnancy and the duration of treatment with CGRP-A during pregnancy.

Conclusions

This study showed that, to date, there are no signals of increased spontaneous reporting with CGRP-A compared to triptans in relation to pregnancy in VigiBase®. Disproportionality analyses depend however on the progressively increasing number of safety reports gathered in the spontaneous reporting system. Therefore, future disproportionality analyses need to be performed and complemented with pregnancy pharmacovigilance studies on patient registries and other investigations on large-scale collaborative projects such as the Innovative Medicines Initiative (IMI) ConcePTION project [20].