Introduction

Migraine, a chronic neurovascular disorder, is a disabling condition characterized by pulsating headache often accompanied by nausea, vomiting, and heightened sensitivity to light or sound. It affects approximately 14.4% of the global population1. The International Headache Society defines chronic migraine as a headache lasting 4 to 72 h, occurring on 15 or more days per month, with at least 8 of those days involving migraines, over a span of more than 3 months2. Affecting about 2% of the population, chronic migraine severely impairs quality of life and is linked to increased anxiety, depression, disability, reduced productivity, and greater absenteeism from daily activities3,4,5.

In recent years, there has been a notable shift in migraine treatment with the introduction of a novel class of drugs targeting the calcitonin gene-related peptide (CGRP) pathway6,7. CGRP is the primary neuropeptide released by the trigeminal nerve, and its signaling is a key mechanism underlying the pathogenesis of migraine attacks, as CGRP is a potent vasodilator8. CGRP transmits nociceptive signals to the brain and spinal cord, promoting neurogenic inflammation and cranial blood vessel dilation, leading to migraines and associated symptoms9. Consequently, the inhibition of CGRP is believed to reduce the likelihood of migraines and effectively alleviate the intensity of acute migraine attacks10.

The United States Food and Drug Administration (US FDA) has recently approved several drugs that target CGRP or its receptor for the acute treatment or prevention of migraines in adults. Fremanezumab and galcanezumab bind to the CGRP ligand, inhibiting its binding to the CGRP receptor, while erenumab binds to the CGRP receptor and antagonizes its function11,12. Orally or sublingually administered small molecule CGRP receptor antagonists include ubrogepant, rimegepant, and atogepant13,14. Atogepant is indicated for episodic migraine prophylaxis, ubrogepant is taken as needed for acute migraine treatment, and rimegepant has been approved for both active treatment and prevention of migraines15.

Additional therapies recommended for migraine prevention include onabotulinumtoxinA, beta-blockers such as propranolol and timolol, and anticonvulsants such as topiramate and valproic acid. Serotonin receptor agonists, also known as triptans, and nonsteroidal anti-inflammatory drugs (NSAIDs) such as celecoxib have been approved for acute or episodic migraine treatment16.

While recent advancements in migraine treatment, such as the development of CGRP inhibitors, have been promising, concerns have arisen regarding potential side effects. A deficiency of CGRP is believed to play a role in Raynaud’s phenomenon17,18. Raynaud’s phenomenon, characterized by abnormal vasoconstriction triggered by cold temperatures or emotional stress, is frequently associated with migraines19,20. Recent reports have indicated that Raynaud’s phenomenon may be induced or exacerbated by CGRP-targeting drugs21,22,23.

The FDA Adverse Event Reporting System (FAERS) is a centralized, computerized information repository used by the FDA and pharmacovigilance experts to conduct postmarketing surveillance for drug safety and detect signals of adverse events (AEs). It comprises reports submitted by manufacturers, as mandated by regulations, as well as voluntary reports directly from consumers and healthcare professionals24.

However, the prescribing information for migraine treatments or prophylaxis does not reference Raynaud’s phenomenon as a potential treatment-related adverse reaction. Therefore, this study aims to assess the association between the use of CGRP inhibitors and the occurrence of Raynaud’s phenomenon by conducting a retrospective disproportionality analysis using the FAERS public database. Additionally, we will compare the incidence of Raynaud’s phenomenon with other drugs used for migraine treatment to further elucidate this potential risk.

Results

Clinical characteristics

A search of the FAERS database identified 186 cases of Raynaud’s phenomenon associated with migraine treatments and prophylaxis. Among these, 89 cases were linked to CGRP inhibitors (Table 1). The breakdown is as follows: 13 cases with fremanezumab (0.42% of AEs), 27 with galcanezumab (0.40% of AEs), 38 with erenumab (0.16% of AEs), 5 with rimegepant (0.18% of AEs), 3 with ubrogepant (0.44% of AEs), and 3 with atogepant (0.36% of AEs). Most patients were female and aged between 18 and 64 years.

Table 1 Clinical characteristics of adverse events associated with Raynaud’s phenomenon.
Full size table

Primary analysis

The CGRP inhibitor class showed significant SDR for Raynaud’s phenomenon compared to all other drugs in the FAERS database (Fig. 1; Table 2; ROR 19.12; 95% CI 15.44–23.69). Among individual CGRP inhibitors, fremanezumab (ROR 32.93; 95% CI 19.06–56.91), galcanezumab (ROR 31.38; 95% CI 21.43–45.94), erenumab (ROR 12.58; 95% CI 9.12–17.36), rimegepant (ROR 12.87; 95% CI 5.34–31.01), ubrogepant (ROR 33.03; 95% CI 10.61-102.81), and atogepant (ROR 23.45; 95% CI 7.54–73.01) all exhibited significant SDRs for Raynaud’s phenomenon relative to all other drugs (Fig. 1; Table 2). Significant SDRs were also observed for triptans (ROR 13.38; 95% CI 9.28–19.29), propranolol hydrochloride (ROR 11.58; 95% CI 8.08–16.59), and celecoxib (ROR 2.29; 95% CI 1.38–3.81). No significant SDRs were found for onabotulinumtoxinA or anticonvulsants.

Fig. 1
figure 1

Reporting Odds Ratios (ROR) with 95% Confidence Intervals (CI) for Raynaud’s Phenomenon Associated with Migraine Therapies: Primary Analysis. (a) CGRP Inhibitors vs. FAERS; (b) Triptans vs. FAERS; (c) Other Drugs vs. FAERS. Blue line indicates positive disproportionate reporting (lower limit of 95% CI > 1.0). Red line marks ROR estimates for all CGRP inhibitors (19.12) and all triptans (13.38).

Full size image
Table 2 Signals of disproportionate reporting for Raynaud’s phenomenon associated with therapies approved for migraine treatment or prophylaxis.
Full size table

Intraclass analysis

No significant SDRs were observed in intraclass comparisons of CGRP antagonists, CGRP receptor antagonists, and triptans, except for galcanezumab (ROR 2.01; 95% CI 1.28–3.17) within the CGRP inhibitor class (Table 3). Additionally, no notable differences were found when comparing CGRP biologics with CGRP inhibitors (ROR 0.91; 95% CI 0.48–1.71). However, a significant SDR for Raynaud’s phenomenon was identified when comparing CGRP antagonists (fremanezumab and galcanezumab) to CGRP receptor antagonists (ROR 2.34; 95% CI 1.54–3.55) (Fig. 2; Table 3).

Table 3 Signals of disproportionate reporting for Raynaud’s Phenomenon:(1) individual CGRP antagonists vs. other CGRP antagonists (2) individual CGRP receptor antagonists vs. other CGRP receptor Antagonists (3) individual CGRP inhibitors vs. other CGRP Inhibitors (4) individual triptans vs. other Triptans (5) CGRP antagonists vs. CGRP receptor Antagonists (6) CGRP Biologics vs. CGRP inhibitors.
Full size table
Fig. 2
figure 2

Intraclass Analysis of Reporting Odds Ratios (ROR) with 95% Confidence Intervals (CI) for Raynaud’s Phenomenon. (a) Individual CGRP Antagonists vs. CGRP Antagonist Class; (b) Individual CGRP Receptor Antagonists vs. CGRP Antagonist Class; (c) Individual CGRP Inhibitors vs. CGRP Inhibitor Class; (d) Individual Triptans vs. Triptan Class; (e) CGRP Antagonists vs. CGRP Receptor Antagonists and CGRP Biologics vs. CGRP Inhibitors. Blue line indicates positive disproportionate reporting (lower limit of 95% CI > 1.0).

Full size image

Interclass analysis

The significant SDR for Raynaud’s phenomenon persisted when CGRP inhibitors were compared with onabotulinumtoxinA (ROR 13.16; 95% CI 4.83–35.85), beta-blockers (ROR 2.20; 95% CI 1.46–3.31), anticonvulsants (ROR 12.57; 95% CI 7.58–20.87), and celecoxib (ROR 7.95; 95% CI 4.60-13.73) (Fig. 3; Table 4). However, no significant difference was observed when CGRP inhibitors were compared with triptans (ROR 0.98; 95% CI 0.64–1.49). These findings suggest a robust association between CGRP inhibitors and Raynaud’s phenomenon, particularly when compared to other migraine therapies.

Fig. 3
figure 3

Interclass analysis of reporting odds ratios (ROR) with 95% confidence intervals (CI) for Raynaud’s phenomenon: CGRP inhibitors vs. comparator groups. This figure compares the SDRs for Raynaud’s phenomenon between the CGRP inhibitor class and various reference groups. Blue line indicates positive disproportionate reporting (lower limit of 95% CI > 1.0).

Full size image
Table 4 Signals of disproportionate reporting for Raynaud’s phenomenon: CGRP inhibitors compared to various comparator groups.
Full size table

Discussion

Among migraine treatments and prophylaxis, significant SDR for Raynaud’s phenomenon were associated with CGRP inhibitors, including fremanezumab, galcanezumab, erenumab, rimegepant, ubrogepant, and atogepant. Triptans, such as sumatriptan, zolmitriptan, rizatriptan, and eletriptan, along with propranolol hydrochloride and celecoxib, also exhibited significant SDRs. The ROR values for CGRP inhibitors ranged from 12.58 to 32.93, while for triptans they ranged from 10.57 to 21.39. Propranolol hydrochloride had an ROR of 11.58, and celecoxib had an ROR of 2.29, indicating that CGRP inhibitors had a more significant association with Raynaud’s phenomenon than other drugs. Despite these significant associations, Raynaud’s phenomenon is not mentioned in the prescribing information for CGRP inhibitors, whereas it is included in the prescribing information for all triptans. This suggests that information regarding the potential risk of Raynaud’s phenomenon should be included in CGRP inhibitor labels for patient safety, as is already done for triptans.

After identifying a potential safety signal through disproportionality analysis, it is crucial to validate this signal using additional methods, such as reviewing individual case reports, epidemiological studies, and comparing known drug safety profiles. Signal validation helps determine whether the observed association is likely causal or due to confounding factors, biases, or random variation.

Our study identified significant SDRs for Raynaud’s phenomenon associated with CGRP inhibitors, consistent with other research utilizing spontaneous AE reporting databases. For example, a WHO pharmacovigilance study identified a significant SDR for Raynaud’s phenomenon in patients using CGRP inhibitors, supporting the hypothesis that these drugs may contribute to its onset​25. In Gérard et al.‘s study, the overall IC for CGRP-targeting drugs associated with Raynaud’s phenomenon was 3.3 (95% CI: 3.0-3.5)​. This is slightly lower than the IC values (4.04 (IC025: 3.69) observed in our analysis for all CGRP inhibitors. They found the highest number of Raynaud’s phenomenon reports associated with erenumab, with an IC of 3.2 (95% CI: 2.8–3.5). In comparison, our analysis showed an IC of 3.42 (IC025: 2.88) for erenumab. Although slightly lower, our results corroborate the substantial association of erenumab with Raynaud’s phenomenon observed in the WHO database.

When comparing CGRP inhibitors with triptans, Gérard et al. reported an IC of 0.4 (95% CI: 0.1–0.6), suggesting a lower disproportionality signal for Raynaud’s phenomenon with CGRP inhibitors relative to triptans25. In contrast, our analysis revealed a ROR of 0.98 (95% CI: 0.64–1.49) for CGRP inhibitors versus triptans, showing no significant difference. This suggests that while both studies identified associations, the relative risk between drug classes may vary depending on the population and the database analyzed. Gérard et al. predominantly found Raynaud’s phenomenon associated with monoclonal antibodies such as erenumab, galcanezumab, and fremanezumab. In our study, the signal was also evident but varied slightly in magnitude, and we observed additional signals for CGRP receptor antagonists like rimegepant and ubrogepant, which were less prominent in the Gérard et al. findings.

Similarly, a study based on the FAERS database found comparable results, reinforcing the association observed in our analysis​26. Sun et al. found galcanezumab (ROR = 12.31; 95% CI: 8.81–17.21), and fremanezumab (ROR = 12.12; 95% CI: 7.16–20.51) to have the strongest signals, which is consistent with our findings. Both studies indicate strong signals for these drugs, reinforcing the association between CGRP inhibitors and Raynaud’s phenomenon. Our study found significant RORs for beta-blockers (ROR = 2.20; 95% CI: 1.46–3.31) and anticonvulsants (ROR = 12.57; 95% CI: 7.58–20.87) when compared to CGRP inhibitors (Table 4). In contrast, Sun et al. primarily focused on CGRP inhibitors without detailed comparisons to these classes. However, our study’s findings suggest that CGRP inhibitors might carry a higher risk for Raynaud’s phenomenon compared to traditional migraine treatments, aligning with the general trend observed in Sun et al.

In addition to these findings, Q. Liang et al. further highlights the association between CGRP inhibitors and Raynaud’s phenomenon, particularly within the gepant class27. The study identified rimegepant as having the strongest signal among vascular disorders, followed by atogepant. In their analysis, rimegepant demonstrated an ROR of 6.51 (ROR025 = 2.92), compared to our analysis, where rimegepant showed a higher ROR of 12.87 (ROR025 = 5.34). Similarly, for atogepant, Q. Liang et al. reported an ROR of 5.98 (ROR025 = 1.93), while our study found a higher ROR of 23.45 (ROR025 = 7.54). Despite these differences in magnitude, both studies consistently highlight the strong association between these gepant-class drugs and Raynaud’s phenomenon, reinforcing a potential class effect. Q. Liang et al. also emphasized that Raynaud’s phenomenon, while rare, remains an important AE to monitor in clinical practice.

Building on these findings, Breen et al. conducted a retrospective cohort study focusing on the microvascular complications of CGRP antagonists in patients with pre-existing Raynaud’s phenomenon, noting that 5.3% of their cohort experienced complications. Their study highlighted that while complications such as worsening Raynaud’s phenomenon and, in severe cases, digital necrosis requiring amputation were rare, they warrant caution when prescribing CGRP inhibitors to patients with underlying Raynaud’s phenomenon22. This reinforces our finding that CGRP inhibitors can exacerbate vascular complications in susceptible individuals.

Moreover, a case report by Bedrin et al. highlighted new onset Raynaud’s phenomenon associated with gepants, specifically rimegepant and ubrogepant. Both cases in their study showed symptoms of Raynaud’s phenomenon after treatment with these small-molecule CGRP antagonists, which further supports the signals we observed for these drugs​28. These findings extend the association of Raynaud’s phenomenon beyond monoclonal antibodies, indicating a broader class effect among CGRP inhibitors. Our study also noted similar signals for rimegepant and ubrogepant, though these were less prominent compared to monoclonal antibodies.

While previous studies identified significant SDRs for Raynaud’s phenomenon associated with CGRP inhibitors, differences in the magnitude of these signals and their comparisons with other migraine therapies could be attributed to variations in the databases used, the populations studied, and the analytical methods employed. However, the overall consistency in the findings across different studies strengthens the evidence linking CGRP inhibitors to Raynaud’s phenomenon, highlighting the importance of continued monitoring and further research into the vascular safety of these medications. Future studies should focus on elucidating the biological mechanisms underlying the association between CGRP inhibition and Raynaud’s phenomenon to better inform clinical practice.

It is important to note, however, that studies explicitly reporting no association or contradictory results for Raynaud’s phenomenon and CGRP inhibitors are currently lacking in the literature. While this may reflect a genuine lack of evidence against such an association, it underscores the need for further research to validate these findings and explore possible inconsistencies across different populations and datasets. Incorporating such studies in future analyses will help ensure a more balanced and comprehensive interpretation of the observed associations.

The exact biological mechanisms through which CGRP inhibitors might induce or exacerbate Raynaud’s phenomenon are not entirely understood, but several plausible hypotheses have been proposed21,22,23,28. CGRP is a potent vasodilator, and its inhibition could lead to unopposed vasoconstriction, particularly in peripheral blood vessels. This mechanism is consistent with the vasoconstrictive nature of Raynaud’s phenomenon, where decreased blood flow to extremities is a hallmark feature. A previous study suggests that the reduced CGRP levels in patients receiving CGRP monoclonal antibodies might decrease vasodilation and contribute to the development or worsening of Raynaud’s phenomenon21.​ Additionally, CGRP is involved in the regulation of vascular tone and neurogenic inflammation, which could further explain its role in pathogenesis of Raynaud’s phenomenon when antagonized​23. CGRP’s role in vasodilation and sensory nerve transmission is well-documented in migraine studies, where it has been shown to mediate arterial vasodilation and activate sensory nerve signaling29,30. Elevated CGRP levels during migraine attacks correlate with headache intensity, and experimental evidence demonstrates that intravenous infusion of CGRP can induce migraine-like headaches31. These findings highlight CGRP’s critical function in maintaining vascular homeostasis. Its inhibition, particularly in peripheral blood vessels, could disrupt this balance, leading to excessive vasoconstriction29. Such mechanisms provide a plausible biological basis for how CGRP inhibitors might exacerbate or induce Raynaud’s phenomenon, especially in patients with predisposing vascular conditions. Supporting this concept, a study reported a significant reduction in CGRP-immunoreactive fibers in the epidermis and subepidermis of the skin in patients with Raynaud’s phenomenon compared to healthy individuals32. This finding suggests that CGRP plays a critical role in maintaining normal vascular function, and its inhibition may contribute to the development or worsening of Raynaud’s phenomenon. Consequently, while anti-CGRP therapies are effective in reducing CGRP release and alleviating migraine symptoms, they may simultaneously increase the risk of inducing or exacerbating Raynaud’s phenomenon in susceptible individuals.

Interestingly, in patients with conditions such as hand-arm vibration injury (HAVS) and Raynaud’s phenomenon, serum levels of biomarkers, including CGRP, were found to be altered33. Studies, including Tekavec et al., demonstrated that serum CGRP levels were elevated in both vascular and neurosensory components of HAVS, suggesting that CGRP plays a vital role in vascular and neural function33,34. In particular, the loss of CGRP-containing nerve fibers in patients with Raynaud’s phenomenon highlights its essential role in maintaining vasodilation and vascular homeostasis17,33. Consequently, CGRP inhibition, such as through monoclonal antibodies, may disrupt this balance, leading to unopposed vasoconstriction. This disruption can exacerbate peripheral vasospasm in conditions like Raynaud’s, contributing to hallmark symptoms such as reduced blood flow, pain, and cold sensitivity in extremities.

Patients with a history of vascular diseases, particularly those with peripheral artery disease or ischemic conditions, may be more susceptible to Raynaud’s phenomenon when treated with CGRP inhibitors35. As Raynaud’s phenomenon involves vasospasm and restricted blood flow to extremities, which can lead to ischemic pain or, in severe cases, tissue damage or necrosis, clinicians should remain vigilant for signs and symptoms such as pain, numbness, or color changes in extremities, particularly after cold exposure36,37. It is recommended that clinicians regularly monitor patients with pre-existing vascular conditions for symptoms of Raynaud’s phenomenon and educate them on recognizing early warning signs. In cases where symptoms arise, immediate medical intervention is necessary to prevent progression to severe ischemic complications. This may involve pharmacological treatment, such as calcium channel blockers or topical vasodilators, as well as lifestyle adjustments to minimize cold exposure and stress38,39,40. Early detection and management are crucial in mitigating the risks associated with CGRP inhibitor use in vulnerable populations.

Despite the signals identified in this study, there are inherent limitations in analyzing spontaneous AE reports. Postmarketing AEs are prone to reporting biases, often characterized by incompleteness, duplication, and significant underreporting. These factors can lead to both false positives and negatives in disproportionality analysis. Moreover, disproportionality analysis may not fully account for potential confounding factors, such as concomitant medications, underlying diseases, and patient characteristics that could influence the occurrence of Raynaud’s phenomenon. To address these limitations, we conducted intra- and inter-class analyses to provide additional layers of validation for the observed associations. While disproportionality measures such as ROR and IC are inherently limited by their inability to adjust for confounding factors (e.g., age, sex, underlying conditions), these supplementary analyses enhance the robustness of our findings41. Specifically, intra-class analysis revealed consistent signals for Raynaud’s phenomenon across individual CGRP inhibitors, including both monoclonal antibodies and small-molecule CGRP receptor antagonists, indicating a broader class effect rather than an association driven by a single agent. Inter-class analysis further contextualized these findings by comparing CGRP inhibitors with other migraine therapies, such as triptans, beta-blockers, and NSAIDs. This approach demonstrated that the signal for Raynaud’s phenomenon with CGRP inhibitors was comparable to or slightly lower than that observed with triptans, a class with well-documented vasoconstrictive effects. By mitigating biases from baseline differences in reporting patterns, these analyses provided a clearer understanding of the associations. Considering diverse and appropriately justified comparators can strengthen the reliability of SDRs, help account for potential confounding factors, and improve their validity and practical utility in clinical settings41,42. Additionally, deduplication procedures and sensitivity analyses were employed to address potential reporting biases, further strengthening the reliability of the results, even in the absence of direct adjustment for confounders. While these efforts improved the robustness of our findings, certain limitations inherent to spontaneous AE reporting systems remain unavoidable. For example, patients with vascular disorders or those on medications with vasoconstrictive properties may have a higher risk of developing Raynaud’s phenomenon, independently of CGRP inhibitor use35,40. Without adjusting for these factors, the drug-event association could be overestimated, complicating the interpretation of results. The submission of a report does not automatically imply a causal relationship between an exposure and an outcome. Utilizing AE data alone, it is not possible to determine the incidence of a specific reaction in a population. Additionally, specific details about Raynaud’s phenomenon, such as whether cases were transient, progressive, or permanent, and information about the size, location, and number of affected areas, were not available. Details regarding the type of medical intervention required to address these adverse reactions were also not reported.

While ROR and IC are widely used and effective measures for signal detection, they are not without limitations, particularly when applied to newer drugs like CGRP inhibitors. These methods are influenced by background reporting rates, making it difficult to distinguish between true signals and events that occur frequently due to their commonality in the population. For AEs with high background incidences, such as cardiovascular events or malignancies, ROR and IC may yield inflated signals, especially if influenced by notoriety bias41,43. Notoriety bias refers to increased media attention or regulatory scrutiny, which can lead to heightened awareness and spur higher reporting rates for newer drugs. This is particularly relevant for CGRP inhibitors, given their recent approval and the potential for artificial inflation in reported AEs due to heightened awareness44. These factors can distort the perceived association between the drug and the AE, making it difficult to establish a true causal relationship44,45. Therefore, spontaneous reporting systems like FAERS may have limitations in detecting signals for common events compared to more robust data sources, such as electronic health records or claims databases, which can reduce reporting biases and improve signal detection46,47. Temporal disproportionality analyses, by evaluating trends in reporting before and after events like regulatory warnings, can also help distinguish genuine increases in AEs from those artificially influenced by external factors, such as media attention or regulatory actions41,48.

To mitigate these issues, we applied deduplication procedures and conducted sensitivity analyses where necessary. However, the potential for bias remains, and the results should be interpreted with caution. It is important to note that disproportionality analysis is primarily a hypothesis-generating tool. The identification of a signal through this method suggests a potential association between a drug and an AE but does not confirm causality. Given the limitations of analyzing spontaneous AE reports, further investigation is necessary to confirm these findings. Subsequent research should involve robust epidemiological studies to establish causality and better understand the risk factors for Raynaud’s phenomenon in patients treated with CGRP inhibitors. In conclusion, this study identified significant SDR for Raynaud’s phenomenon associated with CGRP inhibitors compared to all other drugs and certain migraine therapies in the FAERS database. Despite these significant associations, Raynaud’s phenomenon is not currently mentioned in the prescribing information for CGRP inhibitors. This gap in drug labeling highlights the importance of increasing awareness among healthcare providers regarding the potential association of Raynaud’s phenomenon in patients using CGRP inhibitors, particularly those who may be predisposed to vascular conditions.

Given the potential implications for patient safety, continuous pharmacovigilance and further research are crucial. Future studies should explore the underlying biological mechanisms that might explain the association between CGRP inhibition and Raynaud’s phenomenon and should also involve robust epidemiological investigations to confirm these findings and assess the incidence of this AE in broader patient populations. Finally, as this study is based on an analysis of spontaneous reports and is exploratory in nature, the findings related to CGRP inhibitor-associated Raynaud’s phenomenon need to be validated in large-scale prospective cohort studies. By advancing our understanding of this association, we can better inform clinical practice and potentially update drug safety information to protect at-risk patients.

Methods

Data source and processing

Data for this study were sourced from FAERS, an extensive database that compiles spontaneous reports submitted by healthcare providers, pharmaceutical companies, and consumers. These reports include details on AEs, medication errors, and product quality complaints. We analyzed AEs reported across the entire FAERS dataset, with a particular focus on a subset of reports related to CGRP inhibitors and other therapies approved for migraine treatment or prophylaxis. The subset included triptans, onabotulinumtoxinA, serotonin receptor agonists, beta-blockers, anticonvulsants, and NSAIDs. The comparator groups were selected based on their established roles in migraine management to ensure clinical relevance and reduce indication bias (influence of the underlying condition). This allowed for a balanced comparison within a population with shared underlying conditions, providing a broader context to evaluate whether the observed associations were specific to CGRP inhibitors or extended to other migraine therapies. This study analyzed FAERS reports from the FDA approval year of each drug of interest through August 2023, ensuring a comprehensive dataset for our analysis (Supplementary Table 1 for FDA approval years and corresponding details). The Medical Dictionary for Regulatory Activities (MedDRA) preferred term (PT) “Raynaud’s phenomenon” was used to identify cases.

After data extraction, we performed deduplication to avoid counting multiple versions of the same report before conducting statistical analysis. We examined potential duplicate reports with different case IDs based on suspected product active ingredient, reason for use, reactions, sex, event date, age, and reporting country; reports matching across these seven fields were considered duplicates and were subsequently deduplicated. We used Excel’s advanced filtering and conditional formatting tools to identify potential duplicates. Reports that matched across all seven fields were flagged as duplicates and subsequently removed from the dataset to ensure accuracy in the analysis. These processing steps ensured the accuracy and reliability of the data used for subsequent disproportionality analysis.

Disproportionality analysis

Disproportionality analysis identifies adverse drug reactions (ADRs) reported more frequently than expected in pharmacovigilance. In this study, disproportionality analysis was conducted using two measures, including Reporting Odds Ratios (ROR) and Information Components (IC). These methods are widely used for detecting safety signals in pharmacovigilance databases by comparing the observed number of reports for a specific drug-event combination to the expected number under the assumption of no association41.

The ROR, a frequentist method, calculates the ratio of the odds of reporting a specific AE with a particular drug compared to all other drugs in the database (Supplementary Tables 2 and Supplementary Table 3)49. A signal is considered significant when the lower limit of the 95% confidence interval (CI) of ROR is greater than 1.0 (ROR025 > 1.0)41,50. IC is derived from a Bayesian confidence propagation neural network that compares the observed and expected frequency of a drug-AE pair51. A positive IC value indicates that the event is reported more frequently than expected. A signal is considered significant when the lower limit of the 95% CI of IC is greater than 0 (IC025 > 0)41,50. Together, these methods offer complementary insights into the likelihood of detecting a potential safety signal.

A signal of disproportionate reporting (SDR) refers to a drug-event combination that appears more frequently than expected based on statistical thresholds in disproportionality analysis. An SDR is only considered when there are at least three reports of Raynaud’s phenomenon associated with the drug. A higher signal value (i.e., ROR and IC) indicates a stronger association between the target drug and the suspected AE. The thresholds for identifying an SDR ensure that only statistically robust signals are flagged, suggesting a potential association that warrants further investigation, though they do not confirm causality24,41. With the disproportionality signals identified, further analyses were conducted to explore intra- and inter-class associations.

Intra- and inter-class analysis

We performed several analyses to determine whether CGRP inhibitors are disproportionately associated with Raynaud’s phenomenon in the postmarketing setting. To assess the specificity of the association, we conducted intraclass analyses to compare CGRP inhibitors among themselves, alongside inter-class comparisons with other migraine therapies. The ROR and IC methods were employed to complement these analyses, offering insights into the likelihood of detecting potential safety signals (Supplementary Tables 2 and Supplementary Table 3). A signal was considered significant when ROR025 > 1.0 or IC025 > 0 41,50. We compared Raynaud’s phenomenon cases associated with CGRP inhibitors, both individually and as a class, to those reported for all other drugs in the FAERS database. To mitigate any potential impact of the underlying disease process on the investigated outcome and ensure a balanced study population, Raynaud’s phenomenon reporting associated with the CGRP inhibitor class was also analyzed against active comparator groups composed of other classes of therapies indicated for migraine treatment or prophylaxis. These analyses are crucial for understanding whether the observed association is specific to the CGRP inhibitor class or if it extends to other migraine therapies.

Software, ethics, and reporting guidelines

Statistical analysis and visualization were performed using Microsoft Excel 2019 (Version 2407, Build 17830.20166), GraphPad Prism 8.4.3, and Python 3.12.4. As this observational study relied solely on publicly available, voluntary AE reports and did not involve human subjects, formal ethical approval was not required. This study adhered to the REporting of A Disproportionality Analysis for DrUg Safety Signal Detection Using Individual Case Safety Reports in PharmacoVigilance (READUS-PV) guidelines to ensure transparent, comprehensive, and accurate reporting of our disproportionality analysis52.