Abstract
Background
The present study aimed to investigate the predictive value of calcitonin gene-related peptide (CGRP)-induced migraine attacks for effectiveness to erenumab treatment in people with migraine.
Methods
In total, 139 participants with migraine underwent a single experimental day involving a 20-min infusion with CGRP. Following this, the participants entered a 24-week treatment period with erenumab. The primary endpoints were the predictive value of CGRP-induced migraine attacks on the effectiveness of erenumab, defined as ≥50% reduction in monthly migraine days, or ≥ 50% reduction in either monthly migraine or monthly headache days of moderate to severe intensity.
Results
Among participants with CGRP-induced migraine attacks, 60 of 99 (61%) achieved ≥50% reduction in monthly migraine days during weeks 13–24 with erenumab. Conversely, 13 of 25 (52%) where CGRP infusion did not induce a migraine achieved the same endpoint (p = 0.498). There were no significant differences between the ≥50% reduction in either monthly migraine or monthly headache days of moderate to severe intensity between CGRP-sensitive and non-sensitive participants (p = 0.625).
Conclusions
Our findings suggest that the CGRP-provocation model cannot be used to predict erenumab's effectiveness. It remains uncertain whether this finding extends to other monoclonal antibodies targeting the CGRP ligand or to gepants.
Introduction
The therapeutic landscape of migraine prevention has been reshaped with the advent of monoclonal antibodies (mAbs) and gepants targeting the calcitonin gene-related peptide (CGRP) pathway (1). This novel drug class includes erenumab, a mAb against the CGRP receptor, which is confirmed to be efficacious for migraine prevention in both clinical trials and real-world studies (2–8). Despite its overall clinical benefit and beneficial impact on people with migraine, there remains an unmet need for a subset of individuals without responses (2,3,9), raising interest in finding a reliable biomarker that can predict the therapeutic response (10–14). Real-world investigations have documented various clinical features linked to the responsiveness to CGRP-monoclonal antibodies, encompassing unilateral cranial autonomic features, unilateral headache and responsiveness to triptans. 15 Nonetheless, endeavors to validate these observations have produced inconclusive outcomes, potentially influenced by constraints in sample size, divergences in study methodologies or variations in outcome selection (14,16–20). Consequently, the field of migraine management continues to lack a predictive biomarker suitable for clinical application, thereby delaying the advent of a new era in precision medicine.
A promising area ripe for biomarker exploration is understanding the relationship between CGRP hypersensitivity and the effectiveness of erenumab. Findings from pilot experiments show an ambiguous pattern: people with migraine who experience attacks after intravenous (IV) infusion of CGRP benefit from preventive treatment with erenumab (21). Nevertheless, in this study, a subset of people who experienced CGRP-induced migraine attacks did not respond to treatment with erenumab (21). Thus, it remains uncertain whether experiencing a CGRP-induced migraine attack in a clinical setting can predict a subsequent positive therapeutic effectiveness to erenumab. If such an association is clearly established, a new treatment paradigm could potentially be developed. In this scenario, a CGRP-infusion could be administered for 20 min and evaluated after 24 h (i.e. CGRP-induced migraine attack or not) later by a healthcare professional. The CGRP-infusion test could then serve as a less expensive, safe and fast tool to tailor a personalized treatment plan to avoid a more costly evaluation period.
Here, we present the results of an open-label, single-arm trial involving adult participants diagnosed with migraine. The participants underwent IV infusion of CGRP on one experimental day followed by a 24-week treatment period with erenumab. Our aim was to ascertain whether the induction of CGRP-induced migraine attack might serve as a predictive biomarker for the subsequent effectiveness to preventive treatment with erenumab.
Methods
The data presented herein was collected as part of the larger parental Registry for Migraine (REFORM) study (22). A detailed description of the REFORM study has previously been published (22). In brief, the REFORM study is a prospective, single-center study that involved people screened for migraine at the Danish Headache Center. The study protocol received approval from the relevant ethics committee (H-20047793), and all participants provided written informed consent in accordance with the principles of the Declaration of Helsinki. The study was registered at ClinicalTrials.gov (NCT04592952).
Design
This is a 24-week, single-center, phase IV observational study with a non-randomized design. The study included a single experimental day involving CGRP infusion, a four-week baseline period, followed by a 24-week intervention phase with erenumab (Figure 1). Participants received continuous IV infusion of CGRP (1.5 μg/min) for 20 min in the antecubital fossa on one experimental day using a time- and volume-controlled infusion pump. The dosage of CGRP administered in this study was identical with doses used in previous human experimental studies (23–25). Participants who reported experiencing any headache or taking rescue medications within 24 h prior to infusion start were rescheduled for another experimental day. The study involved four pre-scheduled visits, including the experimental day (CGRP-infusion day), baseline visit, and weeks 12 and 24 following erenumab treatment, as illustrated in Figure 1. The baseline visit was designated as the visit conducted subsequent to the four-week baseline period, during which the participants received their initial erenumab dose.
Study overview. The initial visit served as the experimental day. During this visit, participants underwent a continuous intravenous infusion of 1.5 µg of calcitonin gene-related peptide (CGRP) over a 20-min period. Subsequently, participants entered a 4-week baseline period, during which they were required to maintain a headache diary for at least 21 out of 28 days to be eligible for the initial dose of erenumab. Following the successful completion of the 4-week baseline period, all eligible participants proceeded to the open-label treatment phase. In this phase, they were assigned to receive subcutaneous injections of erenumab at a dosage of 140 mg every four weeks for a total duration of 24 weeks.
Participants
Recruitment of participants took place between October 2020 and January 2022 at the Danish Headache Center. The study population comprised both women and men aged 18 years or older who had a minimum of 12 months of migraine history, with or without aura, meeting the criteria outlined in the International Classification of Headache Disorders, third edition (ICHD-3) (26). Additionally, participants needed to have experienced migraine onset prior to the age of 50 years. To be enrolled in the study, participants were required to have a frequency of at least four days per month on average for three months fulfilling the ICHD-3 criteria for migraine without or with aura. The study protocol allowed for the inclusion of participants, who had been using a stable dosage of migraine preventive medications for a minimum of two months prior to study enrollment. Exclusion criteria for this study included pregnancy, lactation, a history of cluster headache, hemiplegic migraine, post-traumatic headache, or the inability to differentiate between migraine and other types of headaches. A comprehensive description of the study's inclusion and exclusion criteria can be found in the Supplemental material (Table S1).
Procedures
On the experimental day, participants arrived at the Danish Headache Center between 8:00 am and 12:00 pm in a non-fasting state. Prior to the infusion of CGRP, a comprehensive medical and neurological examination was conducted, along with a 12-lead electrocardiogram. In the case of female participants, a urine hCG pregnancy test was administered before the infusion. Participants were informed about the possibility of CGRP inducing headache, although specific details regarding timing or headache type were not disclosed. All study-related procedures were performed with participants lying in a supine position, including the insertion of a venous catheter into an antecubital vein to establish peripheral intravenous access.
Vital signs, headache characteristics, associated symptoms such as nausea, photophobia and phonophobia, usage of any rescue medication, and any adverse events were documented 10 min before the CGRP infusion, and then every 10 min thereafter, up to 60 min after the initiation of the infusion. Following this, participants were discharged and instructed to maintain an hourly headache diary for 12 h from the start of the infusion.
The subsequent day, participants entered a four-week baseline period. During this four-week baseline period, participants were required to complete the headache diary for a minimum of 21 out of 28 days to be eligible for the initial dose of erenumab. The headache diary included various questions such as the presence of any headache, the intensity of the headache measured on a four-point Likert scale (ranging from none to severe based on interference with daily activities), whether the participant considered the headache to be a migraine, the duration of the headache, the presence of aura (visual, sensory or speech/language), associated symptoms, menstrual days and use of rescue medication for headache.
Following the completion of the four-week baseline period, all qualified participants continued into the open-label treatment phase, where they were assigned to receive subcutaneous injections of erenumab 140 mg every four weeks for a duration of 24 weeks.
Case definitions
An experimentally-induced migraine attack was defined as participants satisfying either criteria (1), (2) or (3) during the 12-h observational period.
CGRP-induced migraine attack that meets criteria C and D for migraine without aura, as outlined in ICHD-3
26
or if participants reported experiencing a headache mimicking their typical migraine. Cases were also interpreted as CGRP-induced migraine attacks, if the CGRP-induced headache attack was treated with the participant's usual acute headache medication. CGRP-induced migraine aura was characterized as fully reversible neurological symptoms that either satisfied the ICHD-3 criteria B and C for migraine with aura or closely resembled the typical aura symptoms experienced by the participant.
26
Responders were defined as participants who achieved ≥30%, ≥50% and ≥75% reduction in mean monthly migraine days following erenumab treatment, from baseline compared to week 13–24. Baseline was defined as the four-week baseline period, where participants were required to complete the headache diary for a minimum of 21 out of 28 days.
Outcomes and measures
The two primary endpoints were the value of CGRP-induced migraine attacks on predicting the effectiveness to erenumab, defined as: (1) ≥50% reduction in monthly migraine days following erenumab treatment and (2) ≥50% reduction in either monthly migraine days or monthly headache days of moderate to severe intensity. The two secondary endpoints investigated whether CGRP-induced migraine attacks predicted the ≥30% and ≥75% responder rates for: (1) monthly migraine days and (2) monthly migraine days or monthly headache days of moderate to severe intensity. The exploratory endpoints investigated whether CGRP-induced nausea, CGRP-induced severe intensity (rated 7 to 10 on an 11-point numeric rating scale) and CGRP-induced exacerbation by physical activity could predict the response of erenumab treatment. All analyses were calculated as averages over week 13–24 compared to the baseline period. Monthly migraine days were assessed using both prospective electronic and paper headache diaries, whereas monthly headache days were only evaluated using prospective paper headache diaries.
Statistical analysis
Data that followed a normal distribution were reported using the mean ± SD, whereas data that did not follow a normal distribution were reported using median with interquartile range (IQR). Categorical data were presented as the number and proportion of participants. To assess normality, Shapiro–Wilk's test was used, and, if the group size was greater than 30, normality was assumed based on the central limit theorem. Additionally, the distributions were visually inspected to confirm normality. In case of non-normal distributed data, we used Wilcoxon rank-sum test. A Student's two-sample t-test was used to compare normally distributed data. Fisher's exact test was used to compare categorical data.
To handle missing data in the outcome measures monthly migraine days or monthly headache days, normalization was applied. When diaries contained at least 21 days of data within a four-week timeframe, the available days were normalized by multiplying them with the inverse completion rate. Similarly, for the response period spanning weeks 13–24, data from diaries with at least two available four-week periods were normalized to cover the entire response period.
The purpose of the study was to examine whether response to intravenous CGRP-infusion serves as a predictive biomarker for the treatment effectiveness of erenumab. As such, no sample size calculations were performed. Instead, we intended to enroll a minimum of 400 study participants with migraine and continue to enroll study participants throughout the study duration to maximize the performance of the predictive model.
In November 2021, an interim analysis was conducted using data from 100 participants who had completed the 24-week intervention period. An additional 39 participants had received CGRP infusion and were in weeks 1–24 of the intervention period. The study was terminated following the interim analysis because no significant difference was observed in the ≥50% reduction in monthly migraine days from baseline during weeks 13–24 between those who experienced CGRP-induced migraine attacks and those who did not (45% vs. 35%, p = 0.849). After a comprehensive evaluation, including ethical considerations within the context of the human experimental design where participants can potentially develop headache, a decision was made to terminate the study for safety, ethical and scientific reasons. P < 0.05 was considered statistically significant. All statistical analyses were conducted using R, version 4.1.0 (R Foundation, Vienna, Austria).
Results
Participants
In total, 139 participants were enrolled and completed the experimental day with CGRP infusion. Out of the initial 139 participants, 15 dropped out and consequently did not proceed to the four-week baseline period. As a result, 124 individuals entered the four-week baseline period, with none being excluded or dropping out after being assigned the 140-mg erenumab dose. Consequently, all 124 participants completed the 24-week intervention period with erenumab, providing eligible data for analysis.
Participants’ mean age was 44.1 ± 18 years, 118 (85%) were self-reported females and the mean body mass index was 24.6 ± 4.4 kg/m2. The median number of monthly headache days experienced by the participants was 16.7 (IQR = 5.5) and the median number of days with migraine per month was 10.7 (IQR = 4.3). A history of both migraine with and without aura was reported by 34 (24%) participants. Among the 139 participants, 130 (94%) reported taking triptans as acute medication, and 84 (60%) reported currently taking migraine preventive therapy. Table 1 shows the demographic and clinical characteristics of the entire population, broken down by their response to CGRP infusion.
Demographics and baseline characteristics of the study population.
CGRP = calcitonin gene-related peptide; n = number; IQR = interquartile range.
CGRP-induced migraine attacks
Intravenous infusion of CGRP induced migraine attacks in 110 (79%) out of 139 participants during the 12-h observational period. Thus, 29 (21%) participants did not develop CGRP-migraine attacks. In addition, out of 34 participants diagnosed with migraine with aura, 13 (38%) individuals reported experiencing migraine with aura following to CGRP infusion. In-depth information of CGRP-induced migraine attacks and aura is reported elsewhere (27,28). During the four-week baseline period, 11 (10%) out of 110 participants who experienced CGRP-induced migraine attacks declined to receive treatment with erenumab. In addition, four (14%) out of 29 who did not manifest migraine attacks by CGRP infusion opted not to undergo erenumab treatment. Therefore, in total, 124 participants completed the 24-week intervention period with erenumab.
Effectiveness of erenumab
Ninety-five (76.6%) out of 124 participants achieved a ≥30% reduction in monthly migraine days from baseline to weeks 13–24. Correspondingly, a ≥50% response was observed in 73 out of 124 (58.9%) participants, whereas a ≥75% response was noted in 27 (21.8%) out of 124 participants.
Change from baseline during week 13–24 in mean number of monthly migraine days. Responders: participants developed migraine attack following calcitonin gene-related peptide (CGRP) infusion. Non-responders: participants developed no migraine attack following CGRP infusion. The percentages ≥ 30%, ≥ 50% and ≥ 75% refer to the reduction in monthly migraine days from baseline during weeks 13–24.
Primary endpoint
Among 99 participants with CGRP-induced migraine attacks, 60 (61%) experienced a reduction of ≥50% in monthly migraine days from baseline during weeks 13–24 with erenumab treatment (Figure 2). In comparison, among those who did not experience CGRP-induced migraine attacks, 13 of 25 (52%) participants also achieved a similar reduction in monthly migraine days (p = 0.498, odds ratio (OR) = 1.42, 95% confidence interval (95% CI) = 0.53–3.76).
We had data from 115 participants, focusing on the endpoint ≥50% reduction in either monthly migraine days or monthly headache days of moderate to severe intensity. Among participants who developed CGRP-induced migraine attacks, 65 out of 91 (71.4%) achieved this endpoint. The corresponding figures were 16 of 24 (66.7%) among those who did not manifest by migraine attacks triggered by CGRP infusion. The analysis revealed no statistically significant differences between these two groups (p = 0.625, OR = 1.25, 95% CI = 0.41–3.56).
Secondary endpoints
Among participants who developed CGRP-induced migraine attacks, treatment with erenumab led to a ≥30% monthly migraine days reduction in 78 (79%) and a ≥75% monthly migraine days reduction in 22 (22%) of participants from baseline to week 13–24 (Figure 2). Comparatively, for those who did not experience CGRP-induced migraine attacks, a ≥30% monthly migraine days reduction was achieved in 17 (68%) participants, whereas a ≥75% reduction was seen in five (20%) participants from baseline to week 13–24. The statistical analyses showed no statistically significant differences between these two groups (≥30% monthly migraine days reduction: p = 0.293, OR = 1.74, 95% CI = 0.57–5.00; ≥75% monthly migraine days reduction: p = 1.00, OR = 1.14, 95% CI = 0.36–4.34). Similarly, no significant differences were found between ≥30% and ≥75% reduction in either monthly migraine days or moderate to severe headache days (≥30% reduction: p = 1.00; ≥75% reduction: p = 1.00).
Exploratory endpoints
We investigated whether CGRP-induced nausea, severe intensity, and exacerbation of symptoms due to physical activity, could function as predictors of the response to erenumab. Our analysis revealed that a higher proportion of participants who responded to erenumab treatment reported experiencing CGRP-induced nausea (p = 0.012, OR = 3.08, 95% CI = 1.12–7.89). However, CGRP-induced severe intensity (p = 0.511, OR = 0.70, 95% CI = 0.27–1.80) and CGRP-induced exacerbation by physical activity (p = 0.312, OR = 1.72, 95% CI = 0.57–5.22) did not demonstrate predictive value for the erenumab treatment response.
Discussion
In the present study, we aimed to determine whether hypersensitivity to CGRP, specifically as manifested by migraine attacks triggered by CGRP infusion, could predict the effectiveness of erenumab treatment in people with migraine. Our results indicate that hypersensitivity to CGRP infusion does not correlate with the therapeutic response to erenumab. This suggests that the CGRP-provocation model may not be an appropriate predictive biomarker for erenumab treatment response.
In the present study, a positive response to erenumab treatment was demonstrated in both participants with CGRP induced migraine attacks and participants who did not manifest migraine attacks induced by CGRP infusion. In cases where CGRP infusion did not trigger a migraine attack, it is conceivable that a higher dose or longer infusion duration may induce migraine attacks in these participants. Recently, we investigated potential associations between clinical and sociodemographic factors and the onset of CGRP-induced migraine attacks (27). However, no robust factors significantly linked with the development of CGRP-induced migraine attacks were identified (27). A plausible explanation for this finding may stem from the administration of a high dose of CGRP. This supposition finds support in the observation that a 20-min infusion of vasoactive intestinal polypeptide (VIP) (8 pmol/kg/min) failed to trigger migraine attacks in individuals with migraine (29,30), whereas a 120-min infusion of VIP (8 pmol/kg/min) notably induced migraine attacks compared to placebo (31).
In the present study, a conventional dosage of CGRP was employed; however, it is plausible that utilizing a “sub-threshold dose” of CGRP may lead to a reduction in the incidence of CGRP-induced cases. This leads to the conjecture that participants experiencing CGRP-induced migraine at a sub-threshold dose may be more predisposed to demonstrate a positive response to erenumab treatment. Moreover, it is worth considering that the threshold for CGRP-induced migraine attacks may be influenced by genetic factors. It is pertinent to underscore that these theories are speculative at this point. Preclinical studies indicate a dose-response relationship between elevated CGRP concentrations and a reduction in mean arterial pressure, coupled with an elevation in heart rate, serving as surrogate makers of vasodilation (32). Future research must therefore employ a range of CGRP doses in both animal- and human models of migraine, aiming to elucidate distinctions in nociception and the occurrence of CGRP-induced migraine attacks.
The lack of association between CGRP hypersensitivity and the effectiveness of erenumab treatment could potentially be attributable to the shared impact of both CGRP and erenumab on the AMY1 receptor (33–35). Binding of CGRP to AMY1 receptor induces similar downstream effects (e.g. upregulation of cAMP) as the canonical CGRP receptor (36,37), potentially initiating a migraine attack. Supporting this notion, a recent human experimental study elucidated a partial prevention of CGRP-induced migraine attacks by erenumab (38). It is possible that circulating CGRP may engage other receptors, such as AMY1, ultimately leading to migraine attacks. Thus, the potential involvement of receptors beyond the CGRP receptor, can contribute to our observed variability in erenumab responses among individuals hypersensitive to CGRP.
It is of notable significance to acknowledge that apart from CGRP, there exists other neuropeptides capable of activating the canonical CGRP receptor pathway. These CGRP homologous peptides include amylin and adrenomedullin, both known to induce migraine attacks in people with migraine (36,37). Although these neuropeptides can bind and activate the canonical CGRP receptor, their binding affinity is comparatively lower than that of CGRP (33,34). Conversely, preclinical evidence indicates that erenumab may antagonize not only the canonical CGRP receptor, but also the AMY1 receptor (39–41). In this context, erenumab may be capable of counteracting endogenous amylin by blocking the AMY1 receptor, potentially averting the onset of spontaneous attacks. This theoretical framework holds the potential to explain why participants without CGRP hypersensitivity still responded to erenumab, implying potential hypersensitivity to endogenous amylin and/or adrenomedullin.
The discordance link between CGRP hypersensitivity and the effectiveness of erenumab treatment could be ascribed to various factors. It is conceivable that migraine attacks may stem from CGRP or other circulating molecules in certain instances, whereas, in others, they may occur independently of these molecules. Migraine attacks may likewise be influenced by central subcortical/cortical mechanisms and changes (42–44). Finally, our findings may be influenced by additional contributing factors, such as the absence of a placebo arm during both CGRP infusion and erenumab administration.
Strengths and limitations
This is the first large study (n = 139) aiming to predict the effectiveness of erenumab treatment in people with migraine by CGRP infusion. We implemented rigorous eligibility criteria and maintained a comprehensive headache diary, ensuring consistent and reliable data collection. However, the study is not without limitations.
The study was terminated for futility based on the interim analysis. Although we cannot definitively dismiss the potential marginal predictive significance of CGRP-induced migraine attacks in relation to the response to erenumab treatment, it was considered ethically unacceptable to prolong the study post-interim analysis, which indicated notably insignificant outcomes. As detailed elsewhere (45,46), futility criteria serve to safeguard participants from undue risks (e.g. adverse effects such as CGRP-induced migraine headache) in instances where there exists no reasonable probability of the study achieving its primary endpoints. The small subgroup of participants (n = 29) who did not develop a migraine attack post-CGRP infusion might have limited our statistical power to detect differences in erenumab response. As a result, minor differences in the responder rate to erenumab could have potentially gone undetected. Furthermore, the lack of a placebo-controlled design applies both to the experimental day involving CGRP infusion and within the 24-week intervention period. The inclusion of a placebo arm within the experimental sub-study could have reduced the induction rate of migraine attacks following CGRP infusion. On the other hand, a placebo arm during the 24-week treatment phase could have diminished the number of participants reporting favorable outcomes following treatment with erenumab. Moreover, the vasodilation effect induced by CGRP infusion, resulting in a subsequent flushing, could potentially increase the participant's expectations regarding the development of a CGRP-induced attack. However, it is important to acknowledge that this limitation remains unavoidable, even with the inclusion of a placebo arm. Also, we permitted participants to use other preventive migraine medications, a factor that may have influenced their responsiveness to CGRP infusion and their subsequent response to erenumab. Additionally, we consider CGRP-induced nausea to be an incidental finding because our study was not specifically designed to evaluate this outcome. We also suspect that there might be correlations among the baseline clinical characteristics of our study population that could potentially forecast the response to erenumab. However, our study lacks the statistical power to explore these relationships comprehensively. Lastly, it is worth noting that the majority of participants were recruited from a tertiary referral center, possibly limiting the generalizability of our findings to a broader population of people with migraine.
Conclusions
The study results suggest the inadequacy of the CGRP-provocation model as a biomarker for predicting the response to erenumab treatment. Yet, it remains an open question whether this model's limitations extend to other monoclonal antibodies targeting the CGRP ligand or to gepants. Notably, the study yielded responses to erenumab in people who did not manifest CGRP hypersensitivity, prompting a compelling inquiry into the underlying mechanisms. In this context, the potential role of other neuropeptides, specifically amylin and adrenomedullin, merits consideration. Further research is warranted to unravel these intricate interactions and enhance our understanding of personalized migraine treatment strategies.
Clinical implications
Our findings suggest that there is no correlation between hypersensitivity to CGRP infusion and the therapeutic response to erenumab. This is evidenced by positive responses to erenumab observed even in individuals who did not demonstrate CGRP hypersensitivity. It remains uncertain whether this observation applies to other monoclonal antibodies targeting the CGRP ligand or to gepants.
Supplemental Material
sj-docx-1-cep-10.1177_03331024241258734 - Supplemental material for Hypersensitivity to CGRP as a predictive biomarker of migraine prevention with erenumab
Supplemental material, sj-docx-1-cep-10.1177_03331024241258734 for Hypersensitivity to CGRP as a predictive biomarker of migraine prevention with erenumab by Haidar M Al-Khazali, Håkan Ashina, Rune Häckert Christensen, Astrid Wiggers, Kathrine Rose, Afrim Iljazi, Faisal Mohammad Amin, Messoud Ashina, Josefin Snellman, Tina Maio-Twofoot and Henrik W Schytz in Cephalalgia
Footnotes
Acknowledgements
We thank all of the study participants and the research staff the Danish Headache Center (Mohammad B. A. Lafta, Sarra Al-Khazali, Amir Al-Saoudi, Merete B. Bertelsen and Janne Jensen).
Author contributions
HS, FMA, HA, AI, JS and MA were responsible for concept and design. HMA, RHC, HA, FMA, MA and HS were responsible for acquisition, analysis or interpretation of data. HMA, RHC, HA, AW, KR, MA, JS, TMT and HS were responsible for drafting the manuscript. HMA, RHC, HA, FMA, AI, AW, KR, HS, MA, JS, TMT and HS were responsible for critical revision of the manuscript for important intellectual content. HMA, RHC and HA were responsible for statistical analysis. MA, FMA and HA were responsible for administrative, technical, or material support. MA, FMA, HA and HS were responsible for supervision.
Data availability statement
Upon reasonable request, the corresponding author will provide the necessary data and materials to interested researchers for the purpose of academic scrutiny, reproducibility and further scientific investigation.
Declaration of conflicting interests
The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: HMA reports receiving personal fees from Pfizer outside of the submitted work. HA reports receiving personal fees from Lundbeck and Teva outside of the submitted work. HS reports receiving personal fees from AbbVie, Teva, Lundbeck, Novartis, Eli Lilly, outside of the submitted work. FMA reports receiving personal fees from Pfizer, Teva, Lundbeck, Novartis, and Eli Lilly outside of the submitted work. MA reports receiving personal fees from AbbVie, Amgen, Astra Zeneca, Eli Lilly, GlaxoSmithKline, Lundbeck, Novartis, Pfizer, and Teva Pharmaceuticals outside of the submitted work. MA also serves as an associate editor of Cephalalgia, The Journal of Headache and Pain, and Brain. The remaining authors report no competing interests. JS and TMT are employees of and hold shares in Novartis Pharma AG.
Ethical statement
The study protocol received approval from the relevant ethics committee (H-20047793), and all participants provided written informed consent in accordance with the principles of the Declaration of Helsinki
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was funded by a research grant from the Lundbeck Foundation (R310-2018-3711). This study was funded by and conducted in collaboration with Novartis Pharma AG, Basel, Switzerland. Novartis Pharma, Lundbeck Foundation, (grant number R310-2018-3711).
Supplemental material
Supplemental material for this article is available online.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
