Abstract
Objective
To assess whether erenumab influences cerebral vasomotor reactivity and flow-mediated dilation in migraine patients.
Methods
Consecutive migraineurs prescribed erenumab at our Headache Centre and age and sex-matching controls were invited to participate in this observational longitudinal study. Patients were evaluated for cerebral vasomotor reactivity to hypercapnia (breath-holding index) in middle and posterior cerebral arteries and for brachial corrected flow mediated dilation at baseline (T0), after 2 weeks from the first erenumab injection (T2) and after 2 weeks from the fourth Erenumab injection (T18). Patients displaying a reduction of at least 50% in monthly migraine days after completing the fourth month of therapy were classified as responders.
Results
Sixty patients and 25 controls agreed to participate. Middle and posterior cerebral artery mean flow velocities, breath-holding index and flow-mediated dilation did not differ at T0 and from T0 to T2 in patients and controls. In patients, we neither observed a variation of the explored variables from T0 to T18 nor an interaction between evaluation times (T0–T2 or T0–T18) and chronic condition at T0, responder state or erenumab fourth dose.
Conclusions
Our findings demonstrate that erenumab preserves cerebral vasomotor reactivity and flow-mediated dilation in migraineurs without aura.
Introduction
The discovery of the nociceptive sensitisation of the meningeal vessels is a milestone of migraine pathophysiology (1). The activated trigeminovascular terminals release powerful vasodilator peptides; among these, the calcitonin gene-related peptide (CGRP) is probably the most potent in the cerebral circulation (2). Monoclonal antibodies targeting the CGRP pathway are the first pharmacological therapies specifically developed for migraine prevention (3). Among these, erenumab was the first in class approved by drug agencies. Together with the great excitement for the success of the clinical trials (4), some concerns have been raised for the potential harmful effect of blocking CGRP-mediated vasodilation (5–7). Recently, a case of ischemic stroke associated with Erenumab therapy (8) and anecdotal cases of Raynaud’s phenomenon in patients on CGRP antagonist monoclonal antibodies (9) have been reported.
Among other mechanisms, an impairment in cerebral hemodynamics and endothelial function may explain the increased vascular risk observed in migraineurs (10). These two aspects of the vascular tree physiology can be investigated in humans (11).
Vasomotor reactivity (VMR) is a marker of efficiency of the cerebral circulation since it reflects the potential of the intracranial arterioles to dilate in response to vasodilatory stimuli, such as hypercapnia. However, studies assessing VMR in migraine reported conflicting results (12). Overall, migraineurs present a preserved or even higher cerebral VMR (13–15), unless in a chronic condition (16–17).
Similarly, studies addressing the endothelial reactivity in migraine were inconclusive (13,18). The endothelial reactivity can be studied peripherally as brachial artery flow-mediated dilation (FMD). Brachial FMD reflects the arterial capability to self-regulate its tone by an endothelial response to changes in the local environment (19) and is associated with a higher vascular risk (20).
These two techniques measure different aspects of the arterial regulation and are not related to each other (11). Our study aimed to assess whether erenumab influences cerebral hemodynamics (VMR) and endothelial vascular reactivity (FMD) in patients with migraine.
Methods
Our primary question was to examine whether cerebral VMR and brachial FMD are influenced by monoclonal antibodies antagonising the CGRP receptor. This observational longitudinal study was carried out at the Headache and Neurosonology Unit of the Neurological Clinic at the Campus Bio-Medico University of Rome.
The study was approved by our University Ethical committee (prot 6/19 OSS ComET CBM). All participants provided written informed consent.
Anonymised data will be shared upon request from any qualified investigator.
Design and study population
From March to September 2019, migraine patients, treated at our Headache Centre and receiving the clinical indication for erenumab (Aimovig, Novartis Europharm Limited, Nuremberg, Germany) according to the available guidelines (21) were screened to participate in the study by the senior investigator (FV), head of the Headache Centre. Matched controls were recruited among the university hospital personnel volunteering to perform the BH and FMD tests.
Patients and controls were enrolled according to the following criteria. Inclusion criteria for migraine patients were: Diagnosis of migraine without aura according to the International Classification of Headaches (22). Exclusion criteria for migraine patients were: Diagnosis of migraine with aura according to the International Classification of Headaches (22).
For migraine patients and controls, exclusion criteria were: Previous vascular accident, malignancy, other neurological disorders, cerebral or neck vessels steno-occlusive diseases, ongoing vasoactive therapy (calcium antagonist, beta-blockers) or other migraine prophylactic treatments.
Out of the 72 screened patients, a total of 12 patients were excluded from the study. Six patients were excluded because they also suffered from migraine with aura. Two patients were prescribed erenumab in addition to ongoing migraine prophylactic treatment. Four patients did not want to take part in the study.
The sample size was defined presuming an effect size of 0.6 between pairs, according to a previous study demonstrating the effect of prophylaxis therapy on BHI (14). To achieve a statistical power of 80% and a level of significance of 5% (two-sided), the study would have required a sample size of at least 25 pairs. However, we decided to increase the number of enrolled subjects as, being the first observation of the effect of monoclonal antibodies blocking CGRP pathway in migraine patients, we ought to also detect smaller differences than those observed with other prophylaxis therapies (14).
All patients received erenumab every 4 weeks by subcutaneous injection at the initial dose of 70 mg. The dose was adjusted over time upon clinical indication. Participants requiring any changes in medical therapies were considered as drop-outs. Benefit from erenumab treatment (responder status) was defined as a reduction of at least 50% in monthly migraine days (responder rate) after completing the fourth month of therapy.
Patients were also classified as suffering from chronic (CM) or episodic (EM) migraine according to the International Classification of Headaches (22).
Patients underwent the clinical, cerebral VMR, and FMD evaluations at baseline before erenumab injection (T0), after 2 weeks from the first erenumab injection (T2) and after 2 weeks from the fourth erenumab injection (T18). Evaluation times were selected based on the erenumab half-life (4 weeks) and time to reach the steady state of minimal serum concentration (12 weeks) (www.europa.ema.eu). Controls underwent two evaluations performed 2 weeks apart from each other. Hypertension and diabetes were defined according to WHO guidelines (www.who.int). Smoking habits were also recorded.
Each evaluation included simultaneous VMR assessment of the middle and posterior cerebral arteries and brachial FMD measure. In migraine patients, data on migraine onset age, monthly migraine days (MMDs), and analgesics (including triptans) intake were collected during the first 6 months of therapy. Disease history was measured as years from migraine onset.
Because of circadian variations, the FMD and VMR investigations were performed between 8 and 9 AM in a quiet, temperature-controlled room (20–22°C). Both examinations were performed on the same day 1 hour apart from each other. All subjects were studied after a 12-hour overnight fast. Smokers refrained from smoking during the 12 hours preceding the study. Women of reproductive age were investigated during the first week of the menstrual cycle. No patients had a migraine attack during the 24 hours preceding the examination day. Patients were instructed to avoid the intake of triptans in the 3 days preceding the scheduled examinations No patients took analgesics in the 24 hours before VMR and FMD investigations.
Cerebral VMR
All subjects were screened by continuous-wave Doppler and colour flow B-mode Doppler ultrasound (Philips iU22, Bothell, WA, USA) to exclude carotid, vertebral and intracranial steno-occlusive diseases. Cerebral VMR to hypercapnia of the right posterior cerebral artery (PCA) and left middle cerebral artery (MCA) were measured using the breath-holding test (Multidop-X DWL; Elektronische Systeme GmbH, Germany) (23). With the subject lying supine, two transducers placed on the temporal bone window, held on a headband, and with a stable angle of insonation, were used to obtain a bilateral continuous measurement of the flow velocity of middle and posterior cerebral arteries. The highest flow signal was sought at a depth of insonation ranging from 48–54 mm for the MCA and from 55–70 mm for the PCA. Subjects were requested to hold their breath for a period of 30 sec, which was monitored by a capnometer.
The breath-holding index (BHI) was calculated by dividing the percent increase in mean flow velocity (MFV) occurring during breath holding by the time (30 sec) the subjects held their breath after a normal inspiration ([MFV at the end of breath-holding−rest MFV/rest MFV] × 100/sec of breath holding). The mean BHI observed in other cohorts of healthy subjects is around 1.5%/sec (23); wheareas a cutoff value of less than 0.69%/sec has clearly been demonstrated to have a high predictive value for cerebrovascular disease (24).
Traces from subjects who could not hold their breath for 30 sec were discarded (one patient at T0 evaluation and one patient at T2 evaluation). BHI in the PCA and MCA were calculated off-line by a neuro-sonologist blind to subjects’ diagnosis and time of examination.
Pulsatility Index (PI) in the MCA and PCA was automatically calculated by the Doppler machine as Gosling’s pulsatility index (systolic flow velocity−diastolic flow velocity)/MFV) for the time window selected to measure rest MFV (25).
Brachial FMD
We evaluated endothelial function by measuring the change in forearm blood flow induced by flow-mediated dilation (26). All examinations were performed by a single experienced vascular sonographer, using an ultrasound system (Philips iU22, Bothell, WA, USA) with a broadband 7–14 MHz transducer. With the patient in the supine position, the left brachial artery was scanned over a longitudinal section, 3–5 cm above the elbow. Depth and gain settings were optimised to identify the lumen-to-vessel wall interface. The FMD was assessed by measuring the change in brachial artery diameter, after deflation of a cuff placed around the forearm that had been inflated to 50 mmHg above systolic blood pressure for 5 min. The brachial artery diameters were collected after 50, 60 and 70 sec of reactive hyperemia and compared with baseline measurements The arterial diameter was determined as the internal dimension of the vessel wall from the anterior-to-posterior interface between the lumen and the intima. The mean diameter was calculated from three measurements of arterial diameter performed at the end-diastole incident with the R wave on a continuously recorded electrocardiogram. The response of the vessel diameter to reactive hyperemia (FMD) was expressed as a percent change relative to the diameter before cuff inflation. As previously reported (26), FMD was corrected (cFMD) for the shear stress on the blood vessels, which is directly related to the velocity and the viscosity of the blood but inversely related to the vessel diameter. Peak shear rate, estimated as peak systolic flow velocity (sFV) divided by baseline diameter, was calculated to quantify the FMD stimulus in each subject. FMD responses were normalised by dividing the maximal percentage change in diameter by the peak shear rate.
Statistical analysis
Statistical analyses were performed using SPSS 25.0 (SPSS Inc., Chicago, IL, USA). Differences were considered significant at the p < .05 level. Data distribution was assessed by the Kolmogorov-Smirnov test. Continuous variables are presented as mean values ± standard deviation (SD). Median values with interquartile ranges (IQr) are provided for non-normally distributed variables. According to variable distribution, either t-test or Mann-Whitney U test were adopted for comparison of not-repeated measures between not-related groups. The two-tailed Fisher’s exact test was used for dichotomous variables. Correlations were performed with Pearson’s r or Spearman’s ρ coefficients according to data distribution. A linear generalised model for repeated measures was used to assess the variation of normally distributed continuous variables over time from T0 to T2 and T18, with TIME as within-subject factors and GROUPS (patients compared with controls), DIAGNOSIS (chronic compared with episodic migraineurs), DOSE (erenumab 70 mg or 140 mg) or RESPONDER STATE as between-subject factors. To assess changes over time of variables non-normally distributed, Friedman’s analysis of rank was adopted.
Results
Sixty patients and 25 sex and age-matching controls agreed to participate in the study. MCA and PCA BHI, MCA and PCA MFV at rest, brachial basal and peak systolic flow velocities, shear rate, FMD and cFMD and migraine onset age were normally distributed. Age and MCA and PCA PI, brachial artery diameters, monthly migraine days and drug intake were non-normally distributed.
Baseline characteristics
All control subjects presented appropriate temporal bone windows for ultrasound insonation, except for two patients with right insufficient temporal bone windows for PCA insonation and one bilaterally poor insonation. These patients were included in the analysis as only MCA BHI (in two cases) and FMD measures were collected.
Table 1 summarises the comparison between groups of baseline demographic characteristics and of the analysed parameters along T0-T2 examination times. Vascular risk factors frequency did not differ between groups (consistently p > .1). We observed no difference in MCA and PCA BHI, MCA and PCA MFV at rest, MCA and PCA PI, brachial basal and peak systolic flow velocities, brachial artery diameters, FMD and cFMD in patients compared with controls at T0 and T2.
Demographics, vascular risk factors, VMR and FMD at T0 and T2 in patients and controls.
At T0, 35 out of the 60 patients were affected by chronic migraine (CM). Patients presented a median onset age of 15.0 years (IQr 8.0) with a disease history of 34.5 years (IQr 16.5). At baseline, patients reported in the previous month 15.0 MMDs, (IQr 13.0), 16.0 (IQr 16.0) abortive drug intake, of which 9.0 (IQr 15.0) were triptans.
MCA (1.47%/sec SD .52 vs. 1.63%/sec SD .43) and PCA (1.49%/sec SD .52 vs. 1.5%/sec SD .69) BHI, MCA (64.4 cm/sec SD 12.3 vs. 67.5 cm/sec SD 12.9) and PCA (42 cm/sec SD 8.2 vs. 44.3 cm/sec SD 10.1) MFV at rest, PCA PI (.9 IQr .2 vs. .9.sec IQr .15), FMD (14.1% SD 7.3 vs. 18.0% SD 8.2) and cFMD (.43%.sec SD .27 vs. .59%.sec SD .32) did not differ in CM compared with EM patients (consistently p >.10). MCA PI was non-significantly higher (p = .078) in EM (.9, IQr .14) than in CM patients (.825, IQr .2). We found no correlation between MCA and PCA BHI and cFMD; MCA and PCA BHI, and cFMD did not relate to disease history, MMDs, abortive drug and triptan intake (consistently p > .10).
Longitudinal analysis
Five patients dropped out the therapy for different reasons: One patient changed residency, one lost confidence in the therapy after the first administration and three had to change migraine preventive treatment due to erenumab ineffectiveness (response rate inferior to 30%). Fifty-five patients completed 4 months of therapy; of these, 30 were responders (30/55, 54.5%), and 13 received erenumab 140 mg at the fourth administration. Of the 55 patients completing the therapy, two did not show up at the T18 evaluation. From T0 to T18, patients presented a significant decrease in MMDs (χ2 77,813, df 3, p < .00001) and symptomatic drug intake (χ2 48,317, df 3, p < .00001). We observed no variation from T0 to T2 of all the explored variables (Table 1) both in patients and controls, nor any interaction between TIME (T0–2 or T0–18) and GROUPS, CHRONIC condition at T0, RESPONDER STATE (Figure 1) or erenumab fourth administration dose (Table 2).
Longitudinal changes of VMR and FMD in responders and non-responders. The figure shows the longitudinal variations along T0–T2–T18 times of the studied variables, bars indicate 95% confidence intervals.
Linear generalised model assessing variation of VMR and FMD along examination times.
Groups: patients compared with controls; Dose: erenumab 70 mg compared with erenumab 140 mg at the fourth administration; Chronic: chronic migraine compared with episodic migraine at T0; Responder: responders compared with non-responders after the fourth month of therapy.
Discussion
Blocking the CGRP pathway has emerged as a very promising prevention therapy for migraine. However, concerns have been raised about the hypothetical collateral damage of blocking CGRP mediated vasodilation. Cerebral hemodynamics relay on very refined control under the orchestral action of neurogenic, myogenic, endothelial, and metabolic responses. Neurogenic regulation acts differently in arteries of diverse calipers. Medium- and small-diameter arteries respond to neurotransmitters with vasoactive properties released by sympathetic (e.g. noradrenaline), parasympathetic (e.g. acetylcholine) and sensory (CGRP, serotonin, pituitary adenylate cyclase-activating polypeptide (PACAP), nitric oxide (NO)) neurons. On the other hand, smaller intraparenchymal arterioles’ dilatation is mediated by the astrocyte release of prostaglandins and NO in response to the neuronal firing, also called “neurovascular coupling”. The complex interaction between neurotransmitters released by neuronal activation and vessel regulation is a key point in migraine physiopathology and one of the possible mechanisms explaining the link between migraine and stroke risk (27). Similarly, endothelial activation plays a significant role in vessel caliper regulation by the paracrine secretion of substances such as NO and endothelin-1 that have been largely involved in migraine and cerebrovascular physiopathology (28). The myogenic control induces vasoconstriction when vessels’ transmural pressure increases due to the effect of high blood pressure, or vasodilation when the transmural pressure decreases (i.e. autoregulation). Vasomotor reactivity is a property of cerebral circulation that allows vessels to dilate in response to a local increase in H+ concentration (the metabolic response). It becomes of fundamental importance in forcing the oligoemic lower limit of autoregulation when necessary (29). VMR has proven a consistent marker of cerebral hemodynamics efficiency and may correlate with stroke risk (30). However, the studies investigating cerebral VMR in migraine patients were not consistent, leaving unresolved the question of whether migraneurs have a disfunction of cerebral hemodynamics (1–17).
The main result of our study is that erenumab does not exert any effect on cerebral vasomotor reactivity as assessed by Doppler sonography. This is the first study to show the lack of effect of erenumab on markers of vascular physiology in humans, such as FMD and VMR, as previously proven only experimentally (31). Among oral migraine preventive drugs, previous studies observed that flunarizine can influence VMR (14). Since flunarizine does not directly affect vessels’ smooth muscle cells, it can be hypothesised that it modifies VMR through its action on central structures subserving autonomic vascular control (32). It could be argued that unlike flunarizine, erenumab should not pass the blood-brain barrier due to its high molecular weight, hence its effect on cerebral hemodynamics is not expected. This is certainly one of the main reasons for the clinically observed cerebrovascular safety of erenumab (33). Accordingly, we observed that neither does erenumab have any effect on peripheral (i.e. brachial) flow as measured by flow-mediated dilation. FMD is a measure of arterial endothelial-dependent vessel distention occurring when the flow is suddenly restored after a sustained arrest; FMD reflects the endothelial activation in response to the shear stress induced by the reactive hyperemia and largely depends on the availability of NO. NO is one of the first molecules theorised to play a leading role in the physiopathological cascade occurring during migraine attacks: Migraineurs, especially with aura, were hypothesised to display a super-sensitivity to NO release (34). Moreover, CGRP activation induces increased synthesis of NO (35). Our results support the concept that blocking CGRP does not modify endothelial-dependent NO release under physiological circumstances. Finally, in our population, none of the evaluated variables presented significant variations during the therapy course in responders compared with non-responders (Figure 1), or depending on the erenumab dosing.
Our study also shows that cerebral VMR and brachial FMD do not differ in a population of chronic and high-frequency episodic migraineurs compared with healthy controls. It is worth noting that none of our patients suffered from migraine with aura. This finding is in line with previous observations of preserved VMR in patients with migraine without aura (12–14) but contrasts with some reports of reduced VMR and FMD in a chronic condition (16,17). Compared to these latter studies, it is worth noting that our results are derived from a large population of migraine patients with a long-lasting disease history and are based on repeated measures. Studies assessing cerebral hemodynamics in episodic migraine reported controversial findings, and similar inconsistency may be expected in chronic migraine. One possible explanation for these conflicting results is that in high frequent episodic or chronic migraine, even when the examinations are performed outside the headache phase, it cannot be excluded that patients are in the early prodromal or late postictal phases. More studies addressing this issue are required in larger populations.
In summary, our findings demonstrated that the mechanism of action of erenumab does not interfere with two fundamental homeostatic controls of the vascular tone: The metabolic and endothelial responses, supporting with physiological measures the vascular safety clinically observed.
This observation, however, requires confirmation. This is the first study specifically addressing the effect of blocking the CGRP pathway on hypercapnic vasomotor reactivity and flow-mediated dilation in humans. However, the possible influence on other arterial control mechanisms remains to be investigated. In an animal model, the autoregulatory vasodilation in response to hypotension was found to be attenuated by CGRP receptor desensitisation (36); similar conditions ought to be investigated in humans. Similarly, the observation that erenumab action is uninfluential on hypercapnic cerebral VMR and brachial FMD under physiological circumstances does not imply that in conditions exceeding the cerebral or cardiac ischemic thresholds (e.g. stroke, myocardial infarction), preventing CGRP receptor activation would not interfere with hemodynamics and endothelial rescue mechanisms and worsen the ischemic damage (37). Furthermore, in the present analysis, we did not include subjects with migraine with aura. Peculiar cerebral and peripheral hemodynamics profiles (23,26) and increased stroke risk (38) differentiate these patients from those without aura. Specifically, migraineurs with aura seem to present a more reactive cerebral and peripheral vascular tree (12,23,26). At the moment, it cannot be ruled out that this hyperactive vascular system may be more prone to the inhibition of the CGRP pathway.
Clinical implications
This is the first study aiming to investigate the possible negative effect on arterial dilation of anti-CGRP antibodies in migraine patients. Erenumab does not modify cerebral vasomotor reactivity and brachial flow-mediated dilation. Patients with chronic and episodic migraine and control subjects present comparable cerebral vasomotor reactivity and brachial flow-mediated dilation. Cerebral vasomotor reactivity and brachial flow-mediated dilation cannot be used to predict erenumab effectiveness.
Authorship
CA designed and conceptualised the study, analysed the data and drafted the manuscript for intellectual content. GV and AF had a major role in the acquisition of ultrasound data. CMC selected patients. NB and CF had a major role in the quantification of ultrasound data. MS interpreted the data and revised the manuscript for intellectual content. FV screened the patients, designed and conceptualised the study, and drafted the manuscript for intellectual content.
Footnotes
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: FV received travel grants, honoraria for advisory boards, speaker panels or investigation studies from Allergan, Angelini, Eli-Lilly, Novartis, Teva. MS received speaker’s honoraria from Italfarmaco, PIAM, Boehringer and Novartis Pharmaceuticals. CA, GV, AF, CC, NB and CF report no disclosures.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
