Background
Migraine is the third most prevalent and the second most disabling disease worldwide [
1]. Migraine can be classified as episodic (EM) or chronic (CM) according to the number of monthly headache days [
2]. According to the World Health Organization (WHO), migraine ranks third among the most disabling conditions of the human kind [
3] and is the first cause of disability under 50 years of age [
4]. Despite its significant burden, the available preventive treatments for migraine are not specific [
5,
6] and poorly tolerable [
7,
8], while botulinum toxin A is effective only on CM [
9,
10]. Given this background, the advent of monoclonal antibodies against the calcitonin gene-related peptide (CGRP) or its receptor (CGRPr) represent a breakthrough in migraine prevention [
11‐
13]. Evidence from randomized controlled trials strongly supports the efficacy and safety of those agents in the prevention of both EM and CM [
14,
15]. However, real-life data are needed to confirm the results of clinical trials, provide evidence to meet the needs of common clinical practice, and possibly improve the treatment protocols.
Erenumab, a fully human monoclonal antibody directed against CGRPr, is the first approved migraine-specific treatment [
16], whose efficacy and safety were proven in both EM and CM [
17‐
19]. Erenumab is available in two monthly dosages, namely 70 mg and 140 mg [
16], with a slight numerical advantage of the higher over the lower dosage in terms of efficacy [
20]. Compared with the trials, two real-life data from the USA [
21] and Italy [
22] confirmed the efficacy of erenumab in EM and CM after 2 months, with a higher incidence of constipation that did not lead to drug withdrawal [
21]. However, there are currently no real-life studies assessing the efficacy and safety of erenumab over more than 2 months of treatment.
In the present real-life, multicenter study, we aimed to retrospectively review the efficacy and safety of erenumab in patients with EM and CM.
Discussion
Our real-life multicenter study was performed in a difficult-to-treat population (Table
5). Notably, we treated some patients who were excluded by the trials [
17‐
19,
26], including those with > 4 prior preventive treatment failures. Despite this, we found higher proportions of patients with 50%, 75%, or even 100% response compared with the available trials [
17‐
19,
26], possibly due to longer treatment duration, although we cannot exclude the occurrence of placebo effect.
Table 5Comparison between the randomized controlled trials of erenumab for the prevention of migraine and the present study
General characteristics |
Migraine type | Episodic | Chronic | Episodic | Episodic | Chronic and episodic | 83% chronic, 17% episodic | 93.4% chronic, 6.6% episodic |
Dose (mg) | 70 | 70 | 140 | 70 | 140 | 140 | 70 or 140 | 70 | 70 or 140 |
No. of prior preventive treatment failures | < 2 (no response) | ≤3 (no response) | ≤2 (no response) | 2–4 | – | – | ≤2 |
Follow-up duration, months | 3 | 3 | 6 | 3 | 2 | 2 | 6 |
No. of treated patients | 286 | 191 | 190 | 317 | 319 | 121 | 100 | 78 (13 episodic, 65 chronic) | 89 |
Patient characteristics |
Female, % | 85.7 | 87 | 84 | 84.5 | 85.3 | 80 | 83 | 75% (EM), 80% (CM) | 87.6 |
Mean age, years | 42 | 41.4 | 42.9 | 41.1 | 40.4 | 44.6 | | 47.1 (EM), 47.6 (CM) | 46.8 |
Mean migraine duration, years | 22 | 20.7 | 21.9 | – | – | – | – | 29.1 (EM), 30.2 (CM) | 28.2 |
Medication overuse, % | – | 41 | 41 | – | – | – | – | 61.5 (EM), 84.6 (CM) | 71.9 |
Prior preventive treatment failures, % | 87.3 | 67 | 66 | 40.1 | 36.4 | 100 | 100 | 100 | 100 |
Mean MMDs at baseline | 8.1 | 17.9 | 17.8 | 8.3 | 8.3 | 9.2 | – | 10.9 (EM), 22.0 (CM) | 19.8 |
Outcomes |
MMD decrease, mean days | −2.9 | −6.6 | −6.6 | −3.2 | −3.7 | −1.8 | – | −7 (EM), −15 (CM) | −12.4 |
Triptan use days decrease, mean days | −1.2 | −3.5 | −4.1 | −1.1 | −1.6 | −1.3 | – | – | −5.6 |
50% responders, % | 39.3 | 40 | 41 | 43.3 | 50.0 | 30 | – | 100 (EM), 87.5 (CM) | 74.1 |
Adverse events % | 48.1 | 44 | 47 | 57.3 | 55.5 | 55 | 34 | 1.3 | 22.5 |
Serious adverse events % | 1.1 | 3 | 1 | 2.5 | 1.9 | 2 | 5 | – | 2.2 |
The great majority of our patients had CM, while only 5.6% had EM. Therefore, the results of our study were largely conditioned by patients with CM. Besides, patients with EM included in our study had high migraine frequency and high headache-related impact and disability, which put them close to the clinical status of patients with CM [
27]. The study population likely reflects clinical practice, in which erenumab treatment is given to the most difficult-to-treat patients.
All the available randomized controlled trials of erenumab found variable 50% response rates across the different months of follow-up [
17‐
19,
26]. In our study, we distinguished ‘anytime’ from ‘dose-specific’ responders to account for the variability of response. Although most responders had a significant response within 3 doses, new responders added after each dose of erenumab. Therefore, our data support continuing for at least 3 months and even 6 months before discontinuation. Our data also showed that response to erenumab was persistent in most cases; future studies with larger populations and longer follow-up are needed to assess the course of response to erenumab over time.
According to our data, erenumab decreased the intensity, disability and impact of headache and symptoms of depression and anxiety, which were not specifically addressed in the erenumab trials. Migraine and symptoms of anxiety or depression are linked in a bidirectional fashion [
28]; in the present case series, the improvement of psychiatric symptoms might be explained by the reduction in recurrent disabling migraine attacks which do not respond to treatment.
Our data also showed a reduction of allodynia symptoms, indicating a possible role of erenumab in the reversal of the sensitization to head pain typical of CM. Allodynia is a marker of central sensitization and is typical of CM [
29]. Animal studies suggest that CGRP is implied in generating and maintaining allodynia [
30,
31]; therefore, it is not surprising that allodynia symptoms might have been reversed by erenumab in patients with migraine. Notably, the prevalence of allodynia in our population was lower than in previous reports [
32], suggesting potential underreporting or a mild effect of previous treatments.
We did not use detoxication for patients with medication overuse. A recent randomized controlled trial showed that detoxification alone is effective to improve migraine frequency and may avoid costly medications [
33]. However, as already showed in a subgroup analysis of a randomized controlled trial [
34], our data suggest that erenumab alone might help detoxifying patients with medication overuse.
Another important finding of our study was that a relevant proportion of patients with CM who had failed treatment with botulinum toxin A responded to erenumab. This finding points out that the mechanisms of action of botulinum toxin A and of erenumab are different, with two main consequences. First, erenumab might be offered to patients refractory to botulinum toxin. Second, botulinum toxin A and erenumab might be combined in the future, even at the expense of a high cost, to offer the best possible treatment to patients with severe migraine.
As erenumab is an expensive treatment, factors which predict late response should be identified. In our study, patients not responding to treatment had marginally significant higher median MMDs at baseline and a significantly higher consumption of analgesics compared with responders. We did not systematically assess the response to triptans in our patients.
The response to erenumab, as well as to any migraine treatment, is conditioned by the extreme variability of frequency and severity of migraine itself and cannot be univocally defined. A large proportion of patients not achieving a ≥ 50% response might still have a significant gain in terms of disability, associated symptoms, and drug consumption. Better tools are needed to assess the real improvement of patients with migraine after a treatment.
A further point of debate is the role of erenumab dose escalation from 70 mg to 140 mg monthly. In our observational study, dose escalation was allowed throughout the study period, and 2 patients started treatment with a 140 mg monthly dose. As we performed dose escalation at variable time intervals throughout the study, depending on the patients’ response, we could not assess the exact contribution of dose escalation to the efficacy and safety outcomes. Higher doses of erenumab might have given a substantial contribution to patients with a higher number of baseline MMDs (Table
2) or with medication overuse (Fig.
4).
We found comparable rates of adverse events, and especially serious adverse events, in our study as compared with the available randomized controlled trials (Table
4). The proportion of patients with constipation was higher in our study (13.5%) compared with the trials (0–3.6%) [
17,
18,
26] and even open-label extensions [
35,
36], possibly because patients and their treating physicians expected that event; however, it was mild and well controlled with diet or fibers in all cases and did not lead to discontinuation. Treatment discontinuation in our study was higher than in the available trials [
17‐
19,
26] and mostly due to patients’ preference of discontinuing an ineffective treatment; only one patient discontinued the treatment due to an adverse event. Notably, we did not assess the prevalence of possible anti-drug antibodies; however, the current guidelines for the use of anti-CGRP antibodies do not indicate the routine measurement of anti-drug antibodies [
14]. Overall, the great majority (85.4%) of patients were compliant to the treatment throughout the study period, further supporting the safety and tolerability of erenumab in clinical practice.
The strengths of the present study include its relatively large number of patients as compared with previous real-life studies [
21,
22] and a remarkably longer follow-up of 6 months instead of two. The relatively large number of included patients allowed subgroup efficacy analyses in subgroup of patients with medication overuse and failure of botulinum toxin A. However, our study also has several limitations. Firstly, we could not assess the effect of concurrent oral preventive treatments, their withdrawal and reintroduction, due to heterogeneity and small numbers; however, this is a potential source of bias common to all real-life studies, in which treatments are prescribed on a case-by-case basis according to clinical needs; besides, such treatments were withdrawn or introduced in a minority of patients. Secondly, we did not assess some characteristics potentially linked to erenumab response, including prior response to triptans. Thirdly, the design of our study did not allow us to establish a definite role of dose escalation, as this procedure was decided during different time points according to patient response.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.