Erenumab in the Prevention of High-frequency Episodic and Chronic Migraine

Erenumab in Real Life in Italy (EARLY), The First Italian Multicenter, Prospective Real-life Study

Piero Barbanti MD, PhD; Cinzia Aurilia MD; Gabriella Egeo MD, PhD; Luisa Fofi MD; Sabina Cevoli MD, PhD; Bruno Colombo MD; Massimo Filippi MD; Fabio Frediani MD; Francesco Bono MD; Licia Grazzi MD; Antonio Salerno MD; Bruno Mercuri MD, PhD; Antonio Carnevale MD; Claudia Altamura MD, PhD; Fabrizio Vernieri MD


Headache. 2021;61(2):363-372. 

In This Article


The main finding of this study is that erenumab at the dose of 70 mg is effective in patients affected by HFEM or CM with ≥3 previous preventive therapeutic failures, since it induces a progressive reduction in migraine frequency, use of analgesics, pain severity and disability, and presents a good safety and tolerability profile. Only a small proportion of patients (11%) needed to increase the dose to 140 mg due to no effectiveness or loss of effectiveness.

The efficacy, safety, and tolerability of erenumab have been documented by large RCTs in both EM and CM. In EM (4–14 MMDs), erenumab outperformed placebo in the reduction of MMDs at the dose of 70 mg in the ARISE trial[20] (therapeutic gain: −1.0; p < 0.001) and at both 70 and 140 mg in the STRIVE trial[13] (therapeutic gain: −1.4 and −1.9, respectively; p < 0.001). Moreover, at 140 mg, erenumab induced a 50% or greater reduction from baseline in the number of MMDs compared to placebo (30% vs. 14%, p = 0.002) in migraine patients with 2–4 prior preventive treatment failures in the LIBERTY trial.[21] Erenumab also met all secondary endpoints in the STRIVE and LIBERTY trials and two out of three secondary endpoints in the ARISE study. In CM, erenumab induced a significant reduction in MHDs at both 70 and 140 mg (therapeutic gain over placebo: −2.5; p > 0.0001), also satisfying all secondary endpoints.[12]

Even though the data emerging from an observational study cannot be directly compared with that of the RCTs, it is worth noting that our real-life results show that erenumab 70 mg induced a ≥ 50% reduction in the number of MMDs/MHDs at week 12 compared to baseline in almost 60% of the patients affected by HFEM or CM, a proportion which is nearly twofold that reported in the LIBERTY study in patients with EM using a higher erenumab dose (140 mg).[21] Our data are in keeping with the proportion of erenumab responders (57.5%) identified in an early cohort of Australian patients affected by refractory CM[22] and are even more compelling when considering that the proportion of patients who had previously not benefited from at least three previous migraine were 100% in our study and 63% in the LIBERTY study. We also documented a good tolerability profile with a low number of patients complaining of at least one adverse event (13.7%) at the end of the study. Nevertheless, constipation seems to be more common (8.8%) than previously reported,[13,20,21] it being severe and rated as a serious adverse event in one patient and leading to treatment discontinuation in four patients.

Translating the erenumab scientific evidence on erenumab into clinical practice through real-life studies does not only confirm its effectiveness, tolerability, and safety in a multifaceted clinical setting of patients with diverse comorbidities and multiple prior therapeutic failures, but also addresses different questions in migraine management such as response predictability and unraveling possible specific migraine endophenotypes.[17,23–25]

A placebo-controlled trial reported that erenumab responders are hypersensitive to CGRP infusion.[26] However, a CGRP provocation model is not easily applicable to clinical practice. A new, clinically useful finding emerging from the present real-life study, featuring an in-depth patient clinical characterization, is that easily obtainable clinical features—that is, pain localization, presence of dopaminergic symptoms, psychiatric comorbidities, migraine frequency, and number of prior treatment failures—might be of help in predicting the therapeutic response to erenumab. The contribution of CGRP to the development and maintenance of both peripheral and central sensitization in the trigeminal system[27] could explain the positive association we found between unilateral pain localization (in HFEM) and presence of dopaminergic symptoms (in CM) and the responsiveness to erenumab, a monoclonal antibody, which disrupts the CGRP effects along the trigeminovascular system.[5,6] The pathophysiological mechanism underlying unilateral migraine pain is, in fact, a pronounced peripheral trigeminal sensitization, which is a process probably associated with an increased CGRP release and leading to a pain tracing the anatomical distribution of a hyperactive ophthalmic branch of the trigeminal nerve. Furthermore, migraine patients with unilateral pain, especially when associated with accompanying homolateral parasympathetic symptoms, are not only supposed to be super-responders to trigeminal-targeted treatments, but have also an increased central trigeminal sensitization.[23,25,28,29] It is worth noting that the proportion of erenumab responders, we found (approximately 60%) is in line with that of other trigeminal-targeted treatments and of migraine patients who developed migraine-like episodes following the experimental intravenous administration of trigeminal neuropeptides.[27] Notably, 60% is also the frequency of unilateral pain in migraine patients as reported by Kelman in a large cohort.[30]

The presence of dopaminergic symptoms identifies a putative migraine endophenotype characterized by long-lasting attacks with clinical manifestations related to peripheral and central trigeminal sensitization, namely unilateral cranial autonomic symptoms, allodynia, and osmophobia.[19] The underlying pathophysiological substrate is probably a reduced inhibition exerted by the A11 dopaminergic cells on the neuronal discharge in trigemino-cervical complex, facilitating an undisputed firing of trigeminal nociceptive inputs.[19]

The positive association between basal CM frequency and erenumab responsiveness is in line with the finding of frequency-dependent increase in interictal peripheral blood CGRP levels in migraine patients.[31] Thus, it is conceivable that the higher is the number of headache days at baseline—and therefore—the higher are the CGRP basal levels—the better will be the response to CGRP targeted treatments, at least up to a certain headache frequency. In fact, near daily CM patients often present more relevant psychiatric comorbidities,[32] which could exert a detrimental effect on the effectiveness of specific migraine preventative treatments. This could be due to the fact that the psychological stress triggering effect exceeds the therapeutic properties of the CGRP antagonism.[33]

One major finding of the present study is a high ≥50% response rate in a migraine population usually considered "hard-to-treat," given the number of previous preventive medication failures (Table 1). This notwithstanding, we documented a negative association between the number of prior therapeutic failures and the responsiveness to erenumab, in our probably due—as suggested by daily clinical practice—not only to a less crucial CGRP involvement in these patients, but also to the presence of neglected or inadequately/unsuccessfully treated psychiatric comorbidities.

We acknowledge that the study has some limitations. First, we considered only HFEM and CM referred to third level headache centers, not including patients with lower migraine frequencies. Second, we analyzed only the data of the 230 patients who were treated for at least 12 weeks. Third, our decision at week 8 to increase at week 8 the erenumab dose to 140 mg in nonresponders and in responders who had become nonresponders for at least 4 weeks could be seen as arbitrary. On the contrary, the strengths of the present experience are the involvement of nine headache centers representative of the northern, central, and southern Italian population—a larger patient number compared to other real-world evidence studies[22,34–37]—and the use of a shared semi-structured questionnaire detailing behavioral and sociodemographic factors and all clinical migraine features.