FormalPara Key Summary Points

Why carry out this study?

In patients with migraine, failure of multiple preventive therapies is often required before receiving anti-calcitonin gene-related peptide monoclonal antibody treatment (e.g., eptinezumab); however, guidelines for how many and which therapies must fail vary by country and payor

This study evaluated the efficacy of eptinezumab versus placebo across 24 weeks of treatment in the DELIVER study in patients with migraine and 2–4 prior preventive treatment failures and in subgroups defined by type of prior treatment failure (topiramate, metoprolol/propranolol, amitriptyline, and/or flunarizine)

What was learned from the study?

Regardless of prior preventive treatment failure type, treatment with eptinezumab resulted in greater reduction in monthly migraine days (MMDs) than placebo across 24 weeks of treatment in all subgroups

A larger proportion of patients treated with eptinezumab than with placebo experienced ≥ 50% MMD reduction across all subgroups

Given the differences in preventive migraine treatment guidelines, it is important that the efficacy of eptinezumab does not appear to be impacted by the type or prior preventive medicine that led to treatment failure

Introduction

Migraine is a prevalent neurological disease [1] that is burdensome and costly [2] to individuals, disabling their participation in society [3]. An unmet need existed for treatment that provides sustained migraine prevention to reduce migraine-related burden, particularly because discontinuation of traditional oral preventive treatments due to lack of efficacy, tolerability, and poor adherence is common [4]. The newer class of monoclonal antibodies (mAbs) targeting calcitonin gene-related peptide (CGRP) provides an additional option in the preventive migraine armamentarium to potentially address this unmet need.

Preventive migraine treatment regulations and guidelines can vary by country and payor; thus, patients with migraine may each have an individualized migraine treatment journey prior to finding effective treatment. In Spain, the Ministry of Health and Spanish Agency of Medicines and Medical Devices reimburses anti-CGRP mAbs to patients who have ≥ 8 headache days per month and who have not responded to three or more preventive treatments (one of which must be onabotulinumtoxinA in the case of chronic migraine) [5,6,7]. In the UK, the National Institute for Health and Care Excellence recommends at least three preventive migraine treatment failures (i.e., beta blockers, antidepressants, and/or antiepileptics) prior to prescribing an anti-CGRP mAb [8,9,10,11]. The French National Authority for Health recommends at least two prior preventive treatment failures, with beta blockers (propranolol and metoprolol) and antiepileptics (topiramate) recommended for first- and second-line treatments, respectively [12, 13]. The Danish Medicines Council recommends anti-CGRP mAbs for use in chronic migraine if at least one antihypertensive and one antiepileptic medication have failed for a patient [14]. The Canadian Agency for Drugs and Technologies in Health recommends at least two prior preventive treatment failures as well, with more treatment types mentioned as potential preventive treatment options (i.e., beta blockers, tricyclic antidepressants, verapamil or flunarizine, sodium valproate, topiramate, or gabapentin) despite a lack of clinical evidence [15, 16]. In the US, the Institute for Clinical and Economic Review provides guidance of at least two preventive migraine treatment failures from two different classes prior to prescribing anti-CGRP mAbs, with botulinum toxin A as the first-line treatment for chronic migraine and antiepileptics (divalproex sodium, sodium valproate, topiramate) or beta blockers (metoprolol, propranolol, timolol) as the first-line treatment for episodic migraine [17]. Due to the multiple potential lines of therapy before reaching anti-CGRP treatment in current guidelines, it is important to demonstrate consistent effect of anti-CGRP treatment regardless of prior preventive treatment failure type.

Eptinezumab is the most recent anti-CGRP mAb to be indicated for the preventive treatment of migraine in adults in the US, Europe, and other regions [18, 19]. Multiple large-scale phase 3 clinical trials established the efficacy and safety of eptinezumab for migraine prevention, including its rapid and sustained migraine-preventive effects [20,21,22,23,24,25]. One of these trials, DELIVER, evaluated the efficacy and safety of eptinezumab in patients with 2–4 prior preventive migraine treatment failures [25]. The primary data from the 24-week placebo-controlled period of DELIVER demonstrated reductions in migraine frequency and severity with eptinezumab compared with placebo in this difficult-to-treat population. This post hoc analysis evaluated the efficacy of eptinezumab versus placebo across 24 weeks of treatment in DELIVER subgroups defined by type of prior treatment failure.

Methods

Study Design

The DELIVER study was a multicenter, parallel-group, double-blind, randomized, placebo-controlled phase 3b clinical study conducted at 96 locations across Europe and the US. The detailed protocol, statistical analysis plan, and patient disposition flow chart have been published [25].

The DELIVER study was conducted in accordance with the standards of Good Clinical Practice as defined by the International Conference on Harmonisation and all applicable federal and local regulations. All study documentation was approved by the local review board at each of the 96 study sites across Europe and the US or by a central institutional review board or ethics committee (Supplementary Materials Appendix 1). All patients provided written informed consent prior to any study procedures. The DELIVER trial is registered on ClinicalTrials.gov (NCT04418765) and EudraCT (2019-004497-25).

Patients

Detailed inclusion and exclusion criteria have been published [25, 26], with key criteria summarized here. Eligible patients, aged 18–75 years, were required to have a history of migraine as defined by International Classification of Headache Disorders (3rd edition) diagnostic criteria for ≥ 12 months before screening and onset at or before 50 years of age. Patients had to have documented evidence (medical record or physician confirmation) of 2–4 prior preventive treatment failures due to inadequate efficacy, tolerability, or contraindications within the past 10 years. Prior failure could be with propranolol or metoprolol, topiramate, amitriptyline, flunarizine, candesartan, valproate or divalproex, or botulinum toxins A/B (if taken for chronic migraine). At least one failure had to be due to inadequate efficacy with one of the following medications: propranolol or metoprolol, topiramate, amitriptyline, flunarizine, or candesartan. Patients were ineligible if they had a treatment failure on a previous treatment targeting the CGRP pathway or had a treatment failure on valproate/divalproex or onabotulinumtoxinA and the treatment was not the most recent preventive medication prior to study inclusion.

Randomization and Treatment

DELIVER randomized adults with migraine to receive eptinezumab 100 mg, 300 mg, or placebo via intravenous infusion at baseline (Day 0) and Week 12. Blinded personnel administered infusions over a period of 30 (up to 45) min.

Data Collection and Outcomes

Patients completed a daily electronic diary, which captured headache episodes, migraine attacks, and headache characteristics, including severity, length, and intake of acute medication. Patients attended four physical visits (screening, baseline, and end of Weeks 12 and 24) and four phone contact visits (end of Weeks 4, 8, 16, and 20). The primary efficacy endpoint in the DELIVER study was the change from baseline in the number of monthly migraine days (MMDs) following the first infusion (Weeks 1–12) [25]. A migraine day was defined using International Headache Society criteria [27]. This post hoc analysis summarizes the changes from baseline in MMDs and the percentages of patients experiencing ≥ 50% reduction in MMDs (i.e., ≥ 50% migraine responder rates [MRRs]) in subgroups of patients based on prior treatment type.

Statistical Analyses

Sample size and power estimation were not conducted for this post hoc analysis. Data were analyzed for the following subpopulations of patients with treatment failures on: topiramate, beta blockers (metoprolol or propranolol), amitriptyline, and/or flunarizine. Due to the low number of patients, subgroups defined by prior failure on candesartan, valproate or divalproex, and/or botulinum toxin were not analyzed. Patients may have had prior failures across multiple types of treatment; therefore, subgroups are not mutually exclusive.

The estimated mean changes from baseline in MMDs are from a mixed model for repeated measures that included effects for 4-week periods from Weeks 1–4 to 21–24, country, stratification factor (monthly headache days [MHDs] at baseline: ≤ 14/ > 14), and treatment as factors; baseline MMD score as a continuous covariate; treatment-by-month interaction; baseline score-by-month interaction; and stratum-by-month interaction. Restricted maximum likelihood estimation was used. The ≥ 50% MRRs were analyzed using logistic regression. The model included baseline MMDs as a continuous covariate and treatment and stratification (MHDs at baseline: ≤ 14/ > 14) as factors. P-values (two sided) were not controlled for multiplicity and thus are descriptive.

All statistical analyses were conducted using SAS software (SAS Institute, Inc., Cary, NC, USA) v9.4 or later.

Results

The full analysis set comprised 890 patients (eptinezumab 100 mg, n = 299; eptinezumab 300 mg, n = 293; placebo, n = 298). There were 633 (71.1%) patients for whom topiramate had previously failed, 538 (60.4%) for whom beta blockers failed, 508 (57.1%) for whom amitriptyline failed, and 333 (37.4%) for whom flunarizine failed. Demographics by subgroup are reported in Table 1. Patients were predominantly female (~ 90%), with a mean age of ~ 44 years and 13.7–14.2 mean MMDs at baseline. Rates of comorbid psychiatric (12–14%), vascular/cardiac (11–15%), and nervous system (3–5%) disorders were similar across groups.

Table 1 Demographic characteristics, by prior preventive treatment failure type
Full size table

Reductions in MMDs were greater with eptinezumab than with placebo in all subgroups (Table 2). Differences from placebo in change from baseline were similar across subgroups. In patients treated with eptinezumab across subgroups, larger differences from placebo in change from baseline were consistently observed following a second infusion (Weeks 13–24). Greater reductions in MMDs were observed with eptinezumab versus placebo as early as Weeks 1–4 (Fig. 1), with the effect sustained or further improved through the duration of the 24-week treatment period.

Table 2 Change from baseline in monthly migraine days over 12-week intervals, by prior preventive treatment failure type
Full size table
Fig. 1
figure 1

Change from baseline in monthly migraine days over 4-week intervals, by prior preventive treatment failure type. The estimated LS means are from a mixed model repeated measure with month, country, stratification factor (monthly headache days at baseline: ≤ 14/ > 14), and treatment as factors, baseline score as a continuous covariate, treatment-by-month interaction, baseline score-by-month interaction, and stratum-by-month interaction. BL baseline [mean monthly migraine days], LS least squares

Full size image

Across all subgroups over Weeks 1–12, more than one-third of patients treated with eptinezumab 100 mg and approximately half of patients treated with eptinezumab 300 mg achieved ≥ 50% migraine response compared with < 15% of patients who received placebo (Fig. 2). The proportion of patients achieving ≥ 50% migraine response increased following a second infusion, with more than half of all patients treated with eptinezumab achieving ≥ 50% reduction in MMDs over Weeks 13–24. Similar ≥ 50% MRRs were observed across prior failure subgroups. Across subgroups and time points, both eptinezumab doses had a 3.2–9.4-fold odds of ≥ 50% response compared with placebo (Table 3).

Fig. 2
figure 2

Migraine responder rates of ≥ 50% over 12-week intervals, by prior preventive treatment failure type. The ≥ 50% response threshold is calculated as the average percentage change in monthly migraine days

Full size image
Table 3 Odds ratios of ≥ 50% migraine response over 12-week intervals, by prior preventive treatment failure type
Full size table

Discussion

Given the differences in preventive migraine treatment regulations and guidelines as described earlier, and that eptinezumab is currently a later-line preventive option in many regions, it is important to understand whether the efficacy of eptinezumab is impacted by failure of earlier-line preventive therapies. In this post hoc analysis of patients with migraine and 2–4 prior preventive migraine treatment failures, eptinezumab reduced MMDs more and resulted in larger ≥ 50% MRRs than placebo regardless of the type of prior treatment failure (i.e., topiramate, beta blockers, amitriptyline, and flunarizine). The efficacy of eptinezumab was further improved with a second dose, as evidenced by larger differences from placebo in change in MMDs and higher rates of ≥ 50% migraine response across all subgroups. Patients who received placebo had a lower treatment effect, as seen in the full analysis set [25] compared to the PROMISE-1 [20] and PROMISE-2 [22] studies, which did not differ between subgroups and which did not substantially improve following a second placebo infusion. Psychiatric, vascular/cardiac, and nervous system disorders were similar across subgroups, suggesting no influence of comorbidities on prior treatment failure type. Together, these results provide post hoc evidence that the efficacy of eptinezumab is not meaningfully affected by the type of treatment that had failed prior to eptinezumab initiation, and that efficacy is sustained across 6 months of treatment.

These results are similar to a post hoc analysis of galcanezumab in a similar population [28], suggesting that across anti-CGRP mAbs, the type of prior treatment failure does not substantially alter the efficacy of anti-CGRP treatment. The post hoc galcanezumab analysis used the CONQUER study population of patients with migraine and 2–4 prior preventive treatment failures, and analysis subgroups for topiramate, amitriptyline, propranolol, metoprolol, valproate or divalproex, and onabotulinumtoxinA [28]; across prior failures, the effect with galcanezumab did not differ substantially. While studies with erenumab (LIBERTY) and fremanezumab (FOCUS) have demonstrated efficacy in patients with 2–4 prior treatment failures, the analyses did not evaluate efficacy across individual prior treatment types [29, 30].

The types of medications that must fail before patients may begin treatment with anti-CGRPs belong to different pharmacological classes, and they often show little efficacy and poor tolerability in treatment of migraine [4, 31, 32]. In a study conducted prior to the approval of anti-CGRP mAbs in 2021, discontinuation of these preventive treatments was common [33]. In the 2013 International Burden of Migraine Study, 41% of patients with chronic migraine discontinued one or more preventive medications [4]. The most common reasons for discontinuation were lack of efficacy (39–48%, depending on treatment class) and side effects (34–53%) [4]. In 2021, the ATTAIN study reported that 28.2% of patients discontinued their preventive medication (antidepressants, antihypertensives, antiepileptics, anticonvulsants, botulinum toxin A/B) without consulting their health care provider within the 6-month follow-up period [33].

Prior to anti-CGRP mAbs, no other preventive medications were developed specifically for the preventive treatment of migraine. One potential reason for high discontinuation with traditional oral preventive treatments is that these medications were developed for conditions other than migraine (e.g., onabotulinumtoxinA for pain [34], amitriptyline for depression [35]). The benefits of eptinezumab, regardless of prior preventive medication type that led to treatment failure, together with the lack of rationale and limited efficacy and tolerability of traditional preventive migraine treatments warrant further discussion that the landscape of treatment sequence could be modified and updated.

Limitations

These results are subject to the limitations inherent in post hoc analyses; however, the findings are strengthened by being post hoc analyses of predefined endpoints using predefined methodology. The population of DELIVER comprised predominantly white and female participants, which may limit generalizability to the full migraine population. Because of the range of prior preventive treatment types and prescribing possibilities, dosage and duration of prior preventive medication were considered outside the scope of this manuscript. Additionally, the prior treatment failure subgroups are not mutually exclusive. Subgroups that are independent of each other (i.e., where patients fall into only one category) may lead to different efficacy of anti-CGRP treatment. Analyses did not account for differential lines of therapy within treatment types (i.e., various combinations of prior failures) or time since failure. Candesartan, valproate/divalproex, and onabotulinumtoxinA groups were too small to justify analysis. Patients for whom onabotulinumtoxinA was not the most recent failure were excluded, which may have impacted the inclusion rate. The DELIVER study included a small number of patients with four prior preventive treatment failures (eptinezumab 100 mg, n = 19; eptinezumab 300 mg, n = 14; placebo, n = 27); therefore, differences across number of prior treatment failures were not analyzed [25]. Given regional variations in procedures and treatment sequence requirements as well as differences in access to preventive treatment, this study may be limited in generalizability to the general migraine population with multiple treatment failures.

Conclusions

Regardless of prior preventive treatment failure type, treatment with eptinezumab demonstrated greater reductions in migraine frequency and increases in magnitude of response compared with placebo across all subgroups, with reductions sustained through the 24-week treatment period. Differences from placebo in change from baseline increased following a second infusion with both eptinezumab doses across all subgroups, suggesting that a second dose may provide additional benefit.