European Headache Federation (EHF) consensus on the definition of effective treatment of a migraine attack and of triptan failure

Triptans are serotonin agonists acting on the 5-HT_1B/1D(/1F) receptors, which were specifically designed for the acute treatment of migraine. Their vasoconstrictive properties are mediated by the 5-HT_1B receptor present in arterial smooth muscles [1]. Activation of 5-HT_1D receptor on trigeminal fibers inhibits the release of peripheral vasoactive neuropeptides such as substance P and calcitonin gene-related peptide (CGRP). Some triptans can, with varying affinity, activate the 5-HT_1F receptor, which also inhibits CGRP release and possibly nociceptive modulation and dural neurogenic inflammation [2, 3]. Apart from their 5-HT_1F receptor affinity, all triptans have similar pharmacodynamics, although second generation triptans (almotriptan, eletriptan, frovatriptan, naratriptan, rizatriptan, and zolmitriptan) were designed to have an improved pharmacokinetic profile compared with sumatriptan, including a higher oral bioavailability combined with a comparable or increased plasma half-life or faster rapid onset of action [4, 5]. In general, triptans are presumed too hydrophilic to cross the blood-brain barrier and thus to have a central effect.

The clinical introduction of sumatriptan was a breakthrough for acute migraine treatment. To date, seven triptans in different formulations have demonstrated their superiority to placebo in clinical trials and are currently available for moderate to severe migraine attacks [6]. Triptans are widely prescribed not only in headache centers, and general neurology but also in primary care [7].

In randomized clinical trials (RCTs), several endpoints have been used to assess triptan efficacy and, consequently, response rates. Most trials focused on one single migraine attack and considered substantial pain relief or pain freedom after 1 to 4 h as an adequate response. Despite the general good efficacy, almost one-third of individuals with migraine did not meet such endpoints across trials [8]. However, the lack of response to one triptan during one single attack is not sufficient to determine a general poor response to the whole medication class. In fact, individuals with migraine might respond to the same triptan during a different attack, or to a higher dose and a different formulation of the same triptan or another triptan type. Other factors, such as the timing of dosing relative to the attack onset, accompanying nausea/vomiting, or medication overuse can also contribute to the differences in response [9].

A general definition of triptan failure does not exist. From a clinical point of view, defining criteria for non-response could have several benefits:

Evidence on the precise link between pharmacodynamics, pharmacokinetics, and clinical response to triptans is limited and hindered by inconsistent findings. When comparing responders with non-responders, one study observed that absorption of oral sumatriptan was lower and slower in unsatisfactory responders [10], while other studies found no differences in pharmacokinetics nor genetic diversity of the 5-HT_1B or 5-HT_1F receptor [11, 12] after oral or subcutaneous sumatriptan. Yet, some findings do suggest that pharmacokinetics plays a role: the lack of response to one triptan does not predict responsiveness to another triptan [13], and the highest response rates are observed after subcutaneous sumatriptan. Of interest is a systematic review and meta-analysis which found that women have a higher drug exposure compared with men, without their response being higher [14].

Factors of importance for headache response to triptans may include time of administration and rapid onset of action, such as observed with intranasal or subcutaneous administration [10, 15]. Longer half-life and higher 5-HT_1B receptor potency also influence effectiveness as they are associated with lower headache recurrence [16]. Predictors of response to triptans also include the normalization of elevated CGRP levels in saliva or the extra-jugular vein after drug administration [17, 18], but due to high inter-individual variability, such predictors are not yet useful in the clinical context [19].

The primary aim of the present Consensus was to provide a definition of triptan failure to be used in clinical practice and research. We acknowledge that definition of triptan failure cannot be developed in the absence of a definition of effective treatment of an acute migraine attack and of a definition of triptan-responder and for this reason we agreed upon those definitions before establishing a consensus definition of triptan failure.

The Consensus Panel proposed a definition of response to drugs for the acute treatment of migraine that is patient-centered, with the pivotal concept of patient reported “well-being”.

The initial intention of the Panel was to focus on response, resistance, and refractoriness to triptans; however, when considering each single drug, the same barriers to effective care are present for any drug class. The literature search on triptans faced a high heterogeneity in design, definition of outcomes, and results across the available RCTs and observational studies (Fig. 2). Therefore, most of the proposed definitions resulted from open discussion and personal experience of the Consensus Panel. The new definitions are primarily for use in clinical practice, to select individuals in need for the optimization of acute treatment. However, on the long run, these definitions can have a pathophysiological significance as the identification of group of individuals, resistant or refractory to triptans, could lead to the identification of new pharmacological targets.

When discussing the definitions, the consensus panel faced several issues, including the reporting of drug response or triptan non-response, the elements of variability in response or non-response to drugs, and the difference between resistance and refractoriness to triptans.

To establish efficacy of available treatments in clinic, it is important to use an attack report form (diary). An easy-to-use electronic or paper diary that captures predefined endpoints should be used. Data for acute attack treatment should be entered in real time to limit recall bias. Adverse events may also be collected in the diary. Reporting the degree of response to acute drugs could be important; however, we should be aware that complicated report forms with detailed description of symptoms may be difficult for subjects to fill out during attacks. The quantity and quality of collected data might be inversely proportional. Future clinical research should focus on the balance between a simple user friendly and an extensive complete report of attacks, by individuals with migraine. This applies not only to pain intensity, but also for non-pain symptoms, or adverse events, or the other components of well-being that could be impaired by migraine and restored by acute treatment.

As reported in Fig. 2, the proportion of response to triptans in terms of pain freedom or pain relief was extremely variable across studies. Studies available in the literature assessed some factors that potentially influence response to triptans, including the route and timing of administration. However, other factors, including the clinical presentation of migraine attacks, external triggers such as menstruation or stress, and pharmacogenomics should be further studied to obtain treatment optimization. The Consensus Panel agreed that not only pain, but also non-pain symptoms contribute to the satisfaction of individuals with migraine with their acute treatments. Considering the “most bothersome symptom” together with headache has become a mainstay of RCT for the acute treatment of migraine in recent years [35‐38]. RCTs and observational studies on triptans were mostly focused on migraine pain without considering non-pain symptoms. Migraine-related disability and patient-reported outcomes were also poorly assessed by those studies and should receive more attention from physicians and researchers in the field of headache disorders. An important issue when considering response to drugs for the acute treatment of migraine is the duration of response, which includes the recurrence of pain and the need for rescue medication. It is unclear whether pain recurrence or the use of rescue medication impair the well-being of individuals with migraine or their satisfaction with use of triptans. The Consensus Panel did not reach a conclusion whether these parameters should be included in the definition of triptan response or non-response (Table S8). Pain recurrence and use of rescue medication should be included in the assessment of response to acute medication in clinical practice and undergo further testing before being considered into a definition of response or non-response.

A further issue worth considering is consistency of response to each triptan across multiple attacks. There is uncertainty about the number of attacks to be treated before declaring non-response to a triptan. The Consensus Panel suggested that a triptan could be considered effective if restoring well-being in at least three quarters of migraine attacks (Fig. 3). However, response to acute drugs is influenced by many factors, many of which are independent from the action of triptans, such as hormonal or psychological triggers. Monitoring response to triptans with a standardized tool such as a headache diary could help overcoming the barriers to an optimal acute treatment of migraine. Careful monitoring of that response might lead to the identification of novel strategies for optimization of acute migraine treatment. It could also lead to the rapid identification of individuals eligible to novel treatments or to experimental studies of acute treatments.

The present Consensus not only provided a definition referring to response to each single drug, but also referred to the individual with the introduction of “resistance” and “refractoriness” to triptans (Fig. 4). The cutoffs for resistance and refractoriness were arbitrary and mostly based on the Panel members’ clinical experience. Despite that, they could give an account of clinical conditions of different individuals. In individuals who are resistant to triptans, the drugs could still be used after optimization, while in those who are refractory the drugs should be abandoned in favor of other classes. The difference in response to the drugs could reflect genetically determined differences in the mechanisms and mediators of migraine attacks.

Response and non-response to acute drugs for the treatment of migraine have been historically heterogeneously defined. The Consensus Panel provided experienced-based definitions that simplify the assessment of response to acute drugs and non-response to triptans in clinical practice. The new definitions help to identify suitable individuals for novel treatments and may support studies to gain insights into the pathophysiology of migraine attacks.