What can be learned from the history of recurrence in migraine? A comment

Early on in 1989, during the open phase II studies, attention was drawn to the clinical problem of recurrence. Thus, in an open study on subcutaneous sumatriptan 2–3 mg, ten patients in one Danish centre were given a questionnaire concerning recurrence within 24 h of treatment in the clinic by Dr. Iversen, Gentofte Hospital, Denmark [6]. Five out of ten migraine patients experienced that the migraine headache recurred within 24 h after successful treatment in the first case and these recurrences occurred within the usual duration of the migraine attack [6]. In the other centres, there was no systemic follow-up after the patients left the clinic, and only one recurrence was observed in 101 patients [6]. This clearly demonstrated that in order to observe recurrence one had to look for it by administration of a questionnaire about it.

In the first edition of the guidelines [7] on clinical drug trials in migraine of the International Headache Society from 1991 the proposed primary efficacy parameter was as follows: Number of attacks resolved within 2 h. It was recommended that “number of migraine attacks resolved within 2 h, before any escape medication, should usually be the primary parameter of efficacy. Whenever an attack remits within 2 h, and relapses within 24 h, it is a treatment failure by this criterion” [7]. In practice this parameter is roughly similar to the later in 2000 proposed “sustained pain-free” parameter which is defined “as pain-free within 2 h with no use of escape medication or relapse within 48 h” [8]. This was later modified to 24 h in the Lancet meta-analysis [9]. In the comments from 1991 [7] it was noted that “if a drug is effectively quickly in bringing resolution of the attack, but the attack relapses because of a short duration of action of the drug (as has been observed in patients with longlasting attacks), repeated intake of the same drug can be optional; this requires a special study design” [7]. Thus the committee members were aware of the problem of recurrence most likely by personal experiences from the then ongoing extensive sumatriptan trial programme [2].

In the triptan development programmes, recurrence has been defined as headache relief (a decrease in headache from moderate or severe to mild or none) after 2 h and recurrence of moderate or severe headache within 24 h.

In one of the first randomised clinical trials (RCTs) on subcutaneous sumatriptan 6 mg in 1991 recurrence was not mentioned [10]. It was concluded in this American RCT “that sumatriptan is an effective treatment for patients with migraine. A significant reduction in headaches, clinical disability, nausea, and photophobia occurs within minutes of a subjections injection, with lasting effects for up to 24 h” [10]. However, the patients kept a diary for 48 h after receiving treatment and this conclusion was reached despite the fact that only 34% of the patients remained completely pain free for 24 h [10]. Also in the large oral dose-defining study from 1991 there was no mention of recurrence and no mention of a follow-up after treatment [11].

Even in 1992 it was noted in a paper [12] on CNS adverse events of subcutaneous sumatriptan that these AEs had a short-lasting time profile parallel to the kinetics of the drug whereas” the pharmacodynamic effect with respect to headache, however, last for about 24 h” [10]”.

In contrast, a multiple-dose study of oral sumatriptan from 1991 reported recurrence in 48% of sumatriptan-treated patients [13]. Similarly, in an international RCT on subcutaneous sumatriptan, 6 mg which was done at the same time, it was observed that the migraine recurred in 38% of patients within 24 h after subcutaneous sumatriptan 6 mg [14] Thus even with the most effective way of administering a triptan a considerable recurrence rate was found [15, 16].

In the two comparative RCTs, published in 1991 and 1992, in which oral sumatriptan (recurrence in 41–42%) was compared with ergotamine (30%), and aspirin plus metoclopramide (33%) recurrence was evaluated [17, 18]. From this time on, recurrence was evaluated in almost all RCTs with triptans [16, 19, 20].

From 2000 when IHS [6] recommended sustained pain-free and after the meta-analysis of oral triptans from 2001 in the Lancet [9], most studies have reported on this efficacy measure instead of headache recurrence. In the meta-analysis [9, 21] a rather low sustained pain-free response was found. Thus for sumatriptan 100 mg sustained pain free 2-24 h was 20% and for rizatriptan 10 mg (25%), eletriptan (25%) and almotriptan (27%) it was somewhat higher [21]. Even so, with the best oral treatment at that time less than one-third of patients had a sustained pain-free response.

Whereas addition of a second dose of sumatriptan did not prevent headache recurrence [22‐24] the combination of sumatriptan 85 mg and naproxen 500 mg resulted in more patients (24%) being sustained pain-free than after sumatriptan 85 mg (16%) [25].

The pros and cons of recurrence being attack- or patient dependent versus treatment-dependent are summarised in Table 1. First, a second dose of oral sumatriptan 100 mg was tried as a preventive drug for recurrence [22‐24]. Sumatriptan was given double-blindly 2–4 h after an open-labelled first dose of either subcutaneous [22] or oral sumatriptan [23, 24]. The second dose of sumatriptan did not decrease the incidence of recurrence compared with placebo [22, 23, 25]. This indicated that the incidence of recurrence did not correlate with the pharmacokinetics of sumatriptan. In contrast, sumatriptan was found effective in the treatment of recurrence in four RCTs [16].

In two studies from 1996, Visser et al. [1, 26] investigated the problem of recurrence. In one study in 366 migraine patients risk factors for recurrence were evaluated. Headache recurrence occurred more frequently in patients with more severe attacks and longer untreated attack duration [25] In a second study, Visser et al. [1] could find no correlation between the recurrence of migraine and the pharmacokinetics parameters or the pharmacodynamics parameters (effect on cranial arteries as measured by ultrasound) of subcutaneous sumatriptan studied outside attacks. They concluded that recurrence is most likely patient-dependent [1] and that the results “may imply that novel sumatriptan-like drugs with a more rapid or extensive absorption or a longer plasma half-life may not result in higher initial response rates or prevention of headache recurrence” [1]. Multivariate logistic regression analysis of the eletriptan trial programme identified predictors of headache recurrence [27]. These predictors were age of >35 years, females and severe attacks at baseline [27]. This indicates that the recurrence is mainly patient-dependent.

In 2003, Géraud et al. [28] found a correlation between the half-lives and recurrence rates of oral triptans. However, the incidence of recurrence depends on female gender, age of ≥35 years, and severe baseline characteristics [27], as mentioned above, and these factors were not included in the analysis [28] To illustrate, in two RCTs with zolmitriptan the treated migraine headache was moderate in 73–75% of patients [29, 30], whereas in one RCT with rizatriptan the treated migraine headache was severe in 55% of patients [31]. The recurrence rates in these RCTs with different baseline severity should thus not be compared. Comparison of recurrence rates or sustained pain-free response should thus only be performed in direct comparative RCTs in which randomisation ensures comparable baseline severity of migraine headache and the other predictors for recurrence [27].

In the analysis of recurrence with oral triptans [28] recurrence rates of 17% for frovatriptan 2.5 mg and 33% for sumatriptan 100 mg were used. In contrast, in a direct comparative RCT frovatriptan 2.5 mg (25% recurrence) with a half-life of 26 h did not result in significantly fewer recurrences than sumatriptan 100 mg (31% recurrence) with a half-life of 2 h [32]. This indicates that even with a huge difference in elimination t_½ among two triptans there is no difference in recurrence. The most likely explanation for this is that low triptan levels, as illustrated in Fig. 1, do not influence the risk for recurrence.

In summary, there are thus several pros for recurrence being attack- or patient-dependent.

In contrast, the effect of ergot alkaloids, less recurrence than a triptan in five out of six RCTs in which this parameter was measured [16] speaks strongly against recurrence being patient-dependent. Similarly, the combination of sumatriptan and naproxen [24] resulted in more patients being sustained pain-free (24%) than after sumatriptan (16%) indicating a treatment factor for recurrence.

Some patients have migraine attacks which if untreated last up to 72 h [33]. It is a clinical observation that if they are treated with a triptan they risk multiple recurrences with intake of triptans one to two times a day for several days. This indicates that the migraine process continues despite symptomatic relief by a drug. It has correspondingly been shown with PET scan that even after successful treatment with subcutaneous sumatriptan the brainstem activation found during migraine attacks is persistent [34, 35]. The brain stem activation has been termed the “migraine generator” [36]. Also the postdromes, the most common being tiredness, observed in 68% of patients, indicate [37] that a process is ongoing after the actual attack. Similarly, adverse events such as sedation after triptans occur more frequently after successful treatment indicating demasking of symptoms of the migraine attack [38]. One theoretical way to circumvent this problem is the using of triptan with a very long half-life, e.g., frovatriptan with a t_½ of 26 h (Table 1). However, as suggested in Fig. 1 the terminal t_½ may theoretically not be relevant for recurrence.

Finally, pharmacodynamics may be more important than pharmacokinetics for recurrence. Ergotamine has a kinetic t_½ of 2 h, but a pharmacodynamic t_½ of 10 h [39] due to a tight binding to the arterial receptor. Thus, in vitro the constrictor effect of ergotamine on human temporal and coronary arteries cannot be washed out [40‐42].

In rat middle cerebral artery the contractions induced by ergotamine and dihydroergotamine (DHE) were typically slow in on and off set (about 30–60 min) [43]. The long duration of ergot alkaloids should be investigated further in an attempt to design drugs with less recurrence [43]. DHE has a terminal t_½ of 10 h but, is in my opinion, more likely the tight binding to the receptor that is important [44].

The slow dissociation from the receptor on arteries of DHE and ergotamine also explains the slow onset of action of ergot alkaloids (Fig. 2). The ergot alkaloids’ behaviour, slow onset of action, and long duration of action, fits best with an effect on arteries [39] or veins [44, 45].