Abstract
Objective:
The objective was to determine, through a literature review, whether treatment during the premonitory phase of migraine is a potentially useful migraine management strategy.
Methods:
A general literature review was done with regard to the nature of migraine premonitory symptoms, their frequency, their reliability in predicting migraine attacks, and the effectiveness of medication treatment when given during the premonitory phase.
Results:
Many different symptoms have been reported as premonitory symptoms that occur before migraine attacks. Up to 87% of patients with migraine may experience premonitory symptoms, although some studies have provided estimates as low as 33%. In selected patients, premonitory symptoms may be relatively reliable predictors of a migraine attack to follow. Both naratriptan (open-label study) and domperidone (double-blind, randomized, placebo-controlled study) have been reported to be effective when given during the premonitory phase.
Conclusions:
More research is needed, but there is some evidence that medication treatment during the premonitory phase has the potential to be helpful in selected patients with migraine.
Introduction
A number of migraine sufferers feel that they can predict their attacks with some degree of accuracy. The question arises, therefore, whether such knowledge can be used by those with migraine to treat their attacks more effectively. It is generally accepted that, at least during the pain phase of the migraine attack, early treatment is more effective than treatment delivered later once the attack is fully developed. Could it be, therefore, that treatment during the premonitory phase would be even more effective, and if so, do enough migraine sufferers experience reliable premonitory symptoms that they can make use of for this purpose to make this avenue of therapy worth pursuing?
In this paper, the types of premonitory symptoms which patients report will be discussed, and the literature will be reviewed as to how reliable and potentially useful these are for patients as a predictor of the pain phase of the migraine attack. The literature on treatment during the premonitory phase will also be reviewed to determine whether this might be a useful management approach.
Definition
Premonitory symptoms are symptoms which precede the other symptoms of the migraine attack by 2–48 h, and forewarn of the other symptoms to come. They occur before the aura in migraine with aura, and before the onset of pain in migraine without aura (1, 2).
Types of premonitory symptoms
Premonitory symptoms can take many different forms, and can be quite dramatic. In his book, Migraine, Oliver Sacks (3) describes a particularly interesting case where a man ‘of normally phlegmatic nature’ would ‘have an almost uncontrollable tendency to laugh or sing or whistle or dance’ for 2 or 3 h before the onset of his headaches.
The premonitory symptoms reported by most patients, however, are somewhat more mundane. Kelman (4) classified the premonitory symptoms reported by patients with migraine in the clinical setting into categories of general fatigue, mood changes, gastrointestinal symptoms, and ‘other’. Other authors have compiled more specific lists. As long ago as 1984, Waelkens (5) listed a variety of symptoms based on patient report. Consistent with Kelman’s categories (4), these included fatigue and lethargy, depression, anxiety and irritability, and meteorism (bloating). He also included a number of unusual and difficult to define symptoms such as a feeling of narrowness in the head, slightly bent forward position, and ragged facial expression (5). Many additional symptoms have been listed by various authors including concentration problems, food cravings, physical hyperactivity, sleep problems (6), yawning, asthenia, and somnolence (7). Of interest, a number of symptoms which are part of the International Headache Society diagnostic criteria for migraine (8) have also been reported as occurring during the premonitory phase. These include photophobia and phonophobia (7) and nausea (6). One of the more interesting premonitory symptoms is yawning, which has been reported by a number of authors (6, 7, 9). Additional autonomic symptoms, for example, horripilation (gooseflesh or piloerection) have also been reported as a premonitory symptom (5).
Premonitory symptoms
Patients completed the questionnaire retrospectively with the choice of 12 symptoms. Patients had to indicate the symptom which occurred before at least two-thirds of migraine attacks. bPatients completed a questionnaire prospectively for three attacks with a choice of 28 symptoms. Symptoms also present interictally were not counted as premonitory for individual patients.
Prevalence of premonitory symptoms in patients with migraine
In clinic-based studies, the proportion of patients with migraine that confirm the presence of premonitory symptoms varies widely from study to study, likely reflecting different methodologies used to identify these symptoms. Kelman (4) and Santoro et al. (10) each found that 33% of patients with migraine reported premonitory symptoms, while Schoonman et al. (6) and Quintela et al. (7) found that over 80% reported premonitory symptoms. It would appear, therefore, that a significant proportion of patients with migraine do experience premonitory symptoms, and therefore such symptoms could have therapeutic relevance.
Are premonitory symptoms reliable predictors of the pain phase of the migraine attack?
Whether premonitory symptoms can be reliable predictors of a developing migraine attack in some patients was studied by Giffin et al. (9) in 120 patients with migraine through a prospective electronic diary study. The patient sample in this study was selected according to specific criteria. Only patients who indicated that they experienced premonitory symptoms other than headache, aura, nausea, vomiting, photophobia and phonophobia, and who believed that these predicted headache were included. It was found that, in this selected patient population, patients were able to predict with reasonable accuracy whether the symptom they experienced was likely to be followed by a migraine headache. For example, when patients felt that a migraine attack was very likely to follow, it did so within 72 h 75–95% of the time, and if they felt almost certain, the attack would follow in >95% of the time. Given the migraine frequency in this patient population, a migraine attack would have occurred by chance in only 30% of 72-h time periods.
Overall, when patients recorded a premonitory symptom in their headache diaries, a headache followed within 72 h 72% of the time, which was well above the expected chance level of 30%. Most correct predictions were made within 12 h of headache onset, and many patients were only able to predict the attack 2 h in advance. Nevertheless, from a therapeutic standpoint, it is important to note that in this selected patient population, in 90% of patients in whom a premonitory symptom predicted the attack, the patient had 2 h or more of warning before the pain phase started.
Premonitory symptoms may not all be equally reliable in terms of their ability to predict a migraine attack. Among all premonitory symptoms studied by them, Giffin et al. (9) found that yawning, increased emotionality, difficulty reading, and difficulty speaking were they most reliable predictors.
In an unselected clinic population, Quintela et al. (7) studied how consistently a given premonitory symptom occurred before migraine attacks in individual patients. It was found that in patients who reported concentration difficulties as a premonitory symptom, this symptom occurred consistently before all three of the attacks studied in 53% of patients. Other symptoms were less consistent, however. For patients reporting nausea as a premonitory symptom, this symptom occurred before all three attacks in only 19% of patients. The same authors pointed out that many patients experience similar symptoms to those reported a premonitory symptoms intericatlly, and in their study premonitory symptoms were only accepted as such if that patient did not report them also during the interictal period. For example, 14% of patients reported concentration difficulties interictally. However, as they only accepted premonitory symptoms that occurred in the 24 h prior to the headache attack, it is hard to compare their data with those who accept longer premonitory periods.
In summary, it would appear that a significant number of patients with migraine have premonitory symptoms; many have at least several hours of warning before their headache attack starts, and for an uncertain minority of these patients the premonitory symptoms are reasonably reliable predictors of the oncoming attack. It is of interest that although premonitory symptoms may occur in up to 87% of patients, they are not recognized by most until prompted. As a cautionary note, Kelman (4) found in a clinic population that among the 33% of migraine patients with premonitory symptoms, almost half had less than 1 h of warning before the attack started. Nevertheless, it remains to be seen whether more patients with proper training could predict their headache attacks further in advance.
Treatment during the premonitory phase
There have been few clinical trials where the ability of a medication taken during premonitory symptoms to prevent the expected migraine attack has been tested. Such trials are obviously difficult to do, as patients with premonitory symptoms must first be recruited, the reliability of their premonitory symptoms determined, and then the medication must be tested.
Luciani et al. (11) assessed naratriptan for efficacy when taken in response to premonitory symptoms. They performed an open-label study in 20 patients. To be included, all patients had to have premonitory symptoms which preceded their migraine headache by 4–24 h, and all had to be able to distinguish when these symptoms (fatigue, etc.) were part of a migraine attack prodrome and when not. In the trial, patients recorded diaries for three episodes of premonitory symptoms, and these were followed by headache 100% of the time. The premonitory symptoms recorded during this baseline period included change in energy, change in mood, irritability, food craving, muscle pain/tenderness, sleepiness, change in nausea and change in appetite.
During the treatment phase of the trial, patients were instructed to take naratriptan 2.5 mg during the premonitory phase when they felt headache was inevitable. Each patient treated up to six premonitory phases.
While headache followed the premonitory phase 100% of the time during the baseline phase, during the open-label treatment phase only 40% of the treated premonitory phases were followed by headache. The results also suggested that headache prevention seemed to be more reliable when medication was taken more than 2 h before headache onset, and that those headaches that did occur appeared milder than those during the baseline phase.
Waelkens (12) studied domperidone given during the premonitory phase in a double-blind, placebo-controlled, crossover study. The study population included 19 patients, all with ‘classical’ migraine, who treated 76 attacks during the study. All patients were known to experience warning phenomena (mainly sensory or psychic intolerance) 7–48 (median 24) h before headache onset, and all had severe to incapacitating headache. Medication was taken at the first warning sign and each patient treated four attacks, two with placebo and two with domperidone 30 mg, in random order.
No aura or headache was experienced in 66% of attacks treated with domperidone, as compared to 5% of attacks treated with placebo. There was a statistically significant difference between domperidone and placebo, p < 0001 (Wilcoxon’s matched-pairs signed-ranks test).
In another study, Waelkens (5) examined further the ability of domperidone taken during the premonitory phase to prevent migraine attacks. The study included 19 patients, 18 of whom had migraine with aura, and all of whom had premonitory symptoms which reliably foreboded their migraine attacks. The premonitory symptoms preceded the attacks by 6 h or more in 70% of the cases. The patients took three doses of domperidone (20, 30 or 40 mg) in a blinded fashion and in random order. Each patient treated two attacks with each dose.
Taken at the very first appearance of the early warning symptoms, the 20 mg, 30 mg and 40 mg doses prevented respectively 30%, 58% and 63% of the expected attacks. There was a suggestion that domperidone was more effective when taken early, even 12 h, before attack onset. It did not seem to work nearly as well when taken within 1 h of attack onset.
Waelkens concluded that 30–40 mg of domperidone is required for successful prevention of migraine attacks during the premonitory phase, that headache severity of the remaining attacks is reduced, even by the 20 mg dose, and that timing is crucial: the earlier the medication is taken in the prodrome, the better. From the data presented in the publication, however, it would appear that except perhaps for the 20 mg dose, domperidone was equally effective taken 1–6 h before the attack as compared to earlier time points.
In conclusion, there is evidence that:
Naratriptan can prevent some migraine attacks when taken during the premonitory period (open-label evidence); Domperidone can prevent some migraine attacks when taken during the premonitory period (double-blind, placebo-controlled, crossover evidence); Both drugs appear to work better when taken early (at least 2 h) before headache onset.
Discussion
More research is needed to determine the ability of available drugs to prevent migraine attacks when they are given during the premonitory phase. A better understanding of the pathophysiology of the premonitory phase is also needed in order to guide clinical trial design.
The premonitory phase is probably best considered as part of the migraine attack, although for some premonitory symptoms (for example, anxiety or feeling stressed out) it may be hard to distinguish whether the symptom complained of is a migraine attack trigger or the initial part of the migraine attack itself. If they are part of the migraine attack, it would seem logical to begin symptomatic or acute therapy of the attack when reliable premonitory symptoms first appear. Such a treatment option, if it exists at all, would likely be available to only a minority of migraine sufferers unless careful patient education is able to permit more patients with migraine to accurately identify reliable premonitory symptoms.
Unfortunately, clinical trials with triptans during the aura phase do not lend much support to early treatment of the migraine attack before the onset of the pain phase. Trials with both subcutaneous sumatriptan (13) and oral eletriptan (14) showed treatment responses which were considerably smaller when the medications were given during the aura as compared to those found when these medications were given during the pain phase. Anecdotally, however, personal experience indicates that patients do report success with triptan treatment during the aura, and the eletriptan trials noted above contained small patient numbers.
The premonitory phase might, on the other hand, represent a period of heightened vulnerability for the development of a migraine attack. In this case, clinical trials on the use of the triptans, particularly frovatriptan (15) in the short-term prophylaxis of menstrual migraine might be relevant. Frovatriptan given in a dose of 2.5 mg twice a day starting before migraine onset, during the period of increased migraine vulnerability associated with menstruation in patients with menstrual migraine, significantly reduced migraine attack occurrence (15). Naratriptan and zolmitriptan have shown similar but perhaps lesser degrees of benefit (16, 17). Frovatriptan, or other triptans with a sufficiently long half-life, might be beneficial when taken hours before the onset of the migraine pain phase during early premonitory symptoms.
In addition to research directed at a better understanding of the pathophysiology of premonitory symptoms, it would not be premature to mount clinical trials to study the efficacy of treatment delivered during the premonitory phase. The best candidates for study, based on previous research, would appear to be triptans with a relatively long half-life and dopamine antagonists. The application of behavioral interventions, for example, relaxation exercises, during the premonitory phase could also be considered for study.
While it is might seem counter-intuitive that domperidone, which has poor central nervous system (CNS) penetration (18), could be helpful in migraine treatment during the premonitory phase, it currently does have the best evidence for efficacy as discussed above. It needs to be pointed out, however, that essentially all the patients in the domperidone trials had migraine with aura and whether these results can be generalized to patients with migraine without aura is unknown. It would seem logical, however, to attempt to confirm that domperidone is efficacious when given during the premonitory phase. Also, it might be that a closely related drug with greater CNS penetration, metoclopramide, might be an even better candidate for efficacy.
In summary, more research, including clinical trials, is needed with regard to the pre-emptive treatment of migraine attacks during the premonitory phase. Such trials will need to carefully define their patient population and will need to confirm that in the patients tested the premonitory symptoms do indeed predict the occurrence of the migraine attack with reasonable reliability.
Footnotes
Funding
This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.
Conflict of interest declaration
Dr Becker has served on medical advisory boards for Merck, Allergan, Inc., Pfizer Ltd and AGA Medical. He has received speaker’s honoraria from Teva, Merck Serono S. A., Merck, Allergan and Pfizer Ltd, and research support as part of multicenter clinical trials from Allergan, Medtronic, Inc., AGA Medical and Merck.