Abstract
The prevalence and characterization of premonitory symptoms have not been rigorously studied in children and adolescents. Using a questionnaire, we retrospectively studied the prevalence of 15 predefined premonitory symptoms in a clinic-based population. In 103 children and adolescents fulfilling the International Classification of Headache Disorders, 2nd edn criteria for paediatric migraine, at least one premonitory symptom was reported by 69 (67%). The most frequently reported premonitory symptoms were face changes, fatigue and irritability. The mean number of premonitory symptoms reported per subject was 1.8 (median 2.2). Age, migraine subtype (with or without aura) and mean attack frequency per month had no effect on the mean number of premonitory symptoms reported per subject. In conclusion, premonitory symptoms are frequently reported by children and adolescents with migraine. Face changes seem to be a premonitory symptom peculiar to paediatric migraine.
Introduction
Migraine affects 4–10% of school-aged children (1, 2). It is characterized by periodic, paroxysmal episodes of headache accompanied by nausea, vomiting, abdominal pain and intense desire to sleep. On occasion, dramatic signs such as hemiparesis, acute confusional states, ophthalmoparesis or vertigo may precede or accompany migraine headaches (3). Migraine is characterized by a great variability of phenotypical expressions and it has long been recognized that migraine headaches are frequently associated with a constellation of other non-headache symptoms.
Excluding migraine aura, non-headache symptoms have been associated with the three phases of the attack: the premonitory symptoms, the headache phase and the postdrome. The International Headache Society (IHS) recommends that the terms prodrome and warning symptoms be best avoided because they are often mistakenly used to include aura. In the recently published 2nd edition of the IHS classification (ICHD-II), premonitory symptoms are defined as ‘symptoms preceding and forewarning of a migraine attack by 2–48 h, occurring before the aura in migraine with aura and before the onset of pain in migraine without aura’.
In adults, prevalence rates of patients reporting one or more premonitory symptoms range between 12% (4) and 79% (5). One source of variability in prevalence rate might be differences in study population.
The characterization of premonitory symptoms has been poorly studied in children and adolescents. To our knowledge, no study has been dedicated to the documentation of premonitory symptoms in paediatric migraine. In this study, we assessed the prevalence of 15 premonitory symptoms that have been reported both in adults and paediatric patients using a telephone questionnaire in an unselected clinic-based population of children and adolescents.
Patients and methods
The study population consisted of children and adolescents suffering from migraine according to the IHS criteria. These patients were treated by one of the authors (J-C.C.) in his clinical practice at the neuropaediatric out-patient department of Lille University Hospital Centre and included in a database. This database included all paediatric headache patients seen between March 2003 and May 2008. All evaluations were conducted by telephone. The interviewed person was one of the child/adolescent's parents, preferably the one that usually accompanied the child to the out-patient department, and, as far as possible, the child or adolescent was personally interviewed.
The first part of the questionnaire addressed demographic factors (age, gender) and present migraine characteristics. Migraine-related variables were: migraine subtype (migraine without and/or with aura according to the HIS criteria) and mean attack frequency per month in the last half-year. Eligible subjects were children and adolescents (< 18 years old) who fulfilled the IHS criteria for paediatric migraine with and/or without aura at the time of study, had experienced migraine attacks for at least 1 year before study entry, and had had < 15 headache days per month during last 3 months. The second part of the questionnaire addressed 15 possible premonitory symptoms that were included based on reports in the paediatric and adult literature: fatigue, irritability, nausea, stiff neck/neck pain, yawning, phonophobia, concentration problems, sadness, food craving, anxiety/stressed feeling, hyperactivity, sleep problems, cacosmia, photophobia, face changes. A symptom was considered as premonitory when it preceded the onset of pain, in case of migraine without aura (MoA), or when it preceded the aura, in case of migraine with aura (MA). When aura was present, the patient was asked about the features in order to distinguish premonitory symptoms from aura. A symptom beginning either after aura onset (MA) or after headache onset (MoA) was not considered as premonitory. The duration of the premonitory phase was defined as the 48 h preceding the first symptom of either aura (in case of MA) or headache phase (in case of MoA).
For every possible premonitory symptom patients answered the question ‘how often is a migraine attack preceded by this symptom?’. Answers were categorized as never, rarely (> 0–< 1/3 attacks), often (1/3–2/3 attacks), very often (> 2/3–< 1 attacks) or always. Prevalence of every premonitory symptom was calculated and presented as a percentage. The number of premonitory symptoms per individual was calculated and their repartition according to age, sex, migraine subtype and mean attack frequency per month analysed.
Since patient care was not altered by inclusion in the study, ethics committee approval was not necessary. All analyses were performed using Epi-Info software (6.04 version, CDC, Atlanta, GA, USA). Percentages were rounded to a whole. P-values < 0.05 were considered as statistically significant. Qualitative variables were analysed using χ2 test. Analysis of variance or ANOVA was used for the comparison of prevalence rates.
Results
Data collection was made between 6 and 29 May 2008. Of the 187 subjects with a diagnosis of migraine included in the database, 136 and/or their parent could be contacted by telephone. Of these, 103 actually fulfilled the ICHD-II criteria for paediatric migraine with and/or without aura. All of them agreed to answer the questionnaire. The interviewed person was one of the parents in 70% (n = 72), the child in 8% (n = 8), and both parent and child in 22% (n = 23). The mean age of the subjects was 12.1 years (5–16). Thirty-four patients (33%) reported no premonitory symptom, 69 (67%) reported at least one symptom and 57 (55%) reported two or more (Fig. 1). The characteristics of patients reporting at least one premonitory symptom (n = 69) are shown in Table 1. The mean number of premonitory symptoms reported per subject was 1.8 (median 2.2) and the maximum number of premonitory symptoms reported per subject was seven (Fig. 1). Premonitory symptoms were rarely, often, very often or always associated with migraine attacks in, respectively, 15, 11, 10 and 64% of the 69 subjects who reported at least one premonitory symptom. The most frequently reported premonitory symptoms were: face changes, such as pallor or shadows under the eyes (44%), fatigue (42%) and irritability (24%) (Table 1, Fig. 2). Frequency of premonitory symptoms increased with children's age, but this result was not statistically significant (P = 0.4). Similarly, the effects of gender, migraine subtype and mean monthly attack frequency were not statistically significant (Table 1). Premonitory symptoms were reported in 75% of adolescents aged ≥ 12 years, in 68% of children aged 6–12 years and in 33% of children aged < 6 years. Seventy-two per cent of boys reported premonitory symptoms compared with 65% of girls. Figure 3 shows the prevalence of subjects reporting premonitory symptoms as a function of mean attack frequency per month.
Number of premonitory symptoms per patient. PS, premonitory symptoms.
Prevalence of premonitory symptoms.
Prevalence of premonitory symptoms according to the number of migraine attacks per month.
Migraine and demographic properties of patients reporting at least one premonitory symptom (N = 103)
N, number of patients; OR (CI), odds ratio (confidence interval).
Discussion
The proportion of paediatric migraine patients reporting premonitory symptoms was high: 67% of patients reported at least one symptom. In the absence of another paediatric study pertaining to premonitory symptoms, we can only compare these data with adult studies. This high prevalence rate is comparable to two previous retrospective clinic-based adult studies, where the rates were 79 and 86.9% (5, 6), but in contrast to two other studies, where rates were about 33% (7, 8). In population-based studies rates range from 12% in MoA patients to 18% in MA patients (4). The mean number of premonitory symptoms reported per patient (1.8) is lower than in the study by Schoonman et al., where it was 3.2 (6). One may assume that some of younger children were unable to express a symptom verbally, as already noted by Mortimer et al. (2), who found that prodrome rates were age-dependent and more prevalent in older children; however, there was no effect of age on premonitory symptom frequency in our study. On the other hand, some premonitory symptoms presented by the child were exclusively identifiable by an external observer (such as face changes).
Premonitory symptoms were consistently associated with the migraine attack in 64% of patients, a finding that seems higher than what has been reported in adults. In 460 adult migraine patients attending an out-patient clinic (9), premonitory symptoms preceded migraine attacks in more than 2/3 of events in 46%; in this subgroup premonitory symptoms were followed by an attack in more than 2/3 of cases in ≥ 68% of subjects, which was consistent with other findings reported in adults. In a multicentre, prospective study using a hand-held electronic diary system, premonitory symptoms, when present, were followed by a migraine attack within 72 h on 72% of occasions; in 82% of patients premonitory symptoms were followed by a migraine attack > 50% of the time (10).
The most frequently reported premonitory symptoms were face changes (44%), fatigue (42%) and irritability (24%). Face changes (pallor, shadows under the eyes) seem to be peculiar to children and adolescents, as they have rarely been reported in adult studies. On the other hand, fatigue and irritability have frequently been reported in adult studies, with rates of 72% (10), 46.5% (6) and 25.6% (7) for fatigue, and 23.4% (7) for irritability. By contrast, some premonitory symptoms that were reported in adults, such as mood and behaviour changes, stiff neck, phonophobia, gastrointestinal symptoms and yawning, were rarely reported in our paediatric subjects. There was no difference in the prevalence of premonitory symptoms with gender, contrasting with the study by Schoonman et al. (6), where women reported more premonitory symptoms than men. This difference may be due to the sex ratio, as the majority of patients in adult studies were women, whereas in our study there was a predominance of boys.
Premonitory symptoms may give insight into the pathophysiology of migraine as an episodic dysfunction of trigeminovascular regulation, probably mediated at the level of the brainstem (11, 12). Prodromal symptoms, such as yawning, mood changes, irritability, hyperactivity and sleep disturbances, reflect dopaminergic hypersensitivity, and some of them are mediated by nitric oxide pathways (13). Other symptoms point to a hypothalamic dysfunction. It has been shown that alterations in dopaminergic neurotransmission can modulate clinical susceptibility to migraine and dopamine, at least in a subgroup of migraine patients, and can play an important role in activating the biochemical cascade leading to the premonitory symptoms, and ultimately to the migraine attack (14). Some authors (15) have postulated that many premonitory symptoms might reflect the same biological mechanism underlying the headache phase (disinhibition of brainstem nuclei regulating the amount of pain and other sensory inputs, and their associated motor and autonomic activities). In this setting, the full-blown migraine headache would finally develop when a critical physiological threshold is reached (10).
Premonitory symptoms may offer a unique opportunity for early treatment. In adults, non-randomized trials have suggested that naratriptan (16) and dopamine antagonists (17) may be effective for migraine prevention. Placebo-controlled, randomized trials are lacking to assess whether patient treatment during the premonitory phase prevents migraine. There is also some evidence that the dopamine antagonist domperidone can prevent migraine, but only if taken at least 6 h before the expected attack (18, 19). This question is all the more pertinent in children and adolescents, since migraine attacks are typically shorter in children and adolescents than in adults. However, to our knowledge no study of antimigraine drug during the premonitory phase has been conducted so far in children and adolescents.
This study has some limitations. First, it was a retrospective study. Second, the list of possible premonitory symptoms was based mainly on adult reports of premonitory symptoms and may have missed some symptoms specific to children and adolescents. Third, it was a study of tertiary care patients, and previous studies suggest that premonitory symptoms may reflect patients with more full-blown migraine attacks. However, the response rate was 100%, which eliminates a possible bias that non-responders may have introduced.
Its strong point is the fact that our study was the first paediatric study dedicated to the description of characteristics of premonitory symptoms. It remains that paediatric premonitory symptoms need to be assessed in a prospective design.
In conclusion, our study is the first specifically dedicated to the analysis of premonitory symptoms in paediatric migraine. Premonitory symptoms were frequently reported by children and adolescents with migraine. Among the most frequently reported premonitory symptoms, face changes seemed to be peculiar to children and adolescents, whereas other premonitory symptoms, such as fatigue and irritability, were similar to adults. Sensitivity and specificity of premonitory symptoms for migraine need to be assessed using prospective methods, eventually using electronic diary systems.