The prevalence of triggers in paediatric migraine: a questionnaire study in 102 children and adolescents

Few papers have been dedicated to the study of TF for childhood migraine [2‐4]. All children and adolescents in our study had at least one TF. In the literature, between 75 and 89% of adult migraineurs have at least one TF [5‐8]. In the study by Chakravarty et al. [3], one or more triggers could be detected in 94% of children when studied retrospectively and 100% of them when studied prospectively. The mean number of TF reported per patient [7] in our study was comparable to the adult studies of Kelman (6.7) and Theeler (8.9) [5, 8].

The most common individual trigger was stress which was reported by 75.5% of patients, a figure close to that was found by Chakravarty et al. (78%) [3]. We can but agree with Carod-Artal et al. [9] when they wondered that stress was a TF in different societies, in spite of cultural differences between them. The study by Chakravarty et al. [3], which seems to be the only study specifically devoted to paediatric migraine TF as a whole, available so far, was conducted in India. Climatic, ethnic, dietary and socio-cultural factors are very different in India from those in our country. In adult studies, stress is experienced as a major precipitant of migraine and is reported as a TF by 59–79% of patients [5‐7]. Of note, stress always triggered an attack in 24.5% of patients. We did not distinguish between occurrence during stress and occurrence after stress, adult studies having shown that the former is more often associated with the triggering of an attack than the latter [10]. One may also suspect a confusion bias between lifestyle changes and stress. For example, the start of the new school year may account as a TF as both a lifestyle change and a stress per se. In the Thai study by Visudtibhan et al. [11], some patients had preferentially attacks during the first days of holidays. This may be due to lowering of school pressure or to relaxation after school stress.

Lack of sleep was reported by 69.6% of patients and was generally put aside with fatigue. Late onset of sleep was reported as an occasional TF by 32% of patients in the series of Kelman, whereas 44 to 57% of adult migraineurs report lack of sleep as a TF [5, 6]. Excess of sleep was cited by 21.6% of children and adolescents. Bruni et al. [12] have shown that enhancing sleep quality resulted in the reduction of migraine and tension type headache frequency.

Hot and cold weather were reported by 68.6 and 15.7% of patients, respectively. In adults, climate is cited by 53.2–71% of patients [5, 13]. Hot weather was the main TF triggering an attack “very often” or “often” in our study. In the paediatric study by Chakravarty et al. [3], hot humid weather was a TF in 94% of patients. This study was conducted in Eastern Indian, whose climate is very different from North of France, which is generally mild and rainy. Moreover, there are few days with hot weather during a typical year in North of France. Mukamal et al. [14] have shown that a transient rise of ambient temperature increased the risk of headache requiring emergency department evaluation, with approximate 7.5% higher risk for each 5°C increment in temperature. In a paediatric prospective study using electronic momentary assessment methodology, relative humidity and presence of precipitation were significantly predictive of new headache onset [4]. By contrast, no child reported humidity and rainy days as a TF in our study.

Bright lights were reported by 52.9% of patients, a figure similar to adult literature data (29–61%) [3, 5, 6, 13, 15, 16]. We cannot exclude a possible confusion between warmth, sun exposure and bright lights. In fact, it was difficult during interview to precise what the TF was during hot weather, as patients were concomitantly exposed to these three TF and one cannot exclude that several of TF might summate to trigger the headache.

Dietary precipitants have been much stressed in western literature but were rarely reported by our patients [2, 15]. Chocolate was reported by 11.8% and cheese by 3.9%. This difference between theoretical knowledge and personal experience has been studied in adults by Wöber et al. [17]. They found that, whereas chocolate and cheese were suspected by more than half of adult patients, only 15% of them reported them as real TF. On the other hand, studies on food as a trigger of migraine attacks have given insights on biochemical mechanisms involved in the genesis of the attack. Chocolate contains phenylalanine and cheese tyramine; both have vasoconstrictive properties and may initiate a headache by alteration of cerebral blood flow and release of norepinephrine from sympathetic nerve cells [2].

Triggers are not universal. Moreover, the presence of a TF does not always precipitate an attack in the same individual. Lambert and Zagami [18] hypothesised that most triggers would excite cortical neurons, leading to an inhibitition in the periaqueductal grey matter and the nucleus raphe magnus. These two brainstem nuclei would then release their ongoing descending control of dura matter sensation in the trigeminal nucleus and cause normal sensory traffic from the dura to be perceived as migraine pain. Other hypotheses argue that TF may induce the onset of cortical spreading depression in a preexisting hyper-excitable cortex of a migraine brain, initiating the process of pain generation [19]. Other authors have speculated that common migraine triggers such as stress, skipping a meal, and fatigue may promote the headache through the activation of distinct but converging neuronal pathways that would originate in several brain nuclei that respond to the most common migraine triggers. These include the lateral hypothalamus and perifornical areas that become activated during food and sleep deprivation, and the bed nucleus of stria terminalis and paraventricular hypothalamic nucleus that are involved in regulating stress response. The neuronal outputs from these nuclei converge onto the superior salivatory nucleus, which further projects to the parasympathetic sphenopalatine ganglion [20]. Indirect activation of the sphenopalatine ganglion by triggering factors would result in local meningeal release of parasympathetic-driven vasoactive and algesic mediators such as nitric oxide and acetylcholine that in turn could enhance directly or indirectly the responsiveness of meningeal nociceptors [21].

In the present study 70.6% of patients reporting trigger factors indicated at least one trigger factor that very often or always precipitated an attack of migraine. Suggesting that avoidance of these specific factors could lead to a substantial reduction of attacks is a classical tenet of paediatric migraine management, but has been recently challenged by Martin, that advised instead that clinicians may “need to approach managing headache triggers with more flexibility than simply counselling avoidance of all triggers—think ‘coping with triggers’.” [22].

Limitations of the study are the relatively small number of subjects, the tertiary care recruitment, the retrospective design and the fact that the interviews were performed by phone. In adults, it has been hypothesised that patients with headache report more stress on retrospective self-reports than do headache-free controls because of a tendency to selectively recall more stressful events on retrospective measures [23]. Thus one may consider the possibility of confounded results, due to selective memory biases in the recollection of stressful events. To our knowledge, no paediatric headache study has addressed this issue so far. There might also be an overlap between premonitory symptoms and trigger factors in migraine. It could be that mental stress trigger a migraine attack or that patients perceive more mental stress because they are in the premonitory phase of a migraine attack. This seems very unlikely as premonitory symptoms and trigger factors are very different in children, but difficulties might arise for stress and sleep problems which were reported as premonitory symptoms in respectively 13 and 2% of paediatric migraineurs in a previous study [24].

Strengths of this study are the strict headache diagnosis according to the ICHD-II criteria and the differentiation between consistent and occasional trigger factors. The study also gives data regarding the influence of gender, migraine subtype, or attack frequency on the distribution of triggers.

In conclusion, our study indicates that all children and adolescents with migraine reported at least one trigger factor and that it very often or always precipitated an attack of migraine in 70.6% of them. The most common individual trigger was stress. Of note, mean time elapsed between exposure to a trigger factor and attack onset was inferior to 3 h in most patients. These preliminary data on trigger factors for paediatric migraine need to be assessed using prospective methods, eventually using electronic diary systems, but studying trigger factors for paediatric migraine as a whole appears cumbersome and, to our knowledge, has not even done so far in adult studies.