Background: Variables that are thought to precipitate migraine or tension-type headache episodes in children hitherto have only been studied using retrospective reports. As such, there is little empirical evidence to support the actual predictive association between presumed headache triggers and actual headache occurrence in children.
Objective: The present study sought to determine if fluctuations in weather, a commonly reported headache trigger in children, predict increased likelihood of headache occurrence when evaluated using rigorous prospective methodology ("electronic momentary assessment").
Methods: Twenty-five children (21 girls, 4 boys) between the ages of 8-17 years attending a new patient neurology clinic appointment and having a diagnosis of chronic migraine, chronic tension-type, or episodic migraine headache (with or without aura) participated in the study. Children completed baseline measures on headache characteristics, presumed headache triggers, and mood and subsequently were trained in the use of electronic diaries to record information on headaches. Children then completed thrice daily diaries on handheld computers for a 2-week time period (42 assessments per child) while data on weather variables (temperature, dew point temperature, barometric pressure, humidity, precipitation, and sunlight) in the child's geographic location were recorded each time a diary was completed. Data were analyzed using multilevel models.
Results: Of the weather variables, relative humidity and presence of precipitation were significantly predictive of new headache onset, with nearly a 3-fold increase in probability of headache occurrence during times of precipitation or elevated humidity in the child's area, b = 0.38, t(821) = 2.10, P = .04, and b = 0.02, t(821) = 2.81, P = .01, respectively. These associations remained after accounting for fluctuations in mood, and associations were not significantly stronger in children who at baseline thought that weather was a headache trigger for them. Changes in temperature, dew point temperature, barometric pressure, and sunlight were not significantly predictive of new headache episode occurrence in this sample.
Conclusions: Results of the present study lend some support to the belief commonly held by children with recurrent headaches that weather changes may contribute to headache onset. Although electronic momentary assessment methodology was found to be feasible in this population and to have the potential to identify specific headache triggers for children, it remains to be determined how best (or even whether) to incorporate this information into treatment recommendations.