Chinook winds and migraine attack onset in children and adolescents: A prospective longitudinal clinical cohort study.

objectiveTo explore the relationship between migraine attack onset in children and adolescents and Chinooks, which are dry and warm westerly winds that generally occur in the winter and bring about abrupt weather changes to the east of the Rocky Mountains in Southern Alberta, Canada.

methodsThis was a prospective longitudinal clinical cohort study with recruitment from November 2020 to May 2024. Participants were: 8-18 years old, had migraine as per International Classification of Headache Disorders 3rd edition criteria, had 1-15 headache days/month, lived in the geographical location where Chinook winds occur, and had exposure to at least one pre-Chinook or Chinook day during the study period. Chinook days were defined using Nkemdirim's criteria and Environment Canada data were used to categorize day type as either Chinook, pre-Chinook, or non-Chinook. Weather data were merged with data from daily headache diaries, completed for periods of 8-30 days. The primary outcome was attack onset, defined as a day with a new migraine attack of moderate or severe severity, as per the 4-point scale (0 = none, 1 = mild, 2 = moderate, and 3 = severe). Both univariate and adjusted models were used to determine if there was an association between migraine attack onset and day type (i.e., pre-Chinook, Chinook, or non-Chinook) at the aggregate study sample level. The adjusted models controlled for age and sex, and both models included a random intercept. Subsequently, individual n = 1 models were fitted to explore each individual participant's personal odds of migraine attack onset on both pre-Chinook and Chinook days versus non-Chinook days. Pre-Chinook/Chinook sensitivity values were calculated for each individual by dividing the model's regression coefficient by its standard error. Sensitivity values >1.96 suggest a significant association between pre-Chinook/Chinook days and attack onset.

resultsSixty youth with 1253 days of complete data, of which 144 (12%) were attack onset days, participated in the study. There were 158 Chinook (13%), 124 pre-Chinook (10%), and 971 non-Chinook days (77%). There were 39 female participants (39 of 60; 65%), with a median age of 14 years (quartile [Q] 1 = 12, Q3 = 16), and a median headache frequency of 6.2 days/month (Q1 = 4, Q3 = 11). Neither the univariate nor the adjusted models found any significant association between day type and attack onset at an aggregate level (pre-Chinook adjusted odds ratio [OR], 0.98; 95% confidence interval [CI], 0.54-1.78, p = 0.947; Chinook adjusted OR, 1.15; 95% CI, 0.69-1.91, p = 0.596). No individual participants met the threshold for statistically significant pre-Chinook or Chinook sensitivity.

conclusionWe did not find a relationship between pre-Chinook and Chinook conditions and migraine attack onset. This may be due to the lack of an association between Chinooks and attack onset in youth with migraine, or due to a lack of statistical power in our study. Future studies with greater statistical power should aim to assess for a potential relationship between Chinooks and attack onset, as it could have important treatment implications.