Effects of ambient temperature on pediatric incident seizure: A case-crossover analysis using distributed lag non-linear models

Objective: Emerging evidence supports that brain dysfunction may be attributable to environmental factors. This study aims to examine associations of ambient temperature and temperature variability (TV) with seizure incidence in children, which has not been explored. Material and methods: Data on 2718 outpatient visits due to seizure were collected in Shanghai, China, from 2018 to 2023. Exposure to ambient temperature was estimated at children's residential addresses using spatial-temporal models. A time-stratified case-crossover design with a distributed lag non-linear model (DLNM) was conducted to assess the association between seizure incidence and daily average of ambient temperature over a period of 21 days prior to a case date of disease onset. For a given case date, we selected all dates falling on the same day of the week within the same month as control dates. We calculated a composite index of intra-day and inter-day TV, which was the standard deviation of the daily minimum and maximum temperatures, respectively, over 7 days preceding a case date. We then assessed the association between TV and seizure incidence. Stratified analyses were conducted by age (73.51% < 5 years old and 26.49 % ≥ 5 years old), sex (41.83% female), presence of fever (69.72%), and diagnosis of epilepsy (27.63%). Results: We observed inversed J-shaped temperature-response curves. Lower temperatures had a significant and prolonged effect than higher temperatures. Using 20 °C (with the minimum effect) as the reference, the cumulative odds ratios (ORs) for over 0–21 days preceding the onset at the 5th percentile of the temperature (3 °C) and at the 95th percentile (29 °C) were 3.17 (95% CI: 1.77, 5.68) and 1.54 (95% CI: 0.97, 2.44), respectively. In addition, per 1 °C increases in TV_0–7 was associated with OR of 1.08 (95% CI: 1.01, 1.15). Older children and those experiencing seizure with fever exhibited a higher risk of seizure onset at both lower and higher ambient temperatures. Conclusion: Both low and high temperatures can contribute to the morbidity related to pediatric seizure. Lower temperatures, however, exerted a longer period of effect prior to seizure onset than higher temperatures. An increased risk for incident seizure was significantly associated with temperature variability during preceding 7 days.