Abstract Rationale Extreme heat and cold are risk factors for severe asthma exacerbations. Whether daytime or nighttime exposures are more relevant is unclear. We assessed associations between extreme temperatures and asthma admissions, using census-tract-specific temperature thresholds in a population of adults in the San Francisco Bay Area. Methods Adults aged 18 years and older who were hospitalized for asthma at a large urban hospital (University of California, San Francisco) between 2012-2024 were included. We identified individual-level hospitalizations for asthma using primary encounter ICD-10 codes and aggregated counts at the census tract level. Census-tract-level temperature was estimated from population-weighted centroids of high-resolution gridded surface meteorological data. Daily maximum and minimum temperatures were used to identify “extreme” thresholds ( 95th percentile, 5th percentile) for each census tract separately by year. A time-stratified case-crossover design was used to control for time-invariant confounders. Quasi-Poisson regression with 7-day distributed lag structure was used to estimate associations between temperature and asthma admission risk. A distributed lag structure allowed for evaluation of independent temperature-related risk at specific timepoints. Models were adjusted for daily relative humidity and stratified by county-level coastal proximity to examine coastal climate moderation effects. As an exploratory analysis, daytime extreme heat models were adjusted for nighttime temperatures (and vice versa) to assess the independent effects of heat at specific times of day. Results We identified 2,084 unique asthma admissions over 1,722 census tracts. The 95th percentile for maximum temperature was 33.6 °C (IQR 30.4-36.2) and the 5th percentile for minimum temperature was 2.4 °C (IQR 0.9-4.2) across census tracts. Extreme nighttime heat on the day of admission was associated with significantly reduced hospitalization risk (RR = 0.77, 95% CI 0.62-0.97, p=.03) (Figure). Extreme daytime cold three days before admission was associated with increased risk (RR = 1.39, 95% CI 1.06-1.83, p=.02). Effects were more pronounced for inland counties. Extreme daytime heat was not significantly associated with risk after controlling for nighttime temperatures in exploratory analyses. Exposure to extremely hot nights on lag days 1 and 2 was cumulatively associated with increased risk after controlling for maximum daytime temperatures (RR = 1.39, 95% CI 1.08-1.79, p=.01). Conclusions Delayed temperature effects in specific temporal and spatial patterns were identified for asthma hospitalizations in urban adults. Extreme nightime heat on the day of admission was negatively associated with admission risk, whereas extreme daytime cold on lag day 3 was positively associated with admission risk. Exploratory analysis suggested nighttime heat confers delayed independent risk for hospitalization. This abstract is funded by: None
Beagle et al. (Fri,) studied this question.