Very few researches have concentrated on a variety of time scales to evaluate the association between temperature variation (TV) and childhood asthma (CA), and the evidence for the interaction of air pollutants on this association is lacking. In this study, we aim to estimate the relative risks (RRs) of CA due to TV by following metrics: diurnal temperature range (DTR), temperature changes between neighboring days (TCN), and temperature variability (TV0-t); to quantify the seasonal attributable fraction (AF) and number (AN) of CA due to TV; to examine the interactive effects of the TV and air pollutants on CA in different seasons. We mainly applied distributed lagged nonlinear model (DLNM) and conditional Poisson models to evaluate the associations between TV and outpatient visits for CA during 2014-2019 in Lanzhou, China. Additionally, the bivariate response surface model was used to examine the interplay effect of air pollutants. We found that in warm season, the risks of DTR maximum at lag5 (RR = 1.073, 95% CI: 1.017-1.133); TCN showed protective effect. In cold season, the risks of DTR peaked at lag8 (RR = 1.063, 95% CI: 1.027-1.100); the risks of TCN maximum at lag0 (RR = 1.058 95% CI: 1.009-1.109); the estimation of total cases maximized at TV0-4 in cold season (RR = 1.039 at TV0-3, 95% CI: 1.001, 1.077) and was the lowest at TV0-1 in warm season (RR = 0.999, 95% CI: 0.969, 1.030). In addition, the response surface model graphically pictured ambient air pollutants enhanced the DTR/TV0-4-CA effect for girls. In conclusion, the RRs of CA are markedly increased by TV exposure, particularly during the colder months. A combined evaluation of DTR, TCN, TV0-5∼TV0-6, NO2, SO2, and PM2.5 should be used to identify the adverse effects of TV on CA.
Keywords: Ambient temperature variation; Attributable risk; Childhood asthma; Interaction; Seasonality.
Copyright © 2022 Elsevier Inc. All rights reserved.