Daily particulate matter and temperature from satellite-hybrid models and 1.5 million deaths: A time-stratified case-crossover analysis in Central Mexico

Abstract

BACKGROUND AND AIM: Highly spatio-temporally resolved (daily 1x1-Km) fine particulate matter (PM₂.₅) and air temperature (Ta) models are being used to advance exposure science and refine health effect estimates in developed countries. Limited evidence exists about their use in densely populated areas from middle-income regions, and different exposure conditions. METHODS: We use a time-stratified case-crossover study design, matching cases to control days on day of week, month and year, for 1,516,442 non-external deaths (≥18 years-old) from 586 sub-county areas in Central Mexico (including Mexico City) from 2004-2019. Daily 1x1-Km PM₂.₅ (median=23.4μm³; IQR=17–30.6μm³) and Ta estimates from our new satellite-based models for the region were employed for exposure assessment at the sub-county level using population-weighted centroids. Acute associations of PM₂.₅ and Ta with mortality were estimated with distributed lag conditional logistic models. RESULTS:Mortality was significantly associated with 10μm³ PM₂.₅ increase with stronger effects for lag0 [1.44% (95% CI:1.37–1.50%)] and lag1 [0.67% (0.60–0.75%)], with a 1-week cumulative effect (lag06) of 3.52% (3.41–3.63%). The minimum mortality temperature (MMT) was estimated at 18.4°C (18.3–18.4°C), and used as reference to estimate effects from cold (p5=12°C) and heat (p95=20.6°C) from the daily Ta distribution. Heat effects peaked at lag0 [4.64% (4.40–4.89%)] and persisted for 2-days with a cumulative lag02 effect of 7.07% (6.83–7.32%). CONCLUSIONS:We performed the largest case-crossover study of PM₂.₅ and Ta with mortality in Mexico. Our linked datasets utilizing highly spatio-temporally resolved exposure models will enable inference for developing new concentration-response functions in the Central Mexico region. Additional analyses will investigate effect modification by individual characteristics (e.g., occupational classes, sex, and age) and heterogeneity of associations for specific causes of death. KEYWORDS: Mortality, Particulate matter, Temperature, Short-term exposure