Risk and burden of hospital admissions associated with wildfire-related PM2·5 in Brazil, 2000-15: a nationwide time-series study

Abstract

Background: In the context of climate change and deforestation, Brazil is facing more frequent and unprecedented wildfires. Wildfire-related PM_2·5 is associated with multiple adverse health outcomes; however, the magnitude of these associations in the Brazilian context is unclear. We aimed to estimate the association between daily exposure to wildfire-related PM_2·5 and cause-specific hospital admission and attributable health burden in the Brazilian population using a nationwide dataset from 2000 to 2015.

Methods: In this nationwide time-series analysis, data for daily all-cause, cardiovascular, and respiratory hospital admissions were collected through the Brazilian Unified Health System from 1814 municipalities in Brazil between Jan 1, 2000, and Dec 31, 2015. Daily concentrations of wildfire-related PM_2·5 were estimated using the 3D chemical transport model GEOS-Chem at a 2·0° latitude by 2·5° longitude resolution. A time-series analysis was fitted using quasi-Poisson regression to quantify municipality-specific effect estimates, which were then pooled at the regional and national levels using random-effects meta-analyses. Analyses were stratified by sex and ten age groups. The attributable fraction and attributable cases of hospital admissions due to wildfire-related PM_2·5 were also calculated.

Findings: At the national level, a 10 μg/m³ increase in wildfire-related PM_2·5 was associated with a 1·65% (95% CI 1·51-1·80) increase in all-cause hospital admissions, a 5·09% (4·73-5·44) increase in respiratory hospital admissions, and a 1·10% (0·78-1·42) increase in cardiovascular hospital admissions, over 0-1 days after the exposure. The effect estimates for all-cause hospital admission did not vary by sex, but were particularly high in children aged 4 years or younger (4·88% [95% CI 4·47-5·28]), children aged 5-9 years (2·33% [1·77-2·90]), and people aged 80 years and older (3·70% [3·20-4·20]) compared with other age groups. We estimated that 0·53% (95% CI 0·48-0·58) of all-cause hospital admissions were attributable to wildfire-related PM_2·5, corresponding to 35 cases (95% CI 32-38) per 100 000 residents annually. The attributable rate was greatest for municipalities in the north, south, and central-west regions, and lowest in the northeast region. Results were consistent for all-cause and respiratory diseases across regions, but remained inconsistent for cardiovascular diseases.

Interpretation: Short-term exposure to wildfire-related PM_2·5 was associated with increased risks of all-cause, respiratory, and cardiovascular hospital admissions, particularly among children (0-9 years) and older people (≥80 years). Greater attention should be paid to reducing exposure to wildfire smoke, particularly for the most susceptible populations.

Funding: Australian Research Council and Australian National Health and Medical Research Council.