How air pollution affects cause-specific emergency hospital admissions on days with biomass combustion in Spain?

Although wildfires are growing in number, intensity and extent due to climate change, few studies have been undertaken to analyse their health impact, and fewer still to analyse the impact of variables other than particulate matter. The objective of this study is to analyse the short-term impact of NO, PM, PM, O and temperature in heat waves on emergency hospital admissions in Spain on days when biomass combustion occurs. We conducted an ecological longitudinal time series study across the period 2013-2018. The dependent variables were emergency hospital admissions due to all causes (ICD-10: A00-R99), circulatory causes (ICD-10: I00-I99) and respiratory causes (ICD-10: J00-J99) in 9 representative Spanish provinces. The independent variables were daily mean air pollution concentrations measured at a provincial level. We also included daily maximum temperatures recorded at reference observatories in the respective provinces. Poisson generalised linear models were fitted for days with and without PM advections due to biomass burning smoke. We controlled for trends, seasonalities, the autoregressive nature of the series, Sundays and Public Holidays. Days with biomass burning smoke advections increased PM concentrations in all provinces and ozone in many of them, something that did not occur in equal measure in the case of NO. Nevertheless, the principal impact on admissions was due to O, followed by NO, PM, and lastly, heat-wave temperatures. This pattern was observed for all three causes analysed. The role of PM was relegated to that of a third factor, and the role of temperature to an order of lesser magnitude than that of chemical pollution. To focus the health impacts of wildfires exclusively on the impact of PM concentrations would be to minimise their real impact on population health. It is therefore essential to implement integrated plans that take into account the joint effect of all atmospheric variables affected.