Age standardized heat and cold related mortality impacts across European cities

Background and aim. Heat and cold are important environmental risk factors for mortality with apparent geographical differences in vulnerability. Factors such as the local climate, topology, demography or the socio-economic situation can all affect the mortality burden of heat and cold. This study aims to comprehensively assess of the heat- and cold-related mortality in European cities, including less represented areas such as Scandinavia and Eastern Europe while accounting for the different demographic and vulnerability factors. Methods. We performed a mortality impact assessment related to heat and cold in 801 European cities through a three-stage analysis. First, in a selection of 191 cities, we assessed the exposure-response functions of temperature on mortality for different age groups. Second, these functions were pooled in a meta-analysis model with meta-predictors, including a regional background, age, and 22 vulnerability factors, reduced into a handful of composite indices. Third, this model was used to infer the exposure-response function to the whole 801 city sample, and to compute age-standardized excess mortality rates. Results. Across all examined European cities, we estimated an overall annual age-standardize excess mortality of 116 deaths (95%CI: 104-132) per 100,000 persons due to cold and 15 deaths (95%CI: 13-17) per 100,000 persons due to heat. The burden widely varied across Europe, being lower in western countries and higher in Eastern Europe for both heat and cold. At the same time, southern countries also experienced an important heat-related burden. Conclusions. This study provides a high-resolution assessment of the heat and cold mortality burden across Europe. Results indicate important geographical differences with high vulnerability in the easternmost locations of the continent. These results might give valuable input for designing adaptation policies across Europe. Keywords. Temperature; Mortality; Europe; Meta-analysis; Time series