Temperature-mortality associations by age and cause: a multi-country multi-city study

Abstract

Studies have consistently identified differences in temperature-mortality relationships by age and cause of death. However, evidence is sparse, mostly coming from investigations with limited geographical scope and using disparate epidemiological designs. Here, we assess this issue using data from 537 cities in 33 countries from the Multi-Country Multi-City (MCC) Collaborative Research Network.We apply a flexible two-stage design. In the first stage, we estimated age and cause-specific temperature-mortality associations by location using distributed lag non-linear models. In the second stage, for each cause, we pooled location-specific estimates in novel dose-response multivariate meta-regressions, allowing modelling continuous age-varying risks from different age grouping across locations.Pooled heat-related relative risks (RR) for 99th and 1st percentile versus minimum mortality temperature (MMT), were 1.14 (95%CI:1.09-1.20) for cardiovascular, 1.27 (95%CI:1.16-1.38) for respiratory, and 1.11 (95%CI:1.07-1.15) for other causes. Corresponding RR for cold (1st percentile versus MMT) were 1.39 (95%CI:1.32, 1.46), 1.33 (95%CI:1.24-1.43), and 1.17 (95%CI:1.13-1.20), respectively. Age-specific risks showed differential patterns by cause. Predicted heat-related RR for 25 and 75-year-olds increased from 1.06 to 1.11 for cardiovascular, but they were stable across age groups for respiratory and other causes. The picture was different for cold, with strong age-trends for cardiovascular (1.13 to 1.40) and other causes (1.02 to 1.16), while no variation by age was found for respiratory mortality. There is evidence of higher heat-related risks in high versus low income countries, especially at young age.This study found important differences in temperature-related mortality by age and cause across the globe. Associations were higher for respiratory causes, compared to cardiovascular, but present for other causes as well. Age differences were stronger for cold-related mortality, with older people more at risk in particular for cardiovascular causes, while respiratory mortality showed no age patterns for either heat or cold.On behalf of the MCC Collaborative Research Network