Temporal variation in the association between temperature and cause-specific mortality in 15 German cities

Abstract

Background: There is limited evidence of temporal changes in the association between air temperature and the risk of cause-specific cardiovascular [CVD] and respiratory [RD] mortality.

Method: We explored temporal variations in the association between short-term exposures to air temperature and non-accidental and cause-specific CVD and RD mortality in the 15 largest German cities over 24 years (1993-2016) using time-stratified time series analysis. We applied location-specific confounder-adjusted Poisson regression with distributed lag non-linear models with a lag period of 14 days to estimate the temperature-mortality associations. We then pooled the estimates by a multivariate meta-analytical model. We analysed the whole study period and the periods 1993-2004 and 2005-16, separately. We also carried out age- and sex-stratified analysis. Cold and heat effects are reported as relative risk [RR] at the 1st and the 99th temperature percentile, relative to the 25th and the 75th percentile, respectively.

Result: We analysed a total of 3,159,292 non-accidental, 1,063,198 CVD and 183,027 RD deaths. Cold-related RR for CVD mortality was seen to rise consistently over time from 1.04 (95% confidence interval [95% CI] 1.02, 1.06) in the period 1993-2004 to 1.10 (95% CI 1.09, 1.11) in the period 2005-16. A similar increase in cold-related RR was also observed for RD mortality with risk increasing from 0.99 (95% CI 0.96, 1.03) to 1.07 (95% CI 1.03, 1.10). Cold-related ischemic, cerebrovascular, and heart failure mortality risk were seen to be increasing over time. Similarly, COPD, the commonly speculated driver of heat-related RD mortality was found to have a constant heat-related risk over time. Males were increasingly vulnerable to cold with time for all causes of death. Females showed increasing sensitivity to cold for CVD mortality. Our results indicated a significant increased cold and heat vulnerability of the youngest age-groups (<64) to non-accidental and RD mortality, respectively. Similarly, the older age group (>65) were found to have significantly increased susceptibility to cold for CVD mortality.

Conclusion: We found evidence of rising population susceptibility to both heat- and cold-related CVD and RD mortality risk from 1993 to 2016. Climate change mitigation and targeted adaptation strategies might help to reduce the number of temperature-related deaths in the future.

Keywords: Cardiovascular mortality; Cause-specific mortality; Germany; Respiratory mortality; Temperature; Temporal variation.