Abstract/Summary

While previous research on historical changes in heat-related mortality observed decreasing trends over the recent decades, future projections suggest increasing impact of heat on mortality in most regions of the world. This study aimed to analyse temporal changes in temperature-mortality relationships in Prague, Czech Republic in the warm season (May–September), using a daily mortality time series from 1982 to 2019. To investigate possible effect of adaptation to increasing temperature, we divided the study period into four decades (1980s–2010s). We used conditional Poisson regression models to identify decade-specific relative risk of heat-related mortality and to calculate the annual number of heat-attributable deaths and the heat-attributable fraction of total warm season deaths. We estimated their trends over the whole study period by a generalized additive model with non-parametric smoothing spline. Our results showed that the unprecedentedly hot 2010s was associated with approximately twice as large relative risk of heat-related mortality than in previous decades. This resulted in the reversal of the trend in heat-attributable mortality in the 1990s and its increase during the last two decades. Our findings highlight the importance of further improvement of adaptation measures such as heat-and-health warning systems to protect the heat-susceptible population.