Effect of different heat wave timing on cardiovascular and respiratory mortality in France

Previous epidemiological evidence suggests that the impact of heat waves on mortality may change over time within the summer season. The consideration of heat wave timing could help to optimize the implementation of heat alert systems. We explored the effect of the timing of extreme heat events on mortality risk during the summer season in France. Summertime daily mortality data for 21 French cities from 2000 to 2015 were obtained from the French National Institute of Health and Medical Research. Heat waves were defined according to the official definition of Météo France. The order of heat wave over time, from June to August, was assessed. We also used ambient temperature and considered different summer periods. To quantify mortality risk (for cardiovascular and respiratory causes) for the first and second or later heat waves, quasi-Poisson models were performed. We used distributed lag non-linear models to estimate whether the non-linear exposure–response associations between temperature and mortality differ across different summer periods. Compared with non-heat wave days, the second and later heat waves of the summer season were associated with a higher relative risk (RR) for cardiovascular and respiratory mortality (RR, 95%CI: 1.38, 1.23–1.53; RR, 95%CI: 1.74, 1.45–2.08, respectively) as compared to first heat wave (RR, 95%CI: 1.30, 1.17–1.45, RR, 95%CI: 1.56, 1.33–1.83, respectively). Small increase from the median temperature was associated to an increased risk in mortality in the first stage of the summer (from June to mid-July), while only more extreme temperatures were harmful later in the summer. After the exclusion from the analysis of the August 2003 heat-wave, only results for earlier heat waves episodes and first-period exposures were confirmed. The timing of extreme temperatures modulates heat-related risks in France. Such information could be used to update local heat action plans to optimize health benefits.