Since the heatwave of 2003, France has assessed excess mortality attributable to exposure to heat every summer. Traumatized by the 15,000 additional deaths recorded that summer, following a heatwave episode exceptional in its duration and intensity, the Ministry of Health had instructed Inserm (National Institute of Health and Medical Research) to establish a precise report in order to measure the extent of this excess mortality and to rethink health surveillance systems.
Since then, each summer period, from June 1 to September 15, has been closely monitored, as has each heatwave episode, whether it occurs during or outside this period. This monitoring is carried out by epidemiologists from Public Health France, who analyze mortality data in real time and publish, at the end of each summer, an estimate of the number of deaths attributable to heat. But how is this excess mortality linked to rising temperatures calculated?
This calculation carried out by Public Health France is not based on the causes of death recorded on the death certificates. If the latter sometimes mention “heat stroke”, or “dehydration”, they more often point out other health problems, which the heat may have aggravated. It is then indicated: “heart attack”, “heart failure”, “stroke”, without any link being made with exposure to high temperatures.
During the 2003 heatwave, analyzes showed that a significant part of the excess mortality was recorded due to cardiovascular, respiratory, neurological or poorly defined causes. At the same time, an increase in mentions of “heat stroke” or “hyperthermia” had been observed. Although these certificates therefore constitute an interesting source of information, they remain insufficient to quantify deaths caused by extreme heat.
To estimate the excess mortality attributable to the rise in mercury, Public Health France therefore relies on a statistical approach: by comparing the total number of deaths observed during the summer period – or during a heatwave episode – to the number of deaths expected in the absence of abnormal variations in the thermometer. This calculation is carried out using the EuroMoMo (European Mortality Monitoring) method, used in 25 European countries to monitor episodes of excess mortality (during epidemics such as influenza, Covid or during heatwaves, etc.).
Distinguishing “excess” heat-related deaths
Once this number of additional deaths has been established, it remains to evaluate those which can be attributed to heat and not to other causes (age pyramid, epidemic, etc.). To do this, epidemiologists then use statistical models to estimate the part specifically linked to high temperatures. They correlate the number of people who died, the temperatures they experienced, and how mortality typically varies as temperatures rise.
Thus, during the summer of 2025, 5,700 additional deaths were attributed to heat. This does not mean that 5,700 people died from “heatstroke”, but that if temperatures had remained at a normal level, 5,000 of these deaths would probably not have occurred at that time.
Public Health France also specifically studies each heatwave period. The study of excess mortality linked to heatwaves (prolonged extreme heat) is carried out in two stages. Since 2023, epidemiologists have published a first assessment of excess mortality two weeks after the end of the heatwave. This waiting time is necessary to receive at least 90% of death notifications. The result is a first estimate of excess mortality during this heatwave, all causes observed, that is to say without detailing deaths attributable to the heat.
To have this precision, we must wait until the following autumn, after the end of the monitoring season, for the complete and consolidated assessment of excess summer mortality to appear, which includes additional deaths attributable to heat and details the effects of heatwaves.
According to Public Health France, over the last nine summers (2016-2025), 40,000 additional deaths were recorded during summer surveillance periods, of which 11,700 were specifically attributable to periods of heatwaves. According to epidemiologists, each year, heat represents 1 to 4% of summer mortality and 7 to 12% of mortality during heatwaves, orders of magnitude which have so far remained stable since 2017.
