Introduction:
The number of emergency department visits for heat-related illnesses, such as hyperthermia, is increasing each summer in mainland France. In order to support prevention policies, a better understanding of these impacts is required. This raises the question of how to characterize heat exposure: using the mean temperature only or by taking into account other meteorological variables.

Methods:
This study compares the influence of four thermal indicators (mean temperature (Tm), wet-bulb globe temperature (WBGT), heat index (HI) and universal thermal climate index (UTCI)) on emergency department visits across 33 geographical departments in France during the summer months (June–September) for the period 2015–2019. Five causes were investigated: hyperthermia, dehydration, hyponatremia, isolated fever and fainting. Analyzes were based on quasi-Poisson regressions.

Results:
Irrespective of the thermal indicator, heat was associated with an excess risk of seeking emergency care for hyperthermia, dehydration, hyponatremia, isolated fever and fainting. Relative risks (RR) were found to be comparable across indicators, except for HI. For instance, regarding hyperthermia, an increase in exposure corresponding to the inter-quartile range is associated with an RR of 3.97 (95% confidence interval, [95%CI]: 3.82–4.13) for Tm, 3.81 [3.67–3.95] for WBGT, 4.22 [4.01–4.33] for UTCI, and 1.22 [1.21–2.23] for HI.

Conclusion:
The results confirm that in mainland France, Tm represents a reliable thermal indicator for characterizing heat-related health issues. The more complex indicators that include other meteorological variables do not explain the effects any better. Studies documenting the circumstances of exposure would provide useful insight into the value of exposure indicators in specific contexts, such as sporting events.