Deadly heat: The impacts of extreme heat on mortality, attributable risks and thresholds

Previous studies have reported significant associations between heat and mortality. However, there are knowledge gaps in understanding the exposure-health thresholds, the most sensitive indicators, and demographic variabilities.

We assessed the association between mortality and extreme heat, its intensity/duration and identified the exposure-mortality threshold, the attributable risks, and demographic disparities.

We applied a time-stratified case-crossover design to investigate extreme heat-mortality associations in New York State (NYS) while controlling PM_2.5, O₃, relative humidity, and holiday using conditional logistic regression. Non-accidental mortality and mortality across six subtypes (circulatory, respiratory, hematologic, digestive, infectious, renal) from vital records (2001-2021) were used. We used four definitions to define extreme heat (i.e., >= 90th, 92.5th, 95th, and 97.5th percentile of daily average temperature during warm season), exposure days, and computed risk ratios (RRs), attributable fraction/death number.

We found significant associations between extreme heat and non-accidental mortality (RRs ranged from 1.010 to 1.049) on most lag days, and the risks increased with intensity and duration of heat exposure. The increased risks were consistently observed across different systems except for infectious and renal diseases. The risks were also higher in some demographic groups, but not statistically different. The temperature threshold for significantly increased mortality was 19.016°C. Hematologic mortality has the highest attributable fraction (30.02%), and circulatory mortality had the highest attributable death numbers (26,355 deaths).

Extreme heat was positively associated with increased risks of total mortality and most subtypes with significant exposure-response relationship. While the highest attributable risk is from hematologic mortality, circulatory mortality bears the heaviest burden.