Nonlinear exposure-response associations of daytime, nighttime, and day-night compound heatwaves with mortality amid climate change

Abstract

Heatwaves are commonly simplified as binary variables in epidemiological studies, limiting the understanding of heatwave-mortality associations. Here we conduct a multi-country study across 28 East Asian cities that employed the Cumulative Excess Heatwave Index (CEHWI), which represents excess heat accumulation during heatwaves, to explore the potentially nonlinear associations of daytime-only, nighttime-only, and day-night compound heatwaves with mortality from 1981 to 2010. Populations exhibited high adaptability to daytime-only and nighttime-only heatwaves, with non-accidental mortality risks increasing only at higher CEHWI levels (75th-90th percentiles). In contrast, compound heatwaves posed a super-linear increase in mortality risks after the 25th percentile of CEHWI. Associations of heatwaves with cardiovascular mortality mirrored those with non-accidental mortality but were more pronounced at higher CEHWI levels, while significant associations with respiratory mortality emerged at low-to-moderate CEHWI levels. These results highlight the necessity of considering the nonlinear health responses to heatwaves of different types in disease burden assessments and heatwave-health warning systems amid climate change.

Fig. 1. Pooled exposure-response relationship curves of the cumulative excess heatwave index (CEHWI) for daytime, nighttime, and day-night compound heatwaves in relation to non-accidental mortality, along with the corresponding CEHWI distributions.

a East Asia; b, South Korea; c, Japan; d, China. The solid lines represent the estimated cumulative relative risks (point estimates) of mortality on heatwave days compared to non-heatwave days, while the shaded areas denote the corresponding 95% confidence intervals. The vertical dotted lines, from left to right, indicate the 25th, 75th, and 90th percentiles of the regional CEHWI distribution during heatwaves. Source data are provided as a Source Data file.

Fig. 2. Pooled exposure-response relationship curves of the cumulative excess heatwave index (CEHWI) for daytime, nighttime, and day-night compound heatwaves in relation with cardiovascular disease mortality, along with the corresponding CEHWI distributions.

a East Asia; b South Korea; c Japan; d China. The solid lines represent the estimated cumulative relative risks (point estimates) of mortality on heatwave days compared to non-heatwave days, while the shaded areas denote the corresponding 95% confidence intervals. The vertical dotted lines, from left to right, indicate the 25th, 75th, and 90th percentiles of the regional CEHWI distribution during heatwaves. Source data are provided as a Source Data file.

Fig. 3. Pooled exposure-response relationship curves of the cumulative excess heatwave index (CEHWI) for daytime, nighttime, and day-night compound heatwaves in relation with respiratory disease mortality, along with the corresponding CEHWI distributions.

a East Asia; b South Korea; c Japan; d China. The solid lines represent the estimated cumulative relative risks (point estimates) of mortality on heatwave days compared to non-heatwave days, while the shaded areas denote the corresponding 95% confidence intervals. The vertical dotted lines, from left to right, indicate the 25th, 75th, and 90th percentiles of the regional CEHWI distribution during heatwaves. Source data are provided as a Source Data file.

Fig. 4. Lag structures for the associations of daytime, nighttime, and day-night compound heatwaves with non-accidental, cardiovascular disease, and respiratory disease mortality in East Asia.

a Non-accidental death; b Cardiovascular disease; c Respiratory disease. The solid lines represent the estimated relative risks (point estimates) of mortality for each lag day associated with specific heatwave types compared to non-heatwave days, while the shaded areas indicate the corresponding 95% confidence intervals. Source data are provided as a Source Data file.

Fig. 5. Cumulative relative risks (RRs) and attributable fractions (AFs) of non-accidental, cardiovascular disease, and respiratory disease mortality associated with daytime, nighttime, and day-night compound heatwaves in East Asia.

RRs and AFs are depicted as squares (centres of the error bars), with the error bars representing the 95% confidence intervals. RRs were defined as the risks at the mean of the 90th percentile of cumulative excess heatwave index distributions compared with the risks on non-heatwave days. Source data are provided as a Source Data file. Abbreviation: eCIs empirical confidence intervals.