Effect of temperature change between neighboring days on years of life lost and attribution in Yunnan Province, 2013-2017

Background In the context of global climate change, the health effect of temperature change has been a focus.

Objective This study is designed to investigate the impact of temperature change between neighboring days (TCN) on the years of life lost (YLL) of residents in Yunnan Province, and quantitatively assess the proportion of YLL attributed to TCN.

Methods We collected the daily non-accidental death records and meteorological data of 56 districts or counties in Yunnan Province from 2013 to 2017, and calculated the YLL rate (YLL per 10⁵ inhabitants) based on the life table. We defined April through September as warm season and the other months as cold season, then we established exposure-response relationships between TCN and YLL respectively in the two seasons and calculated attributable fraction (AF) by using two-stage analysis method combining a distributed lag nonlinear model and a metaanalysis.

Results The medians of TCN in both cold and warm seasons were 0.1℃. Compared with TCN=0℃, in warm season, for every 1℃ decrease of TCN, the YLL rate was decreased by 2.00/10⁵ (95% CI:0.28/10⁵-8.71/10⁵); temperature increase between neighboring days (≤ 1.4℃) increased the YLL rate, and when TCN=1.4℃, the YLL rate was increased by 2.15/10⁵ (95% CI:0.04/10⁵-4.26/10⁵). In cold season, temperature decrease between neighboring days (≤ 9.2℃) reduced the YLL rate, and when TCN=9.2℃, the YLL rate was decreased by 8.78/10⁵ (95% CI:0.64/10⁵-16.45/10⁵). The elderly (≥ 65 years old) were more likely to be affected by TCN than the young (0-64 years old). For every 1℃ change of TCN, the YLL rates of the young and elderly residents changed 1.18/10⁵ and 5.18/10⁵ in warm season on average respectively, while 0.79/10⁵ and 8.82/10⁵ in cold season respectively. TCN decrease between neighboring days in warm season had a greater impact on males than on females; however, when the TCN decreased more than 7.61℃ in cold season, females were more susceptible. Temperature decrease between neighboring days gave rise to the YLL decrease attributable to TCN, and temperature increase between neighboring days caused the YLL increase attributable to TCN. Moderate cooling (TCN:P_2.5-0℃) had the greatest effect on AF (warm season, AF=-5.76%, 95%CI:-7.10%--4.49%; cold season, AF=-3.98%, 95%CI:-5.83%--2.07%).

Conclusion Temperature increase between neighboring days has an adverse impact on the years of life lost in selected residents.