Impact Of Climate Warming On Excessive Mortality In Four Chinese Cities

ObjectivesTo understand the characteristics of the effect modes between temperature and mortality and to estimate the excess risk of hot and cold temperatures on cardiovascular (CVD) and respiratory mortality related to future increases in ambient temperatures in Guangzhou, Shanghai, Beijing and Shenyang, representing for the south, east, north and northeast part of China. To estimate the relationship of daily temperature and years of life lost (YLL) of different cause of deaths in Shenyang, and to calculate the YLL attributed to the temperature increase to assess the burden of death for climate change in the future to provide the scientific evidence for adaptative strategies of climate change.MethodsMortality, meteorological and air pollution data were collected in Guangzhou (2004-2008), Shanghai (2001-2008), Beijing (2007-2008) and Shenyang (2005-2008). Distributed lag nonlinear model was used to examine the nonlinear and delayed effects of temperature on mortality. Excess risk of hot and cold temperatures on non-accidental mortality and cardiovascular and respiratory mortality was then calculated. In Shenyang city, YLL due to the increase of future temperatures was also calculated.Results1. Daily counts of deaths in all four cities displayed a strong seasonal pattern, with the highest counts in winter months (from January to February) and the lowest in summer months (from June to August).2. The association between temperature and mortality was U-shaped in Guangzhou, and J-shaped in Shanghai, Beijing and Shenyang. The minimum mortality temperature (MMT) was lower in high latitudes. The association between mortality and temperature seemed linear when temperatures were above the heat threshold (the upper limit of MMT) or below the cold threshold (the lower limit of MMT). Both MMT and estimates of the associateions varied in different cities and age groups.3.Effect modes between temperature and non-accidental mortality of different age groups were similar in all the four cities. Both hot and cold temperatures increased the risk of non-accidental mortality. Significant heat effects appeared in early lag days along with harvest effect and the largest effect appeared on the current day, while cold effects last longer and the largest effect appeared in lag days 3 to 5. The health effects decreases with day lags in both hot and cold temperatures, and the lag days of health effects varied in different cities.4.For the 0-64 year group, we found the highest cold effect in Beijing, with mortality rate increase of 24.63% (95%CI:18.54%-30.25%) per 1℃ decrease. And in 65 and older people, the cold effects of Shenyang showed the strongest. And per 1℃ decrease was associated with increases of 11.09%(95%CI:9.58%-12.56%) in cold related excess deaths. The cold effects were more evident in 65 and older people than 0-64 people in Guangzhou and Shenyang.5. In these four cities, the baseline of annual average heat related excess mortality was highest in Shenyang and the baseline of annual average cold related excess mortality was highest in Guangzhou. With the increase of temperature, the temperature related excess mortality increased in all cities. And Guangzhou was more sensitive to climate change. The temperature related excess mortality showed the lowest effect in Beijing.6. In these cities, the death of cardiovascular disease and respiratory disease showed the evident distribution and seasonality, which the mortality of was higher in winter and lower in summer. This situation was more evident in Beijing and Shanghai.7. Our results showed that the exposure-response relationships of temperature-daily cardiovascular disease death, temperature-daily respiratory disease death, temperature-daily ischemic heart disease (IHD) death, temperature-daily cerebrovascular disease (CVA) death and temperature- daily chronic obstructive pulmonary disease (COPD) death were generally U-shaped in Guangzhou. The effects of cold and heat were higher at lag 0 day. In Shanghai, Shenyang and Beijing, the exposure-response relationships between daily average temperature and daily CVD and respiratory disease death were J-shaped. During lag 0, the effect of heat was more evident than cold.8.The effect modes between daily average temperature and CVD death during these cities was familiar, which showed the strongest heat effect in same day. The characteristic of cold effect was smaller in the same day and climbed to the highest effect during preceding 2-4 days. That is the effect of heat was acute and the cold was lag effect.9. In these four cities, the cold and heat effect threshold of daily average temperature related disease death in southern cities was higher than in northern cities. In each city, the MMT for different cause of deaths was different.10. The effect mode between daily average temperature and respiratory disease death was still showed that the effect of heat was acute with evident harvesting in Guangzhou, Beijing, and Shenyang. The effect of high temperature on COPD lasted longer which lagged up to 11 days with harvesting effects.11. Effects of cold on the CVD death were the strongest in the northern city, and were the smallest in the southern city. When the temperature dropped 1℃, the CVD deaths in Shenyang increased 12.58%, in Beijing increased 7.29%; Among four cities, daily risk of cold effect for IHD deaths were the highest in Beijing, and that of hot effect was highest in Guangzhou.12. Among four cities, temperature related CVD mortality rate was highest in Shenyang (69.3 per 100 thousand people), and lowest in Guanghzou (37.7 per 100 thousand people), which mainly showed hot effect.13. Temperature related moratilty rate for respiratory diseases was highest in Beijing, and lowest in Shenyang. The temperature effects on CVA displayed mainly hot effects in Shenyang, Shanghai, and Guangzhou, and the cold effect was highest in Beijing. The temperature related mortality rate for COPD was highest in Beijing and Shanghai, which was 15.3 and 11.7 per 100 thousand people respectively. The temperature effects displayed hot effect in Beijing, and cold effect in Shanghai.14. With future temperature increase, the change in temperature-related CVD mortality and respiratory disease mortality was different. The excessive CVD mortality rate and respiratory mortality rate decreased when temperature increased in Beijing. With strong heat effect in Guangzhou and Shenyang, temperature-related excessive mortality rate for CVD and respiratory disease increase with future temperature increase.15. The curve for daily mean temperature and YLL for men and respiratory disease was V-shaped, and that for female, deaths from CVD, deaths from IHD and from CVA were J-shaped. Cold effect was strong when temperature below the MMT, and with temperature increase above the MMT, the hot effect decreased, which may be due to the high mortality in elderly people.16. The effect of daily temperature on different gender and people with different disease was not same. The effect mode was similar in males and females, with acute and strong effect for high temperature at lag 0, then decreased. The effect for cold was less than heat, and always with long lags, which was very little at lag 0, and got the highest in 5d then decreased. Hot effect on respiratory death related YLL was acute and lasted for 2d.The cold effect with highest at lag 0, and lasted longer. The hot effect on IHD was same to that on CVD. The cold effect was low at lag 0, and got to the highest in 0-3days, then decreased and lasted longer. The curve of temperature and CVA-related YLL was similar to that for CVD.17. The effects of temperature on male and female displayed different characteristics. Cold effect was related to the "younger age death" in males, while it was mainly related to elderly death in females and CVA patients. Heat related deaths were mainly in aged people, especially with highest risk in aged females or in eldly IHD patients.18. Temperature effects on male and female were mainly hot effect, and the cold effect was higher for males than for females. When the temperature increase 1-3℃, the increases of YLL related to hot effect was offset by cold effect. The net temperature-related would decrease. When the temperature increased 4℃, the temperature-related YLL would increase by 113 man-years per year. With high vulnerable to heat, the YLL for female would increase 464 man-year per year.19. Cold effect on IHD was stronger than heat effect. With temperature increase, temperature-related YLL induced by IHD mortality decreased. The heat effect on CVA was especially high, and the heat effect-related YLL by CVA was 7934 man-year per year, with the proportion 92.74%. The burden of deaths for CVA increased with temperature increasedConclusion1. Daily mean temperature was associated with daily non-accidental mortality. The shape of temperature-mortality curve was different.2. The MMT was different in different cities, which increased with latitude decreased.3.The elderly people aged 65+ years were vulnerable to temperature. The female elderly people were more vulnerable to cold and hot effect than the male.4. The temperature effect on mortality always lasted for several days, the lags for cold and hot effect was different.5.The effect for hot effect was acute, and lasted for 1-2d with harvesting. Cold effect lasted longer up to 30d.6. Chinese cities were sensitive to climate warming. Southern cities were affected more.7. Daily mean temperature was associated with daily deaths from CVD. With high mortality risk on extreme temperature. The risks, effect modes and lags of temperature were different in different cities.8. Daily mean temperature was associated with daily deaths from respiratory disease. The effect was different in south and north cities.9. The daily mean temperature was associated with YLL, with YLL increasing at lower and higher temperatures. The temperature effects in female were mainly on the 65+ aged people. The cold effect in male was mainly on 0-64 aged people.10. With the effects of temperature different in people with different age, sex, and disease, regional adaptative strategies should be made considering the temperature effect modes, aging population composition, and disease spectrum.