Time Series Analysis Of PM_2.5 Impacts On Years Of Life Lost In Chengdu Cardiovascular Disease Patients By DLNM

Background:Distributed lag nonlinear model(DLNM)was developed in recent years for the estimation of cumulative lag effects of air pollution.Crossbasis is the tool of DLNM to deal with the expose-effects and the expose-lag effects;Special geography and specific high humidity made Chengdu unusual in environmental health studies.The health effects of PM2 5 exposures in such high humidity senario for fragile populations is in urgent need to be explored.Objective:This study aims to evaluate the short-term effects of PM2 5 exposures on years of life lost(YLL)of cardiovascular diseases(CVD)in Chengdu populations by the distributed lag nonlinear model(DLNM).Methods:We used the time series analysis to study the short-term effects of PM2.5 exposures on YLL of CVD patients aged 40 years and over in Chengdu with the death statistics from 2013 to 2016 by DLNM.We estimated the YLL by the life-table method recommended by the World Health Organization(WHO).In addition,we checked the effects of indoor-outdoor penetration parameter weighted PM2 5 on the YLL of CVD patients by DLNM also.Results:The main results are as follows:i.The lag effect of PM2 5 on the YLL of CVD:After controlling for the trends of time,day of the week(DOW),holiday,ambient air temperature,and the relative humidity in DLNM,time series analysis showed that short waves of PM2 5 did not have significant lag effects on all type CVD YLL or CVD YLL in all.Short waves of PM2 5 had significant lag effects on YLL of IHD only.Lag2 daily mean YLL of IHD was 0.26(95%CI:0.03-0.50)years for a 10 ?g/m3 increase of PM2 5.The DLNM fitting results showed that short waves of PM2.5 had significant cumulative lag effects on YLL of CVD and IHD(P<0.05),Lag19 daily mean cumulative YLL of CVD was 1.46(80%CI:0.27?2.52)years for a 10 ?g/m3 increase of PM2.5;And Lagl6 daily mean cumulative YLL of CVD was 0.65(80%CI:0.13-1.09)years for a 10 ?g/m3 increase of PM2.5 As for the YLL of cerebrovascular disease and the YLL of hypertension were not found significantly associated with PM2.5 exposures(P>0.05).ii.Circulatory system disease YLL in different PM2 5 concentration scenarios:Compared to the reference exposure level of PM2 5(10?g/m3),DLNM fitting results showed that PM2.5 attibuted CVD YLL was significantly different by the exposure level of PM2 5(P<0.05).As for the lowere level of PM2 5(<37.6?g/m3),only the cumulative YLL of hypertension was significantly associated with PM2 5 level.YLL of CVD in all,YLL of cerebrovascular disease,and YLL of IHD were not significantly affected by PM2.5 in the scale of lower PM2 5(P>0.05).The impacts of lower PM2 5 on hypertension lasted for two days,and the PM2 5 attributed cumulative YLL of hypertension was 1183(95%CI:56-2085)years.As for the higher level of PM2.5(PM2.5>89.9?g/m3),CVD YLL in all and the YLL of IHD were found significantly associated with PM2 5 exposures(P<0.05),while the YLL of cerebrovascular diseases and the YLL of hypertension were not significantly associated with PM2.5 exposures(P>0.05).Higher level of PM2 5 attributed IHD cumulative YLL was 137(95%CI:-69?344)years.?.The YLL of PM2 5 in different temperature scenario:After controlling for the time trends,DOW,holiday,ambient air temperature,and relative humidity,the DLNM fitting results showed that the effects of PM2 5 on YLL of CVD was significantly confounded by temperature(P<0.05).As for the cold weather scenario,cumulative YLL of CVD was not significantly associated with PM2 5 exposures.As for the hot weather,the YLL of IHD was significantly associated with PM2.5 exposures.The PM2 5 effects on cumulative YLL of IHD lasted for 15 days,and the PM2.5 attributed YLL was 2765(95%CI:617-4231)years.iv.Confounding of relative humidity on PM2 5 effects on YLL of CVD:After controlling for the trends of time,DOW,holiday,ambient air temperature,and relative humidity,DLNM fitting results showed that the PM2.5 attributable YLL of CVD was significantly confounded by the effects of humidity(P<0.05).Stratified by humidity,DLNM fitting results showed that the effects of PM2 5 on YLL of CVD lasted for 11 days in the weather with higher humidity(>86%).And the biggest total YLL of CVD was 8.25(95%CI:0.41-20.17)years for every 10 ?g/m3 increase of PM2.5.DLNM fitting results showed that the interaction of air temperature,humidity,and PM2 5 was significantly associated with YLL of CVD.The YLL of CVD for the PM2 5 exposures in the weather of hot and extremely humid(temperature>22?,humidity>86%),was 17.6 folds of that in the weather of hot and dry.No significant effects of PM2 5 on YLL of CVD was found in cold and dry weather or in cold and extremely humid weather(P>0.05).v.Confounding of ozone(O3)on PM2 5 effects on YLL of CVD:After controlling for the trends of time,DOW,holiday,ambient air temperature,and relative humudity,DLNM fitting results showed that the PM2 5 attributable YLL of CVD was significantly confounded by the effects of ozone(P<0.05).The fitting result with the interaction item of PM2 5*ozone by DLNM showed that the YLL of CVD in all decreased 14.99%(19.88years,95CI%:12.86-25.37years),the YLL of cerebrovascular disease decreased 8.48%(5.14years,95CI%:0.09-8.84years),the YLL of IHD decreased 17.34%(8.62years,95CI%:5.25-10.88years),and the YLL of hypertention decreased 17.28%(3.37years,95CI%:1.35-4.45years).vi.Difference in CVD YLL of exposures of outdoor ambient PM2 5(to be shortened as outdoor PM2 5)and indoor infiltrated ambient air PM2 5(to be shortened as indoor infiltrated PM2 5):DLNM fitting results showed that relative risks of CVD YLL for the exposures of indoor infiltrated PM2 5 outdoor PM2 5 were varied by the scale of temperature after controlling for the time of open window,trends of time,ambient air temperature,relative humidity,DOW,and holiday.As for the cold weather,the YLL of CVD due to the exposure of indoor infiltrated PM2.5(YLL=10.53,95%CI:-11.18-36.12 years)was 6.3 folds of that of outdoor PM2.5 exposure(YLL=1.67,95%CI:-1.91-5.34 years).As for the hot weather,the YLL of CVD due to the exposure of indoor infiltrated PM2.5(YLL=57.09,95%CI:16.73-177.95 years)was 6.6 folds of that of outdoor PM2.5 exposure(YLL=8.64,95%CI:-3.05-21.37 years).Conclusion:The results of time-series analysis by DLNM suggested that short wave of PM2.5 might contribute to the cumulative lag effects on YLL of CVD.And the cumulative effects might be confounded by the exposures of ozone and ambient air temperature.