Analysis Of Short-term Effects Of Atmospheric Ozone On Cardiovascular Disease By Using Bayesian Spatio Temporal Model

BackgroundCardiovascular disease（CVD）is currently the leading cause of death in China.In recent years,many epidemiological studies have found that the increase in atmospheric ozone（O₃）concentration can increase the number of deaths from cardiovascular diseases in residents.At present,most studies on the relationship between O₃ and cardiovascular diseases were based on time series analysis or spatial regression analysis methods,few studies considered the effects of time and space at the same time,which may cause bias in study results.ObjectiveApplying the Spatio-temporal Bayesian model to analyze the exposure response relationship between O₃ concentration and the risk of death of cardiovascular diseases,determining the main modifiers of O₃ concentration and the risk of death of cardiovascular disease and its temporal and spatial distribution pattern,and providing information and methodology for the prevention and control of cardiovascular diseases.Method1.Collecting the daily mortality data of 40 districts（counties）in Guangdong Province from 2014 to 2017,and calculating the daily mortality of cardiovascular diseases;Obtaining the studied districts Meteorological data by the interpolation method of meteorological space based on the thin plate spline function;Predicting the air pollutant data of the studied area（county）through the construction of random forest and iterative random forest model.2.Applying The distance-based method to construct the spatial weight matrix,the generalized additive model is established as the benchmark model,and the global Moran’I is used to analyze the spatial correlation of the residuals of the model.3.Constructing Spatio-temporal Bayesian models which included time,space,Spatio- temporal interaction terms and covariates,using DIC criterion（Deviation information criterion）to determine the optimal model,and applying the optimal model to fit the exposure response relationship of O₃ concentration and the risk of death of cardiovascular diseases.4.Analyze the interaction effects of time,space and time and space based on the model to explore the Spatio-temporal distribution of risk of death of cardiovascular disease.5.A stratified analysis method was used to explore the modifiers of the influence of atmospheric ozone concentration on the risk of death of cardiovascular diseases in different populations.Results1.General characteristics:A total of 243,254 deaths from cardiovascular diseases were included in this study,and the average daily death rate from cardiovascular diseases was 0.69 per 100,000 people.The daily average temperature,average relative humidity, average daily concentration of PM_2.5,and 8-hour maximum average concentration of O₃ in the study districts and counties were 22.1,79.2%,36.1μg/m³,and 89.6μg/m³, respectively.The iterative random forest model assesses the O₃ exposure level with R² of 0.78 and RMES of 13.2μg/m³.The benchmark model found that there was a significant positive correlation between O₃ concentration and the risk of cardiovascular disease death.The Moran’I of the model residuals was positive or negative,and results of the model residuals showed that the study districts and counties are spatially correlated.2.Spatio-temporal distribution:the influence of cardiovascular disease death risk has time, space,and temporal and spatial interactions.The average relative risk of death of cardiovascular diseases increased year by year during the study period.The seasonal alternation of each year is a period of high death of cardiovascular caused by O₃;the spatial effect results show that the areas with a higher risk of death are mainly located in Wuhua County.The results showed that there were 2 hot spots（Wuhua County,Shixing County）and 3 cold spots（Luoding City,Lianzhou City and Liannan Yao Autonomous County）..From January to December,the mortality rate of cardiovascular disease caused by O₃ has a dynamic process from convergence to divergence and then convergence.O₃ has a synergistic effect with meteorological factors and other pollutants.3.Exposure response relationship:each 10μg/m³ increase in O₃ concentration（lag04）leads to a 0.44%increase in the risk of death of cardiovascular diseases in the total population （95%CI:0.34%～0.47%）.4.Modifiers:The result of a stratified analysis of gender shows that for every 10μg/m³ increase in O₃（lag03）,the risk of death of cardiovascular disease in men increases by 0.29%（95%CI:0.13%～0.43%）,and in women.The risk of death of cardiovascular disease increased by 0.21%（95%CI:0.11%～0.40%）.The results of age stratification analysis show that for every 10μg/m³ increase in O₃（lag3）,the risk of death of cardiovascular disease in people aged 0-64 increases by 0.10%（95%CI:-0.20%～0.34%）,≥65 years old.The risk of death of cardiovascular diseases in the elderly is increased by 0.31%（95%CI:0.17%%～0.47%）.The results of seasonal stratification analysis showed that for every 10μg/m³ increase in O₃（lag03）,the risk of death of cardiovascular disease in the cold season increased by 0.37%（95%CI:0.17%～0.49%）,and the cardiovascular death risk of the total population in the warm season was increased by 0.37%（95%CI: 0.17%～0.49%）.The risk of death of disease increased by 0.12%（95%CI:- 0.10%%～0.30%）.The results of regional stratification analysis showed that for every 10μg/m³ increase in O₃（lag03）,the risk of death of cardiovascular disease in the Pearl River Delta population increased by 0.25%（95%CI:0.11%～0.44%）.The risk of death of cardiovascular disease increased by 0.27%（95%CI:0.02%%～0.53%）in the non-Pearl River Delta population.ConclusionThe iterative random forest model can improve the accuracy of assessing the exposure level of ambient O₃.Based on the time-space Bayesian model,it is determined that short-term ozone exposure is significantly positively correlated with the risk of death from cardiovascular diseases,and its impact is lagging and cumulative.Among them,short-term ozone exposure has greater impact on females than males,people aged≥65 are more susceptible,cold season people are more sensitive than warm season people,and people in non-PRD regions are more sensitive than people in Pearl River Delta regions.This study combines the environmental characteristics of various places to formulate targeted prevention and control strategies for cardiovascular diseases,which has scientific practical significance.