Method and system for identifying key toxic components of mixed exposure of VOCs (volatile organic compounds) components in atmosphere

Abstract

The invention relates to a method and a system for identifying key toxic components of atmospheric VOCs component mixed exposure. The method comprises the following steps: S1, collecting daily cardiovascular disease emergency admission person times and daily average concentration of atmospheric VOCs components in a certain place within at least one year; s2, grouping the monitored VOCs components according to a chemical structure; s3, constructing a correlation between the VOCs group and the number of admission persons of the cardiovascular disease emergency treatment, and screening out key VOCs components which have great influence on the health outcome; wherein the association between the VOCs groups and cardiovascular disease emergency admission persons is constructed through the WQS model, and the WQS model is as follows: # imgabs0 # imgabs1 # g (u) = beta0 + beta1WQS + z 'phi. According to the method, the association between the mixed exposure of the VOCs components of different chemical structure groups and cardiovascular disease emergency admission is analyzed by using the WQS model based on the actual condition of mixed exposure, so that the combined effect of the components in each VOCs group is researched, and the correlation between the mixed exposure of the VOCs components of different chemical structure groups and the cardiovascular disease emergency admission is also researched; components playing key roles are also screened out.

Description

Technical Field

[0001] The present disclosure relates to the technical field of health risk assessment, and in particular to a method and system for identifying key toxic components of mixed exposure to atmospheric VOCs components. Background Art

[0002] Volatile organic compounds (VOCs) are a class of widely existing compounds. Fuel combustion, traffic emissions, the use of industrial solvents and other activities closely related to people's production and life can produce VOCs, and the diversity of sources also makes the components of atmospheric VOCs diverse. According to the International Agency for Cancer, VOCs components such as benzene, 1,3-butadiene, and formaldehyde are Class I carcinogens to humans. In animal experiments, the carcinogenic hazards and health hazards of high exposure concentrations of VOCs components have been widely studied. In addition, occupational exposure or indoor exposure studies have also shown that exposure to VOCs components affects the health of the body's organs or systems. Therefore, VOCs are one of the main air pollutants that affect the health of the population.

[0003] However, most epidemiological studies focus on exposure to high-concentration VOCs components (400ppb-2000ppb). Nowadays, the application of online monitoring and analysis technology for atmospheric VOCs can achieve fine-scale VOCs component exposure concentration monitoring with high temporal resolution over a continuous period of time. This technology has been applied in many cities or regions. Compared with the high-concentration VOCs exposure in previous studies, the atmospheric VOCs exposure concentrations in these cities or regions are mostly lower than 300ppb. However, there are still few epidemiological studies on the relationship between exposure to low-concentration VOCs components in the atmosphere and health risks to the population.

[0004] Due to the limitations of the analytical methods for the health effects of mixed exposures in the past, the analytical methods used in the few literatures on epidemiological studies on atmospheric VOCs and human health have certain deficiencies. Tsai used a generalized additive model (GAM) to determine that propane, isobutane, and benzene can increase the risk of cardiovascular death, but this method of analyzing the association between a single atmospheric VOCs component and health outcomes does not conform to the actual situation of simultaneous exposure to multiple components. Another study also used the GAM model to analyze the association between the sum of the daily concentrations of 54 atmospheric VOCs components and the number of cardiopulmonary hospitalizations. The results showed that atmospheric VOCs exposure was associated with an increased risk of cardiopulmonary hospitalizations, but the method of summing the concentrations of multiple VOCs components and analyzing health outcomes not only did not consider the interactions between components, but also failed to identify the components that play a key role in the joint effect.

[0005] Therefore, there is currently a lack of a method that can both capture the joint effects of simultaneous exposure to multiple components in real situations and select the components that play a key role. Summary of the invention

[0006] Based on this, the purpose of the present disclosure is to provide a method that can not only analyze the acute combined effects of multi-component mixed exposure of atmospheric VOCs on human health, but also screen out the components that play a key role.

[0007] To achieve the above objectives, the present disclosure adopts the following technical solutions:

[0008] A method for identifying key toxic components of mixed exposure of atmospheric VOCs components comprises the following steps:

[0009] S1. Collect the number of emergency hospital admissions for cardiovascular diseases and the daily average concentrations of atmospheric VOCs components in a certain place for at least one year;

[0010] S2. grouping the monitored VOCs components according to chemical structure to obtain different VOCs groups;

[0011] S3. Establish the association between the VOCs group and the number of emergency admissions for cardiovascular diseases, and screen out the key VOCs components that have a greater impact on health outcomes;

[0012] The association between the VOCs group and the number of emergency admissions for cardiovascular diseases was constructed by the WQS model, and the WQS model was as follows:

[0013]

[0014] g(u)=β0+β1WQS+z'Φ

[0015] Among them, g(u) represents the number of emergency hospitalizations for cardiovascular diseases in the selected population on study day u, β0 is the intercept, z' and Φ represent the covariate matrix and covariate coefficient, c represents the number of components included in each VOCs group, W i represents the weighted coefficient of the VOCs component. The sum of the weighted coefficients of the components in each VOCs group is equal to 1. The weighted coefficient of a single component ranges from 0 to 1. If the weighted coefficient of a single component is > 1 / c, the component is considered to be a key component. β1 is the regression coefficient of the WQS index. i Represents the score of the components in each VOCs group.

[0016] Preferably, in step S1, the number of emergency admissions for cardiovascular diseases of people who have lived in the selected place for less than 6 months is excluded.

[0017] Preferably, in step S1, it also includes collecting the daily average concentration value of the selected atmospheric pollutants and daily meteorological data, the atmospheric pollutants include atmospheric fine particulate matter, ozone and nitrogen dioxide, and the meteorological data include daily average temperature and daily relative humidity data.

[0018] More preferably, it also includes constructing a correlation between the key VOCs components and the number of emergency admissions for cardiovascular diseases, so as to obtain an exposure-response relationship between the short-term exposure of the key VOCs components and the number of emergency admissions for cardiovascular diseases;

[0019] The GLM model is used to construct the association between the key VOCs components and the number of emergency admissions for cardiovascular diseases. The GLM model is as follows:

[0020] Y t =Intercept+β1VOC t +ns(time,df)+ns(tem t ,df)+ns(rh t ,df)+dow+holiday

[0021] Among them, Y t represents the number of emergency admissions for cardiovascular disease on observation day t; VOC t represents the concentration of the key VOCs component on observation day t; β1 represents the exposure-response relationship coefficient between the key VOCs component and the number of emergency admissions for cardiovascular diseases; ns(time, df) is the use of natural spline function to control the time trend; ns(tem t , df) and ns(rh t , df) is the natural spline function used to control meteorological factors; dow and holiday are dummy variables used to control the day of the week effect and holiday effect.

[0022] Preferably, a sensitivity analysis is also included:

[0023] At least two monitoring stations were set up to collect daily average concentrations of atmospheric VOCs components to observe exposure measurement errors;

[0024] The stability of the WQS model was evaluated by changing the temporal degrees of freedom and meteorological degrees of freedom settings in the natural spline.

[0025] More preferably, the sensitivity analysis further includes performing WQS model analysis on the air pollutants and the VOCs respectively to examine the effect of the air pollutants on VOCs; wherein the association between the VOCs group and the number of emergency admissions for cardiovascular diseases is constructed by the WQS model, and the atmospheric pollutants are controlled in the model, and the WQS model is as follows:

[0026]

[0027] g(u)=β0+β poll Poll+β1WQS+z'Φ

[0028] Among them, g(u) represents the number of emergency hospitalizations due to cardiovascular diseases in the selected population on study day u, β0 is the intercept, z' and Φ represent the covariate matrix and covariate coefficient, c represents the number of components included in each VOCs group, W i represents the weighted coefficient of the VOCs component. The sum of the weighted coefficients of the components in each VOCs group is equal to 1. The weighted coefficient of a single component ranges from 0 to 1. If the weighted coefficient of a single component is > 1 / c, the component is considered to be a key component. β1 is the regression coefficient of the WQS index. i represents the score of the components in each VOCs group, β poll is the coefficient of the atmospheric pollutant, and poll is the level of the atmospheric pollutant.

[0029] Preferably, hysteresis analysis is also included:

[0030] The percentage increase in the risk of emergency hospitalization for cardiovascular disease in the population from the day of exposure to the different VOCs groups to the 7th day was calculated, including the calculation of single-day lag effects and moving average lag effects.

[0031] Preferably, the VOCs group includes normal alkanes, alkenes, iso / trans alkanes, aromatic hydrocarbons, aldehydes, ketones, ethers, halogenated aromatic hydrocarbons, halogenated saturated chain hydrocarbons and halogenated unsaturated chain hydrocarbons.

[0032] To achieve the above objectives, the present disclosure also adopts the following technical solutions:

[0033] A system for identifying key toxic components of mixed exposure of atmospheric VOCs components is provided, the system being used to implement the method according to any of the above-mentioned items, comprising a data acquisition module, a grouping module and an analysis module.

[0034] The data collection module collects the number of emergency hospital admissions for cardiovascular diseases and the daily average concentration of atmospheric VOCs components in a certain place for at least one year;

[0035] The grouping module groups the monitored VOCs components according to chemical structures to obtain different VOCs groups;

[0036] The analysis module constructs the association between the VOCs group and the number of emergency admissions for cardiovascular diseases through the WQS model, and screens out key VOCs components that have a greater impact on health outcomes;

[0037] The WQS model formula is as follows:

[0038]

[0039] g(u)=β0+β1WQS+z'Φ

[0040] Among them, g(u) represents the number of emergency hospitalizations due to cardiovascular diseases in the selected population on study day u, β0 is the intercept, z' and Φ represent the covariate matrix and covariate coefficient, c represents the number of components included in each VOCs group, W i represents the weighted coefficient of the VOCs component. The sum of the weighted coefficients of the components in each VOCs group is equal to 1. The weighted coefficient of a single component ranges from 0 to 1. If the weighted coefficient of a single component is > 1 / c, the component is considered to be a key component. β1 is the regression coefficient of the WQS index. i Represents the score of the components in each VOCs group.

[0041] Preferably, it also includes an auxiliary analysis module, which constructs the association between the key VOCs components and the number of emergency admissions for cardiovascular diseases through a GLM model to obtain the exposure-response relationship between the short-term exposure of the key VOCs components and the number of emergency admissions for cardiovascular diseases;

[0042] The GLM model is as follows:

[0043] Y t =Intercept+β1VOC t +ns(time,df)+ns(tem t ,df)+ns(rh t , df)+dow+holiday, where Y t represents the number of emergency admissions for cardiovascular disease on observation day t; VOC t represents the concentration of the key VOCs component on observation day t; β1 represents the exposure-response relationship coefficient between the key VOCs component and the number of emergency admissions for cardiovascular diseases; ns(time, df) is the use of natural spline function to control the time trend; ns(tem t , df) and ns(rh t , df) is the natural spline function used to control meteorological factors; dow and holiday are dummy variables used to control the day of the week effect and holiday effect.

[0044] The technical solution claimed in the present disclosure has achieved the following beneficial effects:

[0045] 1) Based on the actual situation of mixed exposure, this paper uses the WQS model to analyze the association between mixed exposure of VOCs components of different chemical structure groups and emergency hospitalization for cardiovascular disease. It not only studies the joint effects of the components in each VOCs group, but also screens out the components that play a key role.

[0046] 2) The GLM model was used to establish a model of key components and emergency admissions for cardiovascular diseases, and the exposure-response relationship effect value was obtained to further verify the results of the WQS model and make the results more stable and reliable. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the technical solutions in the embodiments of the present disclosure, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are merely embodiments of the present disclosure, and a person of ordinary skill in the art can obtain other drawings based on the provided drawings without paying any creative work.

[0048] Figure 1 Module diagram of the key toxic component identification system for mixed exposure of atmospheric VOCs components.

[0049] Figure 2 Technology roadmap for WQS and GLM model analysis.

[0050] Figure 3 This is a descriptive analysis of atmospheric VOCs exposure in a certain place from 2015 to 2016 (the violin plot shows the average daily total concentration of each VOCs component group, the box, the central red dot and the vertical lines at both ends represent the quartile range, median and 95% CI of the data, respectively, and the curve area represents the distribution of concentration data; the heat map shows the average daily concentration of VOCs components).

[0051] Figure 4 is the Spearman correlation coefficient of each group of VOCs components.

[0052] Figure 5 The combined effect of mixed exposure to each group of VOCs components on emergency hospitalization for cardiovascular disease.

[0053] Figure 6 It is the contribution weight of each VOCs component in the joint effect of emergency hospitalization for cardiovascular disease.

[0054] Figure 7 The combined effect of mixed exposure to each group of VOCs components on cardiovascular disease hospitalization.

[0055] Figure 8 The contribution weight of each VOCs component to the joint effect of cardiovascular disease hospitalization. DETAILED DESCRIPTION

[0056] In order to make the purpose, technical solution and beneficial effects of the embodiments of the present disclosure clearer, the technical solution in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the drawings in the embodiments of the present disclosure. Obviously, the described embodiments are part of the embodiments of the present disclosure, not all of the embodiments. Based on the embodiments of the present disclosure, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present disclosure.

[0057] Example

[0058] This embodiment provides a method for identifying key toxic components in mixed exposure of atmospheric VOCs components. Figure 1 is implemented in the system shown.

[0059] The system includes a data acquisition module, a grouping module and an analysis module. The data acquisition module is used to collect the daily number of emergency hospitalizations for cardiovascular diseases and the hourly concentration data of daily atmospheric VOCs components in the selected area for at least one year, and calculate the daily average concentration of each VOCs component. The grouping module groups the monitored VOCs components according to their chemical structure. The analysis module constructs the association between VOCs groups and emergency hospitalizations for cardiovascular diseases (CVD) through the WQS (Weighted quantile sum) model, and screens out key VOCs components that have a greater impact on health outcomes.

[0060] In a preferred solution, the data collection module is also used to collect the daily average concentration of air pollutants in the selected location. In order to adjust the confounding effect of meteorological factors, daily meteorological data of the selected location is also collected. Among them, air pollutants include fine particulate matter (PM 2.5 ), ozone (Ozone, O3) and nitrogen dioxide (Nitrogendioxide, NO2), and meteorological data include daily average temperature and daily relative humidity data.

[0061] In a more preferred embodiment, the analysis module also includes an auxiliary analysis module, which is used to construct the association between the key VOCs components and the number of emergency admissions for cardiovascular diseases through GLM (Generalize linear model) to obtain the exposure-response relationship between the short-term exposure of the key VOCs components and the number of emergency admissions for cardiovascular diseases.

[0062] The method in this embodiment comprises the following steps:

[0063] S1. Collect the daily number of emergency hospital admissions for cardiovascular diseases and the hourly concentration data of daily atmospheric VOCs components in a certain place for at least one year, and calculate the daily average concentration of each VOCs component;

[0064] S2. Group the monitored VOCs components according to their chemical structures to obtain different VOCs groups;

[0065] S3. Establish the association between VOCs groups and emergency hospitalizations for cardiovascular diseases, and screen out key VOCs components that have a greater impact on health outcomes;

[0066] Among them, the association between the VOCs group and the number of emergency admissions for cardiovascular diseases was constructed by the WQS model, and the WQS model is as follows:

[0067]

[0068] g(u)=β0+β1WQS+z'Φ

[0069] Among them, g(u) represents the number of emergency hospitalizations for cardiovascular diseases in a certain place on study day u, β0 is the intercept, z' and Φ represent the covariate matrix and covariate coefficient, c represents the number of components included in each VOCs group, W i It represents the weighted coefficient of VOCs components. The sum of the weighted coefficients of the components in each VOCs group is equal to 1. The weighted coefficient of a single component ranges from 0 to 1. If the weighted coefficient of a single component is > 1 / c, the component is considered to be a key component. β1 is the regression coefficient of the WQS index. i Represents the scores of components in each VOCs group.

[0070] In a preferred solution, in order to ensure the stability of the population size in the study area and avoid risk estimation errors caused by population fluctuations, in step S1, the number of emergency admissions for cardiovascular diseases of people who have lived in the selected area for less than 6 months is excluded.

[0071] In the preferred scheme, it also includes the step of establishing the association between the key VOCs components and the number of emergency hospitalizations for cardiovascular diseases, so as to obtain the exposure-response relationship between the short-term exposure of the key VOCs components and the number of emergency hospitalizations for cardiovascular diseases;

[0072] Among them, the GLM model was used to construct the association between key VOCs components and the number of emergency admissions for cardiovascular diseases. The GLM model is as follows:

[0073] Y t =Intercept+β1VOC t +ns(time,df)+ns(tem t ,df)+ns(rh t,df)+dow+holiday

[0074] Among them, Y t represents the number of emergency admissions for cardiovascular disease on observation day t; VOC t represents the concentration of key VOCs components on observation day t; β1 represents the exposure-response relationship coefficient between key VOCs components and emergency admissions for cardiovascular diseases; ns(time, df) is the use of natural spline function to control time trend; ns(tem t , df) and ns(rh t , df) is the natural spline function used to control meteorological factors; dow and holiday are dummy variables used to control the day of the week effect and holiday effect.

[0075] In the preferred scheme, sensitivity analysis is also included:

[0076] At least two monitoring stations were set up to collect daily hourly concentration data of atmospheric VOCs components to observe exposure measurement errors;

[0077] The stability of the WQS model is evaluated by changing the temporal degrees of freedom and meteorological degrees of freedom settings in the natural spline.

[0078] WQS model analysis was performed on the air pollutants and VOCs respectively to test the effects of the air pollutants on VOCs. The WQS model was used to construct the association between the VOCs group and the number of emergency admissions for cardiovascular diseases, and the atmospheric pollutants were controlled in the model. The WQS model was as follows:

[0079]

[0080]

[0081] g(u)=β0+β poll Poll+β1WQS+z'Φ

[0082] Among them, g(u) represents the number of emergency hospitalizations due to cardiovascular diseases in the selected population on study day u, β0 is the intercept, z' and Φ represent the covariate matrix and covariate coefficient, c represents the number of components included in each VOCs group, W i It represents the weighted coefficient of VOCs components. The sum of the weighted coefficients of the components in each VOCs group is equal to 1. The weighted coefficient of a single component ranges from 0 to 1. If the weighted coefficient of a single component is > 1 / c, the component is considered to be a key component. β1 is the regression coefficient of the WQS index. i Indicates the score of the components in each VOCs group, β poll is the coefficient of atmospheric pollutants, and poll is the concentration level of atmospheric pollutants.

[0083] In a more preferred embodiment, a hysteresis analysis is also included:

[0084] The percentage increase in the risk of emergency hospitalization for cardiovascular disease in the population from the day of exposure to the different VOCs groups to the 7th day was calculated, including the calculation of single-day lag effects and moving average lag effects.

[0085] In a preferred embodiment, the VOCs group includes normal alkanes, alkenes, iso / trans alkanes, aromatic hydrocarbons, aldehydes, ketones, ethers, halogenated aromatic hydrocarbons, halogenated saturated chain hydrocarbons and halogenated unsaturated chain hydrocarbons.

[0086] Application Examples

[0087] The technical solution claimed in the present disclosure is further described below with specific application examples.

[0088] Ⅰ. Data Collection

[0089] Obtain the number of daily CVD emergency hospital admissions in a certain place from January 1, 2015 to December 31, 2016;

[0090] The atmospheric VOCs online automatic monitoring system (TH-300B system) was used to monitor the concentration of atmospheric VOCs components. The system is a dual-channel system using the GC-FID / MS method, in which the GC-FID channel measures C2 to C5 hydrocarbons, and the GC-MS channel measures oxygen-containing VOCs, halogen compounds, and C5 to C12 hydrocarbons. The monitoring system can analyze nearly 100 compounds through a single sampling and detection. During the study, monitoring systems were placed at at least two different locations in the selected area to obtain hourly concentration data for 98 atmospheric VOCs components. The average daily concentration of each component at each point was calculated, and finally the average daily concentration of the two locations was used to represent the atmospheric VOCs component exposure data in the area;

[0091] Based on fixed environmental monitoring stations in the selected areas, daily atmospheric fine particulate matter, ozone and nitrogen dioxide data were collected. The daily average of the monitoring data at each station was calculated to obtain city-level concentration data. In order to adjust the confounding effect of meteorological factors, the daily average temperature and daily relative humidity data of the area during the study period were also obtained.

[0092] II. Statistical Analysis

[0093] The mean, standard deviation, median, minimum, maximum and interquartile range (IQR) were used to analyze the daily emergency hospitalizations for CVD, air pollutants (PM 2.5, O3 and NO2) and meteorological data were analyzed descriptively. Considering that the properties of a compound are mostly determined by its chemical structure, pollutants with the same chemical structure may have similar toxicity and may also be produced by common emission sources or atmospheric chemical reactions. In this application example, 98 VOCs components were grouped according to their chemical structure, as shown in Table 1. Violin plots and heat maps were combined to show the different VOCs groups and the concentrations of VOCs components. In addition, the Pearson correlation matrix of each group of VOCs components was calculated.

[0094] Table 1. Grouping of 98 VOCs components

[0095]

[0096]

[0097] This application example uses the WQS model based on regression analysis to construct a model between the VOCs group and the number of emergency hospitalizations for cardiovascular diseases, and estimates the joint effect value of the mixed exposure of each group of components on health outcomes. This value is obtained by weighting the weights and concentrations associated with different components and health outcomes. WQS can not only estimate the joint effect of all component mixed exposures on health outcomes by constructing a weighted index, but also explore the relative strength of the contribution of individual exposure factors to the overall index effect under the actual situation of a series of component mixed exposures, thereby screening components that have a greater impact on the outcomes.

[0098] The WQS model formula is as follows:

[0099]

[0100] g(u)=β0+β1WQS+z'Φ

[0101] Where g(u) represents the number of emergency hospitalizations for CVD among the residents in the selected area on study day u, β0 is the intercept, z' and Φ represent the covariate matrix and covariate coefficient, c represents the number of VOCs components in each group included in the analysis, and W i It represents the weighted coefficient of VOCs component. The sum of weighted coefficients of each group of VOCs components is equal to 1. The weighted coefficient of a single component ranges from 0 to 1. If the weighted coefficient of a single component is > 1 / c, the component is considered to be a key component. β1 is the regression coefficient of WQS index. i Indicates the score of the components in each VOCs group. In this application example, the data was randomly divided into a training data set (40%) and a validation data set (60%), and the number of iterations was set to 1000. The regression coefficient of the WQS index was restricted to be positive, and a natural cubic spline with 3 degrees of freedom per year was used to control the air temperature and relative humidity.

[0102] The Poisson generalized linear model was used to construct a model between key VOCs components and CVD emergency hospitalizations to obtain the exposure-response relationship between short-term exposure to key VOCs components and CVD emergency hospitalizations.

[0103] The GLM model is as follows:

[0104] Y t =Intercept+β1VOC t +ns)time,df)+ns(tem t ,df)+ns(rh t ,df)+dow+holiday

[0105] In the above formula, Y t represents the number of emergency admissions for CVD on observation day t; VOC t represents the concentration of key VOCs components on observation day t; β1 represents the exposure-response relationship coefficient between key VOCs components and the number of CVD emergency hospitalizations; ns(time, df) is the use of natural spline function to control the time trend, with 6 degrees of freedom per year; ns(tem t , df) and ns(rh t , df) is the natural spline function used to control meteorological factors, and 3 degrees of freedom are preset; dow and holiday are dummy variables used to control the day of the week effect and holiday effect. Figure 2 The technical roadmap for WQS and GLM model analysis is shown.

[0106] According to previous research results, atmospheric VOCs exposure on the day of exposure and 2 days after exposure is significantly correlated with cardiovascular events. Therefore, this application example considers the day of atmospheric VOC exposure (lag0), single-day lag (lag1), 2-day lag (lag2), 3-day lag (lag3), 4-day lag (lag4), 5-day lag (lag5), 6-day lag (lag6), 7-day lag (lag7), and moving average lag effects (lag01, lag02, lag03, lag04, lag05, lag06, lag07).

[0107] A series of sensitivity analyses were also performed in this application example:

[0108] First, the association between daily atmospheric VOCs exposure and CVD emergency hospital admissions from two monitoring stations (site A and site B) was evaluated to observe exposure measurement errors;

[0109] Secondly, the model stability was evaluated by changing the time degrees of freedom setting (df = 6, 7) and the meteorological degrees of freedom parameter (df = 3, 5) in the natural cubic spline;

[0110] Then, WQS model analysis was performed on other air pollutants and VOCs to test the impact of other pollutants on VOCs. The WQS model was used to construct the association between the VOCs group and the number of emergency admissions for cardiovascular diseases, while controlling for atmospheric pollutants in the model. The WQS model is as follows:

[0111]

[0112] g(u)=β0+β poll Poll+β1WQS+z'Φ

[0113] Among them, g(u) represents the number of emergency hospitalizations due to cardiovascular diseases in the selected population on study day u, β0 is the intercept, z' and Φ represent the covariate matrix and covariate coefficient, c represents the number of components included in each VOCs group, W i It represents the weighted coefficient of VOCs components. The sum of the weighted coefficients of the components in each VOCs group is equal to 1. The weighted coefficient of a single component ranges from 0 to 1. If the weighted coefficient of a single component is > 1 / c, the component is considered to be a key component. β1 is the regression coefficient of the WQS index. i Indicates the score of the components in each VOCs group, β poll is the coefficient of atmospheric pollutants, and poll is the concentration level of atmospheric pollutants.

[0114] Finally, the association between VOCs and CVD hospitalization was estimated.

[0115] All modeling and statistical analyses in this application example were performed in R statistical software (4.1.3).

[0116] Results

[0117] Descriptive Analysis:

[0118] There were 219,751 CVD emergency hospital admissions in the selected regions between January 1, 2015, and December 31, 2016. Descriptive analyses of other air pollutants and meteorological data are shown in Table 2 .

[0119] From January 1, 2015 to December 31, 2016, the average daily concentration of atmospheric VOCs was 54.14±32.94ppb, of which n-alkanes accounted for the majority (31.8% of the total atmospheric VOCs concentration), with an average daily concentration of 17.24±10.71ppb, followed by aldehydes, ketones and ethers (22.0% of the total atmospheric VOCs concentration), with an average daily concentration of 11.91±5.85ppb, and halogenated aromatic hydrocarbons were the group with the smallest concentration (0.1% of the total atmospheric VOCs concentration), with an average daily concentration of 0.07±0.05ppb. The average daily concentrations of each group and component are shown in Figure 2. Figure 3The correlation coefficients of each group of VOCs components are shown in Figure 4 shown.

[0120] Table 2. Descriptive analysis of CVD emergency hospital admissions, other pollutants, and meteorological data in selected regions in 2015-2016

[0121]

[0122] The combined effect of mixed exposure to each group of VOCs components on CVD emergency hospitalization:

[0123] Figure 5 The joint effect of CVD emergency hospitalization related to each VOCs group with each quartile increase in WQS index was shown. Among them, the risk of cardiovascular emergency hospitalization increased by 1.09% (95% CI: 0.13%, 2.06%), 0.98% (95% CI: 0.22%, 1.74%), 0.92% (95% CI: 0.14%, 1.69%), 1.03% (95% CI: 0.14%, 1.93%), 1.69% (95% CI: 0.48%, 2.91%), and 1.85% (95% CI: 0.93%, 2.79%) for each quartile increase in WQS index of n-alkanes, iso / trans-alkanes, aromatic hydrocarbons, halogenated aromatic hydrocarbons, halogenated saturated hydrocarbons, and halogenated unsaturated hydrocarbons. No significant association was found between alkenes, aldehydes, ketones and ethers and the risk of emergency hospitalization for CVD (0.94% (95% CI: -0.14%, 2.02%), 1.03% (95% CI: -0.10%, 2.18%)).

[0124] Hysteresis Analysis:

[0125] The single-day lag and moving average lag effects were calculated, and the results showed that the acute effects of n-alkane, iso / trans-alkane and halogenated saturated chain hydrocarbon group VOCs on CVD emergency hospitalization peaked at lag02, the acute effects of aromatic hydrocarbon group VOCs peaked at lag01 and lag02, the acute effects of halogenated aromatic hydrocarbon group VOCs peaked at lag0, and the acute effects of halogenated unsaturated chain hydrocarbon group VOCs peaked at lag01 (Tables 3 and 4).

[0126]

[0127]

[0128] Screening of key components in each group:

[0129] Figure 6 is the contribution weight of each VOCs component to the joint effect of CVD emergency hospitalization, from Figure 6It can be seen that among the combined effects of n-alkanes on emergency admissions for CVD, the components with higher weights are n-undecane, cyclohexane, n-heptane, cyclopentane, n-pentene, and n-dodecane (weights are 0.17, 0.16, 0.16, 0.15, 0.11, and 0.10, respectively); among iso-transalkanes, the component with higher weights is 3-methylhexane, followed by 3-methylpentane, methylcyclohexane, isopentane, 2,2-dimethylbutane, and methylcyclopentane (weights are 0.29, 0.20, 0.15, 0.10, and 0.08, respectively); Among aromatic hydrocarbons, cumene, m-diethylbenzene, styrene, toluene, and 1,2,4-trimethylbenzene have higher weights, which are 0.27, 0.25, 0.17, 0.12, and 0.07, respectively. Among aromatic hydrocarbons, m-xylene, o-xylene, benzene, cumene, and ethylbenzene have higher weights, which are 0.24, 0.22, 0.13, 0.11, and 0.07, respectively. Among halogenated aromatic hydrocarbons and halogenated unsaturated chain hydrocarbons, 1,4-dichlorobenzene and tetrachloroethylene are the most important substances, with weights of 0.41 and 0.25, respectively. The substance with the highest weight among halogenated saturated chain hydrocarbons is Freon-114 (weight 0.32). After Freon-114, carbon tetrachloride, 1,1,1-trichloroethane, methyl chloride, and dichloromethane have higher weights (weights 0.15, 0.14, 0.07, and 0.07, respectively).

[0130] Exposure-response relationships of key components to CVD emergency hospital admissions:

[0131] Table 5 shows the percentage increase in risk of emergency hospitalization for CVD per IQR increase in key components. n-Undecane, cyclohexane, n-heptane, cyclopentane, n-pentene, n-dodecane, 3-methylhexane, 3-methylpentane, methylcyclohexane, isopentane, isopentane, cumene, n-styrene, toluene, 1,2,4-trimethylbenzene, 1,4-dichlorobenzene, carbon tetrachloride, 1,1,1-trichloroethane, methyl chloride, and dichloromethane were significantly associated with emergency hospitalization for CVD. The risk of CVD emergency hospitalization increased by 8.68% (95% CI: 3.74%, 13.86%), 5.62% (95% CI: 1.35%, 10.06%), 3.84% (95% CI: 1.21%, 6.53%), 3.30% (95% CI: 1.93%, 4.69%), 1.74% (95% CI: 1.10%, 2.39%), and 1.26% (95% CI: 0.66%, 1.87%) for each IQR increase in methylcyclohexane, styrene, dichloromethane, 1,1,1-trichloroethane, toluene, and 1,4-dichlorobenzene, respectively.

[0132] Table 5. Percentage increase in the risk of CVD emergency hospitalization per IQR increase in key components

[0133]

[0134] Sensitivity analysis:

[0135] The results of sensitivity analysis are shown in Tables 6, 7, 8 and Figure 7 and Figure 8 . In order to observe the impact of exposure measurement errors, the association between VOCs components and CVD emergency hospitalization at the two monitoring stations was analyzed. Compared with the main results, the change in the joint effect of the VOCs group that can increase the risk of CVD emergency hospitalization is negligible (Table 6). By changing the degrees of freedom of the time trend and meteorology in the model, the results (Table 6) show that when the degrees of freedom change, almost all groups can increase the risk of CVD emergency hospitalization. When different air pollutant and VOCs models were constructed separately, the results are shown in Table 7. The VOCs group that caused an increased risk of CVD emergency hospitalization was almost the same as the main result. In addition, the association between VOCs components and CVD hospitalization was estimated, and the joint effect of VOCs components and the components that played a key role were similar to CVD emergency hospitalization (Tables 8, Figure 7 and Figure 8 ).

[0136]

[0137] Table 7. Percentage increase in the risk of emergency hospitalization due to CVD in people exposed to short-term exposure to different VOCs groups after controlling for other pollutants

[0138]

[0139] Table 8. Percentage increase in the risk of CVD hospitalization per IQR increase in key components

[0140]

[0141] The scheme of this application case takes into account the exposure data of 98 atmospheric VOCs components and a large amount of CVD emergency hospitalization data, and the results are more reliable. Among them, taking emergency hospitalization as a health outcome overcomes the planned hospitalization defects that may exist in previous studies that use hospitalization as a health outcome. Considering that the sources of VOCs with the same chemical structure may have commonalities, this application case also groups 98 atmospheric VOCs components according to chemical structure, which is conducive to further analysis of chemical structure sources or atmospheric processes. In addition, based on the actual situation of mixed exposure, the WQS model is used to analyze the association between mixed exposure of VOCs components of different chemical structure groups and CVD emergency hospitalization, explore the joint effects of components within each group, and screen out components that play a key role. In addition, the GLM model is used to establish a model of key components and CVD emergency hospitalization, and the exposure-response relationship effect value is obtained to further verify the results of the WQS model and make the results more stable and reliable.

[0142] Unlike the existing scheme that only uses the total concentration of atmospheric VOCs components for analysis without considering the interaction between components, the scheme disclosed in this disclosure considers the joint effect between the various atmospheric VOCs components, indicating that mixed exposure to atmospheric VOCs components can increase the risk of cardiovascular disease, filling the gap in epidemiological evidence of low-concentration atmospheric VOCs exposure. Since there is currently little epidemiological evidence on the association between halogenated aromatic hydrocarbons, halogenated saturated chain hydrocarbons, and halogenated unsaturated chain hydrocarbons and cardiovascular disease, the scheme disclosed in this disclosure helps to fill the epidemiological evidence of atmospheric halogenated aromatic hydrocarbons, halogenated saturated chain hydrocarbons, and halogenated unsaturated chain hydrocarbon VOCs causing cardiovascular damage.

[0143] The method based on WQS and GLM models used in the present disclosure is suitable for analyzing the acute impact of exposure to multi-component mixed atmospheric VOCs on human health and selecting components that play a key role. It also provides a scientific basis for targeted control of the emission of high health risk VOCs components.

[0144] The embodiments and application examples described above are merely illustrative descriptions of the present disclosure and are not intended to limit the scope of the present disclosure. Without departing from the design spirit of the present disclosure, various modifications and improvements made to the technical solutions of the present disclosure by ordinary technicians in this field should all fall within the protection scope determined by the claims of the present disclosure.

Claims

1. A method for identifying key toxic components of mixed exposure of atmospheric VOCs components, characterized in that: The following steps are involved: S1. Collect the number of emergency hospital admissions for cardiovascular diseases and the daily average concentrations of atmospheric VOCs components in a certain place for at least one year; S2. grouping the monitored VOCs components according to chemical structure to obtain different VOCs groups; S3. Establish the association between the VOCs group and the number of emergency admissions for cardiovascular diseases, and screen out the key VOCs components that have a greater impact on health outcomes; The association between the VOCs group and the number of emergency admissions for cardiovascular diseases was constructed by the WQS model, and the WQS model was as follows: g(u)=β0+β1WQS+z'Φ Among them, g(u) represents the number of emergency hospitalizations due to cardiovascular diseases in the selected population on study day u, β0 is the intercept, z' and Φ represent the covariate matrix and covariate coefficient, c represents the number of components included in each VOCs group, W i represents the weighted coefficient of the VOCs component. The sum of the weighted coefficients of the components in each VOCs group is equal to 1. The weighted coefficient of a single component ranges from 0 to 1. If the weighted coefficient of a single component is > 1 / c, the component is considered to be a key component. β1 is the regression coefficient of the WQS index. i Represents the score of the components in each VOCs group.

2. The method according to claim 1, characterized in that In step S1, the number of emergency admissions for cardiovascular diseases of people who have lived in the selected place for less than 6 months is excluded.

3. The method according to claim 1, characterized in that In step S1, it also includes collecting the daily average concentration value of the selected atmospheric pollutants and daily meteorological data, the atmospheric pollutants include atmospheric fine particulate matter, ozone and nitrogen dioxide, and the meteorological data include daily average temperature and daily relative humidity data.

4. The method according to claim 3, characterized in that: It also includes constructing the association between the key VOCs components and the number of emergency admissions for cardiovascular diseases to obtain the exposure-response relationship between the short-term exposure of the key VOCs components and the number of emergency admissions for cardiovascular diseases; The GLM model is used to construct the association between the key VOCs components and the number of emergency admissions for cardiovascular diseases. The GLM model is as follows: Y t =Intercept+β1VOC t +ns(time,df)+ns(tem t ,df)+ns(rh t ,df)+dow+holiday Among them, Y t represents the number of emergency admissions for cardiovascular disease on observation day t; VOC t represents the concentration of the key VOCs component on observation day t; β1 represents the exposure-response relationship coefficient between the key VOCs component and the number of emergency admissions for cardiovascular diseases; ns(time, df) is the use of natural spline function to control the time trend; ns(tem t , df) and ns(rh t , df) is the natural spline function used to control meteorological factors; dow and holiday are dummy variables used to control the day of the week effect and holiday effect.

5. The method according to claim 3, characterized in that: Sensitivity analyses were also included: At least two monitoring stations were set up to collect daily average concentrations of atmospheric VOCs components to observe exposure measurement errors; The stability of the WQS model was evaluated by changing the temporal degrees of freedom and meteorological degrees of freedom settings in the natural spline.

6. The method according to claim 5, characterized in that The sensitivity analysis also includes performing WQS model analysis on the air pollutants and the VOCs respectively to examine the impact of the air pollutants on VOCs; wherein the association between the VOCs group and the number of emergency admissions for cardiovascular diseases is constructed through the WQS model, and the air pollutants are controlled in the model. The WQS model is as follows: g(u)=β0+β poll ·Poll+β1WQS+z'Φ Among them, g(u) represents the number of emergency hospitalizations due to cardiovascular diseases in the selected population on study day u, β0 is the intercept, z' and Φ represent the covariate matrix and covariate coefficient, c represents the number of components included in each VOCs group, W i represents the weighted coefficient of the VOCs component. The sum of the weighted coefficients of the components in each VOCs group is equal to 1. The weighted coefficient of a single component ranges from 0 to 1. If the weighted coefficient of a single component is > 1 / c, the component is considered to be a key component. β1 is the regression coefficient of the WQS index. i represents the score of the components in each VOCs group, β poll is the coefficient of the atmospheric pollutant, and poll is the concentration level of the atmospheric pollutant.

7. The method according to claim 3, characterized in that Also includes hysteresis analysis: The percentage increase in the risk of emergency hospitalization for cardiovascular disease in the population from the day of exposure to the different VOCs groups to the 7th day was calculated, including the calculation of single-day lag effects and moving average lag effects.

8. The method according to claim 1, characterized in that The VOCs group includes normal alkanes, alkenes, iso / trans alkanes, aromatic hydrocarbons, aldehydes, ketones, ethers, halogenated aromatic hydrocarbons, halogenated saturated chain hydrocarbons and halogenated unsaturated chain hydrocarbons.

9. A key toxic component identification system for mixed exposure of atmospheric VOCs components, characterized in that: The system is used to implement the method according to any one of claims 1 to 8, comprising a data acquisition module, a grouping module and an analysis module. The data collection module collects the number of emergency hospital admissions for cardiovascular diseases and the daily average concentration of atmospheric VOCs components in a certain place for at least one year; The grouping module groups the monitored VOCs components according to chemical structures to obtain different VOCs groups; The analysis module constructs the association between the VOCs group and the number of emergency admissions for cardiovascular diseases through the WQS model, and screens out key VOCs components that have a greater impact on health outcomes; The formula of the WQS model is as follows: g(u)=β0+β1WQS+z′Φ Among them, g(u) represents the number of emergency hospitalizations due to cardiovascular diseases in the selected population on study day u, β0 is the intercept, z′ and Φ represent the covariate matrix and covariate coefficient, c represents the number of components included in each VOCs group, W i It represents the weighting coefficient of the VOCs component. The sum of the weighting coefficients of the components in each VOCs group is equal to 1. The weighting coefficient of a single component ranges from 0 to 1. If the weighting coefficient of a single component is > 1 / c, the component is considered to be a key component. β1 is the regression coefficient of the WQS index. qi represents the score of the component in each VOCs group.

10. The system according to claim 9, characterized in that It also includes an auxiliary analysis module, which constructs the association between the key VOCs components and the number of emergency admissions for cardiovascular diseases through a GLM model to obtain the exposure-response relationship between the short-term exposure of the key VOCs components and the number of emergency admissions for cardiovascular diseases; The GLM model is as follows: Y t =Intercept+β1VOC t +ns(time,df)+ns(tem t ,df)+ns(rh t ,df)+dow+holiday Among them, Y t represents the number of emergency admissions for cardiovascular disease on observation day t; VOC t represents the concentration of the key VOCs component on observation day t; β1 represents the exposure-response relationship coefficient between the key VOCs component and the number of emergency admissions for cardiovascular diseases; ns(time, df) is the use of natural spline function to control the time trend; ns(tem t , df) and ns(rh t , df) is the natural spline function used to control meteorological factors; dow and holiday are dummy variables used to control the day of the week effect and holiday effect.

Images