Associations of long-term exposure to ambient air pollution with blood pressure and hypertension de ned by Chinese and ACC/AHA Guideline: The CHCN-BTH cohort study


 Background：The association between long-term air pollutants exposure with blood pressure and hypertension defined by 2017 American College of Cardiology (ACC)/American Heart Association (AHA) Hypertension Guideline is still conflicting. This study was designed to investigate the associations between long-term exposure to air pollutants, blood pressure and hypertension defined by Chinese and ACC/AHA guideline in Chinese adults. Methods: Our study was based on the baseline survey of the Cohort Study on Chronic Disease of Communities Natural Population in Beijing, Tianjin and Hebei (CHCN-BTH) from 2017 to 2019. A spatial statistical model was used to assessed three-year (2014-2016) average pollutant concentrations for PM2.5, and other pollutants concentration (PM10, SO2, NO2) was assessed by data of air monitoring stations. Results: A total of 32,135 adults aged 18-80 years were included, each interquartile range (IQR) increment of PM2.5, PM10, SO2 and NO2 was associated with increases of 0.66mmHg (95%CI: 0.29, 1.03), 040mmHg (0.00, 0.81), 1.38mmHg (0.92, 1.84) and 0.54mmHg (0.16, 0.91) in SBP, respectively. SO2 was associated with increases 0.42mmHg (0.11, 0.72) in DBP. PM2.5, PM10, SO2 and NO2 was associated with an 14% (Odds ratio [OR]:1.14, 95%CI:1.06-1.23), 6% (1.06, 1.00-1.13), 9% (1.09, 1.02-1.17) and 8% (1.08, 1.01-1.16) increase of hypertension defined by Chinese guideline. PM2.5 and SO2 was associated with 16% (1.166, 1.093-1.245) and 10% (1.10, 1.03-1.18) increase of hypertension defined by ACC/AHA guideline. Two-pollutants model, traffic-related pollution model and stratiﬁed analysis yielded similar results. Conclusions: We found that long-term exposure to PM2.5, NO2 and SO2 is associated with increase of blood pressure and hypertension defined by both Chinses and ACC/AHA guideline. PM10 is associated with higher SBP and risk of hypertension defined by Chinese guideline.


Background
Hypertension is a main risk factor for cardiovascular disease that contributed to 17.7 million deaths in the word in 2017 [1,2]. Particularly in China, approximately 27.9% of adults have hypertension, and higher prevalence of hypertension was found in large and medium-sized cities [3].
In recent years, mounting epidemiological studies have shown that exposure to air pollution was related to higher blood pressure and risk of hypertension [4][5][6][7], a recent meta-analysis reviewed the existing researches and reported globally signi cant associations of long-term exposure to PM 2.5 with hypertension and PM 2.5 /NO 2 with DBP [8]. A cross-sectional study in China found PM 2.5 and PM 10 concentration was signi cantly associated with increases of SBP [9]. Another cross-sectional study in India found positive association between PM 2.5 and SBP only in men [10]. However, the result is inconsistent, a Korean study reported no signi cant association between air pollution and SBP [11], a few studies even found a negative association between NO 2 or O 3 and blood pressure [12,13].
In addition, compared with SBP and DBP, much less evidence has been available for air pollution and mean arterial pressure (MAP) as well as pulse pressure (PP). Moreover, rapid economic growth and related emissions have led to increases in air pollution of Beijing-Tianjin-Hebei (BTH) metropolitan region in China recent years [14]. Many previous studies conducted in North America or Europe, where the average PM 2.5 levels were much lower than those in BTH region.
The American College of Cardiology (ACC)/American Heart Association (AHA) Task Force on Clinical Practice Guidelines recently released the 2017 hypertension guideline. One new feature of this guideline is de ning a systolic blood pressure(SBP) of 130 to 139 mmHg or diastolic blood pressure (DBP) of 80 to 89 mmHg as stage 1 hypertension [15], which has been evolving from an SBP/DBP of 140/90 mmHg or higher in 1997 [16], and a further decrease to 130/80 mm Hg in 2017. However, as we known, there is no evidence to prove the association of air pollution and hypertension which de ned by AHA/ACC guideline.
On the other hand, the types of ambient air pollutants are complex, the evidence of association between multi-air pollutants and blood pressure or hypertension was still lack. A retrospective cross-sectional study in Japan showed multi-pollutants, which included suspended particulate matter, SO 2 , NO 2 and photochemical oxidants, could increase blood pressure and the risk of hypertension [17]. Another study in Pakistan also found multi-pollutants were associated with high risk of CVD [18]. We are not found study of multi-pollutants in BTH region, and our study can ll this gap and provide strong evidence of the association between multi-pollutants and blood pressure in BTH region. The aim of the present study was to investigate the association between long-term and high-level exposure of ambient air pollution with hypertension and BP among participants from the baseline survey of the Cohort Study on Chronic Disease of Communities Natural Population in Beijing, Tianjin and Hebei (CHCN-BTH). We further investigated and compared the association of long-term exposure to ambient air pollution and incident of hypertension which were de ned by both Chinese guidelines and AHA/ACC guideline. Moreover, proximity to roadways is considered as a substitute for long-term exposure to tra c-related air pollution, which can better capture exposure to both particle matter and gaseous mixtures. We also investigated several potential effect modi cations by demographic characteristics and health behaviors.

Study population
Our study was based on the CoHort Study on CHronic Disease of Communities Natural Population in Beijing, Tianjin and Hebei (CHCN-BTH, Registration number: ChiCTR1900024725), details of the cohort has been introduced elsewhere [19]. Generally, a multistage strati ed cluster sampling method was used to select participants from 41 communities, 36 townships and 7 functional units in ve cities (i.e., Beijing, Tianjin, and Chengde, Baoding and Shijiazhuang of Hebei Province) from 2017 to 2019. The survey sites covered the central core functional area, the southern expansion area and the northwestern ecological conservation area. We included adults aged 18 to 80, and excluded people had psychiatric disorder or disabled and pregnant women. All participants lived in survey site for more than 3 years. At baseline, participants underwent some physical examinations (e.g., height, weight and blood pressure measurements), questionnaire survey and blood tests. All the participants signed a consent form. This study was reviewed and approved by the ethical committees of Center of Disease Control and Capital Medical University.

Air pollution exposure assessment
A machine learning method was used for estimating the annual concentrations of particles with an aerodynamic diameter of ≤ 2.5µm (PM 2.5 ) from 2014 to 2016 (three years prior to baseline survey) at a resolution of 10km×10km, using high dimensional expansion of numerous predictors (including ground monitored PM 2.5 data, satellite-derived aerosol optical depth (AOD) and other satellite covariates, meteorological variables and chemical transport model simulations) [20]. The estimates were found to be in good agreement with in-situ observations, with determination coe cients (R 2 ) of 0.77. PM 2.5 concentrations was estimated according to geocode residential address of each participant. Then, we calculated the three-year average concentration before the baseline survey (2017) for each participant as the long-term exposure concentration of PM 2.5 .
Air pollutant concentration data is from China National Environmental Monitoring Centre (www.cnemc.cn). We collected air pollution concentration data in 103 monitoring stations of Beijing, Tianjin and Hebei province, included PM 10 SO 2 and NO 2 .The air pollution data is from January 1st 2014 to December 31th 2016. For the calculation of 24-hour mean concentrations of PM 10 , SO 2 and NO 2 , at least 75% of the one-hour values must be available on that particular day. If a station had more than 25% of the values missing for the whole period of analysis, the entire station was excluded from the analysis.

Proximity to major roadway
We calculated residential distance to major roadway based on residential address for neighborhood analysis, which could indirectly represent the tra c-relative pollution. The major roadways, were classi ed as national, provincial and county roadways according to the National Highway Classi cation in China.
We examined the proximity to roadways in categories of proximity < 100m, 100 to < 200m, 200 to < 400m, and 400 to ≤ 1000m. Residential addresses further than 1000 m from a major roadway are not likely an indicator of tra c-related exposure [21]. Therefore, we excluded these addresses (> 1000 m, 8632 participants), which left 23,503 (73.1%) participants for the association analyses of proximity to a major roadway and blood pressure or hypertension.

Outcome assessment
Blood pressure was measured using an electronic sphygmomanometer (OMRON HEM-907, Japan) in the sitting position for three times and the cuff was placed around 2cm above the right arm elbow. Participants were advised not to take hypotensive drugs, smoke, drink alcohol, coffee, or tea, and to abstain from exercising on the test day. Additionally, they were not allowed to talk during the measurement. The average value of three measurements was used as the blood pressure measurement for this study. MAP was calculated as DBP + 1/3 (SBP − DBP); PP was calculated as the difference between SBP and DBP. According to the 2018 Chinese guidelines for the management of hypertension [22], hypertension in our analysis was de ned as having a measured SBP ≥ 140 mm Hg or DBP ≥ 90 mm Hg, and/or having a self-reported physician-diagnosed hypertension or anti-hypertension treatment. We also de ned prevalent hypertension as having the measured SBP ≥ 130 mm Hg or DBP ≥ 80 mm Hg according to the AHA/ACC Hypertension Practice Guidelines 2017 [23].

Statistical analysis
A generalized linear mixed models (GLMMs) was used to estimate the associations between 3-year air pollutants exposure and the blood pressure (SBP, DBP, MAP, PP), covariates including age, sex, BMI, ethnicity, education, marital status, family history of hypertension, self-reported smoking and drinking, physical activity, monthly income, residential area and ambient temperature were included into the models to modify the confounding bias. When air pollutants were considered as a continuous variable, regression coe cient (β) and 95% con dence interval (CI) of BP was calculated by each IQR increment in the 3-year average air pollutants concentration. When the participants were categorized into 4 groups based on the quartiles of pollutants exposure, β value and 95%CI of BP was calculated using the rst quartile as a reference.
A two-level logistic regression model to assess the associations between 3-year air pollutants exposure and hypertension de ned by both Chinese and ACC/AHA guideline, where participants were considered as the rst-level unit and the community as the second-level unit in the model [24]. Odds ratios (ORs) and 95%CI of incident hypertension was reported for each IQR increment in the 3-year average air pollutants concentration. We also estimated the associations of proximity to major roadway with blood pressure and hypertension to representing the tra c-related air pollution. Spearman rank correlation coe cients were performed to assess the relationship between air pollutants, meteorological data and proximity to a major roadway.
We examined whether the association between air pollutants exposure and blood pressure or hypertension was modi ed by age (< 65y and ≥ 65y), gender (male and female), residence (rural and urban), smoking (never and ever smoker), alcohol drinking (never and ever drinking), physical exercise (usual and seldom), then reported the p values for interaction terms.
We also performed several sensitivity analyses. First, we used average air pollutants concentrations for 1-3 years before the survey. Second, participants with no information of hypertensive medication were excluded. Third, we excluded the participants with cardiovascular diseases. Statistical analyses were conducted using SAS version 9.4 (SAS Institute, Inc., Cary, NC). A two-tailed P value of < 0.05 was de ned as statistically signi cant.

Descriptive statistics
A total of 33,391 participants were measured BP, and 1,256 participants were excluded due to missing information on important potential confounders and hypertension history. The basic characteristics of the 32,135 study subjects are summarized in Table 1. Participants were on average 50.0 years and 55.4% of them were women. Most participants were Han, 72.1% of them live in urban, 30.7% of them have been a smoker and 42.7% of them used to drinking. There are 40.8% and 58.5% (Table A1) of the participants diagnosed as hypertension according to Chinese and AHA Hypertension Guideline, respectively. There were signi cant differences in all variables between hypertension de ned by Chinese or AHA/ACC guideline and non-hypertension group. Table 2 shows the 3-year average concentrations of six air pollutants exposure for the 32,135 participants.
The average concentration of SO 2 was 22.9μg/m 3 , which were lower than the National Ambient Air Quality Standards of China (60μg/m 3 ), however, the 3-year average concentration of PM 2.5 , PM 10 , NO 2 were 73.5μg/m 3 , 123.0μg/m 3 and 52.2μg/m 3 , respectively, which exceeded the National Ambient Air Quality Standards of China (35μg/m 3 , 70μg/m 3 and 40μg/m 3 ). The temperature of testing day was 14.7±12.2℃ and the humidity was 55.5±21.5%. The correlations of all air pollutants were moderate with each other.
The mean of proximity to a major roadway was 408.5±282.3m was negatively correlated with all air pollutants.

Associations of air pollutants with blood pressure and hypertension
The associations of long-term exposures to air pollutants with blood pressure are presented in Table 3.
Participants with PM 2.5 and SO 2 values in the fourth quartiles had much higher SBP, DBP, MAP, PP than those in the lowest quartile, and the air pollutants values in the second, third, and fourth quartiles had increased BP levels which compared with the lowest quartile (P for trend<0.01, Table 3). We found similar results between 3-y average NO 2 and SBP or PP. No signi cant association was found between PM 10 and any kinds of blood pressure.
The associations for air pollutants and Chinese/AHA hypertension guideline are shown in Table 4.
For Chinese hypertension guideline, compared with the lowest quintile (Q1) of PM 2.5 , SO 2  In two-pollutants model of air pollutants and blood pressure (Table A5) In two-pollutants model of air pollutants and hypertension (Table A6) 3.3 Associations of proximity to roadway with blood pressure and hypertension Participants whose residential address was ≥1000 m to a roadway were excluded in this association analyses, and 23,503 (73.1%) participants were remained. The main characteristics of this study participants were similar to Tabe1 (Supplementary materials Table A2). The result for proximity to roadway and blood pressure are shown in Figure1. Participants who live within 100m from major roadway are associated with signi cantly higher SBP and PP, and the effect decreased with the distance of proximity to roadway (P for trend=0.004 and 0.01, respectively). Participants who live within 100m from major roadway are associated with an increasing risk of Chinese/AHA hypertension and the OR decreased with the distance of proximity to roadway (P for trend=0.03 and 0.02).

Strati ed analyses for the associations of air pollutants with blood pressure and hypertension
The results of the stratified analysis of air pollution exposure and blood pressure were shown in Figure 2. The associations between PM 2.5 , SO 2 , NO 2 and PP in the age group≥65 years were more signi cant. PM 10 was associated with significantly higher DBP and MAP in the age group≥65 years. We also found higher associations of SO 2 and all four blood pressure values among participants in rural area.
The results of stratified analysis for hypertension are shown in Figure 3. The associations between PM 2.5 and SO 2 and hypertension were significantly higher in the age group<65 years in both Chinese and AHA/ACC hypertension guideline. We also found stronger associations between SO 2 and hypertension in rural participants.

Sensitivity analyses
The results of sensitivity analysis are presented in Supplementary materials. The associations of blood pressure and hypertension to long-term air pollution exposure did not substantially alter in the sensitivity analyses by using 1-to 3-year average concentrations (Table A3, A4). When we excluded the patients with antihypertensive therapy (Figure A1, A2) and cardiovascular cases ( Figure A3, A4) from the data, the results remained robust.

Discussion
To our best knowledge, this is the rst study to investigate long-term air pollution exposure and hypertension de ned by Chinses and AHA/ACC guideline. We found that that 3-year average exposure to SO 2 could signi cantly increase SBP, DBP, MAP and PP; PM 2.5 was associated with elevated SBP, MAP and PP; NO 2 could increase SBP and PP; PM 10 could only signi cantly increase SBP in full adjusted single-pollutant model. All four pollutants were associated with the increment of hypertension de ned by Chinses guideline, only PM 2.5 and SO 2 could increase the risk of hypertension de ned by AHA/ACC guideline. In two-pollutants model we found PM 2.5 and SO 2 combined with other pollutants could signi cantly increase blood pressure; and PM 2.5 could increase hypertension de ned by both Chinese and AHA/ACC guideline with other pollutants. PM 2.5 , NO 2 (only SBP and PP), SO 2 was signi cantly associated with elevated blood pressure (SBP, DBP, MAP, PP) and hypertension which were de ned by both Chinese and AHA guideline. Elder people had larger associations between PM 2.5 , NO 2 and SO 2 exposures and blood pressure (SBP and PP).
Our ndings of a positive association between PM 2.5 , SO 2 and NO 2 exposure and blood pressure are in line with some previously study. Several studies found positive association between PM 2.5 exposure and blood pressure in China [25], India [10] and America [26]. An American cohort study reported that each IQR(3.91μg/m 3 ) increase of 1-year average PM 2.5 exposure was associated with SBP (0.93mmHg; 95CI%,0.05-1.80) and PP (0.89mmHg; 95CI%,0.21-1.58) increased [27]. A large Sister Study in America also found a positive association between each 10μg/m 3 increase in PM 2.5 and SBP (1.4mmHg; 95CI%, 0.6-2.3) as well as PP (1.0mmHg; 95CI%, 0.4-1.7) [28]. And some studies also found positive association between long term SO 2 exposure and blood pressure [29,30]. A Chinese cross-sectional study found 0.80mmHg (95CI%, 0.46-1.14) increase of SBP and 0.31mmHg (95CI%, 0.10-0.51) increase of DBP associated with each IQR (20μg/m 3 ) increase in SO 2 [31] . In addition, a few studies found a positive association between short-term SO 2 exposure and blood pressure which can supplement our results [32,33]. Several studies found similar association between NO 2 exposure and blood pressure as we found in our study [25,34,35]. A recent study found no-liner association between PM 2.5 with SBP and DBP [5], the population (n= 137,809) of this study was from 21 countries and there may remain ecological residual confounding in the analyses to lead no signi cantly results. The results of multi-pollutants model were robust, previous study also showed similar association between particular matter and blood pressure in multi-pollutants model, short-term exposure to PM was signi cantly associated with an increase of SBP and DBP[36].
All four air pollutants were signi cantly associated with hypertension de ned by Chinses guideline. Some previous study also de ned hypertension by SBP≥140 mm Hg or DBP≥ 90 mm Hg as Chinses guideline.
The results in the China-PAR project showed an increased risk of incident hypertension (HR=1.77, 95% CI=1.56-2.00) among participants in the highest quartile exposure of PM 2.5 compared with the lowest quartile [6] .The Andhra Pradesh Children and Parents Study reported that each 1μg/m 3 increase of PM 2.5 was associated 4% higher odds of hypertension [10]. A Canadian study showed a weakly association between each IQR increase in NO 2 and incident hypertension (HR=1.01, 95% CI:1.00, 1.02) [37]. A study in America showed that gestational hypertension risk was signi cantly increased by 18% for SO 2 during the 3 months before conception [38]. Conversely, inconsistent associations were reported in several studies. A cross-sectional study in Taipei showed that 1-year PM 2.5 and NO 2 exposures were not associated with hypertension [39]. Black Women's Health Study in US found an inconsistent association between increasing of NO 2 exposure per IQR (9.7ppb) and hypertension (HR= 0.92, 95%CI: 0.86, 0.98) [12]. In addition, the results of PM 2.5 and SO 2 remain stable when we changed the de nition to SBP≥130 mm Hg or DBP≥ 80 mm Hg as AHA/ACC guideline, PM 10 and NO 2 showed a non-signi cant association. Few previous study reported air pollution and hypertension de ned by AHA/ACC, this is the rst study to investigate air pollution and AHA/ACC hypertension in China, the result suggested that whichever de nition is used, PM 2.5 and SO 2 is the major risk factor of hypertension.
In each city, the location of monitoring stations was not in the direct vicinity of tra c or of industrial sources, and were mandated not to be in uenced by local pollution sources and should also avoid buildings, or housing large emitters such as coal-, waste-, or oil-burning boilers, furnaces, and incinerators.
Thus, we estimated air pollution exposure based on monitoring stations data, at the same time, we considered proximity to a major roadway as tra c-related pollution. Some studies found similar results as we did, a recently study among ≥65 years old Chinese people reported that residential proximity to major roadways was negative associated with hypertension, the ORs for participants living 50 to 100, 101 to 200, and ≥200 meters from major roads were 1.17, 1.21 and 1.22 respectively [40]. We found non signi cantly association between PM 10 with blood pressure or hypertension, which could be explained that the association of particular matters and increasing of blood pressure is mainly affected by PM 2.5 rather than PM 10 . We also found the associations of PM 2.5 and SO 2 on PP appeared to be stronger consistently for participants ≥65 years old, which may be caused by vascular sclerosis in the ≥65 years group [41].
The biological mechanisms through which air pollutants can raise blood pressure are not fully understood. One potential pathway is that air pollutants may raise blood pressure and lead to Our study has several strengths. First, this is the rst study to investigate the air pollution and hypertension de ned by both Chinese and AHA/ACC guideline, and we found a stable correlation between PM 2.5 and SO 2 with hypertension, even though the de nition was changed to SBP≥130 mm Hg or DBP≥ 80 mm Hg. Second, the concentrations of all air pollutants were relatively high in our study regions rather than the low level of air pollution exposures in other countries, for example, the average concentration of PM 2.5 μg/m 3 America[46]19.6μg/m 3 in Danish[47] and 73.5μg/m 3 , therefore, our results could represent the association between high-level and long-term exposures with blood pressure and incident hypertension. Third, we analyzed the association between residential proximity to major roadways with blood pressure values and incident hypertension, which presents the effects of tra c-related pollution.
Fourth, the air pollution of our study included not only particular matters but also gaseous pollutants, furthermore we used two-pollutants model to explore the combination effect of air pollutants. Finally, a large set of covariates were considered in our comprehensive statistical analyses to minimize the impact of confounding variables, including BMI, education attainment, monthly income, gross domestic product, and population density.
There are several limitations in the present study. First, similar to previous cross-sectional studies, we could not establish a causal relationship between air pollutants exposure and blood pressure increasing or hypertension. Second, except for the concentrations of PM 2.5 , where individual levels were estimated by a spatial statistical model, the exposure levels of PM 10 , SO 2 , NO 2 were assigned using data from the nearest air monitoring station, which resulted in an exposure misclassi cation. Third, some covariates were not available, such as indoor air pollution and diet information, we cannot exclude the possible confounding bias by unmeasured factors. Finally, the information of drugs taking was not included in the questionnaire, which was an unmeasured confounding factor in our study.

Conclusion
In conclusion, we found that long-term exposure to higher levels of PM 2.5 , NO 2 and SO 2 is associated with increase of blood pressure values and hypertension de ned by both Chinses and ACC/AHA guideline adjusting for a wide range of covariates. PM10 is associated with higher SBP and risk of hypertension de ned by Chinese guideline. Participant who live within 100m from major roadway is associated with signi cantly higher SBP, PP and risk of hypertension. We also found robust association in strati ed analysis and sensitivity analysis. Our study provides evidence of a role for high levels air pollutants in the blood pressure increase and development of hypertension. However, further longitudinal studies are warranted to con rm our ndings.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. supplementalmaterial.docx