Next Article in Journal
The Quality of Tuberculosis Care in Urban Migrant Clinics in China
Next Article in Special Issue
Fine-Scale Source Apportionment Including Diesel-Related Elemental and Organic Constituents of PM2.5 across Downtown Pittsburgh
Previous Article in Journal
Socio-Cultural Reasons and Community Perceptions Regarding Indoor Cooking Using Biomass Fuel and Traditional Stoves in Rural Ethiopia: A Qualitative Study
Previous Article in Special Issue
Spatial Patterns in Rush-Hour vs. Work-Week Diesel-Related Pollution across a Downtown Core
Article Menu
Issue 9 (September) cover image

Export Article

Int. J. Environ. Res. Public Health 2018, 15(9), 2036; doi:10.3390/ijerph15092036

Article
Relationship of Time-Activity-Adjusted Particle Number Concentration with Blood Pressure
1
Department of Civil and Environmental Engineering, Tufts University School of Engineering, 200 College Ave, Medford, MA 02155, USA
2
Section of Preventive Medicine and Epidemiology, Boston University School of Medicine, 801 Massachusetts Avenue, Suite 470, Boston, MA 02118, USA
3
Department of Public Health and Community Medicine, Tufts University School of Medicine, 136 Harrison Avenue, Boston, MA 02111, USA
4
Health Effects Institute, 75 Federal Street, Suite 1400, Boston, MA 02110, USA
5
Department of Environmental Health, Boston University School of Public Health, 715 Albany St, Boston, MA 02118, USA
6
Tufts University Jonathan M. Tisch College of Civic Life, 35 Professors Row, Medford, MA 02155, USA
*
Author to whom correspondence should be addressed.
Received: 22 August 2018 / Accepted: 13 September 2018 / Published: 18 September 2018

Abstract

:
Emerging evidence suggests long-term exposure to ultrafine particulate matter (UFP, aerodynamic diameter < 0.1 µm) is associated with adverse cardiovascular outcomes. We investigated whether annual average UFP exposure was associated with measured systolic blood pressure (SBP), diastolic blood pressure (DBP), pulse pressure (PP), and hypertension prevalence among 409 adults participating in the cross-sectional Community Assessment of Freeway Exposure and Health (CAFEH) study. We used measurements of particle number concentration (PNC, a proxy for UFP) obtained from mobile monitoring campaigns in three near-highway and three urban background areas in and near Boston, Massachusetts to develop PNC regression models (20-m spatial and hourly temporal resolution). Individual modeled estimates were adjusted for time spent in different micro-environments (time-activity-adjusted PNC, TAA-PNC). Mean TAA-PNC was 22,000 particles/cm3 (sd = 6500). In linear models (logistic for hypertension) adjusted for the minimally sufficient set of covariates indicated by a directed acyclic graph (DAG), we found positive, non-significant associations between natural log-transformed TAA-PNC and SBP (β = 5.23, 95%CI: −0.68, 11.14 mmHg), PP (β = 4.27, 95%CI: −0.79, 9.32 mmHg), and hypertension (OR = 1.81, 95%CI: 0.94, 3.48), but not DBP (β = 0.96, 95%CI: −2.08, 4.00 mmHg). Associations were stronger among non-Hispanic white participants and among diabetics in analyses stratified by race/ethnicity and, separately, by health status.
Keywords:
particle number concentration; ultrafine particulate matter; time-activity adjustment; blood pressure; hypertension; traffic-related air pollution; directed acyclic graph

1. Introduction

Ambient particulate matter (PM) exposure is associated with over four million deaths per year and evidence suggests that certain size fractions of PM are associated with increased risk of hypertension, cardiovascular morbidity, and cardiovascular mortality [1,2,3,4,5,6]. Nonetheless, few epidemiologic studies have considered the cardiovascular impacts of long-term exposure to the smallest size fraction of PM, ultrafine particulate matter (UFP, aerodynamic diameter < 0.1 µm). This is despite toxicologic evidence that UFP may be the most toxic PM size fraction and that UFP may exert cardiovascular effects through mechanisms involving oxidative stress and systemic inflammation [7,8,9,10,11,12]. In the few prospective epidemiologic studies that have considered the cardiovascular consequences of long-term exposure to UFP, there has been reasonable agreement that UFP exposure is associated with adverse cardiovascular impacts including increased risk of ischemic heart disease mortality, increased hypertension risk, increased carotid intima-media thickness, and, in some cases, increased concentrations of biomarkers of inflammation [13,14,15,16,17,18]. Nevertheless, most previous studies have only considered UFP modeled with a spatial resolution of between 200 m and 4 km or have considered the cardiovascular impacts of UFP only in relatively homogenous populations. To inform risk assessment and policy development, studies in heterogeneous populations with highly spatially–resolved UFP exposure estimates are needed.
Modeling UFP with insufficiently high spatial resolution can result in substantial exposure misclassification as UFP concentrations rapidly decline within 100 m of sources, such as major roadways [19]. Similarly, exposure assessment methods that rely on residential average concentrations, rather than methods that also account for time individuals spend away from the home, can result in differential exposure misclassification [20]. We are only aware of one prospective study that modeled UFP at fine spatial resolution (≤20 m). This was our longitudinal analysis of the association of long-term exposure to UFP with blood pressure and C-reactive protein within the Boston Puerto Rican Health Study (BPRHS) [16]. Nevertheless, in that study, we could not account for individual time-activity patterns which could have reduced potential exposure misclassification [20]. Additionally, while there were some indications that certain sub-populations were more vulnerable to the effects of UFP, the BPRHS population consisted only of individuals who identified as Puerto Rican and most BPRHS participants were in generally poor overall health. To understand whether the associations we found in our previous study were generalizable to more diverse populations and whether the associations remained after accounting for time spent in different micro-environments, we used data from the Community Assessment of Freeway Exposure and Health (CAFEH) study.
In the cross-sectional CAFEH study, participants of multiple races/ethnicities were recruited from several communities in the greater Boston area. Using measurements of UFP (as particle number concentration (PNC)—a commonly used and reliable indicator of UFP [21,22]) from each community, we developed a finely resolved (hourly temporal and approximately 20 m spatial resolution) model. We adjusted mean hourly residential estimates using individual data on time spent in different micro-environments to assess time-activity-adjusted PNC (TAA-PNC) [20,23]. We previously found that long-term TAA-PNC was associated with biomarkers of systemic inflammation and with chronic outcomes in the CAFEH population [24,25]. Given evidence that these associations with PNC varied by race/ethnicity and that there were racial disparities in overall health status in this population, we considered effect modification by race/ethnicity and by health factors which could affect susceptibility to UFP [24,26,27,28]. Specifically, our objectives were: (1) to determine whether long-term exposure to TAA-PNC was associated with systolic blood pressure (SBP), diastolic blood pressure (DBP), pulse pressure (PP), and prevalent hypertension; and (2) to determine whether race/ethnicity, statin medication use, diabetes status, or hypertension status modified these associations.

2. Materials and Methods

The cross-sectional, community-based participatory CAFEH study was designed to investigate the relationship between UFP exposure and cardiovascular health. Detailed methods have been published elsewhere [29]. Briefly, all participants were at least 40 years of age and were able to complete a survey in one of six languages. Participants were recruited from four neighborhoods in the Boston metropolitan area (Somerville, Dorchester/South Boston, Chinatown, and Malden). To maximize exposure contrast, individuals residing <100 m, 100–500 m, and >1000 m of either Interstate 90 or Interstate 93 were recruited as part of a stratified random sample. To increase the sample size, individuals residing in elderly housing developments in Somerville and Dorchester/South Boston and individuals residing in the same buildings and floors as participants in the stratified random sample in Chinatown were recruited as part of a convenience sample (18% of participants were part of the convenience sample). Of the 704 participants who completed in-home surveys, 455 participants also attended a clinic visit where their blood pressure was measured. Of these participants, 409 were included in the present analysis since they had complete information on UFP exposure, self-reported race/ethnicity, and blood pressure outcomes. A secondary analysis was also conducted on the 205 participants from Somerville and Dorchester/South Boston who attended a second clinic visit approximately five months after their initial clinic visit (mean time between visits = 138 days, sd = 53 days, min = 35 days, max = 364 days). The Tufts University Health Sciences Institutional Review Board approved the study (protocol # 8468; originally approved in 2008; most recent approval June 13 2017). All participants gave written informed consent.

2.1. Demographics and Health Data

During the in-home visit, participants self-reported age, sex, education (less than high school, high school, or more than high school), race/ethnicity (non-Hispanic white, Asian, or other), country of birth, smoker status (never, former, or current), doctor diagnoses of several health conditions (e.g., hypertension, diabetes), air conditioner use, and time spent in five micro-environments (inside home, outside home, school/work, commuting, and other) on a recent workday/weekday and a recent weekend/non-work day (depending on employment status). Participants also reported frequency of fruit and vegetable consumption (less than or at least seven times per week of both fruits and vegetables), fried food consumption (less than or at least once per week), gas stove use in the home (less than or more than half of the days in the month), and annoyance with traffic sound at home (never, sometimes, often, or always). Participants were also asked to show all of their medications to the field staff member and they were surveyed about physical activity (represented here as the natural log number of minutes per week participants engaged in light or moderate physical activity for consistency with previously published work) [26,30].
At each clinic visit, participants’ height, weight, and seated blood pressure were measured. Body mass index (BMI) was calculated as weight (kg)/[height (m)]2. Blood pressure was measured with an automatic blood pressure machine (Model HEM711ACN2; Omron Healthcare, Kyoto, Japan). For SBP and DBP, measurements taken from the left and right arms were averaged. PP was calculated as the difference between SBP and DBP. Participants were classified as hypertensive if they had a measured SBP above 140, a measured DBP above 90, or if they reported taking medications to treat hypertension.

2.2. Exposure Assessment

Methods for estimating annual average PNC adjusted for individual time-activity data on time spent in different micro-environments (TAA-PNC) have been published previously [20,23,31,32,33]. Briefly, air pollution monitoring was conducted with the Tufts Air Pollution Monitoring Laboratory (TAPL-1) between September 2009 and July 2012 along a fixed route in each of Somerville, Dorchester/South Boston, Chinatown, and Malden. TAPL-1 is a retrofitted gasoline-powered Class-C recreational vehicle equipped with rapid-response instruments, including a butanol-based condensation particle counter (TSI, Model 3775; 4-3000 nm) used to measure PNC with one second averaging time. Spatial coordinates were assigned using a Garmin V GPS receiver (manufacturer-specified accuracy: 3–5 m) [31]. These data were used along with spatial and temporal covariates (e.g., distance from residence to nearest highway and major road, wind speed, wind direction, temperature, day of week, highway traffic volume, and highway traffic speed) to develop a model estimating hourly natural log PNC values at participant residences for each hour of the year in which the participant attended the initial CAFEH study visit [23,32]. Hourly residential PNC estimates were then adjusted for infiltration of PNC into residences, air conditioner use, and participant time-activity based on the amount of workday/weekday and non-workday/weekend time participants reported spending in different micro-environments (inside home, outside home, at work, on highway, and other) [20,33]. For participants who attended a second CAFEH clinic visit, micro-environment time-activity data were consistent between the two study visits. Additional details are given in Appendix A.

2.3. Statistical Analysis

We examined the distribution of demographic factors and exposure estimates in our study population as a whole and stratified by race/ethnicity, statin medication use, and diabetes status. For continuous variables, we calculated the mean and standard deviation. For categorical variables, we calculated proportions. We used independent sample t-tests, one-way ANOVA, and chi-square statistics to examine differences by race/ethnicity, statin medication use, and diabetes status.

2.3.1. Conceptual Model

We constructed the directed acyclic graph (DAG) shown in Figure 1 based on an extensive literature review, as detailed in Appendix B. This type of conceptual model makes assumptions explicit and identifies the minimally sufficient set of variables needed to account for confounding of the exposure–outcome relationship [34]. Since UFP exposure was defined largely by participants’ proximity to roadways, we assumed that proximity was accounted for in all models. Based on this assumption and the relationships in Figure 1, there were four possible minimally sufficient sets of covariates: (1) BMI, diet, physical activity, sex, and smoking; (2) BMI, cooking, physical activity, sex, and smoking; (3) cooking, inhalation rate, and smoking; and (4) diet, inhalation rate, and smoking. Since we did not have any direct measure of inhalation rate, we could not use the third or fourth minimally sufficient sets. Additionally, as our main measure for cooking-related exposures was residential use of a gas stove and this seemed like a crude proxy, we prioritized the first minimally sufficient set for all analyses. In modeling and in tests of the DAG, acculturation/migration was represented by a dichotomous variable for nativity (born in the US or not), proximity was represented by distance to the nearest highway, and diet was represented alternatively as fried food consumption and as fruit and vegetable consumption.

2.3.2. Primary Analysis and Sensitivity Analyses

For each of the four outcomes (SBP, DBP, PP, and hypertension), we constructed separate models (linear models for SBP, DBP, and PP; logistic models for hypertension). We first examined the unadjusted associations between natural log-transformed TAA-PNC exposure and each outcome (Model A). Our main model examined the association between natural log-transformed TAA-PNC exposure and each outcome adjusted for covariates in the first minimally sufficient set identified by the DAG (Model B: BMI, sex, smoking, physical activity, and diet as fried food consumption). We tested collinearity using variance inflation factors (all values were <2). We also examined the normality and homoscedasticity of the residuals. For the hypertension model, we examined the Hosmer–Lemeshow goodness-of-fit test (p = 0.91).
To assess the sensitivity of this model to the choice of the proxy for diet, we used two other proxies for diet (fruit and vegetable consumption in Model C; and race/ethnicity in Model D). We also assessed whether age was a residual confounder in Model E. Since age is an exogenous variable, its inclusion would not be expected to introduce confounding. To assess the sensitivity of our analysis to the assumption that proximity was accounted for within the exposure value, we tested the association between natural log-transformed TAA-PNC exposure and each outcome adjusting for the covariates in Model B as well as covariates that would serve as three different sets of proxies for proximity (Model F: age, race/ethnicity (assumed to define acculturation/migration status), and educational attainment; Model G: Model F covariates as well as a variable representing inclusion in the convenience sample; and Model H: Model B covariates as well as annoyance with traffic sound).

2.3.3. Effect Modification

We examined the associations between natural log-transformed TAA-PNC exposure and each outcome stratified (separately) by race/ethnicity, statin medication use, diabetes status, and hypertension status (only for blood pressure models). We considered unadjusted models (Model A), models adjusted for the primary covariates (Model B), and models adjusted for the primary covariates as well as age (Model E). While we considered the p value for interactions, these were largely un-interpretable because the analysis was not adequately powered to detect interactions.

2.3.4. Consistency of Associations over Time

To determine if the associations we found were stable over time, we compared the effect estimates for the association of natural log-transformed TAA-PNC exposure with each outcome at clinic visit one and at clinic visit two (n = 205). We considered unadjusted models (Model A), models adjusted for the primary covariates (Model B), and models that did not assume that proximity status largely defines exposure status (Model F).

3. Results

Of the 409 study participants, the majority were female (59%) and the average age was 62 years (Table 1). Most participants self-identified as non-Hispanic white (n = 178) or Asian (n = 149). There were 82 participants who self-identified as another race/ethnicity, including 36 black and 26 Hispanic participants. While 87% of the white participants were born in the United States, all of the Asian participants and 59% of the participants of other races/ethnicities were born outside of the United States (Table 1). Most of the Asian participants were born in China (n = 126) or Vietnam (n = 13). Compared to participants who identified as white or as another race/ethnicity, Asian participants were least likely to have attained at least a high school education (p < 0.001), had lower mean BMI (p < 0.001 for both comparisons), and higher mean physical activity levels (p < 0.001 for both comparisons). Non-Hispanic white participants consumed more fruits and vegetables than Asian participants or participants of other races/ethnicities (p = 0.011 and p = 0.016, respectively). Asian participants consumed less fried food than non-Hispanic white or participants of other races/ethnicities (p < 0.001 for both comparisons). Smoking rates significantly differed by sex only for the Asian participants (6% of Asian women had ever smoked versus 61% of Asian men).
Blood pressure and hypertension values differed by race/ethnicity, statin medication use, and diabetes status (Table 1). Asian participants had a significantly higher mean SBP and PP than non-Hispanic white participants (p = 0.002 for both comparisons). Asians also had significantly higher mean PP than participants who self-identified with other racial/ethnic groups (p = 0.007). Nevertheless, both Asians and non-Hispanic whites had significantly lower mean DBP than other participants (p = 0.003 and p < 0.001, respectively). Non-Hispanic whites also had a significantly lower prevalence of hypertension compared to either Asians or other participants (p = 0.009 and p = 0.022, respectively). Participants taking statin medications had higher mean SBP, higher mean PP, and a higher hypertension prevalence than participants who were not taking statins (p < 0.001 for all comparisons). Similarly, participants with diabetes had higher mean SBP, higher mean PP, and a higher hypertension prevalence than participants without diabetes (p = 0.003, p < 0.001, and p < 0.001, respectively).
Annual average levels of TAA-PNC were between 9000 and 35,000 particles/cm3. The mean TAA-PNC was 22,000 particles/cm3 while the median TAA-PNC was 23,000 particles/cm3 (interquartile range = 9000 particles/cm3). Mean natural log-transformed TAA-PNC was significantly higher among Asian participants than among non-Hispanic white participants (p < 0.001; Table 1). Exposure levels were also significantly higher among participants taking statins than among those not taking statin medications (p = 0.022). There was no difference in mean exposure level by diabetes status (p = 0.579).
In our main analysis, we found positive, non-significant associations between natural log-transformed TAA-PNC exposure with SBP (β = 5.23, 95% CI = −0.68, 11.14 mmHg per natural log-unit increase), PP (β = 4.27, 95% CI = −0.79, 9.32 mmHg), and hypertension prevalence (OR = 1.81, 95% CI = 0.94, 3.48; Table 2). We found less evidence for an association with DBP (β = 0.96, 95% CI = −2.08, 4.00 mmHg). For reference, the difference in TAA-PNC concentration for a participant 0.5 natural log-units above the mean compared to a participant 0.5 natural log-units below the mean is approximately 21,000 particles/cm3. The results were not sensitive to the choice of the proxy for diet if fruit and vegetable consumption was used, though the associations were attenuated when race/ethnicity was used as a proxy (Table 2; Models C and D). The results were sensitive to the assumption that proximity was accounted for within the exposure value. Additional adjustment for proximity-related variables attenuated the associations, though adjusting for annoyance with traffic sound did not change the primary results (Table 2; Models F–H).

3.1. Effect Modification

We found some evidence that race/ethnicity modified the relationship between natural log-transformed TAA-PNC and blood pressure (Figure 2). Specifically, there seemed to be stronger associations among non-Hispanic white participants than among other participants. For SBP, PP, and hypertension prevalence, these were stronger positive associations while for DBP, there was a stronger inverse association (Figure 2; OR for hypertension among non-Hispanic whites = 3.47, 95% CI = 0.83, 14.5; OR for Asians = 1.09, 95% CI = 0.42, 2.83; OR for participants of other races/ethnicities = 0.52, 95% CI = 0.03, 10.44). Adjustment for age generally made the effect estimates for non-Hispanic whites slightly stronger while making the effect estimates for other participants slightly weaker (results not shown).
We found little evidence of effect modification of the relations of natural log-transformed TAA-PNC with blood pressure by statin medication use (Figure 3). While the SBP and PP point estimates were slightly higher for participants not on statins, the confidence intervals overlapped completely. There was more evidence that having diabetes and, to a lesser extent, not having hypertension strengthened the relation of natural log-transformed TAA-PNC with SBP and PP (Figure 3). No evidence existed for effect modification of the relation of natural log-transformed TAA-PNC with DBP or hypertension for either statin medications (OR for statin users = 3.54; 95% CI = 0.55, 22.94; OR for non-users = 1.28; 95% CI = 0.61, 2.68) or diabetes status (OR for diabetics = 2.29; 95% CI = 0.31, 16.97; OR for non-diabetics = 1.63, 95% CI = 0.79, 3.38). Additional adjustment for age generally did not substantially change any of the conclusions, although it widened the already wide confidence intervals.

3.2. Consistency of Associations over Time

Of the 205 participants who attended two clinic visits approximately five months apart (all of whom resided in Somerville or Dorchester/South Boston), 67% identified as non-Hispanic white and 4% identified as Asian. Mean SBP decreased by 4.5 mmHg (95% CI = 2.4, 6.5 mmHg decrease), mean DBP decreased by 2.8 mmHg (95% CI = 1.6, 4.1 mmHg decrease), and mean PP decreased by 1.6 mmHg (95% CI = 0.2, 3.1 mmHg decrease). We found that the effect estimates for the association of natural log-transformed TAA-PNC with SBP, PP, and hypertension weakened slightly from clinic visit one to clinic visit two (Figure 4; OR for visit one = 10.3, 95% CI = 1.7, 60.9; OR for visit two = 7.0, 95% CI = 1.4, 36.4). Additional adjustment for proximity did not change these trends (results not shown).

4. Discussion

We found that long-term exposure to UFP (measured as TAA-PNC) was positively, though not significantly, associated with SBP, PP, and hypertension prevalence. The observed associations correspond to differences in SBP and PP that are approximately equivalent to differences observed with an increase of 3–9 years of age [35]. The associations were stronger among participants who identified as non-Hispanic white than among participants who identified as Asian or as another race/ethnicity. Additionally, the associations with SBP and PP were stronger among participants with diabetes than among participants without diabetes. They were also slightly stronger among participants without hypertension than among participants with hypertension. We did not find evidence of an association of UFP with DBP overall or among any sub-group.
As expected, our results were consistent with a previous analysis of the association between long-term UFP exposure and hypertension among adults participating in CAFEH (OR = 1.28, 95% CI = 0.81, 2.02 for the previous analysis compared to our result of OR = 1.81, 95% CI = 0.94, 3.48) [25]. The primary difference in these analyses was how hypertension prevalence was defined; in the previous analysis, elevated SBP and DBP measurements were not considered as part of the diagnostic criteria (which also modestly changed the sample size). Additionally, the covariates included in the previously published paper were not chosen based on a conceptual model, such as a DAG [25].
Our results were also generally consistent with the emerging evidence from longitudinal studies that long-term exposure to UFP is associated with cardiovascular impacts. For example, several longitudinal studies have reported positive associations with biomarkers of inflammation and other sub-clinical cardiovascular markers [14,15,16,17]. Additionally, our finding that long-term exposure to UFP was positively associated with hypertension prevalence is similar to the findings from a Canadian cohort that long-term UFP exposure is associated with incident hypertension, even adjusting for PM2.5 and nitrogen dioxide exposure [18].
Nevertheless, our finding in this cross-sectional study that long-term UFP exposure was positively, but not significantly, associated with SBP and PP was not consistent with our previous finding in a prospective study of Puerto Rican adults residing in eastern Massachusetts. In that study, we found that long-term exposure to UFP was not associated with changes in SBP (β = 0.96; 95% CI = − 0.33, 2.25 mmHg) or PP (β = 0.70; 95% CI = − 0.27, 1.67 mmHg) [16]. Our present study is cross-sectional, and exposures were assessed in the year of the clinical examination (rather than exclusively before blood pressure measurements were taken) and as such, the present study cannot address the question of whether UFP exposure could be considered a causal risk factor for changes in blood pressure. The idea that there could be exposure misclassification or a temporal misalignment in the present study was supported in that the effect estimates were attenuated among the sub-set of participants who attended a second clinic visit approximately five months after their primary clinic visit (Figure 4). Some of the instability of the effect estimates could be due to seasonal differences as particle composition can vary with time of year [36,37] and blood pressure tends to be higher when temperatures are cooler (77% of the first clinic visits occurred between October and March while 96% of the second visits occurred between April and September) [38,39]. It is also possible, however, that some of the differences between the previous study and this one reflect differences in the study populations. In the present analysis, only 6% of participants were Hispanic while in the previous analysis, all participants identified as of Puerto Rican descent.
In the present study, we found evidence that the associations with UFP differed among sub-groups. For example, we found somewhat stronger associations of UFP with SBP, PP, and hypertension among non-Hispanic whites than among other participants, despite higher UFP exposures and increased exposure contrast among participants who identified as Asian or as another race/ethnicity. Given the strong spatial segregation of participants in our study areas by race/ethnicity, it is possible that the differences we observed were attributable to other neighborhood social or environmental characteristics rather than to UFP [40,41,42]. This idea was partially supported in that the associations were attenuated when we controlled for race/ethnicity, which in the CAFEH population is strongly associated with both exposure to social and environmental factors and to differences in health status [26]. It is also possible that differences in general health status modified the associations of UFP with the health outcomes. In previous studies, as in our study, co-morbidities and medication use affected the strength of associations [16,17]. It was a strength of our present study that we were able to recruit a diverse population from several communities in Boston. This allowed us to consider the relationship of UFP with BP outcomes among sub-sets of the population defined by race/ethnicity and by health status.
Another strength of our study was our use of a conceptual model to identify a minimally sufficient adjustment set of covariates. We were able to explicitly state and test our assumptions about the factors that confound the relationship between UFP and BP. For example, if we had included a direct path from sex to cooking (rather than only an indirect path), the minimally sufficient adjustment set of covariates would not change. Similarly, if we included a direct path from age to education to reflect changes in educational attainment patterns over time, the minimally sufficient adjustment set would not change. Nevertheless, if we had included a direct path from age to cooking, age should be included in the minimally sufficient set. As Model E results show, this would slightly attenuate the associations of UFP with SBP and PP.
The use of the conceptual model also allowed us to test the sensitivity of our analyses to the choice of proxy for variables represented within the minimally sufficient adjustment set. For example, we compared the results using fried food consumption and fruit and vegetable consumption as proxies for diet and found that the results were robust against choice of dietary variable. This was important because these two dietary components represented different dietary patterns within our study population (e.g., non-Hispanic whites consumed more fried food and fruits and vegetables, Asians consumed fewer fried foods and fruits and vegetables, and participants of other races/ethnicities consumed more fried foods but fewer fruits and vegetables). Similarly, we were able to quantitatively assess our assumption that since UFP exposure was assessed as a function of distance to roadways, additional adjustment for proximity would result in over-adjusted models. We found that adjustment for annoyance from traffic sound did not meaningfully change the results and that adjustment for other proxies for proximity resulted in attenuated associations, as would be expected if the models were over-adjusted.
Our exposure assessment strategy had both strengths and limitations. Our UFP model was finely resolved in space (20-m resolution) and time (hourly resolution). We further adjusted our exposure estimates for time spent in different micro-environments, avoiding the bias introduced by assuming residential exposure concentrations are representative of long-term UFP exposure [20]. While we did not control for short-term UFP exposures and we did not account for short-term changes in blood pressure, previous work has suggested that controlling for short-term UFP exposure is unlikely to meaningfully change the results [15] and our group has found that daily average UFP exposure is not associated with SBP or PP (though it is associated with DBP) in the CAFEH population [43]. There could still be concerns about the exposure window we used. While UFP exposures were assigned for the calendar year of the study visit, a more relevant time window would include only time preceding the study visit. Additionally, to the extent that exposure misclassification varied by neighborhood or by time period, our analyses of effect modification by race/ethnicity would be strongly impacted due to the spatial segregation of participants and the fact that we collected data from different neighborhoods in different years. Finally, although other air pollutants could act independently of or jointly with UFP to affect blood pressure, we were unable to adjust for exposure to other air pollutants. Beyond the exposure assessment limitations, the major limitations of our study include the cross-sectional design, the fairly small sample size, and the multiple comparisons we made increasing the likelihood of a Type I error. The associations we observed with UFP should not be interpreted causally unless future work addressing the limitations confirms our findings.

5. Conclusions

Overall, we found evidence that UFP was positively associated with SBP, PP, and hypertension prevalence. While our study adds to the growing body of literature on the association between long-term exposure to UFP and cardiovascular outcomes, future longitudinal research should consider the impact of long-term UFP exposure on incident hypertension and changes in BP over time in diverse populations.

Authors contributions

Conceptualization, J.L.D. and D.B.; Formal analysis, L.C. and S.B.; Funding acquisition, L.C. and D.B.; Investigation, A.P.P., K.L., J.L.D. and D.B.; Methodology, L.C., M.W., J.L.D. and D.B.; Project administration, D.B.; Resources, A.P.P., K.L., J.L.D. and D.B.; Software, L.C.; Supervision, M.W., J.L. D. and D.B.; Writing—original draft, L.C.; and Writing—review and editing, L.C., S.B., M.W., A.P.P., K.L., J.L.D. and D.B.

Funding

Funding was provided by the NIEHS (ES015462), the EPA (FP-917349 and FP-917203), the NSF (0966093), the NHLBI (1T32HL125232), a Philanthropic Educational Organization Esther Garrett Edgerton Endowed Scholar Award, Tufts University Vice Provost for Research, Tufts University Department of Civil and Environmental Engineering, Tufts Institute of the Environment, Jonathan M. Tisch College of Civic Life, and the Tisch College Community Research Center at Tufts.

Acknowledgements

We would like to acknowledge the Community Assessment of Freeway and Health team including Wig Zamore. Noelle Henderson provided assistance with the literature review for the directed acyclic graph. We would like to thank all of the participants in the study.

Conflicts of interest

The authors declare no conflict of interest. The founding sponsors had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results. The contents of this article do not necessarily reflect the views of HEI, or its sponsors, nor do they necessarily reflect the views and policies of the EPA or motor vehicle and engine manufacturers.

Appendix A: Exposure Assessment

In each of Somerville, Dorchester/South Boston, Chinatown, and Malden, we drove the Tufts Air Pollution Monitoring Laboratory (TAPL-1) around a fixed route 2–6 times/day on 35 days in the same year in which we collected health information from participants. Monitoring occurred in all four seasons, on all days of the week, and at different times of the day (04:00–22:00). Data quality assurance and control steps have been described previously [23,31,44]. Community-specific multivariate regression models predicting log-transformed hourly PNC were developed using spatial (e.g., side of and distance to highway) and temporal (e.g., wind speed and direction) covariates. Model performance was adequate (model adjusted-R2: Somerville = 0.42, Dorchester/South Boston = 0.35, Chinatown = 0.23, and Malden = 0.31) and stable under leave-one-day-out cross-validation [23].
Using the community-specific PNC models, hourly residential ambient PNC were estimated for each participant for every hour of the year [24]. These estimates were then adjusted for the amount of time individual participants reported spending on a workday/weekday and a non-workday/weekend day in five micro-environments (inside home, outside home, at work, on highway, and other). For hours spent inside the home when local meteorological stations reported ambient temperatures exceeding 21 °C (70 °F), the hourly residential value was modified to account for air conditioner use (window or central air conditioners, depending on survey data). For hours spent at work in occupations that included traffic-related air pollution exposure (e.g., bus drivers), the hourly residential ambient PNC value was replaced with the mean hourly residential PNC of all participants residing ≤ 50 m from the highway. For hours spent at work in occupations without substantial traffic-related air pollution exposure (e.g., nurses) and for time spent in “other” micro-environments, the hourly residential ambient PNC value was replaced with the mean hourly residential PNC of all participants residing > 1000 m from the highway. For time spent on the highway, we replaced the hourly residential PNC with the on-highway concentration predicted by the Somerville model. Individual time-activity-adjusted (TAA) PNC values for each hour were averaged to assign annual TAA-PNC for each participant [20,24].

Appendix B: Directed Acyclic Graph Justification

To develop the directed acyclic graph (DAG), we first identified consensus statements [2,45,46,47,48], meta-analyses [49,50,51], systematic reviews [52,53], reviews [54,55,56,57,58,59,60,61], and longitudinal cohort studies [62,63,64,65,66,67] assessing determinants of blood pressure in adults. We also conducted a PubMed search in January 2017: ((((“blood pressure”[Title]) OR “hypertension”[Title]) AND ((“meta analysis”[Publication Type] OR “review”[Publication Type]))) AND (“risk factor”[Title] OR “risk factors”[Title] OR “determinant”[Title] OR “determinants”[Title] OR “effect of”[Title] OR “influence of”[Title] OR “impact of”[Title])) AND english[Language] Filters activated: Humans, Adult: 19+ years. This search resulted in 94 titles. From a review of the abstracts in Abstrackr [68], we kept 30 articles [69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,87,88,89,90,91,92,93,94,95,96,97]. Articles were excluded if they focused primarily on early life or genetic risk factors, blood pressure or hypertension as an exposure, hypertension treatment or management, blood pressure in children or adolescents, pregnancy-related hypertension, or treatment-resistant hypertension.
From these articles, we identified 21 risk factors for hypertension and high blood pressure (acculturation/migration, age, alcohol, BMI, caffeine/coffee/cocoa, DASH diet, diabetes, dietary fiber/protein supplementation, education/SES, estradiol, glucocorticoids, physical activity, potassium, race, racial discrimination, sex, smoking, sodium intake, statins, stress, and vitamin C). We examined the pairwise relationships between UFP and each of the identified risk factors [19,98,99,100,101,102,103,104,105,106,107]. We prioritized evidence from meta-analyses, systematic reviews, reviews, or consensus statements that addressed direct and indirect (within two steps) determinants of personal exposure to UFP. If these types of articles were not available, we considered evidence from experimental studies, cohort studies, and other study designs (in that order). We used the same evidence prioritization method to assess the relationships among potential covariates [90,108,109,110,111,112,113,114,115,116,117,118,119,120,121,122,123,124,125,126,127,128,129,130,131]. We also considered relationships that were introduced by the study design. For example, convenience sampling was done in specific retirement communities so age was considered an ancestor of UFP exposure. Based on all of the pairwise relationships, we built the DAG in DAGitty v2.3 [132]. The code for the DAG is below.
Of the 99 testable conditional independencies suggested by the DAG, 69 held within our dataset when using the variable for fried food consumption as a proxy for diet and 70 held when using the variable for fruit and vegetable consumption. The most problematic covariate was the proxy for cooking (gas stove use) which is due partially to the fact that none of the Somerville participants responded to this question. Statements of conditional independence with UFP did not hold in our dataset for race, education, and convenience sampling (each with three different sets of covariates) or for acculturation/migration, age, and fruit/vegetable intake (each with one set of covariates).
DAG Code:
Acculturation%2Fmigration 1 @−0.848, −2.923
Age 1 @−4.724, −2.738
BMI 1 @1.013, −1.037
Blood%20pressure O @3.740, 5.031
Convenience%20sample 1 @−4.594, 1.461
Cooking 1 @−2.922, 3.995
Diet 1 @−1.535, 3.699
Education%2FSES 1 @0.966, −2.664
Inhalation%20rate 1 @−3.610, 3.033
Physical%20activity 1 @−1.998, 1.109
Proximity 1 @−4.866, 3.273
Race 1 @−3.302, −1.888
Sex 1 @2.068, −0.630
Smoking 1 @−3.586, 0.813
UFP E @−4.700, 4.920
Acculturation%2Fmigration BMI Blood%20pressure Diet Physical%20activity Proximity Smoking
Age Acculturation%2Fmigration BMI Blood%20pressure Convenience%20sample Diet Physical%20activity Smoking
BMI Blood%20pressure Inhalation%20rate
Convenience%20sample Proximity
Cooking UFP
Diet BMI Blood%20pressure Cooking
Education%2FSES BMI Blood%20pressure Diet Physical%20activity Proximity Smoking
Inhalation%20rate UFP
Physical%20activity BMI Blood%20pressure Inhalation%20rate
Proximity UFP
Race Acculturation%2Fmigration BMI Blood%20pressure Convenience%20sample Diet Education%2FSES Physical%20activity Proximity Smoking
Sex BMI Blood%20pressure Diet Inhalation%20rate Physical%20activity Smoking
Smoking Blood%20pressure Inhalation%20rate UFP
UFP Blood%20pressure

References

  1. Cohen, A.J.; Brauer, M.; Burnett, R.; Anderson, H.R.; Frostad, J.; Estep, K.; Balakrishnan, K.; Brunekreef, B.; Dandona, L.; Dandona, R.; et al. Estimates and 25-year trends of the global burden of disease attributable to ambient air pollution: An analysis of data from the Global Burden of Diseases Study 2015. The Lancet 2017, 389, 1907–1918. [Google Scholar] [CrossRef]
  2. Brook, R.D.; Rajagopalan, S.; Pope, C.A.; Brook, J.R.; Bhatnagar, A.; Diez-Roux, A.V.; Holguin, F.; Hong, Y.; Luepker, R.V.; Mittleman, M.A.; et al. American Heart Association Council on Epidemiology and Prevention, Council on the Kidney in Cardiovascular Disease, and Council on Nutrition, Physical Activity and Metabolism Particulate matter air pollution and cardiovascular disease: An update to the scientific statement from the American Heart Association. Circulation 2010, 121, 2331–2378. [Google Scholar] [CrossRef] [PubMed]
  3. Miller, K.A.; Siscovick, D.S.; Sheppard, L.; Shepherd, K.; Sullivan, J.H.; Anderson, G.L.; Kaufman, J.D. Long-Term Exposure to Air Pollution and Incidence of Cardiovascular Events in Women. N. Engl. J. Med. 2007, 356, 447–458. [Google Scholar] [CrossRef] [PubMed]
  4. Crouse, D.L.; Peters, P.A.; van Donkelaar, A.; Goldberg, M.S.; Villeneuve, P.J.; Brion, O.; Khan, S.; Atari, D.O.; Jerrett, M.; Pope, C.A.; et al. Risk of nonaccidental and cardiovascular mortality in relation to long-term exposure to low concentrations of fine particulate matter: A Canadian national-level cohort study. Environ. Health Perspect. 2012, 120, 708–714. [Google Scholar] [CrossRef] [PubMed]
  5. Fuks, K.; Moebus, S.; Hertel, S.; Viehmann, A.; Nonnemacher, M.; Dragano, N.; Möhlenkamp, S.; Jakobs, H.; Kessler, C.; Erbel, R.; et al. Long-term urban particulate air pollution, traffic noise, and arterial blood pressure. Environ. Health Perspect. 2011, 119, 1706–1711. [Google Scholar] [CrossRef] [PubMed]
  6. Zhang, Z.; Laden, F.; Forman, J.P.; Hart, J.E. Long-Term Exposure to Particulate Matter and Self-Reported Hypertension: A Prospective Analysis in the Nurses’ Health Study. Environ. Health Perspect. 2016, 124, 1414–1420. [Google Scholar] [CrossRef] [PubMed]
  7. Nemmar, A.; Hoet, P.H.M.; Vanquickenborne, B.; Dinsdale, D.; Thomeer, M.; Hoylaerts, M.F.; Vanbilloen, H.; Mortelmans, L.; Nemery, B. Passage of Inhaled Particles into the Blood Circulation in Humans. Circulation 2002, 105, 411–414. [Google Scholar] [CrossRef] [PubMed]
  8. Kreyling, W.G.; Semmler-Behnke, M.; Möller, W. Ultrafine particle-lung interactions: Does size matter? J. Aerosol Med. Off. J. Int. Soc. Aerosols Med. 2006, 19, 74–83. [Google Scholar] [CrossRef] [PubMed]
  9. Oberdörster, G. Pulmonary effects of inhaled ultrafine particles. Int. Arch. Occup. Environ. Health 2000, 74, 1–8. [Google Scholar] [CrossRef]
  10. Delfino, R.J.; Sioutas, C.; Malik, S. Potential Role of Ultrafine Particles in Associations between Airborne Particle Mass and Cardiovascular Health. Environ. Health Perspect. 2005, 113, 934–946. [Google Scholar] [CrossRef] [PubMed]
  11. Araujo, J.A.; Barajas, B.; Kleinman, M.; Wang, X.; Bennett, B.J.; Gong, K.W.; Navab, M.; Harkema, J.; Sioutas, C.; Lusis, A.J.; et al. Ambient particulate pollutants in the ultrafine range promote early atherosclerosis and systemic oxidative stress. Circ. Res. 2008, 102, 589–596. [Google Scholar] [CrossRef] [PubMed]
  12. Lawal, A.O. Air particulate matter induced oxidative stress and inflammation in cardiovascular disease and atherosclerosis: The role of Nrf2 and AhR-mediated pathways. Toxicol. Lett. 2017, 270, 88–95. [Google Scholar] [CrossRef] [PubMed]
  13. Ostro, B.; Hu, J.; Goldberg, D.; Reynolds, P.; Hertz, A.; Bernstein, L.; Kleeman, M.J. Associations of mortality with long-term exposures to fine and ultrafine particles, species and sources: Results from the California Teachers Study Cohort. Environ. Health Perspect. 2015, 123, 549–556. [Google Scholar] [CrossRef] [PubMed]
  14. Aguilera, I.; Dratva, J.; Caviezel, S.; Burdet, L.; de Groot, E.; Ducret-Stich, R.E.; Eeftens, M.; Keidel, D.; Meier, R.; Perez, L.; et al. Particulate Matter and Subclinical Atherosclerosis: Associations between Different Particle Sizes and Sources with Carotid Intima-Media Thickness in the SAPALDIA Study. Environ. Health Perspect. 2016. [Google Scholar] [CrossRef] [PubMed]
  15. Viehmann, A.; Hertel, S.; Fuks, K.; Eisele, L.; Moebus, S.; Möhlenkamp, S.; Nonnemacher, M.; Jakobs, H.; Erbel, R.; Jöckel, K.-H.; et al. Long-term residential exposure to urban air pollution, and repeated measures of systemic blood markers of inflammation and coagulation. Occup. Environ. Med. 2015, 72, 656–663. [Google Scholar] [CrossRef] [PubMed]
  16. Corlin, L.; Woodin, M.; Hart, J.; Simon, M.; Gute, D.; Tucker, K.; Durant, J.; Brugge, D. Longitudinal Associations of Long-term Exposure to Ultrafine Particles with Blood Pressure and Systemic Inflammation in Puerto Rican Adults. Environ. Health 2018, 17. [Google Scholar] [CrossRef] [PubMed]
  17. Pilz, V.; Wolf, K.; Breitner, S.; Rückerl, R.; Koenig, W.; Rathmann, W.; Cyrys, J.; Peters, A.; Schneider, A. C-reactive protein (CRP) and long-term air pollution with a focus on ultrafine particles. Int. J. Hyg. Environ. Health 2018. [Google Scholar] [CrossRef] [PubMed]
  18. Bai, L.; Chen, H.; Hatzopoulou, M.; Jerrett, M.; Kwong, J.C.; Burnett, R.T.; van Donkelaar, A.; Copes, R.; Martin, R.V.; Van Ryswyk, K.; et al. Exposure to Ambient Ultrafine Particles and Nitrogen Dioxide and Incident Hypertension and Diabetes. Epidemiology 2018, 29, 323. [Google Scholar] [CrossRef] [PubMed]
  19. Karner, A.A.; Eisinger, D.S.; Niemeier, D.A. Near-roadway air quality: Synthesizing the findings from real-world data. Environ. Sci. Technol. 2010, 44, 5334–5344. [Google Scholar] [CrossRef] [PubMed]
  20. Lane, K.J.; Levy, J.I.; Scammell, M.K.; Patton, A.P.; Durant, J.L.; Mwamburi, M.; Zamore, W.; Brugge, D. Effect of time-activity adjustment on exposure assessment for traffic-related ultrafine particles. J. Expos. Sci. Environ. Epid. 2015, 25, 506–516. [Google Scholar] [CrossRef] [PubMed]
  21. Baldauf, R.W.; Devlin, R.B.; Gehr, P.; Giannelli, R.; Hassett-Sipple, B.; Jung, H.; Martini, G.; McDonald, J.; Sacks, J.D.; Walker, K. Ultrafine Particle Metrics and Research Considerations: Review of the 2015 UFP Workshop. Int. J. Environ. Res. Public. Health 2016, 13. [Google Scholar] [CrossRef] [PubMed]
  22. HEI Understanding the Health Effects of Ambient UFP: HEI Review Panel on UFP. 2013. Available online: https://www.healtheffects.org/system/files/Perspectives3.pdf (accessed on 14 September 2018).
  23. Patton, A.P.; Zamore, W.; Naumova, E.N.; Levy, J.I.; Brugge, D.; Durant, J.L. Transferability and generalizability of regression models of ultrafine particles in urban neighborhoods in the Boston area. Environ. Sci. Technol. 2015, 49, 6051–6060. [Google Scholar] [CrossRef] [PubMed]
  24. Lane, K.J.; Levy, J.I.; Scammell, M.K.; Peters, J.L.; Patton, A.P.; Reisner, E.; Lowe, L.; Zamore, W.; Durant, J.L.; Brugge, D. Association of modeled long-term personal exposure to ultrafine particles with inflammatory and coagulation biomarkers. Environ. Int. 2016, 92–93, 173–182. [Google Scholar] [CrossRef] [PubMed]
  25. Li, Y.; Lane, K.J.; Corlin, L.; Patton, A.P.; Durant, J.L.; Thanikachalam, M.; Woodin, M.; Wang, M.; Brugge, D. Association of Long-Term Near-Highway Exposure to Ultrafine Particles with Cardiovascular Diseases, Diabetes and Hypertension. Int. J. Environ. Res. Public. Health 2017, 14, 461. [Google Scholar] [CrossRef] [PubMed]
  26. Corlin, L.; Woodin, M.; Thanikachalam, M.; Lowe, L.; Brugge, D. Evidence for the healthy immigrant effect in older Chinese immigrants: A cross-sectional study. BMC Public Health 2014, 14, 603. [Google Scholar] [CrossRef] [PubMed]
  27. Miyata, R.; Hiraiwa, K.; Cheng, J.C.; Bai, N.; Vincent, R.; Francis, G.A.; Sin, D.D.; Van Eeden, S.F. Statins attenuate the development of atherosclerosis and endothelial dysfunction induced by exposure to urban particulate matter (PM10). Toxicol. Appl. Pharmacol. 2013, 272, 1–11. [Google Scholar] [CrossRef] [PubMed]
  28. Schwartz, J.; Park, S.K.; O’Neill, M.S.; Vokonas, P.S.; Sparrow, D.; Weiss, S.; Kelsey, K. Glutathione-S-transferase M1, obesity, statins, and autonomic effects of particles: Gene-by-drug-by-environment interaction. Am. J. Respir. Crit. Care Med. 2005, 172, 1529–1533. [Google Scholar] [CrossRef] [PubMed]
  29. Fuller, C.H.; Patton, A.P.; Lane, K.; Laws, M.B.; Marden, A.; Carrasco, E.; Spengler, J.; Mwamburi, M.; Zamore, W.; Durant, J.L.; et al. A community participatory study of cardiovascular health and exposure to near-highway air pollution: Study design and methods. Rev. Environ. Health 2013, 28, 21–35. [Google Scholar] [CrossRef] [PubMed]
  30. NHIS Data, Questionnaires and Related Documentation. Available online: http://www.cdc.gov/nchs/nhis/data-questionnaires-documentation.htm (accessed on 30 November 2016).
  31. Padró-Martínez, L.T.; Patton, A.P.; Trull, J.B.; Zamore, W.; Brugge, D.; Durant, J.L. Mobile monitoring of particle number concentration and other traffic-related air pollutants in a near-highway neighborhood over the course of a year. Atmos. Environ. 2012, 61, 253–264. [Google Scholar] [CrossRef] [PubMed]
  32. Patton, A.P.; Collins, C.; Naumova, E.N.; Zamore, W.; Brugge, D.; Durant, J.L. An hourly regression model for ultrafine particles in a near-highway urban area. Environ. Sci. Technol. 2014, 48, 3272–3280. [Google Scholar] [CrossRef] [PubMed]
  33. Lane, K.J.; Kangsen Scammell, M.; Levy, J.I.; Fuller, C.H.; Parambi, R.; Zamore, W.; Mwamburi, M.; Brugge, D. Positional error and time-activity patterns in near-highway proximity studies: An exposure misclassification analysis. Environ. Health 2013, 12, 75. [Google Scholar] [CrossRef] [PubMed]
  34. Greenland, S.; Pearl, J.; Robins, J.M. Causal diagrams for epidemiologic research. Epidemiology 1999, 10, 37–48. [Google Scholar] [CrossRef] [PubMed]
  35. Franklin, S.S.; Gustin, W.; Wong, N.D.; Larson, M.G.; Weber, M.A.; Kannel, W.B.; Levy, D. Hemodynamic Patterns of Age-Related Changes in Blood Pressure. Circulation 1997, 96, 308–315. [Google Scholar] [CrossRef] [PubMed]
  36. Daher, N.; Hasheminassab, S.; Shafer, M.; Schauer, J.; Sioutas, C. Seasonal and spatial variability in chemical composition and mass closure of ambient ultrafine particles in the megacity of Los Angeles. Environ. Sci. Process. Impacts 2013, 15, 283–295. [Google Scholar] [CrossRef] [PubMed]
  37. Saffari, A.; Daher, N.; Shafer, M.M.; Schauer, J.J.; Sioutas, C. Seasonal and spatial variation of trace elements and metals in quasi-ultrafine (PM0.25) particles in the Los Angeles metropolitan area and characterization of their sources. Environ. Pollut. 2013, 181, 14–23. [Google Scholar] [CrossRef] [PubMed]
  38. Woodhouse, P.R.; Khaw, K.T.; Plummer, M. Seasonal variation of blood pressure and its relationship to ambient temperature in an elderly population. J. Hypertens. 1993, 11, 1267–1274. [Google Scholar] [CrossRef] [PubMed]
  39. Modesti, P.A. Season, temperature and blood pressure: A complex interaction. Eur. J. Intern. Med. 2013, 24, 604–607. [Google Scholar] [CrossRef] [PubMed]
  40. Mujahid, M.S.; Diez Roux, A.V.; Morenoff, J.D.; Raghunathan, T.E.; Cooper, R.S.; Ni, H.; Shea, S. Neighborhood characteristics and hypertension. Epidemiol. Camb. Mass 2008, 19, 590–598. [Google Scholar] [CrossRef] [PubMed]
  41. Dubowitz, T.; Ghosh-Dastidar, M.; Eibner, C.; Slaughter, M.E.; Fernandes, M.; Whitsel, E.A.; Bird, C.E.; Jewell, A.; Margolis, K.L.; Li, W.; et al. The Women’s Health Initiative: The Food Environment, Neighborhood Socioeconomic Status, BMI, and Blood Pressure. Obesity 2012, 20, 862–871. [Google Scholar] [CrossRef] [PubMed]
  42. Li, F.; Harmer, P.; Cardinal, B.J.; Vongjaturapat, N. Built environment and changes in blood pressure in middle aged and older adults. Prev. Med. 2009, 48, 237–241. [Google Scholar] [CrossRef] [PubMed]
  43. Chung, M.; Wang, D.D.; Rizzo, A.M.; Gachette, D.; Delnord, M.; Parambi, R.; Kang, C.-M.; Brugge, D. Association of PNC, BC, and PM2.5 measured at a central monitoring site with blood pressure in a predominantly near highway population. Int. J. Environ. Res. Public. Health 2015, 12, 2765–2780. [Google Scholar] [CrossRef] [PubMed]
  44. Patton, A.P.; Perkins, J.; Zamore, W.; Levy, J.I.; Brugge, D.; Durant, J.L. Spatial and temporal differences in traffic-related air pollution in three urban neighborhoods near an interstate highway. Atmos. Environ. 2014, 99, 309–321. [Google Scholar] [CrossRef] [PubMed]
  45. Aronow, W.S.; Fleg, J.L.; Pepine, C.J.; Artinian, N.T.; Bakris, G.; Brown, A.S.; Ferdinand, K.C.; Ann Forciea, M.; Frishman, W.H.; Jaigobin, C.; et al. ACCF/AHA 2011 expert consensus document on hypertension in the elderly: A report of the American College of Cardiology Foundation Task Force on Clinical Expert Consensus Documents developed in collaboration with the American Academy of Neurology, American Geriatrics Society, American Society for Preventive Cardiology, American Society of Hypertension, American Society of Nephrology, Association of Black Cardiologists, and European Society of Hypertension. J. Am. Soc. Hypertens. JASH 2011, 5, 259–352. [Google Scholar] [CrossRef] [PubMed]
  46. Appel, L.J.; Brands, M.W.; Daniels, S.R.; Karanja, N.; Elmer, P.J.; Sacks, F.M. American Heart Association Dietary approaches to prevent and treat hypertension: A scientific statement from the American Heart Association. Hypertens. Dallas Tex 1979 2006, 47, 296–308. [Google Scholar] [CrossRef]
  47. Havranek, E.P.; Mujahid, M.S.; Barr, D.A.; Blair, I.V.; Cohen, M.S.; Cruz-Flores, S.; Davey-Smith, G.; Dennison-Himmelfarb, C.R.; Lauer, M.S.; Lockwood, D.W.; et al. Social Determinants of Risk and Outcomes for Cardiovascular Disease: A Scientific Statement from the American Heart Association. Circulation 2015, 132, 873–898. [Google Scholar] [CrossRef] [PubMed]
  48. Whelton, P.K.; He, J.; Appel, L.J.; Cutler, J.A.; Havas, S.; Kotchen, T.A.; Roccella, E.J.; Stout, R.; Vallbona, C.; Winston, M.C.; et al. National High Blood Pressure Education Program Coordinating Committee Primary prevention of hypertension: Clinical and public health advisory from The National High Blood Pressure Education Program. JAMA 2002, 288, 1882–1888. [Google Scholar] [CrossRef] [PubMed]
  49. Guh, D.P.; Zhang, W.; Bansback, N.; Amarsi, Z.; Birmingham, C.L.; Anis, A.H. The incidence of co-morbidities related to obesity and overweight: A systematic review and meta-analysis. BMC Public Health 2009, 9, 88. [Google Scholar] [CrossRef] [PubMed]
  50. Strazzullo, P.; Kerry, S.M.; Barbato, A.; Versiero, M.; D’Elia, L.; Cappuccio, F.P. Do statins reduce blood pressure? A meta-analysis of randomized, controlled trials. Hypertens. Dallas Tex 1979 2007, 49, 792–798. [Google Scholar] [CrossRef] [PubMed]
  51. Linneberg, A.; Jacobsen, R.K.; Skaaby, T.; Taylor, A.E.; Fluharty, M.E.; Jeppesen, J.L.; Bjorngaard, J.H.; Åsvold, B.O.; Gabrielsen, M.E.; Campbell, A.; et al. Effect of Smoking on Blood Pressure and Resting Heart Rate: A Mendelian Randomization Meta-Analysis in the CARTA Consortium. Circ. Cardiovasc. Genet. 2015, 8, 832–841. [Google Scholar] [CrossRef] [PubMed]
  52. Dolezsar, C.M.; McGrath, J.J.; Herzig, A.J.M.; Miller, S.B. Perceived racial discrimination and hypertension: A comprehensive systematic review. Health Psychol. Off. J. Div. Health Psychol. Am. Psychol. Assoc. 2014, 33, 20–34. [Google Scholar] [CrossRef] [PubMed]
  53. Kurian, A.K.; Cardarelli, K.M. Racial and ethnic differences in cardiovascular disease risk factors: A systematic review. Ethn. Dis. 2007, 17, 143–152. [Google Scholar] [PubMed]
  54. Freis, E.D. Age, race, sex and other indices of risk in hypertension. Am. J. Med. 1973, 55, 275–280. [Google Scholar] [CrossRef]
  55. Gillum, R.F. Pathophysiology of hypertension in blacks and whites. A review of the basis of racial blood pressure differences. Hypertension 1979, 1, 468–475. [Google Scholar] [CrossRef] [PubMed]
  56. Winkleby, M.A.; Jatulis, D.E.; Frank, E.; Fortmann, S.P. Socioeconomic status and health: How education, income, and occupation contribute to risk factors for cardiovascular disease. Am. J. Public Health 1992, 82, 816–820. [Google Scholar] [CrossRef] [PubMed]
  57. Kaplan, G.A.; Keil, J.E. Socioeconomic factors and cardiovascular disease: A review of the literature. Circulation 1993, 88, 1973–1998. [Google Scholar] [CrossRef] [PubMed]
  58. Sowers, J.R.; Epstein, M. Diabetes mellitus and associated hypertension, vascular disease, and nephropathy. An update. Hypertens. Dallas Tex 1979 1995, 26, 869–879. [Google Scholar]
  59. Virdis, A.; Giannarelli, C.; Neves, M.F.; Taddei, S.; Ghiadoni, L. Cigarette smoking and hypertension. Curr. Pharm. Des. 2010, 16, 2518–2525. [Google Scholar] [CrossRef] [PubMed]
  60. Narkiewicz, K.; Kjeldsen, S.E.; Hedner, T. Is smoking a causative factor of hypertension? Blood Press. 2005, 14, 69–71. [Google Scholar] [CrossRef] [PubMed]
  61. Rosenthal, T. The effect of migration on hypertension and other cardiovascular risk factors: A review. J. Am. Soc. Hypertens. JASH 2014, 8, 171–191. [Google Scholar] [CrossRef] [PubMed]
  62. Vasan, R.S.; Larson, M.G.; Leip, E.P.; Kannel, W.B.; Levy, D. Assessment of frequency of progression to hypertension in non-hypertensive participants in the Framingham Heart Study: A cohort study. The Lancet 2001, 358, 1682–1686. [Google Scholar] [CrossRef]
  63. Forman, J.P.; Stampfer, M.J.; Curhan, G.C. Diet and lifestyle risk factors associated with incident hypertension in women. JAMA 2009, 302, 401–411. [Google Scholar] [CrossRef] [PubMed]
  64. Conen, D.; Glynn, R.J.; Ridker, P.M.; Buring, J.E.; Albert, M.A. Socioeconomic status, blood pressure progression, and incident hypertension in a prospective cohort of female health professionals. Eur. Heart J. 2009, 30, 1378–1384. [Google Scholar] [CrossRef] [PubMed]
  65. Bowman, T.S.; Gaziano, J.M.; Buring, J.E.; Sesso, H.D. A prospective study of cigarette smoking and risk of incident hypertension in women. J. Am. Coll. Cardiol. 2007, 50, 2085–2092. [Google Scholar] [CrossRef] [PubMed]
  66. Niskanen, L.; Laaksonen, D.E.; Nyyssönen, K.; Punnonen, K.; Valkonen, V.-P.; Fuentes, R.; Tuomainen, T.-P.; Salonen, R.; Salonen, J.T. Inflammation, abdominal obesity, and smoking as predictors of hypertension. Hypertens. Dallas Tex 1979 2004, 44, 859–865. [Google Scholar] [CrossRef] [PubMed]
  67. Crump, C.; Sundquist, J.; Winkleby, M.A.; Sundquist, K. Interactive Effects of Physical Fitness and Body Mass Index on the Risk of Hypertension. JAMA Intern. Med. 2016, 176, 210–216. [Google Scholar] [CrossRef] [PubMed]
  68. Wallace, B.C.; Small, K.; Brodley, C.E.; Lau, J.; Trikalinos, T.A. Deploying an interactive machine learning system in an evidence-based practice center: Abstrackr. In Proceedings of the 2nd ACM SIGHIT International Health Informatics Symposium (IHI), Miami, FL, USA, 28–30 January 2012. [Google Scholar]
  69. Abrams, J.; Vela, B.S.; Coultas, D.B.; Samaan, S.A.; Malhotra, D.; Roche, R.J. Coronary risk factors and their modification: Lipids, smoking, hypertension, estrogen, and the elderly. Curr. Probl. Cardiol. 1995, 20, 533–610. [Google Scholar] [CrossRef]
  70. Alam, S.; Johnson, A.G. A meta-analysis of randomised controlled trials (RCT) among healthy normotensive and essential hypertensive elderly patients to determine the effect of high salt (NaCl) diet of blood pressure. J. Hum. Hypertens. 1999, 13, 367–374. [Google Scholar] [CrossRef] [PubMed]
  71. Bai, Z.; Chang, J.; Chen, C.; Li, P.; Yang, K.; Chi, I. Investigating the effect of transcendental meditation on blood pressure: A systematic review and meta-analysis. J. Hum. Hypertens. 2015, 29, 653–662. [Google Scholar] [CrossRef] [PubMed]
  72. Bønaa, K. Epidemiological and intervention studies on the effect of marine polyunsaturated fatty acids on blood pressure. J. Intern. Med. Suppl. 1989, 731, 105–110. [Google Scholar] [CrossRef] [PubMed]
  73. Cornelissen, V.A.; Fagard, R.H. Effect of resistance training on resting blood pressure: A meta-analysis of randomized controlled trials. J. Hypertens. 2005, 23, 251–259. [Google Scholar] [CrossRef] [PubMed]
  74. Cornelissen, V.A.; Fagard, R.H.; Coeckelberghs, E.; Vanhees, L. Impact of resistance training on blood pressure and other cardiovascular risk factors: A meta-analysis of randomized, controlled trials. Hypertens. Dallas Tex 1979 2011, 58, 950–958. [Google Scholar] [CrossRef] [PubMed]
  75. Gasperin, D.; Netuveli, G.; Dias-da-Costa, J.S.; Pattussi, M.P. Effect of psychological stress on blood pressure increase: A meta-analysis of cohort studies. Cad. Saude Publica 2009, 25, 715–726. [Google Scholar] [CrossRef] [PubMed]
  76. Ha, V.; Sievenpiper, J.L.; de Souza, R.J.; Chiavaroli, L.; Wang, D.D.; Cozma, A.I.; Mirrahimi, A.; Yu, M.E.; Carleton, A.J.; Dibuono, M.; et al. Effect of fructose on blood pressure: A systematic review and meta-analysis of controlled feeding trials. Hypertens. Dallas Tex 1979 2012, 59, 787–795. [Google Scholar] [CrossRef] [PubMed]
  77. Hay, M. Sex, the brain and hypertension: Brain oestrogen receptors and high blood pressure risk factors. Clin. Sci. Lond. Engl. 1979 2016, 130, 9–18. [Google Scholar] [CrossRef] [PubMed]
  78. He, J.; Whelton, P.K. Effect of dietary fiber and protein intake on blood pressure: A review of epidemiologic evidence. Clin. Exp. Hypertens. 1999, 21, 785–796. [Google Scholar] [CrossRef] [PubMed]
  79. Hertz-Picciotto, I.; Croft, J. Review of the Relation between Blood Lead and Blood Pressure. Epidemiol. Rev. 1993, 15, 352–373. [Google Scholar] [CrossRef] [PubMed]
  80. Jee, S.H.; He, J.; Whelton, P.K.; Suh, I.; Klag, M.J. The effect of chronic coffee drinking on blood pressure: A meta-analysis of controlled clinical trials. Hypertens. Dallas Tex 1979 1999, 33, 647–652. [Google Scholar] [CrossRef]
  81. Jee, S.H.; Miller, E.R.; Guallar, E.; Singh, V.K.; Appel, L.J.; Klag, M.J. The effect of magnesium supplementation on blood pressure: A meta-analysis of randomized clinical trials. Am. J. Hypertens. 2002, 15, 691–696. [Google Scholar] [CrossRef]
  82. Johansen, H.L. Hypertension in Canada: Risk factor review and recommendations for further work. Can. J. Public Health Rev. Can. Sante Publique 1983, 74, 123–128. [Google Scholar]
  83. Kaufman, J.; Barkey, N. Hypertension in Africa: An overview of prevalence rates and causal risk factors. Ethn. Dis. 1993, 3 Suppl, S83–S101. [Google Scholar]
  84. Li, G.; Zhang, Y.; Thabane, L.; Mbuagbaw, L.; Liu, A.; Levine, M.A.H.; Holbrook, A. Effect of green tea supplementation on blood pressure among overweight and obese adults: A systematic review and meta-analysis. J. Hypertens. 2015, 33, 243–254. [Google Scholar] [CrossRef] [PubMed]
  85. Luft, F.C.; Miller, J.Z.; Lyle, R.M.; Melby, C.L.; Fineberg, N.S.; McCarron, D.A.; Weinberger, M.H.; Morris, C.D. The effect of dietary interventions to reduce blood pressure in normal humans. J. Am. Coll. Nutr. 1989, 8, 495–503. [Google Scholar] [CrossRef] [PubMed]
  86. Mesas, A.E.; Leon-Muñoz, L.M.; Rodriguez-Artalejo, F.; Lopez-Garcia, E. The effect of coffee on blood pressure and cardiovascular disease in hypertensive individuals: A systematic review and meta-analysis. Am. J. Clin. Nutr. 2011, 94, 1113–1126. [Google Scholar] [CrossRef] [PubMed]
  87. Millard, P.S.; Triplekorn, C.M. Effect of reduced dietary sodium on blood pressure. J. Fam. Pract. 1996, 43, 123–124. [Google Scholar] [PubMed]
  88. Murphy, M.H.; Nevill, A.M.; Murtagh, E.M.; Holder, R.L. The effect of walking on fitness, fatness and resting blood pressure: A meta-analysis of randomised, controlled trials. Prev. Med. 2007, 44, 377–385. [Google Scholar] [CrossRef] [PubMed]
  89. Naing, C.; Aung, K. Prevalence and risk factors of hypertension in myanmar: A systematic review and meta-analysis. Medicine (Baltimore) 2014, 93, e100. [Google Scholar] [CrossRef] [PubMed]
  90. Nissensohn, M.; Román-Viñas, B.; Sánchez-Villegas, A.; Piscopo, S.; Serra-Majem, L. The Effect of the Mediterranean Diet on Hypertension: A Systematic Review and Meta-Analysis. J. Nutr. Educ. Behav. 2016, 48, 42–53. [Google Scholar] [CrossRef] [PubMed]
  91. Ried, K.; Sullivan, T.R.; Fakler, P.; Frank, O.R.; Stocks, N.P. Effect of cocoa on blood pressure. Cochrane Database Syst. Rev. 2012, CD008893. [Google Scholar] [CrossRef]
  92. Roberts, C.; Banning, M. Risk factors for hypertension and cardiovascular disease. Nurs. Stand. 1998, 39–42. [Google Scholar]
  93. Steffen, P.R.; Smith, T.B.; Larson, M.; Butler, L. Acculturation to Western society as a risk factor for high blood pressure: A meta-analytic review. Psychosom. Med. 2006, 68, 386–397. [Google Scholar] [CrossRef] [PubMed]
  94. Steffen, M.; Kuhle, C.; Hensrud, D.; Erwin, P.J.; Murad, M.H. The effect of coffee consumption on blood pressure and the development of hypertension: A systematic review and meta-analysis. J. Hypertens. 2012, 30, 2245–2254. [Google Scholar] [CrossRef] [PubMed]
  95. Wang, M.; Moran, A.E.; Liu, J.; Qi, Y.; Xie, W.; Tzong, K.; Zhao, D. A Meta-Analysis of Effect of Dietary Salt Restriction on Blood Pressure in Chinese Adults. Glob. Heart 2015, 10, 291–299. [Google Scholar] [CrossRef] [PubMed]
  96. Whelton, S.P.; Chin, A.; Xin, X.; He, J. Effect of aerobic exercise on blood pressure: A meta-analysis of randomized, controlled trials. Ann. Intern. Med. 2002, 136, 493–503. [Google Scholar] [CrossRef] [PubMed]
  97. Xu, J.-Y.; Qin, L.-Q.; Wang, P.-Y.; Li, W.; Chang, C. Effect of milk tripeptides on blood pressure: A meta-analysis of randomized controlled trials. Nutr. Burbank Los Angel. Cty. Calif 2008, 24, 933–940. [Google Scholar] [CrossRef] [PubMed]
  98. Residential Proximity to Major Highways—United States, 2010. Available online: https://www.cdc.gov/mmwr/preview/mmwrhtml/su6203a8.htm (accessed on 20 July 2018).
  99. US EPA National Center for Environmental Assessment, W.D.; Moya, J. Exposure Factors Handbook 2011 Edition (Final Report). Available online: https://cfpub.epa.gov/ncea/risk/recordisplay.cfm?deid=236252 (accessed on 20 July 2018).
  100. Buonanno, G.; Stabile, L.; Morawska, L. Personal exposure to ultrafine particles: The influence of time-activity patterns. Sci. Total Environ. 2014, 468–469, 903–907. [Google Scholar] [CrossRef] [PubMed]
  101. Buonanno, G.; Marini, S.; Morawska, L.; Fuoco, F.C. Individual dose and exposure of Italian children to ultrafine particles. Sci. Total Environ. 2012, 438, 271–277. [Google Scholar] [CrossRef] [PubMed]
  102. Morawska, L.; Afshari, A.; Bae, G.N.; Buonanno, G.; Chao, C.Y.H.; Hänninen, O.; Hofmann, W.; Isaxon, C.; Jayaratne, E.R.; Pasanen, P.; et al. Indoor aerosols: From personal exposure to risk assessment. Indoor Air 2013, 23, 462–487. [Google Scholar] [CrossRef] [PubMed]
  103. Wallace, L.; Ott, W. Personal exposure to ultrafine particles. J. Expo. Sci. Environ. Epidemiol. 2011, 21, 20–30. [Google Scholar] [CrossRef] [PubMed]
  104. Int Panis, L.; de Geus, B.; Vandenbulcke, G.; Willems, H.; Degraeuwe, B.; Bleux, N.; Mishra, V.; Thomas, I.; Meeusen, R. Exposure to particulate matter in traffic: A comparison of cyclists and car passengers. Atmos. Environ. 2010, 44, 2263–2270. [Google Scholar] [CrossRef]
  105. Kelly, F. Oxidative stress: Its role in air pollution and adverse health effects. Occup. Environ. Med. 2003, 60, 612–616. [Google Scholar] [CrossRef] [PubMed]
  106. Braniš, M.; Šafránek, J.; Hytychová, A. Exposure of children to airborne particulate matter of different size fractions during indoor physical education at school. Build. Environ. 2009, 44, 1246–1252. [Google Scholar] [CrossRef]
  107. Dennekamp, M.; Howarth, S.; Dick, C. a. J.; Cherrie, J.W.; Donaldson, K.; Seaton, A. Ultrafine particles and nitrogen oxides generated by gas and electric cooking. Occup. Environ. Med. 2001, 58, 511–516. [Google Scholar] [CrossRef] [PubMed]
  108. Salant, T.; Lauderdale, D.S. Measuring culture: A critical review of acculturation and health in Asian immigrant populations. Soc. Sci. Med. 2003, 57, 71–90. [Google Scholar] [CrossRef]
  109. Oza-Frank, R.; Cunningham, S.A. The weight of US residence among immigrants: A systematic review. Obes. Rev. 2010, 11, 271–280. [Google Scholar] [CrossRef] [PubMed]
  110. Ayala; Guadalupe, X.; Baquero, B.; Klinger, S. A Systematic Review of the Relationship between Acculturation and Diet among Latinos in the United States: Implications for Future Research. J. Am. Diet. Assoc. 2008, 108, 1330–1344. [Google Scholar] [CrossRef] [PubMed]
  111. Choi, S.; Rankin, S.; Stewart, A.; Oka, R. Effects of acculturation on smoking behavior in Asian Americans: A meta-analysis. J. Cardiovasc. Nurs. 2008, 23, 67–73. [Google Scholar] [CrossRef] [PubMed]
  112. Serafica, R.C. Dietary acculturation in Asian Americans. J. Cult. Divers. 2014, 21, 145–151. [Google Scholar] [PubMed]
  113. Afable-Munsuz, A.; Ponce, N.A.; Rodriguez, M.; Perez-Stable, E.J. Immigrant generation and physical activity among Mexican, Chinese & Filipino adults in the U.S. Soc. Sci. Med. 2010, 70, 1997–2005. [Google Scholar] [CrossRef] [PubMed]
  114. Gerber, M.; Barker, D.; Pühse, U. Acculturation and physical activity among immigrants: A systematic review. J. Public Health 2012, 20, 313–341. [Google Scholar] [CrossRef]
  115. Gorman, B.K.; Lariscy, J.T.; Kaushik, C. Gender, acculturation, and smoking behavior among U.S. Asian and Latino immigrants. Soc. Sci. Med. 2014, 106, 110–118. [Google Scholar] [CrossRef] [PubMed]
  116. Jamal, A. Current Cigarette Smoking Among Adults—United States, 2005–2015. MMWR Morb. Mortal. Wkly. Rep. 2016, 65. [Google Scholar] [CrossRef] [PubMed]
  117. Sanchez-Vaznaugh, E.V.; Kawachi, I.; Subramanian, S.V.; Sánchez, B.N.; Acevedo-Garcia, D. Differential effect of birthplace and length of residence on body mass index (BMI) by education, gender and race/ethnicity. Soc. Sci. Med. 2008, 67, 1300–1310. [Google Scholar] [CrossRef] [PubMed]
  118. Flegal, K.M.; Carroll, M.D.; Kit, B.K.; Ogden, C.L. Prevalence of Obesity and Trends in the Distribution of Body Mass Index Among US Adults, 1999-2010. JAMA 2012, 307, 491–497. [Google Scholar] [CrossRef] [PubMed]
  119. Chastin, S.F.M.; Buck, C.; Freiberger, E.; Murphy, M.; Brug, J.; Cardon, G.; O’Donoghue, G.; Pigeot, I.; Oppert, J.-M. Systematic literature review of determinants of sedentary behaviour in older adults: A DEDIPAC study. Int. J. Behav. Nutr. Phys. Act. 2015, 12, 127. [Google Scholar] [CrossRef] [PubMed]
  120. Trost, S.G.; Owen, N.; Bauman, A.E.; Sallis, J.F.; Brown, W. Correlates of adults’ participation in physical activity: Review and update. Med. Sci. Sports Exerc. 2002, 34, 1996–2001. [Google Scholar] [CrossRef] [PubMed]
  121. Wareham, N.J.; van Sluijs, E.M.F.; Ekelund, U. Physical activity and obesity prevention: A review of the current evidence. Proc. Nutr. Soc. 2005, 64, 229–247. [Google Scholar] [CrossRef] [PubMed]
  122. Warburton, D.E.R.; Nicol, C.W.; Bredin, S.S.D. Health benefits of physical activity: The evidence. CMAJ 2006, 174, 801–809. [Google Scholar] [CrossRef] [PubMed]
  123. Freisling, H.; Knaze, V.; Slimani, N. A Systematic Review of Peer-Reviewed Studies on Diet Quality Indexes Applied to Old Age: A Multitude of Predictors of Diet Quality. In Diet Quality; Nutrition and Health; Humana Press: New York, NY, USA, 2013; pp. 365–381. ISBN 978-1-4614-7314-5. [Google Scholar]
  124. Satia-Abouta, J.; Patterson, R.E.; Kristal, A.R.; Teh, C.; Tu, S.-P. Psychosocial Predictors of Diet and Acculturation in Chinese American and Chinese Canadian Women. Ethn. Health 2002, 7, 21–39. [Google Scholar] [CrossRef] [PubMed]
  125. Paeratakul, S.; Popkin, B.M.; Keyou, G.; Adair, L.S.; Stevens, J. Changes in diet and physical activity affect the body mass index of Chinese adults. Int. J. Obes. 1998, 22, 424–431. [Google Scholar] [CrossRef]
  126. Weiss, D.R.; O’Loughlin, J.L.; Platt, R.W.; Paradis, G. Five-year predictors of physical activity decline among adults in low-income communities: A prospective study. Int. J. Behav. Nutr. Phys. Act. 2007, 4, 2. [Google Scholar] [CrossRef] [PubMed]
  127. Serra-Majem, L.; Roman, B.; Estruch, R. Scientific Evidence of Interventions Using the Mediterranean Diet: A Systematic Review. Nutr. Rev. 2006, 64, S27–S47. [Google Scholar] [CrossRef] [PubMed]
  128. Villareal, D.T.; Apovian, C.M.; Kushner, R.F.; Klein, S. Obesity in Older Adults: Technical Review and Position Statement of the American Society for Nutrition and NAASO, The Obesity Society. Obes. Res. 13, 1849–1863. [CrossRef] [PubMed]
  129. Morenga, L.T.; Mallard, S.; Mann, J. Dietary sugars and body weight: Systematic review and meta-analyses of randomised controlled trials and cohort studies. BMJ 2013, 346, e7492. [Google Scholar] [CrossRef] [PubMed]
  130. Soltani, S.; Shirani, F.; Chitsazi, M.J.; Salehi-Abargouei, A. The effect of dietary approaches to stop hypertension (DASH) diet on weight and body composition in adults: A systematic review and meta-analysis of randomized controlled clinical trials. Obes. Rev. 17, 442–454. [CrossRef] [PubMed]
  131. Sodowsky, G.R.; Lai, E.W.M.; Plake, B.S. Moderating Effects of Sociocultural Variables on Acculturation Attitudes of Hispanics and Asian Americans. J. Couns. Dev. 70, 194–204. [CrossRef]
  132. Textor, J.; Hardt, J.; Knüpppel, S. DAGitty: A Graphical Tool for Analyzing Causal Diagrams. Epidemiology 2011, 5, 745. [Google Scholar] [CrossRef] [PubMed]
Figure 1. Directed acyclic graph representing the relationships among the exposure (UFP; represented by the green oval with the triangle), outcome (blood pressure; represented by the blue oval with the line), and related factors. Variables represented as pink ovals are ancestors of exposure and outcome while variables represented as green ovals (convenience sample, proximity, inhalation rate, and cooking) are ancestors only of the exposure. Pink lines are biasing paths and the green line between the exposure and outcome is the causal path of interest.
Figure 1. Directed acyclic graph representing the relationships among the exposure (UFP; represented by the green oval with the triangle), outcome (blood pressure; represented by the blue oval with the line), and related factors. Variables represented as pink ovals are ancestors of exposure and outcome while variables represented as green ovals (convenience sample, proximity, inhalation rate, and cooking) are ancestors only of the exposure. Pink lines are biasing paths and the green line between the exposure and outcome is the causal path of interest.
Ijerph 15 02036 g001
Figure 2. Effect modification of the association of ln(TAA-PNC) with blood pressure by race/ethnicity. Open markers represent unadjusted associations (Model A) while solid markers represent adjusted associations (Model B; adjusted for BMI, sex, smoking, physical activity, and diet as fried food consumption).
Figure 2. Effect modification of the association of ln(TAA-PNC) with blood pressure by race/ethnicity. Open markers represent unadjusted associations (Model A) while solid markers represent adjusted associations (Model B; adjusted for BMI, sex, smoking, physical activity, and diet as fried food consumption).
Ijerph 15 02036 g002
Figure 3. Effect modification of the association of ln(TAA-PNC) with blood pressure by statin medication use, diabetes status, and hypertension status. Open markers represent unadjusted associations (Model A) while solid markers represent adjusted associations (Model B; adjusted for BMI, sex, smoking, physical activity, and diet as fried food consumption).
Figure 3. Effect modification of the association of ln(TAA-PNC) with blood pressure by statin medication use, diabetes status, and hypertension status. Open markers represent unadjusted associations (Model A) while solid markers represent adjusted associations (Model B; adjusted for BMI, sex, smoking, physical activity, and diet as fried food consumption).
Ijerph 15 02036 g003
Figure 4. Effect estimates for the association of ln(TAA-PNC) with blood pressure measures for participants in Somerville and Dorchester/South Boston who attended two clinic visits (n = 205). Open markers represent unadjusted associations (Model A) while solid markers represent adjusted associations (Model B; adjusted for BMI, sex, smoking, physical activity, and diet as fried food consumption).
Figure 4. Effect estimates for the association of ln(TAA-PNC) with blood pressure measures for participants in Somerville and Dorchester/South Boston who attended two clinic visits (n = 205). Open markers represent unadjusted associations (Model A) while solid markers represent adjusted associations (Model B; adjusted for BMI, sex, smoking, physical activity, and diet as fried food consumption).
Ijerph 15 02036 g004
Table 1. Study sample characteristics.
Table 1. Study sample characteristics.
TotalWhiteAsianOther
Characteristicn% (n)
or mean (sd)
n% (n)
or mean (sd)
N% (n)
or mean (sd)
N% (n)
or mean (sd)
TAA-PNC * (particles/cm3)40922,000 (6500)17820,000 (4900)14924,000 (7900)8221,000 (5000)
ln[(TAA-PNC) (particles/cm3)]4099.9 (0.35)1789.8 (0.28)14910.0 (0.43)829.9 (0.27)
SBP (mmHg)409137.5 (19.5)178133.9 (18.3)149141.2 (20.6)82138.9 (18.8)
DBP (mmHg)40977.7 (10.3)17876.2 (10.7)14977.3 (9.3)8281.9 (10.4)
PP (mmHg)40959.8 (16.5)17857.6 (14.9)14963.9 (18.4)8257.1 (15.1)
Hypertension40963.8 (261)17855.6 (99)14969.8 (104)8270.7 (58)
Age (years)40961.5 (12.8)17859.8 (11.3)14966.6 (13.4)8256.0 (11.3)
BMI (kg/m2)39327.7 (6.8)16829.5 (6.9)14924.1 (4.1)7630.6 (7.7)
ln[light/moderate physical activity (min/week)]3744.3 (2.2)1643.8 (2.3)1475.1 (1.6)633.5 (2.3)
Female40959.2 (242)17859.6 (106)14956.4 (84)8263.4 (52)
Smoker status398 176 145 77
Current 21.1 (84) 22.2 (39) 14.5 (21) 31.2 (24)
Former 30.7 (122) 43.2 (76) 15.9 (23) 29.9 (23)
Never 48.2 (192) 34.7 (61) 69.7 (101) 39.0 (30)
Fruit and vegetable consumption ≥ 7x/week27538.2 (105)12548.0 (60)9730.9 (30)5328.3 (15)
Fried food consumption ≥ 1x/week40533.8 (137)17645.5 (80)14914.8 (22)8043.8 (35)
Educational Attainment409 178 149 82
< HS 34.2 (140) 11.2 (20) 61.7 (92) 34.2 (28)
HS 31.8 (130) 36.5 (65) 24.2 (36) 35.4 (29)
>HS 34.0 (139) 52.3 (93) 14.1 (21) 30.5 (25)
Born in the USA40445.5 (184)17486.8 (151)1490.0 (0)8140.7 (33)
Statin Medications40029.0 (116)17631.3 (55)14429.2 (42)8023.8 (19)
Hypertension medications40045.0 (180)17635.8 (63)14454.2 (78)8048.8 (39)
Diabetes39920.3 (81)17517.7 (31)14418.8 (27)8028.8 (23)
* Time-activity-adjusted particle number concentration. Italics indicate variable levels.
Table 2. Effect estimates for ln (TAA-PNC).
Table 2. Effect estimates for ln (TAA-PNC).
ModelSBP (mmHg)DBP (mmHg)PP (mmHg)Hypertension
β (95% CI)β (95% CI)β (95% CI)OR (95% CI)
Model A2.87 (−2.60, 8.33)−0.18 (−3.09, 2.72)3.05 (−1.58, 7.68)1.53 (0.86, 2.72)
Model B5.23 (0.68, 11.14)0.96 (2.08, 4.00)4.27 (0.79, 9.32)1.81 (0.94, 3.48)
Model C5.84 (−1.94, 13.61)1.79 (−1.95, 5.52)4.05 (−2.80, 10.90)1.53 (0.68, 3.43)
Model D2.41 (−3.51, 8.32)0.15 (−2.92, 3.21)2.26 (−2.83, 7.35)1.27 (0.64, 2.52)
Model E3.60 (−1.75, 8.95)1.10 (−1.95, 4.15)2.50 (−1.74, 6.74)1.72 (0.84, 3.55)
Model F1.67 (−3.87, 7.22)0.01 (−3.15, 3.18)1.66 (−2.78, 6.11)1.25 (0.57, 2.75)
Model G1.68 (−3.89, 7.24)−0.19 (−3.35, 2.97)1.86 (−2.58, 6.31)1.31 (0.59, 2.91)
Model H5.67 (−0.40, 11.75)1.35 (−1.76, 4.45)4.33 (−0.88, 9.53)1.86 (0.95, 3.66)
(A) Unadjusted (n = 409). (B) Main model adjusted for BMI, sex, smoking, physical activity, and diet (as fried food consumption; n = 347). (C) Model B covariates but using fruit and vegetable consumption for diet (n = 237). (D) Model B covariates but using race/ethnicity for diet (n = 350). (E) Model B covariates as well as age (n = 347). (F) Model B covariates as well as additional adjustment for proximity (as race/ethnicity, age, educational attainment; n = 347). (G) Model B covariates as well as additional adjustment for proximity (as race/ethnicity, age, educational attainment, random or convenience sample participant; n = 347). (H) Model B covariates as well as additional adjustment for proximity (as annoyance at traffic sound; n = 344). Model B is bolded since it is the main model.

© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
Int. J. Environ. Res. Public Health EISSN 1660-4601 Published by MDPI AG, Basel, Switzerland RSS E-Mail Table of Contents Alert
Back to Top