Skip to main content

Association between coarse particulate matter and inflammatory and hemostatic markers in a cohort of midlife women



Exposure to particulate matter air pollution has been associated with cardiovascular disease (CVD) morbidity and mortality; however, most studies have focused on fine particulate matter (PM2.5) exposure and CVD. Coarse particulate matter (PM10–2.5) exposure has not been extensively studied, particularly for long-term exposure, and the biological mechanisms remain uncertain.


We examined the association between ambient concentrations of PM10–2.5 and inflammatory and hemostatic makers that have been linked to CVD. Annual questionnaire and clinical data were obtained from 1694 women (≥ 55 years old in 1999) enrolled in the longitudinal Study of Women’s Health Across the Nation (SWAN) at six study sites from 1999 to 2004. Residential locations and the USEPA air monitoring network measurements were used to assign exposure to one-year PM10–2.5, as well as co-pollutants. Linear mixed-effects regression models were used to describe the association between PM10–2.5 exposure and markers, including demographic, health and other covariates.


Each interquartile (4 μg/m3) increase in one-year PM10–2.5 exposure was associated with a 5.5% (95% confidence interval [CI]: 1.8, 9.4%) increase in levels of plasminogen activator inhibitor-1 (PAI-1) and 4.1% (95% CI: − 0.1, 8.6%) increase in high-sensitivity C-creative Protein (hs-CRP). Stratified analyses suggested that the association with PAI-1 was particularly strong in some subgroups, including women who were peri-menopausal, were less educated, had a body mass index lower than 25, and reported low alcohol consumption. The association between PM10–2.5 and PAI-1 remained unchanged with adjustment for PM2.5, ozone, nitrogen dioxide, and carbon monoxide.


Long-term PM10–2.5 exposure may be associated with changes in coagulation independently from PM2.5, and thus, contribute to CVD risk in midlife women.

Peer Review reports


Cardiovascular disease (CVD) is the leading cause of death for men and women in the US [1]. Risk factors such as sex, age, increased blood pressure, high cholesterol, and smoking, only account for approximately 50% of cardiac events [2]. Evidence suggests that exposure to particulate matter air pollution is associated with cardiovascular morbidity and mortality, potentially through its ability to increase inflammation and coagulation [3].

Given the known adverse health effects, the United States Environmental Protection Agency (USEPA) regulates particulate matter by setting regulatory standards and enforcing mandatory monitoring for particulate matter with an aerodynamic diameter smaller than 2.5 and 10 μm (PM2.5 and PM10, respectively) [4]. Most available research has been focused on the health impacts of PM2.5 exposure; however, potential health effects from exposure to coarse particulate matter with an aerodynamic diameter between 2.5 and 10 μm (PM10–2.5 or coarse particles) might differ from exposure to PM2.5 due to the components, sources, and size. Coarse particles are primarily generated by mechanical grinding and resuspension of solid material and may have a high biological content (such as pollen, fungi, and endotoxins) or crustal matter (such as aluminum and silicon and heavy metals) [5]. As a result of size difference, PM2.5 deposits deeper in the alveolar region of the lungs, while PM10–2.5 mainly deposits higher in the airways [4]. Our understanding of the health impacts from PM10–2.5 exposure is limited, and thus, no specific regulatory standards exist for PM10–2.5 [4].

Evidence also suggests that PM10–2.5 is associated with respiratory and cardiovascular morbidity and mortality [6,7,8,9,10,11], but the majority of the previous studies have focused on short-term exposure. Powell et al. reported a significant positive association between hospitalizations for CVD and same-day PM10–2.5 levels in a Medicare population, aged ≥65 years, in the US [9]. In a systematic review and meta-analysis, respiratory mortality and hospitalizations were increased by 1.4% (95% confidence interval (CI): 0.5–2.4%) and 1.0% (95% CI: 0.1–1.8%), respectively, per 10 μg/m3 increase in short-term PM10–2.5 exposure [12]. However, the associations between long-term PM10–2.5 exposure and mortality/morbidity, which may be derived from different mechanisms from short-term exposure, have been unclear. Early studies reviewed by Brunekreef and Forsberg did not present associations between mortality and morbidity with long-term exposures to PM10–2.5 [13]. The meta-analysis study Adar et al. reported a summary estimate of a 2.1% (95% CI: − 1.6 to 5.8%) higher mortality rate per 10 μg/m3 increment in long-term PM10–2.5 concentration based on six cohort studies, but this association diminished after adjustment for PM2.5 [12]. Recent studies examining specific health endpoints have linked long-term exposure to coarse PM with right ventricular dysfunction and hypertension [14, 15].

Furthermore, few studies have assessed associations between PM10–2.5 and markers of inflammation and coagulation that are predictive of CVD. Investigating the association between particulate matter and CVD biomarkers could help postulate the underlying physiological mechanisms. Adar et al. found that the endotoxin component of the 5-year concentrations of PM10–2.5 was associated with increased inflammation scores, while the copper component of PM10–2.5 was associated with an elevated coagulation score [16].

The Study of Women’s Health Across the Nation (SWAN) cohort was designed to follow a multi-racial/ethnic cohort of midlife women through menopausal transition [17]. With longitudinal biomonitoring across years, this study offers a unique opportunity to study CVD risks associated with long-term PM10–2.5 exposure. With previous work having shown associations between CVD markers and exposures to ambient PM2.5 and gaseous pollutants [18, 19], further research with respect to PM10–2.5 exposure and CVD markers in the SWAN cohort would advance the understanding of whether PM10–2.5 are associated with inflammation and coagulation. Moreover, with the ability to assess other air pollutants, this study could further evaluate confounding by co-pollutants and improve the understanding of the potential independent impact of PM10–2.5 exposure on CVD.


Study population

The SWAN study design and recruitment has been previously described [17]. The present study includes data from six sites: Chicago, Illinois; Detroit, Michigan; Los Angeles, California; Newark, New Jersey; Oakland, California; and Pittsburgh, Pennsylvania. Recruitment included non-Hispanic White women at all sites as well as women from one other racial/ethnic group at each site: African American in Chicago, Pittsburgh, and Detroit; Asian in Oakland (Chinese) and Los Angeles (Japanese); and Hispanic (including women of Central American, Mexican, and Caribbean origin) in Newark. The inclusion criteria for the study at baseline were: being 42 to 52 years of age, having an intact uterus and at least one ovary, no current use of exogenous hormones, not being pregnant or lactating, and having had at least one menstrual period in the last 3 months. Approximately 450 eligible women at each study site have been followed up with annual clinical assessments and interviews.

The present study includes serum samples collected at clinical visits 3 (1999–2000) through visits 7 (2003–2004), since PM2.5, a key component for calculating PM10–2.5 concentration, was not routinely measured nationwide until 1998.

CVD marker measurement and analysis

Fasting blood was drawn at each annual SWAN clinical visit and assayed as described previously [18, 20]. The inflammatory and hemostatic markers assessed in this study included high sensitivity C-reactive protein (hs-CRP), fibrinogen, factor VII coagulant (factor VIIc), tissue-type plasminogen activator antigen (tPA-ag), and plasminogen activator inhibitor Type 1 (PAI-1). They serve as markers of the following processes, including systemic inflammation (hs-CRP), formation of blood clots when endothelial damage occurs (fibrinogen and factor VIIc), and fibrinolysis (tPA-ag and PAI-1) [21].

Hs-CRP was measured in serum by immunonephelometry using Behring reagents. Fibrinogen and Factor VII-C were measured in frozen citrated plasma by an assay in which clotting time is measured and compared to a control on a MLA 1400 C coagulation analyzer. PAI-1 and tPA-ag, both free form and tPA-PAI-1 complexes, were measured in plasma using enzyme-linked immunosorbent assay (ELISA) technique (American Diagnostica). A human single chain tPA-ag was used as a standard calibrated against an international standard (National Institute for Biological Standards and Control, Hertfordshire, United Kingdom). PAI-1 was detected using a solid phased monoclonal antibody and a second enzyme-labeled goat antiserum (American Diagnostica) [22]. PAI-1, tPA-ag and hs-CRP were measured at every visit, while fibrinogen and factor VIIc were measured only for visits 3, 5 and 7. All bioassay analyses were conducted in the Medical Research Laboratory, Lexington Kentucky.

Additionally, hs-CRP values > 10 mg/L were excluded from the analyses because they may be indicative of severe infection, major trauma, or chronic inflammatory disease (0.1% of all observations). For the other inflammatory and hemostatic markers, extreme values (outside the mean ± 3 standard deviations after log transformation) were excluded from study (< 3%) as they may indicate laboratory error.

Exposure assignment

A residential history was obtained for each SWAN participant. Each residence was geocoded, with its coordinate randomly moved up to 400 ft (approximately one block) away to maintain confidentiality. A 20 km circular buffer was created around each address using ArcGIS v10.0 (Environmental Systems Research Institute 2016). The exposure measured by the monitor located in the buffer will be assigned to the participant. For instances in which multiple monitors were located within the buffer, one was chosen based on 1) distance from the residence, and 2) number of visits with exposure data available, which could differ by monitor because of differences in operating dates. Priority was given to the closer monitor to reduce exposure misclassification, but the more distant monitor was chosen if the ratio of (available visits of the distant monitor / available visits of the closer monitor) was greater than the ratio of (the distance to the distant monitor / distance to the closer monitor). If a participant moved between visits, exposure was weighted based on time at each residence using the available move date. If no move date was available, exposure was estimated using the midpoint between the two visits.

Daily PM10–2.5 was calculated as the difference of daily PM2.5 and PM10 concentrations measured at the same monitoring site on the same day, whenever available. Ambient PM2.5 and PM10 concentrations were obtained from the USEPA air monitoring network, which were monitored every third or sixth day, or daily. The data for these measures were in a 24-h average concentration format. There were fewer monitors for PM10, and some monitors take measurements less frequently, which limited the number of measurements of PM10, and thus resulting in fewer calculated PM10–2.5 measurements. Furthermore, given the uncertainty of extreme values, the top and bottom 2.5% of PM10 and of PM2.5 data were trimmed.

To assess the potential confounding effects of ambient gases, ozone (O3), carbon monoxide (CO), nitrogen dioxide (NO2), and sulfur dioxide (SO2) concentrations were also obtained. The available data were daily 8-h maximum concentrations for ozone and CO, and one-hour maximum concentrations for NO2 and SO2. It should be noted that some monitors, specifically in Michigan, only monitored O3 during the summer time. All air pollutant data were downloaded from the USEPA Air Data website (, accessed September 2010).

In this study, we focused on the impact of long-term PM10–2.5 exposure, namely the average prior one-year exposure. The six-month exposure was also calculated and considered as intermediate-exposure for a sensitivity analysis. Daily readings were used to calculate average exposure levels for six-month and one-year prior to each blood draw. To simplify, months were considered to be 30-day increments. A minimum of 9 days was necessary for calculating a one-month average; at least 5 months was necessary for a six-month average; and at least 10 months for a one-year average.


The baseline questionnaire, completed when participants were recruited, collected non-time varying covariates on socioeconomic status, for example, residence address, date of birth (for calculating age), race/ethnicity (categorized into White, African American, Asian, or Hispanic), and education (high school or less, some college, or college graduate).

At annual clinic visits, SWAN participants completed questionnaires providing visit-specific information related to medical history, psychosocial environment, lifestyle behaviors, menstrual bleeding patterns, illness, and use of medications since their last visit. Menopausal status was determined based on self-reported bleeding patterns according to standard definitions, and categorized into pre-, early peri-, late peri-, post-menopause, and unknown [23]. Height and weight were measured to calculate body mass index (BMI). Alcohol consumption was divided into three categories developed by Laura L Schott (EDC Coordinating Center): low = none or < 1 serving/month, moderate = up to 1/week or 0.3/day, and high = 2+/week or > 0.3/day.

Statistical analysis

For each air pollutant and CVD marker, summary statistics were calculated. Biomarker levels were log-transformed to meet the normality assumption. Correlations of markers and air pollutants were calculated based on visit 3 data, as it had the largest sample size among the included visits and serum samples and serves as the baseline for our study.

To study the association between PM10–2.5 and CVD biomarkers, we used linear mixed-effects regression models with each biomarker as a continuous, dependent variable. The average prior one-year PM10–2.5 exposure was included in a single-pollutant model along with covariates. A random intercept was used to account for covariance of measurements, as multiple longitudinal measurements collected from the same woman are highly correlated. Site was included as a fixed effect because participants were nested within each site. First-order ante-dependence structure was specified for repeated measurements from each participant [19].

Potential covariates were evaluated based on statistical significance and Akaike Information Criterion (AIC) value to control for confounding and goodness of model fit, respectively, without adding over fitted covariates. With all potential covariates, a backward elimination of variables, one at a time, was performed. Variables that have been tested but not included in the final model were: season (cold/warm), physical activity score (continuous), depress symptom score (continuous), poverty score of the ZCTA the participant resided (continuous), antilipidemic medication use (yes/no), and depression medication (yes/no). These variables that were not statistically associated with outcome markers and whose elimination did not change AIC were excluded from the model. The final model included study site, race/ethnicity, education, and visit-specific variables including age (continuous), BMI (continuous), menopausal status, active smoking (yes/no), and alcohol consumption.

Visits after major CVD events, including myocardial infarction, coronary heart failure, stroke, percutaneous coronary intervention, and coronary artery bypass graft, were censored from analyses. We also censored visits for which women did not fast 12 h before blood draw. New Jersey data from visits six and seven were censored because only a few participants had serum sample data for those visits.

Potential effect modification was considered by targeting subgroups for stratification. These groups included BMI (below vs. equal to or above 25 kg/m2), current menopause transition stage (early or late peri-menopausal vs. post-menopausal; pre-menopausal was not considered due to small sample size), alcohol consumption category (low vs. medium or high), and education (high school or less vs. some college or more). Each stratified variable was omitted from the base model.

To evaluate the confounding by co-pollutants, we used two-pollutant models, incorporating each of PM2.5, ozone, CO, NO2, and SO2, respectively, along with one-year PM10–2.5 in the final models mentioned above, to evaluate the potential confounding effects of each co-pollutant.

Several sensitivity analyses were run excluding women who: had diabetes, were currently smoking, only had one or two visits, with pre-existing medical conditions (including major CVD events defined above, angina, hypertension, or diabetes) and/or medication use, and had unknown menopausal status and/or hormone use.

Analyses were performed in SAS 9.4 (SAS Institute, Cary, NC). All tests were two-sided and p-values < 0.05 were considered statistically significant. All results were expressed as the percent change in markers per respective interquartile range increase in air pollutant based on averaging times using the formula [100 × (exp(β per unit pollutant  IQR) − 1)].


After applying the exclusions described above, 1694 women with 5982 observations during the study period were available for analyses. Approximately 68% of the women had three or more clinic visits between 1999 and 2004. The race/ethnicity distribution by study site for this population reflected the SWAN sampling strategy discussed previously. Over 50% of women in Chicago, Los Angeles, and Oakland had completed college, while in Detroit, Newark, and Pittsburgh, approximately 27% of the women had completed college. Most women in the study were overweight (25 < BMI ≤ 30) or obese (BMI ≥ 30), with the exception of women from the Oakland and Los Angeles sites. Additionally, most women in this study were neither current smokers nor high alcohol consumers (2+ servings/week or > 0.3 servings/day) (Table 1).

Table 1 Characteristics of the Study Population, SWAN Cohort, 1999–2004

The distributions of CVD biomarkers by risk factor status reflect the potential risk factors associated with inflammation and coagulation (Table 2). Women who were current smokers, had diagnosed diabetes, or were obese had higher levels of the inflammatory and hemostatic markers. African American and Hispanic women tended to have higher levels of the inflammatory/ hemostatic markers while Asian women had lower. The two inflammatory markers, hs-CRP and fibrinogen, were moderately correlated (Pearson correlation coefficient (r) = 0.39), as were the two hemostatic markers, PAI-1 and tPA-ag (r = 0.50).

Table 2 Distribution of inflammatory and hemostatic biomarkers by demographic factors for SWAN cohort, 1999–2004a

The average prior one-year PM10–2.5 exposures for all the SWAN sites were 10.9 ± 3.6 μg/m3 (Table 3). The average prior six-month exposure was similar (10.8 ± 4.0 μg/m3), and had a slightly larger sample size (N = 5466) than the one-year exposure (N = 5175), because PM2.5 measurements started from 1998 which limited the ability to calculate prior one-year exposure for some early visits. Los Angeles site had the highest levels of one-year PM10–2.5 exposure at 17.5 ± 2.3 μg/m3, while Pittsburgh had the lowest levels at 8.5 ± 3.8 μg/m3. Due to the limited availability of PM10 monitors, however, Los Angeles only had a small number of visits with co-located PM10 and PM2.5 monitors, and thus matched PM10–2.5 data. Distributions of one-year average exposure of co-pollutants, including PM2.5, ozone, CO, NO2 and SO2 exposure, can be found in Table S1.

Table 3 Descriptive statistics for prior one-year exposure to coarse particulate matter (PM10–2.5) concentrations (μg/m3) for SWAN sites, 1999–2004

We observed that an interquartile (4 μg/m3) increase in the prior one-year exposure of PM10–2.5 was associated with a 5.5% (1.8, 9.4%) increase in PAI-1 level. This association was consistent for the six-month exposure window, with a 4.0% (0.3, 7.8%) increase in PAI-1 per 4 μg/m3 increase in PM10–2.5. The inflammation marker, hs-CRP, was also associated with PM10–2.5, with a 4.1% (− 0.1, 8.6%) and a 6.0% (1.7, 10.5%) increase in hs-CRP per 4 μg/m3 increase in prior one-year and six-month PM10–2.5 exposure, respectively. Associations with other markers were not observed.

Sensitivity analyses were conducted with some subgroups removed. Positive associations between PM10–2.5 and PAI-1 levels remained statistically significant in models that excluded women with diagnosed diabetes, those who reported current smoking, who did not complete at least three clinic visits, who reported medical preconditions, who reported medication use, and who had unknown menopausal status and hormone use, respectively (Table S2). The marginal association between PM10–2.5 and hs-CRP became statistically significant in the model restricting analysis to women who completed three or more visits and in the model excluding smokers; however, the association disappeared in other sensitivity analyses.

Potential effect modifiers were evaluated for the one-year PM10–2.5 exposure, and some subgroups appeared to have experienced elevated risks (Table 4). For a 4 μg/m3 increase in one-year PM10–2.5, PAI-1 increased by 9.2% (1.4, 17.7%) in women with a BMI <  25, 7.5% (2.2, 13.2%) in peri-menopausal women, 7.6% (0.6, 15.1%) in women with equivalent or less than a high school education, and 7.5% (2.3, 13.0%) in women who reported low alcohol consumption. Additionally, women who reported low alcohol consumption had an 8.8% (2.7, 15.3%) increase in hs-CRP, and women with equivalent or less than a high school education had 2.4% (0.1, 4.7%) decrease in factor VIIc, per 4 μg/m3 increase in PM10–2.5. The stratification results for the other inflammatory and hemostatic markers were not statistically significant.

Table 4 Adjusted associations between PM10–2.5 and cardiovascular disease markers for SWAN cohort, 1999–2004a

The co-pollutant models were conducted in subsets of observations compared with the single pollutant models of PM10–2.5, given the availability of co-pollutant data. For most sites, 70–98% of serum samples had matched co-pollutant exposure, except for ozone in Detroit (0.5%) (not included in the analysis) and Newark (48%), NO2 in Detroit (57%), Chicago (65%), and Newark (65%), and SO2 in Oakland (53%) (Table S1). One-year PM10–2.5 exposure appeared to be moderately correlated with one-year PM2.5 (r = 0.29) and NO2 (r = 0.30). The interquartile ranges for the prior one-year average exposures of co-pollutants was 3 μg/m3 for PM2.5, 0.007 ppm for ozone, 0.3 ppm for CO, 5 ppb for NO2, and 3.0 ppb for SO2.

Our previous studies have found that one-year exposures to PM2.5, NO2, and CO were positively associated with PAI-1 [18, 19]. In two-pollutant models along with PM10–2.5, PM2.5, CO, and NO2 were each still positively associated with PAI-1 levels, and the associations were stronger for PM2.5, CO, and NO2 than PM10–2.5. In contrast, O3 had a negative association with PAI-1. Nevertheless, these co-pollutants did not confound the association between PM10–2.5 and PAI-1 (Fig. 1 and Table S3). Specifically, the positive association remained for PAI-1 and PM10–2.5 from the single pollutant model when considering co-exposure to PM2.5, with a 4.8% (1.0, 8.7%) increase per 4 μg/m3 increase in PM10–2.5. Similarly, PAI-1 increased by 4.8% (0.5, 9.4%) per 4 μg/m3 increase in PM10–2.5 after adjustment for co-exposure to CO, and by 4.3% (0, 8.7%) per 4 μg/m3 increase in PM10–2.5 independent of exposure to NO2. Confounding by SO2 cannot be evaluated, because in the subset of data with matched SO2 measurements, PM10–2.5 showed no association with PAI-1 regardless of including SO2 or not. The availability of SO2 measurements may have introduced bias for testing the association between PM10–2.5 and PAI-1.

Fig. 1
figure 1

Associations between PM10–2.5 and PAI-1, without (dot symbol) and with (triangle symbol) PM2.5, O3, CO, NO2, and SO2 adjusted in the SWAN cohort, 1999–2004. Note: Results shown are percent of change in biomarker level per an interquartile increase of PM10–2.5 exposure, which is 4 μg/m3. Analyses were based on log-transformed biomarker levels, adjusted for study site, age (continuous), race/ethnicity, education, menopausal status, BMI, active smoking status, alcohol consumption category. For each co-pollutant, analyses were run in the subset with non-missing values for both PM10-2.5 and the co-pollutant


This study provided new evidence for associations between long-term exposure to coarse particles and CVD markers. Specifically, we found that prior one-year PM10–2.5 exposure levels were positively associated with the levels of PAI-1, and marginally positively associated with hs-CRP. Elevated PAI-1 increases the risk of thrombosis by inhibiting tPA activity, which initiates the fibrinolysis processes that break down excess blood clotting. Therefore, the associations we found suggested that elevated long-term exposure to coarse particles may increase the risk of thrombosis.

As mentioned earlier, very few studies were conducted on the associations of long-term exposure to PM10–2.5 with CVD markers. Chen and Schwarz observed increased white blood cell count, a potential inflammatory marker, was associated with one-year local PM10 levels (ranging 14.6–78.5 μg/m3) in the US [24]. In contrast, in England, Forbes et al. reported no associations between inflammatory markers, fibrinogen or hs-CRP, and chronic exposures to several outdoor pollutants, including PM10 (ranging 11.0–36.1 μg/m3) [25]. Inconsistencies in study findings could have resulted from differences in air pollutant ranges, geographic and temporal differences in the composition of particles, and/or population differences. The Multi-Ethnic Study of Atherosclerosis estimated 5-year average PM10–2.5 exposures between 2000 and 2006 varied by individuals, with a mean of 5.0 ± 1.7 μg/m3, and found positive, though not statistically significant, associations between long-term PM10–2.5 concentrations and inflammation and coagulation markers [16]. The authors further identified that endotoxin and copper in PM10–2.5 contributed most to inflammation and coagulation effects, respectively, which were robust to the adjustment for PM2.5. As an indicator of abrasive brake wear, copper is expected in substantial amount in PM10–2.5 at most of the SWAN sites, which are major metropolitan areas with intense traffic flow, particularly in Los Angeles [26]. Our study echoes Adar et al’s findings, providing additional evidence on the impact of PM10–2.5 coagulation among different locations and populations.

Our previous studies have identified associations between exposures to PM2.5 and ambient gases with inflammatory and hemostatic markers in the SWAN cohort. One-year exposure to PM2.5 was associated with increases in hs-CRP (% of change [95th% confidence interval]: 21% [6.6, 37%]) and PAI-1 (35% [19, 53%]) per 10 μg/m3 increase in PM2.5 concentrations [18]. One-year exposure to CO and NO2 were also positively associated with PAI-1, though not statistically significant [19]. We note that, while the increased inflammation and coagulation tendency associated with PM10–2.5 was similar to that from PM2.5, PM10–2.5 exposure was weakly correlated with PM2.5, regardless of whether long-term (Spearman r = 0.30, p < 0.01) or short-term (r = 0.15, p < 0.01) exposure. The two-pollutant analyses confirmed the associations with PM2.5 and ambient gases, and further demonstrated that PM10–2.5 created additional burden on inflammation and coagulation. The previous studies did not control for PM10–2.5, so some of the associations observed may have been due to confounding by PM10–2.5.

Considering the variations in chemical composition and the deposition locations within the lungs associated with their sizes, we expect that PM10–2.5 cause health effects through mechanisms different from those of PM2.5. However, it has been observed that coarse PM of low density, such as soil particles, have deposited deeper in the lung as well [27], suggesting that they could cause potential health impacts beyond size alone. A study using in vitro models found out that PM2.5 and PM10–2.5 affected nasal and bronchial epithelial cells and immune response differently, but both increased release of IL-6 in bronchial epithelial cells [28]. Ljubimova et al. observed up regulation of genes encoding inflammatory cytokine pathways (IL13-Rα1 and IL-16) in the brains of rats exposed to PM10–2.5 sourced from the Los Angeles Basin for 1 month, and found that PM10–2.5 was the only particles contributing to this process [29]. A few toxicological studies has indicated that PM10–2.5 may have stronger associations with inflammation and coagulation than PM2.5 [30,31,32,33,34].

Furthermore, risks derived from short-term and long-term exposure may be generated through different mechanisms. Coagulation activation, indicated by the increase of PAI-1 level, is a typical risk that has been associated with long-term exposure [35]. Two potential mechanisms could explain the increased PAI-1. One possibility is that, some PM10–2.5 particles travel deeper in the lung and cause systemic inflammation and oxidative stress similar to that from PM2.5. Generated by the endothelium, PAI-1 is activated by chronic inflammatory conditions and endothelium injury to inhibit fibrinolysis, the process that degrades blood clots; in other words, PAI-1 allows blood clot formation and shortens bleeding time. Another possible pathway is that PM10–2.5 induces respiratory tract inflammation. Inflammatory cytokines produced in the respiratory tract can potentially enter the circulatory system where they can stimulate the liver to release coagulation factors that can alter hemostasis [4]. This could be a “false” coagulation signal that is not accompanied by systemic inflammation or blood vessel damage that initiates fibrinolysis, eventually increasing the potential for thrombosis. Under such circumstances, there may not be clear signs of systemic inflammation and/or endothelium injury. Overall, evidence from epidemiologic or toxicological studies is limited for systemic inflammation and altered hemostasis associated with the long-term exposure to PM10–2.5. Given the marginal/unstable association with hs-CRP and absent relationship with fibrinogen and factor VIIc, we cannot conclude whether inflammation and thrombosis were from the same biological process, or differentiate whether such coagulation reflected vascular inflammation or deep vein thrombosis. Future studies are warranted to explore the physiological consequences and mechanisms involved with chronic exposure to PM10–2.5, especially to the specific chemical species of PM10–2.5.

Our findings also suggest that the associations between PAI-1 and one-year PM10–2.5 exposure appeared particularly strong in some subgroups, including women with a BMI <  25, women who were peri-menopausal, women with equivalent or less than a high school education, and women who reported low alcohol consumption. That is to say, apart from toxicity of chemicals, physiological condition, i.e., BMI and menopausal status, and socioeconomic background, i.e., education, could influence vulnerability to external exposures. For women, the menopausal transition is a complex and multifaceted process that involves multiple organ systems and genetic variability where physiological impacts can manifest differently among individual women [17]. Being in midlife and experiencing a transition through menopause potentially increases women’s vulnerability to environmental exposures, such as air pollution.

This study had several strengths. First, SWAN is a large, multi-site, longitudinal study, which provided coverage of a wide range of ambient exposure levels and particle compositions associated with local sources. Also, with the same participants followed longitudinally and the same sampling/analytical approaches for measurement of biomarkers consistently used, our data provided a unique opportunity to examine health effects associated with long-term PM10–2.5 exposure. Second, because residential history has been maintained for each SWAN participant, we were able to use ambient monitoring data, along with residential history, to assign exposure levels to participants. Finally, substantial demographic and health information collected in SWAN allowed us to examine the impact of effect modifiers, including time-varying menopausal status for this midlife population.

Meanwhile, the limitations of estimating PM10–2.5 exposure need to be noted when interpreting our results. PM10–2.5 was not routinely measured directly, its level was either calculated through the ambient monitor concentrations of PM10 and PM2.5 as we did, or obtained by modeling. Some studies used estimations from land use regression models, which were based on limited cross-sectional measurements to simulate long-term spatial variation [14, 16]. The 5-year average PM10–2.5 exposures of the Multi-Ethnic Study of Atherosclerosis cohort were estimated using land use regression model, ranging from 3.8 ± 1.3 μg/m3 in Winston-Salem, North Carolina to 5.6 ± 1.2 μg/m3 in Chicago, Illinois [16]. Models based on satellite data are not applicable to this study, as most satellite data are only available after early 2000, and therefore, would not cover a significant amount of SWAN visits around 2000. Furthermore, PM10–2.5 concentrations were not commonly estimated through satellite data in the USA, as PM2.5 has been the primary focus. We acknowledge that PM10–2.5 concentration tends to be more heterogeneous across space than PM2.5, resulting in great uncertainty for PM10–2.5 estimations through nearby stationary monitors within a 20 km buffer. The assigned exposure may be overestimated or underestimated depending on whether there are sources nearby. We assume that the exposure assignment error tends to be non-differential [36]. Additionally, in this study, we calculated PM10–2.5 concentrations using data obtained from co-located PM10 and PM2.5 monitors to avoid potential bias. However, PM10 monitoring had been greatly reduced (typically every 6 days for PM10, compared to typically every 3 days for PM2.5), which limited the ability to calculate PM10–2.5 data. This was a particular concern for the Los Angeles site, where only 24% of women had matched PM10–2.5 measurements. The mean average one-year PM10–2.5 exposure in our study ranged from 8.5 ± 3.8 μg/ m3 to 17.5 ± 2.3 μg/m3 at different sites, which aligned with the levels reported in the literature.

A few other limitations also need to be considered. First, as a common issue in cohort studies, loss of follow-up was also observed in SWAN, which reduced statistical power. In our dataset, 35% of participants had completed one or two visits only, including almost all participants at Newark site and half of the participants at Los Angeles site, which was equivalent to 20% of visits. We were not able to run the regression model within these subjects because of the small sample size; however, sensitivity tests excluding these participants confirmed the associations we observed (Table S2), including those from stratified analyses (not shown). Second, potential confounding factors, such as, noise, greenness, household income, were not included in this analysis due to limited available data. Future studies are recommended to include them into considerations. Lastly, heterogeneity of PM10–2.5 compositions by study location may influence the associations with biomarkers, but composition information was not available to use in our study.


The results of this study support the hypothesis that long-term exposure to coarse particulate matter may contribute to inflammation and thrombosis, markers of CVD. The observed associations between long-term PM10–2.5 exposure and PAI-1 were independent, as they were not confounded by ambient PM2.5, ozone, NO2, or CO. Particular subgroups were more sensitive to PM10–2.5 exposure. Further epidemiological and toxicological studies are warranted to identify the specific mechanisms of how PM10–2.5 may affect these biomarkers.

Availability of data and materials

The data that support the findings of this study are available from California Office of Statewide Health Planning and Development (OSHPD) but restrictions apply to the availability of these data, which were used under license for the current study, and so are not publicly available. Data are however available from the authors upon reasonable request and with permission of OSHPD.



Cardiovascular disease

PM2.5 :

Fine particulate matter

PM10–2.5 :

Coarse particulate matter


Study of Women’s Health Across the Nation


United States Environmental Protection Agency


High sensitivity C-reactive protein

factor VIIc:

Factor VII coagulant


Tissue-type plasminogen activator antigen


Plasminogen activator inhibitor Type 1


Enzyme-linked immunosorbent assay

O3 :



Carbon monoxide

NO2 :

Nitrogen dioxide

SO2 :

Sulfur dioxide


Body mass index


Akaike Information Criterion


ZIP code tabulation area


  1. Garcia M, Mulvagh SL, Merz CN, Buring JE, Manson JE. Cardiovascular disease in women: clinical perspectives. Circ Res. 2016;118(8):1273–93.

    Article  CAS  Google Scholar 

  2. Particle Pollution and Cardiovascular Effects [].

  3. Brook RD, Rajagopalan S, Pope CA 3rd, Brook JR, Bhatnagar A, Diez-Roux AV, et al. Particulate matter air pollution and cardiovascular disease: an update to the scientific statement from the American Heart Association. Circulation. 2010;121(21):2331–78.

    Article  CAS  Google Scholar 

  4. USEPA. Integrated Science Assessment (ISA) for Particulate Matter (Final Report, 2019). Washington, DC, EPA/600/R-19/188: U.S. Environmental Protection Agency; 2019.

    Google Scholar 

  5. Pakkanen TA, Loukkola K, Korhonen CH, Aurela M, Mäkelä T, Hillamo RE, et al. Sources and chemical composition of atmospheric fine and coarse particles in the Helsinki area. Atmos Environ. 2001;35(32):5381–91.

    Article  CAS  Google Scholar 

  6. Chen YC, Weng YH, Chiu YW, Yang CY. Short-term effects of coarse particulate matter on hospital admissions for cardiovascular diseases: a case-crossover study in a Tropical City. J Toxicol Environ Health A. 2015;78(19):1241–53.

    Article  CAS  Google Scholar 

  7. Lee H, Honda Y, Hashizume M, Guo YL, Wu CF, Kan H, et al. Short-term exposure to fine and coarse particles and mortality: a multicity time-series study in East Asia. Environ Pollut. 2015;207:43–51.

    Article  CAS  Google Scholar 

  8. Malig BJ, Green S, Basu R, Broadwin R. Coarse particles and respiratory emergency department visits in California. Am J Epidemiol. 2013;178(1):58–69.

    Article  Google Scholar 

  9. Powell H, Krall JR, Wang Y, Bell ML, Peng RD. Ambient coarse particulate matter and hospital admissions in the Medicare cohort air pollution study, 1999-2010. Environ Health Perspect. 2015;123(11):1152–8.

    Article  CAS  Google Scholar 

  10. Yorifuji T, Kashima S, Doi H. Associations of acute exposure to fine and coarse particulate matter and mortality among older people in Tokyo, Japan. Sci Total Environ. 2016;542(Pt A):354–9.

    Article  CAS  Google Scholar 

  11. Zhao Y, Wang S, Lang L, Huang C, Ma W, Lin H. Ambient fine and coarse particulate matter pollution and respiratory morbidity in Dongguan, China. Environ Pollut. 2017;222:126–31.

    Article  CAS  Google Scholar 

  12. Adar SD, Filigrana PA, Clements N, Peel JL. Ambient coarse particulate matter and human health: a systematic review and meta-analysis. Curr Environ Health Rep. 2014;1:258–74.

    Article  Google Scholar 

  13. Brunekreef B, Forsberg B. Epidemiological evidence of effects of coarse airborne particles on health. Eur Respir J. 2005;26(2):309–18.

    Article  CAS  Google Scholar 

  14. D'Souza JC, Kawut SM, Elkayam LR, Sheppard L, Thorne PS, Jacobs DR Jr, et al. Ambient coarse particulate matter and the right ventricle: the multi-ethnic study of atherosclerosis. Environ Health Perspect. 2017;125(7):077019.

    Article  Google Scholar 

  15. Honda T, Eliot MN, Eaton CB, Whitsel E, Stewart JD, Mu L, et al. Long-term exposure to residential ambient fine and coarse particulate matter and incident hypertension in post-menopausal women. Environ Int. 2017;105:79–85.

    Article  Google Scholar 

  16. Adar SD, D'Souza J, Mendelsohn-Victor K, Jacobs DR, Cushman M, Sheppard L, et al. Markers of inflammation and coagulation after long-term exposure to coarse particulate matter: a cross-sectional analysis from the multi-ethnic study of atherosclerosis. Environ Health Perspect. 2015;123(6):541–8.

    Article  CAS  Google Scholar 

  17. Sowers M, Crawford S, Sternfeld B, Morganstein D, Gold E, Greendale G, et al. SWAN: A Multicenter, Multiethnic, Community-Based Cohort Study of Women and the Menopausal Transition. In: JK RAL, Marcus R, editors. Menopause: Biology and Pathobiology. San Diego: Academic Press; 2000. p. 175–88.

    Chapter  Google Scholar 

  18. Green R, Broadwin R, Malig B, Basu R, Gold EB, Qi L, et al. Long- and short-term exposure to air pollution and inflammatory/hemostatic markers in midlife women. Epidemiology. 2016;27(2):211–20.

    Google Scholar 

  19. Wu XM, Basu R, Malig B, Broadwin R, Ebisu K, Gold EB, Qi L, Derby C, Green RS. Association between gaseous air pollutants and inflammatory, hemostatic and lipid markers in a cohort of midlife women. Environ Int. 2017;107:131–9.

    Article  CAS  Google Scholar 

  20. Wang NC, Matthews KA, Barinas-Mitchell EJM, Chang C-CH, El Khoudary SR. Inflammatory/hemostatic biomarkers and coronary artery calcification in midlife women of African-American and white race/ethnicity: the study of Women's health across the nation (SWAN) heart study. Menopause. 2016;23(6):653–61.

    Article  Google Scholar 

  21. Chandler WL, Alessi MC, Aillaud MF, Henderson P, Vague P, Juhan-Vague I. Clearance of tissue plasminogen activator (TPA) and TPA/plasminogen activator inhibitor type 1 (PAI-1) complex: relationship to elevated TPA antigen in patients with high PAI-1 activity levels. Circulation. 1997;96(3):761–8.

    Article  CAS  Google Scholar 

  22. Declerck PJ, Alessi MC, Verstreken M, Kruithof EK, Juhan-Vague I, Collen D. Measurement of plasminogen activator inhibitor 1 in biologic fluids with a murine monoclonal antibody-based enzyme-linked immunosorbent assay. Blood. 1988;71(1):220–5.

    Article  CAS  Google Scholar 

  23. Derby CA, Crawford SL, Pasternak RC, Sowers M, Sternfeld B, Matthews KA. Lipid changes during the menopause transition in relation to age and weight: the study of Women's health across the nation. Am J Epidemiol. 2009;169(11):1352–61.

    Article  Google Scholar 

  24. Chen JC, Schwartz J. Metabolic syndrome and inflammatory responses to long-term particulate air pollutants. Environ Health Perspect. 2008;116(5):612–7.

    Article  Google Scholar 

  25. Forbes LJ, Patel MD, Rudnicka AR, Cook DG, Bush T, Stedman JR, Whincup PH, Strachan DP, Anderson RH. Chronic exposure to outdoor air pollution and markers of systemic inflammation. Epidemiology. 2009;20(2):245–53.

    Article  Google Scholar 

  26. Pakbin P, Ning Z, Shafer MM, Schauer JJ, Sioutas C. Seasonal and spatial coarse particle elemental concentrations in the Los Angeles area. Aerosol Sci Technol. 2011;45(8):949–63.

    Article  CAS  Google Scholar 

  27. Deng Q, Deng L, Miao Y, Guo X, Li Y. Particle deposition in the human lung: health implications of particulate matter from different sources. Environ Res. 2019;169:237–45.

    Article  CAS  Google Scholar 

  28. Onishi T, Honda A, Tanaka M, Chowdhury PH, Okano H, Okuda T, et al. Ambient fine and coarse particles in Japan affect nasal and bronchial epithelial cells differently and elicit varying immune response. Environ Pollut. 2018;242:1693–701.

    Article  CAS  Google Scholar 

  29. Ljubimova JY, Braubach O, Patil R, Chiechi A, Tang J, Galstyan A, et al. Coarse particulate matter (PM2.5–10) in Los Angeles Basin air induces expression of inflammation and cancer biomarkers in rat brains. Sci Rep. 2018;8(1):5708.

    Article  Google Scholar 

  30. Alexis NE, Lay JC, Zeman K, Bennett WE, Peden DB, Soukup JM, Devlin RB, Becker S. Biological material on inhaled coarse fraction particulate matter activates airway phagocytes in vivo in healthy volunteers. J Allergy Clin Immunol. 2006;117(6):1396–403.

    Article  CAS  Google Scholar 

  31. Happo MS, Salonen RO, Halinen AI, Jalava PI, Pennanen AS, Dormans JA, et al. Inflammation and tissue damage in mouse lung by single and repeated dosing of urban air coarse and fine particles collected from six European cities. Inhal Toxicol. 2010;22(5):402–16.

    Article  CAS  Google Scholar 

  32. Kleinman MT, Sioutas C, Chang MC, Boere AJ, Cassee FR. Ambient fine and coarse particle suppression of alveolar macrophage functions. Toxicol Lett. 2003;137(3):151–8.

    Article  CAS  Google Scholar 

  33. Monn C, Becker S. Cytotoxicity and induction of proinflammatory cytokines from human monocytes exposed to fine (PM2.5) and coarse particles (PM10-2.5) in outdoor and indoor air. Toxicol Appl Pharmacol. 1999;155(3):245–52.

    Article  CAS  Google Scholar 

  34. Schins RP, Lightbody JH, Borm PJ, Shi T, Donaldson K, Stone V. Inflammatory effects of coarse and fine particulate matter in relation to chemical and biological constituents. Toxicol Appl Pharmacol. 2004;195(1):1–11.

    Article  CAS  Google Scholar 

  35. Cesari M, Pahor M, Incalzi RA. Plasminogen activator inhibitor-1 (PAI-1): a key factor linking fibrinolysis and age-related subclinical and clinical conditions. Cardiovasc Ther. 2010;28(5):e72–91.

    Article  CAS  Google Scholar 

  36. Wilson WE, Suh HH. Fine particles and coarse particles: concentration relationships relevant to epidemiologic studies. J Air Waste Manag Assoc. 1997;47(12):1238–49.

    Article  CAS  Google Scholar 

Download references


The Study of Women’s Health Across the Nation (SWAN) has grant support from the National Institutes of Health (NIH), DHHS, through the National Institute on Aging (NIA), the National Institute of Nursing Research (NINR) and the NIH Office of Research on Women’s Health (ORWH) (Grants U01NR004061; U01AG012505, U01AG012535, U01AG012531, U01AG012539, U01AG012546, U01AG012553, U01AG012554, U01AG012495). The content of this manuscript is solely the responsibility of the authors and does not necessarily represent the official views of the NIA, NINR, ORWH or the NIH.

The SWAN Repository (U01AG017719).

This publication was supported in part by the National Center for Research Resources and the National Center for Advancing Translational Sciences, National Institutes of Health, through UCSF-CTSI Grant Number UL1 RR024131. Support for address geocoding was provided by California Institute for Energy and Environment, University of California (Subaward No. POB228-X86).

Clinical Centers: University of Michigan, Ann Arbor – Siobán Harlow, PI 2011 – present, MaryFran Sowers, PI 1994–2011; Massachusetts General Hospital, Boston, MA – Joel Finkelstein, PI 1999 – present; Robert Neer, PI 1994–1999; Rush University, Rush University Medical Center, Chicago, IL – Howard Kravitz, PI 2009 – present; Lynda Powell, PI 1994–2009; University of California, Davis/Kaiser – Ellen Gold, PI; University of California, Los Angeles – Gail Greendale, PI; Albert Einstein College of Medicine, Bronx, NY – Carol Derby, PI 2011 – present, Rachel Wildman, PI 2010–2011; Nanette Santoro, PI 2004–2010; University of Medicine and Dentistry – New Jersey Medical School, Newark – Gerson Weiss, PI 1994–2004; and the University of Pittsburgh, Pittsburgh, PA – Karen Matthews, PI.

NIH Program Office: National Institute on Aging, Bethesda, MD – Chhanda Dutta 2016- present; Winifred Rossi 2012–2016; Sherry Sherman 1994–2012; Marcia Ory 1994–2001; National Institute of Nursing Research, Bethesda, MD – Program Officers.

Central Laboratory: University of Michigan, Ann Arbor – Daniel McConnell (Central Ligand Assay Satellite Services).

SWAN Repository: University of Michigan, Ann Arbor – Siobán Harlow 2013 - Present; Dan McConnell 2011–2013; MaryFran Sowers 2000–2011.

Coordinating Center: University of Pittsburgh, Pittsburgh, PA – Maria Mori Brooks, PI 2012 - present; Kim Sutton-Tyrrell, PI 2001–2012; New England Research Institutes, Watertown, MA - Sonja McKinlay, PI 1995–2001.

Steering Committee: Susan Johnson, Current Chair.

Chris Gallagher, Former Chair.

We thank the study staff at each site and all the women who participated in SWAN.

The authors also want to thank Ms. Dharshani Pearson for reviewing the manuscript and providing valuable input.

The content of this manuscript is solely the responsibility of the authors and does not necessarily represent the official views of the Office of Environmental Health Hazard Assessment, California Environmental Protection Agency, or the State of California.

Author information

Authors and Affiliations



ED conducted data analysis and prepared the first draft. BM and RBroadwin helped with the conception of the hypothese and prepared datasets. KE and RBasu provided key input for refining analysis and manuscript. EG, CD, and SP were the PIs of individual SWAN sites, and supervised the data collection at their sites. XW oversaw the whole study, including proposed the hypothesis, designed data analysis plan, supervised the data analysis, conducted results validation, and developed the manuscript based on the first draft. All authors helped with reviewing and finalizing this manuscript. All authors read and approved the final manuscript.

Corresponding author

Correspondence to Xiangmei (May) Wu.

Ethics declarations

Ethics approval and consent to participate

The SWAN protocols and this study of air pollution were approved by the Institutional Review Boards at all six participating sites, and all participants provided signed, written informed consent for the SWAN protocols at study recruitment.

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Additional file 1:

Table S1. Distribution of prior one-year average PM2.5, ozone, CO, NO2, and SO2 concentrations by SWAN site, 1999–2004. Table S2. Results of sensitivity test for the associations between PM10–2.5 and cardiovascular disease markers based on several models for SWAN cohort, 1999–2004. Table S3. Adjusted associations between PM10–2.5 and PM2.5, Ozone, CO, NO2, SO2 and cardiovascular disease markers for SWAN cohort, 1999–2004. (docx 29 kb)

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit The Creative Commons Public Domain Dedication waiver ( applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Davis, E., Malig, B., Broadwin, R. et al. Association between coarse particulate matter and inflammatory and hemostatic markers in a cohort of midlife women. Environ Health 19, 111 (2020).

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: