Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2019 Aug 14.
Published in final edited form as: Environ Int. 2019 Jan 12;124:153–160. doi: 10.1016/j.envint.2018.12.044

Case-crossover Analysis of Short-term Particulate Matter Exposures and Stroke in the Health Professionals Follow-up Study

Jared A Fisher a,b, Robin C Puett b, Francine Laden c,d,e, Gregory A Wellenius f, Amir Sapkota b, Duanping Liao g, Jeff D Yanosky g, Olivia Carter-Pokras a, Xin He a, Jaime E Hart c,d
PMCID: PMC6692897  NIHMSID: NIHMS1044649  PMID: 30641259

Abstract

Background:

Stroke is a leading cause of morbidity and mortality in the United States. Associations between short-term exposures to particulate matter (PM) air pollution and stroke are inconsistent. Many prior studies have used administrative and hospitalization databases where misclassification of the type and timing of the stroke event may be problematic.

Methods:

In this case-crossover study, we used a nationwide kriging model to examine short-term ambient exposure to PM10 and PM2.5 and risk of ischemic and hemorrhagic stroke among men enrolled in the Health Professionals Follow-up Study. Conditional logistic regression models were used to obtain estimates of odds ratios (OR) and 95% confidence intervals (CI) associated with an interquartile range (IQR) increase in PM2.5 or PM10. Lag periods up to 3 days prior to the stroke event were considered in addition to a 4-day average. Stratified models were used to examine effect modification by patient characteristics.

Results:

Of the 727 strokes that occurred between 1999 and 2010, 539 were ischemic and 122 were hemorrhagic. We observed positive statistically significant associations between PM10 and ischemic stroke (ORlag0–3=1.26; 95% CI: 1.03–1.55 per IQR increase [14.46μg/m3]), and associations were elevated for nonsmokers, aspirin nonusers, and those without a history of high cholesterol. However, we observed no evidence of a positive association between short-term exposure to PM and hemorrhagic stroke or between PM2.5 and ischemic stroke in this cohort.

Conclusions:

Our study provides evidence that ambient PM10 may be associated with higher risk of ischemic stroke and highlights that ischemic and hemorrhagic strokes are heterogeneous outcomes that should be treated as such in analyses related to air pollution.

Keywords: Particulate matter, stroke, air pollution, case-crossover

1. Introduction

Stroke is a leading cause of morbidity and mortality in the United States with 795,000 people experiencing a new or recurrent stroke every year (Benjamin et al., 2017). Of these, approximately 140,000 result in mortality, making stroke the fifth leading cause of death in the United States (Benjamin et al., 2017; CDC, 2017).

While a large and growing body of research has established the role of ambient exposures to air pollution and increased risks of cardiovascular mortality and morbidity (Brook et al., 2010), the evidence linking particulate matter (PM) air pollution exposures to the risk of stroke remains equivocal. In the epidemiologic literature, there have been two approaches to examine this relationship: studies of long-term PM exposure as a contributor to higher incidence of stroke and studies of short-term PM exposure as an acute trigger of stroke (Ljungman and Mittleman, 2014; Maheswaran, 2016). Prior studies of short-term PM exposure and stroke have provided inconsistent results, with some (Dominici et al., 2006; Wellenius et al., 2012; Zanobetti and Schwartz, 2009) but not all (Anderson et al., 2001; O’Donnell et al., 2011; Vivanco-Hidalgo et al., 2018) studies finding evidence of increased risk.

A major limitation of much of the prior research in this domain has been the reliance on administrative and hospitalization databases where diagnostic and coding errors are frequent (Leibson et al., 1994; Reker et al., 2002) and where ischemic and hemorrhagic strokes are often considered as a single outcome, despite differing pathophysiological pathways (Ljungman and Mittleman, 2014). Additionally, exposure misclassification in analyses relying on administrative databases can bias the results towards the null by up to 60%, as the timing of the event and the hospitalization are not always congruent (Lokken et al., 2009). For these reasons, a recent review has urged for additional studies using cohort or registry data with well-classified and medically-reviewed outcome data (Wang et al., 2014).

In this case-crossover study, we used a nationwide kriging model to examine the association between short-term ambient exposure to PM less than 10 or less than 2.5 microns in diameter (PM10 and PM2.5) and risk of ischemic and hemorrhagic stroke among participants of the Health Professionals Follow-up Study (HPFS) cohort. Reported strokes were confirmed by medical record review to ensure accuracy and to obtain the exact date of the stroke event. Effect modification by several individual-level characteristics was also examined to identify potentially susceptible subpopulations.

2. Methods

2.1. Study population

The HPFS is an ongoing, prospective cohort study of 51,529 men recruited nationally who were aged 40–75 years at baseline in 1986. Participants were dentists, pharmacists, optometrists, osteopath physicians, podiatrists, or veterinarians at the time of enrollment. Every two years, study participants receive mailed questionnaires at their residential or work address with questions about disease, medical history, and health-related risks and behaviors. Response rates to questionnaires are generally >90%. Our case-crossover study population included all participants who experienced a stroke between 1999 and 2010, the dates for which exposure data were available. Cases were included if both an exact date of the stroke event was known and the participant’s addresses on the questionnaire immediately before the stroke event was in the conterminous US.

The HPFS protocol was approved by the Harvard T.H. Chan School of Public Health Human Subjects Committee and the Brigham and Women’s Hospital Institutional Review Board, and all participants provided implied consent through returning the questionnaires and written informed consent for release of medical records.

2.2. Outcome assessment

Strokes were self-reported on the baseline and follow-up questionnaires by participants. Reported cases were adjudicated by trained study physicians reviewing medical records using a standardized approach. Strokes were confirmed when medical records documented a neurologic deficit with sudden or rapid onset persisting for >24 hours without evidence for other causes, unless death supervened or there was a demonstrable lesion compatible with acute stroke on brain imaging studies. Strokes were classified as ischemic, hemorrhagic, or of undetermined type and the exact date of the stroke event was noted. Death events were ascertained via communication with proxy respondents and/or National Death Index searches. Multiple stroke events from individual participants were included in the analyses if the dates were separated by at least 1 year.

2.3. Exposure assessment

All mailing addresses for the HPFS participants were geocoded to determine latitude and longitude and were used to create subject-specific address histories. Validated national-scale, log-normal ordinary kriging models for PM2.5 and PM10 were used to estimate daily values at the addresses of all HPFS participants. These models have been explained in detail elsewhere (Liao et al., 2006). Briefly, all ambient PM10 and PM2.5 data recorded at over 1,000 monitors operating in the contiguous United States during the study period, 1999–2010, were obtained from the U.S. Environmental Protection Agency Air Quality System (AQS). Longitude and latitude were included in the record for each monitoring site, and monitor-specific daily averages were calculated for monitors with ≥18 hourly measures. Relying on these input data, a semiautomated program built using ArcMap (ESRI, version 8.3) and its Geostatistical Analyst extension (ESRI, Inc. Redlands, CA) was created to produce kriging-estimated daily mean concentrations at each geocoded participant address. The program relied on a spherical model to perform log-normal ordinary kriging at a national-scale and used the weighted least-squares method to estimate semivariograms. Cross-validation statistics were used to evaluate the validity of the estimation method. As detailed in Liao et al., 2006, cross-validation parameters for year 2000, including average prediction error and standardized prediction error near 0, provided evidence of model validity and have justified the use of these models to provide address-specific PM estimates for other population-based health outcome studies (Gondalia et al., 2017; Liao et al., 2008, 2007, 2006; Shih et al., 2011; Whitsel et al., 2009; Zhang et al., 2009). Subsequent validations have shown these cross-validation parameters are consistent for both PM2.5 and PM10 models for years 2000, 2005, and 2010 (Supplemental Table 1). PM2.5 and PM10 estimates were available from these models at each address for each day from January 1, 1999 to December 31, 2010.

2.4. Covariates

Ambient temperature was considered as a potential confounder in this analysis, and daily mean temperature values were predicted for all geocoded addresses from 1999 to 2010. Data on air temperature were obtained from the Modern Era Retrospective-analysis for Research and Applications (MERRA) project (NASA, 2017). MERRA data are available hourly on a grid across the continental US with an approximately 55 km cell size. Hourly gridded data were assigned local time and date based on the time zone of each grid point’s location and then averaged by day. The daily averages were included in spatially-smoothed generalized additive models (GAMs) which were then used for space-time prediction of daily temperature values at any location in the conterminous United States.

Several individual-level characteristics which may modify the relationship between PM exposures and stroke were assessed as potential effect modifiers (O’Donnell et al., 2011; Villeneuve et al., 2012). These variables included: age (<70, ≥70 years old), body mass index (BMI) (< 25.0, < 30.0, or ≥ 30.0 kg/m2), smoking status (current, former, never), Diabetes mellitus (‘yes’ or ‘no’ as to ever having been diagnosed by health professional), hypertension (‘yes’ or ‘no’ as to ever having been diagnosed by health professional), hypercholesterolemia (‘yes’ or ‘no’ as to ever having been diagnosed by health professional), and current aspirin use (‘yes’ as 2+ tablets per week; ‘no’ otherwise). Responses were obtained from the most recent HPFS questionnaire prior to the stroke event.

2.5. Statistical Analysis

A time-stratified case-crossover study design was used to analyze the association between exposure to ambient particulate matter and stroke events. Proposed by Maclure, the case-crossover design has increasingly been used to examine transient effects on the risk of acute events (Maclure, 1991). With this design, each case’s exposure just prior to the event is compared to exposure at other referent periods. In this way, each case serves as his own control, and confounding by invariant and slowly changing risk factors is controlled.

To prevent time-trend bias and to ensure unbiased conditional logistic regression estimates, we used a time-stratified approach with month-long strata for referent period selection (Janes et al., 2005). The case period was defined as day of or days previous to the stroke event, and referent periods (n=3 or 4) were defined as other days in the same calendar month, matched by day of the week.

Using SAS 9.4 (SAS Institute Inc., Cary, NC), conditional logistic regression models were used to obtain estimates of odds ratios (ORs) and 95% confidence intervals (CIs) associated with an interquartile range (IQR) increase in PM2.5 or PM10. PM exposures on the day of the stroke event (lag 0), as well as exposures from 1 to 3 days previous to the stroke event (lag 1 to lag 3) were considered in separate models. Additionally, a 4-day average of PM exposures (lag 0 to lag 3) was included to represent the total exposure to PM in the 4-day period.

Linear terms for mean daily temperature (°C) were included in each model as a control variable. Effect modification of the relationship between PM and both ischemic and hemorrhagic stroke was examined through stratified conditional logistic models. Significance testing between the stratified estimates was assessed using chi-square tests of model heterogeneity. Sensitivity analyses included the examination of effect modification by season and employment status, use of restricted cubic splines to test nonlinear associations, and the extension of lag days to 6 days before the stroke event.

3. Results

A total of 731 stroke cases were medically-reviewed and had a complete record of the date of event; 727 of these occurred in the conterminous US and were successfully geocoded. Five participants had multiple stroke events during the time period (all separated by at least 1 year). Of the 727 cases, 539 (74.1%) were ischemic strokes, 122 (16.8%) were hemorrhagic stokes, and 66 (9.1%) were of undetermined type. Additionally, of the 727 stroke events, 165 (22.7%) were fatal and 562 (77.3%) were non-fatal.

PM exposures on the day of the event as well as descriptive characteristics are presented for total, ischemic, hemorrhagic, and undetermined stroke cases in Table 1. Exposures to PM and outdoor ambient temperature were lower on the day of the event for hemorrhagic stroke than for ischemic stroke using two-sample independent t-tests (p<0.01). Compared to hemorrhagic stroke cases, ischemic stroke cases were slightly younger, had higher BMIs, and were more likely to be current or former smokers. Stroke events were more often fatal among hemorrhagic strokes (49.2%) than among ischemic strokes (10.8%). Over half of the total stroke cases reported having ever had hypertension (61.8%), having ever had high cholesterol (58.5%), or were regularly taking aspirin at the time of the last survey before the stroke (58.5%). Proportions were similar between stroke types for having ever reported hypertension or being a current regular aspirin user. However, a higher proportion of ischemic stroke cases reported having ever had high cholesterol (60.3%) compared to reports from hemorrhagic stroke cases (49.2%).

Table 1.

Descriptive characteristics and event day exposures for men with an ischemic, hemorrhagic or undetermined type stroke event from 1999 to 2010 in the Health Professionals Follow-up Study

Total Stroke
(n=727)		 Ischemic Stroke
(n=539)		 Hemorrhagic Stroke
(n=122)		 Undetermined Type
(n=66)
mean ± SD or N (%) mean ± SD or N (%) mean ± SD or N (%) mean ± SD or N (%)
PM2.5 (μg/m3)a 12.9 ± 7.4 13.1 ± 7.6 11.9 ± 6.7 13.7 ± 7.6
PM10 (μg/m3)a 26.3 ± 12.3 26.8 ± 12.7 23.6 ± 10.1* 26.8 ± 11.9
Temperature (C)a 12.5 ± 10.4 12.9 ± 10.3 10.2 ± 11.2* 13.4 ± 9.5
Age (years) 76.3 ± 8.4 75.2 ± 8.3 77.3 ± 8.0 83.2 ± 6.8
Stroke Outcome
 Fatal 165 (22.7%) 58 (10.8%) 60 (49.2%)* 47 (71.2%)
 Non-fatal 562 (77.3%) 481 (89.2%) 62 (50.8%) 19 (28.8%)
BMI (kg/m2)
 < 25 271 (37.4%) 180 (33.6 %) 54 (44.3 %) 37 (56.1%)
 ≥25 to < 30 353 (48.8%) 277 (51.7%) 54 (44.3%) 22 (33.3%)
 ≥30 100 (13.8%) 79 (14.7%) 14 (11.5%) 7 (10.6%)
Smoking status
 Never 261 (35.9%) 191 (35.4%) 51 (41.8%)* 19 (28.8%)
 Former 347 (47.7%) 270 (50.1%) 46 (37.7%) 31 (47.0%)
 Current 40 (5.5%) 32 (5.9%) 4 (3.3%) 4 (6.1%)
 Missing 79 (10.9%) 46 (8.5%) 21 (17.2%) 12 (18.2%)
Diabetes “Yes”b 109 (15.0%) 81 (15.0%) 18 (14.8%) 10 (15.2%)
Ever Hypertension “Yes”b 449 (61.8%) 337 (62.5%) 73 (59.8%) 39 (59.1%)
Ever High Cholesterol “Yes”b 425 (58.5%) 325 (60.3%) 60 (49.2%)* 40 (60.6%)
Current Regular Aspirin Use “Yes”b 425 (58.5%) 316 (58.6%) 75 (61.5%) 34 (51.5%)
Open in a new tab
a

Daily estimate on the day of the stroke event

b

Self-report from biennial questionnaire: For diabetes, hypertension, high cholesterol: ‘yes’ or ‘no’ is defined as ever having been diagnosed by health professional. For current aspirin use ‘yes’ is defined as 2+ tablets /week

*

Denotes statistically significant (p<0.05) difference between ischemic and hemorrhagic strokes in two-sample t-tests or Chi-square tests

ORs and 95% CIs between short-term PM exposures and total stroke, stroke type, and stroke outcome are presented in Table 2. When all stroke events were considered together, no statistically significant associations were observed with any of the PM exposure measures. For example, the association between same-day PM2.5 and total stroke was OR=1.01 (95% CI: 0.90, 1.14) per IQR [8.91μg/m3] increase, and between same-day PM10 and total stroke the association was OR=1.08 (95% CI: 0.95, 1.22) per IQR [14.46μg/m3] increase.

Table 2.

Odds ratios and 95% confidence intervals per change in IQR for associations between short-term particulate matter exposures and stroke by stroke type and stroke outcome for men in the Health Professionals Follow-up Study

N Lag 0 Lag 1 Lag 2 Lag 3 Lag 0–3 (Avg)
OR (95% CI)a OR (95% CI)a OR (95% CI)a OR (95% CI)a OR (95% CI)a
Total Stroke 727
 PM2.5 1.01 (0.90, 1.14) 0.92 (0.82, 1.03) 0.93 (0.82, 1.04) 1.00 (0.89, 1.12) 0.94 (0.80, 1.10)
 PM10 1.08 (0.95, 1.22) 1.05 (0.93, 1.19) 1.06 (0.93, 1.20) 1.09 (0.96, 1.24) 1.14 (0.96, 1.36)
Ischemic Stroke 539
 PM2.5 1.02 (0.89, 1.16) 0.97 (0.86, 1.11) 0.95 (0.83, 1.09) 1.02 (0.90, 1.16) 0.99 (0.83, 1.18)
 PM10 1.13 (0.98, 1.30) 1.13 (0.99, 1.30) 1.11 (0.96, 1.28) 1.10 (0.95, 1.28) 1.26 (1.03, 1.55)
Hemorrhagic Stroke 122
 PM2.5 0.88 (0.64, 1.20) 0.71 (0.50, 0.99) 0.99 (0.75, 1.31) 1.05 (0.79, 1.38) 0.82 (0.54, 1.24)
 PM10 0.84 (0.60, 1.17) 0.58 (0.39, 0.85) 0.83 (0.59, 1.16) 1.07 (0.79, 1.45) 0.68 (0.42, 1.08)
Undetermined Type 66
 PM2.5 1.29 (0.83, 1.99) 0.67 (0.38, 1.17) 0.53 (0.31, 0.91) 0.70 (0.44, 1.12) 0.60 (0.30, 1.21)
 PM10 1.06 (0.69, 1.64) 1.17 (0.80, 1.71) 1.00 (0.64, 1.56) 0.99 (0.61, 1.60) 1.14 (0.62, 2.09)
Nonfatal Stroke 562
 PM2.5 1.06 (0.93, 1.21) 0.98 (0.86, 1.12) 0.98 (0.86, 1.12) 1.04 (0.92, 1.19) 1.04 (0.86, 1.24)
 PM10 1.16 (1.01, 1.33) 1.12 (0.98, 1.29) 1.11 (0.96, 1.28) 1.10 (0.95, 1.28) 1.28 (1.04, 1.56)
Fatal Stroke 165
 PM2.5 0.87 (0.67, 1.12) 0.69 (0.52, 0.92) 0.76 (0.59, 0.99) 0.86 (0.66, 1.11) 0.67 (0.47, 0.96)
 PM10 0.82 (0.62, 1.08) 0.79 (0.59, 1.05) 0.88 (0.66, 1.16) 1.03 (0.78, 1.36) 0.78 (0.53, 1.15)
Open in a new tab
a

ORs calculated from conditional logistic regression models and presented by change in PM IQR (PM2.5IQR =8.91μg/m3; PM10IQR = 14.46 μg/m3). All ORs control for mean daily temperature.

We observed a statistically significant positive association between 4-day cumulative lag PM10 and ischemic stroke (ORLag0_3=1.26; 95% CI: 1.03–1.55 per IQR increase), as well as suggestive positive associations for the individual lag day models. Statistically significant associations were also observed for an IQR increase in PM10 and nonfatal stroke on the day of stroke event (ORLag0=1.16; 95% CI: 1.01–1.33) and the 4-day cumulative lag period (ORLag0_3=1.28; 95% CI: 1.04–1.56), with suggestive positive associations for the other lag periods.

In contrast to the positive results for ischemic stroke, we observed statistically significant negative associations for hemorrhagic stroke and previous-day IQR increases in PM2.5 (OR=0.71; 95% CI: 0.50 – 0.99) and PM10 (OR=0.58; 95% CI: 0.39 – 0.85) and for undetermined stroke type and PM2.5 two days prior to the stroke event (OR=0.53; 95% CI: 0.31 – 0.91). Likewise, associations for fatal stroke in the same lag periods were also negative, as nearly 65% of fatal outcomes were from hemorrhagic and undetermined stroke events. No significant positive associations with PM2.5 exposures were observed by stroke type or stroke outcome. Generally, ORs for both PM2.5 and PM10 were higher for ischemic strokes compared to hemorrhagic strokes and higher among nonfatal strokes compared to fatal strokes. In tests of heterogeneity, differences between ischemic and hemorrhagic stroke results reached statistical significance in PM10 lag 1 models (ORisc=1.13; 95% CI: 0.99–1.30 vs ORhem=0.58; 95% CI: 0.39 −0.85; p<0.01) and PM10 lag0_3 models (ORisc=1.26; 95% CI: 1.03–1.55 vs ORhem=0.68; 95% CI: 0.42 −1.08; p=0.02).

We observed little evidence of effect modification by age, BMI, smoking status, diabetes, hypertension, hypercholesterolemia, or aspirin use. Forest plots of ORs and 95% CIs for associations between ischemic stroke and 4-day average PM levels are presented in Figure 1. In tests of heterogeneity, no statistically significant differences were observed for either PM10 or PM2.5 exposures. Associations were highest for the PM10 models among never smokers (OR= 1.63; 95% CI: 1.14 – 2.32), those without a history of high cholesterol (OR=1.53; 95% CI: 1.11 – 2.12) and those not currently taking aspirin (OR=1.48; 95% CI: 1.08 – 2.05). Associations were also higher among these same groups in the PM2.5 models. Findings were relatively consistent across individual lag day models for ischemic stroke (Supplemental Table 2).

Figure 1.

Figure 1.

Open in a new tab

Odds ratios and 95% confidence intervals per change in IQR, and p-values for tests of heterogeneity for associationsa between 4-day average particulate matter exposures (Lag0–3) and ischemic stroke by potential effect modifiers for men in the Health Professionals Follow-up Study

aORs calculated from conditional logistic regression models and presented by change in PM IQR (PM2.5IQR =8.91 μg/m3; PM10IQR = 14.46 μg/m3). All ORs control for mean daily temperature. p-value from chi-sq tests of model heterogeneity

b Self-report from biennial questionnaire: For diabetes, hypertension, high cholesterol: ‘yes’ or ‘no’ is defined as ever having been diagnosed by health professional. For current aspirin use ‘yes’ is defined as 2+ tablets /week

Forest plots of ORs and 95% CIs for associations between hemorrhagic stroke and 4-day average PM levels are presented in Figure 2. Due to the low number of cases and subsequent wide confidence intervals in risk estimates, comparisons were difficult between several hemorrhagic stroke subgroups. In particular, meaningful OR estimates could not be obtained for models among current smokers (N=4) and are not presented in Figure 2. Though the number of cases was small (N=14), there was a suggestion of increased risk among those with a BMI ≥30 for an IQR increase in both PM2.5 (OR=2.94; 95% CI: 0.88 – 9.77) and PM10 (OR=1.86; 95% CI: 0.64 – 5.42). Though most subgroup comparisons were similar between PM2.5 and PM10 models, results were less consistent across individual day lag periods for the hemorrhagic stroke models (Supplemental Table 3) compared to the ischemic stroke models. In sensitivity analysis, associations did not vary by employment status, and we saw no evidence of nonlinearity in the associations between PM or temperature with stroke using restricted cubic splines (data not shown). We also observed no evidence of effect modification by season for ischemic (Supplemental Table 4) or hemorrhagic stroke (Supplemental Table 5).

Figure 2.

Figure 2.

Open in a new tab

Odds ratios and 95% confidence intervals per change in IQR, and p-values for tests of heterogeneity for associationsa between 4-day average particulate matter exposures (Lag0–3) and hemorrhagic stroke by potential effect modifiers for men in the Health Professionals Follow-up Study

aORs calculated from conditional logistic regression models and presented by change in PM IQR (PM2.5IQR =8.91 μg/m3; PM10IQR = 14.46 μg/m3). All ORs control for mean daily temperature. p-value from chi-sq tests of model heterogeneity

b Self-report from biennial questionnaire: For diabetes, hypertension, high cholesterol: ‘yes’ or ‘no’ is defined as ever having been diagnosed by health professional. For current aspirin use ‘yes’ is defined as 2+ tablets /week

4. Discussion

We observed positive statistically significant associations between daily changes in ambient PM10 and ischemic and nonfatal stroke events. In contrast, we observed several negative statistically significant associations between short-term exposure to PM and hemorrhagic stroke during several lag periods. Although we did not observe statistically significant evidence of effect modification, associations between PM and ischemic stroke were elevated among nonsmokers, aspirin nonusers, and those having never had a diagnosis of high cholesterol, and the association between PM and hemorrhagic stroke appeared elevated among those in the highest BMI category. Associations were relatively consistent across lag periods for ischemic stroke, but quite variable for hemorrhagic stroke.

Our findings of higher relative odds of ischemic stroke associated with higher daily PM10 levels is consistent with several studies (Tsai et al., 2003; Wellenius et al., 2005; Wordley et al., 1997). Wellenius et al. examined a database of Medicare recipients in 9 U.S. cities and observed a small, but statistically significant, increase in ischemic stroke admissions with higher city-wide PM10 concentrations on the day of admission (1.03% increase per interquartile range increase [23.0μg/m3]; 95% CI: 0.04–2.04) (Wellenius et al., 2005). They did not observe associations between daily PM10 levels and hemorrhagic stroke (Wellenius et al., 2005). Another study by Tsai et al. of stroke admissions in Taiwan, observed that on days of ≥20°C, an interquartile range increase (66.3μg/m3) in PM10 was associated with both ischemic (OR=1.46; 95% CI: 1.32–1.61) and hemorrhagic (OR=1.54; 95% CI: 1.31–1.81) stroke admissions (Tsai et al., 2003). The evidence between short-term exposure to PM10 and ischemic stroke is not consistent, however. Other studies have observed no evidence of a relationship (Andersen et al., 2010; Chan et al., 2006; Henrotin et al., 2007; Mechtouff et al., 2012; Villeneuve et al., 2006).

Similarly, the existing literature between ambient PM2.5 and short-term stroke risk has been mixed, with some studies (Delfino et al., 2009; Dominici et al., 2006; Kettunen et al., 2007; Lisabeth et al., 2008; Wellenius et al., 2012) but not others (Lippmann et al., 2000; Mechtouff et al., 2012; O’Donnell et al., 2011; Villeneuve et al., 2006; Vivanco-Hidalgo et al., 2018) finding positive associations. Several of these studies used medically reviewed stroke cases. Wellenius et al. (2012) performed medical review of 1,705 Boston area patient records to confirm ischemic stroke events, and reported an OR for ischemic stroke of 1.11 (95% CI: 1.03–1.20) (P=0.006) per interquartile range increase in PM2.5 levels (6.4 μg/m3) (Wellenius et al., 2012). In a multicenter cohort study in Lyon, France, Mechtouff et al. used a case-crossover design with 376 medically-reviewed strokes. They observed no association between ischemic stroke and PM2.5 (OR=0.97; 95% CI: 0.83–1.12 per SD [10.4μg/m3] increase)(Mechtouff et al., 2012). Likewise, O’Donnell et al. used data from an acute stroke registry in Ontario, Canada and observed no association with PM2.5 among 9,202 confirmed cases of ischemic stroke (O’Donnell et al., 2011). Vivanco-Hidalgo et al. also used data from a clinical registry of stroke patients and observed no association between PM2.5 and ischemic stroke symptom onset among 2,742 patients in Barcelona, Spain (Vivanco-Hidalgo et al., 2018).

Our findings of statistically significant negative associations between both PM10 and PM2.5 and hemorrhagic stroke on the day prior to the stroke event were unexpected. Though a few prior studies have also reported significant negative associations between PM and stroke (Jalaludin et al., 2006; Talbott et al., 2014), such findings have been limited to certain lag periods, locations, seasons, or other category of total stroke events. Our findings for hemorrhagic stroke were not consistent for individual day lag periods, and the small sample of cases precluded an effective examination across certain subgroups.

Two recent meta-analyses of PM and short-term stroke risk summarized results of prior studies and found significant associations with both PM2.5 and PM10, with suggestions of a stronger effect for ischemic stroke (Shah et al., 2015; Wang et al., 2014). Additionally, Wang et al. 2014 found associations to be stronger in mortality studies, a finding not consistent with our results for fatal stroke. It remains unclear if the relationship between PM and stroke may vary consistently by stroke outcome or severity of stroke, as comparisons are often difficult to make between studies of differing designs. Additional studies are needed that include both fatal and nonfatal strokes or examine the relationship between PM and stroke using previously-defined or objective stroke severity criteria, such as two recent studies by Maheswaran et al. and Wing et al. (Maheswaran et al., 2016; Wing et al., 2017).

As most existing studies examining the short-term associations between stroke and particulate matter air pollution have been studies of administrative datasets, few have been able to examine individual-level variables as potential effect modifiers of the relationship. Such findings are important as they may identify individuals who are more susceptible to the effects of air pollution mediated stroke. Of the covariates we examined in this study, only age, which is available in many administrative datasets, has been tested widely as a potential effect modifier. Similar to our study, most of these studies observed no substantial difference in the PM/stroke association by age group (Anderson et al., 2001; Barnett et al., 2006; Delfino et al., 2009; Larrieu et al., 2007; Linn et al., 2000). Though our results did not reach statistical significance in tests of heterogeneity, we observed consistently elevated ORs for ischemic stroke among nonsmokers. Only a few studies have examined smoking status as a potential effect modifier of the relationship between short-term PM exposure and stroke; each finding no differences by smoking status (O’Donnell et al., 2011; Oudin et al., 2010).

To our knowledge, our study is also the first to examine effect modification by current aspirin usage on the PM-stroke relationship. Aspirin is an antiplatelet medication often given to those at elevated cardiovascular risk and is known to reduce the risk of ischemic stroke (Lei et al., 2016). In this study, we observed elevated ORs of ischemic stroke for aspirin nonusers compared to users across all lag periods and both PM fractions, though heterogeneity tests did not reach statistical significance. Although not directly comparable, Villeneuve et al. examined both current antiplatelet and anticoagulant medication use (medications unnamed) in an analysis of short-term NO2 exposure and stroke (Villeneuve et al., 2012). They observed no differences by antiplatelet use or nonuse, but found a significant association between NO2 exposure and ischemic stroke when limited to anti-coagulant nonusers (Villeneuve et al., 2012). As aspirin is more likely to be taken regularly among those with higher levels of cardiovascular risk, our findings of elevated associations among those with a previous diagnosis of high cholesterol is consistent with these results. Among ischemic stroke cases in our study, 68% and 65% of aspirin users reported having been diagnosed with high cholesterol and hypertension, respectively. Whether these results are evidence of a higher risk from PM-related stroke among those with little previously known cardiovascular risk, or whether medication use may potentially lower PM-related ischemic stroke risk is difficult to disentangle.

Our findings of elevated associations between PM and ischemic stroke among several subgroups not only helps to identify those potentially most vulnerable to PM-related stroke, but the findings may also shed light on pathophysiologic mechanisms by which exposure to PM may lead to stroke. Though the exact mechanisms are not yet fully understood, commonly hypothesized pathways include increased systemic inflammation and oxidative stress leading to procoagulant effects, autonomic irregularities, vascular endothelial dysfunction, and thrombosis which could enhance the likelihood of an ischemic event (Franchini and Mannucci, 2007; Mills et al., 2007). Our finding of reduced associations among smokers might be understood in this context, as cigarette smoke influences and may saturate these same pathways (Barnoya and Glantz, 2005). The specific role of platelet activation as a key mechanism of action has achieved recent attention with several studies showing an association between increased PM exposure and markers of platelet activation (Rich DQ et al., 2012; Wu et al., 2012). Of major interest to our study is a recent finding by Bacarerra et al. (2016) that some of the effects of increased ambient PM on platelet function were mitigated when subjects were taking aspirin (Becerra et al., 2016).

An important strength of this study is the use of high quality outcome and covariate data from HPFS. Many prior studies of short-term PM exposure and stroke risk are based on administrative data, where frequent misclassification of stroke events can bias effect estimates towards the null (Johnsen et al., 2002; Reker et al., 2002). Additionally, our study is one of few to examine patient characteristics, outside of age and gender, as potential effect modifiers. Another major strength of this study is the use of a nationwide kriging model to estimate exposure to ambient PM which may reduce exposure misclassification compared to prior studies that have used area-specific averages or a nearest-monitor approach.

This study also has several limitations. First, although the exposure models used in this study provided estimates of ambient PM at participants’ geocoded addresses, information was not available on how much time each participant spent at the address. Additionally, while studies have shown ambient concentrations of PM correlate relatively well with personal exposures (Avery et al., 2010), it should be noted that an individual’s total exposure to PM represents the sum of their exposure to particles both of outdoor origin as well as particles of indoor origin. For these reasons, some exposure misclassification is unavoidable and may have contributed to our lack of statistical findings. Second, the timing of the stroke event compared to the timing of hospitalization or death is not always congruent. Though we examined potential effects over multiple lag periods, this source of misclassification has been shown to bias effect estimates towards the null (Lokken et al., 2009). Third, although these data came from a large cohort, small sample sizes in this case-only study made an examination of hemorrhagic stroke risk within certain subgroups difficult. Fourth, we were not able to adjust our analyses for relative humidity, which may have led to some residual confounding. However, there is little evidence that relative humidity is an important determinant of stroke risk or an important confounder of the association between PM and stroke. Finally, the associations between PM and stroke may vary by ethnicity and socioeconomic status (Wing et al., 2017, 2015), but the HPFS study population is mostly white and of relatively higher socioeconomic status. Thus, our results may not be generalizable to younger men, women, those of non-Caucasian descent, or to individuals of lower socioeconomic means.

5. Conclusion

Our study provides more evidence that ischemic and hemorrhagic strokes are heterogeneous outcomes and should be treated as such in analyses related to air pollution. Our study also adds to previous results that show PM exposure may increase the risk of ischemic stroke and adds the finding that those of previously unknown cardiovascular risk may be at elevated risk for acute PM-related ischemic stroke events. Whether this is by means of managed risk by medication use or another mechanism requires more investigation.

Supplementary Material

Supplemental 1

Acknowledgements:

This report was supported by the National Institute of Environmental Health Sciences (NIEHS) [grants R01 ES020871, R01 ES017017, R03 ES016619, and P30 ES000002], the National Institute of Cancer [UM1 CA167552], and the National Heart, Lung, and Blood Institute (NHLBI) [R01 HL35464 and F32 HL083648]. The contents of this report are solely the responsibility of the authors and do not necessarily represent the official views of the sponsoring institutions.

Footnotes

The authors declare they have no actual or potential competing financial interests.

References

  1. Andersen ZJ, Olsen TS, Andersen KK, Loft S, Ketzel M, Raaschou-Nielsen O, 2010. Association between short-term exposure to ultrafine particles and hospital admissions for stroke in Copenhagen, Denmark. Eur. Heart J. 31, 2034–2040. 10.1093/eurheartj/ehq188 [DOI] [PubMed] [Google Scholar]
  2. Anderson HR, Bremner SA, Atkinson RW, Harrison RM, Walters S, 2001. Particulate matter and daily mortality and hospital admissions in the west midlands conurbation of the United Kingdom: associations with fine and coarse particles, black smoke and sulphate. Occup Environ Med 58, 504–510. 10.1136/oem.58.8.504 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Avery CL, Mills KT, Williams R, McGraw KA, Poole C, Smith RL, Whitsel EA, 2010. Estimating error in using residential outdoor PM2.5 concentrations as proxies for personal exposures: a meta-analysis. Environ. Health Perspect. 118, 673–678. 10.1289/ehp.0901158 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Barnett AG, Williams GM, Schwartz J, Best TL, Neller AH, Petroeschevsky AL, Simpson RW, 2006. The effects of air pollution on hospitalizations for cardiovascular disease in elderly people in Australian and New Zealand cities. Environ. Health Perspect. 114, 1018–1023. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Barnoya J, Glantz SA, 2005. Cardiovascular Effects of Secondhand Smoke. Circulation 111, 2684–2698. 10.1161/CIRCULATIONAHA.104.492215 [DOI] [PubMed] [Google Scholar]
  6. Becerra AZ, Georas S, Brenna JT, Hopke PK, Kane C, Chalupa D, Frampton MW, Block R, Rich DQ, 2016. Increases in ambient particulate matter air pollution, acute changes in platelet function, and effect modification by aspirin and omega-3 fatty acids: A panel study. Journal of Toxicology and Environmental Health, Part A 79, 287–298. 10.1080/15287394.2016.1157539 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Benjamin EJ, Blaha MJ, Chiuve SE, Cushman M, Das SR, Deo R, Ferranti S.D. de, Floyd J, Fornage M, Gillespie C, Isasi CR, Jiménez MC, Jordan LC, Judd SE, Lackland D, Lichtman JH, Lisabeth L, Liu S, Longenecker CT, Mackey RH, Matsushita K, Mozaffarian D, Mussolino ME, Nasir K, Neumar RW, Palaniappan L, Pandey DK, Thiagarajan RR, Reeves MJ, Ritchey M, Rodriguez CJ, Roth GA, Rosamond WD, Sasson C, Towfighi A, Tsao CW, Turner MB, Virani SS, Voeks JH, Willey JZ, Wilkins JT, Wu JH, Alger HM, Wong SS, Muntner P, Subcommittee, O. behalf of the A.H.A.S.C. and S.S., 2017. Heart Disease and Stroke Statistics—2017 Update: A Report From the American Heart Association. Circulation CIR.0000000000000485. 10.1161/CIR.0000000000000485 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Brook RD, Rajagopalan S, Pope CA, Brook JR, Bhatnagar A, Diez-Roux AV, Holguin F, Hong Y, Luepker RV, Mittleman MA, Peters A, Siscovick D, Smith SC, Whitsel L, Kaufman JD, American Heart Association Council on Epidemiology and Prevention, Council on the Kidney in Cardiovascular Disease, and Council on Nutrition, Physical Activity and Metabolism, 2010. Particulate matter air pollution and cardiovascular disease: An update to the scientific statement from the American Heart Association. Circulation 121, 2331–2378. 10.1161/CIR.0b013e3181dbece1 [DOI] [PubMed] [Google Scholar]
  9. CDC, 2017. FastStats - Leading Causes of Death [WWW Document]. URL http://www.cdc.gov/nchs/fastats/leading-causes-of-death.htm (accessed 9.20.17).
  10. Chan C-C, Chuang K-J, Chien L-C, Chen W-J, Chang W-T, 2006. Urban air pollution and emergency admissions for cerebrovascular diseases in Taipei, Taiwan. Eur. Heart J. 27, 1238–1244. 10.1093/eurheartj/ehi835 [DOI] [PubMed] [Google Scholar]
  11. Delfino RJ, Brummel S, Wu J, Stern H, Ostro B, Lipsett M, Winer A, Street DH, Zhang L, Tjoa T, Gillen DL, 2009. The relationship of respiratory and cardiovascular hospital admissions to the southern California wildfires of 2003. Occup Environ Med 66, 189–197. 10.1136/oem.2008.041376 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Dominici F, Peng RD, Bell ML, Pham L, McDermott A, Zeger SL, Samet JM, 2006. Fine Particulate Air Pollution and Hospital Admission for Cardiovascular and Respiratory Diseases. JAMA 295, 1127 10.1001/jama.295.10.1127 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Franchini M, Mannucci PM, 2007. Short-term effects of air pollution on cardiovascular diseases: outcomes and mechanisms. Journal of Thrombosis and Haemostasis 5, 2169–2174. 10.1111/j.1538-7836.2007.02750.x [DOI] [PubMed] [Google Scholar]
  14. Gondalia R, Avery CL, Napier MD, Méndez-Giráldez R, Stewart JD, Sitlani CM, Li Y, Wilhelmsen KC, Duan Q, Roach J, North KE, Reiner AP, Zhang Z-M, Tinker LF, Yanosky JD, Liao D, Whitsel EA, 2017. Genome-wide Association Study of Susceptibility to Particulate Matter-Associated QT Prolongation. Environ. Health Perspect. 125, 067002 10.1289/EHP347 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Henrotin JB, Besancenot JP, Bejot Y, Giroud M, 2007. Short-term effects of ozone air pollution on ischaemic stroke occurrence: a case-crossover analysis from a 10-year population-based study in Dijon, France. Occup Environ Med 64, 439–445. 10.1136/oem.2006.029306 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Jalaludin B, Morgan G, Lincoln D, Sheppeard V, Simpson R, Corbett S, 2006. Associations between ambient air pollution and daily emergency department attendances for cardiovascular disease in the elderly (65+ years), Sydney, Australia. J Expo Sci Environ Epidemiol 16, 225–237. 10.1038/sj.jea.7500451 [DOI] [PubMed] [Google Scholar]
  17. Janes H, Sheppard L, Lumley T, 2005. Case-crossover analyses of air pollution exposure data: referent selection strategies and their implications for bias. Epidemiology 16, 717–26. [DOI] [PubMed] [Google Scholar]
  18. Johnsen SP, Overvad K, Sørensen HT, Tjønneland A, Husted SE, 2002. Predictive value of stroke and transient ischemic attack discharge diagnoses in The Danish National Registry of Patients. Journal of Clinical Epidemiology 55, 602–607. 10.1016/S0895-4356(02)00391-8 [DOI] [PubMed] [Google Scholar]
  19. Kettunen J, Lanki T, Tiittanen P, Aalto PP, Koskentalo T, Kulmala M, Salomaa V, Pekkanen J, 2007. Associations of Fine and Ultrafine Particulate Air Pollution With Stroke Mortality in an Area of Low Air Pollution Levels. Stroke. [DOI] [PubMed] [Google Scholar]
  20. Larrieu S, Jusot J-F, Blanchard M, Prouvost H, Declercq C, Fabre P, Pascal L, Tertre AL, Wagner V, Rivière S, Chardon B, Borrelli D, Cassadou S, Eilstein D, Lefranc A, 2007. Short term effects of air pollution on hospitalizations for cardiovascular diseases in eight French cities: the PSAS program. Sci. Total Environ. 387, 105–112. 10.1016/j.scitotenv.2007.07.025 [DOI] [PubMed] [Google Scholar]
  21. Lei H, Gao Q, Liu S-R, Xu J, 2016. The Benefit and Safety of Aspirin for Primary Prevention of Ischemic Stroke: A Meta-Analysis of Randomized Trials. Front Pharmacol 7 10.3389/fphar.2016.00440 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Leibson CL, Naessens JM, Brown RD, Whisnant JP, 1994. Accuracy of hospital discharge abstracts for identifying stroke. Stroke 25, 2348–2355. 10.1161/01.STR.25.12.2348 [DOI] [PubMed] [Google Scholar]
  23. Liao D, Peuquet DJ, Duan Y, Whitsel EA, Dou J, Smith RL, Lin H-M, Chen J-C, Heiss G, 2006. GIS approaches for the estimation of residential-level ambient PM concentrations. Environ. Health Perspect. 114, 1374–1380. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Liao D, Peuquet DJ, Lin H-M, Duan Y, Whitsel EA, Smith RL, Heiss G, 2007. National Kriging Exposure Estimation: Liao et al. Respond. Environ Health Perspect 115, A338–A339. 10.1289/ehp.10205R [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Liao D, Whitsel EA, Duan Y, Lin H-M, Quibrera PM, Smith R, Peuquet DJ, Prineas RJ, Zhang Z-M, Anderson G, 2008. Ambient Particulate Air Pollution and Ectopy—The Environmental Epidemiology of Arrhythmogenesis in Women’s Health Initiative Study, 1999–2004. Journal of Toxicology and Environmental Health, Part A 72, 30–38. 10.1080/15287390802445483 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Linn WS, Szlachcic Y, Gong H, Kinney PL, Berhane KT, 2000. Air pollution and daily hospital admissions in metropolitan Los Angeles. Environ. Health Perspect. 108, 427–434. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Lippmann M, Ito K, Nádas A, Burnett RT, 2000. Association of particulate matter components with daily mortality and morbidity in urban populations. Res Rep Health Eff Inst 5–72, discussion 73–82. [PubMed] [Google Scholar]
  28. Lisabeth LD, Escobar JD, Dvonch JT, Sánchez BN, Majersik JJ, Brown DL, Smith MA, Morgenstern LB, 2008. Ambient air pollution and risk for ischemic stroke and transient ischemic attack. Ann. Neurol 64, 53–59. 10.1002/ana.21403 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Ljungman PL, Mittleman MA, 2014. Ambient air pollution and stroke. Stroke 45, 3734–3741. 10.1161/STROKEAHA.114.003130 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Lokken RP, Wellenius GA, Coull BA, Burger MR, Schlaug G, Suh HH, Mittleman MA, 2009. Air pollution and risk of stroke: underestimation of effect due to misclassification of time of event onset. Epidemiology 20, 137–142. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Maclure M, 1991. The case-crossover design: a method for studying transient effects on the risk of acute events. American journal of epidemiology 133, 144–53. [DOI] [PubMed] [Google Scholar]
  32. Maheswaran R, 2016. Air pollution and stroke – an overview of the evidence base. Spatial and Spatio-temporal Epidemiology, Environmental Exposure and Health 18, 74–81. 10.1016/j.sste.2016.04.004 [DOI] [PubMed] [Google Scholar]
  33. Maheswaran R, Pearson T, Beevers SD, Campbell MJ, Wolfe CD, 2016. Air Pollution and Subtypes, Severity and Vulnerability to Ischemic Stroke—A Population Based Case-Crossover Study. PLOS ONE 11, e0158556 10.1371/journal.pone.0158556 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Mechtouff L, Canoui-Poitrine F, Schott A-M, Nighoghossian N, Trouillas P, Termoz A, Porthault-Chatard S, David J-S, Chasles V, Derex L, 2012. Lack of association between air pollutant exposure and short-term risk of ischaemic stroke in Lyon, France. Int J Stroke 7, 669–674. 10.1111/j.1747-4949.2011.00737.x [DOI] [PubMed] [Google Scholar]
  35. Mills NL, Törnqvist H, Robinson SD, Gonzalez MC, Söderberg S, Sandström T, Blomberg A, Newby DE, Donaldson K, 2007. Air pollution and atherothrombosis. Inhal Toxicol 19 Suppl 1, 81–89. 10.1080/08958370701495170 [DOI] [PubMed] [Google Scholar]
  36. NASA, 2017. MERRA Products — GES DISC - Goddard Earth Sciences Data and Information Services Center [WWW Document]. URL https://disc.gsfc.nasa.gov/mdisc/data-holdings/merra/merra_products_nonjs.shtml (accessed 2.10.17). [DOI] [PMC free article] [PubMed]
  37. O’Donnell MJ, Fang J, Mittleman MA, Kapral MK, Wellenius GA, Investigators of the Registry of Canadian Stroke Network, 2011. Fine particulate air pollution (PM2.5) and the risk of acute ischemic stroke. Epidemiology 22, 422–431. 10.1097/EDE.0b013e3182126580 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Oudin A, Strömberg U, Jakobsson K, Stroh E, Björk J, 2010. Estimation of short-term effects of air pollution on stroke hospital admissions in southern Sweden. Neuroepidemiology 34, 131–142. 10.1159/000274807 [DOI] [PubMed] [Google Scholar]
  39. Reker DM, Rosen AK, Hoenig H, Berlowitz DR, Laughlin J, Anderson L, Marshall CR, Rittman M, 2002. The hazards of stroke case selection using administrative data. Med Care 40, 96–104. [DOI] [PubMed] [Google Scholar]
  40. Rich DQ, Kipen HM, Huang W, et al. , 2012. ASsociation between changes in air pollution levels during the beijing olympics and biomarkers of inflammation and thrombosis in healthy young adults. JAMA 307, 2068–2078. 10.1001/jama.2012.3488 [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Shah ASV, Lee KK, McAllister DA, Hunter A, Nair H, Whiteley W, Langrish JP, Newby DE, Mills NL, 2015. Short term exposure to air pollution and stroke: systematic review and meta-analysis. BMJ 350, h1295. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Shih RA, Griffin BA, Salkowski N, Jewell A, Eibner C, Bird CE, Liao D, Cushman M, Margolis HG, Eaton CB, Whitsel EA, 2011. Ambient particulate matter air pollution and venous thromboembolism in the Women’s Health Initiative Hormone Therapy trials. Environ. Health Perspect. 119, 326–331. 10.1289/ehp.1002256 [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Talbott EO, Rager JR, Benson S, Ann Brink L, Bilonick RA, Wu C, 2014. A case-crossover analysis of the impact of PM2.5 on cardiovascular disease hospitalizations for selected CDC tracking states. Environmental Research, Linking Exposure and Health in Environmental Public Health Tracking 134, 455–465. 10.1016/j.envres.2014.06.018 [DOI] [PubMed] [Google Scholar]
  44. Tsai S-S, Goggins WB, Chiu H-F, Yang C-Y, 2003. Evidence for an association between air pollution and daily stroke admissions in Kaohsiung, Taiwan. Stroke 34, 2612–2616. 10.1161/01.STR.0000095564.33543.64 [DOI] [PubMed] [Google Scholar]
  45. Villeneuve PJ, Chen L, Stieb D, Rowe BH, 2006. Associations between outdoor air pollution and emergency department visits for stroke in Edmonton, Canada. Eur. J. Epidemiol 21, 689–700. 10.1007/s10654-006-9050-9 [DOI] [PubMed] [Google Scholar]
  46. Villeneuve PJ, Johnson JYM, Pasichnyk D, Lowes J, Kirkland S, Rowe BH, 2012. Short-term effects of ambient air pollution on stroke: who is most vulnerable? Sci. Total Environ. 430, 193–201. 10.1016/j.scitotenv.2012.05.002 [DOI] [PubMed] [Google Scholar]
  47. Vivanco-Hidalgo RM, Wellenius GA, Basagaña X, Cirach M, González AG, Ceballos P. de, Zabalza A, Jiménez-Conde J, Soriano-Tarraga C, Giralt-Steinhauer E, Alastuey A, Querol X, Sunyer J, Roquer J, 2018. Short-term exposure to traffic-related air pollution and ischemic stroke onset in Barcelona, Spain. Environ. Res 162, 160–165. 10.1016/j.envres.2017.12.024 [DOI] [PubMed] [Google Scholar]
  48. Wang Y, Eliot MN, Wellenius GA, 2014. Short-term Changes in Ambient Particulate Matter and Risk of Stroke: A Systematic Review and Meta-analysis. Journal of the American Heart Association 3, e000983–e000983. 10.1161/JAHA.114.000983 [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Wellenius Burger, MR Coull, BA, et al. , 2012. Ambient air pollution and the risk of acute ischemic stroke. Arch Intern Med 172, 229–234. 10.1001/archinternmed.2011.732 [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Wellenius GA, Schwartz J, Mittleman MA, 2005. Air pollution and hospital admissions for ischemic and hemorrhagic stroke among medicare beneficiaries. Stroke 36, 2549–2553. 10.1161/01.STR.0000189687.78760.47 [DOI] [PubMed] [Google Scholar]
  51. Whitsel EA, Quibrera PM, Christ SL, Liao D, Prineas RJ, Anderson GL, Heiss G, 2009. Heart rate variability, ambient particulate matter air pollution, and glucose homeostasis: the environmental epidemiology of arrhythmogenesis in the women’s health initiative. Am. J. Epidemiol 169, 693–703. 10.1093/aje/kwn400 [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Wing JJ, Adar SD, Sánchez BN, Morgenstern LB, Smith MA, Lisabeth LD, 2015. Ethnic differences in ambient air pollution and risk of acute ischemic stroke. Environ. Res 143, 62–67. 10.1016/j.envres.2015.09.031 [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Wing JJ, Sánchez BN, Adar SD, Meurer WJ, Morgenstern LB, Smith MA, Lisabeth LD, 2017. Synergism of Short-Term Air Pollution Exposures and Neighborhood Disadvantage on Initial Stroke Severity. Stroke 48, 3126–3129. 10.1161/STROKEAHA.117.018816 [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Wordley J, Walters S, Ayres JG, 1997. Short term variations in hospital admissions and mortality and particulate air pollution. Occup Environ Med 54, 108–116. [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Wu S, Deng F, Wei H, Huang J, Wang H, Shima M, Wang X, Qin Y, Zheng C, Hao Y, Guo X, 2012. Chemical constituents of ambient particulate air pollution and biomarkers of inflammation, coagulation and homocysteine in healthy adults: A prospective panel study. Part Fibre Toxicol 9, 49 10.1186/1743-8977-9-49 [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Zanobetti A, Schwartz J, 2009. The effect of fine and coarse particulate air pollution on mortality: a national analysis. Environ. Health Perspect. 117, 898–903. 10.1289/ehp.0800108 [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Zhang Z-M, Whitsel EA, Quibrera PM, Smith RL, Liao D, Anderson GL, Prineas RJ, 2009. Ambient fine particulate matter exposure and myocardial ischemia in the Environmental Epidemiology of Arrhythmogenesis in the Women’s Health Initiative (EEAWHI) study. Environ. Health Perspect. 117, 751–756. 10.1289/ehp.0800046 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplemental 1
NIHMS1044649-supplement-Supplemental_1.docx (36.5KB, docx)

RESOURCES