Air Pollution and Hospital Admissions for Ischemic and Hemorrhagic Stroke Among Medicare Beneficiaries

    the Cardiovascular Epidemiology Research Unit (G.A.W., M.A.M.), Cardiovascular Division, Department of Medicine, Beth Israel Deaconess Medical Center, and Departments of Environmental Health (G.A.W., J.S., M.A.M.) and Epidemiology (J.S., M.A.M.), Harvard School of Public Health, Boston, Mass.

    Abstract

    Background and Purpose— The association between short-term elevations in ambient air particles and increased cardiovascular morbidity and mortality is well documented. Ambient particles may similarly increase the risk of stroke.

    Methods— We evaluated the association between daily levels of respirable particulate matter (aerodynamic diameter 10 μm, PM10) and hospital admission for ischemic and hemorrhagic stroke among Medicare recipients (age 65 years) in 9 US cities using a 2-stage hierarchical model. In the first stage, we applied the time-stratified case-crossover design to estimate the effect of PM10 in each city. We used a 3-day unconstrained, distributed lag model to simultaneously estimate the effect of PM10 0 to 2 days before the admission day and controlled for meteorological covariates in all of the models. In the second stage, we used random-effects metaanalytic techniques to combine the city-specific effect estimates.

    Results— Ischemic (n=155 503) and hemorrhagic (19 314) stroke admissions were examined separately. For ischemic stroke, an interquartile range increase in PM10 was associated with a 1.03% (95% CI, 0.04% to 2.04%) increase in admissions on the same day only. Similar results were observed with CO, NO2, and SO2. For hemorrhagic stroke, no association was observed with any pollutant 0 to 2 days before admission.

    Conclusions— These results suggest that elevations in ambient particles may transiently increase the risk of ischemic, but not hemorrhagic, stroke. Studies with more accurate assessment of timing of stroke onset are necessary to confirm or refute these findings.

    Key Words: ischemic stroke  hemorrhagic stroke  air pollution  environmental exposure  epidemiology

    Introduction

    Epidemiological studies have demonstrated a consistent increased risk for cardiac morbidity and mortality associated with acute exposure to ambient air particles, including triggering of acute myocardial infarction,1,2 discharge of implanted cardioverter defibrillators,3 and hospital admission for acute decompensation of patients with congestive heart failure.4 Although the mechanisms have not been fully elucidated, current evidence suggests that alterations in hemodynamics,5,6 hemostatic factors,7–9 and autonomic function10,11 may underlie the cardiovascular effects of particulate air pollution.

    Exposure to particulate air pollution may similarly increase the risk of stroke. Studies of small-area variation have found a positive association between stroke mortality rates and living in areas of high-ambient pollution.12 Time series studies using hospital discharge summaries report a statistically significant positive association between daily measures of respirable particles (particulate matter with aerodynamic diameter <10 μm, PM10) and cerebrovascular hospitalizations,13–16 but the results have been inconsistent.17–19 Moreover, few studies have distinguished between ischemic and hemorrhagic strokes.15,16,20 Given the putative mechanisms of particulate-related cardiovascular effects, we hypothesized that short-term elevations in PM10 would be associated with increased hospitalizations for ischemic but not hemorrhagic stroke. We evaluated this hypothesis among Medicare beneficiaries 65 years of age in 9 US cities with daily PM10 monitoring.

    Methods

    Study Population

    This study was approved by the Institutional Review Boards of the Harvard School of Public Health and Beth Israel Deaconess Medical Center. We evaluated the association between mean daily PM10 levels and the rate of admission for stroke among Medicare beneficiaries in 9 cities chosen a priori based on their large populations, many years of daily PM10 monitoring, and geographic diversity (Table 1). The observation period in each city was limited by the range of dates for which both hospital admissions and daily PM10 data were available.

    We obtained information on hospital admissions from the Centers for Medicare and Medicaid Services. Ischemic stroke cases were defined as persons admitted from the emergency room with a primary discharge diagnosis of acute but ill-defined cerebrovascular disease or occlusion of cerebral arteries, excluding those without cerebral infarction.21 In a sensitivity analysis, we limited the definition of ischemic stroke to patients admitted from the emergency room with a primary discharge diagnosis of occlusion of cerebral arteries.21 Hemorrhagic stroke cases were defined as persons admitted from the emergency room with a primary discharge diagnosis of intracerebral hemorrhage.21

    Exposure Information

    We obtained daily measures of PM10 and hourly measures of CO, NO2, and SO2 from the US Environmental Protection Agency and computed daily mean concentrations in each city.22 To evaluate the effect of PM10 within the range of exposure deemed acceptable in the United States under current regulations, we excluded from analysis 31 days (of 37 203 possible) on which PM10 levels exceeded 150 μg/m3, the current 24-hour US Environmental Protection Agency standard. We considered pollutant data unavailable for a given city if >10% of days in the observation period had missing values. Using this criterion, data on CO was missing in New Haven, Conn; NO2 in Birmingham, Ala, Salt Lake City, Utah, and Seattle, Wash; and SO2 in Birmingham, Ala. We obtained National Weather Service data on climatologic variables (National Climatic Data Center) and calculated the apparent temperature (an index of human discomfort that incorporates ambient temperature and relative humidity) as described previously.23

    Data on PM10 were missing on 4.0% of the days. In a sensitivity analysis, we used city-specific prediction models to impute single missing values. Specifically, we modeled PM10 as smooth functions of PM10 on the previous and following day, available copollutants, and climatologic variables. City-specific model R2 values ranged from 0.50 to 0.85 (mean, 0.71). We imputed missing values of PM10 for days preceded and followed by nonmissing days. This time-series, referred to as "PM10 with imputed values," had missing values on 1.7% of days.

    Data Analysis

    We used a 2-stage hierarchical model to evaluate the association between ambient pollution and the rate of hospital admission, as described previously.24 In the first stage, we used the time-stratified case-crossover study design25 to separately estimate the effect of air pollution in each city. In this design, each subject’s exposure before a case-defining event (case period) is compared with his or her own exposure experience during 1 control period when the subject did not become a case. Control periods were chosen such that exposures during the case period were compared with exposures occurring on other days of the same month falling on the same day of the week as the case period. Exposure during the case period was defined as pollutant concentrations 0 to 2 days before admission (lags 0 to 2 days). This design has been shown to be effective in controlling for seasonality, time trends, and chronic and slowly varying potential confounders.24

    We performed conditional logistic regression, stratifying on each day, to obtain estimates of odds ratios and 95% CIs. We report effect estimates as the percentage of change in the rate of hospitalization associated with an interquartile range increase in mean daily pollutant levels. For each pollutant, we simultaneously evaluated the effect at lags of 0 to 2 days using unconstrained distributed lag models26 and controlled for meteorological covariates. We modeled the mean apparent temperature at lag 0 as a quadratic function and mean apparent temperature at lags 1 and 2, mean barometric pressure at lags 0 and 1, and pollutants at lags 0 to 2 as linear functions of continuous variables.

    In the second stage, we obtained a combined random-effects estimate from the city-specific effect estimates using standard random-effects meta-analysis methods.27 All of the reported P values are based on 2-sided tests at the =0.05 level. Analyses were performed using SAS V9 and the R statistical package.

    Results

    Ischemic Stroke Admissions

    There were 155 503 hospital admissions from the emergency department with a primary discharge diagnosis of ischemic stroke among Medicare beneficiaries 65 years of age in 9 US cities (Table 1). The overall mean PM10 level was 32.69 μg/m3 (SD±19.75; Table 2). Age on admission ranged from 65 to 117 years with a mean of 78.6±7.7 years. Cases were mostly white (75.4%) and female (60.7%).

    We simultaneously estimated the effect of PM10 at lags of 0 to 2 days in each city and then obtained summary estimates using standard random-effects meta-analysis techniques. A positive association between PM10 and hospital admissions on the same day was observed in 7 of the 9 cities examined (Pheterogeneity=0.71; Figure, A). Overall, an interquartile range increase in PM10 on the day of admission was associated with a 1.03% (95% CI, 0.04% to 2.04%) increase in the rate of hospital admission for ischemic stroke. PM10 levels 1 or 2 days before admission were not associated with increased risk. Using the alternate definition of ischemic stroke yielded a 1.39% (95% CI, 0.26% to 2.52%) increase for an interquartile range increase in PM10 on the admission day and no significant increase associated with PM10 levels 1 or 2 days before admission. Using PM10 with imputed values as the exposure did not materially alter the results.

    City-specific (, height inversely proportional to variance of estimate) and random-effects summary (, horizontal limits indicate 95% CIs) estimates of the percentage change in rate of hospital admission for ischemic (A) and hemorrhagic (B) stroke associated with an interquartile range increase in PM10 (22.96 μg/m3) on the same day (lag 0).

    Ambient measures of CO, NO2, and SO2 were correlated with ambient PM10 levels (Table 2). An interquartile range increase in each pollutant was associated with a qualitatively similar increase in admissions on the same day (Table 3). No association was observed with any pollutant 1 or 2 days before the admission date.

    Hemorrhagic Stroke Admissions

    There were 19 314 hospital admissions with a primary discharge diagnosis of hemorrhagic stroke (Table 1). Compared with patients admitted for ischemic stroke, these patients tended to be younger (mean age, 77.9±7.6 years), more likely to be white (78.2%), and less likely to be female (58.8%).

    PM10 levels on the day of admission were not associated with hemorrhagic stroke admissions although there was evidence of statistical heterogeneity between cities (Pheterogeniety=0.013; Figure, B). PM10 levels 1 to 2 days before admission were also not associated with increased risk. Likewise, no association was observed between hemorrhagic stroke admissions and PM10 with imputed values, CO, NO2, or SO2 (Table 3).

    Discussion

    Previous studies using administrative data from 1 to 3 cities suggest an association between daily fluctuations in PM10 and cerebrovascular hospital admissions.13–16 We evaluated the association between PM10 and hospital admissions for ischemic and hemorrhagic stroke among Medicare beneficiaries 65 years of age in 9 US cities and found that a transient increase in ambient particles was associated with an increased risk of hospital admission for ischemic but not hemorrhagic stroke. Specifically, for ischemic stroke, we found that an interquartile range increase in PM10 was associated with a 1.03% (95% CI, 0.04% to 2.04%) increase in the risk of admission on the same day. In contrast, for hemorrhagic stroke, we found no evidence of an association with PM10 or any other pollutant examined.

    Few published studies have examined the effects of ambient particles specifically on ischemic or hemorrhagic stroke.15,16,20 Using hospital admission records from Taiwan, Tsai et al16 found a statistically significant positive association between PM10 and both ischemic and hemorrhagic strokes after excluding days with mean daily temperature <20°C. Similarly, Hong et al20 analyzed death certificates in Seoul, Korea, with similar statistically significant results for ischemic and hemorrhagic stroke. Linn et al15 also found a significant increase in hospitalization rates for ischemic stroke in metropolitan Los Angeles, but did not consider hemorrhagic strokes. The discrepant results in relation to hemorrhagic stroke may be explained by differences in case definitions, analytic methods, average pollutant levels, or population characteristics between the current and previous studies.

    Exposure to ambient particles may plausibly increase the risk of ischemic stroke by promoting atherosclerotic plaque disruption and thrombosis. This could be mediated by a combination of 3 different mechanisms. First, particulate exposure can induce an acute systemic inflammatory response with an increased number of circulating neutrophils28 and increased levels of C-reactive protein.9 Second, particulate-related changes in hemostatic factors have been reported, including increased levels of fibrinogen7,8 and von Willebrand factor9 and enhanced peripheral arterial thrombosis.29 Third, particulate-related changes in autonomic nervous system activity, as assessed by heart rate variability, are well documented and are consistent with relative sympathoexcitation.10,11

    The observed association between ischemic stroke and air pollution was similar for all of the pollutants considered. Likewise, we consistently failed to find evidence of an association between hemorrhagic stroke and any pollutant considered. Because ambient measures are correlated, separating the effects of individual pollutants is problematic. However, PM10 is relatively spatially homogenous within a metropolitan area,30 and ambient measures at a central site are well correlated with personal exposure measures.31 In contrast, personal exposures to CO, NO2, and SO2 are relatively poorly predicted by ambient measures.32 For example, ambient measures of NO2 are likely more strongly correlated with personal exposure to particulate matter than to personal NO2 exposure.32 Thus, the present study is consistent with the interpretation that sources of pollution that lead to concurrent elevation of these pollutants may be responsible for the observed associations.

    The health effects of exposure to air pollution from indoor sources (eg, from tobacco smoke or cooking) are also of interest, but evaluating these effects requires a different study design and indoor, as well as ambient, pollutant monitors. Although ambient particles have been shown to penetrate indoors, there is little correlation between exposure to particles originating from indoor sources and those from outdoor sources.33

    This study has several potential limitations. First, misclassification of the outcome is expected as a result of diagnostic or coding errors. However, these errors are likely unrelated to pollutant levels and are expected to reduce the precision of our estimates and potentially bias the relative risk toward the null. In fact, using the alternate definition of ischemic stroke, which is likely to be more specific, albeit less sensitive, than the primary definition,34,35 we found a stronger association with PM10. Similarly, we expect that the incidence of clinically unrecognized or nonhospitalized stroke is unrelated to pollutant levels after adjusting for meteorological covariates and month-of-year and day-of-week effects. Therefore, exclusion of these cases is expected to reduce the precision of our estimates but not otherwise bias our results. Second, the use of ambient rather than personal exposure measures is expected to result in exposure misclassification. However, this misclassification is expected to lead one to underestimate the relative risk.36 Third, the date of symptom onset likely preceded the date of admission in a proportion of cases, leading to nondifferential exposure misclassification and bias of the relative risk toward the null. Fourth, residual confounding by short-term respiratory epidemics remains a possibility. However, the time-stratified case-crossover design used in this study has been shown to effectively control for confounding by time trends in both exposure and outcomes.24 In particular, this approach provides effective control for time trends that occur over periods longer than 1 calendar month. Additionally, studies directly examining this possibility suggest that respiratory epidemics are not important confounders of the association between particulate matter and all-cause or cardiovascular mortality.37

    Conclusions

    In the current study, we found that a transient increase in ambient particles was associated with an increased risk of hospital admission for ischemic but not hemorrhagic stroke. Although the relative increase in risk was small, given the large number of people simultaneously at risk for ischemic stroke and exposed to urban pollution, even a small relative risk may be of significant public health interest. Additional studies with more detailed data on the clinical characteristics of subjects and more accurate assessment of the timing of symptom onset are needed to confirm or refute these findings.

    Acknowledgments

    This study was supported by grants ES013804, HL07374, and ES09825 from the National Institutes of Health and by US Environmental Protection Agency research award R827353 from the US Environmental Protection Agency.

    References

    Peters A, Dockery DW, Muller JE, Mittleman MA. Increased particulate air pollution and the triggering of myocardial infarction. Circulation. 2001; 103: 2810–2815.

    D’Ippoliti D, Forastiere F, Ancona C, Agabiti N, Fusco D, Michelozzi P, Perucci CA. Air pollution and myocardial infarction in rome: a case-crossover analysis. Epidemiology. 2003; 14: 528–535.

    Rich DQ, Schwartz J, Mittleman MA, Link M, Luttmann-Gibson H, Catalano PJ, Speizer FE, Dockery DW. Association of short-term ambient air pollution concentrations and ventricular arrhythmias. Am J Epidemiol. 2005; 161: 1123–1132.

    Wellenius GA, Bateson TF, Mittleman MA, Schwartz J. Particulate air pollution and the rate of hospitalization for congestive heart failure among medicare beneficiaries in pittsburgh, pennsylvania. Am J Epidemiol. 2005; 161: 130–136.

    Pope CA 3rd, Dockery DW, Kanner RE, Villegas GM, Schwartz J. Oxygen saturation, pulse rate, and particulate air pollution: A daily time-series panel study. Am J Respir Crit Care Med. 1999; 159: 365–372.

    Zanobetti A, Canner MJ, Stone PH, Schwartz J, Sher D, Eagan-Bengston E, Gates KA, Hartley LH, Suh H, Gold DR. Ambient pollution and blood pressure in cardiac rehabilitation patients. Circulation. 2004; 110: 2184–2189.

    Ghio AJ, Kim C, Devlin RB. Concentrated ambient air particles induce mild pulmonary inflammation in healthy human volunteers. Am J Respir Crit Care Med. 2000; 162: 981–988.

    Schwartz J. Air pollution and blood markers of cardiovascular risk. Environ Health Perspect. 2001; 109 (Suppl III): 405–409.

    Riediker M, Cascio WE, Griggs TR, Herbst MC, Bromberg PA, Neas L, Williams RW, Devlin RB. Particulate matter exposure in cars is associated with cardiovascular effects in healthy young men. Am J Respir Crit Care Med. 2004; 169: 934–940.

    Gold DR, Litonjua A, Schwartz J, Lovett E, Larson A, Nearing B, Allen G, Verrier M, Cherry R, Verrier R. Ambient pollution and heart rate variability. Circulation. 2000; 101: 1267–1273.

    Creason J, Neas L, Walsh D, Williams R, Sheldon L, Liao D, Shy C. Particulate matter and heart rate variability among elderly retirees: The Baltimore 1998 pm study. J Expo Anal Environ Epidemiol. 2001; 11: 116–122.

    Maheswaran R, Haining RP, Brindley P, Law J, Pearson T, Fryers PR, Wise S, Campbell MJ. Outdoor air pollution and stroke in sheffield, united kingdom: a small-area level geographical study. Stroke. 2005; 36: 239–243.

    Wordley J, Walters S, Ayres JG. Short term variations in hospital admissions and mortality and particulate air pollution. Occup Environ Med. 1997; 54: 108–116.

    Moolgavkar SH. Air pollution and hospital admissions for diseases of the circulatory system in three U.S. metropolitan areas. J Air & Waste Manag Assoc. 2000; 50: 1199–1206.

    Linn WS, Szlachcic Y, Gong H Jr, Kinney PL, Berhane KT. Air pollution and daily hospital admissions in metropolitan los angeles. Environ Health Perspect. 2000; 108: 427–434.

    Tsai SS, Goggins WB, Chiu HF, Yang CY. Evidence for an association between air pollution and daily stroke admissions in Kaohsiung, Taiwan. Stroke. 2003; 34: 2612–2616.

    Burnett RT, Smith-Doiron M, Stieb D, Cakmak S, Brook JR. Effects of particulate and gaseous air pollution on cardiorespiratory hospitalizations. Arch Environ Health. 1999; 54: 130–139.

    Le Tertre A, Medina S, Samoli E, Forsberg B, Michelozzi P, Boumghar A, Vonk JM, Bellini A, Atkinson R, Ayres JG, Sunyer J, Schwartz J, Katsouyanni K. Short-term effects of particulate air pollution on cardiovascular diseases in eight European cities. J Epidemiol Community Health. 2002; 56: 773–779.

    Sunyer J, Ballester F, Tertre AL, Atkinson R, Ayres JG, Forastiere F, Forsberg B, Vonk JM, Bisanti L, Tenias JM, Medina S, Schwartz J, Katsouyanni K. The association of daily sulfur dioxide air pollution levels with hospital admissions for cardiovascular diseases in europe (the aphea-ii study). Eur Heart J. 2003; 24: 752–760.

    Hong YC, Lee JT, Kim H, Kwon HJ. Air pollution: a new risk factor in ischemic stroke mortality. Stroke. 2002; 33: 2165–2169.

    Department of Health and Human Services. International Classification of Diseases, 9th Revision, Clinical Modification: ICD-9-CM. 3rd ed. Washington, DC: US Government Printing Office; 1989.

    Zanobetti A, Schwartz J, Dockery DW. Airborne particles are a risk factor for hospital admissions for heart and lung disease. Environ Health Perspect. 2000; 108: 1071–1077.

    O’Neill MS, Zanobetti A, Schwartz J. Modifiers of the temperature and mortality association in seven US cities. Am J Epidemiol. 2003; 157: 1074–1082.

    Schwartz J, Zanobetti A, Bateson TF. Morbidity and mortality among elderly residents in cities with daily pm measurements. In: Revised Analyses of Time-Series Studies of Air Pollution and Health. Special Report. Boston, Mass: Health Effects Institute; 2003: 25–58.

    Levy D, Lumley T, Sheppard L, Kaufman J, Checkoway H. Referent selection in case-crossover analyses of acute health effects of air pollution. Epidemiology. 2001; 12: 186–192.

    Samet JM, Zeger SL, Dominici F, Curriero F, Coursac I, Dockery DW, Schwartz J, Zanobetti A. The national morbidity, mortality, and air pollution study. Part II: Morbidity and mortality from air pollution in the United States. Res Rep Health Eff Inst. 2000; 94: 5–79.

    DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials. 1986; 7: 177–188.

    van Eeden SF, Hogg JC. Systemic inflammatory response induced by particulate matter air pollution: The importance of bone-marrow stimulation. J Toxicol Environ Health A. 2002; 65: 1597–1613.

    Nemmar A, Hoet PH, Dinsdale D, Vermylen J, Hoylaerts MF, Nemery B. Diesel exhaust particles in lung acutely enhance experimental peripheral thrombosis. Circulation. 2003; 107: 1202–1208.

    Ito K, De Leon S, Thurston GD, Nadas A, Lippmann M. Monitor-to-monitor temporal correlation of air pollution in the contiguous us. J Expo Anal Environ Epidemiol. 2005; 15: 172–184.

    Janssen NA, Hoek G, Brunekreef B, Harssema H, Mensink I, Zuidhof A. Personal sampling of particles in adults: Relation among personal, indoor, and outdoor air concentrations. Am J Epidemiol. 1998; 147: 537–547.

    Sarnat JA, Brown KW, Schwartz J, Coull BA, Koutrakis P. Ambient gas concentrations and personal particulate matter exposures: Implications for studying the health effects of particles. Epidemiology. 2005; 16: 385–395.

    Ebelt ST, Wilson WE, Brauer M. Exposure to ambient and nonambient components of particulate matter: A comparison of health effects. Epidemiology. 2005; 16: 396–405.

    Goldstein LB. Accuracy of icd-9-cm coding for the identification of patients with acute ischemic stroke: Effect of modifier codes. Stroke. 1998; 29: 1602–1604.

    Broderick J, Brott T, Kothari R, Miller R, Khoury J, Pancioli A, Gebel J, Mills D, Minneci L, Shukla R. The greater Cincinnati/northern Kentucky stroke study: Preliminary first-ever and total incidence rates of stroke among blacks. Stroke. 1998; 29: 415–421.

    Zeger SL, Thomas D, Dominici F, Samet JM, Schwartz J, Dockery D, Cohen A. Exposure measurement error in time-series studies of air pollution: Concepts and consequences. Environ Health Perspect. 2000; 108: 419–426.

    Touloumi G, Samoli E, Quenel P, Paldy A, Anderson RH, Zmirou D, Galan I, Forsberg B, Schindler C, Schwartz J, Katsouyanni K. Short-term effects of air pollution on total and cardiovascular mortality: The confounding effect of influenza epidemics. Epidemiology. 2005; 16: 49–57.