Abraham, Joseph H. ScD, MS; Baird, Coleen P. MD, MPH
Acute particulate matter (PM) health effects among US military personnel deployed in support of Operation Iraqi Freedom (OIF) and Operation Enduring Freedom (OEF, Afghanistan) have not been well characterized epidemiologically. At the outset of OIF/OEF, PM was identified by the Department of Defense to be the most ubiquitous environmental exposure in the Central Command Area of Operations.1 The World Health Organization,2 the US government,3 and the US Army4 have each identified ambient air pollution as a public health priority. Concern regarding health effects associated with deployment-related environmental exposures has been increasing among soldiers and veterans.5,6 In a 2011 report on the long-term health consequences of exposure to burn pit emissions during deployment to OIF/OEF, the Institute of Medicine7 found insufficient evidence to develop firm conclusions regarding burn pit health effects, but concluded that service in Iraq or Afghanistan might be associated with long-term cardiovascular and respiratory health effects, particularly in highly exposed and/or susceptible populations, mainly because of high ambient concentrations of PM.
A large number of epidemiological studies8–10 of acute PM air pollution effects conducted in numerous countries, using different analytic approaches, have demonstrated that daily changes in 24-hour average levels of ambient PM are associated with subsequent increases in respiratory and cardiovascular disease morbidity and mortality. Nevertheless, acute PM health effects are observed primarily among children, the elderly, and individuals with preexisting chronic health conditions such as asthma, diabetes, and cardiopulmonary disease; not among healthy individuals.11–13 The deployed force, in contrast, is relatively young, with an average age of Army service members being less than 30 years, and 92% being younger than 40 years. The force is also relatively healthy, as military medical standards for enlistment, appointment, or induction exclude many, but not all, potentially susceptible individuals from service.14 That said, young and otherwise healthy individuals are at risk of lung injury from environmental exposures, as evidenced by persistent decrements in lung function among rescue workers who responded to the World Trade Center attacks in 2001.15 Acute PM health effects are also dependent, at least in part, on the composition of the particles.13,16 Although PM levels are comparatively high, the composition of ambient PM in deployed locations17 differs from regions in the United States impacted most strongly by PM.18 In this context, it is unclear whether military personnel, who are generally younger and healthier relative to the general population, are susceptible to health effects of PM exposure during deployment.
A growing number of evaluations of the potential association between deployment and respiratory health are appearing in the published literature. Among participants in the Millennium Cohort Study, a large prospective study of service members, soldiers who had deployed to Iraq and Afghanistan were more likely to report experiencing respiratory symptoms compared with participating soldiers who have never had a deployment to southwest Asia.19 Increasing length of deployment was also positively associated with report of respiratory symptoms among army personnel. Nevertheless, diagnoses of chronic respiratory conditions such as asthma, chronic bronchitis, and emphysema were equally distributed among those deployed to OIF/OEF and never-deployed individuals. Evidence of increased respiratory disease among formerly deployed soldiers has been observed in other populations. In a retrospective review of medical records at the Northport Veterans Affairs Medical Center in Northport, New York (Northport VA),6 the odds of new-onset asthma were higher among those who had deployed in support of OIF/OEF than among former soldiers with no history of overseas deployment. In a subsequent study20 of veterans at the Northport VA by the same group, referral for pulmonary function testing was higher among OIF/OEF veterans than among veterans with no history of OIF/OEF deployment. Among 80 soldiers being evaluated for exertional dyspnea after deployment to Iraq and Afghanistan, 49 underwent biopsy and 38 were subsequently diagnosed with constrictive bronchiolitis, a rare finding characterized by small airways fibrosis and muscular wall thickening.5 Of the soldiers in this case series from Fort Campbell who did not receive a constrictive bronchiolitis diagnosis, two were diagnosed with respiratory bronchiolitis, two had sarcoidosis, two had respiratory bronchiolitis-associated interstitial lung disease, and two had hypersensitivity pneumonitis;11 of the soldiers who did not receive a biopsy were treated for asthma or bronchitis. None of these reports assess associations between a measured, specific, deployment-related environmental exposure and adverse respiratory health outcomes.
The objective of this study was to assess the relationship between ambient PM levels and the occurrence of cardiovascular and respiratory health events occurring during deployment. We evaluated the relationship between short-term exposure to PM air pollution and cardiovascular and respiratory events, utilizing in-theater electronic medical record data obtained during the deployment and measurements of ambient PM levels at multiple deployment sites in southwest Asia. We used a case-crossover study design21,22 to estimate the relative risk of cardiovascular and respiratory outcomes associated with PM less than 2.5 μm in aerodynamic diameter (PM2.5), and that less than 10 μm in aerodynamic diameter (PM10).
We used a time-stratified, bidirectional case-crossover study design to estimate short-term changes in risk of cardiovascular and respiratory health outcomes associated with short-term (0 to 1 day lag) exposure to PM air pollution in a population of military personnel deployed to southwest Asia. The case-crossover study design is analogous to an individually matched case–control study where “control” exposure information corresponding to each “case” record is based on the exposure experience of the case. By matching in this manner, potential confounding by risk factors that change only slowly over time (eg, age, smoking, diet, and underlying health status) is controlled by design. The assumptions implicit in the case-crossover study design are analogous to those of a case–control study. That is, the exposure distribution of referent times (“controls”) is independent of the exposure distribution of hazard times (“cases”), and the exposure distribution of the referent times is representative of the exposure in the source population.
Environmental sampling for this study was conducted between December 2005 and June 2007 at 15 bases in southwest Asia.17,23 The nonidentified source population included military personnel who deployed to one of these bases, and who, had they had a qualifying cardiovascular or respiratory health event, would have had an International Classification of Diseases, Ninth Revision (ICD-9),–identified medical record captured by at least one of the medical records databases utilized for this study.24 To identify health outcomes in this population, we obtained medical record abstracts from the Joint Medical Workstation and the Transportation Command Regulating and Command and Control Evacuation System medical record systems for the time period of the environmental sampling. In addition, we obtained declassified deployment dates and location data from the Deployed Theater Accountability System.
The study population consisted of a case series of military personnel who had a medical encounter for a qualifying cardiovascular or respiratory event recorded in either Joint Medical Workstation or Transportation Command Regulating and Command and Control Evacuation System (see earlier) during the time period of environmental sampling, and for whom deployment data at the time of the health event was known. Case status was defined as having any one of the qualifying cardiovascular (ICD-9 codes 390 to 459; “Diseases of the Circulatory System”) or respiratory (ICD-9 codes 460 to 519; “Diseases of the Respiratory System”) outcomes. In the case of multiple health care encounters for a single individual, outcomes occurring within 7 days of the initial record were excluded from analysis.
Air Pollution and Meteorological Measurements
The environmental sampling being used for this study has been previously described.17 Briefly, PM samples (ambient PM < 2.5 μm in aerodynamic diameter [PM2.5] and < 10 μm in aerodynamic diameter [PM10]) were collected during a period of approximately 1 year at 15 military bases selected to represent areas of potential exposure to military personnel. Sampling sites included six in Iraq (Balad, Baghdad, Tallil, Tikrit, Taji, and Al Asad), four in Kuwait (Northern, Central, Coastal, and Southern regions), two in Afghanistan (Bagram and Khowst), and one each in Qatar and the United Arab Emirates. Low-volume particulate samplers were deployed at each of the 15 sites, operating on a 1-in-6 day sampling schedule.
We obtained hourly meteorological data including temperature, relative humidity, barometric pressure, and wind speed measured and maintained by the Air Force Combat Climatology Center for the selected sampling sites, with the exception of those located in Tikrit and Southern Kuwait, for which meteorological data were unavailable.
Standard stratified data analysis methods were used to analyze the case-crossover data.21 The stratifying (matching) variable was the individual subject experiencing an outcome event. Each risk set, then, consists of one individual as that individual “crosses over” between exposure levels in the referent and hazard time periods (and back again). We matched the hazard period corresponding to each observed outcome event with referent time periods within a 28-day time window centered on the event day. To minimize the influence of autocorrelation between hazard and referent periods, referent PM exposures were selected only from sampling days greater than 7 days on either side of the hazard day.
In the first stage of the analysis, conditional logistic regression analyses were conducted using PROC PHREG and PROC LOGISTIC under the SAS System for Windows, version 9.1.3 (copyright 2002 to 2008, SAS Institute, Inc, Cary, NC). Separate models were constructed for each PM size metric. For each, PM level was entered into the statistical model as an untransformed continuous variable. To consider covariates for multivariate models, the univariate relationship between outcome events and each meteorological parameter was evaluated. Hourly temperature, relative humidity, barometric pressure, and wind speed measurements were used to create daily average values, which were subsequently entered into the statistical models as continuous variables.
In a second analytic stage, the individual sampling site-specific effect estimates were pooled using meta-analytic techniques, with the PM sampling site variable treated as a random effect, and implemented using SAS PROC MIXED. Odds ratios are expressed for a 10-μg/m3 increase in PM concentration.
The 15 sites for which 1-in-6 day sampling of ambient PM was sampled are identified in the map on Fig. 1. Summary demographic information for the study population is presented in Table 1. The median age of the study population was 32 years (min, 18 years and max, 76 years) and over three quarters were men. Twenty-seven percent of the study population was identified as white, 11% black, and 4% “other” race/ethnicity, and data about race/ethnicity were missing for 58% of the study group. Summary distributions of the 24-hour average environmental sampling and meteorological data are presented in Table 2. The 24-hour average PM2.5 observations ranged from a minimum of 4 μg/m3 (Bagram) to a maximum of 922 μg/m3 (Tikrit) with a median level of 52 μg/m3. Likewise, the PM10 observations ranged from 9 μg/m3 (Central Kuwait) to 2570 μg/m3 (Northern Kuwait), with a median of 130 μg/m3. Spearman correlation coefficients (P values) relating the various PM size metrics and meteorological variables are given in Table 3. The sampling site-specific PM metrics correlated moderately to highly, with a Spearman correlation coefficient of 0.76 for PM2.5 and PM10. Correlations between PM levels and average temperature were generally lower, ranging from 0.19 to 0.34. Correlations between PM levels and relative humidity and barometric pressure were low and inverse, and correlations between PM measurements and wind speed were very low.
Summary statistics of qualifying medical records are presented in Table 4. The majority of health encounters (83.5%) were for acute respiratory infections. Of the 343 encounters for chronic obstructive pulmonary disease and allied conditions (ICD-9 490 to 96), 327 (95%) were for asthma (ICD-9 493).
Figure 2 provides a time series of ambient PM levels and health outcome frequency for a base near Baghdad, Iraq, the largest site assessed with respect to population size. Health event data were more sparse for the other sites assessed in this evaluation.
The results from univariate conditional logistic regression models for the two PM size categories at lag 0 and the frequency of a qualifying medical encounter, interpretable as the percent change in the odds of an outcome for a 10-μg/m3 increase in PM, are given in Table 5. The pooled results (odds ratio, 95% confidence interval) from univariate and adjusted models are summarized in Fig. 3. Odds ratios are expressed as the relative increase in risk of cardiovascular or respiratory outcome for a 10-μg/m3 increase in PM. No consistent pattern in the associations between PM and the cardiovascular and respiratory outcome metric was observed. Pooled estimates were not statistically significant, suggesting that the role of chance cannot be excluded as a determinant of the observed associations. Models incorporating 2-day lagged exposures were assessed, as were models run with PM data imputed for missing days using observed PM levels and meteorological data. These sensitivity analyses yielded qualitatively similar results to the primary analyses' results.
Although ambient levels were routinely high at the military bases assessed, elevated levels of PM (PM2.5 and PM10) were not associated with acute cardiovascular and respiratory outcomes among deployed military personnel in this study. There are several explanations for the observed results that deserve discussion. As detailed in a report on the PM exposure assessment conducted by Engelbrecht et al,17,23 the ambient PM in the region consisted primarily of resuspended crustal material, which may be less biologically active than particles from anthropogenic sources.25 It is also possible that this relatively young, healthy population of deployed military personnel is not particularly vulnerable to severe acute cardiovascular and respiratory effects of PM exposure. Severe acute PM health effects are observed primarily among susceptible subpopulations, that is, children, the elderly, and those with underlying chronic health conditions.11 In this evaluation, health outcomes were defined using medical encounters captured by an electronic database. It is possible that PM exposure resulted in adverse health effects, which did not rise to a level of severity that would compel a soldier to seek medical care. The incidence of mild or subacute disease, which would likely not have been captured by this study, may also account for the finding of a positive association between deployment to OIF/OEF and self-reported respiratory symptoms, but not doctor-diagnosed disease in the Millennium Cohort Study population.19
Smoking is a strong risk factor for a wide range of cardiovascular and respiratory health outcomes, and deployment to Iraq and Afghanistan is likely to be positively associated with tobacco use among US military personnel.26,27 Nevertheless, the case-crossover study design inoculates this assessment from confounding by smoking and all other potential confounders that do not vary over short time periods.28 In addition, if smoking is positively correlated with PM exposure, it is unlikely that confounding due to smoking would result in a null association, as observed here, unless PM exposure was protective of cardiovascular and respiratory events. Residual confounding by time-varying factors is a potential source of bias in this study, although adjustment for meteorological conditions and day of week did not affect the observed null finding. We used a health event definition that included as cases individuals with a medical encounter coded as any one of several qualifying respiratory and cardiovascular conditions. More serious limitations of our study, which could result in a falsely negative finding, are: missing PM levels on most days due to the 1-in-6 day sampling scheme; incomplete health outcome capture using the in-theater electronic medical records database; and nondifferential misclassification of troop locations. In addition to these potential sources of bias, the statistical power of our assessment was limited by both the short (1-year) duration of the study and the small population under study, and the likely small magnitude of the effect of PM exposure on the risk of acute respiratory and cardiovascular events in this population. It is possible that PM exposure increases the risk of specific heath events (eg, myocardial infarction), and not others (eg, a medical encounter for influenza). Nevertheless, because of the small size of the study, particularly the small number of specific events, we were not able to assess outcomes individually, a circumstance that could also account for our null finding. Finally, the case-crossover study design is tuned to the evaluation of associations between short-term changes in exposure and acute health events; it cannot be used to evaluate the effects of long-term exposure to high levels of PM on health. As has been previously documented, annual average PM levels at US military bases in southwest Asia are high relative to those in the United States.17 It is possible that this long-term high-level PM exposure results in chronic health effects, and such associations were not assessed in this case-crossover study. The merits and limitations of the PM sampling effort and this epidemiologic association study have been reviewed by a committee convened by the Board on Environmental Studies and Toxicology of the National Academy of Science.29
The epidemiologic literature about the health effects of PM is vast, although there are only a few published reports of PM health effects occurring among deployed military populations. Sanders et al30 conducted a survey of 15,000 military personnel deployed to OEF/OIF and estimated that 69.1% reported experiencing respiratory illnesses, of which 17% required medical care. The frequency of respiratory conditions doubled from precombat to combat operations in this group. In an earlier study among US troops deployed to Bosnia during 1997–1998, Hastings and Jardine31 observed a statistically significant association between weekly maximum PM10 level and weekly frequency of upper respiratory disease. Increases in the frequency of upper respiratory disease with increasing quartiles of average PM10 levels were observed but were not statistically significant. Roop et al32 reported that 5% of troops deployed in support of OEF/OIF and surveyed for a study of respiratory symptoms reported a previous diagnosis of asthma. In that study, asthmatics and nonasthmatics reported statistically significantly increased respiratory symptoms (wheezing, cough, sputum production, chest pain/tightness, and allergy symptoms) during deployment relative to symptom prevalence before deployment. In a retrospective chart review of formerly discharged military personnel enrolled at the Northport VA, Szema et al6 observed a statistically significant positive association between deployment to OIF/OEF areas of operation and new-onset asthma. This group subsequently evaluated the frequency of referrals for pulmonary function testing in the Northport VA population, finding more such referrals among OIF/OEF veterans, relative to veterans with no history of OIF/OEF deployment, although smoking prevalence was also far higher in the deployed group.20
In summary, although the hypothesis under study has biological plausibility and some evidence in support, we found no indication of an association between PM and acute cardiorespiratory outcomes in this deployed soldier population. The null finding of our assessment may indicate that PM does not have a strong effect on the risk of acute adverse respiratory and cardiovascular events, but it may also be due to limitations inherent in the design of the study.
1. Weese CB, Abraham JH. Potential health implications associated with particulate matter exposure in deployed settings in southwest Asia. Inhal Toxicol. 2009;21:291–296.
2. World Health Organization. The World Health Report 2002. Geneva, Switzerland: World Health Organization; 2002.
3. Baucus M. Clean Air Act. Public Law No 101–549; 1990.
4. Mosher D, Lachman B, Greenberg M, Nichols T, Rosen B, Willis H. Green Warriors: Army Environmental Considerations for Contingency Operations From Planning Through Post-Conflict. Santa Monica, CA RAND; 2008.
5. King MS, Eisenberg R, Newman JH, et al. Constrictive bronchiolitis in soldiers returning from Iraq and Afghanistan. N Engl J Med. 2011;365:222–230.
6. Szema AM, Peters MC, Weissinger KM, Gagliano CA, Chen JJ. New-onset asthma among soldiers serving in Iraq and Afghanistan. Allergy Asthma Proc. 2010;31:67–71.
7. Institute of Medicine. Long-term health consequences of exposure to burn pits in Iraq and Afghanistan. In: Press TNA, ed. Washington, DC: Institute of Medicine; 2011.
8. Dominici F, Peng RD, Bell ML, et al. Fine particulate air pollution and hospital admission for cardiovascular and respiratory diseases. JAMA. 2006;295:1127–1134.
9. Samoli E, Peng R, Ramsay T, et al. Acute effects of ambient particulate matter on mortality in Europe and North America: results from the APHENA study. Environ Health Perspect. 2008;116:1480–1486.
10. Samet J, Krewski D. Health effects associated with exposure to ambient air pollution. J Toxicol Environ Health A. 2007;70:227–242.
11. Pope CA III. Epidemiology of fine particulate air pollution and human health: biologic mechanisms and who's at risk? Environ Health Perspect. 2000;108(suppl 4):713–723.
12. Holgate ST, Devlin RB, Wilson SJ, Frew AJ. Health effects of acute exposure to air pollution. Part II: Healthy subjects exposed to concentrated ambient particles. Res Rep Health Eff Inst. 2003:31–50; discussion 51–67.
13. Samet JM. Air pollution risk estimates: determinants of heterogeneity. J Toxicol Environ Health A. 2008;71:578–582.
14. Pfizer. The Health Status of the United States Army: Findings From the Total Army Injury and Health Outcomes Database (TAIHOD). Groton, CT: Pfizer; 2003.
15. Aldrich TK, Gustave J, Hall CB, et al. Lung function in rescue workers at the World Trade Center after 7 years. N Engl J Med. 2010;362:1263–1272.
16. Poschl U. Atmospheric aerosols: composition, transformation, climate, and health effects. Angew Chem Int Ed Engl. 2005;44:7520–7540.
17. Engelbrecht JP, McDonald EV, Gillies JA, Jayanty RK, Casuccio G, Gertler AW. Characterizing mineral dusts and other aerosols from the Middle East–Part 1: ambient sampling. Inhal Toxicol. 2009;21:297–326.
18. Bell ML, Dominici F, Ebisu K, Zeger SL, Samet JM. Spatial and temporal variation in PM(2.5) chemical composition in the United States for health effects studies. Environ Health Perspect. 2007;115:989–995.
19. Smith B, Wong CA, Smith TC, Boyko EJ, Gackstetter GD. Newly reported respiratory symptoms and conditions among military personnel deployed to Iraq and Afghanistan: a prospective population-based study. Am J Epidemiol. 2009;170:1433–1442.
20. Szema AM, Salihi W, Savary K, Chen JJ. Respiratory symptoms necessitating spirometry among soldiers with Iraq/Afghanistan war lung injury. J Occup Environ Med/Am Coll Occup Environ Med. 2011;53:961–965.
21. Maclure M. The case-crossover design: a method for studying transient effects on the risk of acute events. Am J Epidemiol. 1991;133:144–153.
22. Janes H, Sheppard L, Lumley T. Case-crossover analyses of air pollution exposure data: referent selection strategies and their implications for bias. Epidemiology. 2005;16:717–726.
23. Engelbrecht JP, McDonald EV, Gillies JA, Gertler AW. Department of Defense Enhanced Particulate Matter Surveillance Program (EPMSP). Reno, NV: Desert Research Institute; 2008.
24. Rothman KJ, Greenland S, Lash TL. Modern Epidemiology. Philadelphia, PA: Lippincott Williams & Wilkins; 2008.
25. Frumkin H. Understanding the Health Effects of Components of the Particulate Matter Mix: Progress and Next Steps. Cambridge, MA: Health Effects Institute; 2002.
26. Alberg AJ. Cigarette smoking: health effects and control strategies. Drugs Today (Barc). 2008;44:895–904.
27. Smith B, Ryan MA, Wingard DL, Patterson TL, Slymen DJ, Macera CA. Cigarette smoking and military deployment: a prospective evaluation. Am J Prev Med. 2008;35:539–546.
28. Maclure M, Mittleman MA. Should we use a case-crossover design? Ann Rev Public Health. 2000;21:193–221.
29. National Research Council. Review of the Department of Defense Enhanced Particulate Matter Surveillance Program Report (2010). Washington, DC: National Research Council of the National Academies; 2010.
30. Sanders JW, Putnam SD, Frankart C, et al. Impact of illness and noncombat injury during Operations Iraqi Freedom and Enduring Freedom (Afghanistan). Am J Trop Med Hyg. 2005;73:713–719.
31. Hastings DL, Jardine S. The relationship between air particulate levels and upper respiratory disease in soldiers deployed to Bosnia (1997–1998). Mil Med. 2002;167:296–303.
32. Roop SA, Niven AS, Calvin BE, Bader J, Zacher LL. The prevalence and impact of respiratory symptoms in asthmatics and nonasthmatics during deployment. Mil Med. 2007;172:1264–1269.
This article has been cited 3 time(s).
Clinics in Chest MedicineMilitary Service and Lung DiseaseClinics in Chest Medicine
Adverse health consequences of the Iraq War
Journal of Occupational and Environmental MedicineHealth Hazards of Exposures During Deployment to WarJournal of Occupational and Environmental Medicine
©2012The American College of Occupational and Environmental Medicine