Article Text
PDF
Abstract
Objectives Characterise inhalational exposures during deployment to Afghanistan and Southwest Asia and associations with postdeployment respiratory symptoms.
Methods Participants (n=1960) in this cross-sectional study of US Veterans (Veterans Affairs Cooperative Study ‘Service and Health Among Deployed Veterans’) completed an interviewer-administered questionnaire regarding 32 deployment exposures, grouped a priori into six categories: burn pit smoke; other combustion sources; engine exhaust; mechanical and desert dusts; toxicants; and military job-related vapours gas, dusts or fumes (VGDF). Responses were scored ordinally (0, 1, 2) according to exposure frequency. Factor analysis supported item reduction and category consolidation yielding 28 exposure items in 5 categories. Generalised linear models with a logit link tested associations with symptoms (by respiratory health questionnaire) adjusting for other covariates. OR were scaled per 20-point score increment (normalised maximum=100).
Results The cohort mean age was 40.7 years with a median deployment duration of 11.7 months. Heavy exposures to multiple inhalational exposures were commonly reported, including burn pit smoke (72.7%) and VGDF (72.0%). The prevalence of dyspnoea, chronic bronchitis and wheeze in the past 12 months was 7.3%, 8.2% and 15.6%, respectively. Burn pit smoke exposure was associated with dyspnoea (OR 1.22; 95% CI 1.06 to 1.47) and chronic bronchitis (OR 1.22; 95% CI 1.13 to 1.44). Exposure to VGDF was associated with dyspnoea (OR 1.29; 95% CI 1.14 to 1.58) and wheeze (OR 1.18; 95% CI 1.02 to 1.35).
Conclusion Exposures to burn pit smoke and military occupational VGDF during deployment were associated with an increased odds of chronic respiratory symptoms among US Veterans.
- occupational health
- air pollution
- respiratory system
Data availability statement
Data are available on reasonable request. Deidentified data are available for approved investigators with IRB and institutional approval.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
Associations between exposures during military deployment to Afghanistan, Iraq and other countries in Southwest Asia and respiratory health have been of concern but remain ill defined.
WHAT THIS STUDY ADDS
Among US military Veterans recruited randomly from a national deployment roster, responses to a structured exposure questionnaire were used to define deployment-related inhalational exposures, including to burn pits, to other combustion sources and to military job-related vapours gas, dusts or fumes (VGDF), with exposure category assignment supported by factor analysis. Prior burn pit smoke exposures were associated with dyspnoea and chronic bronchitis, while military occupational exposures to VGDF during deployment were associated with dyspnoea and wheeze.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
These findings support the need to better assess the risk of chronic respiratory symptoms and potential disease attributable to exposures in Veterans previously deployed to Afghanistan and Southwest Asia.
Introduction
During land-based deployments to Afghanistan and Southwest Asia (Iraq, Kuwait and other countries in the region), military personnel had the potential for a wide range of inhalational exposures.1 2 Exposure sources included desert dust storms and mechanically generated dust; open-air waste burning (burn pits and other trash burning); other open combustion sources; engine exhaust and exposures related to military occupational activities. Previous analyses of Department of Defence healthcare facility administrative records indicate that returning military personnel have presented with respiratory symptoms, but these were not further characterised diagnostically.3–5 Similarly, among military personnel in the Millennium Cohort Study who self-completed a health survey, those deployed to Afghanistan and Southwest Asia reported persistent or recurring cough or shortness of breath more frequently than non-deployers.6 Further, among Swedish military personnel deployed to that region compared with a civilian referent group, there was an increased prevalence of wheeze and chronic bronchitis; a positive association with time spent in a desert environment with wheeze; and a link between dust storm exposure and chronic bronchitis.7 Case series also support the observation that some Veterans deployed to this region have been diagnosed with chronic respiratory conditions, including asthma and constrictive bronchiolitis.8–15
Although these data are suggestive of an association between deployment and adverse health effects, they allow limited conclusions because they have incomplete characterisation of exposures by type or intensity and are restricted to persons seeking healthcare. To address this knowledge gap, the Veteran Affairs (VA) Cooperative Studies Programme #595, ‘Pulmonary Health and Deployment to Southwest Asia and Afghanistan’ (NCT02825654; also called SHADE (Service and Health among Deployed Veterans)) was initiated to assess the respiratory health of a representative sample of previously deployed Veterans without regard to previous illness or exposure. In this report, we analyse data from the first phase of this study to assess the performance of a structured deployment exposure assessment quesionnaire battery and explore associations between different deployment-related inhalational exposures and postdeployment dyspnoea, wheeze and chronic bronchitis.
Methods
Overview
This analysis uses structured survey data from the Veterans Affairs Cooperative Studies Programme #595. We collected these data during the initial phase of this study, which is currently ongoing. We assessed self-reported exposures with a multi-item battery designed for this study that formed the basis for summary scores for multiple categories of deployment-related environmental and occupational exposures. We also conducted a standardised assessment of chronic respiratory symptoms at the same study visit. Here, we describe associations with deployment-related environmental and occupational exposures. Some of these results were previously reported in abstract form.16 17
Study recruitment
Participants were randomly selected using Defense Manpower Data Center (DMDC) deployment roster records from among Veterans meeting the following study eligibility criteria: separated from active-duty US military service (but still could be a member of the National Guard or Reserve); served between October 2001 and February 2017; had one or more deployments during that period to at least one of seven countries (Afghanistan, Kyrgyzstan, Iraq, Kuwait, Qatar, United Arab Emirates or Djibouti); and was a member of a service branch that had land-based deployments (Air Force, Army, Marine Corps). We also required that at the time of recruitment potential study participants resided within 25 miles of the VA Medical Centre that was a study testing site (Atlanta, Georgia; Boston, Massachusetts; Houston, Texas; Minneapolis, Minnesota; San Diego, California, or Seattle, Washington). We updated address and telephone contact information from the DMDC using LexisNexus.
In the initial phase of the study that we report here, study sites conducted recruitment through a combination of postal mail and telephone to 6913 eligible individuals, of whom 2299 (33.3%) agreed to participate. Of these, 1967 completed onsite visits between April 2018 and March 2020, at which point recruitment for the SHADE study was paused due to the COVID-19 pandemic. All participants completed an in-person, interviewer-administered structured questionnaire that included demographic, smoking and health-based items, military service-related characteristics (including service branch and dates and locations of deployments), and a deployment-related exposure multi-item battery (see the Exposure assessment section).
Exposure assessment
The exposure battery elicited responses regarding 32 occupational or environmental deployment-associated exposures that might have occurred. As delineated in online supplemental tables 1–6, we grouped exposures, a priori, into six domains that broadly represent anticipated deployment-associated exposure sources. The exposure survey items focused on repeated or ongoing exposures that included: 11 questions regarding military occupation-related vapours gas dust or fumes (VGDF) (online supplemental table 1); four questions regarding exposure to burn pit smoke, including regular exercise near the burn pit and work at the burn pit (online supplemental table 2); four questions assessing exposure to open air combustion by-products other than burn pits, including exposures related to combat, other military activities, or oil well and refinery fires (online supplemental table 3); five questions regarding exposure to sources of combustion engine exhaust (online supplemental table 4); five questions regarding mechanically-generated dust exposures (including dust storms) (online supplemental table 5); and three questions regarding exposure to toxicants (including pesticides and chemical warfare agents) (online supplemental table 6). The 32-item exposure battery included an introductory text indicating that the questions referred to ‘heavy’ levels of exposure, such as a sustained or direct exposure or one that could be sensed clearly at the time, for example, through effects on the eyes, throat, or breathing. For affirmative responses, respondents were asked the number of months with exposure and then an estimated number of days in a typical month that exposure occurred. A single item assessing ‘no-fly-days’ due to poor air quality was quantified as the total number of days this occurred over deployment.
Supplemental material
Exposure scoring
We dichotomised self-reported exposures as no exposure vs any reported frequency (online supplemental tables 1–6). We categorised missing item responses as no exposure (3% of responses across the 32 exposure questions). For those with at least some exposure, we quantified intensity based on the frequency of occurrences expressed in person-days of exposure. We calculated the median number of exposure days among those with exposure for each item, generating a three-level ordinal score for each: 0, no exposure; 1, moderate exposure (defined by a frequency less than or equal to median for that item); and 2, heavy exposure (above the median frequency). We assigned a score=1 to any positive response that was not otherwise quantified.
To assess the construct validity of the six exposure domains as conceptualised a priori, we used an iterative process. Initially, we carried out a confirmatory factor analysis (CFA) to assess whether this yielded factor weights consistent with our a priori six domain structure.18 19 The initial CFA indicated that two of six factors (representing the domains of mechanically generated dust/dust storms and combustion engine exhaust exposure) were highly collinear (r=0.978), leading us to collapse these two into a single domain (online supplemental figure 1). The subsequent five-factor CFA identified four items with factor loadings <0.45 (two items in the VGDF domain and two in the newly combined domain of mechanically generated dust/combustion engine exhaust) that were removed from further analyses. Using the 28 remaining items, we calculated summary scores of the five domains. We did this by adding together the values contributed by each respondent (0, 1 or 2) for each item in the domain, scaling each cumulative domain score to a possible maximum of 100 (to standardise the scores among the exposure domains).
Symptom assessment
The health component of the survey included the assessment of current symptoms based on the American Thoracic Society and Division of Lung Diseases, National Heart, Lung, and Blood Insitute (ATS-DLD-78) respiratory questionnaire.20 We defined dyspnoea as a positive response to the question ‘other than strenuous exercise, are you troubled with breathlessness?’ excluding any respondents using an assistive device for mobility (including a crutch or cane). We defined chronic bronchitis as a report of both cough and phlegm for three consecutive months or more during the year and established how many of these had two or more consecutive years of such symptoms. We defined wheeze as any chest wheezing or whistling in the previous 12 months.
Statistical analyses
All analyses were conducted in R V.4.3.0.21 For descriptive characteristics, we used standard summary measures, including means (SD) or medians (IQR) for continuous normally and non-normally distributed variables, respectively. We used generalised linear models with a logit link to estimate the associations of all five factors included in the same model for each of the three respiratory outcomes (dyspnoea, chronic bronchitis and wheeze). We additionally adjusted for personal characteristics known to be associated with respiratory symptoms: age; sex; race (white non-Hispanic vs others); body mass index (BMI) (<25 (referent), 25 up to 30, ≥30); and cigarette smoking (three-level factor: never (referent), former, current). To account for possible correlation among participants recruited from the same study site, we estimated regression parameter CIs using cluster bootstrapping on site, which is a preferable choice over generalised estimating equations when there is a small number of clusters (six sites), especially in the setting of a binary response.22 We scaled the estimated ORs and 95% CIs to a 20 unit change in predictor score. The 20 unit change effectively represents a one fifth (quintile) range within a 100-point scaled score.
We also conducted a sensitivity analysis with separate models adjusting for educational status (<=high school, some college or >=bachelor’s degree) and civilian workplace exposure to VGDF, defined as at least 1 year of time spent in their primary full or part-time civilian job with report of regular exposure to vapours, dust, gas or fumes.
Results
Among 1967 potential participants, we excluded 5 participants who did not complete the questionnaire protocol leaving 1962 who contributed to exposure scoring, and excluded 2 from final analysis who had missing key covariate data. The characteristics of those invited and not invited for study participation (online supplemental table 7) indicated that the distributions of current age, deployment age, sex, race/ethnicity, service branch, educational level at last record of military service and service time were similar between the two groups. The 1960 participants were predominantly male (n=1734; 88.5%) and with a mean age of 40.7±9.7 years (table 1). Most participants at their study visit were overweight or obese (BMI≥25 kg/m2), most had received a Bachelor’s degree or higher degree (51.6%) and 256 (13.1%) were current smokers. Consistent with the study criterion of service branches likely to have land-based deployments, 1183 (60.4%) had been in the Army, with the remainder in the Marines or Air Force. Total deployment duration was a median (IQR) of 11.7 (7.4–16.1) months (range 0.3–148.6), with the study visit completed median (IQR) 120 (89.9–156.7) months postdeployment. Most participants (n=1118; 57.0%) reported a single deployment. There were 1914 persons without an assistive device included in the analysis of dyspnoea. The prevalence of dyspnoea, chronic bronchitis and wheeze was 7.3% (n=139), 8.2% (n=161) and 15.6% (n=305), respectively (table 2). Of the persons defined as having bronchitis for 1 year, 89% (n=144) had ≥2 years of cough and phlegm on most days for three consecutive months of the year. The prevalence of civilian workplace exposure to VGDF was 30.3%.
Participant characteristics (n=1960)
Multivariable models of exposure factor scores and associations with each symptom scaled per 20 unit change in exposure score, unadjusted and adjusted for other characteristics
Overall, 1425 (72.7%) of participants reported exposure to burn pit smoke, with the median duration of exposure to burn pit smoke outdoors, coming into one’s work site or housing, or from regular exercise or other physical exertion beside a burn pit site being up to 5.0 months (table 3 and online supplemental table 2). The median exposure to smoke or fumes from personally operating or working with trash incineration or at a burn pit was 1.67 months, reported by 712 (36.3%) (online supplemental table 2). In contrast, the median exposure to open combustion sources including related to combat or other military sources, reported by 1146 (58.5%), (table 3) ranged from 0.7 to 1.93 months, depending on the individual exposure item surveyed (online supplemental table 3). Exposure to vehicle engine or generator exhaust was of a longer duration, with a median of 8 months in 1310 (66.8%) reporting such exposure (online supplemental table 4). The job with the longest duration of exposure to VGDF was vehicle or aircraft refuelling or maintenance operations (medians of 5.0–6.53 months, online supplemental table 1). Selected toxicant exposures, predominantly from applying pesticide, insecticide or repellent, reported by 860 (43.8%), with a median duration of 2.8 months (online supplemental table 6). The most common single exposure was to a natural dust storm (online supplemental table 5) in 1759 (89.7%) with a median duration of 0.5 months (ie, a median of 15 days during deployment). The cumulative exposure scores among the five domains were positively correlated, with r values ranging from 0.32 to 0.56 (online supplemental table 8).
Exposure score, number exposed (% out of total n=1960) and normalised exposure score distribution (minimum=0, maximum=100)
The results of the crude and adjusted analyses are shown in table 2, presenting the OR and 95% CI associated with each of the 5 exposures per 20 unit increase in score. These analyses, in addition to including together all the exposure scores, also adjust for age, sex, BMI and smoking. Other than for toxicant exposures, there was not substantive change in the estimated ORs observed between analyses that did and did not include those covariates. Military exposures during deployment were associated with increased odds of respiratory symptoms, although the pattern of risk varied by outcome. In the adjusted analysis, burn pit smoke exposure was significantly associated with dyspnoea (OR 1.22; 95% CI 1.06 to 1.47) and chronic bronchitis (OR 1.22; 95% CI 1.13 to 1.44), but not with wheeze (OR 1.04; 95% CI 0.96 to 1.11). Military job-related exposures to VGDF during deployment were significantly associated with dyspnoea (OR 1.29; 95% CI 1.14 to 1.58) and wheeze (OR 1.18; 95% CI 1.02 to 1.35), but had a smaller association with chronic bronchitis (OR 1.14; 95% CI 0.95 to 1.54) that did not exclude the null. There were non-statistically significant estimates of association between ground dust/engine exhaust and chronic bronchitis (OR 1.20; 95% CI 0.96 to 1.58) and with wheeze (OR 1.14; 95% CI 1.00 to 1.43). Exposure to other combustion sources and toxicants were not significantly associated with any of the respiratory outcomes. In a sensitivity analysis adjusting for civilian workplace exposure to VGDF, there were no changes in the associations observed with deployment exposures, and no associations with civilian exposures (online supplemental table 9) other than an elevated OR for chronic bronchitis with wide confidence limits (OR 1.22; 95% CI 0.84 to 1.54). Similarly, adjusting for current education, there were no changes in the associations observed with deployment exposures, and no associations with education (online supplemental table 10).
There were also significantly increased odds of at least two of the three respiratory symptoms with other covariates, including increased odds of each symptom among women and among those with greater BMI, particularly ≥30. As expected, ex-smokers and current smokers had increased odds for each symptom. Other forms of smoking in the absence of a cigarette smoking history were uncommon. Only 33 (1.7%), 14 (0.7%) and 27 (1.4%) of the cohort who had not smoked cigarettes had used e-cigarettes, a pipe or cigars, respectively. As a result, these forms of smoking were not included as separate covariates in regression models.
Discussion
Using a novel exposure scoring scheme based on a multi-item questionnaire battery we characterised a range of inhalational exposures to particulate matter, smoke, gases and fumes experienced by military personnel during deployment to Southwest Asia and Afghanistan. We also identified significant associations between burn pit and military VGDF exposures and chronic respiratory symptoms. Importantly, the pattern of associations differed between burn pits and VGDF for each symptom. Moreover, associations of the other sources of exposures we analysed were weaker, arguing against a generic association with any self-reported exposure and symptoms that would be consistent with recall bias driving these observations.
These findings provide new and potentially important insights. Previous clinical case series and epidemiologic studies assessing deployment-related airborne hazards have considered deployment as a single common exposure or focused on a single source of exposure, such as burn pits. These reports have not considered the range of occupational and environmental inhalational exposures that can be encountered during deployment. Further, we observed the associations we report here in a cohort of US Veterans that were not selected or self-referred because of symptoms or having received healthcare drawn from a roster with demographics representative of the post-deployment Veteran population.
Although our results are based on self-report, which could be a source of bias in reporting associations, the methods we used mitigate against this. By designing a multi-item battery with a priori domains spanning a range of potential exposures and then undertaking a systematic item reduction strategy, we could minimise the impacts of exposure over-reporting and reduce the impacts of intersubject variability. We also provided a standard definition of exposure intensity to the participant. Self-reported heavy exposure to burn pit smoke, although common, reflected a range of responses summed across four items and avoided reduction to a dichotomous metric. The confirmatory factor analysis supports the construct validity of our battery, while providing a rationale for reasonable item reduction. By defining the respiratory outcomes, which we studied using commonly accepted respiratory health questions assessing wheeze, dyspnoea and chronic bronchitis in a standardised way, we also minimised imprecision in symptom-defined outcomes. Finally, because the associations between the scored exposure domains and each respiratory symptom varied it is unlikely that generic reporting bias entirely accounts for the associations we observed.
Our findings are consistent with studies on the effects of relevant exposures in other contexts. For example, there is a robust relationship between non-military occupational exposure to VGDF with dyspnoea and chronic bronchitis.23 24 The presence of chronic respiratory symptoms can be important, as when present in young to middle-aged adults they are associated with an increased risk of future pulmonary disease and reduced pulmonary function.25–27 Chronic exposure to ambient particulate matter to which combustion by products contribute, a key component of burn pit emissions, is also associated with reduced pulmonary function, and potentially chronic obstructive pulmonary disease.28 29 Together these findings suggest that the exposures and symptom-based outcomes that we studied have important implications for respiratory morbidity.
Although different in design, the most informative comparison regarding the prevalence of respiratory symptoms in US Veterans comes from the Millennium Cohort Study where a cohort of 9210 deployers was surveyed regarding persistent or recurring cough or shortness of breath before (2001–2003) and after deployment (2004–2006).6 New-onset cough or shortness of breath was reported by 10.2% and 6.6% following deployment, respectively, similar in magnitude to our report of chronic bronchitis and dyspnoea. Similarly, in Swedish military personnel surveyed 36 months to 5 years postdeployment the prevalence of wheeze (16.3% vs 12.3%) and chronic bronchitis (12.3% vs 6.8%) was significantly higher among soldiers than a referent group.7 Nonetheless, we recognise that data are not directly comparable to ours due to differences in time since deployment, age differences and other cohort characteristics.
Despite the study strengths, potential limitations of this analysis should be noted. We focused on current symptoms and did not assess diagnostic outcomes such as asthma or airflow obstruction, nor did we analyse pulmonary function measurements. We also did not analyse whether previous childhood asthma was related to later postdeployment symptoms, noting that those with ongoing medication-treated asthma were less likely than those without such illness to be deployed. We carried out this analysis using the initial phase of data collection from a larger study whose data collection was suspended during the COVID-19 pandemic, and which will yield a larger study cohort at a later date. Thus, the number of subjects included was not driven by considerations of study size and power. Our exposure battery was developed for this study and has not been externally validated in other cohorts of previously deployed Veterans. We also relied on exposure self-report and not direct measurements of exposures to define exposure categories, for example, from environmental monitoring or personal industrial hygiene sampling. Our participants were recruited from Veterans living in proximity to a select number of urban VA medical centres and thus may not be generalisable to Veterans living in other regions of the USA, such as rural areas. It is also possible that the recruitment pool of Veterans included a higher percentage of participants with symptoms compared with active-duty military personnel not eligible for the study. Although our sensitivity analysis adjusting for civilian occupational exposure to VGDF did not impact the observed associations, a more detailed consideration of civilian occupation would be of interest. We also did not include non-deployment military exposures in the analysis as the patterns and frequency of non-deployment military exposures could not be assessed using the same confirmatory factor analysis. We plan to address this complex question in a later analysis.
Although not designed to evaluate respiratory health outcomes in women, we found that women had increased odds of self-reported respiratory symptoms. This is consistent with previous reports noting that women are more likely to report dyspnoea and possibly other respiratory symptoms compared with men.30–32 It has been speculated that women may be more susceptible to inhalation hazards and the development of future disease, possibly attributable to their smaller airway and lung size compared with men.31 33 It is also noteworthy that we detected associations between an increased BMI and dyspnoea,34 between obesity and wheeze,35 and with smoking and all three outcomes. These are all known associations, which also provides evidence of the internal validity of the symptom outcomes reported.
Our results are consistent with the conclusion that multiple exposures encountered during military deployment increase the likelihood of chronic respiratory symptoms reported years after deployment. These associations may account at least in part for the chronic respiratory problems that have been reported in multiple clinical case series of Veterans who were deployed to Afghanistan and Southwest Asia.8–15 Taken together, our findings support recommendations that postdeployment Veterans be assessed for a wide range of occupational and environmental exposures that may contribute to chronic respiratory symptoms and future chronic respiratory diseases.36
Data availability statement
Data are available on reasonable request. Deidentified data are available for approved investigators with IRB and institutional approval.
Ethics statements
Patient consent for publication
Ethics approval
This study involves human participants and this study was approved by the Department of Veterans Affairs Central Institutional Review Board (protocol number 1613196) and all participants provided informed consent. Participants gave informed consent to participate in the study before taking part.
Acknowledgments
The authors would like to thank the Veterans participating in this study.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
Contributors EG, PDB and SPP designed the data collection tools. AK, AKT, NLS and KN monitored data collection and implemented the study together with EG and PDB who are the guarantors. PDB and EG designed the analysis plan, and AK conducted the statistical analysis with the assistance of AKT. FK, VSF, OI, PRM, ESW, CHW and CW supervised study data collection. JEH, PK and MJ contributed to study design. CAR, CPB, PC, WK and JEH contributed to data analysis and interpretation. EG and PDB drafted the paper and revised the draft paper with the contribution of CAR and AK. All authors approved the final manuscript.
Funding US Department of Veterans Affairs; Veterans Affairs Cooperative Studies Program #595: Pulmonary Health and Deployment to Southwest Asia and Afghanistan, also known as SHADE (Service and Health Among Deployed Veterans) from the Office of Research and Development, Clinical Science Research and Development, Cooperative Studies Program. The contents do not represent the views of the US Department of Veterans Affairs or the US Government.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.