A, Baseline radiograph showing pronounced diffuse mixed alveolar and
interstitial infiltrates consistent with acute lung injury. B, Radiograph
taken approximately 3 months later documents resolution.
See “Methods” section for definitions of severe, nonsevere,
probable, and definite. There were no patients in the “nonsevere, definite”
Customize your JAMA Network experience by selecting one or more topics from the list below.
Shorr AF, Scoville SL, Cersovsky SB, et al. Acute Eosinophilic Pneumonia Among US Military Personnel Deployed in or Near Iraq. JAMA. 2004;292(24):2997–3005. doi:10.1001/jama.292.24.2997
Author Affiliations: Pulmonary, Critical Care,
& Sleep Medicine Service (Dr Shorr) and Allergy and Immunology Service
(Dr Carr), Walter Reed Army Medical Center, Washington, DC; US Army Center
for Health Promotion and Preventive Medicine, Aberdeen Proving Ground, Md
(Drs Cersovsky, Scoville, Shanks, and Petruccelli); and Departments of Communicable
Diseases and Immunology (Dr Ockenhouse) and Preventive Medicine (Dr Smoak),
Walter Reed Army Institute of Research, Silver Spring, Md.
Context Acute eosinophilic pneumonia (AEP) is a rare disease of unknown etiology
characterized by respiratory failure, radiographic infiltrates, and eosinophilic
infiltration of the lung.
Objectives To describe a case series of AEP, illustrate the clinical features of
this syndrome, and report the results of an epidemiologic investigation.
Design, Setting, and Participants Epidemiologic investigation of cases of AEP identified both retrospectively
and prospectively from March 2003 through March 2004 among US military personnel
deployed in or near Iraq. Survivors were offered a follow-up evaluation.
Main Outcome Measure Morbidity and mortality related to AEP.
Results There were 18 cases of AEP identified among 183 000 military personnel
deployed in or near Iraq during the study period, yielding an AEP incidence
of 9.1 per 100 000 person-years (95% confidence interval, 4.3-13.3).
The majority of patients (89%) were men and the median age was 22 (range,
19-47) years. All patients used tobacco, with 78% recently beginning to smoke.
All but 1 reported significant exposure to fine airborne sand or dust. Known
causes of pulmonary eosinophilia (eg, drug exposures or parasitic disease)
were not identified. Epidemiologic investigation revealed no evidence of a
common source exposure, temporal or geographic clustering, person-to-person
transmission, or an association with recent vaccination. Six patients underwent
bronchoalveolar lavage (median eosinophilia of 40.5%). All patients developed
peripheral eosinophilia (range, 8%-42%). Mechanical ventilation was required
in 67% for a median of 7 (range, 2-16) days. Two soldiers died; the remainder
responded to corticosteroids and/or supportive care. Twelve individuals were
reevaluated a median of 3 months after diagnosis. At that point, 3 patients
reported mild dyspnea and 1 reported wheezing. All patients had finished treatment
and had either normal or nearly normal spirometry results. None had recurrent
Conclusions AEP occurred at an increased rate among this deployed military population
and resulted in 2 deaths. Failure to consider AEP in the differential diagnosis
of respiratory failure in military personnel can result in missing this syndrome
and possibly death. The etiology of AEP remains unclear, but the association
with new-onset smoking suggests a possible link.
Eosinophilic lung diseases comprise a variety of processes ranging from
Churg-Strauss syndrome to drug reactions.1 Acute
eosinophilic pneumonia (AEP) is a recently described syndrome characterized
by a febrile illness, diffuse infiltrates on chest radiograph, and pulmonary
eosinophilia.2 Generally, patients with AEP
present with respiratory failure requiring mechanical ventilation.3,4 Clinicians may initially confuse AEP
with severe community-acquired pneumonia, acute respiratory distress syndrome
(ARDS), or both. It is unclear if less-severe forms of AEP exist.
Histopathologically, lung biopsies in patients with AEP reveal both
acute and organizing diffuse alveolar damage with eosinophils filling both
the alveolar and interstitial spaces.5 Peripheral
eosinophilia may be noted in AEP; however, it is often absent at the time
of presentation, which complicates efforts at diagnosis and case identification
and suggests that the initial insult in this disease occurs in the lungs.
Although a number of agents and infections are associated with pulmonary eosinophilia,
the diagnosis of AEP requires the exclusion of known causes of pulmonary eosinophilia.6 Corticosteroids remain the mainstay of therapy for
AEP, and relapses have not been reported.2-4
Unfortunately, no prospective studies describe the natural history of
this disease and no controlled trials exist to guide clinicians in their use
of corticosteroids. Acute eosinophilic pneumonia differs from chronic eosinophilic
pneumonia in that many patients with the latter have a preceding history of
asthma, a median duration of symptoms prior to diagnosis of approximately
7 months, and rarely progress to respiratory failure.7
Acute eosinophilic pneumonia is thought to be a rare disorder and few
cases of AEP have been reported in the medical literature. The etiology of
AEP is unknown, although prior case series have indicated a potential relationship
between tobacco use and AEP.4,8 Severe
pneumonia was previously reported from March through August 2003 among 19
US military personnel who were deployed in support of Operation Iraqi Freedom;
2 of these patients died.9 Ten of these 19
were diagnosed with AEP; an additional 8 patients were diagnosed through March
2004. This article describes these 18 cases of AEP and the clinical features
of this syndrome and reports the results of the epidemiologic investigation
of these cases.
We defined cases of AEP based on a modification of criteria proposed
by Philit et al.4 Specifically, patients had
to report a febrile illness followed by the development of respiratory symptoms
such as cough, dyspnea, or both. Symptoms had to be present for less than
1 month and patients had to have evidence of infiltrates on chest radiograph.
Unlike Philit et al,4 we included individuals
who did and did not develop respiratory failure to identify all possible cases
of AEP. Patients with evidence of pulmonary eosinophilia based on either bronchoalveolar
lavage (BAL) or lung biopsy were classified as definite cases of AEP. Patients
who did not undergo BAL or biopsy but who developed peripheral eosinophilia
(total eosinophil count, >250 cells × 103/mL; percentage
of eosinophils, ≥10% of differential cell count) in the setting of an acute
respiratory illness with new infiltrates were categorized as probable cases
of AEP. Patients needing mechanical ventilation were categorized as “severe”
cases of AEP. In all instances, known causes of both eosinophilia and acute
infection had to be excluded. Laboratory evaluation included complete blood
cell count, metabolic profile, C-reactive protein level, and erythrocyte sedimentation
rate. In addition, patients underwent evaluation for potential parasitic infection,
multiple cultures (sputum, blood, urine, stool) and serologic testing for
other acute infectious processes, and serum studies for diseases known to
be associated with pulmonary eosinophilia (Box). Patients were also interviewed about exposures that may result
in pulmonary eosinophilia, and vaccine histories were reviewed.
After several cases of AEP were identified in July 2003 based on the
criteria outlined above, prospective disease surveillance began both in the
US Central Command area of responsibility (which stretches from the Horn of
Africa to Central Asia; hereafter referred to as “the theater”)
and at US military medical treatment facilities. Based on reviews of military
medical records, we also conducted a retrospective search for cases back to
Serologic Evaluation for Infection
Chlamydia spp (C pneumoniae, C psittaci, and C trachomatis)
Coxsackie B (1-6)
Influenza A and B
Legionella (serum and urine)
Parainfluenza (1, 2, 3)
Respiratory syncytial virus
Rickettsial agents (Rocky Mountain spotted fever group, typhus, Q fever)
Eosinophil cationic protein
Hypersensitivity pneumonitis panel*
*Includes antibodies to Alterneria tenuis, Cephalosporium acremonium, Candida albicans, Micropolyspora faeni, Thermoactinomyces sacchari, Aspergillus spp, Penicillium spp, Rhizopus nigricans, Geotrichum candidum, Fusarium vasinfectum, and pigeon dropping extract.
The major end point of interest was the clinical characterization of
the AEP syndrome. Hence, we collected data regarding patient demographics,
symptom prodrome, and severity of illness at presentation. We also recorded
information regarding initial chest radiograph appearance and temporal trends
in the evolution of peripheral eosinophilia for each case. In addition to
therapeutic interventions, we noted whether patients were treated with corticosteroids
and their responses to these interventions. Specific management decisions
were left to the patients’ primary physicians and were not directed
by protocol. Need for mechanical ventilation, duration of mechanical ventilation,
and mortality also served as outcome measures.
All surviving patients with illness onset during 2003 were offered a
follow-up evaluation by a pulmonary physician (A.F.S.) and an allergist (W.W.C.)
at the Walter Reed Army Medical Center in the fall and winter of 2003. In
addition, 1 patient with illness onset in 2004 received a follow-up evaluation.
Four individuals were not available for follow-up because they had either
returned to the theater or declined evaluation. At these follow-up visits,
patients underwent a complete history and physical examination, allergen testing,
clinical screening for the presence of atopy, repeat chest radiograph, and
pulmonary function testing. We specifically sought evidence for AEP recurrence,
development of chronic eosinophilic pneumonia, or evolution of some new collagen
vascular disease. Many of the initial laboratory studies and serologic tests
were repeated. In addition, eosinophil cationic protein levels were measured.
The patients were also reinterviewed using a standardized questionnaire.
The US Army deployed an investigative team to Germany (July-September
2003) and another to Iraq (August-September 2003) in support of the epidemiologic
investigation of severe pneumonia.9 The Iraq
team visited 5 combat support hospitals in Kuwait and Iraq (Baghdad, Tikrit,
Balad, and Mosul) and reviewed patient records, laboratory results, and radiographs
to identify patients with lower respiratory tract illness. They interviewed
clinical staff and patients and attempted to identify common factors among
patients (eg, time, place, and symptoms) and whether any unusual exposures
or other risk factors may have contributed to these illnesses. Surrogate interviews
using a standardized questionnaire were completed by members of the military
units of the 2 soldiers who died. The Iraq team queried Iraqi Ministry of
Health personnel about pneumonia cases in the local population, specifically
searching for reports of AEP. The Germany team deployed to the Landstuhl Regional
Medical Center, the US military hospital receiving personnel medically evacuated
from the theater. They interviewed patients and designed a laboratory testing
A standardized questionnaire was used to interview patients and capture
demographic, exposure, and clinical data. It was not possible to complete
1 interview because this patient was identified for inclusion in the study
after recovering and returning to the theater. A shorter version of the standardized
questionnaire (absent the clinical collection tool) was self-administered
in Iraq to a convenience sample of 72 members of the respective military units
of the 2 soldiers who died (controls). Patients and controls were questioned
about the following exposures: use of tobacco products (cigarettes and cigars);
dust; sleeping location and duration; petroleum products; bulk ammunition;
solvents or other chemicals; medical waste; close contact with the local population
or prisoners of war; local sources of water; burning vehicles or buildings;
human waste or other refuse; local foods; animals or animal droppings; insects;
over-the-counter and prescription medications; insect repellants; and pesticide,
fungicide, or herbicide application. Patients were also asked if illicit drugs
were available. Smoking status was categorized as nonsmoker, new-onset smoker,
or chronic smoker. New-onset smoking was defined as initiation of smoking
in the theater (or immediately preceding deployment to the theater) among
former nonsmokers or prior smokers (who restarted after at least 1 year of
cessation). Patients who smoked cigarettes or cigars on a regular basis (irrespective
of quantity) prior to arrival in the theater were classified as chronic smokers.
As this was an epidemiologic investigation performed for express public
health reasons and to better define the clinical syndrome, formal institutional
review board procedures were not required. Additionally, the study was performed
at the direction of the Office of the US Army Surgeon General as part of its
responsibility to protect the health of service members. Neither patients
nor controls in the case-control study were required to participate.
The crude rate of AEP was calculated by dividing the number of cases
by an estimate of the average number of military personnel at risk during
the 13-month period. Denominator data were obtained from the US Department
of Defense.10 Comparisons between groups were
tested using the Fisher exact test. Statistical analyses were performed using
SPSS version 10.0 (SPSS Inc, Chicago, Ill); P<.05
represented statistical significance.
Cigarette and cigar samples from the theater were collected and analyzed
for several potential environmental agents. Cigarette samples were collected
by US Army preventive medicine specialists from local merchants and from the
Army and Air Force Exchange Service in Baghdad during late August and early
September 2003. Specifically, 15 different brands (2 cartons of each brand)
were collected. These products were manufactured in a variety of nations including
the United States, France, England, Korea, Jordan, and Iraq. Cigarette control
samples (4 cartons, 2 brands) were collected from the Army and Air Force Exchange
Service at Aberdeen Proving Ground, Md. In addition, a cigar that was partially
smoked by a patient in the theater (in Djibouti) immediately prior to symptom
onset was available for testing. Laboratory analyses included the standard
US Food and Drug Administration pesticides screen with additional screening
for paraquat and diquat; bacterial and fungal colony counts and identification
of prevalent species; and special-threat agents (including ricin, strychnine,
picrotoxinin, and lobeline). Cigarette filters were similarly analyzed. Supporting
analytical laboratories included the US Army Medical Research Institute of
Infectious Diseases, US Department of Agriculture, and the US Food and Drug
Administration. Samples were also sent to clinical investigators at Northwestern
University Medical School, Chicago, Ill, for analyses of allergic-immunologic
cross-reactions with serum samples from the case patients. Serum samples from
patients were tested against a variety of fungal agents and an extract made
from the tobacco purchased in the theater to evaluate patients for evidence
of hypersensitivity pneumonitis.
During the 13-month period, 18 cases of AEP were diagnosed, with 7 meeting
criteria for definitive AEP on the basis of either BAL or lung histology results.
The median age of the cohort was 22 (range, 19-47) years and included 2 women.
There were 2 deaths, resulting in a case-fatality rate of 11%. One individual
died during aeromedical evacuation from Iraq before AEP was diagnosed. The
second death occurred in a patient receiving mechanical ventilation during
treatment for AEP and was due to the development of nosocomial pneumonia.
Respiratory cultures were positive for Klebsiella pneumoniae, and the autopsy confirmed the presence of bacterial superinfection.
Symptoms characterized by half of the patients during the epidemiologic
interview included shortness of breath, fever/chills, fatigue, and/or cough
accompanying the onset of the syndrome (Table
1). No patient reported rash; 3 described joint pain. Time in the
region before becoming ill varied widely (1 day to 11 months). The time between
illness onset and presentation for medical evaluation was a median of 1 day
(range, 1-4 days).
Chest radiographs revealed bilateral alveolar infiltrates in 10 patients;
infiltrates were unilateral in the remainder. The infiltrates were alveolar
in 10 cases and mixed alveolar-interstitial in 8. Pleural effusions were seen
in only 1 individual. A typical chest radiograph is displayed in Figure 1A. Computed tomography scans obtained
on admission to the intensive care unit in 6 of 12 patients demonstrated dense
alveolar consolidation and pulmonary edema consistent with acute lung injury
or ARDS. The pattern of the injury appeared to follow a bronchovascular distribution.
Neither chest radiographs nor computed tomography scans revealed the presence
The proportion of eosinophils in BAL fluid ranged from 25% to 74% (median,
40.5%) (Table 2). Of the 6 BALs performed,
4 were performed within 24 hours of patient arrival at Landstuhl Regional
Medical Center (after evacuation from Iraq), while the remainder were deferred
for approximately 96 hours because of the patient’s clinical status.
On initial evaluation at Landstuhl Regional Medical Center, all patients had
normal peripheral eosinophil counts. Despite evidence of pulmonary eosinophilia
in the patients who underwent BAL, peripheral eosinophilia did not peak until
approximately 3 days after admission. The median peak peripheral eosinophil
count measured 20% and 2500 cells × 103/mL (Table 2).
Twelve of the 18 patients had severe disease requiring mechanical ventilation
(Table 2). Nine of the 12 persons receiving
mechanical ventilation required at least 4 days of ventilatory support, while
2 were extubated while in the intensive care unit at Landstuhl Regional Medical
Center within 48 hours of arrival from Iraq. Excluding the soldier who died
prior to transport out of Iraq, the median duration of mechanical ventilation
was 8 days. Among those 12 treated in the intensive care unit and requiring
mechanical ventilation, the mean ratio of arterial partial pressure of oxygen
(PaO2) to fraction of inspired oxygen (FIO2) measured 170 (range, 88-232). All patients requiring
mechanical ventilation met criteria for ARDS.11 Three
required transient (<12 hours) treatment with vasopressors after air transport
from Iraq. Conventional modes of mechanical ventilation were used, with the
amount of positive end-expiratory pressure never exceeding 15 mm Hg in any
patient. Two patients needed inversion of the inspiratory-expiratory ratio
on the ventilator to maintain adequate oxygen levels. Highly aggressive interventions
for ARDS, such as prone positioning, high-frequency jet ventilation, and extracorporeal
oxygenation were not required. Only 1 patient received paralytic medications,
but this was for less than 12 hours. None of the 17 patients arriving at Landstuhl
Regional Medical Center progressed to respiratory failure if they were evacuated
from Iraq breathing spontaneously.
Serum measures of liver function and renal function were normal on presentation
and remained normal in all survivors. None of the survivors progressed to
refractory shock, organ failure, or both. The soldier who died of nosocomial
pneumonia experienced renal failure and refractory shock, but only after acute
respiratory failure and peripheral eosinophilia had begun to resolve.
Patients with severe AEP were uniformly treated with at least 7 days
of broad-spectrum intravenous antibiotics. The most commonly used regimen
included a combination of imipenem/cilistatin, levofloxacin, and doxycycline.
Nonsevere cases were given intravenous antibiotics in some instances if they
required hospitalization. For those not needing admission, physicians prescribed
oral antibiotics. All but 4 patients received corticosteroids. Patients with
severe AEP were given methylprednisolone intravenously. Nonsevere patients
were treated with oral prednisone. In all instances, corticosteroids were
tapered off over a 4- to 6-week period. Patients given corticosteroids had
improvement in their respiratory status within 96 hours. However, clearing
of their infiltrates lagged. Among these individuals, time to complete radiographic
resolution was longer in those with severe AEP than in those not needing mechanical
ventilation (11 vs 4 days). Individuals with unilateral disease on chest radiograph
also had more rapid radiographic clearing than those with bilateral disease.
However, this likely reflects the fact that no patient with unilateral disease
required mechanical ventilation. There was no apparent association between
severity of respiratory failure as measured by the PaO2/FIO2 ratio and radiographic resolution.
The soldier who died prior to aeromedical evacuation did not receive corticosteroids.
The 3 other patients not treated with corticosteroids improved with supportive
There was no difference between patients receiving mechanical ventilation
and those not receiving mechanical ventilation with respect to their symptom
prodrome or time in region prior to disease onset. All intubated patients
had bilateral infiltrates on chest radiograph, while 3 of 6 less-severely
ill patients presented with unilateral infiltrates (P = .02).
There was no difference in peak peripheral eosinophil counts as a function
of need for mechanical ventilation (mean peak eosinophil count, 2877 (SD,
1538) cells × 103/mL vs 1942 (SD, 1544) cells × 103/mL for patients receiving mechanical ventilation vs those not receiving
Follow-up Evaluation. Twelve of the 16 survivors
were seen for follow-up evaluation a median of 3.5 (range, 1-4) months after
the initial diagnosis of AEP. At this point, no patient was still being treated
with corticosteroids. All individuals had normal chest radiograph results
at reevaluation. Three reported residual dyspnea that was self-graded as mild,
and 1 reported wheezing. Subsequent pulmonary function testing included both
bronchodilator challenge and measurement of carbon monoxide diffusing capacity
(DLCO). Spirometry results were normal in all individuals.
The mean forced vital capacity (FVC) was 97% (SD, 12%) of predicted (range,
76%-114% of predicted), while the mean forced expiratory volume in one second
(FEV1) was 94% (SD, 11%) of predicted (range, 75%-111% of predicted).
No patient had airflow obstruction or a positive bronchodilator response to
albuterol inhalation. The FEV1/FVC ratio varied from 0.77-0.99.
The DLCO was mildly reduced (eg, 65%-75% of predicted) in the
3 patients complaining of residual dyspnea. Among those seen for follow-up
the mean DLCO was 82% (SD, 9%) of predicted (range, 66%-99%
Results of serologic testing for a number of infections and collagen
vascular disorders were negative. Allergy testing demonstrated that all patients
had at least 1 positive percutaneous skin prick test result using a standard
aeroallergen panel of trees, grasses, weeds, molds, and environmental allergen
extracts (US Army Centralized Allergy Extract Laboratory, Silver Spring, Md).
In addition to this panel, a tobacco leaf extract was also used (Greer Labs,
Lenoir, NC). There was no pattern in the antigens resulting in positive skin
test results and no patients tested positive to tobacco leaf extract. A tobacco
smoke extract was developed using albumin-condensated tobacco smoke, and enzyme-linked
immmunosorbent assay testing revealed no evidence of antismoke antibodies
in those patients tested (Northwestern University Medical School). The level
of eosinophilic cationic protein was normal in all 12 of the survivors seen
for follow-up evaluation.
An average of 183 000 military personnel were in or around Iraq
during this time, yielding an AEP incidence of 9.1 per 100 000 person-years
(95% confidence interval, 4.3-13.3). Patient demographics reflected the population
serving in the theater with a median age of 22 years (range, 19-47 years).
Fifteen patients (83%) were in the Army; 2 were in the Navy, and 1 was in
the Marine Corps. All but 2 patients were men and were from multiple military
specialties ranging from infantry to combat medic. Fourteen (78%) were junior
enlisted personnel, 3 were noncommissioned officers, and 1 was an officer.
Patients served in both active and reserve components and were from different
military units. Twelve patients were assigned to units in Iraq, 2 in Kuwait,
2 in Djibouti, 1 in Qatar, and 1 in Uzbekistan.
There was no evidence of a common source exposure or person-to-person
transmission; cases were not clustered temporally. However, AEP incidence
peaked in the summer months, with more than half (55%) of patients becoming
ill from June through August (Figure 2).
The most frequently reported exposures among patients were smoking tobacco
(100%), fine airborne sand or dust (94%), convoy operations (76%), and close
contact with the local population (71%). Nine of 15 patients reported that
illicit drugs were available but only 2 stated they had used them while deployed.
Tobacco smoking was the only exposure that was more common among patients
than controls (Table 3). All of the
patients in the theater reported smoking tobacco and 14 (78%) were new-onset
smokers. In contrast, 48 controls (67%) in the theater reported smoking tobacco
and only 2 reported that they started during this deployment. Therefore, military
personnel who were new-onset smokers had a significantly increased risk (P<.001) of AEP compared with controls (odds ratio, 122;
95% confidence interval, 17-1270).
Patients categorized as new-onset smokers began smoking a median duration
of 1 month (range, 2 weeks to 2 months) prior to illness onset. Of the 4 chronic
smokers, 2 were unique in that their occasional tobacco smoking (prior to
arriving in the theater) increased in quantity while in the theater. The quantity
of cigarettes smoked among new-onset smokers ranged from 2 to 10 cigarettes
per day. No patient reported smoking more than 1 pack of cigarettes per day.
Although 2 patients did not smoke cigarettes, they had recently started smoking
cigars or cigarillos. Members of the military units of the 2 deceased soldiers
reported that one had recently started to smoke and the other was an occasional
smoker who had increased his cigarette quantity in the theater.
Patients used a variety of tobacco brands, all of which were manufactured
in the United States (and purchased at the Army and Air Force Exchange Service
or mailed in packages from family members). Ten of the patients and 32 of
the controls also purchased cigarettes from local merchants (eg, street vendors).
Analysis of tobacco products obtained from the theater revealed no unusual
components, toxins, or pesticide residues.
Specifically, there were no detectable residues of ricin, strychnine,
picrotoxinin, lobeline, or paraquat/diquat or other pesticides in any cigarette
sample (tobacco, paper, and filters). Mold colonies developed on dilution
platings for 3 brands of cigarettes, and in each case the number of colony-forming
units per gram of tobacco was low (<100 colony-forming units per gram of
The majority of patients (67%) underwent vaccination against both smallpox
and anthrax prior to developing AEP. However, 6 patients never received smallpox
vaccine. All patients received at least 1 dose of anthrax vaccine; only 1
patient completed the 6-shot series (median number of anthrax doses received
was 3). The time between vaccination with either agent and onset of symptoms
varied from 3 to 11 months.
Although AEP is thought to be a rare disorder, we identified 18 cases
of AEP among 183 000 military personnel deployed in or near Iraq, with
an incidence of 9.1 per 100 000 person-years. Inquiries to the Iraqi
health officials did not suggest that AEP was occurring in the local population
or that there had been an unusual increase in the incidence of pneumonia of
any kind during the study period.
In our case series, 2 patients died, 1 from a nosocomial superinfection
and the other from rapidly progressive and refractory respiratory failure.
Most often, patients with AEP survive if treated promptly with corticosteroids.
The high case-fatality rate we noted underscores the seriousness of this entity
and the fact that clinicians should consider AEP in the differential diagnosis
of patients presenting with respiratory failure. Additionally, because of
the increased incidence of AEP in this population, we recommend that military
personnel presenting with unexplained respiratory failure undergo bronchoscopy
to exclude AEP and that their clinical syndrome not be assumed to simply represent
severe community-acquired pneumonia. For patients needing mechanical ventilation,
their treatment was not complicated by other organ failures. However, their
respiratory failure was very severe and necessitated use of high levels of
positive end-expiratory pressure and inversion of the inspiratory-expiratory
ratio in certain instances.
All patients with AEP were smokers and 14 of 18 began to use tobacco
shortly after deployment. Comparison with controls suggests an association
between recent-onset smoking and AEP. Previous reports have suggested a link
between recent-onset smoking and AEP.8,12,13 For
example, Nakajima et al12 described several
patients in Japan whose AEP was diagnosed soon after the patients began smoking.
Similarly, Philit et al4 completed a large
retrospective review of AEP in France and noted that 6 of 8 individuals with
AEP who smoked had begun doing so within the 3 months preceding disease onset.
Bolstering the hypothesis that there is a nexus between recent tobacco exposure
and AEP, Shintani et al8 reexposed an individual
with AEP thought to be related to tobacco. After reexposure, the AEP, which
had resolved, returned. Among our patients, 1 had returned to the theater
and restarted smoking 14 months after the onset of AEP. To date, AEP has not
recurred in any of these patients.
In an animal model of hypersensitivity pneumonitis, which is both distinct
from AEP and rare in smokers, nicotine has been shown to alter the expression
of certain cytokines in the lung.14 In turn,
the balance between TH1 and TH2 lymphocytes shifts and
the TH1 phenotype is suppressed. This change can lead to expression
of chemokines that attract eosinophils and might explain biologically any
relationship between smoking and AEP. However, given the prevalence of smoking
in both the general population and the deployed US military population (approximately
1 in 3 service members) as well as the rarity of AEP, it seems that tobacco
alone is unlikely to be either a necessary or sufficient condition for the
development of AEP. Service members have easy access to tobacco, with the
ability to purchase it from local vendors or the military exchanges. Family
members also mail tobacco products to deployed personnel. Efforts are under
way to counsel deployed service members against tobacco use, and the military
has several tobacco cessation initiatives.
Dust may also play some role in the etiology of the AEP cases we observed,
as all but 1 of our patients described significant exposure to fine airborne
sand or dust. Small particles in dust can irritate the airway and cause local
inflammation that could result in a state leading to the production of cytokines
such as interleukin 5, a potent recruiter of eosinophils. Several earlier
cases of AEP have been reported arising after dust exposure.15,16 Recently,
Rom et al17 described a firefighter who developed
AEP after being exposed to dust from the collapse of the New York World Trade
Center. Bronchoalveolar lavage revealed increased levels of interleukin 5
and stimulated CD4+ cells. Mineralogic analysis of the dust showed
high levels of asbestos fibers. Although asbestos was not noted in our cases,
other as-yet unidentified particles may have contributed to AEP. Alternatively,
recent exposure to tobacco may prime the lung in some way such that a second
exposure or injury, eg, in the form of dust, triggers a cascade of events
that culminates in AEP.
Other than tobacco use we could not identify epidemiologically a clear
toxin or exposure to account for the high incidence of AEP. Our patients represented
a variety of occupations and the disease appeared in a variety of locations,
both in and outside of Iraq. Onset of illness developed soon after deployment
in some patients, while others served in the theater for nearly a year before
becoming ill. The list of potential causes of pulmonary eosinophilia is large,
and it may be inappropriate to assume that whatever triggered this disease
was the same in each patient. Different exposures or combinations of exposures
may be responsible for the pulmonary eosinophilia and respiratory failure.
Inability to identify another cofactor, however, does not preclude that one
exists or that the cofactor might vary from patient to patient.
Eosinophilic lung disease has been reported previously among military
personnel.18 In 1997, 2 soldiers from the National
Training Center area at Fort Irwin, Calif, presented with ARDS shortly after
beginning training in this hot, desert environment. Despite broad-spectrum
antibiotics, the soldiers’ respiratory failure did not improve. Bronchoalveolar
lavage revealed eosinophilia, and both patients developed a significant peripheral
eosinophilia several days later. Notably, both soldiers had recently begun
smoking cigarettes. Since March 2004, there have been 4 additional cases of
AEP, all severe, among military personnel deployed in or near Iraq. Our findings
have prompted heightened surveillance for AEP, and from March 2003 to October
2004 we have identified 3 cases of AEP in military personnel not deployed
in or near Iraq (2 in Korea, 1 in Texas).
Many of the clinical features of AEP we noted do not match those previously
described. In earlier case series, normalization of the chest radiograph has
required nearly 1 month.3,4 Our
patients’ radiographs improved more quickly, ranging from 4 to 11 days.
This suggests that variability can be expected in the time to radiographic
clearing. Alternatively, the rapid clearing might reflect the fact that all
patients were expediently evacuated out of the area where they developed their
illness, thus ending their exposure to any potential toxin that might have
triggered the disease. Rapid radiographic improvement could also reflect that
all of our patients were otherwise young and healthy prior to the onset of
Furthermore, we noted 3 patients who improved without corticosteroids.
This has not been reported in previous case series. In fact, initial early
proposals for case definitions for AEP required a positive response to corticosteroids.
The ability of patients to improve without corticosteroids underscores the
possibility that removal from some as-yet unidentified agent or combination
of agents may be crucial for the care of AEP. On the other hand, this observation
implies that the AEP we observed may represent a unique syndrome, different
from traditional forms of AEP. Finally, AEP has always been thought to result
in respiratory failure.1-4 Many
of our patients did not progress to respiratory failure, which underscores
that less-severe forms of AEP likely exist. However, most of our patients
with less-severe AEP never underwent BAL to confirm the diagnosis, which limits
our ability to draw conclusions regarding this possibility.
Our analysis has several important limitations. First, although we specifically
conducted prospective surveillance for AEP, this report is technically retrospective,
with all the accompanying limitations of this approach, including both case
identification and recall bias. Second, as noted above, not all of our patients
underwent BAL. However, being conservative and excluding “probable”
cases still results in a high incidence of AEP. Third, the selection of the
control population for the epidemiologic analysis limits our ability to draw
strong inferences regarding potential causes and associations. The controls
used in this analysis completed a self-administered questionnaire and represented
only 2 military units that were deployed in Iraq. Finally, because patients
and controls both used a combination of different tobacco products manufactured
in different nations (eg, United States, Iraq), we could not determine if
one particular type of tobacco product contributed to AEP. However, the biochemical
and microbiological analyses of the cigarettes revealed no differences between
those made in the United States and those produced elsewhere.
In summary, we describe a case series of AEP among military personnel
deployed in or near Iraq. Patients can present with fulminant respiratory
failure or have less-severe forms of the disease, both of which can mimic
community-acquired pneumonia. Civilian and military physicians should both
consider this diagnosis in military personnel presenting with respiratory
complaints during, or after, a recent deployment or training exercise.
Corresponding Author: Andrew F. Shorr, MD,
MPH, Pulmonary, Critical Care, & Sleep Medicine Service, Walter Reed Army
Medical Center, 6900 Georgia Ave, Washington, DC 20307 (email@example.com).
Author Contributions: Dr Shorr had full access
to all of the data in the study and takes responsibility for the integrity
of the data and the accuracy of the data analyses.
Study concept and design: Shorr, Cersovsky,
Shanks, Smoak, Carr, Petruccelli.
Acquisition of data: Shorr, Cersovsky, Scoville,
Shanks, Ockenhouse, Smoak, Carr.
Analysis and interpretation of data: Shorr,
Scoville, Cersovsky, Shanks, Ockenhouse, Smoak, Carr, Petruccelli.
Drafting of the manuscript: Shorr, Cersovsky,
Shanks, Smoak, Carr.
Critical revision of the manuscript for important
intellectual content: Shorr, Scoville, Cersovsky, Ockenhouse, Smoak,
Statistical analysis: Shorr, Cersovsky, Shanks,
Obtained funding: Petruccelli.
Administrative, technical, or material support:
Cersovsky, Ockenhouse, Smoak, Carr, Petruccelli.
Study supervision: Cersovsky, Smoak, Petruccelli.
Investigating physicians in Iraq: Ockenhouse,
Funding/Support: The US Army Office of the
Surgeon General (OTSG) sponsored the study and provided the personnel dedicated
to this study.
Role of the Sponsor: The OTSG authorized the
submission of the manuscript but was not involved in the design, conduct,
data management, analysis, or manuscript preparation.
Disclaimer: The opinions expressed herein are
not to be construed as official or as reflecting the policies of the Department
of the Army or the US Department of Defense.
Acknowledgment: We thank the many people who
collaborated in and contributed to this investigation, including those working
at the US Centers for Disease Control and Prevention, Armed Forces Institute
of Pathology, Walter Reed Army Institute of Research, US Army Medical Research
Institute of Infectious Diseases, US Department of Agriculture, US Food and
Drug Administration, and Northwestern University Medical School. In particular
we thank Paul Greenberger, MD, Dollene Hemmerlein, BS, Matt McAtee, BS, Stephen
Redd, MD, MPH, and Eric Shuping, MD, MPH. We also acknowledge the important
efforts of the many medics, nurses, physicians, and other health care professionals
who cared for, and continue to care for, US military personnel who are currently
Create a personal account or sign in to: