Enteric Diseases Department, Naval Medical Research Center, Silver Spring, MD, USA ; â Department of .... Atlanta, GA, USA) using a valid double-data entry .... Visit sick call for injury ..... in the deployed environment, such as job descrip-.
392
Acute Gastrointestinal Infection, Respiratory Illness, and Noncombat Injury Among US Military Personnel During Operation Bright Star 2005, in Northern Egypt Mark S. Riddle, MD,* Heather A. Halvorson, MD,† Danny Shiau, MD,† Juliann Althoff, MD, MPH,† Marshall R. Monteville, PhD,‡ Hind Shaheen, PhD,‡ Edward P. Horvath, MD,§ and Adam W. Armstrong, DO‡ *Enteric Diseases Department, Naval Medical Research Center, Silver Spring, MD, USA; †Department of Preventive Medicine & Biometrics, Uniformed Services University, Bethesda, MD, USA; ‡Enteric Diseases Research Program, U.S. Naval Medical Research Unit No. 3, Cairo, Egypt; §U.S. Army, 256th Combat Support Hospital, Brooklyn, OH, USA DOI: 10.1111/j.1708-8305.2007.00159.x
Background. In the fall 2005, approximately 7,500 US military personnel participated in an exercise in the Egyptian desert. The epidemiology of disease and noncombat injury among deployed troops is important in the context of assessing current mitigation strategies and the development of future ones. Methods. To assess the prevalence and impact of diarrhea and enteropathogen distribution, we conducted a case series study. To assess the relative impact of diarrhea compared to respiratory infection and injury, we conducted a postdeployment survey and compared these data to clinic-based syndromic surveillance data. Results. We enrolled 43 patients with acute diarrhea, 21 (49%) having one or more pathogens isolated. Enterotoxigenic Escherichia coli (n = 16), enteroaggregative E coli (n = 3), and Shigella spp. (n = 3) were the most common pathogens identified. Respiratory illness had the highest incidence (73 episodes/100 person-months) compared to diarrhea (35 episodes/100 person-months) and noncombat injury (17 episodes/100 person-months), though noncombat injury more frequently resulted in lost duty days and health-care utilization. Conclusions. Noncombat injuries and illnesses have had a significant impact on military missions and continue to result in force health protection challenges today. Future studies are needed to test and evaluate countermeasures to mitigate these illnesses and injuries to increase the health of the individuals and optimize mission readiness.
H
istorically and in current deployments, disease and nonbattle injury (DNBI) have accounted for a significant amount of morbidity during combat and noncombat operations among US military personnel.1–6 The transformation of the US military to smaller and more mobile units requires greater emphasis on the maintenance and preservation of the individual soldier’s health to be mission ready and capable.7–9 The Department of Defense (DoD) is committed to reducing the incidence of infectious
Corresponding Author: Mark S. Riddle, MD, Enteric Diseases Department, Naval Medical Research Center, Silver Spring, MD 20910, USA. E-mail: markriddlemd@ hotmail.com
diseases and injury in deployed troops. Current strategies focus on provision of safe food and water, education about hygiene, appropriate waste disposal, and injury prevention programs. However, despite these efforts, considerable morbidity from DNBI remains.1 The dynamics of global military deployments must be considered when formulating intervention strategies. A major component of this effort is to perform large-scale epidemiology studies among deployed personnel to assess current strategies, to fully describe the incidence and impact of these illnesses and injuries, and to evaluate how well theater clinic-based syndromic surveillance systems perform to assess DNBI events. A recent military exercise conducted in the Middle East provided such an
© 2007 International Society of Travel Medicine, 1195-1982 Journal of Travel Medicine, Volume 14, Issue 6, 2007, 392–401
Operation Bright Star in Northern Egypt, 2005 opportunity, with a special emphasis on the etiologies of infectious diarrhea. Materials and Methods Study Population, Design, and Data Collection From September through October 2005, approximately 7,500 US military personnel participated in Operation Bright Star, a multinational joint military exercise in the northwestern Egyptian desert. Participating military personnel comprised reserve Army and active duty Marine Corps service members on a 2-week temporary training assignment involving various field exercises in remote sites throughout the area. During this exercise, we conducted a descriptive case series study among a catchment population that was estimated between 2,000 and 4,000 US military personnel who sought care at one of two centrally located field sites (E. P. H.). There were several other medical treatment facilities where patients with diarrhea could have sought treatment, but our research staff were limited to these two because of logistical constraints. Because of the uncertainty in the catchment population and an interest in estimating disease incidence, a postexercise convenience sample survey among departing military personnel was conducted. US military personnel with diarrhea who presented to one of two medical treatment facilities were asked to participate in the study. Diarrhea was defined as three or more loose or liquid stools in a 24-hour period or two or more loose or liquid stools in 24 hours associated with fever or other gastrointestinal (GI) symptoms (nausea, vomiting, cramps, tenesmus, or blood in stool). Volunteers could only be included once in the study and were excluded from participation if they reported antibiotic use within the preceding 24 hours. An anonymous postexercise survey was pretested and administered to US military personnel awaiting air transportation for departure from the country. Based on a convenience sample, they were asked to complete a series of questions related to their diarrhea, respiratory illness, and or injury experience during the entire exercise period. DNBI surveillance is a mandated surveillance system that reports aggregate medical event data categorized into 15 different syndromes (including categories of infectious GI illness, acute respiratory illness, and nonbattle injury) on a weekly basis and reported to a central military public health activity.10 These data, which include aggregate number of initial visits for each syndrome as well as a population denominator, were collected from five clinic sites operating in
393 the exercise and reported using paper-based or electronic templates to the central medical command between the period from September 4 to 27, 2005. These data were obtained and compared to the results for categories of GI infection, acute respiratory illness, and nonbattle injury. Laboratory Testing Field laboratory personnel inoculated stool into Cary–Blair (CB) and buffered glycerol saline transport media and placed in a 4°C container on site. Stool was separated into five 20 cc aliquots: three aliquots were stored at −70°C and two in 10% formalin. Two to three times per week, all specimens were transported to the microbiology laboratory at Naval Medical Research Unit No. 3, Cairo (NAMRU-3), where the stool and serum specimens were processed for testing. Standard laboratory procedures were employed for the isolation and identification of the major enteric bacterial species (Campylobacter, Shigella, Salmonella, and Vibrio). Enterobacteriaceae identification used a biochemical identification scheme (API 20E system; Analytab Products, NY, USA). Commercially available antisera reagents (Difco Laboratories, Detroit, MI, USA) were used to serotype all recovered Shigella isolates. Stool specimen from CB swabs were plated on a modified Skirrow’s medium (Hardy Diagnostics, Santa Maria, CA, USA) and incubated for 48 hours at 42°C in a microaerophilic environment.11,12 Suspicious colonies were further evaluated by Gram’s stain and by oxidase and catalase activity. Hippurate hydrolysis was used to differentiate the Campylobacter isolates into Campylobacter jejuni (hippurate positive) and non-jejuni Campylobacter spp. (hippurate negative).13 MacConkey agar plates were scanned for the presence of lactose-fermenting colonies with colonial morphology similar to that of Escherichia coli. Five isolated colonies similar to E coli were picked from each culture plate and assayed for the synthesis of toxins [heat labile (LT) and heat stable (ST)] using two different enzyme-linked immunosorbent assay formats.14,15 Escherichia coli isolates producing toxins [enterotoxigenic E coli (ETEC)] were further characterized for the presence of surface colonization factors [CFA/I, CS1–CS7, CS8 (CFA/III), CS12 (PCFO159), CS14 (PCFO166), and CS17)] by specific monoclonal antibodies using a dot-blot assay.16,17 DNA templates were prepared from fecal specimens. Enteroaggregative E coli (EAEC) was identified through polymerase chain reaction using the primer set pCVD432, 5′-CTG GCG AAA GAC TGT ATC AT-3′ and 5′-CAA TGT ATA GAA ATC CGC TGT T-3′.18 Thermal cycler (MJ PTC-100) J Travel Med 2007; 14: 392–401
394 was adjusted for the following assay conditions: 94°C (20 s), 59°C (45 s), 72°C (45 s) for 30 cycles, followed by a final extension period at 72°C for 10 minutes. The amplicon (630 bp) was visualized in a 2% agarose gel stained with ethidium bromide. As per manufacturer’s instructions, a standard qualitative antigen detection enzyme immunoassay kit (Techlab, Inc., Blacksburg, VA, USA) was used for the identification of Cryptosporidium parvum from frozen stool for each case. Ova and protozoa were detected microscopically from fecal samples preserved in 10% formalin. Rotavirus (antigen detection) was detected in fecal specimens following manufacturer’s instruction using a commercially available enzyme immunoassay, Rotaclone (Meridian Diagnostics, Inc., Cincinnati, OH, USA). Statistical Analysis Survey data were entered into EpiInfo 6.0 (CDC, Atlanta, GA, USA) using a valid double-data entry method. All statistical analyses were conducted by Stata Version 9 (Stata Corp., College Station, TX, USA). Mantel–Haenszel chi-square or Fisher’s exact tests were used to determine statistical significance among discrete or categorical variables. The t-test (two groups) or analysis of variance (more than two groups) was used for all parametric testing of continuous variables. For nonparametric testing among continuous variables, Kruskal–Wallis test was used. Incidence rates were estimated using Poisson distributions (assumptions met) and evaluated for associations with covariates (demographic, risk behaviors) using Poisson regression. A multivariate Poisson regression model was used to adjust for potential confounding between significant covariates found in the univariate analyses. Statistical significance was set at p < 0.05 for all analyses. This study was approved by the Scientific Review Board protocol and the Institutional Review Board (DoD# NAMRU-3.2005.0012) of the US Naval Medical Research Unit No. 3, Cairo, Egypt. Results Diarrhea/Vomiting Case Surveillance Forty-three patients were enrolled during the surveillance period. The participants were mostly male (81%), with a median age of 30 years [interquartile range (IQR) 24–42 y)] (Table 1). Enlisted personnel made up more than two thirds of the sample (67.5%). Clinical Presentation Of the 43 cases, the median number of loose or liquid stools in the 24 hours preceding the initial clinic J Travel Med 2007; 14: 392–401
Riddle et al. Table 1 Demographic characteristics, clinical symptoms, treatment, and impact associated with acute diarrhea among study participants (N = 43) during Bright Star 2005, Egypt Demographic characteristics Age in years, median (IQR) Male gender, n (%) Rank, n (%) Junior enlisted Senior enlisted Officers Civilians Clinical symptoms on presentation Maximum number of stools per 24 h, median (IQR) Number of loose stools since beginning of episode, median (IQR) Abdominal cramps, n (%) Bloody diarrhea, n (%) Nausea, n (%) Vomiting, n (%) Subjective fever, n (%) Headaches, n (%) Muscle aches, n (%) Joint pain, n (%) Treatment Took any medicines for diarrhea, n (%) Bismuth subsalicylate Loperamide Antibiotic Other Ability to work, n (%) Normal Decreased Not able to work
30 (24–42) 34 (81.0) 15 (34.9) 14 (32.6) 12 (27.9) 2 (4.6) 5 (4–7) 11 (6–20) 38 (88.4) 3 (7.1) 26 (60.5) 7 (16.3) 12 (28.6) 17 (39.5) 15 (34.9) 6 (14.3) 16 (38.1) 5 (31.3) 11 (68.8) 0 3 (18.8) 18 (41.9) 18 (41.9) 7 (16.2)
IQR = interquartile range.
visit was 5, with an IQR of 4 to 7 stools. The median number of loose or liquid stools reported during the entire duration of the current diarrheal episode was 11 (IQR 6–20). The median duration of diarrhea from the start of symptoms to presentation was 2 days (IQR 1–4 d). The most commonly reported associated symptom was abdominal cramping (88%), followed by nausea (61%) and headache (40%) (Table 1). Among the 16 patients (38%) with selftreatment prior to clinic presentation, loperamide was the most common choice (69%). At the initial clinic visit, 16% reported that they were “not able to work,” while an additional 42% reported a decreased ability to work. Pathogens Twenty-one (49%) patients with diarrhea had one or more pathogens identified (Table 2). In 17 (40%) patients, a single pathogen was recovered, while 4 (9%) had two or more pathogens recovered.
395
Operation Bright Star in Northern Egypt, 2005 Table 2 Pathogen distribution among all diarrhea cases (N = 43) collected during Operation Bright Star Cases, n (%) Pathogen type Bacteria ETEC EAEC Campylobacter spp. Shigella spp. Parasite Cryptosporidium parvum Giardia lamblia Virus Rotavirus Negative
Solo
Mixed
12 (27.9) 1 (2.3) 0 3 (7.0)
4 (9.3) 2 (4.7) 0 0
0 0 1 (2.3) 22 (51.2)
1 (2.3) 1 (2.3) 0 —
EAEC = enteroaggregative Escherichia coli; ETEC = enterotoxigenic E coli.
Bacteria were isolated in 47% of the cases. ETEC was the most common bacterial pathogen, identified from 16 patients (12 sole pathogen and 4 mixed), an overall ETEC recovery rate of 37%. ETEC toxin phenotypes identified from this study were 81% ST (13 of 16), 13% LT (n = 2), and a single LTST. A colonization factor was identified in 81% of the ETEC isolated, with CS6 being the most common (n = 9), followed by CS3 (n = 3) and a single CS4 + CS6 isolate. Shigella spp. was the second most common pathogen isolated (n = 3). A total of three cases of EAEC (one sole pathogen and two mixed) were identified. Postexercise Survey A total of 1,458 US military personnel completed a postexercise survey (approximately 20% of deployed forces). Survey respondents had a median age of 32 years (IQR 25–41 y), 79% were male, and all but three were Army (one Navy and two Air Force). Of those reporting their rank, 31% were junior enlisted (E1–E4), 47% senior enlisted (E5– E9), and 22% officers. Median exercise length was 17 days (IQR 14–25 d). Diarrhea/Vomiting Illness Among survey respondents, 24% reported at least one episode of diarrhea at some point during the military exercise and 4% reported a vomiting episode (not mutually exclusive) (Table 3). Of those who experienced diarrhea, only 20% (n = 74) sought medical attention related to diarrhea and 23% of the respondents reported having to stop or significantly decrease work for at least 1 day. We are unable to determine how many of the 74 who sought medical attention were enrolled in our case series, if
Table 3 Descriptive features of postexercise survey evaluating diarrhea/vomiting, respiratory illness, and injury (total surveys analyzed = 1,458) n/N (%) Diarrhea/vomiting illness Diarrhea while deployed Vomiting while deployed Fever with diarrhea No. days with diarrhea/vomiting 5 Visit sick call for diarrhea/vomiting Decrease work performance due to diarrhea/vomiting Members of unit sick with diarrhea/vomiting Respiratory illness Respiratory illness during exercise Fever, muscle aches, or headache associated with illness No. days with respiratory illness 5 Visit sick call for respiratory illness Decrease work performance due to respiratory illness Members of unit sick with respiratory illness Injury Injury while deployed Type of injury Fracture Sprain/strain Burn injury Laceration/open wound Bruise/contusion Other Mechanism of injury Vehicle accident Fall Sports/athletics Machinery/tools Heavy gear/lifting Other No. days with injury 5 Visit sick call for injury Decrease work performance due to injury
343/1,451 (24) 58/1,456 (4) 61/341 (18) 154/330 (47) 126/330 (384) 50/330 (15) 74/361 (20) 82/361 (23) 570/1,394 (41)
683/1,454 (47) 311/681 (46)
168/619 (26) 286/619 (45) 184/619 (29) 162/681 (24) 120/682 (18) 721/1,407 (51)
163/1,454 (11) 2/153 (1) 53/153 (35) 6/153 (4) 17/153 (11) 36/153 (24) 39/153 (25) 3/150 (2) 24/150 (16) 19/150 (13) 15/150 (10) 33/150 (22) 56/150 (37) 40/151 (27) 46/151 (30) 65/151 (43) 95/162 (59) 55/161 (34)
any. While we enrolled 43 subjects who sought medical care for their diarrhea, our research team was only able to staff two military treatment facilities out of several which were set up and able to receive patients during the exercise. Nearly two thirds (66%) of troops reported traveling off base at some time during the exercise and over one third J Travel Med 2007; 14: 392–401
396
Riddle et al.
(34%) reported eating food from non-US military sources. Based on information of diarrheal disease episodes, clinic visit rate, and person-time of deployment among those who did not have missing data (n = 1,445), we calculated an overall diarrhea illness incidence of 35.1 episodes/100 person-months (Poisson 95% CI 31.4–38.9) and a diarrheal incidence where treatment was sought of 7.6 clinic visits/100 person-months (Poisson 95% CI 5.9–9.5). Table 4 describes the Poisson univariate analysis for factors associated with overall and clinic visit incidence of diarrhea. Officer rank [incident rate ratio (IRR) 1.44, 95% CI 1.08–1.92) was associated with increased risk, and a significant trend across ranks was found (p = 0.03). Reporting other members in unit sick with diarrhea was highly associated with increased risk of any diarrhea (IRR 5.79, p
