Travelers’ Diarrhea in Thailand: A Quantitative Analysis Using TaqMan® Array Card

The TaqMan® array card (TAC) was used to detect enteropathogens in stool samples from travelers’ diarrhea cases and asymptomatic controls in a hospital-based surveillance in Thailand. Analyses were performed to determine pathogen–disease association to compare TAC and conventional results.

Abstract

Background

Travelers’ diarrhea (TD) is a common illness experienced by travelers from developed countries who visit developing countries. Recent questionnaire-based surveillance studies showed that approximately 6%–16% of travelers experienced TD while visiting Thailand; however, a majority of TD information was limited mainly to US military populations.

Methods

A TD surveillance study was conducted at Bumrungrad International Hospital in 2012–2014 in Bangkok, Thailand. Enteropathogens were identified using conventional methods and the TaqMan® array card (TAC), which uses real-time polymerase chain reaction for the simultaneous detection of multiple pathogens. Analyses to determine pathogen–disease and symptoms association were performed to elucidate the clinical relevance of each enteropathogen.

Results

TAC identified more pathogens per sample than conventional methods. Campylobacter spp. were the most prevalent, followed by the diarrheagenic Escherichia coli and norovirus GII. These agents had significant pathogen–disease associations as well as high attributable fractions among diarrheal cases. A wide range of pathogen loads for Campylobacter spp. was associated with TD, while heat-labile toxin enterotoxigenic Escherichia coli was associated with an increased pathogen load. Most cases were associated with inflammatory diarrhea, while Campylobacter spp. and Shigella spp. were associated with dysentery.

Conclusions

A pan-molecular diagnostic method such as TAC produces quantifiable and comparable results of all tested pathogens, thereby reducing the variability associated with multiple conventional methods. This allows better determination of the clinical relevance of each diarrhea etiologic agent, as well as their geographical relevance in Thailand.

Travelers’ diarrhea (TD) is one of the most common travel-related illnesses experienced by travelers from developed countries who visit developing countries [1]. Destination plays a major role in developing TD. Africa and South Asia are high-risk regions [1–3], and South America, Southeast Asia, and East Asia have reduced risks due to improved sanitation [3]. However, additional elements contribute to TD, including environmental and host factors [3].

Thailand, located in Southeast Asia, was the top regional tourist destination in 2015 [4]. In recent questionnaire-based surveillance studies, approximately 6%–16% of foreign visitors reported experiencing TD while visiting Thailand [5–7]. Information on the TD etiologic agents was limited to regional- or geographical-based reports [8, 9]; however, specific reports from Thailand were focused on Campylobacter spp. [10, 11]. To account for this limited data, a hospital-based TD case-control surveillance study of travelers to Thailand was conducted in Bangkok from 2012 through 2014. Stool samples were collected for routine laboratory diagnostics to identify potential enteric pathogens.

Conventional diagnostic approaches require multiple sample assessments for all suspected agents. This introduces variability between methods, is limited to the pathogens assessed, is labor intensive, and requires excess time and excess samples. Numerous studies have demonstrated that molecular diagnostic methods to determine diarrhea etiologic agents are sensitive, enable quantification, and estimate pathogen copy numbers required to better ascribe clinical relevance [12–15]. Here, stool samples were retrospectively tested with an enteric pathogen panel TaqMan® array card (TAC), which is a customizable microfluidic card for real-time polymerase chain reaction (PCR) [12, 16–18].

TD etiology results generated from both conventional diagnostic methods and TAC were compared, and TAC was assessed as a potential diagnostic tool for the epidemiological study of TD. Additional evaluation and quantitative analyses were performed to determine pathogen–disease associations applicable to TD in Thailand.

METHODS

Study Enrollment

Foreign travelers who visited Thailand were enrolled in a TD surveillance study from January 2012 through December 2014. Travelers were limited to citizens who originated from North America, Europe, Australia, New Zealand, Japan, Taiwan, and South Korea, were aged ≥18 years, and solicited the inpatient department or outpatient department services of Bumrungrad International Hospital located in Bangkok, Thailand. Diarrhea cases were defined as having 3 or more unformed stools per 24 hours; having 1 additional clinical symptom such as abdominal cramps, fecal urgency, nausea, vomiting, tenesmus, and fever; and waiting 7 days or less before soliciting treatment at the hospital. Controls were participants who visited the hospital for routine health examinations and had no history of diarrhea 2 weeks prior. Controls were matched to cases by age group, for example, 18–29, 30–39, 40–49, etc. Individuals who had resided outside of their home countries for more than 1 year were not eligible for the study. After obtaining written informed consent, a stool sample was provided by the study volunteer. Study participants were asked to complete a questionnaire that captured demographic, clinical, and medical history data. No follow-up was scheduled. The study was approved by the institutional review boards for both Bumrungrad International Hospital in Thailand and the Walter Reed Army Institute of Research, Silver Spring, Maryland.

Conventional Methodology

Stool samples were initially examined for fecal red blood cells (RBCs), white blood cells (WBCs), and ova and parasites by direct microscopic examination. Stool samples were then resuspended and inoculated onto the following selective media and enrichment broth: MacConkey, Hektoen, thiosulfate citrate bile salts sucrose, modified semisolid Rappaport Vassiliadis, modified charcoal cefoperazone deoxycholate agar, buffered peptone water, alkali peptone water, and Preston selective enrichment broth in order to culture Aeromonas, Arcobacter, Campylobacter, Escherichia coli, Plesiomonas, Salmonella, Shigella, and Vibrio spp. as previously described [19]. Subsequently, isolated colonies were identified by standard biochemical and serotype testing. Five lactose-fermenting and 5 nonlactose-fermenting E. coli were identified and selected for identification of the diarrheagenic E. coli by multiplex PCR assays as previously described [20–26]. Norovirus and sapovirus were detected by real-time reverse transcription (RT) PCR [27, 28]. Commercial enzyme-linked immunosorbent assay kits were used to detect parasites (Cryptosporidium, Entamoeba histolytica, and Giardia; TECHLAB, Blacksburg, Virginia) and enteric viruses (adenovirus, astrovirus, and rotavirus; RIDASCREENR, R-Biopharm AG, Darmstadt, Germany) following the manufacturers’ instructions.

Testing With TaqMan® Array Card

Total nucleic acid was extracted using the QiaAmp Fast Stool DNA kit via the QIAcube extraction platform (Qiagen, Valencia, California) [29]. The testing using TAC was performed as previously described [18, 29]. Briefly, 40 µL of extracted nucleic acid from each sample was mixed with 60 µL of reagents from Ag-Path-ID One-Step RT-PCR kit (Applied Biosystems, Foster City, California). This was loaded into each of the 8 ports of the TAC. The card was then sealed and loaded into the ViiA7 instrument (Applied Biosystems) [29]. Each TAC card contained a previously developed enteric pathogen panel [18, 30] that included 12 bacteria (Aeromonas, Bacteroides fragilis, Campylobacter spp. [C. coli and C. jejuni], Clostridium difficile, enteroaggregative Escherichia coli [EAEC], enteropathogenic E. coli [EPEC], enterotoxigenic E. coli [ETEC; heat-labile toxin (LT) and heat-stable toxin (ST)], Helicobacter pylori, Salmonella, Shigella/enteroinvasive E. coli [EIEC], Shiga toxin–producing E. coli [STEC], and Vibrio cholerae); 2 fungi (Encephalitozoon intestinalis and Enterocytozoon bieneusi); 5 nematodes (Ancylostoma duodenale, Ascaris lumbricoides, Necator americanus, Strongyloides stercoralis, and Trichuris trichiuria); 5 protozoan parasites (Cryptosporidium, Cyclospora cayetanensis, E. histolytica, Giardia lamblia, and Cystoisospora belli); and 5 viruses (adenovirus, astrovirus, norovirus GI/GII, rotavirus, and sapovirus). Raw data files were processed using ViiA7 software, version 1.2.2 (Applied Biosystems), and analysis was performed as previously described [17]. An analytical cutoff (lower limit of detection) was applied at a threshold cycle (Ct) >35.

Statistical Analyses

Binary logistic regression and χ² tests were performed to calculate odds ratios (ORs), and the disease–pathogen load association was calculated using binary logistic regression with IBM SPSS Statistics, version 23. Analyses of symptoms and pathogen association included case data only. The attributable fraction (AF) was calculated using the following equation: prevalence × [1 − (1/OR)], and the 95% confidence interval (CI) was estimated using the “attribrisk” package in R. The disease–pathogen load association was performed by stratifying Ct values of each pathogen into the following groups: Ct <20, Ct = 20–25, Ct = 25–30, Ct = 30–35, and all strata were compared against Ct >35 (negative) as previously described [17]. Analyses were performed on all pathogens obtained from available positive cases and controls. Calculated P values <.05 were considered statistically significant.

RESULTS

A total of 338 participants (173 cases and 165 controls) were enrolled; 316 samples (154 cases and 162 controls) were available for testing and analysis by both conventional methods and TAC. The majority of cases were from Europe, and travelers from North America constituted most of the controls (Table 1). The average age was 36.8 years for cases and 37.9 years for controls. More than 80% of the travelers had been in Thailand for less than 30 days at the time of enrollment. Clinical information and stool characteristics are summarized in Table 1. Loose or watery stool was the predominant characteristic among cases, and soft stool was common among controls. All cases experienced at least 1 clinical symptom, with abdominal pain and/or fatigue reported as the primary additional complaint, and 23.4% experienced all clinical symptoms listed (Table 1). Approximately 51% (78/154) of cases administered some form of self-treatment prior to seeking care at the hospital. Of those, 63% (49/78) took antibiotics; however, the percentage difference between antibiotic use and subsequent overall recovery of pathogens was not significantly different (data not shown) from that for cases who did not use antibiotics prior to seeking medical treatment (Table 1).

Table 1.

Demographic and Clinical Information of Cases and Controls of Study Participants in Travelers’ Diarrhea Surveillance Study Who Visited Bumrungrad International Hospital in Bangkok, Thailand, from January 2011 through December 2014

Demographic and Clinical Information	Case, n (%) (N = 154)	Control, n (%) (N = 162)
Gender
Female	55 (35.7)	54 (33.3)
Male	99 (64.3)	108 (66.7)
Origin
North America
Canada	5 (3.3)	5 (3.1)
United States	23 (14.9)	69 (42.6)
Australia/Oceania
Australia	25 (16.2)	19 (11.7)
New Zealand	1 (0.7)	1 (0.6)
East Asia (Japan)	35 (22.7)	1 (0.6)
Europe
Eastern	0	1 (0.6)
Northern	11 (7.1)	7 (4.3)
Southern	1 (0.7)	6 (3.7)
Western	53 (34.4)	53 (32.7)
Age, years
18–29	62 (40.3)	64 (39.5)
30–39	44 (28.6)	44 (27.2)
40–49	12 (7.8)	16 (9.9)
50–59	22 (14.3)	23 (14.2)
60–69	12 (7.8)	14 (8.6)
70–79	2 (1.3)	1 (0.6)
Duration of stay, months
<1	120 (77.9)	136 (84)
1–3	27 (17.5)	13 (8)
>3–6	4 (2.6)	3 (1.9)
6–11	3 (1.9)	10 (6.2)
Symptoms
Abdominal pain	133 (86.4)	NA
Fatigue	134 (87)	NA
Fever	92 (59.7)	NA
Nausea	102 (66.2)	NA
Vomiting	78 (50.6)	NA
Stool characteristics
Formation
Loose	70 (45.5)	0
Watery	74 (48.1)	1 (0.6)
Soft	10 (6.5)	158 (97.5)
Formed	0	3 (1.9)
Microscopic observation
Mucus	66 (42.9)	0
Red blood cells	50 (32.5)	1 (0.6)
White blood cells	122 (79.2)	0
Frequency (per 24 hours)
<3	49 (31.8)	NA
3–4	31 (20.1)	NA
5–6	50 (32.5)	NA
7–8	12 (7.8)	NA
9–10	11 (7.1)	NA
>10	1 (0.7)	NA
Duration of symptoms
≤24 hours	76 (49.4)	NA
2 days	29 (18.8)	NA
3 days	33 (21.4)	NA
4 days	6 (3.9)	NA
5 days	3 (1.9)	NA
6 days	6 (3.9)	NA
7 days	1 (0.7)	NA
Prior self-treatmenta
Antibiotics
Cephalosporins	7 (4.5)	...
Fluoroquinolones	43 (27.9)	...
Macrolides	9 (5.8)	...
Nitroimidazoles	2 (1.3)	...
Tetracycline	1 (0.7)	1 (0.6)
Other	2 (1.3)	...
Oral rehydration salt	1 (0.7)	...
Antimotility	9 (5.8)	...
Absorbent	6 (3.9)	...
Symptoms specific	12 (7.8)	...
Unknown treatment	15 (9.7)	...
Nondiarrhea-related treatment	3 (1.9)	…

Abbreviation: NA, not applicable.

^aPatients treated their symptoms with more than 1 type of medication.

Comparative analyses on the diagnostic results were conducted for pathogens identified by both conventional and TAC methods. The overall detection accuracy by TAC was 95%. The average number of pathogens detected by conventional methods and TAC was 1.4 vs 2.0 for cases and 0.3 vs 0.7 for controls, respectively (Table 2). A total of 143 samples (45.3%) were pathogen negative (25 cases and 118 controls) by routine conventional testing (Table 2). TAC testing resulted in an additional 35 pathogen-positive samples, while 108 samples (16 cases and 92 controls) remained pathogen negative.

Table 2.

Number of Samples Categorized by Number of Pathogens Detected by Conventional Methods and TaqMan® Array Card

Number of Pathogens Identified	Conventional		TaqMan® Array Card
	Case	Control	Case	Control
0	25	118	16	92
1	72	35	47	37
2	34	8	54	20
3	15	1	18	10
4	8	0	13	2
5	0	0	4	1
6	0	0	1	0
7	0	0	1	0

Campylobacter spp. was the most common pathogen detected in cases by both conventional and TAC methods (approximately 30% each; Table 3). The diarrheagenic E. coli (EAEC, EPEC, and ETEC) and norovirus GI and GII were detected more by TAC than by conventional methods for both cases and controls. The comparison of ORs for each pathogen by both methods revealed that Campylobacter spp., Salmonella, and norovirus GI and GII were significantly associated with TD (Table 3). EAEC by TAC and Plesiomonas by conventional methods were also associated with TD (Table 3). EPEC was coidentified in combination with other pathogens including EAEC, Campylobacter spp., and norovirus GI and GII in cases. The calculated ORs of these combinations were similar to the calculated ORs of each pathogen, indicating that EPEC did not significantly add to the association of each pathogen to TD. The AF is a metric that combines prevalence and OR to indicate a proportion of cases that can be attributed to a particular pathogen. Most of the TD cases were attributable to Campylobacter spp. (28%) and norovirus GII (23%; Table 3).

Table 3.

Summary of Detection Percentages in Travelers’ Diarrhea Case and Diarrhea-Free Control Samples, Odds Ratio, and Attributable Fraction of Each Pathogen by Conventional Methods and TaqMan® Array Card

Conventional Methods					TaqMan® Array Card
Pathogen	Case (%) n = 154	Control (%) n = 162	OR (95% CI)	AF (95% CI)	Pathogen	Case (%) n = 154	Control (%) n = 162	OR (95% CI)	AF (95% CI)
Bacteria
Aeromonas	11.7	4.3	2.9 (1.2–7.2)	7.7 (1.5–14.8)	Aeromonas	8.4	3.7	2.4 (0.9–6.5)	4.9 (0–10.8)
Arcobacter a	0.7	0.6	1.1 (0.7–17.0)	0	Arcobacter a	...	...	...	...
Bacteroides fragilis a	...	...	...	...	Bacteroides fragilis a	7.1	9.3	0.8 (0.3–1.7)	0
Campylobacter spp.	31.2	4.3	10.1 (4.4–23.0)	28.1 (19.7–35.6)	Campylobacter spp.	31.2	4.9	8.7 (4.0–19.2)	27.6 (20.2–34.6)
Clostridium difficile a	...	...	...	...	Clostridium difficile a	1.3	1.2	1.1 (0.1–7.6)	0.1 (0–2.64)
EAEC	9.7	11.1	0.9 (0.4–1.8)	0	EAEC	27.3	17.3	1.8 (1.1–3.1)	12.1 (1.6–22.4)
EPEC	5.8	3.7	1.6 (0.6–4.7)	2.2 (0–7.1)	EPEC	29.9	21.6	1.6 (0.9–2.6)	10.5 (0–21.6)
LT-ETEC	2.6	1.2	2.1 (0.4–11.8)	1.4 (0–4.9)	LT-ETEC	5.2	2.5	2.2 (0.6–7.3)	2.8 (0–0.1)
ST-ETEC	2.6	0	... b	2.6	ST-ETEC	5.2	0.6	8.8 (1.1–71.4)	4.6 (1.2–9.4)
LT/ST-ETEC	1.3	0	...	1.3	LT/ST-ETEC	7.8	0	... b	7.8
Helicobacter pylori a	...		...		Helicobacter pylori a	2.6	0	...	2.6
Plesiomonas a	18.8	3.7	6.0 (2.4–15.0)	15.7 (8.5–22.7)	Plesiomonas a	...	...	...	...
Salmonella	19.5	4.3	5.4 (2.3–12.6)	15.8 (8.4–23.3)	Salmonella	8.4	1.2	7.4 (1.6–33.3)	7.3 (3.4–12.6)
Shigella/EIEC	3.9	0	... b	3.9	Shigella	5.8	3.7	1.6 (0.6–4.6)	2.2 (0–6.81)
STEC (stx1)	0	0	...	0	STEC (stx1)	2.0	1.9	1.1 (0.2–5.3)	0.1 (0–3.63)
Vibrio spp.	7.1	0	... b	7.1	Vibrio cholerae	1.3	0	...	1.3
Protozoan parasite
Cryptosporidium	0.7	0	...	0.7	Cryptosporidium	1.3	0	...	1.3
Cyclospora	0	0	...	0	Cyclospora	1.3	0	...	1.3
Entamoeba histolytica	0	0	...	0	Entamoeba histolytica	0	0	...	0
Giardia	0.7	0	...	0.7	Giardia	0.7	1.2	0.5 (0.1–5.8)	0
Virus
Adenovirus	0	0	...	0	Adenovirus	0	0.6	...	0
Astrovirus	0	0	...	0	Astrovirus	0	0	...	0
Norovirus (GI)	2.6	0	...	2.6	Norovirus (GI)	11.0	0.6	20.0 (2.6–152.1)	10.5 (5.6–15.6)
Norovirus (GII)	15.6	0	... b	15.6	Norovirus (GII)	23.4	0.6	49.1 (6.6–393.4)	22.9 (15.9–29.9)
Rotavirus	2.6	0	...	2.6	Rotavirus	2.6	1.2	2.1 (0.4–11.8)	1.4 (0–4.7)
Sapovirus	3.3	0	... b	3.3	Sapovirus	3.9	0.6	6.5 (0.8–54.9)	3.3 (0.1–7)

Abbreviations: AF, attributable fraction; CI, confidence interval; EAEC, enteroaggregative Escherichia coli; EIEC, enteroinvasive E. coli; EPEC, enteropathogenic E. coli; ETEC, enterotoxigenic E. coli; LT, heat-labile toxin; OR, odds ratio; ST, heat-stable toxin; STEC, Shiga toxin-producing E. coli.

^aPathogen was only tested using either TaqMan® array card or conventional method, as indicated.

^bThe P value of Fisher exact test from χ² was significant when the odds ratio was not calculable with SPSS. The pathogens are listed by group in alphabetical order. Bacteria were detected by culture and biochemical methods. Diarrheagenic Escherichia coli were confirmed by polymerase chain reaction (PCR). Protozoan parasites, adenovirus, astrovirus, and rotavirus were detected by enzyme-linked immunosorbent assay. Norovirus and sapovirus were detected by reverse transcription real-time PCR.

In order to determine disease–pathogen load associations, Ct values were evaluated. Campylobacter spp. exhibited a significant disease–pathogen load on all analyzed Ct ranges, with a calculated OR of 9.2 (95% CI, 2.7–31.9), 5.8 (95% CI, 1.2–27.9), and 5.8 (95% CI, 1.6–21.1) for Ct ranges of 20–25, 25–30, and 30–35, respectively. LT-ETEC indicated a significant disease–pathogen load at Ct <20 with a reported OR of 9.5 (95% CI, 1.2–76.9), indicating that a higher pathogen load in combination with lower Ct was significantly associated with TD. Assessments of the remaining identified pathogens did not yield any significant disease–pathogen load association due to a diminished number of samples after stratification of Ct values or the lack of significant disease–pathogen load association (Figure 1).

Figure 1.

Pathogen load and disease association. Positive threshold cycle (Ct) values (<35) were stratified into 4 groups as indicated on the x-axis and analyzed with binary logistic regression using negative results (Ct >35) as the reference group for each pathogen. *A statistically significant odds ratio at respective Ct cycles. Abbreviations: B. fragilis, Bacteroides fragilis; Ct, threshold cycle; EAEC, enteroaggregative Escherichia coli; EIEC, enteroinvasive E. coli; EPEC, enteropathogenic E. coli; ETEC, enterotoxigenic E. coli; LT, heat-labile toxin; ST, heat-stable toxin; STEC, Shiga toxin-producing E. coli.

Using data obtained from both TAC and conventional methods to determine the correlation between pathogen and clinical symptoms demonstrated that among cases, norovirus GII was the only pathogen that was associated with vomiting, and Campylobacter spp. was associated with fever (Figure 2). Approximately half of the cases had watery stool, which had a significant association with infection with Campylobacter spp., Salmonella, and norovirus GII (Table 1 and Figure 2). These pathogens, along with other bacteria such as Plesiomonas, were also associated with the presence of mucus and WBCs in the stool, which are indicative of inflammatory diarrhea. In contrast, dysentery, as defined as a patient having a fever and the presence of RBCs in stool, was associated with Campylobacter spp. and Shigella spp. (Figure 2).

Figure 2.

Associations between clinical symptoms, stool characteristics, and pathogen, as determined by conventional and TaqMan® array card methods. *A statistically significant odds ratio of each pathogen with respect to clinical symptoms and stool characteristics. Abbreviations: Conv., conventional; EAEC, enteroaggregative Escherichia coli; EIEC, enteroinvasive E. coli; EPEC, enteropathogenic E. coli; ETEC, enterotoxigenic E. coli; LT, heat-labile toxin; RBC, red blood cell; ST, heat-stable toxin; STEC, Shiga toxin-producing E. coli; TAC, TaqMan® array card; WBC, white blood cell.

DISCUSSION

This is the first report of the use of the TAC enteric pathogen panel to detect diarrhea etiologic agents from an adult TD population in Bangkok, Thailand. Campylobacter spp. was the most prevalent pathogen detected by both TAC and conventional methods, which aligns with previous reports that Campylobacter spp. were the most prevalent cause of TD in Southeast Asia, particularly in Thailand [9, 31, 32]. ETEC, the most common cause of bacterial TD worldwide, was less prevalent than has been seen in studies from Africa, Latin America, and the Caribbean [2, 3, 33] but was still common. In this study, norovirus was one of the top three pathogens detected using both TAC and conventional methods, highlighting its importance related to TD in Thailand. Norovirus has been underreported in TD studies possibly because most TD studies do not include norovirus diagnostic testing [34].

The clinical features associated with TD in this population were mainly watery stool, as well as the presence of mucus and WBCs in stool, which are characteristics of inflammatory diarrhea. Plesiomonas was also significantly associated with TD with concurrent inflammatory diarrhea, as observed in previous reports [35]. Campylobacter spp. and Shigella were also associated with dysentery, which is consistent with their reported invasive nature [36]. Likewise, the association of noroviruses with vomiting aligns with their reported clinical features [37].

The difference between TAC and conventional methods resulted in different pathogen detection rates. Specifically, the culture-based method is restricted by the enrichment and selection processes; the molecular method focuses on a specific target gene, which limits the number of species being detected. The difference between these two methodologies was apparent with the detection of Salmonella spp. and Vibrio spp. by culture; the detection by TAC was species specific, which could affect the overall detection sensitivity (Table 3). Likewise, the target genes for EAEC and EPEC were similar for conventional PCR and TAC [18, 23, 24]. However, colony selection for downstream molecular characterization of EAEC and EPEC was limited by the initial colony selection with the conventional method. Higher rates of EAEC and EPEC were previously reported with TAC in samples from children in developing countries [17]. Norovirus GI and GII were detected by real-time PCR but with slightly different primers and probes used by both methods [18, 27] and specimens were tested at different times. The pan-molecular diagnostic method by TAC subjects all present pathogens to the same methodology at the same time, resulting in equal analysis and quantitation for each pathogen. While increased TAC detection led to a simultaneous detection of multiple pathogens, clinically relevant detection could be distinguished between high and low quantities, at least for Campylobacter spp. and ETEC (Figure 1) [12].

The increased detection sensitivity also indicated that stool from travelers from developed countries might not be as pathogen free as previously expected. However, a traveler’s susceptibility to acquire pathogens and develop TD increase while traveling, depending on their general behavior, adaptive immunity, gut microbiome characteristics, and their genetic composition to include different histo-binding group antigens targeted by different norovirus strains [33, 38, 39]. However, for cases to remain negative even after being tested with a highly sensitive method such as TAC highlights the importance of novel causes of TD, including toxins and emerging pathogens.

Study limitations included the small population sizes of cases and controls and inclusion of study participants who experienced TD for up to 7 days and who took antibiotics prior to the hospital visit. The small population size did not allow a complete quantitative analysis of the data. Antibiotic use did not demonstrate a significant association with the overall pathogen recovery in this study, but it may enrich for viruses and/or fluoroquinolone-resistant Campylobacter spp. to predominate over other bacteria due to the high usage of fluoroquinolones for the treatment of TD [40]. The combination of these factors may bias the diagnostic results, especially with the culture method, and quantitative analyses to establish associations between disease and symptoms for each pathogen. However, comparative analyses demonstrated that TAC produced similar diagnostic results to conventional methods with noted difference in the diarrheagenic E. coli.

The TAC method was an effective tool for detecting TD etiologic agents, with reduced diagnostic variability, and allowed results to be analyzed quantitatively. However, this does not diminish the importance of conventional methods in studying TD, as culture is still needed to better understand evolution and pathogenicity of bacteria, as well as to determine antibiotic resistance profiles. An expanded TD surveillance study in Thailand and other countries using TAC will provide more information and a better understanding for a cross-comparison between different TD populations from different geographical locations.

Notes

Acknowledgments. The authors acknowledge the contribution and support from the study participants, enrolling nurses, the biostatistician, and laboratory technicians who performed all of the testing at the Department of Enteric Diseases, Armed Forces Research Institute of Medical Sciences.

Disclaimer. The views expressed in this article are those of the authors and do not reflect the official policy of the Department of the Army, Department of Defense, or the US government. Trade names are used for identification purposes only and do not imply endorsement. The investigators have adhered to the policies for protection of human participants as prescribed in AR 70-25.

Funding. This work was supported by the US Military Infectious Diseases Research Program (MIDRP) (Frederick, Maryland).

Potential conflicts of interest. All authors: No reported conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.