Abstract
Objectives
Diarrhoea is a common and well-studied cause of illness afflicting international travellers. However, traveller's diarrhoea can also result from travel between high and low disease transmission regions within a country, which is the focus of this study.Methods
We recruited participants for a case-control study of diarrhoea at four sites along an urban-rural gradient in Northern Ecuador: Quito, Esmeraldas, Borbón and rural communities outside of Borbón. At each of these sites, approximately 100 subjects with diarrhoea (cases) were recruited from Ministry of Health clinics and were age-matched with subjects visiting the same clinics for other complaints (controls).Results
Travellers to urban destinations had higher risk of diarrhoea and diarrhoeagenic Escherichia coli (DEC) infections. Travel to Quito was associated with diarrhoea (aOR = 2.01, 95% CI = 1.10-3.68) and travel to Guayaquil (another urban centre in Ecuador) was associated with Diffuse Adherent E. coli infection (OR = 2.09, 95% CI = 1.01-4.33). Compared to those not travelling, urban origins were also associated with greater risk of diarrhoea in Esmeraldas (aOR = 2.28, 95% CI = 1.20-4.41), and with higher risk of diarrhoeagenic E. coli infections in Quito (aOR = 2.61, 95% CI = 1.16-5.86), with >50% of travel from Quito and Esmeraldas specified to another urban destination.Conclusions
This study suggests that individuals travelling from lower-transmission regions (rural areas) to higher transmission regions (urban centres) within a single country are at a greater risk of acquiring a diarrhoea-related illness. Investments to improve water, sanitation and hygiene conditions in urban areas could have impacts on outlying rural areas within a given country.Free full text
Locals get travellers’ diarrhoea too: risk factors for diarrhoeal illness and pathogenic Escherichia coli infection across an urban-rural gradient in Ecuador
Abstract
OBJECTIVES
Diarrhoea is a common and well-studied cause of illness afflicting international travellers. However, traveller’s diarrhoea can also result from travel between high and low disease transmission regions within a country, which is the focus of this study.
METHODS
We recruited participants for a case-control study of diarrhoea at four sites along an urban-rural gradient in Northern Ecuador: Quito, Esmeraldas, Borbón and rural communities outside of Borbón. At each of these sites, approximately 100 subjects with diarrhoea (cases) were recruited from Ministry of Health clinics and were age-matched with subjects visiting the same clinics for other complaints (controls).
RESULTS
Travellers to urban destinations had higher risk of diarrhoea and diarrhoeagenic Escherichia coli (DEC) infections. Travel to Quito was associated with diarrhoea (aOR = 2.01, 95% CI = 1.10–3.68) and travel to Guayaquil (another urban centre in Ecuador) was associated with Diffuse Adherent E. coli infection (OR = 2.09, 95% CI = 1.01–4.33). Compared to those not travelling, urban origins were also associated with greater risk of diarrhoea in Esmeraldas (aOR = 2.28, 95% CI = 1.20–4.41), and with higher risk of diarrhoeagenic E. coli infections in Quito (aOR = 2.61, 95% CI = 1.16–5.86), with >50% of travel from Quito and Esmeraldas specified to another urban destination.
CONCLUSIONS
This study suggests that individuals travelling from lower-transmission regions (rural areas) to higher transmission regions (urban centres) within a single country are at a greater risk of acquiring a diarrhoea-related illness. Investments to improve water, sanitation and hygiene conditions in urban areas could have impacts on outlying rural areas within a given country.
Introduction
Diarrhoeal diseases are a leading cause of global morbidity and mortality, with an estimated 1.3 million deaths annually, and up to half a million deaths each year among children under five [1-3]. Diarrhoea disproportionally affects people living in low- and middle- income countries (LMICs) [1, 2, 4-6]. In Ecuador, diarrhoeal disease rates ranged from 2.0 to 2.7 cases per 100 000 inhabitants from 2000 to 2005. While disability-adjusted life years (DALYs) have reduced by 82% since 1990 [7, 8], diarrhoeal diseases in Ecuador remain one of the top five communicable diseases causing premature death annually [7]. Repeat episodes and the most severe cases of diarrhoea are often present among children less than 5 years old and can result in cognitive and physical growth deficiencies later in life [4, 9-11]. These latent consequences of diarrhoea contribute to morbidity, therefore targeted interventions are still needed to reduce the transmission of diarrhoeal pathogens.
Diarrhoea is caused by a wide range of aetiological agents, including viruses, bacteria and parasites. Diarrhoeagenic Escherichia coli (DEC) is among the most common aetiologic agents of diarrhoea and can be subdivided into distinct pathotypes based on the presence of specific virulence factors: enterotoxigenic E. coli (ETEC), enteroaggregative E. coli (EAEC), enterohaemorrhagic E. coli (EHEC), Shiga-toxin producing E. coli (STEC), typical enteropathogenic E. coli (EPECt), atypical enteropathogenic E. coli (EPECa), enteroinvasive E. coli (EIEC) and diffusely adherent E. coli (DAEC) [6, 12]. Two recent large studies of aetiologic agents of diarrhoea, the Global Enteric Multicenter Study (GEMS) and the Mal-nutrition and Enteric Disease study (Mal-ED), found ETEC and Shigellae (considered an E. coli pathotype similar to EIEC) to be associated with more severe diarrhoeal disease, and EPECt and ETEC to be associated with increased risk of mortality in infants [13, 14]. The predominant transmission route for these enteric pathogens is faecal-oral (foodborne, waterborne and person-to-person), and the majority of diarrhoeal cases are due to a lack of access to clean water supply, poor sanitation and inadequate personal hygiene behaviours, especially in LMICs [15-19].
While water, sanitation and hygiene (WASH)-related conditions are the most salient factors in determining diarrhoea rates, and the target of most intervention programs, diarrhoeal diseases are complex due to the variety of aetiological agents and the dynamic interaction between personal socio-economic status, environment and behaviour contributing to disease transmission. One factor not traditionally considered in WASH studies is human travel, which can facilitate movement of enteric pathogens from one region to another [20]. Human movement is often considered for other types of infectious diseases [21-23], but for enteric diseases consideration of human travel as a risk factor has mostly been limited to introduction of new strains of cholera to naive populations [24-26], and to “traveller’s diarrhoea,” i.e., the acquisition of diarrhoea by people travelling internationally from low-transmission to high-transmission countries.
International travel is an established risk factor for diarrhoea, the most common illness afflicting travellers [27, 28], afflicting approximately 50% of travellers after a 2-week travel period [29, 30]. “Traveller’s diarrhoea” is often attributed to bacterial infections, particularly ETEC or EAEC [31-35]. Most studies focus on travellers from high-income countries who tend to be more immunologically naïve [31-36]. However, “traveller’s diarrhoea” can also feasibly occur within LMIC countries, as a result of within-country travel between areas of high vs. low transmission [20]. This is important because pathogens can have different prevalence even within small regions [37, 38]. Urban areas with higher transmission may be a source of pathogens for more rural regions.
This study was designed to explore the importance of within-country traveller’s diarrhoea and to understand how risk factors differ in urban vs. rural areas within a given country for diarrhoea and specifically DEC.
Methods
Study design & setting
The EcoZUR (E. coli en Zonas Urbanas y Rurales) study was an age-matched case-control study of diarrhoea conducted in four sites along an urban-rural gradient in northern Ecuador. Figure 1 shows the four study locations: Quito (0°10′50.35″S, 78°28′4.20″W), Esmeraldas (0°57′33.12″N, 79°39′14.29″W), Borbón (1°05′21.13″N, 78°59′23.86″W) and rural villages near the Borbón region. Quito is the capital of Ecuador (population ~1.6 million) [39]; Esmeraldas is the capital of Esmeraldas Province in northwest Ecuador (population ~162 000) [39]; Borbón is a town in Esmeraldas Province located at the junction of Cayapas, Santiago and Onzole rivers (population ~7000); and the rural villages comprise about 125 villages along the three rivers (populations ~50–500). The more urban centres (Quito and Esmeraldas) are more densely populated with greater access to water, sanitation, roads and medical infrastructure; whereas the more rural regions (Borbón and rural villages) are less densely populated, with minimal infrastructure. This study was designed to understand how factors associated with living conditions (e.g., population density, human travel, WASH conditions, animal contact) affect diarrhoea and the enteric pathogens in circulation along this urban-rural gradient. This case-control study design allowed for the capture of both symptomatic and asymptomatic enteric infections.
Region of Study. This map highlights the region (red dashed circle) and four study sites (yellow stars) where the Eco-ZUR study was conducted. These four sites comprise an urban-rural gradient in northern Ecuador: including Quito (pop. ~1.62 M), Esmeraldas (pop. ~162 000), Borbón (pop. ~5000) and the rural villages (pop. ~50–500).
Participant recruitment
Individuals of all ages were recruited from Ecuadorian Ministry of Public Health facilities at each study site between April 2014 and September 2015. Recruitment centres included Centro de Salud N°4 Chimbacalle in Quito, Hospital Delfina Torres de Concha in Esmeraldas, and Hospital Básico in Borbón. Participants from the rural villages were recruited either through scheduled Ministry of Public Health clinical visits on site or at Borbón Hospital if they presented there for medical attention. Through this recruitment strategy we were able to capture subjects from 58 of the 125 villages in the region. At each study location, our goal was to recruit 100 cases of patients presenting with diarrhoea – defined as three or more loose stools within the previous 24 h. Controls were individuals presenting with a non-diarrhoeal complaint at the same facility, without diarrhea or vomiting in the prior 7 days. A one-to-one age-match of cases and controls was carried out in real-time, based on the following age-matching criteria: 0–24 months (±6 months), 25–60 months (±12 months), 61–180 months (±24 months) and >181 months. Rural village control subjects were matched by the location of recruitment (i.e., through Ministry of Health visits to their village or at Borbón Hospital). Participants provided written consent unless they could not read or write, in which case they provided oral consent with a witness. An assent form was used for participants <18 years old, in conjunction with parental consent. Additionally, participants (cases and controls) were required to be a resident of the study location for at least 6 months in order to be included in the study. Any participants who reported usage of antibiotics in the prior week were also excluded.
All consenting participants and/or their guardians completed an electronic survey about demographics, socioeconomic status, medical history, WASH practises and travel history. The surveys were adapted from instruments previously developed for and implemented in this region (e.g. [37, 40]). Our trained study staff conducted these surveys using the Open Data Kit platform (http://opendatakit.org). Additionally, both cases and controls were asked to provide a stool sample for enteric pathogen analysis. Study staff provided the participant with a plastic stool collection container and instructions to return a stool sample within 24 h. If a participant was in diapers, the team member provided additional instructions to assist the parent or guardian in properly handling the stool specimen. A total of 907 subjects enroled successfully completed the survey and 85% of those provided a stool sample.
Escherichia coli pathotype determination
Fresh stool samples were cultured locally upon receipt and tested for DEC pathotypes based on the presence of specific virulence factors. At the Esmeraldas, Borbón and rural village sites samples were streaked for isolation of E. coli in a field laboratory on MacConkey Lactose agar media (MKL), and re-isolated for pure cultures upon arrival at the Universidad San Francisco de Quito (USFQ) laboratory. At the Quito site, Cary-Blair transport media swabs (BD, Franklin Lakes, NJ) were inoculated with faecal material, maintained at 4°C for a maximum of 24 h, and streaked for isolation upon arrival at USFQ on MKL agar. After 24 h of incubation, up to five lactose-positives and one non-lactose fermenting isolate were cultured on Chromocult agar media (Merck, Darmsladt, Germany) and tested for β-glucoronidase activity. Non-lactose-fermenting colonies were further characterised by biochemical tests to detect Shigellae or E. coli (some EIEC are non-fermenters) using the API 20E test (BioMérieux, Marcy l’Etoile, France). The 5–6 colonies were pooled, resuspended in 300 μl of sterile distilled water, boiled for 10 min to release the DNA, and the resulting supernatant was used for PCR testing. Nine singleplex PCR assays were used to detect the presence of virulence genes associated with each diarrhoeagenic E. coli pathotype: It and sta for ETEC [41]; ipaH for EIEC and Shigella [41]; aggR for EAEC [42]; afa for DAEC [43]; bfp for typical EPEC [41]; and eaeA for atypical EPEC [44]. Positive pools for eaeA were subsequently tested for stx1 and stx2 genes for the differentiation of potential EHEC infections [44]. If a pooled sample tested positive for any virulence factor then each of the five isolates were retested individually to identify the positive isolate.
Water and sanitation behaviours
Survey responses regarding water and sanitation behaviours were converted into “improved” vs. “unimproved” categories according to the WHO Joint Monitoring Programme (JMP) for Water Supply and Sanitation guidelines, with the exception of purchased bottled water which was considered improved based on the culture and practises of this region [45].
Defining domestic travel
Domestic travel was considered as any event in which the participant left his or her city/village of origin to visit another destination within Ecuador. Common options for destinations on all forms included Guayaquil, Quito, Santo Domingo, Esmeraldas, San Lorenzo, Borbón and rural villages surrounding the study site. Participants could also input other destinations of travel.
Data analysis
All analysis was completed using R Studio Statistical Software (http://www.rstudio.org/). Unadjusted odds ratios and corresponding P-values were computed using the Pearson’s Chi Square Test. If expected cell counts were less than five, the Fisher’s Exact Test was used.
Multivariate mixed-effect and general logistic regression models were used to compute adjusted odds ratios of the risk of diarrhoeal disease associated with WASH conditions, animal contact and travel. Models were adjusted for other extraneous factors including sex, age, race, government welfare status, and study site was included as a random effect for multi-site models. The DEC cross-sectional models were also adjusted for diarrhoea case status. The lme4 R package was used to run all mixed-effect models that included all study participants [46]. Mixed-effect models were used to assess risk among all participants across the study whereas general logistic regression models were used to assess risks within individual study locations. General logistic regression was also used to examine demographic characteristics associated with travel; variables included in this model were age, sex, government welfare status, education, family employment status and study location (urban vs. rural).
Ethics
Permission for the study and approval of human subjects was obtained from the Ecuadorian Ministry of Public Health (MSP-DIS-2014-0055-O), Emory Institutional Review Board (IRB00065781) and the USFQ Ethical Committee (2013-145M).
Results
The study included a total of 907 participants age 0–85 years old from Quito (n = 253, 28% of subjects), Esmeraldas (n = 209, 23% of subjects), Borbón (n = 243, 27% of subjects) and rural villages (n = 202, 22% of subjects). These four study sites successfully captured both an urban-rural socio-economic gradient and a water and sanitation accessibility gradient, with the urban sites having higher education, employment, improved sanitation and improved drinking water sources (Table 1).
Table 1
Characteristics of study participants by site and disease status in the EcoZUR case-control study of diarrhoea in four sites across an urban-rural gradient in northern Ecuador. This table highlights the expected differences in sociodemographic factors as well as water/sanitation practicss along an urban-rural gradient with fewer urban participants receiving welfare and more receiving higher education, sustaining employment, and overall improved water and sanitation compared to rural participants. However, within each study site there were no significant differences in demographic or sociodemographic characteristics between cases and controls with the expectation of water, sanitation and animal contacts as referenced in the text
| Quito (N = 253) | Esmeraldas (N = 209) | Borbón (N = 243) | Rural Villages (N = 202) | Overall differences across study sites P- value | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Overall, N (%) | Cases, n (%) | Controls, n (%) | Overall, N (%) | Cases, n (%) | Controls, n (%) | Overall, N (%) | Cases, n (%) | Controls, n (%) | Overall, N (%) | Cases, n (%) | Controls, n (%) | ||
| Demographics | |||||||||||||
| 0.01 | |||||||||||||
| 92 (36.4) | 45 (34.6) | 47 (38.2) | 73 (34.9) | 35 (35.0) | 38 (34.9) | 85 (35.0) | 49 (36.8) | 36 (32.7) | 58 (28.7) | 37 (30.8) | 21 (25.6) | ||
| 37 (14.6) | 20 (15.4) | 17 (13.8) | 36 (17.2) | 16 (16.0) | 20 (18.3) | 33 (13.6) | 20 (15.0) | 13 (11.8) | 54 (26.7) | 32 (26.7) | 22 (26.8) | ||
| 43 (17.0) | 20 (15.4) | 23 (18.7) | 47 (22.5) | 23 (23.0) | 24 (22.0) | 45 (18.5) | 24 (18.0) | 21 (19.1) | 41 (20.3) | 22 (18.3) | 19 (23.2) | ||
| 81 (32.0) | 45 (34.6) | 36 (29.3) | 53 (25.4) | 26 (26.0) | 27 (24.8) | 80 (32.9) | 40 (30.1) | 40 (36.4) | 49 (24.3) | 29 (24.2) | 20 (24.4) | ||
| 124 (49.0) | 68 (52.3) | 56 (45.5) | 113 (54.1) | 59 (59.0) | 54 (49.5) | 132 (54.3) | 78 (58.6) | 54 (49.1) | 108 (53.5) | 65 (54.2) | 43 (52.4) | 0.61 | |
| <0.001 | |||||||||||||
| 8 (3.2) | 5 (3.8) | 3 (2.4) | 1 (0.5) | 0 (0.0) | 1 (0.9) | 0 (0.0) | 0 (0.0) | 0 (0.0) | 0 (0.0) | 0 (0.0) | 0 (0.0) | ||
| 3 (1.2) | 2 (1.5) | 1 (0.8) | 87 (41.8) | 38 (38.0) | 49 (45.4) | 133 (54.7) | 72 (54.1) | 61 (55.5) | 108 (53.5) | 58 (48.3) | 50 (61.0) | ||
| 5 (2.0) | 3 (2.3) | 2 (1.6) | 1 (0.5) | 1 (1.0) | 0 (0.0) | 9 (3.7) | 6 (4.5) | 3 (2.7) | 37 (18.3) | 29 (24.2) | 8 (9.8) | ||
| 237 (93.7) | 120 (92.3) | 117 (95.1) | 119 (57.2) | 61 (61.0) | 58 (53.7) | 101 (41.6) | 55 (41.4) | 46 (41.8) | 57 (28.2) | 33 (27.5) | 24 (29.3) | ||
| Sociodemographics | |||||||||||||
| 15 (5.9) | 5 (3.8) | 10 (8.1) | 42 (20.1) | 20 (20.0) | 22 (20.2) | 54 (22.2) | 27 (20.3) | 27 (24.5) | 97 (48.0) | 58 (48.3) | 39 (47.6) | <0.001 | |
| 196 (77.5) | 04 (80.0) | 92 (74.8) | 94 (45.0) | 46 (46.0) | 48 (44.0) | 105 (43.2) | 52 (39.1) | 53 (48.2) | 39 (19.3) | 29 (24.2) | 10 (12.2) | <0.001 | |
| <0.001 | |||||||||||||
| 1 (0.4) | 1 (0.8) | 0 (0.0) | 3 (1.4) | 0 (0.0) | 3 (2.8) | 1 (0.4) | 0 (0.0) | 1 (0.9) | 6 (3.0) | 3 (2.5) | 3 (3.7) | ||
| 26 (10.3) | 11 (8.5) | 15 (12.2) | 14 (6.7) | 9 (9.1) | 5 (4.6) | 14 (5.8) | 10 (7.6) | 4 (3.6) | 62 (30.7) | 36 (30.0) | 26 (31.7) | ||
| 124 (49.0) | 69 (53.1) | 55 (44.7) | 112 (53.8) | 56 (56.6) | 56 (51.4) | 178 (73.6) | 101 (76.5) | 77 (70.0) | 120 (59.4) | 69 (57.5) | 51 (62.2) | ||
| 102 (40.3) | 49 (37.7) | 53 (43.1) | 79 (38.0) | 34 (34.3) | 45 (41.3) | 49 (20.2) | 21 (15.9) | 28 (25.5) | 14 (6.9) | 12 (10.0) | 2 (2.4) | ||
| Water, sanitation and animal contact | |||||||||||||
| 197 (77.9) | 103 (79.2) | 94 (76.4) | 145 (71.1) | 70 (73.7) | 75 (68.8) | 148 (60.9) | 86 (64.7) | 62 (56.4) | 82 (40.6) | 45 (37.5) | 37 (45.1) | <0.001 | |
| 252 (99.6) | 130 (100.0) | 122 (99.2) | 207 (99.5) | 99 (99.0) | 108 (99.1) | 221 (90.9) | 122 (91.7) | 99 (90.0) | 167 (82.7) | 102 (85.0) | 65 (79.3) | <0.001 | |
| 163 (64.4) | 77 (59.2) | 86 (69.9) | 85 (40.9) | 32 (32.0) | 53 (48.6) | 39 (16.1) | 20 (15.0) | 19 (17.3) | 44 (21.8) | 30 (25.0) | 14 (17.1) | <0.001 | |
| 140 (55.3) | 70 (53.8) | 70 (56.9) | 94 (45.0) | 49 (49.0) | 45 (41.3) | 100 (41.2) | 64 (48.1) | 36 (32.7) | 64 (31.7) | 34 (28.3) | 30 (36.6) | <0.001 | |
There were no significant socio-economic or demographic differences between cases and controls within any given site, indicating robust case-control matching (Table 1). Across all sites, there were slightly more males among cases, but this difference was not statistically significant within any single site. Among all participants who provided stool samples (n = 771), 27% (n = 208) were positive for DEC infection, with 43% (n = 90) of these infections present in asymptomatic controls.
Water and sanitation risk factors
Most participants used an improved household drinking water source, with a higher prevalence among urban participants (Quito 99.6%; Esmeraldas 99.5%) than rural participants (Borbón 90.9%; rural villages 82.7%). Subjects from urban sites were more likely to report treating their drinking water (Quito 64.4%; Esmeraldas 40.9%) than subjects in rural sites (Borbón 16.1%; rural villages 21.8%). The prevalence of reported use of improved household sanitation was lower than that of reported improved drinking water. Rural participants were less likely to have improved sanitation (61% in Borbón, 41% in rural villages) than urban participants (78% in Quito, 71% in Esmeraldas).
Multivariable logistic regression exploring associations between WASH practises and diarrhoea case status and DEC infections are shown in Table 2. The use of improved household drinking water sources was not significantly associated with diarrhoea case status or DEC infections, either collectively or within any individual site. However, household drinking water treatment was protective against diarrhoea (adjusted Odds Ratio [aOR] = 0.72, 95% CI = 0.54–0.96, P = 0.03). The coefficients were higher in urban regions (Quito aOR = 0.64, 95% CI = 0.37–1.10, P = 0.11; Esmeraldas aOR = 0.34, 95% CI = 0.17–0.64, P = 0.001), where >40% of participants reported these practises. Of those reporting treatment of household drinking water, nearly all (89%) reported boiling as their form of treatment, with a few reporting use of chlorine (8%) or filtration techniques (2%). Drinking water treatment was not associated with DEC infections. Improved household sanitation was associated with increased risk of diarrhoea in Borbón (aOR = 2.23, 95% CI = 1.14–4.50, P = 0.02) and increased risk of DEC infections in Quito (aOR = 3.67, 95% = 1.17–13.3, P = 0.05), but reduced the risk of DEC infection in Esmeraldas (aOR = 0.39, 95% = 0.15–0.97, P = 0.05).
Table 2
Multivariate logistic regression models for risk of (A) Diarrhoeal Disease and (B) Diarrhoeagenic Escherichia coli infections across an urban-rural gradient: Effects of household water and sanitation practises on disease outcomes. Variables adjusted for in the models are given in the text. Total number of participants reporting diarrhoea is 483. Total number of participants with diarrhoeagenic E. coli infections is 208
| Reported use of improved household sanitation facilities | Reported use of improved household drinking water source | Reported treatment of household drinking water | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Study Participants | n (%) | aOR (95% CI) | P-value | n (%) | aOR (95% CI) | P-value | n (%) | aOR (95% CI) | P-value |
| A. Association with diarrhea case status (n = 483) | |||||||||
| 572 (63.4) | 1.23 (0.88, 1.70) | 0.22 | 847 (93.5) | 1.15 (0.67, 1.98) | 0.62 | 331 (36.6) | 0.72 (0.54, 0.96) | 0.03* | |
| 197 (77.9) | 1.12 (0.54, 2.33) | 0.76 | 252 (99.6) | – | 1.00 | 163 (64.4) | 0.64 (0.37, 1.10) | 0.11 | |
| 145 (71.1) | 1.05 (0.46, 2.40) | 0.91 | 207 (99.5) | – | 0.99 | 85 (40.9) | 0.34 (0.17, 0.64) | 0.001*** | |
| 148 (60.9) | 2.23 (1.14, 4.50) | 0.02* | 221 (90.9) | 0.90 (0.31, 2.54) | 0.84 | 39 (16.1) | 0.77 (0.36, 1.63) | 0.49 | |
| 82 (40.6) | 0.96 (0.50, 1.85) | 0.91 | 167 (82.7) | 1.57 (0.72, 3.43) | 0.25 | 44 (21.8) | 2.05 (0.98, 4.45) | 0.06 | |
| B. Association with diarrheagenic E. coli (n = 208) | |||||||||
| 489 (63.8) | 0.90 (0.60, 1.33) | 0.59 | 770 (93.0) | 1.00 (0.50, 1.97) | 0.99 | 266 (34.5) | 0.79 (0.55, 1.14) | 0.21 | |
| 152 (82.2) | 67 (1.17, 13.3) | 0.03* | 184 (99.5) | – | 1.00 | 116 (62.7) | 1.19 (0.55, 2.65) | 0.66 | |
| 134 (71.7) | 0.39 (0.15, 0.97) | 0.05* | 190 (99.5) | – | 1.00 | 80 (41.9) | 1.18 (0.58, 2.42) | 0.64 | |
| 128 (59.5) | 0.83 (0.37, 1.86) | 0.64 | 196 (91.2) | 1.60 (0.47, 6.17) | 0.47 | 32 (14.9) | 0.79 (0.26, 2.08) | 0.64 | |
| 75 (41.9) | 0.52 (0.22, 1.18) | 0.12 | 146 (81.6) | 1.31 (0.51, 3.75) | 0.59 | 38 (21.2) | 0.55 (0.18, 1.50) | 0.27 | |
aOR, adjusted odds ratio; CI, confidence interval.
Animal exposure risk factors
Forty-four percent of all participants reported contact with animals, including cows, chickens, pigs, dogs, cats and others. A higher proportion of urban study participants reported animal contact, defined as contact with any animal(s) in the past week, (55% in Quito; 45% in Esmeraldas) than rural study participants (41% in Borbón; 32% in rural villages). Urban participants most commonly reported contact with domestic animals, including dogs and cats. Conversely, rural participants reported a more diverse array of animal contacts, including dogs, cats, chickens, pigs, cows and wild game. No significant association was found between reported animal contacts and diarrhoea among the urban participants. However, in Borbón, cases were at least twice as likely to report contact with animals (aOR = 2.08, 95% CI = 1.19–3.67, P = 0.01). No significant association was found between reported animal contact and the presence of DEC at any site.
Domestic travel within Ecuador
Forty-six percent of all participants reported some form of domestic travel in the past year, with a median of four trips. Rural participants reported higher rates of travel (67% in Borbón and 63% in the rural villages) than urban participants (22% in Quito and 33% in Esmeraldas). Figure 2 illustrates the prevalence and distribution of domestic travel in the past year. Urban participants reported more travel to other urban centres (i.e. Guayaquil, Quito, and Santo Domingo) and rural participants reported travel to the more local rural areas (other rural villages, Borbón, and San Lorenzo) (Figure S1). Duration of stay varied based on destination. 62% of participants travelling to urban destinations reported a duration of stay up to 1 week, with those travelling to Guayaquil and Quito staying for up to 2 weeks. Rural destinations on the other hand had shorter visits, with 73% of travellers to rural destinations reporting a visit of just 1 day. The reasons reported for travel are depicted in Figure 3, again with variable differences by destination. There were no significant differences in travel reasons between cases and controls. Across all sites, the most common reason for travelling was for “family” reasons (31% across all sites) followed by “medical” reasons (20% across all sites) and “shopping” (15% across all sites). Few participants (6%) reported travelling within Ecuador in the past week, so we were unable to carry out a robust analysis of these data.
Prevalence of reported travel between study sites in the past year. Travel destination is denoted by the arrow, with the thickness of the arrow weighted by prevalence of reported travel to this destination. (a) Travel from urban/more populated cities to more rural/less populous communities. (b) Travel from rural/less populous communities to urban/more populated cities. It is important to note that not all travel is represented in these figures since data was only collected from participants at the four study locations (yellow stars). Individuals residing in Guayaquil, Santo Domingo and San Lorenzo may engage in more urban to rural travel or rural to urban travel that is not illustrated above.
Multivariate logistic regression model for characterizing Ecuadorian travelers. Demographic characteristics associated with domestic travel in the past year among study participants. All variables listed were included together in the model.
Common characteristics of Ecuadorian travellers
Travel was significantly more common among rural participants than urban participants, with rural participants having eight times higher odds of travel in the past year compared to their urban counterparts (aOR = 8.10, 95% CI = 5.77–11.52, P < 0.001) (Figure 2). Other demographic factors were analysed to highlight characteristics associated with travel in this population (Table 3). Participants over the age of 5 were twice as likely to have travelled in the past year compared to those <5 (aOR = 2.04, 95% CI = 1.51–2.76, P < 0.001). Additionally, individuals who report receiving government subsidised welfare were less likely to report travel in the past year (aOR = 0.51, 95% CI = 0.35–0.75, P < 0.001).
Table 3
Multivariate logistic regression model for characterising Ecuadorian travelers: Demographic characteristics associated with domestic travel in the past year among study participants. All variables listed were included together in the model
| Demographics | n (%) | Reported domestic travel in past year n (%) | No reported domestic travel in the past year n (%) | aOR (95% CI) | P-value |
|---|---|---|---|---|---|
| Study site | |||||
| 462 (50.9) | 125 (30.0) | 337 (68.6) | Ref. | Ref. | |
| 445 (49.1) | 291 (70.0) | 154 (31.4) | 8.10 (5.77, 11.52) | <0.001*** | |
| Age | |||||
| 468 (51.6) | 185 (44.5) | 283 (57.6) | Ref. | Ref. | |
| 439 (48.4) | 231 (55.5) | 208 (42.4) | 2.04 (1.51, 2.76) | <0.001*** | |
| Male | 477 (52.6) | 230 (55.3) | 247 (50.3) | 1.30 (0.96, 1.75) | 0.09 |
| Receives government welfare | 208 (22.9) | 88 (21.1) | 120 (24.4) | 0.51 (0.35, 0.75) | <0.001*** |
| Member of family employed | 434 (48.0) | 192 (46.2) | 242 (49.4) | 1.11 (0.80, 1.56) | 0.53 |
| Highest level of household education | |||||
| 244 (27.0) | 109 (26.3) | 135 (27.6) | Ref. | Ref. | |
| 534 (59.0) | 262 (63.1) | 272 (55.5) | 0.70 (0.48, 1.01) | 0.06 | |
| 116 (12.8) | 40 (9.6) | 76 (15.5) | 0.34 (0.19, 0.59) | <0.001*** | |
| 11 (1.2) | 4 (1.0) | 7 (1.4) | 0.30 (0.07, 1.16) | 0.09 |
Domestic travel risk factors
To assess the potential risk that domestic travel presents to diarrhoeal disease transmission, we analysed reported travel data to any destination in Ecuador in the past year. These data were analysed both by travel origin and destination of travel. All analyses were adjusted for potential confounding by sex, age, race, economic status, study site (random effect for “all sites” models only), and known risk factors for diarrhoea disease transmission, i.e., sanitation and water treatment. DAEC was the most commonly encountered E. coli pathotype, and the only pathotype with sufficient sample size for this analysis. Figure 4 displays the results by origin of travel. Any reported travel in the past year trended towards a positive association with diarrhoea case status across all participants, although this result was not statistically significant (aOR = 1.30, 95% CI = 0.99–1.72, P = 0.06). However, participants travelling from Esmeraldas had more than double the risk of diarrhoea as those not travelling (aOR = 2.28, 95% CI = 1.20–4.41, P = 0.01). Additionally, participants from Quito (but not other sites or all sites combined) reporting travel in the past year had a higher risk of DEC infections (aOR = 2.61, 95% CI = 1.16–8.86, P = 0.02). No associations were observed between domestic travel in the past year and DAEC infections.
Risk by origin of travel. Multivariate logistic regression models for risk of domestic travel in the past year from locations across an urban-rural gradient on (a) diarrhoeal disease (b) diarrhoeagenic Escherichia coli infections and (c) diffuse adherent E. coli (DAEC). Odds ratios are adjusted based on sex, age, race, economic status, study site (random effect for “all sites” models only), diarrhoea case status (DEC models only), and known risk factors for diarrhoea disease transmission, i.e. household sanitation classification and reported household water treatment, with the exception of the DAEC subset, which is presented as unadjusted due to the relatively rare nature of these events.
Figure 5 shows travel risks by destination of travel. Travel destinations included in this analysis, from most populous to least populous are as follows: Guayaquil, Quito, Santo Domingo, Esmeraldas, San Lorenzo, Borbón and rural villages. Participants from each of the four study sites reported destinations visited in the past year, which allowed us to observe trends in which destinations were frequented most by urban vs. rural travellers and assess risks between expected high and low transmission regions (Figure S2). Participants reporting travel to Quito (primarily from Esmeraldas) were more likely to be a diarrhoea case than those who did not report travel to Quito (aOR = 2.01, 95% CI = 1.10–3.68, P = 0.02). Participants reporting travel to San Lorenzo (primarily from Borbón) were nearly twice as likely to be infected with DEC than those who did not travel to this destination (aOR = 1.96, 95% CI = 1.09–3.47, P = 0.02). Participants reporting travel to Guayaquil (primarily from Esmeraldas and Quito) were twice as likely to be infected with DAEC (aOR = 2.09, 95% CI = 1.01–4.33, P = 0.04). However, those reporting travel to Esmeraldas (solely participants from Borbón and rural villages) were significantly less likely to present with DEC infections (aOR = 0.56, 95% CI = 0.33–0.96, P = 0.03), and specifically DAEC infections (aOR = 0.37, 95% CI = 0.17–0.78, P = 0.007), than those who did not travel to Esmeraldas.
Risk by destination of travel. Multivariate logistic regression models for risk of domestic travel in the past year to destinations across an urban-rural gradient on (a) diarrhoeal disease (b) diarrhoeagenic Escherichia coli infections and (c) diffuse adherent E. coli (DAEC). Odds ratios are adjusted based on sex, age, race, economic status, study site (random effect for “all sites” models only), diarrhoea case status (DEC models only), and known risk factors for diarrhoea disease transmission, i.e. household sanitation classification and reported household water treatment, with the exception of the DAEC subset, which is presented as unadjusted due to the relatively rare nature of these events.
Discussion
“Travellers’ diarrhoea” is typically associated with visitors from low-transmission countries visiting high-transmission countries, but this study suggests that this phenomenon can also occur as a result of travel between higher- and lower-transmission locations within a given country. Global urbanisation in recent decades has brought about infrastructure changes that provide increased accessibility to roads and transportation networks, facilitating travel and increasing contact with neighbouring communities. This interconnectivity can be beneficial for rural communities’ access to jobs and medical care. However, it can also facilitate transmission of enteric and other pathogens from high-transmission urban areas, especially when new or more virulent pathogens are introduced to previously naïve populations.
Results from our case-control study across a rural-urban gradient in northern Ecuador suggest that travel in the past year, and particularly travel to urban areas, was associated with an increased risk of illness and/or infection. We found elevated risk of diarrhoea for subjects reporting domestic travel in the past year, and in particular travel from Esmeraldas, 69% of whom listed the more urban sites of Quito and Guayaquil as their destinations. (Figure 4a) Reported travel in the past year from Quito was associated with elevated risk of DEC infection (Figure 4b); 50% of these travellers listed Guayaquil, the only city in Ecuador larger than Quito, as their destination. The models assessing destination provide further insights. The elevated diarrhoea risk associated with travel to Quito (Figure 5a) was primarily driven by travellers from Esmeraldas, who comprised 46% of travellers to Quito. The elevated diarrhoea risk associated with travel to Borbón (Figure 5a) was primarily driven by travellers from rural villages, who comprised 98% of travellers to Borbón; however, this elevated risk was not significant. DEC infection was higher in people reporting travel to San Lorenzo, 90% of who were from Borbón or the rural villages (Figure 5b). Travel to Quito and Guayaquil, the most urban sites, also trended towards elevated risk of DAEC infection (Figure 5c). Interestingly, this pattern did not hold up for travellers reporting visits to Esmeraldas, 100% of whom were from the two rural study sites. Travelling to Esmeraldas was even protective against DEC and DAEC infections. The destination models were adjusted for study site, so the patterns cannot be attributed to specific pathogens circulating within a given site.
In general, we found that significantly more rural than urban participants travel and the movement is towards more urban destinations. This makes sense as people living in rural areas need to travel for basic needs not found locally such as medical purposes, shopping and paperwork (Figure 3). Most of the urban participants reporting travel also went to more urban destinations (Figure S1). Furthermore, travel to urban destinations such as Quito and Guayaquil was often associated with longer durations of stay, which increases the probability of contracting an enteric infection.
Previous studies in Northwestern Ecuador have shown that more rural communities have lower diarrhoea rates, potentially due to reduced contact with outside individuals [37, 38]. Our study has identified a similar pattern but as it relates to travel, with rural participants from lower disease transmission environments being exposed to outside environments with higher disease transmission within urban centres. Urban, more populous regions may act as a hub for transmission of pathogens because of the higher density of people living and moving through the region, increasing intensity of pathogen transmission. Pathogen virulence has been shown to be related to transmission intensity in theoretical, natural and experimental systems [47, 48]. In particular, high-transmission rates, denser host populations, and greater host mobility tend to select for highly virulent pathogens, whereas the reverse conditions select for less-virulent pathogens. This well-described evolutionary relationship may have consequences for diarrhoeal disease management over large scales. If urban-derived pathogen strains cause increased morbidity and mortality amongst rural populations as a function of travel, a cost-effective country-wide strategy may focus on large-scale infrastructural water provision or vaccination efforts in urban areas. In order to better target these interventions, further investigation into regional and even local pathogen-specific disease prevalence is needed to better understand the potential in-country travel risks between areas of higher and lower disease transmission.
In addition to travel, a variety of environmental factors was assessed. We found few and even contradicting associations between improved water and sanitation conditions and diarrhoea case status as well as DEC-specific infections. Nearly all participants in this study had an improved drinking water source, even in rural areas, so we might not have had sufficient variability in this exposure to detect a difference. However, despite the high prevalence of improved drinking water, treatment of household drinking water was protective against diarrhoea, especially in urban areas where this behaviour was more heavily practised. This finding emphasises that “improved” drinking water sources may not necessarily be safe [49, 50], and that the practise and encouragement of water treatment is justified to reduce the transmission of diarrhoeal illness via water consumption.
We also found that recent contact with animals was associated with diarrhoeal disease. While we did not have sufficient data to explore these associations in depth, our results are suggestive that contact with animals in and around the home may increase the risk of diarrhoeal disease. Animals are proving to be important in the transmission of some enteric pathogens and further research is needed [51-53].
This study had several limitations. We focused on reported travel in the past year due to a lack of individuals reporting travel in the past week, limiting our ability to attribute risk to recent travel. We adjusted our models to account for a number of potential confounding factors associated with travel in the past year (e.g., socioeconomic status) but there may be other unmeasured factors influencing rates of travel. Travel itself is variable, with individual differences in destinations, reasons for travel, duration of stay and behavioural practises of which we incorporated into these findings but not all factors could be included in our models. Travellers to urban locations reported longer durations of stay, so the increased risk of disease associated with urban destinations may simply be attributable to increased probability of acquisition of a pathogen over longer time periods. Our data unfortunately does not have the resolution to account for this possibility. However, even if length of stay accounts for the increased risk, the overall finding of increased risk associated with travel to urban area still holds. Additionally, subject recruitment was based on diarrhoea case or control status, so the DEC results were based on a cross-sectional analysis within our case-control study. However, we adjusted the DEC models for diarrhoea case status to account for potential confounding based on disease status. In addition, due to logistical considerations, sampling in rural communities was somewhat different than in the urban communities. In the rural communities, we recruited subjects via two strategies: First, we recruited participants from rural communities who presented at the Borbón Hospital (31% of rural participants). This strategy was limited by the number of diarrhoea cases that presented, and also by the ability to obtain samples from age-matched controls, as visitors to Borbón were eager to return home. To supplement this strategy our staff accompanied Ministry of Health medical teams on their regular field visits to communities and enroled cases and controls during these community clinics (69% of rural participants). Thus, these individuals may have had less severe diarrhoea than those presenting at the hospitals or clinics in the other sites. However, the severity of these participants’ diarrhoea for rural visitors to the Borbón Hospital was likely greater than participants at the other sites, as their illness had to be severe enough to warrant a trip to Borbón. Furthermore, these findings may not be applicable to other regions with more or less interconnectivity than found in Ecuador.
Acknowledgements
We would like to thank the local authorities of Ministry of Public Health for providing us access to the facilities in their hospitals. We would also like to thank Denys Tenorio and Mauricio Ayoví for their assistance with carrying out the field portions of the study. Funding for this study was provided by National Institute for Allergy and Infectious Diseases grant 1K01AI103544. The content is solely the responsibility of the authors and does not necessarily represent the official views of the funders.
Footnotes
Supporting Information
Additional supporting information may be found online in the Supporting Information section at the end of the article:
Figure S1. Distribution of reported travel destinations visited by participants across the four study sites.
Figure S2. Reported Domestic Travel Destination by Study Site.
References
Full text links
Read article at publisher's site: https://doi.org/10.1111/tmi.13183
Read article for free, from open access legal sources, via Unpaywall:
https://onlinelibrary.wiley.com/doi/pdfdirect/10.1111/tmi.13183
Citations & impact
Impact metrics
Citations of article over time
Alternative metrics
Article citations
Multilevel factors drive child exposure to enteric pathogens in animal feces: A qualitative study in northwestern coastal Ecuador.
PLOS Glob Public Health, 4(9):e0003604, 18 Sep 2024
Cited by: 0 articles | PMID: 39292655 | PMCID: PMC11410186
The development and validation of a survey to measure fecal-oral child exposure to zoonotic enteropathogens: The FECEZ Enteropathogens Index.
PLOS Glob Public Health, 4(9):e0002690, 10 Sep 2024
Cited by: 1 article | PMID: 39255298 | PMCID: PMC11386431
Comparison of metagenomic and traditional methods for diagnosis of E. coli enteric infections.
mBio, 15(4):e0342223, 15 Mar 2024
Cited by: 1 article | PMID: 38488359 | PMCID: PMC11005377
Why are so many enteric pathogen infections asymptomatic? Pathogen and gut microbiome characteristics associated with diarrhea symptoms and carriage of diarrheagenic <i>E. coli</i> in northern Ecuador.
Gut Microbes, 15(2):2281010, 22 Nov 2023
Cited by: 1 article | PMID: 37992406 | PMCID: PMC10730187
Distribution of Escherichia coli Pathotypes along an Urban-Rural Gradient in Ecuador.
Am J Trop Med Hyg, 109(3):559-567, 07 Aug 2023
Cited by: 3 articles | PMID: 37549901 | PMCID: PMC10484266
Go to all (10) article citations
Data
Data behind the article
This data has been text mined from the article, or deposited into data resources.
BioStudies: supplemental material and supporting data
Similar Articles
To arrive at the top five similar articles we use a word-weighted algorithm to compare words from the Title and Abstract of each citation.
Travellers' diarrhoea: contemporary approaches to therapy and prevention.
Drugs, 66(3):303-314, 01 Jan 2006
Cited by: 23 articles | PMID: 16526819
Review
Antimicrobial resistance profiles of diarrhoeagenic Escherichia coli isolated from travellers returning to the UK, 2015-2017.
J Med Microbiol, 69(7):932-943, 12 Jun 2020
Cited by: 7 articles | PMID: 32530393
Population genomics of diarrheagenic Escherichia coli uncovers high connectivity between urban and rural communities in Ecuador.
Infect Genet Evol, 113:105476, 29 Jun 2023
Cited by: 2 articles | PMID: 37392822 | PMCID: PMC10599324
Controlled study on enteropathogens in travellers returning from the tropics with and without diarrhoea.
Clin Microbiol Infect, 17(8):1194-1200, 13 Dec 2010
Cited by: 41 articles | PMID: 21054662
Funding
Funders who supported this work.
NIAID NIH HHS (2)
Grant ID: R01 AI137679
Grant ID: K01 AI103544
National Institute for Allergy and Infectious Diseases (1)
Grant ID: 1K01AI103544
National Institute of Allergy and Infectious Diseases (1)
Grant ID: 1K01AI103544