http://orcid.org/0000-0001-7807-5799
Figures
Abstract
Background
Campylobacter is one of the main causes of gastroenteritis worldwide. Most of the current knowledge about the epidemiology of this food-borne infection concerns two species, C. coli and C. jejuni. Recent studies conducted in developing countries and using novel diagnostic techniques have generated evidence of the increasing burden and importance of other Campylobacter species, i.e. non-C. coli/jejuni. We performed a nested case-control study to compare the prevalence of C. coli/jejuni and other Campylobacter in children with clinical dysentery and severe diarrhea as well as without diarrhea to better understand the clinical importance of infections with Campylobacter species other than C. coli/jejuni.
Methodology/Principal findings
Our nested case-control study of 439 stool samples included dysenteric stools, stools collected during severe diarrhea episodes, and asymptomatic stools which were systematically selected to be representative of clinical phenotypes from 9,160 stools collected during a birth cohort study of 201 children followed until two years of age. Other Campylobacter accounted for 76.4% of the 216 Campylobacter detections by qPCR and were more prevalent than C. coli/jejuni across all clinical groups. Other Campylobacter were also more prevalent than C. coli/jejuni across all age groups, with older children bearing a higher burden of other Campylobacter. Biomarkers of intestinal inflammation and injury (methylene blue, fecal occult test, myeloperoxidase or MPO) showed a strong association with dysentery, but mixed results with infection. MPO levels were generally higher among children infected with C. coli/jejuni, but Shigella-infected children suffering from dysentery recorded the highest levels (26,224 ng/mL); the lowest levels (10,625 ng/mL) were among asymptomatic children infected with other Campylobacter. Adjusting for age, sex, and Shigella infection, dysentery was significantly associated with C. coli/jejuni but not with other Campylobacter, whereas severe diarrhea was significantly associated with both C. coli/jejuni and other Campylobacter. Compared to asymptomatic children, children suffering from dysentery had a 14.6 odds of C. coli/jejuni infection (p-value < 0.001, 95% CI 5.5–38.7) but were equally likely to have other Campylobacter infections–odds ratio of 1.3 (0.434, 0.7–2.4). Children suffering from severe diarrhea were more likely than asymptomatic children to test positive for both C. coli/jejuni and other Campylobacter–OR of 2.8 (0.034, 1.1–7.1) and 1.9 (0.018, 1.1–3.1), respectively. Compared to the Campylobacter-free group, the odds of all diarrhea given C. coli/jejuni infection and other Campylobacter infection were 8.8 (<0.001, 3.0–25.7) and 2.4 (0.002, 1.4–4.2), respectively. Eliminating other Campylobacter in this population would eliminate 24.9% of the diarrhea cases, which is almost twice the population attributable fraction of 15.1% due to C. coli/jejuni.
Conclusions/Significance
Eighty-seven percent of the dysentery and 59.5% of the severe diarrhea samples were positive for Campylobacter, Shigella, or both, emphasizing the importance of targeting these pathogens to limit the impact of dysentery and severe diarrhea in children. Notably, the higher prevalence of other Campylobacter compared to C. coli/jejuni, their increasing burden during early childhood, and their association with severe diarrhea highlight the importance of these non-C. coli/jejuni Campylobacter species and suggest a need to clarify their importance in the etiology of clinical disease across different epidemiological contexts.
Author summary
Campylobacter is a major public health concern in developed and developing countries. C. coli and C. jejuni have long been considered to be the major disease-causing species, and clinical microbiologic approaches target these two species. However, less selective diagnostic approaches have shown the increasing importance of other Campylobacter species (i.e. non-C. coli/jejuni). Our case-control study investigated the association between diarrhea, C. coli/jejuni, and other Campylobacter among 439 stool samples from 201 children in peri-urban communities in Loreto, Peru. Three quarters of the 216 Campylobacter detections were associated with other Campylobacter, whose prevalence increased with age and was greater than that of C. coli/jejuni in all age and clinical groups (dysentery, severe diarrhea, and asymptomatic). Despite their lower prevalence, C. coli/jejuni were more strongly associated with higher levels of myeloperoxidase, clinical dysentery, and the presence of leukocytes and blood in the stool compared to other Campylobacter. Other Campylobacter were equally likely as C. coli/jejuni to be detected in severe diarrhea cases–odds ratio of 1.9 (p-value = 0.018, 95% CI 1.1–3.1) and 2.8 (0.034, 1.1–7.1), respectively. Removing C. coli/jejuni in this population would eliminate 15.1% of diarrhea compared to 24.9% if other Campylobacter were eliminated.
Citation: François R, Yori PP, Rouhani S, Siguas Salas M, Paredes Olortegui M, Rengifo Trigoso D, et al. (2018) The other Campylobacters: Not innocent bystanders in endemic diarrhea and dysentery in children in low-income settings. PLoS Negl Trop Dis 12(2): e0006200. https://doi.org/10.1371/journal.pntd.0006200
Editor: Joseph M. Vinetz, University of California San Diego School of Medicine, UNITED STATES
Received: September 22, 2017; Accepted: December 31, 2017; Published: February 7, 2018
This is an open access article, free of all copyright, and may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for any lawful purpose. The work is made available under the Creative Commons CC0 public domain dedication.
Data Availability: All relevant data are within the paper and its Supporting Information files.
Funding: The Etiology, Risk Factors and Interactions of Enteric Infections and Malnutrition and the Consequences for Child Health and Development Project (MAL-ED) is carried out as a collaborative project supported by the Bill & Melinda Gates Foundation (47075), the Foundation for the National Institutes of Health, and the National Institutes of Health, Fogarty International Center. MNK was additionally supported by the Sherrilyn and Ken Fisher Center for Environmental Infectious Diseases of the Johns Hopkins School of Medicine (10POS2015). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Campylobacter are curved gram-negative bacteria and a common cause of enteritis worldwide. The Campylobacter genus is composed of 26 species [1], some of which are thermo-tolerant such as C. jejuni and C. coli, the most commonly reported Campylobacter species affecting humans (82.7% and 17.3%, respectively) [2]. Campylobacteriosis, a food-borne zoonotic infection, is usually diagnosed by selective culture from stool samples, using conditions designed to enhance isolation of C. jejuni and C. coli [3,4]. As a result, most of our understanding of the burden and clinical manifestations of campylobacteriosis is specific to C. coli and C. jejuni. Diarrhea caused by Campylobacter species other than C. coli and C. jejuni is thought to be less severe [1], but this common belief is poorly documented [4].
Several studies have identified a significant burden of disease associated with Campylobacter infection in children less than five years of age. This includes studies that have demonstrated that Campylobacter infection is a risk factor for poor linear growth [5–8]. In the Global Enteric Multicenter Study (GEMS), Campylobacter species had the 5th greatest attributable fraction of moderate-to-severe diarrhea in children under 5, with an attributable fraction as high as 16.4% in some areas [9]. The Etiology, Risk Factors and Interactions of Enteric Infections and Malnutrition and the Consequences for Child Health and Development (MAL-ED) is a birth cohort study conducted in eight different countries. Analyses from MAL-ED have also identified a significant burden of diarrhea associated with Campylobacter using both enzyme immunoassay (EIA) and polymerase chain reaction (PCR) for detection [6,10]. An analysis of a subset of Campylobacter-positive samples showed that roughly 30% of detections by EIA were due to other Campylobacter species, including C. hyointestinalis, C. troglodytis, and C. upsaliensis, which made up an important fraction of the disease burden [6]. This finding is important because other diagnostic tests, both culture and FDA-approved nucleic acid based including tests such as BD Max Enteric Bacterial Panel, BD Diagnostics (Hunt Valley, MD), and BioFire Film Array Gastrointestinal Panel (Salt Lake City, UT) [11], often target C. jejuni and C. coli but do not reliably detect other Campylobacter species. Prior birth cohort studies in Peru using culture-based methods have also shown that Campylobacter is a contributor to the acquisition of postnatal linear growth deficits among children in Peru [5].
Considering the prevalence of Campylobacter in diarrhea among children under five years old as identified by EIA in MAL-ED, we performed this nested case-control study using samples from the MAL-ED birth cohort study in Peru to better understand the pathogenicity of the other Campylobacter species (i.e. non-C. coli/jejuni species) among children living in extreme poverty.
Methods
Study design and population
We identified all dysentery samples in the Peru MAL-ED cohort at the time of this study (n = 99) and then identified two severe diarrhea and two control/asymptomatic samples matched by the child’s age (± 6 months) using a random number generator. In total, we selected 99 dysentery stool samples, 198 stools from severe diarrhea, and 198 control samples. Diarrhea was defined as maternal report of three or more loose stools in a 24-hour period or at least 1 stool with visible blood present [12], and unique diarrheal episodes were separated by at least 2 days without diarrhea [10]. Dysentery was defined as diarrhea with visible blood reported by the child’s mother during the identified illness episode, and severe diarrhea as diarrhea samples with a modified Vesikari severity score (MAL-ED score) of ≥ 5 [13,14]. Control samples were non-diarrhea stools (i.e. monthly MAL-ED stools). The severe diarrhea samples were selected from a pool of 849 samples, and the control samples from a pool of 8,212 samples which were both ordered and selected by random number generator.
Ethics statement
The samples used in this study were collected as part of the larger MAL-ED studies following protocols approved by the Institutional Review Board of the Johns Hopkins Bloomberg School of Public Health (Baltimore, USA) and of the Asociación Benéfica PRISMA (Iquitos, Peru). The data analyzed were anonymized to protect the privacy of the participants (S1 Data).
Stool diagnostic microbiology
Stools were inoculated on Campylobacter agar base supplemented with 10% defribrinated sheep’s blood with Blaser’s supplement (Beckton Dickinson, Sparks, MD) containing vancomycin, cephalothin, trimethoprim, polymixin, and amphotericin B. Plates were incubated for 48 hours at 42°C at 5% O2, 10% CO2, 85% N2. If no growth was observed, agar plates were held at least 72 hours to confirm this finding. Colonies with typical morphology were confirmed by gram stain, and oxidase and catalase testing using standard microbiologic techniques. Campylobacter colonies were identified as C. jejuni if they hydrolyzed hippurate.
Screening for biomarkers of gut inflammation and permeability
Methylene blue was performed as commonly described [15] and fecal occult blood screening was done using commercial cards (Hemoccult, BeckmanCoulter) per manufacturer’s instructions. Samples were also analyzed for myeloperoxidase, MPO, (Alpco, Salem, NH) as described by Kosek et al. to assess neutrophil activity [16].
Culture of Campylobacter spp. as positive and negative controls for quantitative PCR (qPCR) analysis
Campylobacter and Shigella used as controls for qPCR analyses in this study were American Type Culture Collection (ATCC) derived. The lyophilized C. coli ATCC 33559 was rehydrated in tryptic soy broth and grown under micro-aerophilic conditions (5% carbon dioxide, 10% oxygen and balance nitrogen) for 48 hours at 37°C. A C. jejuni culti-loop (ATCC 33291) was grown for 10 min with shaking at 37°C in Luria broth then streaked onto blood agar plates with non-selective medium. The lyophilized C. upsaliensis ATCC 49815 was rehydrated in tryptic soy broth with 5% sheep’s blood and 0.001% pyridoxal-HCl, and cultured on Tryptone Soy Agar (TSA) with 5% sheep’s blood. Both C. jejuni and C. upsaliensis were grown under the same above-mentioned micro-aerophilic conditions for 48 hours at 37°C. Given the reported similarity between some Campylobacter-associated illness and shigellosis [17,18], we also tested the stools for Shigella. S. flexneri ATCC 12022 2b was grown on TSA Congo Red plates overnight at 37°C.
qPCR multiplex assays
DNA was extracted from the control bacterial cultures and stools of children using the MO-BIO DNA Power-Soil extraction kit (Carlsbad, CA). Approximately 0.05–0.10 g of feces were weighted and used for DNA extraction. 16S sRNA, Campylobacter adhesin to fibronectin (cadF), and invasion plasmid antigen H (ipaH) were used to detect all Campylobacter species, C. coli/jejuni, and Shigella by qPCR, respectively. The 25-μl 16S-cadF-ipaH triplex mixtures were prepared using 1 μl of DNA sample, 12.5 μl of Taq Environmental Master Mix 2.0 (ThermoFisher Scientific), 0.375 μl of a primer-probe mix at final concentrations of 0.2 μM for each primer and 0.1 μM for each probe, and nuclease-free water making up the remaining volume. qPCR was performed on a StepOnePlus instrument (Applied Biosystems, Foster City, CA) using the following cycling conditions: 95°C for 10 min, followed by 45 cycles of 95°C for 15 s and 55°C for 1 min.
C. coli and C. jejuni were used as positive controls for cadF; C. upsaliensis and S. flexneri as negative controls for cadF. C. coli, C. jejuni and C. upsaliensis were used as positive controls for 16S, and S. flexneri as negative control. S. flexneri served as the positive control for ipaH. Given the similar results obtained when we previously tested our positive and negative control samples in triplicate, the stool samples were not run in replicate; however, a subset of samples were run in duplicate for quality control purposes. C. coli, C. jejuni, C. upsaliensis and S. flexneri controls and negative controls–which included water only (no primers or master mix) and reactions with no DNA template–were included in every run. A cut-off cycle threshold (Ct) of 38 was used to determine positivity. Samples that were cadF positive were interpreted as positive for C. jejuni/C. coli (all cadF samples were also 16S positive), samples positive for Campylobacter 16S but negative for cadF were designated as positive for other Campylobacter species. The primers and probes used in this study are listed in Table 1.
Statistical analysis
Data analysis was performed using STATA 13.0 (College Station, Texas). Pearson’s chi-squared was used to test the difference between infection status, clinical outcome, age category, and the presence of blood and leukocytes. The difference in MPO levels by infection status and clinical outcome was tested using ANOVA. The odds of Campylobacter detection given clinical outcome (control, dysentery and severe diarrhea) were estimated using logistic regression models. We also calculated the odds of experiencing diarrhea (combining both dysentery and severe diarrhea) given infection status using a logistic regression model in order to compute the population attributable fraction (PAF) according to Blackwelder et al [21]. Age (broken down into interval of 6 months and treated as a categorical variable), sex, and Shigella infections were included as potential confounders based on a priori knowledge [22]. Panel data were analyzed accounting for the clustering at the child level with robust standard errors. Hypothesis testing was done using alpha error of 0.05.
Results
Study population characteristics
Of the 99 dysentery stool samples identified in the Peru MAL-ED cohort, 198 severe diarrhea and 198 control samples age-matched to the dysentery samples, 85, 173, and 181 samples were available for testing and analyzed in each group, respectively. These 439 stool samples were from 201 unique children. Approximately 60% of the samples were from male children; the mean age in each clinical group was 12 months (Table 2). Almost all the children aged 0 to 8 months old (n = 145) were breastfed: 96.2%, 98.1%, and 100.0% among the dysentery, severe diarrhea, and control group, respectively. As expected, the difference in sex, age and breastfeeding status among the three clinical groups was not statistically significant (Table 2).
Prevalence of Campylobacter and Shigella infections across clinical group by qPCR
The highest detection of Campylobacter was reported in the dysentery group –60 of the 85 samples tested positive–followed by the severe diarrhea (93 of 173) and asymptomatic groups (63 of 181). The prevalence of C. coli/jejuni was lower than that of other Campylobacter across all clinical groups. Among the Campylobacter-positive samples, 55.0%, 81.7%, and 88.9% were classified as other Campylobacter among the dysentery, severe diarrhea, and control groups, respectively. Overall, 76.4% of all the Campylobacter infections (n = 216) were other Campylobacter (n = 165).
Shigella infection was also more predominant in the dysentery group where 36.5% of the samples tested positive, compared to 17.9% in the severe diarrhea group and 11.1% in the asymptomatic group. Of the 439 stool samples, 43.1% (n = 189) were negative for infection with Shigella and Campylobacter while 10.9% (n = 48) were co-infected with both bacteria, most of which were Shigella and other Campylobacter co-infections. The infection status by clinical group is presented in Table 3. The relatively higher prevalence of other Campylobacter was corroborated by the culture-based results performed by our collaborators in Iquitos, Peru (Fig 1).
qPCR results showed a higher prevalence of other Campylobacter (non-C. coli/jejuni) than C. coli/jejuni across all clinical groups. It is important to highlight that the culture assay allowed the distinction between C. jejuni and all other hipppurate negative species of Campylobacter, with the latter including C. coli. The results obtained via culture showed a lower prevalence of all three bacteria (C. jejuni, other Campylobacter (non-C. jejuni), and Shigella) than the qPCR results, notably for the other Campylobacter, which is most likely inherent to the lower sensitivity of culture compared to qPCR and the low discriminatory power that culture allows.
Prevalence of Campylobacter and Shigella infection across age groups
Children aged 0–6 months were the least burdened by Campylobacter or Shigella as 29.7% of them tested positive for either bacteria, a lower percentage than the 76.9% observed among children over 18 months old. Across all age groups, the prevalence of other Campylobacter was higher than that of C. coli/jejuni and Shigella and followed an overall increasing pattern over time –19.8% among children aged 0–6 months and 52.7% among those over 18 months. C. coli/jejuni were least common among those over 18 months (2.6%) and most common in the >6-12-month group (19.4%). The pattern of Shigella prevalence across age groups mirrored that of other Campylobacter with the highest prevalence among those over 18 months (44.9%) and lowest among those 0–6 months (2.0%) (Fig 2).
The prevalence of all Campylobacter, other Campylobacter (non-C. coli/jejuni) and Shigella increased over time while that of C. coli/jejuni decreased overall. The isolation frequency for all the bacteria groups increased after 6 months and remained higher than the prevalence observed in the 0–6 month group, except for that of C. coli/jejuni among children over 18 months old which reached levels less than those among the infants.
Assessment of intestinal inflammation and permeability based on disease and infection status
To study the association between infection status, clinical outcome, and injury, we performed the fecal occult blood test in conjunction with the methylene blue test (Table 4). Children suffering from dysentery and those infected with C. coli/jejuni were more likely to have blood present in their stool. The difference across clinical groups for the fecal occult test was only significant among children infected with other Campylobacter as 10.7%, 51.5%, and 23.7% of these children tested positive in the control, dysentery, and severe diarrhea group, respectively.
A similar trend was observed with the methylene blue test used to detect leukocytes in the stools. Fecal leukocytes were detected primarily in the dysentery group and among children infected with C. coli/jejuni. Infection with C. coli/jejuni was associated with an equal likelihood of children having leukocytes present in their stools regardless of their disease status, which contrasts with other Campylobacter infections. Among children who were infected with other Campylobacter, 39.4% of those suffering from dysentery had leukocytes present in their stool compared to 10.7% among the asymptomatic children and 21.1% in the severe diarrhea group. As expected, most of the asymptomatic children tested negative for methylene blue whether they were free of any Campylobacter and Shigella, infected with other Campylobacter, or infected with Shigella (approximately 10%). This percentage was more than quadrupled when asymptomatic children harbored C. coli/jejuni infections (42.9%); this difference in percentage was not significant across clinical groups (51.9% and 41.2% in the dysentery and severe diarrhea group, respectively).
A biomarker of environmental enteropathy of interest was myeloperoxidase (MPO). MPO data were available for 415 samples and were missing completely at random (169/181 controls, 80/85 dysentery samples, and 166/173 severe diarrhea samples). As expected, high neutrophil activity correlated with dysentery and Campylobacter or Shigella infections (Table 4). We aggregated the data presented in Table 4 to compare mean MPO levels across clinical outcome without further breakdown based on infectious status, and mean MPO levels based on infectious status without considering the clinical groups. Dysentery was still associated with the highest MPO levels (23,163 ng/mL), which were significantly different from MPO levels in the control and severe diarrhea groups (12,333 and 11,970 ng/mL, respectively). When we compared the MPO levels between children infected with any Campylobacter species and those who were not, the difference was not statistically significant (15,099 and 13,510 ng/mL, respectively). More specifically, C. coli/jejuni were associated with significantly more inflammation (MPO levels of 19,909 ng/mL) than children who tested negative for Campylobacter (13,510 ng/mL) while the difference in MPO levels between children infected with other Campylobacter (13,538 ng/mL) and Campylobacter-free children was not significant. As expected, the highest MPO levels were noted in children who tested positive for Shigella (19,738 ng/mL) and were significantly different from the MPO levels of 13,071 ng/mL measured among those who tested negative.
Association of Campylobacter infection with clinical outcome
The odds of detecting Campylobacter in a stool sample given the clinical outcome (dysentery, severe diarrhea, control) were modelled separately for each group of Campylobacter species (C. coli/jejuni, other Campylobacter, and all Campylobacter) adjusting for age, sex, and Shigella infection. Compared to the asymptomatic group, the odds of detecting C. coli/jejuni were 14.6 among the dysentery samples (p-value <0.001, 95% CI: 5.5–38.7) and 2.8 (p-value = 0.034, 95% CI: 1.1–7.1) among the severe diarrhea samples. Using the same reference group, the odds of being infected with other Campylobacter were lower than those of C. coli/jejuni: children in the dysentery group were equally likely to have other Campylobacter in their stools compared to the controls (odds of 1.3, p-value = 0.434, 95% CI: 0.7–2.4) while children in the severe diarrhea group were 1.9 times more likely to have other Campylobacter in their stools (p-value = 0.018, 95% CI: 1.1–3.1).
The previous models allowed the comparison of the odds of infection with C. coli/jejuni (or other Campylobacter) in the entire study population (n = 439), i.e. the odds of being infected with C. coli/jejuni (or other Campylobacter) versus not being infected with C. coli/jejuni (or other Campylobacter), with the latter also including stools that were negative for both Campylobacter and Shigella. To compare the odds of infection with C. coli/jejuni to infection with other Campylobacter, another logistic regression model was used, restricting the sample size to the stools that were positive for Campylobacter spp. (n = 216). Based on this new model, the odds of C. coli/jejuni infection were 6.6 among the dysentery samples (p-value = 0.006, 95% CI: 1.7–24.9), and 1.8 among the severe diarrhea samples (p-value = 0.298, 95% CI: 0.6–5.7) compared to the controls. This demonstrates that C. coli/jejuni infections were more likely than other Campylobacter infections to be detected among the dysentery samples, but C. coli/jejuni and other Campylobacter were equally likely to be found among severe diarrhea cases.
The overall odds of Campylobacter infection (both C. coli/jejuni and other Campylobacter infections) in the study population were significantly increased among the dysentery samples (OR = 4.9, p-value < 0.001, 95% CI: 2.5–9.8) and severe diarrhea samples (OR = 2.4, p-value = 0.001, 95% CI: 1.4–4.0) compared to controls. The odds of Shigella infection were also increased in the dysentery samples (OR = 6.4, p-value < 0.001, 95% CI: 2.4–17.3), but the increased odds of 0.9 were not significant in the severe diarrhea samples. There was no significant association between C. coli/jejuni infection and age, but the increased odds of other Campylobacter and Shigella infections were statistically significant across age groups, with infections being more prevalent among older children, supporting the results from the descriptive analysis.
Population attributable fraction (PAF)
To calculate the PAF, the association between diarrhea and infection status was investigated, combining and categorizing the dysentery and severe diarrhea groups as diarrhea (Table 5). Adjusting for sex, age, and Shigella infection, children infected with C. coli/jejuni were 8.8 times more likely to have diarrhea compared to those who tested negative for Campylobacter (p-value <0.001, 95% CI: 3.0–25.7). Using the same reference group, infection with other Campylobacter increased the odds of diarrhea to 2.4 (p-value = 0.002, 95% CI: 1.4–4.2). In other words, children with C. coli/jejuni-positive stools were approximately three times more likely to have diarrhea than children infected with other Campylobacter species. Adjusting for sex, age, and Campylobacter infection, children infected with Shigella were approximately three times more likely to have diarrhea (OR = 3.3, p-value = 0.003, 95% CI: 1.5–7.3).
With other Campylobacter detected in 42.3% of the diarrhea samples and odds of diarrhea of 2.4, approximately a quarter of all diarrhea cases in this study (24.9%, 95% CI: 12.1–32.2%) was attributed to other Campylobacter; this PAF was higher than that of C. coli/jejuni. Indeed, eliminating C. coli/jejuni would only eliminate 15.1% of the diarrhea cases (95% CI: 11.4–16.4%) considering their low prevalence among the diarrhea samples (0.2), despite the odds of diarrhea of 8.8. Overall, 40.8% of the diarrhea cases (95% CI: 27.1–48.6%) could be attributed to Campylobacter (prevalence of 0.6), which is more than double the PAF of 16.7% (95% CI: 7.8–20.7%) due to Shigella (prevalence of 0.2).
Discussion
Our study of the epidemiology of Campylobacter among children in peri-urban communities in the Peruvian Amazon underscores the importance of other Campylobacter species (i.e. non-C. coli/jejuni) in the etiology of diarrhea. Other Campylobacter had the highest attributable fraction for childhood diarrhea (24.9%) and accounted for more infections than C. coli/jejuni in all three clinical groups (asymptomatic, dysentery, and severe diarrhea), comprising more than half of all Campylobacter infections (165 of the 216 Campylobacter detections). Although C. coli/jejuni were more strongly associated with dysentery, other Campylobacter were equally likely as C. coli/jejuni to be detected in severe diarrhea cases. The lower percentage of other Campylobacter detected using culture methods is most likely due to the lower sensitivity of this diagnostic tool optimized for the detection of C. jejuni, and the higher sensitivity of qPCR compared to culture [4,23]. The higher prevalence of other Campylobacter compared to that of C. coli/jejuni is at odds with some of the published literature which has shown that C. coli/jejuni are responsible for more than 80% of all Campylobacter infections and are the only species of major public health importance [2,3]. However, it is important to highlight that the prevalence of C. coli/jejuni does fall within the reported range in developing countries [4,24], suggesting that the discrepancy between our results and published literature stems from the fact that we were detecting more of the non-C. coli/jejuni species [3]. Additionally, others have also reported instances where more than 50% of the Campylobacter isolated were non-C. coli/jejuni species [25]. As more research is being conducted in developing countries using diagnostic techniques that detect other Campylobacter species more reliably, the importance of non-C. coli/jejuni species has been raised [3,4,25,26]. Discordance analysis of BioFire rapid diagnostic system also reported a high amount of culture-negative but probe-positive specimens in a study conducted in different clinics across the USA (58 positive using BioFire compared to 35 positive using culture) [11], suggesting the other Campylobacter may not be entirely restricted to impoverished settings.
Our study also found evidence that the detection of other Campylobacter was more frequent than C. coli/jejuni among children across all age groups and increased with age while that of C. coli/jejuni decreased in the second year of life. Despite the overall increase in diarrhea and infection with all Campylobacter, other Campylobacter, and Shigella from 0 to 18 months, the illness-to-infection ratio decreased over time during the first 18 months. This trend, not observed with C. coli/jejuni, suggests a shift towards more asymptomatic infections which may be explained by the natural immunity developed after early exposure to these bacteria [3,17,24]. Why such trend is not observed with C. coli/jejuni remains unclear and requires additional investigation. The surge in the illness-to-infection ratio observed among the children aged 18–32 months might be due to the relatively small number of children in that age group, and requires additional investigation with a larger study population to validate these findings.
Some of our results about the association between MPO and clinical outcome were unexpected. Since Shigella and Campylobacter are known pathogens that cause inflammatory diarrhea [18], a positive association was expected between the level of MPO and infection with any of these pathogens. Our finding of no significant association between MPO and infection with any Campylobacter species is at odds with McCormick et al.’s study that reported a consistent and positive association between MPO and Campylobacter which had data from all eight MAL-ED sites and greater power to detect associations [7,27]. However, it is important to highlight that we did observe a significant positive association between MPO and C. coli/jejuni compared to the negative-Campylobacter group. Despite this positive association, our overall results of no association between MPO and infection with any Campylobacter species seem to be driven by the lack of significant association between MPO and other Campylobacter, which accounted for most of the Campylobacter infections.
A limitation of our study was a relatively low number of dysentery and severe diarrhea cases, as this study was a community-based study, and most episodes in such cases are mild to moderate. Another limitation was the focus on only two pathogens, Campylobacter and Shigella. Other microbes most likely contributed to symptomatic infections, especially in the severe diarrhea group. Other diarrhea-causing agents that could be added to the assay include enterohemorrhagic E. coli and Salmonella, as well as Yersinia and E. histolytica as alternate causes of bloody and severe diarrhea [28]. These pathogens would not be expected to be very prevalent relative to Shigella and Campylobacter and indeed have been evaluated in this population (10). However, this more inclusive assay would be useful for the evaluation and treatment of individual patients in other epidemiologic settings. Of note, enteroinvasive E. coli would have been classified as Shigella with our PCR assay and is implicitly contained in the data presented. Despite these limitations, our study showed a clear association between clinical outcome and specific Campylobacter species. This finding is consistent with other studies that reported that the highest burden of diarrhea in Loreto, Peru and Venda, South Africa was associated with Campylobacter in the first year of life, and that Campylobacter is the most frequently isolated pathogen in Loreto, Peru [4]. The very few C. coli/jejuni infections detected among children aged 18–32 months (n = 2) might be an artifact of the rather small number of samples from children in that age group. Most importantly, the burden of other Campylobacter was most likely underestimated in our study which defined other Campylobacter as positive for 16S rRNA but negative for cadF. Since infections with C. coli/jejuni and other Campylobacter are not mutually exclusive, it is possible that some of the C. coli/jejuni-positive samples were also positive for other Campylobacter. A comparison of the cadF and 16S rRNA Ct values in the C. coli/jejuni group revealed that 50 out of the 51 samples had a Ct difference greater than 8. Given the average three copies of 16S rRNA per C. coli and C. jejuni genome [29–31], the burden and impact of other Campylobacter is most likely greater than reported. A more refined assessment of the burden of these other Campylobacter species will require sequencing or the development of assays that detect gene sequences specific to these species, which would strengthen our findings. Lastly, the importance of co-infections and how to attribute disease burden in this and other studies of enteric disease in the setting of comprehensive diagnostics remains a challenge, but the random selection of samples from the collection makes alternative co-infections unlikely to undermine study findings.
Overall, our study highlights the importance of other Campylobacter in the etiology of diarrheal disease and contributes to filling the knowledge gap about the epidemiology of Campylobacter infection in developing countries. Although C. coli/jejuni are important causes of dysentery, the severity of disease caused by non-C. coli/jejuni Campylobacter species and the frequency with which they are detected combined with a greater than expected association with moderate to severe diarrhea strongly suggest that they are the cause of an important fraction of the Campylobacter burden in children in the developing world. Further work is underway to determine which of these other Campylobacter species are most prevalent and strongly associated with both clinical disease, enteropathy and the acquisition of linear growth deficits in this population.
Acknowledgments
We particularly want to thank the children and caregivers who participated in the study and fieldworkers for their tireless efforts to make these children’s early life exposures to infectious diseases and the long-term impact of these infections visible to the world.
References
- 1. Kaakoush NO, Castano-Rodriguez N, Mitchell HM, Man SM. Global epidemiology of campylobacter infection. Clin Microbiol Rev. 2015;28(3):687–720. pmid:26062576
- 2. Carattoli A, Dionisi AM, Luzzi I. Use of a LightCycler gyrA mutation assay for identification of ciprofloxacin-resistant Campylobacter coli. FEMS Microbiol Lett. 2002;214(1):87–93. pmid:12204377
- 3. Coker AO, Isokpehi RD, Thomas BN, Amisu KO, Larry Obi C. Human campylobacteriosis in developing countries. Vol. 8, Emerging Infectious Diseases. 2002. p. 237–43. pmid:11927019
- 4. Platts-Mills JA, Kosek M. Update on the burden of Campylobacter in developing countries. Curr Opin Infect Dis [Internet]. 2014;27(5):444–50. Available from: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=4749027&tool=pmcentrez&rendertype=abstract%5Cn http://content.wkhealth.com/linkback/openurl?sid=WKPTLP:landingpage&an=00001432-201410000-00008 pmid:25023741
- 5. Lee G, Paredes Olortegui M, Peñataro Yori P, Black RE, Caulfield L, Banda Chavez C, et al. Effects of Shigella-, Campylobacter- and ETEC-associated Diarrhea on Childhood Growth. Pediatr Infect Dis J. 2014;33(10):1004–9. pmid:25361185
- 6. Platts-Mills J a., Liu J, Gratz J, Mduma E, Amour C, Swai N, et al. Detection of Campylobacter in stool and determination of significance by culture, enzyme immunoassay, and PCR in developing countries. J Clin Microbiol. 2014;52(4):1074–80. pmid:24452175
- 7. Amour C, Gratz J, Mduma E, Svensen E, Rogawski ET, McGrath M, et al. Epidemiology and Impact of Campylobacter Infection in Children in 8 Low-Resource Settings: Results from the MAL-ED Study. Clin Infect Dis. 2016;63(9):1171–9. pmid:27501842
- 8. Black RE, Victora CG, Walker SP, Bhutta ZA, Christian P, De Onis M, et al. Maternal and child undernutrition and overweight in low-income and middle-income countries. Lancet. 2013;382(9890):427–51. pmid:23746772
- 9. Kotloff KL, Nataro JP, Blackwelder WC, Nasrin D, Farag TH, Panchalingam S, et al. Burden and aetiology of diarrhoeal disease in infants and young children in developing countries (the Global Enteric Multicenter Study, GEMS): A prospective, case-control study. Lancet. 2013;382(9888):209–22. pmid:23680352
- 10. Platts-Mills JA, Babji S, Bodhidatta L, Gratz J, Haque R, Havt A, et al. Pathogen-specific burdens of community diarrhoea in developing countries: A multisite birth cohort study (MAL-ED). The Lancet Global Health. 2015;
- 11. Buss SN, Leber A, Chapin K, Fey PD, Bankowski MJ, Jones MK, et al. Multicenter evaluation of the BioFire FilmArray gastrointestinal panel for etiologic diagnosis of infectious gastroenteritis. J Clin Microbiol. 2015;53(3):915–25. pmid:25588652
- 12. Richard SA, Barrett LJ, Guerrant RL, Checkley W, Miller MA. Disease surveillance methods used in the 8-site MAL-ED cohort study. Clin Infect Dis. 2014;59:S220–4. pmid:25305290
- 13. Lee G, Peñataro Yori P, Paredes Olortegui M, Caulfield LE, Sack D a, Fischer-Walker C, et al. An instrument for the assessment of diarrhoeal severity based on a longitudinal community-based study. BMJ Open [Internet]. 2014;4(6):e004816. Available from: http://www.ncbi.nlm.nih.gov/pubmed/24907244 pmid:24907244
- 14. Lee GO, Richard SA, Kang G, Houpt ER, Seidman JC, Pendergast LL, et al. A Comparison of Diarrheal Severity Scores in the MAL-ED Multisite Community-Based Cohort Study. J Pediatr Gastroenterol Nutr. 2016;63(5):466–73. pmid:27347723
- 15. Savola KL, Baron EJ, Tompkins LS, Passaro DJ. Fecal leukocyte stain has diagnostic value for outpatients but not inpatients. J Clin Microbiol. 2001;39(1):266–9. pmid:11136781
- 16. Kosek M, Haque R, Lima A, Babji S, Shrestha S, Qureshi S, et al. Fecal markers of intestinal inflammation and permeability associated with the subsequent acquisition of linear growth deficits in infants. Am J Trop Med Hyg. 2013;88(2):390–6. pmid:23185075
- 17. Kirkpatrick BD, Tribble DR. Update on human Campylobacter jejuni infections. Curr Opin Gastroenterol [Internet]. 2011;27(1):1–7. Available from: http://www.ncbi.nlm.nih.gov/pubmed/21124212 pmid:21124212
- 18. Navaneethan U, Giannella R a. Mechanisms of infectious diarrhea. Nat Clin Pract Gastroenterol Hepatol [Internet]. 2008;5(11):637–47. Available from: http://www.ncbi.nlm.nih.gov/pubmed/18813221 pmid:18813221
- 19. Botteldoorn N, Van Coillie E, Piessens V, Rasschaert G, Debruyne L, Heyndrickx M, et al. Quantification of Campylobacter spp. in chicken carcass rinse by real-time PCR. J Appl Microbiol. 2008;105(6):1909–18. pmid:19016974
- 20. Thiem VD, Sethabutr O, Von Seidlein L, Van Tung T, Canh DG, Chien BT, et al. Detection of Shigella by a PCR Assay Targeting the ipaH Gene Suggests Increased Prevalence of Shigellosis in Nha Trang, Vietnam. J Clin Microbiol. 2004;42(5):2031–5. pmid:15131166
- 21. Blackwelder WC, Biswas K, Wu Y, Kotloff KL, Farag TH, Nasrin D, et al. Statistical methods in the Global Enteric Multicenter Study (GEMS). Clin Infect Dis. 2012;55(SUPPL. 4).
- 22. Singh P, Manning SD. Impact of age and sex on the composition and abundance of the intestinal microbiota in individuals with and without enteric infections. Ann Epidemiol. 2016;26(5):380–5. pmid:27118420
- 23. Liu J, Platts-Mills JA, Juma J, Kabir F, Nkeze J, Okoi C, et al. Use of quantitative molecular diagnostic methods to identify causes of diarrhoea in children: a reanalysis of the GEMS case-control study. Lancet [Internet]. 2016 Sep [cited 2017 Mar 16];388(10051):1291–301. Available from: http://linkinghub.elsevier.com/retrieve/pii/S014067361631529X pmid:27673470
- 24. Calva J, Lopez-Vidal A, Ruiz-Palacios G, Ramos A, Bojalil R. COHORT STUDY OF INTESTINAL INFECTION WITH CAMPYLOBACTER IN MEXICAN CHILDREN. Lancet. 1988;331(8584):503–6.
- 25. Man SM. The clinical importance of emerging Campylobacter species. Nat Rev Gastroenterol Hepatol [Internet]. 2011;8(12):669–85. Available from: http://www.nature.com/nrgastro/journal/v8/n12/pdf/nrgastro.2011.191.pdf pmid:22025030
- 26.
Allos BM, Lastovica AJ. Clinical Significance of Campylobacter and Related Species Other Than Campylobacter jejuni and Campylobacter coli. In: Campylobacter, Third Edition [Internet]. American Society of Microbiology; 2008 [cited 2017 May 14]. p. 123–49. Available from: http://www.asmscience.org/content/book/10.1128/9781555815554.ch07
- 27. McCormick BJJ, Lee GO, Seidman JC, Haque R, Mondal D, Quetz J, et al. Dynamics and Trends in Fecal Biomarkers of Gut Function in Children from 1–24 Months in the MAL-ED Study. Am J Trop Med Hyg [Internet]. 2017 Feb 8;96(2):465–72. Available from: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC5303054/ pmid:27994110
- 28. Barletta F, Mercado EH, Lluque A, Ruiz J, Cleary TG, Ochoa TJ. Multiplex real-time PCR for detection of Campylobacter, Salmonella, and Shigella. J Clin Microbiol. 2013;51(9):2822–9. pmid:23761159
- 29. Taylor DE, Eaton M, Yan W, Chang N. Genome maps of Campylobacter jejuni and Campylobacter coli. J Bacteriol. 1992;174(7):2332–7. pmid:1313002
- 30. Inglis GD, Kalischuk LD. Direct Quantification of Campylobacter jejuni and Campylobacter lanienae in Feces of Cattle by Real-Time Quantitative PCR. Appl Environ Microbiol. 2004;70(4):2296–306. pmid:15066825
- 31. Hansson I, Persson M, Svensson L, Engvall E, Johansson K-E. Identification of nine sequence types of the 16S rRNA genes of Campylobacter jejuni subsp. jejuni isolated from broilers. Acta Vet Scand [Internet]. 2008;50:10. Available from: http://dx.doi.org/10.1186/1751-0147-50-10 pmid:18492293