Characterization of Fecal Microbiota of Children With Diarrhea in 2 Locations in Colombia

Solano-Aguilar, Gloria*; Fernandez, Karem P.; Ets, Hillevi*; Molokin, Aleksey*; Vinyard, Bryan; Urban, Joseph F.*; Gutierrez, Maria Fernanda

Journal of Pediatric Gastroenterology & Nutrition: May 2013 - Volume 56 - Issue 5 - p 503–511
doi: 10.1097/MPG.0b013e318282aa12
Original Articles: Gastroenterology

Objectives: Diarrhea is a leading cause of mortality and morbidity in children younger than 5 years in impoverished regions of the world. Our aim was to compare the fecal microbiota of healthy children with that of children with clinical diarrhea in a population from a tropical highland in Colombia, South America. Our hypothesis was that a reduced prevalence of inherent Lactobacillus and Bifidobacterium species would be associated with enteric viral and bacterial pathogens.

Methods: Children between 1 and 5 years of age from 2 different locations were evaluated for presence of clinical diarrhea. Nucleic acid, isolated from fecal samples, was used to determine by molecular protocols the abundance of inherent bacterial species and presence of enteric pathogens compared with clinically healthy children. The effect of host demographic factors on incidence of diarrhea was also analyzed.

Results: The composition of the fecal microbiota was affected by host demographic factors: age, health status, location, and sex. In partial support of our hypothesis, the relative abundance of commensal Bifidobacterium and Lactobacillus species was inversely correlated with incidence of diarrhea regardless of location.

Conclusions: Our results suggested that changes in fecal microbiota composition of children with clinical diarrhea are associated with certain demographic factors that should be considered before designing a prophylactic intervention. Delivery of certain Lactobacillus species and Bifidobacterium species or a diet rich in bifidogenic components that promote growth of Bifidobacterium species could provide a prophylactic effect to ameliorate the effect of diarrhea in children at risk.

*US Department of Agriculture, Agricultural Research Service, Beltsville Human Nutrition Research Center, Diet, Genomics, and Immunology Laboratory, Beltsville, MD

Facultad de Ciencias, Departamento de Microbiologia Pontificia Universidad Javeriana, Bogota, Colombia

US Department of Agriculture, Henry Wallace Beltsville Agricultural Research Center, Biometrical Consulting Service, Beltsville, MD.

Address correspondence and reprint requests to Gloria Solano-Aguilar, DVM, PhD, Diet Genomics and Immunology Laboratory, Beltsville Human Nutrition Research Center, 10300 Baltimore Ave. BARC-East Bldg 307C-Rm 225, Beltsville, MD 20705 (e-mail:

Received 16 August, 2011

Accepted 11 December, 2012

This study was co-funded by USDA CRIS 1235-51000-055 and Banco de la Republica (Colombia) Project No. 2391. Mention of trade names or commercial products in this publication is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the US Department of Agriculture.

The authors report no conflicts of interest.

Article Outline

Enteric and diarrheal diseases are a leading cause of mortality and morbidity in children younger than 5 years worldwide (1). Despite significant improvement in interventions to treat diarrhea, children experience an average of 3 diarrhea episodes per year, and mortality can be between 1.5 and 2.5 million deaths annually (2). In Colombia, as in many other parts of the world, several reports showed that bacteria, parasites, and viruses are involved in the etiology of diarrhea (3–7); however, for >2 decades, rotavirus (RV) has been the most prevalent agent associated with acute gastroenteritis in Colombian children, with climate-related variation in detection rates exacerbated by poor health services and higher levels of malnutrition (3,6,8). Available prophylactic guidelines including RV vaccination are specific to the locality where they are applied (9). In Colombia, RV vaccination has recently been introduced and cost-effective protection against prevalent serotypes is presently under evaluation (10).

The intestinal microbiota plays an important role in human health by supporting a barrier for colonization of pathogens, providing important metabolic functions, and inducing development and activation of the immune system (11). Bacteroides spp, the most predominant anaerobe in the intestine, maintains a generally beneficial relation with the host (12), but can also be an important pathogen (13). The beneficial health effects attributed to endogenous intestinal bacteria from the genera Lactobacillus and Bifidobacterium are supported by the frequent use of species within these genera that are fed as probiotics to promote health. Both genera are among the first bacteria to colonize the human intestine after birth and decrease in number in adulthood (14). Species within these genera contribute to resistance against pathogens and increase availability of nutrients in the intestine (15). Although the precise mechanisms behind these beneficial effects are largely unknown, bacterial cell surface–associated structures and the interaction of extracellular products with epithelial and mucosal immune cells are considered important (15,16). Delivery of dietary probiotics has been hypothesized as a strategy to prevent diarrhea (17) or as an adjuvant to improve vaccine efficacy (18). Several meta-analyses have suggested that Lactobacillus and Bifidobacterium species may be effective in prevention of gastroenteritis (17,19–21); however, randomized controlled trials conducted mostly in childcare centers from affluent countries have indicated only a modest benefit in preventing acute gastrointestinal infections in healthy infants and children (22–24). Limited and variable outcomes from community-based trials in developing countries suggested some benefit when certain Lactobacillus species were used to prevent the onset of diarrhea (21,25,26). It has been suggested that the variable effectiveness of probiotics can be partially explained by differences in probiotic strain effects, the etiology of diarrhea, level of exposure, host age, diet, and availability of health care in different geographic locations (21,23,25–27). There is presently no information on the species diversity or numbers of Lactobacillus and Bifidobacterium in the intestinal microbiota of children from a tropical setting. An initial step to guide the delivery of probiotics as a prophylactic intervention is to determine the changes in intestinal microbiota that are associated with diarrhea. We hypothesize that commensal Lactobacillus and Bifidobacterium species are differentially affected in children with clinical diarrhea compared with healthy children. This hypothesis was tested in a tropical location where viral and bacterial diarrheas are endemic using a multifactorial analysis of variables associated with diarrhea in a group of children from 2 highland locations in Colombia that have similar socioeconomic levels.

Back to Top | Article Outline


Subjects and Design

A total of 277 Colombian children (1–5 years of age) from 2 locations in the north or southwest of Metropolitan Bogota City were recruited to participate in an observational, descriptive cohort study to evaluate the composition of their fecal microbiota. Children who attended a daycare center or sought medical attention at a local hospital in the northern locality of Chia (location A) and children who attended a local day care center in the southwestern locality of Cuidad Bolivar (location B) were studied from February to August 2009. Children from both locations belong to a similar socioeconomic level with access to basic health and public services, including water and sewage. All caretakers of the children were informed about the study, asked whether they wanted to participate, and, if so, requested to give their written informed consent to provide a fresh fecal sample from each child. The human ethics committee from the Sciences Faculty and Department of Microbiology of Universidad Javeriana in Bogota, Colombia, reviewed and approved the present study. Each caretaker was surveyed about the child's health and nutritional history through questionnaires performed by health professionals to determine clinical diarrhea. The criteria stated by the World Health Organization for diarrhea as liquid or reduced consistency of the stool with increased frequency of >3 times in 24 hours for a period <2 weeks were used for inclusion of children in the diarrhea group. Healthy children from these locations without any clinical signs of chronic disease, vomiting, diarrhea, or fever and with a normal stool consistency and deposition for at least 2 weeks were included in the healthy children group. Exclusion criteria included children that were exclusively breast-fed, had consumed antibiotics up to 1 month before the initiation of the study, or had been hospitalized before collection of the sample. The sample size was powered using data from 2 previous studies where RV prevalence was detected at 13% and 21.4% (28). Based on a binomial exact 95% confidence interval (CI), a sample of 237 children would ensure that the true RV prevalence would fall within these bounds. A total of 5 to 10 g of a recently produced fecal sample was collected from children and immediately refrigerated at 4°C and transported to the Virology Laboratory at the Universidad Javeriana. Upon receipt, samples were separated into 3 identical aliquots. One aliquot was used for detection of Enterobacteriaceae by microbiological culture, and the remaining 2 aliquots were frozen at −80°C until processed for nucleic acid extraction to detect either bacterial DNA or viral RNA, respectively.

Back to Top | Article Outline

Microbiological Isolation

Aliquots for microbiological culture were homogenized, diluted, and enriched in Selenite broth (Difco Laboratories, Detroit, MI) before selective Shigella spp and Salmonella spp growth in XLD or McConkey media (Becton, Dickinson, Franklin Lakes, NJ) for 24 hours at 37°C. Colonies grown were further identified by biochemical tests: triple sugar iron, sulfide motility indol, lysine iron agar, phenylalanine, citrate, urea, and methyl red/Voges Proskauer for the identification of the most common Enterobacteriaceae (29). Positive Salmonella isolates were confirmed for detection of Salmonella-specific O-antigen.

Back to Top | Article Outline

Nucleic Acid Extraction

DNA from feces was isolated using the QIAamp DNA stool minikit (Qiagen, Valencia, CA). Total RNA was extracted from stool suspensions using NucliSENS EasyQ Basic Kit (Biomérieux Inc, Clinical Diagnostics, Marcy l’Etoile, France) (30). A 2-μL aliquot of 10 ng of DNA per microliter or 10 μL RNA (with at least 200 ng of RNA) from each extraction procedure was used as a template for all bacterial and viral quantifications.

Back to Top | Article Outline

Bacterial Detection

The amount of bacterial copy numbers in fecal samples was determined by real-time polymerase chain reaction (PCR) using previously validated primers and probe sets against 16S ribosomal RNA (rRNA) sequences of Lactobacillus and Bifidobacterium species (31–33), Escherichia coli(34), Bacteroides fragilis group (35)Clostridium difficile(34), or specific assays for enteropathogenic E coli(36) and enterotoxigenic B fragilis-ETBF isotypes (bft1-bft3) often described to be responsible for acute diarrhea (37). For determination of the total bacterial load, a set of primers and probe for the 16S rRNA of the domain bacteria were used (31–38) (Table 1). Taqman technology determines the PCR cycle at which the increase in fluorescence reaches a threshold cycle (Ct value), which is proportional to the log of the amount of target DNA (amplicon) (39). A bacterial strain representative for each bacterial species was used as a positive amplification control (Table 1). The base pair length and molecular mass of each amplicon was measured by automated electrophoresis system (DNAchip, Experion, Biorad, Hercules, CA) after purification of DNA fragment with Qiaquick PCR purification kit (Qiagen, Gaithersburg, MD) following the manufacturer's recommendation (40). A linear relation was established between the Ct value and number of DNA fragments ranging between 101 and 1010 copies/μL and this relation was subsequently used to estimate values of log10 fragment copy numbers per 1 g of fecal sample. In all assays used, the amplification efficiency was >90% and the standard curve showed a linear range across at least 5 logs of DNA fragment concentrations with a correlation coefficient >0.9. The lowest detection limit of all assays was between 10 and 100 copies of specific bacterial fragment per reaction.

Back to Top | Article Outline

Viral Detection

RNA extracted from fecal samples was used as a template to determine the presence (positive signal) or absence (no amplification) of norovirus (NV) or RV. Positive diluted controls and unknown samples were included in the real-time PCR assay with specific primers against the viral capside of NV (41) and the structural viral glycoprotein of RV (42).

Back to Top | Article Outline

Statistical Analysis

Differences in bacterial species abundance, determined by Ct values, were converted to values of log10 copy numbers units per gram of feces for comparisons among the groups (diarrhea vs healthy). One-way analysis of variance (ANOVA) was conducted to determine relative differences in bacterial abundance (mean log copy numbers ± SEM) among locations × health status groups. Contrast statements were run to compare the effect of disease (diarrhea or healthy) on different bacterial species by location.

Pathogen prevalence was compared between locations, separately for healthy children and for children with diarrhea, by calculating OR and 95% CI fitting a 2-way (health status × location) generalized linear ANOVA model using a binomial distribution and a logit link to fit pathogen presence/absence.

Three demographic factors (location, health status, and sex) and 3 demographic covariates (age, months with breast-feeding as supplement to diet, and yogurt consumption) were measured in the present study. Analysis of covariance (ANCOVA) models were fit to the data, observed for each pathogen and each bacterium, to examine relationships between diarrhea presence or absence (Y) and bacteria levels present (X); pathogen presence or absence (Y) and bacteria levels present (X); pairs of bacteria levels present; pathogen presence or absence (Y) and demographic covariates (X); and bacteria levels present (Y) and demographic covariates (X). To model the relation between Y and X, each ANCOVA model was initially specified to include all 3 demographic factors (location, health status, and sex), 1 covariate (X), and all interactions among these factors and covariate. In turn, the effect with the largest (ie, least significant) P value was removed from the model, until only significant (P < 0.05) effects remained. Relations between diarrhea (Y) and bacteria (X) or between pathogens (Y) and bacteria (X) or between pathogens (Y) and demographic covariates (X) were modeled in the ANCOVA using logistic regression: Y = exp (intercept + slope × X)/(1 + exp [intercept + slope × X]); all other relations were linear. Odds ratios and confidence intervals were obtained by specifying, in the SAS PROC GLIMMIX model statement, the odds ratio option and by setting the value of X at the integer closest to the average of all observed X. A generalized linear ANOVA model with binary distribution and logit link function was used to determine association between enteropathogenic B fragilis-ETBF and other bacterial species or enteric pathogens. All the modeling and statistical analyses were accomplished using SAS PROC GLIMMIX (SAS Institute Inc, Cary, NC).

Back to Top | Article Outline


After verification of exclusion criteria, 277 children (154 girls and 123 boys) were enrolled in the study (Table 2). Of the 137 and 140 children monitored from locations A and B, respectively, 95% CI for prevalence of diarrhea was 53.3% to 70.2% for location A and 48.4% to 65.3% for location B. Combining all 277 children from both locations, the binomial exact 95% CI for prevalence of diarrhea was 53.6% to 65.4%. For RV, there were 16 clinical cases (11.67%) of diarrhea in location A, whereas 3 (2.2 %) and 4 (2.85%) cases, respectively, came from asymptomatic children in locations A and B. Ten positive NV samples (7.14%) were detected in children with clinical diarrhea in location B and 5 positive samples (3.64%) from location A. Three (2%) additional positive NV samples came from asymptomatic children from location B. For Enterobacter isolates, children with clinical signs of diarrhea from location B had 4 times more cases (14.28%) compared with children from location A (3.65%). Microbiological isolation of Salmonella spp confirmed by serology showed 2 cases (1.43%) at location B and 1 (0.72%) asymptomatic case at location A. Enteropathogenic E coli was confirmed from clinical cases of diarrhea (10.21%) and from asymptomatic healthy children (5.84%) at location A and location B (7.14% for clinical diarrhea and 7.85% for healthy asymptomatic children). Enterotoxigenic B fragilis-ETBF from the bft-1 isotype only were confirmed from clinical cases of diarrhea (8.57%) and from asymptomatic healthy children (4.28%) at location B and only for healthy asymptomatic children from location A (2.18%). There were significantly more cases of C difficile in children with clinical signs of diarrhea in location A (8.76%) compared with children from location B (1.42%) (Table 2).

Copy numbers for total bacteria were reduced in children with diarrhea only in location A (P < 0.05) (Table 3). There was a significant reduction in the abundance of Bifidobacterium spp copy numbers in children with clinical signs of diarrhea compared with healthy asymptomatic children in location A (P < 0.05) with no change in abundance in children from location B. Lactobacillus spp were detected with 2 separate sets of primers/probes with different specificities (31,33). There was no significant change detected in total abundance of Lactobacillus spp copy numbers identified with intergenic Lactobacillus16–23IS rDNA assay in all children regardless of their health status; however, when the Lactobacillus spp assay was used, there was a significant reduction in copy numbers from these Lactobacillus spp in children with diarrhea for both locations compared with healthy asymptomatic children (P < 0.05). The relative abundance of E coli was only affected in location A, where children with clinical diarrhea had increased abundance of E coli copy numbers compared with healthy asymptomatic children (P < 0.05). The abundance of B fragilis group copy numbers was significantly reduced in children with diarrhea compared with healthy asymptomatic children for location A (P < 0.05) and increased for location B (P < 0.0001). The abundance of Bifidobacterium adolescentis copies was reduced (P < 0.05) in children with diarrhea compared with healthy asymptomatic children in location A, but not in location B. No changes in the abundance of B infantis copies were detected in either location (Table 3).

Bifidobacterium spp abundance in fecal samples combined from both locations A and B was inversely related to clinical cases of diarrhea, for boys (P = 0.0425) and for girls (P = 0.0363) with 1 log-standard-unit change in Bifidobacterium spp (from 9 to 10), indicating a proportion of clinical cases of diarrhea with lower Bifidobacterium spp level (odds ratio 0.80 in boys, 0.51 in girls; 95% CI (0.65–0.99) boys and (0.35–0.74) girls (Fig. 1). Lactobacillus spp abundance was also inversely related to clinical cases of diarrhea (P = 0.0067) with 1 log-standard-unit change in Lactobacillus (from 4 to 5), indicating a proportion of clinical cases of diarrhea 0.70 that of the proportion at the lower Lactobacillus level (odds ratio 0.70; 95% CI 0.54–0.91) (Fig. 2). B adolescentis was inversely related to clinical cases of diarrhea (P < 0.0001) with 1 log-standard-unit change in B adolescentis (from 4 to 5), indicating a proportion of clinical cases of diarrhea 0.71 that of the proportion at the lower B adolescentis level (odds ratio 0.71; 95% CI 0.62–0.81). B infantis was inversely related to clinical cases of diarrhea only in location A (P < 0.0001) with 1 log-standard-unit change in B infantis (from 5 to 6), indicating a proportion of clinical cases of diarrhea 0.69 that of the proportion at the lower B infantis level (odds ratio 0.69; 95% CI 0.57–0.83). For the B fragilis group, there was an inverse relation between probability of diarrhea and log10 copies per gram of B fragilis (P < 0.0001). Increasing from 7 to 8, 1 standard log copies per gram unit, diarrhea becomes 0.57 as prevalent (odds ratio 0.57; 95% CI 0.46–0.70).

E coli abundance was inversely related to clinical cases of diarrhea in location B (P = 0.0035) with 1 log-standard-unit change in E coli (from 9 to 10), indicating a proportion of clinical cases of diarrhea 0.46 that of the proportion at the lower E coli level (odds ratio 0.46; 95% CI 0.27–0.77) and with a nonsignificant change at location A (P = 0.2942).

Bifidobacterium spp and Lactobacillus-IS spp were positively correlated in healthy children from location A (r = 0.3; P = 0.03, n = 60), but not from location B (r = 0.03, P = 0.822, n = 52) with no significant correlation in sick children regardless of location. This significant correlation was also found to be sex dependent (r = 0.35, P = 0.01) because overall healthy asymptomatic boys showed a stronger positive correlation than healthy asymptomatic girls, with no effect on children with diarrhea (Table 4). Bifidobacterium spp and Lactobacillus spp were also positively correlated in healthy (r = 0.44, P = 0.001) and sick children from location A (r = 0.31, P = 0.01), with a stronger overall positive correlation in healthy boys (r = 0.47, P < 0.001) (Table 4).

Bifidobacterium spp and the B fragilis group showed an overall positive correlation (r = 0.31, P < 0.0001, n = 242) with a positive correlation observed in both healthy (r = 0.32, P < 0.05, n = 54) and sick (r = 0.29, P < 0.05, n = 65) boys (Table 4).

Logistic regression models involving demographic factors were fit to data on bacterial abundance and virus detection. The abundance of enteropathogenic E coli, enterotoxigenic B fragilis-ETBF, and C difficile sp was difficult to assess because only a limited number of samples had detectable amplification levels. Therefore, specimens with Ct ≥40 were denoted as negative and those <40 were denoted as positive. The 2 statistically significant relations between pathogen versus commensal bacterial species were determined by logistic regression ANCOVA (Table 5); P values identify nonzero slopes. The relations (ie, slopes) between RV and B adolescentis and between Enterobacter and B infantis depended on a combination of location and the interaction of location and health status, respectively (Table 5).

When demographic variables (age, months of breast-feeding as supplement source in diet, and consumption of yogurt) were included as covariates to model pathogen–bacteria associations, C difficile was negatively affected by age, but significant only in boys with diarrhea from location B (slope = −0.089 P = 0.0115) (Table 6). The association between Enterobacter and age differed between healthy children and children with diarrhea (P = 0.0441); Enterobacter was directly affected by age in healthy children and was inversely affected by age in children with diarrhea, regardless of the location. Norovirus prevalence decreased with age (P = 0.0442) for all children from location A, whereas enterotoxigenic B fragilis-ETBF prevalence increased with age (P = 0.0041) for all children from location B (Table 6). Overall, the time of supplementary breast-feeding was inversely related to pathogenic E coli (slope = −0.051, P = 0.04). There was no significant relation to yogurt in relation to pathogens and health status (data not shown).

Prevalence of enterotoxigenic B fragilis ETBF was positively associated with the presence of Enterobacter (P = 0.02) and was dependent on location. All cases occurred in location B (P = 0.0140) (Table 7). Proportion of positive cases of NV and RV were inversely associated with presence of enterotoxigenic B fragilis-ETBF (P < 0.0001). These associations were affected by location for both NV and RV (P < 0.04) and by health status for NV only (P = 0.03) (Table 7).

Back to Top | Article Outline


Our results indicated that there were significant differences in the fecal microbiota of children with clinical diarrhea compared with healthy asymptomatic children. Regression analysis indicated that, overall, the abundance of Bifidobacterium spp was significantly reduced in children with clinical diarrhea from both locations A and B (Fig. 1), whereas certain Lactobacillus spp were reduced in children from both locations (Fig. 2). Abundance of B fragilis, one of the main anaerobic commensal bacterial groups that populate the large intestine and that has been associated with the maintenance of intestinal homeostasis and gut maturation (43), was reduced in children with diarrhea from location A but increased in children with diarrhea from location B (P < 0.05) probably because of the increase in enteropathogenic B fragilis-ETBF for the latter (Table 7); however, overall B fragilis group species were inversely correlated with probability of clinical diarrhea. The relative higher abundance and strong association of Bifidobacterium spp and Lactobacillus spp in healthy asymptomatic children (Table 4) and their inverse correlation with probability of clinical diarrhea (Figs. 1 and 2) suggested that these 2 species may represent a marker of gut homeostasis in the intestinal microbiota of children and that maintenance of basal levels of these bacterial species would be beneficial to a healthy intestine.

Within the same population, children who sought medical assistance for diarrhea in a hospital from location A had a different fecal microbiota compared with children with diarrhea who attended a day care center in location B. In our study, children with diarrhea from location A had a greater increase in the abundance of enteropathogenic E coli (10.21%), C difficile (8.76%), and RV (11.67%) (P < 0.0001) (Table 2). These infectious agents together represent the most common etiologies associated with acute gastroenteritis in children younger than 5 years in developing countries (28,44). Children with diarrhea from location B exhibited a different etiology: higher prevalence of NV (7.14%), Enterobacter (14.3%), enterotoxigenic B fragilis-ETBF (8.57%), and enteropathogenic E coli (7.14%) (Table 2). Differences in age between locations within the same population may explain differences in prevalence of enteric pathogens because the mean age for children with diarrhea at the day care center in location B was significantly higher than the mean age of children who asked for medical attention at location A (Table 2). Logistic regression analysis indicated that age was an important predictor of disease outcome because prevalence of NV, Enterobacter, enterotoxigenic B fragilis-ETBF, and C difficile was affected with age (Table 6). The prevalence of C difficile was significantly different between locations because children with diarrhea from location A, with a lower median age, were estimated to have 7.4 times more C difficile than children with diarrhea from location B (95% CI 1.6–34.5), reported in Table 2 as reciprocal odds ratio location B:A 0.14 (95% CI 0.03–0.63).

Interestingly, the abundance of enteropathogenic E coli was affected by length of time of consumption of breast milk as a supplement to the diet; however, this effect was not associated with any other measured demographic factor (ie, yogurt) or a different pathogen.

Children with no clinical sign of diarrhea from both locations showed some detectable levels of enteric pathogens in their feces. There were a few cases of RV, Enterobacter, enteropathogenic E coli, enterotoxigenic B fragilis-ETBF, and C difficile in fecal samples from asymptomatic healthy children who were statistically similar between locations. The presence of these potentially disease-producing pathogens suggested that assessment of health status by a questionnaire or routine medical examination is not enough to detect the prevalence of enteric disease-positive carriers. Instead, molecular analysis of the feces by PCR technology can provide a more sensitive and quantitative epidemiological tool for the surveillance of enteric pathogens from clinical isolates.

Previous studies that examined the fecal microbiota indicated that anaerobic microbiota is significantly reduced in acute diarrhea, leading to a relative predominance of aerobic bacteria (45,46). In our study, it is possible that the concomitant reduction of anaerobes such as B fragilis group with the overgrowth of bacterial pathogens (ie, enteropathogenic E coli, and Enterobacter) may have contributed to the greater reduction of certain Lactobacillus spp. A transient change in Lactobacillus spp was also reported for a group of children in India who were positive for RV during periods after clinical diarrhea (47). Future longitudinal studies that discriminate among different Lactobacillus spp by sensitive molecular techniques may be required to elucidate the effect of these highly diverse species on diarrhea.

The logistic regression models identified a significant negative relation between prevalence of diarrhea and Bifidobacterium spp (Fig. 1), including B adolescentis and B infantis spp. Administration of bifidobacteria to infants and young children either in the community (48) or to those with acute diarrhea (49) reduced fecal shedding of RV. In addition, studies in experimental animals indicate that the administration of certain bifidobacteria strains resulted in the development of increased titers of IgA antibody in both feces and serum of infected animals, indicating that these probiotic bacteria potentiated the immune response (50,51). In our study, no specific pathogen could be attributed as the major contributor for the negative relation between cases of diarrhea and Bifidobacterium spp, probably because of combined etiology or reduced number of cases.

Taken together, these observations suggested that a dietary intervention that increased the abundance of Bifidobacterium spp in children at risk for diarrhea could reduce the prevalence of diarrhea. It could also contribute to greater abundance of the B fragilis group, which is a major component of the normal anaerobic microbiota and appears to positively correlate with Bifidobacterium spp in healthy children; however, the presence of enterotoxigenic B fragilis-ETBF in children with diarrhea indicated the diversity of this group in the human microbiota. Identification of culture-independent assays that help to detect other fastidious and sometimes unculturable organisms within the B fragilis group should facilitate a better understanding of their effects under health and disease.

Our study demonstrated that demographic factors such as location, age, and dietary supplementation (months of breast-feeding) affected the specific composition of the fecal microbiota of children younger than 5 years. Therefore, the composition of the intestinal microbiota should be characterized a priori to capture all associated factors that may affect the results of a targeted intervention with probiotic bacteria (52). For our specific population in the highlands of Colombia, it is reasonable to hypothesize that administration of Bifidobacterium spp, which has historically shown a good safety record, or a diet rich in bifidogenic prebiotics that promotes growth of Bifidobacterium spp, would help in reducing the prevalence of diarrhea in at-risk children.

Back to Top | Article Outline


1. Black RE, Cousens S, Johnson HL, et al. Global, regional, and national causes of child mortality in 2008: a systematic analysis. Lancet 2010; 375:1969–1987.
2. Boschi-Pinto C, Velebit L, Shibuya K. Estimating child mortality due to diarrhoea in developing countries. Bull WHO 2008; 86:710–717.
3. Ospino DU, Young G, Navarro OA. Viral gastroenteritis and diversity of Rotavirus strains in Colombian children: a systematic review. J Infect Dev Ctries 2008; 2:99–105.
4. Mattar S, Vasquez E. Escherichia coli O157:H7 infection in Colombia. Emerg Infect Dis 1998; 4:126–127.
5. Boeke CE, Mora-Plazas M, Forero Y, et al. Intestinal protozoan infections in relation to nutritional status and gastrointestinal morbidity in Colombian school children. J Trop Pediatr 2010; 56:299–306.
6. Correa A, Solarte Y, Barrera J, et al. [Molecular characterization of rotavirus in the city of Santafe de Bogota, Colombia. Determination of the electrophenotypes and typing of a strain by RT-PCR]. Rev Latinoam Microbiol 1999; 41:167–173.
7. Rugeles LC, Bai J, Martinez AJ, et al. Molecular characterization of diarrheagenic Escherichia coli strains from stools samples and food products in Colombia. Int J Food Microbiol 2010; 138:282–286.
8. O’Ryan M, Perez-Schael I, Mamani N, et al. Rotavirus-associated medical visits and hospitalizations in South America: a prospective study at three large sentinel hospitals. Pediatr Infect Dis J 2001; 20:685–693.
9. Curns AT, Steiner CA, Barrett M, et al. Reduction in acute gastroenteritis hospitalizations among US children after introduction of rotavirus vaccine: analysis of hospital discharge data from 18 US states. J Infect Dis 2010; 201:1617–1624.
10. De la Hoz F, Alvis N, Narvaez J, et al. Potential epidemiological and economical impact of two rotavirus vaccines in Colombia. Vaccine 2010; 28:3856–3864.
11. Reading NC, Kasper DL. The starting lineup: key microbial players in intestinal immunity and homeostasis. Front Microbiol 2011; 2:148.
12. Wexler HM. Bacteroides: the good, the bad, and the nitty-gritty. Clin Microbiol Rev 2007; 20:593–621.
13. Avila-Campos MJ, Liu C, Song Y, et al. Determination of bft gene subtypes in Bacteroides fragilis clinical isolates. J Clin Microbiol 2007; 45:1336–1338.
14. Vaughan EE, Heilig HG, Ben-Amor K, et al. Diversity, vitality and activities of intestinal lactic acid bacteria and bifidobacteria assessed by molecular approaches. FEMS Microbiol Rev 2005; 29:477–490.
15. Sanchez B, Urdaci MC, Margolles A. Extracellular proteins secreted by probiotic bacteria as mediators of effects that promote mucosa-bacteria interactions. Microbiology 2011; 156:3232–3242.
16. van Baarlen P, Troost F, van der Meer C, et al. Human mucosal in vivo transcriptome responses to three lactobacilli indicate how probiotics may modulate human cellular pathways. Proc Natl Acad Sci U S A 2011; 108 (suppl 1):4562–4569.
17. McFarland LV. Evidence-based review of probiotics for antibiotic-associated diarrhea and Clostridium difficile infections. Anaerobe 2009; 15:274–280.
18. Soh SE, Ong DQ, Gerez I, et al. Effect of probiotic supplementation in the first 6 months of life on specific antibody responses to infant hepatitis B vaccination. Vaccine 2010; 28:2577–2579.
19. Guarino A, Lo Vecchio A, Canani RB. Probiotics as prevention and treatment for diarrhea. Curr Opin Gastroenterol 2009; 25:18–23.
20. Floch MH, Madsen KK, Jenkins DJ, et al. Recommendations for probiotic use. J Clin Gastroenterol 2006; 40:275–278.
21. Preidis GA, Hill C, Guerrant RL, et al. Probiotics, enteric and diarrheal diseases, and global health. Gastroenterology 2011; 140:8–14.
22. Thomas DW, Greer FR. Probiotics and prebiotics in pediatrics. Pediatrics 2011; 126:1217–1231.
23. Merenstein D, Murphy M, Fokar A, et al. Use of a fermented dairy probiotic drink containing Lactobacillus casei (DN-114 001) to decrease the rate of illness in kids: the DRINK study, a patient-oriented, double-blind, cluster-randomized, placebo-controlled, clinical trial. Eur J Clin Nutr 2010; 64:669–677.
24. Guandalini S. Probiotics for prevention and treatment of diarrhea. J Clin Gastroenterol 2011; 45 (suppl):S149–S153.
25. Sazawal S, Dhingra U, Hiremath G, et al. Prebiotic and probiotic fortified milk in prevention of morbidities among children: community-based, randomized, double-blind, controlled trial. PLoS One 2010; 5:e12164.
26. Sur D, Manna B, Niyogi SK, et al. Role of probiotic in preventing acute diarrhoea in children: a community-based, randomized, double-blind placebo-controlled field trial in an urban slum. Epidemiol Infect 2011; 139:919–926.
27. Floch MH, Walker WA, Madsen K, et al. Recommendations for probiotic use—2011 update. J Clin Gastroenterol 2011; 45 (suppl):S168–S171.
28. Gutierrez MF, Matiz A, Trespalacios AA, et al. Virus diversity of acute diarrhea in tropical highlands. Rev Latinoam Microbiol 2006; 48:17–23.
29. Winn WC, Koneman EW. Koneman's Color Atlas and Textbook of Diagnostic Microbiology. 6th edPhiladelphia:Lippincott Williams & Wilkins; 2006.
30. Petrich A, Mahony J, Chong S, et al. Multicenter comparison of nucleic acid extraction methods for detection of severe acute respiratory syndrome coronavirus RNA in stool specimens. J Clin Microbiol 2006; 44:2681–2688.
31. Delroisse JM, Boulvin AL, Parmentier I, et al. Quantification of Bifidobacterium spp. and Lactobacillus spp. in rat fecal samples by real-time PCR. Microbiol Res 2008; 163:663–670.
32. Haarman M, Knol J. Quantitative real-time PCR assays to identify and quantify fecal Bifidobacterium species in infants receiving a prebiotic infant formula. Appl Environ Microbiol 2005; 71:2318–2324.
33. Haarman M, Knol J. Quantitative real-time PCR analysis of fecal Lactobacillus species in infants receiving a prebiotic infant formula. Appl Environ Microbiol 2006; 72:2359–2365.
34. Penders J, Vink C, Driessen C, et al. Quantification of Bifidobacterium spp., Escherichia coli and Clostridium difficile in faecal samples of breast-fed and formula-fed infants by real-time PCR. FEMS Microbiol Lett 2005; 243:141–147.
35. Penders J, Thijs C, Vink C, et al. Factors influencing the composition of the intestinal microbiota in early infancy. Pediatrics 2006; 118:511–521.
36. Karns JS, Van Kessel JS, McClusky BJ, et al. Incidence of Escherichia coli O157:H7 and E. coli virulence factors in US bulk tank milk as determined by polymerase chain reaction. J Dairy Sci 2007; 90:3212–3219.
37. Merino VR, Nakano V, Liu C, et al. Quantitative detection of enterotoxigenic Bacteroides fragilis subtypes isolated from children with and without diarrhea. J Clin Microbiol 2011; 49:416–418.
38. Nadkarni MA, Martin FE, Jacques NA, et al. Determination of bacterial load by real-time PCR using a broad-range (universal) probe and primers set. Microbiology 2002; 148:257–266.
39. Bustin SA. A-Z of Quantitative PCR. La Jolla, CA:International University Line; 2004.
40. Solano-Aguilar G, Dawson H, Restrepo M, et al. Detection of Bifidobacterium animalis subsp. lactis (Bb12) in the intestine after feeding of sows and their piglets. Appl Environ Microbiol 2008; 74:6338–6347.
41. Richards GP, Watson MA, Fankhauser RL, et al. Genogroup I and II noroviruses detected in stool samples by real-time reverse transcription-PCR using highly degenerate universal primers. Appl Environ Microbiol 2004; 70:7179–7184.
42. Gouvea V, Glass RI, Woods P, et al. Polymerase chain reaction amplification and typing of rotavirus nucleic acid from stool specimens. J Clin Microbiol 1990; 28:276–282.
43. Mazmanian SK, Liu CH, Tzianabos AO, et al. An immunomodulatory molecule of symbiotic bacteria directs maturation of the host immune system. Cell 2005; 122:107–118.
44. Manrique-Abril FG, Tigne y Diane B, Bello SE, et al. [Diarrhoea-causing agents in children aged less than five in Tunja, Colombia]. Rev Salud Publica (Bogota) 2006; 8:88–97.
45. Fujita K, Kaku M, Yanagase Y, et al. Physicochemical characteristics and flora of diarrhoeal and recovery faeces in children with acute gastro-enteritis in Kenya. Ann Trop Paediatr 1990; 10:339–345.
46. Tazume S, Takeshi K, Saidi SM, et al. Ecological studies on intestinal microbial flora of Kenyan children with diarrhoea. J Trop Med Hyg 1990; 93:215–221.
47. Balamurugan R, Janardhan HP, George S, et al. Molecular studies of fecal anaerobic commensal bacteria in acute diarrhea in children. J Pediatr Gastroenterol Nutr 2008; 46:514–519.
48. Phuapradit P, Varavithya W, Vathanophas K, et al. Reduction of rotavirus infection in children receiving bifidobacteria-supplemented formula. J Med Assoc Thai 1999; 82 (suppl 1):S43–S48.
49. Saavedra JM, Bauman NA, Oung I, et al. Feeding of Bifidobacterium bifidum and Streptococcus thermophilus to infants in hospital for prevention of diarrhoea and shedding of rotavirus. Lancet 1994; 344:1046–1049.
50. Qiao H, Duffy LC, Griffiths E, et al. Immune responses in rhesus rotavirus-challenged BALB/c mice treated with bifidobacteria and prebiotic supplements. Pediatr Res 2002; 51:750–755.
51. Moreno Munoz JA, Chenoll E, Casinos B, et al. Novel probiotic Bifidobacterium longum subsp. infantis CECT 7210 strain active against rotavirus infections. Appl Environ Microbiol 2011; 77:8775–8783.
52. Rijkers GT, Bengmark S, Enck P, et al. Guidance for substantiating the evidence for beneficial effects of probiotics: current status and recommendations for future research. J Nutr 2011; 140:S671–S676.

Bacteroides fragilis; Bifidobacterium; diarrhea; Lactobacillus; norovirus; probiotic; rotavirus

© 2013 by European Society for Pediatric Gastroenterology, Hepatology, and Nutrition and North American Society for Pediatric Gastroenterology,