Etiology and biodistribution of enterobacteria and parasites, and their associated environmental risk factors among children under 5 years old with diarrhea in East-Central Gabon

Better knowledge of endogenous germs and their associated demographic and environmental risk factors in a locality are essential to treat childhood diarrheal diseases. The aim of the study was to identify and characterize bacterial and parasitic pathogens responsible for childhood diarrhea, and to identify environmental risk factors associated with diarrhea in children under 5 years old living in Koula-Moutou, East-Central Gabon. Methods A cross-sectional study was performed from May 2016 to February 2018. One-hundred and thirty-two (132) children under 5 years old were enrolled. The detection of pathogens in stool samples was performed using microscopic examination and MIF concentration for parasites, and conventional culture on selective media for enterobacteria.


Abstract Background
Better knowledge of endogenous germs and their associated demographic and environmental risk factors in a locality are essential to treat childhood diarrheal diseases. The aim of the study was to identify and characterize bacterial and parasitic pathogens responsible for childhood diarrhea, and to identify environmental risk factors associated with diarrhea in children under 5 years old living in Koula-Moutou, East-Central Gabon.
Methods A cross-sectional study was performed from May 2016 to February 2018. One-hundred and thirty-two (132) children under 5 years old were enrolled. The detection of pathogens in stool samples was performed using microscopic examination and MIF concentration for parasites, and conventional culture on selective media for enterobacteria.

Results
The prevalence of diarrheal pathogens was 46.1%, including bacterial enteropathogens (25.5%) and parasites (20.6%). A total of 228 pathogenic organisms were isolated, including 199 bacterial strains (87.3%) and 29 parasites (12.7%). The speci c richness of the isolated enterobacteria was 24 species with a high prevalence of E. coli (39.8%), including 26.7% for the diarrheal cases and 13.1% for the controls. Twelve (12) parasite species were also isolated and the most common types of parasites were rhizopods (44.8%), which accounted for 37.9% of the diarrheal cases and 6.9% of the controls. Univariate analysis showed that the presence of watercourses (OR = 3.37) and domestic animals (OR = 1.65) were signi cant risk factors for diarrhea.

Conclusion
The ndings show a high prevalence of bacterial enteropathogens but a low rate of parasites and bacteria-parasite co-infection in the study area. Risk factors associated with diarrhea among children under 5 years old were the presence of watercourses and domestic animals. These ndings highlight the need to strengthen the routine examination of diarrheic stool samples for the diagnosis of pathogenic organisms. Further analyses are required to better understand the etiologies and risk factors associated with the transmission of bacteria and parasites in rural, semi-urban and urban regions of Gabon.
Background Page 3/27 Diarrheal diseases remain the second cause of preventable childhood morbidity and mortality. They are responsible for 9% of global deaths among children under 5 years old (1,2), despite technological advances and awareness programs issued by the World Health Organization (WHO) and the increased use of oral rehydration therapy in recent decades (3,4). Today, acute diarrhea still represents a major public health issue worldwide and about 80% of diarrheal deaths in children under two years old occur in South-East Asia and sub-Saharan Africa (5,6).
Infectious and contagious diarrhea is common in developing countries where the risk of transmission is even greater because of the appalling environmental conditions. They are the result of non-sanitary elimination of human wastes, very limited access to safe water, poor hygiene practices, and the means of food conservation and cooking (7,8). In addition, the density of urbanization and the increase of populations in cities can also be risk factors contributing to the development and the contraction of diseases in young children (9).
Diarrheal diseases are mostly caused by enteric pathogens, including bacteria, viruses, and protozoa (10). Germs such as rotaviruses, Escherichia coli and other pathogenic bacteria, are the most common etiological agents in young children which induce moderate to severe diarrhea (11,12). In addition, the economic and health impacts due to diarrheal diseases would require monitoring in order to better understand the risk factors leading to their occurrence. Thus, reducing the morbidity of diarrheal diseases requires a comprehensive approach of this pathology that combines the study of therapeutic management, socio-economic, demographic and environmental conditions, the ecology of the pathogens responsible, and knowledge of local communities (5,13).
In Gabon, like everywhere else in sub-Saharan Africa, the epidemiological plan is largely dominated by communicable diseases which are classi ed as the rst group responsible for morbidity. Ninety percent (90%) of the causes of death for children under ve years old in this country are due to malaria (29%), prematurity (15%), acute respiratory infections (11%), HIV (10%), and diarrheal diseases (6%) (14,15). The prevalence of diarrheal diseases was 50% in 2008 and 15.8% in 2012 in the city of Libreville, and the most vulnerable population was children under 15 years old (15,16). However, the national prevalence of endemic diarrhea remains only poorly documented and prospections on the etiology of infectious diarrheas are rarely done routinely. Furthermore, it is also necessary to consider the speci c etiological characteristics of the different cities because bacterial etiology often leads to systematic and abusive antibiotic therapy (17). Therefore, better knowledge of the possible infectious causes and their associated risk factors will improve diarrhea treatment methods in order to preserve effective antibiotics and ght off antibiotic resistance.
The aims of this study were to 1) estimate the prevalence of parasitic and bacterial pathogens, and their associated risk factors in children under 5 years old with diarrhea living in Koula-Moutou and 2) make a map of parasitic and bacterial pathogens responsible for childhood diarrhea in Koula-Moutou, Gabon.

Study area
This study was conducted at the Paul Moukambi Regional Hospital Center (PMRHC) and different districts in Koula-Moutou, Gabon. Koula-Moutou is the administrative capital of the Ogooué-Lolo Province and a semi-urban area in East-Central Gabon (1° 13′ 10″ S and 12° 28′ 0″ E) ( Figure 1). The city of Koula-Moutou had an estimated population of 23,629 inhabitants in 2017 (18). Among the twenty-eight (28) neighborhoods listed in this study, only twenty-seven (27) are currently listed by the municipal authorities of the province (Figure 1).

Study design and population
A cross-sectional study was carried out in Koula-Moutou from May 2016 to February 2018. The different sampling campaigns were carried out intermittently and during the rainy season which corresponds to the peak of diarrheal disease (19). Samples were collected in the pediatric and emergency wards of the PMRHC and different districts of the city. The de nition of "diarrhea case" was based on the recommendations of the World Health Organization (8).
The study population was composed of 132 children aged 0 to 70 months old living in the selected study sites and presenting acute diarrhea or no diarrhea. Children with acute diarrhea, hospitalized patients, children treated on an outpatient or consultant basis and who had not been on antibiotic therapy for more than 24 hours prior to the sampling were included. Children older than 71 months and under treatment for more than 24 hours prior to sampling were excluded from this study. Controls selected during the same period were children who had no history of diarrheal episode over a period of 15 days prior to their recruitment in this study.
Written informed consent was obtained from the parents or guardians of the participating children. For each participant, demographic characteristics (age, sex, origin), clinical characteristics (fever, vomiting, duration of diarrhea before consultation, the type or absence of treatment), and environmental and living conditions (type of drinking water, presence of toilets or a latrine, presence of watercourses, presence of domestic animals, sanitary condition of the plots, proximity of livestock farms) were collected on a standard survey form. Children aged 24 months old or less were considered to be exposed to the disease.

Sample collection
Stool samples were collected in sterile containers from diarrheal pediatric patients. In infants, fecal samples were collected by rectal swab. Each vial was identi ed by the code number of the participant.

Sample examination
Macroscopic examination was performed according to the Bristol scale on each stool sample to note the appearance, consistency, color and possible presence of blood and mucus. A microscopic examination after speci c staining with lugol and merthiolate formalin iodine for the detection of parasites and Gram staining for the appreciation of the morphology and composition of the bacterial ora was carried out on each stool sample.

Detection of parasites
A small amount of the stool sample equivalent to the size of a pea was mixed with a drop of saline (0.9% sodium chloride) and covered with a coverslip. A drop of lugol and merthiolate formalin iodine were also added on a small stool sample and the slide was examined under a light microscope to observe motile parasites. This method is used to detect intestinal protozoa and soil-transmitted helminths (STHs). The sample was considered negative if no intestinal protozoa or helminths were found on the entire slide. The merthiolate-iodine-formaldehyde (thimerosal) concentration (MIFc) method was also used. The MIFc method was carried out as described by M'bondoukwé et al. (20). Samples were considered positive upon the detection of a single parasite.

Isolation and identi cation of bacterial strains
Fecal samples were streaked on selective agar media for the isolation of Gram negative bacilli: Hektoen, Salmonella-Shigella, for the isolation of strains of Salmonella, Shigella; Methylene Blue Eosin (MBE) for the isolation of E. coli and Drigalski for other enterobacteria. Brie y, the stool samples were streaked on the agar surface. The media were incubated at 37°C for 24 hours. After the isolation of pure colonies, the identi cation of the strains was carried out using the standard biochemical system Api 20E (Biomérieux, Marcy l'Etoile, France) and the results were interpreted using the Api web TM stand alone V 1.2.1 software (Biomérieux, Marcy l'Etoile, France). However, no con rmation by serotyping was done for Salmonella, Shigella and Yersinia isolates.

Ethical considerations
The study was approved by the Gabonese National Ethics Committee for Research and the Ministry of Health (PROT N° 0020/2015/SG/CNE). The study protocol was approved and authorized by the Center-East General Direction of Health, the Direction of the Ogooué-Lolo Provincial Academy and the Provincial Direction of Family and Social Welfare. Stool samples were collected from children after obtaining their parents' or guardians' written informed consent.

Statistical analysis
The data collected were typed in an Excel spreadsheet and analyzed using R software version 3.2.2. Heterogeneity between the infectious origins within the two populations and the eco-distribution of enteropathogens were assessed using the chi-square test and Fisher's exact tests. The Shannon diversity index made it possible to assess the bacterial and parasitic diversity observed. Crudes odds ratios (CORs) and 95% con dence intervals (CIs) were used to assess the association between variables and occurrence of diarrhea. In addition, a factorial correspondence analysis was performed to assess the ecodistribution of enteropathogens. The statistical signi cance was set at p<0.05.

Study population
The children's demographic characteristics are summarized in Table 1. A total of 132 children aged from 0 to 70 months old were included, among whom 102 had a diarrheal syndrome and 30 were controls. Overall, the sex ratio was 1.06 (68 males and 64 females). The sex ratios for the children with diarrhea and the control group were 1.13 (54 males and 48 females) and 0.088 (14 males and 16 females), respectively. The mean age of the children included in the study was 32,6 ± 20,6 months old. Diarrheal syndrome was predominantly found in children aged 0 to 24 months old with a prevalence of 44.1%. Prevalence of diarrheal syndrome was 28.4% and 27.5% in children aged 25 to 48 months old and 49 to 70 months old, respectively.
Clinical examination revealed that the most frequent signs associated with diarrhea were fever (26.5%; n = 27/102), the association of fever and vomiting (16.7%; n = 17/102) and vomiting (5.9%; n = 6/102). These signs were assessed according to the usual clinical criteria: weight-to-height and height-to-age ratio, vomiting and fever. In addition to diarrhea, 11.8% of children had other conditions. Thirty-three (32.4%) of the 102 subjects with diarrhea were on treatment; with a prevalence of 15.7% for antiparasitic treatments, 6.9% for traditional treatment, 5.9% for antibiotic therapy and 3.9% with antibioticantiparasitic combination (data not showed).
In the Nematoda class, the highest prevalence was observed with Ascaris lumbricoides (10.3%), followed by Ancylostoma duodenale (6.9%) and Trichuris trichiura (3.4%). Moreover, some parasitic species were also isolated in the controls among which Ancylostoma duodenale (3.4%) and Ascaris lumbricoides (3.4%) for the Nematodes, and Entamoeba coli (6.9%) for the Rhizopoda class. Blastocystis hominis was isolated both in diarrhea cases and in controls with prevalence rates of 10.3% and 3.4%, respectively. The overall diversity of the parasites was assessed by the Shannon index which was valued at 2.2. The difference was not statistically signi cant for all parasites between diarrhea cases and controls.  Speci c richness of Enterobacteriaceae strains isolated from the study population Opportunistic pathogens were more common than strict enteropathogens (Table 4). Strict enteropathogens were found in 18.6% (n = 37/199) of the diarrhea samples. The highest prevalence was for Salmonella spp (8.6%), followed by Salmonella enterica (3.0%; n = 6/199), Salmonella Typhi (2.0%; n = 4/199) and Salmonella Paratyphi A (1.5%) serovars of this species, whereas the genera Shigella and Yersinia were the least represented. The global prevalence of opportunistic pathogens was higher among diarrhea samples (57.8%) compared to control samples (23.6%) (see Table 4). Out of the 17 isolated species, the highest prevalence recorded was for E. coli 1 (25.2% for diarrhea cases and 11.6% for controls), followed by Raoultella ornithinolytica (11.6% for diarrhea cases and 3.0% for controls), Kluyvera spp (2.5% for diarrhea cases and controls, respectively), and Serratia fonticola (2.5% for diarrhea cases).

Distribution between enteropathogens and age groups
The distribution of the different species of enteropathogens according to the different age groups was determined using the correspondence factor analysis (CFA) (Fig. 4). The results of the distribution of enteropathogens according to age groups showed that parasitic enteropathogens, including Entamoeba hystolitica, Dientamoeba fragilis, Schistosoma intercalatum and Schistosoma mansonni were mainly associated with children aged 0 to 24 months old. This same association is found for Salmonella enterica. Enteropathogens of the Shigella genus including shigella spp, shigella sonnei, as well as Yersinia spp and Endolimax nanus were associated with children aged 25 to 48 months old. Salmonella thyphi and Salmonella paratyphi A were associated with children aged 49 to 70 months old. No association between age groups and pathogens was found for Giardia intestinalis, Ancylostoma duodenale and Salmonella spp.

Risk factors among environmental determinants for infections due to enteropathogens
Potential risk factors (drinking water, type of toilet, watercourses, domestic animals, and sanitation of the plot) associated with diarrheal diseases are summarized in Table 6.  Mixed : herbs + garbage bins/farms ; CI at 95% : con dence intervals at 95%, COR : crude odds ratios.
The percentages are will reference to totals given at the top of each column. Odd ratios obtained by logistic regression. * COR and CI values above 1.0 that indicate association with diarrhea The results showed that more than half of the population consumed water from taps (68.6% for diarrheal cases vs 60.0% for controls). Regarding amenities, the majority of children used traditional latrines (59.8% in diarrheal cases and 70.0% in controls, respectively). An unsanitary condition found for plots was the presence of grass (85.3% and 86.7% for diarrheal cases and controls, respectively). Although no signi cant association between the different potential risk factors and the occurrence of diarrhea was observed, the data showed that the presence of a watercourse in the immediate environment increased the risk of diarrhea in children which was multiplied by 4, as well as the presence of domestic animals such as chickens and dogs / cats which multiplied the risk by 2.66 and 2.69, respectively.

Discussion
This study on diarrheal disease in Gabonese children is the rst to assess the etiology of diarrheal syndromes in Koula-Moutou, Gabon. While data on adults and children already exist, children under 5 years old with diarrheal syndromes living in the semi-urban environment of the city of Koula-Moutou have not been previously studied.
In this study, the prevalence of children aged up to 5 years with diarrhea was 77.3%, it is similar to that reported by a previous study in another city in Gabon (21). The results of our study also showed that 41.6% of the children with diarrhea were aged 0 to 24 months old and were the most affected age group with a prevalence of 44%. Our results corroborate those of several studies led in Africa, with a higher prevalence found in children aged 0 to 12 months (8,16,22,23). This predominance could be explained by the fact that generally, between 6 to 23 months of age, children are more vulnerable compared to those aged over 36 months old (24). Indeed, pathologies such as diarrhea are recurrent in this age group due to the gradual decline in maternal antibodies as well as the introduction of new dietary practices sometimes correlated with poor hygienic practices (24,25).
A range of different pathogenic organisms can cause pediatric diarrhea in the world, especially in tropical and developing countries, including rotaviruses and adenoviruses (26), intestinal parasites such as Giardia intestinalis, Entamoeba histolytica (27), and bacteria such as Escherichia coli, Shigella spp, Salmonella, Campylobacter spp (8). This study shows a strong heterogeneity of parasites and bacteria in diarrheal cases in which most of the species are not isolated in control cases. The most common parasitic species isolated in diarrheal cases were Entamoeba spp, Ascaris lumbricoides, Blastocystis hominis, Ancylostoma duodenale, Giardia intestinalis, Endolimax nanus, Schistosoma spp and Trichuris trichiura. This pro le is similar to that reported in previous studies (28,29). In addition, Ascaris lumbricoides, Blastocystis hominis, Giardia intestinalis, Entamoeba spp and Trichuris trichiura appear to be the intestinal parasites most frequently involved in parasitic infections in Gabon (20). These species have a worldwide distribution and are characteristic of some countries (30). Regarding the bacterial group, 11 genera belonging to the Enterobacteriacea family were recorded, a similar pro le was found in  (23,32). This contrast could probably be related to the lack of characterization of diarrheagenic Escherichia coli (DEC) in this study. Furthermore, this study also shown the presence of parasitic pathogens in this population, with a frequency of 20.6%. This prevalence was, however, lower than that recorded in Nigeria (37.1%) (32), Cameroon (59.2%) (7) and Gabon (61.0%) (21). The high prevalence of infections due to bacterial pathogens in these conditions is similar to that found in other studies (8,23).
One of the particularities of this study was the lack of bacterial pathogens isolated from controls. The frequency of bacterial infections (25.5%) is higher than that reported in a previous study by Koko et al. carried out in Libreville (17) who obtained a prevalence of 12.9%. This variability in epidemiological characteristics could undoubtedly be linked to the level of urbanization of the two zones, the environmental conditions and hygiene of the respective populations, which are all factors that in uence the etiology of diarrheal diseases (9, 36). Salmonella spp strains were the most predominant in the bacterial pro le of this study and accounted for 29.4%. This prevalence is in contrast with the one obtained in previous studies concerning developing countries, and particularly countries in the African region where the prevalence was markedly lower (37,38). However, these data corroborate the work of Koko et al. in Gabon (17) and Rathaur et al. in India (39). Indeed, these studies identi ed Salmonella spp as a major etiological agent responsible for diarrhea. In addition, its presence could be correlated with domestic animals like poultry, which could constitute reservoirs of infection and a potential source of Salmonella diarrhea (7). The prevalence rates of the Shigella and Yersinia strains observed in this study, 4.9% and 2.0% respectively, can be explained by the epidemiological heterogeneity of the geographical areas (32,37,40,41).
In this research, multiple infections of enteric pathogens were more found in diarrheal cases than in controls. This has also been reported in other studies particularly in patients with low or average income (35,42). These results would suggest that more than one pathogen was responsible for the diarrheal disease in children living in Koula-Moutou. The presence of mixed infections complicates the diagnosis of a speci c pathogen responsible for the disease and may result in an additive effect, which may lead to new clinical pro les (42). A reliable basic diagnosis must be established for better patient care and treatment.
Factorial correspondence analysis (CFA) reveals a predominance of parasites, Salmonella and Shigella in children over 12 months old. The parasitic data from this study are comparable to the literature which highlights an increase in parasitic infections in subjects over 12 months of age, for which the most likely explanation would be frequent contact with soil (7,43). Likewise, this distribution of bacterial strains has already been reported by other authors (17,23,39).
The transmission and spread of diarrheal diseases are closely linked to environmental factors but also to living conditions, personal hygiene, behavior and domestic environment (44). In this study, univariate analysis of the data showed that the presence of rivers and domestic animals were the risk factors signi cantly associated with the occurrence of diarrhea in the town of Koula-Moutou. These results are consistent with the data reported by Bouba Djourdebbé et al. (24). Watercourses create a habitat conducive to the proliferation of microorganisms. Combined with other parameters, such as the proximity of animals which are potential pathogen reservoirs and the lack of knowledge on good hygiene practices, the presence of rivers in uences the risk of contamination (24,45). Given that diarrheal diseases can be linked to the presence of feces in water, the population of Koula-Moutou can be highly exposed to this health risk. Indeed, this city has a very dense water network which unfortunately serves for domestic use but in which wastewater containing the feces of some animals or even humans can be thrown away.
These practices can obviously contribute to the fairly high prevalence of diarrheal diseases in this area.
This study has some limitations. It used only microbiological analyzes, including stool bacterial cultures, as they are the most common diagnostic routine methods used in developing countries. Stool culture may have a lower sensitivity of detection of bacterial pathogens than PCR, which allows the discrimination of different pathotypes and a better appreciation of the involvement of multiple infections with enteric pathogens associated with diarrhea (42). Moreover, the low proportion of controls in this study may not be representative of the correlation between enteric pathogens and acute diarrhea.

Conclusion
This preliminary study in the city of Koula-Moutou on children with diarrhea reveals a great diversity of enterobacteria and parasites. Although the pro le of pathogenic enterobacteriaceae strains is very classic, this study highlights a high prevalence of unusual bacteria which may be involved in childhood diarrhea. The map of the distribution of pathogens shows no particular characteristic in any area of the city. In addition, the occurrence of diarrhea in this city is linked to environmental determinants such as the presence of rivers and animals. Declarations conceived, designed and coordinated the study, and conducted data analysis. All authors have read and approved the nal manuscript.

Availability of data and materials
The datasets that were used for the analysis and the preparation of this manuscript are available from the corresponding author upon reasonable request.
Ethics approval and consent to participate This is detailed in the "Methods" section.

Consent for publication
Not applicable.