Etiology and pathogenicity of bacterial isolates: a cross sectional study among diarrheal children below five years in central regions of Kenya

Introduction Bacterial agents are among pathogens implicated to cause diarrhea in children resulting to huge mortality and morbidities. Bacterial etiologies causing diarrhea in children below five years are rarely investigated in Central Kenya, which would otherwise guide prescription and target health education. Methods A cross-sectional study approach was applied on 163 randomly selected stool samples from children below five years who presented with diarrhea in Murang`a and Muriranja`s hospitals. The objective was to determine the bacterial agents of diarrhea. Enteric bacterial pathogens were cultured using appropriate media and identified. Statistical analyses were performed using STATA v.13. Chi-square or Fisher exact-test were used to check for evidence of relationship whenever applicable. Results There were nearly equal distributions in gender 86 (52.8%) female vs. 77 (47.2%) male, majority (35.6%) aged between 0-12 months. Bacterial isolates were highly diverse in female than the male, children aged 49-60 months and least among those aged 0-12 months. A total of 188 bacterial isolates belonging to 11 genera were recovered. The predominant bacteria was nonpathogenic Escherichia coli 85 (45.2%), while 13 (6.9%) Escherichia coli were positive for virulence genes, including 8 (4.3%) positive for LT and STp Shiga-like or Enterotoxigenic Escherichia coli, 3 (1.6%) positive for eae and bfpA Enteropathogenic Escherichia coli and 2 (1.1%) positive for Enteroaggregative Escherichia coli gene. Others included: Salmonella 21 (11.2%), Pseudomonas 14 (7.4%), Shigella 14 (7.4%), Klebsiella 12 (6.4%), Aeromonas 8 (4.3%), Enterobacter 7 (3.7%), Proteus 8 (4.3%), Citrobactor 3 (1.6%), Yersinia 2 (1.1%) and Vibrio 1 (0.5%). Conclusion Salmonella was the major bacterial isolate and majority of the bacteria were statistically significant cause of diarrhea (p=0.001).


Introduction
Diarrhea is having loose or watery stools at least three times per day or more frequently than normal for an individual [1]. Despite the efforts in controlling mortality, 9% of all deaths among children below 5 years globally in 2015 were due to diarrhea [2]. This burden of the disease remains unacceptably high. The Millennium Development Goals (MDGs) called for a reduction of child mortality by two thirds between 1990 and 2015. The newly launched and adopted 2030 agenda for Sustainable Development Goals (SDGs) aims to achieve what was not accomplished by MDGs [3]. Globally, diarrhea kills 2,195 children every day, more than AIDS, malaria and measles combined [4]. In Sub-Saharan Africa (SSA), the etiology of diarrhea is seldom known due to the lack of infrastructure for diagnosis. Diarrheal diseases cause 16% of deaths among children below five years in Kenya [5]. Every Kenyan child below the age of five experiences an average of three bouts of diarrhea every year [6]. Under vision 2030, Kenya has committed herself to reduce child mortality by two third among children below 5 years. The prevalence of diarrhea among children below five years in central Kenya stands at 10.4% and 12.1% in Murang`a County [7].
Diarrhea has a myriad of bacterial strains associated with it. The major bacterial pathogens include Escherichia coli, Shigella, Campylobacter, Salmonella, and Vibrio species transmitted mainly through fecal-oral route [1] although other enterobacteriaceae have been linked to cause diarrhea. It is crucial, therefore, to accurately identify the frequency of the broad range of bacterial diarrheal pathogens as well as their virulence genes to better understand bacterial diversity and pathogenicity. Limited continuous surveillance of bacterial etiologies has led to the narrowing arsenal of antibiotic use posing a devastating threat in treatment and management of bacterial-associated diarrhea.

Study site:
The study was carried out in Murang`a North located in Murang`a County, Kenya. Study sites included the two major hospitals; Muranga`a County Referral Hospital and Muriranja`s Level 4 hospital.
Research design: A cross-sectional study approach was used in this study.

Target population:
The research assessed children up to five years of age who reported cases of diarrhea from either of the two hospitals.
Sampling design: Sample selection was done using the systematic random sampling where the first unit (case) was selected randomly in each hospital. The n th case after the starting point followed a systematic selection. The n th case represents the sampling interval which was calculated by dividing the approximate total number of diarrhea cases by the sample size of 163 per facility. Therefore, every 4 th case of diarrhea (Muriranja`s hospital) and 5th (Murang`a Hospital) were selected until a sample size of 163 was reached from both hospitals.

Sample size determination:
Applying the formula for estimating the population proportion with specified relative precision described by Daniel 1999 [8] setting the α at 0.05, and a detection rate of 12.1% for children below five years infected with diarrheal disease in Murang`a County [7], a total of 163 children were recruited to achieve 0.95 power.

Data collection instruments:
The procedure that was used in data collection included structured data collection instruments that involved administering questionnaires directed to child`s caretaker and laboratory request forms.
Validity: Pre-testing was conducted in the two hospitals prior to validate the research methods and tools. Controls were run whenever necessary.
Sample collection: Diarrheal stool samples were collected on the day of presentation at the Hospitals using well labelled sterile leakproof polypots. Stool appearance was recorded in the Laboratory on the study questionnaire entries and request forms that matched the specimen identification number and then cultured in Cary Blair medium (Oxoid, United Kingdom) which was then properly sealed, labeled and stored at 4-8 º C for one day. Samples were disposed as per standard operating procedure of infectious material. On the 2 nd day, the cultured transport media was put in a cool box with frozen ice packs and shipped within 3 hours to the Kenya Medical Research Institute (KEMRI), Center for Microbiology Research in Nairobi.

Bacterial culture, isolation and identification
Escherichia coli (E. coli): E. coli species were detected according to the methods described in the Bacteriological Analytical Manual (BAM) [9]. 25g of sample was weighed into 225mL of Brain Heart Infusion (BHI) broth with a dilution factor of 1:10 and incubated briefly at room temperature (RT) shaking periodically then allowing sample to settle. The medium was decanted into a separate sterile bottle and incubated at 35 º C for 3 hours. The content was then transferred into a sterile container containing 225mL double strength tryptone phosphate broth and incubated for 20 hours at 44 º C. A loop-ful of the broth was then streaked on pre-incubated MacConkey agar plate and incubated for 18-20 hours at 35 º C.
Morphology and biochemical characteristics was followed as described by Feng et al. [9].
Salmonella species: Salmonella species were detected according to the methods described in the BAM [16]. About 25g samples were dissolved in about 225mL of sterilized buffered peptone water (BPW), blended, and incubated at 37 º C for 16-20 hours. About 10mL from the incubated BPW culture was selectively enriched into the 100mL sterilized Selenite Cystine Broth and incubated again at 37 º C for 24-48 hours. After incubation, 1 loop full inoculum from the selective enrichment culture was streaked onto the pre-incubated Bismuth Sulfiite Agar (BSA) and Xylose Lysine Deoxycholate (XLD) agar plate. Morphological identification followed the description outlined by Wallace et al. [16]. Further confirmation of biochemical reactive cultures was done by agglutination test with Salmonella polyvalent (O) somatic antisera as described by Wallace et al. [16].
Shigella species: Shigella species were detected according to the methods described by Andrews and Jacobson [17]. About 25g sample was aseptically weighed into 225 mL Shigella broth in which 0.5ug/mL novobiocin was incorporated and incubated at 37 º C for 18-20 hours. One loop full inoculum from the Shigella broth culture was streaked on the pre incubated MacConkey and XLD agar plate and incubated at 37 º C for 18-24 hours. Then the suspected colonies were identified by their cultural, morphological, and biochemical characteristics as described by Andrews and Jacobson [17].
Vibrio cholera: Vibrio cholera was detected following the procedure as described in the BAM [18]. About 25g samples were blended with 225mL sterilized APW and incubated at 37 º C for [16][17][18] hours. One loop-ful inoculum from the APW culture was streaked on the pre incubated TCBS and CPC agar plate and incubated at 37 º C for 24 hours. The suspected colonies were identified by their cultural, morphological, and biochemical characteristics as described by Kaysner and Angelo [18].
Yersinia species: Yersinia species was detected following the procedure as described in the BAM [19]. About 25g samples were blended with 225mL of Peptone Sorbitol Bile Broth (PSBB), homogenized and immediately incubated at 10 º C for 10 days. The enrichment broth was then recovered from the incubator and mixed lightly. Further, one loop full of enrichment was incorporated into 0.1mL 0.5% potassium hydroxide (KOH) in 0.5% saline then mixed lightly. One loop full was then streaked on MacConkey plate and another to Cefsulodin-Irgasan Novobiocin (CIN) plate and both were incubated at 30 º C for 24-48 hours. Morphological and biochemical identification of Yesinia enterocoloitica was performed following procedure described by Weagant and Feng [19]. Klebsiella species: Klebsiella species was detected following the procedure as described by Cheng et al. [21]. Stool samples were inoculated on both MacConkey and Simmons citrate-inositoltryptophan and bile salts (SCITB) agar media which were then incubated at 37 º C for 48 hours. Typical colonies that appeared yellow on SCITB were picked for spot indole test and TSI.

Results
Bacterial etiology of diarrhea among the study population: Table 2  Bacterial diversity: Figure 1 and Figure 2  weeks prior the Kenya health survey [6]. A much lower isolation rate of 1.4% from what was observed from our study was reported in Beijing [35], probably due to the difference in industrialization.
Aeromonas: Aeromonas species accounted for a prevalence of 8(4.3%) from the study population, almost near to what was reported in a study among Eastleigh refugee children below 5 years [23]. A lower prevalence of 2.0% compared to 4.3% from our finding has been reported among children in China [45]. Isolation of Aeromonas species was specific only to children between 3-4 years and these finding concur data output of a study done in Pakistan and Bangladesh that documented the bacteria been the leading bacterial pathogen as a cause of diarrhea with a peak between 3-5 years [46]. water are the major drivers to the infection [23]. The tendency of toilet training at the age of above 3 years is common which can be hypothesized to elevate fecal oral transmission of this pathogen.

Yersinia enterocolitica:
In conformity with the finding of the study, Yersinia species was recovered from stool samples among infants having diarrhea in Denmark [47]. Infants (below 12 months) were 50% infected with Y. enterocolitica in this study. Likewise, children between 3-4 years were equally 50% infected which can be related to a study that reported children between 3-4 years were more associated to the bacteria [47]. One possibly that this bacteria was found common among infants below one year could be due to their unchallenged immune system and therefore more likely to be inflicted. Infection by other enteric pathogens elevates the infection by Y. enterocolitica as well as domesticated animals such as dogs, cats, pigs and other bovine have been shown to harbor the bacteria [48]. Older children above two years interact more with such Page number not for citation purposes 7 animals which may explain the phenomenon that this group were more infected with the bacteria.
Citrobacter freundii: Citrobacter species are usually thought to be commensal organisms, though some species have acquired specific virulence genes hence enabling them cause diarrhea. More than double-fold prevalence (3.95%) from the finding of the current study was reported in Addis Ababa, Ethiopia and this bacteria was linked to cause childhood diarrhea [39]. Only the female were found infected with C. freundii and those between 1-3 years among study participants. This finding contradicts an observation in a study that showed male (66.7%) were more infected than female (33.3%) [38]. Klebsiella: Male and female participants were equally infected by K. pneumonia with majority of the infected below one year (75%) similar to outcome from a different study that observed children below 12 months were more infected with Klebsiella [38]. K. oxytoca isolation was less common among the male (25%) and only children between 1-2 years (50%) and those between 4-5 years (50%) were found infected. K. pneumonia and K. oxytoca is normal flora of the gut but can cause of diarrhea in human.
Experience from other parts of the world has confirmed that K.
pneumonia was a cause of bloody diarrhea following negative results of other enterobacteriaceae but isolation of K.
pneumonia from pure colonies was evident [49]. K. pneumonia (25%) were isolated from stools that were blood stained and the other 75% isolates were from watery stools (p=0.001).
From data output of this study, it is possible to tentatively assume that K. pneumonia is a potential cause of diarrhea among the study participants unlike K. oxytoca that was found not to have any statistical evidence of significance with the diarrhea (p=0.495).     Table 4: Distribution of pathogenic bacteria by stool appearance Figure 1: Bacterial diversity using Shannon Weaver Diversity Index across participants' age groups