Mixed Aetiology of Diarrhoea in Infants Attending Clinics in the North-West Province of South Africa: Potential for Sub-Optimal Treatment

Routine diagnostic methods for the aetiologic agents of diarrhoea in most developing countries are usually not sensitive enough, leading to under-diagnosis. Thus, this study investigated possible mixed diarrhoeal aetiology by using cultures and real-time polymerase chain reactions (PCR) in children younger than four years old in the Northwest Province, South Africa. In total, 505 stool samples were collected from symptomatic and asymptomatic children who were attending three clinics and the Brits hospital in Madibeng District, between September 2016 and December 2017. Rotavirus, norovirus, Campylobacter, Arcobacter, and diarrhoeagenic Escherichia coli (DEC) were targeted. Campylobacter spp. (24.6%), Arcobacter (15.8%) and DEC (19.6%) were detected using PCR; only Campylobacter spp. (29.7%) and DEC (26.9%) were detected through the culture. Campylobacter jejuni (36%), Campylobacter coli (28%), Campylobacter upsalensis (12%), and Arcobacter butzleri (15.8%) were the only spp. of Campylobacter and Arcobacter identified. The eaeA gene (31.4%) of enteropathogenic E. coli/enterohaemorrhagic E. coli (EPEC/EHEC) was the most prevalent DEC virulence gene (VG) identified. Rotavirus and norovirus were detected at 23.4% and 20%, respectively. Mixed viral aetiology (7.3%) and the co-infection of A. butzleri and Campylobacter (49%) were recorded. A mixed bacterial-viral aetiology was observed in 0.6% of the specimens. Sensitive diagnostic procedures like PCR should be considered to provide the best treatment to children experiencing diarrhoea.


Introduction
Diarrhoea is the most common clinical manifestation of gastrointestinal diseases and is defined as the frequent discharge of unformed, loose, or watery stool, usually three or more times in 24 h [1,2]. Diarrhoea is associated with a high morbidity and mortality and is among the leading causes of the and may not be as sufficiently sensitive as polymerase chain reaction (PCR) assays [32]. Campylobacteraceae are fastidious microorganisms, and their laboratory identification is often challenging because it involves the use of selective and differential culture media, as well as biochemical and serological tests, which may not differentiate the species [33].
The culture of DEC is also problematic since even selective media and most biochemical assays do not distinguish DEC from the non-pathogenic strains. Additionally, serotyping does not always correlate with pathogenicity. However, several virulence factors and DNA sequences associated with DEC have been identified and can only be determined by detecting the presence of genes coding for specific virulence factors that are absent in non-pathogenic strains [19]. These genes have been characterised and can be identified by PCR [34]. Efforts to cultivate noroviruses and rotaviruses in cell culture are not routine, although electron microscopic characterisation and serologic responses are used to identify these viruses; clear and specific identification can only be achieved through PCR [13,35].
In many developing countries, the aetiology associated with diarrhoeal diseases are often unknown [4]. This is mainly due to a failure to isolate the causative agents in the laboratory or isolating one agent and leaving the other; both diagnostic factors affect the management and treatment of the disease. [32,36]. To effectively control diarrhoeal diseases in any community, an accurate understanding of the relative aetiologic agents and early detection of these agents using appropriate diagnostic tools is necessary. Therefore, the goal of this study was to use culture-based and real-time PCR assays to detect rotavirus, norovirus, DEC, Arcobacter spp., and Campylobacter spp. to determine their involvement in the aetiology of gastroenteritis in paediatric patients in the North-West Province, South Africa.

Distribution of Participants According to Age, Mode of Feeding and Presence/Absence of Diarrhoea
A total of 505 stool specimens were collected from children ≤4 years. Of the 505 participants, 350 had diarrhoea or symptoms of gastroenteritis at the time of sample collection, and 155 were healthy controls with no episode of diarrhoea or gastroenteritis on the day sample was collected. The number of children recruited in each category is shown in Table 1. Children under the age of one were the most recruited, and they make up for more than 50% of the gastrointestinal cases.

Prevalence of Aetiologic Agents in Symptomatic Participants
All studied microbial species were more associated with diarrhoea ( Table 2). None of the viruses was isolated from participants with bloody diarrhoea.

Prevalence of Arcobacter and Campylobacter spp. in the Study
The amplification of the Arcobacter using PCR allowed for the identification of Arcobacter in 15.8% of the samples. Arcobacter butzleri was identified more from female participants 52.5% (42/80) than male participants 47.5% (38/80). Arcobacter cryearophilus and Arcobacter skirrowii were tested but were not found in the study population. According to the method of feeding, A. butzleri was more prevalent among babies who received a mixed diet 52/80 (65%) compared to those on an exclusive breast-feeding diet 28/80 (35%). Based on clinical presentation, A. butzleri was significantly associated with infants with diarrhoea 56/80 (70%) compared to those without 30% (24/80) (p = 0.044). Regarding age, A. butzleri was only isolated in children within the 0-24 months age range.

Prevalence of DEC Virulence-Associated Genes
Of all the samples that tested positive for E. coli (PCR amplification of the mdh gene), DEC virulence genes were detected in 88 of these samples. Similarly, from the 136 pure E. coli isolates, PCR identified DEC-associated virulence genes (VGs) in 125 of them. The rate of detection of virulence genes from pure isolates and whole samples is shown in Figure 1.
The ST and LT genes were not amplified in the DNA isolated from the whole samples, and the stx2 and eagg genes were not detected in the pure isolates. The eaeA gene was the most commonly detected DEC-associated VG from both the whole sample and pure culture, while stx1 was the least detected gene. A combination of the stx2 and flicH7 genes was detected in three samples, while the eaeA and stx2 genes were observed in 12 samples. The co-detection of the eagg, stx1, flicH7, and eaeA genes was also recorded.
Additionally, the various DEC VGs were not equally distributed in the children with respect to the presence or absence of symptoms, as well as their mode of feeding (Table 3).

Prevalence of Rotavirus and Norovirus
The prevalence of rotavirus infection in the children recruited into this study was 23.4% (118/505). The age of children affected with rotavirus was restricted to those between the ages of 0 and 25 months. The majority of rotavirus infection, 82.2% (97/118), occurred in children in the 0-12 months age group.
On the other hand, norovirus infection was recorded in 20% (n = 101) of the children participating in the study. The age of the children affected was between 0 and 30 months. The highest rate of infection was noted in children in the age group of 0-12 months (84.2%; n = 85). As opposed to the gender distribution regarding rotavirus infection, more male children were infected with norovirus (53.5%; n = 54), and the gender difference was statistically significant (p = 0.045). Of the 101 detected cases of norovirus infection, 76.2% (n = 77) were from symptomatic cases, while 23.8% (n = 24) were from asymptomatic cases. Vomiting and fever were reported in 39.6% (n = 40) and 54.5% (n = 55) of the norovirus-positive cases, respectively.

Prevalence of Mixed Aetiology
The mixed aetiology of A. butzleri and Campylobacter spp. was found in 9.7% (n = 49) of the samples, while the mixed infection of all the tested bacteria was observed in 1.2% (n = 6) of the samples. Of these, three babies were under the age of six months, all six children were on a mixed feeding diet, and four of the six children (66.6%) had bloody diarrhoea.
Of the children infected with rotavirus (n = 118) and norovirus (n = 101), 16.9% (n = 37) were confirmed to have viral co-infection and were mainly seen in the diarrhoeal cases (75.8%; n = 28). These children were in the age group of 0-24 months. Episodes of fever were recorded in 48.6% (18/37) of the co-infected children, most of whom (70.3%; n = 26) were on a mixed-feeding diet and 51.4% (n = 19) had received the rotavirus vaccine.
Bacterial and viral co-infection was detected in three children who were aged three, four and 18 months, and two of those presented with bloody diarrhoea, fever, and vomiting. One of these three children had received two doses of the rotavirus vaccine, and the other had not been vaccinated against rotavirus.

Discussion
Despite advances in the understanding of pathogen transmission patterns, diarrhoeal diseases remain the leading causes of morbidity worldwide. A vast number of pathogens can cause diarrheal diseases, but only a few account for the majority of the disease burden worldwide [1]. This study was undertaken to determine the aetiologic agents responsible for diarrhoea in the North-West Province of South Africa by analysing faecal samples of symptomatic and asymptomatic children under the age of five years. The study focused on Campylobacter species (C. coli, C. jejuni and C. upsalensis), DEC virulence-associated genes (ST, LT, stx1, stx2, flicH, ipaH, eaeA and eagg), A. butzleri, rotavirus, and norovirus.
Due to the fastidious nature of and special growth requirements for Campylobacteraceae, the current study used both conventional culture methods and real-time PCR assays to detect bacterial agents. Using only the culture method, Campylobacter spp. (29.7%) and E. coli (26.9%) were identified in the collected stool samples. The negative stool samples were further analysed using the 16S rRNA gene for Campylobacter identification, 23S rDNA gene for Arcobacter spp and mdh gene for E. coli identification. This method identified Campylobacter in 24.6% (150), Arcobacter in 15.8% (80), and E. coli in 19.6% from culture-negative samples. These results confirm that, although culture methods are suitable, they are not sufficient to use alone in epidemiological studies, because they might not give a true reflection of the prevalence of the aetiologic agents involved in any given situation. For example, 24.6% of Campylobacter spp., 15.8% of Arcobacter spp., and 19.6% of E. coli would have gone undetected when using only the culture method. These results corroborate the report of Bessede et al. [37], who reported a low sensitivity of culture over PCR. Similarly, Lawson et al. [38] identified 13.1% of Campylobacter spp. from faecal specimens that were previously found to be negative when using the culture method [38]. These differences are due to the higher sensitivity of PCR over culture. Based on the identification of DNA, PCR also identifies the presence of free DNA and viable but non-culturable bacteria that would otherwise not be detected when using culture-dependent techniques. Nevertheless, the detection of these isolates suggests the exposure of the participants to these pathogenic microbes.
The detection frequency of all the aetiologic agents identified in this study showed that Campylobacter spp. were the major agents of diarrhoea within the studied community, followed by E. coli, rotaviruses, noroviruses, and A. butzleri, as they were the most isolated from diarrhoeal samples. In agreement with this study, Campylobacter, E. coli, norovirus, and rotaviruses have been recognised agents of diarrhoea both in community and clinical settings [39][40][41]. The higher detection rate of these aetiologic agents among the symptomatic babies in this study confirms that these agents might have contributed to causing diarrhoea in the studied babies. Higher prevalence of aetiologic agents were noticed more in the samples of babies on mixed feeding diets compared to those on exclusive breastfeeding diets. These results support the previous finding that babies on exclusive breastfeeding diets are less prone to diarrhoea diseases [42,43].

Occurrence of Campylobacter
The prevalence of Campylobacter reported in the present study is slightly higher compared to the 30.4% and 38.9% previously reported in South Africa [44,45]. This difference might be attributed to the two methods used for isolation. However, it reaffirms the presence of Campylobacter infection among children in South Africa and the need to increase surveillance. Three prominent Campylobacter spp. were identified, with C. jejuni being the dominant species. A higher percentage isolation of C. jejuni over other Campylobacter spp. has been reported in studies all over the world. In developed countries such as the USA, 75% of C. jejuni and 4% C. coli have been reported [46]. Similarly, 85% C. jejuni and 15% C. coli have been documented in South Africa [45]. Infection due to C. jejuni, in most cases, can lead to neurological sequalae such as Guillain-Barre syndrome (GBS) [47]. C. jejuni has been noted to be the most prominent of all Campylobacter species worldwide, and it has been reported to be responsible for most human campylobacteriosis [37]. In the current study, although C. jejuni was the dominant species identified in all the samples, a higher prevalence of C. coli was observed among children with clinical symptoms compared to the other Campylobacter spp. Additionally, the 9.9% of identified C. upsalensis was noted more in samples from asymptomatic children. C. upsalensis is mostly associated with household pets such as cats; though rarely found in clinical specimens, it was reported to be the cause of multiorgan failure in a patient in Japan [48].
The trend of Campylobacter infections according to age observed in this study supports earlier reports that Campylobacter infections are prevalent among children under the age of two years in developing countries, and these infections decrease with an increase in the age of the children [49,50]. In studies including age groups, a Campylobacter infection is usually high among the children participants. For example, in Peru, a 41.3% prevalence was reported in children [6] and in India, it was found to be associated with 70% of diarrhoeic children [51]. In South Africa, it was reported in 30% of children under the age of two years [45]. The results of all these studies showed that children are more prone to Campylobacter infection, and control programs targeting the eradication of infections due to Campylobacter should prioritize children and their caregivers, as diarrhoea due to Campylobacter has been reported to be associated with a growth shortfall in children [52].

Prevalence of Arcobacter
Though the presence of other Arcobacter species was investigated in the current study, only A. butzleri was detected. However, this specie was the least detected pathogen in our entire study, with a prevalence of 15.8%, and it was only identified through PCR. These results agree with a case report presented by Arguello et al. in which the authors failed to identify causative pathogens from the stool sample of a patient with bacteraemia through the conventional methods used in their laboratory; however, subsequent testing with molecular assays revealed the presence of A. butzleri [53]. Though there is no defined mortality or morbidity rate for A. butzleri, it has been identified in many studies as an agent of diarrhoea and bacteraemia, particularly in children and immunocompromised individuals [54,55]. Depending on the prevalence of A. butzleri, exposure to risk factors within the community, or the method of identification, different percentages of A. butzleri have been reported. In a previous study in South Africa, 6.5% of A. butzleri was reported [45]; additionally, percentages of 56.7% from diarrhoeic and 45.5% from non-diarrhoeic specimens were reported in Canada [56]. Virulence genes have been reported in A. butzleri in several studies [57][58][59].
The presence of these virulence genes in A. butzleri warranted the international commission of microbiological specification of foods (ICMSF) to categorize A. butzleri as a serious hazard to humans [59,60].

Prevalence of DEC Pathotypes
The pathogenic strains of E. coli are well-known causes of severe diarrhoea outbreaks worldwide [61]. In this study, eight DEC virulence genes representing five E. coli pathotypes were analysed. A higher prevalence of these pathotypes was observed in diarrhoeic stool samples compared to non-diarrhoeic ones. The most prevalent gene was the eaeA (31.4%), which is common in EHEC and EPEC pathotypes. This gene was also more associated with diarrhoea compared to non-diarrhoea samples. For many years, typical EPEC were considered to be a major agent responsible for diarrhoea in children, particularly in developing countries. However, a few sporadic reports have indicated the association of atypical (aEPEC) with children [62][63][64]. In developing countries, it has also been recognised as an agent of infantile diarrhoea, and it can also be found in non-diarrhoea samples [65,66]. In South Africa, a prevalence of 20% was reported in children with diarrhoea [61], while a prevalence of 19% has been reported in Mozambique [67]. In this study, it was observed that the eaeA gene was expressed more from samples of children without fever and vomiting, agreeing with previous a study in Denmark where 44% of infected children did not show any sign or symptom of diarrhoea [68]. Such children could be carriers and transmit the infection to other vulnerable children, especially in resource-poor settings with poor hygiene and sanitation facilities.
Following aEPEC/EHEC, the eagg gene of the EAEC pathotype was the second most identified gene (31.4%) in the DEC-positive samples. It was mostly associated with diarrhoea and a mixed feeding diet. Similar to these findings, EAEC was the prominent pathotype detected in a previous study in South Africa; it was found in 57.3% of those studied [61]. Contrary to the present study, it was detected majorly from non-diarrhoea stools. EAEC pathotypes have emerged as enteric pathogens responsible for acute and persistent diarrhoea and may cause malnutrition and growth defects in children [69,70]. These pathotypes are known agents of traveller's diarrhoea [71,72]. Though they are frequently isolated in asymptomatic cases, reports have shown that even in the absence of diarrhoea, they are associated with intestinal inflammation, which may impact child's development [73]. In developing countries, the prevalence of EAEC pathotypes among diarrhoea cases ranges from 4.5% to 41% [69,71,74,75]. Studies carried out in Europe and Romania have implicated the EAEC pathotype of E. coli as the common bacterial cause of diarrhoea [76][77][78].
In addition to the eagg gene, the EHEC virulence genes stx1, flicH7, and stx2 were expressed in 9%, 17.8%, and 13.6% of those studied, respectively. EHEC pathotypes, also known as Shiga toxin E. coli (STEC), are a public health concern because of the severity of their infections. Their primary virulence factor is the Shiga Toxin (stx), although they produce several other virulence genes, with the stx1 and stx2 genes being the major ones, used to disrupt host protein synthesis, causing apoptotic cell death [79]. Studies performed in other regions have reported low prevalence rates of the STEC pathotypes, such as in Argentina (6.2%), Mexico (2.5%), Colombia (1.8%), and Brazil (0.5%), while in South Africa, a rate of 3.8% has been observed in non-diarrhoea stools [61]. The expression of stx by E. coli is associated with the development of haemolytic-uremic syndrome (HUS) in up to 10% of STEC infections [80,81]. Though HUS can develop in patients of any age, it is a potential life-threatening condition and has been frequently observed in children [81,82]. It is a recognized cause of renal failure and hypertension in children in the UK [83].
The IpaH gene, which is common to Shigella and the EIEC pathotype, was among the least detected genes in this study. It was found in 16.4% of the samples, of which 68.5% were from diarrhoea samples. These results suggest that the EIEC pathotype might be one of the active agents of diarrhoea among the studied population, although to a lower degree. Several studies have reported on the prevalence of the ipaH gene [84,85]. EIEC pathotypes are known to present with diarrhoea and dysentery, similar to Shigella spp. in humans [19].
The ETEC pathotype, produced the ST (heat stable) or in combination with LT (heat labile) genes, is also a major contributor of diarrhoea burden in both developed and developing countries [1,18]. These genes were expressed in 9.3% (ST) and 5.6% (LT) in this study and were more associated with clinical symptoms. These results concur with the report of Mclamb et al. [86]. In their study, they investigated the clinical onset and severity of intestinal disorder exhibited by ETEC in humans, and they observed that ETEC was capable of inducing serious clinical symptoms including persistent diarrhoea, loss in weight, fever, intestinal injury [86], and growth impairment [26].

Prevalence of Viral Aetiology
Viral agents were detected in 36% of the children in the current study. All the rotavirus-infected children were under the age of three years, and 57.6% had received the rotavirus vaccine. The most affected (82.2%) were in the 0-12 months age group. These results correspond with a recent published report on the effectiveness of the rotavirus vaccine in South Africa [7]. Prior to the rotavirus routine immunization programme in South Africa, rotavirus accounted for 35% of all hospital admissions [7]. The current study found a 23.4% prevalence of rotavirus infection, and over 50% of the children had received the rotavirus vaccine. These results supports the fact that the rotavirus vaccine provides protection against rotavirus-associated diarrhoea [7]. Though mortality due to rotavirus has decreased, morbidity seems to be stable, possibly because the available vaccine may not be effective against emerging rotavirus genotypes. Previous studies have reported that, after the introduction of the vaccine in some African countries, genotypes that were not present during the vaccine formulation emerged [87][88][89]. Additionally, during the rotavirus vaccine clinical trial, vaccine efficacy was reported to be low in developing countries compared to developed countries [7].
With the increased use of the rotavirus vaccine, norovirus has been recognised as a leading cause of medically attended diarrhoea in children and adult worldwide. The prevalence of norovirus in many African countries is unknown, with varying percentages of 22% in Botswana [90], 21% in Libya [91] and 18.9% in Kenya [92] being reported. The present study detected norovirus in 20% of the study population. These variations are mostly due to the study duration, population groups, sample size, and inclusion of asymptomatic controls in the studies. In the present study, norovirus was detected only in children younger than two years old; the highest infection (84.2%) was recorded among the children under 12 months of age. These results are in agreement with previous studies in South Africa [93] and Brazil [10], where norovirus infection was frequently detected in children younger than two years old. Though norovirus infections are perceived as self-limiting, infection in children could cause serious symptoms that may require hospitalization. Studies have shown that an infection with norovirus only provides short term immunity, such that pre-existing antibodies will not proffer protection against reinfection [94].

Prevalence of Mixed Aetiology
A high prevalence of mixed aetiology with different DEC pathotypes and Campylobacter spp. was observed in the present study. A mixed aetiology raises the question of whether a single pathogen is responsible for illness or whether several pathogens act in synergy. Some studies have confirmed a mixed aetiology of pathogens in patients. In India, for example, 11.3% of a mixed aetiology of rotavirus and other pathogens was found [63]. In Tanzania, Moyo et al. [95] reported a 20.7% mixed aetiology of DEC among children younger than five years old [95]. Of the 219 children positive for rotavirus or norovirus, 16.9% showed mixed aetiology. In the present study, mixed aetiology was observed both in children with diarrhoea and healthy babies. Other studies have noted that a mixed aetiology of two or more agents and up to five associated pathogens could be found in a host [63]. The present study found that mixed aetiology was more prevalent among children on a mixed feeding diet. This might be a confirmation that gastrointestinal pathogens circulate in the population with ease and that the environment, food, and water may be vehicles of transmission of these agents [96]. It has been reported that the occurrence of a mixed aetiology of infections in a patient is common among individuals living in a community where good sanitation and hygiene is compromised [97]. This ties with the current study, as all the babies that participated were from a rural community; this result could be an indication that environmental factors such as infected stray animals, poor sanitation, and an inconsistent water supply could be some of the reasons for the high mixed aetiology.
Another reason for the detection of a mixed aetiology of an infection could be the method of identification used. In the current study, the culture method and PCR with DNA extracted directly from the samples were used in parallel, which reduced the error of missing positive samples via culturing alone and biochemical identification. However, it should be noted that although PCR is a highly sensitive method, its complexity may be a limitation for its use in routine diagnostic laboratories. As such, the results of this research are meant to stimulate the need for the development of diagnostic methods that would be inclusive of more aetiologic agents than just the ones currently tested.

Specimen Collection
Stool samples were collected from September 2016 to December 2017 from children ≤4 years old, with or without diarrhoea, and attending the Brits District Hospital, Oukasie, Lethabile and Bopang Clinics in the Madibeng District of the North-West Province, South Africa. Diarrhoeal samples were consecutively collected from patients who presented with acute diarrhoea and were not on any antimicrobial therapy, while non-diarrhoeal samples were taken from babies who did not present with any gastrointestinal complaints. A single sample was collected per participant, and participants whose samples had been collected earlier in the study were excluded in order to avoid multiple collections of samples from one participant. Samples were collected in sterile, wide-mouthed, 10 mL bottles with tight-fitting leak-proof lids. These were carefully labelled before being put in a cooler box containing ice packs and transported to the Microbiology Laboratory of the Council for Scientific and Industrial Research (CSIR), Pretoria, within two hours of collection for further processing.
After proper counselling about the study, informed consent was obtained from the parents/guardians attending to the study participants. Detailed personal history, diarrhoeal episode, and associated signs and symptoms were recorded on a questionnaire before collecting the stool samples.

Isolation of Campylobacter and Arcobacter Species
The methods described by Bessede et al. [37] and Shah et al. [98] were employed for isolation of Campylobacter and Arcobacter, with slight modifications. Briefly, stool samples were enriched with Campylobacter growth supplement (ThermoScientific, Johannesburg, South Africa) in a Bolton broth (BB) and incubated at 37 • C for 24 h in a microaerophilic atmosphere (MAE) (CampyGen TM , ThermoScientific, Johannesburg, South Africa). Arcobacter was isolated in an Arcobacter broth (AB) supplemented with CAT (cefoperazone, teicoplanin and amphotericin B; Oxoid, Johannesburg, South Africa) and aerobically incubated at 30 • C for 24 h. After incubation, 10 µL of the BB was dropped on a 0.65 µm filter paper placed on a tryptic blood agar plate (TBA) and allowed to incubate at room temperature for 30 min. The filter papers were carefully removed, and the plates were incubated at 37 • C for 24 h in MAE for the cultivation of Campylobacter. For the growth of Arcobacter, 10 µL of the AB was pipetted onto a 0.65 µm filter paper on a blood agar (BA) plate, incubated for 30 min at room temperature, and then the filter papers were carefully removed and the plates were aerobically incubated at 37 • C for 24 h [98,99].

Preparation of Positive Controls
Bacterial-specific genomic DNA was used as a positive amplification control in the real-time PCR screening process. Genomic DNA was extracted from pure cultures of C. jejuni [American Type Culture Collection (ATCC) 33560] and C. coli (ATCC 33559) using the method described [99]. The diarrhoeic E. coli strains used in this study included EAEC (Strain 3591-87), EIEC (ATCC 43892), EPEC (strain B170) and EHEC (strain O157:H7) as previously reported [100]. These E. coli strains were previously biochemically confirmed in the Department of Microbiology of the University Stellenbosch, South Africa. The presence of the individual virulence genes was also confirmed by using PCR before conducting the experiments.

Genomic DNA Extraction and PCR Identification of Arcobacter and Campylobacter Species
The total faecal genomic DNA was extracted from stool specimens using the ZR Faecal Microbe DNA Miniprep TM kit (Zymo Research Corp., Irvine CA, USA) according to the manufacturer's instructions and from pure culture isolates. The quantity and purity of the DNA obtained were assessed using a Nanodrop 1000 instrument (Nanodrop, San Francisco CA, USA). Extracted DNA was kept at −20 • C and used as template DNA for the real-time PCR.
Genus-and species-specific primers were used for the identification of Campylobacter together with the cycling conditions have previously been described [101,102]. Additionally, the identification of Arcobacter to the genus level and the simultaneous identification of A. butzleri, A. cryaerophilus and A. skirrowii was as previously described by [103].

Amplification of A. butzleri Virulence Gene
All confirmed A. butzleri samples were assessed for the presence of the cadF (Campylobacter invasion protein B) and ciaB genes (Campylobacter adhesin to fibronectin F), with primers and conditions described by Ghunaim et al. [104]. For each reaction, a total of 10 µL of SensiFAST TM HRM Master Mix (Bioline GmbH, Luckenwalde, Germany), to which 0.5 µL (final concentration 0.5 µm) of each of the primer sets (Forward and reverse), 3 µL of template DNA and 1 µL of NF H 2 O was added.

Isolation of Escherichia coli from Stool Samples
The isolation of E. coli from stool samples was carried out using the Colilert-18 Quanti-tray/2000 (IDEXX Laboratories, Inc., Johannesburg, South Africa) Briefly, 2 g of stool sample was thoroughly mixed with 100 mL of distilled water in a sterile vessel, processed and incubated following the manufacturer's instructions. After incubation at 37 • C, presumptive E. coli isolates were harvested from fluorescent Quanti-trays wells and streaked onto an Eosin methylene blue (EMB) agar to obtain pure colonies according to the method described by Abia et al. [105]. Purified colonies were stored at −80 • C in a 50% glycerol for further analysis. 4.7. DNA Extraction and Detection of E. coli Virulence Genes (VGs) from Pure Culture DNA was extracted from selected E. coli colonies by using the heat lysis method and whole stool sample using the ZR Faecal Microbe DNA Miniprep TM kit (Zymo Research Corp., USA) according to the manufacturer's instructions. Extracted DNA templates were tested for the presence of the malate dehydrogenase (mdh) gene, which is found in most E. coli strains. Isolates that harboured the mdh gene were tested for the presence of eight VGs-eaeA (EPEC/EHEC), eagg (EAEC), ipaH (EIEC), ST and LT (ETEC), stx1, stx2 and flicH7 (EHEC)-as previously described [106,107]. All the PCR assays included a positive control consisting of DNA from a reference strains obtained from the Microbiology Laboratories of the Natural Resources and the Environment (NRE) and CSIR, as well as previously characterised by Abia et al. [101]. A reaction mixture void of template DNA was also included in each PCR assay as a no template control.

Identification of Viral Pathogens
Stool samples were prepared by suspending 2 g of stool into 180 µL of phosphate buffered saline (PBS) and homogenized using a vortex. The total ribonucleic acid was extracted from 140 µL of the stool suspension using the QIAamp mini RNA kit (QIAGEN, Tokyo, Japan) according to the manufacturer's instruction.
The identification of rotavirus and norovirus was performed using one-step SensiFast TM SYBER No-ROX kit (Bioline GmbH, Luckenwalde, Germany) and primers published by Onori et al. [107]. A Corbett Life Science Rotor-Gene TM 6000 Cycler (Qiagen, Hilden, Germany) was used for the PCR assays under the following conditions: reverse transcription at 45 • C for 10 min, polymerase activation at 95 • C for 2 min followed by 40 cycles of denaturation at 95 • C for 5 s, annealing at 60 • C for 10 s, and an extension at 72 • C for 5 s. A melt curve analysis of the amplified product was performed using the Rotor-Gene™ real-time analysis software, version 6.1 (build 93) (Corbett Life Science (Pty) Ltd., Sydney, Australia).

Statistical Analysis
The results were analysed using the SPSS (Statistical Package for the Social Sciences; IBM Corporation, Armonk, NY, USA) software, version 20.0. The chi-square test was used to determine the relationship between the different PCR results obtained and other parameters such as age, diarrhoeal status, and mode of feeding. The statistics were considered significant when the p-value was less than 0.05.

Conclusions
The current study indicates that Campylobacter, Arcobacter spp., DEC, rotavirus and norovirus are among the contributors of diarrhoea in the Madibeng District of the North-West Province of South Africa. The results suggest that culture-based methods and a PCR assay should be used simultaneously in the laboratory for the diagnosis of diarrhoea to avoid missing some aetiologic agents, as this could adversely affect treatment outcome. The results also indicated the prevalence of mixed pathogens among children in rural areas and emphasised the need to develop approaches to examine mixed aetiology and to think beyond the one-organism one-disease concept. In addition to the investigated organisms, other bacterial pathogens that are not routinely screened for in the hospital setting may be responsible for a number of episodes of diarrhoea in babies. The prevalence of mixed pathogens in heathy individuals also prove that healthy children can act as carriers of virulent pathogens. There is, therefore, a need for the constant monitoring of aetiologic agents of diarrhoea in children. Knowledge of these agents is needed to set off future research involving inter-microbial interaction, as this could assist in defining the environmental factors that contribute to the mixed aetiology of diarrhoea infections within the same host. Funding: This research received no external funding; however, the student was supported by NRF and UNISA Research Directorate. The APC was funded by Sefako Makgatho Health Science University, South Africa-Prof CL Obi.