The study was conducted in two healthcare facilities, the University College Hospital (UCH) and Adeoyo Maternity Teaching Hospital (AMTH), both in Ibadan, Nigeria. The research was approved by the UI/UCH Ethics Committee with assigned number UI/EC/18/0335 and Oyo State Ethical Review Committee with assigned number AD13/479/1378. Patients’ parents or guardians provided informed consent prior to enrollment.

Children living with HIV and HIV-uninfected controls, aged 0-15 years, were recruited between January 2019 and February 2021. Children Living with HIV were recruited at HIV clinics or wards while HIV-negative controls were recruited at children emergency wards or children outpatient clinics. Recruitments were guided by the national guidelines for screening children for HIV and managing them (Federal Ministry of Health, Abuja, Nigeria, 2016). Briefly, children less than 18 months are screened for HIV using HIV DNA polymerase chain reaction (PCR). Children with positive HIV DNA PCR results are subsequently placed on HAART. Children with negative results get antibody test at 18 months after weaning. For children ≥ 18 months, antibody tests are carried out using Determine™ rapid kit (Abbott Diagnostics, Japan). If reactive, a confirmatory test is done using Stat Pak^® (ChemBio Diagnostic Systems, USA). Uni-Gold™ (Trinity Biotech, Ireland) is used as tie breaker in the event of discordant results.

Eligibility criteria for participation in this study include verified HIV status known to and confirmed by the attending medical personnel, passage of loose stools three or more times within a 24-hour period over the past three days and obtained informed consent from their parents or guardians.

Microscopic examination of the stool samples was done using the method of Winn et al. (2006). Drops of normal saline and iodine were separately placed on a clean glass slide and with an applicator stick, an aliquot of stool sample was picked and placed on the drops of normal saline and iodine to make a smear. A cover slip was placed over each wet preparation which was mounted for microscopic observation using the x10 and x40 objectives.

Modified Ziehl-Neelsen method (Fayer et al., 2000; Tahvildar-Biderouni and Salehi, 2014) was also employed to examine the samples for coccidian parasites Cryptosporidium parvum, Cylospora cayetanensis and Cystoisospora belli. Stool smear was made on clean glass slide and allowed to dry. The dried smear was fixed in methanol for three minutes and subsequently stained with strong carbol fuchsin for 20 minutes and then rinsed in tap water. The smear was counterstained with methylene blue for 60 seconds and rinsed with tap water, then dried. Examination of the slides was done using x40 and x100 objectives.

Detection of occult blood in the stool samples was done using a rapid kit (Cromatest^®, Linear Chemicals, Spain). This was done by transferring an aliquot of the stool sample into the specimen collection tube containing the extraction buffer and mixed together by shaking. Subsequently, three drops of the extracted stool sample were transferred to the test cassette and observed for positivity indicated by a red line showing on the region marked “T” (for test) as well as the region marked “C” (control) on the device. Specimens testing negative lacked the red line on the test region but showed the red line on the control region (Yamamoto and Nakama, 2000).

The stool samples were cultured on MacConkey agar and Eosin Methylene Blue (EMB) agar for the isolation of Escherichia coli and other enteric pathogens. For recovery of Salmonella and Shigella, the stool samples were first enriched overnight in selenite F broth before being sub-cultured onto xylose lysine deoxycholate (XLD) agar. Incubation was done at 37⁰C for 18-24 hours (Medina et al., 2010). For the detection of Yersinia species, stool samples were initially enriched in phosphate buffered saline (PBS) for a 21-day period at 4°C before being sub-cultured onto Yersinia selective agar (YSA) and allowed to incubate at room temperature for 24-48 hours. Isolates were purified via sub-culturing and cryopreserved in 50:50 Luria Broth-glycerol mixture.

Isolate identification was done by conventional biochemical tests and confirmation using the Vitek2 Identification system (Biomerieux Inc., Hazelwood, MO, USA) that allowed suspensions of the isolates in cassettes to undergo a 64-well reagent test for identification. Salmonella spp and diarrhoeagenic E. coli were confirmed/differentiated by PCR.

The DNA of isolates extracted using the Promega^® wizard genomic extraction kit was used to template PCR employing primers for the detection of virulence genes. The genes sought delineating different pathotypes of diarrhoeagenic E. coli and Salmonella spp. listed in Table 1 . Strains were deemed to belong to the respective pathotype if they contained at least one of the gene targets sought for that pathogenic type.

Data obtained in this study were analyzed using Chi square with p set at 0.05. In the many instances where expected values were <5, the Fisher’s Exact test was used.

In this study, only ten children living with HIV who had diarrhoea could be recruited during the study’s 25-month duration and 55 HIV-uninfected children with diarrhoea were recruited as controls. EAEC, EIEC, C. parvum and C. cayetanensis were the most frequently encountered pathogens in the study subjects, while STEC, Shigella and Yersinia were not detected at all. The full and repertoire of bacterial and parasitic pathogens detected/isolated from them are presented on Table 2 .

Other aetiologic studies performed in Ibadan and other locations in Nigeria have found diarrhoeagenic E. coli, Cryptosporidium parvum and Cystoisospora belli commonly associated with diarrhoea (Adesiji et al., 2007; Abdullahi et al., 2010; Saka et al., 2019; Karshima and Karshima, 2021; Akinlabi et al., 2022). These pathogens were all detected in stools from both children living with HIV and the of HIV- uninfected children who served as controls in study ( Table 2 ) with comparable recovery rates seen in both patient groups for almost all pathogens. An important exception is C. parvum, which was recovered from four of ten children living with HIV but only 4 (7.3%) HIV- uninfected children (p = 0.02, Table 2 ). C. parvum, is a major protozoan parasite commonly detected in immunocompromised individuals, particularly people living with HIV, including those in Nigeria (Botero-Garces et al., 2021; Karshima and Karshima, 2021). A meta-analysis of 131 studies by Wang et al. (2018) revealed low prevalence of Cryptosporidium (and Cystoisospora) among People Living with HIV in high income studies and a much higher prevalence, with diarrhoea association in sub-Saharan Africa. Although we found lower frequencies, Cyclospora cayetanensis is also commonly reported from people living with HIV (Stark et al., 2009; Feasey et al., 2011) and Cystoisospora belli is increasingly being recognized as a cause of diarrhoea in this sub-population (Casmo et al., 2018). As shown in Figure 1 at least one parasite was detected in 5/10 of children living with HIV but only 9/55 (16.4%) of children without HIV infection (p = 0.03) showing that parasite detection is associated with HIV.

EAEC was the most frequently detected pathogen in diarrhoeal stools among both children living with HIV and HIV-uninfected children. Reports from within Nigeria (Onanuga et al., 2014; Saka et al., 2019; Akinlabi et al., 2022) and elsewhere (Kotloff et al., 2013; Shah et al., 2016) show that, in the relatively infrequent instances where it is sought, EAEC is commonly the most frequently detected diarrhoeagenic E. coli pathotype in children with diarrhoea although it can also be recovered from children without diarrhoea (Franç et al., 2013; Platts-Mills et al., 2015). Pathogen and host factors may be important determinants of virulence and children without diarrhoea who carry EAEC may be at risk of nutrient malabsorption and other less visible sequelae (Nataro et al., 1992; Nataro et al., 1998; Platts-Mills et al., 2015).

Certain enteric bacterial pathogens, notably Shigella and ETEC have been prioritized for vaccine development. ETEC strains were recovered from children with and without HIV infection in this study, whereas, Shigella, though commonly reported as a pathogen of significance in Africa, was not found in this study (Breurec et al., 2018). While the study size was small, EIEC, which carry the Shigella invasive plasmid were recovered from children with and without HIV. Cabal et al. (2016) reported EIEC association with diarrhoea in patients from Spain. Meanwhile, recent studies in Nigeria have not recovered Shigella or reported relatively low recovery (Joseph et al., 2017; Akinwumi et al., 2021), adding to evidence supporting earlier speculation that this pathogen may be less important in this setting and therefore the importance of understanding pathogen epidemiologies in a range of settings (Lindsay et al., 2015; Okeke et al., 2016).

Overall, the genes delineating EAEC (aap. aggR and aatA) were the most commonly detected by PCR, further confirming the preponderance of EAEC among diarrhoeic children in our setting as previously reported by Saka et al. (2019) and Akinlabi et al. (2022) and elsewhere (Weintraub, 2007). In contrast to EAEC, EPEC and Salmonella spp. were less commonly recovered with other bacterial pathogens. Of the six (6) EPEC strains recovered, including the one isolate from a child living with HIV, three (3) were typical, that is had the EPEC adherence factor plasmid-borne bfpA gene, and three (3) were atypical, that is lacked the gene

Multiple pathogens were detected among five of the ten children living with HIV ( Table 3 ), representing half of the children in the group whereas multiple detection of pathogens was observed in 14 (25.5%) of the 55 HIV-uninfected children. Ledwaba et al. (2018) reported bacteria-bacteria co-infections among diarrhoiec children of rural communities of Vhembe district in South Africa, an observation also made in the current study. In a study done in Uruguay (Varela et al., 2015), mixed infections were observed only in 2 (3.4%) of the diarrhoea patients. Unsurprisingly in this study, there is preponderance of EAEC in the mixed infections, as it was the most frequently encountered pathogen overall. However, certain combinations were encountered more frequently than would be predicted by chance alone. For example, as shown in Table 2 , EAEC and EIEC were seen in combination in three of the ten HIV-uninfected children, whereas probabilities predict that 0-1 children would be expected to carry both pathogens by chance alone based on overall detection rate. Similarly, the combinations of EAEC and Salmonella spp. as well as EAEC, EIEC and ETEC were each detected in two HIV-uninfected children with expected carriage rates of ¾1 child respectively. The detection of multiple pathogens (including combinations of bacterial and parasitic agents, that is bacterial-parasite co-infections) in some of the subjects in this study underscores the multifactorial nature of diarrhoea as was previously reported by Kotloff et al. (2013) in the Global Enteric Multicentre Study (GEMS). Zhang et al. (2016) also reported enteric pathogens co-infections among diarrhoeic children in China. Chattaway et al. (2013) reported common co-infections of EAEC with other pathogens albeit being more significantly recovered from diarrhoeic cases than healthy controls.

Figure 1 shows a plot of the pathogen combinations observed in this study. The repertoire of pathogens among the children living with HIV and HIV-uninfected children appear similar. However, the proportion of children living with HIV who had either parasites or bacteria detected/recovered from their stools was higher than HIV-uninfected children. The detection of similar repertoire of diarrhoeagenic E. coli pathogens from HIV-infected and HIV-uninfected children has previously been reported by Samie et al. (2007). Even though viruses are part of the leading causes of childhood diarrhea (Nguyen et al., 2007), they were not sought in this study and therefore it is possible that more infections might have been mixed than we are able to report.

Gross or occult (hidden) in faeces can have different origins, ranging from injury to the epithelium and mucosal linings of the gastrointestinal tract by invasive pathogens (Bardhan et al., 2000) to colorectal cancer (CRC) (Borges et al., 2018). Blood in the stool delineates acute watery diarrhoea from dysentery, which is important since the two syndromes are managed differently and understanding which pathogens produce either or both syndromes is important for understanding the pathogenesis of diarrhoeal pathogens in hosts of differing susceptibilities. When blood in stool is not macroscopically visible, that is occult, it has to be detected through tests for blood components in stool, which are generally reliable although they may also pick up heme components that have been ingested with food (Rose et al., 1990; Ransohoff and Lang, 1997; Craven, 2001). For infectious diarrhoea, occult blood is a marker of severe disease caused by invasive pathogens such as Shigella, Salmonella and Entamoeba histolytica sensu stricto. Gassama-Sow et al. (2004) had reported high levels of severe diarrhoea among people living with HIV.

In this study, fecal occult blood detection was not associated with number and nature of pathogen recovered but showed significant association with HIV positive status (p = 0.01) as shown in Table 4 . Bacterial pathogens were recovered from 6 of the children living with HIV, 5 of who had occult blood detected in their stool samples whereas of the 20 HIV-negative children from whom bacterial pathogens were isolated, only 5 had their stool samples positive for occult blood (p = 0.02). Similarly, 3 of the 5 parasite-positive children living with HIV also had their stool samples positive for occult blood compared to none (0) of the 9 parasite-positive children without HIV infection in the control group (p = 0.003). Barring earlier stated confounders such as possibility of colorectal cancer and dietary influences among recruited subjects, our data suggest that invasive pathogens may exert the worst toll on children living with HIV.

A limitation of this study, in addition to the unavoidably small sample size, is that it did not investigate other pathogens such as Campylobacter species and viruses that have been implicated in diarrhoea aetiology (Junaid et al., 2011; Imade and Eghafona, 2015). Even though the study focused on a limited number of bacterial and parasitic etiologies of diarrhoea, multiple pathogens in form of bacteria-bacteria, bacteria-parasite and parasite-parasite co-infections were observed in the specimens. Meanwhile, the overall breadth of pathogens detected in specimens from children living with HIV compares with that of the HIV-uninfected children ( Table 4 ). However, bacteria-parasite co-infection in specimens from children living with HIV when compared to those from children without HIV infection (p = 0.004).