Nontyphoidal Salmonella among Children under 5 Years Old in Sub-Saharan Africa and South Asia in the Global Enteric Multicenter Study

ABSTRACT. Factors associated with nontyphoidal Salmonella (NTS) infection have not been well characterized to date. We aimed to compare the associated factors among children under age 5 years with NTS infection in sub-Saharan Africa and South Asia. Data from children having moderate-to-severe diarrhea (MSD) and asymptomatic children with NTS isolated from fecal specimens were extracted from the Global Enteric Multicenter Study (GEMS), conducted from December 2007 to March 2011. Compared with NTS-negative children, NTS-associated MSD cases in South Asia were associated with the presence of goat in the house (adjusted odds ratio [aOR]: 2.15; 95% confidence interval [CI]: 1.25–3.70) and handwashing after handling an animal (aOR: 2.26; 95% CI: 1.36–3.74). In sub-Saharan Africa, children with NTS associated MSD had a greater association with stunting (1.21 95% CI: 1.01–1.45), longer duration of diarrhea (aOR: 1.25 95% CI: 1.19–1.31); presence of cow in house (aOR: 1.54 95% CI: 1.09–2.16), handwashing after handling animal (aOR: 2.41 95% CI: 1.74–3.33). Drinking tube well water (aOR: 0.54 95% CI: 0.32–0.91), availability of toilet facility (aOR: 0.58 95% CI: 0.53–0.65), and handwashing before eating (aOR: 0.76 95% CI: 0.57–1.00) and after defecation (aOR: 0.80 95% CI: 0.69, 0.94) were found to be protective. The differentials between children of both regions having fecal NTS are distinct and underscore the need for policymaking for preventive and control strategies targeting stunted children.


INTRODUCTION
Nontyphoidal Salmonella (NTS) is a Gram-negative bacterium responsible for causing disease in both humans and animals worldwide.3][4] The majority of illnesses in humans caused by NTS are related to gastrointestinal problems, although uncommonly, it also invades the bloodstream.Infants and young children are more susceptible to NTS infection, making them a high-risk population. 5Approximately 5% of the NTS infections occur due to extraintestinal and invasive NTS (iNTS).iNTS found in the bloodstream of children is emerging with new pathogenic features. 6here are existing knowledge gaps regarding the epidemiology and prevalence of NTS, including clinical presentation, risk factors, and antimicrobial-resistance trends in sub-Saharan Africa and South Asia. 7The only study regarding the estimation of the NTS burden in South Asia indicated an incidence of 470 cases per 100,000 person-years. 8NTS infection has been given less attention compared with that of Salmonella enterica serovar Typhi because of the high incidence of S. Typhi in this region.A good understanding of the risk factors for NTS may therefore prevent diarrhea leading to serious adverse outcomes in young children.
In this study, we aimed to compare and differentiate between the associated factors of NTS infection among children of two distinctly different geographic regions, sub-Saharan Africa and South Asia.We also intended to describe the associations among their demographic, environmental, and socioeconomic characteristics, and clinical features.

METHODS
Study site.][11] Study design and study participants.GEMS had a welldefined standardized recruitment protocol. 10The published, 12 working hypothesis, 9 epidemiology, 10 clinical, 13 laboratory, 10 and statistical methods 14 of GEMS have been described elsewhere. 15For our present analysis, we enrolled NTS-positive children from both case and control group of GEMS (N 5 22,567) because NTS was present in stool among the moderate-to-severe diarrhea (MSD) cases (N 5 9,439) and healthy controls (N 5 13,128) of the GEMS.Children having NTS in stool constituted the cases (N 5 378), and children without NTS in stool were considered as the controls (N 5 1134).The control children from each study site were randomly selected from the database of enrolled GEMS by computer-based randomization technique using SPSS version 25.0 (SPSS Inc., Chicago, IL) to prevent any selection bias.To substantiate the statistical power of our analyses, we used a site-specific case-control ratio of 1:3.In this data analysis, we include 1,512 (6.7%) children under 5 years of age from GEMS.Of the 378 (1.68%) NTS-positive cases, 190 were MSD cases with 105 (55.26%) in sub-Saharan Africa and 85 (44.74%) in South Asia.The remaining 188 were NTSpositive, asymptomatic healthy children with 92 (48.94%) from sub-Saharan Africa and 96 (51.06%) from South Asia; 1,134 children were NTS-negative controls (Figure 1).
Stool sample collection and fecal microbiology.Stool specimens were collected from every child at the time of enrollment.Stool samples were processed using the GEMS laboratory procedure protocol. 16,17Bacterial pathogens, viruses, and protozoa were identified using standard laboratory methods. 16TS has been isolated using standard bacteriological methods. 16In our analysis, we did not use the TaqMan molecular diagnostic data that included information on isolates of NTS because TaqMan could not differentiate typhoidal and NTS. 18ariable of interest.Region is categorized as sub-Saharan Africa and South Asia.Clinical manifestations such as vomiting (three or more times/day), fever on admission (temperature measured at least 38 C), and visible blood in stools was assessed based on reports of the primary caretakers. 10iarrhea was defined as the passage of three abnormally loose or watery stools in the previous 24 hours. 10,19Each child was assessed for diarrhea for eligibility of enrollment, and additional criteria for inclusion involved fulfilling at least one of the following conditions for MSD: sunken eyes, loss of skin turgor or very slow, intravenous rehydration administered or prescribed, or admission to hospital with diarrhea or dysentery. 10Height, weight, and mid-upper-arm circumference (MUAC) were measured at enrollment and after rehydration for each child. 20Details of anthropometry have been described elsewhere. 102][23] The interpretation of malnutrition following a WHO guideline was stunted, underweight, and wasted as having HAZ # 2, WAZ # 2, and WHZ # 2. Child breastfeeding status was categorized as breastfed and nonbreastfed.
Sociodemographic information regarding the enrolled children and their household characteristics included mother as a primary caretaker, mother's education, household size, and building materials of the floor as explanatory variables.Further variables addressed included handwashing practices, use of handwashing material, access to the main source of drinking water, available water treatment method, available toilet facility, and animals on the premises.The households were classified into socioeconomic status quintiles based on the wealth quintile index (poor, lower middle, middle, upper middle, and rich).This was done to assess possible factors associated with disease and as indicators for creating a wealth index for each site. 10,24,25The stool was examined for consistency and the presence of pus, mucus, red blood cells, and other copathogens.
Statistical analysis.The statistical analysis was carried out using version 15.0 of the STATA software (Stata Corp., College Station, TX).Categorical variables were expressed in terms of frequency and percentage, and mean and standard deviation (SD) were used to represent continuous variables.To compare the mean differences, a Student's t-test was conducted for continuous variables, and variations in proportions were compared with the chi-square test.The strength of association was determined by estimating odds ratios (ORs) and 95% confidence intervals (CIs) using simple logistic regression.Initially, we performed bivariate analyses of the relevant characteristics to identify factors that were significantly associated with NTS.Eventually, we performed multiple logistic regression analyses to identify the independently associated factors of NTS infection in children aged , 5 years in both regions (sub-Saharan Africa and South Asia), controlling for the site (country) as a cluster for NTS-positive children having MSD and healthy asymptomatic children.Covariates with a P , 0.25 in the bivariate analysis were included in the multiple regression model via a stepwise forward selection technique, as more traditional levels such as P , 0.05 can fail in identifying variables known to be of importance, 26 whereas other relevant variables, such as age and sex, were adjusted for the variable with any P value due to their biological and public health importance.All relevant covariates were included in the subsequent model to obtain an adjusted final model.To detect multicollinearity, the variance inflation factor (VIF) was determined, and none of the variables provided a VIF value . 5. Lastly, a P value , 0.05 was considered statistically significant in multivariable analysis.
Ethical consideration.The study was approved by the ethical committees at the University of Maryland School of Medicine and each of the participating institutes across each of the seven study sites. 10Signed informed consent for the enrollment of each child in the study was collected from the parents or legal guardians (for both sick cases and healthy controls).

RESULTS
Country-specific baseline characteristics of the NTSpositive under 5 children in South Asia and sub-Saharan Africa.The baseline characteristics of the study population having NTS-positive stools are presented in the Supplemental Table 1a.We collected data from 1,512 children enrolled in GEMS.Among them, 41.4% were breastfed; 50.3% of children had MSD, and 14.6% reported the presence of blood in the stool.Approximately 97% of the primary caretakers were mothers, and 70% were literate.Only 23% of children belonged to a middle-class family.Sixty-five percent had fowl/rodent and 44% had a cow in the house, and 50% of families indicated the presence of a dog or cat in the house.One-fourth of the households used tube well water as the main source of drinking water.A toilet facility was available in 87% of the houses.Almost 79% used soap and water during handwashing.Handwashing practice was observed among only 13% after handling an animal.
Characteristic of NTS-positive children with MSD in South Asia and sub-Saharan Africa.Half (50.6%) of the NTS-positive children suffering from MSD were aged between 0 and 11 months, and overall 42.7% were female in South Asia.Less often, they had sunken eyes and loss of skin turgor as indicators of MSD.More children lived in households with a natural floor made of earth, sand, and dung; had a domestic animal in the household; used tube well water as the main source of drinking water; and practiced handwashing more often before eating and cooking, after defecation, and after handling an animal.However, among caregivers and family members, there was less frequent handwashing practice after cleaning the child and less commonly use water and soap for handwashing.Laboratory examinations reported loose, watery stool with the presence of mucus with Giardia as a copathogen (Supplemental Table 1b) less frequently compared with NTS-negative children.Children belonging to both cases and controls were breastfed and used a toilet facility, and the primary caretaker was mostly the mother.
In sub-Saharan Africa, a significantly higher proportion of under 5 children with MSD and NTS-positive stool were stunted and underweight.There were higher durations of diarrhea and more children had the clinical indicators for MSD, but dysentery was less common.The mother was less often the primary caretaker; floors were made of earth, sand, and dung; households had a domestic animal, used non-tube well water as the main source of drinking water, often used treated drinking water.Families less frequently had a toilet facility at the house and were less likely to practice handwashing before eating and after defecation; however, they practiced handwashing more often before nursing the child, after cleaning the child, and after handling animal.On examination stool was loose and watery with fecal mucus compared with stool specimens of the control group.Campylobacter and Giardia were reported as the more frequent copathogens (Supplemental Table 1b) in the children of the control group compared with children from the NTS-positive group.
Characteristics of asymptomatic NTS-positive healthy children in South Asia and sub-Saharan Africa.Compared with asymptomatic NTS-negative children in South Asia, asymptomatic NTS-positive children were more often aged between 0 and 11 months, were more frequently underweight but less often stunted.These aforementioned asymptomatic NTS-positive children had more frequent episodes of fever; were less likely to live in households with floors made of earth, sand, and dung or have a cow at the house; more frequently reported that there was no treatment of drinking water.They also practiced handwashing after defecation more frequently, but were less likely to do so before eating or after handling an animal.Campylobacter was detected as a copathogen (Supplemental Table 1b) more often.
In sub-Saharan Africa among NTS-positive cases, more than half of the children were aged 0 to 11 months; they were more likely to be breastfed and malnourished; use tube well water as the main source of drinking water, and have a toilet facility for fecal disposal.Caregivers were less likely to practice handwashing before eating and or nursing a child, but more frequently did so before cooking.Stool samples collected from these children were less commonly mixed with mucus, and Campylobacter and Giardia were more often isolated among the NTS-negative control group (Supplemental Table 1b).

DISCUSSION
In our study, approximately half of the NTS cases were children aged between aged 0 and 11 months in both distinct geographic regions.Our findings support previous studies indicating that young children, particularly infants, are more susceptible to NTS infection. 27,28ompared with South Asia, stunting was significantly associated with both symptomatic and asymptomatic NTS infections among children from sub-Saharan Africa.The previous study from Africa reported malnutrition to be a risk factor for NTS infection. 29We assume that many factors may be attributable to the discrepancy in findings between sub-Saharan Africa and South Asia, including lower levels of the immunocompromised children in South Asia, owing to fewer cases with advanced HIV infection and a lower burden of various comorbidities, including malaria.Studies have indicated that NTS infection is more often invasive in immunocompromised patients compared with healthy patients.The role of the interleukin-17 (IL-17) in NTS infection has been suggested to increase vulnerability to invasive NTS infection; this is being documented by human studies with IL-12p40 (inherently agnostic cytokine) deficiency, 30 suggesting that IL-12 can play a critical role in protecting against NTS.There is reduced IL-12 expression in malnourished children compared with well-nourished controls, 31 which may make them more vulnerable to NTS infection.
In sub-Saharan Africa, the increased duration of diarrhea is a commonly accompanying clinical feature among NTS-positive children.In South Asia, we did not find any association between NTS-positive MSD and the duration of diarrhea.Case-control studies from the United States, 32 Palestine, 33 Denmark, 34 and Brazil 35 have reported an association with acute diarrhea and NTS, whereas no such association was found in Thailand 36 or Bangladesh. 37The duration of diarrhea was not associated with NTS in South Asia, probably due to the low prevalence of HIV. 38n sub-Saharan Africa, caregivers had fewer complaints about visible blood in the stool.No association was found with NTS-positive MSD and visible blood in stool in South Asia.Such observations are concomitant with the findings of other studies. 39,40Similarly, dysentery as a criterion suggesting MSD was significantly related to NTS infection in sub-Saharan Africa but not in South Asia.2][43] The low prevalence of HIV and malaria in South Asia is a likely explanation for the absence of any significant association between dysentery and NTS infection.Since 1966, NTS has been reported as a cause of bacteremia in 33 of 54 African countries, spanning the five geographic areas  of Africa, but mainly in sub-Saharan Africa, according to a systematic review. 42However, data suggest that iNTS is a significantly less serious concern in Asia than typhoid fever.The high frequency of falciparum malaria in Africa could be one explanation for this discrepancy between continents. 44n our research, we observed that NTS isolation is associated with an increased need for IV rehydration, and there is evidence of increased inflammation of the gut, including higher levels of stool mucus in sub-Saharan Africa.These observations are suggestive of the pathogenicity of NTS causing secretory diarrhea, resulting in a fluid-electrolyte deficit in the study children; these findings are consistent with a study conducted in Bangladesh. 45In sub-Saharan Africa, NTS-positive children had a frequent loss of skin turgor, which was indicative of dehydration and the need for oral rehydration therapy instead of intravenous saline infusions.
We also report that children with MSD and associated NTS infection were more likely to drink tube well water in sub-Saharan Africa.In South Asia, NTS was not linked with treatment of drinking water at the home.Previous findings from another study in Asia indicated that the type of water was not a risk factor for NTS infection. 45A research in Bangladesh found that patients with S. typhi bacteremia have a greater likelihood of unboiled water intake than those with NTS bacteremia. 46This does not rule out the remote possibility of using untreated water by NTS-positive children.
In Africa, mothers' handwashing before food preparation, after defecation, and before eating were correlated with less diarrhea. 47Different studies have shown that handwashing before eating has a positive impact on preventing shigellosis 48 and improved water, sanitation, and hygiene reduced childhood diarrhea. 49In our study, we also found primary caretakers' handwashing before eating and after defecation had a protective effect on MSD children with NTS infection in sub-Saharan Africa.There was no association between handwashing practices before eating and NTS in South Asia, perhaps because handwashing practice before eating was much lower than after defecation. 50In this region, handwashing practices with soap and water are poor, but most women from slums and rural areas rub their hands on the ground or use soil and rinse them with water after defecation during handwashing. 51Practicing handwashing after handling animals was significantly associated with NTSassociated MSD in children in both regions, probably for the same reason as the mothers washing hands only with water but not soap after handling animals.In our study, we found no association between improved sanitation and NTS in South Asia but had a protective effect in sub-Saharan Africa.In South Asia, gross diversity in access to improved sanitation progress is prevalent. 52n sub-Saharan Africa, NTS-positive cases were significantly associated with the presence of a cow or goat in South Asia, but the nature of the role of animals in the transmission of NTS is not clear.However, research at the GEMS site in Kenya found that a moderately high proportion of fecal chicken pools were positive for NTS. 53Among the asymptomatic NTS-positive children, presence of a cow was found to be protective in South Asia.Such evidence make the interaction between NTS and domestic animals at home an interesting topic for future studies.
NTS cases often had stools with mucus in sub-Saharan Africa, but among asymptomatic children, mucus was found less often in the stool.There have been reports of elevated intestinal inflammation, including higher levels of mucus and red blood cells.These data may suggest invasive pathogenicity of NTS in children under 5 years of age, which is consistent with other studies. 45e found that Giardia was less likely to be observed as a copathogen among NTS-positive children in sub-Saharan Africa and South Asia.Conversely, in South Asia Campylobacter was significantly associated with NTS-positive cases.In other studies, the prevalence and concentration of Campylobacter in household pets was lower than that of farm animals that did not come in contact with the public. 54Campylobacter was highly associated with pets. 55A hospitalbased surveillance study reported a lack of association of Campylobacter and Giardia as copathogens with NTS cases in Bangladesh. 45Additionally, we did not find any association of NTS infection with the primary caretaker's education, socioeconomic context, handwashing with soap, and overcrowding at the household level.
The strengths of our study involved unbiased sampling, a large sample size, and high-quality standard laboratory performance.The study looked for the factors associated with both symptomatic and asymptomatic NTS infection among children under 5 years from both regions, which further enriches the findings of the study.
Our study also has some limitations.We could not determine the effect of HIV and factors related to HIV infection, and other associated factors such as HIV data in children were absent for South Asia.We assume that a stronger predisposing factor for NTS could be immunocompromised status (e.g., HIV infection) of children with severe malnutrition, which we could not explore in our analysis.CONCLUSION Distinct differences in sociodemographic and clinical characteristics among children with NTS between South Asia and sub-Saharan Africa were observed.Further studies are required to explore diversity within and among regions and emphasize the importance of policy-making for the prevention of nutritional disorders including stunting among children under 5 years in sub-Saharan Africa and South Asia.
Acknowledgments: We acknowledge with gratitude the commitment of the Global Enteric Multicenter Study (GEMS) in collaboration with the Center for Vaccine Development, University of Maryland School of Medicine, Baltimore, MD, USA.GEMS received a grant from the Bill & Melinda Gates Foundation.We are grateful to the GEMS staff and the parents and children for their contributions.This research protocol was funded by core donors who provide unrestricted support to International Center for Diarrheal Disease Research, Bangladesh (icddr,b) for its operations and research.Current donors providing unrestricted support include the Governments of Bangladesh, Canada, Sweden, and the United Kingdom.We thank our core donors for their support and commitment to icddr,b's research efforts.A publicly available GEMS dataset was analyzed in this study.These data can be obtained from ClinEpiDB (https://clinepidb.org/ce/app/record/dataset/DS_841a9f5259).Authors' addresses: Rina Das, Md Ahshanul Haque, Mohammod Jobayer Chisti, Tahmeed Ahmed, and Abu Syed Golam Faruque, Nutrition and Clinical Services Division, International Center for Diarrheal Disease Research, Bangladesh (icddr,b), Dhaka, Bangladesh.Emails: rina.das@icddrb.org,ahshanul.haque@icddrb.org,chisti@ icddrb.org,tahmeed@icddrb.org,and gfaruque@icddrb.org.
This is an open-access article distributed under the terms of the Creative Commons Attribution (CC-BY) License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Financial
support: This work was supported, in whole or in part, by the Bill & Melinda Gates Foundation (grant no.INV-002050).Under the grant conditions of the Foundation, a Creative Commons Attribution 4.0 Generic License has already been assigned to the Author Accepted Manuscript version that might arise from this submission.

Total children enrolled in GEMS N= 22,567 GEMS Controls (without diarrhoea/ healthy children) N= 13,128 GEMS Cases (with moderate to severe diarrhoea) MSD N= 9,439 NTS (-) healthy children n= 12,940 NTS (+) healthy children n= 188 NTS (-) MSD children n= 9,249 NTS (+) MSD children n= 190 NTS (-) healthy children from Sub-Saharan Africa n= 354 NTS (-) healthy children from South Asia n= 319 NTS (+) healthy Children from South Asia n= 96 NTS (+) healthy children from Sub-Saharan Africa n= 92 NTS (-) MSD children from Sub-Saharan Africa n= 237 NTS (-) MSD children from South Asia n= 224 NTS (+) MSD children from South Asia n= 85 NTS (+) MSD children from Sub-Saharan Africa n= 105 Inclusion criteria for cases in GEMS study • Child is 0-59 months of age • Child belongs to the demographic surveillance system population at the site • Child is not currently enrolled as a case • Child meets case definition of diarrhea (≥3 abnormally loose stools in the previous 24 h • Diarrhea episode is: -Acute (onset within 7 d of study enrollment) and -Represents a new episode (onset after ≥7 diarrhea-free days) • Diarrhea is moderate-to-severe, meaning that the child met at least 1 of the following criteria: -Sunken eyes, confirmed by parent/primary caretaker as more than normal -Loss of skin turgor (determined by abdominal skin pinch (slow return [≤2 s] or very slow return [>2 s]) -Intravenous rehydration administered or prescribed -Dysentery (visible blood in a loose stool) -Hospitalized with diarrhea or dysentery Inclusion Criteria for Controls in GEMS study • Resides in demographic surveillance system area • Matched to the index case as follows: Age: ±2 months for cases 0-11 months ±4 months for cases 12-59 months May not exceed the stratum boundaries of the case, eg, a control for an 11-months-old case must be between the ages of 9 and 11 months and a control for a 13-months-old must be between the ages of 12 and 17 months -Same sex -Residence: lives in the same or nearby village/neighborhood as the case -Time: enrolled within 14 days of presentation of the case • No diarrhea in the previous 7 days NTS (-) control NTS (+) Case NTS (+) Case NTS (-) control FIGURE
1. Study profile of enrolled children.GEMS 5 Global Enteric Multicenter Study; MSD 5 moderate-to-severe diarrhea; NTS 5 nontyphoidal Salmonella.

TABLE 1
Factors associated with MSD among children having NTS present in stool aOR 5 adjusted odds ratio; CI 5 confidence interval; MSD 5 moderate-to-severe diarrhea; NTS 5 nontyphoidal Salmonella; OR 5 odds ratio.* P values were calculated using multiple logistic regression and all variables are adjusted for age, gender, wealth index, presence of copathogen, and site.Stunting: height/length-for-age z-score # 2, (%; for , 5 years of age).

TABLE 2
Results of multiple logistic regression in exploring the factors of NTS infection in MSD children aged , 5 years in sub-Saharan Africa and South Asia aOR 5 adjusted odds ratio; CI 5 confidence interval; MSD 5 moderate-to-severe diarrhea; NTS 5 nontyphoidal Salmonella; OR 5 odds ratio.* P values were calculated using multiple logistic regression; adjusted for age, gender, and site as a cluster, variables with P value , 0.05 were considered for inclusion in the final model.

TABLE 3
Results of multiple logistic regression in exploring the factors of asymptomatic NTS-positive children aged , 5 years in sub-Saharan Africa and South Asia 5 adjusted odds ratio; CI 5 confidence interval; MSD 5 moderate-to-severe diarrhea; NTS 5 nontyphoidal Salmonella; OR 5 odds ratio.* P values were calculated using multiple logistic regression; adjusted for age, gender, wealth index, presence of copathogen and site as cluster; variables with P value , 0.05 were considered for inclusion in the final model.Stunting: height/length-for-age z-score # 2 (%; for , 5 years of age). aOR

TABLE 4
Results of multiple logistic regression in exploring the factors of NTS infection in children aged , 5 years in both sub-Saharan Africa and South Asia (both MSD and asymptomatic NTS) P values were calculated using multiple logistic regression and all variables are adjusted for age, gender, wealth index, presence of co pathogen and site.Stunting: height/length-for-age z-score # 2 (%; for , 5 years of age).