I have read the journal's policy and Vianney Tricou has the following competing interests: He was employed at Takeda pharmaceutical company in singapore, Malaysia one year after the end of the study. This does not alter our adherence to all PLOS policies on sharing data and materials.
Conceived and designed the experiments: PB LC JCG LBLN PS MV. Performed the experiments: SB PB AF LF TF JCG MO LBLN CR RR VT. Analyzed the data: NV LC LT MV. Contributed reagents/materials/analysis tools: SB AF LF TF MO CR RR VT. Wrote the paper: SB NV TGV LT VT PS MV.
In Sub-Saharan Africa, infectious diarrhea is a major cause of morbidity and mortality. A case-control study was conducted to identify the etiology of diarrhea and to describe its main epidemiologic risk factors among hospitalized children under five years old in Bangui, Central African Republic.
All consecutive children under five years old hospitalized for diarrhea in the Pediatric Complex of Bangui for whom a parent’s written consent was provided were included. Controls matched by age, sex and neighborhood of residence of each case were included. For both cases and controls, demographic, socio-economic and anthropometric data were recorded. Stool samples were collected to identify enteropathogens at enrollment. Clinical examination data and blood samples were collected only for cases.
A total of 333 cases and 333 controls was recruited between December 2011 and November 2013. The mean age of cases was 12.9 months, and 56% were male. The mean delay between the onset of first symptoms and hospital admission was 3.7 days. Blood was detected in 5% of stool samples from cases. Cases were significantly more severely or moderately malnourished than controls. One of the sought-for pathogens was identified in 78% and 40% of cases and controls, respectively. Most attributable cases of hospitalized diarrhea were due to rotavirus, with an attributable fraction of 39%. Four other pathogens were associated with hospitalized diarrhea: Shigella/EIEC, Cryptosporidium parvum/hominis, astrovirus and norovirus with attributable fraction of 9%, 10%, 7% and 7% respectively. Giardia intestinalis was found in more controls than cases, with a protective fraction of 6%.
Rotavirus, norovirus, astrovirus, Shigella/EIEC, Cryptosporidium parvum/hominis were found to be positively associated with severe diarrhea: while Giardia intestinalis was found negatively associated. Most attributable episodes of severe diarrhea were associated with rotavirus, highlighting the urgent need to introduce the rotavirus vaccine within the CAR’s Expanded Program on Immunization. The development of new medicines, vaccines and rapid diagnostic tests that can be conducted at the bedside should be high priority for low-resource countries.
Infectious diarrhea is a major cause of illness and death among children under five years from low-income country. In order to identify infectious agents associated with diarrhea, we conducted a case-control study in the Pediatric Complex of Bangui, the sole public pediatric hospital from Central African Republic (CAR). A total of 333 hospitalized children with diarrhea and 333 controls were included, controls being pair matched to the cases according to age, sex and neighborhood. At least one of the sought-for pathogens was identified in 80% of hospitalized children, and approximately one in ten cases presented mixed bacterial/viral co-infections. Five pathogens were positively associated with hospitalized diarrhea, namely rotavirus, norovirus, astrovirus,
In 2013, 6.3 million children under the age of five years died, 578,000 of them from diarrheal diseases. Nearly half of these diarrhea-related deaths were reported from Sub-Saharan Africa [
Nowhere are these problems of childhood diarrhea and shortfalls in its management more evident than in the poorest and most unstable countries of Sub-Saharan Africa. The Central African Republic (CAR) is a resource-limited country in equatorial Africa (ranked 180/187 according to the Human Development Index in 2013). Mortality among under five year-old children was 179/1000 in 2010 [
The Global Enteric Multicenter Study (GEMS) study is a matched case-control study of moderate-to-severe diarrhea in children aged 0–59 months which aimed to estimate the pathogen-specific disease burden in populations from four sites in Africa and three in Asia. The GEMS showed that preventive strategies targeting five pathogens (rotavirus,
Our study, a matched case-control study of diarrhea among hospitalized children under five years was conducted at the Pediatric Complex (PCB) in the CAR’s capital city, Bangui. It was performed in collaboration with the Institut Pasteur de Paris (IPP), the Institut Pasteur de Bangui (IPB) and the PCB.
The study’s primary objective was to identify pathogens associated with diarrhea in hospitalized children under five years of age. Secondary objectives were i) to describe the clinical symptoms of severe diarrhea among hospitalized children, ii) to identify the risk factors associated with severe diarrhea (anthropometric, socio-economic, environmental characteristics), iii) to describe the management of diarrhea before and during hospitalization, and iv) to describe the vital status of children with severe diarrhea during hospitalization and two months after discharge.
Our study was a matched case-control study conducted in Bangui, CAR from December 2011 to November 2013 at the PCB, the country’s sole public pediatric hospital. Cases were children under 60 months of age, hospitalized for diarrhea. Other inclusion criteria were residence in one of Bangui’s eight districts and a general health condition that would support blood and stool sampling. Exclusion criterion was being positive for human immunodeficiency virus (HIV). Controls, identified from the community, were pair matched to the cases according to age (±2 months for infants (0–11 months), ±3 months for toddlers (12–23 months) and ±6 months for children (24–59 months)), sex and neighborhood. To be eligible, controls had to be in good general health, with no history of diarrhea or antibiotic use during the seven days before sampling. HIV status was not systematically tested in controls, but if parents spontaneously declared a child to be seropositive, that child was not included. Cases and controls could not be included more than once.
The research protocol was approved by the Scientist Committee of the Sciences and Health University of Bangui, the CoRC (Clinical Research Committee of Institut Pasteur), the CCTIRS (Comité Consultatif sur le Traitement de l’Information en matière de Recherche dans le domaine de la santé) and the CNIL (Commission Nationale de l’Information et des Libertés) in France. Written informed consent was obtained from all children’s parents or legal guardians for both cases and controls. This study was conducted according to the protocol and ethical principles with their origins in the Declaration of Helsinki. The project provided treatment and laboratory testing free of charge. The CN/CNLS (Coordination Nationale du Comité National de Lutte contre le Sida), with financial support from the Global Fund, covered all costs for HIV treatment.
A pediatrician from the CPB carried out the clinical examination of potential cases and decided on the need for hospitalizing each case for severe diarrhea. Clinical observations, including signs of severe dehydration, hemodynamic shock and extra-digestive symptoms, were collected. At the inclusion of each case, a trained nurse conducted a standardized questionnaire with parents or caregivers so as to gather demographic, environmental and clinical information. Anthropometric measures were also collected: weight, length or height and mid-upper arm circumference (MUAC). Treatments used before and during hospitalization were recorded, as well as the duration of hospitalization and status at discharge. Two months after the child was discharged, parents were contacted by phone so that we could assess the child’s vital status.
At the inclusion of each case, the same trained nurse conducting the standardized questionnaire for all cases was asked to select age- and gender-matched children without diarrhea over the previous seven days, living in the same urban neighborhood as the case. Two nurses were responsible for control enrollments. Parents or primary caretakers underwent the same standardized questionnaire to collect demographic and epidemiological data. The same anthropometric measures were also gathered. Each control provided a fresh stool sample at home in the morning. The sample was returned within 30 minutes of collection to the Institut Pasteur laboratory. Mayors and chiefs of Bangui’s urban districts were informed of the implementation of the study; in turn, they informed the populations living in these districts. Among controls, the vital status at two months was not collected.
Bacterial agents (
Five putative
DNA from faecal samples was extracted with the QIAamp DNA Stool Mini Kit (Qiagen, Courtaboeuf, France). To improve assessment of the involvement of
The stool specimens obtained from enrolled children were later tested for rotavirus, astrovirus, adenovirus and norovirus antigens by using commercial enzyme immunoassays. Suspensions of the stool specimens were prepared according to the manufacturer’s recommendations. For the detection of rotaviruses of the group A, astroviruses and adenoviruses, the ProSpecT Rotavirus, Astrovirus and Adenovirus tests (Oxoid, Thermo Fisher Scientific, Basingstoke, UK) were respectively used according to the manufacturer’s instructions. To detect noroviruses of the genogroups (GG) I and II, the IDEIA Norovirus kit (Oxoid, Thermo Fisher Scientific, Basingstoke, UK) was used according to the manufacturer’s instructions.
A portion of stool was concentrated by the merthiolate iodine formaldehyde concentration technique and examined for helminth eggs and protozoa cysts. Differentiation of pathogenic
A blood sample of 10 mL maximum was collected from each case and sent at room temperature to the IPB within one hour. Blood smears were prepared, stained with 4% Giemsa and analyzed under a light microscope (×100 oil immersion) to detect forms of
To show an odds ratio of 2 characterizing the association between a given pathogen and severe diarrhea, with a pathogen prevalence of 5%, a power of 80% and a two-sided α = 0.05, a sample size of 600 cases and 600 controls was necessary.
The variables collected are defined in
Variables | Definitions |
---|---|
Severe dehydration | One of the following signs: lethargy, unconsciousness or apathy; or two of the following signs: sunken and dry eyes, absence of tears, dry or sticky mouth, inability to drink, abdominal skin pinch with very slow recoil |
Hemodynamic shock | One shock criteria according to age following the references of the francophone group of pediatric resuscitation [ |
Nutritional status |
Normal nutrition: MUAC ≥ 125mm |
Moderate acute malnutrition (MAM): MUAC < 125mm and ≥ 115mm |
|
Severe acute malnutrition (SAM): MUAC < 115 mm |
|
Weight for height Z-score (WHZ) | Calculated according to Myatt et al. [ |
Socio-economic level | Lowest income: no cell phone |
Middle income: cell phone but no car, refrigerator in working condition, or modern sanitation (flushing toilets inside the house) | |
Highest income: car or refrigerator in working condition or modern sanitation (flushing toilets inside the house) | |
Seasons | Rainy season: from May to October |
Dry season: from November to April | |
Cohabiting parents | The child’s mother and father living in the same house |
Improved water | Water from fountains or running water |
a The mid-upper arm circumference (MUAC) was chosen instead of the usual weight-for-height ratio for age because children with severe diarrhea can lose more than 10% of their body weight. In addition, the MUAC is less prone to errors, and thus a better indicator of mortality risk associated with malnutrition than weight-for-height [
b According to WHO standards [
Comparison between cases and controls were made by univariate conditional logistic regression to take into account the matching of cases and controls. Pathogens potentially associated with severe diarrheas in univariate analysis with a p-value<0.25 were included in a backward conditional logistic regression, adjusted on the presence of other pathogens. Results are reported as adjusted OR (aOR) with 95% CI. The attributable fraction (AF) was calculated for pathogens with significant aOR with the following formula: aAF = P (pathogens) among cases* (aOR-1)/aOR, that is, the proportion of severe diarrhea attributable to this specific pathogen. When the association was significantly negative with an aOR<1, the protective fraction was calculated with the following formula: aPF = (1- aOR) x p (events) among controls, that is, the proportion of severe diarrhea avoided by the presence of this specific pathogen.
For cases, associations between socio-economic, anthropometric and clinical data with the use of ORS and antibiotics before and during hospitalization, and association with vital status were determined by univariate analysis. All variables associated with a p value<0.25 were included in a backward logistic regression. The final model includes only variables with a p-value<0.05. Interactions were tested and the goodness-of-fit of the model was studied using the Hosmer-Lemeshow statistic.
General characteristics of cases are described in
a Severe acute malnutrition: MUAC<115mm
b Moderate acute malnutrition: 115≥MUAC≤125mm
c Weight for length Z-score
The mean time since onset of diarrhea before cases were presented at the hospital was 3.7 (±1.8) days. Cases presented the following clinical symptoms: severe dehydration (N = 239, 72%); hemodynamic shock (N = 216, 65%); serious neurological injury (N = 239, 72%); fever (body temperature ≥38°) (N = 264, 79%); vomiting (N = 267, 80%); and extra digestive signs (N = 94, 28%), including 69 cases of upper respiratory tract infection, 5 cases of pulmonary signs, 8 cases of digestive signs and 12 cases of other signs. 26 cases (8%) were positive for malaria, all of them due to
Demographic characteristics of cases and controls are summarized in
The results are summarized in
Cases n(%) | Controls n(%) | aOR |
aAF |
||
---|---|---|---|---|---|
Infants | n = 195 | n = 195 | |||
Rotavirus | 90 (46.1) | 8 (4.1) | 44.8 (10.6–189.5) | 45.0 (41.6–45.7) | |
Astrovirus | 25 (12.8) | 8 (4.1) | |||
Norovirus | 18 (9.2) | 9 (4.6) | 3.2 (1.1–9.8) | 6.3 (0.8–8.1) | |
Adenovirus | 11 (5.6) | 9 (4.6) | |||
17 (8.7) | 14 (7.2) | ||||
2 (1.0) | 1 (0.5) | ||||
1 (0.5) | 4 (2.1) | ||||
3 (1.5) | 5 (2.6) | ||||
Diarrhoeagenic |
24 (12.3) | 17 (8.7) | |||
0 (0) |
1 (0.5) |
||||
0 (0) | 6 (3.1) | ||||
32 (16.4) | 7 (3.6) | 4.6 (1.8–11.7) | 12.5 (7.1–14.6) | ||
Toddlers | n = 103 | n = 103 | |||
Rotavirus | 43 (41.7) | 2 (1.9) | |||
Astrovirus | 7 (6.8) | 7 (6.8) | |||
Norovirus | 11 (10.7) | 5 (4.8) | 2.6 (0.8–8.5) | ||
Adenovirus | 6 (5.8) | 6 (5.8) | |||
23 (22.3) | 14 (13.6) | ||||
7 (6.8) | 1 (1.0) | ||||
8 (7.7) | 3 (2.9) | ||||
2 (1.9) | 3 (2.9) | ||||
Diarrhoeagenic |
10 (9.7) | 13 (12.6) | |||
1 (1.0) |
3 (2.9) |
||||
2 (1.9) | 12 (11.6) | 0.16 (0.03–0.7) | 9.7 (3.5–11.3) | ||
7 (6.8) | 2 (1.9) | 2.9 (0.6–14.8) | |||
Children | n = 35 | n = 35 | |||
Rotavirus | 1 (2.8) | 1 (2.8) | |||
Astrovirus | 1 (2.8) | 0 (0) | |||
Norovirus | 4 (11.4) | 0 (0) | |||
Adenovirus | 2 (5.7) | 3 (8.6) | |||
10 (28.6) | 2 (5.7) | 4.1 (0.9–19.4) | |||
2 (5.7) | 2 (5.7) | ||||
4 (11.4) | 1 (2.9) | 2.3 (0.2–23.4) | |||
0 (0) | 1 (2.9) | ||||
Diarrhoeagenic |
3 (8.6) | 5 (14.3) | |||
4 (11.4) |
2 (5.7) |
||||
1 (2.8) | 8 (22.9) | 0.17 (0.02–1.4) | |||
3 (8.6) | 0 (0) | ||||
Total | n = 333 | n = 333 | |||
Rotavirus | 134 (40.4) | 11 (3.3) | 51.8 (15.6–172.6) | 39.2 (37.4–39.8) | |
Astrovirus | 33 (9.9) | 15 (4.5) | 3.4 (1.05–10.8) | 7.0 (0.5–8.9) | |
Norovirus | 33 (9.9) | 13 (3.9) | 3.0 (1.3–7.2) | 6.6 (2.3–8.5) | |
Adenovirus | 19 (5.7) | 18 (5.4) | |||
50 (15.0) | 30 (9.0) | 2.4 (1.3–4.5) | 8.7 (3.5–7.7) | ||
11 (3.3) | 3 (0.9) | 3.4 (0.7–17.6) | |||
13 (3.9) | 8 (2.4) | ||||
5 (1.5) | 9 (2.7) | ||||
Diarrhoeagenic |
37 (11.1) | 35 (10.5) | |||
5 (1.5) |
6 (1.8) |
||||
3 (0.9) | 26 (7.8) | 0.2 (0.05–0.7) | 6.2 (2.3–7.4) | ||
42 (12.6) | 9 (2.7) | 6.0 (2.4–14.6) | 10.5 (7.3–11.7) |
a Adjusted on the presence of others pathogens
b Determined by ipaH PCR
c Determined by culture
d Trophozoite
e Cystic form
Other pathotypes (EPEC, ATEC, EIEC and STEC) were found in less than 3% of the children.
In multivariate analyses, when adjusted on the presence of other pathogens, five pathogens were positively associated with diarrhea: rotavirus, norovirus, astrovirus,
In cases, the prevalence of pathogens varied according to age categories (
Viruses were more frequently identified during the dry season, with 65% (115/179) of cases compared to 43% (66/154) in rainy season, p<0.001. In contrast,
Before hospitalization, ORS and zinc supplementation were prescribed to 38% and 0.9% children, respectively. Antiparasitic treatments and antibiotics were administered to 44% and 34% of children, respectively. One on four children (25%) received traditional treatments that consisted mainly in infusions or herbal decoctions (72%), herbal enemas (16%) or fruit porridge (9%). During hospitalization, 99% of children received ORS, 87% intravenous rehydration, 70% antibiotics, 66% antiparasitic treatments, 55% zinc, and none received traditional treatments (
The mean duration of hospitalization was 4.8 (±2.8) days. Four percent of children (12/333) died during hospitalization and 1% (3/271; 50 missing values) died during the two months after discharge. Among the 12 deaths, 5 were infants (< 11 months), 5 were toddlers (12 to 23 months), and 2 were children (24–59 months). Three cases were positive for
This study is the first case-control study conducted in CAR that provides the etiology and clinical outcome of children of less than five years that were hospitalized for diarrhea in Bangui. A large prospective study, the GEMS study, used a similar approach with matched case-control in children between 0 to 59 months suffering from moderate-to-severe diarrhea in seven countries, including four in Africa: The Gambia and Mali (West Africa), Mozambique (South Africa) and Kenya (East Africa) [
HIV positive children were excluded from the present study. Indeed, HIV (co-) infection increases the risk of severe diarrhea by impairing the immune system, making problematic the interpretation of results in the context of case/control comparison. Moreover, it was not possible to test the community controls for HIV without returning to provide individual announcement and counseling, which was not possible in our study. Even if the HIV status of the controls was unknown, we supposed that it was low because only 22 hospitalized cases were tested positive for HIV and the children selected as controls were in good health. Among the 406 HIV-negative cases, only 333 were included in our analysis because they were well-matched with the controls as shown in the flow chart. The main criterion for impaired matching was age. The majority of these cases were borderline, with only few days or weeks of difference between cases and controls. The addition of the wrongly matched children for age and sex did not modify the results. In the current analysis, we prefer to maintain the criteria that were defined in the protocol.
At least one of the sought-for pathogens was identified in around 80% of children hospitalized for diarrhea, and approximately one in ten cases presented mixed bacterial/viral co-infections, consistent with findings in other African countries [
In case-control studies, diarrhoeagenic
As reported in previous studies [
Our findings also shed light on the management of severe childhood diarrhea. Although the WHO recommends the use of antibiotics in children with bloody diarrhea (5% of cases in our study), suspected cholera, or associated sepsis, we found that 40% of children before their arrival and 70% during hospitalization received an antibiotic treatment. Combined with the widespread use of antiparasitic treatments before consultation (44%), this finding could partially account for the relatively low number of bacteria or parasites found among cases. However, the bias was minimized for
Furthermore, the uncontrolled consumption of antimicrobial agents is cause for concern in countries like the CAR with inadequate healthcare systems, because it favors the spread of antimicrobial resistance [
Despite guidelines that recommend the use of ORS and zinc supplementation for all children, only 40% of children received ORS and less than 1% took zinc before their hospitalization. Although the pre-hospitalization ORS finding is higher than that found in other studies (ORS before hospitalization in 16% of cases according to UNICEF in CAR between 2008 and 2012, 20% in Senegal), it is essential to continue education of mothers on the importance of rehydration and zinc in home management of diarrhea.
The WHO also recommends exclusive breastfeeding for the first six months of life. Only 16% of children under six months old were exclusively breastfed, whereas others were exposed to putative pathogens in weaning foods or inadequate diversity of complementary foods.
Our study has limitations. We could not complete the planned inclusion of 600 cases and 600 controls, mainly due to security problems in Bangui. Only 2 or 3 children were recruited per day during the study-period. This low recruitment can be explained by the cost associated with transport to the hospital, hospitalization, treatment and care. As previously described [
The choice of immunological methods to detect rotavirus, astrovirus, adenovirus and norovirus is questionable. Indeed, some more sensitive molecular methods have been used in other epidemiological studies on diarrhea [
The data reported here are particularly important, given the significance of childhood diarrhea in countries with inadequate healthcare systems and long-term instability, as well as the lack of high-quality data and the difficulty of carrying out such studies in these contexts. Rotavirus, norovirus, astrovirus,
(DOCX)
We would like to express our most sincere thanks to the children and parents participating in this study. We also extend our thanks to all the staff from the Complexe Pédiatrique de Bangui and the Institut Pasteur in Bangui and its director Dr Mirdad Kazanji who made this work possible despite the difficult conditions, particularly Ernest José Gbao and Natacha Zato. We would also like to acknowledge Kathleen Victoir from the Institut Pasteur of Paris for her support as administrative program manager.