Factors associated with anaemia among preschool- age children in underprivileged neighbourhoods in Antananarivo, Madagascar

Background: Anaemia occurs in children when the haemoglobin level in the blood is less than the normal (11 g/dL), the consequence is the decrease of oxygen quantity in the tissues. It is a prevalent public health problem in many low-income countries, including Madagascar, and data on risk factors are lacking. We used existing data collected within the pathophysiology of environmental enteric dysfunction (EED) in Madagascar and the Central African Republic project (AFRIBIOTA project) conducted in underprivileged neighbourhoods of Antananarivo to investigate the factors associated with anaemia in children 24 to 59 months of age. Methods: Children included in the AFRIBIOTA project in Antananarivo for whom data on haemoglobin and ferritin concentrations were available were included in the study. Logistic regression modelling was performed to identify factors associated with anaemia. Results: Of the 414 children included in this data analysis, 24.4% were found to suffer from anaemia. We found that older children (adjusted OR: 0.95; 95% CI: 0.93–0.98) were less likely to have anaemia. Those with iron deficiency (adjusted OR: 6.1; 95% CI: 3.4–11.1) and those with a high level of faecal calprotectin (adjusted OR: 2.5; 95% CI: 1.4–4.4) were more likely to have anaemia than controls. Conclusions: To reduce anaemia in the children in this underprivileged area, more emphasis should be given to national strategies that improve children’s dietary quality and micronutrient intake. Furthermore, existing measures should be broadened to include measures to reduce infectious disease burden. This study conducts a secondary analysis of data collected as part of the AFRIBIOTA project, a translational study of the pathophysiology of EED performed in the two African cities of Antananarivo (Madagascar) and Bangui (Central African Republic). Details of the project objectives and methodology of the AFRIBIOTA project are provided elsewhere [14]. AFRIBIOTA is a case–control study of stunting in which 260 stunted children and 200 age- and sex-matched nonstunted children were recruited in each country. Data collection for the AFRIBIOTA project was conducted from November 2016 to March 2018. Children from 24 to 59 months of age with no obvious signs of severe disease and with negative HIV serology were recruited. The recruitment was mainly community-based and was conducted in underprivileged areas of the Urban Commune of Antananarivo (Andranomanalina Isotry, Ankasina and their surrounding neighbourhoods) and in three health care facilities (The Centre de Santé Maternelle et Infantile de Tsaralalana (CSMI), the Centre Hospitalier Universitaire Mère Enfant de Tsaralalana and the paediatric surgery department of the Centre Hospitalier Universitaire Joseph Ravoahangy Andrianavalona). This secondary data analysis focuses on the children living in Antananarivo who were recruited from the community setting. Children included in the AFRIBIOTA project in Antananarivo and for whom data on haemoglobin and ferritin concentrations were available were included in this secondary analysis (Fig. 1). Flow chart of the study participants Data were collected by interviewing mothers/closest caregivers and using a standardized questionnaire. Anthropometric measurements were performed by trained health professionals; blood and stool samples were also collected. Screening and recruitment were conducted at the community level with the support of community health workers. The interviews and the collection of biological samples were conducted at the hospital centres: Centre Hospitalo-Universitaire Mère Enfant de Tsaralalana and Centre Hospitalo-Universitaire Joseph Ravoahangy Andrianavalona. Each child’s weight was measured twice to the nearest 0.1 kg using an electronic scale (KERN, ref. MGB 150 K100 and EKS, People’s Republic of China). When the difference in the two measurements exceeded 0.1 kg, another measurement was performed until the last three values did not differ by more than 0.1 kg. Each child’s height was measured to the nearest 0.1 cm with the child in a standing position using collapsible height boards (ShorrBoard® Infant/Child/Adult Measuring Board, MD, USA). The same procedure was followed for each child to ensure consistent measurement. For both indicators, the mean of the two or three values obtained was reported. Venous blood samples (2 mL) were collected and used in complete blood count, C-reactive protein (CRP), ferritin and citrulline analysis. They were collected in Microtainer® tubes containing ethylenediamine tetraacetic acid (EDTA) and sent at + 4 °C to the Clinical Biology Center of the Institut Pasteur de Madagascar (IPM) within 1 hour after blood collection. One hundred microlitres (100 μL) of plasma was extracted from each sample of whole blood, stored at − 80 °C and sent to the Hôpital Universitaire Necker-Enfants Malades, Paris for citrulline testing. A clean, dry plastic container was given to the mother/caregiver of each child for stool sample collection with detailed instructions on how to collect fresh stool samples. Part of each stool sample was sent to the Unité de Bactériologie expérimentale at IPM as soon as possible for the detection of intestinal parasites. The remainder of each stool sample was stored in liquid nitrogen in the field and shipped to IPM for storage at − 80 °C. An aliquot of each sample was shipped on dry ice to the Service de Coprologie Fonctionnelle, Hôpital Salpétrière Paris for measurement of calprotectin and alpha-antitrypsin levels. The questionnaire collected individual data about each child (diseases requiring hospital admission during the year prior to the survey, feeding practices (age at introduction of complementary feeding, age at cessation of breastfeeding, 24-hour recall)) and about the child’s mother (education level, nutritional status). Household data, including type of housing and amount of household assets, were also collected. A detailed description of the questionnaire is given in [15]. The complete blood count, including haemoglobin assessment, was performed on a SYSMEX autoanalyser (XN 1000 or XT-2000 i) (Landskrona, Sweden) using the fluorocytometric technique. Plasma CRP concentrations were assessed using an enzyme-linked immunosorbent assay (ELISA). Plasma ferritin concentrations were assessed on the ARCHITECT machine (Abbott, IL, USA) using a chemiluminescent microparticle immunoassay (CMIA). These analyses were performed according to standard procedures at the Clinical Biology Centre of IPM (ISO18189 certification). Citrulline was measured by liquid chromatography coupled to tandem mass spectrometry (UPLC–MS/MS) at the Laboratoire de Biochimie Métabolomique et Protéomique, Hôpital Universitaire Necker-Enfants Malades, Paris. For accurate quantification, a stable isotope internal standard of the same structure (purchased from Eurisotop, Saint Aubin, France) was added to the sample before protein precipitation. Before analysis, the samples were derivatized using the AccQ Tag™ Ultra (Waters Corporation, Milford, MA, USA) according to the manufacturer’s recommendations. Amino acid separation was performed on an Acquity™ UPLC system using a CORTECS™ UPLC C18 column (1.6 μm, 2.1 × 150 mm) coupled to a microTQS™ tandem mass spectrometer (Waters Corporation, Milford, MA, USA). Faecal calprotectin was assayed using a “sandwich”-type ELISA that uses a polyclonal Ab system (Calprest; Eurospital). The concentration of α1 antitrypsin (AAT) in faeces was measured using an immunonephelemetric method adapted on the BN ProSpec system (Siemens) [16]. The analysis of these faecal biomarkers was conducted at the Service de Coprologie Fonctionnelle, Hôpital Salpétrière Paris. All faecal samples were physically examined and screened for intestinal parasites as previously described [17]. The main variable of interest was the occurrence of anaemia. Anaemia was defined according to the WHO criteria [18] as Hb less than 110 g/l (adjusted for altitude). Age, sex and height and the 2006 WHO Child Growth Standards for children 24 to 59 months of age [19] were used to calculate children’s height-for-age z scores, which were used to define stunting and normal growth. Stunting and normal growth were defined as height-for-age z score < − 2 SD and height-for-age z score > − 2 SD, respectively. Anaemia was defined as severe when the child’s Hb level was less than 70 g/l and moderate at Hb levels between 70 g/l and 99 g/l. Anaemia was defined as mild if the child’s Hb level was between 100 g/l and 109 g/l [1]. A dietary diversity score (DDS) was calculated by counting the number of food groups consumed by the child during the 24-hour period prior to the survey. The WHO recommends basing the DDS on seven food groups: (1) grains, roots and tubers; (2) legumes and nuts; (3) dairy products; (4) flesh foods (meats/fish/poultry); (5) eggs; (6) vitamin A-rich fruits and vegetables; and (7) other fruits and vegetables. A diverse diet is defined as one that has a DDS of at least four. Accordingly, children with a DDS < 4 were classified as having low dietary diversity; otherwise, they were considered to have an adequate diet [20]. The body mass index (BMI) of the mothers was assessed by dividing their weight (in kilograms) by the square of their body height (in metres). Mothers were classified as underweight if their BMI was < 18.5 kg/m2 and as not underweight if their BMI was ≥18.5 kg/m2. Pregnant mothers were classified according to the categories proposed by Ververs et al. [21]. A wealth index based on a minimal set of assets was created, allowing separation of the subjects into three distinct groups based on principal component analysis (PCA). The minimal set of assets included housing materials (floor and wall materials, ownership of an automobile, telephone, bicycle, motorcycle), access to specific utilities (electricity, plumbing, cooking location), and family size. We defined three household wealth categories according to the clusters observed: the poorest, middle and wealthiest categories. Details of the wealth index have been described previously [15]. Iron deficiency was defined as a plasma ferritin concentration < 12 μg/l in the absence of inflammation [22]. To eliminate the influence of inflammation on ferritin plasma concentrations, a correction factor of 0.67 was used to adjust the ferritin plasma concentration value in the presence of inflammation [23]. A CRP value > 6 mg/l was considered an indicator of inflammation. For citrulline, a value below 7 μmol/l was considered too low, and a value above 43 μmol/l was considered too high according to the normal values provided by the Hôpital Necker Enfants Malades. According to the thresholds used in routine diagnostics at the Hôpital Pitié-Salpêtrière, the threshold for AAT was 1.25 mg/g dry weight, and values above this threshold were considered elevated. For calprotectin, the normal value was equal to or less than 150 μg/g for children 2–3 years of age and equal to or less than 100 μg/g for those between 3 and 5 years of age; children who had values above these thresholds were classified as having elevated values. Statistical analysis was performed using R statistical software (version 3.4.3; The R Foundation for Statistical Computing, Vienna, Austria). Descriptive analysis was performed using proportions for categorical variables and means or medians with interquartile ranges for continuous variables according to their distributions. We used binomial logistic regression model analysis to identify independent predictors of the occurrence of anaemia. A bivariate analysis was performed to identify the explanatory variables to be included in the multivariate analysis. All explanatory variables with p value < 0.20 in the bivariate analysis were included in the logistic regression model. A backwards stepwise logistic regression was applied to obtain the variables associated with the occurrence of anaemia. Explanatory variables included the following: 1) biological characteristics: iron status, presence of intestinal parasites, alpha-antitrypsin and calprotectin levels, status of intestinal damage and repair (citrulline levels in blood); 2) child characteristics: age, gender, nutritional status, occurrence of dental caries or symptoms such as dermatitis, cough, runny nose, or clogged nose, age at introduction of the first complementary food, weaning age, and dietary diversity status; 3) maternal characteristics: body mass index; and 4) household characteristics: wealth index. This study was conducted within the framework of the AFRIBIOTA project, which has been approved by the Ethics Committee for Biomedical Research at the Ministry of Public Health in Madagascar (N°104-MSANP/CE - 12/09/2016) and the Institutional Review Board of the Institut Pasteur (2016–06/IRB). Parents or caregivers were informed about the study and signed the informed consent form before the inclusion of their children. The biological analyses were performed free of charge. Treatments were given to infected and anaemic children according to the national recommendation; the cost of the treatment was covered by the project.