Prevalence, predictors of low birth weight and its association with maternal iron status using serum ferritin concentration in rural Eastern Ethiopia: a prospective cohort study

Introduction: Low birth weight (LBW) is one of the major predictors of perinatal survival, infant morbidity, and mortality, as well as the risk of developmental disabilities and illnesses in future lives. The effect of the nutritional status of pregnant women on birth outcomes is becoming a common research agenda, but evidence on the level of low birth weight (LBW) and its association with prenatal iron status in Ethiopia, particularly among rural residents, is limited. Thus, this study aimed to assess the prevalence, predictors of LBW, and its association with maternal iron status using serum ferritin concentration in Haramaya district, eastern Ethiopia, 2021. Methods: A community-based prospective cohort study design was conducted. Of a total of 427 eligible pregnant women followed until birth, 412 (96.48%) were included in the final analysis. Iron status was determined using serum ferritin (SF) concentration from venous blood collected aseptically from the ante-cubital veins analyzed on a fully automated Cobas e411 (German, Japan Cobas 4000 analyzer series) immunoassay analyzer. Iron deficiency(ID) and iron deficiency anemia (IDA) were classified as having SF less than 15 μg/L and SF less than 15 μg/L and Hb level of < 11.0 g/dl during the first or third trimester or < 10.5 g/dl during the second trimester as well, respectively. Birthweight was measured within 72 h of birth and < 2500 g was considered LBW. Birthweight was measured within 72 h of birth and < 2500 g was considered as LBW. A Poisson regression model with robust variance estimation was used to investigate the factors associated with LBW and the association between maternal iron status and LBW. An adjusted prevalence ratio with a 95% confidence interval was reported to show an association using a p-value < 0.05. Results: About 20.2% (95% CI: 16%-24%) of neonates were born with LBW. The prevalence of LBW was 5.04 (95% CI = 2.78–9.14) times higher among women who were iron deficient during pregnancy compared to those who were normal. The neonates of women who were iron deficient during pregnancy had lower birth weight (aPR=5.04; 95% CI = 2.78–9.14) than the neonates of women who were normal. Prevalence of LBW was higher among mothers who were undernourished (MUAC < 23cm) (aPR = 1.92; 95% CI= 1.33–2.27), stunted (height <145cm) (aPR=1.54; 95% CI=1.04–2.27) and among female neonates (aPR=3.70; 95% CI= 2.28–6.00). However, women who were supplemented with iron and folic acid (IFAS) during pregnancy had a 45% decreased chance of delivering low birth weight (aPR= 0.55; 95% CI=0.36–0.84). Conclusion: We found that LBW is of public health significance in this predominantly rural setting. ID during pregnancy is found to have a negative effect on birth weight. IFA supplementation, the maternal under-nutrition, height, and sex of neonates were identified as predictors of low weight at birth. To improve maternal nutritional status, health interventions must address targeted strategies promoting desirable food behavior and nutritional practices. These include; promoting the consumption of diversified and rich iron food to improve the maternal nutritional status. A continued effort is needed in enhancing universal access and compliance with IFA supplementation to improve maternal health. Intervention strategies that are complementary and comprehensive across the vulnerable periods for women during pregnancy and their neonates that are based on a life-cycle approach are suggested. The study was embedded into the Haramaya Health Demographic Surveillance and Health Research Centre (HDS-HRC), which was established in 2018. The HDS-HRC covers 12 rural kebeles (the lowest administrative unit in Ethiopia) out of 33 kebeles found in the district located approximately 500 KM away from the capital city, Addis Ababa. From 5252 pregnant women in the district during the study period, 2306 were under follow-up of the HDS-HRC [27, 28]. A community-based prospective cohort study was conducted. All pregnant women whose serum ferritin and hemoglobin concentrations were determined at their second or third trimester and who later gave singleton live births were eligible for the study. Neonates reached later on 72 h of birth were excluded. We also excluded pregnant women whose exposure was determined in the first trimester, as the fetal weight gain in the first trimester is known to be minimal. This study is a longitudinal study that obtained birth outcome information of pregnant women, the exposure of LBW was determined from Charan and Biswas's study [29], with the input of 95% confidence level, 5% marginal error, 80% power, and 1:1 ratio between exposed and non-exposed subjects. The expected prevalence of LBW in exposed and non-exposed subjects was taken from prior studies [30]. As a result, 432 sample sizes were computed. Nevertheless, by adding a 10% non-response rate 475 participants were included. After constructing a sampling frame from the HDS-HRC database, simple random sampling was applied to the eight randomly selected kebeles and then the eligible women were selected using the computer-generated lottery method. Nevertheless, for this specific paper, 427 pregnant women who were in their second or third trimester during the baseline survey were considered. Figure 1 displays a detailed description a flow diagram for study into two phases. Study flow shows date and duration, study design, and data collected As described in the previous papers [31, 32], from 475 calculated sample sizes, 448 pregnant women were included in the baseline survey. For this specific paper, 427 whose serum ferritin and hemoglobin concentrations were determined at their second or third trimester during the baseline survey were considered. Of 427 followed pregnant women, birth weight was taken from 412 neonates, as lost to follow-up [2], multiple births [3] and early neonatal death [2], birth weight measurement taken after 72 h of birth [3], and fetal loss [5] were excluded. We recruited participants selected from the HDS-HRC database voluntarily from the community during home-to-home visits. We obtained days in which antenatal care services were not held for pregnant women in the nearby health facilities. On these dates, researchers visited the randomly selected kebeles and eligible pregnant women required for recruitment. Although special attention was given to involving all eligible pregnant women for participation, some of them did not volunteer to participate in the cohort. We made announcements using health extension workers to invite pregnant women to select kebeles in the district. Upon recruitment and signing of informed consent, we assessed socio-economic characteristics, dietary practices, anthropometric measurements, and collected blood samples for hematology and biochemical analysis. We carried out the face-to-face interviews to collect knowledge, attitudes, and obstetric and other pregnancy-related data on house-to-house visits. The two-day training was given by research experts to train data collectors, laboratory professionals, and supervisors before the pre-test on each data collection tool. A day field pre-testing session followed the training in a nearby rural district. These trained data collectors carried out all data collection for this study at the randomly selected kebeles in the HDS-HRC. After the baseline study, participants' telephone numbers and addresses were collected and entered into a book, and we entered the expected date of delivery for each participant. The occurrence of births in the cohort was promptly found and informed by pre-established local community health promoters which the country labeled as ''Women Developmental Army (WDA)" and health extension workers (HEW). We used a structured questionnaire to gather data on the birth experiences of each participant. Birth outcomes such as gestational age at birth, birth weight, birth delivery method, neonatal mortality, stillbirth, abortions, and postpartum morbidity of mothers were obtained. Among all the birth outcomes data, birth weight was used as the outcome variable in this study. Birth weight was gauged within 72 h of birth by trained and well-experienced data collectors who were employed by HDS-HRC. Among all the birth outcomes data, birth weight was used as the outcome variable in this study. As described in the previous papers [31, 32], in the baseline survey, data were collected through face-to-face interviews by trained research assistants using a standard pretested questionnaire (food frequency questionnaire) translated to the local language. The questionnaire contains data on socio-economic, obstetric, maternal perception, food consumption, dietary diversity, knowledge, attitude, and practices of pregnant women. Dietary diversity was assessed using the validated Food Frequency Questionnaire (FFQ), containing 27 commonly the list of food items consumed [33–37]. In addition, Mid-Upper Arm Circumference (MUAC) and maternal height measurements were taken. The questionnaire was initially prepared in English language and translated to the local language (Afan Oromo) by individuals with good command of both languages. It was also pre-tested on 10% of the sample in Kersa District before actual implementation. A 5 ml venous blood was aseptically drawn from the antecubital veins and aliquoted into plain test tubes without anticoagulants. The blood samples were centrifuged, followed by separation of serum, stored frozen at -80 °C, and analyzed at the national chemistry laboratory in Ethiopian Public Health Institute (EPHI). We measured SF and serum high-sensitive C-reactive protein (hsCRP). SF was analyzed on a fully automated Cobas e411 (German, Japan Cobas 4000 analyzer series) immunoassay analyzer by the electro-chemiluminescence (ECL) method using commercial kits supplied by Roche Company, Germany at National Clinical Chemistry Reference Laboratory, EPHI. Whereas, highly sensitive C-protein reactive (hsCRP) was analyzed by Roche/Hitachi Cobas 6000 (c501): (German, Japan Cobas 6000 series of Roche) fully automated clinical chemistry analyzer [38]. The tests were performed by trained and experienced medical laboratory technologists. Two levels of quality control (QC) samples were performed at least once every 24 h when the test is in use, once per reagent kit, and following each calibration to evaluate the functionality of the instrument and reagent, and the results of QC were evaluated using the Levey–Jennings chart (Wesgard rules). The calibration method has been standardized against the WHO International Standard NIBSC code: 03/178, 1st International Standard (IS) NIBSC (National Institute for Biological Standards and Control) "Reagent for Ferritin (human liver)" 80/602, and Reference preparation of the IRMM (Institute for Reference Materials and Measurements) BCR470/CRM470 (RPPHS-Reference Preparation for Proteins in Human Serum) for serum ferritin and serum hsCRP respectively. Calibration was performed as per the standard operating procedures (SOPs).The high ferritin cut-off point (SF < 15 μg/L) recommended by WHO for developing countries was used to define ID to compensate for the effect of infection, which can lead to elevation of the level of ferritin. Serum ID, moderate iron depletion, and ID were termed as SF less than 15 μg/L, 15–30 μg/L, and > 30 μg/L [39]. Serum CRP levels higher than 5 mg/L were termed as high CRP [33]. Anemia was defined as a Hb level of < 11.0 g/dl during the first or third trimester or < 10.5 g/dl during the second trimester [39]. Hemoglobin concentration was measured at each study site by well-trained medical technologists from capillary blood using a portable HemoCue Hb 301®, which is a gold standard for fieldwork. Hemoglobin values were adjusted for altitude as per the Center for Disease Prevention and Control (CDC) recommendation [40]. The significance of babies was computed to the nearest 100 g using calibrated Docbel BRAUNH scale). A LBW was defined as a live birth baby born with a birth weight of < 2500 g, and control was defined as a live birth baby born with a birth weight of ≥ 2500 g [41]. To estimate the economic level of families, a wealth index was employed. The wealth dispersion was generated by applying principal component analysis (PCA). The index was calculated based on the ownership of latrine, agricultural land and size, selected household asset, a quantity of livestock, and source of water for drinking containing 41 household variables. As described in the previous paper [31] nutritional knowledge and attitude towards consumption of an iron-rich diet were evaluated with questions using the Likert scale using PCA and the factor scores were totaled and classified into terciles. Women's autonomy was evaluated by seven validated questions which were adopted from the Ethiopian demographic health survey [12]. For each question, the response was coded as "one" when the decision is made by the woman alone or jointly with her husband, or "zero" otherwise. The detailed methods of data quality assurances used are described in previous papers (3,321), in the baseline survey. Quality assurance during laboratory analysis was monitored in the National Reference Laboratory for Clinical Chemistry at the Ethiopian Public Health Institute (EPHI). The EPHI laboratory is accredited by the Ethiopian National Accreditation Office (ENAO) to conduct tests under ISO 15189:2012, Quality and Competence Medical Laboratory Requirements (accreditation no. M 0025) by well-trained and experienced laboratory professionals and standard operating procedures were strictly followed for respective parameters. Data were double entered using EPiData version 3.1 software. Data were cleaned, coded, and checked for missing and outliers, for further analysis and exported to STATA version 14 (College Station, Texas 77,845 USA) statistical software. The outcome variable was dichotomized as LBW (= 1) if newborns were born with a birth weight of < 2500 g or normal (= 0), otherwise. Thus, the Poisson regression analysis model with a robust variance estimate was fitted to identify predictors of LBW. Bivariate analysis and multivariable analyses were done to identify the association between independent variables and LBW. The backward regression was fitted with selected socio-economic and fertility-related variables. The goodness of fit was checked by Hosmer–Lemeshow statistic and omnibus tests. Possible interactions between covariates were tested. Akaike's information criterion (AIC) and Bayesian information criterion (BIC) were used to test for model fitness. All variables with p < 0.25 in the binary analyses were included in the multivariable analysis after checking for multi-collinearity using variance inflation factors. Adjustments were done for independent variables (age, Mid-Upper Arm Circumference, maternal height, sex of neonates, hemoglobin level, C-reactive protein) for assessing iron status and birth weight. The direction and strength of statistical association were measured using adjusted prevalence ratio (aPR) along with their corresponding 95% confidence interval (CI). An adjusted prevalence ratio (aRR with a 95% confidence interval was reported to show an association at a p-value < 0.05. To estimate the economic level of families, a wealth index was employed. The wealth dispersion was generated by applying principal component analysis (PCA). The index was calculated based on the ownership of the latrine, agricultural land and size, selected household assets, the quantity of livestock, and source of water for drinking containing 41 household variables. As described in the previous paper [31], nutritional knowledge and attitude towards consumption of an iron-rich diet were evaluated with questions using the Likert scale using PCA, and the factor scores were totaled and classified into tertiles. Women's autonomy was evaluated by seven validated questions which were adopted from the Ethiopian demographic health survey [37]. For each question, the response was coded as "one" when a decision is made by the woman alone or jointly with her husband or "zero" otherwise. All methods of this study were carried out under the Declaration of Helsinki's ethical principle for medical research involving human subjects [12]. Ethical approval was obtained from the Institutional Health Research Ethics Review Committee (ref no: IHRERC/266/2020) College of Health and Medical Sciences, Haramaya University before the commencement of the study. Written informed consent was obtained from each participant or legally authorized representatives for those below 16 years of age. The interview was conducted in private and confidentiality of the participants' information was maintained.