Prevalence and determinants of gestational weight gain among pregnant women in Niger

Low gestational weight gain (GWG) and low mid-upper arm circumference (MUAC) are associated with adverse pregnancy outcomes. We aimed to assess the prevalence and determinants of low GWG and low MUAC among pregnant women in rural Zinder, Niger. A community-based survey was conducted among 1,384 pregnant women in the catchment areas of 18 integrated health centers in the region of Zinder, Niger. Weight and MUAC were measured during an in-home visit and again 1 month later, when haemoglobin concentration and micronutrient status were also assessed. The prevalence of low GWG was defined based on the 2009 United States Institute of Medicine (U.S. IOM) guidelines (<0.35 kg/week) and less than the third centile of the International Fetal and Newborn Growth Consortium for the 21st Century (INTERGROWTH-21st) standards. Factors associated with GWG and MUAC were identified using logistic regression models adjusting for season, village, and gestational age. The median (interquartile range) age was 25.0 (20.7, 30.0) years, and 16.4% were ≤19 years. The prevalence of low GWG were 62.9% and 27.5% according to 2009 IOM and less than the third INTERGROWTH-21st centile, respectively; 24.9% had low MUAC. Higher α-1-acid glycoprotein (OR = 1.7, 95% CI [1.1, 2.8]) and C-reactive protein (OR = 1.2, 95% CI [1.02, 1.50]) increased the odds of low GWG. Adolescents (OR = 2.7, 95% CI [1.8, 4.0]), housewives (OR = 1.97, 95% CI [1.36, 2.86]), and those who reported recent food assistance (OR = 1.80, 95% CI [1.04, 3.11]) had higher odds of low MUAC. Prevalence of low GWG and low MUAC was high among pregnant women. Determinants of GWG and MUAC included socio-economic, demographic, and biological factors, although only markers of inflammation were consistent predictors across different definitions of low GWG. The present study was a community‐based survey conducted as part of the baseline assessment for the Niger Maternal Nutrition (NiMaNu) Project. This project was a programmatic intervention to improve antenatal care (ANC) services and compared pre‐ and post‐intervention cohorts of pregnant women. Methods for the baseline survey assessments have been described in detail elsewhere (K. Begum et al., 2018; Wessells et al., 2017). Briefly, the community‐based baseline survey was conducted from March 2014 to September 2015. We enrolled pregnant women in 68 rural villages belonging to 18 integrated health centers (IHCs) in the catchment area of two health districts (Mirriah and Zinder), in the Zinder region, Niger. The 18 IHCs were selected based on their accessibility and their distance to Zinder, the regional capital, and because of the limited number and scope of interventions that were being implemented in their catchment area. Within the catchment area of each IHC, the village containing the IHC was automatically included in the survey, as well as one additional randomly selected village among those with a health post. Among the remaining villages in the catchment area of each IHC, four villages ≤10 km and four villages >10 km were randomly selected and randomized to order of participation. Pregnant women from the first two selected villages (IHC‐village and health post village), and the first two villages from the subsequent randomization were enrolled, with a target of enrolling approximately 16–20 women per village and a sample size of approximately 77 women per IHC. When the target number of women was not met by the first four villages in each IHC, women were included from the remaining villages of each IHC following the order of the randomization list. Within each village, participants were identified using a random walk method (United Nations, 2008). The enrolment of participants was implemented continuously over a period of 18 months with approximatively one new IHC surveyed each month. All identified pregnant women (regardless of gestational week) were eligible for study participation, if they had resided in a participating village for at least 6 months prior to enrolment and had no plans to move out of the study area within the next 2 months. A woman was excluded if she had a severe illness warranting immediate hospital referral or was unable to provide consent due to impaired decision‐making ability. Each enrolled woman participated in two study visits. During the first contact (Visit 1), we obtained written informed consent and interviewed women using a structured questionnaire to collect information regarding SES, demographics, and knowledge, attitude, and practices relating to diet, health, pregnancy (current and previous), and ANC attendance. Pregnant women were weighed in light clothing in duplicate to 50‐g precision (SECA 874). Women’s height (SECA 213, Seca, Hamburg, Germany), MUAC (ShorrTape© Measuring Tape), and symphysis‐fundal height (ShorrTape© Measuring Tape, Weigh and Measure, Olney, MD) were measured in duplicate to 0.1‐cm precision. A third measurement was performed, and the mean of the two closest measurements was calculated when the first two measurements were >0.2 kg (weight) or >0.5 cm apart (height, MUAC, and symphysis‐fundal height). Approximatively 1 month later (Visit 2), each participating pregnant woman was invited to a follow‐up assessment. The structured interviews and anthropometric measurements were repeated. Capillary blood samples were drawn to assess haemoglobin concentration by HemoCue® Hb 201+ (Hemocue, Inc; Lake Forest, CA). As described elsewhere (Wessells et al., 2017), venous blood samples (7.5 ml) were collected in a subgroup of participants for the measurement of folate, vitamin B12, retinol binding protein, plasma ferritin, soluble transferrin receptor (sTfR), zinc, α‐1‐acid glycoprotein (AGP), C‐reactive protein (CRP), and histidine‐rich protein II (HRP2) concentrations. GA was estimated as a weighted average of the following obtained information: reported last menstrual period (by estimated number of months, lunar cycles, and/or proximity to a religious or cultural event), time elapsed since quickening, and two fundal height measurements taken approximately 1 month apart (Hess & Ouedraogo, 2016). Three proxy indices (housing quality, household assets, and household livestock) were used to estimate the household SES, as previously described (K. Begum et al., 2018). Household food insecurity was assessed using the Household Food Insecurity Access categories (Coates, Swindale, & Bilinsky, 2007). Pregnant women’s dietary practices were assessed using a list‐based food frequency questionnaire, and those who reported consuming at least five of 10 defined food groups in the previous 24 hr were considered to meet the Minimum Dietary Diversity for Women (FAO and FHI 360, 2016). The outcomes of this study included GWG per week (in kilograms), low GWG, MUAC (in centimeters), and low MUAC. GWG per week was calculated by subtracting weight at Visit 1 from the weight recorded on Visit 2 divided by the number of elapsed days and multiplied by seven. Adequacy of GWG was assessed by comparing GWG of the study participants with the 2009 U.S. IOM guidelines for GWG and the INTERGROWTH‐21st standards, as described in more detail below. Low MUAC was defined as <23 cm (Ververs et al., 2013). The mean GWG recommended by the U.S. IOM is 0.45 kg/week, with ranges from 0.44 to 0.58 kg/week for underweight and 0.35 to 0.50 kg/week for normal weight women, respectively (IOM and NRC, 2009). These guidelines are based on pre‐pregnancy BMI, which are not known for the participants in the present study. Thus, considering that the majority of women in the present study were likely to be either underweight or of normal weight (Institut National de la Statistique & ICF International, 2007), a GWG of 0.35–0.58 kg/week was considered within the IOM guidelines. Less than 0.35 kg/week was considered as GWG below the IOM guidelines (or a proxy for low GWG) and >0.58 kg/week as GWG above the IOM guidelines (or a proxy of excessive GWG). Because the IOM guidelines apply to women in the second and third trimester of gestation, women in their first trimester of pregnancy were excluded from this classification. Considering that the majority of women had low or adequate GWG in the present study population, GWG was transformed in a dichotomous variable (i.e., GWG < 0.35 kg and GWG ≥ 0.35 kg) to assess factors associated with low GWG. We calculated the observed GWG per week for each pregnant woman according to her estimated GA in weeks. Because the published INTERGROWTH‐21st standards represent cumulative GWG, they cannot be directly compared with the observed GWG per week. Rather, the observed GWG was compared with the GWG rate per week of women enrolled in the Fetal Growth Longitudinal Study of the INTERGROWTH‐21st Project (personal communication). Observed GWG less than the third centile of the GWG by GA of the INTERGROWTH‐21st standards was considered to be GWG below the INTERGROWTH‐21st standards (or a proxy for low GWG). Observed GWG between the third and the 97th centile was considered GWG within the standards (or a proxy of normal weight gain), and observed GWG > 97th centile of expected GWG was considered GWG above the INTERGROWTH‐21st standards (or a proxy of excessive GWG). Observed GWG per week of the participants was also compared with the median of the INTERGROWTH‐21st standards and categorically defined as being above or below the INTERGROWTH‐21st median. The overall sample size for the NiMaNu project was specified to be able to detect with 80% power a difference of 10% in the prevalence of anaemia as the primary outcome of the programmatic intervention (Hess & Ouedraogo, 2016). Assuming an initial anaemia prevalence of 50%, a significance level of 0.05, power of 0.80, and a design effect of 2 to account for the cluster sampling design, a sample size of 768 was needed, which was then inflated by 17% for attrition, yielding a target sample size of 925 pregnant women for the baseline survey. However, for the impact assessment of the main trial described elsewhere (Hess & Ouedraogo, 2016), the baseline survey was extended for 6 months to allow statistical models to account for the potential seasonal effects of participants’ enrolment on the outcome measures. Based on the same assumption as above, the additional sample size required for the baseline survey was estimated at 77 pregnant women per month for a total of 463 over 6 months. In total, 1,388 pregnant women were needed to be enrolled in the baseline survey to provide 80% power to detect a difference of 10% in the prevalence of anaemia in the primary impact assessment. Data were successfully obtained from 1,385 pregnant women. This sample size was adequate to estimate the prevalence of low GWG + 3.5% (95% CI), assuming a prevalence of 50%. Data were double‐entered and compared using EpiData Entry version 3.1 (Odense, Denmark). All statistical analyses were performed using the SAS System version 9.4 (SAS Institute, Cary, NC, USA). We analysed available data from the baseline survey of the NiMaNu project. Data were examined using univariate analysis (graphical plotting) to look for outliers. Outliers that were clearly impossible or implausible values were corrected if possible, or trimmed when correction was not possible, which was the case for one GWG and one MUAC measurement. A detailed statistical analyses plan is available (Hess & Ouedraogo, 2016). GWG per week and MUAC were assessed for conformance to the normal distribution. Predictors not normally distributed (i.e., ferritin, sTfR, CRP, AGP, and folate and vitamin B12) were natural log transformed. Descriptive analysis of initial characteristics of study participants was performed. Factors associated with low GWG and low MUAC, as well as GWG per week and MUAC as continuous variables, were identified using generalized estimating equation models, in SAS proc glimmix to permit adjusting for cluster effects by village. All models were minimally adjusted to include year, season, and village, and analyses were performed using robust standard errors. All predictors were run in individual models, and the minimally adjusted odds ratio (for low GWG and MUAC as binary outcomes) and the minimally adjusted mean difference (for GWG and MUAC as continuous outcomes) from each individual model were reported. Potential predictors were identified based on a literature review and background knowledge and prespecified in the statistical analyses plan (Hess & Ouedraogo, 2016). These included maternal age, education, number of pregnancies, number of living children, height, occupation, SES, household food insecurity, reported increase or decrease in the number of meals per day and quantity of food consumed due to pregnancy, reported receipt of food assistance, adequate minimum dietary diversity, micronutrient status (plasma ferritin, zinc, and retinol binding protein adjusted for inflammation; Wessells et al., 2017; sTfR, vitamin B12, and folate), markers of inflammation (AGP and CRP), and malaria antigenemia (HRP2). To explore which predictors were consistently and significantly associated with GWG per week and low GWG, we ran five independent analyses including GWG per week and GWG < 0.35 kg/week adjusting for women's GA, and GWG per week, GWG less than the third centile and GWG < 50th centile INTERGROWTH‐21st standards not adjusting for GA following the methods of the respective standards (Cheikh Ismail et al., 2016; Hutcheon & Bodnar, 2018; IOM and NRC, 2009). If a predictor was associated with at least three GWG (GWG per week and/or different definitions of low GWG) or both of the MUAC outcomes (MUAC in centimeters or low MUAC), it was considered to be a consistent predictor of low GWG or undernutrition. A P value <.05 was considered as statistically significant for the all tests performed. This study was part of the NiMaNu Project, which was approved by the National Ethical Committee in Niamey (Niger) and the Institutional Review Boards of the University of California, Davis (USA). The study was registered at www.clinicaltrials.gov as {"type":"clinical-trial","attrs":{"text":"NCT01832688","term_id":"NCT01832688"}}NCT01832688. In the presence of a neutral witness, consent materials were presented in both written and oral formats. Informed consent was obtained and documented with a written signature or a fingerprint.