Prenatal supplementation with multiple micronutrient supplements or medium-quantity lipid-based nutrient supplements has limited effects on child growth up to 24 months in rural Niger: A secondary analysis of a cluster randomized trial

Background: Prenatal multiple micronutrient supplementation (MMS) and lipid-based nutrient supplementation (LNS) can improve birth outcomes relative to iron-folic acid supplementation (IFA); however, effects on child postnatal growth remain unclear. Objectives: The aim was to compare the effect of prenatal MMS, medium-quantity LNS (MQ-LNS), and IFA on child growth up to 2 y of age. Methods: We conducted a cluster randomized controlled trial of prenatal nutritional supplementation in Madarounfa, Niger. Villages were randomly assigned for pregnant women to receive IFA (17 villages, 1105 women), MMS (18 villages, 1083 women) or MQ-LNS (18 villages, 1144 women). Pregnant women received nutritional supplements weekly until delivery, and children were followed up monthly from 6-8 wk to 24 mo of age. We assessed the effect of prenatal MMS and MQ-LNS compared with IFA and the effect of prenatal MMS compared with MQ-LNS on child length-for-age z scores (LAZ), weight-for-age z scores (WAZ), and weight-for-length z scores (WLZ) at 24 mo of age using generalized linear models. In secondary analyses, we used mixed-effects models to assess the trajectories of anthropometric z scores longitudinally from 6-8 wk to 24 mo. Results: Compared with IFA, MMS and MQ-LNS had no effect on child LAZ, WAZ, or WLZ at 24 mo of age (P > 0.05). Children in the MQ-LNS arm had significantly higher mid-upper arm circumference at 24 mo than children in the MMS arm: mean difference 0.50 cm (95% CI 0.10, 0.91 cm). WAZ and WLZ trajectories were more negative in the MQ-LNS arm compared with IFA and MMS, with lower z scores from 14 to 20 mo of age. However, WAZ and WLZ trajectories converged after 20 mo of age, and there were no differences by 24 mo of age. Conclusions: Prenatal MMS and MQ-LNS had limited effect on anthropometric measures of child growth up to 24 mo of age as compared with IFA in rural Niger. This study was conducted in the Madarounfa Health District, Maradi region of south-central Niger. Maternal and child nutritional status in the region was poor, with 54% of children <5 y being stunted, 19% being wasted, and 43% being underweight (21). Approximately 20% of women of reproductive age were underweight (BMI <18.5) and 43% were anemic (21). We conducted a double-blind, placebo-controlled randomized phase III clinical trial to assess the efficacy of Rotasiil (Serum Institute of India, Pvt Limited), a live, oral rotavirus vaccine against severe rotavirus gastroenteritis (22). Details on the study design, participants, and procedures of the vaccine efficacy trial have been previously published (22). Given evidence of lower efficacy of oral vaccines in high-mortality settings and the potential of nutritional supplementation to boost immunogenicity (23, 24), we nested a cluster randomized controlled trial (RCT) within the parent vaccine trial to test the effect of the type of prenatal nutritional supplementation on infant immune response to 3 doses of a live, oral rotavirus vaccine (immunogenicity trial) (25). By design, the nested immunogenicity trial was conducted concurrently with the parent vaccine trial, drawing from the same population but with separate enrollment and outcome assessment (Figure 1). The unit of randomization in the immunogenicity trial was the village (n = 53). Randomization of village clusters in the immunogenicity trial was stratified by village size (<100, 100–249, ≥250 nonpregnant women of reproductive age), and block randomization with permutated blocks of random sizes was used to allocate villages in a 1:1:1 ratio to 1 of 3 prenatal supplementation arms: 1) IFA, 2) MMS, and 3) LNS. After providing consent for village participation, the head of each village selected the name of 1 of the 3 supplements from a jar, which served as randomized village assignment. Nonpregnant women of reproductive age in participating villages provided informed consent for monthly pregnancy surveillance. Women with a confirmed pregnancy (based on a urine test) were screened for eligibility to enroll in the immunogenicity trial and begin prenatal supplementation. Inclusion criteria for pregnant women in the immunogenicity trial were as follows: <30 weeks’ gestation at the time of enrollment, intended to remain in the study area through delivery and for 2 y thereafter, and did not have a chronic health condition, severe illness, evident pregnancy complications (moderate to severe edema, hemoglobin (Hb) <7 g/dL, or diastolic blood pressure >90 mmHg), or known peanut allergy at the time of enrollment. Eligible women who provided informed consent were enrolled and received the supplement until pregnancy outcome. Women were enrolled into the immunogenicity trial from March 2015 to November 2016. Sequence of events in the parent vaccine trial and the immunogenicity substudy. At 6–8 wk after birth, infants were screened for eligibility for enrollment in the parent double-blind, placebo-controlled phase III vaccine trial (22). Infant inclusion criteria for the vaccine trial were 6–8 wk of age, able to swallow and no history of vomiting within the past 24 h, parent/guardian intended to remain in the study area for 2 y, and parent/guardian provided written informed consent. Infants enrolled in the vaccine trial were followed up monthly until they reached 24 mo of age. Women in the IFA arm received tablets containing 60 mg iron and 400 μg folic acid (Remedica Ltd) as the standard of care. Women were instructed to take 1 tablet daily. Women in the MMS arm received capsules containing a daily dose of 30 mg iron, 400 μg folic acid, and 20 other micronutrients (DSM Nutritional Products). The capsules provided 2 times the RDA for each micronutrient, except for iron, folic acid, calcium, phosphorous, potassium, and magnesium. This dose was more effective in improving birth weight in Guinea Bissau as compared with routine IFA relative to 1 time the RDA compared with IFA (26). Women in the LNS arm received a daily 40-g sachet of fortified, ready-to-use food made of peanuts, oil, dried skimmed-milk powder, and sugar (Nutriset S.A.S.). The LNS contained the same 22 micronutrients as the MMS. Due to its size, the product is classified as a medium-quantity LNS (MQ-LNS) (18). Detailed nutritional composition of the 3 study supplements is shown in Supplemental Table 1. Formative work conducted prior to the start of the trial showed that the 3 supplements were well accepted by the communities and pregnant women (27). A study midwife provided the first package of supplements and instructions for use and storage at the time of enrollment. Community health assistants thereafter conducted weekly home visits to distribute a 10-d supply of the supplements: 7 days’ supply to be consumed until the next scheduled weekly home visit and 3 days’ extra supply. The extra supply was provided in case of loss, damage, or unexpected delay until the next home visit and was returned to the community health assistant at the next home visit if unused. Each week, the community health assistants reviewed supplement adherence, discussed health events and concerns since the last distribution, and provided the next 10-d supply. Since supplements were not identical in appearance, participants and study staff were not blinded to intervention allocation. Data analysts remained blinded to intervention allocation until the analysis was completed. At enrollment, a study midwife collected data on maternal and household socioeconomic and demographic characteristics, conducted a physical and obstetric examination, and assessed maternal anthropometry, Hb, and malaria infection. Maternal weight was assessed using an electronic scale, and underweight defined as a BMI <18.5. Maternal Hb concentration was assessed from a finger-prick sample using a HemoCue machine (HemoCue Hb 301), and anemia was defined as Hb <11 g/dL. Malaria infection was assessed using a rapid diagnostic test [SD Bioline Malaria Ag Pf (HRP-2)]. Adherence to the supplementation regimen was defined as the mean percentage of supplements consumed by the woman, based on used supplement counts made by community health assistants during each home visit divided by the total number of supplements that should have been consumed from enrollment into the trial until delivery. A household wealth index was constructed using principal components analysis of 10 items describing asset and livestock ownership and housing quality. Food security was assessed using the household hunger scale (28). Improved sanitation was defined as household having access to a flush toilet, improved pit latrine, or slab latrine. Improved water source was defined as household using a covered or protected ground well for drinking water. At infant enrollment in the vaccine trial at 6–8 wk of age, study staff assessed child growth at the health facilities. From 3 to 24 mo of age, community health assistants conducted monthly home visits to assess child growth, health, and nutrition. Child weight, length, and mid-upper arm circumference (MUAC) were assessed using standard protocols (29). Child weight was measured to the nearest 0.1 kg using a SECA scale until 6 mo and Salter scale thereafter. Recumbent length was measured to the nearest 0.1 cm using a wooden height board. We calculated anthropometric z scores according to the 2006 WHO child growth standards (30): length-for-age z score (LAZ), length-for-weight z score (WLZ), and weight-for-age z score (WAZ). Extreme values (< -6 z or > 6 z) for all anthropometric z scores were excluded. Stunting was defined as LAZ < -1, underweight as WAZ < −2, and wasting as WLZ < -2 (30). The study was approved by the Comité Consultatif National d'Ethique in Niger, the Comité de Protection des Personnes in France, the Commission d'Ethique de la Recherche sur l'Etre Humain, Hôpitaux Universitaires de Genève in Switzerland, the Research Ethics Review Committee of the WHO in Switzerland, and the Western Institutional Review Board in Olympia, WA, USA. An independent Data Safety and Monitoring Board, established prior to the start of the parent trial, conducted safety reviews for adverse and serious adverse events after half the pregnancies were enrolled and every 6 mo thereafter. The parent trial was registered with ClinicalTrials.gov, identifier {"type":"clinical-trial","attrs":{"text":"NCT02145000","term_id":"NCT02145000"}}NCT02145000. The primary endpoint of the immunogenicity trial designed to test the effect of prenatal nutritional supplementation on infant immune response was anti-rotavirus IgA seroconversion, defined as a ≥3-fold rise in serum titer of anti-rotavirus IgA from Rotasiil dose 1 to 28 d post–Rotasiil dose 3. The immunogenicity trial's sample size was based on power calculations to detect a 20% absolute difference in the proportion of children that seroconvert between nutritional supplements with 90% power and 0.05 α, assuming a 30% seroconversion rate in the IFA arm, 20% nonaccessibility, and 30% exclusion due to detection of rotavirus disease between Rotasiil doses (25). In our primary analysis, we evaluated the effect of prenatal MMS and MQ-LNS compared with IFA and the effect of prenatal MQ-LNS compared with MMS on postnatal growth of singleton children at 24 mo of age. Multiple births were excluded from the analysis. We used generalized linear models to assess differences in continuous LAZ, WLZ, and WAZ at 24 mo of age and log-binomial models to assess the relative risk of stunting, wasting, and underweight at 24 mo of age. We present unadjusted mean differences (MDs) for continuous outcomes and relative risks (RRs) for binary outcomes with their 95% CIs. All models accounted for clustering at the village level using cluster-robust SEs. In secondary analyses, we examined differences in LAZ, WAZ, and WLZ trajectories from 6–8 wk to 24 mo of age using linear mixed-effects models. Trajectory analyses included all singleton live births with at least 1 anthropometric measurement after birth. The models included the intervention arm, month of assessment, and an interaction term between these variables. The trajectory models accounted for clustering by village and a compound symmetry correlation structure for within-subject correlation. We tested for difference in trajectory over time for each group comparison: MMS vs. IFA, MQ-LNS vs. IFA, and MQ-LNS vs. MMS. If the test for difference in trajectory for a group comparison (interaction of intervention arm and assessment month) was statistically significant, we presented differences in mean z scores at each month of assessment from 6–8 wk to 24 mo. Differences were tested applying a Tukey-Kramer adjustment for multiple comparisons. As a sensitivity analysis of the primary and secondary analyses, we estimated multivariate models to account for potential imbalance between randomized arms at enrollment and to potentially increase precision (31). We controlled for the following prespecified covariates, which are known predictors of child growth: household wealth, size, and food security; maternal age, education, and underweight; and child age and sex. We also controlled for maternal anemia, malaria infection, and whether the woman was enrolled into the immunogenicity trial during the hunger season (May–September). Multivariate models also accounted for whether the child was randomly assigned to the vaccine or placebo group of the parent vaccine trial. Finally, we explored potential effect modification of MMS and MQ-LNS, relative to IFA, on postnatal growth outcomes by prespecified enrollment factors: maternal education, anemia, underweight, and season of enrollment into the trial; household wealth, food security, improved sanitation, and improved water source; child sex; and adherence to supplements. Interactions were considered statistically significant at P < 0.10 based on a Wald test. All analyses used the intention-to-treat principle and were conducted in Stata version 16 (StataCorp LP) (32).