Maternal supplementation with small-quantity lipid-based nutrient supplements compared with multiple micronutrients, but not with iron and folic acid, reduces the prevalence of low gestational weight gain in semi-urban Ghana: A randomized controlled trial

Background: It is unclear whether maternal supplementation with small-quantity lipid-based nutrient supplements (SQ-LNSs; 118 kcal/d) affects maternal weight. Objective: We compared several secondary anthropometric measures between 3 groups of women in the iLiNS (International Lipid-based Nutrient Supplements)-DYAD trial in Ghana. Methods: Women (n = 1320; < 20 wk of gestation) were randomly assigned to receive 60 mg Fe + 400 mg folic acid/d (IFA), 18 vitamins and minerals/d [multiple micronutrients (MMNs)], or 20 g SQ-LNSs with 22 micronutrients/d (LNS) during pregnancy and a placebo (200 mg Ca/d), MMNs, or SQ-LNSs, respectively, for 6 mo postpartum. Weight, midupper arm circumference (MUAC), and triceps skinfold (TSF) thickness at 36wk of gestation and 6mo postpartumwere analyzed, as were changes from estimated prepregnancy values.We assessed the adequacy of estimated gestational weight gain (GWG) by using Institute of Medicine (IOM) and International Fetal and Newborn Growth Standards for the 21st Century (INTERGROWTH-21st) guidelines. Results: The estimated prepregnancy prevalence of overweight or obesity was 38.5%. By 36 wk of gestation, women (n = 1015) had a mean 6 SD weight gain of 7.4 6 3.7 kg and changes of 21.0 6 1.7 cm in MUAC and 22.8 6 4.1 mm in TSF thickness. The LNS group had a lower prevalence of inadequateGWGon the basis of IOM guidelines (57.4%) than the MMN (67.2%) but not the IFA (63.1%) groups (P = 0.030), whereas the prevalence of adequate (26.9% overall) and excessive (10.4% overall) GWG did not differ by group. The percentages of normal-weight women (in kg/m2: 18.5 < body mass index < 25.0; n = 754) whose GWG was less than the third centile of the INTERGROWTH-21st standards were 23.0%, 28.7%, and 28.5% for the LNS, MMN, and IFA groups, respectively (P = 0.36). At 6 mo postpartum, the prevalence of overweight or obesity was 45.3%, and the risk of becoming overweight or obese did not differ by group. Conclusion: SQ-LNS supplementation is one potential strategy to address the high prevalence of inadequateGWGinwomen in settings similar to Ghana, without increasing the risk of excessive GWG. We have previously described the setting, design, and participants of the iLiNS-DYAD efficacy trial in Ghana (11). Briefly, the study was conducted in a semi-urban settlement (Somanya-Odumase-Kpong area) in the Yilo Krobo and the Lower Manya Krobo districts ∼70 km north of Accra, Ghana. It was designed as a parallel, individually randomized, controlled trial with 3 equal-size groups. The study protocol was approved by the ethics committees of the University of California, Davis; the Ghana Health Service; and the University of Ghana Noguchi Memorial Institute for Medical Research and was registered at clinicaltrials.org (identifier {"type":"clinical-trial","attrs":{"text":"NCT00970866","term_id":"NCT00970866"}}NCT00970866). Participants were pregnant women aged ≥18 y, at ≤20 wk of gestation, who were attending usual antenatal clinics in the 4 main health facilities in the area between December 2009 and December 2011, and who had antenatal cards complete with history and examination. A woman was excluded if any of the following applied: not residing in the area, intention to move out of the area within the next 2 y, milk or peanut allergy, unwillingness to receive fieldworkers or take study supplement, participation in another trial, gestational age >20 wk before completion of enrollment, and antenatal card indicating HIV infection, asthma, epilepsy, tuberculosis, or malignant disease. We previously described (11) that after baseline assessments, eligible women were randomly assigned to receive 1 of 3 treatments: 1) 60 mg Fe + 400 μg folic acid/d during pregnancy and placebo (200 mg Ca/d) during the first 6 mo postpartum (IFA group), 2) MMN capsules containing 18 vitamins and minerals (including 20 mg Fe)/d during pregnancy and the first 6 mo postpartum (MMN group), and 3) 20 g/d of SQ-LNSs containing micronutrients similar to those in the MMN supplement plus calcium, phosphorus, potassium, and magnesium as well as energy (118 kcal/d) and macronutrients (e.g., protein and essential FAs) during pregnancy and the first 6 mo postpartum (SQ-LNS supplement or LNS group). The SQ-LNS (individual 20-g sachets) was supplied by Nutriset S.A.S. and the IFA and MMN (10 capsules/blister pack) by DSM South Africa. We previously published the nutrient contents of all supplements (11) and the rationale for the concentrations of the nutrients (16). Apart from iron, the vitamin and mineral contents of the MMN and SQ-LNS supplements were either 1 or 2 times the RDA for pregnancy or, in a few cases, the maximum amount that could be included in the supplement given technical and organoleptic constraints (16). Group allocations were developed by the study statistician with the use of a computer-generated (SAS version 9.4) scheme in blocks of 9, and randomization was performed by the study nurse who offered 9 sealed, opaque envelopes at a time, one of which was picked by the participant to reveal the allocation. Allocation information was securely kept by one field supervisor and the study statistician only. At enrollment, the study nurse gave each woman a 2-wk supply of the assigned supplement with instructions on how to consume it (IFA and MMN, with water after a meal, 1 capsule/d; SQ-LNSs, mixed with any prepared food, one 20-g sachet/d) and a standard nutrition message (“Do not forget to eat meat, fish, eggs, fruits, and vegetables whenever you can; you still need these foods even as you take the supplement we have given you”) designed to reflect Ghana Health Service’s nutritional advice for women during pregnancy (17) and local food availability. Thereafter, fieldworkers visited women in their homes biweekly, at which time they delivered fresh supplies of supplement and monitored supplement intakes and morbidity (11). After the women gave birth, fieldworkers visited them and their infants each week, but the delivery of supplement and monitoring of intakes and morbidity were done biweekly as before until women exited the study at 6 mo postpartum. During follow-up, women were told not to consume >1 capsule (IFA and MMN groups) or sachet (LNS group)/d, even if they forgot to take the supplement the previous day or days. Women were told to take their assigned supplement with them if they wanted to travel out of the study area. Those who would not return before the next biweekly visit were given an extra supply for the period they intended to be away. To maintain blinding, the IFA and MMN supplements were color-coded (3 different colors for IFA and 3 for MMN supplements) and were therefore known to the study team and participants only by the colors. It was not possible to blind fieldworkers and participants to the IFA and MMN supplements compared with the SQ-LNS supplement due to their apparent differences, but the anthropometrists who measured the women were blinded to the group assignments, and no one apart from the study statistician had knowledge of group assignment until all preliminary analyses had been completed. The outcome measures evaluated at 36 wk of gestation (last laboratory visit before delivery) were as follows: total gestational weight (kilograms), midupper arm circumference (MUAC; centimeters), and triceps skinfold (TSF) thickness (millimeters) gain; GWG per week; percentage adequacy of GWG achieved; percentage of women whose GWG was inadequate (less than the lower cutoff of recommendations), adequate (within the recommended range), or excessive (more than the upper cutoff of recommendations) according to the IOM GWG recommendations (13); and the percentage of normal-weight (in kg/m2; 18.5 < BMI < 25.0) women whose estimated GWG was less than the third centile or >97th centile of the INTERGROWTH-21st standards (14). Outcomes evaluated at 6 mo postpartum (maternal endpoint of the study) were weight, MUAC, TSF thickness, and BMI; change in weight, MUAC, TSF thickness, and BMI from prepregnancy; and the percentage of women who became overweight or obese out of those who were not overweight or obese at prepregnancy. We collected background demographic and socioeconomic information at enrollment by using a questionnaire and completed anthropometric and laboratory assessments at enrollment, 36 wk of gestation, and 6-mo postpartum. Weight (Seca 874; Seca), height (Seca 217; Seca), MUAC, and TSF thickness (Holtain calipers) were measured with the use of standard procedures. Blood hemoglobin concentration was measured by HemoCue (HemoCue AG), and malaria parasitemia was assessed by using a kit (Vision Biotech) (11). As described previously (11), gestational age was determined mostly by ultrasound biometry (Aloka SSD 500). The sample size for the iLiNS-DYAD Ghana study (11) was based on detecting a small-to-moderate (18) effect size (Cohen’s d) of 0.3 between any 2 groups for any continuous outcome measure, with a 2-sided 5% test and 80% power. As previously reported (11), a total of 1320 pregnant women were enrolled into the study. We also previously reported (11) a temporary mislabeling of IFA and MMN capsules, as a result of which 170 women who had been assigned to the IFA group inadvertently received the MMN capsule either throughout pregnancy (n = 85) or during part of their pregnancy (n = 85) before receiving the intended IFA capsule for the rest of follow-up, and another 170 women assigned to the MMN group also received the IFA capsule either throughout pregnancy (n = 78) or during part of their pregnancy (n = 92) before receiving the intended MMN capsule. In this current analysis that covered both pre- and postnatal periods of the intervention, we included all of the women enrolled into the study without discarding any data from those who received the unintended supplements, because the unintended exposure occurred only in the prenatal period. We used this same approach in our previous publication (19) in which we evaluated the impact of the intervention spanning pre- and postnatal periods on the attained growth of 18-mo-old children. We estimated that the actual percentage of follow-up days (13%) during which the women in the IFA and MMN groups had the unintended exposure was relatively small (19), and in addition, no women in the LNS group were exposed to any other supplement apart from the intended SQ-LNSs. At 36 wk of gestation and 6-mo postpartum, we had anthropometric data for 1015 and 1073 women, respectively. With these sample sizes, we had >97% power to detect an effect size of 0.3 between any 2 groups for any continuous outcome at each of the time points. We developed our statistical analysis plan and posted it on our website (www.ilins.org) before data analysis. The secondary outcomes in the present analysis were prespecified in the statistical analysis plan. Statistical analysis was performed on an intention-to-treat basis by using SAS for Windows Release 9.4. Thus, women were included in the analysis regardless of adherence to treatment. To address the protocol violation as a result of the consumption of mislabeled capsules by some women during pregnancy, we analyzed the data by using 2 scenarios as done previously (19). In the first, intervention groups were based on the supplement that women were intended to receive when they were enrolled, and in the second, intervention groups were based on the supplement women that actually received when they were enrolled. The variables used to assess the outcome measures were obtained as follows: first, because it was not possible to obtain women’s prepregnancy weight (needed to apply the IOM GWG recommendations), we used a third-degree polynomial regression model with 1 predictor variable (gestational age at enrollment) to estimate prepregnancy weight on the basis of weight and gestational age at the time of enrollment. In this case, the shape of the true nonlinear response function of maternal weight to gestational age was unknown (or complex), and a polynomial function was a good approximation of the true function (20). The procedure was accomplished by first determining the best transformation of the weight at enrollment that achieved a normal distribution and regressing the transformed weight on gestational age, gestational age squared, and gestational age cubed in order to generate predicted and residual values. We then inspected the regression curve to determine the earliest gestational age before the CI expanded, assuming that weight gain before that time was minimal. Next, we determined the mean of the predicted values at the selected time point in early gestation, added this mean value to the residual for each individual, and then back-transformed the result to obtain the estimated prepregnancy weight for the individual. We used the same approach to estimate the prepregnancy MUAC and TSF thickness on the basis of the values measured at enrollment. We estimated prepregnancy BMI as estimated prepregnancy weight divided by the square of height measured at the time of enrollment. We calculated the total GWG, gestational MUAC gain, and gestational TSF thickness gain by subtracting the estimated prepregnancy values from those measured at the last prenatal visit to the laboratory at ∼36 wk of gestation, an approach used by other investigators (21). The rate of GWG by 36 gestational weeks was calculated as total GWG divided by completed weeks of gestation (22). The percentage adequacy of GWG as a continuous variable was calculated by dividing the total GWG by the expected GWG (i.e., the amount of weight a woman was supposed to gain according to the IOM recommendation when her weight was measured at ∼36 wk of gestation) and multiplying the result by 100 (21). For the expected GWG, we used the following formula: expected GWG = expected first-trimester total weight gain + [(gestational age at the last weight measurement at ∼36 wk of gestation − 13 wk) × recommended rate of GWG for the second and third trimesters] (21, 23, 24). The expected first-trimester total weight gain was assumed to be 2 kg for underweight and normal-weight women, 1 kg for overweight women, and 0.5 kg for obese women (21); and the recommended rates of GWG for the second and third trimesters were 0.5, 0.4, 0.3, and 0.23 kg/wk for underweight, normal-weight, overweight, and obese women, respectively (25). Because the IOM recommends a range of total GWG for each prepregnancy BMI group, we classified the percentage of weight-gain recommendations met as inadequate, adequate, or excessive (26). For each BMI-specific range, we divided the lower and upper limits of the recommended weight-gain range by the expected weight gain by 40 wk of gestation and multiplied the result by 100 to obtain the corresponding range of the recommended percentage of expected weight gain (26). For example, for women with a normal prepregnancy BMI, the expected GWG by 40 wk of gestation is as follows: 2.0 kg + [(40 wk − 13 wk) × 0.4 kg/wk] = 12.8 kg. For the IOM’s recommended total weight-gain range of 11.5–16 kg, the lower and upper limits of the corresponding range of the recommended percentage of expected weight gain are as follows: 11.5 kg/12.8 kg × 100 and 16 kg/12.8 kg × 100 = 90% − 125% of the 12.8-kg expected weight gain. Hence, we classified inadequate, adequate, and excessive weight gain as <90%, 90–125%, and >125% of recommendations, respectively. For a normal-weight woman who gained a total of 9.0 kg and whose weight was last measured at 37 wk of gestation, the expected GWG would be 11.7 kg and the percentage of recommendations met would be 9.0 kg/11.7 kg × 100 = 77%, which would be classified as inadequate GWG. Changes in weight, MUAC, TSF thickness, and BMI from prepregnancy to 6 mo postpartum were calculated by subtracting the estimated prepregnancy values, from the values measured at 6 mo postpartum. To examine how the GWG assessment with the use of the IOM’s recommendation would compare with a similar assessment with the use of the INTERGROWTH-21st standards, we calculated the 3rd and the 97th centiles of expected GWG at the last antenatal measurement at ∼36 wk of gestation. For the third centile of GWG, we used the following formula: where GA = weeks of gestation at the last antenatal measurement (14). For the 97th centile of GWG, we replaced the −1.88 in the formula for the 3rd centile of expected GWG by 1.88 (14). We used actual GWG less than the calculated 3rd centile of the expected GWG as a proxy for inadequate GWG and actual GWG greater than the calculated 97th centile of expected GWG as a proxy for excessive GWG. Because the INTERGROWTH-21st standards are appropriate for normal-weight women, only women in our sample who had an estimated prepregnancy BMI between 18.5 and 24.99 were included in the analysis involving those standards. We summarized the background characteristics at enrollment (11) and the number of days from the last prenatal measurements (∼36 wk of gestation) to delivery as means ± SDs or frequencies (percentage) by using the group assignment based on supplements that women were intended to receive when they were enrolled. At 36 wk of gestation and 6 mo postpartum, we calculated descriptive statistics for the overall sample, before comparing the 3 treatment groups by using general linear models (continuous outcomes) and logistic regression models (binary) with Tukey-Kramer adjustment for multiple comparisons. Along with the group comparisons, we calculated pairwise mean differences (continuous outcomes) and RRs (binary outcomes) with their 95% CIs and P values. RRs were calculated by using Poisson regression (27). These comparisons were performed twice, first without any covariate adjustments and then with adjustment for covariates that were significantly associated (P < 0.10) with the outcome in question in a bivariate analysis. Potential covariates were specified before analysis and included primiparity (yes or no), season at enrollment (wet or not wet), anemia (yes or no), age, gestational age at enrollment, and assets index, housing index, and Household Food Insecurity Access Scale score derived by using principal components analysis (28). Finally, as done previously (19) as a possible option for addressing the protocol violation, we performed a secondary analysis of the outcome variables, in which we combined women in the IFA and MMN groups to conduct a 2-group comparison with those who consumed the SQ-LNSs. We considered that for these assessments in a clinical trial, in which all of the outcomes were secondary, prespecified, and highly correlated, correcting for multiplicity was unnecessary (29). Statistics in the texts are means ± SDs (continuous outcomes) or percentages (binary outcomes). Women’s adherence to supplement intake, defined as the percentage of follow-up days that women self-reported consuming the supplements (pregnancy/lactation) was as follows: 88.1%/85.7% for the IFA group, 87.0%/85.0% for the MMN group, and 83.7%/80.0% for the LNS group, as reported previously (30). Data on morbidity have yet to be analyzed, but we reported previously (11, 19) that serious adverse events were evenly distributed across the 3 groups.