Fortified balanced energy–protein supplementation during pregnancy and lactation and infant growth in rural Burkina Faso: A 2 × 2 factorial individually randomized controlled trial

Background Optimal: nutrition is crucial during the critical period of the first : 1,000 days from conception to 2 years after birth. Prenatal and postnatal supplementation of mothers with multimicronutrient-fortified balanced energy–protein (BEP) supplements is a potential nutritional intervention. However, evidence on the long-term effects of BEP supplementation on child growth is inconsistent. We evaluated the efficacy of daily fortified BEP supplementation during pregnancy and lactation on infant growth in rural Burkina Faso. Methods and findings A 2 × 2 factorial individually randomized controlled trial (MISAME-III) was implemented in 6 health center catchment areas in Houndé district under the Hauts-Bassins region. From October 2019 to December 2020, 1,897 pregnant women aged 15 to 40 years with gestational age <21 completed weeks were enrolled. Women were randomly assigned to the prenatal intervention arms receiving either fortified BEP supplements and iron–folic acid (IFA) tablets (i.e., intervention) or IFA alone (i.e., control), which is the standard of care during pregnancy. The same women were concurrently randomized to receive either of the postnatal intervention, which comprised fortified BEP supplementation during the first 6 months postpartum in combination with IFA for the first 6 weeks (i.e., intervention), or the postnatal control, which comprised IFA alone for 6 weeks postpartum (i.e., control). Supplements were provided by trained village-based project workers under direct observation during daily home visits. We previously reported the effect of prenatal BEP supplementation on birth outcomes. The primary postnatal study outcome was length-for-age z-score (LAZ) at 6 months of age. Secondary outcomes were anthropometric indices of growth (weight-for length and weight-for-age z-scores, and arm and head circumferences) and nutritional status (prevalence rates of stunting, wasting, underweight, anemia, and hemoglobin concentration) at 6 months. Additionally, the longitudinal prevalence of common childhood morbidities, incidence of wasting, number of months of exclusive breastfeeding, and trajectories of anthropometric indices from birth to 12 months were evaluated. Prenatal BEP supplementation resulted in a significantly higher LAZ (0.11 standard deviation (SD), 95% confidence interval (CI) [0.01 to 0.21], p = 0.032) and lower stunting prevalence (−3.18 percentage points (pp), 95% CI [−5.86 to −0.51], p = 0.020) at 6 months of age, whereas the postnatal BEP supplementation did not have statistically significant effects on LAZ or stunting at 6 months. On the other hand, postnatal BEP supplementation did modestly improve the rate of monthly LAZ increment during the first 12 months postpartum (0.01 z-score/month, 95% CI [0.00 to 0.02], p = 0.030), whereas no differences in growth trajectories were detected between the prenatal study arms. Furthermore, except for the trend towards a lower prevalence of underweight found for the prenatal BEP intervention at 6 months (−2.74 pp, 95% CI [−5.65 to 1.17], p = 0.065), no other secondary outcome was significantly affected by the pre- or postnatal BEP supplementation. Conclusions This study provides evidence that the benefits obtained from prenatal BEP supplementation on size at birth are sustained during infancy in terms of linear growth. Maternal BEP supplementation during lactation may lead to a slightly better linear growth towards the second half of infancy. These findings suggest that BEP supplementation during pregnancy can contribute to the efforts to reduce the high burden of child growth faltering in low- and middle-income countries. Our research was reported using the Consolidated Standards of Reporting Trials (CONSORT) 2010 checklist [26]. The MISAME-III study protocol has been described elsewhere in detail [27]. In brief, MISAME-III is an individually RCT evaluating the efficacy of a combination of a daily fortified BEP supplement and iron–folic acid (IFA) tablet, as compared to IFA supplementation alone, during pregnancy and lactation. The study follows a 2 × 2 factorial design where women were enrolled into the study to receive BEP supplementation during pregnancy (prenatal intervention) and/or the first 6 months of lactation (postnatal intervention). Therefore, women were randomized into one of the 4 study groups: (1) prenatal BEP and IFA supplementation; (2) postnatal BEP and IFA supplementation; (3) both pre- and postnatal BEP and IFA supplementation; or (4) both pre- and postnatal IFA only supplementation. The study was implemented in 6 rural health center catchment areas in the district of Houndé in the Hauts-Bassins region of Burkina Faso. The study area is characterized by a Sudano-Sahelian climate with a dry season running between September/October and April. Agricultural activity is the main livelihood in the study area, with dominantly cotton and maize production. The habitual diet during pregnancy is nondiverse [28], predominantly based on maize with a complement of leafy vegetables [29]. Hence, among a subsample of MISAME-III women, dietary micronutrient intakes of the base diet (i.e., excluding supplements) during pregnancy were inadequate to cover the Estimated Average Requirements (EARs) and the mean daily energy intake was estimated to be approximately 1,940 kcal at the end of the preharvest season [30]. Furthermore, findings from MISAME-III indicate high burden of adverse birth outcomes, including small-for-gestational age (26%) and low birthweight (10%) [25]. In addition, a survey in the same region found a high prevalence of child stunting (21%) and wasting (7%), whereas the rate of EBF was reported to be 62% [31]. Before MISAME-III commenced, women aged between 15 and 40 years living in the study area were identified through a door-to-door census (n = 10,165). Between October 2019 and December 2020, a network of female village workers trained for the project conducted follow-up home visits of these women every 5 weeks for early identification of pregnancy through self-reported amenorrhea. Women who reported amenorrhea were referred to the nearest health center for a urinary pregnancy test and their first antenatal care consultation. Following a positive test result, women were asked for their informed written consent to participate in the study and randomized into the pre- and postnatal intervention arms. Within 14 days of enrollment, pregnancy was confirmed and gestational age was determined by ultrasonography. Study exclusion criteria were nonconfirmed pregnancies by ultrasound, pregnancies ≥21 completed weeks of gestation at enrollment, or multifetal pregnancies (i.e., postrandomization exclusion). Furthermore, women who planned to leave or deliver outside the study area and women who reported to be allergic to peanuts were not enrolled. The MISAME-III study was approved by the ethics committee of the University Hospital of Gent, Belgium and the ethics committee of Centre Muraz, Burkina Faso. The trial is registered at ClinicalTrials.gov with registration number {"type":"clinical-trial","attrs":{"text":"NCT03533712","term_id":"NCT03533712"}}NCT03533712. A stratified permuted block randomization schedule was used to assign women to either the intervention or control arms of the pre- and postnatal interventions. Randomization blocks were generated per health center in blocks of 8 (i.e., 4 controls and 4 interventions arms for each of the pre- and postnatal intervention) using Stata 15.1 (Statacorp, Texas, USA) by a research analyst not involved in the study. The allocation was coded with the letters A (control) or B (intervention) for the prenatal intervention arms, and X (control) and Y (intervention) for the postnatal intervention arms. The allocation letter codes were concealed in sequentially numbered separate sealed opaque envelops by the study employees who had no direct contact with the study participants. At enrolment during the first antenatal visit, study midwives assigned the women to the study groups by drawing the next sealed envelope with the letter code. The composition and daily dose of the pre- and postnatal BEP intervention supplements were identical; only the duration of supplementation differed. The BEP supplement, an LNS in the form of an energy-dense peanut paste fortified with multiple micronutrients, was selected based on a two-phase formative study conducted in the same study community [32,33]. The complete nutritional composition of a daily dose of the fortified BEP is provided in S1 Table. The product is ready to use, does not require a cold chain, and is highly stable with a long shelf life. A daily dose of the BEP (72 g) provided an energy top up of 393 kcal (36% energy from lipids, 20% from protein, and 32% from carbohydrates) and covered at least the EARs of pregnant women for 11 micronutrients, except calcium, phosphorous, and magnesium [34]. A daily dose of an IFA tablet (Sidhaant Life Sciences, Delhi, India) contained 65 mg of iron [form: FeH2O5S] and 400 μg folic acid [form: C19H19N7O6], in accordance with the standard of care in Burkina Faso. The prenatal intervention provided either a daily BEP supplement plus IFA tablet (intervention) or IFA alone (control) starting from enrollment, during the first/early second trimester (average gestational age of 11.8 weeks) to birth. In the postnatal intervention, BEP was provided to women in the intervention group for 6 months starting from birth. Women in the intervention and control group both received IFA tablets for 6 weeks postpartum. Once enrolled, women were assigned to the study’s village-based project workers, who were responsible for conducting daily home visits to provide the supplements and conduct compliance assessment through direct observation of supplement intake. During these visits, pregnant women were encouraged to attend all antenatal care visits, give birth at the health center, and attend monthly postnatal health center visits. Furthermore, the importance of a healthy diet during pregnancy, as well as the importance of optimal infant feeding practices, was discussed. Supervision of village-based workers was conducted on monthly bases throughout the study using a Lot Quality Assurance Sampling scheme and empty sachet counts to ensure that study participants were visited according to the project protocol. After delivery, mothers were invited to attend monthly growth monitoring sessions organized by study midwives at their nearest health center until all study children reached the age of 6 months. Subsample of children who were born earlier in the study cohort had a continued followed-up through home visits at approximately 9 months (n = 1,345) and approximately 12 months (n = 955) of age. During the growth monitoring sessions, midwives counselled the mother on EBF practices and carried out infant anthropometry measurements. Infant weight was measured with a digital Seca 384 scale to the nearest 10 gm, whereas length was measured with Seca 416 length board to the nearest 1 mm. Head and mid-upper arm circumference (MUAC) measurements were taken to the nearest 1 mm with a Seca 212 measuring tape. Measurements were taken in duplicates, and when there is a large discrepancy between the duplicate measures (>10 mm for length, >200 g for weight, and >5 mm for head and arm circumferences), a third measurement was carried out, and, finally, the average of the two closest measurements was used for analysis. The occurrence of common childhood morbidities (fever, diarrhea, vomiting, runny nose, cough, difficulty of breathing, grunting, and skin lesions) was assessed by asking mothers how many days over the past 7 days their infant experienced each of these morbidity symptoms. Diarrhea was defined as the occurrence of ≥3 liquid or semisolid stools within a day. Furthermore, armpit body temperature measurements were taken by study midwives, and fever was diagnosed as a body temperature ≥37.5°C. At the age of 6 months, infant hemoglobin concentration (g/dL) was determined from a finger prick capillary blood sample using HemoCue Hb 201+ instrument. Information about infant feeding practices was also gathered by maternal 24-hour recall during the monthly anthropometric measurements. Duration of EBF was estimated by asking mothers the date the infant started receiving liquid/solid foods. Minimum dietary diversity for children was assessed at 9 and 12 months of age and achieved when an infant consumed foods and beverages from at least 5 out of 8 defined food groups during the previous day (i.e., (i) breast milk; (ii) grains, roots, tubers, and plantains; (iii) pulses (beans, peas, lentils), nuts, and seeds; (iv) dairy products (milk, infant formula, yogurt, cheese); (v) flesh foods (meat, fish, poultry, organ meats); (vi) eggs; (vii) vitamin A–rich fruits and vegetables; and (viii) other fruits and vegetables) [35]. Infants with acute illnesses and acute malnutrition were referred to the nearest health center. The primary outcome of this study was children’s LAZ at the age of 6 months. Secondary outcomes included a suite of indices of growth and nutritional status at the age of 6 months, such as weight-for-length (WLZ) and weight-for-age (WAZ) z-scores, MUAC and head circumference, prevalence rates of stunting (LAZ <−2 SD), wasting (WLZ <−2 SD), and underweight (WAZ <−2 SD), hemoglobin concentration, and anemia (<11 g/dL). Additional secondary outcomes included the longitudinal prevalence of common childhood morbidities and the incidence of wasting and EBF during the 6 months follow-up. Furthermore, growth trajectories of infants were considered as additional secondary outcomes by evaluating monthly changes in the aforementioned anthropometric indices during the 12 months follow-up period. We previously reported the primary and secondary outcomes at birth [25]. In the absence of dependency between the effects of the pre- and postnatal interventions (see below), a sample size of 588 children per study arm (n = 1,176) was required to detect a difference of 0.18 z-score (with an SD of 1.1) in LAZ between study arms at the age of 6 months with an 80% power and a 95% confidence level [36]. A total of 1,891 pregnant women were required to accommodate for a potential 26% loss of information during the prenatal period due to a combination of abortions, miscarriages, stillbirths, multifetal pregnancies, out-migrations, maternal deaths and incomplete data as detected in the previous MISAME trial [37], and a further anticipated 16% lost-to-follow-up from birth to 6 months postpartum. All analyses in this study were documented a priori in the MISAME-III statistical analysis plan, which was published in the study website on November 3, 2020 (https://www.misame3.ugent.be/resource-files/MISAME-III_SAP_v1_102019.pdf). Analyses were conducted using Stata 17.1 (Statacorp, Texas, USA) and a two-sided statistical significance was considered at alpha <0.05. Descriptive statistics are reported as means ± SD for the continuous variables and as percentages for the nominal variables. Anthropometric indices of LAZ, WLZ, and WAZ were calculated based on the WHO 2006 Child Growth Standards [38]. The analysis strategy to evaluate the effects of BEP supplementation on infant outcomes was established based on the presence or absence of dependency (i.e., statistically significant interaction at p < 0.1) between the effects of the pre- and postnatal interventions [39]. Accordingly, due to the lack of a statistically significant interaction on the study outcomes (e.g., LAZ p-interaction = 0.767), we followed a factorial approach evaluating the two interventions separately. In this approach, the effect of each of the pre- and postnatal interventions on the study outcomes were evaluated independently of the effect of the other intervention (i.e., estimating the effect of one intervention by adjusting for the allocation to the other intervention). Group differences in growth and nutritional status of infants at the age of 6 months (also at 9 and 12 months in a subsample) were estimated by fitting linear regression models for the continuous outcomes and linear probability models with robust variance estimation for the binary outcomes. Poisson regression models with robust variance estimation were fitted to estimate risk ratios between study arms in terms of the number of months infants received EBF and the number of months infants were diagnosed with wasting during the 6 months postpartum. All models contained the health center and randomization block as fixed effects to account for any possible clustering by the study design. Adjusted models further included prespecified known prognostic factors of the study outcomes such as maternal age, height, body mass index (BMI), MUAC, hemoglobin concentration, gestational age, and parity at study enrollment. Our main analysis followed a modified intention-to-treat (mITT) principle in which all infants who had birth anthropometry measurements were included in the analyses at 6 months. For this purpose, we conducted multiple imputations of missing data using chained equations under the “missing-at-random assumption.” Fifty imputations of missing data were conducted to estimate the regression coefficients. We further assessed the robustness of our findings by conducting sensitivity analyses such as a complete cases analysis (i.e., by including only infants with available measurements at 6 months) and a per-protocol analysis (i.e., by including only participants with BEP adherence of at least 75%). As a secondary analysis, we modelled growth trajectories of infants from birth to 6 months, which continued up to 9 and 12 months on a subsample of infants. Mixed effects models with a random intercept for the individual infant and a random slope for the age of the child (in months) were fitted to estimate group differences on average monthly changes in LAZ, WLZ, WAZ, MUAC, and head circumference. We explored the best model fit for our data by assessing different potential relationships of the study outcomes with time, by visual inspection of graphs and comparing model fit indices (AIC and BIC). We applied linear models (WLZ and WAZ), quadratic model (LAZ), and restricted cubic spline models with 6 knots (MUAC and head circumference) according to the model fit. Fixed effects in the model included the main effect of group, time, and group by time interaction, which the later estimates difference between groups on monthly changes in the outcomes. All models included the clustering indicators (i.e., health center and randomization block), whereas adjusted models additionally included the aforementioned known maternal prognostic factors of infant growth. Furthermore, we explored potential intervention effect modification on the primary outcome LAZ at 6 months by prespecified subgroup factors, such as maternal BMI (<18.5 kg/m2), MUAC (<23 cm), hemoglobin (<11 g/dl), height (<155 cm), age (<20 years), completion of primary education, possible and probable prenatal depression (Edinburgh Postnatal Depression Scale ≥10 points and ≥13 points), primiparity, household food insecurity (Household Food Insecurity Access Scale), newborn sex, season of delivery (lean season: June to September), and interpregnancy interval (<18 months). The longitudinal prevalence rates of common childhood morbidities during the 6 months follow-up were calculated using “the total number of days that the outcome was reported” as numerator, and “the total days observed or assessed” as denominator. We used Poisson regression models with robust variance estimation to estimate risk ratios comparing study arms by the occurrence of morbidity outcomes. Finally, to inform on the potential effect of the combination of the pre- and postnatal BEP interventions, we conducted supplementary analysis using a four intervention arms approach evaluating LAZ and stunting at 6 months and LAZ trajectories from birth to 12 months comparing the control group for both intervention periods (IFA/IFA) against each of the prenatal only BEP (BEP/IFA), the postnatal only BEP (IFA/BEP), and the combination of the pre- and postnatal BEP (BEP/BEP) groups. Additional information regarding the ethical, cultural, and scientific considerations specific to inclusivity in global research is included in S1 Supporting Information.