Effects of prenatal small-quantity lipid-based nutrient supplements on pregnancy, birth, and infant outcomes: a systematic review and meta-analysis of individual participant data from randomized controlled trials in low- and middle-income countries

Abstract

Background: Undernutrition during pregnancy increases the risk of giving birth to a small vulnerable newborn. Small-quantity lipid-based nutrient supplements (SQ-LNSs) contain both macro- and micronutrients and can help prevent multiple nutritional deficiencies.

Objectives: We examined the effects of SQ-LNSs provided during pregnancy compared with 1) iron and folic acid or standard of care (IFA/SOC) or 2) multiple micronutrient supplements (MMSs) and identified characteristics that modified the estimates of effects of SQ-LNSs on birth outcomes.

Methods: We conducted a 2-stage meta-analysis of individual participant data from 4 randomized controlled trials of SQ-LNSs provided during pregnancy (n = 5273). We generated study-specific and subgroup estimates of SQ-LNS compared with IFA/SOC or MMS and pooled the estimates. In sensitivity analyses, we examined whether the results differed depending on methods for gestational age dating, birth anthropometry, or study design.

Results: SQ-LNSs (compared with IFA/SOC) increased birth weight [mean difference: +49 g; 95% confidence interval (CI): 26, 71 g] and all birth anthropometric z-scores (+0.10-0.13 standard deviation); they reduced risk of low birth weight by 11%, newborn stunting by 17%, newborn wasting by 11%, and small head size by 15%. Only 2 trials compared SQ-LNSs and MMSs; P values for birth outcomes were >0.10 except for head circumference (e.g., z-score for gestational age: +0.11; 95% CI: -0.01, 0.23). Effect estimates for SQ-LNSs compared with IFA/SOC were greater among female infants and, for certain outcomes, among mothers with body mass index <20 kg/m², inflammation, malaria, or household food insecurity. Effect estimates for SQ-LNSs compared with MMSs were greater for certain outcomes among female infants, first-born infants, and mothers <25 y.

Conclusions: SQ-LNSs had positive impacts on multiple outcomes compared to IFA/SOC, but further research directly comparing SQ-LNSs and MMSs is needed. Targeting SQ-LNSs to vulnerable subgroups may be worth considering.

Clinical trial registry: This study was registered at PROSPERO as CRD42021283391.

Keywords: antenatal interventions; balanced energy protein supplementation; fetal growth restriction; infant stunting; infant wasting; low birth weight; maternal nutrition; preterm birth.

FIGURE 1

Study flow diagram. IFA, iron and folic acid supplement; IPD, individual participant data; MMS, multiple micronutrient supplement; RCT, randomized controlled trial; SOC, standard of care; SQ-LNS, small-quantity lipid-based nutrient supplement.

FIGURE 2

Forest plots of the effect of SQ-LNS compared with that of IFA/SOC on (A) birth weight (g), (B) LBW, (C) newborn LAZ, (D) newborn HCZ, (E) newborn BMIZ, and (F) duration of gestation. Individual study estimates were generated from log-binomial regression for binary outcomes and linear regression for continuous outcomes controlling for baseline measure when available and using robust standard errors for cluster-randomized trials. Pooled estimates were generated using inverse-variance weighting with both fixed and random effects. BMIZ, body mass index z-score; CI, confidence interval; HCZ, head circumference-for-age z-score; IFA, iron and folic acid; LAZ, length-for-age z-score; LNS, lipid-based nutrient supplement; LBW, low birth weight; MD, mean difference; RR, relative risk; SOC, standard of care; SQ-LNS, small-quantity lipid-based nutrient supplement.

FIGURE 3

Pooled effects of SQ-LNS compared with those of IFA/SOC on birth outcomes, stratified by selected effect modifiers. (A, B) Sex. (C, D) Material BMI. (E, F) Maternal age. (G, H) Maternal inflammation. (I, J) Food insecurity. Individual study estimates (not shown) were generated from log-binomial regression for binary outcomes and linear regression for continuous outcomes controlling for baseline measures when available and using robust standard errors for cluster-randomized trials. Pooled estimates (shown here) were generated using inverse-variance weighting with fixed effects. CI, confidence interval; HCZ, head circumference-for-age z-score; IFA, iron and folic acid supplement; LNS, lipid-based nutrient supplement; MUAC, mid-upper arm circumference; P-for-interaction, P value for the interaction indicating the difference in effects of SQ-LNS between the 2 levels of the effect modifier; SGA, small-for-gestational age; SOC, standard of care; SQ-LNS, small-quantity lipid-based nutrient supplement.

FIGURE 3

Pooled effects of SQ-LNS compared with those of IFA/SOC on birth outcomes, stratified by selected effect modifiers. (A, B) Sex. (C, D) Material BMI. (E, F) Maternal age. (G, H) Maternal inflammation. (I, J) Food insecurity. Individual study estimates (not shown) were generated from log-binomial regression for binary outcomes and linear regression for continuous outcomes controlling for baseline measures when available and using robust standard errors for cluster-randomized trials. Pooled estimates (shown here) were generated using inverse-variance weighting with fixed effects. CI, confidence interval; HCZ, head circumference-for-age z-score; IFA, iron and folic acid supplement; LNS, lipid-based nutrient supplement; MUAC, mid-upper arm circumference; P-for-interaction, P value for the interaction indicating the difference in effects of SQ-LNS between the 2 levels of the effect modifier; SGA, small-for-gestational age; SOC, standard of care; SQ-LNS, small-quantity lipid-based nutrient supplement.

FIGURE 4

(A, B) Pooled effects of SQ-LNS compared with those of IFA/SOC on anthropometric outcomes at 6 mo of age, stratified by infant sex. Individual study estimates (not shown) were generated from log-binomial regression for binary outcomes and linear regression for continuous outcomes controlling for baseline measures when available and using robust standard errors for cluster-randomized trials. Pooled estimates (shown here) were generated using inverse-variance weighting with fixed effects. CI, confidence interval; HCZ, head circumference-for-age z-score; IFA, iron and folic acid supplement; LNS, lipid-based nutrient supplement; MUAC, mid-upper arm circumference; P-for-interaction, P value for the interaction indicating the difference in effects of SQ-LNS between the 2 levels of the effect modifier; SOC, standard of care; SQ-LNS, small-quantity lipid-based nutrient supplement.

FIGURE 5

Forest plots of the effect of SQ-LNS compared with that of MMS on (A) birth weight (g), (B) newborn LAZ, and (C) newborn HCZ. Individual study estimates were generated from log-binomial regression for binary outcomes and linear regression for continuous outcomes controlling for baseline measures when available and using robust standard errors for cluster-randomized trials. Pooled estimates were generated using inverse-variance weighting with both fixed and random effects. CI, confidence interval; HCZ, head circumference-for-age z-score; LAZ, length-for-age z-score; LNS, lipid-based nutrient supplement; MD, mean difference; MMS, multiple micronutrient supplement; SQ-LNS, small-quantity lipid-based nutrient supplement.