Postnatal growth and body composition up to two years of age of term and preterm infants with placental insufficiency and/or small size for gestational age at birth

Abstract

Background: Growth in the first thousand days can affect lifelong health, yet joint clinical management between prenatal, perinatal and postnatal healthcare providers is often limited. Preterm birth and/or foetal growth restriction (FGR), often presenting as small-for-gestational age (SGA), affect short- and long-term outcomes. Placental insufficiency affects foetal growth and body composition (BC) even in appropriate-for-gestational age (AGA) newborns. Preterm birth, SGA and placental insufficiency are prevalent in South Africa. Aims: This thesis aimed to (a) develop an integrated framework for foetal/infant growth assessment, (b) compare the predictive value of two commonly used preterm infant growth charts, (c) describe one-year growth and its early life predictors in preterm infants with/without SGA (d) describe two-year growth and BC, and early-life predictors thereof, in term infants with/without placental insufficiency. Methods: Framework development (objective a): An interdisciplinary (obstetricians, paediatricians and dietitians) iterative think-tank approach, supported by published literature, was used for framework development. Cohort 1, a preterm historical cohort (objectives b and c), utilised patient records (N=321, 111 SGA, 310 AGA) from the kangaroo mother care follow-up clinic at a tertiary South African hospital. Using anthropometric data up to 12 months, z-scores were calculated with the Fenton Growth Chart (FGC), INTERGROWTH-21ST Newborn Size Standards (IG-NBSS), and INTERGROWTH-21ST Postnatal Growth Standards for Preterm Infants (IG-PPGS). Birth weight z-score (BWZ, FGC vs. IG-NBSS) and weight gain up to 50 weeks postmenstrual age (ΔWZ, FGC vs. IG-PPGS) were compared (Cohen’s Kappa) for association with one-year anthropometry (malnutrition). Cohort 2, the UmbiBaby cohort (objective d), included 81 term-born infants with Doppler-derived umbilical artery resistance index (UmA-RI) assessed at 28-34 weeks’ gestation (55 normal, 26 abnormal UmA-RI). During eight follow-up visits over two years, anthropometric measurements were taken, and fat-free mass (FFM) and fat mass (FM) assessed using deuterium dilution. Z-scores were calculated for FM (FMZ), FFM (FFMZ), FM index (FMIZ) and FFM index (FFMIZ). For both cohorts, z-scores were calculated for weight-for-age (WAZ), length-for-age (LAZ), weight-for-length (WLZ), BMI-for-age (BMIZ), MUAC-for-age (MUACZ) and HC-for-age (HCZ) using WHO Anthro (age-corrected for preterm infants). Underweight (WAZ< 2), stunting (LAZ< 2), wasting (WLZ< 2) and overweight (BMIZ>+2) rates were calculated at last visit. Outcomes were compared for SGA (birth weight-for-GA <10th percentile) vs. AGA (birth weight-for-GA ≥10th - ≤90th percentile) infants (preterm cohort), and normal vs. abnormal UmA-RI (UmbiBaby cohort). Longitudinal growth was characterised using latent class trajectory modelling (LCTM). Multivariable analysis investigated early-life predictors of growth trajectories (both cohorts) and one-year malnutrition (preterm cohort). Results: Objective a: An integrated framework of measurements, indices and indicators used by various health care providers in antenatal, perinatal and postnatal care and research informed all subsequent investigations. Objective b (preterm cohort): FGC and IG-PPGS produced similar ΔWZ (IG-PPGS= 0.26±1.23, FGC= 0.11±1.14; P=0.153) and performed similarly in multivariable analysis. Using ΔWZ<-1, FGC predicted more underweight (42.0% vs. 36.0%), more wasting (43.5% vs. 39.1%), and equal stunting (33.3%), while ΔWZ>+1 on FGC predicted more overweight (57.1% vs. 38.1%). There was substantial agreement between the charts in terms of number of infants with ΔWZ<-1, 1≤ΔWZ≤+1 and ΔWZ>+1 (Κ=0.647) and the association between these classes and malnutrition outcomes (Κ=0.734 to 0.627) Objective c (preterm cohort): At one year, SGA infants had lower anthropometric z-scores and more stunting (34.2% vs. 9.1%; P<0.001), underweight (31.2% vs. 7.2%; P<0.001) and wasting (12.6% vs. 4.3%, P=0.012), despite larger first-year WAZ gains (+0.70±1.30 vs. +0.05±1.30, P<0.001). In multivariable analysis, birth weight z-score (BWZ) predicted one-year undernutrition better than being born SGA. LCTM identified three WAZ and LAZ trajectories (faltering, gradual gain, catch-up), and two WLZ trajectories (faltering, gain). Lower BWZ was associated with WAZ and LAZ catch-up but WLZ faltering. Larger ΔWZ was associated with WAZ catch-up and gradual LAZ gain. WAZ and WLZ faltering were associated with more underweight (49.1%, 22.4%), stunting (45.5%, 23.5%) and wasting (21.8%, 10.3%), while WAZ catch-up and WLZ gain were associated with more overweight (24.4%, 17.6%; all P<0.001). Gradual LAZ gain was associated with the least underweight (2.0%), stunting (2.1%) and wasting (2.1%, all P<0.001). Objective d (UmbiBaby cohort): Infants with abnormal UmA-RI had lower WAZ up to 18 months (mean±SD [ 0.6±0.82 to 0.2±1.12] vs. [0.1±1.18 to 0.6±1.09]; P=0.037-0.017 for measurements at different ages), LAZ up to 14 weeks ([ 1.3±1.25 to 0.9±0.87] vs. [ 0.2±1.04 to 0.1±1.00]; P=0.004-0.021); and FFMZ up to 9 months ([ 0.1±0.82 to 0.7±0.71] vs. [0.7±1.00 to 1.3±0.85]; P=0.002-0.028). LCTM identified three WAZ, LAZ, WLZ, BMIZ, HCZ and FFMZ trajectories, and two MUACZ, FMZ, FMIZ and FFMIZ trajectories. While FMZ and FMIZ trajectories converged around 2 years, FFMZ and FFMIZ trajectories declined. In multivariable analysis, lower BWZ (or SGA) predicted lower WAZ, WLZ, BMIZ, FMZ and FFMZ trajectories, while higher (or abnormal) UmA-RI predicted lower LAZ and FFMZ trajectories. Conclusions and recommendations: The complex associations of prenatal and early postnatal growth with growth and BC outcomes at 1-2 years underscore the importance of an integrated approach to growth in the first thousand days. Careful documentation on the Road-to-Health Booklet of pregnancy conditions (including UmA-RI screening results, when available) and accurately measured birth anthropometry can facilitate interdisciplinary communication. For preterm infants, FGC and IG-PPGN perform similarly, as long as the change in z-score (rather than an absolute z-score) is used to assess growth. Abnormal UmA-RI predicts lower LAZ and FFMZ trajectories up to 24 months, while lower BWZ predicts lower WAZ, LAZ and WLZ at 12 months in preterm infants (accompanied by higher rates of underweight, stunting and wasting), and at 24 months in term infants. Though the cause of the high observed rates of abnormal UmA-RI in the South African population is currently unknown, interventions to support healthy pregnancies and foetal growth should be prioritised. Early and ongoing assessment of WAZ, LAZ and WLZ can guide nutrition interventions for optimal longer-term growth; faltering growth in any anthropometric index should prompt further assessment of health conditions and nutrition status, followed by appropriate caregiver counselling, referral or other interventions.