320 - Patterns of Free Amino Acids in Early Human Milk Relate to Growth at Discharge in Term and Preterm Infants

Publication Number: 3309.320

Meaghan Molloy, Yale School of Medicine, Middletown, CT, United States; Laura D. Brown, University Of Colorado Anschutz Medical Campus, Aurora, CO, United States; Veronika Shabanova, Yale School of Medicine, New Haven, CT, United States; Clare Keeney, Frank H. Netter MD School of Medicine at Quinnipiac University, North Haven, CT, United States; Catherine O. Buck, Yale School of Medicine, New Haven, CT, United States

Background: Human milk is a bioactive substance that influences infant growth and metabolic health. The composition reflects the interplay between the dynamic needs of the infant and the maternal metabolic state. Of the components in human milk, free amino acid (FAA) concentrations appear to be tightly regulated. The influence of human milk composition on the long-term metabolic health of the infant is currently unknown.

Objective: To examine FAA patterns in human milk from women with diabetes (DM) and/or obesity (OB) and their relationship to infant adiposity measures.

Design/Methods: Single-center prospective cohort study of term (> 37 weeks) and preterm (30 to 36 weeks) infants enrolled after birth. FAA in human milk samples were quantified using high-performance liquid chromatography. Infant adiposity measures were obtained prior to discharge. Exploratory factor analysis was performed to identify underlying patterns among FAA concentrations, compare patterns by metabolic health in pregnancy exposure groups, and using quantile regression, relate patterns with infant adiposity measures.

Results: Among 150 infants in the initial cohort, 48 with human milk samples were included. 56% were born to mothers with diabetes and/or obesity and 67% were preterm (Table 1). Milk samples were collected at a median of 3.3 days. Exploratory factor analysis revealed two factors. Isoleucine, leucine, methionine, phenylalanine, proline, tyrosine, and valine loaded on factor 1. Glutamine and alanine loaded on factor 2. Factor 1 scores were higher for the DM/OB group (p=0.04). Higher factor 1 scores were associated with greater weight z-score, including in a sensitivity analysis limited to preterm infants (Tables 2 and 3). Factor 2 scores were inversely associated with length z-score. The slope between factor 1 and head/abdominal circumference and between factor 2 and triceps skin fold thickness differed by maternal DM/OB exposure (interaction term DM/OB*factor; p< 0.15).

Conclusion(s): In this cohort of term and preterm infants, patterns of FAA in human milk differed among women with and without DM and/or OB in pregnancy. The branch chain amino acids, methionine, proline, phenylalanine, and valine were positively associated, while glutamine and alanine were inversely associated with adiposity measures at discharge, with some effect modification by DM/OB exposure. These results provide direction for future studies to investigate how human milk composition influences infant growth and development longitudinally.

Table 1. Patient Characteristics by exposure to diabetes mellitus and/or obesity in pregnancy (N=48)
Data presented as N (%) or median (IQR)
NICU, Neonatal Intensive Care Unit
Diabetes Mellitus (includes type 1/type 2 pre-pregnancy diabetes or gestational diabetes)
Obesity (defined as pre-pregnancy or first-trimester body mass index ≥30 kg/m2)
Percent maternal milk exposure calculated by feeds with all maternal milk and direct breast feeds divided by all enteral feeds during their inpatient stay.
No meaningful differences observed between exposure groups when compared using chi-square or Wilcoxon rank-sum tests.

Table 2. Quantile regression† estimates of median infant adiposity measures at discharge (N=48). Beta values represent median change in adiposity outcomes at discharge for each unit change in human milk factor scores created from free amino acids.
† Models adjusted for postmenstrual age at time of last growth measure prior to discharge, and interaction term (factor score x maternal diabetes/obesity) if p-value for interaction term < 0.15.
SFT (skin fold thickness), DM (diabetes mellitus), OB (obesity)
Z-score calculated using Fenton growth reference (2013)
** p-value <0.05, * p-value <0.15
Factor 1 contains isoleucine, leucine, methionine, phenylalanine, proline, tyrosine, and valine (factor loadings >0.8)
Factor 2 contains glutamine and alanine (factor loadings >0.8)


Table 3. Sensitivity analysis utilizing quantile regression† of infant adiposity measures at discharge among preterm infants only (N=32). Beta values represent median change in adiposity outcomes at discharge for each unit change in human milk factor scores created from free amino acids.
† Models adjusted for postmenstrual age at time of last growth measure prior to discharge, and interaction term (factor score x maternal diabetes/obesity) if p-value for interaction term < 0.15.
SFT (skin fold thickness), DM (diabetes mellitus), OB (obesity)
Z-score calculated using Fenton growth reference (2013)
** p-value <0.05, * p-value <0.15
Factor 1 contains isoleucine, leucine, methionine, phenylalanine, proline, tyrosine, and valine (factor loadings >0.8)
Factor 2 contains glutamine and alanine (factor loadings >0.8)