Circadian Variations in Energy and Macronutrient Content of Preterm Human Milk

Abstract

Background: Breast milk has long been named as liquid gold. Preterm infants particularly benefit from the immunoprotective and nutritional properties of breast milk early in life. Human milk analysis has been proven to be a potentially valuable tool to individualize the preterm infant’s diet, but the superior collection of 24-h milk samples might not be feasible in every clinical practise. An understanding of possible circadian variations in the nutrient content of preterm milk could provide knowledge in the nutritional care of preterm infants. Objectives: The primary objective of the study was to compare the content of energy and macronutrients in preterm human milk from a pooled 24-h milk sample with a single morning sample at one week postpartum. Subjects and methods: Milk samples were collected from 30 mothers of preterm infants admitted to the Neonatal Intensive Care Unit (NICU) at St. Olavs Hospital. The majority of infants were born with low birth weight (LBW<2500 g). The mother expressed her milk every three to four hour during a 24- h period at one week postpartum. A subgroup (n=6) collected 24-h and morning milk samples at three weeks postpartum. The content of energy (kcal/dL), protein (g/dL), carbohydrate (g/dL) and fat (g/dL) in preterm milk was analysed using a mid-infrared (MIR) human milk analyser (HMA). The nutritional values were compared with current reference values for energy and macronutrient content in preterm human milk. Results: The content of energy and protein were lower in the single morning samples compared to pooled 24-h samples at one week postpartum, 3 kcal/dL (p=0.016) and 0.17 g/dL (p<0.001), respectively. Morning samples showed a tendency to underestimate fat concentration in 24-h samples (0.25 g/dL, p=0.054). Limits of agreement were wide for the energy, protein and fat, but not for carbohydrate, illustrated with Bland-Altman plots. Protein concentration decreased with 29% (p<0.001) from the 1st to 3rd week postpartum in the subgroup. Wide mother-to-mother variations in nutrient content in preterm milk were demonstrated at both one and three weeks postpartum. The present median energy and macronutrient content was greater compared to current reference values. Conclusion: Single morning samples underestimated mean energy and protein concentration compared with 24-h pooled samples of preterm human milk. Therefore, spotted morning milk samples could not be used to provide accurate values of preterm human milk composition. Effort should be made to collect 24-h milk samples prior to human milk analyses for individualized nutritional care. Based on our findings, morning spot samples may provide better estimates of the true milk composition compared to the use of current reference values.