Background
In very preterm infants, adequate early nutritional support is of utmost importance for the quality of growth and neurodevelopmental outcomes in the short-, medium- and long-term. Human milk (HM) has well-known advantages over infant formulas, including for brain development.
Objectives
To determine, in a homogeneous sample of HM-fed very preterm infants, the associations of in-hospital measured protein, energy, and protein-to-energy ratio (PER) intake with weight gain velocity, body composition and head circumference (HC) at term corrected age (CA), and with neurodevelopmental outcome at 18 months CA.
Methods
A cohort study was conducted, being eligible consecutive inborn neonates with less than 33 weeks of gestation, who were exclusively or predominantly HM-fed (own’s mother milk and/or donor human milk).
The study was approved by the Hospital and Medical School ethics committees and is registered at the ISRCTN (ID: 27916681). Informed written consent was obtained from the parents or legal representative of each infant.
Our unit nutrition protocol, based on international and national recommendations, was followed. A standard fortification method with the blinded addition of modular protein and/or fat supplements was used, considering the lowest reported HM protein content and the minimum recommended intake for weight. A mid-infrared analyzer was used to measure the macronutrients content of administered HM. Anthropometry was performed using the recommended techniques. Body composition assessment, using air displacement plethysmography (ADP), was scheduled after discharge, at 40 weeks CA; fat mass percentage (FM%) and fat mass index (FMI) were used as surrogates of adiposity. The assessment of the Mental Developmental Index (MDI) and Psychomotor
xviii
Developmental Index (PDI), using the Bayley Infant Development Scales version II, were scheduled at 18 months CA.
Statistical analysis: required samples of 70 and 75 infants were estimated to detect significant differences in body composition and neurodevelopmental outcomes, respectively. Univariate analysis, using parametric or nonparametric tests as adequate, assessed the associations of cumulative in-hospital protein, energy, and PER intake with weight gain velocity, fat mass (FM), fat-free mass (FFM), FM%, FMI, HC, MDI, and PDI. The same statistical methods were used to assess potential confounding variables, using p<0.10 for inclusion in models. Linear mixed models were used to input missing measured values of own’s mother milk composition and linear multiple regression analyses were used to assess the adjusted effect between independent and dependent variables. A nested case-control analysis was used to determine the associations of lower (≤ -1 z-score) and higher (≥ +1 z-score) adiposity with protein, energy, and PER intake.
Results
Thirty-three infants were included in the cohort, with a median (interquartile range) gestational age of 30 (28-31) weeks and birthweight of 1175 (1010-1408) g. Compared with the 56-excluded formula-fed infants, the 33 infants who completed the study had significantly lower gestational age, lower prevalence of twins and stayed longer in hospital.
Eight hundred and thirty-two pooled HM samples were analyzed, representing 65.0% of the total administered samples. After disclosing the HM macronutrients measurements, it was found that the minimum recommended intake for weight were achieved in 63.6% of infants for protein, 15.2% for energy, and 93.9% for PER. The median daily protein, energy, and PER intake from birth to 35 weeks CA ranged from 2.7-4.2 g/kg, 53.7-109.2 kcal/kg, and 3.4-5.6, respectively.
The mean (standard deviation - SD) in-hospital weight gain velocity was 10.1 (3.8) g/kg/day. At mean (SD) 39.9 (1.9) weeks, body mass was of 2817.6 (504.3) g, FM of 441.5 (184.0) g, FFM of 2376.1 (376.0) g, FM% of 15.3 (4.8), and FMI of 2.0 (0.7).
xix
Neurodevelopment was assessed at 20 months CA. Overall, the mean (SD) score for MDI was 100.2 (11.5) and for PDI 97.4 (8.0). The mean MDI score was below normal in 6.2% infants, normal in 78.1%, and accelerated in 15.6%; the mean PDI score was below normal in 6.2% infants and normal in 93.8%.
In multivariate analysis, only gestational age was associated with low weight gain velocity (p<0.0001). After adjustment for gestational age, only FFM was associated with low protein (p=0.008) and energy (p=0.001) intake. In the nested case-control analysis, in infants with lower adiposity, a FM% ≤ -1 z-score was associated with low energy and protein intake (p=0.050) and a FMI ≤ -1 z-score was associated with low PER intake (p=0.026); in infants with higher adiposity, a FMI ≥ +1 z-score was associated with low energy intake (p<0.0001) and high PER intake (p<0.0001).
In multivariate analysis, it was found that GA and sex were predictors of high HC at term CA, adjusted for protein (p=0.010), energy (p=0.013) and PER intake (p=0.013). In-hospital cumulative protein, energy, and PER intake were neither significantly correlated with any MDI or PDI scores at mean 20 months CA, nor met the defined criteria to enter multivariate analysis.
Conclusions
In this cohort of exclusively or almost exclusively HM-fed very preterm infants, the cumulative in-hospital protein, energy, and PER intake were weakly-to-moderately correlated with weight gain velocity, but not with body composition at term CA in the entire sample. Analyzing infants with extremes of adiposity, those with lower adiposity received significantly lower energy, protein, and PER intake, while infants with higher adiposity received significantly lower energy intake but higher PER intake, compared with the remaining infants. The GA and sex were significant predictors of high HC at term CA, adjusted for protein, energy and PER intake. In-hospital cumulative protein, energy, and PER intake were not significantly correlated with MDI or PDI scores at a mean of 20 months CA. The method of standard fortification with blinded modular protein and fat supplements resulted in insufficient energy and protein intake.
The undersized sample might be insufficient to test the hypothesized associations of macronutrient intake with body composition and neurodevelopmental outcome.
xx
Notwithstanding, our analyses have relied on measured protein and energy HM content and not on its assumed composition, which is a strength of the study.
Key-words: body composition, energy intake, head circumference, human milk, neurodevelopmental outcome, protein intake, very preterm infants, weight gain velocity.