Effect of Target Fortified Human Milk on the Growth Quality of Very Preterm Infants (TargetFort)

Adiposity, Body Composition, Breast Milk Collection, Feeding Patterns, Growth, Nutrient Deficiency, Prematurity

It is reported that the standard method for fortification of human milk (HM) overestimates the energy and protein densities of HM (Macedo MHNP 2018), thus originating infant undernutrition (Macedo AJP 2018). The target fortification, based on analysis of HM composition, is considered the gold-standard method (Rochow 2015, McLeod 2016). This observational mixed cohort study aims to assess if very preterm infants fed HM with target fortification have greater growth during hospital stay and better body composition at term post-menstrual age (PMA), than those fed HM with standard fortification.

Study design: observational mixed cohort study, comparing growth and body composition in a contemporary cohort of very preterm infants fed HM with target fortification, the currently adopted clinical practice, with a historical cohort of very preterm infants fed HM with standard fortification (Macedo AJP 2018). Study periods: historical cohort from 1 February 2014 to 28 February 2015; contemporary cohort: start February 2020, estimated recruitment period of 16 months; Settings: Neonatal Care Unit (NICU) and Human Milk Bank at Maternidade Dr. Alfredo da Costa and Nutrition Laboratory at Hospital Dona Estefânia, Centro Hospitalar Universitário de Lisboa Central. Product to be analyzed: HM, including mother's own milk (MOM) and donor's milk (DHM). Donor HM is pasteurized using the Holder method (Peila 2016) which is adopted by the Human Milk Bank of Maternidade Dr. Alfredo da Costa (Macedo MHNP 2018). Demographic variables: Gestational age, sex, singleton or twin, birth weight, small-, appropriate- or large-for-gestational age (\<3rd percentile, ≥3rd percentile and ≤97th percentile, \>97th percentile, respectively) (Fenton 2013), severity index (SNAPPE II) (Richardson 2001), use of prenatal corticosteroids, diagnosis of late sepsis (Modi 2009), necrotizing enterocolitis (grade ≥ 3) (Bell 1971), intraperiventricular hemorrhage (grade ≥ 3) (Papile 1978), multicystic periventricular leukomalacia (de Vries 1992), and chronic lung disease (Becker 1984). Method of collection and analysis of HM: as described in a previous study (Macedo MHNP 2018), in order to minimize daily variability of breast milk composition, mothers are asked to save milk collected through 24 hours in the same container. The HM composition (MOM and DHM) are analyzed using the Miris human milk analyzer (Miris AB, Uppsala, Sweden), following the method described in a previous study (Macedo MHNP 2018). The composition is expressed in densities: Kcal/dL of energy and g/dl of fat, raw and true protein, carbohydrates and ashes. Energy and macronutrient composition of the multi-component HM fortifier and modular protein and fat supplements. For this purpose, an Excel program to calculate modular protein and fat supplements to be added to fortified HM was developed and registered (Nona R, Cardoso M, Portuguese Directorate of Intellectual Property Services, IGAC-DSPI, nº 480/2020, 26 February 2020). Daily intakes of energy (Kcal/kg), protein (g/kg) and protein:energy ratio (P:E), based on administered volume of milk (ml/kg). Anthropometric and body composition measurements in infants: According to the previously described method (Macedo AJP 2018), during the hospital stay, the same observer (MMC) measures daily the body weight (allowing the calculation of weight gain velocity), and weekly the length and the head circumference. Within the first week after discharge, body composition will be assessed using displacement plethysmography, to evaluate fat mass (FM), fat free mass (FFM), percentage of FM (%FM), percentage of FFM (%FFM) and FM index (FMI), as described in a previous study (Macedo AJP 2018). Both FMI and %FM are used as indicators of adiposity. Estimate of sample size: The study sample size was calculated to detect a difference of 2 g/kg/day in growth velocity with a standard deviation 2.6 (Macedo, 2018; Tremblay, 2017) for normally distributed variables, a significance level of 0.05, and an 80% power; thus, a required sample of 67 infants (n1=33; n2+20%=34) was estimated.

Portugal