Study Design and Patients
We retrospectively analyzed the brain Magnetic Resonance Imaging (MRI) - acquired at TEA - of a cohort of preterm infants with gestational age (GA) ≤28 weeks and/or birthweight (BW) <1000 grams, admitted to the neonatal intensive care unit of the Fondazione Policlinico Universitario Agostino Gemelli, IRCCS in Rome, Italy, between December 1, 2015 and May 31, 2019. The enrolled infants were a cohort of subjects randomly assigned in our previous trial to receive a MLE (SMOFlipid 20%, Fresenius Kabi, Italy) or SLE (Intralipid 20%, Fresenius Kabi, Italy).8
Infants with congenital or chromosomal malformations, inborn errors of metabolism, and/or congenital infections of the central nervous system, and with major brain pathology such as intraventricular hemorrhage (IVH)9 grade 3 or higher, periventricular leukomalacia (PVL)10 and/or parenchymal hemorrhagic infarction on cerebral ultrasound were excluded.
The study protocol was approved by Institutional Review Board (Prot. ID 2759). Written informed consent was obtained from the parents.
Data collection
Baseline data based on sex, birthweight (BW) and BW z-score, HC and HC z-score, length and length z-score, GA (based on the last menstrual period, ultrasound in early pregnancy, and postnatal physical examination), the number of prenatal and postnatal steroid course were recorded. Furthermore, BW, HC, length, and their z-scores at time of MRI scan were included.
Clinical outcomes such as respiratory distress syndrome (RDS); pharmacologically treated patent ductus arteriosus (PDA), bronchopulmonary dysplasia (BPD)11, necrotizing enterocolitis (NEC)12, retinopathy of prematurity (ROP)13, incidences of late-onset sepsis and length of hospital stay were recorded for each infant.
Nutritional intake
The composition of the two LEs are detailed in Table 1. Daily total parenteral and enteral protein, fat, and caloric intakes for the first 28 days of life were available for all infants.
Intravenous lipids were administered at a dose of 1.5 g/Kg/die within first 24 hours of life and were gradually increased by 0.5 g/Kg/die until the dose of 3 g/Kg/die was reached within the first week of life. According to our local protocol, the parenteral lipid intake was reduced by 25% to 50% when plasma triglycerides concentrations were between 265-442 mg/dL and was stopped when they exceeded 442 mg/dL. Minimal enteral feeding was started on the first day of life. Enteral feeding was increased by 20-30 ml/Kg/day. All infants received human milk (own mother’s milk or donor milk); human milk was enriched with fortifier when enteral feeding of 100 ml/Kg/day was reached.
Magnetic Resonance Imaging
All MRI investigations were performed according to a standard protocol on a 1.5 Tesla MRI system (Philips Healthcare) with a standard head coil.
Infants were sedated using Sevoflurane in O2 and administered by vaporization or intravenous midazolam before the examination as the clinical local protocol.
Heart rate, transcutaneous oxygen saturation and respiration rate were continuously monitored, and an expert anesthetist was present throughout the entire examination.
The scanning protocol included T2 and T1-weighted imaging. Parameters of the scanning protocol included: axial three-dimensional (3D) T1-weighted image, coronal 3D T1-weighted image, axial T2-weighted image, and coronal T2-weighted image.
We used axial T2-weighted MRI to measure Supratentorial Volume (SuV), Cerebellar Volume (CeV), Brainstem Volume (BsV), Total Brain Volume (TBV) and Cerebellar Volume corrected for TBV.
SuV was defined as the volumes of the cerebral hemispheres, including cortical gray matter, basal ganglia, and white matter, without the lateral ventricles, the III ventricle and the cerebrospinal fluid (CSF) spaces. BsV was defined as the volumes of the midbrain, pons, and medulla oblongata without the CSF spaces, the IV ventricle and the Silvian acqueduct. CeV was defined as the volume of the entire cerebellum (vermis and hemispheres), without the CSF spaces and the IV ventricle. TBV was defined as the volume of all brain structures, that is, intracranial volume without the volume of the ventricles and CSF (Figure 1). All measures were corrected for PMA at the time of scan with linear regression analysis.
Measurements were performed twice, with an automatic segmentation method and then with a semiautomatic segmentation open-source program called ITK-SNAP (version 3.8.0, University of Pennsylvania, Philadelphia, PA, USA), using the Cavalieri principle14,15
Two neonatologists with experience in the field of neonatal neuroimaging (MLT and CC) and a neuroradiologist (GDA) evaluated the MRI scans for brain injury according to Kidokoro et al.16
Primary and secondary outcomes
The primary outcome of the study was the CeV valued on MRI acquired at TEA.
Secondary outcomes included TBV, SuV, BsV and Cerebellar volume corrected for TBV evaluated on MRI acquired at TEA.
Sample size and Statistical Analyses
The sample size calculation was based on the literature data according to which premature infants with a mean GA of 28 weeks, and without severe IVH (grade 3-4), have a mean CeV measured with MRI equal to 18.3 ± 3.2 cm3.17 Expecting an 18% gain in CeV (equal to 3.3 cm3) in premature infants who received the MLE, a sample of at least 16 infants per arm is required, considered an alpha error of 0.05 and a study power of 80%.18
Data were analyzed with Statistical software SPSS for Windows version 25.0 (SPSS, Inc, Chicago, IL, USA). Continuous variables are expressed as mean and standard deviation or as median and interquartile range, whereas categorical variables are expressed as numbers and percentages. Continuous variables were first tested for normal distribution by using the Shapiro-wilk test. Categorical variables were compared using the Chi-squared test or the Fisher exact test. Continuous variables were compared using T student test for independent samples (to compare data with normal distribution) or Mann-Whitney U test (to compare the others continuous data).
Because the volumes increase with PMA, we corrected the volume with linear regression analysis according to the following equation: corrected volume = measured volume + the slope x (40 – postmenstrual age on MR imaging). The level of significance was set at p < .05.