Milk Temperature Control and Necrotizing Enterocolitis Risk in Extremely Preterm Infants

Although the pathophysiology of NEC is not fully understood, epidemiological studies strongly suggest a multifactorial cause, involving infection and inflammation, premature birth, hypoxic-ischemia, improper feeding, and cold exposure. As for cold exposure, it encompasses inhaling cold air through the respiratory tract, contact with cold environments through the skin, and consuming cold food through the digestive tract. Such exposure increases the risks of cardiovascular hospitalization, temperature-related mortality, allergic diseases including asthma and atopic dermatitis, and neonatal death. A recent study by Lyu et al. suggested that admission hypothermia is associated with an increased incidence of NEC. However, no research has systematically explored how preventing cold exposure, such as through milk feeding via the digestive tract, could potentially reduce the development of NEC among EPIs/VPIs.

Covariates collected using medical records throughout study conduction included maternal age, neonatal sex, birthweight and gestational age at delivery, and weekly weight until discharged. Gestational diabetes mellitus was diagnosed based on a 2-hour 75g three-time-point oral glucose tolerance test was performed at the clinic according to the International Association of Diabetes and Pregnancy Study (IADPSG) criteria adopted by Chinese Obstetrics and Gynecology guidelines between 24 and 28 weeks of gestation. The investigators collected information on clinical diagnosis of hypertension disorders during pregnancy (HDP) from medical records, which was defined by systolic blood pressure (SBP) ≥140 mm Hg and/or diastolic blood pressure (DBP) ≥ 90 mm Hg, with positive proteinuria, at any timepoints between week 20 and delivery.Other pregnancy outcomes or complications were also retrieved from the medical databased, such as the admission of antenatal corticoids, placenta previa, premature rupture of the membrane, and intrahepatic cholestasis of pregnancy that were diagnosed based on national guidelines. As for infants, neonatal critical case score (NCIS) were assessed within 24 hours of admission and divided into non-critical group (>90), critical group (70-90) and extremely critical group (<70) according to the Chinese Pediatric guidelines. Others neonatal medical conditions during hospitalization were also recorded if diagnosed based on international guidelines, namely small-for-gestational age (SGA), respiratory distress syndrome (RDS) and early onset of sepsis (EOS) within three days after birth. In addition, the days to diagnosis of NEC, length of stay during hospitalization, and cases of surgery among NEC cases were collected from thermostatic feeding and standard feeding groups.

Inclusion criteria

• gestational age at delivery between 24+0 and 31+6 weeks
• initiation of enteral nutrition within 24 hours after birth.

Exclusion criteria

• parental decision to not participate
• presence of major congenital anomalies
• requirement for surgical intervention prior to randomization.

The subject would be considered censored if the study ended due to one of the following conditions:

• 1. Death
• 2. Parents' decision to withdraw participation
• 3. Discharge based on doctors' recommendations.

Once enrolled, the neonates received minimal enteral nutrition within the first 24 hours after birth for the initial three days. Following this period, both groups of neonates were fed expressed milk using a pumping system.

the milk temperature was measured using the following steps:

• 1) During the first week after birth, each preterm infant had two identical pumping setups. One pump was used for feeding according to group allocation, while the second, set to zero velocity, was used to measure milk temperature
• 2) The temperature from the second pump, taken at the same time milk entered the stomach from the first pump, was recorded as the final milk temperature
• 3) After the first week, only one pump was used per infant, and the temperature of the last 1 ml of milk was measured as the final temperature.

Both groups were supplemented with human milk fortifier (HMF) for preterm infants weighing 1800 g or less. The neonates were fed 50 ml/kg/day of breast milk, following the Chinese Expert Consensus on the Use of Breast Milk Fortifiers in Premature Infants. The milk was sourced from the infant's mother. Initially, infants received half-strength fortified milk for 3-7 days before transitioning to full-strength fortified milk. HMF was discontinued once the infant's body weight reached the 25th-50th percentile for appropriate-for-gestational-age infants or the 10th percentile for small-for-gestational-age infants, depending on sex and gestational age. The choice of formula, HMF, and the decision to add lactase was at the discretion of the attending neonatologist.

Weaning from thermostatic feeding would occur if any of the following conditions were met:

• 1) If no signs of feeding intolerance (FI) appeared after administering pumped milk within 15 minutes for 1-2 days, thermostatic feeding would transition to standard feeding
• 2) The feeding duration could be gradually reduced in 10-minute increments down to a minimum of 15 minutes. Weaning from standard feeding would occur if no FI was observed after oral feeding within 30 minutes for 1-2 days.

Due to the lack of existing intervention-based evidence for reference, we based our sample size calculation on an RCT that examined the impact of feeding patterns (human milk-based diet vs. bovine milk-based products). According to the published data, the incidence of NEC was significantly lower in infants receiving an exclusively human-milk-based diet (4.5%) compared to those receiving bovine milk-based products (15.9%; p = 0.04), with approximately 10% in difference of incidence. Using previously published data on NEC prevention, we conducted a power analysis for a superiority clinical trial. To achieve 80% power with a 5% Type I error rate (α = 0.05), we estimated a sample size of 100 participants per group (n₁ = 100, n₂ = 100). Considering a 20% loss to follow-up, we adjusted the sample size to 120 participants per group to ensure adequate power to detect an approximately 10% difference in NEC incidence between the intervention and control groups (R version 4.2.2).