Predictive model for feeding intolerance in neonates with hypoxic ischemic encephalopathy during therapeutic hypothermia

Abstract

To construct a nomogram for feeding intolerance (FI) during therapeutic hypothermia (TH) in neonates with hypoxic-ischemic encephalopathy (HIE). 179 neonates with HIE were recruited between March 2017 and July 2023 and clinical data subjected to least absolute shrinkage and selection operator (LASSO) regression and multivariable logistic regression analysis. A predictive model was constructed and verified by receiver operating characteristic (ROC) curve analysis, calibration plots and decision curve analysis (DCA). Neonatal infection, 5-min Apgar score, hypoglycemia, time of enteral nutrition initiation, initial enteral feeding volume (15–30 mL/kg/day) and rate of feeding advancement (1–5 mL/kg/day) were found to be independent predictors for FI. Earlier initiation, larger initial volume and rapid feeding progression increased FI risk and slow advancement was protective. ROC analysis gave an area under the curve (AUC) of 0.83 (95% CI: 0.77–0.89) and internal verification concordance index (C-index) was 0.829. DCA showed a favorable net clinical benefit for the FI predictive model. The predictive model may identify the causes of FI at an early stage and inform clinical decisions.

Introduction

Therapeutic hypothermia (TH) is increasingly used in cases of neonatal hypoxic-ischemic encephalopathy (HIE) and appears beneficial for neurological development and mortality rate¹. However, neonatal feeding intolerance (FI) has received little attention although it has been observed in about one-third of critically ill patients². Asphyxia-induced hypoxia and ischemia may cause blood flow redistribution which affects the gastrointestinal tract and results in FI, characterized by gastric retention, bloating and regurgitation. TH treatment reduces inflammation and metabolic damage, attenuating systemic oxidative stress and protecting the gastrointestinal tract^3,4.

The reduction in intestinal blood flow produced by TH may increase the risk of necrotizing enterocolitis (NEC) and the delay or halting of feeding is a response to reduce clinical risk and ensure patient safety^5,6. However, enteral feeding during TH may help to maintain intestinal function and microbial diversity⁷. Minimal enteral feeding (MEN) has been shown to shorten hospital stay and reduce reliance on parenteral nutrition following TH⁸. Neonates who received MEN tended to transition more quickly to complete enteral nutrition and had a lower risk of NEC^8,9 which is considered the most severe gastrointestinal complication of TH. Gale et al. reported a low incidence (< 5/3,236) of severe NEC following TH but concerns remain over the safety of initiating early enteral nutrition and the relationship between TH and FI remains uncertain. For example, Hu et al.¹¹ found an approximately 28% incidence of FI during TH in a single randomized clinical trial with a small sample size of 92 neonates and using stringent screening standards.

FI has been linked to sepsis, NEC, malnutrition, metabolic abnormalities and aspiration pneumonia and presents a potential risk to neurodevelopmental processes and growth throughout the late neonatal period^12,13. However, FI is normally identified from the clinical characteristics of > 50% gastric retention¹⁴, abdominal distension¹², emesis¹⁵ and diarrhea¹⁶ despite being affected by many factors with no single factor accurately predicting its extent in a given individual. Birth weight, gestational age¹⁷, disease factors^18,19 and oxygenation abnormalities²⁰ are all considered to have an impact. The early identification of FI risk during TH in neonates with HIE is critical, enabling a proactive approach to improve nutrition, shorten hospital stay and improve long-term quality of life. The current study aimed to identify FI characteristics and potential predictors in neonates. A predictive model was developed to inform clinicians and aid the estimation of FI likelihood, enabling individualized decisions on nutrition.

Materials and methods

Study design and patient population

Clinical data, including demographics, perinatal factors, laboratory results, feeding practices and FI episodes were collated from medical records of neonates treated at Chongqing Medical University Children’s Hospital between March 2017 and July 2023. FI was assessed during therapeutic hypothermia and rewarming phase (72–96 h). Inclusion criteria were: (1) Neonates diagnosed with HIE according to the Neonatal Group of Science guidelines, a branch of the Chinese Medical Association, and treated with TH (http://guide.medive.cn/); (2) admission to the neonatology department within 12 h of birth; (3) gestational age ≥ 35 weeks and birth weight ≥ 1800 g; (4) administration of enteral feeding during TH. Exclusion criteria were: (1) congenital gastrointestinal abnormalities; (2) congenital genetic metabolic disorders; (3) death within 72 h of birth or treatment abandonment for personal reasons; (4) no administration of enteral feeding during TH.

Patient flow chart

A total of 222 neonates with HIE who received TH and enteral feeding were initially enrolled. Six were excluded because the duration of TH was less than 48 h; 27 due to missing data when ≥ 9/46 clinical variables were unavailable and 10 because TH was initiated > 12 h after birth. 179 neonates were included in the final analysis (Fig. 1).

Fig. 1 [Images not available. See PDF.]

Patient inclusion flowchart

Data collection

Birth data, maternal obstetrical characteristics and aEEG were recorded and clinical definitions were according to Practical Neonatology (5th ed., Shao et al.)²¹. (1) HIE severity was determined by clinical assessment using the Sarnat staging system²². (2) Advanced maternal age was defined as ≥ 35 years. (3) Placental abnormalities included placental abruption, placental calcification, placenta previa and velamentous placenta. (4) Neonatal hypokalemia was diagnosed if serum potassium fell below 3.5 mmol/L within the first 72 h of admission. (5) Neonatal hyperlactatemia was diagnosed if arterial blood gas lactate was > 4.8 mmol/L on admission. (6) Neonatal hypoglycemia was defined as plasma glucose concentration < 40 mg/dL (2.2 mmol/L) in the first 48 h of life or < 45 mg/dL (2.5 mmol/L) after 48 h.

aEEG data were reviewed by neuroelectrophysiologists and neonatal aEEG activity classified according to upper and lower boundary amplitudes and continuity using the five-category method proposed by Hellström-Westas et al.²³.

Statistical power was tested by the a posteriori method of sample size calculation and power analysis was performed on neonatal infections with the lowest significance level (OR = 0.36, at the level of α = 0.05). The power value calculated for 179 samples was 0.9, above the threshold of 0.80, indicating that statistical power met the criteria and that the sample size met the requirements.

Feeding protocol

• Time of enteral nutrition initiation: All infants underwent systematic gastrointestinal tolerance assessments, comprising daily abdominal circumference measurement and bowel sound auscultation, to determine readiness for enteral feeding.

• Types of milk: Maternal breast milk was the first choice and hydrolyzed formula, amino acid-based formula, preterm formula or 5% glucose solution was used, contingent on the infant’s gastrointestinal tolerance, if maternal breast milk was unavailable.

• Initial enteral feeding volume: Enteral feeds were initiated at 10–20 mL/kg/day or 15–30 mL/kg/day, depending on the infant’s birth weight and clinical condition²⁴.

• Rate of feeding advancement: Feeding volume was increased by approximately 1–5 mL/kg/day or 6–10 mL/kg/day, depending on individual tolerance. FI was routinely assessed by monitoring gastric residuals, vomiting and abdominal distension.

Definition

A clear definition of FI has yet to be established. The presence of at least one of the following criteria is generally considered to constitute FI²: (1) gastric residual volume > 50% of the previous feed; (2) emesis of gastric contents more than twice in 24 h; (3) diarrhea more than six times per day with loose stools; (4) abdominal distension, defined as an increase in abdominal circumference of > 2 cm from baseline; (5) hematochezia with observed rectal bleeding; (6) feeding plan failure, including decreased, delayed or interrupted enteral feedings.

Statistical analysis

Continuous variables are reported as mean ± SD or median (IQR) and categorical variables as frequencies and percentages. Inter-group differences were analyzed by Mann-Whitney U test, t-test or chi-square test. Least absolute shrinkage and selection operator (LASSO) regression was performed to select optimal predictive factors. Cross-validation was performed to determine the optimal value of λ. Multivariate logistic regression analysis was conducted to identify statistically significant predictors. A nomogram was developed internally and validated using 1,000 bootstrap resamples. The C-index was calculated and a calibration curve plotted to evaluate calibration capacity and assess discriminatory performance. Model performance was assessed by receiver operating characteristic (ROC) curve and decision curve analysis (DCA) used to assess net benefit to neonates. A value of p < 0.05 was considered to indicate statistical significance.

Results

Participants

The complete medical histories of 179 neonates with HIE undergoing TH between March 2017 and July 2023 were evaluated and participants divided into FI (n = 63, 35.2%) and non-FI (n = 106) groups. Mean ± SD weight was 3228.86 ± 450.25 g, 118 (65.6%) were male and 61 (36.3%) were female. Median gestational age was 39 weeks and median time for TH initiation was 3 h (Table 1).

Table 1. Characteristics of HIE neonates during TH.

Variables		Overall (n = 179)	Feeding intolerance (n = 63)
HIE grade, n (%)	Severe	67 (37.43%)	23 (36.51%)
	Moderate	103 (57.54%)	36 (57.14%)
	Mild	9 (5.03%)	4 (6.35%)
Basic information
Sex, n (%)	Male	118 (65.92%)	41 (65.08%)
	Female	61 (34.08%)	22 (34.92%)
Gestational age (weeks) (median [IQR])		39.0[38.0,40.0]	39.0 [37.5, 40.0]
Time of TH initiation (h) (median [IQR])		3.00[2.00, 4.00]	3.00 [2.00, 4.50]
Birth weight (g) (mean ± SD)		3228.86 ± 450.25	3214.44 ± 424.86
1-min Apgar score (mean ± SD)		6.02 ± 2.10	5.88 ± 2.25
5-min Apgar score (mean ± SD)		8.09 ± 1.88	8.05 ± 2.12
10-min Apgar score (mean ± SD)		8.95 ± 1.32	8.87 ± 1.46
Maternal and perinatal data
Advanced maternal age (≥ 35years), n (%)	Yes	24 (13.41%)	9 (14.29%)
Gestational hypertension, n (%)	Yes	19 (10.61%)	3 (4.76%)
Gestational diabetes mellitus, n (%)	Yes	34 (18.99%)	10 (15.87%)
Mode of delivery, n (%)	Cesarean section	67 (37.43%)	27 (42.86%)
Premature rupture of membranes, n (%)	Yes	18 (10.06%)	10 (15.87%)
Meconium-stained amniotic fluid, n (%)	Yes	53 (29.61%)	17 (26.98%)
Intrauterine distress, n (%)	Yes	69 (38.55%)	24 (38.10%)
Nuchal cord, n (%)	Yes	49 (27.37%)	15 (23.81%)
Placental abnormalities, n (%)	Yes	14 (7.82%)	8 (12.70%)
Bedside monitoring equipment
aEEG, n (%)	Moderate-to-severe	119 (66.48%)	39 (61.90%)
Laboratory examination
Hyperlactatemia, n (%)	Yes	113 (63.13%)	35 (55.56%)
Hypokalemia, n (%)	Yes	15 (8.38%)	6 (9.52%)
Other
Use of sedatives, n (%)	Yes	137 (76.54%)	48 (76.19%)
Time of enteral nutrition initiation (h) (median [IQR])		17.00 [4.50, 44.50]	8.00 [3.50, 26.50]
Ventilator support, n (%)	Yes	45 (25.14%)	10 (15.87%)
Diagnosis
Neonatal brain injury, n (%)	Yes	117 (65.36%)	40 (63.49%)
Neonatal infections, n (%)	Yes	73 (40.78%)	20 (31.75%)
Neonatal hyperbilirubinemia, n (%)	Yes	37 (20.67%)	16 (25.40%)
Neonatal pneumonia, n (%)	Yes	71 (39.66%)	25 (39.68%)
Respiratory failure, n (%)	Yes	43 (24.02%)	12 (19.05%)
Neonatal intracranial, n (%)	Yes	39 (21.79%)	12 (19.05%)
PPHN, n (%)	Yes	29 (16.20%)	15 (23.81%)
NRDS, n (%)	Yes	3 (1.68%)	3 (4.76%)
Neonatal shock, n (%)	Yes	4 (2.23%)	0 (0.00%)
Pulmonary hemorrhage, n (%)	Yes	3 (1.68%)	3 (4.76%)

PPHN persistent pulmonary hypertension of the newborn, NRDS respiratory distress syndrome.

Identification of independent risk factors

A preliminary screening by LASSO regression reduced the 46 candidate features initially identified to 18 potential predictors (Fig. 2A). The initial reduction was based on non-zero coefficients under the optimal penalty parameter, λ, determined by cross-validation (Fig. 2B). Multivariate logistic regression analysis of the 18 variables with backward stepwise selection allowed those variables without statistically significant associations with FI to be removed. Neonatal infection, neonatal hyperbilirubinemia, NRDS, pulmonary hemorrhage, ventilator support, 5-min Apgar score, gestational hypertension, premature rupture of membranes, hypoglycemia, hypokalemia, time of enteral nutrition initiation, infant formula, hydrolyzed formula, amino acid formula, preterm infant formula, initial enteral feeding volume (15–30 mL/kg/day), rate of feeding advancement (1–5 mL/kg/day) and rate of feeding advancement (6–10 mL/kg/day) were included.

Fig. 2 [Images not available. See PDF.]

Selection of characteristics by LASSO regression. (A) LASSO coefficient profiles of 46 candidate variables. Each curve represents variable coefficient trajectory as the penalty parameter lambda (λ) increases. (B) Ten-fold cross-validation was performed to determine the optimal value of λ using minimum criteria. The vertical dashed line indicates λ value resulting in minimum mean cross-validated error. Variables with non-zero coefficients at this point were selected for model construction.

Development of an individualized prediction model

Appropriate adjustments were made considering clinical expertise and statistical relevance. Variables were selected by LASSO regression to enable variable selection via coefficient shrinkage controlled by the penalty parameter, λ. Ten variables with non-zero coefficients were identified and included in multivariate logistic regression analysis which highlighted five independent predictors of FI (Table 2): neonatal infection (p = 0.018, 95% CI: 0.15–0.82), 5-min Apgar score (p = 0.007, 95% CI: 0.59–0.92), hypoglycemia (p = 0.005, 95% CI: 1.8–23.18), time of enteral nutrition initiation (p < 0.001, 95% CI: 0.93–0.98), initial enteral feeding volume (15–30 mL/kg/day) (p = 0.003, 95% CI: 1.54–8.27) and rate of feeding advancement (1–5 mL kg/day) (p = 0.001, 95% CI: 0.07–0.51). A nomogram was developed for risk prediction by incorporating these variables into the forecasting model (Fig. 3).

Table 2. Multifactorial logistic regressors of feeding intolerance.

Variables	β	SE	P	OR (95% CI)
Neonatal infection	− 1.018	0.432	0.018*	0.36 (0.15–0.82)
Neonatal hyperbilirubinemia	0.903	0.509	0.076	2.47 (0.91–6.84)
Ventilator support	− 0.927	0.496	0.062	0.40 (0.14–1.01)
5-min Apgar score	− 0.304	0.112	0.007*	0.74 (0.59–0.92)
Gestational hypertension	− 1.309	0.773	0.090	0.27 (0.05–1.11)
Premature rupture of membranes	1.168	0.610	0.056	3.22 (0.98–11.01)
Hypoglycemia	1.818	0.645	0.005*	6.16 (1.80–23.18)
Hypokalemia	1.039	0.701	0.138	2.83 (0.70–11.34)
Time of enteral nutrition initiation	− 0.045	0.011	0.000*	0.96 (0.93–0.98)
Initial enteral feeding volume (15–30 mL/kg/day)	1.252	0.427	0.003*	3.50 (1.54–8.27)
Rate of feeding advancement (1–5 mL kg/day)	− 1.614	0.503	0.001*	0.20 (0.07–0.51)

Fig. 3 [Images not available. See PDF.]

Nomogram for FI prediction in neonates with HIE.

Each predictor corresponds to a row in the nomogram. A patient’s score is determined by locating the value for each variable and drawing a vertical line to the “Points” axis to assign a number of points. Summing of point values for all predictors gives the total score which may be mapped to the “Risk” axis to estimate the probability of FI.

Performance of the model

The logistic regression model had an area under the curve (AUC) value after ROC curve analysis of 0.83 (95% CI: 0.77–0.89), indicating good discrimination. Model sensitivity was 0.762 and specificity was 0.750 (Fig. 4). Optimally corrected C-index was calculated to be 0.829 from 1,000 bootstrap resamples, indicating good discrimination. A calibration plot (Fig. 5) showed close agreement between predicted and observed FI probabilities, suggesting acceptable model calibration. DCA demonstrated greater net clinical benefit for the nomogram than either the “treat all” or “treat none” strategy across a threshold probability range of 5–80% (Fig. 6). The nomogram was considered to have potential clinical utility for risk-guided feeding decisions.

Fig. 4 [Images not available. See PDF.]

ROC curve for the FI predictive model.

Fig. 5 [Images not available. See PDF.]

Calibration curves for the FI nomogram.

The curve shows agreement between predicted and observed probabilities with closeness to the 45-degree diagonal indicating good calibration.

Fig. 6 [Images not available. See PDF.]

DCA of FI prediction nomogram for neonates with HIE during TH.

The DCA curve demonstrated the model’s net clinical benefit with the height of the curve indicating its clinical value.

Discussion

Enteral nutrition has a substantial influence on clinical prognosis in the critically ill neonate. The incidence of FI was 35.2% (63/179) in the current cohort. Neonatal infection, 5-minute Apgar score, hypoglycemia, time of enteral nutrition initiation, initial enteral feeding volume (15–30 mL/kg/day) and rate of feeding advancement (1–5 mL kg/day) were identified as predictors of FI by multivariable logistic regression analysis. A prediction model is presented for the estimation of FI probability in neonates with HIE. The previous literature on neonatal FI incorporates many variations in definitions and research approaches and further work on neonates with HIE during TH is required to consolidate the current findings.

Potential FI risk factors were investigated following a review of relevant literature and insights from clinical experts. The most pertinent characteristics were identified by LASSO analysis to address the challenges of multiple correlated variables. Variable multicollinearity was assessed by variance inflation factor (VIF), indicating that multicollinearity was not a major concern. The FI nomogram and predictive model may facilitate early identification of neonates with HIE informing clinical decisions regarding maintenance and promotion of feeding programs.

5-min Apgar score was found to be an independent factor for FI, consistent with previous studies which have found Apgar score to be a valuable diagnostic tool for neonatal asphyxia^25,26. A lower Apgar score indicates a requirement for additional intervention and a higher risk of NEC²⁷. Positive-pressure ventilation and endotracheal intubation are used to resuscitate neonates with cardiorespiratory failure and these patients are unlikely to maintain spontaneous breathing. The treatment increases airway pressure, abdominal pressure and blood flow redistribution in the acute phase of asphyxia but Havranek et al. reported a negative correlation between continuous positive airway pressure and microvascular oxygen saturation of the gastric mucosa²⁸. In addition, positive pressure ventilation may induce an inflammatory response and activate the renin-angiotensin system, reducing blood flow to internal organs, including the gastrointestinal tract, which may affect bowel movement. A persistent reduction in blood circulation associated with FI²⁹. Neonates prone to stress-induced metabolic dysfunction and hypoxia or ischemia may affect the intestinal mucosa, increasing inflammatory chemicals and decreasing gastric emptying³⁰. The 5-minute Apgar score is considered a better reflection of sustained perinatal compromise and correlates more strongly with adverse neonatal outcomes than the 1- or 10-minute scores³¹. While the 1-minute score may reflect transient depression, and the 10-minute score may be influenced by postnatal interventions, the 5-minute score more reliably indicates prolonged physiological stress³², which may impair splanchnic perfusion and contribute to FI.

The enteral feeding initiation time is critical for the prevention of FI in infants with HIE. Multiple organ dysfunction syndrome may be present, causing systemic injuries and severe hemodynamic disturbances which affect the gastrointestinal tract^33,34. Enteral nutrition may increase the risk of developing NEC and FI during TH³⁵ and congenital intestinal malformations or intestinal obstruction are clear contraindications for enteral feeding whereas asphyxia-hypoxia is not considered an absolute contraindication.

The optimal timing for nutritional support in neonates with HIE undergoing TH is difficult to define. The TOBY registry manual recommends introducing enteral nutrition “with caution” during TH after correction of biochemical and metabolic abnormalities³⁶. Prolonged fasting may cause atrophy of gastrointestinal mucosa, reduced diastase activity, weakened intestinal mucosal barriers and increased susceptibility to infection³⁷. Delayed feeding is known to hinder gastrointestinal maturation in preterm infants, slowing tract development and reducing hormone secretion which increases FI risk. A delay in the start of enteral feeding may increase the risk of intestinal inflammation and contribute to higher morbidity rates. Normal digestive function is impaired and the intestines become more susceptible to inflammation and associated complications. In general, delayed enteral feeding increases the risk of adverse health outcomes arising from gastrointestinal dysfunction³⁸. Protective effects on the gastrointestinal system have also been suggested for enteral feeding during TH⁹. Sakhuja et al. reported no change in the speed of blood flow in the bowel, assessed by ultrasound, during TH³⁹. Boo et al.⁴⁰ found the timing of milk introduction to be the sole risk factor for FI and early feeding has been suggested to enhance gastrointestinal tract maturation, stimulate gastrointestinal hormone secretion and activation and improve gastrointestinal motility^41,42. A 2017 UK hospital survey revealed significant variations in nutrition regimens during TH⁶. 59% hospitals routinely provided enteral nutrition with 45% initiating feeding on day 1, 41% on day 2 and 14% during the rewarming process. Decisions to initiate early feeding during TH were primarily based on the presence of gastrointestinal malformations and the infant’s hemodynamic stability. The timing of feeding initiation affects physiological and neurological development, recovery and the occurrence of complications in the neonate with HIE treated with TH. The current study highlights the importance of individualized feeding plans to optimize recovery and development.

Early minimal enteral feeding (MEN) with a gradual increase in milk feed may prevent the escalation of systemic inflammation and associated complications in neonates with HIE during TH⁴³, supplying nutrients and energy, regulating gastrointestinal motility, accelerating gastric emptying, improving gastrointestinal tolerance and reducing bacterial colonization of the gastrointestinal tract⁸. The biological properties of MEN have a greater favorable impact on gastrointestinal function than the nutritional properties⁴⁴. Moderate underfeeding in the early stages of acute illness, followed by a gradual increase to meet nutritional targets was suggested to be beneficial for critically ill neonates²⁴. In addition, the likelihood of emesis within the first 24 h after birth was greater in late preterm infants whose initial feeding volumes were > 8 mL/kg with the probability of FI increasing by 40% for volumes > 10 mL/kg. Indeed, a rapid increase in feeding volume was associated with an increased likelihood of FI⁴⁵ and a gradual increase protected against unfavorable feeding outcomes^35,46.

High plasma insulin levels often accompany neonatal asphyxia, resulting in hypoglycemia and poor outcomes⁴⁷. HIE classification, adverse outcomes and abnormal blood glucose levels have been linked and⁴⁸ hypoglycemia associated with severity and prognosis⁴⁹, as it exacerbates the brain damage caused by asphyxia⁵⁰. TH may reduce the extent of brain damage but coordination abnormalities, such as impaired suck-swallow-breathe, may persist and contribute to FI. Specifically, hypoglycemia may aggravate neurological injury, impair autonomic gastrointestinal regulation, and delay feeding coordination, thereby predisposing neonates to FI. Hypoglycemia was an independent risk factor for FI in the current cohort of neonates with HIE undergoing TH.

We recommend cautious and individualized enteral nutrition strategies for neonates with HIE undergoing TH. Early initiation of feeding may be considered within 24–48 h after TH initiation on condition that the infant is clinically stable. An initial feeding volume of 10–20 mL/kg/day with a slow rate of increase of 1–5 mL/kg/day was associated with reduced FI risk. Close monitoring for clinical signs of intolerance is essential and breast milk remains the preferred nutritional source. These recommendations, in conjunction with the predictive model, may assist clinicians in optimizing nutritional management of this high-risk population.

We acknowledge some limitations to the current work. First, it was a single-center retrospective study with a small sample size and a lack of external validation. The model showed good internal performance via bootstrap resampling but studies in larger, prospective, multicenter cohorts are required for confirmation. Second, selection bias and unmeasured confounders may have affected predictor identification and some clinically relevant variables, such as cord arterial pH and BE, were not included since > 20% data was missing due to inconsistent sampling during emergency or out-of-hospital admissions. Third, the definition of FI was adapted from criteria used for preterm infants and may not fully represent its manifestation in neonates with HIE. Lastly, only infants able to initiate enteral feeding during TH were included which may have excluded the most severely ill patients, leading to an underestimate of FI incidence and attenuating the association between HIE severity and feeding outcomes.

Conclusions

In conclusion, a nomogram to predict feeding intolerance in neonates with hypoxic-ischemic encephalopathy undergoing therapeutic hypothermia was developed and internally validated. The model enables early risk identification, allowing clinicians to implement preventive strategies such as delaying enteral feeding or adjusting feeding protocols based on individual risk levels. It also supports more standardized and personalized nutritional management with the potential to improve outcomes and reduce complications. Incorporating this tool into clinical workflows may aid decision-making and enhance communication with caregivers.

Acknowledgements

The authors acknowledge all members of Chongqing Medical University Children’s Hospital.

Author contributions

Conception and design: Huayin Yin, Jianhui Wang and Huayun He. Data collection: Ying Liu, Zhongping Shui, and Haimei Duan. Analysis and interpretation of data: Zhongping Shui and Qiuyi Sun. Drafting of the manuscript: Zhongping Shui. Writing, review, and editing: Zhongping Shui, Qiuyi Sun, and Huan He. All authors approved the final manuscript for submission and agreed to be accountable for all aspects of this study.

Funding

The current study was supported by the Department of Nursing, Chongqing Medical University Children’s Hospital (Grant No.: CHCQMU2023.11).

Data availability

The raw data supporting the conclusions of this study will be made available by the corresponding author upon reasonable request.

Declarations

Competing interests

The authors declare no competing interests.

Ethical approval

The Ethics Committee of Chongqing Medical University Children’s Hospital reviewed and approved this study (Approval Number: No. 482, Approval Date: 31 October 2023). Research involving human subjects has been conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from all subjects and/or their legal guardians.

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

1. Jacobs, S. E. et al. Cooling for newborns with hypoxic ischaemic encephalopathy. Cochrane Database Syst. Rev.2013, CD003311. https://doi.org/10.1002/14651858.CD003311.pub3 (2013).

2. Weeks, CL; Marino, LV; Johnson, MJ. A systematic review of the definitions and prevalence of feeding intolerance in preterm infants. Clin. Nutr.; 2021; 40, pp. 5576-5586. [DOI: https://dx.doi.org/10.1016/j.clnu.2021.09.010]

3. Stefanutti, G et al. Moderate hypothermia protects against systemic oxidative stress in a rat model of intestinal ischemia and reperfusion injury. Shock; 2005; 24, pp. 159-164.1:CAS:528:DC%2BD2MXpslGntrg%3D [DOI: https://dx.doi.org/10.1097/01.shk.0000168871.60531.6f]

4. Thornton, K. M. et al. Effects of whole body therapeutic hypothermia on gastrointestinal morbidity and feeding tolerance in infants with hypoxic ischemic encephalopathy. Int. J. Pediatr. 643689. https://doi.org/10.1155/2014/643689 (2014).

5. Battersby, C et al. Study protocol: optimising newborn nutrition during and after neonatal therapeutic hypothermia in the united kingdom: observational study of routinely collected data using propensity matching. BMJ Open.; 2018; 8, e026739. [DOI: https://dx.doi.org/10.1136/bmjopen-2018-026739]

6. Hazeldine, B et al. Survey of nutritional practices during therapeutic hypothermia for hypoxic-ischaemic encephalopathy. BMJ Paediatrics Open.; 2017; 1, e000022. [DOI: https://dx.doi.org/10.1136/bmjpo-2017-000022]

7. Markus, M et al. Fluid supply and feeding practices in cooled asphyxiated newborns. CHILDREN-BASEL; 2021; 8, 899. [DOI: https://dx.doi.org/10.3390/children8100899]

8. Chang, LL et al. Enteral feeding as an adjunct to hypothermia in neonates with Hypoxic-Ischemic encephalopathy. Neonatology; 2018; 113, pp. 347-352. [DOI: https://dx.doi.org/10.1159/000487848]

9. Kimberly, M. et al. Effects of whole body therapeutic hypothermia on gastrointestinal morbidity and feeding tolerance in infants with hypoxic ischemic encephalopathy. Int. J. Pediatr. 643689–643689. https://doi.org/10.1155/2014/643689 (2014).

10. Gale, C et al. Feeding during neonatal therapeutic hypothermia, assessed using routinely collected National neonatal research database data: a retrospective, UK population-based cohort study. Lancet Child. Adolesc. Health; 2021; 5, pp. 408-416. [DOI: https://dx.doi.org/10.1016/S2352-4642(21)00026-2]

11. Hu, Y et al. Early versus delayed enteral nutrition for neonatal hypoxic-ischemic encephalopathy undergoing therapeutic hypothermia: a randomized controlled trial. Ital. J. Pediatr.; 2022; 48, 146. [DOI: https://dx.doi.org/10.1186/s13052-022-01342-2]

12. Eveleens, RD et al. Definitions, predictors and outcomes of feeding intolerance in critically ill children: A systematic review. Clin. Nutr.; 2020; 39, pp. 685-693.1:STN:280:DC%2BB3M%2FjtFOnug%3D%3D [DOI: https://dx.doi.org/10.1016/j.clnu.2019.03.026]

13. Valentine, CJ. Nutrition and the developing brain. Pediatr. Res.; 2020; 87, pp. 190-191. [DOI: https://dx.doi.org/10.1038/s41390-019-0650-y]

14. Jiang, L et al. The effects of an enteral nutrition feeding protocol on critically ill patients: A prospective multi-center, before-after study. J. Crit. Care; 2020; 56, pp. 249-256. [DOI: https://dx.doi.org/10.1016/j.jcrc.2020.01.018]

15. Martinez, EE et al. Gastric emptying in critically ill children. J. Parenter. Enter. Nutr.; 2017; 41, pp. 1100-1109. [DOI: https://dx.doi.org/10.1177/0148607116686330]

16. Haney, A; Burritt, E; Babbitt, CJ. The impact of early enteral nutrition on pediatric acute respiratory failure. Clin. Nutr. ESPEN; 2018; 26, pp. 42-46. [DOI: https://dx.doi.org/10.1016/j.clnesp.2018.04.017]

17. Ortigoza, EB. Feeding intolerance. Early Hum. Dev.; 2022; 171, 105601. [DOI: https://dx.doi.org/10.1016/j.earlhumdev.2022.105601]

18. Reintam Blaser, A; Bachmann, KF; Deane, AM. Gastrointestinal function in critically ill patients. Curr. Opin. Clin. Nutr. Metab. Care; 2023; 26, 463. [DOI: https://dx.doi.org/10.1097/MCO.0000000000000955]

19. Heyland, DK; Incidence, et al. Risk factors, and clinical consequence of enteral feeding intolerance in the mechanically ventilated critically ill: an analysis of a multicenter, multiyear database. Crit. Care Med.; 2021; 49, 49. [DOI: https://dx.doi.org/10.1097/CCM.0000000000004712]

20. Reintam Blaser, A. et al. Enteral Feeding intolerance: updates in definitions and pathophysiology. Nutr. Clin. Pract.36, 40–49. https://doi.org/10.1002/ncp.10599 (2021).

21. Shao, X. M., Ye, H. M. & Qiu, X. S. Practical Neonatology, 5 edn. https://opac.nlc.cn/F?RN=167689320 (Accessed 8 May 2025) (People’s Medical Publishing House, 2019).

22. Levene, ML; Kornberg, J; Williams, TH. The incidence and severity of post-asphyxial encephalopathy in full-term infants. Early Hum. Dev.; 1985; 11, pp. 21-26.1:STN:280:DyaL2M3jtVaksw%3D%3D [DOI: https://dx.doi.org/10.1016/0378-3782(85)90115-x]

23. Hellström-Westas, L et al. Amplitude-integrated EEG classification and interpretation in preterm and term infants. NeoReviews; 2006; 7, pp. e76-e87. [DOI: https://dx.doi.org/10.1542/neo.7-2-e76]

24. Moltu, SJ et al. Nutritional management of the critically ill neonate. J. Pediatr. Gastroenterol. Nutr.; 2021; 73, pp. 274-289. [DOI: https://dx.doi.org/10.1097/MPG.0000000000003076]

25. James, LS et al. The acid-base status of human infants in relation to birth asphyxia and the onset of respiration. J. Pediatr.; 1958; 52, pp. 379-394.1:STN:280:DyaG1c%2Fot1Kksg%3D%3D [DOI: https://dx.doi.org/10.1016/S0022-3476(58)80058-X]

26. Michel, A. Review of the reliability and validity of the Apgar score. Adv. Neonatal Care; 2022; 22, pp. 28-34. [DOI: https://dx.doi.org/10.1097/ANC.0000000000000859]

27. Kordasz, M et al. Risk factors for mortality in preterm infants with necrotizing enterocolitis: a retrospective multicenter analysis. Eur. J. Pediatr.; 2022; 181, pp. 933-939.1:CAS:528:DC%2BB38XksFantrg%3D [DOI: https://dx.doi.org/10.1007/s00431-021-04266-x]

28. Havranek, T; Madramootoo, C; Carver, JD. Nasal continuous positive airway pressure affects pre- and postprandial intestinal blood flow velocity in preterm infants. J. Perinatol.; 2007; 27, pp. 704-708.1:STN:280:DC%2BD2snjtlekuw%3D%3D [DOI: https://dx.doi.org/10.1038/sj.jp.7211808]

29. Ukleja, A; Altered,. GI motility in critically ill patients. Nutr. Clin. Pract.; 2010; 25, pp. 16-25. [DOI: https://dx.doi.org/10.1177/0884533609357568]

30. Duan, H et al. The role of neuropeptide Y and Aquaporin 4 in the pathogenesis of intestinal dysfunction caused by traumatic brain injury. J. Surg. Res.; 2013; 184, pp. 1006-1012.1:CAS:528:DC%2BC3sXmsFeqtbc%3D [DOI: https://dx.doi.org/10.1016/j.jss.2013.03.096]

31. Iliodromiti, S et al. Apgar score and the risk of cause-specific infant mortality: A population-based cohort study. Lancet; 2014; 384, pp. 1749-1755. [DOI: https://dx.doi.org/10.1016/S0140-6736(14)61135-1]

32. Casey, BM; McIntire, DD; Leveno, KJ. The continuing value of the Apgar score for the assessment of newborn infants. N. Engl. J. Med.; 2001; 344, pp. 467-471.1:STN:280:DC%2BD3M7hvF2gtw%3D%3D [DOI: https://dx.doi.org/10.1056/NEJM200102153440701]

33. Polglase, GR; Ong, T; Hillman, NH. Cardiovascular alterations and multiorgan dysfunction after birth asphyxia. Clin. Perinatol.; 2016; 43, pp. 469-483. [DOI: https://dx.doi.org/10.1016/j.clp.2016.04.006]

34. Alward, CT et al. Effects of asphyxia on cardiac output and organ blood flow in the newborn piglet. Pediatr. Res.; 1978; 12, pp. 824-827.1:STN:280:DyaE1c3kvVSltw%3D%3D [DOI: https://dx.doi.org/10.1203/00006450-197808000-00003]

35. Oddie, SJ; Young, L; McGuire, W. Slow advancement of enteral feed volumes to prevent necrotising Enterocolitis in very low birth weight infants. Cochrane Db Syst. Rev.; 2017; 8, CD001241. [DOI: https://dx.doi.org/10.1002/14651858.CD001241.pub7]

36. UK TOBY Cooling Register Clinician’s Handbook 2010 Version 4. https://www.npeu.ox.ac.uk/.

37. Gale, C et al. Administration of parenteral nutrition during therapeutic hypothermia: a population level observational study using routinely collected data held in the National neonatal research database. Arch. Dis. Child. Fetal Neonatal Ed.; 2021; 106, pp. 608-613. [DOI: https://dx.doi.org/10.1136/archdischild-2020-321299]

38. Konnikova, Y. et al. Late enteral feedings are associated with intestinal inflammation and adverse neonatal outcomes. PLoS ONE10, e0132924. https://doi.org/10.1371/journal.pone.0132924 (2015).

39. Sakhuja, P et al. Gastrointestinal hemodynamic changes during therapeutic hypothermia and after rewarming in neonatal hypoxic-Ischemic encephalopathy. Pediatr. Neonatol; 2019; 60, pp. 669-675. [DOI: https://dx.doi.org/10.1016/j.pedneo.2019.04.003]

40. Boo, NY; Soon, CC; Lye, MS. Risk factors associated with feed intolerance in very low birthweight infants following initiation of enteral feeds during the first 72 hours of life. J. Trop. Pediatr.; 2000; 46, pp. 272-277.1:STN:280:DC%2BD3M%2Fks1Gguw%3D%3D [DOI: https://dx.doi.org/10.1093/tropej/46.5.272]

41. Liu, K et al. Effectiveness of early enteral feeding on health outcomes in preterm infants: an overview of systematic reviews. Eur. J. Clin. Nutr.; 2023; 77, pp. 628-636. [DOI: https://dx.doi.org/10.1038/s41430-022-01223-1]

42. Gunes, AO et al. Factors associated with the transition time to full enteral feeding in newborns with hypoxic ischemic encephalopathy. Arch. Iran. Med.; 2022; 25, pp. 547-551. [DOI: https://dx.doi.org/10.34172/aim.2022.87]

43. Bozkurt, O et al. Prolonged minimal enteral nutrition versus early feeding advancements in preterm infants with birth weight ≤ 1250 g: a prospective randomized trial. J. Maternal-Fetal Neonatal Med.; 2022; 35, pp. 341-347.1:CAS:528:DC%2BB3cXhvVSktbs%3D [DOI: https://dx.doi.org/10.1080/14767058.2020.1716723]

44. Ordaz-Jiménez, MR et al. [Gastrointestinal hormones during minimal enteral feeding of sick premature infants]. Rev. Invest. Clin.; 1998; 50, pp. 37-42.

45. Kumar, N. et al. Impact of feeding volumes in the first 24 h of life on neonatal feeding intolerance. Front. Pediatr. 11. https://doi.org/10.3389/fped.2023.1245947 (2023).

46. Masoli, D et al. Enteral feeding and necrotizing enterocolitis: does time of first feeds and rate of advancement matter??. J. Pediatr. Gastroenterol. Nutr.; 2021; 72, 763.1:CAS:528:DC%2BB3MXhs12kt7%2FL [DOI: https://dx.doi.org/10.1097/MPG.0000000000003069]

47. Madar, J et al. European resuscitation Council guidelines 2021: newborn resuscitation and support of transition of infants at birth. Resuscitation; 2021; 161, pp. 291-326. [DOI: https://dx.doi.org/10.1016/j.resuscitation.2021.02.014]

48. Guellec, I et al. Glycemia and neonatal encephalopathy: outcomes in the lytonepal (Long-Term outcome of neonatal hypoxic encephalopathy in the era of neuroprotective treatment with Hypothermia) cohort. J. Pediatr.; 2023; 257, 113350. [DOI: https://dx.doi.org/10.1016/j.jpeds.2023.02.003]

49. Basu, SK et al. Hypoglycaemia and hyperglycaemia are associated with unfavourable outcome in infants with hypoxic ischaemic encephalopathy: a post hoc analysis of the CoolCap study. Archives Disease Child. - Fetal Neonatal Ed.; 2016; 101, pp. F149-F155. [DOI: https://dx.doi.org/10.1136/archdischild-2015-308733]

50. Parmentier, CEJ et al. Hypoglycemia in infants with Hypoxic-Ischemic encephalopathy is associated with additional brain injury and worse neurodevelopmental outcome. J. Pediatr.; 2022; 245, pp. 30-38e1. [DOI: https://dx.doi.org/10.1016/j.jpeds.2022.01.051]