Abstract
This study aimed to develop an effective individualized predictive nomogram for the occurrence of necrotizing enterocolitis (NEC) in premature infants with early-onset sepsis (EOS). A total of 238 premature infants meeting the inclusion criteria of gestational age < 37 weeks and EOS diagnosis, including 71 with NEC and 167 without NEC (NEC incidence: 29.8%), treated at the First Hospital Affiliated to Army Medical University from January, 2016, to September, 2024 were retrospectively enrolled as a modeling cohort. Additionally, 205 preterm with EOS (53 with NEC and 152 with non-NEC, NEC incidence: 25.9%), who were treated at Liaocheng People’s Hospital from January, 2014, to September, 2024 were retrospectively enrolled as a validation cohort to assess the predictive efficacy of the model. LASSO-Logistic regression analysis were applied to screen independent predictors, which were subsequently incorporated into a nomogram constructed using R software. Model performance was assessed through receiver operating characteristic analysis, calibration curves, decision curve analysis (DCA), and clinical impact curves (CIC). Lasso-logistic regression identified four independent predictors of NEC in premature infants with EOS: chorioamnionitis (OR = 3.07, 95% CI: 1.26–7.48, p = 0.013), neonatal respiratory distress syndrome (OR = 2.20, 95% CI: 1.10–4.41, p = 0.027), lactate level (OR = 1.96, 95% CI: 1.48–2.58, p < 0.001), and white blood cell (WBC) count (OR = 0.89, 95% CI: 0.83–0.95, p < 0.001). These factors were integrated into the nomogram. The nomogram demonstrated excellent discriminative ability with the area under the receiver operating curve of 0.848 (95% CI: 0.793–0.903, sensitivity: 0.820, specificity: 0.761) in the modeling cohort and 0.825 (95% CI: 0.764–0.887, sensitivity: 0.750, specificity: 0.755) in the validation cohort, enabling early risk stratification for targeted clinical monitoring. Calibration curves confirmed good agreement between predicted and observed NEC probabilities (modeling cohort: χ2 = 3.539, df = 8, p = 0.896; validation cohort: χ2 = 12.769, df = 8, p = 0.120). DCA and CIC further verified the nomogram’s high net clinical benefit, confirming its utility in guiding clinical decision-making. This study establishes a nomogram based on four readily accessible variables to predict NEC in premature infants with EOS. With robust predictive performance, this tool enables early risk stratification of high-risk infants, facilitating timely and targeted monitoring and intervention.
Keywords: Early-onset sepsis, Necrotizing enterocolitis, Nomogram, Premature infants
Subject terms: Paediatric research, Gastroenterology, Gastrointestinal diseases
Introduction
Necrotizing enterocolitis (NEC) represents a devastating gastrointestinal emergency predominantly affecting preterm and low birth weight infants. NEC is one of the main cause of neonatal death, with a mortality as high as 20%-30%. Even survivors often experience short or long-term complications, such as intestinal stenosis, short bowel syndrome and neurological sequelae1,2, which seriously affects their quality of life.
The pathogenesis of NEC involves many factors, among which infectious mechanisms play an important role3,4. Sepsis has been proven to be a high-risk factor for NEC5,6. Notably, NEC complicating sepsis correlates with severe sequelae, such as intestinal perforation and peritonitis, which leads to increased surgical difficulty, poor prognosis, and a mortality up to 24.9%7. These outcomes emphasize the imperative for early risk stratification in clinical practice.
Several studies have developed prediction models for NEC in neonates, but these tools primarily focus on general neonatal populations or late-onset sepsis (LOS)- associated NEC, rather than the EOS subpopulation8–10. For example, Tao et al. identified risk factors for NEC in LOS neonates, failing to account for the unique timing of infection (≤ 72 h postnatal)10. Similarly, De et al. summarized NEC risk factors but did not focus on EOS-specific variables, limiting the utility of the model for EOS patients11. These general models often overlook the pathophysiological distinctiveness of EOS-associated NEC—where early infection impairs intestinal epithelial barrier function and triggers inflammatory responses.
NEC has an insidious onset, lacks specific diagnostic methods, and progresses rapidly, which may lead to serious consequences in a short period of time. Therefore, exploring risk factors acting in the prenatal and early neonatal periods and constructing a targeted risk prediction model is critical to optimizing strategies to reduce morbidity and mortality. However, to date, there is no research on risk factors associated with NEC in preterm infants with EOS, nor is there a perinatal risk prediction model to predict NEC occurrence in this specific population. To address this gap, we retrospectively analyzed clinical data of preterm infants with EOS and developed a risk prediction nomogram by integrating multiple early accessible indicators, aiming to provide a validated tool for early risk stratification of NEC in EOS preterm infants—ultimately facilitating timely, targeted intervention.
Methods
Study subjects
This retrospective multicenter study was conducted at two tertiary-care teaching hospitals in China: the First Hospital Affiliated to Army Medical University in Chongqing and Liaocheng People’s Hospital in Shandong. A total of 238 premature infants with EOS who were treated from January 1, 2016, to September 1, 2024 at First Hospital Affiliated to Army Medical University and who were eligible for inclusion based on the inclusion/exclusion criteria were enrolled as a modeling cohort. 205 preterm with EOS who were treated at Liaocheng People’s Hospital from January 1, 2014, to September 1, 2024, were enrolled as the validation cohort to assess the predictive efficacy of the model. No infants in either cohort received probiotics during hospitalization.
Inclusion criteria: (1) gestational age < 37 weeks; (2) diagnostic criteria for EOS according to the Chinese expert consensus on diagnosis and treatment of neonatal sepsis (2019)12. Exclusion criteria: (1) Bell stage I; (2) congenital malformations; (3) NEC diagnosed within 48 h of EOS confirmation; (4) incomplete medical records. According to whether developed proven NEC (Bell’s stage⩾II), infants with EOS were signed into NEC group and non-NEC group (modeling cohort: 71 infants with NEC and 167 infants with non-NEC; validation cohort: 53 infants with NEC and 152 infants with non-NEC). The fowchart of patient enrollment is shown in Fig. 1.
Fig. 1.
Screening flowchart of NEC in infants with EOS in the modeling cohort and validation cohorts.
This study was approved by the Ethics Committee of the First Hospital Affiliated of Army Medical University, and the study was conducted in accordance with the principles of the Helsinki Declaration. Due to the retrospective nature of the study, the need for informed consent was waived by First Hospital Affiliated of Army Medical University.
Clinical definitions
For the diagnosis of EOS, we referenced the expert consensus on the diagnosis and treatment of neonatal sepsis (2019)12 by the Subspecialty Group of Neonatology, Chinese Medical Association. This consensus defines EOS based on clinical signs and laboratory findings within the first 72 h of life, and explicitly includes both culture-proven and clinically-diagnosed cases: (1) Confirm diagnosis: Blood culture or culture of aseptic body cavity was positive (modeling cohort: 40/238, 16.8%; validation cohort: 25/205, 12.2%). (2) Clinical diagnosis: Blood nonspecific test was positive (at least 2 items); or abnormal cerebrospinal fluid examination; or DNA of pathogenic bacteria detected in blood (modeling cohort: 198/238, 83.2%; validation cohort: 180/205, 87.8%). The nonspecific blood tests included the following: ① blood cell count analysis: a white blood cell (WBC) count < 5 × 109/L or a WBC > 20 × 109/L; ② cell classification: immature neutrophils/total neutrophil (I/T) ratio ≥ 0.16; ③ platelet (PLT) count < 100 × 109/L; and ④ a C-reactive protein (CRP) level ≥ 8 mg/L. The diagnostic criteria for EOS are as follows: fever or low body temperature, less crying, poor mental response and other clinical manifestations. At the same time, it is necessary to meet at least one of the two diagnostic criteria.
For the diagnosis of NEC, refer to the fifth edition of Practical Neonatology13. NEC diagnostic criteria is neonates with vomiting, bloody stools, bloating, apnea, unstable body temperature, bradycardia and other symptoms, and X-rays suggest signs of intestinal wall gas accumulation, portal vein gas accumulation or intestinal perforation. Neonatal respiratory distress syndrome(NRDS) diagnostic criteria: the infant develops progressively worsening respiratory distress within 6 h after birth, and chest X-ray shows markedly decreased bilateral lung translucency with a ground-glass appearance, air bronchograms, or “white-out” lungs. Polycythemia referred to the venous hematocrit > 65%. Hemodynamically significant PDA (hsPDA) was defined as PDA with heart failure with clinical and/or radiological evidences. Preterm severe anemia was defined as hemoglobin (Hb) or hematocrit (Hct) level below the 1th percentile for the same gestational age and postnatal age. Septic shock was defined as sepsis with cardiovascular dysfunction. Asphyxia referred to 1 min or 5 min Apgar score ≤ 7, and umbilical artery pH < 7.2 at birth. The above diseases were diagnosed based on the 5th Edition of Practical Neonatology 13.
Data collection
Demographic data for all the study patients were obtained from the electronic medical records and included: (1) general information: gestational age, birth weight, gender, vaginal birth, asphyxia, small for gestational age (SGA), multiple births, breast feeding, meconium-stained amniotic fluid, etc.; (2) perinatal characteristics: maternal age, perinatal complications (hypertension, diabetes), chorioamnionitis, premature rupture of membranes (PROM) ≥ 18 h, use of glucocorticoids, and use antibiotics for the first 3 days before birth, etc.; (3) complications: NRDS, polycythemia, hsPDA, septic shock, severe anemia; (4) treatment measures: umbilical vein catheterization, transfusion of red blood cells, mechanical ventilation. (5) laboratory tests on the first day of life: white blood cell, hemoglobin, platelets, MPV, procalcitonin, lactate, albumin and blood culture.
The average age of diagnosis was calculated as the mean postnatal day of NEC diagnosis in the NEC group. The clinical data of non-NEC group were collected before the average age of diagnosis of NEC. When suspecting the presence of EOS, we used third-generation cephalosporins combined with penicillin for treatment, and then adjusted the anti-infection treatment measures based on blood culture or infection indicators.
Statistical analysis
SPSS 26.0 was used for data analysis. Continuous variables of normal distribution was described using mean ± standard deviation. Non-normal distributed are expressed as the median (interquartile range). Differences in continuous variables were assessed for significance using T test or Mann–Whitney U-tests. Categorical variables were presented as proportions and were compared using chi-squared test and Fisher’s exact test.
Development and assessment of the nomogram
Firstly, all collected variables were initially put into least absolute shrinkage and selection operator (LASSO) regression to screen the possible factors for predicting the occurrence of NEC in premature infants with EOS. The LASSO model was validated ten times using minimum criteria to determine the best parameter (lambda) so that the lambda value with the smallest cross-validation error would be the optimal model. Then, the variables identified by LASSO were incorporated into univariate and multiple logistic regression analysis to to determine associated the independent factors for NEC in preterm infants with EOS. The survminer package of R software was applied to draw forest plots, and the rms package of R was used to transform the final regression model into a nomogram. The receiver operating curve (ROC) and C-index analysis were performed for the validation and determine its discrimination ability of the model. The calibration curves of the modeling and validation cohort were plotted to evaluate the consistency between the predicted and observed probabilities. Decision curve analysis (DCA) and clinical impact curve (CIC) analysis was performed to evaluate the nomogram model’s potential for clinical application. All statistically signifcant tests were two-tailed, and p < 0.05 was regarded as statistically signifcant.
Sample size
Sample size was determined using the “10-events-per-variable (EPV)” empirical rule—a standard method for regression-based prediction model studies—to ensure parameter estimation stability and avoid overfitting. Through preliminary variable screening (Section “Development and assessment of the nomogram”), we anticipated a maximum of 5 potential predictors for NEC. Thus, the minimum required number of NEC cases (outcome events) was 5 × 10 = 50 per cohort.
Results
Patient’s characteristics
This study eventually included 443 premature infants with EOS, with 238 patients in the modeling cohort (71 patients with NEC and 167 patients with non-NEC) and 205 patients in the validation cohort (53 cases with NEC and 152 patients with non-NEC cases) (Fig. 1). During the study period, the modeling cohort had 6709 preterm infants born, among which 238 met the EOS diagnostic criteria and inclusion/exclusion criteria (the incidence of EOS: 3.54%). The validation cohort had 6384 preterm infants born, among which 205 met the EOS diagnostic criteria and inclusion/exclusion criteria (EOS incidence: 3.21%).The median postnatal day of NEC diagnosis was 11.11 (7.00, 13.00) in the modeling cohort and 12.57 (8.00, 16.00) in the validation cohort. The general information, perinatal characteristics, complications, treatment measures and laboratory indicators between the two groups or two cohorts were summarized in Tables 1, 2, and 3. In the modeling cohort, birth weight, gestational age, WBC count, PLT count and albumin level were significantly lower among the patients with NEC than the patients with non-NEC group, and lactate level and the proportion of PROM > 18 h, chorioamnionitis, NRDS, septic shock and the use of mechanical ventilation were significantly higher in NEC group (Table 1). All of the above mentioned changes were also observed in the validation cohort (Table 2), with the exception of PLT count and the use of mechanical ventilation, implying that these indicators were not significantly different between NEC and non-NEC group in the validation cohort (Table 3). Comparisons between the modeling and validation cohort suggested that only small differences were showed between the general characteristics, complications, laboratory results and treatment measures (Table 3).
Table 1.
The baseline characteristics of neonates with EOS in the modeling cohort.
| Non-NEC (n = 167) | NEC (n = 71) | p | |
|---|---|---|---|
| General characteristics | |||
| Gestational age (weeks) | 33.1 (30.7, 35.4) | 31.9 (29.6, 33.7) | 0.004 |
| Birth weight (g) | 1800.0 (1355.0, 2270.0) | 1405.0 (1110.0, 1840.0) | < 0.001 |
| Gender (male, n%) | 100 (59.9) | 47 (66.2) | 0.359 |
| Vaginal birth (n%) | 68 (40.7) | 28 (39.4) | 0.854 |
| Multiple birth (n%) | 46 (27.5) | 21 (29.6) | 0.750 |
| Prenatal corticosteroids (n%) | 107 (64.1) | 53 (74.6) | 0.112 |
| PROM > 18h (n%) | 38 (22.8) | 26 (36.6) | 0.027 |
| Meconium-stained amniotic fluid (n%) | 5 (3.0) | 2 (2.8) | 0.941 |
| 1 min Apgar | 10.0 (8.0, 10.0) | 9.00 (8.0, 10.0) | 0.120 |
| 5 min Apgar | 10.0 (10.0, 10.0) | 10.0 (10.0, 10.0) | 0.893 |
| SGA (n%) | 18 (10.8) | 14 (19.7) | 0.064 |
| Breastfeeding (n%) | 76 (45.5) | 37 (52.1) | 0.351 |
| Perinatal characteristics | |||
| Gestational hypertension (n%) | 20 (12.0) | 13 (18.3) | 0.196 |
| GDM (n%) | 35 (21.0) | 20 (28.2) | 0.227 |
| Intrahepatic cholestasis of pregnancy (n%) | 10 (6.0) | 2 (2.8) | 0.306 |
| Chorioamnionitis (n%) | 18 (10.8) | 20 (28.2) | 0.001 |
| Use antibiotics for the first 3 days before birth (n%) | 37 (22.2) | 21 (29.6) | 0.222 |
| Laboratory results | |||
| WBC (× 109/L) | 10.6 (7.1, 14.7) | 5.5 (3.4, 10.4) | < 0.001 |
| Hemoglobin (g/L) | 163.0 (141.0, 178.0) | 161.0 (137.0, 184.0) | 0.877 |
| PLT (× 109/L) | 236.0 (166.0, 294.0) | 201.0 (153.0, 254.0) | 0.018 |
| MPV (fL) | 10.3 (9.9, 10.9) | 10.4 (9.7, 11.2) | 0.688 |
| PCT(ng/ml) | 0.3 (0.1, 1.5) | 0.4 (0.1, 1.6) | 0.299 |
| Lactate (mmol/L) | 1.5 (1.2, 2.2) | 3.1 (2.1, 3.8) | < 0.001 |
| Alb (g/L) | 30.4 (26.2, 34.5) | 28.5 (24.7, 31.2) | 0.002 |
| Blood culture (positive, n%) | 30 (18.0) | 10 (14.1) | 0.464 |
| Complications | |||
| Asphyxia (n%) | 8 (4.8) | 3 (4.2) | 0.849 |
| NRDS (n%) | 54 (32.3) | 43 (60.6) | < 0.001 |
| Severe anemia (n%) | 12 (7.2) | 8 (11.3) | 0.299 |
| Polycythemia (n%) | 9 (5.4) | 4 (5.6) | 0.939 |
| hsPDA (n%) | 13 (7.8) | 8 (11.3) | 0.386 |
| Septic shock (n%) | 5 (3.0) | 9 (12.7) | 0.004 |
| Treatment measures | |||
| UVC (n%) | 32 (19.2) | 19 (26.8) | 0.191 |
| Transfusion of red blood cell (n%) | 15 (9.0) | 8 (11.3) | 0.585 |
| Mechanical ventilation (n%) | 44 (26.3) | 30 (42.3) | 0.015 |
Alb albumin, EOS early-onset sepsis, GDM gestational diabetes mellitus, hsPDA hemodynamically signifcant PDA, Lac lactate, MPV mean platelet volume, NEC necrotizing enterocolitis, NRDS neonatal respiratory distress syndrome, PCT Procalcitonin, PLT platelets, PROM premature rupture of membranes, SGA small for gestational age, UVC umbilical vein catheterization, WBC white blood cell.
Table 2.
The baseline characteristics of neonates with EOS in the validation cohort.
| Non-NEC (n = 152) | NEC (n = 53) | p | |
|---|---|---|---|
| General characteristics | |||
| Gestational age (weeks) | 31.9 (31.0, 34.2) | 31.6 (30.4, 32.4) | 0.030 |
| Birth weight (g) | 1595.0 (1342.5, 2400.0) | 1450.0 (1265.0, 1600.0) | 0.005 |
| Gender (male, n%) | 99 (65.1) | 36 (67.9) | 0.712 |
| Vaginal birth (n%) | 57 (37.5) | 24 (45.3) | 0.318 |
| Multiple birth (n%) | 23 (15.1) | 12 (22.6) | 0.211 |
| Prenatal corticosteroids (n%) | 131 (86.2) | 48 (90.6) | 0.409 |
| PROM > 18h (n%) | 30 (19.7) | 19 (35.8) | 0.018 |
| Meconium-stained amniotic fluid (n%) | 10 (6.6) | 1 (1.9) | 0.192 |
| 1 min Apgar | 10.0 (7.0, 10.0) | 10.0 (6.0, 10.0) | 0.748 |
| 5 min Apgar | 10.0 (9.0, 10.0) | 10.0 (9.0, 10.0) | 0.542 |
| SGA (n%) | 14 (9.2) | 12 (22.6) | 0.011 |
| Breastfeeding (n%) | 58 (38.2) | 23 (43.4) | 0.502 |
| Perinatal characteristics | |||
| Gestational hypertension (n%) | 50 (32.9) | 9 (17.0) | 0.028 |
| GDM (n%) | 31 (20.4) | 9 (17.0) | 0.589 |
| Intrahepatic cholestasis of pregnancy (n%) | 5 (3.3) | 2 (3.8) | 1.000 |
| Chorioamnionitis (n%) | 12 (7.9) | 13 (24.5) | 0.001 |
| Use antibiotics for the first 3 days before birth (n%) | 44 (28.9) | 22 (41.5) | 0.092 |
| Laboratory results | |||
| WBC (× 109/L) | 11.4 (8.2, 14.7) | 6.9 (5.0, 8.9) | < 0.001 |
| Hemoglobin (g/L) | 170.0 (156.3, 184.8) | 167.0 (153.0, 176.0) | 0.114 |
| PLT (× 109/L) | 236.5 (200.3, 272.3) | 239.0 (200.0, 261.5) | 0.625 |
| MPV (fL) | 10.3 (9.7, 10.9) | 10.2 (9.7, 10.6) | 0.095 |
| PCT(ng/ml) | 0.4 (0.2, 1.2) | 0.3 (0.2, 1.0) | 0.280 |
| Lac (mmol/L) | 1.9 (1.4, 3.4) | 3.2 (2.0, 4.4) | < 0.001 |
| Alb (g/L) | 29.0 (26.0, 32.8) | 25.0 (23.0, 29.0) | 0.030 |
| Blood culture (positive, n%) | 16 (10.5) | 9 (17.0) | 0.216 |
| Complications | |||
| Asphyxia (n%) | 14 (9.2) | 7 (13.2) | 0.409 |
| NRDS (n%) | 49 (32.2) | 33 (62.3) | < 0.001 |
| Severe anemia (n%) | 21 (13.8) | 7 (13.2) | 0.912 |
| Polycythemia (n%) | 12 (7.9) | 5 (9.4) | 0.726 |
| hsPDA (n%) | 8 (5.3) | 5 (9.4) | 0.283 |
| Septic shock (n%) | 5 (3.3) | 8 (15.1) | 0.002 |
| Treatment measures | |||
| UVC (n%) | 44 (28.9) | 13 (24.5) | 0.536 |
| Transfusion of red blood cell (n%) | 19 (12.5) | 9 (17.0) | 0.413 |
| Mechanical ventilation (n%) | 60 (39.5) | 23 (43.4) | 0.616 |
Alb albumin, EOS early-onset sepsis, GDM gestational diabetes mellitus, hsPDA hemodynamically signifcant PDA, Lac lactate, MPV mean platelet volume, NEC necrotizing enterocolitis, NRDS neonatal respiratory distress syndrome, PCT Procalcitonin, PLT platelets, PROM premature rupture of membranes, SGA small for gestational age, UVC umbilical vein catheterization, WBC white blood cell.
Table 3.
The baseline characteristics of neonates with EOS in modeling cohort and validation cohort.
| Modeling cohort (n = 238) | Validation cohort (n = 205) | p | |
|---|---|---|---|
| General characteristics | |||
| Gestational age (weeks) | 32.8 (30.4, 35.0) | 31.7 (31.0, 33.5) | 0.172 |
| Birth weight (g) | 1660.0 (1250.0, 2138.8) | 1520.0 (1310.0, 2200.0) | 0.925 |
| Gender (male, n%) | 147 (61.8) | 135 (65.9) | 0.372 |
| Vaginal birth (n%) | 96 (40.3) | 81 (39.5) | 0.860 |
| Multiple birth (n%) | 67 (28.2) | 35 (17.1) | 0.006 |
| Prenatal corticosteroids (n%) | 160 (67.2) | 179 (87.3) | < 0.001 |
| PROM > 18h (n%) | 64 (26.9) | 49 (23.9) | 0.472 |
| Meconium-stained amniotic fluid (n%) | 7 (2.9) | 11 (5.4) | 0.197 |
| 1 min Apgar | 10.0 (8.0, 10.0) | 10.0 (10.0, 10.0) | 0.147 |
| 5 min Apgar | 10.0 (8.0, 10.0) | 10.0 (9.0, 10.0) | 0.115 |
| SGA (n%) | 32 (13.4) | 26 (12.7) | 0.812 |
| Breastfeeding (n%) | 113 (47.5) | 81 (39.5) | 0.092 |
| Perinatal characteristics | |||
| Gestational hypertension (n%) | 33 (13.9) | 59 (28.8) | < 0.001 |
| GDM (n%) | 55 (23.1) | 40 (19.5) | 0.358 |
| Intrahepatic cholestasis of pregnancy (n%) | 12 (5.0) | 7 (3.4) | 0.399 |
| Chorioamnionitis (n%) | 38 (15.9) | 25 (12.2) | 0.338 |
| Use antibiotics for the first 3 days before birth (n%) | 58 (24.4) | 57 (27.8) | 0.411 |
| Laboratory results | |||
| WBC (× 109/L) | 8.8 (5.6, 13.1) | 9.8 (6.9, 13.3) | 0.086 |
| Hemoglobin (g/L) | 162.0 (141.0, 179.0) | 170.0 (156.0, 181.0) | 0.001 |
| PLT (× 109/L) | 228.5 (159.0, 289.5) | 237.0 (200.0, 265.0) | 0.138 |
| MPV (fL) | 10.3 (9.8, 11.0) | 10.2 (9.7, 10.8) | 0.247 |
| PCT(ng/ml) | 0.3 (0.1, 1.5) | 0.4 (0.2, 1.0) | 0.157 |
| Lac (mmol/L) | 1.8 (1.3, 2.9) | 2.1 (1.5, 3.4) | 0.098 |
| Alb (g/L) | 29.4 (25.9, 33.5) | 29.0 (26.0, 32.0) | 0.095 |
| Blood culture (positive, n%) | 40 (16.8) | 25 (12.2) | 0.171 |
| Complications | |||
| Asphyxia (n%) | 11 (4.6) | 21 (10.2) | 0.023 |
| NRDS (n%) | 97 (40.8) | 82 (40.0) | 0.872 |
| Severe anemia (n%) | 20 (8.4) | 28 (13.7) | 0.076 |
| Polycythemia (n%) | 13 (5.5) | 17 (8.3) | 0.237 |
| hsPDA (n%) | 21 (8.8) | 13 (6.3) | 0.328 |
| Septic shock (n%) | 14 (5.9) | 13 (6.3) | 0.840 |
| Treatment measures | |||
| UVC (n%) | 51 (21.4) | 57 (27.8) | 0.119 |
| Transfusion of red blood cell (n%) | 23 (9.7) | 28 (13.7) | 0.189 |
| Mechanical ventilation (n%) | 74 (31.1) | 83 (40.5) | 0.039 |
Alb albumin, EOS early-onset sepsis, GDM gestational diabetes mellitus, hsPDA hemodynamically signifcant PDA, Lac lactate, MPV mean platelet volume, NEC necrotizing enterocolitis, NRDS neonatal respiratory distress syndrome, PCT Procalcitonin, PLT platelets, PROM premature rupture of membranes, SGA small for gestational age, UVC umbilical vein catheterization, WBC white blood cell.
Risk factors identification for the occurrence of NEC in premature infants with EOS
A three-step analytical approach-LASSO regression, univariate, and multivariate logistic regression was employed to identify NEC-associated risk factors. LASSO was performed to screen out the nonzero features from the modeling cohort. Consequently, the partial likelihood deviance curve with the log (lambda) was plotted. In addition, two dotted vertical lines were drawn at the optimal values by determining the minimum criteria and the 1 standard error (SE) of the minimum criteria (Fig. 2A). The number of potential predictors reduced from 34 to 7, including Birth weight, Chorioamnionitis, 5 min apgar, NRDS, WBC, Lac and Blood culture) (Fig. 2B). Thereafter, the above-mentioned 7 variables were applied to univariate logistic analysis, and the results showed that birth weight, chorioamnionitis, NRDS, WBC count, lactate level may be the risk factors for the occurrence of NEC in premature infants with EOS (p < 0.05) (Table 4). Subsequent multivariate logistic regression analysis confirmed three independent risk factors: the presence of chorioamnionitis (OR = 3.07, 95% CI: 1.26–7.48, p = 0.013), NRDS (OR = 2.20, 95% CI: 1.10–4.41, P = 0.027), and high lactate level (OR = 1.96, 95% CI: 1.48–2.58, p < 0.001) were independent risk factors for the development of NEC in premature infants with EOS (p < 0.05) (Table 4), while a lower WBC count (OR = 0.89, 95% CI: 0.83–0.95, p < 0.001) was independently associated with an increased risk of NEC (Fig. 3).
Fig. 2.
Selection of predictor features in the modeling cohort using least absolute shrinkage and selection operator regression. (A) The selected optimal parameter (lambda) in the least absolute shrinkage and selection operator model was tenfold cross-validation based on the minimum criteria. (B) Least absolute shrinkage and selection operator coefficient profiles of the 34 characteristics. Seven features with nonzero coefficients were finally obtained.
Table 4.
Logistic regression analysis results.
| Variables | Univariate analysis | Multivariates analysis | ||
|---|---|---|---|---|
| OR (95% CI) | p | OR (95% CI) | p | |
| Birth weight | 1.00 (1.00–1.00) | < 0.001 | ||
| Chorioamnionitis | 3.25 (1.59–6.61) | 0.001 | 3.07 (1.26–7.48) | 0.013 |
| 5 min Apgar | 1.06 (0.81–1.40) | 0.659 | ||
| NRDS | 3.21 (1.81–5.71) | < 0.001 | 2.20 (1.10–4.41) | 0.027 |
| WBC | 0.86 (0.80–0.92) | < 0.001 | 0.89 (0.83–0.95) | < 0.001 |
| Lac | 2.03 (1.56–2.63) | < 0.001 | 1.96 (1.48–2.58) | < 0.001 |
| Blood culture | 2.03 (1.56–2.63) | 0.465 | ||
GDM gestational diabetes mellitus, Lac lactate, NRDS neonatal respiratory distress syndrome, WBC white blood cell.
Fig. 3.
Forest chart of the analysis of independent risk factors in the modeling cohort.
Development of the predictive nomogram model for the occurrence of NEC in premature infants with EOS
Based on the above results of the multivariate regression analysis, the independent risk predictors for the occurrence of NEC in premature infants with EOS from the modeling cohort were applied to R software to construct the nomogram model (Fig. 4). The sum of the corresponding scores for chorioamnionitis, NRDS, WBC count and lactate level in this nomogram was associated with an increased risk of NEC. As an example case in which the nomogram was applied, a premature infants with EOS had chorioamnionitis and NRDS, a WBC count of 3 × 109/L, and lactate level of 3 mmol/L. The predictive variables produced the following number of points in the nomogram model: chorioamnionitis: 11 points; NRDS: 9 points; WBC count: 53 points; lactate level: 20 points. Therefore, this patient had a total score of 93 points. Corresponding to the total score at the bottom of the chart, the probability of NEC was approximately 88.0%.
Fig. 4.
Nomogram model for early prediction of NEC in premature infants with EOS. A nomogram for NEC in premature infants with EOS was developed based on the presence of chorioamnionitis, NRDS, WBC and lactate level. EOS, early-onset sepsis; NRDS, neonatal respiratory distress syndrome; NEC, necrotizing enterocolitis; WBC, White blood cell.
Validation and evaluation of the predictive nomogram model for the occurrence of NEC in premature infants with EOS
The area under the receiver operating curve (AUC) for the nomogram in the modeling cohort was 0.848 (95% CI: 0.793–0.903) (sensitivity: 0.820, specificity: 0.761, NPV: 0.643 and PPV: 0.890), C-index = 0.848 (Fig. 5A), and the calibration curve of the model (χ2 = 3.539, df = 8, p = 0.896) revealed a good agreement between the nomogram predicted probability and the actual observed result of NEC (Fig. 5B). In addition, the probability of NEC in premature infants with EOS from the validation cohort was calculated by the nomogram model. The area under the curve (AUC) for the validation cohort was 0.825 (95% CI: 0.764–0.887) (sensitivity: 0.750, specificity: 0.755, NPV: 0.513 and PPV: 0.898) (Fig. 5C), which also had a well-calibrated curve (χ2 = 12.769, df = 8, p = 0.120) in the assessment of risk (Fig. 5D). Furthermore, the optimal risk threshold for predicting NEC in the modeling cohort was identified as 0.307, using the ROC curve and maximum Youden index (0.581).
Fig. 5.
The ROC curves and calibration curves of the nomogram model in modeling and validation cohort. (A) Modeling cohort ROC curve. (B) Validation cohort ROC curve. (C) In the training cohort, the calibration curve for the nomogram mode was showed. (D) In the validation cohort, the calibration curve for the nomogram mode was showed. The dotted line represents the performance of the model. The diagonal line represents the ideal prediction. ROC, receiver operating characteristic.
Finally, we used DCA and CIC curve analysis to evaluate the nomogram model’s potential for clinical application. DCA on the modeling and validation cohort showed that the capacity of the nomogram to predict the occurrence of NEC was more beneficial than either the “treat-all” or “treat-none” strategy, with a threshold probability of 10–85% and 10–70%, respectively (Fig. 6A and B). The CIC curves showed that the nomogram has high efficiency in the identification of NEC in premature infants with EOS at a relatively higher risk threshold (Fig. 6C and D). All the DCA and CIC curves indicated that the nomogram model might be promising in clinical decision-making.
Fig. 6.
The decision analysis curves and clinical impact curves of the nomogram model in modeling and validation cohort. (A) The DCA curve for the model of the modeling cohort. (B) The DCA curve for the model of the validation cohort. A Net Benefit is showed on the y-axis, the x-axis is labeled Risk Threshold Probability. The black line indicates the net benefit assuming no participants develop NEC, whereas the light grey curve signifies the net benefit assuming all participants are at risk of developing NEC. The clinical utility of the model is illustrated by the area between the red curve (nomogram) and the light grey curve (treat-all) and black line (treat-none). The larger the area between the model curve and the black and light gray line, the better the nomogram’s clinical value. (C) The CIC curve based on risk factor risk models in the modeling cohort. (D)The CIC curve based on risk factor risk models in the validation cohort. Out of 1,000 neonates, in the solid red line, the number of babies considered high risk at each risk threshold is showed, whereas in the blue dashed line, the number of newborns who were considered true positives (cases) is showed. The close alignment between the red (number of high-risk infants) and blue (number of true positives) lines at lower thresholds indicates good precision of the model. DCA, decision curve analysis; CIC, clinical impact curve.
Discussion
Necrotizing enterocolitis is a life-threatening disease that can lead to intestinal perforation, peritonitis, and even death. Studies have shown that sepsis is an important risk factor for NEC14. When sepsis occurs, bacteria can directly damage intestinal epithelial cells, and the endotoxins and other products they produce can also cause intestinal necrosis15. At the same time, the body is in an inflammatory activated state during severe infection, producing a variety of inflammatory factors, including platelet activating factor, tumor necrosis factor-α, etc., all of which can directly or indirectly lead to the damage of intestinal mucosa, and are involved in the occurrence and development of NEC15,16. In addition, infection can cause intestinal vascular constriction, intestinal ischemia and hypoxia, accumulation of metabolic products, which ultimately trigger NEC17,18.
As a consequence of the fulminant nature of NEC, and its symptoms and signs in the early stages are non-specific, often leading to delayed diagnosis and proper treatment, resulting in a poor prognosis. It is unlikely that new diagnostic and therapeutic strategies will achieve major breakthroughs in reducing the mortality and morbidity. Early identification of possible risk factors is more likely to achieve better results. Therefore, this study aims to analyze the specific risk factors for the development of NEC in infants with EOS. Our results revealed that the presence of chorioamnionitis, NRDS, low WBC count and high lactate level were independent risk factors for NEC in neonates with EOS. Based on these risk factors, we constructed a predictive nomogram model for the occurrence of NEC. The AUC of the ROC curves for the modeling and validation cohort were 0.848 (95% CI: 0.793–0.903) and 0.825 (95% CI: 0.764–0.887), respectively, indicating that the performance of the nomogram model was satisfactory with good discrimination and prediction accuracy. Meanwhile, DCA and CIC curve demonstrated that the nomogram model had a high overall net benefit in both cohorts. The four variables used to establish the nomogram model are easily obtained at admission, which can help to improve our ability to identify at-risk infants, thereby reducing the incidence of this devastating disease in this vulnerable population.
The incidence of chorioamnionitis in the modeling and validation cohort of this study were 15.96% (38/238) and 12.19% (25/205), respectively, indicating that pregnant with chorioamnionitis requires sufficient attention. Gantert et al.19 pointed out that chorioamnionitis can lead to developmental changes in a range of organs, including the intestine, and is associated with adverse outcomes in neonates. Several studies have reported that chorioamnionitis is associated with an increased risk for the development of NEC14,20,21. The mechanism of NEC caused by chorioamnionitis is still unclear, and the possible reasons are as follows: (1) Chorioamnionitis can prevent intestinal development. Study have found that inflammation in utero can affect the maturation of the fetal intestine, ultimately leading to postnatal intestinal pathology22; (2) Chorioamniotis may lead to impaired development of intestinal innate immune defense, tight junction distribution, and vascular function in the uterine gut, making it easy for microorganisms and toxins to enter the mucosa and intestinal lining19.
A study from 26 tertiary hospitals in Guangzhou Province showed that NRDS increased the risk of early neonatal death23. Hong et al.24 confirmed that grade III ~ IV NRDS were high-risk factors for late-onset NEC occurring between 8-14d. The inflammatory response that occurs during EOS is not conducive to fetal lung maturation. In an animal model study of intrauterine infection in neonatal rats, it was found that the number and volume of alveoli decreased in the intrauterine infection group, indicating that intrauterine infection may affect alveolar development25. When the balance of inflammatory reaction in the body is broken, a large amount of inflammatory mediators may cause extensive lung tissue damage, thus increasing the incidence rate and severity of NRDS. Study provide evidence that in premature infants with grade III-IV NRDS, immature lungs can impair gas exchange function and intestinal tissue oxygenation disorders, leading to the occurrence of NEC26. In addition, most children with NRDS require invasive mechanical ventilation, which has a significant impact on changes in intrathoracic pressure and affects systemic venous return, which may affect intestinal blood supply and increase the risk of NEC27.
This study found that elevated lactate level was an independent risk factor for the occurrence of NEC. As an important indicator of systemic perfusion and oxygen metabolism, elevated lactate level reflects the increase of anaerobic metabolism under low perfusion28. It has already been shown that hyperlactatemia serves as a possible predictor of poor outcome in critically ill children29. Valpacos et al.30 showed that increased lactate levels increased the likelihood of NEC for 45% of respondents. Srinivasjois et al.31 Found that plasma lactate level may predict progression of definite NEC to surgery or death in preterm neonates. Kordasz et al.32 found that high lactate level at disease onset and during disease correlated with NEC severity and mortality. The pathogenesis can be explained as follows: the systemic inflammatory response syndrome triggered by EOS leads to the release of vasoconstrictive mediators and impaired autoregulation, resulting in splanchnic vasoconstriction. In the immature intestinal vasculature of preterm infants, this significantly compromises microcirculatory flow. The consequent shift to anaerobic metabolism within intestinal tissues generates lactic acid, directly reflected by rising blood lactate levels. This localized intestinal hypoxia and acidosis damage the mucosal barrier, facilitating bacterial translocation and initiating the inflammatory cascade characteristic of NEC15,17. Furthermore, in sepsis, hepatic blood flow and function are often impaired, reducing the liver’s capacity to clear lactate, which further exacerbates the hyperlactatemia33. Thus, an elevated lactate level in an infant with EOS is not merely a marker of illness severity but a direct warning sign of gut hypoperfusion and ischemic injury, providing a plausible biological pathway to NEC.
Our model identifies lower WBC count as a significant risk factor for NEC, a finding that, while seemingly counterintuitive in the context of infection, is pathophysiologically sound. This result should be interpreted as leukopenia signifies a state of immunologic dysfunction, which increases vulnerability to NEC. A depressed WBC count during sepsis often signifies a more ominous state of bone marrow suppression, or impaired immune response to the overwhelming infection34,35. An infant who cannot maintain an adequate WBC count is thus failing to mount an effective immune response. This state of immunologic dysfunction may impair the ability to contain the septic insult and to protect the gut from secondary bacterial invasion and translocation, thereby increasing the risk of developing subsequent complications like NEC. A study found that compared to survivors, infants who died had significantly lower WBC, and WBC < 4 × 109/L was significantly associated with mortality from late-onset sepsis in very low birth weight infant36. Yu et al.37 found that NEC neonates necessitating surgery or death show significantly lower WBC level. Consequently, a relatively higher WBC count in infants with EOS may indicate a greater reserved capacity.
While the individual predictors are informative, the principal strength of this nomogram lies in its integrative and synergistic capacity, providing a holistic assessment of risk that reflects the multifactorial pathogenesis of NEC following EOS. The model captures a plausible pathophysiological cascade: Chorioamnionitis represents a prenatal ‘first hit,’ priming the fetal gut through inflammatory exposure in utero, which impairs intestinal barrier function and innate immunity19,22. After birth, the presence of NRDS constitutes a significant postnatal ‘second hit,’ characterized by hypoxia and the hemodynamic consequences of respiratory support, which can compromise splanchnic perfusion and exacerbate intestinal injury26,27. This cascade of events is quantified by the lactate level, an objective biomarker that integrates the effects of systemic hypoxia and tissue hypoperfusion resulting from both the septic process and respiratory failure28,33. Finally, a low WBC count signals the infant’s inability to mount an effective response to these insults, indicating bone marrow suppression or immune exhaustion in the face of overwhelming inflammation 34,35. Thus, a high score on the nomogram does not merely indicate the presence of multiple risk factors but identifies an infant undergoing a synergistic vicious cycle of prenatal priming, postnatal physiological compromise, tissue hypoxia, and immune depletion. This integrative approach allows for a more nuanced and accurate risk stratification than any single variable could achieve, effectively translating the complex biology of EOS-associated NEC into a practical clinical tool.
It is noteworthy that while both gestational age and birth weight were significant in univariate analysis, they were not retained as independent predictors in the final multivariate model. This is a key finding that underscores the value of our model. It suggests that the influence of these fundamental demographic risk factors is mediated through the downstream clinical complications that are more directly captured by our final predictors. For instance, lower gestational age predisposes infants to NRDS, and both immaturity and growth restriction contribute to hemodynamic vulnerability that is reflected in elevated lactate levels. The inflammatory insult of chorioamnionitis and the immune status reflected by the WBC count are more proximal events in the causal pathway to NEC following EOS. Therefore, our model, by incorporating these potent proximal mediators of risk, effectively integrates the vulnerability conferred by lower gestational age and birth weight but in a form that is more directly related to the active pathophysiological state of the infant. This parsimonious model achieved excellent predictive performance without them, highlighting the utility of the selected variables in this specific clinical context.
Several prediction models for NEC in preterm infants have been previously developed, utilizing a variety of prenatal, clinical, and microbiological factors. These models have shown promising predictive abilities, with reported AUCs often ranging between 0.80 and 0.90. It is important to situate our work within this existing literature. Key previous models, such as those by Tao et al.10 for NEC in late-onset sepsis (AUC: 0.860) or by Zhang et al.9 for general preterm populations (AUC: 0.805), differ fundamentally from our study in their target population. To the best of our knowledge, the present nomogram is the first specifically developed and validated for predicting NEC in the high-risk subpopulation of preterm infants with confirmed EOS. This distinction is critical, as the pathophysiology and risk profile may differ in infants already experiencing a systemic inflammatory response from early-onset infection. While direct numerical comparison of AUC values with models derived from general preterm cohorts is not appropriate, the discriminatory power of our model (AUCs of 0.848 and 0.825) is robust and comparable to these previous efforts, suggesting its utility within this specific context. A key advantage of our model is its reliance on only four variables that are readily available early in the clinical course, facilitating rapid risk stratification at the bedside.
The DCA indicated that the nomogram provided a superior net benefit for clinical decision-making compared to the ‘treat all’ or ‘treat none’ strategies across a wide range of clinically relevant risk thresholds (10–85% in the modeling cohort and 10–70% in the validation cohort). We propose a threshold probability of 0.307 as a balanced, clinically actionable cut-off. This corresponds to a total point value of approximately 67 on the nomogram. Infants scoring above this threshold should be considered for intensified monitoring protocols or feeding adjustments. The CIC confirmed that at this threshold, the number of infants identified as high-risk is clinically manageable and contains a high proportion of true cases, underscoring the model’s efficiency in targeting interventions.
However, this study has some limitations. Firstly, it is a retrospective study that may lead to a certain degree of selection bias. Some potential risk factor data that may affect the occurrence of NEC in premature infants with EOS have not been fully preserved. Secondly, a potential limitation of this study is the extended time frame of data collection (2014–2024), which may lead to temporal bias. While we have ensured consistent application of inclusion/exclusion criteria and diagnostic standards across the entire study period, we acknowledge that unmeasured temporal variations in supportive care (e.g., ventilation strategies, feeding protocols) cannot be fully excluded and might influence the incidence and management of NEC. Thirdly, our study utilized a definition of EOS that encompasses clinically diagnosed cases in addition to culture-proven cases. While this reflects pragmatic clinical practice, it is broader than the bacteremia-proven definitions used in some studies (e.g., by NICHD). This inclusion may introduce heterogeneity, as the inflammatory state in culture-negative infants could have non-infectious etiologies. Fourthly, our study did not differentiate between medical and surgical NEC in the predictive model. While this was beyond the scope of our primary aim, we recognize that identifying risk factors for the most severe, surgical form of the disease is of paramount clinical importance. The limited number of surgical NEC cases in our cohorts precluded a robust sub-analysis for this outcome. We propose that validating our model’s ability to predict surgical NEC, and identifying additional predictors specific to disease severity, constitute critical objectives for future large-scale, prospective studies. Finally, there were significant differences in some baseline characteristics (prenatal corticosteroid administration, Asphyxia, gestational hypertension, and mechanical ventilation) between the modeling and validation cohorts. These differences reflect real-world variations in patient populations and clinical practice across different medical centers. While such discrepancies could potentially threaten the external validity of a prediction model, the fact that our nomogram maintained strong discriminatory power (AUC = 0.825) in the validation cohort is reassuring and speaks to its robustness. This is likely because the final model is driven by a core set of pathophysiologically grounded predictors that were similarly distributed between both cohorts and were not confounded by the management-related variables that differed. Nevertheless, future prospective validations in broader, multi-center cohorts are warranted to further confirm the generalizability of our nomogram across diverse clinical settings.
Conclusions
This study developed and validated a clinical prediction nomogram for the occurrence of NEC in premature infants with EOS. The visual nomogram enables early risk stratification, allowing clinicians to identify high-risk neonates requiring intensified monitoring and guide prophylactic interventions based on modifiable parameters. Future implementation studies should evaluate its effectiveness in reducing NEC incidence through targeted surveillance protocols.
Acknowledgements
This work was supported by the National Natural Science Foundation of China (Grant/Award Number: 82301956) and Science and Health Joint Medical Research Project of Chongqing (Grant Number: 2025MSXM037).
Author contributions
H.Y.L. and N.C.: conceptualization, methodology. Y.W. and L.L.S.: formal analysis, investigation. H.Y.L. and Y.W.: investigation, data curation. H.Y.L. and N.C.: writing-original draft, visualization. Y.W. and L.L.S.: supervision. N.C. and Y.W.: project administration. All authors reviewed and approved the final version of the manuscript.
Data availability
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.
Declarations
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Huifan Li and Yu Wang have equally contributed to this work.
References
- 1.Neu, J. & Pammi, M. Pathogenesis of NEC: impact of an altered intestinal microbiome. Semin. Perinatol.41(1), 29–35 (2017). [DOI] [PubMed] [Google Scholar]
- 2.Frost, B. L. et al. New medical and surgical insights into neonatal necrotizing enterocolitis: A review. JAMA Pediatr.171(1), 83–88 (2017). [DOI] [PubMed] [Google Scholar]
- 3.Zhang, L. P. et al. Risk factors for necrotizing enterocolitis in very preterm infants: A case-control study in southwest China. J. Matern Fetal. Neonatal Med.32(6), 896–901 (2019). [DOI] [PubMed] [Google Scholar]
- 4.Wertheimer, F., Arcinue, R. & Niklas, V. Necrotizing enterocolitis: enhancing awareness for the general practitioner. Pediatr. Rev.40(10), 517–527 (2019). [DOI] [PubMed] [Google Scholar]
- 5.Mackay, C. A. et al. Epidemiology and outcomes of neonatal sepsis: Experience from a tertiary Australian NICU. Neonatology121(6), 703–714 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Multicenter Study Collaborative Group for Evaluation of Outcomes in Very Low Birth Weight Infants. A multicenter prospective cohort study of late-onset sepsis and its poor prognosis in very low birth weight infants. Zhonghua Er Ke Za Zhi. 61(3): 228–234 (2023). [DOI] [PubMed]
- 7.Bizzarro, M. J., Ehrenkranz, R. A. & Gallagher, P. G. Concurrent bloodstream infections in infants with necrotizing enterocolitis. J. Pediatr164(1), 61–66 (2014). [DOI] [PubMed] [Google Scholar]
- 8.Liu, K., Guo, J., Zhu, Y., Yang, J. & Su, Y. Analysis of risk factors and establishment of predictive models for neonatal necrotizing enterocolitis: a retrospective study. Ital J. Pediatr.51(1), 80 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Zhang, B. et al. Establishing a nomogram for predicting necrotizing enterocolitis in premature infants: A retrospective multicenter cohort study. Dig Liver Dis.57(1), 231–240 (2025). [DOI] [PubMed] [Google Scholar]
- 10.Tao, M. et al. Development and validation of a nomogram model for predicting the occurrence of necrotizing enterocolitis in premature infants with late-onset sepsis. Eur. J. Med. Res.30(1), 595 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.De Bernardo, G. et al. Clinical picture, diagnosis, management of NEC, and effects of probiotics on its prevention: A narrative review. Curr. Pediatr. Rev.21(2), 104–110 (2025). [DOI] [PubMed] [Google Scholar]
- 12.Subspecialty Group of Neonatology, the Society of Pediatrics, Chinese Medical Association, et al. Expert consensus on the diagnosis and management of neonatal sepsis (version 2019). Zhong hua Er Ke Za Zhi. 57(4): 252–257 (2019). [DOI] [PubMed]
- 13.Shao, X. M., Ye, H. M. & Qiu, X. S. Practice of neonatology 5th edn. (People’s Medical Publishing House, 2018). [Google Scholar]
- 14.Lu, Q., Cheng, S. P., Zhou, M. & Yu, J. L. Risk factors for necrotizing enterocolitis in neonates: A retrospective case-control study. Pediatr. Neonatol.58(2), 165–170 (2017). [DOI] [PubMed] [Google Scholar]
- 15.Neu, J. & Pammi, M. Necrotizing enterocolitis: The intestinal microbiome, metabolome and inflammatory mediators. Semin. Fetal. Neonatal Med.23(6), 400–405 (2018). [DOI] [PubMed] [Google Scholar]
- 16.Tirone, C. et al. Gut and lung microbiota in preterm infants: Immunological modulation and implication in neonatal outcomes. Front. Immunol.10, 2910 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Rose, A. T. & Patel, R. M. A critical analysis of risk factors for necrotizing enterocolitis. Semin. Fetal. Neonatal Med.23(6), 374–379 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Samuels, N. et al. Risk factors for necrotizing enterocolitis in neonates: A systematic review of prognostic studies. BMC Pediatr.17(1), 105 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Gantert, M. et al. Chorioamnionitis: A multiorgan disease of the fetus?. J. Perinatol.30(1), 21–30 (2010). [DOI] [PubMed] [Google Scholar]
- 20.Been, J. V. et al. Chorioamnionitis as a risk factor for necrotizing enterocolitis: a systematic review and meta-analysis. J. Pediatr.162(2), 236–42.e2 (2013). [DOI] [PubMed] [Google Scholar]
- 21.Duci, M. et al. Maternal and placental risk factors associated with the development of necrotizing enterocolitis (NEC) and its severity. J. Pediatr. Surg.54(10), 2099–2102 (2019). [DOI] [PubMed] [Google Scholar]
- 22.Neu, J., Mshvildadze, M. & Mai, V. A roadmap for understanding and preventing necrotizing enterocolitis. Curr. Gastroenterol. Rep.10(5), 450–457 (2008). [DOI] [PubMed] [Google Scholar]
- 23.Collaborative Study Group for Extremely Preterm and Extremely Low Birth Weight Infants. Short-term outcomes and their related risk factors of extremely preterm and extremely low birth weight infants in Guangdong province. Zhonghua Er Ke Za Zhi. 57(12):934–942 (2019). [DOI] [PubMed]
- 24.Hong, K. Y. & Zhu, Y. Chinese multicenter EUGR Collaborative Group, et al. The role of nutrition in analysis of risk factors and short-term outcomes for late-onset necrotizing enterocolitis among very preterm infants: A nationwide, multicenter study in China. BMC Pediatr.24(1), 172 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Muratore, C. S. et al. Endotoxin alters early fetal lung morphogenesis. J. Surg. Res.155(2), 225–230 (2009). [DOI] [PubMed] [Google Scholar]
- 26.McPherson, C. & Wambach, J. A. Prevention and treatment of respiratory distress syndrome in Preterm neonates. Neonatal Netw.37, 169–177 (2018). [DOI] [PubMed] [Google Scholar]
- 27.Lamia, B., Molano, L. C., Muir, J. F. & Cuvelier, A. Cardiopulmonary interactions in the course of mechanical ventilation. Rev. Mal. Respir.33, 865–876 (2016). [DOI] [PubMed] [Google Scholar]
- 28.Caplan, M. S., Simon, D. & Jilling, T. The role of PAF, TLR, and the inflammatory response in neonatal necrotizing enterocolitis. Semin Pediatr Surg.14, 145–151 (2005). [DOI] [PubMed] [Google Scholar]
- 29.Patki, V. K., Antin, J. V. & Khare, S. H. Persistent hyperlactatemia as the predictor of poor outcome in critically ill children: a single-center, prospective, observational cohort study. J. Pediatr. Intensive Care.6, 152–158 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Valpacos, M. et al. Diagnosis and management of necrotizing enterocolitis: An international survey of neonatologists and pediatric surgeons. Neonatology113, 170–176 (2018). [DOI] [PubMed] [Google Scholar]
- 31.Srinivasjois, R. et al. Prediction of progression of definite necrotising enterocolitis to need for surgery or death in preterm neonates. J. Matern. Fetal. Neonatal. Med.23, 695–700 (2010). [DOI] [PubMed] [Google Scholar]
- 32.Kordasz, M. et al. Risk factors for mortality in preterm infants with necrotizing enterocolitis: a retrospective multicenter analysis. Eur. J. Pediatr.181(3), 933–939 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Ha, T. S. et al. Lactate clearance and mortality in septic patients with hepatic dysfunction. Am. J. Emerg. Med.34, 1011–1015 (2016). [DOI] [PubMed] [Google Scholar]
- 34.Aird, W. C. The hematologic system as a marker of organ dysfunction in sepsis. Mayo. Clin. Proc.78, 869–881 (2003). [DOI] [PubMed] [Google Scholar]
- 35.Belok, S. H., Bosch, N. A., Klings, E. S. & Walkey, A. J. Evaluation of leukopenia during sepsis as a marker of sepsis-defining organ dysfunction. PLoS ONE16(6), e0252206 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Goh, G. L. et al. Risk factors for mortality from late-onset sepsis among preterm very-low-birthweight infants: A single-center cohort study from Singapore. Front. Pediatr.31(9), 801955 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Yu, M., Liu, G., Feng, Z. & Huang, L. Combination of plasma white blood cell count, platelet count and C-reactive protein level for identifying surgical necrotizing enterocolitis in preterm infants without pneumoperitoneum. Pediatr. Surg. Int.34(9), 945–950 (2018). [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.