📄 medicine

Early risk prediction of necrotizing enterocolitis using 48-hour laboratory indicators and pre-symptom clinical-course variables: a single-center retrospective study

This single-center retrospective study developed and internally validated a nomogram incorporating five early postnatal predictors (5-minute Apgar score, parenteral nutrition, feeding advancement rate, neutrophil percentage, and procalcitonin) to effectively stratify the risk of necrotizing enterocolitis in neonates, achieving an AUC of 0.839.

Original authors: Yu Lang, Yuedi Hu, Ling Yan, Leilei Shen

Published 2026-07-07

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Yu Lang, Yuedi Hu, Ling Yan, Leilei Shen

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine a newborn baby's gut is like a delicate, untested bridge. Sometimes, this bridge starts to show cracks, leading to a serious condition called Necrotizing Enterocolitis (NEC). The problem is that the first signs of these cracks—like a little vomiting or a bloated tummy—are very vague. They look exactly like the bridge just being "picky" about its food (feeding intolerance) rather than actually breaking. Doctors often have to wait until the bridge is clearly collapsing (a confirmed diagnosis) to know for sure, but by then, it might be too late to prevent the worst damage.

This study is like a team of engineers trying to build a predictive weather forecast for that bridge. Instead of waiting for the storm to hit, they wanted to see if they could spot the warning signs before the bridge starts to crack.

The "Weather Report" (The Study Goal)

The researchers at Southwest Hospital in China looked back at the records of 248 babies who had tummy troubles. They wanted to create a simple tool—a scorecard (called a nomogram)—that could tell them, "This baby has a high chance of the bridge breaking soon," while "That baby is probably just having a tummy ache."

The Ingredients (The Predictors)

To build this scorecard, they didn't use complicated, expensive, or rare tests. They used five pieces of information that are usually already available in the first two days of a baby's life:

The 5-Minute Apgar Score: Think of this as the baby's "starting line health check." A lower score here is like a car that sputtered a bit when the engine was first turned on; it suggests the baby started life with a bit more stress.
Parenteral Nutrition (IV Feeding): This is when a baby gets food through a vein instead of a bottle or tube. Needing this early on is like a car needing a tow truck immediately after starting; it suggests the baby's system is fragile or having trouble handling normal food.
Feeding Advancement Rate: This measures how fast the doctors were able to increase the baby's milk. If they had to go very slowly (like creeping along in heavy traffic), it was a warning sign. If they could speed up the feeding (driving on an open highway), it was a good sign.
Neutrophil Percentage: These are white blood cells that act like the body's "security guards." If there are too many of them in the blood early on, it's like seeing a massive police presence before a crime has happened—it suggests the body is already on high alert for trouble.
Procalcitonin: This is another chemical marker that rises when the body is fighting inflammation. Think of it as a smoke detector that is starting to beep before you can even see the fire.

How the Scorecard Works

The researchers plugged these five numbers into a mathematical formula to create a risk score.

The Result: The scorecard was quite good at its job. It could correctly identify about 71% of the babies who would actually get the serious disease (NEC) and correctly rule out 86% of the babies who would be fine.
The "Tipping Point": They found a specific score (0.353) that acts like a threshold. If a baby's score is above this line, the model says, "Be very careful; this baby is high-risk." If it's below, the baby is likely lower risk.

What the Study Doesn't Say (Important Limits)

The authors are very careful not to overpromise. Here is what this tool is not:

It is not a crystal ball: It doesn't guarantee a baby will get sick or stay healthy. It just gives a better guess than a doctor's gut feeling alone.
It is not a final diagnosis: The study explicitly states this tool is for stratifying risk (sorting babies into groups) before the disease is fully confirmed. It doesn't replace the need for doctors to look at X-rays or examine the baby.
It is not ready for the whole world yet: This was a "single-center" study, meaning they only looked at babies from one specific hospital in China. The authors say this tool needs to be tested in other hospitals and with different groups of babies before it can be used as a standard rule everywhere.
It doesn't change the past: The study looked back at old records. It didn't test if using this scorecard actually saved lives or changed how doctors treated the babies in real-time.

The Bottom Line

Think of this study as building a prototype alarm system. The researchers found that by combining a baby's early birth health, how they were fed, and two simple blood tests, they could build a tool that sounds an alarm earlier than usual. It's a promising first step that suggests we might be able to spot the "cracks" in the bridge before the whole structure fails, but the alarm system still needs to be tested in other buildings before we trust it to save the day.

Technical Summary: Early Risk Prediction of Necrotizing Enterocolitis

Problem Statement
Necrotizing enterocolitis (NEC) is a life-threatening gastrointestinal disease in neonates, yet early identification remains a significant clinical challenge. The initial manifestations of NEC—such as vomiting, regurgitation, abdominal distension, and feeding intolerance—are nonspecific and frequently overlap with benign feeding intolerance or other gastrointestinal disorders. Consequently, distinguishing neonates at high risk of progressing to definitive NEC (Modified Bell stage II or III) from those with suspected or non-progressive conditions (Stage I or other manifestations) is difficult. Current risk assessment relies heavily on serial clinical judgment and imaging, but no single sign or laboratory marker has proven sufficiently reliable for early stratification. This study addresses the need for a predictive tool that operates within the "ambiguous window" before definite NEC progression occurs.

Methodology
This was a single-center, retrospective study conducted at the neonatal intensive care unit (NICU) of Southwest Hospital, Army Medical University, involving neonates admitted between January 2018 and January 2024.

Cohort: The study included 248 neonates who developed suspected NEC or feeding intolerance-related gastrointestinal manifestations. The cohort was divided into a case group (77 neonates with Modified Bell stage II or III NEC) and a control group (171 neonates with Stage I disease or non-NEC gastrointestinal manifestations).
Predictors: The study evaluated perinatal characteristics, maternal complications, feeding-related variables, pre-symptom clinical-course variables, and routine laboratory indicators measured within 48 hours after birth.
Model Development: Univariable logistic regression identified candidate predictors ( $P < 0.05$ ), which were then entered into a multivariable logistic regression using a forward conditional procedure. A nomogram was constructed based on the final model.
Validation and Analysis: Model performance was assessed using discrimination (ROC curve), calibration (Hosmer–Lemeshow test), and decision curve analysis (DCA). Internal validation was performed via bootstrap resampling (1,000 repetitions). A LASSO logistic regression was conducted as a sensitivity analysis to test the robustness of predictor selection.
Exclusions: Neonates with congenital gastrointestinal malformations, spontaneous intestinal perforation, or incomplete data were excluded. No imputation was performed; a complete-case analysis was utilized.

Key Contributions
The primary contribution of this work is the development and internal validation of a nomogram designed specifically for early risk stratification in neonates presenting with ambiguous gastrointestinal symptoms. Unlike models predicting outcomes after NEC is established, this tool integrates:

Early Postnatal Laboratory Indicators: Specifically neutrophil percentage and procalcitonin measured within 48 hours of birth.
Pre-Symptom Clinical Variables: Including feeding advancement rate and parenteral nutrition usage prior to the onset of gastrointestinal symptoms.
Perinatal Factors: Specifically the 5-minute Apgar score.

The study provides a quantitative framework to estimate individualized NEC risk before the disease progresses to a definitive stage, aiming to guide decisions regarding intensified monitoring rather than immediate therapeutic intervention.

Results

Predictors: The final model retained five independent predictors: 5-minute Apgar score, parenteral nutrition, feeding advancement rate, neutrophil percentage, and procalcitonin. Lower Apgar scores, use of parenteral nutrition, slower feeding advancement rates, higher neutrophil percentages, and higher procalcitonin levels were associated with increased NEC risk.
Performance: The nomogram demonstrated good discrimination with an Area Under the Curve (AUC) of 0.839 (95% CI, 0.785–0.893).
Thresholds: Using the Youden index, an optimal cutoff probability of 0.353 was identified. At this threshold, the model achieved a sensitivity of 71.4% and a specificity of 86.0%. The negative predictive value was 87.0%.
Validation: Calibration curves indicated acceptable agreement between predicted and observed risks (Hosmer–Lemeshow $P = 0.720$ ). Decision curve analysis suggested the model provided a net benefit over "treat-all" or "treat-none" strategies across a range of threshold probabilities.
Sensitivity: LASSO analysis confirmed the robustness of the five selected predictors, although it also identified additional variables (formula feeding, absolute neutrophil count, calcium) that were not retained in the primary parsimonious model.

Significance and Claims
The authors claim that this nomogram offers a clinically useful method for risk stratification during the early, uncertain phase of NEC development. By combining early biological signals (inflammatory markers) with pre-symptom clinical course data, the model aims to help clinicians identify neonates who require closer monitoring, repeated abdominal assessments, and earlier imaging reassessment.

The study emphasizes that the model is not intended to replace clinical diagnosis or dictate specific therapeutic interventions (such as surgery) but rather to provide an individualized risk estimate to support clinical vigilance. The authors explicitly state that the tool requires external and prospective validation before clinical implementation, noting limitations related to the single-center design, the specific gestational age profile of the cohort (which included term and late preterm infants), and the retrospective nature of the data. The work underscores the value of integrating routinely available data to refine early risk assessment in a challenging clinical window.