A multinational collaboration led by Northwestern University and KU Leuven introduces an XGBoost-based clinical decision tool to predict acute kidney injury and survival in neonates treated with therapeutic hypothermia. By integrating gestational age, birth weight, postnatal age, and early serum creatinine trends, the model achieves AUC 0.95 and 75% accuracy on cross-validated multicenter data, enabling timely risk stratification and individualized neonatal management.
Key points
XGBoost classifier uses gestational age, birth weight, postnatal age, and daily serum creatinine to predict five neonatal outcome classes.
Trained on 1,149 hypothermia-treated neonates and 801 controls with stratified 10-fold cross-validation and patient-level data splits.
Achieves mean AUC 0.95 and 75.1% overall accuracy, outperforming existing neonatal AKI biomarkers for early risk stratification.
Why it matters:
This high-accuracy AI tool enables clinicians to identify at-risk neonates under therapeutic hypothermia earlier, potentially improving interventions and outcomes.
Q&A
How does the XGBoost model handle serial creatinine data?
Why is predicting AKI in cooled neonates challenging?
What does an AUC of 0.95 signify?
What is therapeutic hypothermia in neonates?
Read full article
Academy
Therapeutic Hypothermia in Neonatal Care
Therapeutic hypothermia is a medical intervention that lowers a newborn’s body temperature to 33–34°C for up to 72 hours. It is applied to infants who suffered perinatal asphyxia—a lack of oxygen around the time of birth—to reduce brain injury and improve neurological outcomes.
Oxygen deprivation, or hypoxia, can cause a cascade of cellular damage in the brain and other organs. Cooling slows metabolism, reduces inflammation, and stabilizes cell membranes, limiting the spread of injury. Hypothermia is now the standard of care for moderate to severe hypoxic–ischemic encephalopathy (HIE) in high-income neonatal intensive care units.
Mechanisms and Benefits:
- Metabolic Reduction: Cooling decreases the brain’s energy requirements by up to 50%, preserving ATP stores and reducing excitotoxicity.
- Inflammation Control: Hypothermia dampens inflammatory cytokine release, limiting secondary injury in the brain and peripheral organs.
- Apoptosis Inhibition: Lower temperatures slow programmed cell death pathways, reducing neuronal loss over the treatment period.
Despite clear neurological benefits, hypothermia affects multiple organ systems. In neonates, the kidney’s response to cooling is critical because impaired renal perfusion and delayed creatinine clearance can lead to acute kidney injury (AKI). Early detection of AKI allows timely adjustments to fluid management, drug dosing, and supportive therapies.
Monitoring Kidney Function in cooled neonates
In cooled infants, traditional markers like serum creatinine and urine output are harder to interpret:
- Serum Creatinine Dynamics: At birth, an infant’s creatinine reflects maternal levels, then falls over days. Cooling further alters clearance rates.
- Urine Output Variability: Cooling and resuscitation fluids can cause transient oliguria or polyuria, which do not always correlate with kidney injury severity.
As a result, clinicians need more robust tools combining multiple clinical variables and trend analysis rather than single-value thresholds.
Integrating Machine Learning
Machine learning models, such as XGBoost, can ingest time-series data—including daily creatinine measurements, postnatal age, birth weight, and gestational age—to learn complex patterns that precede AKI. These models outperform static cutoff methods and can provide probability estimates for risk, enabling earlier interventions.
Clinical Application and Outcomes
By integrating an AI-based decision support tool into neonatal care workflows, healthcare teams can:
- Automatically flag high-risk neonates within the first 72 hours of cooling.
- Tailor fluid and drug dosing regimens based on individualized risk profiles.
- Plan closer monitoring or earlier renal protective measures for infants predicted to develop AKI.
Ongoing research aims to refine these models with additional biomarkers (e.g., NGAL, KIM-1) and larger datasets, ultimately improving both neurological and renal outcomes in vulnerable newborns.
