Nutrition for the Sick Preterm: Can We Make It More Precise?

Area of Science:

• Neonatal intensive care
• Nutritional science
• Biomedical informatics

Background:

• Historically, neonatal nutrition relied on clinical intuition, leading to varied outcomes.
• Current evidence-based guidelines, while improving care, often fail to address individual preterm infant needs due to population-based statistics.
• Existing malnutrition scoring systems offer reactive identification rather than proactive nutritional guidance.

Purpose of the Study:

• To highlight the limitations of current neonatal nutritional strategies.
• To introduce the potential of precision-based approaches for optimizing infant nutrition.
• To explore the role of AI and ML in developing proactive, individualized nutritional interventions.

Main Methods:

• Review of historical and current neonatal nutritional practices.
• Discussion of limitations in population-based and scoring-based approaches.
• Exploration of emerging AI/ML technologies for risk stratification and mechanistic investigation.

Main Results:

• Current generalized nutritional guidelines inadequately serve the heterogeneous preterm infant population.
• AI and ML offer novel capabilities for predictive analytics and risk categorization.
• Multiomic integrations can elucidate mechanistic pathways and identify biomarkers for preventative strategies.

Conclusions:

• Precision nutrition, powered by AI/ML, promises to revolutionize neonatal intensive care by enabling proactive, individualized nutritional strategies.
• This approach moves beyond population averages to address the unique needs of each preterm infant.
• Future research should focus on integrating multiomics and AI/ML for biomarker discovery and targeted preventative interventions.