First-Day Glycemic Exposure and 28-Day Mortality in the ICU: A Multicohort Study.

Area of Science:

• Critical Care Medicine
• Endocrinology
• Health Informatics

Background:

• Glycemic control in Intensive Care Units (ICUs) is crucial but often assessed using single glucose readings.
• Early glycemic exposure metrics are needed for timely interventions and personalized targets.

Purpose of the Study:

• To investigate the association between first-day time-weighted average glucose (TWAG) and 28-day mortality.
• To evaluate GlucoSurvAI, an interpretable AI model for early risk stratification in ICUs.

Main Methods:

• Retrospective cohort study of 8,048 adult ICU patients from 7 U.S. hospitals.
• Analysis of first-day TWAG (<100, 100-139, 140-179, ≥180 mg/dL) and 28-day mortality using Cox models.
• Performance evaluation of GlucoSurvAI using AUROC and Brier score, with SHAP for interpretability.

Main Results:

• Higher first-day TWAG was associated with increased 28-day mortality (e.g., 140-179 mg/dL: aHR 1.42; ≥180 mg/dL: aHR 1.41).
• GlucoSurvAI demonstrated high accuracy (AUROC 0.967) and low Brier score (~0.026).
• First-day vasopressors, corticosteroids, and insulin use were also linked to higher mortality.

Conclusions:

• A first-day TWAG threshold near 140 mg/dL is associated with elevated 28-day ICU mortality.
• GlucoSurvAI provides accurate, interpretable first-day risk prediction for ICU patients.
• Findings support individualized glycemic targets and early interventions based on exposure-aware metrics.