AKI Subtyping and Prognostic Analysis Based on Serum Electrolyte Features in ICU

Area of Science:

• Nephrology
• Critical Care Medicine
• Data Science in Healthcare

Background:

• Acute Kidney Injury (AKI) is a common complication in intensive care units (ICUs).
• Heterogeneity exists among AKI patients, suggesting potential for distinct subtypes.
• Current AKI classification may not fully capture variations in mortality risk and treatment response.

Purpose of the Study:

• To identify distinct subtypes of ICU patients with AKI using serum electrolyte data.
• To assess the association between these AKI subtypes and in-hospital mortality risk.
• To investigate how treatments like diuretics and renal replacement therapy (RRT) affect mortality risk within each subtype.

Main Methods:

• Utilized the eICU Collaborative Research Database (eICU-CRD) and a Chinese tertiary hospital dataset.
• Identified AKI patients based on KDIGO guidelines.
• Employed K-Medoids clustering on initial serum electrolyte measurements to define AKI subtypes.
• Applied logistic regression to analyze mortality risk and treatment associations within subtypes.
• Validated findings using a separate critical care dataset.

Main Results:

• Three distinct AKI subtypes were identified in over 15,000 ICU patients.
• Subtype 1 (normal electrolytes) had the lowest mortality risk; Subtype 2 (high chloride) had moderate risk; Subtype 3 (high phosphate, low bicarbonate) had the highest risk.
• Treatment associations with mortality varied significantly: diuretics and RRT reduced mortality in Subtype 3 but increased it in Subtype 2.
• Findings were consistently validated in an independent Chinese cohort.

Conclusions:

• Serum electrolyte profiles can effectively differentiate AKI subtypes with varying mortality risks in ICUs.
• The effectiveness and impact of treatments like diuretics and RRT are subtype-dependent.
• AKI subtyping holds potential for developing personalized management strategies to improve patient outcomes.