Noninvasive Prediction Models of First Decompensation in Compensated Advanced Chronic Liver Disease: A Meta-Analysis

Area of Science:

• Hepatology
• Internal Medicine
• Clinical Prediction Models

Background:

• Compensated advanced chronic liver disease (cACLD) carries a significant risk of hepatic decompensation.
• Accurate identification of high-risk patients is crucial for timely intervention.

Purpose of the Study:

• To critically appraise noninvasive prediction models (NITs) for identifying cACLD patients at increased risk of hepatic decompensation.
• To determine the accuracy and reliability of existing NITs.

Main Methods:

• Systematic review and meta-analysis of published articles until February 2025.
• Included studies evaluated prognostic models combining at least two noninvasive markers.
• Extracted data included C-statistics, AUC, and model calibration.

Main Results:

• 30 studies with 47,647 participants were included, evaluating 39 prognostic models.
• Models specifically developed and validated in cACLD, such as SAVE (C-statistic=0.87) and ABC scores (C-statistic=0.85), demonstrated superior prediction.
• Limited validation and calibration were noted across most models.

Conclusions:

• cACLD-specific NITs are the preferred choice for predicting decompensation.
• Future research must prioritize robust validation, calibration, and external validation with standardized endpoints for clinical reliability.