[Predicting very early rebleeding after acute variceal bleeding based in classification and regression tree analysis...

Area of Science:

• Hepatology
• Gastroenterology
• Internal Medicine

Context:

• Variceal bleeding (VB) is a critical complication in cirrhosis, leading to significant mortality.
• Early rebleeding (30-50%) necessitates accurate risk stratification for timely intervention.
• Current prognostic models for very early rebleeding (VER) are limited.

Purpose:

• Identify risk factors for VER after acute VB.
• Develop and validate a novel prognostic model using Classification and Regression Tree Analysis (CART).
• Compare CART model performance against established MELD and Child-Pugh scores.

Summary:

• A study of 60 cirrhotic patients with acute VB identified serum albumin, creatinine, and initial blood transfusion volume as key predictors of VER.
• The CART model, utilizing these variables, demonstrated superior predictive accuracy (AUC=0.82) compared to MELD (AUC=0.46) and Child-Pugh (AUC=0.50).
• Specific thresholds identified by CART include albumin < 2.85 mg/dL, creatinine > 1.65 mg/dL, and transfusion of >= 2 units of packed red cells within 24 hours.

Impact:

• The CART algorithm provides a simple yet accurate tool for predicting VER in cirrhotic patients.
• This facilitates more aggressive and personalized treatment strategies for high-risk individuals.
• Improved risk assessment can potentially reduce mortality associated with variceal bleeding in cirrhosis.