Dynamic liver test patterns do not predict bile duct stones

Area of Science:

• Gastroenterology
• Diagnostic Imaging
• Predictive Modeling

Background:

• Choledocholithiasis prediction models often perform suboptimally.
• Accurate prediction is crucial to minimize adverse events from unnecessary endoscopic retrograde cholangiopancreatography (ERCP) and reduce costs of advanced imaging like magnetic resonance cholangiopancreatography (MRCP) and endoscopic ultrasound (EUS).

Purpose of the Study:

• To identify clinical predictors with high positive and negative predictive value for choledocholithiasis.
• To evaluate the performance of the American Society for Gastrointestinal Endoscopy (ASGE) risk stratification algorithm in predicting bile duct stones.

Main Methods:

• Retrospective review of 604 inpatients undergoing ERCP for suspected bile duct stones.
• Analysis of liver enzyme patterns, transabdominal ultrasound, clinical features, and the ASGE risk stratification algorithm.
• Assessment of MRCP's predictive value for choledocholithiasis.

Main Results:

• Bile duct stones were confirmed in 67.9% of patients.
• No combination of liver enzymes, clinical features, or basic imaging yielded highly predictive algorithms.
• The ASGE high-risk criterion had a positive predictive value of only 68%; MRCP demonstrated the highest predictive value for stones.

Conclusions:

• Existing clinical and biochemical models are insufficient for accurately predicting choledocholithiasis.
• The ASGE risk stratification criteria were found to be insensitive in this population.
• MRCP or EUS may be considered to avoid unnecessary ERCP and associated complications.