Back to 2017 Posters

Objectives: Common bile duct stones (CBDS) are present in 4-20% of patients undergoing cholecystectomy for cholecystitis. Biochemical markers, including bilirubin, and ultrasound findings are often used as initial markers for CBDS. In equivocal cases bilirubin is often repeated after several hours to assess trend and plan further imaging. We hypothesized that trends in total bilirubin could be used to improve the accuracy of CBDS prediction.
Design: Retrospective comparison of various methods to analyze sequential serum bilirubin to predict CBDS. Presence of stones was confirmed by MRCP, ERCP or cholangiogram.
Setting: University Hospital
Patients: 59 patients, over a five-year period with acute cholecystitis (mean age 61.2, 57% female) presenting with an elevated bilirubin.
Interventions: None
Main Outcome Measure(s): Multiple measures were tested using receiver-operator characteristic (ROC) curves to determine the best method for predicting CBDS.
Results: The total bilirubin level at presentation, average over 24 hours and average over 48 hours (3.74 mg/dl vs. 2.29 mg/dl, p=0.005; 3.72 mg/dl vs. 2.40 mg/dl, p=0.009; 2.41 mg/dl vs. 1.47 mg/dl, p<0.001) respectively, were all higher in those with CBDS. When comparing the ROCs the maximum AUC was 0.726 (table). There was no statistically significant difference between the ROC curves.
Conclusions: In patients who present with elevated serum bilirubin, delaying the decision to pursue surgical intervention or additional imaging to obtain follow-up bilirubin does not improve predictive value. If bilirubin is trended over time for other reasons, the maximum bilirubin in 24 hours should be used to predict CBDS.