Objective: To establish a diagnostic algorithm, magnetic resonance cholangiopancreatography (MRCP) vs. endoscopic retrograde cholangiopancreatography (ERCP), to evaluate for common bile duct (CBD) stones in patients presenting with acute gallstone pancreatitis (AGP).Methods: A retrospective review of 125 consecutive patients presenting with AGP from 2002-7 was performed. Significance was determined by unpaired t-test and multivariate analysis. Results: Of the 125 patients, 26 underwent MRCP. 5 patients had positive MRCP and required ERCP. 38 patients underwent ERCP, 20 (52.6%) had CBD stones. Comparison of AGP patients with and without CBD stones demonstrated increased CBD size on ultrasound (US) (10.7±6 vs. 6±3 mm), admission alkaline phosphatase (AP) (341± 249 vs.161±102 u/l), total bilirubin (TB) (5.32±6.0 vs. 2.02±3 mg/dl) and direct bilirubin (DB) (4.5±6 vs.1.2±2 mg/dl) as significant predictors of CBD stone (p<0.05).Patients were stratified into low, intermediate and high probability groups based on mean and standard deviation of the 4 predictive criteria. Sensitivity, specificity and positive predictive value (PPV) for CBD stone were then calculated.(Table 1)Patients meeting clinical criteria for high probability group had a 100% PPV and specificity of CBD stone and should undergo initial ERCP. AGP patients with CBD stone are effectively captured by the clinical criteria comprising the intermediate group (90% sensitivity) but the low PPV (37.5%) indicates that the majority will not have CBD stones and therefore not require intervention. These patients should undergo initial MRCP. Patients meeting low probability criteria are unlikely to have a stone and may be evaluated by cholangiogram at time of cholecystectomy. Conclusion: Patients who meet high probability criteria should undergo ERCP as they have 100% PPV and specificity of CBD stone. Intermediate group patients have a 90% sensitivity but low PPV (37.5%) of CBD stone and should undergo MRCP. Low probability patients may be evaluated by cholangiogram at time of cholecystecomy. Application of this algorithm effectively screens for patients with CBD stones and stratifies them to appropriate initial testing; potentially decreasing the amount of unnecessary studies and interventions.
Criteria for Determining Probability of CBD stone in Patients with Acute Gallstone Pancreatitis

CBD stone probability	CBD size + AP + TB + DB on US (mm) (U/L) (mg/dl) (mg/dl)	Sensitivity	Specificity	PPV
Low	<3.0 + <200 + <1.5 + <0.5	5%	79%	4%
Intermediate	≥3.0 + ≥200 + ≥1.5 + ≥0.5	90%	69%	37.5%
High	≥9.0 + ≥270 + ≥5.0 + ≥3.0	35%	100%	100%