Back to 2011 Program

Background: Detection of persistent common bile duct (CBD) stones in patients presenting with acute biliary pancreatitis (ABP) remains challenging, resulting in high rates of negative endoscopic retrograde cholangiopancreatography (ERCP) and magnetic resonance cholangiopancreatography (MRCP). In a study previously conducted by our institution, an algorithm for selective management of patients with ABP was developed. In this study, multivariate analysis identified 5 clinical variables which significantly correlated with persistent stone. Optimal values were: CBD size on ultrasound≥9mm; alkaline phoshatase≥250U/L; gamma-glutamyl transferase≥350U/L; total bilirubin≥3mg/dl; and direct bilirubin≥2 mg/dl. We hypothesized that grouping variables would increase reliability of detection. Receiver operator curve (ROC) curve analysis and linear regression were applied to obtain optimal and equitable predictive values, and variables combined. Based on results, a 5-point scoring system was devised and recommendations for biliary evaluation suggested. Table 1 demonstrates recommendations according to probability of CBD stone as assessed by score. Purpose: To prospectively study an algorithm developed for evaluation for CBD stones in patients presenting with ABP.Method: Following institutional board approval, all patients presenting with a diagnosis of ABP were offered enrollment into a prospective study. Informed consent was obtained for all enrolled patients. Patients not enrolled were considered controls. Enrollment was at the discretion of the patient and attending physician. To date, 45 patients have been captured of which data collection is complete for 34. Figure 1 demonstrates study design. Results: Of the 34 patients, 17 were in-protocol and 17 were controls. Twenty patients were low probability (0-2), 10 patients intermediate (3-4) and 4 patients high probability (5). CBD stones were present in 15% of low probability, 30% intermediate and 100% of high probability patients (p<0.01). A total of 13 MRCP and 6 ERCP were performed. Overall, 69% of MRCP performed were negative for CBD stone and 33% of ERCP were negative. Comparison of patients demonstrated decreased performance of MRCP (23% vs. 53%) and rate of negative MRCP (50% vs. 78%) as well as decreased performance of ERCP (12% vs. 23%) and decreased rate of negative study (0% vs. 50%) for protocol versus control patients. For the low probability protocol patients with CBD stone, all stones were detected at time of cholangiogram and extracted via ERCP. No postoperative complications ensued.Conclusion: Preliminary results demonstrate that the algorithm correlates with probability of CBD stone. This algorithm may be utilized to stratify patients to appropriate initial testing; decreasing the amount of unnecessary studies and interventions.
Table 1: Positive predictive value (PPV), associated odds ratio (OR) with 95% confidence interval [CI] and recommendations for biliary evaluation based on score:

Score	PPV	OR+95%[CI]	P-value	Recommended Biliary Evaluation
0	0%	0 [0-0.16]	<0.001	IOC
1	11%	0.33 [0.09-1.1]	0.076	IOC
2	25%	1.05 [0.3-3.4]	1.00	IOC
3	39%	2.30 [0.8-6.4]	0.125	MRCP
4	50%	3.83 [1.3-11.1]	≤0.01	MRCP
5	92%	53.1 [6.8-581.2]	≤0.01	ERCP

*IOC = intraoperative cholangiogram