Bedside score predicting retained common bile duct stone in acute biliary pancreatitis

BACKGROUND Retained common bile duct (CBD) stone after an acute episode of biliary pancreatitis is of paramount importance since stone extraction is mandatory. AIM To generate a simple non-invasive score to predict the presence of CBD stone in patients with biliary pancreatitis. METHODS We performed a retrospective study including patients with a diagnosis of biliary pancreatitis. One hundred and fifty-four patients were included. Thirty-three patients (21.5%) were diagnosed with CBD stone by endoscopic ultrasound (US). RESULTS In univariate analysis, age (OR: 1.048, P = 0.0004), aspartate transaminase (OR: 1.002, P = 0.0015), alkaline phosphatase (OR: 1.005, P = 0.0005), gamma-glutamyl transferase (OR: 1.003, P = 0.0002) and CBD width by US (OR: 1.187, P = 0.0445) were associated with CBD stone. In multivariate analysis, three parameters were identified to predict CBD stone; age (OR: 1.062, P = 0.0005), gamma-glutamyl transferase level (OR: 1.003, P = 0.0003) and dilated CBD (OR: 3.685, P = 0.027), with area under the curve of 0.8433. We developed a diagnostic score that included the three significant parameters on multivariate analysis, with assignment of weights for each variable according to the co-efficient estimate. A score that ranges from 51.28 to 73.7 has a very high specificity (90%-100%) for CBD stones, while a low score that ranges from 9.16 to 41.04 has a high sensitivity (82%-100%). By performing internal validation, the negative predictive value of the low score group was 93%. CONCLUSION We recommend incorporating this score as an aid for stratifying patients with acute biliary pancreatitis into low or high probability for the presence of CBD stone.