Patient selection and timing of dynamic computed tomography in acute pancreatitis.

Abstract

Dynamic computed tomography (CT) is the most accurate method for determining the extent of necrosis in acute pancreatitis. Debate exists, however, regarding patient selection and the optimal timing of CT. This study examined selection based on biochemical and/or clinical criteria and the influence of delayed dynamic CT performed 5-10 days after the onset of an attack. A total of 120 patients with acute pancreatitis were studied. Dynamic CT was performed if any of the following criteria were identified: a biochemically severe attack (according to Glasgow criteria) in the first 24 h, C-reactive protein (CRP) level above 120 mg/l in the first 3 days or failure of clinical resolution within 7 days. Of 42 patients selected for CT, five died from multisystem organ failure before day 5. There were no deaths or delayed complications in the 78 patients not selected for scanning. Positive Glasgow criteria alone in the scanned group had a sensitivity for predicting necrosis (as recognized by CT) of 22 per cent and a specificity of 20 per cent. Measurement of CRP level alone had a sensitivity of 26 per cent and specificity of 80 per cent. Failure of clinical resolution had a sensitivity of only 7 per cent but a specificity of 100 per cent. The combination of Glasgow criteria and CRP level had a sensitivity of 44 per cent and specificity of 100 per cent. Delayed complications (pseudocyst, five; infection, six) occurred only in patients with necrosis, and there were two deaths. An 'at-risk' group can be identified for CT on the basis of biochemical and clinical observations. Neither routine nor emergency dynamic CT in acute pancreatitis seems justified.