A simplified method for computed tomographic estimation of prognosis in acute pancreatitis

Background: Computed tomography is valuable for the diagnosis of acute pancreatitis. Although CT-based prognostic scoring systems are available, they are complex and impractical for routine clinical use. We examined the validity of a simplified CT-based scoring system in a cohort of patients with acute pancreatitis. Method: Observational study based on correlation of CT findings with clinical outcomes. Seventy patients admitted to the Royal Infirmary of Edinburgh from January 1991 to December 1997 with a diagnosis of acute pancreatitis and undergoing CT with intravenous contrast during the first 3 to 10 days after admission were included in the study. Results: Multivariate logistic regression analysis demonstrated that the finding of mesenteric oedema and free peritoneal fluid on CT were independent early predictive factors of adverse outcome. Allocating one point each for either mesenteric oedema (MO) or peritoneal fluid (P) (giving a maximum score of 2), a simple MOP score was derived. Compared with the Glasgow and APACHE multiple-factor scoring systems and the Helsinki and Balthazar CT-based scoring systems areas under ROC curves were: admission Apache II 0.57, admission Glasgow 0.62, Balthazar score 0.79, Helsinki score 0.85 and MOP score 0.87. Conclusions: The presence of mesenteric oedema or peritoneal fluid on CT appears to be a simple and widely applicable predictor of disease severity in acute pancreatitis.