Is it possible to predict the outcome of acute pancreatitis?

Abstract

During a 4.5 year period 318 patients presenting to the Department of Surgery, Queen Mary Hospital, University of Hong Kong have been scored using four disease severity systems. Two scores were calculated; on admission clinical and laboratory data and 48 hours later using the APACHE II, Stevens sepsis score (SSS). Mannheim peritonitis index (MPI) and Elebute and Stoner (MRC) system. Scores of the patients who died or who suffered systemic complications were statistically significantly higher than in those who had a uncomplicated recovery using all four systems. The 48 hour data showed the larger differences. Accuracy, sensitivity and specificity for deaths were between 60-100% using cut off points suggested by the originators of the scoring systems. Again 48 hour data tended to be more accurate and the APACHE II system was the most useful overall. Similar results were found in prediction of complications but sensitivity was unreliable. As in other scoring systems which use clinical or laboratory derived parameters, these four disease severity systems can identify patients who are at risk. Patients who need intensive resuscitation and treatment can be selected early but with no greater accuracy than by using systems specifically designed for acute pancreatitis.