Abstract

The aim of this paper is to evaluate the predictability of various scoring systems and serum markers in the assessment of the severity of acute pancreatitis. All consecutive patients with acute pancreatitis were prospectively studied. Body mass index was measured at the time of admission. The demographic data, etiology, mean hospital; stay, clinical, radiological, biochemical findings, morbidity, and mortality were recorded. The relations between these parameters, scoring systems (Ranson, APACHE II, Imrie, and various serum markers) and patients' outcome were determined by using appropriate tests. Ninety seven (fifty men and forty seven women) patients were incorporated in the study; mean age was 51 years. Biliary pancreatitis was the most common etiological factor, followed by idiopathic pancreatitis (60 and 29%, respectively). Seventy (72%) patients had severe pancreatitis and 27 (28%) cases had mild disease. Ranson (p=0.2), Glasgow (p=0.4), and APACHE II (p=0.5) appeared insignificant predictors of the severity of acute pancreatitis by multivariate analysis. More reliable serum markers were pancreatic amylase (p ≤ 0.001), neutrophil elastase (p ≤ 0.001), serum albumin (p ≤ 0.02), and C-reactive protein (p ≤ 0.001). Results turned out to be more homogenous when CT scan findings were added together. Not a single parameter achieved statistically significant predictive value when used alone. Ranson, Imrie score, and APACHE II are not accurate predictors of the severity of acute pancreatitis. Serum markers are better predictors to elucidate the severity of disease.

Key words: Acute pancreatitis, Ranson criteria, Imrie score, APACHE II, serum amylase, serum C-reactive protein.