Objectives

We aimed to compare the prognostic value of various predictors and complex scoring systems for prediction of severe acute pancreatitis (SAP) according to the revised Atlanta classification.

Methods

C-reactive protein (CRP) and procalcitonin were obtained on admission, and CRP level 24 hours after admission (CRP2) was measured. Various scoring systems including Ranson, Acute Physiology and Chronic Health Examination (APACHE II), the Bedside Index for Severity in Acute Pancreatitis, and Computed Tomography Severity Index (CTSI) were calculated.

Results

There were 146 patients with acute pancreatitis (mean age, 50.6±18.3 years; 63% male), of which 43 patients (29.5%) received a diagnosis of moderately severe AP, and 17 patients (11.6%) received a diagnosis of SAP. In patients with moderately severe acute pancreatitis to SAP, CTSI (odds ratio [OR], 10.46; 95% confidence interval [CI], 4.3–25.43; P< 0.001), APACHE II (OR, 3.87; 95% CI, 1.18–12.64; P= 0.025), and CRP2 (OR, 4.5; 95% CI, 1.53–13.1; P= 0.006) were strongly related to moderately severe acute pancreatitis and SAP. In patients with SAP compared with mild to moderately severe AP, procalcitonin (OR, 4.36; 95% CI, 1.01–18.96; P= 0.049) was the only factor strongly associated with SAP.

Conclusions

Procalcitonin was the best predictor for patients with SAP; CTSI, APACHE II, and CRP2 were valuable predictors for patients with moderately severe acute pancreatitis and SAP.