[The comparison of different clinical scoring systems for predicting prognosis in acute pancreatitis based on the revised Atlanta classification].

OBJECTIVE: To compare the predictive value of BISAP (bedside index for severity in acute pancreatitis), APACHE II (acute physiology and chronic health evaluation II), and Ranson scoring system in persistent organ failure (POF) and mortality in patients diagnosed as acute pancreatitis (AP) based on the revised Atlanta classification.

METHODS: Demographic, clinical and laboratory data of 350 consecutive AP patients admitted to the First Affiliated Hospital of Nanchang University were prospectively collected from November, 2009 to January, 2012. A retrospective analysis was performed and 310 patients finished the follow-up. The median age of whole population was (50.5 ± 16.4) years old. Patients were classified into early phase group ( ≤ 7 days) and late phase group ( > 7 days) based on the interval between onset of AP and admission. Demographics and clinical data were collected to calculate Ranson, APACHE II and BISAP scores during the first 3 days of hospitalization. Poor prognosis was defined as POF or death.

RESULTS: The three scoring systems similarly demonstrated modest accuracy for predicting POF or death in early phase group [area under the receiver operating characteristic curve (AUCROC):0.68-0.84], but failed to predict the prognosis of AP patients in late phase group. Daily scoring of APACHE IIand BISAP on the first 3 days after admission demonstrated modest to high predictive accuracy to poor prognosis (AUCROC:0.69-0.95), but this was not statistically significant (P > 0.05) .

CONCLUSIONS: These three clinical scoring systems show modest accuracy for predicting POF or death in AP patients on the early phase based on the revised Atlanta classification. The BISAP scoring system has similar prognostic value to APACHE II and Ranson. However, due to the simplicity and convenience, BISAP scoring system is more popular in clinical practice. Daily scoring on the first 3 days after admission fails to predict the prognosis accurately.