Association between severity and the determinant-based classification, Atlanta 2012 and Atlanta 1992, in acute pancreatitis: a clinical retrospective study.

Recently, the determinant-based classification (DBC) and the Atlanta 2012 have been proposed to provide a basis for study and treatment of acute pancreatitis (AP). The present study aimed to evaluate the association between severity and the DBC, the Atlanta 2012 and the Atlanta 1992, in AP. Patients admitted to our center with AP from January 2007 to July 2013 were reviewed retrospectively. Patients were assigned to severity categories for all the 3 classification systems. The primary outcomes include long-term clinical prognosis (mortality and length-of-hospital stay), major complications (intraabdominal hemorrhage, multiple-organ dysfunction, single organ failure [OF], and sepsis) and clinical interventions (surgical drainage, continuous renal replace therapy [CRRT] lasting time, and mechanical ventilation [MV] lasting time). The classification systems were validated and compared in terms of these abovementioned primary outcomes. A total of 395 patients were enrolled in this retrospective study with an overall 8.86% in-hospital mortality. Intraabdominal hemorrhage was present in 27 (6.84%) of the patients, multiple-organ dysfunction in 73(18.48%), single OF in 67 (16.96%), and sepsis in 73(18.48%). For each classification system, different categories regarding severity were associated with statistically different clinical mortality, major complications, and clinical interventions (P < 0.05). However, the Atlanta 2012 and the DBC performed better than the Atlanta 1992, and they were comparable in predicting mortality (area under curve [AUC] 0.899 and 0.955 vs 0.585, P < 0.05); intraabdominal hemorrhage (AUC 0.930 and 0.961 vs 0.583, P < 0.05), multiple-organ dysfunction (AUC 0.858 and 0.881 vs 0.595, P < 0.05), sepsis (AUC 0.826 and 0.879 vs 0.590, P < 0.05), and surgical drainage (AUC 0.900 and 0.847 vs 0.606, P < 0.05). For continuous variables, the Atlanta 2012 and the DBC were also better than the Atlanta 1992, and they were similar in predicting CRRT lasting time (Somer D 0.379 and 0.360 vs 0.210, P < 0.05) and MV lasting time (Somer D 0.344 and 0.336 vs 0.186, P < 0.05). All the 3 classification systems accurately classify the severity of AP. However, the Atlanta 2012 and the DBC performed better than the Atlanta 1992, and they were comparable in predicting long-term clinical prognosis, major complications, and clinical interventions.