Albumin–bilirubin score is associated with in-hospital mortality in critically ill patients with acute pancreatitis

Objective Acute pancreatitis is one of the most common causes of inpatient care among gastrointestinal conditions. Few easy and commodious biomarkers are used in clinical practice to predict the prognosis of patients with acute pancreatitis. This study was aimed at examining the association of albumin–bilirubin (ALBI) score and in-hospital mortality in critically ill patients with acute pancreatitis. Methods We enrolled all critically ill patients with acute pancreatitis retrospectively in Monitoring in Intensive Care Database III database. Clinical data and demographic information were collected for each patient in our study. Multivariate logistic regression models and smooth curve fitting were used to determine whether ALBI score could be an independent indicator for the prognosis of patients with acute pancreatitis. Predictive performance of ALBI was assessed by receiver operating characteristic analysis. Results A total of 284 patients with acute pancreatitis met the inclusion criteria, and 35 patients died in hospital. The ALBI in nonsurvived group was much higher than survived group (−1.0 vs. −1.5; P < 0.001). The association of ALBI and in-hospital mortality was almost linear by smooth curve fitting ( P < 0.001) and positive associations were observed between ALBI and RDW and WBC in patients with acute pancreatitis. Multivariate logistic regression indicated ALBI could be independent risk factors to predict the prognosis of patients with acute pancreatitis (odds ratios = 1.60; P = 0.02). The area under curve of in-hospital mortality prediction (0.86; P < 0.001) were superior to Sequential Organ Failure Assessment (SOFA) score (0.72; P < 0.001), Simplified Acute Physiology Score II (SAPS-II) (0.71; P < 0.001), Acute Physiology and Chronic Health Evaluation II (APACHE-II, 0.83; P < 0.001), Ranson score (0.75; P < 0.001) and Glasgow score (0.72; P < 0.001). Conclusion ALBI could be a useful marker of in-hospital mortality for patients with acute pancreatitis, which was better than SOFA, SAPS-II, APACHE-II, Ranson score and Glasgow score in our study.