The performance of organ dysfunction scores for the early prediction and management of severity in acute pancreatitis: an exploratory phase diagnostic study

Abstract

Objective: To evaluate contemporary organ dysfunction scoring systems for early prediction of severity in acute pancreatitis (AP). Methods: In a consecutive cohort of 181 patients with AP, organ dysfunction scores (logistic organ dysfunction system [LODS] score, Marshall organ dysfunction score, and sequential organ failure assessment score) were collected at 24 and 48 hours. Acute Physiology and Chronic Health Evaluation II (APACHE II) scores were calculated on admission and 24 and 48 hours and C-reactive protein level measured at 48 hours. Patients who died or used critical care facilities (level 2/3) during admission were classed as severe. Results: Area under curve for APACHE II score at admission was 0.78 (95% confidence interval, 0.69-0.86). At 24 hours, area under curve for LODS, Marshall organ dysfunction system, sequential organ failure assessment, and APACHE II scores were 0.82, 0.80, 0.80, and 0.82, respectively. The LODS score at cutoff of 1 achieved 90% sensitivity and 69% specificity, corresponding to a positive predictive value of 38%. Acute Physiology and Chronic Health Evaluation II score as a rule-out for selection of mild cases at a test threshold of 9 (scores <= 8 being selected) gives homogeneity of 91% and efficiency of 79%. Conclusions: Contemporary organ dysfunction scoring systems provides an objective guide to stratification of management, but there is no perfect score. All scores evaluated here perform equivalently at 24 hours. Acute Physiology and Chronic Health Evaluation II may have practical clinical value as a rule-out test.