The Value of RANSON Score Combined with BMI in Predicting the Mortality in Severe Acute Pancreatitis: A Retrospective Study

Objective: To explore the value of modified RANSON score in predicting mortality from severe acute pancreatitis (SAP). Methods: In this retrospective study, 461 SAP patients hospitalized from January 2016 to January 2020 were enrolled. AP (acute pancreatitis) patients from our hospital were employed as the training set. In addition, AP patients from the affiliated hospital of Nantong University were set as the validation set. The clinical characteristics of patients were compared between the two sets. The independent risk factors for SAP were determined through logistic regression. Moreover, the risk factors were derived for various prediction models by logistic regression. Multiple methods were adopted to assess the predictive ability of various models. Results: A total of 338 patients were assigned into the training set, while 123 patients were assigned into the validation set. The patients in the training and validation sets showed the consistent distribution trends (P>0.05). In the training set, significant differences between patients in the non-survival and survival groups were BMI, PCT, platelets (PLT), direct bilirubin (DBil) and RANSON scores (P<0.05). In further multivariate analysis, BMI, PCT and RANSON score were found as the independent risk factors for the mortality of SAP (OR=1.12, 1.25, 1.28, 95% CI:1.06-1.19, 1.08-1.44, 1.12-1.47, P<0.05). In the training set and validation set, ROC curve analysis showed that AUC of BMI+RANSON score was 0.778 and 0.789, respectively. In the calibration curve, the fitting degree of RANSON score+BMI and ideal assessment model was 0.975 and 0.854, respectively. The decision curve suggested that the net benefit per patient increased with the lengthening of the RANSON score+ BMI model curve. As revealed by the results of NRI and IDI indicators, RANSON score+BMI was optimized based on RANSON score (P<0.05). Conclusion: BMI+RANSON was confirmed as a modified model effective in predicting the mortality from SAP.