Analysis of risk factors and construction of a predictive model for severe acute pancreatitis complicated by sinistral portal hypertension.

Abstract

Objective: Sinistral portal hypertension (SPH) is a common complication of severe acute pancreatitis (SAP). Patients with SPH often present asymptomatic, but are at risk of gastrointestinal bleeding and abdominal bleeding due to the presence of varices of the corresponding vessels, which are often fatal. However, there is no prediction model for SAP combined with SPH. This study aimed to identify the risk factors of SAP combined with SPH and to construct a relevant predictive model using independent risk factors.

Materials and methods: The clinical data of 431 SAP patients were collected in this study. According to the presence or absence of SPH, the patients were divided into SPH group (n = 126) and non-SPH group (n = 305), and 431 patients were randomly assigned to the training set and validation set. Univariate logistics regression analysis was used to screen out the variables with significant differences, and then backward stepwise regression method was used for multivariate logistic regression analysis to determine the independent risk factors of SAP combined with SPH. Then a prediction model was constructed and represented by a nomogram, and the model was verified by internal validation. Receiver operating characteristic (ROC) curve and calibration curve were used to evaluate the predictive ability and accuracy of the model, and decision curve analysis (DCA) was used to evaluate the clinical value of the model.

Results: Multivariate logistic regression analysis showed that male, MCTSI score, white blood cell count (WBC), and portal venous system vascular lesions (PVPSL) were independent risk factors for SAP complicated with SPH. The area under the working curve (AUC) of the clinical nomogram in the training set was 0.95 (95% CI: 0.92-0.97),and the P value of the Hosmer-Lemeshow test of the calibration curve was 0.969. The AUC in the validation set was 0.98 (95%CI: 0.96-1.00), and the P value of the Hosmer-Lemeshow test of the calibration curve was 0.963. The DCA in the training set and the validation set showed good clinical applicability of the model.

Conclusion: Male, MCTSI score, WBC and PVPSL are independent risk factors for SAP complicated with SPH. The establishment of prediction model for SAP complicated with SPH is of great significance for the prevention and treatment of SPH in clinical practice.