| Objective:Based on the key factors involved in the pathophysiological process of acute pancreatitis(AP),including oxidative stress,inflammation and metabolic status,a model for early prediction of AP severity was established to provide some reference for the early diagnosis and treatment of AP and the improvement of prognosis.Methods:A retrospective analysis was performed on AP patients treated in the Department of Gastroenterology,Affiliated Hospital of Yangzhou University from January 2017 to December 2020.Patients were divided into mild acute pancreatitis group(MAP group),moderately severe acute pancreatitis group(MSAP group)and severe acute pancreatitis group(SAP group)according to the Atlanta classification.The general clinical data and laboratory indicators were analyzed by single factor analysis.General clinical data included the patient’s age,sex,Body Mass Index(BMI),smoking history,alcohol consumption history,past medical history,AP etiology,and length of hospitalization.Laboratory indicators included oxidative stress related indicators:retinol-binding protein(RBP)and superoxide dismutase(SOD).Inflammation related indicators:white blood cell(WBC),neutrophil(NEUT),monocytes(MONO),Creactive protein(CRP),blood platelet(PLT),lactate dehydrogenase(LDH).Metabolism related indicators:hemoglobin(HB),total bilirubin(TB),direct bilirubin(DB),albumin(ALB),alanine aminotransferase(ALT),aspartate aminotransferase(AST),glutamyl transpeptidase(GGT),alkaline phosphatase(ALP),glucose(GLU),blood urea nitrogen(BUN),creatinine(Cr),Triglyceride(TG),cholesterol(CHO),lipoprotein a(Lp),high-density lipoprotein(HDL),low-density lipoprotein(LDL),serum calcium(Ca),serum natrium(Na),serum potassium(K),and serum chlorine(CI).The laboratory indicators with statistical significance were selected for multifactor Logistic regression analysis,and the independent predictors of MSAP+SAP were screened to establish the Logistic regression prediction model.Check the calibration and differentiation of the prediction model.The evaluation value of AP severity was compared between the prediction model and the common scoring system(MCTSI score,BISAP score,and Ranson score).To evaluate the value of the prediction model in evaluating the severity of AP caused by different etiologies(acute biliary pancreatitis,alcoholrelated acute pancreatitis,hypertriglyceridemia acute pancreatitis).To evaluate the predictive model for systemic inflammatory response syndrome(SIRS)and organ failure(OF).Results:1.A total of 286 AP patients were enrolled in the study,including 210 MAP patients,49 MS AP patients and 27 SAP patients.The length of hospitalization in MSAP group and SAP group was significantly longer than that in MAP group(P<0.05).There was no statistical significance in age,sex,BMI,smoking history,alcohol consumption history,past medical history or AP etiology among the three groups(P>0.05).According to different etiological classification,66 patients with acute biliary pancreatitis,30 patients with alcohol-related acute pancreatitis,and 145 patients with hypertriglyceridemia acute pancreatitis were included in this study,and there was no significant statistical difference in disease severity among the three groups(P>0.05).2.Oxidative stress related indicators SOD and RBP,inflammation related indicators WBC,NEUT,MONO,CRP and LDH,metabolism related indicators GLU,Ca and CI were correlated with the severity of AP(P<0.05).3.Multivariate Logistic regression analysis showed that SOD,NEUT,LDH,GLU and Ca were independent predictors of AP severity(P<0.05).4.Logistic regression model based on oxidative stress,inflammation and metabolism was established by combining SOD,NEUT,LDH,GLU and Ca to predict MSAP and SAP.Prediction model:4.882-0.014SOD(u/ml)+0.014NEUT(×109/L)+0.007LDH(U/L)+0.147 GLU(mmol/L)-3.158Ca(mmol/L).5.Hosmer-Lemeshow test(goodness of fit)was used to evaluate the calibration degree of the prediction model,and the result showed that the goodness of fit effect was good.There was no statistical difference between the actual clinical observation and the prediction model.6.The area under ROC curve was used to evaluate the differentiation of prediction models.The AUC of the prediction model to evaluate the severity of AP was 0.818(95%CI 0.768-0.861),the best cutoff value was-1.17,the sensitivity was 82.89%,and the specificity was 72.38%.7.By comparing the area under the curve between the prediction model and the MCTSI score,BISAP score,and Ranson score on the severity of AP,we found that the prediction model had no difference in evaluating AP severity compared with MCTSI score and BISAP score,and was significantly better than Ranson score.It shows that the prediction model has good value for evaluating the severity of AP.8.The prediction model has good value in evaluating the severity of AP caused by different etiologies.The AUC of the prediction model to evaluate the severity of acute biliary pancreatitis was 0.848(95%CI 0.738-0.924),the sensitivity was 94.12%,and the specificity was 69.39%.The AUC of the prediction model to evaluate the severity of alcohol-related acute pancreatitis was 0.881(95%CI 0.710-0.970),the sensitivity was 100%,and the specificity was 63.64%.The AUC of the prediction model to evaluate the severity of hypertriglyceridemia acute pancreatitis was 0.845(95%CI 0.776-0.900),the sensitivity was 85.00%,and the specificity was 74.29%.9.The prediction model has a good value in evaluating the adverse prognosis of AP.The AUC of the prediction model to evaluate the SIRS was 0.731(95%CI 0.676-0.782),the sensitivity was 67.19%,and the specificity was 69.82%.The AUC of the prediction model to evaluate the OF was 0.844(95%CI 0.797-0.884),the sensitivity was 93.94%,and the specificity was 67.19%.Conclusion:1.The prediction model based on oxidative stress,inflammation and metabolism established in combination with SOD,NEUT,LDH,GLU and Ca has certain reference value for evaluating the early severity and adverse prognosis of AP.2.Prediction model:4.882-0.014SOD(u/ml)+0.014NEUT(×109/L)+0.007LDH(U/L)+0.147GLU(mmol/L)-3.158Ca(mmol/L). |
PDF Full Text Request