Establish A Nomogram For Early Prediction Of Severe Acute Pancreatitis And Prognostic Analysis Based On R Language

Objective: Establish and verify a visual nomogram with early predictive value for the first onset of severe acute pancreatitis(SAP)and explore prognostic-related risk indicators.Methods: Retrospectively included 1860 patients with first acute pancreatitis(AP)who were admitted to the Affiliated Hospital of Southwest Medical University from January 2013 to December 2020.The R software caret package was used to divide the patients into the original cohort and the 7:3 ratio.In the verification cohort,patients in the two cohorts were divided into nonsevere acute pancreatitis group(NSAP group)and severe acute pancreatitis group(SAP group)according to the 2012 Atlanta grading standard,divided into death group and non-death group according to whether the patient died on the first day of admission.Statistic and analyze the general data and laboratory examination results of the two groups of patients in the original cohort,and perform intergroup comparisons,single-factor and multi-factor Logistic regression analysis of the included relevant clinical indicators,and obtain regression models.Spearman method was used to analyze the correlation between indicators and scoring system.R software was used to visualize the models to obtain the columns.Line graph,and verify the model separately in the original queue and the verification queue.Results:(1)There were 1302 people in the original cohort and 558 people in the verification cohort.There was no significant difference in baseline data between the two cohorts.(2)The comparison between groups in the original cohort and the univariate Logistic regression analysis showed that diabetes,hypertension,creatinine,urea,white blood cell count(WBC),red blood cell distribution width(RDW),C-reactive protein(CRP),albumin(ALB),D-Dimer(D-Di),Apolipoprotein B(Apo B),Apolipoprotein A1(Apo A1),and BMI are related factors for the onset of SAP.Multivariate regression analysis results show WBC,RDW,CRP,D-Di,BMI were independent risk factor for the onset of SAP,and positively correlated with scoring system.Apo A1 and ALB were independent protective factors,and negatively correlated with the scoring system.(3)Construct a nomogram based on multivariate regression analysis.If the total score of the nomogram exceeds 120 points,there is a possibility of SAP,and the total score of more than 200 points may cause SAP as high as 90%.(4)In the verification of the original cohort,the consistency index of the nomogram was 0.946(95%CI,0.930~0.0.962).The calibration curve showed that the nomogram had good calibration ability.The verification results of the verification cohort were basically consistent with the original cohort.The ROC curve displayed nomogram in the original cohort had better ability to predict SAP than the MCTSI,Ranson,and BISAP scoring systems.The decision curve analysis showed that the clinical benefit of the nomogram was higher than that of the MCTSI,Ranson,and BISAP scoring systems.Verify the ROC curve displayed in the cohort.The ability of nomogram to predict SAP was better than that of MCTSI scoring system,which was equivalent to Ranson and BISAP scoring system.Decision curve analysis showed that the clinical benefit of nomogram was higher than MCTSI scoring and was equivalent to Ranson and BISAP scoring system.(5)Multivariate regression analysis showed that HCT and CRP were independent risk factors for death in SAP patients within 24 hours after admission,and ALB was an independent protective factor.Conclusion:(1)WBC,RDW,CRP,D-Di,BMI,Apo A1,ALB are all good predictors of the onset of SAP,there was a certain correlation with MCTSI,Ranson,BISAP scoring system.(2)The nomogram established in this study can effectively predict the onset of SAP,which is simpler and more accurate than other scoring systems,but requires external verification.(3)HCT,CRP,and ALB may be good predictors of early death in SAP patients.