The Predictive Value Of Non-invasive Nomogram Model For First Variceal Bleeding In Liver Cirrhosis

Objective: Esophagogastric varices bleeding(EVB)is a common complication in patients with cirrhosis,which is critical and has a high mortality rate.Prophylactic treatment of EVB in clinical practice is still a challenging problem,Identifying the risk of developing a first EVB and developing individualized treatment programs is an important part of the management of patients with cirrhosis.NOW,the endoscopy and the measurement of HVPG(hepatic venous pressure Gradient)are still the gold standard for assessing the degree of varices and portal vein pressure,which can calculate esophagogastric varices bleeding.However,both gastroscope and hepatic venous pressure gradient are invasive operations,especially the HVPG,which is difficult to promote clinical application and the acceptance of patients is poor.Therefore,this study in oder to establish a non-invasive nomogram for the first bleeding of esophageal and gastric varices in patients with liver cirrhosis and verify the model.Methods: The Clinical data of patients with liver cirrhosis without gastrointestinal bleeding were collected in the Department of Infection and Gastroenterology of Southwest Medical University from January 2016 to May2021.The patients were divided into modeling group and verification group in a ratio of 7:3.And the modeling group was further divided into the bleeding group and the non-bleeding group according to the occrrence of bleeding followed up for at least 1 year.In the modeling group,Using the univariate and multivariate logistic regression to analyze the risk factors of first bleeding of esophageal and gastric varices in patients with cirrhosis,and were used for constrction of the nomogram prediction model.The area under ROC curve,sensitivity and specificity in the modeling group and validation group were used to evaluate the prediction accuracy of the model,evaluating the measures of dispersion between calibration curve and standard curve,and the clinical practicability by the DCA curve.Results: A total of 358 liver cirrhosis patients were included,including254 patients in the modeling group and 104 patients in the validation group.Child-Turcotte-Pugh(CTP),neutrophil-lymphocyteratio(NLR)ratio,hemoglobin and platelet count had a good correlation with EVB(CTP:OR=1.287,NLR:OR=1.145,HGB:OR= 0.971,PLT:OR= 0.982).Based on CTP,HGB,PLT and NLR,the area under ROC curve(AUC)of the model group was0.823(95% CI: 0.771-0.876)The sensitivity is 0.75 and the specificity is 0.77,the area under ROC curve(AUC)was 0.812(95% CI: 0.729-0.894),sensitivity is 0.69 and the specificity is 0.83 in the validation group.The distribution of the calibration curve and the standard curve is relatively consistent in the train set(mean error 0.017),the uniformity was slightly weaker on the validation set in the validation set(mean error 0.049),This model has been proved to have a relatively good clinical practicability by DCA curve analysis,the probability threshold of 0.423 in the train set and 0.379 in the validation set,and the model had an acceptable net benefit.Conclusion: Compared with the non-bleeding group,they had higher CTP score,higher NLR ratio,and lower HGB and platelet count,There was no statistical difference between 2 groups.Child-Turcotte-Pugh,NLR,hemoglobin and platelet count were independent risk factors for the EVB in patients with cirrhosis.The Non-invasive nomogram model is helpful for clinical identification of high-risk patients with the first bleeding of esophageal and gastric varices in patients with cirrhosis.Both internal and external validation demonstrate that the model has good prediction accuracy and clinical application efficacy.