Analysis Of Risk Factors And Etiological Characteristics For Patients With Liver Cirrhosis Complicated With Spontaneous Bacterial Peritonitis

Objective:To explore the risk factors and etiological characteristics of spontaneous bacterial peritonitis（SBP）associated with ascites in patients with cirrhosis,so as to detect SBP early and rationally apply antibiotics to reduc the risk of death.Methods:Retrospective analysis of 488 cases of cirrhosis ascites hospitalized in Quanzhou first hospital from April 2009 to October 2019.They were divided into the SBP group（n=272）and non-SBP group（n=216）accroding to the presence of SBP.The SPSS 25.0 statistical software was used to perform a univariate analysis of the clinical data of the two groups,and the statistically different data was further subjected to multivariate logistic regression analysis to obtain the independent risk factors of cirrhosis ascites with SBP.A binary logistic regression model was established based on independent risk factors and their regression coefficients,and the predictive value of the model was determined by applying the receiver operating characteristic curve（ROC curve）.While the distribution characteristics of pathogenic bacteria isolated from patients with SBP and the drug sensitivity of commonly used antibacterial drugs were analyzed.Results:1.The results of univariate analysis showed that there existed sixteen factors that were statistically significant（P<0.05）between two groups,including previous history of SBP,upper gastrointestinal bleeding,liver failure,sepsis,peripheral blood leukocytes,peripheral blood neutrophil percentage,platelets,serum albumin,serum total bilirubin,prothrombin time,prothrombin activity,International standard prothrombin time,ascites poly morphonuclear（PMN）,ascites albumin,end-stage liver disease model score（MELD score）,Child-pugh grade.2.The results of multivariate logistic regression analysis of meaningful single factors showed that only previous history of SBP,upper gastrointestinal bleeding,liver failure and elevated PMN counts were associated with the occurrence of SBP.3.Based on the above four independent risk factors and their regression coefficients,a binary logistic regression model was established.The equation of model Y is as follows:Y=1.451×（previous history of SBP:1 for yes,0 for no）+0.511×（complicated upper gastrointestinal bleeding:1 for yes,0for no）+0.669（complicated liver failure:1 for yes,0 for no）+1.282×ascites PMN count(×10⁹·L^-1)-1.547.The area under the subject’s working（AUC）predictived by this model is similar to ascites PMN[0.823（95%CI:0.786-0.856）vs.0.854（95%CI:0.820-0.884）,P>0.05].When Y=-1.43,the sensitivity and specificity of the prediction of SBP were 91.91%and 64.81%,respectively.When the ascites PMN count was≥0.25×10⁹·L^-1,the sensitivity and specificity of predicting SBP were 40.07%and97.69%,respectively.4.The positive rate of ascites bacterial culture was18.56%（49/264）.A total of 55 strains of pathogens were isolated,including54.6%（30/55）strains of Gram-negative bacteria,43.6%（24/55）strains of Gram-positive bacteria,and 1.8%.（1/55）strains of fungi.Among the strains detected in the past 5 years,the proportion of Gram-positive bacteria was 51.22%,and Gram-negative bacteria was48.78%.The bacteria producing extended spectrum beta-lactamases（ESBLs）include Escherichia coli And Klebsiella pneumoniae,the production rate of ESBLs of Escherichia coli was 28.57%（4/14）and the production rate of ESBLs of Klebsiella pneumoniae was 25.0%（1/4）.ESBLs-producing strains detected in the past 5 years,the detection rate of ESBLs+strains produced by Escherichia coli was 44.4%,and the detection rate of ESBLs+strains produced by Klebsiella pneumoniae was 33.3%.Conclusion:1.Previous history of SBP,upper gastrointestinal bleeding,liver failure,and elevated PMN counts are the independent risk factors of cirrhosis with ascites accompanied by SBP.2.The prediction model of SBP constructed based on the above four risk factors is similar to the prediction efficiency of ascites PMN count,but the two have higher sensitivity and specificity respectively,and the combination of the two can effectively predictive SBP.3.It′s difficult to separate bacteria in patients with SBP,and the positive bacteria obtained in clinical practice are main Gram-negative bacilli,but the proportion of Gram-positive bacteria has increased in recent years,while the ESBLs had an increasing trend.