Bacteriology And Quick Screening Model With Its Clinical Use For Symptomatic Bacterascites Based On Cirrhosis Database

Background and objectives: Symptomatic bacterascites(SB)is a variant of spontaneous bacterial peritonitis(SBP)with signs of infection but an ascites polymorphonuclear leukocyte(PMN)count <250/mm3,and it can be only confirmed by a positive ascites culture.It is difficult to diagnose SB as soon as it appears because several days are usually taken to get ascites culture results.To date,early indicators for SB have not been determined.The present study aims to summarize the bacteriology and explore a quick screening model with its clinical application for SB by building a cirrhosis database.Methods: 1.Creation of cirrhosis database Refering to the domestic and international guidelines for management of cirrhosis,content of the database to be collected was determined after the discussion by experts.Then cirrhosis database was created by Epi InfoTM 7 software and clinical data were entered into the database.2.Bacteriology and antibiotic resistance for symptomatic bacterascites Patients with SBP and positive ascites culture were selected from the created database and then divided into SB group and conventional SBP group(ascites PMN cell count ≥ 250/mm3).Differences between the two groups about the bacterial features,antibiotic resistance profiles and prognosis were compared.3.Serum procalcitonin(PCT)and C-reactive protein(CRP)in the screening of symptomatic bacterascites.Eligible patients with cirrhosis were selected from the created database and then divided into SB group and non-infective ascites group.Receiver operating characteristic(ROC)curve and area under curve(AUC)were used to assess the screening value of PCT and CRP for SB.4.Building and clinical application of quick screening model for SB.Eligible patients with cirrhosis were selected from the created database and then divided into SB group and non-infective ascites group.Independent variables were screened out by univariate analysis and binary logistic regression,and then used to build a quick screening model for SB.New hospitalized patients with cirrhosis and ascites were screened by the model above,and then the positive ones were divided into test group(treated with antibiotics)and control group(treated without antibiotic)randomly.Therapeutic response rate,disease progression and prognosis were compared prospectively between the two groups.Results: 1.Creation of cirrhosis database The created cirrhosis database consisted of essential information interface,medical history data interface,complaints and physical examination interface,laboratory examination interface,imaging and endoscopy interface,medical treatment interface,intervention therapy interface,disease evaluation and prognosis interface,complications interface and discharge follow-up interface.Functions such as data entry,data self checking,automatic calculation and so on,were available in this database.Currently,clinical data of total 568 hospitalized patients with cirrhosis were enrolled into the database.2.Bacteriology and antibiotic resistance for symptomatic bacterascites In total,103 patients were enrolled in SB group and 110 patients were enrolled in conventional SBP group.SB cases were mainly caused by Gram-positive bacteria(55.3%,53/103)and conventional SBP cases were mainly caused by Gram-negetive bacteria(71.8%,79/110),and the difference was significant(c2=16.18,P<0.01).The resistance grade of third generation cephalosporin without enzyme inhibitor was safety for Gram-positive bacteria and alarm for Gram-negetive bacteria in SB group,and that was safety for Gram-positive bacteria but no empirical use for Gram-negetive bacteria in conventional SBP group.Mortality rates at 30 days follow-up were 36.9%(38/103)and 43.6%(48/110)in SB group and conventional SBP group,respectively,and the difference was not significant(c2=1.005,P=0.316). 3.Serum procalcitonin(PCT)and C-reactive protein(CRP)in the screening of symptomatic bacterascites.In total,30 patients were enrolled in SB group and 51 patients were enrolled in non-infective ascites group.In the diagnosis of SB,the AUCs were 0.725,0.848,0.737 and 0.806 for PCT,CRP,series test and parallel test,respectively,and there was no significant difference between the any two AUCs(P>0.05).The optimal cutoff value of PCT was ≥ 0.43ng/ml and that of CRP was ≥ 12.76mg/L.For PCT,CRP,series test and parallel test,according to the optimal cutoff values,the sensitivities were 70.0%,70.0%,53.3% and 86.7%,respectively,and the specificities were 76.5%,88.2%,94.10% and 74.5%,respectively.4.Building and clinical application of quick screening model for SB In total,103 patients were enrolled in SB group and 204 patients were enrolled in non-infective ascites group retrospectively.Clinical data of these patients were analysed and then a quick screening model for SB was constructed based on body temperature,abdominal tenderness,blood neutrophil percentage,blood total bilirubin,prothrombin time,and ascites nucleated leukocyte count.The AUC of this model for diagnosing SB was 0.939,and a screening score ≥ 0.328 was the best cutoff value for diagnosing SB with a sensitivity of 86.4% and a specificity of 92.2%.Prospectively,12 patients were enrolled in test group and 12 patients were enrolled in control group.In test group,6 patients were treated with ceftazidime throughout the study,and another 6 received antibiotic adjustment for poor response.At the endpoint of this study,therapeutic response rates were 66.7%(8/12)and 16.7%(2/12),respectively,and the difference was significant(Fisher exact test,P=0.036).Incidence rates of events including ascites PMN cell count ≥ 250/mm3,ascites nucleated leukocyte count ≥250/mm3,positive ascites culture,confirmed SB,fever with abdominal pain,abdominal tenderness and/or rebound pain,and hepatic encephalopathy,were all lower in test group than those in control group,however,no difference was significant(P>0.05).No death event occurred in the two groups.Conclusion: 1.Building cirrhosis database with Epi Info TM 7 software is feasible.Doctors can carry out clinical researches efficiently with the assistance of this database.2.Bacterial features and antibiotic resistance in SB are both different from that in conventional SBP,and the prognosis of SB is similar to that of conventional SBP.With regard to initial empirical therapy,third generation cephalosporin without enzyme inhibitor is suitable for SB but is not a reasonable choice for conventional SBP.3.Serum PCT,CRP and combination of the two seem to be satisfactory diagnostic biomarkers for patients with SB,and each of them has its own superiority.Which one to be chosen should depend on clinical practice needs.4.Quick screening model for SB can identify patients with SB efficiently.Positive patients according to this model will benefit from antibiotic therapy.