| PartⅠ:Development and validation of a novel non-invasive assessment model for liver fibrosis in patients with chronic hepatitis BBackground:Chronic Hepatitis B(CHB)is a serious global public health problem.It is reported that there are still 86 million of chronic hepatitis B virus(HBV)infection in China,and about 400,000 people die each year from HBV-related complications,including cirrhosis,liver cancer or liver failure.Antiviral therapy is an effective measure to block liver fibrosis,reduce liver cirrhosis,liver cancer or liver failure.According to CHB guidelines,patients should be considered for antivirus therapy when they have moderate to severe fibrosis(METAVIR≥F2),though ALT is normal and HBVDNA is positive.Therefore,it is important to stage fibrosis in CHB patients.Liver biopsy is the gold standard for staging fibrosis.However,its limitations prevent its wide application.In recent years,although non-invasive liver fibrosis has made great progress,most of the fibrosis models are built on the basis of chronic hepatitis c,and the applicability of CHB is still controversial.Object:We aimed to development and validation an non-invasive liver fibrosis model for CHB patients by analyzing routnely available clinical indicator including 4 fibrosis markers(LN、HA、PⅢNP、CIV).Therefore,we will provide a new assessment method for the liver fibrosis staging in CHB patients.Methods:1.Patients with CHB who were treatment naive and who underwent liver biopsy at the First Affiliated Hospital,College of Medicine,Zhejiang University,from 1 January 2012 to 31 January 2019 were enrolled.Patients who were treatment between January 2012 and 30 september2017 were enrolled in training group.And Patients who were treatment between 1 October 2017 to 31 January 2019 were enrolled in validation group.The demographic,clinical and laboratory data were recorded,including age,gender,basic diseases,height,weight,and laboratory data(mainly including blood routine,biochemistry,HBVDNA,AFP,coagulation function and 4 fibrosis markers).The histopathological results of the patients were reviewed by the same experienced pathologist to determine the liver fibrosis stageing.2.Clinical characteristic between training and validation group,between significant fibrosis and insignificant fibrosis in training group were compared by Student’s t-test or nonparametric test(Mann-Whitney test).Liver fibrosis predictors were analyzed using univariate and multivariate logistic regression.And fibrosis model was then developed by multivariate logistic regression.3.The diagnostic value of fibrosis model was assessed by the area under the receiver operating characteristic(ROC)curves.And then compare to other models:FIB-4,APRI,Fibroindex,Fom,Lok and GPR.4.The new fibrosis model was applied to the validation set to test for accuracy by measuring the areas under the ROC curves.And also compare to other models:FIB-4,APRI,Fibroindex,Fom,Lok and GPR.5.The accuracy of the new model was assessed in all the CHB patients and compared with liver fibrosis factors(LN,HA,PIIINP,CIV).6.The cut-off of the new model were established,sensitivity and specificity,positive predictive value and negative predictive value were calculated,and tested in the validation group.Results:1.A total of 339 CHB patients were enrolled in this study,223 patients were enrolled in training group,and 116 patients were enrolled in validation group.There was no significant difference in age,gender,BMI and basic diseases between the two groups(P>0.05).According to METAVIR score system,339 CHB patients were classified into fibrosis staging,including 74 patients with F0,155 patients with F1,67 patients with F2,31 patients with F3,and 12 patients with F4.There was no significant difference in the distribution of liver fibrosis between the two groups(P>0.05).2.Of the 223 patients in the training group,75 had significant fibrosis or above.Compare to CHB patients with insignificant fibrosis,CHB patients with significant fibrosis had higher AFP,ALT,AST,r-GT,HA,PIIINP,CIV,HBcAb values,lower PLT,ALB,A/G,CHE values(P<0.05).After univariate and multivariate logistic regression analysis,we found that AST,HA,PIIINP were the predictors of significant fibrosis(P<0.05).Using previous studies for reference,PLT is also selected for multivariate logistic regression analysis to obtain the new fibrosis model:-3.928+0.015×AST(U/L)+0.028×HA(ng/ml)+0.089×PⅢNP(ng/ml)-0.006×PLT(109/L).In order to facilitate the calculation,the model is simplified into the following formula:Fibrosis=(AST(U/L)×HA(ng/mL)×PⅢNP(ng/mL))/(PLT(109/L)×1000)3.With the progression of fibrosis,the value of the new fibrosis model presented a gradually increasing in the training group(P<0.05).The AUROC of the new fibrosis moedl(Fibrosis)for the diagnosis of significant fibrosis in training group was 0.826(0.770-0.874)had a significantly greater AUROC than did the APRI,FIB-4,Fibroindex,Forn,GPR,Lok models(P<0.05).The AUROC of the new fibrosis model(Fibrosis)for the diagnosis of advanced fibrosis in training group was 0.737(0.674-0.793),had a significantly greater AUROC than did the APRI,FIB-4,Forn GPR,Lok models(P<0.05),tended greater than Fibroindex(P=0.0598).The AUROC of the new fibrosis model(Fibrosis)for the diagnosis of liver cirrhosis fibrosis in training group was 0.740(0.677-0.796),had a significantly greater AUROC than did the Forn models(P<0.05),but comparable with the APRI,FIB-4,Fibroindex,GPR,Lok models(P>0.05).4.With the progression of fibrosis,the value of the new fibrosis model presented a gradually increasing in the validation group(P<0.05).The AUROC of the new fibrosis model(Fibrosis)for the diagnosis of significant fibrosis in validation group was 0.878(0.804-0.931),had a significantly greater AUROC than did the APRI,FIB-4,Fibroindex,Forn,Lok models(P<0.05),also tender greater than GPR(P=0.0506).The AUROC of the new fibrosis model(Fibrosis)for the diagnosis of advanced fibrosis in validation group was 0.895(0.825-0.945),had a significantly greater AUROC than did the Fibroindex model(P<0.05),also tender greater than FIB-4(P=0.0639),but comparable with the APRI,Forn,GPR,Lok models(P>0.05).The AUROC of the new fibrosis model(Fibrosis)for the diagnosis of liver cirrhosis in validation group was 0.895(0.825-0.945),comparable with the APRI,FIB-4,Fibroindex,Forn,GPR,Lok models(P>0.05).5.The AUROC of the new fibrosis model(Fibrosis)for the diagnosis of significant fibrosis in all 339 CHB patients was 0.843(0.800-0.880),had a significantly greater AUROC than did LN,HA,PIIINP and CIV(P<0.05).The AUROC of the new fibrosis model(Fibrosis)for the diagnosis of advanced fibrosis in all 339 CHB patients was 0.767(0.718-0.811),had a significantly greater AUROC than did LN,HA(P<0.05),also tender greater than CIV(P=0.0501),and comparable with PIIINP(P=0.2345).The AUROC of the new fibrosis model(Fibrosis)for the diagnosis of liver cirrhosis in all 339 CHB patients was 0.807(0.761-0.848),had a significantly greater AUROC than LN,HA,PIIINP(P<0.05),and comparable with CIV(P=0.7425).6.Based on the ROC of Fibrosis in training group,we defined the cut-off of fibrosis staging.Fibrosis≤0.13 was suggested to exclude significant fibrosis with negative predictive value(NPV)of 93.10%and sensitivity of 92.00%.Fibrosis>0.45 was suggested to predict significant fibrosis with positive predictive value(PPV)of 75.56%and spesificity of 92.57%.Using cutoff values of ≤0.13 and>0.45,51.57%of patients could be accurately excluded or diagnosed as significant fibrosis.Fibrosis≤0.17 as suggested to exclude advanced fibrosis with negative NPV of 94.55%and sensitivity of 80.65%.Fibrosis>0.54 was suggested to predict advanced fibrosis with PPV of 32.43%and spesificity of 86.98%.Using cutoff values of<0.17 and>0.54,52.02%of patients could be accurately excluded or diagnosed as advanced fibrosis.Fibrosis≤0.30 was suggested to exclude liver cirrhosis with NPV of 98.69%and sensitivity of 71.42%.Fibrosis>3.02 was suggested to predict liver cirrhosis with PPV of 50.00%and spesificity of 99.07%.Using cutoff values of ≤0.30 and>3.02,69.50%of patients could be accurately excluded or diagnosed as liver cirrhosis.Cut-off was used in the validation group(116 patients).We found that:57.76%(67 patients),59.48%(69 patients),and 71.55%(83 patients)of CHB patients could be accurately excluded or diagnosed as significant fibrosis,advanced fibrosis and liver cirrhosis,respectively.Conclusions:The new Fibrosis model can evaluate the liver fibrosis stage in patients with CHB,especially significant fobrosis,which is superior to the Fibrosis models(APRI,FIB-4,Fibroindex,Forn,Lok,GPR)and the traditional direct serum markers of liver fibrosis(LN,HA,PIIINP,CIV).Application of this model may decrease the need for staging liver fibrosis by liver biopsy,and provide basis for the treatment of CHB.Part Ⅱ:Accuracy of International Guidelines for Identifying Significant Fibrosis in Patients with chronic Hepatitis BBackground:Hepatitis B virus infection can cause a range of clinical manifestations,from asymptomatic hepatitis B surface antigen(HBsAg)carriers with histologically normal to HBV-induced liver fibrosis,cirrhosis,liver cancer or liver failure.Therefore,it is essential to dynamically assess the disease activity and progression of patients with HBV-infected.And ALT and HBVDNA are the common assessment indicators.Although the three guidelines(APASL,EASL,AASLD)have not reached a unified level of ALT and HBV DNA in initiating antiviral therapy,they have reached a unified level in liver histopathology(≥F2).Although these guidelines do not state that HBV DNA and ALT be used as markers of underlying fibrosis,this rationale is implicit in the recommendation for initiating antiviral therapy.Object:We aimed to assess the accuracy of the combined ALT and HBV DNA criteria within the 3 international guidelines in identifying significant fibrosis in patients with chronic hepatitis B(CHB).Methods:1.Patients with CHB who were treatment naive and who underwent liver biopsy at the First Affiliated Hospital,College of Medicine,Zhejiang University,from 1 January 2012 to 31 March 2018 were enrolled.The demographic,clinical and laboratory data were recorded,including age,gender,basic diseases,height,weight,and laboratory data(mainly including blood routine,biochemistry,HBVDNA,AFP,coagulation function).The histopathological results of the patients were reviewed by the same experienced pathologist to determine the liver fibrosis stageing.2.Patients were grouped according to the criteria of the 3 international guidelines.Patients were classified by fulfillment of APASL,EASL and AASLD guideline criteria.APASL group:patients with HBVDNA>20000IU/ml(HBeAg positive),HBVDNA>2000IU/m 1(HBeAg negative),and elevated ALT above twice the upper limited normal(ULN)(ULN as 40U/L).EASL group:patients with HBVDNA>2000IU/ml,andelevated ALT above the ULN(ULN as 40U/L).AASLD group:patients with HBVDNA>20000IU/ml(HBeAg positive),HBVDNA>2000IU/ml(HBeAg negative),and ALT twice the ULN or above(ULN as 35U/L for male,25U/L for female).3.Clinical characteristic of CHB patients between significant fibrosis and insignificant fibrosis,and between patients who fulfilled any guideline criteria and did not fulfill any guideline criteria,were compared byStudent’s t-test or Nonparametric test(Mann-Whitney test).Clinical characteristic of CHB patients among the APASL,EASL and AASLD group were compared by one-way ANOVA or Nonparametric test(Mann-Whitney test).4.The accuracy of the 3 international guidelines for identifying significant fibrosis in patients with CHB was assessed by the area under the receiver operating characteristic(ROC)curves.And the sensitivity and specificity,positive predictive value and negative predictive value were calculated.Results:1.A total of 601 CHB patients were enrolled in this study.Overall,101 patients(16.81%)were classified by the APASL guideline criteria,258(42.93%)by the EASL criteria,and 150(24.96%)by the AASLD criteria.171 patients(28.45%)reached to significant fibriosis(≥F2),of which 41 patients(23.98%),89 patients(52.05%)and 59 patients(34.50%)could be accurately diagnosed by APASL,EASL and AASLD guidelines.2.Compared with patients with insignificant fibrosis,the patients with significant fibrosis had lower HBsAg titers,higher HBcAb titers,higher ALT,AST,r-GT and alkaline phosphatase(AKP)levels,lower albumin,albumin-globulin ratio,cholinesterase(CHE)and PLT levels,higher AFP levels and higher INR and PT values(P<0.05).Compared with patients who did not fulfill any criteria,the patients who fulfilled any guideline criteria were younger,had a higher male proportion,had more HBeAg positive patients,had more serious liver function impairment,and had more patients with significant fibrosis(P<0.05).Compared with patients who did not meet APASL,EASL or AASLD guidelines,patients who met APASL,EASL or AASLD guidelines had a higher proportion of significant fibrosis(P=0.003,P=0.004,P=0.001).3.The AUC of the 3 international guidelines for the diagnosis of significant fibrosis(≥F2)was 0.559(0.519-0.600)for APASL,0.578(0.538-0.618)for EASL and 0.581(0.540-0.621)for AASLD,respectively.The accuracy of AASLD guidelines for identifying significant fibrosis was better than APASL guidelines and EASL guidelines(P<0.05),and EASL guidelines was better than APASL guidelines(P<0.05).4.APASL guidelines can accurately diagnose 41 patients(23.98%)or exclude 370 patients(86.05%)of fibrosis.The sensitivity,specificity,positive predictive value and negative predictive value are 23.98%,86.05%,40.59%and 74.00%,respectively.EASL guidelines can accurately diagnose 89 patients(52.05%)or exclude 261 patients(60.70%)of fibrosis.The sensitivity,specificity,positive predictive value and negative predictive value are 52.05%,60.70%,34.50%and 76.09%,respectively.AASLD guideline can accurately diagnose 59 patients(34.50%)or patients 339 patients(78.84%)of fibrosis.The sensitivity,specificity,positive predictive value and negative predictive value are 34.50%,78.84%,39.33%and 75.17%.Conclusions:Although the 3 international guidelines(APASL,EASL and AASLD guidelines)have different antiviral criteria(HBVDNA and ALT levels),they can reflect the significant fibrosis of patients with CHB(AASLD works best),especially the superiority in eliminating significant fibrosis. |
PDF Full Text Request