Changes Of Antiviral Therapy In Hospitalized Patients With Chronic HBV Infection During The Past 16 Years And The Clinical Characteristics Of Recurrent Chronic Hepatitis B After Therapy Withdrawal With Stringent Cessation Criteria

Background and AimsApproximately 240 million persons are chronically infected with hepatitis B virus(HBV)in the world.Persistent replication of HBV is the major cause of liver cirrhosis,hepatic decompensation and hepatocellular carcinoma(HCC).In China,an estimated 100 million people are chronically infected with HBV and 20 million are chronichepatitis B(CHB)patients.For CHB patients,the most important thing is antiviral therapy.With the emergency and application of more and more antiviral drugs including interferon and nucleos(t)ide analogues,the efficacy and safety is better in CHB and hepatitis B related liver cirrhosis(LC-B)patients.International and domestic organizations recommend and update guidelines for management of chronic hepatitis B,In real world,the proportion of antiviral therapy and the actual application of antiviral drugs are not satisfied.The corresponding changes of proportion and application about antiviral treatment in CHB and LC-B patients in our region are lacking.Hepatitis B e antigen(HBeAg)negative chronic HBV infection is a worldwide increase in recent ten years,but the difference in the prevalence rate is very large and long-term longitudinal studies are lacking.Most HBeAg seroconverters remain in sustained remission,but HBeAg-negative hepatitis occurs in the remaining HBeAg seroconverters;the latter is a critically important subgroup in which progression of liver disease often continues.In fact,the majority of cirrhosis complications and HCC occur in this population of HBeAg-negative CHB.Furthermore,comparing with HBeAg-positive patients,longer time of treatment and higher recurrent rate after therapy withdrawal were found in HBeAg-negative CHB patients.Thus,more attention should be paid in HBeAg-negative CHB and LC-B patients.Nucleos(t)ide analogues(NAs)are administered in CHB and LC-B patients because of their convenient oral administration,powerful antiviral potency and safety.The ideal therapeutic endpoints in both HBeAg-positive and HBeAg-negative patients are loss or seroconversion of hepatitis B surface antigen(HBsAg).However,NAs cannot directly affect and eliminate covalently closed circular DINA,and thus long-term,even lifelong,NAs therapy is usually required.Indefinite therapy,however,can result in poor adherence,unaffordable costs,and unpredictable safety issues.Due to the possible finite duration of the etiological treatment,international organizations including the Asian Pacific Association for the Study of the Liver(APASL)recommend rules for the cessation of NAs in CHB patients.However,in real life and clinical practice,relapse has often been observed in CHB patients despite the adoption of stringent cessation criteria.Furthermore,stopping NA therapy before attaining ideal criteria may increase the fear of the risk of reactivation,cirrhosis,liver decompensation,liver failure,HCC and developing resistance.Most previous studies have focused on the durability of NAs after treatment cessation and the risk factors of relapse.However,no study focused on the clinical characteristics and safety profile of relapsers systematically and in detail,including the distribution of relapse time,the difference between early and late relapsers,the appropriate follow-up time,the degree of abnormal liver function,HBeAg reversal or not,the necessity and efficacy of re-antiviral therapy intervention,the possibility of liver failure and decompensation,etc.In fact,the above characteristics are still of great concerns to the clinicians,esp.to those in endemic areas.ln addition,in real life and clinical practice,we found a subgroup of HBeAg-positive patients who cannot experience e antigen seroconversion and only appear e antigen seroclearance,even if HBsAg disappear.For this subgroup of patients,we have also conducted exploratory studies after the approval of ethical committee and informed consent of patients.In this study,a systematic investigation and analysis was made about the changes of HBeAg-negative proportion and antiviral therapies in the past 16 years in our region.Furthermore,based on the now more than ten years of follow-up data,the objective of this study was to investigate the clinical characteristics and outcomes of recurrent CHB patients who met cessation criteria and experienced relapse after NAs withdrawal.These results will provide reference for antiviral therapy of local CHB and LC-B patients and evidence for domestic and international guidelines about optimal follow-up time and safety profile of relapsers.MethodsA retrospective survey was investigated in hospitalized CHB and LC-B cases from 2001,2006,2011 and 2016.All cases met the diagnostic criteria for CHB or LC-B,and patients with coexisting HCV,HDV or HIV infection,alcoholism,autoimmune hepatitis,or drug induced liver damage were excluded.The results of all the liver function and HBV DNA detection were abnormal.The study about drug withdrawal enrolled a prospective cohort of patients who met the corresponding cessation criteria and discontinued NAs treatment since 2001.All patients met the criteria for diagnosis and treatment outlined by 2008 APASL guidelines.For HBeAg seroconversion and HBeAg-negative patients,the cessation criteria met 2008 APASL guidelines.For HBeAg seroclearance,the cessation criteria were defined as follows:HBeAg seroclearance with undetectable HBV DNA and normal alanine aminotransferase(ALT)documented on 2 separate occasions at least 6 months apart and a duration of total therapy ≥18 months.In total,312 CHB patients who met the cessation criteria discontinued oral antiviral agents(including lamivudine,adefovir dipivoxil,telbivudine and entecavir).After cessation of oral antivirals,Patients’ liver function and HBV DNA were evaluated every month for the first 4 months and at month 6,9,12 and then once every 3-6 months thereafter.One hundred and twenty-five CHB patients who met the relapse criterion(viral relapse:defined as serum HBV DNA>104 copies/mL in two consecutive samples at least two weeks apart)were included in this study.Early relapse was defined as viral relapse confirmed within 3 months after cessation and late relapse as more than 3 months after cessation.Statistical analyses were performed using Student’s t-test,Analysis of Variance,Mann-Whitney test,chi-square test,Fisher’s exact test,logistic regression analysis as appropriate.Differences were considered statistically significant at P<0.05.Results1.The result about 16 years survey in hospitalized CHB and LC-B cases showed that a total of 5355 cases were included,and LC-B cases accounted for 34.6%(272/787),28.9%(455/1575),42.4%(675/1591),44.7%(626/1402)in 2001,2006,2011 and 2016,the difference was statistically siganificant(χ2 ＝99.086,P<0.001).A further comparation showed that the proporation of LC-B in 2011 and 2016 was higher than that of 2001 and 2006(χ2 =21.206,79.679 and 13.590,63.193,all P<0.001).In CHB patients,the proportion of HBeAg-negative cases was about 22-25%since 2001,but the proportion in 2016(31.4%)was higher than that of 2001,2006 and 2011(χ2 =6.955,21.491 and 12.740;P =0.008,0.000 and 0.000).In LC-B patients,the proportion of HBeAg-negative cases has increasing year and year.The proportion in 2016(69.6%)was higher than that of 2001(42.3%),2006(52.3%)and 2011(51.6%)(χ2 =59.905,33.752 and 44.399;all p<0.001).There was no statistical significance between 2006 and 2011 in the proportion of HBeAg-negative cases(χ2＝0.062,P＝0.804),and the proportion in 2006 and 2011 was higher than that of 2001(χ2=6.854,6.676;P =0.009,0.010).In all LC-B patients,the median age was 50(15-82)and 52(11-80)in HBeAg-positive and HBeAg-negative patients,respectively(Z=-4.955,P<0.001);In all CHB patients,the median age was 33(1-73)and 41(5-77)in HBeAg-positive and HBeAg-negative patients,respectively(Z =-12.781,P<0.001).2.The proportion of antiviral therapy was all increasing in CHB,LC-B,HBeAg-positive and HBeAg-negative patients in the past 16 years.The proportion of antiviral therapy had reached 95%in 2016.The proportion of antiviral therapy in CHB(56.1%and 74.3%)was higher than that of LC-B(39.0%and 62.9%)in 2001 and 2006(χ2=20.931,20.518;all P<0.001),but the proportion of antiviral therapy in CHB(91.3%and 94.8%)was lower than that of LC-B(94.1%and 97.1%)in 2011 and 2016(χ2=4.393,4.539;P =0.036,0.033).In CHB patients,the proportion of antiviral therapy in HBeAg-negative was lower than that of HBeAg-positive patients in 2001,2006 and 2011,the results was 42.5%vs.60.6%,65.4%vs.76.8%,87.1%vs.92.6%,respectively(χ2 =12.655,12.892,6.202;P =0.000,0.000,0.013),but there was no statistical difference in 2016(95.1%vs.94.7%)(χ2=0.041,P=0.840).In LC-B patients,the proportion of antiviral therapy in HBeAg-negative was lower than that of HBeAg-positive patients in 2001,2006 and 2011,the results was 30.4%vs.45.2%,53.8%vs.72.8%,91.1%vs.97.2%,respectively(χ2 =6.104,17.606,11.459;P =0.013,0.000,0.001).but higher proportion was found in HBeAg-negative in 2016(98.6%)than that of HBeAg-positive patients(93.7%)(χ2 =11.562,P=0.001).In CHB patients,no matter HBeAg-positive or HBeAg-negative,the proportion of lamivudine(34.9%,29.4%.7.3%,0.8%and 48.1%,37.9%,13.2%.0%)was decreasing(χ2＝274.370,132.428;all P<0.001),and oral first-line drugs(entecavir or tenofovir)(0%,4.5%,24.6%,52.4%and 0%,5.0%,26.5%,55.2%)was increasing(χ2=459.687,144.253;all P<0.001)in the past 16 years.In LC-B patients,no matter HBeAg-positive or HBeAg-negative.the proportion of lamivudine(100%,63.3%,16%,2.2%and 100%,74.2%,20.2%,4.7%)was decreasing(χ2＝336.098,383.815;all P<0.001),and oral first-line drugs(entecavir or tenofovir)(0%,7.6%,40.6%,88.8%and 0%,7.0%,30.9%,79.5%)was increasing(χ2=289.595,325.653;all P<0.001)in the past 16 years.3.A total of 125 patients relapsed after discontinuing treatment and were eligible for this analysis.In the HBeAg seroconversion group,the number of relapse cases after discontinuing lamivudine(LAM),adefovir dipivoxil(ADV),telbivudine(LdT)and entecavir(ETV)were 29,4,4 and 1,respectively;in the HBeAg-negative group,there were 38,9,0 and 2 cases,respectively;and in the HBeAg seroclearance group,there were 23,9,3 and 3 cases,respectively.The median follow-up time after relapse was19,26 and 31 months,and the longest follow-up time after drug withdrawal was all more than 10 years in each group.In the HBeAg seroconversion group and HBeAg-negative group,the earliest relapse time points were month 2 and month l,the latest relapse time points were month 134 and month 96,respectively.For the two groups,the most common relapse time point were both month 2.In the HBeAg seroconversion group,44.7%,65.8%,76.3%and 89.5%of the relapse cases occurred within 3 months,6 months,12 months and 48 months,respectively,while in the HBeAg-negative group,44.9%,51.0%,77.6%and 91.8%occurred within 3 months,6 months 12 months and 36 months,respectively.The median relapse times of HBeAg seroconversion and HBeAg-negative patients were 4 months(2-134)and 6 months(1-88),respectively.There was no statistical difference between the two groups in relapse time(Z =-0.488,P =0.625).In the HBeAg seroclearance group,the earliest relapse time point was month l,the latest relapse time point was month 30 and the most common relapse time point was month 2.In this group,71.1%,86.8%and 94.7%of the relapse cases occurred within3 months,6 months and 12 months,respectively.Half of the relapse cases occurred within 2 months and no relapse case occurred in 3 years.The median relapse time of HBeAg seroclearance was 2.5(1-30)months,which was higher than HBeAg seroconversion patients(Z =-2.726,P =0.006).4.For the recurrent patients who met the cessation criteria outlined by 2008 APASL guidelines,the time to undetectable HBV DNA in early relapse group[3(1-10)months]was later than that of late relapse group[2(1-7)months](P =0.026).The time to undetectable HBV DNA was a predictive factor of early relapse,odds ratio(OR)was 1.311,95%confidence interval(Cl)was 1.014-1.694,P=0.039,and the optimal cutoff value to discriminate the early relapsers from late ones was 2.5 months(sensitivity was 59%,specificity was 64.6%).For those with a longer time(>2.5 months)to undetectable HBV DNA,more relapsers were found within 3 months after cessation.5.Of the 87 recurrent patients meeting the cessation criteria outlined by 2008 APASL guidelines,viral relapses were accompanied by elevated ALT in 70(80.5%)patients.Peak ALT five times over the upper limit of normal(ULN)after relapse was observed in 34.2%of the HBeAg seroconversion and 40.8%of the HBeAg-negative patients.Hepatic decompensation and liver failure were not observed.Different salvage treatments were administered according to the patients’ drug history,the available drugs,the severity of hepatic injury,patients’ economic status,and patient willingness.Of the 62 patients receiving antiviral retreatment,61 showed complete viral response within 3(1-9)months,and the remaining 1 patient showed a partial viral response and later LAM resistance.Of note,LAM resistance was observed at 10(8-38)months in 11 of the 22 lamivudine retreatment patients.6.Of 23 patients with abnormal ALT and no antiviral retreatment,ten patients obtained spontaneous viral remission 3.5(1-19)months later.The clinal characteristics and laboratory data including age at treatment cessation,gender,pretreatment ALT,pretreatment HBV DNA,total treatment duration,time to undetectable HBV DNA and relapse time were comparable between spontaneous viral remission and non-spontaneous viral remission.Only no family history was the predictive factor of spontaneous viral remission.7.In HBeAg seroclearance group(exploratory group),viral relapses were accompanied by elevated ALT in 24(63.2%)patients.A peak ALT five times over ULN after relapse was observed in 55.3%of the HBeAg seroclearance patients.Compared to HBeAg seroconversion group,earlier relapse time[2.5(1-30)months vs.4(2-134)months,Z=-2.726,P=0.006]and higher proportion of HBeAg reversion cases(78.9%vs.28.9%,χ2＝19.119,P=0.000)were observed in HBeAg seroclearance.Tending to be higher peak ALT[248.5(20-682)IU/L vs.121.5(35-1022)IU/L,Z =-1.766,P =0.077]and higher proportion of peak ALT five times over ULN after relapse[55.3%vs.34.2%,χ2=3.406,P=0.065]were observed in HBeAg seroclearance.8.The factors including age at treatment cessation,gender,pretreatment ALT,initial treatment,family history,total treatment duration,time to undetectable HBeAg were comparable between HBeAg reversion and non-HBeAg reversion.Univariate analysis revealed that HBeAg reversion was related to pretreatment HBV DNA(7.38± 0.67 log10 copies/ml vs.6.97 ± 0.93 log10 copies/mL,t =2.174,P=0.034),time to undetectable HBV DNA[3(1-15)months vs.2(1-8)months,Z =-2.843,P=0.004],relapse time[3(1-30)months vs.4(2-134)months,Z =-2.280,P =0.023]and proportion of HBeAg seroconversion(26.8%vs.77.1%,χ2=19.119,P=0.000).Multivariate analysis revealed that pretreatment HBV DNA and HBeAg seroconversion were predictive factors of HBeAg reversion.The higher the pretreatment HBV DNA and the lower the proportion of HBeAg seroconversion,the higher the HBeAg reversion occurred.9.The factors including age at treatment cessation,gender,pretreatment ALT,initial treatment,pretreatment HBV DNA,family history,total treatment duration,consolidation treatment duration,time to undetectable HBV DNA and relapse time were comparable between abnormal ALT at viral relapse and normal ALT at viral relapse.Univariate analysis revealed that abnormal ALT at viral relapse was related to HBV DNA at relapse(5.66 ± 1.20 log10 copies/ml vs.4.94 ± 0.82 log10 copies/mL,t=-3.725,P=0.000)and baseline positive HBeAg(54.3%vs.77.4%,χ2＝5.212,P=0.022).The multivariate analyses showed that abnormal ALT at viral relapse was related to pretreatment HBeAg status.That means baseline negative HBeAg patients were more likely accompanied by abnormal ALT after relapse.The factors including age at treatment cessation,gender,pretreatment ALT,initial treatment,pretreatment HBV DNA,pretreatment HBeAg status,family history,total treatment duration,consolidation treatment duration,time to undetectable HBV DNA and relapse time were comparable between hepatitis flare after relapse and non-hepatitis flare after relapse.Univariate analysis revealed that hepatitis flare after relapse was related to ALT at viral relapse[107(16-818)IU/L vs.61(16-184)IU/L,Z＝-3.791,P ＝0.000]and the time of re-antiviral therapy after relapse[1(0-24)months vs.0(0-10)months,Z =-2.514,P＝0.012].The multivariate analysis showed that hepatitis flare after relapse was related to ALT at viral relapse.Further,receiving operating characteristic curve(ROC)showed that ALT cutoff value was 64.5 IU/L,the area under the curve was 0.698,95%Cl was 0.603-0.793,P<0.001,the sensitivity was 68.5%,and the specificity was 53.5%.That means the higher the ALT at viral relapse,the higher possibility of the peak ALT over 5 ULN,and ALT at viral relapse>64.5IU/L was more likely to become hepatitis flare.10.The factors including age at treatment cessation,gender,pretreatment HBeAg status,family history,ALT and HBV DNA at viral relapse,the proportion of abnormal ALT and peak ALT after viral relapse was comparable between LAM treatment and ADV treatment groups.The lower pretreatment HBV DNA(6,37 ± 1.22 log10 copies/mL vs.6.97 ± 0.96 log10 copies/mL,t =2.460,P=0.015),lower pretreatment ALT[143(83-660)IU/L vs.233(80-916)IU/L,Z =-1.802,P =0.072],less initial treatment(27.3%vs.78.9%,χ2=21.923,P=0.000)and longer total treatment duration[48(24-94)months vs.28(12-65)months,Z=-4.774,P=0.000]were observed in ADV group.Compared to LAM group,longer time to undetectable HBV DNA[3(1-15)months vs.3(1-9)months,Z=-2.500,P =0.012]and earlier viral relapse[2(1-44)months vs.4(1-134)months,Z=-2.109,P=0.035]were observed in ADV treatment patients.Conclusions1.During the past 16 years in our region,the proportion of LC-B cases was increasing year by year in hospitalized patients with chronic HBV infection.The proportion of HBeAg-negative cases in CHB patients was still in the minority.However,in LC-B patients,the proportion of HBeAg-negative cases was in the majority and it was increasing year by year during the past 10 years.2.In our region,the proportion of antiviral therapy was all increasing in CHB,LC-B,HBeAg-positive and HBeAg-negative patients in the past 16 years.Professional doctors have paid more attention to antiviral therapy of chronic HBV infection,especially LC-B patients and HBeAg-negative patients.The proportion of high efficiency and high-genetic barrier agents(oral first-line drugs)was increasing especially in LC-B patients.3.Different distribution of relapse time and clinical characteristics were observed in three groups(including HBeAg seroconversion,HBeAg-negative and HBeAg seroclearance).For the three groups,the most common relapse time point were all month 2,the earliest relapse time was month 2,month 1 and month 1,respectively.90%of relapse cases occurred within 48 months.36 months and 12 months,respectively.Thus at least 4 years,3 years and 1 year of close follow-up are necessary.4.For those patients who met cessation criteria outlined by 2008 APASL guidelines,time to undetectable HBV DNA was the only predictive factor of early relapse.For those with a longer time(>2.5 months)to undetectable HBV DNA,more attention should be paid to the early stages(within 3 months)after cessation.5.For recurrent patients meeting the cessation criteria outlined by 2008 APASL guidelines and recurrent patients in HBeAg seroclearance group(exploratory group),NA withdrawal is generally safe,although close monitoring and timely intervention are needed.Patients with ALT at viral relapse>64.5IU/L were more likely to become hepatitis flare,thus more attention should be paid to these recurrent patients.6.Not all recurrent patients need antiviral retreatment because spontaneous viral clearance can be observed in some patients especially in non-family history ones.7.For pretreatment HBeAg-positive patients,the higher the pretreatment HBV DNA and the lower the proportion of HBeAg seroconversion,the higher the HBeAg reversing to HBeAg-positive occurred.