Sustained Responses and Loss of HBsAg in HBeAg-Negative Patients With Chronic Hepatitis B Who Stop Long-Term Treatment With Adefovir

GASTROENTEROLOGY 2012;143:629 636 Sustained Responses and Loss of HBsAg in HBeAg-Negative atients With Chronic Hepatitis B Who Stop Long-Term Treatment With Adefovir STEHANOS J. HADZIYANNIS,*, VASSILIOS SEVASTIANOS,* IRENE RATI,* DIMITRIOS VASSILOOULOS, and EMILIA HADZIYANNIS *Department of Medicine and Hepatology, Henry Dunant Hospital, 2nd Academic Department of Medicine, Hippokration Hospital, and the Molecular Biology Laboratory of the Liver Unit at the Evgenidion Hospital, National and Kapodistrian University of Athens, Athens, Greece BACKGROUND & AIMS: Little is known about the biochemical and virological effects of stopping long-term nucleos(t)ide analogue therapy for hepatitis B e antigen (HBeAg)-negative patients with chronic hepatitis B (CHB). METHODS: We performed a cohort observational study, following 33 HBeAg-negative patients with CHB, undetectable serum HBV DNA, and normal levels of aminotransferases after long-term (4 or 5 years) treatment with adefovir dipivoxil (ADV). All patients were followed for 5.5 years; follow-up visits included measurements of serum alanine aminotransferase (ALT), hepatitis B surface antigen (HBsAg), and HBV DNA monthly for the first 6 months and every 3 6 months thereafter. Various factors were measured at baseline, the end of treatment (EOT), and following treatment to identify those associated with clearance of HBsAg. RESULTS: During the first few months of the postdiscontinuation period, all patients experienced virological and 25 (76%) had biochemical relapse. During the follow-up period, 18 patients (55%) who had discontinued antiviral therapy achieved sustained response (HBV DNA level 2000 IU/L, persistently normal level of ALT). Among these, 13 (72%) cleared HBsAg. Fifteen patients (45%) with virological and/or biochemical relapse were re-treated with oral antiviral agents (11 during the first 18 months and 4 after the third year), without evidence of liver decompensation; only 1 lost HBsAg (6%). Higher pretreatment and EOT levels of ALT, no previous treatment with interferon, and lower level of HBsAg at the EOT were significantly associated with HBsAg clearance based on multivariate analysis. CONCLUSIONS: In HBeAg-negative patients with CHB, it is safe and effective to discontinue ADV therapy after 4 or 5 years; 55% of patients have sustained responses, and 39% of patients lose HBsAg. Keywords: rospective Clinical Trial; Replication Suppression; Liver Disease. Hepatitis B e antigen (HBeAg)-negative chronic hepatitis B (CHB) is currently the predominant type of CHB in Europe, the Mediterranean basin, and many other geographical regions of the world. 1 5 Although it represents an immune active phase in the course of chronic hepatitis B virus (HBV) infection, 6,7 the host immune mechanisms not only fail to eliminate the virus or suppress its replication, but they also lead to progressive immune-mediated chronic liver damage. 8 Ideally, treatment of CHB should be aiming at HBV elimination, but because this is not a goal easily achievable with currently available therapies, a generally accepted treatment approach is potent and durable suppression of HBV replication, which could lead to prevention of cirrhosis and hepatocellular carcinoma (HCC). 9 However, in HBeAgnegative CHB, discontinuation of finite treatments of up to 3-year duration with oral nucleos(t)ide analogues (NUCs) is followed by virological and biochemical relapses in the majority of patients and the benefits gained by therapy are lost. 10 18 Therefore, current treatment guidelines 4,19 for HBeAg-negative CHB recommend longterm/indefinite oral antiviral therapy without development of HBV resistance. 2,20 What would be the outcome of both chronic HBV infection itself and the underlying liver disease if such long-term therapies with NUCs were discontinued? This remains hitherto unknown because such a treatment strategy has not been thus far attempted. On the other hand, long-term follow-up data from cohorts of HBeAgnegative patients treated with conventional or pegylated interferons have shown that sustained biochemical and virological remissions followed by loss of hepatitis B surface antigen (HBsAg) do occur in a good proportion of patients with HBeAg-negative CHB after stopping such treatment. 21 25 In this prospective long-term follow-up study, the rate of biochemical and virological responses, including HBsAg loss, was assessed in a cohort of patients with HBeAg-negative CHB after discontinuation of effective long-term treatment (4 or 5 years) with adefovir dipivoxil (ADV). 26,27 atients and Methods Study Design This was a prospective, cohort, observational study initiated immediately after stopping a 192- or 240-week ADV treatment regimen in patients with HBeAg-negative CHB included in a randomized placebo-controlled trial (GS-98-438), the Abbreviations used in this paper: ADV, adefovir dipivoxil; CHB, chronic hepatitis B; EOT, end of treatment; NUC, nucleos(t)ide analogue; OR, odds ratio; CR, polymerase chain reaction; SR, sustained response; ULN, upper limit of normal. 2012 by the AGA Institute 0016-5085/$36.00 http://dx.doi.org/10.1053/j.gastro.2012.05.039

630 HADZIYANNIS ET AL GASTROENTEROLOGY Vol. 143, No. 3 Figure 1. Flow chart showing the disposition of patients in the different cohorts of the open-label ADV study. results of which have already been published in detail. 14,26,27 The original study design consisted of an initial double-blinded phase of 96 weeks, followed by an open-label, long-term safety and efficacy period of 144 weeks. In this international prospective trial, 185 patients were enrolled initially in the year 2000 at 32 sites. 26 atients were randomly assigned to receive 10 mg ADV (n 123) or placebo (n 62). At week 49, 119 patients initially treated with ADV were randomized either to continue treatment (ADV-ADV group) or switch to placebo (ADV-placebo group) and patients initially given placebo were switched to ADV (placebo-adv group), all in a double-blind manner. From week 96 and on, all patients in the GS-98-438 study continued longterm ADV therapy. Treatment ended at study week 240. Thus, patients in the ADV-ADV group (n 70) received treatment for 240 weeks (5 years) and those in the placebo-adv group (n 55) for up to 192 weeks (4 years); they are herein also referred as cohorts A and B, respectively. At week 240, a liver biopsy was performed for the patients of both cohorts according to the protocol and ADV treatment was terminated. atients Forty-seven patients with HBeAg-negative CHB from Athens, Greece, were randomly allocated to one of the 2 cohorts of long-term ADV treatment (ADV-ADV, n 27; placebo-adv, n 20). From the third year of the long-term ADV trial and on, all patients were treated and followed up in the Liver Unit of the Henry Dunant Hospital of Red Cross of Athens (Figure 1 and Table 1). Ten patients (21%) experienced virological and biochemical breakthrough during ADV therapy and had been entered in rescue protocols, while 37 (79%) achieved longterm HBV suppression (undetectable serum HBV DNA) that was maintained up to the end of treatment (EOT) without development of genotypic HBV resistance. Four of the 37 patients did not agree to stop therapy and continued to receive ADV that was already commercially available. All 33 remaining patients accepted treatment discontinuation and signed an informed consent that also included acceptance of the follow-up protocol of long-term observation and possible re-treatment. Their overall characteristics are depicted in Table 1. There was no significant difference by the Mann Whitney test in the characteristics of the patients in the 2 cohorts of long-term ADV treatment. Follow-up rotocol atients were evaluated every month for the first 6 months and every 3 6 months (or more frequently if clinically indicated) during the posttreatment follow-up period. A complete physical examination and laboratory assessment including Table 1. Number and Characteristics of all atients in the ADV GS-98-438 Study of 5- or 4-Year Duration and Those From Athens Who Gave Informed Consent for Discontinuation of Treatment and Long-term Follow-up All patients in the GS-98-438 study Athens cohort n 125 47 Consent for ADV discontinuation EOT virologic response, n (%) 84 (67) 37 (79) 33 Age (y), median (range) 47 (18 65) 53 (29 69) 52 (29 69) Male/female 103/22 38/9 27/6 Baseline ALT s (U/L), median (range) 99 (24 742) 109 (44 457) 100 (44 408) Serum HBV DNA (log 10 copies/ml), median (range) 7.12 (3.7 9.5) 7.27 (4.8 8.5) 7.15 (4.8 8.4) Knodell necroinflammation score, median (range) 7 (2 17) 7 (4 17) 7 (4 17) Ishak fibrosis score, median (range) 2 (1 4) 3 (1 4) 3 (1 4) 5/4-year treatment duration, n 70/55 27/20 22/11 HBV genotype All D

September 2012 SUSTAINED AND HBsAg LOSS IN CHB RESONSES 631 alanine aminotransferase (ALT), aspartate aminotransferase, alkaline phosphatase, -glutamyl transpeptidase, albumin, globulin, prothrombin time, blood cell count, blood urea nitrogen, creatinine, lipid profile, and glucose were performed in each visit. -Fetoprotein and serologic markers of HBV infection (including a semiquantitative assay of antibody to hepatitis B core antigen [anti-hbc] immunoglobulin M) were repeatedly tested. Ultrasonography of the upper abdomen and power Doppler of the Haller s tripod were performed every 6 months. Laboratory Methods Virological, serological, and biochemical assays were performed by methods described in previous publications of the GS-98-438 study. 14,27 Quantification of serum HBsAg was performed by the Architect analyzer (chemiluminescent immunoassay, Abbott Laboratories, Abbott ark, Illinois, cutoff, 0.04 IU/mL). A Roche real-time HBV DNA polymerase chain reaction (CR) assay (Cobas TaqMan) with a lower limit of quantitation of 169 copies/ml or 29 IU/mL was used during the last year of the GS-98-438 study, at the EOT (as described in detail in previous publications of the GS-98-438 trial), and during the first 4 months of the posttreatment follow-up. 14,27 All available subsequent sera that were initially tested by an in-house real-time CR (detection limit, 1000 copies/ml or 200 IU/mL) were retested by the current automated Cobas Amplirep TaqMan assay (Roche Molecular Systems Inc, leasanton CA) with a lower limit of quantitation of 20 IU/mL. Definitions of Events The events of interest were defined as follows during follow-up. EOT virological response. Undetectable serum HBV DNA (by Cobas TaqMan Real Time CR; detection limit, 29 IU/mL). Virological remission. Decline of serum HBV DNA levels to nondetectability or 2000 IU/mL (by a real-time CR assay with a detection limit of 200 IU/mL as standardized on the basis of the Cobas TaqMan real-time CR test). Virological relapse. Increase in HBV DNA level 2000 IU/mL or increase by 1 log 10 from nadir in 2 consecutive measurements at least 3 months apart. Sustained response. ersistently not detectable HBV DNA or 2000 IU/mL combined with persistently normal ALT s (upper limit of normal [ULN], 40 U/L) following posttreatment month 6 sustained to the end of follow-up. Biochemical relapse. Increase in ALT level 1.2 times ULN. Hepatitis flare. Increase in ALT level 10 times ULN. Criteria for Re-treatment of atients Experiencing a Relapse After Discontinuation of Long-term ADV Therapy After discontinuation of therapy, re-treatment with ADV or other NUCs was offered to patients in biochemical and virological relapse with ALT level 2.5 times ULN persisting for more than 4 months without any spontaneous decline toward normal as well as to patients with posttreatment ALT flares exceeding 10-fold times ULN that persisted for more than 2 months in consecutive assays. For safety reasons, depending on the clinical and laboratory findings, patients were subjected to repeated clinical assessment and to laboratory testing with additional examinations scheduled on an individual basis. Figure 2. The number of patients who experienced a relapse and received antiviral therapy (relapsers) or lost HBsAg among the remaining patients during follow-up. Statistical Analysis Data entry and statistical analysis were performed with the statistical package IBM SSS Statistics (version 19.0; SSS Inc, Chicago, IL). In all cases, tests of significance were 2 tailed with a level at.05. Corrected 2 or 2-tailed Fisher exact test and t test or Mann Whitney test were used for qualitative and quantitative data when appropriate. Logistic regression models were used for multivariate analysis, and the Kaplan Meier method was used to estimate the rates of HBsAg loss and the log-rank as a test for significance. Quantitative variables that were normally distributed were expressed as mean s SD, and those non-normally distributed were expressed as median s (range). Multivariate analysis (backward stepwise logistic regression analysis) of variables with.10 at univariate analysis was performed to identify independent factors associated with sustained response (SR) and HBsAg loss. The study was conducted in compliance with the 1975 Declaration of Helsinki and approved by local regulatory bodies. Authors contributed to the design, execution, analysis, or interpretation of the study, and each author contributed to the writing and approved the final manuscript. Results atients Outcome All 33 patients who had discontinued ADV (Figure 1) completed a 5.5-year period of posttreatment observation, with a median duration of 69 months (range, 67 72 months). Their baseline characteristics in comparison to the initial cohorts of the GS-98-438 study are shown in Table 1. All patients are alive and in excellent general condition, and none of them have developed cirrhosis or HCC. Eighteen patients (55%) achieved sustained biochemical and virological remission with 13 of them also clearing HBsAg (39%), thus resulting in a 5.5-year rate of HBsAg clearance of 72% among sustained biochemical/ virological responders who stopped ADV treatment and of 35% among all 37 patients from Athens with EOT response (Figure 2). atients with HBsAg loss have hitherto continued to be followed up regularly at 6-month intervals. None of them have reversed to HBsAg positivity, and 9 (69%) have stably seroconverted to antibody to HBsAg. It is noteworthy that none of the 4 individuals who did not consent to stop therapy have hitherto experienced HBsAg loss, while one of them has developed

632 HADZIYANNIS ET AL GASTROENTEROLOGY Vol. 143, No. 3 Figure 3. The course of (A and B) ALT and (C and D) HBV DNA levels during the first 12 months off treatment in patients with biochemical relapse (n 25) who either resumed antiviral therapy (n 15; A and C) or remained without antiviral therapy (n 10; B and D). For patients who resumed antiviral treatment during the first year (n 9), ALT and HBV DNA s are shown until the onset of therapy (A and C). In E, the course of HBV DNA in the 8 patients who did not receive antiviral therapy and did not show biochemical relapse is also shown. Undetectable HBV DNA s are assigned a of 1. HCC. HBsAg loss has occurred in 8 of 22 (36.4%) and 5 of 11 (45.5%) patients in cohorts A and B, respectively ( NS). Virological and Biochemical rofiles During Follow-up Serum HBV DNA was negative at EOT in all patients (N 33) but became again detectable in the first month after treatment discontinuation. In the majority of patients (29/33 [88%]), posttreatment serum HBV DNA levels peaked either in the first (n 22) or second posttreatment month (n 7). The posttreatment reappearance of serum HBV DNA was also observed in low levels in 3 patients with hardly detectable HBsAg at EOT ( 5.23 IU/mL), but HBV DNA became again undetectable at month 2 and has remained persistently negative thereafter up to the end of follow-up. In 25 of the 33 patients (76%), HBV reactivation was associated with transient or persisting biochemical relapse (Figure 3). The increase in ALT level occurred either concomitantly (11/25 [44%]) with the reappearance of HBV DNA or 1 2 months later. The duration of the biochemical relapse was transient in most of the patients and by month 9 of follow-up, ALT s had declined to normal levels. In Figure 3, the individual serial ALT and HBV DNA titers among patients with (n 25) or without (n 8) biochemical relapse who were either re-treated (n 15; Figure 3A and C) or not (n 10, Figure 3B and D) are shown. The change in HBV DNA titers is also depicted for the 8 untreated patients who did not show biochemical relapse (Figure 3E). At the end of the first year, 15 patients who did not receive re-treatment with antivirals were already in SR and their number increased to 16 at the end of year 2, while beyond year 3 until the end of follow-up, 18 patients remained in SR (55%; Figure 2 and Supplementary Figure 1). The rate of HBsAg loss among those untreated patients with SR gradually increased from 20% (3/15) in year 1 to 37.5% in year 2 (6/16), 44% in year 3 (8/18), 50% in year 4 (9/18), 55.5% in year 5 (10/18), and 72% at the end of follow-up (13/18; Figure 2 and Supplementary Figure 1). Among the 5 patients with SR who did not lose

September 2012 SUSTAINED AND HBsAg LOSS IN CHB RESONSES 633 HBsAg, 3 had undetectable HBV DNA at the end of follow-up and 2 had low HBV DNA levels ( 2000 IU/mL; Supplementary Figure 1). Treatment of Relapsing atients During follow-up, 15 of the 33 patients (45%) have experienced biochemical and virological relapse (Figures 2 and 3) and have received salvage therapy either with lamivudine or ADV. Six patients during the early follow-up period (asymptomatic, without any laboratory evidence of hepatic decompensation) were re-treated with NUCs immediately after the first increase in ALT level at their own request, without waiting for the results of retesting (as required by the protocol). All of these early treated patients were considered relapsers, and none of them lost HBsAg during long-term follow-up. Nine more patients displayed persistent virological and biochemical relapse during the following 5 years of follow-up (Figure 2). Among the 15 relapsing patients who were re-treated with NUCs for indefinite duration, only one has hitherto achieved HBsAg loss (6%). redictors of HBsAg Loss The following variables were assessed as possible determinants/predictors of HBsAg loss: age, sex, ALT (pretreatment, EOT, peak ALT during the first posttreatment year), previous interferon treatment, serum HBV DNA levels (before ADV treatment), duration of HBV DNA detectability during ADV treatment, liver necroinflammation and fibrosis (pretreatment, EOT), HBsAg (EOT, not enough sera were available for HBsAg titration at baseline), timing of ALT and HBV DNA peak during the first posttreatment year, and duration of ADV treatment (4 or 5 years). These different pretreatment, EOT, and posttreatment variables were examined as potential predictors of HBsAg loss either among the total patient population (N 33 Table 2) or among non re-treated patients only (n 18) (data not shown). Among all patients who discontinued long-term effective ADV treatment (N 33), lower pretreatment HBV DNA levels (odds ratio [OR], 0.99;.04), higher ALT level at EOT (OR, 1.18;.007), lower HBsAg titer at EOT (OR, 1;.02), no previous interferon treatment (OR, 0.07;.002), and no re-treatment with antivirals (OR, 0.03;.002) were associated by univariate analysis with HBsAg loss (Table 2). It should be mentioned that all patients who lost HBsAg were male (14/14) compared with those who remained HBsAg positive (13/19;.03 by Fisher exact test), although this could not be tested by univariate analysis. Using multivariate analysis at the different time points of follow-up (pretreatment, EOT, first posttreatment year), higher pretreatment (OR, 1.02;.027) and EOT (OR, 1.29;.019) ALT levels, lower HBsAg levels at EOT (OR, 0.99;.033), and no re-treatment with antivirals after discontinuation of ADV (OR, 0.027;.002) were associated with HBsAg loss. Using a model in which the most significant variables by multivariate analysis at the different time points were included, none of these variables showed a statistically significant association with HBsAg loss (Table 2). The same analysis was performed in the 18 patients who did not resume antiviral therapy during posttreatment follow up (data not shown). By univariate and multivariate analysis, only ALT level at EOT (OR, 1.31; 95% CI, 1.02 1.69;.035) was associated with HBsAg loss. It should be mentioned that none of the patients who remained HBsAg positive exhibited biochemical relapse during posttreatment follow-up (0/5), compared with 77% of those who lost HBsAg (10/13;.007 by Fisher exact test; Supplementary Figure 1), while all patients who lost HBsAg were male (13/13 vs 3/5;.06 by Fisher exact test). These 2 variables could not be tested by univariate analysis. HBsAg Titers HBsAg levels, measured at the end of long-term ADV therapy, were lower in patients with SR who finally cleared HBsAg (mean SD, 1435 2188; median/range, 161/0.1 5904 IU/mL) compared with relapsers (mean SD, 4794 3367; median/range, 4837/983 11,543 IU/ ml;.002; Figure 4) and patients with SR who did not clear HBsAg (mean SD, 5385 5062; median/range, 5242/124 11,233 IU/mL), although the difference did not reach statistical significance with the latter group (.07; Figure 4). In Supplementary Figure 1, the individual changes in HBsAg titers among non re-treated patients (n 18) are shown both for those who remained HBsAg positive (n 5; Supplementary Figure 1A, C and E) or lost HBsAg (n 13; Supplementary Figure 1B, D and F). Discussion Stopping courses of up to 3-year duration of antiviral treatment with NUCs in HBeAg-negative CHB is followed by biochemical relapses in the vast majority of patients, induced by resumed replicative activity of HBV and reflected by increases in ALT and serum HBV DNA levels, respectively. 10 18 This was also the case in the mother placebo-controlled GS-98-438 study with the ADV-placebo group, in which patients were left untreated despite virological and biochemical relapses, without any significant adverse events. 14 Nevertheless, for safety reasons, stringent criteria were applied in the present study regarding the management of relapses after stopping long-term treatment with ADV, even though all included patients had compensated liver disease, normal liver function test results, and significant improvement in liver necroinflammation and fibrosis. 27 Despite the fact that all 33 patients who were followed up were in long-term remission under ADV treatment, HBV DNA became again detectable in serum in all of them soon after treatment was discontinued. However, the increase in serum HBV DNA level was transient and rather low. These findings are not surprising in view of our previous observations showing that HBV covalently

Table 2. Comparison Between atients Without (n 19) and With HBsAg Loss (n 14) Among the Whole atient opulation (N 33) Characteristic No HBsAg loss (n 19) HBsAg loss (n 14) Univariate analysis Multivariate analysis a Multivariate analysis (.1) b Multivariate analysis c OR (95% CI) OR (95% CI) OR (95% CI) OR (95% CI) Before/during ADV treatment Age 53.7 12.2 55.3 11.71 1.012 (0.95 1.08).701 1.341 (0.90 2.00).151 revious interferon treatment 15/4 3/11.002 0.073 (0.01 0.39).002 0.000 (0.00 2.21).067 0.025 (0.00 0.48).015 0.015 (0.00 0.74).084 (yes/no) ALT (U/L) 121.7 63.7 138.3 103.3.9 1.003 (0.99 1.01).562 1.048 (0.99 1.11).108 1.022 (1.00 1.41).027 1.019 (1.00 1.39).066 HBV DNA (IU/mL), median 6.38 (3.90 6.91) 5.93 (3.93 6.55).06 0.999 (0.99 1.00).043 0.999 (0.99 1.00).095 0.999 (0.99 0.99).033 0.999 (0.99 1.00).094 (range) Liver necroinflammation 7.6 2.5 8.1 2.8.78 1.077 (0.81 1.42).604 2.374 (0.63 8.95).201 (Knodell) d Liver fibrosis (Knodell) d 2.2 1.2 2.5 1.2.44 1.267 (0.69 2.33).447 24.267 (0.36 1637.80).138 2.718 (0.85 8.74).093 9.027 (0.73 111.92).087 Duration of ADV therapy (4/5 6/13 6/8.5 1.625 (0.39 6.82).507 0.001 (0.00 16.90).165 years) Duration of HBV DNA 48.5 6.7 48.6 7.8.78 1.003 (0.91 1.11).945 1.847 (0.85 4.01).122 undetectability (mo) End of ADV treatment (EOT) ALT (U/L) 23.7 6.6 32.6 8.5.007 1.182 (1.05 1.34).007 1.485 (0.92 2.39).103 1.295 (1.04 1.61).019 HBsAg titer (IU/mL) 4905 3783 1733 2457.02 1.000 (0.99 1.00).022 0.999 (0.99 1.00).137 0.999 (0.99 0.99).040 Liver necroinflammation 3.2 1 2.6 0.6.14 0.422 (0.15 1.20).105 0.185 (0.02 2.15).178 (Ishak) e Liver fibrosis (Ishak) e 2.7 0.7 2.5 1.5.68 0.806 (0.40 1.61).540 1.615 (0.47 5.58).449 ost-adv treatment follow-up eak ALT after stopping ADV 175.5 218.5 128.5 110.2.83 0.998 (0.99 1.00).110 1.054 (0.99 1.13).346 (first year) Timing of peak ALT (mo) 3.95 2.88 3.14 3.46.23 0.915 (0.72 1.16).860 1.042 (0.66 1.64).432 eak HBV DNA (IU/mL), 5.15 (3.65 7.11) 4.67 (2.28 5.58).08 0.999 (0.99 1.00).128 0.999 (0.99 1.00).185 median (range) Timing of peak HBV DNA 3.42 3.7 1.43 0.85.09 0.648 (0.36 1.16).344 0.496 (0.12 2.12).123 (mo) Re-treatment after stopping ADV (yes/no) 14/5 1/13.0003 0.027 (0.00 0.27).002 0.000 (0.00 3.92).993 0.027 (0.00 0.27).002 0.002 (0.00 0.29).093 NOTE. A number of variables before, during, at the end of, and after ADV treatment were compared between patients who finally achieved HBsAg loss and those who did not in the whole population of patients (N 33). Normally distributed variables are presented as mean SD, and non-normally distributed variables are presented as median (range). Qualitative data were compared using 2-tailed Fisher exact test and quantitative data using t test or Mann Whitney test (as appropriate), and results are shown in the third column. Results of univariate and multivariate analysis were also performed and s and ORs with 95% confidence intervals (95% CIs) are depicted (in bold;.05). a Results of multivariate logistic regression analysis of HBsAg loss using all available variables for each of the 3 periods (before/during ADV treatment, EOT, and after ADV treatment [first year]) are shown. b Results of multivariate logistic regression analysis of HBsAg loss using only variables with.1 at the multivariate analysis for each of the 3 periods as in a. c Results of multivariate logistic regression analysis of HBsAg loss including all variables with.1. Here all available variables from all periods were taken into account. d Liver biopsy data available for 31 patients before ADV treatment. e Data available for 24 patients at EOT. 634 HADZIYANNIS ET AL GASTROENTEROLOGY Vol. 143, No. 3

September 2012 SUSTAINED AND HBsAg LOSS IN CHB RESONSES 635 Figure 4. EOT HBsAg levels in patients with SR and HBsAg loss (n 13) vs patients with virological and biochemical relapse (n 15;.002) or patients with SR without HBsAg loss (n 5;.07). SR indicates persistently not detectable HBV DNA or 2000 IU/mL combined with persistently normal ALT level following posttreatment month 6 sustained until the end of follow-up. closed circular DNA in the liver, although greatly reduced, still remain detectable despite effective long-term suppression of HBV replication, 28 while the same can be observed even in patients who clear HBsAg. 29 The spontaneous decrease of posttreatment serum HBV DNA levels to low ( 200 IU/mL) or nondetectable levels combined with long-term biochemical remission and subsequent loss of HBsAg indicate that the host immune mechanisms that had been modulated during the long period of HBV suppression by ADV treatment reacted effectively to the resumed HBV replication, eventually clearing those hepatocytes that had started to re-express HBV proteins. Such an immune-mediated clearance of HBV-infected hepatocytes represents the main mechanism of CHB remission and HBsAg loss following interferon alfa therapy. 19,24,30 However, for such an immune clearance to be mounted, the immune system needs not only to become modulated but also to be re-exposed to the antigens of the replicating virus. HBsAg loss is considered the closest to cure outcome of chronic HBV infection and the golden goal of current therapies, particularly in difficult-to-treat patients with HBeAg-negative CHB. 4,19,24 Therefore, our results from the multivariate logistic regression analysis showing that lower HBsAg levels at EOT were associated with a higher probability of future HBsAg loss indicate that low HBsAg levels at EOT could predict a subsequent HBsAg loss. Although a specific cutoff point with a high negative or positive predictive for HBsAg loss could not be established (probably due to the small number of patients), this observation is of practical clinical relevance and needs to be confirmed in future large-scale studies. Among the other pretreatment variables studied, no previous interferon treatment, higher ALT levels, and lower HBV DNA levels were also associated with HBsAg loss in our multivariate logistic regression analysis. Similarly, among the EOT variables, higher ALT levels were associated with HBsAg loss, along with the HBsAg levels. However, the mean ALT s in those who lost HBsAg compared with those who did not were within the normal range. As expected, no re-treatment with antivirals after discontinuation of ADV was the strongest predictor of HBsAg loss among the whole patient population. Among the population of non re-treated patients, only a higher ALT level at EOT was associated with HBsAg loss in the multivariate logistic regression analysis. Collectively, these findings suggest that among patients whose immune system is able to recognize and lyse HBV-infected hepatocytes (probably reflected by a higher ALT level), a significant proportion can eventually achieve HBsAg clearance. Ongoing studies assessing viral expression (total HBV DNA, covalently closed circular DNA, RNA transcripts, HBsAg, hepatitis B core antigen) in the liver as well as host immune responses (serum immunoglobulin M anti-hbc level and type of liver inflammation) during the posttreatment period could provide additional information regarding the pathophysiology and potential predictors of HBsAg clearance in this population. Due to its surprisingly good results, this long-nuc treatment/discontinuation study actually represents an emerging new treatment paradigm that could be carefully applied to patients with HBeAg-negative CHB who have compensated liver disease but not to patients with cirrhosis or advanced liver fibrosis who, in case of severe posttreatment biochemical relapse, may be at risk for development of liver failure. A limitation of this study is the small number of included patients. However, the vast majority of patients from our center who participated in the international long-term mother study (33/37) gave informed consent for stopping therapy and all of them have been followed up for approximately 6 years. Another potential limitation could be that all of our patients had genotype D infection. 7 However, this may be also viewed as an advantage because the outcome of chronic genotype D HBV infection appears to be worse compared with genotype A, which is also encountered in Europe and similar to other HBV genotypes. 7 In conclusion, our study shows that effective 4- to 5-year treatment with ADV followed by drug discontinuation leads to HBsAg loss in approximately 40% of patients with well-compensated HBeAg-negative CHB. This is the highest rate of off-treatment HBsAg loss that has been achieved with any type of antiviral therapy in patients with chronic CHB so far. Supplementary Materials Note: To access the supplementary material accompanying this article, visit the online version of

636 HADZIYANNIS ET AL GASTROENTEROLOGY Vol. 143, No. 3 Gastroenterology at www.gastrojournal.org, and at http:// dx.doi.org/10.1053/j.gastro.2012.05.039. References 1. Funk ML, Rosenberg DM, Lok AS. World-wide epidemiology of HBeAg-negative chronic hepatitis B and associated precore and core promoter variants. J Viral Hepat 2002;9:52 61. 2. Hadziyannis SJ, apatheodoridis GV. Hepatitis B e antigen-negative chronic hepatitis B: natural history and treatment. Semin Liver Dis 2006;26:130 141. 3. Manesis EK. HBeAg-negative chronic hepatitis B: from obscurity to prominence. J Hepatol 2006;45:343 346. 4. European Association for the Study of the Liver. EASL Clinical ractice Guidelines: management of chronic hepatitis B. J Hepatol 2009;50:227 242. 5. Liaw YF, Brunetto MR, Hadziyannis SJ. The natural history of chronic HBV infection and geographical differences. Antivir Ther 2010;15(Suppl 3):2040 2058. 6. Hoofnagle JH, Doo E, Liang TJ, et al. Management of hepatitis B: summary of a clinical research workshop. Hepatology 2007;45: 1056 1075. 7. Hadziyannis SJ. Natural history of chronic hepatitis B in Euro- Mediterranean and African Countries. J Hepatol 2011;55:183 191. 8. Hadziyannis SJ, Vassilopoulos D. Immunopathogenesis of hepatitis B e antigen negative chronic hepatitis B infection. Antivir Res 2001;52:91 98. 9. Kwon H, Lok AS. Does antiviral therapy prevent hepatocellular carcinoma? Antivir Ther 2011;16:787 795. 10. Tassopoulos NC, Volpes R, astore G, et al. Efficacy of lamivudine in patients with hepatitis B e antigen-negative/hepatitis B virus DNA-positive (precore mutant) chronic hepatitis B. Lamivudine recore Mutant Study Group. Hepatology 1999;29:889 896. 11. Santantonio T, Mazzola M, Iacovazzi T, et al. Long-term follow-up of patients with anti-hbe/hbv DNA-positive chronic hepatitis B treated for 12 months with lamivudine. J Hepatol 2000;32:300 306. 12. Huang YH, Wu JC, Chang TT, et al. Analysis of clinical, biochemical and viral factors associated with early relapse after lamivudine treatment for hepatitis B e antigen-negative chronic hepatitis B patients in Taiwan. J Viral Hepat 2003;10:277 284. 13. Fung SK, Wong F, Hussain M, et al. Sustained response after a 2-year course of lamivudine treatment of hepatitis B e antigennegative chronic hepatitis B. J Viral Hepat 2004;11:432 438. 14. Hadziyannis SJ, Tassopoulos NC, Heathcote EJ, et al. Long-term therapy with adefovir dipivoxil for HBeAg-negative chronic hepatitis B. N Engl J Med 2005;352:2673 2681. 15. Chan HL, Wang H, Niu J, et al. Two-year lamivudine treatment for hepatitis B e antigen-negative chronic hepatitis B: a double-blind, placebo-controlled trial. Antivir Ther 2007;12:345 353. 16. Shouval D, Lai CL, Chang TT, et al. Relapse of hepatitis B in HBeAg-negative chronic hepatitis B patients who discontinued successful entecavir treatment: the case for continuous antiviral therapy. J Hepatol 2009;50:289 295. 17. Liang Y, Jiang J, Su M, et al. redictors of relapse in chronic hepatitis B after discontinuation of anti-viral therapy. Aliment harmacol Ther 2011;34:344 352. 18. Liu F, Wang L, Li XY, et al. oor durability of lamivudine effectiveness despite stringent cessation criteria: a prospective clinical study in hepatitis B e antigen-negative chronic hepatitis B patients. J Gastroenterol Hepatol 2011;26:456 460. 19. Lok ASF, McMahon BJ. AASLD ractice Guidelines. Chronic hepatitis B: update 2009. Available at: www.aasld.org. Accessed September 12, 2011. 20. Keeffe EB, Dieterich DT, Han SH, et al. A treatment algorithm for the management of chronic hepatitis B virus infection in the United States. Clin Gastroenterol Hepatol 2004;2:87 106. 21. Manesis EK, Hadziyannis SJ. Interferon alpha treatment and retreatment of hepatitis B e antigen-negative chronic hepatitis B. Gastroenterology 2001;121:101 109. 22. Brunetto MR, Moriconi F, Bonino F, et al. Hepatitis B virus surface antigen levels: a guide to sustained response to peginterferon alfa-2a in HBeAg-negative chronic hepatitis B. Hepatology 2009; 49:1141 1150. 23. Moucari R, Korevaar A, Lada O, et al. High rates of HBsAg seroconversion in HBeAg-positive chronic hepatitis B patients responding to interferon: a long-term follow-up study. J Hepatol 2009;50: 1084 1092. 24. Marcellin, Bonino F, Lau GK, et al. Sustained response of hepatitis B e antigen-negative patients 3 years after treatment with peginterferon alpha-2a. Gastroenterology 2009;136:2169 2179. 25. Rijckborst V, Hansen BE, Cakaloglu Y, et al. Early on-treatment prediction of response to peginterferon alfa-2a for HBeAg-negative chronic hepatitis B using HBsAg and HBV DNA levels. Hepatology 2010;52:454 461. 26. Hadziyannis SJ, Tassopoulos NC, Heathcote EJ, et al. Adefovir dipivoxil for the treatment of hepatitis B e antigen-negative chronic hepatitis B. N Engl J Med 2003;348:800 807. 27. Hadziyannis SJ, Tassopoulos NC, Heathcote EJ, et al. Long-term therapy with adefovir dipivoxil for HBeAg-negative chronic hepatitis b for up to 5 years. Gastroenterology 2006;131:1743 1751. 28. Laras A, Koskinas J, Dimou E, et al. Intrahepatic levels and replicative activity of covalently closed circular hepatitis B virus DNA in chronically infected patients. Hepatology 2006;44:694 702. 29. Levrero M, ollicino T, etersen J, et al. Control of cccdna function in hepatitis B virus infection. J Hepatol 2009;51:581 592. 30. apatheodoridis GV, Manesis E, Hadziyannis SJ. The long-term outcome of interferon-alpha treated and untreated patients with HBeAg-negative chronic hepatitis B. J Hepatol 2001;34:306 313. Received January 24, 2012. Accepted May 24, 2012. Reprint requests Address requests for reprints to: Stephanos J. Hadziyannis, MD, Department of Medicine and Hepatology, Henry Dunant Hospital, 107 Mesogion Avenue, 115 26 Athens 11526, Greece. e-mail: hadziyannis@ath.forthnet.gr; fax: (30) 210-697-2974. Conflicts of interest The authors disclose no conflicts. Funding Sponsored in part by research grants from the Special Account for Research Grants, National and Kapodistrian University of Athens (Athens, Greece), and Gilead Sciences. The sponsors supported financially the performance of the laboratory tests for the first 2 years of the long-term follow-up observation and provided assistance in the statistical analysis of the results.

September 2012 SUSTAINED AND HBsAg LOSS IN CHB RESONSES 636.e1 A. B. 100000.00 10000.00 100000.00 10000.00 HBsAg (IU/L) 1000.00 100.00 10.00 1.00 0.10 HBsAg (IU/L) 1000.00 100.00 10.00 1.00 0.10 0.01 0.01 0 1 2 3 4 6 9 12 18 24 30 36 42 48 54 60 66 0 1 2 3 4 6 9 12 18 24 30 36 42 48 54 60 66 ALT (U/L) C. D. 480 440 400 360 320 280 240 200 160 120 80 40 E. F. ALT (U/L) 0 0 0 1 2 3 4 6 9 12 18 24 30 36 42 48 54 60 66 0 1 2 3 4 6 9 12 18 24 30 36 42 48 54 60 66 480 440 400 360 320 280 240 200 160 120 80 40 100000000.00 10000000.00 1000000.00 100000.00 100000000.00 10000000.00 1000000.00 100000.00 HBV DNA (IU/mL) 10000.00 1000.00 100.00 10.00 1.00 HBV DNA (IU/mL) 10000.00 1000.00 100.00 10.00 1.00 0 1 2 3 4 6 9 12 18 24 30 36 42 48 54 60 66 0 1 2 3 4 6 9 12 18 24 30 36 42 48 54 60 66 Supplementary Figure 1. The time course of serial HBsAg, ALT, and HBV DNA titers during the posttreatment follow-up period in the 18 patients with SR who did not resume antiviral therapy are shown (A, C, and E, patients without HBsAg loss [n 5]; B, D, and F, patients with HBsAg loss [n 13]). The time interval for testing was every 3 months for the first 12 months and every 6 months (//) after month 12.