An estimated 350 million individuals are chronically

Transcription

1 Long-Term Effect of Antiviral Therapy on Disease Course After Decompensation in Patients With Hepatitis B Virus Related Cirrhosis Jeong Won Jang, 1,9 Jong Young Choi, 1,9 Young Seok Kim, 2,9 Hyun Young Woo, 3,9 Sung Kyu Choi, 4,9 Chang Hyeong Lee, 5,9 Tae Yeob Kim, 6,9 Joo Hyun Sohn, 6,9 Won Young Tak, 7,9 and Kwang-Hyub Han 8,9 The effect of viral suppression on long-term disease outcome after decompensation in patients with hepatitis B virus (HBV)-related cirrhosis has not been established. The aim of this study was to determine the long-term effect of antiviral therapy (AVT) in patients with HBV-related decompensated cirrhosis. This was a multicenter, prospective, inception cohort study of 707 patients who presented with first-onset decompensated complications, including 284 untreated and 423 antiviral-treated patients (58 previously treated, 253 with early treatment, and 112 with delayed treatment). The primary endpoint was 5-year liver transplantation (LT)-free survival. Secondary endpoints included virological response (VR) and serological response and improvement in liver function. Despite baseline high HBV activity and worse liver function, antiviral-treated patients had significantly better transplant-free survival than untreated patients (5-year survival rates of 59.7% vs. 46.0%, respectively), with more apparent benefits from antivirals in Child-Turcotte-Pugh class B/C and high-viremia groups. The rate of VR and hepatitis B e antigen seroconversion at 5 years in antiviral-treated patients was 14.2% and 49.1%, respectively. A significant improvement in liver function was observed in treated versus untreated patients, with 33.9% of treated patients delisted for LT. Patients with early treatment had better clinical outcomes than those with delayed treatment. Survival was dependent on antiviral response, being significantly better in responders than in nonresponders or untreated cases. The initial benefit of AVT was negated over time in nonresponders. Antiviral treatment and maintained VR remained independently predictive of survival. The study results were corroborated by propensity score-matching analysis. Conclusion: AVT significantly modifies the natural history of decompensated cirrhosis, improving liver function and increasing survival. The results underscore the importance of promptly administering potent antiviral drugs to patients under consideration for LT. (HEPATOLOGY 2015;61: ) An estimated 350 million individuals are chronically infected with hepatitis B virus (HBV) worldwide. 1 The annual incidence of cirrhosis among chronic HBV carriers was reported to be approximately 2.1%-6.0%. 2 After advancement to cirrhosis, liver disease may continue to progress and Abbreviations: ADV, adefovir; ALT, alanine aminotransferase; AST, aspartate aminotransferase; AVT, antiviral therapy; CHB, chronic hepatitis B; CI, confidence interval; CTP, Child-Turcotte-Pugh; CVD, clevudine; ETV, entecavir; HBV, hepatitis B virus; HBeAg, hepatitis B e antigen; HBsAg, hepatitis B surface antigen; HCC, hepatocellular carcinoma; HR, hazard ratio; LCCRC, Liver Cirrhosis Clinical Research Center; LdT, telbivudine; LT, liver transplantation; LVD, lamivudine; MELD, Model for End-Stage Liver Disease; PCR, polymerase chain reaction; SBP, spontaneous bacterial peritonitis; VB, variceal bleeding; VR, virological response. From the 1 Department of Internal Medicine, The Catholic University of Korea, College of Medicine, Seoul, Korea; 2 Department of Internal Medicine, Soonchunhyang University College of Medicine, Bucheon, Korea; 3 Department of Internal Medicine, Pusan National University Hospital, Busan, Korea; 4 Department of Internal Medicine, Chonnam National University Medical School, Gwangju, Korea; 5 Department of Internal Medicine, Catholic University of Daegu School of Medicine, Daegu, Korea; 6 Department of Internal Medicine, Hanyang University Guri Hospital, Guri, Korea; 7 Department of Internal Medicine, Kyungpook National University Hospital, Daegu, Korea; 8 Department of Internal Medicine, Yonsei University College of Medicine, Seoul, Korea; and 9 Liver Cirrhosis Clinical Research Center, Seoul, Korea Received April 30, 2014; accepted January 22, Additional Supporting Information may be found at onlinelibrary.wiley.com/doi/ /hep.27723/suppinfo. 1809

2 1810 JANG ET AL. HEPATOLOGY, June 2015 decompensated complications can occur, especially in those with active HBV replication. Hepatic decompensation among chronic HBV carriers is associated with high mortality and has a reported 5-year survival rate of 14%-35%, which is significantly lower than the rate of 80%-85% among those with compensated cirrhosis. 1-4 A substantial proportion of patients with cirrhosis still have active HBV replication, as evidenced by positive hepatitis B e antigen (HBeAg) or high HBV- DNA levels, 5 which are the strongest factors for disease progression. 6,7 Although understanding the natural history of HBV-related diseases may help to optimize the care of chronic HBV carriers, information on the course of illness after the onset of decompensation is too scanty or, at best, inconclusive because of the tremendous heterogeneity in patient inclusion criteria across studies. Clinical studies to date have revealed that oral antivirals are generally safe and effective in HBV suppression, with an improvement in liver disease in patients presenting with HBV-related decompensation. 2,8 Recent studies also showed that profound viral suppression with newer, more potent antivirals showed short-term efficacy as assessed within 1 or 2 years However, it remains unclear whether these benefits actually translate into parallel gains with respect to avoidance of liver transplantation (LT) or a reduction in the number of transplants, or in improved longterm outcomes in individuals with decompensated cirrhosis, because most previous studies were retrospective, noncomparative, lacked a contemporary control group, and included heterogeneous patients with various stages of cirrhosis at study entry and only shortterm follow-up. 2,5,13 To address these issues, we conducted a prospective, longitudinal study that included patients with decompensated cirrhosis from seven medical centers in Korea. The aim of this study was to evaluate the long-term effect of antiviral therapy (AVT) and maintained HBV suppression in this cohort over a followup period of 7 years after onset of decompensated complications. Patients and Methods Study Population. This multicenter, prospective, observational, inception cohort study was conducted as part of a study project of the Liver Cirrhosis Clinical Research Center (LCCRC) in Korea. 14 Between 2005 and 2012, patients at seven Korean referral centers who met the following criteria were entered into the study: age 18 years; hepatitis B surface antigen (HBsAg) carrier for 6 months; pathological or clinical evidence of cirrhosis, including nodularity/splenomegaly on liver imaging and/or thrombocytopenia; presentation with the first episode of liver decompensation defined as the occurrence of complications, such as ascites, variceal bleeding (VB), encephalopathy, spontaneous bacterial peritonitis (SBP), or hepatorenal syndrome. Patients with hepatocellular carcinoma (HCC) at baseline, a history of other malignancy, and any other form of liver disease, including hepatitis C virus, serious alcoholic disease (consumption >20 g/ day), or autoimmune hepatitis, were excluded. This study (KC10RIMI0483) was approved by the local ethics committees in accord with the 1975 Declaration of Helsinki. Each patient provided informed consent. Assessments and Follow-up. Baseline and onstudy laboratory assessments included routine hematological tests, hepatobiliary function tests, and virological/serological tests for HBV. Patients were followed up with HBV virological/serological tests at the discretion of the local treating physicians at least once every 3-6 months. Serum alpha-fetoprotein and liver ultrasonography were performed regularly every 6 months to detect HCC. Additionally, computed tomography/ magnetic resonance imaging were performed if clinically warranted. Child-Turcotte-Pugh (CTP) and Model for End-Stage Liver Disease (MELD) scores were followed at 6-month intervals. HBsAg, HBeAg, and antibodies to the antigens were measured using commercially available enzyme immunoassays. Serum HBV-DNA level was quantified by either VERSANT 3.0, Bayer HealthCare (detection limit: 2,000 copies/ ml) until 2007 or real time polymerase chain reaction This study was supported by a grant of the Korea Healthcare Technology R&D Project, Ministry of Health and Welfare, Republic of Korea (HI10C2020). The statistical consultation was supported by the Catholic Research Coordinating Center of the Korea Health 21 R&D Project (A070001), Ministry of Health & Welfare, Republic of Korea. Address reprint requests to: Jong Young Choi, M.D., Division of Hepatology, Department of Internal Medicine, Seoul St. Mary s Hospital, College of Medicine, The Catholic University of Korea, #222 Banpo-daero, Seocho-gu, Seoul , Republic of Korea. jychoi@catholic.ac.kr; fax: Copyright VC 2015 by the American Association for the Study of Liver Diseases. View this article online at wileyonlinelibrary.com. DOI /hep Potential conflict of interest: Y.S.K. has served as an advisory committee member for Gilead. He served as a member for study sponsored by Bristol-Myers Squibb (BMS), Gilead, Roche, and MSD. W.Y.T. has served as a consultant and speaker for BMS, Gilead, and Roche. J.Y.C. has served as an advisory committee member or a member for a study sponsored by GSK, BMS, and Roche. J.W.J. has served as a consultant and speaker for BMS, Gilead, and Roche.

3 HEPATOLOGY, Vol. 61, No. 6, 2015 JANG ET AL (PCR)-based methods (detection limit: copies/ ml) from 2008, according to local availability. HBV genotypes were determined by restriction fragment mass polymorphism, as previously described. 15 For all patients, complications were appropriately managed according to current guidelines in all centers. 16,17 A low-salt diet and diuretics were prescribed for patients with ascites. Patients with VB received endoscopic band ligation/sclerotherapy or transjugular intrahepatic portosystemic shunt, and were discharged with b-blocker secondary prophylaxis, unless contraindications were documented. Antibiotic therapy was initiated appropriately for bacterial infection, including SBP. Prophylactic antibiotics were administered for patients with previous SBP and VB. LT allocation was based on the Korean Network for Organ Sharing criteria: status 1, intensive care unit-bound fulminant liver failure with expected survival <7 days; status 2, continuously hospitalized, with a CTP score 10 and liver-related complications; and status 3, requiring continuous medical care, with a CTP score 7. Endpoints and Definitions. The primary endpoint was 5-year LT-free survival. Secondary endpoints included virological response (VR), HBeAg seroconversion, and improvement in CTP and MELD scores at index time points, as measured by mean change from baseline. Early mortality was defined as death occurring within 6 months after decompensation. 8 VR was defined as a maintained undetectable HBV-DNA level (<400 IU/mL) during therapy. 18 Patients who had drug resistance at enrollment, but subsequently achieved maintained VR with rescue therapy were classified as responders. Patients with drug resistance or primary nonresponse 19 and those who were nonevaluable for VR because of early death, LT, or loss to follow-up were classified as nonresponders. The diagnosis of HCC was made by imaging techniques and/or biopsy, as suggested by the regional guideline. 20 AVT. AVT was considered for all participants with detectable viremia, but, in practice, was primarily administered for patients who met the Korean national insurance policy, which, until 2010, covered only patients with alanine aminotransferase (ALT) levels >80 IU/L and HBV DNA >20,000 IU/mL. These cut-off levels for patients with cirrhosis were modified to ALT >40 IU/L and HBV DNA >2,000 IU/mL in For patients who did not meet the insurance criteria, AVT was delayed until follow-up ALT/HBV-DNA tests met the strict criteria owing to the high costs of antivirals in Korea, unless the patient was willing to receive antivirals without reimbursement. Early treatment was defined as antiviral treatment initiated within 3 months of decompensation and delayed treatment if treatment was initiated after 3 months of decompensation. In Korea, lamivudine (LVD) was the only first-line agent until 2007, entecavir (ETV) and clevudine (CVD) have been available since 2007, and telbivudine (LdT) since Adefovir (ADV) was approved only as rescue therapy for LVD resistance and given as monotherapy or in combination with continued LVD, depending on the patient s liver function. Statistical Analysis. Data are expressed as the mean 6 standard deviation or median (range). Categorical variables were compared with the chi-square test or Fisher s exact test, means were analyzed by the Student t test, and medians by Wilcoxon s rank-sum test. Survival time was calculated as the time interval from the date of first-episode decompensation to LT, death, or end of follow-up. LT-free survival curves were constructed using Kaplan-Meier s method, and differences were assessed by log-rank test. Bonferroni s correction was used for multiple comparisons (LT-free survival according to response). The cumulative probability of HBeAg seroconversion and HCC development was determined for the groups, with death and loss to follow-up being considered as competing risks. Continuous variables were transformed into two-level categorical data, based on their median values. The prognostic significance of demographic characteristic was determined using uni- and multivariate Cox s proportional hazard regression analyses. Variables that were significant (P < 0.05) on univariate analysis were entered into the multivariate Cox regression model. We performed rigorous adjustment for differences in baseline characteristics of patients using propensity score. We calculated propensity using multivariable logistic regression to model a dichotomous outcome of the antiviral treatment or control groups for patients in the sample. A full nonparsimonious model was developed that included 13 covariates. This model yielded a c statistic of 0.851, indicating a good ability to differentiate between antiviral treatment and control. One-to-one matching was performed by the Greedy matching method using a macro (available online in the public domain at: proceedings/sugi26/p pdf). We were able to match 127 patients with antiviral treatment to 127 patients without treatment. Stratified Cox s regression with patients matched on propensity score was used to estimate the treatment effect on survival. A two-sided P value <0.05 was considered statistically significant. All analyses were performed using SAS software (version 9.3; SAS Institute Inc., Cary, NC).

4 1812 JANG ET AL. HEPATOLOGY, June 2015 Fig. 1. Flow diagram of patient enrollment. Abbreviations: ARDS, adult respiratory distress syndrome; FU, follow-up. Results Baseline Characteristics. Up until December 2012, the LCCRC cohort included 1,515 patients with decompensated complications. Of these, 808 patients who did not fulfill the diagnosis of chronic hepatitis B (CHB) were excluded and a total of 707 HBsAg-positive patients were consecutively included in this study and followed until December 2013 (Fig. 1). Baseline characteristics of the 707 study subjects are shown in Table 1. The entire patient population consisted of 471 males and 236 females (age, years), and mean baseline CTP and MELD scores were and , respectively. Because HBV genotype C is the universal type in Korean chronic carriers, 21 HBV genotyping is not routinely performed. In our study, genotyping was performed in 86 patients, and all (100%) had genotype C. Of the study subjects, 423 (59.8%) patients had antiviral therapy and 284 (40.2%) did not. Antiviral therapy was initiated early within 3 months in 253 patients (early treatment group) and postponed for >3 months (median, 16.67; interquartile range: ) in 112 until their follow-up data met the reimbursement criteria (delayed treatment group). The remaining 58 patients had received anti-hbv therapy before study entry (previous treatment group). Antivirals used included LVD (203), ETV (198), ADV (7), CVD (10), and LdT (5). On the whole, the antiviral-treated group showed poorer liver function and higher HBV replication activity at baseline than the untreated group (Table 1). VR and Serological Response. Overall, VR after treatment was observed in 245 (57.9%) patients, HBV drug resistance in 104 (24.6%), and primary nonresponse in 13 (3.1%). The remaining 61 (14.4%) were

5 HEPATOLOGY, Vol. 61, No. 6, 2015 JANG ET AL Table 1. Baseline Characteristics of the Study Population Entire Cohort Propensity Score-Matched Cohort All Patients (n 5 707) Treated Group (n 5 423) Untreated Group (n 5 284) P Value* Treated Group (n 5 127) Untreated Group (n 5 127) P Value* Standardized Difference in % Sex, male:female 471 (66.6):236 (33.4) 270 (63.8):153 (36.2) 201 (70.8):83 (29.2) < (70.1):38 (29.9) 89 (70.1):38 (29.9) > Age, years < Mode of decompensation Ascites 356 (50.4) 240 (56.7) 116 (40.8) 58 (45.7) 56 (44.1) 3.21 VB 174 (24.6) 88 (20.8) 86 (30.3) 41 (32.3) 35 (27.6) Encephalopathy 36 (5.1) 17 (4.0) 19 (6.7) 7 (5.5) 6 (4.7) 3.63 SBP 14 (2.0) 11 (2.6) 3 (1.1) 2 (1.6) 2 (1.6) 0.00 Hepatorenal syndrome 4 (0.6) 2 (0.5) 2 (0.7) 1 (0.8) 1 (0.8) 0.00 Combined 123 (17.4) 65 (15.4) 58 (20.4) 18 (14.2) 27 (21.3) HBeAg seropositivity 297 (50.9) 233 (59.4) 64 (21.5) < (52.0) 64 (50.4) HBV DNA (log IU/mL) < HBV genotype C (n 5 86/86; 100) C (n 5 59/59; 100) (n 5 27/27; 100) >0.999 C (n 5 15/15; 100) C (n 5 9/9; 100) > AST (IU/L) 70 (10-7,040) 78 (10-7,040) 55 (11-2,070) < (42-140) 63 (41-113) ALT (IU/L) 43 (5-2,673) 51 (9-2,673) 33 (5-174) < (26-94) 38 (23-68) Total bilirubin (mg/dl) 1.9 ( ) 2.1 ( ) 1.7 ( ) ( ) 1.8 ( ) Albumin (g/dl) PT (INR) CTP score CTP class A 101 (14.3) 48 (11.3) 53 (18.7) 19 (15.0) 18 (14.2) 2.26 B 363 (51.3) 228 (53.9) 135 (47.5) 60 (47.2) 61 (48.0) C 243 (34.4) 147 (34.8) 96 (33.8) 48 (37.8) 48 (37.8) 0.00 MELD score Figures in parentheses indicate percentage. *P value for comparison between treated and untreated groups. Abbreviations: PT, prothrombin time; INR, international normalized ratio.

6 1814 JANG ET AL. HEPATOLOGY, June 2015 Fig. 2. Cumulative rate of HBeAg seroconversion and transplant-free survival in (A and B) the whole cohort and (C and D) early treatment group. *Adjusted for age and sex. (E) Survival according to treatment response in the whole cohort. **Bonferroni-adjusted P value. not eligible for assessment of response because of early death, LT, or loss to follow-up after treatment initiation. The rate of VR in the whole antiviral-treated group was 41.4% and 14.2% at 12 and 60 months after study inclusion, respectively. Of the 117 evaluable nonresponders, 19 eventually switched to AVT and 17 to ETV (1.0 mg), 50 had AVT add-on therapy, and 31 had no change in antivirals at further follow-up.

7 HEPATOLOGY, Vol. 61, No. 6, 2015 JANG ET AL Fig. 3. Mean changes from baseline in (A) Child-Pugh score and (B) MELD score at 6, 12, 36, and 60 months for the control and each treatment group. *P < Abbreviation: Tx, treatment. The most common resistance pattern observed in these patients was rtl180m 1 rtm204i/v (79.4%; Supporting Table 1). Normalization of ALT at 12 months was achieved in 63.7% (170 of 267) of antiviral-treated patients with a baseline ALT above the upper limit of normal; the remainder showed no ALT normalization (57) or were lost to follow-up (8) or LT/death (32). Overall, HBeAg seroconversion was achieved in 81 (27.3%) of 297 HBeAg-positive patients. Probability of HBeAg seroconversion was 13.4%, 34.7%, and 49.1% at 12, 36, and 60 months for the whole antiviral-treated group, whereas corresponding probability in the untreated group was 10.9%, 15.7%, and 35.1%, respectively; the rate of HBeAg seroconversion was not different between the two groups (P ; Fig. 2A). However, when treatment time was taken into account, the early treatment group showed a significantly higher rate of HBeAg seroconversion than the untreated group, with a corresponding rate of 17.8%, 40.8%, and 57.4%, respectively (P ; Fig. 2C), whereas the previous or delayed treatment groups did not show an increase in HBeAg seroconversion (P > 0.05; Supporting Fig. 1). Change in Hepatic Function. Over 60 months, mean change from baseline in CTP score was and for the antiviral-treated and control groups, respectively, and the corresponding change in MELD score was and The treatment group showed a greater reduction in both CTP and MELD scores over time among patients who had tests performed at each time point (Fig. 3A,B). Regarding time of treatment initiation, the early treatment group exhibited a greater decrease in CTP and MELD scores than the delayed treatment group (Fig. 3A,B). Among 606 patients with baseline CTP scores 7, 14.7% (55 of 375) of antiviral-treated patients achieved a 2-point reduction in CTP score up to 60 months, compared with only 2.2% (5 of 231) of untreated patients (P < 0.001). At 60 months, 45 (12.0%) and 4 (1.7%) of antiviral-treated and untreated patients, respectively, remained in CTP class A (P < 0.001). In total, 33.9% (127 of 375) of antiviral-treated patients were removed from the LT waiting list within 12 months. Transplant-Free Survival. During the mean follow-up of months, 260 patients died,

8 1816 JANG ET AL. HEPATOLOGY, June 2015 Table 2. Factors for Transplant-Free Survival Entire Cohort (n 5 707) Propensity Score-Matched Cohort (n 5 254) Univariate Multivariate Univariate Multivariate HR (95% CI) P Value Adjusted HR (95% CI) P Value HR (95% CI) P Value Adjusted HR (95% CI) P Value Male sex 1.43 ( ) ( ) ( ) Age >55 years 1.39 ( ) ( ) ( ) ( ) Multiple complications 1.98 ( ) < ( ) ( ) ( ) HBeAg seropositivity 1.10 ( ) ( ) High HBV activity 1.15 ( ) ( ) AST >70 IU/L 1.32 ( ) ( ) ( ) ALT >40 IU/L 1.15 ( ) ( ) CTP class C 1.43 ( ) < ( ) < ( ) < ( ) MELD score > ( ) < ( ) < ( ) < ( ) <0.001 AVT (2) 1.53 ( ) < ( ) < ( ) < ( ) < underwent LT, 6 were transferred to another hospital, 57 were lost to follow-up, and the remaining 320 were still alive at the last visit. Causes of death were liver related in 242 (93.1%) and liver unrelated in 18 (6.9%) patients (Fig. 1). The majority of patients (77.1%) who died of liver-related diseases or underwent LT after the first 6 months were those who had been untreated or failed to achieve maintained VR under therapy. Kaplan-Meier s survival curve showed an early and steep decline within 6 months for both treatment and control groups of the whole cohort. Overall, the treatment group showed better survival than the untreated group (P ; Fig. 2B). Early mortality was not significantly different between the antiviral-treated and untreated groups (13.8% vs. 19.7%; P ). Given that the timing of antiviral initiation varied among treated patients, 253 who had early treatment at study entry were subjected to further survival analyses. As a result, the early treatment group had significantly better LT-free survival than the untreated group, with 5-year survival probabilities of 59.7% versus 46.0%, respectively (P ; Fig. 2D). This effect was more significant among patients with advanced disease, including CTP class B (P )/C (P < ) and high MELD scores 13 (P ) and among those with high viremia, defined as positive HBeAg or HBV DNA >2,000 IU/ ml (P < ; Supporting Figs. 2, 4). The better survival in the treatment group was still maintained after excluding 48 patients with hepatitis flares (aminotransferase levels >400 IU/L; P ; data not shown) and also after excluding 101 patients suffering early deaths (P ; data not shown). Next, we investigated the effect of antiviral response on survival by dividing patients into three groups: responder, nonresponder, and untreated control groups. Patients achieving VR had significantly longer survival than the other patients (P < ; Fig. 2E). Probability of LT-free survival at 12, 36, and 60 months was 94.0%, 81.4%, and 73.0% in the responder group; 76.1%, 58.2%, and 42.7% in the nonresponder group; and 73.2%, 55.8%, and 46.0% in the control group, respectively (Fig. 2E). Table 3. Factors for Transplant-Free Survival Among the Antiviral-Treated Group Entire Cohort (n 5 423) Propensity Score-Matched Cohort (n 5 127) Univariate Multivariate Univariate Multivariate HR (95% CI) P Value Adjusted HR (95% CI) P Value HR (95% CI) P Value Adjusted HR (95% CI) P Value Male sex 1.28 ( ) ( ) Age >55 years 1.60 ( ) ( ) ( ) ( ) Multiple complications 1.69 ( ) ( ) ( ) HBeAg seropositivity 1.18 ( ) ( ) High HBV activity 1.37 ( ) ( ) AST >70 IU/L 1.14 ( ) ( ) ALT >40 IU/L 1.04 ( ) ( ) CTP class C 1.29 ( ) ( ) ( ) MELD score > ( ) < ( ) < ( ) ( ) VR (2) 1.77 ( ) < ( ) < ( ) ( ) Delayed treatment 1.10 ( ) ( ) 0.601

9 HEPATOLOGY, Vol. 61, No. 6, 2015 JANG ET AL Fig. 4. Analysis of propensity score-matched cohort. (A) Survival between treated and untreated cohorts. (B) Survival according to treatment response. *Bonferroni-adjusted P value. Prognostic Factors. To evaluate factors that affected survival, the clinical variables listed in Table 2 were included in the analysis. Cox s proportional hazard regression analysis showed that sex, age, decompensated complications, CTP class, MELD score, and AVT were independently predictive of LT-free survival. When analyzed within the antiviral-treated group, age, MELD scores, and response to AVT remained independently predictive of survival (Table 3). Adjusted survival was significantly better in antiviral-treated patients than in untreated patients (hazard ratio [HR]: 1.68; 95% confidence interval [CI]: ; P < 0.001). In particular, adjusted survival was significantly better in responders than in nonresponders (HR, 2.43; 95% CI: ; P ) or untreated patients (HR, 2.79; 95% CI: ; P < 0.001). Propensity Score-Matching Analysis. To confirm the validity of the treatment effects, we conducted a propensity score-matching analysis. A total of 254 patients were obtained after propensity score matching and well matched for demographic and clinical characteristics (Table 1). Among the propensity score-matched cohort, LT-free survival was significantly longer in the antiviraltreated group than the untreated group, with 5-year survival rates of 51.6% and 31.0%, respectively (P ; Fig. 4A). We then performed Cox s regression analysis to identify factors for survival in the matched cohort. Together with age and MELD score, AVT was identified as an independent prognostic factor for survival (HR, 4.11; 95% CI: ; P < 0.001; Table 2). When analyzed within antiviral-treated patients of the propensity score-matched cohort, response to AVT (HR, 1.75; 95% CI: ; P ) was the strongest predictor of survival (Table 3). When treated as a continuous variable, MELD score was also independently predictive of survival in antiviral-treated patients of the matched cohort (Supporting Table 2). Responders had significantly better survival than the other patients (P ; Fig. 4B). The survival rate of the nonresponder group appeared marginally better than that of controls for the first few years, but this difference was negated during continued follow-up and survival was comparable to that of untreated controls after 5 years (P ). Finally, we analyzed incidence of HCC in this propensity score-matched cohort. Overall, 56 patients of the cohort developed HCC; 46 developed HCC after 1 year, whereas 10 developed HCC within 1 year. These 10 were excluded from the analysis, because they were presumed to have preexisting HCC that was not detected at baseline. The cumulative incidence of HCC at 5 years was 35.1% and 29.4% in the antiviral-treated and untreated groups, respectively (P ). When the whole cohort was taken into account, the corresponding incidence was 32.3% and 27.2%, respectively (P ). For both cohort analyses, there was no difference in the incidence of HCC between patients with and without treatment. Discussion The major strength of our study is that it represents the largest prospective study ever reported and one of the longest followed cohort studies of HBV-related decompensated cirrhosis. Moreover, the current study enlisted subjects at a uniform point in the course of liver disease (first-onset decompensated complications) under strict inclusion criteria. The inclusion of an inception cohort may provide methodological rigor for this observational study on the natural history of disease, because decompensated cirrhosis is a heterogeneous, multistage

10 1818 JANG ET AL. HEPATOLOGY, June 2015 disease entity and patient prognosis principally depends on the stage of cirrhosis. The findings of this study provide convincing evidence for the long-term benefits of AVT for HBV-related decompensated cirrhosis, which have, thus far, remained incomplete specifications in practice guidelines. 1,18,19 Additionally, our conclusion is reliable given that the main findings were corroborated by propensity score-matched analysis. The present study, which evaluated 707 patients with HBV-related complications followed for up to 7 years, demonstrated that effective AVT is associated with long-term benefits in all outcome measures. First, maintained HBV suppression with antivirals resulted in long-term improvement in liver function, as evidenced by a significant decrease in CTP or MELD scores at 5 years. In contrast to earlier studies in this setting with follow-up limited to 1-2 years, 9-12 our study presented long-term results, indicating that the improvement in liver function by treatment can be maintained for up to 5 years, thus delaying or avoiding the need for LT in approximately 60% of patients who were otherwise considered for LT. Specifically, 33.9% of antiviral-treated patients were eventually delisted for LT within 1 year. Last, AVT significantly prolonged survival, which was better in patients with than without response or untreated patients. The overall benefits of AVT appear even more solid, given that the treated cohort had unfavorable baseline characteristics, including more patients with higher HBV activity and worse liver biochemical profiles, compared to untreated controls. This underscores the fundamental need for antiviral administration in these patients. Although regional practice guidelines recommend early AVT for HBV-related decompensated cirrhosis, 1,18,19 it is not certain if initiating treatment earlier would improve the rate of response and translate into a meaningful longterm outcome. Importantly, our findings addressed this issue and confirmed the superior effects of early antiviral intervention in these patients. Efficacy measures were significantly better in patients with early initiation of antivirals upon presentation of decompensation than in those with delayed therapy. Intriguingly, our study is the first to demonstrate a higher rate of HBeAg seroconversion in patients who received early antiviral treatment, compared to untreated controls, which has not been observed in previous studies of short-term LVD therapy. 22 Additionally, a reduction in CTP and MELD scores over time was more apparent with early treatment than with delayed therapy. In this regard, our results support the guidelines that recommend early initiation of potent antivirals for decompensated patients to allow a better chance of clinical benefit. Another interesting and heretofore unreported finding from this study is the difference in long-term outcome across the extent of viral suppression in patients with decompensated cirrhosis. In our results, antiviral response was the strongest predictor of long-term outcome, with significantly better survival in maintained virological responders than in nonresponders or untreated cases. In particular, failure to achieve maintained response ultimately negated the initial benefits of antivirals over extended follow-up (>5 years), with an overall lack of survival gain compared to untreated patients (Fig. 4B). These extend the results of LVD for patients with compensated diseases to the setting of decompensated cirrhosis, 23 demonstrating that the antiviral effect remains valid even after developing decompensation. Considering that maintaining potent suppression of HBV led to better long-term results and that HBeAg, HBV DNA, and aminotransferase levels had little impact on outcomes, our findings again support the current guidelines that recommend tenofovir or ETV as a preferred first-line choice for patients with liver cirrhosis, even when HBV-DNA and aminotransferase levels are only modest. 1,18,19 Despite these long-term beneficial effects of treatment, not all treated patients survived an episode of decompensation, especially during the early period of therapy. In our results, 13.4% (34 of 253) of patients, mostly those with severe hepatic dysfunction, died within 6 months, compatible with previous observations. 8 This implies that, because it takes several months for liver function to recover after initiating antivirals, 5,13 any patient with severe decompensation should be considered for LT even if antiviral therapy is initiated. Recently, anti-hbv therapy was shown to reduce the risk of HCC in patients with compensated cirrhosis However, we found no protective effect in patients with decompensated cirrhosis. This might be owing to the relatively short survival of these patients, but another possible explanation is that genetic alterations in cirrhotic nodules may have preexisted before initiation of antivirals. Thus, patients presenting with decompensation should be started early with AVT and monitored closely with regular HCC surveillance while under consideration for LT. It is of interest that the 5-year survival rate of 46% observed in our untreated cohorts appears better than the previously reported rates of 14%-35%. 1-4 This may be owing to improved medical management of cirrhosis in current practice. Alternatively, it is possible that patients with persistently low HBV/ALT levels who were at low risk of disease progression were eventually categorized as untreated, because our practice

11 HEPATOLOGY, Vol. 61, No. 6, 2015 JANG ET AL was to administer antivirals to initially untreated patients with detectable viremia later if they met the insurance criteria on follow-up (delayed treatment group). Of note is that our results stress the significantly superior benefits of AVT compared with this low-risk control group with good prognosis. The modest therapeutic gains observed in the low viremia or CTP class A groups, compared to the high viremia or class B/C groups, were likely to be associated with the small number of treated cases under our stringent insurance criteria and the relatively longer survival observed in untreated cases. Indeed, there were only five deaths other than early mortality in these groups with early treatment (data not shown). Thus, the beneficial effect of antivirals in these low-risk groups will likely become obvious with more extended follow-up. It may also be argued that the treatment group included more patients with acute decompensation secondary to hepatitis flares, who were likely to have a better prognosis 2 and thus contributed to the increased survival, compared to untreated controls. However, exclusion of patients with hepatitis flares from the analysis did not significantly change the results. It would not be ethically justifiable to include untreated controls in studies of cirrhosis. However, reimbursement for anti-hbv agents strictly limited to specified indications in Korea provided us a unique and valuable opportunity to explore the natural history after decompensation and the long-term effects of AVT with the inclusion of contemporary untreated controls. Our study has additional limitations. Changes in hepatic function were assessed in surviving patients with tests at index time points. Such a select picture may lead to the potential for overestimating treatment effects by excluding those with missing data. The cut-off value defining VR (400 IU/mL) was arbitrary and insensitive because non-pcr assays were used for HBV-DNA testing during the first few years of the study initiation. However, by definition, most of our patients achieving response were ETV treated or successfully rescued patients who continued to have undetectable HBV-DNA levels, even using PCR methods, during extended follow-up. It is possible that a subset of critically ill patients who were unable to take oral medications might have been unintentionally categorized into the untreated group and subsequently suffered early deaths, producing overestimates of power. However, significant benefits from antivirals remained valid even when we separately analyzed only patients surviving longer than 6 months. Finally, this is a nonrandomized study with different baseline characteristics in the two cohorts, which might affect the results. However, the treated cohorts had rather unfavorable factors, such as greater levels of aspartate aminotransferase (AST)/ALT, poorer hepatic reserve, and higher HBV activity, compared to untreated cohorts. Therefore, the beneficial effects of antivirals might be even more than we observed. Furthermore, the effect was confirmed using propensity score-matched analysis to minimize potential confounders. These indicate that the present results are solid and robust. In conclusion, anti-hbv therapy significantly modifies the natural history of decompensated cirrhosis, improving liver function and increasing survival. Early intervention with potent antivirals provides better virological and serological responses in these patients. Maintained VR in patients under AVT leads to better long-term LT-free survival, compared to nonresponders or untreated patients. Liver function continues to improve and is maintained for up to 5 years, delaying the need for LT, especially in responders. Thus, antiviral interventions to achieve potent and durable suppression of HBV should be promptly initiated in patients with decompensated cirrhosis under consideration for LT. Acknowledgment: The authors thank Prof. Hyeon Woo Yim and Mi Sun Park for their helpful statistical assistance. The study protocol was approved by the institutional review boards at all of the participating hospitals (approval no.: KC10RIMI0483; title of the study, Clinical Study for Epidemiologic, Clinical Features, Survival, and Prognosis of the Patients with Liver Cirrhosis in Korea ). The results presented herein are part of the aforementioned study. The overall study projects were supported by the LCCRC, which has been supervised by the National Strategic Coordinating Center for Clinical Research, Ministry of Health and Welfare, Republic of Korea (HI10C2020). Additional information on the study projects is available at: References 1. European Association For The Study Of The Liver. EASL clinical practice guidelines: management of chronic hepatitis B virus infection. J Hepatol 2012;57: Chu CM, Liaw YF. Hepatitis B virus-related cirrhosis: natural history and treatment. Semin Liver Dis 2006;26: de Jongh FE, Janssen HL, de Man RA, Hop WC, Schalm SW, van Blankenstein M. Survival and prognostic indicators in hepatitis B surface antigen-positive cirrhosis of the liver. Gastroenterology 1992;103: Fattovich G, Giustina G, Schalm SW, Hadziyannis S, Sanchez-Tapias J, Almasio P, et al. Occurrence of hepatocellular carcinoma and decompensation in western European patients with cirrhosis type B. The EUROHEP Study Group on Hepatitis B Virus and Cirrhosis. HEPATOLOGY 1995;21:77-82.

12 1820 JANG ET AL. HEPATOLOGY, June Peng CY, Chien RN, Liaw YF. Hepatitis B virus-related decompensated liver cirrhosis: benefits of antiviral therapy. J Hepatol 2012;57: Yang HI, Lu SN, Liaw YF, You SL, Sun CA, Wang LY, et al. Hepatitis B e antigen and the risk of hepatocellular carcinoma. N Engl J Med 2002;347: Iloeje UH, Yang HI, Su J, Jen CL, You SL, Chen CJ, et al. Predicting cirrhosis risk based on the level of circulating hepatitis B viral load. Gastroenterology 2006;130: Fontana RJ, Hann HW, Perrillo RP, Vierling JM, Wright T, Rakela J, et al. Determinants of early mortality in patients with decompensated chronic hepatitis B treated with antiviral therapy. Gastroenterology 2002;123: Shim JH, Lee HC, Kim KM, Lim YS, Chung YH, Lee YS, et al. Efficacy of entecavir in treatment-naive patients with hepatitis B virusrelated decompensated cirrhosis. J Hepatol 2010;52: Liaw YF, Sheen IS, Lee CM, Akarca US, Papatheodoridis GV, Suet- Hing Wong F, et al. Tenofovir disoproxil fumarate (TDF), emtricitabine/tdf, and entecavir in patients with decompensated chronic hepatitis B liver disease. HEPATOLOGY 2011;53: Liaw YF, Raptopoulou-Gigi M, Cheinquer H, Sarin SK, Tanwandee T, Leung N, et al. Efficacy and safety of entecavir versus adefovir in chronic hepatitis B patients with hepatic decompensation: a randomized, open-label study. HEPATOLOGY 2011;54: Chan HL, Chen YC, Gane EJ, Sarin SK, Suh DJ, Piratvisuth T, et al. Randomized clinical trial: efficacy and safety of telbivudine and lamivudine in treatment-naive patients with HBV-related decompensated cirrhosis. J Viral Hepat 2012;19: Bae SH, Yoon SK, Choi JY, Jang JW, Cho SH, Yang JM, et al. Timing of lamivudine administration according to Child class in patients with decompensated cirrhosis. J Gastroenterol Hepatol 2005;20: Woo HY, Jang JW, Choi JY, Kim SG, Lee CH, Shon JH, et al. Mode of cirrhosis related complication in Korean patients with liver cirrhosis a prospective multicenter cohort study. HEPATOLOGY 2012;56:951A. 15. Jang JW, Chun JY, Park YM, Shin SK, Yoo W, Kim SO, et al. Mutational complex genotype of the hepatitis B virus X /precore regions as a novel predictive marker for hepatocellular carcinoma. Cancer Sci 2012;103: de Franchis R. Evolving consensus in portal hypertension. Report of the Baveno IV consensus workshop on methodology of diagnosis and therapy in portal hypertension. J Hepatol 2005;43: European Association for the Study of the Liver. EASL clinical practice guidelines on the management of ascites, spontaneous bacterial peritonitis, and hepatorenal syndrome in cirrhosis. J Hepatol 2010;53: Liaw YF, Leung N, Kao JH, Piratvisuth T, Gane E, Han KH, et al. Asian-Pacific consensus statement on the management of chronic hepatitis B: a 2008 update. Hepatol Int 2008;2: Lok AS, McMahon BJ. Chronic hepatitis B: update HEPATOLOGY 2009;50: Bruix J, Sherman M; American Association for the Study of Liver Diseases. Management of hepatocellular carcinoma: an update. HEPATO- LOGY 2011;53: Bae SH, Yoon SK, Jang JW, Kim CW, Nam SW, Choi JY, et al. Hepatitis B virus genotype C prevails among chronic carriers of the virus in Korea. J Korean Med Sci 2005;20: Singal AK, Fontana RJ. Meta-analysis: oral anti-viral agents in adults with decompensated hepatitis B virus cirrhosis. Aliment Pharmacol Ther 2012;35: Liaw YF, Sung JJ, Chow WC, Farrell G, Lee CZ, Yuen H, et al. Lamivudine for patients with chronic hepatitis B and advanced liver disease. N Engl J Med 2004;351: Papatheodoridis GV, Lampertico P, Manolakopoulos S, Lok A. Incidence of hepatocellular carcinoma in chronic hepatitis B patients receiving nucleos(t)ide therapy: a systematic review. J Hepatol 2010;53: Sung JJ, Tsoi KK, Wong VW, Li KC, Chan HL. Meta-analysis: treatment of hepatitis B infection reduces risk of hepatocellular carcinoma. Aliment Pharmacol Ther 2008;28: Supporting Information Additional Supporting Information may be found at onlinelibrary.wiley.com/doi/ /hep.27723/suppinfo.