Graphical Abstract
INTRODUCTION
Hepatitis B virus (HBV) infection remains a significant global health challenge [1], with approximately 257.5 million people currently living with HBV, resulting in over 850,000 deaths annually [2], primarily from late complications such as cirrhosis and liver cancer [3]. The number of deaths from HBV infection is projected to increase rather than decrease through 2030 [2]. Alarmingly, patients with chronic hepatitis B (CHB) are severely underdiagnosed and undertreated worldwide, especially severe in resource-constrained regions [4], but also common even in high-income countries [5,6].
The indications for antiviral therapy in CHB patients have been debatable. While international guidelines provide recommendations based on clinical assessments such as serum alanine aminotransferase (ALT) levels, HBV DNA levels, and liver fibrosis severity [7-9], a sizable portion of patients fall into a “gray zone” for which antiviral treatment is not conventionally indicated [10]. These patients often present with minimally elevated ALT levels, fluctuating viral load, or mild liver fibrosis [11]. For patients in the gray zone, the recommendations have typically been frequent monitoring without antiviral therapy. However, recent studies demonstrated the risks for disease progression and clinical complications in patients ineligible for antiviral therapy by the current criteria [11-13], challenging the traditional paradigms of treatment initiation.
In response to the alarming forecasts of disease burden and the calls to close the gaps in the care cascade of patients with CHB, several recent updates in clinical guidelines have aimed to simplify and expand treatment eligibility [14,15]. Nevertheless, the broader treatment strategies are not unanimously agreed upon among experts and professional associations [16,17], underscoring the necessity of empirical evidence to examine the treatment effects in patients conventionally considered ineligible for treatment.
To address the knowledge gap, we have conducted a double-blind randomized placebo-controlled trial to evaluate the efficacy of antiviral therapy on histological outcomes in CHB patients with minimally raised ALT (the TORCH-B trial), and demonstrated that treatment was efficacious in reducing the risk of fibrosis progression [18]. Building upon the TORCH-B trial, the current study is a rollover phase where all eligible patients were transitioned to open-label tenofovir disoproxil fumarate (TDF) treatment for an extended period of three years. The primary objective of this follow-up study is to assess the changes in liver histopathology from placebo to TDF or following prolonged TDF therapy, providing long-term data on the impact of antiviral treatment in this specific patient population.
MATERIALS AND METHODS
Study design and participants
This study is an open-label extension phase of the TORCH-B trial that has been fully published [18]. In brief, the TORCH-B trial screened adult CHB patients with serum HBV DNA >2,000 IU/mL and ALT levels between one to two folds above the upper limit of normal (ULN) at six hospitals in Taiwan (E-Da Hospital, Kaohsiung; Ditmanson Medical Foundation Chiayi Christian Hospital, Chiayi; Taichung Veterans General Hospital, Taichung; National Taiwan University Hospital Yunlin Branch, Yunlin; Chi Mei Medical Center Liouying Hospital, Tainan; Taitung Mackay Memorial Hospital, Taitung). Patients were excluded for viral co-infection with human immunodeficiency virus (HIV), hepatitis C virus, or hepatitis D virus, cirrhosis, hepatic decompensation, malignancy, organ transplant, or severe comorbidities limiting life expectancy to less than a year. Pregnant or lactating women were not eligible as well. Enrolled patients were randomized to receiving TDF or placebo for 3 years.
All patients who had completed the TORCH-B trial and remained seropositive for hepatitis B surface antigen (HBsAg) were invited to participate in the current study. Both hepatitis B e antigen (HBeAg)-positive and HBeAg-negative patients were eligible. All participants gave written informed consent for this open-label rollover study, which was conducted in compliance with Good Clinical Practice and the principles of the Declaration of Helsinki. The study protocol (EMRP35103N) was approved by all the participating hospitals and it was registered in ClinicalTrials.gov with the identifier of NCT02463019.
Intervention and follow-up
All enrolled patients received TDF (300 mg) once daily for 3 years (156 weeks). The treatment continued from the blinded intervention in the TORCH-B trial without interruption. Upon enrollment, participants filled out a structured questionnaire to collect demographic data, socioeconomic status, lifestyle information, and medical history. The baseline assessment also included physical examination, hemogram, blood biochemistry, HBV serology, serum viral load, and abdominal sonography.
Participants were followed up at a frequency of every 12 weeks. At each follow-up visit, any adverse events were evaluated and recorded, in addition to measurement of serum liver enzymes (ALT and aspartate aminotransferase [AST]). Additionally, patients were instructed to return unused medication to count the tablets for assessing drug adherence.
More comprehensive laboratory tests that included serum biochemistry (ALT, AST, bilirubin, phosphorus, potassium, and creatinine), urinalysis, alpha-fetoprotein, HBV DNA, HBeAg, anti-HBe, and prothrombin time were measured every 24 weeks. Abdominal sonography was also performed every 24 weeks. The blood cell counts and serological tests of HBsAg and anti-HBs were assessed annually.
Percutaneous liver biopsy was scheduled at the end of the 3-year open-label trial, within the window of one month. The procedure was performed with a core biopsy needle (SuperCore™ 18Gx15cm; Argon Medical, Frisco, TX, USA) under real-time sonographic guidance with the biopsied site confined to the right anterior section of the liver.
All serological markers of HBV (HBsAg, anti-HBs, HBeAg, and anti-HBe) were measured using commercially available immunoassays (ABBOTT GmbH & Co., Wiesbaden, Germany). Serum samples with HBsAg levels exceeding the range of automated quantitation (>250 U/L) were manually diluted for a quantitative result without upper bounds. Serum HBV DNA levels were quantified using the Roche COBAS TaqMan Assay, with the ranges between 20 and 170,000,000 IU/mL.
Outcome assessment
The primary outcomes were changes in liver histopathology according to the Ishak and Knodell scoring systems after 3 years of TDF treatment [19,20]. For the tissues to be evaluable, the biopsies should contain at least 6 portal tracts. Two pathologists independently reviewed the histological slides to stage liver fibrosis by the Ishak system and grade necroinflammation by the Knodell system. A third pathologist was consulted on the occasion of disagreement to achieve a final report. The changes in histopathology were determined with the liver biopsy taken at the end of the TORCH-B trial (within one month before enrollment to the current roll-over study) as the reference.
Secondary outcomes, assessed within one month after the 3-year study period, included biochemical normalization (serum ALT <40 U/L), virological remission (undetectable serum HBV DNA), and serological endpoints comprising quantitative HBsAg levels, HBsAg loss, and HBeAg loss for HBeAg-positive patients.
Statistical analyses
Descriptive statistics were presented as median with interquartile range (IQR) for continuous data and as counts with percentages for categorical variables. Within a single group, the Chi-square test of goodness of fit was employed to examine how proportions were distributed with an even distribution as the null hypothesis. Paired comparisons within a single group utilized the Wilcoxon signed-rank test for continuous data and the Stuart-Maxwell test for categorical data. Because HBsAg and HBV DNA data were logarithmically transformed, values below the detection limits were imputed as 1 for the purpose of computation.
By the modified intention-to-treat principle, histological changes were evaluated in all patients with paired liver biopsies taken at both the beginning and end of the study. Analyses of biochemical, serological, and virological outcomes included all patients with end-of-study blood tests, regardless of liver biopsy availability. We performed post-hoc analyses to explore factors associated with changes in Ishak fibrosis stage and Knodell inflammation score. Additional post-hoc analyses were conducted to compare histopathological, virological, and biochemical parameters at the end of the current study, as well as their changes from the TORCH-B’s baseline to the conclusion of this open-label study, according to the treatment allocation in the TORCH-B trial. Statistical analyses were performed using Stata software (version 18; Stata Corp., College Station, TX, USA). A two-tailed P-value below 0.05 was considered statistically significant.
RESULTS
Characteristics of the study population
Among the 160 participants of the TORCH-B trial, a total of 146 patients were eligible and enrolled to receive open-label TDF between January 10, 2015 and November 17, 2018 (Fig. 1). Among them, 73 patients were rolled over from placebo and the other half had been taking TDF for 3 years. Baseline characteristics of the study population were summarized in Table 1. Overall, they were predominantly male (n=118, 80.8%) with a median age of 47 years (IQR 41–57). As would be expected, most of the patients initially randomized to TDF had undetectable levels of serum HBV DNA. Conversely, the median viral load in the prior placebo group was 3.95 log IU/mL (95% CI 2.27–5.33). Similarly, significant fibrosis (Ishak stage ≥3) was less prevalent in the prior TDF receivers (n=15, 20.5%) compared to the prior placebo counterparts (n=27, 37.0%). Significant necroinflammation (Knodell score ≥4) was also less common in patients previously assigned to TDF (n=9, 12.3%) than those in the placebo group (n=27, 37.0%).
During the study period, 11 patients were unwilling to continue the study or lost to follow-up. Serious illness, which included hepatocellular carcinoma (n=5), lung cancer (n=1), lymphoma (n=1), and HIV superimposed infection (n=1) occurred in 8 patients. Besides, 4 patients refused liver biopsy at the end of the study. Therefore, 123 patients completed the open-label trial with paired liver biopsy. The study ended on December 3, 2021 when the last patient underwent liver biopsy (Fig. 1).
Changes in hepatic histopathology after 3 years of open-label TDF treatment
The results of liver histopathology were summarized in Table 2. As shown in Figure 2, the distributions of fibrosis stage and inflammatory severity following treatment were significantly different from the baseline (P<0.0001 for the distribution of fibrosis stages and P=0.0082 for the distribution of inflammatory scores; Stuart–Maxwell test).
The changes in liver fibrosis and necroinflammation from baseline to 3 years after TDF treatment are illustrated in Figure 3. Overall, significantly more patients decreased in the Ishak fibrosis stage (n=74, 60.2%) than those registering no changes (n=32, 26.0%) or any increases (n=17, 13.8%) (P<0.0001, Chi-square test). The proportion of patients with an Ishak score ≥ 3 points significantly decreased from 26.8% (n=33) to 9.8% (n=12) (P=0.0002, Stuart–Maxwell test). The post-hoc analysis found that the viral remission rate appeared to be lower (58.8% or 10/17 vs. 86.8% or 92/106; P=0.07) in patients with any progression in the fibrosis stage (Supplementary Table 1).
The changes in the Knodell inflammatory scores were also significant with more patients achieving a decrease in the score (n=58, 47.2%), compared to no changes (n=29, 23.6%) or an increase (n=36, 29.3%) (P=0.0038, Chi-square test). The proportion of patients with a Knodell score ≥ 4 points decreased from 20.3% (n=25) to 10.6% (n=13) (P=0.032, Stuart–Maxwell test). In the post-hoc analysis for factors associated with worsening in inflammation (Supplementary Table 2), patients with any increase in the Knodell inflammation scores had a significantly higher rate of fatty liver (50% or 18/36 vs. 27.6% or 24/87; P=0.02).
In the sensitivity analysis, which assumed all dropouts due to severe adverse events had experienced progression in both fibrosis and inflammation, the results remained consistent with the main analysis (Supplementary Fig. 1).
Histopathological outcomes according to the prior treatment status
In the subgroup switching from placebo, both fibrosis stages and necroinflammation scores were significantly improved (Fig. 4A). The Ishak fibrosis stage decreased, did not change, and increased in 42 (68.9%), 13 (21.3%), and 6 (9.8%) patients, respectively (P<0.0001, Chi-square test). The Knodell score decreased, did not change, and increased in 37 (60.7%), 10 (16.4%), and 14 (22.9%) patients, respectively (P<0.0001, Chi-square test).
The changes in histological outcomes were less pronounced in patients who continued TDF from the previous trial (Fig. 4B). Among them, there were more patients with a decrease in the Ishak fibrosis stage (n=32, 51.6%) than those with no changes (n=19, 30.7%) or an increase (n=11, 17.7%) (P=0.0044, Chi-square test). However, no significant differences were observed for the changes in necroinflammatory scores, with similar proportions of patients with decreases (n=21, 33.9%), not changes (n=19, 30.7%), and increases (n=22, 35.4%), respectively (P=0.89, Chi-square test).
Virological, biochemical, serological, and safety outcomes
The secondary outcomes were reported from 127 patients retained at the end of the open-label trial, regardless of liver biopsy or not. Among them, 105 patients (82.7%) achieved virological remission with serum HBV DNA below detectable limits, significantly increasing from 61 (48.0%) patients at baseline (P<0.0001, Stuart–Maxwell test). Similarly, the proportion with ALT normalization (<40 U/L) significantly increased from 54.3% (n=69) to 69.3% (n=88) after t reatment (P=0.0071, Stuar t–Maxwel l test). Quantitatively, serum HBV DNA levels significantly decreased from 1.36 log IU/mL (95% CI, undetectable to 4.16) to undetectable levels (P<0.0001, signed rank test). Serum ALT also decreased from 29 U/L (95% CI 37–50) to 25 U/L (95% CI 31–44) in the entire study population (P=0.048, signed rank test). However, the improvements were con-fined to the patients initially assigned to placebo. No significant changes were observed in patients already on TDF treatment (Table 3).
No patients cleared HBsAg during the 3-year study period. The serum levels of HBsAg did not significantly change, either overall or in each subgroup according to the previous intervention (Table 3). Four of the 27 HBeAg-positive patients were lost to follow-up and 5 of the remaining 23 (21.7%) patients became negative for HBeAg at the study end.
There were no serious adverse events attributable to the study medication (i.e., TDF). Eight patients were removed from the study because of developing hepatocellular carcinoma (n=5, 3 previously assigned to placebo and 2 to TDF), lung cancer (n=1), lymphoma (n=1), and HIV infection (n=1). Another 11 patients chose to drop out of the study because of personal choices unrelated to adverse events.
Post-hoc comparisons by treatment allocation in the TORCH-B trial
Baseline data at randomization in the TORCH-B trial for the current study participants are summarized by the randomized allocation in Supplementary Table 3. In the post-hoc analysis comparing changes in virological, biochemical, and histologic parameters from the TORCH-B baseline to the end of the current study, no significant differences were observed based on treatment allocation (Supplementary Table 4). Similarly, the two treatment groups showed no significant differences in the virological, biochemical, and histologic parameters measured at the end of this open-label study (Supplementary Table 5).
DISCUSSION
In this roll-over study from the TORCH-B trial, we found significant improvements in hepatic histopathology after treatment with TDF for 3 years. Over 60% of the study participants exhibited a decrease in the fibrosis stage, and the proportion of patients with an Ishak score ≥3 decreased from 26.8% to 9.8%. Besides, nearly half of the participants showed a decrease in necroinflammatory scores, and the proportion with a Knodell score ≥4 points decreased from 20.3% to 10.6%. The changes in histological outcomes were more pronounced in patients switching from placebo. In patients who continued TDF from the TORCH-B trial, there were significant decreases in the fibrosis stage without changes in the inflammation scores. Serum levels of HBV DNA and ALT significantly decreased in the placebo switchers and remained stable without significant changes for patients who had been taking TDF. Nevertheless, serum levels of HBsAg did not change and no patients cleared HBsAg. The safety analyses were unremarkable, with serious adverse events attributable to the study medication.
Our study demonstrated the treatment effects of TDF in patients with minimally raised ALT, a population for whom the treatment indications remain contentious [7-9]. In a study population characterized by minimal or mild fibrosis (104 of 146 or 71.2% with an Ishak score <3 points at baseline), we found that the severity of fibrosis was reduced in 60.2% of participants following 3 years of antiviral treatment. Notably, even among patients who had already been treated with TDF for 3 years in the TORCH-B trial, 51.6% showed further reduction in liver fibrosis severity after an additional 3 years of treatment. This finding underscores the beneficial effects of continued antiviral therapy on histopathological outcomes.
While the significant decreases in inflammatory scores, viral loads, and serum ALT levels in the placebo switchers clearly suggest treatment effects in treatment-naïve patients [21], the absence of such changes in patients already on TDF should not be interpreted as a lack of benefit. After 3 years of antiviral therapy in the TORCH-B trial, hepatic inflammation was clinically insignificant in most of these TDF-experienced patients (Knodell score <4 in 64 out of 73, or 87.7%, of this subgroup). Additionally, most of them had already achieved virological remission with serum ALT levels within normal limits [22]. Therefore, the lack of significant changes in these outcomes in the TDF-experienced group may be interpreted as maintenance of treatment effects, rather than a decline in efficacy, in patients continuing antiviral therapy.
We found that histopathology worsened in a minority of patients on antiviral therapy, consistent with a recent study that reported fibrosis progression in 189 of 733 (25.8%) treated patients, based on paired liver biopsies on therapy [23]. The reasons for progression despite treatment are likely multifactorial. For instance, the treatment may not completely shut down intracellular viral activities [24], which can lead to cellular damages in the absence of clinically overt hepatitis. The risk of incomplete viremia remission contributing to fibrosis progression is well documented [25]. Our post hoc analysis suggested fatty liver may also be contributory to histopathological progression. Additional factors such as alcohol consumption, medication use, or other environmental exposures could also affect the outcomes. Notably, although patients with alcoholic hepatitis were excluded from entering the TORCH-B trial, participants’ drinking behaviors might have changed during the 6-year study period.
While TDF therapy effectively improved liver histology, it did not significantly affect HBsAg levels, and no patients achieved HBsAg clearance in this open-label trial. This finding is consistent with existing literature regarding the limited effects of NUC therapy on serological endpoints [26]. The persistence of HBsAg is likely due, at least in part, to integrated HBV DNA continuing to produce HBsAg in patients on long-term NUC therapy [27]. Therefore, novel agents or strategies are needed to achieve functional cure of CHB [28,29], as current NUC monotherapy, despite its other benefits, falls short in this therapeutic efficacy [30].
Our findings align with emerging studies that reported clinical effectiveness of antiviral therapy for CHB patients without typical features of active hepatitis or inactive infection, often referred to as the “gray zone”. In a multicenter cohort study involving 855 adults with CHB managed at 14 centers across America, Europe, and Asia, Huang and colleagues found that antiviral therapy was associated with a significantly lower risk of HCC, with an adjusted hazard ratio of 0.3 (95% CI 0.1–0.6, P=0.001) [31]. Moreover, the preliminary results of a multinational multicenter randomized trial led by Lim and colleagues (the ATTENTION trial) indicated significant reduction in composite clinical events (including HCC, death, and hepatic decompensation; hazard ratio 0.21, 95% CI 0.05–0.97; P=0.027) in patients randomized to antiviral therapy [32]. These lines of evidence, along with the current study and the TORCH-B trial, support the recent calls for expanding treatment eligibility to include patients in the gray zones [14,15,33,34].
Expanding treatment eligibility will carry significant implications for resource allocation, particularly in resourceconstrained regions where existing gaps in access to antiviral therapy are already profound [1]. Therefore, parallel efforts in addition to expansion of treatment eligibility are essential to prevent exacerbating existing inequalities and to ensure that expanded criteria translate into improved outcomes across all affected populations. Drawing from global HIV control efforts [35], we suggest emphasizing sustainable funding models, access to generic medications, decentralized healthcare, industry partnerships, and public awareness campaigns to scale up HBV treatment.
It is important to note that neither the TORCH-B trial nor the current open-label study was specifically designed to assess the urgency of initiating treatment immediately in “gray zone” patients. Our post-hoc analyses might suggest that the study outcomes were not significantly different whether treatment was delayed by three years of placebo, provided that placebo recipients were closely monitored and promptly treated in cases of acute hepatitis flares [18]. Nevertheless, the study designs, including sample size estimations and outcome definitions, were not dedicated to such comparisons. Therefore, specifically designed studies― preferably large, randomized controlled trials with long-term follow-up focused on clinical outcomes―are needed to evaluate whether, or how, the immediacy of antiviral therapy might impact outcomes in “gray zone” patients.
Our study has several strengths that include the use of paired liver biopsies, a clearly specified study population, and a high retention rate through the study. However, the following limitations are acknowledged. Firstly, the open-label, single-arm design could potentially introduce bias. It is important to note that this design was a consequence of the study being a rollover extension from the original TORCH-B trail. Secondly, the lack of a control group in this extension phase makes it challenging to draw definitive causal inferences. Although we can compare outcomes to baseline measurements with reasonable assumptions, we cannot completely rule out the influence of natural disease progression or regression that might have occurred independently of treatment. Thirdly, the study endpoints are arguably not “hard” clinical outcomes such as occurrence of liver cancer or resultant mortality. Nevertheless, liver histopathology (particularly fibrosis) is known to predict clinical complications in patients with CHB [36]. Lastly, the relatively small sample size may limit the generalizability of our findings to all patients in the gray zone of CHB treatment. Indeed, our study population was predominantly male and Asian. Given that the manifestations of CHB and treatment responses can vary across different ethnicities and genders, our results should be interpreted cautiously when considering application to more diverse populations.
In conclusion, this study demonstrates favorable outcomes in histopathological, virological, and biochemical endpoints following three years of open-label TDF therapy for CHB. These findings, building upon our previous placebo-controlled trial that focused on patients whose ALT levels did not fulfill treatment eligibility criteria, provide further evidence to support the initiation of antiviral therapy in this patient population and the continuation of treatment in those already receiving it. Our results suggest that patients with minimally raised ALT, often considered to be in a “gray zone”, may benefit from antiviral therapy, potentially preventing disease progression and improving long-term outcomes.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
Supplement1
Print
