Background
Approximately 240 million people worldwide are affected by chronic hepatitis B (CHB) [
1]. Complications occur in approximately 20 to 30% of CHB patients, resulting in a CHB-related mortality rate of 650,000/year globally. In China, 90 million people are affected, accounting for almost 7% of the total population. The annual hepatitis B virus (HBV)-related cancer mortality rate in China is 330,000/year [
2,
3]. Because CHB is a significant risk factor for cirrhosis and hepatocellular carcinoma (HCC), intensive efforts have focused on the development of effective therapeutic strategies that aim for the clearance of HBV as measured by hepatitis B surface antigen (HBsAg), hepatitis B e antigen (HBeAg), viral DNA, normalization of hepatic function indicated by alanine transaminase (ALT), and slowing the progression of cirrhosis and inflammation in the liver [
4]. Current therapeutic guidelines recommend the use of oral nucleosides/nucleotide analogues (NAs) and interferon-based therapy with interferon-alpha or pegylated interferon-alpha for patients with HBeAg-positive CHB [
5,
6].
NAs effectively suppress viral replication and reduce disease-related morbidity and mortality. NAs require long-term treatment and are associated with the development of drug resistance and high relapse rates [
7]. Interferon-based therapy has a finite treatment course with no associated drug resistance, although more frequent adverse events and the route of administration are of clinical concerns [
8‐
10]. The major advantage of pegylated interferon (PEG-IFN) therapy is the induction of a robust off-treatment response along with increased virological response during the follow-up period, reflecting a post-treatment delayed immune response [
11]. Randomized controlled trials (RCT) have demonstrated PEG-IFN treatment resulted in HBeAg seroconversion in approximately one-third of HBeAg-positive CHB patients, including those who were previously treated [
12‐
14].
Studies investigating the long-term clinical responses to PEG-IFN in Europe, Asia, and the USA have reported sustained response rates ranging from 21 to 60%, depending on the treatment regimen, the definition of response, and the length of follow-up [
1,
15‐
17]. Some CHB patients, nevertheless, had treatment relapse after a certain period of post-treatment observation, especially among those co-infected with HBV and HCV [
14,
18,
19]. Thus, evaluating the sustainability of treatment response to PEG-IFN treatment in the long term is in need for patients with CHB.
In China, evidence of a long-term follow-up (LTFU) on PEG-IFN off-treatment response among HBeAg positive patients is very limited. Findings from an Asian multicenter RCT (P05170) favored PEG-IFN alfa-2b dosing scheme by 1.5 μg/kg/wk. for 48 weeks over other dosing schemes by 1.0 or 1.5 μg/kg/wk. for 24 weeks for HBeAg seroconversion among patients with CHB (with > 95% were Chinese) [
20]. Therefore, this RCT provided us an opportunity to investigate the sustainability of complete response (CR) and serological response (SR) in Chinese CHB patients treated with the different PEG-IFN alfa-2b dosing schemes in a relatively longer off-treatment duration [
20].
Material and methods
Study design and patients
This study was designed as a LTFU for the original multicenter, randomized, controlled study (P05170) [
20] in which 657 interferon-naïve CHB patients were randomized to receive different PEG-IFN alfa-2b regimens (1.0 μg/kg/week for 24 weeks; 1.5 μg/kg/week for 24 or 48 weeks) followed by 24 weeks of post-treatment observation [
20]. The LTFU was conducted at 21 out of 25 P05170 sites in China. Enrollment began in May 2014 and ended in November 2015 due to no potential to recall any more subjects from P05170. MSD China sponsored the study and designed and analyzed the study in collaboration with leading investigators. This study was conducted in accordance with the guidelines of the Declaration of Helsinki and the principles of Good Clinical Practice (ICH-E6) and was approved by independent ethics committees of all study sites before the study enrollment commenced. Medical Ethics Committee of NanFang Hospital of Southern Medical University (NFEC-201402-Y2).
Patients were eligible to the LTFU if they were randomized and treated at the sites from P05170 in mainland China and willing to give consent. Patients were excluded if they were pregnant or nursing or had a history of any illness that, in the opinion of the investigator, might affect the completion of LTFU. Subjects qualified for the LTFU were contacted by telephone or physical visit upon the approval from the local ethics committees and were screened by the investigator at an outpatient hepatology clinic in the participating sites. All patients completed the LTFU by one onsite visit once included. A separate and specific signed informed consent was obtained from all patients prior to study screening for participation in the LTFU.
Clinical and laboratory measurements
The clinical and laboratory measurements of P05170 were reported previously [
20]. Briefly, assays for HBV-DNA, HBV genotyping and virology, and clinical assessment, including liver biopsies, were conducted at baseline, at the end of treatment, and at a 24-week post-treatment visit for all randomized patients.
During the LTFU, the investigator assessed clinical signs, symptoms, and complications of liver disease (i.e., hepatocellular carcinoma, ascites, variceal bleeding, encephalopathy, jaundice, or liver transplant), and the administration of (any other) antiviral therapy after the end of P05170. Prior and concomitant medications at the LTFU for the treatment of CHB and/or other active medical conditions were documented. Co-medications dated during the period of original study observation were cross-checked by data records from P05170 [
20]. Any supporting documents, including previous test results, were provided to the investigator to assess the clinical significance of additional antivirals after the end of the original study.
Blood samples for virological biomarkers (HBV-DNA, HBsAg/Ab, and HBeAg/Ab) were collected at the LTFU and sent to a central laboratory for measurement (LabCorp, Beijing, China). HBV DNA quantification at LTFU was performed using Taqman-based polymerase chain reaction (PCR) assays (COBAS AmpliPrep/COBAS TaqMan48, Roche Molecular Systems Inc.) with a lower detection limit of 20 IU/mL. Quantification of HBsAg levels were measured with Elecsys HBsAg II quantitative electrochemiluminescence immunoassay kits (Roche Diagnostics, Indianapolis, IN) using Modular Analytics E170 (Roche Diagnostics, Switzerland) with a range of 5–52,000 IU/mL for up to 400-fold dilutions of samples. HBeAg levels were quantified using the Elecsys HBeAg II quantitative assay using Modular Analytics E170 with a lower quantification limit of 5 PEIU/mL. Hematology and chemistry were measured at the participating sites using local laboratory standards.
Clinical endpoints
The study endpoints were sustained complete response (CR) and sustained serological response (SR) at the LTFU. Sustained serological response was defined as HBeAg loss and seroconversion to anti-HBe at the end of the P05170 study and at the LTFU. Sustained complete response was defined as HBeAg loss and seroconversion to anti-HBe as well as HBV-DNA < 2000 IU/mL at the end of P05170 study and at the LTFU.
Statistical analysis
Sample size
The rate of subject recall from P05170 determined the sample size for the LTFU rather than any sample size consideration based on inference statistic and power calculation. It was anticipated that 60% of subjects from the full analysis set (FAS) in P05170 would be included in the LTFU, representing a total sample size of 395 subjects. Study enrollment stopped when 322 subjects were screened and included in the LTFU, as the study team concluded that there was no potential to find any more subjects from the original study, albeit all practiced efforts.
Analysis of endpoints
A point estimate (proportion) for the study endpoints was presented by different treatment schemes groups. Off-treatment sustainability was calculated as the number of patients who had sustained SR or CR at the LTFU divided by the number of patients who had SR or CR at EOS (of P05170). The duration of follow-up after the end of the P05170 study was calculated as the interval between the date of the last visit in P05170 and the date of LTFU visit. Missing data for study endpoints were not imputed by any means. Subjects with missing data for study endpoint assessment were considered non-responders for study endpoints. Subjects who were clinically re-treated with any antiviral therapy before and at the time of the LTFU were deemed non-responders for study endpoints. Univariate Cox regression analysis was used to identify factors associated with a sustained SR or CR in LTFU. Hazard ratio (HR) and 95% confidence interval (CI) were calculated to quantify the strength of associations. Variables with a p < 0.25 by Wald test for regression coefficient were used to perform multivariate Cox regression analysis using a forward selection procedure.
The subject’s clinical and laboratory characteristics are summarized and presented by different time points in P05170 and at the LTFU. Continuous variables are presented as the mean (standard deviation, SD); categorical variables are expressed as count and frequency (%). Group differences were compared using analysis of variance (ANOVA) or χ2 test whenever appropriate. All statistical analyses were performed using NCSS 10 software (NCSS, LLC, Kaysville, UT). All statistical tests were two-sided and performed at the 0.05 level of significance unless otherwise specified.
Discussion
This long-term follow-up study was an extension of our previous study [
20] and compared the sustainability of three PEG-IFN alfa-2b dosing schemes for the treatment of Chinese HBeAg-positive patients with CHB six years post-treatment. Findings from the study confirmed a good sustainability of PEG-IFN alfa-2b off-treatment response in the long term and the most efficacious was seen in the patient group treated with 1.5 μg/kg/wk. for 48 weeks. Other important findings suggested an increase in the rates of HBe seroconversion over time in patients receiving PEG-IFN alfa-2b 24-week dosing schemes and overall low-to-moderate rates of SR and CR in patients at the LTFU in all treatment groups with or without additional NA therapies.
The findings of the present study are consistent with earlier studies that investigated the sustainability of treatment response in which duration of treatment was found as an important factor associated with the long-term durability of PEG-IFN therapy in HBeAg-positive patients [
8,
14,
15,
21‐
24]. The response rates in this study are comparable to those reported previously [
8,
14,
15,
21‐
24]. Wong et al. (2010) found that the response (HBe seroconversion with HBV-DNA < 10,000 IU/mL) to PEG-IFN alpha-2b was sustained for five years in HBeAg-positive CHB patients; 82 and 57% had sustained HBeAg seroconversion and virologic response, respectively, at the end of follow-up [
21]. Okanoue et al. (2016) showed that patients with CHB infection (
N = 137) sustained a good response (normal ATL levels and low-level of hepatitis B virus) five years post-PEG-IFN alpha-2a treatment, and a 48-week administration of PEG IFN-α-2a showed a better response rate (26.4%) than a 24-week administration (18.0%) [
22]. However, the study revealed that > 60% of patients received additional antiviral therapy due to viral reactivation [
22]. In a small Japanese study, Masaki et al. (2015) also found that a 48-week regimen of PEG-IFN alfa-2a had significantly higher rates of HBeAg clearance, ALT normalization, and HBV-DNA loss at the end of five years [
15].
Considering additional antivirals before the LTFU, we evaluated the relapse rates in patients who achieved SR or CR at the end of the original study and the clinical endpoints in patient subgroups with or without additional NAs at the LTFU time point. Patients treated with the 48-week PEG IFN alfa-2b dosing scheme had the lowest need for additional NA therapy. However, additional NA treatment after initial PEG-IFN dosing showed better response rates in normalizing ALT and slightly higher response rates in reducing HBsAg load for patients receiving 24-week PEG-IFN dosing schemes. It should be cautious in the clinical practice that overall rates of SR and CR were generally not satisfactory for CHB patients at the LTFU. On the other hand, we found SR/CR response rates between 18 and 30% at the LTFU for patients who had failed such responses at the end of original study but were free of additional NAs, which were similar to results reported from a previous study [
25]. In particular, patients who received a 24-week dosing scheme had an increase in HBe seroconversion from the end of previous study treatment through the LTFU. These findings further supported a longer-term immunomodulatory effect with PEG-IFN treatment in at least in a small fraction of HBeAg-positive CHB patients.
A number of studies have investigated baseline and early on-treatment predictive factors associated with the durability of response to PEG-IFN therapy. Early HBeAg loss after PEG-IFN alfa-2b therapy was associated with viral DNA suppression and HBsAg loss at a 3-year LFTU visit [
26]. Sonneveld et al. (2010) found that 97% of HBeAg-positive CHB patients treated with PEG-IFN who showed no decrease in serum HBsAg levels at 12 weeks of treatment showed a non-response at three years [
27]. Additionally, Sonneveld et al. (2012) demonstrated that PEG-IFN alfa-2b-treated CHB patients who had a complete response (HBeAg loss along with HBV DNA < 10,000 copies/mL) showed a sustained HBsAg decline after three years [
28]. Ma et al. (2010) showed that on-treatment serum HBsAg and HBeAg levels were significant predictors of sustained response to PEG-IFN alfa-2b treatment for CHB [
24] which was supported by a study from Wang et al. (2016) [
29].
Virological load has also been shown to be a significant predictor for sustained treatment responses [
21]. Okanoue et al. found that low HBV-DNA levels at the end of treatment were independently associated with favorable five-year response rates [
22]. A long-term follow-up study of 21 Taiwanese HBeAg-positive patients with CHB at 6.5–12.5 years of IFN-alpha post-treatment found that pre-treatment with HBV DNA load (< 2 × 10
8 copies/mL) predicted sustained cumulative virological response [
30]. Other factors including younger age and baseline ALT levels were explored to have an association with a sustained virological response [
31].
We created Cox regression models to explore baseline and on-treatment predictors, including demographics and CHB biomarkers for sustained SR and CR. Patients carrying genotype B HBV were significantly less likely to have a sustained SR compared to those carrying other genotypes (genotype C: 54.0%). In addition, the dosing scheme by 1.5 μg/kg/wk. for 48 weeks was the only associated factor for sustained CR at the LTFU. Our results indicate that genotype B is associated with a reduced chance of sustained SR compared with other genotypes, which differs from the findings of previous studies. Zhao et al. (2007) found that genotype B and younger patient age (< 25 years) were independently associated with sustained CR and a better response to low-dose therapy among Chinese CHB-infected patients who were HBeAg positive (
N = 230) and had received PEG-IFN alpha-2b therapy for 24 weeks [
32]. In addition, Buster et al. (2008) found that the durable loss of HBeAg three years after PEG-IFN treatment was greater in patients carrying genotypes A and B than in those carrying genotypes C and D [
26]. A pooled analysis of two large trials of HBeAg-positive patients treated with PEG-IFN found that higher levels of ALT and lower levels of HBV-DNA were predictive of sustained response for patients infected with genotypes A, B, and C [
33] and genotype D had the lowest chance of sustained response, irrespective of HBV-DNA or ALT levels. In addition, Chen et al. (2011) found that genotype B was associated with better immediate, late, and sustained response at 24 weeks of follow-up in HBeAg-positive patients treated with PEG-IFN alpha-2a for six months [
34]. The reasons for the differences between our findings and those of previous studies are not clear but may reflect the fact that most of our patients were infected with HBV of genotype B or C (98.4%); hence, our results primarily reflect differences between these two genotypes.
To the best of our knowledge, the LTFU is the first study evaluating long-term clinical data in Chinese CHB patients treated with PEG-IFN alfa-2b who were followed through a 6-year (312-week) period. Our study patients were selected from those who had completed our previously described P05170 study [
20]. Sustained serological and complete responses were pre-defined and assessed throughout the RCT (P05170) until the LTFU. The present study was able to capture clinical information on additional antiviral use during the follow-up period. Patients were deemed non-responders if additional NA therapy was initiated after the completion of the original study or had any missing data regarding clinical response assessment.
This LTFU study has certain limitations. First, fewer than half of patients from P05170 were recalled and included in the LTFU. Patients with uncontrolled HBV infections may have sought additional medical treatment at other facilities. Moreover, those who had good long-term responses may have chosen to ignore this follow-up contact. Therefore, the sustainability and long-term clinical outcomes may be over- or underestimated. The LTFU was designed as one follow-up visit rather than as a continuous follow-up protocol by specific time-intervals. We did not have the opportunity to assess the exact time point for relapse in patients who achieved previous favorable responses or patients who had favorable responses despite their previously failed clinical outcomes. Lastly, due to lower-than-anticipated recall rates for patients treated by different PEG IFN dosing schemes, liver-related events were not sufficiently assessed and thus, liver related complications may be underreported.