SEARCH

SEARCH BY CITATION

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Background

The roles remain unclear of early on-treatment quantitative serum HBsAg and hepatitis B virus (HBV) DNA levels in the prediction of a sustained response (SR) to peginterferon alfa-2a therapy in HBeAg-negative chronic hepatitis B (CHB) patients infected with genotype B or C.

Aims

To determine their roles in HBeAg-negative CHB patients infected with genotype B or C.

Methods

Sixty-one patients were treated with peginterferon alfa-2a for 48 weeks. Serum HBsAg levels were quantified using the Abbott Architect HBsAg QT assay throughout treatment. Multiple regression analyses were performed to identify independent predictors of SR.

Results

Nineteen patients (31%) achieved SR with serum HBV DNA levels <312 copies/mL at 24 weeks post-treatment. Serum HBsAg levels at 12 (OR 31.9; 95% CI 4.8–209.6; = 0.0003) and 24 weeks of therapy (OR 8.8; 95% CI 2.0–38.0; = 0.0035), and HBV DNA levels at baseline (OR 7.0; 95% CI 1.3–36.2; = 0.0203), 12 (OR 7.9; 95% CI 1.2–48.4; = 0.0249) and 24 weeks of therapy (OR 22.3; 95% CI 2.2–224.0; = 0.0083) were early independent predictors of SR. A serum HBsAg cut-off of 150 IU/mL at week 12 had an AUC, sensitivity, specificity and positive and negative predictive values of 0.75, 63%, 95%, 86% and 85% with respect to predicting SR respectively.

Conclusions

A quantitative serum HBsAg level at 12 weeks of therapy can be used for the early prediction of SR to peginterferon therapy in HBeAg-negative CHB patients infected with genotype B or C.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Chronic hepatitis B (CHB) is one of the most important health issues, affecting more than 350 million people worldwide.[1] In the natural history of CHB, hepatitis B e antigen (HBeAg) seroconversion usually leads to resolution of clinical hepatitis activity.[1] However, some patients develop hepatitis B virus (HBV) reactivation at various intervals following HBeAg seroconversion and evolve into HBeAg-negative CHB.[1, 2] Reactivation of HBV is a significant predictor of progression to liver cirrhosis.[1, 2]

The treatment options for HBeAg-negative CHB patients include oral nucleos(t)ide analogues and interferon.[3-6] Although nucleos(t)ide analogues are potent inhibitors of HBV replication, easy to administer and without significant side effects, long-term maintained viral suppression is generally believed to be required for effective control of disease progression.[3-7] Interferon functions through both immunomodulatory and antiviral activities. Although interferon is commonly associated with side effects, interferon only requires treatment for a finite duration.[3-6] Studies on peginterferon alfa-2a have shown that approximately 25% of patients remained in remission, 3 years after the cessation of 48 weeks of therapy.[8, 9] Moreover, peginterferon alfa-2a therapy can achieve hepatitis B surface (HBsAg) seroclearance in 8% of patients,[9] which is the desirable parameter closest to a clinical cure of CHB and associated with a reduced incidence of cirrhosis and hepatocellular carcinoma (HCC) and better survival.[3-6]

Recently, the quantitative serum HBsAg level or the magnitude of decline in serum HBsAg level during peginterferon therapy has been shown to be predictive of a sustained response (SR) to peginterferon therapy in HBeAg-negative CHB patients.[10-12] A combination of serum HBsAg and HBV DNA levels was further proposed as a strong negative predictor to peginterferon therapy in HBeAg-negative CHB patients.[13] However, these studies were conducted in CHB patients mostly infected with genotype A or D. HBV genotypes have been shown to influence serum HBsAg kinetics during peginterferon therapy in HBeAg-negative CHB patients.[14] The predictive roles of serum HBsAg and HBV DNA levels in patients infected with genotype B or C are less clear.

The purpose of this study was to determine the roles of early on-treatment serum HBsAg and HBV DNA levels in the prediction of a SR to peginterferon alfa-2a therapy in HBeAg-negative CHB patients infected with genotype B or C.

Patients and methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Patients

Consecutive CHB patients who visited the Liver Unit at the China Medical University Hospital between November 2004 and July 2009 were considered for peginterferon alfa-2a therapy if they met the following criteria: (i) age ≥18 years and <70 years; (ii) positive serum HBsAg for at least 6 months; (iii) negative for serum HBeAg with or without the presence of anti-HBe antibody for at least 3 months; (iv) serum alanine aminotransferase (ALT) levels >2X (except in patients with cirrhosis) and ≦10× the upper limit of normal (ULN; 40 IU/L) twice 3 months apart; (v) serum HBV DNA levels ≥10 000 copies/mL; and (vi) a liver biopsy showing findings compatible with CHB within 12 months prior to evaluation. The exclusion criteria were as follows: decompensated liver disease; evidence of HCC; co-existing serious medical or psychiatric disease; co-infection with hepatitis C virus, hepatitis D virus or human immunodeficiency virus; history of alcohol or drug abuse; history of previous therapy with interferon; and previous nucleos(t)ide therapy for CHB within 6 months prior to evaluation. All patients signed written informed consent, and this retrospective study was approved by the Institutional Review Board of the hospital.

Treatments

All patients were prospectively followed and received peginterferon alfa-2a therapy at a dose of 180 μg weekly for 48 weeks with 24 weeks of post-treatment follow-up. Serum ALT levels were measured biweekly during the first 3 months, then monthly until 12 weeks after the end of therapy and at 24 weeks of post-treatment follow-up. Serum HBeAg and anti-HBe were assayed with commercially available enzyme immunoassays (Vitros ECi; Ortho-Clinical Diagnostics, High Wycombe, Buckinghamshire, UK) at baseline, and then every 24 weeks until the end of follow-up. Serum HBsAg levels at baseline, during therapy (12, 24 and 48 weeks) and at 24 weeks post-treatment were retrospectively quantified using the Abbott Architect HBsAg QT assay (dynamic range, 0.05–250 IU/mL). Serum HBV DNA levels were measured at baseline, during therapy (12, 24 and 48 weeks) and at 24 weeks post-treatment with the Cobas Amplicor HBV Monitor Test (Roche Diagnostics, lower limit of detection, 312 copies/mL). HBV genotyping was performed using the polymerase chain reaction (PCR)-restriction fragment length polymorphism method, as previously described.[15] Liver histology was graded and staged according to the METAVIR scoring system.[16]

Efficacy measures

Efficacy analyses included the percentages of patients achieving serum HBV DNA levels <10 000 copies/mL or the lower limit of detection (312 copies/mL), mean HBV DNA levels and HBV DNA declines from baseline at the end of therapy and at 24 weeks post-treatment. A SR was defined as PCR-undetectable serum HBV DNA (<312 copies/mL) at 24 weeks post-treatment. Relapse was defined as those patients who achieved PCR-undetectable serum HBV DNA at the end of therapy, but showed subsequent detectable serum HBV DNA at 24 weeks post-treatment. Nonresponse was defined as detectable serum HBV DNA at the end of therapy. Patients with relapse and nonresponse were grouped as non-SR.

Statistical analysis

Continuous data are expressed as the mean ± standard deviation (s.d.) or the median ± interquartile range (IQR) as appropriate. Normality of the data was examined using the Anderson–Darling test. Categorical data are expressed by frequency and proportion. The baseline and on-treatment factors were univariately considered using logistic regression for assessing the association with the SR in terms of odds ratios (ORs) and 95% confidence intervals (CIs). The area under the receiver operating characteristic (ROC) curve (AUC) as well as sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) were calculated for evaluating the discriminatory ability. A stepwise selection approach with or without potential confounders in the logistic regression model was used to select key predictors among the baseline and on-treatment factors. The same on-treatment factor (i.e. HBsAg levels at various time points) was picked with or without confounders in the model. All statistical analyses were performed using SAS v9.2 software (SAS Institute, Cary, NC, USA). A P value <0.05 was considered statistically significant.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Baseline characteristics of the patients

There were 61 patients, including 50 men and 11 women with a mean age of 43.5 ± 12.1 years. The median serum ALT level was 78.0 ± 89.5 IU/L. At the time of biopsy, the median necroinflammatory score was 1.00 ± 1.00 and the median fibrosis score was 2.00 ± 1.00. Forty-eight (79%) and 13 patients (21%) were infected with HBV genotypes B and C respectively. The mean serum HBV DNA level was 6.33 ± 1.74 log10 copies/mL. The median serum HBsAg level was 3.03 ± 0.64 log10 IU/mL. The baseline characteristics between patients infected with genotypes B and C were not statistically different (Table S1).

Therapeutic efficacy

At the end of 48 weeks of treatment, the median serum HBV DNA level and decline in HBV DNA from baseline was 2.49 ± 0.0 and 3.62 ± 2.95 log10 copies/mL respectively. Forty-seven (77%) and 56 (91%) patients had serum HBV DNA levels <312 and <10 000 copies/mL respectively (Table 1). The median serum HBsAg level and decline in HBsAg from baseline was 2.52 ± 1.22 and 0.34 ± 0.83 log10 IU/mL respectively (Table 1). At the end of the 24-week follow-up period, the median serum HBV DNA level and decline in HBV DNA from baseline was 3.71 ± 2.70 and 1.75 ± 3.63 log10 copies/mL respectively. Twenty-eight of the 47 patients with undetectable serum HBV DNA at the end of treatment had a virological relapse. Nineteen (31%) and 32 (52%) patients had serum HBV DNA levels <312 and <10 000 copies/mL respectively (Table 1). The median serum HBsAg level and decline in HBsAg from baseline was 2.46 ± 1.05 and 0.45 ± 0.76 log10 IU/mL respectively (Table 1). Two patients (3%) infected with genotype B virus lost HBsAg during study. At baseline, the serum HBV DNA levels were 8.37 and 4.41 log10 copies/mL, respectively, and the HBsAg levels were 3.19 and 1.0 log10 IU/mL respectively. Both achieved PCR-undetectable serum HBV DNA at 12 weeks of treatment and remained PCR-undetectable throughout follow-up. They became negative for serum HBsAg at 24 and 48 weeks of treatment, respectively, and remained negative throughout follow-up. The serum HBV DNA levels, a decline in HBV DNA from baseline, serum HBsAg levels and a decline in HBsAg from baseline among patients infected with genotypes B and C throughout treatment and follow-up were not statistically different (Table S1).

Table 1. Therapeutic efficacy of peginterferon alfa-2a in HBeAg-negative chronic hepatitis B
 On-treatmentFollow-up
12 weeks24 weeks48 weeks24 weeks
  1. Data are presented as the median ± interquartile range or proportion.

HBV DNA (log10 copies/mL)2.49 ± 0.792.49 ± 0.252.49 ± 0.03.71 ± 2.70
HBV DNA decline (log10 copies/mL)3.17 ± 2.813.32 ± 3.233.62 ± 2.951.75 ± 3.63
HBsAg (log10 IU/mL)2.89 ± 0.832.75 ± 1.032.52 ± 1.222.46 ± 1.05
HBsAg decline (log10 IU/mL)0.06 ± 0.370.18 ± 0.610.34 ± 0.830.45 ± 0.76
HBV DNA <312 copies/mL (%)60707731
HBV DNA <104 copies/mL (%)85859152

Univariate association between baseline or on-treatment factors and SR

Factors associated with SR were analysed. The factors included age, gender, fibrosis, viral genotype, baseline serum ALT, HBV DNA and HBsAg levels, on-treatment serum HBV DNA and HBsAg levels and magnitude of decline from baseline at weeks 12, 24 and 48. Age, gender, fibrosis, viral genotype and baseline serum ALT level were not significantly associated with SR (Table 2). The baseline serum HBsAg level was marginally associated with SR. The baseline serum HBV DNA level (OR = 7.5, ≦4.3 vs. >4.3 log10 copies/mL; 95% CI 1.6–33.9; = 0.0081) was significantly associated with SR. Using cut-off values of 2.15, 2.05 and 1.91 log10 IU/mL for serum HBsAg levels, and 0.6, 0.94 and 1.25 log10 IU/mL for a decline in HBsAg from baseline at 12, 24 and 48 weeks of treatment, respectively, serum HBsAg levels at 12 (OR = 34.2, ≦2.15 vs. >2.15; 95% CI 6.2–187.4; < 0.0001), 24 and 48 weeks of treatment, and a decline in HBsAg at 12 and 48 weeks of treatment were significantly associated with SR (Table 2). Using cut-off values of 3.0, 2.5 and 2.5 log10 copies/mL for serum HBV DNA levels, and 4.5, 5.2 and 5.2 log10 copies/mL for a decline in HBV DNA from baseline at 12, 24 and 48 weeks of treatment, respectively, serum HBV DNA levels at 24 weeks of treatment (OR = 12.2, ≦2.5 vs. >2.5; 95% CI 1.4–100.4; = 0.0197), and a decline in HBV DNA at 24 and 48 weeks of treatment were significantly associated with SR (Table 2). Using different cut-off values for serum HBsAg and HBV DNA levels and magnitude of decline from baseline from those used for association with SR, serum HBsAg levels at baseline, 12, 24 and 48 weeks of treatment, a decline in HBsAg at 48 weeks of treatment, serum HBV DNA levels at baseline, 24 and 48 weeks of treatment, and a decline in HBV DNA at 12 and 48 weeks of treatment were significantly associated with the achievement of serum HBV DNA level <10 000 copies/mL at 24 weeks post-treatment (Table S2).

Table 2. Univariate analysis of baseline and on-treatment factors associated with sustained response
FactorsOR95% CIP value
  1. CI, confidence interval; OR, odds ratio (with the latter category as the reference).

  2. NA: odds ratio cannot be estimated because of an empty cell (all SR subjects with log10 HBV DNA <2.5 copies/mL at week 48).

Age: ≦32 vs. >322.7(0.7, 10.1)0.1239
Gender: male vs. female0.7(0.1, 2.9)0.6805
Fibrosis: 3/4 vs. 1/21.9(0.6, 6.2)0.2403
Genotype: B vs. C1.6(0.4, 6.9)0.4819
Baseline ALT: >103 vs. ≦103 IU/L2.5(0.8, 7.8)0.1074

Baseline HBsAg

≦2.71 vs. >2.71 log10 IU/mL

3.5(0.9, 12.4)0.0534

Baseline HBV DNA

≦4.3 vs. >4.3 log10 copies/mL

7.5(1.6, 33.9)0.0081

HBsAg at week 12

≦2.15 vs. >2.15 log10 IU/mL

34.2(6.2, 187.4)<0.0001

HBsAg at week 24

≦2.05 vs. >2.05 log10 IU/mL

10.1(2.7, 37.5)0.0005

HBsAg at week 48

≦1.91 vs. >1.91 log10 IU/mL

11.9(3.2, 44.3)0.0002

HBV DNA at week 12

≦3.0 vs. >3.0 log10 copies/mL

2.9(0.7, 11.8)0.1243

HBV DNA at week 24

≦2.5 vs. >2.5 log10 copies/mL

12.2(1.4, 100.4)0.0197

HBV DNA at week 48

≦2.5 vs. >2.5 log10 copies/mL

NA
HBsAg decline at week 12>0.6 vs. ≦0.6 log10 IU/mL7.5(1.6, 33.9)0.0081
HBsAg decline at week 24>0.94 vs. ≦0.94 log10 IU/mL3.3(0.7, 14.4)0.0988
HBsAg decline at week 48 >1.25 vs. ≦1.25 log10 IU/mL7.6(1.8, 30.4)0.0042
HBV DNA decline at week 12>4.5 vs. ≦4.5 log10 copies/mL2.8(0.9, 9.0)0.0707
HBV DNA decline at week 24>5.2 vs. ≦5.2 log10 copies/mL4.3(1.1, 16.1)0.0298

HBV DNA decline at week 48

>5.2 vs. ≦5.2 log10 copies/mL

5.5(1.3, 22.2)0.0157

Independent predictors for SR

Multivariate logistic regression analyses were performed to identify independent predictors for SR. Age, gender, fibrosis, viral genotype and baseline serum ALT and HBsAg levels were not predictive of SR. Serum HBsAg levels at 12 (OR = 31.9, ≦2.15 vs. >2.15 log10 IU/mL; 95% CI 4.8–209.6; = 0.0003), 24 and 48 weeks of treatment, and serum HBV DNA levels at baseline, 12 and 24 weeks (OR = 22.3, ≦2.5 vs. >2.5 log10 copies/mL; 95% CI 2.2–224.0; = 0.0083) of treatment were independent predictors of SR (Table 3). When the magnitude of a decline in serum HBsAg and HBV DNA from baseline at 12, 24 and 48 weeks of treatment were analysed individually in the same logistic regression model along with age, gender, fibrosis, viral genotype, baseline serum ALT, HBsAg and HBV DNA levels, baseline serum HBV DNA level remained as an independent predictor of SR. In addition, a decline in serum HBsAg from baseline at 12 (OR = 24.8, >0.6 vs. ≦0.6 log10 IU/mL; 95% CI 2.1–281.4; = 0.0095) and 48 weeks of treatment, and a decline in serum HBV DNA from baseline at 12, 24 and 48 weeks of treatment were independent predictors of SR (Table 3). Multivariate logistic regression analyses were also performed to identify independent predictors for the achievement of serum HBV DNA level <10 000 copies/mL at 24 weeks post-treatment. Serum HBsAg levels at baseline, 12, 24 and 48 weeks of treatment, a decline in serum HBsAg from baseline at 12 weeks of treatment and serum HBV DNA levels at baseline and 48 weeks of treatment were independent predictors (Table S3).

Table 3. Multivariate analysis of baseline and on-treatment factors associated with sustained response
FactorsOR95% CIP value
  1. HBV, hepatitis B virus; CI, confidence interval; OR, odds ratio (with the latter category as the reference).

  2. NA: odds ratio cannot be estimated because of empty cell (all SR subjects with log10 HBV DNA <2.5 copies/mL at week 48).

  3. a

    Adjusted for age, gender, fibrosis, genotype and baseline alanine aminotransferase.

  4. b

    Adjusted for age, gender, fibrosis, genotype, baseline alanine aminotransferase, HBsAg and HBV DNA.

Baseline factorsa

HBsAg

≦2.71 vs.>2.71log10IU/mL

3.5(0.8, 15.2)0.0952

HBV DNA

≦4.3 vs.>4.3log10 copies/mL

7.0(1.3, 36.2)0.0203
On-treatment factorsa

HBsAg at week 12

≦2.15 vs.>2.15log10IU/mL

31.9(4.8, 209.6)0.0003

HBsAg at week 24

≦2.05 vs.>2.05log10IU/mL

8.8(2.0, 38.0)0.0035

HBsAg at week 48

≦1.91 vs.>1.91log10IU/mL

13.5(2.8, 65.0)0.0011

HBV DNA at week 12

≦3.0 vs.>3.0log10 copies/mL

7.9(1.2, 48.4)0.0249

HBV DNA at week 24

≦2.5 vs.>2.5log10 copies/mL

22.3(2.2, 224.0)0.0083

HBV DNA at week 48

≦2.5 vs.>2.5log10 copies/mL

NA
On-treatment factors†b

HBsAg decline at week 12

>0.6 vs. ≦0.6log10IU/mL

24.8(2.1, 281.4)0.0095

HBsAg decline at week 24

>0.94 vs. ≦0.94log10IU/mL

3.1(0.5, 19.2)0.2134

HBsAg decline at week 48

>1.25 vs. ≦1.25log10IU/mL

8.9(1.4, 55.1)0.0183

HBV DNA decline at week 12

>4.5 vs. ≦4.5log10 copies/mL

13.9(1.8, 104.5)0.0104

HBV DNA decline at week 24

>5.2 vs. ≦5.2log10 copies/mL

10.3(1.6, 63.1)0.0113

HBV DNA decline at week 48

>5.2 vs. ≦5.2log10 copies/mL

13.5(2.1, 85.4)0.0057

Serum HBV DNA and HBsAg levels throughout treatment among patients with SR, relapse and nonresponse

We compared the serum HBV DNA and HBsAg levels throughout treatment among patients with SR (n = 19), relapse (n = 28) and nonresponse (n = 14). The serum HBV DNA and HBsAg levels were not significantly different among the three groups at baseline. The serum HBV DNA levels were significantly different between the nonresponse and the SR groups at 12 (3.97 ± 1.55 vs. 2.77 ± 0.61 log10 copies/mL, = 0.02), 24 (4.25 ± 1.87 vs. 2.53 ± 0.15 log10 copies/mL, < 0.001) and 48 weeks of treatment (4.12 ± 1.57 vs. 2.63 ± 0.46 log10 copies/mL, < 0.001). The serum HBV DNA levels were also significantly different between the nonresponse and the relapse groups at 12 (3.97 ± 1.55 vs. 2.84 ± 0.64 log10 copies/mL, = 0.02), 24 (4.25 ± 1.87 vs. 2.72 ± 0.54 log10 copies/mL, < 0.001) and 48 weeks of treatment (4.12 ± 1.57 vs. 2.49 ± 0.00 log10 copies/mL, < 0.001). The serum HBV DNA levels were similar between the SR and relapse groups during treatment. At 24 weeks post-treatment, the serum HBV DNA levels were significantly lower in the SR group than in the relapse (2.49 ± 0.00 vs. 4.70 ± 1.64 log10 copies/mL, < 0.001) and nonresponse groups (2.49 ± 0.00 vs. 5.33 ± 1.70 log10 copies/mL, < 0.001; Figure 1a). The serum HBsAg levels were significantly lower in the SR group than in the relapse group at 12 (2.15 ± 0.98 vs. 2.96 ± 0.49 log10 IU/mL, < 0.01), 24 (1.79 ± 1.40 vs. 2.74 ± 0.61 log10 IU/mL, < 0.01) and 48 weeks of treatment (1.43 ± 1.64 vs. 2.59 ± 0.67 log10 IU/mL, < 0.01) and at 24 weeks post-treatment (1.76 ± 1.36 vs. 2.69 ± 0.68 log10 IU/mL, < 0.01; Figure 1b). The serum HBsAg levels were not significantly different between the relapse and nonresponse groups (Figure 1b). During treatment, although the serum HBV DNA levels were similar between the SR and relapse groups, the serum HBsAg levels were significantly different between them.

image

Figure 1. Serum hepatitis B virus (HBV) DNA (a) and HBsAg (b) kinetics throughout treatment and follow-up among patients with sustained response (SR), relapse and nonresponse. (a) nonresponse vs. SR: ☆, image, ◇, □; nonresponse vs. relapse: ▽, ○, △; SR vs. relapse: ◎. (b) SR vs. relapse: ☆, image, ◇, □; SR vs. nonresponse: ○, △.

Download figure to PowerPoint

Predictive values of serum HBsAg and HBV DNA kinetics for SR

As the serum HBV DNA and HBsAg levels during treatment were significant predictors of SR, we determined the predictive values of serum HBV DNA and HBsAg levels at baseline, 12 and 24 weeks of treatment on SR to identify any clinically useful guide for early treatment decisions. The sensitivity, specificity, PPV and NPV of these predictors are summarised in Table 4. Based on an analysis of the AUC, a cut-off value of serum HBsAg level at 2.15 log10 IU/mL (=150 IU/mL) at 12 weeks of treatment gave the best AUC (=0.75) with a sensitivity of 63%, a specificity of 95%, a PPV of 86% and a NPV of 85%. A cut-off value of serum HBV DNA level at 3.0 log10 copies/mL at 12 weeks of treatment only gave an AUC of 0.58 with a sensitivity of 84%, a specificity of 35%, a PPV of 37% and a NPV of 83%. Combining these two predictive factors conferred a limited further increase in predictive accuracy (AUC, 0.78; sensitivity, 52%; specificity, 95%; PPV, 83%; and NPV, 82%; Table 4). The predictive values of the decline in serum HBV DNA and HBsAg from baseline at12 and 24 weeks of treatment were inferior to the absolute levels (Table S4). The predictive values of serum HBV DNA and HBsAg levels at baseline, 12 and 24 weeks of treatment and the decline in serum HBV DNA and HBsAg from baseline at 12 and 24 weeks of treatment for the achievement of serum HBV DNA <10 000 copies/mL at 24 weeks post-treatment were also determined. The serum HBsAg levels at 12 (cut-off: 2.3 log10 IU/mL) and 24 weeks (cut-off: 1.9 log10 IU/mL) of treatment had a lower predictive value for the achievement of serum HBV DNA <10 000 copies/mL at 24 weeks post-treatment than for SR (Table S5). Combining the serum HBV DNA level conferred no further increase in predictive accuracy (Table S5). Again, the predictive values of the decline in serum HBV DNA and HBsAg from baseline at 12 and 24 weeks of treatment were inferior to the absolute levels (Table S6).

Table 4. Predictive values of baseline and on-treatment HBV DNA and HBsAg levels for sustained response
Predictive factorsCut-offaAUCSESPPPVNPV
  1. HBV, hepatitis B virus; AUC, area under the receiver operating characteristic curve; NPV, negative predictive value; PPV, positive predictive value; SE, sensitivity; SP, specificity.

  2. a

    Units for HBV DNA and HBsAg are log10 copies/mL and log10 IU/mL respectively.

Log HBV DNA at baseline4.30.5080.3680.9290.700.76
Log HBV DNA at week 123.00.5810.8420.3570.370.83
Log HBV DNA at week 242.50.6820.9470.4050.420.94
Log HBsAg at baseline2.710.6040.3680.8570.540.75
Log HBsAg at week 122.150.7560.6320.9520.860.85
Log HBsAg at week 242.050.7340.5790.8810.690.82
Log HBV DNA/Log HBsAg at week 123.0/2.150.7830.5260.9520.830.82
Log HBV DNA/Log HBsAg at week 242.5/2.050.7970.5790.8810.690.82

Four of the 19 patients achieving serum HBV DNA levels <312 copies/mL and 6 of the 32 patients achieving serum HBV DNA levels <10 000 copies/mL did not have their serum ALT normalised at 24 weeks post-treatment. Obesity was suspected to account for the elevated serum ALT levels among them. We reasoned that the serum HBV DNA level alone may perform better as a criterion than the combined endpoint. On the basis of an analysis of the AUC, we compared the relative discriminatory abilities of serum HBsAg levels at baseline, 12 and 24 weeks of treatment for the achievement of (i) serum HBV DNA <312 copies/mL, (ii) serum HBV DNA <312 copies/mL and normal ALT, (iii) serum HBV DNA <10 000 copies/mL, and (iv) serum HBV DNA <10 000 copies/mL and normal ALT, respectively, at 24 weeks post-treatment (Table S7). The serum HBsAg level at 12 weeks of treatment had a higher predictive value for the achievement of serum HBV DNA <312 copies/mL than for serum HBV DNA <312 copies/mL and normal ALT at 24 weeks post-treatment. The serum HBsAg level at 12 weeks of treatment had a similar predictive value for the achievement of serum HBV DNA <10 000 copies/mL and for serum HBV DNA <10 000 copies/mL and normal ALT at 24 weeks post-treatment (Table S7).

We also used the algorithm proposed by Rijckborst et al.[13] to examine its applicability to our patient cohort. The patients with no decline in serum HBsAg or less than 2 log10 copies/mL decline in serum HBV DNA at 12 weeks of treatment had 25% chance of achieving serum HBV DNA <312 copies/mL, <10 000 copies/mL or <10 000 copies/mL with normal ALT at 24 weeks post-treatment. The patients with any decline in serum HBsAg or ≥2 log10 copies/mL decline in serum HBV DNA or both at 12 weeks of treatment had 30–35%, 49–66% and 38–66% chances of achieving serum HBV DNA <312 copies/mL, <10 000 copies/mL and <10 000 copies/mL with normal ALT at 24 weeks post-treatment respectively (Figure 2).

image

Figure 2. Chances of achieving sustained response, hepatitis B virus (HBV) DNA <10 000 copies/mL or HBV DNA <10 000 copies/mL and normal ALT at 24 weeks post-treatment according to the decline from baseline in serum HBsAg and HBV DNA levels at week 12.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Peginterferon alfa-2a is an established therapy for HBeAg-negative CHB. The current study demonstrated that 52% and 31% of patients achieved serum HBV DNA levels <10 000 and <312 copies/mL, respectively, at 24 weeks after 48 weeks of therapy, which was superior to the initial registration trial in which 43% and 19% of patients achieved serum HBV DNA levels <20 000 and <400 copies/mL, respectively, at 24 weeks after 48 weeks of therapy.[8] Sixty-one per cent of patients enrolled in the global registration trial were Asian. Among them, 50% achieved serum HBV DNA levels <10 000 copies/mL, which was comparable to our results.[17] Therefore, our results support the notion that peginterferon is one therapeutic option for HBeAg-negative patients who seek to treat the disease with a finite course of therapy.

According to the REVEAL study, a serum HBV DNA level at 10 000 copies/mL is regarded as a threshold level above which long-term progressive liver diseases, such as cirrhosis and HCC, can occur.[18, 19] Further analysis showed that the risk of liver-related mortality is increased, even in patients with serum HBV DNA levels between 300 and 10 000 copies/mL, when compared with patients with serum HBV DNA levels <300 copies/mL.[20] Therefore, sustained suppression of serum HBV DNA to below detectable levels appears to be the most desirable therapeutic endpoint one should aim for when treating HBeAg-negative patients.

An analysis of the 315 HBeAg-negative patients with long-term follow-up following peginterferon therapy showed that younger age, high baseline ALT and low baseline HBV DNA were significant predictors of a SR (HBV DNA <10 000 copies/mL) at 24 weeks post-treatment.[9, 21] The ORs associated with the predictors were mild. In this study, we identified the baseline HBV DNA as a significant predictor of SR and the baseline HBV DNA and HBsAg as significant predictors of achieving HBV DNA <10 000 copies/mL at 24 weeks post-treatment. Furthermore, we identified the on-treatment serum HBsAg and HBV DNA levels as even better predictors of achieving SR or HBV DNA <10 000 copies/mL at 24 weeks, after 48 weeks of peginterferon therapy in HBeAg-negative patients infected with genotype B or C. It would be desirable if a baseline predictor such as HBsAg or HBV DNA can be used to identify potential responders before treatment is initiated. However, the predictive value of serum HBsAg only became moderate early during treatment (AUC = 0.50 at baseline vs. 0.75 at week 12). Early on-treatment HBsAg level is a more clinically useful predictor of treatment response.

Recently, serum HBsAg levels were shown to positively correlate with the transcriptional activity of the intrahepatic HBV covalently closed circular DNA (cccDNA) and serum HBV DNA levels in HBeAg-positive, but not in HBeAg-negative CHB patients.[22, 23] Synthesis of HBsAg from integrated HBV genomic sequences and/or dissociation of HBsAg transcription/secretion from viral replication are proposed as possible causes of this dissociation in HBeAg-negative patients.[22, 23] Intrahepatic cccDNA levels declined in parallel with the reduction of serum HBsAg levels during antiviral therapy and intrahepatic cccDNA levels at the end of antiviral therapy were shown to predict a SR after the cessation of therapy.[24, 25] Despite the imperfect correlation between serum HBsAg and intrahepatic cccDNA levels in HBeAg-negative CHB, it was shown that the end-of-treatment HBsAg level correlated strongly with a SR (HBV DNA <400 copies/mL) at 6 months post-treatment in HBeAg-negative patients undergoing peginterferon therapy,[11] which is in agreement with our observation. More importantly, we demonstrated that the early on-treatment serum HBsAg level serves as a better predictor for SR in HBeAg-negative patients undergoing peginterferon therapy.

Moucari et al.[12] showed that a decline of 0.5 and 1.0 log10 IU/mL in serum HBsAg levels at weeks 12 and 24 of therapy, respectively, had high predictive values for SR (NPV [90%], PPV [89%] for week 12; and NPV [97%], PPV [92%] for week 24) in a cohort of 48 HBeAg-negative patients (56% for genotype A or D). They concluded that early on-treatment serum HBsAg levels can be used as a useful marker for predicting a sustained off-treatment response. However, Rijckborst et al.[13] showed that the on-treatment decline in HBsAg alone was of limited value in the prediction of achieving HBV DNA <10 000 copies/mL and normal ALT at 24 weeks post-treatment, but the combination of a decline in serum HBsAg and HBV DNA at week 12 gave a better prediction in 102 HBeAg-negative patients (93% for genotype A or D). An absence of a decline in serum HBsAg and HBV DNA (<2 log10 copies/mL) at week 12 of therapy provided a strong discontinuation criterion for HBeAg-negative patients. In this study consisting of 61 HBeAg-negative patients infected with genotype B or C, we found that the absolute value of the on-treatment serum HBsAg level was a better predictor of SR, achieving HBV DNA <10 000 copies/mL or achieving HBV DNA <10 000 copies/mL and normal ALT at 24 weeks post-treatment than the magnitude of the decline in HBsAg, suggesting that the remaining cccDNA level is the more critical determinant of a sustained response. The serum HBsAg level at week 12 of therapy had a moderate predictive value for SR, but a lower predictive value for achieving HBV DNA <10 000 copies/mL or for achieving HBV DNA <10 000 copies/mL and normal ALT at 24 weeks post-treatment. Combining the serum HBsAg and HBV DNA levels at week 12 of therapy yielded a limited further increase in predictive accuracy. Furthermore, the algorithm proposed by Rijckborst et al.[13] did not perform well in our patient cohort. Therefore, although the general concept of monitoring the serum HBsAg levels during peginterferon therapy is useful, the predictive rule and value may vary depending on the patient population studied.

One major limitation of the current study was its relative small sample size, which might create the possibility of model overfitting. Further studies enrolling a larger cohort of patients from the Asia-Pacific region are warranted to validate our findings before this predictive rule can be applied universally to patients infected with genotype B or C.

There are advantages associated with the early identification of potential sustained responders to peginterferon therapy. First, it allows us to persuade candidate patients to remain on therapy. Secondly, it guides us to terminate therapy in patients with a poor early response, particularly in those who experienced troublesome side effects associated with interferon therapy. Thirdly, it provides a strategy whereby we can use serum HBsAg levels to stratify patients for further clinical studies investigating the potential roles of combination therapy with oral antiviral agents to improve therapeutic outcomes. Therefore, our study provides evidence supporting the routine use of quantitative serum HBsAg measurement in therapeutic monitoring of HBeAg-negative patients infected with genotype B or C who undergo peginterferon therapy. Although we only demonstrated the predictive value of early serum HBsAg measurement for SR 24 weeks post-treatment, all the 19 sustained responders had serum HBV DNA <312 copies/mL at 48 weeks post-treatment. Therefore, the predictive value for a SR derived from our study can be extended to 1 year post-treatment. Nonetheless, we need a larger patient cohort and a longer follow-up to assess the predictive value for a long-term SR or HBsAg seroclearance.

In conclusion, peginterferon alfa-2a therapy showed efficacy for HBeAg-negative CHB with 52% and 31% of patients achieving serum HBV DNA levels <10 000 and <312 copies/mL, respectively, at 24 weeks after 48 weeks of therapy. Serum HBsAg levels at 12, 24 and 48 weeks of therapy and HBV DNA levels at baseline, 12 and 24 weeks of therapy were independent predictors of SR at 24 weeks post-treatment. The serum HBsAg levels were significantly different during treatment between patients with SR and relapse. A quantitative serum HBsAg level at week 12 can be used for early prediction of a sustained off-treatment response to peginterferon therapy to optimise management in HBeAg-negative CHB patients infected with genotype B or C.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Declaration of personal interests: Cheng-Yuan Peng has acted as a consultant for Roche, Bristol-Myers Squibb and Novartis. Hsueh-Chou Lai, Yu-Fen Li, Wen-Pang Su, Po-Heng Chuang and Jung-Ta Kao have no conflict of interest to declare. Declaration of funding interests: This study was supported by grants DMR-96-105 and DMR-98-011 from the China Medical University Hospital in Taichung, Taiwan.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information
FilenameFormatSizeDescription
apt4973-sup-0001-TableS1.docWord document64KTable S1. Patient characteristics by hepatitis B virus genotype.
apt4973-sup-0002-TableS2.docWord document60KTable S2. Univariate analysis of baseline and on-treatment factors associated with achieving HBV DNA <10 000 copies/mL at 24 weeks post-treatment.
apt4973-sup-0003-TableS3.docWord document57KTable S3. Multivariate analysis of baseline and on-treatment factors associated with achieving HBV DNA <10 000 copies/mL at 24 weeks post-treatment.
apt4973-sup-0004-TableS4.docWord document38KTable S4. Predictive values of on-treatment HBV DNA and HBsAg decline levels for sustained response.
apt4973-sup-0005-TableS5.docWord document39KTable S5. Predictive values of baseline and on-treatment HBV DNA and HBsAg levels for achieving HBV DNA <10 000 copies/mL at 24 weeks post-treatment.
apt4973-sup-0006-TableS6.docWord document35KTable S6. Predictive values of on-treatment HBV DNA and HBsAg decline levels for achieving HBV DNA <10 000 copies/mL at 24 weeks post-treatment.
apt4973-sup-0007-TableS7.docWord document41KTable S7. AUCs for different definitions of therapeutic endpoints.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.