Efficacy and Safety of Peginterferon Alfa-2a (40KD) in Children With Chronic Hepatitis B: The PEG-B-ACTIVE Study

STUDY DESIGN

PEG-B-ACTIVE (YV25718, NCT01519960) was a randomized, controlled, open-label, multicenter, phase III study conducted at 37 sites in China, Russian Federation, Ukraine, Australia, Israel, Belgium, Poland, United Kingdom, Bulgaria, Germany, United States, and Italy.

Patients without advanced fibrosis (AF; METAVIR F0-F2 or equivalent) were randomized (2:1) to 48 weeks of PegIFN alfa-2a (Group A) or an untreated control group (Group B; Supporting Fig. S1). Patients with AF, but not cirrhosis (METAVIR F3 or equivalent), were assigned to 48 weeks of PegIFN alfa-2a (Group C), because it was considered clinically and ethically unacceptable to delay treatment in patients at risk of disease progression. Randomization was stratified by HBV genotype (A vs. non-A) and ALT (<5 × upper limit of normal [ULN] vs. ≥5 × ULN) and performed centrally by Interactive Voice/Web Response System (IxRS).

Group B patients who completed the 48-week observation period without HBeAg seroconversion could switch, at investigator and patient discretion, at any time up to 1 year after the observation period, to receive 48 weeks of PegIFN alfa-2a. This accommodated regional differences in clinical practice: in some regions, pediatricians adopt adult treatment strategies for children in the immune-active phase to avoid treatment delays, whereas in other regions a watch-and-wait approach is preferred. All patients entered long-term follow-up, lasting up to 5 years after the end of treatment/principal observation period.

The study was conducted in accord with the principles of the Declaration of Helsinki and Good Clinical Practice, and the study protocol was approved by the relevant health authority and ethical committee at each site. Written informed consent was obtained from the parent/legal guardian of all patients, from the patient if legally considered an adult by national legislation, and assent from the patient, where appropriate. Patient safety data were monitored by an independent data safety and monitoring board on an ongoing basis throughout the study conduct.

PATIENTS

Patients aged 3 to <18 years were eligible if they were HBeAg-positive (≥6 months), had detectable HBV DNA (>2,000 IU/mL), and elevated ALT (>1 × but ≤10 × ULN) at screening. Patients must have had a liver biopsy within 2 years (9 months if they had AF). Liver biopsy was performed and assessed at participating sites, within 6 months of enrollment in 89% of patients (median 18 days). Patients were ineligible if they had cirrhosis (METAVIR F4 or equivalent), tested positive for hepatitis A, C, or D or for HIV, or had a history or evidence of chronic liver disease other than CHB, or suspicion of HCC. Patients with a history of psychiatric disorder, significant chronic pulmonary or cardiac disease, poorly controlled thyroid disease or diabetes, previous solid organ or stem-cell transplant, or evidence or history of malignancy were also excluded.

Patients in Groups A and C received PegIFN alfa-2a subcutaneously once-weekly for 48 weeks, with dosing based on body surface area (BSA) categories previously approved for use in pediatric CHC patients in the EU: 0.51-0.53 m², 45 μg; 0.54-0.74 m², 65 μg; 0.75-1.08 m², 90 μg; 1.09-1.51 m², 135 μg; and >1.51 m², 180 μg. This dosing regimen is based upon the formula of (dose = 180 × BSA/1.73 m²), where a BSA of 1.73 m² is considered the average in adults, and is designed to achieve equivalent PegIFN alfa-2a exposures in children compared with adults while minimizing potential dosing errors. Dose adjustments during the study were allowed for adverse event (AEs) or changes in BSA.

EFFICACY ASSESSMENTS

The primary efficacy endpoint was proportion of patients with HBeAg seroconversion (loss of HBeAg and presence of antibody to hepatitis B e antigen [anti-HBe] antibody) 24 weeks after the end of treatment/observation in Groups A and B. Secondary efficacy endpoints included the proportion of patients with HBeAg clearance, HBsAg clearance, HBsAg seroconversion, HBV DNA <20,000 IU/mL, HBV DNA <2,000 IU/mL, undetectable HBV DNA, and normalized ALT 24 weeks after the end of treatment/observation and at other time points (e.g., end of treatment). Changes in virological and serological biomarkers over time were also assessed as secondary endpoints.

Serum levels of HBeAg, HBsAg, anti-HBe, anti-HBs (by enzyme-linked immunosorbent assay [Siemens Centaur, Siemens Healthcare Diagnostics, Camberley, UK]) and HBV DNA (by real-time quantitative PCR [qPCR]; Roche COBAS TaqMan [Roche Diagnostics, Welwyn, UK] 48; lower limit of detection, 29 IU/mL), were measured centrally at baseline and every 12 weeks thereafter.

SAFETY ASSESSMENTS

Safety evaluations included incidence of AEs, laboratory assessments, growth parameters, and vital signs. Severity of AEs was assessed (mild, moderate, or severe) and causality determined by the investigator. Dose reductions could be made for moderate-severe AEs, and specific dose-adjustment guidance was provided for management of treatment-emergent neutropenia, thrombocytopenia, or elevated ALT (Supporting Table S1).

Safety was assessed at baseline, weeks 1, 2, 4, 8, and 12, and every 6 weeks thereafter during treatment, and as appropriate during follow-up. Standard hematology and blood chemistry were analyzed locally or centrally, while urinalysis was conducted locally. Sex-specific Centers for Disease Control and Prevention Growth Charts were used to calculate height/weight for age percentiles and deviation from the mean (Z-scores).27

SUBSTUDIES

Patients at eight centers could participate in an optional liver elasticity substudy, where transient elastography (TE; FibroScan) assessments were performed at baseline, end of treatment, and week 24 and year 2 of follow-up. Correlation between liver elasticity/liver stiffness (LS) and liver fibrosis (LF) at baseline, and longitudinal changes in LS were assessed.

Patients who received PegIFN alfa-2a (including Group B patients who switched) could participate in an optional pharmacokinetic (PK) substudy. PK blood samples were collected at weeks 1 and 24 (predose and 24-48, 72-96, and 168 hours postdose) and at weeks 4, 8, and 12 (predose). PegIFN alfa-2a concentrations were analyzed using a population PK approach (see Supporting Material).

STATISTICAL ANALYSIS

The primary endpoint was assessed using the Cochran-Mantel-Haenszel test, adjusted for stratification factors of genotype (A vs. non-A) and ALT (<5 × ULN vs. ≥5 × ULN). The Breslow-Day test was used to assess homogeneity of odds ratios (ORs) across strata, and sensitivity analyses were performed using unstratified Pearson’s chi-square and Fisher’s exact tests. Assuming a 32% HBeAg seroconversion rate for Group A and 10% for Group B, a sample size of 145 patients was required to provide ≥80% power at the 0.05 level of significance with a two-sided test.

Descriptive statistics of secondary endpoints were summarized by treatment group. For binary endpoints, P values were calculated by unstratified Fisher’s exact test. Continuous endpoints were summarized descriptively with 95% confidence intervals (CIs). No interim analyses were performed.

Response rates were calculated for all patients who received ≥1 dose of study drug. Patients with missing values were classified as nonresponders, including Group B patients who switched to PegIFN alfa-2a before the primary endpoint at 24 weeks posttreatment. Efficacy and safety endpoints are reported up to 24 weeks’ follow-up. Long-term follow-up and data from Group B patients who switched will be reported separately, once available.

PATIENTS

Of the 162 enrolled patients (June 2011-September 2014), 151 were randomized to Group A (n = 101) and Group B (n = 50), and 10 patients with AF were allocated to Group C. An additional patient, initially allocated to Group C but subsequently found not to have AF, withdrew without receiving any study drug.

Baseline characteristics were well balanced between Groups A and B (Table 1). Most patients (64%) were male, 56% were Asian, and mean age was 10.7 years (range, 3-17) with approximately 15% <5 years. Genotype C (34%) and D (32%) infections were more common, and 13 (9%) patients had normal ALT at baseline (despite having elevated levels at screening).

Most patients in Group A (98%; 99 of 101) completed 48-week treatment and 24-week follow-up; 2 patients did not complete treatment (Supporting Fig. S2). In Group B, 12% of patients withdrew and 58% switched to PegIFN alfa-2a before 24-week follow-up.

EFFICACY

After 24 weeks of follow-up, the percentage of patients with HBeAg seroconversion was significantly higher in Group A (25.7%; 26 of 101) versus Group B (6% [3 of 50]; OR, 5.43; 95% CI, 1.54-9.20; P = 0.0043; Fig. 1). The Breslow-Day test (P = 0.3732) demonstrated homogenous ORs across strata, and the stratified result was supported by the unstratified Pearson’s chi-square (P = 0.0038) and Fisher’s exact (P = 0.0038) tests.

A higher proportion of patients in Group A versus Group B achieved secondary efficacy endpoints, including HBsAg clearance (8.9% vs. 0%; P = 0.03), HBV DNA <2,000 IU/mL (28.7% vs. 2%; P < 0.001), and ALT normalization (51.5% vs. 12%; P < 0.001; Table 2). Levels of HBV DNA, HBeAg, and HBsAg decreased during treatment in Group A while remaining stable in Group B (Fig. 2).

Subgroup analyses, although limited by small patient numbers, suggested higher treatment-induced HBeAg seroconversion in younger patients, those with genotypes B and C, and patients with lower HBV DNA, HBeAg, and HBsAg at baseline (Fig. 3). Multiple logistic regression analyses identified PegIFN alfa-2a therapy, lower baseline HBeAg levels, and HBV-DNA levels as significant predictors of HBeAg seroconversion (Supporting Table S2). Notably, baseline age and HBV genotype (A vs. non-A) were not identified as significant predictors. Of the 9 patients who achieved HBsAg loss, 6 (67%) were <12 years, whereas 4 had genotype B infection, 4 genotype C, and 1 genotype A.

Because 69% of Group B patients (35/51) were imputed as nonresponders attributed to withdrawal (n = 6) or switching to PegIFN alfa-2a (n = 29), additional sensitivity analyses were conducted for the primary endpoint. Using a prespecified last-observation-carried-forward approach, the difference between groups remained significant (27 of 101 [26.7%] vs. 4 of 50 [8.0%]; P = 0.008). Similarly, extrapolating the Group B HBeAg seroconversion rate, from 6% at the end of the principal observation period to 9% at follow-up week 24, still resulted in a considerably lower rate than that observed in Group A. Furthermore, in a post-hoc analysis of Group B data at follow-up week 24 regardless of whether patients switched to PegIFN alfa-2a, the difference between groups remained significant (26 of 101 [25.7%] vs. 4 of 50 [8.0%]; P = 0.011).

The proportion of patients with HBeAg seroconversion during PegIFN alfa-2a treatment was 8% compared with the 6% in untreated patients. However, the HBeAg seroconversion rate increased to 26% during the 24 weeks posttreatment in Group A while remaining steady in Group B. Virological response (VR; HBV DNA <2,000 IU/mL) increased to 31% during treatment and remained steady during the 24 weeks posttreatment (Fig. 4). Quantitative levels of HBV DNA, HBsAg, and HBeAg declined during treatment, and the decline was significantly more pronounced (as early as week 12) in patients who achieved HBeAg seroconversion (Supporting Fig. S3). These findings are similar to adult data and suggest a potential for response-guided therapy.

TE

Seventy-five patients from Groups A (n = 44), B (n = 25), and C (n = 6) participated in the liver elasticity substudy (Supporting Table S3). There was a modest correlation between LS and LF at baseline, and an accurate cutoff for significant fibrosis (SF) was not identified (Supporting Table S4; Supporting Fig. S4). The cutoff recommended by the European Association for the Study of the Liver and Asociación Latinoamericana para el Estudio del Hígado guidelines for excluding SF (<6.0 kilopascals [kPa])28 showed poor diagnostic performance (sensitivity, 0.65; specificity, 0.67; positive predictive value, 0.55; negative predictive value, 0.76) and wrongly excluded SF in 24% (9 of 37) of patients.

PegIFN alfa-2a treatment was associated with a modest decrease in LS. Mean LS scores from baseline to the end of treatment/principal observation period decreased by 0.49 and 1.52 kPa in Groups A and C, respectively, with an increase of 0.38 kPa in Group B (Supporting Fig. S4; Supporting Fig. S5). Within Group A, the decline in mean LS scores during treatment was similar between patients with/without HBeAg seroconversion at 24 weeks posttreatment (0.55 vs. 0.46 kPa, respectively).

To assess the possible confounding effect of inflammation on LS measurements, a linear regression analysis was performed (Supporting Fig. S6). In Group A, changes in LS scores from baseline to 24 weeks posttreatment were not associated with changes in ALT (correlation coefficient, –0.1392). Further details are provided in the Supporting Material and have been presented elsewhere.29

PK

Thirty patients from Groups A (n = 16), B (n = 9; after switching to PegIFN alfa-2a), and C (n = 5) participated in the PK substudy (Supporting Table S5). PK data were best described by a one-compartment disposition population model with a sequential zero–first-order absorption (Supporting Fig. S7). Estimated mean PegIFN alfa-2a exposures (area under the curve) in children was 3,484 h.ng/mL, which is consistent with exposures estimated in adults (Supporting Figs. S8 and S9). Exposures in pediatric patients were also comparable across BSA dosing categories (Supporting Fig. S9). Five of 21 patients in Groups A and C who participated in the PK substudy achieved HBeAg seroconversion at follow-up week 24. No relationship between response and PegIFN alfa-2a exposure could be identified in this small sample. Further details are provided in the Supporting Material and have been presented elsewhere.30

SAFETY

Consistent with previous reports of IFN alfa therapy, the most common AEs included pyrexia, headache, and abdominal pain (Table 3). Most AEs were mild or moderate and rarely resulted in treatment discontinuation (1%; 1 patient for elevated ALT and aspartate aminotransferase [AST]). No notable difference in safety was observed between younger and older children. Seven serious AEs (SAEs), which resolved without sequelae, were experienced by 6 patients in Group A (progression of hepatitis B, tonsillitis, scalp microsporum infection, latent tuberculosis, elevated AST, elevated ALT, and osteochondrosis). In Group B, 1 patient experienced an SAE (suicidal ideation) and withdrew from the study.

Depression or depressed mood, a potential concern of IFN alfa therapy, was reported infrequently. Twelve patients (12%) in Group A and 1 (2%) in Group B reported a neuropsychiatric disorder during the study. All events in Group A were mild or moderate, with the most common being insomnia (3 patients; 3%). Depression and depressed mood were reported by 1 patient each. In Group B, 1 participant reported suicidal ideation as described above.

Laboratory abnormalities during treatment were common but manageable with dose modifications. Treatment-emergent decreases in neutrophils, leukocytes, hemoglobin, and platelets reversed shortly after cessation of therapy. Neutropenia in Group A was moderate (0.50-0.75 × 10⁹ cells/L) in 16 patients (16%) and severe (<0.5 × 10⁹ cells/mL) in 3 patients (3%; none <0.25 × 10⁹ cells/mL). Of these 19 patients, 7 reported an AE of infection, only one of which was severe or serious. This patient, with a lowest neutrophil count (0.75 × 10⁹ cells/L) on study day 119, was diagnosed with severe and serious acute tonsillitis on study day 129. This was unrelated to study treatment, and the dose was not modified. Following penicillin treatment, the event resolved on study day 137. In the remaining 6 patients, the start of the infection AEs did not correlate with the timing of their lowest neutrophil counts, and the AEs were considered unrelated to study treatment, resolving without dose modification.

Incidence of ALT flares (>5 × ULN and 2× baseline) was higher during (37%) versus posttreatment (16%); most were associated with decreasing HBV-DNA levels (>1 log₁₀ IU/mL); none was accompanied by elevated bilirubin (>2 × ULN) nor resulted in hepatic decompensation. Two patients had AEs of increased thyroid-stimulating hormone in Group A. These were mild and had resolved by the end of follow-up. No patient in Group A reported AEs of hypothyroidism or hyperthyroidism or required dose modification for thyroid function abnormalities. In Group B, 1 case of hypothyroidism was reported (the patient did not switch to PegIFN alfa-2a).

Short-term effects on growth appeared minimal. By the end of treatment, mean change in height-for-age Z-scores was –0.10 and proportion of patients with >15% drop in height-for-age percentile was 6% (Table 4; Supporting Fig. S10). Similar changes were observed in Group B untreated patients (–0.01 and 2%, respectively) and also in weight-for-age Z-scores. Patients will continue to be monitored up to 5 years after the end of treatment to assess any long-term effects.

EFFICACY AND SAFETY OUTCOMES IN GROUP C

Although limited by small cohort size, the efficacy and safety outcomes in 10 patients with AF, 4 (40%) of whom were aged <5 years (Table 1), were broadly consistent with Group A. At follow-up week 24, 3 patients (30%) achieved HBeAg seroconversion, 3 (30%) had undetectable HBV DNA, 7 (70%) HBV DNA <2,000 IU/mL, and none achieved HBsAg clearance (Supporting Table S6). LS decreased, with a mean decrease in LS score of 1.52 kPa, by the end of treatment. No SAEs, or AEs leading to study or treatment discontinuation, were reported in Group C.