Reply

We read with interest the article by Sonneveld et al., who reported an association between on-treatment hepatitis B surface antigen (HBsAg) levels and a sustained response to peginterferon alfa-2b in hepatitis B e antigen (HBeAg)–positive patients (n 1⁄4 221). No HBsAg decline by week 12 of therapy was associated with a low chance of a sustained response (97% probability of nonresponse) and was proposed as an early stopping rule for peginterferon therapy. Because this rule needs to be validated in other studies, we investigated how the rule would have performed in HBeAg-positive patients treated with peginterferon alfa-2a during two independent, large-scale studies. HBeAg-positive patients received peginterferon alfa-2a (180 lg/ week) with or without lamivudine (100 mg/day) for 48 weeks as part of a phase 3 study (n 1⁄4 542) or peginterferon alfa-2a (180 lg/week) for 48 weeks as part of the Nephrotic Syndrome Study Network (NEPTUNE) study (n 1⁄4 136). Overall, the rates of HBeAg loss and hepatitis B virus (HBV) DNA levels < 10,000 copies/mL in the phase 3 and NEPTUNE peginterferon alfa-2a studies were similar (25% and 24%, respectively), and they were higher than those in Sonneveld et al.’s analysis (19%). In accordance with Sonneveld et al.’s data, the HBsAg decline was more pronounced in patients with a response 6 months post-treatment versus nonresponders. Patients with no HBsAg decline from the baseline to week 12 had 82% (80/97) and 71% (22/31) probabilities of nonresponse in the phase 3 and NEPTUNE studies, respectively; these were considerably lower than the probability of 97% in Sonneveld et al.’s study (Fig. 1). The probabilities of response in patients with no HBsAg decline were 18% (17/97) and 29% (9/ 31), respectively. Applying the stopping rule would have resulted in premature treatment discontinuation in some patients (17 and 9, respectively) who would have responded. HBeAg seroconversion 6 months post-treatment, rather than HBeAg loss and HBV DNA levels <10,000 copies/mL, was the primary endpoint in the peginterferon alfa-2a studies. Using this more robust indicator of sustained immune control would have resulted in some patients in the phase 3 and NEPTUNE studies (30 and 12, respectively) discontinuing their treatment prematurely if the stopping rule had been applied. Differences in the study populations could explain the varying response rates and the fact that the proposed stopping rule could not be validated by the peginterferon alfa-2a analyses. Sonneveld et al.’s analysis was a European study in which only 20% of the patients were Asian, whereas the populations of the phase 3 and NEPTUNE peginterferon alfa-2a studies were predominantly Asian (>80%). This influenced the genotype distribution; Sonneveld et al.’s study had a high proportion of genotype A or D patients, whereas the peginterferon alfa-2a studies included predominantly genotype B and C patients. In combination with the differences in the treatment regimens (peginterferon alfa-2a versus peginterferon alfa-2b and 48 weeks of therapy versus 52 weeks) and in the numbers of patients included in the analyses, this may account for the differences in the results. Monitoring HBsAg levels during peginterferon therapy provides a good indication of the treatment response and helps in identifying early success. However, it is clear that further analysis is required either to identify an early stopping rule for peginterferon therapy that is valid for all genotypes or to develop genotype-specific algorithms.