Letter: No Biochemical Relapse Is Associated With the Highest Off-Therapy HBsAg Loss Rate—Authors' Reply

Abstract

We thank Professors Jeng and Liaw for their interest in our study [1, 2]. We agree that ALT elevation following cessation of nucleos(t)ide analogues (NAs) in patients with chronic hepatitis B (CHB) is not associated with higher HBsAg seroclearance rates. In our study, we found that elevated serum ALT analysed as a time-varying variable was not associated with HBsAg seroclearance regardless of whether the threshold was set at the upper limit of normal (ULN), two-time ULN (clinical relapse) or five-time ULN (acute flares). These findings further examine and validate the results reported in their study [3].

They noted that the 10-year cumulative incidence of HBsAg seroclearance in the current study (12.4%) was lower than that in our previous research and other studies [3, 4]. We acknowledge that the incidence of HBsAg seroclearance could be underestimated in this real-world retrospective study because HBsAg was not regularly measured following a prespecified protocol in routine clinical practice, unlike in a prospective research setting with selected participants. Moreover, the apparent inconsistency might result from differences in analytical approaches. For instance, observation for HBsAg loss was censored on the resumption of antiviral treatment without considerations of competing risks in our prior study [4]. This approach would yield a significantly higher estimate of HBsAg loss rate as shown in the current study (figure S2). Since retreatment is unlikely independent of the probability of HBsAg seroclearance, analysing it as non-informative censoring may lead to an overestimation of HBsAg seroclearance incidence [5]. Regardless, our sensitivity analyses employing different approaches to estimate HBsAg seroclearance incidence consistently confirmed no association between ALT elevation and HBsAg loss.

The current study did not specifically investigate the association between retreatment and subsequent HBsAg seroclearance. Nevertheless, a recent randomised study reported that delaying retreatment by adopting a higher threshold did not increase the chance of HBsAg seroclearance in HBeAg-negative patients who stopped NA therapy after a minimum 24 months of viral suppression. They found no patients with an end-of-treatment HBsAg level greater than 1000 IU/mL would clear HBsAg during follow-up, irrespective of the assignment to higher or lower retreatment thresholds [6]. Only patients with EOT HBsAg level below 1000 IU/mL had a chance of HBsAg seroclearance, suggesting that the likelihood of HBsAg seroclearance was already determined by the HBsAg level at treatment cessation, whether or not antiviral therapy was resumed for clinical relapses.

We fully agree on the importance of rigorous monitoring and timely retreatment to ensure the safety of patients who stop NA therapy [7]. However, acute flare-ups induced by NA withdrawal can progress rapidly, potentially leading to devastating outcomes even in patients meticulously monitored with weekly intervals [8, 9]. Therefore, the practice of finite NA therapy relies on accurately identifying suitable candidates who are likely to achieve maximal benefit with minimal risk. Given that biochemical relapse following NA cessation is not associated with the incidence of HBsAg seroclearance, it is inadvisable to discontinue NA therapy in patients with a high risk of biochemical relapse and a low probability of HBsAg loss, as indicated by established risk predictors [10].

Ying-Nan Tsai: conceptualization, writing – original draft, investigation, visualization, project administration. Yao-Chun Hsu: conceptualization, writing – review and editing, supervision, investigation, visualization, project administration.

Ying-Nan Tsai reported no conflicts of interest. Yao-Chun Hsu has received research grants from Gilead Sciences; lecture fees from Abbvie, Bristol-Myers Squibb, Gilead Sciences, Grifols and Roche; and has served as an advisory committee member for Gilead Sciences and Sysmex.

This article is linked to Tsai et al papers. To view these articles, visit https://doi.org/10.1111/apt.18515 and https://doi.org/10.1111/apt.70061.