Serum HBV RNA is a Potential Predictor of Hepatitis B Surface Antigen Reversion

Abstract

Potential conflict of interest: Nothing to report. Because sustained hepatitis B surface antigen (HBsAg) seroclearance with or without anti‐HB appearance represents a functional cure of chronic hepatitis B (CHB), HBsAg loss has been recommended as an optimal endpoint of antiviral treatment.1 However, composite reversion including HBsAg reversion and/or hepatitis B virus (HBV) DNA reappearance may still occur after drug withdrawal, even after achieving HBsAg loss.2 Therefore, a novel marker predicting the risk of composite reversion is needed. Because serum HBV RNA has been suggested as a potential predictor of nucleos(t)ide analogue (NA) off‐therapy viral rebound in our previous study,3 we wondered whether it could be used to predict the risk of composite reversion. Methods and Results In this study, 32 CHB patients who achieved seroclearance of both HBsAg and HBV DNA after receiving pegylated interferon (Peg‐IFN) treatment alone or in combination with NA therapy were enrolled, including 12 patients with composite reversion and 20 patients without. Posttherapy monitoring interval was 12‐24 weeks. The levels of HBV serum markers and liver function examination were tested during the monitoring process. Among the 12 patients with composite reversion, 2 patients occurred at 4 weeks following drug withdrawal, 8 patients at 24‐48 weeks, and the last 2 patients at more than 48 weeks. The clinical characteristics of the patients are given in Table 1 and the Supporting Information. Serum HBV RNA levels at the end of treatment were compared between the composite reversion and the noncomposite reversion groups. As indicated in Table 1, 7 of 12 patients were positive for serum HBV RNA at the end of treatment in the composite reversion group, and 0 of 20 patients were positive in the noncomposite reversion group. In other words, all 7 patients who were positive for HBV RNA at the end of treatment had composite reversion occur after drug withdrawal, whereas composite reversion occurred in only 5 of the 25 patients who were negative for HBV RNA (Table 1). The positive predictive value and negative predictive value of serum HBV RNA levels at the end of treatment for the prediction of composite reversion was 100% and 80%, respectively. This indicates that being serum HBV RNA–positive at the end of treatment could predict the risk of composite reversion after drug withdrawal. Table 1 - Demographic and Clinical Characteristics of the Enrolled Patients at the End of Treatment Characteristics Composite Reversion (n = 12) No Composite Reversion (n = 20) P Value* Age, median (years) 37 (26‐56) 37 (21‐61) 1.000 Male sex, n (%) 8 (66.67) 13 (65) 1.000 HBV DNA < 20 IU/mL, n (%) 12 (100) 20 (100) 1.000 HBsAg < 0.05 IU/mL, n (%) 12 (100) 20 (100) 1.000 Anti‐HBs level (IU/mL), median (range) 403.4 (38.53‐1000) 478.15 (2‐1000) 0.709 Anti‐HBs level < 100 IU/L, n (%)† 3 (25) 2 (10) 0.338 HBeAg < 1.0 (S/CO), n (%) 12 (100) 20 (100) 1.000 Anti‐HBe‐positive (S/CO), n (%)‡ 10 (83.33) 20 (100) 0.133 ALT < 40 IU/mL, n (%) 10 (83.33) 15 (75) 0.683 AST < 40 IU/mL, n (%) 10 (83.33) 15 (75) 0.683 Antiviral drugs: Peg‐IFN combined with NAs, n (%) Peg‐IFN alone, n (%) 9 (75) 3 (25) 13 (65) 7 (35) 0.703 0.703 Duration of HBsAg loss before drug withdrawal (weeks),§ median (range) 42.5 (24‐84) 43 (0‐100) 0.585 Follow‐up time after drug withdrawal (weeks),§ median (range) 48 (20‐92) 171 (48‐273) < 0.001 HBV RNA positive, n (%) 7 (58.33) 0 (0) < 0.001 Abbreviations: ALT, alanine aminotransferase; AST, aspartate aminotransferase; HBeAg, hepatitis B e antigen; and S/CO, signal‐to‐cutoff ratio.*Mann‐Whitney U test or Fisher’s exact test.†Because it has been reported that anti‐HBs ≥ 100 IU/mL is important to prevent relapse,4 the number of patients with anti‐HBs ≥ 100 IU/mL or < 100 IU/mL in both groups was also analyzed.‡Positive anti‐HBe is defined as anti‐HBe < 1.0 S/CO according to the manufacturer’s instructions.§Drug withdrawal in each case was decided by the attending physician with use of similar but nonuniform criteria. Next, the association between serum HBV RNA and the type of HBV composite reversion was analyzed in the composite reversion group. Among the 12 composite reversion patients, 6 patients showed HBsAg reversion, 3 showed HBV DNA reappearance, and 3 showed both HBsAg reversion and HBV DNA reappearance. As given in Table 2 and Supporting Table S1, all 6 patients who experienced HBsAg reversion were serum HBV RNA–positive at the end of treatment, whereas only 1 of 3 patients with both HBsAg reversion and HBV DNA reappearance, and none of the 3 patients with HBV DNA reappearance, were serum HBV RNA–positive. This result indicates that being serum HBV RNA–positive at the end of treatment could predict a high risk of HBsAg reversion after drug withdrawal. Moreover, HBV S gene mutation was analyzed in a patient who had an HBV DNA level (7,020 IU/mL) high enough for the polymerase chain reaction (PCR)‐based sequence analysis. The result reveals that multiple point mutations, including I126T/S, T140I, and D144A in the S gene, were found in this patient. Table 2 - Association Between Serum HBV RNA and the Type of HBV Composite Reversion After Drug Withdrawal Group HBsAg Reversion HBsAg Reversion and HBV DNA Reappearance HBV DNA Reappearance Total (n) P Value* HBV RNA‐positive† 6 1 0 7 HBV RNA‐undetectable 0 2 3 5 0.006 Total (n) 6 3 3 12 *Fisher’s exact test.†Positive HBV RNA is defined as the presence of positive amplification signal (appeared amplification curve). Discussion It has been reported that longer durations of HBsAg loss before NA withdrawal may result in more durable HBsAg seroclearance.5 However, perhaps because of the small cohort in this study, the association between the duration of HBsAg loss and the risk of HBsAg reversion was not observed. In this study, all 7 CHB patients with positive HBV RNA at the end of treatment developed composite reversion after drug withdrawal, indicating that the intrahepatic covalently closed circular DNA (cccDNA) in these patients were present and still transcriptionally active. In other words, at least in some CHB patients who appeared to achieve “functional cure,” the template of HBV replication–cccDNA may still be transcriptionally active. Based on previous reports, negative serum HBsAg might be caused by the S gene mutation–mediated poor detectability of HBsAg rather than the real HBsAg seroclearance.6 In this study, the S gene mutation was found in a patient with HBV DNA reappearance and HBV DNA level high enough for the PCR‐based sequence analysis. Furthermore, the positive serum HBV RNA with negative HBV DNA indicated that intrahepatic cccDNA might be in a transcriptional active state, but with the decreased reverse transcriptional activity of HBV DNA polymerase (p‐protein) mediated by either NAs treatment3 and/or the P gene mutation.7 Accordingly, for the diagnosis of occult HBV infection (OBI), we suggest that the persistence of HBV RNA should also be taken into consideration. Because serum HBV RNA as a virological marker showed certain superiority in reflecting cccDNA activity, especially among those patients under NA treatment, the OBI should be redefined as the persistence of HBV DNA and/or HBV RNA (HBV nucleic acid positive) in the serum or liver of individuals with undetectable HBsAg. Moreover, because HBV DNA was reverse transcribed from pregenomic RNA (pgRNA) by p‐protein, and the reverse transcriptional activity of p‐protein might decrease for those patients with positive serum HBV RNA at the end of treatment, it might lead to the phenomenon that patients with HBV DNA reappearance have less HBV RNA positivity at the end of treatment. In summary, the current study suggested that serum HBV RNA might be a potential viral marker in predicting composite reversion, especially HBsAg reversion, after antiviral drug withdrawal. Like serum HBV RNA, serum hepatitis B core‐related antigen, another viral marker associated with intrahepatic cccDNA, may complement HBV RNA in predicting composite reversion after drug withdrawal. However, multicenter, large‐scale, and prospective studies should be conducted to validate these findings and presumptions in the future. Author contributions L.F., C.X., and W.J. designed the research; W.J., C.X., W.Y., C.Z., W.L., and H.H. performed the research; all authors analyzed the data; W.J. wrote the paper; and L.F. and C.X. revised the paper. Author names in bold designate shared co‐first authorship.