Background and Aims: Long-term nucleos(t)ide analog (Nuc) treatment can reduce HCC in patients with HBV-related liver cirrhosis (HBV-LC). Earlier small cohort studies showed a comparable 5-year incidence of HCC in HBeAg-negative patients with HBV-LC who stopped and those continued Nuc therapy. This study aimed to validate these findings using a large cohort with 10-year follow-up. Approach and Results: From 2 centers, 494 HBeAg-negative patients with HBV-LC who stopped (finite group) and 593 who continued (continuous group) Nuc therapy were recruited. HCC, HBsAg loss, liver-related mortality/transplantation, and overall survival rates were compared between 2 groups with 1:1 propensity score matching of sex, treatment history, types of Nuc, age, transaminases, platelet count, and HBsAg levels at end of therapy in finite group or 3-year on-therapy in continuous groups. During a median follow-up of 6.2 (3.4–8.9) years, the annual and 10-year HCC incidence were lower in finite group (1.6 vs. 3.3%/y and 10-y 15.7% vs. 26.8%, respectively; log-rank test, p<0.0001). The finite group showed greater HBsAg decline/year (−0.116 vs. −0.095 log₁₀ IU/mL, p=0.0026) and 7.6 times higher 10-year incidence of HBsAg loss (22.7% vs. 3%, p<0.0001). Multivariate Cox regression showed finite therapy an independent factor for HBsAg loss (adjusted HR: 11.79) but protective against HCC (adjusted HR: 0.593), liver-related mortality/transplantation (adjusted HR: 0.312), and overall mortality (adjusted HR: 0.382). Conclusions: Finite Nuc therapy in HBeAg-negative HBV-LC may reduce HCC incidence, increase HBsAg loss, and improve survival. Greater HBsAg decline/loss may reflect enhanced immunity and contribute to the reduction of hepatic carcinogenesis.

Off-therapy hepatitis flare may be detrimental or, conversely, facilitate hepatitis B surface antigen (HBsAg) decline. Retreatment decisions are crucial. HBsAg was quantified before and during flares, at peak/retreatment start and at Months 6 and 12 in 336 entecavir/tenofovir-retreated and 105 non-retreated hepatitis B e antigen (HBeAg)-negative patients. Increasing HBsAg during ALT flare defined a ‘virus-dominating flare’ and decreasing HBsAg a ‘host-dominating flare’. Two hundred and eighty-eight retreated patients with a virus-dominating flare showed greater 1 year HBsAg decline (−1.0 versus −0.01 log₁₀ IU/mL; P < 0.0001), more frequent rapid decline (69.8% versus 8.3%; P < 0001) and higher 3 year incidence of HBsAg < 100 IU/mL (32% versus 12%; P = 0.026) than 48 patients with a host-dominating flare, of whom 16 (33.3%) showed 3.8-fold (2 to 52-fold) HBsAg rebound on retreatment (versus 2/288; P < 0.0001). Compared with non-retreated controls, 1 year HBsAg decline was greater (−1.0 versus −0.47 log₁₀ IU/mL; P < 0.0001) and faster (69.8% versus 42.5%; P < 0.0001) in patients with a virus-dominating flare, whereas 1 year HBsAg decline (−0.01 versus −0.16 log₁₀ IU/mL) and 3 year HBsAg loss rate (0% versus 21%; P = 0.009) were lower in patients with a host-dominating flare. Entecavir/tenofovir retreatment effectively decreases HBsAg level in patients with a virus-dominating flare but is ineffective/worse in patients with a host-dominating flare. These results support the use of combined HBsAg/ALT kinetics for the decision to retreat patients with a virus-dominating flare and withhold retreatment for patients with a host-dominating flare.