Long-term outcomes after nucleos(t)ide analogue cessation in chronic hepatitis B – follow-up from the RETRACT-B cohort

Abstract Background and aims Stopping nucleos(t)ide analogues (NA) in chronic hepatitis B (CHB) patients can increase HBsAg loss rates. However, long-term follow-up (FU) data after NA cessation among a multiethnic population are lacking. This study aimed to assess the long-term outcomes after NA cessation in a large global cohort. Methods Extended FU data were collected from patients enrolled in the RETRACT-B cohort. New patients were added if meeting the inclusion criteria (virally suppressed, HBeAg negative and HBsAg positive at end of therapy [EOT]). The primary outcome was the 10-year off-treatment HBsAg loss rate. Secondary outcomes were retreatment and adverse events. Results In total, 2,029 patients were included. At EOT, mean age was 52 years, 81% Asian, 15% White, and 4% Black/other, mean HBsAg at EOT was 2.7 log10 IU/mL (<100/100-1,000/≥1,000 IU/mL: 16/46/38%). Median FU was 65 [range 35-99] months. The 10-year off-treatment HBsAg loss rate was 17% (<100/100-1,000/≥1,000: 52/13/6%, p<0.001) and 59% was retreated (<100/100-1,000/≥1,000: 28/62/71%). HBsAg seroreversion was observed in 9 out of 171 (5.3%) patients, of whom 6 re-lost HBsAg. The 10-year risk of flares, HCC and decompensation were 30%, 3.9%, and 3.3%. A history of cirrhosis was associated with a higher risk of decompensation (10yr: 5.7% vs. 3.0%) and HCC (10 yr: 12% vs. 2.4%), both p<0.001. Conclusion At 10 years, 17% of patients achieved off-treatment HBsAg loss, whereas more than half has been retreated. Patients with HBsAg levels <100 IU/mL benefit the most from NA cessation with a HBsAg loss rate of 52%. Based on our data, finite NA therapy is not recommended if HBsAg ≥1,000 IU/mL or in cirrhotic patients. Impact and implications Based on our study, we can conclude that finite NA therapy can be considered in patients with HBsAg levels <100 IU/mL when aiming for HBsAg loss, since they have the highest likelihood of achieving this endpoint (52%) and lowest need for retreatment (28%) during long-term follow-up, but only when strict monitoring can be guaranteed. In contrast, finite therapy is not recommended in patients with HBsAg levels ≥1,000 IU/mL or those with a history cirrhosis, especially as cirrhosis is a known risk factor for hepatic decompensation and HCC. In addition, our data can be used to further improve patient monitoring after finite NA therapy, especially in those who successfully remain off-therapy for five to ten years. Clinical trial number not applicable.