Background
Stopping nucleos(t)ide analogue (NA) therapy in patients with chronic hepatitis B (CHB) may trigger a beneficial immune response leading to HBsAg loss, but clinical trials on re-start strategies are lacking.

Aim
To assess whether it is beneficial to undergo a prolonged flare after NA cessation.

Methods
One-hundred-and-twenty-seven patients with HBeAg negative, non-cirrhotic CHB with at least 24 months of viral suppression on NA therapy were included. All study participants stopped antiviral therapy and were randomised to either low-threshold (ALT > 80 U/L and HBV DNA > 2000 IU/mL) or high-threshold (ALT > 100 U/L for >4 months, or ALT > 400 U/L for >2 months) for the re-start of therapy. The primary endpoint was HBsAg loss within 36 months of stopping antiviral treatment. The primary analysis was based on intention-to-treat allocation with last observation carried forward.

Results
There was a numerical but not statistically significant difference in HBsAg loss between the low-threshold (3 of 64; 4.7%) and the high-threshold (8 of 63; 12.7%) group (risk difference: 8.0%, 95% CI: −2.3 to 19.6, p = 0.123). None of the patients with end-of-treatment HBsAg > 1000 IU/mL achieved HBsAg loss; among those with end-of-treatment HBsAg < 1000 IU/mL, 8 of 15 (53.3%) achieved HBsAg loss in the high-threshold group compared to 3 of 26 (11.5%) in the low-threshold group.

Conclusions
We could not confirm our hypothesis that a higher threshold for restart of therapy after NA withdrawal improves the likelihood of HBsAg loss within 36 months in patients with HBeAg negative CHB. Further studies including only patients with HBsAg level <1000 IU/mL and/or larger sample size and longer follow-up duration are recommended.