- 1Division of Hepatogastroenterology, Department of Internal Medicine, Kaohsiung.
- 2Chiyai and Linkou Chang Gung Memorial Hospital, Chang Gung University College of Medicine, Taiwan.
- 3Linkou Chang Gung Memorial Hospital, Chang Gung University College of Medicine, Taiwan. Electronic address: Rachel.firstname.lastname@example.org.
- 4Division of Hepatogastroenterology, Department of Internal Medicine, Kaohsiung,. Electronic address: email@example.com.
Background & aims: We investigated the incidence and factors associated with relapse of hepatitis B virus (HBV) infection in patients with levels of HB surface antigen (HBsAg) below 100 IU/mL after cessation of entecavir or tenofovir disoproxil fumarate (TDF) treatment.
Methods: We collected data from patients with chronic HBV infection without cirrhosis treated with entecavir from 2007 and 2015 or TDF from 2011 through 2016 in Taiwan. We identified 135 patients with levels of HBsAg below 100 IU/mL at the end of treatment (79 entecavir and 56 TDF) and collected data from them for a median 87 weeks (interquartile range, 48-161 weeks) for use as the development set. We collected data from 108 patients from separate medical centers in Taiwan, followed for a median 126 weeks (interquartile range, 61-214) weeks, and used these as the validation group. Post-treatment virologic relapse was defined as a serum level of HBV DNA above 2000 IU/mL and clinical relapse was defined as level of alanine aminotransferase (ALT) greater than 40 U/L and level of HBV DNA above 2000 IU/mL.
Results: In the development group, the 5-year incidences of virologic relapse, clinical relapse, and HBsAg loss were 40.9%, 32.5%, and 47%, respectively. Baseline level of HBV DNA and end of treatment level of HBsAg were independently associated with relapse. In the development group, 17.3% of patients with end of treatment levels of HBsAg below 40 IU/mL had a virologic relapse within 5 y whereas 67.6% of patients with a level of HBsAg of 40 IU/mL or more had a virologic relapse within 5 y (P<.001); proportions of patients with clinical relapses were 10.2% (HBsAg below 40 IU/mL) and 57.6% (HBsAg ≥40 IU/mL; P<.001). In the validation groups, for patients with end of treatment HBsAg levels below 40 IU/mL or ≥40 IU/mL, the rates of virologic relapse 5 y were 31.1% and 80.5% (P<.001); rates of clinical relapse were 14.2% and 50.3% (P<.001), respectively. Rates of virologic and clinical relapse within 5 y were low (below 10%) in patients with a combination of end of treatment level of HBsAg below 40 IU/mL and baseline HBV DNA level below 5×104 IU/mL, or baseline hepatitis B core-related antigen level below 4 log U/mL in the development group.
Conclusions: An end of treatment level of HBsAg of 40 IU/mL or lower is optimal for stopping nucleos(t)ide analog therapy. Waiting to stop therapy until patients have a combination of baseline levels of HBV DNA of 5×104 IU/mL or HBcrAg of 4 log U/mL and end of treatment level of HBsAg of 40 IU/mL might reduce the risk of HBV relapse.