Background and Aims:
Among patients with chronic hepatitis B (CHB), the rate of spontaneous hepatitis B surface antigen (HBsAg) loss is low, ~1%. How race affects HBsAg loss is unclear as prior studies have been restricted to homogenous racial cohorts. To investigate the influence of race on rates of HBsAg loss we assessed the incidence and factors associated with HBsAg loss in a multiracial cohort of CHB patients.
Methods:
Data from 3 prospective, longitudinal cohorts of CHB patients were analyzed. The HEP-B-AK cohort included patients followed at the Alaska Native Tribal Health Consortium (ANTHC); The Hepatitis B Research Network (HBRN) cohort included patients followed at 28 centers across the U.S. and Canada; and the NIH cohort included patients followed at the Liver Disease Branch, NIH. Patients ≥18 years with CHB followed for at least 1 year were included. Patients with HIV, HCV and HDV co-infection were excluded. Cohorts were analyzed separately and together to assess incidence of HBsAg loss in two ways: Any HBsAg loss defined as ever testing HBsAg negative, or sustained HBsAg loss defined as two consecutive HBsAg negative tests at least 6 months apart. Time-to-event analysis was performed to determine rate of HBsAg loss in person years (PYs). Cox regression analysis was used to identify demographic, virological and clinical factors associated with HBsAg loss.
Results:
Among 2,920 patients, mean age at first assessment was 36.8 years, 53% were males, 46% patients were Asian, 34% Alaska Native, 10% White and 10% Black/African-American. Mean follow-up was 14.6 years. At initial presentation, patients in the HEP-B-AK cohort were younger, more likely to be HBV DNA negative and test positive for HBV genotypes D and F, compared to the HBRN and NIH cohorts (Table 1). The cumulative incidence of any HBsAg loss was 1.45 per 100 PYs; being 1.35, 1.54 and 1.96 per 100 PYs among the HEP-B-AK, HBRN and NIH cohorts, respectively. Rates of any HBsAg loss varied significantly by race and were 3.59, 1.91, 1.35 and 1.14 per 100 PYs among Caucasian, Black/African-American, Alaska Native and Asian race, respectively. On multivariate analysis, predictors of any HBsAg loss included age (HR 1.29 per decade, CI 1.20-1.38), male gender (HR 1.29, CI 1.08-1.54), White vs Asian race (HR 1.71, CI 1.19-2.45), HBV DNA (HR 0.52, CI 0.43-0.63) and genotypes B vs A (HR 0.63, CI 0.41-0.96) or E vs A (HR 0.10, CI 0.01-0.75). The multivariate analysis for sustained HBsAg loss yielded age, gender, race, HBV DNA and HBV genotype as significant predictors of HBsAg loss.
Conclusions:
In this multiracial cohort of CHB patients, at initial presentation among multiple demographic, virologic and clinical factors, genotype and race had the strongest association with any and sustained HBsAg loss, respectively.

Chriss Homan

Appearances