Josiah Brown Poster Abstract


Tina D. Kantaria
Debika Bhattacharya, MD MSc
Jonathan Grotts, MA, Chi-Hong Tseng, PhD
Clinical Outcomes in HIV/HBV Co-infected Individuals in a South African Cohort


HIV and Hepatitis B virus (HBV) co-infection poses a challenge in many regions around the world where both viruses are highly endemic. In Africa, it is estimated that 10% of HIV-infected individuals are HBV co-infected, and may be as high as 25% in countries that are highly endemic. HIV/HBV co-infection is associated with increased risk of hepatotoxicity after initiating ART, however, there is equipoise whether this is apparent across all countries and what the effect of HBV infection has on HIV clinical and virologic outcomes. Understanding what factors may play a role in each of these outcomes and how they can be modified may help minimize the risk of adverse events and optimize HIV and HBV treatment in co-infected individuals.


To assess the impact of HBV co-infection on the incidence of hepatotoxicity and immunologic and virologic outcomes in a South African cohort of HIV-infected individuals after initiating ART.


We conducted a post-hoc analysis of data from participants enrolled in a randomized trial of ART administration in South Africa (CIPRA-SA.) Participants were classified as having HIV mono-infection, HIV/HBV co-infection (further classified as HBeAg positive and HBeAg negative) and occult HBV infection (defined as HBsAg-negative but with detectable HBV viremia.) In participants with baseline liver enzyme levels within normal ranges (ALT 31, AST 35) hepatotoxicity was defined as: grade 0 (1.25x ULN); grade 1 (1.25-2.5x ULN); grade 2 (2.6-5x ULN); grade 3 (5.1-10x ULN); and grade 4 (10x ULN.) In participants with elevated baseline levels, hepatotoxicity was defined as: grade 0 (1.25x baseline); grade 1 (1.25-2.5x baseline); grade 2 (2.6-3.5x baseline); grade 3 (3.6-5x baseline); and grade 4 (.5x baseline). Clinical and virologic outcomes were measured at 6, 12, 24, 48, 72, and 96 weeks. The relationships between hepatotoxicity and HBV co-infection was evaluated with a Cox regression survival analysis, with the primary outcome identified as the first highest hepatotoxic event.  Preliminary analysis of the association between HIV/HBV co-infection and CD4 gain and HIV VL suppression was performed using a univariate Firth logistic regression model.



812 HIV-infected individuals were included in the analysis; 752 (93%) were HIV mono-infected, 48 (5%) were HIV/HBV co-infected, and 12 (1%) had occult HBV infection. Of the HIV/HBV coinfected, 28 (58.3%) were HBeAg positive and 20 (41.7%) were HBeAg negative. HBV viral load above 20,000 IU/mL was reported in 92.9% of HBeAg positive participants, 26.3% of HBeAg negative participants, and 18.21% of occult HBV participants. At baseline, the HIV/HBV co-infected group had higher ALT and AST levels (33 and 42, respectively) compared to the HIV mono-infected group (30 and 36.5, respectively.) The HBeAg positive group had higher median baseline ALT and AST levels (38.5 and 47.5, respectively) compared to HBeAg negative (30 and 36.5, respectively, p<0.001).

292 (39.2%) HIV mono-infected and 28 (59.6%) HIV/HBV co-infected participants experienced a grade 1-4 hepatotoxic event (p=0.009). 33 (4.4%) HIV mono-infected and 10 (21.3%) of HIV/HBV co-infected participants experienced grade 3-4 hepatotoxic event (p<0.001.) In a time-to-event analysis, HIV/HBV co-infection HBeAg positivity was significantly associated with hepatotoxicity (HR 7.03, 95% CI 3.24-15.22) compared to HBeAg negativity (HR 2.44, 95% CI 0.58-10.15, p<0.001.) NVP and EFV-containing ART regimens had HR of 4.18 (95% CI 0.54-32.17) and 2.4 (95% CI, respectively, but this did not reach statistical significance.

When assessing the impact of HIV/HBV co-infection on CD4 gain, HBeAg positivity showed a lower overall CD4 gain at 24 weeks, 48 weeks, and 96 weeks than HBeAg negative, HIV mono-infected, and occult HBV groups, but the difference was not statistically significant.


HBV co-infection occurred in 5% of HIV-infected participants in HIV clinical trial in South Africa, of which 58.3% were HBeAg positive. Grade 3-4 hepatotoxicity occurred in 4.4% and 20.8% of HIV mono-infected and HIV/HBV co-infected participants, respectively (p<0.001). In a univariate time-to-event analysis of hepatotoxicity, HIV/HBV co-infection HBeAg positivity was significantly associated with hepatotoxicity (HR 7.03, 95% CI 3.24-15.22, p<0.001.) No statistically significant differences in CD4 gain were found between mono and co-infected groups. More attention to hepatotoxicity monitoring in HIV/HBV co-infected HBeAg positive patients may be warranted.