An Evaluation of Hepatitis B Virus Diagnostic Methods and Responses to Antiretroviral Therapy among HIV-Infected Women in Thailand

Abstract

Coinfection with HIV and hepatitis B virus (HBV) is common in resource-limited settings but is frequently not diagnosed. The authors retrospectively tested specimens for HBV in HIV-infected Thai women who had participated in an antiretroviral therapy (ART) clinical study. A substantial proportion (27 of 211; 13%) of HIV-infected women were HBV coinfected. Among HIV/HBV-coinfected women, the authors observed similar rates of antiretroviral-associated liver toxicity (despite nevirapine [NVP] use) and CD4 count reconstitution as observed in HIV-monoinfected women. Hepatitis B surface antigen (HBsAg) screening detected the majority (81%) of HBV coinfections, including all 5 HBV-coinfected women who did not suppress HBV despite 48 weeks of lamivudine (3TC)-containing ART and could be used to tailor ART for patients diagnosed with HBV coinfection in accordance with World Health Organization guidelines. Although HBsAg screening did not diagnose 5 occult HBV coinfections, these women achieved HBV suppression on 3TC-containing ART, suggesting that not detecting occult HBV coinfection would have limited clinical impact.

Introduction

Although hepatitis B virus (HBV) infection is vaccine preventable, many regions of southeast Asia, including Thailand, have a high prevalence of HBV infection. ¹ HIV infection adversely impacts the natural history of hepatitis B disease by increasing the replication rate of HBV and accelerating progression to cirrhosis, hepatocellular carcinoma, and other liver-related morbidity. ² While HBV cure is often not attainable in HIV-infected adults, several HIV antiretroviral drugs also have anti-HBV activity, and HBV can be effectively treated in many coinfected patients. ³ The antiretroviral nucleoside analog, lamivudine (3TC), is a component of many first-line antiretroviral therapy (ART) regimens. Lamivudine has anti-HBV activity, but single-drug anti-HBV therapy with 3TC is not recommended ⁴ because 3TC resistance can develop. ⁵ Thailand’s ART guidelines now recommend screening for HBV and treating HIV/HBV-coinfected patients with an antiretroviral regimen that includes 2 drugs with anti-HBV activity (eg, tenofovir [TDF] plus 3TC). ³ Although the 2010 World Health Organization (WHO) ART guidelines also recommend treating HIV/HBV-coinfected patients with an antiretroviral regimen that includes 2 drugs with anti-HBV activity, these guidelines do not recommend specifically that HIV-infected patients be routinely screened for HBV coinfection or specify which diagnostic assays should be used to detect HBV. ⁶ To evaluate the frequency of HBV coinfection, optimal diagnostic methods, and response to 3TC-containing ART, we retrospectively tested stored specimens for HBV in a cohort of HIV-infected Thai women who had participated in an ART clinical study.

Material and Methods

Women included in this analysis had participated in a multicountry cohort study designed to assess the effectiveness of ART in HIV-infected women previously exposed to a single-intrapartum dose of nevirapine (NVP). ⁷ Eligible participants were women aged 18 years or older who were ART-naive and qualified to start ART (CD4 count <200/µL or an AIDS-defining illness), according to the Thai 2005 to 2007 national guidelines. The present analysis focusing on HBV coinfection was restricted to participants who provided written informed consent to long-term storage of blood specimens for later testing.

From 2005 to 2007, we enrolled HIV-infected women in Thailand initiating ART and followed their response to therapy for 48 weeks. Participants initiated 3TC and stavudine (d4T) plus either NVP or efavirenz (EFV). Stored specimens were tested for serologic markers of HBV infection (at 2 and 48 weeks) using enzyme immunoassays (Murex; Abbott Diagnostics, Dartford, United Kingdom) and for HBV DNA (at baseline and at 48 weeks) using a viral load assay (COBAS Amplicor Monitor v2.0; Roche Molecular Diagnostics, Pleasanton, California). Serum transaminase levels were measured at baseline and at 2, 4, 8, 16, and 24 weeks to detect evidence of antiretroviral-associated liver toxicity, defined as an alanine aminotransferase (ALT) or aspartate aminotransferase (AST) level ≥250 IU/mL. ⁸

We defined a participant as having “HBV coinfection” if she was seropositive for hepatitis B surface antigen (HBsAg) or had detectable HBV DNA. Hepatitis B surface antigen-positive participants were also tested for hepatitis B e antigen (HBeAg). A participant was defined as “HBV immune” if she was seronegative for HBsAg but seropositive for hepatitis B surface antibody (HBsAb), regardless of hepatitis B core antibody (HBcAb) status. We defined a participant as “isolated core antibody positive” if she was seropositive for HBcAb but seronegative for HBsAg and HBsAb; and as “never HBV infected” if she was seronegative for HBsAg and HBcAb. We performed HBV viral load testing on the baseline and 48-week specimens of HBsAg-positive and isolated core-antibody-positive participants. Hepatitis B viral load was defined as high if it was ≥1000 copies/mL, as low if it was 100 to 999 copies/mL, and as undetectable (or suppressed if the participant was receiving anti-HBV treatment) if it was below the assay’s limit of detection (100 copies/mL). Using previously established methods, ⁹ we amplified a 447-basepair fragment from the HBV polymerase gene from all the specimens with a high-HBV viral load to determine the HBV genotype and to detect the emergence of mutations associated with 3TC resistance. For each specimen, we sequenced approximately 48 clones derived from the amplicons. Differences in clinical and demographic variables among participants with and without HBV coinfection were analyzed using SAS version 9.1 (SAS Institute Inc, Cary, North Carolina), applying the Wilcoxon rank-sum test for continuous variables and the chi-square test or Fisher exact test for categorical variables.

Results

Overall, 211 (97%) of 217 Thai study participants consented to stored specimen testing. At baseline, participants had a median age of 31 years (interquartile range [IQR]: 27-36 years), a median CD4 count of 145 cells/mm³ (IQR: 49-220 cells/mm³), and a median HIV RNA viral load of 108 000 copies/mL (IQR: 39 000-309 000 copies/mL). Additionally, 22 (11%) had abnormal baseline transaminases (ALT or AST ≥50 IU/mL).

At baseline, 22 (11%) women were HBsAg-positive (Figure 1). A high HBV viral load was detected in 10 (50%) of 20 women whose sera were available for testing, of whom 6 had an HBV viral load >1 000 000 copies/mL. All 10 women were also HBeAg positive. Among the 9 women with baseline specimens available for sequencing, 8 were HBV genotype C and 1 woman was dually infected with HBV genotypes B and E; no mutations associated with 3TC resistance were detected at baseline. The other 10 HBsAg-seropositive women who had a low or undetectable HBV DNA viral load were HBeAg negative. Among the remaining 189 HBsAg-negative women, 60 (28%) women were HBV immune, 36 (17%) women were isolated core antibody positive, and 93 (44%) women had no evidence of prior HBV infection. Among the 34 women who were isolated core antibody positive and whose plasma was available for molecular testing, a high HBV viral load was detected in 2 (6%) women and a low HBV viral load was detected in 3 (9%) women (Figure 1). In total, 27 (13%; 95% confidence interval [CI]: 9%-18%) HBV coinfections were detected including 22 women seropositive for HBsAg and 5 women who were isolated core antibody positive with a detectable HBV viral load. Hepatitis B surface antigen testing detected 22 (81%) of 27 women with HBV coinfections and 10 (83%) of 12 HBV coinfections with a high HBV viral load.

Figure 1.

Hepatitis B virus (HBV) serology and viral load results among HIV-infected Thai women (N = 211). *All participants with a low-HBV viral load achieved HBV viral load suppression at 48 weeks; **16 (55%) remained isolated core antibody positive, 11 (38%) became HBsAb positive, and 2 (7%) became HBcAb negative. HBV indicates hepatitis B virus; HBsAg, hepatitis B surface antigen; HBsAb, hepatitis B surface antibody; HBcAb, hepatitis B core antibody.

Baseline median CD4 counts were not significantly different among women with and without HBV coinfection (156 versus 142 cells/mm³, respectively; P = .89). Baseline serum ALT levels were also not significantly different (20 versus 19 IU/mL, respectively; P = .49), and none of the 22 women with abnormal baseline transaminases had HBV coinfection. After 48 weeks of ART, HIV/HBV-coinfected women did not have significantly different HIV clinical responses compared with HIV-monoinfected women, as measured by increases in median CD4 count (156 versus 154 cells/mm³ increase; P = .52), rates of HIV viral load suppression (96% versus 96% with HIV RNA viral load <400 copies/mL; P = 1.00), and rates of ART-associated liver toxicity (7% versus 10%; P = 1.00).

Among the 12 HIV/HBV-coinfected women with high pre-ART HBV viral loads, 5 (42%; 95% CI: 19%-69%) failed to suppress HBV replication after 48 weeks of ART, and all 5 had been HBsAg seropositive at baseline. The remaining 7 suppressed HBV (ie, to <100 copies/mL) including 3 who became HBeAg seronegative. Compared with the 5 women who failed to suppress HBV, the 7 suppressers had a higher baseline median ALT level (26 versus 16 IU/mL; P = .04), experienced a higher peak ALT level on ART (89 versus 40 IU/mL; P = .02), and were significantly more likely to experience at least one abnormal ALT level (>50 IU/mL) on ART (86% versus 0%; P = .02), respectively. Baseline median CD4 counts were lower (69 versus 160 cells/mm³; P = .20), and median increases in CD4 count after 48 weeks of ART were higher (209 versus 106 cells/mm³; P = .53) among women who suppressed HBV compared with women who failed to suppress HBV, but these differences did not reach statistical significance. All 13 women with a low or undetectable pre-ART HBV viral load suppressed HBV after 48 weeks of ART.

Specimens were available for sequencing at 48 weeks for 4 of the 5 women who failed to suppress HBV. Mutations associated with 3TC resistance were detected at low frequencies (2%-5% of sequenced clones) in 2 of these women (note 1). HIV RNA viral load was suppressed (<400 copies/mL) in 4 of the 5 women who failed to suppress HBV, suggesting adequate medication adherence.

Discussion

We found that a substantial proportion of HIV-infected Thai women (13%) were HBV coinfected. These high rates (approximately 2-3 times the HBV prevalence in the general Thai adult population ¹⁰ ) may result from HIV-associated immunosuppression that increases the risk of reactivating HBV and developing chronic infection, ¹¹ or a greater likelihood of HBV exposure. HBV coinfection is not routinely screened for in HIV-infected patients in many resource-limited settings. ¹² A strategy that screens HIV-infected patients for HBsAg and then tailors ART specifically to patients identified as being HBV coinfected should improve the effectiveness of treatment. ¹³ In this study, HBsAg testing detected the majority of HBV coinfections, but a significant proportion (19%) of HBV coinfections were only detected with HBV DNA testing, which is often not widely available in resource-limited settings. These occult HBV infections (ie, HBsAg negative but detectable HBV DNA) ¹⁴ were detected among women who were isolated core antibody positive, a status that may have resulted from the loss of HBsAb. Although HBsAg screening did not detect 5 occult HBV coinfections, all 5 women achieved HBV viral load suppression on 3TC-containing ART. This finding suggests that not detecting occult HBV coinfection would have limited clinical impact.

World Health Organization and Thai guidelines for HIV treatment now recommend the use of 2 antiretroviral agents with dual anti-HIV/anti-HBV activity, such as 3TC and TDF, in HIV/HBV-coinfected patients to achieve better rates of HBV suppression. ^3,6 We retrospectively observed that a significant proportion of HIV/HBV-coinfected women with a high pre-ART HBV viral load failed to suppress HBV replication while prescribed ART containing 3TC but without TDF, which is consistent with previous studies. ¹⁵ The majority of women who suppressed HBV replication in our study had a transient increase in ALT, suggesting that ART-induced immune reconstitution is an important contributor to suppression of HBV replication. In women who failed to suppress HBV replication, we observed little evidence of HBV resistance, suggesting that 3TC monotherapy lacked potency to suppress HBV in highly viremic women. HBV 3TC resistance, however, often emerges after longer (2-5 years) durations of monotherapy. ⁵

It has been proposed that HBV coinfection may adversely affect the natural history of HIV disease by increasing susceptibility to antiretroviral-associated liver toxicity and by impairing the immunological response to HIV treatment. ² Among the HIV/HBV-coinfected women in this analysis, we observed similar rates of antiretroviral-associated liver toxicity as observed in HIV-monoinfected women despite the predominant use of NVP. In addition, we did not observe differences in the rates of CD4 count reconstitution or rates of HIV viral load suppression to suggest that HBV coinfection was impacting the response to HIV treatment.

There were several limitations to our study. First, all of the participants in this study were Thai women; and therefore, the findings cannot necessarily be extrapolated to men or non-Thai populations. In addition, HBV genotype C predominated in this study, and this genotype has been associated with more serious liver disease, higher HBV viral loads, and lower rates of spontaneous HBeAg seroconversion compared with other genotypes. ¹⁶ It will be important to confirm these results in settings with other HBV genotypes (eg, in sub-Saharan Africa where genotypes A1 and D predominate). Finally, the number of HIV/HBV-coinfected women with a high baseline HBV viral load was small, which limited the precision of some of our estimates.

In conclusion, HBV coinfection was common among HIV-infected Thai women in this cohort. HBsAg testing detected the majority of (but not all) HBV coinfections and could be used to tailor ART for patients diagnosed with HBV coinfection in accordance with WHO guidelines. ⁶ Although the clinical implications of occult HBV coinfection missed by HBsAg testing alone were limited in our study, these results (ie, the frequency of occult HBV coinfection and response to treatment) should be confirmed in other settings.