|Home | About | Journals | Submit | Contact Us | Français|
Isolated antibody to hepatitis B core antigen (anti-HBc) is a common serologic finding in persons infected with human immunodeficiency virus (HIV), but the outcome and clinical significance are uncertain.
We performed repeated hepatitis B virus (HBV) serologic tests on women who participated in the Women’s Interagency HIV Study and who had isolated anti-HBc at study entry.
Repeated serologic tests were performed for 322 women (282 HIV-infected and 40 HIV-uninfected) at a median of 7.5 years after study entry. Seventy-one percent of women retained isolated anti-HBc serologic status, 20% acquired antibody to hepatitis B surface antigen (anti-HBs), and 2% acquired hepatitis B surface antigen (HBsAg). In unadjusted analysis, increasing age, injection drug use, and hepatitis C viremia were negatively associated with acquisition of anti-HBs. For HIV-infected women, predictors of acquisition of anti-HBs were an increase in CD4 cell count and the use of highly active antiretroviral therapy (HAART). Receipt of drugs with activity against HBV and self-reported HBV vaccination did not predict anti-HBs acquisition. In the multivariable regression model, HAART use remained a significant predictor of anti-HBs acquisition, whereas women with hepatitis C viremia were more likely to retain isolated anti-HBc serologic status.
Isolated anti-HBc status remained stable over time for the majority of women, especially women with chronic hepatitis C virus infection. Development of anti-HBs was predicted by HAART use and an increase in CD4 cell count. We conclude that a proportion of HIV-infected women with isolated anti-HBc have prior natural HBV infection with anti-HBs that is at an undetectable level because of immune dysfunction. Isolated anti-HBc in the presence of chronic hepatitis C virus infection may be attributable to a different phenomenon, such as dysfunctional antibody production.
The presence of antibody to hepatitis B core antigen (anti-HBc) in the absence of both hepatitis B surface antigen (HBsAg) and antibody to hepatitis B surface antigen (anti-HBs) is a common serologic finding, but the clinical significance is uncertain [1, 2]. Isolated anti-HBc could represent active hepatitis B virus (HBV) infection in the absence of hepatitis B surface antigen (HBsAg), resolved infection in which anti-HBs titers are below the level that defines positive, a false-positive test result, or least likely, the “window period” between the resolution of HBs antigenemia and the development of anti-HBs. The significance of each of these scenarios is quite different, and clinicians are faced with uncertainty about whether patients are susceptible to HBV, have active disease that requires monitoring or treatment, or have resolved disease.
Isolated anti-HBc is very common among persons infected with human immunodeficiency virus (HIV) [3–5]. It is hypothesized that, in HIV disease, a proportion of persons with isolated anti-HBc have had natural infection with HBV but have lost anti-HBs in the setting of deteriorating immune function. A number of early investigations have shown that HIV-infected persons lose anti-HBs as HIV disease progresses [6, 7]. This hypothesis is supported by a cross-sectional study in the Women’s Interagency HIV Study (WIHS) that shows that isolated anti-HBc serologic status is independently associated with HIV infection as well as higher HIV RNA levels among those HIV-infected . What remains unclear is what proportion of HIV-infected persons with isolated anti-HBc have cleared natural infection and what the relationship of hepatitis C virus to isolated anti-HBc serologic status is over time. To better understand this phenomenon, we performed follow-up serologic tests on the well-characterized HIV-infected and HIV-uninfected women who participated in the WIHS and who had isolated anti-HBc at study entry.
The WIHS is a longitudinal study of HIV-infected and HIV-uninfected women at risk that enrolled 2054 HIV-infected women and 569 HIV-uninfected women at 6 sites (Chicago, IL; San Francisco Bay Area and Los Angeles, CA; Brooklyn and the Bronx and/or Manhattan, NY; and Washington, DC ) from October 1994 through November 1995. From October 2001 through September 2002, an additional 737 HIV-infected women and 406 uninfected women were enrolled. Informed consent was obtained from all participants in accordance with the guidelines of the US Department of Health and Human Services and the institutional review boards of participating institutions. Women are assessed semiannually by means of an interview, a physical examination, and the collection of blood and genital specimens. The cohort was designed to reflect the demographics of the HIV epidemic among US women. Details of cohort recruitment, retention, and demographic characteristics are published elsewhere [9, 10].
Included in our study were women who had isolated anti-HBc at baseline (i.e., study entry) and serum samples available from a visit >12.5 years after baseline serologic testing. HIV and hepatitis C virus (HCV) seroconverters were excluded. To estimate the incidence of de novo acquisition of hepatitis B, we also repeated serologic tests for a group of women with negative hepatitis B serologic test results at baseline. We chose women who had serum samples available at time points that matched those of the study group and who had a degree of either sexual or injection drug use risk during the follow-up period that was similar to that of the women with isolated anti-HBc. All HBV-seronegative women who used injection drugs (IDU) during the WIHS were included to approximate the incidence of IDU among women with isolated anti-HBc. In addition, women with incident high-risk sexual behavior, defined as ≥2 male sexual partners in a 6-month period, were included at a rate similar to that of the isolated anti-HBc group; the remainder of the group was chosen randomly from women with serum samples available at matching time points.
Of 493 WIHS women who had isolated anti-HBc at baseline, 452 were tested for the presence of HBV DNA ; of these, 8 (2%) of 400 HIV-infected women and none of 52 HIV-uninfected women had detectable HBV DNA. Four of these viremic women had adequate duration of follow-up and are included in the present study.
At the baseline WIHS visit (i.e., study entry), we performed hepatitis B serologic tests for anti-HBs (Ausab enzyme immunoassay [EIA]; Abbott), total anti-HBc (Corzyme EIA; Abbott), and HBsAg (Auszyme Microparticle EIA; Abbott). Of the 3766 WIHS women, 1620 did not have serologic tests for anti-HBs performed at baseline; for many of these women, stored serum samples from the first, second, or third WIHS visit were tested using Vitros ECi (Ortho Diagnostics). Because the Abbott assays were no longer available, all follow-up serologic tests were performed using Vitros ECi for anti-HBs, anti-HBc, and HBsAg. The Vitros ECi has shown excellent correlation with results from Abbott assays for HBsAg and anti-HBs [12, 13].
Hepatitis C antibody testing was performed at baseline by means of EIA 2.0 (Abbott) or EIA 3.0 (Ortho Diagnostics). Qualitative HCV RNA testing was performed for HCV antibody–positive women by means of COBAS Amplicor HCV, version 2.0 (Roche Diagnostics), with a lower limit of 100 IU/mL.
Unadjusted analyses were performed using the χ2 test, Fisher’s exact test, or Student’s t test, as appropriate. Candidate variables that had a statistical significance of P <.10 in unadjusted analysis were entered into multivariable logistic regression models. Time since baseline was retained in all multivariable models. Final multivariable models were created by removing variables in a backward step-wise fashion until only variables significant at P <.05 remained. All analyses were performed using SAS, version 9.1.3 (SAS).
Variables that were explored for an association with serologic outcomes were HIV status, age at baseline, race (white or non-white), absolute CD4 cell count (categorized as <200, 200–500, and >500 cells/mm3), change in CD4 cell count from baseline to follow-up, plasma HIV RNA (copies/mL), HCV antibody (HCV Ab) status (HCV Ab positive/ RNA positive vs. HCV Ab negative and HCV Ab positive/RNA negative), time since baseline, self-reported incident HBV vaccine, use of any anti-viral with activity against HBV, HCV treatment (any treatment with interferon or ribavirin during follow-up), use of highly active antiretroviral therapy (HAART; as defined by the contemporary Department of Health and Human Services guidelines ) at >50% of visits, and any IDU or high-risk sexual behavior between baseline and follow-up serologic tests.
Vaccination history is ascertained in the WIHS by self-report. To estimate the validity of self-report, we performed a review on a convenience sample of medical charts that included those of HIV-infected women who accessed care at sites where WIHS research was conducted. The entire available outpatient record was reviewed to estimate the sensitivity and specificity of self-report. HBV vaccine history was corrected for those women who were found to have reported vaccine history incorrectly; this partially corrected vaccine history is the variable used in analyses and presented in tables.
Of the 493 women who had isolated anti-HBc at the WIHS baseline visit, 322 are included in the study; 112 (23%) died before adequate follow-up, 33 (6.7%) were lost to follow-up, 2 (0.4%) were HIV seroconverters, 20 (4%) had no serum samples available from an appropriate follow-up visit, and 4 (0.8%) had specimens not tested for technical reasons. Follow-up serologic tests were performed at a median of 7.5 years after baseline (range, 2.5–10.4 years). The results of the follow-up serologic tests are presented in figure 1. Seventy-one percent of women with isolated anti-HBc at baseline continued to have only anti-HBc at follow-up, 20% acquired anti-HBs, and 2% acquired HBsAg. The 4 women with occult hepatitis B viremia at baseline retained isolated anti-HBc status. Of 178 women with hepatitis C viremia, 21 (12%) acquired anti-HBs over time, and 32 (18%) had any change in HBV serology (data not shown).
Three hundred twenty-four women who had negative HBV serologic test results at baseline were also retested to estimate the rate of acquisition of HBV among WIHS women (figure 1). These women had a duration of follow-up and a degree of composite sexual and IDU risk during follow-up that were similar to those of women with isolated anti-HBc, although the proportion of these women who had HIV and IDU between serologic assessments was lower than the corresponding proportion in the isolated anti-HBc group. Of these 324 women, 250 (77%) had negative HBV serologic test results at follow-up; one woman acquired anti-HBc and HBsAg, 13 (4%) acquired both anti-HBc and anti-HBs, 56 (17%) acquired anti-HBs alone, and 4 (1%) acquired isolated anti-HBc. The incidence of HBV acquisition, assuming that acquisition of anti-HBc with either HBsAg or anti-HBs reflects incident HBV infection, was 4.3% (14 of 324 women; 0.56 incident cases per 100 person-years). The inclusion of women who acquired isolated anti-HBc increases the incidence to 5.6% (18 of 324 women; 0.75 incident cases per 100 person-years).
We analyzed predictors of conversion from isolated anti-HBc to the presence of both anti-HBs and anti-HBc in both unadjusted and multivariable models. In the unadjusted analysis (table 1), increasing age and active HCV were associated with stable isolated anti-HBc status. Among HIV-infected women, an increase in the CD4 cell count from baseline and HAART use at >50% of visits were associated with acquisition of anti-HBs. Conversion to anti-HBs was not associated with a higher number of sexual partners or IDU during follow-up. Neither self-reported incident hepatitis B vaccination nor treatment with antivirals with activity against HBV or HCV were associated with conversion to the presence of both anti-HBs and anti-HBc.
In multivariable analysis (table 2), only HAART at >50% of visits (odds ratio [OR], 2.11; 95% confidence interval [CI], 1.16–3.82) was associated with an increased likelihood of anti-HBs acquisition. Active HCV was strongly predictive of stable isolated anti-HBc (OR for conversion, 0.25; P <.001).
We performed an analysis of predictors of conversion to HBsAg positivity (n = 7) among women with isolated anti-HBc at baseline, with the candidate predictors named in the Methods section (data not shown). The predictors in both the unadjusted and multivariable model of acquisition of HBsAg were CD4 cell count <200 cells/mm3 at baseline (but not at follow-up), compared with other CD4 cell strata (multivariable OR, 11.32; 95% CI, 1.80–71.1), and HCV RNA negativity (multivariable OR, 12.75; 95% CI, 1.29–126.6). Neither IDU nor high-risk sexual behavior predicted acquisition of HBsAg.
Of the women with isolated anti-HBc, there was no difference in rates of change in anti-HBs status between women whose baseline measurements were performed using the Ausab EIA (19% [31 of 165]) or the Vitros ECi (23% [36 of 157]; OR, 0.78; 95% CI, 0.45–1.3).
We reviewed outpatient medical records for 119 women and found the sensitivity and specificity of self-reported incident HBV vaccine were 0.79 and 0.83, respectively. Forty-five of 322 women with baseline isolated anti-HBc reported incident HBV vaccine. Of these 45 women, 2 converted to anti-HBs alone, 10 converted to the presence of both anti-HBs and anti-HBc, 31 retained isolated anti-HBc status, and 2 had negative HBV serologic test results at follow-up. By HCV status, 3 of 28 women with HCV who reported incident vaccination acquired anti-HBs, as opposed to 9 of 17 women without HCV who reported incident vaccination (relative risk, 0.2; P = .004).
Of the 324 women with negative HBV serologic test results at baseline, 221 were HIV-infected and therefore candidates for HBV vaccination. Of the 221 HIV-infected women, 64 reported HBV vaccination during WIHS follow-up; of these, 24 women acquired anti-HBs alone, 38 remained HBV seronegative, 1 acquired HBsAg and anti-HBc, and 1 acquired both anti-HBs and anti-HBc. By HCV infection status, 3 of 11 vaccinated women with HCV infection acquired anti-HBs, compared with 32 of 73 women without HCV infection (relative risk, 0.62; P = .34).
In this large follow-up study of women with isolated anti-HBc, we found that most women, particularly those with chronic hepatitis C infection, retained isolated anti-HBc status over a median of 7.5 years of follow-up.
The most common change in HBV serology among women with isolated anti-HBc was acquisition of anti-HBs while maintaining anti-HBc (58 of 282 HIV-infected women). We hypothesize that the majority of these women had remote infection with HBV with levels of anti-HBs below the laboratory definition of positive. Evidence that conversion to the presence of both anti-HBc and anti-HBs did not, in most instances, represent incident natural infection with HBV is the statistically significant lower rate of seroconversion among a demographically and behaviorally similar cohort over the same time frame (63 of 322 patients vs. 14 of 324 patients; P <.001) and the fact that acquisition of anti-HBs was not associated with IDU or risky sexual behavior. We found that use of HAART was independently associated with acquisition of anti-HBs after several years of follow-up. Others have found that an increase in the CD4 cell count predicted acquisition of anti-HBs over time among HIV-infected persons with isolated anti-HBs , and early studies found that anti-HBs is lost in persons as immune function declines [6, 7].
HCV viremia, in our study, was strongly associated with continued isolated anti-HBc serologic status. Only 21 (12%) of 178 women with HCV infection acquired anti-HBs over time, compared with 40 (28%) of 144 women without HCV infection, and only 32 (18%) of the 178 women with HCV infection had any change in HBV serology. Although hepatitis C and hepatitis B are epidemiologically linked, it is unclear what proportion of persons with HCV infection and isolated anti-HBc have actually been exposed to HBV. Several lines of study suggest that persons with HCV have an impaired ability to produce anti-HBs. A number of investigators have shown that the production of anti-HBs in response to vaccination is impaired with HCV infection among HIV-infected and HIV-uninfected persons, even in the absence of significant liver disease [16–19]. Leroy et al.  found a correlation between the level of hepatitis C viremia and both nonresponse to HBV vaccination and lower anti-HBs titers in patients with HCV infection without cirrhosis. In a study of hepatitis B vaccination among HIV-infected persons, HCV infection was associated with lack of development of anti-HBs in response to vaccination and lower anti-HBs titers in those who did respond . It is not clear whether lack of production of anti-HBs in response to vaccination and lack of production of anti-HBs in natural infection are analogous.
Investigations into the effect of hepatitis C on the immune system have found a number of manifestations that may be relevant to the production of anti-HBs. Auffermann-Gretzinger et al.  found that persons with active hepatitis C, but not those who had been successfully treated, had impaired dendritic (antigen-presenting) cell maturation. Dendritic cell dysfunction has been demonstrated for both myeloid and plasmacytoid cells in HCV-infected persons .
Another possible explanation for persistent isolated anti-HBc in HCV infection is aberrant antibody production. HCV infection is associated with a number of autoimmune phenomena, most notably mixed cryoglobulinemia but also glomerulonephritis, Sjogren syndrome, thyroiditis, and lichen planus . Patients with HCV infection also produce rheumatoid factor and other autoantibodies . HCV is known to infect lymphocytes and to cause clonal B cell expansion, and this event leads in some patients to B cell non-Hodgkin lymphoma [24, 25]. Whether any of these phenomena are at play in patients with isolated anti-HBc who have not been exposed to HBV is unknown. Our data do not suggest that isolated anti-HBc is caused by false-positive results from laboratory error, given the long-term persistence of anti-HBc among those who originally had the antibody.
Many questions remain with regard to the treatment of patients with isolated anti-HBc, but the WIHS data allow a number of clinically relevant observations. In the WIHS, HBV DNA testing was low yield; only 2% of HIV-infected women with isolated anti-HBc had hepatitis B viremia at baseline, and the majority of these had low-level intermittent viremia . Clearly, isolated anti-HBc in the absence of hepatitis C may represent prior resolved infection and may not require intervention. However, given that 6 women with isolated anti-HBc without HCV infection developed HBsAg during follow-up, we recommend HBV vaccination for those with significant behavioral risk of HBV. Our finding that acquisition of anti-HBs was associated with HAART use and with an increase in the CD4 cell count suggests that vaccine response may improve after immune reconstitution. With HCV infection, there was little change in hepatitis B serologic test results over the course of the study, and the serologic response to vaccination was poor; vaccination in this setting would seem to be of less utility.
There are several notable findings with regard to the women who were serologically negative at baseline. Seventy-five percent of the HIV-infected women remained susceptible to HBV, and only 18% developed isolated anti-HBs consistent with successful hepatitis B vaccination. At least 4.3% of the women who were serologically negative at baseline developed serologic test results consistent with incident HBV infection during the median 7.5 years of follow-up (a rate of 0.56 incident cases per 100 person-years). Of the HIV-infected women with originally negative HBV serologic results, 29% reported incident vaccination. HBV vaccination is recommended for all HIV-infected persons , but compliance with this recommendation is incomplete. In a large review of sites of the HIV Outpatient Study, Tedaldi et al.  found that 32% of eligible HIV-infected patients received hepatitis B vaccine. Our findings support the recommendation to vaccinate all susceptible HIV-infected persons against HBV and suggest that follow-up serology after vaccination may be warranted, especially for those at particularly high risk of HBV acquisition.
Our results should be considered in light of the limitations of our data. For some of the women, different assays were used to detect HBV serologic tests at baseline and follow-up. Although this could have influenced our findings, the stability of HBV serologic test results over time and the lack of difference in rates of seroconversion to anti-HBs between women who had baseline anti-HBs measured by the Ausab EIA and women who had baseline anti-HBs measured by the Vitros ECi led us to believe that the majority of the changes in HBV serologic test results reflect biologic events rather than differences in laboratory assays. The history of hepatitis B vaccination was ascertained by self-report. Although we confirmed that self-report was 79% sensitive for incident vaccination, all of our findings with regard to vaccination response should be considered preliminary observations that require confirmation by studies with more rigorous vaccination ascertainment. Finally, HCV RNA status was ascertained only at baseline. However, we believe that even if some women have been misclassified, our findings are robust, because spontaneous clearance of hepatitis C viremia is rare in established infection, acquisition of hepatitis C was found to be rare in the WIHS in an early study , and the effect of HCV viremia on serologic outcome was highly significant.
In this long-term serologic follow-up study, we found that 71% of women with isolated anti-HBc at the initial WIHS visit retained that serologic status over a median of 7.5 years. Twenty percent of women with isolated anti-HBc acquired anti-HBs over time. Acquisition of anti-HBs was associated with greater exposure to HAART among HIV-infected women, and this fact implies that some women with isolated anti-HBc, particularly those without HCV infection, had remote self-limited HBV infection but had lost anti-HBs with immune function decline. Active HCV infection was significantly associated with maintenance of isolated anti-HBc (P <.001).Whether HCV-infected women with isolated anti-HBc had remote natural HBV infection or isolated anti-HBc from dysfunctional antibody production remains unclear.
Data were collected by the Women’s Interagency HIV Study (WIHS) Collaborative Study Group.
Financial support. The National Institute of Allergy and Infectious Diseases (UO1-AI-35004, UO1-AI-31834, UO1-AI-34994, UO1-AI-34989, UO1-AI-34993, and UO1-AI-42590 to WIHS); the National Institute of Child Health and Human Development (UO1-HD-32632 to WIHS); the National Cancer Institute; the National Institute on Drug Abuse; the National Institute on Deafness and Other Communication Disorders; and the National Center for Research Resources (MO1-RR-00071, MO1-RR-00079, and MO1-RR-00083).
Potential conflicts of interest. All authors: no conflicts.