Effect of Hepatitis G Virus Infection on Progression of HIV Infection in Patients with Hemophilia FREE

Hepatitis G virus (HGV), also known as GB virus C, is a newly discovered member of the Flaviviridae family and is therefore distantly related to hepatitis C virus (HCV) (1). It is transmitted by transfusion, from mother to infant, and probably with low efficiency by sexual contact (1–5). Persistent HGV infection, defined as detection of viremia (HGV RNA) in serum, is present in 1% to 2% of healthy volunteer blood donors and 10% to 20% of populations at high risk for parenteral exposure (1–2, 6). The prevalence of antibody to the viral envelope (anti-E2) is twofold to foURFold higher. Remarkably, HGV does not seem to cause clinically significant hepatitis (2, 6).

Researchers searching for an HGV-related disease have suggested that persons infected with both HIV and HGV progress to AIDS more slowly than those infected only with HIV (7–9). Toyoda and coworkers (7) reported lower HIV viral load and AIDS incidence with detection of HGV RNA in serum. Conversely, Sabin and colleagues (8) found an increased risk for AIDS and death with detection of HGV RNA or anti-E2 antibodies. The differences were not statistically significant in either of these studies. Lefrère and coworkers (9) noted significantly lower HIV viral load, higher CD4⁺ lymphocyte count, and better AIDS-free survival in patients with HGV RNA than in those without HGV RNA or anti-E2 antibodies.

Our objective was to determine the effect of HGV on AIDS-free survival in a well-characterized cohort of patients with hemophilia, controlling for age, HIV viral load, and the highly influential 32-nucleotide deletion in the C-C chemokine receptor 5 polymorphism (CCR5 Δ 32) (10–12).

Between 1982 and 1996, patients with hemophilia or a related coagulation disorder were invited to join the Multicenter Hemophilia Cohort Study (MHCS), as reported elsewhere (10–11). The study particularly sought to understand the cause and natural history of HIV infection and AIDS. Our analysis focused on the effect of HGV on AIDS-free survival in a defined subset of HIV-positive hemophilic men (5, 13). Patients were evaluated semiannually with a standardized physical examination, abstraction of medical records, and phlebotomy. Serum or plasma was separated by centrifugation and stored at −70 °C until used. Samples for our analyses were taken between 1986 and 1996 (median, 1988), when 120 patients were receiving no antiretroviral therapy, 10 were receiving only nucleoside reverse transcriptase inhibitors, and 1 was receiving highly active combination therapy (14). Institutional review boards at each participating institution reviewed and approved the protocol of the MHCS, and signed informed consent was obtained from each patient.

Antibodies to HIV and HCV were detected by enzyme immunoassay and immunoblotting, and viral loads were quantified with commercial kits (Amplicor HIV-1 Monitor, Roche Molecular Systems, Inc., Branchburg, New Jersey; Quantiplex HCV RNA 2.0 Assay [branched DNA], Chiron Corp., Emeryville, California) (10–11, 13). Genotyping of CCR5 was performed by using single-stranded conformation polymorphism–heteroduplex analysis (12). Hepatitis G virus RNA was detected by using reverse transcriptase polymerase chain reaction with the Enzymun-Test DNA (Boehringer Mannheim Corp., Indianapolis, Indiana), and anti-E2 antibodies were detected by using the Enzymun-Test Anti-HGenv (Boehringer Mannheim Corp.) (5).

The date of primary HGV infection could not be determined or imputed because patients with hemophilia were repeatedly exposed to bloodborne viruses and only one sample from each patient was tested. Therefore, HGV status (HGV-positive with detectable HGV RNA or anti-E2 antibodies, or HGV-negative) was assumed to be fixed at the time of HIV seroconversion. We defined AIDS as the development of a life-threatening opportunistic infection or malignant condition. The Kaplan–Meier method was used to estimate the effect of HGV infection on AIDS-free survival, and median values were used to define subgroups for further analysis. Groups were compared by using log-rank tests. The effects of other variables were evaluated with multivariate proportional hazards models. All models conformed to the proportional hazards assumption. Continuous variables were described with mean values and 95% CIs. Analyses were performed by using SAS software, version 6.12 (SAS Institute, Cary, North Carolina), and BMDP software, version 7.0 (Statistical Software, Inc., Los Angeles, California).

The funding source exercised no direct control over the analysis or the decision to submit the paper for publication.

Of the 131 patients with hemophilia who were infected with HIV and HCV, HGV RNA and anti-E2 antibodies were detected in 3, HGV RNA alone was detected in 19, and anti-E2 alone was detected in 38. Therefore, 60 of 131 patients (46%) were infected with HGV. Hepatitis G virus was present in 4 of the 11 patients receiving antiretroviral therapy (36%); 3 had anti-E2 antibodies, and 1 had HGV RNA. At the approximate time of HGV sampling, HGV-positive patients had significantly higher CD4⁺ lymphocyte counts (mean, 529 cells/mm³ compared with 318 cells/mm³; mean difference, 211 cells/mm³ [95% CI, 88 to 333 cells/mm³]) than HGV-negative patients. We found that CD4⁺ lymphocyte counts were similar in patients who were positive for HGV RNA (mean, 510 cells/mm³) and those who were positive for anti-E2 antibodies (mean, 528 cells/mm³). Patients who were HGV-positive had marginally higher CD8⁺ lymphocyte counts than HGV-negative patients (878 cells/mm³ compared with 716 cells/mm³; mean difference, 161 cells/mm³ [CI, −69 to 392 cells/mm³]) and were older (30 years compared with 27 years; mean difference, 3.5 years [CI, −0.2 years to 7.3 years]).

Viral loads for HCV were similar in patients who were positive for HGV RNA (mean, 6.91 log₁₀ equivalents/mL) and those who were positive for anti-E2 antibodies (mean, 6.92 log₁₀ equivalents/mL). However, HGV-negative patients had slightly lower HCV viral loads (mean, 6.65 log₁₀ equivalents/mL; mean difference, 0.27 log₁₀ equivalents/mL [CI, −0.03 to 0.56 log₁₀ copies/mL]) than those who were HGV-positive. Viral loads for HIV were similar in HGV-negative patients (mean, 4.32 log₁₀ copies/mL), HGV RNA–positive patients (mean, 4.21 log₁₀ copies/mL), and anti-E2 antibody–positive patients (mean, 4.35 log₁₀ copies/mL).

During a mean follow-up of 11.2 years, when most patients received no antiretroviral treatment or zidovudine alone, 72 patients developed an AIDS-defining opportunistic illness. In univariate Kaplan–Meier analysis, risk for AIDS was significantly lower among HGV-positive patients (P = 0.03) (Figure); AIDS incidence in these men was 3.9 per 100 person-years compared with 5.9 per 100 person-years in HGV-negative patients (difference, 1.9 per 100 person-years [CI, −0.3 to 4.2 per 100 person-years]). The 19 patients who were positive for HGV RNA and the 38 patients who were positive for anti-E2 antibodies had better 12-year cumulative AIDS-free survival rates (72% [CI, 50% to 93%] and 66% [CI, 51% to 82%], respectively) than HGV-negative patients (40% [CI, 28% to 52%]).

After adjustment for age, HIV viral load, and CD4⁺ and CD8⁺ lymphocyte counts in a proportional hazards model, risk for AIDS was approximately 40% lower for HGV-positive patients than for HGV-negative patients (Table, model 1). When added to this model, HCV viral load was not predictive of AIDS and did not substantially affect relations of the other variables (data not shown). One hundred of 131 patients had available CCR5 genotypes. In a proportional hazards model that included CCR5 genotype, age, and HIV viral load (Table, model 2), risk for AIDS was significantly reduced in patients with HGV infection (hazard ratio, 0.48 [CI, 0.26 to 0.87]).

We found that HGV infection decreased the risk for AIDS among HIV-infected men with hemophilia, an association that was independent of age and CCR5 genotype. Because few patients were receiving antiviral drugs, the association was not confounded by therapy.

Our findings are similar to and expand on previous findings. In a study of Japanese patients with hemophilia, HIV viral load and AIDS incidence were lower in those who had HGV RNA in serum. However, the difference was not statistically significant (7), and patients with HGV RNA were slightly younger, a characteristic associated with better prognosis (10–11). In age-matched and age-unmatched analyses, HIV-infected French adults with HGV RNA had significantly lower HIV viral loads, higher CD4⁺ lymphocyte counts, and better AIDS-free survival rates than patients who were negative for HGV RNA and anti-E2 antibodies (9). We found a similar reduction of AIDS risk in patients with HGV RNA or anti-E2 antibodies. Contrary results were found in British patients with hemophilia (8), but the reason for this is unknown. Prevalent cohort studies like ours cannot determine whether HGV truly decreases risk for AIDS among HIV-infected patients or is merely a surrogate for some other causal event. Nonetheless, we found no evidence of bias by age or duration of follow-up that might provide a trivial explanation.

The rate at which immune deficiency and AIDS develop varies greatly among HIV-infected patients. Younger patients and those with CCR5 Δ 32 heterozygosity have substantially better rates of AIDS-free survival (10–12), as do patients receiving highly active antiretroviral therapy and Pneumocystis carinii pneumonia prophylaxis. Because the MHCS was begun early in the AIDS epidemic, relatively few participants received antiretroviral therapy or prophylaxis until recently (14). In our study, the effect of HGV infection on AIDS-free survival was independent of CCR5 genotype. Nonpharmacologic exogenous factors that affect AIDS risk remain poorly defined. Co-infection with cytomegalovirus, for example, seems to have little or no effect, except possibly in infants (15–16).

Hepatitis G virus viremia is cleared in more than 50% of acute infections. This is followed by the appearance of anti-E2 antibodies, which usually persist over the long term (6). Remarkably, although HGV viremia often lasts for years, it does not seem to cause chronic disease (2, 6, 17). Like HCV, its distant flavivirus relative, HGV infects but has little replication in lymphocytes (18) and could therefore interfere with HIV within lymphocytes. Alternately, HGV could indirectly affect AIDS risk through induction of various chemokines and other soluble factors (19) or by altered expression of chemokine receptors, which are essential co-receptors for HIV (20). Investigation of the direct and indirect effects of HGV infection on chemokines, other soluble factors, and chemokine receptors could provide a mechanism for our observation of decreased risk for AIDS.

The following institutions and investigators participated in the Multicenter Hemophilia Cohort Study:

J.J. Goedert, T.R. O'Brien, P.S. Rosenberg, C.S. Rabkin, E.A. Engels, M. Hisada, E. Maloney, M.H. Gail, S.J. O'Brien, M. Dean, M. Carrington, M. Smith, and C. Winkler, National Cancer Institute, Rockville, Maryland, and Frederick, Maryland; B. Konkle, Cardeza Foundation Hemophilia Center, Philadelphia, Pennsylvania; M. Manco-Johnson, Mountain States Regional Hemophilia and Thrombosis Program, University of Colorado, Aurora, Colorado; D. DiMichele and M.W. Hilgartner, Hemophilia Treatment Center, New York Presbyterian Hospital, New York, New York; P. Blatt, Christiana Hospital, Newark, Delaware; L.M. Aledort and S. Seremetes, Hemophilia Center, Mount Sinai Medical Center, New York, New York; K. Hoots, Gulf States Hemophilia Center, University of Texas at Houston, Houston, Texas; A.L. Angiolillo, N.L.C. Luban, Hemophilia Center, Children's Hospital National Medical Center, Washington, D.C.; A. Cohen and C.S. Manno, Hemophilia Center, Children's Hospital of Philadelphia, Philadelphia, Pennsylvania; C. Leissinger, Tulane University Medical School, New Orleans, Louisiana; G.C. White II, Comprehensive Hemophilia Center, University of North Carolina, Chapel Hill, North Carolina; M.M. Lederman, S. Purvis, and J. Salkowitz, Case Western Reserve University School of Medicine, Cleveland, Ohio; C.M. Kessler, Georgetown University Medical Center, Washington, D.C.; A. Karafoulidou and T. Mandalaki, Hemophilia Center, Second Regional Blood Transfusion Center, Laikon General Hospital, Athens, Greece; A. Hatzakis and G. Touloumi, National Retrovirus Reference Center, Athens University Medical School, Athens, Greece; W. Schramm and F. Rommel, Medizinische Klinik Innerstadt der Maximilian, Universitaet Muenchen, Munich, Germany; P. de Moerloose, Haemostasis Unit, Hôpital Cantonal Universitaire, Geneva, Switzerland; S. Eichinger, University of Vienna Medical School, Vienna, Austria; K.E. Sherman, University of Cincinnati Medical Center, Cincinnati, Ohio; D. Whitby and D. Waters, Scientific Applications International Corp., Frederick, Maryland; and V. Lamprecht and B.L. Kroner, Research Triangle Institute, Rockville, Maryland.