Impact of chronic hepatitis C and/or D on liver fibrosis severity in patients co-infected with HIV and hepatitis B virus

The histological study on the reciprocal influence of chronic hepatitis C (HCV) and/or delta (HDV) on liver damage in a cohort of 134 HIV-HBV co-infected patients concluded on a significant association between HDV co-infection (noted in 13 patients) and Metavir F3-F4 liver fibrosis score [odds ratio (OR) = 7.08, 95% confidence interval (CI) = 1.06-47.28 for HBV-HDV, OR = 10.02, 95% CI = 1.03-97.42 for HBV-HCV-HDV, compared to OR = 1.76, 95% CI = 0.50-6.17 for HBV alone]. Co-treatment of other multiple viral hepatitis infections should also be taken into consideration, especially in the case of chronic HDV.

Overlapping routes of transmission in hepatitis B (HBV), delta (HDV) and/or C (HCV) viruses induce a high prevalence of multiple hepatic co-infection in HIV-infected patients, which contributes to an increased liver-related morbidity ^[1,2]. Several studies have focused on the role of HIV-induced immunosupression and hepatic viral interference as co-factors of liver damage ^[3,4], but none has specifically studied the implication of each hepatic virus in the fibrogenic process. The present study aimed to determine and quantify the respective contribution of HCV and/or HDV on liver fibrosis among HIV-HBV co-infected patients.

Patients in this cross-sectional study were prior enrolled in the HIV-HBV Cohort Study, for which the design has been described previously ^[5]. Serum antibodies to HCV and HDV, with quantification of HCV-RNA or HDV-RNA and genotyping, were determined. Four mutually exclusive groups were defined per positive concordant serology 6 months prior to and at inclusion: (i) HBV; (ii) HBV-HCV; (iii) HBV-HDV; and (iv) HBV-HCV-HDV. Liver fibrosis was assessed by histological evaluation of liver paraffin-embedded biopsies obtained within 18 months of inclusion (2002-2003). Samples were stained with hematoxylin-phloxin-saffran and picrosirius red and were then blindly read by four pathologists who were unaware of the clinical and biological data. Histological fibrosis was scored using the METAVIR classification ^[6] and trichotomized on levels of fibrosis (F0-F1 versus F2 versus F3-F4). An ordinal logistic regression model was fitted and adjusted odds ratios (ORs), with 95% confidence interval (CI) and P-values, were calculated for each variable. Variables included a priori in the model were age, body mass index, log₁₀ viral copies of HBV DNA, CD4+ cell count, current treatment with Lamivudine and Tenofovir, alcoholic consumption, and sex; which are known factors affecting liver fibrosis.

Among the 134 patients with at least one liver biopsy at inclusion (median age = 40.2 years, interquartile range (IQR) = 35.8-44.4; sex ratio = 18), 23 patients were identified with additional hepatitis co-infection (HBV-HCV, n = 10; HBV-HDV, n = 6; HBV-HCV-HDV, n = 7). Levels of median CD4+ cells (423/mm³, IQR = 283-545) and combined antiretroviral therapy (cART) exposure (80.8% of all patients and 76% treated with a drug of dual HIV-HBV activity) were similar across all groups. HIV-RNA was above the detection limit (50 copies/ml) in 69.2% of HDV-co-infected patients (9/13) versus 57.8% of non-HDV co-infected patients (70/121). Active HBV replication (> 200 copies/ml) was noted in 100 of 134 (74.6%) patients. HBV genotypes were distributed as follows (n = 97): A = 63 (66.7%), D = 10 (10.8%), E = 7 (7.5%), G = 13 (14.2%). Replicative HCV infection was noted in eight of 17 patients (47.1%), revealing a predominance of genotype 1 (n = 4), followed by genotype 3 (n = 2) and genotype 4 (n = 2). Active HDV replication was also observed in ten of 12 patients (nine patients with genotype 1 and one patient with genotype 5). A METAVIR fibrosis score of 2 or greater was found in 83.4% (5/6) of HBV-HDV and 100% (7/7) of HBV-HCV-HDV co-infected patients. Prevalence of activity in the liver (METAVIR activity score of 2 or greater) in patients with HCV and/or HDV co-infection was 65% versus 35% in the HBV group (P = 0.007). In these HIV-infected patients, the adjusted odds of liver fibrosis increased in patients with HBV-HDV and HBV-HCV-HDV (OR = 7.73, 95% CI = 1.11-53.78; OR = 8.54, 95% CI = 1.24-56.10, respectively, for fibrosis stage F3-F4), but not in HBV-HCV co-infected patients (OR = 1.76, 95% CI = 0.50-6.17) compared to patients with HBV alone. A secondary analysis conducted after regrouping HDV co-infection groups (Table 1, model 2) yielded similar results with an 8.17-fold increased odds of liver fibrosis compared to only HBV (95% CI = 1.94-34.28, P = 0.004), whereas the increase of odds related to HBV-HCV was not significant (OR = 1.76, 95% CI = 0.50-6.18).

Table 1

Outside the context of HIV, studies regarding multiple hepatic co-infections focused on viral interactions in which HDV played a dominant role with a subsequent inhibition of HBV and HCV replication ^[7,8]. In HIV-infected patients, data on viral interference are controversial and those reporting histological evidence of a dominant viral influence over another are non-existent. Studies predating the extensive use of cART either demonstrated an absence of an inhibitory effect of HDV or an inhibitory role of HDV replication over HBV and HCV ^[9], the latter of which was confirmed by two recent studies in a less immunocompromised population ^[10,11]. The level of viral replication in chronic HCV might not be a determinant to the degree of liver fibrosis ^[12]; however the putative direct cytopathic effect of replicating HDV ^[13] may hasten the liver fibrosis process, leading to a more rapid fibrosis evolution in the case of HDV infection compared to HCV alone ^[14]. Genotype 1 was the most prevalent genotype in HCV and HDV co-infection. This is consistent with the geographical repartition of such genotypes in HIV-negative patients born in Europe ^[15,16], whereas the only patient infected with an HDV genotype 5 strain, identified as a newly-characterized African clade ^[17], came from Mali. These data show that the prevalences of HCV and HDV genotypic characteristics in HIV-infected patients are similar to those of HIV-negative populations and influenced by geographical origin.

Two main limitations arise in this study. First, liver progression in these groups could not be established due to the cross-sectional design. However, a recent longitudinal study comparing HIV-HBV and HIV-HBV-HDV co-infections demonstrated significantly higher number of cirrhosis and hepatic decompensation cases with HDV infection ^[11]. Second, although our population was well characterized, the patient sample size was relatively small, in line with other reports ^[10,11,18] and reflecting the difficulty of identifying these rare multiinfected patients. This problem was attenuated by using exact ordinal logistic models with all HDV-infected patients merged together.

Recently published guidelines on HIV-HBV co-infection management outlined the importance of early treatment of chronic HBV ^[19]. Co-treatment of all multiple viral hepatitis infections should be taken into consideration ^[18]. However, although efficient antiviral drugs are available to treat HCV in HIV-infected patients ^[20,21], the treatment of HDV is still an unsolved issue ^[22]. One clinical trial using high doses of interferon in HIV-negative, HBV-HDV co-infected patients showed promising results on survival and fibrosis regression ^[23]. Furthermore, Lamivudine with Peg-Interferon may be more efficient than Peg-Interferon alone in controlling HBV and HDV replication ^[24]. Aside from Lamivudine, more effective antiviral treatments (i.e. Tenofovir) must be examined longitudinally to assess any indirect effect on HDV in multiple hepatic co-infection.

Acknowledgement

We thank Professor F. Zoulim for his collaboration on HBV genomics, P. Tran and N. Algans for managing the logistics of the French HIV-HBV Cohort, and G. Pannetier and V. Massari for their help in the data management. The authors are also grateful to the patients and the clinical teams for their commitment to the French HIV-HBV Cohort Study.

Sponsorship: This work was supported by fundings from SIDACTION, ANRS (Agence Nationale de Recherche sur le Sida) and promoted by IMEA (Institut de Médecine et d'Epidémiologie Appliquée).