CROI: Choice of nucleoside analogue in co-infection: tenofovir associated with better response to HCV therapy

In people with HIV/HCV co-infection, drug interactions between nucleoside analogues and ribavirin may lead to poorer responses to hepatitis C (HCV) treatment. Two poster presentations of retrospective data by Spanish study teams at the Fifteenth Conference on Retroviruses and Opportunistic Infections showed that co-infected patients who were receiving tenofovir had the best chances of sustained response to HCV therapy.

The first poster, by José Mira and team, described a retrospective multicentre study of 256 patients, all of whom started first-line HCV therapy with pegylated interferon (peg-IFN) and ribavirin while receiving ART for HIV infection. All patients were receiving a three-drug ART regimen including one protease inhibitor (PI) or one NNRTI, plus a dual-nucleoside backbone of either: abacavir and lamivudine (3TC), tenofovir and 3TC, or tenofovir and FTC.

The study group data was drawn from 256 patients in fifteen Spanish hospitals. Median age was 42, 78% were male, median baseline HCV viral load was 5.9 log₁₀ and median CD4 cell count was 473 cells/mm³.

All were treated with a combination of subcutaneous peg-IFN alpha-2a (180 µg/week) or alpha-2b (1.5 µg/kg/week) plus oral ribavirin (RBV) at 600 to 1500 mg/day. Treatment duration was 48 weeks for patients with HCV genotype 1 or 4, and 24 or 48 weeks at the treating physicians' discretion for genotypes 2 and 3. Temporary discontinuations or dose reductions of peg-IFN and/or RBV were also at the physicians' discretion. Such dose reductions were made for 10% to 12% of patients, with no significant difference between those on tenofovir or abacavir.

In an intention-to-treat analysis, significantly better responses to HCV therapy were seen with tenofovir than with abacavir. Lower SVR in patients taking abacavir was most pronounced and significant in those who received lower ribavirin doses; at higher doses, the trend remained but became statistically insignificant. SVR was as follows:

(*not significant)

In multivariate analysis, a tenofovir-containing regimen independently predicted SVR (adjusted odds ratio [OR], 2.6; 95% confidence interval [CI], 1.05 to 6.9; p=.03). SVR was also greater for those with HCV genotype 2 or 3 (OR, 8.9; 95% CI, 4 to 20; p<.001), baseline LDL cholesterol levels ≥100 mg/dL (OR, 3.06; 95% CI, 1.4 to 6.7; p=.004), lower baseline plasma HCV RNA load (OR, 1.85 per log₁₀; 95% CI, 1.1 to 3.1; p=.016) and undetectable baseline HIV viral load (OR, 3.5; 95% CI, 1.01 to 12.5; p=.003).

The other poster, by González-Garcia and team, described another, larger retrospective multicentre study. This study included data on 719 patients from 35 sites who initiated first-line HCV therapy between January 2003 and November 2005. In this case, however, their concomitant ART therapy consisted of dual NRTIs plus an NNRTI or a PI, or triple-nucleoside therapy including abacavir.

The analysis in this case compared the SVR between two groups: the 238 tenofovir-receiving patients, and the 481 not receiving tenofovir. The TDF group were taking TDF plus 3TC or FTC; the non-TDF group included those on AZT+3TC (n=265), d4T+3TC (n=164), or ABC+3TC (n=52). Patients on triple-nucleoside therapy (AZT+3TC+ABC) were counted in the AZT+3TC group (number not reported). Those receiving didanosine (ddI), or TDF in combination with either AZT, d4T or ABC, were excluded.

Baseline characteristics were similar to the first group: median age was 41 years and 75% were male. Other baseline characteristics were well-matched between the TDF and non-TDF groups except for somewhat lower CD4 cell counts (535 vs. 601 cells/mm³; p=.003), exposure to more antiretroviral regimens (7.2 vs. 5.7, p<.001), and more lipodystrophy (30.3% vs. 22.9%, p=.033) in the TDF group.

Ribavirin doses were reduced more often in those who did not receive TDF than in those that did (12.8% vs. 19.5%, p=.03), especially in those who received AZT (23.2%, p=.003)

The raw intention-to-treat (ITT) analysis found no differences in SVR between the TDF-treated and non-TDF-treated groups (45% vs. 39%, p=.12). Multivariate analysis adjusted for the five factors which predicted response in univariate analysis: HCV genotype, alcohol intake > 50 g/day, and baseline HCV viral load (<500,000 IU/mL), HIV viral load (<50 copies/mL), and AST/ALT ratio. Results from this analysis were as follows:

(*not significant)

Both studies, therefore, were in agreement that tenofovir was associated with better sustained virologic responses to first-line HCV treatment. Previous reports, however, have found that poorer SVR may result in patients on abacavir-containing regimens - an issue these posters did not entirely clarify.

The Mira study, drawing on previous findings of poorer HCV response in patients taking abacavir, specifically addressed whether AZT might have had a confounding effect in this scenario. By excluding patients on AZT, Mira found that the preferential effect of tenofovir over abacavir (in the presence of 3TC or FTC) persisted.

The González-Garcia study reached slightly different conclusions: that tenofovir was associated with improved HCV SVR, but that AZT – not abacavir – was associated with poorer tolerability and efficacy. There was not, in this study, a significantly different outcome for patients on dual-nucleoside therapy including abacavir. In this study, an unspecified number of patients on abacavir-containing triple-nucleoside therapy were grouped with the other patients on AZT, not abacavir; and change of ART was considered as treatment failure. It might have been illuminating to see the triple-nucleoside patients included in the abacavir arm for analysis.

Both studies were also retrospective analyses. Further, prospective studies may be needed to fully distinguish the impact of AZT and abacavir on pegylated interferon/ribavirin treatment for HCV.