Sofosbuvir for hepatitis C works well despite multiple negative predictive factors

Hepatitis C treatment using sofosbuvir (Sovaldi) is highly effective even for people with multiple factors traditionally associated with poor response. Having four or more negative predictive factors, however, raises the risk of post-treatment relapse, according to a report at the 49th annual meeting of the European Association for the Study of the Liver (EASL) held recently in London.

The advent of direct-acting antiviral agents (DAAs) has begun to revolutionise treatment for chronic hepatitis C. But several aspects of treatment – including the optimal duration of therapy and the best regimens for patients traditionally considered difficult to treat – are not yet fully understood.

Graham Foster of Queen Mary University of London and colleagues looked at the influence of various host and virus factors traditionally associated with poor response to interferon-based therapy.

Interferon, formerly the mainstay of treatment for chronic hepatitis C, works by stimulating the body's own immune response against the virus. Direct-acting antivirals directly interfere with different steps of the hepatitis C virus (HCV) lifecycle. Studies have shown that some of the factors associated with poor response to interferon may not have as much effect when using DAAs.

Traditional negative predictive factors include:

• Viral genotype: HCV genotypes 1 and 4 have been considered harder to treat than genotypes 2 or 3.
• Viral load: high pre-treatment HCV RNA (>800,000 IU/ml) is associated with poorer response.
• Sex: men do not respond as well as women.
• Age: older people do not respond as well as younger people.
• Race/ethnicity: people of African descent (and Hispanic/Latino people in some studies) do not respond as well as Caucasian or Asian people.
• Body weight: heavier people, especially those classified as obese (body mass index >30), do not respond as well.
• IL28B status: people who carry the TT or CT genetic variations do not respond as well as those with the favourable CC pattern.
• Liver damage: people with advanced fibrosis (stage F3) and especially cirrhosis (F4) do not response as those with absent to moderate fibrosis (F0-F2).
• HIV/HCV co-infection: people with both HIV and HCV do not respond as well as those with HCV alone.
• Prior treatment: People who previously showed modest or minimal response to prior interferon-based therapy (partial and null responders) do not respond as well as previously untreated (treatment-naive) people or those who previously responded but relapsed after finishing treatment.

Sofosbuvir leads to high response rates for most patient populations. However, a small proportion of people do not achieve sustained virological response, or undetectable HCV RNA at 12 weeks post-treatment (SVR12), which is considered a cure. In most cases this is due to relapse, or viral rebound after treatment is finished.

Foster and his team performed a retrospective meta-analysis of data from phase 2 and 3 clinical trials of sofosbuvir, used with either pegylated interferon and ribavirin (ATOMIC, NEUTRINO) or ribavirin alone (FISSION, POSITRON, FUSION, VALENCE).

The researchers first performed a univariate analysis to see the effect of each factor by itself, then did a multivariate analysis to look at how these factors interact. Finally, they calculated cure rates for people with increasing numbers of negative predictive factors.

The analysis included 339 treatment-naive patients with genotype 1 chronic hepatitis C treated with sofosbuvir plus pegylated interferon and ribavirin for 12 weeks, 285 treatment-experienced and treatment-naive patients with genotype 2 treated with sofosbuvir plus ribavirin for 12 weeks, and 247 treatment-naive and treatment-experienced patients with genotype 3 treated with sofosbuvir plus ribavirin for 24 weeks.

Overall, about 60% of participants were men and about two-thirds were age 50 or older. Fifteen per cent of people with genotype 1, 8% of people with genotype 2 and no patients with genotype 3 were black. Approximately 60 to 70% had unfavourable IL28B gene variants. About 16% of people with genotype 1 and 2 had cirrhosis, rising to 24% for those with genotype 3. One-third with genotype 2 and 58% with genotype 3 were treatment-experienced (previously treated people with genotype 1 were not included). More than three-quarters had high baseline HCV viral load.

In a univariate regression analysis of the combined data set, factors found to be significantly associated with post-treatment relapse (p < 0.05) were: liver cirrhosis (odds ratio [OR] 4.3, or more than 4-fold higher risk), baseline HCV RNA >800,000 IU/ml (OR 3.9), male sex (OR 3.5), weight over 75kg (OR 3.2), IL28B non-CC (OR 2.8), prior treatment (OR 2.8) and age over 50 years (OR 1.9). The effects of black race, Hispanic ethnicity and elevated baseline ALT were not statistically significant.

Surprisingly, while having HCV genotype 3 rather than 2 was a significant risk factor (OR 2.5), the differences between genotype 1 vs 2 and between 3 vs 1 did not reach statistical significance (p = 0.18 and 0.06, respectively). Nor did HCV subtype 1a vs 1b turn out to be significant in a more restricted analysis of  people with genotype 1 (p = 0.14).

In a multivariate regression analysis of the combined data set, only six factors were independently associated with relapse: high baseline viral load (OR 4.7), cirrhosis (OR 4.0), IL28B non-CC (OR 3.4), weight over 75kg (OR 2.5), prior treatment (OR 2.3) and male sex (OR 2.3).

SVR12 rates were 100% for people with zero or one negative predictive factor, above 99% for those with two factors and 94% for those with three factors. After this, efficacy dropped off, falling to 88% for those with four negative factors, 68% for those with five factors and 57% for those with six factors.

Cure rates were 90% or higher for all genotype groups with zero to three negative predictive factors. For those with four or more factors, the effect was most pronounced among people with HCV genotypes 1 or 3 (in fact, no genotype 1 patients had zero risk factors).

"Sofosbuvir-based regimens were highly effective, even in patients with a combination of multiple negative factors," the researchers concluded. "SVR12 rates were comparatively lower in patients with five or six of the negative predictors."

"Patients need to have a cluster of poor predictors to be at risk for failure," Graham Foster explained. He suggested that these factors "may help guide us in an interferon-free world" when deciding on what regimens to use and for how long.

Notably, this analysis is specific to sofosbuvir, an HCV polymerase inhibitor. Other trials indicate that the negative predictive factors which remain relevant may differ among the various DAAs.

Although HIV and HCV co-infection was not included in this analysis, other studies have shown that people with HIV can respond to interferon-free regimens as well as people with HCV alone, leading some experts to suggest that people with co-infection should no longer be separated out as a 'special' or difficult-to-treat population.