Response to Antivirals for HCV in HIV/HCV Co-infection
Response to Antivirals for HCV in HIV/HCV Co-infection
Among 619 HIV/HCV co-infected patients treated with PEG-IFN and ribavirin in the VA healthcare system between 2002 and 2009, SVR was achieved in 16.7% (82/491) of patients infected with genotype 1 HCV, 44% (51/116) of patients infected with genotypes 2 or 3 and 8% (1/12) of patients infected with genotype 4. Treatment characteristics that independently predicted SVR were the starting dose of ribavirin and use of erythropoietin in genotype 1-infected patients and use of erythropoietin in genotype 2-infected and genotype 3-infected patients.
Pivotal clinical trials using PEG-IFN and ribavirin for HIV/HCV co-infected patients reported SVR rates of 14–29% in patients with genotype 1 and 44–73% in patients with genotype 2 or 3 HCV. SVR rates in the VA setting are therefore at the lower range of those reported in clinical trials. Other observational studies have demonstrated that SVR rates achieved in 'real world' clinical settings are lower than those reported in clinical trials. For example, among 226 HCV/HIV co-infected patients of all genotypes who received combination treatment with interferon alpha or PEG-IFN and ribavirin at the Johns Hopkins HIV clinic, only 36 (15.9%) achieved SVR, although rates were not categorised by genotype or receipt of pegylated vs. regular interferon. Eligibility criteria in clinical trials lead to highly selected study populations. It is likely that multiple viral, patient, provider and facility factors account for the differences in SVR rates between clinical trials and observational studies. It is also likely that observational studies such as ours and the study from Johns Hopkins reflect the SVR rates in general clinical practice in the United States more closely than clinical trials.
Patients included in our study had a high prevalence of predictors associated with low SVR, particularly Black race and early discontinuation of treatment. Black patients, who constituted the majority of genotype 1-infected patients in this study, have lower response rates to interferon-based therapies, partly due to higher prevalence of the IL28B genotypes C/T or T/T. We observed higher SVR rates among White patients (21.1% for genotype 1 and 47.1% for genotype 2 and 3) – although even these were lower than those reported in clinical trials. Only 32% of patients received >36 weeks of treatment and only 20% received ≥48 weeks of treatment, which is the recommended duration. Lower SVR rates could possibly be explained by high rates of premature treatment discontinuation in VA-based patients, as has been observed in studies of HCV mono-infected populations. It is possible that higher SVR rates than those we describe here are observed in clinical settings that have a lower prevalence of these adverse predictors of SVR.
Our study confirms that SVR rates achieved by PEG-IFN and ribavirin are lower in HIV/HCV co-infected patients than in HCV mono-infected patients. HCV mono-infected patients treated in the VA between 2001 and 2007 with PEG-IFN and ribavirin were reported to have SVR rates of 26% for genotype 1, 62% for genotype 2 and 52% for genotype 3 (when patients without a post-treatment HCV RNA were counted as nonresponders).
The most important independent predictor of SVR across all genotypes was use of erythropoietin, which is commonly used off-label to treat treatment-induced anaemia. Although associations between erythropoietin and SVR have been described in HCV mono-infected patients, the associations we observed in HIV/HCV co-infected patients were greater in magnitude. For example, in HCV mono-infected patients in VA care, use of erythropoietin was associated with higher SVR with an AOR of 1.27, whereas among HIV/HCV co-infected patients, we observed an AOR of ≈3 (Table 3). Also, the association between erythropoietin and SVR persisted whether the decline in haemoglobin occurred early or late in anti-viral treatment (Table 4), whereas in HCV mono-infected patients, erythropoietin seemed to have a profound effect on SVR only in early-onset anaemia. The greater impact of erythropoietin in HIV/HCV co-infected patients may be because treatment-induced anaemia is a more important problem in HIV/HCV co-infected patients than in HCV mono-infected patients due to bone marrow suppression by HIV or certain antiretrovirals, such as zidovudine. Zidovudine was used in 31% of patients during their anti-viral treatment in this time period (2002–2009) and was, in fact, significantly associated with lower SVR among genotype 1 patients in multivariate analysis (AOR 0.41, 95% CI 0.2–0.9), confirming the findings of a recent study from Spain. Zidovudine is much less commonly used now than in the study period.
In HCV mono-infected patients, treatment-induced anaemia was 'paradoxically' shown to be associated with higher SVR rates, possibly by being a surrogate marker for high plasma ribavirin concentration that may, in turn, lead to higher SVR rates. As treatment-induced anaemia is associated with use of erythropoietin and may also be associated with SVR, it can be a potential confounder of the association between erythropoietin and SVR. However, we carefully adjusted for both drop in haemoglobin or minimum haemoglobin as well as ribavirin dose and other potential confounders and found that the association between erythropoietin and SVR persisted (Table 4). Interestingly, we found no association between treatment-induced anaemia and SVR in HIV/HCV co-infected patients as was described in mono-infected patients. This suggests that, in HIV/HCV co-infected patients, the negative effects of treatment-induced anaemia (early discontinuation, dose reduction in PEG-IFN and ribavirin) may counterbalance any positive effects of high plasma ribavirin concentrations.
Early trials of therapy in HIV/HCV co-infected patients used relatively low doses of ribavirin due to concerns for anaemia. Subsequent studies used higher, 'weight-based' ribavirin regimens irrespective of genotype, such as 1000 mg/day if body weight was <75 kg and 1200 mg/day if >75 kg in the PRESCO trial. The package insert for Copegus (ribavirin) states that all HIV/HCV co-infected patients should receive 800 mg/day. A recent RCT ('PARADIGM') found that SVR was slightly higher (60/275 or 22%) in genotype 1 HIV/HCV co-infected patients randomised to 1000/1200 mg/day of weight-based ribavirin than those randomised to 800 mg/day (26/135 or 19%) in combination with PEG-IFN alpha 2a, but the difference was not statistically significant. The 2009 AASLD guidelines recommend that the starting dose of ribavirin for genotypes 2 and 3 should be 800 mg/day, while for genotypes 1 and 4, it should be 1000 mg/day (body weight <75 kg) or 1200 mg/day (body weight ≥75 kg). Our findings support the need for this higher ribavirin dose in genotype 1 HCV, as patients who received the recommended dose were twice as likely to achieve SVR as patients who received less than the recommended dose (AOR 2.1, 95% CI 1.1–3.8). Our findings also support the AASLD recommendation for 800 mg/day in genotype 2 or 3 HCV/HIV co-infected patients, as patients who received a higher dose did not have a significantly higher SVR rate.
We found that patients treated by mid-level providers had an equivalent chance of SVR as patients treated by physicians. Similarly, patients treated by gastroenterologists had equivalent SVR rates as those treated by infectious disease specialists and primary care providers. These observations support findings by Arora et al. that mid-level providers and clinicians who are not gastroenterologists can provide equivalent anti-viral treatment, especially if given in a supportive environment. Thus, expansion of anti-viral treatment to appropriately trained mid-level providers and nongastroenterologists seems to be a reasonable way to increase access to treatment.
There are several limitations to the current study. First, as an observational study, we can only conclude an association, and not causation, between specific treatment characteristics (such as erythropoietin use or ribavirin dose) and SVR. Second, the population included a majority of Black patients, which is uncommon in the United States and Europe. Lastly, the small sample size for genotype 2 and 3 patients reduced our power to identify significant associations with SVR.
Hepatitis C virus NS3/4A PIs (boceprevir and telaprevir) used in combination with PEG-IFN and ribavirin is now the standard of care for HCV genotype 1 mono-infected patients. However, the safety, efficacy and tolerability of such triple therapy regimens have not been established in HIV/HCV co-infected patients. Furthermore, significant drug–drug interactions exist with HIV antiretroviral medications. The findings from phase 2 studies are encouraging, with a reported SVR of 74% in patients treated with telaprevir. The low rates of SVR in the current and past studies highlight the great need for the development of direct-acting anti-viral regimens suitable for HIV/HCV co-infected patients, particularly those with fewer side effects and drug–drug interactions. Furthermore, erythropoiesis-stimulating agents will probably continue to be important adjuncts during HCV treatment if direct-acting anti-virals are approved for HIV/HCV co-infected patients, and may take on an even larger role due to high rates of anaemia with boceprevir and telaprevir.
Discussion
Among 619 HIV/HCV co-infected patients treated with PEG-IFN and ribavirin in the VA healthcare system between 2002 and 2009, SVR was achieved in 16.7% (82/491) of patients infected with genotype 1 HCV, 44% (51/116) of patients infected with genotypes 2 or 3 and 8% (1/12) of patients infected with genotype 4. Treatment characteristics that independently predicted SVR were the starting dose of ribavirin and use of erythropoietin in genotype 1-infected patients and use of erythropoietin in genotype 2-infected and genotype 3-infected patients.
Pivotal clinical trials using PEG-IFN and ribavirin for HIV/HCV co-infected patients reported SVR rates of 14–29% in patients with genotype 1 and 44–73% in patients with genotype 2 or 3 HCV. SVR rates in the VA setting are therefore at the lower range of those reported in clinical trials. Other observational studies have demonstrated that SVR rates achieved in 'real world' clinical settings are lower than those reported in clinical trials. For example, among 226 HCV/HIV co-infected patients of all genotypes who received combination treatment with interferon alpha or PEG-IFN and ribavirin at the Johns Hopkins HIV clinic, only 36 (15.9%) achieved SVR, although rates were not categorised by genotype or receipt of pegylated vs. regular interferon. Eligibility criteria in clinical trials lead to highly selected study populations. It is likely that multiple viral, patient, provider and facility factors account for the differences in SVR rates between clinical trials and observational studies. It is also likely that observational studies such as ours and the study from Johns Hopkins reflect the SVR rates in general clinical practice in the United States more closely than clinical trials.
Patients included in our study had a high prevalence of predictors associated with low SVR, particularly Black race and early discontinuation of treatment. Black patients, who constituted the majority of genotype 1-infected patients in this study, have lower response rates to interferon-based therapies, partly due to higher prevalence of the IL28B genotypes C/T or T/T. We observed higher SVR rates among White patients (21.1% for genotype 1 and 47.1% for genotype 2 and 3) – although even these were lower than those reported in clinical trials. Only 32% of patients received >36 weeks of treatment and only 20% received ≥48 weeks of treatment, which is the recommended duration. Lower SVR rates could possibly be explained by high rates of premature treatment discontinuation in VA-based patients, as has been observed in studies of HCV mono-infected populations. It is possible that higher SVR rates than those we describe here are observed in clinical settings that have a lower prevalence of these adverse predictors of SVR.
Our study confirms that SVR rates achieved by PEG-IFN and ribavirin are lower in HIV/HCV co-infected patients than in HCV mono-infected patients. HCV mono-infected patients treated in the VA between 2001 and 2007 with PEG-IFN and ribavirin were reported to have SVR rates of 26% for genotype 1, 62% for genotype 2 and 52% for genotype 3 (when patients without a post-treatment HCV RNA were counted as nonresponders).
The most important independent predictor of SVR across all genotypes was use of erythropoietin, which is commonly used off-label to treat treatment-induced anaemia. Although associations between erythropoietin and SVR have been described in HCV mono-infected patients, the associations we observed in HIV/HCV co-infected patients were greater in magnitude. For example, in HCV mono-infected patients in VA care, use of erythropoietin was associated with higher SVR with an AOR of 1.27, whereas among HIV/HCV co-infected patients, we observed an AOR of ≈3 (Table 3). Also, the association between erythropoietin and SVR persisted whether the decline in haemoglobin occurred early or late in anti-viral treatment (Table 4), whereas in HCV mono-infected patients, erythropoietin seemed to have a profound effect on SVR only in early-onset anaemia. The greater impact of erythropoietin in HIV/HCV co-infected patients may be because treatment-induced anaemia is a more important problem in HIV/HCV co-infected patients than in HCV mono-infected patients due to bone marrow suppression by HIV or certain antiretrovirals, such as zidovudine. Zidovudine was used in 31% of patients during their anti-viral treatment in this time period (2002–2009) and was, in fact, significantly associated with lower SVR among genotype 1 patients in multivariate analysis (AOR 0.41, 95% CI 0.2–0.9), confirming the findings of a recent study from Spain. Zidovudine is much less commonly used now than in the study period.
In HCV mono-infected patients, treatment-induced anaemia was 'paradoxically' shown to be associated with higher SVR rates, possibly by being a surrogate marker for high plasma ribavirin concentration that may, in turn, lead to higher SVR rates. As treatment-induced anaemia is associated with use of erythropoietin and may also be associated with SVR, it can be a potential confounder of the association between erythropoietin and SVR. However, we carefully adjusted for both drop in haemoglobin or minimum haemoglobin as well as ribavirin dose and other potential confounders and found that the association between erythropoietin and SVR persisted (Table 4). Interestingly, we found no association between treatment-induced anaemia and SVR in HIV/HCV co-infected patients as was described in mono-infected patients. This suggests that, in HIV/HCV co-infected patients, the negative effects of treatment-induced anaemia (early discontinuation, dose reduction in PEG-IFN and ribavirin) may counterbalance any positive effects of high plasma ribavirin concentrations.
Early trials of therapy in HIV/HCV co-infected patients used relatively low doses of ribavirin due to concerns for anaemia. Subsequent studies used higher, 'weight-based' ribavirin regimens irrespective of genotype, such as 1000 mg/day if body weight was <75 kg and 1200 mg/day if >75 kg in the PRESCO trial. The package insert for Copegus (ribavirin) states that all HIV/HCV co-infected patients should receive 800 mg/day. A recent RCT ('PARADIGM') found that SVR was slightly higher (60/275 or 22%) in genotype 1 HIV/HCV co-infected patients randomised to 1000/1200 mg/day of weight-based ribavirin than those randomised to 800 mg/day (26/135 or 19%) in combination with PEG-IFN alpha 2a, but the difference was not statistically significant. The 2009 AASLD guidelines recommend that the starting dose of ribavirin for genotypes 2 and 3 should be 800 mg/day, while for genotypes 1 and 4, it should be 1000 mg/day (body weight <75 kg) or 1200 mg/day (body weight ≥75 kg). Our findings support the need for this higher ribavirin dose in genotype 1 HCV, as patients who received the recommended dose were twice as likely to achieve SVR as patients who received less than the recommended dose (AOR 2.1, 95% CI 1.1–3.8). Our findings also support the AASLD recommendation for 800 mg/day in genotype 2 or 3 HCV/HIV co-infected patients, as patients who received a higher dose did not have a significantly higher SVR rate.
We found that patients treated by mid-level providers had an equivalent chance of SVR as patients treated by physicians. Similarly, patients treated by gastroenterologists had equivalent SVR rates as those treated by infectious disease specialists and primary care providers. These observations support findings by Arora et al. that mid-level providers and clinicians who are not gastroenterologists can provide equivalent anti-viral treatment, especially if given in a supportive environment. Thus, expansion of anti-viral treatment to appropriately trained mid-level providers and nongastroenterologists seems to be a reasonable way to increase access to treatment.
There are several limitations to the current study. First, as an observational study, we can only conclude an association, and not causation, between specific treatment characteristics (such as erythropoietin use or ribavirin dose) and SVR. Second, the population included a majority of Black patients, which is uncommon in the United States and Europe. Lastly, the small sample size for genotype 2 and 3 patients reduced our power to identify significant associations with SVR.
Hepatitis C virus NS3/4A PIs (boceprevir and telaprevir) used in combination with PEG-IFN and ribavirin is now the standard of care for HCV genotype 1 mono-infected patients. However, the safety, efficacy and tolerability of such triple therapy regimens have not been established in HIV/HCV co-infected patients. Furthermore, significant drug–drug interactions exist with HIV antiretroviral medications. The findings from phase 2 studies are encouraging, with a reported SVR of 74% in patients treated with telaprevir. The low rates of SVR in the current and past studies highlight the great need for the development of direct-acting anti-viral regimens suitable for HIV/HCV co-infected patients, particularly those with fewer side effects and drug–drug interactions. Furthermore, erythropoiesis-stimulating agents will probably continue to be important adjuncts during HCV treatment if direct-acting anti-virals are approved for HIV/HCV co-infected patients, and may take on an even larger role due to high rates of anaemia with boceprevir and telaprevir.
