The advent of triple therapy (TT) with first-generation protease inhibitors boceprevir (BOC) and telaprevir (TVR) in addition to pegylated interferon and ribavirin (PEG-IFN/RBV) has resulted in a significant improvement in the sustained virological response (SVR) rate and potentially in life years gained compared to dual therapy (DT), when treating naïve or treatment-experienced patients with genotype 1 (G1) chronic hepatitis C (CHC). This benefit is partly offset by the increased complexity of treatment, and the increased costs and risks of therapy, making it necessary to optimize the indications for TT. Naïve patients with mild fibrosis and the IL28B CC polymorphism and/or with a rapid virological response (RVR) to DT can still benefit from DT, while TT is preferable in all others. Phase 3 trials have clearly shown that a 1 log10 decrease in HCVRNA after 4 weeks of DT associated with a favourable IL28B genotype and a low stage of fibrosis, and a pattern of previous response to DT in treatment-experienced patients are the strongest predictors of an SVR to TT. Moreover, an extended rapid virological response (eRVR) is associated with an SVR rate >90%, so that the overall duration of treatment can be shortened in a high proportion of patients. Further efforts to optimize the current TT regimens both by increasing efficacy and improving tolerance are still needed. Most important, in the future, treatment can probably be personalized based on data from post-marketing surveillance of TT providing information about patient groups that were underrepresented in phase 3 studies such as those with cirrhosis.
In the past few years, management of naïve and treatment-experienced genotype1 chronic hepatits C (G1 CHC) patients has changed because of the development of triple therapy (TT), which combines first-generation direct acting antiviral agents (DAAs) and PEG-IFN/RBV. Although there are numerous DAAs in the pipeline, the two currently approved nonstructural serine (NS3/4) protease inhibitors, telaprevir (TVR) and boceprevir (BOC) will be the mainstay of therapy in the next 2–3 years . Triple therapy with TVR or BOC with PEG-IFN/RBV significantly increases the rate of SVR not only in naïve [2-4], but also in treatment-experienced [5-7] G1 CHC patients. In particular phase 3 trials of TT with BOC (SPRINT-2)  or TVR (ADVANCE and ILLUMINATE) [3, 4] showed SVR rates ranging from 63 to 75% in previously untreated G1 CHC patients emphasizing an increase in the SVR rate of about 25% compared to DT. Similarly, phase 3 trials combining TT with BOC (RESPOND-2 and PROVIDE) [5, 6] or TVR (REALIZE)  in previously treated G1 CHC patients divided according to the previous virological response to DT showed that SVR rates progressively increased from 75 to 86% in relapse patients (RR), 52 to 57% in partial responders (PAR) (PAR-HCVRNA drop>2log10 at week 12, but never undetectable) and 31 to 37% in non-responders (NR) (NR-HCVRNA drop<2log10at week 12-) resulting in an increase in SVR of about 55, 40 and 30% respectively compared to DT.
Although these results are encouraging, the use of TT in clinical practice must still be optimized, first by identifying patients in whom the SVR with TT will be clearly better than that with DT, but also by evaluating host and viral factors that can predict the response to TT at baseline and during the early phase of treatment. In this way, the best expected results could be obtained and risks would be reduced by shortening treatment in early responders, and stopping therapy in NR as early as possible. Figure 1 summarizes the key points assessed in this review of the optimization of TT in G1 CHC patients.
Triple therapy, dual therapy, or ‘informed deferral’?
Even if the use of TT as the new standard of therapy in naïve and experienced G1 CHC patients is cost-effective compared to DT [8, 9], it requires investment of medical and economic resources, and will also result in significant impairment in patient quality of life during treatment because of more severe complications from adding PI to DT. Thus to avoid or reduce lack of resources and useless side effects in patients, patients who will benefit from TT should be identified.
Predictors of sustained virological response to dual therapy
In treatment-naïve G1 CHC patients, various results clearly show that both baseline and on-treatment variables can accurately identify patients who are likely to achieve an SVR with DT. Specifically, IL28B genetic status is the best baseline predictor of SVR. In particular, the probability of an SVR to DT in G1 CHC patients without severe liver fibrosis and with the IL28B CC genotype is more than 80% . Therefore, the similar SVR rates observed in previously untreated IL28B CC patients who received TVR (90%) or BOC-based (80%) TT suggest that TT should be avoided in these subgroups of patients, because the only advantage of TT compared to DT is a potential shortening of treatment duration despite higher costs and more side effects .
Rapid virological response
Many studies in the literature including observational studies, clinical trials and meta-analyses, have conclusively shown that a rapid virological response (RVR), is a highly relevant predictor of SVR in G1 CHC patients. The SVR rate in these subjects is more than 80%, and the duration of DT can sometimes be shortened . Thus, treatment with TT does not benefit this group of patients. Finally, it should be noted that although a baseline evaluation of the IL28B genotype allows a pre-treatment allocation of patients to DT or BOC/TVR-based TT, if an RVR-guided strategy is to be used, DT must be begun in all G1 treatment naïve patients and HCVRNA must be evaluated after 4 weeks. The absence of a RVR in these cases provides relevant data about the sensitivity of the patient to DT. Triple therapy can then be continued with BOC-based strategies because a lead-in phase of 4 weeks of DT is part of BOC-based treatments, while it has not been tested in TVR trials in naïve patients.
Although IL28B status and RVR identify naïve G1 CHC patients in whom the SVR rate with DT is similar to that with TT, no easy-to-treat sub-groups of previously treated patients have been identified. Thus, retreatment of RR or NR patients to DT is not recommended [11, 12], and a rational use of TT is suggested. Specifically, to optimize treatment in experienced patients, patients requiring TT who have a high probability of achieving an SVR (i.e. cirrhotic RR) must be identified compared to those with a low probability of achieving an SVR who should wait for more effective drugs (i.e. –NR- with minimal liver damage).
The notion of ‘informed deferral’
Finally, it should be emphasized that even if assigning patients to DT, TT, or deferred treatment with new antiviral agents is of clinical and economic interest, this process has drawbacks. For example a recent editorial introduced the new notion of ‘Informed Deferral’ , which stated that the management strategy must be discussed individually with the patient. Specifically, the pros and cons of each therapeutic approach should be discussed, in particular the limitations of each strategy because of errors in the diagnosis of liver damage and in predicting the progression of fibrosis, the changes in the patient profile which may affect access and tolerance to future therapies, and to potential new drugs.
Optimization of triple therapy: the role of eRVR and of stopping rules
If the on-treatment response is evaluated in naïve and experienced G1 CHC patients who receive TT, treatment duration can be adapted in case of virological response, and therapy can be stopped in case of non-response, thus optimizing treatment and reducing the side effects and costs.
Extended rapid virological response
Randomized controlled trials (RCT) evaluating BOC and TVR in G1 CHC patients [2-4] introduced the notion of an extended rapid virological response (eRVR), defined as undetectable HCV-RNA at week 8 maintained through to week 24 for BOC, and as undetectable HCV-RNA at week 4 maintained through to week 12 for TVR. Achieving an eRVR is a key point in the management of naïve G1 CHC patients who receive TT. The ADVANCE study  showed that patients who reach an eRVR obtain an SVR in 89% of cases compared to 54% of patients who do not. Similarly, in the SPRINT-2 study , nearly 90% of patients who achieved an eRVR achieved an SVR compared to 60% in those who did not. The ADVANCE  and SPRINT-2  studies also clearly showed that, in patients without cirrhosis who achieve an eRVR, treatment can be shortened from a fixed duration (12 weeks of TT followed by 36 weeks of DT for TVR; or 4 weeks of DT, followed by 24 weeks of TT, followed by 20 weeks of TT for BOC) to a shorter duration (12 weeks of TT followed by 12 weeks of DT for TVR; or 4 weeks of DT, followed by 24 weeks of TT for BOC), without reducing the probability of an SVR. Finally, most of the patients with an eRVR who received TVR-based TT (54% and 63% in ADVANCE and ILLUMINATE respectively) [3, 4], and a high proportion of those treated with BOC-based regimens (44% of patients in SPRINT-2) were naïve subjects (). Interestingly, in BOC-related therapy, SPRINT-1 study showed that a lead-in phase increased the rate of patients who achieved an eRVR (62% vs. 38%) .
Overall, these results clearly show that naïve patients with G1 CHC who undergo TT frequently achieve an eRVR and that the probability of achieving an SVR is greater than 90% in these patients with a shortened treatment schedule. Thus, the FDA and the EMA have authorized a shortened therapeutic regimen in G1 CHC patients without cirrhosis who achieve an eRVR during TT [15-18]. The clinical relevance of an eRVR has also been investigated in previously treated patients who underwent TT. In particular, the RESPOND-2 study  compared a fixed duration (4 weeks of DT, followed by 44 weeks of BOC-based TT), to response-guided therapy (RGT) in patients who underwent 4 weeks of DT followed by 32 weeks of BOC-based TT, with, in patients without an eRVR, a further 12 weeks of DT. The SVR rate in the fixed schedule was slightly better than that in the RGT. These results together with those in the REALIZE study , in which all previously treated patients received fixed duration therapy were reassessed by the EMA and FDA [15-18], which recommended a fixed duration regimen of BOC-based therapies in all non-cirrhotic patients including 4 weeks of DT, followed by 32 weeks of TT, and another 12 weeks of DT based on retrospective studies. In contrast, and although RCTs were not performed, the same regulatory authorities [15-18], recommended treating both RR patients without cirrhosis and naïve patients by RGT TVR-based therapy, with fixed duration therapy in all other previously treated patients. This is particularly important in clinical practice as an eRVR was obtained in 68–76% of RR patients treated with TVR-based TT.
These results show that accurate evaluation of eRVR is crucial to optimize therapy in G1 CHC patients who undergo TT. Recent retrospective analysis of RCTs with BOC and TVR has evaluated the clinical relevance of on-treatment detectable but below the lower limit (detectable/BLOQ) HCVRNA, and undetectable HCVRNA . Phase 3 RCTs of TVR and BOC assessed HCVRNA with the Roche COBAS TaqMan HCV 2.0 assay [2-7]. The lower limit of plasma HCVRNA quantification (LLOQ) was considered to be 25 international units per ml (IU/ml) in these trials, and the limit of detection (LOD) was considered to be 9.3–10 IU/ml. The latter was used to assess eligibility for a shortened treatment regimen. Based on these RCTs it is not clear if the difference between undetectable and detectable/BLOQ HCVRNA is relevant for clinical practice. Retrospective analysis of BOC RCTs shows that at the recommended RGT decision time points (week 8 for BOC and week 4 for TVR), the presence of detectable/BLOQ HCVRNA was observed in about 20%–30% of cases and was associated with a reduction in SVR of approximately 20% compared to patients with undetectable HCVRNA . Moreover, data from retrospective analysis of phase 2 studies (PROVE 1 and 2 for TVR; SPRINT 1 for BOC) showed that patients with detectable/BLOQ HCVRNA at week 4 with TVR, and at week 8 with BOC had lower SVR rates when they were treated with a shortened than with a fixed duration treatment regimen (38% vs. 75% for BOC; 40% vs. 57% for TVR) . These results suggest that the presence of detectable/BLOQ HCVRNA is associated with lower SVR rates than with undetectable HCVRNA, and TT treatment should not be shortened in these cases.
To optimize therapy, in addition to identifying patients who can benefit from a short therapeutic regimen, patients who do not respond to TT must also be identified to reduce costs and risks from both side effects and virological resistances if SVR is improbable. Phase 3 studies of TVR have shown stopping rules with HCV RNA levels >1000 IU at week 4, and a <2log10 decrease at week 12 (stop all drugs) in naïve patients [3, 4]; and HCV RNA levels >100 IU at weeks 4, 6, and 8, or a <2log10 decrease at week 12 in previously treated patients . Nevertheless the EMA recommended stopping therapy in both naïve and experienced G1 CHC patients treated with TVR if HCVRNA >1000 at week 4 (stop only TVR) or at week 12 (stop all drugs) based on reassessment of a few patients with HCVRNA at weeks 4 and 12 between 100 and 1000 IU who achieved SVR . On the other hand, AISF Italian guidelines took a more conservative approach in experienced patients, and suggested a cut-off of 100 IU . For BOC, the SPRINT-2  and RESPOND-2  studies recommended stopping therapy if HCVRNA was undetectable at week 24 in naïve patients and at week 12 in previously treated G1 CHC patients. EMA also modified stopping rules for TVR in both naïve and experienced patients, and suggests stopping therapy if HCVRNA >100 IU at week 12 . This recommendation was also the subject of a recent scientific study to identify uniform stopping rules in all BOC-treated patients. In accordance with EMA recommendations, this study identified >100 IU at week 12 as the best stopping rule  for both naïve and experienced patients. The benefit of this new cut-off is that a greater proportion of naïve NR-TT patients can stop therapy, thus reducing the development of viral resistances without reducing overall SVR. In addition based on an analysis of 6 NR patients, this cut-off, which is slightly less restrictive than that in the REALIZE study, should recover certain patients who can achieve an SVR, i.e. patients with detectable/BLOQ HCVRNA.
Predictors of response to triple therapy
The strongest predictor of SVR in patients who undergo TT is achieving an eRVR, especially naïve patients. Nevertheless, for TT as well as for DT, it is important to identify baseline and early on-treatment predictors of response (Table 1).
Table 1. Factors associated with likelihood of sustained virological response to triple therapy in both naive and experienced patients with genotype 1 chronic hepatitis C
Treatment-naive genotype 1 chronic hepatitis C patients
Experienced genotype 1 chronic hepatitis C patients
≥1Log10 HCVRNA drop after 4 weeks of DT (only for BOC)
Lower BMI (only for BOC)
Absence of cirrhosis
Absence of cirrhosis
≥1Log10 HCVRNA drop after 4 weeks of DT (only for BOC)
Baseline predictors of response: liver fibrosis and IL28B genotype
In naïve patients, SPRINT-2  and ADVANCE  for BOC and TVR respectively showed that the severity of liver fibrosis and IL28B single nucleotide polymorphisms (SNP) also affect treatment outcome in patients who undergo triple therapy. In particular, SPRINT-2  showed that the SVR was higher in patients without severe fibrosis (67% vs. 52% in the BOC-fixed duration arm) and in patients with IL28B CC (80%) compared to CT (71%) and TT (59%) patients. Thus, it appears that TT is only more effective than DT in genotypes CT and TT, but not in CC. (). Similarly, lower SVR rates were found in patients with severe fibrosis in the ADVANCE study  than in those without (62% vs. 78%), also showing a progressive increase in the SVR rate from IL28B CC (90%), to CT (71%) and TT (73%) patients. In the IL28B genotype, the results of triple therapy were only better than DT in CT and TT patients . In addition, other factors that negatively affected SVR were genotype subtype 1a and Black ethnic origin for both BOC and TVR-based therapies.
A retrospective study was recently published evaluating SVR predictors in naïve patients who underwent BOC-based TT. This study also included an evaluation of the IL28B SNP, which was not evaluated in the phase 3 protocols. Interestingly, this retrospective study identified low baseline HCV viral load (≤400 000 IU), IL28B rs 12979860 CC genotype, the absence of cirrhosis, HCV genotype 1b, and non-Black ethnic origin as independent positive predictors of SVR . In addition, Poordad et al.  evaluated the results of RCTs on the role of a 1log10 decrease in HCVRNA after 4 weeks of DT as a predictor of SVR in BOC-treated patients. They included this variable in the model of SVR. Interestingly, the author observed that the decrease in HCVRNA remained significantly associated with SVR, independent of the IL28B genotype UNCLEAR . It is necessary to determine whether metabolic factors, which may influence whether an SVR is achieved with DT are also negative predictors of SVR in TT in naïve patients. For example, the abovementioned retrospective analysis of SPRINT-2 showed that obesity negatively influenced both the 4 week virological response and an SVR (in the model including a 4 week decrease in HCVRNA) . In contrast, a retrospective analysis  of the study by Marcellin et al.  assessing the efficacy of TVR-based TT in naïve G1 CHC patients with TVR administered two to three times daily, using PEG-IFN alfa2a or 2b, did not show any association between virological response and metabolic parameters . Specifically, the authors did not find any relationship between HOMA, the expression of insulin resistance, the decrease in HCVRNA at 4 weeks, on-therapy and end-of-therapy response, and SVR . In contrast, higher LDL serum levels, and the absence of severe fibrosis, which are well-known predictors of SVR in DT, were associated with a higher probability of SVR in TVR-based TT . Nevertheless, the results of that study seem to suggest that metabolic factors do not significantly influence SVR in TVR-based TT. Although these results could support a strong direct antiviral activity for PI, further data are needed before conclusions can be drawn. The results by Serfaty et al.  were not included in the SVR model. They were obtained in a cohort of young patients (mean age 44 years), with a low prevalence of diabetes (3%), most of whom were normal weight (61%) without data on IL28B SNPs, and in whom liver steatosis was only evaluated in 41% of patients. All these variables could affect SVT to DT.
The importance of the previous response in treatment-experienced patients
Data from the RCTs RESPOND-2  and REALIZE  in previously treated G1 CHC patients, clearly show that the pattern of the previous response to DT strongly affects the probability of achieving an SVR after TT, with a progressive increase in SVR rates from NR, to PAR and then RR. The key role of the virological response profile in treatment experienced patients, as a predictor of SVR, was also recently confirmed in a study on TVR, in whom patients in TVR control arms were retreated with TVR-based TT . Because the previous response to DT has a significant impact on SVR rates in case of retreatment with TT and because of the limited availability of data in patients in clinical practice, on-treatment predictors of response to TT would be very useful. Results from the lead-in arm of REALIZE , and from RESPOND-2  evaluating all previously treated G1 CHC patients, showed significantly lower SVR rates in patients with <1log10 decrease in HCV RNA after 4 weeks of DT (33% for TVR, and 34% for BOC), compared with those with a >1log decrease (82% for TVR, and 79% for BOC). Interestingly, the trials also confirmed these differences in patients stratified according to the previous response, even if higher SVR rates were observed in RR, followed by PR and NR. Specifically, the REALIZE study  showed that the percentage of SVR in patients with a 1log10IU/ml decrease at week 4 was 54% vs. 15% in NR, 59% vs. 56% in PR and 88% vs. 62% in RR respectively. Similarly, the RESPOND-2 study  showed that the SVR rate in patients with a decrease in HCVRNA at week 4 was 61% vs. 37% in PR, and 81% vs. 37% in RR.
The importance of fibrosis
When evaluating the predictors of response to TT in previously treated G1 CHC patients, the severity of liver fibrosis must be considered in addition to the previous response and the on-treatment response. In particular, in the REALIZE study , the SVR in patients who received TT was 74% in those with F0–F2 fibrosis, 66% in those with F3 fibrosis, and 47% in those with cirrhosis. However, the impact of fibrosis on the SVR was greater in PAR (F0–F2: 72%, F3: 56%, F4: 34%) and in NR (F0–F2: 41%, F3: 39%, F4: 14%), while in RR, the stage of fibrosis had no impact on SVR (F0–F2: 86%, F3: 85%, F4: 84%). Similar data were reported in the RESPOND-2 RCT, in which SVR rates in the BOC-fixed duration arm were 75% and 83% in RR F0-F2 and F3-F4 patients respectively, and 55% and 46% in PR F0-F2 and F3-F4 patients respectively . However, these data must be interpreted with caution because of the very few patients with severe fibrosis included in the study. Finally, while IL28B genotype does not affect SVR in previously treated G1 CHC patients [15, 16, 22], the sub-type of the viral genotype must also be assessed in this group of patients receiving TT. In fact, a lower SVR rate (59%) in patients with genotype 1a was found in the REALIZE study  compared with those with genotype 1b (71%). In the retrospective study by Poordad et al. of the results of the RESPOND-2 RCTs  in previously treated patients, after correction for genetic (IL28B), clinical-metabolic, viral and histological variables, the pattern of previous response remained independently associated with SVR after BOC-base TT . The 1log10 decrease in HCVRNA at week 4 associated with the pattern of response to previous DT was also an independent predictor .
The retrospective studies by Poordad et al. of the SPRINT-2 and RESPOND-2 studies for predictors of response to TT also showed that IL28B CC genotype, low HCV viral load, low BMI, and the absence of cirrhosis in naïve patients, and the pattern of previous response in treatment experienced patients, were the strongest predictors of a 1log10decrease in HCVRNA at week 4 (97% ‘good response’ in naïve IL28B CC, and 94% in experienced IL28B CC) .
Overall, these results show that well-known predictors of SVR to DT are also predictors for TT, but to a lesser extent: (i) IL28B genotype and liver fibrosis in naïve patients, and the pattern of previous response in experienced patients, are the strongest predictors of SVR; (ii) a 1log10 decrease in HCVRNA after 4 weeks of DT, mostly observed in IL28B CC patients is the strongest predictor of SVR. Finally, it is worth stressing that all data from retrospective studies should be interpreted with caution as these studies were performed in subgroups of the entire study cohort (62% from SPRINT-2 and 66% from RESPOND-2 for the Poordad study), and the prevalence of variables found associated with response rates, i.e. low viral load and cirrhosis, were low.
Also, it is important to understand the role of IL28B polymorphism in treatment response with functional studies . Furthermore, in most future trials with direct acting antivirals , the association between the IL28B polymorphism and SVR will be analysed, to identify whether this predictive factor remains.
Results of RCTs and retrospective studies of TVR or BOC-based therapies in naïve or experienced G1 CHC patients show that SVR is improved and life years and quality-adjusted life years are gained with TT compared to DT. However, due to the complexity of administration and the high rate of toxicity of these new therapeutic strategies in clinical practice, efficacy must be optimized to reduce costs and risks. We suggest the following:
Clearly identify G1 CHC patients who can still benefit from DT (IL28BCC and/or RVR patients), those who need effective TT (RR patients), and those who should wait for new, more potent drugs (NR patients with no significant liver damage).
Clearly identify baseline and on-treatment factors associated with the probability of response to TT, such as IL28B genotype, severity of liver fibrosis and a decrease in HCVRNA at week 4 in naïve patients, and the pattern of previous response to therapy and a decrease in HCVRNA at week 4 in treatment experienced patients.
Establish the therapeutic schedule to be applied to individual patients based on baseline and on-treatment SVR predictors to achieve optimal results and reduce treatment duration, costs and risks.
In the future, further optimization of treatment and ‘personalized’ TT may be possible, based on the effect of other variables such as vitamin D levels on SVR. Moreover, data from post-marketing evaluation of TT should provide information on patient groups that were underrepresented in phase 3 studies such as those with cirrhosis, making truly personalized treatment possible in these patients.
The authors declare no conflict of interest and have no disclosure.