Potential conflict of interest: Nothing to report.
The contents of the manuscript represent the authors' personal opinions and do not necessarily reflect any position of the US Food and Drug Administration.
See Editorial on Page 875
The purpose of this report is to illustrate the US Food and Drug Administration's rationale for approving response-guided therapy (RGT) for telaprevir (TVR) in combination with pegylated interferon-α and ribavirin (P/R) for the treatment of adults with genotype 1 chronic hepatitis C who were prior relapsers. RGT was prospectively evaluated in two registration trials of treatment-naïve subjects. In these studies, RGT allowed subjects who achieved undetectable hepatitis C virus RNA from weeks 4 and 12, known as extended rapid virologic response (eRVR), to stop all treatments at 24 weeks. A patient without eRVR received an additional 36 weeks of P/R after 12 weeks of a TVR triple regimen (total of 48 weeks). However, RGT in prior P/R relapsers was not prospectively evaluated. Empirical cross-trial data indicated high sustained virologic response rates (>90%) in prior relapsers achieving eRVR, irrespective of P/R duration (24 or 48 weeks). Further analyses demonstrated that interferon responsiveness does not change in P/R-experienced subjects with a second round of P/R. The comparability in interferon responsiveness across treatment courses allowed us to bridge data between treatment-naïve and P/R-experienced subjects to support the approval of RGT in prior relapse subjects. (HEPATOLOGY 2013)
On May 23, 2011, the US Food and Drug Administration (FDA) approved telaprevir (TVR) for use in combination with pegylated interferon-α and ribavirin (P/R) for the treatment of genotype 1 chronic hepatitis C in adults who are either P/R treatment-naïve or -experienced.1 TVR is a hepatitis C virus (HCV) NS3/4A protease inhibitor and represents a new class of small molecules that directly targets HCV replication.2 The approved dosing regimen for TVR is 750 mg given three times daily for 12 weeks (T12) in combination with P/R. After reviewing the TVR new drug application, which contained three phase 3 studies, the FDA approved response-guided therapy (RGT) for treatment-naïve patients and prior P/R relapsers.3
The RGT schema was prospectively evaluated in two registration trials of treatment-naïve subjects: Study 108 (ADVANCE Trial: 8 versus 12 weeks of TVR)4 and Study 111 (ILLUMINATE Trial: 12 weeks of TVR and 24 versus 48 weeks of P/R treatment in subjects achieving extended rapid virologic response as defined below).5 In these TVR studies, the RGT schema allowed subjects who achieved undetectable HCV RNA from week 4 through 12, known as extended rapid virologic response (eRVR), to stop all treatments at week 24.6 A patient with eRVR received 12 weeks of a triple drug regimen (TVR+P/R) followed by an additional 12 weeks of P/R treatment. A patient without eRVR received an additional 36 weeks of P/R after 12 weeks of a TVR triple regimen (total of 48 weeks).
RGT was not prospectively evaluated in prior P/R-relapse subjects. Here, P/R-relapse subjects refer to patients who had undetectable HCV RNA at the end of P/R treatment, but did not achieve sustained virologic response (SVR). In the registration trial Study C216 (REALIZE Trial)7 evaluating P/R-experienced subjects, all subjects received 12 weeks of TVR concurrently with 48 weeks of P/R (T12/PR48) irrespective of eRVR status. However, two phase 2 studies, Study 106 (PROVE3 trial8) and Study 107 investigated 12 weeks of TVR with 24 weeks of P/R (T12/PR24) in a small number of prior relapse subjects.
A critical decision for the FDA was whether to approve RGT or the full course (48 weeks) of P/R treatment for prior P/R-relapse patients. Approval of the full course of P/R treatment would have been straightforward as it was the regimen studied in the registration trials. In contrast, the choice to approve RGT would depend on a small number of subjects in early studies and a cross-trial comparison. To assist in the evaluation of RGT in this patient population, the FDA review team identified an important link between the treatment-naïve and P/R-experienced HCV populations.9 Conceptually, if the P/R-experienced population is considered to represent a subpopulation within the treatment-naïve population, then data could be bridged to inform dosing recommendations between populations. Therefore, the FDA used data in the phase 2 studies and the knowledge derived from bridging the treatment-naïve and P/R-experienced populations to support dosing recommendations for prior relapsers who achieve eRVR.
Bridging knowledge from reliable resources to provide clinical evidence of effectiveness and to derive dosing recommendations is acceptable from a regulatory perspective.10 There is precedent for the FDA using bridging studies and supportive data to approve a dosing regimen that has not been directly evaluated in phase 3 trials. For example, oxcarbazepine and topiramate monotherapy for the treatment of partial seizures was approved in pediatric patients without a controlled clinical trial.11, 12 Similar examples include clevidipine for short-term intravenous management of blood pressure in adults, canakinumab for the treatment of cryoprin-associated periodic syndrome in adults and pediatric patients aged 4 years and above, levofloxacin for the treatment of postexposure inhalational anthrax, and argatroban for prophylaxis or treatment of thrombosis in pediatrics.13-17
This report summarizes the rationale to support TVR RGT dosing recommendations in prior P/R relapsers.
Data From Phase II and III Studies
In the TVR development program, phase 3 Studies 108 (ADVANCE),4 111 (ILLUMINATE),5 and C216 (REALIZE)7 formed the basis of approval in treatment-naïve and P/R-experienced patients (Fig. 1). The primary endpoint of all phase 3 trials was the proportion of subjects achieving SVR, demonstrated by having undetectable HCV RNA 24 weeks after last planned dose of study drug (SVR24).
In the ADVANCE trial, 1,088 previously untreated subjects were randomized to an initial regimen of either 8 or 12 weeks of telaprevir (T8 or T12) in combination with response-guided P/R therapy based on eRVR status (24 or 48 weeks) or a control regimen of P/R for 48 weeks. The main objective of the ADVANCE trial was to establish the superiority of a TVR+P/R regimen compared with a standard P/R regimen in a treatment-naïve population.
The ILLUMINATE trial prospectively studied RGT in 322 treatment-naïve subjects who achieved eRVR after 12 weeks of triple drug regimen (TVR+ P/R) by randomized comparison between 24 (T12/PR24) and 48 (T12/PR48) weeks of P/R treatment. The main objective of the ILLUMINATE trial was to establish noninferiority between T12/PR24 and T12/PR48 regimens.
The REALIZE trial was designed to establish the superiority of a T12/PR48 regimen to standard P/R treatment in P/R-experienced subjects. A total of 662 subjects who received prior P/R treatment (stratified by prior relapsers, prior partial responders, and prior null responders18) were randomized either to 12 weeks of TVR in combination with 48 weeks of P/R (T12/PR48), a 4-week delayed start (DS) of 12 weeks of TVR in combination with 48 weeks of P/R [T12(DS)/PR48], or a standard regimen of 48 weeks of P/R. This study did not evaluate RGT.
Studies 106 and 107 were phase 2 studies in which T12/PR24 was investigated in P/R-experienced subjects (including some subjects who achieved eRVR status). However, these studies were exploratory and did not provide a direct within-trial comparison to T12/PR48 treatment in this population.
DS, delayed start; EOS, end of study; eRVR, extended rapid virologic response; FDA, US Food and Drug Administration; HCV, hepatitis C virus; P/R, pegylated interferon-α and ribavirin; RGT, response-guided therapy; SVR, sustained virologic response; TVR, telaprevir.
Evidence to Support Regulatory Decisions
Data from phase 2 exploratory studies (Studies 106 and 107) included a total of 52 prior relapse subjects who achieved eRVR and were treated with a total of 24 weeks of treatment (T12/PR24). These data were retrospectively compared with prior relapse subjects with eRVR treated for 48 weeks [T12/PR48 or T12(DS)/PR48] in the REALIZE trial. As shown in Fig. 2, SVR rates were high in prior relapsers (>90%), irrespective of P/R duration (24 or 48 weeks). The SVR rates were also high after T24/PR48 treatment in a small number of subjects in Study 106.
Given the limitations of the data and the caveats associated with a cross-trial comparison without a prospective hypothesis, there was a need for additional supportive evidence to facilitate the approval of T12/PR24 in prior relapse subjects.
The supportive evidence was derived from bridging data between treatment-naïve and P/R-experienced subjects. For treatment-naïve subjects, the distribution of change in HCV RNA at week 4 classified by the end of study (EOS) outcome in the P/R48 arm in the ADVANCE trial was compared with that of P/R-experienced subjects classified by the prior (baseline) response to interferon in the combined P/R48 and T12(DS)/PR48 arms in the REALIZE trial. As shown in Fig. 3 and Table 1, interferon responsiveness (defined by mean decline in HCV RNA at week 4) was −3.5 log10 IU/mL in responders, followed by −2.8 log10 IU/mL in relapsers, −1.6 log10 IU/mL in partial responders, and −0.6 log10 IU/mL in null responders. Notably, interferon responsiveness in subgroups of P/R-experienced subjects in the REALIZE trial was well-matched to the corresponding subgroups of treatment-naïve subjects (prior relapsers, −2.7 log10 IU/mL; prior partial, −1.4 log10 IU/mL; and prior null, −1.0 log10 IU/mL). We have proposed that the characteristics of these populations were similar because of the apparent lack of resistance to P/R.9 Based on this interferon responsiveness data, it was apparent that potential future relapsers, if treated with P/R, could be considered a subgroup within a treatment-naïve population. Toward that end, the following arguments were considered in order to bridge the data and support our regulatory decision:
1In the ADVANCE trial, the EOS outcome after P/R treatment was 44% responders, 18% relapsers, 19% partial responders, and 9% null responders. These outcomes were similar to historical data after P/R treatment.
2In the ADVANCE trial, 60% of subjects treated with TVR and P/R achieved eRVR and were eligible for RGT with a shorter duration regimen of T12/PR24. Of those, approximately 90% achieved SVR.
3We can reasonably assume that of the 60% of subjects who achieved eRVR while receiving TVR and P/R, 44% would have been P/R responders according to the proportions of outcomes in treatment-naïve subjects who achieved eRVR. The remaining 16% subjects who achieved eRVR with TVR and P/R likely represent those who potentially would not respond to P/R if treated with P/R alone. Among the nonresponder categories, the interferon responsiveness of relapsers is most similar to responders; therefore, it is most likely that the additional subjects achieving eRVR and ultimate successful treatment would have been potential relapsers if receiving P/R alone.
Table 1. Comparison of Interferon Responsiveness at Week 4 Between Treatment-Naïve Subjects (ADVANCE Trial) and P/R-Experienced Subjects (REALIZE Trial)
Data represent the change from baseline in HCV RNA (log10) and are expressed as the mean (median [interquartile range]).
−2.8 (−2.9 [−3.6 to −1.5])
−2.7 (−2.4 [−3.6 to −1.7])
−1.6 (−1.4 [−2.0 to −0.9])
−1.5 (−1.3 [−1.8 to −0.8])
−0.6 (−0.5 [−0.9 to −0.4])
−1.0 (−0.9 [−1.2 to −0.5])
Therefore, a treatment that was deemed appropriate for future relapsers should be appropriate for prior relapsers.
To further support the bridging between treatment-naïve and P/R-experienced subjects, the internal consistency between the ADVANCE and REALIZE trials was checked through external validation. The observed SVR rate in the P/R-experienced subjects from the REALIZE trial was applied to derive the expected SVR rate for treatment-naïve subjects after T12/PR in the ADVANCE trial. As shown in Table 2, the predicted overall SVR (74%) for treatment-naïve subjects using data from P/R-experienced subjects matched closely to the observed overall SVR rate in the ADVANCE trial (75%). Using the same approach, the predicted SVR rate for T12/PR treatment in treatment-naïve subjects with eRVR in the ADVANCE trial was derived to be 92%, which also matched closely the observed 88-92% SVR rate with T12/PR in the ADVANCE trial and the ILLUMINATE trial (data not shown). These external validation tests further supported the assumption that P/R-experienced subjects are represented within subgroups of treatment-naïve subjects.
Table 2. SVR Rates in P/R-Experienced Subjects Treated With TVR Triple Therapy (REALIZE Trial) Inform About SVR Rates in Treatment-Naïve Subjects Treated With TVR (ADVANCE Trial)
Naïve Population After Potential P/R Treatment
Distribution of EOS Response (%) to P/R Treatment in Treatment-Naïve Subjects*
% SVR in P/R Responders† and Experienced Subjects‡ Receiving TVR
Derived % SVR Among Treatment-Naïve Subjects Receiving TVR
Data in treatment-naïve subjects with the P/R (control) arm from the ADVANCE trial. Subjects not accounted for in the ADVANCE trial are those who did not complete the study for the following reasons: adverse event, death, lost to follow-up, withdrawal of consent, and other (e.g., noncompliance with study drug, other noncompliance, refused further treatment). These numbers are comparable between P/R and T12/PR treatment.
The SVR rate is assumed to be 100% with TVR triple therapy for P/R responders because they can already achieve SVR with the double therapy.
Observed in P/R-experienced subjects with the TVR arm from the REALIZE trial.
Similar to the observed 75% TVR response rate in treatment-naïve subjects from the ADVANCE trial.
Similar arguments, which support RGT for prior relapsers, could also be applied to prior partial and null responders For example; baseline prior partial and null responders are similar in characteristics to that of “potential” prior and null responders. If this is the case, these subgroups of P/R-experienced subjects should also be eligible to receive the RGT dosing regimen. There were limited clinical data from Study 107, which included 22 prior partial and 3 null P/R responders with eRVR who received T12/PR24. The cross-study comparison indicated that SVR rates were 62-77% in prior partial responders and 62-71% in prior null responders who achieved eRVR, irrespective of P/R duration (24 or 48 weeks), suggesting the RGT regimen might also be suitable in prior partial and null responders. Study 106 did not differentiate null responders and partial responders at baseline. For the overall group of nonresponders, the SVR rate was 68%, irrespective of P/R duration (24 or 48 weeks). Although the FDA review team decided that these limited data did not rise to the extent of adequate evidence for a regulatory decision, the data were helpful in forming hypotheses. From the risk-benefit viewpoint, given poor interferon responsiveness for these populations, it is prudent to maximize therapy using knowledge of previous response to P/R. Shortened treatment duration is attractive to reduce side effects due to P/R treatment, but these subjects have a 1 in 3 chance of failing treatment with TVR+P/R, even after achieving eRVR. Our data suggest that the chance of treatment failure is likely to be similar despite a longer treatment duration, and that RGT would be a good option if the virus is cleared early. It may be possible to prospectively study the RGT hypotheses in prior partial and null responders; however, the study size and enrollment time for such a trial is likely to be impractical given the rapid pace of HCV drug development.
The bridging analyses indicate that from a clinical decision perspective, previous P/R exposure should not alter new treatment decisions involving interferon-based triple therapy, because the interferon responsiveness to a second course of interferon is comparable to the first course.9 This is unlike human immunodeficiency virus, in which treatment failure frequently leads to resistance and lower efficacy during subsequent courses of treatment with drugs in the same class. Instead, the outcomes of HCV interferon-based therapy are primarily affected by baseline factors (e.g., fibrosis, age, race, HCV genotypes, and baseline viral load). Theoretically, one would expect that retreatment would result in the same interferon responsiveness and 0% SVR in previously P/R failed subjects. However, that is an unrealistic assumption as several patient-related factors (e.g., motivation, adherence, tolerability, etc.) cannot be replicated, and the categories for prior partial and null responders are arbitrary. For example, a 1.9-log reduction in viral load by week 12 would be a null responder, but a 2.1-log change on a retreatment would be a partial responder. Although the response cannot be exactly replicated in retreatment, several studies have shown that previous relapsers achieve SVR at a higher frequency than partial responders and null responders.19-22 This is additional evidence supporting the hypotheses that patients do not lose interferon responsiveness upon treatment. Therefore, our primary focus should be on whether the previous P/R treatment had resulted in a lower response or caused resistance, rather than an exactly same response in a retreatment. Current clinical experience has not identified a resistance to interferon-based therapy.
Under the Food, Drug, and Cosmetics Act, the FDA approves drugs based on substantial evidence of safety and effectiveness. The risk/benefit assessment for RGT in certain populations must also consider the adequacy of efficacy data in some subgroups (e.g., blacks/African Americans and patients with cirrhosis). Overall, the SVR rate with telaprevir in treatment-experienced Blacks/African Americans is similar to that in Caucasians, although the number of black patients in the trials was small (≈4% of treatment-experienced subjects). For treatment-naïve patients with cirrhosis, the SVR rate (≈70%: 19/27 from the ADVANCE and ILLUMINATE trials) was numerically lower in patients achieving eRVR with a 24-week treatment (T12/PR24) compared with the 92% (11/12 from ILLUMINATE trial) SVR rate in those with a 48-week treatment (T12/PR48). This observation, though based on a small number of subjects, led to labeling language that treatment-naïve patients with cirrhosis may benefit from a 48-week total treatment, even for those patients who had achieved eRVR.1 For prior relapsers with cirrhosis, SVR was achieved in all six patients (from Studies 106 and 107) who achieved eRVR with a 24-week treatment compared with the 90% (31/42 from ADVANCE trial) SVR rate in those with a 48-week treatment; however, the number of subjects was small. The bridging analysis demonstrates that a treatment that was deemed appropriate for potential relapsers (represented within the treatment-naïve population) should also be appropriate for prior relapsers. With very limited information, a 48-week total treatment may be better for both potential relapsers (in the treatment-naïve population) and prior relapsers with cirrhosis who have similar interferon-response characteristics.
In conclusion, FDA-approved telaprevir RGT for prior relapsers based on knowledge from phase 2 and phase 3 studies and bridging data between treatment-naïve and P/R-experienced subjects. Further, data from the REALIZE trial independently reproduced data from the ADVANCE trial and supported the assertion that data from treatment-experienced subjects may serve as supportive evidence in treatment-naïve subjects.
Data for these analyses were submitted to the US Food and Drug Administration by Vertex as part of a new drug application.