FDA Perspective on Sofosbuvir Therapy for Patients With Chronic Hepatitis C Virus Genotype 1 Infection Who Did Not Respond to Treatment With Pegylated Interferon and Ribavirin
2014; Elsevier BV; Volume: 147; Issue: 6 Linguagem: Inglês
10.1053/j.gastro.2014.10.027
ISSN1528-0012
AutoresPoonam Mishra, Jeffry Florian, Karen Qi, Wen Zeng, Lisa K. Naeger, Eric Donaldson, Sarah Connelly, Jules O’Rear, Dionne Price, Jeffrey S. Murray, Debra Birnkrant,
Tópico(s)Chronic Lymphocytic Leukemia Research
ResumoSee related commentary, Ghany and Liang, on page 1201. See related commentary, Ghany and Liang, on page 1201. On December 6, 2013, the U.S. Food and Drug Administration (FDA) approved sofosbuvir (SOF) for the treatment of chronic hepatitis C (CHC) viral infection as a component of an antiviral treatment regimen. SOF is an inhibitor of the hepatitis C virus (HCV) NS5B RNA-dependent RNA polymerase and inhibits HCV replication. SOF was designated a breakthrough therapy because preliminary clinical evidence indicated substantial improvement over available therapies for the treatment of CHC infection.1FDA Presentations at the October 25, 2013, Meeting of the Antiviral Drugs Advisory Committee. Available at http://www.fda.gov/AdvisoryCommittees/CommitteesMeetingMaterials/Drugs/AntiviralDrugsAdvisoryCommittee/ucm387736.htm.Google Scholar It was the first drug with a breakthrough therapy designation approved for treatment of CHC. One of the key questions the FDA faced during the review process was whether there was sufficient evidence to support use of SOF in genotype 1 (GT1) patients who have failed previous therapy with pegylated interferon and ribavirin (PEG+RBV). This subgroup of GT1 patients with prior documented failure to a PEG+RBV regimen was not specifically included in the registrational trial of SOF in combination with PEG+RBV. Noting that GT1 treatment-naïve subjects treated with SOF and PEG+RBV experienced a high sustained virologic response (SVR) rate 12 weeks after completing therapy (SVR12) coupled with an improved safety profile and shorter treatment duration, the FDA review team estimated the SVR rate in the GT1 treatment-experienced population using available data from registrational trials, published literature, and historical data. This perspective summarizes 3 scientific rationales that support the efficacy of SOF in combination with PEG+RBV for treatment of patients who did not achieve SVR (nonresponders) after a prior course of PEG+RBV. Clinical trials evaluating treatments for CHC viral infection have generally empirically categorized patients as treatment naïve or treatment experienced based upon prior virologic response to a PEG+RBV regimen. HCV GT1 patients who failed prior treatment with PEG+RBV were not specifically studied in the SOF clinical development program. FDA reviewers recognized that this population was in need of new therapies with improved efficacy and safety profiles. Once a new drug application is submitted by an Applicant (sponsor) to the FDA for marketing approval in the United States, an FDA review team evaluates whether the data demonstrate that the drug is safe and effective for its proposed use. The review team performs independent analyses of the trial results to determine whether the submitted data support the proposed treatment regimen or whether additional information is needed to make a regulatory decision. The FDA reviewers may perform additional analyses to evaluate a drug's effectiveness in patient subpopulations not analyzed in the original submission. The FDA conducted 3 exploratory analyses to estimate the SVR12 rate in prior PEG+RBV nonresponders using existing trial data in the treatment-naïve population studied in NEUTRINO. NEUTRINO was an open-label, single-arm, Phase III trial that evaluated 12 weeks of treatment with SOF in combination with peginterferon alfa 2a (PEG) and RBV in treatment-naïve subjects with GT 1, 4, 5, or 6 HCV infection compared with a prespecified historical control (Figure 1). The primary objectives of the trial were to determine the efficacy of SOF+PEG+RBV as measured by the proportion of subjects with SVR12 and to evaluate the safety and tolerability of SOF+PEG+RBV. The majority of the subjects (89%) in this trial had GT1 HCV infection. In the primary efficacy analyses, 89% of subjects achieved SVR12 in the NEUTRINO trial. Previous FDA analyses have indicated that ≤50% of treatment-naïve GT1 subjects fail PEG+RBV therapy and are subsequently categorized as PEG+RBV nonresponders.1FDA Presentations at the October 25, 2013, Meeting of the Antiviral Drugs Advisory Committee. Available at http://www.fda.gov/AdvisoryCommittees/CommitteesMeetingMaterials/Drugs/AntiviralDrugsAdvisoryCommittee/ucm387736.htm.Google Scholar, 2Liu J. Florian J. Birnkrant D. et al.Interferon responsiveness does not change in treatment-experienced hepatitis C Subjects: implications for drug development and clinical decisions.Clin Infect Dis. 2012; 55: 639-644Crossref PubMed Scopus (7) Google Scholar, 3Liu J. Jadhav P.R. Amur S. et al.Response-guided telaprevir therapy in prior relapsers? The role of bridging data from treatment-naïve and experienced subjects.Hepatology. 2013; 57: 897-902Crossref PubMed Scopus (14) Google Scholar Therefore, treatment-naïve response rates >50% are evidence that a direct-acting antiviral (DAA)+PEG+RBV regimen will be effective in patients who would potentially be classified as PEG+RBV nonresponders. The observed high overall SVR rate in HCV GT1 treatment-naïve subjects from NEUTRINO led the review team to explore whether the data may support use of SOF+PEG+RBV for 12 weeks in patients who have failed a prior PEG+RBV regimen. An analysis was done to predict the SVR rate in HCV GT1 treatment-experienced subjects based on the observed SVR rate for PEG+RBV treatment in historical trials. The historical SVR rate for PEG+RBV in treatment-naïve subjects ranged from 40% to 50%. As a conservative assessment, it was assumed that 50% of the HCV GT1 treatment-naïve subjects in NEUTRINO could be PEG+RBV treatment failures and that all subjects who failed to respond to SOF+PEG+RBV (11%) in NEUTRINO would have also been PEG+RBV treatment failures. Based on these assumptions, it implies that the remaining 39% (ie, 50% – 11%) of subjects who would have failed PEG+RBV alone responded to the SOF+PEG+RBV treatment in NEUTRINO. Using this approach, we predicted an SVR rate of 78% (39%/50%) for SOF+PEG+RBV (12 weeks of treatment) in the HCV GT 1 prior PEG+RBV nonresponders (Figure 2). In addition, this exploratory analysis assumes that PEG+RBV responsiveness remains relatively constant across first and subsequent PEG+RBV–based regimens. This assumption is supported by a lack of identified resistance to PEG treatment, similarity between on-treatment responses after initial or subsequent courses of PEG+RBV treatment, and previous FDA analyses bridging observations between treatment-naïve and prior PEG+RBV treatment failures.2Liu J. Florian J. Birnkrant D. et al.Interferon responsiveness does not change in treatment-experienced hepatitis C Subjects: implications for drug development and clinical decisions.Clin Infect Dis. 2012; 55: 639-644Crossref PubMed Scopus (7) Google Scholar, 3Liu J. Jadhav P.R. Amur S. et al.Response-guided telaprevir therapy in prior relapsers? The role of bridging data from treatment-naïve and experienced subjects.Hepatology. 2013; 57: 897-902Crossref PubMed Scopus (14) Google Scholar, 4Florian J. Jadhav P.R. Amur S. et al.Boceprevir dosing for late responders and null responders: the role of bridging data between treatment-naïve and -experienced subjects.Hepatology. 2013; 57: 903-907Crossref PubMed Scopus (13) Google Scholar Finally, the analysis assumes baseline characteristics are similar between subjects in the historical PEG+RBV studies and in NEUTRINO. The review team identified that NEUTRINO consisted of an older patient population and more subjects with HCV GT 1a, cirrhosis, and higher baseline HCV RNA compared with historical PEG+RBV studies (Table 1). These observations suggest that the patient demographics represented within NEUTRINO were similar or even slightly less favorable than patient characteristics observed in previous treatment-naïve trials, making the analysis more conservative.Table 1Baseline Characteristics for Hepatitis C Virus (HCV) Genotype 1 Treatment-Naïve Subjects Who Received Pegylated Interferon and Ribavirin in Historical Studies and NEUTRINO StudyCharacteristicHistorical StudiesaIncluding pegylated interferon and ribavirin treatment arms in ACHIEVE-1, IDEAL, PROVE 1, PROVE 2, ADVANCE, SPRINT-II studies. (n = 3374)NEUTRINO (n = 292)Age Mean (standard deviation)47 (9)52 (10) Median (Q1, Q3)49 (43, 53)54 (48, 59)Male (%)5965Black (%)1517Genotype 1a (%)6277Cirrhosis (%)518Baseline HCV viral load ≥800,000 IU/mL (%)7683a Including pegylated interferon and ribavirin treatment arms in ACHIEVE-1, IDEAL, PROVE 1, PROVE 2, ADVANCE, SPRINT-II studies. Open table in a new tab A review of published literature and HCV treatment guidelines identified several baseline factors that predispose to PEG+RBV treatment failure in HCV GT1 treatment-naive patients: High baseline viral load, fibrosis score of F3 or F4, presence of steatosis, pretreatment fasting glucose of >5.6 mmol/L, high baseline alanine aminotransferase levels, age >40 years, and African-American race.5Shiffman M.L. Di Bisceglie A.M. Lindsay K.L. et al.HALT-C Trial GroupPeginterferon alfa-2a and ribavirin in patients with chronic hepatitis C who have failed prior treatment.Gastroenterology. 2004; 126: 1015-1023Abstract Full Text Full Text PDF PubMed Scopus (472) Google Scholar, 6McHutchison J.G. Lawitz E.J. Shiffman M.L. et al.IDEAL Study TeamPeginterferon alfa-2b or alfa-2a with ribavirin for treatment of hepatitis C infection.N Engl J Med. 2009; 361: 580-593Crossref PubMed Scopus (1133) Google Scholar, 7Ghany M.G. Strader D.B. Thomas D.L. et al.Diagnosis, management, and treatment of hepatitis C: an update.Hepatology. 2009; 49: 1335-1374Crossref PubMed Scopus (2669) Google Scholar, 8Ghany M.G. Nelson D.R. Strader D.B. et al.An update on treatment of genotype 1 chronic hepatitis C virus infection: 2011 practice guideline by the American Association for the Study of Liver Diseases.Hepatology. 2011; 54: 1433-1444Crossref PubMed Scopus (984) Google Scholar, 9American Association for the Study of Liver Diseases (AASLD), and Infectious Diseases Society of America (IDSA), in collaboration with International Antiviral Society-USA (IAS–USA). Recommendations for testing, managing, and treating hepatitis C 2014. Available at http://www.hcvguidelines.org.Google Scholar Recently published analyses have also identified that interleukin (IL)-28B is associated with response to PEG+RBV treatment, and may partly explain the impact of race.10Ge D. Fellay J. Thompson A.J. et al.Genetic variation in IL28B predicts hepatitis C treatment-induced viral clearance.Nature. 2009; 461: 399-401Crossref PubMed Scopus (3145) Google Scholar Based on these baseline predictors, the accrued knowledge has shown that the observed SVR rates between the harder-to-treat, treatment-naïve population and documented PEG+RBV partial/null responders in the treatment-experienced population are similar. One of the important questions during the review process was whether these 3 baseline factors (ie, IL-28B host genotype non-CC, baseline viral load >800,000 IU/mL, and Metavir fibrosis score F3–F4) predict a poor response to PEG+RBV in GT1 treatment-naïve patients. Analyses were performed using these same baseline factors with data from previously submitted clinical trials containing PEG+RBV treatment-naïve control arms. Although only limited data were available (particularly for IL-28B data), these analyses demonstrated that treatment-naïve subjects with these 3 baseline factors who received PEG+RBV had an overall treatment response (9% SVR [n/N: 4/43]) comparable with the observed SVR rate in prior documented PEG+RBV treatment failures. However, it should be noted that many different combinations of baseline factors can be used to predict treatment-naïve patients who may potentially fail PEG+RBV treatment, and that the selected factors may have varying predictive ability. This analysis was then extended to the observed treatment results in GT1 treatment-naïve subjects from NEUTRINO. Specifically, we selected the subjects with the 3 baseline factors noted, which were associated with a lower SVR rate to PEG+RBV treatment. It was hypothesized that subjects with these baseline factors (eg, with baseline non-CC IL-28B genotype, higher baseline HCV RNA, and cirrhosis) would be representative of the response rate of PEG+RBV nonresponders. Additional baseline factors are anticipated to impact treatment response with DAAs, particularly viral subtype. For completeness, the response rates based on GT subtype are also included in the table, but a multifactorial analysis with genotype subtype (GT1b) and other baseline factors could not be conducted, because the number of GT1b subjects with non-CC IL-28B genotype and/or cirrhosis was too small to provide any informative comparisons. As shown in Table 2, non-CC IL-28B subjects had lower SVR rates than CC subjects, subjects with higher viral load at baseline had lower SVR rates than those with lower baseline levels, and subjects with higher Metavir scores or cirrhosis had lower SVR rates that subjects with lower Metavir scores or no cirrhosis at baseline. In addition, subjects with GT1b HCV had a lower response rate of 82% compared with 92% for subjects with GT1a HCV.Table 2Comparison of SVR12 Rates Among Subjects with Various Baseline Predictors of Response in NEUTRINOBaseline FactorsSOF+PEG+RBV 12 Weeks SVR12, % (n/N)IL-28B CC98 (84/86) Non-CC86 (177/206)Baseline HCV RNA 800,000 IU/mL88 (214/243)Metavir score F0-F291 (128/141) F3-F482 (68/83) Missing96 (65/68)Cirrhosis status Cirrhotic81 (42/52) Noncirrhotic91 (216/237) Missing100 (3/3)Genotype subtype GT1a92 (206/225aOne subject had genotype 1a/1b mixed infection.) GT1b82 (54/66)Non-CC/high baseline viral load/Metavir F3-F471 (37/52)IL = interleukin; PEG = pegylated interferon; RBV = ribavirin; SOF = sofosbuvir; SVR12 = sustained virologic response rate 12 weeks after completing therapy.a One subject had genotype 1a/1b mixed infection. Open table in a new tab IL = interleukin; PEG = pegylated interferon; RBV = ribavirin; SOF = sofosbuvir; SVR12 = sustained virologic response rate 12 weeks after completing therapy. In subjects with multiple baseline factors associated with poor response (non-CC IL-28B, high baseline viral load, and Metavir F3–F4), the SVR12 rate was 71% (95% CI, 57%–83%) which greatly exceeds the PEG+RBV response rate previously observed for null and partial responders (5%–20% SVR rate).5Shiffman M.L. Di Bisceglie A.M. Lindsay K.L. et al.HALT-C Trial GroupPeginterferon alfa-2a and ribavirin in patients with chronic hepatitis C who have failed prior treatment.Gastroenterology. 2004; 126: 1015-1023Abstract Full Text Full Text PDF PubMed Scopus (472) Google Scholar, 11PEGINTRON (peginterferon alfa-2b) injection, for subcutaneous use. US Prescribing Information. Available at http://www.accessdata.fda.gov/drugsatfda_docs/label/2014/103949s5288lbl.pdf.Google Scholar, 12INCIVEK (telaprevir) tablets, for oral use. US Prescribing Information. Available at http://www.accessdata.fda.gov/drugsatfda_docs/label/2013/201917s012lbl.pdf.Google Scholar, 13VICTRELIS (boceprevir) capsules, for oral use. US Prescribing Information. Available at http://www.accessdata.fda.gov/drugsatfda_docs/label/2014/202258s014lbl.pdf.Google Scholar A within-study verification of this result is not possible as NEUTRINO did not include a PEG+RBV control arm. Similarly, a comparison between this predicted response rate in a treatment-naïve subset and response rates in PEG+RBV nonresponders could not be performed because PEG+RBV nonresponders were not evaluated with SOF+PEG+RBV. It should be noted that although these analyses provide supportive evidence for treating prior PEG+RBV nonresponders with SOF+PEG+RBV, it is not sufficient to determine the optimal regimen in such patients. It is evident from other approved HCV regimens that on-treatment response is associated with treatment outcome and can be used as a measure for informing treatment duration (ie, response-guided therapy). When a similar analysis was conducted for NEUTRINO, the data suggested that subjects with detectable HCV RNA at treatment week 2 (TW2) were less likely to achieve SVR12 (86% SVR12; 128/149) than subjects with HCV RNA target not detected at TW2 (94% SVR12; 131/140). Treatment-naïve subjects with multiple poor baseline predictive factors (such as high baseline viral load, non-CC IL-28B, advanced fibrosis/cirrhosis) were more likely to have HCV RNA detectable at TW2 (71%; 37/52) than other treatment-naïve GT1 subjects (47%; 112/237). Because the primary reason for treatment failure was relapse, this observation indicates that the optimal treatment duration for a subset of GT1 subjects (treatment-naïve with poor baseline predictive factors and prior PEG+RBV nonresponders) may be different. However, it should be noted that overall the number of subjects with detectable HCV RNA at TW2 was low and that SVR12 was achieved in a majority of subjects despite TW2 response, making it a poor predictor of treatment outcome. The prior PEG+RBV treatment classification (eg, naïve, nonresponders) for HCV treatment failures will be less applicable as a PEG+RBV regimen alone is no longer the standard of care for HCV GT1 patients. Alternative definitions of these patient populations based on baseline characteristics or other undefined factors will likely have greater importance in the future to help guide treatment recommendations. It should be noted that there may be other baseline factors (eg, HCV genotype subtype) or combinations of factors that are predictive of patient response among DAA+PEG+RBV regimens as well as emerging interferon-free regimens. This exploratory analysis was based on the assumption that the odds ratio or relative risk between the treatment-naïve population and prior null responders, defined as patients having a <2-log10 decrease in HCV RNA at week 12 of PEG+RBV therapy, was the same in the SOF+PEG+RBV regimen as that observed in previous HCV programs including boceprevir and telaprevir (Table 3). The PEG+RBV null responders were chosen for the analysis because they represented the most difficult-to-treat subset in the PEG+RBV treatment failures. As shown in Table 3, the odds ratios between treatment-naïve and prior PEG+RBV null responders were different among the HCV programs, but the relative risks were fairly consistent.Table 3SVR in Genotype 1 (GT1) Treatment-Naïve Subjects and Prior Pegylated Interferon and Ribavirin (PEG+RBV) Null Responders in Boceprevir and TelaprevirDrugSVR Rate in GT1 Treatment NaïveSVR Rate in GT1 Prior PEG+RBV Null RespondersOdds Ratio Between Treatment-naïve and Null RespondersRelative Risk Between Treatment naïve and Null RespondersBocepreviraData source: Approved drug labels for telaprevir and boceprevir. Available at Drugs@FDA.66% (242/366)38% (20/52)3.10.5TelapreviraData source: Approved drug labels for telaprevir and boceprevir. Available at Drugs@FDA.79% (285/363)32% (47/147)7.80.3SVR = sustained virologic response.a Data source: Approved drug labels for telaprevir and boceprevir. Available at Drugs@FDA. Open table in a new tab SVR = sustained virologic response. The extrapolation used the observed SVR12 rates for the HCV GT1 subjects from either the telaprevir and/or boceprevir treatment-naïve and prior null responder trials and NEUTRINO (treatment-naïve subjects) to derive the SVR12 rate for the 12-week SOF+PEG+RBV treatment in the GT1 prior PEG+RBV null responders. Specifically, the extrapolation of the SVR12 rate for the 12 weeks of SOF+PEG+RBV treatment in prior PEG+RBV null responders was performed by equating the odds ratios or relative risks and solving for the unknown SVR 12 rate. Table 4 summarizes the analysis results. The extrapolation based on equivalent odds ratios resulted in lower predicted SVR rates ranging from 52% to 73%, compared with 66% to 81% predicted SVR rates based on equivalent relative risks.Table 4Predicted Sustained Virologic Response (SVR) Rates for 12-Week SOF+PEG+RBV in Prior PEG+RBV Null Responders Based on ExtrapolationComparator DrugPredicted SVR Rate for 12-week SOF+PEG+RBV in Prior PEG+RBV Null Responders Based on Odds Ratio (95% CI)Predicted SVR Rate for 12-week SOF+PEG+RBV in Prior PEG+RBV Null Responders Based on Relative Risk (95% CI)Boceprevir73% (57%, 85%)81% (71%, 87%)Telaprevir52% (38%, 66%)66% (50%, 76%)PEG = pegylated interferon; RBV = ribavirin; SOF = sofosbuvir. Open table in a new tab PEG = pegylated interferon; RBV = ribavirin; SOF = sofosbuvir. One of the limitations of the analyses is that it assumes that the relative efficacy between treatment-naïve and null responders is the same for the SOF-containing regimen and other regimens for treatment of CHC infection in GT1 subjects. This may not be plausible because SOF is a nucleotide analog inhibitor, and boceprevir and telaprevir are HCV protease inhibitors. Another limitation of the approach was that it did not consider the different treatment durations for treatment-naïve and prior PEG+RBV null responders in the comparator drugs. Response-guided therapy was used in the treatment-naïve trials of boceprevir and telaprevir but not in the PEG+RBV null-responder trials. Therefore, the treatment-naïve subjects may have a shorter treatment duration compared with the prior PEG+RBV null responders in the comparator drugs. As a result, the predicted SVR rates for the SOF regimen in the prior PEG+RBV null responders may be an overestimate. The advantages of SOF in combination with PEG+RBV for 12 weeks in the HCV GT1 population include an improved safety profile, shorter treatment duration, and low pill burden. Specifically, once-daily dosing of SOF in addition to lower interferon injection burden owing to shorter treatment duration may improve patient adherence to therapy. In addition, there are fewer drug–drug interactions with SOF compared with previously approved HCV protease inhibitors. Based on the cumulative evidence gathered by the analyses conducted by the FDA review team, as well as the clinical benefit considerations, SOF was indicated in all GT1 patients irrespective of their previous treatment experience. The analyses shown herein, which bridge information from varying sources, provide evidence that SOF would have at least similar effectiveness in prior PEG+RBV nonresponders as other approved regimens. Although these analyses provide data to support SOF+PEG+RBV regimen as an effective therapeutic option for prior PEG+RBV nonresponders, the available data are not sufficient to support estimates of the SVR12 rates in specific subgroups of PEG+RBV nonresponders, such as prior null responders and prior partial responders. Moreover, although the data support the recommended regimen is effective, additional data may be helpful in determining the optimal regimen and treatment duration in this harder to treat population. Acknowledging the limitations of the analyses presented herein, there exists substantial evidence of safety and efficacy, in addition to the benefits of shorter treatment duration and improved tolerability profile, which supports SOF use in combination with PEG+RBV as a therapeutic option for a subgroup of the GT1 patient population previously classified as PEG+RBV nonresponders. Bridging analyses have allowed us to utilize our understanding of the biology of response to therapy in CHC viral infection and thus provides a new tool to facilitate timely access of novel therapies to patients in need. Data for these analyses were submitted to the United States Food and Drug Administration by Gilead Sciences Inc as part of a new drug application. Building Bridges and Providing Transparency to the Hepatitis C Virus Drug Approval ProcessGastroenterologyVol. 147Issue 6PreviewChronic hepatitis C is a major cause of cirrhosis and hepatocellular carcinoma worldwide1 and the leading indication for adult liver transplants in the United States and Europe.2 Although the incidence of chronic hepatitis C is declining in the developed world, the number of deaths owing to complications of the infection is estimated to increase over the next decade.3 A parallel rise in the cost of managing complications of advanced chronic hepatitis C is projected to increase from a current $6.5 billion to approximately $9.1 billion in 2024 in the United States. Full-Text PDF
Referência(s)