1.      Abbreviations:

SVR        :               Sustained virologic response

CHC       :               Chronic Hepatitis C

TVR       :               Telaprevir

BVR       :               Boceprevir

DAA       :               Direct Acting Anti-Viral therapy

eRVR     :               Extended Rapid Virologic Response

EOT VR :               End of treatment virologic response

DMID    :               Division of Microbiology and Infectious Diseases

2.      Introduction

Chronic Hepatitis C (CHC) infection is a major cause of chronic liver disease, hepatocellular carcinoma and is a common indication for liver transplantation [1]. Telaprevir (TVR) and Boceprevir (BVR), the first FDA approved NS3/4A protease inhibitorschanged the landscape of CHC therapy in 2011. Despite being discontinued in the US market after 3 years of FDA approval, both protease inhibitors left a great impact in the treatment approach of CHC: ushering in a new era of Direct Acting Anti-Viral (DAA) therapy for CHC. These agents achieved higher sustained virologic responses (SVR) in clinical trials of CHC genotype 1 patients (up to 70%-80%). This is a dramatic improvement from the old dual regimen of Pegylated interferon and ribavirin which had lower SVR rates (14%-29%) [2]. A higher SVR predicts long-term response to Treatment [3] and reduces liver-specific as well as all-cause mortality [4]. Despite the enhanced efficacy of TVR/BVR based triple therapy, there were associated significant adverse outcomes and toxicities.

Randomized clinical trials have important characteristics but may have limited generalizability in terms of efficacy, safety, and tolerability of these drugs in everyday outpatient setting outside of a controlled trial. We report our experience of 113 CHC patients, with emphasis in our analysis on the predictors of SVR and treatment failure in everyday life outside of a controlled trial setting.

3.      Methods

A single center prospective cohort of CHC patients (treatment naïve, experienced and Cirrhotics) treated at the Liver Associates of Texas Hepatology Clinics between July 2011 and Jan 2014. Patients were non-randomly assigned to two treatment groups, either TVR or BVR based triple regimen with Pegylated interferon and ribavirin. Response guided therapy was adopted. Futility rules were applied. Endpoints of the study included comparison of Extended Rapid Virologic Response (eRVR), end of treatment virologic response (EOT VR), SVR at 12 weeks (SVR12) and 24 weeks (SVR24.Outcomes and treatment discontinuations were analyzed, with emphasis on 8 potential predictors of SVR: Gender, Age, BMI, Genotype, Ethnicity, eRVR, Prior treatment experience and Pre-treatment HCV RNA levels.

Statistical analysis was performed using statistical software package R (version 3.1.0, R Foundation for Statistical Computing) and SPSS statistics v19. Correlation matrices were built to compare maintenance of treatment response between the TVR and BVR groups at ETR, SVR12, and SVR24. Logistic regression models for treatment response at ETR, SVR12, and SVR24 were built to identify predictors of treatment response. Reduction of predictor variables for the logistic regression models was performed through a regression variable selection method (Akaike’s information criterion).

4.                  Results

4.1.              Patient characteristics and Virologic responses

113 CHC patients were treated with either TVR (85/113) or BVR (28/113) based triple regimen. Patient characteristics and potential predictors of treatment response are shown in Table 1. Virologic responses are summarized in Table 2, and are compared to SVR rates in clinical trials in table 3. SVR12 and SVR24 data were available for 107 and 105 patients, respectively. Compensated cirrhosis was present in 41 CHC patients based on liver biopsy (35 received TVR and 6 received BVR).

4.2.DAAs treatment Failure

Rates of treatment failure with DAAs are presented in table 4. Overall, CHC therapy was discontinued in 38/85 (44.7%) patients in the TVR cohort and 9/28 (32.1%) patients in the BVR cohort. This is due to either viral breakthrough during therapy, non-response to DAAs or secondary to serious adverse events. 4/85 (5%) relapsed to TVR-based therapy and 8/85 (9.4%) experienced a non-response. 3/28 (11%) relapsed to BVR-based therapy and 21.4% (6/28) experienced a non-response.

Serious adverse events included hematological changes (Thrombocytopenia, anemia or neutropenia - DMID toxicity grades 3-4), hepatic decompensation, DRESS syndrome, GI bleeding, myocardial Infarction, acute renal failure, depression with suicidal ideation or fever and retroperitoneal inflammation.

In CHC patients with cirrhosis, non-response to TVR was observed in 8.6% of cirrhotic patients and with BVR in 33%. 6% of cirrhotic TVR patients and 17% of the cirrhotic BVR patients relapsed. 14% of TVR patients with cirrhosis experienced treatment discontinuation secondary to viral breakthrough. Serious adverse events were seen in 20% of TVR patients.

4.3.              Maintenance of Virologic Response

Analysis of correlation matrices revealed no significant difference in maintenance of treatment response at ETR, SVR12, and SVR24 between TVR or BVR treatment groups (p > 0.05).

4.4.              Predictors of EOT VR, SVR12 and SVR24 (presented in Table 5)

Due to low subject to variable ratio in our single center study, we implemented regression variable selection using Akaike’s Information Criterion (AIC). Applied to our model, regression variable selection revealed age, genotype, BMI, eRVR, and prior treatment experience as the most relevant predictor variables for treatment response at end of treatment and SVR12 (Pre-treatment HCV RNA viral load, Gender, and Ethnicity were dropped). For SVR24, the AIC-selected predictors were age, genotype, eRVR, prior treatment experience and low pretreatment HCV RNA (BMI, Gender, and Ethnicity were dropped).

Predictors of EOT: Logistic regression models for EOT VR with all 8 predictors revealed no significant predictors of response at end of treatment. In addition, based on AIC-selected predictors, regression variable selection showed that none of these variables were significant predictors of response.

Predictors of SVR12: Logistic regression models with 8 predictors of SVR12, showed that age was a positive predictor, although the odds ratio was very small (OR: 1.07, p=0.0476).

However, regression variable selection showed that none of the variables were significant predictors of SVR12

Predictors of SVR24: Logistic regression models with 8 predictors of SVR24 showed that age was a positive predictor of treatment response, although the odds ratio is very small (OR: 1.08, p=0.0444). However, based on AIC-selected predictors, none of the variables were significant predictors of SVR24.

5.      Discussion

The efficacy of DAAs in our experience in everyday outpatient setting outside of a controlled trial differed from that of previously reported to the clinical trials, with major differences in SVR rates. This supports the importance of studying FDA approved pharmaceutical agents in real life as clinical trials data can have limited generalizability.In everyday outpatient hepatology practice, outcomes and causes of discontinuations varied significantly among both treatment groups. The TVR group experienced higher rates of serious adverse events leading to treatment discontinuation, while the BVR group failed therapy mainly secondary to non-response or relapse to DAA. The higher rates of viral breakthrough in the TVR group could be explained by earlier treatment discontinuation secondary to serious adverse event occurrence.The same outcomes apply to patients with liver cirrhosis.

Of the three logistic regression models built for predictor variables at EOT VR, SVR12 and SVR24, only the SVR12 and SVR24 models had a significant predictor of response.  Age was a positive predictor of response at SVR12 and SVR24. However, given that the odds ratio was very small, and that advanced age rarely leads to better clinical outcomes, this is likely statistical noise due to small power rather than a real effect of the predictor. This was confirmed when regression variable selection was performed to improve model stability and subject-to-variable ratio, where the small positive effect of age on treatment outcome did not exist for both SVR12 and SVR24 models.As a result, based on AIC-selected predictors, regression variable selection showed that none of the variables significantly predicted end of treatment response, SVR12 and SVR24. This does not correspond with data from recent studies [5-8].

In summary, our study demonstrates the importance of everyday outpatient hepatology practice data for new drugs that have obtained FDA approval.

6.      Financial support

None.

7.      Potential competing interests:

The authors declare that there is no conflict of interest regarding the publication of this paper, except for the following author:

Victor Ankoma-Sey, MD:

-          Speaking: Abbvie, Intercept, Gilead, Merck, Salix.

-          Research: Gilead, Hologic, Conatus, Intercept, Novo Nordisk, Genfit, Octeta, Inovio, Abbvie, Mayo, Exalenz, EiSAI.

Table 1: Characteristics of Chronic Hepatitis C patients and potential predictors of response to therapy.

Table 2: Virologic responses to DAA therapy; TVR vs BVR-based triple regimen.

Table 3: Comparison of SVR rates in everyday outpatient practice vs clinical trials.

Table 4: Treatment failure rates with DAA therapy.

Table 5:Predictors of EOT VR, SVR12 and SVR24 based on regression variable selection.

1.       Tsoulfas G, Goulis I, Giakoustidis D, Akriviadis E, Agorastou P, et al. (2009) Hepatitis C and liver transplantation. Hippokratia 13: 211-215.

2.       Falck-Ytter Y, Kale H, Mullen KD, Sorescu L, et al. (2002) Surprisingly small effect of antiviral treatment in patients with hepatitis C. Ann Intern Med136:288-292

3.       Ghany MG, Strader DB, Thomas DL, Seeff LB (2009) American Association for the Study of Liver Diseases. Diagnosis, management, and treatment of hepatitis C: an update. Hepatology49:1335-1374.

4.       Backus LI, Boothroyd DB, Phillips BR, et al. (2011) A sustained virologic response reduces risk of all-cause mortality in patients with hepatitis C. Clin Gastroenterol Hepatol9:509-516.

5.       Werner CR,Franz C,Egetemeyr DP,Beck R,Malek NP, et al. (2015) First-generation protease inhibitor-triple therapy: SVR 24, safety, and predictors of response in a large single center cohort. Virol J12: 37.

6.       Manzano-RobledaMdel C, Ornelas-Arroyo V, Barrientos-Gutiérrez T, Méndez-Sánchez N, Uribe M, et al. (2015) Boceprevir and telaprevir for chronic genotype 1 hepatitis C virus infection. A systematic review and meta-analysis.Ann Hepatol 14: 46-57.

7.       Butt AA,Yan P,Shaikh OS,Freiberg MS,Lo Re V, et al. (2015) Virologic response and haematologic toxicity of boceprevir- and telaprevir-containing regimens in actual clinical settings. J Viral Hepat22:691-700.

8.       Sterling RK, Kuo A, Rustgi VK, Sulkowski MS, Stewart TG, et al. (2015) Virological outcomes and treatment algorithms utilisation in observational study of patients with chronic hepatitis C treated with boceprevir or telaprevir. Aliment PharmacolTher41:671-685.