Asunaprevir

Favorable outcome of retreatment by direct-acting antivirals for hepatitis C patients with daclatasvir plus asunaprevir combination therapy failure

Aim: In patients with hepatitis C virus, treatment failure of daclatasvir plus asunaprevir combination therapy (DCV + ASV) seems to become intractable due to the induction of resistance-associated substitutions. This study aimed to investi- gate the outcomes of retreatment with direct-acting antivirals (DAAs) in patients with DCV + ASV therapy failure, as well as changes in drug resistance mutations.
Methods: We retrospectively analyzed 44 patients re-treated with DAAs after DCV + ASV failure between December 2015 and April 2018. All patients were analyzed for amino acid substi- tutions, and additional treatment regimens were selected based on the results and current treatment guidelines. Results: The sustained virological response rate with second- line treatment was 81.8% (36/44), and relapse occurred in five of 16 patients who received sofosbuvir/ledipasvir and three of seven patients who received DCV/ASV/beclabuvir. Third- and fourth-line treatments were also tried in relapsed cases, and the overall sustained virological response rates were 90.9% (40/ 44) and 93.2% (41/44), respectively. A high rate of viral clearance was eventually observed. Before second-line treatment, the prevalence of mutations in the NS5A and NS3/4A regions was 100% (44/44) and 86.4% (38/44), respectively. There was no signif- icant increase in the number of amino acid substitutions in patients for whom second-line treatment failed. Conclusions: Amino acid substitutions were frequently ob- served in patients with DCV + ASV failure, but most patients achieved a sustained virological response after retreatment with DAAs. Although the spread of drug-resistant viruses due to unsuccessful DAA treatment was a matter of concern, most cases of DCV + ASV failure were overcome with additional treatment.

INTRODUCTION
COMBINATION THERAPY WITH daclatasvir, an NS5A inhibitor, and asunaprevir, an NS3/4A inhibitor, (DCV + ASV) was the first interferon (IFN)-free directantiviral agent (DAA) regimen used to treat hepatitis C virus (HCV) infection in Japan.1 This treatment achieved a high sustained virological response (SVR) rate for HCV genotype 1b patients.2,3 The SVR rate was non-inferior to (and higher than) telaprevir plus peg-IFN/ribavirin.4 InCorrespondence: Dr Satoru Kakizaki, Department of Gastroenterology and Hepatology, Gunma University Graduate School of Medicine, 3-39-15 Showa- machi, Maebashi, Gunma 371-8511, Japan. Email: [email protected] of interest: S Kakizaki received research funding from BMS and Gilead Sciences. K Sato received research funding from AbbVie.Financial support: None declared.Received 13 August 2019; revision 23 October 2019; accepted 1 November 2019.© 2019 The Japan Society of Hepatology 303the clinical trial, the SVR rates at 12 weeks and 24 weeks after treatment were 85.1% and 84.7%, respectively.2 The incidence of adverse effects was lower in comparison with that of patients receiving conventional IFN therapy, making it possible to treat patients who were difficult to treat with IFN-containing regimens.2–5 In Japan, DCV+ ASV has been available for HCV genotype 1 patients since 2014, and is administered to many patients.

It is reported that the factors associated with a non-SVR in patients receiving DCV + ASV therapy include poor drug adherence and resistance-associated substitutions (RASs) of NS5A or NS3/4A, such as mutations of L31 and Y93. In a phase III study, the SVR rate of patients with L31 and/or Y93 mutations at baseline was as low as 40.5%.2 It has been reported that RASs appeared after unsuccessful treatment, and that RASs in the NS5A region persist for>1 year.5,6 Thus, the induction of RASs due to DCV+ ASV failure might be associated with the development of multidrug resistance that is difficult treat.In fact, combination therapy with sofosbuvir (SOF) and ledipasvir (LDV) (SOF/LDV) was highly effective against HCV genotype 1 in clinical trials, with an SVR12 rate of 98–100%.7,8 However, the SVR rate decreased in patients with a history of DCV + ASV treatment, and a history of DCV + ASV treatment was considered to be a risk factor for treatment failure.9–13 It is also reported that IL28B rs8099917 non-TT type is correlated with non-SVR in SOF/LDV treatment as salvage therapy for prior NS5A in- hibitor regimens.9,10Glecaprevir (GLE), NS3/4A inhibitor, and pibrentasvir (PIB), NS5A inhibitor, combination therapy (GLE/PIB) has recently been approved, and is expected to achieve a high SVR12 rate of 93% in patients with DCV + ASV fail- ure.14,15 Although this regimen is expected to be used as a salvage therapy, there are still few real-world data on pa- tients with DCV + ASV failure. In this study, we retrospec- tively investigated the overall outcomes of patients with DCV + ASV failure who were retreated with DAAs.Patient selection and treatment regimens ROM DECEMBER 2015 to April 2018, a total of 87 pa- tients with DAA failure were referred to our institution(Gunma University Hospital, Maebashi, Japan) for DAAretreatment.

There were 47 patients with DCV + ASV fail- ure. We analyzed 44 traceable cases among them. Patients who met the following criteria were excluded from further treatment: (i) decompensated cirrhosis (clinical findings, Child–Pugh grade B or higher); (ii) active hepatocellular carcinoma (HCC); and (iii) problematic complicationsinvolving other organs. Only patients with inactive HCC that were treated with curative intent were included. The amino acid substitutions (AASs) in the NS3/4A and NS5A regions were analyzed in all patients for which additional DAA treatment was indicated, and additional treatment regimens were selected based on the results and treatment guidelines or clinical trials. Additional treat- ments were selected for each patient: SOF 400 mg + LDV 90 mg, or ombitasvir 12.5 mg + paritaprevir 78.5 mg + ri- itonavir 50 mg (OBV/PTV/r), or DCV 15 mg + ASV 100 mg + beclabuvir 37.5 mg (DCV/ASV/BCV), or GLE 100 mg + PIB 40 mg, or SOF 400 mg + velpatasvir (VEL) 100 mg + ribavirin (SOF/VEL + RBV). We used these five regimens in this study, but OBV/PTV/r and DCV/ASV/ BCV are no longer available in Japan. The duration of treatment and the dosage of the drug were determined in accordance with the guidelines, health insurance policies, and clinical trials. If additional treatment was unsuccess- ful, the AASs were reanalyzed and additional treatment was given if possible.This study was approved by the institutional review board, and the need for written informed consent was waived because of the retrospective nature of the study.

The treatment outcome was judged based on the achieve- ment of a SVR12. Cases that were HCV-RNA-negative until the end of treatment and which became positive during the follow-up period were classified as relapsed cases. Cases that became HCV-RNA-negative then became HCV-RNA-positive during treatment were classified as breakthrough cases. A “non-response” was defined by the absence of HCV-RNA negativity during or after treatment.Laboratory tests and drug resistance mutationsLaboratory data were analyzed at the start of additional treatment, at the end, and at 12 weeks after the additional treatment. The Albumin-Bilirubin (ALBI) score was calcu- lated for evaluating hepatic functional reserve.16 HCV- RNA was measured by real-time polymerase chain reac- tion. AASs were analyzed by direct sequencing (LSI Medience, Tokyo, Japan). AASs in the NS3/4A and NS5A regions were analyzed in all cases. In patients treated with regimens containing NS5B polymerase inhibitors, AASs in the NS5B region, were also analyzed.An intention to treat analysis was performed based on the outcome of the additional treatment, and the final thera- peutic effect was evaluated. Continuous variables werecompared using the Mann–Whitney U-test and Kruskal– Wallis test. Nominal variables were compared by Fisher’s exact test. A post-hoc test of the comparisons of more than three groups was carried out using the Holm method. The EZR software program (R Foundation, Vienna, Austria) was used to carry out the statistical analyses.

RESULTS
HE FOLLOWING REGIMENS were selected for the secondary treatment of the 44 patients: SOF/LDV (SOF/LDV group; n = 16), OBV/PTV/r (OBV/PTV/r group; n = 1), DCV/ASV/BCV (DCV/ASV/BCV group; n = 7),GLE/PIB (GLE/PIB group; n = 17), and SOF/VEL/RBV (SOF/VEL/RBV group; n = 3; Figure 1). The background characteristics and laboratory data of the 44 patients are presented in Table 1. There were 16 men and 28 women. The median age was 68 years (range 53–84 years). A total of 23 patients were diagnosed with chronic hepatitis, and21 were diagnosed with liver cirrhosis. Seven patients were L31 mutation-positive before secondary treatment,and 13 patients were Y93 mutation-positive. A total of 23 patients had a history of IFN treatment, and seven had a history of IFN-based DAA treatment. Seven pa- tients had been treated for HCC. The prevalence of liver cirrhosis in the SOF/LDV group was higher than that in the other groups (75.0% [12/16] vs. 32.1% [9/28], P = 0.011). There were no significant differences in other factors.Treatment outcomesThe outcomes of additional treatment are shown in Figure 1a. The SVR rate in patients receiving second-line treatment was 81.8% (36/44), whereas the rates in patients receiving third-line and fourth-line treatment were 90.8% (40/44) and 93.2% (41/44), respectively, indicat- ing good results. All cases achieved an SVR12 in the OBV/PTV/r (1/1), GLE/PIB (17/17), and SOF/VEL + RBV(3/3) groups.

In the SOF/LDV group (n = 16), 11 patients achieved an SVR12, and five patients relapsed. One case of deletion of P32, known as a very strong drug-resistant mutation, was observed in the SOF/LDV group. Despitethe P32 deletion, an SVR12 was achieved. Four patients with relapsed disease in the SOF/LDV group received third-line treatment. For third-line treatment, three pa- tients received GLE/PIB and all patients achieved an SVR12. One patient who received DCV/ASV/BCV as third-line treatment relapsed; however, this patient was treated with GLE/PIB as a fourth-line treatment and achieved an SVR12. The overall SVR12 rate in the SOF/LDV group was 93.8% (15/16). In the DCV/ASV/ BCV group (n = 7), four patients achieved an SVR12. Two of these patients received third-line treatment with GLE/PIB, and one patient achieved an SVR12; one patient relapsed. The remaining relapsed patient was under treat- ment with SOF/VEL + RBV. The SVR12 rate in the DCV/ASV/BCV group was 71.4% (5/7). Finally, the overall SVR12 rate in all DCV + ASV failure cases was 93.2% (41/ 44). The SVR rates in the prior IFN-based DAA treatmenthistory present group was significantly lower than the ab- sent group in second-line treatments (28.6% [2/7] vs. 91.9% [34/37], P < 0.001), as shown in Figure 1b. Therewere no significant differences in the SVR rate betweenthe patients with liver cirrhosis and without liver cirrhosis in second-line treatments.Pre-existence and emergence of AASs Table 2 shows the positive rates for AASs in NS5A, NS3/4A, and NS5B. All patients had AASs in the NS5A region beforesecondary treatment. Mutations of the NS3/4A region wererecognized in 86.4% of patients. Thus, mutations in both the NS5A and NS3/4A regions were found in 86.4% of the patients. In the NS5A region, mutations in Y93, L31, A92, and P32 were found in 61.4% (27/44), 54.5% (24/44), 4.5% (2/44), and 4.5% (2/44) of the patients, re- spectively. The mutations in P32 included a deletion (n = 1) and a non-analyzable mutation (n = 1). A total of 23 patients (52.3%) had both L31 and Y93 RASs. In the NS3/4A region, 16 patients had a D168 mutation. Two or more mutations were detected in the NS5A region in 90.9% (40/44), and in the NS3/4A region in 70.5% (31/44) of patients.There were unsuccessful cases in the SOF/LDV and DCV/ASV/BCV groups. In the SOF/LDV group, the inci- dence of mutations in Y93 was significantly higher in the unsuccessful group (P = 0.007). The incidence of muta- tions in D168 was also significantly higher in the unsuc- cessful group (P = 0.034). There were no significant mutations in the DCV/ASV/BCV group (Fig. 2). Due to the use of NS5B polymerase inhibitor, AASs in the NS5B region were also analyzed. A mutation in NS5B C316 was found in 80% (4/5) of the patients in the SOF/LDV group, and 100% (2/2) of the patients in the DCV/ASV/ BCV group.In the 30 patients in which NS5A L31 and Y93 muta- tions were analyzed before DCV + ASV, L31 mutations were found in 23.3% (7/30) of the patients before primary treatment. This rate increased significantly to 53.3% (16/ 30) after DCV + ASV failure (P = 0.016). The incidence of Y93 mutations also increased from 41.9% (13/31) to 51.6% (16/31); however, the difference was not statisti- cally significant. In patients who received three or more DAA treatments, there was no significant increase in the number of AASs in the NS5A or NS3/4A regions before the start of second-line and third-line treatment (Fig. 3). Although the number of AASs increased after unsuccessful treatment after DCV+ASV failure, the number of AASs did not increase after the failure of other regimens in the present study.In patients who achieved an SVR12, the serum alanine aminotransferase (P < 0.001) and aspartate aminotrans- ferase (P < 0.001) levels significantly decreased. The serumalbumin level did not change in the overall population (Table 3). However, it was significantly improved in the patients who showed decreased serum albumin level at the start of additional treatment. The total bilirubin level, prothrombin time and platelet count did not changebefore or after treatment (data not shown). However, the ALBI score, a recently reported method for evaluating the hepatic functional reserve, also improved in the albumin- lowering group.16Unsuccessful casesThere were three unsuccessful cases in this study. The reasons for abandoning additional treatments included complications (lung cancer and pneumothorax), carcino- genesis in the liver, and waiting for a new treatment regimen. The last case is currently being treated with SOF/VEL + RBV. DISCUSSION N PATIENTS WITH DCV + ASV failure, a high SVR rate of 93.2% was finally achieved. However, the SVR rate was as low as 81.8% with second-line treatment, especiallyin patients who were retreated with SOF/LDV or DCV/ASV/BCV, who had SVR rates of 68.8% and 57.1%, respectively. The SVR rate with SOF/LDV in patients with a history of DCV + ASV treatment has been reported to be 66–87%.9–13 Iio et al. reported that the accumulation of RASs due to repeated treatment was the main reason for lower SVR rates.13 A similar cause was suggested in the present study, because the prevalence of L31 mutations in patients with DCV + ASV failure was increased in com- parison with before primary treatment, and 90.9% of the patients had two or more AASs in NS5A. Among patients with SOF/LDV failure, a high rate of C316 mutations was observed in the NS5B region. C316 mutations are reported to be found in 46.9% of Japanese HCV genotype 1b- infected patients, and it is reported to reduce the SVR rates of genotype 1b patients treated with SOF/LDV.18–21The SOF/LDV group showed a significantly higher prev- alence of liver cirrhosis in comparison with the other groups (P = 0.011). In 2015, SOF/LDV was approved in Japan (after DCV + ASV) for patients with chronic hepati- tis C and compensated liver cirrhosis. For this reason, among patients in whom DCV + ASV failed, many patients with advanced liver fibrosis who were at high risk for HCC were treated with SOF/LDV. Previous studies have re- ported that the progression of liver fibrosis reduces the SVR rate in patients treated with DAAs.10,22 Akuta et al. re- ported that a high FIB-4 index is associated with failure of SOF/LDV treatment,10 suggesting that the treatment out- come is reduced in patients with advanced hepatic fibrosis. Although it did not reach statistical significance in the pres- ent study, the higher incidence of advanced liver fibrosis in the SOF/LDV group might have also been a factor in the higher incidence of treatment failure.It is reported that the factors associated with a non-SVR in patients receiving DCV + ASV therapy include poor drug adherence and RASs of NS5A or NS3/4A. In the present study, 15.9% of patients had prior IFN-DAA therapy, and the Y93 mutation-positive rate was 48.5% before treat- ment. In addition, nearly half (47.7%) of the patients had cirrhosis. These are considered to be the main causes of DCV + ASV failure in the present study. Just four patients discontinued treatment, and poor adherence was consid- ered inappropriate as a cause of treatment failure in the present study.In clinical trials, DCV/ASV/BCV showed good results in patients with and without L31 and Y93 mutations.23,24 However, prior DAA failure, particularly DCV + ASV fail- ure, has been reported to be associated with resistance to DCV/ASV/BCV.25 In the present study, three patients treated with DCV/ASV/BCV who had relapsed after DCV+ ASV achieved an SVR, whereas three patients with break- through failed. BCV was expected to have an effect, and the treatment was carried out; however, the effect was insuffi- cient. DCV + ASV is approved in Japan for 24 weeks, whereas DCV/ASV/BCV is approved for 12 weeks. Thus, there is a possibility that the treatment period was insufficient, despite the additional therapeutic effect of BCV. In any case, caution should be exercised in reusing agents from unsuccessful regimens. The outcomes of additional treatment with GLE/PIB were satisfactory. Among the 44 patients, 21 patients treated with GLE/PIB eventually achieved an SVR, and the SVR rate, as determined by the intention to treat analysis, was 95.5% (21/23). The SVR rate in patients who received additional treatment with GLE/PIB after un- successful DAA treatment was previously reported to be 93–100%. The results of the present study are considered to be comparable. Although the GLE/PIB group had simi- lar RAS profiles to the SOF/LDV group, GLE/PIB treatmentachieved a higher SVR rate. PIB has decreased antiviral activity against genotype 1b HCV in vitro when P32 is deleted, and GLE has decreased antiviral activity when D168 is mutated (Mavyret full prescribing information https://www.rxabbvie.com/pdf/mavyret_pi.pdf). In the present study, mutations in D168 were found in 48% (12/25) of the patients treated with GLE/PIB, but no P32 mutations were detected, and an SVR was achieved in most patients. The exclusion of cases in which a therapeutic ef- fect could not be expected (based on RAS profiles) might have been one of the factors associated with the good results.The achievement of an SVR improved the aspartate aminotransferase and alanine aminotransferase levels, and was associated with an increased albumin level in patients with decreased albumin levels at baseline. The ALBI score improved in the SVR cases, who had low albumin levels. The ALBI score was originally proposed as an indicator for hepatic functional reserve in patients with HCC.16 However, the ALBI score was recently reported to indicate liver fibrosis and prognosis in patients with chronic hepa- titis C.26 We calculated it for evaluating hepatic functional reserve. It has been reported that improvement of the synthetic ability of the liver and increased albumin were observed in HCV-infected patients who obtained an SVR with DCV + ASV.27,28 Liver stiffness has been reported to be improved in patients with advanced liver fibrosis who achieve an SVR with DAAs.29,30 Furthermore, the achieve- ment of an SVR with SOF/VEL is reported to be associated with an improved liver function in patients with decom- pensated cirrhosis,31 and there is no doubt that viral elimination with DAAs has a favorable effect on the hepatic function. Even in patients who fail first-line treatment, viral elimination might improve the hepatic function, especially in patients with a decreased hepatic reserve. Furthermore, viral elimination with DAAs might reduce the incidence of HCC similar to IFN-based therapy.32Various DAA regimens have been developed since the introduction of DCV + ASV, and the ability to strongly sal- vage cases in which treatment is unsuccessful, particularly with GLE/PIB, might have contributed to curing patients with DCV + ASV failure. It was feared that the accumula- tion of RASs after multiple DAA treatments would make HCV infection intractable. In the present study, failure of first-line treatment was associated with an increased preva- lence of L31 mutations, but there was no significant increase in the number of AASs after second-line and third-line treatment. In addition, RASs in Y93 and L31 mutations are generally being overcome due to the devel- opment of therapeutic agents, and additional treatment for patients with a history of unsuccessful DAA treatmentis promising treatment. With the advent of SOF/VEL, DAA salvage therapy can be expected with a corresponding im- provement in treatment outcomes.The present study was associated with several limita- tions. First, the treatment regimens differed according to the time in which the patients were treated, and no addi- tional treatment regimens have been decided based on the same criteria. Second, this study did not account for the low rate of measurement of IL28B gene polymor- phism, which is associated with treatment failure. Third, the analysis of RASs before DCV + ASV was insufficient, es- pecially in the NS3/4A region, and the accumulation of RASs in patients undergoing treatment with DCV + ASV was not adequately assessed. In conclusion, treatment failure with DCV + ASV might lead to the development of RASs; however, advances in therapy could overcome the effects of DCV + ASV failure. The hepatic function reserve might improve, and retreatment Asunaprevir should be attempted if possible.