Direct-acting antivirals (DAAs) are currently the treatment of choice for chronic hepatitis C virus (HCV) infection [1]. Several treatment options are available for chronic HCV such as nucleoside polymerase inhibitors like sofosbuvir (SOF), non-nucleoside polymerase inhibitors such as dasabuvir (DSV), protease inhibitors (PIs) such as simeprevir (SMV), paritaprevir boosted with ritonavir (PTVr), and grazoprevir (GZR), and NS5A inhibitors such as ledipasvir (LDV), ombitasvir (OBV), daclatasvir (DCV), elbasvir (EBV), and velpatasvir (VEL) have proven highly effective for eradicating HCV [1-3]. Another PI, asunaprevir (ASV), has been approved in some Asian countries [3]. These drugs are usually given in combination to increase their effectiveness and prevent the emergence of harbored HCV resistance-associated substitutions (RASs) [2]. The following DAA combination regimens are currently used in clinical practice for patients infected with HCV genotypes (GT) 1 and 4: 1) SMV plus SOF ± ribavirin (RBV); 2) SOF plus an NS5A inhibitor such as DCV or LDV or VEL (the two latter in fixed-dose combinations; 3) an NS3-4A protease inhibitor (PI), paritaprevir boosted with ritonavir (PTV/r) plus an NS5A inhibitor, OBV, in a fixed-dose combination with a non-nucleoside polymerase inhibitor, DSV (in GT4, DSV is not included in the regimen); 4) GZV/EBV in a fixe-ddose combination; and 5) DCV and ASV in some Asian countries. In patients infected by GT2 and GT3, the only approved regimens are SOF and RBV, and SOF with an NS5 inhibitor either DCV or VEL. In GT5 and GT6, the approved drugs are SOF combined with an NS5A inhibitor. Whichever of these regimens is used, sustained virologic response (SVR) rates in the registration studies involving GT1 and GT4 patients are higher than 90%, but SVR is lower in GT3 treatment-experienced patients. The safety profiles of these combinations, particularly RBV-free ones are excellent.

The results of treatment with these DAA regimens in various real-world cohorts show high SVR rates (80%-90%). Variations in the SVR depend on the number of patients with decompensated liver disease and cirrhosis included in the studies [3-12]. For patients with GT2 or GT3 infection, the first approved interferon-free regimen was SOF plus RBV for 12 to 24 weeks, which yielded SVR rates of 68% to 90% [13-15]. In addition, three DAA combinations are available for GT3: SOF plus DCV, SOF/LDV and SOF/VEL [3]. The extensive use of these regimens foresees that a fair number of patients will fail these newer antiviral treatments in the near future.

This review focusses on the management of patients who fail DAA regimens, paying special attention to NS5 failures. The types of failures, the characteristics of patients who fail, the role of HCV drug resistance testing, and the available rescue strategies are reviewed.

Hepatitis C virus has a high turnover rate and an estimated halflife of only 2 to 5 hours: 10¹⁰ to 10¹² virions are produced and cleared per day in an infected patient [16-18]. Because of the lack of proof-reading activity of the HCV-RNA-dependent RNA polymerase (NS5B) and the high replication activity of HCV, a large number of viral variants are continuously produced during infection (error rate, approximately 10^-3 to 10^-4 mutations per nucleotide per genomic replication) [18]. Most of these variants are cleared by the host’s immune system or are unable to replicate because of a functional loss in the encoded proteins [19,20], but a large number survive and maintain a chronic infection. The frequency of HCV isolates depends on their replication efficacy and other known and unknown viral and host factors [21]. The complex mixture of different but closely related genomes comprising the HCV population is known as a quasispecies. Typically, a dominant strain (wild-type) is detectable within the viral quasispecies, along with strains present at lower frequencies. Therapy with DAAs that incompletely suppress viral replication enables selection of preexisting strains with reduced susceptibility to the drugs administered, which play a role in treatment failure. The full clinical impact of HCV drug resistance, its implications in retreatment options, and the potential role of baseline resistance are issues lacking complete answers.

Failures to DAA combinations in pivotal trials are usually related to relapse, which is defined as serum HCV RNA below the level of quantitation at completion of therapy followed by a rebound to pretreatment levels once therapy is discontinued. Very few patients fail by virological breakthrough, defined by increases in previously undetectable HCV RNA, while they are still on antiviral treatment.

In both situations, virological relapse and breakthrough, a common finding is selection of RASs within the viral quasispecies [22]. Treatment-emergent RASs with specific amino acid substitutions result in decreases in the effectiveness of a particular drug and represent a developing therapeutic challenge for the clinicians managing these patients. Treatment-emergent RASs will persist depending on the DAAs used. RASs selected during treatment with a PI such as simeprevir are transient and disappear in a few months after discontinuing therapy. For example, in a phase-III trial including patients without cirrhosis, 17% and 3% of patients experienced a relapse in the groups receiving 8 weeks and 12 weeks of SMV and SOF respectively [23]. Among patients with compensated cirrhosis treated for 12 weeks without RBV, 16% did not achieve SVR, including 3 with viral breakthrough and 13 relapsers. Among 14 patients for whom sequencing data were available, 11 had RASs in the NS3 protease at the time of failure. No sofosbuvir RASs were detected [24].

RASs related to NS5A inhibitors persist over time and have an impact on the choice and efficacy of future treatment. The typical case is the fixed-dose combination of SOF/LDV. In an integrated analysis of various studies using this regimen, virological failure occurred in 2.4% of 2144 patients [25]. NS5A RASs were present in 74% of 51 patients at the time of failure. Most patients included had HCV GT1a. RASs were detected at several positions: M28T, Q30R, L31M, H58D, and Y93H (Table 1).

In some Asian countries such as Korea and Japan, the combination of ASV plus DCV is widely used for patients with HCV GT1b infection. This combination is not recommended for patients with GT1a or other genotypes, and it has not been approved in Europe or the United States (Table 2). The main reason is the impact that baseline NS5A RASs have on the SVR. In a pooled analysis including 979 patients with GT1b infection treated with ASV and DCV for 24 weeks, baseline RASs at positions L31F/I/M/V had a considerable effect on SVR rates: 42% in 43 patients with RASs versus 88% in 936 patients without RASs at these positions [26]. RASs in the NS3 protease did not affect SVR. These results indicate that patients with HCV GT1b with preexisting NS5A RASs at the L31 position should not be treated with this regimen. The rate of virological failures to ASV and DCV was 14% in 437 patients without cirrhosis and 16% in 206 patients with cirrhosis. A relevant finding of the studies analyzed was that two-thirds of patients experienced on-treatment breakthrough or no virological response at the end of treatment, whereas one-third relapsed [26]. In a multinational, phase-III, multicohort study with ASV and DCV the most frequent RASs selected at treatment failure in GT1b patients were at position D168 of the NS3 protease and at position Y93 of the NS5A complex which were present in approximately 75% of patients [27].

Treatment failure has been related to host factors, viral factors, and treatment-related factors. A combination of two or more factors is needed to result in treatment failure [28], as the presence of a single factor has a limited effect on the SVR. Among the host factors, interleukine 28B polymorpyism and cirrhosis are the most imporatntSeveral viral factors have been associated with a low SVR, such as GT1a, GT3, high viral load, and presence of some baseline RASs, but the impact of these factors depends on the DAA regimen used. The treatment-related factors include previous therapy, short therapy duration, and adherence (Fig. 1).

The genetic diversity of viral populations and resistance-associated variants has been traditionally assessed by population sequencing (ie, direct sequencing of a PCR product) or by clonal sequencing (ie, cloning of a PCR product into plasmid vectors, followed by direct sequencing). The lower limit of detection of direct sequencing ranges from 15% to 25% [29]. Clonal sequencing is more sensitive, but also more expensive and time-consuming. These limitations are overcome by a technique known as deep sequencing, which allows detection of minor variants present at frequencies as low as 0.5% [20].

At present, none of these techniques are commercially available in Europe, but a number of laboratories are able to perform sequence analysis of the NS3-4A, NS5A, and NS5B regions. In the United States, these tests are commercially available, but they are costly and may not be covered by all insurances companies. Currently, baseline RAS testing to individually tailor therapy is not recommended for treatment-naïve patients. In treatment-experienced patients with some genotypes (eg, GT1a) treated with certain combinations such as GZR and EBR, baseline RAS testing is recommended to decide on therapy duration and whether to include ribavirin in the regimen [3]. In patients who have failed DAA treatment, testing for resistance-associated variants that confer decreased susceptibility to NS3 protease inhibitors and to NS5A inhibitors is recommended for HCV GT1 patients (regardless of the subtype) who have compensated cirrhosis or a reason for urgent retreatment [3]. The results of these tests can help in selecting the retreatment drugs and therapy duration.