Prevalence of Drug Resistance Associated Substitutions in Persons With Chronic Hepatitis C Infection and Virological Failure Following Initial or Re-treatment With Pan-genotypic Direct-Acting Antivirals: A Systematic Review and Meta-analysis Authors: Seth Inzaule, Philippa Easterbrook, Ashley Latona, Nathan P Ford, William Irving, Philippa C Matthews, Marco Vitoria, Chris Duncombe, Amalia Giron, Suzanne McCluskey, Olufunmilayo Lesi, Serge Tchamgoue, Rachel Halford, Danjuma Adda, Emma Thomson, Geoff Dusheiko, Michael R Jordan Published: 2024 ABSTRACT AbstractBackgroundThe advent of short-course, curative treatment with direct-acting antivirals (DAA) has given promise for the global elimination of hepatitis C virus (HCV) infections by 2030. Virological failure occurs in 2%–12% of persons receiving curative DAA treatment and may be presaged by pre-existing polymorphisms or result from selection of drug resistant variants during therapy.MethodsWe conducted a systematic review to assess the prevalence of HCV resistance associated substitutions (RAS) among individuals with chronic hepatitis C infection who had virological failure following initial or re-treatment with pan-genotypic DAA regimens. We included 34 and 22 studies assessing RAS in people with virological failure published between January 2014 and July 2023. Pooled RAS prevalence was estimated using random-effects meta-analysis.ResultsThe pooled prevalence of RAS in people with virological failure following initial DAA treatment was 78.0% (95% confidence interval [CI]: 62.0–92.0) for sofosbuvir/velpatasvir, 81.0% (95% CI: 67.0–93.0) for sofosbuvir/daclatasvir, and 79.0% (95% CI: 70.0–87.0) for glecaprevir/pibrentasvir, with a high prevalence of resistance to the NS5A inhibitors. Among those with virological failure following re-treatment regimens, RAS were present in 93.0% (95% CI: 83.0–99.0) for sofosbuvir/velpatasvir/voxilepravir and in 100% (95% CI: 92.0–100) for glecaprevir/pibrentasvir, with resistance driven by RAS to NS5A inhibitors.DiscussionAt least 1 RAS is present in a high proportion of the few individuals with virological failure following initial or re-treatment with pan-genotypic DAA regimens. There is a need for ongoing surveillance for DAA-associated resistance, to assess risk factors for their development and clinical impact to inform best practice strategies for re-treatment. The introduction of oral direct-acting antivirals (DAA) in 2014, which provide potent and well-tolerated short-course curative treatment, has revolutionized hepatitis C virus (HCV) management and given promise for the global elimination of HCV infection [1]. World Health Organization (WHO) recommends dual-DAA combination therapy targeting at least 2 distinct non-structural proteins. The recommended pan-genotypic regimens for the initial treatment consist of the NS5B nucleoside polymerase inhibitor (NPI) sofosbuvir (SOF) in combination with either the NS5A replication complex inhibitors (NS5A) daclatasvir (DAC) or velpatasvir (VEL) or a dual combination of the NS5A inhibitor, glecaprevir (GLE) with the NS3/4A polymerase inhibitor (NS3/4A) PI pibrentasvir (PIB) [2, 3]. The pan-genotypic DAAs recommended by WHO are highly effective with cure rates (defined as sustained virological response at 12 weeks after completion of treatment) of generally >95%; however, around 2% to 12% experience virological failure [4–12]. Factors associated with treatment failure include prior treatment experience with interferon regimens, poor adherence to treatment, presence of cirrhosis, and pre-existing resistance-associated substitutions (RAS), or polymorphisms linked to certain subtypes which confer lower sensitivity to the existing DAAs [13, 14]. Virological failure is often associated with selection of RAS related to the drugs used in the therapeutic regimen [15–17]. Although there are concerns that RAS may limit retreatment options, some studies indicate that increasing the duration of treatment with the same regimen or alternative regimens may achieve cure [18–20]. Currently, only 1 pan-genotypic regimen (sofosbuvir [SOF]/velpatasvir [VEL]/voxilaprevir [VOX]) targeting 3 drug target sites is recommended by several professional societies for retreatment after failure of initial DAA regimens [2, 3, 21, 22]. The pan-genotypic combination GLE/PIB administered alone or in combination with SOF is also proposed for use in those with virological failure following use of SOF-containing regimens and or either a PI or an NS5A inhibitor (but not both) [2, 3]. Over the last 10 years, there have been considerable efforts to scale up HCV diagnosis and expand access to treatment in low- and middle-income countries (LMICs). As of 2022, WHO estimates that 36% of those with chronic hepatitis C infection have been diagnosed, and 20% treated. However, progress has been uneven, and diagnosis and treatment coverage is lower in low- and middle-income countries, especially in Sub-Saharan Africa (SSA) (<1%) and Southeast Asia (5%) [1]. The availability of generic DAAs, particularly the combinations of SOF/DAC and SOF/VEL, has helped improve access to treatment due to their lower cost [23]. Although treatment failure rates are low ranging from 2% to 12% [22], there are concerns that in settings and populations with a high burden of infection and/or where there are also ongoing high rates of transmission [24], such treatment failures may be consequential, particularly in the presence of clinically significant resistance [22]. Overall, the potential for selection of RAS during virological failure and possible subsequent onward transmission [15–17] in the context of limited options for retreatment regimens in LMICs needs attention. Information on prevalence of DAA-associated drug resistance mutations in those with virological failure following initial DAA treatment or re-treatment and their clinical significance remains scarce in resource-limited settings and is important to inform re-treatment and appropriate prevention and monitoring strategies for LMICs. We conducted a systematic review to determine the prevalence of RAS among persons with chronic hepatitis C infection with virological failure after initial or re-treatment with pan-genotypic DAA regimens. METHODS Search Strategy and Selection Criteria We conducted a systematic review by searching the available English language literature in PubMed in accordance with PRISM guidelines following a predefined study protocol [25]. The search terms and procedures are described in Supplementary Appendix pp1. Two authors (S. C. and A. L.) independently screened titles and abstracts from the search and subsequently screened the full texts of potentially eligible articles. We included prospective cross-sectional studies, prospective observational studies, and randomized trials but excluded case studies and studies that did not clearly report a denominator. We assessed the risk of bias for each record using a modified Johanna Briggs institute critical appraisal tool for systematic reviews of prevalence studies [26] (Supplementary Appendix). Our primary outcome was the prevalence of RAS following either treatment failure (ie, failure to achieve SVR after treatment duration) or breakthrough infection (ie, having detectable HCV RNA on treatment after initially developing undetectable HCV RNA levels during treatment). RAS was defined as any clinically relevant mutation based on EASL 2020 mutation list [22] or by the geno-2-pheno database [27]. To estimate the prevalence of RAS after initial treatment failure to WHO-recommended pan-genotypic regimens we included data from people experiencing virological failure defined as having viral breakthrough or viral relapse after treatment with either SOF/VEL, SOF/DAC, or GLE/PIB with or without co-administration of ribavirin. We excluded studies or subjects who had a history of prior DAA treatment but included those who had been pre-treated with pegylated interferon and/or ribavirin before starting DAA treatment. We also excluded studies that did not report RAS at time of failure. For assessment of the prevalence of RAS in people experiencing failure after retreatment, we included studies from people in whom prior DAA treatment had failed and who were being retreated with SOF/VEL/VOX or GLE/PIB. We excluded studies using SOF/VEL/VOX or GLE/PIB as initial treatment and those that did not report RAS information. We extracted RAS prevalence information, as well as study characteristics (eg, study design and region) and HCV genotypes in Microsoft Excel. Reported RAS were re-analyzed using the EASL 2020 drug resistance mutations list [22]. In cases where analyses had been done based on the geno-2-pheno resistance algorithm [27], no additional re-analysis was performed. In studies that also reported the presence of RAS present at low abundance, we only included RAS in our analysis if they were reported at a threshold of ≥15% as recommended by other treatment guidelines [21, 28] because the clinical relevance of low abundant RAS remains uncertain [29]. Data Analyses The proportions and corresponding 95% confidence intervals (CIs) of RAS were first transformed using Freeman–Tukey Double Arcsine transformation to stabilize on the variances due to sparse data [30]. The pooled prevalence was calculated using random effects meta-analysis because high between-study variability was expected. We evaluated between-study heterogeneity using T2 statistic. We also estimated pooled RAS prevalence by HCV genotype (GT) in cases where we had at least 2 studies. Supplementary Material