Prevalence of Y93H resistance-associated substitution in treatment-naïve hepatitis C virus genotype 1b patients: a systematic review and meta-analysis

Abstract

Background

Hepatitis C virus (HCV) genotype 1b infection remains a leading cause of chronic liver disease despite advances in direct-acting antivirals (DAAs). Resistance-associated substitutions (RASs) in the non-structural 5 A (NS5A) region, particularly the Y93H mutation, have been associated with reduced antiviral susceptibility and variable treatment outcomes. The global baseline prevalence of this substitution in treatment-naïve populations has not been systematically characterized.

Methods

A systematic review and meta-analysis were conducted in accordance with PRISMA 2020 guidelines and registered with PROSPERO. Observational and randomized studies reporting baseline NS5A Y93H prevalence in treatment-naïve HCV genotype 1b patients were identified through PubMed, Embase, and Web of Science. Data were synthesized using a random-effects model, and heterogeneity was quantified by the I² statistic. Sensitivity and publication bias analyses were performed using leave-one-out and DOI plot methods.

Results

Nine studies comprising 1,560 treatment-naïve patients from Asia and Europe, were included. The pooled prevalence of Y93H was 12% (95% CI, 6%–20%), with substantial heterogeneity (I² = 82.4%). Sensitivity analyses confirmed result stability, and minor asymmetry in the DOI plot suggested low publication bias.

Conclusion

The Y93H substitution is a frequent naturally occurring variant in untreated HCV genotype 1b infection. Despite reduced clinical relevance with pangenotypic DAAs, its persistence warrants continued monitoring. Standardized sequencing and global resistance surveillance remain essential for optimizing treatment strategies and informing antiviral policy.

Clinical trial number

Not applicable.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12879-026-14091-y.

Introduction

Hepatitis C virus (HCV) infection remains a major global health challenge despite the remarkable success of direct-acting antivirals (DAAs) that have transformed its management and prognosis [1, 2]. The widespread use of DAAs has markedly improved cure rates and reduced liver-related morbidity, yet the emergence and persistence of viral resistance continue to pose important challenges for sustaining these gains [3]. Resistance-associated substitutions (RASs) that arise naturally or under antiviral pressure may reduce treatment efficacy, especially in the absence of regimen tailoring or baseline resistance testing [4].

Among the targets of DAAs, the non-structural 5 A (NS5A) protein plays a central role in viral replication, assembly, and interaction with host-cell pathways [5]. NS5A inhibitors are key components of multiple therapeutic combinations, valued for their potency and favourable tolerability profiles.⁵ However, their effectiveness can be compromised by baseline RASs that diminish inhibitor binding or replication suppression [6]. Of particular interest is the Y93H substitution in the NS5A gene, a single amino acid change that has been consistently linked to reduced susceptibility to several NS5A inhibitors, including daclatasvir, ledipasvir, velpatasvir, and elbasvir, especially in genotype 1b (GT-1b) infection [7, 8].

The Y93H variant is regarded as one of the most clinically consequential resistance-associated substitutions in HCV therapy.⁸ Its presence has been linked to diminished virologic response and relapse with earlier DAA regimens, particularly when occurring alongside additional NS5A mutations such as L31M or P58S [9–11]. Although newer pan-genotypic inhibitors demonstrate greater resilience to resistance, the persistence of Y93H within circulating viral populations continues to warrant attention for both clinical surveillance and baseline testing [12]. Notably, the distribution of Y93H is highly variable across regions and study populations, reflecting not only underlying viral genetic diversity but also differences in sequencing platforms, detection thresholds, and analytic methodologies. Conventional population sequencing captures only dominant variants, whereas next-generation sequencing (NGS) can detect minor quasispecies that may lack clinical relevance yet elevate apparent prevalence rates [13]. The absence of standardized assay thresholds and uniform reporting frameworks further amplifies inconsistencies across studies, underscoring the need for harmonized evaluation and systematic synthesis of available evidence.

Evidence on the NS5A Y93H substitution in treatment-naïve HCV GT-1b infection remains scattered and methodologically inconsistent. Many studies have mixed genotypes, combined naïve and treated cohorts, or lacked defined sequencing thresholds, limiting clarity on true baseline resistance [7, 14, 15]. Most reports are region-specific and small in scope, offering an incomplete view of global variation. These inconsistencies have hindered data synthesis and delayed standardized approaches to resistance surveillance. Establishing reliable prevalence estimates in untreated GT-1b populations is therefore vital to guide testing strategies, optimize therapy selection, and strengthen long-term monitoring of antiviral resistance.

Despite the clinical and epidemiologic significance of this substitution, no prior systematic review has comprehensively synthesized its global prevalence in treatment-naïve HCV GT-1b patients. Previous meta-analyses addressing NS5A resistance have either included mixed genotypes or lacked stratification by detection method, thereby obscuring genotype-specific patterns. Addressing these gaps is essential to generate standardized and interpretable evidence that can guide both clinical decision-making and resistance monitoring initiatives. Accordingly, the present study was undertaken to provide a rigorous and comprehensive synthesis of available data on the prevalence of NS5A Y93H resistance-associated substitution in treatment-naïve HCV GT-1b infection. By consolidating the evidence base and quantifying methodological influences, this study aims to clarify the global landscape of NS5A Y93H resistance and strengthen the foundation for ongoing surveillance, clinical interpretation, and antiviral policy planning in HCV management.

Methods

This systematic review and meta-analysis included observational studies and randomized controlled trials assessing the NS5A Y93H substitution in treatment-naïve HCV GT-1b infection. The study followed PRISMA 2020 guidelines [16] and was prospectively registered with PROSPERO. This approach ensured methodological transparency and contextual relevance in synthesizing resistance data across diverse settings.

Eligibility criteria

Eligible studies enrolled DAA naïve individuals with chronic hepatitis C and reported the baseline prevalence of the NS5A Y93H substitution in genotype 1b. We included randomized clinical trials, prospective or retrospective cohort studies, and case–control designs that confirmed genotype 1b using standardized HCV diagnostic methods and provided extractable data to compute a prevalence proportion, as the analysis focused strictly on HCV genotype 1b infection. We included full-text articles published in English, with studies searched from database inception through 2025 (Table S1). We excluded studies that involved previously treated patients; HIV and HBV co-infections; mixed-genotype cohorts without separable genotype 1b data; assessments of mutations other than NS5A Y93H or pooled RAS results without site-specific estimates; reports lacking primary data (editorials, commentaries, reviews), case reports or case series, qualitative studies, and any article without an accessible full text or not published in English.

Search strategy

A comprehensive literature search was performed across PubMed, Embase, and Web of Science to identify studies evaluating the NS5A Y93H resistance-associated substitution in treatment-naïve HCV genotype 1b infection. The strategy combined controlled vocabulary and free-text terms related to Hepatitis C virus, viral drug resistance, and mutations, together with specific keywords “Y93H” and “tyrosine 93.” Boolean operators were applied to optimize sensitivity and specificity, enabling comprehensive retrieval of relevant experimental and clinical evidence on resistance mutations across NS5A, NS3, and NS5B regions (Table S2).

Screening and data extraction

The process of study selection was carried out in the Nested Knowledge (NK) platform, which allows transparent tracking of all review stages. Screening was conducted in two successive phases following the predefined protocol. In the first phase, two independent reviewers examined the titles and abstracts of all identified records to exclude clearly irrelevant reports. During the second phase, the same reviewers assessed the full-text versions of potentially eligible articles to confirm inclusion according to the established criteria. Any differences in judgment were discussed and resolved by consensus; when disagreement persisted, a third reviewer provided an independent opinion to reach a final decision. Data extraction was also performed within the NK platform using a structured spreadsheet that had been developed before analysis began. The extraction form captured detailed information on study design, country, sample size, participant characteristics, detection approach, genotype confirmation, and reported prevalence of the NS5A Y93H variant. Each entry was checked by both reviewers to ensure internal consistency and accuracy before data synthesis.

Quality assessment

The methodological quality and risk of bias of the included studies were appraised independently by two reviewers, using tools appropriate to each study design. For observational studies including cross-sectional, cohort, and case–control designs the Newcastle–Ottawa Scale (NOS) was applied [17]. This instrument evaluates three domains: the adequacy of participant selection, comparability of study groups, and accuracy of exposure or outcome assessment. Each study could achieve a maximum of nine points; scores of 7–9 indicated high quality, 4–6 moderate quality, and fewer than 4 low quality. For randomized controlled trials, the Cochrane Risk of Bias 2.0 tool was used [18]. This framework assesses potential bias in randomization processes, deviations from intended interventions, completeness of outcome data, measurement reliability, and selective reporting. Any differences between reviewers were discussed, and consensus was reached before final scoring to ensure objectivity and methodological rigor. Detailed quality assessment rating showed in Table S3.

Statistical analysis

All statistical analyses were carried out using R software (v4.3.0). Quantitative synthesis employed the “meta” packages, which are widely used for meta-analytical computations in biomedical research. The pooled prevalence of the NS5A Y93H resistance-associated substitution was estimated as a proportion with corresponding 95% confidence intervals (CIs) [19]. Because methodological and regional variability was anticipated across studies, a random-effects model was applied to generate summary estimates that account for both within- and between-study variance. The extent of statistical heterogeneity was quantified using the I² statistic, with conventional thresholds of approximately 25%, 50%, and 75% interpreted as representing low, moderate, and high heterogeneity, respectively [20]. When substantial heterogeneity was identified, subgroup analyses and sensitivity checks were planned to explore potential sources of variation. Publication bias was assessed visually using Doi plots, which provide a more balanced representation of asymmetry than traditional funnel plots [21]. All analyses were conducted using two-tailed tests, and a p value < 0.05 was considered statistically significant.

Results

Study selection

A total of 625 records were retrieved through electronic database searches, comprising 159 from PubMed, 332 from Embase, and 134 from Web of Science (Fig. 1). Prior to screening, 178 duplicate records, 72 records identified as ineligible by automated tools, and 50 records removed for other reasons were excluded. Following deduplication, 325 unique records remained for title and abstract screening. Of these, 250 records were excluded for not meeting the predefined eligibility criteria. Full-text articles were sought for 50 reports; however, 34 could not be retrieved owing to restricted access or incomplete publication details. The remaining 16 reports underwent full-text evaluation, during which 7 studies were excluded five for not reporting the outcome of interest and two for unavailability of full text. Ultimately, nine studies fulfilled all inclusion criteria and were incorporated into the systematic review and meta-analysis.

Fig. 1.

PRISMA flow chart showing study inclusion process

Summary of study characteristics

Nine studies published from 2012 to 2021 were finally included in this review and meta-analysis (Table 1). Each study evaluated treatment-naïve patients carrying hepatitis C virus genotype 1b. The study populations came from Japan, China, Italy, Ireland, and several multinational settings. Among the included reports, five used a cross-sectional design, two were cohort investigations, and two were randomized clinical trials. The number of participants in individual studies ranged between 46 and 435, reflecting variability in study scope and population characteristics. The detection of the NS5A Y93H substitution was performed using different sequencing techniques, such as Sanger sequencing, population-based assays, and next-generation or deep sequencing. Studies that applied deep sequencing methods, including those by Fujii et al. [22] and Li et al. [23], identified minor resistant variants with greater precision. Randomized trials, such as the EGALITE trial by Hagiwara et al. [24] and the AVIATOR study by Krishnan et al. [25], provided robust baseline data generated through validated resistance-testing protocols.

Table 1.

Summary characteristics of included studies

Study	Study Design	Country	Sample Size	Male percentage	Method Used to Assess Study Outcome	NOS score
Bertoli et al. 2018 [43]	Cross-sectional	Italy	435	Not specified	Population based sequencing of NS3/NS5A/NS5B regions	Moderate
Fujii et al. 2015 [22]	Cross-sectional	Japan	296	46.6%	Deep next generation sequencing of NS5A region	High
Hagiwara et al. 2019 [24]	Randomized controlled trial	Japan	82	39%	Sanger and next generation sequencing (NGS) of NS5A region	High
Ikeda et al. 2014 [7]	Observational cohort	Vietnam	59	Not specified	Sanger sequencing of NS5A domain I (pre-treatment)	Moderate
Krishnan et al. 2015 [25]	Randomized controlled trial	Japan	130	Not specified	Population sequencing of NSA5 region	High
Li et al. 2017 [23]	Cross-sectional	China	70	51.4%	Deep sequencing of NSA5 gene region	High
Nguyen et al. 2016	Cross-sectional	China	46	Not specified	Population sequencing of NS3/NSA5 regions	Low
Suzuki et al. 2012 [44]	Observational	Ireland	294	Not specified	Sanger sequencing of NS5A/B regions	Moderate
Wang et al. 2015 [45]	Cross-sectional	Japan	148	55.4%	Sanger sequencing of NSA5 region	Moderate

Prevalence of Y93H resistance-associated substitution

Across nine studies including 1,560 treatment-naïve patients with hepatitis C virus genotype 1b, the pooled prevalence of the NS5A Y93H resistance-associated substitution among treatment-naïve patients with hepatitis C virus genotype 1b was 12% (95% CI, 6%–20%), based on data from nine studies comprising 1,560 participants (Fig. 2). The 95% prediction interval ranged from 0% to 36%, indicating considerable variability in expected prevalence across future studies. Reported estimates in individual studies varied widely, from 5% (95% CI, 2%–11%) to 36% (95% CI, 24%–49%), with most studies reporting rates between 7% and 15%. Substantial between-study heterogeneity was observed (I² = 82.4%; P <.001), likely reflecting differences in patient populations, geographic origin, and sequencing methodology. Study weights were relatively balanced, ranging from 9% to 13%, suggesting that no single study disproportionately influenced the pooled estimate.

Fig. 2.

Forest plot showing prevalence of the Y93H NS5A resistance-associated substitution among hepatitis C virus

A leave-one-out sensitivity analysis was conducted to assess the influence of individual studies on the pooled prevalence. Excluding any single study did not significantly alter the summary estimate, which remained at 12% (95% confidence interval: 6–20%). Heterogeneity values (I²) varied minimally between 67% and 85%. These results indicate that the overall findings were stable and not driven by any individual study (Fig. 3).

Fig. 3.

Leave-one-out sensitivity analysis for prevalence of the Y93H NS5A resistance-associated substitution among hepatitis C virus

In the subgroup analysis by geographic region, the prevalence of hepatitis C varied across settings. Studies from Europe (2 studies; n = 729) reported a pooled prevalence of 9%, whereas studies from Asia (7 studies; n = 831) demonstrated a higher pooled prevalence of 14%. These findings indicate notable regional variation in pooled prevalence estimates (Table 2).

Table 2.

Sub-group analysis of hepatitis c patients

Sub group	Type	No. of studies	Sample size (N)	Pooled prevalence (95% CI)	Heterogeneity	P Value
Geographic region	Europe	2	729	0.09 (0.07–0.12)	NA	0.38
	Asia	7	831	0.14 (0.07–0.22)	86%	< 0.001

Publication bias

Assessment of publication bias using a DOI plot showed a nearly symmetrical distribution of effect sizes around the pooled estimate, with an LFK index of 1.76, indicating minor asymmetry (Fig. 4). This suggests a low likelihood of significant publication bias, although some small-study effects may be present. Overall, the observed asymmetry was within acceptable limits and did not materially affect the robustness of the meta-analytic findings.

Fig. 4.

DOI plot assessing publication bias in included studies

Discussion

In this systematic review and meta-analysis of nine studies including 1,560 treatment-naïve patients infected with HCV genotype 1b, the pooled prevalence of the NS5A Y93H resistance-associated substitution (RAS) was estimated at 12%. Considerable between-study heterogeneity (I² = 82.4%) underscores the variability in underlying viral populations and detection methods. Sensitivity analysis demonstrated that the summary estimate was robust to the exclusion of any individual study, confirming the stability of the pooled prevalence. Taken together, these findings provide a quantitative synthesis of the global distribution of Y93H in untreated GT-1b infection, highlighting its persistent presence even before antiviral exposure.

The prevalence estimate obtained in this meta-analysis was consistent with reports from Japan and other East Asian regions, where baseline Y93H occurrence has been described as relatively common among untreated genotype 1b infections [26]. Earlier studies conducted before the widespread use of DAAs reported similar figures, suggesting that this substitution likely represents a stable polymorphism within certain viral lineages rather than a transient, treatment-driven mutation [27–29]. Reports from European and North American cohorts have documented lower prevalence rates, often below, reflecting population-specific viral diversity and differential selective pressures.

Previous studies that combined NS5A resistance-associated substitutions across different genotypes often overlooked the distinct molecular and epidemiologic patterns unique to genotype 1b [30, 31]. By restricting the analysis to treatment-naïve populations with confirmed genotype 1b infection, the present study provides a more accurate characterization of baseline NS5A resistance in this clinically relevant subgroup [31, 32]. The findings are consistent with prior investigations by Ikeda and Fujii, which documented comparable frequencies of the Y93H variant using deep-sequencing approaches capable of detecting minor viral populations. The inclusion of contemporary randomized clinical trials, such as the EGALITE study and the AVIATOR trial, further reinforces the validity of these observations through the use of standardized genotyping methods aligned with current diagnostic practice [33–35]. Collectively, this synthesis refines the existing evidence base by delineating genotype-specific resistance patterns and underscoring the reproducibility of Y93H detection across diverse analytical platforms and geographic settings. The persistence of Y93H in treatment-naïve individuals suggests that it may confer limited fitness cost to the virus, allowing it to circulate in the absence of drug pressure. The substitution occurs within domain I of the NS5A protein, a region critical for replication complex formation and interaction with host factors [9].

The observed geographic heterogeneity may also reflect founder effects or population-specific transmission patterns. East Asian lineages of GT-1b exhibit higher baseline Y93H frequencies, possibly due to historical clustering of resistant quasispecies that achieved stable circulation. Moreover, differences in sequencing depth and reporting thresholds can influence apparent prevalence [36, 37]. Deep or next-generation sequencing (NGS) platforms detect minority variants at levels below 1%, whereas conventional Sanger sequencing typically identifies substitutions present in at least 15%–20% of viral populations [38]. Studies using NGS therefore tend to report higher frequencies of Y93H, even when many of these low-abundance variants may not carry phenotypic significance. Such methodological variation likely contributes to the wide prediction interval observed in this analysis.

Although the advent of pangenotypic direct-acting antivirals has substantially diminished the clinical relevance of single resistance-associated substitutions, baseline NS5A variants continue to warrant consideration in the management of genotype 1b infection. The prevalence estimated in this analysis indicates that a meaningful proportion of untreated patients carry the Y93H substitution before initiation of therapy [9, 12]. In regimens primarily dependent on NS5A inhibition, particularly earlier combinations such as daclatasvir with asunaprevir or ledipasvir with sofosbuvir, the presence of Y93H has been linked to lower rates of sustained virologic response and an increased risk of post-treatment relapse [39, 40]. Although newer pangenotypic combinations, including glecaprevir with pibrentasvir and sofosbuvir with velpatasvir, demonstrate higher genetic barriers to resistance, baseline Y93H may still influence regimen selection, treatment duration, or the need for intensified therapeutic monitoring in certain clinical contexts [41, 42].

Understanding the population-level distribution of the Y93H substitution holds important implications for public health planning and antiviral resistance control. In regions where this variant occurs more frequently, incorporating targeted genotypic screening before therapy initiation may improve regimen selection, particularly where first-generation NS5A inhibitors remain in use [9]. Ongoing surveillance of naturally occurring resistance-associated substitutions can also elucidate the molecular epidemiology of HCV genotype 1b infection and inform tracking of resistance evolution as antiviral coverage widens. Adoption of standardized sequencing protocols and harmonized reporting frameworks would enhance cross-study comparability and strengthen the evidence base for global treatment recommendations.

The strength of this meta-analysis lies in its focused design and methodological rigor, which allowed a clear estimation of baseline NS5A resistance in treatment-naïve HCV genotype 1b infection. Restricting inclusion to untreated populations minimized confounding from prior antiviral exposure, while the consistent use of a random-effects model and sensitivity analyses strengthened the reliability of the pooled prevalence. Geographic representation was uneven, with a predominance of data from East Asia and limited evidence from other regions. In addition, restricting inclusion to English-language publications may have excluded potentially relevant studies, introducing a degree of selection bias. Despite these constraints, the synthesis provides the most comprehensive and methodologically coherent estimate to date of baseline Y93H prevalence in treatment-naïve HCV genotype 1b infection, establishing a foundation for future global surveillance and resistance research.

Future investigations should emphasize prospective, geographically representative studies that apply standardized deep-sequencing methods and harmonized reporting criteria to ensure comparability across settings. Incorporating real-world data from national and international surveillance systems may further refine prevalence estimates and reveal temporal shifts in the circulation of resistant variants as antiviral access expands. Experimental research exploring the replication fitness, persistence, and transmissibility of Y93H in the absence of therapeutic pressure is also needed to clarify its virologic and epidemiologic relevance. In addition, linking prevalence data with treatment outcome analyses would help define clinically meaningful thresholds for baseline resistance testing and guide the selective use of pretreatment genotyping in contemporary care.

Conclusion

This systematic review and meta-analysis provide the most comprehensive synthesis to date of baseline NS5A Y93H resistance among treatment-naïve patients with HCV genotype 1b. The pooled prevalence of approximately 12% indicates that this substitution remains a common naturally occurring variant, independent of antiviral exposure. Despite substantial heterogeneity across regions and sequencing platforms, the findings consistently highlight the persistence of Y93H within global viral populations. Although the advent of pangenotypic direct-acting antivirals has mitigated the clinical impact of individual resistance variants, the presence of Y93H continues to warrant consideration in regimen selection, particularly in regions where older NS5A inhibitors are used. Standardized deep-sequencing methods, harmonized reporting, and global surveillance networks are essential to refine prevalence estimates and clarify the clinical threshold for baseline resistance testing in contemporary hepatitis C management.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

HCV

Hepatitis C virus

DAA

Direct-acting antiviral

RAS

Resistance-associated substitution

NS5A

Non-structural protein 5 A

GT-1b

Genotype 1b

CI

Confidence interval

Author contributions

Conceptualization: D. K. K., D. Y. L., P. S. Study design and analytical framework: N.K., R. K., P. S. Data curation and validation: D. K. K., D. Y. L., S. L., S. D., S. N. K. Methodology refinement and interpretation: R. A., A. M. G. Critical review and guidance: S. D., S. K., D. K. K. Manuscript drafting: M. S., V. K., P. S. Manuscript revision and finalization: All authors.

Funding

This study received no funding.

Data availability

All data generated or analyzed during this study are included in this published article (and its Supplementary information files).

Declarations

Ethics approval and consent to participate

Not applicable, as there were no human participants involved in this study.

Consent for publication

Not applicable, as this study does not involve any individual person’s data in any form.

Competing interests

The authors declare no competing interests.