Background

WHO has called for research into predictive factors for selecting persons who could be successfully treated with shorter durations of antiviral therapy for Hepatitis C. We evaluated early virological response as a means of shortening treatment and explored host, viral and pharmacokinetic contributors to treatment outcome.

Methods

Duration of sofosbuvir and daclatasvir (SOF/DCV) was determined according to day 2 (D2) virologic response for HCV genotype (gt) 1- or 6-infected adults in Vietnam with mild liver disease. Participants received 4 or 8 weeks of treatment according to whether D2 HCV RNA was above or below 500 IU/ml (standard duration is 12 weeks). Primary endpoint was sustained virological response (SVR12). Those failing therapy were retreated with 12 weeks SOF/DCV. Host IFNL4 genotype and viral sequencing was performed at baseline, with repeat viral sequencing if virological rebound was observed. Levels of SOF, its inactive metabolite GS-331007 and DCV were measured on day 0 and 28.

Findings

Of 52 adults enrolled, 34 received 4 weeks SOF/DCV, 17 got 8 weeks and one withdrew. SVR12 was achieved in 21/34 (62%) treated for 4 weeks, and 17/17 (100%) treated for 8 weeks. Overall, 38/51 (75%) were cured with first-line treatment (mean duration of 37 days). Despite a high prevalence of putative NS5A-inhibitor resistance-associated substitutions (RAS), all first-line treatment failures were cured after retreatment (13/13). We found no evidence treatment failure was associated with host IFNL4 genotype, viral subtype, baseline RAS or DCV levels. SOF metabolite levels were higher in those failing 4-week therapy.

Interpretation

Shortened SOF/DCV therapy with retreatment if needed, reduces DAA use while maintaining high cure rates. D2 virologic response alone does not adequately predict SVR12 with 4 weeks of treatment.

At screening, HCV genotype and subtype were determined using NS5B, Core and 5’ UTR sequencing, according to the method described by Chau et. al. To evaluate the impact of HCV subtypes and resistance-associated substitutions on treatment outcome, whole genome sequencing (WGS) was additionally performed on all enrolled participants’ virus at baseline, and upon virological rebound and at start of retreatment in participants failing therapy. WGS of the HCV viral genome was attained using Illumina MiSeq platform. The de novo assemblies nucleotide sequences were translated into amino acid and were aligned to H77 HCV reference (GenBank ID: NC_038882.1) and the NS5A and NS5B protein regions were extracted.  We only looked for RAS that were present in at least 15% of the reads in the sample and had a read count of greater than 10.

We used the Public Health England (PHE) HCV Resistance Group’s definition for resistance-associated substitutions (RAS). For genotype 1, we looked for RASs defined specifically for genotype 1 as they are well studied. For genotype 6, we looked for all RASs defined across all genotypes, as little work has been done on RASs in genotype 6.  

For DCV, we looked for 24R, 28T, 30E/K/T, 31M/V, 32L, 58D, and 93C/H/N/R/S/W in genotype 1 infection and additionally looked for 28S, 30R and 31F in genotype 6 infection. For SOF, we looked for 159F, 237G, 282T, 315H/N, 321A/I in genotype 1 infection and additionally looked for 289I in genotype 6 infection.  

In addition to viral sequencing, we evaluated host genetic polymorphisms within the interferon lambda 4 (IFNL4) gene of all participants at baseline. Genotyping of IFNL4 rs368234815 was performed on host DNA using the TaqMan® SNP genotyping assay and primers described previously with Type‐it Fast SNP Probe PCR Master Mix (Qiagen).

The consensus viral sequences generated by the above method are provided in a fasta format data.