What do we need to know about RAVs clinically?

14 th European HIV & Hepatitis Workshop Rome, 25-27 May, 2016 What do we need to know about RAVs clinically? Stefan Zeuzem, MD University of Frankfurt Germany

Background Resistance associated variants (RAVs) or resistance associated polymorphisms (RAPs) with reduced drug susceptibility are observed in patients with chronic HCV Due to the high replication rate of HCV (10 12 / day) Error-prone RNA-dependent RNA polymerase (10-4 to 10-5 per copied nucleotide) All possible single and double mutants are predicted to be generated multiple times each day Not all RAVs are fit enough to persist (but potentially under selection pressure of a respective DAA). Most NS3A RAVs disappear after EoT (HCV1b > HCV1a), while many NS5A RAVs persist The level of resistance in vivo may differ between a RAV in previously untreated patient and the same RAV selected by a DAA (= TEV) (e.g. due to compensatory mutations) Rong et al., Sci Transl Med 2010

How to define a RAV? RAV per specific drug or drug class (e.g. resistance against elbasvir or any NS5A inhibitor) Genotypic backbone in vitro and in vivo (e.g. genotype 1a or 1b replicons, other genotypes) Group EBR RAVs GT K24 M28 (L28) 1a - A, G, T Q30 (R30)* Position L31 P32 S38 H58 (P58) A92 Y93 D, E, H, G K, L, R F, M, V - - D - C, H, N, S 1b - Any Any Any L - D K Any NS5A Class RAVs 1a or 1b Any Any Any Any Any Any Any Any Any Wild type amino acid at position 28 is M in GT1a and L in GT1b *Wild-type amino acid at position 30 is Q in GT1a and R in GT1b. Wild type amino acid at position 58 is H in GT1a and P in GT1b No EBR RAVs identified at this position Note that EBR RAVs are a subset of the NS5A class RAVs Zeuzem et al., Ann Intern Med 2015 Jacobson et al., AASLD 2016

How to define a RAV? RAV per specific drug or drug class (e.g. resistance against ledipasvir or any NS5A inhibitor) Genotypic backbone in vitro and in vivo (e.g. genotype 1a or 1b replicons, other genotypes) Level of reduced susceptibility in vitro (e.g. any aa change vs >2.5-fold vs >100-fold)

NS5A RAVs - Overview GT 1a GT1b M28T Q30R L31M/V Y93H/N L31V Y93H/N Ledipasvir >20x >100x >100x >1,000x >1,000x/- Ombitasvir >1,000x >100x Daclatasvir >100x >1,000x Elbasvir >10x >10x <3x >100x >100x >1000x >10x >100x Velpatasvir <10x <3x >20x >10,000x <10x >50x >1,000x >10,000x >20x >20x >100x >1,000x Odalasvir >20x <10x <3x >1,000x <3x ABT-530 <3x <3x <3x <10x <3x <3x >100x <10x >10x/- <3x/- MK-8408 <10x <10x <10x <10x <10x <10x Cheng et al., EASL 2012; Wong et al., AAC 2013; Krishnan et al., AAC 2015; Fridell et al., Hepatology 2011; Liu et al., AAC 2015; Cheng et al., EASL 2013; Patel et al., EASL 2015; Ng et al., CROI 2014; Asante-Appiah et al., AASLD 2014 <3x <10x EC 50 fold-change compared to WT replicon

How to define a RAV? RAV per specific drug or drug class (e.g. resistance against ledipasvir or any NS5A inhibitor) Genotypic backbone in vitro and in vivo (e.g. genotype 1a or 1b replicons, other genotypes) Level of reduced susceptibility in vitro (e.g. any aa change vs >2.5-fold vs >100-fold) Methodology of RAV detection (sensitivity threshold) (e.g. deep sequencing with a cut-off at 1% or populations sequencing with a cut-off at 15-25%)

Methodology of RAV detection GT1a N=3509 GT1b N=1887 30 30 Class RAVs Prevalence (%) 20 10 Class RAVs LDV RAVs Y93H Prevalence (%) 20 10 LDV RAVs Y93H 0 0 5 10 15 20 25 30 35 40 45 50 Sensitivity Threshold (%) 0 0 5 10 15 20 25 30 35 40 45 50 Sensitivity Threshold (%) GT1a Position LDV RAVs NS5A Class RAVs 24 G/N/R G/N/R 26 E 28 A/G/T A/G/T/V 30 E/G/H/L/K/R/T C/E/G/H/I/L/K/R/S/T/Y 31 I/F/M/V I/F/M/V 32 L L 38 F F 58 D D/L 92 K/T K/T 93 C/F/H/N/S C/F/H/L/N/R/S/T/W GT1b Position LDV RAVs NS5A Class RAVs 28 M 31 I/F/M/V I/F/M/V 32 L L 58 D D 92 K K 93 C/H/N/S C/H/N/S Date on file, Gilead Hongmei Mo, personal communication

Frequency and characteristics of RAVs in treatment-naive and DAA-experienced patients - European RAVs database: Serum samples of 3549 European HCV-infected patients Population-based sequencing for NS3, NS5A, and NS5B Considered relevant if associated with treatment failure or shown to confer >2-fold changed drug susceptibility in comparison to reference strain RAVs in SOF/PR and SOF/RBV failures No RAVs detected in G1a or G2 G1b Frequency of RAVs* NS5B RAVs in DAA-naive pts (53%) RAVs in SOF/PR and SOF/RBV pts (61%) 8 6 2 16% DAA-naive (n=706) SOF/PR; SOF/RBV (n=55) L159F 32% C316H/N EC50 fold change: 1.6 (L159F together with C316N) EC50 fold change: 1.3 Dietz J, et al. EASL 2016, Barcelona. #PS007 *Double variants were calculated as single RAVs. Thus, sum of single RAVs frequencies is not identical with rate of patients with RAVs. 55% G3 NS5B RAVs in DAA-naive pts () RAVs in SOF/PR and SOF/RBV pts (11%) 8 6 2 DAA-naive (n=313) SOF/PR; SOF/RBV (n=41) L159F 11%

Frequency and characteristics of RAVs in treatment-naive and DAA-experienced patients - European RAVs database: NS3 G1a Frequency of RAVs* Frequency of RAVs* 8 6 2 8 6 2 RAVs in DAA-naive pts (36%) RAVs in SMV/SOF pts (82%) 34% RAVs in SMV/SOF failures <1% 1% 3% <1% <1% 1% 3% 1% F43V Q80K 59% DAA-naive (n=706) Q80K/R 41% 1% <1% <1% R155 G/K/T 44% EC50 fold change: 10->100 Dietz J, et al. EASL 2016, Barcelona. #PS007 S122R DAA-naive (n=766) SMV/SOF (n=18) A156T/V SMV/SOF (n=33) R155K D168E/V EC50 fold change: 10 20 100 20 100 NS3 G1b RAVs in DAA-naive pts (3%) RAVs in SMV/SOF pts (44%) A156 S/TV 24% D168 E/N/V Frequency of RAVs 8 6 2 EC50 fold change: *Double variants were calculated as single RAVs. Thus, sum of single RAVs frequencies is not identical with rate of patients with RAVs. G4 Q80R 25% NS3 RAVs in DAA-naive pts (4%) RAVs in SMV/SOF pts (75%) DAA-naive (n=99) SMV/SOF (n=4) 4% D168E/V no data in G4 5

Frequency of RAVs* NS5A DCV/SOF 10 8 6 2 RAVs in DAA-naive pts (13%) RAVs in DCV/SOF pts (83%) 67% RAVs in DCV/SOF versus LDV/SOF failures (G1) 25% 17% 8% 1% 6% 8% 4% 3% 8% <1% <1% 1% K24R M28 A/T/V Q30 E/H/L/R L31M DAA-naive (n=766) DCV/SOF (n=14) H58D A92T Y93 C/F/N EC50 fold change: 2 10 20 100 2 100 >100 Dietz J, et al. EASL 2016, Barcelona. #PS007 G1a *Double variants were calculated as single RAVs. Thus, sum of single RAVs frequencies is not identical with rate of patients with RAVs. 10 8 6 2 RAVs in DAA-naive pts (13%) RAVs in LDV/SOF pts (82%) 59% NS5A LDV/SOF 34% 1% 2% 6% 7% 4% 3% 9% 2% <1% 2% <1% 1% K24R M28T/V Q30 E/H/L/R L31M DAA-naive (n=766) LDV/SOF (n=47) EC50 fold change: >1000 >100 >100 >100 >100 Frequency of RAVs* Frequency and characteristics of RAVs in treatment-naive and DAA-experienced patients - European RAVs database: 10 8 6 2 <1% <1% L28V RAVs in DAA-naive pts (2) RAVs in DCV/SOF pts (82%) DAA-naive (n=706) R30H DCV/SOF (n=14) 73% 9% 4% 7% L31 F/I/M/V A92T Y93H 82% G1b 10 8 6 2 <1% <1% L28V R30H 3% P32L H58D A92T RAVs in DAA-naive pts (2) RAVs in LDV/SOF pts (93%) DAA-naive (n=706) LDV/SOF (n=41) 9% 4% 3% 7% L31 F/M/V A92T Y93H Y93 C/F/N 83%

Frequency and characteristics of RAVs in treatment-naive and DAA-experienced patients - European RAVs database: Frequency of RAVs* 10 8 6 2 Dietz J, et al. EASL 2016, Barcelona. #PS007 RAVs in DCV/SOF versus LDV/SOF failures (G3 & 4) NS5A DCV/SOF RAVs in DAA-naive pts (9%) RAVs in DCV/SOF pts (95%) DAA-naive (n=313) DCV/SOF (n=21) 7% 11% 2% A30K Y93H 84% G3 G4 10 8 6 2 NS5A LDV/SOF RAVs in DAA-naive pts (9%) RAVs in LDV/SOF pts (7%) DAA-naive (n=313) LDV/SOF (n=15) 7% 7% 2% EC50 fold change: >1000 no data Frequency of RAVs* 10 8 6 2 RAVs in DAA-naive pts (66%) RAVs in DCV/SOF pts (10) DAA-naive (n=99) DCV/SOF (n=2) 2% L28M 84% L30R/S 10 10 8 6 2 2% L28M EC50 fold change: A30K DAA-naive (n=99) 64% 57% Q30R M31V 29% Y93H RAVs in DAA-naive pts (66%) RAVs in LDV/SOF pts (86%) LDV/SOF (n=9) Y93C/H 43% 2 10 >100

Frequency and characteristics of RAVs in treatment-naive and DAA-experienced patients - European RAVs database: RAVs in 3D failures Dietz J, et al. EASL 2016, Barcelona. #PS007

How to define a RAV? RAV per specific drug or drug class (e.g. resistance against ledipasvir or any NS5A inhibitor) Genotypic backbone in vitro and in vivo (e.g. genotype 1a or 1b replicons, other genotypes) Level of reduced susceptibility in vitro (e.g. any aa change vs >2.5-fold vs >100-fold) Methodology of RAV detection (sensitivity threshold) (e.g. deep sequencing with a cut-off at 1% or populations sequencing with a cut-off at 15-25%) Impact of RAVs on SVR rates (e.g. combination partner, treatment duration, stage of disease)

Simeprevir + Sofosbuvir in GT1a-infected patients (TN+TE): Impact of the NS3 RAV Q80K 100 80 SVR w/o Q80K SVR with Q80K 97 96 SVR (%) 60 40 20 0 SMV + SOF Tx duration 12 weeks Cirrhosis No Q80K frequency Kwo et al., OPTIMIST-1, EASL 2015; Lawitz et al., OPTIMIST-2, Hepatology 2015

Simeprevir + Sofosbuvir in GT1a-infected patients (TN+TE): Impact of the NS3 RAV Q80K SVR w/o Q80K SVR with Q80K 100 80 84 97 96 SVR (%) 60 40 73 20 0 SMV + SOF SMV + SOF Tx duration 8 weeks 12 weeks Cirrhosis No No Q80K frequency 42% Kwo et al., OPTIMIST-1, EASL 2015; Lawitz et al., OPTIMIST-2, Hepatology 2015

Simeprevir + Sofosbuvir in GT1a-infected patients (TN+TE): Impact of the NS3 RAV Q80K SVR w/o Q80K SVR with Q80K SVR (%) 100 80 60 40 84 73 97 96 92 74 20 0 SMV + SOF SMV + SOF SMV + SOF Tx duration 8 weeks 12 weeks 12 weeks Cirrhosis No No Yes Q80K frequency 42% 47% Kwo et al., OPTIMIST-1, EASL 2015; Lawitz et al., OPTIMIST-2, Hepatology 2015

SVR12 Rates by Treatment Regimen (LDV/SOF)and Duration: Patients without Cirrhosis With NS5A RAVs No NS5A RAVs 100 98 99 99 99 99 90 80 SVR12 (%) 60 40 20 0 30/32 107/108 187/ 189 504/509 79/88 298/300 LDV/SOF 8 Weeks LDV/SOF 12 Weeks LDV/SOF 12 Weeks TN < 6M TN TE Studies included for analysis: LDV/SOF 8 weeks: GS-US-337-0118 (LONESTAR 1), GS-US-337-0108 (ION-3); LDV/SOF 12 Wks TN: GS-US-GS-US-334-1274 (Bleeding Disorder), GS-US-337-0102 (ION-1), GS-US-337-0108 (ION-3), GS-US-337-0113 (Japan 1), GS-US-337-0115 (ION-4), GS-US-337-0122 (Electron 2), GS-US-337-0131(China), GS-US-337-0118 (LONESTAR 1), GS-US-337-1406, GS-US-337-1468 (LEPTON); LDV/SOF 12 Wks TE: GS-US-337-0109 (ION-2), GS-US-337-0113 (Japan 1), GS-US-337-0115 (ION-4), GS-US-337-0124 (SOLAR-2), GS-US-334-1274 (Bleeding Disorder), GS-US-337-0118 (LONESTAR 1), GS-US-337-0131(China), GS-US-337-1406, GS-US-337-1468 (LEPTON) Sensitivity threshold at 1% (deep sequencing) Zeuzem et al., AASLD 2015

SVR12 Rates by Treatment Regimen (LDV/SOF) and Duration: TN Patients with Cirrhosis With NS5A RAVs No NS5A RAVs SVR12 (%) 100 80 60 40 20 0 96 96 100 100 100 88 26/27 65/68 10/10 27/27 8/9 19/19 LDV/SOF LDV/SOF+RBV LDV/SOF 12 weeks 24 weeks Studies included for analysis: LDV/SOF 12 Wks: GS-US-334-1274 (Bleeding Disorder), GS-US-337-0102 (ION-1), GS-US-337-0113 (Japan 1), GS-US-337-0115 (ION-4), GS-US-337-0122 (ELECTRON-2), GS-US-337-0131(China), GS-US-337-1406; LDV/SOF+RBV 12 Wks: GS-US-337-0102 (ION-1), GS-US-337-0113 (Japan 1), GS-US-337-0122 (ELECTRON-2); LDV/SOF 24 Wks: GS-US-337-0102 (ION-1), GS-US-334-1274 (Bleeding Disorder) Sensitivity threshold at 1% (deep sequencing) Zeuzem et al., AASLD 2015

SVR12 Rates by Treatment Regimen and Duration: TE Patients with Cirrhosis With NS5A RAVs No NS5A RAVs 100 80 96 100 89 87 SVR12 (%) 60 40 20 0 59/66 206/214 13/15 84/84 LDV/SOF+RBV LDV/SOF 12 weeks 24 weeks Studies included for analysis: LDV/SOF+RBV 12 Wks: GS-US-337-0109 (ION-2), GS-US-337-0113 (Japan 1), GS-US-337-0118 (LONESTAR-1), GS-US-337-0122 (ELECTRON-2), GS-US-337-0123 (SOLAR-1), GS-US-337-0124 (SOLAR-2), GS-US-337-1118 (Retreatment), P7977-0523 (ELECTRON); LDV/SOF 24 Wks: GS-US-337-0109 (ION-2), GS-US-337-0121 (SIRIUS), GS-US-334-1274 (Bleeding Disorder) Sensitivity threshold at 1% (deep sequencing) Zeuzem et al., AASLD 2015

How to define a RAV? RAV per specific drug or drug class (e.g. resistance against ledipasvir or any NS5A inhibitor) Genotypic backbone in vitro and in vivo (e.g. genotype 1a or 1b replicons, other genotypes) Level of reduced susceptibility in vitro (e.g. any aa change vs >2.5-fold vs >100-fold) Methodology of RAV detection (sensitivity threshold) (e.g. deep sequencing with a cut-off at 1% or populations sequencing with a cut-off at 15-25%) Impact of RAVs on SVR rates (e.g. combination partner, treatment duration, stage of disease)

Grazoprevir + Elbasvir for 12 wks in TN/relapse GT-1a patients: Impact of NS5A RAVs on SVR rates Population Sequencing Deep Sequencing at 1% Sens. Threshold EBR RAVs NS5A Class RAVs EBR RAVs NS5A Class RAVs PREVALENCE No RAVS: 414/438 (95%) No RAVS: 352/438 (8) No RAVS: 396/439 (9) No RAVS: 289/439 (65%) 5% 2 1 35% SVR12 10 8 6 2 98% 98% 86% 98% 98% 58% 72% 91% 405 414 14 24 345 352 74 86 EBR RAVs NS5A class RAVs EBR RAVs NS5A class RAVs 389 396 31 43 284 289 136 150 Population sequencing Patients without RAVs Patients with RAVs Next generation sequencing NGS with 1% ST supplemented by Population Sequencing when NGS not available. One GT1a was missing baseline population sequencing data but had baseline NGS data EBR RAV List = For GT1a: M/L28T/A/G, Q/R30E/H/R/G/K/L/D, L31M/V/F, H58D, or Y93C/H/N/S NS5A Class RAV List = Any variant from reference strain at NS5A position 24, 28, 30, 31, 32, 38, 58, 92 or 93 Jacobson et al., AASLD 2015

Effect of RAVs at Specific Baseline Positions on RAV Position Likelihood to Achieve SVR 12 GT1a-Infected TN/TE Subjects given EBR/GZR 12 weeks (no RBV) SVR12 Subjects with RAVs (1% ST NGS) SVR12 Subjects With RAVs (Population Sequencing) 24 15/18 (83.3%) 4/4 (100.) 28 61/68 (89.7%) 29/33 (87.9%) 30 14/23 (60.9%) 4/10 (40.) 31 15/23 (65.2%) 5/13 (38.5%) 32 1/1 (100.) -- 38 9/9 (100.) -- 58 75/77 (97.4%) 48/49 (98.) 92 6/6 (100.) 3/3 (100.) 93 9/14 (64.3%) 5/8 (62.5%) NGS using 1% ST supplemented by Population Sequencing when NGS not available NS5A Class RAV List = Any variant from reference strain at NS5A position 24, 28, 30, 31, 32, 38, 58, 92 and 93

Grazoprevir + Elbasvir for 12 wks in TN/relapse GT-1b patients: Impact of NS5A RAVs on SVR rates Population Sequencing Deep Sequencing at 1% Sens. Threshold EBR RAVS NS5A Class RAVS EBR RAVS NS5A Class RAVS PREVALENCE No RAVS: 220/265 (83%) No RAVS: 184/265 (69%) No RAVS: 211/265 (8) 17% 31% 2 No RAVS: 172/265 (65%) 35% SVR12 10 8 6 2 10 98% 10 99% >99% 98% 99% 99% 219 220 44 45 183 184 80 81 EBR RAVs NS5A class RAVs EBR RAVs NS5A class RAVs 210 211 53 54 171 172 92 93 Population sequencing Patients without RAVs Patients with RAVs Next generation sequencing NGS 1% ST supplemented by Population Sequencing when NGS not available EBR RAV List = For GT1b: P32L, P58D or A92K or any variant at NS5A position 28, 30, 31 or 93 NS5A Class RAV List = Any variant from reference strain at NS5A position 24, 28, 30, 31, 32, 38, 58, 92 or 93 Jacobson et al., AASLD 2015

Grazoprevir + Elbasvir for 12 wks in non-responder GT-1a patients: Impact of NS5A RAVs on SVR rates Population Sequencing Deep Sequencing at 1% Sens. Threshold EBR RAVS NS5A Class RAVS EBR RAVS NS5A Class RAVS PREVALENCE No RAVS: 61/68 (9) No RAVS: No RAVS: 54/68 59/68 (79%) (87%) 1 21% 13% No RAVS: 47/68 (69%) 31% SVR12 10 8 6 2 97% 96% 97% 96% 64% 44% 29% 59 61 2 7 52 54 9 14 EBR RAVs NS5A class RAVs EBR RAVs NS5A class RAVs 57 59 4 9 45 47 76% 16 21 Population sequencing Patients without RAVs Patients with RAVs Next generation sequencing NGS 1% ST supplemented by Population Sequencing when NGS not available. One GT1a was missing baseline population sequencing data but had baseline NGS data EBR RAV List = For GT1a: M/L28T/A/G, Q/R30E/H/R/G/K/L/D, L31M/V/F, H58D, or Y93C/H/N/S NS5A Class RAV List = Any variant from reference strain at NS5A position 24, 28, 30, 31, 32, 38, 58, 92 or 93 Jacobson et al., AASLD 2015

Grazoprevir + Elbasvir + RBV for 16/18 wks in NR GT- 1a patients: Impact of NS5A RAVs on SVR rates Population Sequencing EBR RAVS NS5A Class RAVS Deep Sequencing at 1% Sens. Threshold EBR RAVS NS5A Class RAVS PREVALENCE No RAVS: 51/52 (98%) 2% No RAVS: 44/52 (85%) No RAVS: 48/52 15% (92%) 8% No RAVS: 38/52 (73%) 27% SVR12 10 8 6 2 10 10 10 10 10 10 10 10 51 51 1 1 44 44 8 8 EBR RAVs NS5A class RAVs EBR RAVs NS5A class RAVs 48 48 4 4 38 38 14 14 Population sequencing Patients without RAVs Patients with RAVs Next generation sequencing NGS 1% ST supplemented by Population Sequencing when NGS not available. One GT1a was missing baseline population sequencing data but had baseline NGS data EBR RAV List = For GT1a: M/L28T/A/G, Q/R30E/H/R/G/K/L/D, L31M/V/F, H58D, or Y93C/H/N/S NS5A Class RAV List = Any variant from reference strain at NS5A position 24, 28, 30, 31, 32, 38, 58, 92 or 93 Jacobson et al., AASLD 2015

How to define a RAV best clinically? RAVs (single and combination) per specific drug Genotypic backbone in vitro and in vivo Level of reduced susceptibility in vitro >100-fold sensitivity threshold at 15-25% (population sequencing) Impact of RAVs on SVR rates in defined patient population and a defined regimen (DAA combination, treatment duration) and there is even more complexity...

Y93H and IL28B CC prevalence by region 60 GT1a/b 60 54 Prevalence of Y93H (%) 40 20 4 8 6 17 Prevalence of IL28B CC (%) 40 20 22 22 30 0 144 3262 North America 73 929 21 360 104 597 Europe Oceania Asia Pacific 0 713 3262 North America 203 929 109 360 324 597 Europe Oceania Asia Pacific Zeuzem et al., AASLD 2015

Association of Y93H and IL28B genotype CC Non CC Y93H Prevalence (%) 30 25 20 15 10 5 0 p <0.0001 p <0.0001 p <0.0001 25 13 12 5 5 1 178/1348 179/3800 36/774 26/2538 142/574 153/1262 GT 1 GT 1a GT 1b 5,148 GT1 patients in Gilead HCV clinical trials with both NS5A sequencing and IL28B genotype data Zeuzem et al., AASLD 2015

SVR12 rates by Y93H or any NS5A RAV and IL28B genotype With Y93H No Y93H With RAVs No RAVs 100 80 97 100 98 100 97 98 89 93 SVR12 (%) 60 40 20 0 74 76 429 431 153 156 IL28B CC 350 351 54 61 1164 1196 259 280 IL28B non-cc SVR12 data pooled from all guideline recommended LDV/SOF regimens 959 977 Studies included for analysis: GS-US-337-0118 (LONESTAR 1), GS-US-GS-US-334-1274 (Bleeding Disorder), GS-US-337-0102 (ION-1), GS-US-337-0108 (ION-3), GS-US-337-0113 (Japan 1), GS-US-337-0115 (ION-4), GS-US-337-0122 (Electron 2), GS-US- 337-0131(China), GS-US-337-1406, GS-US-337-1468 (LEPTON), GS-US-337-0109 (ION-2), GS-US-337-0124 (SOLAR-2), GS-US-334-1274 (Bleeding Disorder), GS-US-337-1406 Zeuzem et al., AASLD 2015

SVR12 Rates by Treatment Regimen and Y93H/IL28B: Patients without Cirrhosis IL28B CC IL28B non-cc With Y93H No Y93H 100 80 100 100 100 97 100 100 100 98 100 94 97 79 SVR12 (%) 60 40 20 0 3 3 31 31 1 1 102 105 51 51 239 239 LDV/SOF 8 Weeks LDV/SOF 12 Weeks LDV/SOF 12 Weeks TN < 6M TN TE 23 23 468 479 15 16 85 85 11 14 266 273 Studies included for analysis: 30 GS-US-337-0118 (LONESTAR 1), GS-US-GS-US-334-1274 (Bleeding Disorder), GS-US-337-0102 (ION-1), GS-US-337-0108 (ION-3), GS-US-337-0113 (Japan 1), GS-US-337-0115 (ION-4), GS-US-337-0122 (Electron 2), GS-US- 337-0131(China), GS-US-337-1406, GS-US-337-1468 (LEPTON), GS-US-337-0109 (ION-2), GS-US-337-0124 (SOLAR-2), GS-US-334-1274 (Bleeding Disorder), GS-US-337-1406

It is useful to detect RAVs? There is no YES or No answer First and most important: Methodology for detection of RAVs must standardized (and automated) Pharmaceutical industry must fully publish available RAV data in collaboration with academia Usefulness of RAV testing will be patient population and treatment regimen dependent RAV testing most likely not required in patient populations with SVR rates > 99% RAV testing most likely be clinical useful and cost-effective in population with suboptimal SVR rates (definition < 95% / < 9?) & if population large enough Regimens w/o a very high barrier to resistance drug Treatment-experienced patients (in particular when exposed to DAAs) Patients with cirrhosis When the shortest possible treatment duration is economically important IL28B CC genotype is significantly associated with a higher prevalence of Y93H. This association is functionally not yet understood The relevance of (in particular NS5A) RAVs, baseline HCV RNA levels, and IL28B genotype on treatment outcome requires further intensive clinical research

Treatment decisions in patients with RAVs Switch drug class (exception: nucleosidic polymerase inhibitors) Treat longer (up to 24 weeks) Add ribavirin Use triple therapies (NI + PI + NS5A-Inhibitor) Second generation PIs and NS5A-Inhibitors (e.g. Glecaprevir, Pibrentasvir) Peginterferon (??)