1. Role of resistance and resistance testing for managing HCV
3rd International HIV/Viral Hepatitis Co-Infection Meeting, AIDS 2016, Sunday 17th July 2016, Durban, South Africa
Jürgen Kurt Rockstroh
Department of Medicine I,
University Hospital Bonn,
Bonn, Germany

2. Conflict of Interest Jürgen Rockstroh has received:
Honoraria for lectures and/or consultancies from Abbott, AbbVie, Bionor, BMS, Cipla, Gilead, Janssen, Merck and ViiV.
Research grants from Dt. Leberstiftung, DZIF, NEAT ID.

3. General aspects of DAA resistance

4. General aspects
The advent of direct-acting agents (DAAs) has improved treatment of HCV but may be limited by primary drug resistance and also development of RAS (resistance-associated substitutions) in the setting of virological failure

5. Not All Direct-Acting Antivirals are Created Equal
Characteristic
Protease Inhibitor*
Protease Inhibitor**
NS5A
Inhibitor
Nuc Polymerase Inhibitor
Non-Nuc
Polymerase Inhibitor
Resistance profile
Pangenotypic efficacy
Antiviral potency
Adverse events
NS5A, nonstructural protein 5A; Nuc, nucleotide.
So, if we think about these different classes of drugs, it’s relevant to consider how they match up to each other on different categories. And so, what you can see is the different categories of drugs across the top and some of these characteristics in the different rows. If we start with the first-generation protease inhibitors, telaprevir and boceprevir, you can see that these do not do all that well on any of the categories with a poor resistance profile—they only cover genotype 1 effectively. Although they’re reasonably potent antivirals, they are associated with a number of adverse events. So, you can see a lot of red lights in that category.  
The second-generation protease inhibitors are really improved in terms of their adverse event profile and they are more potent, but they still have only a moderate expansion of the genotypic coverage and still have a limited barrier to antiviral resistance. And similarly, the NS5A inhibitors are extremely potent and very well tolerated but have a low barrier to resistance and, again, a suboptimal complete coverage of the different genotypes.
Now, you see a fair number of green lights in the nucleoside polymerase inhibitor category, and this is because this group of direct-acting antivirals has a very high genetic barrier to resistance, which I’ll discuss more later, and it seems to cover all of the 6 major HCV genotypes. They’re extremely potent and in fact, at least to date, seem to be very well tolerated. The reason this still gets an average adverse event profile and not a green light is because, at least in this class of drugs, as they’ve been developed we have seen that although they seem to be well tolerated, there have been some important and quite significant toxicities of drugs identified in small numbers of patients taking drugs from this class in clinical trials. And in some cases, this has led to the discontinuation of potential nuc polymerase inhibitors from being developed. So, at least this still leaves a slight question mark that there may be unrecognized toxicities as these get used in larger populations of patients. But at least to date, the AE profile actually looks quite good.
And finally, we come to the nonnuc polymerase inhibitors, and you see that there’s really no green lights in this category. They really suffer from a fairly poor resistance profile, a limited genotypic coverage, limited potency, and there have been some issues with AEs, as well. And this class of drugs can only be considered, at least at this stage, as a third DAA or an add-on DAA and cannot really be counted upon as a DAA to do most of the heavy lifting and, really, can only be added to one of the other classes.
Good profile
Average profile
Least favorable profile
*First generation. **Second generation.

6. Ombitasvir-specific RAVs detected in this study:
Prevalence of Baseline GT1a NS5A RAVs: Impact of RAV Definition and Sensitivity of Detection
NS5A Inhibitor Class RAVs detected in this study at amino acid positions:
M28(all), Q30(all), L31(all), P32L, H58D/R, and Y93(all)
Ombitasvir-specific RAVs detected in this study:
M28T/V, Q30E/R, H58D, Y93C/F/H/L/N
1% Detection Threshold
1% Detection Threshold
15% Detection Threshold
15% Detection Threshold
Sulkowski M, et al. 23rd CROI; Boston, MA; February 22-25, Abst. 539LB.

7. Stopping Rules—The Facts: Multiple Mutations May Be More Troublesome
Loss of detectable resistance in patients with resistant variant(s) at failure of TVR + pegIFN/RBV (analysis includes only patients with follow-up data)
1.0
0.8
V36M + R155K (n = 124)
R155K alone (n = 41)
0.6
Probability
0.4
V36M alone (n = 22)
0.2 2 4 6 8 10 12 14 16 18
Mos After Treatment Failure
And, you can see on the next slide how this has a very significant impact on the rate of decline of the resistant virus.
What we’re seeing here is a number of different hepatitis C drug–associated variants that confer some degree of resistance to therapy, and you can see that, when treatment is stopped, these mutations slowly disappear and the virus returns to become predominantly the wild-type, i.e., the sensitive viral strain. What you can see is that some mutations disappear more quickly and others disappear more slowly, so what we don’t want to do is to encourage the virus to form the very resistant, very persistent variants that may take much longer to clear. So, the take-home message is that we do not want to breed resistance into the virus that will make future courses of treatment more difficult.
V36M Alone*
R155K Alone†
V36M + R155K
% of 1a failures (WT: 16%) 10 20 46
Median mos to loss (95% CI)
6 (4-9)
10 (9-13)
13 (10-13)
*Comparison of V36M vs V36M + R155K: P < †Comparison of R155K vs V36M + R155K: P = .48.
Sullivan J, et al. EASL Abstract 8.

8. Long-Term Persistence of HCV NS5A Variants After Treatment With LDV
NS5A RAVs in patients who failed HCV treatment with ledipasvir (LDV) in the absence SOF
Positions 24, 28, 30, 31, 32, 58, 93 that confer >2.5-fold reduced susceptibility to LDV in vitro were included
Majority of RAVs Detected After 96 Weeks (> 1% of Population)
Registry Study
62/63
58/58
42/43
45/45
52/55
50/58
Patients without NS5A RAVs
Patients with NS5A RAVs
NS5A RAVs persisted in majority of patients for 96 weeks
Almost all patients developed NS5A RAVs at treatment failure
Wyles D, et al. 50th EASL; Vienna, Austria; April 22-26, Abst. O059.

9. Persistence of RAVs in Patients who Relapsed after 3D
67/2510 patients with genotype 1a and virologic failure after 3D
24 wks post-treatment
48 wks post-treatment
97%
96%
100 90
75% 80 70
57% 60
Prevalence of RAVs (%)
46% 50 40 30 20 9% 10
N=67
N=57
N=70
N=51
N=44
N=35
NS3 RAVs
(paritaprevir)
NS5A RAVs
(ombitasvir)
NS5B RAVs
(dasabuvir)
(Krishnan et al., EASL 2015)

10. What do the guidelines say?

11. EASL 2015

12. What does the label say?

13. What does the label say?
When considering OLYSIO (Simeprevir) combination treatment with peginterferon alfa and ribavirin in HCV genotype 1a patients, patients should be tested for the presence of virus with the NS3 Q80K polymorphism before starting treatment
Zepatier label USA: Genotype 1a: Testing for the presence of virus with NS5A resistance-associated polymorphisms is recommended.
Daklinza® (daclatasvir) and Sunvepra® (asunaprevir) licensed in Japan

14. QUEST: No Benefit of Simeprevir if Q80K Positive
100
SMV + P/R 85
P/R 84 80 60 58 53
SVR12 (%) 52 43 40 20
228/
267
70/
133
138/
165
36/ 83
49/ 84
23/ 44
GT, genotype; P/R, peginterferon/ribavirin; SMV, simeprevir; SVR12, sustained virologic response 12 weeks posttreatment.
So, if we actually look at that in a little more detail on the next slide, what you see is in the patients with genotype 1b, you can see that there was an 85% SVR rate in those getting simeprevir and peg/riba compared to only 53% in the control arm. And similarly, in patients with genotype 1a without the Q80K mutation, they had a very similar SVR rate of 84%. But strikingly, when the Q80K mutation was present at baseline, you can see that suddenly that SVR rate in the simeprevir-treated patients drops down to just 58%, very similar to that seen in the control arms.  
So, really, you can see that when this polymorphism is present, unfortunately simeprevir adds little benefit over the peg/riba control, and unfortunately, Q80K is actually relatively common in genotype 1a, which is also the more common subtype found in North America. You can see that in this trial, 34% of patients with genotype 1a had the Q80K mutation, and this contrasts to only about 1% or fewer with genotype 1b. So, this is really a genotype 1a issue and really very little or no benefit to simeprevir when this mutation is found.
n/N =
GT1b
GT1a no Q80K
GT1a + Q80K
Q80K present in 34% of GT1a patients No benefit of simeprevir if Q80K positive
Jacobson I, et al. AASLD Abstract 1122.

15. OPTIMIST 2: SMV + SOF in HCV Mono-Infx with GT1 and Cirrhosis
SVR12 Rates (95% CI) by:
Prior Treatment History
IL28B Genotype
(94.0; 100)
(98.8; 100)
(87.0; 100)
(92.1; 100)
(80.1; 100)
44/50
42/53
25/29
40/54
15/19
Platelets, Albumin, and
FibroScan Score (Intent-to-treat Population)
HCV Geno/Subtype, and Baseline Q80k
(93.1; 100)
(74.0; 93.2)
(93.1; 100)
(92.8; 100)
(82.7; 100)
(89.1; 100)
(60.1; 86.9)
(68.2; 85.1)
(93.1; 100)
(90.7; 100)
*Treatment-experienced patients included prior relapcers, prior non-responders, IPN-intolerant, and other patients. CI, confidence interval; GT, genotype: HCV, hepatitis C virus: IPN interaction SMV. Imeprevir: SOF, sodbavir, SVR12, sustained virologic response 12 weeks after end of treatment.
60/72
25/34
35/34
26/31
13/19
73/54
32/53
47/50
12/15
11/11
Platelets <90,000/mm1
Platelets >90,000/mm1
FibroScan Score
>20 kPa
Albumin
<4 g/dL
Albumin
>4 g/dL
FibroScan
Score
>12.5-<20kPa
SMV + SOF 12 weeks
Lawitz E, et al; 8th IAS, Vancouver, Canada, July 19-22, 2015; Abst. TULBPE11.

16. C-EDGE TN + C-EDGE CO-INFECTION: NS3/4A Resistance Associated Variants
Resistance analysis population (290 GT1a; 172 GT1b)†
RAV Status in Patients with Baseline Sequence % (n/m)
SVR12
All Patients % (N/n)
NS3 RAVs
≤5-fold potency loss
>5-fold
potency loss
Genotype 1a RAVS
Baseline NS3 RAVS
53.4
(155/290)
96.1
(149/155)
0/0
No baseline NS3 RAVs
46.6
(135/290)
93.3
(126/135) —
Genotype 1b RAVS
Baseline NS3 RAVS
17.4
(30/172)
96.7
(29/30)
96.1
(25/26)
100
(4/4)
No baseline NS3 RAVs 82
(142/172)
99.3
(141/142) —
PN060 GT1a:
Baseline NS3 RAVS: 57% (86/151); 97% (83/86); 97%(83/86); 0/0
No baseline NS3 RAVs: 43%(65/151); 89% (58/65)
PN061 GT1a RAVs
baseline NS3 RAVs: (69/139); SVR12 (66/69); SVR12 NS3 RAVs </= 5-fold potency loss (66/69); SVR12 NS3 RAVs >5-fold potency loss n/a (0/0)
no baseline NS3 RAVs: 70/139 ;SVR12 (68/70)
PN060 Genotype 1b RAVS
Baseline NS3 RAVS:19%(25/129); 96% (24/25); 96% 21/22; 100% 3/3
No baseline NS3 RAVs: 81% (104/129); 100%(104/104)
PN061 GT1b RAVs
baseline NS3 RAVs: (5/43); SVR12 (5/5); SVR12 NS3 RAVs </= 5-fold potency loss (4/4); SVR12 NS3 RAVs >5-fold potency loss (1/1)
no baseline NS3 RAVs: (38/43); SVR12 (37/38)
†The resistance analysis population includes all patients from the full analysis set who have sequencing data available and who either achieved SVR12 or met criteria for virologic failure
N = number of patients who achieved SVR12
m = number of patients with evaluable baseline sequence
n = number of patients with or without a baseline RAV
Signature NS3 loci included the substitutions V36A/G/L/M/I, T54A/C/G/S, V55A/I, Y56H, Q80K/R, V107I,
122A/G/R, I132V, R155X, A156S/T/V/F/G/L, V158I, D168X, I/V170A/F/T/V, and M175L.
The following NS3 RAV(s) are considered to have >5-fold resistance to GZR based on GT1a replicons:
Y56H, R155G/T/W, A156G/T/V/L, D168A/G/T/V/L/I/F/Y/E/H/K/R .

17. C-EDGE TN + C-EDGE CO-INFECTION: NS5A Resistance Associated Variants
Resistance analysis population (294 GT1a; 173 GT1b)†
RAV Status in Patients with Baseline Sequence % (n/m)
SVR12
All Patients % (N/n)
NS5A RAVs
≤5-fold potency loss
>5-fold potency loss
Genotype 1a RAVS
Baseline NS5A RAVS
9.9
(29/294)
65.5
(19/29)
87.5
(14/16)
38.5
(5/13)
No baseline NS5A RAVs
90.1
(265/294)
98.1
(260/265) —
Genotype 1b RAVS
Baseline NS5A RAVS
13.3
(23/173)
95.7
(22/23)
100
(1/1)
95.5
(21/22)
No baseline NS5A RAVs
86.7
(150/173)
99.3
(149/150) —
PN060 GT1a RAVS
Baseline NS5A RAVS: 12%(19/154); 58%(11/19); 90%9/10; 22%2/9
No baseline NS5A RAVs: 88%(135/154); 99%(133/135)
PN061 GT1a
Baseline RAVs: 10/140; 8/10; 5/6; 3/4
No baseline RAVS: 130/140; 127/130
PN060 Genotype 1b RAVS
Baseline NS5A RAVS:14%(18/130); 94% (17/18); 100%1/1; 94%16/17
No baseline NS5A RAVs: 86% (112/130); 100%(112/112)
PN061 GT1b
Baseline RAVs:5/43; 5/5 (all greater than 5-fold)
No baseline RAVs: 38/43; 37/38
†The resistance analysis population includes all patients from the full analysis set who have sequencing data available and who either achieved SVR12 or met criteria for virologic failure
N = number of patients who achieved SVR12
m = number of patients with evaluable baseline sequence
n = number of patients with or without a baseline RAV
Signature NS5A loci included the substitutions M28T/V/A, Q30E/H/R/G/K/L/D, L31M/V/F, H58D, and Y93C/H/N/S for GT1a and the substitutions L28T/V/A, R30E/H/G/K/L/D, L31M/V/F, P58D and Y93C/H/N/S for GT1b.
Based on the available GT1a replicon data, the following variants are considered to have >5-fold
resistance to EBR: M/L28T/A, Q/R30E/H/R/G/K/L/D, L31M/V/F, H58D, Y93C/H

18. Daclatasvir + Asunaprevir: High SVR across all patient types without baseline L31 and/or Y93H polymorphisms
McPhee et al. APASL Poster 1549; Kao JH..Ahn SH. The Liver Week 2015.

19. SVR to Sofosbuvir/Ledipasvir According to NS5A RAVs (513 cirrhotic patients)
Sarrazin et al., EASL 2015

20. Impact of Baseline GT1a NS5A Class RAVs and Ombitasvir-specific RAVs on SVR Rate
OBV-specific RAVs
OBV-specific RAVs
Class RAVs
Class RAVs 89 94
230
236 70 74
249
256 48 50
271
280 36 38
283
292
Similar SVR rates were observed irrespective of the presence or absence of baseline variants
Sulkowski M, et al. 23rd CROI; Boston, MA; February 22-25, Abst. 539LB.

21. Can a resistance test guide treatment decision making in patients with prior failure of DAA based therapy?

22. SVR according to baseline parameters Prior Treatment Duration
Retreatment of Patients Who Failed 8 or 12 Weeks of LDV/SOF-Based Regimens With LDV/SOF for 24 Weeks
LDV/SOF 24 Weeks N=41
Mean age, y (range)
58 (35-71)
Male, n (%)
34 (83)
Black/African American, n (%)
10 (24)
IL28B non-CC, n (%)
38 (93)
GT 1a, n (%)
Mean HCV RNA, log10IU/mL (range)
6.2 ( )
Cirrhosis, n (%)
19 (46)
Presence of NS5A RAVs
15 (79)
Prior HCV treatment, n (%)
LDV/SOF ± RBV
33 (80)
LDV/SOF + GS-9669
8 (20)
Prior HCV treatment duration, n (%)
8 weeks
30 (73)
19 (63)
12 weeks
11 (27)
11 (100)
SVR according to baseline parameters
No Yes
Cirrhosis
8 wks 12 wks
Prior Treatment Duration
No Yes
Baseline
NS5A
RAVs
Lawitz E, et al. 50th EASL; Vienna, Austria; April 22-26, Abst. O005.

23. SVR12 by Baseline NS5A RAVs GT 1 Retreatment
Results and Analysis
SVR12 by Baseline NS5A RAVs GT 1 Retreatment
Prior to re-treat
No NS5B resistance associated (S282T) or treatment-emergent (L159F, V321A) variants were detected
At second virologic failure
4 of 12 (33%) patients had NS5B variants detected
S282T (n=2)
L159F (n=1)
Double-mutant S282T + L159F (n=1)
SVR12 (%)
Lawitz E, et al. 50th EASL; Vienna, Austria; April 22-26, Abst. O005.

24. Re-treatment after failure to LDV/SOF
9 patients without SVR in ION-4 after 12 weeks of LDV/SOF
SVR in 8/9; 1 relapse 4 weeks after EOT: GT1a, no cirrhosis
LDV/SOF Failure
N=9
LDV/SOF + RBV
SVR12
Wk 0
Wk 12
Wk 24
Wk 36 GT
NS5A RAVs Before Primary Study (%)
NS5A RAVs at Virologic Rlapse After Primary Study (5)
SRV12 1a
None
Yes
L31M (>99), H58D (92)
Y93F (1), Y93N (10)
Y93N (<99)
L31M (>99), Y93N (<25)
L31M (>99), Y93N (>99)
1a*
Y93N (>99) 1b
Y93H (>99)
L31I (11), Y93H (>99)
L31V (>99)
L31M (>99) No
SVR in 8/9; 1 relapse 4 weeks after EOT: GT1a, no cirrhosis
Cooper C, et al. 23rd CROI; Boston, MA; February 22-25, Abst. 573.

25. Assess HCV guidelines approach to re-treatment:
Retreatment of Patients Who Failed DAA-combination Therapies - Real-world Experience From a Large Hepatitis C Resistance Database
N = 310
SMV/SOF ± RBV
N=55
LDV/SOF ± RBV n=114
DCV/SOF ± RBV
n=51
RTV/OBV ± RBV
n=30
SOF + RBV
n=60
Mean age, y (range)
58 (43-75)
57 (34-77)
55 (31-71)
55 (34-58)
52 (27-65)
Men, n (%)
43 (78)
94 (82)
42 (82
26 (87)
47 (78)
Cerrhosis, n (%)
37 (71)
62 (57)
33 (70)
11 (37)
21 (44)
+RBV, n (%)
10 (55)
39 (34)
8 (16)
19 (63)
60 (100)
Prior IFN Therapy
39 (76)
67 (67)
20 (76)
21 (70)
27 (63)
G, n (%) 1
49 (89)
90 (79)
29 (57)
27 (90) - 2
27 (45) 3
1 (2)
15 (13)
20 (39)
33 (55) 4
5 (9)
9 (8)
2 (4)
3 (10)
Treatment duration 8/12/24 weeks, n
-/53/1
12/80/20
-/25/26
-/29/1
-/26/29
Subset of the European resistance database (n=3549) with persons who failed DAAs outside of clinical trials (N=310) – drug-class specific RASs (NS3, NS5A, NS5B) associated with > 2-fold increase in EC50
Assess HCV guidelines approach to re-treatment:
Use active DAAs
Add RBV
Longer duration
Vermehren J, et al. 51st EASL; Barcelona, Spain; April 13-17, Abst. PS103.

26. Retreatment after DAA Failure
22 of 119 patients with G1 and NS5A treatment failure (DCV or LDV + SOF)
Retreatment with with PI containing regimen
SMV/SOF +/- RBV or 3D +/- RBV for 12 or 24 weeks
SVR12 in 6 of 7 patients (limited data)
27 of 49 patients with G1 and SMV/SOF treatment failure
Retreatment with NS5A containing regimen
LDV/SOF (n=23) or 3D (n=4) +/- RBV
SVR12 in 20 of 22 patients
Frequency of RAVs
Frequency of RAVs
No NS5B RAVs
Vermehren J, et al. 51st EASL; Barcelona, Spain; April 13-17, Abst. PS103.

27. Retreatment after DAA Failure (cont’d)
14 of 23 patients with genotype 3 and SOF/RBV treatment failure
Retreatment with DCV (n=13) or LDV (n=1) + SOF +/- RBV
SVR12 in 7 of 7
DAA failures – male sex, cirrhosis, prior PR
73 patients (28%) started RAS-driven retreatment
Preliminary data promising
5 of 27 patients with G1 and 3D treatment failure
Retreatment with SOF based regimen
No data on outcomes
Frequency of RAVs
Frequency of RAVs
Vermehren J, et al. 51st EASL; Barcelona, Spain; April 13-17, Abst. PS103.

28. JR37
Summary
After IFN-free treatment failure, HCV variants resistant to protease inhibitors progressively disappear by population sequencing, replaced by wild-type virus
In contrast, viruses resistant to NS5A inhibitors and to NNIs persist for years
In most patients who fail to achieve an SVR on an IFN-free regimen, viruses that are resistant to one or more of the DAAs administered are present as the dominant species at the time of relapse
By means of population sequencing, HCV RAVs at baseline may have an impact on the rate of SVR with IFN-free regimens in patients with negative host factors
The addition of ribavirin and extending treatment duration appears to minimize the impact of pre-existing RAVs on SVR