Combination therapy: Navigating the minefield of safety vs efficacy Jordan J Feld MD MPH Toronto Centre for Liver Disease Sandra Rotman Centre for Global Health University of Toronto
Combination therapy:
Navigating the minefield of
safety vs efficacyJordan J Feld MD MPH
Toronto Centre for Liver Disease
Sandra Rotman Centre for Global Health
University of Toronto
Disclosures
• Consulting: Abbvie, Gilead, Janssen, Merck
• Research: Abbvie, Gilead, Janssen, Merck,
Regulus
Outline
• What combinations are possible?
• Rational choices
• Approach to evaluation – lessons from other
diseases
– Efficacy
– Safety
Approaches to therapyViral Targets - DAA
• Viral entry
• cccDNA
formation/transcription/de
gradation
• RNA intermediates
• Encapsidation
• DNA replication
• Assembly
• Release
Immune Targets
• Innate immune response
– IFN
– TLR agonists
– RIG-I agonists
• Adaptive immune
response
– Anti-antagonists
(checkpoint inhibitors)
– Vaccination
Combination Approaches
Viral target
A+ Viral target
A NA + NA
Viral target
A+ Viral target
B NA + RNAi
Immune
Target A + TLR7 + VaccineImmune
Target B
Viral target
A+
NA + TLR7
Immune
Target A
Attractive combinations
Nuc
HBV DNAsuppression
+ +Immunoe
Target
RNAi
Nucleic Acid Polymers
cccDNAi
+/- cccDNAi
+/- entry inhibitor
+/- RNAi
+/- CpAM
TLR/RIG-I agonist
αPD1/PDL1
Therapeutic
vaccine
May not need all 3 ‘classes’…mix and match
Viral ProteinDepletion(s, x, core)
Key Questions
• Efficacy
– How much proof do we need for each agent in a
combination approach?
– Do we need clear ‘value’ for each agent?
– Do we need different endpoints for different targets?
• Safety
– Safety for each agent vs each combination?
‘Easy’ scenario – combine approved
drugs
Drug A
Proven activity
Proven safety
Drug B
Proven activity
Proven safety
Trial of A + B vs A vs B
(Nuc + PegIFN)
Even here…some challenges
- Population
- Duration
- Endpoints
- Big trial (2 control arms)
More difficult if new agents…Assume the 3-target paradigm is correct…
Drug A
Dose 1
Dose 2
Dose 3
Measurement
of activity & safety
Drug B
Dose 1
Dose 2
Dose 3
Measurement
of activity & safety
Drug C
Dose 1
Dose 2
Dose 3
Measurement
of activity & safety
Drug A + B
Drug A + C
Drug B + C
Drug A + B + C
Markedly more complicated if interactions, preferred
dose/duration/pop’n not clear or different endpoints by
MOA
What does combination mean?1. Sequential therapies (in either direction)
2. Concomitant from the start
3. Sequential + concomitant
• Allows assessment of each drug ‘alone’ and in combination
(and see effect on HBV DNA)
– Efficacy
– Safety
• More complicated if a third drug or 2 novel drugs
Drug A
Dose 1
Dose 2
Dose 3
12 weeks 12 weeks
Drug A + NA
A few lessons/cautions from HCV
Good examples
1.Rapid assessment of efficacy – ideal scenario
2.Better than expected (Peg/RBV)
Bad examples
1.Value of each agent in a combo (Merck)
2.Insisting on your own pipeline (BMS)
3.Unexpected toxicity (BMS)
Day0 2 4 6 8 10 12 14
-7
-6
-5
-3
-2
-1
0
1
Mean
lo
g10
HC
V R
NA
ch
an
ge f
rom
baselin
e
(IU
/mL
)
-4
Placebo
Polymerase then protease
Protease then polymerase
Combo
Gane Lancet 2010
Good Example 1 - DAAs
• Brief monotherapy followed by combo therapy
• Should be more effective in HBV – higher barrier to resistance
• But non-DNA endpoints take a long time to see an effect!
• Requires a permissive regulatory agency (thx NZ) – monotherapy…
Do you need all parts of your combo?
Prior Treatment Experience Treatment-experienced Cirrhosis
Grazoprevir (PI)/Ruzasvir (NS5A)/3682 (Nuc) x 8-16w G3 HCV
PP
SV
R1
2
• Highly effective…should have led to a Phase 3 study
• But FDA said…before you go to 3, are you sure 2 would not be
adequate?
• Back to the drawing board with NS5A + Nuc…may have killed their
HCV program Lesson: Make sure you need all of the combo
Lawitz AASLD 2016 – Abstract 110
Recognize your weaknesses…BMS HCV therapies
1.Daclatasvir – excellent pan-genotypic well tolerated NS5A
2.Asunaprevir – a weak protease inhibitor with toxicity issues
3.Beclabuvir – weak non-nuc polymerase inhibitor with toxicity
issues
Great portfolio or 1 great drug?
Lesson: Insistence on going alone can kill good drugs
Good Example 2 – Better than expected
IFN RBV
The IFN/RBV example
• Challenging precedent
• Very modestly effective therapy combined with a therapy with no activity as monotherapy
• Why was it ever tried? How would we know to do this again?
• Argues against the ‘must prove each component works’ approach
• Makes the case for ‘rational combinations’ for which each component is safe, whether effective alone or not
But endpoints and populations
may be the biggest issue
Endpoints
• With HCV or HIV simple…viral load
• Easy to measure– Changes rapidly
– Goal is undetectability in the body serum is pretty good surrogate for liver (HCV), less so for HIV
• HBV – not so simple– HBV DNA
– HBsAg
– HBeAg/anti-Hbe
– cccDNA
– HBV RNA or HBcrAg
– ALT, fibrosis…and the list goes on
Is there consensus?
Sterilizing cure(cccDNA loss)
Too hard to achieve
Sustained Virological Response(sAg +ve, DNA negative, off therapy)An advance but not enough of one
Functional Cure(sAg loss with undetectable DNA
& Normal ALT)Challenging but achievable goal
88% of attendees at EASL/ASSLD HBV Endpoints conference chose Functional Cure as the preferred goal
for future therapies
HBsAg• Preferred endpoint…but some issues
• 1. May mean different things with different targets• Natural clearance = cccDNA loss/silencing…ie what we want
• cccDNA target – probably same as above
• siRNA – blocks translation…probably no effect on cccDNA transcription unless it induces immune control
• Other antiviral – probably what we want but may depend
• Immune target – immune control - probably what we want
• False negative integrated DNA could produce sAg even in people who have cleared/silenced all cccDNA
2. Decline is SLOW! • Need long-term therapy to see an effect – probably more true with optimal
response above – major issue to avoid excluding a good drug
• If surrogate, what decrease in quant HBsAg matters? 0.5 log, 1 log, more?
3. Similar challenges likely with HBV RNA, crAg etc• Potentially other issues given uncertainty of ‘what they mean’ and whether that
differs by MOA of therapy
Different target…different
endpoint• May be a major challenge with different mechanisms of
action
– Virological target - multiple
– Immunological target – multiple
• May need ‘proof of concept’ early endpoints…possibly
including liver biopsy…for each target (Phase 1, 2a/2b)
eg. RIG-I activator works (ISG induction, even if
limited sAg decline)
• Clinically relevant endpoint for approvals (e.g. HBsAg)
(Phase 3)
Populations
• HCV/HIV – pretty easy
– Genotype
– Cirrhosis vs non-cirrhotic (start non-cirrhotic)
• HBV
– Which phase of disease for which approach?
– Certain strategies may not work in all phases
• e.g. Immunotherapy may not work in IT patients
Nuc suppressed or not nuc
suppressed – a key question
Nuc SuppressionPros
• Safety – intuitively ‘makes sense’… less likely to flare
• Clinical need – patients on therapy need therapy (?)
• Additive/synergistic effect –DNA suppression may help whatever other mechanism is being evaluated
• Easiest patients to find!!
Cons
• Safety – flares may still occur but significance less clear
• May miss an effect…if a therapy requires ‘active disease’ with inflammation…
• Lose assessment of HBV DNA as an endpoint
• Additional DDIs to consider
DNA can be useful…
DNA levels sAg levels
• Much greater effect seen on DNA than sAg
• If treat nuc suppressed patients…may miss this activity
Korolowicz PLoS One 2017
SB-9000 (RIG-I agonist) given to infected woodchucks
Bottom line on efficacy
• Not simple to measure
– Different agents with different mechanisms very
challenging
– Cannot always predict outcome of combinations
– The combination should not just be what you have in
house (a fatal mistake in HCV!!)
• Endpoints and populations remain a challenge
Safety
Step 1: Drug Interactions
• Relatively straightforward (but studies not cheap!)
• Most DDIs can be predicted once metabolism and
elimination of agents are known
• Initial healthy volunteer studies with standard PK
measurements
• Do we need to do DDI studies in patients with liver
disease? Can be done…but not easy, may not
always be necessary
• Do they need to be done for all agents & NAs?
Step 2: Toxicity
1. Flares
– Complicated!
2. Everything else
– Similar to non-HBV therapies
– Challenge is sorting out the culprit
Non-hepatic AEDrug A + Drug B + Drug C
Non-hepatic AE not seen with monotherapy studies
What’s the conclusion?
1.Type of AE may help – e.g. immune mediated more likely to
be PD1 than CpAM (but not certain!)
2.Is it an idiosyncratic reaction from A, B or C missed in BRIEF
monotherapy studies or from the combination?
3.Is it a class effect or a drug-specific effect?
BMS HCV Nuc – cardiac toxicity 1 death – killed the program
Lesson: Chronic animal tox studies critical before combo
ALT flares…a special problem
Same issues of attribution apply as with non-hepatic AE with
combo therapy
Immune-mediated flare
(“good flare”)
Autoimmune Hepatitis
(“bad flare”)
DILI
(“bad flare”)
- Viral control
- Suppression of HBV
DNA (if not on NA)
- May lead to HBeAg
and/or HBsAg loss
- Usually self-limited but
can be severe/fatal
- ‘What we want’ as long
as not too severe
- NA rescue may help
- Idiosyncratic or dose
dependent
- Most mechanisms
unknown
- Hy’s Law hard to
apply with elevated
ALT at baseline
- Can be severe/fatal
- No specific therapy
- Likely only with
immunotherapy
(anti-PD1/PDL1)
- Rapid onset
- Can be severe/fatal
- Steroid responsive
The flares we ‘want to see’
1
1 0
1 0 0
1 0 0 0
1 0 0 0 0
0
1 0 0
2 0 0
3 0 0
1 4 8 12 16 20 242 3B L
HB
sA
g, IU
/mL A
LT
, U/L
IFN
-γ S
po
ts/ 1
06
Cells
Week
Nivolumab (anti-PD1) + TDF
ALT flare leading to HBsAg clearance
Gane EASL 2017
Sorting out DILI in HBV
• Workshop to develop stopping rules for HBV
flares…no consensus reached!
• Suggested approach
Baseline ALT On-treatment change
1-2 x ULN >5 x baseline & >10 x ULN
2-5 x ULN >3 x baseline
≥ 5 x ULN >2 x baseline
This may be one argument for NA-suppressed patients
Kullak-Ublick Drug Safety 2014
Would biopsy help
• Unknown
• FDA feels likely of ‘limited value’
– Invasive
– Too non-specific to distinguish
• Not totally sure I agree…
– May be valuable as we learn more
– May give evidence of activity or lack thereof
HBV Forum EASL 2017
ALT flares during combo therapyHBRN IT Trial: Entecavir lead-in x 8w + PegIFN x 40 in IT adults
• ALT flares occurred in 68%, none icteric
• Associated with sAg decline in some but not all
• Are they good or bad? What do they mean?
Feld AASLD 2017
ALT flares can be helpful
Suresh PLoS One 2017
• ALT flare suggests ‘activity’
• Only occur during SB 9200…resolve with Nuc
• Determined approach in trials…SB9200 followed by ETV
RIG-I agonist before or after ETV sequential therapy in woodchucks
Some guidance from the FDA
FDA plans to have a ‘HBV Drug Development
Guidance Document’ - 2018
N Brown & T Block Summary of FDA/HBF/HBV Forum meeting
Key points
• Populations to study– Naïve ideal but NA-suppressed easier to find both
• DDI studies– Not required for all add-ons to NAs if no DDI expected
– Required for all other types of novel therapies
• Duration– Allow longer than standard 4w Ph 1b trials eg. 12w (if
pre-clinical/animal data supportive)
• Combos– Combo tox studies in animals first if >1 new agent
N Brown & T Block Summary of FDA/HBF/HBV Forum meeting
Summary
• Both efficacy and safety are a major challenge
with multiple new agents developed at once
• Added complexity with various endpoints and
populations to study
• Choosing the right combo, in the right patients,
for the right duration will be challenging
• Even if the science is perfect…innovative
approaches to trial design will be needed…
• Dr Chan how are we going to get there?