Presentation by Dr Andrea Jorgensen, Senior Lecturer - Biostatistics, University of Liverpool: Personalised Medicine in the treatment of AF at The future of innovation in AF and stroke prevention in the NWC, 27 June 2018, Haydock Park Racecourse
Dr Andrea Jorgensen - The future of innovation in atrial fibrillation and stroke prevention in the NWC
1. Personalised Medicine in
the Treatment of AF –
Genotype Guided Dosing
Dr Andrea Jorgensen
Department of Biostatistics
University of Liverpool
2. Warfarin
• Despite newer anticoagulants becoming increasingly
used, warfarin still commonly used for AF
• Works by interfering with body’s Vit K production–
prothrombin (hence clot) formation slowed down
• Efficacy depends on maintaining INR in therapeutic
range (typically 2-3). Difficult because:
i. Warfarin has narrow therapeutic index – dose for therapeutic
anticoagulation (‘therapeutic dose’) very close to dose causing
over-anticoagulation
ii. Warfarin has very large inter-individual variability in
therapeutic dose requirements: 0.5mg/day to >10mg/day
3. Consequences of getting
dose wrong
• Most feared adverse effect of warfarin treatment is bleeding
• Major bleeding: 7.2 events per 100 patient years1
• Fatal bleeding: 1.3 events per 100 patient years1
• Risk of bleeding higher when INR above therapeutic range
• Important to achieve and maintain therapeutic range as soon as
possible
1. Linkins LA, Choi PT, Douketis JD. Ann Intern Med 2003; 139: 893-900
4. Current practice
• INR measurement-dose change cycle repeated until stable anticoagulation
(INR achieved and maintained in therapeutic range)
• INR measurements thereafter approximately once monthly
Day 1-3: Loading dose
– e.g.10mg; 5mg; 5mg/
5mg; 5mg; 5mg in
elderly
Day 4: INR
measurement
INR entered into
dose calculating
software (e.g.
DAWN/RAID) –
new dose suggested
where necessary
INR measured
INR measurement-dose change cycle
5. Problems with current
practice
• Patient inconvenience: regular clinic visits/blood
tests until therapeutic INR maintained
• NHS costs of regular monitoring
• Time to achieving therapeutic range can be lengthy
• Increased risk of bleeding if therapeutic range
exceeded
• Increased risk of clotting if therapeutic range under-
achieved
6. Factors influencing
therapeutic dose
• Considerable amount of research into factors
associated with therapeutic dose requirements
• Factors identified include:
o Age
o Gender
o Weight
o Interacting medications
o Liver disease
o Alcohol intake
o Vitamin K intake
o Genetic factors…
7. Genetic variants associated
with therapeutic dose
Gene CYP2C9 involved in metabolism of
Warfarin
Genotype at variants known as CYP2C9*2 and
CYP2C9*3 associated with warfarin dose
Gene VKORC1 involved in Vitamin K cycle
Genotype at several variants in VKORC1
associated with warfarin dose requirements
9. CYP2C9*2
Jorgensen AL, FitzGerald RJ, Oyee J, Pirmohamed M, et al. (2012) PLoS ONE 7(8): e44064.
*1/*2 vs *1/*1 *2/*2 vs *1/*1
Mutant allele
carriers need
less dose
Mutant allele
carriers need
less dose
Mutant allele
carriers need
more dose
Mutant allele
carriers need
more dose
Mutant carriers need almost
1mg/day less
Mutant carriers need almost
1.5mg/day less
10. CYP2C9*3
Jorgensen AL, FitzGerald RJ, Oyee J, Pirmohamed M, et al. (2012) PLoS ONE 7(8): e44064.
*1/*3 vs *1/*1 *3/*3 vs *1/*1
Mutant allele carriers
need more dose
Mutant allele carriers
need more dose
Mutant allele carriers
need less dose
Mutant allele carriers
need less dose
Mutant carriers need over
1.5mg/day less
Mutant carriers need over
2.5mg/day less
11. CYP2C9*2 and *3 combined
Jorgensen AL, FitzGerald RJ, Oyee J, Pirmohamed M, et al. (2012) PLoS ONE 7(8): e44064.
1 mutant allele
vs none
2 mutant
alleles vs none
Mutant allele carriers
need less dose
Mutant allele carriers
need more dose
Mutant allele carriers
need less dose
Mutant allele carriers
need more dose
Mutant carriers need over
1 mg/day less
Mutant carriers need over
3.5 mg/day less
12. VKORC1 (rs9923231)
Jorgensen AL, FitzGerald RJ, Oyee J, Pirmohamed M, et al. (2012) PLoS ONE 7(8): e44064.
1 mutant allele
vs none
2 mutant
alleles vs none
Mutant allele carriers
need less dose
Mutant allele carriers
need more dose
Mutant allele carriers
need less dose
Mutant allele carriers
need more dose
Mutant carriers need almost
1.5 mg/day less
Mutant carriers need almost
3 mg/day less
13. International Warfarin
Pharmacogenetic Consortium
(IWPC)
International collaborative effort
to develop dosing algorithm for
warfarin
Contribution of data on 5052
patients from several different
primary studies worldwide
Built linear regression model
(derive and validate)
Ref: Estimation of the warfarin dose with
clinical and pharmacogenetic data. N Engl J
Med (2009) 360: 753-764.
14. Clinical utility ?
IWPC dosing algorithm predicted substantial proportion of variability
in dose requirements ~50%
Warfarin labelling changed to recommend that, if CYP2C9/VKORC1
genotype known, could assist in guiding dose
But:
genetic testing not made compulsory prior to starting warfarin treatment
routine genetic testing pre-treatment not implemented in practice
General consensus – needed to prove clinical utility
How do we know approach better than current practice ?
Gold-standard to prove clinical utility: randomised controlled trial
15. EU-PACT Trial
The European Pharmacogenetics of
Anticoagulant Therapy Trial
Funded by EU Seventh Framework Programme
Five centres:
Liverpool, UK
Newcastle, UK
St. Helens, UK
Uppsala, Sweden
Enkoping, Sweden
A randomized trial of genotype-guided dosing of warfarin. N Engl J Med (2013)
369: 2294-2303.
16. EU-PACT Trial (cont…)
• AIM: Determine whether genotype-guided dosing of
warfarin superior to standard clinical care during initial 3
months of treatment in patients with AF or VTE,
previously naïve to warfarin
• DESIGN: Pragmatic single-blind parallel group
randomized controlled trial
17. • Primary outcome: %
time in therapeutic
range first 12 weeks
• Secondary outcomes
included:
- INR≥4
- Time to therapeutic
INR
- Time to stable dose
- Major/minor bleeds
• Block
randomisation
• Stratified
according to centre
and indication
• Single-blind
allocation
• Warfarin-naïve
• Atrial fibrillation/
Venous thromboembolism
• Therapeutic INR
range: 2-3
CONTROL ARM
INTERVENTION ARM
Study Design
Eligibility &
Consent
Randomise
Days 1-3:
Genotype-
guided
loading dose
regime
Repeated
INRs
followed by
genotype-
guided
dose
adjustment
Compare
outcome
between two
trial arms
Repeated
INRs
followed by
non
genotype-
guided dose
adjustment
Days 1-3:
Non
genotype-
guided
loading dose
regime
19. Patient follow-up
• Genotype guided arm, n=227
• Control arm, n=228
• Caucasian ethnic background
• 3 month follow-up
• INR measured day 1, 4, 6, 8, 15, 22, 57, 85
• Additional clinic visits and INR measurements according to
clinical need
20. Point of care genotyping assay
Time <2 hours
Results for:
• CYP2C9*2
• CYP2C9*3
• VKORC1 (-1639G>A)
21. Baseline Variables
• Well balanced between arms
• Males: 61.0%; n=277
• Caucasian: 98.5%; n=447
• Age: mean 67.3 years (sd 13.7)
• AF: 72.1%; n=328
• Genotype frequencies consistent with literature
22. Results – Primary outcome
• Percentage time in therapeutic range first
12 weeks:
o 67.4% (genotype guided) vs 60.3% (control)
o 7% difference (p<0.001)
Difference in mean
INR greatest soon after
initiation, then reduces
Difference in % time in
range becomes apparent
between 5-10 days after
initiation, then tapers off
after about 9 weeks
23. Results – Secondary
outcomes
• INR≥4: 27%(genotype-guided) vs 36.6%(control); p=0.03
• Time to therapeutic INR: 21 days (genotype-guided) vs
29 days (control); p<0.001
• Time to stable dose: 44 days(genotype-guided) vs 59
days (control); p=0.003
• Bleeding events: 37%(genotype-guided) vs
38% (control); p=0.87
24. Details of bleeding events
• No major bleeding events according to the ISTH
classification were reported in the trial
• Three bleeding events were classified as clinically
significant and required admission to the hospital -
all in control group
• Majority of the minor bleeding episodes consisted of
bruising
• Only one thromboembolic event – in
control group
25. Conclusions
• Genotype-guided approach:
increased time patients spend within therapeutic INR
reduced likelihood of excessive anticoagulation (INR≥4)
reduced time taken to achieve therapeutic INR
reduced time to stable dose
• Reduced intervention
• Improved safety
• Cost implications
• No difference in bleeding outcomes – translation
to improved clinical outcomes unclear…
• …BUT INR≥4 increases bleeding risk
27. Clinical Implementation
Study (GGD Study)
Aim: determine whether genotype guided dosing (GGD) is
beneficial and feasible in clinical practice
Six clinics from North West of England
Recruited patients commencing warfarin for AF
Three clinics used GGD approach to determine dose for
first 5 days (GGD arm)
Three clinics used local standard approach to determine
dose for first 5 days (control arm)
From day 6 onwards, standard clinic care
28. GGD Study – follow-up
Follow-up for 3 months
Genotyping on point-of-care test developed by LGC;
results available within 1 hour
All INR and dose changes recorded
All adverse events recorded
Quality of life questionnaires completed at baseline
and 3 months (cost-effectiveness)
Patients and healthcare providers in GGD arm
completed questionnaires about experience
29. GGD Study – outcomes
Primary outcome: % time in target INR range first 3 months
Secondary outcomes:
i. INR ≥ 4 in first week
ii. INR < 2 in first week
iii. Number of clinic visits first 3 months
iv. Adverse events
v. Staff opinion of GGD
vi. Patient opinion of GGD
vii. Cost-effectiveness of GGD implementation
Primary outcome and secondary outcomes i)-iv) compared
between GGD arm (n=132)and control arm (n=93)
Control arm enriched by ‘dashboard data’(n=640)to
increase power
30. GGD Study – results
% Time in target INR range: 61.9% (GGD) vs 55.4%
(control); p=0.002
INR ≥ 4: 2.46% (GGD) vs 7.37% (control); p=0.06
Patients viewed GGD positively; majority grading various
aspects of the approach as “Very acceptable” or
“Acceptable”
Staff largely positive about their experience, with only some
expressing concern about the additional time patients
waited to know their dose
Cost-effectiveness analysis underway
31. GGD study - conclusions
• Further evidence of benefits of GGD in terms of
anticoagulation control
• Suggests only minor adjustments required to the process
to allow smooth implementation into clinical practice
• Cost-effectiveness analysis results will provide further
insight into feasibility
32. Acknowledgements
• EU-PACT Investigators (PI Professor Sir Munir Pirmohamed)
• GGD study investigators (PI Professor Sir Munir Pirmohamed)
• All patients who took part in the trial and the study
• All centres who recruited patients for trial and study
• All nurses, data managers and monitors for help in running
trial and study
• LGC for the point of care genotyping platform
• European Commission (EU-FP7) for funding trial
• CLAHRC and Innovation Agency North West Coast for
supporting study
33. What about the newer
anticoagulants ?
• Newer anticoagulants on the market e.g. dabigatran, rivaroxaban,
apixaban
• Non-inferior to warfarin; licensed by NICE for avoidance of
stroke/systemic embolism in non-valvular AF (rivaroxaban also
licensed for VTE)
• But…
• …costly
• …no biomarker for easy monitoring (of particular importance in
elderly where interacting medication usage high)
• …no known antidote in case of bleeding
• …all affected by renal function, particular problem in elderly (approx.
10% of individuals over 80 require anticoagulation)
• Evidence of clinical utility in those with mechanical heart valves is
lacking