Transcript eSlide - P6780 ACC Epstein Presentation

Effect of Genotyping Warfarin Patients on Outcomes:
Results from The National Community-based MedcoMayo Warfarin Effectiveness Study (MM-WES)
Robert S Epstein MD MS, Thomas P. Moyer PhD, Ronald E. Aubert PhD,
Dennis J. O’Kane PhD, Fang Xia PhD, Robert R. Verbrugge PhD,
Brian F. Gage MD MS, J. Russell Teagarden DMH, RPh
Medco Health Solutions, Franklin Lakes, NJ;
Mayo Clinic, Rochester, MN;
Washington University, St Louis MO
Funding sources: Medco Health Solutions, Mayo Clinic Center for Individualized Therapy
Manuscript is “in press” in Journal of American College of Cardiology
Medco is a registered trademark of Medco Health Solutions, Inc.
© 2010 Medco Health Solutions, Inc. All rights reserved.
Background
Warfarin exhibits large inter-individual dosing requirements
Warfarin is a leading cause of morbidity and mortality
Two genes account for ~33% of variance in dosing
 Cytochrome P450 2C9 (CYP2C9) – pharmacokinetics
 VKORC1 – pharmacodynamics
Meta-analysis of 3 clinical trials of warfarin genotyping showed
a 32% decrease in major bleeding (RR 0.68, CI 0.22-2.06)*
*Eckman MH, Rosand J, Geenberg SM, Gage BF: Cost-effectiveness of using pharmacogenetic
information in warfarin dosing for patients with nonvalvular atrial fibrillation. Ann Int Med
2009;150(2):73-83.
2
© 2010 Medco Health Solutions, Inc. All rights reserved.
Study question:
Does use of CYP2C9/VKORC1 testing
reduce the risk of hospitalization during the
first 6 months of warfarin treatment?
3
© 2010 Medco Health Solutions, Inc. All rights reserved.
Study design: Comparative effectiveness
6 month follow-up on all patients initiating warfarin in all groups
Medco-Mayo
Warfarin Effectiveness Study
23 benefit plan sponsors
56 benefit plan sponsors
Historical control
Intervention
group
External
historical control
External
concurrent control
July 2006 –
June 2007
n=2688
July 2007 –
February 2009
n=896
July 2006 –
June 2007
n=2688
July 2007 –
February 2009
n=2688
Primary comparison
4
© 2010 Medco Health Solutions, Inc. All rights reserved.
Comparison of general trends in practice
Statistical Methods of Outcomes
Comparisons
Unadjusted comparisons – Kaplan-Meier and log-rank tests
Adjusted comparisons – propensity scores to handle
participant/non-participant differences, indications for therapy,
specific concomitant drugs, medical conditions, prior history of
hospitalization or history of bleed/thromboemboli.
ANALYSES
Intention-to-treat – all outcomes even if adverse event
preceded genotype
Per-protocol – only those outcomes counted if post-genotype
5
© 2010 Medco Health Solutions, Inc. All rights reserved.
Flow of the genotyping arm
Medco identified
‘new starts’ to
warfarin on any
given day of
the week
Medco contacted
‘new starts’ to
solicit verbal
informed consent
Medco contacted
physician for
clinical information
and consent for
patient to receive
genotype test
First half of
enrollment – Medco
arranged for home
blood draw –
received written
informed consent,
sent blood to Mayo
Mayo completed lab test –
supplied report to physician
and results to Medco
6
© 2010 Medco Health Solutions, Inc. All rights reserved.
Sample Mayo Clinic Laboratory Report
Sample Lab Report: Warfarin Genotype Results
Medco Health Solutions
Mayo/Medco Warfarin Protocol
Attn: Accounts Payable
100 Parsons Pond Drive
Franklin Lakes, NJ 07417
Accession #:
Patient Name:
Birth Date:
Medical Rec #:
Client Accn #:
A1234567
DOE, JANE
09/13/1942 Age:65 Gender:F
1234
123456789
Ordering Phys:
Collect Date:
Received Date:
11/07/2007
11/08/2007
SMITH, JOHN
10:40 AM
7:19 AM
---------------------------------------------------------------------------------------------------------------
Test Requested
Result
Units
Ref Range
Perform Site *
------------------------------------Rapid DNA Extraction
Comment
Genomic DNA was extracted.
MCR
============================================================================
CYP2C9 + VKORC1 Genotype, Warfarin
CYP2C9 430C>T(*2)
C/T
MCR
CYP2C9 1075A>C(*3)
A/C
MCR
CYP2C9 1076T>C(*4)
T/T
MCR
CYP2C9 1080C>G(*5)
C/C
MCR
CYP2C9 818delA(*6)
A/A
MCR
VKORC1 -1639G>A
A/A
MCR
Interpretation
MCR
This genotype is rare and has very high sensitivity to
warfarin. Warfarin dose decrease and frequent INR
monitoring should be considered.
7
© 2010 Medco Health Solutions, Inc. All rights reserved.
Participants were from 49 US states
8
© 2010 Medco Health Solutions, Inc. All rights reserved.
Results: Baseline characteristics
Characteristic
Mean age, yrs (SD)
Male(%)
Medications (%)
Amiodarone
Statins
Sulfamethoxazole
Fluconazole
NSAID
Clopidogrel
Steroids
Conditions (%)
GI bleed
Atrial fibrillation
Pulmonary embolism
Deep vein thrombosis
Hypertension
Diabetes
Prior hospitalizations (%)
Any cause
Bleeding or thromboembolism
9
Historical control
(n=2688)
65.2 (8.0)
60.5%
Intervention group
(n= 896)
65.2 (8.3)
60.5%
P-value
0.921
1.000
4.0%
14.5%
4.4%
2.4%
19.6%
10.8%
12.4%
3.2%
16.9%
5.2%
2.6%
19.9%
10.2%
13.6%
0.313
0.071
0.268
0.803
0.865
0.574
0.354
3.6%
40.4%
11.0%
24.6%
47.0%
15.3%
4.0%
41.1%
11.8%
25.8%
54.2%
11.6%
0.539
0.709
0.501
0.489
<0.001
0.007
54.4%
23.6%
52.8%
24.8%
0.405
0.469
© 2010 Medco Health Solutions, Inc. All rights reserved.
Results: (n=424)
Warfarin Rxs within 21 days post-genotyping
10
Warfarin
sensitivity
% patients
Mean weekly
dose change (SE)
P-value
< Normal
29.0%
+6.65 mg (1.98)
<0.01
Normal
28.1%
+1.10 mg (1.40)
0.50
Mild
11.6%
+3.21 mg (3.41)
0.21
Moderate
25.0%
-3.65 mg (1.56)
<0.01
High
4.0%
-10.14 mg (3.18)
0.04
Very high
2.4%
-17.33 mg (4.54)
<0.01
© 2010 Medco Health Solutions, Inc. All rights reserved.
Results:
Six month hospitalization rates
>=1 hospitalization per 100 patients/6months
25.52
Historical control (n=2688)
Intervention group (n=896)
18.45
8.13
5.97
All cause
Bleed or thromboembolism
Intention to treat (ITT)
p-value
<0.001
0.039
Genotyping associated with 28% decrease in all-cause
hospitalizations and 27% decrease in bleed or thrombo-emboli
11
© 2010 Medco Health Solutions, Inc. All rights reserved.
Results: All-cause hospitalization rate
Intention-to-treat analyses
All cause
Hospitalization risk
30%
IG
HC
25%
20%
HR: 0.69 (CI: 0.58, 0.82)
P<0.001
15%
10%
5%
0%
0
10
20
30
40
50 60
70
80
90 100 110 120 130 140 150 160 170 180
Days after treatment onset
Controlled for age, comorbid conditions, drugs, propensity score, indications, prior GI bleed or
VTE, history of prior hospitalization
12
© 2010 Medco Health Solutions, Inc. All rights reserved.
Results:
Hospitalization rate for bleed / thromboembolism. Intention-to-treat analysis.
Bleed or thromboembolism
Hospitalization risk
10%
8%
6%
IG
HC
HR: 0.72 (CI: 0.53, 0.97)
P=0.029
4%
2%
0%
0
10
20
30
40
50 60
70
80
90 100 110 120 130 140 150 160 170 180
Days after treatment onset
Controlled for age, comorbid conditions, drugs, propensity score, indications, prior GI bleed or VTE,
history of prior hospitalization
13
© 2010 Medco Health Solutions, Inc. All rights reserved.
Results: Same time frames
External control group comparison – pre vs. post
Hospitalization rates (unadjusted)
 Any cause
 Bleeds/thromboemboli
22.8% vs. 22.7%
7.8% vs. 7.2%
Hospitalization rates (adjusted) H.R. (95% CI)
 Any cause
 Bleeding or thromboemboli
HR 0.98 (0.88-1.1)
HR 0.92 (0.76-1.1)
No difference in hospitalization rates over the
same 2 time periods in the external controls
14
© 2010 Medco Health Solutions, Inc. All rights reserved.
Limitations
Non-randomized trial design
 Entire population comparison before and after population
intervened upon (quasi-experiment)
 Intention-to-treat biased against genotyping by including events
that occurred before genotype conducted
 Participation rate high – by both participants and physicians
Hawthorne effect
 Kept the intervention to a minimum
 No payment to providers who participated
Administrative Claims Data for endpoints
 Peer-reviewed literature for algorithm with ICD-9 codes
15
© 2010 Medco Health Solutions, Inc. All rights reserved.
Conclusions
Genotyping warfarin patients resulted in ~30% risk reduction
in hospitalizations for all-cause and for bleeds/thromboemboli
Having genotyping closer to therapy initiation was even better
Physicians changed warfarin prescriptions in the direction
suggested by the Mayo genotype laboratory report
Physicians were very willing (75% agreement) to order the
genotype tests
16
© 2010 Medco Health Solutions, Inc. All rights reserved.
Future Suggestions
Further research warranted to confirm and extend our findings
 Elderly vs. non-elderly
 Role of genotyping to prevent bleeds or thromboemboli as
individual endpoints
 Impact of genotyping on INR measurements
 Economic evaluation of genotyping vs. usual care
17
© 2010 Medco Health Solutions, Inc. All rights reserved.