Charisma Genomics

CHARISMA Genomics

Deepak L. Bhatt MD, MPH, Katy L. Simonsen PhD, Eileen S. Emison PhD, Keith A. A. Fox MBChB, P. Gabriel Steg MD, Gilles Montalescot MD, PhD, Nihar Bhakta MD, Werner Hacke MD, Marcus D. Flather MD, Patrice Cacoub MD, Mark A. Creager MD, Peter B. Berger MD, Steven R. Steinhubl MD, Gurunathan Murugesan PhD, Kandice Kottke-Marchant MD, PhD, A. Michael Lincoff MD, Eric J. Topol MD, on Behalf of the CHARISMA Executive Committee and Investigators

Disclosure for Dr. Bhatt Dr. Bhatt has served as a consultant to: Arena, Astra Zeneca, Bristol-Myers Squibb, Cardax, Cogentus, Daiichi Sankyo, Eli Lilly, Eisai, Glaxo Smith Kline, Johnson & Johnson, Medtronic, Millennium, Otsuka, Paringenix, PDL, Philips, Portola, sanofi aventis, Schering Plough, Takeda, The Medicines Company, Vertex. Principal Investigator for several potentially related studies. His institution has received funding from Bristol-Myers Squibb, Eisai, Ethicon, Heartscape, Sanofi Aventis, The Medicines Company. This presentation discusses off-label and/or investigational uses of various drugs and devices. This CHARISMA study was funded by sanofi aventis and Bristol-Myers Squibb. The genetic analyses were done by Bristol-Myers Squibb’s Human Genetics and Statistical Genetics Groups.

Variability in Clopidogrel Responsiveness in a Diverse Population of 544

Serebruany VL, Steinhubl SR, Berger PB, et al. JACC 2005; 45:246 –51.

Genetic Variations and Clopidogrel Response

Mega JL, Close SL, Wiviott SD, et al. N Engl J Med. 2009;360:354-362.

Genetic Variations and Clopidogrel Response Heterozygotes vs Homozygotes •

Probability of all-cause death, recurrent myocardial infarction, and stroke (%) according to CYP2C19 loss of function variant allele polymorphisms in the FAST MI registry

Simon T, et al. N Engl J Med. 2009;360:363-375.

Primary Efficacy Results (MI/Stroke/CV Death) by Pre-Specified Entry Category Population

RR (95% CI)

p value

Qualifying CAD, CVD or PAD *

0.88 (0.77, 0.998) 0.046

(n=12,153)

Multiple Risk Factors *

1.20 (0.91, 1.59) 0.20

(n=3,284)

Overall Population†

0.93 (0.83, 1.05) 0.22

(n=15,603) 0.4

0.6 0.8

Clopidogrel + ASA Better

1.2

1.4 1.6

Placebo + ASA Better

* A statistical test for interaction showed marginally significant heterogeneity (p=0.045) in treatment response for these prespecified subgroups of patients † 166 patients did not meet any of the main inclusion criteria

Bhatt DL, Fox KA, Hacke W, et al. NEJM 2006; 354:1706-17.

Cumulative event rate (%)

Overall Population: Principal Secondary Efficacy Outcome (MI/Stroke/CV Death/Hospitalization)† 20

Placebo + ASA* 17.9%

15

Clopidogrel + ASA* 16.7%

10

RRR: 7.7% [95% CI: 0.5%, 14.4%] p = 0.04

5

0

0

6

12 18 24 Months since randomization§

30

First Occurrence of MI, Stroke, CV Death, or Hospitalization for UA, TIA, or Revascularization *All patients received ASA 75-162 mg/day §The number of patients followed beyond 30 months decreases rapidly to zero †

Bhatt DL, Fox KA, Hacke W, et al. NEJM 2006; 354:1706-17.

Primary Endpoint (MI/Stroke/CV Death) in Patients With Previous MI, IS, or PAD* “CAPRIE-like Cohort”

* Post hoc analysis.

Bhatt DL, Flather MD, Hacke W, et al. J Am Coll Cardiol. 2007;49:1982-1988.

Aims •

To determine the effect of CYP2C19 polymorphisms on CV death, MI, stroke in patients randomized to clopidogrel versus placebo in a stable CV population

•

To assess the impact on severe or moderate GUSTO bleeding

•

To assess the impact on CV death, MI, stroke, or hospitalization for ischemic events (more events)

•

To examine effects in higher risk subgroups

•

To assess the impact on any GUSTO bleeding (more events)

Methods •

A subset (N=4862) of the 15,603 patients enrolled into CHARISMA who consented were genotyped for CYP2C19*2, *3, and *17 alleles

•

Genotyping details –

*2 and *17 were genotyped by Restriction Fragment Length Polymorphism

–

*3 was genotyped by Taqman allelic discrimination assay

–

Accuracy: 11% of samples were genotyped in duplicate; no errors were identified in the replicates

•

Statistical methods: –

A Cox model of time to primary event, with treatment, genotype, and treatment by genotype interaction terms, was used to elucidate the genotype effect

–

Covariates used in the genetic analyses were: inclusion criterion (symptomatic versus asymptomatic), prior use of statins, prior use of calcium channel blockers, and smoking

–

Bleeding events were analyzed with logistic regression using the same terms

Genotype Frequencies n = 4862

K-M Survival Curves by Genotype

Uncorrected p-value for testing the difference of time to primary event between genotype groups, from the log-rank test: 0.166

Uncorrected p-value for testing the difference of time to primary event between genotype groups, from the log-rank test: 0.162

Subjects with genotype frequencies < 5 were excluded from the analysis. Survival curves are truncated at 30 months.

Cox Model HRs by Treatment Arm Effect Treatment vs Placebo

p value 0.33

HR

95% CI for HR

1.209

(0.83, 1.77)

Effect of Genotype in Placebo Group *2/WT vs WT/WT

0.54

0.852

(0.51, 1.42)

*2/*2 vs WT/WT

0.19

1.852

(0.74, 4.65)

*2/*17 vs WT/WT

0.47

1.285

(0.65, 2.54)

*17/WT vs WT/WT

0.052

1.486

(1.00, 2.21)

*17/*17 vs WT/WT

0.71

0.841

(0.33, 2.11)

Effect of Genotype in Clopidogrel Group *2/WT vs WT/WT

0.63

1.112

(0.72, 1.71)

*2/*2 vs WT/WT

0.022

2.383

(1.14, 5.00)

*2/*17 vs WT/WT

0.36

1.322

(0.72, 2.42)

*17/WT vs WT/WT

0.72

0.927

(0.61, 1.40)

*17/*17 vs WT/WT

0.44

0.696

(0.28, 1.74)

*2/*2 has increased risk in clopidogrel arm, but much of this risk is also present in placebo arm

Hazard Ratios

* Primary Analysis

Caveat: Small Number of Primary Events Placebo Genotype

Number (%) of Subjects with Primary Event

Clopidogrel Total

Genotype

Number (%) of Subjects with Primary Event

Total

WT/WT

49 (5.05%)

971

WT/WT

57 (6%)

950

*2/WT

21 (4.29%)

489

*2/WT

33 (6.73%)

490

*2/*2

5 (8.77%)

57

*2/*2

8 (13.79%)

58

*2/*17

10 (6.29%)

159

*2/*17

13 (7.65%)

170

*17/WT

48 (7.48%)

642

*17/WT

37 (5.75%)

643

*17/*17

5 (4.42%)

113

*17/*17

5 (4.42%)

113

*3/WT

0 (0%)

1

*3/WT

0 (0%)

2

*2/*3

0 (0%)

2

*2/*3

0 (0%)

2

Secondary Endpoint: Hazard Ratios

Symptomatic Subpopulation (N=3666): Hazard Ratios

CAPRIE-like Subpopulation (N=2773): Hazard Ratios

GUSTO moderate and severe: Logistic Regression ORs by Treatment Arm Effect Treatment vs Placebo

p value

OR

95% CI for OR

0.407

1.228

(0.756, 2.007)

Effect of Genotype in Placebo Group *2/WT vs WT/WT

0.007

0.313

(0.106, 0.742)

*2/*2 vs WT/WT

0.420

1.707

(0.400, 4.985)

*2/*17 vs WT/WT

0.644

0.785

(0.231, 2.018)

*17/WT vs WT/WT

0.302

0.723

(0.377, 1.330)

*17/*17 vs WT/WT

0.709

0.801

(0.190, 2.291)

Effect of Genotype in Clopidogrel Group *2/WT vs WT/WT

0.297

1.322

(0.777, 2.214)

*2/*2 vs WT/WT

0.033

0.000

–

*2/*17 vs WT/WT

0.823

0.905

(0.339, 2.029)

*17/WT vs WT/WT

0.906

1.031

(0.619, 1.714)

*17/*17 vs WT/WT

0.054

0.217

(0.012, 1.018)

• Significantly decreased bleeding risk for *2/wt compared with wt/wt on placebo • Result is based on very few events • No conclusions for *2/*2 (3 vs 0 events)

All GUSTO bleeding events: Logistic Regression ORs by Treatment Arm Effect

p value

OR

95% CI for OR

Treatment vs Placebo

<0.0001

2.4110

(1.978, 2.943)

Effect of Genotype in Placebo Group *2/WT vs WT/WT

0.6392

1.0640

(0.821, 1.374)

*2/*2 vs WT/WT

0.3528

0.7250

(0.340, 1.400)

*2/*17 vs WT/WT

0.5803

1.1180

(0.748, 1.641)

*17/WT vs WT/WT

0.8618

0.9790

(0.769, 1.243)

*17/*17 vs WT/WT

0.3477

0.7910

(0.469, 1.278)

Effect of Genotype in Clopidogrel Group *2/WT vs WT/WT

0.1712

0.8550

(0.683, 1.070)

*2/*2 vs WT/WT

4 x 10-4

0.3290

(0.160, 0.619)

*2/*17 vs WT/WT

0.0652

0.7270

(0.513, 1.020)

*17/WT vs WT/WT

0.9222

1.0100

(0.824, 1.238)

*17/*17 vs WT/WT

0.2236

1.2770

(0.860, 1.891)

• Significantly more bleeding events in wt/wt subjects on clopidogrel compared with placebo • Significantly reduced risk of bleeding events for the *2/*2 genotype vs wt/wt on clopidogrel • Risk in *2/*2 is not significantly different between

Limitations •

Genotyped patients differed from non-genotyped patients

•

Overall trial was negative for the primary endpoint, so power is somewhat limited – Secondary endpoint was positive and more than doubles events, but still no relationship with heterozygotes and outcomes – Higher risk subgroups also did not find relationship with heterozygotes and outcomes – Still, more events in a placebo controlled trial than any other study to date

•

Stable population, so results seen here may not apply to ACS – But perhaps more relevant for chronic therapy

Conclusions •

No relationship seen between CYP2C19*2 heterozygotes and outcomes

•

Small % of homozygotes may have worse outcome, though this is noted to an extent in the placebo arm as well

•

First large study to establish potential relationship between less bleeding and genotype

•

Further prospective study needed to determine clinical relevance of CYP2C19 polymorphisms on efficacy/bleeding - and appropriate clinical action to take - before routine testing can be recommended

Acknowledgements •

Bristol-Myers Squibb Human Genetics Group for the genotyping – Terrye A. DelMonte, B.S. – Lester E. Hui, B.S.

•

Bristol-Myers Squibb Statistical Genetics and Biomarkers Group for statistical analyses – Oksana Mokliatchouk, Ph.D. – Lisa R. Denogean, Ph.D.

•

Lionel Thevathasan, M.D., sanofi aventis, for reviewing

•

Stuart Wakelin, Ph.D. and George Clarkson, Ph.D. from Alpha-Plus Medical Communications Ltd for graphical support (funded by sanofi aventis)

Q+A

Backup Slides

GUSTO moderate and severe: Counts and Interactions by Treatment Arm Genotype

WT/WT

*2/WT

*2/*2

*2/*17

*17/WT

*17/*17

*3/WT

*2/*3

Treatment

No. (%) with Mod / severe bleeds

Total

p-value Rel to wt

Clopidogrel

37

(3.89%)

950

Placebo

31

(3.19%)

971

Clopidogrel

25

(5.10%)

490

0.297

Placebo

5

(1.02%)

489

0.007

Clopidogrel

0

(0%)

58

0.033

0.407

Placebo

3

(5.26%)

57

0.420

Clopidogrel

6

(3.53%)

170

0.823

Placebo

4

(2.52%)

159

0.644

Clopidogrel

26

(4.04%)

643

0.906

Placebo

15

(2.34%)

642

0.302

Clopidogrel

1

(0.88%)

113

0.054

Placebo

3

(2.65%)

113

0.709

(0%)

2

(0%)

1

(0%)

2

(0%)

2

Clopidogrel Placebo Clopidogrel Placebo

0 0 0 0

p-value interaction

0.005

0.027

0.839

• Significantly decreased bleeding risk for *2/wt compared with wt/wt on placebo • Result is based on very few events

0.388

0.239

• No conclusions for *2/*2 (3 vs 0 events)

All GUSTO bleeding events: Counts and Interactions by Treatment Arm Genotype

WT/WT

*2/WT

*2/*2

*2/*17

*17/WT

*17/*17

*3/WT

*2/*3

Treatment

No. (%) with bleeding event

Total

p-value Rel to wt

Clopidogrel

392

(41.3%)

950

Placebo

220

(22.7%)

971

Clopidogrel

184

(37.6%)

490

0.171

Placebo

116

(23.7%)

489

0.639

Clopidogrel

11

(19.0%)

58

0.0004

Placebo

10

(17.5%)

57

0.353

Clopidogrel

58

(34.1%)

170

0.065

Placebo

39

(24.5%)

159

0.580

Clopidogrel

266

(41.4%)

643

0.922

Placebo

143

(22.3%)

642

0.862

Clopidogrel

53

(46.9%)

113

0.224

Placebo

21

(18.6%)

113

0.348

Clopidogrel

0

(0%)

2

Placebo

1

(100%)

1

Clopidogrel

1

(50%)

2

Placebo

1

(50%)

2

p-value interaction <0.0001

0.211

0.113

• Significantly more bleeding events in wt/wt subjects on clopidogrel compared with placebo

0.107

0.845

0.134

• Significantly reduced risk of bleeding events for the *2/*2 genotype vs wt/wt on clopidogrel • Risk in *2/*2 is not significantly