Insights from Bleeding in TRITON TIMI 38 Stephen

Insights from Bleeding in TRITON – TIMI 38 Stephen D. Wiviott, MD Cardiovascular Division Brigham and Women’s Hospital Harvard Medical School Investigator, TIMI Study Group

Stephen D. Wiviott, MD § Consulting Fees – ARENA – sanofi-aventis – Bristol-Myers Squibb – Astra. Zeneca § Honoraria – Eli Lilly and Company – Daiichi Sankyo – Astra. Zeneca – Novartis AG

§ Grants/Contracted Research – Eli Lilly and Company – Daiichi Sankyo – Merck & Co. , Inc. My discussion will include off label discussions. I intend to discuss results of clinical trials of novel antiplatelet agents

Lessons From TRITON – TIMI 38 More intensive treatment with antiplatelet therapy (with prasugrel) can be expected to reduce stent thrombosis and increase bleeding in all groups of patients. Striking a balance in individual patients results from weighing these competing risks.

Balance of Efficacy and Safety 15 138 events Clopidogrel 12. 1 CV Death / MI / Stroke 9. 9 Endpoint (%) 10 Prasugrel 5 TIMI Major Non. CABG Bleeds Prasugrel Clopidogrel 35 events 2. 4 1. 8 0 0 30 60 90 180 Days 270 360 HR 0. 81 (0. 73 -0. 90) P=0. 0004 NNT = 46 450 HR 1. 32 (1. 03 -1. 68) P=0. 03 NNH = 167 Wiviott SD, Braunwald E, Mc. Cabe CH et al NEJM 2007

Major Bleed (non CABG) (Landmark Analysis - 30 days) Major Bleed (%) 3 HR 1. 19 P=0. 34 2 HR 1. 48 P=0. 028 Prasugrel 1. 03 Clopidogrel 0. 87 Prasugrel 1. 42 Clopidogrel 0. 97 1 0 0 5 10 15 Early Risk 20 25 30 306090 Days 180 270 Late Risk 360 450

Serious Bleeding Prasugrel (n = 6713) Clopidogrel (n = 6707) HR for prasugrel (95% CI) P-value 303 (4. 51) 231 (3. 44)* 1. 32 (1. 11 -1. 56) 0. 002 146 (2. 17) 112 (1. 67)* 1. 31 (1. 02 -1. 67) 0. 03 Fatal TIMI major bleeding 21 (0. 31) 6 (0. 09)* 3. 50 (1. 41 -8. 68) 0. 01 Non-fatal TIMI major bleeding 125 (1. 86) 106 (1. 58) 1. 20 (0. 93 -1. 56) 0. 17 TIMI minor bleeding 164 (2. 44) 124 (1. 85) 1. 33 (1. 05 -1. 68) 0. 02 Transfusion 178 (2. 65) 134 (2. 00) 1. 33 (1. 07 -1. 67) 0. 01 Serious bleeding TIMI major bleeding

Minimal Bleeding Minimal bleeding 460 (6. 85) 313 (4. 68) 1. 48 (1. 29 -1. 71) <0. 001 115 (1. 71) 83 (1. 24) 1. 39 (1. 05 -1. 84) 0. 02 Retroperitoneal 3 (0. 04) 1. 00 (0. 20 -4. 96) 1. 00 Gastrointestinal 94 (1. 40) 72 (1. 07) 1. 31 (0. 96 -1. 78) 0. 09 Hemoptysis 12 (0. 18) 13 (0. 19) 0. 93 (0. 42 -2. 03) 0. 85 Hematuria 43 (0. 64) 28 (0. 42) 1. 54 (0. 96 -2. 48) 0. 08 Epistaxis 101 (1. 50) 53 (0. 79) 1. 92 (1. 37 -2. 67) <0. 001 Other 101 (1. 50) 70 (1. 04) 1. 45 (1. 07 -1. 96) 0. 02 Location: Puncture Site

Predictors of Serious Bleeding Hazard ratio (95% CI) P-value Strength of association with bleeding* 1. 77 (1. 44 -2. 18) <0. 001 28. 79 1. 59 (1. 29 -1. 95) <0. 001 19. 33 1. 07 (1. 04 -1. 10) <0. 001 17. 98 1. 22 (1. 09 -1. 38) <0. 001 11. 07 1. 34 (1. 12 -1. 60) 0. 001 10. 19 0. 70 (0. 53 -0. 92) 0. 01 8. 62 0. 73 (0. 53 -1. 00) 0. 05 8. 62 1. 35 (1. 10 -1. 66) 0. 005 7. 98 1. 60 (1. 07 -2. 39) 0. 02 5. 23 1. 05 (1. 01 -1. 09) 0. 03 4. 84 0. 82 (0. 68 -0. 99) 0. 04 4. 34 1. 24 (1. 00 -1. 54) 0. 05 4. 02 Predictors of any serious bleeding Female sex GPIIb/IIIa inhibitor used Duration of intervention (per 10 min. intervals) Age (by decade) Assignment to prasugrel (vs. clopidogrel) Eastern Europe (vs. North America) Middle East, Africa, or Asia–Pacific (vs. North America) ST-elevation myocardial infarction Femoral access Creatinine clearance (per 10 ml/min decrease) Hypercholesterolemia Hypertension

Non-CABG TIMI Major Bleeding (After 3 days) for Prasugrel Group Impact of Weight and Age FDA Advisory Board Presentation, Washington DC Feb 2

Balance of Efficacy and Safety in Patients < 75 Yrs, ≥ 60 kg, and without Prior TIA/Stroke 16 CV death, NF MI, or NF stroke 14 Endpoint (%) 12 10 8 Clopidogrel 11. 0% Hazard Ratio, 0. 75 (95% CI, 0. 66 - 0. 84) P < 0. 001 Prasugrel 8. 3% 6 4 TIMI major bleeding Hazard Ratio, 1. 240 (95% CI, 0. 91 - 1. 69) P = 0. 17 2 Prasugrel 2. 0% Clopidogrel 1. 50% 0 0 30 90 180 270 360 450 Days 2095. 01 Modified from Wiviott SD et al. NEJM 2007; 357: 2001 -2015

Femoral vs. Radial Artery Access Instrumented TIMI Non-CABG Bleeding Femoral = 12273 Radial = 1120 Percentage P=0. 002 P=0. 004 P=0. 013 P=0. 058

Femoral Artery Access: Prasugrel vs. Clopidogrel Instrumented TIMI Bleeding Percentage P=0. 0017 N. S. P value for interaction N. S.

Radial Artery Access: Prasugrel vs. Clopidogrel Instrumented TIMI Bleeding Percentage All comparisons N. S. P value for interaction N. S.

Time Dependence of Bleeding Mortality Association • Multivariable Cox model including treatment, baseline and procedural variables, and a propensity score for bleeding • Multiple iterative landmark analyses starting from day 1 up to day 180 after the bleeding event. 95%-confidence interval HR for landmark day 1 (= day 1 after bleeding event compared to subjects without bleeding event 1 out of their qualifying event) HR for landmark day 2 HR for landmark day 3 HR for landmark day 4. . . Hochholzer, W AHA 2010

Serious bleeding (Major plus Minor) HR for this landmark day 95%-confidence interval 17

Spontaneous serious bleeding HR for this landmark day 95%-confidence interval 18

Instrumented or traumatic serious bleeding HR for this landmark day 95%-confidence interval 19

Indirect Comparaison of Loading Doses of P 2 Y 12 inhibitors Silvain J. and Montalescot G. Circ Cardiovasc Interv 2011; 4; 95 -103

TIMI major non-CABG related bleeding Silvain J. and Montalescot G. Circ Cardiovasc Interv 2011; 4; 95 -103

Intracranial and Fatal bleeding Silvain J. and Montalescot G. Circ Cardiovasc Interv 2011; 4; 95 -103

Lessons • Prasugrel led to more bleeding of all types across multiple subgroups • Patient and procedural characteristics are strong correlates of bleeding in addition to therapy • Serious bleeding was associated with increased mortality in the short/intermediate term, with a greater and longer duration risk from spontaneous bleeding • Bleeding increases were consistent across otherapies with increased IPA

Other bleeding categories 24

Conclusions • The impact of serious bleeding events on mortality was dynamic, with a significant increase in risk within the first few days. However, the impact of bleeding on mortality waned rapidly and the HR was not significantly different from 1. 0 after 30 days. • The impact of instrumented or traumatic serious bleeding on mortality was of short duration (i. e. , less than a week). • Serious spontaneous bleeding tended to have a longer impact on mortality with a significantly elevated HR for approximately one month and a non-significant trend beyond that. The more prolonged impact of spontaneous serious bleeding might in part be explained by the underlying conditions contributing to the bleeding event that were not fully represented by the available patient characteristics. 25

Femoral vs. Radial Artery Access Femoral artery access (FAA) – N = compared to radial artery access (RAA) FAA (N=12 273) % RAA (N=1 120) % HR P Value 2. 2 1. 3 1. 70 (0. 98 -2. 98) 0. 0582 Instrumented 0. 7 0. 3 2. 40 (0. 76 -7. 60) 0. 1244 Spontaneous 1. 3 1. 0 1. 45 (0. 74 -2. 85) 0. 2736 TIMI Major & Minor non-CABG 4. 5 2. 4 1. 94 (1. 29 -2. 93) 0. 0012* Instrumented 1. 7 0. 6 3. 09 (1. 37 -6. 95) 0. 0042* Spontaneous 2. 3 1. 8 1. 34 (0. 82 -2. 19) 0. 2379 3. 6 1. 8 2. 05 (1. 28 -3. 29) 0. 0023* TIMI Major non-CABG Transfusion * significant

Bleeding, Transfusion & Method of Arterial Access: Prasugrel vs. Clopidogrel Prasugrel % Clopidogrel % HR P Value* N=4228 N=4200 TIMI Major non-CABG Instrumented 0. 7 0. 5 1. 31 (0. 75 -2. 28) 0. 3352 TIMI Major & Minor non-CABG Instrumented 1. 7 1. 5 1. 14 (0. 81 -1. 60) 0. 4509 Transfusion 4. 2 2. 9 1. 46 (1. 15 -1. 86) 0. 0017 N=1920 N=1922 TIMI Major non-CABG Instrumented 0. 7 0. 87 (0. 41 -1. 83) 0. 7133 TIMI Major & Minor non-CABG Instrumented 2. 2 1. 6 1. 31 (0. 81 -2. 10) 0. 2711 Transfusion 4. 1 3. 3 1. 19 (0. 84 -1. 69) 0. 3236 N=564 N=556 TIMI Major non-CABG Instrumented 0. 4 0. 2 1. 97 (0. 18 -21. 78) 0. 5712 TIMI Major & Minor non-CABG Instrumented 0. 5 0. 7 0. 99 (0. 20 -4. 90) 0. 9893 Transfusion 1. 5 2. 2 0. 62 (0. 24 -1. 61) 0. 3252 FAA-NC FAA-CD RAA *Interaction P-values not significant