Pitfalls of the Current Bleeding Definitions Sunil V. Rao MD Assistant Professor of Medicine Duke University Medical Center Durham VA Medical Center Duke Clinical Research Institute

Sunil V. Rao, MD DISCLOSURES Consulting Fees Honoraria The Medicines Company, Bristol-Myers Squibb, Astra Zeneca Honoraria sanofi-aventis U.S. LLC Grants/Contracted Research Cordis, a Johnson & Johnson company, Momenta Pharmaceuticals, Inc., Portola Pharmaceuticals, Inc. I intend to reference unlabeled/ unapproved uses of drugs or devices in my presentation. I intend to reference Bivalirudin and Fondaparinux for ACS; Enoxaparin for PCI; Clopidogrel for stenting.

Disclosures Consultant, Speakers’ Bureau, Honoraria Sanofi-Aventis, Bristol Myers Squibb The Medicines Company Astra Zeneca Research funding Cordis Corporation Momenta Pharmaceuticals Portola Pharmaceuticals Off-label uses of drugs/devices may be discussed Fondaparinux in NSTE ACS Enoxaparin in PCI Bivalirudin in NSTE ACS

Bleeding definitions Pitfalls Bleeding definitions and Data elements Bleeding and outcomes What are the definitions designed to do? What should the definitions do? Future directions

Major bleeding definitions – a sampling TIMI1 GUSTO1 CURE2 ACUITY3 TRITON GUSTO, CHARISMA CURE ACUITY, HORIZONS-AMI Fatal / life threatening (related to instrumentation, spontaneous, trauma), ICH, Hb ↓ ≥5 g/dL, or absolute HCT ↓ ≥15% ICH, or causes haemodynamic compromise and requires intervention Fatal / Life threatening Fatal, ICH, requires surgery, hypotension requiring inotropes, Hb ↓ ≥5 g/dL, or transfusion ≥4 U Other major Disabling, intraocular with vision loss, or transfusion 2-3 U Intracranial, Retroperitoneal, Intraocular, Access site hemorrhage requiring surgical intervention, 5cm diameter hematoma, Reduction in hemoglobin concentration of 4g/dL without an overt source of bleeding or 3g/dL with an overt source of bleeding, Re-operation for bleeding, Use of any blood product transfusion 1. Rao SV, et al. J Am Coll Cardiol. 2006;47:809-816. 2. Yusuf S, et al. N Engl J Med. 2001;345:494-502. 3. Stone GW, et. Ak. N Engl J Med 2006;355:2203-2216

Comparing “major” bleeding across ACS trials: Is it the patients, the trial, or the definition? 9.1 1 2 3 4 5 Percent 3.7 2.4 1.7 1.5 Eptifibatide + UFH (GUSTO severe) ASA + clopidogrel Enoxaparin ASA + prasugrel The PURSUIT Investigators. N Engl J Med. 1998 ; Yusuf S, et al. N Engl J Med. 2001 SYNERGY Trial Investigators. JAMA. 2004; Wiviott SD, et al. N Eng J Med. 2007 Bhatt DL, et al. N Engl J Med. 2006;354:1706-1717.

Kaplan Meier Curves for 30-Day Death, Stratified by Bleed Severity Bleeding & Outcomes N=26,452 pts from PURSUIT, GUSTO IIb, PARAGON A & B Kaplan Meier Curves for 30-Day Death, Stratified by Bleed Severity log rank p-value for all four categories <0.0001 log-rank p-value for no bleeding vs. mild bleeding = 0.02 log-rank p-value for mild vs. moderate bleeding <0.0001 log-rank p-value for moderate vs. severe <0.001 Rao SV, et al. Am J Cardiol. 2005

TIMI bleeding and mortality N=10,974 PCI patients In-hospital death TIMI Major bleeding Transfusion ≥ 2U 1-year death Transfusion ≥ 2U Odds ratio for death Kinnaird et al. Am J Cardiol 2003

CURE Bleeding and Outcomes N=34,146 ACS Patients from OASIS Registry, OASIS-2, and CURE 25 20 15 10 5 Mortality, % 30 60 90 120 180 150 No bleeding Minor bleeding Major bleeding Life threatening Days OASIS = Organization to Assess Strategies for Ischemic Syndromes. Eikelboom JW, et al. Circulation. 2006;114:774-782.

ACUITY major bleeding and outcomes N=13,819 Manoukian SV, et. al. JACC 2007

Transfusion in ACS N=24,111 pts from PURSUIT, PARAGON B, GUSTO IIb Rao SV, et. al., JAMA 2004

Effect of Bleeding Definition on the Association of Bleeding and Death/MI N=15,858 ACS patients from PURSUIT & PARAGON B 30-day death/MI GUSTO Mild GUSTO Moderate TIMI Minimal TIMI Minor TIMI Major 1.00 Increased Risk Decreased Risk 1.00 (0.85, 1.20) 1.08 (0.85, 1.36) 1.13 (0.97, 1.33) 3.48 (2.56, 4.73) 1.92 (1.58, 2.34) 1.19 (1.01, 1.41) GUSTO Severe Rao et al. JACC 2006.

Major bleeding data elements and outcomes: N=22,000 pts from REPLACE-2, ACUITY, HORIZONS-AMI Event Hazard ratio (95% CI) Deaths within 1 y, n p TIMI major bleed 4.85 (3.56–6.60) 53 <0.001 Non-TIMI major bleed with transfusion 2.98 (2.10–4.24) 40 Non-TIMI major bleed without transfusion 1.79 (1.09–2.93) 17 0.021 Large (≥5 cm) hematoma only 1.30 (0.58–2.92) 6 0.53 Mehran R. European Society of Cardiology 2009 Congress; September 2, 2009; Barcelona, Spain.

Hematomas and outcomes: N=3342 pts from STEEPLE White HD, et. al. AHJ 2009

“Nuisance” Bleeding and Drug Discontinuation N=2360 unselected pts. receiving DES Prospective data collection Major events adjudicated Serebruany bleeding classification* % discontinued Even “nuisance’ bleeding can impact adherence. In this study, 11% of patients with nuisance bleeding discontinued their clopidogrel despite having a DES *Alarming bleeding = ICH, life-threatening, + transfusion Internal bleeding = hematoma, epistaxis, mouth or vaginal, Melena, IO, hematuria or hematemesis Nuisance bleeding = bruising, petechiae, ecchymosis Overall rate of bleeding=32.4% Roy P, et. al. AJC 2008

Bleeding and Evidence-based Therapies N=2498 ACS patients from the PREMIER Registry Discharge ASA and thienopyridine Pts. with bleeding vs. pts. without bleeding Aspirin OR (95% CI) Thienopyridine OR (95% CI) 0.45 (0.31, 0.64) 0.62 (0.42, 0.91) Discharge 1 Month 6 Months 1 Year 0.68 (0.50, 0.92) 0.83 (0.59, 1.17) 0.63 (0.46, 0.87) 1.06 (0.78, 1.45) Regardless of the bleeding definition that is used, the occurrence of a hemorrhagic complication is associated with a significant decrease in long-term medication adherence as seen in this study from the PREMIER registry 0.94 (0.66, 1.34) 1.12 (0.81, 1.55) 0.5 1.0 2.0 0.5 1.0 2.0 Wang TY, et. al. Circulation 2008

Deconstructing Definitions: Importance of data elements Bleeding definition Data Elements Clinical Laboratory Consequences ICH Hematoma Hemodynamic compromise Transfusion Fatal Hgb decrease Severity Classification

What are bleeding definitions designed to do? Assess the bleeding risk of a therapeutic strategy BUT…the “risk” can be dialed up or down depending on the definition So…what should bleeding definitions do? They should allow clinicians to weight the risks and benefits of one therapy against another

What data elements should bleeding definitions use? Rao SV, et. al. AHJ 2009

ABC Recommendations Rao SV, et. al. AHJ 2009

Pitfalls of the current bleeding definitions Little overlap in data elements Some “major” definitions are considered “minor” by other scales and vice versa Risk of “gaming” safety data Although many appear to be associated with mortality, very few studies have directly compared two definitions in terms of prognosis

Arguments for and against a universal bleeding definition Minimizes “gaming” Allows for comparison across therapies Would include clinically important data elements There is precedent (MI, stent thrombosis) Many definitions are associated with mortality Doesn’t allow tailoring of trial designs Medical Rx vs. PCI Allows for post-hoc (or prespecified) reconstruction of traditional bleeding definitions

Pitfalls of the current bleeding definitions Conclusions Many definitions are associated with adverse outcomes Data elements are important Some overlap across definitions; Many do not “Minor” bleeding appears important in terms of medication adherence Some studies call this “Major” bleeding

Pitfalls of the current bleeding definitions Conclusions There are arguments for and against a universal bleeding definition But, the time has probably come for a consensus definition BARC meeting – 2/25/10

Duke Univ. Medical Center Duke Clinical Research Institute