Presenter Disclosure Information DISCLOSURE INFORMATION: The following relationships exist related to this presentation Stock options None; Consults for None Research support Schering Plough, Guilford Pharmaceuticals, Aventis Atul Aggarwal, MD Characteristics and Outcomes of Patients Taking Warfarin Prior to Percutaneous Coronary Intervention

Atul Aggarwal, 1 David Dai, 2 John S. Rumsfeld, 3 Lloyd W. Klein, 4 and Matthew T. Roe, 2 on behalf of the American College of Cardiology – National Cardiovascular Data Registry (NCDR) Nebraska Heart Institute, Hastings, NE, 1 Duke Clinical Research Institute, Durham, NC, 2 Denver VA Medical Center/ University of Colorado, Denver, CO 3 and Rush Medical College, Chicago, IL 4 Characteristics and Outcomes of Patients Taking Warfarin Prior to Percutaneous Coronary Intervention

Background There is little information available about clinical characteristics, peri-procedural management, and in-hospital outcomes of patients taking warfarin who undergo percutaneous coronary intervention (PCI) Prior reports suggest increased risk of bleeding with use of warfarin after PCI

Methods Patients undergoing PCI in American College of Cardiology – National Cardiovascular Data Registry from January 1 st, 2004 till March 30, 2006 were evaluated (n=307,443) Complete definitions of all variables were prospectively defined and are available at the web site

Data collection Information collected for patients taking warfarin at home prior to PCI and compared with those not taking prior warfarin Patients stratified according to the urgency of the procedure –Urgent PCI defined as cardiogenic shock at admission, presentation with ST elevation myocardial infarction with onset of symptoms within 24 hours of performance of PCI, or primary, rescue or facilitated PCI –All other PCI procedures categorized as elective

Event Definitions Bleeding defined as during or after catheterization laboratory visit until discharge, and required a transfusion and/or that prolonged the hospital stay, and/or associated with hemoglobin drop >3.0 gm/dl Bleeding events reported by sites; not centrally adjudicated Bleeding at percutaneous entry site could be external or hematoma >10 cm femoral access, or >2 cm for radial access, or >5 cm for brachial access A derived composite bleeding endpoint utilized based upon site reported bleeding location - bleeding at percutaneous entry site / retroperitoneal / gastrointestinal / genitourinary bleeding / bleeding site not classified Intracranial bleeding events not specified separately Patients who died on the day of performance of PCI excluded from this outcome

Statistical Analysis Continuous/ordinal variables analyzed with use of chi-square rank based group means score statistics, and categorical variables by Pearson chi-square tests Primary outcome measures of mortality and that of composite bleeding complications were analyzed among patients taking warfarin compared with those not taking warfarin Patients within a hospital were more likely to be similar, so analysis performed with generalized estimating equation (GEE) models to account for correlations among clustered responses (that is within-hospital correlations) after adjusting for baseline clinical differences

Results

Clinical Characteristics Of the 307,443 patients who underwent PCI, 11,173 (3.6%) were receiving warfarin before PCI, and 44,443 patients (15%) underwent urgent PCI as previously defined

Elective PCI (n=263,000)Urgent PCI (n=44,443) Warfarin (n=10002) No Warfarin (n=252998) P-valueWarfarin (n=1171) No Warfarin (n=43272) P-value Demographics Age (years)70± ±11.9< ± ±13.0< Male sex Risk factors for Coronary Artery Disease Diabetes mellitus (%) < < Hypertension (%) < < Tobacco abuse (%) < < Family history (%) < < Dyslipidemia (%) < Medical History Old MI >7 days (%) < < History of CHF (%) < < Previous PCI (%) < Previous CABG (%) < < Prior valve surgery (%)8.50.8< < Cerebrovascular disease (%) < < Cardiogenic shock (%)1.20.8< < Results – Clinical Characteristics

Concomitant Medications

Elective PCIUrgent PCI WarfarinNo Warfarin P-valueWarfarinNo Warfarin P-value Aspirin (%) < GP IIb-IIIa antagonist (%) < < Theinopyridine (%) < Any heparin (%) < Any heparin plus any DTI (%) Concomitant Medications

Procedural Characteristics

Elective PCI (n=263000)Urgent PCI (n=44443) Warfarin No Warfarin P-valueWarfarinNo Warfarin P-value Ejection fraction (%) 48.1±1553.8±12< ±1446.6±13< IABP use (%) < Multi-vessel (%) < High risk (type C, %) Vein graft lesion (%) < <0.001 Stents per procedure 1.5± ± DES use (%) < < Post procedure TIMI 3 flow (%) No-reflow (%) Procedural Characteristics

Clinical Outcomes

Unadjusted

Numbers in percentages W W NoW Unadjusted In-Hospital Mortality urgent elective

Numbers in percentages W W NoW Unadjusted In-Hospital Bleeding urgent elective

Elective PCIUrgent PCI WarfarinNo Warfarin P-valueWarfarinNo Warfarin P-value In-hospital Mortality (%) < < Periprocedural MI (%) Cardiogenic shock (%) < CHF (%)1.60.7< < Stroke (%)0.80.5< Renal failure (%) < Patients who died on the same day as PCI are excluded (n=786) Unadjusted Clinical Outcomes

Elective PCI (85%)Urgent PCI (15%) WarfarinNo Warfarin P-valueWarfarinNo Warfarin P-value Bleeding, composite (%) < < Bleeding at access site (%) < Retroperitoneal bleeding (%) Gastrointestinal bleeding (%) < Genitourinary bleeding (%) < Bleeding site not classified (%) < < Cardiac tamponade (%) Unadjusted Clinical Bleeding Outcomes

Adjusted Clinical Outcomes 1.07 ( ) 1.26 ( ) 0.90 ( ) 1.42 ( ) In-hospital BleedingIn-hospital Mortality Elective PCI Urgent PCI Elective PCI Urgent PCI Odds Ratio P Value

Registry did not collect specific indication for warfarin use before and after PCI Does not include prothrombin time prior to PCI or timing of last dose of warfarin before PCI Don’t have data on specific doses of heparins or glycoprotein IIb-IIIa antagonists used during the peri-procedural period; these medications frequently overdosed in elderly population also more likely to be taking warfarin prior to PCI Unclear whether these medications dose adjusted for renal function that was worse in patients taking warfarin Thus, bleeding complications may have been exaggerated in patients taking warfarin prior to performance of PCI since we cannot account for overdosing of concomitant antithrombotic medications Limitations

Conclusions Patients taking warfarin prior to elective and urgent PCI were at increased risk of bleeding complications, despite the fact that they were less likely to receive aspirin, theinopyridines and glycoprotein IIb-IIIa antagonists during PCI No association was observed between warfarin use and risk- adjusted in-hospital mortality Warfarin use is largely a marker for co-morbidities with regard to in-hospital mortality risk Additional studies needed to determine if the higher bleeding risk translates into worse longer-term outcomes, to delineate the optimal timing of elective PCI in patients taking chronic warfarin, and to ascertain optimal use / dosing of anti-platelet and anti-thrombin agents for patients on home warfarin