1. BRAVO and oral GP IIb/IIIa Clinical Trial Commentary Dr Eric Topol Chairman and Professor, Department of Cardiology Director of the Joseph J Jacobs Center for Thrombosis and Vascular Biology at the Cleveland Clinic Dr Robert Califf Professor of Cardiology Associate Vice Chancellor for Clinical Research at Duke University

2. Why BRAVO was stopped BRAVO Adverse events at safety review (n=9200) OutcomeLotrafiban %Placebo %p-value Mortality2.72.00.022 Thrombocytopenia2.20.5 Major bleeding4.21.3<0.0001

3. Meta-analysis of oral IIb/IIIa trials EXCITE OPUS SYMPHONY Pooled 1.0% 1.4% 1.8% 1.4% 1.35% 1.95% 2.00% 1.80% Death Odds Ratio & 95% CI Trial Placebo Fiban Placebo BetterFiban Better 7,232 10,302 9,169 26,703 N p = 0.616 Breslow- Day homogeneity 00.511.52 Xemilofiban Orbofiban Sibrafiban p = 0.023 1.36 1.40 1.12 1.27 BRAVO

4. What is the cause? 3 leading explanations Partial agonist effect: Less than 80% blockade, resulting in activation and a pro-thrombotic effect. Inflammatory process: Sub-threshold receptor blockade could release CD-40 ligand and selectin, provoking inflammation. Apoptosis: Direct inducement of cardiomyocyte apoptosis, causing arrhythmias. BRAVO

5. Partial agonist effect Can’t reach 80% blockade with oral therapy because of bleeding complications. Oral GP IIb/IIIa inhibition is left at 30-50% blockade. Intravenous administration has been able to reach 80-90% in short-term, acute use. This sub-threshold level results in platelet activation. BRAVO

6. Thrombotic path What is activated? Prothrombin complex is activated. (But no corresponding increase of thrombotic events has been shown) P-selectin is activated. (adhesion receptor on platelet surface) The thrombosis supposition is suspect. (Many negative papers studying this effect have not yet been published) BRAVO

7. Inflammatory path CD-40 ligand shedding caused by sub- threshold receptor blockade CD-40 is a pro-inflammatory mediator, stimulates white cells, and could also be involved in a thrombotic pathway Studies presented at AHA 2000 are not yet published. BRAVO

8. Apoptosis Fitzgerald found that GP IIb/IIIa inhibitors can provoke cell death in rat cardiomyocytes. Effect was dose dependent, which doesn’t fit with the partial agonist idea. If this was the operative mechanism, why did it take a year for the deaths to show up in the BRAVO trial? BRAVO

9. Questions raised 1)What does this mean for the one ongoing trial? 2) What does this mean for the intravenous IIb/IIIa inhibitors? 3) Does this mean anything for the other so- called antiplatelet drugs? What is a class? BRAVO

10. PLA polymorphism Desmond Fitzgerald has suggested that the OPUS trial saw outcomes dependent on PLA polymorphism PLA-2 is present in 20%-25% of patients and is already associated with worse outcomes Data on any interaction not yet in print BRAVO

11. “I've never been more convinced of a bad drug class than this. Because this is so intensively studied. But I think it would be obviously only with the oral IIb/IIIa. […] but this class kind of reeks of badness from the biology, and obviously worse than that, much worse than that, is the consistent death risk of 35% increase in death in 43 000 patients, 4 agents, 5 trials. I don't know how much more convincing one would need.” Dr Eric Topol Chairman and Professor, Department of Cardiology Director of the Joseph J Jacobs Center for Thrombosis and Vascular Biology at the Cleveland Clinic BRAVO Oral GP IIb/IIIa inhibitors

12. What is a class? How do you define a class or a class effect? Along what characteristics do you divide up the agents? Are there any exceptions? BRAVO

13. What we have learned Even with extreme caution, it is very difficult to stop a trial in time to prevent increased mortality. Phase II trials can be very misleading, and give strong misimpressions. BRAVO

14. “I don't think treating diabetes is not a realm for cardiologists. I think obviously it's critically interrelated […] To me one of the instructive things about the development and problems of this class of drugs is that in phase III studies were done capable of detecting differences in mortality that are way below the threshold that most therapies today that are chronically given to people with a variety of diseases could even come close to picking up.” Dr Robert Califf Professor of Cardiology Associate Vice Chancellor for Clinical Research at Duke University BRAVO Diabetes

15. Class Classes are an easy point of reference for guidelines Talking about individual drugs often gets contentious Must be very careful before assigning things to a “class” and assuming they all act alike BRAVO