Lessons learned from a clinical trial.

Remarkable advances in cardiovascular care have been substantially mediated by large-scale, randomized clinical trials. These trials not only have identified treatments resulting in major improvements in patient outcomes but also have enhanced our understanding of the natural history of contemporary disease and the impact of risk factors and complications. The process by which phase III clinical trials in cardiovascular medicine are created, implemented, completed, analyzed, presented, and published has evolved dramatically over the past decade. Historically, government and academic alliances took center stage in this process because of their intellectual equity, opinion leadership, access to patients, and allocation of public interest–related research tax dollars. This axis has now shifted. Community practitioners and groups of physicians, often organized regionally, now provide the majority of subjects for clinical trials. Furthermore, substantial scientific expertise, at both the basic and clinical levels, resides within multinational pharmaceutical firms. Additionally, contract research organizations have seized the business opportunity afforded by the need for timely and efficient operational aspects of clinical trials. Although communitybased trials, efficiency, and expertise are prerequisites for a major clinical trial, the extraordinary costs of completing them squarely places the sponsor in a dominant position.1 Our participation in the failed large-scale attempt to develop a novel oral glycoprotein IIb/IIIa inhibitor, sibrafiban, for secondary prevention of coronary heart disease has stimulated us to reflect on issues that arose in the design and conduct of this trial.2,3 Using the Sibrafiban Versus Aspirin to Yield Maximum Protection From ischemic Heart Events Post-Acute Coronary Syndromes (SYMPHONY) and 2nd SYMPHONY trials as examples, we review issues here that should be of interest to investigators, clinical practitioners, participating institutions, sponsors, data and safety monitoring boards (DSMBs), and the patients whom we serve.4 Our purpose is to encourage discussion so that standards can be enhanced for this form of collaborative science. The SYMPHONY Experience The oral platelet glycoprotein IIb/IIIa inhibitor sibrafiban met every criteria then known to qualify as a prototypical inhibitor of platelet aggregation.5 It seemed reasonable to assume that an agent with a more potent and predictable ability to inhibit platelet aggregation would have a more beneficial effect than aspirin on reducing cardiovascular events.6,7 This assumption was further buttressed by our generally favorable experience in the development of intravenous glycoprotein IIb/IIIa inhibitors for acute coronary syndromes.8,9 Extending this benefit with orally administered, longer-acting preparations was an obvious broadened paradigm. Phase I and II studies demonstrated that doses accounting for creatinine clearance could reliably produce serum sibrafiban levels within the range dictated by intestinal absorption with predictable inhibition of platelet aggregation and a doseresponse relationship with bleeding sustainable in the acceptable range.5,10 Although the peak-to-trough ratio of 2 necessitated twice-daily dosing, this seemed acceptable, especially given the belief that a reversible platelet inhibitor would be beneficial. Hence, we undertook a large-scale phase III international trial of 9233 patients from 670 sites in 33 countries to define the effects of sibrafiban on clinical outcomes.11 After extensive discussions, the sponsor and the international Steering Committee initiated the protocol designed to evaluate the efficacy and safety of sibrafiban. To explore potential interactions between aspirin and sibrafiban, the Steering Committee preferred a 3-arm design of combination aspirin and sibrafiban, aspirin alone, and sibrafiban alone. However, the sponsor was concerned that an aspirinsibrafiban combination might demonstrate safety concerns that would be unfavorably linked to its novel compound and was hopeful that 1 of the sibrafiban doses would demonstrate clear superiority. Thus, the sponsor mandated that trial funding would proceed only if there were 2 experimental arms, ie, lowand high-dose sibrafiban (neither containing aspirin) and aspirin alone. Because the initial trial, 1st SYMPHONY, did not contain a combined aspirin-sibrafiban