A higher rate of clinical events in poor clopidogrel and/or aspirin responders was documented by using different methods to measure platelet function, but no conclusive data about the appropriate methodology to explore platelet reactivity are available. A total of 746 patients included in the cohort of the RECLOSE trial who had successful drug-eluting stent implantation were assessed for post-treatment residual platelet reactivity (RPR) in platelet-rich plasma by 10 microM adenosine 5'-diphosphate (ADP), 1 mM arachidonic acid (AA) and 2 microg/ml collagen-induced platelet aggregation and in whole blood by the PFA-100 system. At six-month follow-up, RPR by two stimuli (ADP and AA or ADP and collagen) and by three stimuli (ADP, AA and collagen) is significantly associated with higher percentage of primary (definite or probable stent thrombosis) and secondary (cardiac mortality and stent thrombosis) end-points than RPR by ADP, AA, collagen and PFA-100 system. According to the primary and secondary end points, the specificity values for RPR identified by two (ADP and AA:94%; ADP and collagen:97%) and three stimuli were higher with respect to RPR by ADP (88%), or RPR by AA (83%) or RPR by collagen (90%). The positive likelihood ratio values of RPR by three stimuli (9.55) or of RPR by ADP and collagen (8.08) were higher than those of RPR by ADP (2.59), by AA (2.05), by collagen (4.73), or by PFA-100 (2.63). This prospective study documents that the evaluation of platelet reactivity addressed to identify patients at risk of thrombotic events on dual antiplatelet treatment has to be carried out by methods able to explore different pathways.