OBJECTIVES: TRITON-TIMI 38 demonstrated reduced thrombotic events associated with increased bleeding in patients taking prasugrel compared with clopidogrel.  Treatment compliance has previously been reported for the overall trial population but not for the lower bleeding risk (LBR) group of patients characterized by age < 75 years, weight ≥ 60 kg, and no prior TIA/stroke.  The objective of this analysis was to compare persistence, discontinuation rates, and reasons for discontinuation among LBR patients receiving prasugrel versus clopidogrel. METHODS:    A total of 10,727 of 13,608 patients from TRITON-TIMI 38 were identified to be in the LBR population (prasugrel: N=5390, clopidogrel: N=5337).  Patients were followed for up to 15 months.  Persistence was measured as the time from randomization to the first gap of >14 days in which the patient was not known to be taking study drug and comparison made between treatment arms using a Cox proportional hazards model and controlling for demographics and medical history. RESULTS:   Similar persistence levels were observed for prasugrel-treated patients and clopidogrel-treated patients (327 vs. 329 days, p=0.856).  Sensitivity analyses using 7-day and 30-day gaps confirmed this finding.  Among the studied population, 15.8% prasugrel patients and 15.7% clopidogrel patients (p=0.923) prematurely discontinued their study medication.  Among patients who permanently discontinued their medication, the most common reason recorded was “patient decision” (52.4% for prasugrel, 54.0% for clopidogrel, p=0.517).  Discontinuation due to hemorrhagic adverse events was more common among prasugrel than clopidogrel patients (11.3% vs. 6.7%, respectively, p=0.001); discontinuation due to non-hemorrhagic adverse events was not significantly different between drugs (25.6% vs. 28.3%, respectively, p=0.219). CONCLUSIONS:   Overall, similar persistence and discontinuation rates were observed for prasugrel and clopidogrel for this LBR patient population.   In both groups, patient decision was more likely to contribute to discontinuation than were adverse events.