Real-world observations with prasugrel compared to clopidogrel in acute coronary syndrome patients treated with percutaneous coronary intervention in the United States

Objectives: To compare 30 and 90 day real-world acute myocardial infarction (AMI) and bleeding related rehospitalization rates in acute coronary syndrome (ACS) patients receiving percutaneous coronary intervention (PCI; ACS-PCI) treated with clopidogrel or prasugrel. Research design and methods: Using the Premier hospital database, ACS-PCI patients receiving a drug-eluting (DES) or bare-metal (BMS) stent and clopidogrel or prasugrel from July 2009 to June 2011 were analyzed. Patients were included based on the prasugrel US prescribing information (USPI), excluding patients with a history of transient ischemic attack/stroke and patients >= 75 years without diabetes or prior MI. The primary endpoint was 30 day adjusted AMI rehospitalization rate. Secondary endpoints included 90 day AMI and 30 and 90 day bleeding-related rehospitalization rates. Treatment comparisons were adjusted using propensity score stratification. Results: At the index event, prasugrel patients (N = 9404) differed from clopidogrel patients (N = 74,163) by having a lower risk of comorbid conditions associated with bleeding, being more likely younger and male, having ST-elevation MI and receiving a DES. For clopidogrel and prasugrel, respectively, the observed AMI-related rehospitalization rates were 4.7% and 3.9% at 30 days (p<0.0001) and 6.3% and 5.1% at 90 days (p<0.0001). After adjustment, prasugrel was associated with similar to 10% lower odds of AMI-related rehospitalization (30 days: OR = 0.892 [95% CI: 0.798, 0.998]; 90 days, OR = 0.901 [95% CI: 0.817, 0.994]). Adjusted bleeding-related rehospitalization rates were similar to each other (OR = 1.035 at 30 days [95% CI: 0.765, 1.399]; OR = 0.922 at 90 days [95% CI: 0.725, 1.172]). Study limitations: Treatment adherence was not assessed. Bleeding events not resulting in a hospitalization (e. g. office, outpatient, or emergency room visits), deaths outside the hospital, or readmissions to a hospital outside of the Premier alliance were not captured in the database. Conclusions: The different patient characteristics between prasugrel- and clopidogrel-treated patients suggest physicians are more selective in choosing patients for prasugrel than recommended in the prasugrel USPI. However, after adjustment for these differences, 30 and 90 day AMI rehospitalization rates were lower for prasugrel-treated patients compared to clopidogrel-treated patients, with no difference in adjusted bleeding-related rehospitalization rates.