Abstract TP404: Applying ARISTOTLE Trial to Actual Practice

BACKGROUND: Clinical event rates may differ among patients treated in real world (RW) clinical practice compared to patients enrolled in randomized controlled trials (RCTs). OBJECTIVE: To estimate the absolute effect of apixaban compared to warfarin on stroke and major bleed rates in a managed care, nonvalvular atrial fibrillation (NVAF) population. METHODS: We selected patients with a diagnosis of NVAF during 2007-2010 from Medco health plans. Pharmacy claims were used to define periods of warfarin exposure. Strokes and major bleeds were identified using diagnosis codes on the medical claims, and their rates were calculated during warfarin exposure. Estimated real world event rates for patients treated with apixaban were calculated by applying the ARISTOTLE relative risk reductions (RRRs) to the reference RW rates observed during warfarin exposure. Absolute risk reductions and associated number needed to treat (NNT) or number needed to harm (NNH) relative to warfarin were calculated using the estimated RW event rates. RESULTS: The RW population had a mean age of 74.9 years and mean CHADS 2 score of 2.8. Stroke and major bleeding event rates among NVAF patients treated with warfarin in RW were higher compared to those enrolled in ARISTOTLE clinical trial (Stroke: 5.61 vs. 1.51 per 100 PYs; Major Bleed: 11.14 vs. 3.09 per 100 PYs). Similarly, we estimate the absolute stroke and major bleed rates observed in the RW for apixaban would also be higher compared to the RCT (Stroke: 4.43 vs. 1.19 per 100 PYs; Major Bleed: 7.68 vs. 2.13 per 100 PYs). At the same time, the use of apixaban instead of warfarin would be associated with lower NNTs in order to prevent a stroke (85 vs. 313) and a major bleed (29 vs. 105) in the RW vs. RCT. CONCLUSIONS: The numbers needed to treat to avoid a stroke and major bleeding with apixaban instead of warfarin in the real world are likely to be substantially lower than in a randomized controlled trial setting.