BACKGROUND: Intravenous recombinant tissue plasminogen activator (rtPA) therapy has been established by randomized clinical trials for administration within 4.5 hours. Both the MERCI Retrieval Device and the Penumbra Aspiration System are options for endovascular mechanical embolectomy based on single-arm studies of treatment initiated within 8 hours. Both devices receive Centers for Medicare and Medicaid Services (CMS) reimbursement at a higher rate than IV rtPA. In this context, we sought to determine if embolectomy use is increasing over time, and further characterize this group. METHODS: We reviewed the Premier Perspective Database for ischemic stroke-related hospitalizations (ICD-9 codes 433.x1, 434.x1 and 436) of patients of 18+ years during 2008, 2009 and 2010 fiscal years. This database includes all payor sources, and additional demographic and drug utilization data, from a sampling of 15% of US hospitals. We identified 187,764 (92%) from such hospitalizations representing first admissions. Use of reperfusion therapies was determined through the thrombolysis procedure code (99.10), pharmacy records (use of 50 or 100 mg Alteplase vials), hospital billing information (notation of the Penumbra or MERCI device), and/or the mechanical thrombectomy procedure code (39.74). We also compared characteristics of those treated with mechanical embolectomy versus thrombolysis alone. RESULTS: An increasing proportion of ischemic strokes were treated with reperfusion therapies each year: 4.1%( 2482/61096) in 2008, 4.7% (2983/63093) in 2009, and 5.4% (3451/63575) in 2010 (p<0.001). Embolectomy use almost doubled (191%), and also increased as a proportion of reperfusion treatments by 37%, over the last three years: 11.1% (275/2482) in 2008, 11.9% (354/2983) in 2009, and 15.2% (524/3451) in 2010 (p<0.001). Overall, those who received embolectomy, as compared to rtPA alone, were younger (66.7 vs 69.5 mean years; p<0.001), with higher mortality rates (28% vs 11%; p<0.001), at larger hospitals (43% vs 37% >500 beds; p<0.001), at teaching hospitals (68% vs 49%; p<0.001), at urban hospitals, (95.5% vs 93.2%;p=0.003), on Medicare (63% vs 57%; p<0.001), and of Caucasian race (72.45 vs 67.4%;p<0.001). Multimodal imaging use, based on CT-perfusion on patient bills, was more common among those treated with embolectomy than rtPA alone (4.0 vs 1.8% in 2008, p=0.021; 17.8 vs 5.6% in 2009, p<0.001; 24.2 vs 6.7% in 2010, p<0.001). CONCLUSIONS: Mechanical embolectomy use has almost doubled in the US over the last three years, and is associated with a substantially higher rate of CT-based multimodal imaging use. As healthcare resources are increasingly being used to treat patients with mechanical embolectomy, it becomes imperative that the effectiveness of mechanical embolectomy and associated CT-based multimodal imaging selection be determined in definitive clinical trials.