OBJECTIVE Emboli generated during cardiac surgery have been associated with aortic clamping and manipulation. Proximal anastomotic devices are thought to be less traumatic by eliminating partial clamping, potentially resulting in fewer adverse outcomes. Intra-aortic filtration has been shown to effectively capture particulate debris. We compared the amount of debris released using intra-aortic filtration and the clinical outcomes between conventionally handsewn and automated proximal anastomoses. METHODS Seventy-seven patients undergoing primary coronary artery bypass grafting with cardiopulmonary bypass were enrolled in a prospective randomized study. Patients were assigned to the anastomotic device Group I (Symmetry Aortic Connector, n = 39) or the conventional handsewn anastomosis control Group II (n = 38). Proximal anastomoses were performed before cardiopulmonary bypass in both groups. Intra-aortic Filter 1 (EMBOL-X)was deployed prior to partial clamping or puncturing the aorta for device application and removed after the proximal anastomosis was completed. Prior to cross-clamp removal, a second filter was inserted (Filter 2). A core laboratory performed quantitative and histologic analyses of the debris captured. Clinical outcomes included adverse events, neurocognitive test scores, graft patency, and mortality. RESULTS Preoperative variables and risk factors were not significantly different between Groups I and II (EuroSCORE 3.9+/-2.6 vs. 4.2+/-2.5). Filter analyses showed no significant difference between Groups I and II in Filter 1 or 2 for either surface area of particles or total number of particles (P> 0.005). There was a significant decrease between Filters 1 and 2 in both Groups for surface area of particles (Group I: 18.5+/-23.8 mm2 vs. 10.7+/-16.3 mm2, P = 0.017; Group II: 15.0+/-15.4 mm2 vs. 6.9+/-.6.5 mm2, P = 0.004), and for total number of particles in Group II (8.6+/-3.7 vs. 7.1+/-2.4, P = 0.023). No significant differences were observed between Group I (device) and Group II (control) outcomes for myocardial infarction, neurocognitive deficit, stroke, length of stay, graft occlusion, or mortality. CONCLUSIONS The application of proximal aortic connectors without partial clamping does not reduce particulate emboli or affect clinical outcomes compared with conventional anastomoses. Cross-clamping during cardiopulmonary bypass produces less particulate debris than conventional or automated proximal anastomoses performed off-pump, suggesting a major source of emboli is the anastomotic process.