Recent developments in coronary artery bypass graft surgery (CABG) without cardiopulmonary bypass made the sole use of high thoracic epidural anesthesia (TEA) in conscious patients feasible. Previously, TEA has been reported only for single-vessel CABG via lateral thoracotomy. We investigated the feasibility and complications of sole TEA in 20 patients undergoing beating-heart arterial revascularization via partial lower sternotomy for single-vessel disease (minimally invasive direct coronary artery bypass grafting [MIDCAB] technique; n = 10) or complete median sternotomy for multivessel disease (off-pump coronary artery bypass grafting [OPCAB] technique; n = 10). An epidural catheter was inserted at the T1-2 or T2-3 interspace. An epidural infusion of ropivacaine 0.5% and sufentanil 1.66 micro g/mL was started to establish anesthetic levels at C5-6 for OPCAB and at T1-2 for MIDCAB. Nine OPCAB and eight MIDCAB procedures were completed while patients were awake and spontaneously breathing during the entire procedure. Because of surgical pneumothorax (OPCAB), insufficient anesthesia, or phrenic nerve palsy (both MIDCAB), three patients required intraoperative conversion to general anesthesia. The heart rate decreased significantly (P < 0.05) by 10%-15% in both groups during the procedure. Compared with baseline (B), mean arterial blood pressure (mm Hg) was decreased significantly only during coronary anastomosis (CA) (B(OPCAB), 95 +/- 11; CA(OPCAB), 68 +/- 9; B(MIDCAB), 86 +/- 10; CA(MIDCAB), 73 +/- 10; P not significant between groups). PaCO(2) increased from 42 +/- 2 mm Hg to 46 +/- 7 mm Hg (P < 0.05) throughout the perioperative course during OPCAB, whereas it remained almost unaltered during MIDCAB procedures. All patients rated TEA as "good" or "excellent." In conclusion, we demonstrated that the sole use of TEA for MIDCAB and OPCAB procedures was feasible and provided a high degree of patient satisfaction in our small and highly selected cohorts. IMPLICATIONS. The sole use of high thoracic epidural anesthesia was studied in 20 patients who underwent beating-heart coronary artery bypass grafting using either median or partial lower sternotomy while awake.