It has been suggested that coronary artery bypass grafting (CABG) performed in the setting of emergent failure of percutaneous transluminal coronary angioplasty causes minimal increased risk compared with routine CABG. We reviewed the records of 103 patients undergoing emergency CABG for failed percutaneous transluminal coronary angioplasty (group 1) and compared them with an identical number of consecutive CABG patients from 1987 (group 2). Group 1 had a lower risk profile evidenced by lower mean age (p less than 0.01), fewer diseased vessels (p less than 0.0001), better ventricular function (p less than 0.001), fewer left main lesions (p less than 0.0001), and fewer patients with acute ischemia requiring intravenous administration of nitroglycerin (p less than 0.01). Despite these differences, the group 1 patients had a higher mortality rate (11% versus 1%; p less than 0.01) and a higher rate of perioperative infarctions (new Q wave) (22% versus 6%; p less than 0.01). An analysis of risk factors was performed in the group 1 patients using 36 preoperative and operative variables. Multivariate analysis revealed that left ventricular score (p less than 0.0001), preoperative (after percutaneous transluminal coronary angioplasty) need for inotropic support (p less than 0.005), and age (p less than 0.025) were independent predictors of operative mortality. In conclusion, emergency CABG after failed percutaneous transluminal coronary angioplasty carries a significantly greater risk of operative death and perioperative infarction than elective CABG.