Previous reports of emergency coronary artery bypass grafting often included cues that were not true surgical emergencies, thereby creating inappropriately favorable results. To accurately investigate this important subgroup of patients, we analyzed our recent experience with truly emergent coronary artery bypass grafting. From January 1984 to January 1989, 117 patients underwent true emergency bypass grafting for acute refractory coronary artery ischemia. Clinical deterioration was associated with failure of percutaneous angioplasty in 37 patients and instability during diagnostic catheterization in 13 patients. Refractory ischemia developed in the remaining patients while on the ward or in the intensive care unit. All operations were performed within four hours of surgical consultation, most within one hour. Overall in-hospital operative mortality was 14.5% ( 17 / 117 ), and 76.5% of deaths ( 13 / 17 ) were due to cardiac-related causes. Major morbidity occurred in 35.9% ( 42 / 117 ). Univariate analysis isolated ejection fraction, extent of coronary artery disease, previous myocardial infarction, hypertension, need for inotropic support, use of an intraaortic balloon pump, and cardiopulmonary resuscitation as risk factors for operative mortality. Stepwise multivariate analysis confirmed that previous myocardial infarction, hypertension, cardiopulmonary resuscitation, and reoperation were independently significant risk factors. Age, sex, diabetes, left main disease, and peripheral vascular disease had no significant impact on the prognosis. The 4% operative mortality ( 2 / 50 ) for patients taken directly to the operating room from the catheterization suite was significantly lower than the 22.4% mortaiity ( 15 / 67 ) associated with emergerncies arising on the ward or intensive care unit ( p < 0.01). A logistic risk equation developed from this population accurately modeled operative mortality. All 9 patients predicted to have a greater than 90% risk of operative death died, whereas only 1 (1.5%) of the 65 patients predicted to have a less than 5% risk died. We conclude that true emergency coronary artery bypass grafting carries a particularly high risk, with an operative mortality in the range of 15%. This finding merits full consideration when making therapeutic decisions concerning this group of critically ill patients.