Coronary artery bypass grafting is the preferred therapy for left main trunk coronary stenoses because of a demonstrated survival advantage compared with medical therapy. Recently, however, advocates of percutaneous intervention have suggested that stenting or atherectomy may provide acceptable results for these patients. To determine the spectrum of surgical outcome among patients with left main trunk coronary stenoses, we analyzed 14 potential covariates of in-hospital and 3-year mortality for a consecutive series of patients without prior coronary artery bypass grafting treated at our institution from January 1990 to December 1995. In-hospital mortality (2.3%) for 1585 consecutively treated patients correlated independently with renal dysfunction, age, and class III to IV heart failure. A model was constructed that allowed prediction of outcome (c-statistic = 0.77, p < 0.0001), and only the 9% of patients at highest risk were found to be at risk for death >5%. Mortality at 3 years (15.6% ± 2.2%) in a previously specified subgroup (n = 285) correlated independently with baseline age, renal dysfunction, and chronic pulmonary disease. A model was constructed that allowed prediction of outcome for the validation sample (four risk groups with 3-year mortality rates of 4.5% ± 2.5%, 6.5% ± 2.8%, 20.0% ± 4.2%, and 39.8% ± 8.5%; p < 0.0001. These data provide a contemporary benchmark of clinical outcome against which possible competing therapies may be compared. (Am Heart J 1998;135:335-8.)