Methods: To analyze the long-term outcome of dynamic cardiomyoplasty, we retrospectively studied 127 consecutive patients who underwent this procedure in Paris, France ( n = 76), São Paulo, Brazil ( n = 37), and Portland, Oregon ( n = 14). Preoperative data were collected for patients operated on between January 1985 and June 1994 and examined with respect to effect on long-term survival. Patients had a mean age of 50 ± 13 years and were predominantly male (82%). In 46% the cause of disease was ischemic. Concomitant operations were performed in 22 patients. Results: Operative mortality was 12% (15/127). Kaplan-Meier survival ± standard error at 1 through 5 years was 73% ± 4%, 57% ± 5%, 49% ± 6%, 44% ± 6%, and 40% ± 7%, respectively. There was a distinct improvement at 6 months in New York Heart Association functional class (3.2 ± 0.05 vs 1.7 ± 0.07, p < 0.0001) and a small but significant increase in left ventricular ejection fraction (20% ± 0.8% vs 23% ± 1.5%, p = 0.04). Ninety-day mortality was associated with low right ventricular ejection fraction, a blunted hemodynamic response to exercise testing, and requirement for an intraaortic balloon pump at the time of the operation. Using a stepwise Cox regression method of multivariable survival analysis ( n = 101), we determined that atrial fibrillation, New York Heart Association class IV, high pulmonary capillary wedge pressure, and balloon pump use were independent variables simultaneously associated with poor overall survival. When metabolic testing variables were added to this model, peak oxygen consumption eliminated both pulmonary capillary wedge pressure and functional class from the model, albeit with fewer ( n = 74) patients. Conclusion: Dynamic cardiomyoplasty is an evolving therapy for symptomatic congestive heart failure, the results of which may be enhanced by intelligent, risk-sensitive patient selection. (J T horac C ardiovasc S urg 1996;112:1640-50)