Little is known about the association of echocardiographic estimates of right ventricular (RV) function with survival, in relation to hemodynamic and exercise-derived predictors of outcome in congestive heart failure. We prospectively studied 40 patients (age 55+/-10 years, in New York Heart Association functional class III [70%] and IV [30%]), with left ventricular (LV) ejection fraction <30%. At enrollment, all patients underwent echocardiographic evaluation of LV dimensions and function. RV shortening was measured as the difference of the end-diastolic distance - the end-systolic distance between the tricuspid annulus and the RV apex. Thirty-five patients (88%) were able to perform a maximal symptom-limited exercise test. Peak oxygen consumption (peak VO2) and percent peak age- and gender-adjusted predicted oxygen consumption (%peak VO2) were calculated. Of 40 patients, 10 died during a mean follow-up period of 14+/-10 months. On univariate analysis, nonsurvivors had lower RV shortening (p=0.0001), higher pulmonary artery wedge pressure (p=0.009), higher pulmonary vascular resistance (p=0.02), and lower mean aortic pressure (p=0.05). Cox proportional-hazards model revealed that the only independent associate of survival was RV shortening (p=0.0005), with a trend toward significance for mean aortic pressure (p=0.08). The best cutoff point of RV shortening identified by the receiver-operating curve was 1.25 cm. This value had a sensitivity of 90%, specificity of 80%, and overall predictive accuracy of 83% to distinguish survivors from nonsurvivors. In patients with advanced heart failure, preserved RV function as indicated by an echocardiographically derived RV shortening > 1.25 cm is a strong predictor of survival.