To determine the prognostic value of hemodynamic variables at rest and during exercise, 49 patients with chronic congestive heart failure undergoing hemodynamic evaluation at rest and during symptom-limited exercise were followed for 1 year. One-year mortality rate was 33%. On univariate analysis, nonsurvivors differed significantly from survivors in pulmonary arterial wedge pressure at rest (22 +/- 10 vs 15 +/- 10 mm Hg; p = 0.01) and during exercise (32 +/- 9 vs 24 +/- 9 mm Hg; p = 0.003), stroke work index at rest (19 +/- 6 vs 25 +/- 9 g-m/m2; p = 0.03) and during exercise (20 +/- 7 vs 32 +/- 14 g-m/m2; p = 0.001) and exercise-induced increment in stroke work index (0.5 +/- 0.4 vs 7 +/- 8 g-m/m2; p = 0.004), but not with respect to left ventricular ejection fraction, exercise duration, peak oxygen consumption or peak left ventricular hydraulic power. Patients with a peak exercise stroke work index less than 20 g-m/m2 had a 66% mortality rate compared with a mortality rate of 13% in patients with a peak exercise stroke work index greater than 20 g-m/m2 (p = 0.0001). Multiple logistic regression analysis identified pulmonary arterial wedge pressure at rest and peak exercise stroke work index as the only independent predictors of mortality. A receiver-operating characteristic curve analysis revealed that peak exercise stroke work index provided significant incremental prognostic information over the resting hemodynamic variables.(ABSTRACT TRUNCATED AT 250 WORDS)