Fourteen patients with critical aortic stenosis (valve area 0.4 cm2/m2), a history of advanced congestive heart failure, left ventricular ejection fraction less than 0.45 (mean 0.28 d 0.03) and no other valvular lesions or obstructive coronary artery disease were studied to assess prognosis with aortic valve replacement. Eleven of 14 (79%) survived surgery; 10 of these 11 showed major clinical improvement postoperatively and form group 1. The three patients who died and the patient who did not improve form group 2. Although group 2 had higher preoperative values for aortic valve area and left ventricular end-diastolic volume and lower ejection fraction and cardiac output than group 1, none of these factors alone reliably predicted outcome. The mean systolic gradient was an important predictor ofoutcome: Nopatient with a mean systolic gradient 30 mm Hg had a good outcome, irrespective of valve area or other hemodynamic variables. Ejection fraction was plotted against left ventricular wall stress for both groups. For group 1, there was a close linear relation that could be extrapolated back to normal wall stress and normal ejection fraction. This suggested afterload mismatch as a major cause for this group's depressed ejection fraction. In group 2 ejection fraction was lower for any given wall stress, suggesting depressed contractility, rather than afterload mismatch, as the cause of the left ventricular dysfunction. Thus, either afterload mismatch or depressed contrac-tility may result in depressed ejection fraction in patients with aortic stenosis; which one predominates may have major prognostic importance.

THE NATURAL HISTORY of critical aortic stenosis is well defined. 1 2 Patients with this disease who have symptoms of congestive heart failure and re-main untreatedby aortic valve replacement have an average life expectancy of approximately 2 years. Although recent studies3-5 have shown that the ma-jority of patients with critical stenosis and congestive heart failure respond well to corrective surgery, some patients with this condition are not helped by aortic valve replacement. Patients with aortic stenosis and congestive heart failure who responded well and those who responded poorly to aortic valve replacement may represent two distinct groups, rather than op-posite ends of a spectrum. To test this hypothesis and to identify the defining characteristics of these two groups, we examined our experience at the Peter Bent Brigham Hospital in patients with pure aortic stenosis and severe congestive failure who underwent cardiac catheterization and aortic valvereplacement over a 5-year period.