Relations between preoperative data and status at one- and 10-year control and left ventricular performance (radionuclide angiocardiography) at follow-up 10-17 years (mean 12 years) after valve replacement (1965-73) for aortic stenosis were studied. Marked improvements at one-year were noted for New York Heart Association (NYHA) functional classes, cardiothorac index (CTI), and ECG hypertrophy (Romhilt-Estes) score. One-year status did not influence that at 10 years. In 59 patients who were in NYHA class I at one year, class I was maintained in 75% of those who were in class II preoperatively, compared with 47% of those who had been in class III, and 23% of those who had been in class IV, (P less than 0.05). The same relation to preoperative level was found for patients with CTI less than or equal to 0.50 at one-year control. The ECG score did not change after the first year. Left ventricular hypertrophy (ECG score greater than or equal to 4) was present in 90% and 84% of the patients at one and 10 years, respectively. Ejection fraction (EF) and peak filling rate (PFR) at follow-up were (mean +/- SD) 65 +/- 13% and 3.14 +/- 1.05 end-diastolic volume s-1, respectively; only 13 patients had a radionuclide study within the normal range for all parameters measured. A normal study was found predominantly in patients with complete regression of hypertrophy. Multiple linear regression models incorporating preoperative data allowed calculation of predicted EF (65 +/- 9) and PFR (3.17 +/- 0.92) which did not differ from the measured values. Functional status and left ventricular systolic and diastolic performance late after the operation depended closely on preoperative status. Improvements noted at one year were only temporary in patients with preoperative advance disease. Normal left ventricular performance late after valve replacement was related to complete reversibility of hypertrophy. Our results argue for operation in minimally symptomatic patients.