BACKGROUND AND AIMS OF THE STUDY: The study aim was to assess the value of exercise stress testing in identifying asymptomatic patients with moderate or severe valvular aortic stenosis (AS). These patients generally develop symptoms during follow up, and require valvular replacement surgery (VRS) at one to three years after single symptom-limited exercise stress testing. Limited data are available on predictors of outcome in asymptomatic patients with valvular AS. A single symptom-limited exercise stress test might offer more precise risk stratification of patients referred for cardiological evaluation. METHODS: The safety and diagnostic accuracy of exercise testing to predict symptom development and need for surgery was assessed prospectively in 30 asymptomatic patients (mean age 62+/-14 years) with valvular AS. Twenty patients had moderate AS (mean Doppler gradient 30-49 mmHg), and 10 severe AS (gradient > or =50 mmHg). Patients underwent a symptom-limited maximal exercise test with upright bicycle ergometry. There were no complications during and after exercise testing. All patients were followed up for at least 36 months. RESULTS: Ergometry was abnormal in 18 patients (60%); two patients had a fall in systolic blood pressure, one patient had a fall in systolic blood pressure with angina and ECG signs of myocardial ischemia, one had angina and ECG signs of myocardial ischemia, three patients had ECG signs of myocardial ischemia without symptoms, and 11 had dyspnea at low workload. During the following 12 months all patients with a normal exercise test remained asymptomatic (negative predictive value 100%). Ten of the 18 patients with abnormal exercise test experienced symptoms and required VRS, but eight did not (positive predictive value 55%). After 36 months, only two of 12 patients with a normal exercise test developed symptoms and required VRS (negative predictive value 83%); among subjects with abnormal exercise test, four of 18 required VRS. At three years after exercise testing, 10 patients with a normal exercise stress test and four with an abnormal test did not require VRS as they remained asymptomatic (positive predictive value 78%). There was no statistically significant difference in valvular aortic area, maximal and mean gradient between patients with normal and abnormal exercise tests. CONCLUSION: Exercise stress testing may be performed safely in asymptomatic patients with moderate or severe valvular AS. Tests which meet criteria for normal patients allow physicians confidently to postpone VRS and to suggest a simple, cost-effective method of follow up in such cases. An abnormal test may reveal symptoms or identify a population for closer follow up.