Noninvasive Coronary Flow Reserve Predicts Response to Exercise in Asymptomatic Severe Aortic Stenosis

Background: In patients with asymptomatic aortic stenosis (AS), exercise stress echocardiography (ESE) provides additional prognostic information beyond baseline. The coronary flow velocity reserve (CFVR) is impaired in AS, but its link with exertion is unknown in this setting. The aim of this study was to test the hypothesis that CFVR could predict exercise capacity and abnormal exercise test results in AS. Methods: Noninvasive CFVR and symptom-limited semisupine ESE were prospectively performed the same day in 43 patients with asymptomatic isolated severe AS (mean age, 68.5 +/- 11 years; 26% women; mean aortic valve area, 0.8 +/- 6 0.16 cm(2); mean left ventricular ejection fraction, 70 +/- 7%). CFVR was performed in the distal part of the left anterior descending coronary artery using intravenous adenosine infusion (140 mg/kg/min over 2 min), and ESE was performed at an initial workload of 25 W with a 20-to 25-W increase at 2-min intervals. An abnormal result on ESE was defined as onset of symptoms at <75% of maximum predicted workload, electrocardiographic ST-segment depression >= 2 mm during exercise, increase of systolic blood pressure <20 mm Hg or decrease in blood pressure, and complex ventricular arrhythmia. Seventeen patients with isolated severe asymptomatic AS, unable to exercise because of extracardiac conditions, served as a comparative group. Results: Resting, hyperemic left anterior descending coronary artery flow velocity and CFVR (2.45 +/- 6 0.8 vs 2.46 +/- 0.8) were similar between the group unable to perform ESE and the ESE group (P = NS for all). Compared with patients with normal results on ESE, those with abnormal results on ESE (n = 22) were older, had higher E/e' ratios, had higher resting left anterior descending coronary artery flow velocities (39 +/- 12 vs 31 +/- 8 cm/sec), and had lower CFVR (2.01 +/- 6 0.3 vs 2.85 +/- 6 0.7; P <.01 for all). Furthermore, CFVR was significantly correlated with age, changes in transvalvular pressure gradient and left ventricular ejection fraction with exercise, workload (in watts), and exercise duration (P<.05 for all). After adjusting for other variables, CFVR remained independently correlated with exercise duration, workload, and abnormal results on ESE (P <.01 for all). On receiver operating characteristic curve analysis, CFVR <2.3 was the best cutoff to predict abnormal results on ESE (area under the curve = 0.88 +/- 6 0.06, P <.01). Conclusions: In patients with asymptomatic severe AS, noninvasive CFVR is correlated with exercise duration and workload, and low CFVR predicts abnormal results on ESE with good accuracy.