BACKGROUND: Multidetector computed tomography is useful for determining the appropriate transcatheter heart valve (THV) size in patients with severe aortic stenosis who are suboptimal surgical candidates. The relationship between adherence to the recommended CoreValve sizing algorithm and clinical outcomes is not known. METHODS AND RESULTS: We evaluated 1023 patients with severe aortic stenosis deemed high or extreme risk for surgery treated with the CoreValve THV. All patients underwent preprocedural multidetector computed tomography, and the scans were reviewed at a central analysis center using standardized software. Compliance to a recommended sizing algorithm was used to identify patients with below-range, in-range, and above-range THV sizing. A device annular sizing ratio (DAR) was also calculated based on the native annulus perimeter and perimeter of the selected THV. Clinical end points included the presence of paravalvular aortic regurgitation evaluated by an independent echocardiographic laboratory. Adherence to the sizing algorithm was highest with a 31-mm THV (92.6%) and lowest with the 23-mm THV (38.5%). Below-range sizing was associated with a higher rate of moderate or severe paravalvular aortic regurgitation (15.3%) than in-range (6.5%) or above-range (10.0%; P<0.001) sizing. Higher DARs were associated with lower rates of moderate or severe paravalvular aortic regurgitation: DAR ≤10%, 17.6%; DAR 10% to 15%, 9.9%; DAR 15% to 20%, 6.3%; and DAR >20%, 4.9%; P<0.001. There was no increase in clinical events associated with higher DARs. CONCLUSIONS: Adherence to a sizing algorithm guided by multidetector computed tomography resulted in lower rates of paravalvular aortic regurgitation after self-expanding transcatheter valve replacement without an increase in complications. CLINICAL TRIAL REGISTRATION: URL: http://www.clinicaltrials.gov. Unique identifier: NCT01240902.