Conformational Pulsatile Changes of the Aortic Annulus Impact on Prosthesis Sizing by Computed Tomography for Transcatheter Aortic Valve Replacement

Objectives This study sought to investigate pulsatile changes of the aortic annulus and their impact on prosthesis selection by computed tomography (CT). Background Precise noninvasive prosthesis sizing is a prerequisite for transcatheter aortic valve replacement. Methods A total of 110 patients with severe aortic stenosis (mean age: 82.9 +/- 8 years, mean aortic valve area: 0.69 +/- 0.18 cm(2)) underwent electrocardiogram-gated CT. Aortic annulus dimensions were planimetrically quantified as area-derived diameter (D-A = 2 X root (CSA/pi), where CSA is the cross-sectional area) and perimeter-derived diameter (D-P = P/pi, where P is the length of the perimeter) in 5% increments of the RR interval. Hypothetical prosthesis sizing was based on D-A and D-P (23-mm prosthesis for <22 mm; 26 mm: 22 to 25 mm; 29 mm: > 25 mm) and compared between maximum and traditional cardiac CT reconstruction phases at 35% and 75% of RR. Agreement for prosthesis selection was calculated by kappa statistics. Results D-A and D-P were increased and eccentricity was reduced during systole, with DA-MAX and DP-MAX most often observed at 20% of RR. D-P was consistently larger than D-A. Average net differences were 2.0 +/- 0.6 mm and 1.7 +/- 0.5 mm by DA-MIN versus DA-MAX and DP-MIN versus DP-MAX. Agreement for prosthesis sizing was found in 93 of 110 patients (kappa = 0.75) by DA-75% and in 80 of 110 patients (kappa = 0.53) by DA-MAX compared with DA-35%; and in 94 of 110 patients (kappa = 0.73) by DP-75% and in 93 of 110 patients (kappa = 0.73) by DP-MAX compared with DP-35%. With sizing by DA-75% or DP-75%, nominal prosthesis diameter was smaller than DA-MAX or DP-MAX in 15 and 6 patients respectively. Conclusions Aortic annulus morphology exhibits conformational pulsatile changes throughout the cardiac cycle due to deformation and stretch. These changes affect prosthesis selection. Prosthesis selection by diastolic perimeter- or area-derived dimensions harbors the risk of undersizing. (J Am Coll Cardiol Intv 2012;5:984-94) (c) 2012 by the American College of Cardiology Foundation