Abstract
Background
Aortic stenosis severity is defined by the calculated aortic valve area (AVA) using the Doppler-derived continuity equation (CE; AVACE = 0.785 x (LVOTd)2 x LVOT VTI/AV VTI (LVOTd = left ventricular outflow tract diameter; AV = aortic valve; VTI = velocity-time integral)). The LVOTd is the “Achilles heel” due to limitations of conventional 2D imaging and unstandardized measurement sites (0.5–1cm “apical” in the LVOT (LVOTdApi) vs “annulus” level (LVOTdAnn)) with no consensus even in recent guidelines. Geometrical differences in the LVOT and annulus of bicuspid (BAV) and tricuspid (TAV) aortic valves are further confounders. There remains a paucity of evidence to guide best practice. Erroneous LVOTd values will result in inaccurate AVACE estimations, triggering inappropriate intervention, or inaction.
Purpose
To define the optimal LVOTd by transthoracic echocardiography (TTE; LVOTdApi vs LVOTdAnn by 2D vs 3D) for AVACE calculation, and to identify factors accounting for any observed differences between BAVs and TAVs. AVA measured by 3D TTE planimetry (AVA3D) was the reference standard.
Methods
TTEs with 3D datasets of the AV complex in patients with BAVs and TAVs were included. 2D-LVOTdApi and 2D-LVOTdAnn were measured from 2D parasternal long-axis images. 3D-LVOTd values were calculated from the measured circumference, area, and average of minor/major dimensions at the “apical” (3D-LVOTdApi-Circ; 3D-LVOTdApi-Area; 3D-LVOTdApi-Ave) and “annulus” (3D-LVOTdAnn-Circ; 3D-LVOTdAnn-Area; 3D-LVOTdAnn-Ave) levels using multiplanar reconstruction. LVOTVTI and AVVTI were traced from standard spectral Doppler waveforms. The ratio of minor:major dimensions was used as an eccentricity index (EI) of the “apical” (EIApi) and “annulus” (EIAnn) sites.
Results
53 BAVs and 52 TAVs were included. Mean BAV-AVA3D and TAV-AVA3D were 3.48±0.93cm2 and 3.41±0.69cm2 respectively. In BAVs, estimated AVACE using 3D-LVOTdApi-Circ (3.43±1.05cm2; Intraclass correlation (ICC) 0.971) and 3D-LVOTdAnn-Circ (3.40±1.01cm2; ICC 0.968) correlated best with AVA3D. Conversely in TAVs, 3D-LVOTdApi-Area (3.29±0.69cm2; ICC 0.983) and 3D-LVOTdAnn-Area (3.24±0.70cm2; ICC 0.975) performed optimally. 3D-LVOTdApi-Circ and 3D-LVOTdAnn-Circ overestimated TAV-AVA3D due to the greater ellipticity of the “apical” and “annulus” sites in TAVs (TAV vs BAV: EIApi 0.78 vs 0.83, P=0.04; EIAnn 0.84 vs 0.90, P<0.005). 2D-LVOTd-derived AVACE were predictable underestimations of the AVA3D but the “annulus” level outperformed “apical” in both BAVs and TAVs (BAV-ICC 0.857 vs 0.771; TAV-ICC 0.917 vs 0.889).
Conclusions
The “apical” LVOT and “annulus” are more circular in BAVs where 3D circumference (3D-LVOTdApi-Circ or 3D-LVOTdAnn-Circ) yielded the best AVACE. The more elliptical geometry in TAVs resulted in 3D area (3D-LVOTdApi-Area or 3D-LVOTdAnn-Area) the preferred technique. Resolving the quandary of measurement site, 2D-LVOTdAnn is favored over 2D-LVOTdApi for both BAVs and TAVs.
Modified continuity equation using left ventricular outflow tract three-dimensional imaging for aortic valve area estimation
