Introduction:
Transapical-aortic valve implantation (TA-AVI) has evolved into a routine procedure in selected elderly high-risk patients. However, more-than-mild paravalvular leaks remain a significant drawback of current TAVI systems. The aim was to study the impact of native aortic valve calcification on paravalvular leaks in cardiac contrast-enhanced computed tomography (CT).
Methods:
The degree and distribution of native valve calcification were quantified using an Aortic Valve Calcium Score (AVCS) for each cusp separately (3mensio Valves™ workstation, version 7.0, 3mensio Medical Imaging B.V., Netherlands). To exclude an artificial increase of the AVCS due to the presence of contrast material, we used a threshold for density [mean aortic density + 2*D] and volume [0, 3, 5, 25 and 50 mm
3
] of calcification. AVCS was compared to the rate of paravalvular leaks, assessed intraoperatively by echocardiography and root angiography.
Results:
Eighty-eight consecutive patients prior to TA-AVI with preoperative CT aged 80.0 ± 5.7 years, 51.1 % female were included. Three prosthesis sizes were used for annular diameters up to 23 mm (n = 29), 26 mm (n = 45) and 29 mm (n = 14). Mean log. EuroSCORE was 24.6 ± 15.4 % and mean STS-Score was 8.4 ± 8.3 %. The mean AVCS in patients without paravalvular leaks (n = 42) was 606.4 ± 374.3; with mild paravalvular leaks (n = 39) was 761.2 ± 530.4; and with moderate paravalvular leaks (n = 5) was 792.4 ± 515.3 with the highest calcification in the non-coronary cusp. There was no significant association between the total AVCS and paravalvular leaks (χ
2
-statistic = 2.9; P = 0.13, 551 hounsfield units). The additional use of the volume-based threshold did not lead to an increase of the association between the AVCS and paravalvular leakages. Paravalvular leaks were significantly associated with the degree (r
Spearman
= 0.34; χ
2
-statistic = 10.0; P = 0.02) and location of eccentric calcified plaques.
Conclusions:
Quantification of aortic valve calcification in contrast enhanced computed tomography shows only a weak correlation with paravalvular leakage and is therefore not reliable as a predictor, respectively. The degree of eccentric cusp calcification was significantly associated with the occurrence and location of paravalvular leaks.