Predictors of calcification distribution in severe tricuspid aortic valve stenosis

We investigated aortic valve calcification (AVC) distribution and predictors for leaflet calcification patterns in patients with severe tricuspid aortic valve stenosis undergoing transcatheter aortic valve replacement (TAVR). Patients undergoing routine multi-sliced computed tomography (MSCT) for procedural planning were enrolled. MSCT data were transferred to a dedicated workstation for evaluation (3mensio Structural Heart™, Pie Medical Imaging BV, Maastricht, The Netherlands) and analyzed. Participants were separated into asymmetrical (AC) and symmetrical (SC) leaflet calcification and potential predictors for calcification distribution were identified with univariate and multivariate regression analysis. 567 Participants with severe tricuspid AS were divided into asymmetrical (AC, n = 443; 78.1%) and symmetrical (SC, n = 124; 21.9%) AVC. In AC, the non-coronary cusp was the most calcified cusp (n = 238; 57.7%). SC is more common in females (AC/SC: 49.2% vs. 67.7%; p < 0.0001). AVC was more severe in patients with AC, who also have larger aortic root dimensions. Multivariate analysis depicted, inter alia, left ventricular outflow tract (LVOT) calcification < 25 Agatston units (OR 1.81 [1.09–3.00], p = 0.021), a mean pressure gradient < 36 mmHg (OR 1.77 [1.03–3.05], p = 0.039), and an annulo-apical angle > 67° (OR 1.68 [1.00–2.80], p = 0.049) as predictors for SC, although with only moderate predictive value. Data from this retrospective analysis indicate that SC occurs more frequently in females. The cumulative leaflet calcification burden is higher in patients with AC, who also present with larger aortic root dimensions. The predictive value for prominent calcification of different aortic valve cusps in AC patients was only low to moderate. Trial registration number: NCT01805739. Supplementary Information The online version contains supplementary material available at 10.1007/s10554-021-02248-6.


Introduction
Aortic valve stenosis (AS) is the most common valvular heart disease in western countries [1]. The extent of aortic valve calcification (AVC), measured by multislice computed tomography (MSCT), correlates well with the degree of AS severity [2,3], and is an integral part of current guideline recommendations for the management of patients with valvular heart disease [4]. While sex-and severity-related distribution of aortic valve calcification are well-researched [5,6], very little is known about predictors for (a)symmetrical leaflet calcification, which can be frequently observed in daily clinical practice.
Therefore, we performed an in-depth analysis of calcium distribution patterns and favoring, underlying conditions in patients with severe, tricuspid AS.

Study population
We retrospectively enrolled 567 patients with severe tricuspid AS, who underwent routine pre-procedural planning for 1 3 transcatheter aortic valve replacement (TAVR) at the Heart Center Düsseldorf. Patients with prior aortic valve replacement and bicuspid aortic valve were excluded to guarantee comparability between groups. AS severity was defined according to the current European guideline recommendations for the management of patients with valvular heart disease [6]. Patients were separated into an asymmetrical (AC) and a symmetrical (SC) leaflet calcification pattern. In case of AC, participants were further divided by the dominant calcified cusp (non-coronary; NCC, left coronary; LCC and right coronary cusp; RCC). Asymmetrical leaflet calcification was defined as a difference of > 150 Agatston Units (AU) in-between the three leaflets, which is also visibly subsumable.
All patients provided written informed consent for the use of clinical, procedural, and follow-up data for research. The study procedures are in accordance with the Declaration of Helsinki. All necessary ethical oversight was secured: the Local Ethics Committee approved the study protocol (4080) and the study is registered at clinical trials (NCT01805739).

Statistical analysis
Continuous data are described by means with standard deviation, median with upper and lower 95% confidence interval (CI) or interquartile ranges (IQR). Categorical variables are expressed by frequencies and percentages of a whole. Sexrelated differences are analysed with the 2-sided Student's t test for continuous variables and the chi-square test (Fisher's exact test) for categorical variables. Univariate and multivariate analysis are used to identify dependent and independent predictors for asymmetric and symmetric calcification distribution. Others calcification distribution patterns were excluded due to the small sample size (NCC/LCC; NCC/ RCC; LCC/RCC) Only co-variates with a p-value below 0.1 in the univariate analysis qualified for multivariate binary logistic regression. Receiver-operating-characteristic (ROC) curves are described as c-indices (area-under-the-curve) with 95% CI. The data analysis was performed with SPSS (version 22.0, SPSS, Inc., Chicago, IL) and GraphPad Prism (version 6.0, GraphPad Software, San Diego, CA). All statistical tests were 2-sided, and a p-value < 0.05 was considered statistically significant.

MSCT imaging acquisition protocol and threedimensional (3D) reconstruction
Pre-procedural cardiac MSCT was routinely performed as native and contrast-enhanced multi-slice CT in all patients. CT data were obtained using a 128-slice, single source CTscanner with temporal resolution of 150 ms and a collimation of 128 × 0.6 mm ("SOMATOM Definition AS+", Siemens Healthcare, Forchheim, Germany). Images were taken in accordance with TAVR-related standardized recommendations for CT image acquisition [7]. They were transferred to a dedicated workstation for evaluation (3mensio Structural Heart™, Pie Medical Imaging BV, Maastricht, The Netherlands) and reconstructed in the coronal, sagittal and axial planes. After identification of the virtual annular plane, three hinge points were set on the axial plane, and 3D volume-rendered reconstruction initiated. We assessed the calcium amount of the aortic valve and the upper left ventricular outflow tract (LVOT) within 1 cm below the annulus. The calcification was re-calculated in Agatston Units (AU) for the overall-and separated leaflet calcium burden assessment. Calcium originating from extra-valvular structures, such as the mitral valve annulus, the ascending aorta, and the coronary arteries was cropped.

Asymmetric calcification with dominating NCC calcification
Univariate analysis depicted the presence of malignancy, male gender, the absence of a porcelain aorta, previous coronary artery bypass grafting, and a larger aortic anatomy as possible influencing factors for dominant NCC calcification. Multivariate analysis identified the absence of a porcelain aorta (OR 2.03 [1.07-3.86], p = 0.031) and previous coronary artery bypass grafting (OR 1.95 [1.14-3.32], p = 0.0014) as independent predictors for a pronounced NCC-calcification. However, c-statistics-even combinedremained only in a poor range ( Table 2).

Asymmetric calcification with dominating LCC calcification
Prominent LCC-calcification was linked to a porcelain aorta, a smaller aortic valve area, higher pressure gradients and pronounced LVOT calcification. Multivariate analysis offered no independent predictor for dominating LCC-calcification in patients with AC pattern.

Asymmetric calcification with dominating RCC calcification
Concerning RCC calcification univariate analysis identified, inter alia, the absence of a permanent pacemaker (PPM) at *Significant level p < 0.05 (bold, italics) Values are mean ± SD, mean ± 25th and 75th percentile or n (%)   (Table 2). A graphical illustration of the read-out is given in Fig. 1. Detailed results of uni-und multivariate regression analysis can be found in the Supplement (Table 3).

Discussion
To our knowledge, this is the first study with a systematic and in-depth analysis of AVC distribution, which includes comorbidities, hemodynamic parameters, and anatomical pre-dispositions of patients with severe tricuspid AS undergoing TAVR.
Our retrospective study revealed that: • A symmetrical calcification pattern is more frequently observed in females who have narrow aortic root dimensions. • Cumulative leaflet calcification is higher in patients with asymmetrical leaflet calcification patterns. • Independent predictors for prominent calcification of single aortic valve cusps vary widely and have only poor to moderate predictive value: • Independent predictors for symmetrical leaflet calcification are COPD, a less pronounced LVOT calcification, lower mean pressure gradients, and a more horizontal aorta. • Independent predictors for pronounced NCC calcification are previous CABG and the absence of a porcelain aorta. • Independent predictors for a prominent RCC calcification are the absence of a PPM, a less elliptical annulus index, a higher RCA/RCC ratio, and a less horizontal aorta. .

Asymmetrical vs. symmetrical AVC distribution
Koshkelashvili et al. retrospectively analyzed 318 non-contrast axial chest CT scans of subjects aged over 65 years in an all-comers cohort. They could also show that the LCC was frequently the most calcified cusp [8]. The study was not performed in patients solely presenting with AS but provided important information on race-related differences in early calcification patterns. In our study, race-differences were not considered to be relevant since over 99% of the study population were Caucasian. Overall, the number of studies concerning AVC patterns is limited since AVC is pre-dominantly examined in total and semi-quantitatively during pre-procedural planning for TAVR [9,10]. It has already been shown that AVC distribution is linked to important clinical endpoints such as the occurrence of conduction disturbances, the risk of annular rupture, coronary occlusion or paravalvular leakage [11][12][13][14]. Especially asymmetric calcification, possibly leading to increased cardial forces on the NCC and RCC and adjacent structures such as the intramembranous septum, and pronounced LCC calcification are relevant variables for a permanent pacemaker implantation post TAVR [15,16]. Current studies do not provide insight in pre-dispositioning factors, which are linked to the presented calcification pattern in patients with severe tricuspid AS. Furthermore, fibrosis is an essential factor in degenerative AS, leading to higher-grade stenosis under non-severe AVC-thresholds, especially in women. Sex-related differences in the progression and clinical phenotypes of aortic valve stenosis are well researched. High AVC is more likely in men and shows a strong correlation with the severity of AS in both sex [6].
Woman have a slower progression rate of AS but lower calcium levels are already sufficient to create the same hemodynamic impairment as in men with higher AVC load [17]. This might be explained by the fact that women are more prone to higher levels of valvular fibrosis and dense connective tissue at the same degree of hemodynamic relevant aortic stenosis severity as compared to their male counterparts [18]. Since the SC group primarily consists of women, the lower aortic valve gradients and the lower overall AVC load may be explained by the aforementioned facts.
Shear forces, aortic root entrance angles, and flow-patterns may also play an important role in AVC distribution and should be further analyzed. A horizontal aorta, the extreme form of increased aortic root angulation and the annulo-apical entrance angles, is known to limit maneuverability of self-expandable devices [19,20] but also means a potential shear force trigger, probably leading to flowdependent calcification. However, flow is pre-dominantly linked to myocardial function. Surprisingly, neither cardiac output, higher-grade reduced left ventricular ejection fraction or other hemodynamic characteristics in this context took a particular influence on calcification distribution, are displayed with independent predictors-identified by multivariate analysis. CABG coronary artery bypass grafting,COPD chronic obstructive pulmonary disease, dPmean mean pressure gradient, PPM permanent pacemaker, Ratio RCA /RCCb ratio of the right coronary artery to the base of the right coronary cusp supposing that calcification may be more significantly linked to aortic flow characteristics.

Dominant leaflet calcification in patients with AC
Permanent ventricular pacing was shown to be associated with alterations in regional myocardial and coronary perfusion [21] and may, therefore, be linked to altered calcification patterns. Coronary artery disease was no dependent or independent predictor in our analysis, but the distribution of coronary artery disease and the distance of the coronary arteries might have an impact on AVC patterns through altered sinus perfusion and connected ostial calcification areas. Effacement of the sinuses may impair coronary flow and increase the mechanical stress and thus structural degeneration of the aortic valve leaflets [22]. In bicuspid valves, deterioration of aortic blood flow or eccentric flow is well-known to increase radial pressure and shear stress on the aortic wall [23], also causing enhanced dilatation of the aortic root, closing the circle towards aortic entrance angles and a horizontal aorta in tricuspid valves. A sub-analysis of the TAVI-WIN Registry [24] could show that increased calcium volumes of the RCC were an independent predictor for new pacemaker implantation after TAVR, whereas increased calcium volume of the NCC had a protective effect. However, findings are highly contradictive throughout current literature [14,25]. Pathophysiologically, it remains unclear how different calcification patterns are created: in the initiation phase of disease, endothelial damage allows lipid infiltration and subsequent inflammation. In the propagation phase pro-fibrotic pathways and microcalcification smooth the way for further calcification [26]. Biomechanical studies could link high strain to the formation of calcific noduli and disease progression [27,28]. Therefore, variable cusp geometries and congestive differences in length may also contribute to different calcification patterns, rather then underlying comorbidities.

Limitations
This is only a single-center analysis and limited to the retrospective quality of available data, which is also reflected by c-statistics. Furthermore, several important factors, like pressure recovery and ventriculo-arterial impedance, were not analyzed. This study of pre-disposing factors for different calcification patterns lacks translational value and does not influence clinical decision making.

Future directions
This is the first study trying to determine predictive factors for different types of aortic calcification patterns. Predictive factors vary widely throughout baseline characteristics and only showed poor to moderate correlation. This study may encourage clinicians to perform quantitative calcium assessment measurements in pre-TAVR MSCTs to create further evidence in this field.

Conclusions
Data from this retrospective analysis indicate that SC occurs more frequently in female patients and narrow aortic root anatomies, whereas cumulative leaflet calcification seems to be higher in AC patterns. Independent predictors for SC are COPD, a less pronounced LVOT calcification, lower mean pressure gradients and a more horizontal aorta. However, the correlation of different baseline characteristics with certain calcification patters were only in poor to moderate range. This is the first study addressing predictors for different calcification patterns, which are known to impact clinical outcomes of patients undergoing TAVR.
provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.