Fig. 9.1
Hemodynamic assessment of a patient (a) with and (b) without PVL. (a) AR index patient A = (37 − 18)/111 × 100 = 17. (b) AR index patient B = (42 − 10)/105 × 100 = 30
Fig. 9.2
Overview of different hemodynamic indices with their definition, cutoff values, and specificity. N.A. not available
ARI ratio correlates ARI before and after transcatheter valve implantation. The ARI ratio with a cutoff <0.60 improved the specificity for the prediction of more than mild PVL and 1-year mortality from 75.1% to 93.2% and from 75.0% to 93.3%, respectively (Fig. 9.2) [13].
The diastolic pressure-time (DPT) index is calculated by measuring the area between the aortic and left ventricular pressure-time curves during diastole and divided by the duration of diastole. DPT index is adjusted for the SBP (DPT indexadj = (DPT index/SBP) × 100) [14].
DPT indexadj decreases with significant PVL (grade ≥2), and a value ≤27.9 seems associated with 1-year mortality (hazard ratio 2.5, 95% confidence interval; 1.3–6.4); p < 0.001) (Fig. 9.2) [14].
9.2.4 Echocardiographic Assessment
The Valve Academic Research Consortium-2 (VARC-2) recommends Doppler echocardiography for the quantitative and semiquantitative assessment of PVL [15]. Color Doppler echocardiography can distinguish between trans- and paravalvular leakage; for the evaluation of PVL, Color Doppler should be performed just below the valve stent, whereas for the evaluation of transvalvular leakage, it should be performed at the coaptation point of the leaflets [15]. All imaging windows should be assessed in order to ensure complete visualization of PVL; however the parasternal short-axis view is critical in assessing the number and severity of paravalvular jets [15].
Transesophageal echocardiography (TEE) may improve PVL assessment in patients in whom poor images are obtained by transthoracic echocardiography (TTE); however TEE is more invasive.
Current trends to perform TAVI under local anesthesia or (mild) conscious sedation limit TEE feasibility. Furthermore TTE assessment in the cath lab is challenging because the patient is in the supine position (no left lateral decubitus). In addition, TTE may mask PVL jets located posteriorly, whereas TEE may mask jets located anteriorly.
9.2.4.1 Limitations
Most echocardiographic parameters (Fig. 9.3) used for the assessment of PVL are based on surgical heart valves and are not validated in transcatheter heart valves. In addition, several studies suggest that echocardiography underestimates the severity of PVL when compared to cardiac magnetic resonance (CMR) [16, 17].
Fig. 9.3
VARC-2 echocardiography criteria for the quantification of PVL adapted from Kappetein et al. PW pulsed wave, EROA effective regurgitation orifice area
Recently, Geleijnse et al. showed that the parasternal short-axis analysis of the circumferential extent of PVL, which is recommended by the VARC-2 and is considered critical in assessing PVL, was false negative in 14% of cases. This may imply underestimation of PVL in prior studies relying on circumferential PVL extent [18].
9.2.5 Cardiac Magnetic Resonance (CMR)
Cardiac magnetic resonance is a noninvasive imaging modality allowing accurate and reproducible quantification of aortic regurgitation (AR) by using phase-contrast velocity mapping technique [16, 17]. A phase-contrast view in a short-axis plane just above the THV is obtained for quantification of the forward and reversed flow volumes (Fig. 9.4) [16]. The regurgitation fraction (RF), which is defined as the diastolic reversed flow volume/systolic forward volume × 100, can be used as a parameter for the stratification of the severity of PVL. None/trivial corresponds with a RF of <8%; mild corresponds with a RF of 9–20%; moderate corresponds with a RF of 21–39%; severe corresponds with a RF of >40% [16, 17].
Fig. 9.4
Example of aortic regurgitation quantification with cardiac magnetic resonance (CMR) by using phase-contrast velocity technique. (a and b) Coronal and three-chamber views (white line represents the level of flow measurement and asterisk (*) the valve in aortic position), (c and d) Phase-contrast velocity and anatomic images, (e) Graphic of flow measurement showing a regurgitation fraction of 66%. Image courtesy of Raluca Chelu, MD, Department of Radiology, Erasmus Medical Center
9.2.5.1 Limitations
Since CMR is not available in the catheterization room, intra-procedural assessment of PVL is not possible and thereby not contributing in the decision-making whether to perform additional corrective maneuvers. In addition, CMR does not differentiate between transvalvular and paravalvular leakage. The cutoff values of the RF, used for the stratification of PVL, are not validated.
Also, TAVI-induced conduction abnormalities may require a permanent pacemaker or implantable cardioverter defibrillator (ICD) which is at least a relative contraindication for CMR (even for the MR-compatible devices).
9.2.6 Biomarkers
Recently van Belle et al. demonstrated that changes in von Willebrand factor during TAVI can predict the presence of PVL [19]. Defects in von Willebrand factor high-molecular-weight (HMW) multimers occur in patients with PVL, through turbulent blood flow caused by paravalvular leakage. The HMW multimer conformation changes lead to proteolytic cleavage [19]. This may shorten HMW multimers that are less hemostatically competent and cause a prolongation of the closure time with adenosine diphosphate (CT-ADP).
CT-ADP decreased in patients with no regurgitation post-TAVI from 235 ± 62 (baseline) to 129 ± 54 s (end of procedure), while in patients with persistent AR, CT-ADP remained high throughout the procedure. In the corrected regurgitation group (i.e., post-balloon dilatation or second valve), the CT-ADP did not change markedly from 250 ± 53 (baseline) to 223 ± 49 s (after valve implantation) but decreased after the corrective procedure to 124 ± 59 s. These findings were also confirmed in a validation cohort: The CT-ADP at the end of the procedure was significantly higher in patients with aortic regurgitation than in those without regurgitation (244 ± 64 s vs. 118 ± 53 s, p < 0.001 [19].
9.3 Determinants of PVL
9.3.1 Patient-Related Factors
Native Aortic Valve Calcification
In contrast to surgical aortic valve replacement, the calcified native aortic valve is not excised with TAVI. In fact, valvular calcification is needed to ensure anchoring of the THV. We previously demonstrated that patients with valve dislodgement had significantly less aortic root calcification (Agatston score median 1951 AU (IQR, 799–3103) vs. 3289 AU (IQR 2097–4481), p = 0.016) with an Agatston score <2359 AU as a single independent predictor for valve dislodgement (OR 3.10, 1.09–8.84 [20]. However, excessive calcification of the aortic annulus (Fig. 9.5) might lead to frame under expansion and incomplete circumferential apposition (of the THV) to the native annulus [21–23]. Amount and distribution of annular calcification are a predictor for PVL [24–27]. A study on [27] 112 consecutively treated patients confirmed a significant association between the aortic valve calcium score (AVCS) and PVL [odds ratio (OR; per AVCS of 1000), 11.38; 95% confidence interval (CI) 2.33–55.53; p = 0.001)]. The mean AVCS in patients without PVL (n = 66) was 2704 ± 151, 3804 ± 2739 (p = 0.05) in mild PVL (n = 31), and 7387 ± 1044 (p = 0.002) with PVL (n = 4). An increase of the Agatston calcium score with 100 HU is associated with increased risk for PVL (odds ratio 1.09; 95% confidence interval 1.01–1.17; p = 0.029) [25].
Bicuspid Aortic Valve
Bicuspid aortic valve (BAV) phenotype (Fig. 9.6) is the most common congenital valvular abnormality, occurring in 0.5–2% of the general population [28], and is associated with accelerated valve degeneration. BAV has so far been an exclusion criterion in randomized TAVI trials, so limited data about TAVI in BAV is available [1, 2]. TAVI in BAV may suffer from uneven frame expansion and subpar function, including PVL [29]. A systematic review on TAVI in BAV reported a 31% incidence of ≥ moderate PVL [30]. The rate of at least moderate PVL post TAVI seems consistently higher with BAV vs. tricuspid aortic stenosis (25% vs. 15%, p = 0.05) [31]. BAV tends to have a higher degree of root calcification (Agatston score 1262.7 ± 396.0 vs. 556.4 ± 461.9, p < 0.01) [32]. The self-expandable Medtronic CoreValve seems more underexpanded in BAV than in degenerated tricuspid aortic valves (underexpansion at base of the stent frame in 81.7% ± 14.9% vs. 94.7% ± 15.0%, p = 0.06; at annulus level, 74.3% ± 16.7% vs. 89.9% ± 10.5%, p = 0.03; at leaflet level, 64.6% ± 13.1% vs. 81.2% ± 13.2%, p < 0.01) [32].
Fig. 9.5
MSCT image of a severely calcified tricuspid aortic valve. NC noncoronary cusp, RC right coronary cusp, LC left coronary cusp
Fig. 9.6
MSCT image of a calcified bicuspid aortic valve type I L-R, with fusion of the left and right coronary cusp. NC noncoronary cusp, RC right coronary cusp, LC left coronary cusp
9.3.2 Procedural Factors
Valve Type
Several meta-analyses suggest that the frequency of PVL is higher with self-expandable valves (SEV) than with the balloon-expandable valves (BEV) [7, 33]. In the randomized Comparison of Transcatheter Heart Valves in High-Risk Patients With Severe Aortic Stenosis: Medtronic CoreValve Versus Edwards SAPIEN XT (CHOICE) trial, PVL assessed by contrast aortography and TTE was more frequent with Medtronic CoreValve SEV as compared to SAPIEN XT [34]. The nitinol SEV frame has lower radial force than the stainless steel BAV frame [35] which may explain a more ellipsoid and underexpanded frame configuration with SEV by rotational angiography and a higher incidence of ≥ moderate PVL [36].
Patient Prosthesis Mismatch
Sizing for TAVI relies on a detailed aortic root assessment by noninvasive imaging techniques. Oversizing relative to the native annulus may provoke conduction abnormalities or more rarely annulus rupture and coronary obstruction, whereas undersizing may increase the risk for valve embolization and PVL.
Three-dimensional, volume-rendered multi-sliced computed tomography (MSCT) is currently “the gold standard” for aortic annulus measurement and device sizing. Echocardiography typically underestimates annular dimensions and may thus predispose to valve undersizing and PVL [37, 38]. Indeed MSCT-guided annular sizing reduced the incidence of >mild PVL when compared with two-dimensional TEE-guided annular sizing (7.5% vs. 21.9%; p = 0.045) [38].
Prosthesis Malpositioning
Appropriate positioning of THV is essential. Various THVs have a sealing mechanism (i.e., skirt) (Fig. 9.7), located at the lower part of the frame, to minimize retrograde blood flow into the LV. However, in too deep implantations (too ventricular) (Fig. 9.8a), the sealing fabric ends up below the native annulus. In case of a too high (aortic) implantation (Fig. 9.8b), the THV may not cover the native annulus.
Fig. 9.7
Example of a sealing mechanism at the inflow portion of the frame of the transcatheter heart valve
Fig. 9.8
Angiographic view of (a) a too deep (too ventricular) and (b) a too high (too aortic) implantation of a transcatheter aortic heart valve. Yellow line: native aortic annulus
9.3.3 Post-procedural Factor
9.3.3.1 Prosthetic Valve Endocarditis
Prosthetic valve endocarditis (PVE) is diagnosed according to the modified Duke criteria [39]. PVE is a rare but serious complication after TAVI, with an incidence varying in the literature from 0.6% to 3.4% [1, 40, 41]. A large multicenter registry reported a 1.13% PVE incidence [42]. PVE may damage the leaflets and/or framework and extend into paravalvular tissue causing AR (transvalvular and/or paravalvular). A multicenter study reported new or worsening AR in 15.1% of TAVI patients with PVE [43].
9.4 Treatment
9.4.1 Balloon Postdilatation
Balloon postdilatation may (partly) correct frame underexpansion (Fig. 9.9). Balloon postdilatation can improve frame expansion and reduce PVL in the majority of patients with ≥ moderate PVL [44].
