Two-Dimensional Transesophageal Echocardiography Assessment of the Major Aortic Annulus Diameter in Patients Undergoing Transcatheter Aortic Valve Replacement.

BACKGROUND: Multidetector computed tomography (MDCT) is the gold standard in annulus sizing before transcatheter aortic valve replacement (TAVR). However, MDCT has limited applicability in specific subgroups of patients, such as those with atrial fibrillation and chronic kidney disease. Two-dimensional transesophageal echocardiography (2DTEE) has traditionally been limited to the long-axis measurement of the anteroposterior diameter of the aortic annulus. We describe a new 2DTEE approach for the measurement of the major diameter of the aortic annulus.
METHODS: Seventy-six patients with symptomatic severe aortic valve stenosis and high surgical risk underwent MDCT and 2DTEE before TAVR. A modified five-chamber view was used to measure the major aortic annulus diameter. This was obtained starting from a mid-esophageal four chamber and retracting the TEE probe up until the left ventricular outflow tract and the left and noncoronary aortic cusps were visualized: major aortic annulus diameter was measured as the distance between their insertion points in systole.
RESULTS: Major aortic annulus diameters measured at 2DTEE showed good correlation with MDCT diameter (r = 0.79; P < 0.001) and perimeter (r = 0.87; P < 0.0001). Using factsheet-derived sizing criteria, 2DTEE alone would have allowed accurate sizing in 75% of patients, with 21% of oversizing predominantly with smaller annuli.
CONCLUSIONS: We describe a new method for 2DTEE measurement of the major aortic annulus diameter; this approach is simple, correlates with MDCT, and allows adequate TAVR sizing in most patients. These findings may help in the assessment of patients with contraindications to or inadequate MDCT images. Copyright:

BACKGROUND

Degenerative aortic stenosis (AS) is the most common native valve disease and its prevalence is increasing dramatically with the aging of the population.[12] Surgical aortic valve replacement is the standard therapy for symptomatic patients with severe AS, while in the last few years, transcatheter aortic valve replacement (TAVR) has been established as a feasible therapeutic alternative for patients at high or intermediate surgical risk.[34] TAVR requires noninvasive assessment of aortic annular dimensions for determining the size of prosthesis to be implanted.

Current guidelines recommend routine use of cardiac multidetector computed tomography (MDCT) in the preoperative assessment of aortic and annular anatomy[56] since both transthoracic (TTE) and transesophageal echocardiography (TEE) underestimate the true annular diameters. Indeed, the aortic annulus is often oval shaped, allowing an easy echocardiographic measurement only of the minor annular diameter.[78] However, MDCT may also display limited accuracy and/or feasibility in some clinical scenarios (due to technical issues or comorbidities) leaving echocardiographic techniques (TTE and TEE) an important complementary role in these instances.

The aim of this study was to evaluate the applicability of a new two-dimensional (2D) TEE approach for the measurement of aortic annular major diameter, its relationship with MDCT, and its usefulness to predict prosthesis size in patients undergoing TAVR.

METHODS

Study population

This was an observational, prospective, investigator-initiated, single-center study. Between January 2011 and December 2015, 76 patients affected by symptomatic severe AS and at high predicted surgical risk (EuroScore II ≥15%)[9] were selected for TAVR at a tertiary private hospital (missing institution name). The only exclusion criterion was the inability to perform TEE due to contraindications or patient preference. All procedures performed in the study were in accordance with the ethical standards of the National Research Committee. The study was approved 221/CE4/18 by our Ethics Committee and all patients signed informed consent.

Study procedures

The most relevant clinical and demographic data were collected including age, sex, weight, height, and presence of cardiovascular risk factors. As part of preoperative evaluation, patients underwent clinical examination, 2DTTE and 2DTEE, and MDCT.

Transthoracic two-dimensional-echocardiography

A complete TTE was performed in all patients using a commercially available ultrasound system (IE-33 Matrix Philips) with a 3.5 MHz probe. All measurements were performed according to the European Society of Cardiology Recommendations for Chamber Quantification.[10] In particular, annulus diameter was measured at end-systole using zoom mode at the level of the right and noncoronary cusps from the parasternal long-axis view.

Transesophageal echocardiography

All patients included into the present study underwent 2DTEE using a commercially available ultrasound system (IE 33 Matrix Philips). The anteroposterior diameter of the aortic annulus was measured as usual, at end-systole, in a standard mid-esophageal long-axis view (three-chamber view at approximately 120°). The anteroposterior (minor) aortic annulus diameter was measured from the junction of the aortic leaflet with the septal endocardium to the junction of the leaflet with the mitral valve posteriorly, inner edge to inner edge. To measure the major diameter of the aortic annulus, we used the following nonconventional approach: Starting from a mid-esophageal four chamber at 0°, a modified five-chamber view was obtained by retracting the TEE probe up until the left ventricular outflow tract and the left and noncoronary aortic cusps were visualized [Figure 1a]. The distance between the insertion points of the two cusps, measured at end-systole, was considered the major diameter. In a subgroup of 10 patients, three-dimensional (3D) TEE reconstructions were also obtained to confirm the anatomic premises of this approach using customized section planes to reproduce our modified view in relation to other orthogonal planes [Figure 1b].[511] All data were analyzed offline by two trained physicians blinded to patient characteristics and computed tomography (CT) scan findings.

Figure 1

Modified transesophageal echocardiography five-chamber view for the measurement of major aortic annular diameter. (a) From a standard mid-esophageal 0° four-chamber view (left), the probe is pulled up until the base of the left and noncoronary cusps is visualized (right), and the distance between the insertion points of the two cusps is measured at end-systole. (b) Three-dimensional transesophageal echocardiography reconstructions showing the anatomical premises of the approach; green panel (upper left) demonstrates the modified five-chamber view described above; red panel (upper right) represents an orthogonal-plane reconstruction showing correct alignment with the true major annulus diameter; a grid on the far right serves as geometrical model of the anatomy of the aortic annulus, valve, and root

Multidetector computed tomography

Multidetector, electrocardiogram (ECG)-gated cardiac CT was performed using a commercially available scanner (GE Revolution EVO 128 slices; GE Healthcare, Waukesha, WI, USA). Patients with atrial fibrillation (AF) with a beat-to-beat variability exceeding 20 bpm during baseline breathing were excluded from CT scanning. Patients with heart rate >70 bpm were pretreated with oral β-blockers, administered approximately 1 h before the examination. Iodixanol (Visipaque®, GE), a nonionic, iso-osmolar contrast agent, was injected through an antecubital vein followed by normal saline. CT examinations were performed using retrospective ECG-gated spiral scanning with a pitch value of 0.2. Scanning was performed using an X-ray tube potential of 120 kVp and an effective tube current–time product of 850 mA. The detector collimation was 128 mm × 0.6 mm. Images were reconstructed using a single segment image reconstruction algorithm. Image sets were reconstructed using data from 10% to 90% of R-R interval with 10% increments; the best dataset to image the aortic annulus was chosen by the interpreting radiologist. A 3D workstation (GE 4.7 software version) was used to obtain images in the plane of the aortic valve and the following measures were recorded: maximum and minimum annulus diameters, annulus perimeter, sinotubular junction and ascending aorta maximum diameter, height of the sinus of Valsalva, and maximum diameter of the Valsalva's sinuses; eccentricity of the annulus was defined as the ratio between the maximum and minimum annulus diameters at MDCT. All data were analyzed offline by a trained radiologist blinded to patient characteristics and TEE features.

Transcatheter aortic valve replacement procedure

The CoreValve ReValving™ prosthesis (Medtronic Inc., Minneapolis, MN, USA) was implanted in all instances through the femoral approach according to the common technique. All procedures were carried out under local anesthesia, with additional sedation and/or analgesia, in a standard catheterization laboratory with surgical backup. Transfemoral access was obtained percutaneously or after surgical cutdown. The femoral access was closed surgically or percutaneously (Prostar XL, Abbott).

Statistical analysis

Continuous variables are reported as mean (standard deviation) or median (interquartile range) and compared by means of unpaired t-test or Mann–Whitney U-tests, when appropriate. Categorical variables are reported as raw numbers (%) and compared by means of Pearson Chi-square test, Fisher exact test, or log-rank tests, when appropriate. Correlation between variables was assessed by univariate linear regression analysis. Multivariable regression analysis was performed to identify independent predictors of final prosthesis size: variables found to be significantly correlated with prosthesis size at univariate analysis were included in a stepwise, forward multiple regression model. Agreement between TEE and CT measurements of major annular diameter was explored using Bland–Altman analysis. Inter- and intraobserver variability was explored on recorded loops, from a set of five repeated measures by two experienced physicians over a sample of 10 patients and expressed as an intraclass coefficient (ICC). All probability values were two-tailed and P < 0.05 was considered significant. Data were analyzed using SPSS for Windows, release 13.0 (SPSS Inc., Chicago, Illinois).

RESULTS

Seventy-six patients with severe AS scheduled for TAVR were included in the study. Clinical and echocardiographic characteristics are shown in Table 1. The mean age was 81.3 ± 5.3 years (range: 67–95) and 30 (39%) patients were male. Left ventricular ejection fraction was 54.4 ± 11.7%. EuroScore II was 18.2 ± 5.1%. All patients underwent 2D TTE and 2D TEE in absence of significant reported adverse reactions and in all patients minor and major aortic annulus diameters were successfully measured.

Table 1

Descriptive characteristics of the study population

n	76
Age, years	81±5
Males, n (%)	30 (39)
Height, cm	164±7
Weight, kg	71±11
EuroScore II, %	18.2±5.1
NYHA Class II/III/IV, n (%)	25/35/16 (33/46/21)
Atrial fibrillation, n (%)	34 (45)
Aortic valve area (continuity equation), cm2	0.7±0.2
Mean aortic gradient, mmHg	46±6
MDCT major diameter, mm	26.6±2.2
MDCT perimeter, mm	76.8±6.6
MDCT area, mm2	475±88
2DTEE minor diameter, mm	21.0±3.7
2DTEE major diameter, mm	26.5±4.8
Final prosthesis size, mm	27.6±2.2

NYHA=New York Heart Association, MDCT=Multidetector computed tomography, 2DTEE=Two-dimensional transesophageal echocardiography

All patients underwent transfemoral implantation of a self-expandable aortic bioprosthesis (CoreValve or CoreValve Evolut R, Medtronic Inc., Minneapolis, MN, USA). Choice of prosthesis size was based on MDCT findings and on intraprocedural considerations by the interventional cardiologist, including balloon valvuloplasty sizing if deemed appropriate.

Mean major aortic annulus diameter measured by 2DTEE was 26.5 ± 4.8 mm, mean minor diameter at 2DTEE was 21.0 ± 3.7 mm; mean major aortic annulus diameter measured through MDCT and final prosthesis size were respectively 26.6 ± 2.2 mm and 27.6 ± 2.2 mm. Figure 2 shows examples of MDCT and TEE measurement in three different patients.

Figure 2

Major diameters measured by multidetector computed tomography (left column) and two-dimensional transesophageal echocardiograph (right column); the three cases refer to patients with small (21 mm), intermediate (26 mm), and large (31 mm) CoreValve size selection

2DTEE major diameter showed good intraobserver and interobserver reproducibility with ICC of respectively 0.88 and 0.77. Figure 3 depicts the correlation between 2DTEE major annulus diameter measurement, MDCT measurements, and final prosthesis size. Very good correlation was also found between 2DTEE major diameter and MDCT perimeter (r = 0.87; P < 0.0001). Bland–Altman analysis showed a good agreement between the two methods across the whole range of aortic annulus diameters [Figure 4]. Patients with disagreeing measurements (≥2 mm difference between MDCT and 2DTEE) did not differ from those with agreeing values in terms of MDCT diameter, annulus eccentricity index, final prosthesis size, body surface area (BSA), and AF prevalence.

Figure 3

Correlation between two-dimensional transesophageal echocardiograph major diameter and multidetector computed tomography major diameter (Panel A) and final prosthesis size (Panel B)

Figure 4

Bland–Altman chart depicting agreement between multidetector computed tomography and two-dimensional transesophageal echocardiograph major annular diameters; solid line defines mean bias, dashed lines define upper and lower LOA (±1.96 standard deviation)

The accuracy of prosthesis sizing by 2DTEE was explored according to prosthesis factsheet-derived sizing charts [Figure 5]. 2DTEE alone would have allowed for accurate prediction in 75% of our patient population. Most oversizing would have happened in patients with smaller annuli (23–26 mm prostheses). However, most noncorresponding measurements between the two methods were found early in the study period, probably at the beginning of a learning curve.

Figure 5

Prosthesis size prediction by two-dimensional transesophageal echocardiograph stratified by actual final prosthesis size

2DTEE aortic diameter (r = 0.70, P < 0.001), MDCT aortic diameter (r = 0.73, P < 0.001), and BSA (r = 0.49, P < 0.01) were significantly related to the final size of the aortic prosthesis. Using a stepwise multiple linear regression analysis, including age and BSA, 2DTEE aortic diameter (beta-coefficient = 0.46, P < 0.01) confirmed to be an independent predictor of the final prosthesis size (model R2 = 0.6, P < 0.001).

DISCUSSION

TAVR has imposed as a new standard in high-risk patients with AS, and it is increasingly being used in intermediate and low-risk populations.[1213] Appropriate sizing of the implanted prosthesis is crucial to minimize procedural complications that worsen postoperative outcome (most importantly, paravalvular leaks and conduction disease).[141516]

While MDCT provides unequaled spatial resolution and reproducibility in the preoperative assessment of aortic anatomy, representing the current gold standard, researchers in other imaging modalities are striving to develop alternative approaches in this context. Despite being easily applicable in the vast majority of TAVR patients, MDCT may be limited by the need for iodinated contrast media administration, and its accuracy may decrease in patients with difficulties in breath holding and arrhythmia (such as AF). Echocardiographic imaging may complement or replace MDCT in all such cases. Conceptually, 3D TEE is the natural counterpart to MDCT, allowing for single-beat 3D reconstruction of the whole annulo-aortic region with good spatial and temporal resolution. Several studies, indeed, have demonstrated an acceptable correlation between 3DTEE and MDCT, and some degree of standardization exists on how to measure the aortic annulus with 3D.[5171819] On the other hand, 2DTEE has been routinely used for valve sizing at the beginning of the TAVR era, restricting to traditional, “minor” diameter measurements from a long-axis view. Early in the TAVR experience – and thanks to the growing knowledge of the noncircular anatomy of the aortic annulus derived from the increasing use of MDCT – it became clear that standard 2DTEE brought to frequent undersizing of the prosthesis, in particular when the aortic annulus has a very oval shape (thus with a larger discrepancy between the minimum and maximum diameters).[78] A shift away from TEE in TAVR has been observed since then, also due to issues concerning patient comfort and anesthesiologic management (especially for intraoperative monitoring). We still adopt TEE (both with 2D and 3D approaches) as a complement in the preoperative evaluation of patients with inadequate MDCT images.

Our approach stems directly from these limitations of standard TEE. While 120° three-chamber views allow the measurement of the distance between the hinge points of the right and noncoronary cusps, thus resulting in an estimate of the minor diameter of the annulus, a more posterior section obtained at 0° in a slightly more cranial position will allow visualization on the left and noncoronary cusps, thus approximating aortic annular major diameter, as confirmed by our 3D reconstructions. Unlike other 2DTEE approaches to the aortic annulus (for example, deep transgastric views), this modified five-chamber view is technically simple, universally feasible, and well reproducible.

Coherently with these premises, our measurements correlated well with MDCT data (both diameters and perimeter) and with final prosthesis size; in most cases in our series, major 2DTEE diameter alone would have proved sufficient in establishing adequate prosthesis sizing. Some of the discrepant findings between 2DTEE and MDCT may have derived from an initial learning curve with the new measurement and from the extent of calcification of the aortic annulus.

Using 2DTEE alone for sizing in our patient population would have brought to 75% of correct sizing. Unsurprisingly, most oversizing would have been observed in patients with smaller annular (and prosthesis) size, a finding inherently related to the echo technique – smaller, more circular annuli produce worse images with our approach, especially if heavily calcified. It must also be acknowledged; however, that the actual size of prosthesis adopted in our patients reflects real-life decisions including intraprocedural considerations and balloon sizing and did not rely exclusively on MDCT measurements.

To our knowledge, this is the first study aiming to standardize a new, feasible, and reliable 2DTEE approach to the measurement of major aortic annulus diameter. In this first experience on a relatively small sample, our method seemed to correlate well with CT measurements and to accurately guide TAVR prosthesis sizing in the majority of subjects. Relying on low-tech equipment and simple mid-esophageal views, this approach may complement information from other imaging modalities in patients unable to undergo adequate MDCT scanning, especially in contexts with scarce availability of newer technologies, such as 3D echocardiography.

Study limitations

This study has several limitations: first, this is a single-center study. Moreover, the small sample size may have included a relatively narrow variety of aortic annulus anatomies. The study only included patients treated with a self-expandable prosthesis. Finally, unlike MDCT and 3D echo-based approaches, the proposed 2DTEE method aims at measuring diameters but not annular area and perimeter.

CONCLUSIONS

2D TEE assessment of aortic annulus diameter using a modified five-chamber view seems to be accurate and shows good correlation with MDCT. Further studies are needed to confirm the potential usefulness of this novel approach in the assessment of patients undergoing TAVR, especially in patients with inadequate MDCT images when 3D echo is not available.

Financial support and sponsorship

The study was investigator-sponsored.

Conflicts of interest

There are no conflicts of interest.

Ethical clearance

221/CE4/18.