Echocardiography has been debated as an adjunct for transcatheter aortic valve replacement (TAVR). The aim of this prospective study was to comparatively evaluate intraprocedural guidance using intracardiac echocardiography (ICE) and transesophageal echocardiography (TEE).
Fifty high-risk patients with severe aortic stenosis scheduled for TAVR were randomized to either guidance using ICE (group 1; n = 25) or monitoring using TEE (group 2; n = 25).
In contrast to TEE, ICE allowed continuous monitoring. The need for probe repositioning during the procedure was much lower in group 1 (0.1 ± 0.3 vs 5.7 ± 0.7 maneuvers, P < .001). Compared with TEE, the transcaval intracardiac echocardiographic view provided higher coaxiality with the ascending aorta expressed as the length of ascending aorta depicted (4.9 ± 1.2 vs 6.1 ± 1.2 cm, P = .003). Both coronary ostia were more frequently visualized in group 1 (18 vs 2 cases, P < .001). ICE-derived annular measurements were correlated closely with preinterventional readings on TEE ( n = 25, r 2 = 0.90, P < .001). TEE underestimated intraprocedural pressure gradients in comparison with preinterventional measurements (mean difference, −10.2 ± 11.1 mm Hg; n = 11, P = .012), but ICE did not (mean difference, −0.3 ± 14.1 mm Hg; n = 25, P = .913). ICE and TEE detected newly grown thrombi (2 vs 1 case). Severe complications (e.g., annular dissection, pericardial effusion) were not observed.
ICE, which is compatible with sedation and local anesthesia, can be considered an alternative to TEE for intraprocedural guidance during TAVR. It also seems to match the required work flow during TAVR better than TEE.
Although conventional heart surgery remains the treatment of choice in patients with symptomatic aortic stenosis, transcatheter aortic valve replacement (TAVR) has been demonstrated to be a viable alternative in a select high-risk population and, when successful, results in marked hemodynamic and clinical improvement. However, the technique has been limited by technical constraints and procedural risks. Transesophageal echocardiography (TEE) is not only essential for preprocedural sizing of the aortic annulus and for evaluation of valvular morphology and function but also greatly facilitates the detection of intraprocedural complications such as thrombus formation, the development of severe aortic regurgitation after valvuloplasty or pericardial effusion suggestive of annular dissection and myocardial perforation, interference of the prosthesis with anterior mitral leaflet motion and consecutive development or worsening of mitral regurgitation, new wall motion abnormalities due to compression of a coronary ostium, and aortic dissection or intramural hematoma formation. Although TEE has emerged as a vital adjunct to fluoroscopy, its capabilities for guiding the entire procedure have been found to be not optimal. This also applies to the recently introduced guidance of TAVR using real-time three-dimensional TEE. Because TAVR alone can now be performed under only sedation and local anesthesia, TEE is often omitted. In contrast, intracardiac echocardiography (ICE) does not require general anesthesia, will most likely not interfere with fluoroscopy, and can provide simultaneous guidance. In case studies and preclinical tests, ICE was shown to be an effective alternative guidance method for TAVR.
Hypothesizing that both imaging techniques have advantages and limitations, ICE and TEE were compared with respect to the intraprocedural diagnostic information particularly important for safety and technical success immediately before device deployment, during implantation, and thereafter. Another goal of this investigation was to evaluate how well the methods fit into the procedural work flow of TAVR.
Because they were very poor candidates for standard surgical valve replacement, a total of 53 consecutive patients with symptomatic aortic stenosis and aortic valve areas < 0.8 cm 2 were to be treated either by transapical antegrade or transfemoral retrograde TAVR in accordance with the joint recommendations of the European Association of Cardiothoracic Surgery and the European Society of Cardiology. Three patients were excluded because transapical access ( n = 1) or transfemoral access ( n = 2) could not be established as intended. Thus, the final study population comprised 50 subjects in whom the Edwards Sapien (Edwards Lifesciences, Irvine, CA) transcatheter heart valve was deployed with the Ascendra delivery system, the RetroFlex II catheter, or the NovaFlex catheter (all Edwards Lifesciences). The interventional procedure was carried out under general anesthesia with endotracheal intubation in all patients. Before the procedure, consecutive patients scheduled for TAVR were randomized by alternately assigning them to either guidance using ICE (group 1; n = 25) or monitoring using TEE (group 2; n = 25). Splitting the population into two groups was necessary because interactions make simultaneous ICE and TEE in the same patient infeasible. All patients provided written informed consent for both intervention and follow-up. Data collection was prospective and was approved by the ethics committee of Innsbruck Medical University (Innsbruck, Austria).
Echocardiography Before TAVR
The diagnosis of severe aortic stenosis was based on standard parameters, including ejection fraction measured by transthoracic echocardiography. On the basis of transthoracic five-chamber views, and in some cases the suprasternal notch or right parasternal view, mean instantaneous pressure gradients were calculated from velocities measured in the left ventricular outflow tract and the aorta using the modified Bernoulli equation. The aortic valve effective orifice area was calculated using the continuity equation. In all patients, preinterventional TEE served to detect cases of bicuspid aortic valve and to measure the maximum posterior to anterior aortic valve annular diameter in a zoomed early systolic midesophageal long-axis view at 110° to 150° rotation at the level of leaflet insertion. The size of the prosthesis was in all cases based on transesophageal echocardiographic measurements. An annular diameter of 18 to 21 mm was considered appropriate for a 23-mm prosthesis, and a diameter of 22 to 25 mm was considered suitable for a 26-mm prosthesis.
Computed Tomography and Invasive Diagnostics
Computed tomography was routinely performed to calculate the x-ray system detector angulation. Optimal fluoroscopic visualization was achieved by aligning the beam parallel to the annular plane (i.e., by getting the cusps to superimpose on the image). This view facilitates precise alignment of the prosthetic valve with the native annulus. Computed tomography was also used to evaluate the morphology of the entire aorta and the aortic valve and to measure the diameters of the iliac and femoral arteries. A vessel diameter of 8 mm was considered adequate for a 22Fr transfemoral sheath to accommodate the 23-mm diameter prosthetic valve and a diameter of 9 mm sufficient for a 24Fr sheath to accommodate the 26-mm diameter prosthetic valve (RetroFlex II catheter). For the recently available NovaFlex catheter, a vessel diameter of ≥7 mm was considered sufficient. Short segments of noncalcified focal stenoses were not considered exclusion criteria for transfemoral access. All individuals underwent left-heart and right-heart catheterization to measure standard hemodynamic parameters and instantaneous transvalvular pressure gradients.
For ICE performed, with an Acuson X300 PE ultrasound unit (Siemens Medical Solutions, Erlangen, Germany), an 8Fr AcuNav catheter (Siemens Medical Solutions) was introduced into the femoral vein and advanced through the inferior vena cava to the entrance of the superior vena cava into the right atrium. The highest frequency available (10 MHz) was used. Turning the intracardiac echocardiographic catheter counterclockwise, the transducer was aimed at the neighboring ascending aorta and was locked with a slight anterior tilt to obtain the longitudinal transcaval/transatrial view ( Figure 1 A) for continuously displaying the sinotubular junction, aortic bulb, aortic valve, and left ventricular outflow tract. Transcaval and transatrial longitudinal intracardiac echocardiographic views were used to measure the posterior-to-anterior annular diameter, as in preinterventional TEE, the maximum length of the ascending aorta depicted together with the aortic valve in one view as a measure of coaxiality between cut plane and ascending aorta, and transvalvular pressure gradients before intervention, after predilatation and after deployment of the prosthetic valve. Spectral Doppler analyses were performed without angle correction. All data were averaged from three measurements. The ICE operator was blinded to the preinterventional transesophageal echocardiographic measurement of annular size.
Longitudinal views were also used for qualitative evaluation of the native valve and aortic morphology, for implantation of the valve prosthesis, and for confirming the patency of the coronary ostia thereafter. The longitudinal view from the entrance of the superior vena cava toward the right atrium was used as the primary intraprocedural standard view on ICE. This view was helpful during guidewire passage through the native valve, balloon predilatation, final adjustment, and deployment of the valve prosthesis during TAVR. It also served to detect possible complications and to evaluate prosthetic valve function, especially to depict any paravalvular or transvalvular leaks. The severity of any regurgitation was qualitatively evaluated with a multiparametric approach based on color and continuous-wave Doppler, as recommended by the American Society of Echocardiography. Any repositioning and minor readjustments of the intracardiac echocardiographic catheter required for maintaining an adequate longitudinal view were recorded. At the end of the interventional procedure, the tip of the catheter was manipulated into the tricuspid valve orifice at the level of the annulus to obtain the short-axis standard view of the aortic valve to rule out aortic annular dissection and to again check for transvalvular and paravalvular leaks ( Figure 1 B). Alternatively, a short axis could be adjusted from the right atrium with the catheter tip tilted anteriorly. Thereafter, the intracardiac echocardiographic catheter was advanced into the right ventricle for a transventricular long-axis view of the left ventricle to assess the presence of pericardial effusion and to visually estimate left ventricular ejection fraction as a final check after prosthetic valve deployment. Manipulation of the catheter was performed by the interventional operator, who is also an experienced echocardiographic examiner meeting the level III criteria of the American Society of Echocardiography. The ultrasonographic unit was operated by an experienced ultrasonographic technician.
TEE was performed by another level III examiner using the same ultrasound unit used for ICE at a frequency of 7.0 MHz. Midesophageal long-axis views at 110° to 150° rotation served as the main intraprocedural views during TAVR and were used to facilitate wire passage through the native valve and for comparative measurements of annular size, length of the section of ascending aorta depicted together with the aortic valve to estimate coaxiality, and for visual estimation of left ventricular ejection fraction as well as for qualitative assessment of valvular morphology and function. All data were averaged as described for ICE. Transesophageal short-axis views were used to confirm proper prosthetic valve opening after deployment and to depict the coronary ostia. To obtain transvalvular pressure gradients before and after predilatation as well as after prosthetic valve deployment, deep transgastric or transgastric longitudinal views of the aortic valve were used whenever feasible, and the Doppler beam was aligned parallel to the left ventricular outflow tract. TEE was also used to exclude possible complications. As described for ICE, any repositioning and minor readjustments of the probe were recorded.
Patient characteristics are summarized as frequencies and percentages or as mean ± SD. Qualitative data were analyzed using χ 2 tests with one degree of freedom. Differences between frequencies were tested using the binomial test. The Kolmogorov-Smirnov test was used to assess normality. Differences between two unpaired groups were evaluated using t tests or Mann-Whitney U tests and for more than two unpaired groups using analysis of variance or Kruskal-Wallis tests. If necessary, a subinvestigation was done with a post hoc analysis (Bonferroni) or Mann-Whitney U test. Differences between more than two paired groups were calculated using the Friedmann test and, if necessary, further subinvestigations using Wilcoxon’s tests. Results are also displayed as Bland-Altman graphs. For annular diameters comparatively measured in the same patients (group 1), the paired sign test was used to assess the significance of the difference in observer variability between ICE and TEE. All reported P values were two sided, and a type I error level of 5% was used. The P value was adjusted with Bonferroni’s correction. Calculations were performed using SPSS version 17 for Windows (SPSS, Inc., Chicago, IL) and StatView version 5 for Macintosh (SAS Institute Inc., Cary, NC).
Data from all patients were included in the variability study. To determine intraobserver variability, one observer analyzed and measured intracardiac echocardiographic data, and another observer measured transesophageal echocardiographic data. All measurements were repeated after a 4-week interval. Observers were blinded to the results of the first measurements. To determine the interobserver variability of both methods, a third observer independently analyzed and measured intracardiac echocardiographic data and a forth one transesophageal echocardiographic data. From these four observers, the second ICE and the second TEE observer were each blinded to the results of the first observer. Analyses and measurements were performed offline on the same images. Intraobserver and interobserver variability were determined for aortic annular diameter, maximum length of the ascending aorta depicted together with the aortic valve, and transvalvular pressure gradients. Results are expressed as percentile differences (100% × absolute difference between the measurements/mean of the measurements).
Baseline characteristics of the whole patient population as well as the two groups are summarized in Table 1 . Overall, the patients represented an extremely high-risk population with respect both to their expected natural courses and for standard surgical valve replacement. Groups 1 and 2 did not differ with respect to age, gender, body mass index, New York Heart Association functional class, hemodynamics, effective aortic valve area, left ventricular ejection fraction before TAVR, logistic European System for Cardiac Operative Risk Evaluation score, size of implanted valve prostheses, and access course.
|Variable||Overall ( n = 50)||Group 1 ( n = 25)||Group 2 ( n = 25)||P (groups)|
|Age (y)||81.3 ± 6.1||79.9 ± 6.6||82.8 ± 5.4||.14|
|Men||18 (36%)||9 (36%)||9 (36%)||NA|
|BMI (kg/m 2 )||25.2 ± 4.3||24.5 ± 4.3||26.0 ± 4.3||.28|
|NYHA functional class||3.18 ± 0.59||3.25 ± 0.55||3.10 ± 0.64||.43|
|Mean pressure gradient (mm Hg) ∗||48.1 ± 12.4||45.9 ± 10.2||50.4 ± 14.2||.25|
|EAVOA (cm 2 ) †||0.63 ± 0.12||0.66 ± 0.11||0.60 ± 0.13||.14|
|LVEF (%) ∗||46.0 ± 15.7||48.9 ± 16.9||43.1 ± 14.3||.25|
|EuroSCORE||30.8 ± 15.5||29.7 ± 16.9||31.9 ± 14.3||.66|
|26-mm prosthesis implanted||24 (48%)||12 (48%)||12 (48%)||NA|
|Transfemoral implantation||19 (38%)||10 (40%)||9 (36%)||.79|
Procedural Outcome and Follow-Up After TAVR
Procedural success was achieved in all patients. One (2%) procedure-related embolic stroke was observed after transfemoral TAVR. One (2%) iliac artery dissection occurred after transfemoral TAVR and was successfully treated by stent implantation. Overall, there were three deaths (6%) occurring within 30 days (days 4, 17, and 24) after transapical prosthetic valve deployment because of respiratory, renal, and multiple-organ failure. No deaths after transfemoral TAVR have been observed. Second-degree atrioventricular block did not occur, complete atrioventricular block requiring permanent pacemaker implantation occurred in one patient (2%), and new left bundle branch block without need for a permanent pacemaker was seen in three patients (6%). Neither intraprocedural ST-segment elevation myocardial infarction nor cases of endocarditis were observed.
Comparative Echocardiographic Monitoring
No complications related to ICE or TEE occurred. Calculation of detector angulation revealed positions between 5° right anterior oblique, 15° left anterior oblique, 10° cranial, and 10° caudal to be optimal. So as not to interfere with optimal fluoroscopic viewing, transesophageal echocardiographic monitoring had to be interrupted for preinterventional and postinterventional angiography of the ascending aorta, one to three balloon dilatations of the native valve, for final positioning and if necessary repositioning of the balloon catheter carrying the valve prosthesis immediately before valve deployment, for definitive implantation ( Figure 2 ), and if necessary for postdilatation of the implanted prosthetic valve. Each withdrawal and repositioning of the transesophageal probe was explicitly requested by the interventional operator. A few additional minor readjustments were needed. In group 1, minor readjustments were required to the same extent, but almost no repositioning of the intracardiac echocardiographic catheter ( Table 2 ). Consequently, in group 1, continuous ICE permitted simultaneous echocardiographic and fluoroscopic viewing throughout the TAVR procedure, including assessment of posterior-to-anterior aortic valve annular diameter ( Figure 3 ), guidewire crossing and predilatation of the stenotic valve, final positioning of the valve prosthesis, and subsequent implantation. The functional result was obvious ( Figure 4 ) immediately after balloon evacuation and withdrawal (e.g., the effectiveness of predilatation was clearly evident by significant regression of the pressure gradient [ Figures 5 and 6 ], and prosthetic valve opening and paravalvular or transvalvular leakages were easily depicted [ Table 2 ]). With TEE (group 2), in some instances, the stent containing the valve and the balloon on which the valve is mounted were hard to differentiate because of the lower image resolution of TEE compared with ICE because of the lower frequency of the transesophageal probe. On longitudinal intracardiac echocardiographic images, coaxiality between the aorta and the valve balloon system was found to be better than on the standard transesophageal view. The maximum length of the ascending aorta depicted by ICE in group 1 together with the aortic valve exceeded the length obtainable by TEE in group 2 ( Table 2 ). In group 2, Doppler quantitation of transvalvular flow velocities was less often feasible in group 2 than in group 1, because obtaining a stable transducer position for deep transgastric or transgastric longitudinal transesophageal views with adequate alignment of the Doppler beam toward the left ventricular outflow was challenging and often took several minutes. Transvalvular Doppler quantitation could be carried out in all group 1 patients but was possible in only 11 group 2 subjects (44%). TEE underestimated transvalvular pressure gradients in comparison with ICE ( Figure 6 ). Preinterventional echocardiography-derived transvalvular pressure gradients were confirmed in group 1 by intracardiac echocardiographic measurements before predilatation (mean difference, −0.3 ± 14.1 mm Hg; n = 25, P = .913; Table 1 , Figure 6 ). In contrast, intraprocedural TEE before predilatation revealed lower pressure gradients than preinterventional transthoracic echocardiography (mean difference, −10.2 ± 11.1 mm Hg; n = 11, P = .012). Wire crossing of the native valve could be directly observed and assisted by either ICE or TEE in 12 patients (48%) in group 1 and 7 patients (28%) in group 2.
|Parameter||Group 1 (ICE) ( n = 25)||Group 2 (TEE) ( n = 25)||P|
|Repositioning of intracardiac echocardiographic catheter or transesophageal echocardiographic probe ∗||0.1 ± 0.3||5.7 ± 0.7||<.001|
|Minor readjustments of intracardiac echocardiographic catheter or transesophageal echocardiographic probe||3.7 ± 1.2||3.1 ± 0.8||.077|
|Severity of paravalvular leakage †||1.3 ± 0.8||1.2 ± 0.9||.853|
|Severity of transvalvular leakage †||0.2 ± 0.4||0.1 ± 0.2||.328|
|Depicted length of ascending aorta (cm) ‡||6.1 ± 1.2||4.9 ± 1.2||.003|
|Visibility of coronary ostia|
|RCA ostium visible||23 (92%)||3 (12%)||<.001|
|LCA ostium visible||19 (76%)||18 (72%)||.736|
|LCA and RCA ostia visible||18 (72%)||1 (4%)||<.001|