The use of transcatheter aortic valve replacement (TAVR) is increasing worldwide. We present our 6-year experience using three-dimensional (3D) transesophageal echocardiography (TEE) and investigate whether different sizing methods of the aortic annulus lead to different prosthesis size that may impact outcome.
We investigated 262 patients who underwent TAVR and had 3D TEE data sets of the aortic annulus. We have used the area-derived diameter ( D area = 2 <SPAN role=presentation tabIndex=0 id=MathJax-Element-1-Frame class=MathJax style="POSITION: relative" data-mathml='(area/π)’>(area/π)−−−−−−−√(area/π)
( a r e a / π )
) and the circumference-derived diameter ( D circ = Circumference /π ) to size the prosthesis in separate populations in different time periods.
The D circ method is correlated with lower incidence of paravalvular aortic regurgitation (PVAR; odds ratio = 0.44, 95% confidence interval, 0.23-0.85; P = .015). Other factors associated with PVAR were the cover index, area-mismatch index, and circumference-mismatch index. Retrospectively, for the purposes of the study, we used the Edwards-Sapien 3 Valve 3D sizing guide in all patients, to predict the hypothetical valve size with each method. In the whole population, the calculated D circ was higher in all cases ( D circ = 23.4 ± 2.3 mm vs D area = 22.9 ± 2.3 mm; P < .001). The two methods had good agreement in predicting the valve size (kappa = 0.600). In total, 192 (73.3%) patients were assigned for the same prosthesis size, whereas 70 (26.7%) would be eligible for a different size, of which 44 (16.7%) would definitely have had a different valve implanted.
Using the aortic annulus area or circumference to calculate the annular diameter provides different values. Comparing the two methods, a different prosthesis size could have been implanted in 26.7% of patients. In our series the use of circumference-derived diameter resulted in lower incidence of PVAR. The findings of this study may be independent of the imaging modality and may therefore also apply to computed tomography-based aortic annulus measurements, but this needs to be further investigated.
As the aortic annulus becomes more ellipsoid, the area shrinks but the circumference remains the same.
Using annulus diameter derived by area vs. circumference could potentially lead to different valve size in TAVR.
In our cohort, the two methods would have resulted in a potentially different valve size in 26.7% of patients.
Using aortic annulus circumference to size the TAVR prosthesis resulted in lower prevalence of paravalvular AR.
Transcatheter aortic valve teplacement (TAVR) is a validated treatment for symptomatic severe aortic stenosis (AS) in inoperable or high-risk surgical patients. Accurate sizing of the aortic valve (AV) prosthesis is paramount to ensure acute success and a durable long-term outcome. Many procedures are performed using two-dimensional (2D) transthoracic (TTE) or transesophageal (TEE) echocardiographic measurements of the aortic annulus. These measurements assume that the shape of the aortic annulus is circular. However, since the aortic annulus is usually oval shaped, it has become apparent that making a single diameter measurement from 2D images will result in erroneous dimensions.
This has led to the use of multislice computed tomography (CT) and three-dimensional (3D) TEE imaging to size the aortic annulus, since they can both account for irregular, noncircular annular shapes. Planimetry of aortic annulus in 3D orientated images improves annulus measurements, and good agreement has been reported between multislice CT and 3D TEE. The sizing guide for the most recent commercially available balloon-expandable valves is based on 3D annular area and 3D area-derived diameter. However, as the shape of the annulus becomes more ellipsoid, the area shrinks but the circumference remains unchanged. In this study we test the hypothesis that sizing the aortic annulus using the circumference, as opposed to the area, may lead to different prosthesis size choice, especially in patients with a significantly ellipsoid shape of the aortic annulus, and we seek to investigate possible clinical implications.
A total of 294 patients were referred to our center for TAVR between September 2009 and May 2014. They presented with symptomatic, severe valvular calcific AS and were reviewed by a multidisciplinary team. A total of 11 patients were excluded from the study. Seven of them presented with a failing AV bioprosthesis and were considered for a valve in valve procedure. One had a metallic mitral valve prosthesis that did not allow adequate TEE imaging and underwent a gated, multislice CT for prosthesis sizing. Three patients developed an iatrogenic ventricular septal defect or Gerbode defect, and therefore paravalvular aortic regurgitation (PVAR) could not be accurately quantified. The implanted valves were the Edwards-Sapien XT (Edwards Lifesciences Ltd., Berkshire, UK; 179 patients), the Edwards-Sapien 3 (Edwards Lifesciences Ltd., Berkshire, UK; 83 patients), the Boston Scientific Lotus valve (Boston Scientific, Marlborough, MA; 17 patients), and the Jenavalve (JenaValve Technology Ltd., CA; 4 patients). The Lotus and Jenavalve valves are deployed with a controlled mechanical expansion as opposed to the Edwards balloon-expanded valves. This difference along with the different sizing recommendations may pose a bias to the explored outcome (PVAR); therefore the patients with a Lotus or Jenavalve prosthesis were excluded, and we only studied patients with Edwards-Sapien valves. All patients signed an informed consent for the therapeutic and diagnostic procedures, which included 3D TEE.
Transesophageal echocardiography was performed in all patients as part of the preprocedural workup to ensure suitability for the intervention, as is standard practice in our center. In addition, all patients had periprocedural TEE for guidance and monitoring of complications. The ultrasound systems iE33 and Epiq7 (Philips Healthcare, Andover, MA) with a matrix array probe (X7-2t) and the Vivid 9 (GE Healthcare, Hertfordshire, UK) with the 6VT-D transducer, to allow for 2D and 3D live images, were used. The AV and the ascending aorta were studied in short-axis (midesophageal, 40°-70°) and long-axis views (midesophageal, 110°-135°), and 2D images with and without color Doppler were obtained and stored in the hospital cardiology imaging archive (Philips, Xcelera). Using the “3D zoom” feature, 3D data sets of the AV, including the left ventricular outflow tract (LVOT) and proximal ascending aorta (sinuses and sinotubular junction), were acquired. A single-beat acquisition was used.
Echo Image Analysis
The 3D data sets were analyzed offline with the Philips Qlab (QLAB cardiac 3DQ, Philips Medical Systems, ver. 9, 10) or online on GE machines (GE Healthcare, Hertfordshire, UK). The measurements were made in a midsystolic frame. Three multiplanar reconstruction planes (MPRs) were used. Two were bisecting the long axis of the LVOT and were orthogonal to each other in the sagittal and coronal planes. The third one was transverse to the previous ones and was intersecting the AV in its short axis at the level of the annulus ( Figure 1 ). Special care was taken to position the transverse MPR at the level just below the hinge point of all three aortic cusps. One experienced operator performed a manual tracing of the aortic annulus in this transverse plane and also measured the maximum and minimum diameters ( D max and D min, respectively). This analysis methodology is analogous to that performed using CT data sets.
The AV was assessed with both 2D and 3D echo data sets, and the degree of calcification was graded as moderate or severe. There were no cases with mild valvular calcification. The AV was considered as severely calcified if all cusps demonstrated considerably increased echogenicity in more than 50% of their tip surface and there was also significant calcification affecting the commissures between cusps or there was eccentric calcification on cusps measuring more than 5 mm ( Figure 2 ). The cases with lower echogenicity and thickening of the aortic cusps were graded as having moderate calcification.
The postprocedural PVAR was graded by echocardiography (both TEE and TTE) according to recently published guidance. Using the semiquantitative method of circumferential extent of the PVAR jets, the quantitative width of jet at its origin, and the ratio of the jet width to LVOT diameter, the PVAR was classified in a five-grade scheme (mild, mild to moderate, moderate, moderate to severe, and severe).
Measurements and Calculations
The effective annulus diameter was calculated from the annular area and circumference according to the equations:
Area-derived effective diameter ( D area ) : ( D area = 2 ( a r e a / π ) )
Circumference-derived effective diameter ( D circ ) : D circ = C i r c u m f e r e n c e / π