Transesophageal Echo Leads to Wrong CoreValve Sizing in Nearly Half of Patients

Download this article's Factoid (PDF & PPT for Gold Subscribers)


Transesophageal echocardiography (TEE) fails to appropriately oversize CoreValve in approximately 50% of patients, according to a study published in the June 2014 issue of JACC: Cardiovascular Interventions.

Nicolo Piazza, MD, PhD, of the German Heart Centre Munich (Munich, Germany), and colleagues say computed tomography (CT) assessment is the better option for achieving appropriate sizing and reducing risk of paravalvular leak post procedure.

 

Methods
The researchers examined data on 157 high- or excessive-risk patients who underwent TAVR with CoreValve (Medtronic; Minneapolis, MN) at their institution between January 6, 2009 and June 6, 2010. Valve sizing was based on TEE measurement of the annular diameter and standard sizing criteria; 66 patients (42%) received a 26-mm valve and 91 (58%) received a 29-mm valve.
Using CT datasets from vascular screening, annular diameters were retrospectively calculated to gauge the percentage of patients in whom the manufacturer’s recommended oversizing criteria were met. This strategy is necessary to generate interference between the prosthesis and the annulus, providing adequate anchoring and sealing. CT-based oversizing calculations included area, mean, and perimeter.


Sizing Incorrect More Often with TEE

Substantial differences were seen when TEE and CT perimeter sizing were compared. TEE classified 95.5% of patients suitable for either the 26- (35.0%) or 29-mm (60.5%) valve, 10.8% suitable for the 31-mm valve, and another 10.8% suitable for the 23-mm valve. In contrast, CT perimeter classified 76.6% of patients suitable for either the 26- (15.3%) or 29-mm valve (61.3%), 21.7% suitable for the 31-mm valve, and none suitable for the 23-mm valve. Additionally, in the CT perimeter analysis, 8.9% of patients were judged to have annuli too large for currently available valve sizes.

With TEE, the average oversizing was 20.1 ± 8.2%. However, when CT data were applied, the actual oversizing was 10.4 ± 7.8%, which translates to an approximately 48% overestimation with TEE vs CT.

With TEE, 81.0% of patients achieved the recommended oversizing and thus were considered to have received the appropriate valve size, while 11.4% had excessive oversizing and 7.6% had insufficient oversizing. When CT perimeter data were applied, only 51% of patients achieved recommended oversizing, meaning 49% of patients received an inappropriate valve size.

A total of 38 patients (24.2%) experienced significant paravalvular leak. When patients met CT perimeter criteria, however, the rate was 21% lower than with TEE (13.8% vs 35.1%; P = .003).

On multivariable analysis, adherence to CT perimeter-based oversizing was an independent predictor of reduced risk of paravalvular leak (OR 0.36; 95% CI 0.14-0.90; P = .029). ROC curve analysis confirmed that TEE appeared to be the least efficacious imaging modality.

Further analysis identified CT perimeter-based oversizing of 9% for the 26-mm prosthesis and 9.6% for the 29-mm device as the optimal thresholds for predicting paravalvular leak.

TEE-Based Recommendations Faulty 

Dr. Piazza and colleagues observe that Medtronic’s 2007 oversizing recommendations, which assumed that TEE sizing is accurate, led them to believe they were achieving approximately 20% oversizing. They were surprised to find that it was in fact half that number.

“Obviously, this information would have had a substantial impact on [transcatheter heart valve] size selection: up to one-half of all patients were deemed to have received the incorrect CoreValve size and approximately 30% would have been deemed ineligible for the available CoreValve prostheses at that time,” they write. Additionally, the investigators note that in patients who developed significant paravalvular leak, oversizing was 3-fold greater with TEE (19%) than with CT perimeter (6.2%).

“This reinforces the message that TEE led us to select valves that were too small for patients’ anatomy, especially in those with significant [paravalvular leak],” Dr. Piazza and colleagues observe. “Our ROC analysis further reinforces the hypothesis that TEE-based sizing was a poor predictor of [paravalvular leak] when compared with CT measures.”

They add that CT perimeter sizing “appears to be the most sensitive CT-based measure for predicting [paravalvular leak] and is recommended for… all patients undergoing CoreValve implantation.”

 


Source:
Mylotte D, Dorfmeister M, Elhmidi Y, et al. Erroneous measurement of the aortic annular diameter using 2-dimensional echocardiography resulting in inappropriate CoreValve size selection: a retrospective comparison with multislice computed tomography. J Am Coll Cardiol Intv.

 

Related Stories:

Disclosures
  • Dr. Piazza reports serving as a proctor and consultant for Medtronic.

Comments