The percutaneous closure of mitral paravalvular leaks has been reported in patients who are poor operative candidates. Unsuccessful percutaneous closure of leaks may be related to morphologic characteristics of the defects.
Ten patients were selected from a database for mitral dehiscence closure, in whom two-dimensional transesophageal echocardiography revealed inadequate leak closure. Another 4 patients with optimal results were also selected. Real-time three-dimensional transesophageal echocardiography (3DTEE) was performed in all of them.
Real-time 3DTEE enabled the determination of the locations and number of the leaks, as well as their shapes, lengths, widths, areas, and extent. We were also able to observe the position of the device (or devices) implanted during percutaneous closure.
According to this preliminary study, 3DTEE can improve understanding of the causes underlying failure of these techniques to reduce regurgitation secondary to a defect. This could improve patient selection and procedure results, but further studies are needed.
Perivalvular dehiscence is a complication that may occasionally occur in patients with prosthetic valves, especially if these are mechanical. They occur in approximately 2% to 17% of cases and in up to 10% of cases of prosthetic valve reoperation, with a prevalence of approximately 10% to 15% in follow-up studies using standard two-dimensional (2D) transesophageal echocardiography (2DTEE). The surgical closure of paravalvular leaks is usually advised in severely symptomatic patients and in those requiring blood transfusions for persistent hemolysis. Operative mortality is 6% to 14%. The percutaneous closure of severe periprosthetic leaks has been reported in patients who are poor operative candidates, with different devices and techniques. Our group recently published results from the longest series of cases presented in the literature of periprosthetic mitral valve leak closure, in which total or a highly significant reduction in the severity of regurgitation was shown. Unsuccessful percutaneous closure of leaks may be related to their large sizes and shapes, which prevent closure using simple devices. Two-dimensional TEE allows only an approximate assessment of the morphology and extent of a leak. However, the recently developed real-time three-dimensional (3D) TEE (3DTEE) offers excellent imaging of the different cardiac structures, basically (being a transesophageal technique) of the mitral valve, left atrium, and prosthetic mitral valves.
The purpose of this study was to evaluate the utility of 3DTEE in analyzing the underlying causes of unsuccessful leak closure. To do this, the technique was used to analyze a cohort of patients who were found to have unsuccessful periprosthetic leak closure with persistent severe regurgitation at a 2-month postoperative follow-up study.
The database of our group for percutaneous prosthetic mitral valve closure currently consists of 52 patients, 27 of whom formed part of the previous study. All percutaneous interventions were performed at the same institution and by the same operator (E.G.). The devices used in all interventions were patent ductus arteriosus occluders (Amplatzer; AGA Medical Corporation, Minneapolis, MN). All paravalvular leak closure procedures were performed under general anesthesia with radiographic and 2D transesophageal echocardiographic guidance. Monitoring with 2DTEE was used to locate the optimal region of the interatrial septum for transseptal puncture and help the hemodynamist introduce the guide through the defect and correctly deliver the device. Two-dimensional TEE was also used to monitor the onset of possible complications (interference with prosthetic valve leaflets) and to assess the final result of the procedure (measurement of the grade of residual periprosthetic insufficiency).
Additional 2DTEE was performed ≥1 or 2 months after the procedure in all patients with successful device implantation. This is the minimal period of time considered necessary for the possible onset of thrombosis and device endothelialization.
Devices were successfully deployed in 33 of the 52 patients (63.5%). In 2 such patients, in whom mitral regurgitation remained significant, second devices were implanted. Implantation of the devices was able to reduce the severity of regurgitation in 17 patients (51.5% of patients with successful implantation) at final follow-up. In the other 16 patients, the degree of regurgitation remained unchanged despite ≥1 implanted device. Those patients in whom device implantation failed to reduce the degree of regurgitation constituted our main study group. We selected these interventions (with implanted devices but with unchanged or significant residual regurgitation) in the last 2 years prior to study design (15 patients). These patients were contacted by phone to propose the development of 3DTEE. Three patients died, another was reoperated on for mitral prosthesis replacement, and another declined to be included in the study. We also proposed the study to those patients who showed complete reductions in regurgitation via 2DTEE in the last year (4 patients).
Ultimately, our study population consisted of 14 patients. Ten of these patients had inadequate leak closure on 2DTEE with persistent severe mitral insufficiency 2 months after the operation. The other 4 patients showed excellent results on 2DTEE and no further mitral regurgitation. All patients agreed to participate and signed informed consent forms.
The study was conducted using the Philips iE33 ultrasound system and the X7-2 t matrix transesophageal real-time transducer with 2600 elements and PureWave crystal technology (Philips Medical Systems, Andover, MA). Once the best plane has been located using the standard 2D image, the 3D Zoom feature displays a biplane image on which two orthogonal regions of interest must be placed and adjusted in length, width, and depth to fit the target structure. Engaging the 3D Zoom feature again displays the live 3D image. Once the best plane has been located using the standard 2D Doppler color image, the Full Volume feature presents a biplane image on which two orthogonal color regions of interest must be placed and adjusted in length, width, and depth to fit the target structure. By then pressing Acquire, 7 subvolumes are captured, and the synchronized electrocardiogram is displayed. All 3D volume data sets were processed and cropped using QLAB (Philips Medical Systems) to obtain the most beneficial view.
The number of acquisitions during the first 4 cases was as follows: two 3D Zoom acquisitions, including the complete mitral prosthetic valve and walls of the left atrium; another two 3D Zoom acquisitions focused on the particular location of the paravalvular leak; and two color Full Volume acquisitions, including the complete mitral prosthetic valve. Image quality was optimal in all cases. Once we reached the learning curve, the number of acquisitions for the remaining 10 cases was reduced as follows: one 3D Zoom acquisition, including the complete mitral prosthetic valve and walls of the left atrium; one 3D Zoom acquisition focused on the particular location of the perivalvular leak; and 2 color Full Volume acquisitions, including the complete mitral prosthetic valve. Image quality was again optimal in all cases.
In a manner similar to that described in the methodology of our previous study, the mitral annulus was divided into 4 quadrants for the purpose of classifying location. To define the origin and extent (in terms of degree along the prosthetic annulus) of the periprosthetic regurgitation, the echocardiographic mitral annulus was divided into sections similar to those used in heart surgery: anterior (between 9 and 12 o’clock on the annulus, corresponding to the area situated between the aortic root and the exit of the left atrial ear), septal or inner side (between 12 and 3 o’clock, along the interatrial septum), posterior (between 3 and 6 o’clock, when the regurgitation flows along the free wall of the left atrium when the echocardiographic probe is rotated between 90° and 120°), and lateral or outer side (between 6 and 9 o’clock, with the periprosthetic jet flowing along the free wall of the atrium when the probe is at 0° rotation). The severity of regurgitation was determined by studying the characteristics of regurgitant flow (area, length) on color Doppler and flow in the pulmonary veins and by calculating actual regurgitant area. The degree of regurgitation was classified into 4 grades: I (mild), II (mild to moderate), III (moderate), and IV (severe).
We also studied (using 3DTEE) the number of defects, as well as the shape, width, length, and area of each dehiscence. We measured width and length in millimeters. Area was measured in square millimeters. The measurement of anatomic characteristics (area, width, and length) of each dehiscence leak involved the extraction of a 2D plane from the volume using the multiplanar reconstruction tool. These measurements were performed offline using QLAB.
In all cases, 2DTEE allowed the determination of the severity of regurgitation on the basis of standard criteria. However, this technique does not help assess the exact anatomic characteristics of the residual dehiscence. Two-dimensional TEE could not perform measurements of either the area or the length of a dehiscence, nor could it provide a delineation of its shape.
The 3D transesophageal echocardiographic technique enabled us to the determine location of the periprosthetic dehiscence in each of the patients, as well as the number of such defects. We also studied shape and extent of the dehiscence, in terms of degree, along the prosthetic annulus. We were further able to observe the position of the device (or devices) implanted during percutaneous closure inside the dehiscence, and we were able to take measurements to estimate the length, width, and area of the dehiscence. None of the patients had any complications as a result of 3DTEE. Figures 1 to 3 show images obtained during the studies conducted on our group of patients.
Table 1 summarizes the characteristics of echocardiographic dehiscence after the completion of percutaneous closure of the defects in the 14 patients included in our series. The table compares information obtained using 2DTEE and 3DTEE.
|Patient||Severity ∗||Device implanted||Location||Extent||Shape||Location||Extent||Shape||Width (mm)||Length (mm)||Area (mm 2 )|