Seven adult first cross merino ewes were used for 20 implantations of Impella CP®. Sheep characteristics are presented in Table E1. Haemodynamic variables prior to the implantation of the Impella CP® are presented in Table E2.
Initial position guidance:
Three-dimensional en-face images of the aortic valve (Figure 1, Movie 1) offered delineation of the three leaflets and assessment of their motion to exclude pre-existing aortic valve pathology. The quality of AV assessment was found to be adequate in all cases. The aortic root and proximal ascending aorta were imaged from the right atrial position of the 3D ICE catheter in all cases. Significant intracardiac or aortic anatomic abnormalities were not found in any animals. Activation of the three-dimensional Color Doppler for assessment of aortic valve incompetence did not reveal underlying pathology beyond trivial aortic regurgitation in two animals (Figure 2, Movie 2). Advancement of the catheter further into right atrium provided views of the left ventricular outflow tract and left ventricular cavity including both papillary muscles (Figure 3, Movie 3). The 2D and 3D cut-plane views offered long-axis views of the left ventricle. Minor adjustments in catheter position with retroflexion (posterior steering tilt) were often required to optimize imaging of the mid-and apical portions of the left ventricle. This view offers potential identification of left ventricular thrombus. Anticlockwise rotation opened the ventricular view of the mitral annulus (Figure 4, Movie 4) and en-face view of the mitral valve in all cases.
The J-tip 0.035-inch stiff access guidewire was visible in all cases within the aortic root and ascending aorta (Figure 5, Movie 5). A J-wire loop forming in the aortic root in one case was immediately identifiable on both 2D and 3D imaging. The position of the cut-plane in 3D ICE volume had to be at times adjusted to include lateral parts of the left ventricular cavity for better visualization of the wire. The wire was clearly identified by 2D ICE, but fan-like rotational manipulations of the ICE catheter were required to place the wire within 2D plane. The spatial relationship between the wire and cardiovascular structures was significantly better appreciated with 3D ICE in all cases. However, some tilting of the 3D volume on the screen was helpful for identification of the wire within left ventricular cavity due to reverberation artifacts. Some reduction in dynamic range settings was helpful for sharper and faster wire visualisation. Inappropriately high gain hindered identification of the wire, especially when it was positioned in a proximity and parallel to the walls of the left ventricle.
A diagnostic catheter was inserted over the stiff guidewire and was also clearly identifiable on 3D ICE in all cases and did not require 2D ICE imaging for clarification. It appeared slightly more echogenic than the wire when compared side by side, but not significantly different when imaged by itself. A side-lobe artifact arising from the guidewire and to the lesser degree from the catheter was noted on several instances (Figure 6, Movie 6). 2D ICE was sufficient to demonstrate the diagnostic catheter traversing the aortic valve and entering the LV cavity. However, the position of the catheter tip was not always clearly visible (Figure 7) with 3D ICE providing significantly better positional visualisation (Movie 7). The use of 3D ICE identified catheter malposition retrogradely entering the left atrium via the mitral valve prompting repositioning of the catheter in two sheep (Figure 8).
Thin soft 0.018-inch guidewire was more difficult to visualise in most cases with inconsistent identification of the forming loops, especially with 3D ICE.
The advancement of the Impella pump over the guidewire into the ventricle could be visualized during all Implantations with minor requirement for adjusting the cut-plane position, and occasional rotational adjustment of the ultrasound probe. The reinforced catheter of the Impella correctly appeared in the image as a double-walled structure with cut-plane positioned along the catheter. It created prominent reverberation artifact (Figure 9, Movie 8). The tear-drop appearance of the metal cap between blood inlet area and plastic pigtail was highly echogenic, making it an ideal 3D marker for positioning (Figure 10, Movie 9). The Impella plastic pigtail was difficult if not impossible to visualise in all cases. Strong reverberation artifact arising from the tear-drop metal cap at the inflow further complicated visualisation of the plastic pigtail. The relationship between the Impella CP® catheter inflow portion of the system and surrounding cardiac structure was superior with 3D ICE when compared to 2D ICE in all cases. However, superior spatial and temporal resolution of 2D imaging offered better appreciation of the aortic valve leaflets and mitral subvalvular apparatus in relation to the Impella. One of the insertions identified the tip of the catheter being impacted under the posteromedial papillary muscle prompting repositioning with slight withdrawal of the catheter (Figure 11). In 8 implantations the tear drop initially could not be clearly identified due to the impaction into the apex. Slow gradual withdrawal of the Impella under 3D ICE guidance was undertaken until tea-drop became obvious in the mid-LV-cavity.
Overall, the image quality of highly relevant for Impella® implantation cardiac structures with 3D ICE was good to excellent (scored 8.6 on a scale 0-10). The quality of imaging components of the procedural equipment with 3D ICE was good (scored 7.8 on a scale 0-10) with exception of thin soft guidewire, which allowed for barely adequate quality.
3D Colour Doppler produced extensive “colour bleeding” artifact at the standard settings and required significant reduction in colour gain to identify the blood inlet in the catheter (Figure 12). Identification of the inlet on 3D Colour Doppler was achieved in 18 implantations and served as an additional confirmation of correct tea-drop identification and inflow site relative position to the cardiac structures. The “3.5 cm” rule applicable for human adults for the position of inflow below aortic valve was not used due to the different anatomical characteristics of the ovine left ventricle. ICE was sufficient in all cases to ensure adequate flows and absence of inflow obstruction as detected by Impella® Controller.
The outflow could be visualized in all cases as an extensive turbulence within the proximal ascending aorta above the aortic root. Reduction in colour gain diminished colour “bleeding” over the myocardial tissue in all cases and offered good confirmation of the pump outflow position within the aorta.
3D Colour Doppler was used to reassess the aortic valve for potential incompetence following implantation of the Impella CP®. Mild and moderate peri-catheter aortic regurgitation was easily identified on five occasions (Figure 13, Movie 10).
Postprocessing with adjustments in dynamic range, gain, and colour priority and transparency for each individual image presents better visualization and appreciation of spatial relationships between the catheter and cardiovascular structures.
Insertion of the guidewire and the diagnostic catheter inside of the left ventricle frequently caused ectopic cardiac beats due to the direct myocardial irritation by the devices.
There were two failures of the Impella implantation to achieve pump flow:
Case 1. The guidewire became tangled around intra-aortic pressure catheter and could not be removed after insertion of the pump across the aortic valve. It occurred in sheep 2 during the third implantation. Both 2D ICE and 3D ICE failed to identify the cause of the problem. Removal of the Impella® demonstrated severely kinked soft guidewire.
Case 2. The excessive portion of the guidewire within left ventricular cavity became kinked and tangled around Impella catheter and could not be removed after insertion of the pump across the aortic valve. An attempt by the operator to apply extra pulling pressure on the guidewire resulted in kinking of the Impella catheter approximately 10 cm above the pump motor housing causing perforation of the carotid artery and resulting in catastrophic bleeding. It occurred in sheep 5 during the second implantation. Both 2D ICE and 3D ICE failed to identify the cause of the problem.
In one sheep the ICE probe entered the right internal thoracic vein and was identified by unusual imaging with the view resembling right ventricular-centric apical transthoracic echocardiographic view. Multiple attempts at repositioning of the probe were required before it was successfully negotiated into the right atrium. This complication occurred in sheep 3 during the first implantation and is likely to be idiosyncratic to sheep.