The Role of Echocardiography in the Management of Patients Undergoing a Ventricular Assist Device Implantation and/or Transplantation The Role of Echocardiography in the Management of Patients Undergoing a Ventricular Assist Device Implantation and/or Transplantation

Heart transplantation (HTx) is a curative treatment for patients with advanced heart failure (HF); however, since transplant opportunities are severely limited due to donor shortage, the left ventricular assist device (LVAD) has become a standard therapy for patients awaiting HTx. The role of echocardiography as a primary imaging modality to monitor the allograft function in transplant recipients as well as to optimize LVAD settings in LVAD recipients has been expanding. The purpose of this review is to highlight the clinical role of echocardiography in the management of patients under‐ going LVAD implantation and/or HTx. In particular, we overview (1) how to detect LVAD malfunction and device‐associated complication in LVAD recipients and (2) echocardiographic assessments of cardiac allograft rejection in transplant recipients.


Introduction
Heart transplantation (HTx) provides considerable survival benefits for patients with endstage heart failure, but it is available for only a small fraction of such patients all over the world due to donor shortage [1]. Therefore, many heart transplant candidates require longterm support by a left ventricular assist device (LVAD) while they await transplantation [1,2]. More recently, mechanical circulatory support has evolved into a standard therapy for patients with advanced heart failure, not only as a bridge to cardiac transplantation but also as a destination therapy or a bridge to myocardial recovery [3].
Echocardiography is a primary imaging modality in the assessment of cardiac structure and function in patients with advanced HF. In addition, echocardiography can be performed at the patient's bedside, and results are immediately available. In this review, we highlight the effectiveness of echocardiography in the management of patients undergoing LVAD implantation and/or HTx.

Echocardiography in LVAD recipients
A growing number of heart transplant candidates require long-term support by an LVAD while they await cardiac transplantation. Further, LVAD therapy has become a standard therapy for patients with advanced HF, not only as a bridge to cardiac transplantation but also as a destination therapy or a bridge to myocardial recovery. Here, we focused on the usefulness of echocardiography in patients undergoing LVAD implantation.

Preoperative assessment
It is important to assess the LVAD eligibility and rule out any contraindications against LVAD surgery prior to an operation. Several structural issues that can be surgically corrected at the time of LVAD implantation should be carefully evaluated prior to the LVAD surgery. The presence of clots, especially at the apex, should be carefully assessed because it will increase the risk of inflow cannula obstructions and/or perioperative stroke. Intracardiac shunts, including patent foramen ovale, should also be carefully assessed before and during surgery. Intracardiac shunts must be closed at the time of LVAD surgery. Further, coexisting valvular heart disease should be assessed prior to the LVAD procedure. Concomitant valvular surgery can be performed at the time of LVAD implantation; however, although such an additional approach can provide possible benefits, data regarding its long-term effect are limited, and the indications are still controversial. Another important issue to be carefully evaluated preoperatively includes right ventricular (RV) function because right ventricular failure (RVF) after LVAD placement is associated with increased morbidity and mortality.

Preoperative valvular assessment
Regarding tricuspid regurgitation (TR), several previous papers have revealed that tricuspid annular dilatation is highly associated with post-LVAD right ventricular failure [4]. Kukucka M et al. reviewed 122 patients without significant TR at the time of VAD implant and found that a tricuspid annulus diameter >43 mm was an independent predictor of survival after LVAD (Figure 1). On the other hand, whether the TR should be surgically managed at the time of LVAD surgery is controversial. Dunlay et al. performed a literature search of randomized controlled trials and observational studies (including 3249 patients) that compared the outcome of concomitant tricuspid valve surgery at the time of LVAD with that of LVAD alone [5]. They found that the addition of valvular surgery at the time of the LVAD procedure prolonged cardiopulmonary bypass times by an average of 31 minutes, but no differences were found between the groups for acute renal failure, early mortality, or the need for a right ventricular assist device. Having said that, a recent paper from Columbia University suggested that concomitant tricuspid valve procedures at the time of LVAD surgery can be performed safely and protect against worsening tricuspid regurgitation during the first two years of support [6]. In either case, the severity of TR and annular size need to be assessed preoperatively. Surgeons should also bear in mind that preexisting severe TR, especially with annular size >43 mm, is at higher risk of adverse events after surgery.  [4]). Kaplan-Meier survival curves of patients with tricuspid valve (TV) annulus diameter <43 mm (blue) and >43 mm (red). Censored patients are represented by vertical marks. Numbers of patients at risk at 0, 12, 24, and 36 months of follow-up are presented above the x-axis (log-rank test, p = 0.007).
Aortic insufficiency (AI) occurs in up to 50% of patients within 1 year after continuous flow LVAD implantation. Although de novo AI can be commonly seen postoperatively, the presence of more than mild AI as well as any structural abnormality, as detected by transthoracic echocardiography (TTE), should be reported to the LVAD surgery team. In cases with poor TTE images, the results derived from intraoperative transesophageal echocardiography (TEE) should be carefully discussed. The valvular morphology, valvular calcification, possible fusion, and myxomatous changes should also be reported to the surgeons. The recently published comprehensive review of AI post-LVAD by Cowger et al. suggested the importance of intraoperative TEE to detect unmasked AI [7]. During the initiation of continuous flow LVAD support, as LV filling pressures drop with early unloading, the gradient between the aortic root and the LV increases, potentially exposing significant AI that was previously unrecog-nized. Because AI severity can be associated with an increase in pump speeds, we can quantitatively assess AI severity at different pump speeds to consider the necessity of concomitant aortic valve surgery in an operating room. This review summarized the risk and benefits of aortic valve surgery at the time of LVAD (Table 1) Mitral valve insufficiency has fewer effects on postoperative outcome compared with aortic and tricuspid valve insufficiency. Indeed, a significant number of patients who had severe mitral regurgitation due to annulus dilatation and tethered pupillary muscle preoperatively showed a remarkable decrease in mitral regurgitation flow under LVAD support [8]. Although mitral valve surgery at the time of LVAD implant to correct severe mitral regurgitation does not affect postoperative mortality or cause other adverse events, the procedure can be considered in cases undergoing an LVAD procedure as a bridge to recovery. In addition, concomitant mitral valve repair can decrease pulmonary vascular resistance [9]. Kitada et al. investigated preoperative echocardiographic features associated with persistent mitral regurgitation after LVAD implantation (Figure 2) [10]. They found that the posterior displacement of the coapta-tion point of a mitral leaflet (30 vs. 24 mm), papillary muscle distance (49 vs. 43 mm), and tethering area (353 vs. 299 mm 2 ) before surgery were greater in patients who had persistent moderate to severe mitral regurgitation post-LVAD than those in patients who did not have significant MR postoperatively. A multivariate analysis showed that the posterior displacement was the only independent predictor for persistent MR.

Preoperative and perioperative right ventricular assessment
Right ventricular failure (RVF) remains a major cause of morbidity and mortality following LVAD surgery. The incidence of RVF post-LVAD is 10-30% despite the recent improvements in device technology and postoperative patient management. Under LVAD support, right ventricular (RV) preload increases as a result of increased circulatory volume, whereas RV afterload is expected to decrease, secondary to improvement in pulmonary vascular resistance [11]. A sepal wall shift induced by LVAD alters the RV structure, which may worsen RV contractile and relaxation abnormalities. Therefore, when considering RV systolic and diastolic reserve before and also after surgery, it is important to identify which patients may need RVspecific mechanical and medical support post-LVAD [12].    [13]. In addition to these risk scores, serial echocardiographic assessments are helpful in evaluating RV functional reserve prior to surgery. Previously reported echocardiographic parameters associated with the risk for developing RVF after LVAD implantation have included tricuspid annular dilation (>43 mm) [4], tricuspid annular motion (8 vs. 15 mm) [14], and RV-to-LV end-diastolic diameter ratio (>0.72) [15]. However, it is sometimes technically difficult to obtain ideal RV images that allow quantitative assessments of patients with advanced heart failure, particularly if the patients are severely congested, intubated, and/or have a markedly enlarged left ventricle (LV) that obscures the right ventricle (RV) [16]. Kato et al. focused only on left-sided 2D echo parameters that can predict RVF post-LVAD. They showed that patients with relatively small LV size, preserved LV contraction, and a dilated left atrium were at higher risk for RVF after LVAD surgery (Figure 3) [16]. In addition to the conventional echo parameters, Grant et al. reported that the incremental role of RV strain to predict RVF [17]. More recently, Kato et al. reported that serial echocardiograms using tissue Doppler imaging (TDI) and speckle tracking echocardiography (STE) before and soon after (within 72 hours) LVAD surgery may aid in identifying the need to initiate targeted RVFspecific therapy [12]. In this study, RV stiffness (as reflected by TDI-derived E/E′) and decreased RV contractility (as reflected by TDI-derived S′ and RV longitudinal strain) before and soon after LVAD surgery were found to be useful parameters to include in the perioperative management of LVAD patients (Figure 4).

Perioperative assessment
Other than the speed adjustment to avoid RV failure due to excessive RV preload by LVAD support, several important points should be evaluated by intraoperative TEE. First, de-airing of the heart chamber should be confirmed. Careful observation of trapped air at the site of anastomosis sites and around the LVAD inflow/outflow cannula is required [18]. Second, adjusting LV speed to maintain appropriate LV unloading without a septal shift under TEE guidance is required. The positioning of the inflow cannula at the apex should be monitored by TEE as well. Third, as mentioned above, the existence of valvular diseases and intracardiac shunts, which can be corrected simultaneously at the time of LVAD implantation, should be communicated to the surgeons. Finally, pericardial effusion and its amount should also be carefully observed by TEE. Cardiac tamponade can occur relatively often because patients under LVAD support require sufficient anticoagulation soon after surgery to prevent clot formation at the cannula and inside the device. Table 3 illustrates the checklist that will help sonographers/echocardiologists to perform an LVAD echo. In general, we can simply summarize the purposes of echo in LVAD recipients as follows: (1) to carefully monitor device malfunction, (2) to adjust appropriate LVAD setting/ speed (appropriate peripheral perfusion and RV preload), and (3) to evaluate myocardial recovery and to seek optimal timing for LVAD weaning.

Postoperative assessment
The points to be evaluated by TEE on a periodic basis are as follows: the location and thrombus at the inflow cannula; LV cavity diameters; septal position; RV function; valvular regurgitation, especially about the aortic valve opening/intervals and regurgitation. • Positioning (against LV wall) Table 3. Echo LVAD checklist.

General postoperative assessment in LVAD recipients
Recommendations for device speed adjustment include the target measures of mean arterial pressure above 65 mmHg, maintaining the position of interventricular septum and shape, and intermittent aortic valve opening, under the condition of no more than mild mitral regurgitation to ensure appropriate unloading of the LV. Optimization of speed settings is extremely important to prevent several of the key complications associated with chronic LVAD support. The importance of ensuring the middle septal position for optimal RV function has been well established [19,20].
Serially monitoring the timing and its interval of aortic valve opening in all LVAD recipients are necessary. Also, adjusting the LVAD speed to maintain the aortic valve opening is important to prevent the development of aortic valve regurgitation. At least 10 cardiac cycles should be recorded to evaluate the aortic valve opening. Because the interval of aortic valve opening, LV diameter, and grade of MR entirely depend on the degree of LV unloading, the LVAD setting together with the echo report needs to be recorded ( Figure 5). Aortic regurgitation is sometimes seen with atypical timing (Figure 6) or continuously, both during the diastolic and systolic phases [21].
Cardiac output using RV outflow-derived Doppler estimation can be calculated as follows: cardiac output = stroke volume × heart rate, stroke volume = π × (RV outflow diameter/2) 2 × time velocity integral at RV outflow. In patients who have at least an intermittent aortic valve opening, RV cardiac output minus LV outflow-derived cardiac output is equivalent to the estimated pump flow. Figure 5. Representative images in a LVAD recipient with different LVAD speeds. This patient received HeartMate II (Thoratec Corp) implantation. Under 8400 rpm, the aortic valve opened intermittently, and the ejection time was only 75 ms. When we set the speed down to 8200, the aortic valve opened every beat, but due to less unloading, the LV diameter increased and the amount of mitral regurgitation also increased. Figure 6. Aortic regurgitation during systolic phase accompanied by mitral regurgitation in patients with a continuous-flow left ventricular assist device (quoted from Ref. [22]). Echo images obtained from a patient undergoing LVAD implantation who showed systolic-phase aortic regurgitation (AR). The timing of the regurgitation jet started at the mid-systolic phase and ended at the early diastolic phase (A). The AR occurred slightly after the onset of mitral regurgitation (MR) (B), and both MR and AR timings were consistent with the systolic phase. No remarkable AR jet was documented during the diastolic phase. The AV was mostly closed throughout the cycles, which opens once every 8-10 beats (C). The mean pressure gradients of the trans-AV and trans-mitral valve based on the continuous wave Doppler measurements of AR (D) and MR flow (E) were 3.7 and 24.3 mmHg, respectively. The morphology of AV annulus changes through the cycles irrespective of the AV opening, with the AV annulus abnormally distorted and dilated during early mid-systole (F), whereas the septum wall as well as the AV annulus edge slightly pushed toward the LV during diastole (G).
Serial assessments of pulmonary artery pressure by Doppler-derived TR pressure gradients are also important. In general, LVAD support can successfully unload LV, which results in the correction of pulmonary hypertension due to left-sided heart failure. However, some patients have showed residual pulmonary vascular resistance post-LVAD; therefore, echo-guided optimal medical therapy, including the necessity of pulmonary dilators such as PDE5 inhibitors (sildenafil®, etc.), is required.

Detection of LVAD malfunction
The careful observation of the inflow cannula is critically important. By using multiple views, including nonstandard ones, the thrombus or other causes of obstruction should be ruled out. The direction of the inflow cannula should also be reported. The direction may sometimes change after the surgery and direct toward the lateral wall, which may cause suctioning or inadequate LVAD support. Contrast echocardiography can provide additional information.
Detecting the outflow cannula obstruction by echocardiography is difficult, but practitioners should try to find a good echo window and investigate any abnormality, including kinking (Figure 7) [22,23]. The protocol for a ramp study was established by Uriel N [20]. It is useful in optimizing LVAD settings and in diagnosing device malfunctions. Ramp test echocardiography can be performed at the time of discharge for speed optimization and/or if device malfunction is suspected (Figure 8) [24]. The patient's left ventricular size, the frequency of the aortic valve opening, valvular insufficiency, blood pressure, and continuous flow-LVAD parameters should be recorded according to the increments of the device speed. Serial assessments of ramp tests are also helpful to detect LVAD clots [24]. Figure 8. A representative case with device thrombosis was detected by a ramp echocardiography device (quoted from Ref. [24]). A case of a 29-year-old woman undergoing HeartMate II LVAD implantation; serial ramp studies were used to diagnose intradevice thrombus after device implantation. The first ramp study on postoperative day (POD) 26 revealed an adequate reduction in ventricular size according to the increase in LVAD. The patient was discharged home and received routine anticoagulation maintenance therapy. However, a second ramp test was performed on POD 56 due to increased lactase dehydrogenase and brain natriuretic peptide levels and showed a marked increase in the LV chamber size without an adequate response to the LVAD speed changes. Given the suspicion for partial pump thrombosis, the patient was immediately hospitalized and received intravenous heparin infusion. The patient eventually underwent cardiac transplant successfully, and the partial clot was found inside of the pump (lower panel).

Assessment of native cardiac function
It is important to assess native LV function, especially in patients receiving LVAD as a bridge to recovery. We cannot assess LV function without turning off the LVAD because it drastically affects preload and afterload; therefore, we need to reduce the LVAD speed under adequate anticoagulation during weaning test echocardiography. Strain assessment has been reported to be more sensitive in evaluating the myocardial systolic and diastolic reserve, and 2D speckle tracing echocardiography for the assessment of myocardial recovery in LVAD recipients may be useful [25].

Donor heart evaluation
Evaluating a donor heart as accurately as possible at the time of procurement provides essential information to a recipient team leading the delicate posttransplant management of the heart [26]. If an organ procurement team has a cardiologist or sonographer who knows which patient is going to receive the heart, the team can gather detailed information by bedside echocardiography on the donor in light of the potential recipient's conditions at the organ procurement.
Measuring the heart size of the donor from bedside echocardiography at the time of organ procurement can provide useful information for judging the appropriateness of proceeding with the heart transplant in the case of a donor-recipient size mismatch. The wall thickness of the donor heart may also be useful information for optimizing the medical therapy after transplantation, as well as for deciding whether or not to use the organ. Information regarding the presence or absence of a septal defect would be of help to surgeons planning the additional procedure of septal closure at the time of transplantation. Information about the coronary flow in the left anterior descending artery of the donor heart, especially in cases with coronary risk factors, is useful for judging the availability of the heart, as well as for considering issues related to posttransplant medical management. Finally, information about preexisting localized wall motion abnormalities from bedside echocardiography is useful for speculating on the possibility of rejection or other reasons for wall motion abnormality after transplant surgery.
According to such information, the team can make a final decision whether or not to harvest the heart. For example, the donor heart may be relatively small for the potential recipient. If a donor heart with a lower limit of normal systolic function shows decreased coronary flow and localized right heart wall motion abnormality, the heart should be declined in cases where the potential recipients have moderately high pulmonary vascular resistance. Such recipients need to receive a donor heart with good right ventricular function.

A noninvasive rejection diagnosis
Advances in immunosuppressive therapy have resulted in a marked decrease in the incidence of acute allograft rejection in heart transplant recipients; however, acute rejection still remains an important determinant factor for long-term morbidity and mortality. Acute rejection can result in not only the immediate risk of graft loss or heart failure but also of subsequent allograft vasculopathy [27]. Therefore, early diagnosis of rejection and consequent timely treatment are crucial for the early and long-term care in heart transplant recipients. Detection of allograft rejection based on the findings derived from endomyocardial biopsy (EMB) is still a gold standard; however, EMB is invasive, cost and time consuming, and may have a possibility of sampling error and interobserver variability. Although many noninvasive modalities, including radionuclide imaging, MRI, and gene expression profiling, have been investigated for their potential to detect rejection, none of them have been found to be sufficient for replacing EMB. Echocardiography has been routinely used in the management of cardiac transplant recipients. Indeed, it is an easily applicable, repeatable, and powerful noninvasive tool in the management of posttransplant recipients [28].   Table 4 summarizes the conventional echocardiographic parameters associated with acute cellular rejection [28]. Conventional echocardiography soon after the surgery can provide information about global systolic and diastolic functions, wall motion abnormality, and the hemodynamics of the transplanted hearts. Any apparent abnormal findings such as remarkable systolic and/or diastolic impairment may acute or hyperacute rejection, including antibody-mediated rejection, although primary graft failure, donor-related graft dysfunction, and any perioperative accidents should also be considered. The ability of conventional echo parameters to detect rejection is still limited to severe clinically detectable rejection. However, the findings are still useful for assessing responsiveness to treatment. In general, patients with rejection develop restrictive physiology accompanied by various degrees of systolic dysfunction. Valantine HA et al. reported that a 15% decrease in mitral deceleration time or isovolumic relaxation time (IVRT) is associated with biopsy proven rejection [29]. More recently, Sun et al. reported that a combination of IVRT less than 90 ms, a mitral E/A ratio more than 1.7, and other clinical parameters is independently associated with rejection [30]. However, because transplant recipients usually have higher resting heart rates than the nontransplant population due to denervation, their mitral E and A waves can be fused. Indeed, it is difficult to obtain clear Doppler waves from transplant recipients. They frequently have extended adhesion of the transplanted heart to the chest cavity, which hinders the acquisition of an appropriate Doppler angle. The TEI index or myocardial performance index (MPI), which is a parameter of a Doppler-derived combination of systolic and diastolic time intervals, is a useful parameter in patients with E-A fusion and high heart rate; therefore, the MPI has the potential to detect rejection more accurately than traditional Doppler indices [31]. Representative conventional 2D echo images associated with and without rejection are shown in Figure 9.  • Values can be calculated offline using stored 2D images A′, late diastolic mitral annular velocity, E′, early diastolic mitral annular tissue velocity. * LV torsion = (apical endsystolic rotation) -(basal end-systolic rotation). Table 5. Tissue-Doppler imaging and 2D-speckle-tracking echocardiography-derived variables associated with rejections (quoted from Ref. [29]).

Tissue Doppler imaging and speckle tracking echocardiography
Tissue Doppler imaging (TDI) enables the measurements of systolic and diastolic velocities within the myocardium. Several studies have evaluated the usefulness of TDI-derived mitral annular velocities to detect allograft rejection, which are summarized in Table 5 [28]. Strain rate analysis has a potential to detect even mild rejection. Kato TS et al. reported that the attenuation of LV longitudinal strain and the diastolic strain rate derived from TDI were associated with conventional ISHLT (International Society for Heart and Lung Transplantation) grade 1b or higher rejection without hemodynamic alterations (Figure 10) [32]. Marciniak et al. found significantly lower LV longitudinal and radial peak systolic strain and strain rate values in patients with conventional ISHLT grade 1b or higher rejection. TDI-derived strain and strain rate potentially reflect abnormalities [33]. Figure 10. Representative TDI-derived strain analysis in a patient with and without rejection (quoted from Ref. [33]). Representative pathological findings for EMB specimens (a, b) and strain analysis (c, d) of HTx with ISHLT grade 0 (a, c) or 3a (b, d) rejection. Sections in (a) and (b) were stained with hematoxylin-eosin; scale bars, 100 mm. EMB, endomyocardial biopsies; HTx, heart transplant recipients; ISHLT, International Society for Heart and Lung Transplantation.
Two-dimensional speckle-tracking echocardiography (2D-STE) was developed as an angleindependent echocardiographic modality to evaluate cardiac mechanical function. The 2D-STE-derived parameters associated with rejection are also shown in Table 5 [28]. The association between LV torsional deformation and rejection in transplant recipients has been reported since the 1980s. Sato et al. reported that 2D-STE-derived LV torsion values are decreased in patients with rejection, and the serial assessments of an intra-patients comparison showed that a cut-off value of a 25% reduction of LV torsion from the baseline is associated with ISHLT grade 2 or higher rejection, which returns to the baseline after adequate rejection treatment (Figure 11) [34]. LV global strains are also calculated using 2D-STE in an angleindependent manner. Sera F et al. reported that 2D-STE-derived LV global longitudinal strain was associated with treatment-requiring rejection [35] (Figure 12). In addition to its major advantage of angle independency, 2D-STE has other advantages over TDI, such as spatial resolution, translational artifacts, the sensitivity to signal noise, the time needed for data acquisition, and the necessity of employing expert readers. Three-dimensional (3D) STEs are useful echocardiographic modalities to assess various strain and rotation parameters more accurately than 2D-STE by tracking the same speckle throughout the cardiac cycle. However, it will take several years for the validation studies of 3D-STE to be performed to verify the value of rejection-detecting tools in heart transplant recipients. Figure 11. Representative 2D speckle-tracking echocardiogram and analysis of torsion in a patient with and without rejection. (quoted from Ref. [35]). Representative 2D-STE imaging with rotation curves obtained from the same recipient (a 32-year-old man) at LV short-axis views of the apex (a, c) and the base (b, d). Each color of the deformational curve represents one segment of the LV, and the dashed white curve depicts the mean rotation of six segments. The LVtor, defined as the difference between apical basal end-systolic rotation when the patient had ISHLT grade 2 rejection, was 10.9 degrees (a, b). The LV torsion accompanied with ISHLT grade 0 after rejection treatment was 15.6 degrees (c, d). The % change of LV torsion in this patient at the time of rejection was approximately 30% decreased from his baseline. ISHLT, International Society for Heart and Lung Transplantation; LV, left ventricular; 2D-STE, 2D speckle-tracking echocardiography; EMB, endomyocardial biopsies; HTx, heart transplant recipients.

Transplant vasculopathy and echocardiography
Echocardiography is a helpful and an ideal noninvasive tool to detect transplant vasculopathy or chronic rejection as well. Dobutamine or/and exercise stress echocardiography has been used to detect allograft vasculopathy, especially for pediatric patients or those with renal insufficiency [36]. Decreases in strain and strain rates at rest and with dobutamine stress are also useful to detect significant transplant vasculopathy. Contrast echocardiography is another useful method.