Current Cardiology Reports

, 16:468

The Practical Role of Echocardiography in Selection, Implantation, and Management of Patients Requiring LVAD Therapy

Authors

  • Maria Chiara Todaro
    • Cardiology Unit, Department of Clinical and Experimental MedicineUniversity of Messina
  • Bijoy K. Khandheria
    • Aurora Cardiovascular ServicesAurora Sinai/Aurora St. Luke’s Medical Centers, University of Wisconsin School of Medicine and Public Health
  • Timothy E. Paterick
    • Aurora Cardiovascular ServicesAurora Sinai/Aurora St. Luke’s Medical Centers, University of Wisconsin School of Medicine and Public Health
  • Matt M. Umland
    • Aurora Cardiovascular ServicesAurora Sinai/Aurora St. Luke’s Medical Centers, University of Wisconsin School of Medicine and Public Health
    • Aurora Cardiovascular ServicesAurora Sinai/Aurora St. Luke’s Medical Centers, University of Wisconsin School of Medicine and Public Health
Echocardiography (RM Lang, Section Editor)

DOI: 10.1007/s11886-014-0468-5

Cite this article as:
Todaro, M.C., Khandheria, B.K., Paterick, T.E. et al. Curr Cardiol Rep (2014) 16: 468. doi:10.1007/s11886-014-0468-5
Part of the following topical collections:
  1. Topical Collection on Echocardiography

Abstract

Viable treatment options for advanced heart failure have not emerged as the number of people afflicted with this condition has grown. Although heart transplantation is the only curative strategy for patients with end-stage heart failure, the relative shortage of donors has led to a worldwide plateau of this option over the past 20 years. The result is an unacceptably high mortality rate among patients with advanced heart failure. Interest in developing alternative curative strategies based on chronic circulatory support, with the aim of prolonging and improving quality of life for these patients, has grown. Patients supported with left ventricular assist devices require structured longitudinal care from a team of providers. An integrated approach using basic echocardiography is critical to patient selection, implantation, and continued surveillance and success of patients with left ventricular assist devices.

Keywords

Left ventricular assist deviceHeart failureDoppler echocardiographyRight ventricular failure

Introduction

Although overall cardiovascular mortality continues to decline because of improved treatments for acute and chronic cardiovascular conditions, the incidence and prevalence of heart failure are increasing. This trend is driven by an increase in the elderly population [1]. In the United States, 10,000 people will turn 65 every day until the year 2030, and three-quarters of all heart failure can be found within this demographic. A worldwide heart failure epidemic is anticipated during this same time interval, and, unfortunately, despite more than 25 years of primarily medical therapy, there have been marginal improvements in 5-year survival [2]. Most patients with symptomatic heart failure will be faced with disease progression characterized by functional limitations, deterioration of quality of life, and early death. Therefore, strategies beyond medical therapy have been, and will continue to be, created to address this growing need. Heart transplantation has been the traditional approach to the care of patients with end-stage heart diseases. However, despite the relative improvement in organ availability, the number of heart transplantations annually has never met the growing need [3]. In this regard, the single best approach to the care of patients in the end stages of heart failure is left ventricular assist device (LVAD) therapy.

Disabling symptoms and improving quality of life and prognosis are central to all successful therapies for advanced heart failure. Within the past 25 years, LVAD therapy has evolved to achieve these goals and, in some populations, is the most viable option for end-stage heart diseases. The REMATCH (Randomized Evaluation of Mechanical Assistance for the Treatment of Congestive Heart Failure) clinical trial was the first to randomly assign patients ineligible for transplantation (destination therapy) to either continued optimal medical therapy or electronically driven LVAD therapy (HeartMate XVE, Thoratec Corp., Pleasanton, CA), and demonstrated a significant survival benefit of LVAD therapy at 12 months compared with medical therapy alone (52 % vs 25 %, respectively) [4]. Subsequent studies have demonstrated incremental improvements in survival, device durability, and patient satisfaction parameters for both destination therapy and bridge-to-transplantation indications [46]. Current LVAD statistics rival 1-year survivorships with heart transplantation in select patient populations, reaching 80 % at 1 year, and 70 % at 2 years among continuous flow pumps, spurring further LVAD use [7•].

In general, patients receive LVAD therapy based on clinical need, with the most common indications including bridge to transplant (BTT), bridge to candidacy, destination therapy (DT), and bridge to recovery. Approximately 80 %–90 % of LVADs are implanted in transplant-eligible candidates (BTT) whose clinical condition has deteriorated either acutely, necessitating circulatory support as a life-saving intervention, or chronically, such that the risks of urgent transplantation are deemed intolerable secondary to poor end-organ function related to impaired native cardiac function [8]. In both scenarios, the implantation of an LVAD provides critical circulatory support with the hopes of restoring end-organ function and thus lowering the risks for urgent transplantation. Destination therapy is generally considered for patients who are ineligible for heart transplantation, either because of age, comorbidities, or personal preference [9]. The goal of destination LVAD therapy is quality and quantity of life without the long-term option of transplantation. The average length of LVAD support for BTT indication ranges from weeks to years and generally months to years for DT indication [10-12].

Miniaturization of pumps has obviated the need for a deformable blood sac and mechanical valves, and most pumps implanted are designated continuous-flow devices (either axial or centrifugal) [13]. Therefore, with codification of disease center-specific designation and destination therapy device indications approved by Centers for Medicare and Medicaid Services, there has been a dramatic increase in device use in the U.S. Although comfort levels for LVAD therapy continue to increase, it is clear that program-based approaches to patient selection, perioperative management, and long-term care will garner the greatest success. Furthermore, with the expansion of DT indication resulting in extended durations of device support, ongoing surveillance strategies beyond clinical assessments must be defined, and transthoracic Doppler echocardiography (TTDE) has been used as a means to manage the long-term patient device experience [14].

Patient Selection for LVAD Therapy

Transthoracic Doppler echocardiography carries a Class I indication for the evaluation of patients with heart failure [15, 16], and a comprehensive Doppler analysis allows the assessment of central hemodynamics with >90 % accuracy [17, 18]. The purpose of preoperative LVAD assessments with TTDE is 2-fold (Table 1):
Table 1

Preoperative transthoracic echocardiography checklist

Parameter

Echocardiographic view

Methods, technique

LV EF and volumes

Apical 4- and 2-chamber

Simpson method

RV function

Apical 4-chamber

Sa, TAPSE, fractional area change

TR quantification

Apical 4-chamber,

 parasternal short-axis

Qualitative and quantitative approach* (color Doppler, CW Doppler)

PAPs

Apical 4-chamber

Modified Bernoulli equation from peak TR velocity (color Doppler, CW Doppler)

AR quantification

Parasternal long-axis,

 apical 5-chamber

Qualitative and quantitative approach* (color Doppler, CW Doppler)

MR quantification

Apical 4- and 2-chamber

Qualitative and quantitative approach* (color Doppler, CW Doppler)

Interatrial shunt

Subcostal,

 apical 4-chamber

Color Doppler, contrast, bubble study

Intracardiac clots

Apical 4- and 2-chamber,

 off-axis

2-dimensional echocardiography, contrast

AR aortic regurgitation, CW continuous wave, EF ejection fraction, LV left ventricular, MR mitral regurgitation, PAPs systolic pulmonary artery pressure, RV right ventricular, Sa tissue Doppler tricuspid annulus peak systolic velocity, TAPSE tricuspid annular plane systolic excursion, TR tricuspid regurgitation

*Reference [67]

  1. 1)

    Patient prognosis and, thereby, potential patient selection; and

     
  2. 2)

    Evaluation of cardiac and great vessel anatomic abnormalities that would change or complicate operative management.

     

Heart failure, as it pertains to LVAD therapy, is the result of a series of temporally acute and chronic cardiovascular injuries sufficient to impair the ability of the heart to adequately meet the circulatory demands of the body. Understanding individual risk is paramount to determining the most suitable patient for LVAD therapy. Therefore, despite the common moniker, heart failure has tremendous heterogeneity in clinical risk [19-21]. Among patients with heart failure with reduced ejection fraction (HFREF), TTDE can segregate those with the absolute highest risks (estimated 6-month mortality approaching 25 % or greater) in whom LVAD therapy would have demonstrable survival benefits. Several comprehensive reviews of TTDE variables, which independently identify patients at high risk of morbidity and mortality, have been published [22-25]. More recently, investigators retrospectively combined both clinical and TTDE variables to develop a heat failure score (0–5) that independently provided prognosis. The authors found that mortality rates were directly related to an increase of the score; a score of 5 had a 4-fold increased risk of death compared with patients in the low-risk group, and predicted a 1-year mortality of 40 %. This would identify a population of advanced HFREF where a consideration for LVAD therapy may be warranted [26••].

We advocate the following Doppler echocardiographic variables to help identify patients at high risk: left ventricular ejection fraction (LVEF) and left ventricular (LV) geometry, LV filling pattern, right ventricular (RV) function, and pulmonary systolic pressure [27, 28] (Table 2). Among these variables, potentially the strongest marker of prognosis in patients with HFREF, is Doppler assessment of the mitral inflow pattern during diastole. The specific pattern, ratio of peak velocities, and measures of time intervals provide accurate estimates of left atrial and pulmonary wedge pressures [29]. Static measurement of mitral inflow in patients with HFREF has confirmed the presence of a restrictive filling pattern predicts worse New York Heart Association (NYHA) functional class and exercise capacity despite similar demographic and other echocardiographic variables [30]. Furthermore, among transplant-eligible candidates with peak oxygen consumption <14 ml/kg/min, the presence of a restrictive filling pattern independently segregated patients with the highest risk of mortality (48 % vs 20 %, restrictive vs not restrictive, respectively) [31]. Finally, the presence of a persistent restrictive filling pattern between 3 and 6 months after targeted medical therapy demonstrated a 35 % annualized mortality [32].
Table 2

Echocardiographic predictors of prognosis in heart failure

Parameter

Reference

Number of patients (mean EF)

Endpoint

Echocardiographic parameters(independent predictors of prognosis)

P-value

LV remodeling

Grayburn PA, et al. J Am Coll Cardiol. 2005 [95]

336

Primary endpoint of death. Secondary endpoints of death, HF hospitalization or transplant

LVEDVI >120 ml/m2

0.0005

Carluccio E, et al. Eur J Heart Fail. 2013 [26••]

747 (<45 %)

Death from any cause

LVESVI ≥ 84 ml/m2

<1.0001

Mitral inflow and diastolic function

Dokainish H, et al. J Am Coll Cardiol. 2005 [96]

116 (<50 %)

Cardiac death or re-hospitalization for CHF

E/Ea >15

0.0001

Doughty RN, et al. Eur J Heart Fail. 2008 [97]

3,540

All-cause mortality

Restrictive filling pattern

<0.0001

Somaratne JB, et al. J Am Soc Echocardiogr. 2009 [98]

887

All-cause death

Pseudonormal filling vs. abnormal or normal relaxation

 

Rossi A, et al. Eur J Heart Fail. 2009 [99]

1,157

Death or hospitalization for worsening HF

LA area >20.5 cm2or >9.7 cm2/m2, DT of <140 ms

<0.0001

Ghio S, et al. Congest Heart Fail. 2012 [29]

601

All-cause mortality median follow-up of 32 mos.

Restrictive filling pattern

 

Mornoş C, et al. Int J Cardiovasc Imaging. 2013 [18]

345 (<35 %)

Cardiac death or readmission due to HF worsening in 35.1 ± 8.7 mos. follow-up

E/(Ea × Sa) >1.6

0.001

Carluccio E, et al. Eur J Heart Fail. 2013 [26••]

747

(<45 %)

Death from any cause

LAVI ≥ 45 ml/m2 DT ≤ 140 ms

<0.0001

0.025

RV

size/function

Ghio S, et al. Am J Cardiol. 2000 [50]

140

(<35 %)

Death from any cause and major cardiovascular events

TAPSE ≤ 14 mm

<0.0001

Meluzin J, et al. Int J Cardiol. 2005 [55]

177

Cardiac-related deaths and nonfatal cardiac events (hospitalizations for heart failure or malignant arrhythmias for a follow-up of 16 mos.

RVTDI Sa <10.8 cm/s RVTDI Ea <8.9 cm/s

<0.001

Meluzin J, et al. J Am Soc Echocardiogr. 2005 [54]

140 (<40 %)

Cardiac-related deaths and nonfatal cardiac events (hospitalizations for heart failure or malignant arrhythmias over a follow-up of 17 mos.

RVTDI Sa <10.8 cm RVTDI Ea <8.9 cm/s Annular acceleration during isovolumic contraction of ≤ 2.52 m/s

<0.001

Damy T, et al. J Card Fail. 2012 [34]

1,547

Death from any cause during a median follow-up of 63 (41–75) mos.

TAPSE <15.9 mm

0.0001

Carluccio E, et al. Eur J Heart Fail. 2013 [26••]

747 (<45 %)

Death from any cause

TAPSE <16 mm

<0.0001

Pulmonary hypertension

Carluccio E, et al. Eur J Heart Fail. 2013 [26••]

747 (<45 %)

Death from any cause

PAPs ≥ 45 mm Hg

<0.001

Merlos P, et al. Eur J Intern Med. 2013 [65]

1,210

1-year all-cause mortality

PAPs >60 mm Hg

0.038

 

Ghio S, et al. Eur J Heart Fail. 2013 [100]

658 (<45 %)

Death, urgent heart transplantation and ventricular fibrillation

PAPs ≥ 40 mm Hg TAPSE ≤ 14 mm

 

CHF congestive heart failure, DT deceleration time of E wave, E pulsed wave Doppler early peak velocity of transmitral filling pattern, Ea tissue Doppler early diastolic peak velocity, EF ejection fraction, HF heart failure, LA left atrial, LAVI left atrial volume index, LV left ventricular, LVEDVI left ventricular end-diastolic volume index, LVESVI left ventricular end-systolic volume index, PAPs systolic pulmonary arterial pressure, RV right ventricular, RVTDI right ventricular tissue Doppler imaging, Sa tissue Doppler systolic peak velocity, TAPSE tricuspid annular plane systolic excursion

Right Ventricular Assessment

Right ventricular evaluation is pivotal in patient selection for LVAD implantation. Right ventricular size and function are independent risk factors among both patients with HFREF [33-35] and among those who receive LVAD therapy [36, 37]. The latter is related to the fact that RV failure post-LVAD implantation contributes to substantial morbidity and mortality [38-40, 41••]. Therefore, the goal of TTDE is to systematic identify RV dysfunction to both assign risk among HFREF patients and selectively identify those individuals who would pose high risk for RV failure post-LVAD implantation. We advocate a stepwise assessment of RV size and function before consideration of LVAD implantation using commonly assessable parameters of RV size and function (Table 3).
Table 3

Echocardiographic predictors of right ventricular failure

Variable

Normal RV function (lower-upper limit)

RV dysfunction

High risk of RV failure*

TAPSE, mm

16–30

≤ 16

<7.5

Pulsed wave Doppler velocity at the annulus, cm/s

10–19

≤ 10

<8

Fractional area change, %

35–63

35–20

<20

Free-wall LS

–9.5–14.9

–9.0 –11.4

<−9.2

IVAT, m/s2

2.2–5.2

---

<2.52

RVEDD-to-LVEDD ratio

0.72 ± 0.15

0.74 ± 0.17

>0.75

RV sphericity index

0.60 ± 0.10

---

0.64 ± 0.09

TR velocity, m/s

---

---

Severe TR >2.5 m/s

IVAT isovolumic acceleration time, LS longitudinal strain, LVEDD left ventricular end-diastolic diameter, RV right ventricular, RVEDD right ventricular end-diastolic diameter, TAPSE tricuspid annular plane systolic excursion, TR tricuspid regurgitation

*Need of postoperative RV assist device or the use of inotropic agents for >14 d

Due to the shape and location of the right ventricle, multiple TTDE views are often employed to obtain RV size [42••]. Among patients with LV dilatation and failure, further displacement and deformation of the normal size RV chamber adds to the technical challenges of visualization. However, once RV chamber dilation occurs (often in late-stage HFREF) there is generally improved visualization by standard apical 4-chamber, parasternal short-axis, and subcostal views. Visual inspection of RV size is often relative to LV chamber size and, therefore, quantification is preferred. Most RV assessments can be achieved from the standard apical 4-chamber views. Slight medial angulation or shift of the transducer and cropping of the sector to include the entire RV can assist quantification algorithms (Fig. 1). Once image quality is satisfactory, RV fractional area change (FAC) is quantified [42••]. This method allows 2-dimensional estimations of RV size and function. A cutoff RVFAC <0.32 predicted a 31.2 % composite death or hospitalization rate among patients with HFREF [43]. Moreover, RV dysfunction expressed as a FAC <26 has been shown as an independent predictor of cardiovascular mortality, total mortality, and HF with each 5 % decrease in RVFAC associated with a 16 % increased odds of cardiovascular mortality [44]. More recent data indicates that RVFAC values ranges between 20 %–30 % are more likely to require LVAD or transplantation among patients with HFREF [45]. However, once RVFAC reaches <20 % the incidence of postoperative RV failure dramatically increases [46]. Concordant data among 2 groups found that once preoperative RV/LV diameter ratio exceeded 0.72 and 0.75, there was an 11.4- and 5-fold increased risk for RV failure, respectively, among patients supported with continuous-flow LVAD [47, 48]. Given the difficulties in imaging the entire RV, other surrogate methods to assess RV function have been investigated. Tricuspid annular plane excursion (TAPSE) can be quantified for nearly all patients from the apical 4-chamber view [49]. Right ventricular dysfunction as expressed by a low TAPSE (<14 mm) was identified as an independent marker of risk among a cohort of heart failure patients and added significant prognostic information to NYHA Class III or IV, LVEF (<20 %) and mitral deceleration time (<125 ms) [50]. These data were later confirmed after 63 months of follow-up in a larger population of patients (>1,000) with both HFREF and heart failure with preserved ejection fraction, in which a low TAPSE (15.9 mm) was identified as independent prognostic marker and predicted a 47 % mortality at 12 months [34]. Among patients who received a continuous-flow LVAD, a depressed TAPSE correlated with RV failure postoperatively, and a 7.5 mm excursion value had 91 % specificity and 46 % sensitivity in defining patients who required prolonged inotropic support or pulmonary vasodilators (>14 days) [36]. Right ventricular contraction pressure index is a novel parameter of RV function and incorporates TAPSE and Doppler estimates of pressure gradients. The relationship is described by the equation: TAPSE x (RV systolic pressure - mean right atrial pressure). The estimates of pressure can be derived from standard Doppler assessments. Right ventricular contraction pressure index correlates well with RV stroke work index and, therefore, provides an independent noninvasive tool to assess RV function [35]. Its utility has not been assessed in patients who receive LVAD support.
https://static-content.springer.com/image/art%3A10.1007%2Fs11886-014-0468-5/MediaObjects/11886_2014_468_Fig1_HTML.gif
Fig. 1

Approach to right ventricular (RV) assessment prior to left ventricular assist device (LVAD) implantation should focus on easily accessible 2-dimensional and Doppler parameters that explore RV size and function. When RV is normal size and tricuspid annular plane systolic excursion (TAPSE) >7.5, then one can proceed to LVAD implantation without further assessments [34]. However, if RV size is designated mildly/moderately or severely enlarged, a RV/LV ratio should be explored [47, 48]. Further exploration of RV function using TASPE among these groups defines populations in whom LVAD implantation can proceed, proceed with caution, and proceed with planned RV support. We advocate understanding and optimizing central hemodynamics prior to LVAD implantation among groups designated “caution” and “plan for RV support” as these patients may have intermediate and high risk of RV failure after LVAD implantation

Right ventricular longitudinal regional and global strains can be obtained with spectral Doppler techniques, but have been associated with greater challenges regarding angle dependence, accuracy, and reproducibility. However, regional strain, specifically RV lateral annular tissue Doppler imaging (TDI) from the apical 4-chamber view, has been validated to add prognostic information among chronic heart failure patients [51, 52•]. A RV TDI systolic wave velocity of <10.8 cm/s and an early RV diastolic TDI velocity <8.9 cm/s predicted cardiac-related death and nonfatal cardiac events (hospitalizations for heart failure or malignant arrhythmias) in several cohorts of patients with HFREF [53-55]. The combination of end-diastolic short-/long-axis ratio ≥ 0.6, tricuspid annulus peak systolic velocity <8 cm/s, and peak systolic longitudinal strain rate <0.6/s carries high predictive value for LVAD postoperative RV failure. In addition, preoperative RVFAC, estimated right atrial pressure and left atrial volume index were able to predict 5-fold relative risk of postimplant RV dysfunction and hemodynamic instability [46]. Ideally, these TTDE parameters should be combined to develop a scoring system that would provide additional risk stratification prior to LVAD implantation.

Due to RV geometry and imaging planes, global RV assessment can be challenging; however, important information regarding RV function after LVAD implantation has been gleaned using novel velocity vector mapping analysis. Grant et al. studied a cohort of 117 patients who underwent continuous-flow LVAD implantation and found that peak RV free wall longitudinal strain of –9.6 % had 76 % specificity and 68 % sensitivity in predicting the need for postoperative RV assist device or the use of inotropic agents for >14 days [56••]. In a multivariate logistic regression analysis including variables from the established Michigan RV risk score, peak strain remained an independent predictor of RV failure. Concordant data from Cameli et al. demonstrated that patients with a continuous-flow LVAD who presented with low preoperative RV free wall longitudinal strain (–9.2 % ± 1.9 %) had a progressive decline of RV strain after LVAD implantation, whereas those with greater values at baseline (–15.5 % ± 3.6 %) increased these measures of RV performance [57].

Pulmonary Artery Hypertension and Tricuspid Regurgitation Estimation

Pulmonary artery hypertension (PAH) is a consequence and complicating feature of end-stage HFREF, dramatically impacting prognosis even after LVAD support [58, 59]. The presence of or the inability to reverse either significant pre-existing PAH or elevations in pulmonary vascular resistance (PVR) identifies patients at high risk for RV failure and perioperative mortality during heart transplantation [60]. Importantly, chronic LVAD therapy has been observed to reverse PVR and, therefore, is an important consequence of the BTT strategy [61, 62]. Therefore, PAH and elevated PVR both identify populations of patients with end-stage HFREF in whom LVAD therapy may be leveraged but can also be associated with perioperative RV failure.

Doppler echocardiographic estimates of pulmonary artery pressure correlate well with invasive measures and provide independent prognostic information among patients with HFREF [58, 63-65]. Among 108 consecutive patients with symptomatic systolic heart failure, a stepwise logistic regression model identified the presence of a maximal tricuspid regurgitant jet velocity >2.5 m/s as the most important prognostic variable for the identification of death and hospitalization over a 28-month follow-up period [58].

Although grade or severity of tricuspid regurgitation is directly related with PAH, it can independently influence outcome after LAD implantation. For example, patients with severe, compared with mild, tricuspid regurgitation have required longer post-LVAD implant intravenous inotropic support, temporary RV assist device, and a longer hospital stay [66]. We advocate standard techniques of grading tricuspid regurgitation [67], and patients with greater than moderate tricuspid regurgitation undergo surgical correction with either tricuspid valve annuloplasty or replacement. Finally, since multiple factors influence RV performance, it is not surprising that in a multivariable logistic regression analysis of 218 patients who underwent LVAD therapy, the presence of central venous pressure >15 mm Hg, severe RV dysfunction, severe tricuspid regurgitation, preoperative intubation, and heart rate >100 were major criteria predictive of postoperative biventricular support [68].

Valvular Heart Diseases Assessment

Determining the severity and significance of valvular heart disease among LVAD-eligible patients is an important aspect of optimal long-term patient outcome. Specific lesions that may require surgical correction include aortic regurgitation, mitral regurgitation, mitral stenosis (either native or prosthetic), and tricuspid regurgitation. Standard methods to estimate severity with hemodynamic caveats associated with severe HFREF must be performed [67, 69-71]. At times, transesophageal echocardiography (TEE) or invasive hemodynamic assessments may be required to optimally assess valve physiology [72, 73]. The significance of valvular heart disease must be understood in the context LVAD physiology, which is anticipated to provide profound hemodynamic LV unloading, increased mean aortic pressure, and augmented venous return to the right heart. Therefore, even moderate degrees of valvular heart disease can be accentuated and, thus, may require corrective surgical intervention at the time of LVAD implantation. The decision to correct specific valvular lesions must be balanced with LVAD indication (BTT or DT), anticipated duration of support and the specific added surgical risks. In a cohort of 281 patients who underwent implantation of a HeartMate II (Thoratec Corp., Pleasanton, CA) as BTT, an overall 30-day mortality of 5.8 % was demonstrated in the group of patients who received LVAD therapy alone vs an overall 30-day mortality of 11.3 % in the group of patients who underwent concurrent cardiac procedures in conjunction with LVAD implantation. Subgroup analysis demonstrated that simultaneous patent foramen ovale closure was not associated with an increased 30-day mortality rate, but concurrent valvular procedures increased the risk to 8.5 % (30-day mortality rate for aortic valve procedure was 25 %, concurrent mitral valve procedure, 0 %, and tricuspid repair, 3.3 %) [74].

Aortic Valve

Aortic valve stenosis generally poses little significance to LVAD surgery or post-LVAD implantation patient care. However, hemodynamic changes associated with LVAD therapy (ie, decreased LV pressure and increased mean aortic pressure) are expected to worsen aortic insufficiency acutely, and changes to aortic valve structure noted after long-term continuous-flow support further compromises aortic valve function and result in even greater degrees of regurgitation [14]. In extreme circumstances, this pathologic spiral can lead to closed-loop circulation between the aorta and the LVAD, compromising end-organ perfusion [75]. Several investigators have advocated repair, replacement or closure (over-sewing) of the aortic valve at the time of LVAD implantation when moderate aortic insufficiency is present [76-78]. Percutaneous transcatheter closure of the aortic valve has been performed in select centers to treat LVAD-associated aortic insufficiency in patients who are poor candidates for repeat operation, however, the safety and durability of percutaneous options will need larger and longer-term cohorts [79, 80].

Special Anatomic Considerations for LVAD Patients

The aforementioned physiologic changes associated with continuous-flow LVADs (ie, low LV pressure throughout systole and diastole) predispose to worsening of right to left shunting and, thus, investigation of intracardiac communication at any chamber should be a mainstay of all preoperative patient evaluations. Patent foramen ovale is estimated to affect approximately 25 % of the general population, true atrial septal defects <1 %, and acquired ventricular septal defects (ie, postmyocardial infarction) <0.1 %. Frequent screening for these lesions can be accomplished with TTDE using agitate saline from peripheral or central venous access [81]. Irrespective of the communication, special attention should be dedicated to identifying the significance of the lesion in the presence of low left-sided cardiac pressures. This may be best accomplished by intraoperative TEE, when the adjustments of anesthesia and patient afterload optimize left-sided pressures, for which the use of agitate saline can identify both the presence and location [82-84].

Perioperative Echocardiography Evaluation

Early perioperative echocardiographic evaluation, generally by TEE, should focus on optimization of hemodynamic parameters, reassessment of valvular lesions and intracardiac shunts (preferably before sternotomy has been repaired), and documenting presence of pericardial effusion. The position of the apical cannula from the 4- and 2-chamber midesophageal long-axis views by TEE should be documented, and, while locations may vary, optimal cannula position should be directed perpendicular to the plane of the mitral annulus and toward the P2 segment of the mitral valve. An inflow cannula directed toward the septum, posterior wall, or adjacent to a prominent papillary muscle should be avoided as these positions can result in varying degrees of inflow obstruction and ineffective LV unloading. Although surgical repositioning can be challenging, we advocate TEE-guided management to ensure proper cannula positioning prior to separation from cardiopulmonary bypass or repair of sternotomy. Subsequent echocardiographic evaluations will be dictated by clinical status, but we advocate at least 1 TTDE shortly after central hemodynamic access has been withdrawn and the patient is initiating ambulation. Ideally, pump adjustments can be made at the time of TTDE; the following concepts should be kept in mind:
  1. 1.

    Unloading of central hemodynamics should be the initial goals of therapy. If alterations in pump speed do not result in requisite change in direct or surrogate measures of hemodynamics then pump malfunction should be suspected.

     
  2. 2.

    Maximization of blood flow through LVAD. Most devices have an optimal range (RPMs) for operation, for which maximal pump washout and minimal stagnation of blood in either the inflow or outflow conduits can be achieved. In this regard, predominant closure of the aortic valve throughout the cardiac cycle and minimization of mitral regurgitation should be associated with maximal blood flow through the pump.

     
  3. 3.

    Continuous-flow LVADs are preload dependent. Many devices have algorithms that rely on intrapump pressure and power use to trigger down titration of RPMs to prevent low preload (suction) events. In this regard, the relative size of the LV chamber and location of the septum can influence the changes in pump speed.

     
Figure 2 and Movies 16 are examples of the influence of pump speed on various Doppler echocardiographic parameters in the same patient during various clinical settings. Although pump speeds are not usually changed during outpatient clinical care, we advocate periodic (at 3 and 6 months) TTDE evaluation. Alternatively, we have employed a physician-directed handheld echocardiographic approach to outpatient LVAD management using Vscan devices (GE Healthcare, Waukesha, WI). Although these devices do not provide continuous, pulse wave, or tissue Doppler techniques nor have robust analytical tools, they are good for detecting several specific features of LVAD supported hearts, including relative size of the LV, location of the septum, aortic valve opening, positioning of the apical cannula, size of the right ventricle, and presence of pericardial effusion. The color Doppler feature can be used to screen for aortic, mitral, and tricuspid valve regurgitation. We have used these devices to adjust pump speeds during clinical evaluation, and changes to cardiac parameters typically are observed within 5–7 minutes.
https://static-content.springer.com/image/art%3A10.1007%2Fs11886-014-0468-5/MediaObjects/11886_2014_468_Fig2_HTML.gif
Fig. 2

Approach to left ventricular assist device (LVAD) interrogation, adjustment, and optimization using 2-dimensional and Doppler echocardiographic parameters. Three separate scenarios defining pump speeds that are either “too high,” “just right” and “too low” are demonstrated, and recommendation to pump speed are based on these parameters, recognizing that pump adjustment should always be made in the context of patient clinical status. The echocardiographic features of adequate LVAD support include: decompression of the left ventricle, intermittent or midsystolic septal shift, aortic valve opening intermittently or not at all, and low Doppler estimates of left atrial pressure (LAP) and pulmonary artery systolic pressure (PASP). Left ventricular decompression can be estimated by measures of the left ventricular diameter at end diastole (LVIDd) from the parasternal long-axis view and compared with prior or baseline studies; often decreases in LVIDd >1 cm are sufficient. Septal shift reflects relative pressure in the left and right ventricles and can be used to gauge over- or under-decompression of the left ventricle. Intermittent movement of the septum leftward timed with early midsystole indicates sufficient LV unloading without deforming right ventricular (RV) geometry and predisposing to RV suction cascade (see text). Aortic valve opening intermittently requires that at least 5–10 cardiac cycles be recorded; the aortic valve opening with every cycle may indicate low pump speeds, LV recovery or LVAD malfunction, and a ramp study should be performed. Finally, standard Doppler measures of mitral inflow and tricuspid regurgitation velocity can be used to noninvasively estimate LAP and PASP; ideally these should reflect adequate unloading of central hemodynamics

Often the clinical scenario or technical artifacts generated by the LVAD or cannula require dedicated TTDE or TEE assessments of LVAD-supported patients. A systematic focused protocol to acquire datasets should be followed when imaging LVAD patients (Table 4). We advocate documenting device type and pump speed along with clinical concerns before starting echocardiographic imaging. If necessary, repeat measurements after LVAD speed adjustments (such as ramp speed protocol) can be used to document appropriate LVAD function or for the optimization of support.
Table 4

Postoperative transthoracic echocardiography checklist

Parameter

Echocardiographic view

Methods, technique

LV EF and volumes

Apical 4- and 2-chamber,

 parasternal long- and/or short-axis

Modified Simpson Biplane (if possible), Quinones method, fractional shortening, FAC, AV opening status and apical inflow cannula velocity variation index (systolic-to-diastolic velocity ratio at the level of inflow cannula)

LVAD output

Right parasternal or parasternal short-axis (RVOT)

Outflow graft cross-sectional area × the outflow graft TVI or RVOT cross-sectional area × RVOT TVI if AV does not open

LV unloading

Apical 4-chamber

Interventricular and interatrial septum shift (neutral or slight leftward = adequate LV unloading; rightward = inadequate LV unloading; extreme leftward = excessive LV unloading

RV function

Apical 4-chamber

Sa, TAPSE, FAC

TR quantification

Apical 4-chamber,

 parasternal short-axis

Qualitative and quantitative approach* (color Doppler, CW Doppler)

PAPs

Apical 4-chamber

Modified Bernoulli equation from TR velocity peak (color Doppler, CW Doppler)

AR quantification

Parasternal long axis,

 apical 5-chamber

Qualitative and quantitative approach* (color Doppler, CW Doppler) and outflow graft cross-sectional area × the outflow graft TVI – systemic stroke volume (RVOT cross-sectional area × RVOT TVI)

MR quantification

Apical 4- and 2-chamber

Qualitative and quantitative approach* (color Doppler, CW Doppler)

Interatrial shunt

Subcostal,

 apical 4-chamber

Color Doppler, contrast, bubble study

Intracardiac clots

Apical 4- and 2-chamber and off-axis

2-dimensional echocardiography and contrast

Outflow cannula velocity peak and pulsatility, 0.5–2.0 m/s

High left parasternal,

 right parasternal

Color Doppler, CW Doppler

Inflow cannula velocity peak and pulsatility, 0.7–2.0 m/s

Apical 4-chamber

Color Doppler, CW Doppler

Pericardial space

Subcostal

2-dimensional echocardiography, M-mode

AR aortic regurgitation, AV aortic valve, CW continuous wave, EF ejection fraction, FAC fractional area change, LV left ventricular, LVAD left ventricular assist device, MR mitral regurgitation, PAPs systolic pulmonary artery pressure, RV right ventricular, RVOT right ventricular outflow tract, Sa Tissue Doppler tricuspid annulus peak systolic velocity, TAPSE tricuspid annular plane systolic excursion, TR tricuspid regurgitation, TVI time-velocity integral

*Reference [67]

Left Ventricular Morphology and Function

Traditional LV function measurements, such as LVEF, are artificial in LVAD-supported patients as preload and afterload should both be low and may only be relevant when assessing patients for consideration of weaning LVAD therapy for possible explantation [85, 86]. Traditional LVEF, volumes or mass calculations by various methods may be challenging from apical windows due to cannula and/or device artifact [52•]. Surrogates of LV function have been used to assess for native LV function, including the frequency and extent of aortic valve opening (Fig. 3) by M-mode Doppler and inflow cannula velocity variation. Since the LVAD and the native heart are in series, the intrinsic contractility of the LV will directly influence inflow maximal velocity at peak systole [87•]. The inflow velocity variation (difference between the systolic and diastolic apical velocity) has been related to LVEF. Importance of LV morphology should be emphasized after LVAD support and is generally a good reflection of hemodynamic unloading both acute and chronically [88•]. Specific attention to parasternal long-axis measures of left ventricle internal diameter during diastole (LVIDd) and the position of the interventricular septum throughout the cardiac cycle should be ascertained. Adequate unloading by these parameters is reflected by a reduction then stabilization of LVIDd. Since the interventricular septum is integral to RV function, a neutral or slight leftward shift during peak systole indicates adequate LV decompression [89, 90]. Leftward septal shift should be avoided as this may result in impaired RV performance, worsening of tricuspid regurgitation through tethering of the septal leaflet, and predispose to clinical right heart failure [91]. Similarly, a predominant rightward shift of the septum is a marker of inadequate LV unloading, raising concerns for LVAD malfunction.
https://static-content.springer.com/image/art%3A10.1007%2Fs11886-014-0468-5/MediaObjects/11886_2014_468_Fig3_HTML.gif
Fig. 3

Aortic valve opening status. LVAD left ventricular assist device

Right Ventricular Assessment After LVAD Implantation

Achievement of LV unloading should be balanced with RV function since augmenting pump speeds can increase venous return and potentially worsen RV performance. This phenomenon is most challenging in the immediate postoperative setting where up to a third of patients have varying degrees of RV dysfunction, the most extreme cases leading to major causes of postoperative morbidity and mortality [92•]. Therefore, the echocardiographic methods used to assess RV function after LVAD implantation should follow a similar protocol proposed for preoperative evaluations [52•]. The focus should be on RV morphology, function, and severity of tricuspid regurgitation. Hemodynamic data should include estimates of RV systolic pressure (peak tricuspid regurgitation velocity) and mean right atrial pressure (inferior vena cava interrogation) [52•]. The advantage of TTDE is the noninvasive serial assessment after LVAD speed adjustments are made. One specific scenario involving the interplay of LV unloading and RV function is worthy of mention. The “suction cascade” has been described with excessive unloading of left heart chambers and is the result of a complex biphasic effect on RV function. Initially, LV decompression and reduction of pulmonary artery pressure reduces RV afterload and improves function as measured by cardiac output. Unfortunately, as venous return is augmented by improved cardiac output RV dilation and tricuspid regurgitation may ensue, thus reducing effective right to left flow. Compounding effects occur if the interventricular septum shifts leftward, which worsens tricuspid regurgitation and reduces RV function [89]. The overall effect is reduced LVAD flow, suction event alarms, elevated right atrial pressure, and systemic end-organ hypoperfusion. Transthoracic Doppler echocardiography plays a critical role in the early diagnosis and management of this clinical scenario. Paradoxically, reducing LVAD pump speeds reverses many of the structural and functional changes that occur with the “suction cascade” and TTDE can be used to follow these changes in real time.

Inflow Cannula

The apical inflow cannula is usually positioned lateral to the left anterior descending coronary artery from an epicardial location and oriented anteriorly to posteriorly within the left ventricle toward the P2 segment of the mitral valve. The orientation can be confirmed with 4- and 2-chamber views. This ideal location also allows interrogation of the inflow cannula with color flow and Doppler techniques from the apical views. Variations in cannula position may require off-axis views for interrogation; attempts should be made to align pulsed and continuous wave Doppler signals parallel with the inflow cannula to document laminar, unidirectional, and low-peak velocity flows with no regurgitation. In axial-flow LVADs, such as the HeartMate II (Thoratec Corp.), peak velocity in the inflow cannula ranges from 0.7 to 2.0 m/s, according to preload and intrinsic residual function of the native heart. Generally, a pulsatile inflow pattern is characterized by phasic changes in flow throughout the cardiac cycle, reaching a maximum during systole and minimum during diastole [92•]. Abnormally high velocities and abbreviated or spiked systolic flows are characteristics of inflow cannula obstruction and can be caused by tissue ingrowth, pump migration with reorientation of the inflow or clot [93]. Malalignment of Doppler flow will underestimate peak velocities and may erroneously reassure clinicians that inflow obstruction is not present. Intermittent suction events with obstruction from the LV walls can be documented with TTDE [92•]. Regurgitant flow is distinctly abnormal and suggests major pump malfunction (pump stoppage or no flow).

Outflow Cannula

The outflow graft exits the distal end of the LVAD (often below and to the right of the lower sternum) and traverses right parasternally and is end-to-side anastomosed along the right anterolateral aspect of the aorta, usually 5–7 cm above the sinus of Valsalva. Optimal visualization of the outflow cannula by TTDE is technically challenging and, if clinically important, may require TEE. Generally, transthoracic echocardiography in the high left parasternal long-axis and right parasternal off-axis views, with the patient lying on his right side, shows the long-axis of the aortic outflow cannula. Color flow, pulse wave, and continuous wave Doppler are used to evaluate flow patterns. Flow velocity in the outflow graft should be recorded with the pulsed wave Doppler sample volume at least 1 cm proximal to the aortic anastomosis [52•]. In axial-flow pumps, the flow in the outflow graft is unidirectional with a peak velocity usually ranging from 0.5 to 2.0 m/s and slightly pulsatile, according to LVAD output, speed (RPM), and angle of insertion into the native aorta [94].

Hemodynamic Instability in Early Postoperative Days

Acute hemodynamic decompensation may occur in early postoperative days, and most times it is caused by hypovolemia (from bleeding), acute RV dysfunction, cardiac tamponade, pulmonary embolism, or LVAD thrombosis [93]. “Suction cascade” is the most typical expression of acute RV dysfunction. It must be suspected whenever echocardiographic examination shows a dilated and hypocontractile RV, significant functional tricuspid regurgitation, small left ventricle, and an intermittent inflow cannula obstruction by the collapsed LV wall. A differential diagnosis of pulmonary embolism should be considered as it also can cause acute RV dysfunction and is associated with greater right-sided pressures.

Cardiac tamponade is sometimes difficult to diagnose since the typical approach for assessing interventricular independence is not possible and blood collections may be loculated and confined to a small area, compressing a particular chamber [52•]. LVAD dysfunction or thrombosis should be suspected when a combination of clinical (cardiogenic shock, low LVAD output), echocardiographic (rightward deviation of the interventricular and interatrial septum, significant functional mitral regurgitation, aortic valve opening every cardiac cycle), and laboratory clues (intravascular hemolysis) are detected (Fig. 4).
https://static-content.springer.com/image/art%3A10.1007%2Fs11886-014-0468-5/MediaObjects/11886_2014_468_Fig4_HTML.gif
Fig. 4

Concomitant conditions in left ventricular assist device (LVAD) dysfunction. LDH lactate dehydrogenase

Moreover, if impaired LVAD output results in the diastolic aortic pressure exceeding the LV diastolic pressure, reverse flow from the ascending aorta through the outflow and inflow cannulas can be observed [88•].

Conclusions

Chronic LVAD therapy is one of the few clinically proven options for patients with advanced heart failure. Current LVAD statistics rival 1-year survivorships with heart transplantation in select patient populations, reaching 80 % at 1 year and 70 % at 2 years among continuous-flow pumps, and spurring further LVAD use [7••]. Echocardiography (TTDE and TEE) is an integral aspect of the evaluation, selection, and management of patients prior to, during and after LVAD therapy. Modern comprehensive digital-based echocardiography laboratories with standardized reporting systems that are field-specific database elements are capable of generating queries based on age, ejection fraction, and hemodynamic elements, thus potentially allowing rapid population screening. An emphasis on high-risk echocardiographic Doppler markers can assist in optimal patient selection, and specific attention to valvular lesions and intracardiac shunts, in the context of LVAD physiology, help to identify surgical approach and risk. We advocate a comprehensive evaluation of the right heart, which often incorporates invasive imaging modalities. Specific TTDE, both qualitative and quantitative, parameters have been established to assist in defining risk of postoperative RV failure. We advocate patients at high risk for RV failure by TTDE undergo optimization and reassessment prior to LVAD implantation, although prospective evaluation of this approach has not been established. Finally, the role of serial assessments of LVAD patients with TTDE (even with simple handheld devices) to assist in the identification and management of complications are evolving. We advocate programs that incorporate a robust and integrated echocardiographic Doppler approach in the management of patients with LVADs, with the goals of individualizing LVAD therapy to patient-centric outcomes.

Acknowledgments

The authors gratefully acknowledge Joe Grundle and Katie Klein of Aurora Cardiovascular Services for the editorial preparation of the manuscript and Brian J. Miller and Brian Schurrer of Aurora Sinai Medical Center for help with the figures.

Compliance with Ethics Guidelines

Conflict of Interest

Maria Chiara Todaro declares that she has no conflict of interest. Bijoy K. Khandheria declares that he has no conflict of interest. Timothy E. Paterick declares that he has no conflict of interest. Matt M. Umland declares that he has no conflict of interest. Vinay Thohan declares that he has no conflict of interest.

Human and Animal Rights and Informed Consent

This article does not contain any studies with human or animal subjects performed by any of the authors.

Supplementary material

11886_2014_468_MOESM1_ESM.asf (1.1 mb)
Movie 1Color Doppler parasternal long-axis view at 9,400 rpm of a left ventricular assist device-supported patient demonstrating features of pump speeds that are too high: left ventricle (LV) completely decompressed, interventricular septum shifted toward the LV, aortic valve not opening, flow acceleration mid-LV cavity, and no or trace mitral regurgitation. (ASF 1141 kb)
11886_2014_468_MOESM2_ESM.asf (919 kb)
Movie 2Color Doppler apical 4-chamber view at 9,400 rpm of a left ventricular assist device-supported patient demonstrating features of pump speeds that are too high: left ventricle (LV) completely decompressed, interventricular septum shifted toward the LV, aortic valve not opening, flow acceleration mid-LV cavity, and no or trace mitral regurgitation. (ASF 919 kb)
11886_2014_468_MOESM3_ESM.asf (613 kb)
Movie 3Color Doppler parasternal long-axis view at 9,200 rpm of a left ventricular assist device-supported patient demonstrating features of pump speeds that are adequately supported: left ventricle (LV) decompressed, interventricular septum shifts intermittently toward the LV, aortic valve periodically opening, no flow acceleration mid-LV cavity, and mild mitral regurgitation. (ASF 612 kb)
11886_2014_468_MOESM4_ESM.asf (575 kb)
Movie 4Color Doppler apical 4-chamber view at 9,200 rpm of a left ventricular assist device-supported patient demonstrating features of pump speeds that are adequately supported: left ventricle (LV) decompressed, interventricular septum shifts intermittently toward the LV, aortic valve periodically opening, no flow acceleration mid-LV cavity, and mild mitral regurgitation. (ASF 575 kb)
11886_2014_468_MOESM5_ESM.asf (957 kb)
Movie 5Color Doppler parasternal long-axis view at 8,600 rpm of a left ventricular assist device-supported patient demonstrating features of pump speeds that are too low: left ventricle (LV) not decompressed, interventricular septum shifts toward the right ventricle, aortic valve periodically opening, no flow acceleration mid-LV cavity, and worsening mitral regurgitation. (ASF 956 kb)
11886_2014_468_MOESM6_ESM.asf (847 kb)
Movie 6Color Doppler apical 4-chamber view at 8,600 rpm of a left ventricular assist device-supported patient demonstrating features of pump speeds that are too low: left ventricle (LV) not decompressed, interventricular septum shifts toward the right ventricle, aortic valve periodically opening, no flow acceleration mid-LV cavity, and worsening mitral regurgitation. (ASF 847 kb)

Copyright information

© Springer Science+Business Media New York 2014