Chapter 66. Long-Term Mechanical Circulatory Support

Heart failure (HF) remains one of the most common causes of death in the United States. Preventive measures and treatment options for HF have, however, improved significantly. The 2005 ACC/AHA guidelines introduced a comprehensive and systematic method of assessing and managing patients with risk factors and those with HF.1 The suggested staging system is ambitiously inclusive, covering a wide range of patients with cardiovascular problems. It categorizes patients into Stages A (at high risk for HF but without structural heart disease or symptoms of HF) to D (refractory HF requiring specialized interventions). Surgical interventions play a significant role at each stage, and their role becomes more important as the stage advances. The surgical options described in the guidelines include, but are not limited to, coronary artery bypass grafting (CABG), valve replacement or repair, ventricular restoration surgery, heart transplantation, permanent mechanical support, and experimental surgery. A wide range of cardiac surgical procedures are covered, implying that HF is becoming one of the most important fields for cardiac surgeons.

Among the available surgical interventions for HF, mechanical circulatory support (MCS) is distinct from the others in that it, at least partially, replaces the pump function of the failing heart, similar to heart transplantation. Unlike heart transplantation, however, it does not rely on a human donor supply, which is a well-known limiting factor. As technologic advances have accrued, MCS has evolved into a reliable and well-described treatment option.

MCS may include intra-aortic balloon pump (IABP), ventricular assist device (VAD), total artificial heart (TAH), cardiopulmonary support (CPS) such as venoarterial (VA), extracorporeal membrane oxygenation (ECMO), and additional circulatory support systems. In this chapter we will focus on VAD, TAH, and CPS. Among these, VAD therapy as a method of ventricular support and TAH as a replacement of the heart are gaining particular attention. In this chapter, VAD therapy will be discussed in detail in relationship to other technologies.

Mechanical support of the cardiopulmonary system was first clinically introduced by John Gibbon in 1953, when he first successfully employed cardiopulmonary bypass for the repair of an atrial septal defect.2 In 1963, DeBakey implanted the first VAD in a patient who suffered a cardiac arrest following aortic valve replacement. The patient subsequently died on postoperative day 4. In 1966, DeBakey reported the first successful bridge to recovery with implantation of a pneumatically driven VAD in a patient suffering from postcardiotomy shock. The patient was supported for 10 days, and ultimately survived to discharge.3 Soon thereafter, Cooley reported the first successful bridge to transplantation (BTT) using a pneumatically driven implantable artificial heart.4 During this time, the National Heart, Lung and Blood Institute (NHLBI) began funding initiatives to further the development of VADs and a TAH. DeVries and colleagues reported the successful implantation of the Jarvik-7-100 TAH in 1984.5 Despite initial success, a high incidence of thromboembolic and infectious complications led to a moratorium on the use of the TAH in 1991. During this period, however, continued advances in the development of left ventricular assist devices (LVADs) were made. Early survival data of patients undergoing LVAD insertion revealed greater than 30% hospital mortality. Advancements in perioperative care and technology have led to improvements in outcomes,6 and this culminated in FDA approval of an LVAD as a BTT in 1994.

The feasibility of a mechanical-based approach to the treatment of end-stage HF was validated by the Randomized Evaluation of Mechanical Assistance for the Treatment of Congestive Heart Failure (REMATCH) trial in 2001.7 This landmark prospective randomized trial demonstrated a marked survival benefit in patients receiving an LVAD for the treatment of end-stage HF when compared with medical management alone. At the same time, the limitations of device technology using a volume displacement pump were highlighted by the high incidence of adverse events related to mechanical support, such as infection, device failure, and thromboembolic events.

Newer-generation VADs have been developed to overcome these limitations. These pumps are characterized by continuous flow driven by a rotary pump. A recently completed randomized trial demonstrated that a rotary pump (HeartMate II) compared with a volume displacement pump (HeartMate XVE) significantly improved the probability of survival free from stroke and device failure at 2 years.8

These continuous-flow second- and third-generation pumps generate low-grade pulsatility as opposed to the first-generation “pulsatile” pumps with volume displacement. With the growing number of patients receiving mechanical circulatory support device (MCSD) therapy, a national registry, the Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS), was initiated in 2005. This registry was devised as a joint effort of the NHLBI, the Centers for Medicare & Medicaid Services (CMS), the FDA, clinicians, scientists, and industry representatives in conjunction with the University of Alabama at Birmingham and the United Network for Organ Sharing (UNOS). This registry has and will continue to help improve the clinical outcomes of MCSD therapy, expedite new device clinical trials, and promote research.

VA ECMO

Oxygenation of the blood using a membrane oxygenator can be achieved either with a venovenous (VV) or VA circuit. VV ECMO is used for respiratory failure relying on native heart function to move oxygenated blood through the systemic circulation. In order to support a failing heart, VA ECMO is necessary, and is established using a venous cannula(s) for inflow and an arterial cannula for outflow with a perfusion pump and an oxygenator bypassing the heart and lungs. These basic mechanisms are essentially the same as the cardiopulmonary bypass machine. The cannulation sites can be either central or peripheral depending on the situation. The ease of establishing VA ECMO makes this device suitable for emergent circumstances but only for short-term support (days). This device is also widely used in the pediatric population with success. Patients on VA ECMO need strict anticoagulation and in general have to be kept on a ventilator.

VAD

VADs are classified as short-term (or nonimplantable) and long-term (or implantable) devices from the duration of the support perspective, and univentricular (right or more commonly left ventricular) and biventricular assist devices from the configuration of the assist perspective. From the mechanism of driving pump system point of view, they are also divided into volume displacement (or pulsatile) devices and rotary (continuous-flow) devices. An appropriate device is chosen based on the individual patient's clinical status and the surgeon's comfort. The majority of these devices are placed surgically although recent advances have enabled percutaneous placement of certain devices, such as the Impella and TandemHeart.9–11 The major limitations of percutaneous VADs as of 2009 are their somewhat limited flow capacity (up to approximately 5 L/min).

TAH

TAH is discussed elsewhere in this book, but in short, it replaces the patient's own heart in the orthotopic position and provides biventricular support. It has an advantage over the VAD in that conditions of the native heart, such as aortic insufficiency (AI), tricuspid regurgitation (TR), patent foramen ovale (PFO), mechanical valves, and cardiac tumors, have no influence on the function and maintenance of a TAH once it is implanted. The major drawbacks are the complexity of the implant procedure, and its size (body surface area [BSA] >1.7m2 is required for the CardioWest TAH).

Currently used MCSDs are listed in Table 66-1, with some of their features described below.

First-Generation VADs

The HeartMate XVE (Fig. 66-1A) is the only FDA-approved device for destination therapy (DT) use. This device generates pulsatile flow through a pusher plate placed in a relatively large housing, which precludes implantation in small patients (BSA <1.5 kg/m2). The unique textured inner surface allows patient management with no anticoagulation. It is used worldwide and has been tested extensively in large clinical trials, demonstrating superior outcome to medical management.7 However, its long-term use is limited by high probability of device malfunction and infection. The Thoratec PVAD and Berlin Heart Excor enable implantation in small patients including the pediatric population. They can also provide BiVAD support. A major disadvantage is the exteriorized pump.

Second-Generation VADs

This group of VADs is currently gaining popularity. Miniaturization of mechanical support technology using a rotary pump of axial-flow design contributed to limiting the surface area exposed to blood allowed for the elimination of valves, air vents, and compliance chambers. The presence of fewer moving parts reduced device malfunction rates. Initial concerns over the potential influence of the continuous-flow feature on patient physiology appear to have been assuaged based on midterm outcomes. The HeartMate II (Fig. 66-1B) was shown in a randomized control trial to be superior to the HeartMate XVE.8 Clinical experience has been accumulated with the MicroMed-DeBakey VAD and the Jarvik 2000 as well.12,13 A unique power delivery system based on cochlear implant technology is adopted in the Jarvik 2000.13 A titanium pedestal is screwed into the temporal bone behind the ear and conveys the electrical system through highly vascular scalp skin (Fig. 66-2). Device infection might be reduced with this technology because of immobility and the absence of subcutaneous fat.

Third-Generation VADs

The main feature of the third-generation VADs is that they eliminated bearings by using either hydrodynamic or electromagnetic suspension of an impeller to reduce mechanical wear and trauma to blood cells. The DuraHeart (Fig. 66-3) is a centrifugal pump with a magnetically levitated impeller that is composed of a magnetic bearing, an impeller, a housing, and a direct current brushless motor. The pump provides contact-free rotation of the impeller without any material wear, which predicts a high level of durability. The pump weighs 540 g and has a diameter of 72 mm and a height of 45 mm. The HeartWare (Fig. 66-4) is another example of a centrifugal pump that sits within the pericardial space. The left ventricular (LV) apical inflow cannula is an integral part of the pump itself. The device weighs only 145 g and yet provides up to 10 L/min of flow. These and other devices are currently undergoing clinical trials.

When considering MCSD placement, the cardiac surgeon and the team must be clear of their goal. This is particularly important in order for the surgeon and the team to be able to select the appropriate patient, the appropriate timing for placement, and the appropriate device to use. Five major goals are: “BTT,” “DT,” “bridge to recovery,” “bridge to decision,” and “bridge to bridge” (Fig. 66-5).

BTT

Patients who are evaluated by a multidisciplinary transplant team and are determined to be suitable candidates for heart transplantation may require MCSD placement while on the waiting list. With progressive lengthening of waiting times on transplantation lists, the use of a VAD as a BTT has become standard therapy in end-stage HF. This decision is made once it is felt that the patient is unlikely to survive to transplantation or if he/she is developing additional organ dysfunction which may compromise transplant eligibility. Currently, there are no widely accepted guidelines as to the indication and timing for VAD placement as a BTT, and therefore the decision has to be individualized by weighing the balance among three main factors: the estimated waiting time for a heart, the estimated chance of death on the waiting list, and the risk of the operation. For example, patients with blood type O have much longer waiting times on the transplant list compared with others in the United States, and therefore the threshold for device placement is lowered.

MCSD placement may also be considered for patients with borderline eligibility for heart transplantation, such as those with an elevated pulmonary vascular resistance (PVR) or moderate end-organ dysfunction. LVAD support was demonstrated to decrease the PVR of patients who were otherwise ineligible for transplantation with successful posttransplant outcomes.14–19 This indication is known as “bridge to eligibility.”

DT

Patients with end-stage HF who are not eligible for heart transplantation may benefit from MCSD placement because of prolongation and/or improvement in quality of life. They are expected to receive circulatory support from the MCSD for life and this indication is called “DT.” The REMATCH trial demonstrated that LVAD therapy is superior to medical management for this category of patients.7 Based on this result the FDA approved the use of an LVAD (HeartMate VE) as DT. Currently accepted indications refer to the eligibility criteria for the REMATCH trial. As a reference, the current CMS criteria20 are: patients who have chronic end-stage HF (New York Heart Association [NYHA] Class 4 end-stage LV failure for at least 90 days with a life expectancy of less than 2 years) are not candidates for heart transplantation and meet all of the following conditions:

1. The patient's Class 4 HF symptoms have failed to respond to optimal medical management, including dietary salt restriction, diuretics, digitalis, beta-blockers, and angiotensin-converting enzyme (ACE) inhibitors (if tolerated) for at least 60 of the last 90 days.

2. The patient has an LV ejection fraction (EF) <25%.

3. The patient has demonstrated functional limitation with a peak oxygen consumption of <12 mL/kg/min, or the patient has a continued need for intravenous inotropic therapy owing to symptomatic hypotension, decreasing renal function, or worsening pulmonary congestion.

4. The patient has the appropriate body size (BSA ≥1.5m2) to support the VAD implantation.

Of importance, one-third of BTT patients lost their transplant candidacy and were delisted with outcomes parallel to those of DT, and as many as 17% of DT patients eventually underwent heart transplantation.21,22

Bridge to Recovery

With an MCSD unloading the ventricle, some patients demonstrate recovery of myocardial function followed by successful MCSD explantation. MCSD placement for this purpose is called bridge to recovery. Only a few patients who received LVAD support demonstrated shifts in ventricular and myocardial properties back toward normal (myocardial “reverse remodeling”) and successfully underwent explantation of the device with sustained cardiac function. The process of ventricular “remodeling,” which is characterized by progressive ventricular dilatation and dysfunction, is mediated by many factors on molecular, biochemical, metabolic, cellular, extracellular matrix, and ventricular structural levels.23 After the introduction of the LVAD, it was appreciated that many of these abnormalities could be reversed, at least to some degree.

At this point it is rather difficult to predict which patient will experience recovery of cardiac function with MCSD placement. MCSD placement may be undertaken for another indication with subsequent recovery of myocardial function. However, there have been promising reports on promoting myocardial recovery with pharmacologic or regenerative therapies in conjugation with MCSD placement, and this indication will likely continue to grow.

With wider interpretation of the definition, bridge to recovery has come to encompass placement of a temporary right ventricular assist device (RVAD) simultaneous with or shortly after an implantable LVAD placement, and this indication is not uncommon.

Bridge to Decision

MCS may be required for an acutely ill patient with cardiogenic shock. Acute hemodynamic decompensation may not permit the thorough evaluation required for BTT or DT. A short-term MCSD may be applied as a “bridge to decision.”

Acute hemodynamic collapse can be caused by various processes, including, but not limited to acute myocardial infarction (AMI), cardiomyopathy, myocarditis, acute on chronic HF, severe rejection of a transplant heart, postcardiotomy shock, and unsuccessful percutaneous coronary intervention. Regardless of the etiology, the patient will need continuous assessment and may require progressively escalating support of care. If the patient deteriorates or does not improve despite inotrope/vasopressor administration, nitric oxide inhalation, and/or IABP placement, MCSD placement is considered. Another clinical scenario necessitating MCSD use as a bridge to decision is the patient who requires continuous cardiopulmonary resuscitation (CPR). In this setting, a clinical decision is made immediately, and once it has been decided to proceed with MCSD placement, the cardiac surgeon/cardiologist should establish circulatory support using the quickest modality available at his or her facility.

Bridge to decision is a strategy used to defer the decision of whether to proceed with longer-term and more definitive therapy while the patient is supported with a short-term device. The patient, while on the device, may recover from or succumb to the decompensation in cardiac function, or may continue to rely on MCS. It has been proposed that the success of long-term implantable LVAD support relies on the optimization of the preoperative status of its candidates. In other words, LVAD implantation in moribund patients results in dismal outcomes. Bridge to decision use of short-term MCS allows selection of those patients who will benefit from a long-term device. Evaluation for candidacy for a longer-term strategy such as heart transplantation or DT is initiated as soon as the MCSD is placed.

Bridge to Bridge

In a small number of patients, a short-term MCSD becomes necessary to replace a previously placed short-term MCSD while the patient is awaiting a more definitive plan. This indication is called bridge to bridge. The most common clinical scenario involves replacing a short-term MCSD that is providing inadequate flow, or a VA ECMO, with a more reliable short-term MCSD in order to facilitate further improvement in the clinical status of the patient who is too ill to undergo a long-term implantable device placement. Bridge to bridge is usually applied for exchange of a short-term device to a long-term device.

A total of 420 patients from 75 institutions were entered into the INTERMACS database from June 2006 to December 2007.24 The indication for MSCD placement was BTT in 336 patients (179 for BTT, 83 for “likely to be listed,” 44 for “moderately likely to be listed,” and 30 for “unlikely to be listed”) and DT in 63 patients.

The success of an MCSD is heavily dependent on appropriate patient selection.21,25,26 Potential candidates are evaluated from a hemodynamic and nonhemodynamic standpoint. In 2006, the International Society for Heart and Lung Transplantation proposed Guidelines for the Care of Cardiac Transplant Candidates, in which recommendations for MCSD use are described.18

From a hemodynamic perspective, patients considered for MCSD placement can no longer sustain adequate systemic oxygen delivery to maintain normal end-organ function despite maximal medical therapy and/or IABP support. Such hemodynamic compromise can be the result of a variety of cardiac conditions. Traditional hemodynamic criteria for device placement include: blood pressure (BP) less than 80 mm Hg, mean arterial pressure less than 65 mm Hg, cardiac index less than 2.0 L/min/m2, pulmonary capillary wedge pressure greater than 20 mm Hg, and a systemic vascular resistance (SVR) greater than 2100 dynes.sec/cm.27 Malignant ventricular arrhythmias that cannot be adequately controlled with permanent pacemakers and/or implantable defibrillators are an additional hemodynamic indication for MCS.

Determining the hemodynamic indication for MCSD placement, however, is not always straightforward. Although the need for MCS is rather evident for a patient in acute cardiogenic shock, conflict may exist when a patient demonstrates an insidious course with congestive heart failure (CHF). Examples include: a hospitalized patient with CHF who worsens or fails to improve over a period of days; a patient managed on an outpatient basis for known CHF who slowly deteriorates, requiring incremental medical treatment; or a transplant candidate who is managed with an outpatient inotrope infusion. In order not to miss the appropriate window for MCSD placement, the clinician should be aware of the worrisome signals of CHF decompensation. These signals include worsening symptoms (recurrent admission, refractory to medical therapy, symptomatic at rest), medication issues (intolerance to or lower doses of ACE inhibitors/angiotensin receptor blockers [ARBs] and beta-blockers, increasing doses of diuretics), hyponatremia, and unresponsiveness to cardiac resynchronization therapy. Inotrope dependence is associated with a less than 50% 6-month survival.28,29 In these situations, in addition to assessment with right heart catheterization and echocardiogram, end-organ function, in particular renal and hepatic function, is followed closely. It is becoming evident that outcomes after MCSD placement are better if the device is placed before the development of end-organ dysfunction. For patients with HF refractory to medical management, or for transplant candidates on inotropic support, a surgical consult should be obtained sooner than later, and the appropriate indication and timing of MCSD placement should be determined through a thorough discussion between the cardiac surgeon and cardiologist.

The INTERMACS registry proposed a series of seven clinical profiles with three modifiers for patient selection. These profiles help define the acuity and severity of illness and simplify the assessment of implant risk30–32 (Table 66-2). The modifiers include arrhythmia, temporary circulatory support, and frequent flyer.32 Approximately 60 to 80 % of indications for LVAD placement were INTERMACS level 1 (“crash and burn”) or level 2 (“sliding on inotropes”).21,32

Of note, because only LVAD support has been approved for DT, significant impairment of right ventricular (RV) function and high PVR may exclude patients from eligibility for DT.

Patients in acute cardiogenic shock usually cannot wait for a thorough evaluation and thus the decision to proceed with MCSD may be made purely from a hemodynamics standpoint.

However, several additional factors discussed previously have to be taken into consideration to ensure the success of a long-term MCSD. Again, the treating team has to be clear of the goal of MCSD placement. The majority of long-term MCSDs are implanted either for BTT or DT purposes, each of which has different criteria for appropriate patient selection. Specific requirements for DT were mentioned in the previous section. The following are general considerations applicable to patient selection for both treatment goals.

Assessment of Operative Risk

Operative mortality constitutes a large portion of mortality after MCSD placement. The common causes of early postoperative death are bleeding, RV failure, sepsis, multiple organ failure (MOF), and stroke. Although several screening scales, including one developed at our own program, have been developed to predict early mortality after long-term LVAD placement, they are not always helpful for the assessment of procedural difficulty.22,33 The responsible surgeon has to address the surgical complexity of each case. Certain anatomical scenarios, such as the patient with multiple previous sternotomies for complex congenital heart disease, might possess a prohibitive operative risk regardless of the remainder of the patient's risk profile.

Assessment of Organ Function

Patients with end-stage HF commonly have significant comorbidities. Life-limiting comorbidities preclude eligibility for long-term MCSD placement. On the other hand, in the setting of impaired end-organ function, some degree of recovery may be seen after MCSD placement. Renal and hepatic functions were demonstrated to improve with both pulsatile and continuous-flow LVAD support.34–40 Assessing and defining the severity and reversibility of vital organ function is critical in the patient selection process.

Renal dysfunction, which commonly coexists in HF patients, is strongly associated with poor outcomes after VAD implantation.41,42 Patients who are on chronic dialysis are generally excluded from implantable MCSD placement. Furthermore, patients on continuous venovenous hemofiltration (CVVH) or those with a serum creatinine of greater than 3.0 mg/dL are excluded from LVAD use for DT at our program.

Liver function is also carefully addressed. We have proposed a screening scale to predict survival after LVAD insertion.33,43 Univariate and multivariate analyses of our patient cohort confirmed a prothrombin time of greater than 16 seconds to be one of the most significant variables included in the summation score. Reinhartz et al. demonstrated that patient survival after LVAD placement decreased (82% to 56% to 33%) as preoperative direct bilirubin levels increased (<1.2 to 1.2–3.6 to >3.6 mg/dL, respectively).44 Liver function strongly correlates with patient outcomes after a VAD. For DT, we exclude patients whose intern ational normalized ratio is greater than 2.5 (even with efforts to reverse it), and/or patients whose ALT and AST are more than three times normal values.

Patients with advanced pulmonary disease who are on prolonged ventilatory support or whose FEV1 is less than 1 L are not candidates for long-term MCSD. Other considerations include severe impairment of cognitive function, a history of neurologic events with significant residual deficits, active infection, active gastrointestinal bleeding, severe peripheral vascular disease (such as critical limb ischemia with rest pain or ulceration, or past limb amputation), and severe malnutrition. Malignancy is not necessarily a contraindication.

Other Considerations

MCSD operation and maintenance requires a large commitment on the part of the patient and his/her support structure. Our program requires that all patients referred for MCSD evaluation undergo a comprehensive multidisciplinary assessment. A psychiatrist assesses any psychiatric impairment that may contribute to the patient's inability to adhere to a postdevice regimen. In addition, neurocognitive function is evaluated by neuroscientists. The psychosocial evaluation performed by qualified healthcare personnel explore: social, personal, vocational, financial and environmental supports. A mental health history is also taken, assessing for substance abuse and alcohol use or abuse and its potential impact on the success or failure of device support. Nutritional status is screened with referral to a nutritionist as needed.

Lastly, age itself is not an absolute contraindication to MCSD placement.18 Our program addresses the feasibility of device placement on an individual case basis with particular attention paid to end-organ functional reserve and psychosocial issues in older patients. It is a Class 1 recommendation from the ISHLT that patients older than 60 years of age undergo a thorough evaluation for the presence of other clinical risk factors.18

Difficulties in predicting outcomes after long-term MCSD placement prompted several working groups to develop screening scales.

The revised Columbia screening scale published in 2003 offers a way of stratifying prospective device patients according to their risk of operative death from LVAD insertion based on several clinical factors: mechanical ventilation, postcardiotomy; prior LVAD insertion; central venous pressure greater than 16 mm Hg; and prothrombin time greater than 16 seconds.43 Each factor is given a weight (4, 2, 2, 1, 1, respectively), with a cumulative score of >5 predicting an operative mortality of 46%, versus 12% for a score ≤5. This scoring system is a revision of a prior scale developed in conjunction with the Cleveland Clinic Foundation.33 The revised scoring scale was the result of an analysis of 130 patients undergoing implantation of the HeartMate vented electrical device.

Lietz et al., focusing on DT, analyzed outcomes from 280 of 309 patients who underwent LVAD (HeartMate XVE) implantation as DT since the completion of the REMATCH trial. Nine variables, which were found to be significantly related to 90-day in-hospital mortality, were assigned a weighted risk score equal to the odds ratio rounded to the nearest integer (Table 66-3). The sum of the weighted risk scores is suggested as a risk score. Scores are categorized as low being risk score of 0 to 8, medium to high (9 to 18), and very high (>19). These risk factors are representative of severe deterioration of the general medical condition of the patient, as evidenced by poor nutritional status with a low serum albumin level, impaired renal function, markers of right heart failure such as low pulmonary artery pressures, or congestive elevation of liver enzymes. Probable infection, as evidenced by leukocytosis and coagulation abnormalities such as declining platelet count or prolonged international normalized ratio, anemia, and small body size further worsen the chance of operative survival. This screening scale was shown to correlate with 90-day and 1-year survival after LVAD implantation (Fig. 66-6). Other preexisting screening scales such as the APACHE II and Seattle Heart Failure Model were also shown to be predictive of midterm success of pulsatile LVAD therapy.45,46

Although the above scoring systems guide the preoperative risk assessment, they were developed based on the experience with the first-generation LVADs, and may not be directly applicable to newer devices. Using a cohort of 86 patients receiving a continuous-flow LVAD at the Johns Hopkins Hospital, Schaffer et al.47 compared various risk screening scales including the Leitz-Miller, Columbia, APACHE II, INTERMACS, and Seattle Heart Failure Model risk scores to assess their ability to predict on 30-day, 90-day, and 1-year mortality, and found that the Seattle Heart Failure Model best differentiated low- and high-risk patients at all mortality end points.

Determining whether biventricular support will be necessary prior to MCSD placement is imperative in choosing an appropriate device or in determining candidacy for DT. It usually becomes evident which ventricle predominantly needs to be supported, and most often the it is the LV. Previously mentioned CMS criteria well summarize the indications for LVAD support: NYHA Class 4 HF refractory to optimal medical management, LVEF less than 25%, functional limitation with a peak oxygen consumption of less than 12 mL/kg/min or the patient has a continued need for intravenous inotropic therapy.20 In addition, patients with malignant arrhythmias as well as anyone on the transplant waiting list could be considered for LVAD therapy.

The difficulty lies in determining whether another ventricle, usually the RV, needs to be supported or not. RV failure occurs in 20 to 40% of LVAD recipients and it correlates with worse outcomes.48–50 In general, RV failure occurs in “sicker” patients reflecting more advanced HF with coexisting end-organ dysfunction owing to low organ perfusion pressures (low BP and high venous pressure). Despite technologic advances, Patel et al. demonstrated that the incidence of RV dysfunction remained constant with an incidence of 35% after HeartMate XVE implantation compared with 41% after HeartMate II implantation. However, fewer HeartMate II patients required RVAD placement and fewer required pure inotropic support for right heart failure.49 Based on 68 patients who developed RV failure out of 197 LVAD recipients, the group from the University of Michigan proposed an RV failure risk score consisting of a vasopressor requirement, AST ≥80 IU/L, bilirubin ≥2.0 mg/dL, and creatinine ≥2.3 mg/dL.51

Several studies attempted to identify predictors of BiVAD need, identifying more than 20 variables. The results have been inconsistent. Ochiai et al. from the Cleveland Clinic reviewed 245 patients receiving pulsatile LVADs, among whom 23 patients required RVAD support, and reported that the need for pre-LVAD circulatory support, female gender, and nonischemic etiology were the most significant predictors of RVAD requirement after LVAD placement.52 Fitzpatrick et al. reported a large series from the University of Pennsylvania. They retrospectively analyzed 266 LVAD recipients, of whom 99 (37%) required RVAD placement after LVAD placement. Cardiac index ≤2.2 L/min/m2 (odds ratio 5.7), RV stroke work index ≤0.25 mm Hg/L/m2 (odds ratio 5.1), severe preoperative RV dysfunction on echocardiographic assessment (odds ratio 5.0), preoperative creatinine ≥1.9 mg/dL (odds ratio 4.8), previous cardiac surgery (odds ratio 4.5), and systolic BP ≤96 mm Hg (odds ratio 2.9) were shown to be the best predictors of RVAD need. They proposed a risk score: 18 × (CI) + 18 × (RVSWI) + 17 × (creatinine) + 16 × (previous cardiac surgery) + 16 × (RV dysfunction) + 13 × (SBP), with a maximum possible score of 98. Applying the risk score to their cohort, a score of ≥50 predicted the need for BiVAD support with a sensitivity and specificity of 83% and 80%, respectively.53 These scoring systems, however, remain to be validated.

Most malignant arrhythmias improve with LVAD support with or without automated implantable defibrillators/permanent pacemakers. BiVAD support, however, should be considered for recurrent sustained ventricular tachycardia or ventricular fibrillation in the presence of an untreatable arrhythmogenic pathologic substrate (eg, giant cell myocarditis).18

Isolated RVAD support is rarely indicated, especially as long-term support.54 Potential indications include isolated RV infarct, isolated RV failure after heart transplantation or pulmonary embolism, or postcardiotomy RV failure.

INTERMACS data from 420 patients at 75 institutions entered into the INTERMACS database from June 2006 to December 2007demonstrated that 314 patients received LVADs, 5 RVADs, 77 BiVADs, and 24 TAHs.24

Device selection starts with consideration of the goal of MCS placement from among the five goals described in the “Goal of Device Support” section. Also of importance is to determine which ventricle(s) will be supported.

For a bridge to decision purpose, a short-term MCSD is appropriate as they are quicker to place and less expensive. When a patient is too unstable to go to the operating room, VA ECMO via percutaneous cannulation may be the most realistic option. It is relatively easy and quick to establish, and is familiar to cardiac surgeons and paramedics compared with other types of devices. These features make it an excellent device of choice in emergent situations at a wide variety of institutions. This is also the only device available for biventricular support in extreme emergencies. The major shortcoming of VA ECMO is its mandatory use of an oxygenator, which requires strict anticoagulation and careful observation to prevent thromboembolic/hemorrhagic complications. Patients also have to remain on mechanical ventilation. Its use is limited to a relatively short period of support (hours to days). Percutaneous LVADs such as the Impella or TandemHeart for LV failure might be a reasonable option, especially when the deterioration occurs in the cardiac catheterization suite. These percutaneous devices may not be able to provide enough flow for a large patient or a very sick patient, in which case they may have to be replaced with some other device in the operating room. The options at this point include the CentriMag, Thoratec PVAD, and ABIOMED AB or BVS5000. These devices are easy to place, provide adequate flow, and can be used either as a BiVAD or a univentricular assist device. These devices are suitable in the rare bridge-to-bridge situation described earlier, in which an initial temporary device needs to be replaced (percutaneous VADs or ECMO). Another application is for temporary RVAD support at the time of or shortly after implantable LVAD placement. Our current preference for short-term MCS is to use the CentriMag, which enables establishment of high flow support through simple cannulation techniques almost equivalent to regular cannulation for routine cardiac surgery. In our experience, the CentriMag has been reliably used with support periods of greater than 1 month.

For BTT, unless the estimated waiting time to transplantation is short, in which case one of the above mentioned short-term devices may be used, a long-term implantable device is the device of choice. The Thoratec PVAD and IVAD are the only FDA-approved devices at this point for biventricular support with implantable features. FDA-approved and currently purchasable LVADs include the HeartMate XVE, HeartMate II, and Novacor LVAD. Many other devices are under clinical investigation and the options will continue to grow. Currently there is no comparison data among devices, except the HeartMate II was shown to be superior to the HeartMate XVE in a randomized controlled trial.8 We currently use the HeartMate II or other investigational devices as part of clinical trials.

The only device currently approved by the FDA for DT is the HeartMate XVE. Again, several clinical investigations are undergoing, and likely more options will become available in the near future.

Anticoagulation

Continuous-flow devices generally require antiplatelet therapy and anticoagulation. Patients with contraindication to these regimens are considered for the HeartMate XVE.

Body Size

Patient body habitus is a very important factor in choosing a device. The size of an implantable VAD dictates the lower limitation of body size, which is usually described in BSA. For example, a BSA of less than 1.5 m2 is a contraindication for conventional pulsatile devices, such as the HeartMate XVE whereas the HeartMate II may be placed in patients with a BSA as low as 1.2 m2. For smaller or pediatric patients, paracorporeal devices may have to be chosen.

Significant semilunar valve regurgitation is repaired at any type of MCSD insertion. Important anatomical abnormalities that require surgical correction at LVAD implantation include PFO, AI, TR, and LV apical thrombus. Intraoperative transesophageal echocardiogram (TEE) plays an important role here.

An LVAD may create pressure gradient from the right to left atrium which exaggerates right to left shunt through a PFO, resulting in profound desaturation.55 Any PFO is closed with bicaval cannulation. We also routinely ask the anesthesiologist to perform a bubble study with TEE after LVAD placement to rule out a small PFO that might have remained undetected.

The presence of AI creates a circuit from the ascending aorta, into the LV, and back to the ascending aorta through the LVAD. LVADs, especially ones with continuous flow, expose the aortic valve to a constant pressure gradient across the valve with the LV being suctioned through the inflow cannula, and a nonphysiologic reversal of blood flow coupled with stagnant blood flow at the base of the aortic root. These unique hemodynamic features may potentiate AI. Significant regurgitation is controlled surgically. The aortic valve can either be repaired or sewn shut. The aortic valve closure leaves patients with no LV outflow, which may become of consequence in the case of device malfunction. We therefore prefer to repair the aortic valve whenever it is feasible. Our procedure of choice involves approximating the center of the leaflets, the nodes of Arantis, with a 4-0 polypropylene suture reinforced with small bovine pericardial pledgets. Stagnant blood around the aortic root can be thrombogenic especially when patients have a preexisting mechanical aortic prosthesis, in which case we close off the aortic valve with a patch. The other option is to replace the mechanical valve with a bioprosthesis.

MCSD candidates are prone to develop TR because of dilatation of the tricuspid annulus, RV dysfunction, and the presence of pacing leads. We aggressively repair significant TR at the time of LVAD implantation to maximize forward flow from the RV. The repair is usually readily accomplished by placing a ring, but additional repair techniques may be necessary for leaflet pathology created by pacing leads.

LV mural thrombi occur in one-third of Q-wave AMI cases, 50% of LV aneurysms, and up to 18% of hearts with dilated cardiomyopathy.56 Cardioplegic arrest may be required for removal of the LV thrombus to prevent systemic embolization.

We describe our technique of long-term LVAD implantation. Most of the steps are shared by many implantable LVADs. Essential components of the operation include (1) mediastinal exposure and creation of the device pocket; (2) outflow graft construction to the ascending aorta; (3) placement of device inflow, usually at the LV apex; (4) de-airing of the device; and (5) device actuation. Preoperative antibiotic prophylaxis covering both gram-positive and gram-negative bacteria is administered and usually continued for 48 to 72 hours postoperatively. Our preference is the combination of rifampin, trimethoprim-sulfamethoxazole, and fluconazole. Autologous blood is drawn by the anesthesiologist before heparinization. Patients tolerate phlebotomy of up to 1 L very well as they are usually fluid overloaded. Administration of this blood helps with hemostasis at the completion of the procedure. We have abandoned serine protease inhibitor (eg, aprotinin) use.

A vertical midline incision is made with extension onto the abdominal wall. After standard median sternotomy, a preperitoneal pocket is created in the preperitoneal space in the left upper quadrant.57 Sometimes the preperitoneal plane is very thin and attenuated, so the posterior rectus sheath is entered and a plane superficial to this is developed. The device is brought onto the field and positioned in the preperitoneal pocket. The driveline is tunneled percutaneously and brought out of the skin through the first counter-incision in the right upper quadrant and finally through the second punched-out incision placed in the left upper quadrant (Fig. 66-7). After systemic heparinization, the patient is cannulated for cardiopulmonary bypass. Venous cannulation can be achieved with a single dual-stage cannula placed in the right atrium, or with standard bicaval cannulation if concomitant procedures are being performed (ie, tricuspid valve repair or closure of a PFO). The ascending aorta is cannulated distally with consideration for future reoperation for cardiac transplantation. An effort is made to minimize cardiopulmonary bypass times in this ill group of patients. If the hemodynamic status permits, an attempt is made to perform the outflow graft anastomosis off-pump using a partial occluding clamp placed on the ascending aorta (Fig. 66-8). The outflow graft is measured and cut with enough length to allow for a gentle curvature toward the right chest without redundancy that may lead to outflow graft kinking. A longitudinal aortotomy is created and the anastomosis performed using continuous 4-0 polypropylene suture. The anastomosis is reinforced with BioGlue (CryoLife Inc., Kennesaw, GA). Cardiopulmonary bypass is initiated and attention is turned to the placement of the LV apical inflow cannula. This is usually done under conditions of normothermia with a beating heart. A specialized coring knife is used to create the apical core. Care must be taken to correctly identify the apex of the left ventricle and avoid deviation into the septum or the lateral wall. After coring of the apex, care is taken to ensure a clear inflow tract. Any excess trabeculations or myocardium that may impinge on the inflow cannula are excised. Ventricular thrombus is carefully removed. Sutures of 2-0 braided polyester reinforced with felt pledgets are then placed around the circumference of the core in a horizontal mattress fashion and passed through the sewing ring of the inflow cuff (Fig. 66-9). The cuff is secured and then attached to the inflow cannula of the device. The heart is allowed to fill with blood and the device is de-aired through a venting hole placed in the outflow graft. Evacuation of air is confirmed with TEE. The patient is then weaned from cardiopulmonary bypass, and the device is initiated. After ensuring adequate hemostasis, drains are placed in the mediastinum, pleural spaces, and device pocket. A synthetic pericardial membrane, such as Gore-Tex membrane (Gore Medical Products, Flagstaff, AZ), is placed over the anterior mediastinal structures and secured to the pericardial edges to minimize the risk of injury during reoperative median sternotomy. The sternum and soft tissues are then closed in the standard fashion. Particular attention is given to secure abdominal fascial closure to prevent wound dehiscence and subsequent device infection. In certain patients in whom excessive mediastinal bleeding is encountered secondary to coagulopathy, our preference is to pack the mediastinum and perform delayed sternal closure after approximately 24 hours, once bleeding has subsided and the patient has been adequately resuscitated.

Postoperative care is crucial to successful outcomes in patients undergoing MCSD implantation, and management starts before surgery. Again appropriate patient selection is above all the most important key to success. Also important is identifying the appropriate timing of MCSD insertion; patients must first receive aggressive medical management as necessary with IABP to optimize end-organ function. Preexisting infection needs to be identified and treated. Appropriate MCSD configuration (LVAD, BiVAD, or TAH, etc) and device selection follow.

Meticulous surgical technique is the key to preventing unnecessary postoperative complications. Blood product transfusion with platelets, fresh frozen plasma, and cryoprecipitate is given as needed. Preheparinization harvest of autologous blood seems helpful in hemostasis. We are cautious about using recombinant factor 7 in this group of patients because of unknown interaction between the mechanical device and this product. Adequate hemostasis is achieved before leaving the operating room, because massive blood transfusion contributes to the development of RV failure after LVAD insertion. After device initiation, proper device function is ensured in the operating room to restore and maintain adequate end-organ perfusion. Inadequate device flows may be the result of mechanical problems such as inflow cannula obstruction or malposition, outflow graft kinking, or cardiac tamponade. Optimization of RV function (for isolated LVAD cases) is initiated before coming off CPB in concert with the anesthesiologist. Continuous assessment of RV function throughout the operation is accomplished with direct visualization by the surgeon and TEE by the anesthesiologist. The following seven hemodynamic factors are to be taken into consideration and optimized: CVP (central ventricular pressure) (preload), PVR (afterload), RV contractility, interventricular interaction, heart rate (HR), rhythm, and SVR (for coronary perfusion to the RV). RV distension is avoided as it leads to increased stress on the thin RV wall, worsened TR, decreased end-organ perfusion, and right-to-left shunting via a PFO (if untreated). A CVP of over 15 mm Hg is an alarming sign. Inotropic agents such as phosphodiesterase inhibitors, dobutamine, and epinephrine as well as inhaled nitric oxide are used liberally to support the RV.58 Dynamic interventricular interaction through the ventricular septum significantly changes after LVAD initiation and this may compromise RV function, especially with continuous-flow devices. With excessive unloading by the device, the ventricular septum is shifted toward the LV impairing the contraction of the RV. Flow through the device is adjusted to maintain an appropriate position of the septum based on TEE. The appropriate shape of the LV in TEE short-axis view is “round shape” whereas flattened septum (D-shaped LV) indicates excessive suction. HR and rhythm are controlled with pacing and/or administration of antiarrthymic agents. Maintaining SVR and adequate perfusion pressure (mean systemic BP approximately 80 mm Hg) to ensure RV coronary perfusion is achieved with vasopressors. Arginine vasopressin has been shown to be effective in treating vasodilatory hypotension in this group of patients.59 Excessive afterload, however, may impair pump afterload and should be treated accordingly particularly with continuous-flow devices. If optimization of RV function can not be achieved, RVAD support is considered.

Postoperative care in the intensive care unit (ICU) is the continuation of intraoperative care, which is centered on optimization of RV function. The focus of the first several hours is on monitoring for blood loss and RV function. (Near)-normalization of coagulation status is achieved with proper warming and blood product transfusion. The surgeon should have a low threshold for mediastinal reexploration for excessive postoperative bleeding. RV function is assessed with hemodynamic parameters keeping the above-mentioned seven factors (CVP, PVR, RV contractility, interventricular interaction, HR, rhythm, and SVR) in mind. In a typical stable patient after LVAD insertion, a satisfactory mean BP of above 80 mm Hg and CVP below 10 are achievable with minimum to moderate doses of inotropes/vasopressors. Proactive and aggressive treatment of subtle signs of RV failure, such as increasing vasopressor requirement, elevated CVP with relatively low pulmonary artery pressures, low output from the LVAD (measured by the device) or the RV (measured by thermodilution method with a pulmonary artery [PA] catheter), low urine output, and low mixed venous oxygen saturation should be undertaken. This can help avoid deterioration of RV function and the subsequent cascades that ensues, including device low-output and eventual end-organ dysfunction. The treatment of RV dysfunction consists of liberal increase in dosage and/or addition of inotropes, vasopressors, and inhaled nitric oxide, aggressive diuresis, instituting CVVH, optimization of ventilator settings, and controlling the HR and rhythm with pacing and/or antiarrhythmic agents. Also, the interventricular effect of the LVAD needs to be addressed. This is a frequent cause of hemodynamic instability in the first 24 to 48 hours after LVAD insertion (especially with continuous-flow devices). Temporarily decreasing the pump speed usually restores hemodynamics. Some of the devices provide certain indices. For instance, the HeartMate II provides a pulse index (PI), which is a measure of the amount of pulsatility seen by the pump over interval of 15 seconds. PI relates to the amount of unloading of the LV and is a helpful guide for the adjustment of pump speed. Overdecompression of the LV might also manifest as ventricular arrhythmia caused by mechanical irritation of the ventricular wall with the inflow cannula. Reducing the speed will reshape the LV with resolution of the arrhythmia.

Once stabilized, the patient is weaned from sedatives and then from the ventilator support. Unexpected significant hypoxia could be related to an undiscovered PFO. Vasopressors/inotropes are weaned off as tolerated, maybe except for one of the inotropes (our preference is milrinone), which may be kept on for several more days to support the RV and help diuresis. Aggressive diuresis is initiated to maintain appropriate preload to the RV. A continuous IV infusion of furosemide is preferred in the ICU and it is converted to intermittent IV bolus. Bumetanide and nesiritide are added as needed. Temporary CVVH may be required. Inotropes and diuretics are used liberally until the patient excretes the excess fluid, which is assessed by daily measurement of weight and physical exam (pretibial edema, lung sounds). Transthoracic echocardiogram is obtained periodically to guide the amount of LV unloading (reflected by the LV end-diastolic dimension) and RV function. Should the patient not respond to escalating treatment for RV failure at any point, RVAD placement should proceed without delay. Once complete hemostasis is confirmed, usually by postoperative day 2 or 3, antiplatelet therapy and/or anticoagulation are started, based on the recommendation for the device used. We no longer routinely initiate anticoagulation with an IV heparin infusion because of high incidence of bleeding complications. Currently anticoagulation is started with oral warfarin approximately postoperative day 3 to 5 when the patient is ready to be transferred out of the ICU. Enteral feedings are resumed early in the postoperative course. Once the patient is extubated, attention is focused on physical therapy and rehabilitation.60 Should the patient develop anemia with hyperbilirubinemia, hemolysis needs to be ruled out although it is very rare.61

Special Postoperative Care after BiVAD Insertion

Obviously RV failure is not a concern with BiVAD support. Sometimes supernormal output from the BiVAD is required to meet the metabolic needs of various end organs because, generally speaking, patients who require BiVAD support are sicker than those who require isolated LVAD support. Flow, however, has to be balanced between the LVAD and the RVAD. The left side of the heart receives more return than the right because of bronchial circulation. RVAD flow is adjusted so as to not overflow the LVAD one unless flow through the aortic valve is maintained; otherwise pulmonary congestion will result.

Overall clinical outcomes have been improving owing to advances in technology and the accumulation of clinical experience.8,39,62–64

BTT

Success rates with pulsatile LVADs as a BTT increased from below 65% in the 1990s to 72% in early the 2000s at our program.65 Further improvement was noticed with newer-generation LVADs. John et al. reported six month survival rate of 87% for 32 patients who received the HeartMate II as a BTT.66 A prospective, multicenter study of 133 patients who received the HeartMate II as a BTT showed that at 6 months 75% reached the principal outcomes of heart transplantation (42%), cardiac recovery (1%), or ongoing device support (32%) with a mortality rate of 19%.62 Overall actuarial survival for patients continuing to receive pump support was 68% at 12 months. Over the past decade, the number of heart transplant recipients supported by a VAD at the time of transplantation has more than doubled to over 400 per year.67 Studies have concluded that long-term survival is not affected by pretransplant VAD support.68,69 Using UNOS registry data, Russo et al. compared the long-term survival after heart transplantation between recipients who received a VAD as a BTT and those who did not. The VAD group was further divided into extracorporeal, intracorporeal, and paracorporeal VADs. Although the extracorporeal group had diminished graft survival in the first 90 days (likely related to the more morbid pre-VAD status of the recipients), there were not significant differences in risk-adjusted graft survival at up to 5 years between recipients who received any type of VAD as a BTT and those who did not.67

Success rates for BTT with BiVAD support, on the other hand, were reported to be only 25 to 60%.48,51–53,70–72 Poorer success was reported specifically among those who required an RVAD following LVAD placement. Early planned institution of BiVAD support might improve outcomes.73

DT

Outcomes for LVADs placed for DT are well-summarized in two landmark randomized studies.7,8 LVADs were shown to be superior to optimal medical management with regard to survival and quality of life. The REMATCH trial demonstrated a 48% reduction in the risk of death at 1 year with a pulsatile LVAD (HeartMate VE). Subsequently, a continuous-flow device (HeartMate II) was shown to improve survival at 2 years (58% vs 24% for a pulsatile LVAD, the HeartMate XVE). Interestingly, the 2-year survival with the pulsatile-flow device in these two randomized studies remained roughly the same, and this suggests that the main contributor to improvements in survival is technologic advancement.

Bridge to Recovery

Few LVAD recipients show recovery of ventricular function, and even if they do, they frequently suffer from recurrence of HF after LVAD explantation. We have shown that significant clinical recovery occurred in only 5 out of 111 patients, and in these patients, symptoms frequently returned and many died of consequences of HF.74 In large case series, recovery with removal of LVAD support has been seen in 4 to 30% of patients with a 20 to 50% incidence of recurrent CHF.75–78 The results of Birks et al. regarding intensive pharmacologic treatment are most promising.79 Their patients underwent institution of LVAD support and were treated with lisinopril, carvedilol, spironolactone, and losartan to enhance reverse remodeling. Once regression of LV enlargement had been achieved, the beta2-adrenergic receptor agonist clenbuterol was administered to prevent myocardial atrophy. Eleven of fifteen patients had sufficient myocardial recovery to undergo explantation of the LVAD. The cumulative rate of freedom from recurrent HF among the surviving patients was 100% and 88.9%, 1 and 4 years after explantation, respectively.

INTERMACS provides standardized definitions of adverse events. The defined events include major bleeding, cardiac arrhythmia, pericardial fluid collection, device malfunction, hemolysis, hepatic dysfunction, hypertension, major infection, myocardial infarction, neurologic dysfunction, psychiatric episode, renal dysfunction, respiratory failure, right heart failure, non–central nervous system (CNS) thromboembolic event, venous thromboembolism event, wound dehiscence, and other adverse events. Among these, CNS events, MOF, RV failure, and arrhythmias remain the most common cause of death.24 The cumulative incidence rate of adverse events in the first 60 days after MCSD insertion can be as high as 89%.61

Operative Mortality

Recently reported 30-day survival rates after MCSD insertion are approximately 80 to 90%.22 Higher INTERMACS profiles were shown to correlate with higher mortality,80 and operative risk categories developed by Lietz et al. for DT patients estimate 90-day in-hospital mortality after LVAD placement at 2% in low-risk patients and 81% in very high-risk patients.22 Estimated operative mortality rates from the revised Columbia screening scale, developed with a focus on the operative mortality, are approximately 5% for a score of 0–1, approximately 15% for a score of 2–4, and approximately 45% for a score of ≥5.43 Better patient selection has contributed to decreasing operative mortality, and proceeding to MCSD insertion before a patient becomes too sick is the key. Postcardiotomy support remains one of the most challenging indications. The Thoratec registry data for PVAD placements by March 2005 showed a survival of 44.7% with LVAD support (n = 253) and 32.4% with BiVAD support (n = 241) for this indication.81

Major Bleeding

Bleeding is the most common complication after MCSD implantation and mediastinal reexploration is required in 20% up to 60% of MCSD recipients.82 Resuscitation with massive blood product transfusion is associated with increased morbidity and mortality. A recent study demonstrated that the number of bleeding events requiring packed red blood cell transfusion was 1.66 per patient-year for a continuous-flow device and 2.45 for a pulsatile-flow device.8 Bleeding risk persists even after the acute postoperative period, and this seems to be above what is normally observed in patients receiving anticoagulation treatment. Late bleeding most commonly manifests as gastrointestinal bleeding, and continuous-flow devices appear to have a higher incidence of this complication compared with pulsatile-flow devices.83 Development of arteriovenous malformations and impaired von Willebrand factor–dependent platelet aggregation are the proposed mechanisms.83,84

Right Heart Failure

RV failure occurs in 20 to 40% of LVAD recipients and approximately 5% of cases require RVAD support. A detailed discussion is found in Univentricular or Biventricular Support.

Device Malfunction

Device durability is an important limitation to the use of currently approved pulsatile-flow LVADs because valve or bearing failures occur frequently by 18 months.7 The advent of continuous-flow devices dramatically reduced the incidence of this complication. In a recent randomized study, there were no primary-pump or bearing failures in patients with a continuous-flow LVAD.8 It was suggested that redesign of the percutaneous lead and development of modular components may further reduce the infrequent need for replacement, which occurred in only 9% of patients. The University of Michigan group experienced an 11% incidence of device malfunction with BiVAD support compared with 9% with LVAD support.61

Major Infection (Device-Related)

Sepsis and device malfunction were the most common causes of death in patients who received a pulsatile-flow LVAD in the REMATCH trial. Within 3 months after implantation, the probability of device infection was 28%. Infections mostly involved the drive-line tract and pocket and were treated with local measures and antibiotics. The continuous-flow device reduced the rate of device-related infection to nearly half that seen with the pulsatile-flow device.8

Neurologic Dysfunction

MCSDs predispose recipients to the risk of thromboembolism, with ischemic stroke being the most devastating consequence. The combined risk of ischemic and hemorrhagic CVA was reported to be 8 to 20%. The type of device may correlate with the risk of stroke. Tsukui et al. reported an actuarial freedom from CVA at six months of 75%, 64%, and 33% with the HeartMate (X)VE and II device, Thoratec BiVAD, and Novacor device, respectively.85

Table 66-4 summarizes the current probability of adverse events from the INTERMACS registry between June 2006 and March 2008.86

Advances in technology have led to the creation of more durable and versatile devices, resulting in improved clinical outcomes for the MCSD recipients. Minimally invasive approaches for MCSD insertion will be developed. MCSDs may be used across a wider spectrum of patients and purposes, and patients with less advanced HF may receive partial circulatory support to improve quality of life. Approaches in which MCS combined with one or more addtitional treatment modalities, such as a drug therapy or cell therapy to prevent post-LVAD explant remodeling, may prove fruitful. In fact, clinical trials in these areas are ongoing.

Accumulation of clinical experience will allow consensus guidelines to be developed. However, at the current rapid pace of technologic progress, such guidelines may be outdated at the time of their publication, somewhat similar to the experience with coronary stents. Therefore, cardiac surgeons performing MCSD surgery should remain updated.

MCS is still a relatively new technology in the treatment of HF, and much research remains to be done to elucidate its role in clinical medicine and also to better understand HF itself. Both basic and clinical investigations are warranted with support from nonprofit organizations and industry.

• Patient selection is the most important component to achieve good outcomes with MCSD implantation.
• Elucidating the goal of MCSD placement (BTT, DT, bridge to recovery, bridge to dicision, bridge to bridge) is crucial for appropriate patient and device selection.
• Optimizing RV function is the key after LVAD implantation.