Scope of the Problem

According to the American Heart Association, approximately 6 million people in the United States have congestive heart failure (CHF). Available statistics indicate that the incidence of CHF in the population approaches 15.2 per 1000 after age 65, 31.7 per 1000 after age 75, and 65.2 per 1000 after age 85, with 1,106,000 hospital discharges for heart failure (HF) in 2006 alone.¹ HF is the leading cause of hospitalization in patients older than 65,² with a reported associated cost of $24 to $50 billion annually.^1–4 On a global scale, HF reportedly affects 0.4% to 2.0% of the adult population.⁵

Despite great advances in the understanding of the pathophysiology of HF and the development of medications that can potentially attenuate the progression of that pathophysiology, morbidity and mortality from this disease remain high. The incidence rate of hospitalization for HF increased by 70% during the 1990s,² and patients with New York Heart Association (NYHA) Class IV symptoms currently have a reported 1-year mortality rate of 30% to 50%.⁶ By comparison, the corresponding rates for NYHA Class I-II patients and Class II-III patients are 5% and 10% to 15%, respectively. (Table 27-1 defines the NYHA symptomatic classes.⁷) Thus, one of the major goals in the management of HF is the prevention of progression to advanced stages.

TABLE 27-1 New York Heart Association (NYHA) Functional Capacity

From The Criteria Committee of the New York Heart Association: Nomenclature and Criteria for Diagnosis of Diseases of the Heart and Great Vessels, 9th edition. Boston: Little, Brown & Co, 1994, pp 253–256.

Although many patients successfully achieve temporary relief of HF symptoms with medical management, the underlying pathophysiology inevitably progresses, and pharmacologic interventions alone eventually will become inadequate in most patients. A variety of surgical procedures can be performed to improve cardiac function and potentially arrest (or even reverse) the progression to severe dysfunction; but until very recently, surgical intervention (short of transplantation or placement of a ventricular assist device [VAD]) was considered contraindicated in patients with advanced HF. Surprisingly, good outcomes with “corrective” interventions, however, now have resulted in patients presenting for surgical treatment of their HF on a regular basis. This chapter describes the procedures typically performed in this population and the anesthetic considerations for patients with advanced HF.

Brief Review of the Pathophysiology

The current understanding of the pathophysiology of chronic HF maintains that initial increases in end-diastolic ventricular volume and pressure trigger the release of endogenous natriuretic peptides that promote diuresis.^8–10 Concurrent activation of the sympathetic nervous system causes peripheral vasoconstriction and increases the inotropic state of the myocardium. Initially, these mechanisms act to decrease excessive preload (which restores wall tension to normal) and maintain cardiac output (CO) and arterial blood pressure (BP) in the face of mildly depressed ventricular function. Eventually, however, the carotid, ventricular, and aortic arch baroreceptors are activated by the relative hypovolemia, which leads to further activation of the sympathetic nervous system (via the medullary vasomotor regulatory center), as well as the renin-angiotensin-aldosterone axis, and the release of vasopressin. The resultant peripheral vasoconstriction, mild fluid retention, and further increases in heart rate and inotropy will again compensate for the failing heart. Ultimately, however, chronic sympathetic stimulation causes myocardial β₁-adrenergic receptors to downregulate, and as ventricular function deteriorates, left ventricular (LV) end-diastolic volumes and pressures again increase, resulting in increased ventricular wall tension. During this time, increased levels of angiotensin II result in adverse myocardial remodeling. Remodeling is a key event in the progression of HF and refers to changes in not only ventricular geometry (e.g., dilation) but also myocardial composition (e.g., myocyte hypertrophy, lengthening, hyperplasia, fibrosis). In addition, increased circulating levels of angiotensin II may enhance myocyte apoptosis (programmed cell death) via a protein kinase C–mediated increase in cytosolic calcium levels.¹¹

Alterations in chamber geometry (dilation) and myocardial remodeling lead to progressively decreased forward CO, perpetuating the vicious cycle of adverse neurohumoral activation and transient, tenuous compensation. Myocardial oxygen demand increases, whereas oxygen supply potentially decreases because of shortened diastolic periods and increased diastolic wall tension. This developing diastolic dysfunction (ventricular “stiffness”) leads to increased left atrial and pulmonary pressures, pulmonary congestion, and increased right ventricular (RV) afterload. This eventually may progress to signs and symptoms of right-sided HF.

Modern medical management of chronic CHF, therefore, uses agents that have been shown to decelerate the progression to severe failure, reduce adverse myocardial remodeling, and enhance survival (e.g., angiotensin-converting enzyme inhibitors,^12–14 β-blockers,^15–18 and aldosterone antagonists^19,20), in combination with other agents that improve the symptoms but have not been shown to improve long-term survival alone (e.g., diuretics, digoxin). (See Chapter 10.)

Limitations of Current Medical Management

Given that there are currently between 300,000 and 800,000 patients in the United States who have progressed to NYHA Class III and IV status despite modern medical management,²¹ it appears that current treatment strategies have significant limitations. Part of the failure of medical management alone to control the progression of the disease may be that treatment has traditionally focused on systolic dysfunction. Four classic stages of HF have been described: (1) an initial cardiac injury or insult, (2) activation of specific neurohormonal axes with resultant cardiac remodeling, (3) compensatory fluid retention and peripheral vasoconstriction, and (4) ultimate contractile failure.²² In contrast with this traditional conception, it is now known that diastolic dysfunction (decreased lusitropic function) is the primary problem in an estimated 30% to 50% of patients with HF. Pharmacologic interventions aimed at improving diastolic dysfunction and attenuating (if not reversing) remodeling currently are the subject of randomized trials worldwide.

Revascularization

Coronary artery disease has become the most common cause of HF.²⁹ Of those patients currently listed for heart transplantation, 36% carry a primary diagnosis of ischemic heart disease, and 31% of those transplanted in 2007 had ischemia as their primary indication. Commonly used terms describing the extent of myocardial injury are defined in Table 27-2.

TABLE 27-2 Commonly Used Terms Describing the Extent of Myocardial Injury and the Potential for Recovery

Where viable myocardium and feasible targets exist, revascularization of chronically ischemic, hibernating myocardium can improve ventricular function, downgrade NYHA functional class, and improve prognosis.^30,31 Although the primary benefit of revascularization appears to be functional improvement of the left ventricle, reducing ischemic substrate for arrhythmias and retarding adverse myocardial remodeling are important secondary benefits.²⁴

Despite an increased perioperative risk for morbidity and mortality in this population, the world’s literature reports current survival rates between 57% and 75% at 5 years, with in-hospital mortality rates between 1.7% and 11%.³¹ A review reported an 83.5% survival rate at 2 years after revascularization compared with only 57.2% survival in patients with CHF who were not revascularized.³⁰ In general, morbidity and mortality tend to correlate inversely with EF and directly with NYHA functional class. Additional factors predisposing patients to greater morbidity and mortality include advanced age, female sex, hypertension, diabetes, and emergent operations³² (see Chapter 18). The decreases in morbidity and mortality after revascularization in this high-risk population in recent years are at least partially attributable to improvements in surgical technique and myocardial protection, but the concurrent performance of mitral valve repair and ventricular reshaping address the adverse ventricular loading conditions present and also may contribute to improved outcomes. The results of ongoing clinical trials evaluating combinations of surgical procedures (e.g., revascularization plus ventricular reshaping vs. revascularization alone) are discussed in detail later.

The importance of determining the viability of myocardium in the area to be revascularized cannot be overstated because the potential for recovery of function depends on residual contractile reserve, integrity of the sarcolemma, and metabolically preserved cellular function.³¹ Methods to detect viable myocardium include dobutamine stress echocardiography, single-photon emission computed tomography, positron emission tomography, and cardiac magnetic resonance imaging (see Chapters 1 and 2). Although dobutamine stress echocardio-graphy often has been shown to have the greatest predictive accuracy,³³ important limitations need to be taken into account. Dobutamine stress echocardiography does not demonstrate viability directly, but improvement in mechanical contraction under pharmacologic stimulation. An example for a false-negative result can be seen if there is loss of contractile proteins in the presence of preserved function of the muscle fiber membrane.³¹ Some centers report using only the intraoperative assessment of myocardial wall thickness and contractility to determine potential viability with revascularization of the target region.³⁴ Regardless of the specific method used, the important point is that there needs to be viable tissue to revascularize, and the best results will be obtained in properly selected individuals. Overwhelmingly, encouraging results worldwide suggest that, when feasible, revascularization is of benefit and provides survival advantage to patients with significant ventricular dysfunction.

Correction of Mitral Regurgitation

The mitral valve is a complex apparatus consisting of the anterior and posterior leaflets, the mitral annulus, the chordae tendineae, the papillary muscles, and the wall of the left ventricle. The posterior portion of the annulus is only rudimentarily developed and flexible. This explains why this portion of the annulus is prone to dilation during pathologic volume-overloaded states and mandates some form of mechanical stabilization when surgical valve repair is undertaken. The normal annulus has a three-dimensional (3D) saddle shape that is exacerbated during systole because of apical displacement of the commissures.³⁵ In addition, the aortic root bulges posteriorly during systole. These dynamic phenomena lead to the ability of the annulus to change its shape during the cardiac cycle. During systole, an elliptical shape is assumed that facilitates coaptation of the leaflets. During diastole, a more circular form increases orifice dimensions, decreasing resistance to LV inflow.³⁶ Consequently, the LV free wall, papillary muscles, and chordae play an important role in the competence of the valve, as well as in LV function during systole.

The mechanism responsible for mitral regurgitation can best be understood by utilizing the Carpentier classification (Table 27-3). This classification describes the motion of the mitral leaflets and position of the coaptation zone relative to the annular plane. The mitral regurgitation seen in patients with CHF is most often functional, primarily because of apical displacement of the papillary muscles resulting in tethering of the leaflets leading to systolic restriction of leaflet motion (type IIIb).

TABLE 27-3 Carpentier Classification of Mitral Regurgitation

From Carpentier A, Chauvaud S, Fabiani JN, et al: Reconstructive surgery of mitral valve incompetence: Ten-year appraisal. J Thorac Cardiovasc Surg 79:338, 1980.

Historically, many physicians considered mitral regurgitation advantageous for the failing left ventricle. It was believed that a low-pressure atrial “pop off” allowed the failing ventricle to protect itself from the high afterload of the systemic circulation and gave the illusion that the heart had a better overall contractile state than really existed. This misconception was “supported” by the fact that surgical replacement of the mitral valve was associated with a very high mortality rate in patients with depressed LV function.³⁷

Romano and Bolling’s³⁸ work disproved this misconception, showing that despite increased operative risk, mitral valve repair or replacement was beneficial to patients with severely depressed LV function, CHF, and mitral regurgitation. Operative mortality rates of 5% were reported, with 1- and 2-year survival rates of 80% and 70%, respectively.³⁸ Not only was long-term mortality reduced, but the increase in LV systolic function (on average by 10%) enabled a downgrading of NYHA class and resulted in an improved quality of life (see Chapter 19).

Although the majority of patients with end-stage HF will exhibit functional mitral regurgitation (as discussed earlier), there may be additional concurrent valvular pathology present in a given patient. An intraoperative transesophageal echocardiography (TEE) evaluation of the valvular anatomy, the mechanism of the mitral regurgitation, and direct surgical inspection will determine the feasibility of repair. Data obtained from patients suffering from organic mitral valve regurgitation have been extrapolated to the functional mitral regurgitation group in the belief that valve repair is preferable to valve replacement, because there are demonstrated hemodynamic advantages associated with preservation of the subvalvular apparatus,³⁹ and long-term anticoagulation is not required. However, review of the literature cannot unequivocally support this assumption. Magne et al⁴⁰ were unable to show a survival advantage between MV repair and replacement in patients with ischemic mitral regurgitation. Gillinov et al⁴¹ showed that a survival benefit could be obtained by repairing the mitral valve as opposed to replacing it, especially in lower-risk patients. However, in high-risk patients, defined as patients with extremely low EF and dilated ventricles, this survival benefit was lost, prompting them to recommend MV replacement.⁴¹ Braun et al⁴² showed that end-diastolic diameter larger than 65 mm was associated with poorer survival in 108 patients undergoing restrictive mitral annuloplasty and CABG. Reverse remodeling was seen in all patients with end-diastolic diameter smaller than 65 mm; however, it was seen in only 25% with end-diastolic diameter larger than 65 mm, making them question whether MV repair/CABG is justified in grossly dilated ventricles. Most likely a ventricular solution will need to be considered in this subset of patients.

A common dilemma that is encountered by the perioperative team is what to do with a patient with ischemic cardiomyopathy and moderate mitral regurgitation. In the setting of severe mitral regurgitation, most would agree to perform concurrent mitral regurgitation repair, and in the setting of mild mitral regurgitation, to leave the mitral valve dysfunction unaddressed. Penicka et al’s⁴³ study potentially could help shed light on this topic. They found improvement in moderate ischemic mitral regurgitation only in patients in whom viable myocardium in the region of the papillary muscles could be identified by preoperative single-photon emission computed tomographic testing. Patients with nonviable myocardium and dyssynchrony between the papillary muscles showed postoperative worsening of ischemic mitral regurgitation. Consequently, the authors argued to perform annuloplasty in all patients with nonviable myocardium and perform isolated CABG in patients with viable myocardium.⁴³

The primary repair technique in the setting of type IIIb dysfunction consists of downsized annuloplasty. Although much debate exists regarding whether the annuloplasty should be performed with the aid of a complete semirigid ring or band, most experts agree that the annulus needs to be remodeled and stabilized. Although this technique results in excellent short-term results, most surgeons realize that addressing a ventricular problem at the annular level cannot represent the ideal solution. Magne et al⁴⁴ showed that the use of annuloplasty rings eliminated severe mitral regurgitation; however, this was at the cost of leaving patients with functional mitral stenosis. One valvular disorder could be exchanged for another by undersizing the annulus (see Chapter 19).

After the procedure, TEE is used to assess the adequacy of the repair and potential improvement of overall cardiac function. Often, hemodynamic improvement is not immediately apparent, and TEE is used to optimize preload and guide pharmacologic interventions (see Chapters 12 to 14, 32, and 34).

Left Ventricular Restoration

In 1985, building on work by Cooley, Jatene, and others,^45–47 Dor et al introduced a ventricular reshaping procedure intended to improve systolic performance by excluding akinetic/dyskinetic and aneurysmal portions of the left ventricle with a circular stitch at the transitional zone between contractile and noncontractile myocardium. A small patch was used within the ventricular cavity as needed to reestablish ventricular wall continuity at the level of the purse-string suture.⁴⁸

This procedure generally was performed in conjunction with revascularization (and included mitral repair/replacement as needed) and thus helped establish important principles for a surgical ventricular restoration: revascularization of ischemic myocardium, decreasing of ventricular volume, and the restoration of ventricular shape. A subsequent publication by the same group described the results of this procedure in 130 patients (35% of whom had “heart failure” as the indication for the operation) and reported a 6% in-hospital mortality rate and a 3% late mortality rate because of recurrence of cardiac failure.⁴⁹

In 1996, Batista et al⁵⁰ introduced a procedure for the reshaping of the nonischemic, dilated, and failing left ventricle of NYHA Class IV patients through resection of a wedge of normal myocardium from the LV apex to the base (laterally, between the papillary muscles). This “partial left ventriculectomy” restored more normal ventricular geometry and decreased wall tension. Functional mitral regurgitation also was addressed during the Batista procedure by a mitral valve replacement or repair. Although many patients did benefit initially from this procedure (reduction of NYHA functional class to NYHA Class I in 57% and NYHA Class II in 33.3%),⁵¹ perioperative mortality was high (20% in both Batista’s own series and in the large Cleveland Clinic experience).^51,52 In addition, the experience of several centers was that many patients required rescue mechanical circulatory assistance after the procedure, and many patients experienced a redilatation of their left ventricle resulting in a return to NYHA Class IV status.^53–55 Thus, despite the short-lived period of initial enthusiasm in the mid- to late-1990s, the Batista procedure essentially has been abandoned. The concept of ventricular reshaping, however, remains of interest.

The modified Dor procedure (endoventricular circular patch plasty) has been used successfully to reshape large, dilated, spherical left ventricles of patients who have had an anterior wall myocardial infarction (MI) with resulting aneurysm and akinesis/dyskinesis. Essentially, a Dacron patch is placed within the LV cavity to exclude the large akinetic/dyskinetic area of the anterior wall. This restores LV geometry to a more normal elliptical shape and improves systolic function. When performed concurrently with CABG, significant early and late improvements in both NYHA functional class and EF have been demonstrated, with an in-hospital mortality rate of 12%.^56,57 A trial of 439 patients undergoing this procedure found an improved in-hospital mortality rate of 6.6% and an 18-month survival rate of 89.2%. In this series, CABG was performed concurrently in 89%, mitral valve repair in 22%, and mitral valve replacement in 4%.⁵⁸

Surgical ventricular restoration is the modern name for a modified Dor procedure, in which, in addition to exclusion of the akinetic aneurysmal segment of the anterior and/or septal wall and revascularization, a sizing balloon (or other sizer) is used to create an elliptical left ventricle of 30% to 40% smaller volume than baseline. Despite a worldwide surgical ventricular restoration database already in place including more than 5000 surgical ventricular restoration patients, a randomized, controlled trial was designed named “The Surgical Treatment for Heart failure (STICH),” with the intention to scientifically prove the efficacy of this procedure as an adjunct to revascularization. The results of this important trial have been eagerly awaited for years. Unfortunately, STICH did not demonstrate that the addition of surgical ventricular restoration to revascularization was associated with a greater improvement in symptoms or exercise tolerance, or with a reduction in the rate of death or hospitalization for cardiac causes when compared with revascularization alone.⁵⁹ The negative results from this trial have caused much controversy and the validity questioned.⁶⁰ Noted areas of concern revolve around participant eligibility for anatomic reasons, lack of standardized volumetric assessments in large numbers of participants, and alteration of the primary objective after trial commencement. In essence, the expert opinion is that the STICH trial was “well conceived, but poorly executed.” At the time of this writing, surgical ventricular restoration continues to be performed for true LV aneurysms but has fallen out of favor as a routine adjunct for the dilated end-stage left ventricle.

Cirillo’s⁶¹ recent publication described a further modification of the Dor concept wherein the shape and orientation of the patch, as well as the manner of suturing, ostensibly maintained more physiologic orientation of the myocardial fibers. Named the “KISS” procedure (Keep fIbers orientation with Strip patch reShaping), this methodology reportedly allowed for a return of normal apical rotation and ventricular torsion to optimize systolic performance. Long-term results of this new approach await further study.

Cardiac Resynchronization Therapy and Implantable Cardioverter-Defibrillators

The progression of disease resulting in advanced cardiac failure is typically accompanied by conduction defects and arrhythmias, and pacemakers and implantable cardioverter-defibrillators (ICDs) are common in this population. In addition to the well-known defects in sinus or atrioventricular node function, intraventricular conduction defects delay the onset of RV or LV systole in 30% to 50% of patients with advanced HF.^62,63 This lack of coordination of LV and RV contractions further impairs CO^64–67 and has been reported to increase the risk for death in this population^68,69 (see Chapters 4 and 25).

Cardiac resynchronization therapy (CRT) entails biventricular pacing (not to be confused with dual-chamber atrioventricular sequential pacing) to optimize the timing of RV and LV contractions. The right atrium is paced by a lead in the right atrium, the right ventricle by a lead in the right ventricle, and the left ventricle by a lead in a coronary vein (accessed via the coronary sinus). Although CRT is more an interventional cardiology procedure than a surgical procedure per se, anesthesiologists frequently are asked to provide sedation (if not general anesthesia) for these sometimes lengthy implantation procedures.

Studies have shown that atrial-synchronized biventricular pacing (pacing the left and right ventricles in a carefully timed manner) can resynchronize RV and LV contraction, improving CO and overall hemodynamics. This enhances these patients’ ability to exercise, which improves their NYHA functional class, and decreases the length and frequency of their hospitalizations, which improves their quality of life.^70–74

Sudden death from ventricular fibrillation accounts for approximately 350,000 deaths annually in the United States.¹ Patients with advanced HF experience ventricular fibrillation with a frequency six to nine times that of the general population,¹ and ventricular fibrillation causes 40% of all deaths in this population even in the absence of apparent disease progression based on symptoms.⁷⁵ Thus, ICDs commonly are indicated for patients with advanced cardiac failure. An ICD is a device capable of arrhythmia detection and automatic defibrillation. ICDs successfully terminate ventricular fibrillation in greater than 98% of episodes, and studies have demonstrated that an ICD increases survival and decreases the risk for sudden death in patients with ischemic cardiomyopathy and decreased LV function.^76–78

The COMPANION trial (Comparison of Medical Therapy, Pacing, and Defibrillation in Heart Failure) studied 1500 patients with NYHA Class III/IV HF, a QRS interval of greater than 120 milliseconds, PR interval greater than 150 milliseconds, and an LVEF less than or equal to 35%. Compared with optimal pharmacologic therapy alone, CRT decreased the risk of the combined end point of death from or hospitalization for HF by 34%. The combination of CRT and ICD implantation reduced these risks by 40%.⁷⁹

The CARE-HF (Cardiac Resynchronization in Heart Failure) trial enrolled 800 patients with NYHA Class III/IV HF, a QRS interval of more than 150 milliseconds, a QRS interval of more than 120 milliseconds with echocardiographic evidence of dyssynchrony, and an LVEF of 35% or less. Compared with optimal medical therapy alone, CRT (without ICD functionality) reduced all-cause mortality by 36%. In addition, CRT showed significant improvement of cardiac dyssynchrony, ventricular function, and mitral regurgitation based on echocardiographic criteria.⁶⁸

Most recent guidelines give a Class I recommendation for placement of a CRT device with or without ICD in patients with LVEF ≤ 35%, QRS ≥ 120 milliseconds, presence of sinus rhythm, and NYHA Class III/IV symptoms on optimal medical therapy.⁸⁰ However, two recent trials have investigated the effects of CRT on patients with NYHA Class I/II HF. The REVERSE trial (Resynchronization Reverses Remodeling in Systolic Left Ventricular Dysfunction) randomized 610 NYHA Class I/II patients with QRS ≥ 120 milliseconds and LVEF ≤ 40% to receive a CRT device (±ICD) that was either active (CRT-ON) or disabled (CRT-OFF). This study showed a significant delay in time to first hospitalization and improvement in measures of LV remodeling in the CRT-ON group.⁸¹ The MADIT-CRT trial (Multicenter Automatic Defibrillator Implantation Trial–Cardiac Resynchronization Therapy) randomized 1820 NYHA Class I/II patients with QRS ≥ 130 milliseconds and LVEF ≤ 30% to receive CRT and ICD or ICD alone. The CRT-ICD group had a significant 41% decrease in risk for first HF event, significant reduction in LV volumes, and increase in LVEF as compared with the ICD only group.⁸² Thus, although CRT previously has been about reduction in symptoms, it is anticipated that results such as those reported from MADIT-CRT will lead to an increased utilization of CRT in relatively asymptomatic patients with developing HF to prevent the progression of the disease.

Ventricular Assist Devices: Implementation of Support

VADs can be used to take over the pumping function of the failing ventricle and provide effective CO to the arterial circulation downstream from the failing ventricle. In the case of the failing right ventricle, a right ventricular assist device (RVAD) can divert the deoxygenated venous return to the heart and pump it directly to the pulmonary arterial circulation. In the case of the failing left ventricle, a left ventricular assist device (LVAD) can divert the oxygenated blood returning to the left side of the heart and pump it directly into the aorta. Figure 27-1 demonstrates common cannulation strategies in the heart and great vessels by which MCS is implemented. Although there are a few notable exceptions (discussed later), these basic cannulation strategies are common to all manufacturers’ devices currently in use, regardless of the type of output they produce.

Figure 27-1 The Thoratec ventricular assist device (VAD) and three cannulation approaches for univentricular left-heart support (A) and biventricular support (B, C). Potential cannulation strategies for VADs are shown. B, The most commonly used cannulation approaches for most currently available VADs. Cannulas are placed in the heart and great vessels to divert blood returning to the failing side of the heart to the pump. The blood collected in the VAD blood chamber is then ejected into the arterial circulation immediately downstream of the failing ventricle. AO, aorta; Apex, left ventricular apex; IAG, cannula inserted via the interatrial groove and directed toward the LA roof; LA, left atrial appendage; PA, pulmonary artery; RA, right atrium.

(Reproduced from Thoratec Corporation, Pleasanton, CA.)

In addition to maintaining the circulation, decompression (or emptying) of the failing ventricle provides additional benefit because decompression will decrease the wall tension (and, therefore, myocardial oxygen demand) in the volume- and pressure-overloaded failing ventricle, potentially allowing for recovery of ventricular function. This is particularly relevant to short-term support of the acutely failing ventricle, but recovery of function (to varying degrees) also can be seen with chronically failing ventricles.

Mechanical Circulatory Support: Modern Practice

In prior years, VADs were regarded as either a last-ditch hope for recovery after an acute cardiac event that resulted in refractory cardiogenic shock (“bridge-to-recovery”), or as a “bridge-to-transplantation” for patients with chronic HF who were doing poorly. This has all changed now. More widespread acceptance of MCS, the recognition that outcomes are better when the support is instituted earlier rather than later, the development and use of risk scores to appropriately select patients and support strategies,^83–94 the demonstration that there is often significant improvement in multiorgan function during the time spent on VAD support,^95–97 the advent of new and improved devices, and a substantially decreased rate of VAD-associated complications^98–101 have allowed for an increased and an improved clinical utilization of this life-saving technology in the management of both acute and chronic HF.

Short-Term Ventricular Assist Device Use

The main indication for the short-term use of VADs always has revolved around the concept that the acutely failed myocardium could potentially recover after a short period of support. Depending on the cause of the acute failure (e.g., ischemic stunning, viral myocarditis), early experience revealed that myocardial recovery typically took a week or two in the majority of well-managed patients, who were then theoretically able to be weaned from support. Most short-term VAD use in prior decades was with extracorporeal, pulsatile devices, although standard centrifugal pumps producing continuous flow were also occasionally used, especially for the pediatric population and for extracorporeal membrane oxygenation (ECMO). In reality, however, regardless of the device used as a bridge-to-recovery, support often was initiated too late, complications (e.g., bleeding, thromboembolism, infection, multisystem organ failure) were frequent, and the percentage of patients who were able to be weaned from support (much less discharged from the hospital) was often disappointing.

Early experience also demonstrated that specific patient conditions or comorbidities (Table 27-4) potentially increased the risk for complications during VAD support or made support difficult to implement. It also was proposed that patients who were not potential transplant candidates should probably not be considered for VAD support. Finally, it commonly was believed for many years that patients unlikely to survive regardless of the reestablishment of effective systemic perfusion should probably not even be considered for temporary VAD support. However, society currently demands that everyone (no matter how critically ill) be given an opportunity to recover, and the new and improved MCS technology is often now used with the understanding that improvement in clinical status will allow for a continuation of support, whereas a worsening of the clinical picture will prompt a discontinuation of support. Nevertheless, it is the advancements of the technology and an increased patient management experience that have brought a new flexibility to the arena of short-term MCS.

TABLE 27-4 Conditions or Comorbidities That Make Ventricular Assist Device Placement or Use Difficult, Make the Patient More Likely to Have Major Complications, or Make Meaningful Recovery Unlikely

BSA, body surface area; ESLD, end stage liver disease; ESRD, end stage renal disease.

The traditional concept of the temporary or short-term use of a VAD as a “bridge-to-recovery” has thus been expanded to include concepts such as “bridge-to-immediate survival,” “bridge-to-next decision,” and “bridge-to-a-bridge.”

Figure 27-2 The Tandem Heart percutaneous Ventricular Assist Device (pVAD).

This recently introduced device provides circulatory support with a centrifugal pump head and the potential advantages of percutaneous cannula insertion. Partial left-heart bypass is accomplished via an interatrial septal cannulation into the left atrium across the area of the fossa ovalis, and oxygenated blood is returned to the femoral artery.

(Courtesy of CardiacAssist Inc., Pittsburgh, PA.)

Figure 27-3 The Impella.

The Impella 2.5 is a minimally invasive, catheter-based cardiac assist device designed to unload the left ventricle (thereby reducing myocardial oxygen consumption), while increasing overall cardiac output to maintain systemic perfusion. The Impella 2.5 can be percutaneously inserted into the left ventricle from the femoral artery. The tip of the catheter rests in the left ventricle during support, generating flows up to 2.5 L/min in the ascending aorta.

(Reproduced from Abiomed, Inc., Danvers, MA.)

Figure 27-4 A, The Abiomed BVS5000. B, Biventricular support with the Abiomed BVS5000. Cannulas are placed in the heart and great vessels to divert blood to the extracorporeal BVS5000 pumps. Blood collected in the pumps is ejected into the ascending aorta (left ventricular assist device [LVAD]) and main pulmonary artery (right ventricular assist device [RVAD]). Note that the pumps fill by gravity drainage and are, therefore, generally mounted below the level of the heart. Adjustments can be made to pump height to optimize filling and pump ejection. For example, lowering a pump can augment device filling but will increase the afterload against which the device must eject. Conversely, raising a pump can facilitate pump emptying but will decrease filling of the device.

(Reproduced from Abiomed Corporation, Danvers, MA.)

Figure 27-5 The Thoratec pVAD (percutaneous ventricular assist device).

(Reproduced from Thoratec Corporation, Pleasanton, CA.)

Figure 27-6 The Abiomed AB5000 ventricle.

Hookups for the inflow and outflow cannulas are at the right. Although arrows indicate the direction of flow, they also serendipitously point at the artificial unidirectional valves that mandate anticoagulation during support with this device. The driveline, which alternately provides vacuum to assist filling and compressed air to accomplish ejection, is at the left.

(Reproduced from Abiomed Corporation, Danvers, MA.)

Figure 27-7 The Thoratec CentriMag.

(From Thoratec Corporation, Pleasanton, CA.)

Though useful conceptually, the apparent distinctions between these indications can sometimes appear somewhat artificial in the clinical arena, and the current practice of short-term MCS is really more of a continuous spectrum beginning with the immediate survival of the patient.

Bridge-to-Immediate Survival

Since the late 1960s, the IABP has been the most commonly used “bridge-to-immediate survival” because it simultaneously increases myocardial O₂ supply and decreases O₂ demand, interrupting the otherwise inexorable cycle leading to ventricular failure. It is estimated that 5% to 10% of patients will experience development of cardiogenic shock after an acute MI, and early survival rates for these patients are on the order of 5% to 21%.¹⁰² However, 75% of such patients who are unresponsive to pharmacologic interventions will exhibit hemodynamic improvement with IABP therapy alone,¹⁰³ and early survival rates in these patients are reported to approach 93% when treated with IABP counterpulsation.¹⁰⁴

Although a balloon pump can improve the output from an acutely stunned ventricle, it only can augment forward CO by about 25% to 30% at maximum depending on the afterload,^105,106 and it is clear that there will be no augmentation of forward CO if there is a complete absence of LV function; thus, by itself, an IABP cannot be expected to rescue a patient from catastrophic myocardial failure (see Chapter 32).

ECMO has heretofore been the mainstay of emergent temporary MCS when there is intractable cardiorespiratory failure, and ECMO is making a comeback as an integral part of resuscitative protocols in many institutions, but the development of effective devices that can be deployed rapidly, at the first recognition of refractory ventricular insufficiency, has made “bridge-to-immediate survival” more of a reality than ever.

A key determinant of the overall survivability of acute cardiac failure is the rapidity with which the failing ventricle can be decompressed and resumption of adequate systemic perfusion assured. Although immediate implementation of support with conventional short-term VADs could potentially achieve these goals, sternotomy and CPB typically are utilized for the implantation of the requisite cannulae in the heart and great vessels. Furthermore, delays may be experienced while awaiting the availability of the operating room and necessary perioperative staff, prolonging the period of time during which the failing ventricle is pressure and volume overloaded, and the splanchnic beds and peripheral tissues are underperfused. It was considerations such as these that drove the development of several new and innovative short-term assist devices. Once immediate survival is assured, such a device conceivably can be changed later to another capable of providing longer term support.

Bridge-to-Recovery, Bridge-to-Next Decision

It is now well-known that the success of “bridge-to-recovery” with a VAD hinges on appropriate patient selection and prompt intervention. If the myocardium is going to recover after an acute insult, it generally tends to do so within a week or two (although the process may take longer than a month in some patients), and the patient may then be weaned from MCS. Although dismal in the past, rates of successful weaning from short-term support have improved. In the postcardiotomy cardiogenic shock population, survival rates approaching 50% have been reported with the Abiomed BVS5000 and AB5000 ventricle in “experienced centers with well-defined protocols for patient selection and timing of intervention.”^122,123 Survival from acute MI cardiogenic shock has been reported at 42% with the AB5000 ventricle (mean number of days supported, 25.4) and 27% with the BVS 5000 (mean number of days supported, 5.2).¹²⁴ Limited information is available at the time of this writing regarding current rates of recovery from short-term support with the new devices (discussed later), but the strategies of support have changed dramatically in recent years and current data may no longer be directly comparable with the pure “bridge-to-recovery” strategy used in prior decades.

Regardless, prompt intervention to restore adequate systemic perfusion and careful patient selection are the key considerations if a patient is to be successfully bridged to recovery with any VAD. Despite the new technology currently available and the new strategies surrounding mechanical circulatory assistance, any patient who is unlikely to survive regardless of the reestablishment of effective systemic perfusion should not even be considered for VAD support. In contrast, if the myocardial injury is deemed likely to recover and the clinical assessment is otherwise favorable, VAD support may be considered. Although time is of the essence, a balance must be struck in the decision-making process and clinicians cannot wait until cardiogenic shock has resulted in multisystem organ failure before initiating mechanical circulatory assistance because experience has shown that the patient is unlikely to survive.

Early information¹²⁵ based on the experience with the Abiomed BVS5000 indicated that the best outcomes tended to occur in the following situations:

Poor outcomes occur in the following situations:

Although the use of the BVS5000 device has decreased in recent years, the experience with this pioneering “bridge-to-recovery” device formed the basis of the current understanding that careful patient selection and prompt intervention to restore adequate systemic perfusion are critical if a patient is to be successfully bridged to recovery with MCS.

Additional considerations revolve around whether univentricular or biventricular support is required, but often this is a decision now made in retrospect in the arena of short-term support, once LVAD support has been engaged and the clinical situation is reassessed. Severe RV dysfunction has been reported to occur in up to 30% of LVAD-supported patients^126,127 because of unfavorable alterations in RV geometry (e.g., leftward shift of the interventricular septum) resulting in increased RV compliance and decreased RV contractility in the presence of increased RV preload¹²⁸ and potentially increased RV afterload. Although the overall incidence of RV failure in the setting of LVAD support appears to be decreased with continuous flow not intended to completely decompress the left ventricle during support, the perioperative management of RV preload and afterload continues to play an enormous role in the potential for RV dysfunction after implementation of LVAD support. The issue of RV failure during LVAD support is more of an issue during long-term support and is discussed in more detail later.

When severe cardiopulmonary failure is present, and there is uncertainty about the recoverability of the situation (or the neurologic status of the patient), ECMO again has become a popular circulatory support strategy.¹²⁹ ECMO has significant disadvantages, however, including a somewhat limited potential duration of support, the need for dedicated personnel to manage the flows and the anticoagulation, and a high incidence of complications as the duration of support increases.

Until recently, pulsatile, extracorporeal devices such as the Abiomed BVS5000 (FDA-approved in 1992), the Thoratec pVAD (FDA-approved in 1998), and the more recently approved AB5000 ventricle (FDA-approved in 2003) were the standard short-term bridge-to-recovery devices for patients with refractory cardiac failure. However, advances in technology and device engineering have supplanted the time-honored, pulsatile devices, and the CentriMag rapidly is becoming the short-term support device of choice.

Long-Term Ventricular Assist Device Use

Permanent replacement of the failing heart was the original intent of the research and development in the field of mechanical circulatory assistance, and this dream is alive and well in the latest version of the total artificial heart, the AbioCor Implantable Replacement Heart (Abiomed; Figure 27-8), but most long-term VAD use has been as a bridge-to-transplantation.

Figure 27-8 A, The AbioCor Implantable Replacement Heart. B, Orthotopic implantation of the AbioCor Implantable Replacement Heart. The native failed heart is removed, and the AbioCor is implanted orthotopically, anastomosed to cuffs of native atria and the great vessels. Transcutaneous energy transfer technology eliminates the need for percutaneous wires.

(From Abiomed, Inc., Danvers, MA.)

The traditional concept of the intermediate or long-term use of a VAD as a “bridge-to-transplantation” has now been expanded to include the concept of “destination therapy” (DT), a final management strategy for end-stage HF in patients who are not transplant eligible. However, as mentioned earlier, time spent on VAD support often improves multisystem organ function and can potentially convert high-risk transplant-ineligible patients into transplant-eligible patients. Thus, another intermediate or perhaps long-term use for VADs might be termed “bridge-to-improved candidacy.”

Current FDA-approved devices used as a bridge-to-transplantation in the United States include the Thoratec pVAD (Thoratec Laboratories; see Figure 27-5), the HeartMate XVE (Thoratec Corporation, Woburn, MA; Figure 27-9), the IVAD (implanted vascular assist device; Thoratec Laboratories; Figure 27-10), the HeartMate II (Thoratec Laboratories; Figure 27-11), and the CardioWest TAH (total artificial heart; SynCardia Systems, Tucson, AZ; Figure 27-12). Currently, the HeartMate XVE and the HeartMate II are the only devices FDA approved in the United States for DT. The Novacor LVAS (World Heart, Ottawa, Canada) was approved as a bridge-to-transplantation in 1998 but is no longer being implanted. Despite arguably superior engineering regarding device longevity in comparison with other devices available at the time, as well as comparable rates of successful bridging to transplantation, the Novacor was associated with a high incidence of thromboembolic complications. Table 27-6 summarizes the basic characteristics of devices used for long-term support.

Figure 27-9 The HeartMate XVE.

(From Thoratec Corporation, Pleasanton, CA.)

Figure 27-10 The IVAD (implanted vascular assist device).

(From Thoratec Corporation, Pleasanton, CA.)

Figure 27-11 The HeartMate II.

(From Thoratec Corporation, Pleasanton, CA.)

Figure 27-12 A, The CardioWest Total Artificial Heart. B, Orthotopic implantation of the CardioWest Total Artificial Heart. The failed native heart is removed, and the CardioWest TAH is implanted orthotopically, anastomosed to cuffs of native atria and the great vessels.

(Reproduced from SynCardia Systems, Tucson, AZ.)

Intermediate- or long-term VAD support potentially is indicated as a bridge-to-transplantation in situations in which no myocardial recovery is expected (e.g., end-stage cardiomyopathy) or when an acutely stunned or infarcted left ventricle fails to recover despite support with a short-term VAD. By providing effective CO in place of the failed native heart, this technology can stave off the end-organ damage resulting from a rapidly deteriorating CO and allows severely decompensated transplant-eligible patients to potentially survive long enough to receive a donor heart. An additional benefit of this application of VADs is an improved quality of life, often as an outpatient, while awaiting a new heart. Further, significant improvements in multiorgan function have been demonstrated during the time spent on VAD support,^95–97 and it is unusual nowadays for a patient to present for heart transplantation without an LVAD in situ.

The HeartMate I (World Heart, Oakland, CA; see Figure 27-9) has heretofore been the most commonly used bridge-to-transplantation device for patients with advanced LV failure. The original pneumatically powered HeartMate IP was FDA-approved for this indication in 1994, and the electrically powered version (the VE, vented electric) was approved in 1998. Data from prior years indicated a 67% success rate of bridging-to-transplantation with the HeartMate VE.¹³⁵ The major advantage of the pulsatile HeartMate LVAS always has been the “antithrombogenic” lining of its blood chamber that obviated the need for formal anticoagulation with warfarin once a neointima was established. The major disadvantages of the HeartMate included infections of the large percutaneous lead and in the preperitoneal pocket where the device was implanted, as well as a limited durability beyond 18 months.

The current incarnation (the XVE) is the result of improvements to the device, and it has been in use since approximately 2002. According to the manufacturer, more than 4500 patients have been implanted with the HeartMate XVE in 186 centers worldwide. The average age has been 51 years old, with a range from 8 to 74, and the longest duration of support (ongoing patient on one device) has been 1854 days.¹³⁶

Patients with biventricular failure generally have been bridged to transplantation with the Thoratec pVAD (FDA-approved for this indication in 1995) or, potentially, the IVAD (the implantable, titanium-coated version of the Thoratec device). According to the manufacturer, more than 4000 patients have been supported by the Thoratec pVAD at more than 240 medical centers in 26 countries. The longest duration of support is reportedly 1204 days, with 858 of those patients discharged to home, and the rate of successful bridge-to-transplantation is reported at 69%.¹³⁶ For the IVAD, more than 500 patients reportedly have been implanted at 95 medical centers in 9 countries, with the longest duration of support being 979 days,¹³⁶ and a reported rate of successful bridge-to-transplantation of 69%.¹³⁷

The CardioWest total artificial heart (SynCardia Systems; see Figure 27-12) is available in select centers internationally as a bridge-to-transplantation for patients with biventricular failure, and a resurgence of interest in this device has been seen recently. A successful bridge-to-transplantation rate of 79% was observed in prior years.¹³⁸

Complications such as infection of percutaneous drivelines, perioperative bleeding, RV failure, sepsis, and multisystem organ failure were always an enormous issue with the first-generation, pulsatile VADs. Thromboembolism was less prevalent with the HeartMate because of the antithrombogenic lining of its blood chamber, but it did occur. A review of 228 patients on long-term support with the Thoratec, Novacor, and HeartMate as a bridge to transplantation reported cerebral embolism in 24%, 39%, and 16%, respectively, despite adherence to recommended anticoagulation protocols.¹³⁹

Updated information from recent clinical experiences is not available for these first-generation devices at the time of this writing, but it may be moot because the nonpulsatile HeartMate II was approved as a bridge-to-transplantation in April 2008, and it rapidly has become an intermediate and long-term support device of choice.

Permanent Ventricular Assist Device Use: Destination Therapy

A relatively new indication for long-term VAD use is called destination therapy (DT), which refers to the intentionally permanent use of an LVAD as a permanent management solution for end-stage cardiac failure. From a certain perspective, the field of MCS has come full circle because the original concept of ventricular support devices was for the permanent replacement of the failing heart. From a certain perspective, this dream is alive and well in the latest version of the total artificial heart, the AbioCor Implantable Replacement Heart discussed later.

The REMATCH trial (Randomized Evaluation of Mechanical Assistance for the Treatment of Congestive Heart Failure) established that the use of a left-sided VAD was not only an effective tool to treat patients with advanced HF but resulted in more than twice the survival rate and an improved quality of life in comparison with optimal medical management.¹⁴⁶ Based on the results of REMATCH, the FDA approved the HeartMate VE in November 2002 for transplant-ineligible patients as DT, but improvements to the HeartMate device resulted in the HeartMate XVE by the time FDA approval was granted, and it is the XVE that has been implanted for DT patients in the United States. Although use of the XVE quickly was associated with fewer adverse events and a better survival rate than in the REMATCH trial,¹⁴⁷ increased experience with patient management likely also played a role in the observed improved outcomes.

To date, DT has been indicated only for transplant-ineligible patients, but it may be anticipated that this indication may be broadened going forward because although cardiac transplantation remains the gold standard therapy for end-stage disease, the number of donor organs is severely limited in comparison with the number of patients who would benefit, and most patients with end-stage cardiac disease cannot realistically expect to be transplanted. Another factor that may lead to a potential broadening of the indication is the FDA approval of the HeartMate II for DT in January 2010, after the completion of the HeartMate II DT pivotal clinical trial.¹⁴⁸ In this prospective, randomized trial, the HeartMate II was pitted against the HeartMate XVE on a 2:1 basis. Not only was the incidence of adverse events, including infection, sepsis, right-heart failure, and mechanical problems, significantly lower in HeartMate II patients compared with patients implanted with the XVE, but HeartMate II patients experienced shorter hospital stays. After approval of the HeartMate II for DT, the number of patients implanted with the XVE as DT has dropped off significantly, though the VXE may remain useful for patients who cannot be anticoagulated.

New Risk Scores: Optimizing Survival while Minimizing Risk during Long-Term Ventricular Assist Device Support

Just as with the short-term devices used as a short-term bridge-to-recovery, it has been demonstrated that the best outcomes with long-term devices used as DT are obtained through careful patient selection and earlier implantation. Large retrospective analyses have allowed for a new understanding of how specific clinical derangements play out in terms of perioperative morbidity and mortality and have provided insight into how intensive medical therapy can optimize patients before VAD use.

Ventricular Assist Device Use and the Potential for Myocardial Recovery

In cases of acute ventricular failure, removal of the sudden volume and pressure overload and improvement in the balance of myocardial oxygen supply and demand may allow for myocardial recovery and weaning from mechanical support. Clinical experience has shown that if recovery from an acute insult is going to occur, it generally does so over the course of 1 to 2 weeks. In cases of long-standing, progressive failure, decompression of the dilated and chronically failing left ventricle by an LVAD also may allow for some degree of recovery over time.

Although echocardiographic and histologic support for this general premise has been available since the mid 1990s,¹⁵⁶ the underlying biochemical mechanisms of remodeling (and its reversal) are only now being elucidated. Patten et al showed that therapy with a VAD normalizes inducible nitric oxide synthase expression in association with decreased cardiomyocyte apoptosis.¹⁵⁶ Decompression of the left ventricle by an LVAD has been reported to allow for normalization of LV geometry, regression of myocyte hypertrophy,¹⁵⁷ favorable changes in LV collagen content,¹⁵⁸ and normalized expression of genes controlling excitation-contraction coupling and the calcium content of the sarcoplasmic reticulum.¹⁵⁹ It has been reported that maximum structural reverse remodeling is complete after around 40 days of LV decompression, with reversal of some of the molecular aspects of remodeling by 20 days.¹⁶⁰

It should be appreciated, however, that although a great deal of knowledge regarding myocardial remodeling (and its reversal) has been elucidated, long-lasting ventricular recovery after chronic HF is a relatively infrequent event, reportedly occurring primarily in patients with dilated cardiomyopathy, and the actual number of patients who recover sufficiently to have their long-term VAD explanted is small. Worldwide efforts to understand the nature of adverse ventricular remodeling and its reversal are ongoing, and the relevant literature is rapidly burgeoning.

Abiocor Implantable Replacement Heart

From the original pneumatically driven devices with their massive external control consoles to the totally implantable computer-controlled AbioCor Implantable Replacement Heart (Abiomed) now in clinical trials, the mechanical TAH has been the subject of intensive research and development for decades.

The first TAH was a pneumatically driven biventricular pump developed by Dr. Domingo Liotta and colleagues in the 1960s. This device (the Liotta TAH) was implanted in a 47-year-old patient with severe HF by Dr. Denton Cooley on April 4, 1969, and was used for 64 hours as a bridge-to-heart transplantation.¹⁶⁰ The patient died of Pseudomonas pneumonia 32 hours after his transplantation, but the Liotta heart proved that a mechanical device could be successfully used clinically to sustain a patient.

The second human implantation of a TAH also was performed by Dr. Cooley. In July 1981, the Akutsu III TAH was used successfully for 55 hours as bridge-to-transplantation in a 36-year-old patient with end-stage HF.¹⁶¹ The Jarvik-7 TAH was first implanted as a permanent replacement heart in August 1985 in a 61-year-old man with primary cardiomyopathy and chronic obstructive pulmonary disease.¹⁶² Although the patient survived only 112 days, the duration of his survival was encouraging. Since 1991, the Jarvik-7 has been known as the CardioWest TAH. As discussed earlier, this device is still in use today as a bridge-to-human heart transplantation in selected centers in the United States, France, and Canada.

The AbioCor Implantable Replacement Heart (see Figure 27-8) represents a major advance in artificial heart technology because it is truly totally implantable; there are no percutaneous cables, conduits, or wires. The device is motor driven, so a source of compressed air to drive the pumping action is not required, allowing patients complete mobility. The device itself weighs approximately 2 pounds and is orthotopically implanted. Transcutaneous energy transfer is used (in lieu of a percutaneous cable) to supply the motor-driven hydraulic pumping of the artificial ventricles with power and system control. Artificial unidirectional valves within the device mandate anticoagulation during support.

Initial implantations of the AbioCor were performed in 14 patients between 2001 and 2003. Although the device performed well, with one patient supported longer than 1 year, there were a lot of strokes and only one patient actually was discharged to home. Results from the initial implants were submitted to the FDA in September 2004 for marketing approval, but the FDA ultimately denied the application, citing concerns about patient inclusion criteria and device labeling, potential benefit versus risk of the device, anticoagulation protocols, and quality-of-life versus quantity-of-life issues.

However, because of its limited market, in September 2006, the AbioCor was FDA approved as a Humanitarian Use Device. Currently, the AbioCor Implantable Replacement Heart is FDA approved for patients younger than 75 years with end-stage biventricular failure who are not transplant eligible and who cannot be treated with an LVAD alone as DT. No new information is available at the time of this writing regarding recent implantations of the AbioCor Implantable Replacement Heart.

Cellular Transplantation into the Myocardium

A novel approach to treating severe systolic dysfunction is the injection of harvested autologous skeletal muscle cells into the failing myocardium. This procedure can be performed either surgically at the end of a revascularization procedure or percutaneously in the catheterization laboratory.

The use of cell transplantation is not a novel approach to treating disease. Skin and bone- marrow transplantations were the first replacement therapies described. These were followed by embryologic neuron transplantation in patients with Parkinson’s disease and islet cell transplantation for the treatment of diabetes mellitus.¹⁶⁴

The basic understanding of the remodeling process is that viable and contractile cardiomyocytes undergo apoptosis and become replaced by noncontractile tissue; this, in turn, leads to systolic and diastolic dysfunction. In an attempt to restore functionality, contractile cells are injected into this region. In clinical practice, myoblasts from the patient’s quadriceps muscle have been used. Using the patient’s own tissue has several advantages. First, the complications of pharmacologic immunosuppression are avoided. Second, there are no ethical problems in contrast with those frequently observed when fetal cells are used. Finally, the ease of harvest and processing make this tissue ideal for this purpose.

Skeletal muscle cells, however, are histologically different from native cardiomyocytes. Adhesion molecules, which are found in native cardiomyocytes, are not found in skeletal myocytes (N-cadherin and connexin-43).¹⁰⁴ These adhesion molecules are important for adhesion to the extracellular matrix and for intercellular communication.

Clinically, a Phase I study from Poland, the POZNAN trial,¹⁶⁵ showed that myoblasts can be implanted safely via a percutaneous route into a scarred region of the left ventricle. Only 10 patients were enrolled into this study. Nine were transplanted. One could not be transplanted because of technical reasons. Six patients of the nine were followed up. Improvement of NYHA class and EF were observed in all of them. Similar results were seen by Menasche et al.¹⁶⁴ A frequently encountered occurrence in the phase I studies was the fact that many patients had episodes of ventricular tachycardia after the procedure. They were successfully treated with amiodarone or electric cardioversion. Phase II studies are now in progress in the United States and Europe. Many other cell types are now being experimentally injected into the myocardium or given intravenously in an attempt to regrow cardiac myocytes.^166–168 These cells include adult bone marrow stem cells, embryonic stem cells, and cardiac progenitor cells found mainly in the atrium. They have been injected alone or with multiple growth factors such as granulocyte-macrophage colony-stimulating factor, vasoactive endothelial growth factor, and angiopoietan-1, which mobilize progenitor cells and induce new cell growth. The transplanted cells may morph into new cardiac muscle cells, or they may improve cardiac function by boosting the growth of new blood vessels, or releasing other growth factors that encourage cell proliferation and survival. Any of these effects could explain some of the early positive results seen to date.

Gene Transfer in Cardiac Myocytes

Gene therapy has received much interest by the media. Multiple research groups are trying to cure or at least alleviate symptoms brought on by CHF through gene therapy. It is necessary to inoculate the cell with DNA to change the genetic programming of a cell. Vectors are being used to achieve this goal. Vectors commonly used range from plasmid DNA to different virus types (e.g., adenovirus, herpes virus). The two cellular pathways that are being targeted are the sarcoplasmic reticulum and the β-adrenergic pathway.¹⁶⁹ The targets of gene therapy are to increase β-adrenergic function, adenylyl cyclase, the V₂ vasopressin receptor, and sarcoplasmic reticulum Ca²⁺ ATPase (SERCA2a), while also decreasing phospholamban and β-adrenoreceptor kinase (BARK 1). All of these new therapies hold great promise, and many trials are under way.¹⁷⁰

Essential Considerations

Cardiac failure can be the common outcome of a variety of underlying causative factors, and patients may, therefore, present with widely varying clinical status. Some may appear quite compromised, whereas others may appear surprisingly well-compensated despite significant underlying pathophysiology. Regardless of what is planned surgically, achieving hemodynamic stability during the induction and maintenance of anesthesia requires a preoperative knowledge of the status of the coronary arteries (including prior bypass grafts or stents), the extent of any valvular regurgitation and/or stenosis that may be present, and whether there is significant pulmonary hypertension. In addition, all patients with significant ventricular dysfunction require a thorough preoperative evaluation of the major organ systems (particularly the renal, hepatic, and central nervous systems) for impairment. The anesthetic plan can then take any existing organ dysfunction into consideration.

Preoperative Optimization

Patients with cardiac failure are medically optimized by pharmacologic manipulations of afterload, β-blockade, and diuresis. Most medications should be continued throughout the perioperative period, but the decision whether to withhold angiotensin-converting enzyme inhibitors and diuretics should be made on an individual basis. Patients with HF often require preinduction optimization of intravascular volume status, pharmacologic manipulations of inotropy and afterload, adjustments to pacemaker settings (where present), and, on occasion, elective placement of an IABP. It is prudent to provide supplemental oxygen and monitor vital signs during the preoperative period. This is especially important if anxiolytic medications are given because this population will not tolerate sudden decreases in sympathetic tone, hypoxemia, or the potentially increased pulmonary vascular resistance that may accompany a respiratory acidosis if hypoventilation results from anxiolysis.

Intraoperative Considerations

Most procedures intended to improve cardiac function will require CPB, but the potential availability of bypass cannot be taken for granted because the function of even a severely depressed ventricle can still get worse, and it is well-known that having to emergently institute CPB significantly increases morbidity and mortality. In the prebypass period, further depression of cardiac function and significant increases in ventricular afterload must be avoided because these will increase myocardial oxygen demand and may cause ischemia. Preload, afterload, heart rate, and contractility must be continuously optimized for each patient at all times.

Where CPB will be used, it is generally prudent to consider using an antifibrinolytic agent (e.g., ε-aminocaproic acid or tranexamic acid) in an attempt to decrease bleeding in the postbypass period. In some patients, CPB may not be necessary to insert an LVAD.

Anesthetic Agents and Technique

Although the usual sedative and hypnotic agents may be tolerated in patients with mild cardiac failure, the failing heart is chronically compensated by a heightened adrenergic state, and removal of that sympathetic tone may lead to rapid decompensation with cardiovascular collapse during anesthetic induction. Patients with severely decreased ventricular function tend to decompensate quickly from physiologic and hemodynamic aberrations (e.g., hypercarbia, hypoxemia, hypotension, bradycardia/tachycardia, sudden alterations in volume status, and loss of sinus rhythm); agents should be chosen and used in a manner likely to maintain hemodynamic stability. In addition, agent selection should take into account any coexisting renal or hepatic insufficiency. Intravascular volume status needs to be carefully considered and continuously optimized for each individual patient. Inotropic and vasoactive agents including ephedrine, phenylephrine, dopamine, epinephrine, milrinone, vasopressin, nitroglycerin, and nitroprusside should be available and judiciously used at the first sign of refractory hemodynamic instability.

There can be no cookbook approach to these patients. Despite the perceived similarity of one patient with cardiac failure to another, each individual’s underlying pathophysiology must be carefully considered and then anesthetic agents chosen that will best maintain the hemodynamic goals for that patient.

Traditionally, a technique based on high-dose opioid (e.g., total fentanyl dose, 50 to 100 μg/kg, or total sufentanil dose, 5 to 10 μg/kg), together with a neuromuscular blocking agent, has been used for patients with severely depressed cardiac function. Although such a technique likely will result in many hours of hemodynamic stability, potential disadvantages of this technique are that amnesia may not be adequate and the bradycardia and initial chest wall rigidity that typically accompany such an induction must be pharmacologically countered (see Chapter 9).

Etomidate (0.2 to 0.3 mg/kg intravenously) is usually the induction agent of choice in these patients because it causes neither a significant reduction in surgical ventricular restoration nor a significant decrease in myocardial contractility. The decreases in vascular tone and myocardial contractility that accompany induction with propofol make this drug unsuitable for those with severely depressed cardiac function. Similarly, thiopental, with its propensity to cause myocardial depression and venodilation, with consequent decreases in CO, is not often used for these patients.

As a general rule, high doses of the potent inhalation agents are poorly tolerated in this population. Although all of the inhalation agents (including nitrous oxide) are myocardial depressants to varying extents, enflurane and halothane are particularly potent in this regard and generally are avoided in patients with depressed ventricular function. Isoflurane, sevoflurane, and desflurane are more likely to be compatible with hemodynamic stability in the well-optimized patient, although isoflurane and desflurane must be used with caution because of their particular tendency to decrease systemic vascular resistance. In comparison with the other currently available agents, sevoflurane appears to cause less myocardial depression and decrease in surgical ventricular restoration. In addition to direct myocardial depression and vasodilation, the inhaled anesthetic agents also may affect myocardial automaticity, impulse conduction, and refractoriness, potentially resulting in reentry phenomena and arrhythmias.

Although its use in adults has decreased dramatically in recent years, ketamine remains an extremely useful agent in patients with severely decreased ventricular function. A ketamine induction (1 to 2.5 mg/kg intravenously or 2.5 to 5 mg/kg intramuscularly), followed by a maintenance infusion (50 to 100 μg/kg/min), usually will provide excellent hemodynamic stability while assuring adequate analgesia and amnesia. Where feasible, midazolam generally is provided before giving ketamine in an attempt to lessen the potential postemergence psychiatric side effects that may occur in some patients. Additional small doses (1 to 2 mg every 2 to 3 hours) or an infusion of midazolam (0.5 μg/kg/min) often are provided when a ketamine infusion is in use. For adults and older pediatric patients, a small dose of glycopyrrolate (e.g., 0.2 mg IV) generally is provided to act as an antisialagogue. Atropine (10 μg/kg) is used for this purpose in neonates and infants. Once on CPB, the ketamine infusion can be stopped, and moderate-to-high doses of narcotics administered.

Central venous access and pulmonary artery catheterization (PAC) are extremely useful (if not mandatory) in this patient population for several reasons. First, pharmacologic interventions are frequently necessary, and potent inotropic and vasoactive agents are preferably administered to the circulation through a central route. Second, despite recent controversy regarding the usefulness and potentially increased morbidity and mortality with a PAC in critically ill patients, the ability to follow and optimize trends of CO and other hemodynamic indices, as well as the ability to assess the efficacy of pharmacologic interventions to manipulate pulmonary vascular resistance, cannot be overlooked.¹⁷¹ The studies critical of routine PAC use do not directly address the cardiac surgical population and thus cannot reasonably be used to exclude this patient population from their use. Third, an extraordinarily useful monitor for evaluating the adequacy of oxygen delivery is measurement of mixed venous oxygen saturation (see Chapter 14). It might be argued that a central venous catheter alone can be used to estimate central filling pressures, and Mangano et al¹⁷² demonstrated the ability to assess LV filling pressures using a central venous catheter, but only if the EF was greater than 40%. The population in question, however, will (by definition) present to the operating room with severely depressed LV function, justifying the use of a PAC in the majority of cases.

TEE is now considered one of the main monitoring devices used by a cardiac anesthesiologist. In addition to all the other 2D and Doppler information obtainable, the main advantage of TEE over a PAC lies in the ability to directly visualize and optimize filling volumes. When using a PAC, the clinician assumes a certain filling volume by measuring filling pressures. However, this assumed correlation is accurate only when the pressure-volume curve is known, and this is rarely the case in clinical practice. This patient population with severe CHF, usually presenting with concurrent diastolic dysfunction, will have pathologic compliance curves, requiring higher than normal filling pressures to obtain normal filling volumes. In addition, the Doppler capabilities of modern TEE systems can give the clinician a wealth of knowledge to help fine-tune pharmacologic interventions. There are, however, certain limitations to using TEE. In addition to the fact that initial acquisition cost and maintenance of equipment are considerable, TEE is an invasive monitoring technique, and injuries to the pharynx and esophagus are well-known (though uncommon) complications. Perhaps the biggest limiting factor in using TEE as a routine monitoring device, however, is the fact that interpretation of ultrasound images can be complex and requires an experienced echocardiographer.

Nowhere is TEE a more invaluable intraoperative tool than during surgical procedures intended to improve cardiac function, because the success of many of these procedures depends on specific information provided by the echocardiographer. For example, TEE visualization of the precise mechanism and location of mitral regurgitation often determine the feasibility of valve repair. TEE is used to assess the anatomy of the valve overall, as well as to specifically evaluate the leaflets for abnormal thickening, calcification, mobility, and points of coaptation with respect to the annular plane. Doppler analyses and color-flow mapping complement the 2D evaluation and may provide additional information. The use of a mechanistic classification of mitral regurgitation greatly facilitates communication with the surgeon. The Carpentier classification of mitral regurgitation (see Table 27-3) is often used because it mechanistically distinguishes valves with normal leaflet motion (type I), excessive leaflet motion (type II), and restricted leaflet motion (type III).¹⁷³

Transesophageal Echocardiography and Ventricular Assist Devices

The role of TEE where mechanical VADs are concerned really begins before placement of the device, with a TEE evaluation focused on detecting or ruling out specific anatomic pathologies that may prevent the device from functioning as intended or lead to preventable complications, and often must be surgically addressed before starting support by the device.

To know what to look for, it is important to understand the basics of the device, including where the device or its components reside during support, how they are placed there, and the principles by which the device functions to assist the ventricle and support the circulation. As described earlier, the goal of such assist devices is twofold: (1) to decompress, or to empty, the failing ventricle by diverting blood from the failed side of the heart into the pump; and (2) once the pump is full, the stroke volume in the pump must be ejected into the arterial circulation immediately downstream of the failing ventricle to provide effective forward CO.

Regardless of the specific manufacturer’s device attached to the inflow and outflow cannulae, TEE is an invaluable tool before, during, and after the placement of VADs (Table 27-7).

TABLE 27-7 Common Perioperative Echocardiographic Assessment of Patients Undergoing Left Ventricular Assist Device Insertion

LVAD, left ventricular assist device.

From Castillo JG, Anyanwu AC, Adams DH, et al: Real-time 3-dimensional echocardiographic assessment of current continuous-flow rotary left ventricular assist devices. J Cardiothorac Vasc Anesth 23:702–710, 2009, Table 3.

Before LVAD placement, TEE is used to detect specific anatomic pathologies that will:

During LVAD placement, TEE is used to:

After LVAD placement, TEE is used to:

For a more detailed description of the utility of TEE before and during VAD placement and in the perioperative period, the interested reader is referred to the many published reviews available in the literature and major textbooks of echocardiography.^174–178 (See Chapters 12 and 13.)

Three-dimensional Transesophageal Echocardiography and Ventricular Assist Devices

Given the increasing number of VADs being implanted worldwide, it is progressively expected of cardiac anesthesiologists to become comfortable providing anesthetic care to this extremely challenging subset of patients. This includes the echocardiographic assessment of VAD placement, as well as the timely detection of potential catastrophic events that are unique to VAD surgery. Three-dimensional echocardiography (3DE) is now a well-established imaging modality. It should never be viewed in isolation, but complementary and supplementary to 2D imaging. Strengths of 3D imaging lie in the interrogation of the positioning and alignment of cannulae, as well as an improved accuracy and reproducibility of ventricular volumes and function leading to a better understanding of ventricular spatial relations.

Volume assessment by 3DE has been shown to be rapid, accurate, and superior to conventional standardized 2D methods. Ventricular volume and mass obtained by 3DE have even compared favorably with those obtained from studies with magnetic resonance imaging, further demonstrating advantages in efficacy and accuracy in assessing volumes in remodeled ventricles after MI.^179–181

Inflow Cannula

3DE has made it substantially easier to inspect and visualize the orientation of the inflow cannula, commonly entering the left ventricle from the apex. The echocardiographic examination of the inflow cannula position and orientation using 2DE required at least two orthogonal views (four-chamber and two-chamber long axis). A dataset can be acquired and spatially oriented when using 3DE so that the imager views the mitral valve en face from the left atrial perspective. The cropping tool can now be used to edit away the mitral valve and basal regions of the left ventricle, enabling the echocardiographer to obtain an en face view of the outflow cannula as it enters the LV apex. The cannula orifice should be centrally located entering the apex of the ventricle, aligned to the LV inflow tract (mitral valve orifice), not abutting any ventricular structures (Figure 27-13). Often the cannula ends up being slightly angled toward the anteroseptal ventricular wall. As long as the deviation is less than 30 degrees, no hindrance of ventricular drainage should be encountered. Figure 27-14 depicts the proper positioning of the inflow cannula of the Impella VAD entering the left ventricle in a retrograde fashion through the aortic valve.

Figure 27-13 The cannula orifice should be centrally located entering the apex of the ventricle, aligned to the left ventricular inflow tract (mitral valve orifice), not abutting any ventricular structures. LV, left ventricle; RV, right ventricle.

Figure 27-14 The proper positioning of the inflow cannula of the Impella VAD (ventricular assist device) entering the left ventricle (LV) in a retrograde fashion through the aortic valve. LA, left atrium.

Inflow cannula patency is obviously critical in achieving adequate CO. In general, patients undergoing LVAD insertion who present with severely dilated ventricles are less prone to cannula malalignment-induced hindrance of ventricular drainage. In contrast, patients with ventricles of normal dimension (e.g., acute myocarditis, acute MI) are more dependent on proper cannula alignment. In general, deviations of less than 30 degrees from the LV inflow axis are well tolerated.

In conjunction with color-flow Doppler, the echocardiographer can check for unidirectional laminar flows through the ventricle to the device. The presence of abnormal high-velocity turbulent flows or an aliasing flow at the cannula orifice suggests cannular obstruction. The differential diagnosis can include hypovolemia, a thrombotic episode, malalignment with partial obstruction by ventricular walls, or compression of the interventricular septum (“suck-down” effect; Figure 27-15). Because the treatment for these disorders is different, it is important to have excellent echocardiographic imaging capabilities. In the scenario of malalignment, the surgeon may be able to reposition the cannula by moving the flexible conduit and reassessing the LVAD hemodynamics, routinely done at the time of chest closure. A “suck-down” effect is treated by primarily reducing device flows and volume loading of the patient, as well as potentially providing RV support.

Figure 27-15 Compression of the interventricular septum.

LV, left ventricle; RV, right ventricle.

Postoperative Considerations

By and large, improvements in ventricular function will not be immediately apparent in this population after the majority of surgical procedures described in this chapter. In fact, ventricular function is often worse after a major cardiotomy because there often has been some degree of myocardial stunning during CPB despite the best of myocardial protective techniques with modern cardioplegia. Generally, it will be necessary to optimize intravascular volume status and use pharmacologic manipulations of afterload and contractility. Temporary pacing with epicardial wires placed during surgery often is used to optimize heart rate. In addition, meticulous management of electrolytes, coagulation status, and red blood cell mass is necessary.

One area that is often neglected in this population with sometimes tenuous hemodynamics is postoperative pain management. Patients with severely decreased ventricular function will not tolerate the stress response and tachycardia that accompany postoperative pain because of the increased myocardial oxygen demand (potentially leading to ischemia) and decreased diastolic filling time (potentially leading to decreased stroke volume). This combination is especially deleterious in patients with poor ventricular function and will exacerbate hemodynamic instability.

Most often, postoperative pain management in this population is with intermittent boluses of opioids delivered by a nurse or via patient-controlled analgesia pumps. Regional techniques (e.g., continuous epidural infusions and single-shot intrathecal opioids) are becoming popular, although concern for central neuraxial hematomas in light of intraoperative heparinization and the coagulopathy that results from CPB still limit their use in the adult population in many centers.

Placement of preservative-free morphine in the subarachnoid space (7.5 to 10 μg/kg) or the epidural space (75 to 100 μg/kg) generally will be well tolerated and can provide adequate analgesia for approximately 16 to 24 hours after cardiac surgical procedures. The main side effects of this technique are itching (usually controllable by a low-dosage infusion of naloxone, 0.5 to 1 μg/kg/hr, titrated to effect) and sedation with potential late respiratory depression (see Chapter 38).