Chapter 5. Cardiovascular Pathology

In the past several decades, cardiovascular pathology has been launched into the forefront of high-quality patient care as a direct result of the virtual explosion in the number and scope of cardiovascular surgical and interventional diagnostic and therapeutic procedures and devices. Cardiovascular pathology informs the practice of cardiovascular surgery and medicine through defining the pathologic anatomy in individual patients and patient cohorts, and in developing and testing procedures and prosthetic devices that facilitate evidence-based choices among surgical or catheter-based interventional options and optimize short- and long-term patient management. Importantly, beyond implicit clinical benefit for individual patients, the discipline of cardiovascular pathology is a cornerstone of modern cardiovascular research and the preclinical development and clinical implementation of innovative drugs, devices, and other therapeutic options.

This chapter summarizes pathologic considerations most relevant to surgery and catheter-based interventions used to diagnose and treat the major forms of acquired cardiovascular disease. It emphasizes pathologic anatomy, clinico-pathologic correlations, and pathophysiologic mechanisms in the various forms of structural heart disease. In view of space limitations, several important areas (eg, aortic disease) are necessarily omitted from discussion and others (eg, cardiac assist and replacement devices) are abbreviated, as they are discussed in greater detail elsewhere in this book. Moreover, although we have not included the key pathologic considerations herein, we are mindful that the number of adults with congenital heart disease is increasing rapidly and that they have unique and important clinical and pathologic concerns.1,2

Myocardial Hypertrophy

Hypertrophy is the compensatory response of the cardiac muscle (the myocardium), to increased work (Fig. 5-1).3 This structural and functional adaptation accompanies nearly all forms of heart disease, and its consequences often dominate the clinical picture. Hypertrophy induces an increase in the overall mass and size of the heart that reflects increased size of individual myocytes largely through addition of contractile elements (sarcomeres) and associated cell and tissue elements. Functionally beneficial augmentation of myocyte number (hyperplasia) in response to stress or injury has not been demonstrated in the adult heart.

The pattern of hypertrophy reflects the nature of the stimulus (see Fig. 5-1B). Pressure overload (eg, in systemic hypertension or aortic stenosis) induces concentric hypertrophy with an increased ventricular mass, increased wall thickness, and increased ratio of wall thickness to cavity radius without dilation. In contrast, volume-overload (eg, in aortic or mitral regurgitation, myocardial infarction, or dilated cardiomyopathy) promotes hypertrophy accompanied by chamber dilation, in which both ventricular radius and total mass are increased (sometimes called eccentric hypertrophy). In myocardial infarction, where there is regional myocyte necrosis and loss, hypertrophy occurs only in noninfarcted regions of myocardium (termed compensatory hypertrophy).

In contrast, the chamber wall is affected globally by the increased chamber pressure of hypertension, the pressure or volume overload of valvular heart disease, and in dilated cardiomyopathy. The constellation of regional changes that occurs in both infarcted and noninfarcted myocardium or more globally in pressure and volume overload is called ventricular remodeling.4 At a cell level, pressure overload promotes augmentation of cell width via parallel addition of sarcomeres; in contrast, volume overload and/or dilation stimulate augmentation of both cell width and length via both parallel and series addition of sarcomeres.

The changes in the heart described in the preceding initially increase the efficiency of the heart, enhance function, and are thereby adaptive. However, when these changes are excessive and prolonged, they may ultimately become deleterious and contribute to cardiac failure. This can occur via several mechanisms. Because the vasculature does not proliferate commensurate with increased cardiac mass, hypertrophied myocardium is usually relatively deficient in blood vessels and thereby particularly vulnerable to ischemic damage. Moreover, myocardial fibrous tissue is often increased. Also important are the molecular changes that accompany and likely mediate enhanced function in hypertrophied hearts, and which may subsequently promote development of heart failure. For example, hypertrophy induces a gene expression profile in cardiac myocytes that is generally similar to that of proliferating cells generally and particularly that of fetal cardiac myocytes during development. Differentially expressed proteins may be less functional and/or more or less abundant than normal. Hypertrophied and/or failing myocardium also may have mechanical disadvantage, reduced adrenergic responsiveness, decreased calcium availability, impaired mitochondrial function, and microcirculatory spasm. Apoptosis of cardiac myocytes may also contribute to heart failure. Novel therapeutic strategies for heart failure treatment based on molecular mechanisms are under investigation.5–7

Thus, owing to the totality of the changes described above, cardiac hypertrophy comprises a tenuous balance. The adaptive changes may be overwhelmed by potentially deleterious structural, functional, and biochemical/molecular alterations, including quantitative or qualitative alterations in cardiac configuration, metabolic requirements of an enlarged muscle mass, protein synthesis, decreased capillary/myocyte ratio, fibrosis, microvascular spasm, and impaired contractile mechanisms. Hypertrophy also decreases myocardial compliance and may thereby hinder diastolic filling. In addition, left ventricular hypertrophy is an independent risk factor for cardiac mortality and morbidity, especially sudden death.8

Heart failure can occur with pressure or volume overload of many causes, owing to both regional and global lesions (Fig. 5-2). Although left ventricular hypertrophy regresses in many cases following removal of the stimulus, the extent of resolution in an individual is unpredictable. Moreover, progressive cardiac failure may ensue following cardiac surgery and despite revascularization or hemodynamic adjustment, for example, by valve replacement or repair (see Fig. 5-2B). In addition, markedly increased cardiac muscle mass may compromise intraoperative myocardial preservation.

Cardiomyopathies

Cardiomyopathies are diseases in which the primary cardiovascular abnormality is in the myocardium. A primary cardiomyopathy comprises a condition solely or predominantly confined to heart muscle, whereas a secondary cardiomyopathy (often called specific heart muscle disease) implies myocardial involvement as a feature of a generalized systemic or multisystem disorder, such as amyloidosis, hemochromatosis, other infiltrative and storage diseases, drug and other toxic reactions, sarcoidosis, various autoimmune and collagen vascular diseases, or neuromuscular/neurologic disorders (eg, Duchenne-Becker muscular dystrophy). Primary cardiomyopathies are subdivided into genetic, mixed, and acquired categories, underscoring the role of recently elucidated molecular and genetic mechanisms.9 Genetic causes of primary cardiomyopathy include hypertrophic cardiomyopathy, arrhythmogenic right ventricular cardiomyopathy, left ventricular noncompaction, the ion channel disorders (eg, long QT and Brugada syndromes), and some cases of dilated cardiomyopathy. Acquired causes of primary cardiomyopathy include myocarditis (inflammatory cardiomyopathy), and stress-provoked (tako-tsubo), tachycardia-induced, and peripartum cardiomyopathy. The terms ischemic heart disease, valvular heart disease, and hypertensive heart disease are preferred over the terms ischemic cardiomyopathy, valvular cardiomyopathy, and hypertensive cardiomyopathy, because these conditions more likely reflect compensatory and remodeling changes induced by another cardiovascular abnormality. The epicardial coronary arteries are usually free of significant obstructions in patients with cardiomyopathies.

In some cases (eg, myocarditis, sarcoidosis, amyloidosis, and hemochromatosis) the cause of a cardiomyopathy may be revealed by light and/or electron microscopic examination of an endomyocardial biopsy specimen, and other conditions such as arrhythmogenic right ventricular cardiomyopathy and hypertrophic cardiomyopathy have characteristic gross and microscopic features demonstrable upon evaluation of the heart at the time of transplantation or autopsy. In endomyocardial biopsy, also used in the management of the ongoing surveillance of cardiac transplant recipients,10 a bioptome is inserted into the right internal jugular or femoral vein and advanced under fluoroscopic or echocardiographic guidance through the tricuspid valve, to the apical half of the right side of the ventricular septum, yielding 1- to 3-mm fragments of myocardium. Correlation between right- and left-sided findings generally is good in most myocardial diseases and transplant rejection.

Common variants of primary cardiomyopathy are illustrated in Fig. 5-3.

Dilated Cardiomyopathy

Dilated cardiomyopathy is characterized by cardiomegaly usually to two to three times normal weight, and four-chamber dilation (see Fig. 5-3A).11 The primary functional abnormality in dilated cardiomyopathy is impairment of left ventricular systolic function. Mural thrombi are sometimes present, predominantly in the left ventricle and are a potential source of thromboemboli. The histologic changes in dilated cardiomyopathy, comprising myocyte hypertrophy and interstitial fibrosis, are nonspecific and indistinguishable from those in failing muscle in ischemic or valvular heart disease (see Fig. 5-3B). Moreover, the severity of the morphologic changes does not necessarily correlate with the severity of dysfunction or the patient's prognosis.

Dilated cardiomyopathy has a genetic and often familial basis in approximately 30 to 50% of cases, with autosomal dominant (most common), autosomal recessive, X-linked and mitochondrial inheritance reported. Mutations most commonly involve genes encoding proteins of the cardiac myocyte cytoskeleton. Secondary cardiomyopathies yielding a dilated phenotype also occur as a result of alcoholism and pregnancy-associated nutritional deficiency or immunologic reaction.12

Hypertrophic Cardiomyopathy

Hypertrophic cardiomyopathy is characterized macroscopically by massive myocardial hypertrophy without dilation (see Fig. 5-3C), often with disproportionate thickening of the ventricular septum relative to the free wall of the left ventricle (ratio >1.3) (termed asymmetric septal hypertrophy). In some patients, the basal septum is markedly thickened at the level of the mitral valve, and the outflow of the left ventricle may be narrowed during systole, yielding left ventricular outflow tract obstruction. In such cases, contact between the left ventricular outflow tract and the anterior mitral leaflet during ventricular systole (observed by echocardiography as systolic anterior motion of the mitral valve) results in LV outflow tract endocardial thickening, in a configuration that mirrors the anterior leaflet of the mitral valve. The most important microscopic features in hypertrophic cardiomyopathy include: (1) disorganized myocytes and contractile elements within cells (myofiber disarray) typically involving 10 to 50% of the septum; (2) extreme myocyte hypertrophy, with myocyte diameters frequently more than 40 μm (normal approximately 15 to 20 μm); and (3) interstitial and replacement fibrosis (see Fig. 5-3F).

The clinical course of hypertrophic cardiomyopathy is extremely variable. Potential complications include atrial fibrillation with potential mural thrombus formation and embolization, infective endocarditis of the mitral valve, intractable cardiac failure, and sudden death. Sudden death is common, with risk related to the degree of hypertrophy and specific gene mutations.13 Reduced stroke volume results from decreased diastolic filling of the massively hypertrophied left ventricle. Patients with left ventricular obstruction may benefit from septal reduction by surgical myotomy/myectomy or chemical ablation.14 End-stage heart failure can be accompanied by cardiac dilation.

Hypertrophic cardiomyopathy has a genetic basis in virtually all cases.15 In many patients the disease is familial, with autosomal dominant transmission and variable expression; remaining cases are sporadic. Hundreds of mutations have been identified in 11 genes; almost all genes are for sarcomeric proteins, most commonly β-myosin heavy chain. Nevertheless, the mechanism by which defective sarcomeric proteins produce the phenotype of hypertrophic cardiomyopathy is yet incompletely defined.

Restrictive Cardiomyopathy

Restrictive cardiomyopathy is characterized by impeded diastolic relaxation and left ventricular filling, often with preserved systolic function (called diastolic dysfunction). Any disorder that interferes with ventricular filling can cause restrictive cardiomyopathy (eg, eosinophilic endomyocardial disease, amyloidosis, storage diseases, or postirradiation fibrosis) or mimic it (eg, constrictive pericarditis or hypertrophic cardiomyopathy). In restrictive cardiomyopathy, biatrial dilation but not ventricular dilation may be prominent. Distinct morphologic patterns may be revealed by endomyocardial biopsy, particularly in deposition processes such as amyloid or Fabry's disease.16

Arrhythmogenic Right Ventricular Cardiomyopathy

Arrhythmogenic right ventricular cardiomyopathy is characterized by a dilated right ventricular chamber and severely thinned right ventricular wall, with extensive fatty infiltration, loss of myocytes with compensatory myocyte hypertrophy, and interstitial fibrosis (see Fig. 5-3G–H).17 Clinical features include right-sided heart failure and arrhythmias. Arrhythmias are often brought on by exertion, and this condition is associated with sudden death in athletes. Right ventricular cardiomyopathy is associated with mutations in several genes involved in cell-cell adhesion, including plakoglobin, desmoplakin, plakophilin-2, and desmoglein-2.

Myocardial ischemia is a physiologic condition that occurs when perfusion via coronary flow is inadequate to meet metabolic needs, thus interfering with the cellular delivery of oxygen and nutrients to cardiac myocytes, and removal of waste products. Myocardial ischemia is most often caused by diminished perfusion of the myocardium resulting from obstruction or narrowing of a coronary artery secondary to atherosclerosis. Nonatherosclerotic epicardial coronary artery obstructions can also occur in autoimmune diseases (eg, systemic lupus erythematosus and rheumatoid arthritis), progressive systemic sclerosis (scleroderma), vasculitis (eg, Buerger disease and Kawasaki disease), fibromuscular dysplasia, and as a result of dissection, spasm, or embolism. Obstruction of the small intramural coronary arteries occurs in diabetes, the Fabry disease, and amyloidosis and in cardiac allografts (see later). Decreased coronary flow leading to global hypoperfusion and myocardial ischemia can also occur as a result of systemic hypotension and during cardiopulmonary bypass. Ischemia can also result from increased cardiac demand secondary to exercise, tachycardia, hyperthyroidism, or ventricular hypertrophy and/or dilation. Moreover, the effects of ischemia are potentiated when oxygen supply is decreased secondary to anemia, hypoxia, or cardiac failure.

Atherosclerosis

Atherosclerosis is a chronic, progressive, multifocal disease of the vessel wall, beginning in the intima, whose characteristic lesion is the atheroma or plaque that forms through intimal thickening (mediated predominantly by smooth muscle cell proliferation and matrix production) and lipid accumulation (mediated primarily by insudation of lipid into the arterial wall followed by monocyte/macrophage phagocytosis).18 Atherosclerosis primarily affects the large elastic arteries and large- and medium-sized muscular arteries of the systemic circulation, particularly near branches, sharp curvatures, and bifurcations. Coronary arterial atherosclerosis involves especially the epicardial branches of the left anterior descending (LAD) and circumflex arteries, and the right coronary diffusely, but generally not their intramural branches. Most atheromas in the coronary arteries are segmental and eccentric, with plaque-free segments longitudinal and circumferentially. In early lesions, the plaque bulges outward at the expense of the media with the arterial lumen remaining circular in cross-section at essentially the same original diameter (ie, the vessel wall outer diameter enlarges, a process termed vascular remodeling).19

Although veins are usually spared from atherosclerosis, venous bypass grafts interposed within branches of the arterial system (such as saphenous vein grafts) frequently develop intimal thickening and ultimately atherosclerotic obstructions. Paradoxically, some arteries used as arterial grafts, including the internal mammary artery, are largely spared.

Pathogenesis

The prevailing theory of lesion formation in atherosclerosis centers on interactions among arterial wall endothelial and vascular smooth muscle cells, circulating monocytes, platelets, and plasma lipoproteins. A key contributor to atherosclerosis is endothelial cell injury induced by chronic hypercholesterolemia, homocystinemia, chemicals in cigarette smoke, viruses, localized hemodynamic forces, systemic hypertension, hyperglycemia, or the local effects of cytokines. These factors cause phenotypic and hence functional changes in endothelial cells, called endothelial dysfunction.20 Endothelial dysfunction causes: (1) vasoconstriction owing to decreased production of the vasodilator nitric oxide; (2) increased permeability to lipoproteins; (3) expression of tissue factor leading to thrombosis; and (4) expression of certain injury-induced adhesion molecules leading to adherence of platelets and inflammatory cells.

Progression from an early, subendothelial lesion (called a fatty streak) to a complex atheromatous plaque involves the following processes: (1) monocytes adhere to endothelial cells, migrate into the subendothelial space, and transform into tissue macrophages; (2) smooth muscle cells migrate from the media into the intima, proliferate, and secrete collagen and other extracellular matrix constituents; (3) lipids accumulate via phagocytosis in macrophages (forming foam cells) and smooth muscle cells, as well as extracellularly; (4) lipoproteins are oxidized in the vessel wall leading to generation of potent biologic stimuli such as chemoattractants and cytotoxins; (5) persistent chronic inflammation; (6) cellular necrosis with release of intracellular lipids (mostly cholesterol esters); and often (7) calcification. Mature atherosclerotic plaques consist of a central core of lipid, cholesterol crystals, macrophages, smooth muscle cells, foam cells, and lymphocytes along with necrotic debris, separated from the lumen by a fibrous cap rich in collagen.

Clinical manifestations of advanced coronary arterial atherosclerosis occur through encroachment of the lumen leading to progressive stenosis, or to acute plaque disruption with thrombosis (see the following). In the absence of significant coronary blockages, myocardial perfusion is adequate at rest, and compensatory vasodilation provides flow reserve that is more than sufficient to accommodate increased metabolic demand during vigorous exertion. When the coronary luminal cross-sectional area is decreased by approximately 75%, blood flow becomes limited during exertion; with 90% reduction, coronary flow may be inadequate at rest. However, occlusions that develop slowly may stimulate the formation of collateral vessels that protect against distal myocardial ischemia. Aneurysms may form secondary to atherosclerosis as a result of destruction of the media beneath a plaque, a process most common in the aorta and other large vessels (where plaque does not easily cause obstruction). The natural history, morphologic features, key pathogenetic events, and clinical complications of atherosclerosis are summarized in Figs. 5-4 and 5-5.

Role of Acute Plaque Change

The onset and prognosis of ischemic heart disease are not well predicted by the angiographically determined extent and severity of luminal obstructions.21 The conversion of chronic stable angina or an asymptomatic state to an acute coronary syndrome (ie, myocardial infarction, unstable angina, and sudden coronary death) is dependent on dynamic vascular changes, such as fracture or rupture of the fibrous cap, exposing deep plaque constituents. Less commonly, the change is fissuring, and/or ulceration of the fibrous cap, which results in flow disruption and exposure of the luminal blood to highly thrombogenic surfaces, thereby setting the stage for platelet aggregation and mural or total thrombosis.22

Plaques having a high propensity to rupture are known as vulnerable plaques. Such lesions: (1) have a thin collagenous fibrous cap and few smooth muscle cells (the cells that produce collagenous matrix); (2) contain many macrophages producing matrix metalloproteinases (enzymes that degrade the collagen that ordinarily lends strength to the fibrous cap); and (3) contain large areas of foam cells, extracellular lipid and necrotic debris. Inflammation may contribute to coronary thrombosis by altering the balance between prothrombotic and fibrinolytic properties of the endothelium. There is evidence that lipid lowering by diet or drugs such as statins (HMG CoA reductase inhibitors) reduces accumulation of macrophages expressing matrix-degrading enzymes and thereby stabilizes plaque by increasing the thickness and strength of the fibrous cap.23

Vulnerable plaques are not necessarily significantly obstructive. Indeed, pathologic and clinical studies show that plaques that rupture and lead to coronary occlusion often produced only mild to moderate luminal stenosis (and often no symptoms) prior to acute plaque change. Thus, there is great interest in identifying vulnerable plaques in individuals at a potentially therapeutic stage. Calcification of the coronary arteries detected noninvasively by electron beam computed tomography predicts the extent of atherosclerotic disease overall but does not predict plaque instability. The most relevant features of plaque structure can be evaluated using intravascular ultrasound, optical coherence tomography (OCT), and potentially noninvasive molecular imaging.24,25

The events that trigger abrupt changes in plaque configuration and superimposed thrombosis and the efficacy and safety of interventional therapies depend on influences both intrinsic (eg, structure and composition, as described in the preceding) and extrinsic (eg, blood pressure, vasospasm, and platelet reactivity) to the plaque.26,27 Potential outcomes for ruptured plaques include progression to thrombotic occlusion, nonocclusive thrombosis, healing at the site of plaque erosion, atheroembolization or thromboembolization, organization of mural thrombus (plaque progression), and organization of the occlusive mass with recanalization.

Ischemic Myocardial Injury

The clinical manifestations of ischemic heart disease most frequently reflect the downstream effects of a complex and dynamic interaction among fixed atherosclerotic narrowing of the epicardial coronary arteries, intraluminal thrombosis overlying a ruptured or fissured atherosclerotic plaque, platelet aggregation, vasospasm, and the responses of the myocardium to ischemia.

Progression of Injury

The changes in the myocardium following the onset of myocardial ischemia are sequential and the cellular consequences are primarily determined by the severity and duration of flow deprivation (Table 5-1). Within seconds of onset, ischemia induces a transition from aerobic metabolism to anaerobic glycolysis in cardiac myocytes, leading to inadequate production of high-energy phosphates such as ATP, and the accumulation of metabolites such as lactic acid, causing intracellular acidosis. Myocardial function is exquisitely sensitive to these biochemical consequences, with total loss of contraction within 2 minutes in severe ischemia. Nevertheless, ischemic changes in an individual cell are not immediately lethal, and injury is potentially reversible if the duration of ischemia is short (<20 minutes). Irreversible injury of cardiac myocytes marked by cell membrane structural defects occurs only after 20 to 40 minutes within the most severely ischemic area (Fig. 5-6). Lethal injury to groups of cells owing to severe prolonged ischemia causes myocardial infarction.

Within the region of myocardium vulnerable to die if ischemia is not relieved in a timely manner (termed the area at risk), loss of perfusion is not uniform, and not all cells in the area at risk are equally affected. The most severely affected myocytes, and therefore the first to become necrotic, reside in the subendocardium and in the papillary muscles furthest from lateral regions of adequate perfusion. If uninterrupted ischemia progresses, there is a wavefront of cell death outward from the mid-subendocardial region toward, eventually encompassing the lateral borders and less severely ischemic subepicardial and peripheral regions of the area at risk. In myocardial infarction, approximately 50% of the area at risk becomes necrotic in approximately 3 to 4 hours. The final transmural extent of an infarct is generally established within 6 to 12 hours. The key principle is that if perfusion is restored prior to the onset of irreversible changes, cell death can be prevented. Thus, restoration of flow to a severely ischemic area via therapeutic intervention can alter the outcome, depending on the interval between the onset of ischemia and restoration of blood flow (reperfusion).28

Neither reversible ischemia nor necrosis existing less than 6 or more hours before patient death are visible by routine gross or microscopic pathologic analysis of a cardiac specimen at autopsy. However, in a patient who died at least 2 to 3 hours following the onset of infarction, the presence of a necrotic region may be detected as a staining defect with triphenyl tetrazolium chloride (TTC), a dye that turns viable myocardium a brick-red color on reaction of the dye with intact myocardial dehydrogenases.29 The earliest observable microscopic features of infarction are intense eosinophilia, nuclear pyknosis, and loss of myocytes in clusters; some may be stretched and wavy. Short-term ischemia insufficient to cause frank necrosis cannot be reliably demonstrated by pathologic examination.

Tissue Repair Following Myocardial Infarction

The inflammation and repair sequence in response to myocardial cell death follows a largely stereotyped sequence similar to tissue repair following injury at extracardiac sites. The inflammatory reaction is characterized by early exudation of polymorphonuclear leukocytes, seen after 6 to 12 hours and maximal within 1 to 3 days. Subsequently (3 to 5 days), the infiltrate consists predominantly of macrophages that remove the necrotic tissue and set the stage for collagen production by fibroblasts accompanied by neovascularization beginning approximately 7 to 10 days at the margins of preserved tissue. The gross appearance reflects the progressive microscopic changes (Fig. 5-7). Ultimately, the infarcted tissue is replaced by dense collagenous scar, which is fully mature by about 6 to 8 weeks. Healing of myocardial infarction may be altered by reperfusion (see the following), mechanical stress, gender, and neurohumoral and other factors.30

Although cardiac myocytes are traditionally thought incapable of regeneration, a growing body of evidence suggests that regeneration of cardiac myocytes can occur under certain circumstances, including at the viable borders of myocardial infarcts.31 Whether this capacity for renewal can be harnessed to therapeutic advantage is not yet known (discussed later in this chapter).

Effects of Reperfusion

Reperfusion of an ischemic zone can alter the progression of myocardial ischemic injury, in both jeopardized and already necrotic myocardium. Reperfusion occurring before the onset of irreversible injury prevents infarction altogether. Reperfusion later (ie, following some cell death) may limit infarct size through salvage of myocytes located outside the leading edge of the “wavefront,” provided that these myocytes are only reversibly injured at the time reperfusion occurs.32 Thus, the potential for recovery of viable tissue decreases with increasing severity and duration of ischemia (see Fig. 5-6). Owing to the typical progression of ischemic injury, reperfusion 3 to 4 hours following onset of ischemia is considered the practical limit for significant myocardial salvage.

Reperfused previously ischemic myocardium may have hemorrhage (owing to microvascular damage; see the following) and necrotic myocytes with transverse eosinophilic lines (called contraction bands) that represent hypercontracted sarcomeres (Fig. 5-8) caused by ischemic cell membrane damage followed by a massive cellular influx of calcium derived from the restored blood flow. In reperfusion after especially severe global ischemia, the left ventricle may undergo a massive tetanic contraction (stone heart syndrome).33

Microvascular injury can follow myocyte injury in severe, prolonged myocardial ischemia, eventuating in hemorrhage during reperfusion, visible grossly and microscopically (resulting from vascular wall incompetence), and/or occlusion (resulting from endothelial or interstitial edema and/or plugging by platelet or neutrophil aggregates). Microvascular occlusions may inhibit the reperfusion of damaged regions (no-reflow phenomenon).34 Moreover, reperfusion may damage some of the ischemic but still viable myocytes that were not irreversibly injured (called reperfusion injury). Arrhythmias may occur after reperfusion, possibly secondary to myocyte damage by oxygen free radicals and/or increased intracellular calcium.

Myocardial Stunning, Hibernation, and Preconditioning

Although reperfusion may salvage ischemic but not necrotic myocardium, metabolic and functional recovery is usually not instantaneous; indeed, reversible postischemic myocardial dysfunction (myocardial stunning) may persist for hours to days following brief periods of ischemia.35 Myocardial stunning may occur after percutaneous coronary intervention, cardiopulmonary bypass, or ischemia related to unstable angina or stress.

Regions of viable myocardium with chronically impaired function may exist in the setting of chronically reduced coronary blood flow (termed hibernating myocardium).36 Myocardial hibernation is characterized by: (1) persistent wall motion abnormality; (2) low myocardial blood flow; and (3) evidence of viability in at least some of the affected areas. Contractile function of hibernating myocardium can improve if blood flow returns toward normal or if oxygen demand is reduced. Correction of this abnormality is likely responsible for the reversal of longstanding defects in ventricular wall motion observed following coronary bypass graft surgery or percutaneous coronary intervention. Morphologically, sublethal chronic ischemic injury often manifests as myocyte vacuolization, particularly in the subendocardium.37

Adaptation to short-term transient ischemia (ie, duration insufficient to cause cell death) may induce tolerance against subsequent, more severe ischemic insults (ischemic preconditioning).38 Thus, short (5-minute) periods of cardiac ischemia followed by reperfusion can protect myocardium against injury during a prolonged period of subsequent ischemia. Understanding the yet uncertain mechanisms of this protection could lead to targets for “pre-emptive” pharmacologic stimulation of these pathways.

Myocardial Infarction and Its Complications

Coronary atherosclerosis with acute plaque rupture and superimposed thrombosis often results in a transmural (Q-wave) myocardial infarct, in which ischemic necrosis involves at least half and usually a nearly full thickness of the ventricular wall in the distribution of the involved artery. In contrast, a subendocardial (nontransmural, non–Q wave) infarct constitutes an area of ischemic necrosis that is limited to the inner third to half of the ventricular wall, which can occur in the setting of episodic hypotension, global ischemia, or hypoxemia, or from interruption by reperfusion of the evolution of a transmural infarct. Subendocardial infarction can be multifocal, often extending laterally beyond the perfusion territory of a single coronary artery, especially when associated with diffuse stenosing coronary atherosclerosis without acute plaque rupture or thrombosis.

The short-term mortality rate from acute myocardial infarction has declined from 30% in the 1960s to 7 to 10% or less today, especially for patients who receive aggressive reperfusion/revascularization and pharmacologic therapy. However, half of all deaths from myocardial infarction occur within the first hour after the onset of symptoms, often before the victims can reach the hospital. Poor prognostic factors include advanced age, female gender, diabetes mellitus, and a previous myocardial infarction. Left ventricular function and the extent of obstructive lesions in vessels perfusing viable myocardium are the most important prognostic factors.

Important complications of myocardial infarction include ventricular dysfunction, cardiogenic shock, arrhythmias, cardiac rupture, infarct extension and expansion, papillary muscle dysfunction, right ventricular involvement, ventricular aneurysm, pericarditis, and systemic arterial embolism; the cardiac rupture syndromes and ventricular aneurysm are illustrated in Fig. 5-9. Outcome after myocardial infarction is dependent on infarct size, location, and transmurality. Patients with transmural anterior infarcts are at greatest risk for regional dilation and mural thrombi and have a worse clinical course than those with inferior-posterior infarcts. In contrast, inferior-posterior infarcts are more likely to have serious conduction blocks and right ventricular involvement.

Myocardial infarcts produce functional abnormalities approximately proportional to their size. Large infarcts have a higher probability of cardiogenic shock and congestive heart failure. Nonfunctional scar tissue resulting from previous infarcts and areas of stunned or hibernating myocardium may also contribute to overall ventricular dysfunction. Cardiogenic shock following myocardial infarction is generally indicative of a large infarct (often >40% of the left ventricle). The high mortality of myocardial dysfunction and cardiogenic shock has been alleviated somewhat in recent years by the use of intra-aortic balloon pumps and ventricular assist devices to bridge patients through phases of prolonged reversible ischemic dysfunction.

Although many patients have cardiac rhythm abnormalities following myocardial infarction, the conduction system is involved by necrosis only in a minority, and heart block following myocardial infarction usually is transient. Tachyarrhythmias usually originate in areas of severely ischemic or necrotic myocardium, owing to electrically unstable ischemic myocardium, often at the edge of an infarct. Ischemia without frank infarction can lead directly to lethal arrhythmias owing to myocardial irritability; indeed, autopsy studies of sudden death victims and clinical studies of resuscitated survivors of cardiac arrest show that only a minority of patients with ischemia-induced malignant ventricular arrhythmias develop a full-blown acute myocardial infarction.

Cardiac rupture syndromes comprise three entities (see Fig. 5-9 A–C): (1) rupture of the ventricular free wall (most common), usually with hemopericardium and cardiac tamponade; (2) rupture of the ventricular septum (less common), leading to an acquired ventricular septal defect with a left-to-right shunt; and (3) papillary muscle rupture (least common), resulting in the acute onset of severe mitral regurgitation. Acute free-wall ruptures usually are rapidly fatal and are the cause of death in 8 to 10% of patients with fatal acute transmural myocardial infarcts; repair is rarely possible.39 Ruptures tend to occur relatively early following infarction with as many as 25% presenting within 24 hours (mean interval 4 to 5 days), most commonly through the lateral free wall. A pericardial adhesion that arrests the rapidly moving blood front and aborts a rupture may result in a contained rupture with a hematoma communicating with the ventricular cavity (false aneurysm); half of false aneurysms eventually rupture.

Postinfarction septal ruptures with acute ventricular septal defect complicate 1 to 2% of infarcts. They are of two types: (1) single or multiple sharply localized, jagged, linear passageways that connect the ventricular chambers (simple type), usually involving the anteroapical aspect of the septum; and (2) defects that tunnel serpiginously through the septum to a somewhat distant opening on the right side (complex type), usually involving the basal inferoseptal wall.40 In complex lesions, the tract may extend into regions remote from the site of the infarct, such as the right ventricular free wall. Without surgery, the prognosis is poor for patients with infarct-related ventricular septal defects.

Papillary muscles are particularly vulnerable to ischemic injury and rupture, particularly the posterior medial papillary muscle. Papillary muscle rupture can occur later than other rupture syndromes (as late as 1 month after MI). Because tendinous cords arise from the heads of the papillary muscles and provide continuity with each of the valve leaflets, interference with the structure or function of either papillary muscle can result in dysfunction of both mitral valve leaflets.

Isolated right ventricular infarction is rare, but involvement of the right ventricle by extension of an inferoseptal infarct occurs in approximately 10% of transmural infarcts and can have important functional consequences, including right ventricular failure with or without tricuspid regurgitation and arrhythmias.

Infarct extension is characterized by incremental new or recurrent necrosis in the same distribution as a completed recent infarct. Extension most often occurs between 2 and 10 days after the original infarction, forms along the lateral and subepicardial borders of a recent infarct, and histologically appears younger than the previously necrotic myocardium. In contrast, infarct expansion is a disproportionate thinning and dilation of the infarcted region occurring through a combination of: (1) slippage between muscle bundles, reducing the number of myocytes across the infarct wall; (2) disruption of myocardial cells; and (3) tissue loss within the necrotic zone. Infarct expansion does not lead to additional necrotic myocardium per se, but may promote intracardiac mural thrombus formation. The increase in ventricular volume caused by regional dilation increases the wall stress and thereby the workload of noninfarcted myocardium. Infarct expansion is often the substrate for late aneurysm formation. Occurring in approximately 30% of transmural infarcts, infarct expansion increases morbidity and mortality.

Ventricular aneurysms are large scars that paradoxically bulge during ventricular systole and often result from a large transmural infarct that undergoes expansion (see Fig. 5-9D).41 Although frequently as thin as 1 mm, ventricular aneurysms rarely rupture owing to their walls of tough fibrous or fibrocalcific tissue. Hypertrophied myocardial remnants as well as necrotic but inadequately healed myocardium often are present in aneurysm walls and mural thrombus is common. Some pharmacologic approaches to limiting infarct size, such as steroids and other anti-inflammatory agents, could exacerbate infarct expansion and aneurysm formation.

Ventricular remodeling comprises the collective structural changes that occur both in the necrotic zone and uninvolved areas of the heart, including left ventricular dilation, wall thinning by infarct expansion, compensatory hypertrophy of noninfarcted myocardium, and potentially late aneurysm formation.42 Congestive heart failure secondary to coronary artery disease occurs when the overall function of viable myocardium can no longer maintain an adequate cardiac output or regions of hyperfunctioning residual myocardium suffer additional ischemic episodes.

Revascularization

The rationale for revascularization in early acute myocardial infarction is as follows: (1) prolonged thrombotic occlusion of a coronary artery causes transmural infarction; (2) the extent of necrosis during an evolving myocardial infarction progresses as a wavefront and becomes complete only 6 to 12 hours or more following coronary occlusion (see Fig. 5-6); (3) both early- and long-term mortality following acute myocardial infarction correlate strongly with the amount of residual functioning myocardium; and (4) early reperfusion rescues some jeopardized myocardium.43 Thus, the benefits of thrombolytic therapy or early percutaneous coronary intervention depend on and are assessed by the amount of myocardium salvaged, recovery of left ventricular function, and resultant reduction in mortality. These clinical endpoints are largely determined by the time interval between onset of symptoms and a successful intervention adequacy of early coronary reflow, and the degree of residual stenosis of the infarct vessel. Spontaneous recanalization via endogenous thrombolysis can be beneficial to left ventricular function but occurs in fewer than 10% of patients within the critical interval.

Percutaneous Coronary Intervention

Percutaneous coronary intervention (PCI) is used in patients with stable angina, unstable angina, or acute myocardial infarction to restore blood flow through a diseased portion of the coronary circulation obstructed by atherosclerotic plaque and/or thrombotic deposits, and in obstructions in saphenous vein grafts, internal mammary artery grafts, and occasionally coronary arteries in transplanted hearts.

Percutaneous Transluminal Coronary Angioplasty

In percutaneous transluminal coronary angioplasty (PTCA), the plaque splits at its weakest point and enlargement of the lumen and increased blood flow occurs by plaque fracture (the predominant mechanism), and by embolization, compression, redistribution of the plaque contents, and overall mechanical expansion of the vessel wall.44 The split extends at least to the intimal-medial border and often into the media, is accompanied by variable circumferential and longitudinal medial dissection, and can induce a flap that impinges on the lumen. These changes result in local flow abnormalities and generation of new, thrombogenic blood-contacting surfaces (to some extent similar to what is observed with spontaneously disrupted plaque). These changes contribute to the propensity for acute thrombotic closure that occurs in up to 5% of patients.

The long-term success of PTCA is limited by the development of progressive restenosis, which occurs in 30 to 50% of patients, most frequently within the first 4 to 6 months.45 Although vessel wall recoil and organization of thrombus likely contribute, the major process leading to restenosis is excessive medial smooth muscle migration to the intima, proliferation and secretion of abundant extracellular matrix as an exaggerated response to angioplasty-induced injury.

Coronary atherectomy of primary or restenosis lesions mechanically removes obstructive tissue by excision. Deep arterial resection, including medial and even adventitial elements, occurs frequently but has not been associated with acute symptomatic complications. The morphology of arterial vessel healing after directional or rotational atherectomy is similar to that following angioplasty.

Stents

Stents are expandable tubes of metallic or polymeric mesh that are used to splint open the vessel wall at the site of balloon angioplasty and thereby mitigate the negative sequelae of PTCA.46 Stents preserve luminal patency and provide a larger and more regular lumen by acting as a scaffold to support the disrupted vascular wall and minimize flow disruption and flow disruption and thrombus formation. Stenting has been shown to be superior to angioplasty alone in vessels greater than 3 mm in diameter, chronic total occlusions, stenotic vein grafts, restenotic lesions after angioplasty alone, and patients with myocardial infarction.47

Stent technology has undergone a rapid evolution, including sequentially: (1) bare-metal stents (BMS); and (2) drug-eluting stents (DES), both used extensively in clinical interventional cardiology; and more recently, (3) completely resorbable/biodegradable stents (RBS), which are presently in preclinical development and clinical trials. Bare-metal stents for coronary implantation are short tubular segments of metal mesh composed of balloon-expandable 316L stainless steel or nickel-titanium alloy (Nitinol) that range from approximately 2.5 to 3.5 mm in diameter and approximately 1 to 3 cm in length. Development has focused on permitting stents to become more flexible and more easily delivered and deployed, allowing the treatment of a greater number and variety of lesions.

Key stent complications are thrombosis, usually occurring early, and late proliferative restenosis (Fig. 5-10). Thrombotic occlusion, occurring in 1 to 3% of patients within 7 to 10 days of the procedure (see Fig. 5-10A), has largely been overcome by aggressive multidrug treatment with antiplatelet agents such as clopidogrel, aspirin, and glycoprotein IIb/IIIa inhibitors. The major long-term complication of bare-metal stenting is in-stent restenosis, which occurs in 50% of patients within 6 months.48 The causes of stent thrombosis and restenosis are complex and are largely owing to stent-tissue interactions, including inflammation, which may interfere with healing and re-endothelialization.49 Damage to the endothelial lining and stretching of the vessel wall stimulate adherence and accumulation of platelets, fibrin, and leukocytes. Stent wires may eventually become completely embedded in an endothelium-lined intimal fibrosis layer composed of smooth muscle cells in a collagen matrix (see Fig. 5-10B). This tissue may thicken secondary to the release of growth factors, chemotactic factors, and inflammatory mediators from platelets and other inflammatory cells that result in increased migration and proliferation of smooth muscle cells, and increased production of extracellular matrix molecules, narrowing the lumen and resulting in restenosis.

DES effectively inhibit in-stent restenosis.50,51 The drugs used most widely are rapamycin (sirolimus)52 and paclitaxel53 in the Cypher (Cordis) and Taxus (Boston Scientific) stents, respectively. Rapamycin, a drug used for immunosuppression in solid organ transplant recipients, inhibits proliferation, migration, and growth of smooth muscle cells and extracellular matrix synthesis. Paclitaxel, a drug used in the chemotherapeutic regimens for several types of cancer, also has similar anti–smooth muscle cell activities. These drugs are embedded in a polymer matrix (such as a copolymer of poly-n-butyl methacrylate and polyethylene–vinyl acetate or a gelatin-chondroitin sulfate coacervate film) that is coated onto the stent.

Data from recent studies suggest a small but significant increased risk of late (>1 year postimplantation) stent thrombosis causing myocardial infarction and/or sudden death in patients who have DES.54 The significance and causes of later stent thrombosis are subjects of active controversy and investigation. However, animal studies and some clinical data suggest that late stent thrombosis may be related to DES-induced inhibition of stent endothelialization.

RBS ultimately resorb and facilitate removal of foreign material that may potentiate a thrombotic event, permit more versatility in subsequent therapies, and do not interfere with the diagnostic evaluation by noninvasive imaging such as cardiac magnetic resonance and CT. Several RBS are in development or in clinical trials. The key challenge is to achieve an optimal practical balance among biocompatibility, and control the kinetics of stent degradation at a rate appropriate to maintain mechanical strength to limit recoil.55

Coronary Artery Bypass Graft Surgery

Coronary artery bypass graft (CABG) surgery improves survival in patients with significant left main coronary artery disease, three-vessel (and possibly two-vessel) disease, or reduced ventricular function, and prolongs and improves the quality of life in patients with left main equivalent disease (proximal left anterior descending and proximal left circumflex), but does not protect them from the risk of subsequent myocardial infarction.56 The principal mechanism for these benefits is thought to be the restoration of blood flow to hibernating myocardium.

The hospital mortality rate for CABG surgery is approximately 1% in low-risk patients, with fewer than 3% of patients suffering perioperative myocardial infarction. The most consistent predictors of mortality after CABG are urgency of operation, age, prior cardiac surgery, female gender, low left ventricular ejection fraction, degree of left main stenosis, and number of vessels with significant stenoses. The most common mode of early death after CABG is acute cardiac failure leading to low output or arrhythmias, owing to myocardial necrosis (often with features of reperfusion described in the preceding), postischemic dysfunction of viable myocardium, or a metabolic cause, such as hypokalemia. As in cardiac surgery in general, the risk of perioperative myocardial infarction increases with the degree of cardiomegaly, and in patients who have had preoperative infarction.

Early thrombotic occlusion of the graft vessel may occur, usually potentiated by inadequate distal run-off from extremely small and/or atherosclerotic distal native coronaries. Additional factors may include dissection of blood into the graft or native vessel at the anastomotic site, atherosclerosis or arterial branching at the anastomotic site, or distortion of a graft that is too short or too long for the intended bypass. In some cases, thrombosis occurring early postoperatively involves only the distal portion of the graft, suggesting that early graft thrombosis was initiated at the distal anastomosis. Most patients who die early after CABG have patent grafts.

The patency of saphenous vein grafts is reported as 60% at 10 years; occlusion results from (with increasing postoperative interval) thrombosis, progressive intimal thickening, and/or obstructive atherosclerosis.57 Between 1 month and approximately 1 year, graft stenosis is usually caused by intimal hyperplasia with excessive smooth muscle proliferation and extracellular matrix production (similar to that seen in restenosis following angioplasty and stenting). Atherosclerosis becomes the predominant mechanism in graft occlusion beyond 1 to 3 years after CABG, and earliest in those patients with the most significant atherosclerotic risk factors. Plaques in grafts often have poorly developed fibrous caps with large necrotic cores, and can develop secondary dystrophic calcific deposits that extend to the lumen (Fig. 5-11); thus, the potential for disruption, aneurysmal dilation and embolization of atherosclerotic lesions in vein grafts exceeds that for native coronary atherosclerotic lesions, and balloon angioplasty, stenting or intraoperative manipulation of grafts may potentiate atheroembolism.

In contrast to saphenous vein grafts, internal mammary artery (IMA) grafts have greater than 90% patency at 10 years (Fig. 5-12).58 Multiple factors likely contribute to the remarkably higher long-term patency of IMA grafts compared with vein grafts. Although free saphenous vein grafts sustain disruption of their vasa vasora and nerves, endothelial damage, medial ischemia, and acutely increased internal pressure, an internal mammary artery graft generally has minimal preexisting atherosclerosis and requires minimal surgical manipulation, maintains its nutrient blood supply, is adapted to arterial pressures, needs no proximal anastomosis, and has an artery-to-artery distal anastomosis. Graft and recipient vessel have comparable sizes with the internal mammary artery but are disparate (graft substantially larger) with saphenous vein. Radial and gastroepiploic arteries are occasionally used successfully. The advent of off-pump and minimally invasive coronary artery bypass grafting has stimulated efforts to facilitate sutureless anastomosis of the graft to the aorta.59

Structure-Function Correlations in Normal Valves

Normal valve function requires structural integrity and coordinated interactions among multiple anatomical components. For the atrioventricular valves (mitral and tricuspid), these elements include leaflets, commissures, annulus, tendinous cords (chordae tendineae), papillary muscles, and the atrial and ventricular myocardium. For the semilunar valves (aortic and pulmonary), the key structures are the cusps, commissures, and their respective supporting structures in the aortic and pulmonary roots.

The anatomy of the mitral and aortic valves is illustrated in Fig. 5-13.

Mitral Valve

Of the two leaflets of the mitral valve (see Fig. 5-13A), the anterior (also called septal, or aortic) leaflet is roughly triangular and deep, with the base inserting on approximately one-third of the annulus. The posterior (also called mural, or ventricular) leaflet, although more shallow, is attached to about two-thirds of the annulus and typically has a scalloped appearance. The mitral leaflets have a combined area approximately twice that of the annulus; they meet during systole with apposition to approximately 50% of the depth of the posterior leaflet and 30% that of the anterior leaflet. Each leaflet receives tendinous cords from both anterior and posterior papillary muscles. The mitral valve orifice is D-shaped, with the flat anteromedial portion comprising the attachment of the anterior mitral leaflet in the subaortic region. This part of the annulus is fibrous and noncontractile; the posterolateral portion of the annulus is muscular and contracts during systole to asymmetrically reduce the area of the orifice. The edges of the mitral leaflets are held in or below the plane of the orifice by the tendinous cords, which themselves are pulled from below by the contracting papillary muscles during systole. This serves to draw the leaflets to closure and maintain competence. The posterior leaflet is more delicate, has a shorter annulus-to-free-edge dimension than the anterior, and is therefore more prone to postinflammatory fibrous retraction and deformation owing to myxomatous degeneration. The posterior leaflet has distinct scallops that are designated P1, P2, P3, respectively, beginning from the anterolateral toward the posteromedial commissure. The orifice of the tricuspid valve is larger and less distinct than that of the mitral; its three leaflets (anterior, posterior, and septal) are larger and thinner than those of the mitral valve.

Aortic Valve

Although grossly less obviously complex than the mitral valve and apparatus, the aortic valve has structural complexity at several levels.60 The three aortic valve cusps (left, right, and noncoronary) attach to the aortic wall in a semilunar fashion, ascending to the commissures and descending to the base of each cusp (see Fig. 5-13B). Commissures are spaced approximately 120 degrees apart and occupy the three points of the annular crown, representing the site of separation between adjacent cusps. Behind the valve cusps are dilated pockets of aortic root, called the sinuses of Valsalva. The right and left coronary arteries arise from individual orifices behind the right and left cusps, respectively. At the midpoint of the free edge of each cusp is a fibrous nodule called the nodule of Arantius. A thin, crescent-shaped portion of the cusp on either side of the nodule, termed the lunula, defines the surface of apposition of the cusps when the valve is closed (approximately 40% of the cuspal area). Fenestrations (holes) near the free edges commonly occur as a developmental or degenerative abnormality, are generally small (<2 mm in diameter), and have no functional significance, because the lunular tissue does not contribute to separate aortic from ventricular blood during diastole. In contrast, defects in the portion of the cusp below the lunula are associated with functional incompetence; such holes also suggest previous or active infection. When the aortic valve is closed during diastole, there is a back pressure on the cusps of approximately 80 mm Hg. The pulmonary valve cusps and surrounding tissues have architectural similarity to but are more delicate than those of the corresponding aortic components, and lack coronary arterial origins.

All four cardiac valves essentially have the same microscopically inhomogeneous architecture, consisting of well-defined tissue layers, covered by endothelium. Using the aortic valve as the paradigm (see Fig. 5-13C), the ventricularis faces the left ventricular chamber, and is enriched in radially aligned elastic fibers, which enable the cusps to have minimal surface area when the valve is open but stretch during diastole to form a large coaptation area. The spongiosa is centrally located and is composed of loosely arranged collagen and abundant proteoglycans. This layer has negligible structural strength, but accommodates relative movement between layers during the cardiac cycle and absorbs shock during closure. The fibrosa provides structural integrity and mechanical stability through a dense aggregate of circumferentially aligned, densely packed collagen fibers, largely arranged parallel to the cuspal free edge. Normal human aortic and pulmonary valve cusps have few blood vessels; they are sufficiently thin to be perfused from the surrounding blood. In contrast, the mitral and tricuspid leaflets contain a few capillaries in their most basal thirds.

With specializations that include crimp of collagen fibers along their length, bundles of collagen in the fibrous layer oriented toward the commissures, and grossly visible corrugations, cusps are extremely soft and pliable when unloaded, but taut and stiff during the closed phase. Moreover, the orientation of connective tissue and other architectural elements is nonrandom in the plane of the cusp, yielding greater compliance in the radial than circumferential direction. The fibrous network within the cusps effectively transfers the stresses of the closed phase to the annulus and aortic wall. This minimizes sagging of the cusp centers, preserves maximum coaptation, and prevents regurgitation. During diastole, adjacent cusps of the aortic valve coapt over a substantial area (as much as one-third of the cuspal area). For the mitral valve, the subvalvular apparatus including tendinous cords and papillary muscles is the critical mechanism of valve competency.

Valvular Cell Biology

Recent studies have fostered an emerging picture of how valves form embryologically, mature in the fetus, and function, adapt, maintain homeostasis, and change throughout life. These essential relationships facilitate an understanding of valve pathology and mechanisms of disease, foster the development of improved tissue heart valve substitutes, and inform innovative approaches to heart valve repair and regeneration.61

During normal development of the heart, the heart tube undergoes looping, following which the valve cusps/leaflets originate from mesenchymal outgrowths known as endocardial cushions.62 A subset of endothelial cells in the cushion-forming area, driven by a complex array of signals from the underlying myocardium, changes their phenotype to mesenchymal cells and migrates into the acellular extracellular matrix called cardiac jelly. Likely regulated by TGFβ and VEGF, the transformation of endocardial cells to mesenchymal cells is termed transdifferentiation or endothelial-to-mesenchymal transformation (EMT).

Subsequent to morphogenesis, embryonic fetal valves possess a dynamic structure composed of a nascent ECM and cells with characteristics of myofibroblasts.63 Changes in cell phenotype and ECM remodeling continue throughout human fetal and postnatal development, and throughout life, leading to ongoing changes in properties, as evidenced by increasing valve stiffness with increasing age.64

Two types of cells are present in the fully formed aortic valve: endothelial cells located superficially and interstitial cells located deep to the surface. Aortic valve endothelial cells (VEC) have a different phenotype than those in the adjacent aorta and elsewhere,65,66 but the implications of these differences are not yet known. The second cell type comprises the valvular interstitial cells (VICs). VICs have variable properties of fibroblasts, smooth muscle cells, and myofibroblasts; they maintain the valvular extracellular matrix, the key determinant of valve durability. To maintain integrity and pliability throughout life, the valve cusps and leaflets must undergo ongoing physiologic remodeling that entails synthesis, degradation and reorganization of its ECM, which depends on matrix degrading enzymes such as matrix metalloproteinases (MMPs). Although VIC are predominantly fibroblast-like in normal valves, VIC can become activated when exposed to environmental (ie, mechanical and chemical) stimulation. Activated VIC assume a myofibroblast-like phenotype and mediate connective tissue remodeling.

Pathologic Anatomy of Valvular Heart Disease

Cardiac valve operations utilizing replacement or repair usually are undertaken for dysfunction caused by calcification, fibrosis, fusion, retraction, perforation, rupture, stretching, infection, dilation, or congenital malformations of the valve leaflets/cusps or associated structures. Valvular stenosis, defined as inhibition of forward flow secondary to obstruction caused by failure of a valve to open completely, is almost always caused by a primary cuspal abnormality and is caused by a chronic disease process. In contrast, valvular insufficiency, defined as reverse flow caused by failure of a valve to close completely, may result from either intrinsic disease of the valve cusps or from damage to or distortion of the supporting structures (eg, the aorta, mitral annulus, chordae tendineae, papillary muscles, and ventricular free wall) without primary cuspal pathology. Regurgitation may appear either acutely, as with rupture of cords, or chronically, as with leaflet scarring and retraction. Both stenosis and insufficiency can coexist in a single valve. The most commonly encountered types of valvular heart disease are illustrated in Figs. 5-14 and 5-15.

Calcific Aortic Valve Stenosis

Calcific aortic stenosis (AS), the most frequent valvular abnormality requiring surgery, is usually the consequence of calcium phosphate deposition in either an anatomically normal aortic valve or in a congenitally bicuspid valve (see Fig. 5-14A,B).67 Stenotic, previously normal tricuspid valves present primarily in the seventh to ninth decades of life, while bicuspid valves with superimposed calcification generally become symptomatic earlier (usually sixth to seventh decades).68

Calcific aortic stenosis is characterized by heaped-up, calcified masses initiated in the cuspal fibrosa at the points of maximal cusp flexion (the margins of attachment); they protrude distally from the aortic aspect into the sinuses of Valsalva, inhibiting cuspal opening. However, although the ventricular surfaces of the cusps usually remain smooth, the calfic depoosits often extend close to ventricular surface (see Fig. 5-14C). The calcification process generally does not involve the free cuspal edges, appreciable commissural fusion is absent, and the mitral valve generally is uninvolved. Calcified material resembling bone is often present (see Fig. 5-14D). Aortic valve sclerosis comprises a common, earlier, and hemodynamically less significant stage of the calcification process. Nevertheless, aortic sclerosis is associated with an increase of approximately 50% in the risk of death from cardiovascular causes and the risk of myocardial infarction, even in the absence of hemodynamically significant obstruction of left ventricular outflow.69

Aortic stenosis induces a pressure gradient across the valve, which may reach 75 to 100 mg Hg in severe cases, necessitating a left ventricular pressure of 200 mg Hg or more to expel blood. Cardiac output is maintained by the development of concentric left ventricular hypertrophy secondary to the pressure overload. The onset of symptoms such as angina, syncope, or heart failure in aortic stenosis heralds the exhaustion of compensatory cardiac hyperfunction, and carries a poor prognosis if not treated by aortic valve replacement. Other complications of calcific aortic stenosis include embolization that may occur spontaneously or during interventional procedures, hemolysis, infective endocarditis, and extension of the calcific deposits into the ventricular septum causing conduction abnormalities.

Aortic valve calcification has been traditionally considered a degenerative, dystrophic, and passive process. However, recent studies suggest active regulation of calcification in aortic valves similar in some respects to that in the atherosclerotic process in arteries, with mechanisms that include inflammation, lipid infiltration, and phenotypic modulation of VIC to an osteoblastic phenotype,70 and overlapping risk factors. Similarities to atherosclerosis have stimulated interest in the possibility that statin drugs may decrease the rate of aortic stenosis progression; however, benefit of statins for AS has not been supported by clinical studies.71

Congenitally Bicuspid Aortic Valve

With a prevalence of approximately 1%, bicuspid aortic valve (BAV) is a frequent abnormality.72 The two cusps are typically of unequal size, with the larger (conjoined) cusp having a midline raphe, representing an incomplete separation or congenital fusion of two cusps. Less frequently, the cusps are of equal size (see Fig. 5-14B). Neither stenotic nor symptomatic at birth or throughout early life, BAV are predisposed to accelerated calcification; ultimately, almost all become stenotic. Aortic pathology, including dilation and/or dissection, commonly accompanies BAV. BAV and other congenital valve abnormalities underlie over two-thirds of aortic stenosis in children and almost 50% in adults. Infrequently, they become purely incompetent, or complicated by infective endocarditis, even when the valve is hemodynamically normal. Only rarely is an uncomplicated BAV encountered incidentally at autopsy.

Recent studies have confirmed previous reports of familial clustering of BAV and left ventricular outflow tract obstruction malformations, and their association with other cardiovascular malformations.73 For example, mutations in the signaling and transcriptional regulator NOTCH1 caused a spectrum of developmental aortic valve abnormalities and severe calcification in two families with nonsyndromic familial aortic valve disease.74

Mitral Annular Calcification

Calcific deposits also can develop in the ring (annulus) of the mitral valve of elderly individuals, especially women. Although generally asymptomatic, the calcific nodules may lead to regurgitation by interference with systolic contraction of the mitral valve ring or, very rarely, stenosis by impairing mobility of the mitral leaflets during opening. Occasionally, the calcium deposits may penetrate sufficiently deeply to impinge on the atrioventricular conduction system to produce arrhythmias (and rarely sudden death). Patients with mitral annular calcification have an increased risk of stroke, and the calcific nodules, especially if ulcerated, can be the nidus for thrombotic deposits or infective endocarditis.

Rheumatic Heart Disease

Rheumatic fever is an acute, often recurrent, inflammatory disease that generally follows a pharyngeal infection with group A beta-hemolytic streptococci, principally in children. In the past several decades, rheumatic fever and rheumatic heart disease have declined markedly but not disappeared in the United States and other developed countries. Evidence strongly suggests that rheumatic fever is the result of an immune response to streptococcal antigens, inciting either a cross-reaction to tissue antigens, or a streptococcal-induced autoimmune reaction to normal tissue antigens.75

Chronic rheumatic heart disease most frequently affects the mitral and to a lesser extent the aortic and/or the tricuspid valves. Usually dominated by mitral stenosis,76 chronic rheumatic valve disease is characterized by fibrous or fibrocalcific thickening of leaflets and tendinous cords, and commissural and chordal fusion (see Fig. 5-15A and B). Stenosis results from leaflet and chordal fibrous thickening and commissural fusion, with or without secondary calcification. Regurgitation usually results from postinflammatory scarring-induced retraction of cords and leaflets. Combinations of lesions may yield valves that are both stenotic and regurgitant. Although considered the pathognomonic inflammatory myocardial lesions in acute rheumatic fever, Aschoff nodules are found infrequently in myocardium sampled at autopsy or at valve replacement surgery, most likely reflecting the extended interval from acute disease to critical functional impairment.

Myxomatous Degeneration of the Mitral Valve (Mitral Valve Prolapse)

Myxomatous mitral valve disease (mitral valve prolapse) causes chronic, pure, isolated mitral regurgitation.77 Owing to improved imaging technology and large community studies, a prevalence of mitral valve prolapse of approximately 2% has been established with a rate of the serious complications, including heart failure, mitral regurgitation, infective endocarditis, stroke, or other manifestation of thromboembolism, progressive congestive heart failure, sudden death, or atrial fibrillation of approximately 3%. Mitral valve prolapse is the most common indication for mitral valve repair or replacement.

In mitral valve prolapse (MVP), one or both mitral leaflets are enlarged, redundant, or floppy and will prolapse, or balloon back into the left atrium during ventricular systole (see Fig. 5-15C). The three characteristic anatomic changes in mitral valve prolapse are: (1) intercordal ballooning (hooding) of the mitral leaflets or portions thereof (most frequently involving the posterior leaflet), sometimes accompanied by elongated, thinned, or ruptured cords; (2) rubbery diffuse leaflet thickening that hinders adequate coaptation and interdigitation of leaflet tissue during valve closure; and (3) annular dilation, with diameters and circumferences that may exceed 3.5 and 11.0 cm, respectively (see Fig. 5-15D). Pathologic mitral annular enlargement predominates in and may be confined to the posterior leaflet, because the anterior leaflet is firmly anchored by the fibrous tissue at the aortic valve and is far less distensible. The key microscopic changes in myxomatous degeneration are attenuation or focal disruption of the fibrous layer of the valve, on which the structural integrity of the leaflet depends, and focal or diffuse thickening of the spongy layer by proteoglycan deposition (which gives the tissue an edematous, blue appearance on microscopy called myxomatous by pathologists).78 These changes in tissue structure and composition weaken the leaflet. Concomitant involvement of the tricuspid valve is present in 20 to 40% of cases, and the aortic and pulmonary valves also may be affected.

Secondary changes may occur, including: (1) fibrous thickening along both surfaces of the valve leaflets; (2) linear thickening of the subjacent mural endocardium of the left ventricle as a consequence of friction-induced injury by cordal hamstringing of the prolapsing leaflets; (3) thrombi on the atrial surfaces of the leaflets, particularly in the recesses behind the ballooned leaflet segments; (4) calcification along the base of the posterior mitral leaflet; and (5) cordal thickening and fusion that can resemble postrheumatic disease.

The pathogenesis of myxomatous degeneration is uncertain, but this valvular abnormality is a common feature of Marfan's syndrome and occasionally other hereditary connective tissue disorders such as Ehlers-Danlos syndrome, suggesting an analogous connective tissue defect. In these heritable disorders of connective tissue, including Marfan's syndrome, MVP is usually associated with mutations in fibrillin-1 (FBN-1); recent evidence also has implicated abnormal TGF-β signaling (similar to the aortic abnormalities in the pathogenesis of Marfan's syndrome and related disorders).79 Although it is unlikely that more than 1 to 2% of patients with MVP have an identifiable connective tissue disorder, studies utilizing genetic linkage analysis have mapped families with autosomal dominant mitral valve prolapse to specific chromosomal abnormalities, several of which involve genes that could be involved in valvular tissue remodeling.

Ischemic Mitral Regurgitation

In ischemic mitral regurgitation (IMR), also called functional mitral regurgitation, the leaflets are intrinsically normal, whereas myocardial structure and function are altered by ischemic injury.80 Present in 10 to 20% of patients with coronary artery disease, IMR worsens prognosis following myocardial infarction, with reduced survival directly related to the severity of the regurgitation. Mechanisms of IMR include an ischemic papillary muscle that fails to tighten the cords during systole, and fibrotic, shortened papillary muscle that fixes the chordae deeply within the ventricle. Nevertheless, papillary muscle dysfunction alone is generally insufficient to produce IMR, and regional dysfunction and dilation with an increasing spherical shape of the LV, which pulls the papillary muscles down and away from the center of the chamber, usually contributes. There is substantial interest in developing surgical and/or percutaneous approaches to the repair of ischemic mitral regurgitation.81

Carcinoid and Drug-Induced Valve Disease

Patients with the carcinoid syndrome often develop plaquelike intimal thickenings of the endocardium of the tricuspid valve, right ventricular outflow tract, and pulmonary valve superimposed on otherwise unaltered endocardium.82 The left side of the heart is usually unaffected. These lesions are related to elaboration by carcinoid tumors (most often primarily in the gut) of bioactive products, including serotonin, which cause valvular endothelial cell proliferation but are inactivated by passage through the lung.

Left-sided but similar valve lesions, have been reported to complicate the administration of fenfluramine and phentermine (fen-phen), appetite suppressants used for the treatment of obesity, which may affect systemic serotonin metabolism (see Fig. 5-15E).83 Typical diet drug-associated plaques have proliferation of myofibroblast-like cells in a myxoid stroma. Similar left-sided plaques may be found in patients who receive methysergide or ergotamine therapy for migraine headaches; these serotonin analogs are metabolized to serotonin as they pass through the pulmonary vasculature. Moreover, drug-related valve disease has been reported in patients taking pergolide mesylate, an ergot-derived dopamine receptor agonist used to treat Parkinson's disease and restless leg syndrome.84

Infective Endocarditis

Infective endocarditis is characterized by colonization or invasion of the heart valves, mural endocardium, aorta, aneurysmal sacs, or other blood vessels, by a microbiologic agent, leading to the formation of friable vegetations laden with organisms (Fig. 5-16).85 Although virtually any type of microbiologic agent can cause infective endocarditis, most cases are bacterial.

The clinical classification into acute and subacute forms is based on the range of severity of the disease and its tempo, on the virulence of the infecting microorganism, and presence of underlying cardiac disease. Acute endocarditis is a destructive infection by a highly virulent organism, often involving a previously normal heart valve and leading to death within days to weeks in more than 50% of patients if left untreated. In contrast, in a more indolent lesion, called subacute endocarditis, organisms of low virulence cause infection on previously deformed valves; in this situation, the infection may pursue a protracted course of weeks to months during which the infection may be undetected and untreated.

The modified Duke criteria provide a standardized assessment of patients with suspected infective endocarditis that integrates factors predisposing patients to the development of infective endocarditis, blood-culture evidence of infection, echocardiographic findings, and clinical and laboratory information.86 The previously important clinical findings of petechiae, subungual hemorrhages, Janeway's lesions, Osler's nodes, and Roth's spots in the eyes (secondary to retinal microemboli) have now become uncommon owing to the shortened clinical course of the disease as a result of antibiotic therapy.

Vegetations composed of fibrin, inflammatory cells, and organisms are the hallmark of endocarditis. Staphylococcus aureus is the leading cause of acute endocarditis and produces necrotizing, ulcerative, invasive, and highly destructive valvular infections. The subacute form is usually caused by Streptococcus viridans. Cardiac abnormalities, such as chronic rheumatic heart disease, congenital heart disease (particularly anomalies that have small shunts or tight stenoses creating high-velocity jet streams), myxomatous mitral valves, bicuspid aortic valves, and artificial valves and their sewing rings predispose to endocarditis. In intravenous drug abusers, left-sided lesions predominate, but right-sided valves are commonly affected. In about 5 to 20% of all cases of endocarditis, no organism can be isolated from the blood (culture-negative endocarditis), often because of prior antibiotic therapy or organisms difficult to culture.87

The complications of endocarditis include valvular insufficiency (rarely stenosis), abscess of the valve annulus (ring abscess), suppurative pericarditis, and embolization. With appropriate antibiotic therapy, vegetations may undergo healing, with progressive sterilization, organization, fibrosis, and occasionally calcification. Cusp or leaflet perforation, cordal rupture, or fistula formation from a ring abscess into an adjacent cardiac chamber or great vessel can cause regurgitation. Ring abscesses are generally associated with virulent organisms, and a relatively high mortality rate.

Valve Reconstruction and Repair

Reconstructive procedures to repair mitral insufficiency of various etiologies and to minimize the severity of rheumatic mitral stenosis are now highly effective and commonplace. A recent survey of practice in the United States showed that 69% of mitral valve operations for mitral regurgitation currently use repairs.88 Reconstructive therapy of selected patients with aortic insufficiency and aortic dilation may also be done occasionally, but repair of aortic stenosis has been notably less successful. The major advantage of repair over replacement relates to the elimination of both prosthesis-related complications and the need for chronic anticoagulation therapy. Other reported advantages include a lower hospital mortality, better long-term function owing to the ability to maintain the continuity of the mitral apparatus, and a lower rate of postoperative endocarditis. Figures 5-17 and 5-18 illustrate the pathologic anatomy of various open and catheter-based mitral valve reconstruction procedures. Figure 5-19 illustrates the difficulty of surgical repairs for aortic stenosis.

Percutaneous endovascular repair approaches to valvular heart disease include balloon valvuloplasty, percutaneous placement of a mitral annular constraint device in the coronary sinus, and double-orifice edge-to-edge mitral valve repair without cardiopulmonary bypass for the treatment of mitral regurgitation. Catheter-based endovascular valve repair is most likely to be used in patients with severe disease deemed inoperable and in patients with early-stage regurgitant lesions in whom valve repair may prevent progressive ventricular dilation. Percutaneous transluminal balloon dilation of stenotic valves has been used successfully to relieve some congenital and acquired stenoses of native pulmonary, aortic, and mitral valves, and stenotic right-sided porcine bioprosthetic valves.

Mitral Stenosis

Commissurotomy may be employed in the operative repair of some stenotic mitral valves in which fibrosis and shortening of both cords and leaflets have markedly decreased leaflet mobility and area. Factors that compromise the late functional results of or technically prevent mitral commissurotomy and thereby necessitate valve replacement include: (1) left ventricular dysfunction; (2) pulmonary venous hypertension and right-sided cardiac factors, including right ventricular failure, tricuspid regurgitation, or a combination of these; (3) systemic embolization; (4) coexistent cardiac disorders, such as coronary artery or aortic valve diseases; (5) residual or progressive mitral valve disease, including valve restenosis, residual (unrelieved) stenosis, or regurgitation induced at operation; (6) advanced leaflet (especially commissural) calcification; (7) subvalvar (predominantly chordal) fibrotic changes; and (8) significant regurgitation owing to retraction.

Percutaneous balloon mitral valvuloplasty has been used to treat mitral stenosis for over two decades, with excellent success in patients with suitable valvular and subvalvular morphology.89 However, because balloon valvuloplasty largely involves separation of the fused leaflets at the commissures, this procedure is also unlikely to provide significant benefit to patients with the valve features summarized in the preceding obviating surgical commissurotomy.

Mitral Regurgitation

Reconstructive techniques are widely used to repair mitral valves with nonrheumatic mitral regurgitation.90–92 Structural defects responsible for chronic mitral regurgitation include: (1) dilation of the mitral annulus; (2) leaflet prolapse into the left atrium with or without elongation or rupture of chordae tendineae; (3) redundancy and deformity of leaflets; (4) leaflet perforations or defects; and (5) restricted leaflet motion as a result of commissural fusion in an opened position, and leaflet retraction, or chordal shortening or thickening.

Following surgical resection of excess anterior or posterior leaflet tissue in valves with redundancy, annuloplasty with or without a prosthetic ring is used to reduce the annulus dimension to correspond to the amount of leaflet tissue available. Edge-to-edge (Alfieri stitch) mitral valve repair has also been used.93 Tissue substitutes such as glutaraldehyde-pretreated xenograft or autologous pericardium can be used to repair or enlarge leaflets. Ruptured or elongated cords may be repaired by shortening or replacement with pericardial tissue or thick suture material.

Percutaneous approaches currently being evaluated for mitral regurgitation attempt to emulate one or more of the components of surgical mitral valve repair, including annular reduction and edge-to-edge mitral leaflet apposition.94 However, leaflet resection and cordal modification cannot be easily done via catheter, and there is considerable anatomical variability of the coronary sinus.95 Percutaneous approaches in various stages of development and clinical evaluation include implantation of a device in the coronary sinus, left atrium (or both), or by device placement behind the posterolateral leaflet of the mitral valve (see Fig. 5-18A and B). The goal is to plicate or straighten the posterior mitral annulus. Additional annuloplasty approaches, presently in preclinical testing, include a suture annuloplasty from the ventricular side of the mitral annulus, thermal modification of the annulus to obtain shrinkage, and a percutaneous ventricular restraint system that attempts to reshape the left ventricle. Another catheter-based approach uses an edge-to-edge clip prosthesis simulating the edge-to-edge surgical (Alfieri stitch) repair in which the midportions of the anterior and posterior mitral leaflets are clipped together (see Fig. 5-18C and D).

Aortic Stenosis

In balloon dilation of calcific aortic stenosis, individual functional responses vary considerably and data suggest a modest early incremental benefit, high early mortality, and high restenosis rate. Improvement derives from commissural separation, fracture of calcific deposits, and stretching of the valve cusps (see Fig. 5-19A). The major complications include cerebrovascular accident secondary to embolism, massive regurgitation owing to valve trauma, and cardiac perforation with tamponade. Fractured calcific nodules can themselves prove dangerous.96 In pediatric cases in which the cusps are generally pliable, cuspal stretching, tearing, or avulsion may also occur.

In calcific aortic stenosis the calcific deposits arise deep in the valve fibrous layer (see Fig. 5-14C). Their removal by sharp dissection or ultrasonic debridement generally removes a considerable fraction of the valve substance, resulting in severe compromise of mechanical integrity (see Fig. 5-19B and C).97

Valve Replacement

Severe symptomatic valvular heart disease other than pure mitral stenosis or incompetence is most frequently treated by excision of the diseased valve(s) and replacement by a functional substitute. Five key factors determine the results of valve replacement in an individual patient: (1) technical aspects of the procedure; (2) intraoperative myocardial ischemic injury; (3) irreversible and chronic structural alterations in the heart and lungs secondary to the valvular abnormality; (4) coexistent obstructive coronary artery disease; and (5) valve prosthesis reliability and host-tissue interactions.

Valve Types and Prognostic Considerations

Cardiac valvular substitutes are of two types, mechanical and biologic tissue (Fig. 5-20 and Table 5-2).98,99 Prostheses function passively, responding to pressure and flow changes within the heart. Mechanical valves are composed of nonphysiologic biomaterials and employ a rigid, mobile occluder (composed of pyrolytic carbon in contemporary valves), in a metallic cage (cobalt-chrome or titanium alloy) as in the Bjork-Shiley, Hall-Medtronic, or Omniscience valves, or two carbon hemidisks in a carbon housing (as in the St. Jude Medical, CarboMedics CPHV, and On-X prostheses). Pyrolytic carbon has high strength and fatigue and wear resistance and good thromboresistance. Tissue valves resemble natural semilunar valves, with pseudoanatomical central flow and biological material. In the past decade, innovations in tissue valve technologies and design have expanded indications for their use. Contemporary utilization of bioprosthetic tissue valves is estimated to be about 80% of all aortic and 69% of all mitral substitute heart valves.99–101 Most tissue valves are bioprosthetic xenografts fabricated from porcine aortic valve or bovine pericardium, which have been preserved in a dilute glutaraldehyde solution, and a small percentage are cryopreserved allografts.

In a recent compilation of risk models of isolated valve surgery by the Society for Thoracic Surgeons, the overall mortality was 3.4% (3.2% aortic and 5.7% mitral) and varied strongly with case mix.102 The majority of early deaths are caused by hemorrhage, pulmonary failure, low cardiac output, and sudden death with or without myocardial necrosis or documented arrhythmias. Potential complications related to mitral valve insertion include hemorrhagic disruption and dissection of the atrioventricular groove, perforation or entrapment of the left circumflex coronary artery by a suture, and pseudoaneurysm or rupture of the left ventricular free wall.

Improvement in late outcome has occurred predominantly from earlier referral of patients for valve replacement, decreased intraoperative myocardial damage, improved surgical technique and enhanced valve prostheses. Following valve replacement with currently used devices, the probability of 5-year survival is about 80% and of 10-year survival about 70%, depending on overall functional state, preoperative left ventricular function, left ventricular and left atrial size, and extent and severity of coronary artery disease.

Valve-Related Complications

Although early prosthetic valve–associated complications are unusual, prosthetic valve–associated pathology becomes an important consideration beyond the early postoperative period. Late death following valve replacement results predominantly from either cardiovascular pathology unrelated to the substitute valve or prosthesis-associated complications. In the few randomized studies of mechanical prosthetic and bioprosthetic valves (comprising Bjork-Shiley mechanical and porcine aortic bioprosthetic valves), approximately 60% or more patients had an important device-related complication within 10 years postoperatively. Moreover, long-term survival was better among patients with a mechanical valve, but with an increased risk of bleeding.103,104 Valve-related complications frequently necessitate reoperation, now accounting for approximately 10 to 15% of all valve procedures, and they may cause death. Four categories of valve-related complications are most important: thromboembolism and related problems, infection, structural dysfunction (ie, failure or degeneration of the biomaterials comprising a prosthesis), and nonstructural dysfunction (ie, miscellaneous complications and modes of failure not encompassed in the previous groups) (Table 5-3).105

Thrombosis and Thromboembolism

Thromboembolic complications are the major valve-related cause of mortality and morbidity after replacement with mechanical valves, and patients receiving them require lifetime chronic therapeutic anticoagulation with warfarin derivatives.106,107 Thrombotic deposits on a prosthetic valve can immobilize the occluder(s) or cusps, or shed emboli (Fig. 5-21). Owing to their biologic material that comprises the cusps and central flow, tissue valves are less thrombogenic than mechanical valves; their recipients generally do not require long-term anticoagulation in the absence of another specific indication, such as atrial fibrillation. Nevertheless, the rate of thromboembolism in patients with mechanical valves on anticoagulation is not widely different from that in patients with bioprosthetic valves without anticoagulation (2 to 4% per year). Chronic oral anticoagulation also carries a risk of hemorrhage. Anticoagulation is particularly difficult to manage in pregnant women.108 The risk of thromboembolism is potentiated by preoperative or postoperative cardiac functional impairment.

“Virchow's triad” of factors promoting thrombosis (surface thrombogenicity, hypercoagulability, and locally static blood flow) largely predicts the relative propensity toward and locations of thrombotic deposits.109 For example, with caged-ball prostheses, thrombi form distal to the poppet at the cage apex. Tilting disk prostheses are particularly susceptible to total thrombotic occlusion or shedding emboli from small thrombi, with the thrombotic deposits generally initiated in a flow stagnation zone in the minor orifice of the outflow region of the prosthesis. In contrast, bileaflet tilting disk valves are most vulnerable to thrombus formation near the hinges where the leaflets insert into the housing (see Fig. 5-21A). Late thrombosis of a bioprosthetic valve is marked by large thrombotic deposits in one or more of the prosthetic sinuses of Valsalva (see Fig. 5-21B), and no causal underlying cuspal pathology can usually be demonstrated by pathologic studies. Some valve thromboemboli, especially early postoperatively with any valve type, can be initiated at the valve sewing cuff before it is healed, thus providing the rationale for early antithrombotic therapy for all types.

As with other devices in which nonphysiologic artificial surfaces are exposed to blood at high fluid shear stresses, platelet deposition dominates initial blood-surface interaction, and prosthetic valve thromboembolism correlates strongly with altered platelet function. Nevertheless, although platelet-suppressive drugs largely normalize indices of platelet formation and partially reduce the frequency of thromboembolic complications in patients with mechanical prosthetic valves, antiplatelet therapy alone is generally considered insufficient to adequately prevent thromboembolism. The lack of vascular tissue adjacent to thrombi that form on bioprosthetic or mechanical valves retards their histologic organization and may prolong the susceptibility to embolization as well as render the age of such thrombi difficult to determine microscopically. Nevertheless, this feature may allow thrombolytic therapy to be an option in some cases.110

Prosthetic Valve Endocarditis

Prosthetic valve infective endocarditis (Fig. 5-22) occurs in 3 to 6% of recipients of substitute valves.111 Infection can occur early (<60 days postoperative) or late. The microbial etiology of early prosthetic valve endocarditis is dominated by the staphylococcal species, Staphylococcus epidermidis and S. aureus, even though prophylactic regimens used today target these microorganisms. The clinical course of early prosthetic valve endocarditis tends to be fulminant, and accompanied by valvular or annular destruction or persistent bacteremia. In the generally less virulent late endocarditis, a source of infection and/or bacteremia can often be found; the most frequent initiators are dental procedures, urologic infections and interventions, and indwelling catheters. The most common organisms in these late infections are S. epidermidis, S. aureus, S. viridans, and enterococci. Rates of infection of bioprostheses and mechanical valves are similar, and previous endocarditis on a natural or substitute valve markedly increases the risk.

Infections associated with mechanical prosthetic valves and some with bioprosthetic valves are localized to the prosthesis-tissue junction at the sewing ring, and accompanied by tissue destruction around the prosthesis (see Fig. 5-22A). This comprises a ring abscess, with potential paraprosthetic leak, dehiscence, fistula formation, or heart block caused by conduction system damage. Bioprosthetic valve infections may involve, and are occasionally limited to, the cuspal tissue, sometimes causing secondary cuspal tearing or perforation with valve incompetence or obstruction (see Fig. 5-22B and C). Surgical reintervention usually is indicated for large highly mobile vegetations or cerebral thromboembolic episodes, or persistent ring abscess.

Structural Valve Dysfunction

Prosthetic valve dysfunction owing to materials degradation can necessitate reoperation or cause prosthesis-associated death (Fig. 5-23). Durability considerations vary widely for mechanical valves and bioprostheses, specific types of each, different models of a particular prosthesis (utilizing different materials or having different design features), and even for the same model prosthesis placed in the aortic rather than the mitral site. Mechanical valve structural failure is often catastrophic and may be life threatening; in contrast, bioprosthetic valve failure generally causes progressive symptomatic deterioration.

Fractures of metallic carbon components (disks or housing) are unusual in most contemporary mechanical valves.112 Fractures of bileaflet tilting disk valves are rare (see Fig. 5-23A). However, of approximately 86,000 Bjork-Shiley 60- and 70-degree Convexo-Concave heart valves implanted, a cluster of more than 500 cases has been reported in which the two attachment points of the welded outlet strut fractured because of metal fatigue, leading to disk escape and often death.113 Single leg strut fractures have been noted at a presymptomatic stage in some patients.114 In contrast, structural dysfunction of tissue valves is the major cause of failure of the most widely used bioprostheses (flexible-stent-mounted, glutaraldehyde-preserved porcine aortic valves, and bovine pericardial valves) (see Fig. 5-23B–D).115 Within 15 years following implantation, 30 to 50% of porcine aortic valves implanted as either mitral or aortic valve replacements require replacement because of structural dysfunction manifested as primary tissue failure. Cuspal mineralization is the key mechanism with regurgitation through secondary tears, the most frequent failure mode, particularly in porcine aortic bioprosthetic valves. Nevertheless, there is increasing recognition that noncalcific structural damage owing to collagen fiber disruption (independent of calcification) contributes to bioprosthetic heart valve failure.116 Pure stenosis owing to calcific cuspal stiffening occurs less frequently. Calcific deposits are usually localized to cuspal tissue (intrinsic calcification), but calcific deposits extrinsic to the cusps may occur in thrombi or endocarditic vegetations. Calcification is markedly accelerated in younger patients, with children and adolescents having an especially accelerated course. Bovine pericardial valves can also suffer tearing and calcification with abrasion of the pericardial tissue an important contributing factor in some designs.117

The morphology and determinants of calcification of bioprosthetic valve tissue have been widely studied in experimental models. The process is initiated primarily within residual membranes and organelles of the nonviable connective tissue cells that have been devitalized by glutaraldehyde pretreatment procedures, and involves reaction of calcium-containing extracellular fluid with membrane-associated phosphorus. The pathologic changes in bioprosthetic valves that occur following implantation are largely rationalized on the basis of changes induced by the preservation and manufacture of a bioprosthesis, including: (1) denudation of surface cells, including endothelial cells in porcine aortic valves, and mesothelial cells in bovine pericardium; (2) loss of viability of the interstitial cells; and (3) locking of the cuspal microstructure in a static geometry.

Nonstructural Dysfunction

Extrinsic (nonstructural) complications of substitute heart valves are illustrated in Fig. 5-24. Paravalvular defects may be clinically inconsequential may aggravate hemolysis or may cause heart failure through regurgitation. Early paravalvular leaks may be related to suture knot failure, inadequate suture placement, or separation of sutures from a pathologic annulus in endocarditis with ring abscess, myxomatous valvular degeneration, or calcified valvular annulus as in calcific aortic stenosis or mitral annular calcification. Late small paravalvular leaks usually are caused by anomalous tissue retraction from the sewing ring between sutures during healing. Paravalvular defects tend to be small and difficult to locate by surgical or pathologic examination (see Fig. 5-24A).

Extrinsic factors can mediate late prosthetic valve stenosis or regurgitation, including a large mitral annular calcific nodule, septal hypertrophy, exuberant overgrowth of fibrous tissue (see Fig. 5-24B and C), interference by retained valve remnants (such as a retained posterior mitral leaflet or components of the submitral apparatus; Fig. 5-24D), or unraveled, long, or looped sutures or knots (see Fig. 5-24E). With bioprosthetic valves, cuspal motion can be restricted by sutures looped around stents, and suture ends cut too long may erode into or perforate a bioprosthetic valve cusp.

Valvular Allografts/Homografts

Aortic or pulmonary valves (with or without associated vascular conduits) transplanted from one individual to another have exceptionally good hemodynamic profiles, a low incidence of thromboembolic complications without chronic anticoagulation, and a low reinfection rate following valve replacement for endocarditis.118 Contemporary cryopreserved allografts, in which freezing is performed with protection from crystallization by dimethyl-sulfoxide and storage until use at −196°C in liquid nitrogen, have demonstrated freedom from degeneration and durability equal to or better than those of conventional porcine bioprosthetic valves (approximately 50 to 90% valve survival at 10 to 15 years compared with 40 to 60% for bioprostheses).

Morphologic changes are summarized in Fig. 5-25. Cryopreserved human allograft heart valves/conduits show gross changes of conduit calcification and cuspal stretching (see Figs. 5-25A and B). Microscopically, there is progressive loss of normal structural demarcations and cells beginning in days. Long-term explants are devoid of both surface endothelium and deep connective tissue cells; they have minimal inflammatory cellularity (Fig. 5-25C).119

Pulmonary Valvular Autografts

Often called the Ross operation in recognition of its originator, Sir Donald Ross, pulmonary autograft replacement of the aortic valve yields excellent hemodynamic performance, avoids anticoagulation, and carries a low risk of thromboembolism.120 Explanted pulmonary autograft cusps show: (1) near-normal trilaminar structure; (2) near-normal collagen architecture; (3) viable endothelium and interstitial cells; (4) usual outflow surface corrugations; (5) sparse inflammatory cells; and (6) absence of calcification and thrombus (see Fig. 5-25).121 However, the arterial walls show considerable transmural damage (probably perioperative ischemic injury caused by disruption of vasa vasorum) with scaring and loss of medial smooth muscle cells and elastin. The early necrosis and healing with probable resultant loss of strength/elasticity of the aortic wall may potentiate late aortic root dilation.122,123

Stentless Porcine Aortic Valve Bioprostheses

Nonstented (stentless) porcine aortic valve bioprostheses consist of glutaraldehyde-pretreated pig aortic root and valve cusps that have no supporting stent.124 The most widely used models, St. Jude Medical Toronto SPV (St. Jude Medical Inc., St. Paul, MN), Medtronic Freestyle (Medtronic Heart Valves, Santa Ana, CA) and Edwards Prima (Edwards Life Sciences, Irvine, CA), bioprostheses differ slightly in overall configuration, details of glutaraldehyde fixation conditions, and anticalcification pretreatment. The principal advantage of a stentless porcine aortic valve is that it generally allows for the implantation of a larger bioprosthesis (than stented) in any given aortic root, which is hypothesized to enhance hemodynamics and thereby regression of hypertrophy and patient survival.125

The available evidence suggests that the durability of stentless bioprostheses is comparable with that of contemporary stented bioprostheses. However, nonstented porcine aortic valves have greater portions of aortic wall exposed to blood than in currently used stented valves, and calcification of the aortic wall and inflammation at the junction of aortic wall within the recipient's tissue, are potentially deleterious, owing to the large area of this interface. Calcification of the wall portion of a stentless valve could stiffen the root, cause nodular calcific obstruction potentiate wall rupture, or provide a nidus for emboli. Analyses of explanted nonstented valves show pannus and tissue degeneration, manifest as tears and cuspal calcification, but not substantial aortic wall calcification.126,127

Catheter-Based Valve Replacement

New catheter techniques for inserting foldable prosthetic valves within stenotic aortic and pulmonary valves, and for emulating surgical repair of regurgitant mitral valves are in various stages of preclinical development and early clinical use.128,129 Presently, catheter-based, often percutaneous valve replacement is most widely used in patients with severe aortic stenosis disease deemed otherwise inoperable as a bridge to valve replacement in patients in whom surgery needs to be delayed, and in congenital heart disease, in which percutaneous pulmonary valve replacement may find a distinct niche to obviate the morbidity of reoperation to replace malfunctioning pulmonary conduits.

Catheter-based valve replacement uses a device which has two components: (1) an outer stentlike structure and (2) leaflets; these two components together constitute a functioning valvular prosthesis. Representative designs are illustrated in Fig. 5-26. The stent holds open a valve annulus or segment of a prosthetic conduit and resists the tendency of a vessel, valve annulus and diseased native leaflets to recoil following balloon dilation, supports the valve leaflets, and provides the means for seating of the prosthesis in the annulus or vessel.

The devices generally consist of biologic tissue such as bovine, equine, or porcine pericardium and bovine jugular venous valves mounted within a collapsable stent. The stents can be made from self-expandable or shape-memory materials such as nickel-titanium alloys (eg, Nitinol), or from balloon-expandable materials such as stainless steel, platinum-iridium, or other alloys. For a balloon-expandable device the delivery strategy involves collapsing the device over a balloon and placing it within a catheter-based sheath. The catheter containing the device can be inserted into the femoral artery (or vein) using essentially the same technique for deployment of coronary artery stents. In aortic stenosis, the device is passed from the femoral artery retrograde up the aorta to the aortic valve and deployed between the cusps of the calcified aortic valve, pushing the diseased cusps out of the way. Alternatively, the device can be deployed in an antegrade fashion through a minimally invasive surgical approach exposing the apex of the left ventricle. The transapical approach is favored in patients with significant atherosclerotic disease of the femoral artery and aorta. These devices may also play a role in the treatment of surgically implanted bioprosthetic valves that are failing because of stenosis or regurgitation in a so-called “valve-in-valve” application, in which a new valve is placed via catheter into the lumen of the existing valve.

Several devices are currently in various stages of development and clinical use in the aortic and pulmonary position. The two transcatheter aortic valves with the largest clinical experience are the Edwards SAPIEN device and the CoreValve system.130,131 The SAPIEN device is composed of a balloon expandable stainless steel stent that houses a bovine pericardial valve. The stent has a low profile and is designed to be placed in the subcoronary position. There is a polymer skirt circumferentially attached to the stent to reduce paravalvular leaks. The CoreValve device is composed of a self-expandable Nitinol stent that houses a porcine pericardial trileaflet valve. The CoreValve stent is longer and is meant to be placed in the left ventricular outflow tract extending into the aortic root. These devices have been used in approximately 4000 patients worldwide, the first-in-man experience was reported in 2002, and clinical trials are ongoing. The Medtronic Melody transcatheter pulmonary valve is composed of a balloon expandable platinum-iridium alloy stent which houses a segment of bovine jugular vein containing its native venous valve. The Melody was designed to be used in children or young adults with congenital heart disease who have received surgically implanted right ventricular outflow tract conduits that are failing either because of stenosis or regurgitation.

Catheter-based stent-mounted prosthetic valves present novel challenges. Valved stents are significantly larger than most existing percutaneous cardiac catheters and devices, and are presently on the order of 22 to 24 Fr. In the aortic position, there is the potential to impede coronary flow, or interfere with anterior mitral leaflet mobility or the conduction system or the native diseased leaflets. Stent architecture may also preclude future catheter access to the coronaries for possible interventions. Secure seating within the aortic annulus or a pulmonary conduit and long-term durability of both the stent and the valve tissue are also major considerations.

Cardiac Transplantation

Cardiac transplantation provides long-term survival and improved quality of life for many individuals with end-stage cardiac failure.132 Overall predicted 1-year survival is presently approximately 86% and 5-year survival is about 70%.133 The most common indications for cardiac transplantation, accounting for 90% of the patients, are idiopathic cardiomyopathy and end-stage ischemic heart disease; other recipients have congenital, other myocardial, or valvular heart disease.

Hearts explanted at the time of transplantation may also have previously undiagnosed conditions and unexpected findings.134 Most frequent is eosinophilic or hypersensitivity myocarditis, seen in 7 to 20% of explants and characterized by a focal or diffuse mixed inflammatory infiltrate, rich in eosinophils, and generally associated with minimal myocyte necrosis. In virtually all cases, the hypersensitivity is a response to one or more of the many heart failure medications taken by transplant candidates, including dobutamine. Several diseases responsible for the original cardiac failure can recur in and cause dysfunction of the allograft, including amyloidosis, sarcoidosis, giant-cell myocarditis, acute rheumatic carditis, and Chagas' disease.

Recipients of heart transplants undergo surveillance endomyocardial biopsies on an institution-specific schedule, which typically evolves from weekly during the early postoperative period, to biweekly until 3 to 6 months, and then approximately one to four times annually, or at any time when there is a change in clinical state. Histologic findings of rejection frequently precede clinical signs and symptoms of acute rejection. Optimal biopsy interpretation requires four or more pieces of myocardial tissue; descriptions of technical details and potential tissue artifacts are available.135

The major sources of mortality and morbidity following cardiac transplantation are perioperative ischemic injury, infection, allograft rejection, lymphoproliferative disease, and obstructive graft vasculopathy.

Early Ischemic Injury

Ischemic injury can originate from the ischemia that accompanies procurement and implantation of the donor heart. Several time intervals are potentially important: (1) the donor interval between brain death and heart removal, perhaps partially related to terminal administration of pressor agents or the release of norepinephrine and cytokines associated with brain death; (2) the interval of warm ischemia between donor cardiectomy to cold storage; (3) the interval during cold transport; and (4) the interval during rewarming, trimming, and implantation. Hypertrophy and coronary obstructions tend to promote ischemia, whereas decreased tissue temperature and cardioplegic arrest mitigate against ischemic injury. As in other situations of transient myocardial ischemia, frank necrosis, or prolonged ischemic dysfunction of viable myocardium or both may be present. Myocardial injury can cause low cardiac output in the perioperative period.

Perioperative myocardial ischemic injury may be detectable in endomyocardial biopsies. Owing to the anti-inflammatory effects of immunosuppressive therapy, the histologic progression of healing of myocardial necrosis in transplanted hearts may be delayed (Fig. 5-27). Therefore, the repair phase of perioperative myocardial necrosis frequently confounds the diagnosis of rejection in the first postoperative month, and in some cases, for as long as 6 weeks. Late ischemic necrosis suggests occlusive graft vasculopathy (see the following).

Rejection

Improved immunosuppressive regimens in heart transplant patients have substantially decreased the incidence and severity of rejection episodes. Hyperacute rejection occurs rarely, most often when a major blood group incompatibility exists between donor and recipient, and acute rejection is unusual earlier than 2 to 4 weeks postoperatively. Although acute rejection episodes occur largely in the first several months after transplantation, rejection can occur years postoperatively, rationalizing the practice of many transplant centers to continue late surveillance biopsies at late but widely spaced intervals.

Acute cellular rejection is characterized histologically by an inflammatory cell infiltrate, with or without damage to cardiac myocytes; in late stages, vascular injury may become prominent (Fig. 5-28). Until recently, the International Society for Heart and Lung Transplantation (ISHLT) working formulation from 1990 was the most widely accepted and used to guide immunosuppressive therapy in heart transplant recipients.136 In 2004, the grading system was revised as follows (with the “R” reflecting the revised system): Grade 0R—no rejection (no change from 1990), Grade 1R—mild rejection (1990 grades 1A, 1B, and 2), Grade 2R—moderate rejection (1990 grade 3A), and Grade 3R—severe rejection (1990 grades 3B and 4). The 1990 and 2004 formulations are compared in Table 5-4.

Acute antibody-mediated rejection (AMR) remains a controversial entity with a highly varied incidence among transplant centers without consensus on recognition, diagnosis either by histopathologic or immunologic testing, and clinical impact. AMR is diagnosed by nonspecific clinicopathologic findings, including: (1) cardiac dysfunction in the absence of cellular rejection or ischemic injury; (2) histologic features of interstitial edema, endothelial cell swelling, and intravascular macrophages on endomyocardial biopsy; (3) positive immunoperoxidase staining for C4d; and (4) the presence of circulating antidonor antibodies. AMR most often is diagnosed in sensitized patients (including those with previous transplantation, transfusion or pregnancy, and previous ventricular assist device use) and is associated with worse graft survival. The contribution of AMR to allograft coronary disease is unknown. Some transplant centers treat AMR with plasmapheresis.

Findings in surveillance endomyocardial biopsies that must be distinguished from rejection include lymphoid infiltrates either confined to the endocardium or extending into the underlying myocardium and often accompanied by myocyte damage (so-called Quilty lesions, which have no known clinical significance), old biopsy sites, and healing ischemic injury either in the perioperative period or resulting from graft vasculopathy. Lymphoproliferative disorders and infections also may be seen in biopsies.

Infection

The immunosuppressive therapy required in all heart transplant recipients confers an increased risk of infection with bacterial, fungal, protozoan, and viral pathogens, with cytomegalovirus (CMV) and Toxoplasma gondii remaining common opportunistic infections in this setting. Prophylaxis is typically given to patients at high risk of primary CMV infection (donor seropositive, recipient seronegative). Viral and parasitic infections can present a challenge in endomyocardial biopsies as the multifocal lymphocytic infiltrates with occasional necrosis seen with these infections can mimic rejection.

Graft Vasculopathy (Graft Coronary Arteriosclerosis)

Graft vasculopathy is the major limitation to long-term graft and recipient survival following heart transplantation.137 Up to 50% of recipients have angiographically evident disease 5 years after transplantation, whereas intravascular ultrasound identifies graft vasculopathy in 75% of patients at 3 years posttransplant. However, graft vasculopathy may become significant at any time and can progress at variable rates. We have encountered graft coronary disease in several patients as early as 6 to 12 months postoperatively at the Brigham and Women's Hospital.

Graft vasculopathy (Fig. 5-29) occurs diffusely and ultimately involves both intramyocardial and epicardial allograft vessels, potentially leading to myocardial infarction, arrhythmias, congestive heart failure, or sudden death. Although this process has been called accelerated atherosclerosis, the morphology of the obstructive lesion of graft vasculopathy is distinct from that of typical atherosclerosis (Table 5-5).

The vessels involved have concentric occlusions characterized by marked intimal proliferation of myofibroblasts and smooth muscle cells with deposition of collagen, ECM, and lipid (see Fig. 5-29A and B). Lymphocytic infiltration varies from almost none to quite prominent. The internal elastic lamina often is almost completely intact, with only focal fragmentation. The resulting myocardial pathology includes subendocardial myocyte vacuolization (indicative of sublethal ischemic injury) and myocardial coagulation necrosis (indicative of infarction).

Although the precise mechanisms of graft vasculopathy are not definitely established, there is mounting evidence that chronic allogenic immune response to the transplant and nonimmunologic factors contribute to the vascular injury. Hyperlipidemia, advanced donor age, CMV infection, perioperative graft ischemia, diabetes, and hyperhomocysteinemia have been associated with graft vasculopathy, but their roles in the evolution of the disease remain uncertain. There is no apparent difference in the frequency with which graft vasculopathy develops in patients who were transplanted for end-stage coronary artery disease and those operated for idiopathic cardiomyopathy.

Early diagnosis of graft vasculopathy is limited by the lack of clinical symptoms of ischemia in the denervated allograft, by the relative insensitivity of coronary angiography, which frequently underestimates the extent and severity of this diffuse disease, and by the exclusive or predominant involvement of small intramyocardial vessels. Histologic changes of chronic ischemia such as subendothelial myocyte vacuolization can be seen on surveillance biopsies as evidence of myocardial ischemia and may suggest graft vasculopathy (see Fig. 5-29C and D). Obstructive vascular lesions are not usually amenable to PCI or coronary artery bypass grafting because of their diffuse distribution. For most cases, retransplantation is the only effective therapy for established graft atherosclerosis.

Posttransplant Lymphoproliferative Disorders

Posttransplant lymphoproliferative disorders (PTLDs) are a well-recognized and serious complication of the high-intensity long-term immunosuppressive therapy required to prevent rejection in cardiac allografts. Several factors increase the risk of PTLD, including pretransplant Epstein-Barr virus (EBV) seronegativity (10- to 75-fold increase), young recipient age, and CMV infection or mismatching (donor-positive, recipient-negative).138

PTLDs can present as an infectious mononucleosis-like illness or with localized solid tumor masses, especially in extranodal sites (eg, heart, lungs, gastrointestinal tract). The vast majority (>90%) of PTLDs derive from the B-cell lineage and are associated with EBV infection, although T- or NK-cell origin and late-arising EBV-negative lymphoid malignancies have been described. There is strong evidence that the lesions progress from polyclonal B-cell hyperplasias (early lesions) to lymphomas (monomorphic PTLDs) in a short period of time, in association with the appearance of cytogenetic abnormalities. Therapy centers on a stepwise approach of antiviral treatment and reduction in immunosuppression, and then progression to lymphoma chemotherapy.

Cardiac Assist Devices and Total Artificial Hearts

Continuing and increasing discrepancy between the number of available donor hearts and therefore transplants performed (<2300 annually in the United States, and declining) and the number of patients in the terminal phase of heart failure and refractory to medical management (estimated at 250,000 to 500,000 in the United States, and rising) has prompted efforts in the development of ventricular assist devices (VADs), total artificial hearts, and other therapies.

Mechanical cardiac assist devices and artificial hearts have traditionally been used in two settings: for ventricular augmentation sufficient to permit a patient to survive postcardiotomy or postinfarction cardiogenic shock while ventricular recovery is occurring, and as a bridge to transplantation when ventricular recovery is not expected and the goal is hemodynamic support until a suitable donor organ is located. More recently, left ventricular assist devices (LVADs) have been shown to provide long-term cardiac support with survival and quality-of-life improvement over optimal medical therapy in patients with end-stage congestive heart failure who are not candidates for transplantation. LVADs are also being investigated as a “bridge-to-recovery” in patients with congestive heart failure to induce reverse ventricular remodeling leading to an improvement in cardiac function that would eventually allow device removal. Many types of pumps are currently under development or in clinical use as ventricular assist devices.139 Mechanical devices that entirely replace the native heart are currently in use or under development as bridge to transplant (eg, CardioWest Total Artificial Heart)140 or as destination therapy (eg, AbioCor).141

The major complications of cardiac assist devices are hemorrhage, thrombosis/thromboembolism, infection, and durability limitations (Fig. 5-30). Hemorrhage continues to be a problem in device recipients, although the risk of major hemorrhage has been decreasing with improved devices, therapies, and methods. Many factors predispose to perioperative hemorrhage, including: (1) coagulopathy secondary to hepatic dysfunction, poor nutritional status, and antibiotic therapy; (2) platelet dysfunction and thrombocytopenia secondary to cardiopulmonary bypass; and (3) the extensive nature of the required surgery.

Nonthrombogenic blood-contacting surfaces are essential for a clinically useful cardiac assist device or artificial heart. Indeed, thromboembolism occurred in most patients having long-term implantation of the Jarvik-7 artificial heart and is a major design consideration for current devices. Thrombi form primarily in association with crevices and voids, especially in areas of disturbed blood flow such as near connections of conduits and other components to each other and the natural heart (see Fig. 5-30A). Most pulsatile pumps have used smooth polymeric pumping bladders; these are frequently associated with thromboembolic complications. One approach to the design of the blood pump is the use of textured polyurethane and titanium surfaces, which accumulate a limited platelet/fibrin pseudointimal membrane that is resistant to thrombosis, allowing only antiplatelet therapy for anticoagulation in most device recipients (HeartMate I, Thoratec). Newer, continuous-flow VADs are carefully designed to minimize thrombosis, but anticoagulation therapy is typically required.

Accounting for significant morbidity and mortality following the prolonged use of cardiac assist devices, infection can occur either within the device or associated with percutaneous drive lines (see Fig. 5-30B). Susceptibility to infection is potentiated by not only the usual prosthesis-associated factors (see later), but also by the multisystem organ damage from the underlying disease, the periprosthetic culture medium provided by postoperative hemorrhage, and by prolonged hospitalization with the associated risk of nosocomial infections. Assist device–associated infections are often resistant to antibiotic therapy and host defenses, but are generally considered not an absolute contraindication to subsequent cardiac transplantation. Novel device designs, including alternative sites for driveline placement and the elimination of the driveline altogether with transcutaneous energy transmission technology, may play a role in further decreasing infection.

Other complications include hemolysis, pannus formation around anastomotic sites, calcification, and device malfunction.142 Device failure can result from fracture or tear of one of the prosthetic valves within the device conduits in pulsatile VADs, as this application provides a particularly severe test of valve durability (see Fig. 5-30C), or secondary to damage or dehiscence to the pumping bladder (see Fig. 5-30D). There is evidence that patients on VAD are more likely to develop allosensitization, which can pose a significant risk to post-transplantation outcome in the bridge-to-transplant patients.143 These complications not only have significant morbidity and can be fatal, but they may make a previously suitable patient ineligible for future transplantation.

Long-term LVADs are primarily used as a bridge to transplantation because cardiac transplantation currently offers a better long-term outlook for most patients. In a subset of patients, support by LVAD sporadically results in improved cardiac function, with heart transplantation no longer necessary even after removal of the LVAD (bridge to recovery). Left ventricular assist device support is now recognized to offer potential for myocardial recovery, a favorable outcome that is further enhanced by combination with pharmacologic therapy. LVADs lead to lowered cardiac pressure and volume overload in the myocardium followed by decreased ventricular wall tension, reduced cardiomyocyte hypertrophy, improved coronary perfusion, and decreased chronic ischemia.

To what extent recovery of myocardial function can occur during VAD implantation is uncertain. Many pathophysiologic changes occur during the progression to end-stage heart failure, ranging from the subcellular (eg, abnormal mitochondrial function and calcium metabolism) to the organ and system level (eg, ventricular dilation, decreased ejection fraction, and neurohormonal changes), leading to the signs and symptoms of congestive failure. Implantation of an LVAD can reverse many of these changes (reverse remodeling), leading to increased cardiac output, decreased ventricular end-diastolic volume, and normalization of neurohormonal status such that a small fraction of patients can be weaned from the device without the need for subsequent cardiac transplantation. Current research focuses on the mechanisms of cardiac recovery, identification of patients who could achieve recovery, and specifics such as the timing and duration of therapy; a key goal is to identify potential predictors and novel therapeutic targets capable of enhancing myocardial repair.144

Arrhythmias generally occur as a result of disorders of electrical impulse formation, disorders of electrical impulse conduction, or a combination of the two. The underlying anatomic substrates for arrhythmogenesis are many. Many of the primary cardiomyopathies can present with arrhythmias, including the genetic (eg, hypertrophic cardiomyopathy, arrhythmogenic right ventricular cardiomyopathy, and the ion channelopathies), mixed (eg, dilated cardiomyopathy), and acquired (eg, myocarditis) groups. Many secondary cardiomyopathies also may have arrhythmias as a predominant feature (eg, sarcoidosis). A common cause of arrhythmias and sudden death (especially in the older adult population) is ischemic heart disease, both in patients with and without a prior myocardial infarction. Myocardial hypertrophy and fibrosis of any etiology (eg, secondary to valvular heart disease, hypertension, or a remote infarction) also can provide the anatomical and functional substrate for the development of an arrhythmia. These underlying processes and pathologic anatomy increase the risk of spontaneous lethal arrhythmias in the setting of acute initiating events such as acute ischemia, neurohormonal activation, changes in electrolytes, and other metabolic stressors.145

Treatments for arrhythmias and their complications include pharmacologic therapy, and device therapy146 (eg, pacemakers, defibrillators), and ablation therapy.

Pacemakers and Implantable Cardioverter-Defibrillators

Modern cardiac pacing is achieved by a system of interconnected components consisting of: (1) a pulse generator that includes a power source and electric circuitry to initiate the electric stimulus and to sense normal activity; (2) one or more electrically insulated conductors leading from the pulse generator to the heart, with a bipolar electrode at the distal end of each; and (3) a tissue, or blood and tissue, interface between electrode and adjacent stimulatable myocardial cells, which is of critical importance in the proper functioning of the pacemaker. Typically, a layer of nonexcitable fibrous tissue forms around the tip of the electrode. This fibrosis may be induced by the electrode itself or may result from myocardial scarring from some other cause, most commonly a healed myocardial infarction. The thickness of this nonexcitable tissue between the electrode and excitable tissue determines the stimulus threshold, or the strength of the pacing stimulus required to initiate myocyte depolarization, and thus the amount of energy required from the pacemaker. Attempts to reduce the thickness of this layer (thereby extending battery life) include improved lead designs147 with active fixation and the use of slow, local release of corticosteroids from the lead tip. Implantable cardioverter-defibrillators (ICDs) are used in the treatment of patients who have life-threatening ventricular arrhythmias that are refractory to medical management and unsuitable for other surgical or ablative therapy, and have similar components to the pacemaker described in the preceding. These devices sense arrhythmias that can lead to sudden death and deliver therapy in the form of rapid ventricular pacing and/or a defibrillation current to terminate the dysrhythmic episode. ICDs also must overcome the barrier posed by the interfacial fibrosis at the electrode tip.

Complications from the use of pacemakers include lead displacement, vascular or cardiac perforation leading to hemothorax, pneumothorax or tamponade, lead entrapment, infection, erosion of the device into adjacent tissues owing to pressure necrosis, rotation of the device within the pocket, thrombosis and/or thromboembolism, and lead fracture, in addition to malfunction of the device itself. If the lead needs to be extracted for chronic infection or device-related defects, damage to the myocardium and/or tricuspid valve may occur secondary to encasement of the lead in fibrous tissue. Similar complications affect ICDs, with additional considerations being the consequences of repeated defibrillations on the myocardium and vascular structures, including myocardial necrosis, and the risk of oversensing (with resultant unnecessary shocks) or undersensing (with resultant sudden death). Increased attention has been paid of late to malfunctioning ICDs because of electrical flaws in a specific model that resulted in failure to terminate fatal arrhythmias.148

Ablation

Ablation involves the directed destruction of arrhythmogenic myocardium, accessory pathways, or conduction system structures to control or cure a variety of arrhythmias, including atrial flutter and fibrillation, ventricular tachycardias, and paroxysmal supraventricular tachycardias that are refractory to medical management.149,150 Ablation can be carried out as part of a surgical procedure, or through percutaneous catheter means. Electrophysiologic (EP) studies can be used to (1) provide information on the type of rhythm disturbance, (2) terminate a tachycardia by electrical stimulation, (3) evaluate the effects of therapy, (4) ablate myocardium involved in the tachycardia, and (5) identify patients at risk for sudden cardiac death.

Radiofrequency ablation acutely produces coagulation necrosis within the myocardium directly underneath the source tip, eliminating the source or pathway of the arrhythmia. The characteristic histologic changes include loss of myocyte striations, loss or pyknosis of nuclei, hypereosinophilia, and contraction bands. The edge of the fresh lesion is often hemorrhagic with interstitial edema (Fig. 5-31) and inflammation. The area undergoes the usual progression of healing similar to that in an infarct, with an early neutrophilic infiltrate, followed by macrophages to handle the necrotic debris, followed by granulation tissue formation and eventual scarring. Through the use of irrigated, cooled catheters, the lesions can penetrate deeply into the myocardium. Techniques using other forms of energy, including cryoablation, microwave, and lasers, have been used to create lesions in the myocardium for the treatment of arrhythmias.

Although metastatic tumors to the heart are present in 1 to 3% of patients dying of cancer, primary tumors of the heart are unusual.151,152 The most common tumors, in descending order of frequency, are: myxomas, lipomas, papillary fibroelastomas, angiomas, fibromas, and rhabdomyomas, all benign and accounting for approximately 80% of primary tumors of the adult heart. The remaining 20% are malignant tumors, including angiosarcomas, other sarcomas, and lymphomas. Many cardiac tumors have a genetic basis.

The most frequent cardiac tumors are illustrated in Fig. 5-32.

Myxoma

Myxomas are the most common primary tumor of the heart in adults, accounting for about 50% of all benign cardiac tumors. They typically arise in the left atrium (80%) along the interatrial septum near the fossa ovalis. Occasionally, myxomas arise in the right atrium (15%), the ventricles (3 to 4%), or valves. These tumors arise more frequently in women and usually present between the ages of 50 and 70 years. Sporadic cases of myxoma are almost always single, whereas familial cases can be multiple and present at an earlier age.

Myxomas range from small (<1 cm) to large (up to 10 cm) and form sessile or pedunculated masses that vary from globular and hard lesions mottled with hemorrhage to soft, translucent, papillary, or villous lesions having a myxoid and friable appearance (see Fig. 5-32A). The pedunculated form frequently is sufficiently mobile to move into or sometimes through the ipsilateral atrioventricular valve annulus during ventricular diastole, causing intermittent and often position-dependent obstruction. Sometimes, such mobility exerts a wrecking ball effect, causing damage to and secondary fibrotic thickening of the valve leaflets.

Clinical manifestations are most often determined by tumor size and location; some myxomas are incidentally detected in patients undergoing echocardiography for other indications, whereas others may present with sudden death. Symptoms are generally a consequence of valvular obstruction, obstruction of pulmonary or systemic venous return, embolization, or a syndrome of constitutional symptoms. Intracardiac obstruction may mimic the presentation of mitral or tricuspid stenosis with dyspnea, pulmonary edema, and right-sided heart failure. Fragmentation of a left-sided tumor with embolization may mimic the presentation of infective endocarditis with transient ischemic attacks, strokes, and cutaneous lesions; emboli from right-sided lesions may present as pulmonary hypertension. Constitutional symptoms such as fever, erythematous rash, weight loss, and arthralgias may result from the release of the acute-phase reactant interleukin-6 from the tumor leading to these inflammatory and autoimmune manifestations. Echocardiography, including transesophageal echocardiography, provides a means to noninvasively identify the masses and their location, attachment, and mobility. Surgical removal usually is curative with excellent short- and long-term prognosis. Rarely, the neoplasm recurs months to years later, usually secondary to incomplete removal of the stalk.

The Carney complex is a multiple neoplasia syndrome featuring cardiac and cutaneous myxomas, endocrine and neural tumors, as well as pigmented skin and mucosal lesions. Previously described cardiac myxoma syndromes such as LAMB (lentigines, atrial myxoma, mucocutaneous myxomas, and blue nevi) and NAME (nevi, atrial myxomas, mucinosis of the skin, and endocrine overactivity) are now encompassed by the Carney complex. The Carney complex is inherited as an autosomal dominant trait, and is associated with mutations in the PRKAR1 gene encoding the R1 regulatory subunit of cyclic AMP–dependent protein kinase A.

Histologically, myxomas are composed of stellate or globular cells (myxoma cells), often in formed structures that variably resemble poorly formed glands or vessels, endothelial cells, macrophages, mature or immature smooth muscle cells, and a variety of intermediate forms embedded within an abundant acid mucopolysaccharide matrix and covered by endothelium. Although it had long been questioned whether cardiac myxomas were neoplasms, hamartomas, or organized thrombi, it is now widely believed that they represent benign neoplasia. These tumors are thought to arise from remnants of subendocardial vasoformative reserve cells or multipotential primitive mesenchymal cells, which can differentiate along multiple lineages, giving rise to the mixture of cells present within these tumors.

Other Cardiac Tumors and Tumor-Like Conditions

Cardiac lipomas are discrete masses that are typically epicardial, but may occur anywhere within the myocardium or pericardium. Most are clinically silent, but some may cause symptoms secondary to arrhythmias, pericardial effusion, intracardiac obstruction, or compression of coronary arteries. Magnetic resonance imaging is useful in the diagnosis of adipocytic lesions because of its ability to identify fatty tissues. Histologically, these tumors are composed of mature adipocytes, identical to lipomas elsewhere. A separate, non-neoplastic condition called lipomatous hypertrophy of the interatrial septum is characterized by accumulation of unencapsulated adipose tissue in the interatrial septum that can lead to arrhythmias. Histologically, this tissue is composed of a mixture of mature and immature adipose tissue and cardiac myocytes, in contrast to the pure mature adipose tissue of a proper lipoma.

Papillary fibroelastomas153 are usually solitary and located on the valves, particularly the ventricular surfaces of semilunar valves and the atrial surfaces of atrioventricular valves. The most common site is the aortic valve, followed by the mitral valve. They constitute a distinctive “sea anemone” cluster of hairlike projections up to 1 cm or more in length, and can mimic valvular vegetations echocardiographically (see Fig. 5-32B).154 Histologically, they are composed of a dense core of irregular elastic fibers, coated with myxoid connective tissue, and lined by endothelium. They may contain focal platelet-fibrin thrombus and serve as a source for embolization, commonly to cerebral or coronary arteries. Surgical excision is recommended to eliminate these embolic events. Although classified with neoplasms, fibroelastomas may represent organized thrombi, similar to the much smaller, usually trivial, whisker-like Lambl's excrescences that are frequently found on the aortic valves of older individuals.

Rhabdomyomas comprise the most frequent primary tumor of the heart in infants and children.155 They are usually multiple and involve the ventricular myocardium on either side of the heart. They consist of gray-white myocardial masses up to several centimeters in diameter that may protrude into the ventricular or atrial chambers, causing functional obstruction. These tumors tend to spontaneously regress, so surgery is usually reserved for patients with severe hemodynamic disturbances or arrhythmias refractory to medical management. Most cardiac rhabdomyomas occur in patients with tuberous sclerosis, the clinical features of which also include infantile spasms, skin lesions (hypopigmentation, shagreen patch, subcutaneous nodules), retinal lesions, and angiomyolipomas. This disease, in its familial form, exhibits autosomal dominant inheritance, but about half the cases are sporadic owing to new mutations. Histologically, rhabdomyomas contain characteristic “spider cells,” which are large, myofibril-containing, rounded or polygonal cells with numerous glycogen-laden vacuoles separated by strands of cytoplasm running from the plasma membrane to the centrally located nucleus.

Cardiac fibromas, although also occurring predominantly in children and presenting with heart failure or arrhythmias, or incidentally, differ from rhabdomyomas in being solitary lesions that may show calcification on a routine chest radiograph.156 Fibromas are white, whorled masses that are typically ventricular. There is an increased risk of cardiac fibromas in patients with Gorlin's syndrome (nevoid basal cell carcinoma syndrome),157 an autosomal dominant disorder characterized by skin lesions, odontogenic keratocysts of the jaw, and skeletal abnormalities. Histologically, fibromas consist of fibroblasts showing minimal atypia and collagen with the degree of cellularity decreasing with increasing age of the patient at presentation. Although they are grossly well circumscribed, there is usually an infiltrating margin histologically. Calcification and elastin fibers are not uncommon in these lesions.

Sarcomas, with angiosarcomas, undifferentiated sarcomas, and rhabdomyosarcomas being the most common, are not distinctive from their counterparts in other locations. They tend to involve the right side of the heart, especially the right atrioventricular groove (see Fig. 5-32C). The clinical course is rapidly progressive as a result of local infiltration with intracavity obstruction and early metastatic events.

Peculiar microscopic-sized cellular tissue fragments have been noted incidentally as part of endomyocardial biopsy or surgically removed tissue specimens, either free-floating or loosely attached to a valvular or endocardial mass.158 Termed mesothelial/monocytic incidental cardiac excrescences (MICE), they appear histologically largely as clusters and ribbons of mesothelial cells and entrapped erythrocytes and leukocytes, embedded within a fibrin mesh. These “masses” are considered artifacts of no clinical significance formed by compaction of mesothelial strips (likely from the pericardium) or other tissue debris and fibrin, which are transported via catheters or around an operative site on a cardiotomy suction tip.

Biomaterials are synthetic or modified biologic materials that are used in implanted or extracorporeal medical devices to augment or replace body structures and functions.159,160 Biomaterials include polymers, metals, ceramics, carbons, processed collagen, and chemically treated animal or human tissues, the latter exemplified by glutaraldehyde-preserved heart valves and pericardium. Biomaterials in medical devices interact with the surrounding tissues. The first generation of biomedical materials (eg, metals used for early valve substitutes) was generally designed to be inert; the goal was to reduce the host inflammatory responses to the implanted material. By the mid-1980s, a second generation of technology, comprising bioactive biomaterials, was emerging that could interact with the host in a beneficial manner (eg, biodegradable polymer sutures, drug delivery systems, textured bladder surfaces in ventricular assist devices). In the recent past, with a greater understanding of material-tissue interactions at the cellular and molecular levels, regenerative materials are being designed to stimulate specific cellular and tissue responses at the molecular level.161,162

Biomaterial-tissue interactions comprise effects of both the implant on the host tissues and the host on the implant, and are important in mediating prosthetic device complications (Fig. 5-33).163 These interactions have local and potentially systemic consequences. Complications of cardiovascular medical devices, irrespective of anatomical site of implantation, can be grouped into six major categories: (1) thrombosis and thromboembolism; (2) device-associated infection; (3) exuberant or defective healing; (4) biomaterials failure (eg, degeneration, fracture); (5) adverse local tissue interaction (eg, toxicity, hemolysis); and (6) adverse effects distant from the intended site of the device (eg, biomaterials/device migration, systemic hypersensitivity).

Blood-Surface Interaction

Thromboembolic complications of cardiovascular devices cause significant mortality and morbidity. Thrombotic deposits can impede the function of a prosthetic heart valve, vascular graft, or blood pump, or cause distal emboli. As in the cardiovascular system in general, surface thrombogenicity, hypercoagulability, and locally static blood flow (called Virchow's triad), present individually or in combination, determine both the relative propensity toward thrombus formation and the location of thrombotic deposits with specific devices. No known synthetic or modified biologic surface is as thromboresistant as the normal, unperturbed endothelium. Like a blood vessel denuded of endothelium, foreign materials on contact with blood spontaneously and rapidly (within seconds) absorb a film of plasma components, primarily protein, followed by platelet adhesion.164 If conditions of relatively static flow are present, macroscopic thrombus can ensue. Considerable evidence implicates a primary role for blood platelets in the thrombogenic response to artificial surfaces.165 The specific physical and chemical characteristics of materials that regulate the outcomes of blood-surface interaction are incompletely understood.

Coagulation proteins, complement products, other proteins, and platelets are activated, damaged, and consumed by blood-material interactions. The clinical approach to control of thrombosis in cardiovascular devices is generally through systemic anticoagulants, particularly Coumadin (warfarin) or antiplatelet agents. Coumadin inhibits thrombin formation but does not inhibit platelet-mediated thrombosis.

Hemolysis (damage to red blood cells) in implants and extracorporeal circulatory systems results from both cell-surface contact and turbulence-induced shear forces. Erythrocytes have reduced survival following cardiopulmonary bypass. Also, with earlier model heart valve prostheses, renal tubular hemosiderosis or cholelithiasis occurred in many patients, indicative of chronic hemolysis. Implantation of synthetic vascular grafts has been demonstrated to lower platelet survival, a defect that is largely alleviated by antiplatelet therapy. Also, late following implantation, platelet survival in recipients of such grafts is largely restored; possibly as a result of graft surface “passivation” by adsorbed protein.

Tissue-Biomaterials Interaction

Synthetic biomaterials elicit a foreign body reaction, a special form of nonimmune inflammatory response with an infiltrate predominantly composed of macrophages.166,167 For most biomaterials implanted into solid tissue, encapsulation by a relatively thin fibrous tissue capsule (composed of collagen and fibroblasts) resembling a scar ultimately occurs, often with a fine capillary network at its junction with normal tissue. Ongoing mild chronic inflammation associated with this fibrous capsule is common with clinical implants; however, a substantial inflammatory infiltrate consisting of monocytes/macrophages and multinucleated foreign body giant cells in the vicinity of a foreign body generally suggests persistent tissue irritation. Although immunologic reactions have been proposed rarely for synthetic biomaterial-tissue interactions,168 and antibodies can be elicited by implantation of some materials in finely pulverized form, proven clinical cardiovascular device failure owing to immunologic reactivity is rare.

Healing of a Vascular Graft, Heart Valve Sewing Cuff, or Endovascular Stent

Healing of a vascular graft, heart valve sewing cuff, or endovascular stent derives principally from overgrowth from the host vessel across anastomotic sites; this tissue is called pannus. Grafts, fabrics, and stents used as cardiovascular implants heal primarily by ingrowth of endothelium and smooth muscle cells from the cut edges of the adjacent artery, and contact points with vessels and/or myocardium; here endothelium is present, and the tissue is called a neointima. Two additional potential mechanisms of endothelialization exist: (1) tissue ingrowth through the fabric of a graft with interstices large enough to permit ingrowth of fibrovascular elements arising from capillaries extending from outside to inside the graft, which may permit endothelial cells to migrate to the luminal surface at a large distance from the anastomosis; and (2) deposition of functional endothelial cell progenitors from the circulating blood.169

Exuberant fibrous tissue can occur at a vascular anastomosis as an overactive but physiologic repair response (Fig. 5-34). Synthetic and biologic vascular grafts often fail because of generalized or anastomotic narrowing mediated by connective tissue proliferation in the intima, and heart valve prostheses can have excessive pannus that occludes the orifice. Intimal hyperplasia results primarily from smooth muscle cell migration, proliferation, and extracellular matrix elaboration following and possibly mediated by acute or ongoing endothelial cell injury. Important contributing factors include: (1) surface thrombogenesis; (2) delayed or incomplete endothelialization of the fabric; (3) disturbed flow across the anastomosis; and (4) mechanical “mismatch” at the junction of implant and host tissues.

Diminished healing also is clinically important in certain circumstances, such as periprosthetic leak associated with heart valve prostheses. Moreover, markedly diminished endothelialization owing to toxicity of the cytostatic agents, paclitaxel and sirolimus, has been implicated in late thrombosis of drug-eluting coronary stents. For uncertain reasons, humans have a limited ability to completely endothelialize cardiovascular prostheses, and full endothelialization of clinical grafts (yielding an intact neointima) usually does not occur. Endothelial cell coverage generally is restricted to a zone near an anastomosis, typically 10 to 15 mm, thereby allowing healing of intracardiac fabric patches and prosthetic valve sewing rings, but not long vascular grafts. Thus, except adjacent to an anastomosis of a vascular graft, a compacted platelet-fibrin aggregate (pseudointima) comprises the inner lining, even after long-term implantation. Because firm adherence of such linings to the underlying graft may be impossible, dislodgment of the lining and formation of a flap-valve can occur and cause obstruction.170 Current research focuses on novel vascular graft materials that enhance endothelial cell attachment, grafts, and other implants preseeded with unmodified or genetically engineered endothelial cells, attempts to block smooth muscle cell proliferation, and engineered tissue vascular grafts (see the following).171

Infection

Infection is a common complication of implanted prosthetic devices and a frequent source of morbidity and mortality.172,173 Early implant infections (<1 to 2 months postoperatively) most likely result from intraoperative contamination or early postoperative wound infection. In contrast, late infections generally occur by a hematogenous route, and can be initiated by bacteremia induced by therapeutic dental, gastrointestinal, or genitourinary procedures. Antibiotics given prophylactically at the time of device implantation and shortly before subsequent diagnostic and therapeutic procedures may protect against implant infection.

The presence of a foreign body potentiates infection in several ways. Microorganisms may inadvertently be introduced into deep tissue locations by contamination at device implantation, bypassing natural barriers against infection. Some devices, such as many of the current left ventricular assist devices, require a percutaneous driveline, providing a continuous potential means of entry for microorganisms. The production of a biofilm (ie, a multicellular consortium of microbial cells that is irreversibly associated with a material surface and enclosed in a self-produced extracellular matrix composed primarily of polysaccharides) by the organisms is protective against the humoral and cellular immune response of the host.174 Infections associated with medical devices are characterized microbiologically by a high prevalence of organisms capable of forming protective biofilms,175 including gram-positive bacteria such as S. epidermidis, S. aureus,Enterococcus faecalis, and S. viridans, gram-negative bacteria such as Escherichia coli and Pseudomonas aeruginosa, and fungi such as Candida albicans. Some of these organisms, especially S. epidermidis and S.viridans, are of relatively low virulence in the absence of a foreign body, but are frequent causes of medical device infection. Biofilm inhibits the penetration and effectiveness of opsonizing antibodies, inflammatory cells, and antibiotics. Moreover, vasculature may be diminished and there may be necrotic tissue in the vicinity of an implant. Consequently, an implant-associated infection often persists until the device is removed.

Tissue Engineering and Cardiovascular Regeneration

Tissue engineering comprises therapeutic approaches that use living cells together with natural materials or synthetic polymers to develop or regenerate functional tissues.176–178 In the most widely used generic strategy to engineering tissues, the first step involves the seeding of cells on a synthetic polymer scaffold (usually a bioresorbable polymer in a porous configuration) or natural material (such as collagen or chemically treated tissue), each in the desired geometry for the engineered tissue, and a tissue is matured in vitro. The cells may be either fully differentiated or stem cells. The in vitro phase is done in a metabolically and mechanically supportive environment with growth media (a bioreactor), the cells proliferate, and an elaborate extracellular matrix is created in the form of a “new” tissue (the construct). In the second step, the construct is implanted in vivo into the appropriate anatomical location. Remodeling of the construct in vivo following implantation is intended to recapitulate normal functional architecture of an organ or tissue. Another collection of methods uses the in vivo environment as the bioreactor and thereby aims to recruit endogenous cells to build tissues from within. Key processes occurring during the in vitro and in vivo phases of tissue formation and maturation are: (1) cell proliferation, sorting, and differentiation; (2) extracellular matrix production and organization; (3) degradation of the scaffold; and (4) remodeling and potentially growth of the tissue.

Active research yielding progress has occurred with tissue-engineered vascular grafts, myocardium, and heart valves, as discussed in the following.

Engineered Vascular Grafts

Because of the failure of small-diameter synthetic vascular grafts, engineered blood vessels of small caliber are actively being investigated.179,180 Vascular cells have been applied onto tubular resorbable polymer scaffolds matured in vitro in a bioreactor prior to in vivo implantation.181 Exposure to pulsatile physical forces generally enhances graft properties; pulsed grafts are thicker, have greater suture retention, and higher cell and collagen density than nonpulsed engineered grafts, and have a histologic appearance similar to that of native arteries. Other investigators have fabricated mechanically sound engineered tissue vascular grafts by constructing a cohesive cellular sheet of smooth muscle cells, rolling this sheet to form the vessel media, analogous to a jelly roll, wrapping a sheet of human fibroblasts around the media to serve as an adventitia, and seeding endothelial cells in the lumen.182 Vessel “equivalents” composed of collagen and cultured bovine fibroblasts, and smooth muscle and endothelial cells have been investigated, but despite reinforcement with a Dacron mesh, such grafts have been unable to withstand burst strengths for in vivo applications.183 Another approach to vascular graft engineering utilizes naturally derived matrices with or without cell repopulation prior to implantation.184 Vascular grafts fabricated from small intestine submucosa used as experimental vascular grafts in dogs were reported to be completely endothelialized and histologically similar to arteries.185

Another approach to vascular engineering extends the concept of Sparks' silicone mandril-grown graft used clinically in the 1970s, in which a collagenous tube was formed as the fibrous capsule reactive to an implanted foreign body (silicone mandril) adjacent to a diseased vessel; the mandril was subsequently removed yielding an autologous tissue tube.186 However, owing to the variability of the quality of tissue generated in older patients in areas of circulatory insufficiency, such vascular replacements developed aneurysms when used clinically. Grafts grown as the reactive tissue that forms around silicone tubing inserted into the peritoneal cavity of rats, rabbits, and dogs, everted (so that mesothelial cells become the blood-contacting surface), and grafted into the carotid artery of the same animal remained patent up to 4 months.187

Clinical series of tissue engineered blood vessels as pulmonary artery segments and hemodialysis access grafts have been reported.188,189 Clinically employed conventional ePTFE vascular grafts have been seeded with endothelial cells at the time of implantation.190

Regeneration of Cardiac Tissue

The classical view is that the adult heart responds to mechanical overload by hypertrophy (increase in cell size) and to severe ischemic or other injury by cell death (see earlier). Functional increase in cardiac myocyte number by cell regeneration or hyperplasia has generally not been considered possible. Nevertheless, several lines of research raise enthusiasm that clinically important cardiac regeneration may indeed be possible.

Recent evidence that myocyte regeneration and death occur physiologically, and that these cellular processes are enhanced in pathologic states, has challenged the view of the heart as a postmitotic organ. Heart homeostasis may be regulated by a stem cell compartment in the heart characterized by multipotent cardiac stem cells that possess the ability to acquire the distinct cell lineages of the myocardium. They are capable of regenerating myocytes and coronary vessels throughout life and have the capacity to adapt to increases in pressure and volume loads.191 Moreover, adult bone marrow cells may be able to differentiate into cells beyond their own tissue boundary and create cardiomyocytes and coronary vessels. Myocardial tissue generated from cells, scaffolds, and bioreactors has been encouraging with respect to construct survival, vascularization, and integration, but significant functional improvement has yet to be demonstrated.192,193 Application of cyclic mechanical stretch and electrical signals have been shown to enhance cell differentiation and force of contraction.194

Animal experiments and some early-phase clinical trials lend credence to the notion that cell-based cardiac repair may be an achievable therapeutic target, including: (1) the possibility that the heart may have endogenous mechanisms of repair; (2) cell-based therapies in which fetal or adult cardiomyocytes, skeletal myoblasts or non–muscle stem or differentiated cells are injected into the heart; and (3) experimental generation of functional myocardium by cell-scaffold-bioreactor tissue engineering approaches yielding a prosthetic myocardial patch.195–197 Clinical trials have used cells derived from skeletal muscle and bone marrow, and basic researchers are investigating sources of new cardiomyocytes, such as resident myocardial progenitors and embryonic stem cells to achieve structural and functional integration of the graft with the host myocardium. Nevertheless, the most appropriate form of cellular therapy for myocardial injury remains to be identified, and it is unclear at present whether the beneficial effects observed in some studies are a result of functional and electrically integrated myocytes (the ideal), enhancement of angiogenesis owing to a local and nonspecific inflammatory reaction at the site of injection, or a “paracrine” effect, whereby transplanted cells produce growth factors, cytokines, and other local signaling molecules that are beneficial to the infarct through neovascularization and/or scar remodeling.

Engineered Heart Valves

Recent scientific and technological progress has stimulated the goal of generating a living valve replacement device that would obviate the complications of conventional valve replacement, adapt to changing environmental conditions in the recipient, and potentially grow with a growing patient.198–200 Innovative work toward this objective is active in many laboratories, and may eventually lead to clinical application. The long-term success of a tissue engineered (living) valve replacement will depend on the ability of its living cellular components (particularly VIC) to assume normal function with the capacity to repair structural injury, remodel the ECM, and potentially grow.

Tissue engineered heart valves (TEHV) grown as valved conduits from autologous cells (either vascular wall cells or bone marrow–derived mesenchymal stem cells) seeded on biodegradable synthetic polymers grown in vitro have functioned in the pulmonary circulation of growing lambs for up to 5 months.201,202 These grafts evolved in vivo to a specialized layered structure that resembled that of native semilunar valve. A recent study has shown that pulmonary vascular walls fabricated from vascular wall cells and biodegradable polymer and implanted into very young lambs enlarged proportionally to overall animal growth over a 2-year period.203

To eliminate the need for in vitro cell seeding and culture steps, an alternative tissue engineering strategy has used either a scaffold of a decellularized naturally derived biomaterial (such as animal or human allograft valve, or decellularized sheep intestinal submucosa [SIS]) implanted without prior seeding but with the intent of harnessing intrinsic circulating cells to populate and potentially remodel the scaffold.204 Tissue-derived scaffolds must possess desirable three-dimensional architecture, mechanical properties, and potential adhesion/migration sites for cell attachment and ingrowth. Decellularized porcine valves implanted in humans had a strong inflammatory response and suffered structural failure.205

Translation of heart valve tissue engineering and regenerative medicine from the laboratory to the clinical realm has exciting potential but also formidable challenges and uncertainties. Key hurdles include selection and validation of suitable animal models, development of guidelines for characterization and assurance of the quality of an in vitro fabricated tissue engineered heart valve for human implantation, and strategies to understand, monitor, and potentially control patient variability in wound healing and tissue remodeling in vivo.

Frederick J. Schoen is or has been a consultant to Celxcel, Corazon, Cordis, Direct Flow Medical, Ethicon, Edwards Lifesciences, Medtronic, Pi-R-Square, Sadra Medical, Sorin Medical, St. Jude Medical, and Sulzer Carbomedics, in the past 5 years.

Robert Padera is or has been a consultant to Atria, Direct Flow, Du Pont, Medtronic, Mitral Solutions, Mitralign, Sadra, Sorin, St. Jude, and Transmedics.