Although the term “pericarditis” is applied to many, indeed most, pericardial disorders, relatively few of the disorders actually are associated with inflammatory-cell reaction, and therefore probably most do not warrant the designation “pericarditis.” Fibrinous deposits without inflammatory cells, as occur in most patients with uremia, for example, do not justify a diagnosis of “pericarditis.” The presence of inflammatory cells in the pericardia rather than an acute clinical course or a pericardial friction rub should determine whether the term “pericarditis” is appropriate. “Pericardial heart disease” is a better general term.

Although its exact frequency from either clinical or necropsy standpoints is not certain, pericardial heart disease, as Osler emphasized, is observed more frequently at necropsy than during life. Pericardial heart disease has been found clinically in >1% of patients admitted to one large general hospital (2) and in about 5% of consecutive necropsies at another large general hospital (3). These figures were produced in the era before cardiac catheterization or operation, renal dialysis, or widespread use of antibiotics. The clinical and necropsy frequencies of pericardial heart disease in recent years are unknown. Better diagnostic tools, particularly echocardiography, and prolonged longevity in many systemic illnesses have almost certainly increased the clinical recognition of pericardial heart disease in recent years.

Nevertheless, as shown in Figure 1, pericardial heart disease is more often subclinical than clinical. Even when clinical manifestations are present, evidence of myocardial constriction is uncommon. Myocardial constriction from pericardial disease may result from pericardial effusion without pericardial thickening, from pericardial thickening without pericardial fluid, or from both effusion and thickening. As shown in Figure 2 and 3, among patients with “pericardial” constriction (in actuality, it is the myocardium that is constricted, not the pericardium), the larger the amount of pericardial fluid, the thinner the pericardia and, conversely, the thicker the pericardia, the smaller the amount of pericardial fluid.

Figure 1 The clinical and morphologic spectrum of pericardial heart disease.

Figure 2 Pericardial fluid accumulation is, in general, not proportional to pericardial thickening: the more the pericardial fluid, the less the pericardial thickening and vice versa. The dashed lines on either side of the uninterrupted line indicate that the (more ...)

Figure 3 More pericardial fluid or thicker pericardia is necessary to constrict a hypertrophied heart than a normal-sized one. RV indicates right ventricle; LV, left ventricle.

The type of clinical presentation of pericardial heart disease, i.e., acute or chronic, does not necessarily correlate with the type of morphologic derangement. The person with a thick constricting pericardium (obvious evidence of chronic disease) may present clinically with an acute picture (Figure 4). Probably more often than not, however, the clinical course follows the morphologic picture.

Figure 4 Relationship between the clinical and morphologic features in pericardial heart disease. Although clinical and morphologic acuteness and clinical and morphologic chronicity often go together, a chronic morphologic process can present acutely.

The normal parietal pericardium (fibrous sac or fibrous pericardium), the bag that encloses the heart, consists of a 1-mm-thick layer of dense collagen (fibrous layer) with sparse interspersed elastic fibrils, and it is covered by a layer of mesothelial cells (serous layer) (Figure 5). The visceral pericardium (epicardium or serous pericardium) is the surface of the heart itself and consists of a thin (>1 mm) layer of loose fibrous tissue covered by mesothelial cells (serous layer). The serous layer of parietal pericardium is continuous with the serous layer of visceral pericardium at or near sites of attachment of the great vessels to the heart. The portion of epicardium that covers the vessels is arranged in the form of 2 tubes: the ascending aorta and pulmonary trunk are enclosed in the first (arterial mesocardium), and the superior and inferior venae cavae and the 4 pulmonary veins are enclosed in the second (venous mesocardium). The attachment of the latter to the parietal pericardium is in the shape of an inverted U, and the cul-de-sac enclosed between the limbs of the U lies behind the left atrium and is known as the oblique sinus. The passage between the venous and arterial mesocardia, i.e., that between the aorta and pulmonary trunk anteriorly and the atria posteriorly, is the transverse sinus.

Figure 5 Normal parietal pericardium. A single layer of mesothelial cells covers the layer of dense fibrous tissue on the cardiac side. On the mediastinal side of the fibrous layer is adipose tissue, and in it are located vascular channels and nerves. Masson stain, (more ...)

The flask-shaped bag of parietal pericardium is closed at its neck by fusion with the adventitia of the great vessels; its base is attached to the tendinous and muscular portions of the left-sided diaphragm. It is also attached to the posterior surface of the sternum by the superior (to manubrium) and inferior (to xiphoid process) pericardiosternal ligaments. Between the left main pulmonary artery and subjacent pulmonary vein is a triangular fold of visceral pericardium known as the ligament of the left vena cava, a remnant of the lower part of the left superior vena cava. The sinus node is located near the attachment of the parietal pericardium to the superior vena cava, and consequently it may be “irritated” by a pericardial process, producing an arrhythmia or a conduction disturbance.

Electron microscopic studies of the normal parietal pericardium show that the tissue is composed of 3 layers: 1) the serosa, consisting of a surface layer of mesothelial cells and a narrow submesothelial space, 2) the fibrosa, containing variously oriented layers of collagen fibrils and small elastic fibers, and 3) the epipericardial connective tissue, containing mainly large coarse bundles of collagen and forming part of the pericardiosternal ligament (4). Scanning electron microscopic studies show the mesothelial cells to contain single cilia covered with microvilli. The latter appear to bear friction and to increase the surface area for fluid transport.

Beneath the visceral pericardium is either myocardium or adipose tissue. The amount of subepicardial adipose tissue increases, up to a point, with age. Large quantities are rare in persons under 10 years of age and over 90 years of age. The deposits are largest in the atrioventricular sulci and, next, around the anterior and posterior descending coronary arteries. At times, the entire surface of the heart is covered by fat (Figure 6). In general, the amount of subepicardial adipose tissue is proportional to the amount of fat present in other body tissues (5, 6). At times, however, considerable quantities of subepicardial adipose tissue are present in persons of normal body weight, and the amount may not be excessively increased in persons of huge body weight (<135 kg [300 lb]). Corticosteroid therapy causes the amount of subepicardial tissue to increase (7, 8). In persons with debilitating diseases such as malignant neoplasms or chronic starvation, the subepicardial fat may atrophy, giving it a watery, gelatinous appearance.

Figure 6 Massive increase in subepicardial adipose tissue in an 82-year-old man with idiopathic hemochromatosis. The myocardium was completely covered by fat. The heart weighed 500 gm. (a) Anterior view. (b) Posterior view. (c) Longitudinal section showing the (more ...)

When subepicardial adipose tissue is excessive, strands of fat cells also extend into adjacent myocardium. This occurrence is most frequent in the walls of the right ventricle and right atrium. The fat may extend through the myocardial wall and infiltrate the endocardium. Although they had no hemodynamic documentation, many writers in the 1800s believed that excessive subepicardial adipose tissue could cause cardiac dysfunction. Osler, writing in 1892, stated that excessive subepicardial fat “occasionally leads to dangerous or even fatal impairment of the contractile power of the heart” (1). Sudden nontraumatic death in the late 1800s was often attributed to a “fatty heart.”

Although normally the serous layer of parietal pericardium contacts the serous layer of visceral pericardium, up to 50 mL of fluid, which has a chemical composition similar to that of serum, normally is present in the pericardial sac (9). The pericardial fluid, in essence, serves as a “lubricating oil” to diminish the amount of friction between the moving heart and the adjacent tissues.

In the adipose tissue beneath the visceral pericardium and on the mediastinal aspect of the parietal pericardium are located vascular channels, including arteries, veins, and lymphatics, and also nerves. None of these structures is located within the fibrosa component of the parietal pericardium. The arteries of the pericardium arise from the internal mammary artery and its musculophrenic branch and from the descending thoracic aorta. Branches of the vagus and phrenic nerves and from the sympathetic trunk supply the pericardium.

Normally, the visceral pericardium is translucent and the underlying subepicardial tissue or myocardium is readily visible. In many individuals, the epicardium is focally white (Figure 7), and then the underlying structures are obscured. These foci (variously called milk spots, soldiers' spots, tendinous patches, and maculae tendineae) most commonly are present over the anterior surface of the right ventricle, often are multiple, and increase in incidence and size with age and with cardiac enlargement. Histologically, these foci consist of dense collagen, occasionally with underlying small collections of mononuclear cells (lymphocytes). Consequently, they may most appropriately be called “collagen plaques.” They are usually absent in the first decade of life and nearly always present by the sixth decade.

Figure 7 Epicardial collagen plaques (soldier's plaques, milk spots) in the hearts of 4 different patients. (a) Collagen plaques are present over the right ventricle and over the left ventricular apex. (b) A discrete collagen plaque is visible over the right ventricle. (more ...)

When the right ventricle is enlarged, the collagen plaques over its anterior surface may be quite prominent and presumably result from contact of the beating heart with the undersurface of the sternum. In conditions producing considerable enlargement of the right atrium—for example, atrial septal defect, chronic lung disease, or cardiomyopathy—the collagen plaques may occur over the right atrium (10). In kyphoscoliosis, the plaques often are found on the posterior surface of the heart, presumably due to contact with the vertebral bodies during cardiac motion. In conditions that cause left ventricular enlargement, particularly systemic hypertension and left ventricular outflow tract obstruction, collagen plaques often are present over the apex of the left ventricle. Plaques in this location may be anatomic expressions of the “point of maximal impulse” and result from contact of the left ventricular apex with the undersurface of the anterior ribs during ventricular systole. In patients with large amounts of subepicardial adipose tissue, collagen plaques may be less numerous, presumably since the epicardial fat acts as a cushion or absorbing vehicle for the underlying myocardium.

Although pericardial disease has many causes, the morphologic reaction of the pericardia to injury is rather limited. The pericardium reacts to acute injury by exuding fluid, fibrin, or cells or combinations of these three. The type of fluid and/or cells exuded is determined by the cause of the pericardial disease. There is no pericardial reaction to either serous fluid or to blood in the pericardial sac. If, however, the lipid component of pooled lysed erythrocytes is injected into the pericardial sac, a fibrinous reaction may result and fibrous or even cholesterol pericardial heart disease may develop (11). Blood in the pericardial space of injured pericardia, however, may lead to fibrous pericardial adhesions. Isolated injury to the serosal pericardia of rabbits without associated bleeding causes only a fibrinous reaction; if blood is injected after the serosal injury, pericardial adhesions may result (12). Deposits of fibrin alone on the pericardial surfaces apparently cause no permanent reaction.

If microorganisms enter the pericardial space or tissues, the reaction appears to depend on the type of infecting organism. Viruses, the presumed cause of “acute benign pericarditis,” generally produce only a transient pericardial reaction, which usually resolves (13). Acid-fast organisms produce a mononuclear cell reaction and generally go on to cause severe fibrous thickening of the pericardia. Pyogenic organisms cause a violent polymorphonuclear cell reaction that also may progress to fibrous thickening with or without constriction.

General morphologic responses to injury are listed in Tables 1 and 2. Although their occurrence in pure form is unusual, the recognition of the dominant morphologic component is useful in determining the cause of pericardial heart disease in any particular patient. The causes of each of the primary morphologic components are discussed more thoroughly in the remainder of this report.

Table 1 Morphologic classification of pericardial heart disease

Table 2 Etiologic and morphologic classifications of pericardial heart disease

Fibrin deposits on the pericardial surfaces are observed commonly at necropsy. The deposits may be focal or diffuse, and they may or may not be associated with an increased quantity of pericardial fluid. Every type of pericardial disease probably is associated with fibrinous deposits at one time, and probably most cases resolve without residua. Fibrinous deposits are not capable by themselves of causing myocardial constriction. There are 6 general causes of fibrinous pericardial heart disease (Table 3).

Table 3 Etiology of fibrinous pericardial heart disease

Figure 8 Diffuse fibrinous pericarditis from acute myocardial infarction. This 56-year-old man developed a precordial friction rub on day 6 following infarction and died on day 16. (a) A deposit of fibrin covers the mildly thickened epicardium. (b) The left circumflex (more ...)

Figure 9 (a) False vs (b) true left ventricular aneurysms with adherent parietal pericardia. The wall of the false aneurysm is parietal pericardium, which had adhered to the epicardium before the left ventricle ruptured. In the true aneurysm, the parietal pericardium (more ...)

Irradiation to the mediastinum and chemical agents initially cause a fibrinous pericardial reaction that also may heal by fibrous obliteration (18).

Acute rheumatic fever usually is associated with a subclinical fibrinocellular cell reaction (27, 28). This pericardial reaction appears to resolve because evidence of previous pericardial disease is rarely observed at necropsy in patients with rheumatic heart disease. Calcific pericardial deposits also may occur in patients with rheumatic heart disease, but they too are rare and, when present, are focal and never associated with myocardial constriction.

In the precorticosteroid era, fibrinous pericardial heart disease occurred in about 60% of patients with systemic lupus erythematosus (7, 8). Today, a fibrous obliterative pericardial reaction is observed more commonly (8). A fibrinous pericardial reaction also is observed at necropsy in about 40% of patients with rheumatoid arthritis (29).

The morphologic features of pericardial disease resulting from hypersensitivity states and drug reactions are not certain, but it is presumed that the reaction is of a fibrinous type.

Excessive (<50 mL) effusion into the pericardial sac is extremely common, probably far more common than generally appreciated, because the amounts rarely are measured accurately at necropsy. Serous effusions up to 200 mL, as occur commonly in patients with congestive cardiac failure or hypoalbuminemia, may be overlooked entirely at necropsy.

The amount of pericardial fluid necessary to cause ventricular constriction is highly variable and dependent on several factors: 1) the time period required to accumulate the fluid, 2) the amount of thickening of the underlying or overlying pericardia, 3) the amount of muscle mass (weight) of the cardiac ventricles, and 4) the blood volume. The more rapidly fluid accumulates, the greater the chance that myocardial dysfunction (constriction) will result. Fluid volumes many times the normal amount may be present without myocardial constriction if the accumulation is slow (Figure 2). The greater the thickness of the parietal or visceral pericardia or both, the smaller the amount of fluid needed to produce signs or symptoms of constriction. The thicker the ventricular walls, the greater the amount of fluid required for constriction. More fluid is required, for example, to constrict the hypertensive ventricle than the normotensive ventricle (see Figure 3).

The effusions can be serous, bloody, or lymphatic/chylous (Table 4).

Table 4 Etiology of effusion (<50 mL) into pericardial space

Figure 10 Hemorrhagic pericardial heart disease in chronic renal disease. This 26-year-old woman had renal failure and pericardial effusion on admission 3 weeks before death. Pericardiocentesis was performed twice, but cardiac tamponade progressed. At necropsy, (more ...)

Figure 11 Hemorrhagic pericardial heart disease in a 30-year-old man with Gaucher's disease and pulmonary hypertension (from Gaucher cells plugging pulmonary capillaries). Chest radiographs (a) 2 months before death and (b) day of death, demonstrating marked enlargement (more ...)

Figure 12 Focal epicardial hemorrhages in hypothrombocytopenia. The anterior surface of the heart shows focal hemorrhages in a 56-year-old man with acute myelogenous leukemia. The epicardial hemorrhages seen here are typical of those observed in many patients with (more ...)

Infective pericarditis is the presence of pus, microorganisms, or both in the pericardial sac or tissues. Nontuberculous infective pericarditis is infrequent. In the preantibiotic and prethoracic surgical eras, the infection was usually spread from the lung, pleura, mediastinum, or abdomen or was a complication of a generalized septicemia. Today, the most common predisposing factors are cardiothoracic operations, immunosuppressive therapy, trauma, rupture of the esophagus into the pericardial sac secondary to neoplasm (Figure 13), and infective endocarditis with rupture of a ring abscess, septic coronary embolus, or rupture of a myocardial abscess (43) (Figure 14).

Figure 13 Purulent pericarditis in the exterior of the heart in a 56-year-old man with squamous cell carcinoma arising in the floor of the mouth. He developed a neoplastic abscess cavity in the left posterior mediastinum, and it opened into the esophagus, bronchus, (more ...)

Figure 14 Schematic portrayal of the pathogenesis of pericarditis in infective endocarditis. Ao indicates aorta; LV, left ventricle; LA, left atrium.

The expected increase in frequency of purulent pericarditis with the advent of cardiothoracic operations has been offset by the widespread use of antibiotics and the decrease in bacterial infections in general. Indeed, purulent pericarditis is less frequent today than in the past (44). The occurrence of infective pericarditis as a consequence of infective endocarditis is not widely appreciated (45). In a review by Buchbinder and Roberts of 1284 patients with fatal infective endocarditis, 172 (13%) had pericarditis at necropsy. The pericarditis in this circumstance need not be purulent (45). Figure 14 shows several mechanisms of formation of pericarditis in patients with infective endocarditis.

The organisms responsible for nontuberculous infective pericarditis have changed considerably since the introduction of antibiotics, cardiothoracic surgery, and immunosuppression. In the preantibiotic era, a gram-positive bacterium was responsible in 80% and a gram-negative bacterium in 4% of cases (Table 5). These figures were from the study by Cabot of 186 necropsy patients with purulent pericarditis (28). In a later study from the same hospital (Massachusetts General) of 26 patients with purulent pericarditis, gram-positive bacteria were responsible in 42%, gram-negative bacteria in 39%, and fungi in 19% of cases (43) (Table 5). In none of Cabot's patients were fungi the causative agent. Furthermore, the types of gram-positive bacteria causing infective pericarditis have changed considerably. Streptococcus and pneumococcus were most common in the preantibiotic era and staphylococcus in the recent antibiotic era (Table 5).

Table 5 Organisms responsible for nontuberculous infective pericardial heart disease

Purulent pericarditis with or without antibiotic treatment has a poor prognosis. It nearly always indicates a widespread infection of the mediastinum and lung. Although tamponade and constriction may occur, drainage of the pericardial pus usually is not lifesaving because the infective process is so extensive.

Although parasites can infect pericardium, all are rare. About 100 patients with amebic pericarditis have been reported, and in each spread was from a hepatic abscess (46). Before actual perforation into the pericardial space, amebic pericarditis is described classically as “anchovy-sauce” pus, but Entamoeba histolytica organisms are infrequently found in it (46). A rapid course to constriction may occur. Pericardial involvement by toxoplasmosis and echinococcosis is extremely rare, and cases of pericardial involvement by dracunculosis have been reported (47).

Fungal pericarditis has increased in frequency with the use of immunosuppressive agents. Although usually granulomatous, fungal pericarditis may be purulent. Coccidioides immitis has been found in pericardium, and clinical evidence of pericarditis occurs in about 15% of patients with pulmonary coccidioidomycosis (48). Pericarditis occurred in 2% of 470 necropsied patients with actinomycosis (49), but sulfur granules were rare. Pericardial calcific deposits may occur in patients who react to histoplasmosis but not to other fungal antigens, but acute pericarditis has not been reported in patients with histoplasmosis. Pericardial candidiasis has been observed in patients with leukemia receiving immunosuppressive drugs (Figure 15).

Figure 15 Candida tropicalis pancarditis in a 7-year-old girl with acute lymphocytic leukemia and a 21-month illness. The child never had signs or symptoms of cardiac disease but died of disseminated candidiasis. (a) Section of the right atrial wall with fungi (more ...)

Pericarditis has been seen in association with several viral diseases, including infectious mononucleosis, atypical pneumonia, mumps, measles, smallpox, and influenza. Although viral titers may rise, morphologic confirmation of a viral etiology is lacking.

Pericarditis probably is most frequent from Coxsackie viruses (50). Biopsy of parietal pericardium, however, has not revealed intracellular virions, although viral antigens may be present. Indeed, almost no morphologic information is present on viral pericarditis in humans.

Fibrous pericardial heart disease is thickening of the pericardia or the presence of adhesions between the visceral and parietal pericardia or both. Either the thickening or the adhesions may be focal or diffuse. The thickening may involve both visceral and parietal pericardia or only one of them. The collagen plaque, for example, is a form of fibrous pericardial heart disease with focal thickening, usually limited to the epicardium. Pericardial thickening may occur in the absence of pericardial adhesions, adhesions may occur in the absence of pericardial thickening, or the two may occur together. When pericardial adhesions are diffuse, the pericardial space is obliterated, and this circumstance has been called “obliterative pericarditis.”

“Obliterative pericarditis” may occur in the absence of thickening of the parietal pericardium. Obviously, since its space no longer is present, pericardial effusion, by definition, cannot be present in the patient with “obliterative pericarditis.” A pericardial effusion may be present in patients with focal pericardial adhesions or in patients with diffuse pericardial thickening with either no adhesions or only focal adhesions.

There are many causes of pericardial thickening and pericardial adhesions. The most common are listed in Table 6.

Table 6 Etiology of fibrous (adhesive) pericardial-heart disease

Mediastinal irradiation in high doses in humans, with resulting pericardial injury, occurs primarily in patients with Hodgkin's disease (18, 52, 53), because it is the most common mediastinal tumor irradiated in sufficient doses to cause cardiac disease (Figure 16). Pericardial injury probably occurs in most patients receiving ≥4000 cGy to the mediastinum. Fibrous thickening of the pericardia combined with pericardial effusion (usually serous) has been the most frequent consequence of this irradiation. Clinical evidence of cardiac dysfunction resulting from pericardial disease occurs in about 5% of patients with Hodgkin's disease treated with high-dose mediastinal irradiation (31). Cardiac tamponade may occur, usually a result of combined pericardial effusion and fibrous thickening of the parietal pericardium (54). The coronary arteries lying in the subepicardial adipose tissue also may be damaged by irradiation, with resulting coronary arterial luminal narrowing and symptomatic or fatal ischemic heart disease (52, 53).

Figure 16 Radiation pericarditis. Stage IIIB Hodgkin's disease was diagnosed in this 54-year-old man, and he then received radiation therapy to the mediastinum. Eighteen months later, he developed recurrent pericardial effusion and signs of pericardial constriction, (more ...)

A diagnostic dilemma occurs in patients who have mediastinal irradiation for a tumor and then develop evidence of pericardial effusion. Is the pericardial effusion radiation induced or is it due to contiguous spread of the neoplasm to the pericardium? It is of some value to know that pericardial effusion of sufficient magnitude to be detected by echocardiogram commonly is found after mediastinal irradiation, that it may lead to pericardial tamponade, and that the peak evidence of the pericardial reactions appears from 2 to 8 months after radiation therapy, although a wide range of time from irradiation to onset of pericardial symptoms has been reported.

At necropsy, nearly 50% of patients with rheumatoid arthritis have fibrous obliterative pericardial heart disease and, rarely, calcific or cholesterol pericardial disease (55). Usually, no functional consequence results from the obliterative pericardial heart disease.

Among 36 patients with systemic lupus erythematosus (SLE) studied at necropsy (7, 8), 19 had pericardial heart disease: fibrinous in 4, purulent in 2, and fibrous in 13. Before corticosteroids were introduced, the fibrinous type was the most common seen in patients with SLE. Among patients with SLE treated with corticosteroids, the fibrous type is the most common. Among 19 SLE patients with pericardial heart disease at necropsy, none had evidence of myocardial dysfunction as a consequence of the pericardial involvement. Eleven of the 19 had pericardial friction rubs. In patients with SLE, pericardial heart disease serves as an excellent marker of the presence of associated endocarditis: endocarditis was present at necropsy in 18 of the 19 SLE patients with and in none of the 17 SLE patients without pericardial disease.

Although uncommon clinically, pericardial heart disease is common at necropsy in patients with scleroderma (56). The fibrous obliterative type is the most common form in scleroderma. The fibrinous type, seen less commonly, may result from associated renal failure rather than scleroderma per se (56). Among the scleroderma patients I studied at necropsy, a third had pericardial involvement, fibrous in each.

Table 7 Etiology of constrictive pericardial heart disease in the USA treated by pericardiectomy

An interesting entity called “occult constrictive pericardial disease” has been described. This syndrome consists of nonspecific symptoms of fatigue, dyspnea, and chest pain in patients, most of whom had a history of pericarditis in the past. They had abnormal hemodynamics at cardiac catheterization until they were challenged by an infusion of 1 L of warm saline. They then developed hemodynamic findings characteristic of constrictive pericarditis. These findings were absent in control patients with heart disease similarly challenged with volume expansion. Pericardiectomy relieved the symptoms and altered the hemodynamic findings in response to saline infusion in all patients. The pericardium was abnormal in all subjects.

Administration of procainamide, methysergide, or practolol has been well documented to result in constrictive pericarditis. One wonders whether a significant percentage of patients who have “idiopathic constrictive pericarditis” may have developed it as a complication of drug therapy.

As seen in Figure 17, the heart is involved in approximately 10% of patients with malignant neoplasms, and 85% of the patients with cardiac involvement have neoplastic involvement of the pericardium (59). Furthermore, only about 10% of patients with cardiac involvement have clinical evidence of cardiac disease, and in 90% of those the clinical dysfunction is the result of pericardial involvement, usually constriction by either pericardial effusion or neoplastic thickening of pericardia or both. The most common neoplasms in absolute numbers with cardiac metastases are lung (in men) and breast (in women), followed by leukemia and lymphoma (Table 8). Nearly any tumor, however, may metastasize to the heart. Among patients with carcinoma of the breast or lung, about 10% have neoplastic involvement of the heart (59). Because spread of these 2 tumors to the heart appears to be by direct extension, the pericardia virtually always are involved.

Figure 17 The frequency of pericardial involvement in malignant neoplasms.

Table 8 Etiology of neoplastic pericardial heart disease

Among specific malignant neoplasms, the ones with the highest percentage of metastases to the heart are melanoma (70%) (60), leukemia (37%) (61), and lymphoma (24%) (62). As with the tumors that spread by direct extension, melanoma, leukemia, and lymphoma invade the pericardia in 85% of the patients with cardiac metastases, and spread of these neoplasms appears to be by the hematogenous routes.

Although pericardial effusions, often bloody, are frequent in patients with cardiac metastases, pericardial effusions also are common in patients with malignant neoplasms without cardiac metastases (59, 62). Pericardial effusions (<50 mL) were recorded in 17% of 48 patients with cardiac lymphoma and in 14% of 148 patients with lymphoma but without cardiac metastases (62). In both groups, the effusions were serous and probably the result of hypoalbuminemia (>3.5 gm/dL), which was present in 95% of the 196 necropsy patients with lymphoma (62).

Pericardial heart disease may occur in patients with malignant neoplasms who do not have involvement of the heart (including pericardia) by neoplasm. Varying types of the disease were present in 21% of 48 patients with cardiac lymphoma and in 7% of 148 patients with lymphoma but without metastases to the heart (62).

Symptoms and signs in patients with cardiac metastases often are attributed to cardiac involvement by the neoplasm, but this may not be justified. Dyspnea, for example, was present in 27% of 48 patients with and in 52% of 148 patients without cardiac lymphoma (62). Likewise, precordial or substernal chest pain (6%), precordial murmurs (35%), and ventricular gallops (5%) were present in the same frequencies among patients with lymphoma irrespective of the presence or absence of cardiac metastases (62). Surprisingly, electrocardiographic abnormalities were present in identical frequency among patients with lymphoma with and without cardiac tumor deposits: 62% in each of the 2 groups (62).

Primary pericardial neoplasms are extremely rare, and about half of them are benign (59). The most common is teratoma (63). Although usually histologically benign, teratoma can produce fatal compression of a cardiac chamber. Hemangioma, leiomyofibroma, lipoma, and fibroma are other “benign” pericardial neoplasms. The hemangioma may rupture to cause hemorrhagic pericardial effusion with or without tamponade.

Among the primary malignant pericardial neoplasms, mesothelioma is by far the most frequent. It may totally encase the heart, resulting in fatal myocardial constriction.

Table 9 Etiology of granulomatous pericardial heart disease

The frequency of tuberculous pericarditis has diminished in recent years, presumably paralleling the reduction in incidence of pulmonary tuberculosis. Osler observed tuberculosis in 215 (22%) of 1000 necropsies, and 7 (3%) of the 215 had pericardial involvement (65). Subsequent studies (66) have shown pericarditis to be present at necropsy in 8% of patients with pulmonary tuberculosis.

The pathogenesis of pericardial involvement by tuberculosis is not clear. Pericardial disease probably results from early dissemination following a primary infection. The infection appears to reach the pericardium either from the blood (miliary) or, more commonly, by retrograde lymphatic spread from infected mediastinal glands. The pericardial infection rarely arises from direct spread from the lung or pleura.

Four stages of tuberculous pericarditis have been described (67). The fibrinous stage is characterized by diffuse deposits of fibrin associated with a granulomatous reaction. This stage is observed rarely, and there is some debate as to whether it can resolve or whether it always goes on to the second stage, which is characterized by effusion into the pericardial sac. The effusions, which may be massive (<2000 mL), usually are serous or serosanguineous but may be bloody or turbid. Many lymphocytes may be present in the effusion, but tuberculous organisms are uncommonly observed morphologically in the fluid or cultured from it.

By stage 3, the parietal pericardium is considerably thickened by both fibrous tissue and granulomas, and most of the fluid has resolved. If the parietal pericardium becomes markedly thickened, myocardial constriction may occur, and this is considered stage 4. By this time, the pericardial space may be completely obliterated by diffuse fibrous adhesions. Eventually the granulomas apparently can be replaced entirely by fibrous tissue and, if observed morphologically at this point, the pericardial disease would have to be called idiopathic rather than granulomatous (tuberculosis). Calcific deposits with or without bone formation may occur in the markedly thickened pericardia. Progression from stage 3 to stage 4 may occur despite antituberculous therapy.

Tuberculous pericarditis by no means always progresses to the constrictive fibrous stage (Figure 18). Heimann and Binder (68) described 31 patients with granulomatous pericarditis; none had constriction and all had tuberculous hilar lymph nodes. Among patients with “chronic” constrictive pericarditis, however, tuberculosis remains an important cause, accounting for nearly 10% of the cases in the USA (Table 7). In some areas of the world, a tuberculous cause of constrictive pericarditis is more common than a nontuberculous cause. Among 61 patients with constrictive pericarditis undergoing pericardiectomy in India, the etiology was tuberculous in 37 (61%) and nontuberculous in 24 (39%) (69).

Figure 18 Tuberculous pericarditis in a 45-year-old dyspneic man whose chest radiograph showed complete loss of volume of the left lung and massive left pleural effusion. Congestive cardiac failure and respiratory insufficiency led to death. At necropsy, the entire (more ...)

Fungi and parasites are rare causes of granulomatous pericarditis.

Figure 19 Talc-induced granulomatous pericarditis in a 92-year-old woman who was stated to have had angina pectoris beginning at age 63. At age 67, she underwent talcum-powder sprinkling on her pericardium (Beck or Thompson procedure). Immediately preoperatively, (more ...)

Figure 20 Sarcoidosis of the pericardium in a 47-year-old woman who was well until 16 months before death, when exertional dyspnea appeared. Thereafter, there was clinical evidence of recurring pleural and pericardial effusions. She died of progressive respiratory (more ...)

The occurrence of calcific deposits in pericardia constitutes calcific pericardial heart disease. The calcific deposits may vary in size from microscopic, i.e., visible only by histologic examination, to massive, i.e., encircling all or most of the heart and readily visible by radiographic examination. Calcific deposits appear to represent the end stage of organization of a pericardial process and in themselves are not indicative of any specific etiology. Histologic examination of calcified pericardia rarely provides specific diagnoses. Conditions predisposing to calcific pericardial heart disease are listed in Table 10. The etiology is based on historical data or an associated abnormality rather than on specific pericardial alteration. There is strong suggestive evidence that large calcific pericardial deposits represent “burnt out” pericardial tuberculosis.

Table 10 Etiology of calcific pericardial heart disease

At times, huge calcific deposits may be present in the pericardia without evidence of myocardial constriction or other myocardial dysfunction (76) (Figure 21). For myocardial constriction to occur, the calcific deposits must encircle both ventricles (Figure 22). Localized bands of calcium, however, may cause cardiac dysfunction. The occurrence of a large band of calcium transversing the right ventricular conus has led to right ventricular outflow obstruction. An anular band of calcium in the right or left atrioventricular sulcus may delay or inhibit ventricular filling and consequently simulate tricuspid or mitral stenosis.

Figure 21 Calcific pericardial heart disease in a 78-year-old man, an alcoholic with cirrhosis, esophageal varices, and chronic pancreatitis, who died of a gastrointestinal hemorrhage. No evidence of cardiac dysfunction was ever present clinically. At necropsy (more ...)

Figure 22 Calcific constrictive pericardial heart disease in a 68-year-old man who had had pericardiectomies for relief from constriction 15 and 7 years before death. Terminally, evidence of myocardial constriction reappeared. Chylous ascites secondary to obstruction (more ...)

Clinically, cholesterol pericardial heart disease most often is characterized by slowly developing, nonconstricting large effusions (77). Although the pericardial fluid may be turbid or clear, cholesterol crystals give the effusion a “gold paint” appearance. The parietal and visceral pericardia may become markedly thickened; histologic examination shows giant cells containing cholesterol clefts, mononuclear cells including foam cells, and fibrous tissue (Figure 23).

Figure 23 Cholesterol pericardial heart disease in a 34-year-old woman with hypothyroidism. She was short and stocky and had enlarged earlobes, bilateral hypoplastic fourth toes, receding hairline, small tongue, and no menstruation. She had orthopnea, pedal edema, (more ...)

Cholesterol pericardial heart disease may be seen in many different conditions (Table 11) but is most common in patients with hypothyroidism, rheumatoid arthritis, and tuberculosis (77).

Table 11 Etiology of cholesterol pericardial heart disease

The cause of the increased cholesterol content of the pericardial fluid in these patients is not clear, but several theories have been presented: 1) that necrosis of superficial cells of the pericardia liberates intracellular cholesterol, 2) that lysis of erythrocytes following hemopericardium yields the cholesterol, and 3) that pericardial inflammation decreases lymphatic drainage of pericardia, causing decreased reabsorption of cholesterol, with ultimate precipitation and crystallization (77). The cholesterol crystals then incite a vigorous cellular reaction, enhancing fluid production and pericardial thickening. Chronic exudation and diminished reabsorption of cholesterol might increase its concentration and allow crystallization. The total lipid content of the pericardial fluid, however, is approximately the same as that of serum.

In patients with hypothyroidism, the primary effusion may contain sufficient cholesterol to incite a local pericarditis, which then further increases the cholesterol concentration as less is reabsorbed. Large effusions with resulting high levels of cholesterol are produced. In these patients, the pericarditis appears to be a secondary phenomenon whereas cholesterol pericardial heart disease in euthyroid patients must be associated with an actual pericardial inflammation for the cholesterol to accumulate. Experimental studies have shown that small amounts of blood in the presence of pericardial damage from any cause will produce dense adhesions; this phenomenon probably accounts for cases of constriction in cholesterol pericardial disease.

Congenital anomalies include cysts and absence of all or portions of parietal pericardia. These cases are rare. Cysts can be true or false. The true ones are located within the pericardial sac but have no communication with it; the false ones, or diverticulae, are protrusions of parietal pericardium and consequently have direct communication with it. Of 72 cysts reported by Wychulis and associates (78), 88% were true and 12% were false. The cysts are thin-walled and translucent, are lined by endothelial cells, contain watery yellow fluid, usually are attached by a thick fibrous pedicle, and most commonly are located in the right cardiophrenic angle (Figure 24). In one series, the cysts ranged in size from 2 to 16 cm in largest diameter. Although they can be multilobulated, 99% are unilobular. The fluid within the cysts is similar to normal pericardial fluid. The cysts produce signs or symptoms of cardiac dysfunction in 25% of the patients and no evidence of dysfunction in the other 75%. The larger the size, the more likely symptoms will result. The mechanism of formation of pericardial cysts is not clear. Failure of the pericardial mesenchymal lacunae to unite with the pericardial coelom has been suggested (79).

Figure 24 Pericardial cyst, an incidental necropsy finding in a 75-year-old woman. The cyst, which contained serous fluid, overlay the right atrium and arose from a pedicle attached to the right main pulmonary artery (RPA). SVC indicates superior vena cava.

Absence of, or defects in, parietal pericardium may be classified into 3 major types (80): 1) heart and left lung in a common cavity (60%), 2) defect or foramen in parietal pericardium providing communication between pericardial sac and pleural sac (20%) (nearly always the left one) and 3) totally absent or rudimentary parietal pericardium (20%). These defects may be associated with anomalies of the heart (30%), lungs, pleura, peritoneum, or kidney and are more common in males. Cardiac enlargement occurs in about half of the patients, and commonly the heart is displaced to the left and abnormally mobile. Total absence of parietal pericardium usually produces no functional disturbance. Partial absence, however, may result in herniation of the entire heart or a part of the left atrium through the defect or allow direct extension of a lung infection to the pericardium via the defect.

Complete absence of the parietal left pericardium over the left side of the heart presents a clinical syndrome that is usually diagnostic. It consists of a physical finding of left ventricular enlargement and hypertrophy (apical impulse forceful and displaced to the left), a systolic ejection murmur at the base, an electrocardiogram that shows right axis deviation (as in the normal individual assuming the left lateral decubitus position), and a chest radiograph that shows levo position of the heart. This condition may be confused with a cardiomyopathy. Partial absence of the partial pericardium has been confused with a diagnosis of congenital heart disease with the prominent left atrial appendage misinterpreted as enlargement of the pulmonary trunk.

Constrictive pericardial heart disease—from pericardial effusion, thickened pericardia, or both—may be confused with a number of conditions (Table 12). An occasional patient with idiopathic dilated cardiomyopathy or extensive cardiac amyloidosis has been subjected in the past to thoracotomy because of suspected pericardial constriction. Echocardiography, of course, makes it easier to differentiate pericardial heart disease from these other conditions.

Table 12 Differential diagnosis of constrictive pericardial heart disease

An ideal classification of pericardial heart disease would take into account clinical, etiologic, and morphologic features of this condition, but a single classification combining these 3 components is lacking. Pericardial heart disease is relatively uncommon clinically. When present at necropsy, it usually had not been recognized during life. The term “pericarditis” is inaccurate because most pericardial diseases are noninflammatory. Morphologically, chronic pericardial heart disease symptoms are present; however, few patients develop evidence of cardiac dysfunction (constriction). When pericardial “constriction” occurs, it is the result of increased pericardial fluid or increased pericardial tissue or both. Increased fluid is treated by drainage; increased tissue is treated by excision. In most patients with chronic constrictive “pericarditis,” the etiology is not apparent even after histologic examination of pericardia.