Pericardial disease is most often manifested clinically as the pain syndrome caused by pericarditis or by the hemodynamic compromise caused by increased pericardial fluid (tamponade) or by a thickened or calcified pericardium (constrictive pericarditis). Other clinically important diseases of the pericardium are much less common (e.g., purulent or tubercular pericarditis) but also must always be considered, as discussed below.
Pericarditis commonly presents with a characteristic syndrome that includes central, pleuritic chest pain that is worse when lying down and less severe when standing upright or leaning forward. Frequently, there is no pericardial friction rub; however, the presence of a friction rub or an abnormal electrocardiogram or echocardiogram may suggest pericarditis in patients without symptoms of pericarditis who are being assessed for other conditions such as cancer, myocardial infarction, uremia, or a collagen vascular disease. The pain associated with pericarditis may be atypical and difficult to differentiate from pain caused by myocardial ischemia. The electrocardiogram is often normal or not diagnostic; the characteristic type of diffuse ST elevation (concave upward), not restricted to leads reflecting a region supplied by one coronary artery, when present, may be difficult to distinguish from early repolarization (normal). PR-segment depression is very suggestive of pericarditis. Repolarization changes may evolve over hours to days; in contrast to changes due to ischemia or infarction, the ST segment returns to baseline before the T wave inverts. No Q waves develop, and cardiac enzymes usually are normal but may increase in some patients.
In general, it is not difficult to discriminate between myocardial ischemia and pericarditis. However, occasionally, this can be very difficult and the two may coexist. The classic symptoms of pericarditis, unassociated with symptoms, signs, or electrocardiographic changes (Fig. 28-3) suggesting ischemia or infarction, can be confidently attributed to acute pericarditis. This may be confirmed by the presence of classic ST changes on the electrocardiogram (frequently absent) and/or a pericardial friction rub (also frequently absent). If the patient has known or suspected chronic or acute coronary artery disease, the disease may coexist with pericarditis so that the potential for pericarditis after infarction should be considered. Other characteristics may be helpful in discerning the cause of the pain, such as its duration, the response to anti-ischemic or anti-inflammatory medications, or the results of laboratory tests. Minor elevations in cardiac enzymes including troponin T or I may not be indicative unless serial changes indicate either diagnosis. It is not rare for the dilemma to be resolved in the cardiac catheterization laboratory. The association of myocarditis with pericardial pain may make the diagnosis even more difficult to resolve.
Twelve-lead electrocardiogram of a patient with acute pericarditis. Note PR segment depression prominent leads 2, 3, and aVF. ST-segment elevation is present globally (except in lead aVR) and is concave upward.
Benign idiopathic pericarditis is appropriately named except for occasional cardiac tamponade23 or refractory chest pain. Progression from clinically apparent idiopathic pericarditis to constrictive pericarditis is very uncommon.
Unless the clinical characteristics suggest a condition other than benign idiopathic pericarditis or there is a hemodynamically significant effusion, a conservative approach is generally appropriate and diagnostic pericardiocentesis is usually not indicated. Treatment is usually with nonsteroidal antiinflammatory drugs or colchicine; steroids should be reserved for only the most refractory symptoms, and pericardiectomy is only rarely indicated.23 If the cause of the pericarditis is known, specific treatment for that disease is indicated.
Small pericardial effusions are relatively common in patients with heart disease or in critically ill patients. The presence or absence of a pericardial friction rub correlates poorly with the volume of fluid. Tamponade is caused by accumulated fluid usually distributed throughout the pericardial space over the ventricles and the right atrium and, to a lesser extent, behind the left atrium. Localized collections of fluid, particularly early after open heart surgery, may compress individual chambers with similar consequences. The rate of fluid accumulation is very important; if slow, the pericardium stretches over time and large volumes (up to several liters) may collect before causing hemodynamic difficulty. However, if the fluid accumulates quickly, the noncompliant pericardium does not stretch and tamponade can be caused by as little as 150 to 200 mL. As pericardial pressure increases, transmural ventricular pressures decrease despite increased intracavitary pressures.
Cardiac tamponade develops when constraint is sufficient to substantially reduce cardiac filling and, hence, output. When this happens, diastolic pressures equilibrate in all chambers so that RA and left atrial pressures, RVEDP, LVEDP and pulmonary artery diastolic pressures become equal or nearly equal (<3- to 5-mm Hg difference).24 Commonly, cardiac tamponade is apparent at pericardial pressures of 15 to 20 mm Hg and with a jugular venous pressure usually greater than 10 cm of water above the sternal angle. It is characteristically associated with tachycardia, decreased cardiac output, and hypotension, although hypertension occasionally is present. The normal inspiratory decrease in arterial blood pressure (<10 mm Hg) is exaggerated (pulsus paradoxicus); this is not specific for tamponade and can occur in other circumstances, particularly when there is increased inspiratory effort. However, it is almost always present in cardiac tamponade (in the absence of aortic regurgitation, atrial septal defect, or severe LV dysfunction).24 The normal inspiratory increase in venous return persists in tamponade; the resultant increase in RVEDV causes a decrease in LVEDV and output because the pericardium cannot expand to accommodate the increased RVEDV. The decrease in LVEDV causes blood to accumulate in the lungs during inspiration. Persistent constraint to filling throughout diastole is responsible for the decreased y descent (early diastole) in the jugular venous pulse and RA pressure tracing and the slowly rising RV early diastolic pressure (Figs. 28-4 and 28-5). The Kussmaul sign and the square root (dip-plateau) sign in the jugular venous pulse are not seen in tamponade.
Right atrial pressure in a patient with cardiac tamponade during mechanical ventilation. Note the absent y descent and the blunting of the waveforms, especially during positive-pressure ventilation.
Right ventricular pressure in a patient with cardiac tamponade. Note the slow rise in ventricular pressure in early diastole, corresponding to decreased passive filling.
Once considered, the diagnosis of cardiac tamponade is generally straightforward. In a critically ill patient, a pericardial infusion sufficiently large to cause hemodynamic problems may be discovered fortuitously when echocardiography is performed for various reasons. Increased right-sided filling pressures or pulsus paradoxicus should initiate a systematic hemodynamic assessment that will ultimately lead to the correct diagnosis. Physical examination may show an obvious cause for the increased right-sided filling pressure. However, causes of increased filling pressures are often best evaluated initially by transthoracic or transesophageal echocardiography. Either modality helps to detect valvular or myocardial disease and pericardial causes of increased RA pressure. The assessment may be complicated by the presence of other conditions. One may also be faced with the management of tamponade and the need to deal with the specific cause of the effusion, for example, tuberculosis or neoplasm.
The characteristic hemodynamic features of tamponade (equilibration of diastolic pressures) may be readily apparent in hemodynamically monitored patients. Electrocardiographic features of tamponade include low voltage and electrical alternans, but these findings need not be present. Electrical alternans suggests the presence of tamponade (Fig. 28-6); it is generally associated with swinging of the heart within the pericardium. Hemodynamically significant effusions are characterized echocardiographically25–30 by the presence of moderate or large effusions (unless accumulated rapidly), early diastolic RV collapse, right atrial collapse (Fig. 28-7), the presence of fluid behind the left atrium, a dilated inferior vena cava, and increased respiratory variation in tricuspid and mitral Doppler flow velocities. However, the diagnosis of tamponade remains a clinical diagnosis.31 Differentiation from constrictive pericarditis is not an issue, although hemodynamic comparisons are interesting (see below). The chest x-ray can be useful in identifying patients with cardiac tamponade, especially when associated with a quiet precordium (Fig. 28-8). There may be a characteristic “water-bottle” appearance, with rapid tapering of the bulging pericardial silhouette. Pulmonary edema is almost never seen in uncomplicated tamponade. Decreased QRS voltage associated with an enlarged cardiac silhouette on radiography or electrical alternans may suggest the presence of a large effusion. Other radiographic signs are not relevant because of the greater sensitivities of other imaging modalities. Computed tomography and magnetic resonance imaging are useful to diagnose effusions and may indicate the nature of the effusion.32
Electrocardiogram of a patient with cardiac tamponade demonstrates low voltages in the limb leads and marked electrical alternans.
Left. An early diastolic frame of a two-dimensional echocardiogram in a patient with cardiac tamponade. There is a very large pericardial effusion and right atrial and right ventricular collapse. Right. A late diastolic frame from the same cardiac cycle; both ventricles are underfilled.
Chest radiogram (anterior) of a patient with rapidly increasing pericardial effusion and hemodynamic evidence of cardiac tamponade. Note how the cardiac silhouette is rounded in its lower portion and tapers at the base of the heart, resembling a plastic bag filled with water sitting on a table.
The contribution of a small effusion to a patient's condition may be difficult to ascertain and is based on the principles described above. Even a small effusion may be significant if it accumulated rapidly; examples include trauma, recent invasive cardiac procedures, and other conditions that may cause bleeding into the pericardium, such as aortic dissection.
In the absence of tamponade, even large effusions may not need to be drained.33 If the effusion is believed to be benign, a careful conservative approach may be sufficient. The fundamental problem in cardiac tamponade is constrained ventricular filling, so removal of the fluid is central to effective treatment. Unless the patient's volume has been depleted, there is no potential benefit from volume loading because filling cannot be increased. There is also limited potential benefit from inotropic agents because there is already substantial neurohumoral stimulation in these patients. The method of pericardial drainage is not standardized.34–38 Unless the etiology of the effusion indicates otherwise, needle drainage should be performed with echocardiographic guidance. Large effusions are usually safely drained, whereas small effusions are inherently associated with greater risk. Electrocardiographic monitoring with a lead connected to the needle (using ST-segment elevation on the electrocardiogram to indicate contact with the myocardium) is not as safe as echocardiographic monitoring. Drainage of the first 50 to 100 mL of fluid usually relieves much of the hemodynamic compromise. A catheter is usually left in place under suction to drain the pericardial space for at least 2 to 3 days. Clotting of the catheter may be prevented with a slow infusion of heparinized saline.39 The drained fluid should be sent for appropriate diagnostic assessment even if the underlining etiology is known or suspected, despite a low diagnostic yield.40 The presence of blood in the effusion is not helpful in establishing the cause of the effusion.41 Recurrent effusions are not uncommon, particularly if the original effusion was large. Other different techniques are used initially or subsequently. These include balloon pericardiotomy,42 thoracoscopic drainage (which improves diagnostic accuracy),43,44 and a pericardial window.35 Surgical drainage should be considered initially in patients with smaller effusions because of the higher risk when drained by needle. Long-term survival rate is dependent on the underlying disease rather than on the method of drainage.34,35,45,46 Large effusions due to metastatic disease may be best treated surgically, whereas benign effusions may be adequately treated with a single-needle drainage procedure. Chemotherapeutic or sclerosing agents or steroids may be installed in the pericardium after drainage.47
Whereas substantial pericardial effusions have been associated with anticoagulation, the presence of a friction rub or a very small amount of pericardial fluid detected by echocardiography should not deter the clinician from continuing anticoagulation when there is a clear indication for it. Examples include myocardial infarction and the presence of a prosthetic heart valve. Patients can be safely monitored clinically and echocardiographically for the unlikely development of tamponade. Anticoagulants should be discontinued in Dressler syndrome (which is increasingly rare after myocardial infarction).
Up to 40% of patients with chronic renal failure develop pericardial effusions during the course of their disease. Effusions may occur before dialysis and during maintenance dialysis.48 There is some question as to whether heparin contributes to the development of effusion, but patients on dialysis using heparin-free solutions still develop effusions. Because preload is reduced during hemodialysis, the risk of rendering an asymptomatic effusion hemodynamically significant is real. The probability that dialysis will cause an effusion to clear is highest when the effusion predates the start of intensive hemodialysis or develops early in its course (within weeks).49 The size of the effusion is the single most important determinant of need for surgical therapy. Hemodialysis is more effective than peritoneal dialysis for reducing effusions and should be initiated early if there is evidence of enlargement. Indomethacin and steroids are of uncertain value. Failure to resolve or progression of an effusion during the course of intense hemodialysis may dictate the need for drainage.
Pericardial abnormalities are common after cardiac surgery. Early after surgery, hemopericardium and hemomediastinum can cause tamponade, even though the pericardium is commonly left open. When there is postoperative hemodynamic compromise, it is necessary to consider the possibility of tamponade. Although echocardiography is often very helpful in this regard, tamponade is sometimes caused by a localized collection of fluid or blood that may be difficult to identify, even by transesophageal echocardiography. A careful synthesis of hemodynamic and echo data may require reoperation without a clear diagnosis beforehand. The presence of unexplained hemodynamic deterioration (e.g., with good ventricular function) suggests the possibility of localized tamponade, even if it is difficult to demonstrate echocardiographically.
The postpericardiotomy syndrome occurs several weeks postoperatively in 10% to 20% of patients and is characterized by fever, chest pain, and the presence of a friction rub.50 In many ways, it is analogous to the Dressler syndrome, which occurs weeks after myocardial infarction. Good evidence supports an autoimmune pathogenesis, with antiheart antibodies and circulating immune complexes present in many patients.51 Electrocardiographic changes identical to those seen in acute pericarditis may or may not be present. Aspirin is the treatment of choice. Corticosteroids may be considered in patients who do not respond to aspirin or colchicine.
Pericardial effusions are much more likely to be sterile than infected. Purulent pericarditis is most often the result of local spread from empyema, mediastinitis, endocarditis, prior pericardiotomy, or burn injury.52 It is rare, tends to occur in immunocompromised patients, and may be difficult to treat. Although drainage is necessary, it is not clear what the best approach is. It is commonly first discovered when pericardiocentesis is performed by needle; continuous catheter drainage should be initiated then, but surgery likely will be necessary. Patients with the human immunodeficiency virus may present with tamponade, and an infectious cause is frequent; this is associated with a very high mortality rate.53
Purulent pericarditis may develop in the setting of severe infectious disease with a wide variety of causative organisms. When pericardial fluid is evident on echocardiography in a patient who is at risk for purulent pericarditis, a definitive diagnostic procedure should be considered because its mortality rate is high. The best approach is removal of the fluid. Some centers perform a subxiphoid procedure and proceed to pericardiectomy, if infection is present. Thoracoscopic pericardiectomy also may be useful in such cases.
It is clearly important to discriminate between purulent and sterile effusions. Electrocardiography and echocardiography are not useful. Gallium 67 (67Ga) uptake, when used with single-photon emission computed tomography, has been reported to have a resolution that can distinguish pericardial from myocardial inflammation and that may be able to detect inflammation associated with secondary infection of the pericardial space. There are several reports of tuberculous pericarditis diagnosed by planar 67Ga imaging. In patients with complex disease who may have multiple areas of 67Ga uptake in the region, simultaneous use of technetium 99 and 67Ga with subtraction imaging may increase the specificity of the test.
In constrictive pericarditis, the pericardium becomes thickened, may be calcified, and prevents normal ventricular filling. Signs of constriction may become evident as effusive pericarditis resolves, and normal hemodynamics may return spontaneously or with conservative treatment.54,55 Forward output is compromised and the increased diastolic pressures cause edema. Early diastolic ventricular filling is rapid in constriction and is responsible for the classic dip-plateau configuration of atrial and ventricular pressure tracings; there is a rapid y descent with an early nadir and a sharp rise in pressure as the chamber reaches its limits of expansion (Fig. 28-9). After atrial systole, there is a rapid x descent so that an M-shaped jugular venous pulsation may be observed (Fig. 28-10). The fixed maximum volume of the pericardium results in equalization of end-diastolic pressures, similar to that in cardiac tamponade. If the central volume has been decreased, as might result from excessive diuresis, equalization of pressures may not be observed; volume expansion, however, will cause this to occur.
Simultaneous right ventricular and left ventricular pressures in a patient with constrictive pericarditis. The pressures are elevated and nearly equal, with little rise in right ventricular pressure during the last two-thirds of diastole. LV, left ventricle; RV, right ventricle.
Right atrial pressure in a patient with constrictive pericarditis. Note the prominent x and y descents and the M shape. The a and v waves and the x and y descents are as described in text.
The Kussmaul sign (inspiratory increase in the jugular venous pressure) is common56 but not specific for constriction; it is seen most often in RV infarction and other conditions affecting the right ventricle. There are several echocardiographic findings associated with constrictive pericarditis (Fig. 28-11), none of which is pathognomonic.29 The pericardium may appear thickened, but this usually cannot be discerned with confidence; transesophageal studies may be more accurate in defining pericardial thickness. Because rapid ventricular filling occurs early and stops abruptly, M-mode evidence of absent LV filling in the last half of diastole is common (flat posterior wall tracing); continued expansion of the left ventricle throughout diastole suggests that constriction is absent. During early rapid filling, there is a “septal bounce,” and there may be premature opening of the pulmonary valve (Fig. 28-12). During inspiration, there is substantial leftward septal shift in keeping with the unchanged or increased RVEDP and decreased LVEDP. Doppler echocardiography adds diagnostic accuracy but is not widely accepted as definitive.57,58 Chest radiography is often helpful; the absence of signs of left heart failure in the presence of right heart failure helps to narrow the focus. The cardiac silhouette can be small or large, depending on antecedent heart disease. Egg shell–like calcification may be apparent, usually more easily seen in the lateral projection; the absence or presence of calcification is not useful to rule in or out constriction.59 Nonspecific ST-T-wave abnormalities are often present on the electrocardiogram, but this is of little value.
A. Simultaneous left and right ventricular pressures superimposed on an M-mode echocardiogram in a patient with constrictive pericarditis. At end inspiration, the interventricular septum shifts leftward, thereby enlarging the right ventricle and shrinking the left ventricle. Associated with this shift are increased right ventricular systolic pressure and decreased left ventricular systolic pressure. B. Simultaneous left and right ventricular pressures superimposed on the left ventricular inflow Doppler signal. Left ventricular inflow velocity decreases during inspiration, corresponding to the decrease in left ventricular filling noted in the top panel. This change is associated with a decrease in generated left ventricular pressure and an increase in right ventricular pressure. C. Aortic pressure superimposed on an M-mode echocardiogram through the aortic valve. At end inspiration, opening of the aortic valve is markedly reduced, as is aortic pressure. Ao, aorta; EXP, expiration; INSP, inspiration; IVS, interventricular septum; LV, left ventricle; PW, posterior wall; RV, right ventricle.
M-mode echocardiogram in a patient with constrictive pericarditis. Note the flattened posterior wall of the left ventricle during diastole, consistent with little ventricular filling. The early rush of blood into the ventricle results in a shuttering motion of the interventricular septum and creates a notch (arrow). ECG, electrocardiogram; IVS, interventricular septum; LV, left ventricle; LVPW, left ventricular posterior wall; RV, right ventricle.
Pericardial thickening is detected readily and with high sensitivity by computed tomography and magnetic resonance imaging;32 however, in a large series, it was not present in almost one in five patients with constriction.60 Because fast computed tomography and magnetic resonance imaging are capable of visualizing dynamic chamber processes and patterns of filling, they can also provide physiologic information.
The characteristic presentation of constrictive pericarditis is that of right heart failure, often severe, without another apparent cause. The diagnosis may be difficult to prove in the presence of other forms of heart disease; this is increasingly problematic because constriction after cardiac surgery is not rare. Hepatosplenomegaly sometimes incorrectly suggests the presence of metastases. The jugular venous pulse may suggest constriction; there may be a Kussmaul sign and the characteristic M-shaped pulsation highlighted by the rapid y descent. A pericardial knock (an early diastolic, sometimes loud, sound) may be heard.
In some instances, it is difficult to distinguish constrictive pericardial disease from restrictive cardiomyopathy.28,29,58 Cardiac catheterization is required for simultaneous measurement of right and left filling pressures; when low, these pressures may differ but will converge if fluid is administered (<5-mm Hg difference). In restrictive cardiomyopathy, diastolic pressures are generally different and pulmonary artery pressure is often higher than in constriction. Volume loading is suggested if the diagnosis is not evident from the initial hemodynamics.61 Similar measurements can be obtained with a flow-directed pulmonary artery catheter. Doppler echocardiography also may be helpful but remains sufficiently unreliable to preclude it as the only diagnostic test.62 Myocardial biopsy can be helpful.
Constrictive pericarditis may be a complication of cardiac surgery, occurring in fewer than 1% of patients.63 The time to onset can be as short as 6 weeks or as long as years after surgery. In one series, almost 65% of patients with postoperative constriction were thought to have developed the postpericardiotomy syndrome earlier in their course.64 Pericardiectomy is made more difficult by the presence of bypass grafts.
In constrictive pericarditis, pericardiectomy is the only effective treatment. This is often very difficult and sometimes results are poor. As a result, some patients will have persistently increased venous pressure, and others, after initial improvement, may have recurrent constriction.65
Effusive-constrictive pericarditis is a combination of a thickened, potentially constricting pericardium with a variable amount of pericardial fluid. It occurs most often in patients with malignancies.66 The hemodynamic characteristics are similar to those of tamponade and may evolve into those due to constriction as the volume of fluid is decreased.