|
|
|
Pericardiotomy under direct vision was first done in 1815, and in 1840, the first blind approach using a trocar was carried out successfully on a patient with tamponade from malignancy. (1) By the end of the 19th century, the trocar-and-cannula method of pericardiocentesis was commonly used. The subxiphoid approach was first described in 1911.
Blind, electrocardiography (ECG)-assisted pericardiocentesis has a significant morbidity rate, reportedly as high as 15 to 20%. (2) (3) For this reason ultrasound diagnosis of pericardial effusion with fluoroscopic or ultrasound guidance has become the norm for elective pericardiocentesis because of its lower (0.5 to 3.7%) (4) (5) incidence of complications. The ECG-assisted blind pericardiocentesis technique remains the standard procedure for truly emergent pericardiocentesis when a lengthy delay may be associated with obtaining and organizing ultrasound or fluoroscopic assistance.
The medical literature concerning pericardiocentesis tends to address two distinct categories of pericardial fluid collection: acute hemopericardium (largely secondary to trauma) and pericardial effusion from other causes. This separation is not entirely artificial, as these two clinical entities are quite different in their time course, etiology, and treatment.
Acute hemopericardium has several causes, including coagulopathies, cardiovascular catastrophes, and acute injury resulting from either blunt or penetrating trauma. All of these causes result in rapid accumulation of whole blood in the pericardial sac. The fluid accumulates too fast for the relatively inelastic pericardial sac to stretch and accommodate the fluid. The result is cardiac tamponade produced by small fluid volumes and with an essentially normal pericardial size.
Traumatic tamponade due to penetrating trauma may result from obvious external injury such as knife or gunshot wounds, or it may be insidious, as seen with iatrogenic cardiac perforation during cardiac or vascular procedures.
In external penetrating trauma, tamponade is most commonly the result of a stab wound to the heart, (6) presumably because the pericardium seals itself after a stab wound but cannot do so after gunshot wounds. (6) (7) Approximately 80 to 90% of stab wounds to the heart demonstrate tamponade, (6) (8) compared with 20% of gunshot wounds. Larger pericardial wounds from gunshots generally drain into the pleural space and produce a hemothorax. (9) Cardiac tamponade is often suspected with anterior chest wounds, but it is imperative to remember that any penetrating wound of the lateral chest, back, or upper abdomen also may involve the heart.
Iatrogenic causes of cardiac tamponade are relatively uncommon but well-known complications of invasive or diagnostic procedures. Pacemaker insertion (either transthoracic or transvenous) and cardiac catheterization, including valvuloplasty and angioplasty, are two of the main causes, resulting in tamponade when cardiac chambers or coronary vessels are inadvertently penetrated. (10) (11) (12) Such penetration of vascular structures is common during transthoracic pacemaker placement. (13) Tamponade is also seen as a complication after cardiac surgery, although it is usually anticipated, and
Cardiac tamponade may result from perforation of the right atrium or, less commonly, of the right ventricle or superior vena cava by a central venous pressure (CVP) catheter or subclavian hemodialysis catheter. (17) This event is usually not diagnosed early and is therefore often fatal. (18) Perforation may occur during placement or, more commonly, 1 to 2 days later, when the catheter erodes through tissue, particularly if a catheter made of stiff material is used or when the left internal jugular vein approach is used. (19) Tamponade from CVP line placement is seldom seen in the emergency department but must always be considered when there is sudden decompensation in a patient with a CVP line in place. Tamponade should always be considered when a patient deteriorates hemodynamically after an invasive diagnostic or therapeutic procedure involving the heart.
Blunt trauma may cause hemopericardium. Usually it occurs in the presence of a major chest injury with associated bruises or rib and sternal fractures. Cases have been reported, however, in which tamponade occurred in blunt trauma with no obvious signs of injury to the thorax. (20) Such incidents may be more common than is clinically recognized, judging by the reports of constrictive pericarditis and pericardial defects found months to years later in trauma patients who were not originally noted to have effusion. Pericardial effusion due to blunt trauma may also be a late finding, becoming symptomatic 12 to 15 days after trauma. (21)
Severe deceleration injury may cause tamponade as a result of aortic dissection. This appears to be an uncommon development, with two case series reporting tamponade in 3.6% (1 of 28 patients) and 2.3% (1 of 43 patients) of victims of aortic injury. (22)
Theoretically, cardiopulmonary resuscitation (CPR) can cause pericardial effusion secondary to the blunt trauma of chest compressions, broken ribs, or intracardiac injections. Early studies reported pericardial effusion in 1 to 3% of CPR survivors. (23) Echocardiographic studies showed small cardiac effusions (but not tamponade) in 12% of survivors, only 4% of whom had received intracardiac injections. (24) Thus, although case reports of tamponade exist, (25) (26) it would appear that CPR and intracardiac drug injections are unlikely to cause significant effusion, much less tamponade.
Nontraumatic but acute hemopericardium caused by a bleeding diathesis, aortic dissection and ventricular rupture behaves much like traumatic tamponade because of its acute nature. These types of hemopericardium are less obvious in etiology than hemopericardium caused by external trauma.
Bleeding diathesis may cause spontaneous bleeding into the pericardial sac. The incidence of spontaneous pericardial tamponade in anticoagulated patients has been reported to range from 2.5 to 11%. (10) (27) Thrombolytic therapy has also been implicated in tamponade secondary to bleeding diathesis. Among 392 patients, only 4 (1%), all with large anterior myocardial infarctions, developed tamponade secondary to hemopericardium without ventricular rupture. (28)
A dissection of the ascending aorta may extend around the base of the aorta into the pericardial sac, causing dramatic, rapid, and often fatal tamponade. This pathologic abnormality may be due to conditions such as syphilis, Marfan syndrome, or atherosclerosis. Infection may create pseudoaneurysms of the aorta, which also can present as tamponade. (29)
Ventricular rupture after myocardial infarction is a common source of acute hemopericardium. Although the prognosis is often grim, survival is possible with prompt recognition and definitive treatment. (30) (31)
The etiology of nonhemorrhagic pericardial effusions determines the aggressiveness, type, and speed of treatment required. Patients with nonhemorrhagic etiologies usually accumulate effusions slowly, which allows the pericardium to stretch and accommodate up to 2000 mL of fluid. (32) This slower accumulation, often over weeks to months, means that even in a moderately hypotensive patient, more time may be available for evaluation and treatment. (33) (34) In many cases of small nonhemorrhagic effusion, tamponade does not occur, and the effusion may resolve with treatment of the underlying disease or may be managed successfully by elective pericardiocentesis.
Many disease processes, ranging from the common to the rare (Table 15-1) , can cause pericardial effusion. The cause of nonhemorrhagic tamponade may not be obvious on examination in the emergency department, and tamponade is frequently misdiagnosed as congestive heart failure or respiratory disease. Although neoplasm has generally been the most common underlying cause of nonhemorrhagic effusion, (35) (27) some reports (36) (37) have identified infectious complications of the human immunodeficiency virus (HIV) as a common etiology of large nonhemorrhagic pericardial effusion and tamponade (Table 15-2) .
HIV-related effusions have been ascribed to many opportunistic bacterial and viral infections, with mycobacterial infections being the most common. (36) (37) (38) Noninfectious pericardial effusions in HIV have been caused by Kaposi sarcoma and lymphoma. (39) (40)
Cancer is a prominent cause of nonhemorrhagic effusions; the pericardium is involved in 20% of patients with disseminated tumors (41) and 8% of all patients with cancer. (42) There is primary pericardial involvement in 69% of acute leukemias, in 64% of malignant melanomas, and in 24% of lymphomas; however, the incidence of actual tamponade in these malignancies is not known. Of metastases to the pericardium, 35% originate in the lung, 35% in the breast, 15% in lymphomas, and <3% in each of the other cancers. (42) Thus, any patient who is known to have one of these malignancies should be considered at risk for tamponade. Metastasis to the heart is usually a late finding in cancer, and other foci located elsewhere are usually evident. (43) Classic findings of tamponade, such as pulsus paradoxus, are frequently absent in cancer patients with tamponade, and their symptoms are usually attributed to their malignancy. (42)
Radiation pericarditis, particularly after treatment for Hodgkin s disease, is a common cause of effusion. (32) Effusion occurs in approximately 5% of those patients who receive 4000 rad to the heart.
Approximately 15 to 20% of patients on dialysis for renal failure develop pericarditis, and 35% of those with pericarditis develop tamponade. (44) (45) Up to 7% of patients on
Thirty percent of myxedema patients may have pericardial effusions,
but few have tamponade. (35)
Most of the other
| Neoplasm |
Mesothelioma Lung Breast Melanoma Lymphoma |
| Pericarditis |
Radiation (especially after Hodgkin s disease) Viral Bacterial Staphylococcus Peumococcus Haemophilus Fungal Tuberculosis Amebiasis Toxoplasmosis Idiopathic |
|
Connective tissue disease |
Systemic lupus erythematosus Scleroderma Rheumatoid arthritis Acute rheumatic fever |
|
Metabolic disorders |
Myxedema Uremia Cholesterol pericarditis Bleeding diatheses |
| Cardiac disease |
Acute myocardial infarction Dissecting aortic aneurysm Congestive heart failure Coronary aneurysm |
| Drugs |
Hydralazine Phenytoin Anticoagulants Procainamide Minoxidil |
| Trauma |
Blunt Major trauma Closed-chest CPR Penetrating Major penetrating trauma Intracardiac injections Transthoracic and transvenous pacing wires Pericardiocentesis Cardiac catheterization CVP catheter |
| Miscellaneous |
Serum sickness Chylous effusion Loffler syndrome Reiter syndrome Behcet syndrome Pancreatitis Postpericardiotomy Amyloidosis Ascites |
|
|
Krikorian(27)
(120 Patients) (%) |
Guberman(35)
(56 Patients) (%) |
|---|---|---|
| Neoplastic disease | -- | 32 |
| Pericardial invasion | 16 | -- |
| Radiation pericarditis | 7.5 | 4 |
| Etiology uncertain | 18 | -- |
| Traumatic hemopericardium | 9 | -- |
| Hemopericardium, nontraumatic | 2.5 | -- |
| Rheumatic disease | 12 | 2 |
| Uremia/dialysis | 5 | 9 |
| Bacterial infection | 2.5 | 12.5 |
| Congestive heart failure | 1.5 | -- |
| Uncertain etiology | 12.5 | -- |
| Idiopathic pericarditis | 13.5 | 14 |
| Cardiac infarction | -- | -- |
| Iatrogenic diagnostic procedures | -- | 7.5 |
| Myxedema | -- | 4 |
| Aneurysm | -- | 4 |
| Anticoagulation and cardiac disease | -- | 11 |
| Postpericardiotomy | -- | 2 |
An interesting but rare cause of cardiac tamponade is pneumopericardium. Pneumopericardium is most commonly seen with pneumothorax and pneumomediastinum as a complication of respiratory therapy in infants, but it may also occur from similar barotrauma in adults. (46) Pneumopericardium also occurs spontaneously in asthma, (47) after blunt chest injury, (48) and even after high-speed motorcycle rides. (49) Pneumopericardium is usually benign, but tension pneumopericardium has been reported as a cause of life-threatening tamponade after blunt chest trauma (50) and after pericardiectomy. (51) The appearance of life-threatening pneumopericardium and tamponade has also been described immediately (52) and 6 days after penetrating chest trauma. (53)
The pericardium is a tough, leathery sac that encloses the heart. It normally contains about 25 to 35 mL of serous fluid. (54) The pericardium is not rapidly elastic, although it does demonstrate stress relaxation within minutes of increased intrapericardial pressure, providing a slight ability to accommodate sudden increases in fluid. (55) As fluid accumulates, the first 80 to 120 mL is easily accommodated without significantly affecting pericardial pressure (Fig. 15-1) (Figure Not Available) . (56) However, if an additional 20 to 40 mL is rapidly accumulated, the intrapericardial pressure almost doubles, often leading to sudden decompensation. With chronic effusions
Pericardial compliance varies considerably in different individuals and various disease states. This compliance helps to determine the pressure-volume response curve (Fig. 15-2) (Figure Not Available) . (54) The pressure-volume relationship demonstrates hysteresis; the withdrawal of a quantity of fluid drops the pressure more than the addition of the fluid raised it.
As pericardial fluid accumulates, the increased intrapericardial pressure is transmitted across the myocardial wall and causes compression of the atria and perhaps the vena cava and pulmonary veins. This reduces right ventricular filling in diastole, producing decreased stroke volume and cardiac output. (57) Pulse pressure narrows as reflex sympathetic stimulation increases. Severe tamponade is produced with intrapericardial pressures of 15 to 20 mm Hg. (58)
Because stroke volume is decreased, heart rate increases to maintain cardiac output. Sympathetic discharge causes both arterial and venous vasoconstriction. (58) (59) Vasoconstriction increases venous pressure, which helps to restore the normal venous-atrial and atrioventricular filling gradients. These compensatory mechanisms are often effective and may permit establishment of a new homeostasis with normal cardiac output.
In chronic effusion and in early tamponade, cardiac contractility is not affected, and myocardial perfusion is normal. (57) (60) (61) As pressure continues to increase, coronary perfusion pressure drops, so in its later stages, tamponade is an ischemic event for the heart. By the time hypotension is measurable, left ventricular blood flow has already decreased 37%. (62) For comparable degrees of hypotension, experimental animals in hemorrhagic shock have five times greater coronary blood flow than animals in cardiac tamponade. (62) Severe
Figure 15-1 (Figure Not Available) Production of cardiac tamponade by injections of saline into the pericardial sac. Although pericardial space can acutely accommodate 80 to 120 mL of fluid without a significant increase in pericardial pressure, note steep increases in pressure and drop in blood pressure at about 200 mL of saline. Once critical volumes are reached, very small increases cause significant hemodynamic compromise. (From Fowler NO: Physiology of
cardiac tamponade and pulsus paradoxus. II: Physiological, circulatory, and
pharmacological responses in cardiac tamponade. Mod Concepts Cardiovasc Dis
47:116, 1978. Reproduced by permission of the American Heart Association,
Inc.)
Figure 15-2 (Figure Not Available) Relationship of intrapericardial pressure to volume of pericardial fluid. Note that pressure drops more rapidly when fluid is removed than when it accumulates. (From Pories W, Gaudiani V: Cardiac tamponade. Surg Clin North Am 55:573, 1975. Reproduced by permission.)
As intrapericardial pressure continues to rise, the heart s compensatory mechanisms fail. Myocardial ischemia and perhaps lactic acidosis from poor tissue perfusion may be the triggering events that disrupt the uneasy equilibrium. (64) Atrial pressure rises rapidly (Fig. 15-3) (Figure Not Available) . The atria and pulmonary circulation, being at much lower pressure than the systemic arterial pressure, are more vulnerable to the rising intrapericardial pressure. A ""pressure plateau"" occurs in which right atrial pressure, right ventricular diastolic pressure, pulmonary artery diastolic pressure, and pulmonary capillary wedge pressure are virtually identical.
This equalization of pressures leads to the echocardiographic hallmark of tamponade: right ventricular collapse. At this point hypotension is severe, bradycardia is common, and pulseless electrical activity (PEA) may occur. Unless intrapericardial pressure is immediately decreased, pulmonary blood flow ceases, and cardiac arrest follows. (64)
Total blood volume affects cardiac compensation, and it is possible to encounter a ""low-pressure"" cardiac tamponade. (65) The hypovolemic patient with tamponade has a decreased venous pressure, which not only decreases cardiac output, but also may obscure the diagnosis, because distended neck veins or an elevated CVP are not present. In a patient with a chronic pericardial effusion, the onset of hypovolemia can lower filling pressure enough to precipitate tamponade, and conversely, providing additional volume may temporarily offset increased pericardial pressure.
Ventilation and blood CO2 levels have significant effects on cardiac tamponade. This is of particular significance, because trauma and other patients with tamponade may also have respiratory impairment. Pericardial pressure decreases 3 to 6 mm Hg with a hypocarbia of 24 torr and increases 2 to 4 mm Hg when the PCO2 reaches 57 torr. (66) This degree of hypercarbia-induced pericardial pressure rise can decrease
Figure 15-3 (Figure Not Available) Summary of physiologic changes in tamponade. RV, right ventricle. (From Shoemaker WC, Carey JS, Yao ST, et al: Hemodynamic monitoring for physiological evaluation, diagnosis, and therapy of acute hemopericardial tamponade from penetrating wounds. J Trauma 13:36, 1973; and Spodick D: Acute cardiac tamponade: Pathologic physiology, diagnosis, and management. Prog Cardiovasc Dis 10:65, 1967. Reproduced by permission.)
The clinical diagnosis of pericardial effusion can be difficult. In contrast, pericardial tamponade is a clinical diagnosis, but specific clinical signs are often inaccurate. Particularly in the setting of acute hemorrhagic tamponade, the time from the first signs of tamponade to full arrest may be brief, and a high level of suspicion to invoke ultrasound or invasive diagnostic testing must be maintained. (68)
Classic clinical findings have been described for tamponade. However, these findings are often clearly present only when the patient is in fulminant tamponade. Ideally, tamponade is diagnosed early, when the patient suffers no more than dyspnea, weakness, or sometimes right heart failure. A common pitfall is to attribute respiratory symptoms (e.g., dyspnea on exertion) to a more common condition such as heart failure or pulmonary pathology and to overlook pericardial effusion until the classic late signs (e.g., hypotension) appear. (69)
Acute pericardial tamponade may resemble tension pneumothorax, acute hemothorax, hypovolemia, pulmonary edema, or pulmonary embolism. Severe right ventricular contusion can mimic the findings of tamponade. (70) The patient is often agitated or panic stricken, confused, uncooperative, restless, cyanotic, diaphoretic, and acutely short of breath. In the late stages, the patient is moribund. Hypotension in the presence of severe cyanosis and distended neck veins is a helpful but late finding.
The classic physical findings of tamponade were first characterized by Beck in 1935. He described two triads, one for acute and one for chronic compression. (71) The chronic compression triad consists of high CVP; ascites; and a small, quiet heart. The triad in acute compression consists of high CVP, decreased arterial pressure, and muffled heart sounds. Unfortunately, in most major trauma series, only about one third of patients demonstrate the complete acute triad, (64) (72) although almost 90% have one or more signs. (6) The simultaneous occurrence of all three physical signs is a very late manifestation of tamponade and is usually seen only shortly before cardiac arrest (see Fig. 15-3) (Figure Not Available) .
Careful hemodynamic monitoring reveals much earlier changes that indicate the progression of tamponade (Table 15-3) (Table Not Available) . (73) In grade I tamponade, cardiac output and arterial pressure are normal, but CVP and heart rate are increased. In grade II tamponade, blood pressure is normal or slightly decreased, CVP is increased, and tachycardia persists. In grade III tamponade, the classic findings of Beck s acute triad occur. Although this sequence represents the natural history of acute tamponade, the time course varies. Some patients are stable at a given stage for hours; others proceed to cardiac arrest within minutes. (64) (73) Unfortunately, not all patients with early tamponade respond with a predictable pattern of change in vital signs. Brown and coworkers found that 6 of 18 patients with tamponade, defined through right heart catheterization, responded to tamponade with elevated systolic blood pressure. (74) After pericardiocentesis, these patients had a marked reduction in systolic blood pressure accompanied by increased cardiac output. All of these patients had previously been hypertensive.
The measurement of pulsus paradoxus is a procedure in its own right (see Chapter 70) . Pulsus paradoxus is defined as an exaggeration of the normal inspiratory fall in blood pressure. (59) (72) A paradoxical pulse (pressure) is one of the classic physical signs of tamponade, but it is not pathognomonic. It is also caused by pulmonary emphysema, asthma, labored respirations, obesity, cardiac failure, constrictive pericarditis, pulmonary embolism, and cardiogenic shock. (6) (33) (64) Measuring the paradoxical pulse is difficult and time consuming, and any frightened, hypotensive patient with labored breathing can demonstrate this finding (Fig. 15-4) .
If the difference between inspiratory and expiratory systolic blood pressures is >12 mm Hg, the paradoxical pulse is abnormally high. (75) Most patients with proven tamponade will demonstrate a difference of 20 to 30 mm Hg or more during the respiratory cycle. (6) (33) (64) This may not be true of patients with very narrow pulse pressures (typical of grade III tamponade); they will have a ""deceptively small"" paradoxical pulse of 5 to 15 mm Hg. The decrease in magnitude of pulsus paradoxus with hypotension occurs because the paradoxical pulse is a function of actual pulse pressure, and the inspiratory systolic pressure may be below the level at which
Pulsus paradoxus has been correlated with the amount of impairment of cardiac output by tamponade. In atraumatic patients, a 15% pulsus paradoxus in the face of relative hypotension was found in 97% of patients with moderate or severe tamponade and only 6% of patients with absent or mild tamponade. (75) A similar study of right ventricular diastolic collapse by echocardiography found that an abnormal pulsus paradoxus had a sensitivity of 79%, a specificity of 40%, a positive predictive value of 81%, and a negative predictive value of 40%. (76)
The absence of a paradoxical pulse does not rule out tamponade. Although the mean paradoxical pulse was 49 mm Hg in one series of nonhemorrhagic tamponade, (35) 23% of the patients had a paradoxical pulse of <20 mm Hg, and 1 patient had no measurable paradoxical pulse. An abnormal pulsus paradoxus has been reported to be absent in tamponade when there is an atrial septal defect, aortic insufficiency, localized collections of pericardial blood, or extreme tamponade with hypotension. (65) It may also be absent when left ventricular diastolic pressure is intrinsically elevated owing to poor left ventricular compliance. This was seen in one
Figure 15-4 Normally systolic blood pressure drops slightly during inspiration. To measure pulsus paradoxus, the patient breathes normally while lying at a 45° angle. The blood pressure cuff is inflated well above systolic pressure and slowly deflated. When the pulse is first heard only during expiration, this is the upper value. The cuff is deflated until the pulse is heard during both inspiration and expiration, and this is the lower value. The difference in the two values is the amount of pulsus paradoxus. A difference of more than 12 mm Hg is abnormal.
Although the absence of a paradoxical pulse rules against severe tamponade, it does not completely rule it out. Whether time is taken to determine pulsus paradoxus depends on the patient s status. If the patient is moribund or rapidly deteriorating, taking time to check this parameter is obviously a poor choice of priorities.
Venous distention, reflecting increased CVP, is also a late sign in
cardiac tamponade (see Fig. 15-3) (Figure Not Available)
.
It may be masked by venoconstriction as a result of vasopressors (e.g., dopamine),
intrinsic sympathetic discharge, or hypovolemia. (33)
(64)
(73)
(78)
Neck vein distention may be obvious clinically, but the measured CVP is more
reliable than the presence of venous distention. The CVP reading should take
into account positive-pressure ventilation and the effects of a Valsalva
maneuver. Most patients with significant tamponade will have a CVP of 12
to 14 cm H2
O. (78)
Hypovolemia changes the intrapericardial pressure-volume curve in tamponade
and will lower the CVP reading at any given stage in the tamponade process.
Animal studies have documented that right atrial pressure can be normal in tamponade when hypovolemia is present. One case of low-pressure cardiac tamponade was reported in a patient with no jugular venous distention, no paradoxical pulse, and a right atrial pressure of 8 mm Hg. (65) Thus although the initial CVP reading is useful and diagnostic if grossly elevated (e.g., 20 to 30 cm H2 O), (51) (78) a series of CVP readings looking for an upward trend is the most sensitive diagnostic tool. (78) A rising CVP, especially when there is persistent hypotension, is extremely suggestive of tamponade in the trauma patient. In the rare case of the hypovolemic patient in whom tamponade is suspected but who demonstrates a low CVP, a fluid challenge will help clarify the situation and will also improve cardiac output at least temporarily. (65)
Routine chest radiographs and ECGs may be useful in increasing the level of suspicion for pericardial effusion and
Chest radiographs are not useful in the diagnosis of acute traumatic tamponade, because the cardiac size and shape do not change acutely. However, the radiographs may reveal hemothorax, bullet location, or even pneumopericardium.
In the nontrauma patient with chronic effusion, a chest film often reveals an enlarged, sac-like ""water bottle"" cardiac shadow. Unfortunately, it is difficult to differentiate pericardial from myocardial enlargement, and radiographs cannot be used to distinguish between simple pericardial effusion and tamponade. One finding that is useful in identifying effusion on the plain chest film is the epicardial fat pad sign. The water density space between the radiolucent epicardial fat and the mediastinal fat represents the pericardial tissues and is normally <2 mm. An increase in this width suggests pericardial fluid or thickening (Fig. 15-5) . This sign may be seen in 41% of upright lateral and 23% of frontal chest films in proven pericardial effusion. (81) The diagnostic value may be enhanced by using a supine rather than upright cross-table, lateral chest radiograph. Obtaining a supine lateral film increases the sensitivity of the epicardial fat pad sign from 31 to 51%. (82)
ECGs may suggest, but should not be
used to diagnose, pericardial effusion or cardiac tamponade. Most ECG changes,
such as PR-segment depression, low-voltage QRS complexes, and electrical
alternans, have acceptable specificity but poor sensitivity for pericardial
effusion or tamponade. (33)
(38)
Low voltage is defined as a QRS amplitude 5 mV in all limb leads (or
a sum of the limb lead QRS amplitude
30 mV), and PR depression is defined
as
1 mV in at least 1 lead other than aVR. In 1 study correlating the
ECG with echocardiographic evaluation, ECG signs had an overall sensitivity
of only 1 to 17% and a specificity of 89 to 100% for pericardial effusion. (33)
Others have demonstrated significantly higher sensitivity,
i.e., in the range of 32 to 68% for voltage criteria. (84)
PR-segment depression is the most common ECG finding in pericardial tamponade,
and low voltage is most commonly associated with a moderate to large effusion.
It is important to note that none of the ECG findings differentiate tamponade
from effusion.
Electrical alternans is caused by pendular motion of the heart within the pericardial sac. (85) Alternans of the QRS complex has been seen in about 22% of medical tamponade cases (69) but in only 5% of cancer patients with tamponade. (42) Electrical alternans of both the P wave and the QRS complex (total electrical alternans) is a rare finding but when seen is thought to be pathognomonic of tamponade (Fig. 15-6) (Figure Not Available) . (33) (86) Alternans does not always appear in the standard ECG leads; a bipolar chest lead (Lewis lead) may be needed to detect it.
Echocardiography is the best available tool for diagnosing pericardial effusion and has the further advantage of being noninvasive. (87) Echocardiography is very sensitive in the diagnosis of pericardial effusion and tamponade. (88) (89)
The disadvantages of echocardiography are that it requires ultrasound equipment and is dependent on a skilled operator who is specifically trained in echocardiography. Even when immediately available, echocardiography may take at least 5 minutes, which may be too much time for a patient who is deteriorating rapidly.
Figure 15-5 Epicardial fat pad sign. The water density space between the radiolucent epicardial fat and the mediastinal fat represents the pericardium and its contents and should be 2 mm or less. An increase suggests pericardial fluid or thickening. A, Left anterior-oblique chest film. B, Lateral chest film. On acute tamponade, the chest radiograph has very minimal diagnostic value.
Figure 15-6 (Figure Not Available) Overall, the ECG has a low sensitivity for pericardial effusion or tamponade, but PR depression, low voltage, or electrical alternans may be seen. Lewis lead electrocardiogram (ECG) showing total electrical alternans of both amplitude and configuration of P and QRS complexes. This is rarely seen but is almost pathognomonic of tamponade. Note that electrical alternans may not be evident in standard ECG leads. (From Sotolongo RP, Horton JD: Total electrical alternans in pericardial tamponade. Am Heart J 101:854, 1981. Reproduced by permission.)
The diagnosis of effusion is easily made by visualizing a large area of fluid, often best seen behind the heart. This view will also demonstrate the ventricular wall and quickly distinguishes the patient with a large effusion from the patient with congestive heart failure and a dilated ventricle. False-positive results may be obtained if a large amount of subepicardial fat is present. (90)
Effusion is not synonymous with tamponade, and the volume of fluid needed to produce tamponade in an individual depends on the thickness of the ventricular myocardium, the rate of fluid accumulation, and the distensibility of the pericardium. Echocardiographic diagnosis of tamponade is best made on the observation of right ventricular diastolic collapse. (91) (92) Other useful echocardiographic suggestions of tamponade are observation of the heart swinging rhythmically in the pericardial effusion, excessive respiratory variation in the size of the ventricles, greatly decreased right ventricular size, and pseudoprolapse of the mitral valve. (77)
Although right ventricular or atrial collapse has been touted to be
highly specific and sensitive for tamponade, (76)
when the definition of tamponade is based on clinical criteria, the picture
becomes less clear. Whereas echocardiography is very sensitive for tamponade,
it may be nonspecific. (89)
(93)
(94)
For example, in a group of 50 patients with an echocardiograph
diagnostic for tamponade, 94% had a systolic blood pressure 100 mm Hg
and 58% had a cardiac index
2.3 L/min/m (2)
. (89)
It is unclear whether the subset of patients with electrocardiographically
""proven"" tamponade but no clinical signs of tamponade have a
natural history that differs from those with a large, nontamponade pericardial
effusion. One author has suggested, ""In some of these patients, when
the etiology is known and the disease can be treated effectively with medication
. . . pericardial drainage may not be necessary."" (94)
If echocardiographic determination of tamponade is overly sensitive, do the echocardiographic signs associated with tamponade provide predictive value regarding outcome? In a retrospective evaluation of 187 hospitalized patients with pericardial effusion, the effusion size alone was found to have a strong predictive value for patient outcome when measured in terms of eventual tamponade or the combined end point of tamponade, surgical drainage, or pericardiocentesis. (95) The addition of criteria for tamponade, right atrial and ventricular collapse, and inferior vena cava (IVC) plethora added little, if any, prognostic value. Wall and coworkers, in an analysis of 57 patients with new, large pericardial effusions, found that these additional findings did not differentiate patients with and without tamponade, although cardiac tamponade occurred in almost 50% of patients, regardless of these findings. (96)
The risk of performing pericardiocentesis without echocardiographic determination of pericardial fluid was demonstrated by the Krikorian series. (27) Of patients with a clinical picture of tamponade, 17% actually had constrictive pericarditis, 16% had congestive heart failure and fluid overload, and 5% had obstruction of the superior vena cava. None of these patients could be expected to benefit from pericardiocentesis, and all were at risk for complications.
Although in most circumstances computed tomography (CT) is less readily available than echocardiography and requires that the patient be transported to the site of the CT equipment, CT is effective in defining the presence and extent of pericardial effusion in the stable patient. (97) In certain circumstances, CT can provide a more definitive diagnosis than echocardiography. In one series, eight equivocal echocardiograms were evaluated by CT. (98) Two patients thought to have pericardial effusion by ultrasound were found by CT to have pleural effusions. Another patient with pericardial effusion by ultrasound was found by CT to have an epicardial lipoma. Three loculated pleural effusions not seen by ultrasound were defined by CT. A final 2 patients had hemopericardium visualized by CT but not ultrasound. In circumstances where the patient is stable and ultrasound produces equivocal results or is not available, CT may provide a definitive diagnosis of pericardial effusion.
There are two indications for pericardiocentesis: (1) to diagnose the cause or presence of a pericardial effusion and (2) to relieve tamponade. The former is an elective procedure and ideally should be accomplished under ultrasound guidance. The latter may be semi-elective and performed with ultrasound guidance or emergent and performed blind with ECG assistance.
The use of pericardiocentesis for diagnosis of the etiology of nonhemorrhagic effusions is widespread, although opinions of its utility vary. (32) (99) (100) Neoplastic cells, blood, bacteria, viruses, and chyle can be sought. Measurement of pericardial fluid pH can be helpful, because inflammatory fluid is significantly more acidotic than noninflammatory fluid. (101) When a specific etiology is suspected, additional diagnostic testing may be useful (e.g., adenosine deaminase in tuberculosis, and carcinoembryonic antigen in suspected malignancy). (102)
The diagnostic accuracy of pericardiocentesis varies greatly from series to series, depending on the vigor with which a definitive etiology was sought and the prevalence of certain etiologies in the patient population under consideration. In one large series, fluid was obtained in 90% of the taps, but a specific etiologic diagnosis was obtained in only 24% of the fluid specimens. (27) Certain diagnoses are unlikely to be made from pericardial fluid. Pericardial fluid has been shown to give false-negative cytologic results in several cases of lymphoma and mesothelioma. (27) In HIV patients, effusions caused by Kaposi sarcoma and cytomegalovirus have been diagnosed by pericardial biopsy after fluid studies were nondiagnostic. (103) (104)
An alternative diagnostic tool is subxiphoid pericardiotomy. This technique, performed in the operating suite, obtains both fluid and a pericardial biopsy specimen. It is more likely to provide a definite diagnosis and has been performed safely without general anesthesia. (105) (106) In a prospective series of 57 patients, 36% obtained a definitive diagnosis; 40%, a probable diagnosis; 16%, a possible diagnosis; and 7% remained undiagnosed with subxiphoid pericardiotomy. (107) Although it is uncertain whether this technique is safer than ultrasound-guided pericardiocentesis, published reports show a low rate of complications. (107)
Regardless of technique, the need to sample small effusions or obtain pericardial tissue has been questioned. A prospective series found a diagnostic rate of 6% with pericardial fluid and 5% with pericardial tissue when a small persistent effusion was sampled for the specific purpose of diagnosis. In contrast, when patients from the same population had therapeutic intervention for tamponade, the yields from fluid and tissue were 54% and 22%, respectively. (100)
The use of pericardiocentesis as a diagnostic tool in traumatic tamponade is limited. When used diagnostically to determine the presence of pericardial bleeding in trauma, the procedure has a false-negative rate of between 20% and 40%. (78) (108) (109) (110) The reason for the high false-negative rate (defined as no blood aspirated) is well demonstrated by typical stab wounds of the heart. (8) (111) Ninety-six percent of the patients had blood in the pericardium, but it was clotted in 41% of the patients and partially clotted in another 24%. In only 19% was the blood completely fluid and thus capable of giving a true-positive result on pericardiocentesis.
The most common reason for performing pericardiocentesis in the emergency department (ED) is as part of the treatment for cardiac arrest or in peri-arrest situations. In particular, the presentation of pulseless electrical activity (PEA) with elevated jugular venous pressure should cause immediate consideration of pericardiocentesis. In this setting, blind, ECG-guided pericardiocentesis can be life-saving. However, the overwhelming majority of patients with PEA have neither significant effusion nor tamponade, and other etiologies for the PEA also should be sought. Pericardiocentesis also may be considered in other presentations of effusion with existing or incipient tamponade.
Pericardiocentesis is often, at least temporarily, therapeutic in cardiac tamponade. Most nonhemorrhagic effusions are liquids that can be drained easily through a small needle. Removal of even a small amount of fluid can immediately and dramatically improve blood pressure and cardiac output. Pericardiocentesis relieves tamponade due to nonhemorrhagic effusions in 60 to 90% of cases. (35) (27) (42) Patients in whom it fails often have purulent pericarditis or malignant invasion of the pericardium.
Pericardiocentesis without catheter placement may be much less useful for long-term management of these patients; 26% of the patients in the study by Guberman and coworkers eventually required pericardial resection. (35) In Krikorian s series, 24% of the patients were managed successfully with one pericardiocentesis, 37% needed multiple taps or an indwelling catheter, and 39% required surgical drainage. (27) Fifty-five percent of the last group had traumatic hemopericardium.
Patients with renal failure and pericardial effusion may be better managed by dialysis than pericardiocentesis. In one series, 63% of these patients were successfully managed with dialysis alone, and only 6% needed surgical treatment over the long term. (27) Tamponade is less frequent with pericarditis when it occurs within the first months of dialysis, and such patients are much more likely to be successfully managed without invasive intervention. (44) When invasive treatment is needed for dialysis patients, pericardiocentesis is probably a poor choice; 9 of 10 patients who received it had complications in one series, and it was the only invasive treatment that resulted in death. (44)
An algorithm for the urgent management of nonhemorrhagic cardiac tamponade is shown in Figure 15-7 .
Pericardiocentesis is never the definitive treatment in hemorrhagic tamponade. (112) (113) Although aspiration of a small quantity of fluid may cause dramatic improvement, blood usually reaccumulates. (33) (86) Thus, patients with pericardial hemorrhage ultimately require thoracotomy to explore and repair the cardiac injury.
One of the greatest potential drawbacks of pericardiocentesis in traumatic tamponade is that it may delay thoracotomy. In one study of 25 trauma patients with cardiac injury, (112) all of those who were operated on within 2 hours of injury survived, regardless of age or type of wound. With greater delay, none survived. Sugg and colleagues, in a study of 459 similar patients, found a mortality rate of 43% when pericardiocentesis was the sole treatment, but only 16% when surgery was performed. (109) Most investigators agree that with early thoracotomy and little or no reliance on pericardiocentesis, the number of deaths due to stab wounds has decreased. (78) (110) (113) (114) Sugg and associates reported that 10 of 18 patients with traumatic tamponade who were managed by repeated pericardiocentesis alone died within 1 to 2 hours. (109) At autopsy, all patients had completely repairable wounds.
Nonetheless, while other temporizing treatments are instituted (see discussion below) and arrangements for definitive surgical treatment are being made, pericardiocentesis may temporarily improve the patient s hemodynamic situation (Fig. 15-8) . Some clinical evidence supports the usefulness of pericardiocentesis as a temporizing measure. In a study of 174 patients with tamponade from penetrating trauma, 96 underwent operating room thoracotomy, 44 underwent emergency department thoracotomy, and 34 received
Figure 15-7 Management of nontraumatic cardiac tamponade. IV, intravenous line; CVP, central venous pressure; ECG, electrocardiogram.
Figure 15-8 Management of traumatic cardiac tamponade. IV, intravenous lines; CVP, central venous pressure; ECG, electrocardiogram; R/O, rule out.
When a trauma patient s condition is relatively stable, but a high level of suspicion for a penetrating cardiac wound is present, an alternative to thoracotomy is the subxiphoid pericardial window. (105) (115) (116) The procedure has been done under local anesthesia, and although it is possible to perform the procedure in the ED, (116) most authors believe the procedure should be reserved for the operating suite. (117) (118) (119)
There is no absolute contraindication to pericardiocentesis. It should not be performed when better treatment modalities are immediately available (e.g., dialysis for uremic patients and immediate surgery for trauma patients). For diagnostic or nonemergency pericardiocentesis, the absence of echocardiographic or CT diagnosis is a relative contraindication, because the complication rate increases dramatically under such circumstances. If ultrasound or fluoroscopic guidance is available, it should be used in nonemergent situations.
Pericardiocentesis is ideally performed in the cardiac catheter laboratory under fluoroscopic guidance. In the emergency department, as well as elsewhere in the hospital, echocardiography is useful for directing pericardiocentesis. With ultrasound, the area of the heart with the greatest fluid accumulation can be accurately identified and its relationship to the body wall clarified. (120) (121) An entry site and angle of penetration can then be chosen that have the greatest likelihood of obtaining fluid while simultaneously avoiding vital structures. Ultrasound also can be used to identify when the needle tip enters the pericardial space, although visualizing the needle with ultrasound can be difficult. (122) Specially made guides are available to allow visualization of the needle during the procedure and reduce complications. (123) Without
Figure 15-9 Equipment for emergent pericardiocentesis: long, 18-ga spinal needle; wire with alligator clips for connection to the electrocardiograph machine; and syringe (three-way stopcock optional). Sterile skin preparation and local anesthetic are also required.
Figure 15-10 An example of the contents of a prepackaged pericardiocentesis set: finder needle, Seldinger wire, dilator, catheter guide, and pigtail catheter. Sterile skin preparation and local anesthetic are also required.
Although the procedure can be performed with only a syringe and a spinal needle, ECG monitoring is desirable. An alligator clamp is used to connect the needle to any of the precordial leads (V leads) of a properly grounded ECG device (Fig. 15-9) . Generally the V lead (usually V1 or V5 ), which permits a continuous display during rhythm monitoring, is used.
The traditional needle choice has been a 7.5- to 12.5-cm (3- to 5-inch), 18-ga spinal needle with an obturator. It is best to leave the obturator in the needle during initial passage through the skin to avoid obstruction of the needle lumen. More recently, the shorter Teflon-sheathed Intracath needle has been used. Alternatively, the clinician can use the guide wire (Seldinger) technique, inserting a plastic catheter over a flexible guide or J wire. With this technique, an 18-ga, thin-walled needle is used for placement of the wire. The catheter (after removal of the accompanying introducer) may be left in place for prolonged drainage, if needed. (124) (125)
For drainage of blood, pus, or other viscous effusions, a large catheter such as a No. 7 to 9 Fr Cordis sheath should be inserted. (126) Alternatively, the guide wire technique can be used to insert a radiopaque, 16-ga, flexible, fenestrated, central venous catheter, which can then be connected to closed suction drainage and left in place for long periods of time. (127) Pigtail catheters with side and end holes or nephrostomy drainage catheters also can be used. (125) Multilumen catheter patency can be maintained by slow continuous flush with heparinized saline solution. (125) Complete ""sets"" containing necessary equipment for placing a catheter using the guide wire technique are commercially available (Fig. 15-10) , including sets designed for pediatric use. (128)
A three-way stopcock may be attached to the needle or catheter to allow removal of more than one filled syringe without much movement of the needle. The continuous motion of the heart may require minor changes in needle or catheter position during the procedure. Lengthy or repeat drainage is much safer if the steel needle is withdrawn and a plastic catheter is left in place.
While preparing for pericardiocentesis in the unstable patient or attempting to stabilize the patient while the operating suite is readied for thoracotomy or subxiphoid pericardiotomy, temporizing measures should be considered. In the patient with suspected tamponade and without jugular venous distention, the administration of a fluid bolus may improve hemodynamics. (65) In the setting of nonpenetrating tamponade, a fluid challenge has been recommended (56) (129) ; animal experiments have found this to be beneficial, with or without nitroprusside for afterload reduction. (130) However, a follow-up prospective evaluation in patients with tamponade found no benefit from either fluid challenge or nitroprusside; cardiac output remained unchanged at a mean of 5.1 L/min, in contrast to 9.1 L/min after pericardiocentesis. (131) In the trauma patient with penetrating cardiac injury, fluid resuscitation may produce improvement or deterioration. Animal experiments indicate that the response depends on whether fluid infusion produces recurrent bleeding from the cardiac wound. (132) One report found that dextran solution for volume expansion produced significant hemodynamic improvement in patients with subacute ventricular free wall rupture after acute myocardial infarction. (30) In summary, judicious volume expansion may produce temporary beneficial hemodynamic results, but this is not uniformly true.
Pressors also have been recommended as a temporizing measure in tamponade. Dopamine, dobutamine, norepinephrine, and isoproterenol have been evaluated. Norepinephrine produced increased cardiac output in animal models of tamponade (133) (134) but failed to increase cardiac output in patients with malignant effusion. (133) Isoproterenol produced increased cardiac output in animal models but detrimentally affected cardiac blood flow. (133) Both dopamine and dobutamine have produced increased cardiac output and other improvements in hemodynamics in the setting of tamponade. (20) (134) Either of these agents may be helpful as a temporizing agent in tamponade, but dobutamine may be preferable on theoretical grounds because of its greater beta activity. (134)
All necessary equipment must be checked and laid out in advance. Full resuscitation equipment must be on hand, including a defibrillator. The patient must have an IV line in place and be attached to a cardiac monitor. The nonemergency patient may require sedation, but in an emergency, pericardiocentesis is usually performed on patients who are already obtunded or unresponsive as a result of low cardiac output. Use of sedation in these patients not only is unnecessary, but also carries a high risk of hemodynamic or respiratory deterioration. Premedication of the patient with atropine may help to prevent vasovagal reactions. When possible, the presence of pericardial effusion and the optimal anatomic approach should be determined in advance by echocardiography. If surgery may be needed, preparations should already be under way to ensure prompt availability of both an operating room and a surgeon.
If the patient s clinical condition permits, the chest should be elevated at a 45° angle to bring the heart closer to the anterior chest wall. If the abdomen is distended because of gastric contents or previous positive-pressure ventilation, a nasogastric tube should be used to decompress the stomach. The entire lower xiphoid and epigastric area should be carefully prepared with 10% povidone-iodine solution and sterilely draped, if time permits.
If the patient is awake, the skin and the proposed route of the pericardial needle should be anesthetized by infiltration with 1% plain lidocaine or 0.5% bupivacaine. Note that the pericardium is very sensitive and should be anesthetized in patients who are awake. (124)
The choice of anatomic approach in the past has been governed largely by conjecture and theory, not by actual study of patients with pericardial effusion. Traditionally the subxiphoid approach was preferred and is widely touted in most texts and articles as by far the optimal choice. However, two-dimensional echocardiography allows direct visualization in the individual patient of both the areas of maximal effusion and the location of vital structures, and studies of echocardiography-directed pericardiocentesis have found that the intercostal space near the heart apex is usually the best site for puncture, not the traditional subxiphoid approach. (121) (126) Careful cadaver studies have corroborated this finding, demonstrating greater safety with a parasternal approach in the fifth intercostal space and showing that the greatest number of injuries (usually to the right atrium) occurred with any variant of the subxiphoid approach. (13) In contrast, studies of intracardiac injection using the same routes have found an increased incidence of pneumothorax when parasternal or intracostal approaches are used (see discussion of complications of pericardiocentesis). This risk may increase with underlying lung disease. Whenever time and the patient s condition permit, clinicians should rely on echocardiography to define the extent of and optimal approach to pericardial effusion. When time or circumstances prevent the use of ultrasound, the clinician should use the approach with which he or she is most familiar.
In this approach, the needle is inserted perpendicular to the skin in the left fifth intercostal space medial to the border of cardiac dullness (Fig. 15-11) . Older texts identify the puncture site as being at least 3 to 4 cm lateral to the sternal border to avoid the internal mammary artery. However, anatomic studies indicate that penetration immediately lateral to the sternum is less likely to cause this complication. (13)
In the traditional subxiphoid approach, the needle is inserted between the xiphoid process and the left costal margin at a 30° to 45° angle to the skin (Fig. 15-12) . Because the heart
Figure 15-11 Parasternal approach for pericardiocentesis. The patient is depicted in a supine position, although a preferable position would be sitting at a 45° angle, if the patient s clinical condition permits.
Figure 15-12 A and B, Xiphosternal approach for pericardiocentesis. The needle is aimed for the sternal notch or the left shoulder. Note the electrocardiography monitor. Although the patient is shown in a supine position, a preferable position would be sitting at a 45° angle, if the patient s condition permits. This general approach is also used for intracardiac injection of advanced cardiac life support drugs.
In the less commonly used apical approach, the needle is inserted 1 cm outside the apex beat in the intercostal space below the apex beat and within the area of cardiac dullness, and aimed toward the right shoulder. (124) If the apex cannot be palpated, the needle is inserted just inside the area of cardiac dullness. This area is close to the lingula and left pleural space, and pneumothorax is more frequent; a concomitant pleural effusion may be inadvertently tapped. In theory, this technique is used because the coronary vessels are small at the apex, and if a ventricle is entered, it is the thick-walled left ventricle, which is more likely to seal off any ventricular injury. Data are insufficient to say whether these theoretical advantages are important.
After the skin has been punctured but before the pericardial needle is advanced, any obturator in the needle is removed, and an aspirating syringe is attached. At this time, ECG monitoring is begun by attaching a sterile cord with alligator clips (see Fig. 15-9) from the pericardial needle to any precordial lead (V lead) of the ECG machine. The V lead is then recorded, as the needle becomes an ""exploring electrode."" The machine must be properly tested and internally grounded. Small current leaks can induce dysrhythmias. (33) The purpose of the ECG monitoring is to prevent ventricular puncture. When the needle touches the epicardium, a current-of-injury pattern is noted on the ECG (Fig. 15-14) . This current of injury may be local and could be missed if a lead other than a V lead is monitored or if a cardiac monitor (which has a lower frequency response than the ECG machine) is used. Usually one notes ST-segment elevation on contact with the heart or pericardium in the absence of an effusion, but a premature contraction or other ventricular dysrhythmia also may be induced by direct mechanical stimulation of the ventricular epicardium by the needle. Contact with the atrium can cause atrial dysrhythmias, marked elevation of the PR segment, or atrioventricular dissociation. (15) If there is abnormal myocardial scarring secondary to infarction or other diseases or if there is malignant infiltration of myocardium, no current of injury may be generated. (16) Thus, ECG monitoring is not infallible in preventing myocardial penetration. In addition, the incessant motion of the heart makes it almost impossible to merely touch the epicardium.
With constant ECG monitoring, the operator slowly advances the needle and syringe while gently aspirating with the syringe. The needle will penetrate the pericardium (a barrier whose penetration usually cannot be palpated) at about 6 to 8 cm below the skin in adults and 5 cm or less below the skin in children. (54) The patient who is awake may complain of sharp chest pain as the sensitive pericardium is entered. When pericardial content is aspirated, the needle should not be advanced further. If a current of injury is noted (see Fig. 15-14) , the needle is touching epicardium and can easily lacerate myocardium or coronary vessels. The needle should be withdrawn a few millimeters until the
Figure 15-13 (Figure Not Available) During ""blind"" pericardiocentesis, the subxiphoid approach is
recommended. A short needle (16- or 18-ga) is inserted into the left xiphocostal
angle perpendicular to the skin and 3 to 4 mm below the left costal margin
(A). After advancing the needle to the inner
aspect of the rib cage, the needle s hub is depressed so that the needle
points toward the patient s left shoulder. The needle is then cautiously
advanced about 5 to 10 mm until fluid is reached (B). The fingers may sense a distinct ""give"" when the
needle penetrates the parietal pericardium. Successful removal of fluid
confirms the needle s position. The syringe is then disconnected from the
needle, and the flexible tip of the guide wire is advanced into the pericardial
space under fluoroscopic guidance (C). The
needle is withdrawn and replaced with a soft, multihole pigtail catheter
(No. 6 to 8 Fr) using the Seldinger technique. After dilating the needle
tract, the catheter is advanced over the guide wire into the pericardial
space (D). Once the catheter is properly
positioned, aspiration of fluid should result in rapid improvement in blood
pressure and cardiac output, a decrease in atrial and pericardial pressures,
and a decrease in the degree of any paradoxical pulse (E). Electrical alternans, if present, also decreases or disappears.
(From Spodick DH: The technique of pericardiocentesis. J Crit Illness 2:91-96,
1987.)
Figure 15-14 Current of injury. There is an obvious
change in the electrocardiogram when the pericardiocentesis needle touches
the epicardium. Following slight withdrawal (arrow), the ST-segment elevation diminishes.
An attempt is then made to drain pericardial fluid or blood. If blood is obtained, the possibility of cardiac puncture should be entertained. If fluoroscopy is available, the injection of a small amount of contrast will quickly disclose intracardiac placement. In other circumstances, the needle may need to be repositioned and the aspirate reexamined. Laboratory tests may help distinguish circulatory blood from hemorrhagic pericardial fluid. The latter should have a lower hematocrit measurement than venous blood. Substantially different hematocrit values rule out the possibility that the needle was in a cardiac chamber. Hemorrhagic pericardial fluid usually is about 0.10 pH unit more acid than simultaneously obtained arterial blood. (101) Bloody pericardial fluid may clot, particularly in traumatic situations in which bleeding is brisk, so clotting of the aspirated blood does not eliminate the possibility of a pericardial source. Nonclotting blood is indicative of defibrinated pericardial blood. Practically, however, there is rarely time for such analysis.
If an indwelling catheter is to be placed, a guide wire should be advanced through the needle (see Fig. 15-13) (Figure Not Available) . A dilator is passed over the needle to expand the needle tract. The guide wire should be maintained in sight and stabilized at all times. If intracardiac placement of the needle or guide wire is suspected, positioning must be verified by ultrasound or fluoroscopy or by using the techniques described above before the needle tract is dilated. Once the tract has been dilated, the pigtail catheter is placed over the guide wire. If the dilator is not used, particularly with the subxiphoid approach, the pigtail catheter tip may hang in the subcutaneous (SQ) tissue, making placement difficult.
After the catheter is placed, or if a decision is made to do a single aspiration, as much fluid as possible should be aspirated from the pericardium. The removal of even 30 to 50 mL may result in marked clinical improvement in patients with tamponade. The catheter may be placed for continuous or intermittent drainage. A chest film should be obtained after the procedure to rule out iatrogenic pneumothorax. Patients should be monitored closely for 24 hours for signs of reaccumulating fluid or iatrogenic complications from the procedure. Repeat ultrasound examination is wise. Diagnostic evaluation of nonhemorrhagic fluid is similar to the analysis of pleural fluid (see Chapter 8) .
The failure of pericardiocentesis to yield fluid (""dry tap"") may be considered a complication, as the procedure has failed to achieve its desired result. If a dry tap is considered a complication, it is by far the most frequent one. In addition, the pericardial needle can injure any organ within its reach, causing pneumothorax or myocardial or coronary vessel laceration and, thus, hemopericardium. (114) Venous air embolism may be caused by air entering the heart. (135) The pericardial needle can also induce dysrhythmias from direct irritation of the epicardium or from small currents leaking from the connected ECG machine. (2)
Assessing the frequency of complication from pericardiocentesis is not as simple as it might appear. Changes in diagnosis of effusion by ultrasound and guidance of the procedure by ultrasound or fluoroscopy have greatly reduced the likelihood of complication. (100) (126) For example, Wong
Six studies summarized in Table 15-4 demonstrate that the risks of pericardiocentesis remain quite significant. The major complications will be discussed individually.
Cardiac arrest and death occurred in approximately 2% of patients in the larger series listed in Table 15-4 , but in none of a series of patients whose pericardiocentesis was directed solely by ultrasound. However, an exact causal relationship between pericardiocentesis and sudden death is difficult to substantiate. For example, in one series of 52 patients, the only death occurred in a patient in cardiogenic shock who had a nonproductive pericardiocentesis and who, on postmortem examination, had severe arteriosclerotic heart disease, not tamponade. (3) An additional case of cardiac arrest (successfully resuscitated) in this series was in a patient with a nonproductive pericardiocentesis; the cause of the arrest was not discussed. (3)
In two studies comprising 197 patients who underwent pericardiocentesis under echocardiographic guidance, there have been no cardiac arrests or deaths. (100) (126) In a series of 352 pericardiocenteses performed under fluoroscopic guidance, only two deaths resulted. (4) Ultrasound or CT confirmation of effusion was used in all but 15 cases. The two deaths occurred during or after the procedure, but whether they should be attributed to the procedure is unclear. One patient with aortic rupture penetrating into the pericardial space died of cardiac arrest immediately after the puncture. The other death, in a postmyocardial infarction patient with left ventricular aneurysm, was due to ventricular fibrillation that occurred about 15 minutes after the procedure.
Cardiac chamber, vessel, or lung laceration occurs in 6 to 9% of patients, even in the hands of experienced physicians under controlled situations. Nonfatal cardiac puncture, pneumothorax, and pneumoperitoneum have been reported, (45) as well as suppurative costochondritis. Most cardiac perforations occur in the right ventricle, but left ventricular (4) as well as atrial punctures have been reported. (15)
In Krikorian and Hancock s series, 13 of 123 patients developed hemopericardium as a result of pericardiocentesis, 1 as a result of a lacerated coronary artery. (27) One patient died from a punctured ventricle. Surgical control was necessary for 4 patients who developed tamponade, whereas 8 patients with hemopericardium did not develop tamponade and were managed conservatively. Several cases of induced tamponade occurred in patients with platelet counts >50 × 10 (9) /L.
Guberman and coworkers reported 3 right ventricular lacerations in 46 patients; one was fatal. (35) Wong and colleagues found 5 right ventricular punctures, 4 in patients with nonproductive pericardiocentesis, but none causing any adverse sequelae. (3) In a series of dialysis patients, 9 of 10 receiving pericardiocentesis had serious complications, including 3 deaths and 2 myocardial lacerations. (44) Duvernoy and coworkers (4) reported 23 penetrations (all right ventricular except 2 in which the right and left ventricles had been perforated), along with 4 cases of significant arterial bleeding in a series of 352 procedures.
Researchers differ in their opinions as to the adverse effects of ventricular
puncture. Most ventricular punctures during the procedure occur in the lower
aspect of the right ventricle. Because the pressure is lower there than in
the left ventricle, (54)
there should be less bleeding;
however, the ventricular wall is also thinner and more vulnerable to laceration.
In a series of patients whose pericardiocentesis was exclusively directed
by ultrasound, ventricular puncture still occurred in 1.5% but was without
consequence due to small needle size. (126)
In another
study, right ventricular laceration occurred in 1 patient despite the use
of echocardiography, producing tamponade and necessitating emergency surgery. (100)
|
|
Permanyer- Miralda(100) (1985) |
Wong(3)
(1979) |
Guberman(35)
(1981) |
Krikorian(27)
(1978) |
Kwasnik(45)
(1978) |
Callahan(126)
(1985) |
|---|---|---|---|---|---|---|
| No. of cases | 80 | 52 | 56 | 123 | 34 | 117 |
| Environment | Cardiology service in Spain, all echoed | Cath lab with echo, fluoroscopy, no trauma | Cardiology service with echo, fluoroscopy, etc., no trauma | University hospital, most in cath lab, 9% trauma | All uremic patients | Mostly medical, all echo-directed |
| Success in obtaining fluid (%) | 88 | 69 | 87 | 86 | -- | 98 |
| Diagnosis from taps (%) | 19 | 50 | 60 (malignancy only) | 18 | -- | -- |
| Cardiac arrest (% resuscitated) | 0 | 2 | 2 | -- | -- | 0 |
| Death (%) | 0 | 2 | 2 | 1.6 (3.2 * ) | -- | 0 |
| Ventricular puncture or laceration (%) | 1.2 | 9 | 6.5 | -- | -- | 1.5 (minor) |
| False-negative taps (%) | 5.3 | 7.6 | -- | -- | -- | 0.8 |
| Surgery needed for tamponade (%) | 1.2 | -- | 26 | 39 | -- | -- |
| New hemopericardium (%) | 1.2 | -- | -- | 10.5 | -- | 0 |
| Major dysrhythmias (%) | 0 | -- | -- | 0.08 | -- | 0 |
| Hypotensive episode (% vasovagal) | -- | -- | -- | 2 | -- | 0.8 |
| Pneumothorax (%) | 0 | -- | -- | -- | 3 | 0.8 |
| Pneumoperitoneum (%) | 0 | -- | -- | -- | 3 | 0.8 |
A small number of pneumothoraces and pneumopericardium have been reported in various series (see Table 15-4) but have been without clinical consequence. A single case of tension pneumothorax has been reported after pericardiocentesis, but the cause-and-effect relationship was unclear. (136)
Serious dysrhythmias induced by pericardiocentesis are rare. Premature ventricular contractions (PVCs) occur commonly during the procedure and are benign in most cases. Wong and coworkers, (3) Guberman and colleagues, (35) Callahan and investigators, (126) and Kwasnik and coworkers (45) reported no dysrhythmias. Krikorian and Hancock reported only one episode of ventricular tachycardia and several ""hypotensive vasovagal reactions,"" which were associated with bradycardia and responded to atropine and fluid loading. (27) Duvernoy and colleagues (4) reported 1 case of ventricular tachycardia and 1 case of atrial fibrillation among 352 procedures.
There have been a few case reports of adverse consequences even when pericardiocentesis inflicts no injury. Most of these have to do with the fact that during pericardiocentesis, the stroke volume of the previously collapsed right ventricle increases 77% with the first 200 mL of fluid removed. (57) Generally, this increase in stroke volume is greater initially than that demonstrated by the left ventricle. This can have significant consequences for both right and left ventricular function. In 3 of 6 patients in whom large effusions were removed by pericardiocentesis, there was right ventricular dilation and overload, with abnormal septal motion and either no increase in right ventricular ejection fraction or a decrease. (137) These patients returned to normal hemodynamic status slowly.
Sudden pulmonary edema also has been reported after pericardiocentesis, presumably due to a sudden increase in venous return to the left ventricle at a time when peripheral vascular resistance is still high from compensatory catecholamine secretion. (138) (139) (140) Supporting evidence for this explanation is that right ventricular stroke volume increased 76% after relief of tamponade, greater than the stroke volume increase of the left ventricle. (60) Circulatory collapse with persistently low arterial blood pressure has been reported in a patient who was drained of 700 mL of clear fluid at a rate of 100 mL/min. (141) These authors suggest that ketamine anesthesia may have played a role, but the relative ischemia created by tamponade, coupled with the sudden increase in left-sided preload, created a persisting imbalance. They recommend that pericardial drainage rate not exceed 50 mL/min. Given the rare occurrence of pulmonary edema or primary cardiac compromise, it is unclear whether this limitation is justified.
A case of brief profound bradycardia and rebound hypertension has also been reported after surgical relief of tamponade. (142) Such responses have not been noted in large series of patients receiving pericardiocentesis.
In nontraumatic patients, tamponade should always be considered in the differential diagnosis of shock, especially in patients who are on anticoagulants, who have had recent myocardial infarction, or who have had pericardial disease; when malignancy is present; when aortic dissection is suspected; or when a CVP catheter is in place. Tamponade should also be considered in the differential diagnosis when hypotension persists following closed-chest CPR or attempts at cardiac pacing.
In any patient with blunt or penetrating chest or upper abdominal trauma, the possibility of traumatic tamponade must also be considered. If clinical deterioration occurs in the emergency department pending operative care, temporizing pericardiocentesis should be considered if other therapy fails. When such a patient arrives with no obtainable blood pressure or in profound shock and unconscious, immediate thoracotomy and pericardiotomy are indicated after intubation. (113) (143) (144) Pericardiocentesis may cause a dangerous delay in this situation and has a low success rate.
Management of tamponade requires a sound understanding of pathophysiology; an ever-vigilant evaluation; and the willingness, if necessary, to perform relatively high-risk procedures such as pericardiocentesis in critically ill or injured patients.
|
|
|
"