Case 22-2004 — A 30-Year-Old Woman with a Pericardial Effusion
http://www.100md.com
《新英格兰医药杂志》
Presentation of Case
A 30-year-old woman was admitted to the hospital because of fever, cough, shortness of breath, and a pericardial effusion.
Approximately one month earlier, the woman had noticed an increased heart rate and fluttering in her neck when she bent over; both were relieved when she stood upright. During these episodes, her heart rate was 100 to 120 beats per minute. One week later, she noticed shortness of breath after walking up one flight of stairs. Since the age of 19 years, she had had intermittent atrioventricular nodal reentrant tachycardia, with heart rates of up to 180 beats per minute, and she had had approximately three episodes a year that resolved with the Valsalva maneuver. She had also had three episodes that had required visits to the emergency department and treatment with verapamil and adenosine. Because her current symptoms reminded her of these episodes, she consulted her cardiologist and wore a Holter monitor for a 24-hour period. No abnormal rhythms were revealed.
One week later, the woman noticed increased fatigue, and several days later fever (temperature up to 38.0°C) developed, with a dry cough and increased shortness of breath. One episode of substernal chest pain occurred; it lasted 30 minutes and was relieved with acetaminophen. She visited her primary care doctor one week after this episode and was found to have anemia. The thyrotropin level was 0.65 μU per milliliter and the thyroxine level was 1.14 mg per deciliter (14.67 nmol per liter). Four days later, she went to the emergency department of another hospital because of persistent cough and fever (temperature up to 38.1°C). A chest radiograph showed a right pleural effusion and a possible infiltrate. A 10-day course of levofloxacin was begun. Two days later, the formal report of the chest radiograph was issued and noted an enlarged cardiac silhouette. An echocardiogram showed a pericardial effusion with right ventricular diastolic collapse — findings consistent with tamponade. She came to this hospital and was admitted.
The patient was a physician in training. She did not smoke, drink alcohol, or use illicit drugs. A sister had Graves' disease, and her mother had hyperthyroidism. Eleven months before admission, she had traveled to Nairobi, Kenya, for a month-long medical rotation. Since then, she had not traveled outside of the United States or to Martha's Vineyard, Massachusetts. A tuberculin skin test performed five months after she had returned from Africa was negative. She had had multiple exposures to children with viral illnesses in the course of her work. There was no history of rash or joint swelling.
On physical examination, she appeared well. The blood pressure was 130/83 mm Hg, the heart rate 98 beats per minute, the respiratory rate 20 breaths per minute, and the temperature 36.1°C. There was a pulsus paradoxus of 8 mm Hg (140 mm Hg decreasing to 132 mm Hg). The hemoglobin oxygen saturation was determined by pulse oximetry to be 100 percent while the patient was breathing ambient air. An examination of the woman's head, eyes, ears, nose, and throat showed no abnormalities. An examination of her neck showed a jugular venous pressure of 12 cm of water with bilateral carotid bruits and no lymphadenopathy. The breath sounds were clear. An examination of the heart showed no murmurs but did reveal a possible pericardial friction rub at the left lower sternal border. Abdominal and neurologic examinations showed no abnormalities, and the arms and legs were warm, with symmetric pulses and no edema. Laboratory test results are shown in Table 1.
Table 1. Laboratory Values on Admission.
A chest radiograph disclosed small, bilateral pleural effusions and an enlarged cardiac silhouette (Figure 1A). An electrocardiogram showed a normal sinus rhythm at 98 beats per minute; QRS intervals of 0.138, 0.075, and 0.405 second; a normal axis; diffuse T-wave inversions; and a downsloping J point in the V2 lead. An echocardiogram showed a large, circumferential pericardial effusion with right atrial inversion (Figure 2), diastolic inversion of the right ventricular wall, and respiratory variation in the intracardiac flows — findings consistent with increased intrapericardial pressure.
Figure 1. Chest Radiographs.
On a chest radiograph obtained on admission (Panel A), there are small bilateral pleural effusions and an enlarged cardiac silhouette. A chest radiograph obtained two months after discharge (Panel B) shows resolution of the pleural effusions and the development of new, small nodular opacities in the right upper lobe.
Figure 2. Echocardiograms Obtained on Admission.
In an apical four-chamber view (Panel A and Video Clip 1 of the Supplementary Appendix), there is a large pericardial effusion (bar) and inversion of the right atrium (arrow), caused by elevated pericardial pressure, in late diastole and early systole. A parasternal short-axis view (Panel B and Video Clip 2 of the Supplementary Appendix) shows that the right ventricular outflow tract is compressed in diastole because of the elevated pericardial pressure (arrow). The bar shows the pericardial effusion. A Doppler study of flow across the mitral valve (Panel C) shows respiratory variation in the mitral inflow E wave, a sign of tamponade. The arrowheads denote expiration, the asterisks inspiration, E diastolic mitral E wave, and A diastolic mitral A wave. The scale to the right indicates flow velocity, measured in centimeters per second.
Pericardiocentesis was performed, and 370 ml of bloody fluid was removed. The pericardial pressure decreased from 7 to 0 mm Hg. A catheter drained decreasing amounts of fluid during the next 48 hours and was removed. A cytologic examination of the pericardial fluid showed blood and no malignant cells. A Gram's stain showed red cells but no neutrophils or organisms, and a culture showed no growth of bacteria. A smear and a culture were negative for acid-fast bacilli and mycobacteria. An echocardiogram obtained on the second hospital day showed no reaccumulation of the pericardial effusion. The patient was discharged on the third day. A tuberculin skin test was performed during the hospitalization, and 16 mm of induration developed two days after discharge.
Echocardiograms obtained one and two months after discharge showed a minimal pericardial effusion, diffuse fibrin deposition on the visceral pericardium, and no evidence of tamponade. Two months after discharge, the patient was seen in the pulmonary clinic for evaluation of the positive tuberculin skin test. On physical examination, the lungs were normal on auscultation. Smears and cultures of three specimens of sputum were negative for acid-fast bacilli and mycobacteria. A radiograph of the chest showed resolution of the pleural effusions but two new, small nodular opacities in the right upper lung zone (Figure 1B). Computed tomographic (CT) scanning of the chest, performed without the intravenous administration of contrast material (Figure 3A and Figure 3B), showed a cluster of small nodules, 3 to 5 mm in diameter, in the right upper lobe, as well as multifocal mediastinal lymphadenopathy, a confluent mass of subcarinal lymph nodes, and an enlarged right hilar lymph node. There was minimal pericardial thickening and slight thickening of the right posterior pleura.
Figure 3. CT Scans of the Chest.
A CT scan of the chest obtained without the use of intravenous contrast material two months after discharge shows mediastinal lymphadenopathy (Panel A) and a cluster of small nodules in the right upper lobe (Panel B). A CT scan of the chest obtained with the use of intravenous contrast material four months after discharge reveals a decrease in the size of the mediastinal lymph nodes and in the size of the nodules in the right upper lobe (Panel C) and a new right pleural effusion (Panel D).
One week later, bronchoscopy was performed, with transbronchial fine-needle aspiration of the subcarinal lymph nodes, transbronchial biopsy of the right upper lobe, and bronchoalveolar lavage of the right upper lobe. Cytologic examination of the specimen obtained by fine-needle aspiration disclosed no malignant cells, and flow cytometry showed no evidence of lymphoma. Pathological examination of the specimen from the transbronchial lung biopsy showed bronchial epithelium with no evidence of malignant cells or granulomas. Cultures of the bronchial-lavage fluid yielded normal flora, no mycobacteria, and no fungi. The serum angiotensin-converting–enzyme level was 38 U per liter (normal range, 7.0 to 46.0).
Two months later, another CT scan of the chest (Figure 3C and Figure 3D) showed a decrease in the size of the anterior mediastinal lymph nodes, no change in the size of the lymph nodes elsewhere, and nodular opacities in the right upper lobe, one of which had decreased in size. A new right pleural effusion had developed.
A diagnostic procedure was performed.
Differential Diagnosis
Dr. Edward A. Nardell: May we review the echocardiograms and the imaging studies of the chest?
Dr. Dali Fan: Normally, the space between the parietal and visceral pericardium is not visualized on an echocardiogram. In this patient, there is an echolucent space behind the posterior wall and anterior to the right ventricular outflow tract, indicating the presence of a pericardial effusion (Figure 2A). The effusion completely surrounds the heart, with a separation of at least 1 cm between the parietal and visceral pericardium, indicating that the fluid collection has an estimated volume of more than 500 ml.
In tamponade, cardiac filling is compromised by elevated pericardial pressure. The right atrium, the chamber with the lowest pressure, is the first to be affected. In the apical four-chamber view (Figure 2A and Video Clip 1 of the Supplementary Appendix, available with the full text of this article at www.nejm.org), the contour of the right atrium is inverted in late diastole and early systole, when the internal pressure is lowest. This is a sensitive but relatively nonspecific sign of tamponade, since pulmonary diseases can cause a similar phenomenon.1 The right ventricle and outflow tract are also vulnerable to elevated pericardial pressure. The parasternal short-axis M-mode echocardiogram (Figure 2B and Video Clip 2 of the Supplementary Appendix) shows that the right ventricular outflow wall is deflected posteriorly after the mitral valve opens because of the elevated pericardial pressure. The sensitivity of this finding is lower, but the specificity is higher than that of right atrial collapse.2 Video Clip 3 of the Supplementary Appendix shows that the heart swings dramatically within the large pericardial effusion. This movement occurs in both the anteroposterior and mediolateral directions and may include an exaggerated twisting of the heart around its long axis.3,4,5
The normal increase in flow across the mitral valve during expiration and the decrease during inspiration are exaggerated in tamponade.6 On the Doppler study of mitral inflow (Figure 2C), the variation of mitral valve filling with respiration (the peak E wave) is greater than 25 percent. In summary, anatomical evaluation shows that this patient has a large pericardial effusion, and a physiological evaluation reveals that she has changes consistent with increased pericardial pressure and tamponade.
Dr. Jo-Anne O. Shepard: The chest radiograph obtained at the time of admission (Figure 1A) reveals small bilateral pleural effusions and an enlarged cardiomediastinal silhouette. As the chest radiograph obtained two months after the patient's discharge (Figure 1B) shows, the pleural effusions subsequently resolved and the cardiac silhouette decreased in size. There are new, faint, irregular nodular opacities in the right upper lobe. A chest CT scan also obtained at that time (Figure 3A) shows lymphadenopathy in the anterior mediastinum in the paratracheal region and in the subcarinal and right hilar regions. There is minimal right lateral pleural thickening. There is a cluster of nodules, 3 to 5 mm in diameter, in the periphery of the right upper lobe (Figure 3B). A chest CT scan obtained four months after discharge shows a decrease in the mediastinal lymphadenopathy, and the nodular opacities in the right upper lobe have slightly decreased in size (Figure 3C). The subcarinal and right hilar lymphadenopathy remains stable. A new right pleural effusion is present (Figure 3D), and the nodular opacities in the right upper lobe have slightly decreased in size.
Dr. Nardell: This 30-year-old woman came to the hospital with pleural and pericardial effusions, tamponade, fever, pulmonary nodules, and mediastinal lymphadenopathy. Of these clinical features, the most immediately life-threatening is pericardial effusion with tamponade, and I will build my discussion around this relatively uncommon finding in a young, otherwise healthy woman.
The pericardial effusion along with the progressive tamponade explains many of the symptoms that culminated in the patient's hospitalization. As the pericardial sac fills with fluid, diastolic filling is increasingly compromised, especially in the low-pressure right ventricle. The echocardiogram also showed right ventricular diastolic collapse, a finding that is said to correlate with a reduction in cardiac output of 15 to 25 percent. With tamponade, stroke volume decreases, and cardiac output is maintained by an increased heart rate. Central venous pressure increases and aids diastolic filling. Bending over could have transiently reduced her venous return and cardiac filling by reducing the hydrostatic pressure in the superior vena cava or by compressing the inferior vena cava. Either one of these mechanisms could result in tachycardia and reduced cardiac output. Her blood pressure was normal, which is probably a result of physiologic compensation. These findings suggest a chronic, rather than an acute, process.
The evidence of pericardial tamponade quickly reversed when 370 ml of fluid was removed. The finding of so much fluid is in itself helpful in establishing the cause. Pericardial fluid can accumulate in large amounts only if it does so slowly. Traumatic effusions, effusions due to acute bacterial infections, and other acute effusions compromise cardiac function and can rapidly lead to death because there is no time for the pericardium to stretch to accommodate the fluid. Pericardial effusions due to viral infections, lupus, and rheumatoid arthritis also tend to be small but can occur rapidly and lead to tamponade.
This patient also had bilateral pleural effusions. These could be part of the underlying illness, or they could be secondary to her effective heart failure due to tamponade. The most common causes of bilateral pleural effusions are congestive heart failure and renal failure. Systemic inflammatory illnesses, such as lupus and widespread metastatic cancer, can also cause bilateral pleural effusions, but they are less common. At this point, I will assume that the pleural effusions were due to heart failure. However, I will attribute the new, unilateral effusion that developed later in the course of her illness to the underlying process responsible for the pericarditis.
The causes of pericarditis vary among populations. In a general hospitalized population, unselected cases of pericarditis are predominantly idiopathic or due to cancer (Table 2),7 and the diagnostic yield of pericardiocentesis is low in the absence of tamponade or signs of infection.9,10,11 In contrast, among patients admitted to a university tertiary-care center with new, large, symptomatic pericardial effusions, infectious causes predominate, followed by a malignant condition and radiation-induced pericarditis.8
Table 2. Causes of Pericarditis.
Noninfectious Causes of Pericarditis
Among noninfectious causes of pericarditis, tumors are most common, with lung cancer, breast cancer, and Hodgkin's disease heading the long list of possibilities. Given this patient's mediastinal lymphadenopathy, it is important to consider cancer, especially lymphoma. Factors that argue against this diagnosis are the absence of a known primary tumor, the negative results of the cytologic evaluation of the pericardial fluid, and the negative results of examination of the bronchoscopic washings and the transbronchial needle–biopsy specimen. A pericardial biopsy performed at the time of the pericardiocentesis would have been useful for ruling out cancer and certain other diagnoses.
Several noninfectious, inflammatory causes of pericarditis — radiation therapy, drugs, myocardial infarction, and metabolic factors — are not supported by the clinical history. Remaining are the collagen vascular diseases and sarcoidosis. Almost all of the collagen vascular diseases have been associated with pericarditis, but systemic lupus erythematosus and rheumatoid arthritis are the most common causes. The prevalence of collagen vascular diseases is increased among young women, such as this patient. Although pericardial disease is said to be the most common cause of symptomatic cardiac involvement in lupus, most pericardial effusions in patients with lupus are small and asymptomatic and are associated with a polyserositis and a pericardial friction rub on examination. In one study, pericardial effusion was seen in 19 percent of patients and tamponade in 2.5 percent, and it was the initial presentation in 1 percent.12 Mediastinal lymphadenopathy can occur in lupus, but nodular lung infiltrates are not consistent with a diagnosis of lupus. The low titer of antinuclear antibodies and the negative test for double-stranded DNA also make such a diagnosis unlikely. In one study of complications of rheumatoid arthritis, only 1.25 percent of patients with the disease had pericarditis13 and 0.5 percent had tamponade. Furthermore, pericarditis occurs late in the course of rheumatoid arthritis. This patient had no joint symptoms, and a test for rheumatoid factor was negative. Still's disease in adults has been associated with pericarditis, and tests for both antinuclear antibodies and rheumatoid factor are usually negative, but this patient did not have the high fever, polyarthralgias, rash, leukocytosis, abnormal aminotransferase levels, and elevated serum ferritin levels that are characteristic of that disease.14
Sarcoidosis can cause mediastinal lymphadenopathy and commonly affects the heart, causing arrhythmias, conduction abnormalities, heart failure, and pericardial disease. Pericardial involvement is rare,15 but tamponade has been reported.16 The level of angiotensin-converting enzyme was normal in the patient under discussion.
Infectious Causes of Pericarditis
This process of elimination leads me to the infectious causes of pericarditis. Most bacterial infections of the pericardium are fulminant, result in small effusions, and are rapidly fatal. They are usually caused by common bacteria such as staphylococcus, streptococcus, Escherichia coli, salmonella, and Neisseria meningitidis, and most will be evident on Gram's staining and will grow in culture. This patient started taking levofloxacin just before the pericardiocentesis was performed, and this drug could have interfered with growth in culture, although on staining, purulent fluid and organisms would probably still be evident. The clinical course, however, is not consistent with an acute bacterial process.
Viral infection is believed to account for the majority of idiopathic cases of pericarditis, the most common organisms being coxsackieviruses A and B, echovirus, adenovirus, and the human immunodeficiency virus (HIV). Pericarditis may occur during the viral syndrome or one to three weeks later. The effusions are usually small but can be associated with tamponade. Pain and a pericardial friction rub are common, and the myocardium may be involved, with unexplained arrhythmias or conduction defects.
The list of other agents that can cause pericarditis includes Rickettsia rickettsii, Chlamydia psittaci, Borrelia burgdorferi, Treponema pallidum, actinomyces, Mycoplasma pneumoniae, and nocardia. Most of these bacteria cause acute, small pericardial effusions, and infections occur after specific exposures in endemic areas. This patient had not recently visited Martha's Vineyard, an area associated with both Lyme disease and rickettsial disease. Chlamydial pericarditis can cause large effusions, but there had been no exposure to sick birds or other likely sources. Cases of fungal pericarditis (histoplasmosis and coccidioidomycosis) are usually limited to areas where these infections are endemic and occur in immunocompromised hosts.
Exposure to infectious agents in regions where they are endemic is always an important clue with respect to infection. This patient had spent a month on medical rotation in Nairobi, where tuberculosis is endemic. Five months after her return (four months before the onset of her symptoms), a tuberculin skin test was negative, but after her admission to this hospital, the induration in reaction to a skin test measured 16 mm, a finding highly predictive of infection with Mycobacterium tuberculosis. Both the possible exposure in Kenya and the new, positive reaction to a tuberculin skin test push tuberculosis to the top of the list of diagnoses.
Tuberculous Pericarditis
Although tuberculous pericarditis is an uncommon complication in countries where tuberculosis rates are low, in countries where the prevalence of tuberculosis is high it is among the most common causes of heart failure. In the Transkei region of South Africa, for example, tuberculous pericarditis is such a common cause of heart failure it is called "Transkei heart."17 The presence of mediastinal lymphadenopathy, pulmonary lesions, and a pleural effusion — all on the same side — fits the clinical pattern of tuberculous pericarditis.
As in cases of tuberculous pleuritis, the pericardial inflammation in tuberculous pericarditis appears to occur in response to relatively few organisms or to antigenic material in the pericardial space. Smears to test for acid-fast organisms in the fluid are usually negative, and cultures are positive in only about half the cases. The quinolone antibiotic used to treat presumed pneumonia may have reduced the chance of a positive culture for M. tuberculosis. A biopsy of the pericardium at the time of the pericardiocentesis might have increased the diagnostic yield.18 However, viral pericarditis was probably presumed to be the diagnosis, and a biopsy was deemed unnecessary. For most causes of pericardial effusion, the yield of pericardiocentesis with or without biopsy is low, and evidence-based indications for biopsy have not yet been established.
To confirm the diagnosis of tuberculosis and to rule out a tumor, another specimen would be needed. Although a fine-needle aspiration biopsy of a lung nodule or a thoracoscopic biopsy of the pleura might provide the necessary information for a definitive diagnosis, the most direct approach would be mediastinoscopy and biopsy of the mediastinal lymph nodes. This approach would provide sufficient tissue for all the required diagnostic procedures for tuberculosis, tumor, and the other, less likely possibilities.
Could this serious infection have been prevented? With increasing global travel, the risks of exposures to infectious agents are increasing.19 There is little that one can do to avoid tuberculosis in high-risk medical settings. Tuberculin skin testing of travelers before and two months after a visit to areas where tuberculosis is endemic is important, and conversion to a positive result should prompt treatment for latent infection.20 This patient had a negative tuberculin skin test five months after returning from Africa. Tuberculin skin-test anergy due to immunosuppression associated with other illnesses, such as infections with HIV, or drugs, such as corticosteroids, could explain this finding, but neither disease nor medication use was known to be present. The tuberculin skin test can have technical problems, ranging from injections that are subcutaneous instead of intracutaneous to tuberculin material that is defective due to mishandling. Transient specific anergy to tuberculin during an acute infection is well recognized and has been attributed to the release of cytokines that inhibit delayed hypersensitivity, but this woman was not acutely ill.
In conclusion, on the basis of the presumptive diagnosis of tuberculous pericarditis, there are two lessons from this case: first, always consider tuberculosis in persons who return from endemic areas, and second, always consider tuberculosis in persons with large pericardial effusions. In addition, consider performing a biopsy at the time of the pericardiocentesis.
Dr. Nancy Lee Harris (Pathology): Dr. Medoff, can you summarize your thoughts at the time of the diagnostic procedure?
Dr. Benjamin Medoff (Pulmonary Medicine): Because of the patient's history of pericardial disease and the abnormal chest radiograph, I was concerned about active tuberculosis. I sent sputum samples for culture and ordered a CT scan of the chest. After I received the report of the chest CT scan, I broadened the differential diagnosis to include autoimmune disease, sarcoidosis, and lymphoma. A bronchoscopy was nondiagnostic. The patient delayed further evaluation for two months for personal reasons. I then ordered another CT scan, and I referred her to Dr. John Wain, of Thoracic Surgery, for bronchoscopy and mediastinoscopy.
Clinical Diagnosis
Tuberculosis.
Dr. Edward A. Nardell's Diagnosis
Tuberculosis.
Pathological Discussion
Dr. Eugene J. Mark: The bronchoscopy showed widening of the carina, but the mucosa was normal. Biopsy specimens were taken from the mediastinal nodes. Intraoperative evaluation of frozen sections and subsequent permanent sections showed necrotizing granulomas with caseation and multinucleated giant cells (Figure 4). Stains for acid-fast bacilli were negative, but cultures of both the bronchoscopic washings and the lymph-node specimens yielded M. tuberculosis, and the organism was susceptible to all the tested agents.
Figure 4. Histologic Sections of a Mediastinal Lymph Node.
The mediastinal lymph node (Panel A) shows extensive caseating granulomas. A giant cell is seen at higher magnification (Panel B).
The tubercle bacillus reaches the pericardium by direct spread from the mediastinal lymph nodes, the likely route in this case. Miliary spread from the lungs is less common; direct spread from other nearby organs, such as the lungs, or the spine or sternum is even rarer. Tuberculous pericarditis develops in stages. The early stage is manifested by a fibrinous serosal exudate that contains lymphocytes and an associated effusion. The middle stage is cellular; granulomas with necrosis are present, and organisms are likely to be visible. In the late stage, fibrosis develops — often in association with calcification — and obliterates the pericardial cavity.
Pericardial tuberculosis must be distinguished from cardiac tuberculosis, which is much less common and more ominous.21 Cardiac tuberculosis occasionally occurs by direct extension from the pericardium, but more often it reflects miliary spread of infection from tuberculosis elsewhere in the body. Necrotizing granulomatous nodules are present, usually in the left ventricle, and can also lead to mycotic aneurysms of the heart.
In addition to the detection of organisms by acid-fast staining, a polymerase-chain-reaction assay for the IS6110 sequence22 is being evaluated and appears to be more sensitive than staining of the organisms by a considerable factor. Assays for adenosine deaminase may have diagnostic value.23 However, culture is still the gold standard for the identification of mycobacterial infection.
Dr. Harris: Dr. Medoff, can you tell us about the patient's treatment and follow-up?
Dr. Medoff: After the diagnosis was established, I initiated a four-drug regimen with isoniazid, rifampin, pyrazinamide, and ethambutol given for six months. She has now finished her drug therapy and has had no other problems.
Dr. Harris: I spoke with the patient about her negative tuberculin skin test after her return from Africa. She reported that the site had in fact become red and pruritic, which was in contrast to her reactions to prior tuberculin skin tests, but that because it was not indurated, it was officially interpreted as negative. In retrospect, it seems possible that the test was positive.
Anatomical Diagnosis
Tuberculosis.
Source Information
From the Division of Pulmonary Medicine, Cambridge Hospital, Cambridge, Mass. (E.A.N.); the Departments of Cardiology (D.F.), Radiology (J.O.S.), and Pathology (E.J.M.), Massachusetts General Hospital; and the Departments of Medicine (E.A.N.), Cardiology (D.F.), Radiology (J.O.S.), and Pathology (E.J.M.), Harvard Medical School.
References
Gillam LD, Guyer DE, Gibson TC, King ME, Marshall JE, Weyman AE. Hydrodynamic compression of the right atrium: a new echocardiographic sign of cardiac tamponade. Circulation 1983;68:294-301.
Armstrong WF, Schilt BF, Helper DJ, Dillon JC, Feigenbaum H. Diastolic collapse of the right ventricle with cardiac tamponade: an echocardiographic study. Circulation 1982;65:1491-1496.
Matsuo H, Matsumoto M, Hamanaka Y, et al. Rotational excursion of heart in massive pericardial effusion studied by phased-array echocardiography. Br Heart J 1979;41:513-521.
Feigenbaum H, Zaky A, Grabhorn LL. Cardiac motion in patients with pericardial effusion: a study using reflected ultrasound. Circulation 1966;34:611-619.
Gabor GE, Winsberg F, Bloom HS. Electrical and mechanical alternation in pericardial effusion. Chest 1971;59:341-344.
Appleton CP, Hatle LK, Popp RL. Cardiac tamponade and pericardial effusion: respiratory variation in transvalvular flow velocities studied by Doppler echocardiography. J Am Coll Cardiol 1988;11:1020-1030.
Sagrista-Sauleda J, Merce J, Permanyer-Miralda G, Soler-Soler J. Clinical clues to the causes of large pericardial effusions. Am J Med 2000;109:95-101.
Corey GR, Campbell PT, Van Trigt P, et al. Etiology of large pericardial effusions. Am J Med 1993;95:209-213.
Merce J, Sagrista-Sauleda J, Permanyer-Miralda G, Soler-Soler J. Should pericardial drainage be performed routinely in patients who have a large pericardial effusion without tamponade? Am J Med 1998;105:106-109.
Permanyer-Miralda G, Sagrista-Sauleda J, Soler-Soler J. Primary acute pericardial disease: a prospective series of 231 consecutive patients. Am J Cardiol 1985;56:623-630.
Zayas R, Anguita M, Torres F, et al. Incidence of specific etiology and role of methods for specific etiologic diagnosis of primary acute pericarditis. Am J Cardiol 1995;75:378-382.
Kahl LE. The spectrum of pericardial tamponade in systemic lupus erythematosus: report of ten patients. Arthritis Rheum 1992;35:1343-1349.
Escalante A, Kaufman RL, Quismorio FP Jr, Beardmore TD. Cardiac compression in rheumatoid pericarditis. Semin Arthritis Rheum 1990;20:148-163.
Moder KG, Miller TD, Allen GL. Cardiac tamponade: an unusual feature of adult onset Still's disease. J Rheumatol 1995;22:180-182.
Pesola G, Teirstein AS, Goldman M. Sarcoidois presenting with pericardial effusion. Sarcoidosis 1987;4:42-44.
Zelcer AA, LeJemtel TH, Jones J, Stahl J. Pericardial tamponade in sarcoidosis. Can J Cardiol 1987;3:12-13.
Strang JI. Tuberculous pericarditis in Transkei. Clin Cardiol 1984;7:667-670.
Seferovic PM, Ristic AD, Maksimovic R, Tatic V, Ostojic M, Kanjuh V. Diagnostic value of pericardial biopsy: improvement with extensive sampling enabled by pericardioscopy. Circulation 2003;107:978-983.
Ko G, Thompson KM, Nardell EA. Estimation of tuberculosis risk on a commercial airliner. Risk Anal 2004;24:379-388.
Jasmer RM, Nahid P, Hopewell PC. Latent tuberculosis infection. N Engl J Med 2002;347:1860-1866.
Rose AG. Cardiac tuberculosis: a study of 19 patients. Arch Pathol Lab Med 1987;111:422-426.
Park do Y, Kim JY, Choi KT, et al. Comparison of polymerase chain reaction with histopathologic features for diagnosis of tuberculosis in formalin-fixed, paraffin-embedded histologic specimens. Arch Pathol Lab Med 2003;127:326-330.
Koh KK, Kim EJ, Cho CH, et al. Adenosine deaminase and carcinoembryonic antigen in pericardial effusion diagnosis, especially in suspected tuberculous pericarditis. Circulation 1994;89:2728-2735.
Related Letters:
Case 22-2004: A 30-Year-Old Woman with a Pericardial Effusion
Cook V. J., FitzGerald J. M., Nardell E. A.(Edward A. Nardell, M.D., )
A 30-year-old woman was admitted to the hospital because of fever, cough, shortness of breath, and a pericardial effusion.
Approximately one month earlier, the woman had noticed an increased heart rate and fluttering in her neck when she bent over; both were relieved when she stood upright. During these episodes, her heart rate was 100 to 120 beats per minute. One week later, she noticed shortness of breath after walking up one flight of stairs. Since the age of 19 years, she had had intermittent atrioventricular nodal reentrant tachycardia, with heart rates of up to 180 beats per minute, and she had had approximately three episodes a year that resolved with the Valsalva maneuver. She had also had three episodes that had required visits to the emergency department and treatment with verapamil and adenosine. Because her current symptoms reminded her of these episodes, she consulted her cardiologist and wore a Holter monitor for a 24-hour period. No abnormal rhythms were revealed.
One week later, the woman noticed increased fatigue, and several days later fever (temperature up to 38.0°C) developed, with a dry cough and increased shortness of breath. One episode of substernal chest pain occurred; it lasted 30 minutes and was relieved with acetaminophen. She visited her primary care doctor one week after this episode and was found to have anemia. The thyrotropin level was 0.65 μU per milliliter and the thyroxine level was 1.14 mg per deciliter (14.67 nmol per liter). Four days later, she went to the emergency department of another hospital because of persistent cough and fever (temperature up to 38.1°C). A chest radiograph showed a right pleural effusion and a possible infiltrate. A 10-day course of levofloxacin was begun. Two days later, the formal report of the chest radiograph was issued and noted an enlarged cardiac silhouette. An echocardiogram showed a pericardial effusion with right ventricular diastolic collapse — findings consistent with tamponade. She came to this hospital and was admitted.
The patient was a physician in training. She did not smoke, drink alcohol, or use illicit drugs. A sister had Graves' disease, and her mother had hyperthyroidism. Eleven months before admission, she had traveled to Nairobi, Kenya, for a month-long medical rotation. Since then, she had not traveled outside of the United States or to Martha's Vineyard, Massachusetts. A tuberculin skin test performed five months after she had returned from Africa was negative. She had had multiple exposures to children with viral illnesses in the course of her work. There was no history of rash or joint swelling.
On physical examination, she appeared well. The blood pressure was 130/83 mm Hg, the heart rate 98 beats per minute, the respiratory rate 20 breaths per minute, and the temperature 36.1°C. There was a pulsus paradoxus of 8 mm Hg (140 mm Hg decreasing to 132 mm Hg). The hemoglobin oxygen saturation was determined by pulse oximetry to be 100 percent while the patient was breathing ambient air. An examination of the woman's head, eyes, ears, nose, and throat showed no abnormalities. An examination of her neck showed a jugular venous pressure of 12 cm of water with bilateral carotid bruits and no lymphadenopathy. The breath sounds were clear. An examination of the heart showed no murmurs but did reveal a possible pericardial friction rub at the left lower sternal border. Abdominal and neurologic examinations showed no abnormalities, and the arms and legs were warm, with symmetric pulses and no edema. Laboratory test results are shown in Table 1.
Table 1. Laboratory Values on Admission.
A chest radiograph disclosed small, bilateral pleural effusions and an enlarged cardiac silhouette (Figure 1A). An electrocardiogram showed a normal sinus rhythm at 98 beats per minute; QRS intervals of 0.138, 0.075, and 0.405 second; a normal axis; diffuse T-wave inversions; and a downsloping J point in the V2 lead. An echocardiogram showed a large, circumferential pericardial effusion with right atrial inversion (Figure 2), diastolic inversion of the right ventricular wall, and respiratory variation in the intracardiac flows — findings consistent with increased intrapericardial pressure.
Figure 1. Chest Radiographs.
On a chest radiograph obtained on admission (Panel A), there are small bilateral pleural effusions and an enlarged cardiac silhouette. A chest radiograph obtained two months after discharge (Panel B) shows resolution of the pleural effusions and the development of new, small nodular opacities in the right upper lobe.
Figure 2. Echocardiograms Obtained on Admission.
In an apical four-chamber view (Panel A and Video Clip 1 of the Supplementary Appendix), there is a large pericardial effusion (bar) and inversion of the right atrium (arrow), caused by elevated pericardial pressure, in late diastole and early systole. A parasternal short-axis view (Panel B and Video Clip 2 of the Supplementary Appendix) shows that the right ventricular outflow tract is compressed in diastole because of the elevated pericardial pressure (arrow). The bar shows the pericardial effusion. A Doppler study of flow across the mitral valve (Panel C) shows respiratory variation in the mitral inflow E wave, a sign of tamponade. The arrowheads denote expiration, the asterisks inspiration, E diastolic mitral E wave, and A diastolic mitral A wave. The scale to the right indicates flow velocity, measured in centimeters per second.
Pericardiocentesis was performed, and 370 ml of bloody fluid was removed. The pericardial pressure decreased from 7 to 0 mm Hg. A catheter drained decreasing amounts of fluid during the next 48 hours and was removed. A cytologic examination of the pericardial fluid showed blood and no malignant cells. A Gram's stain showed red cells but no neutrophils or organisms, and a culture showed no growth of bacteria. A smear and a culture were negative for acid-fast bacilli and mycobacteria. An echocardiogram obtained on the second hospital day showed no reaccumulation of the pericardial effusion. The patient was discharged on the third day. A tuberculin skin test was performed during the hospitalization, and 16 mm of induration developed two days after discharge.
Echocardiograms obtained one and two months after discharge showed a minimal pericardial effusion, diffuse fibrin deposition on the visceral pericardium, and no evidence of tamponade. Two months after discharge, the patient was seen in the pulmonary clinic for evaluation of the positive tuberculin skin test. On physical examination, the lungs were normal on auscultation. Smears and cultures of three specimens of sputum were negative for acid-fast bacilli and mycobacteria. A radiograph of the chest showed resolution of the pleural effusions but two new, small nodular opacities in the right upper lung zone (Figure 1B). Computed tomographic (CT) scanning of the chest, performed without the intravenous administration of contrast material (Figure 3A and Figure 3B), showed a cluster of small nodules, 3 to 5 mm in diameter, in the right upper lobe, as well as multifocal mediastinal lymphadenopathy, a confluent mass of subcarinal lymph nodes, and an enlarged right hilar lymph node. There was minimal pericardial thickening and slight thickening of the right posterior pleura.
Figure 3. CT Scans of the Chest.
A CT scan of the chest obtained without the use of intravenous contrast material two months after discharge shows mediastinal lymphadenopathy (Panel A) and a cluster of small nodules in the right upper lobe (Panel B). A CT scan of the chest obtained with the use of intravenous contrast material four months after discharge reveals a decrease in the size of the mediastinal lymph nodes and in the size of the nodules in the right upper lobe (Panel C) and a new right pleural effusion (Panel D).
One week later, bronchoscopy was performed, with transbronchial fine-needle aspiration of the subcarinal lymph nodes, transbronchial biopsy of the right upper lobe, and bronchoalveolar lavage of the right upper lobe. Cytologic examination of the specimen obtained by fine-needle aspiration disclosed no malignant cells, and flow cytometry showed no evidence of lymphoma. Pathological examination of the specimen from the transbronchial lung biopsy showed bronchial epithelium with no evidence of malignant cells or granulomas. Cultures of the bronchial-lavage fluid yielded normal flora, no mycobacteria, and no fungi. The serum angiotensin-converting–enzyme level was 38 U per liter (normal range, 7.0 to 46.0).
Two months later, another CT scan of the chest (Figure 3C and Figure 3D) showed a decrease in the size of the anterior mediastinal lymph nodes, no change in the size of the lymph nodes elsewhere, and nodular opacities in the right upper lobe, one of which had decreased in size. A new right pleural effusion had developed.
A diagnostic procedure was performed.
Differential Diagnosis
Dr. Edward A. Nardell: May we review the echocardiograms and the imaging studies of the chest?
Dr. Dali Fan: Normally, the space between the parietal and visceral pericardium is not visualized on an echocardiogram. In this patient, there is an echolucent space behind the posterior wall and anterior to the right ventricular outflow tract, indicating the presence of a pericardial effusion (Figure 2A). The effusion completely surrounds the heart, with a separation of at least 1 cm between the parietal and visceral pericardium, indicating that the fluid collection has an estimated volume of more than 500 ml.
In tamponade, cardiac filling is compromised by elevated pericardial pressure. The right atrium, the chamber with the lowest pressure, is the first to be affected. In the apical four-chamber view (Figure 2A and Video Clip 1 of the Supplementary Appendix, available with the full text of this article at www.nejm.org), the contour of the right atrium is inverted in late diastole and early systole, when the internal pressure is lowest. This is a sensitive but relatively nonspecific sign of tamponade, since pulmonary diseases can cause a similar phenomenon.1 The right ventricle and outflow tract are also vulnerable to elevated pericardial pressure. The parasternal short-axis M-mode echocardiogram (Figure 2B and Video Clip 2 of the Supplementary Appendix) shows that the right ventricular outflow wall is deflected posteriorly after the mitral valve opens because of the elevated pericardial pressure. The sensitivity of this finding is lower, but the specificity is higher than that of right atrial collapse.2 Video Clip 3 of the Supplementary Appendix shows that the heart swings dramatically within the large pericardial effusion. This movement occurs in both the anteroposterior and mediolateral directions and may include an exaggerated twisting of the heart around its long axis.3,4,5
The normal increase in flow across the mitral valve during expiration and the decrease during inspiration are exaggerated in tamponade.6 On the Doppler study of mitral inflow (Figure 2C), the variation of mitral valve filling with respiration (the peak E wave) is greater than 25 percent. In summary, anatomical evaluation shows that this patient has a large pericardial effusion, and a physiological evaluation reveals that she has changes consistent with increased pericardial pressure and tamponade.
Dr. Jo-Anne O. Shepard: The chest radiograph obtained at the time of admission (Figure 1A) reveals small bilateral pleural effusions and an enlarged cardiomediastinal silhouette. As the chest radiograph obtained two months after the patient's discharge (Figure 1B) shows, the pleural effusions subsequently resolved and the cardiac silhouette decreased in size. There are new, faint, irregular nodular opacities in the right upper lobe. A chest CT scan also obtained at that time (Figure 3A) shows lymphadenopathy in the anterior mediastinum in the paratracheal region and in the subcarinal and right hilar regions. There is minimal right lateral pleural thickening. There is a cluster of nodules, 3 to 5 mm in diameter, in the periphery of the right upper lobe (Figure 3B). A chest CT scan obtained four months after discharge shows a decrease in the mediastinal lymphadenopathy, and the nodular opacities in the right upper lobe have slightly decreased in size (Figure 3C). The subcarinal and right hilar lymphadenopathy remains stable. A new right pleural effusion is present (Figure 3D), and the nodular opacities in the right upper lobe have slightly decreased in size.
Dr. Nardell: This 30-year-old woman came to the hospital with pleural and pericardial effusions, tamponade, fever, pulmonary nodules, and mediastinal lymphadenopathy. Of these clinical features, the most immediately life-threatening is pericardial effusion with tamponade, and I will build my discussion around this relatively uncommon finding in a young, otherwise healthy woman.
The pericardial effusion along with the progressive tamponade explains many of the symptoms that culminated in the patient's hospitalization. As the pericardial sac fills with fluid, diastolic filling is increasingly compromised, especially in the low-pressure right ventricle. The echocardiogram also showed right ventricular diastolic collapse, a finding that is said to correlate with a reduction in cardiac output of 15 to 25 percent. With tamponade, stroke volume decreases, and cardiac output is maintained by an increased heart rate. Central venous pressure increases and aids diastolic filling. Bending over could have transiently reduced her venous return and cardiac filling by reducing the hydrostatic pressure in the superior vena cava or by compressing the inferior vena cava. Either one of these mechanisms could result in tachycardia and reduced cardiac output. Her blood pressure was normal, which is probably a result of physiologic compensation. These findings suggest a chronic, rather than an acute, process.
The evidence of pericardial tamponade quickly reversed when 370 ml of fluid was removed. The finding of so much fluid is in itself helpful in establishing the cause. Pericardial fluid can accumulate in large amounts only if it does so slowly. Traumatic effusions, effusions due to acute bacterial infections, and other acute effusions compromise cardiac function and can rapidly lead to death because there is no time for the pericardium to stretch to accommodate the fluid. Pericardial effusions due to viral infections, lupus, and rheumatoid arthritis also tend to be small but can occur rapidly and lead to tamponade.
This patient also had bilateral pleural effusions. These could be part of the underlying illness, or they could be secondary to her effective heart failure due to tamponade. The most common causes of bilateral pleural effusions are congestive heart failure and renal failure. Systemic inflammatory illnesses, such as lupus and widespread metastatic cancer, can also cause bilateral pleural effusions, but they are less common. At this point, I will assume that the pleural effusions were due to heart failure. However, I will attribute the new, unilateral effusion that developed later in the course of her illness to the underlying process responsible for the pericarditis.
The causes of pericarditis vary among populations. In a general hospitalized population, unselected cases of pericarditis are predominantly idiopathic or due to cancer (Table 2),7 and the diagnostic yield of pericardiocentesis is low in the absence of tamponade or signs of infection.9,10,11 In contrast, among patients admitted to a university tertiary-care center with new, large, symptomatic pericardial effusions, infectious causes predominate, followed by a malignant condition and radiation-induced pericarditis.8
Table 2. Causes of Pericarditis.
Noninfectious Causes of Pericarditis
Among noninfectious causes of pericarditis, tumors are most common, with lung cancer, breast cancer, and Hodgkin's disease heading the long list of possibilities. Given this patient's mediastinal lymphadenopathy, it is important to consider cancer, especially lymphoma. Factors that argue against this diagnosis are the absence of a known primary tumor, the negative results of the cytologic evaluation of the pericardial fluid, and the negative results of examination of the bronchoscopic washings and the transbronchial needle–biopsy specimen. A pericardial biopsy performed at the time of the pericardiocentesis would have been useful for ruling out cancer and certain other diagnoses.
Several noninfectious, inflammatory causes of pericarditis — radiation therapy, drugs, myocardial infarction, and metabolic factors — are not supported by the clinical history. Remaining are the collagen vascular diseases and sarcoidosis. Almost all of the collagen vascular diseases have been associated with pericarditis, but systemic lupus erythematosus and rheumatoid arthritis are the most common causes. The prevalence of collagen vascular diseases is increased among young women, such as this patient. Although pericardial disease is said to be the most common cause of symptomatic cardiac involvement in lupus, most pericardial effusions in patients with lupus are small and asymptomatic and are associated with a polyserositis and a pericardial friction rub on examination. In one study, pericardial effusion was seen in 19 percent of patients and tamponade in 2.5 percent, and it was the initial presentation in 1 percent.12 Mediastinal lymphadenopathy can occur in lupus, but nodular lung infiltrates are not consistent with a diagnosis of lupus. The low titer of antinuclear antibodies and the negative test for double-stranded DNA also make such a diagnosis unlikely. In one study of complications of rheumatoid arthritis, only 1.25 percent of patients with the disease had pericarditis13 and 0.5 percent had tamponade. Furthermore, pericarditis occurs late in the course of rheumatoid arthritis. This patient had no joint symptoms, and a test for rheumatoid factor was negative. Still's disease in adults has been associated with pericarditis, and tests for both antinuclear antibodies and rheumatoid factor are usually negative, but this patient did not have the high fever, polyarthralgias, rash, leukocytosis, abnormal aminotransferase levels, and elevated serum ferritin levels that are characteristic of that disease.14
Sarcoidosis can cause mediastinal lymphadenopathy and commonly affects the heart, causing arrhythmias, conduction abnormalities, heart failure, and pericardial disease. Pericardial involvement is rare,15 but tamponade has been reported.16 The level of angiotensin-converting enzyme was normal in the patient under discussion.
Infectious Causes of Pericarditis
This process of elimination leads me to the infectious causes of pericarditis. Most bacterial infections of the pericardium are fulminant, result in small effusions, and are rapidly fatal. They are usually caused by common bacteria such as staphylococcus, streptococcus, Escherichia coli, salmonella, and Neisseria meningitidis, and most will be evident on Gram's staining and will grow in culture. This patient started taking levofloxacin just before the pericardiocentesis was performed, and this drug could have interfered with growth in culture, although on staining, purulent fluid and organisms would probably still be evident. The clinical course, however, is not consistent with an acute bacterial process.
Viral infection is believed to account for the majority of idiopathic cases of pericarditis, the most common organisms being coxsackieviruses A and B, echovirus, adenovirus, and the human immunodeficiency virus (HIV). Pericarditis may occur during the viral syndrome or one to three weeks later. The effusions are usually small but can be associated with tamponade. Pain and a pericardial friction rub are common, and the myocardium may be involved, with unexplained arrhythmias or conduction defects.
The list of other agents that can cause pericarditis includes Rickettsia rickettsii, Chlamydia psittaci, Borrelia burgdorferi, Treponema pallidum, actinomyces, Mycoplasma pneumoniae, and nocardia. Most of these bacteria cause acute, small pericardial effusions, and infections occur after specific exposures in endemic areas. This patient had not recently visited Martha's Vineyard, an area associated with both Lyme disease and rickettsial disease. Chlamydial pericarditis can cause large effusions, but there had been no exposure to sick birds or other likely sources. Cases of fungal pericarditis (histoplasmosis and coccidioidomycosis) are usually limited to areas where these infections are endemic and occur in immunocompromised hosts.
Exposure to infectious agents in regions where they are endemic is always an important clue with respect to infection. This patient had spent a month on medical rotation in Nairobi, where tuberculosis is endemic. Five months after her return (four months before the onset of her symptoms), a tuberculin skin test was negative, but after her admission to this hospital, the induration in reaction to a skin test measured 16 mm, a finding highly predictive of infection with Mycobacterium tuberculosis. Both the possible exposure in Kenya and the new, positive reaction to a tuberculin skin test push tuberculosis to the top of the list of diagnoses.
Tuberculous Pericarditis
Although tuberculous pericarditis is an uncommon complication in countries where tuberculosis rates are low, in countries where the prevalence of tuberculosis is high it is among the most common causes of heart failure. In the Transkei region of South Africa, for example, tuberculous pericarditis is such a common cause of heart failure it is called "Transkei heart."17 The presence of mediastinal lymphadenopathy, pulmonary lesions, and a pleural effusion — all on the same side — fits the clinical pattern of tuberculous pericarditis.
As in cases of tuberculous pleuritis, the pericardial inflammation in tuberculous pericarditis appears to occur in response to relatively few organisms or to antigenic material in the pericardial space. Smears to test for acid-fast organisms in the fluid are usually negative, and cultures are positive in only about half the cases. The quinolone antibiotic used to treat presumed pneumonia may have reduced the chance of a positive culture for M. tuberculosis. A biopsy of the pericardium at the time of the pericardiocentesis might have increased the diagnostic yield.18 However, viral pericarditis was probably presumed to be the diagnosis, and a biopsy was deemed unnecessary. For most causes of pericardial effusion, the yield of pericardiocentesis with or without biopsy is low, and evidence-based indications for biopsy have not yet been established.
To confirm the diagnosis of tuberculosis and to rule out a tumor, another specimen would be needed. Although a fine-needle aspiration biopsy of a lung nodule or a thoracoscopic biopsy of the pleura might provide the necessary information for a definitive diagnosis, the most direct approach would be mediastinoscopy and biopsy of the mediastinal lymph nodes. This approach would provide sufficient tissue for all the required diagnostic procedures for tuberculosis, tumor, and the other, less likely possibilities.
Could this serious infection have been prevented? With increasing global travel, the risks of exposures to infectious agents are increasing.19 There is little that one can do to avoid tuberculosis in high-risk medical settings. Tuberculin skin testing of travelers before and two months after a visit to areas where tuberculosis is endemic is important, and conversion to a positive result should prompt treatment for latent infection.20 This patient had a negative tuberculin skin test five months after returning from Africa. Tuberculin skin-test anergy due to immunosuppression associated with other illnesses, such as infections with HIV, or drugs, such as corticosteroids, could explain this finding, but neither disease nor medication use was known to be present. The tuberculin skin test can have technical problems, ranging from injections that are subcutaneous instead of intracutaneous to tuberculin material that is defective due to mishandling. Transient specific anergy to tuberculin during an acute infection is well recognized and has been attributed to the release of cytokines that inhibit delayed hypersensitivity, but this woman was not acutely ill.
In conclusion, on the basis of the presumptive diagnosis of tuberculous pericarditis, there are two lessons from this case: first, always consider tuberculosis in persons who return from endemic areas, and second, always consider tuberculosis in persons with large pericardial effusions. In addition, consider performing a biopsy at the time of the pericardiocentesis.
Dr. Nancy Lee Harris (Pathology): Dr. Medoff, can you summarize your thoughts at the time of the diagnostic procedure?
Dr. Benjamin Medoff (Pulmonary Medicine): Because of the patient's history of pericardial disease and the abnormal chest radiograph, I was concerned about active tuberculosis. I sent sputum samples for culture and ordered a CT scan of the chest. After I received the report of the chest CT scan, I broadened the differential diagnosis to include autoimmune disease, sarcoidosis, and lymphoma. A bronchoscopy was nondiagnostic. The patient delayed further evaluation for two months for personal reasons. I then ordered another CT scan, and I referred her to Dr. John Wain, of Thoracic Surgery, for bronchoscopy and mediastinoscopy.
Clinical Diagnosis
Tuberculosis.
Dr. Edward A. Nardell's Diagnosis
Tuberculosis.
Pathological Discussion
Dr. Eugene J. Mark: The bronchoscopy showed widening of the carina, but the mucosa was normal. Biopsy specimens were taken from the mediastinal nodes. Intraoperative evaluation of frozen sections and subsequent permanent sections showed necrotizing granulomas with caseation and multinucleated giant cells (Figure 4). Stains for acid-fast bacilli were negative, but cultures of both the bronchoscopic washings and the lymph-node specimens yielded M. tuberculosis, and the organism was susceptible to all the tested agents.
Figure 4. Histologic Sections of a Mediastinal Lymph Node.
The mediastinal lymph node (Panel A) shows extensive caseating granulomas. A giant cell is seen at higher magnification (Panel B).
The tubercle bacillus reaches the pericardium by direct spread from the mediastinal lymph nodes, the likely route in this case. Miliary spread from the lungs is less common; direct spread from other nearby organs, such as the lungs, or the spine or sternum is even rarer. Tuberculous pericarditis develops in stages. The early stage is manifested by a fibrinous serosal exudate that contains lymphocytes and an associated effusion. The middle stage is cellular; granulomas with necrosis are present, and organisms are likely to be visible. In the late stage, fibrosis develops — often in association with calcification — and obliterates the pericardial cavity.
Pericardial tuberculosis must be distinguished from cardiac tuberculosis, which is much less common and more ominous.21 Cardiac tuberculosis occasionally occurs by direct extension from the pericardium, but more often it reflects miliary spread of infection from tuberculosis elsewhere in the body. Necrotizing granulomatous nodules are present, usually in the left ventricle, and can also lead to mycotic aneurysms of the heart.
In addition to the detection of organisms by acid-fast staining, a polymerase-chain-reaction assay for the IS6110 sequence22 is being evaluated and appears to be more sensitive than staining of the organisms by a considerable factor. Assays for adenosine deaminase may have diagnostic value.23 However, culture is still the gold standard for the identification of mycobacterial infection.
Dr. Harris: Dr. Medoff, can you tell us about the patient's treatment and follow-up?
Dr. Medoff: After the diagnosis was established, I initiated a four-drug regimen with isoniazid, rifampin, pyrazinamide, and ethambutol given for six months. She has now finished her drug therapy and has had no other problems.
Dr. Harris: I spoke with the patient about her negative tuberculin skin test after her return from Africa. She reported that the site had in fact become red and pruritic, which was in contrast to her reactions to prior tuberculin skin tests, but that because it was not indurated, it was officially interpreted as negative. In retrospect, it seems possible that the test was positive.
Anatomical Diagnosis
Tuberculosis.
Source Information
From the Division of Pulmonary Medicine, Cambridge Hospital, Cambridge, Mass. (E.A.N.); the Departments of Cardiology (D.F.), Radiology (J.O.S.), and Pathology (E.J.M.), Massachusetts General Hospital; and the Departments of Medicine (E.A.N.), Cardiology (D.F.), Radiology (J.O.S.), and Pathology (E.J.M.), Harvard Medical School.
References
Gillam LD, Guyer DE, Gibson TC, King ME, Marshall JE, Weyman AE. Hydrodynamic compression of the right atrium: a new echocardiographic sign of cardiac tamponade. Circulation 1983;68:294-301.
Armstrong WF, Schilt BF, Helper DJ, Dillon JC, Feigenbaum H. Diastolic collapse of the right ventricle with cardiac tamponade: an echocardiographic study. Circulation 1982;65:1491-1496.
Matsuo H, Matsumoto M, Hamanaka Y, et al. Rotational excursion of heart in massive pericardial effusion studied by phased-array echocardiography. Br Heart J 1979;41:513-521.
Feigenbaum H, Zaky A, Grabhorn LL. Cardiac motion in patients with pericardial effusion: a study using reflected ultrasound. Circulation 1966;34:611-619.
Gabor GE, Winsberg F, Bloom HS. Electrical and mechanical alternation in pericardial effusion. Chest 1971;59:341-344.
Appleton CP, Hatle LK, Popp RL. Cardiac tamponade and pericardial effusion: respiratory variation in transvalvular flow velocities studied by Doppler echocardiography. J Am Coll Cardiol 1988;11:1020-1030.
Sagrista-Sauleda J, Merce J, Permanyer-Miralda G, Soler-Soler J. Clinical clues to the causes of large pericardial effusions. Am J Med 2000;109:95-101.
Corey GR, Campbell PT, Van Trigt P, et al. Etiology of large pericardial effusions. Am J Med 1993;95:209-213.
Merce J, Sagrista-Sauleda J, Permanyer-Miralda G, Soler-Soler J. Should pericardial drainage be performed routinely in patients who have a large pericardial effusion without tamponade? Am J Med 1998;105:106-109.
Permanyer-Miralda G, Sagrista-Sauleda J, Soler-Soler J. Primary acute pericardial disease: a prospective series of 231 consecutive patients. Am J Cardiol 1985;56:623-630.
Zayas R, Anguita M, Torres F, et al. Incidence of specific etiology and role of methods for specific etiologic diagnosis of primary acute pericarditis. Am J Cardiol 1995;75:378-382.
Kahl LE. The spectrum of pericardial tamponade in systemic lupus erythematosus: report of ten patients. Arthritis Rheum 1992;35:1343-1349.
Escalante A, Kaufman RL, Quismorio FP Jr, Beardmore TD. Cardiac compression in rheumatoid pericarditis. Semin Arthritis Rheum 1990;20:148-163.
Moder KG, Miller TD, Allen GL. Cardiac tamponade: an unusual feature of adult onset Still's disease. J Rheumatol 1995;22:180-182.
Pesola G, Teirstein AS, Goldman M. Sarcoidois presenting with pericardial effusion. Sarcoidosis 1987;4:42-44.
Zelcer AA, LeJemtel TH, Jones J, Stahl J. Pericardial tamponade in sarcoidosis. Can J Cardiol 1987;3:12-13.
Strang JI. Tuberculous pericarditis in Transkei. Clin Cardiol 1984;7:667-670.
Seferovic PM, Ristic AD, Maksimovic R, Tatic V, Ostojic M, Kanjuh V. Diagnostic value of pericardial biopsy: improvement with extensive sampling enabled by pericardioscopy. Circulation 2003;107:978-983.
Ko G, Thompson KM, Nardell EA. Estimation of tuberculosis risk on a commercial airliner. Risk Anal 2004;24:379-388.
Jasmer RM, Nahid P, Hopewell PC. Latent tuberculosis infection. N Engl J Med 2002;347:1860-1866.
Rose AG. Cardiac tuberculosis: a study of 19 patients. Arch Pathol Lab Med 1987;111:422-426.
Park do Y, Kim JY, Choi KT, et al. Comparison of polymerase chain reaction with histopathologic features for diagnosis of tuberculosis in formalin-fixed, paraffin-embedded histologic specimens. Arch Pathol Lab Med 2003;127:326-330.
Koh KK, Kim EJ, Cho CH, et al. Adenosine deaminase and carcinoembryonic antigen in pericardial effusion diagnosis, especially in suspected tuberculous pericarditis. Circulation 1994;89:2728-2735.
Related Letters:
Case 22-2004: A 30-Year-Old Woman with a Pericardial Effusion
Cook V. J., FitzGerald J. M., Nardell E. A.(Edward A. Nardell, M.D., )