A Sailor's Heartbreak
http://www.100md.com
《新英格兰医药杂志》
A 24-year-old sailor presented to the emergency department describing a seven-day history of shortness of breath, occasional substernal chest pain, increasing abdominal girth, and nausea and vomiting. He reported that his legs had swelled and he had gained 14 kg (31 lb) during the previous three weeks, when he had been sailing from the Caribbean to Nantucket Island. He reported that before the onset of these symptoms, he had been in his usual state of good health. He used no medications. He had had no known contact with sick persons and no known exposure to tuberculosis, and he had no risk factors for infection with the human immunodeficiency virus.
A Pulmonologist: This young, previously healthy man describes dyspnea in the setting of substantial weight gain. The degree of weight gain and the time course suggest that this is a subacute process. The differential diagnosis is broad and includes cardiac, pulmonary, hepatic, and renal dysfunction.
On physical examination, the patient's vital signs were as follows: blood pressure, 127/82 mm Hg; heart rate, 102 beats per minute; temperature, 99.3°F (37.4°C); and respiration, 26 per minute. The oxygen saturation was 96 percent while the patient was receiving 100 percent supplemental oxygen. No pulsus paradoxus was detected. The patient appeared to be in mild distress but spoke in complete sentences. His sclerae were anicteric. The jugular venous pressure was not elevated. The lungs had decreased breath sounds and dullness to percussion approximately halfway up. The first and second heart sounds were normal, with no rub, murmurs, or gallops. Abdominal examination revealed mild tenderness in the right upper quadrant and a fluid wave without hepatosplenomegaly. The patient had 3+ edema bilaterally and no joint abnormalities, rashes, jaundice, or lymphadenopathy. His mental status was normal. Chest radiography revealed an enlarged cardiac silhouette, with moderate bilateral pleural effusions with no infiltrate or interstitial edema (Figure 1).
Figure 1. Anteroposterior Chest Radiograph Showing an Enlarged Cardiac Silhouette with Moderate Bilateral Pleural Effusions and No Infiltrate or Interstitial Edema.
A Pulmonologist: The patient has evidence of fluid in the pleural and peritoneal spaces and, according to the results of chest radiography, a pericardial effusion as well. Thoracentesis is warranted, with examination of the pleural fluid to determine whether it is transudative, owing to increased hydrostatic or reduced oncotic pressures, or exudative, which would suggest an inflammatory or infectious process or cancer. Cardiac disorders causing increased hydrostatic pressures and hepatic or renal problems causing reduced serum protein levels may lead to diffuse fluid accumulation in pleural, pericardial, peritoneal, and interstitial spaces. Inflammatory processes cause interstitial edema infrequently; although multiple serosal surfaces can be involved, localized fluid collections are more common.
Although a patient with viral infection can present with diffuse serositis, the time course of this man's illness and the absence of any myalgias or constitutional symptoms are atypical. Bacterial causes of diffuse serositis are uncommon, but occasionally, pneumonia with a parapneumonic effusion can occur with a pericardial effusion; ascites, however, would be unlikely. Tuberculosis may develop insidiously, with pleural, pericardial, and peritoneal effusions. Another common cause of diffuse serositis is collagen vascular disease, including Still's disease, systemic lupus erythematosus, and mixed connective-tissue disease. Familial Mediterranean fever would need to be considered if the patient had a history of recurrent episodes of serositis and fever.
The presence of ascites raises the possibility of cirrhosis with elevated portal pressures. A normal jugular venous pulse argues against cardiac dysfunction leading to increased hydrostatic pressures. I would be interested in an analysis of the pleural fluid, a complete blood count, a test for antinuclear antibodies, and urinalysis.
The serum electrolytes were within normal limits. The glucose level was 109 mg per deciliter (6.0 mmol per liter); blood urea nitrogen, 20 mg per deciliter (7.1 mmol per liter); creatinine, 1.9 mg per deciliter (168.0 μmol per liter); aspartate aminotransferase, 855 U per liter; alanine aminotransferase, 822 U per liter; total bilirubin, 1.2 mg per deciliter (20.5 μmol per liter); alkaline phosphatase, 99 U per liter; albumin, 3.0 g per deciliter; and lactate dehydrogenase, 52 U per liter. The white-cell count was 11,100 per cubic millimeter, the hemoglobin level was 13.1 g per deciliter, and the platelet count was 285,000 per cubic millimeter. The international normalized ratio (INR) of prothrombin time was 1.5. The antinuclear-antibody test was negative. Urine dipstick analysis and microscopy revealed no abnormalities.
Analysis of the pleural fluid showed that the pH was 8.0, the glucose level 83 mg per deciliter (4.6 mmol per liter), the lactate dehydrogenase level 98 U per liter, and the protein level 2.9 g per deciliter. The white-cell count was 220 per cubic millimeter, with 8 percent neutrophils, 68 percent lymphocytes, 13 percent macrophages, and 11 percent mesothelial cells. Bacterial cultures, cytologic tests, and a smear for acid-fast bacilli were negative. A transthoracic echocardiogram showed preserved right and left ventricular function with a moderate-to-large pericardial effusion but no echocardiographic evidence of cardiac tamponade.
A Pulmonologist: The relatively normal serum albumin level rules out reduced oncotic pressure as a cause of the fluid collections. A connective-tissue disorder is unlikely given the negative antinuclear-antibody test and the absence of other characteristic signs and symptoms. A ratio of pleural-fluid lactate dehydrogenase to serum lactate dehydrogenase that is greater than 0.6 is consistent with the presence of an exudative effusion. The pH of the pleural fluid is higher than normal, possibly indicating the presence of an air bubble in the syringe used to collect the specimen. A predominance of lymphocytes is seen in tuberculosis and cancer and argues against bacterial infection. The presence of mesothelial cells and the negative smear for acid-fast bacilli reduce the likelihood of tuberculosis, but the smear has a very low sensitivity. Cancer remains a consideration, despite negative results on serologic analysis, since a single sample of pleural fluid has a yield of less than 50 percent in the presence of pleural effusion associated with cancer. Pleural biopsy could increase the diagnostic yield.
The patient was treated with intravenous diuretics during the next four days, with improvement in his dyspnea and edema. On hospital day 6, he vomited twice and had an abrupt onset of shortness of breath and abdominal pain. He became hypotensive and tachycardic. Computed tomographic (CT) angiography showed small emboli within several right subsegmental pulmonary arteries as well as within the left main pulmonary artery. Treatment with intravenous heparin was instituted.
A Pulmonologist: Pulmonary embolus alone does not explain all of his findings and is most likely a complication of his hospitalization or an underlying disorder. Although pulmonary infarction can cause an exudative pleural effusion, acute pulmonary emboli are not associated with pericardial effusion and ascites. Chronic thromboembolism with pulmonary hypertension and increased right-heart filling pressures could result in a clinical picture of peripheral edema, ascites, and pleural effusions, but a transudative effusion would be expected in that case. At this point, cancer is an important concern, but infectious or collagen vascular causes have not been ruled out.
Ultrasonography of the lower extremities showed no deep venous thrombosis. The patient was able to maintain oxygen saturations of 92 to 93 percent while receiving 100 percent supplemental oxygen.
A Cardiologist: The patient was dyspneic, tachycardic, and hypotensive, with pericardial effusion and pulmonary emboli. There was no evidence of profound right ventricular dysfunction on examination or echocardiography, so the hypotension is unlikely to be from pulmonary emboli. We do not have a recent assessment of jugular venous pressure. It is difficult to determine whether the pericardial effusion contributed to his instability. Although the absence of echocardiographic signs of tamponade is reassuring, such signs could be masked by pulmonary hypertension due to a combination of hypoxic vasoconstriction and pulmonary emboli. The size of the pericardial effusion suggests a subacute process; the effusion may have enlarged with anticoagulation.
The cause of pericardial effusion can be elusive. Possibilities include infection, cancer, radiation, collagen vascular disease, hypothyroidism, trauma, and uremia. I would suggest invasive cardiac hemodynamic measurements to help confirm or rule out tamponade. Pericardiocentesis in patients with tamponade would be therapeutic and is more likely to yield a specific diagnosis than in patients with an uncomplicated effusion.
Pericardiocentesis was performed. The initial pericardial pressure was 24 mm Hg (normal range, –3 to –6). Measurements of pressure in the right side of the heart were not performed because of concern that pulmonary emboli would be dislodged during placement of a pulmonary-artery catheter. After 800 ml of bloody fluid was removed, the pericardial pressure dropped to 3 mm Hg. A pericardial drainage catheter was placed.
Analysis of the pericardial fluid revealed that the glucose level was 83 mg per deciliter (4.6 mmol per liter), the lactate dehydrogenase level 163 U per liter, the protein level 4.7 g per liter, and the hematocrit 41.5 percent. The white-cell count was 4900 per cubic millimeter, with 49 percent neutrophils, 1 percent bands, 30 percent lymphocytes, 12 percent monocytes, 1 percent eosinophils, 6 percent macrophages, and 1 percent nucleated red cells.
A Cardiologist: A hematocrit of 41 percent indicates the presence of frank blood; the hematocrit would be lower if slow intrapericardial bleeding complicated an initially nonbloody effusion. The clinical history is not compatible with hemopericardium complicating myocardial infarction, aortic dissection, or chest trauma. An aggressive inflammatory process, bleeding diathesis, or erosion of a primary or metastatic tumor is a concern. An initial worry when fluid is grossly bloody is that the pericardiocentesis needle entered the right ventricle and perforated a chamber or lacerated an artery, but one would not expect the pericardial pressure to drop, hemodynamic variables to improve, or the effusion to resolve if these complications had occurred.
Echocardiography confirmed resolution of the effusion. Dyspnea and tachycardia improved, and the blood pressure remained stable. A heparin infusion was restarted. Bacterial cultures of the pericardial fluid, cytologic analysis, and a smear for acid-fast bacilli all were negative. On hospital day 8, chest CT was repeated.
A Cardiologist: The decision to restart heparin administration with the drainage catheter in place is appropriate if there is concern for further hemodynamic and respiratory derangement associated with an untreated pulmonary embolism. With a functioning drainage catheter, reaccumulation and recurrent cardiac tamponade would be unlikely. The unexplained hemopericardium should prompt continued investigation to identify a unifying process.
CT angiography showed no obstructive pulmonary emboli in the main pulmonary arteries. Multiple small filling defects within the segmental arteries supplying the left upper lobe and right upper lobe were present. The pericardial catheter was removed 48 hours later, after drainage ceased. Warfarin was started and heparin discontinued after a therapeutic INR was achieved.
A transthoracic echocardiogram revealed only a small circumferential pericardial effusion with no signs of tamponade. With diuresis, the patient's body weight returned to baseline, and the dyspnea resolved completely. An evaluation for hypercoagulability revealed that the patient was heterozygous for the factor V Leiden mutation.
The consensus of the clinical service was that the pericardial effusion was postviral or idiopathic, with bleeding into the pericardial space caused by anticoagulation. Pulmonary embolism was due to decreased mobility in the setting of a hypercoagulable state. Fluid retention, ascites, and pleural effusions were attributed to elevated filling pressures in the right side of the heart in the setting of subacute cardiac tamponade. The patient was discharged on hospital day 19.
A Hematologist-Oncologist: One cannot argue about short-term anticoagulation, since the patient does not appear to be bleeding further into his pericardial space. The question becomes whether lifelong anticoagulation is necessary. He has a hypercoagulable state due to a genetic alteration. This change results in the inability of protein C to work as a physiologic anticoagulant. Patients with the factor V Leiden mutation are more prone to deep venous thrombosis and pulmonary embolus. More than half of spontaneous deep venous thromboses are related to the factor V Leiden mutation in some populations. This estimate exceeds what is seen in other populations, and the patient is a 24-year-old man whose clinical condition is in a state of flux. He is a young, active sailor, and a discussion with him is needed about whether he wishes to continue anticoagulation after six months.
Three weeks later, the patient noted new, right-sided pleuritic chest pain and was sent to the emergency department. The INR was 1.6. CT angiography revealed resolution of the pleural effusions, a loculated pericardial effusion adjacent to the right side of the heart, and no new pulmonary embolus. Transthoracic echocardiography revealed enlargement of the pericardial effusion, with echocardiographic evidence of cardiac tamponade. Plans were made for a pericardial biopsy and the placement of a pericardial fenestration.
Transesophageal echocardiography, which was performed before the procedure, revealed a loculated pericardial effusion and a lobulated, echodense mass around the right atrium, with communication into the right atrial cavity. Color-flow Doppler imaging that was performed after the administration of contrast material revealed a communication between the right atrium and the pericardial space (Figure 2 and Video Clip 1 of the Supplementary Appendix, which is available with the full text of this article at www.nejm.org). The space measured 4 cm by 9 cm, and the defect in the right atrium measured 5 mm. The placement of a pericardial fenestration was aborted.
Figure 2. Color-Flow Doppler Image Obtained during Transesophageal Echocardiography.
The image shows flow through a communication between the right atrium (solid arrow) and the pericardial space, where a mass can be seen (open arrow).
A Hematologist-Oncologist: Metastatic tumors are by far the most common tumors in the heart. They occur 20 to 40 times as frequently as do primary tumors. Lung cancer is the most common cancer to metastasize to the heart, followed by breast and renal carcinoma, as well as melanoma and lymphoma. Germ-cell tumors are considered in a young person because they occur in a bimodal distribution with a peak in the teens and 20s. The most common primary benign cardiac tumor is myxoma, which has a predilection for the left atrium. With respect to malignant tumors that are primary to the heart, the various histologic subtypes of sarcoma are the most common. The invasive nature of this mass suggests that it is a malignant neoplasm.
One day later, the patient returned to the operating room for repair of the right atrial perforation and pericardiectomy. Several thrombi and a large mass were identified within the pericardial space. The mass extended from the pericardial space through a right atrial perforation and into the right atrium. No involvement of the pleura was noted.
Hematoxylin and eosin staining of the pericardium and right atrial mass revealed organizing thrombus and areas with a papillary configuration showing anastomosing vascular channels lined with atypical cells with hyperchromatic nuclei (Figure 3).
Figure 3. Specimens of a Mass in the Pericardium and Right Atrium (Hematoxylin and Eosin).
An organizing thrombus and areas with a papillary configuration with anastomosing vascular channels are visible (Panel A). At a higher magnification, vascular channels that are lined with atypical cells with hyperchromatic nuclei can be seen (Panel B).
Immunohistochemical staining for CD15 (Leu-M1), cytokeratins, S-100 protein, chromogranin, and calretinin were negative, which effectively ruled out malignant mesothelioma. Staining for factor VIII, Ulex europaeus agglutinin, CD34, and CD31 were all positive.
A Pathologist: Angiosarcoma and malignant papillary mesothelioma were highest on the histologic differential diagnosis. The results of immunohistochemical staining, the degree of cytologic atypia, and the infiltrating nature of the lesion suggest a diagnosis of angiosarcoma.
A Hematologist-Oncologist: Sarcomas are particularly aggressive, and unfortunately, no good therapies for this disease exist. Soft-tissue sarcomas, like this one, do respond to doxorubicin and ifosfamide with partial and sometimes complete regression. The treatment of primary cardiac sarcomas with transplantation has been reported in the literature but is controversial. Transplantation is not an option here, since the patient's disease is not localized. It is quite possible that the tumor began outside the heart and eroded into it, which led to the embolism that was seen in the patient's lungs.
The patient returned to Britain, where he lived, to undergo chemotherapy with doxorubicin and ifosfamide. Despite therapy, he died six months later.
Commentary
The clinicians were certainly surprised by the uncommon cause of this young man's illness. Factors making the diagnosis even more challenging were an initial presentation that was atypical for an underlying cardiac cancer, the lack of visualization of the tumor on transthoracic echocardiography and CT, and the concomitant pulmonary embolism. The exudative pleural effusion and the presence of frank blood in the pericardial space without a clear cause suggested a more insidious, unifying diagnosis of underlying cancer, rather than idiopathic pericardial effusion exacerbated by anticoagulation and complicated by cardiac tamponade, as was originally presumed.
Cardiac angiosarcoma accounted for many of this patient's findings, such as ascites and peripheral edema due to obstruction of the right side of the heart, hemopericardium, chamber perforation, and what were most likely tumor emboli and malignant pleural effusions, yet the diagnosis remained elusive until the tumor was seen on transesophageal echocardiography. Most cardiac tumors can be seen with transthoracic echocardiography, but the location of this tumor adjacent to the right atrium limited its visualization on conventional echocardiography and CT. Although transesophageal echocardiography has been shown to have increased sensitivity for the detection of right atrial tumors1 and intracardiac invasion and growth,2 routine use without suspicion of cardiac tumor is not advocated in the evaluation of pericardial effusion unless the result of a transthoracic study is inconclusive or suboptimal.
More than three quarters of primary cardiac tumors are benign myxomas, fibromas, or lipomas. Malignant primary cardiac tumors are extremely rare, and cardiac angiosarcomas account for approximately 25 percent of these tumors.3 They are rare tumors of mesenchymal origin that are composed of malignant cells that form vascular channels. Angiosarcomas appear most often in the right side of the heart, particularly in the right atrium or the pericardium, with a peak incidence between the third and fifth decades of life.4 An intracavitary-mass effect is common, as are local metastases. Hemopericardium or obliteration of the pericardial space by tumor cells and thrombus can be complications.5,6 Spontaneous rupture of an angiosarcoma is extremely rare.7,8 The mean survival after diagnosis is less than one year,4 but successful treatment with combined radiotherapy and chemotherapy followed by transplantation has been reported.9 The extent of our patient's disease at the time of presentation precluded consideration of this approach.
We are indebted to Dr. Wendy Stead and Dr. Jeffrey Goldsmith for their assistance in the preparation of the manuscript.
Source Information
From the Divisions of Cardiology (D.S.P.), Infectious Diseases (B.M.B.), and Pulmonary and Critical Care Medicine (R.M.S.), and the Department of Medicine (C.C.S.), Beth Israel Deaconess Medical Center and Harvard Medical School, Boston.
Address reprint requests to Dr. Pinto at the Interventional Cardiology Section, Beth Israel Deaconess Medical Center, 185 Pilgrim Rd., Boston, MA 02215, or at dpinto@bidmc.harvard.edu.
References
Leibowitz G, Keller NM, Daniel WG, et al. Transesophageal versus transthoracic echocardiography in the evaluation of right atrial tumors. Am Heart J 1995;130:1224-1227.
Geibel A, Kasper W, Keck A, Hofmann T, Konstantinides S, Just H. Diagnosis, localization and evaluation of malignancy of heart and mediastinal tumors by conventional and transesophageal echocardiography. Acta Cardiol 1996;51:395-408.
Lam KY, Dickens P, Chan AC. Tumors of the heart: a 20-year experience with a review of 12,485 consecutive autopsies. Arch Pathol Lab Med 1993;117:1027-1031.
Burke AP, Cowan D, Virmani R. Primary sarcomas of the heart. Cancer 1992;69:387-395.
Oshima K, Ohtaki A, Kano M, et al. Primary cardiac angiosarcoma associated with cardiac tamponade: case report. Jpn Circ J 1999;63:822-824.
Rodriguez-Vera FJ, Sobrino-Marquez JM, Garcia Moreno JM, Colchero Fernandez J. Tamponade as the clinical onset of a cardiac angiosarcoma. Rev Port Cardiol 2000;19:361-364.
Serra V, Moura L, Almeria C, Perez de Isla L, Zamorano J. Right atrial rupture. Rev Port Cardiol 2004;23:731-736.
Corso RB, Kraychete N, Nardeli S, et al. Spontaneous rupture of a right atrial angiosarcoma and cardiac tamponade. Arq Bras Cardiol 2003;81:611-613.
Baay P, Karwande SV, Kushner JP, Olsen S, Renlund DG. Successful treatment of a cardiac angiosarcoma with combined modality therapy. J Heart Lung Transplant 1994;13:923-925.(Duane S. Pinto, M.D., Bar)
A Pulmonologist: This young, previously healthy man describes dyspnea in the setting of substantial weight gain. The degree of weight gain and the time course suggest that this is a subacute process. The differential diagnosis is broad and includes cardiac, pulmonary, hepatic, and renal dysfunction.
On physical examination, the patient's vital signs were as follows: blood pressure, 127/82 mm Hg; heart rate, 102 beats per minute; temperature, 99.3°F (37.4°C); and respiration, 26 per minute. The oxygen saturation was 96 percent while the patient was receiving 100 percent supplemental oxygen. No pulsus paradoxus was detected. The patient appeared to be in mild distress but spoke in complete sentences. His sclerae were anicteric. The jugular venous pressure was not elevated. The lungs had decreased breath sounds and dullness to percussion approximately halfway up. The first and second heart sounds were normal, with no rub, murmurs, or gallops. Abdominal examination revealed mild tenderness in the right upper quadrant and a fluid wave without hepatosplenomegaly. The patient had 3+ edema bilaterally and no joint abnormalities, rashes, jaundice, or lymphadenopathy. His mental status was normal. Chest radiography revealed an enlarged cardiac silhouette, with moderate bilateral pleural effusions with no infiltrate or interstitial edema (Figure 1).
Figure 1. Anteroposterior Chest Radiograph Showing an Enlarged Cardiac Silhouette with Moderate Bilateral Pleural Effusions and No Infiltrate or Interstitial Edema.
A Pulmonologist: The patient has evidence of fluid in the pleural and peritoneal spaces and, according to the results of chest radiography, a pericardial effusion as well. Thoracentesis is warranted, with examination of the pleural fluid to determine whether it is transudative, owing to increased hydrostatic or reduced oncotic pressures, or exudative, which would suggest an inflammatory or infectious process or cancer. Cardiac disorders causing increased hydrostatic pressures and hepatic or renal problems causing reduced serum protein levels may lead to diffuse fluid accumulation in pleural, pericardial, peritoneal, and interstitial spaces. Inflammatory processes cause interstitial edema infrequently; although multiple serosal surfaces can be involved, localized fluid collections are more common.
Although a patient with viral infection can present with diffuse serositis, the time course of this man's illness and the absence of any myalgias or constitutional symptoms are atypical. Bacterial causes of diffuse serositis are uncommon, but occasionally, pneumonia with a parapneumonic effusion can occur with a pericardial effusion; ascites, however, would be unlikely. Tuberculosis may develop insidiously, with pleural, pericardial, and peritoneal effusions. Another common cause of diffuse serositis is collagen vascular disease, including Still's disease, systemic lupus erythematosus, and mixed connective-tissue disease. Familial Mediterranean fever would need to be considered if the patient had a history of recurrent episodes of serositis and fever.
The presence of ascites raises the possibility of cirrhosis with elevated portal pressures. A normal jugular venous pulse argues against cardiac dysfunction leading to increased hydrostatic pressures. I would be interested in an analysis of the pleural fluid, a complete blood count, a test for antinuclear antibodies, and urinalysis.
The serum electrolytes were within normal limits. The glucose level was 109 mg per deciliter (6.0 mmol per liter); blood urea nitrogen, 20 mg per deciliter (7.1 mmol per liter); creatinine, 1.9 mg per deciliter (168.0 μmol per liter); aspartate aminotransferase, 855 U per liter; alanine aminotransferase, 822 U per liter; total bilirubin, 1.2 mg per deciliter (20.5 μmol per liter); alkaline phosphatase, 99 U per liter; albumin, 3.0 g per deciliter; and lactate dehydrogenase, 52 U per liter. The white-cell count was 11,100 per cubic millimeter, the hemoglobin level was 13.1 g per deciliter, and the platelet count was 285,000 per cubic millimeter. The international normalized ratio (INR) of prothrombin time was 1.5. The antinuclear-antibody test was negative. Urine dipstick analysis and microscopy revealed no abnormalities.
Analysis of the pleural fluid showed that the pH was 8.0, the glucose level 83 mg per deciliter (4.6 mmol per liter), the lactate dehydrogenase level 98 U per liter, and the protein level 2.9 g per deciliter. The white-cell count was 220 per cubic millimeter, with 8 percent neutrophils, 68 percent lymphocytes, 13 percent macrophages, and 11 percent mesothelial cells. Bacterial cultures, cytologic tests, and a smear for acid-fast bacilli were negative. A transthoracic echocardiogram showed preserved right and left ventricular function with a moderate-to-large pericardial effusion but no echocardiographic evidence of cardiac tamponade.
A Pulmonologist: The relatively normal serum albumin level rules out reduced oncotic pressure as a cause of the fluid collections. A connective-tissue disorder is unlikely given the negative antinuclear-antibody test and the absence of other characteristic signs and symptoms. A ratio of pleural-fluid lactate dehydrogenase to serum lactate dehydrogenase that is greater than 0.6 is consistent with the presence of an exudative effusion. The pH of the pleural fluid is higher than normal, possibly indicating the presence of an air bubble in the syringe used to collect the specimen. A predominance of lymphocytes is seen in tuberculosis and cancer and argues against bacterial infection. The presence of mesothelial cells and the negative smear for acid-fast bacilli reduce the likelihood of tuberculosis, but the smear has a very low sensitivity. Cancer remains a consideration, despite negative results on serologic analysis, since a single sample of pleural fluid has a yield of less than 50 percent in the presence of pleural effusion associated with cancer. Pleural biopsy could increase the diagnostic yield.
The patient was treated with intravenous diuretics during the next four days, with improvement in his dyspnea and edema. On hospital day 6, he vomited twice and had an abrupt onset of shortness of breath and abdominal pain. He became hypotensive and tachycardic. Computed tomographic (CT) angiography showed small emboli within several right subsegmental pulmonary arteries as well as within the left main pulmonary artery. Treatment with intravenous heparin was instituted.
A Pulmonologist: Pulmonary embolus alone does not explain all of his findings and is most likely a complication of his hospitalization or an underlying disorder. Although pulmonary infarction can cause an exudative pleural effusion, acute pulmonary emboli are not associated with pericardial effusion and ascites. Chronic thromboembolism with pulmonary hypertension and increased right-heart filling pressures could result in a clinical picture of peripheral edema, ascites, and pleural effusions, but a transudative effusion would be expected in that case. At this point, cancer is an important concern, but infectious or collagen vascular causes have not been ruled out.
Ultrasonography of the lower extremities showed no deep venous thrombosis. The patient was able to maintain oxygen saturations of 92 to 93 percent while receiving 100 percent supplemental oxygen.
A Cardiologist: The patient was dyspneic, tachycardic, and hypotensive, with pericardial effusion and pulmonary emboli. There was no evidence of profound right ventricular dysfunction on examination or echocardiography, so the hypotension is unlikely to be from pulmonary emboli. We do not have a recent assessment of jugular venous pressure. It is difficult to determine whether the pericardial effusion contributed to his instability. Although the absence of echocardiographic signs of tamponade is reassuring, such signs could be masked by pulmonary hypertension due to a combination of hypoxic vasoconstriction and pulmonary emboli. The size of the pericardial effusion suggests a subacute process; the effusion may have enlarged with anticoagulation.
The cause of pericardial effusion can be elusive. Possibilities include infection, cancer, radiation, collagen vascular disease, hypothyroidism, trauma, and uremia. I would suggest invasive cardiac hemodynamic measurements to help confirm or rule out tamponade. Pericardiocentesis in patients with tamponade would be therapeutic and is more likely to yield a specific diagnosis than in patients with an uncomplicated effusion.
Pericardiocentesis was performed. The initial pericardial pressure was 24 mm Hg (normal range, –3 to –6). Measurements of pressure in the right side of the heart were not performed because of concern that pulmonary emboli would be dislodged during placement of a pulmonary-artery catheter. After 800 ml of bloody fluid was removed, the pericardial pressure dropped to 3 mm Hg. A pericardial drainage catheter was placed.
Analysis of the pericardial fluid revealed that the glucose level was 83 mg per deciliter (4.6 mmol per liter), the lactate dehydrogenase level 163 U per liter, the protein level 4.7 g per liter, and the hematocrit 41.5 percent. The white-cell count was 4900 per cubic millimeter, with 49 percent neutrophils, 1 percent bands, 30 percent lymphocytes, 12 percent monocytes, 1 percent eosinophils, 6 percent macrophages, and 1 percent nucleated red cells.
A Cardiologist: A hematocrit of 41 percent indicates the presence of frank blood; the hematocrit would be lower if slow intrapericardial bleeding complicated an initially nonbloody effusion. The clinical history is not compatible with hemopericardium complicating myocardial infarction, aortic dissection, or chest trauma. An aggressive inflammatory process, bleeding diathesis, or erosion of a primary or metastatic tumor is a concern. An initial worry when fluid is grossly bloody is that the pericardiocentesis needle entered the right ventricle and perforated a chamber or lacerated an artery, but one would not expect the pericardial pressure to drop, hemodynamic variables to improve, or the effusion to resolve if these complications had occurred.
Echocardiography confirmed resolution of the effusion. Dyspnea and tachycardia improved, and the blood pressure remained stable. A heparin infusion was restarted. Bacterial cultures of the pericardial fluid, cytologic analysis, and a smear for acid-fast bacilli all were negative. On hospital day 8, chest CT was repeated.
A Cardiologist: The decision to restart heparin administration with the drainage catheter in place is appropriate if there is concern for further hemodynamic and respiratory derangement associated with an untreated pulmonary embolism. With a functioning drainage catheter, reaccumulation and recurrent cardiac tamponade would be unlikely. The unexplained hemopericardium should prompt continued investigation to identify a unifying process.
CT angiography showed no obstructive pulmonary emboli in the main pulmonary arteries. Multiple small filling defects within the segmental arteries supplying the left upper lobe and right upper lobe were present. The pericardial catheter was removed 48 hours later, after drainage ceased. Warfarin was started and heparin discontinued after a therapeutic INR was achieved.
A transthoracic echocardiogram revealed only a small circumferential pericardial effusion with no signs of tamponade. With diuresis, the patient's body weight returned to baseline, and the dyspnea resolved completely. An evaluation for hypercoagulability revealed that the patient was heterozygous for the factor V Leiden mutation.
The consensus of the clinical service was that the pericardial effusion was postviral or idiopathic, with bleeding into the pericardial space caused by anticoagulation. Pulmonary embolism was due to decreased mobility in the setting of a hypercoagulable state. Fluid retention, ascites, and pleural effusions were attributed to elevated filling pressures in the right side of the heart in the setting of subacute cardiac tamponade. The patient was discharged on hospital day 19.
A Hematologist-Oncologist: One cannot argue about short-term anticoagulation, since the patient does not appear to be bleeding further into his pericardial space. The question becomes whether lifelong anticoagulation is necessary. He has a hypercoagulable state due to a genetic alteration. This change results in the inability of protein C to work as a physiologic anticoagulant. Patients with the factor V Leiden mutation are more prone to deep venous thrombosis and pulmonary embolus. More than half of spontaneous deep venous thromboses are related to the factor V Leiden mutation in some populations. This estimate exceeds what is seen in other populations, and the patient is a 24-year-old man whose clinical condition is in a state of flux. He is a young, active sailor, and a discussion with him is needed about whether he wishes to continue anticoagulation after six months.
Three weeks later, the patient noted new, right-sided pleuritic chest pain and was sent to the emergency department. The INR was 1.6. CT angiography revealed resolution of the pleural effusions, a loculated pericardial effusion adjacent to the right side of the heart, and no new pulmonary embolus. Transthoracic echocardiography revealed enlargement of the pericardial effusion, with echocardiographic evidence of cardiac tamponade. Plans were made for a pericardial biopsy and the placement of a pericardial fenestration.
Transesophageal echocardiography, which was performed before the procedure, revealed a loculated pericardial effusion and a lobulated, echodense mass around the right atrium, with communication into the right atrial cavity. Color-flow Doppler imaging that was performed after the administration of contrast material revealed a communication between the right atrium and the pericardial space (Figure 2 and Video Clip 1 of the Supplementary Appendix, which is available with the full text of this article at www.nejm.org). The space measured 4 cm by 9 cm, and the defect in the right atrium measured 5 mm. The placement of a pericardial fenestration was aborted.
Figure 2. Color-Flow Doppler Image Obtained during Transesophageal Echocardiography.
The image shows flow through a communication between the right atrium (solid arrow) and the pericardial space, where a mass can be seen (open arrow).
A Hematologist-Oncologist: Metastatic tumors are by far the most common tumors in the heart. They occur 20 to 40 times as frequently as do primary tumors. Lung cancer is the most common cancer to metastasize to the heart, followed by breast and renal carcinoma, as well as melanoma and lymphoma. Germ-cell tumors are considered in a young person because they occur in a bimodal distribution with a peak in the teens and 20s. The most common primary benign cardiac tumor is myxoma, which has a predilection for the left atrium. With respect to malignant tumors that are primary to the heart, the various histologic subtypes of sarcoma are the most common. The invasive nature of this mass suggests that it is a malignant neoplasm.
One day later, the patient returned to the operating room for repair of the right atrial perforation and pericardiectomy. Several thrombi and a large mass were identified within the pericardial space. The mass extended from the pericardial space through a right atrial perforation and into the right atrium. No involvement of the pleura was noted.
Hematoxylin and eosin staining of the pericardium and right atrial mass revealed organizing thrombus and areas with a papillary configuration showing anastomosing vascular channels lined with atypical cells with hyperchromatic nuclei (Figure 3).
Figure 3. Specimens of a Mass in the Pericardium and Right Atrium (Hematoxylin and Eosin).
An organizing thrombus and areas with a papillary configuration with anastomosing vascular channels are visible (Panel A). At a higher magnification, vascular channels that are lined with atypical cells with hyperchromatic nuclei can be seen (Panel B).
Immunohistochemical staining for CD15 (Leu-M1), cytokeratins, S-100 protein, chromogranin, and calretinin were negative, which effectively ruled out malignant mesothelioma. Staining for factor VIII, Ulex europaeus agglutinin, CD34, and CD31 were all positive.
A Pathologist: Angiosarcoma and malignant papillary mesothelioma were highest on the histologic differential diagnosis. The results of immunohistochemical staining, the degree of cytologic atypia, and the infiltrating nature of the lesion suggest a diagnosis of angiosarcoma.
A Hematologist-Oncologist: Sarcomas are particularly aggressive, and unfortunately, no good therapies for this disease exist. Soft-tissue sarcomas, like this one, do respond to doxorubicin and ifosfamide with partial and sometimes complete regression. The treatment of primary cardiac sarcomas with transplantation has been reported in the literature but is controversial. Transplantation is not an option here, since the patient's disease is not localized. It is quite possible that the tumor began outside the heart and eroded into it, which led to the embolism that was seen in the patient's lungs.
The patient returned to Britain, where he lived, to undergo chemotherapy with doxorubicin and ifosfamide. Despite therapy, he died six months later.
Commentary
The clinicians were certainly surprised by the uncommon cause of this young man's illness. Factors making the diagnosis even more challenging were an initial presentation that was atypical for an underlying cardiac cancer, the lack of visualization of the tumor on transthoracic echocardiography and CT, and the concomitant pulmonary embolism. The exudative pleural effusion and the presence of frank blood in the pericardial space without a clear cause suggested a more insidious, unifying diagnosis of underlying cancer, rather than idiopathic pericardial effusion exacerbated by anticoagulation and complicated by cardiac tamponade, as was originally presumed.
Cardiac angiosarcoma accounted for many of this patient's findings, such as ascites and peripheral edema due to obstruction of the right side of the heart, hemopericardium, chamber perforation, and what were most likely tumor emboli and malignant pleural effusions, yet the diagnosis remained elusive until the tumor was seen on transesophageal echocardiography. Most cardiac tumors can be seen with transthoracic echocardiography, but the location of this tumor adjacent to the right atrium limited its visualization on conventional echocardiography and CT. Although transesophageal echocardiography has been shown to have increased sensitivity for the detection of right atrial tumors1 and intracardiac invasion and growth,2 routine use without suspicion of cardiac tumor is not advocated in the evaluation of pericardial effusion unless the result of a transthoracic study is inconclusive or suboptimal.
More than three quarters of primary cardiac tumors are benign myxomas, fibromas, or lipomas. Malignant primary cardiac tumors are extremely rare, and cardiac angiosarcomas account for approximately 25 percent of these tumors.3 They are rare tumors of mesenchymal origin that are composed of malignant cells that form vascular channels. Angiosarcomas appear most often in the right side of the heart, particularly in the right atrium or the pericardium, with a peak incidence between the third and fifth decades of life.4 An intracavitary-mass effect is common, as are local metastases. Hemopericardium or obliteration of the pericardial space by tumor cells and thrombus can be complications.5,6 Spontaneous rupture of an angiosarcoma is extremely rare.7,8 The mean survival after diagnosis is less than one year,4 but successful treatment with combined radiotherapy and chemotherapy followed by transplantation has been reported.9 The extent of our patient's disease at the time of presentation precluded consideration of this approach.
We are indebted to Dr. Wendy Stead and Dr. Jeffrey Goldsmith for their assistance in the preparation of the manuscript.
Source Information
From the Divisions of Cardiology (D.S.P.), Infectious Diseases (B.M.B.), and Pulmonary and Critical Care Medicine (R.M.S.), and the Department of Medicine (C.C.S.), Beth Israel Deaconess Medical Center and Harvard Medical School, Boston.
Address reprint requests to Dr. Pinto at the Interventional Cardiology Section, Beth Israel Deaconess Medical Center, 185 Pilgrim Rd., Boston, MA 02215, or at dpinto@bidmc.harvard.edu.
References
Leibowitz G, Keller NM, Daniel WG, et al. Transesophageal versus transthoracic echocardiography in the evaluation of right atrial tumors. Am Heart J 1995;130:1224-1227.
Geibel A, Kasper W, Keck A, Hofmann T, Konstantinides S, Just H. Diagnosis, localization and evaluation of malignancy of heart and mediastinal tumors by conventional and transesophageal echocardiography. Acta Cardiol 1996;51:395-408.
Lam KY, Dickens P, Chan AC. Tumors of the heart: a 20-year experience with a review of 12,485 consecutive autopsies. Arch Pathol Lab Med 1993;117:1027-1031.
Burke AP, Cowan D, Virmani R. Primary sarcomas of the heart. Cancer 1992;69:387-395.
Oshima K, Ohtaki A, Kano M, et al. Primary cardiac angiosarcoma associated with cardiac tamponade: case report. Jpn Circ J 1999;63:822-824.
Rodriguez-Vera FJ, Sobrino-Marquez JM, Garcia Moreno JM, Colchero Fernandez J. Tamponade as the clinical onset of a cardiac angiosarcoma. Rev Port Cardiol 2000;19:361-364.
Serra V, Moura L, Almeria C, Perez de Isla L, Zamorano J. Right atrial rupture. Rev Port Cardiol 2004;23:731-736.
Corso RB, Kraychete N, Nardeli S, et al. Spontaneous rupture of a right atrial angiosarcoma and cardiac tamponade. Arq Bras Cardiol 2003;81:611-613.
Baay P, Karwande SV, Kushner JP, Olsen S, Renlund DG. Successful treatment of a cardiac angiosarcoma with combined modality therapy. J Heart Lung Transplant 1994;13:923-925.(Duane S. Pinto, M.D., Bar)