Logo
Search for

Volume 21, Issue 1, Pages 7-13 (January 2008)


View previous. 7 of 28 View next.

Echocardiography in Pericardial Disease

Samuel Wann, MD, FASEaCorresponding Author Informationemail address, Edward Passen, MDb

Since its introduction into clinical cardiology in the 1960’s, echocardiography has become an essential and integral part of the evaluation of the cardiac patient. Dr. Harvey Feigenbaum had a seminal role in popularizing echocardiography. Detection of pericardial effusion was one of the first applications of echocardiography to gain widespread acceptance. Over the past 40 years, echocardiography has become a ubiquitous, first line tool for evaluating the pericardium and other cardiac structures.

Article Outline

Abstract

Pericardial Effusion

Cardiac Tamponade

Constrictive Pericarditis

Effusive Constrictive Pericarditis

Summary

References

Copyright

By his own account, Harvey Feigenbaum1 was not the first to describe the use of reflected ultrasound to detect pericardial effusion. Inge Edler,2 an early pioneer in the use of high-frequency sound waves for cardiac diagnosis, published the first ultrasound images of an anterior pericardial effusion in 1961.3 Other investigators also described the use of cardiac ultrasound to detect pericardial fluid,4, 5, 6, 7 but Dr Feigenbaum8 was clearly one of the first to recognize the enormous clinical potential of ultrasound for detecting pericardial effusion. His early studies using A-mode (Figure 1), M-mode, and later two-dimensional echocardiography, as well as his vision and passion, were instrumental in promoting the widespread use of echocardiography, a term he helped popularize, as a practical tool for accurate, noninvasive detection of pericardial effusion.8, 9, 10, 11, 12 These early observations also introduced clinicians everywhere in the broader potential of echocardiography at a time when the diagnosis of mitral stenosis was its primary but limited role.


View full-size image.

Figure 1. A-mode echocardiogram from Dr. Feigenbaum’s 1965 publication8 recorded in a patient before (top panel) and after (bottom panel) pericardiocentesis. “PW” was thought to be the posterior wall of the left ventricle, separated from “P,” the pericardium-fluid interface, by a pericardial effusion. This separation disappeared following pericardiocentesis. Similar findings were demonstrated during intrapericardial saline injection in 5 mongrel dogs.


Before the introduction of echocardiography, detection of pericardial effusion was often a difficult diagnostic challenge, sometimes requiring potentially hazardous procedures, such as cardiac catheterization or even blind pericardiocentesis, to confirm a clinical suggestion. Although the history, physical examination, electrocardiogram, and chest radiograph are useful in the investigation of pericardial disease, all lack the sensitivity and specificity of the echocardiogram. Introduction of the echocardiogram revolutionized the evaluation of pericardial disease and echocardiography continues to be the primary diagnostic tool for assessing the spatial distribution and functional significance of pericardial diseases.

As shown in Figure 2, the pericardium surrounds the heart and extends to insert on the proximal great vessels. The pericardium consists of a visceral layer, which is contiguous with the epicardium of the heart, and a parietal layer, which forms a sac around the heart. The phylogenetic purpose of the pericardium is unclear.13, 14 Anatomically, the pericardium isolates the heart from the rest of the mediastinum and thorax. Physiologically, the pericardium may have little if any significant role under normal circumstances, but it is affected by a variety of disease conditions often with major clinical consequences.15


View full-size image.

Figure 2. CT “pericardiogram.” Iodinated contrast material was inadvertently injected into the pericardial space during an attempted pulmonary angiogram. The catheter perforated the coronary sinus. Note that the contrast in the pericardial space is seen not only around the ventricles and atria, but extends into the pericardial reflection on the proximal pulmonary artery and aorta. Catheter based pulmonary angiography has largely been abandoned in favor of CT pulmonary angiography with peripheral venous injection of contrast.


Echocardiography can be useful in detecting a variety of conditions that affect the pericardium including: (1) partial or complete absence of the pericardium; (2) primary pericardial cysts and both primary and metastatic pericardial tumors; (3) inflammation with secondary effusion as a result of bacterial or viral infection, myocardial infarction, trauma, uremia, neoplasm, hypothyroidism, or high-dose radiation exposure; (4) pericardial tamponade that occurs when the fluid accumulation causes an increase in pressure in the pericardial sac, pressing on the heart and interfering with normal filling/function; and (5) pericardial constriction as a result of thickening, fibrosis, and adherence between the layers of the pericardium, constricting the heart, and impairing diastolic filling.

Pericardial Effusion 

return to Article Outline

Echocardiography is usually the primary diagnostic method used for initial detection of pericardial effusion. As shown in Figure 3, pericardial fluid is recognized as an echo-free space between the visceral and parietal pericardium surrounding the heart. Pericardial fluid first accumulates posterior to the heart, when the patient is examined in the supine position, as the heart itself usually floats in transudative pericardial fluid. Small pericardial effusions are generally visualized as small echo-free spaces in the posterior atrioventricular groove, and small amounts of fluid can be present even in healthy individuals. As the effusion increases, it extends laterally and with large effusions the echo-free space expands to surround the entire heart. Large effusion also extends behind the left atrium (between the posterior atrial wall and the aorta) (Figure 4) and can, thus, be differentiated from pleural effusions, which do not usually appear immediately posterior to the left atrium.


View full-size image.

Figure 3. Large pericardial effusion, presenting as an echo-free space anterior and posterior to the heart.



View full-size image.

Figure 4. Large pleural effusion lies posterior to the heart and the descending thoracic aorta (DA). The patient also has a pericardial effusion (PE), which extends behind the left atrium, anterior to the descending aorta. There is also a small amount of pericardial effusion anterior to the right ventricle. AO – Ascending aorta. LA – left atrium. LV – left ventricle.


Two-dimensional echocardiography is generally quite sensitive and specific in detecting pericardial effusion (both free and loculated), and can be used to estimate the amount of fluid present. Very small pericardial effusions may be most obvious during systole. An echo-free space appearing only anterior to the heart is usually caused by pericardial fat, not effusion (Figure 5). Differentiating pericardial fat from localized pericardial effusion can be difficult in some patients. Computed tomography may be helpful.16


View full-size image.

Figure 5. An echo-free clear space seen only anterior to the heart (arrow) is usually due to pericardial fat, not pericardial effusion, as the heart “floats” in pericardial fluid.


It is difficult to characterize the origin of fluid in the pericardial space by its echocardiographic appearance alone, but hemorrhagic or purulent fluid (Figure 6) may be more echogenic than simple serous fluid. Fibrinous bands and mass lesions can sometimes be seen in the pericardial space. Percutaneous needle pericardiocentesis may be safely performed under real-time echocardiography guidance17 (Figure 7, A). Successful removal of the pericardial fluid can be readily documented (Figure 7, B and C).


View full-size image.

Figure 6. An echogenic space surrounding the heart (arrow) is seen in this patient who has a bloody pericardial effusion.



View full-size image.

Figure 7. (A) Echocardiographically guided pericardiocentesis. The needle (arrow) is seen within the pericardial effusion (PE). (B) Two-dimensional echocardiogram prior to pericardiocentesis. Note echo-free space around the heart and invagination of the right atrial free wall. (C) Two-dimensional echocardiogram taken immediately after completion of pericardiocentesis, showing resolution of the pericardial effusion and no right atrial collapse.


Cardiac Tamponade 

return to Article Outline

The pericardial sac can gradually stretch to accommodate increasing volume, but at any point in time, the total intrapericardial volume is relatively fixed throughout the cardiac cycle. When the ability of the pericardium to stretch is exceeded by rapid or massive accumulation of fluid, any additional fluid causes the pressure within the pericardial sac to increase. When the increasing intrapericardial pressure exceeds the intracardiac pressure, the positive transmural pressure gradient compresses the adjacent cardiac chamber or chambers. Because intracardiac pressures vary throughout the cardiac cycle, the pericardial pressure will exceed the intracardiac pressure within different chambers at different points in the cardiac cycle. The chambers with the lowest instantaneous pressures are affected first. Hence, right atrial inversion (during ventricular systole, while the atrium is relaxed) is usually an early sign of compression, followed by diastolic compression of the right ventricular outflow tract. Respiration also affects intracardiac pressures, particularly those on the right side of the heart. During inspiration, intrathoracic and intrapericardial pressures decrease, resulting in augmented flow into the right atrium and right ventricle, with decreased flow out of the pulmonary veins into the left atrium and left ventricle. Reciprocal changes occur during expiration, and can be documented by Doppler echocardiographic changes in mitral and tricuspid inflow as well as pulmonary and systemic outflow (Figure 8).18, 19 In the presence of normal intrapericardial pressure, the normal respiratory variation in filling of the left and right heart causes a small (<10 mm Hg) inspiratory decrease in systemic arterial systolic blood pressure.


View full-size image.

Figure 8. (A) Mitral valve Doppler recording; (B) aortic valve Doppler recording.


Echocardiography is a powerful tool for investigating the complex relationships among anatomy, pressure and flow, and relative ventricular volume, which can be affected by interaction among the pericardium, the pleura, and the heart.20, 21, 22 When intrapericardial pressure exceeds right atrial pressure, the thin free wall of the right atrium invaginates, collapsing toward the right atrial cavity (Figure 9). Similarly, the right ventricular free wall shows signs of collapse when the intrapericardial pressure exceeds right ventricular intracavitary pressure (Figure 10). Very small differences in pressure result in brief late diastolic invagination of the right atrial free wall in the absence of clinically significant tamponade. The longer the duration of right atrial invagination relative to the length of the cardiac cycle, the greater the likelihood of significant hemodynamic compromise.


View full-size image.

Figure 9. A patient with a large pericardial effusion (PE) shows evidence of right atrial (RA) collapse (arrow). The right ventricle (RV), left ventricle (LV), and left atrium (LA) are normal. Right atrial collapse by itself is not an indicator of significant hemodynamic embarrassment due to cardiac tamponade, but is an early sign that intra-pericardial pressure exceeds intra-cardiac pressure.



View full-size image.

Figure 10. Subcostal view showing pericardial effusion (PE), right atrium (RA), left ventricle (LV), and left atrium (LA).


Right ventricular diastolic collapse generally requires a larger pressure difference between the intrapericardial space and the intracardiac chambers than does right atrial collapse. Right ventricular diastolic collapse is, thus, a more specific but less sensitive indicator of hemodynamically significant pericardial compression. In experimental studies, diastolic inversion of the right ventricular free wall begins abruptly. At the first appearance of right ventricular collapse, intrapericardial and intracardiac pressures are essentially equal, with an associated moderate decline in stroke volume. As intrapericardial pressure increases relative to intracardiac pressure, the duration and severity of right ventricular collapse progresses to the point of near obliteration of the right ventricular cavity throughout diastole, and ultimate circulatory collapse.

Both right atrial and right ventricular collapses are affected by changes in the state of hydration, central blood pressure, and volume. Pulmonary hypertension and right ventricular hypertrophy can delay right ventricular collapse until very high intrapericardial pressure is present.21, 22

Tamponade also produces reciprocal respiration-related changes in right and left ventricular volume, diastolic filling, and systolic emptying that can be documented by echocardiography. These findings underlie the physiology of the exaggerated decrease in systemic aortic pressure or pulsus paradoxus that is the clinical hallmark of tamponade. The exaggerated decrease in arterial pressure occurs because, in the setting of a fixed intrapericardial volume, the inspiratory increase in right ventricular filling causes the interventricular septum to shift to the left, exaggerating the normal decrease in LV filling volume and, hence, stroke volume (ventricular interdependence). With expiration, left ventricular filling and stroke volume increase, as right ventricular filling is relatively reduced, leading to the exaggerated variation in stroke volume and arterial pressure.

Doppler echocardiography is particularly useful in demonstrating the exaggerated phasic variation in ventricular inflow and outflow caused by tamponade.19 Respiratory variation in tricuspid and pulmonary flow is more dramatic than mitral and aortic flow, but there is progressive impairment in all intracardiac flow as the degree of tamponade worsens. Plethora of the inferior vena cava is a useful indicator of elevated right atrial pressure and is usually present in both cardiac tamponade and constrictive pericarditis (Figure 11).


View full-size image.

Figure 11. (A) Subcostal two-dimensional echocardiogram showing ascites (As) above the liver, separated from a pericardial effusion (PE) by the diaphragm. The inferior vena cava (IVC) is dilated as it enters the right atrium (RA), indicative of elevated central venous pressure. The hepatic veins (HV) are also dilated. (B) An M-mode echocardiogram of the inferior vena cava (IVC) shows the absence of respiratory variation in the diameter of the dilated IVC.


These echocardiographic findings identify changes in cardiac structure and function and an associated decrease in cardiac output that often occur well before the onset of pulsus paradoxus and, thus, can be an important indicator of hemodynamic compromise before it becomes clinically apparent.

Constrictive Pericarditis 

return to Article Outline

Constrictive pericarditis is an uncommon, easily misdiagnosed disease. Constrictive pericarditis should be in the differential diagnosis for all patients who present with dyspnea, ascites, edema, and elevated jugular venous pressure, especially when left ventricular systolic function is normal.

Chronic inflammation of the pericardium results in thickening, fibrosis, and fusion of both visceral and parietal layers causing impaired diastolic cardiac filling.23 Echocardiography may be useful in recognizing an increase in pericardial thickness (Figure 12),24 but precise measurements are difficult because of reverberations of reflected ultrasound or shadowing caused by calcification. The first clue that constrictive pericarditis may be present is often the finding of a dilated vena cava indicative of increased central venous pressure (Figure 11) in the presence of normal right and left ventricular systolic function. In classic constrictive pericarditis, pressures in all four cardiac chambers equalize during diastole. At the onset of diastole, the rate of ventricular filling is often increased as a result of the elevated atrial pressures, but is rapidly arrested by the pericardial constraint, resulting in a rapid increase in ventricular diastolic pressure creating the characteristic dip-and-plateau pattern. Because the total ventricular volume is limited by the pericardium, increased filling of one ventricle occurs at the expense of the other (ventricular interdependence). This ventricular interdependence underlies many of the echocardiographic signs of constriction including the respiratory shift in the position of the interventricular septum (Figure 13), and exaggeration of the normal respiratory changes in mitral and tricuspid flow (tricuspid flow increases more than normal while mitral flow decreases more than normal with inspiration).


View full-size image.

Figure 12. Thickened pericardium (arrow) in a patient with constrictive pericarditis.



View full-size image.

Figure 13. Constrast enhanced two-dimensional echocardiogram in a patient with constrictive pericarditis. (A) During inspiration the interventricular septum (IVS) curves toward the right ventricle (RV) as the left ventricle (LV) expands. (B) During inspiration, the interventricular septum is shifted toward the left ventricle due to fixed intrapericardial volumes, as the right heart gets larger, the left heart must get smaller.


Restrictive cardiomyopathy can share many of the clinical features of constrictive pericarditis and a number of echocardiographic features have been described to separate these entities. Causes of restrictive cardiomyopathy such as amyloid heart disease may be suggested on echocardiography when the cardiac walls are markedly thickened. Other features suggesting that a patient has restriction include the presence of severe pulmonary hypertension and more severe elevation of left ventricular filling pressures than right ventricular filling pressures, as these findings are not typical in constriction.

Doppler echocardiography may be especially useful in differentiating constrictive pericarditis from restrictive cardiomyopathy.24, 25, 26, 27, 28, 29, 30, 31, 32 The marked respiratory variation in left and right ventricular inflow velocities seen in constrictive pericarditis is not typically present in restrictive cardiomyopathy. Enhanced respiratory variation in pulmonary vein diastolic flow may be augmented by rapid intravenous volume loading in patients with constrictive pericarditis.

Tissue Doppler echocardiography and color Doppler M-mode have also been reported to be of value in differentiating restrictive cardiomyopathy from constrictive pericarditis.32 Em, a tissue Doppler parameter of early relaxation, is normal in constriction, but reduced in restriction as a result of intrinsic myocardial disease. Flow propagation velocity into the left ventricle by color Doppler M-mode is greater than 45 cm/s in constriction but reduced to less than 45 cm/s in restriction.33, 34

Effusive Constrictive Pericarditis 

return to Article Outline

Effusive constrictive pericarditis combines features of cardiac tamponade with those of constrictive pericarditis. Figure 14 shows the echocardiogram of such a patient, who has both a large pericardial effusion pressing on the heart and a markedly thickened pericardium impairing diastolic filling. Both the hemodynamic and Doppler echocardiographic findings may be intermediate, showing features of both cardiac tamponade and constriction.


View full-size image.

Figure 14. Effuso-constrictive pericarditis. (A) shows both a pericardial effusion (small arrow) and thickened pericardium (large arrow). (B) shows respiratory variation in mitral inflow consistent with pulsus paradoxicus.


Summary 

return to Article Outline

Current widespread use of echocardiography to diagnose pericardial disease has proven that Harvey Feigenbaum was right in his early enthusiasm for this technology. Ultrasound is harmless, relatively inexpensive, and widely available. The ability of echocardiography to elucidate the functional and structural consequences of pericardial disease is especially powerful. Although other noninvasive technologies including cardiac magnetic resonance and computed tomography have been developed to provide even more detailed information about the heart and the pericardium, echocardiography remains the first and often only diagnostic method needed to make a definitive diagnosis and guide appropriate treatment in patients with pericardial effusion, cardiac tamponade, or constrictive pericarditis. The development of echocardiography for use in pericardial disease has been a landmark development of the last half-century of cardiology.

References 

return to Article Outline

1. 1Feigenbaum H. History of echocardiography. Available from: http://www.asecho.org/freepdf/FeigenbaumChapter.pdf.

2. 2Edler I. Diagnostic use of ultrasound in heart disease. Acta Med Scand. 1955;308:32–38.

3. 3Edler I, Gustafson A, Karlefors T, Christensson B. Ultrasound cardiography. Acta Med Scand. 1961;370:68–74.

4. 4Goldberg BB, Ostrum BJ, Isard JJ. Ultrasonic determination of pericardial effusion. J Am Med Assoc. 1967;202:103–108.

5. 5Klein JJ, Segal BL. Pericardial effusion diagnosed by reflected ultrasound. Am J Cardiol. 1968;22:57–60. MEDLINE | CrossRef

6. 6Moss A, Bruhn F. The echocardiogram: an ultrasound technique for the detection of pericardial effusion. New Engl J Med. 1966;274:380–388. MEDLINE

7. 7Rothman J. Ultrasonic diagnosis of pericardial effusion. Circulation. 1967;35:358–363. MEDLINE

8. 8Feigenbaum H, Waldhausen JA, Hyde LP. Ultrasound diagnosis of pericardial effusion. JAMA. 1965;107:108–111.

9. 9Feigenbaum H, Zaky A, Waldhausen JA. Use of ultrasound in the diagnosis of pericardial effusion. Ann Intern Med. 1966;65:443–450. MEDLINE

10. 10Feigenbaum H. Echocardiographic diagnosis of pericardial effusion. Am J Cardiol. 1970;26:475–479. MEDLINE | CrossRef

11. 11Feigenbaum H, Zaky A, Waldhausen JA. Use of reflected ultrasound in detecting pericardial effusion (false negative with loculated effusion). Am J Cardiol. 1967;19:84. MEDLINE | CrossRef

12. 12Feigenbaum H. Diagnostic ultrasound as an aid to the management of patients with pericardial effusion. Chest. 1969;55:59–65. CrossRef

13. 13Haeckel E. Riddle of the universe at the close of the nineteenth century. 1899;Available from: http://en.wikipeia.org/wiki/Recapitulation_theory..

14. 14Kishore KS. The heart of structural development: the functional basis of the location and morphology of the human vascular pump. J Postgrad Med. 2003;49:282–284. MEDLINE

15. 15Maisch B, Seferovic PM, Ristic AD, Erbel R, Rienmuller R, Adler Y, et al. Guidelines on the diagnosis and management of pericardial diseases executive summary: the task force on the diagnosis and management of pericardial diseases of the European Society of Cardiology. Eur Heart J. 2004;587–610.

16. 16Ansari A, Rholl AO. Pseudopericardial effusion: echocardiographic and computerized tomographic correlations. Clin Cardiol. 1986;9:551. MEDLINE | CrossRef

17. 17Drummond JB, Seward JB, Tsang TS, Hayes SN, Miller FA. Outpatient two-dimensional echocardiography-guided pericardiocentesis. J Am Soc Echocardiogr. 1998;11:433–435. Abstract | Full Text | Full-Text PDF (155 KB) | CrossRef

18. 18Gonzalez MS, Basnight MA, Appleton CP. Experimental pericardial effusion: relation of abnormal respiratory variation in mitral flow velocity to hemodynamics and diastolic right heart collapse. J Am Coll Cardiol. 1991;17:239–248. MEDLINE

19. 19Picard MH, Sanfilippo AJ, Newell JB, Rodriguez L, Guerrero JL, Weyman AE. Quantitative relation between increased intrapericardial pressure and Doppler flow velocities during experimental cardiac tamponade. J Am Coll Cardiol. 1991;18:234–242. MEDLINE

20. 20Gillam 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. MEDLINE

21. 21Leimgruber P, Klopfenstein HS, Wann LS, Brooks HL. The hemodynamic derangement associated with right ventricular diastolic collapse in cardiac tamponade: an experimental echocardiographic study. Circulation. 1983;68:612–620. MEDLINE

22. 22Singh S, Wann LS, Schuchard GH. Right ventricular and right atrial collapse in patients with cardiac tamponade–a combined echocardiographic and hemodynamic study. Circulation. 1984;70:966–972. MEDLINE

23. 23Schnittger I, Bowden RE, Abrams J, Popp RL. Echocardiography: pericardial thickening and constrictive pericarditis. Am J Cardiol. 1978;42:388–395. MEDLINE | CrossRef

24. 24Hatle LK, Appleton CP, Popp RL. Differentiation of constrictive pericarditis and restrictive cardiomyopathy by Doppler echocardiography. Circulation. 1989;79:357–370. MEDLINE

25. 25Oh JK, Tajik AJ, Seward JB. Diagnostic role of Doppler echocardiography in constrictive pericarditis. J Am Coll Cardiol. 1994;23:154–162. MEDLINE

26. 26Oh JK, Tajik AJ, Appleton CP, Hatle LK, Nishimura RA, Seward JB. Preload reduction to unmask the characteristic Doppler features of constrictive pericarditis: a new observation. Circulation. 1997;95:796–799. MEDLINE

27. 27Rajagopalan N, Garcia MJ, Rodriguez L. Comparison of new Doppler echocardiographic methods to differentiate constrictive pericardial heart disease and restrictive cardiomyopathy. Am J Cardiol. 2001;87:86–94. Abstract | Full Text | Full-Text PDF (1008 KB) | CrossRef

28. 28Hatle LK, Appleton CP, Popp RL. Differentiation of constrictive pericarditis and restrictive cardiomyopathy by Doppler echocardiography. Circulation. 1989;79:357–370. MEDLINE

29. 29Von Bibra H, Scjpber K, Jenni R. Diagnosis of constrictive pericarditis by pulsed Doppler echocardiography of the hepatic vein. Am J Cardiol. 1989;63:483–488. MEDLINE | CrossRef

30. 30Schninger I, Bowden RE, Abrams J, Popp RL. Echocardiography: pericardial thickening and constrictive pericarditis. Am J Cardiol. 1978;42:388–395. MEDLINE | CrossRef

31. 31Sun JP, Abdalla IA, Yang XS. Respiratory variation of mitral and pulmonary venous Doppler flow velocities in constrictive pericarditis before and after pericardiectomy. J Am Soc Echocardiogr. 2001;14:1119–1126. Abstract | Full-Text PDF (1682 KB) | CrossRef

32. 32Abdalla IA, Murray RD, Lee JC. Does rapid volume loading during transesophageal echocardiography differentiate constrictive pericarditis from restrictive cardiomyopathy?. Echocardiography. 2002;19:125–134. MEDLINE

33. 33Rajagopalan N, Garcia MJ, Rodriguez L, Murray RD, Apperson-Hansen C, Stugaard C, et al. Comparison of new Doppler echocardiographic methods to differentiate constrictive pericardial heart disease and restrictive cardiomyopathy. Am J Cardiol. 2001;87:86–94. Abstract | Full Text | Full-Text PDF (1008 KB) | CrossRef

34. 34Oh JK, Seward JB, Tajik AJ. The echo manual. In: Philadelphia: Lippincott, Williams and Wilkins; 2006;p. 302–304.

a Wisconsin Heart Hospital, Milwaukee, Wisconsin

b Advocate Lutheran General Hospital, Chicago, Illinois.

Corresponding Author InformationReprint requests: Samuel Wann, MD, Wisconsin Heart Hospital, 10000 Bluemound Rd, Milwaukee, WI 53226.

PII: S0894-7317(07)00808-5

doi:10.1016/j.echo.2007.11.003


View previous. 7 of 28 View next.