Coronary Sinus Compression

ORIGINAL ARTICLE
Coronary Sinus Compression: An Early

Computed Tomographic Sign of Cardiac Tamponade
Menachem M. Gold, MD,* Hugo Spindola-Franco, MD,* Vineet R. Jain, MD,* Daniel M. Spevack, MD,
and Linda B. Haramati, MD*
Objective: To determine retrospectively the distinguishing features of
cardiac tamponade on conventional chest computed tomography (CT).
Materials and Methods: Blinded retrospective analysis of CT scans
from 14 patients (6 women, 8 men; age range, 49Y93 yrs; mean age, 71
yrs) with echocardiographic evidence of tamponade and 15 controls
(11 women, 4 men; age range, 37Y96 yrs; mean age, 66 yrs) without
tamponade was performed by 3 cardiothoracic radiologists. Computed
tomographic scans were analyzed for right ventricular flattening,
contrast reflux into the azygos vein, and coronary sinus compression.
Inferior vena cava (IVC) and superior vena cava short-axis diameter
and pericardial fluid attenuation were recorded. If the pericardium or
pericardial fluid was sampled, results were noted. Case and control
group variables were compared using the Fisher exact test and the
t test. Results were also subjected to logistic regression analysis.
Results: Coronary sinus compression was present in 46% (6/13)
patients with tamponade and in no controls (P = 0.006). Trends
toward IVC dilatation and elevation of pericardial fluid attenuation in
cases of tamponade did not reach statistical significance. A specific
pathological diagnosis was made in 88% (7/8) of tamponade cases
and 29% (2/7) of controls (P = 0.04).
Conclusions: The detection of coronary sinus compression on CT is
an early specific indicator of cardiac tamponade. Dilatation of the
IVC and the presence of elevated pericardial fluid attenuation are CT
signs suggestive of the diagnosis.
Key Words: tamponade, coronary sinus, computed tomography
(J Comput Assist Tomogr 2008;32:72Y77)
ardiac tamponade is a potentially life-threatening clinical

condition caused by compression of the heart by fluid in
the pericardial space, resulting in hemodynamic compromise.1
Increasing intrapericardial pressure leads to restriction of
ventricular filling and a reduction of stroke volume and cardiac
output. The diagnosis is typically suspected based on a
characteristic constellation of clinical findings, including
tachycardia, hypotension, distended neck veins, pulsus paradoxus, and indistinctness of cardiac sounds.2 In patients with
suspected tamponade, transthoracic echocardiography has
From the Departments of *Radiology and Internal Medicine, Division of

Cardiology, Albert Einstein College of Medicine, Montefiore Medical
Center, Bronx, NY.
Received for publication July 19, 2006; accepted August 31, 2006.
Reprints: Linda B. Haramati, MD, Department of Radiology, Albert Einstein
College of Medicine, Montefiore Medical Center, 111 210th St, Bronx,
NY 10467 (e-mail: lharamat@montefiore.org).
Copyright * 2008 by Lippincott Williams & Wilkins
72
emerged as the diagnostic study of choice because of its

portability, high sensitivity for pericardial fluid, and real time
visualization of the cardiac chambers.
The clinical manifestations of tamponade depend on the
rapidity of fluid accumulation and the effectiveness of
compensatory mechanisms. Additionally, most symptoms
and physical findings in tamponade are nonspecific.2,3
Consequently, patients with tamponade are often initially
imaged with modalities other than echocardiography such as
chest computed tomography (CT). Although conventional CT
provides only limited hemodynamic information, it has been
shown to diagnose pericardial effusions accurately, and the
attenuation value can point to the nature of the fluid.4
Several CT signs suggestive of tamponade have been
described in case reports and small series. Doppman et al5
concluded that pericardial effusions producing tamponade do
not deform the right ventricular contour but always enlarge the
inferior vena cava (IVC). Subsequent authors have countered
with examples of tamponade causing flattening of the anterior
heart border6 and severe compression of the right atrium and
right ventricle with little, if any, enlargement of the IVC.7
Others assert that secondary CT signs of elevated central
venous pressure can serve as a clue to the diagnosis. These
include reflux of contrast into the azygos vein8 and IVC
dilatation.9 Finally, coronary sinus compression as a specific
indicator of tamponade has been described on coronary
arteriography.10
To our knowledge, a systematic comparison of the CT
findings of pericardial effusion with and without tamponade
has not been performed. The purpose of this study, therefore,
was to determine retrospectively the distinguishing features of
cardiac tamponade on conventional chest CT.
MATERIALS AND METHODS

Study Patients
Sixty-three cases of cardiac tamponade were retrospectively identified by review of transthoracic echocardiogram reports between February 2000 and December 2004. Of
these, 15 cases had chest CT scans within 72 hours of the
echocardiogram available for analysis. Over the same period,
15 control cases of pericardial effusions and no tamponade
with chest CT scans within 72 hours of the echocardiogram
were identified. Upon further review, 2 of the 15 tamponade
studies were recognized as belonging to the same patient who
had 2 episodes of tamponade separated by a period of
13 months. The second tamponade episode was therefore
excluded from the data. As a result, 14 CT studies performed
J Comput Assist Tomogr
& Volume 32, Number 1, January/February 2008
Copyright @ 2008 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Coronary Sinus Compression
TABLE 1. Comparison of CT Findings in Patients With Pericardial Tamponade and Controls

Characteristics
Right ventricular flattening
Conrast reflux into azygos vein
Coronary sinus compression
IVC diameter (cm), mean T SD
SVC diameter (cm), mean T SD
Effusion attenuation (HU), mean T SD
Specific pericardial diagnosis
Temponade Group
Control Group
36% (5/14)
14% (1/7)
46% (6/13)
2.39 T 0.38
1.74 T 0.28
19.40 T 10.21
88% (7/8)
7% (1/15)
40% (2/5)
0% (0/14)
2.18 T 0.20
1.83 T 0.32
13.58 T 9.31
29% (2/7)
0.081
0.546
0.006*
0.067
0.415
0.059
0.041*
*Statistical significance.
in 14 patients (6 women, 8 men; age range, 49Y93 yrs; mean

age, 71 yrs) with tamponade and 15 CT studies performed in
15 control patients (11 women, 4 men; age range, 37Y96 years;
mean age, 66 years) were available for review and analysis.
Our institutional review board approved this retrospective
study; patient informed consent was not required.
Control cases were matched with the tamponade cases
according to the sizes of the effusions. This was done to ensure
that the size of the effusion did not present a bias in determining the presence or absence of tamponade. Six effusion
size categories were used based on the size description
given in the echocardiogram report. The breakdown of effusion sizes for each group was as follows: minimal (n = 1),
small (n = 1), small-to-moderate (n = 3), moderate (n = 4),
moderate-to-large (n = 2), and large (n = 3). The 15th control
had a small effusion.
Chart review of all the cases and controls demonstrated
no significant change in the clinical status of the patients
between the time of the echocardiogram and the chest CT.

Furthermore, no intervention was performed between the
echocardiogram and the CT scan. None of the patients
suffered acute trauma or had recent heart surgery. Only one of
the tamponade patients was clinically unstable at presentation.
None of the control patients was clinically unstable.
Image Acquisition
Computed tomographic scans were performed on a
variety of helical scanners with a variety of techniques.
Collimation ranged from 1.3 to 7 mm and interscan spacing
from 0.7 to 7 mm in both groups. Seven of the tamponade
cases and 5 of the controls were performed with administration
of intravenous contrast.
Image Analysis
Blinded retrospective analysis of each CT was jointly
performed by a group of 3 fellowship-trained cardiothoracic
radiologists. Differences were resolved by consensus. Each CT
scan was evaluated for the presence or absence of the following
previously described signs of tamponade: right ventricular wall
flattening, reflux of contrast into the azygos vein, and coronary
sinus compression. Right ventricular wall flattening and
coronary sinus compression were assessed subjectively. The
diameters of the IVC and superior vena cava (SVC) and
the attenuation of the effusion in Hounsfield units (HU) were
recorded for each case. Inferior vena cava and SVC short-axis
vessel diameters were measured just below and just above their
junctions with the right atrium, respectively. If the pericardium
or pericardial fluid was sampled, results were recorded.
Statistical Analysis
FIGURE 1. A 66-year-old man with a moderate pericardial

effusion and echocardiographic evidence of tamponade.
Axial contrast-enhanced CT image demonstrates compression
of the coronary sinus (arrows).
The Fisher exact test was used to compare case and

control groups with regard to categorical variables. The
unpaired Student t test was used to compare continuous
variables between both groups. Multiple logistic regression
analysis was performed with true diagnosis (tamponade or
absence of tamponade) as the dependent variable with the
following independent variables: flattening of the right heart
border (present or absent), reflux of contrast into the azygos
vein (present or absent), coronary sinus compression (present
or absent), IVC diameter, SVC diameter, and effusion
attenuation. Differences were considered statistically significant at a P G 0.05. All statistics were calculated by using a
statistical software package (SAS version 9.1.2; SAS Institute,
Cary, NC).
* 2008 Lippincott Williams & Wilkins
73
Gold et al
RESULTS
Patient Characteristics
There was no significant statistical difference between
the tamponade and control groups with regard to patient age
(P = 0.333) or sex (P = 0.139).
Effusion Characteristics
Among the categorical CT features analyzed (Table 1),
only the presence of coronary sinus compression was found
to be statistically significant (P = 0.006). The coronary sinus
was narrowed in 46% (6/13) of tamponade cases (Figs. 1 and 2)
and in none of the controls (Figs. 3 and 4). The coronary sinus
was not well visualized in 1 case and in 1 control. Among the
continuous variables assessed, IVC dilatation approached, but
did not reach, statistical significance. Logistic regression
analysis demonstrated a correlation between increasing effusion attenuation and tamponade that approached significance.
The mean time differential between the echocardiogram
and CT in both case and control groups was 24 hours. Of the
6 patients in whom coronary sinus compression was identified,
3 received contrast and 3 did not. Coronary sinus compression
was noted on CT scans with slice thicknesses of 1.5 (n = 1), 5
(n = 4), and 7 mm (n = 1).
Among those whose pericardial fluid or tissue was
sampled, a specific pathological diagnosis was made in 88%
(7/8) of tamponades and 29% (2/7) of controls (P = 0.04 1). Of
the 8 tamponade cases sampled, 4 had inflammatory
pericarditis (reactive mesothelial cells [n = 2], acute inflammatory cellular infiltration [n = 1], perivascular lymphocytic
infiltration [n = 1]), 3 had metastatic carcinoma, and 1 was
negative. Among the 7 controls sampled, 5 were negative and
2 had inflammatory pericarditis (acute pericarditis [n = 1],
chronic inflammatory cells [n = 1]).
Computed tomographic imaging of the single clinically
unstable patient with tamponade demonstrated right ventricu-
FIGURE 2. An 83-year-old man with a small-to-moderate

pericardial effusion and echocardiographic evidence of
tamponade. Unenhanced CT image shows narrowing of the
coronary sinus (arrows).
74
FIGURE 3. A 74-year-old woman with a moderate pericardial

effusion but no echocardiographic evidence of tamponade.
Axial contrast-enhanced CT image demonstrates a normal
coronary sinus (arrows).
lar flattening, coronary sinus compression, and IVC dilatation

(Fig. 5). Pericardial fluid attenuation in this patient was 23 HU,
and fluid analysis showed acute inflammatory cells.
DISCUSSION
Cardiac tamponade is a potentially life-threatening
condition that requires prompt recognition and intervention.
The ability to establish the correct diagnosis necessitates
integration of symptomatic manifestations, physical signs, and
imaging features. The broad spectrum of findings in tamponade is a direct reflection of its underlying pathophysiology.
Our understanding of tamponade pathophysiology has
evolved. In their original classic description, Reddy et al11
proposed that tamponade was an Ball-or-none[ phenomenon. In
this model, pericardial pressure began to rise independently of
right and left ventricular filling pressures. As long as pericardial
pressure did not exceed a critical level, namely, right ventricular
filling pressure, no hemodynamic changes occurred.
Based on subsequent hemodynamic observations, Reddy
et al12 revised their original model. The new model depicts
tamponade as a continuum of hemodynamic abnormalities that
progresses through 3 phases. In phase 1, as pericardial pressure
begins to rise, both the right and left ventricles begin to stiffen
and demonstrate higher-than-normal filling pressures. At this
early stage, the right and left ventricular filling pressures remain
higher than the pericardial pressure. Right atrial pressure also
rises due to compensatory increases in systemic venous
pressure. The compensatory elevated venous pressure generally
maintains a normal cardiac output.
Phase 2 is heralded by the equilibration of pericardial
pressure with the elevated right ventricular filling pressure.
Despite further compensatory increases in venous pressure
and elevation of heart rate, ventricular filling cannot be
maintained, and cardiac output declines. Inspiratory decreases
in systolic blood pressure are often manifested in this phase
and may even reach diagnostic criteria for pulsus paradoxus.
FIGURE 4. A 68-year-old woman with a minimal pericardial

effusion and no echocardiographic evidence of tamponade.
Unenhanced CT image shows a normal coronary sinus
(arrows).
When pericardial pressures rise high enough to equilibrate

with left ventricular filling pressures (phase 3), hemodynamic
collapse is imminent.
In this study, 13 of the 14 cases with echocardiographic
evidence of tamponade were clinically stable at the time of
initial imaging. It is well documented that right atrial and right
ventricular collapse may be seen on echocardiography in
patients who have pericardial effusions but no clinical evidence
of cardiac tamponade.13Y15 This can be explained in light of our
current understanding of tamponade physiology as a continuum of hemodynamic abnormalities. In the early phases of
tamponade, patients may have subtle, compensated, hemodynamic changes that are visualized on echocardiography but are
not manifested clinically. As the effusion becomes more

hemodynamically embarrassing, compensatory neurohormonal mechanisms intensify, leading to the clinical findings of
tachycardia, dyspnea, and peripheral hypoperfusion.16
The current study, consisting of patients in the early
stages of tamponade, did not confirm the reliability of any
previously reported CT signs of tamponade. The only sign
specific for tamponade in our series was coronary sinus
compression, which was present in 6 of 13 patients with
tamponade (46%) and in no case of isolated pericardial
effusion (0/14). In evaluating the coronary sinus, we used a
qualitative visual impression, rather than a quantitative
measurement on a single slice for several reasons. In an
early description of the coronary sinus on CT, Micklos and
Proto17 concluded that the tubular sinus can have a variable
orientation with respect to the cross-sectional image. Based on
its orientation, the sinus may have any number of appearances,
sizes, and shapes on a given slice. Subsequent detailed
anatomical analysis of the coronary sinus has demonstrated
marked variability in the drainage pattern of its tributaries,
which contribute to its caliber at varying points along its
span.18 Based on the above findings, we decided that because a
standardized diameter measurement of the coronary sinus is
not feasible on an axial CT slice, a visual assessment of the
sinus would more accurately reflect its overall caliber.
Coronary sinus compression has previously been
described in a case report as an angiographic sign of
tamponade that can be seen even before hemodynamic or
echocardiographic evidence of tamponade.10 Those authors
hypothesized that as intrapericardial pressure rises, the thinwalled midportion of the coronary sinus is compressed before
the more muscular right atrium and ventricle. Based on the
above model of tamponade pathophysiology, an alternate
hypothesis for how this sign can precede hemodynamic or
clinical signs of tamponade can be proffered. Although both
the coronary sinus and right atrium are subject to the same
rising external pericardial pressure, the right atrial distending
pressure also rises initially as a result of the body_s
FIGURE 5. A 49-year-old man who presented with hypotension posthemodialysis. Transthoracic echocardiography demonstrated
a moderate pericardial effusion and tamponade. A, Axial-enhanced CT image demonstrates dilatation of the IVC (white arrows)
and right ventricular flattening (black arrows). B, Axial-enhanced CT image at more caudal level (than A) shows compression of
the coronary sinus (black arrows).
75
Gold et al
compensatory elevations in systemic venous pressure. The

elevated right atrial distending pressure may not be transmitted
to the coronary sinus, however, due to the thebesian valve, a
small fold of endocardium at the ostium of the coronary sinus.
This valve helps to prevent reflux into the coronary sinus from
the right atrium.19 Thus, the coronary sinus would be the first
structure to collapse under rising pericardial pressure.
In 1991, Jatene et al published an anatomical study of
the thebesian valve in 94 human hearts.20 They found that the
valve was absent in 16% of specimens, Bresidual[ (G15% of
coronary sinus orifice area) in 31%, Bpartial[ (915% of orifice
area) in 45%, and trabeculated or double in 8% of specimens.
If our hypothesis regarding coronary sinus compression is
correct, the reason it is not seen in all cases of tamponade may
be because of absence or hypoplasia of the thebesian valve in a
significant proportion of patients, as reported by Jatene et al. A
widely patent ostium would allow transmittal of elevated
venous pressures into the coronary sinus, thus preventing its
early collapse.
In our series, tamponade also correlated with the
presence of a specific pathological diagnosis (P = 0.04). Of
patients who had their pericardium or pericardial fluid
sampled, those with a specific diagnosis were 3.5 times more
likely to have tamponade. Similarly, the higher the attenuation
of the pericardial fluid on CT, the greater the likelihood the
patient had tamponade, although this trend did not quite reach
statistical significance (P = 0.06).
Discrepancies in the literature exist as to the association
between IVC dilatation and tamponade on CT.5,7 Our data
demonstrate a trend toward IVC dilatation in cases of
tamponade that did not reach statistical significance (P =
0.07). Within the continuum of hemodynamic changes in
tamponade, elevation of systemic venous pressures is one of
the earliest compensatory responses. As a relatively high
compliance vascular structure with a nonmuscular wall, the
IVC dilates early. This sign has limited practical value in
diagnosing tamponade when prospectively evaluating a CT for
2 reasons. One, a large overlap in IVC diameter exists between
the tamponade and control groups. In addition, this finding is
nonspecific and can be seen with elevated right-sided heart
pressures due to any cause.
There was no statistically significant difference in mean
SVC diameter between the 2 groups, confirming the lack of
correlation between SVC diameter and tamponade previously
reported by Doppman et al.5 The reason the SVC does not
dilate as readily as the IVC in the setting of elevated central
pressure may be anatomical in origin. Examination of human
pathological specimens has shown that the wall of the SVC is
fixated to the deep layer of the serous pericardium, contributing to growth in resistance during the elevation of venous
pressure and making possible the preservation of integrity of
the SVC.21
The limitations of this series include its retrospective
nature, small sample size, and range of CT techniques.
Additionally, our series did not include any cases of
tamponade due to trauma, a major cause of tamponade.22
Because only 1 patient in our series had clinical manifestations
of tamponade, the scope of our conclusions is limited to those
patients in the early stages of tamponade. In such patients, not
76
all of the previously reported CT signs of tamponade may yet

be manifested. Indeed, perusal of the radiology literature to
date reveals that in almost all the CT tamponade cases
demonstrating flattening of the right heart border or secondary
signs of elevated central venous pressure such as contrast
reflux into the azygos vein, overt clinical signs of tamponade
were present in the setting of blunt trauma.
Important information concerning the heart can be
gleaned from a thoracic CT even when performed without
cardiac gating.23,24 It is clear from our series that previously
described CT signs of cardiac tamponade provide limited
assistance in differentiating between patients with pericardial
effusions and no tamponade and those in the early phases of
tamponade. However, the detection of coronary sinus
compression on CT is an early specific indicator of cardiac
tamponade. Dilatation of the IVC and the presence of elevated
pericardial fluid attenuation are CT signs suggestive of the
diagnosis, but are nonspecific. Patients with such findings
should be closely monitored to ensure that hemodynamic
compromise does not progress.
ACKNOWLEDGMENT
The authors thank Katherine D. Freeman, DrPH, for her
help with the statistical analysis.
REFERENCES
1. Starling EH. Some points in the pathology of heart disease. Lancet.
1897;1:652Y655.
2. Spodick DH. Acute cardiac tamponade. N Engl J Med. 2003;349:
684Y690.
3. Spodick DH. Pericardial diseases. In: Braunwald E, Zipes DP, Libby P,
eds. Heart Disease: A Textbook of Cardiovascular Medicine.
Philadelphia, PA: W.B. Saunders, 2001:1823Y1876.
4. Tomoda H, Hoshiai M, Furuya H, et al. Evaluation of pericardial
effusion with computed tomography. Am Heart J. 1980;99:701Y706.
5. Doppman JL, Rienmuller R, Lissner J, et al. Computed tomography in
constrictive pericardial disease. J Comput Assist Tomogr. 1981;5:1Y11.
6. Hernandez-Luyando L, Calvo J, Gonzalez de las Heras E, et al. Tension
pericardial collections: sign of flattened heart in CT. Eur J Radiol.
1996;23:250Y252.
7. Jones BV, Vu D. Diagnosis of posttraumatic pericardial tamponade by
plain film and computed tomography and control of bleeding by
embolotherapy of the left inferior phrenic artery. Cardiovasc Intervent
Radiol. 1993;16:183Y185.
8. Harries SR, Fox BM, Roobottom CA. Azygos reflux: a CT sign of
cardiac tamponade. Clin Radiol. 1998;53:702Y704.
9. Goldstein L, Mirvis SE, Kostrubiak IS, et al. CT diagnosis of acute
pericardial tamponade after blunt trauma. AJR Am J Roentgenol.
1989;152:739Y741.
10. Steiner MA, Marshall JJ. Coronary sinus compression as a sign of
pericardial tamponade. Cathet Cardiovasc Intervent. 2000;49:455Y458.
11. Reddy PS, Curtiss EL, O_Toole JD, et al. Cardiac tamponade:
hemodynamic observations in man. Circulation. 1978;58:265Y272.
12. Reddy PS, Curtiss EL, Uretsky BF. Spectrum of hemodynamic changes
in cardiac tamponade. Am J Cardiol. 1990;66:1487Y1491.
13. Gillam LD, Guyer DE, Gibson TC, et al. Hemodynamic compression
of the right atrium: a new echocardiographic sign of cardiac tamponade.
Circulation. 1983;68:294Y301.
14. Armstrong WF, Schilt BF, Helper DJ, et al. Diastolic collapse of the
right ventricle with cardiac tamponade. Circulation. 1982;65:1491Y1496.
15. Shiina A, Yaginuma T, Kondo K, et al. Echocardiographic evaluation
of impending tamponade. J Cardiogr. 1978;9:555.
16. Goldstein JA. Cardiac tamponade, constrictive pericarditis, and
restrictive cardiomyopathy. Curr Probl Cardiol. 2004;29:503Y567.
17. Micklos TJ, Proto AV. CT demonstration of the coronary sinus.

J Comput Assist Tomogr. 1985;9:60Y64.
18. Sousa-Rodrigues CF, Alcantara FS, Olave E. Topography and biometry of
the coronary venous system and its tributaries. Int J Morphol.
2005;23:177Y184.
19. Williams PL, Warick R, Dyson M, et al, eds.
Gray_s Anatomy. 37th ed. New York, NY: Churchill-Livingstone,
1989:703.
20. Jatene M, Jatene F, Costa R, et al. Anatomic study of the coronary sinus
valve-Thebesius valve. Chest. 1991;100(Suppl.):905.
21. Pilfer CR, Puffer MI, Santi FP, et al. Structural characteristics of
the superior venae cavae wall at the intrapericardium segment in

adults and aging individuals. Okajimas Folia Anat Jpn. 1996;73:
89Y100.
22. Rizzo AG, Steinberg SM, Flint LM. Prospective assessment of the
value of computed tomography for trauma. J Trauma. 1995;38:338Y342.
23. Bruzzi JF, Remy-Jardin M, Delhaye D, et al. When, why, and how to
examine the heart during thoracic CT: Part 1, basic principles. AJR Am J
Roentgenol. 2006;186:324Y332.
24. Bruzzi JF, Remy-Jardin M, Delhaye D, et al. When, why, and how to
examine the heart during thoracic CT: Part 2, clinical applications. AJR
Am J Roentgenol. 2006;186:333Y341.
77

Coronary Sinus Compression

Încărcat de

Informații document

Drepturi de autor

Formate disponibile

Partajați acest document

Partajați sau inserați document

Opțiuni de partajare

Vi se pare util acest document?

Este necorespunzător acest conținut?

Drepturi de autor:

Formate disponibile

Coronary Sinus Compression

Încărcat de

Drepturi de autor:

Formate disponibile

ORIGINAL ARTICLE