Improved detection with transesophageal George D. Birmingham, Madison,
of infective endocarditis echocardiography
MD, Peter S. Rahko,
MD, and Ford Ballantyne
Infective endocarditis continues to be a serious problem with a significant morbidity and mortality despite current therapy.l The signs and symptoms are varied leading to a wide differential diagnosis. History, physical examination, and laboratory assessment are frequently inconclusive, and the diagnosis is often made by combining clinical and laboratory data. Blood cultures may be negative in 5 % to 15% of cases, frequently because of prior use of antibiotics.2-5 The clinical course can change rapidly, necessitating an early diagnosis. These issues confirm the need for a sensitive and specific test that can accurately diagnose endocarditis. Echocardiography is a widely available and useful modality for evaluating suspected endocarditis. Standard M-mode and two-dimensional sensitivities for visualization of vegetations range from 30% to 80% .6-11 Echocardiography also gives valuable information on associated complications, as well as on cardiac function. Echocardiographic identification of a valvular vegetation in patients with active infective endocarditis is associated with a higher incidence of complications including embolization, congestive heart failure, need for surgical intervention, and death.g-15 Thus identifying a vegetation by echocardiography not only confirms the diagnosis but also allows for the timely institution of appropriate therapy. There are significant limitations to standard transthoracic echocardiography for the detection of vegetations. Both false positive and false negative findings lessen its clinical usefulness.ls l6 Echocardiograms of satisfactory visual quality as to be clinically useful are difficult to obtain in some patients, particularly those with obesity, chest wall deformity, lung From The Section Wisconsin Medical
of Cardiology, School.
Received
for publication
July
Reprint University
requests: Peter of Wisconsin
S. Rahko, Hospital,
4/l/34516 774
III,
MD.
Wis.
Department 30, 1991;
accepted
MD, Section 600 Highland
of Medicine, Sept.
University
of
20, 1991.
of Cardiology, H6/334 CSC, Ave., Madison. WI 53792.
disease, and the inability to be properly positioned for adequate interrogation. For patients with prosthetic valves even high-quality findings may be difficult to interpret, since prosthetic devices cause imaging artifacts and masking that can interfere with the ultrasonic interrogation of portions of the heart.17 False positive findings of vegetations are a wellknown hazard in the evaluation of myxomatous mitral valves seen in the mitral valve prolapse syndrome.ls The sonographic detection of a vegetation is influenced by several factors: echogenicity of the lesion, size of the vegetation, presence of preexisting valvular disease, presence of a prosthetic valve, sensitivity of the instrument, skill and experience of the sonographer, and the a priori likelihood of endocarditis.l, 14,lg As a consequence 20 % to 30 % of patients may have nondiagnostic transthoracic echocardiograms. The recent introduction of transesophageal echocardiography has provided a new approach to the diagnosis of endocarditis. Several recent studies have demonstrated the usefulness of this technique in diagnosing not only vegetations but also complications of endocarditis.8, lg, 20-27 Our study was designed to directly compare the sensitivity and specificity of transesophageal echocardiography with that of standard transthoracic echocardiography in the evaluation of endocarditis. We also sought to determine the incremental value of transesophageal echocardiography over transthoracic studies in various clinical situations. METHODOLOGY
The population reviewed consisted of all patients referred to the Adult Echocardiography Laboratory of the University of Wisconsin for evaluation of suspected infective endocarditis between July 1988 and August 1990. An initial transthoracic study was performed on all patients. If this initial study was sufficiently diagnostic, we did not proceed to a transesophageal study because we felt it would be unethical to do SO. Transesophageal echocardiography was
Volume Number
123 3
Table I. Clinical data evaluated for determination presence of infective endocarditis Historical information History of preexisting valvular disease Presence of a prosthetic valve History of prior endocarditis History of intravenous drug abuse Recent procedure likely to cause transient or sustained remia Antibiotic use within 2 weeks before onset of symptoms Symptoms Fatigue Fever/Chills Sweats Myalgia/Arthralgia Headache/Neurologic change Anorexia/Weight loss Physical examination findings Fever 038.3 C) Any murmur Congestive heart failure (rales, Ss, increased central pressure) Janeway’s lesions Osler’s nodes Roth’s spots Splinter hemorrhage Splenomegaly Embolization Laboratory data Hematuria Proteinuria Anemia (hemoglobin lO,OOO K/wl) Elevated erythrocyte sedimentation rate (>25 mm/hr) Increased platelets (>400,000 K/ml) Abnormal ECG Abnormal chest x-ray Positive blood cultures
TEE detection
of the
endocarditis
775
Table II. Sensitivity and specificity of clinical factors that were significantly more common in patients with endocarditis
compared
to those
Factor
bacte-
of
P*
Symptoms of fever Positive blood culture Any murmur Preexisting valvular disease Headache/Neurologic change Sweats Physical findings of embolization Splinter hemorrhage Janeway’s lesions *Represents
comparison
without
between
0.005 0.001 0.0005 0.003 0.02 0.04 0.01 0.05 0.04
Sensitivity (967) 90 84 81 71 52 48 26 23 13
patients with and without
Specificity (%) 40 57 63 67 77 23 97 100 100 endocarditis.
venous
performed if the diagnostic quality of the transthoracic study was inadequate, when the transthoracic study was inconclusive, or when the presence of an abnormality required further evaluation. Thus the study population consisted of 61 patients who had paired transthoracic and transesophageal echocardiograms, as well as clinical information available for review. Ages of the patients ranged from 17 to 89 (mean 52) years. There were 38 men and 23 women. Echocardiographic examinations were performed up to 22 days apart (mean 4.3 days). Studies were performed on Hewlett-Packard Sonos 500 or 1000 ultrasound imaging systems (Hewlett-Packard Co. Medical Products Co., Andover, Mass.) with a 2.5 or 3.5 MHz transducer. Transesophageal studies were performed with a 5.0 MHz transducer mounted on the tip of a gastroscope. After giving informed consent patients fasted for at least 4 hours before the proce-
dure, and a local anesthetic was applied to the pharynx before insertion of the transducer. Sedation was used at the discretion of the physician performing the study. No complications occurred during any of the transesophageal studies. The medical record of each patient was reviewed for specific historical, symptomatic, physical, and laboratory data (Table I). The diagnosis of infective endocarditis was established by means of these clinical data and the criteria of von Reyn et a1.3 One modification of the von Reyn criteria was made since the patients studied were predominantly a referral population. The presence of any murmur was substituted for the finding of a new regurgitant murmur. Eight patients were evaluated by surgical or autopsy criteria. The frequency with which positive individual clinical signs occurred was calculated. Results of each paired transthoracic and transesophageal study were reviewed by two experienced echocardiographers blinded to the clinical data. The studies were scored for image quality as poor (l+), fair (2+), good (3+), or excellent (4+). Findings in the studies were also graded as positive (definitely present) or negative (possibly or not present) for the presence of a vegetation. When vegetations were visualized they were classified according to location and size. Dimensions were graded as small (lO mm). Underlying valve disease and complications resulting from infective endocarditis were also noted. Statistical analysis of the categoric data was performed by means of chi-square analysis or Mann-Whitney U test where appropriate. Sensitivity was defined as (true positive)/(true positive + false negative) X 100 and specificity as (true negative)/(true negative + false positive) X 100.
March American
Rahko, and Ballantyne
776
Birmingham,
Table
III. Blood culture results of clinically positive casesof infective endocarditis Affected Blood
Mitral
culture
1
2
3 6 2
4 4 1
12
1 1 2 ii?
&treptococcus Group C streptococcus Group D streptococcus Staphylococcus aureus Coagulase-negative staphylococcus Bacillus species Cardiobacterium hominus Candida species Culture negative
were Staphylococcus
1992 Journal
valve Device
Tricuspid
Total 3 1
1
11 4 1 1 3 2 33*
1 1
*In the two patients with infections on two valves, the organisms since infections were classified by location.
Table
Aortic
Heart
2
i
4
aureu~ and group D streptococcus.
These organisms are counted twice
IV. Comparisonof image quality between transesophagealand transthoracic echocardiographicstudies Study Transthoracic Transesophageal
Poor
Fair
7 (11%) 0
12 (20%) 3 (5%)
OBSERVATIONS
Of the 61 patients, 31(51% ) had findings that were positive for infective endocarditis according to the criteria of von Reyn et a1.3There were 33 vegetations in the 31 patients, inasmuch as two patients had infection involving both mitral and aortic valves. All eight patients documented as having infective endocarditis at surgery or autopsy also had positive clinical findings. The frequency of positive findings for each clinical factor shown in Table I was determined. Table II shows clinical factors that were significantly more common in patients with infective endocarditis. Of the items listed under historical information, only a history of preexisting valvular disease was significantly related to the identification of a patient with infective endocarditis. Our sample contained few patients with a prosthetic valve or a prior history of endocarditis and no patients with a history of intravenous drug abuse. From the list of symptoms commonly associated with infective endocarditis, only fever/chills, headache/neurologic changes, and sweating were significantly correlated with the presence of infective endocarditis. From the physical examination items reviewed, any type of murmur, Janeway lesions, and systemic embolization were found significantly more often in those with infective endocarditis. Only one Roth’s spot was noted and no Osler’s nodes were found. Of the items exam-
Good 26 (43 ‘% ) 8 (13%)
Excellent 16 (26%) 50 (82%)
ined in the laboratory data, only positive blood cultures were significantly related to the presence of infective endocarditis. For the mitral valve, aortic valve, and all locations together, Table III shows that Staphylococcus species were the most common organisms followed by Streptococcus species. Table IV lists the grading of image quality for each transthoracic and transesophageal examination. Overall, transesophageal echocardiographic image quality was superior to transthoracic image quality in 39 patients (64 ‘%). In 20 patients (33 %) image quality was equivalent. The average quality score for all transthoracic studies (2.8 f 1.0) was significantly lower thanfortransesophagealstudies (3.8 f 0.5,~ < 0.01). A total of 33 vegetations were visualized (Figs. 1 and 2). Two patients had vegetations involving two valves. Both were identified correctly by transesophageal echocardiology and confirmed at surgery. Table V lists the sensitivity of transthoracic and transesophageal studies for identifying vegetations by location and for the entire group. The specificity for the 30 patients without infective endocarditis was 100 % by the transthoracic method. There was one false positive finding by the transesophageal method. For the mitral valve, transesophageal echocardiography was 100% sensitive in locating vegetations versus 50% for transthoracic studies. For the aortic valve, the sensitivity of transesophageal and transthoracic
Volume Number
123 3
TEE detection
of
endocarditis
777
Fig. 1. A, Transthoracic echocardiogramof mitral valve infective endocarditis. Underlying lesion is mitral valve prolapse.No vegetation is visible. AL, Anterior leaflet; AO, aorta; LA, left atrium; LV, left ventricle; PL, posterior leaflet; RV, right ventricle. B, Transesophagealechocardiogramof mitral valve infective endocarditis (samepatient as in A). Vegetation is clearly evident. AV, Aortic valve; LA, left atrium; MV, mitral valve; VEG, vegetation.
echocardiography was 88 5%and 25%) respectively. Overall sensitivity for transesophageal studies was 88 5’; compared with 30 5%for transthoracic studies. If only studies rated good or excellent for image quality were evaluated, the sensitivity of transthoracic echocardiography only improves to 36 95, whereas it remains 88% for a transesophageal study. Table VI lists the vegetation size of the patients with clinically positive findings. On the mitral valve, all vegetations seen by transthoracic echocardiography were large. This was in full agreement with the corresponding transesophageal results. Vegetations not detected by
transthoracic studies but visualized by the transesophageal method were medium or large. For the aortic valve two of three vegetations located by both transthoracic and transesophageal echocardiography were graded as the same size; the other case differed by one grade. Among the vegetations seen only by transesophageal echocardiography, all three sizes were represented. No comparison can be made concerning vegetations on the tricuspid valve or intravascular devices, inasmuch as none was observed by transthoracic echocardiography. There were six complications found by transesoph-
March
778
hnerican
Birmingham, Rahko, and Balluntyne
Heart
1992 Journal
Fig. 2. A, Transthoracic echocardiogram of aortic valve infective endocarditis. One leaflet of aortic valve is thickened; remainder of valve is normal. AV, Aortic valve; LA, left atrium; RVOT, right ventricular outflow tract. B, Transesophageal echocardiogram of aortic valve infective endocarditis (same patient as in A). Leaflet is more clearly visualized and vegetation is well defined. LA, Left atrium; LV, left ventricle; MV, mitral valve; RV, right ventricle; VEG, vegetation.
V. Comparison of the sensitivity transthoracic echocardiography
of vegetation identification
Table
Study TT TE
6/12 12/12
4/16
(50 0; ) (100%)
valve;
AV, aortic
valve;
TV
(25 % )
14/16
of infective endocarditis echocardiography
MD, Peter S. Rahko,
MD, and Ford Ballantyne
Infective endocarditis continues to be a serious problem with a significant morbidity and mortality despite current therapy.l The signs and symptoms are varied leading to a wide differential diagnosis. History, physical examination, and laboratory assessment are frequently inconclusive, and the diagnosis is often made by combining clinical and laboratory data. Blood cultures may be negative in 5 % to 15% of cases, frequently because of prior use of antibiotics.2-5 The clinical course can change rapidly, necessitating an early diagnosis. These issues confirm the need for a sensitive and specific test that can accurately diagnose endocarditis. Echocardiography is a widely available and useful modality for evaluating suspected endocarditis. Standard M-mode and two-dimensional sensitivities for visualization of vegetations range from 30% to 80% .6-11 Echocardiography also gives valuable information on associated complications, as well as on cardiac function. Echocardiographic identification of a valvular vegetation in patients with active infective endocarditis is associated with a higher incidence of complications including embolization, congestive heart failure, need for surgical intervention, and death.g-15 Thus identifying a vegetation by echocardiography not only confirms the diagnosis but also allows for the timely institution of appropriate therapy. There are significant limitations to standard transthoracic echocardiography for the detection of vegetations. Both false positive and false negative findings lessen its clinical usefulness.ls l6 Echocardiograms of satisfactory visual quality as to be clinically useful are difficult to obtain in some patients, particularly those with obesity, chest wall deformity, lung From The Section Wisconsin Medical
of Cardiology, School.
Received
for publication
July
Reprint University
requests: Peter of Wisconsin
S. Rahko, Hospital,
4/l/34516 774
III,
MD.
Wis.
Department 30, 1991;
accepted
MD, Section 600 Highland
of Medicine, Sept.
University
of
20, 1991.
of Cardiology, H6/334 CSC, Ave., Madison. WI 53792.
disease, and the inability to be properly positioned for adequate interrogation. For patients with prosthetic valves even high-quality findings may be difficult to interpret, since prosthetic devices cause imaging artifacts and masking that can interfere with the ultrasonic interrogation of portions of the heart.17 False positive findings of vegetations are a wellknown hazard in the evaluation of myxomatous mitral valves seen in the mitral valve prolapse syndrome.ls The sonographic detection of a vegetation is influenced by several factors: echogenicity of the lesion, size of the vegetation, presence of preexisting valvular disease, presence of a prosthetic valve, sensitivity of the instrument, skill and experience of the sonographer, and the a priori likelihood of endocarditis.l, 14,lg As a consequence 20 % to 30 % of patients may have nondiagnostic transthoracic echocardiograms. The recent introduction of transesophageal echocardiography has provided a new approach to the diagnosis of endocarditis. Several recent studies have demonstrated the usefulness of this technique in diagnosing not only vegetations but also complications of endocarditis.8, lg, 20-27 Our study was designed to directly compare the sensitivity and specificity of transesophageal echocardiography with that of standard transthoracic echocardiography in the evaluation of endocarditis. We also sought to determine the incremental value of transesophageal echocardiography over transthoracic studies in various clinical situations. METHODOLOGY
The population reviewed consisted of all patients referred to the Adult Echocardiography Laboratory of the University of Wisconsin for evaluation of suspected infective endocarditis between July 1988 and August 1990. An initial transthoracic study was performed on all patients. If this initial study was sufficiently diagnostic, we did not proceed to a transesophageal study because we felt it would be unethical to do SO. Transesophageal echocardiography was
Volume Number
123 3
Table I. Clinical data evaluated for determination presence of infective endocarditis Historical information History of preexisting valvular disease Presence of a prosthetic valve History of prior endocarditis History of intravenous drug abuse Recent procedure likely to cause transient or sustained remia Antibiotic use within 2 weeks before onset of symptoms Symptoms Fatigue Fever/Chills Sweats Myalgia/Arthralgia Headache/Neurologic change Anorexia/Weight loss Physical examination findings Fever 038.3 C) Any murmur Congestive heart failure (rales, Ss, increased central pressure) Janeway’s lesions Osler’s nodes Roth’s spots Splinter hemorrhage Splenomegaly Embolization Laboratory data Hematuria Proteinuria Anemia (hemoglobin lO,OOO K/wl) Elevated erythrocyte sedimentation rate (>25 mm/hr) Increased platelets (>400,000 K/ml) Abnormal ECG Abnormal chest x-ray Positive blood cultures
TEE detection
of the
endocarditis
775
Table II. Sensitivity and specificity of clinical factors that were significantly more common in patients with endocarditis
compared
to those
Factor
bacte-
of
P*
Symptoms of fever Positive blood culture Any murmur Preexisting valvular disease Headache/Neurologic change Sweats Physical findings of embolization Splinter hemorrhage Janeway’s lesions *Represents
comparison
without
between
0.005 0.001 0.0005 0.003 0.02 0.04 0.01 0.05 0.04
Sensitivity (967) 90 84 81 71 52 48 26 23 13
patients with and without
Specificity (%) 40 57 63 67 77 23 97 100 100 endocarditis.
venous
performed if the diagnostic quality of the transthoracic study was inadequate, when the transthoracic study was inconclusive, or when the presence of an abnormality required further evaluation. Thus the study population consisted of 61 patients who had paired transthoracic and transesophageal echocardiograms, as well as clinical information available for review. Ages of the patients ranged from 17 to 89 (mean 52) years. There were 38 men and 23 women. Echocardiographic examinations were performed up to 22 days apart (mean 4.3 days). Studies were performed on Hewlett-Packard Sonos 500 or 1000 ultrasound imaging systems (Hewlett-Packard Co. Medical Products Co., Andover, Mass.) with a 2.5 or 3.5 MHz transducer. Transesophageal studies were performed with a 5.0 MHz transducer mounted on the tip of a gastroscope. After giving informed consent patients fasted for at least 4 hours before the proce-
dure, and a local anesthetic was applied to the pharynx before insertion of the transducer. Sedation was used at the discretion of the physician performing the study. No complications occurred during any of the transesophageal studies. The medical record of each patient was reviewed for specific historical, symptomatic, physical, and laboratory data (Table I). The diagnosis of infective endocarditis was established by means of these clinical data and the criteria of von Reyn et a1.3 One modification of the von Reyn criteria was made since the patients studied were predominantly a referral population. The presence of any murmur was substituted for the finding of a new regurgitant murmur. Eight patients were evaluated by surgical or autopsy criteria. The frequency with which positive individual clinical signs occurred was calculated. Results of each paired transthoracic and transesophageal study were reviewed by two experienced echocardiographers blinded to the clinical data. The studies were scored for image quality as poor (l+), fair (2+), good (3+), or excellent (4+). Findings in the studies were also graded as positive (definitely present) or negative (possibly or not present) for the presence of a vegetation. When vegetations were visualized they were classified according to location and size. Dimensions were graded as small (lO mm). Underlying valve disease and complications resulting from infective endocarditis were also noted. Statistical analysis of the categoric data was performed by means of chi-square analysis or Mann-Whitney U test where appropriate. Sensitivity was defined as (true positive)/(true positive + false negative) X 100 and specificity as (true negative)/(true negative + false positive) X 100.
March American
Rahko, and Ballantyne
776
Birmingham,
Table
III. Blood culture results of clinically positive casesof infective endocarditis Affected Blood
Mitral
culture
1
2
3 6 2
4 4 1
12
1 1 2 ii?
&treptococcus Group C streptococcus Group D streptococcus Staphylococcus aureus Coagulase-negative staphylococcus Bacillus species Cardiobacterium hominus Candida species Culture negative
were Staphylococcus
1992 Journal
valve Device
Tricuspid
Total 3 1
1
11 4 1 1 3 2 33*
1 1
*In the two patients with infections on two valves, the organisms since infections were classified by location.
Table
Aortic
Heart
2
i
4
aureu~ and group D streptococcus.
These organisms are counted twice
IV. Comparisonof image quality between transesophagealand transthoracic echocardiographicstudies Study Transthoracic Transesophageal
Poor
Fair
7 (11%) 0
12 (20%) 3 (5%)
OBSERVATIONS
Of the 61 patients, 31(51% ) had findings that were positive for infective endocarditis according to the criteria of von Reyn et a1.3There were 33 vegetations in the 31 patients, inasmuch as two patients had infection involving both mitral and aortic valves. All eight patients documented as having infective endocarditis at surgery or autopsy also had positive clinical findings. The frequency of positive findings for each clinical factor shown in Table I was determined. Table II shows clinical factors that were significantly more common in patients with infective endocarditis. Of the items listed under historical information, only a history of preexisting valvular disease was significantly related to the identification of a patient with infective endocarditis. Our sample contained few patients with a prosthetic valve or a prior history of endocarditis and no patients with a history of intravenous drug abuse. From the list of symptoms commonly associated with infective endocarditis, only fever/chills, headache/neurologic changes, and sweating were significantly correlated with the presence of infective endocarditis. From the physical examination items reviewed, any type of murmur, Janeway lesions, and systemic embolization were found significantly more often in those with infective endocarditis. Only one Roth’s spot was noted and no Osler’s nodes were found. Of the items exam-
Good 26 (43 ‘% ) 8 (13%)
Excellent 16 (26%) 50 (82%)
ined in the laboratory data, only positive blood cultures were significantly related to the presence of infective endocarditis. For the mitral valve, aortic valve, and all locations together, Table III shows that Staphylococcus species were the most common organisms followed by Streptococcus species. Table IV lists the grading of image quality for each transthoracic and transesophageal examination. Overall, transesophageal echocardiographic image quality was superior to transthoracic image quality in 39 patients (64 ‘%). In 20 patients (33 %) image quality was equivalent. The average quality score for all transthoracic studies (2.8 f 1.0) was significantly lower thanfortransesophagealstudies (3.8 f 0.5,~ < 0.01). A total of 33 vegetations were visualized (Figs. 1 and 2). Two patients had vegetations involving two valves. Both were identified correctly by transesophageal echocardiology and confirmed at surgery. Table V lists the sensitivity of transthoracic and transesophageal studies for identifying vegetations by location and for the entire group. The specificity for the 30 patients without infective endocarditis was 100 % by the transthoracic method. There was one false positive finding by the transesophageal method. For the mitral valve, transesophageal echocardiography was 100% sensitive in locating vegetations versus 50% for transthoracic studies. For the aortic valve, the sensitivity of transesophageal and transthoracic
Volume Number
123 3
TEE detection
of
endocarditis
777
Fig. 1. A, Transthoracic echocardiogramof mitral valve infective endocarditis. Underlying lesion is mitral valve prolapse.No vegetation is visible. AL, Anterior leaflet; AO, aorta; LA, left atrium; LV, left ventricle; PL, posterior leaflet; RV, right ventricle. B, Transesophagealechocardiogramof mitral valve infective endocarditis (samepatient as in A). Vegetation is clearly evident. AV, Aortic valve; LA, left atrium; MV, mitral valve; VEG, vegetation.
echocardiography was 88 5%and 25%) respectively. Overall sensitivity for transesophageal studies was 88 5’; compared with 30 5%for transthoracic studies. If only studies rated good or excellent for image quality were evaluated, the sensitivity of transthoracic echocardiography only improves to 36 95, whereas it remains 88% for a transesophageal study. Table VI lists the vegetation size of the patients with clinically positive findings. On the mitral valve, all vegetations seen by transthoracic echocardiography were large. This was in full agreement with the corresponding transesophageal results. Vegetations not detected by
transthoracic studies but visualized by the transesophageal method were medium or large. For the aortic valve two of three vegetations located by both transthoracic and transesophageal echocardiography were graded as the same size; the other case differed by one grade. Among the vegetations seen only by transesophageal echocardiography, all three sizes were represented. No comparison can be made concerning vegetations on the tricuspid valve or intravascular devices, inasmuch as none was observed by transthoracic echocardiography. There were six complications found by transesoph-
March
778
hnerican
Birmingham, Rahko, and Balluntyne
Heart
1992 Journal
Fig. 2. A, Transthoracic echocardiogram of aortic valve infective endocarditis. One leaflet of aortic valve is thickened; remainder of valve is normal. AV, Aortic valve; LA, left atrium; RVOT, right ventricular outflow tract. B, Transesophageal echocardiogram of aortic valve infective endocarditis (same patient as in A). Leaflet is more clearly visualized and vegetation is well defined. LA, Left atrium; LV, left ventricle; MV, mitral valve; RV, right ventricle; VEG, vegetation.
V. Comparison of the sensitivity transthoracic echocardiography
of vegetation identification
Table
Study TT TE
6/12 12/12
4/16
(50 0; ) (100%)
valve;
AV, aortic
valve;
TV
(25 % )
14/16
