Neurol India Home 

Year : 2000  |  Volume : 48  |  Issue : 4  |  Page : 333--7

Clinical features of cardioembolic strokes identified by transesophageal echocardiography.

S Kaul, SJ Kittner, MA Woznaik, MA Sloan, M Corretti, JR Hebel, TR Price 
 Departments of Neurology, Epidemiology and Preventive Medicine and Medicine, University of Maryland School of Medicine, Baltimore, USA., USA

Correspondence Address:
S Kaul
Departments of Neurology, Epidemiology and Preventive Medicine and Medicine, University of Maryland School of Medicine, Baltimore, USA.


Clinical and CT scan features predictive of a cardiac source of embolism (CSOE) are helpful in planning appropriate investigations in ischaemic strokes. The currently described predictors of CSOE were determined before the availability of trans esophageal echocardiography (TEE). After the advent of TEE, many new CSOE were discovered. The present study was planned to investigate if the previously described predictors of CSOE are also valid for patients with CSOE detectable only with TEE (TEE-detected CSOE). From 1992-1995, 485 consecutive patients of ischemic stroke were enrolled in the Maryland Stroke Data Bank (MSDB). Patients with CSOE identified only by TEE and not by clinical, electrocardiographic or transthoracic echocardiographic (TTE) examination were compared to patients with a CSOE with respect to the features of the history, neurologic examination and CT scan. Of 485 patients with cerebral infarction, 132 (27%) patients had CSOE. In 21/132 (16%), diagnosis of high risk CSOE could be established only by TEE. The most discriminating clinical findings in TEE-detected CSOE patients were visual field deficit (OR 2.9; 95% CI, 1.1-7.4) and neglect (OR 3.4; 95% CI,1.2-9.3). Less strong associations were also found with other clinical features described previously for CSOE. No significant differences were found for features of the initial CT scan. In summary, presence of visual field defect and hemineglect may suggest a higher likelihood of finding a CSOE by TEE, even if the clinical cardiac examination and TTE are normal.

How to cite this article:
Kaul S, Kittner S J, Woznaik M A, Sloan M A, Corretti M, Hebel J R, Price T R. Clinical features of cardioembolic strokes identified by transesophageal echocardiography. Neurol India 2000;48:333-7

How to cite this URL:
Kaul S, Kittner S J, Woznaik M A, Sloan M A, Corretti M, Hebel J R, Price T R. Clinical features of cardioembolic strokes identified by transesophageal echocardiography. Neurol India [serial online] 2000 [cited 2023 Feb 5 ];48:333-7
Available from:

Full Text

  ::   IntroductionTop

Many studies have shown that transesophageal echocadiography (TEE) improves the identification of potential cardiac sources of embolism (CSOE).[1],[2],[3] However, unless a cardioembolic etiology is suspected, the patients may not be subjected to TEE and, therefore, a CSOE may be missed. Certain features of history, neurological examination and CT scan have been suggested to be related to an underlying CSOE in patients with ischemic stroke, specifically, systemic embolism, e.g. abrupt onset, diminished level of consciousness, certain cortical findings (visual field deficit, neglect, aphasia), and presence of superficial and deep infarcts greater than lobe in size on CT scan. These studies were done before the availability of transesophageal echocardiography.[4],[5],[6],[7],[8],[9] It is not known if these predictors are also valid for patients in whom the CSOE is detectable only with TEE. The present study was undertaken to address this issue.

  ::   Material and methodsTop

The study population consisted of 485 consecutive patients with ischaemic stroke, prospectively enrolled in the Maryland Stroke Data Bank (MSDB) from January 1992 to May 1995. The MSDB utilized the same methods as the National Institute of Neurological Disorders and Stroke (NINDS) which examined risk factors, presentation, course and outcome of cerebrovascular disease.[1] History and physical examination were obtained on all patients by MSDB neurologists, generally within 24 hours of admission. Transthoracic echocardiography (TTE) in 337 patients and TEE in 122 patients. The patients were subjected to TEE only if the etiology of stroke was not clear and there was a suspicion of cardioembolism by the attending physician. The transthoracic studies were performed by standard techniques with a 2.5 MHz phased array transducer (Hewlett Packard Omniplane, Sonos 1500, Andover, MA). Views were obtained in the parasternal, apical, and subcoastal windows. The transesophageal studies were obtained with a 5 MHz monoplaner transducer (Hewlett Packard 1500) until November 1992, after which all studies were done with multiplanar transducer. All patients had 12-lead electrocardiograms. Holter monitoring was done when there was a suspicion of arrhythmias. After reviewing the clinical features, neuroimaging and other laboratory data, a final diagnosis was made on each patient by the stroke team. Based on clinical, electrocardiographic, holter and echocardiographic (TTE, TEE or both) evidence, patients were classified into CSOE and non-CSOE categories, depending on the presence or absence of an underlying cardioembolic source. Among those with CSOE, we identified a group of patients in whom the CSOE was identified only by TEE (TEE-detected CSOE) and who had no clinical, electrocardiographic or holter evidence of atrial fibrillation, atrial flutter or sick sinus syndrome, no history of recent (<6 months) myocardial infarction and who had normal or non-diagnostic TTE. Patients with cerebral infarcts in whom no CSOE could be detected by clinical, TTE and TEE evaluation, were included in non-CSOE category. Only those cardiac characteristics, which are recognized to be more strongly associated with embolic risk,[4],[10] were considered to be CSOE. This included, a history of valvular surgery, history or electrocardiographic evidence of atrial fibrillation, atrial flutter or sick sinus syndrome, TTE or TEE evidence of ventricular hypokinesis, cardiomyopathy, atrial septal aneurysm with patent foramen ovale, isolated atrial septal aneurysm (>1.5 cm), spontaneous echo contrast and complex aortic arch debris which included mobile, protruded (>4 mm) or ulcerated echodensity. Unlike our previous study on the impact of TEE on management,[11] where we included all the TEE detected CSOE irrespective of their importance, we included only definite and universally accepted TEEdetected CSOE for this study, so as to be as accurate as possible in our selection of cardioembolic strokes. Hence, we did not consider isolated patent foramen ovale, in the absence of deep venous thrombosis or small atrial septal aneurysm, as a CSOE. Their emboligenic potential, though reported, is still controversial.[10]

TEE-detected CSOE and non-CSOE patients were compared. The historical features analyzed were those previously reported as having a positive association with cardiogenic brain embolism such as history of systemic embolism, abrupt onset, and diminished level of consciousness at onset.[4] A history of systemic embolism included either a history or an episode of systemic embolism during hospitalization. Strokes with abrupt onset were defined as a deficit not present on awakening, with onset to peak deficit occurring during the first one hour and no subsequent deterioration in the first 24 hours. Diminished level of consciousness at onset included drowsiness, stupor or coma. The initial neurologic examination features analyzed included diminished level of consciousness and certain cortical findings such as visual field deficit, neglect and aphasia which have previously been reported as positively associated with CSOE.[5] The initial CT scan features analyzed were the presence of superficial and deep infarcts, and infarcts greater than lobe in size, as these have previously been shown to be positively associated with cardioembolic strokes.[6]

Statistical Methods : Fisher's exact test was used to determine the statistical significance of the association between the occurrence of specific characteristics in history, neurological examination and CT scan features in the TEE - detected CSOE and non-CSOE categories. The Odd's ratio was used to measure the strength of the association. Ninety five percent confidence intervals on the Odd's ratio were calculated using the method given by Woolf.[12]

  ::   ResultsTop

Of the 485 patients, 132 (27%) had cardioembolic strokes. We identified 21 patients in whom only TEE demonstrated a high risk CSOE, with the preceding TTE being normal or nondiagnostic. Patients with TEE-detected CSOE had a mean age of 63 (range 2685 years) and a male : female ratio of 9:12. There were 353 patients in the non-CSOE category. The mean age was 60 ( range 16-93) and a male : female ratio of 166:187. [Table I] shows the abnormalities detected by TEE. These included, complex aortic arch echodensities in 7 (6 with mobile plaques and 1 with an immobile but protruding plaque in the arch; ulceration was noted in 1), left atrial appendage thrombus in 3, atrial septal aneurysm with patent foramen ovale having right-on-left shunt in 3, aortic valve thrombus / vegetations in 4, isolated atrial septal aneurysm in 1, multiple valve vegetations in 1, right atrial thrombus with patent foramen ovale in 1, and spontaneous echo contrast in 1 patient. The percent occurrence of features on history, neurologic examination and CT scan is shown in [Table II]. Although most of the features of history and neurological examination showed a higher percent of occurrence in TEE-detected CSOE category, only the presence of visual field deficit (OR 2.9; 95% CI, 1.1 - 7.4) and neglect (OR 3.4; 95% CI, 1.2 - 9.3) was significantly related to TEE-detected CSOE. The CT scan features did not show any trend to be associated with the TEE - detected CSOE.

  ::   DiscussionTop

With the advent of TEE, many patients who would have been considered to have a 'stroke of unknown etiology' have been determined to have a potential CSOE.[13] In 16% of cardioembolic strokes in our study, the CSOE was identified only by TEE. In another study, 18% of patients with ischemic stroke, without clinical evidence of heart disease and a normal TTE, had a CSOE demonstrable only by TEE.[14] We have called this subset of cardioembolic stroke, 'TEE detected CSOE'. Unless suspected and investigated by TEE, the CSOE is likely to be missed in these patients. Unfortunately, not all patients undergo TEE due to the lack of established indications for performing the TEE.[10] Some authors have suggested to perform this test only if TTE is normal, and the patient satisfies the previously described clinical criteria of cardioembolic strokes.[15] However, there are no available data on the applicability of such criteria to TEE-detected CSOE. We found a higher occurrence of previously reported historical features of cardioembolic strokes i.e., systemic embolism, abrupt onset, and diminished level of consciousness in infarcts associated with TEE-detected CSOE as compared with non-CSOE infarcts, though the difference was not statistically significant. In the neurologic examination, the presence of visual field deficit (OR 2.9; 95%, CI, 1.1 - 7.4) and neglect (OR 3.4; 95% CI, 1.2 - 9.3) were more frequent in the TEE- detected CSOE patients as compared to those without a CSOE. It is interesting to note that, although a history of diminished level of consciousness was more frequent in TEE-detected CSOE group compared to non-CSOE group, same was not found on examination. It is likely that the patients might have improved by the time they were examined. Shrinking deficits are well known in embolic strokes because of the distal migration of the embolism.[16] The CT scan features found previously to be associated with a CSOE did not show any trend to be associated with TEE-detected CSOE[6] suggesting thereby that the infarcts caused by TEE-detected CSOE were smaller in size than what was described previously for cardioembolic strokes. This is further substantiated by the rapid improvement in sensorium and presence of only fewer cortical deficits than expected of a cardioembolic strokes in the TEEdetected strokes. Based on our experience, it appears that TEE-detected CSOE, cause smaller infarcts and have a milder clinical picture than other cardioembolic strokes. Although, there are no other studies to confirm this observation, it is possible that there are different clinical subtypes, even within the category of cardioembolic strokes. It is known that there are different sizes of cardiogenic emboli depending on the fibrin content, and it has been suggested that the smaller emboli may be associated with minor cerebral events.[15] The current knowledge about clinical features of cardioembolic strokes is based on previous studies where atrial fibrillation was the most frequent underlying cardio-embolic source. Atrial fibrillation is known to be associated with heavy red clots and could have strongly influenced the conclusion of those studies.[4],[5],[6],[7],[8] In contrast, all patients with atrial fibrillation were excluded in this study. It must be admitted, however, that these analysis were particularly limited by the small numbers with wide confidence intervals and limited statistical power and the results obtained in this study could represent random variation associated with a small sample size. To the best of our knowledge, there is no data available on the clinical features of TEEdetected strokes for comparison and further studies are needed to be done on this subject.

We did not find any significant difference in the mean age of patients with TEE-detected CSOE and those without a CSOE. This emphasizes the utility of TEE in all age groups as opposed to only the younger patients as suggested by some studies.[10] This is further supported by the fact that the most frequent abnormality in TEE-detected CSOE was mobile, pedunculated thrombus in the aortic arch and ascending aorta, found almost exclusively in older age.

In summary, in 16% of cardioembolic strokes, a definitive CSOE could be found only after TEE. Aortic arch echodensities were the usual abnormalities. Although our data suggest that patients with TEE-detected CSOE may have many of the same clinical features as other CSOE patients, the presence of visual field deficit and sensory neglect in particular provide a reasonable clue to the presence of an underlying CSOE, thereby guiding the clinician to determine as to who should undergo a TEE.


1Pearson AC, Labovitz AJ, Tatineni S et al : Superiority of transesophageal echocardiography in detecting cardiac sources of embolism in patients with cerebral ischemia of uncertain etiology. J Am Coll Cardiol 1991; 17 : 66-72.
2Lee RJ, Batzokis T, Yeoh TK et al : Schnittger I. Enhanced detection of intracardiac sources of cerebral emboli by transesophageal echocardiography. Stroke 1991; 22 : 734-739.
3Aschenberg W, Schluter M, Kremer P et al : Transesophageal two-dimensional echocardiography for the detection of left atrial appendage thrombus. J Am Coll Cardiol 1986; 7 : 163-166.
4Kittner SJ, Sharkness CM, Price TR et al : Infarcts with a cardiac source of embolism in the NINDS Stroke Data Bank; Historical features. Neurology 1990; 40 : 281-284.
5Kittner SJ, Sharkness CM, Sloan MA et al : Infarcts with a cardiac source of embolism in the NINDS Stroke Data Bank : Neurologic examination. Neurology1992; 42 : 299-302.
6Kittner SJ, Sharkness CM, Sloan MA et al : Features on initial computed tomography scan of infarcts with a cardiac source of embolism in the NINDS Stroke Data Bank. Stroke 1992; 23 : 1748-1751.
7Bogousslavsky J, Cachin C, Regil F et al : Cardiac sources of embolism and cerebral infarction - clinical consequences and vascular concomitants: The Lausanne Stroke Registry Neurology1991; 41 : 855-859.
8Ramirez-Lassepas M. Cipolle RJ, Bjork RJ et al : Can embolic stroke be diagnosed on the basis of neurologic clinical criteria ? Arch Neurol 1987; 44 : 87-89.
9Foulkes MA, Wolf PA, Price TR et al : The Stroke Data Bank design methods and baseline characteristics. Stroke 1988; 19 : 547-554.
10DeRook FA, Comess KA, Albers GW et al : Transesophageal echocardiography in the evaluation of stroke. Ann In Med 1992; 117 : 922-932.
11Kaul S, Wozniak MA, Sloan MA et al : Impact of transesophageal echocardiography on the long term management of ischemic stroke: experience from a university hospital. Annals of Indian Academy of Neurology 1999; 2 : 161-165
12Woolf B : On estimating the relation between blood group and disease. Ann Hum Genet 1955; 19 : 251-253.
13Amarenco P, Cohen A, Baudrimont M et al : Transesophageal echocardiographic detection of aortic arch disease in patients with cerebral infarction. Stroke1992; 23 : 1005-1009.
14Cujec B, Polasek P, Voll C et al : Transesophageal echocardiography in the detection of potential cardiac source of embolism in stroke patients. Stroke1991; 22 : 727-733.
15Kittner SJ, Hawk MW : Emboli of cardiac origin . In : Johnson, Griffin (eds). Current therapy in neurologic disease. 4th edition. St. Louis, Mosby. Missour. 1993; 174-178.
16Streifler JY, Furlan AJ, Barnett JHM : Cardiogenic brain embolism: Incidence, varieties, treatment. In: Stroke: Pathophysiology, diagnosis and management. 2nd.ed., Barnett HJM, Mohr JP, Stein BM et al (eds). Churchill Livingstone, New York. 1992; 967-994.