Mitral Valve Strands in Patients With Focal Cerebral Ischemia
Background and Purpose Filamentous strands attached to the mitral valve are a recently described finding in occasional patients undergoing transesophageal echocardiography (TEE), but the frequency and clinical significance of these strands remain poorly defined. The purpose of the present study was to review the prevalence of mitral valve strands in patients undergoing TEE examination and to explore the relation of these strands to cardioembolic cerebral ischemia.
Methods All patients with native mitral valves referred for clinically indicated TEE over a 2-year period at our institution were evaluated for the presence of mitral valve strands (defined as highly mobile filamentous masses <1 mm thick attached to the atrial surface of mitral leaflets).
Results Of 968 study patients, mitral valve strands were identified in 22 individuals (2.3%). Mitral valve strands were significantly more common in patients referred for TEE as a result of a recent ischemic cerebrovascular event compared with patients referred for other study indications (6.3% versus 0.3%, respectively; P<.00001). Among patients ≤50 years of age with likely cardioembolic stroke or transient ischemic attack, 16% were found to have mitral valve strands on TEE examination. In 9% of these young patients, no other TEE finding associated with cardioembolic risk was present.
Conclusions Filamentous strands attached to the mitral valve appear to represent another risk factor for embolic cerebral ischemia, particularly in patients ≤50 years of age.
Cardiogenic cerebral embolism is believed to be responsible for 15% to 30% of all ischemic strokes.1 Transthoracic echocardiography is useful for identifying cardiac sources of embolism in only a small percentage of unselected patients with unexplained cerebral ischemia.2 Recently, transesophageal echocardiography (TEE) has emerged as a more useful tool in the evaluation of such patients. TEE has been shown to have a substantially higher yield than transthoracic echocardiography alone in detecting potential cardiac embolic sources, particularly in patients without clinical evidence of cardiac disease.3 4 5 6 7 8 9 Nevertheless, a significant percentage of patients with presumed cardiogenic cerebral embolism will not have a recognized embolic source apparent on TEE examination.3 4 5 6 7 8 9
During TEE examinations in our laboratory, we have noted the occasional presence of mobile filamentous strands on the mitral valve. The purpose of the present study was to review the significance of these strands in patients with unexplained focal cerebral ischemia.
Subjects and Methods
All TEE examinations performed between May 1990 and May 1992 at our institution were prospectively entered into a computerized database. Patients with mechanical mitral prostheses were identified and excluded from inclusion in the present study. The patient population for this investigation therefore consisted of 968 subjects. There were 548 males and 420 females, with a mean age of 52±17 years (range, 14 to 87 years). All patients were referred for TEE by their personal physician for clinical indications. Clinical information was collected at the time of TEE evaluation through patient interviews, consultation-form review, and, if available, inpatient-chart examination. Special emphasis was given to the recognition of recent (within 3 months of TEE examination) cerebral ischemic events. The study protocol was reviewed and approved by our institutional human subjects review board, and all patients gave written informed consent.
Complete TEE examinations were performed in subjects in the fasting state after topical anesthesia of the oropharynx with lidocaine spray and intravenous sedation with meperidine and midazolam. Images were obtained using a Hewlett-Packard Sonos 500, 1000, or 1500 ultrasonograph equipped with a 5.0-MHz single or biplane transducer (Hewlett-Packard models 21362A and 21363A, respectively). Standard transesophageal and transgastric views were obtained. The mitral valve was carefully interrogated in all patients independent of the study indication. The zoom feature was used at the discretion of the echocardiographer performing the study. All patients undergoing TEE as a result of a recent cerebral ischemic event received a venous saline contrast injection during visualization of the interatrial septum for detection of an interatrial communication. In patients with a negative saline contrast study during normal respiration, a repeat injection was performed during a Valsalva maneuver. Studies were recorded on 12.7-mm (0.5-in) S-VHS videotape for review.
TEE studies were reviewed at the time of the examination by one of two experienced echocardiographers. Mitral valve strands were defined as thin (<1 mm in width), highly mobile, filamentous projections attached to the atrial surface of the mitral leaflets (Fig 1⇓). Valvular masses >1 mm in width were not considered strands. Mitral valves were also evaluated for abnormal thickening (>5 mm), masses consistent with infective vegetations, and mitral valve prolapse. All information was subsequently entered into the report database.
Classification of Cerebral Ischemic Events
Data from patients with a recent cerebrovascular ischemic event as the indication for their TEE examination were carefully reviewed by two neurologists. On the basis of available clinical and neuroimaging information, such patients were classified into one of three groups: (1) likely cardioembolic stroke or transient ischemic attack (TIA); (2) unlikely cardioembolic stroke or TIA, with other or more accepted mechanism(s) identified; and (3) uncertain etiology or incomplete workup.
A complete workup was considered to include a detailed history and physical examination, CT and/or MRI scan of the brain, carotid ultrasound and/or angiogram, transthoracic echocardiogram, and a thrombotic profile. Study neurologists were blinded to findings on TEE for purposes of patient classification. Features used to classify patients into the likely cardioembolic category included (1) abrupt onset of maximal neurological deficit; (2) loss of consciousness at event onset; (3) absence of antecedent TIAs; (4) concomitant systemic embolism; (5) presence of atrial fibrillation, acute myocardial infarction, or mitral valve stenosis; (6) cortical distribution of infarcts in the territory of the middle and posterior cerebral arteries; (7) hemorrhagic cerebral infarction; and (8) absence of carotid stenosis on ultrasound or angiography.10
Unlikely cardioembolic stroke or TIA was diagnosed when features suggestive of a cerebrovascular etiology were present, including (1) stuttering onset of neurological deficit; (2) history of antecedent TIAs; (3) carotid stenosis of >50% on ultrasound or angiogram; (4) findings of cerebral arterial vasculitis or dissection; (5) risk factors for lacunar infarction, including diabetes mellitus and hypertension; and (6) lacunar infarction on CT or MRI scan.10 Patients in whom clinical and imaging procedures did not clearly favor either cardioembolic or cerebrovascular causes were considered to have had events of uncertain etiology.
Group comparisons for categorical values were made using the χ2 test. Continuous variables were compared using unpaired Student's t tests. A value of P<.05 was considered significant.
Mitral Valve Strands
Mitral valve strands were identified on 22 of 968 TEE examinations (2.3%). These 22 patients with mitral valve strands included 13 men and 9 women, with a mean age of 50±15 years (range, 28 to 87 years). Of the 968 study patients, 318 (33%) were being studied as a result of a recent ischemic cerebrovascular event. Mitral valve strands were present in 20 of these 318 patients (6.3%) compared with only 2 of 650 patients (0.3%) being studied for other clinical indications (P<.00001). All patients with mitral valve strands were in sinus rhythm. There was no relation of mitral valve strands to gender or the presence of hypertension.
Cerebral Ischemic Events in Patients With Mitral Valve Strands
Of the 20 patients with mitral valve strands in whom the study indication was a recent ischemic cerebrovascular event, 3 (15%) were categorized by study neurologists as having an uncertain etiology or incomplete workup. Ten patients (50%) were characterized as having unlikely cardioembolic stroke, with other or more accepted mechanisms identified (2 lacunar strokes, 2 large-vessel cerebrovascular disease, 1 fibromuscular dysplasia, 1 migraine, and 4 other). The remaining 7 patients (35%) with mitral valve strands were believed to have likely cardioembolic stroke or TIA.
Mitral Valve Strands in Young Patients With Focal Cerebral Ischemia
The 318 patients undergoing TEE evaluation because of a recent cerebral ischemic event included 107 patients ≤50 years of age. Study neurologists classified 44 (41%) of these individuals as having likely cardioembolic stroke or TIA. Fig 2⇓ shows the TEE findings for these 44 patients. Mitral valve strands were identified in 7 (16%) of these individuals. In 3 patients with mitral valve strands, other TEE findings associated with cardioembolic risk were present (2 patent foramina ovale and 1 mitral valve prolapse). However, in 4 (9%) individuals with mitral valve strands, no other cardiac abnormality was identified on TEE examination.
The major finding in the present study was the identification by TEE examination of highly mobile filamentous strands on the mitral valve in approximately 6% of patients referred for evaluation of a possible cardiac source of cerebral embolism. Among patients ≤50 years old with recent focal cerebral ischemia thought to be of cardioembolic origin, 16% were found to have mitral valve strands. In 9% of this young group of patients, no other explanation for stroke or TIA could be identified. Mitral valve strands were seen only rarely in individuals undergoing TEE for reasons other than recent cerebral ischemia. These findings suggest an etiologic role for mitral valve strands in ischemic stroke and TIA, particularly in young patients with a high clinical likelihood of a cardioembolic event.
Origin of Mitral Valve Strands
The pathological composition of mitral valve strands is not currently known. However, the appearance of these strands is highly suggestive of Lambl's excrescences. Lambl11 first described these small filiform processes on the ventricular surface of aortic valve leaflets in 1856. He identified them in 2% of an autopsy series but was unable to attribute any clinical significance to them.
Magarey12 reported the presence of similar excrescences on the atrial surface of the mitral valve in an autopsy series of 250 patients. The majority of excrescences were found to be <1 mm thick and 1 to 10 mm long, similar in size to the strands visualized in the present study with TEE. Histologically, Magarey found excrescences to be composed of a cellular connective tissue core covered by endothelium. On the basis of his pathological observations, Magarey speculated that Lambl's excrescences originate from fibrinous deposits on valve leaflets.
None of the patients with mitral valve strands in the present study underwent mitral valve surgery or autopsy study. Consequently, pathological confirmation of TEE-identified mitral valve strands as Lambl's excrescences was not possible. Lee et al3 described a patient with mitral valve strands seen on TEE who had mitral valve replacement because of severe mitral regurgitation. Pathological examination of the resected valve demonstrated filiform processes consistent with Lambl's excrescences. Fitzgerald et al13 reported a case of lower extremity occlusive embolism thought to be of cardiac origin. The embolic material recovered at surgery was composed of a conglomerate of fine threadlike structures. After histological examination, it was concluded that the embolus represented a conglomeration of multiple Lambl's excrescences. This limited evidence is consistent with the hypothesis that mitral valve strands represent Lambl's excrescences and that they are capable of peripheral embolization.
Comparison With Previous Studies
Very few reports of mitral valve strands and their clinical significance are available. Lee et al3 was the first to report the detection of threadlike strands on the mitral valve seen on TEE. In a series of 50 consecutive patients with suspected cerebral emboli and an unremarkable transthoracic echocardiogram, 11 individuals (22%) were noted on TEE examination to have mitral valve strands. In none of these patients was an alternative explanation for cerebral embolism found.
More recently, Freedberg et al14 retrospectively reviewed a series of 1559 patients who underwent TEE over a 2-year period. Fine threadlike strands were identified on the mitral valve in 63 patients (4.0%). In addition, 26 patients (1.7%) were noted to have similar strands on the aortic valve. Strands were much more common in patients referred for TEE as a consequence of a recent embolic event (10.6%) compared with patients referred for other reasons (2.3%). The authors concluded that TEE-detected mitral valve strands are strongly associated with systemic embolization.
The results of the present study support findings from these previous reports. Mitral valve strands were identified far more commonly in our patients undergoing TEE assessment for cardiac sources of embolism than in patients referred for alternative indications (6.3% versus 0.3%, respectively; P<.00001). Furthermore, more than half of our patients with probable cardioembolic cerebral ischemia in whom mitral valve strands were present had no other cardiac abnormality that might explain cerebral embolism.
The overall incidence of mitral valve strands in our study was similar to that observed by Freedberg et al.14 Lee et al3 reported a much higher percentage of patients with mitral valve strands (22%) in their study population. This difference likely reflects the selective composition of the patient group examined by Lee et al. All patients included in their study had an unexplained cerebral ischemic event and additionally had a transthoracic echocardiogram that failed to reveal a potential cardiac source of embolism.
Patients with prosthetic mitral valves were not included in the present study. Our laboratory recently reported the presence of strands on 26% of 214 consecutive patients with prosthetic mitral valves examined with TEE.15 These strands were much more common in patients with a suspected cardioembolic event than in those referred for other indications (53% versus 15%, respectively; P=.0004). Similarly, Isada et al16 reported a relationship between prosthetic mitral valve strands and cardioembolism in a smaller series of 87 patients. Prosthetic valve strands seen with TEE strongly resemble those seen on native mitral valves, but it is unclear whether they share a common origin. Limited pathological examination of these strands from surgically excised prosthetic valves at our institution has demonstrated fibrin to be a significant component of strand composition.15 In conjunction with observations by Magarey12 that Lambl's excrescences originate from fibrinous deposits on valve leaflets, it seems likely that similar pathological processes are active in strand formation on both native and prosthetic mitral valves.
Relationship of Mitral Valve Strands and Cardioembolism to Age
A notable observation in the present study was that 16% of patients ≤50 years old with embolic cerebral ischemia were found to have mitral valve strands on TEE. In 9% of these individuals, mitral valve strands represented the sole identifiable cardiac abnormality. This relatively high prevalence of mitral valve strands in young patients with presumed embolic focal cerebral ischemia has not been previously reported.
In his autopsy study, Magarey12 noted that Lambl's excrescences were progressively more common with increasing age. He concluded that they were an accompaniment of the aging process as a result of wear and tear on the valve. The observation in our study of an increased incidence of mitral valve strands in younger patients with embolic cerebral ischemia may reflect identification of individuals with premature degenerative valvular changes. Such changes could conceivably lead to fibrin deposition and subsequent organization into visible strands. No echocardiographic evidence of degenerative valvular changes was seen in the present study, however, since most mitral valves with strands appeared to be otherwise structurally normal. Similarly, Freedberg et al14 reported that more than 90% of native valves with strands were morphologically normal on TEE. To date, we have not had the opportunity at our institution to histologically examine a mitral valve with strands to assess microscopic degenerative changes.
Several limitations to the present study should be mentioned. Most importantly, TEEs were not interpreted in a blinded manner but were instead performed at the time of the study. In addition, TEE operators were not blinded to the study indication during test performance. It could therefore be argued that patients referred for source of embolism evaluation had a more extensive examination of the mitral valve, with more zoom use and more angles of interrogation. However, the present study information was collected in a prospective manner, with complete evaluation of the mitral valve for strands being conducted in all patients. Although zoom use was not standardized, no patient was found to have strands that were not visible without the zoom feature in the present study.
There is no gold standard test for identifying patients as having either cardioembolic or cerebrovascular cerebral ischemia. Therefore, some of the patients in the present study may have been misclassified. This limitation is inherent to all studies of cerebral ischemia. Nevertheless, the present investigation used the expertise of neurologists with substantial experience in evaluating stroke patients, and radiographic and transthoracic echocardiographic imaging were available for review for nearly all patients.
It is not possible from the present study to establish cause and effect for mitral valve strands and cerebral embolism. This limitation is shared by many of the presumed cardiac sources of embolism identifiable by echocardiography (eg, mitral valve prolapse, atrial septal aneurysm, patent foramen ovale, and spontaneous echo contrast). It is conceivable that mitral valve strands simply represent a marker for other etiologic mechanisms, such as a hypercoagulable state. Nevertheless, our study supports a limited number of prior studies that suggest a strong relationship between valvular strands and cardioembolic events. Further investigation is therefore needed, including prospective studies and studies to define optimal therapy.
The authors would like to thank Teresa Mrkvicka Henderson for her expert secretarial assistance.
- Received January 15, 1996.
- Revision received March 14, 1996.
- Accepted March 14, 1996.
- Copyright © 1996 by American Heart Association
American College of Cardiology/American Heart Association Task Force on Assessment of Diagnostic and Therapeutic Cardiovascular Procedures (Subcommittee to Develop Guidelines for the Clinical Application of Echocardiography). ACC/AHA guidelines for the clinical application of echocardiography. J Am Coll Cardiol. 1990;16:1505-1528.
Lee RJ, Bartzokis T, Yeoh TK, Grogin HR, Choi D, Schnittger I. Enhanced detection of intracardiac sources of cerebral emboli by transesophageal echocardiography. Stroke. 1991;22:734-739.
Pop G, Sutherland GR, Koudstaal PJ, Sit TW, de Jong G, Roelandt JRTC. Transesophageal echocardiography in the detection of intracardiac embolic sources in patients with transient ischemic attacks. Stroke. 1990;21:560-565.
DeRook F, Comess K, Albers G, Primozich J, Atwood JE, Bolger A. Prevalence of transesophageal echocardiographic findings in stroke patients with and without carotid stenosis. J Am Soc Echocardiogr. 1991;4:293. Abstract II-G.
Cujec B, Polasek P, Voll C, Shuaib A. Transesophageal echocardiography in the detection of potential cardiac source of embolism in stroke patients. Stroke. 1991;22:727-733.
Zenker G, Erbel R, Kramer G, Mohr-Kahaly S, Drexler M, Harnoncourt K, Meyer J. Transesophageal two-dimensional echocardiography in young patients with cerebral ischemic events. Stroke. 1988;19:345-348.
Lambl VA. Papillare exkreszenzen an der semilunar-klappe der aorta. Wien Med Wochenschr. 1856;6:244-247.