Intracranial Microembolic Signals in 500 Patients With Potential Cardiac or Carotid Embolic Source and in Normal Controls
Background and Purpose We undertook this study to evaluate the prevalence and clinical correlations of Doppler microembolic signals (MES) in stroke-prone patients.
Methods Patients with potential cardiac (n=300) or carotid (n=100) embolic source and control subjects (n=100) were monitored with transcranial Doppler sonography for MES. Transthoracic (n=192) and/or transesophageal (n=134) echocardiography and carotid studies (continuous-wave Doppler, n=181; color-coded duplex, n=47) were performed in all patients with potential native cardioembolic source. Carotid disease was evaluated by means of continuous-wave Doppler (n=87), color-coded duplex (n=70), or intra-arterial angiography (n=24) in patients with potential carotid embolic source.
Results Overall MES prevalence was 23% in patients with potential native cardioembolic source (infective endocarditis [n=7] 43%, left ventricular aneurysm [n=38] 34%, intracardiac thrombus [n=23] 26%, dilative cardiomyopathy [n=39] 26%, nonvalvular atrial fibrillation [n=24] 21%, valvular disease [n=80] 15%), 55% in patients with prosthetic cardiac valves (mechanical [n=77] 58%, porcine [n=7] 43%, homografts [n=5] 20%), 28% in patients with carotid disease (symptomatic [n=46] 52%, asymptomatic [n=54] 7%; P<.01), and 5% in control subjects. No relationship between MES counts and patients’ age, sex, or actual medication was noted. The sensitivity and specificity of MES detection in identifying patients with potential embolic sources were 31% and 95%, respectively.
Conclusions Our study confirmed the reported clinical significance of MES in patients with carotid disease and the high specificity of this technique. The demonstrated low sensitivity of MES detection could be due to short monitoring duration or application of antihemostatic treatment. Prospective large-scale studies are needed to determine the definitive value of MES detection as a diagnostic method in patients with potential cardioembolic source.
Embolism accounts for 15% to 30% of all strokes, whereby embolic material originates mostly from the heart or the brain-supplying arteries.1 2 3 4 While several investigators have examined the significance of various cardiovascular lesions in both retrospective and prospective studies, evaluation of individual risk profiles is currently not feasible. Such information would be particularly helpful in clinical decision making, especially concerning mode and intensity of antihemostatic treatment.
Detection of intracranial MES by means of TCD has been described in several stroke-prone patient groups. Although several recent reports suggested a prognostic value of MES detection in patients with occlusive carotid disease,5 6 7 8 no such results were obtained in patients with potential cardioembolic source; both studies published to date describing 289 and 10010 patients (atrial fibrillation and prosthetic cardiac valves9 and various cardiac lesions,10 respectively) failed to provide definitive evidence on this matter. At the same time, no comparative data are available for patients with cardiac or carotid disease, since no single study has examined both groups. Comparison of results of different studies is not very reliable, in particular because of methodological variations (especially concerning MES identification and monitoring time) and discrepancies in the classification of patients.
We undertook this study to evaluate (1) the prevalence of MES in patients with potential cardiac or carotid embolic source and in normal control subjects, (2) the relation of MES to clinical parameters, and (3) the potential value of MES in identifying patients with potential embolic sources.
Subjects and Methods
A total of 500 patients were examined in this study (potential cardioembolic source, n=300, 11.3% symptomatic; carotid disease, n=100, 46% symptomatic; normal controls, n=100).
Patients With Potential Native Cardioembolic Source
All medical inpatient records of the Department of Cardiology were reviewed twice weekly, and patients with potential cardioembolic source were recruited into this study.
Acute infective endocarditis. All patients were monitored within 2 weeks of diagnosis while being treated with intravenous heparin. Cardiac rhythm was sinus in 5 and atrial fibrillation in 2 cases. Lesions affected the aortic valve in 3 and the mitral valve in the remaining 4 cases.
Intracardiac thrombus. Thrombus formation was localized in the left ventricle in 9 and in the left atrium in 14 cases. The underlying disease was dilative cardiomyopathy in 12 patients. Cardiac rhythm was sinus in 17 and atrial fibrillation in 5 patients; 1 patient carried a cardiac pacemaker. Antihemostatic treatment at the time of monitoring consisted of intravenous heparin (n=9), warfarin (n=12), or both (n=2).
Left ventricular aneurysm. This resulted from anterior myocardial infarction in all cases. Time interval between myocardial infarction and TCD monitoring ranged between 1 and 98 months (mean±SD, 26±4 months). Cardiac rhythm was sinus in 30 and atrial fibrillation in 8 cases. Thirty-one patients were receiving antihemostatic treatment at the time of TCD monitoring (intravenous heparin, n=2; warfarin, n=3; both, n=1; aspirin, n=25).
Dilative cardiomyopathy. This was ischemic in origin in 18 and idiopathic in 21 cases. Intermittent ventricular tachycardia was evident in 34 patients (87%); underlying cardiac rhythm was atrial fibrillation in 27 and sinus in 12 cases. Twenty-four patients were receiving antihemostatic treatment at the time of TCD monitoring (intravenous heparin, n=3; warfarin, n=4; aspirin, n=17).
Valvular heart disease. Underlying valvular lesions affected the mitral valve in 29 (pure stenosis, n=1; pure regurgitation, n=23; combined lesions, n=5) and the aortic valve in 45 cases (pure stenosis, n=18; pure regurgitation, n=10; combined lesions, n=17). Both mitral and aortic valves were affected in the remaining 6 patients. Cardiac rhythm was sinus (n=56), atrial fibrillation (n=21), or continuous ventricular pacing (n=3). Thirty-two patients were recruited while awaiting cardiac surgery for valve replacement. At the time of TCD monitoring, 10 patients were receiving warfarin and 6 intravenous heparin.
Nonvalvular atrial fibrillation. This diagnosis was based on the finding of atrial fibrillation in three electrocardiograms performed within 6 months before investigation and the absence of additional valvular lesions on transthoracic (n=21) and/or transesophageal echocardiography (n=9). Eight patients were taking aspirin and 2 warfarin at the time of TCD monitoring.
Patients With Prosthetic Cardiac Valves
Patients were either enrolled when attending their routine follow-up examination (n=40) or randomly selected from the database of the Department of Cardiac Surgery and asked to undergo TCD monitoring by means of a standardized letter.
Seventy-seven patients carried a mechanical valve (Tecna, n=36; Sorin monostrut, n=10; Saint-Jude medical, n=12; ATS, n=13; other types, n=6 [Carbomedics, n=5; Star-Edwards, n=1]), 7 a porcine valve (Carpentier-Edwards supra-annular, n=4; Hancock, n=2; Saint-Jude, n=1), and 5 a homograft valve. The prosthetic valve had been inserted in the aortic position in 59, in the mitral in 24, and in both aortic and mitral positions in 6 cases. Patients were examined 13±2 months after valve replacement (mean±SE). Cardiac rhythm was sinus in 72, atrial fibrillation in 14, and paced in 3 cases. All patients with mechanical, as well as 3 patients with porcine valves, were receiving warfarin at the time of TCD monitoring.
Patients With Carotid Disease
One hundred patients with occlusive carotid disease were recruited. These were (1) symptomatic patients awaiting carotid endarterectomy, (2) acutely symptomatic patients, or (3) asymptomatic patients identified during routine vascular screening. Antiplatelet treatment was administered in 38 (aspirin, n=36; ticlopidine, n=2) and intravenous heparin in 23 cases.
Normal Control Subjects
One hundred patients without history of cerebrovascular or cardiac disease served as normal control subjects. None of these subjects was receiving any antihemostatic treatment.
Evaluation of the carotid arteries by means of continuous-wave Doppler (181) and/or color-coded duplex (n=47) was performed in all patients with potential native cardioembolic source. Diagnosis of a >50% stenosis of the ICA served as exclusion criterion in this group. The degree of stenosis in patients with carotid disease was diagnosed on continuous-wave Doppler (n=87), color-coded duplex (n=70), and/or intra-arterial subtraction angiography (n=24).
Transthoracic (n=192) and/or transesophageal (n=134) echocardiography were performed in patients with potential cardioembolic source, with the exception of those with prosthetic cardiac valves. Additionally, cardiac catheterization was performed in 35 patients with valvular disorders as part of the standard preoperative workup.
Transcranial Doppler Studies
Bilateral (patients with carotid disease and 75 patients with potential cardioembolic source) or unilateral TCD monitoring was performed over the MCA(s) using 2-MHz probes of a pulsed Doppler machine (Pioneer TC-4040, EME, or Multi-Dop X-4 [DWL]) for 30 minutes per patient in the presence of an experienced examiner. MES counts and prevalence in patients with potential cardioembolic source undergoing bilateral monitoring were evaluated on the basis of the results in the right MCA. All monitoring sessions were recorded on DAT (n=398) and/or hard disk (n=93) for later evaluation.
In patients with carotid disease and prosthetic cardiac valves, MES identification was performed on-line during patient monitoring. The same was true for approximately 30% of patients with potential native cardioembolic source. MES of the remaining patients were counted off-line by a blinded examiner. Off-line reevaluation of MES counts by an independent examiner blinded to patient identity and diagnosis was performed for all patients with carotid disease and 50 patients with potential cardioembolic source. These were randomly selected based on the DAT numbers (first patient in DATs 1 to 50).
Criteria for MES identification were (1) characteristic acoustic properties, (2) short duration (<0.3 millisecond), (3) random appearance in the cardiac cycle, (4) unidirectional signal, and (5) intensity increase at least 3 dB above the background. The lowest sample volume possible was used during monitoring (mostly between 6 and 8 mm). At the same time, gain was reduced until the MCA signal had a pale blue color, corresponding to a background intensity of 3 to 6 dB.
Classification of Patients
A standardized neurological questionnaire evaluating the occurrence of symptoms suggestive of cerebral ischemia was obtained from all study participants. Amaurosis fugax, transient or permanent limb or facial weakness, and transient or permanent speech deficit were considered suggestive of anterior circulation ischemia; diplopia, hemianopsia, blurred vision, vertigo, or impaired balance were considered suggestive of ischemia of the posterior circulation. The medical records of all study participants were also reviewed for potential neurological events.
Patients with potential native cardioembolic source and prosthetic cardiac valves were classified as symptomatic if symptoms suggestive of cerebral ischemia (anterior or posterior territory) occurred within 6 months before TCD monitoring or since valve implantation, respectively. Patients with carotid disease were classified as symptomatic if neurological symptoms suggesting an effect on the territory supplied by the diseased ICA had occurred within 6 months before TCD monitoring.
Twenty-five normal control subjects between 20 and 35 years old were enrolled from medical or nursing staff. The same questionnaire was applied in these cases, without any additional examinations. Remaining control subjects (n=75) were neurological inpatients without symptoms or signs of cerebrovascular disease. This was evaluated on the basis of the medical records of each case. Absence of carotid disease and sinus rhythm on electrocardiography were the only inclusion criteria in this group. No other examinations and in particular no transthoracic or transesophageal echocardiography were routinely performed.
Normally and nonnormally distributed data were expressed as mean±SE and as median (95% CIs) or median, interquartile range, respectively. Unpaired t test was used to compare normally distributed data, and Mann-Whitney, Wilcoxon, and Kruskal-Wallis tests were used for nonnormally distributed data. Distribution of frequencies was evaluated using χ2 test. Interobserver variability was calculated by comparing the results of the two different observers for patients with carotid disease (n=100) and patients with potential cardioembolic source (n=50) using Student’s t test. Interobserver agreement on the presence of MES was also evaluated (number of all MES-positive plus number of all MES-negative patients identified as such by both observers divided by 100). Additionally, the sensitivity and specificity of MES detection in identifying patients with potential (cardiac or carotid) embolic source were calculated (number of MES-positive patients divided by total number of patients and number of MES-negative controls divided by all controls, respectively). Significance was declared at the P<.05 level.
No significant differences in MES counts were noted in the 75 bilaterally examined patients with potential cardioembolic source: 0 (0 to 1) versus 0.5 (0 to 0.5), median (95% CI), right and left MCA, respectively (P=.83, Mann-Whitney). Nevertheless, MES were detected unilaterally in 3 patients (0/2, 1/0, 1/0 for right/left MCA, respectively).
Comparison of the results of the two observers provided satisfactory results: patients with carotid disease (P=.96, two-sample t test), agreement over the presence of MES in 85% of cases; patients with potential cardioembolic source (P=.86, two-sample t test), agreement over the presence of MES in 82% of cases. Total MES counts in the 50 DATs were 48 and 44 signals for observers 1 and 2, respectively. A single MES was detected by one of the two observers in 8 of 9 cases not unanimously characterized as MES positive or negative, while 2 MES were detected by one observer in the remaining case.
Both prevalence and counts of MES were significantly higher in patients compared with healthy control subjects (31.3% versus 5% and 4 [3 to 6] versus 1 [1 to 2], respectively, both P<.01, χ2 and Mann-Whitney tests; Table 1⇓). MES counts were significantly higher in patients with carotid compared with those with native cardiac lesions (11 [7 to 19] versus 2 [2 to 3], respectively, P<.01, Mann-Whitney and Wilcoxon tests) without corresponding differences in MES prevalence (28% versus 23.2%, P<.05, χ2 test). No influence of mode or intensity of antihemostatic treatment, cardiac rhythm, or patients’ age or sex on MES prevalence and counts was evident in any group.
Patients With Potential Native Cardioembolic Source
The prevalence of MES showed significant differences among the different groups (P<.01, Kruskal-Wallis) and was highest in patients with infective endocarditis (42.9%), followed by those with left ventricular aneurysm (34.2%), intracardiac thrombus (25.6%), dilative cardiomyopathy (26.1%), nonvalvular atrial fibrillation (20.8%), and valvular disease (15%) (Table 1⇑). Although MES prevalence was higher in symptomatic (n=24) compared with asymptomatic patients, this difference did not reach statistical significance (29.6% versus 22.5%, P<.05, χ2 test). MES counts of the various groups are presented in Table 1⇑, except for those in which MES prevalence was too low to allow statistical analysis (infective endocarditis: 3 patients with 2, 4, and 5 MES; nonrheumatic atrial fibrillation: 5 patients with 1, 1, 1, 2, and 3 MES). Time period between onset of neurological symptoms was 2.7±0.4 months (mean±SE, within 72 hours in 5 cases). Eleven patients had suffered a transient ischemic attack and 13 a complete stroke.
Patients With Prosthetic Cardiac Valves
The prevalence of MES was highest in patients with mechanical valves, followed by those with porcine and homograft valves (58%, 43%, and 20%, respectively). MES counts in patients with SJM, Tecna, BSM, and ATS valves were 17.5 [1 to 40.5], 4 [2 to 9], 2 [1 to 3], and 5.5 [1 to 15], respectively (all median [95% CI], all P<.05, Mann-Whitney and Wilcoxon tests). MES counts in the 6 patients with dual valve replacement were higher compared with those with sole aortic or mitral replacement, but this difference did not reach statistical significance (13.5 [0 to 34] versus 1.5 [1 to 3] and 2 [0.5 to 5.5], respectively, all P<.05, Mann-Whitney and Wilcoxon tests). No correlation between valve size and MES counts was evident in any group. Similarly, no significant differences in MES counts were noted between symptomatic (n=10) and asymptomatic (n=79) patients (4 [1 to 59] versus 5 [2.5 to 9.5], P<.05, Mann-Whitney and Wilcoxon tests). Time interval between TCD monitoring and symptoms onset was 16±6 months. Three patients had suffered a transient ischemic attack and 7 a complete stroke.
Patients With Carotid Disease
The degree of carotid disease was significantly higher in symptomatic compared with asymptomatic patients (P<.05, χ2 test; Table 2⇓) without evident corresponding differences in age or sex (64.2±1 versus 64.6±1 years and 34.8% versus 20.3% women, respectively; P<.05, paired two-sample t and χ2 tests).
Symptomatic patients had suffered complete stroke (n=15), transient ischemic attack (n=19), or amaurosis fugax (n=12). Mean time between TCD monitoring and onset of symptoms was 2.6±0.3 months (within 48 hours in 16 cases).
Significant differences in MES prevalence and counts were noted in the MCAs supplying symptomatic (n=46) compared with those supplying asymptomatic (n=154) cerebral hemispheres (14 [8 to 22] and 52% versus 3 [2 to 6] and 3.9%; P<.01, Mann-Whitney and χ2 tests, respectively). The same difference was evident when comparing symptomatic to asymptomatic patients (MES prevalence 50% versus 7%; P<.01, χ2 test). Additionally, both parameters were higher in the 27 patients examined within 1 month of onset of symptoms compared with the remaining 19 symptomatic patients (4.5 [2 to 9.5] versus 0.5 [0 to 1.5], P<.02, Mann-Whitney; 59.3% versus 36.8%, P<.05, χ2 test).
Normal Control Subjects
Prevalence of MES in normal control subjects was 5% (Table 1⇑). All MES-positive subjects were older than 46 years.
The specificity and sensitivity of MES detection in identifying patients with potential (cardiac or carotid) embolic sources were 95% and 31.3%, respectively.
To the best of our knowledge, the present study describes MES monitoring of the largest patient population to date and is the first large-scale study examining patients with both potential cardiac and potential carotid embolic sources, thus providing valuable comparative data on MES prevalence in these patient groups. Still, several methodological drawbacks must be taken into account before analyzing our results: (1) patients were rarely examined immediately after the onset of neurological symptoms, (2) most patients were receiving antihemostatic treatment at the time of TCD monitoring, (3) control subjects were selected on the basis of clinical history alone, and (4) the monitoring duration of 30 minutes may be too short for certain patient groups. Application of antihemostatic treatment and the interval between symptom onset and TCD monitoring constitute serious methodological problems, since several studies have demonstrated the close relationship between these factors and MES counts.12 13 14 Nabavi et al15 demonstrated a striking clinical relevance of MES detection in patients with left ventricular assist devices (Novacor), using a very rigid protocol comprising almost daily monitorings. This approach is feasible only in long-term hospitalized patients, although its time requirements allow a widespread use. Thus, while points 1 and 2 surely influence the results of MES monitoring, they appear to be unsolvable problems, even in prospective studies. Although a monitoring duration of 30 minutes was used in several studies to date, this remains a suboptimal compromise, mostly because of lack of patient cooperation, which is caused mainly by limited tolerance of the probe.
The first interesting finding of our study was the high specificity of MES detection, which would probably be further improved by application of imaging studies to exclude occult embolic sources in normal control subjects. The demonstrated low sensitivity obscures this finding and argues against clinical use of this technique as a screening method, at least with the monitoring duration of 30 minutes.
Our results concerning MES prevalence in patients with cardiac lesions must be treated with caution, since the various groups were inhomogeneous, in particular concerning treatment and cardiac rhythm. Apart from the already mentioned limitations, the lack of significant differences between symptomatic and asymptomatic patients in this group can additionally be due to selective administration or intensification of antihemostatic treatment in symptomatic cases. Sliwka et al10 recently published data on MES detection in 100 patients with potential cardioembolic source. Their results are only partially comparable with ours because of differences in the classification of patients. While the overall MES prevalence reported in this study was similar to our results (36% and 32%, respectively), significant differences were noted in patients with nonvalvular atrial fibrillation (7 of 11 patients [63.6%] reported by Sliwka et al10 compared with 4 of 24 patients [21%] in the present study). This discrepancy could be coincidental or due to differences in treatment between the two studied groups.
Prevalence of MES and the lack of influence of treatment and cardiac rhythm on their counts in patients with mechanical prosthetic valves are in accordance with findings in previous publications.16 17 18 19 We failed to demonstrate any significant differences in MES counts between symptomatic and asymptomatic patients. While this issue is still a matter of debate, the low number of symptomatic patients in the present study prohibits definitive statements. MES detection in patients with homograft valves has not been previously reported. Our initial observations suggest a lower MES prevalence compared with patients with mechanical or porcine valves.
Our results in patients with carotid lesions add weight to the growing evidence that MES are a prognostic marker in these cases.5 6 7 8 14 The finding of significantly higher MES counts in acutely symptomatic patients is in accordance with previous reports.5 14 No echocardiography studies were performed in these patients, mainly because most of them presented either as outpatients (asymptomatic cases) or as inpatients of surgical departments. Although this point constitutes a methodological weakness of the present study, it also applies to all published reports of MES detection in patients with carotid disease.
The observation of unilateral MES detection in 3 of the 75 bilaterally monitored cases suggests that monitoring over both MCAs provides more reliable results, in particular in patients with low MES counts. Additionally, the fact that MES were detected unilaterally in 99% of patients with carotid disease supports the hypothesis that bilateral TCD monitoring can aid the differentiation between cardiac and carotid embolic sources.
In conclusion, this study confirmed previous reports on the clinical relevance of MES in patients with carotid disease and the high specificity of MES detection. While no direct clinical significance of MES in patients with potential cardioembolic sources was demonstrated, prospective studies including long-term patient follow-up and neuropsychological evaluation are warranted to equivocally clarify this issue.
Selected Abbreviations and Acronyms
|ICA||=||internal carotid artery|
|MCA||=||middle cerebral artery|
|TCD||=||transcranial Doppler sonography|
- Received January 14, 1997.
- Revision received April 4, 1997.
- Accepted April 4, 1997.
- Copyright © 1997 by American Heart Association
Bogousslavsky J, Melle GV, Regli F. The Lausanne Stroke Registry: analysis of 1000 consecutive patients with first stroke. Stroke. 1988;19:1083-1092.
Caplan LR, Hier DB, D’Cruz I. Cerebral embolism in the Michel Reese Stroke Registry. Stroke. 1983;14:530-537.
Siebler M, Sitzer M, Rose G, Benfeldt D, Steinmetz H. Silent cerebral embolism caused by neurologically symptomatic high-grade carotid stenosis: event rates before and after carotid endarterectomy. Brain. 1993;116:1005-1015.
Siebler M, Nachtmann A, Sitzer M, Rose G, Kleinschmidt A, Rademacher J, Steinmetz H. Cerebral microembolism and the risk of ischemia in asymptomatic high-grade internal carotid artery stenosis. Stroke. 1995;26:2184-2186.
Siebler M, Kleinschmidt A, Sitzer M, Steinmetz H, Freund HJ. Cerebral microembolism in symptomatic and asymptomatic high-grade internal carotid artery stenosis. Neurology. 1994;44:615-618.
Babikian VL, Hyde C, Pochay V, Winter MR. Clinical correlates of high-intensity transient signals detected on transcranial Doppler sonography in patients with cerebrovascular disease. Stroke. 1994;25:1570-1573.
Tong DC, Bolger A, Albers GW. Incidence of transcranial Doppler–detected microemboli in patients referred for echocardiography. Stroke. 1994;25:2138-2141.
Sliwka U, Job FP, Wissuwa D, Diehl R, Flachkampf FA, Hanrath P, Noth J. Occurrence of transcranial Doppler high-intensity transient signals in patients with potential cardiac sources of embolism: a prospective study. Stroke. 1995;26:2067-2070.
Consensus Committee of the Ninth International Cerebral Hemodynamics Symposium: basic identification criteria of Doppler microembolic signals. Stroke. 1995;26:1123.
Georgiadis D, Hill M, Zunker P, Stögbauer F, Ringelstein EB. Anticoagulation monitoring with transcranial Doppler. Lancet. 1994;344:1373-1374.
Forteza AM, Babikian VL, Hyde C, Winter M, Pochay V. Effect of time and cerebrovascular symptoms on the prevalence of microembolic signals in patients with cervical carotid stenosis. Stroke. 1996;27:687-690.
Nabavi DG, Georgiadis D, Mumme T, Schmid C, Scheld HH, Ringelstein EB. Prognostic impact of Doppler microembolic signals in patients with left ventricular assist devices. Stroke. 1996;27:891-896.
Sliwka U, Diehl RR, Meyer B, Schöndube F, Noth J. Transcranial Doppler ‘high intensity transient signals’ in the acute phase and long-term follow-up of mechanical heart valve implantation. J Stroke Cerebrovasc Dis. 1995;5:139-146.
Braekken KS, Russell D, Brucher R, Svennevig J. Incidence and frequency of cerebral embolic signals in patients with a similar bileaflet mechanical heart valve. Stroke. 1995;26:1225-1230.
Sturzenegger M, Beer JH, Rihs F. Monitoring combined antithrombotic treatments in patients with prosthetic heart valves using transcranial Doppler ultrasound. Stroke. 1995;26:63-69.
Georgiadis D, Grosset DG, Kelman AW, Faichney A, Lees KR. Incidence and characteristics of intracranial microemboli signals in patients with different types of prosthetic cardiac valves. Stroke. 1994;25:587-592.
van Zuilen EV, Moll FL, Vermeulen FE, Mauser HW, van Gijn J, Ackerstaff RG. Detection of cerebral microemboli by means of transcranial Doppler monitoring before and after carotid endarterectomy. Stroke. 1995;26:210-213.