Incidence and Frequency of Cerebral Embolic Signals in Patients With a Similar Bileaflet Mechanical Heart Valve
Background and Purpose The aim of this study was to determine the incidence and frequency of cerebral embolic signals in a patient population with the same mechanical heart valve using transcranial Doppler examination. Furthermore, it aimed to identify patient and valve characteristics that correlated with the occurrence of these signals.
Methods Ninety-two patients with Carbomedics valves and 15 healthy control subjects took part in the study. Thirty-six patients were examined before and immediately after valve implantation (group 1), 34 patients 1 year after surgery (group 2), and 22 patients 5 years after surgery (group 3). Cerebral embolic signals were detected using transcranial Doppler monitoring of the right middle cerebral artery.
Results Asymptomatic cerebral embolic signals were detected in 87% of the total 92 patients, in 77.8% of group 1 patients, in 91.2% of group 2 patients, and in 95.5% of group 3 patients. No cerebral embolic signals were detected in group 1 patients before surgery or in control subjects. The incidence (P=.04) and frequency (P=.002) of cerebral embolic signals increased significantly with longer duration of valve implantation. A significant positive correlation was also found between frequency of cerebral embolic signals and valve size (r=.4326, P=.00001). Median frequency of embolic signals in patients with a history suggestive of cerebrovascular events (n=14) was 60 signals per hour compared with 11 signals per hour in those with no such history (n=42; P=.04).
Conclusions The incidence and frequency of cerebral embolic signals increased with the duration of valve implantation. The frequency of these signals also was dependent on valve size. Patients who had experienced cerebrovascular symptoms had a higher frequency of cerebral embolic signals compared with those with no such signals. These results should be interpreted with caution but suggest that this method could be of help in assessing the risk of stroke in prosthetic heart valve patients and that prospective clinical studies should now be carried out.
Although many advances have been made in the design and construction of prosthetic heart valves since their introduction over 30 years ago, none of these protheses can compare with the normal human valve with regard to hemodynamic functioning or freedom from complications.1 2 3
Mechanical prosthetic valves offer satisfactory hemodynamic functioning and long-term durability, but they are thrombogenic and patients who receive them require long-term anticoagulation with warfarin. The frequency of thromboembolism in mechanical valve patients is 2% to 4% per patient-year despite anticoagulation and is twice as high in patients with a valve implanted in the mitral compared with the aortic position.2 3 The majority of these thromboembolic events involve the central nervous system,4 5 and their incidence is related to valve type2 3 4 and the presence of atrial fibrillation.5 Most valves appear to show the highest risk for thromboembolic events in the first 6 to 12 months after implantation.2 4
Advances in Doppler technology have made it possible to detect not only gaseous emboli6 7 but also emboli composed of solid elements that are frequently involved in cerebral embolism.7 8 9 This method has now been applied to mechanical heart valve patients, and clinical studies strongly suggest that subclinical cerebral embolic signals may be detected in the majority of patients.10 11 12 13 There are, however, considerable differences in the reported incidence and frequency of these signals, findings that at present remain unexplained.
The aims of this study were (1) to determine the incidence and frequency of cerebral embolic signals in a population of patients with similar mechanical heart valves and (2) to identify patient and valve characteristics that may correlate with the incidence and frequency of these signals.
Subjects and Methods
A total of 92 prosthetic heart valve patients were admitted to the study after giving informed consent. The group consisted of 64 men and 28 women aged from 25 to 82 years, with a median of 64 years. Thirty-six patients took part in the study both before and after valve implantation, whereas the remaining 56 patients were examined on one occasion, 1 or 5 years after implantation surgery. Patients were excluded if Doppler examination of the precerebral or intracranial vessels showed stenosis or occlusion of the carotid, vertebral, or major intracranial arteries.
The preoperative cardiac diagnoses were as follows: aortic stenosis (n=28), aortic insufficiency (n=25), a combination of aortic stenosis and insufficiency (n=13), mitral stenosis (n=6), mitral insufficiency (n=8), a combination of mitral stenosis and insufficiency (n=4), or combined aortic and mitral valve disease (n=8). In 21 cases (22.8%), coronary artery surgery was also performed. All patients received a mechanical bileaflet Carbomedics valve14 ranging in diameter from 19 to 33 mm (median, 25 mm). The patients were divided into three groups.
Group 1 included 36 patients, 27 men and 9 women aged 30 to 82 years (median, 63 years), who received an aortic (n=28), mitral (n=5), or aortic and mitral (n=3) valve. Before valve implantation, the mean left atrial diameter was 43 mm, the mean New York Heart Association (NYHA) functional class was 2.8, and 3 (8.3%) patients were in atrial fibrillation. In these patients, transcranial Doppler monitoring (TDM) was carried out 2 days before and again from 2 to 4 days after surgery.
Group 2 included 34 patients, 23 men and 11 women aged 25 to 82 years (median, 67 years), who had received an aortic (n=28), mitral (n=5), or aortic and mitral valve (n=1) 1 year previously. Before valve implantation, the mean left atrial diameter was 44 mm, the mean NYHA class was 2.8, and 2 (5.9%) patients were in atrial fibrillation.
Group 3 consisted of 22 patients, 14 men and 8 women aged 29 to 77 years (median, 63 years), who had received an aortic (n=13), mitral (n=8), or aortic and mitral valve (n=1) 5 years previously. Preoperatively, the mean left atrial diameter was 47 mm, the mean NYHA class was 2.9, and 5 (22.7%) patients were in atrial fibrillation.
Fifteen elderly healthy individuals, 9 men and 6 women aged 67 to 84 (median, 70) years, with no history of cerebrovascular or cardiac disease served as control subjects. All had normal clinical neurological findings, and Doppler examination of the precerebral and major intracranial arteries was normal.
Transcranial Doppler Monitoring
Doppler examination of the right middle cerebral artery (MCA) was performed in all patients and control subjects with a TC2000S (Nicolet/EME) with a 2-MHz flat monitoring probe. Each monitoring period lasted for 30 minutes in patients and 60 minutes in the control subjects. A stable probe position was maintained using a Müller fixation system (Nicolet/EME) at a depth of 48 to 54 mm and at the angle giving the strongest Doppler signal from the MCA. All Doppler findings were recorded on videotape (Panasonic AG 7355, Matsushita Electrical Industrial).
Embolic signals were recognized using the following criteria: (1) audible short-lasting increases of the Doppler signal power, (2) red spots in the Doppler signal when its intensity was assessed using a color-coded signal intensity scale, and (3) an enhanced power increase (relative power increase above the background Doppler signal) of at least 4 dB measured with a specially designed algorithm.15 The number of embolic signals was presented as the median frequency of signals per hour with interquartile range defined as the 25th to the 75th centiles. Artifacts were assessed as being present when signal intensity increased simultaneously in both directions around the zero velocity line.
Neurological Symptoms and Signs
All patients were questioned regarding cerebrovascular symptoms and specifically asked whether he or she had experienced episodes with the following symptoms: (1) sudden difficulties with speech or understanding simple sentences, (2) sudden weakness or clumsiness in one or both limbs on the same side, (3) sudden numbness, tingling, or loss of feeling in one or both limbs on the same side, (4) sudden loss of balance or sensation of spinning, (5) sudden loss of vision in one eye or a part of one eye, and (6) sudden double vision. A clinical neurological examination was carried out after each TDM.
Coagulation and Hemolysis
After surgery all patients were treated with warfarin, and none were receiving antiplatelet drugs. Anticoagulation activity was measured in all patients with the thrombotest and converted into international normalized ratio (INR). Hematocrit (percent), lactate dehydrogenase (LDH; units per liter), haptoglobin (grams per liter), and hemoglobin in plasma (milligrams per deciliter) were measured, and reticulocytes (percent) were counted in groups 2 and 3. All blood samples were taken within 24 hours of the TDM.
All patients had a clinical cardiological examination including an electrocardiogram (ECG) and an echocardiogram before valve implantation. In the early postoperative period, heart rhythm was monitored continuously on a scope, and ECGs were taken when a possible arrhythmia was observed. A clinical cardiological examination including an ECG was performed within 1 week of surgery and annually thereafter. Echocardiography was carried out if a patient had clinical symptoms or signs suggestive of valve dysfunction.
Three independent groups of patients and two outcome variables, incidence and frequency of embolic signals, were considered. Group comparisons regarding the proportion of patients with embolic signals were performed using the Mantel-Haenszel test of linear trend with increasing odds ratio (OR) by time of exposure, ie, the time elapsed since valve implantation.16 We considered group 1 as the reference group with an OR of 1. As the distribution of the frequency of embolic signals was skewed, the Mann-Whitney or Kruskal-Wallis nonparametric test was used to compare groups of patients.
The following variables were analyzed for an association with the frequency of embolic signals: valve size, patient age at the time of Doppler examination, anticoagulation intensity, and blood parameters indicating hemolysis. The correlation coefficient measurements were performed using the nonparametric Spearman’s correlation coefficient because of the nongaussian distribution of the variables.17 After univariate analysis, multivariate linear regression was used to identify independent prognostic factors.18 When analyzing the association between frequency of embolic signals and valve size, we used the multivariate regression model where size of the valves (19 mm to 33 mm) was dichotomized into large (>23 mm) versus small (≤23 mm).
There was no significant difference between the three patient groups with regard to the preoperative characteristics of age, NYHA functional class, and left atrial diameter.
Assessment of the individual groups showed that none of the 36 patients in group 1 had embolic signals before valve implantation, whereas 28 (77.8%) did so 2 to 4 days after surgery. Cerebral embolic signals were also detected in 31 (91.2%) of the 34 group 2 patients 1 year after implantation and in 21 (95.5%) of the 22 group 3 patients examined 5 years after valve implantation (Table 1⇑).
The risk for developing cerebral embolic signals with longer duration of valve implantation was assessed by comparing the proportion of patients with embolic signals in the three patient groups. Patient group 1 was used as a reference group and was given an OR of 1. Calculation of the OR for group 2 gave a value of 2.95 and a value of 6 for group 3. This risk of developing cerebral embolic signals increased significantly with longer duration of valve implantation (linear trend of P=.04) (Fig 2⇓).
The median frequency of cerebral embolic signals was 6 per hour in group 1 (interquartile range [25th to 75th centiles], 2.5 to 24) and 6 (interquartile range, 2 to 66.5) in group 2 patients. In group 3 patients, the median frequency was 41 signals per hour (interquartile range, 10 to 112.5) (Table 1⇑). The difference in frequency was significant when group 3 was compared with group 1 (P=.002). The difference in frequency was not significant between groups 2 and 3 (P=.09) and groups 2 and 1 (P=.3).
There was no significant difference when valve size was compared in the three patient groups (Table 2⇓), except for valves in group 2 and group 3 patients (P=.04). The median frequency of embolic signals for the total patient population was significantly correlated to the size of the valve (Spearman’s r=.4326, P=.00001) (Table 2⇓, Fig 3⇓). This was found to be independently true for patients with aortic valve replacement (Spearman’s r=.4180, P=.0004). This correlation was not examined for the 18 patients with mitral valve prostheses because of the small number of patients. A multivariate analysis was performed where size of the valves and age were associated with frequency of embolic signals. This showed an association between size of the valves and the frequency of embolic signals when adjusted for age.
For all patients, there was a significantly higher median frequency of embolic signals among the mitral valve patients (37 signals per hour) compared with the aortic valve patients (6 signals per hour) (P=.02) (Table 1⇑). However, this difference could be explained by the larger diameters of the mitral valves (median diameter, 29 mm) compared with those of the aortic valves (median diameter, 24 mm).
The frequency of embolic signals was inversely correlated with the age of the patient (Spearman’s r=−.227, P=.03). This association was present when adjusted for valve position. Our results showed that valve size in younger patients was significantly (P=.04) larger than in older patients, which may explain the observed inverse association.
Atrial fibrillation was present during the postoperative TDM in 7 group 1 patients (19.4%) and in 2 group 2 patients (5.9%), and 10 group 3 patients (45.5%) were in atrial fibrillation (Table 1⇑). In the total patient population, the median frequency of embolic signals was 22 signals per hour in 19 patients with atrial fibrillation compared with 14 signals per hour in 73 patients with sinus rhythm (Table 1⇑). This difference was not statistically significant (P=.4).
None of the patients experienced new cerebrovascular symptoms during TDM. Cerebral embolic signals were not detected in 2 group 1 patients who were temporarily disoriented for time and place in the early postoperative period. Six (17.6%) of the 34 group 2 patients experienced acute cerebrovascular symptoms within 1 year of valve implantation (Table 1⇑): 4 had transient ischemic attacks (TIAs), 1 a combination of amaurosis fugax and TIA, and 1 a hemispheric ischemic stroke with a residual mild hemiparesis. Eight (36.4%) of the 22 group 3 patients experienced acute cerebrovascular symptoms within 5 years of valve implantation, including 5 patients with TIA, 2 with amaurosis fugax, and 1 with a hemispheric ischemic stroke resulting in a slight aphasia (Table 1⇑). This corresponds to a stroke incidence of 1.4 per 100 patient-years. The median frequency of embolic signals was 60 (interquartile range, 20.4 to 134) signals per hour in the 14 (25%) patients from groups 2 and 3 combined (n=56) who had experienced cerebrovascular symptoms compared with 11 (interquartile range, 2 to 66.5) signals per hour in the remaining 42 patients without symptoms (P=.04).
The median INR in group 2 patients was 2.6, 2.5 for those with aortic and 2.8 for those with mitral valves (P=.6). Among group 3 patients, the median INR was 2.6 for both valve locations (P=.9). The frequency of cerebral embolic signals was not associated with the anticoagulation intensity in the total patient population or when each of the three groups was assessed separately (Table 2⇑). There was a significant positive correlation between the frequency of embolic signals and LDH levels for patients from groups 2 and 3 combined (Spearman’s r=.4591, P=.0004) (Table 2⇑). Furthermore, a significant correlation was found between LDH levels and valve size (Spearman’s r=.5183, P=.00001). No correlation was found between the frequency of embolic signals and levels of hematocrit, reticulocytes, haptoglobin, or hemoglobin in plasma in these patients.
At the time of the TDM, the mean NYHA class was 1.1 for group 2 and 1.5 for group 3 patients (P=.01). Transthoracic echocardiography was performed in 6 group 2 patients. The examined valve prostheses were found to be functioning normally. A paravalvular leak was found in 3 of the 8 (13.6%) group 3 patients who had transthoracic echocardiography.
In this study, we found that the majority of patients with Carbomedics heart valves had asymptomatic cerebral embolic signals when examined with transcranial Doppler. Furthermore, we found that both the incidence and frequency of these embolic signals appeared to be at least partly dependent on the duration of valve implantation. These results provide further support for previous studies that report a high incidence of asymptomatic cerebral embolic signals in mechanical heart valve patients.10 11 12 13
There is discussion as to the exact Doppler characteristics of cerebral emboli. We therefore used three main criteria in this study: first, the audible signal and, second, the color-coded intensity scaling of the Doppler signal. Our third parameter was enhanced power, which was measured using a specially designed algorithm.15 Enhanced power is a measure of the relative power increase caused by an embolus compared with the normal background power due to flow in the MCA. We set a level of 4 dB for embolic signals, since this value exceeds the variations in enhanced power that are found in the normal MCA signal.
Our findings strongly suggest that the embolic signals were the result of valve implantation because these signals were not present in patients who were also examined before surgery, and they were not detected in the control subjects. Cerebral embolic signals have also been detected in patients with internal carotid artery stenosis or occlusion,19 20 but such patients were excluded from this study.
The composition and size of cerebral emboli in prosthetic heart valve patients is unknown. Cavitation-induced gas or vapor bubbles have been reported during in vitro testing of mechanical heart valve prostheses.21 22 23 24 25 26 27 However, these bubbles collapsed in less than 1 millisecond; it is therefore very unlikely that they last long enough to reach the cerebral circulation.22 Another possibility is that these signals are due to whole-blood emboli, in which case we would expect that their frequency would correlate with the degree of anticoagulation with warfarin. We did not find such a correlation, which is in agreement with previous reports.10 11 12 13 Finally, aggregation of red blood cells or platelets may represent the origin of these microembolic signals. It has been shown that there is an increased tendency for such aggregation in patients with mechanical heart valves.28 Furthermore, a clinical study in this patient group has shown that treatment with warfarin together with antiplatelet medication (acetylsalicylic acid) may reduce the frequency of cerebrovascular events when compared with treatment with warfarin and placebo.29
We found an increase in the incidence and frequency of cerebral embolic signals with longer duration of valve implantation. In vitro and in vivo studies have shown that cavitation may induce pitting, microcracking, and erosion of the pyrolytic carbonated surface of prosthetic heart valves.23 24 When cavitation occurs on a bileaflet valve, it may appear at the same localized area of the leaflet during each heart cycle, producing a cumulative effect on the valve surface.25 26 Erosions on the valve surface may then predispose to thrombus and embolus formation. This cumulative damaging effect caused by cavitation is one possible explanation for the increased incidence and frequency of cerebral embolic signals with duration of valve implantation.
Another main finding in this study was the positive correlation between signal frequency and valve diameter. Valve cross-sectional area may in itself be of importance, since the exposure of blood to artificial surfaces initiates a complex series of interrelated reactions, among which are adhesion and activation of platelets and white blood cells and the blood coagulation/fibrinolytic system, which may lead to thrombus formation.30 Two recent studies have shown that larger valves have lower cavitation thresholds,24 26 which may be of importance if cavitation plays a direct or indirect role in the origin of embolic signals. A correlation was found between signal frequency and the degree of hemolysis, as expressed by the LDH level. This may be explained by a positive correlation between LDH level and valve size.28 Cavitation is one mechanism that may cause blood damage and hemolysis.27
Although there is some evidence that the incidence and frequency of cerebral embolic signals may differ with the different mechanical valve types,10 12 this was not confirmed in a recent study.13 This issue is therefore unresolved and requires the completion of studies in which an adequate number of different valve types are compared, with similar sizes and duration of valve implantation, and in which the same criteria for embolus detection are used.
In our patient population with a similar valve type, we observed a large variation in the frequency of cerebral embolic signals, ranging from 0 to more than 300 signals per hour, suggesting that there are yet unidentified additional individual factors responsible for emboli formation.31
Patients did not experience acute cerebral symptoms during TDM, even though two patients had more than 300 cerebral embolic signals per hour. This suggests that the signal size or composition is such that they do not cause obstruction of the microcirculation of the brain. Although it may seem unlikely that these signals lead to a cumulative effect on neuropsychological function, this possibility has not been excluded. A decline in cognitive function has been reported in patients with Bjørk-Shiley valves 5 years after implantation,32 but this study did not include a control population.
The incidence of stroke in this study was 1.4 per 100 patient-years, which is similar to or less than that reported previously. In a study of patients with Medtronic Hall valves, the incidence of stroke was 1.4 per 100 patient-years,33 whereas in a population with different heart valve types, the incidence was 2.6 per 100 patient-years.34
An important question is the possible prognostic significance of cerebral embolic signals in patients with prosthetic heart valves. In this study, we found that patients who had experienced TIAs, amaurosis fugax, or stroke had a higher frequency of cerebral embolic signals compared with those who had not experienced such symptoms. This association has not been found in previous studies.10 11 12 We stress that these results should be assessed with caution, since all studies to date have included only retrospective clinical data. However, the results of this study do not exclude the possibility that this examination method may provide important prognostic information for this patient group. The results strongly suggest that prospective studies should now be designed in which the aim is to determine whether the detection of cerebral embolic signals may be of help in assessing the risk of stroke in prosthetic heart valve patients.
This study was supported by The Norwegian Council on Cardiovascular Diseases.
- Received November 17, 1994.
- Revision received March 23, 1995.
- Accepted March 27, 1995.
- Copyright © 1995 by American Heart Association
Karp RB, Sand ME. Mechanical prosthesis: old and new. In: Heart Valve Replacement and Reconstruction. Chicago, Ill: Year Book Medical Publisher Inc; 1987:235-253.
Spencer MP, Campbell SD, Sealey JL, Lindenbergh J. Experiments on decompression bubbles in the circulation using ultrasonic and electromagnetic flowmeters. J Occup Med. 1969;2:238-244.
Russell D, Madden KP, Clark WM, Sandset PM, Zivin JA. Detection of arterial emboli using Doppler ultrasound in rabbits. Stroke. 1991;22:253-258.
Russell D. The detection of cerebral emboli using Doppler ultrasound: theoretical, experimental and clinical aspects. In: Newell DW, Aaslid R, eds. Transcranial Doppler Ultrasound. New York, NY: Raven Press Publishers; 1991:207-213.
Spencer MP, Thomas GI, Nicholls SC, Sauvage LR. Detection of middle cerebral artery emboli during carotid endarterectomy using transcranial Doppler ultrasound. Stroke. 1990;21:415-423.
Grosset DG, Georgiadis D, Stirling S, Cowburn P, Kelman AW, Faichney A, Lees KR. Subclinical embolism in patients undergoing cardiac valve implantation and coronary artery bypass surgery. Eur J Cardiothorac Surg. 1994;8:63-66.
Georgiadis D, Grosset DG, Kelman A, Faichney A, Lees KR. Prevalence and characteristics of intracranial microemboli signals in patients with different types of prosthetic cardiac valves. Stroke. 1994;25:587-592.
Brucher R, Russell D. The Doppler characteristics of solid and gaseous emboli. Cerebrovasc Dis. 1994;4(suppl 3):88. Abstract.
Schlesselman J. Case-Controlled Studies. New York, NY: Oxford University Press; 1982:203-206.
Altman DG. Practical Statistics for Medical Research. London, England: Chapman and Hall; 1993:286-288.
Altman DG. Practical Statistics for Medical Research. London, England: Chapman and Hall; 1993:336-349.
Siebler M, Sitzer M, Steinmetz H. Detection of intracranial emboli in patients with symptomatic carotid artery disease. Stroke. 1992;23:1652-1654.
Georgiadis D, Grosset DG, Lees KR. Transhemispheric passage of microemboli in patients with unilateral internal carotid artery occlusion. Stroke. 1993;24:1664-1666.
Kafesjian R, Howanec M, Ward GD, Diep L, Wagstaff LS, Rhee R. Cavitation damage of pyrolytic carbon in mechanical heart valves. J Heart Valve Dis. 1994;3(suppl 1):S2-S7.
Richard G, Beavan A, Strzepa P. Cavitation threshold ranking and erosion characteristics of bileaflet heart valve prostheses. J Heart Valve Dis. 1994;3(suppl 1):S94-S101.
Shu MCS, Leuer LH, Armitage TL, Schneider TE, Christiansen DR. In vitro observation of mechanical heart valve cavitation. J Heart Valve Dis. 1994;3(suppl 1):S85-S93.
Garrison LA, Lamson TC, Deutsch S, Geselowitz DB, Gaumond RP, Tarbell JM. An in-vitro investigation of prosthetic heart valve cavitation in blood. J Heart Valve Dis. 1994;3(suppl 1):S8-S24.
Salzman EW, Merrill EW. Interaction of blood with artificial surfaces. In: Colman RW, Hirsh J, Marder VJ, Salzman EW, eds. Haemostasis and Thrombosis: Basic Principles and Clinical Practice. 2nd ed. Philadelphia, Pa: JB Lippincott; 1987:1335-1347.
Sotaniemi KA, Mononen H, Hokkanen TE. Long-term cerebral outcome after open-heart surgery: a five-year neuropsychological follow-up study. Stroke. 1986;17:410-416.