Author + information
- Received January 15, 2008
- Revision received March 14, 2008
- Accepted April 16, 2008
- Published online September 2, 2008.
- ↵⁎Reprint requests and correspondence:
Dr. Marco R. Di Tullio, Columbia University Medical Center, PH3-342, 622 West 168th Street, New York, New York 10032
Objectives Our goal was to assess the effect of hypercoagulability on the risk of stroke in patients with aortic plaques.
Background Atherosclerotic plaques in the aortic arch are a risk factor for ischemic stroke. Their relationship with blood hypercoagulability, which might enhance their embolic potential and affect treatment and prevention, is not known.
Methods We performed transesophageal echocardiography in 255 patients with first acute ischemic stroke and in 209 control subjects matched by age, gender, and race/ethnicity. The association between arch plaques and hypercoagulability, and its effect on the stroke risk, was assessed with a case-control design. Stroke patients were then followed prospectively to assess recurrent stroke and death.
Results Large (≥4 mm) arch plaques were associated with increased stroke risk (adjusted odds ratio [OR]: 2.4, 95% confidence interval [CI]: 1.3 to 4.6), especially when ulcerations or superimposed thrombus were present (adjusted OR: 3.3, 95% CI: 1.4 to 8.2). Prothrombin fragment F 1.2, an indicator of thrombin generation, was associated with large plaques in stroke patients (p = 0.02), but not in control subjects. Over a mean follow-up of 55.1 ± 37.2 months, stroke patients with large plaques and F 1.2 over the median value had a significantly higher risk of recurrent stroke and death than those with large plaques but lower F 1.2 levels (230 events per 1,000 person-years vs. 85 events per 1,000 person-years; p = 0.05).
Conclusions In patients presenting with acute ischemic stroke, large aortic plaques are associated with blood hypercoagulability, suggesting a role for coagulation activation in the stroke mechanism. Coexistence of large aortic plaques and blood hypercoagulability is associated with an increased risk of recurrent stroke and death.
The presence of large plaques in the proximal segment of the aorta has been shown to be associated with an increased risk of ischemic stroke. The association between aortic plaques and stroke, first discovered in pathology studies (1), has been subsequently confirmed in vivo by case-control (2–4) and prospective studies (5–7) using transesophageal echocardiography (TEE). A relationship between plaque thickness and stroke risk has been identified, with a cutoff of 4 mm widely acknowledged as clinically relevant for risk stratification (2,6,8). Morphologic features of the plaque, such as presence of ulcerations or mobile components, have been identified that also contribute to the embolic risk (4,9–11).
Despite these findings, the best preventive approach to decrease the cerebral embolic risk of aortic plaques has not been identified. While there is general agreement that plaques with mobile components, which are often proven to be thrombi at pathology, may require systemic anticoagulation (8,12,13), there is no consensus on which treatment should be used in case of large but nonmobile plaques. It has been hypothesized that superimposed thrombotic material could at times contribute to the plaque thickness that is commonly measured by TEE (2), and that such a thrombotic component could enhance the plaque's embolic potential. However, the relationship between aortic plaques and hypercoagulability, which would conceivably predispose to thrombus formation on a plaque, has not been studied. A positive association between plaque presence/thickness and hypercoagulability could provide insights on stroke mechanism and potential for preventive measures. The aim of the present study was, therefore, to assess the relationship between proximal aortic plaques presence/thickness and hypercoagulability, and to assess the potential impact on the risk of ischemic stroke and death.
As part of the National Institute of Neurological Disorders and Stroke-funded APRIS (Aortic Plaques and Risk of Ischemic Stroke) study, 255 patients over the age of 55 years with acute ischemic stroke and 209 stroke-free control subjects were enrolled in the study between 1997 and 2002. Stroke patients were consecutive patients residing in northern Manhattan, admitted to Columbia University Medical Center with a first acute ischemic stroke, defined according to TOAST (Trial of Org 10172 in Acute Stroke Treatment) criteria (14). Control subjects were selected from the participants of the NOMAS (Northern Manhattan Study) trial, and matched to cases by age (within 5 years), gender, and race/ethnicity. Informed consent was obtained from cases and controls, or their health proxies when necessary. The study was approved by the Institutional Review Board of Columbia University Medical Center.
Hypertension was defined as a systolic blood pressure recording ≥140 mm Hg or a diastolic blood pressure recording ≥90 mm Hg based on the mean of 2 readings, a patient's self-report of a history of hypertension, or antihypertensive medication use. Diabetes mellitus was defined by a patient's self-report of such a history, insulin use, oral hypoglycemic use, or a fasting glucose ≥126 mg/dl. Hypercholesterolemia was defined as total serum cholesterol >240 mg/dl, a patient's self-report of hypercholesterolemia, or the presence of lipid-lowering treatment. Smoking was defined by current use of tobacco. The presence of atrial fibrillation (AF) was documented based on the results of a current or past electrocardiogram. Coronary artery disease was defined as history of myocardial infarction, coronary artery bypass grafting or percutaneous coronary intervention, typical angina, and use of anti-ischemic medications. Carotid duplex Doppler examination to assess for stenosis was performed in 216 stroke patients (84.7%) and in 168 control subjects (80%).
Assessment of coagulation parameters
Since January 1, 1998, fibrinogen, lupus anticoagulant, and prothrombin fragment F 1.2 levels were measured upon presentation in stroke patients and at enrollment in control subjects. Patients in whom coagulation parameters could not be obtained at admission or who were on systemic anticoagulation at the time of the blood drawing were not considered in the analyses involving coagulation. Therefore, all coagulation parameters were available for analysis in 134 stroke patients and in 142 control subjects.
Detection of aortic plaques
Transesophageal echocardiography was performed in stroke patients within 3 days of stroke onset and in control subjects upon enrollment. Transesophageal echocardiographic examination of the aorta was performed in a systematic fashion as previously described (15) using a multiplane transducer. The aortic arch was defined as the portion of aorta comprised between the curve at the end of the ascending portion and the takeoff of the left subclavian artery. A plaque was defined as a discrete protrusion of the intimal surface of the vessel at least 2 mm in thickness, different in appearance and echogenicity from the adjacent intact intimal surface. The presence and location of any plaque was recorded. In cases of multiple plaques, the most advanced lesion was considered. Plaque thickness was measured as a continuous variable. Plaques were classified as small (<4 mm) or large (≥4 mm). The presence of ulcerations or mobile components (Fig. 1) was also recorded. An ulceration was defined as a discrete indentation of the luminal surface of the plaque with base width and maximum depth of at least 2 mm each. Plaques with ulceration and/or mobile components were defined as complex plaques, according to a previously published definition (4). Interpretation of the echocardiographic studies was performed by a single experienced echocardiographer (M.D.T.), blinded to case-control status.
Follow-up and outcome evaluation
All subjects were followed annually by telephone. Any vascular event or acknowledgment of neurological or cardiac symptoms during the standardized interview triggered an in-person assessment. All subjects who screened positive for possible stroke were assessed in person by a neurologist. In addition, active hospital surveillance of admission and discharge International Classification of Diseases-9th Revision codes was performed. Stroke was defined by the first symptomatic occurrence of any type of stroke as defined by TOAST criteria (14). Diagnosis of ischemic stroke was determined by 2 neurologists independently, and the NOMAS principal investigator (R.L.S.) adjudicated any disagreements. Ischemic stroke, fatal and nonfatal, and death were used as the primary outcomes of the study.
Data are presented as mean values ± 1 SD for continuous variables and as proportions for categorical variables. For the case-control component of the study, the frequency of aortic plaques was compared in stroke patients and control subjects. Differences between proportions were assessed by the chi-square test, replaced by the Fisher exact test when the expected cell count was <5. Differences between mean values were assessed by the unpaired Student t test. Unadjusted odds ratios (ORs) for the association between the various definitions of aortic plaque and ischemic stroke were calculated.
Multivariate logistic regression analysis was used to test the independent association between aortic plaques (independent variable) and ischemic stroke (dependent variable) after entering age, gender, and stroke risk factors as potential confounding factors in the model. Variables significantly associated with ischemic stroke at univariate analysis (arterial hypertension, diabetes mellitus, hypercholesterolemia, AF, carotid stenosis ≥60%) were entered as independent variables in the model, along with pertinent demographics (age, gender). The effect of plaque morphology was assessed using indicator variables, which defined plaque thickness and complexity as a series of binary terms. To test the effect of hypercoagulability on the results, coagulation parameters were entered as independent variables in the multivariate models. Interaction terms were created to assess the combined effect of plaque thickness and coagulation abnormalities on stroke risk. Adjusted ORs and 95% confidence intervals (CIs) were calculated from the beta coefficients and the standard errors. The relationship between aortic plaque presence/thickness and coagulation parameters in stroke patients and control subjects was examined by the analysis of variance approach and F-test.
For the prospective component of the study, event rates (stroke and death) were calculated in stroke patients and in control subjects, and again in subjects with no/small plaques or with large plaques. In stroke patients with large plaques, analyses were then stratified by prothrombin fragment F 1.2 levels (below or above the 50th percentile of the distribution). Kaplan-Meier event-free curves were constructed in patients with different F 1.2 levels. The difference between groups was evaluated by means of the log-rank test. SAS statistical package (version 9.1.3, SAS Institute, Cary, North Carolina) was used in the analyses. A 2-tailed p value of ≤0.05 was considered significant for all analyses.
The study population was comprised of 255 patients with first-ever ischemic stroke and 209 stroke-free control subjects matched to patients by age, gender, and race/ethnicity. Demographics and cardiovascular risk factors in stroke patients and in control subjects are shown in Table 1. Stroke patients were significantly older than control subjects, and had a significantly higher prevalence of hypertension and diabetes. History of AF and carotid stenosis ≥60% was also more frequent in stroke patients than in control subjects.
Large (≥4 mm) and complex (i.e., ulcerated or mobile) plaques in the aortic arch were more frequent in stroke patients than in control subjects (Table 1). Ulceration or mobile components were almost exclusively seen in large plaques, with only 1 occurrence among small plaques. Among coagulation parameters, fibrinogen and prothrombin fragment F 1.2 levels were also significantly higher in stroke patients (Table 1).
Table 2 illustrates the relationship between plaque thickness/morphology and ischemic stroke. Plaques ≥4 mm thick were associated with stroke in both univariate analysis and after adjustment for demographics and other stroke risk factors. The presence of plaque ulceration or mobile components further enhanced the strength of the association of plaques with stroke.
Table 3 summarizes the relationship between plaque presence/thickness and coagulation parameters. In stroke patients, an increase in F 1.2 levels was observed with increasing plaque thickness (p = 0.07). In particular, a significant elevation in F 1.2 levels was observed in patients with large plaque compared with patients with no plaque (p = 0.02). A similar trend, but no statistical significance, was observed for fibrinogen, whereas no relationship with plaque thickness was observed for lupus anticoagulant. Since AF can cause both blood stagnation and activation of coagulation, the same analysis was performed after exclusion of subjects with history of AF. This subanalysis showed similar results, with a significant elevation in F 1.2 levels in stroke patients with large plaques compared with patients with no plaque (from 1.23 ± 0.42 ng/ml to 2.01 ± 1.61 ng/ml; p = 0.02), and also an elevation in fibrinogen levels (from 331.9 ± 96.9 mg/ml to 398.5 ± 115.4 mg/ml; p = 0.004).
In multivariate logistic analysis adjusted for other stroke risk factors, there was a trend toward a significant effect of the interaction between large plaque and F 1.2 levels (as continuous variable) on the risk of stroke (adjusted OR: 1.61, 95% CI: 0.93 to 2.79). No trends toward interactions were observed between plaque thickness and the other coagulation parameters. No significant relationship between plaque presence/thickness and any of the coagulation parameters was observed in control subjects.
Table 4 illustrates the combined effect of plaque thickness/complexity and F 1.2 levels (below or above the median value) on the stroke risk. The effect of large plaques was enhanced by the associated elevation of F 1.2.
Study subjects were followed-up for stroke or death for an average 55.1 ± 37.2 months. Follow-up information was available in 219 stroke patients (86%) and in all 209 control subjects. A total of 89 events occurred, 64 in stroke patients (29.2%) and 25 in control subjects (12.0%; p < 0.001) (Table 5). Incidence rates of stroke or death were 133 events per 1,000 person-years in stroke patients and 17 events per 1,000 person-years in control subjects (p < 0.001).
In subjects who had all necessary data, we correlated the event incidence rates with the presence and thickness of aortic plaque and with F 1.2 levels. In stroke patients, large plaques were associated with increased event rates in subjects with F 1.2 levels above the median value compared with those seen in subjects with F 1.2 levels below the median value (230 events per 1,000 person-years vs. 85 events per 1,000 person-years; p = 0.05). In patients with no or small arch plaques, no significant difference was observed between subjects with F 1.2 levels above or below the median value (174 events per 1,000 person-years vs. 131 events per 1,000 person-years; p = 0.58). No significant differences were observed for these comparisons in control subjects.
Figure 2 illustrates the Kaplan-Meier event-free survival probability in stroke patients according to initial plaque thickness and F 1.2 levels. Subjects with large aortic plaques and F 1.2 levels above the median value had significantly worse outcome than those with large plaques but lower F 1.2 levels (p = 0.046).
To our knowledge, this is the first study to report on the impact of the activation of coagulation parameters on the risk of stroke in patients with aortic arch plaques. In our study, patients presenting with acute strokes and large aortic arch plaques showed an elevation in F 1.2 levels that paralleled the plaque thickness, and was not observed in stroke-free control subjects of similar age and with similar plaque thickness. Stroke patients with large plaques and F 1.2 elevation had a significantly worse risk of recurrent stroke or death than those with large plaques but lower F 1.2 levels. These combined observations suggest that hypercoagulability may contribute to the increase in stroke risk associated with large aortic plaques.
Aortic plaques, hypercoagulability, and ischemic stroke
In the present study, large aortic arch plaques were associated with a 2.4-fold increase in the risk of ischemic stroke after adjustment for other stroke risk factors, and with a 3.3-fold increase in risk when ulcerations or mobile components were present on the plaque surface. This finding is consistent with previous data in the literature published by us and others (2–9,11). The potential for cerebral embolization of large and complex plaque has also been demonstrated by detecting the actual delivery of microemboli to the brain with transcranial Doppler monitoring (16). The novel aspect of our study is the association between plaque thickness and hypercoagulability, and their combined effect on the stroke risk. The embolic potential of an aortic plaque is related to its lipid content and/or to the presence of superimposed thrombus. The presence of thrombus, visible in most cases on TEE as a mobile, faintly echogenic component, has occasionally been confirmed at surgery (17–19). In an autopsy study (20), superimposed thrombus was detected on 17 of 120 plaques (14%). Moreover, thromboembolism appears to be much more frequent than atheroembolism in patients with severe aortic arch plaques, with a frequency reported to be as high as 33% at 1 year (5) compared with only 0.7% for atheroembolism (13). It is, therefore, conceivable that the coexistence of a hypercoagulable state in a patient with aortic arch plaque may increase the likelihood of superimposed thrombus formation and further enhance the embolic potential of the plaque. In a recently published study (21), plaque progression over time was associated with increased risk of recurrent stroke and was predicted by the levels of homocysteine, whose procoagulant effects are well known (22). Among coagulation parameters, lupus anticoagulant was chosen in our study because of preliminary data on its possible association with aortic plaque thickness (23). Fibrinogen, a coagulation factor, has also been shown to be independently associated with thoracic aortic atherosclerosis (24). Prothrombin fragment F 1.2 is considered the best indicator of thrombin generation (25). Fragment F 1.2 is the byproduct of the conversion of prothrombin to thrombin, which is the final pathway of both the intrinsic and extrinsic coagulation mechanisms. Therefore, fragment F 1.2 conceivably represents a good indicator of the net result of all procoagulant and anticoagulant influences in the body. In our study, fibrinogen levels were significantly higher in stroke patients than in control subjects, not surprisingly for an acute phase reactant. Fragment F 1.2 levels were also significantly higher in stroke patients, also as reported before (26,27). However, a significant increase in F 1.2 levels with increasing aortic plaque thickness was observed in stroke patients, but not in control subjects. A similar trend was also observed for plaque thickness and fibrinogen levels. This observation is compatible with the hypothesis that the embolic potential of a plaque is enhanced in the presence of an activation of the coagulation process, and that such activation during an acute disease is proportional to the severity of the plaque. Moreover, a trend toward an interaction of plaque thickness and F 1.2 levels on the stroke risk was observed, suggesting that the combined effect of these 2 conditions on the stroke risk may be greater than that of either condition alone. No relationship between plaque thickness and F 1.2 or fibrinogen levels was observed in control subjects, suggesting that plaques, even of large size, which are incidentally detected in otherwise healthy individuals, do not carry the same high embolic potential as those found in stroke patients, possibly because of the absence of a coexisting activation of the coagulation process.
These observations of the case-control study were corroborated and enhanced in the follow-up component of our study. Plaques ≥4 mm in thickness were associated with a significantly increased incidence of recurrent stroke and death, but only in the subgroup with initial F 1.2 levels above the median value (Fig. 2). This finding further reinforces the concept that the combination of plaque thickness and coagulation activation, rather than plaque thickness alone, may be associated with worse outcomes, and that the reported association between aortic plaques and increased stroke risk may have been partly mediated through activation of the coagulation process.
Comparison with previous studies and therapeutic implications
To our knowledge, no previous studies have examined the effect of hypercoagulability on the risk of ischemic stroke in patients with aortic plaques. As a consequence, the use of antithrombotic treatment in stroke patients with aortic plaques has largely been empirical, and no specific recommendations to prevent recurrent stroke are contained in the current American Heart Association/American Stroke Association guidelines (28), except for the general recommendations for lipid-lowering treatment that are applicable to all patients with evidence of atherosclerosis. When a mobile thrombus is visualized on the plaque, systemic anticoagulation is usually initiated to prevent recurrent stroke, although data on its efficacy have been obtained in small, nonrandomized studies (12) or in patients with AF (13). In case of plaques ≥4 mm thick but without superimposed mobile thrombus, antiplatelet agents are sometimes prescribed without real evidence-based data. Only one study in the literature (29) reported a beneficial effect of warfarin treatment over aspirin or ticlopidine on the risk of embolic events in patients with plaques ≥4 mm without mobile thrombus. That study, however, was small, and the treatment was not randomized.
The results of the present study, while not sufficient to suggest the use of systemic anticoagulation in this patient group, do indicate that stroke patients with large nonmobile aortic plaque should be tested for hypercoagulability, and especially prothrombin fragment F 1.2 elevation, because of the increased risk of recurrent stroke and death that is associated with the combination of these 2 conditions. Specifically designed treatment trials are necessary to evaluate whether systemic anticoagulation may decrease the risk of subsequent events in this patient population.
Strengths and limitations
Our study is the first to assess in a systematic fashion the association of proximal aortic atherosclerosis with hypercoagulability, and its impact on the risk of recurrent ischemic stroke and death. The results may be of importance to provide context for planning treatment trials to decrease vascular events in this patient population. Although the study involved the performance of the semi-invasive TEE, control subjects were community-based stroke-free volunteers rather than patient control subjects, which eliminated the possibility of referral bias and allowed for a more accurate estimation of the differences with stroke patients.
The study has some limitations. Carotid duplex Doppler was not performed in all subjects. Coagulation variables were not available or interpretable (because of pre-existing anticoagulant treatment) in all subjects. Homocysteine levels were not measured. Also, 14% of stroke patients were lost to follow-up; therefore, their influence on the results is unknown. Finally, the relatively small sample size forced the use of a composite end point of stroke and death for the prospective component of the study.
Our study shows that hypercoagulability is associated with aortic plaque thickness in patients presenting with an acute ischemic stroke, and that the combination of large aortic plaques and hypercoagulability is associated with an increased risk of recurrent stroke and death. Further studies are necessary to explore the hypothesis that systemic anticoagulation may decrease the risk of events in this patient population.
The authors thank Inna Titova, MPH, and Gabrielle Gaspard, BS, for their assistance in the collection of the data.
The study was supported by R01 NS36286 from the National Institute of Neurological Disorders and Stroke (NINDS). Dr. Di Tullio was the recipient of an NINDS Mid-Career Award in Patient-Oriented Research (K24 NS02241).
- Abbreviations and Acronyms
- atrial fibrillation
- confidence interval
- odds ratio
- transesophageal echocardiography
- Received January 15, 2008.
- Revision received March 14, 2008.
- Accepted April 16, 2008.
- American College of Cardiology Foundation
- Tunick P.A.,
- Rosenzweig B.P.,
- Katz E.S.,
- Freedberg R.S.,
- Peres J.L.,
- Kronzon I.
- Cohen A.,
- Tzourio C.,
- Bertrand B.,
- Chauvel C.,
- Bousser M.G.,
- Amaranco P.
- Kronzon I.,
- Tunick P.A.
- Dressler F.A.,
- Craig W.R.,
- Castello R.,
- Labovitz A.J.
- Adams H.P. Jr..,
- Bendixen B.H.,
- Kappelle L.J.,
- et al.
- Di Tullio M.R.,
- Sacco R.L.,
- Homma S.
- Rundek T.,
- Di Tullio M.R.,
- Sciacca R.R.,
- et al.
- Vaduganathan P.,
- Ewton A.,
- Nagueh S.F.,
- Weilbaecher D.G.,
- Safi H.J.,
- Zoghbi W.A.
- Khatibzadeh M.,
- Mitusch R.,
- Stierle U.,
- Gromoll B.,
- Sheikhzadeh A.
- Sen S.,
- Hinderliter A.,
- Sen P.K.,
- et al.
- Di Tullio M.,
- Sacco R.L.,
- Santoni-Rugiu F.,
- et al.
- Kargman D.E.,
- Hurlet-Jensen A.,
- Gu Q.,
- Sacco R.L.
- Sacco R.L.,
- Adams R.,
- Albers G.,
- et al.
- Ferrari E.,
- Vidal R.,
- Chevallier T.,
- Baudouy M.