If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
There is an apparent link between thrombogenesis and inflammation. We hypothesized that systemic inflammation [as indicated by C-reactive protein (CRP)] would be related to the presence of left atrial (LA) thrombus in patients with atrial fibrillation (AF). To test this hypothesis, we evaluated the relationship between CRP and LA thrombus in patients with non-rheumatic AF.
Methods and results
Between October 2004 and December 2008, 190 patients with non-rheumatic AF (122 males, age 71 ± 10 years) who underwent transesophageal echocardiography (TEE) were enrolled and analyzed. All patients were examined for presence or absence of LA thrombus by TEE. CRP was measured within 1 week before the TEE examination. LA thrombus was detected in 19 patients (10%). Hypertension, hypertensive heart disease (HHD), valvular heart disease, ticlopidine, and CRP were univariate correlates of LA thrombus. By multivariate analysis, HHD (p < 0.01), ticlopidine (p = 0.01), and CRP (p = 0.03) were independently associated with LA thrombus. A cut-off CRP value for identifying LA thrombus was 0.21 mg/dl (sensitivity: 84%, specificity: 60%, positive predictive value: 19%, and negative predictive value: 97%).
Conclusion
A high CRP is related to LA thrombus in patients with non-rheumatic AF.
]. We hypothesized that systemic inflammation [as indicated by C-reactive protein (CRP)] would be related to the presence of LA thrombus in patients with AF. To test this hypothesis, we evaluated the relationship between CRP and LA thrombus in patients with non-rheumatic AF.
Methods
Study patients
The study population was identified from a retrospective database of 278 consecutive patients with AF who underwent TEE from October 2004 to December 2008. Patients with rheumatic valvular heart diseases (n = 13), prosthetic valve after operation of mitral stenosis (n = 11), and those in whom only the aorta was observed because of aortic dissection (n = 2) were excluded from this study. In 29 cases with multiple TEE examinations (total 88), only the first examination was included in this study. Finally, a total of 190 patients (122 males, 68 females, mean age 71 ± 10 years, range 36–95 years) with AF were enrolled and analyzed. Among these patients, 6 patients had infective endocarditis, 3 patients had aortic dissection, 24 patients had recent embolic events that occurred within 2 weeks (including acute ischemic stroke).
According to the TEE result, the study patients were divided into 2 groups depending on the presence (n = 19) or absence (n = 171) of LA thrombus. In patients with LA thrombus, 10 patients had recent embolic events that occurred within 2 weeks (including acute ischemic stroke).
In addition, patients treated with warfarin therapy [the target prothrombin time-international normalized ratio (PT-INR) value of between 2.0 and 3.0 for patients aged < 70 years or PT-INR value of between 1.6 and 2.6 for patients aged ≥ 70 years] [
Optimal intensity of warfarin therapy for secondary prevention of stroke in patients with nonvalvular atrial fibrillation: a multicenter, prospective, randomized trial Japanese Nonvalvular Atrial Fibrillation-Embolism Secondary Prevention Cooperative Study Group.
Optimal intensity of international normalized ratio in warfarin therapy for secondary prevention of stroke in patients with non-valvular atrial fibrillation.
] for more than 3 weeks at the time of TEE (n = 73, 38.4%) were also divided into 2 groups, with (n = 8, 11.0%) and without (n = 65, 89.0%) LA thrombus.
Transthoracic echocardiography (TTE)
All echocardiographic examinations were performed by using Sonos 7500 (Philips Ultrasound, Bothell, WA, USA) with an S3 probe. All subjects underwent a standard transthoracic 2 dimensional (2D) and Doppler echocardiographic examinations. In addition to routine conventional echocardiographic indices, we measured LA volume. LA volume was measured by the prolate-ellipsoid method, LA volume was calculated as reported previously [
A T6H probe was used for 2D TEE images by using Sonos 7500 (Philips Ultrasound). In TEE, the following features were specifically assessed: (1) presence or absence of thrombus in the LA or LA appendage (LAA); (2) presence or absence of SEC within the LA or LAA; (3) peak emptying velocity of the LAA; (4) presence or absence of significant (≥ moderate) mitral regurgitation (MR).
Thrombus were defined as highly echogenic masses adjacent to the endocardial surface and clearly differentiated from normal structures such as the pectinate muscles. SEC was defined as slowly swirling, smoke-like echoes inside the LA or LAA. Gain was continuously adjusted to ensure good visualization and to avoid noise artifacts. The LAA peak emptying velocity was obtained by Doppler echocardiography. Two experienced cardiologists interpreted the TEE blinded to the laboratory data.
Data collection
In each patient, the following information had been collected as the initial clinical parameters: gender, age, body mass index (BMI), type of antithrombotic therapy (warfarin, heparin, aspirin, or ticlopidine), coexisting conditions including valvular heart disease which includes significant MR, aortic valve regurgitation and aortic valve stenosis, ischemic heart disease, hypertrophic cardiomyopathy, dilated cardiomyopathy, hypertensive heart disease (HHD), congenital heart disease, hypertension, diabetes mellitus, and hyperlipidemia. Lone AF was defined as AF occurring in the absence of structural heart disease and hypertension under the age of 65 years.
Blood samples
Blood samples were taken within 1 week before the TEE examination. In our hospital, laboratory data including CRP were measured routinely before TEE examination. The serum CRP was measured by latex nephelometry (LT Auto Wako CRP, Osaka, Japan). We used latex as the reagent and Hitachi 7500 analyzer (Hitachi, Tokyo, Japan) as the measurement system. The lowest detection CRP limit of this test was <0.02 mg/dl. After blood samples were taken, medications including anticoagulation were not changed until TEE examination was performed.
Statistical analysis
Data are expressed as mean value ± SD or as median value with interquartile range. Differences in clinical features and plasma markers between patients with and without LA thrombus were evaluated with an unpaired Student t-test for normally distributed continuous variables, Mann–Whitney U-test for nonparametrically distributed continuous variables, and chi-square tests for categorical variables. Factors significantly associated with LA thrombus on univariate analysis (p < 0.05) were entered into a stepwise logistic regression analysis to determine independent associates of LA thrombus. Statistical analyses were done with StatView 5.0 software (SAS Institute, Cary, NC, USA). A p-value of <0.05 was considered statistically significant. Inter- and intra-observer agreements for the detection of LA thrombus was assessed by using kappa statistics. A kappa index above 0.8 was considered to indicate very good agreement.
Results
LA thrombus was detected in 19 of 190 patients (10%). Table 1 summarizes the clinical characteristics of patients with and without LA thrombus. There were significant differences in HT, ticlopidine, HHD, valvular heart disease, and CRP between the 2 groups. Duration and intensity [PT-INR and activated partial thromboplastin time (APTT)] of anticoagulation therapy were not different between the 2 groups.
Table 1Clinical characteristics and research indices in patients with and without left atrial (LA) thrombus.
LA thrombus (−)
LA thrombus (+)
p
N
171
19
Age (years)
70.9 ± 9.8
75.4 ± 10.1
0.06
Male (%)
63.7
68.4
0.69
Body mass index (kg/m2)
22.8 ± 3.5
22.5 ± 3.2
0.68
Smoker (%)
36.8
34.5
0.84
Diabetes mellitus (%)
22.2
42.1
0.09
Hypertension (%)
62.0
89.5
0.02
Hyperlipidemia (%)
22.2
42.1
0.09
ACEI/ARB (%)
33.9
47.4
0.24
Warfarin (≥21 days) (%)
9.4
11.0
0.73
Aspirin (%)
22.8
36.8
0.26
Ticlopidine (%)
3.5
21.1
0.01
Statin (%)
18.1
31.6
0.22
Ischemic heart disease (%)
10.5
15.8
0.45
Dilated cardiomyopathy (%)
2.3
0.0
1.00
Hypertrophic cardiomyopathy (%)
6.4
5.3
1.00
Hypertensive heart disease (%)
16.4
63.2
<0.01
Congenital heart disease (%)
3.5
0.0
1.00
Valvular heart disease (%)
45.0
21.1
0.04
Lone atrial fibrillation (%)
2.9
0.0
1.00
White blood cells (/μl)
6225 ± 2195
7824 ± 4041
0.05
Red blood cells (×104/μl)
417 ± 622
447 ± 75
0.05
Hematocrit (%)
39.8 ± 7.5
41.9 ± 5.8
0.08
Platelet (×104/μl)
20.0 ± 5.6
20.1 ± 6.1
0.72
PT-INR
1.7 ± 0.8
1.7 ± 0.6
0.46
APTT (s)
36.2 ± 14.7
34.8 ± 12.3
0.64
Creatinine (mg/dl)
0.93 ± 0.50
0.98 ± 0.27
0.11
C-reactive protein (mg/dl)
0.13 (0.05–0.72)
0.90 (0.33–2.72)
<0.01
Total-cholesterol (mg/dl)
182.4 ± 42.0
196.6 ± 59.2
0.65
Triglycerides (mg/dl)
105.7 ± 60.5
104.9 ± 67.3
0.57
LDL-cholesterol (mg/dl)
109.5 ± 33.5
117.9 ± 50.3
0.52
HDL-cholesterol (mg/dl)
48.6 ± 15.8
49.2 ± 16.3
0.95
Values are mean ± SD, median (interquartile range), or percentage. ACEI, angiotensin-converting enzyme inhibitors; ARB, angiotensin II receptor blocker; PT-INR, prothrombin time-international normalized; APTT, activated partial thromboplastin time; LDL, low density lipoprotein; HDL, high density lipoprotein.
Fig. 1 shows a box plot of CRP in patients with and without LA thrombus. CRP in patients with LA thrombus was significantly higher than in patients without LA thrombus [0.90 (0.33–2.72) vs. 0.13 (0.05–0.72) mg/dl, p < 0.01].
Figure 1Comparison of C-reactive protein (CRP) between patients with and without left atrial (LA) thrombus.
Table 2 shows the TTE and TEE findings. There were significant differences in LA dimension, intraventricular septum, and posterior wall thickness between the 2 groups. In addition, in patients with LA thrombus, the LAA velocity was significantly lower and significant MR was less frequently observed than in patients without LA thrombus. In this study, SEC was found in all patients with LA thrombus. The kappa statistic for inter- and intra-observer agreements for LA thrombus by TEE were 0.94 and 1.00, respectively.
Table 2Findings on transthoracic echocardiography and transesophageal echocardiography.
LA thrombus (−)
LA thrombus (+)
p
TTE findings
LA dimension (cm)
4.5 ± 0.8
4.8 ± 0.9
0.04
LA volume (ml)
70.7 ± 58.1
70.9 ± 36.7
0.53
E prime (cm/s)
7.0 ± 2.5
6.8 ± 1.1
0.58
Intra ventricular septum (cm)
1.1 ± 0.2
1.4 ± 0.4
<0.01
Posterior wall (cm)
1.1 ± 0.2
1.3 ± 0.3
<0.01
LVDd (cm)
4.7 ± 0.8
4.6 ± 0.9
0.63
LVDs (cm)
3.2 ± 0.9
3.1 ± 1.2
0.39
Ejection fraction (%)
57.4 ± 12.2
54.2 ± 15.4
0.56
TEE findings
Spontaneous echo contrast (%)
26.9
100.0
<0.01
LAA velocity (cm/s)
36.7 ± 21.0
18.3 ± 5.9
<0.01
Significant MR (%)
33.7
10.5
0.04
LA, left atrium; TEE, transesophageal echocardiography; TTE, transthoracic echocardiography; LVDd, left ventricular diastolic dimension; LVDs, left ventricular systolic dimension; LAA, left atrial appendage; MR, mitral regurgitation.
By multivariate analysis, HHD (p < 0.01), ticlopidine (p = 0.01), and CRP (p = 0.03) were independent clinical predictors of LA thrombus (Table 3). According to the receiver operating characteristic (ROC) analysis, a cut-off CRP value for identifying LA thrombus was 0.21 mg/dl (sensitivity: 84%, specificity: 60%, positive predictive value: 19%, and negative predictive value: 97%) (Fig. 2).
Table 3Multivariate analysis of characteristics and C-reactive protein levels.
Figure 2Receiver operating characteristics curve to determine the cut-off value of C-reactive protein for left atrial thrombus detection by transesophageal echocardiography.
Among the patients who were receiving warfarin therapy at the time of TEE (n = 77), CRP was also significantly higher in patients with LA thrombus [n = 8, 0.98 (0.67–1.45) vs. 0.11 (0.04–0.36) mg/dl, p < 0.01] (Fig. 3) despite similar INR values between the two groups (2.02 ± 0.71 vs. 2.09 ± 0.78, p = 0.92). Optimal anticoagulation (defined as PT-INR value of between 2.0 and 3.0 for patients aged < 70 years or PT-INR value of between 1.6 and 2.6 for patients aged ≥ 70 years) [
Optimal intensity of warfarin therapy for secondary prevention of stroke in patients with nonvalvular atrial fibrillation: a multicenter, prospective, randomized trial Japanese Nonvalvular Atrial Fibrillation-Embolism Secondary Prevention Cooperative Study Group.
Optimal intensity of international normalized ratio in warfarin therapy for secondary prevention of stroke in patients with non-valvular atrial fibrillation.
According to the ROC analysis, a cut-off CRP value for identifying LA thrombus in patients receiving warfarin therapy was 0.62 mg/dl (sensitivity: 88%, specificity: 81%, positive predictive value: 35%, and negative predictive value: 98%).
On the other hand, among the patients who were receiving heparin therapy (n = 70, 36.8%), LA thrombus was found in 10 patients. CRP showed a trend toward being higher in patients with LA thrombus than those without [0.78 (0.26–4.81) vs. 0.17 (0.08–0.53) mg/dl, p = 0.03] despite similar APTT values between the two groups (38.4 ± 14.7 vs. 38.0 ± 15.6 s, p = 0.82).
Repeated TEE examinations were performed in 10 of 19 patients (53%) with LA thrombus. LA thrombus resolved in 4 patients, whereas LA thrombus persisted in the remaining 6 patients. The baseline CRP at the time of the first TEE was comparable between patients whose LA thrombus disappeared and remained. During follow-up, CRP showed a trend toward decrease in both groups [resolved group: 0.47 (0.15–1.06) to 0.20 (0.04–0.36) mg/dl, p = 0.28, persisted group: 1.24 (0.61–4.26) to 0.77 (0.25–1.24) mg/dl, p = 0.25]. On the other hand, PT-INR significantly increased in patients with resolved LA thrombus (1.51 ± 0.34 vs. 2.63 ± 0.80, p = 0.03), but not in patients with persistent LA thrombus (1.92 ± 0.80 vs. 2.26 ± 0.73, p = 0.48).
Twenty-nine (15%) of 190 patients had evidence of inflammatory diseases that could explain a higher CRP. Fourteen patients had infectious diseases (infective endocarditis, pneumonia, sepsis, and infectious arthritis) and 15 patients had other systemic inflammatory diseases (rheumatoid arthritis, hyperthyroidism, aortitis, glomerulonephritis, carcinoma, and appendicitis). After excluding these patients, CRP was still significantly higher in patients with LA thrombus than in patients without LA thrombus [0.65 (0.26–1.12) vs. 0.12 (0.04–0.42) mg/dl, p < 0.01].
Discussion
To the best of our knowledge, this is the first study demonstrating the relationship between systemic inflammation and LA thrombus in patients with non-rheumatic AF. The predilection of LA thrombus formation in AF patients has long been known. However, the pathogenesis of auricular thrombosis has not been entirely specified yet. Virchow identified a triad of components implicated in the process of thrombosis [
]: abnormal conditions of blood flow, vessel wall damage, and abnormal blood constituents. In AF, there are variable relations to these three components: (1) LAA velocity; (2) atrial endothelial cells; and (3) prothrombotic state. Previous studies have suggested an association between decreased blood flow in the LAA as demonstrated by reduced LAA flow velocities and the loss of atrial contraction in AF [
Transesophageal echocardiographic correlates of thromboembolism in high-risk patients with nonvalvular atrial fibrillation. The Stroke Prevention in Atrial Fibrillation Investigators Committee on Echocardiography.
]. The CHADS2 [Congestive heart failure, Hypertension, Age, Diabetes, Stroke (Doubled)] score is widely used as a simple and reliable clinical score to identify those with high likelihood of ischemic stroke among patients with AF [
]. In fact, HHD was independently associated with LA thrombus in our present study population. Although previous studies have suggested that LA thrombus was more frequently found in patients with low left ventricular (LV) ejection fraction [
], our data did not support these previous reports, possibly because of small sample size. Our study population mainly consisted of patients with preserved LV systolic function and therefore did not show impact of LV systolic function on LA thrombus formation.
In our present study, incidence of significant MR in the LA thrombus group was significantly lower than in patients without LA thrombus, concordant with previous reports. Previous studies suggested that significant MR may be protective against the formation of SEC/LA thrombus [
Significant mitral regurgitation is protective against left atrial spontaneous echo contrast and thrombus as assessed by transesophageal echocardiography.
A significant difference was present between LA thrombus and ticlopidine. This was an unlikely finding that may have been due to chance alone, considering the small number of patients who were on ticlopidine.
Several studies demonstrate that patients with AF show significant increases in plasma fibrinogen and D-dimer levels, suggesting the presence of a hypercoagulable or pro-thrombotic state [
]. In this study, plasma D-dimer level was measured only in 70 patients (37%) at the time of TEE. Plasma D-dimer levels were significantly higher in patients with LA thrombus than in those without LA thrombus [2.20 (1.20–8.15) vs. 0.50 (0.50–1.50) μg/ml, p < 0.01].
In addition, a previous study has demonstrated that there is a possible link between CRP and deep vein thrombosis [
]. Therefore, CRP may play some role in LA thrombus formation. Furthermore, CRP has been suggested as a biomarker to predict cardiovascular events in apparently healthy subjects as well as patients with coronary artery disease or valvular heart disease [
Impact of statin therapy on systemic inflammation, left ventricular systolic and diastolic function and prognosis in low risk ischemic heart disease patients without history of congestive heart failure.
]. However, the exact mechanisms by which CRP affects cardiovascular disease are still poorly understood and controversial.
In the present study, we found that elevated CRP was independently associated with the presence of LA thrombus. Our findings are consistent with those of previous studies reporting associations between inflammation and coagulation [
]. It has been reported that CRP promotes platelet adhesion to endothelial cells and emphasize the possible role of CRP in linking inflammation and thrombosis and provide a potential mechanism for the high incidence of vascular events associated with high CRP level [
]. Furthermore, a recent report from the Stroke Prevention in Atrial Fibrillation (SPAF)-III study demonstrated that CRP was positively correlated to stroke risk and related to stroke risk factors and prognosis in 880 patients with AF [
High-sensitivity C-reactive protein and soluble CD40 ligand as indices of inflammation and platelet activation in 880 patients with nonvalvular atrial fibrillation: relationship to stroke risk factors, stroke risk stratification schema, and prognosis.
]. Our present results may explain the possible link between stroke risk and systemic inflammation in patients with AF.
An epidemiological study as well as a large-scale randomized trial consistently demonstrated that the incidence of stroke in patients with AF is related to the size of the LA [
Predictors of thromboembolism in atrial fibrillation: II. Echocardiographic features of patients at risk. The Stroke Prevention in Atrial Fibrillation Investigators.
]. Therefore, it is possible that some prothrombotic alterations in the endocardium may occur during the process of structural remodeling in the LA wall. Yaron et al. have shown that CRP directly affected the endothelial phenotype promoting thrombosis [
]. Therefore, in addition to the impact of the LA enlargement, CRP may modulate LA endothelial function leading to thrombogenic status in patients with AF.
Acute embolic events may also be related to elevated CRP levels. In this study, 24 (13%) of 190 patients had recent embolic events that occurred within 2 weeks. Among these patients, 10 had LA thrombus. After excluding these patients with recent embolic events, CRP was still significantly higher in patients with LA thrombus than in patients without LA thrombus [0.90 (0.50–1.24) vs. 0.13 (0.05–0.49) mg/dl, p < 0.01].
The efficacy of oral anticoagulant therapy in reducing the risk of thromboembolic events has been demonstrated in patients with AF. To optimize the intensity of anticoagulation, as indicated by the INR, a target INR needs to achieve the best balance between the prevention of thromboembolic events and the occurrence of bleeding complications [
Adjusted-dose warfarin versus low-intensity, fixed-dose warfarin plus aspirin for high-risk patients with atrial fibrillation: Stroke Prevention in Atrial Fibrillation III randomised clinical trial.
]. Interestingly, among our study patients on warfarin, INR value did not differ between patients with and without LA thrombus. In fact, there were some cases with LA thrombus even under an acceptable INR value as recommended by guidelines. In addition, there was no correlation between CRP and INR (r = 0.16, p = 0.73). Therefore, inflammation may be related to LA thrombus independent of anticoagulation states. A multicenter, prospective and randomized study from Japan demonstrated that low intensity warfarin treatment (INR 1.5–2.1) for prevention of stroke recurrence was safer than conventional intensity treatment (INR 2.2–3.5) in the elderly [
Optimal intensity of warfarin therapy for secondary prevention of stroke in patients with nonvalvular atrial fibrillation: a multicenter, prospective, randomized trial Japanese Nonvalvular Atrial Fibrillation-Embolism Secondary Prevention Cooperative Study Group.
]. Yasaka et al. demonstrated a sharp rise in the incidence of severe hemorrhage in INR ≥ 2.6 and reported that most patients suffering severe hemorrhage were elderly [
Optimal intensity of international normalized ratio in warfarin therapy for secondary prevention of stroke in patients with non-valvular atrial fibrillation.
]. Therefore, an INR value of between 1.6 and 2.6 seems optimal to prevent major ischemic or hemorrhagic events in elderly Japanese non-valvular AF patients. However, we found that, some cases, even if they had maintenance of the INR within the range which guidelines recommended, had complicated LA thrombus. In such cases, CRP levels were significantly higher than in those without LA thrombus. Therefore, presence of inflammation as evident by high CRP level may also help stratify high-risk patients for LA thrombus formation and possibly stroke and peripheral embolism in non-rheumatic AF. In this study, we identified the cut-off CRP value for identifying LA thrombus as 0.21 mg/dl. In patients receiving warfarin therapy, the cut-off CRP value was 0.62 mg/dl. Based on our results, the INR value alone may not be enough to predict LA thrombus formation among patients with high CRP. Although LA thrombus resolved in 4 of 10 patients who underwent serial TEE examination, CRP level did not significantly change in these patients. The small sample size and a difference in intensity of anticoagulation possibly explain the lack of relationship between serial changes in CRP and LA thrombus.
Limitations
First, this was a single-center retrospective study performed in a relatively small number of patients. Second, antithrombotic therapies were not randomized. The formation of LA thrombus is critically dependent upon anticoagulation status and anticoagulation duration. Third, incidence of LA thrombus in patients with recent embolic events may be underestimated because LA thrombus had already been embolized and thus disappeared. Fourth, although LA thrombus was related to high CRP level, causal relationship between prothrombotic state and inflammation is unclear. In addition, although we found a cut-off CRP value for identifying LA thrombus, inflammation is not the only predictor of LA thrombus. Therefore, a cut-off CRP value may not be used for predicting LA thrombus but for excluding LA thrombus. Finally, it is unknown whether our present results can be applicable to all patients with AF.
Conclusion
Systemic inflammation is related to LA thrombus formation in patients with non-rheumatic AF. Our results indicate that presence of systemic inflammation as evident by high CRP level may help stratify high-risk patients for LA thrombus formation and possibly stroke and peripheral embolism in non-rheumatic AF.
Disclosures
None.
References
Wolf P.A.
Abbott R.D.
Kannel W.B.
Atrial fibrillation as an independent risk factor for stroke: the Framingham Study.
Optimal intensity of warfarin therapy for secondary prevention of stroke in patients with nonvalvular atrial fibrillation: a multicenter, prospective, randomized trial Japanese Nonvalvular Atrial Fibrillation-Embolism Secondary Prevention Cooperative Study Group.
Optimal intensity of international normalized ratio in warfarin therapy for secondary prevention of stroke in patients with non-valvular atrial fibrillation.
Transesophageal echocardiographic correlates of thromboembolism in high-risk patients with nonvalvular atrial fibrillation. The Stroke Prevention in Atrial Fibrillation Investigators Committee on Echocardiography.
Significant mitral regurgitation is protective against left atrial spontaneous echo contrast and thrombus as assessed by transesophageal echocardiography.
Impact of statin therapy on systemic inflammation, left ventricular systolic and diastolic function and prognosis in low risk ischemic heart disease patients without history of congestive heart failure.
High-sensitivity C-reactive protein and soluble CD40 ligand as indices of inflammation and platelet activation in 880 patients with nonvalvular atrial fibrillation: relationship to stroke risk factors, stroke risk stratification schema, and prognosis.
Predictors of thromboembolism in atrial fibrillation: II. Echocardiographic features of patients at risk. The Stroke Prevention in Atrial Fibrillation Investigators.
Adjusted-dose warfarin versus low-intensity, fixed-dose warfarin plus aspirin for high-risk patients with atrial fibrillation: Stroke Prevention in Atrial Fibrillation III randomised clinical trial.
00060-2&id=gr1.jpg){kind=link}
00060-2&id=gr2.jpg){kind=link}
00060-2&id=gr3.jpg){kind=link}