Atrial fibrillation is associated with ischemic stroke because of thrombi that form within the left atrial appendage (LAA). The aim of this study was to develop a new parameter for LAA function that is easily performed using transthoracic echocardiography (TTE).
TTE and transesophageal echocardiography were performed in 106 patients with stroke. LAA wall motion velocity (TTE-LAWV) was measured using Doppler tissue imaging at the LAA tip.
TTE-LAWV was significantly lower in patients with atrial fibrillation and LAA thrombus than in those with atrial fibrillation and no LAA thrombus and in sinus rhythm (7.5 ± 1.9 vs 10.0 ± 3.4 and 13.8 ± 5.7 cm/s, respectively, P < .05). TTE-LAWV was significantly correlated with LAA emptying flow velocity ( R = 0.462, P < .05). The multivariate logistic regression analysis showed that TTE-LAWV < 8.7 cm/s was an independent predictor of LAA thrombus formation (odds ratio, 9.473; 95% confidence interval, 1.172-76.55; P < .05).
TTE-LAWV can noninvasively evaluate LAA dysfunction and assist in the detection of LAA thrombus.
The National Institute of Neurological Disorders and Stroke has characterized cardioembolic stroke as an important clinical issue because it is the most common cause of death in patients with acute ischemic stroke. It is well known that the left atrial appendage (LAA) is a major thromboembolic source in patients with stroke and atrial fibrillation (AF). Many clinical reports have shown a close relation between LAA thrombus formation and left atrial mechanical remodeling on the basis of findings from transesophageal echocardiography (TEE). It was reported that the presence of spontaneous echocardiographic contrast (SEC) or LAA peak flow velocity measured by TEE is useful for the detection of LAA dysfunction, which causes LAA thrombus formation. However, TEE may not be performed for screening to quantify the risk for stroke in patients with AF, because it is a semi-invasive procedure. Although transthoracic echocardiography (TTE) can be widely used for screening because of its noninvasive nature, it is thought to be difficult to detect LAA thrombus and evaluate LAA dysfunction on TTE. Doppler tissue imaging (DTI) reflects regional myocardial function. Recently, DTI was also reported to be useful for evaluation of LAA function. In that study, using DTI with TEE, it was shown that LAA dysfunction plays a role in LAA thrombus formation.
We hypothesized that DTI velocity from the LAA on TTE (TTE-LAWV) is a feasible parameter for the risk stratification of patients with AF. In the present study, we compared TTE-LAWV with conventional markers of LAA dysfunction and investigated whether TTE-LAWV could predict thrombus formation.
We performed TTE and TEE in 106 patients referred for the treatment of acute cerebral infarction. TTE and TEE were done within 7 days of onset (mean, 6 ± 1 days). We excluded patients with malignant disease (n = 6), chronic disseminated intravascular coagulation (n = 3), failure of TEE (n = 5), and failure of TTE-LAWV measurement (n = 10). The admission assessment included determining the risk factors for cerebral infarction, clinical ischemic stroke category (National Institute of Neurological Disorders and Stroke ), and disease severity using the National Institutes of Health Stroke Scale.
We defined patients who had no histories of AF and no documented AF on continuous electrocardiographic monitoring during hospitalization as the sinus rhythm group, patients who had histories of AF before admission and/or continuously documented AF on continuous electrocardiographic monitoring during hospitalization as having chronic AF, and patients who had AF at admission and recovered to normal sinus rhythm during hospitalization and/or showed sinus rhythm at admission and documented transient AF on continuous electrocardiographic monitoring during hospitalization as having paroxysmal AF. Six patients (6%) had paroxysmal AF (all were in sinus rhythm at the time of TEE), and 48 patients (45%) had chronic AF. Previous anticoagulation before onset was performed in 2 patients with paroxysmal AF (33%) and 25 patients (52%) with chronic AF. Patients with paroxysmal AF and those with chronic AF were defined as the AF group.
The study subjects were classified into 3 groups on the basis of the presence of AF and LAA thrombus (group A: patients in sinus rhythm [n = 52]; group B: patients with AF and no LAA thrombi [n = 29]; and group C: patients with AF and LAA thrombi [n = 25]). We compared between transthoracic and transesophageal parameters among the 3 groups retrospectively. The local ethics committee approved the study protocol, and informed consent was given by all subjects.
TTE was done using a Hewlett-Packard Sonos 7500 ultrasound instrument equipped with a sector transducer (carrier frequency, 2.5 or 3.75 MHz) (Hewlett-Packard Corporation, Palo Alto, CA). A 5-MHz phased-array multiplane probe was used for TEE. We examined the following parameters and findings using standard views and techniques: left atrial dimension (LAD), left ventricular end-diastolic dimension, left ventricular percentage fractional shortening measured by TTE, and the presence of an atrial septal aneurysm, a patent foramen ovale, SEC, or LAA thrombus evaluated by TEE. LAA thrombus was diagnosed when a fixed or mobile echogenic mass could be clearly differentiated from the wall of the left atrium or LAA. In patients with AF, echocardiographic measurements were obtained as the mean of 5 consecutive cardiac cycles. All findings were evaluated by two independent experienced echocardiologists who did not know the patients’ clinical and other characteristics, and all echocardiographic measurements obtained by the two echocardiologists had good reproducibility.
TEE-LAWV, defined as LAA peak wall velocity, was measured using tissue Doppler with the sample volume placed at the LAA tip by TEE, as reported previously. Peak wall velocity within each RR interval at diastole was obtained by scanning the appendage at 0°, 30°, 60°, and 90° and was averaged. We could also observe the LAA from parasternal the short-axis view on TTE and measure TTE-LAWV using a similar method as for TEE-LAWV measurement ( Figure 1 ). We could evaluate TTE-LAWV in most patients (92%), except those with poor echocardiographic image quality due to obesity, chronic obstructive pulmonary disease, or emaciation.
LAA emptying flow velocity (LAA eV) was assessed using pulsed-wave Doppler with the sample volume placed 1 cm distal from the mouth of the appendage by TEE. Peak flow velocity within each RR interval at diastole was obtained by scanning the appendage at 0°, 30°, 60°, and 90° and was averaged. We analyzed the correlations between TTE-LAWV and other conventional parameters.
Aortic and Carotid Echocardiographic Studies
Aortic images were obtained after the cardiac examination by TEE. The prevalence of protrusion ≥5 mm and/or mobile plaques in the arch were examined.
Bilateral carotid artery imaging was performed with a 7.5-MHz linear transducer connected to a Sonos 7500 system. The carotid intima-media thickness without protruding atheromatous plaques was measured at end-diastole according to the method reported by Pignoli et al and was obtained as the mean of the bilateral common carotid arteries.
Blood samples were collected to determine the serum hemostatic marker levels at the time of the echocardiographic studies. Antithrombin III, fibrinogen, fibrin-monomer, plasminogen, α 2 -plasmin inhibitor, fibrinogen degradation products, and d -dimer were assessed. General biochemical parameters were measured using routine laboratory methods.
Results are expressed as mean ± SD for continuous variables and as percentages of the total number of patients for categorical variables. Skewed variables are shown as medians and interquartile ranges. Statistical analysis was conducted using StatView version 5.0 (SAS Institute Inc, Cary, NC). Patient characteristics, echocardiographic parameters, and hemostatic markers were compared among groups A, B, and C using Student’s t test for unpaired continuous variables and the χ 2 test for categorical variables. If data were not distributed normally, the Mann-Whitney U test was used. P values < .05 were considered significant. Receiver operating characteristic (ROC) curves were constructed to determine the relevant TTE-LAWV cutoff values for predicting LAA thrombus on the basis of the optimal sensitivity and specificity. Bland-Altman analysis was performed to provide an analysis of agreement between TTE-LAWV and TEE-LAWV. To determine independent predictors of the presence of LAA thrombus for all patients, logistic regression analysis was performed. Significant variables selected in univariate logistic regression analysis ( P < .05) were entered into the multivariate analysis. The intraobserver reliability and the interobserver reliability of TTE-LAWV and TEE-LAWV were assessed in 10 patients by two echocardiologists, each repeated once. By intraclass correlation coefficient, the mean intraobserver reliabilities of TTE-LAWV and TEE-LAWV were 98.6% and 98.1%, respectively. The mean interobserver reliabilities of TTE-LAWV and TEE-LAWV were 98.9% and 97.6%, respectively.
There were no significant differences among the 3 groups in sex distribution; the prevalence of hypertension, diabetes mellitus, and hyperlipidemia; history of previous stroke; and the use of oral antiplatelet medications ( Table 1 ). Group C was significantly older than group A. The incidence of cardioembolic stroke was significantly higher in patients with AF (groups B and C) compared with patients in sinus rhythm (group A). Moreover, patients with AF and thrombi (group C) had a higher incidence of cardioembolic stroke compared with patients with AF without thrombi (group B). CHADS 2 scores were significantly higher in group C than in group B. Medication with anticoagulants was significantly higher in groups B and C than in group A ( Table 1 ).
|Variable||Group A (n = 52)||Group B (n = 29)||Group C (n = 25)|
|Age (y)||67 ± 15||71 ± 13||76 ± 10 ∗|
|Heart rate (beats/min)||73 ± 13||82 ± 12 ∗||81 ± 17 ∗|
|Hypertension||40 (71%)||22 (76%)||18 (72%)|
|Diabetes mellitus||15 (29%)||12 (41%)||8 (32%)|
|Hyperlipidemia||30 (58%)||8 (28%)||6 (24%)|
|Current smoking||29 (56%)||15 (52%)||12 (48%)|
|Previous stroke||12 (23%)||3 (10%)||7 (28%)|
|Atrial fibrillation||0 (0%)||29 (100%)||25 (100%)|
|National Institutes of Health Stroke Scale score||4.0 (0.5-8)||4.0 (2-10)||5.5 (2-14)|
|CHADS 2 score||1.9 ± 0.9||2.2 ± 1.4||3.0 ± 1.4 ∗ †|
|National Institute of Neurological Disorders and Stroke clinical category|
|Cardioembolic stroke||3 (6%)||11 (38%) ∗||23 (92%) ∗ †|
|Atherothrombotic stroke||12 (23%)||5 (17%)||0 (0%)|
|Lacunar stroke||9 (17%)||3 (10%)||2 (8%)|
|Others or undetermined||28 (54%)||10 (35%)||0 (0%) ∗ †|
|Medication before onset|
|Antiplatelets||20 (38%)||12 (41%)||9 (36%)|
|Anticoagulants||4 (8%)||15 (52%) ∗||12 (48%) ∗|
|Prothrombin time–international normalized ratio||1.41 ± 0.39||1.51 ± 0.34 ∗||1.55 ± 0.4 ∗|
LAA thrombi were not detected in group A patients. There were no significant differences in the presence of atrial septal aneurysm, patent foramen ovale, and significant aortic or carotid atherosclerotic plaques among the 3 groups. Groups B and C had significantly larger LADs and smaller left ventricular percentage fractional shortening compared with group A ( Table 2 ). Groups B and C had significantly smaller LAA eV, a higher prevalence of SEC, smaller TEE-LAWV, and smaller TTE-LAWV compared with group A ( Table 2 ). Furthermore, group C had a significantly higher prevalence of SEC, smaller TEE-LAWV, and smaller TTE-LAWV compared with group B ( Table 2 ). As shown in Figure 2 , TTE-LAWV decreased with advancing CHADS 2 score. This result suggested that LAA function was impaired in patients with higher thromboembolic risks.
|Variable||Group A (n = 52)||Group B (n = 29)||Group C (n = 25)|
|LAD (mm)||34 ± 5||46 ± 8 ∗||46 ± 9 ∗|
|LV end-diastolic diameter (mm)||46 ± 6||49 ± 7 ∗||48 ± 6|
|LV fractional shortening (%)||37 ± 6||34 ± 8||32 ± 8 ∗|
|LAA eV (cm/s)||70 ± 12||58 ± 14 ∗||16 ± 10 ∗ †|
|SEC||1 (2%)||11 (38%) ∗||19 (76%) ∗ †|
|LAA thrombus||0 (0%)||0 (0%)||25 (100%) ∗ †|
|Atrial septal aneurysm||5 (10%)||1 (3%)||0 (0%)|
|Patent foramen ovale||3 (6%)||3 (10%)||2 (8%)|
|Aortic plaque ‡||4 (8%)||4 (14%)||3 (12%)|
|Aortic IMT (mm)||2.8 ± 1.3||3.0 ± 1.5||3.3 ± 1.8|
|Carotid plaque §||9 (17%)||4 (24%)||2 (8%)|
|Carotid IMT (mm)||0.8 ± 03||0.8 ± 0.1||0.7 ± 0.2|
|TEE-LAWV (cm/s)||16.5 ± 8.5||10.9 ± 4.4 ∗||7.1 ± 1.4 ∗ †|
|TTE-LAWV (cm/s)||13.8 ± 5.7||10.0 ± 3.4 ∗||7.5 ± 1.9 ∗ †|
|TTE-LAWV < 8.7 cm/s||3 (6%)||10 (34%) ∗||18 (72%) ∗ †|
|C-reactive protein (mg/dL)||0.2 (0.1-0.8)||0.1 (0.1-1.2)||2.1 (0.2-5.6) ∗ †|
|Antithrombin III (%)||102 ± 23||98 ± 21||102 ± 22|
|Fibrinogen (mg/dL)||395 ± 115||423 ± 182||535 ± 242 ∗ †|
|Fibrin-monomer (mg/mL)||5.5 (3.3-8.5)||5.5 (2.9-9.6)||5.8 (2.8-13.5)|
|Plasminogen (%)||94 ± 18||91 ± 13||96 ± 16|
|α 2 -plasmin inhibitor (%)||101 ± 14||105 ± 14||116 ± 18 ∗ †|
|d -dimer (mg/ml)||1.3 (0.5-3.1)||1.5 (0.5-4.3)||1.7 (0.8-4.7)|
|Fibrinogen degradation products (mg/mL)||4.7 (3.5-6.2)||3.9 (2.9-6.8)||5.3 (3.9-9.4)|
There were no significant differences in hemostatic markers between group A and group B ( Table 2 ). However, group C had higher levels of hemostatic markers such as fibrinogen, α 2 -plasmin inhibitor, and C-reactive protein compared with groups A and B.
There was no significant difference in prothrombin time-international normalized ratio between patients with LAA thrombi and patients without LAA thrombi who were taking warfarin (1.55 ± 0.40 vs 1.52 ± 0.33, P = .745; Table 3 ).
|LAA thrombus||No LAA thrombus||LAA thrombus||No LAA thrombus|
|Variable||(n = 12)||(n = 19)||(n = 14)||(n = 61)|
|Prothrombin time–international normalized ratio||1.55 ± 0.40||1.52 ± 0.33||—||—|
|TTE-LAWV (cm/s)||6.9 ± 1.5||11.2 ± 4.7 ∗||8.4 ± 2.1||13.0 ± 5.6 †|