Prediction of Left Atrial Appendage Thrombi in Non-Valvular Atrial Fibrillation

Seiji Habara; Keigo Dote; Masaya Kato; Shota Sasaki; Kenji Goto; Hiroaki Takemoto; Daiji Hasegawa; Osamu Matsuda

Eur Heart J.  2007;28(18):2217-2222.  2007 Oxford University Press
Posted 10/30/2007

Abstract and Introduction

Abstract

Aims: There is little knowledge about the predictors of left atrial appendage (LAA) thrombi in non-valvular atrial fibrillation (NVAF). We investigated the ability of D-dimer to predict LAA thrombi.
Methods and results: In this study, 925 patients with NVAF were enrolled. At the time of transoesophageal echocardiography (TEE), D-dimer levels were measured simultaneously. Significant independent predictors of LAA thrombi were the presence of congestive heart failure [odds ratio (OR) 3.10, 95% confidence interval (CI) 1.775.50, P < 0.0001), a history of recent embolic events (OR 3.39, 95% CI 1.906.04, P < 0.0001), and D-dimer levels (OR 97.6, 95% CI 17.3595.8, P < 0.0001). Receiver operating characteristic analysis yielded an optimal cutoff value of 1.15 g/mL for D-dimer to detect LAA thrombi. LAA thrombi were detected in 21.8% of patients with higher D-dimer values, whereas it was detected in only 3.1% of patients with lower D-dimer values. D-dimer cutoff level of 1.15 g/mL had a negative predictive value of 97% for identifying LAA thrombi.
Conclusion: In patients with NVAF, D-dimer may be helpful for predicting the absence of LAA thrombi. D-dimer level was clinically useful to guide the management of patients with NVAF, especially for those complicated with congestive heart failure and/or recent embolic events.

Introduction

Atrial fibrillation, the most common sustained cardiac rhythm disturbance, increases in prevalence as the population ages.[1,2] Embolic stroke is one of the most feared complications of atrial fibrillation. Compared with patients with sinus rhythm, those with atrial fibrillation have a four- to five-fold increased risk of ischaemic stroke.[2] Transoesophageal echocardiography (TEE) allows thrombi in the left atrial appendage (LAA) to be detected with a high degree of accuracy.[3,4] There is no laboratory test to aid the diagnosis of LAA thrombi. D-dimers appear to be a useful parameter for assessing the degrees of hypercoagulability. Therefore, it may have a potential role as a non-invasive marker of the presence of thrombi and risk for thrombo-embolism. In the investigation and management of deep venous thrombosis, pulmonary embolism, and acute aortic dissection, the use of D-dimer is well established.[57] A recent study with a small number of patients shows that D-dimer is a reliable parameter to exclude the presence of atrial thrombi in patients with atrial fibrillation.[8] We investigated the ability of D-dimer to predict LAA thrombi in several clinical situations.

Methods

Study Population

Between February 1992 and February 2005, 1200 consecutive patients with non-valvular atrial fibrillation (NVAF) were eligible. The definition and classification of atrial fibrillation used in this study were based on published guidelines from the American College of Cardiology-American Heart Association and the European Society of Cardiology.[9] At the time of TEE, plasma D-dimer level was measured simultaneously. Excluded were patients with organic valvular heart diseases, presence of prosthetic valve, aortic aneurysm, left ventricular thrombus, aortic dissection, deep vein thrombosis, and pulmonary embolism. All patients were assessed by a nine-item prediction rule to preliminarily categorize their pretest probability of deep vein thrombosis.[10] If patients were categorized as 'high pretest probability', we examined deep vein thrombosis using venous ultrasonography or enhanced computed tomography. Informed consent was obtained from all patients after explanation of the study, and the study was approved by the institutional Ethics Committee.

Ultrasound Evaluation

All patients underwent transthoracic echocardiography and TEE. A Philips medical ultrasonograph (Model SONOS 2500 or SONOS 5500, Philips Medical Systems, MA, USA) with a 2.5 MHz transducer and a 5 MHz omniplane probe was used to obtain transthoracic echocardiogram and TEE. Before TEE study, the procedure was explained in greater detail, and written informed consent was obtained. Lidocaine hydrochloride spray was routinely used for local anaesthesia, which should cover the posterior pharynx and tongue. We used 5-10 mg diazepam intravenously for sedation, if necessary. Sedative was used for ∼30% of our patients. We usually performed TEE within 15 min. The failure rate of TEE was <1% (10 patients, owing to the failure of oesophageal intubation). A careful survey of the entire atria, including the appendages, was performed by TEE. Maximal LAA areas were measured by tracing a line from the top of the upper pulmonary vein limbus along the entire endocardial LAA border. Left atrial cavity and LAA were closely inspected for the presence of thrombi, and spontaneous echo contrast (SEC) was graded from 0 to 4 following the classification of Fatkin et al.[11] The LAA-emptying peak flow velocity signals within each R-R interval were averaged over a minimum of eight cardiac cycles. LAA thrombi were recognized by identifying a mobile or sessile, irregularly shaped, grey, textured density that was clearly separate from the lining of the atrial appendage.[12] It was a clearly defined echo-dense intracavitary mass, which was acoustically distinct from the endocardium.[12] Echocardiographers were blinded to D-dimer levels. In patients with paroxysmal atrial fibrillation (PAF), TEE was performed before they recovered to normal sinus rhythm. All TEE data were analysed by two independent cardiologists. When the results were different between the two cardiologists, a final decision was made by consensus. The interobserver intraclass correlation coefficient for the detection of thrombi was 0.99 [95% confidence interval (CI) 0.98-0.99], and that for the evaluation of SEC grade by TEE was 0.97 (95% CI 0.96-0.98). The intraobserver intraclass correlation coefficient for the detection of thrombi was 0.99 (95% CI 0.98-0.99), and that for the evaluation of SEC grade by TEE was 0.99 (95% CI 0.98-0.99).

Blood Sampling and Assays

All assays were performed in the laboratory of our institute, and investigators and laboratory personnel were blinded to the clinical status. For the quantitative determination of D-dimers in sodium citrate plasma, a latex-enhanced photometric immunoassay (LPIA) (Mitsubishi Kagaku Iatron, Tokyo, Japan) was used with a pathfast immunoassay analyser (Mitsubishi Kagaku Iatron). Within 90 min of collection, processing and analysis were performed. There was no change in the measuring method of fibrin D-dimer over the study period. Intra-assay variability was determined by assessing 30 individual blood samples, which was 3.9%.

Statistical Analysis

The SPSS statistical software package (version 11.0, SPSS Inc., Chicago, IL, USA) was used for all statistical calculations. The data were expressed as mean SD, and categorical variables as percentages. The optimal cutoff points were determined by receiver operating characteristic (ROC) curves, and the sensitivity and specificity of D-dimer levels for group distinction were determined according to: sensitivity = true positives/(true positives + false negatives) 100%, specificity = true negatives/(true negatives + false positives) 100%, positive predictive value = true positives/(true positives + false positives) 100%, and negative predictive value = true negatives/(true negatives + false negatives) 100%. Continuous variables were compared by unpaired t-test. Discrete variables were compared by χ2 test with Yates' correction. Univariate logistic regression analysis was used to assess the association of clinical and laboratory variables with the presence of LAA thrombi. Stepwise multivariable logistic regression analysis was then applied to individuate the variables independently associated with the presence of LAA thrombi. Only variables with a value of P < 0.05 on univariate analysis were included in the multivariable model. Values of P < 0.05 were considered to be statistically significant.

Results

Patients Population

We identified 925 patients (68.8 10.3 years) with NVAF (including 250 patients with PAF, 84 with persistent atrial fibrillation, and 591 with permanent atrial fibrillation). Among these patients, 583 patients (63.0%) had hypertension, 208 (22.5%) had congestive heart failure defined as New York Heart Association (NYHA) class II or higher, and 208 (22.5%) had recent embolic events that occurred within 2 weeks (including acute ischaemic stroke); 158 patients had cerebral embolic events which were classified as cerebrovascular accident in 103 and transient ischaemic attack in 55, and 50 patients had peripheral (n = 21), mesenteric artery (n = 11), splenic artery (n = 4), coronary artery (n = 4), or renal artery (n = 10) embolism.

Left Atrial Appendage Thrombi

Eighty patients had thrombi in LAA, and only three had them in the left atrial cavity. The clinical and echocardiographic characteristics of patients with and without LAA thrombi are shown in Table 1 . Patients with LAA thrombi were older and had greater prevalence of hypertensive diseases than those without LAA thrombi. Compared with patients without LAA thrombi, those with thrombi had larger left atrial dimension, lower ejection fraction, greater prevalence of moderate-to-severe SEC, and lower LAA velocities.

Multivariable Analysis

On the basis of the results of univariate analysis, the following variables were entered into the multivariate model: age, presence of hypertension, non-PAF, presence of congestive heart failure, a history of recent embolic events, and D-dimer levels. Independent predictors of LAA thrombi are summarized in Table 2 . Multivariable analysis revealed that non-PAF, presence of congestive heart failure, a history of recent embolic events, and D-dimer levels were the independent predictors of LAA thrombi. The most powerful predictor was plasma D-dimer levels.

Plasma D-Dimer Level

Plasma D-dimer levels were significantly higher in patients with LAA thrombi than in those without thrombi (median of 2.33 and interquartile range of 1.15-6.30 vs. median of 0.59 and interquartile range of 0.35-1.22, P < 0.0001). Distribution of D-dimer level is shown in Figure 1. ROC analysis yielded an optimal cutoff value of 1.15 g/mL for D-dimers to detect the LAA thrombi (Figure 2). LAA thrombi were detected in 21.8% of patients whose plasma D-dimer levels were not <1.15 g/mL (n = 289) (patients with higher D-dimer levels), whereas in only 3.1% of patients, plasma D-dimer levels were <1.15 g/mL (patients with lower D-dimer levels) (n = 636). Although the plasma D-dimer cutoff value of 1.15 g/mL had a sensitivity of 76% (95% CI 65-85), a specificity of 73% (95% CI 70-76), and a positive predictive value of 22% (95% CI 17-27) for identifying thrombi in LAA, it had a negative predictive value of 97% (95% CI 95-98). Patients with higher plasma D-dimer levels were older (73.2 9.4 years vs. 66.8 10.1 years, P < 0.05) and had a greater prevalence of hypertensive heart diseases (71 vs. 59%, P < 0.05) than those with lower plasma D-dimer levels. The average of the peak LAA velocities in patients with higher plasma D-dimer levels was significantly lower than that in those with lower plasma D-dimer levels (33.8 21.1 vs. 38.5 24.2 cm/s, P < 0.05). Increasing grades of left atrial SEC were associated with increasing D-dimer levels (Figure 3).

Figure 1. 

Distribution of D-dimer levels. Data of the same concentration are presented as just one data point (e.g. six patients had a D-dimer level of 0.1 g/mL).

     

Figure 2. 

Calculation of optimal cutoff value by receiver operating characteristic curve analysis with respect to D-dimers. Receiver operating characteristic analysis yielded an optimal cutoff value of 1.15 g/mL for D-dimers. Area under the receiver operating characteristic curve is 0.80 (95% CI 0.75-0.85).

     

Figure 3. 

D-dimer levels in non-valvular atrial fibrillation patients with increasing grades of severity of left atrial spontaneous echo contrast. Increasing grades of left atrial spontaneous echo contrast were associated with increasing D-dimer levels. *P < 0.05, 0 vs. 2+, 3+, and 4+; **P < 0.05, 1+ vs. 2+, 3+, and 4+; #P < 0.05, 4+ vs. 2+ and 3+.

     

Paroxysmal Atrial Fibrillation

Twelve (5%) out of 250 patients with PAF and 71 (11%) out of 675 patients without PAF had thrombi in LAA. The clinical and echocardiographic characteristics of patients with and without PAF are shown in Table 3 . Patients with PAF had smaller left atrial dimension, smaller left ventricular dimension in the end-diastole, higher ejection fraction, less prevalence of moderate-to-severe SEC, higher LAA velocities, and smaller LAA area than those without PAF. In patients with PAF, although the plasma D-dimer cutoff value of 1.15 g/mL had a sensitivity of 75% (95% CI 47-91), a specificity of 70% (95% CI 64-76), and a positive predictive value of 11% (95% CI 5-20) for identifying thrombi in LAA, it had a negative predictive value of 98% (95% CI 95-100). In patients without PAF, although the plasma D-dimer cutoff value of 1.15 g/mL had a sensitivity of 76% (95% CI 64-85%), a specificity of 74% (95% CI 71-78), and a positive predictive value of 26% (95% CI 20-32) for identifying thrombi in LAA, it had a negative predictive value of 96% (95% CI 94-98).

Patients' Complications and Left Atrial Appendage Thrombi

Among patients complicated with congestive heart failure (n = 208), 33 (15.9%) had LAA thrombi. Although the plasma D-dimer cutoff value of 1.15 g/mL had a specificity of 66% (95% CI 59-73) and a positive predictive value of 32% (95% CI 23-43) for identifying thrombi in LAA, it had a sensitivity of 85% (95% CI 68-95) and a negative predictive value of 96% (95% CI 91-98).

Among patients complicated with recent embolic events (n = 208), 42 (20.2%) had LAA thrombi. Patients with recent embolic events appeared to have higher D-dimer levels (median of 2.85 and interquartile range of 1.06-7.85 vs. median of 0.83 and interquartile range of 0.45-2.10, P < 0.0001). The plasma D-dimer cutoff value of 1.15 g/mL had a sensitivity of 91% (95% CI 77-97) and a negative predictive value of 97% (95% CI 91-99), although it had a specificity of 66% (95% CI 58-73) and a positive predictive value of 40% (95% CI 30-51) for identifying thrombi in LAA. In patients without recent embolic events (n = 717), the plasma D-dimer cutoff value of 1.15 g/mL had a sensitivity of 78% (95% CI 62-89), a specificity of 76% (95% CI 73-79), and a positive predictive value of 16% (95% CI 12-22) for identifying thrombi in LAA, and it had a negative predictive value of 98% (95% CI 97-99). Furthermore, 11 (34.4%) of 32 patients who were complicated with congestive heart failure and recent embolic events had LAA thrombi. Among patients with congestive heart failure or recent embolic events (n = 384), the plasma D-dimer cutoff value of 1.15 g/mL had a specificity of 67% (95% CI 62-72) and a positive predictive value of 35% (95% CI 28-43) for identifying thrombi in LAA, and it had a sensitivity of 89% (95% CI 79-95) and a negative predictive value of 97% (95% CI 94-99).

Anticoagulant Therapy

Seventy-three patients received an adequate anticoagulation therapy (international normalized ratios ≥ 2.0). LAA thrombi were detected in only two patients (2.7%) receiving the adequate anticoagulation therapy. These two patients had higher D-dimer levels (≥1.15 g/mL).

Discussion

D-Dimer Levels and Left Atrial Appendage Thrombi

This study demonstrated the ability of D-dimer to predict LAA thrombi in NVAF. D-dimer has gained widespread clinical use as a parameter for the detection of fibrin formation. Previous studies demonstrate that patients with atrial fibrillation show significant increases in plasma fibrinogen and D-dimer level,[13,14] suggesting the presence of a hypercoagulable or pro-thrombotic state. Our data clearly showed that, in patients with atrial fibrillation, the use of a single laboratory test could exclude the possibility of LAA thrombi formation. The negative predictive value of D-dimer measurement with a cutoff of 1.15 g/mL in our study corresponded to the 98% rate reported by Somloi et al.[8] However, D-dimer concentration of not less than 1.15 g/mL could not be used for positive diagnosis because the false-positive rate was 27% (positive predictive value of 22%). Several disease conditions such as infectious or inflammatory diseases and cancer were associated with fibrin formation and degradation. In a population with D-dimer concentration of not less than 1.15 g/mL, patients with these conditions could potentially be included. Although false-positive results (27%) and false-negative results (24%) were relatively often observed, we thought it was important for patients with atrial fibrillation to do risk stratification. We should perform the immediate and adequate anticoagulant therapy and careful observation for patients with LAA thrombi to prevent serious sequela. From this point of view, we thought our results are useful for risk stratification. We believed our results might lead to prompt detection of LAA thrombi and prevention of serious sequela.

Most thrombi in NVAF originate in LAA,[15] suggesting the origin of thrombo-embolism.[16] A case of a disappearing LAA thrombus resulting in a stroke is reported.[17] Therefore, we thought it was important to detect LAA thrombi before they were embolized. Among the patients other than those with recent embolic events, 41 (5.7%) had LAA thrombi in our study. The immediate treatments with adequate anticoagulant therapy and preventive percutaneous or surgical intervention should be performed, if needed. Several randomized, prospective trials involving patients with NVAF have confirmed a significant reduction in the risk of stroke with warfarin.[18-20] These trials have provided a basis of important guidelines for the use of warfarin in such patients.[9,21] A high D-dimer level that is associated with an increased risk of stroke in patients with atrial fibrillation may also warrant the institution of anticoagulant therapy. Three recent randomized studies[22-24] suggest that the strategy used to treat atrial fibrillation (rate control vs. rhythm control) does not have a substantial effect on the quality of life or cardiovascular endpoints. However, in highly symptomatic patients whose heart rate cannot be controlled, rhythm control may still be preferable. When cardioversion was being considered, determination of absence of LAA thrombi by D-dimer might be utilized.

Optimal Cutoff Level

In this study, we identified the 'optimal' cutoff by using ROC analysis. The best cutoff has the highest sensitivity and lowest 1-specificity and is therefore located as high up on the vertical axis and as far left on the horizontal axis as possible (upper left corner). When investigating whether D-dimer can be used to eliminate the possibility of LAA thrombi formation in patients with NVAF, a decision-maker should choose a cutoff criterion from the higher right-hand portion of the ROC curve to minimize false negatives. We compared a cutoff level of 1.15 g/mL (closest to the left upper corner of the ROC curve) with that of 0.80 g/mL (higher right-hand portion of the ROC curve). At a cutoff level of 0.80 g/mL, the sensitivity was 81.9% (95% CI 72-90), specificity 62.1% (95% CI 59-65), positive predictive value 17.6% (95% CI 14-22), and negative predictive value 97.2% (95% CI 95-98%) for identifying thrombi in LAA. At a cutoff level of 1.15 g/mL, the sensitivity was 75.9% (95% CI 65-85), specificity 73.1% (95% CI 70-76%), positive predictive value 21.7% (95% CI 17-27%), and negative predictive value 97.0% (95% CI 95-98) for identifying thrombi in LAA. We established the 'optimal' cutoff level as 1.15 g/mL because a negative predictive value was almost the same between cutoff levels of 1.15 and 0.80 g/mL.

We evaluated the possibility of LAA thrombi formation by dichotomized D-dimer levels after confirming the ability of D-dimer to predict LAA thrombi in patients with NVAF. In clinical settings, this categorization has been useful for risk stratification of patients with atrial fibrillation. However, there is uncertainty about the 'optimal' cutoff level in patient populations other than those with atrial fibrillation because the cutoff levels are known to be very susceptible to changes in individual populations.

D-Dimer Levels and Left Atrial Spontaneous Echo Contrast

Left atrial SEC may provide visual demonstration of blood stasis within the left atrial cavity. A high prevalence of left atrial SEC has been found in patients with LAA thrombi.[25] The presence of SEC in the left atrium has been proposed as a marker of increased thrombo-embolic risk. Our study elucidated an increase in D-dimer levels with grades of left atrial SEC. A more or less compensated hypercoagulable state exists in many patients with atrial fibrillation. Measurement of D-dimer level was a means for the detection of LAA thrombi and the hypercoagulable state in our population. However, there are few prospective data that describe the degree of predictability provided by hypercoagulability for the risk of having thrombo-embolic diseases.

Congestive Heart Failure and Recent Embolic Events

Recent congestive heart failure, a history of hypertension, and previous arterial thrombo-embolism are significantly and independently associated with a substantial risk of thrombo-embolism.[26] Multivariable analysis of selected clinical variables showed that non-PAF, presence of congestive heart failure, a history of recent embolic events, and D-dimer levels were associated with a significantly increased risk of LAA thrombi that might reflect the incidence of thrombo-embolic events. In our study, by the measurement of D-dimer levels, LAA thrombi could be sensitively detected in patients with NVAF complicated with congestive heart failure and/or recent embolic events.

Limitations

First, in this study, D-dimer was measured by LPIA, and data and conclusions could be drawn for this specific D-dimer assay only. D-dimer assay included LPIA, manual latex agglutination assays, sandwich-type enzyme-linked immunoassays (ELISA), automated ELISA systems, and membrane-based immunofiltration assay systems. According to the results of a FACT study,[27] however, D-dimer LPIA displays a response similar to that in most of the other assays. Secondly, antithrombotic therapies were not randomized. The majority of patients were not anticoagulated. Thirdly, our results of patients complicated with a history of recent embolic events were likely to underestimate the number of patients with LAA thrombi because LAA thrombi had already been embolized when we performed TEE.

Conclusions

In a selected cohort of patients with NVAF, D-dimer may be helpful for predicting the absence of LAA thrombi owing to its high negative predictive value. D-dimer level was clinically useful to guide the management of patients with NVAF, especially for those complicated with congestive heart failure and/or recent embolic events.


Table 1. Baseline Characteristics Among Patients With and Without Left Atrial Appendage Thrombi


  LAA thrombi (-) (n = 842) LAA thrombi (+) (n = 83) P-value
Age 68.5 10.3 72.0 9.9 <0.01
Male (%) 68 61 0.20
Diabetes mellitus (%) 20 20 0.93
Hypertension (%) 62 73 0.04
Hyperlipidaemia (%) 33 36 0.83
LAD (mm) 43.5 6.9 46.6 9.3 0.0006
LVDd (mm) 47.7 10.6 49.2 7.6 0.22
Ejection fraction (%) 60.9 12.3 53.9 16.1 0.0001
LAA area (cm2) 6.3 2.9 7.0 2.6 0.082
LAA velocity (cm/s) 38.5 23.6 21.2 12.4 <0.0001
≥Moderate SEC in LA (%) 50 83 <0.0001
≥Moderate SEC in Ao (%) 41 61 0.008
Ao atheroma ≥3 mm (%) 21 25 0.46
Recent embolic events (%) 20 51 <0.0001
Heart failure (%) 27 48 0.0004
Adequate anticoagulation therapy (%) 8 2 <0.0001

LAD, left atrial dimension; LVDd, left ventricular dimension in end-diastole; LAA, left atrial appendage; SEC, spontaneous echo contrast; LA, left atrial; Ao, aorta.


    Table 2. Results of the Multivariable Analysis: Independent Predictors of Left Atrial Thrombi


    Variables OR 95% CI P-value
    Non-paroxysmal atrial fibrillation 2.13 1.15-4.19 0.02
    Presence of congestive heart failure 3.10 1.77-5.50 <0.0001
    A history of recent embolic events 3.39 1.90-6.04 <0.0001
    D-dimer levels 97.6 17.3-595.8 <0.0001


      Table 3. Baseline Characteristics in Patients With and Without Paroxysmal Atrial Fibrillation


        With PAF (n = 250) Without PAF (n = 675) P-value
      Age 69.2 9.8 68.6 10.5 0.44
      Male (%) 66 68 0.41
      Diabetes mellitus (%) 22 19 0.43
      Hypertension (%) 64 62 0.73
      Hyperlipidaemia (%) 29 34 0.45
      LAD (mm) 40.0 6.2 45.2 6.8 <0.0001
      LVDd (mm) 46.1 6.3 47.6 7.4 0.0067
      Ejection fraction (%) 64.7 10.6 58.6 13.2 <0.0001
      LAA area (cm2) 5.4 1.8 6.7 3.2 <0.0001
      LAA velocity (cm/s) 53.1 25.5 31.2 19.5 <0.0001
      ≥ Moderate SEC in LA (%) 25 63 <0.0001
      ≥ Moderate SEC in Ao (%) 31 46 0.0006
      Ao atheroma ≥ 3 mm (%) 20 21 0.61
      Recent embolic events (%) 24 22 0.18
      Heart failure (%) 18 33 <0.0001
      LAA thrombi (%) 5 11 0.004

      PAF, paroxysmal atrial fibrillation; LAD, left atrial dimension; LVDd, left ventricular dimension in end-diastole; LAA, left atrial appendage; SEC, spontaneous echo contrast; LA, left atrial; Ao, aorta.




        References

        1. Go AS, Hylek EM, Phillips KA, Chang Y, Henault LE, Selby JV, Singer DE. Prevalence of diagnosed atrial fibrillation in adults: national implications for rhythm management and stroke prevention: the AnTicoagulation and Risk Factors in Atrial Fibrillation (ATRIA) Study. JAMA (2001) 285:2370-2375.
        2. Kannel WB, Wolf PA, Benjamin EJ, Levy D. Prevalence, incidence, prognosis, and predisposing conditions for atrial fibrillation: population-based estimates. Am J Cardiol (1998) 82:2N-9N.
        3. Manning WJ, Weintraub RM, Waksmonski CA, Haering JM, Rooney PS, Maslow AD, Johnson RG, Douglas PS. Accuracy of transesophageal echocardiography for identifying left atrial thrombi: a prospective, intraoperative study. Ann Intern Med (1995) 123:817-822.
        4. Fatkin D, Scalia G, Jacobs N, Burstow D, Leung D, Walsh W, Feneley M. Accuracy of biplane transesophageal echocardiography in detecting left atrial thrombus. Am J Cardiol (1996) 77:321-323.
        5. Palareti G, Legnani C, Cosmi B, Valdre L, Lunghi B, Bernardi F, Coccheri S. Predictive value of D-dimer test for recurrent venous thromboembolism after anticoagulation withdrawal in subjects with a previous idiopathic event and in carriers of congenital thrombophilia. Circulation (2003) 108:313-318.
        6. Eggebrecht H, Naber CK, Bruch C, Kroger K, von Birgelen C, Schmermund A, Wichert M, Bartel T, Mann K, Erbel R. Value of plasma fibrin D-dimers for detection of acute aortic dissection. J Am Coll Cardiol (2004) 44:804-809.
        7. Rathbun SW, Whitsett TL, Vesely SK, Raskob GE. Clinical utility of D-dimer in patients with suspected pulmonary embolism and nondiagnostic lung scans or negative CT findings. Chest (2004) 125:851-855.
        8. Somloi M, Tomcsanyi J, Nagy E, Bodo I, Bezzegh A. D-dimer determination as a screening tool to exclude atrial thrombi in atrial fibrillation. Am J Cardiol (2003) 92:85-87.
        9. Fuster V, Ryden LE, Asinger RW, Cannom DS, Crijns HJ, Frye RL, Halperin JL, Kay GN, Klein WW, Levy S, McNamara RL, Prystowsky EN, Wann LS, Wyse DG, American College of Cardiology. ACC/AHA/ESC guidelines for the management of patients with atrial fibrillation. A report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines and the European Society of Cardiology Committee for Practice Guidelines and Policy Conferences (Committee to develop guidelines for the management of patients with atrial fibrillation) developed in collaboration with the North American Society of Pacing and Electrophysiology. Eur Heart J (2001) 22:1852-1923. American Heart Association; European Society of Cardiology; North American Society of Pacing and Electrophysiology.
        10. Wells PS, Anderson DR, Bormanis J, Guy F, Mitchell M, Gray L, Clement C, Robinson KS, Lewandowski B. Value of assessment of pretest probability of deep-vein thrombosis in clinical management. Lancet (1997) 350:1795-1798.
        11. Fatkin D, Loupas T, Jacobs N, Feneley MP. Quantification of blood echogenicity: evaluation of a semiquantitative method of grading spontaneous echo contrast. Ultrasound Med Biol (1995) 21:1191-1198.
        12. Catherine M, Otto MD. The Practice of Clinical Echocardiography (2002) 2nd ed. Philadelphia, PA: WB Saunders Co.
        13. Lip GY, Lowe GD, Rumley A, Dunn FG. Increased markers of thrombogenesis in chronic atrial fibrillation: effects of warfarin treatment. Br Heart J (1995) 73:527-533.
        14. Lip GY, Lowe GD, Rumley A, Dunn FG. Fibrinogen and fibrin D-dimer levels in paroxysmal atrial fibrillation: evidence for intermediate elevated levels of intravascular thrombogenesis. Am Heart J (1996) 131:724-730.
        15. Blackshear JL, Odell JA. Appendage obliteration to reduce stroke in cardiac surgical patients with atrial fibrillation. Ann Thorac Surg (1996) 61:755-759.
        16. Hart RG, Halperin JL. Atrial fibrillation and stroke. Revisiting the dilemmas. Stroke (1994) 25:1337-1341.
        17. Parekh A, Jaladi R, Sharma S, Van Decker WA, Ezekowitz MD. The case of a disappearing left atrial appendage thrombus: direct visualization of left atrial thrombus migration, captured by echocardiography, in a patient with atrial fibrillation, resulting in a stroke. Circulation (2006) 114:e513-e514.
        18. Ezekowitz MD, Bridgers SL, James KE, Carliner NH, Colling CL, Gornick CC, Krause-Steinrauf H, Kurtzke JF, Nazarian SM, Radford MJ, Rickles FR, Shabetai R, Deykin D. Warfarin in the prevention of stroke associated with nonrheumatic atrial fibrillation. N Engl J Med (1992) 327:1406-1412.
        19. Atrial Fibrillation Investigators. Risk factors for stroke and efficacy of antithrombotic therapy in atrial fibrillation: analysis of pooled data from five randomized controlled trials. Arch Intern Med (1994) 154:1449-1457.
        20. Hart RG, Benavente O, McBride R, Pearce LA. Antithrombotic therapy to prevent stroke in patients with atrial fibrillation: a meta-analysis. Ann Intern Med (1999) 131:492-501.
        21. Albers GW, Dalen JE, Laupacis A, Manning WJ, Petersen P, Singer DE. Antithrombotic therapy in atrial fibrillation. Chest (2001) 119((suppl.)):194S-206S.
        22. Hohnloser SH, Kuck K-H, Lilienthal J. Rhythm or rate control in atrial fibrillationPharmacological Intervention in Atrial Fibrillation (PIAF): a randomised trial. Lancet (2000) 356:1789-1794.
        23. Van Gelder IC, Hagens VE, Bosker HA, Kingma JH, Kamp O, Kingma T, Said SA, Darmanata JI, Timmermans AJ, Tijssen JG, Crijns HJ, Rate Control versus Electrical Cardioversion for Persistent Atrial Fibrillation Study Group. A comparison of rate control and rhythm control in patients with recurrent persistent atrial fibrillation. N Engl J Med (2002) 347:1834-1840.
        24. Wyse DG, Waldo AL, DiMarco JP, Domanski MJ, Rosenberg Y, Schron EB, Kellen JC, Greene HL, Mickel MC, Dalquist JE, Corley SD, Atrial Fibrillation Follow-up Investigation of Rhythm Management (AFFIRM) Investigators. A comparison of rate control and rhythm control in patients with atrial fibrillation. N Engl J Med (2002) 347:1825-1833.
        25. Fatkin D, Kelly RP, Feneley MP. Relations between left atrial appendage blood flow velocity, spontaneous echocardiographic contrast and thromboembolic risk in vivo. J Am Coll Cardiol (1994) 23:961-969.
        26. The Stroke Prevention in Atrial Fibrillation Investigators. Predictors of thromboembolism in atrial fibrillation. I. Clinical features of patients at risk. Ann Intern Med (1992) 116:1-5.
        27. Dempfle CE, Zips S, Ergul H, Heene DL, FACT study group. The fibrin assay comparison trial (FACT): correlation of soluble fibrin assays with D-dimer. Thromb Haemost (2001) 86:1204-1209.
        Reprint Address

        Seiji Habara Department of Cardiology, Kurashiki Central Hospital, 1-1-1 Miwa, kurashiki-shi, Okayama 710-8602, Japan. Tel: + 81 86 422 0210; fax: + 81 86 421 3424. E-mail: sh10461@kchnet.or.jp


        Seiji Habara, Keigo Dote, Masaya Kato, Shota Sasaki, Kenji Goto, Hiroaki Takemoto, Daiji Hasegawa, Osamu Matsuda
        Department of Cardiology, Hiroshima City Asa Hospital, Hiroshima, Japan