Does Left Atrial Appendage Morphology Influence Left Atrial Appendage Flow Velocity?

Abstract

Background: The shape of the left atrial appendage (LAA) might affect thrombus formation. The chicken wing-type LAA (CW) has been reported as unlikely to influence stroke events in atrial fibrillation (AF) patients, so we investigated whether LAA shapes could influence LAA function.

Methods and Results: We studied 102 patients (64 men, age 65±9 years) who underwent transthoracic echocardiography, transesophageal echocardiography (TEE), and cardiac computed tomography prior to catheter ablation (CA) for AF. LAA morphology were classified into 2 types: (1) CW: LAA with a bend in its shape and (2) non-CW type (NCW): LAA without any bends. All patients were classified into these groups using a cutoff value of LAA flow velocity (LAAFV). Patients with LAAFV <35 cm/s were classified as the low LAAFV group (Low FV, n=37). The patients with LAAFV >35 cm/s were classified as normal LAAFV group (Normal FV, n=65). The NCW type was detected in 25/102 patients (25%). In multivariate analysis, the patients with Low FV were associated with NCW type (P=0.0429, odds ratio [OR] 9.664, 95% confidence interval [CI] 1.075–86.900) and higher B-type natriuretic peptide (BNP) (P=0.0350, OR 1.012 for each 1 pg/ml increase in BNP, 95% CI 1.001–1.022).

Conclusions: The NCW-type LAA and higher BNP were associated with lower LAAFV. One reason for the frequent cardiogenic stroke in patients with the NCW-type LAA may be the lower LAAFV. (Circ J 2015; 79: 1706–1711)

Atrial fibrillation (AF)-associated strokes are devastating and a major contributor to healthcare costs. A previous study has shown that AF is a predictive factor for severe stroke and early death.¹ AF is also associated with a 5-fold increase in stroke risk.² In a Japanese clinical study, the incidence of ischemic stroke events in 20,000 patients with AF was reported by Tomita et al as 4.6% during a 1.7-year follow-up.³

The left atrial appendage (LAA) is known to be a major source of cardiac thrombus in patients with AF.^4,5 A previous study has shown that a mobile, ball-type thrombus is an important risk factor for cerebral/arterial embolism⁶ and a recent study reported that transcatheter LAA closure is safe and feasible.⁷ In other study, reverse remodeling of the LAA after catheter ablation (CA) was reported,⁸ so LAA morphology has received a lot of attention in recent years. Recently, some studies reported that 29–50% in AF patients have a “chicken wing (CW)” type LAA, but that this shape is less likely to influence stroke events in patients with AF.^9,10 However, it remains uncertain why the incidence of stroke/transient ischemic attack (TIA) is higher in patients with the non-CW (NCW) type LAA than in patients with a CW LAA. We hypothesized that the shape of the LAA is related to thrombus formation via a decrease in LAA flow velocity (FV), and in the present study we investigated the influence of LAA shape on LAA function.

Methods

Patient Population

This study included consecutive 107 patients with nonvalvular AF who underwent transthoracic echocardiography (TTE), transesophageal echocardiography (TEE) and cardiac computed tomography (CCT) prior to CA at Hyogo College of Medicine between October 2012 and November 2014. In order to elicit a rhythm effect, patients whose rhythms at TTE, TEE, and CCT were not the same were excluded. If a patient underwent multiple TTE, TEE or CCT, we used the initial examination results as the index study. All patients were classified into 2 groups using a cutoff value (35 cm/s) of LAAFV:¹¹ patients with LAAFV <35 cm/s were classified as the low LAAFV group (Low FV, n=37) and those with LAAFV >35 cm/s were the normal LAAFV group (Normal FV, n=65). All patients gave written informed consent before CA and had to be >18 years old. The research protocol was approved by the locally appointed ethics committee.

A paroxysmal arrhythmia was defined as recurrent AF that terminated spontaneously within 7 days, and a persistent arrhythmia was defined as recurrent AF sustained beyond 7 days, or lasting <7 days but requiring pharmacological or electrical cardioversion.¹²

The CHADS₂ scoring scheme is a validated method for estimating the risk of stroke in patients with AF,¹³ so we calculated it for all patients at the time of TEE.

LAA Morphology and CCT

All patients underwent ECG-gated 64-slice multidetector CT (SOMATOM Definition Edge, Siemens, Erlangen, Germany) within 1 month prior to CA. The slice acquisition thickness was 0.75 mm. An independent workstation (Ziostation version 2.1, Ziosoft, Tokyo, Japan) was used for analysis. The 3D structures of the left atrium (LA) and LAA were constructed using the volume-rendered post-processing technique. In all cases, 40–70 ml of intravenous contrast agent (Imeron 350, Bracco Imaging, Konstanz, Germany) was injected at 5 ml/s and then followed by a 30-ml saline bolus at 5 ml/s. CT images of the LA, LAA and pulmonary vein (PV) were reconstructed at a separate workstation and integrated using Ensite Verismo (St. Jude Medical, Minneapolis, MN, USA) before CA. LAA volume was also obtained using EnSite Verismo.

In our study, LAA shape was classified into 2 types using CCT. The CW was defined as LAA with an obvious bend in the proximal or middle part of the dominant lobe, or the LAA folding back on itself at some distance from the perceived LAA orifice (Figure A). The NCW was defined as the LAA without any bends (Figure B). In the previous study, the NCW included unique LAA shapes such as the “cauliflower”, “cactus” and “wind sock”.⁶ Because we considered that those shapes had defined subjectively, we classified LAA morphology into only 2 types: CW and NCW. The LAA morphology was classified by 2 cardiologists who were blinded to the clinical data. The patients with low-quality CCT scans and/or TEE/TTE images were excluded (n=2).

Figure.

Typical examples of left atrial appendage morphology are shown: (A) chicken wing type and the different shapes of the non-chicken wing type: (B-a) cactus, (B-b) cauliflower and (B-c) wind sock.

Echocardiographic Imaging

All patients also underwent TTE within 1 month before the CA procedure using a Prosound F75 (Hitachi Aloka Medical, Tokyo, Japan) with a 3.88-MHz transducer probe. We assessed the left atrial diameter (LAD), LA volume index (LAVI), E wave, e’ wave, E/e’ ratio, deceleration time, left ventricular (LV) dimension during end diastole, LV hypertrophy (LVH), LV ejection fraction (LVEF), heart rate and rhythm (sinus rhythm [SR] or AF rhythm) during TTE. The LAV at the time just before the mitral valve opening on the apical 4- and 2-chamber views was determined off-line using Simpson’s biplane disc summation method.¹³ LAV was indexed for body size by dividing by body surface area calculated with the DuBois formula.¹⁴ LVH was defined as a posterior wall or interventricular septal thickness >12 mm. If less than 12 mm, the LV mass index was calculated. If that was >95 (female) or 115 (male), the patient was considered to have LVH. This definition of LVH was based on the American Society of Echocardiography’s guideline.¹⁵

All patients underwent TEE within 48 h of the CA procedure using an iE 33 (Philips Medical System, Andover, MA, USA) with a 4-MHz omniplane probe. After local pharyngeal anesthesia with lidocaine spray, the patient was placed in the left lateral position, and the transesophageal transducer was inserted in the esophagus. We assessed the LAAFV, the LAA orifice area, and we looked for LA thrombus (LAT) and spontaneous echocardiography contrast (SEC). The LAAFV was obtained at the outlet of the LAA by the pulsed Doppler method. The sample volume was placed within the LAA cavity >1 cm away from the LA cavity. The peak emptying velocities were analyzed and averaged for 5 cardiac cycles. The LAA orifice diameter was measured by 2D-imaging on a 45-degree plane. The LAA orifice area was measured by 2D imaging, which was reconstructed according to 3D imaging, by planimetry. An LAT was defined as being present if there was a mass that could be distinguished from the surrounding endocardium or pectinate muscles.¹⁶ SEC was defined as a pattern of dynamic “smokelike”, slowly swirling, intracavitary echo-densities imaged with gain settings adjusted to eliminate background noise.¹⁶ Patients with LAT were excluded from the study (n=3). No complications from any of the TEE procedures were reported.

Statistical Analysis

All data are expressed as the mean value±standard deviation. Statistical comparisons were made by Student’s t-test. Differences between categorical variables were evaluated by chi-square analysis. A P value <0.05 was considered statistically significant. The univariate analyses were performed using Student’s t-test or chi-square analysis. The variables that were found to be significant in the univariate analysis were entered into a multivariate analysis. The independent association with the Low FV group was evaluated using a forward stepwise logistic regression analysis. All analyses were performed with StatView version 5.0 software (SAS Institute, Cary, NC, USA).

Results

Patient Population

The baseline characteristics of the patients in our study are described in Table 1. Of the 107 patients who initially enrolled, 2 were excluded because of low-quality CT images and 3 because of LAT. The remaining 102 patients were registered in this study. Low FV was detected in 37/102 patients (36.3%); 11 of them (29.7%) had structural heart disease, which included 9 with non-ischemic cardiomyopathy and 8 with ischemic heart disease. Compared with the patients in the Normal FV group, there was a higher rate of prior congestive heart failure (CHF: 27% vs. 3%, P=0.0006) and persistent arrhythmia (59% vs. 28%, P=0.0016) in the patients with low FV. The mean CHADS₂ score did not differ between the Normal and Low FV groups.

Table 1. Baseline Characteristics of the Patients Undergoing Catheter Ablation of AF

	Normal FV (n=65)	Low FV (n=37)	P value
Male sex, n (%)	39 (60)	25 (68)	0.4472
Age (years)*	64±10	66±7	0.1405
Height (m)*	1.6±0.1	1.6±0.1	0.7971
Weight (kg)*	63±13	63±11	0.9936
BMI (kg/m2)*	23±3	24±3	0.7284
HT, n (%)	38 (58)	23 (62)	0.7140
Dlp, n (%)	22 (34)	12 (32)	0.8842
DM, n (%)	9 (14)	10 (27)	0.1002
SHD, n (%)	13 (20)	11 (30)	0.2654
Prior CHF, n (%)	2 (3)	10 (27)	0.0006
Prior stroke/TIA, n (%)	8 (12)	3 (8)	0.5109
Vascular disease, n (%)	25 (38)	18 (49)	0.3165
CHADS2 score*	1.1±1.0	1.4±1.0	0.1094
Persistent arrhythmia, n (%)	18 (28)	22 (59)	0.0016
PT-INR*	1.4±0.6	1.5±0.5	0.5084
APTT (s)*	38±19	37±8	0.8658
CRP (mg/dl)*	0.3±0.7	0.1±0.2	0.3015
BNP (pg/ml)*	72±78	149±170	0.0021
HANP (pg/ml)*	63±137	86±73	0.3352
Creatinine (mg/dl)*	0.8±1.2	1.2±1.8	0.2529
HbA1c (NGSP) (%)*	6.0±0.6	6.0±0.6	0.7458

*Mean±SD. AF, atrial fibrillation; APTT, activated partial thromboplastin time; BMI, body mass index; BNP, B-type natriuretic peptide; CHF, congestive heart failure; CRP, C-reactive protein; Dlp, dyslipidemia; DM, diabetes mellitus; FV, flow velocity; HANP, human atrial natriuretic peptide; HT, hypertension; NGSP, National Glycohemoglobin Standardization Program; PT-INR, prothrombin time-international normalized ratio; SHD, structural heart disease; TIA, transient ischemic attack.

The blood samples of the patients are described in Table 1. Compared with the patients with normal FV, patients with low FV showed higher concentrations of B-type natriuretic peptide (BNP) (149±170 pg/ml vs. 72±78 pg/ml, P=0.0021). The prothrombin time-international normalized ratio, activated partial thromboplastin time, C-reactive protein, human atrial natriuretic peptide, serum creatinine, and HbA1c (national glycohemoglobin standardization program) values did not differ between the Normal and Low FV groups.

Medication status is described in Table 2. It did not differ between the 2 groups; 26 patients received Warfarin and 68 received a non-vitamin K antagonist oral anticoagulant (dabigatran: 26 patients, rivaroxaban: 35 patients, apixaban: 7 patients).

Table 2. Medication Profile of the Patients Undergoing Catheter Ablation of AF

	Normal FV (n=65)	Low FV (n=37)	P value
β-blocker, n (%)	31 (48)	24 (65)	0.0944
ACEI/ARB, n (%)	27 (42)	16 (43)	0.7383
CCB, n (%)	27 (42)	12 (32)	0.3629
Statin, n (%)	14 (22)	10 (27)	0.5298
Anticoagulant, n (%)	60 (92)	34 (92)	0.9401
NOAC/warfarin, n (%)	46/14 (71/22)	22/12 (6/32)	0.2371
Antiplatelet drug, n (%)	8 (12)	3 (8)	0.7419
AAD, n (%)	40 (62)	25 (68)	0.5426

*Mean±SD. AAD, antiarrhythmic drug; ACEI, angiotensin-converting enzyme inhibitor; ARB, angiotensin II receptor blocker; CCB, calcium-channel blocker; NOAC, non-vitamin K antagonist oral anticoagulant. Other abbreviations as in Table 1.

CCT and Echocardiographic Characteristics

The echocardiographic characteristics are described in Table 3. Of the 102 patients who were registered in the study, the NCW LAA was detected in 25 (24.5%) by CCT. The NCW morphology included the following distribution of the 3 specific types: 6 patients (5.9%) had the cactus shape, 7 (6.9%) had the wind sock, and 12 (11.8%) had the cauliflower. Compared with the patients in the Normal FV group, patients with low FV showed lower LAAFV, lower LVEF, larger LAD, larger LAVI, higher E wave, higher E/e’ ratio, and larger LAA orifice area. Compared with the patients with normal FV, there was a higher rate of NCWL, AF rhythm during TEE, and prevalence of SEC in the patients with low FV.

Table 3. Echocardiographic and Computed Tomography Characteristics of the Patients Undergoing Catheter Ablation of AF

	Normal FV (n=65)	Low FV (n=37)	P value
LAD (mm)*	39.6±5.9	43.3±6.4	0.0040
LAVI (ml/m2)*	36.2±12.1	46.3±14.3	0.0003
E wave (cm/s)*	69.2±17.3	91.6±33.8	<0.0001
e’ wave (cm/s)*	7.4±2.4	9.7±11.2	0.1170
E/e’ ratio*	10.1±3.5	12.8±6.2	0.0062
DcT (ms)*	197±47	194±85	0.8086
LVDd (mm)*	47.8±4.1	48.6±5.1	0.3888
LVH, n (%)	7 (11)	6 (16)	0.4277
LVEF (%)*	67.1±9.5	61.7±12.9	0.0174
AF rhythm during TEE, n (%)	20 (31)	23 (62)	0.0020
HR during TEE (beats/min)*	73±20	78±17	0.1945
LAAFV (cm/s)*	53.0±13.2	23.9±6.9	<0.0001
SEC, n (%)	3 (5)	6 (16)	0.0470
LAA orifice area (cm2)*	3.9±1.9	5.3±1.4	0.0023
NCW, n (%)	6 (9)	19 (51)	<0.0001
LAA volume (ml)*	11.0±4.9	11.2±5.6	0.6838

*Mean±SD. DcT, decelaration time; HR, heart rate; LAA, left atrial appendage; LAD, left atrial diameter; LAVI, left atrial volume index; LVDd, left ventricular diameter during end diastole; LVEF, left ventricular ejection fraction; LVH, left ventricular hypertrophy; NCW, non-chicken wing type LAA; SEC, spontaneous echocardiography contrast; SR, sinus rhythm; TEE, transesophageal echocardiography. Other abbreviations as in Table 1.

Univariate and Multivariate Analyses

Univariate and multivariate analyses are described in Table 4. Patients with low FV were significantly associated with lower LVEF, larger LAD, larger LAVI, higher E wave, higher E/e’, larger LAA orifice area, higher BNP, prior CHF, persistent arrhythmia, AF rhythm during TEE, SEC, and NCW LAA. In the multivariate analysis, the patients with low FV were associated with NCW LAA (P=0.0429, OR 9.664, 95% CI 1.075–86.900) and higher BNP (P=0.0350, OR 1.012 for each 1 pg/ml increase in BNP, 95% CI 1.001–1.022).

Table 4. Analysis of Predictors of Low FV in the LAA of Patients With AF

	Univariate	Multivariate
	P value	P value	OR (95% CI)
Prior CHF	0.0006	0.1345	10.924 (0.477–250.338)
Persistent arrhythmia	0.0016	0.8806	0.862 (0.125–5.969)
BNP	0.0021	0.0350	1.012 (1.001–1.022)
LAD	0.0040	0.1486	0.870 (0.721–1.051)
LAVI	0.0003	0.5780	1.024 (0.942–1.114)
E wave	<0.0001	0.1343	1.053 (0.984–1.126)
E/e’ ratio	0.0062	0.3621	0.840 (0.577–1.222)
LVEF	0.0174	0.9329	1.004 (0.916–1.101)
AF rhythm during TEE	0.0020	0.6337	1.673 (0.202–13.885)
SEC	0.0470	0.4619	0.278 (0.009–8.397)
LAA orifice area	0.0023	0.1252	1.418 (0.907–2.216)
NCW shape	<0.0001	0.0429	9.664 (1.075–86.900)

CI, confidence interval. Other abbreviations as in Tables 1,3.

Discussion

Main Findings

The object of this study was to investigate the influence of LAA shape on LAA function. In our multivariate analysis, the NCW type of LAA and higher BNP concentration were independent factors related to low FV in. To the best of our knowledge, this is the first study to show that the shape of the LAA affects LAA function. The low FV could be what causes patients with the NCW type of LAA to have a higher incidence of stroke/TIA.

Influence of LAA Shape on LAA FV

The mechanisms of stroke/TIA in patients with the CW and NCW LAA shapes are still unclear. Although lower LAAFV is most likely to potentiate thrombus formation in the case of cardiogenic stroke in patients with the NCW LAA, other possible factors should be considered. The first is the complex structure of the NCW LAA. The LAA has a long, tubular and hooked structure with deferent lobes. The number of lobes is variable in each patients. Therefore, it is difficult to assess the number of lobes in the CW type and NCW type. The next is the position of the LAA position. In a previous study, the LAA position was classified into 3 groups (low, middle, and high) according to comparison of the superior aspect of the LAA orifice with that of the left superior PV.¹⁷ It was reported that s high takeoff position of the LAA had a lower rate of stroke, but there is limited evidence about the effect of the LAA’s position and further investigation is necessary.

Compared with the patients with normal FV, the prevalence of NCW LAA was higher in patients with low FV in our study. In a previous study, the NCW type showed a tendency to decrease LAAFV, but there was no significant difference.¹⁷ Various factors have been shown to influence LAAFV. For instance, LV diastolic filling pressure (E/e’),¹⁸ LA pressure,¹⁹ and LA size,²⁰ as well as ventricular rate,²¹ and regularity.²¹ However, whether these factors associate with LAAFV is controversial.

We assessed the following mechanism of lower FV in patients with the NCW than the CW type of LAA. First, the inferomedial portion of the NCW LAA may be externally compressed between the LA free wall and the pericardium during ventricular diastole because the CW has a longer lobe and bend. Second, patients with CW LAA may have greater muscle mass to contract the LAA. Ito et al reported that LAA function changes mainly by alterations in its intrinsic contractile property based on loading conditions.²² In the other words, it may be possible that dysfunction is related to intrinsic conditions, rather than being an indirect result of changes in LA pressure and/or volume overload. Third, high LA pressure or low LA compliance might change the shape of the LAA. For instance, it is possible to change from CW to wind sock or from cactus to cauliflower. However, these hypotheses are uncertain and further investigation is necessary.

This study, to the best of our knowledge, revealed for the first time the relationship between LAA shape and LAAFV. However, there was no difference in the LAAFV between the CW and NCW types of LAA in previous studies.^9,17 The differences in findings may be explained by the following. First, differences in the patients’ characteristics. For instance, the mean LAAFV in the other studies^9,17 was higher than that in our study, so it is possible that the other studies did not include patients with slower LAAFV. Second, the method of measuring LAAFV in our study may be different from that used in the other studies. We placed the sample volume within the LAA cavity >1 cm away from the LA cavity. However, the definition of LAA orifice was ambiguous in most studies. One study reported disparities in FV within the LAA.²³

Previous study showed that AF rhythms decrease the LAAFV.²¹ In our study, we performed statistical analysis with different rhythms (AF rhythm or SR) at the time of TEE, but the results of LAAFV did not change (AF at TEE, CW type: n=26, 43.8 cm/s vs. Non-CW type: n=17, 23.6 cm/s, P<0.0001; SR at TEE, CW type: n=51, 49.5 cm/s vs. Non-CW type: n=8, 32.7 cm/s, P=0.0084). Therefore, LAA shape may influence LAAFV regardless of heart rhythm.

Increasing BNP concentration was associated with low FV in our study. In a previous study by Tamura et al, the association between BNP and LAA function in patients with acute ischemic stroke was reported.²⁴ They showed that plasma BNP levels significantly correlated with LAAFV regardless of SR or AF rhythm, which was similar to the result in our study. BNP is secreted into the plasma mainly in response to mechanical stretching of the LV, LA and LAA myocardium. The LA is exposed to high pressure caused by AF rhythm or LV diastolic filling pressure, and the long-term increase in LA pressure leads to LA dilatation and LAA dysfunction. Compared with the patients with normal FV, there was a higher rate of AF rhythm at the time of blood collection (62% vs. 31%, P=0.0020) in the patients with low FV in our study. Therefore, it is possible that AF rhythm might increase the serum BNP level. However, it is not certain how the plasma BNP concentration is associated with LAAFV and further investigation is necessary.

The NCW LAA is suspected to show a strong predisposition for cerebral infarction in AF patients, there fore the LAA type should be included in the risk stratification of cardiogenic stroke. And it may also be advisable to administer anticoagulants according to LAA morphology.

Study Limitations

First, this was a retrospective study with a small number of patients in a single center. Therefore, further prospective study with a larger cohort is necessary. Second, the LAA morphology was defined subjectively and classified into 4 types in previous studies,^9,17 but we classified them into only 2 types to minimize interobserver variability. Our suggestion is not a complete one, so strict definition of LAA morphology may be necessary. Third, our analysis of blood coagulation was inadequate. We did not analyze blood markers such as D-dimer, soluble fibrin, fibrin/fibrinogen degradation products, thrombin-antithrombin complex or plasmin α2-plasmin inhibitor complex. Moreover, there is great variability among the LAA shapes in the effect of anticoagulants. Therefore, we could not assess blood markers and further investigation of this aspect is necessary. Fourth, we considered the influence of the cardiac cycle. Previous study showed that the diameter, perimeter, and area of the LAA orifice undergoes no significant change during the various cardiac phases.²⁵ Therefore, we consider that the cardiac cycle has a very small effect. Fifth, Fatkin et al reported the various mechanisms of LAAFV. In patients with SR, LAAFV is mainly caused by LAA contraction, whereas in patients with AF, LAAFV is caused by the following mechanisms. In the first type, LAAFV is associated with vigorous or little fibrillatory motion of the LAA. This is the most frequent pattern in patients with AF. In the next type, prolonged low LAAFV is associated with minimal changes in LAA diameter. In this type, passive emptying of the LAA occurs by inward and outward displacement of its medial wall by the adjacent LV wall during ventricular filling. Thus, the cause of LAAFV is different between AF rhythm and SR. This is a limitation of our study. However, there is currently no appropriate method of measuring LAAFV, so we used the pulsed Doppler method of measuring LAAFV. Finally, we did not assess the LAA emptying fraction (LAAEF), because the method for measuring LAAEF by TEE is very complex and we could not measure it in all patients.

Conclusions

The NCW-type LAA and higher BNP were associated with lower LAAFV in the present study. One reason for the frequent cardiogenic stroke in patients with the NCW type may be lower LAAFV.

Funding

No funding was received for the study.

Disclosures

There are no conflicts of interest on the part of the authors.

References

• 1.    Kimura  K,  Minematsu  K,  Yamaguchi  T; Japan Multicenter Stroke Investigators’ Collaboration (J-MUSIC). Atrial fibrillation as a predictive factor for severe stroke and early death in 15,831 patients with acute ischemic stroke. J Neurol Neurosurg Psychiatry 2005; 76: 679–683.
• 2.    Wolf  PA,  Abbott  RD,  Kannel  WB. Atrial fibrillation as an independent risk factor for stroke: The Framingham Study. Stroke 1991; 22: 983–988.
• 3.    Tomita  F,  Kohya  T,  Sakurai  M,  Kaji  T,  Yokoshiki  H,  Sato  M, et al. Prevalence and clinical characteristics of patients with atrial fibrillation: Analysis of 20,000 cases in Japan. Jpn Circ J 2000; 64: 653–658.
• 4.    Al-Saady  NM,  Obel  OA,  Camm  AJ. Left atrial appendage: Structure, function, and role in thromboembolism. Heart 1999; 82: 547–554.
• 5.    Watson  T,  Shantsila  E,  Lip  GY. Mechanisms of thrombogenesis in atrial fibrillation: Virchow’s triad revisited. Lancet 2009; 373: 155–166.
• 6.    Abe  Y,  Asakura  T,  Gotou  J,  Iwai  M,  Watanabe  Y,  Sando  M, et al. Prediction of embolism in atrial fibrillation: Classification of left atrial thrombi by transesophageal echocardiography. Jpn Circ J 2000; 64: 411–415.
• 7.    Matsuo  Y,  Sandri  M,  Mangner  N,  Majunke  N,  Dähnert  I,  Schuler  G, et al. Interventional closure of the left atrial appendage for stroke prevention. Circ J 2014; 78: 619–624.
• 8.    Machino-Ohtsuka  T,  Seo  Y,  Ishizu  T,  Yanaka  S,  Nakajima  H,  Atsumi  A, et al. Significant improvement of left atrial and left atrial appendage function after catheter ablation for persistent atrial fibrillation. Circ J 2013; 77: 1695–1704.
• 9.    Di Biase  L,  Santangeli  P,  Anselmino  M,  Mohanty  P,  Salvetti  I,  Gili  S, et al. Does the left atrial appendage morphology correlate with the risk of stroke in patients with atrial fibrillation? Results from a multicenter study. J Am Coll Cardiol 2012; 60: 531–538.
• 10.    Anselmino  M,  Scaglione  M,  Di Biase  L,  Gili  S,  Santangeli  P,  Corsinovi  L, et al. Left atrial appendage morphology and silent cerebral ischemia in patients with atrial fibrillation. Heart Rhythm 2014; 11: 2–7.
• 11.    Goldman  ME,  Pearce  LA,  Hart  RG,  Zabalgoitia  M,  Asinger  RW,  Safford  R, et al. Pathophysiologic correlates of thromboembolism in nonvalvular atrial fibrillation. I: Reduced flow velocity in the left atrial appendage (The Stroke Prevention in Atrial Fibrillation [SPAF-III] study). J Am Soc Echocardiogr 1999; 12: 1080–1087.
• 12.    Cappato  R,  Calkins  H,  Chen  SA,  Davies  W,  Iesaka  Y,  Kalman  J, et al. Worldwide survey on the methods, efficacy, and safety of catheter ablation for human atrial fibrillation. Circulation 2005; 111: 1100–1105.
• 13.    Gage  BF,  Waterman  AD,  Shannon  W,  Boechler  M,  Rich  MW,  Radford  MJ. Validation of clinical classification schemes for predicting stroke: Results from the national registry of atrial fibrillation. JAMA 2001; 285: 2864–2870.
• 14.    Ujino  K,  Barnes  ME,  Cha  SS,  Langins  AP,  Bailey  KR,  Seward  JB, et al. Two-dimensional echocardiographic methods for assessment of left atrial volume. Am J Cardiol 2006; 98: 1185–1188.
• 15.    Lang  RM,  Bierig  M,  Devereux  RB,  Flachskampf  FA,  Foster  E,  Pellikka  PA, et al; Chamber Quantification Writing Group; American Society of Echocardiography’s Guidelines and Standards Committee; European Association of Echocardiography. Recommendations for chamber quantification: A report from the American Society of Echocardiography’s Guidelines and Standards Committee and the Chamber Quantification Writing Group, developed in conjunction with the European Association of Echocardiography, a branch of the European Society of Cardiology. J Am Soc Echocardiogr 2005; 18: 1440–1463.
• 16.    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.
• 17.    Kimura  T,  Takatsuki  S,  Inagawa  K,  Katsumata  Y,  Nishiyama  T,  Nishiyama  N, et al. Anatomical characteristics of the left atrial appendage in cardiogenic stroke with low CHADS2 scores. Heart Rhythm 2013; 10: 921–925.
• 18.    Iwakura  K,  Okamura  A,  Koyama  Y,  Date  M,  Higuchi  Y,  Inoue  K, et al. Effect of elevated left ventricular diastolic filling pressure on the frequency of left atrial appendage thrombus in patients with nonvalvular atrial fibrillation. Am J Cardiol 2011; 107: 417–422.
• 19.    Tabata  T,  Oki  T,  Fukuda  N,  Iuchi  A,  Manabe  K,  Kageji  Y, et al. Influence of left atrial pressure on left atrial appendage flow velocity patterns in patients in sinus rhythm. J Am Soc Echocardiogr 1996; 9: 857–864.
• 20.    Khurram  IM,  Dewire  J,  Mager  M,  Maqbool  F,  Zimmerman  SL,  Zipunnikov  V, et al. Relationship between left atrial appendage morphology and stroke in patients with atrial fibrillation. Heart Rhythm 2013; 10: 1843–1849.
• 21.    Obel  OA,  Luddington  L,  Maarouf  N,  Aytemir  K,  Ekwall  C,  Malik  M, et al. Effects of ventricular rate and regularity on the velocity and magnitude of left atrial appendage flow in atrial fibrillation. Heart 2005; 91: 764–768.
• 22.    Ito  T,  Suwa  M,  Kobashi  A,  Yagi  H,  Hirota  Y,  Kawamura  K. Influence of altered loading conditions on left atrial appendage function in vivo. Am J Cardiol 1998; 81: 1056–1059.
• 23.    Goldberg  YH,  Gordon  SC,  Spevack  DM,  Gordon  GM. Disparities in emptying velocity within the left atrial appendage. Eur J Echocardiogr 2010; 11: 290–295.
• 24.    Tamura  H,  Watanabe  T,  Nishiyama  S,  Sasaki  S,  Wanezaki  M,  Arimoto  T, et al. Elevated plasma brain natriuretic peptide levels predict left atrial appendage dysfunction in patients with acute ischemic stroke. J Cardiol 2012; 60: 126–132.
• 25.    Wang  Y,  Di Biase  L,  Horton  RP,  Nguyen  T,  Morhanty  P,  Natale  A. Left atrial appendage studied by computed tomography to help planning for appendage closure device placement. J Cardiovasc Electrophysiol 2010; 21: 973–982.