European Heart Journal - Cardiovascular Imaging Advance Access published May 5, 2015 European Heart Journal – Cardiovascular Imaging doi:10.1093/ehjci/jev117
Correlation between left atrial appendage morphology and flow velocity in patients with paroxysmal atrial fibrillation Keiko Fukushima1†*, Noritoshi Fukushima1,2†, Ken Kato1†, Koichiro Ejima1, Hiroki Sato2, Kenji Fukushima 3, Chihiro Saito 1, Keiko Hayashi 1, Kotaro Arai 1, Tetsuyuki Manaka 1, Kyomi Ashihara1, Morio Shoda 1, and Nobuhisa Hagiwara1 1 Department of Cardiology, Tokyo Women’s Medical University, 8-1 Kawada-cho, Shinjuku-ku, Tokyo 162-8666, Japan; 2Department of Preventive Medicine and Public Health, Tokyo Medical University, Tokyo, Japan; and 3Department of Diagnostic Imaging and Nuclear Medicine, Tokyo Women’s Medical University, Tokyo, Japan
Received 2 February 2015; accepted after revision 15 April 2015
Aims
----------------------------------------------------------------------------------------------------------------------------------------------------------Keywords
left atrial appendage † morphology † paroxysmal atrial fibrillation † transoesophageal echocardiography † multidetector computed tomography
Introduction Thromboembolic events in patients with atrial fibrillation (AF) are deleterious,1,2 with .90% of embolic strokes caused by thrombi that formed in the left atrial appendage (LAA).3,4 It was recently reported that LAA morphology correlates with transient ischaemic attacks (TIA) and strokes in patients with AF.5 – 7 Additionally, reduced LAA flow velocity (LAAFV) is a well-established risk factor for thromboembolism.8 – 10 However, the relationship between LAAFV and LAA morphology has not been fully characterized in
patients with paroxysmal AF. We hypothesized that LAAFV varies with LAA morphology. The aim of this study was to evaluate the associations between LAAFV and types of LAA morphology.
Methods Patients This retrospective study included 144 consecutive patients who had been referred for radiofrequency catheter ablation for symptomatic
* Corresponding author. Tel: +81 3 3353 8111; Fax: +81 3 3356 0441, E-mail: [email protected] †
The authors contributed equally to this work.
Published on behalf of the European Society of Cardiology. All rights reserved. & The Author 2015. For permissions please email: [email protected].
Downloaded from by guest on July 11, 2015
Reduction of left atrial appendage (LAA) flow velocity (FV) is a risk factor for thrombus formation and increases the risk of stroke in patients with atrial fibrillation (AF). Furthermore, LAA morphology is correlated with stroke in patients with AF. The aim of this study was to correlate LAAFV with LAA morphology in patients with AF. ..................................................................................................................................................................................... Methods We studied 96 patients (age 59.0 + 10.2 years, 75% male) referred for radiofrequency catheter ablation for paroxysmal and results AF. All patients underwent computed tomography (CT) and transthoracic and transoesophageal echocardiography during sinus rhythm. LAA morphology was classified as one of the four types (chicken wing, windsock, cactus, and cauliflower) on CT images. There were significant differences in LAAFV among LAA morphologies (chicken wing 73.7 + 21.9 cm/s, windsock 61.9 + 19.6 cm/s, cactus 55.3 + 14.1 cm/s, cauliflower 52.7 + 18.1 cm/s, P ¼ 0.008). Post hoc multiple comparisons showed that LAAFV was higher in patients with chicken wing than in those with cactus (P ¼ 0.006, vs. chicken wing) and cauliflower (P ¼ 0.006, vs. chicken wing), but not with windsock (P ¼ 0.102). After adjustment for clinical and LAA anatomical covariates (orifice area, volume, and trabeculation), multiple linear regression analyses revealed that LAA morphology was an independent determinant of LAAFV [chickens wing: standardized partial regression coefficients (b) ¼ 0.317, P ¼ 0.0014; windsock: b ¼ 0.303, P ¼ 0.038]. ..................................................................................................................................................................................... Conclusion LAA morphology is a significant determinant of LAAFV, suggesting an underlying mechanism for the association between LAA morphology and embolic events.
Page 2 of 8 drug-refractory paroxysmal AF (PAF) from October 2010 to March 2014. All patients had undergone transthoracic echocardiography (TTE), transoesophageal echocardiography (TEE), and multidetector computed tomography (MDCT) using a 64- or 320-slice CT scanner. Patients who had undergone either TEE or MDCT during AF rhythm were excluded from this study (n ¼ 12), as were patients who had undergone MDCT without contrast enhancement because of renal dysfunction (n ¼ 25). Additionally, 11 of the patients who had undergone contrast enhancement CT were excluded, because the quality of the CT images was too poor to permit LAA visualization. Thus, 96 of the 144 patients had undergone enhanced CT, TTE, and TEE during sinus rhythm and were included in this study. In our institution, patients who undergo radiofrequency catheter ablation for AF are routinely anticoagulated with warfarin or non-vitamin K antagonist oral anticoagulants for at least 4 weeks before the ablation, even for those with CHADS2 scores of ≤1. All patients gave their written informed consent.
Echocardiographic data analysis
Measurements of LAAFV LAAFV was measured during sinus rhythm by TEE using pulsed-wave Doppler imaging, as previously reported by our group.14 Briefly, the sample volume (4 mm) was positioned 1 cm from the LAA orifice in the longitudinal view of that appendage. LAAFV was defined as the late diastolic emptying velocity, namely the peak of the outflow velocity after the P-wave on an electrocardiogram, and was measured over three cardiac cycles.
Cardiac MDCT imaging analysis From October 2010 to March 2012, contrasted-enhanced cardiac CT imaging was performed with a 64-slice MSCT scanner (n ¼ 33; Aquilion 64, Toshiba Medical Systems, Otawara, Japan) and from April 2012 to March 2014 with a 320-slice CT scanner (n ¼ 63; Aquilion ONE, Toshiba Medical Systems). The slice acquisition thickness was 1.0 mm. With the 64-slice scanner, a contrast agent was injected through a 20-gauge catheter into the antecubital vein at 0.06 × body weight (kg) mL/s for scan duration +8 s, whereas with the 320-slice scanner, contrast agent was injected at 0.06 × body weight (kg) mL/s for 12 s. Threedimensional LAA structures were reconstructed using an image analysis system (ZIOSTATION2, Ziosoft, Tokyo, Japan). Standard measurements of LAA volume, length, trabeculation, and the angle of the first LAA bend were studied in the resultant three-dimensional images. The LAA orifice was measured manually on multiplanar reformatted images, its size defined by its narrowest portion.15 The formula 0.785 × LAA long diameter × LAA short diameter was used to calculate a surrogate value for the area of the LAA orifice in accordance with its elliptical shape.15 All CT scans were analysed independently by two trained observers blinded to clinical data; conflict was resolved by common agreement with a third observer. Mild trabeculations were defined as minimal or no indentation on the LAA wall, moderate trabeculations as trabeculations
on part of the LAA wall with portions of the LAA showing minimal or no indentation, and extensive trabeculations as diffuse indentation throughout the LAA wall.6
Classification of LAA morphology The types of LAA morphology (chicken wing, windsock, cactus, and cauliflower)5,7 are shown in Figure 1 . As there are numerous anatomical features of LAA,16 a definite classification of LAA morphology is still being developed. Therefore, although the division of LAA morphology into four types was originally designed to help practical planning for a transcatheter LAA closure device placement,17 use of this classification is rapidly expanding.5 – 7,17,18 In the present study, the classification of LAA morphology reported by Kimura et al. was intentionally used as it is based on these four classification types and contains more objective measurements to distinguish the LAA morphologies.7 In this classification, LAAs are classified into the following four types: (i) ‘chicken wing’, a .4-cm-long main lobe with a folded angle of ,1008 (Figure 1A); (ii) ‘windsock’, a main lobe .4 cm long with a folded angle of .1008 (Figure 1B); (iii) ‘cactus’, a ,4-cm-long main lobe with more than two lobes over 1 cm (Figure 1C); and (iv) ‘cauliflower’, a ,4-cm-long main lobe with no forked lobes (Figure 1D).
Statistical analysis All variables were tested for normal distribution using the Shapiro– Wilk test. Continuous variables are presented as mean + standard deviation and were compared using one-way analysis of variance (ANOVA) with Tukey or Dunnett test for multiple comparisons, or the Kruskal – Wallis test, as appropriate. Categorical variables were compared using the x 2 test or Fisher’s exact test. Pearson correlation was used to assess the associations between two quantitative variables. Multiple linear regression analyses were performed to identify significant determinants of LAAFV and then expanded in the following manner. First, by using a dummy variable, the four LAA morphologic types were entered in Model 1. Second, relevant clinical characteristics (i.e. sex, age) were added to Model 1 (Model 2). Third, LAA anatomical characteristics (i.e. LAA orifice area, volume, and trabeculation) were added to Model 2 (Model 3). Because the diameter of the LAA orifice was used to calculate its area, the diameter was excluded from the analysis to avoid collinearity. Finally, the presence of hypertension and dyslipidaemia were added to Model 3 (Model 4). R 2 and adjusted R 2 were calculated in each step to assess the model fit. The inter-observer agreement between readers was evaluated using Cohen’s kappa. A good level of agreement was defined as k ≥ 0.61.19 A P-value of ,0.05 was considered statistically significant. Statistical analyses were performed using IBM SPSS Statistics for Windows, Version 19.0 (IBM Corp., Armonk, NY, USA).
Results Relevant clinical characteristics The baseline general characteristics of the 96 patients are shown in Table 1. Overall, the mean age was 59 + 10 years, 71 (75%) were males, mean body mass index (BMI) was 23.3 + 2.7 kg/m2, and 10 (10.4%) had a history of stroke/TIA. The LAA morphologies were distributed as follows: chicken wing (n ¼ 12, 12.5%), windsock (n ¼ 31, 32.3%), cactus (n ¼ 37, 38.5%), and cauliflower (n ¼ 16, 16.7%). The prevalence of hypertension and dyslipidaemia was lower in patients with chicken wing type than in patients with other types of LAA morphology. However, there were no significant differences between LAA morphology types in age, sex, BMI, prevalence of
Downloaded from by guest on July 11, 2015
All patients had received TTE and TEE within 2 days prior to the scheduled ablation procedure. TTE and TEE were performed with an iE33 ultrasound system (Philips Healthcare, Bothell, WA, USA) using a broadband S4 transducer (2 – 4 MHz) and an X7-2t TEE transducer by two cardiologists who were blinded to the clinical details and results. All images were stored digitally for playback and analysis. The comprehensive echocardiographic measurements by TTE and TEE conformed to the recommendations of the American Society of Echocardiography.11,12 The presence of a left atrial spontaneous echo contrast (SEC) was diagnosed by detecting a characteristic swirling motion distinct from a white noise artifact in the atrial cavity.13
K. Fukushima et al.
Page 3 of 8
Correlation between left atrial appendage morphology and flow velocity
Table 1
Clinical characteristics Overall
Chicken wing
Windsock
Cactus
Cauliflower
P-value
............................................................................................................................................................................... n (%)
12 (12.5)
31 (32.3)
37 (38.5)
16 (16.7)
Age (years)
59.0 + 10.2
96
54.4 + 15.3
59.1 + 10.9
59.8 + 7.9
60.7 + 8.7
0.380
Sex, male (%) BMI
72 (75.0) 23.3 + 2.7
10 (83.3) 22.9 + 3.4
21 (67.7) 23.7 + 2.7
29 (78.4) 23.6 + 2.6
12 (75.0) 22.0 + 2.4
0.673 0.181a
8 (50)
0.020
7 (43.8) 2 (12.5)
0.015 0.537
Hypertension, n (%)
46 (47.9)
1 (8.3)
19 (61.3)
18 (48.6)
Dyslipidaemia, n (%) Diabetes mellitus, n (%)
34 (35.4) 12 (12.5)
2 (16.7) 1 (8.3)
17 (54.8) 6 (19.4)
8 (21.6) 3 (8.1)
History of CV event, n (%)
10 (10.4)
1 (8.3)
2 (6.5)
4 (10.8)
3 (18.8)
0.620
8 (83.3)
1 (8.3)
4 (12.9)
2 (5.4)
1 (6.3)
0.766 0.158
0
45 (46.9)
10 (83.3)
11 (35.5)
16 (43.2)
8 (50.0)
1 ≥2
32 (33.3) 19 (19.8)
1 (8.3) 1 (8.3)
12 (38.7) 8 (25.8)
15 (40.5) 6 (16.2)
4 (25.0) 4 (25.0)
SHD, n (%) CHADS2 score, n (%)
Values are mean + standard deviation, or number (%). BMI, body mass index; CV, cerebral vascular; SHD, structural heart disease. a P-values for continuous variables were obtained by one-way ANOVA and by Kruskal–Wallis test.
structural heart disease or diabetes mellitus, history of cerebral vascular events (stroke/TIA), or CHADS2 score. Furthermore, all patients had received oral anticoagulation. Eighty-nine patients (nine with chicken wing, 29 with windsock, 36 with cactus, and 15
with cauliflower morphology) had received warfarin, and seven patients (three with chicken wing, two with windsock, one with cactus, and one with cauliflower morphology) had received nonvitamin K antagonist oral anticoagulants. The mean value for
Downloaded from by guest on July 11, 2015
Figure 1 Classification of LAA morphology. Representative three-dimensional reconstruction images of each type of LAA are shown. (A) Chicken wing, (B) windsock, (C) cactus, and (D) cauliflower.
Page 4 of 8
K. Fukushima et al.
international normalized ratio of prothrombin time (PT-INR) was 1.94 + 0.44 in patients who received warfarin, while there were no differences between the LAA morphology groups in PT-INR values (chicken wing: 1.98 + 0.34; windsock: 1.88 + 0.38; cactus: 1.98 + 0.52; cauliflower: 1.95 + 0.40; P ¼ 0.517, Kruskal– Wallis test).
Echocardiographic characteristics The patients’ baseline echocardiographic characteristics are shown in Table 2. There were no significant differences between LAA morphology types in left ventricular ejection fraction, left atrial volume index, or E/e′ [ratio of early diastolic transmitral flow velocity (E) to the mitral annular velocity at the early diastolic phase on tissue Doppler (e′ )]. SEC was detected in only five patients, with no significant difference in the prevalence of SEC among LAA morphology types. No participants had thrombus formation.
Cardiac MDCT imaging variables
Table 2
Correlations between LAA type and LAAFV According to ANOVA, there was a significant difference between morphologic types in LAAFV (chicken wing: 73.7 + 21.9 cm/s; windsock: 61.6 + 19.6 cm/s; cactus: 55.3 + 14.1 cm/s; cauliflower: 52.7 + 18.1 cm/s; P ¼ 0.008). Dunnett’s post hoc multiple comparisons showed that LAAFV was significantly higher in patients with chicken wing than in those with cactus (P ¼ 0.006, vs. chicken wing) and cauliflower (P ¼ 0.006, vs. chicken wing), but not with windsock morphology (P ¼ 0.102) (Figure 2). Agreement on LAA morphology was good with a k score of 0.80.
Relationship between LAAFV and associated variables Regression coefficients, standard error, standard partial regression coefficients (b), R 2, and adjusted R 2 are shown in Table 4. Multiple regression analysis showed that LAA morphology was an independent
Echocardiographic variables Overall
Chicken wing
Windsock
Cactus
Cauliflower
P-value
55.7 + 4.9 30.0 + 10.3
57.3 + 5.1 26.1 + 6.1
54.8 + 5.4 32.1 + 9.7
55.7 + 4.9 30.7 + 9.7
55.8 + 3.9 27.0 + 9.9
0.367a 0.252a
9.3 + 3.2
8.5 + 2.7
10.4 + 3.3
9.1 + 3.4
8.3 + 1.9
0.120b
5 (5.2)
1 (8.3)
............................................................................................................................................................................... LVEF (%) LAVI (mL/m2) E/e′ SEC, n (%)
1 (3.2)
1 (2.7)
2 (12.5)
0.337
Values are mean + standard deviation, or number (%). LVEF, left ventricular ejection fraction; LAVI, left atrial volume index; SEC, spontaneous echo contrast. a Kruskal –Wallis test. b One-way ANOVA.
Table 3
MDCT variables Overall
Chicken wing
Windsock
Cactus
Cauliflower
P-value
............................................................................................................................................................................... Orifice area (cm2)
3.7 + 1.0
3.7 + 1.1
4.0 + 1.0
3.3 + 0.9
3.6 + 1.1
0.038a
16.9 + 5.9
18.6 + 5.4
20.1 + 5.4
14.7 + 4.5
14.9 + 7.2
,0.001a
29 (30.2)
3 (25.0)
9 (29.0)
12 (32.4)
5 (31.3)
Moderate
42 (43.8)
8 (66.7)
12 (38.7)
17 (47.2)
5 (31.3)
Extensive
25 (26.0)
1 (8.3)
10 (32.3)
8 (22.2)
6 (37.5)
LAA volume (mL) b
LAA trabeculation Mild
0.524
Values are mean + standard deviation, or number (%). MDCT, multidetector computed tomography; LAA, left atrial appendage. a One-way ANOVA. b Mild trabeculations were defined by minimal or no indentation on the LAA wall, moderate trabeculations were defined by trabeculations on part of the LAA wall with portions of the LAA showing minimal or no indentation, while extensive trabeculations were defined by diffuse indentation throughout the LAA wall.
Downloaded from by guest on July 11, 2015
The patients’ baseline cardiac MDCT imaging variables are shown in Table 3. By ANOVA, there were significant differences between morphologic types in LAA orifice area and volume (P ¼ 0.04, P , 0.001, respectively). The LAA orifice area was the largest in patients with the windsock type (4.0 + 1.0 cm2), and the smallest in patients with the cactus type (3.3 + 0.9 cm2). Tukey’s post hoc multiple comparisons showed a significant difference of LAA orifice area between patients with windsock and with cactus (P ¼ 0.022). Furthermore, the LAA
volume was the largest in patients with the windsock type (20.1 + 5.4 mL) and the smallest in patients with the cactus type (14.7 + 4.5 mL). Tukey’s post hoc multiple comparisons showed that LAA volume was significantly larger in patients with windsock than those with cactus (P ¼ 0.001, vs. windsock) and cauliflower (P ¼ 0.014, vs. windsock). There were no other significant differences by Tukey’s post hoc multiple comparisons.
Correlation between left atrial appendage morphology and flow velocity
Page 5 of 8
and significant determining factor for LAAFV (chickens wing: b ¼ 0.385, P ¼ 0.003) after adjusting for age, sex, orifice area, LAA volume, and degree of trabeculation of LAA in Model 3 (Table 4). Furthermore, because the prevalence of hypertension and dyslipidaemia was different among types of LAA morphology, we adjusted these variables in Model 4. In the full multiple regression model, LAA morphology was independently associated with LAAFV (chickens wing: b ¼ 0.317, P ¼ 0.0014; windsock: b ¼ 0.303, P ¼ 0.038).
Discussion Our study showed that LAAFV during sinus rhythm differed significantly between LAA morphologic types, that LAAFV was significantly higher in patients with chicken wing than cactus and cauliflower, but not with windsock morphology, and that LAA morphology was an independent and the only significant determining factor for LAAFV after adjusting for age, sex, presence of hypertension and dyslipidaemia, LAA orifice area, volume, and degree of LAA trabeculation. Thromboembolic events are significant clinical problems in patients with AF, with previous studies focusing on the clinical value of LAAFV as a predictor of thromboembolism.8 – 10 Our findings of a relationship between LAAFV and the LAA morphology provide additional clinical implications for risk stratification. LAAFV was significantly higher in patients with chicken wing than cactus and cauliflower, but not with windsock morphology. Recently, Di Biase et al. 5 reported that patients with chicken wing morphology were less likely to have cerebral embolic events. Our results support
their findings and suggest a potential mechanism: higher LAAFV is associated with less thrombus formation and stroke,8 – 10 and we found that patients with chicken wing morphology had higher LAAFV. These data suggest that LAA morphology might provide a notable factor for risk stratification in patients with PAF who have low CHADS2 scores, which may help to decide upon prescribing oral anticoagulation therapy or LAA closure in patients who cannot receive anticoagulation drugs because of haemorrhage risk. Whether determination of LAA morphology is reproducible between different observers remains controversial. Moreover, diversity in the prevalence of each LAA morphology (chicken wing, windsock, cactus, and cauliflower) is observed among different studies.6 Specifically, the prevalence for chicken wing ranges from 17.5 to 48%, windsock from 19 to 46.7%, cactus from 5 to 30%, and cauliflower from 3 to 29.1%.5 – 7,17,18 These variations may reflect a difference in the sampling method of the studied patients (i.e. sampling bias) or subtle differences of definition of LAA morphologies used in various studies. Taken together, these findings suggest that an assessment of LAA morphology should be performed by a relevant appropriate classification. The definition of LAA morphology is subjective, although we minimized inter-observer variability by using rigorous definitions that incorporated more objective information for the types of LAA.7 Indeed, agreement on LAA morphology was good with a k score of 0.80. Although complicated definitions of LAA morphology have limitations for use in clinical practice, our stringent definitions were appropriate for study purposes and our findings provide accurate information with important clinical implications.
Downloaded from by guest on July 11, 2015
Figure 2 Correlations between LAAFV and LAA morphology. According to ANOVA, LAAFV differs significantly between types of LAA morphology (P ¼ 0.008). Dunnett’s post hoc multiple comparisons showed that LAAFV was higher in patients with chicken wing than in those with cactus (P ¼ 0.006, vs. chicken wing) and cauliflower morphology (P ¼ 0.006, vs. chicken wing), but not with those with windsock (P ¼ 0.102).
Page 6 of 8
Table 4
K. Fukushima et al.
Multiple linear regression for LAAFV in patients with PAF Regression coefficients
Standard error
Standardized partial regression coefficients (b)
P-value
R2
Adjusted R 2
............................................................................................................................................................................... Model 1
0.120
0.092
0.149
0.102
0.201
0.127
0.245
0.155
Morphology Cauliflower (reference) Chicken wing Windsock
– 20.928
– 6.729
– 0.376
– 0.002
8.821
5.424
0.224
0.107
Cactus Model 2
2.586
5.272
0.068
0.625
Sex, male
5.634
4.268
0.133
0.190
20.153
0.184
20.084
0.409
Age Morphology Cauliflower (reference)
–
–
–
19.501 8.987
6.785 5.414
0.351 0.228
0.005 0.100
Cactus
2.254
5.246
0.06
0.668
Model 3 Sex, male
4.382
4.582
0.103
0.341
20.163
0.185
20.089
0.380
Chicken wing Windsock
Age Morphology Cauliflower (reference)
–
–
–
21.388
–
6.900
0.385
0.003
Windsock Cactus
10.556 2.553
5.723 5.261
0.268 0.067
0.069 0.629
Orifice area
20.020
0.023
20.107
0.383
LAA volume Degree of LAA trabeculation
20.115 4.811
0.435 2.549
20.036 0.194
0.793 0.063
3.607 20.152
4.594 0.184
0.085 20.083
0.435 0.411
Model 4 Sex, male Age Hypertension
27.157
3.759
20.193
0.060
Dyslipidaemia Morphology
23.749
4.093
20.097
0.362
Cauliflower (reference) Chicken wing Windsock Cactus Orifice area LAA volume Degree of LAA trabeculation
–
–
–
17.643 11.923
–
7.003 5.665
0.317 0.303
0.014 0.038
1.458
5.269
0.038
0.783
20.200 20.147
0.023 0.429
20.109 20.047
0.373 0.733
4.783
2.527
0.193
0.062
Sex: Female ¼ 0, Male ¼ 1; Degree of LAA trabeculation: Mild ¼ 1, Moderate ¼ 2, Extensive ¼ 3. LAAFV, left atrial appendage flow velocity; PAF, paroxysmal atrial fibrillation; LAA, left atrial appendage.
The design of the present study did not allow us to identify associations between the type of LAA morphology and LAA thrombogenic milieu, such as SEC and thrombus formation. The prevalence of SEC in patients with AF reportedly varies from 12 to 67%.20,21 In this study, we found a low prevalence of SEC (5.2%). Although we have no data for haemostatic factors (e.g. erythrocyte sedimentation rate, fibrinogen level, thrombin-antithrombin 3 complex, or prothrombin fragment 1 and 2),20 this apparent discrepancy may be partially explained as follows. The prevalence of SEC is reportedly higher in patients with
persistent and long-standing AF than in patients with PAF,18 and all patients in this study had PAF. Furthermore, there were no patients with thrombus formation in LAA. Thus, a possible explanation is that all patients had been receiving oral anticoagulants from 4 weeks prior to the scheduled ablation procedure in accordance with our institution’s protocols. Consequently, our Japanese patients with AF had PT-INR values that were adequate for preventing thromboembolic events.22 In this study, we found a significant difference between LAA morphologic types in the LAA orifice area. Furthermore, the smallest
Downloaded from by guest on July 11, 2015
–
Chicken wing
Page 7 of 8
Correlation between left atrial appendage morphology and flow velocity
Study limitations This study has several limitations. First, it was a single-centre study with a relatively small sample, which may affect the prevalence of LAA morphologic types. Second, our results were obtained from patients with PAF. Thus, the findings cannot be extrapolated to patients with persistent and long-standing AF. In this regard, LAAFV is dependent on basal rhythm (i.e. sinus or AF rhythm), which is reduced during AF rhythm,10 preventing differences between LAA morphology. Therefore, focusing on LAAFV during sinus rhythm is optimal for examining the relationship between LAAFV and LAA morphology. Third, we did not assess LAA function, such as implementing strain and strain rate imaging by TEE.29 To date, there are no established means of measuring LAA function, and trabeculated LAA with pectinate muscles interferes with tracing the shape of the LAA.12 However, it remains possible that there are differences between LAA function and LAA morphology, and associations between LAA function and LAAFV, which should be confirmed in future studies.
Conclusions LAAFV differs significantly between LAA morphologic types, and LAA morphology is a significant determinant of LAAFV. These findings suggest an underlying mechanism for the association between LAA morphology and embolic events in patients with AF.
Acknowledgements The authors thank Mr Tomoaki Hirano, Mr Hiroshi Iimura, Mr Issei Takano, and Mr Fumio Kurokawa for technical support of cardiac MDCT imaging and TEE. Conflict of interest: None declared.
References 1. Wolf PA, Abbott RD, Kannel WB. Atrial fibrillation as an independent risk factor for stroke: the Framingham study. Stroke 1991;22:983 –8. 2. Scherr D, Dalal D, Chilukuri K, Dong J, Spragg D, Henrikson CA et al. Incidence and predictors of left atrial thrombus prior to catheter ablation of atrial fibrillation. J Cardiovasc Electrophysiol 2009;20:379 –84. 3. Blackshear JL, Odell JA. Appendage obliteration to reduce stroke in cardiac surgical patients with atrial fibrillation. Ann Thorac Surg 1996;61:755 – 9. 4. Beigel R, Wunderlich NC, Ho SY, Arsanjani R, Siegel RJ. The left atrial appendage: anatomy, function, and noninvasive evaluation. JACC Cardiovasc Imaging 2014;7: 1251 –65. 5. 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 – 8. 6. 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 – 9. 7. 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 –5. 8. Kamp O, Verhorst PM, Welling RC, Visser CA. Importance of left atrial appendage flow as a predictor of thromboembolic events in patients with atrial fibrillation. Eur Heart J 1999;20:979–85. 9. 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 –9. 10. Handke M, Harloff A, Hetzel A, Olschewski M, Bode C, Geibel A. Left atrial appendage flow velocity as a quantitative surrogate parameter for thromboembolic risk: determinants and relationship to spontaneous echocontrast and thrombus formation –a transesophageal echocardiographic study in 500 patients with cerebral ischemia. J Am Soc Echocardiogr 2005;18:1366 –72. 11. Lang RM, Bierig M, Devereux RB, Flachskampf FA, Foster E, Pellikka PA et al. 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 –63. 12. Hahn RT, Abraham T, Adams MS, Bruce CJ, Glas KE, Lang RM et al. Guidelines for performing a comprehensive transesophageal echocardiographic examination: recommendations from the American society of echocardiography and the society of cardiovascular anesthesiologists. J Am Soc Echocardiogr 2013;26:921 –64. 13. Vincelj J, Sokol I, Jaksic O. Prevalence and clinical significance of left atrial spontaneous echo contrast detected by transesophageal echocardiography. Echocardiography 2002;19:319 –24. 14. Fukushima K, Fukushima N, Ejima K, Kato K, Sato Y, Uematsu S et al. Left atrial appendage flow velocity and time from p-wave onset to tissue doppler-derived a’ predict atrial fibrillation recurrence after radiofrequency catheter ablation. Echocardiography 2014. doi: 10.1111/echo.12823. 15. Lee JM, Shim J, Uhm JS, Kim YJ, Lee HJ, Pak HN et al. Impact of increased orifice size and decreased flow velocity of left atrial appendage on stroke in nonvalvular atrial fibrillation. Am J Cardiol 2014;113:963–9. 16. Veinot JP, Harrity PJ, Gentile F, Khandheria BK, Bailey KR, Eickholt JT et al. Anatomy of the normal left atrial appendage: a quantitative study of age-related changes in 500 autopsy hearts: implications for echocardiographic examination. Circulation 1997; 96:3112 –5. 17. 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 – 82. 18. Petersen M, Roehrich A, Balzer J, Shin DI, Meyer C, Kelm M et al. Left atrial appendage morphology is closely associated with specific echocardiographic flow pattern in patients with atrial fibrillation. Europace 2014. doi:10.1093/europace/euu347. 19. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977;33:159 –74. 20. Patel SV, Flaker G. Is early cardioversion for atrial fibrillation safe in patients with spontaneous echocardiographic contrast? Clin Cardiol 2008;31:148 –52.
Downloaded from by guest on July 11, 2015
LAA orifice areas were found in patients with the cactus type of morphology, consistent with a previous report.6 Although previous studies have reported that LAA orifice size is associated with LAAFV and stroke/TIA in patients with AF,6,15,23 the orifice area was not a significant determinant of LAAFV in the present study. This may be explained by the fact that our study did not include patients with persistent and long-standing AF. First, it was reported that LAAFV is lower in patients with persistent and long-standing AF than in those with PAF.18 The lower LAAFV in patients with persistent and long-standing AF directly affects the relationship between LAAFV and the orifice area. Second, patients with persistent and long-standing AF have higher LAA volumes than those with PAF.24 In this regard, we found a significant relationship between LAA orifice area and volume, even in patients with PAF. It has not been definitely established that a larger orifice area accompanies the greater LAA volume when the basal rhythm changes from paroxysmal to persistent and long-standing AF. However, greater LAA volumes may have a causal effect on the relationship between orifice area and LAAFV. Further studies are required to evaluate whether serial changes in basal rhythm affect the relationship between LAA orifice area and volume in parallel with changes in LAAFV. Although we did not assess the presence of LA structural abnormalities in this study, accessory LAA, left atrial diverticulum, and atrial septal aneurysm (ASA) are well-recognized anatomical variants by recent cardiac imaging technology.25 – 27 Moreover, the presence of these LA structure abnormalities might directly or indirectly lead to formation of the thrombosis and changes in LAAFV. Indeed, previous studies reported that the presence of ASA decreases LA function, which affects LAAFV.25,28 As it remains unclear whether the presence of the accessory LAA and left atrial diverticula also modify LA function and affect LAAFV, further studies are required to confirm this challenging issue.
Page 8 of 8 21. Bernhardt P, Schmidt H, Hammerstingl C, Luderitz B, Omran H. Patients with atrial fibrillation and dense spontaneous echo contrast at high risk a prospective and serial follow-up over 12 months with transesophageal echocardiography and cerebral magnetic resonance imaging. J Am Coll Cardiol 2005;45:1807 –12. 22. Inoue H, Okumura K, Atarashi H, Yamashita T, Origasa H, Kumagai N et al. Target international normalized ratio values for preventing thromboembolic and hemorrhagic events in Japanese patients with non-valvular atrial fibrillation: results of the J-rhythm registry. Circ J 2013;77:2264 –70. 23. Agmon Y, Khandheria BK, Meissner I, Petterson TM, O’Fallon WM, Wiebers DO et al. Are left atrial appendage flow velocities adequate surrogates of global left atrial function? A population-based transthoracic and transesophageal echocardiographic study. J Am Soc Echocardiogr 2002;15:433–40. 24. Park HC, Shin J, Ban JE, Choi JI, Park SW, Kim YH. Left atrial appendage: morphology and function in patients with paroxysmal and persistent atrial fibrillation. Int J Cardiovasc Imaging 2013;29:935 –44.
K. Fukushima et al.
25. Na JO, Shin SY, Lim HE, Choi CU, Kim SH, Kim JW et al. Impaired transport function of the left atrium and left atrial appendage in cryptogenic stroke patients with atrial septal aneurysm and without patent foramen ovale. Eur J Echocardiogr 2011;12: 140 –7. 26. Duerinckx AJ, Vanovermeire O. Accessory appendages of the left atrium as seen during 64-slice coronary CT angiography. Int J Cardiovasc Imaging 2008;24:215 –21. 27. Abbara S, Mundo-Sagardia JA, Hoffmann U, Cury RC. Cardiac CT assessment of left atrial accessory appendages and diverticula. Am J Roentgenol 2009;193:807 –12. 28. Goch A, Banach M, Piotrowski G, Szadkowska I, Goch JH. Echocardiographic evaluation of the left atrium and left atrial appendage function in patients with atrial septum aneurysm: implications for thromboembolic complications. Thorac Cardiovasc Surg 2007;55:365 –70. 29. Sevimli S, Gundogdu F, Arslan S, Aksakal E, Gurlertop HY, Islamoglu Y et al. Strain and strain rate imaging in evaluating left atrial appendage function by transesophageal echocardiography. Echocardiography 2007;24:823 –9.
Downloaded from by guest on July 11, 2015
Correlation between left atrial appendage morphology and flow velocity in patients with paroxysmal atrial fibrillation Keiko Fukushima1†*, Noritoshi Fukushima1,2†, Ken Kato1†, Koichiro Ejima1, Hiroki Sato2, Kenji Fukushima 3, Chihiro Saito 1, Keiko Hayashi 1, Kotaro Arai 1, Tetsuyuki Manaka 1, Kyomi Ashihara1, Morio Shoda 1, and Nobuhisa Hagiwara1 1 Department of Cardiology, Tokyo Women’s Medical University, 8-1 Kawada-cho, Shinjuku-ku, Tokyo 162-8666, Japan; 2Department of Preventive Medicine and Public Health, Tokyo Medical University, Tokyo, Japan; and 3Department of Diagnostic Imaging and Nuclear Medicine, Tokyo Women’s Medical University, Tokyo, Japan
Received 2 February 2015; accepted after revision 15 April 2015
Aims
----------------------------------------------------------------------------------------------------------------------------------------------------------Keywords
left atrial appendage † morphology † paroxysmal atrial fibrillation † transoesophageal echocardiography † multidetector computed tomography
Introduction Thromboembolic events in patients with atrial fibrillation (AF) are deleterious,1,2 with .90% of embolic strokes caused by thrombi that formed in the left atrial appendage (LAA).3,4 It was recently reported that LAA morphology correlates with transient ischaemic attacks (TIA) and strokes in patients with AF.5 – 7 Additionally, reduced LAA flow velocity (LAAFV) is a well-established risk factor for thromboembolism.8 – 10 However, the relationship between LAAFV and LAA morphology has not been fully characterized in
patients with paroxysmal AF. We hypothesized that LAAFV varies with LAA morphology. The aim of this study was to evaluate the associations between LAAFV and types of LAA morphology.
Methods Patients This retrospective study included 144 consecutive patients who had been referred for radiofrequency catheter ablation for symptomatic
* Corresponding author. Tel: +81 3 3353 8111; Fax: +81 3 3356 0441, E-mail: [email protected] †
The authors contributed equally to this work.
Published on behalf of the European Society of Cardiology. All rights reserved. & The Author 2015. For permissions please email: [email protected].
Downloaded from by guest on July 11, 2015
Reduction of left atrial appendage (LAA) flow velocity (FV) is a risk factor for thrombus formation and increases the risk of stroke in patients with atrial fibrillation (AF). Furthermore, LAA morphology is correlated with stroke in patients with AF. The aim of this study was to correlate LAAFV with LAA morphology in patients with AF. ..................................................................................................................................................................................... Methods We studied 96 patients (age 59.0 + 10.2 years, 75% male) referred for radiofrequency catheter ablation for paroxysmal and results AF. All patients underwent computed tomography (CT) and transthoracic and transoesophageal echocardiography during sinus rhythm. LAA morphology was classified as one of the four types (chicken wing, windsock, cactus, and cauliflower) on CT images. There were significant differences in LAAFV among LAA morphologies (chicken wing 73.7 + 21.9 cm/s, windsock 61.9 + 19.6 cm/s, cactus 55.3 + 14.1 cm/s, cauliflower 52.7 + 18.1 cm/s, P ¼ 0.008). Post hoc multiple comparisons showed that LAAFV was higher in patients with chicken wing than in those with cactus (P ¼ 0.006, vs. chicken wing) and cauliflower (P ¼ 0.006, vs. chicken wing), but not with windsock (P ¼ 0.102). After adjustment for clinical and LAA anatomical covariates (orifice area, volume, and trabeculation), multiple linear regression analyses revealed that LAA morphology was an independent determinant of LAAFV [chickens wing: standardized partial regression coefficients (b) ¼ 0.317, P ¼ 0.0014; windsock: b ¼ 0.303, P ¼ 0.038]. ..................................................................................................................................................................................... Conclusion LAA morphology is a significant determinant of LAAFV, suggesting an underlying mechanism for the association between LAA morphology and embolic events.
Page 2 of 8 drug-refractory paroxysmal AF (PAF) from October 2010 to March 2014. All patients had undergone transthoracic echocardiography (TTE), transoesophageal echocardiography (TEE), and multidetector computed tomography (MDCT) using a 64- or 320-slice CT scanner. Patients who had undergone either TEE or MDCT during AF rhythm were excluded from this study (n ¼ 12), as were patients who had undergone MDCT without contrast enhancement because of renal dysfunction (n ¼ 25). Additionally, 11 of the patients who had undergone contrast enhancement CT were excluded, because the quality of the CT images was too poor to permit LAA visualization. Thus, 96 of the 144 patients had undergone enhanced CT, TTE, and TEE during sinus rhythm and were included in this study. In our institution, patients who undergo radiofrequency catheter ablation for AF are routinely anticoagulated with warfarin or non-vitamin K antagonist oral anticoagulants for at least 4 weeks before the ablation, even for those with CHADS2 scores of ≤1. All patients gave their written informed consent.
Echocardiographic data analysis
Measurements of LAAFV LAAFV was measured during sinus rhythm by TEE using pulsed-wave Doppler imaging, as previously reported by our group.14 Briefly, the sample volume (4 mm) was positioned 1 cm from the LAA orifice in the longitudinal view of that appendage. LAAFV was defined as the late diastolic emptying velocity, namely the peak of the outflow velocity after the P-wave on an electrocardiogram, and was measured over three cardiac cycles.
Cardiac MDCT imaging analysis From October 2010 to March 2012, contrasted-enhanced cardiac CT imaging was performed with a 64-slice MSCT scanner (n ¼ 33; Aquilion 64, Toshiba Medical Systems, Otawara, Japan) and from April 2012 to March 2014 with a 320-slice CT scanner (n ¼ 63; Aquilion ONE, Toshiba Medical Systems). The slice acquisition thickness was 1.0 mm. With the 64-slice scanner, a contrast agent was injected through a 20-gauge catheter into the antecubital vein at 0.06 × body weight (kg) mL/s for scan duration +8 s, whereas with the 320-slice scanner, contrast agent was injected at 0.06 × body weight (kg) mL/s for 12 s. Threedimensional LAA structures were reconstructed using an image analysis system (ZIOSTATION2, Ziosoft, Tokyo, Japan). Standard measurements of LAA volume, length, trabeculation, and the angle of the first LAA bend were studied in the resultant three-dimensional images. The LAA orifice was measured manually on multiplanar reformatted images, its size defined by its narrowest portion.15 The formula 0.785 × LAA long diameter × LAA short diameter was used to calculate a surrogate value for the area of the LAA orifice in accordance with its elliptical shape.15 All CT scans were analysed independently by two trained observers blinded to clinical data; conflict was resolved by common agreement with a third observer. Mild trabeculations were defined as minimal or no indentation on the LAA wall, moderate trabeculations as trabeculations
on part of the LAA wall with portions of the LAA showing minimal or no indentation, and extensive trabeculations as diffuse indentation throughout the LAA wall.6
Classification of LAA morphology The types of LAA morphology (chicken wing, windsock, cactus, and cauliflower)5,7 are shown in Figure 1 . As there are numerous anatomical features of LAA,16 a definite classification of LAA morphology is still being developed. Therefore, although the division of LAA morphology into four types was originally designed to help practical planning for a transcatheter LAA closure device placement,17 use of this classification is rapidly expanding.5 – 7,17,18 In the present study, the classification of LAA morphology reported by Kimura et al. was intentionally used as it is based on these four classification types and contains more objective measurements to distinguish the LAA morphologies.7 In this classification, LAAs are classified into the following four types: (i) ‘chicken wing’, a .4-cm-long main lobe with a folded angle of ,1008 (Figure 1A); (ii) ‘windsock’, a main lobe .4 cm long with a folded angle of .1008 (Figure 1B); (iii) ‘cactus’, a ,4-cm-long main lobe with more than two lobes over 1 cm (Figure 1C); and (iv) ‘cauliflower’, a ,4-cm-long main lobe with no forked lobes (Figure 1D).
Statistical analysis All variables were tested for normal distribution using the Shapiro– Wilk test. Continuous variables are presented as mean + standard deviation and were compared using one-way analysis of variance (ANOVA) with Tukey or Dunnett test for multiple comparisons, or the Kruskal – Wallis test, as appropriate. Categorical variables were compared using the x 2 test or Fisher’s exact test. Pearson correlation was used to assess the associations between two quantitative variables. Multiple linear regression analyses were performed to identify significant determinants of LAAFV and then expanded in the following manner. First, by using a dummy variable, the four LAA morphologic types were entered in Model 1. Second, relevant clinical characteristics (i.e. sex, age) were added to Model 1 (Model 2). Third, LAA anatomical characteristics (i.e. LAA orifice area, volume, and trabeculation) were added to Model 2 (Model 3). Because the diameter of the LAA orifice was used to calculate its area, the diameter was excluded from the analysis to avoid collinearity. Finally, the presence of hypertension and dyslipidaemia were added to Model 3 (Model 4). R 2 and adjusted R 2 were calculated in each step to assess the model fit. The inter-observer agreement between readers was evaluated using Cohen’s kappa. A good level of agreement was defined as k ≥ 0.61.19 A P-value of ,0.05 was considered statistically significant. Statistical analyses were performed using IBM SPSS Statistics for Windows, Version 19.0 (IBM Corp., Armonk, NY, USA).
Results Relevant clinical characteristics The baseline general characteristics of the 96 patients are shown in Table 1. Overall, the mean age was 59 + 10 years, 71 (75%) were males, mean body mass index (BMI) was 23.3 + 2.7 kg/m2, and 10 (10.4%) had a history of stroke/TIA. The LAA morphologies were distributed as follows: chicken wing (n ¼ 12, 12.5%), windsock (n ¼ 31, 32.3%), cactus (n ¼ 37, 38.5%), and cauliflower (n ¼ 16, 16.7%). The prevalence of hypertension and dyslipidaemia was lower in patients with chicken wing type than in patients with other types of LAA morphology. However, there were no significant differences between LAA morphology types in age, sex, BMI, prevalence of
Downloaded from by guest on July 11, 2015
All patients had received TTE and TEE within 2 days prior to the scheduled ablation procedure. TTE and TEE were performed with an iE33 ultrasound system (Philips Healthcare, Bothell, WA, USA) using a broadband S4 transducer (2 – 4 MHz) and an X7-2t TEE transducer by two cardiologists who were blinded to the clinical details and results. All images were stored digitally for playback and analysis. The comprehensive echocardiographic measurements by TTE and TEE conformed to the recommendations of the American Society of Echocardiography.11,12 The presence of a left atrial spontaneous echo contrast (SEC) was diagnosed by detecting a characteristic swirling motion distinct from a white noise artifact in the atrial cavity.13
K. Fukushima et al.
Page 3 of 8
Correlation between left atrial appendage morphology and flow velocity
Table 1
Clinical characteristics Overall
Chicken wing
Windsock
Cactus
Cauliflower
P-value
............................................................................................................................................................................... n (%)
12 (12.5)
31 (32.3)
37 (38.5)
16 (16.7)
Age (years)
59.0 + 10.2
96
54.4 + 15.3
59.1 + 10.9
59.8 + 7.9
60.7 + 8.7
0.380
Sex, male (%) BMI
72 (75.0) 23.3 + 2.7
10 (83.3) 22.9 + 3.4
21 (67.7) 23.7 + 2.7
29 (78.4) 23.6 + 2.6
12 (75.0) 22.0 + 2.4
0.673 0.181a
8 (50)
0.020
7 (43.8) 2 (12.5)
0.015 0.537
Hypertension, n (%)
46 (47.9)
1 (8.3)
19 (61.3)
18 (48.6)
Dyslipidaemia, n (%) Diabetes mellitus, n (%)
34 (35.4) 12 (12.5)
2 (16.7) 1 (8.3)
17 (54.8) 6 (19.4)
8 (21.6) 3 (8.1)
History of CV event, n (%)
10 (10.4)
1 (8.3)
2 (6.5)
4 (10.8)
3 (18.8)
0.620
8 (83.3)
1 (8.3)
4 (12.9)
2 (5.4)
1 (6.3)
0.766 0.158
0
45 (46.9)
10 (83.3)
11 (35.5)
16 (43.2)
8 (50.0)
1 ≥2
32 (33.3) 19 (19.8)
1 (8.3) 1 (8.3)
12 (38.7) 8 (25.8)
15 (40.5) 6 (16.2)
4 (25.0) 4 (25.0)
SHD, n (%) CHADS2 score, n (%)
Values are mean + standard deviation, or number (%). BMI, body mass index; CV, cerebral vascular; SHD, structural heart disease. a P-values for continuous variables were obtained by one-way ANOVA and by Kruskal–Wallis test.
structural heart disease or diabetes mellitus, history of cerebral vascular events (stroke/TIA), or CHADS2 score. Furthermore, all patients had received oral anticoagulation. Eighty-nine patients (nine with chicken wing, 29 with windsock, 36 with cactus, and 15
with cauliflower morphology) had received warfarin, and seven patients (three with chicken wing, two with windsock, one with cactus, and one with cauliflower morphology) had received nonvitamin K antagonist oral anticoagulants. The mean value for
Downloaded from by guest on July 11, 2015
Figure 1 Classification of LAA morphology. Representative three-dimensional reconstruction images of each type of LAA are shown. (A) Chicken wing, (B) windsock, (C) cactus, and (D) cauliflower.
Page 4 of 8
K. Fukushima et al.
international normalized ratio of prothrombin time (PT-INR) was 1.94 + 0.44 in patients who received warfarin, while there were no differences between the LAA morphology groups in PT-INR values (chicken wing: 1.98 + 0.34; windsock: 1.88 + 0.38; cactus: 1.98 + 0.52; cauliflower: 1.95 + 0.40; P ¼ 0.517, Kruskal– Wallis test).
Echocardiographic characteristics The patients’ baseline echocardiographic characteristics are shown in Table 2. There were no significant differences between LAA morphology types in left ventricular ejection fraction, left atrial volume index, or E/e′ [ratio of early diastolic transmitral flow velocity (E) to the mitral annular velocity at the early diastolic phase on tissue Doppler (e′ )]. SEC was detected in only five patients, with no significant difference in the prevalence of SEC among LAA morphology types. No participants had thrombus formation.
Cardiac MDCT imaging variables
Table 2
Correlations between LAA type and LAAFV According to ANOVA, there was a significant difference between morphologic types in LAAFV (chicken wing: 73.7 + 21.9 cm/s; windsock: 61.6 + 19.6 cm/s; cactus: 55.3 + 14.1 cm/s; cauliflower: 52.7 + 18.1 cm/s; P ¼ 0.008). Dunnett’s post hoc multiple comparisons showed that LAAFV was significantly higher in patients with chicken wing than in those with cactus (P ¼ 0.006, vs. chicken wing) and cauliflower (P ¼ 0.006, vs. chicken wing), but not with windsock morphology (P ¼ 0.102) (Figure 2). Agreement on LAA morphology was good with a k score of 0.80.
Relationship between LAAFV and associated variables Regression coefficients, standard error, standard partial regression coefficients (b), R 2, and adjusted R 2 are shown in Table 4. Multiple regression analysis showed that LAA morphology was an independent
Echocardiographic variables Overall
Chicken wing
Windsock
Cactus
Cauliflower
P-value
55.7 + 4.9 30.0 + 10.3
57.3 + 5.1 26.1 + 6.1
54.8 + 5.4 32.1 + 9.7
55.7 + 4.9 30.7 + 9.7
55.8 + 3.9 27.0 + 9.9
0.367a 0.252a
9.3 + 3.2
8.5 + 2.7
10.4 + 3.3
9.1 + 3.4
8.3 + 1.9
0.120b
5 (5.2)
1 (8.3)
............................................................................................................................................................................... LVEF (%) LAVI (mL/m2) E/e′ SEC, n (%)
1 (3.2)
1 (2.7)
2 (12.5)
0.337
Values are mean + standard deviation, or number (%). LVEF, left ventricular ejection fraction; LAVI, left atrial volume index; SEC, spontaneous echo contrast. a Kruskal –Wallis test. b One-way ANOVA.
Table 3
MDCT variables Overall
Chicken wing
Windsock
Cactus
Cauliflower
P-value
............................................................................................................................................................................... Orifice area (cm2)
3.7 + 1.0
3.7 + 1.1
4.0 + 1.0
3.3 + 0.9
3.6 + 1.1
0.038a
16.9 + 5.9
18.6 + 5.4
20.1 + 5.4
14.7 + 4.5
14.9 + 7.2
,0.001a
29 (30.2)
3 (25.0)
9 (29.0)
12 (32.4)
5 (31.3)
Moderate
42 (43.8)
8 (66.7)
12 (38.7)
17 (47.2)
5 (31.3)
Extensive
25 (26.0)
1 (8.3)
10 (32.3)
8 (22.2)
6 (37.5)
LAA volume (mL) b
LAA trabeculation Mild
0.524
Values are mean + standard deviation, or number (%). MDCT, multidetector computed tomography; LAA, left atrial appendage. a One-way ANOVA. b Mild trabeculations were defined by minimal or no indentation on the LAA wall, moderate trabeculations were defined by trabeculations on part of the LAA wall with portions of the LAA showing minimal or no indentation, while extensive trabeculations were defined by diffuse indentation throughout the LAA wall.
Downloaded from by guest on July 11, 2015
The patients’ baseline cardiac MDCT imaging variables are shown in Table 3. By ANOVA, there were significant differences between morphologic types in LAA orifice area and volume (P ¼ 0.04, P , 0.001, respectively). The LAA orifice area was the largest in patients with the windsock type (4.0 + 1.0 cm2), and the smallest in patients with the cactus type (3.3 + 0.9 cm2). Tukey’s post hoc multiple comparisons showed a significant difference of LAA orifice area between patients with windsock and with cactus (P ¼ 0.022). Furthermore, the LAA
volume was the largest in patients with the windsock type (20.1 + 5.4 mL) and the smallest in patients with the cactus type (14.7 + 4.5 mL). Tukey’s post hoc multiple comparisons showed that LAA volume was significantly larger in patients with windsock than those with cactus (P ¼ 0.001, vs. windsock) and cauliflower (P ¼ 0.014, vs. windsock). There were no other significant differences by Tukey’s post hoc multiple comparisons.
Correlation between left atrial appendage morphology and flow velocity
Page 5 of 8
and significant determining factor for LAAFV (chickens wing: b ¼ 0.385, P ¼ 0.003) after adjusting for age, sex, orifice area, LAA volume, and degree of trabeculation of LAA in Model 3 (Table 4). Furthermore, because the prevalence of hypertension and dyslipidaemia was different among types of LAA morphology, we adjusted these variables in Model 4. In the full multiple regression model, LAA morphology was independently associated with LAAFV (chickens wing: b ¼ 0.317, P ¼ 0.0014; windsock: b ¼ 0.303, P ¼ 0.038).
Discussion Our study showed that LAAFV during sinus rhythm differed significantly between LAA morphologic types, that LAAFV was significantly higher in patients with chicken wing than cactus and cauliflower, but not with windsock morphology, and that LAA morphology was an independent and the only significant determining factor for LAAFV after adjusting for age, sex, presence of hypertension and dyslipidaemia, LAA orifice area, volume, and degree of LAA trabeculation. Thromboembolic events are significant clinical problems in patients with AF, with previous studies focusing on the clinical value of LAAFV as a predictor of thromboembolism.8 – 10 Our findings of a relationship between LAAFV and the LAA morphology provide additional clinical implications for risk stratification. LAAFV was significantly higher in patients with chicken wing than cactus and cauliflower, but not with windsock morphology. Recently, Di Biase et al. 5 reported that patients with chicken wing morphology were less likely to have cerebral embolic events. Our results support
their findings and suggest a potential mechanism: higher LAAFV is associated with less thrombus formation and stroke,8 – 10 and we found that patients with chicken wing morphology had higher LAAFV. These data suggest that LAA morphology might provide a notable factor for risk stratification in patients with PAF who have low CHADS2 scores, which may help to decide upon prescribing oral anticoagulation therapy or LAA closure in patients who cannot receive anticoagulation drugs because of haemorrhage risk. Whether determination of LAA morphology is reproducible between different observers remains controversial. Moreover, diversity in the prevalence of each LAA morphology (chicken wing, windsock, cactus, and cauliflower) is observed among different studies.6 Specifically, the prevalence for chicken wing ranges from 17.5 to 48%, windsock from 19 to 46.7%, cactus from 5 to 30%, and cauliflower from 3 to 29.1%.5 – 7,17,18 These variations may reflect a difference in the sampling method of the studied patients (i.e. sampling bias) or subtle differences of definition of LAA morphologies used in various studies. Taken together, these findings suggest that an assessment of LAA morphology should be performed by a relevant appropriate classification. The definition of LAA morphology is subjective, although we minimized inter-observer variability by using rigorous definitions that incorporated more objective information for the types of LAA.7 Indeed, agreement on LAA morphology was good with a k score of 0.80. Although complicated definitions of LAA morphology have limitations for use in clinical practice, our stringent definitions were appropriate for study purposes and our findings provide accurate information with important clinical implications.
Downloaded from by guest on July 11, 2015
Figure 2 Correlations between LAAFV and LAA morphology. According to ANOVA, LAAFV differs significantly between types of LAA morphology (P ¼ 0.008). Dunnett’s post hoc multiple comparisons showed that LAAFV was higher in patients with chicken wing than in those with cactus (P ¼ 0.006, vs. chicken wing) and cauliflower morphology (P ¼ 0.006, vs. chicken wing), but not with those with windsock (P ¼ 0.102).
Page 6 of 8
Table 4
K. Fukushima et al.
Multiple linear regression for LAAFV in patients with PAF Regression coefficients
Standard error
Standardized partial regression coefficients (b)
P-value
R2
Adjusted R 2
............................................................................................................................................................................... Model 1
0.120
0.092
0.149
0.102
0.201
0.127
0.245
0.155
Morphology Cauliflower (reference) Chicken wing Windsock
– 20.928
– 6.729
– 0.376
– 0.002
8.821
5.424
0.224
0.107
Cactus Model 2
2.586
5.272
0.068
0.625
Sex, male
5.634
4.268
0.133
0.190
20.153
0.184
20.084
0.409
Age Morphology Cauliflower (reference)
–
–
–
19.501 8.987
6.785 5.414
0.351 0.228
0.005 0.100
Cactus
2.254
5.246
0.06
0.668
Model 3 Sex, male
4.382
4.582
0.103
0.341
20.163
0.185
20.089
0.380
Chicken wing Windsock
Age Morphology Cauliflower (reference)
–
–
–
21.388
–
6.900
0.385
0.003
Windsock Cactus
10.556 2.553
5.723 5.261
0.268 0.067
0.069 0.629
Orifice area
20.020
0.023
20.107
0.383
LAA volume Degree of LAA trabeculation
20.115 4.811
0.435 2.549
20.036 0.194
0.793 0.063
3.607 20.152
4.594 0.184
0.085 20.083
0.435 0.411
Model 4 Sex, male Age Hypertension
27.157
3.759
20.193
0.060
Dyslipidaemia Morphology
23.749
4.093
20.097
0.362
Cauliflower (reference) Chicken wing Windsock Cactus Orifice area LAA volume Degree of LAA trabeculation
–
–
–
17.643 11.923
–
7.003 5.665
0.317 0.303
0.014 0.038
1.458
5.269
0.038
0.783
20.200 20.147
0.023 0.429
20.109 20.047
0.373 0.733
4.783
2.527
0.193
0.062
Sex: Female ¼ 0, Male ¼ 1; Degree of LAA trabeculation: Mild ¼ 1, Moderate ¼ 2, Extensive ¼ 3. LAAFV, left atrial appendage flow velocity; PAF, paroxysmal atrial fibrillation; LAA, left atrial appendage.
The design of the present study did not allow us to identify associations between the type of LAA morphology and LAA thrombogenic milieu, such as SEC and thrombus formation. The prevalence of SEC in patients with AF reportedly varies from 12 to 67%.20,21 In this study, we found a low prevalence of SEC (5.2%). Although we have no data for haemostatic factors (e.g. erythrocyte sedimentation rate, fibrinogen level, thrombin-antithrombin 3 complex, or prothrombin fragment 1 and 2),20 this apparent discrepancy may be partially explained as follows. The prevalence of SEC is reportedly higher in patients with
persistent and long-standing AF than in patients with PAF,18 and all patients in this study had PAF. Furthermore, there were no patients with thrombus formation in LAA. Thus, a possible explanation is that all patients had been receiving oral anticoagulants from 4 weeks prior to the scheduled ablation procedure in accordance with our institution’s protocols. Consequently, our Japanese patients with AF had PT-INR values that were adequate for preventing thromboembolic events.22 In this study, we found a significant difference between LAA morphologic types in the LAA orifice area. Furthermore, the smallest
Downloaded from by guest on July 11, 2015
–
Chicken wing
Page 7 of 8
Correlation between left atrial appendage morphology and flow velocity
Study limitations This study has several limitations. First, it was a single-centre study with a relatively small sample, which may affect the prevalence of LAA morphologic types. Second, our results were obtained from patients with PAF. Thus, the findings cannot be extrapolated to patients with persistent and long-standing AF. In this regard, LAAFV is dependent on basal rhythm (i.e. sinus or AF rhythm), which is reduced during AF rhythm,10 preventing differences between LAA morphology. Therefore, focusing on LAAFV during sinus rhythm is optimal for examining the relationship between LAAFV and LAA morphology. Third, we did not assess LAA function, such as implementing strain and strain rate imaging by TEE.29 To date, there are no established means of measuring LAA function, and trabeculated LAA with pectinate muscles interferes with tracing the shape of the LAA.12 However, it remains possible that there are differences between LAA function and LAA morphology, and associations between LAA function and LAAFV, which should be confirmed in future studies.
Conclusions LAAFV differs significantly between LAA morphologic types, and LAA morphology is a significant determinant of LAAFV. These findings suggest an underlying mechanism for the association between LAA morphology and embolic events in patients with AF.
Acknowledgements The authors thank Mr Tomoaki Hirano, Mr Hiroshi Iimura, Mr Issei Takano, and Mr Fumio Kurokawa for technical support of cardiac MDCT imaging and TEE. Conflict of interest: None declared.
References 1. Wolf PA, Abbott RD, Kannel WB. Atrial fibrillation as an independent risk factor for stroke: the Framingham study. Stroke 1991;22:983 –8. 2. Scherr D, Dalal D, Chilukuri K, Dong J, Spragg D, Henrikson CA et al. Incidence and predictors of left atrial thrombus prior to catheter ablation of atrial fibrillation. J Cardiovasc Electrophysiol 2009;20:379 –84. 3. Blackshear JL, Odell JA. Appendage obliteration to reduce stroke in cardiac surgical patients with atrial fibrillation. Ann Thorac Surg 1996;61:755 – 9. 4. Beigel R, Wunderlich NC, Ho SY, Arsanjani R, Siegel RJ. The left atrial appendage: anatomy, function, and noninvasive evaluation. JACC Cardiovasc Imaging 2014;7: 1251 –65. 5. 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 – 8. 6. 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 – 9. 7. 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 –5. 8. Kamp O, Verhorst PM, Welling RC, Visser CA. Importance of left atrial appendage flow as a predictor of thromboembolic events in patients with atrial fibrillation. Eur Heart J 1999;20:979–85. 9. 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 –9. 10. Handke M, Harloff A, Hetzel A, Olschewski M, Bode C, Geibel A. Left atrial appendage flow velocity as a quantitative surrogate parameter for thromboembolic risk: determinants and relationship to spontaneous echocontrast and thrombus formation –a transesophageal echocardiographic study in 500 patients with cerebral ischemia. J Am Soc Echocardiogr 2005;18:1366 –72. 11. Lang RM, Bierig M, Devereux RB, Flachskampf FA, Foster E, Pellikka PA et al. 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 –63. 12. Hahn RT, Abraham T, Adams MS, Bruce CJ, Glas KE, Lang RM et al. Guidelines for performing a comprehensive transesophageal echocardiographic examination: recommendations from the American society of echocardiography and the society of cardiovascular anesthesiologists. J Am Soc Echocardiogr 2013;26:921 –64. 13. Vincelj J, Sokol I, Jaksic O. Prevalence and clinical significance of left atrial spontaneous echo contrast detected by transesophageal echocardiography. Echocardiography 2002;19:319 –24. 14. Fukushima K, Fukushima N, Ejima K, Kato K, Sato Y, Uematsu S et al. Left atrial appendage flow velocity and time from p-wave onset to tissue doppler-derived a’ predict atrial fibrillation recurrence after radiofrequency catheter ablation. Echocardiography 2014. doi: 10.1111/echo.12823. 15. Lee JM, Shim J, Uhm JS, Kim YJ, Lee HJ, Pak HN et al. Impact of increased orifice size and decreased flow velocity of left atrial appendage on stroke in nonvalvular atrial fibrillation. Am J Cardiol 2014;113:963–9. 16. Veinot JP, Harrity PJ, Gentile F, Khandheria BK, Bailey KR, Eickholt JT et al. Anatomy of the normal left atrial appendage: a quantitative study of age-related changes in 500 autopsy hearts: implications for echocardiographic examination. Circulation 1997; 96:3112 –5. 17. 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 – 82. 18. Petersen M, Roehrich A, Balzer J, Shin DI, Meyer C, Kelm M et al. Left atrial appendage morphology is closely associated with specific echocardiographic flow pattern in patients with atrial fibrillation. Europace 2014. doi:10.1093/europace/euu347. 19. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977;33:159 –74. 20. Patel SV, Flaker G. Is early cardioversion for atrial fibrillation safe in patients with spontaneous echocardiographic contrast? Clin Cardiol 2008;31:148 –52.
Downloaded from by guest on July 11, 2015
LAA orifice areas were found in patients with the cactus type of morphology, consistent with a previous report.6 Although previous studies have reported that LAA orifice size is associated with LAAFV and stroke/TIA in patients with AF,6,15,23 the orifice area was not a significant determinant of LAAFV in the present study. This may be explained by the fact that our study did not include patients with persistent and long-standing AF. First, it was reported that LAAFV is lower in patients with persistent and long-standing AF than in those with PAF.18 The lower LAAFV in patients with persistent and long-standing AF directly affects the relationship between LAAFV and the orifice area. Second, patients with persistent and long-standing AF have higher LAA volumes than those with PAF.24 In this regard, we found a significant relationship between LAA orifice area and volume, even in patients with PAF. It has not been definitely established that a larger orifice area accompanies the greater LAA volume when the basal rhythm changes from paroxysmal to persistent and long-standing AF. However, greater LAA volumes may have a causal effect on the relationship between orifice area and LAAFV. Further studies are required to evaluate whether serial changes in basal rhythm affect the relationship between LAA orifice area and volume in parallel with changes in LAAFV. Although we did not assess the presence of LA structural abnormalities in this study, accessory LAA, left atrial diverticulum, and atrial septal aneurysm (ASA) are well-recognized anatomical variants by recent cardiac imaging technology.25 – 27 Moreover, the presence of these LA structure abnormalities might directly or indirectly lead to formation of the thrombosis and changes in LAAFV. Indeed, previous studies reported that the presence of ASA decreases LA function, which affects LAAFV.25,28 As it remains unclear whether the presence of the accessory LAA and left atrial diverticula also modify LA function and affect LAAFV, further studies are required to confirm this challenging issue.
Page 8 of 8 21. Bernhardt P, Schmidt H, Hammerstingl C, Luderitz B, Omran H. Patients with atrial fibrillation and dense spontaneous echo contrast at high risk a prospective and serial follow-up over 12 months with transesophageal echocardiography and cerebral magnetic resonance imaging. J Am Coll Cardiol 2005;45:1807 –12. 22. Inoue H, Okumura K, Atarashi H, Yamashita T, Origasa H, Kumagai N et al. Target international normalized ratio values for preventing thromboembolic and hemorrhagic events in Japanese patients with non-valvular atrial fibrillation: results of the J-rhythm registry. Circ J 2013;77:2264 –70. 23. Agmon Y, Khandheria BK, Meissner I, Petterson TM, O’Fallon WM, Wiebers DO et al. Are left atrial appendage flow velocities adequate surrogates of global left atrial function? A population-based transthoracic and transesophageal echocardiographic study. J Am Soc Echocardiogr 2002;15:433–40. 24. Park HC, Shin J, Ban JE, Choi JI, Park SW, Kim YH. Left atrial appendage: morphology and function in patients with paroxysmal and persistent atrial fibrillation. Int J Cardiovasc Imaging 2013;29:935 –44.
K. Fukushima et al.
25. Na JO, Shin SY, Lim HE, Choi CU, Kim SH, Kim JW et al. Impaired transport function of the left atrium and left atrial appendage in cryptogenic stroke patients with atrial septal aneurysm and without patent foramen ovale. Eur J Echocardiogr 2011;12: 140 –7. 26. Duerinckx AJ, Vanovermeire O. Accessory appendages of the left atrium as seen during 64-slice coronary CT angiography. Int J Cardiovasc Imaging 2008;24:215 –21. 27. Abbara S, Mundo-Sagardia JA, Hoffmann U, Cury RC. Cardiac CT assessment of left atrial accessory appendages and diverticula. Am J Roentgenol 2009;193:807 –12. 28. Goch A, Banach M, Piotrowski G, Szadkowska I, Goch JH. Echocardiographic evaluation of the left atrium and left atrial appendage function in patients with atrial septum aneurysm: implications for thromboembolic complications. Thorac Cardiovasc Surg 2007;55:365 –70. 29. Sevimli S, Gundogdu F, Arslan S, Aksakal E, Gurlertop HY, Islamoglu Y et al. Strain and strain rate imaging in evaluating left atrial appendage function by transesophageal echocardiography. Echocardiography 2007;24:823 –9.
Downloaded from by guest on July 11, 2015
