2022 年 86 巻 8 号 p. 1263-1272
Background: The left atrial appendage (LAA) is a therapeutic target for preventing cardioembolic stroke in patients with non-valvular atrial fibrillation (NVAF). A large LAA ostium limits percutaneous LAA closure. This study investigated the characteristics and factors associated with a large LAA ostium in Japanese patients with NVAF.
Methods and Results: In 1,102 NVAF patients, the maximum LAA diameter was measured using transesophageal echocardiography (TEE). A large LAA ostium was defined by a maximum diameter of >30 mm. Forty-four participants underwent repeated TEEs, and changes in LAA size under lasting AF were assessed. A large LAA ostium was observed in 3.1% of all participants and 8.9% of patients with long-standing persistent AF (LSAF). The large LAA group had greater CHA2DS2-VASc (P=0.024) and HAS-BLED scores (P=0.046) and a higher prevalence of LAA thrombus (P=0.004) than did the normal LAA group. LSAF, moderate or severe mitral regurgitation, left atrial volume ≥42 mL/m2, E/E’ ratio ≥9.5, and left ventricular mass ≥85 mg/m2 were independently associated with a large LAA ostium (P<0.001, P<0.001, P=0.009, P=0.009, and P=0.032, respectively). In 44 patients with lasting AF, the LAA ostial diameter increased over time (P<0.001).
Conclusions: NVAF patients with a large LAA ostium may have a higher risk of stroke and bleeding. LSAF and factors leading to LA overload may be closely associated with LAA ostial dilatation and can promote it.
Due to global aging, the worldwide prevalence and incidence of non-valvular atrial fibrillation (NVAF) has progressively increased.1 Approximately 90% of non-rheumatic AF-related left atrial (LA) thrombi arise from the left atrial appendage (LAA),2 and strokes caused by LAA thrombus are associated with high mortality and morbidity.3 Because its prevention is mandatory to improve the outcome of patients with NVAF, LAA is considered an important therapeutic target in NVAF.4,5 Catheter-based LAA closure (LAAC) technology was developed to permanently seal off the LAA and prevent cardioembolism. Current available devices for percutaneous LAAC can accommodate >95% of the LAA anatomy;5 however, patients with unsuitable anatomical features of the LAA do not benefit from this procedure. LAA morphology is highly variable, and LAA ostium maximal diameters have been reported to range from 10 to 40 mm.6 Although the required LAA morphology partly varies among the LAAC devices, a large LAA ostium (maximum diameter of >32 mm) is one of the most common characteristics limiting LAAC due to inadequate stabilization and increased risk of device embolization.4,5 Thus, excessive enlargement of the LAA ostium might be an important clinical issue. Clinical trials of WATCHMANTM (Boston Scientific, Natick, MA, USA), the leading device for percutaneous LAAC, indicated that the LAA ostium in Japanese patients might be larger than that in Western patients.7,8 However, characteristics of LAA with large ostia have not been fully investigated in Japanese patients. Accordingly, the present study aimed to examine the prevalence, clinical features, and factors associated with a large LAA ostium in Japanese patients with NVAF.
We retrospectively screened 1,968 consecutive patients who underwent transesophageal echocardiography (TEE) and transthoracic echocardiography following clinical indication at the University of Tsukuba Hospital between January 2017 and December 2018. As the present study focused on potential candidates for percutaneous LAAC,4,9 patients with no history of AF (n=385), moderate or greater mitral valve stenosis (n=17), history of mechanical mitral valve replacement (n=38), and history of percutaneous LAAC or surgical LAA resection/ligation (n=34) were excluded from the analysis. Patients with inadequate TEE image quality or a lack of fundamental images of LAA (n=392) were also excluded. Finally, a total of 1,102 patients with non-valvular AF (defined as AF in the absence of moderate-to-severe mitral stenosis or a mechanical heart valve9) were enrolled in the present study, as illustrated in Figure 1. All patients received adequate anticoagulation therapy when TEE was performed. Of these patients, 44 underwent subsequent TEEs (range, 2–3 times) for LAA thrombus evaluation before the treatment of recurrent persistent AF (n=40) and recurrent stroke under LSAF (n=4). These longitudinal data were used to assess the change in LAA size over time under an AF burden. In the present study, we defined a large LAA ostium to have a maximal diameter of >30 mm for the following reasons. During LAAC, the LAA ostial diameters should be confirmed with appropriate volume-loading conditions (mean LA pressure of >12 mmHg).4 In standard TEEs performed in the echo laboratory, patients undergo examination after fasting for at least 6 h. Therefore, LAA ostial diameters are occasionally underestimated because of the patient being subjected to dehydrating conditions. In contrast, patients can receive intravenous extracellular fluid (approximately 1,000 mL/day) from the day before LAAC to avoid dehydration. Therefore, LAA ostium measurements obtained in the operation room prior to LAAC can be larger than those obtained in the echo laboratory because of the preoperative volume load. In fact, it has been reported that appropriate volume loading before device sizing for percutaneous LAAC commonly increases the LAA ostial dimensions by approximately 2 mm on average.10 Current commercially available percutaneous LAAC devices cannot be used if the LAA ostial diameter is >32 mm.4 For example, the upper limit of the LAA ostial diameter to use each LAAC device safely are as follows: WATCHMAN FLXTM (Boston Scientific) ≤31.5 mm, WaveCrest® (Biosense Webster, Irvine, CA, USA) ≤32 mm, and Amplatzer AmuletTM (Abbott Vascular, Santa Clara, CA, USA) ≤31 mm. Therefore, if the LAA ostium measured by TEE in the examination room under fasting conditions is >30 mm, the true LAA size in the operation room might be occasionally >32 mm and this limits percutaneous LAAC. In real-world practice, if the LAA had a maximum ostial diameter of >30 mm, LAA closure with a single device is considered very challenging.11 In the present study, we investigated the proportion of patients with a large LAA orifice in the overall cohort, each AF type, and in patients with both a CHADS2 score ≥2 and a HAS-BLED score ≥3, fulfilling the indication for percutaneous LAAC in Japan.12
Flowchart of the study population. A total of 1,102 patients, including 34 with a large LAA ostium with a maximum diameter of >30 mm and 1,068 with a normal LAA ostium a with maximum diameter of ≤30 mm, were enrolled. AF, atrial fibrillation; LAA, left atrial appendage; LAAC, left atrial appendage closure; NVAF, non-valvular atrial fibrillation; TEE, transesophageal echocardiography.
AF was defined as paroxysmal (all AF episodes lasting <7 consecutive days), persistent (AF lasting for >7 days, including episodes that are terminated by cardioversion after ≥7 days), or long-standing persistent AF (LSAF; continuous AF lasting for ≥1 year) according to the current guidelines.9 Congestive heart failure was defined as “recent decompensated heart failure (irrespective of the ejection fraction [EF], or presence (even if asymptomatic) of moderate-severe left ventricular [LV] systolic dysfunction on imaging, or hypertrophic cardiomyopathy” based on the current guidelines.9
This study was conducted in accordance with the principles of the Declaration of Helsinki and approved by the Medical Ethics Review Committee of the University of Tsukuba Hospital (R02-105). Informed consent was obtained from all patients using the opt-out method on our department’s website (http://www.md.tsukuba.ac.jp/clinical-med/cardiology/images/research_group/07/clinical_research/kenkyuu103.pdf).
Transthoracic EchocardiographyTransthoracic echocardiography was performed using commercially available ultrasound systems (General Electric Medical Systems, Milwaukee, WI, USA; Siemens Medical Solutions USA, Inc., Malvern, PA, USA; Toshiba Medical Systems, Tochigi, Japan). Each echocardiographic measurement was obtained according to the current guidelines of the American Society of Echocardiography.13,14 Quantification and grading of mitral regurgitation (MR) were performed using an integrated method, as recommended by the American Society of Echocardiography.15 All measurements were averaged over 3 cardiac cycles during the sinus rhythm. In the AF rhythm, an index beat, which is the beat after the almost equal preceding and pre-preceding intervals, was used for each measurement.
Transesophageal EchocardiographyTEE was performed using Vivid E95 machines (GE Vingmed Ultrasound, Horten, Norway) or the EPIQ CVx/X8-2t ultrasound system (Philips Medical Systems, Andover, MA, USA). All patients underwent TEE after a minimum of 6 h of fasting. A topical anesthetic agent was used along with intravenous sedation. Using the 2D TEE mode, gray-scale images during the cardiac cycles at the 0°, 45°, 90°, and 135° planes were recorded to assess the dimensions of the LAA. The line from the left circumflex coronary artery to 1–2 cm beneath the left superior pulmonary vein ridge was chosen as the anatomic landmark to measure the orifice diameter of the LAA.16,17 The maximum and minimum diameters of the LAA ostium were measured from orthogonal planes, and the ratio of both was calculated as an eccentricity index. The depth of the LAA was defined as the distance between the orifice and apex of the LAA. After clearly displaying the LAA, a 3D image of the LAA was obtained from a region of interest. The 3D images were recorded with R-wave gating over 4–6 beats in sinus rhythm, whereas they were recorded in a single beat during AF. Care was taken to minimize the sector width and length to improve the temporal resolution, and the 3D images were recorded at a volume rate of at least 10 images per second. The adequate quality of 3D images was obtained and analyzed for 642 patients. To assess the LAA ostial diameter, the multiplanar reconstruction mode was used to clearly display the cross-section of the LAA orifice, as described previously.17 The largest LAA diameter obtained from both 2D and 3D analyses was chosen for later analysis for the following reasons: 2D TEE underestimates the LAA maximal diameter when the measurement lines do not correctly pass through the center of LAA orifice; 3D TEE underestimates the LAA maximal diameter when the volume rates of 3D images are not sufficient to capture the LAA’s most expanded phase. In order to complement the weaknesses of each method, we adopted the maximum measurements obtained from both 2D and 3D images. LAA emptying and filling flow velocity were measured in the longitudinal view of the LAA using pulse-wave Doppler, and 5 consecutive emptying flow velocities were averaged. LAA thrombi were defined as round, oval, or irregularly shaped masses within the LAA, were distinct from the underlying LAA endocardium and pectinate muscles and were present in multiple imaging planes. The anteroposterior diameter of the mitral annulus and vena contracta area of MR was assessed by multiplanar reconstruction of 3D images, as previously described.15,18
Evaluation of the LAA Morphology Using Multidetector Computed TomographyOf the entire study cohort, 480 patients underwent contrast-enhanced electrocardiogram-gated computed tomography imaging of the LAA using a 320-slice computed tomography scanner (Aquilion ONE Genesis Edition; Canon Medical Systems, Otawara, Japan) and 350 mg iodine/mL iomeprol (Iomeron 350; Bracco Imaging, Milan, Italy) within 1 month prior to TEE. An independent workstation (Ziostation 2; Ziosoft, Tokyo, Japan) was used for the analysis. The number of LAA lobes and LAA morphology (chicken wing, windsock, cauliflower, and cactus) were evaluated using multiplanar reconstruction and 3D volume-rendering images, as previously described.19
Statistical AnalysisContinuous variables are presented as the mean±standard deviation or median with the interquartile range in cases of skewed distribution. Categorical variables are presented as numbers and percentages. Differences between groups were evaluated using the Student’s paired t-test, unpaired t-test, Chi-squared test, Fisher’s exact probability test, or the Mann-Whitney U-test, as appropriate. To compare ≥3 groups, the Kruskal-Wallis and Dunn-Bonferroni post-hoc tests were used. Multivariable linear regression was performed among the variables with P values of <0.10 using univariate correlation analysis. A logistic regression model was used to evaluate the association between clinical variables and an LAA ostium diameter of >30 mm. In advance, receiver operator characteristic curve analysis was performed to determine the best cut-off values for the CHADS2 score, B-type natriuretic peptide (BNP), LV mass index, LA volume index, and E/E’ ratio for predicting a large LAA ostium. The value with the best discriminatory power was determined based on the Youden index. The LA volume index and MR severity were closely associated; therefore, they were divided into different models to avoid multicollinearity in the multivariable analysis. Linear mixed-effect models under the assumption of data missing at random were used to assess the changes in LAA size over time under lasting AF. The reproducibility of LAA measurements by TEE was tested in 40 randomly selected patients. The interobserver agreement in differentiating all 4 LAA types evaluated by computed tomography (CT) was determined using the kappa coefficient. Statistical significance was set at P<0.05. All statistical analyses were performed using SPSS (version 27.0; SPSS Inc., Chicago, IL, USA) and JMP 16 software (SAS Institute, Cary, NC, USA).
The general characteristics of the patients are summarized in Table 1. Of the 1,102 study subjects, 275 (25.0%) were female, with an overall mean age of 65±11 years. According to our definition, 34 patients (3.1%) were classified as having a large LAA ostium. The prevalence of a large LAA in each AF type was 5 (0.8%) with paroxysmal AF, 9 (3.3%) with persistent AF, and 20 (8.9%) with LSAF. In the present study cohort, 124 patients met the indication for LAAC approved in Japan (CHADS2 ≥2 and HAS-BLED ≥3). Of these, 14 patients (11.3%) had a large LAA orifice of >30 mm (Supplementary Figure). Compared with the normal LAA group, the large LAA ostium group had higher CHADS2, CHA2DS2-VASc, and HAS-BLED scores and a higher serum BNP level, with a greater proportion of patients with congestive heart failure, prior stroke/transient ischemic attack/thromboembolism, vascular disease, and long-standing persistent AF. Regarding the parameters of transthoracic echocardiography, greater values of the LV mass index, LA diameter, LA volume index, and E/E’ ratio were observed in the large LAA group. Of the 90 patients with more than moderate MR, 23 had degenerative MR and 67 had ventricular or atrial functional MR. The proportion of patients with more than moderate MR was higher in the large LAA group than in the normal group. With respect to the TEE findings, the large LAA group showed a lower LAA velocity and higher prevalence of LA/LAA spontaneous echo contrast and LAA thrombus than did the normal LAA group despite adequate anticoagulation. The diameter of the LAA ostium obtained from 3D images tended to be greater than that of 2D images; however, the difference was not statistically significant. The large LAA group tended to have a lower eccentricity index, representing a rounder shape of the LAA orifice, than did the normal group. Computed tomography analysis revealed that the large LAA group had a larger number of LAA lobes and a higher prevalence of cauliflower morphology than did the normal group.
| Total (n=1,102) |
Large LAA ostium with a maximum diameter of >30 mm (n=34) |
Normal LAA ostium with a maximum diameter of ≤30 mm (n=1,068) |
P value | |
|---|---|---|---|---|
| Age, years | 65±11 | 68±11 | 65±10 | 0.057 |
| Female | 275 (25.0) | 6 (17.6) | 269 (25.2) | 0.317 |
| Body surface area, m2 | 1.73±0.20 | 1.72±0.16 | 1.73±0.20 | 0.705 |
| Body mass index, kg/m2 | 24.42±3.68 | 23.62±3.23 | 24.43±3.69 | 0.203 |
| CHADS2 score | 1.2±1.1 | 2.0±1.6 | 1.2±1.1 | 0.006 |
| CHA2DS2-VASc score | 2.3±1.6 | 3.1±2.2 | 2.2±1.5 | 0.024 |
| HAS-BLED score | 1.5±1.0 | 2.1±1.6 | 1.5±1.0 | 0.046 |
| Congestive heart failure | 133 (12.1) | 8 (23.5) | 125 (11.7) | 0.043 |
| Hypertension | 684 (62.1) | 19 (55.9) | 665 (62.3) | 0.450 |
| Diabetes | 175 (15.9) | 4 (11.8) | 171 (16.0) | 0.505 |
| History of stroke/TIA/thromboembolism | 110 (10.0) | 14 (41.2) | 96 (9.0) | <0.001 |
| Vascular disease | 141 (12.8) | 10 (29.4) | 131 (12.3) | 0.007 |
| Abnormal renal function | 13 (1.2) | 0 (0) | 13 (1.2) | 0.664 |
| Abnormal liver function | 3 (0.3) | 1 (2.9) | 2 (0.2) | 0.090 |
| History of major bleeding or predisposition to bleeding |
33 (3.0) | 5 (14.7) | 28 (2.6) | <0.001 |
| BNP, pg/mL | 76 [32, 163] | 101 [80, 210] | 71 [31, 163] | 0.016 |
| AF pattern | ||||
| Paroxysmal | 607 (55.1) | 5 (14.7) | 602 (56.4) | <0.001 |
| Persistent | 271 (24.6) | 9 (26.5) | 262 (24.5) | 0.796 |
| Long-standing persistent | 224 (20.3) | 20 (58.8) | 204 (19.1) | <0.001 |
| Duration of persistent AF, years (n=495) | 1.1 [0.4, 3.9] (range, 0.1–27) |
5.0 [0.7, 8.0] (range, 0.2–25) |
1.0 [0.4, 3.1] (range, 0.1–27) |
0.001 |
| Status during echocardiography | ||||
| Normal sinus rhythm | 703 (63.8) | 18 (52.9) | 685 (64.1) | 0.181 |
| Systolic arterial pressure, mmHg | 127±19 | 128±22 | 127±19 | 0.748 |
| Diastolic arterial pressure, mmHg | 74±13 | 73±12 | 74±13 | 0.664 |
| Heart rate, beats/min | 71±18 | 66±15 | 71±18 | 0.068 |
| Transthoracic echocardiography | ||||
| LV ejection fraction, % | 63.5±11.4 | 60.3±14.1 | 63.8±11.6 | 0.173 |
| LV end-diastolic volume, mL/m2 | 56.1±20.2 | 60.6±28.1 | 56.0±19.9 | 0.341 |
| LV end-systolic volume, mL/m2 | 22.0±16.0 | 27.2±22.9 | 21.8±15.7 | 0.184 |
| LV mass index, g/m2 | 90.9±27.1 | 108.8±28.9 | 90.3±26.8 | <0.001 |
| LA diameter, mm/m2 | 39.8±6.9 | 45.2±7.2 | 39.7±6.8 | <0.001 |
| LA volume, mL/m2 | 40.5±16.6 | 60.8±34.1 | 39.8±15.4 | 0.001 |
| Mitral inflow E wave, cm/s | 71.4±22.8 | 81.8±27.2 | 71.0±22.5 | 0.007 |
| E/E’ ratio | 8.5±3.3 | 10.9±4.5 | 8.5±3.3 | 0.006 |
| Moderate or severe MR | 90 (8.2) | 13 (38.2) | 77 (7.2) | <0.001 |
| Transesophageal echocardiography | ||||
| LAA emptying velocity, cm/s | 49.4±22.2 | 35.5±16.7 | 49.8±22.2 | <0.001 |
| LAA filling velocity, cm/s | 50.1±19.7 | 41.5±21.3 | 50.3±19.6 | 0.012 |
| LA/LAA spontaneous echo contrast | 160 (14.5) | 11 (32.4) | 149 (14.0) | 0.006 |
| LAA thrombus | 22 (2.0) | 4 (11.8) | 18 (1.7) | 0.004 |
| 2D parameters | ||||
| Maximum diameter of LAA ostium, mm | 22.5 [19.6, 24.6] | 30.7 [30.3, 35.7] | 22.1 [19.5, 24.3] | <0.001 |
| Minimum diameter of LAA ostium, mm | 17.1 [14.7, 19.2] | 25.0 [23.3, 25.5] | 16.7 [14.6, 19.0] | <0.001 |
| Eccentricity index of LAA ostium | 1.28 [1.20, 1.42] | 1.21 [1.18, 1.40] | 1.29 [1.20, 1.43] | 0.279 |
| Maximal LAA depth, mm | 29.3 [26.6, 32.0] | 32.3 [32.0, 34.6] | 28.9 [26.4, 31.7] | <0.001 |
| 3D parameters (n=642) | (n=22) | (n=620) | ||
| LAA orifice area, cm2 | 1.24 [0.94, 1.55] | 1.73 [1.66, 2.06] | 1.26 [1.17, 1.38] | <0.001 |
| Maximum diameter of the LAA ostium, mm | 22.7 [19.9, 24.9] | 31.0 [30.4, 36.6] | 22.4 [19.6, 24.8] | <0.001 |
| Minimum diameter of the LAA ostium, mm | 17.6 [14.8, 20.0] | 25.3 [23.6, 25.9] | 17.1 [14.7, 19.9] | <0.001 |
| Eccentricity index of the LAA ostium | 1.28 [1.18, 1.41] | 1.23 [1.17, 1.39] | 1.29 [1.20, 1.43] | 0.352 |
| LAA morphology using MDCT (n=480) | (n=19) | (n=461) | ||
| Number of lobes | 1.8±0.8 (range, 1–5) |
2.7±1.0 (range, 1–5) |
1.8±0.8 (range, 1–4) |
<0.001 |
| Chicken wing type | 80 (16.7) | 0 (0) | 80 (17.4) | 0.029 |
| Windsock type | 220 (45.8) | 4 (21.1) | 216 (46.9) | 0.027 |
| Cauliflower type | 152 (31.7) | 15 (78.9) | 137 (29.7) | <0.001 |
| Cactus type | 28 (5.8) | 0 (0) | 28 (6.1) | 0.312 |
Results are shown as the mean±SD, median [interquartile range], or as the number (percentage in the same group). P values indicate the statistical difference between the 2 groups; differences were evaluated using an unpaired t-test, Chi-squared test, Fisher’s exact probability test, or the Mann-Whitney U-test, as appropriate. 2D, two-dimensional; 3D, three-dimensional; AF, atrial fibrillation; BNP, B-type natriuretic peptide; LA, left atrial; LAA, left atrial appendage; LV, left ventricular; MDCT, multidetector computed tomography; MR, mitral regurgitation; TIA, transient ischemic attack.
Table 2 shows the results of the linear regression analysis of the factors associated with the maximum diameter of the LAA ostium. AF type, MR severity, and LA volume index showed higher standardized regression coefficients and were identified as having an independent and closer association with the LAA ostial diameter than did other variables.
| Variables | Univariate analysis | Multivariable analysis | ||||
|---|---|---|---|---|---|---|
| Model including the clinical variables (R2=0.121) |
Model including the echocardiographic variables (R2=0.138) |
|||||
| r | P value | β | P value | β | P value | |
| Age | 0.053 | 0.080 | ||||
| Body surface area | 0.063 | 0.038 | 0.063 | 0.039 | ||
| Body mass index | −0.023 | 0.478 | ||||
| AF pattern* | 0.316 | <0.001 | 0.250 | <0.001 | ||
| Duration of persistent AF (n=489) | 0.234 | <0.001 | ||||
| CHADS2 score | 0.051 | 0.098 | 0.072 | 0.023 | ||
| CHA2DS2-VASc score | 0.002 | 0.958 | ||||
| HAS-BLED score | 0.032 | 0.290 | ||||
| Log BNP | 0.168 | <0.001 | 0.108 | 0.001 | ||
| MR severity† | 0.200 | <0.001 | 0.137 | <0.001 | ||
| LV ejection fraction | −0.115 | <0.001 | ||||
| LV end-diastolic volume | 0.085 | 0.005 | ||||
| LV end-systolic volume | 0.114 | <0.001 | ||||
| LV mass index | 0.160 | <0.001 | 0.107 | 0.009 | ||
| LA diameter | 0.247 | <0.001 | ||||
| LA volume index | 0.309 | <0.001 | 0.292 | <0.001 | ||
| E/E’ ratio | 0.065 | 0.040 | 0.091 | 0.008 | ||
r indicates Pearson’s correlation coefficient or Spearman’s rank correlation coefficient, as appropriate. β and adjusted R2 indicate the standardized regression coefficient and adjusted coefficient of determination, respectively. *AF pattern was entered as a categorical variable, with 5 possible levels (i.e., 0=paroxysmal AF; 1=persistent AF; 2=longstanding persistent AF with a duration of 1–4 years; 3=longstanding persistent AF with a duration of 5–10 years; 4=longstanding persistent AF with a duration of ≥10 years). †MR severity was entered as a categorical variable, with 4 possible levels (i.e., 0=no MR, 1=mild MR, 2=moderate MR, and 3=severe MR). Abbreviations as in Table 1.
Figure 2 shows the distributions of the maximum LAA ostial diameter in the subgroups of patients. A more persistent form, longer duration of AF, and more advanced MR were significantly associated with enlargement of the LAA ostium. Moreover, as LA dilatation and LV hypertrophy progressed, enlargement of the LAA ostium also progressed significantly.
Distributions of the maximal diameter of the left atrial appendage (LAA) ostium. Beeswarm and box plots showing the distributions of the maximum LAA ostial diameter grouped by (A) type and duration of atrial fibrillation (AF), (B) severity of mitral regurgitation (MR), (C) left atrial (LA) volume index tertiles, and (D) left ventricular (LV) mass index tertiles. The boxes represent medians and interquartile ranges, and whiskers represent 95% confidence intervals. The numbers below the whiskers indicate the median values. The P values at the top of the figures indicate statistical differences (evaluated using a Kruskal-Wallis test). The following P values were derived via Dunn-Bonferroni post-hoc tests. *P<0.01 vs. paroxysmal AF (A), no MR (B), and the first tertile (C,D). †P<0.05 vs. persistent AF (A), mild MR (B), and the second tertile (C). ‡P<0.05 vs. long-standing persistent AF with a 1- to 4-year duration.
Table 3 presents the clinical and echocardiographic determinants of the large LAA ostium. Multivariable analysis revealed that a CHADS2 score of ≥3 points, LSAF, moderate or severe MR, LV mass index of ≥85 mg/m2, LA volume index of ≥42 mL/m2, and E/E’ ratio of ≥9.5 were independently associated with a maximum LAA ostial diameter of >30 mm.
| Univariate | Multivariable | |||||
|---|---|---|---|---|---|---|
| Model including the clinical variables |
Model including the echocardiographic variables |
|||||
| OR (95% CI) | P value | OR (95% CI) | P value | OR (95% CI) | P value | |
| Age (per 1-SD increase) | 1.534 (0.988–0.2383) |
0.057 | ||||
| Body surface area (per 1-SD increase) |
0.936 (0.667–1.315) |
0.704 | ||||
| CHADS2 score ≥3 points | 4.756 (2.323–9.739) |
<0.001 | 3.064 (1.405–6.684) |
0.005 | ||
| Longstanding persistent AF | 4.804 (2.408–9.586) |
<0.001 | 5.061 (2.450–10.454) |
<0.001 | ||
| BNP ≥80 pg/mL | 1.652 (0.807–3.380) |
0.170 | ||||
| Moderate or severe MR | 7.967 (3.841–16.525) |
<0.001 | 5.580 (2.543–12.244) |
<0.001 | ||
| LV ejection fraction (per 1-SD increase) |
2.136 (1.054–4.329) |
0.035 | ||||
| LV mass index ≥85 mg/m2 | 5.667 (2.171–14.797) |
<0.001 | 3.374 (1.113–10.229) |
0.032 | ||
| LA volume index ≥42 mL/m2 | 6.624 (2.858–15.353) |
<0.001 | 3.603 (1.376–9.434) |
0.009 | ||
| E/E’ ratio ≥9.5 | 4.720 (2.217–10.052) |
<0.001 | 2.895 (1.310–6.401) |
0.009 | ||
CI, confidence interval; OR, odds ratio; SD, standard deviation. Other abbreviations as in Table 1.
The serial evaluation of LAA ostial diameters in patients with persistent AF who underwent repeated TEE over time is shown in Figure 3. The periods between the first to second and first to third TEE session were 1.4±0.6 and 2.8±0.4 years, respectively. There were gradual increases in the LAA ostial maximum and minimum diameters. Although the eccentricity index tended to decrease, it did not reach statistical significance. In addition, analysis of 3D TEE revealed the anteroposterior diameter of the mitral annulus and vena contracta area of MR were increased in the last TEE compared with the first TEE (first TEE vs. last TEE: 0.20±0.07 cm2 vs. 0.22±0.07 cm2, P=0.004 and 3.5±0.6 vs. 3.6±0.5 cm, P=0.03, respectively), suggesting deterioration of MR due to mitral annuls dilatation.
Longitudinal changes in the echocardiographic left atrial appendage (LAA) ostial size. Changes in the LAA ostial maximum diameter (A), minimum diameter (B), and eccentricity index (C) in patients with persistent atrial fibrillation who underwent repeated transesophageal echocardiography (TEE) are shown. The numbers below the lines indicate the mean±standard deviation at each time point.
Details on the reproducibility of LAA measurements by TEE are provided in the Supplementary File.
To prevent cardioembolic stroke in patients with NVAF, the LAA is the primary therapeutic target, and its size is important when considering percutaneous LAAC. Significant dilatation of the LAA orifice leads to a contraindication to the use of any percutaneous LAAC device because of the risk of device embolization. Compared with other studies, this study included the largest number of patients with NVAF to report the prevalence and clinical characteristics of patients with large LAA ostium. In the present study, we demonstrated that: (1) approximately 3% of the study cohort, 9% of patients with LSAF, and 11% of patients meeting the indication for percutaneous LAAC had an enlarged LAA ostium diameter (>30 mm); (2) patients with LAA ostial dilatation had higher stroke and bleeding risk scores and a higher prevalence of LAA thrombus than did those without; (3) more persistent form and longer duration of AF were significantly associated with a large LAA ostium; (4) in the subgroup with lasting AF, the LAA diameter increased over time; and (5) LSAF, increased LA volume index, high E/E’ ratio, significant MR, and LV hypertrophy were independent determinants of the maximum diameter of the LAA ostium.
Potential Clinical Problems in Patients With a Large LAA OstiumPatients with an LAA ostial diameter >32 mm cannot benefit from the current percutaneous LAAC because it exceeds the upper size limit of any device.4 In the present study, the proportion of patients with a large LAA orifice of >30 mm was as high as 9% in patients with long-standing persistent AF; furthermore, it reached 11% in those with both a CHADS2 score ≥2 and HAS-BLED score ≥3, who met the indication for percutaneous LAAC in Japan.12 These patients might be at risk of not benefiting from percutaneous LAAC due to a large LAA ostium, considering an approximately 2 mm increase in the ostium diameter due to appropriate volume loading.10 We also found that patients with a dilated LAA ostium had higher stroke and bleeding risk scores, slower LAA flow velocity, and a higher prevalence of LAA thrombus. Although these patients need anticoagulants, any anticoagulant carries a bleeding risk and may be contraindicated in certain patients. Percutaneous LAAC is a therapeutic option to address the high risk of both. However, our results suggest that patients with a high need for LAAC are at risk of not being treated with LAAC because of their large ostial size. In current clinical practice, if the LAA ostial size exceeds the limit, there is no way to close the LAA percutaneously, except for off-label attempts.11 Therefore, to avoid losing the chance of proper percutaneous LAAC, understanding the factors that might promote LAA dilatation is important.
Causal Relationship Between Dilatation of the LAA Ostium and Clinical FactorsThe present study showed that LSAF, an increased LA volume index and LV mass index, moderate or greater MR, and a high E/E’ ratio representing LV diastolic dysfunction14 were independent factors associated with LAA ostial dilatation of >30 mm. In addition, through longitudinal observation in a subgroup, we found that long-lasting AF promotes LAA remodeling over time. These results are almost in line with those of previous studies that investigated LAA size.20,21 Although the associations between LAA dilatation and chronic AF, LA dilatation, and LV hypertrophy have been suggested in previous studies,17,22,23 to our knowledge, this is the largest study to examine the independent associations of these clinical factors in patients with AF. Considering our previous data on LAA morphology,24 atrial MR,18 LV diastolic dysfunction, and LA remodeling25 in AF, together with the results of the present study, LAA ostial dilatation might be a compensatory consequence of multiple interactive etiologies (Figure 4). For example, long-lasting AF results in LA structural remodeling, which leads to mitral annulus dilatation and causes MR via worsening of leaflet coaptation.18 A significant MR increases the volume/pressure overload to the LA. Also, LV hypertrophy causes LV diastolic dysfunction and increased LA pressure, both of which may aggravate LA remodeling and contribute to the arrhythmogenic substrate and development and maintenance of AF.25 LAA is more distensible than the LA main chamber, likely because of its elastic properties via myofibrillar orientation or histological characteristics resembling skeletal muscle.26 The high compliance of LAA may be beneficial when the LA pressure and/or volume overload is increased and LA distensibility is decreased.27 Thus, LAA dilatation might be a phenomenon that buffers and compensates for the LA overload in AF patients. In patients with factors related to LAA dilatation, rhythm control via catheter ablation might be a promising therapeutic strategy to reversibly reduce the LAA size, as we previously reported.24
Schematic presentation of the possible course of developing a large left atrial appendage (LAA) ostium. The schematic illustration indicates the possible mechanism for the development of a large LAA ostium via atrial fibrillation (AF) and the interactions between factors (yellow boxes) determined in the present study. The black arrows indicate the plausible relationships. White lines and numbers in the transesophageal echocardiography image indicate the measurement site and the diameter of the LAA ostium.
LAA has a wide variety of individual morphologies. It has been reported that the size of the LAA orifice varies considerably.6 In addition to individual differences, several studies have indicated that there may be racial differences in terms of LAA morphology. In the PREVAIL trial, which is a clinical trial of WATCHMAN devices conducted in the United States, most participants were treated with the 24- or 27-mm device.8 In contrast, the 30-mm device was implanted most frequently in the SALUTE trial conducted in Japan.7 Compared with non-Asians, Asians, including Japanese, have a larger LAA size7,28,29 and more complicated LAA morphology with many lobes and trabeculation.28,30 In the present study, which included Japanese patients, the proportion of large LAA with complicated cauliflower-type structures was higher than that reported in previous studies conducted in non-Asian countries.19,31 These results suggest that not only individual, but also racial, differences influence the LAA ostial size. Thus, racial differences in LAA morphology and size should be considered when developing future devices. However, further investigations are required.
Clinical ImplicationsPatients with LSAF with LA dilatation, LV hypertrophy, LV diastolic dysfunction, and progressed MR have a higher risk of not benefiting from percutaneous LAAC. Care should be taken to identify the optimal timing of LAAC, and early intervention should be considered in patients with multiple risk factors related to LAA dilatation. Elimination of AF by catheter ablation might also be an important option to prevent LAA dilatation or reversibly reduce the size of the LAA.24 In addition, our findings suggest that, in patients with LSAF and factors associated with large LAA, careful sizing of the LAA using preoperative imaging is particularly important and should be performed to determine whether LAAC is possible with a commercially available device.
Study LimitationsFirst, the participants mainly consisted of patients undergoing catheter ablation or electrical cardioversion for AF in a single center; therefore, whether the results are applicable to a general population with NVAF is unknown. Second, this is a retrospective case-control study; therefore, the causal relationship between a large LAA ostium and clinical factors leading to LAA overload cannot be fully clarified. In addition, there might be a potential selection bias. To overcome these limitations, a multicenter prospective study with a larger number of participants may be required. Third, 3D images of LAA were recorded in a single beat if AF continued during TEE; this might have affected the LAA measurements because of lower frame rates; however, the methodology of 3D image analysis was in line with that in a previous study that investigated the LAA ostial size in patients with AF.17 Furthermore, the differences in vendor equipment might affect the echocardiographic results of this study. Because this is a retrospective observational study, no echocardiographic data obtained by different vendors simultaneously in the same patients were available; therefore, the validation across different vendors and software for each measurement could not be performed. Future investigations might be necessary to address this point.
Among Japanese patients with NVAF, those with a large LAA ostium had high risk scores for both stroke and bleeding, and they might be potential candidates for percutaneous LAAC. Although further investigations are required to clarify the exact causal relationship, LSAF and multiple factors leading to LA overload may be closely associated with LAA ostial dilatation and can promote it. Attention should be paid to the treatment and follow up of patients with LSAF to avoid missing the opportunity for percutaneous LAAC.
This work was supported by JSPS KAKENHI (grant number: 19K17586).
T.M.-O. had endowed departments by Boston Scientific Japan, Japan Lifeline Co., Ltd., Nihon Cohden Corporation, Biotronik Japan, Inc., Toray Industries, Inc., Abbott Vascular Japan Co., Ltd., and ASTEC Co., Ltd. All other authors declare that that they have no conflicts of interest.
This study was approved by the Medical Ethics Review Committee of the University of Tsukuba Hospital (R02-105), and informed consent was obtained from all patients using the opt-out method on our Department’s website (http://www.md.tsukuba.ac.jp/clinical-med/cardiology/images/research_group/07/clinical_research/kenkyuu103.pdf).
The deidentified participant data will not be shared.
Please find supplementary file(s);
http://dx.doi.org/10.1253/circj.CJ-22-0053
