1 INTRODUCTION

Left atrial appendage (LAA) morphology can be determined by cardiac computed tomography angiography (CCTA). However, the role of the LAA in the initiation and perpetuation of AF has to be defined further. Pulmonary vein electrical isolation (PVI) is the standard strategy at the initial ablation procedure for the treatment of paroxysmal and persistent atrial fibrillation (AF).1 Despite continuing technical improvements,2 a significant proportion of patients have recurrence after the first ablation procedure.3 Several mechanisms causing recurrence have been identified. Electrical reconnection of the PVs is the main mechanism, and gap closure in the second procedure remains the standard of care.1, 4, 5 With the introduction of next generation ablation catheters and smart ablation protocols, the majority of PVs show durable isolation.4, 6, 7 Additional lesions beyond PV isolation (PVI) (e.g., empirical lines, complex fractionated electrograms, ganglionated plexi, triggers) have been studied, but until now, these measures have not demonstrated additional benefit as compared to PVI alone in the initial ablation procedure.8 Clinical predictors of AF recurrence after catheter ablation are LA size, AF type, female sex, and in-hospital AF relapse, as well as comorbidities such as impaired cardiac and renal function.9-11 LAA itself might also be a source of extra-PV AF triggers12 or might serve as the substrate of perpetuating AF, though it is unknown whether the LAA anatomy correlates with the recurrence rate after PVI. LAA is a complex anatomical structure with substantial variation in size and morphology; it is not possible to determine LAA morphology or LAA size through transthoracic echocardiography (TTE). However, CCTA is the best modality to determine LAA morphology and atrial dimensions.13 Other benefits from preprocedural cardiac computed tomography angiography (CCTA) include ruling out possible thrombi, detecting underlying coronary artery disease (CAD), and gaining advanced information of individual PV anatomy.1, 14

The present study sought to investigate whether LAA morphology and detailed measurement of a variety of LAA and LA parameters, determined by CCTA, can predict atrial arrhythmia recurrence after initial cryoballoon PVI in symptomatic AF patients.

2 METHODS 2.1 Study design

This prospective single-centre registry study enrolled consecutive patients undergoing second-generation CBA between May 2012 and September 2016. Informed consent was obtained from all patients. The study was approved by the regional ethics review board and was conducted in accordance with the Declaration of Helsinki.

2.2 Study participants

Consecutive symptomatic patients scheduled for the initial AF ablation procedure aiming at PV isolation were treated with the second-generation cryoballoon (Arctic Front Advance™, Medtronic Inc., MN, USA) and were prospectively enrolled into the institutional observational registry. If a recent preprocedural CCTA of sufficient quality to assess the LAA anatomy was available, the patient was considered for the present blinded analysis. The clinical indications for CCTA were, firstly, exclusion of CAD, and secondly, determination of LA and PV anatomy prior to the CBA. No patient was excluded from cryoballoon ablation based on variations in LA, PV, LAA anatomy, or LA volume, as determined by CCTA. None of the consecutive patients were treated by means of point-by-point radiofrequency (RF) ablation for PVI. Standard exclusion criteria for AF ablation were applied.15 Baseline characteristics were collected prospectively.

2.3 Objectives and endpoints

The primary objective of the study was to evaluate the impact of LAA anatomy and morphology on the recurrence of AF after CBA. The secondary objective was the general determination of possible independent clinical risk factors for AF recurrence.

2.4 Preprocedural investigations

Prior to the ablation procedure, all patients underwent transthoracic and transesophageal echocardiography to exclude possible LA thrombus formation. Additionally, individual left atrial anatomy was revealed using a 64-slice CT scanner (Brilliance 64, Philips Medical Systems, Cleveland, OH, USA) with retrospective electrocardiography (ECG) gating and 3D reconstruction prior to the procedure. Scanning was performed at 120 kVp, with an effective tube current of 600 mAs. The slice collimation was 64 × 0.625 mm, with a gantry rotation time of 0.4 s and a pitch of 0.2. Images were reconstructed at 0.9 mm slice thickness at increments of 0.45 mm. Contrast enhancement with 80 ml of contrast agent (Imeron 400 MCT, Iomeprol 81.65 g/100 ml, Bracco, Konstanz, Germany) was injected at a flow rate of 5 ml/s and followed by a 50-ml saline flush. A blinded observer (JP) analyzed, segmented, and measured each CCTA image visually and quantitatively with regard to the defined parameters and the morphological classification of the LAA using EnSite Precision™ (Abbott Medical GmbH, Eschborn, Germany). Multiplan volume-rendered post-processing was used to acquire a 3D perspective. After anatomical segmentation into PV, LA, and LAA, all 2-dimensional (2D) and 3D measurements were conducted.

2.5 Data acquisition, management and quality control

The main 2D parameters included the maximal length, height, and depth of the LAA. The volume of the LAA and LA was computed automatically after anatomical segmentation. Major parameters analyzed three-dimensionally were the LAA length, roof top and bottom lines, distance from the mitral valve annulus to the middle of the LA roof, depth of the LA, and septum orifice distance (see Figure S3). Major subsequent computed parameters were the ellipsoidal area of the LAA ostium, perimeter calculated by Ramanujan's formula, and trapezoid area of the posterior wall calculated by the roof top line, and roof and bottom lines and the distance between those lines. Based on the criteria established by Wang et al. and Kimura et al., the LAA was classified into one of four types: windsock, chicken-wing, cactus, and cauliflower.16, 17 After initial classification, there was a reassessment of all images among a team of four physicians for objective validation. CBA was performed as previously described15, 18 (Appendix).

All ablation procedures were performed by experienced interventional electrophysiologists (EH, UD, FS, MW). Following the intervention, all patients were monitored with continuous ECG for at least 24–48 h. TTE was performed to exclude pericardial effusion. In the event of symptoms, additional ECG and Holter studies were continued for up to 7 days. Holter recordings after 1, 3, 6, and at least 12 months were organized to screen for symptomatic or asymptomatic atrial arrhythmias. Follow-up was also ensured in cooperation with referring physicians, and detailed questionnaires (15 questions) were administered for each case via mail and telephone calls. If there was any suspicion of recurrence, the referring physician was contacted to validate the diagnosis. Only recurrences outside of the 90-day blanking period were categorized as failures.

2.6 Statistical analysis

Statistical analysis was performed using Microsoft Excel 2016 (Microsoft Corp., Redmond, WA, USA) and SPSS version 25 (IBM Corp., Armonk, NY, USA). Categorical variables are reported as numbers and percentages. In accordance with the Shapiro–Wilk test, continuous variables are expressed as means with standard deviations (SD) or as medians with quartiles. To assess the relationship between LAA parameters and the recurrence of AF after CBA, univariate Cox regressions and Kaplan–Meier plots with log-rank tests were used. To avoid problems regarding multicollinearity in multivariate models Pearson's and Spearman's correlation coefficients were computed to identify intervariable relationships.Main independent risk factors were determined using a stepwise multivariate Cox regression model with bidirectional elimination including only parameters of highest univariate significance. Parameters highly correlated with these main features were eliminated. Receiver operating characteristic (ROC) analysis with the corresponding area under the curve (AUC) was performed to determine the specific cut-off values. Statistical significance was defined as p ≤ 0.05. All data were analyzed using the SPSS version 25.

3 RESULTS 3.1 Study population and procedural results

From May 2012 to September 2016, 1103 patients underwent PVI for symptomatic paroxysmal or persistent AF. CCTA was performed in 725 (65.7%) patients, and 473 (42.9%) patients had sufficient image quality for LAA measurements. Figure S1 represents a flow diagram of the study. The mean age of patients who underwent PVI was 66.2 ± 9.5 years, of whom, 189 (40%) were females. The majority of patients (58.6%) had symptomatic paroxysmal AF. Table 1 illustrates the baseline characteristics of the study population in terms of recurrence and non-recurrence.

TABLE 1. Baseline characteristics All patients With recurrence Without recurrence p-value Value Analyzed Value Analyzed Value Analyzed Age, years 66.25 ± 9.47 473 67.36 ± 9.28 166 65.65 ± 9.53 307 0.05 Females (%) 189 (40.0) 473 74 (15.6) 166 115 (24.3) 307 0.08 Persistent AF (%) 196 (41.4) 473 82 (17.3) 166 114 (24.1) 307 0.006 Mitral regurgitation ≥ °II 18 (3.8) 473 10 (2.1) 166 8 (1.7) 307 0.005 Valve disease ≥ °II 25 (5.3) 473 13 (2.7) 166 12 (2.5) 307 0.007 LA diametera mm 43 [40; 47] 440 45 [41; 50] 159 42 [39; 47] 281 0.002 Ejection fractiona (%) 56.88 ± 6.1 459 56.39 ± 5.64 162 57.14 ± 6.32 297 0.72 Hypertension 309 (66.2) 467 112 (24.0) 164 197 (42.1) 303 0.51 Hypertensive heart diseasea 103 (21.8) 467 43 (9.2) 164 60 (12.8) 303 0.10 Cardiomyopathy 16 (3.4) 466 3 (0.6) 163 13 (2.8) 303 0.2 Prior myocardial infarction 8 (1.7) 465 3 (0.6) 162 5 (1.1) 303 0.98 Structural heart disease 170 (35.9) 473 69 (14.6) 166 101 (21.3) 307 0.042

Overweight

(BMI > 25)

233 (61.6) 378 80 (21.2) 137 153 (40.4) 241 0.21 Obesity (BMI > 30) 67 (17.7) 378 21 (5.6) 137 46 (12.2) 241 0.21

Obesity °II/III

(BMI > 35)

20 (5.3) 378 9 (2.4) 137 11 (3.0) 241 0.64 Left common ostiumb 101 (21.4) 473 37 (7.8) 166 64 (13.5) 307 0.81 Accessory veinsb 83 (17.5) 473 37 (7.8) 166 46 (9.7) 307 0.053 Note: n (%), mean ± SD, or median (IQR). Abbreviations: AF, atrial fibrillation; BMI, body mass index; IQR, interquartile range; LA, left atrium; SD, standard deviation. a Determined by transthoracic echocardiography. b Determined by contrast enhanced cardiac computer tomography.

All patients underwent CBA, and all PVs were successfully isolated with the cryoballoon. No additional RF or Cryo-Tip touch-up ablations were performed. The median procedural time was 130 (110; 155) min with an LA time of 90 (75; 110) min. The fluoroscopy time was 22 (17; 27) min, and the median dose area product was 1829 (1044; 3099) cGycm2. Of all 473 patients undergoing CBA, 166 (35.1%) experienced AF recurrence during follow-up. The median follow-up time was 19 months.

3.2 LAA morphology and outcome

The distribution of the LAA morphological types is shown in Figure 1. Among the 166 recurrence events, chicken-wing morphology had the highest chance of recurrence at 37.8%, followed by windsock at 36.5%, cauliflower at 32.0%, and cactus at 28.8%. None of these categories was found to have a statistically significant impact on the AF recurrence rate (log-rank; p = 0.596). The corresponding Kaplan–Meier plot is illustrated in Figure 2A.

Classification of LAA morphology. This figure depicts the LAA morphological types as defined by Wang [22] with Kimura's quantitative qualifiers [23]. According to the measured LAA length and number of lobes, LAA morphology was classified into one of four types: windsock, chicken-wing, cactus, and cauliflower. The prevalent LAA type distribution in the study is provided here. LAA, left atrial appendage

Outcome of cryoballoon ablation related to clinical predictors and LA/LAA anatomy. Kaplan–Meier curves demonstrate freedom of AF after cryoballoon ablation according to (A): LAA morphology, which showed no statistical impact on the recurrence rate of AF after cryoballoon ablation (log-rank; p = 0.596). (B): LA volume with a cut-off level of ≥122.7 ml: Larger LA volumes demonstrated a highly significant impact on the recurrence rate of AF after cryoballoon ablation (p < 0.001). (C): LAA volume cut-off level of ≥11.25 ml: Larger LAA volumes demonstrated a highly significant impact on the recurrence rate of AF after cryoballoon ablation (p < 0.001). (D): mitral valve regurgitation ≥°II. Its presence was significantly related to a higher recurrence rate (log-rank; p = 0.003). (E): AF-type: Persistent type AF demonstrated a significant impact on the recurrence rate of AF after cryoballoon ablation (p = 0.001). (F): female sex on its own showed a non-significant trend on the recurrence rate of AF after cryoballoon ablation (p = 0.07). AF, atrial fibrillation; LAA, left atrial appendage; LA, left atrium

3.3 LAA volume and the correlation to LA volume

LAA and LA volumes showed statistically significant correlations. We conducted linear regression to quantify the correlation of the LA volume with the LAA volume, which demonstrated that the LAA volume increased by 0.70 ml per 10 ml increase in LA volume (p < 0.001, see Figure S2).

3.4 LAA morphology

Chicken-wing morphology had the largest LAA volume with 9.9 [7.98; 12.83] ml, followed by windsock morphology with an LAA volume of 9.65 [7.73; 13.08] ml. Cactus and cauliflower morphologies were smaller, with LAA volumes of 5.4 [4.6; 7.5] ml and 5.6 [4.4; 7.6] ml, respectively (p < 0.001).

3.5 Outcome predictors: Univariate analysis

In univariate Cox regression models, four continuous parameters were significant predictors of AF recurrence, with LA volume in ml demonstrating the highest predictive value (hazard ratio [HR] 1.01; 95% confidence interval [CI] [1.006–1.015]; p < 0.001). The LA volume determined by CCTA seemed to be superior compared to the LA diameter determined by transthoracic echocardiography and showed a higher significance (HR 1.037; 95% CI [1.015–1.060]; p = 0.001). The second most important parameter was the septum orifice distance (HR 1.053; 95% CI [1.028–1.08]; p < 0.001), followed by the trapezoid area of the posterior LA wall (HR 1.001; 95% CI [1–1.001]; p < 0.001) and the LAA volume (HR 1.051; 95% CI [1.025–1.078]; p < 0.001). All significant results are listed in Table 2 and Table S1.

TABLE 2. Univariate analysis of baseline characteristics and measurement data Univariate analysis HR 95% CI p-value Baseline characteristics Female sex 1.31 0.97–1.78 0.08 Age, years 1.02 1.000–1.034 0.05 Persistent AF 1.54 1.140–2.090 0.006 Mitral regurgitation ≥ II 2.5 1.31–4.75 0.005 Structural heart disease 1.38 1.01–1.88 0.04 LA diametera, mm 1.037 1.015–1.060 0.002 LAA morphology Chicken-wing 1.13 0.781–1.624 0.53 Windsock 1.09 0.802–1.476 0.59 Cactus 0.73 0.440–1.201 0.21 Cauliflower 0.94 0.604–1.459 0.78 Significant LAA measurements LAA max width, mm 1.03 1.003–1.049 0.03 LAA maximal depth, mm 1.02 1.000–1.036 0.05 LAA volume, ml 1.05 1.025–1.078 <0.0001 LAA Dmax 3D, mm 1.06 1.021–1.107 0.003 LAA Dmin 3D, mm 1.02 1.006–1.032 0.004 LAA Length, mm 1.02 1.002–1.041 0.03 Perimeter LAA ostium, mm 1.04 1.008–1.078 0.02 Area LAA ostium, mm2 1.20 1.100–1.300 0.001 Significant LA measurements LA volume, 10 ml 1.100 1.060–1.150 <0.000001 Roof top line, mm 1.03 1.013–1.050 0.001 Roof bottom line, mm 1.03 1.012–1.054 0.002 Posterior wall box height, mm 1.03 1.003–1.065 0.03 Distance MVA-LA roof, mm 1.04 1.014–1.064 0.002 Depth of the LA, mm 1.04 1.013–1.067 0.003 Septum orifice distance, mm 1.05 1.028–1.080 <0.0001 Perimeter LIPV ostium, mm 1.02 1.001–1.042 0.04 Trapezoid area of the posterior LA wall, cm2 1.08 1.040–1.100 <0.0001 Note: This table illustrates the results of the univariate analysis of all important baseline characteristics and CCTA derived data. Abbreviations: AF, atrial fibrillation; CCTA, cardiac computed tomography angiography; CI, confidence interval; Dmax, maximal ostial diameter; Dmin, minimal ostial diameter; HR, hazard ratio; LA, left atrium; LAA, left atrial appendage; LIPV, left inferior pulmonary vein; MVA, mitral valve annulus. a Measured by transthoracic echocardiography. 3.6 Outcome predictors: Correlation analysis

To avoid problems of multicollinearity in multivariate models, correlation analysis was performed for LA volume and all baseline parameters and measurement data according to the Pearson and Spearman tests (see Table S2, Table S3). Among the baseline parameters, mitral regurgitation, AF type, sex, age, structural heart disease, and hypertensive heart disease were significantly correlated to LA volume (the correlation matrix is provided in the Table S2). All CCTA results that related logically to the LA volume, such as the septum orifice distance, depth of the LA, distance of the mitral valve annulus to the LA roof, and trapezoid area of the posterior left atrial wall, showed significant positive correlations, as did the LAA volume and its companion parameters (e.g., perimeter of the LAA ostium or area of the LAA ostium). Details are provided in the Table S3. LA volume was the best CCTA-derived measurement parameter for predicting AF recurrence and was, therefore, included in multivariate regression analysis.

3.7 Outcome predictors: Multivariate analysis

After precise analysis of intervariable correlation, a stepwise multivariate Cox-regression model that included all significant baseline parameters and continuous CCTA parameters was performed. The independent parameters we evaluated were LA volume (HR 1.012; 95% CI [1.008–1.016]; p < 0.001, mitral regurgitation ≥°II (HR, 2.27; 95% CI [1.189 to 4.333]; p = 0.013), and female sex (HR 1.648; 95% CI, 1.196 to 2.271; p = 0.002). The exact values are presented in Table 3. All independent risk factors were evaluated using Kaplan–Meier Survival curves for AF recurrence during follow-up (see Figure 2).

TABLE 3. Evaluation of independent risk factors Univariate analysis Multivariate analysis HR 95% CI p-value HR 95% CI p-value Baseline characteristics Female sex 1.31 0.97–1.78 0.08 1.648 1.196–2.271 0.002 Age, years 1.02 1.000–1.034 0.05 Persistent AF 1.54 1.140–2.090 0.006 Mitral regurgitation ≥°II 2.5 1.31–4.75 0.005 2.270 1.189–4.33 0.013 Structural heart disease 1.38 1.01–1.88 0.04 LA diametera, mm 1.037 1.015–1.060 0.002 CCTA measurements LA volume, 10 ml 1.100 1.060–1.150 <0.001 1.012 1.008–1.016 <0.001 Abbreviations: AF, atrial fibrillation; CCTA, cardiac computed tomography angiography; CI, confidence interval; HR, hazard ratio; LA, left atrium. a Determined by transthoracic echocardiography. b