Impact of left atrial geometric remodeling on late atrial fibrillation recurrence after catheter ablation

Aims To quantitatively investigate the impact of left atrial geometric remodeling on atrial fibrillation recurrence after catheter ablation. Methods A retrospective analysis of 105 patients with atrial fibrillation who underwent coronary computed tomographic angiography before catheter ablation. Risk factors for atrial fibrillation recurrence were identified by multivariable logistic regression analysis and used to create a nomogram. Results After at least 12 months of follow-up, 30 patients (29%) developed recurrent atrial fibrillation. Patients with recurrence had higher left atrial volume, left atrial sphericity, and lower left atrial ejection fraction (LAEF) (P < 0.05). There was no significant difference in asymmetry index between the two groups (P = 0.121). Multivariable regression analysis showed that left atrial minimal volume index (LAVImin) [odds ratio (OR): 1.026, 95% confidence interval (CI): 1.002–1.050, P = 0.034], left atrial sphericity (OR: 1.222, 95% CI: 1.040–1.435, P = 0.015) and CHADS2 score (OR: 1.511, 95% CI: 1.024–2.229, P = 0.038) were independent predictors of atrial fibrillation recurrence. The combined model of the left atrial sphericity to the LAVImin substantially increased the predictive power for atrial fibrillation recurrence [area under the curve (AUC) = 0.736, 95% CI: 0.627–0.844, P < 0.001], with a sensitivity of 80% and a specificity of 61%. A nomogram was generated based on the contribution weights of the risk factors; the AUC was 0.772 (95% CI: 0.670–0.875) and had good internal validity. Conclusion The CHADS2 score, left atrial sphericity, and LAVImin were significant and independent predictors of atrial fibrillation recurrence after catheter ablation. Furthermore, the nomogram had a better predictive capacity for atrial fibrillation recurrence.


Introduction
Atrial fibrillation is associated with a high risk of thromboembolic stroke, heart failure, and premature death. It is the commonest cardiac arrhythmia and is also associated with high cost burdens on healthcare systems. 1 Although catheter ablation is a widely accepted treatment for drugrefractory atrial fibrillation with a success rate of 50-80%, 2,3 atrial fibrillation recurrence remains an important issue. Beyond pulmonary vein-associated triggers, it has been suggested that the underlying mechanisms instigating atrial fibrillation recurrence are complex. Various imaging modalities, including echocardiography, computed tomography (CT), and cardiac MRI (CMRI), have been used to demonstrate the important role of left atrial geometry in assessing atrial fibrillation incidence. 4,5 Recent research has focused on precise assessments of the left atrial shape, which may prove to be a more precise marker than size alone. 6 This study aimed to investigate whether left atrial geometric remodeling has a prognostic impact on atrial fibrillation recurrence after catheter ablation and to develop a precise predictive nomogram.

Study population
This study retrospectively analyzed 105 consecutive atrial fibrillation patients who underwent CT angiography (CTA) before atrial fibrillation ablation in the Department of Medical Imaging at the Second Hospital of Hebei Medical University, Shijiazhuang, China, between October 2017 and September 2018. According to the 2014 AHA/ACC/HRS guidelines, 7 paroxysmal atrial fibrillation was self-limiting within 7 days of onset, and persistent atrial fibrillation lasted longer than 7 days or required cardioversion. Patients were eligible for inclusion if they had documented symptomatic atrial fibrillation and no previous ablation treatment. The exclusion criteria were valvular heart disease, cardiomyopathy, congenital cardiac abnormalities, and the presence of left atrial thrombus. The institutional review board of the Second Hospital of Hebei Medical University granted ethical approval. Written informed consent for participation was not required for this study in accordance with national legislation and institutional requirements.

Image acquisition
All images were generated using a 256-slice CT scanner (Brilliance iCT, Philips Healthcare, Cleveland, Ohio, USA) and retrospective electrocardiography (ECG)gated spiral data acquisition. Metoprolol (50-100 mg) was given 30-60 min before CCTA to patients with a mean heart rate at least 65 beats/min. The scan parameters were as follows: tube voltage, 120 kV; automatic tube current, 280-350 mA; detector collimation, 128 Â 0.625 mm; slice thickness, 0.625 mm; section interval, 0.33 mm; gantry rotation time, 330 ms; beam pitch, 0.18. The contrast material of iohexol (Omnipaque 350; 1.0 ml/kg) was injected into the ulnar vein at the rate of 4-5 ml/s.

Image analysis
Retrospective ECG-gated reconstruction generated 10 phases with an interval of 10% over an R-R interval at 5-95%, and then image data were transferred to the Philips EBW4.5 workstation for postprocessing. Three-dimensional images of the left atrium and pulmonary vein were automatically or semi-automatically segmented with the workstation. The left atrial volume (LAV) was obtained after the atrial appendage and the pulmonary veins were excluded. Left atrial maximum volume (LAVmax), left atrial minimum volume (LAVmin), and left atrial appendage volume (LAAV) were automatically calculated. Body surface area (BSA) correction was used to determine left atrial maximum volume index (LAVImax) and left atrial minimum volume index (LAVImin). Left atrial ejection fraction (LAEF) and left ventricular ejection fraction (LVEF) were calculated using cardiac function software. According to guidance by Wadell, 8 the maximal left atrial diameter was assessed on the rendered image, assuming that the volume of the sphere was the same as that of the left atrial, and then the diameter of the corresponding sphere was calculated. The left atrial sphericity was the ratio of the diameter of the corresponding sphere to the maximum left atrial diameter (Fig. 1a). Next, LAV was divided into anterior (LA-A) and posterior (LA-P) volumes using a section parallel to the posterior wall between the pulmonary vein ostia and the LAA. The asymmetry index (ASI) was the ratio of LA-A to LAV (Fig. 1b). 9 All the parameters were observed and measured from multiple angles by two radiologists after relevant training.

Ablation procedure
According to the uniform standard of circumferential isolation of the pulmonary vein, catheter ablation was performed using a 3.5 mm tip irrigated ablation catheter (Navistar Thermocool, Biosense-Webster, Diamond Bar, catheter ablation, USA) placed at the ostia of the pulmonary veins to record pulmonary vein potentials. The left atrial geometry reconstruction was guided by electroanatomic mapping (CARTO-3 system, Biosense-Webster, Diamond Bar, catheter ablation, USA). The bidirectional conduction block from the atrium to the pulmonary veins  was the endpoint of the pulmonary vein isolation, as confirmed using a Lasso catheter (Biosense-Webster, Diamond Bar, catheter ablation, USA). If atrial fibrillation was not terminated, additional linear ablation and complex fractionated atrial electrogram ablation were performed.

Follow-up
Follow-up was performed at 3, 6, and 12 months by telephone interviews with patients and/or family members regarding the condition of the heart. Atrial fibrillation recurrence, defined as any episode greater than 30 s identified by 12-lead ECG or Holter after a 3-month blank period, signified the end of follow-up.

Statistical analysis
Normally distributed continuous variables are expressed as means AE standard deviations, and skewed data are expressed as medians with interquartile ranges. Categorical variables are expressed as frequencies and percentages. Chi-square analysis was used for categorical variables, and Student's t-test or the Mann-Whitney U-test was used for continuous variables.
Multivariable logistic regression analysis with a backward stepwise procedure were performed to evaluate risk factors for atrial fibrillation recurrence. The model included variance inflation factor (VIF) calculations to assess collinearity. The criteria used for variables in the model were P less than 0.1 for inclusion and P greater than 0.2 for removal, and odds ratio (OR) and 95% confidence interval (CI) calculations were performed. Receiver-operating characteristic (ROC) curve analysis determined the optimal cut-off values with the highest sensitivity and specificity of left atrial structure parameters for predicting atrial fibrillation recurrence.
The multivariable logistic regression analysis informed the construction of a nomogram. The discriminating ability of the nomogram was evaluated using concordance index calculation and calibration plots.
All statistical analyses were performed using SPSS Statistics for Windows, version 24 (IBM Corp., Armonk, New York, USA) and R, version 4.0.3 (R Foundation for Statistical Computing, Vienna, Austria). All P values were two-sided, and P values less than 0.05 were considered statistically significant.

Study sample
The baseline patient characteristics are summarized in Table 1. We enrolled 105 patients, including 66 men (63%), with a mean age of 59.8 AE 10.5 years. After at least 12 months of follow-up, 30 patients (29%) experienced recurrent atrial fibrillation. There were no significant differences between the patients who experienced atrial fibrillation recurrence (recurrence group) and those who did not (nonrecurrence group) in terms of age, sex, atrial fibrillation duration, BMI, estimated glomerular filtration rate, hypertension status, diabetes mellitus status, vascular disease status, coronary artery disease status, or history of stroke/transient ischemic attack. There was also no significant intergroup difference in terms of medication intake. However, the recurrence group had significantly higher proportions of patients with heart failure (P ¼ 0.013) and persistent atrial fibrillation (P ¼ 0.045).
Impact of left atrial geometric remodeling Guo et al. 911 The recurrence group also had significantly higher CHA 2 DS 2 -VASc and CHADS 2 scores (P ¼ 0.009 and 0.006, respectively) than the nonrecurrence group.
Univariable analysis and multivariable logistic analyses of clinical and CTA parameters for recurrent atrial fibrillation are summarized in  (Table 4).
In the nomogram (Fig. 2) based on the contribution weights of the selected variables in the regression model, each predictive value was assigned a point on the corresponding variables line and added to form a scale from 0 to 100% to find out the corresponding risk of atrial fibrillation recurrence. Left atrial sphericity was the strongest predictor of recurrence after ablation, followed by LAVImin and CHADS 2 score. The nomogram model could accurately predict atrial fibrillation recurrence with an AUC of 0.772 (95% CI: 0.670-0.875) (Fig. 3) and had good internal validity (Fig. 4).

Discussion
We investigated the impact of left atrial structural remodeling on outcomes of catheter ablation for atrial fibrillation. The main study findings were as follows: LAVImin, left atrial sphericity, and CHADS 2 score were independent predictors of atrial fibrillation recurrence after catheter ablation; LAVImin greater than 46.4 ml/m 2 and left atrial sphericity greater than 81% were associated with a high risk of atrial fibrillation recurrence; the integration of left atrial sphericity data with the LAVImin data provided additional prognostic information regarding atrial fibrillation recurrence; we devised a nomogram combining left atrial sphericity, LAVImin, and CHADS 2 score data to predict the risk of atrial fibrillation recurrence. Our findings may help improve atrial fibrillation therapy by facilitating more appropriate and earlier patient selection.
Atrial remodeling, known as structural and functional remodeling, is a response to increased pressure and volume overload, resulting in changes in the myocardium (myocyte hypertrophy and interstitial fibrosis). A more fibrotic left atrium would trigger areas of slow conduction and altered repolarization dynamics, shifting the focus of atrial fibrillation initiation and maintenance to the left atrium from the pulmonary veins and requiring additional substrate ablation. 10 Radiofrequency ablation in such patients may be insufficient, leading to increased recurrence rates. Furthermore, a high fibrotic burden is associated with left atrial dysfunction, 11 and an impaired left atrial emptying fraction has been correlated with recurrent atrial fibrillation. 12 Sustained atrial fibrillation is associated with heart failure, [13][14][15] which, in turn, increases left atrial pressure and dilatation, further aggravating the development and progression of atrial fibrillation. In other words, atrial fibrillation and heart failure frequently coexist synergistically. 16,17 This might explain why the prevalence of heart failure was higher in the recurrence group than that in the nonrecurrence group in our study.
Left atrial enlargement is regarded as an independent predictor of atrial fibrillation onset and atrial fibrillation recurrence after radiofrequency ablation. 18,19 However, given the asymmetrical nature of atrial enlargement, the Impact of left atrial geometric remodeling Guo et al. 913  left atrial diameter is no longer regarded to reflect the true LAV. Lang et al. 20 demonstrated that the LAV (and not anterior-posterior left atrial diameter) was associated with atrial fibrillation recurrence after radiofrequency ablation. In a meta-analysis, the authors demonstrated that higher LAV and LAVI were associated with atrial fibrillation recurrence, and these factors emerged as independent predictors. 4 Our findings aligned with this observation.
Recently, several researchers have shifted their focus from left atrial size to shape, which can be assessed as a marker of maladaptive remodeling in atrial fibrillation and is associated with poor ablation outcomes and prognosis. 6,21-23 Bisbal et al. 24 propose a novel quantity termed left atrial sphericity index to describe the morphological remodeling of the left atrium and suggest that the shape of the left atrium changes from disk-shaped to spherical in response to atrial fibrillation. They offered a physiological explanation for this phenomenon that the sphere has the best volume-to-surface ratio and a more stable form; therefore, the left atrial wall accommodates the increased volume and pressure by spherical expansion under low stress. 25 They also presented evidence that patients with a higher sphericity index had an 11-fold higher risk of recurrence than patients with discoid left atria. 26 However, in contrast to previous work, den Uijl et al. studied the effects of the LAV, left atrial sphericity, and fibrosis on the prognosis of atrial fibrillation recurrence after catheter ablation in 83 patients and demonstrated the LAV as the strongest predictor. 27 Moon et al.
observed that the prognosis of spherical remodeling was better among patients with LAV less than 125 ml, whereas patients with LAV at least 125 ml had a better prognosis. 8 In our study, LAV and sphericity were both stronger predictors of atrial fibrillation recurrence after catheter ablation, and we further found that the integration of the left atrial volumetric and geometric quantification provided better prognostic information to predict atrial fibrillation recurrence beyond that obtainable using left atrial minimum volume or sphericity individually.
Nedios et al. 28 demonstrated that atrial fibrillation progression led to asymmetrical dilatation, mainly at the anterior part of the atrium. They claimed that an ASI greater than 60% was associated with a low success rate after catheter ablation but the ASI was less pronounced in association with persistent atrial fibrillation, which already had facilitated asymmetric atrial dilatation. In contrast, we found that ASI was not a strong prognostic predictor of atrial fibrillation ablation outcomes. One explanation for our results could be that asymmetrical deformities may occur before chamber dilatation.   Calibration plots of the nomogram for atrial fibrillation recurrence. The xaxis represents the predicted probability of atrial fibrillation recurrence, and the y-axis represents the observed probability.
other studies, the presence of persistent atrial fibrillation was not associated with atrial fibrillation recurrence in our study. 30,31 This might be why atrial fibrillation progression was associated with hypertension, diabetes mellitus, and heart failure, which played important roles in influencing structural remodeling and ablation outcomes, and high CHADS 2 scores, which, combined with these systemic diseases, may explain the influence of persistent atrial fibrillation. 32 A nomogram is a kind of graphical model that can estimate specific outcomes or survival in association with certain risk factors. Nomograms are widely used in oncology and medicine. 33 Given that catheter ablation for atrial fibrillation patients can be associated with ischemic, bleeding, and thromboembolic events, 34 it was necessary to identify patients who were unlikely to benefit from pulmonary vein isolation. Our model was composed of left atrial size, shape, and clinical findings, which played different roles and interacted with each other. The nomogram could provide a precise probability of atrial fibrillation recurrence. To our knowledge, ours was the first attempt to establish a nomogram to calculate the likelihood of atrial fibrillation recurrence after ablation.

Limitations
This study had a few limitations. First, this single-center, retrospective study had a small sample size. Second, patients with common concomitant diseases were not excluded. Third, atrial fibrillation recurrence may be asymptomatic and cannot be detected by regular and occasional ECG recordings, so our findings need to be confirmed by continuous rhythm monitoring during longterm follow-up. Furthermore, this nomogram model needs further external validation.

Conclusion
The CHADS 2 score, LAVImin, and left atrial sphericity are strongly associated with recurrent atrial fibrillation after catheter ablation. The integration of left atrial size and shape could substantially increase certainty when predicting atrial fibrillation recurrence. A novel nomogram could potentially provide better prognostic information, improve patient selection, and avoid unnecessarily aggressive medical therapy.