The Effect of Catheter Ablation on Left Atrial Size and Function for Patients with Atrial Fibrillation: An Updated Meta-Analysis

Background Catheter ablation (CA) for atrial fibrillation (AF) is now an important therapeutic modality for patients with AF. However, data regarding changes in left atrial (LA) function after CA have indicated conflicting results depending on the AF types, follow-up period, and the analytical imaging tools. The objective of this review was to analyze the effect of CA on the LA size and function for patients with AF. Methods We searched for studies regarding LA size and function pre- and post-ablation in PubMed, Embase, the Cochrane Library, and Web of Knowledge through May 2014. LA function was measured by LA ejective fraction (LAEF), LA active ejective fraction (LAAEF), or both. Total and subgroup analyses were implemented using Cochrane Review Manager Version 5.2. Weighted mean differences with 95% confidence intervals were used to express the results of continuous outcomes using fixed or random effect models. I2 was used to calculate heterogeneity. To assess publication bias, Egger’s test and Begg’s funnel plot were performed using Stata 12.0. Results Twenty-five studies (2040 enrolled patients) were selected for this meta-analysis. The LA diameter (LAD), maximum LA volume, and minimal LA volume were significantly decreased post-ablation, as compared with those at a pre-ablation visit. Compared with the pre-ablation outcomes, we found no significant differences in LAEF/LAAEF at a post-ablation follow-up. Decreases in LA volume and LAEF remained significant post-ablation for paroxysmal AF (PAF); however, the LAEF was insignificant changes in persistent AF (PeAF). Heterogeneity was significant in spite which individual study was excluded. A publication bias was not found. In a meta-regression analysis, we did not find any factor that contributed to the heterogeneity. Conclusion With CA, LA volumes and LAD were decreased significantly in patients with AF; LAEF was not significant changes in patients with PeAF but decreased in those with PAF.


Introduction
Atrial fibrillation (AF) is the most commonly sustained tachyarrhythmia in clinical practice. It is associated with an increase in disease-related hospitalizations; a reduction in quality of life; complications such as congestive heart failure (HF), thromboembolism, and stroke; and an increased mortality risk [1][2][3][4]. Catheter ablation (CA) is considered an efficient mainstream therapy and potentially curative treatment for drug-refractory symptomatic AF [5,6]. After successful ablation, patients with AF would experience improved left atrial (LA) function because of a reduction in AF burden. Nevertheless, it is noted that extensive atrial scar tissue formation produced by CA may result in adverse reactions in atrial function in AF patients. Although Jeevanantham et al. [7] reported successful CA for AF patients does not appear to adversely impact LA function, recent studies that investigated the impact of CA on LA function reported inconsistent results. Therefore, the purpose of this study was to update evidence regarding the effect of CA on the LA size and function in patients with AF.

Search Strategy
We performed a search for articles pertaining to CA in AF patients using the key words "atrial fibrillation," "catheter ablation," "atrial size," "left atrial function," and "left atrium function. " We searched for all relevant studies, without any language limitations, in PubMed, Embase, the Cochrane Library, and Web of Knowledge through May 2014. Manual searches were also performed of the bibliographies.
Exclusion criteria were as follows: (1) surgical ablation; (2) left ventricular ejection fraction (LVEF) of <50% or included HF patients in each enrolled study; (3) significant valvular disease including a stenotic valvular lesion or moderate-to-heavy regurgitation after valvular replacement; (4) heart dysfunction was caused by structural heart disease or another disease; (5) the LA parameters, as detailed previously, were not reported either pre-or post-ablation; and (6) median and inter-quartile range outcomes were reported.

Data Extraction and Quality Evaluation
Two reviewers (Xiong and Li) assessed the quality of each study and then independently extracted data from the included studies; another author (Wang) checked the data. The extracted information were: (1) basic information regarding those studies, including country and publication year; (2) the number of patients in the study; (3) patient characteristics; (4) type of catheter ablation performed for the treatment group; and (5) outcome measures, as previously defined. Any disagreement was resolved by discussion with a third party (Wang).
To evaluate the quality of the included studies, the following aspects had been performed, including (1) research design; (2) the representativeness of the enrolled patients; (3) the bias of loss to follow-up; and (4) other biases and limitations.

Statistical Analysis
Cochrane Review Manager Version 5.2 and Stata 12.0 were used to perform the statistical analysis. Weighted mean differences (WMDs) with 95% confidence intervals (CIs) were used for expressing continuous outcomes. Statistical heterogeneity was tested using the χ 2 test and was quantified using the I 2 statistic; significant heterogeneity was defined as a P of <0.10 or an I 2 of >50%. Data were pooled using a fixed effect or random effect model, based on whether the absence of significant heterogeneity existed. If the absence of heterogeneity was significant, the fixed effect model was performed, but if not, the random effect model was performed. Publication bias was evaluated using Egger's test and Begg's funnel plot with Stata 12.0; statistical significance was defined as a P of <0.05.

Study Characteristics
We identified 1566 references from electronic databases using the previously described strategy. According to the inclusion criteria, 92 citations were retrieved and required further evaluation after screening the title, abstract, or both. Forty-one reviews and 14 case reports were excluded. Two studies reported median and inter-quartile range outcomes; 6 studies included surgical ablation; 2 studies had a follow-up of <3 months; and 2 studies included HF patients. Finally, 25 studies (2040 enrolled patients) were selected for this meta-analysis . The selection process is demonstrated in a flow chart (Fig 1). The characteristics of each included study are listed in Table 1. The primary results of each included study are shown in Table 2.
Subsequently, we implemented a subgroup analysis on the basis of AF recurrence. There were significant decreased in LAD (WMD, -1.63 mm; 95%CI, from -3.01 mm to -0.24 mm, P = 0.02, S1

Heterogeneity Analysis
After performing a heterogeneity test, the existence of heterogeneity among those studies should not be ignored. To demonstrate the origin of the heterogeneity, a meta-regression analysis and sensitivity analysis were performed. Heterogeneity was significant in spite which individual study was excluded. As previously illustrated, a subgroup analysis was performed; each Effect of CA on LA Size and Function for AF Patients outcome was analyzed based on the AF type (PAF or PeAF). A random effects model was used to combine the effect size because significant heterogeneity was shown as an all-total consequence.

Publication Bias Analysis
Egger's test and Begg's funnel plot were used to evaluate publication bias. There were no significant risks of publication bias according to an analysis using Stata 12.0 (the P value for each test was >0.05; Table 3). The funnel plot was generally symmetrical, and it indicated that the publication bias for the studies was controlled.

Discussion
In the present review, we found that the LA volumes and LAD were significantly decreased after CA therapy during follow-up imaging. Nonetheless, we did not find any significant changes in LA function (included LAEF and LAAEF) after ablation treatment during followup imaging. Furthermore, there were significant decreases in the LA volumes and LAEF with  paroxysmal AF after CA treatment. However, we did not find any significant changes in outcomes, as previously detailed, for persistent AF after ablation therapy, except for LA volumes.
CA is a therapeutic method for terminating the underlying electrophysiological mechanism of AF. The substrate and trigger foci are isolated by freezing (cryoablation) or radiofrequency energy and then terminate the electrical conduction from the pulmonary vein (PV) to LA. Currently, CA is approved by the Food and Drug Administration (FDA) for managing paroxysmal AF. Although this practical strategy is also used for managing non-paroxysmal AF, unfortunately, it is not yet approved by the FDA [2,35]. The resumption of a sinus rhythm with CA is a perfect consequence, but the amount of LA scarring produced by CA could influence LA structural and functional remodeling, especially with repeated ablation. Structural remodeling includes increasing LA size and a change in LA strain. Several studies [10,18,19] reported that the enlargement could be reversed after successful ablation therapy that is defined as the maintenance of a sinus rhythm during follow-up [34]. Thus, LA reverse remodeling may become a robust sign of successful CA for patients with AF. Further studies should be conducted to evaluate the effects on LA function for patients with AF after a repeat ablation treatment.
There was a significant decrease in LAEF after CA treatment in studies with paroxysmal AF; however, we did not find similar outcomes in studies with persistent AF. Rodrigues et al. [17] reported a degradation in LAEF after CA for patients with paroxysmal AF at a follow-up duration of about 8 months after performing transthoracic echocardiography (TTE). Hof et al. [11] found a similar outcome using three-dimensional computed tomography (CT). However, Erdei et al. [10] and Machino-Ohtsuka et al. [13] described that the LAEF was preserved and even increased in patients without an AF recurrence at a follow-up of >12 months; however, it had decreased in AF recurrence patients after TTE and CMR. Why did this phenomenon occur in these studies? Several reasons for this variance should be considered, including the  follow-up duration after CA therapy; the chronicity of AF; the different clinical outcomes; and the different degrees of tissue damage related to the different ablation strategies, tools, or both. More consideration should be given to the follow-up duration regarding studying LA function and size. Studies with a long follow-up (not less than 12 months) [12,13,19] have illustrated significant increases in LAEF after ablation treatment in AF patients; however, insignificant changes in LAEF with paroxysmal AF in 3 months follow-up [14,16]. Further, a sensitivity analysis based on the follow-up duration was not persuasive because of a lack of detailed individual patient data. McGann et al. [36] reported that the quantification and detection of left atrial wall scarring would be applicable 3 months after CA in patients with AF. As we known, there is a phenomenon of atrial "stunning" during in 1 month follow-up after catheter ablation therapy, and either the LA size or the LA function is unstable change. After the "blanking period" (about 3 months), which the rate of recurrence AF is highest [37], the quantification and detection of LA size and function is more credible and accurate. Therefore, it is important that a longer follow-up duration should be performed for evaluating LA function. Moreover, because the LA function was assessed using only the sinus rhythm (SR), it is difficult to evaluate LA function in permanent AF patients.
In addition, the imaging technique is another important factor. Many different methods were performed in the studies, including TTE [9, 10, 13-15, 17-20, 22, 24, 26-32], CMR [11,12,16,25,33], and CT [14,21,23,27]. As an established method in cardiac imaging modalities, TTE can identify the size of each chamber, as well as the ejection fraction of the LA and left ventricle (LV). However, a limitation occurs if patients are obese and have serious obstructive pulmonary disease with poor acoustic windows [38]. Multidetector Computed Tomography (MDCT) has a prominent temporal and spatial resolution for measuring LA volumes. CMR can concurrently discover pre-ablation fibrosis and post-ablation scar tissue and measure PV anatomy in patients who undergo CA therapy [39]. As we known, different analytical methods or image tools would obtain different results. Compare to TTE, using CMR and CT have a prominent temporal and spatial resolution for measuring LA volumes and EF, the results of CMR or CT should be more accurate than that of TEE. Due to this important issue, we performed subgroup analysis based on the variant methods of image. Subsequently, after excluding studies using TTE, only 9 studies (enrolled 635 patients) [11,12,14,16,21,23,25,27,33] were included in the subgroup analysis on the basis of detecting by CMR and/or CT. The LA volumes significantly decreased, LAEF/LAAEF insignificant changed. The explanations for this phenomenon as follow. First, CMR and CT/MDCT are more accuracy and improve reproducibility in measurement of LA volumes and functions compare to TTE. Second, the numbers of included studies were decreased, and then it may influence the pooled data. Therefore, compared with MDCT and CMR, TTE may underestimate the true LA size and function.
Nonetheless, there is no gold standard for measuring LA function. In the present review, the LAEF was used to define LA function in 15 studies [10-17, 19, 23, 27-29, 31, 33], and only 6 studies used LAAEF to define LA function [11,15,16,24,29,31]. Furthermore, A wave velocity [9,13,17,[27][28][29]32] and A' wave velocity [10,17,28,32] were used to define LA contractile function. The A wave velocity involves the peak velocities of the late transmitral flow, as measured by pulsed-wave DE, reflecting LA systolic function from hemodynamics, but it is not sensitive because it can be affected by the LV diastolic function and preload. However, compared with the A wave velocity, the A' wave velocity, as detailed previously, is an easy and effective means to assess LA systolic function from tissue motion because it is correlated with changes in the LA systolic area and volume [8]. Therefore, further studies should be conducted to assess this method for evaluating LA function.
Beyond these, the treatment strategy and energy of catheter ablation are another factor. In this meta-analysis, the majority of included studies were used RFCA, only one study [10] performed cryoablation. As we known, different treatment strategies, such as SPVI and PVAI, lead to different outcomes, and diverse ablation temperature and power resulted in different damages for atrium. The included studies used RFCA were set at a similar value of the ablation temperature and power, and therefore, the results have consistency and comparability. After excluding the study performed cryoablation, there were similar pooled data compare to previous detailed. Due to there was no more available data, further studies should be focused on evaluating the effects on LA function and size for patients with AF after cryoablation treatment.
Heterogeneity is an important issue for explaining the outcomes of this review, and significant heterogeneity was found in this meta-analysis. Subsequently, sensitivity analyses were performed, and heterogeneity was significant in spite which individual study was excluded. We did not find any contributing factor for the heterogeneity with a meta-regression. The quality of the included articles may be the origin of heterogeneity.
Moreover, our review had some limitations. First, we did not consider any randomized control trial in this meta-analysis; the sample sizes of the included studies were small, and most were single center and either a prospective or retrospective study that may have added potential biases to such studies. Second, it is difficult to draw decisive conclusions regarding LA functional change after ablation therapy, because of inconsistencies regarding individual patient data, the imaging method, and the follow-up duration. Third, although publication bias was not significant after performing an Egger's test and a Begg's funnel plot, the influence of bias in this article could not be thoroughly excluded, as only studies published in English were included. Forth, we have tried addressing an issue but indirectly regarding the effectiveness of CA for AF by looking at LA size and function, however, it is a pooled data and it has its own set of issues which precludes us from providing any more clarity. Moreover, another limitation is the lack of a gold standard to measure LA function among these involved studies. Currently, MDCT and CMR are considered relatively accurate methods for measuring LA function and size. Finally, although several studies reported that the LA volumes and sizes are predictors of AF recurrence after CA therapy [40][41][42], our review did not perform an analysis based on AF recurrence in different types of AF. Therefore, we do not know the relationship between AF recurrence and LA function/size among different types of AF.
In conclusion, With CA, LA volumes and LAD were decreased significantly in patients with AF; LAEF was not significant changes in patients with PeAF but decreased in those with PAF.