In this study, we aimed to comprehensively assess LA morphofunctional remodeling using 3D STE which was shown to be feasible and reproducible. Our main finding is that AF is independently associated with LA enlargement as well as impairment of LA reservoir and pump function but not conduit function. Besides, 3D STE derived function parameters, especially LAVmin, LAEF, aLAEF, LASrc and LASctc can be used to differentiate PAF from non-AF subjects with good sensitivity and specificity.

Feasibility of 3D STE in assessing LA function

In our study, the success rate of 3D STE analysis in each groups was 100, 96 and 83% respectively. In terms of reproducibility, both intraobserver and interobserver reproducibility were excellent when conducted under sinus rhythm, however, in patients with Per-AF, the interobserver reproducibility was less satisfying. The reasons were as follows: 1) LA enlargement was more significant in Per-AF and thus wider sector angle was needed during acquisition to include the entire LA which lowered the frame rates and image quality, 2) multi-beat acquisition might be feasible under AF rhythm, but the reconstruction effect would be largely compromised, 3) only single-beat acquisition can be used with obvious heart rate variation which significantly lowered the frame rates, 4) patients with Per-AF tended to have higher heart rates as shown by our study, and thus higher frame rates were needed in order to improve temporal resolution which again compromised image quality. To sum up, practically it’s more difficult to obtain 3D images with good quality in subjects with Per-AF, which lowers the accuracy of automatic LA contour recognition and tracing by the software, and therefore more manual correction is needed which is not always easy and introduces more error in measurements since there still lack widely accepted standard operational guide of this new technique, especially in subjects with Per-AF.

Echopac 4D Auto LAQ by GE Health, as a 3D STE software package dedicated to LA, was not available until recent years. Experience in using this software is scarce. One study demonstrated LA functional changes in end stage renal dysfunction in a total of 71 subjects, but didn’t discuss the usage of the software at the same time [25]. To our knowledge, our study is the first to analyze the feasibility and reproducibility of this software as well as its utility to assess LA morpho-functional remodeling in patients with both PAF and Per-AF, at the meantime with a decent sample size and thus filled certain gap in this field.

LA structural and functional changes associated with AF and AF pattern

As LA function is affected by many factors such as age, gender, BMI, comorbidity, LV function status, etc., therefore, in order to demonstrate LA remodeling independently associated with AF, confounders were adjusted based on univariant linear regression and clinical judgement. Our study showed that impairment of LA reservoir and contraction function was independently associated with PAF, additionally, LA reservoir function was further decreased in Per-AF patients. However, pLAEF and LAScd were not independently associated with PAF or Per-AF, suggesting that AF has relatively small direct impact on LA conduit function. Interestingly, we noticed that pLAEF was even higher in Per-AF group compared to PAF group, which we believe was caused by different calculating methods of pLAEF between 2 groups as it was considered equal to LAEF under AF rhythm. Therefore, comparison of LA conduit function between subjects under different rhythm is not suitable.

Our results are consistent with previous studies, demonstrating association between AF and enlarged LA volumes as well as impaired LA function indicated by conventional 2D echocardiography derived volume measurements [10, 26] or longitudinal LA strain and strain rate derived from 2D STE [27] or DTI [28]. However, 3D STE, as a novel technique combining real-time echocardiography with speckle tracking analysis, is a lot less applied in LA function assessment compared to LV. Mochizuki et al. used 2D STE and 3D STE to compare atrial deformation and synchrony in healthy subjects and patients with AF and found that 3D STE was less time consuming and more sensitive [16]. Significant decreases in longitudinal, circumferential and area strain across AF types were detected, however, confounders were not adjusted in their study in despite of some significant imbalances at baseline and thus couldn’t yet manifest AF-induced changes in LA strain as they tried to address in their article. The results of our study, with a much greater sample size, is a valuable supplement to theirs and more conclusive as explicating LA structural and functional changes independently associated with AF. A sub-study of the MYGYAR-path study conducted by Chadaide et al. compared 3D STE derived LA strain parameters in 20 AF patients and 11 healthy controls. Decrease of peak circumferential and radial strain was shown in all LA segments while reduction of peak longitudinal strain was only detected in mid and superior LA segments, and no difference was found in terms of LAEF [29]. However, their results may hold certain bias due to small sample size. Furukawa et al. sought to assess the effect of PAF on top of hypertension on LA function using 3D STE (44 PAF + HTN vs. 50 non-AF HTN), and demonstrated decreased peak atrial longitudinal strain, LAEF and synchrony compared to hypertensive patients without PAF which indicated diminished LA compliance and electro-mechanical abnormality [30]. The above studies shared one common limitation that they used 3D STE software originally designed for LV which was prone to introducing errors. Good news is that GE Health has recently launched 3D STE software package dedicated to LA, and to our knowledge, our study is the first to explore its application in AF population including subjects with Per-AF whom were usually excluded in many studies, and thus provides valuable experience.

Relationship between LA volumetric function parameters and strain

Echopac 4D Auto LAQ allows simultaneous analysis of LA volume and strain, and therefore we were able to compare the relationship between different LA function parameters. The three components of LA function can be either indicated by volumetric function parameters (LAEF, pLAEF, and aLAEF) or strain parameters (LASr/LASrc, LAScd/LAScdc, and LASct/LASctc). It’s not surprising to find strong correlation between LA emptying fraction and its corresponding strain parameter that represents the same LA function, however, on top of that, we found that the correlation between LA emptying fraction and its corresponding circumferential strain was always significantly stronger than that with the longitudinal strain, which we believe suggest that changes of LA deformation in the circumferential direction might have greater impact on LA global hemodynamic function. To our knowledge, our study is the first to compare the relationship between LA volumetric function parameters and longitudinal strain with the relationship between LA volumetric function parameters and circumferential strain since few studies were able to provide all these parameters at the same time using other techniques. At present, many studies are focusing on 2D STE derived LA longitudinal strain and its prognostic value for various adverse cardiovascular outcome, however, we assume that LA circumferential strain might possibly hold greater value, and therefore we call for more studies with this index in the future.

AF related fibrosis and LA remodeling

LA enlargement and fibrosis are hallmark of AF which not only initiate but also contribute to the maintenance and progression of AF [4, 31]. The degree of LA fibrosis can be evaluated via delayed-enhancement cardiac magnetic resonance imaging (DE-CMR), however its use in daily clinical practice is limited due to its high cost, operational complexity and significant time consumption [32]. Till now, echocardiography remains the most useful and convenient tool to evaluate cardiac structure and function. Since it’s now believed that fibrosis is associated with LA remodeling and decrease of LA compliance, LA function parameters assessed by echocardiography can be considered as a surrogate for LA fibrosis. There have been some studies supporting this idea. Cameli et al. found that peak atrial longitudinal strain (PALS) derived from 2D STE had strong inverse relationship with the extent of LA fibrosis (r = − 0.82) based on histological sample obtained from valve replacement surgery in 46 AF patients with severe mitral regurgitation while LAVImax and LAEF showed medium association with LA fibrosis (r = 0.51 and − 0.61, respectively) [33]. Another study by Kuppahally et al. demonstrated that strain value of LA mid-lateral wall was inversely related to the extent of LA fibrosis detected by DE-CMR as well as AF burden [34]. Therefore, comprehensive assessment of LA function via 3D STE can provide doctors with more disease information of different AF patients.

Identification of PAF using LA function parameters

Another important finding of our study is that LA volume and function parameters can be used to distinguish PAF from non-AF subjects with great accuracy, sensitivity and specificity.

According to current guideline, diagnosis of AF should be made based on a standard 12-lead ECG or a single-lead ECG [35]. The current prevalence of AF is estimated to be 2–4% [1], however, since some AF episodes are completely asymptomatic, or instantly self-terminate before ECG examinations are conducted, many AF patients are undetected. AF can lead to some fatal and highly-disabling complications such as stroke and congestive heart failure, which can be prevented or managed through medication. Therefore, it’s of great importance to identify those PAF patients who visited the clinics under sinus rhythm and make timely diagnosis. In the REVEAL AF study, insertable cardiac monitor (ICM) was used to screen for AF in patients with high risk for a mean observation period of 22.5 months. The detection rate of AF at 18 months was 29.3% which further increased to 40.0% if the monitoring period was prolonged to 30 months, revealing a surprising amount of undetected AF via conventional 12-lead ECG or regular 24 h ambulatory ECG monitoring [36]. The CRYSTAL-AF study aimed to detect underlying AF in patients with cryptogenic stroke using ICM and the detection rate at 6, 12, and 36 months was 8.9, 12.4 and 30%, respectively [37]. It’s easy to understand that the more intense the screening strategy is, the more undiagnosed AF can be detected, yet with higher cost. Therefore, the major issue is to determine the high risk population that truly needs prolonged rhythm monitoring. In our opinion, LA function parameters are useful under this scenario because unlike ECG abnormalities of PAF patients which is “paroxysmal”, changes of LA function are stable and can be detected by echocardiography under sinus rhythm, even before morphology changes take place, and thus can help to identify subjects with potentially underlying AF. In this study, we calculated the PPV and NPV of LA function parameters to identify PAF patients in populations with different AF prevalence. As for the general public with an estimated PAF prevalence of 1% [1, 38], the NPV of various LA volume/function parameters was excellent (> 99%), and therefore can be used to effectively rule out PAF and so we know who don’t need intensified ECG monitoring. On the other hand, in populations with a stroke history where the prevalence of AF is approximately 20% [39], the PPV of LAEF and LASrc for identification of PAF exceeded 50%, which is of great value because current commonly used tools for investigating cardiogenic stroke are echocardiography and 24 h ambulatory electrocardiogram with the purpose to find relevant structure abnormalities (such as patent foramen ovale, LA thrombus, endocardial vegetations, etc.) and AF. However, it’s highly likely that no AF episodes occur during monitoring period after the stroke event, which may lead to misclassification of the stroke type as cryptogenic stroke and delay the initiation of appropriate anticoagulant therapy. What we propose is that comprehensive assessment of LA volume and function using 3D STE can further suggest whether the patients still hold great chance (> 50%) of underlying AF even when Holter results are negative, and therefore require prolonged ECG monitoring. On the contrary, intensified AF screening can be reasonably saved when LA volume and function are within normal ranges. It needs to be point out that our aim is not to use LA function parameters derived from echocardiography to directly diagnose AF, but to help determine the optimal screening strategy for different individuals. In this way, valuable medical resources can be reasonably allocated.

Advantages and limitations of 3D STE

STE is the current preferred method for strain measurement compared to TDI, as the latter is angle dependent, time-consuming due to requirement of wall-by-wall sampling, relatively more dependent on pre-load, and poor in reproducibility. Up till now, 2D STE is the most commonly used method to obtain strain, however, it faces problems such as lacking consensus on acquisition and offline analysis, foreshortening of the 2D images, the out-of-plane phenomenon, etc.. The Echopac 4D Auto LAQ we used in our study is a 3D STE software package dedicated to the LA, which technically overcomes the inherent disadvantages of 2D STE. Its high degree of automation saves the need to manually trace LA endocardium, which not only reduces analysis time, but also greatly improves reproducibility. In addition, the information it provides is more elaborate, including reliable LA phasic volumes which is not based on geometric assumption, LA volumetric function parameters as well as LA strain in multiple directions. On the other hand, 3D STE faces certain challenges as well. As it’s a rather new technique, it lacks operation consensus and experience. Besides it has a relatively higher requirement for image quality that might lead to unsuccessful analysis or unreliable results if not fulfilled. Specific software and transducer are needed to perform 3D STE and values vary between venders which limit its application in clinical practice. However, since 3D echocardiography is a promising tool in the future, and the operation of 3D STE is actually easier compared to 2D STE, yet providing a lot more information, we believe this technique hold great value in both clinical practice and scientific research.


Our study is a single centre, cross-sectional study took place in a public teaching hospital and though we tried our best to enrol patients with AF continuously during the study period, some might still be missed due to various reasons, and thus our study may carry certain bias. The classification of PAF and Per-AF was made based on patients’ description of their symptoms combining with their previous ECG and Holter reports, and therefore misclassification between groups might exist. We detected stronger correlations between LA circumferential strains and volumetric function parameters, but the underlying mechanism couldn’t yet be drawn from this study. LA circumferential strains might possibly hold greater value than longitudinal strains, but more studies will be needed in the future.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit The Creative Commons Public Domain Dedication waiver ( applies to the data made available in this article, unless otherwise stated in a credit line to the data.


This article is autogenerated using RSS feeds and has not been created or edited by OA JF.

Click here for Source link (