Left Atrial Strain

Corresponding Author

The left atrium (LA) has long been regarded as a pressure balloon passively reflecting events in the main heart chamber. Much of our approach to heart imaging has been based on this concept, which has undeniable advantages in the interpretation of left ventricular (LV) function.

However, a large body of evidence has accumulated along the years linking LA function and deformation to invasive hemodynamics1, 2, 3 and to prognosis in a wide array of heart conditions.3, 4, 5, 6, 7 The LA strain integrates information on LV contractility, filling pressures, and LA myocardial properties and has lately been standardized.8, 9, 10

In this issue of JACC: Asia, Chiang et al¹¹ propose not only that the LA function has prognostic value for the cardiac patient, but that the LA deformation may even surpass the LV strain as prognostic tool. In a monocentric head-to-head comparison between deformation parameters, they provide statistically strong evidence that the most sensitive and best correlated deformation parameter extracted from bidimensional studies seems to be the LA strain. This is a rather surprising result, from a theoretical and clinical viewpoint.

LA deformation cannot be studied without taking into account the LV function,^9,12, 13, 14 because the longitudinal excursion of the atrioventricular annulus is the main driver of this deformation (bilateral/reciprocal influence on the relatively rigid atrioventricular annulus) (Figure 1). LV global longitudinal strain (GLS) has already been established as a solid parameter that estimates the global heart function and is currently an integrated part of clinical evaluation.^14,15

Figure 1.

Schematic Representation of the Atrioventricular System as a Physical System Oscillating Between 2 Elastic Springs, of Which One Has a Young’s Modulus Largely Greater Than the Second

The thicker, stiffer left ventricle (LV) is represented by the upper larger spring, while the thin left atrial (LA) wall is represented by the lower smaller spring. The relatively rigid atrioventricular annulus (blue rectangle) has a longitudinal motion under the reciprocal influence of these 2 springs (red arrows). For small deformations compared with the maximal distensibility, the global motion can be approximated to the presence of only 1 (stronger) spring. As the weaker spring becomes stronger (like the left atrium getting stiffer in advancing disease states), its role in the system can no longer be neglected. Potential LA radial distension vectors (question marks, orange arrows) may contribute to the LA longitudinal strain.

This strong dependence of the LA function on the LV properties is obvious when we analyze the LA and LV deformation curves, performed on the same cardiac cycle, in the same patient (Figure 2). We notice that the 2 curves mirror each other, with differences in amplitude produced by the different reference length of the 2 chambers, for the same longitudinal deformation.¹²

Figure 2.

Left Atrial Global Deformation Curve and GLS of the Left Ventricle, Performed on the Same Apical 4-Chamber Cardiac Cycle in a Normal Subject

Left atrial global deformation curve (A, B) and global longitudinal strain (GLS) of the left ventricle (C, D), performed on the same apical 4-chamber cardiac cycle, in a normal subject. The specific phasic function components are noted: the reservoir (elongation during left ventricular systole), the conduit/recoil (pre-A diastolic filling, when the left atrium also acts like an elastic conduit between the pulmonary veins and the left ventricle), and the active atrial pump. Notice that despite differences in scale (caused by the different length of the 2 cavities), the curves mirror each other almost perfectly.

It has been demonstrated in clinical studies that the left atrial reservoir strain (LASr) has a clear correlation with long-term prognosis, potentially integrating information on LV systolic and diastolic function, chamber dimensions, and hemodynamics in a single measurement.^12,15 One of the possible reasons for this correlation with prognosis, in some studies even stronger than the LV GLS,^5,12,16 could be the left atrial thin-walled structure, highly sensitive to overload, and rapidly remodeling (in the form of dilatation and replacement fibrosis), compared with the thicker, more resistant LV walls.

In a real-life demonstration of the theoretical concepts discussed in the previous text, Chiang et al¹¹ investigated a large and diverse patient’s population, followed up to 4 years, finding that LASr had better long-term prognostic value than GLS or right ventricular free-wall longitudinal strain (RVfwLS). This very diversity of pathology may, however, have had an influence on outcomes. The conflicting results obtained in previous studies on well-defined pathologies seem to indicate that the type of disease can have a driving effect on the results. In this cohort comprising 33% with ischemic heart disease, 30% with secondary cardiomyopathy, and 9% with dilated cardiomyopathy, the progressive diastolic dysfunction was probably accurately detected by reduced LA compliance in the reservoir phase. Even with rather large numbers, this monocentric study cannot limit the potential selection bias of the predominant local type of heart disease.

One very important strength of this demonstration comes from its confirmation from multimodality deformation analysis, since the results are similar in cardiac magnetic resonance feature tracking and echocardiographic 2-dimensional strain analysis, performed using the same postprocessing software. Care was also taken to assess the influence of automated tracking vs manual correction in echocardiography, limiting the operator-dependent changes. Consistent results were reported across all analysis.

Most interestingly, in this wide array of pathologies, and maybe because of this diversity, the authors do not find a strong correlation between LV GLS and the LASr, which comes as a surprise given the theoretical considerations presented in the previous text and the results of previous reports.12, 13, 14 One very likely reason is that the study was retrospective and on echocardiographic apical 4-chamber views that were not optimized for the left atrium and hence quite often foreshortened LA views. This potential foreshortening may also limit the technical strength of the analysis.

The fact that the RVfwLS had a lesser impact on prognosis is less surprising.¹¹ The strength of the RV analysis seems to be more obvious in right heart pathology^16,17 or heart failure (HF). In a recent study in patients with HF with preserved ejection fraction, Rothschild et al¹⁸ demonstrated the incremental value of RVfwLS to LV GLS and LASr for predicting all-cause death. However, in the same type of patients with HF with preserved ejection fraction, Freed et al¹⁹ found that LASr outperformed LV GLS and RVfwLS in predicting cardiovascular death and hospitalization.

Although this interesting data encourages us to systematically perform LA strain in clinical patients, it does not help us answer the question on the threshold, which may be used to define prognosis.^8,9,15 In other words, we are pretty confident that, on a statistical level, the evaluation of the LASr correlates with prognostic data in heart patients. However, for the individual patient, the value of the LASr derived from the echocardiography tissue tracking or cardiac magnetic resonance feature tracking has a lesser impact, while GLS values are already quite robust and clinically meaningful.^9,14,15

Moreover, the LA strain itself is problematic for several reasons. Most importantly, the regional deformation in the recommended apical 4-chamber view^14,15 (Figure 2) is hindered by the LA anatomy: LA roof, pulmonary veins, and the proximity of the left atrial appendage, which usually do not participate to the elastic properties of the wall. Therefore, it is theoretically possible that LA deformation, as well as the prognostic information it carries, may be reduced to a linear longitudinal deformation, or even simpler, a longitudinal annulus motion (mitral annular plane systolic excursion, tissue Doppler imaging, or annulus tracking^13,20,21). In this context, we also need to remember that the most important recent technical standardization of longitudinal strain²² managed to align intervendor results, with the price of sacrificing local deformation in favor of the global longitudinal variation in length (closer to the original Lagrangian strain¹⁴).

Because the GLS and the LASr investigate the same longitudinal atrioventricular annulus motion, in the same moment of the heart cycle, their significance should be the same, and their prognostic value also. The difference in amplitude, related to the difference in initial length of the 2 chambers, leads to a wider range of values for the LASr (as it is evident in the wider distribution of this variable in the recent study by Chiang et al¹¹), but the position of the reference points in the far field (LA roof) may also lead to more error compared with the GLS. The contribution of the small deformable portions of the LA walls to the annulus motion during ventricular systole, to the point that it may drive prognosis, and also that it should create a difference in a physical system dominated by the much larger LV, seems unsupported by mathematical and physics considerations, although suggested by statistical clinical data.¹¹

Could it be that the complex longitudinal motion of the atrioventricular annulus, with its asymmetrical tilt and slight systolic rotation,¹⁴ added to the nonlongitudinal atrial expansion during ventricular systole (Figure 1), as well as the “pressure escape” in the pulmonary veins, together exert a sufficient effect making the LASr more sensitive to pathological evolution than the reciprocal LV deformation? Designing an in vivo experiment that would answer this question seems excessively intricate. On the other hand, a multicenter clinical investigation, with very strict methodology and core laboratory analysis of the imaging data may be more feasible in resolving this dilemma.

Because these theoretical considerations seem supported by recent clinical data, suggesting the predominant driving role of the LV in LA deformation,¹² the practically sound approach would be the systematic study of the GLS as a measure of heart function and main prognostic tool, while the LASr should complement these studies, especially in cases when the left atrial dysfunction plays a major role (like in mitral valve disease, atrial fibrillation, HF with preserved ejection fraction, and diastolic dysfunction in general).^23,24 RVfwLS should be performed in cases with RV enlargement or dysfunction, significant tricuspid regurgitation, or pulmonary hypertension.¹⁵

Funding Support and Author Disclosures

The author has reported that he has no relationships relevant to the contents of this paper to disclose.