Background

In recent years, greater attention has been focused on patients with typical heart failure (HF) signs and symptoms but with normal values of left ventricular (LV) ejection fraction (EF) usually >50% [1]. This subset represents the half of subjects that access the emergency room and even more in ambulatory setting [2]. Prevalence of HF with preserved EF (HFpEF) is increasing. Morbidity and mortality are still elevated [3], and differently from HF with reduced EF (HFrEF), outcomes have not improved in last decades, mainly because of under diagnosis and lacks of effective therapies [4].

The pathophysiological mechanism that leads from asymptomatic diastolic dysfunction (DD) to HFpEF is still not completely clear. Despite a maintained cardiac output, left-sided cardiac chambers go through important remodeling which alters cardiomyocytes and interstitial ultrastructure [5] and then filling function [6]. In particular, a key role in preserving diastolic phase and preventing the onset of HF symptoms seems to be played by the left atrium (LA). Echocardiography allows to accurately and non-invasively evaluate LA dimensions and function. This could improve correct diagnosis, therapy and prognostication in patients with DD and HFpEF.

The purpose of this review is to highlight the importance of LA function in patients suffering from DD and HFpEF and to describe the usable methods for its assessment. We base our paper on the research on PubMed literature, using mainly the following terms: “left atrial function”, “left atrium and diastolic dysfunction”, “left atrial strain and diastolic dysfunction”, “left atrium and heart failure preserved ejection fraction”, “left atrium and diastolic heart failure”, “left atrial strain and heart failure preserved ejection fraction”, “left atrial strain and diastolic heart failure”, “left atrium heart failure symptoms”, and “left atrium heart failure prognosis”. We did not decide to pose a time limit in this research, selecting the most appropriate and innovative papers.

Left atrium in diastolic dysfunction

The LA is considered a reservoir and a conduit of blood, placed between pulmonary and systemic circulation, and ensures correct ventricular filling without raising pulmonary venous pressure. The LA cavity is in direct contact with the LV cavity when the mitral valve opens, during diastolic phase, so it is exposed to LV pressures during every diastole. For this reason, increase in LA size and alteration in its function could represent a picture of average LV pressure history and may indicate an index of DD [7, 8] differently from E/E′ ratio that can be considered a snapshot of pressures at the time of the evaluation [9]. In patients with normal LV filling pressure at rest, it has been demonstrated how LA volume index can predict rise in filling pressures during exercise echocardiography, supporting the hypothesis of a cumulative effect [10]. Therefore, LA volume has been defined as the “glycosylated hemoglobin of diastolic dysfunction”.

Evaluation of left atrium size

Transthoracic echocardiogram represents the most used method for assessing LA dimensions. The most commonly used linear dimension is LA anteroposterior measurement that could be obtained in M-mode or, preferably, two-dimensions (2D) in parasternal long-axis view. Despite its high reproducibility, anteroposterior dimension is considered inaccurate because it might not be representative of the real LA size [11] and should never be assessed as the only LA dimensional parameter. The LA area, measured in apical two- or four-chamber view, may be obtained but barely used in common practice. LA volume, in particular, LA volume index (that is, LA volume indexed to BSA, body surface area), is the preferable method for its evaluation because it relies on minor geometric assumptions and permits to detect enlargements along different space axes. The American Society of Echocardiography recommends the assessment of LA volumes in two- and four-chamber apical views by Simpson’s method, with an upper normal revised cut-off value of 34 ml/m2 [12]. In patients with raised LV filling pressures, LA dimensions tend to increase, as already mentioned. While in grade I DD, LA volume is still normal, the atrial chamber dilates in higher grades. It has been widely demonstrated how LA volume expresses the intensity and the chronicity of DD [13] and proportionally increases with increasing severity of DD as defined by invasive hemodynamic study [14]. LA macroscopic remodeling is fundamental because it guarantees the best ventricular filling despite elevated wall stiffness and high diastolic pressures, until very advanced stages of dysfunction. On the other hand, the LA preserves pulmonary capillary circulation from hemodynamic overload [15] for a long time and, consequently, preserves the patients from symptoms’ onset. Exertional dyspnoea is the typical clinical presentation in subjects with DD; between all echocardiographic variables, LA volume index has been shown to be an independent predictor of exercise capacity [16].

Assessment of left atrium function

LA physiology during every cardiac cycle consists of three different phases that modulate LV filling. The reservoir phase starts with mitral valve closure: the LA fills up also thanks to the movement of atrial floor towards the heart apex. As LA pressure falls, its volume increases facilitating the passage of blood from the pulmonary circulation. After mitral valve opening, during early LV filling, the energy stored in LA walls is transferred to the LV and, jointly with the suction effect exerted from the ventricle itself, allows blood flow to the LV and the reduction of LA volume. At the same time, a direct flow from pulmonary veins through the LA into the LV occurs (conduit phase). In the last phase (atrial systole), the LA contracts determining the motion of mitral valve plan towards the ventricular apex and the additional transfer of blood into the ventricular chamber. LA function can be evaluated by measuring chamber volumes during cardiac cycle in different time points that reflect the mentioned phases (at end-systole, before mitral valve opening, that is maximal volume; at the onset of P wave on ECG, that is pre-atrial contraction volume; at end-diastole, before mitral valve closure, that is minimal volume) and calculating, with specific formulas, its reservoir, conduit and pump fractions [17]. In addition to volumetric evaluation, Doppler echocardiography has been used for the same purpose. Peak A wave by transmitral pulsed-wave (PW) Doppler is considered an index of LA function [18], but it is influenced by loading conditions and patient’s age [19]. Also, tissue Doppler imaging (TDI) A′ wave has been correlated to atrial function in numerous studies, proportionally with the grade of DD [20, 21]. However, this approach does not allow to evaluate regional LA function.

Speckle tracking echocardiography (STE)-derived LA strain currently constitutes the best method to assess atrial function. The software was born for the application to the LV, to analyze its longitudinal, radial and circumferential deformation, but STE has demonstrated great feasibility and reproducibility also when applied to the LA chamber [22]. It represents a direct measurement of intrinsic LA myocardial deformation, relatively independent of loading conditions and geometric assumptions [23]. After the acquisition of atrial 2D images in apical four and two chambers, strain curves are obtained by an off-line semi-automated analysis using specific software. In the current ASE/EAE consensus [24], two different techniques have been proposed to quantify atrial deformation by STE. The most widely applied method is performed using QRS onset as a reference point. According to this measurement, LA strain is composed of two peaks that trace atrial activity during cardiac cycle. During reservoir phase, the LA fills and stretches so strain increase, reaching a positive peak at the end of chamber filling: at this point, it is possible to measure peak atrial longitudinal strain (PALS); after mitral valve opening, the LA quickly empties and its volume falls so strain decreases up to a plateau, in diastasis phase. Then, in the short period before atrial contraction, only in patients with sinus rhythm, LA strain increases again up to a second positive peak, that is peak atrial contraction strain (PACS) [22] (Fig. 1). The other modality of quantification has been proposed by Saraiva et al. and considers P wave as a reference point: it measures a first negative peak atrial longitudinal strain curve (atrial systole, called ε negative), a second positive peak atrial strain curve (corresponding to reservoir function, ε positive) and their sum [25]. It has been widely demonstrated how LA function, assessed by STE, deteriorates even without chamber dilation [26], and then, a reduction in strain value suggests the presence of a microstructural remodeling that appears long before macroscopic one. Once extensive fibrosis has developed, the remodeling process might become irreversible so it is fundamental to detect it very early in order to intervene therapeutically. LA strain has been correlated with the degree of atrial fibrosis evaluated by Masson trichrome staining on an LA tissue sample [27]. Cardiac magnetic resonance (CMR) with late Gadolinium enhancement (CMR) allows to evaluate the presence of myocardial inflammation and fibrosis thanks to the relative accumulation of Gadolinium as the result of a slower washout and an increased extracellular volume [28]. The detection of fibrosis with CMR is challenging for the LA mostly because of its thin walls; however, experts suggest that new 3D CMR sequences, despite being time consuming and technically difficult, have obtained better results [29]. Interesting data have emerged on the inverse correlation between LA strain and the amount of LA fibrosis detected by 3D CMR [30], but they require further evaluation and validation. DD is characterized by altered LV relaxation and reduced compliance that lead to an increase in LV end-diastolic pressure, mean LA pressure and pulmonary capillary wedge pressure (PCWP), all referred to as LV filling pressures [31]. PALS represents a great tool in the non-invasive evaluation of these pressures (Fig. 2), with the greatest diagnostic accuracy among all other echocardiographic parameters [32–34]. Moreover, LA strain can be used to estimate LA stiffness, both invasively, with PCWP, obtained by heart catheterization, and non-invasively, with E/E′ ratio, got by PW Doppler and TDI. A low atrial function is an independent predictor of estimated metabolic equivalents (METs) in patients with DD [35].

Fig. 1
figure 1

Assessment of peak atrial longitudinal strain (PALS) at the end of reservoir phase and peak atrial contraction strain (PACS) before the atrial contraction. The dashed curve represents the average atrial longitudinal strain during the cardiac cycle [51]

Full size image
Fig. 2
figure 2

Left atrial strain and ventricular filling pressure: differences in pulsed-wave Doppler of mitral flow velocities, tissue Doppler imaging and strain between a patient with normal filling pressures (above) and one with increased ones (below) [24]

Full size image

Left atrium role in HFpEF

HFpEF may represent the evolution of a chronic DD, but completely understanding its pathophysiology still represents a work in progress. In fact, while DD has a high prevalence in the community and increases with age, the main part of these patients does not successively evolve towards HF with typical signs and symptoms [36]. Comorbidities, as age, hypertension, diabetes and coronary artery disease, play an important role in HFpEF [37]. Making an accurate diagnosis of HFpEF remains a challenging topic too. For a long period of time, in HF patients, attention has been focused on LV structure, dimensions and function, relegating the LA to a marginal position. In recent years, however, the atrial chamber has acquired increasing importance and its fundamental role has been shown both in modulating ventricular filling and in providing diagnostic and prognostic clues.

Since the very first stages of LV filling impairment, the LA tries to implement a sort of compensating action in order to support the dysfunctional ventricle chamber. Over time, nevertheless, the remodeling process profoundly alters the atrium, so much that compensatory mechanisms fail and a real “LA failure” appears [38]. This is the moment in which HFpEF becomes evident.

The LA is frequently dilated in HF patients if compared with controls, despite a preserved EF [39], and its size tends to increase proportionately with LV mass [40] and stiffness [41]. Atrial chamber volume is more useful in detecting patients with HFpEF than TDI E′-related parameters [42], and subjects with bigger LA show higher risk of developing HF [43]. Clearly, finding an enlarged LA can be secondary to other causes such as mitral regurgitation [44], atrial fibrillation [45], anemia [46], age and obesity, but in subjects with sinus rhythm and without mitral valve diseases, it is mostly related to elevated filling pressures.

On the other hand, in front of a normal volume, it cannot be excluded that the LA function has already begun to deteriorate. In DD, raised LV filling pressures cause a progressive increase in atrial chamber wall, stretch atrial cardiomyocytes, determine natriuretic peptide secretion and lead to LA dysfunction [47]. LA strain shows diagnostic accuracy in this setting. HFpEF patients have lower LA reservoir, conduit and pump function than healthy controls, and strain decreases independently from LA size or history of atrial fibrillation (AF) [48] (Fig. 3). Moreover, PALS is able to distinguish between subjects with DD and those with HFpEF; LA stiffness is typically elevated in patients with DD and even more in HFpEF [49]. LA strain is an important correlate of symptoms’ appearance [50]. Atrial dysfunction has been pointed out, in particular, in patients with new-onset dyspnoea; therefore, it is mandatory to conduct a non-invasive evaluation of strain in this setting in order to guarantee a correct differential diagnosis [51].

Fig. 3
figure 3

Left atrial strain in diastolic dysfunction and heart failure with preserved ejection fraction: comparison of peak atrial longitudinal strain (PALS) between a hypertensive subject (left) and a patient accessing the emergency room because of worsening dyspnoea with following diagnosis of HFpEF (right)

Full size image

Atrial chamber beyond HF diagnosis

Heart failure severity, recurrence and mortality

LA strain has shown a good correlation with functional capacity during effort [52], a decreased peak oxygen consumption (peak VO2) at cardiopulmonary exercise testing [53], NHYA class [50] and NT-proBNP plasma levels [54]. Then, atrial dysfunction could represent a real marker of severity [48]. One of the main goals of HF therapy is to prevent recurrence of symptoms and hospital admissions. The discharge from a HF hospitalization is followed by a readmission within a month in almost 24% of cases [55], and the readmission rate in patients with preserved EF is not less than those with reduced EF forms [56]. Subjects already diagnosed with HF show different clinical courses depending on the LA reshaping: higher volumes and non-reversible remodeling with specific treatments correlate with major recurrence of HF during follow-up [57]. LA volume index is capable to predict hospital admission similarly to LV EF [58]. Also, LA-impaired function independently predicts re-hospitalization and might be useful for HF risk stratification [59]. A recent large meta-analysis demonstrated a strong association between LA enlargement and HF mortality, also after adjustment for EF, PW Doppler pattern, symptom status and age [60]; nevertheless, strain during reservoir phase has appeared to be superior to any other echocardiographic parameter in predicting cardiovascular death [61] also after adjustment for atrial fibrillation, LA volume and LV mass [53].

Major cardiovascular events

In addition to hospitalization and mortality, HF patients are exposed to a high risk of other cardiovascular events, mainly AF, stroke, transient ischemic attack (TIA) and myocardial infarction (MI). LA volume index is a more robust marker of patients’ outcome over and above LA area or diameter [62]. In general, the higher the LA volume, the poorer the prognosis [13, 38, 63], and atrial size should always be considered into the risk stratification of the patient [64]. However, the LA represents an additional prognostic tool in this setting, as strongly demonstrated in an important prospective study, extending previous results [61] (Fig. 4). AF is the most common consequence of LA remodeling: reactive deposition of collagen fibers in the interstitium causes massive wall fibrosis [65], with consequent alterations in normal electrical conduction [66], and then, increasing progressively, determines evolution to permanent form. Assessment of the grade of fibrosis by Gadolinium late enhancement CMR has high costs and side effects [67]. As mentioned before, an inverse relationship has been assessed between the grade of atrial fibrosis detected by 3D CMR and LA strain, so this could be a promising tool that requires further investigations, for a non-invasive and rapid evaluation [30]. Consequently to the onset of AF, these patients have high risk to develop a stroke. PALS has been used to predict atrial stasis [68] and is typically reduced in subjects with history of ischemic events, being independently associated with stroke [69]. Lastly, atrial function evaluated by STE is an important predictor of AF recurrence both after cardioversion [70] and ablation [71, 72].

Fig. 4
figure 4

Left atrial strain and prognosis: Kaplan–Meier event-free survival curves in patients with different values of global PALS [51]

Full size image

Limitations and future perspectives

Assessment of LA function by STE could represent the missing added value for a correct evaluation of patients with DD and HFpEF. Limitations include the lack of dedicated software for the analysis of atrial strain and the presence on market of software supplied by vendors that use different analysis algorithms and may create biases. Moreover, LA shape could be difficult to obtain if the acoustic view of the patient is suboptimal and signal components arising from structures placed around LA may contaminate the feasibility of the analysis. However, application of STE analysis is continuously growing in different systemic fields and heart disease. LA strain could be included in diagnostic algorithm to ensure a correct diagnosis both in DD (Fig. 5) and HFpEF (Fig. 6), in addition to the current standard echocardiographic indexes. The promising results of LA strain obtained in DD and HFpEF should be expanded for diagnostic, therapeutic and prognostic aims. Numerous studies and trials regarding HFpEF therapy and drug efficacy are still ongoing. Precise target values of PALS might help in the assessment of the efficacy of a pharmacological strategy in reducing high filling pressures and in reversing the adverse remodeling process. In prognostication field, extension of previous promising studies regarding the correlation between major cardiovascular events and LA strain will rise accuracy of patient management.

Fig. 5
figure 5

Algorithm for the echocardiographic prediction of elevated left ventricular filling pressures (modified from Nagueh et al. [31])

Full size image
Fig 6
figure 6

Algorithm for the diagnosis of HFpEF in patients presenting with suspected symptoms and/or signs of heart failure

Full size image

Conclusion

The LA has shown to play a major role in patients from DD to HFpEF. Shifting the focus from the LV, it can be understood how the LA is not a passive chamber but it is capable to guarantee a correct ventricular diastolic filling also in advanced stage of dysfunction and, at the same time, to protect pulmonary capillary circulation from raised pressures. In order to carry out all these functions, the LA goes through an important remodeling until its wall structure is so altered that it starts to lose function and typical HF symptoms appear. Assessment of atrial size and, even more, myocardial deformation by STE allows a complete evaluation of patients in many steps, from diagnosis to prognostication, risk assessment and prediction of major events. Despite a great improvement in HFrEF outcome, thanks to target therapeutic strategies, no treatment has been shown to reduce morbidity and mortality in patients with HFpEF [73]. It is now clear how preserving a correct LA function represents a goal in these subjects and both pharmacological and non-pharmacological therapies to invert chamber remodeling should be investigated.