TTE is a reliable method to assess the anatomy of the LA. The measurement of LA anteroposterior diameter from M-mode echocardiography in the parasternal long-axis or short-axis window is commonly utilized in clinical work. This conventional method has a tendency to underestimate true LA size, because the shape of the LA is not spherical and enlargement of the LA is often asymmetrical.6) Hence, 2D derived LA volume assessment using the biplane area-length method or the Simpson's method provides more accurate measures of LA size.7) A horizontal line needs to be drawn across the mitral annular plane, and the LA area does not include the funnel of the mitral valve leaflets (Fig. 1). When tracing the outline of the LA, the LAA and the confluence of the PVs are excluded from the measurement. The reference value for 2D echocardiographic maximum LA volume indexes is 22 ± 6 mL/m².7)8) A LA volume index of 29 mL/m² is usually considered the upper limits of the normal value.7) More recently, three-dimensional (3D) echocardiography for assessing LA volume has been available and has demonstrated good correlation with biplane 2D measurements.9) However, studies on 3D echocardiographic measurement of LA volume have been too small, and there is no consensus on the methods or comparisons with established normal values. LA size has been established as a prognostic marker for both adverse cardiovascular events as well as overall outcomes.10-12) According to the data from the Framingham study, a 5-mm increase in LA dimension was associated with a 39% increased risk for subsequent development of AF.4) Psaty et al.13) demonstrated that subjects in sinus rhythm with a LA dimension over 5.0 cm had about four times the risk of developing AF during the following period of surveillance. Increased LA volume, likewise LA dimension, is predictive of subsequent AF11) and it has been confirmed that LA volume is a superior measure over LA dimension for predicting outcomes of AF.14) LA enlargement is associated with a low probability of successful cardioversion for chronic AF or maintenance of sinus rhythm.15)16) An increased LA volume index can predict recurrence of AF after RFCA.17) Although AF promotes further LA enlargement, cardioversion and maintenance of sinus rhythm may reverse this process. Van Gelder et al.18) reported that after restoration and long-term maintenance of sinus rhythm, echocardiographic atrial dimensions may decrease. Reant et al.19) also demonstrated that RFCA for isolated AF reverse morphological remodeling of the LA after restoration of sinus rhythm.

In addition to providing information about atrial anatomy, echocardiography is also a potent tool for assessment of atrial function. With the loss of atrial booster pump function in AF, the LA-LV pressure gradient during early LV filling is increased by elevation of the LA pressure to maintain stroke volume.20) Thus, a reduction in both LA compliance and volume has been observed with the onset of AF that further decreases cardiac function and increases the risk of thromboembolism. The change of the LA function in different phases can be evaluated non-invasively by echocardiography, utilizing not only usual methods including transmitral flow and changes in LA area and volume, but also novel techniques such as tissue Doppler imaging (TDI) and strain imaging.

Transmitral inflow patterns by pulsed wave Doppler provides information of LA mechanical function. The peak transmitral inflow velocity of the late diastolic filling wave (A) is commonly used as a measure of LA mechanical function,21) but it cannot be measured in AF because of the absence of the atrial waveform. The peak A wave velocity has been used in the serial follow-up of LA function in patients with AF after cardioversion22) and RFCA.23) The time-velocity integral of the A wave and the atrial fraction (i.e., the ratio of the time-velocity integral of the mitral A wave to that of the total diastolic transmitral flow) may be used as a parameter of LA function.24)

TDI allows the quantification of the low-velocity, high-amplitude, long-axis intrinsic myocardial velocities in both systole and diastole, and provides a relatively load-independent measure of both LV systolic and diastolic function.25) The peak velocity at the mitral annulus in late diastole, following atrial contraction (A' or Aa velocity) can be utilized as a rapid and accurate marker of atrial function.26) The A' velocity correlates with other parameters of atrial function such as peak A velocity of mitral inflow, atrial fraction and atrial ejection force.26) Another study also demonstrated that the tissue Doppler A' velocity correlates with LA fractional area and volume change.27)

Doppler strain and strain rate imaging are derivatives of tissue Doppler velocities. TDI quantifies global tissue motion velocity whereas strain and strain rate represent the extent of local tissue deformation and its rate, respectively (Fig. 2). Strain measures the myocardial deformation during a cardiac cycle, and strain rate measures the tissue velocity gradient within the myocardium. Strain and strain rate imaging may overcome the major limitation of TDI, and may be largely independent of translational effects due to tethering by neighboring myocardial segments.

Using TDI and/or strain imaging techniques, it was found the decreased compliance of LA walls, the impairment of the reservoir and conduit function of LA, and the loss of the booster pump function in patients with AF.28)29) With the degree of LA enlargement, LA mechanical dysfunction has emerged as an independent predictors for the maintenance of sinus rhythm after successful cardioversion29) and RFCA.30)

Atrial stunning is characterized by reduced atrial mechanical function after restoration of sinus rhythm from AF, which may last several weeks and which is associated with increased risk of thromboembolism. This phenomenon is well recognized with peak A velocity of transmitral inflow22)24) as well as TDI or strain imaging.31) This suggests that a dissociation of electrical and mechanical recovery occurs after successful restoration of sinus rhythm, with a delay in gradual improvement of atrial mechanical function.

These parameters, mentioned above, can only be easily measured in sinus rhythm and are often not applied to the state of AF. Other LA functional parameters, which are rhythm independent, have been recently suggested as being required.32) More studies and trials to be accepted and validate these parameters are needed to use routinely in clinical practice.

The risk of systemic emboli, probably arising in the LA cavity or LAA as a result of circulatory stasis, is an important consideration in AF. AF is considered to be responsible for almost half of all cardiogenic thromboembolism. One of the primary goals of management of patients with AF is the prevention of thromboembolic events. TTE has a low sensitivity for detection of thrombi in LA and especially LAA.51) TEE provides excellent visualization of posterior cardiac structures and is the modality of choice for detecting LA or LAA thrombi with a sensitivity and specificity of approximately 95% to 100% (Fig. 8).52) Thrombus identification is also challenging even if the appendage is visualized adequately. It is difficult to distinguish small thrombi from artifacts, including prominent trabecular structures, duplication artifacts, and adipose tissue within the transverse sinus. It is necessary to attempt to differentiate any suspicious abnormalities from thrombus in multiple views.

In the absence of formed thrombi, a dense SEC has been demonstrated to be a strong predictor of ischemic strokes (Fig. 9). When blood flow velocities are reduced in cardiac chambers and especially in the LAA, "smoke-like" echoes swirling in the cavity may be seen. An annual thromboembolic event rate of 12% has been observed in patients with SEC compared to 3% in patients without it.53) The finding of thrombus on TEE represents a poor outcome. In one study, it was associated with an embolic risk up to 10.4% per year and a death risk of 15.8% in a series of patients, most of whom were receiving anticoagulation therapy.54)

LAA was sometimes regarded as just a minor extension of the atrium. However, the importance of its function has been brought up because of its contribution to formation of thrombus. The mechanical function of LAA is best assessed with TEE utilizing pulsed wave Doppler measurement of LAA emptying and filling velocities (Fig. 10). In patients with sinus rhythm, LAA contracts once per cardiac cycle, and flow velocities at the ostium demonstrate a biphasic pattern with peak velocities generally exceeding 50 cm/s.55) Low LAA emptying flow velocities (< 20 cm/s) in AF correlate strongly with the presence of SEC and thrombus formation,56) whereas LAA emptying flow velocities > 40 cm/s predict greater likelihood to remain in sinus rhythm 1 year after successful cardioversion.53) For patients with AF, TEE risk factors of clinical thromboembolism include LAA thrombus, severe SEC, decreased (< 20 cm/s) flow velocities at LAA ostium, and complex aortic plaque (Table 1).56)57) A TEE performed 3 to 6 months after AF ablation can also evaluate thromboembolic risk and need for long-term anticoagulation, as echocardiographic risk factors may be present even if restoration of sinus rhythm is successful.

In addition to LAA Doppler assessment, measuring LAA area and ejection fraction, TDI, and 3D TEE are emerging as the new complementary techniques.58-60) These are performed rarely, however, and lack of data suggesting additional prognostic value beyond simple flow velocity assessment.

One of the primary goals of therapy in AF is the restoration and maintenance of sinus rhythm, with the aim of reducing symptoms and thromboembolic events. Cardioversion of AF to sinus rhythm can be performed with high rates of immediate success but unfortunately it may be complicated by embolic events in up to 7% of patients if the procedure is not preceded by several weeks of anticoagulation.61)62) Conventional care of patients with AF of prolonged (more than 48 hours) of unknown duration had required that these patients should be treated with oral anticoagulation for at least 7 weeks (3 weeks before and 4 weeks after cardioversion).63) The conventional approach to antithrombotic management before trying cardioversion has several potential disadvantages. This has promoted the recent suggestion of alternative approaches, such as TEE-guided early cardioversion. The advantages of TEE-guided early cardioversion with short term anticoagulation over the conventional strategy include the following: 1) In patients with no thrombus identified by TEE, the total duration of anticoagulation can be reduced by weeks, potentially reducing the risk of bleeding. 2) A shorter duration of AF before cardioversion is a strong predictor of AF recurrence following cardioversion. Therefore, early cardioversion with a TEE-guided approach might prevent the atrial remodeling and enable higher rates of sinus conversion and maintenance of sinus rhythm. 3) A TEE-based strategy could lower total costs by limiting laboratory tests, reducing preprocedural bleeding events, and improving clinical outcomes (Table 2).

However, thromboembolic events following cardioversion from AF were documented among patients with absent or subtherapeutic anticoagulation despite a negative TEE for thrombus,64) possibly caused by tiny thrombi that are undetected at TEE or as a result of new thrombus formation caused by LA/LAA stunning after cardioversion.24) Therefore, in current practice, TEE-guided early cardioversion includes screening for an existing thrombus and therapeutic anticoagulation at the time of TEE, as well as post-cardioversion anticoagulation prophylaxis to avoid new thrombus formation because of LA or LAA stunning (Fig. 11). The assessment of cardioversion using transesophageal echocardiography (ACUTE) multi-center randomized study demonstrated that embolic events and maintenance of sinus rhythm were similar in the two treatment strategies (conventional vs. TEE-guided early cardioversion), although the rate of hemorrhagic events was lower with the TEE-guided strategy.65) A new antithrombotic strategy of TEE-guided cardioversion with low molecular weight heparin has been suggested. The ACUTE II pilot study66) and randomized trial,67) which compared short-term antithbombotic therapy using low molecular weight heparin (LMWH) with intravenous unfractionated heparin (UFH), showed similar safety and efficacy with a lower length of hospitalization and more sinus rhythm in the LMWH group. In concert with precardioversion therapeutic anticoagulation with UFH, LMWH, or warfarin, TEE-guided early cardioversion is comparable to 4 weeks of precardioversion warfarin with a risk of clinical thromboembolism < 1%.65-67)