Assessment of atrial regional and global electromechanical function by tissue velocity echocardiography: a feasibility study on healthy individuals

The enlargement of left atrial (LA) diameter is associated with cardiovascular disease and is a risk factor for atrial fibrillation, stroke and death. [1‐6] LA function reliably predicts exercise capacity in patients with recent myocardial infarction[7] or non-ischemic dilated cardiomyopathy[8] and differs in patients with ischemic and dilated cardiomyopathy.[9] Moreover, LA volume is an independent prognostic factor in several subsets of patients. [10‐12] Although commonly used, LA size assessed by M-mode echocardiography does not correlate well with LA volumes, so several methods to estimate LA volumes have been developed.[13, 14] The LA reservoir, conduit and pump functions may be estimated from volume measurements. [15‐19] However, the reliability and clinical usefulness of those methods have been poorly studied. Pulsed-wave Doppler interrogation of the blood flow velocity during atrial contraction, the peak mitral inflow A wave, and its velocity time integral have also been used as surrogate markers of atrial function. [20‐22] These variables represent the diastolic properties of the LV [23‐25] and do not accurately reflect atrial mechanical properties. Tissue velocity echocardiography (TVE) has now been developed as a valuable tool for the evaluation of left and right ventricular systolic and diastolic functions.[26, 27] Furthermore, this technique has also been used to assess the regional functions of the left and right atrium.[28, 29] Although atrial anatomy was described more than a century ago, a new interest in atrial anatomy and its relations with atrial electromechanical function has only recently emerged.[30, 31] Conventionally, atrial electrical function has been evaluated from resting electrocardiography (ECG), and more accurately by invasive electrophysiology techniques. [32‐34] The rapid development of these invasive techniques has improved not only diagnostic capabilities,[35, 36] but also our understanding of how the electrical impulse spreads through atrial tissues,[37, 38] and has led to improvements in the treatment of supra-ventricular tachy-arrhythmias. [38‐40] However, the invasive nature and the high costs of these procedures limit their widespread use and repeatability. Therefore, the development of noninvasive, safe, accurate and repeatable methods that might provide similar information is necessary. We aimed here to find simple and repeatable methods to assess both electrical and mechanical regional atrial functions by means of TVE.

All demographic features and measures of standard 2-D and pulsed-wave Doppler echocardiography data are shown in Table 1. Of interest, no differences were found between measures of RA and LA functions, as assessed by short or long axes, or among volumes and volume-derived indices. TVE-derived variables assessing the RV and LV long-axis systolic and diastolic functions are shown in Table 2. No significant differences among LV walls were found for any index of systolic and diastolic function. TVE-derived variables obtained from the RV free wall were significantly different from each LV wall.

Table 3 shows several time intervals. The PA-start interval (P-Aa' start) was longer at low atrial levels in each atrial wall than at the mid atrial level and shorter at the RA than for all LA walls (Fig. 3). Some statistical significant differences among different LA walls were also found. The PA-peak interval (P-Aa' peak) was similar at low and mid atrial levels in almost all atrial walls with exceptions in the inferior and posterior LA walls. This interval was shorter for the IAS, inferior and posterior LA walls than for the lateral and anterior LA walls at mid and low levels (Fig. 4). The A-wave duration (Aa'-dur.) was shorter at the low atrial level than at the mid atrial level in each atrial wall, but not in the inferior and posterior LA walls. The total electromechanical activity (TEMA) was similar in all RA and LA atrial walls measured at low and mid levels, and no differences were found between any of the LA walls.

Table 4 shows several velocities and velocity-derived variables: The peak velocity during atrial contraction (Aa' peak vel.) was higher at low than mid levels in each atrial wall, but no significant differences were found between any LA walls. This variable was higher in RA than in all LA walls. Similar results were found for the atrial displacement occurring during atrial contraction (Aa' disp.) and the ratio of atrial displacement measured at atrial level to the total LV myocardial displacement measured at ventricular level (Aa' cont.). The strain rate (Aa' SR) was higher in the RA than in all LA-walls, and lower in the IAS than in the lateral and posterior LA walls. The strain (Aa' S) was higher in the RA than in all LA-walls, and no differences were found between LA walls.

Tables 5 and 6 show the correlation coefficients between 2-D-derived and TVE-derived variables of LA and RA global mechanical function. There were no correlations between 2-D- and TVE-derived variables, apart from modest correlations between LA diameter and LA displacement, between RA long axis diameter and RA displacements, and between RA ejection fraction and strain rate.

The inter- and intra-observer measurement error and repeatability, as expressed by the British Standards Institution guidelines and coefficients of variation are presented in Table 7. The PA-start interval and the PA-peak interval, which mainly express atrial electrical function showed the largest inter- and intra-observer measurement errors and variability. However, the A-wave duration and the total electromechanical activity, which express a combination of the atrial electrical and mechanical functions, had better values of measurement error and repeatability. The same was true for all the TVE-derived variables that express regional and global atrial mechanical function.

The main new findings of this study of healthy young individuals are as follows. (1) Some TVE-derived variables indirectly reflect the atrial electrical activation that follows the known activation process as revealed by invasive electrophysiology. (2) The regional and global atrial mechanical function is explained by an upward movement of the atrial walls at the region near the A-V ring with a continuous reduction of this movement towards the upper levels of atrial walls. (3) The atrial mechanical function is quite similar in all LA walls; however, all indices of mechanical function were higher in the RA than in the LA. (4) There were no correlations between the 2-D- and TVE-derived variables expressing atrial mechanical function. (5) Values of measurement error and repeatability were good for atrial mechanical function, but only acceptable for electrical function.

Atrial electrical activation, as assessed by the PA-start interval, began at the RA and followed through the IAS, to the inferior and posterior LA walls. This is the known normal electrical activation process, as obtained by invasive electrophysiology techniques.[32, 37] In the present study, there were no statistical significant differences in the PA-start interval between IAS and the inferior and posterior LA walls, indicating that the activation process could indistinctly occur through any of these walls, as demonstrated by the presence of preferential conduction pathways nearby the IAS, the posterior LA wall and the coronary sinus.[33, 34, 37]. In a recent study, using M-mode color tissue Doppler registrations of the tricuspide and mitral rings, an abnormal time interval from the onset of P wave until the backward motion of the left atrio-ventricular ring was used to indirectly detect abnormal atrial electromechanical coupling in patients with paroxysmal atrial fibrillation.[43]

The relation between atrial anatomy and its mechanical function has been poorly studied. The present study showed that all atrial walls actively moved upwards from the region of the A-V ring at late diastole, with a reduction of this movement towards the upper parts, thus empting the atria and contributing towards the last part of filling of the LV. This longitudinal movement of the atrial walls is probably related to the longitudinal endocardial muscular fibers along the walls of the LA and RA. The more pronounced longitudinal movement in the RA may be explained in part by the larger pectinate muscles in the RA, but also by the lower pressures in the heart's right side. To what extent circumferential contraction of the atrial muscle fibers might contribute to atrial mechanical function is unknown. Anatomically, the large amount of circumferential muscle fibers present in the vestibules of the RA and LA[30, 31, 44] might imply some kind of circumferential or radial contraction of the atria. However, no movement of the posterior LA wall at late diastole can be observed by conventional M-mode echocardiography. Other circumferential fibers, such as Bachman's bundle located at the subepicardium joining the RA and LA, seem to play a critical role for electrical impulse spreading[37] rather than in circumferential atrial contraction. The assessment of circumferential atrial mechanical function by conventional echocardiography and TVE remains elusive.

No correlations were found between 2-D- and TVE-derived variables of atrial mechanical function, as was also found in a previous study[29]. Although 2-D-derived variables measure volumes and volume-derived indices that might indicate some kind of atrial mechanical force, it was surprising to find no correlations between the variables obtained by the two different techniques. This might indicate that the velocities and the displacements registered from all atrial walls by TVE are less dependent on volume loading conditions than 2-D-derived variables and therefore could be used as reliable measurement of pure atrial mechanical contraction or inotropism. In fact, Donald et al. showed that LA function assessed by TVE was relatively independent of LV function.[45] It should also be considered that movements of the heart not related to atrial contraction might partly contribute to the velocities and displacements registered from all atrial walls. Therefore, 2-D- and TVE-derived variables might not be used interchangeably to assess atrial mechanical function.

Some measures of atrial electrical function, for example the PA-start interval and the PA-peak interval, had only fair measures of repeatability and measurement error. However, most of the TVE-derived variables expressing atrial mechanical function had good values of repeatability and measurement error. Assessing atrial mechanical function by measuring volumes is time-consuming and depends on age, gender, and body surface area[14, 19] In addition, atrial volume indices are also dependent on loading conditions[46, 47] and are not necessarily more reproducible than TVE-derived variables.

The identification of an abnormal electrical activation process could be of interest in some patients with atrial fibrillation or other supra-ventricular tachy-arrhythmias, in whom the premature atrial contraction acting as a triggering factor could be aggravated by local delayed conduction (reviewed in[48, 49]). Further refinement of the TVE technique are necessary not only to identify the mechanical activation atrial sequence during normal sinus rhythm, but also to identify the origin and the activation sequence of supra-ventricular ectopic beats and in patients with RA, IAS or bi-atrial pacing. Thus, TVE could be an excellent adjunct to invasive electrophysiological techniques in selecting adequate patients and in the evaluation of atrial electromechanical consequences of RA, IAS or bi-atrial pacing.

The assessment of pure mechanical atrial function by means of atrial wall movements may give more concrete clues about the recovery process of atrial electromechanical function after conversion for atrial fibrillation and flutter and can give additional pathophysiological insights on the thromboembolic process that occur in some of those patients.[50] TVE-derived parameters may also give additional pathophysiological information on the process of atrial electromechanical remodeling that occurs in patients with sustained supra-ventricular tachy-arrhythmias.[51]

Several studies have shown the independent prognostic value of atrial function measurements in subsets of patients.[6, 11, 12] TVE-derived variables of atrial mechanical function may have an additional role for facilitating the assessment of atrial function and consequently in the process of risk stratification.

The results of the present study refer only to a group of young healthy individuals and the values for each of the studied variables are, therefore, only applicable to that population group. As discussed, the measures of atrial electrical function showed only fair values of repeatability and measurement error. There were two reasons: the image acquisition rate (less than 100 frames per second) means an implicit measurement error of 10 ms; it was also difficult to identify the beginning of the P-wave in the ECG from the monitor in the echocardiography machine. Improving temporal resolution by image acquisition at more than 200 frames per second, and improving and adjusting the ECG quality in the present equipment may help solve or decrease this problem. The velocities and displacements registered by atrial walls do not only represent the process of atrial contraction, but also the translational movement of the heart. Until now, no appropriate algorithms that correctly deal with this problem have been found. Presently, it is not possible by means of TVE to simultaneously record the electromechanical function of all atrial walls in one heartbeat. The development of three-dimensional TVE may help resolve this difficulty.

TVE is a noninvasive bedside tool that requires further refinements to provide reproducible, repeatable and potentially clinically useful data on atrial electromechanical function in health and disease.