Assessment of Left Atrial Size

A wealth of imaging and hemodynamic data have documented that the left atrium (LA) is not only a simple conduit for left ventricular filling; it also represents a critical structure for overall cardiac performance. It not only acts as a contractile pump that delivers 15% to 30% of the entire left ventricular filling volume but also acts as a reservoir that collects pulmonary venous return during ventricular systole allowing passage of the atrial stored blood into the left ventricle (LV) during early ventricular diastole. Given its important physiological role, changes to normal LA size or function have been associated with adverse cardiovascular outcomes. Previously, only LA maximal size was considered to be a clinically relevant prognostic marker. Recently, both LA minimum volume and phasic function parameters have been reported to be powerful predictors of outcome in various cardiac conditions. LA volume is a surrogate marker of the severity and chronicity of diastolic dysfunction. Additionally, maximal LA volume (LAV _max ) is a biomarker for adverse cardiac events in healthy individuals and in various cardiovascular conditions, including myocardial infarction, heart failure, stroke, degenerative mitral regurgitation, and atrial fibrillation (AF). Echocardiography, traditionally two-dimensional echocardiography (2DE), and more recently three-dimensional echocardiography (3DE), is the most commonly used noninvasive imaging modality to evaluate LA size. ^, Measurement of LA size is best accomplished with transthoracic echocardiography, which allows complete visualization of the LA. With transesophageal echocardiography, the LA cannot be fit in the image sector thereby precluding the ability to measure its actual size.

Historically, the anteroposterior (AP) LA diameter by M-mode or 2D images from the parasternal long-axis view was used to estimate LA size ( Fig. 38.1 ). From this approach, the AP dimension of the LA can be measured. Because this measurement is easy to perform and highly reproducible, it has become the most frequently used metrics of LA size by echocardiography laboratories worldwide. However, LA AP diameter underestimates LA size. LA enlargement is asymmetrical, occurring in the mediolateral and superoinferior axes, with relatively limited enlargement in the AP dimension because of the constraints of the spine and the sternum ( Fig. 38.2 ). LA AP dimension identified only 49% of patients with an enlarged LA versus 76% identified by evaluating LA volume. Accordingly, the use of the LA AP diameter to estimate LA size is currently discouraged by the guidelines except in patients with hypertrophic cardiomyopathy in whom this parameter is included in the score to stratify the risk of sudden cardiac death.

Open full size image

Leading edge-to-leading edge linear measurement (dotted blue line) of left atrial anteroposterior diameter using two-dimensional (left) and M-mode (right) echocardiography. Ao, Aorta; LA, left atrium; LV, left ventricle; RVOT , right ventricular outflow tract.

Open full size image

Schematic showing that the anteroposterior dimension of the left atrium is constrained between the sternum and the spine and therefore the largest expansion may only occur in the superoinferior dimension (red dashed double arrow) . Normal left atrium with three-dimensional reconstruction of actual volume size (bottom left) and a dilated left atrium in the same patient (bottom right) showing the change in left atrial size and shape.

Biplane LA volume by 2DE is the currently recommended measurement to evaluate LA size because it is a stronger predictor of outcomes than linear dimensions. However, accurate measurement of LA volume by 2DE requires dedicated acquisitions of apical views optimized for the LA with proper endocardial border tracing. The long axis of the LV is not parallel to the long axis of the LA, and this has been more clearly demonstrated with the adoption of 3DE for LA assessment ^, ( Fig. 38.3 ). Accordingly, measurements should be obtained from dedicated apical four- and two-chamber views. In the correct imaging plane and time of the cardiac cycle, the base of the LA should be at its largest size, indicating that the imaging plane passes through the maximal short-axis dimension. The LA length should also be maximized to ensure alignment along the true long axis of the LA and the length of the LA long-axes measured in the two- and four-chamber views should be similar.

Open full size image

Apical views of the heart obtained from a three-dimensional echocardiography dataset to illustrate the fact that the long axis of the left ventricle (red pointed line) and left atrium (yellow pointed line) do not lie in the same plane. Left atrial (LA) size displayed in the four-chamber view optimized for the left ventricle (4CH) is significantly shorter than the LA size displayed in the two-chamber (2CH) and apical longa-axis (LAX) views.

Measurements should be taken at the end of left ventricular systole because this is when the LA chamber is at its greatest dimension. When tracing the endocardial border, the LA appendage, the confluence of the pulmonary veins and the space between the mitral valve leaflets and annulus should be excluded from LA volume measurements ( Fig. 38.4 ). Phasic LA volumes can be calculated by measuring LA volumes at various times of the cardiac cycle: maximal LA volume is measured just before mitral valve opening, LA pre A volume at the onset of the P wave on the electrocardiographic tracing, and minimal LA volume at end-diastole (before mitral valve closure) ( Fig. 38.5 ). Parameters of phasic LA function (active and passive emptying volume and fraction and conduit volumes) are calculated from these volumes ( Table 38.1 ).

Open full size image

Two-dimensional measurements to obtain the left atrial volume using the apical four- (left) and two-chamber (right) views (see text for details). Using the same endocardial border tracings, the left atrial volume obtained with the area–length method (LAESV A-L) was larger than the one obtained with the Simpson’s (MOD BP) algorithm. LAA, Left atrial appendage; LLPV, left lower pulmonary vein; LUPV, left upper pulmonary vein; RUPV, right upper pulmonary vein.

Open full size image

Phasic left atrial volumes. From the top: spectral Doppler of left ventricular filling, electrocardiography tracing, three-dimensional left atrial surface and volume–time curves to show the time (red lines) , and volumes of the left atrium at left ventricular end-systole (LA Vmax), at end-diastole (LA Vmin), and before the P-wave on the EKG (LA VpreA).

Using 2DE, biplane LA volume can be measured using two algorithms: the modified Simpson’s method of disk summation and the area–length method. With the Simpson’s method, the LA endocardial border is traced and volume computed by adding the volume of a stack of 20 cylinders of height equal to L/20 (in which L is the length of the LA) and bases calculated by orthogonal minor and major transverse axes (a _i and b _i ) assuming an oval shape

Alternatively, a biplane calculation could also be performed using the area–length algorithm with LA areas and lengths obtained from both the apical four- (A1) and two-chamber (B1) views. LA volume is calculated as:

The length (L) of the LA is the shortest distance between the midline of the plane of the mitral annulus to the opposite superior side (roof) of the LA measured in either the four- or two-chamber views. To avoid significant miscalculations caused by foreshortening of one of these two views, the difference between L measured in the two- and four-chambers views should be less than 1 cm. Although the area–length method still assumes an ellipsoidal LA shape, it has the advantage of reducing linear dimensions to a single measurement. Both methods have been reported to be accurate compared with measurements obtained using computed tomography. The biplane area–length method systematically yields larger LA volumes than the disk summation method. ^, However, both methods have comparable prognostic power.

Body size is a major determinant of LA size, with absolute LA volumes being larger in men than in women. However, indexation of LA volumes to body surface area (LAVI) leads to similar values between men and women and corrects for the effect of gender. ^, Reference values for LA volumes derived from 2DE have been similar in population-based studies and in normative studies of healthy volunteers. The effects of healthy aging on LA volume have been controversial, with some studies reporting an increase in LA volumes only at extremes of age and others demonstrating a progressive age-related increase in LA volumes. Differences in LA size according to ethnicity suggested larger LA size for Europeans compared with South and East Asians.

The threshold value to diagnose LA enlargement by 2DE has recently been revised in the 2015 European Association of Cardiovascular Imaging/American Society of Echocardiography guidelines for chamber quantification. Thus, the previous cut-off of maximal LA volume (LAV _max ) greater than 28 mL/m ² has been raised to 34 mL/m ² based on pooled data coming from larger cohorts of healthy participants. This revision aligns the “enlarged LA” cut-off value with the one recommended in the algorithm to diagnose LV diastolic function. Moreover, the revised value of greater than 34 mL/m ² used to define LA enlargement, is perhaps more clinically relevant because a LAV _max greater than 32 mL/m ² was associated to adverse outcomes in ischemic stroke, diabetes, and heart failure. ^, Finally, the revised threshold value to define an enlarged LA allowed a reclassification into normal LA size for 21% of patients previously reported as having an enlarged LA, without any loss of the prognostic power associated with an enlarged LA.

The problem with currently recommended partition values (mild LA enlargement – LAV _max = 35–41 mL/m ² , moderate LA enlargement LAV _max = 42 to 48 mL/m ² , and severe LA enlargement – LAV _max >48 mL/m ² ) is the narrow range of the varying grades of LA volume enlargement. Thus, even small measurement errors can result in misclassification of the grade of LA enlargement. The recent multicenter Normal Reference Ranges for Echocardiography (NORRE) study, which included 734 healthy individuals, suggested upper normal limits for LAV _max to be even larger than the current 34-mL/min cut-off (42 mL/m ² using the area–length method and 37 mL/m ² using the Simpson’s method).

2DE LA volume correlates with LA volumes obtained using 3DE, CT, and cardiac magnetic resonance (CMR), with 2DE demonstrating a systematic underestimation of LA volumes. ^, This is likely caused by some foreshortening of the LA in the absence of dedicated acquisitions which maximize the LA long axis. Moreover, 2DE is associated with more difficult endocardial border definition, particularly in the two-chamber view, with consequent suboptimal accuracy of 2DE method. Despite these limitations, the ease of use and wide availability of 2DE makes it a clinically powerful tool, and it boasts the largest body of evidence on the alterations in LA volumes, as well as on its prognostic value (discussed in detail later).

3DE is fast becoming the modality of choice to measure cardiac chamber volumes. Its lower interobserver variability and higher test–retest reproducibility compared with 2DE make 3DE particularly important and the preferred method for serial measurements. With the most recent technological advances, 3DE datasets of the LA are easily obtained with acceptable frame rate using single-beat acquisition. Regarding 3DE work flow, both semiautomated and fully automated contour detection of 3DE datasets have shown good correlation with manual tracing methods, with significant reduction in analysis times and increase in measurement reproducibility. In a multicenter study, the agreement for classification of an enlarged LA using a cut-off of 34 mL/m ² had a κ coefficient of interrater agreement of 0.88 (4 false negatives and 7 false positives) between 3DE LA volume and CMR compared with a κ coefficient of 0.71 (25 false negatives and 2 false positives) for 2DE. Although Simpson’s method of disks was previously applied, more recently, the 3DE speckle-tracking and pattern recognition methods have been developed to achieve a fully automated identification of endocardial border to measure LA volumes, though these newer algorithms require further clinical validation (see Fig. 38.5 ).

3DE LA volumes correlate better than 2DE with the volumes obtained by either CT or CMR ^, because 3DE includes no geometrical assumptions regarding LA chamber shape. Moreover, 3DE-derived LA volumes have demonstrated better reproducibility than those derived from 2DE. Two studies sought to define reference values of 3D LA volumes by adapting 3DE software algorithms developed for the LV to measure the LA. ^, A recent study of 276 of healthy participants used a 3D software package specific for the LA, showing that 3DE LA phasic volumes were significantly larger than those obtained by 2DE ¹⁵ ( Table 38.1 ). Similar to 2DE LA volume, indexation to body surface area of 3DE LA volumes eliminated gender differences, and a small yet significant increase in 3DE LA volume was observed with aging.

3DE may be the way to measure LA volume in the future because novel fully automated, dedicated software packages are available on high-end ultrasound systems ( Fig. 38.6 ), which will allow more robust test–retest reproducibility for serial measurements. The availability of fully automated quantitation algorithms using single-beat volume datasets allows a reliable LA 3DE quantification even in patients with AF. The major limitations at present are the limited spatial resolution of 3DE datasets and the relative paucity of data for both normative values, as well as prognostic value of 3DE LA volumes and phasic function indices.

Open full size image

Automated measurements of the LA volumes using three-dimensional echocardiography. Top, Auto LA Q (GE Vingmed) that provides also automated measurements of longitudinal and circumferential strain, in addition to volumes. Bottom, HeartModel (Philips Medical Systems). EF, Emptying fraction; EV, emptying volume; LA, left atrial; LASr, left atrial longitudinal strain reservoir; LAScd, left atrial longitudinal strain conduit; LASct, left atrial longitudinal strain contraction; LASr_c, left atrial circumferential strain reservoir; LAScd_c, left atrial circumferential strain conduit; LASct_c, left atrial circumferential strain contraction; LV, left ventricular; Vmax, maximal volume; Vmin, minimal volume; VpreA, volume before atrial contraction.