Key points are not available for this paper at this time.
The history of recording atrioventricular (AV) annular motion using ultrasound can be traced back to the 1960s1. The first papers describing the feasibility of recording mitral annular motion using M-mode ultrasonography and demonstrating that these recordings are distinctly different from those of the mitral valve leaflets, and instead resemble left ventricular volume curves, were published in 19672, 3. Furthermore, it was found that the amplitude of mitral annular plane systolic excursion (MAPSE) correlated with stroke volume and was reduced in patients with heart failure3. However, these important observations were somehow forgotten and the technique has not been popular among cardiologists, who mostly prefer using the ejection fraction (EF) to assess systolic ventricular function. In 1932, Hamilton and Rompf4 made two important observations on the basis of classical animal experiments: a) that the pumping function of the heart is determined mostly by the reciprocating caudocephalad movements of the AV ‘septum’ (annulus) and b) that the base of the heart moves towards a relatively fixed apex during systole. The importance of AV annular plane displacement (AVPD) in the longitudinal axis for the pumping function of the heart was later emphasized by Lundbäk5, who developed the concept of a ‘dynamic adaptive piston pump’6. These basic concepts subsequently led to the revival of AVPD as a measure of global cardiac function in adult cardiology7-11. In fetal cardiology, the M-mode technique for measuring AVPD was introduced in 2001 by Carvalho et al.12; in a small pilot study, they showed it to be feasible to evaluate long-axis function of the heart prenatally using anatomical (angular/steerable) M-mode. Gestational age-specific cross-sectional reference ranges were published by Gardiner et al.13 in 2006. However, despite its relative simplicity, the M-mode technique of measuring AVPD for assessing heart function prenatally has not gained popularity. Two related articles on this subject are published in the current issue of the Journal. One of them14 compares tricuspid annular plane systolic excursion (TAPSE) measured by the conventional M-mode method to that measured by spatiotemporal image correlation (STIC) M-mode, whereas the other15 compares the ability of M-mode-derived AVPD to that of pulsed-wave tissue Doppler-derived annular velocities in detecting cardiac dysfunction in growth-restricted fetuses. While these papers are a welcome addition to the scarce literature available on the feasibility and applicability of AVPD in fetal cardiology, it is difficult to predict if these methods will be widely used for evaluating fetal heart function. However, to apply AVPD measurements either in research or in clinical practice, it is important to have an understanding of some physiological concepts and to be aware of the technical strengths and limitations of the different ultrasound modalities and echocardiographic techniques that can be used to assess AV annular plane motion. Although left ventricular EF remains a clinically useful parameter of systolic function in adults and children, it is not particularly popular among perinatal cardiologists as a reliable measure of fetal heart function. EF is significantly preload- and afterload-dependent. In the fetus, due to the parallel arrangement of the circulation, the two cardiac ventricles function under different loading conditions. Therefore, it is important to measure the EF of each ventricle separately, but reliable measurements can be difficult to obtain due to poor resolution of endocardial borders and variable fetal position. Furthermore, the EF reflects mainly radial (short-axis) function of the heart and may remain normal despite overt changes in longitudinal and circumferential myocardial function. Therefore, the search for other reliable non-invasive methods of evaluating fetal cardiac function continues. During the cardiac cycle, the AV annulus moves towards the cardiac apex during ventricular contraction and towards the base during relaxation, while the outer contours of the ventricles remain relatively unchanged. These reciprocating movements of the AV annulus lead to its rhythmic displacement longitudinally. The AVPD is the distance covered by the AV plane between its positions farthest from the apex at the beginning of ventricular contraction and closest to the apex at the end of contraction. The AVPD has been shown to be a major contributor to ventricular pumping5, 16. Movement of the AV annulus towards the relatively fixed apex is a result of longitudinal fiber shortening during myocardial contraction17. Therefore, intuitively, AV plane motion should provide information on the long-axis function of the heart if reliable measurements could be performed. However, several physiological factors have to be considered for appropriate interpretation of AVPD values. In contrast to the EF, which is comparable among healthy individuals irrespective of differences in body size, AVPD is size-dependent18. It has been shown in experimental studies using an animal model that the deformation can be different in different-sized hearts, while the contractility remains the same19. In smaller hearts the annulus has a shorter distance to travel towards the apex compared with in larger hearts. This may explain the observed differences in AVPD values between men and women, which are likely to be a reflection of gender-related differences in ventricular size. It is therefore reasonable to expect AVPD values in the fetus to vary according to gestational age and fetal size. Indeed, the amplitude of displacement has been shown to increase with gestational age in cross-sectional studies using both M-mode13 and color tissue Doppler imaging (TDI)20, whereas AVPD values normalized by left ventricular length seem to decrease with advancing gestation20, but serial evaluation has not been performed and gestational age-specific longitudinal reference values are lacking. TAPSE is higher than MAPSE in fetuses at any given gestation13. This might be because the right ventricle is the dominant ventricle in the fetus. Differences in myocardial fiber orientation and other factors associated with the right ventricular structure and function21 may also be responsible for this observed difference. In healthy adults, longitudinal AVPD accounts for 80% of the right ventricular stroke volume, compared with 60% for the left ventricle16, and TAPSE is approximately 55% higher than MAPSE, with a constant MAPSE/TAPSE ratio of 0.66 which is not affected by body size22. However, the exact relationship in the fetus and whether the ratio changes during fetal maturation is not known. Furthermore, it is important to keep in mind that AV annular motion can be affected by tethering and translational motion from other parts of the myocardium, and it is unlikely that AVPD alone will provide a complete picture of cardiac function. Fetal cardiac dysfunction may be global (e.g. in twin–twin transfusion syndrome, severe placental insufficiency and fetal infections) or regional (e.g. in outflow tract obstructions and Ebstein's anomaly). The AVPD is a parameter of global cardiac function, but deformation imaging is required to assess regional function. Deformation indices can be particularly useful in assessing ventricular rotation and torsion. The study by Cruz-Lemini et al.15 in this issue showed that the AVPD is reduced in growth-restricted fetuses. However, whether AVPD measurements correlate with the severity of growth restriction is not known. In adults, AVPD is known to correlate well with serological markers of cardiac dysfunction, such as NT-proBNP (N-terminal prohormone of brain natriuretic peptide)23, 24. Intrauterine growth-restricted fetuses with severe placental insufficiency are known to have altered biochemical markers of cardiac function25, 26. Whether AVPD values in growth-restricted fetuses correlate with umbilical cord serum markers of cardiac dysfunction, blood gases, acid–base status and/or perinatal outcome needs to be investigated. The motion of any object can be described in terms of displacement, time, velocity and acceleration/deceleration. The AV plane annular motion during a cardiac cycle can be evaluated using a variety of ultrasound modalities, for example M-mode, color or pulsed-wave TDI and two-dimensional speckle tracking, and is usually measured in a four-chamber apical view. In the fetus, the AVPD can be recorded at the mitral, septal or tricuspid region of the AV annulus. One major advantage of using the M-mode technique is its high sampling rate. It is best to position the M-mode cursor through the lateral annulus, parallel to the interventricular septum, keeping the angle of insonation to the long axis of the heart as low as possible and including the tissue–blood border for easier identification of annular motion. However, using M-mode, pre- and postejection phases of the cardiac cycle may be difficult to identify accurately in the absence of a fetal electrocardiogram. As standard M-mode evaluation of AVPD is performed measuring the maximum systolic excursion as a distance between the nadir and the zenith of the annular motion profile (Figure 1), the movements that occur during the isovolumic phases of the cardiac cycle may or may not be included. M-mode is angle-dependent, so obtaining a standard view and insonating at a low angle can be challenging due to variable fetal position. Using anatomical M-mode may help to overcome this problem12, 27. However, it is important to remember that the axial resolution of anatomical M-mode is affected by the lateral resolution of the B-mode image. Real-time and anatomical M-modes correlate well, although anatomical M-mode appears to give slightly higher AVPD values27. Post-processing tools that allow STIC and adjustment of image orientation in a desired plane may also facilitate M-mode measurements in overcoming difficulties related to fetal position14, 28. Two-dimensional imaging and automated measurement of AV annulus motion using a speckle-tracking algorithm is another solution to the problem, as this method is angle-independent and rapid24, 29. It can utilize routinely acquired two-dimensional images for offline analysis, but requires good-quality images with high frame rates30. A number of speckle tracking software packages are available commercially, for example Aloka 2D Tissue Tracking, GE 2D Strain (Automated Functional Imaging), Philips QLAB, Siemens syngo Velocity Vector Imaging (VVI), Toshiba 2D3D Wall Motion Tracking31, that allow calculation of displacement, velocity and deformation. However, these may differ in their performance as they use different tracking algorithms, and none of them is yet validated for fetal application. Matsui et al.30 compared AVPD values measured using standard M-mode echocardiography and the high-frame-rate VVI technique, and found them to be similar. The motion of the AV plane can also be assessed using pulsed-wave TDI32 (Figure 2) or color TDI20 (Figure 3). In their paper, Cruz-Lemini et al.15 compare the diagnostic ability of M-mode-derived AVPD with that of TDI-derived annular velocities to detect cardiac dysfunction in intrauterine growth restriction. Although comparing the diagnostic ability of two different methods of assessing AV plane motion may be acceptable, it is important to be aware of the intrinsic differences related to dimensional heterogeneity when making direct comparisons, as the velocities are measured in cm/s whereas the displacement is measured in cm. However, displacement can also be obtained by integrating the velocity curve (displacement = velocity × time). Therefore, the systolic velocity curve of annular motion obtained during the ejection phase of the cardiac cycle can be traced to measure AVPD, as the velocity time integral (VTI) of the systolic waveform (S′ wave) represents systolic annular displacement (Figure 2). Thus, at least in theory, if one can record reliable myocardial velocities using TDI, AVPD can also be measured, using the same velocity curves avoiding the need for further acquisition with another ultrasound modality. In addition to AVPD, the TDI method allows evaluation of other components of AV plane motion, i.e. longitudinal myocardial velocities, acceleration/deceleration and time intervals of the different phases of the cardiac cycle. The TDI method is also angle-dependent, but has the advantage of delineating clearly the cardiac cycle events. The sampling frequency of TDI can be affected by several factors including the imaging sector angle (width), depth and pulse repetition frequency. However, reasonable sampling frequency can be achieved with online recording of myocardial motion using pulsed-wave spectral TDI33 and can provide information on velocities and time intervals, including events during the isovolumic (pre- and postejection) phases of the cardiac cycle34. However, spectral broadening is a recognized problem with this technique35 and changing offline gain settings can significantly alter the measurements. In contrast to pulsed-wave TDI, which requires tracing of the systolic (S′) wave contour to measure VTI, color TDI allows instantaneous calculation of VTI at any period of the cardiac cycle using tissue tracking, and displacement curves can be generated automatically by integration of color TDI velocity data over time (Figure 4). The effect of temporal filtering on color TDI measurements is insignificant, which may increase its clinical applicability. However, as the temporal resolution is 10 ms for a frame rate of 100 frames/s and 5 ms for a frame rate of 200 frames/s, sampling frequencies of > 150–200 frames/s are needed to measure different cardiac cycle events reliably using color TDI36. Therefore, it is important to reduce the ultrasound sector width while recording color TDI images in order to achieve the highest possible frame rates. Another way of measuring AVPD is by overlaying the M-mode cursor on the color TDI image (real-time or a stored cineloop) of the fetal heart in an apical four-chamber view. M-mode color-coded tissue imaging has high spatiotemporal resolution and may allow better identification of changes in direction of AV plane movements due to color-coding of the tissue motion (Figure 5). Left ventricular AVPD measured using M-mode, temporal integration of pulsed-wave TDI-derived systolic waveform and color TDI-derived tissue tracking correlate, but effective concordance between these methods was found to be suboptimal, in adults37. Therefore, these techniques might not be interchangeable. Whether AVPD measurements in the fetus using different ultrasound modalities give similar results needs to be investigated. Which method is most accurate, valid and reliable is also not known. Furthermore, as the AV plane systolic excursion is size-dependent, it is important to use normalized AVPD values rather than absolute values. As fetal heart size may not always correlate with gestational age (e.g. the heart may be large in proportion to body size in growth-restricted fetuses due to relative cardiomegaly), it is better to normalize AVPD values by heart size rather than gestational age or estimated fetal weight. Ventricular base-to-apex longitudinal diameter in end-diastole32, biventricular diameter or cardiac circumference/area could be used for this purpose. However, which parameter is best remains to be evaluated. AVPD is a measure of ventricular long-axis function and can be evaluated in the fetus using a variety of echocardiographic methods. Although some of these techniques may be relatively simple to learn and use, they lack experimental and clinical validation for fetal application. Furthermore, whether there is effective concordance between different methods and whether AVPD correlates with other recognized markers of fetal cardiac dysfunction remains to be investigated. Longitudinal studies investigating physiological maturational changes throughout gestation are needed before the usefulness of AVPD to predict or diagnose fetal cardiac dysfunction can be determined.
Ganesh Acharya (Mon,) studied this question.