Digital Subtraction High-Frame-Rate Echocardiography in Detecting Delayed Onset of Regional Left Ventricular Relaxation in Ischemic Heart Disease
Background Because left ventricular (LV) diastolic function is impaired before systolic function in patients with ischemic heart disease and because ischemic heart disease is constituted of regional rather than global abnormalities of the left ventricle, measures of LV regional diastolic dysfunction, if possible, should provide the most sensitive assessment of the coronary involved region. The objectives of this study are to clarify whether high-frame-rate two-dimensional echocardiography, combined with digital subtraction image processing, may be used to visualize regional LV relaxation abnormalities in patients with ischemic heart disease and to clarify whether this technique provides a measure for the noninvasive assessment of the coronary involved region.
Method and Results In 30 normal subjects and 59 patients with ischemic heart disease, two-dimensional echocardiograms obtained at a rate of 60 frames per second were provided on line for digital subtraction analysis, with which digitized images were continuously subtracted on a frame-by-frame basis. The subtracted images were analyzed to determine the onset of the segmental outward motion of the LV wall in early diastole in each of 16 segments per subject. Regional relaxation index, defined as the interval from the second heart sound to the onset of outward wall motion, was significantly prolonged in the coronary involved segments compared with the normal segments (36.3±18.0 versus 101.2±34.0 ms, P<.01). The prolongation in the regional relaxation index was observed even in the coronary involved segments without reduction in systolic wall motion. When a cutoff level of 50.0 ms was used, coronary involved segments could be distinguished from normal or border segments with a sensitivity of 92% and a specificity of 81%.
Conclusions Digital subtraction high-frame-rate echocardiography may be used to visualize regional LV relaxation abnormalities in patients with ischemic heart disease. The time interval from the second heart sound to the onset of the segmental outward motion of the LV wall (regional relaxation index) obtained with this technique provides a noninvasive and accurate measure for assessing coronary involved regions.
Left ventricular (LV) diastolic function is frequently impaired before systolic function in patients with ischemic heart disease. A number of noninvasive indexes of diastolic function have been proposed by the analysis of M-mode echocardiograms, Doppler LV filling velocity patterns, and radionuclide ventriculograms. Diastolic function as assessed with these indexes is frequently abnormal even in patients with ischemic heart disease in whom systolic function is maintained within normal ranges.1 2 These indexes are particularly valuable in the evaluation of the outcome of coronary angioplasty, the inference of restenosis, and/or the estimate of prognosis in patients with ischemic heart disease.3 4 5 Conversely, several groups analyzed regional rather than global LV diastolic function. Specifically, cine or radionuclide angiography was used in patients with ischemic heart disease to demonstrate that the beginning of the regional LV outward wall motion in the isovolumic relaxation phase is delayed in the coronary involved region.6 7 8 Another study showed that rapid atrial pacing induced delay of the regional outward wall motion, and the delay was observed before hypokinesis developed.9 Because the impaired region is confined within the perfused area of the involved coronary arteries and because an ischemia-related event, if small, may not necessarily impair global LV diastolic function,10 the measure of the LV regional diastolic function may provide a more sensitive estimate of the coronary involved region than the measure of global LV function.11 LV regional diastolic function has been assessed mainly by the analysis of cine or radionuclide angiograms. Conventional two-dimensional echocardiography has not been used for assessing LV regional diastolic function in particular because the temporal resolution of currently available echocardiography is too poor to detect the subtle motion in the regional LV wall or the difference in the timing of wall motion among the segments. Temporal resolution of two-dimensional imaging, however, has been improved by the recent advent of high-frame-rate echocardiography. Analysis of high-frame-rate two-dimensional echocardiograms with a digital subtraction technique should further improve the temporal resolution.
In this study, high-frame-rate echocardiography was used in combination with a digital subtraction technique to assess regional LV wall motion in the isovolumic relaxation period in normal subjects and in patients with ischemic heart disease. The objectives of this study are to clarify whether the new technique may be used to visualize regional LV relaxation abnormalities in patients with ischemic heart disease and to clarify whether this technique provides a measure for the noninvasive assessment of the coronary involved region. We also related the index of regional LV diastolic function obtained with the new technique to the conventional echo indexes of LV global diastolic filling.
Principle of Digital Subtraction Analysis of High-Frame-Rate Echocardiograms
Two-dimensional echocardiograms can be obtained at a frame rate of 60 frames per second (16.7-ms interval) with a commercially available high-frame-rate echograph (EUB-565, Hitachi Medical Co). For this analysis, two-dimensional echocardiograms are recorded into a cine-loop memory. Then, echocardiograms are replayed with a conventional video frame rate (30 frames per second), and RS-170 video output signal is continuously digitized and processed with a real-time image processing system (Series 151, Imaging Technology Inc). The system is used to obtain subtraction images by performing subtraction in two successive digital echo images on a frame-by-frame basis (Fig 1⇓). The difference in the pixel intensity between a pair of successive images is caused by changes in backscatter intensity, movement of scattering objects, or both. If high-frame-rate rather than conventional echocardiography were used, the effects of the change in backscatter intensity in a certain pixel should be small. Thus, the difference, the residue of subtraction, is more likely to be a sign of the movement of the ultrasound-reflecting object than a sign of the change in backscatter intensity. In our system, positive residues of subtraction are displayed in white, and negative residues of subtraction, in black. We can monitor subtraction images in real time, and adequate-quality subtraction images are recorded into a disk array system and on videotapes for further analysis.
Because subtraction extracts the difference between two images, moving parts in echocardiograms are enhanced in subtraction echocardiograms. When an echogenic portion, eg, the endocardium, moves in a certain direction, the pixels in front of the echogenic portion get brighter and the pixels in the back get darker. Thus, the movement of an echogenic portion is displayed as an enhanced black-and-white pattern, and other static areas are displayed in a neutral gray color in the subtraction images. Thus, we can identify a minute wall motion and its direction by detecting a black-and-white pattern. In this study, we applied this image processing technique to delineate the outward motion of the LV wall in early diastole.
Candidates for this study were 75 consecutive patients with a history of exercise-inducible angina pectoris referred for elective coronary angiography and 36 healthy volunteers. Four patients and 6 volunteers were excluded because of inadequate ultrasound imaging. Twelve patients were excluded because of LV hypertrophy as assessed by septal wall thickness of >12 mm. None of the other patients had a history of arrhythmia, intraventricular conduction disturbance, hypertension, valvular heart disease, congenital heart disease, or heart failure. Patients with recent episodes of myocardial infarction or those treated with coronary angioplasty or bypass graft surgery within 3 months were also excluded. Thus, the study population consisted of 59 patients with ischemic heart disease and 30 volunteers. Previous myocardial infarction was present in 37 of 59 patients. Systolic regional wall motion abnormalities were evident in these 37 patients. No systolic wall motion abnormalities were found in the remainder. No ECG or echocardiographic abnormalities were present at rest or at exercise stress testing in any normal volunteers.
Patients underwent ultrasound examination within 1 month before or after coronary angiography. Echocardiograms were obtained in the left lateral position at end expiration. No stress was placed on the subjects. Patients were not given β-blocking agents or calcium channel blocking agents for 48 hours before the examination. Other medications were suspended for 24 hours before the examination.
We obtained high-frame-rate two-dimensional echocardiograms with a 3.5-MHz sector probe. We took four views: the parasternal long-axis view, parasternal short-axis view at the papillary muscle level, apical four-chamber view, and apical right anterior oblique view. The gray-scale gamma correction curve was maintained at a linear (1:1) relation between the gray level and echo amplitude. The echo gain was adjusted to the appropriate level. High-frame-rate echocardiograms were analyzed with a digital subtraction technique as described above.
Digital subtraction high-frame-rate echocardiographic study was followed by conventional M-mode, two-dimensional, and Doppler echocardiographic studies. A phonocardiogram was recorded simultaneously with M-mode and Doppler echocardiographic tracings with a microphone placed on the right sternal border of the second intercostal space. LV end-diastolic and end-systolic dimensions were measured in LV M-mode echocardiograms in a standard fashion.12 Isovolumic relaxation time was determined as the interval from the second heart sound to the onset of mitral valve opening in the M-mode recordings. We also recorded the pulsed Doppler transmitral flow velocity pattern at the mitral tip using a 2.5-MHz sector transducer. The transmitral flow velocity pattern was provided for the determination of peak early diastolic LV filling velocity (E), peak atrial filling velocity (A), and their ratio (E/A) as previously described.13 14
Assessment of LV Segmental Wall Motion
Regional wall motion abnormalities were assessed with a 16-segment model (Fig 2⇓). Regional systolic wall motion was scored by conventional two-dimensional echocardiograms on a three-point scale: 1, normal (>5-mm endocardial excursion); 2, hypokinetic (2- to 5-mm excursion); and 3, akinetic (<2-mm excursion).15 A segment scored 2 or 3 was considered to be abnormal.
In all patients, each of 16 segments was classified as one of the following three categories on the basis of the coronary distribution and the location of coronary stenosis: a normal segment (supplied with a coronary artery without significant stenosis), a border segment, or a coronary involved segment (supplied with a coronary artery with significant [>75%] stenosis). This classification of segments was done independently by an angiographer who was unaware of the results of ultrasound studies.
Analysis of Subtraction Images
Subtraction images were analyzed to determine the timing of the onset of outward endocardial motion in early diastole in each segment. This analysis was done in a blinded fashion without knowledge of patient profile or the results of ultrasound studies. Outward motion of the endocardium is delineated as a black contour of the edge of the endocardium (Fig 1⇑). An interval from the second heart sound to the appearance of the outward motion was measured in each segment and was defined as a regional relaxation index. To avoid misinterpretation of artifacts and noises that are not related to ventricular relaxation, any outward motion before the second heart sound and inward motion after that was ignored. Further, temporal continuity of the black contour of the endocardial edge after the first appearance was taken into consideration in the determination of the onset of outward motion. Regional relaxation index was measured in each LV segment. Because the frame rate is 60 frames per second, the value of the regional relaxation index is a multiple of 16.7 ms. For segments that could be visualized in multiple planes, averaged values were provided for further analysis.
Each patient had 16 values of regional relaxation index, and an average value of these was defined as global relaxation index. Asynchronized relaxation is considered to exist if there is a considerable variation in the timing of regional LV relaxation among segments. Thus, the SD of regional relaxation index was determined from the data of 16 segments in each patient and was provided as asynchrony index. If all 16 segments could not be assessed with echocardiography, only assessable segments were taken into account in this analysis (mean, 15 segments).
Contour Analysis of Left Ventriculograms
To validate the relaxation index, left ventriculography was performed in the right anterior oblique mode at a frame rate of 60 frames per second in 15 randomly selected patients, including 7 with reduced LV wall motion. LV contour was manually traced on serial ventriculographic frames from aortic valve closure to end diastole. Each contour was superimposed by alignment of the LV long axis and provided for segment length analysis with a cardiac function analyzing system (CARDIO 500, Kontron Instruments). Right anterior oblique LV contour was divided into six regions as done in the echocardiographic analysis. In each region, a time–segment length curve was drawn after smoothing of the data with the first to the seventh harmonics of the Fourier expansion. The time from aortic valve closure to the first peak of the first derivative of the time–segment length curve (peak dL/dt) was determined as an angiographic index of regional LV relaxation and was compared with the relaxation index.
Reproducibility of Measurements
Six patients with ischemic heart disease and four healthy subjects were randomly selected to determine reproducibility of the measurements based on regional relaxation index. Subtraction echo images were analyzed by one observer on two occasions (intraobserver variability) and by two independent observers (interobserver variability). Mean values of the absolute difference in the measurements were 9.7±11.3 ms (intraobserver variability) and 8.3±19.1 ms (interobserver variability) for regional relaxation index, 3.3±2.0 ms (intraobserver variability) and 9.4±4.5 ms (interobserver variability) for global relaxation index, and 3.1±2.9 ms (intraobserver variability) and 3.8±2.9 ms (interobserver variability) for asynchrony index.
Values are expressed as mean±SD. Statistical analyses of the difference in the parameters among groups were performed with one-way ANOVA and Scheffé’s method. We considered a value of P<.05 significant. Linear regression analysis was used to examine the relation between two variables.
LV Relaxation Index in Normal Subjects
Regional relaxation indexes were successfully obtained in 455 of 480 segments (95%) of healthy subjects. The minimum value of the regional relaxation index was 0.0 ms, indicating that regional relaxation starts at the same time as the second heart sound. The maximum value was 83.3 ms, in the fifth frame after the second heart sound. Regional relaxation index was shorter than their own isovolumic relaxation time in all healthy subjects.
Segmental variation in the regional relaxation index was not observed in healthy subjects. The maximum regional difference in the regional relaxation index was observed between the mid anterior septal segment and the basal-lateral segment. However, there were no significant differences in the average values of the regional relaxation index among 16 segments (Table 1⇓). The absolute difference was smaller than the time resolution (16.7 ms) of the present system. Thus, the regional relaxation index was not different among LV segments in healthy subjects, and the asynchrony index was 11.6±3.7 ms.
The global relaxation index, which was determined as an averaged value of regional relaxation index over 16 segments in each subject, was 33.0±10.9 ms. The global relaxation index did not correlate with age (r=.00, P=NS), nor did the asynchrony index (r=.09, P=NS).
LV Relaxation in Patients With Ischemic Heart Disease
Regional relaxation index was successfully obtained in 855 of 960 segments (89%) in patients with ischemic heart disease. Analyzable segments included 212 healthy segments, 264 border segments, and 380 coronary involved segments. Systolic segmental wall motion was reduced in 184 of the 380 coronary involved segments.
There was no significant difference in the regional relaxation index between the normal segments of patients with ischemic heart disease and the segments of the healthy subjects (36.3±18.0 versus 32.6±16.0 ms, P=NS). Inward regional LV wall motion was sometimes observed even after the second heart sound in the coronary involved segments. In such segments, the onset of the outward motion was obviously delayed. The regional relaxation index was significantly prolonged in the coronary involved segments (Figs 3⇓ and 4⇓). Regional relaxation index was significantly prolonged even in the coronary involved segments in which the systolic segmental wall motion was normal (101.2±34.0 ms, P<.01 versus normal segments). Regional LV relaxation was even more prolonged in the coronary involved segments in which systolic wall motion was reduced (121.3±46.4 ms, P<.01 versus normal segments).
The regional relaxation index of the normal segments ranged from 0.0 to 100 ms. The regional relaxation index of the coronary involved segments ranged from 16.7 to 250 ms. If a cutoff level of 50.0 ms (mean+SD value of healthy subjects) was used, the sensitivity of the prediction of the coronary involved segment was 92%, and specificity was 81%. If the same criterion was used, the sensitivity of the prediction of the border or coronary involved segment was 75%, and specificity was 94%.
The asynchrony index, defined as the variation of regional relaxation index, was significantly greater in patients with ischemic heart disease than in healthy subjects (35.1±16.5 versus 11.6±3.7 ms, P<.01). The asynchrony index was greater even in patients without old myocardial infarction than in healthy subjects (24.3±8.2 ms, P<.01 versus healthy subjects).
Relation of Digital Subtraction High-Frame-Rate Echocardiographic Indexes With Conventional Indexes of LV Diastolic Filling
A single peak was found in the LV inflow velocity pattern in 3 patients, and these patients were excluded from the analysis of the relation of the regional index with conventional indexes of LV filling. The E/A ratio tended to be lower and isovolumic relaxation time was significantly longer in patients with ischemic heart disease than in healthy subjects (Table 2⇓). Thus, these parameters of LV diastolic filling were compared with the indexes derived from digital subtraction high-frame-rate echocardiograms. Global relaxation index tended to be longer in subjects with more prolonged isovolumic relaxation time (r=.44, P<.01, n=86). However, global relaxation index did not correlate with the E/A ratio (r=.00, P=NS, n=86).
Isovolumic relaxation time was more prolonged in patients with a greater asynchrony index (r=.41, P<.01, n=86) (Fig 5⇓). If patients with myocardial infarction were excluded, the correlation coefficient was slightly improved (r=.71, P<.01, n=52). The E/A ratio did not correlate with asynchrony index (r=.13, P=NS, n=86).
Comparison With Angiographic Index of LV Regional Relaxation
The time from the aortic valve closure (the second heart sound) to peak dL/dt ranged from 30 to 237 ms. There was a significant correlation between the time from aortic valve closure to peak dL/dt and regional relaxation index (r=.61, P<.01) (Fig 6⇓).
Left Ventricular Wall Motion in the Isovolumic Relaxation Period in Healthy Subjects
The regional relaxation index was shorter than isovolumic relaxation time in any segment in all healthy subjects, suggesting that the outward motion of the LV wall motion starts in the isovolumic relaxation period in normal hearts. There was no regional difference in the regional relaxation index in healthy subjects, indicating that regional LV relaxation is homogeneous in normal hearts. This finding is in contrast to previous findings,16 because Hammermeister and colleagues showed heterogeneity in the regional LV relaxation in the normal heart. Further, several studies demonstrate that LV diastolic function is impaired with aging.13 17 18 They showed a significant correlation between age and indexes of LV global diastolic function in subjects without heart disease. In the present study, however, any digital subtraction high-frame-rate echocardiographic parameters of LV relaxation did not correlate with age in healthy subjects. The difference between our study and previous studies may be explained by the complexity of the diastolic function and/or the differences in the index of diastolic function measured. Although most of the previous parameters of LV diastolic function are affected by LV diastolic behavior not only in the isovolumic period but also in the early diastolic filling period, the indexes measured in this study are unlikely to be affected by LV diastolic behavior in the early diastolic filling period. Furthermore, our regional relaxation index reflects the timing of LV relaxation but is unlikely to reflect the rate of LV relaxation.
Regional Delay of the Onset of Outward Wall Motion in Patients With Ischemic Heart Disease
In the present study, the regional relaxation index was significantly prolonged in the coronary involved segments. Abnormal regional LV relaxation has been observed frequently in patients with ischemic heart disease and is often accompanied by abnormal simultaneous regional inward motion elsewhere in the LV wall6 19 ; the results of this study are consistent with these previous findings. The mechanism of abnormal LV regional relaxation in the coronary involved region has been explained by the presence of regional ischemia, alternation in the coronary flow, loading conditions, and cell-level changes, eg, action potential duration, time course of cytosolic calcium transient, intracellular phosphate, modification of myofilaments, and so on.20 21 22 23 The regional relaxation index was often longer than isovolumic relaxation time in the coronary involved segments, indicating that outward wall motion may start even after the mitral opening in such regions. The segmental prolongation of the regional relaxation index may indicate regional delay of the onset of relaxation. Such regional abnormalities in LV relaxation were present even in patients with ischemic heart disease but with normal systolic function and could be detected noninvasively with digital subtraction high-frame-rate echocardiography.
The detection of the prolongation of regional LV relaxation by digital subtraction high-frame-rate echocardiography may be useful for the estimation of coronary involved segments because the difference in the regional relaxation index between the coronary involved segment and the normal segment was >60 ms. This value is far greater than the frame interval of the present system (16.7 ms). In patients with ischemic heart disease, abnormal relaxation can persist with normal systolic function, suggesting that diastolic function is a more sensitive indicator than systolic function. This concept was supported by previous studies. Smalling et al24 measured segmental shortening and lengthening in a chronically instrumented canine ischemic heart model. They showed no significant relative difference between reduced coronary flow region and control region in segmental shortening but a twofold to fourfold difference in segment lengthening between ischemic and nonischemic regions. Other groups showed that regional diastolic asynchrony was corrected by treatment.25 26 Although regional abnormal relaxation is a nonspecific phenomenon, detection of segmental abnormalities of relaxation may be a very sensitive indicator of ischemia. In this study, as well as in most previous studies, measures of “normal systolic function” were very gross. Studies that used load-independent measures of LV contractility show that LV contraction and relaxation are mathematically related and that systolic function is really not “normal” in the setting of abnormalities of diastolic function.27 28 29 Thus, it may be possible that regional relaxation index is at least partially related not only to diastolic function but also to systolic function.
In this study, a cutoff level of 50 ms (mean±SD value of normal subjects) was used in the calculation of sensitivity (92%) and specificity (81%). To represent the 95% confidence interval, this value should be 2 SD. When a cutoff level of 67 ms (nearly mean±2 SD value of normal subjects) was used, the sensitivity and the specificity were 81% and 98%, respectively (Table 3⇓). For screening, a method with a high sensitivity despite a low specificity is preferable to the reverse. Thus, from this viewpoint, 50 ms may be more practical than 67 ms as a cutoff level.
Heterogeneity of Relaxation (Asynchrony) in Patients With Ischemic Heart Disease and its Relation to Global LV Diastolic Function
The E/A ratio was decreased and isovolumic relaxation time prolonged in patients with ischemic heart disease. These changes were associated with an increase in asynchrony index in patients with ischemic heart disease, implying a close relation between asynchronized relaxation and global diastolic dysfunction in patients with ischemic heart disease. In the correlation study, isovolumic relaxation time was significantly correlated with asynchrony index. Thus, asynchronized relaxation appeared to be one of the important factors affecting isovolumic relaxation time. This finding is consistent with previous findings because heterogeneity of regional diastolic function has been considered to be one of the important determinants of global LV diastolic dysfunction both in animals and in humans.8 11 30 31 32 33 The difference in segment lengthening between the anterior and posterior walls influenced isovolumic relaxation rates in dogs.20 In patients with ischemic heart disease, a significantly prolonged ventricular relaxation time was associated with segmental early relaxation,34 and impairment of global diastolic filling was associated with diastolic asynchrony.25 35 36 37
Asynchronized relaxation was evident even in patients without systolic wall motion abnormalities. The correlation coefficient between isovolumic relaxation time and asynchrony index was improved if patients with old myocardial infarctions were excluded, suggesting that asynchronized relaxation might not be important as a contributor to isovolumic relaxation time in patients with old myocardial infarctions. In other words, abnormal asynchrony may be a more important determinant of isovolumic relaxation time in patients without systolic wall motion abnormalities than in those with systolic wall motion abnormalities. Although a decrease in the E/A ratio was associated with an increase in asynchrony index in our patients with ischemic heart disease, these indexes correlate only weakly with each other. The influence of asynchronized LV relaxation may be less on early diastolic filling than on the index of the isovolumic phase. In other words, effects of asynchronized relaxation on the global LV diastolic function may get smaller between the isovolumic relaxation period and the early to late diastolic filling period. The difference between our study and previous studies may be explained by the difference in the cardiac phase when the index of asynchrony was measured.
Digital subtraction high-frame-rate echocardiography has several limitations. First, temporal resolution depends on the frame rate of two-dimensional imaging. Although temporal resolution has been improved in the present system compared with the conventional ultrasound system, it might still be suboptimal. The frame rate of 60 frames per second was the same as that of cine angiography; however, this frame rate might not provide the same temporal resolution as cine angiography, because two-dimensional echograms are constituted by echo-beam scanning. Although the frame rate of 60 frames per second was useful to detect the prolongation of the regional LV relaxation in the coronary involved region, the possible physiological asynchronized LV relaxation was not detectable with the present system. A higher frame rate may be required for visualization of the physiological asynchronized wall motion.
Second, noises in the original two-dimensional echocardiograms significantly lowered the accuracy of the measurements of quantitative indexes. It is difficult to identify even anatomic orientation if noisy echo images are used for subtraction. The measurement of regional relaxation index is no doubt inaccurate if determined in noisy echograms with low image quality. The quality of subtraction images depends in part on the contrast of gray level in the original two-dimensional echocardiograms.
Third, translation of the left ventricle may affect measurements of indexes. The spatial displacement of the LV wall is affected not only by the ejection and filling process but also by the translation of the entire left ventricle. Regional difference of indexes was not significantly different from results of the ventriculographic study that was aligned by position of the LV long axis. If translation of the entire left ventricle greatly affected the regional relaxation index, it might cause a regional difference in the regional relaxation index even in healthy subjects. Lack of regional difference in the regional relaxation index implies that the translation may be minimal in the isovolumic relaxation period. Because an influence of the entire LV motion depends on the magnitude of its translation,38 this kind of influence may not have been enough to affect the measurements of the regional relaxation index. Translation of the heart due to respiration was minimal because the examination was done with breath held at end expiration.
Fourth, effects of the magnitude of the excursion of the LV wall on the measurements of the regional relaxation index are noted. In this study, the timing of the LV outward motion was measured in a quantitative fashion. The black-and-white pattern in the subtraction image was more easily recognized if the excursion of the LV wall motion was larger. In the presence of reduced LV wall excursion, the occurrence of regional relaxation may be judged to be later than the actual occurrence. Thus, this may partially account for the longer relaxation index, particularly in the segments with reduced wall motion excursion. However, it is noted that the relaxation index was prolonged even in ischemic heart disease patients with normal systolic excursion of the LV wall.
To assess the overall effects of the technical limitations of our method on the measurement, relaxation index was compared with angiographic measures. The regional relaxation index was certainly more prolonged in the region in which wall motion was retarded in phase and slow in velocity. Although the correlation between the relaxation index and the time from aortic valve closure to peak dL/dt might be modest, this does not diminish the value of regional relaxation index, because angiographic and echocardiographic measures cannot be considered to be perfectly identical. In particular, echography provides information on a slice plane of the left ventricle, whereas angiography provides information on the silhouette of the LV cavity. In addition, the velocity of the wall motion may affect the time from the aortic valve closure to peak dL/dt more than the relaxation index. Because our measure has not been validated against any established comparable measure, the effects of potential problems of misregistration artifacts due to motion of the heart relative to the transducer during diastole, respiration, patient movement, and so on have not been fully assessed.
Fifth, there was a considerable overlap between healthy subject and patient groups in global relaxation index that is the averaged value of regional relaxation index in each individual, which may limit the use of this index for the noninvasive diagnosis of the disease. Such an overlap is reasonable when patients with regional LV disease are compared with healthy subjects, particularly because regional abnormalities may not be large enough to deviate from the average value in some patients. Separation might be more clear if the maximal rather than averaged value of regional relaxation index in individual patients were used. However, use of the maximal value is likely to increase vulnerability to the potential problems of artifacts described earlier.
Finally, it is noted that asynchronized relaxation is not specific for ischemic heart disease. Although we showed that digital subtraction high-frame-rate echocardiographic assessment of regional relaxation index is highly sensitive and specific for identifying the coronary involved regions, regional nonuniformity related to impaired LV relaxation has been also observed in patients with hypertrophic cardiomyopathy.39 40 In fact, LV relaxation is influenced by a variety of factors: ventricular hypertrophy, fibrosis, heart failure, alternation of loading condition, and catecholamines.20 21 22 Thus, the effect of these factors on the relaxation index should be assessed in future studies to apply this method to ischemic patients with concomitant hypertension, hypertrophy, or other disorders.
Use of high-frame-rate echocardiography and a digital subtraction technique improves both the spatial and temporal resolution of two-dimensional echocardiography. Thus, this technique was used to visualize segmental retardation of LV relaxation in this study. Asynchronous LV relaxation may exist in patients with other cardiac diseases, such as idiopathic cardiomyopathy and left ventricular hypertrophy. The present technique may also contribute to detecting impaired regional rather than global LV diastolic function and its heterogeneity (asynchrony) in such patients.
The coronary involved region could be estimated noninvasively even in patients without LV systolic wall motion abnormalities by use of digital subtraction high-frame-rate echocardiography. Prolongation of the regional relaxation index was highly sensitive for the assessment of the coronary involved area. Although delayed relaxation might not be specific for ischemic heart disease, our method should provide a clinically useful estimate of the ischemia-related regions, particularly in patients with possible ischemic heart disease. It is noted that the regional relaxation index, an index of regional LV function, can be obtained by counting the frame number from the second heart sound or the aortic valve closure to the appearance of the LV regional outward motion in subtraction images and that its measurement does not require any exercise or pharmacological stress.
This study was supported by a Grant-in-Aid for Developmental Scientific Research (A 03507001) from the Ministry of Education, Science, and Culture. We are grateful to Kunio Ono, PhD, Hirokazu Ihara, PhD, and Hiroshi Kanda, PhD, of Hitachi Medical Corp for their advice and modification of the equipment. We also thank Koji Umeshita, MD, for recruiting the healthy volunteers.
- Received August 10, 1994.
- Accepted October 18, 1994.
- Copyright © 1995 by American Heart Association
Upton MT, Gibson DG, Brown DJ. Echocardiographic assessment of abnormal left ventricular relaxation in man. Br Heart J. 1976;38:1001-1009.
DeMaria AN, Wisenbaugh TW, Smith MD, Harrison MR, Berk MR. Doppler echocardiographic evaluation of diastolic dysfunction. Circulation. 1991;84(suppl I):I-288-I-295.
Snow FR, Gorcsan J III, Lewis SA, Cowley MJ, Vetrovec GW, Nixon JV. Doppler echocardiographic evaluation of left ventricular diastolic function after percutaneous transluminal coronary angioplasty for unstable angina pectoris or acute myocardial infarction. Am J Cardiol. 1990;65:840-844.
Oh JK, Ding ZP, Gresh BJ, Bailey KR, Tajik AJ. Restrictive left ventricular diastolic filling identifies patients with heart failure after acute myocardial infarction. J Am Soc Echocardiogr. 1992; 5:497-503.
Ludbrook PA, Byrne JD, Tiefenbrunn AJ. Association of asynchronous protodiastolic segmental wall motion with impaired left ventricular relaxation. Circulation. 1981;64:1201-1211.
Gibson D, Mehmel H, Schwarz F, Li K, Kubler W, Mehmel H. Asynchronous left ventricular wall motion early after coronary thrombosis. Br Heart J. 1986;55:4-13.
Bonow RO, Bacharach SL, Green MV, Kent KM, Rosing DR, Lipson LC, Leon MB, Epstein SE. Impaired left ventricular diastolic filling in patients with coronary artery disease: assessment with radionuclide angiography. Circulation. 1981;64:315-323.
Dawson JR, Gibson DG. Regional left ventricular wall motion in pacing induced angina. Br Heart J. 1988;59:309-318.
Gibson D, Mehmel H, Schwarz F, Li K, Kubler W. Changes in left ventricular regional asynchrony after intracoronary thrombolysis in patients with impending myocardial infarction. Br Heart J. 1986;56:121-130.
Sahn DJ, DeMaria A, Kisslo J, Weyman AE, the Committee on M-Mode Standardization of the American Society of Echocardiography. Recommendations regarding quantitation of M-mode echocardiography: results of survey of echocardiographic measurements. Circulation. 1978;58:1072-1083.
Yamamoto K, Masuyama T, Tanouchi J, Doi Y, Kondo H, Hori M, Kitabatake A, Kamada T. Effects of heart rate on left ventricular filling dynamics: assessment from simultaneous recordings of pulsed Doppler transmitral flow velocity pattern and haemodynamic variables. Cardiovasc Res. 1993;27:935-941.
Hammermeister KE, Gibson DG, Hughes D. Regional variation in the timing and extent of left ventricular wall motion in normal subjects. Br Heart J. 1986;56:226-235.
Hirota YA. A clinical study of left ventricular relaxation. Circulation. 1980;62:756-763.
Gillebert TC, Lew WY. Nonuniformity and volume loading independently influence isovolumic relaxation rates. Am J Physiol. 1989;257:H1927-H1935.
Blaustein AS, Gaasch WH. Myocardial relaxation, VI: effects of beta-adrenergic tone and asynchrony on LV relaxation rate. Am J Physiol. 1983;244:H417-H422.
Grossman W. Diastolic dysfunction and congestive heart failure. Circulation. 1990;81(suppl III):III-1-III-7.
Silverman HS, Ninomiya M, Blank PS, Hano O, Miyata H, Spurgeon HA, Lakatta EG, Stern MD. A cellular mechanism for impaired posthypoxic relaxation in isolated cardiac myocytes: altered myofilament relaxation kinetics at reoxygenation. Circ Res. 1991;69:196-208.
Bonow RO, Vitale DF, Bacharach SL, Fredrick TM, Kent KM, Green MV. Asynchronous left ventricular regional function and impaired global diastolic filling in patients with coronary artery disease: reversal after coronary angioplasty. Circulation. 1985; 71:297-307.
Humphrey LS, Topol EJ, Rosenfeld GI, Borkon AM, Baumgartner WA, Gardner TJ, Maruschak G, Weiss JL. Immediate enhancement of left ventricular relaxation by coronary artery bypass grafting: intraoperative assessment. Circulation. 1988; 77:886-896.
Eichhorn EJ, Willard JE, Alvarez L, Kim AS, Glamann DB, Risser RC, Grayburn PA. Are contraction and relaxation coupled in patients with and without congestive heart failure? Circulation. 1992;85:2132-2139.
Little WC, Rassi A, Freeman GL. Comparison of effects of dobutamine and ouabain on left ventricular contraction and relaxation in closed-chest dogs. J Clin Invest. 1987;80:613-620.
Courtois M, Mechem CJ, Barzilai B, Ludbrook PA. Factors related to end-systolic volume are important determinants of peak early diastolic transmitral flow velocity. Circulation. 1992;85:1132-1138.
Lew WY, Rasmussen CM. Influence of nonuniformity on rate of left ventricular pressure fall in the dog. Am J Physiol. 1989; 256:H222-H232.
Momomura S, Bradley AB, Grossman W. Left ventricular diastolic pressure-segment length relations and end-diastolic distensibility in dogs with coronary stenoses: an angina physiology model. Circ Res. 1984;55:203-214.
Yamagishi T, Ozaki M, Kumada T, Ikezono T, Shimizu T, Furutani Y, Yamaoka H, Ogawa H, Matsuzaki M, Matsuda Y, Arima A, Kusukawa R. Asynchronous left ventricular diastolic filling in patients with isolated disease of the left anterior descending coronary artery: assessment with radionuclide ventriculography. Circulation. 1984;69:933-942.
Bonow RO. Regional left ventricular nonuniformity: effects in left ventricular diastolic function in ischemic heart disease, hypertrophic cardiomyopathy, and the normal heart. Circulation. 1990;81(suppl III):III-54-III-65.
Clayton PD, Jeppson GM, Klausner SC. Should a fixed external reference system be used to analyze left ventricular wall motion? Circulation. 1982;65:1518-1521.
Hayashida W, Kumada T, Kohno F, Noda M, Ishikawa N, Kojima J, Himura Y, Kawai C. Left ventricular regional relaxation and its nonuniformity in hypertrophic nonobstructive cardiomyopathy. Circulation. 1991;84:1496-1504.