Author + information
- Received June 6, 2011
- Revision received September 14, 2011
- Accepted September 27, 2011
- Published online January 1, 2012.
- Toshihiko Asanuma, MD, PhD⁎ (, )
- Yumi Fukuta, MSc,
- Kasumi Masuda, PhD,
- Ayana Hioki, BSc,
- Mariko Iwasaki, MSc and
- Satoshi Nakatani, MD, PhD
- ↵⁎Reprint requests and correspondence:
Dr. Toshihiko Asanuma, Division of Functional Diagnostics, Department of Health Sciences, Osaka University Graduate School of Medicine, 1-7 Yamadaoka, Suita, Osaka, 565-0871, Japan
Objectives The aim of this study was to evaluate which regional myocardial parameters derived from speckle tracking echocardiography could demonstrate myocardial ischemic memory in a brief ischemia-reperfusion dog model.
Background Myocardial ischemic memory imaging, denoting the visualization of abnormalities provoked by ischemia and sustained even after restoration of perfusion, can convey important clinical information. We previously reported that post-systolic shortening (PSS) remains in the risk area after recovery from brief ischemia. However, it is still unclear whether abnormalities in other regional deformation parameters persist after relief from brief ischemia.
Methods Echocardiographic data were chronologically acquired from 11 dogs during 2 min of coronary occlusion followed by reperfusion. Regional systolic and diastolic deformation parameters, including parameters related to PSS, were measured from radial and circumferential strain and from strain rate analyzed in the risk and normal areas. Strain imaging diastolic index (SI-DI), which had been proposed as a parameter for assessing ischemic memory, was also calculated.
Results Peak systolic strain, end-systolic strain, and peak systolic strain rate decreased in the risk area during occlusion but recovered to the baseline level immediately after reperfusion. Strain rate during early diastole decreased during occlusion; however, the decrease did not persist after reperfusion. Post-systolic strain index (PSI) and time-to-peak strain index, which are parameters of PSS, increased during occlusion. These increases persisted until 10 to 20 min after reperfusion (circumferential PSI: 0.02 ± 0.04 [baseline] vs. 0.08 ± 0.04 [20 min], p < 0.05). SI-DI did not show a significant change during occlusion because of a large variation.
Conclusions Although abnormalities of PSS-related parameters alone persisted after recovery from 2-min occlusion, abnormalities of other deformation parameters, such as strain rate during early diastole, did not. These data suggest that assessment of PSS by speckle tracking echocardiography is useful for detecting myocardial ischemic memory.
Acute chest pain often resolves before a patient can present to the hospital for further evaluation, which makes it difficult to accurately determine whether the chest pain was due to myocardial ischemia. Stress testing can be used to provoke myocardial ischemia in such cases, but this testing is difficult to perform in unstable patients. In such situations, imaging technology with the capabilities to detect recent myocardial ischemia (i.e., ischemic memory) would be of tremendous utility.
Post-systolic shortening (PSS) is myocardial contraction that occurs after end-systole and is a sensitive marker of myocardial ischemia (1–3). PSS can be easily identified by tissue Doppler and speckle tracking echocardiography. In a previous study using a dog model with brief ischemia, we reported that PSS persists even after a rapid recovery in systolic strain abnormalities and that assessment of PSS is useful for detection of myocardial ischemic memory (4).
Meanwhile, Ishii et al. (5) evaluated the strain imaging diastolic index (SI-DI), calculated from changes in myocardial strain from end-systole to the first one-third of diastole, before and after percutaneous coronary intervention and showed that the abnormality of SI-DI in the risk area was still present 24 h after intervention. From this result, they concluded that regional diastolic dysfunction persisted after intervention, and they called this phenomenon “diastolic stunning.”
However, the transition point from contraction to relaxation in regional myocardium is not always consistent with that from systole to diastole in global left ventricular (LV) function, which is defined hemodynamically. Especially in the myocardium with PSS, because the regional transition point is delayed after global end-systole, the beginning of regional relaxation may be during the first one-third of diastole or after that. This implies that although SI-DI is a diastolic parameter, it may not reflect regional relaxation in the myocardium with PSS. A different parameter, such as strain rate during early diastole (SRe), which reflects regional relaxation after the end of PSS, should be assessed to clarify this issue, but it is still unclear whether regional deformation parameters other than PSS can convey ischemic memory after brief ischemia. We therefore evaluated whether abnormalities of regional systolic and diastolic deformation parameters, including SRe, derived from speckle tracking echocardiography persist after relief from brief myocardial ischemia.
All animal studies were performed in accordance with guidelines for the care and use of laboratory animals at our institution. A total of 11 open-chest mongrel dogs (14.6 ± 1.0 kg) were used. Dogs were anesthetized using intravenous pentobarbital sodium (35 mg/kg), intubated, and ventilated with room air using a respirator pump. Anesthesia with pentobarbital sodium was maintained throughout the experiment (6.8 to 9.0 mg/kg/h). The electrocardiogram was monitored continuously, and LV pressure was measured by a 5-F micromanometer (Millar Instruments, Houston, Texas). The heart was suspended in a pericardial cradle through a left lateral thoracotomy in the right recumbent position. The proximal portion of the left circumflex coronary artery (LCX) was dissected free from surrounding tissues, and a vascular occluder was placed there. A perivascular ultrasonic flow probe (Transonic Systems, Ithaca, New York) was set at the distal site of the occluder and connected to a digital flow meter.
Echocardiography for the speckle tracking analysis was performed using an Aplio ultrasound system with a PST-50AT transducer (Toshiba, Otawara, Japan). The transmitting and receiving frequency was set at 5.0 MHz, and the frame rate was set at 118 frames/s. The LV short-axis view at the papillary muscle was visualized with a water bath as a standoff, and the position of the transducer was fixed by a mechanical arm. Two-dimensional images were digitally captured over 3 consecutive cardiac cycles in each data acquisition.
Because ischemic memory after 15 and 5 min of coronary occlusion has been tested in the previous study (4), we assessed ischemic memory after a briefer duration of occlusion in the present study. The LCX was thus occluded for 2 min, followed by reperfusion for 60 min. Before LCX occlusion, heparin (100 U/kg) and lidocaine (2 mg/kg) were intravenously administrated to prevent coronary thromboembolism and ventricular arrhythmia. LCX occlusion and reperfusion was confirmed by flow measurement. Echocardiographic data for the speckle tracking analysis were acquired at baseline, at the end of occlusion, and at 10, 20, 30, and 60 min after the onset of reperfusion. LV pressure was recorded at the same time as the acquisition of the echocardiographic data. LV systolic pressure, LV end-diastolic pressure, maximum and minimum time derivatives of LV pressure (dP/dtmax, dP/dtmin), and time constant of LV pressure decay during isovolumic relaxation period (τ) were averaged from 5 consecutive cardiac cycles. At the end of the protocol, real-time myocardial contrast echocardiography (MCE) was performed during transient LCX occlusion to confirm the size and location of ischemic risk area, as previously described (4). The size of the risk area was calculated from the percentage of the area of contrast defect to whole LV area on the image.
The speckle tracking analysis was performed by offline software (Toshiba, Otawara, Japan) (6). The endocardial border, excluding papillary muscles, was visually identified, and a contour was manually traced on an end-systolic frame. The epicardial border was determined by setting an even width of myocardium. In the data obtained during LCX occlusion, the epicardial border was also traced manually because the end-systolic thickness of ischemic myocardium is markedly thinner than that of nonischemic myocardium. Afterward, endocardial and epicardial borders were automatically tracked for all frames in the clip. Myocardium in the short axis was automatically divided into 6 segments. Myocardial strain and strain rate profiles were analyzed in a segment within the risk area derived from MCE and in an opposite segment within the normal area. End-diastole was defined at peak R-wave of the electrocardiogram, and end-systole was defined at the timing of LV dP/dtmin (7). The reference point (i.e., zero strain) was set at end-diastole.
Radial and circumferential strain and strain rate parameters were calculated from 3 consecutive cardiac cycles and averaged (Fig. 1). With this software, values of radial strain and strain rate were calculated from the endocardial and epicardial tracking points (i.e., transmural strain), and those of circumferential strain and strain rate were calculated from only the endocardial tracking points. Strain values were color coded and displayed on whole myocardium in radial strain and on the inner half of myocardium in circumferential strain. Peak systolic strain, end-systolic strain, and peak systolic strain rate (SRs) were measured as parameters of regional systolic function, and SRe and strain rate during atrial contraction (SRa) were measured as parameters of regional diastolic function. SRe was defined as a peak strain rate in the direction of myocardial lengthening (i.e., negative peak in radial SRe and positive peak in circumferential SRe) between end-systole and the beginning of atrial contraction.
When PSS was observed in the strain profile, post-systolic strain, which is peak strain after end-systole, was measured. The post-systolic strain index (PSI) was calculated from a ratio, ([post-systolic strain] − [end-systolic strain])/(maximum strain change during the cardiac cycle), as a parameter of PSS. Time-to-peak strain, which is the time from the R-wave to the peak strain during the cardiac cycle, was also measured, and time-to-peak strain index was calculated from a ratio, (time-to-peak strain)/(R-R interval duration), as another parameter of PSS.
Systolic stretch, which is passive wall motion induced by an increase of LV pressure, occurs in myocardium with delayed or impaired contraction (8). Because delayed contraction is often observed in the ischemic myocardium, it was analyzed as a parameter other than PSS. Systolic stretch was calculated from a ratio: (peak extension strain during systole)/(maximum strain change during systole). In accordance with the method of Ishii et al. (5), SI-DI was calculated from a ratio: ([end-systolic strain] − [first one-third strain value of diastolic duration])/(end-systolic strain).
One image clip was randomly selected from each dog, and a total of 11 clips were used to assess interobserver and intraobserver variability for radial and circumferential peak systolic strain, SRe, and PSI.
Data are expressed as mean ± SD. Multiple comparisons in each hemodynamic parameter were performed by 1-way analysis of variance with the Dunnett post hoc test. Because of heterogeneity of variance, multiple comparisons in each deformation parameter were made against values at baseline by 1-way repeated measures analysis of variance with the Bonferroni post hoc test (using nonpooled error terms) in the risk and normal areas separately. Interobserver and intraobserver variability was determined by Bland-Altman analysis. Values of p < 0.05 were considered to represent statistical significance.
Hemodynamics and risk area size
LV end-diastolic pressure was significantly higher during LCX occlusion than at baseline. LV systolic pressure and dP/dtmin tended to decrease and τ tended to be prolonged during occlusion relative to baseline. However, these parameters recovered to baseline levels by 10 min after reperfusion (Table 1). The size of the risk area derived from MCE was 31.5 ± 6.9%.
Strain and strain rate profiles
Circumferential strain and strain rate profiles in the risk and normal areas are shown in Figure 2. In the strain profiles, systolic stretch and PSS were demonstrated in the risk area during LCX occlusion. Although systolic strain recovered to the baseline level by 10 min after reperfusion, PSS persisted until 30 min after reperfusion.
In the strain rate profiles, SRs and SRe decreased in the risk area during occlusion; however, the decrease did not persist after reperfusion. In particular, SRe increased beyond baseline levels early after reperfusion.
Peak systolic strain and end-systolic strain
Radial peak systolic strain significantly decreased in the risk area and increased in the normal area during occlusion but recovered to baseline levels by 10 min after reperfusion. Circumferential peak systolic strain significantly decreased in the risk area and tended to increase in the normal area but also recovered to baseline after reperfusion. Radial and circumferential end-systolic strain showed results similar to those seen with peak systolic strain (Fig. 3).
SRs, SRe, and SRa
Radial SRs tended to decrease, and circumferential SRs significantly decreased in the risk area during occlusion, but these values recovered to baseline levels by 10 min after reperfusion.
SRe and SRa were analyzed in 9 of 11 dogs because the values could not be separately delineated in 2 of the dogs. Radial and circumferential SRe significantly decreased in the risk area during occlusion but recovered after reperfusion. At 10 min after reperfusion, radial and circumferential SRe significantly increased beyond the baseline levels. Radial and circumferential SRa did not change significantly in the risk and normal areas (Fig. 4).
PSI and time-to-peak strain index
In radial and circumferential strain, PSI significantly increased in the risk area during occlusion. The significant increase of radial PSI persisted until 10 min after reperfusion, and the significant increase of circumferential PSI continued until 20 min after reperfusion. Radial and circumferential PSI tended to be higher than baseline levels, even 30 min after reperfusion.
In radial and circumferential strain, time-to-peak strain index as well as PSI significantly increased in the risk area during occlusion. The significant increase of time-to-peak strain index persisted until 10 min after reperfusion in circumferential strain. Time-to-peak strain index also tended to be higher than baseline levels, even 30 min after reperfusion (Fig. 5).
Systolic stretch and SI-DI
Radial and circumferential systolic stretch significantly increased in the risk area during occlusion but recovered to baseline levels by 10 min after reperfusion. In radial and circumferential strain, SI-DI changed in the risk area during occlusion, but this change did not reach the level of statistical significance because of a large standard deviation during occlusion. Radial and circumferential SI-DI tended to decrease in the risk area at 10 min after reperfusion when compared with baseline, but this difference did not reach the level of statistical significance. On the contrary, radial and circumferential SI-DI significantly increased in the normal area during occlusion and at 10 min after reperfusion (Fig. 6).
Interobserver and intraobserver variability
Interobserver and intraobserver variability for radial and circumferential peak systolic strain, SRe, and PSI are shown in Table 2. The limits of agreement in the radial direction tended to be larger than those in the circumferential direction.
In this study, chronological changes of regional systolic and diastolic deformation parameters derived from speckle tracking echocardiography were evaluated in a 2-min coronary occlusion-reperfusion model. With this brief ischemia, not only systolic parameters such as peak systolic strain, end-systolic strain, and SRs, but also diastolic parameters such as SRe and SRa did not demonstrate ischemic memory. By contrast, we found that the PSS-related parameters PSI and time-to-peak strain index are effective parameters for assessing ischemic memory. Because SI-DI did not significantly change during occlusion and after reperfusion, our data suggest that it is not a proper parameter for ischemic memory.
Evaluation of ischemic memory by echocardiography
The ability to detect myocardial abnormalities that persists after the resolution of ischemic chest pain would be invaluable in the clinical setting. It has been reported that the suppression of fatty acid metabolism due to myocardial ischemia persists after relief from ischemia (9–11). Using cardiac single-photon emission computed tomography with β-methyl-p-[123I]-iodophenyl-pentadecanoic acid, Dilsizian et al. (10) detected abnormal fatty acid metabolism 30 h after stress testing, and they concluded that the metabolic imaging could be used for detecting recent ischemic insults. This is a promising method for ischemic memory imaging but must be performed in a radiation-controlled area. By contrast, ischemic memory imaging by echocardiographic techniques would have greater utility in the clinical setting and could even facilitate bedside examination.
PSS is a sensitive marker of myocardial ischemia, and the diagnostic accuracy of PSS for detecting ischemia is superior to that of systolic strain parameters (2). Moreover, assessment of PSS allows assessment of ischemic memory because PSI persists after relief from brief ischemia despite a rapid recovery of systolic strain (4). However, prior to the present study, it was unclear whether abnormalities of deformation parameters other than PSI could demonstrate ischemic memory after such brief ischemia.
Systolic parameters after brief ischemia-reperfusion
In the present study, LV systolic pressure tended to decrease during LCX occlusion but recovered to the baseline level after reperfusion. Regional systolic parameters, peak systolic strain, end-systolic strain, and SRs, decreased during occlusion but also recovered after reperfusion.
When severe ischemia occurs in a transient fashion, myocardial systolic dysfunction can persist despite normalization of blood flow; this phenomenon is known as “myocardial stunning.” In myocardial stunning, the functional, biochemical, and microstructural abnormalities that occur following ischemia are reversible, and contractile force is gradually restored (12). Systolic dysfunction in myocardial stunning can thus be used as a measure of ischemic memory. However, in the present study, abnormalities of the systolic parameters did not persist after reperfusion because the occlusion time was relatively short. This result suggests that these systolic parameters are not adequate measures of ischemic memory following such brief ischemia.
Diastolic parameters after brief ischemia-reperfusion
Regional systolic dysfunction and regional diastolic dysfunction both occur during ischemia. Animal and clinical studies using tissue Doppler echocardiography have revealed that SRe decreases in the risk area during ischemia induced by coronary occlusion (13,14). However, whether abnormalities of regional diastolic parameters persist after brief ischemia has not been fully investigated.
In the present study, LV end-diastolic pressure increased during occlusion, dP/dtmin tended to decrease, and τ tended to be prolonged. However, these global diastolic parameters recovered to baseline levels after reperfusion. Further, the regional diastolic parameter, SRe, decreased during occlusion but increased beyond the baseline level early after reperfusion. This result suggests that SRe does not convey ischemic memory following brief ischemia as well as systolic parameters. The reason for the increase of SRe early after reperfusion may be due to the persistence of PSS. As shown in Figure 2, the slope of strain profiles after PSS became steeper, resulting in increased SRe. Because regional relaxation in the myocardium with PSS occurs mostly after isovolumic relaxation, it seems to be influenced by a tethering effect of the surrounding nonischemic myocardium.
PSS-related parameters, systolic stretch, and SI-DI
PSI and time-to-peak strain index significantly increased in the risk area until about 10 to 20 min after reperfusion. These results are consistent with our previous study (4) and suggest that PSS-related parameters can convey ischemic memory after such brief ischemia. In our results, ischemic memory of PSS-related parameters persisted longer in circumferential than radial strain. Smaller standard deviations of values in circumferential strain seem to be one of the reasons. Moreover, circumferential strain is derived from only the endocardial tracking points in the software we used. This may be another reason of longer ischemic memory in circumferential strain because subendocardial myocardium is greatly affected by ischemia than subepicardial myocardium. Although we used time-to-peak strain index in the present study, time to zero-crossing in strain rate (15) would be also useful for detecting ischemic memory.
The mechanism underlying PSS remains unclear. Animal experiments conducted by Skulstad et al. (7) suggested that PSS was due to passive recoil and active contraction, and these investigators concluded that PSS in dyskinetic segments was due to passive recoil, whereas PSS in hypokinetic and akinetic segments was related to active contraction. Meanwhile, Claus et al. (16) suggested that PSS developed as an attempt to restore differences in end-systolic myocardial thickness between the ischemic area and the adjacent nonischemic area. In any case, these results indicate that PSS reflects systolic dysfunction more than diastolic dysfunction. Although systolic stretch is often observed in the ischemic myocardium, it did not convey ischemic memory in the present study.
Ishii et al. (5) showed that abnormalities in SI-DI persisted after coronary intervention, and they termed this prolonged dysfunction “diastolic stunning.” SI-DI indicates the strain change during the first one-third of diastole. Theoretically, its value should reflect PSS and be able to demonstrate ischemic memory. However, unlike PSI, SI-DI did not show a significant change during occlusion and after reperfusion in the present study. This reason seems to be a large variation of SI-DI during occlusion. Division by end-systolic strain is needed in the calculation of SI-DI. Because end-systolic strain approaches zero in the risk area during occlusion, SI-DI was prone to vary widely in this time frame.
Actually, SI-DI tended to decrease in the risk area at 10 min after reperfusion when compared with that at baseline; however, the decrease did not reach the level of statistical significance. When the data during occlusion in our experiment were excluded from the analysis, circumferential SI-DI significantly decreased in the risk area at 10 min after reperfusion compared with the baseline. Although this result suggests that SI-DI has the capability to demonstrate ischemic memory, the reason of the decreased SI-DI values seems to be due to large PSS at 10 min after reperfusion. The assessment during the first one-third of diastole may not be appropriate for evaluating regional relaxation.
Although decreased SI-DI values persisted 24 h after intervention in the report by Ishii et al. (5), these values almost completely recovered to the baseline level by 30 min after reperfusion in the present study. In the results of Ishii et al. (5), not only SI-DI, but also end-systolic strain, continued to decrease even 24 h after reperfusion (i.e., systolic stunning). This result suggests that the ischemic insults of myocardium in their study were unexpectedly severe despite brief occlusion. The cause of the systolic stunning after the brief occlusion is unknown. However, microvascular and myocardial damage incidental to coronary intervention such as distal emboli by the atherosclerotic plaque may be 1 of the reasons in the clinical setting.
In the normal area, radial and circumferential SI-DI significantly increased during occlusion and at 10 min after reperfusion. Lucats et al. (17) have reported that regional relaxation in the normal area is accelerated when PSS is induced in the risk area. This phenomenon seems to be related to the increase of SI-DI in the normal area. The reason is unclear but may be due to compensatory hyperkinetic motion.
Interobserver and intraobserver variability
In a prior report, interobserver and intraobserver limits of agreement of longitudinal strain measurements by speckle tracking echocardiography were ±6.4% and ±6.0%, respectively, in the animal study and ±8.6% and ±5.2%, respectively, in the clinical study (18). The limits of agreement in our data seem to be consistent with this result.
Because only the LV short axis was evaluated in the present study, strain and strain rate parameters were analyzed in the radial and circumferential directions but could not be analyzed in the longitudinal and transverse directions. The fact that only the LCX territory was examined is another limitation.
Although PSS persisted for about 20 min after brief ischemia-reperfusion, this duration may be too short for ischemic memory imaging. We assessed ischemic memory after 2-min occlusion in the present study. However, because the duration of ischemic memory changes with coronary occlusion duration and severity, this phenomenon would likely be of longer duration in more severe cases of ischemia (4).
The reason why only PSS-related parameters persisted after brief ischemia is unclear. The persistency of PSS may reflect subtle systolic stunning that cannot be detected by conventional systolic parameters, and a reversible metabolic disorder (9–11) may be one possible mechanism of this phenomenon, but further investigation is necessary.
Stress echocardiography offers better diagnostic accuracy for detecting myocardial ischemia but is difficult to perform on unstable patients. Because the assessment of PSS by speckle tracking echocardiography can be performed in a wider variety of clinical scenarios, we believe that this method would be suitable for ischemic memory imaging.
PSS is a sensitive indicator of myocardial ischemia but can also be observed in normal myocardium (19,20), which may complicate the determination of ischemic memory. However, serial examinations, with repeat examination 10 to 20 min later, may help detect acute chronological changes in PSS that would indicate the presence of ischemia rather than normal myocardium.
Of the regional myocardial deformation parameters derived from speckle tracking echocardiography, only PSS-related parameters demonstrated ischemic memory based on their persistency after relief from brief ischemia. By contrast, the abnormality of the regional diastolic parameter SRe did not convey ischemic memory.
This study was supported in part by research grants from Japan Society for the Promotion of Science (KAKENHI 20500418) and the Shimadzu Science Foundation. Dr. Nakatani has received a research grant from Toshiba Medical Systems. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- minimum time derivative of left ventricular pressure
- left circumflex coronary artery
- left ventricular
- myocardial contrast echocardiography
- post-systolic strain index
- post-systolic shortening
- strain imaging diastolic index
- strain rate during atrial contraction
- strain rate during early diastole
- peak systolic strain rate
- Received June 6, 2011.
- Revision received September 14, 2011.
- Accepted September 27, 2011.
- American College of Cardiology Foundation
- Pislaru C.,
- Belohlavek M.,
- Bae R.Y.,
- Abraham T.P.,
- Greenleaf J.F.,
- Seward J.B.
- Voigt J.U.,
- Exner B.,
- Schmiedehausen K.,
- et al.
- Asanuma T.,
- Uranishi A.,
- Masuda K.,
- Ishikura F.,
- Beppu S.,
- Nakatani S.
- Ishii K.,
- Suyama T.,
- Imai M.,
- et al.
- Skulstad H.,
- Edvardsen T.,
- Urheim S.,
- et al.
- Coppola B.A.,
- Covell J.W.,
- McCulloch A.D.,
- Omens J.H.
- Kawai Y.,
- Tsukamoto E.,
- Nozaki Y.,
- Morita K.,
- Sakurai M.,
- Tamaki N.
- Dilsizian V.,
- Bateman T.M.,
- Bergmann S.R.,
- et al.
- Kontos M.C.,
- Dilsizian V.,
- Weiland F.,
- et al.
- Bolli R.
- Abraham T.P.,
- Belohlavek M.,
- Thomson H.L.,
- et al.
- Amundsen B.H.,
- Helle-Valle T.,
- Edvardsen T.,
- et al.