Author + information
- Received January 12, 2009
- Revision received July 8, 2009
- Accepted July 21, 2009
- Published online November 1, 2009.
- Toshihiko Asanuma, MD, PhD⁎ (, )
- Ayumi Uranishi, MSc,
- Kasumi Masuda, PhD,
- Fuminobu Ishikura, MD, PhD,
- Shintaro Beppu, MD, PhD and
- Satoshi Nakatani, MD, PhD
Reprint requests and correspondence:
Dr. Toshihiko Asanuma, Department of Functional Diagnostic Science, Division of Health Sciences, Osaka University Graduate School of Medicine, 1-7 Yamadaoka, Suita, Osaka, 565-0871, Japan
Objectives We sought to investigate the time course of post-systolic thickening (PST) and systolic abnormality after recovery from brief myocardial ischemia.
Background Myocardial ischemic memory imaging, denoting the visualization of abnormalities provoked by ischemia and sustained even after restoration of perfusion, is desirable and allows after-the-fact recognition of ischemic insult. PST offers a sensitive marker of myocardial ischemia, but whether this abnormal thickening remains after relief from brief ischemia is unclear.
Methods Tissue strain echocardiographic data were acquired from 27 dogs under 2 different conditions of myocardial ischemia induced by either brief coronary occlusion (15 or 5 min) followed by reperfusion (Protocol 1) or by dobutamine stress during nonflow-limiting stenosis (Protocol 2). Peak systolic strain and post-systolic strain index (PSI), a parameter of PST, were analyzed.
Results In Protocol 1, peak systolic strain was significantly decreased in the risk area during occlusion. This decrease in peak systolic strain in the 15-min group did not completely recover to baseline levels even 120 min after reperfusion, whereas the decrease in the 5-min group recovered immediately after reperfusion. We found that PSI was significantly increased during occlusion, but increased PSI in the 5-min group remained until 30 min after reperfusion (−0.19 ± 0.18 [baseline] vs. 0.19 ± 0.14 [30 min], p < 0.05) despite the rapid recovery of peak systolic strain. In Protocol 2, increased PSI was sustained until 20 min after the end of dobutamine infusion (−0.26 ± 0.11 [baseline] vs. −0.16 ± 0.10 [20 min], p < 0.05), although peak systolic strain recovered by 5 min after the end of dobutamine infusion.
Conclusions PST remained longer than abnormal peak systolic strain after recovery from ischemia. Assessment of PST may be valuable for detecting myocardial ischemic memory.
The identification of myocardial ischemia in patients with a history of chest pain remains challenging, particularly in the absence of abnormalities on electrocardiogram or elevation of serum biomarkers induced by myocardial damage. Because failure of identification causes inappropriate diagnosis, the development of myocardial ischemic “memory” imaging, i.e., the visualization of the abnormalities provoked by ischemia and sustained even after restoration of perfusion, is desirable. Although the suppression of fatty acid metabolism due to myocardial ischemia persists after the resolution of chest pain (1,2), myocardial function recovers relatively soon except in cases of myocardial stunning. Conventional assessment by echocardiography thus cannot be used for myocardial ischemic memory imaging.
Recent developments in tissue strain echocardiography (TSE) have permitted the quantitative assessment of regional myocardial function and precise detection of subtle abnormalities in wall motion such as post-systolic thickening (PST), which is defined as myocardial contraction after aortic valve closure (AVC) (3). PST is a sensitive marker of myocardial ischemia, and this evaluation improves the diagnostic accuracy for ischemic heart disease (4,5). This abnormal thickening has been suggested to continue after restoration of perfusion, along with decreased systolic wall motion in the stunned myocardium (6). However, persistence after recovery from a short duration of myocardial ischemia that does not cause stunning remains unclear. We hypothesized that PST would remain longer than abnormalities of systolic wall motion after relief from brief ischemia and could be used for ischemic memory imaging. We therefore investigated chronological changes in PST and systolic abnormality after recovery from ischemia. For this purpose, 2 different conditions of ischemia were used: 1) brief coronary occlusion followed by reperfusion (supply ischemia); and 2) dobutamine stress during nonflow-limiting stenosis (demand ischemia).
All animal studies were performed in accordance with guidelines for the care and use of laboratory animals at our institution. A total of 27 open-chest dogs (14.4 ± 1.4 kg) were used. Dogs were anesthetized with the use of intravenous pentobarbital sodium (35 mg/kg), intubated, and ventilated with room air by use of a respirator pump. Anesthesia with pentobarbital sodium was maintained throughout the experiment (6 to 8 mg/kg/h). A 5-F catheter was placed in the ascending aorta to monitor blood pressure, and a 7-F double-lumen catheter was placed in the inferior vena cava for continuous infusion of drugs. The electrocardiogram was monitored continuously.
In the model of brief coronary occlusion followed by reperfusion (Protocol 1), the heart was suspended in a pericardial cradle through a midline thoracotomy in the supine position. The proximal portion of the left anterior descending artery (LAD) 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 flowmeter. In the model of nonflow-limiting stenosis (Protocol 2), left lateral thoracotomy was performed in the right recumbent position, and a vascular occluder and flow probe were placed in the proximal portion of the left circumflex artery (LCX).
Tissue Strain Echocardiography
Tissue strain echocardiography was performed by the use of an ultrasound system (Aplio; Toshiba, Otawara, Japan). The left ventricular short-axis view at the papillary muscle was visualized with the use of a water bath as a standoff. Position of the ultrasound transducer was fixed by the use of a mechanical arm. Tissue harmonic 2-dimensional images and tissue harmonic Doppler data were simultaneously captured during 3 consecutive cardiac cycles. The frame rate was set at 103 frames/s. A microphone for the phonocardiogram was placed directly on the base of the aorta. Timing of AVC was determined by the aortic component of the second heart sound derived from the phonocardiogram (7).
Myocardial radial strain was analyzed from tissue Doppler data by the use of online software (TDI-Q, Toshiba) (7). The strain value between 2 points on a beam was determined by the use of the strain-definition formula. The distance between the 2 points was set at 3 mm. After the center of contraction was manually set on the image, Doppler angle correction was performed to evaluate regional radial strain in the short-axis view. The reference point (i.e., zero strain) was set at the peak R-wave. Assigned strain values in myocardium were color-coded and displayed on 2-dimensional images. To evaluate regional strain, a 3- or 4-mm circular region of interest (ROI) was set on the endocardium and manually tracked, frame by frame, during cardiac cycles.
Myocardial Contrast Echocardiography (MCE)
Real-time MCE (24 frames/s) was performed by use of the pulse subtraction mode to confirm the size and position of ischemic risk area. Definity (Bristol-Myers Squibb Medical Imaging, Billerica, Massachusetts) was diluted 1:10 in normal saline and administered intravenously at a rate of 0.75 ml/min. The same short-axis images as used for TSE were captured digitally from before the burst frames throughout the replenishment of microbubbles within the myocardium during transient coronary occlusion. One clear end-diastolic image was selected during the replenishment of contrast to evaluate the risk area. The size of the risk area was measured by planimetry and expressed as the percentage of nonperfused area in left ventricular myocardium.
Changes in regional strain over time were investigated in 13 open-chest dogs with brief LAD occlusion followed by reperfusion in Protocol 1 (Fig. 1). We performed MCE during 20 s of LAD occlusion to estimate risk area, and an interval of >30 min was provided for recovery from wall motion abnormality induced by the transient occlusion. We performed TSE after complete recovery of myocardial strain values in the risk area. The LAD was then occluded for 15 min (n = 5) or 5 min (n = 8), followed by reperfusion for 180 min. Before the LAD occlusion, heparin (100 U/kg) and lidocaine (2 mg/kg) were intravenously administrated to prevent coronary thromboembolism and ventricular arrhythmia. Lidocaine was maintained with continuous infusion (5 mg/kg/h) throughout the protocol. We confirmed LAD occlusion and reperfusion by LAD flow measurement. We acquired TSE data at baseline; at the end of occlusion; immediately after reperfusion (0 min); and 10, 20, 30, 60, 90, and 120 min after reperfusion. At the end of the experiment (180 min after reperfusion), the heart was excised, and a myocardial cross section corresponding to the echocardiographic image was obtained. This section was incubated in a 2% solution of 2,3,5-triphenyltetrazolium chloride for 15 min at 37°C, and the infarcted myocardium was identified as the region failing to demonstrate red staining.
Changes in regional strain were investigated before, during, and after dobutamine infusion in 14 open-chest dogs with nonflow-limiting stenosis of the LCX in Protocol 2 (Fig. 1). The risk area was evaluated by MCE in the same manner as in Protocol 1. After intravenous administration of heparin and lidocaine, LCX flow was measured before and during hyperemia to assess coronary flow reserve under conditions without stenosis. Hyperemia was induced by intravenous infusion of adenosine triphosphate (280 to 560 μg/kg/min), and the infusion rate was adjusted in each dog to achieve maximum hyperemia without a decrease in blood pressure. Nonflow-limiting stenosis was then created, and LCX flow was again measured before and during the same adenosine triphosphate infusion to evaluate coronary flow reserve under stenotic conditions. After measuring coronary flow reserve, dobutamine was incrementally infused at 10, 20, 30, and 40 μg/kg/min (>3 min each) as in clinical setting. The TSE data were acquired at baseline; under peak dobutamine stress; and 5, 10, 15, 20, 30, 60, and 120 min after the end of dobutamine infusion.
The TSE data were used to calculate peak systolic strain, post-systolic strain, defined as the peak strain coincident with PST after AVC, and maximal strain, defined as maximum length change during a cardiac cycle (Fig. 2). Post-systolic strain index (PSI) was calculated from a ratio, (post-systolic strain – peak systolic strain)/maximal strain, as a parameter of PST. When no peak was observed after AVC in the strain profile, the strain value at regional onset of myocardial thinning caused by early mitral filling was used as post-systolic strain. Onset of thinning was assessed by the use of regional tissue velocity from TSE data. Values from 3 consecutive cardiac cycles were averaged to obtain each index. Peak systolic strain and PSI were analyzed in the risk area (LAD region) in Protocol 1 and in both the risk area (LCX region) and the opposite normal LAD region in Protocol 2. Because relative wall thickening is important to detect ischemia in the dobutamine stress echocardiography, the ratio of peak systolic strain in the LCX region to that in the normal LAD region was measured in Protocol 2.
Interobserver and intraobserver correlations
Ten image clips were randomly selected from the total clips analyzed in Protocol 1 to assess interobserver and intraobserver correlations for TSE indexes. To determine the intraobserver correlation, the analysis was repeated by a second observer who was blinded to the values obtained by the first observer. To determine the intraobserver correlation, analysis was repeated 2 weeks later by the same observer.
Data are expressed as mean ± SD. Comparisons among hemodynamics and TSE indexes in each protocol were performed by analysis of variance followed by the Dunnett post-hoc test. Comparisons of sizes of the risk area between the 15- and 5-min occlusion groups were performed by use of the unpaired t test. Comparisons of hemodynamics and coronary flow data before and after the creation of nonflow-limiting stenosis were performed by use of the paired t test. Interobserver and intraobserver correlations were determined by use of Bland-Altman analysis. Values of p < 0.05 were considered statistically significant.
Hemodynamics in Protocol 1
Heart rate and systolic/diastolic blood pressure did not differ significantly throughout coronary occlusion and reperfusion in either the 15- or 5-min occlusion group (Table 1).
Risk area size and myocardial necrosis in Protocol 1
The size of the risk area derived from MCE did not differ significantly between the 15- and 5-min occlusion groups (26.5 ± 4.3% [15 min] vs. 29.1 ± 8.1% [5 min], p = 0.520). Myocardial necrosis was not detected by 2,3,5-triphenyltetrazolium chloride staining in any dogs from either group.
Chronology of changes in TSE indexes in Protocol 1
Peak systolic strain in the risk area was significantly decreased at the end of occlusion compared with that at baseline and began to normalize after reperfusion in the 15-min occlusion group; however, it had not completely recovered to baseline levels even at 120 min after reperfusion (Fig. 3A). Although peak systolic strain at the end of occlusion in the 5-min occlusion group decreased to the same level as that in the 15-min occlusion group, this decrease recovered to baseline levels immediately after reperfusion (Fig. 3B).
The PSI in the risk area was significantly increased at the end of occlusion in both occlusion groups as the result of the occurrence of PST. These increases were almost the same in each group. The PSI in the 15-min occlusion group decreased after reperfusion but had not reached baseline levels by 120 min after reperfusion (Fig. 4A). The PSI in the 5-min occlusion group recovered to baseline levels by 120 min after reperfusion, but the significant increase in PSI remained until 30 min after reperfusion despite the rapid recovery of peak systolic strain (Fig. 4B).
Strain profiles in the risk area derived from dogs with 15 or 5 min of occlusion are shown in Figure 5. Systolic wall thinning and PST after AVC were demonstrated at the end of occlusion in both groups. Complete recovery of peak systolic strain was not observed by 120 min after reperfusion, and PST was sustained in dogs with 15 min of occlusion (Fig. 5A). In contrast, peak systolic strain recovered immediately after reperfusion in dogs with 5 min of occlusion, but PST remained present until 30 min after reperfusion (Fig. 5B).
Nonflow-limiting stenosis in Protocol 2
The LCX flow after creating nonflow-limiting stenosis was unchanged compared with that before stenosis (17.3 ± 4.9 ml/min [before] vs. 17.8 ± 6.2 ml/min [after], p = 0.468), whereas the coronary flow reserve of LCX after stenosis was significantly decreased compared with that before stenosis (2.18 ± 0.48 [before] vs. 1.31 ± 0.24 [after], p < 0.001).
Hemodynamics in Protocol 2
Although heart rate and systolic blood pressure were significantly increased during peak dobutamine stress compared with levels at baseline, no significant increases were identified as of 10 min after stopping dobutamine infusion (Table 2).
Chronology of changes in TSE indexes in Protocol 2
Peak systolic strain increased during peak dobutamine stress in the risk and normal areas; however, the increase in the risk area tended to be smaller than that in the normal area because of reduced coronary reserve. Consequently, the ratio of peak systolic strain in the risk area to that in the normal area tended to decrease during peak dobutamine stress compared with that at baseline (0.75 ± 0.21 [baseline] vs. 0.63 ± 0.21 [peak stress], p = 0.604). This ratio recovered to baseline levels by 5 min after the end of dobutamine infusion (0.75 ± 0.21 [baseline] vs. 0.81 ± 0.24 [5 min], p = 0.977). In contrast, PSI in the risk area was significantly increased during peak dobutamine stress, and this increase was sustained until 20 min after the end of dobutamine infusion. The PSI in the normal area was unchanged during and after dobutamine stress (Fig. 6).
Interobserver and intraobserver correlations
For interobserver and intraobserver correlations in the measurement of peak systolic strain, mean differences were 0.019 and 0.002, and limits of agreement (±1.96 SD) were ±0.097 and ±0.060, respectively. Mean differences in the PSI measurement were 0.005 and 0.001, and limits of agreement were ±0.217 and ±0.263, respectively.
In the present study we examined the diagnostic value of PST for ischemic memory imaging. The PSI was significantly increased during 5 min of coronary occlusion, and increased PSI remained until 30 min after reperfusion despite the rapid recovery of peak systolic strain. Increased PSI induced by dobutamine infusion during nonflow-limiting stenosis also was sustained until 20 min after the end of infusion.
Recovery of PST after relief from myocardial ischemia
Deterioration of wall thickening or peak systolic strain induced by myocardial ischemia usually recovers soon after restoration of blood flow, except in cases of myocardial stunning. The present data revealed delayed recovery of peak systolic strain in the 15-min occlusion group, whereas recovery was observed immediately after reperfusion in the 5-min occlusion group, suggesting that 5 min of occlusion is not long enough to cause myocardial stunning.
Although PST is well known to be provoked during ischemia, the details of recovery after relief from ischemia remain unclear. Jamal et al. (6) have suggested that PST continues after reperfusion along with decreased systolic strain. In addition, in the 15-min occlusion group of our model, PSI as well as peak systolic strain did not recover to baseline levels even at 120 min after reperfusion. However, in the 5-min occlusion group, PSI continued to increase until 30 min after reperfusion, notwithstanding the immediate recovery of systolic strain. This, to the best of our knowledge, represents the first report that shows PST persists longer than systolic strain abnormalities after reperfusion.
In nonflow-limiting stenosis, the recovery of peak systolic strain was demonstrated by 5 min after ceasing dobutamine infusion, whereas increased PSI in the risk area remained until 20 min after cessation. Because the effect of dobutamine continues for a short time after the infusion is stopped, the timing of relief from ischemia should be interpreted carefully. However, heart rate and blood pressure recovered to baseline levels by 20 min, suggesting that increased PSI after 20 min is not induced by demand ischemia. Assessment of PST thus seems to be useful for detecting ischemic memory even in cases of demand ischemia.
Mechanisms of the occurrence of PST are still controversial. Skulstad et al. (8) have demonstrated that PST occurs by passive recoil in dyskinetic segments but is caused by active contraction in hypokinetic or akinetic segments. Claus et al. (9) have shown that PST originates from the end-systolic inhomogeneous state where neighboring segments have a different wall thickness as a passive phenomenon that is the result of elastic segment interaction. Although we could not clarify the mechanisms underlying the delayed recovery of PST in the present study, PST might reflect ischemic cellular or subcellular dysfunction more sensitively than peak systolic strain because chronological changes of PST did not parallel those of peak systolic strain. Prolonged suppression of fatty acid metabolism after myocardial ischemia despite the restoration of blood flow has been identified in recent studies (1,2) in which the authors used β-methyl-p-[123I]-iodophenyl-pentadecanoic acid. It should be further studied whether PST detects such a metabolic disorder.
Two studies (10,11) have reported that PST can assess myocardial viability. Because PSI was similarly high at the end of occlusion in the 15- and 5-min occlusion groups and myocardium in both groups was viable, our data are consistent with those previous studies. However, prediction of different recovery processes between the 15- and 5-min occlusion groups by PSI values immediately before reperfusion appears to be difficult.
In patients with chest pain, symptoms have often largely resolved by the time they come to the emergency department. An electrocardiogram and biochemical markers such as cardiac troponin can be helpful in diagnosing myocardial injury in such patients; however, the sensitivity of these methods for detecting myocardial ischemia in the absence of myocardial cell destruction is unsatisfactory. Stress echocardiography offers better sensitivity and specificity but is difficult to perform on unstable patients. Cardiac single-photon emission computed tomography with β-methyl-p-[123I]-iodophenyl-pentadecanoic acid appears to represent a promising method for ischemic memory imaging but must be performed in a radiation-controlled area.
If ischemic memory imaging could be performed with the use of echocardiography, this technique would be practical for diagnosing ischemic heart disease at the bedside in the emergency department. We think that the assessment of PST is promising as a method of ischemic memory imaging because this abnormality is prolonged compared with peak systolic strain. Because PST is sometimes observed even in healthy subjects (12,13), the judgment of ischemic memory may be complicated in discriminating between normal and abnormal PST. We would recommend that echocardiography for diagnosing ischemic memory should be performed twice in 10- or 15-min intervals because PST as a sign of ischemic memory would decrease over time whereas normal PST would remain at the same level. If PST tends to disappear in this sequence, it suggests recent but nonsustained myocardial ischemia.
The effect of rich collaterals in dogs should be considered when our results are translated in the clinical use. Even 5 min of occlusion would be more severe in humans because they have fewer collaterals. Therefore, PST might remain longer in the clinical settings compared with our results.
Our data also imply the advantages of evaluating PST after stress test for diagnosing stable angina. Tachycardia due to dobutamine often misleads physicians or sonographers into falsely identifying wall motion abnormality, but this would be avoided if assessment could be performed after the end of dobutamine infusion when the heart rate becomes slower.
The use of TSE is frequently hampered by the angle dependency of the Doppler technique. Despite the angle correction technique in this system, accurate strain values cannot be measured in the 2- to 3-o'clock and 9- to 10-o'clock positions of the short-axis view. The ROIs must be dislocated from these regions, and so they were not always placed in the center of the risk area. We therefore performed MCE and confirmed that all ROIs were carefully set to be included within the risk area.
Lucats et al. (14) have demonstrated that PST derived from sonomicrometry in a dog model is observed in the posterior wall even at baseline. In our results, however, PST was hardly detected in the LCX regions at baseline. Although the cause of this discrepancy is uncertain, the location of ROIs in our analysis might be different from that of Lucats's model because we could not place them in the center of the risk area as mentioned previously.
Absolute values of Doppler-derived strain are affected by the derivative pitch of strain measurement, the size and transmural location of ROIs, and the like. For this reason, a strain value derived from TSE should be normalized by a reference value. However, PSI can partly cancel this issue because it is a ratio of PST against the maximal strain.
Wang et al. (15) have shown that PST is more prominent in the subendocardial layer than in the subepicardial layer. Because we could not observe strain in each layer, it is still unknown whether PST in the subendocardial layer remains longer after recovery from ischemia than that in the entire layer.
In this work we found that PST remained longer than abnormal peak systolic strain after recovery from myocardial ischemia induced by brief coronary occlusion or dobutamine stress during nonflow-limiting stenosis. Therefore, the assessment of PST by TSE may facilitate the detection of myocardial ischemic memory.
This study was supported in part by research grants from the Fukuda Foundation for Medical Technology and Japan Society for the Promotion of Science (KAKENHI 20500418). Dr. Nakatani received a research grant from Toshiba Medical Systems.
- Abbreviations and Acronyms
- aortic valve closure
- left anterior descending artery
- left circumflex artery
- myocardial contrast echocardiography
- post-systolic strain index
- post-systolic thickening
- region of interest
- tissue strain echocardiography
- Received January 12, 2009.
- Revision received July 8, 2009.
- Accepted July 21, 2009.
- American College of Cardiology Foundation
- Kawai Y.,
- Tsukamoto E.,
- Nozaki Y.,
- Morita K.,
- Sakurai M.,
- Tamaki N.
- Dilsizian V.,
- Bateman T.M.,
- Bergmann S.R.,
- et al.
- Pislaru C.,
- Belohlavek M.,
- Bae R.Y.,
- Abraham T.P.,
- Greenleaf J.F.,
- Seward J.B.
- Kukulski T.,
- Jamal F.,
- Herbots L.,
- et al.
- Voigt J.U.,
- Exner B.,
- Schmiedehausen K.,
- et al.
- Jamal F.,
- Strotmann J.,
- Weidemann F.,
- et al.
- Skulstad H.,
- Edvardsen T.,
- Urheim S.,
- et al.
- Takayama M.,
- Norris R.M.,
- Brown M.A.,
- Armiger L.C.,
- Rivers J.T.,
- White H.D.
- Lyseggen E.,
- Skulstad H.,
- Helle-Valle T.,
- et al.