The recent development of accurate methods to measure two-dimensional strain during dobutamine stress echocardiography has reactivated the debate as to the respective value of myocardial perfusion versus myocardial function assessment in detecting coronary stenoses. The aim of our study was to compare the effects of progressive coronary constrictions on two-dimensional strain and myocardial contrast echocardiography parameters during stress conditions.
Nine open-chest pigs were studied in the setting of various degrees of coronary constrictions. Two-dimensional strain and myocardial contrast echocardiography with Flash refilling sequence acquisitions were obtained at rest and during dobutamine infusion. Values of longitudinal strain (LS), circumferential strain (CS), radial strain (RS), and wall thickening, as well as myocardial perfusion parameter (A.b), were then calculated.
At rest, accuracy for detecting coronary stenosis was higher for CS, LS, and A.b (74%, 67%, and 69%, respectively) than for RS or wall thickening (62% and 64%, respectively). Dobutamine stress echocardiography increased the accuracy of A.b and LS to 77% and to 73%, respectively. Sensitivity during stress was higher for CS (93%) and A.b (77%), whereas specificity was higher for LS (89%) than for other parameters. Combined evaluations (CS+A.b, CS+LS, and LS+A.b) during dobutamine stress echocardiography improved both sensitivity and accuracy for detecting coronary stenosis.
Quantitative evaluation of contraction by LS and CS analysis and perfusion (A.b) during stress echocardiography resulted in similar accuracy levels, whereas the radial component was less accurate. Maximal sensitivities and accuracies were obtained by combined evaluations during stress.
Identification of coronary stenosis by qualitative assessment of left ventricular wall thickening (WT) during dobutamine stress echocardiography is still highly dependent on the expertise of the interpreter. According to the ischemic cascade theory, perfusion abnormalities should be detected at lesser degrees of coronary stenosis than radial function abnormalities. It has therefore been hypothesized that evaluation of perfusion may improve the detection of coronary artery disease. Myocardial contrast echocardiography using a real-time Flash-refilling sequence has been validated for the detection of perfusion defects. Comparison of quantitative myocardial contrast echocardiography with radial contraction evaluated by M-mode confirmed the superiority of myocardial perfusion assessment in detecting coronary stenosis both at rest and during stress.
A recently developed ultrasonic technique to measure two-dimensional (2D) strain based on speckle tracking allows for simultaneous assessment of circumferential, longitudinal, and radial myocardial deformations. This technique overcomes the angle-dependent limitations of strain assessment by tissue Doppler by analyzing greyscale B-mode images independently, yielding a more robust dataset for quantifying myocardial contractility. In both experimental and clinical studies, 2D strain has been shown to be feasible and accurate in the detection of contraction changes during ischemia under control conditions and during dobutamine infusion.
The availability of enhanced myocardial contrast echocardiography and strain techniques has rekindled the debate as to the respective value of assessing myocardial perfusion versus myocardial function in detecting coronary stenosis during stress echocardiography. This experimental open chest study conducted in piglets was designed to assess the effects of progressive coronary artery constriction on myocardial contraction assessed using WT and 2D strain, and to analyze these effects in relation with myocardial perfusion evaluated by myocardial contrast echocardiography.
Materials and Methods
The experimental protocols were handled in compliance with the Guiding Principles in the Use and Care of Animals (National Institutes of Health, No.85-23, revised 1996). Nine male York pigs, weighing 40 ± 3 kg, were sedated with ketamine hydrochloride (20 mg/kg) plus acepromazine (1 mL) and anesthetized with sodium pentobarbital (10 mg/kg) and ketamine (500 mg/h) during surgery. After tracheal intubation, the pigs were ventilated with a respirator (Siemens Servo B, [Siemens, Erlangen, Germany] room air supplemented with oxygen). Two fluid-filled catheters were advanced: one into the proximal aorta via the left carotid artery for monitoring systemic arterial pressure, and one into the right jugular vein for drug and fluid infusions. After sternotomy, the heart was suspended in a pericardial cradle, and the left anterior descending coronary artery was isolated proximally to the first diagonal branch. An ultrasonic flow probe (Transonic Systems Inc, Ithaca, NY) was positioned around the coronary artery for measuring mean coronary blood flow, and a screw-occluder was then placed around.
Assessment of Contraction by Echocardiography
Echocardiographic data were acquired with a Vivid 7 (General Electric Medical System, Horten, Norway) using a 4-MHz transducer placed directly on the epicardium. The transducer was fixed in a saline-filled latex bag. For speckle-tracking strain analysis, routine B-mode gray-scale images (mean frame rate 75 ± 5 Hz) were recorded in parasternal and apical views by manual application of the transducer. The recordings were analyzed offline using Echopac PC software (General Electric Medical System). Segmental circumferential and end-systolic radial strains, as well as WT, were quantified in the parasternal mid-level short-axis view. Apical views were used to measure longitudinal strain (LS) in risk area (RA, anterior wall) and in control area (CA, inferior wall). The initial contour of endocardial border was manually delineated in end systole and automatically tracked frame by frame with myocardial wall. The quality of the tracking could be verified by the operator, and both the contour and the region of interest were readjusted for optimal evaluation.
Assessment of Myocardial Perfusion by Contrast Echocardiography
Myocardial contrast echocardiography was performed using an HP Sonos 5500 (Philips Healthcare, Best, The Netherlands) capable of low-energy, real-time imaging. Short-axis midpapillary muscle images were obtained with the transducer fixed on a saline-filled latex bag positioned on the left ventricular anterolateral wall. Sonovue (Bracco, Mano, Switzerland) was continuously infused at a rate of 1 mL/min. Flash sequences were performed consisting of 10 high energy (1.7 mechanical index) consecutive pulses that produced bubble destruction, followed by low-energy (0.1 mechanical index) image acquisition of 20 cardiac cycles that enabled visualization of subsequent refilling. End-systolic images were selected for subsequent quantitative analysis. Perfusion sequences were analyzed using QLAB software (Philips). Two regions of interest were manually traced in the mid-wall myocardium in short-axis view in RA and CA. Myocardial contrast echocardiography quantification parameters were obtained from the analysis of the refilling sequence according to the fitted exponential function: y = A(1-exp -bt )+C (A: peak plateau amplitude or myocardial blood volume, b: rate constant of signal intensity increase or red blood cell velocity). The product of A and b (A.b) provides a measure of myocardial blood flow, which correlated well with microsphere evaluation in previous studies.
After baseline recordings, 4 grades of coronary artery constriction of increasing severity were applied: 2 stages of non–flow-limiting stenosis at rest (NFLS) and 2 stages of flow-limiting stenosis at rest (FLS). NFLS were applied to obtain a 40% and 70% reduction in coronary artery flow during hyperemia produced by 140 μg/kg/min adenosine infusion. FLS were calibrated to induce 25% and 50% reductions in coronary artery baseline resting flow. Mean left anterior descending coronary artery flow was measured at rest and during hyperemia induced by adenosine infusion. Adenosine infusion was only used to cause hyperemia and to augment flow to calibrate NFLS, and not as a diagnostic stress test. The dose of adenosine was adjusted to maximize hyperemia without causing systemic hypotension. The 4 grades of coronary constriction of increasing severity were done in the same sequence to avoid myocardial preconditioning, which could have modified the results. Aortic pressure and echocardiographic data were obtained at baseline and during the 4 constriction conditions both before and after a 15-minute period of continuous intravenous dobutamine infusion at a rate of 30 μg/kg −1 /min −1 . Dobutamine was then stopped during 30 minutes to allow for full recovery of function and perfusion. Coronary perfusion area was determined in the left ventricle short-axis as the myocardial area devoid of opacification after contrast infusion during complete left anterior descending coronary artery occlusion and corresponded to anterior and anteroseptal walls. The animals were sacrificed under deep anesthesia.
Summary data are expressed as means ± standard deviation. The Statistical Package for the Social Sciences version 11.5 (SPSS Inc, Chicago, IL) and Statel software (ad Science, Paris, France) were used for statistical analysis. Hemodynamic and echocardiographic measurements were compared using the Mann–Whitney test. P value less than .05 was considered significant. Normal values were obtained during baseline measurements. We defined baseline stage as normal state and considered all stages of coronary stenosis abnormal. The optimal value for detecting coronary constriction by strain, WT, and A.b measurements was determined from receiver operating characteristic (ROC) curve analysis. The optimal cutoff point was chosen as the whole value giving the best composite of specificity and sensitivity. Area under the ROC curves was used to compare the accuracies of tested modalities using the Hanley and McNeil method. Accuracy was calculated as the total number of true positive and true negative tests divided by the total number of measurements. Intra- and interobserver variability of 2D strain was measured at rest and during dobutamine stress. To reduce intraobserver variations, the main interpreter performed all measurements twice. To reduce interobserver variations, all measurements were performed separately by a second blinded interpreter.
The basic hemodynamic parameters obtained during all experimental conditions are summarized in Table 1 . Creation of NFLS or FLS did not induce any significant changes in heart rate or arterial blood pressure. Infusion of dobutamine was followed by a significant increase in heart rate (mean +29%, P < .01), and this effect was observed to the same extent both in the absence and presence of left anterior descending coronary artery constriction.
|Heart rate, beats/min|
|Rest||117 ± 9||116 ± 14||114 ± 18||115 ± 18||111 ± 17|
|Dobutamine||150 ± 9 d||153 ± 9 d||146 ± 14 d||144 ± 12 d||148 ± 7 d|
|Systolic blood pressure, mm Hg|
|Rest||116 ± 26||111 ± 31||119 ± 35||120 ± 40||125 ± 31|
|Dobutamine||138 ± 41||128 ± 35||140 ± 59||143 ± 68||143 ± 35|
|Diastolic blood pressure, mm Hg|
|Rest||80 ± 40||78 ± 46||82 ± 41||78 ± 41||82 ± 40|
|Dobutamine||85 ± 49||76 ± 47||86 ± 46||84 ± 43||90 ± 36|
|Mean coronary flow, mL|
|Rest||24.6 ± 6.8||25.1 ± 6.9||24.3 ± 5.5||18.2 ± 5.7 a||12.2 ± 5.4 b|
|Adenosine hyperemia||65.2 ± 21 c||44.7 ± 15.7 a c||38.6 ± 11.2 b c||26.5 ± 5 b||15.5 ± 6.5 b|
Effects of Ischemia on Absolute Values of Strain Observed at Rest
Figure 1 summarizes the measurements of 2D strains and WT in RA and CA, before and during dobutamine infusion, in the absence (base) and presence of coronary constrictions. At rest, NFLS40% and NFLS70% did not induce any significant change in function. FLS25% significantly reduced LS and circumferential strain (CS) at rest in the RA compared with the CA (−26% and −30%, respectively, P < .05), whereas WT and radial strain (RS) were significantly reduced by the more severe degree of constriction (FLS50%). In CA, baseline values for strain were similar to those of RA, and we did not observe any change in response to any level of constriction. Figure 2 A further illustrates the changes in A.b, LS, CS, and RS in RA observed at rest for each coronary constriction level compared with “base.”
Effects of Dobutamine on Absolute Changes in Strain in Conditions of Coronary Artery Constriction
Compared with values observed at rest in the absence of coronary artery stenosis, dobutamine infusion increased CS by 16%, LS by 11%, RS by 35%, and WT by 40%. The effects of coronary constriction on individual strain parameters were more pronounced during dobutamine-induced pharmacologic stress than at rest, and occurred at a lower magnitude of coronary artery constriction. A significant reduction in LS and CS was observed at NFLS70% (−12% and −14%, respectively, P < .05), whereas a significant decrease in absolute RS and WT did not occur until the FLS50% was imposed ( P < .05) ( Figure 1 ). Figure 2 B summarizes percent changes in A.b, LS, CS, and RS in RA during dobutamine infusion for each level of coronary artery constriction in comparison with baseline.
Effects of Coronary Constriction on Myocardial Perfusion (A.b)
Figure 3 A summarizes the measurements of A.b in RA, before and during dobutamine, in the absence and presence of coronary constriction. At rest in the RA, A.b was 6.3 ± 0.9, increasing to 11.7 ± 2.2 during dobutamine infusion (+86%, P < .01). FLS25% significantly reduced A.b parameter compared with baseline (no coronary constriction) (4.8 ± 0.5 vs 6.3 ± 2.7, P < .05). During dobutamine infusion, NFLS70% significantly reduced A.b from 11.7 ± 2.2 to 7.5 ± 1.6 ( P < .01). Figure 2 shows the percent change of A.b induced by each coronary artery constriction level compared with baseline, at rest ( Figure 2 A) and during dobutamine infusion ( Figure 2 B). The effects of coronary artery constriction on individual perfusion parameters were more pronounced during dobutamine infusion than at rest and occurred at a lower magnitude of constriction. Figure 3 B summarizes the measurements of A.b RA/CA ratio at rest and during dobutamine infusion. At rest, the values started to decline when FLS25% was imposed, whereas during dobutamine infusion, NFLS70% was sufficient to induce a significant decrease in ratio values (−25% compared with baseline, P < .05).
Comparison of the Different Parameters in the Diagnosis of Ischemia
Table 2 shows area under ROC curves allowing determination of optimal cutoff values for sensitivity, specificity, and accuracy for each parameter in detecting ischemia at rest and during dobutamine stress echocardiography. At rest, the area under ROC curve values for A.b, LS, CS, RS, and WT were 0.66, 0.66, 0.76, 0.53, and 0.55, respectively ( Table 2 , Figure 4 A). Area under ROC curve of RS was significantly lower than for CS ( P < .05). Optimal cutoff value was 5.7 for A.b, −13.5% for LS, −22% for CS, 43% for RS, and 47% for WT. On the basis of these optimal cutoff values, sensitivity was higher for CS (77%) than for the other parameters. Specificity was higher for LS (89%) than for A.b (75%), CS (71%), or radial parameters. Accuracy was 69% for A.b, 67% for LS, and 74% for CS. The combinations of CS + A.b and CS+LS provided higher sensitivities (88% and 85%, respectively) and accuracies (83% and 80%, respectively) ( Table 3 ).
|Parameters||CS Rest||CS Dobutamine infusion||LS Rest||LS Dobutamine infusion||RS Rest||RS Dobutamine infusion||WT Rest||WT Dobutamine infusion||A.b Rest||A.b Dobutamine infusion|