|Home | About | Journals | Submit | Contact Us | Français|
Chronic ischemic left ventricular (LV) dysfunction is present in a number of clinical syndromes in whom myocardial revascularization results in an improvement of LV function, patients functional class and their survival. Early diagnosis of and treatment of viability is essential. Coronary arteriography is of limited value in diagnosis of viability. Non-invasive testing is essential for diagnosis which can be matched to the pathophysiologic changes that occur in hibernating myocardium. However, no single test has a perfect, or near perfect, sensitivity and specificity, and thus, a combination of tests are usually needed. Algorithms are developed to integrate these tests in clinical decision making.
Scientific discoveries tend to be made
not by those who seek to prove a hypothesis,
but by those who keep their eyes open……..Isaac NewtonAlbert Einstein 1
It was believed that left ventricular (LV) dysfunction (LVD) at rest was the result of ongoing ischemia or myocardial infarction. 33 years ago studies of patients undergoing coronary artery bypass graft surgery for angina led to the discovery of improvement or even normalization of the LVD with myocardial revascularization, that is, there was viable myocardium [Hibernating Myocardium (HM)] in areas of LV dysfunction.2,3 It took almost 12+ years for the concept of HM to be clinically acceptable. In the ensuing years certain clinical issues have been recognized:
These findings have emphasized the need for and importance of non-invasive tests to diagnose and quantitate the viable myocardium in areas of LV dysfunction.
This review is limited to assessment of myocardial viability only in chronic LV dysfunction and will not attempt to differentiate between stunning and hibernation. It will focus on the important issues for the non-invasive cardiac imager, namely: 1) Pathophysiology of myocardial hibernation as it pertains to the imager and imaging targets; 2) Identification of viable myocardium and prediction of functional outcomes by various imaging modalities; 3) Integrated imaging schema which can be used in the clinical setting to identify viable myocardium; and 4) Issues involved in clinical decision making.
The fundamental derangement is reduced myocardial blood flow (MBF), specifically subendocardial MBF (SE-MBF). In normal animals, the subendocardium governs transmural contraction and the subendocardium receives more MBF per unit of muscle than the epicardium.14 In conscious dogs, there is a sensitive coupling between SE-MBF and function so that only a to 20% reduction in SE-MBF15 can cause severe regional dysfunction, and the relationship between systolic thickening and SE-MBF is more or less linear and dependent on the hemodynamic situation.16 There is a 2 to 1 relationship in the reduction of SE-MBF to reduction of transmural MBF (TM-MBF).16-19 For example, a 25% reduction in TM-MBF results in a 50% reduction in SE-MBF (Figure 1). A 50% reduction in TM-MBF in anesthetized dog20 and 75% reduction in SE-MBF in the conscious dog results in akinesis of LV wall whereas subepicardial MBF does not correlate to transmural LV wall function14. In animals with coronary arterial obstruction following relief of ischemia, the ischemia-induced vasodilatation is maldistributed in the myocardium supplied by the obstructed coronary artery possibly due to microcirculatory changes. As a result, the increased volume of MBF from the hyperemia is directed to a greater extent to the epicardium and the subendocardium remains ischemic at least for a period of time.21 Thus, relief of ischemia and even restoration of TM-MBF does not necessarily mean SE-MBF is normal or has been restored to pre-ischemic levels.
Acute and short-term animal studies have shown that episode(s) of ischemia lead to the development of HM.22
Experimental studies of adaptation to chronic fixed coronary stenosis have shown: a) MBF may be normal or only mildly diminished in the early period but regional contraction may already be reduced (“stunning”) but over the subsequent time period reduction in MBF and function are matched (“hibernation”)14 ; and b) Progression to matched decreases in flow and function (HM) also occurs very early when a 15-minute partial coronary occlusion is followed by reperfusion through a critical stenosis (“hibernation”).10
The few studies that showed blood flow was “not reduced” measured only TM-MBF by PET; at that time it was not possible to measure SE-MBF in humans. The weight of evidence shows patients with HM have reduced TM-MBF at rest 14 which implies SE-MBF is significantly reduced to a much greater extent.
Until recently it was not possible to measure SE-MBF in humans with any degree of precision. Using cardiovascular magnetic resonance imaging (CMR), Selvanayagam and colleagues24 have documented in patients with HM, SE-MBF is reduced at rest when compared to areas without significant coronary stenosis in the same patients and following revascularization SE-MBF and LV function are normalized/improved (Figure 2). This led Klocke to conclude this study “…provides convincing evidence for the hibernation paradigm.” 25
Coronary flow reserve (CFR) is reduced in people with hyperlipidemia, in those with coronary atherosclerosis, and in patients with obstructive coronary artery disease. Thus, reduced CFR is not just a phenomenon of HM or stunning.
Irrespective of rest flow, CFR is almost always reduced in viable but dysfunctional myocardium, albeit more severely in segments with low flow at rest. The severity of CFR reduction directly impacts on the ability of viable myocardium to improve its contraction upon inotropic stimulation, as this requires increases in MBF. Viable segments without contractile reserve (CR) usually have lower CFR than viable segments with CR. The response of viable segments to these stimuli varies greatly from segment to segment as well as with the intensity and the duration of stimulation. Many viable segments in HM exhibit a biphasic response when challenged by increasing levels of inotropic stimulation.26 At low levels of stimulation, and provided that sufficient CFR is present, systolic wall thickening usually increases and starts earlier in systole. At higher levels, when the increase in demand cannot be matched anymore by further increases in MBF, systolic function again deteriorates and can even become worse than at baseline. The observation of this biphasic pattern is important to rule out other causes of regional dysfunction such as the presence of a subendocardial scar or remodelled myocardium, these two conditions being characterized by a sustained CR at both low and high levels of inotropic stimulation. About 20-25% of viable segments do not improve functionally during inotropic stimulation. These segments usually have mild reduction of resting MBF, avidly take up glucose under fasting conditions and have an almost completely exhausted MBF reserve. This prevents them from increasing their oxygen consumption upon inotropic stimulation and hence to increase their contractile function. Apart from the exhaustion of MBF reserve, other factors, like the presence and severity of cardiomyocytes alterations or down regulation of ß-adrenoreceptors, may also contribute to the lack of CR in apparently viable segments.
Besides the changes in resting MBF and CFR, several structural alterations affect cardiomyocytes and extra cellular matrix in patients with chronic HM.22,27 These structural changes are mostly encountered in the areas of dysfunction, but can also be seen in remote normally contractile regions. Most of the available information on these structural changes has been gathered from studies in which human myocardial biopsy specimens that were harvested at the time of coronary bypass surgery and also from experimental studies of HM.
Histological analysis of human dysfunctional myocardium demonstrated the microvasculature was better preserved in segments that improved in function after revascularization than in those that remained dysfunctional. This is particularly true for the capillaries whose density and cross-sectional area are usually within the normal range in viable segments.28 In contrast, there is greater heterogeneity among persistently dysfunctional segments despite revascularization, and approximately half of them show significant capillary rarefaction.
The primary alteration is the depletion of contractile elements in the cardiomyocytes. In some cells, this is limited to the vicinity of the nucleus, whereas in others it is very extended, leaving only few or no sarcomeres at the cell periphery. The space previously occupied by the myofilaments is usually filled with glycogen. The mitochondria are increased in number and display alterations in size and shape. Nuclei are usually tortuous, and show uniformly dispersed heterochromatin. Sarcoplasmic reticulum is virtually absent, as are T-tubules. At the molecular level, the expression of myosin, actin, titin and α-actinin is usually reduced. Cytoskeletal proteins such as desmin, tubulin and vinculin are disorganized. The expression of connexin-43, a major gap junction protein, and that of nuclear A-type lamins are also reduced.
Shows increased amounts of type I collagen, type III collagen and fibronectin in chronic LVD. Within the widened interstitium, an increased number of vimentin-positive cells (endothelial cells and fibroblasts) and macrophages can usually be seen.
The severity of the structural changes that affect chronically dysfunctional myocardium are likely to impact on its ability to respond to an inotropic stimulus and on the speed at which it may or may not recover after revascularization. Segments with less fibrosis or with less severe cardiomyocyte alterations are more likely to improve in function following the administration of dobutamine or after revascularization. Current data suggests a spectrum of myocardial dysfunction may exist in patients with coronary artery disease.
Based on the understanding of the pathophysiology of myocardial viability, a number of these targets can be clinically imaged. The following section describes each of the imaging techniques that can be used to image the key pathphysiologic substrates of: a) Fibrosis and/or cellular viability; b) Flow-metabolism; c) Microcirulation; and d) Contractile reserve.
The most important determinant of the return of resting contraction following revascularization is the severity of the underlying tissue fibrosis, whether interstitial or infarct-related. Imaging methods can directly assess the presence of tissue fibrosis, such as delayed enhancement by CMR, or its inverse, i.e. the mass of viable cardiomyocytes, such as thallium and metabolic imaging using positron emission tomography (PET), or consequences of fibrosis such as wall thickness/chamber size (Echocardiography and CMR). However, in the absence of flow or CR measurements, images obtained at rest while sensitive, cannot distinguish between viable and remodeled myocardium (especially in the remote myocardium) and therefore lack specificity.
LV volume measurement is a guide to the degree of irreversible LV damage. A very dilated LV (end-diastolic volume greater than twice the upper limit of normal) is unlikely to demonstrate significant global functional recovery (eg. improvement of LV ejection fraction >5%), as this degree of LV remodeling is usually caused by a large number of scarred segments. 2D echocardiography underestimates LV volumes, LV opacification improves the reliability of these measurements. 3 dimensional echocardiography permits more accurate volume measurements without needing to make geometric assumptions which is desirable if the LV has been involved in multiple previous myocardial infarctions.29
The involvement of ≥4 ventricular wall segments by scar (ie thinned segments) also identifies the LV that is unlikely to show global functional recovery after revascularization. Extensive infarction also reduces LV compliance, producing a restrictive filling pattern. A short mitral E wave deceleration time is associated with a small number of viable segments or a large number of scar segments, and deceleration time is directly related to the degree of LV ejection fraction change after revascularization, with >5% improvement being unusual with a deceleration time of <180 ms.30 Segments with significant transmural extent of infarction become retracted and fibrotic as the infarct heals. LVend-diastolic wall thickness of ≤0.5-0.6 cm is associated with akinesis or even dyskinesis, and these segments usually demonstrate no CR in response to dobutamine.30,31 A thinned segment is very unlikely to recover (≤ 5% probability to recover); thinning to ≤ 0.5-0.6 cm has a sensitivity of 94% for infarction but its specificity is much lower. A segment of > 0.5-0.6cm may recover (specificity 48%) and is more likely to do so if it augments in response to dobutamine; the combination of wall thickness with augmentation at dobutamine stress (see below) gives a sensitivity of 88% and specificity of 77% for prediction of functional recovery.30,31 Dobutamine stress and single photon emission computed tomography (SPECT) provide similar degree of additional information to wall thickness.31
Increasing transmural extent of scar is inversely related to the likelihood of functional recovery after revascularization and correlates with the degree of reduction of radial function. Wall motion scoring is too insensitive to distinguish minor gradations of function. The thickness of subendocardial scar may be identified using myocardial contrast echocardiography; the relative thickness of scar and perfused myocardium corresponds with transmural extent of scar by MRI. New echocardiographic indices of LV deformation are sensitive to the reductions in function caused by non-transmural scar. Tissue velocity-based longitudinal strain and strain-rate are reduced with subendocardial scar, although the relationship is non-linear because of the importance of the subendocardium to longitudinal function. Studies using speckle-based (“two-dimensional”) strain32 have shown radial or circumferential strain to be impaired in proportion to the transmural extent of scar.
Myocardial deformation is impaired in proportion to the extent of myocardial fibrosis, so the magnitude of strain is an index of the amount of chronically dysfunctional tissue. Unfortunately, this association is not sufficiently robust to be used in decision-making, probably because the parameters are load-dependent. Although tissue velocity is less useful than strain for the assessment of regional changes because it is less site-specific and influenced by tethering of adjacent segments, its high temporal resolution (especially using pulsed-wave Doppler) may allow the measurement of pre-ejection velocity, which may be a load-independent marker. Preserved pre-ejection velocity has been reported to be a reasonable resting marker of viability, and is a predictor of the likelihood of functional recovery and favorable outcome.33
Gadolinium chelates are extracellular/interstitial contrast agents that enhance T1 relaxation in regions of infarction as a consequence of the increased volume of distribution of the chelates within collagenous scar and their delayed washout due to reduced capillary density.34 The end result is that regions of myocardial infarction (MI) appear bright on inversion recovery images acquired 10-20 minutes after gadolinium administration. The spatial extent of LGE closely mirrors the distribution of myocyte necrosis early after MI and that of collagenous scar seen at 8 weeks.35 Furthermore, studies have shown that in regions of the heart subjected to reversible injury, the retention of contrast does not occur.36 Regions of chronically dysfunctional myocardium often consist of an admixture of reversibly injured and irreversibly injured (infarcted) myocardium.37 The power of LGE resides in its ability to distinguish these two states within the same segment of myocardium (Figure 3).
The ability of LGE to identify and characterize myocardial scar has been directly compared to both SPECT and PET. In a study of 91 patients with suspected or known coronary artery disease, SPECT failed to correctly identify nearly half of the subendocardial infarcts that were identified by LGE.38 Another study compared LGE to PET in 31 patients with ischemic HF (10). Infarct mass correlated well between the two modalities (r=0.81, p<0.0001), but LGE more frequently identified subendocardial scar than PET did, again reflecting its superior spatial resolution. In fact, when the transmural extent of scar is <37%, PET defines the segment as viable due to the presence of sufficient signal from remaining viable midwall and subepicardium.39 When scar is >37% transmural, PET defines the segment as nonviable.
The utility of LGE for identifying the likelihood of functional recovery after revascularization was demonstrated in a study of 50 patients imaged before and after revascularization.40 Eighty percent of patients had some region of LGE with a mean signal intensity that was 530% of that seen in normal segments. Functional recovery inversely correlated with the transmural extent of scar. In segments with no LGE, 78% demonstrated recovery of function. Only 1 of 58 segments with >75% transmural extent, however, showed any improvement following revascularization, thus demonstrating the powerful negative predictive value of this finding. Several subsequent studies have further supported the LGE imaging approach for predicting functional recovery. 41,42 The standard CMR pulse sequence used for LGE imaging is an inversion recovery (IR) gradient echo technique.43 which has been carefully validated in a canine model. This pulse sequence requires selection of an inversion time that optimally nulls normal myocardium. A recently developed phase-sensitive inversion recovery (PSIR) technique is less sensitive to the choice of an incorrect inversion time and may therefore be more efficient.44 Additional novel approaches include a subtractive inversion recovery technique using both a long and short inversion time,45 that improves infarct-blood pool contrast by 247 ± 136% compared to magnitude inversion recovery while maintaining signal difference-to-noise ratio for infarct-myocardium. However, subtraction techniques are prone to misregistration. Another approach to improve the differentiation of the blood pool and infarcted subendocardium is a multi-contrast method using a combination of T1 and T2-weighted imaging.46 For patients who cannot hold their breath, a single shot sub second technique was recently validated with slightly lower sensitivity and overall accuracy but still quite useful in the acutely ill patient nonetheless.47
Studies with both echocardiography and cine CMR have established end-diastolic wall thickness as an important parameter in the assessment of myocardial viability, although its accuracy, using fluorodeoxyglucose (FDG) PET as the reference standard, is less than that of CR in response to dobutamine.48 In one study, segments that failed to improve following revascularization had a significantly lower end-diastolic wall thickness (6mm vs. 9.8mm, p <0.001) than those that showed functional improvement.49 More important than end-diastolic wall thickness alone may be the residual thickness of the unenhanced rim of viable myocardium beyond the area of LGE in the subendocardium. One study compared en-diastolic wall thickness and viable rim with FDG PET in a group of 22 patients with ischemic cardiomyopathy.39 Receiver-operator characteristic analysis demonstrated greater area under the curve for viable rim compared to end-diastolic wall thickness (0.95 vs. 0.86, respectively). Optimal cutoff values for viability were 5.4 mm for end-diastolic wall thickness and 3.0 mm for the viable rim.
Tracers used with SPECT or PET to image MBF are commonly referred to as “perfusion tracers”, their uptake and retention mechanisms require viable myocellular membranes. Thus, uptake and visualization of myocardial regions with perfusion tracers requires the presence of working, viable myocytes. This information is the converse of direct and indirect fibrosis imaging described above by echocardiography and CMR. Thallium-201 redistribution, which reflects myocardial potassium space with SPECT, and rest sestamibi imaging (with or without nitrates), which reflects mitochondrial membrane integrity are the two most common “perfusion” tracers used to image cellular viability.
The evaluation of myocardial ischemia and viability by thallium scintigraphy has occupied a rather unique place in the management of patients with known or suspected coronary artery disease since late 1970’s.50,51 Like potassium, thallium is transported across the sarcolemma membrane via the Na-K ATPase system. The initial extraction and distribution of thallium in the myocardium is primarily a function of blood flow (either during stress or at rest) and is unaffected by hypoxia, chronic hypoperfusion (hibernation), or postischemic dysfunction (stunning), unless myocardial infarction is present. The later distribution of thallium (3-4 hr or 24 hr after stress or rest injection), termed redistribution phase, is flow-independent and is a function of regional blood volume and myocardial potassium space reflecting cellular viability.51,52 Thus, thallium defects on early rest images that “fill-in” on delayed, redistribution phase (termed reversible defect) represents a scintigraphic pattern of HM. In contrast, because thallium is not actively taken up in regions of scarred myocardium, defects on rest images that persist on redistribution images (termed irreversible or fixed defect) represent scarred myocardium.
There are a number of thallium protocols that are used clinically for the detection of myocardial viability. Among the range of choices, 2 protocols are optimized for viability detection: 1) rest-redistribution and 2) stress-4-hour redistribution-reinjection imaging (Figure 4). The former assesses myocardial viability alone53,54 while the latter assesses myocardial ischemia and viability.55,56 A pooled analysis of rest-redistribution and stress-redistribution-reinjection thallium studies reported a relatively high sensitivity (80%-90% range) and modest specificity (54%-80% range) for the prediction of recovery of regional function after revascularization.57,58 However, these conclusions must be viewed in the context of the limitations of pooled data analysis. When taking into consideration regions with reversible defects (ischemia) and success of revascularization (reexamining regional perfusion or vessel patency after revascularization) stress-redistribution-reinjection thallium imaging yields excellent positive and negative predictive accuracy (both in the 80%-90% range) for recovery of function after revascularization.55,56
In the case of technetium-99m based perfusion tracers, both sestamibi and tetrofosmin are lipophilic cationic complexes that are retained within the mitochondria due to a large negative transmembrane potential. Because accumulation and retention of sestamibi and tetrofosmin are related to energy-dependent processes which maintain mitochondrial membrane polarization, they may also serve as markers of cellular viability. The administration of nitrates to improve resting MBF prior to injection of sestamibi or tetrofosmin appears to improve slightly the ability of these tracers to detect cellular viability.59
With regard to prediction of improvement in LV function after revascularization, both thallium and technetium-99m based techniques have been shown to be reasonably accurate. Reported sensitivities for recovery of ventricular function are in the 75-85% range, with positive- and negative-predictive values of about 70% and 90%, respectively.57,58,60 With additional enhancements to imaging protocols, such as nitrate-enhanced rest perfusion imaging and gated SPECT for assessment of function, the accuracy of these techniques is even higher.59,61
Viable myocardial segments display a variety of perfusion patterns.
Decreased regional myocardial tracer uptake at rest could reflect either lack of cell membrane integrity in an area of scarred myocardium or reduced MBF secondary to HM (dysfunctional but viable). Therefore, in patients with chronic ischemic LV dysfunction, myocardial “perfusion” tracers alone may not reliably differentiate hibernating from scarred myocardium. In that setting, flow-independent probes that assess intact cellular metabolic processes may be used as an adjunct to resting myocardial blood flow.
Despite the excellent flow kinetics and biological properties of thallium, attenuation of photons is a potential limitation for thallium in patients with large body habitus. PET systems have generally better sensitivity and spatial resolution than SPECT systems, and provide more accurate attenuation correction. Thus, metabolic imaging with FDG PET may provide incremental information to thallium regarding myocardial viability, especially in patients with severely impaired LVD.62 High-energy phosphates, such as ATP, are generated in the myocardium by oxidative phosphorylation and glycolysis. In the normal myocardium, utilization of free fatty acids is the preferred metabolic pathway for ATP production. In the setting of myocardial ischemia, where local oxygen supply is reduced, the myocardium shifts ATP production from fatty acid metabolism (which occurs in the mitochondria and is oxygen-dependent) to glucose utilization (which occurs in the cytoplasm and is oxygen-independent).
The principle of using a metabolic tracer for assessing myocardial viability is based on the concept that viable tissue is metabolically active, while scarred tissue is metabolically inactive. The glucose analogue FDG may be preserved or increased in hypoperfused but viable myocardium, termed metabolism/perfusion mismatch (Figure 5) 62a. Conversely, FDG uptake will be decreased or absent in hypoperfused and scarred myocardium, termed metabolism/perfusion match. The metabolism/perfusion mismatch pattern represents a scintigraphic marker of HM. PET imaging in dysfunctional myocardium had positive and negative predictive values of 85% and 92% respectively for recovery of function after revascularization.63-69 A subsequent European multicenter trial in 178 patients confirmed the high sensitivity of PET mismatch pattern for predicting functional recovery after revascularization70.
The prognostic significance of perfusion-metabolism mismatch pattern has been demonstrated in several nonrandomized, retrospective studies with PET. Patients with perfusion-metabolism mismatch pattern (HM) who were treated with revascularization had lower ischemic events and deaths when compared to those treated medically.7,71,72 Moreover, the extent of the PET mismatch pattern correlated with improvement in LV function, clinical course and magnitude of heart failure symptoms after revascularization.73 Patients with matched defects (concordant reduction in regional perfusion and metabolism), indicating scarred myocardium displayed no such difference in outcomes or clinical benefit from revascularization.7
The principal limitation for the widespread application of PET imaging for the assessment of myocardial metabolism is its availability. Currently in the United States, there are approximately 1,600 PET cameras versus 12,000 SPECT cameras. Thus, PET is not readily available for most practicing cardiologists. Because SPECT cameras are more prevalent than PET cameras in the community and cardiology offices, investigators have recently focused their attention on developing gamma-emitting metabolic radiotracers, such as β-Methyl-p-[123I]-iodophenyl-pentadecanoic acid (BMIPP), in order to make the assessment of myocardial metabolism readily available to patients presenting with symptoms of heart failure or LVD.74
BMIPP is a fatty acid analogue that provides insight into myocardial metabolism using SPECT cameras. In a recent study, delayed recovery of myocardial metabolism after ischemic injury, termed ischemic memory, was shown with BMIPP.+ Fatty acid metabolism remained suppressed for a prolonged time (up to 30 hours following an ischemic episode) even after perfusion had returned to normal. 74,75
Uptake of BMIPP from the plasma into myocardial cells occurs via CD36 transporter protein present on the sarcolemmal membrane. Retention of BMIPP in the myocardium most likely reflects activation of fatty acids by coenzyme A, and indirectly, of cellular ATP production resulting from fatty acid metabolism. Thus, in the setting of myocardial ischemia, reduction in ATP production secondary to diminished fatty acid metabolism is mirrored by decreased myocardial BMIPP uptake. Although BMIPP is approved for clinical use in Japan, it has not yet received approval by the Food and Drug Administration in the United States. In the clinical setting, the finding of persistent and prolonged disturbances in BMIPP uptake, long after resolution of ischemic symptoms, may provide a scintigraphic imprint as to the underlying cause of chronic left ventricular dysfunction.
The coronary microcirculation is preserved in myocardium with chronic ischemic LV dysfunction. The micro bubbles in echocardiographic contrast agents are intravascular tracers and the intensity of their reflected signal can be used to demonstrate the presence and relative size of the microcirculation, while their rate of accumulation is related to coronary flow. Myocardial contrast enhancement of dysfunctional segments at rest is therefore a potential marker for viable myocardium. The value of contrast echocardiography for this purpose has been reported in over ten studies, with sensitivities ranging from 62-92%, and specificity from 67-87%. Nonetheless, this technique remains technically challenging and may be difficult to use in all myocardial segments due to issues of shadowing and attenuation.
Resting scar imaging and perfusion imaging does not permit distinction between viable and remodeled myocardium. Assessment of MBF or CR is therefore mandatory combined with perfusion, metabolic or scar imaging.
Viable myocardium has been shown to augment function in response to inotropes (eg. dobutamine, amrinone), coronary vasodilators (eg. dipyridamole), arterial vasodilators (eg. levosimandan) and low level exercise. However, the most widely used stressor to augment function is low-dose dobutamine (LDDE), with an average sensitivity of 75-80% and a specificity of 80-85% for the prediction of functional recovery both early after infarction as well as in the setting of chronic LV dysfunction. Augmentation of function in response to low-dose dobutamine is a more reliable predictor of recovery than improvement at peak dose (eg. 40 mcg/kg/min). Nonetheless, the peak-dose response is important, because the most reliable predictor of functional recovery is the biphasic response (ie. augmentation at low-dose with deterioration at peak-dose) indicating that the tissue is not only viable but also supplied by a stenosed infarct-related artery (Figure 6A-6C). The extent of viability is an important determinant of the likelihood of recovery of overall LV function (eg. characterized by an LVEF improvement of >5%). The presence of ≥4 segments with a biphasic response has a specificity and sensitivity of 80-90% for prediction of global functional recovery. Patients demonstrating a significant LVEF improvement with LDDE are also likely to show global functional recovery and reverse remodelling.76
Echocardiography is an ideal initial screening test in these patients, because of its wide availability, low cost, and similar levels of accuracy to other more expensive modalities.58 Like any non-invasive test, there may be false-negative and false-positive responses with LDDE. The reliability of LDDE is influenced by a number of other factors apart from the extent of viable tissue, and false negative responses may be due to reduced substrate supply (compromised if the tissue is ischemic), loss of the myocardial contractile apparatus and excessive fibrosis (which may splint the segment and prevent it from shortening). If flow is severely compromised, the tissue may become ischemic before exhibiting an augmentation response. For this reason, LDDE should be performed with multiple low doses (eg. 5 and 10 mcg/kg/min - some authors have used a very low dose of 2.5 mcg/kg/min and doses higher than 10 mcg/kg/min), continually assessing the response, as augmentation may be very transient and ischemia is increasingly likely as the heart-rate increases. The chronotropic response to dobutamine usually starts at doses >10 mcg/kg/min, although the widespread use of long-acting beta blockers for HF has moved the “low-dose” threshold towards 20 mcg/kg/min. Another cause of false-negative responses is damage to the contractile apparatus within viable myocytes, caused by recurrent episodes of ischemia and stunning. Failure to respond to dobutamine may not indicate lack of viability, which may still be identified using PET; in these situations, prolonged follow-up may be needed to demonstrate functional recovery after revascularization. Non-transmural scar influences the likelihood of recovery of resting function, and may cause false positive responses. Subendocardial scar is a major contributor to resting dysfunction, so segments with non-transmural scar may respond to dobutamine because of augmentation of the subepicardium, but not recover resting function after revascularization. There are risks in the performance of peak dose dobutamine stress in individuals with severe LVD. The underlying reserve of the LV may be limited and hypotension, heart failure and serious arrhythmias may ensue if severe ischemia develops. These tests should be performed by experienced personnel, and the test may need to be stopped at the development of ischemia, and nitrates and oxygen may need to be administered. Nonetheless, there is no good evidence to indicate that there is a substantial increase of risk of dobutamine stress in the situation, compared to the usual risk of a significant adverse event in 3:1000 studies.
A problem relates to the challenges of interpreting LDDE. As with all stress echocardiographic techniques, wall motion scoring is limited by subjectivity and technical challenges. The measurement of myocardial deformation may prove to be a more reliable means of quantitation of this response. Strain-rate at LDDE correlates with the presence of myocardial viability evidenced by FDG imaging, with an optimal cut-point for this purpose being a SR increment of >0.23/second.77 A strain-rate increment of >0.25/second also predicts functional recovery after revascularization (sensitivity 80%, specificity 75%).78
The demonstration of CR is an established predictor of myocardial viability58 As with echocardiography, CMR assessment of CR is performed using the infusion of low doses of dobutamine (5-10 μg/kg/min). One study compared low dose dobutamine CMR to FDG-PET in 35 patients with chronic MI and segmental LVD.48 Using both EDWT and improvement in systolic thickening as markers of viability, CMR had a sensitivity, specificity, and diagnostic accuracy of 88%, 87%, and 92% respectively. Using improvement in LV function post-revascularization as the reference standard in the same patient group,49 dobutamine CMR had a sensitivity and specificity of 89% and 92%.
The utility of incorporating myocardial tagging for the assessment of intramyocardial functional reserve has also been demonstrated. In acute MI, the quantitative response to dobutamine is predictive of functional recovery, although the response in the subendocardium is quite complex.79 Studies in chronic ischemic heart disease before and after multi-vessel revascularization using tagged dobutamine CMR demonstrate that half of dysfunctional segments recover resting function, and, of the remainder, half demonstrate rest dysfunction but CR.80 Half of these latter segments in turn recover rest function when examined 3 years after revascularization.80
Some controversy exists as to which test performs best for predicting functional recovery among CMR-based techniques. LGE clearly identifies scar, but the relationship between scar and functional recovery in infarcts with 1-75% transmural infarction is complex. The performance of dobutamine CMR has been compared to LGE in this scenario.81 Although no difference was seen in the identification of viability in segments without LGE or those with scarring ≥75%, dobutamine CMR was superior in predicting recovery in zones in which LGE demonstrated between 1% and 75% transmural scar. 82 However, the performance of LGE varied greatly depending on the cutoff values used to define viability.82 Complementary use of dobutamine CMR and LGE may prove to be the optimal strategy for predicting post-revascularization functional recovery, as shown in a study of 15 patients studied before and 20 weeks after coronary artery bypass surgery.83 In segments with 1-50% infarct transmurality, as shown by LGE, dobutamine response was predictive of functional recovery, whereas the predictive value of LGE alone was intermediate. Wall thickening with dobutamine CMR before revascularization correlated with wall thickening at rest at follow-up.
This combined approach to defining viability by CMR may be warranted in certain clinical scenarios. While it may not be practical to perform low dose dobutamine infusion on all patients evaluated for viability, decision-making in those with multiple segments having 1-50% transmurality may benefit. In patients with no late enhancement or >50% late enhancement, there is little to gain by adding dobutamine, and LGE remains the preferred technique because of its ease of use.
Beyond the assessment of mere presence or absence of myocardial viability and predictive values of the various tests for recovery of regional or global LV function, it is perhaps equally important (if not more important) to demonstrate a survival advantage offered by revascularization for patients with ischemic LV dysfunction with viable myocardium. In a pooled analysis consisting of 3088 patients in 24 studies, long-term survival after revascularization or medical therapy was determined independent of the type of viability testing, which included SPECT radionuclide imaging, PET or dobutamine echocardiography (7). In patients exhibiting predominantly viable myocardium, follow up on medical therapy was associated with very high risk, a 16% annual mortality. On the other hand, in similar patients, revascularization was associated with an 80% reduction in annual mortality (16% vs. 3.2%, p< 0.0001), compared with medical therapy (Figure 7). Importantly, patients with the most severe LV dysfunction derived the greatest benefit from revascularization (Table 1). With worsening LV EF, the survival benefit associated with revascularization of patients with viable myocardium increased proportionately (Figure 8) (7).
None of the currently available imaging tools can address all the different aspects of the complex pathophysiology of myocardial viability and hibernation. Accordingly, they should often be used in combination in order to achieve the highest possible level of diagnostic accuracy. The clinician and non-invasive imager should note the following in the evaluation of reversible ischemic dysfunction:
There are several steps in the management of these patients.
Dr. Kramer was supported in part by RO1 HL075792 from National Institutes of Health, Bethesda, MD;
Dr. Marwick was supported by project grant 210217 from the National Health and Medical Research Council, Canberra, Australia (THM).
DISCLOSURE Dr. Rahimtoola has received Honoraria for educational lectures from American College of Cardiology Foundation; American College of Physicians; University of California Los Angeles; University of California Irvine; Cornell University; Creighton University; Thomas Jefferson University; Cedars-Sinai Medical Center; Harvard Medical School; University of Wisconsin; University of Hawaii; Cardiologists Association of Hong Kong, China; ATS; St. Jude Medical; Carbomedics; Merck; Pfizer; Edwards Life Sciences.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.