Only Hibernating Myocardium Invariably Shows Early Recovery After Coronary Revascularization
Background The aims of this study were to identify hibernating myocardium (hypocontractile, hypoperfused viable myocardium that regains contractility after revascularization) in the clinical setting and to predict functional outcome in patients with coronary artery disease after coronary revascularization.
Methods and Results Preoperative data related to the anterior free wall of the left ventricle were collected in 50 coronary bypass surgery candidates (positron emission tomography [PET], [13N]NH3 for flow, and [18F]FDG for metabolism [MET]; equilibrium-gated nuclear angiography [EGNA] for regional ejection fraction [REF]; and histological data from myocardial biopsies for percentage fibrosis and viable myocytes). Three months after surgery, the patients had follow-up PET and EGNA investigations. A principal-components analysis identified four patient clusters. Cluster 1 (n=9) had normal viable myocardium. Cluster 2 (n=18) had viable hypocontractile myocardium (REF, 39±12%) showing a PET mismatch pattern. Cluster 3 (n=16) had viable hypocontractile myocardium associated with morphological myocyte injury showing a matched moderate decrease in flow (66±11%) and MET (70±11%). Cluster 4 (n=7) had hypocontractile myocardium with mainly scar tissue (fibro-sis, 74±12%). After surgery, only cluster 2, with hibernating myocardium, showed significant improvement in REF (from 39±12% to 50±13%, P<.05). Cluster 3, with sites of morphological myocyte injury, showed no recovery. The stepwise logistic regression showed a combination of low preoperative REF and high MET to be the best predictor of functional recovery (P<.008).
Conclusions Multivariate analysis identifies hibernating myocardium showing early postrevascularization recovery, as opposed to viable but myolytic myocardium with no early recovery. Postrevascularization recovery can be predicted (combination of low REF and high MET) by noninvasive techniques.
A number of earlier studies1 2 3 4 have documented that significant abnormalities in the left ventricular segmental wall motion may also occur in patients without infarction. There were early reports of sometimes dramatic improvement in left ventricular function after coronary artery bypass surgery,5 6 7 and it was suggested that ischemic noninfarcted myocardium can exist in a state of “functional hibernation.”8
Using multivariate analysis, Flameng et al9 showed that in fact, a group of patients exists who have significant coronary artery stenosis and impaired wall motion in the absence of myocardial infarction. It was shown from biopsy material obtained during surgery that these regions consisted of viable myocardium that showed a variable degree of reduced volume fraction of sarcomeres and accumulation of glycogen. There was functional recovery of these regions after revascularization, and the patients showed a superior cardiac and total survival 2 years after bypass surgery compared with patients with postinfarction hypokinesis.
When Rahimtoola10 further advocated the term “hibernating myocardium,” he also referred to patients who had severe coronary disease and severe left ventricular dysfunction. This dysfunction was seen to reverse immediately after revascularization. It is a “smart heart,” he said, that can conserve its energy and remain viable in conditions of low perfusion.10 Indeed, hibernation is thought to be a state in which the viable myocardium adapts itself to low-flow conditions by downregulating its function. This is potentially to be recovered if adequately revascularized. With the development of the PET scan, which gives simultaneous information about perfusion and metabolism of a given region, it is possible to identify viable regions of the myocardium that could benefit from revascularization.11 12 13 Recently, more insight has been gained into the metabolic state and the regional blood flow in myocardium that is thought to be chronically ischemic, functionally inactive, but still viable.14 15 16 There have not been many studies to date, however, that could assess the postrevascularization functional and PET outcomes simultaneously and correlate these data with the morphological viability assessment in the same patients. Finally, little is known about the extent and timing of recovery of such functionally downregulated myocardium after revascularization.
We studied coronary bypass surgery candidates with a severe LAD stenosis or occlusion with or without regional dysfunction and tried to identify hibernating myocardium (defined as functional deficit in the presence of low perfusion and a normalization of function after revascularization) and predict the postrevascularization functional recovery using the currently available noninvasive techniques of the PET scan and EGNA in combination with transmural needle biopsies obtained during surgery.
Fifty patients (47 men and 3 women) undergoing coronary artery bypass graft surgery were included in the study. The study was approved by the institutional ethical committee, and all subjects gave informed consent. The selection criteria included (1) a diseased LAD and (2) only anterior wall infarction, if any. Ages ranged from 44 to 76 years (median, 62 years). The preoperative ECGs were classified according to the Minnesota code for Q and QS patterns in the anterior region. Fifteen of the 50 patients showed a Q-wave pattern in the anterior region. All patients took 100 to 160 mg/d aspirin; two thirds also took a β-blocker, and a smaller number also took other medications (one third, a calcium antagonist; another one third, various other medications for pulmonary disease, gout, hyperlipidemia, etc; and fewer than one quarter, nitrates and ACE inhibitors). After surgery, nitrates were withdrawn from the medications. There were no diabetics in our population. All patients underwent the following investigative procedures.
One to 3 months before surgery, all patients had a contrast angiography and ventriculography. No premedication was used. Coronary angiography was done with the Sones technique and was visually interpreted by experienced cardiologists. Significant CAD was defined as ≥80% stenosis of the coronary luminal diameter. The degree of collateralization was visually interpreted. After measurement of the left ventricular pressure, a biplane left ventriculography at 30° right anterior oblique projection was done. The ventricular image size was calibrated with a grid filmed at the level of the left ventricular cavity. Wall motion was measured from the end-diastolic and end-systolic outlines of the left ventricle in the right anterior oblique projection by use of routinely available software.
One to 2 days before surgery, all patients had an EGNA. Red blood cells were labeled with 20 mCi 99mTc by an in vitro technique. After centrifugation and removal of the supernatant, the labeled red blood cells were injected into the patients.17 An EGNA acquisition was obtained in the LAO 45° projection for 10 minutes, and two additional acquisitions were obtained for 5 minutes (LAO 70° and anteroposterior). A small field gamma camera (PHO/GAMMA V, Siemens) fitted with a high-resolution parallel collimator was used and connected to a dedicated computer (Sophy 20 workstation of Sopha Medical Benelux). At LAO 45° projection, there is minimum overlap of the left and the right ventricles, and the left ventricle is divided into eight sectors. Global EFs and REFs of the anterior wall were calculated, the normal value of REF being 60±22%. Three months after surgery, 42 of 50 patients had a control EGNA.
Positron Emission Tomography
One to 2 days before surgery, all patients had a PET scan. A whole-body positron emission tomograph (model 931-08/12, CTI Siemens) with eight detector rings allowing the acquisition of 15 planes with an interplane spacing of 6.75 mm was used. A small cyclotron (Cyclone 10/5, Ion Beam Applications) and auxiliary chemical equipment were used to produce [18F]FDG and [13N]NH3. A 2-minute rectilinear scan, used for positioning the heart within the field of view, was followed by a 15-minute transaxial transmission scan with a 68Ge ring source used for photon attenuation correction.
PET Myocardial Perfusion Imaging
[13N]NH3 (20 mCi) in 5 mL saline was slowly infused at a constant rate of 10 mL/min followed by a 20-mL flush with saline at the same rate. Acquisition was started simultaneously with the injection of [13N]NH3. Twenty-one frames were recorded in each patient, with a total acquisition time of 20 minutes.
PET Myocardial Metabolic Imaging
Metabolic studies were done after a 12-hour overnight fast by use of the euglycemic hyperinsulinemic clamp technique,18 which results in a postabsorptive steady state. A constant intravenous infusion of 20% glucose and insulin was started, and the glycemia was determined from serial blood samples withdrawn every 5 minutes. The rate of glucose infusion was adjusted to obtain normoglycemia, and the insulin infusion rate was adjusted according to the patient's weight. The glucose clamp was optimized during myocardial perfusion imaging. After stabilization of the glycemia between 85 and 95 mg/dL and after time was allowed for isotope decay (not earlier than 50 minutes after [13N]NH3 injection), 10 mCi of [18F]FDG was injected as a bolus. The acquisition was started immediately after the injection, with 22 frames being recorded in each patient. The total acquisition time was 70 minutes. The insulin infusion was stopped 15 minutes before the end of acquisition.
The 19 frames of the flow study were reconstructed by use of a Hanning filter (cutoff frequency, 0.3). A three-dimensional delineation of the left ventricular wall was used to construct a polar map of every frame of the dynamic study. A correction for spillover and recovery is based on the measured point spread function and on the epicardial and endocardial boundaries provided by the delineation. Since these same resolution effects preclude accurate estimation of the wall thickness, we imposed a constant thickness of 13 mm. The polar maps were divided into 33 regions: 1 apical and 4 rings of 8 regions. For every pixel of the polar map, the corresponding position in the left ventricular wall was known. This allowed us to locate the biopsy area in the polar maps according to the indications of the surgeon. We applied a three-compartment model19 20 to calculate absolute flow values from the tracer uptake curves of each region. After absolute values were obtained with this model, the flow in the anterior wall or biopsy area was expressed as a percentage of the maximum flow in the area considered to be normal (≤50% stenosis in the left circumflex or right coronary artery). Segments with flow values of >80% of flow in the reference segment (flow index, >0.8) were considered to be normal. The 22 frames of the metabolic study were reconstructed with a Hanning filter (cutoff frequency, 0.4). After creation of polar maps identical to these constructed for the flow studies, regional glucose utilization values were estimated from a Patlak graphical analysis.21 The flow reference zone was used as a reference region for FDG. The metabolism in the biopsy area is also expressed as a percentage of the glucose utilization in the reference zone. Three months after surgery, 28 of 50 patients had a PET scan.
Histology: Light Microscopy
During surgery, a transmural needle biopsy (Travenol Laboratories) was taken from the anterior wall of the left ventricle, 3.5 to 4 cm from the apex, in the perfusion domain of the LAD. The biopsy was immediately fixed in 3% glutaraldehyde buffered with 90 mmol/L KH2PO4 and adjusted to pH 7.4 with 0.1N KOH. After fixation for 2 hours at room temperature, the biopsy was rinsed in the above buffer, postfixed in 2% OsO4 solution buffered with 0.05 mol/L veronal acetate, dehydrated in graded series of ethanol, and embedded in epoxy resin. Semithin 2-μm sections were examined with light microscopy. Toluidine blue staining was used to determine the myocardial viability and degree of Tm fib in all biopsies by morphometry,22 expressed as a percentage of the entire area analyzed. The morphometric analysis was done as follows. A grid consisting of vertical and horizontal lines providing 121 intersections (points) was used. Counting the number of points overlying a certain structure results in a quantitative determination of the volume of the structure under investigation in relation to the volume of the entire tissue under the square grid. The total number of points is considered to be 100%, and the points counted in the connective tissue are expressed as a percentage of the entire tissue within the limits of the grid. The axis of the grid is then rotated ≈45°, and the points are counted again. The same procedure is repeated on a different area of the same section. Longitudinal sections at a magnification of ×250 are evaluated. Blood vessels and perivascular interstitial cells are excluded from the connective tissue. To quantify the degree of cellular changes, a minimum of 100 myocytes were evaluated per section, and only cells with a visible nucleus were included. PAS staining was used to show the presence of glycogen. Myocytes were planimetrically scored for the degree of glycogen. Percentage loss of sarcomeres was determined in the same way for every cell. A loss of at least 10% of myofibrillar material was necessary for cells to be classified as “affected,” “altered,” or “myolytic” cells (M cells). In this way, the average percentage of M cells per biopsy was calculated. Areas of special interest were selected for detailed electron microscopic examination.
All numerical data are presented as average values±SD. A univariate analysis was done on preoperative and postoperative data obtained from PET (flow and MET), EGNA (REF), and the biopsy material (Tm fib, M cells, and glycogen content) to find a correlation between these various parameters measured. A Pearson correlation coefficient was calculated and the significance determined (P<.05 taken as significant). Multiple logistic regression analysis was used to predict postrevascularization recovery (yes or no). The independent variables examined were the preoperative PET, EGNA, and biopsy data and their interac-tions. The probability of postrevascularization (P) recovery was determined by the formula log P/(1−P)=β0+β1×REF PRE+β2×MET PRE+β3×flow PRE. Principal-components analysis was used to obtain a biplot of the preoperative data of all patients. A biplot is a two-dimensional graphical display of the multivariate data.23 The term “biplot” is derived from the representation on the same graph of two features of the data, eg, the individual patients (by points) and the variability and correlation of the preoperative variables (by vectors). The length of the vector represents the SD of the respective variable. The angle between two vectors represents the correlation between the two variables. Furthermore, the distance between two points in the biplot (patients) represents the Mahalanobis distance between the original profiles of these patients. The orthogonal projection of a point on a vector gives an indication about the original value of that particular patient for that particular variable (eg, close to the center or a more outlying observation for that variable). In addition, a multivariate cluster analysis was done to divide the patients into nonoverlapping subgroups. This was a two-stage procedure. In the first step, the centroids of the clusters were chosen, and then all patients were classified to the nearest cluster. In the second step, a certain clustering criterion was checked, and depending on that criterion, patients were reallocated to another cluster. All statistical calculations were done with the statistical analysis system (SAS).24
Thirty-four of the 50 patients had three-vessel disease, 7 had two-vessel disease, and 9 had single-vessel disease of the LAD. In 9 patients, the LAD was totally occluded. The catheterization was done 1 to 3 months before surgical revascularization. During this period, the anterior wall motion could worsen, improve, or remain unchanged. Therefore, these data were disregarded during the analysis of the preoperative data.
The amount of connective tissue in the myocardium was assessed morphometrically. Although the biopsy was divided into a subepicardial and a subendocardial part and the percentage fibrosis was determined in these two parts, only the average value will be given as Tm fib. The Tm fib ranged from 4% to 88% in the patients. To compare the amount of connective tissue in biopsies of patients with CAD with that in biopsies from nonischemic hearts, we assessed transmural biopsies from 6 patients undergoing surgical correction of atrial septal defect. In these patients, the average Tm fib was 3±3%. Univariate analysis of data from the 50 patients showed that there was poor correlation between Tm fib and M cells (r=.13, P=.35).
The bulk of myocardial cells had structural features similar to those seen in nonischemic hearts. However, two important structural changes were repeatedly observed in the myocardium of CAD patients. First, a substantial number of myocytes showed variably decreased volume fraction of sarcomeres without showing any signs of atrophy. The loss of contractile material in these cells was limited to the vicinity of the nucleus. Some myocytes showed extensive loss of sarcomeres. Second, in the regions of myolysis, there was an accumulation of glycogen (strongly PAS positive) intermingled with numerous mitochondria. Signs of purely degenerative changes, such as acute necrosis of myocytes, intracellular edema, and abnormal lipid storage, were not seen. The percentage of normal myocytes and the percentage of M cells were assessed morphometrically per biopsy. Myolytic cells were defined as viable cells showing at least a 10% reduction of sarcomeres. Such M cells ranged from 2% to 81% among the patients. The degree of myolysis, however, was variable in these cells. The degree of glycogen accumulation provides an index of the severity of myolysis, because it appears that the space left by the dissolution of sarcomeres is partially replaced by glycogen (see Fig 1⇓).
PET and EGNA Data
All patients underwent a PET and EGNA study before surgery. The individual preoperative data ranged from very low to very high values, as follows: REF from 9% to 95% (mean±SD, 46±22), flow from 22% to 100% (67±18), and MET from 15% to 122% (81±24) compared with a control region. Forty-two of the 50 patients had EGNA again 2.5±0.5 months after surgery, and 28 of 50 patients had a PET scan. When the postoperative REF (45±24%), flow (78±20%), and MET (80±20%) were compared with the preoperative data, no significant changes were observed (P>.05). When the preoperative data are considered, there appears to be a strong positive correlation between REF and flow (r=.71, P<.0001) and a weaker correlation between REF and MET (r=.40), denoting that flow is a very good predictor of preoperative function. The postoperative data show a good correlation between all three variables (P<.0001, see Table 1⇓), denoting that the three variables are closely linked and that each can predict the other two. When the preoperative versus the postoperative data are considered, there is a definite improvement in the correlation between REF and MET (r=.68, P<.0001) and between flow and MET (r=.75, P<.0001), denoting the presence of viable tissue that benefits from revascularization (see Fig 4⇓). An increase of 20% or more may be considered to be an improvement in REF.
PET, EGNA, and Histology Data
The preoperative data show a negative correlation between REF, flow, and MET versus Tm fib (see Table 1⇑), with MET having the best negative correlation with Tm fib (r=−.83, P<.0001). After surgery, there is a better correlation between the above variables (P<.0001), denoting that Tm fib plays an important role in determining the postoperative regional outcome in terms of function, flow, and metabolism. A poor correlation was observed regarding the myocyte morphological changes and the preoperative and postoperative EGNA and PET data.
For prediction of postrevascularization functional recovery, a stepwise logistic regression analysis showed that low REF in association with high MET had the best predictive value for functional recovery (P<.008). A combination of the data obtained from the histology did not appear to have a very high predictive value regarding functional recovery.
Principal-components analysis was performed using preoperative data of REF, flow, and MET and the histological data of Tm fib, M cells, and glycogen. From these data, four subpopulations could be identified. These subpopulations are defined as follows, and for each group the mean values±SD were calculated for the different variables (see Fig 2⇓ and Table 2⇓).
Cluster 1 patients (n=9) had normal REF, flow, and MET (77±13%, 91±10%, and 101±8%, respectively). The global EF was also normal at 68±8%. Histological analysis showed a low degree of Tm fib (8±4%), a low number of cells with myofibrillar loss (M cells, 12±8%), and low glycogen content contained in central plaques. Three months after surgery, all patients had an EGNA, and 7 of 9 had a PET scan. These patients did not show a significant change in REF, flow, or MET (71±16%, 92±10%, and 95±7%, respectively, P>.05).
Cluster 2 patients (n=18) had low REF and moderately decreased flow but almost normal MET (39±12%, 67±10%, and 97±13%, respectively). The global EF was low at 51±10%. There was little Tm fib (13±10%), and only some myocytes showed subcellular alterations (M cells, 24±10%, P>.05 versus cluster 1), with low glycogen content. Three months after surgery, 15 of 18 patients had an EGNA, and 14 of 18 had a PET scan. These patients had the best postoperative outcome, showing a significant recovery of REF (50±13%, P<.05) after 3 months (see Fig 3⇓). The flow was also significantly higher (78±8%, P<.05), denoting successful revascularization. There was no significant change in the MET at 89±21% (P>.05).
Cluster 3 patients (n=16) had low REF and a matched moderate decrease in flow and MET (44±18%, 66±11%, and 70±11%, respectively). The global EF was low at 53±15%. The Tm fib was increased at 34±10% (P<.05 versus clusters 1 and 2), with a substantial number of myocytes showing a variable loss of myofibrils (M cells, 44±20%, P<.05) and substantial glycogen accumulation. Three months after surgery, 15 of 16 patients had an EGNA and 5 of 16 had a PET scan. There was no significant improvement in the REF for the group as a whole (47±17%). Only 7 of 15 patients showed an improvement in the REF. With regard to regional flow, MET, and histological aspects, these patients had preoperative characteristics similar to those who did not recover.
Cluster 4 patients (n=7) had low REF, flow, and MET (29±13%, 39±14%, and 39±13%, respectively); a very high degree of Tm fib (74±12%); and a substantial degree of M cells (22±19%) and glycogen content. Three months after surgery, 5 of 7 patients had an EGNA and 2 of 7 had a PET scan. There was no significant change in the REF (19±14%, P>.05).
For correlation between myocardial fibrosis and function, flow, and metabolism, the transmural fibrosis appears to be the major determinant of function, flow, and metabolism before and after surgery, showing a significant correlation with these parameters (see Table 1⇑ and Fig 4⇓). After surgery, there is a further improvement in this correlation.
The most important issue in the treatment of CAD is the effect on survival. Hibernating myocardium, defined originally as a dysfunctional myocardial state caused by chronic hypoperfusion, has aroused great interest because of its potential to improve function and consequently also survival after revascularization.
We studied 50 patients with CAD, all of them with a severe LAD stenosis. This study provides detailed regional data, including preoperative function, perfusion, and metabolic and histological analyses. A resting flow of <80% compared with the control region was considered to be a flow deficit. In cluster 1 patients (with normal preoperative EGNA and PET data), there appears to be perfect coupling between perfusion-metabolism and function. These patients were considered to be candidates for coronary artery bypass graft surgery to provide symptomatic relief from the episodes of ischemia that occur when there is increased demand on the heart during daily activity, an increase in microvascular tone, or reductions in blood pressure or with a superimposed vasoconstriction over a given stenosis. After surgery, as one might expect, there were no significant changes in the REF, flow, or MET. In cluster 2 patients with a low degree of Tm fib, we observed a downregulation of function in accordance with the reduced resting flow. The MET, however, was high, showing a state of relative mismatch between flow on the one hand and function on the other. In this group, we had two criteria to define viability: the low Tm fib and the near-normal MET. These patients appeared to benefit the most from revascularization, showing a significant recovery of function in the presence of improved perfusion as measured 3 months later. This improvement might have been evident earlier had we measured the function at an earlier stage. Functional recovery, however, was not complete, even though these patients had predominantly normal structure and very little loss of contractile material. The reason for this incomplete recovery is indeed very puzzling. Early improvement in function has been described in experimental situations.25 26 However, no one as yet has succeeded experimentally in reproducing a chronic situation that exists in patients with either constant or intermittent low-flow conditions and their metabolic consequences.
A minimal level of blood flow is required to sustain viability.27 Severe flow reductions can accurately define myocardium as nonviable, but mild to moderate reductions in blood flow discriminate poorly between viable and nonviable tissue. In our study, viability data were obtained from two sources in the same patient, namely, the FDG uptake and the histology. In cluster 3 patients showing a matched moderate decrease in flow and metabolism, 7 of 15 patients showed an improvement in the REF. However, no recovery was observed 3 months after revascularization for the group as a whole. These patients differed from all others in having increased myocyte glycogen accumulation and a high number of M cells showing a variable loss of contractile material (44±20%). The inconsistent functional recovery observed in these patients once again raises the question of whether these myolytic cells could be responsible for the lack of functional recovery. A substantial number of the altered cells in cluster 3 patients had lost almost all the contractile material and stored a lot of glycogen. Such a degree of morphological myocardial injury may imply that the myocytes might require a very long time to rebuild the contractile machinery, if at all. Evidence for further morphological alterations comes from Ausma et al,28 29 who showed that there is an alteration in the distribution of titin in cells with myolysis. Titin is one of the earliest markers of cardiac differentiation, and the changes in distribution observed in myocytes with myolysis resemble those occurring during muscle cell differentiation, but then in reversed order.28 Furthermore, α-smooth muscle actin, which gradually disappears from adult cardiac myocytes, becomes reexpressed in adult myocytes found at sites showing the morphological alterations described above.29 In human dilated cardiomyopathic hearts, Hein et al30 used double-staining procedures to show that defects of contractile apparatus were accompanied by a simultaneous reduction of titin. Abnormalities of titin expression may be important, because this protein influences sarcomeric elastic behavior and is necessary as a template for the organization of newly synthesized myosin and actin filaments.
With regard to the energy content of myolytic cells, it appears that the high-energy phosphate content of these viable myocytes is relatively preserved. Flameng et al31 studied myocardial biopsies from patients with severe LAD stenosis and abnormal wall motion and without ECG evidence of infarction. The adenylate pool in myocardial biopsies taken from these patients was only partially depleted, but the mitochondrial function was intact, and the myocyte morphology, although altered, was that of viable cells. This was in contrast to patients with evolving infarctions who had a decreased adenylate pool (up to 52%), accumulation of nucleotides (inosine), altered mitochondrial function, and myocytes showing irreversible ultrastructural changes.
The degree of myocardial fibrosis remains, in the end, an absolute proof of viability. In cluster 4 patients with a high degree of Tm fib (74±12%), once again the coupling between perfusion-metabolism and function is observed. There is a matched decrease in function, flow, and MET. These patients did not benefit from revascularization in any way.
The dynamic nature of CAD contributes to the difficulties in defining hibernating myocardium and the underlying pathophysiological mechanisms. A regional flow reduction either could involve the entire myocardial wall or could be due to the coexistence of scar tissue in the subendocardium and normal myocardium in the subepicardium. A restoration of blood flow will improve wall motion in the first instance but not in the second. A variable functional outcome after revascularization may therefore be expected.32 Flow and metabolic patterns can provide accurate indexes of the myocardial state. Schelbert and coworkers11 12 33 have done some pioneering work in this field and have provided us with early evidence that it is possible to identify a subpopulation of patients with hypoperfused (or even normally perfused) dysfunctional but viable tissue (as identified by PET) that benefits from revascularization. In our study, tissue viability was identified by use of a combination of the “gold standard,” which is histology, and PET data. Our work confirms the findings of Schelbert and coworkers and identifies another entity of patients with viable myocardium showing a moderate matched decrease in flow and metabolism and with a variable functional outcome. The mechanisms involved here might be different from those involved in the “classic myocardial hibernation” (which shows intact structure and early recovery after revascularization), since these patients show definite subcellular alterations of viable myocardium and increased interstitial fibrosis. In our opinion, only a multivariate approach can identify such an entity.
To summarize, this study provides detailed regional data in 50 patients, including preoperative functional, perfusion, metabolic, and histological characteristics and a postoperative functional follow-up in the same patients. Regional viability was determined by a combination of histology (gold standard) and PET data. Three important observations are made in this study: (1) It is possible to identify a subpopulation of patients that satisfies the requirements of hibernating myocardium and shows early functional benefit from revascularization. Here the subcellular structure is preserved. The underlying mechanism for regional dysfunction could be a multiplicity of ischemic insults (superimposed on myocardium that is hypoperfused or normally perfused in the resting state), causing repetitive stunning. (2) A subset of patients exists with viable myocytes showing extensive morphological alterations. The underlying mechanism involved in this structural adaptation might be speculated to be an initial prolonged severe ischemic insult with subsequent chronic hypoperfusion or subsequent multiplicity of ischemic insults, the outcome being a coexistence of increased subendocardial fibrosis and a persistent stunning effect. These patients did not show early recovery but might be speculated to show a late recovery after revascularization. It is possible that without revascularization, such a state of affairs might evolve to cellular degeneration and increased regional connective tissue. (3) Low REF in association with high MET was the best predictor of functional recovery. This further confirms the early observation made by Tillisch et al.33 A further addition of Tm fib to the stepwise logistic analysis did not improve the predictive value. In our study, we were able to double-check viability because of the transmural myocardial biopsies and the MET obtained from the PET study. Statistically, there was a very good negative correlation (P<.0001) between the Tm fib and MET, ie, the higher the MET, the lower the Tm fib, with a very good predictive value (R2=.83). In most studies in which it would not be possible to obtain myocardial biopsies, metabolic data obtained from PET studies would therefore appear to be a reliable parameter (in combination with function) of myocardial viability. Finally, our data further confirm that the preoperative function, flow, and metabolism and the postoperative outcome are largely determined by the degree of transmural fibrosis (see Fig 4⇑).
The very nature of our study (biopsy material obtained from living patients during surgery) obviously imposes sampling limitations. Therefore, only one transmural needle biopsy was taken from the anterior wall of each patient. The histological analysis, although not representative of the entire anterior wall, nevertheless provides us with more than just an indication of viability and structural adaptation.
Only 28 of the 50 patients had a PET scan after surgery. This cumbersome investigation was refused by most of the remaining 22 patients (because there were no therapeutic consequences and because of the time factor). Our major interest was postoperative functional recovery; thus, an incomplete postoperative PET follow-up did not interfere with our goal.
Our study shows that a multivariate approach using data obtained from currently available noninvasive techniques can identify different subpopulations with specific defining characteristics in patients with chronic CAD. In our patient population, it was exclusively cluster 2 (functionally downregulated viable myocardium) that showed early functional improvement after revascularization. In cluster 3 patients with myocytes showing extensive myofibrillar loss (structurally downregulated viable myocardium), there was no recovery 3 months after revascularization. A delayed normalization of the function might be speculated if a complete resynthesis of the lost myofibrils were possible. The preoperative data obtained from PET and EGNA (low REF in combination with high metabolism) were very useful in predicting recovery if there was adequate revascularization. The degree of fibrosis plays an important role in the determination of preoperative and postoperative function, flow, and metabolism.
The authors wish to thank F. Thone´ for his help with the histology and I. Verhaegen and S. Vleugels for their help with the acquisition of the PET studies.
Selected Abbreviations and Acronyms
|CAD||=||coronary artery disease|
|EGNA||=||equilibrium-gated nuclear angiography|
|LAD||=||left anterior descending coronary artery|
|LAO||=||left anterior oblique|
|PET||=||positron emission tomography|
|REF||=||regional ejection fraction|
|Tm fib||=||transmural fibrosis|
Reprint requests to Prof Dr Willem Flameng, Center for Experimental Surgery and Anesthesiology, Minderbroeder Str 17, B-3000 Leuven, Belgium.
- Received November 6, 1995.
- Revision received January 16, 1996.
- Accepted January 22, 1996.
- Copyright © 1996 by American Heart Association
Bjork L, Culhead I, Hallen A. Cine´ angiographic studies of the left ventricle in patients with angina pectoris. Circulation. 1967;36:868-877.
Chatterjee K, Swan HJC, Parmley WW, Sustaita H, Marcus H, Matlott J. Depression of left ventricular function due to acute myocardial ischemia and its reversal following aorto-coronary saphenous vein bypass. N Engl J Med. 1972;282:1117-1172.
Bodenheimer M, Banka V, Hermann G, Trent R, Pasdar H, Helfant R. Reversible asynergy: histopathologic and electrocardiographic correlations in patients with coronary artery disease. Circulation. 1976;53:792-796.
Chatterjee K, Swan HJC, Parmley W, Sustaite H, Marcus H, Matloff J. Influence of direct myocardial revascularization on left ventricular asynergy and function in patients with coronary heart disease. Circulation. 1973;47:276-286.
Rutherford BD, Gan GT, Danielson GK. Left ventricular hemody-namics before and soon after saphenous vein bypass graft operation for angina pectoris. Br Heart J. 1972;34:1156-1162.
Flameng W, Wouters L, Sergeant P, Lewi P, Borgers M, Thone´ F, Suy R. Multivariate analysis of angiographic, histologic and electrocardiographic data in patients with coronary heart disease. Circulation. 1984;70:7-17.
Brunken R, Tillisch J, Schwaiger M, Child J, Marshall R, Mandelkern M, Phelps M, Schelbert H. Regional perfusion, glucose metabolism, and wall motion in patients with chronic electrocardiographic Q-wave infarction: evidence for persistence of viable tissue in some infarcted regions by positron emission tomography. Circulation. 1986;73:951-963.
Eitzman D, Al-Aouar Z, Kanter H, vom Dahl J, Kirsh M, Deeb GM, Schwaiger M. Clinical outcome of patients with advanced coronary artery disease: outcome of patients with advanced coronary artery disease after viability studies with positron emission tomography. J Am Coll Cardiol. 1992;20:559-565.
Schelbert H. Positron emission tomography for the assessment of myocardial viability. Circulation. 1991;84(suppl I):I-122-I-131.
Vanoverschelde JL, Wijns W, Depre´ C, Essamri B, Heyndrickx G, Borgers M, Bol A, Melin J. Mechanisms of chronic regional postischemic dysfunction in humans: new insights from the study of noninfarcted collateral-dependent myocardium. Circulation. 1993;87:1513-1523.
Shivalkar B, Mortelmans L, Schiepers C, Borgers M, Nuyts J, Bormans G, Flameng W. Glycogen storage in chronically hypoperfused myocardium associated with maintained FDG uptake. Circulation. 1992;86(suppl I):I-417. Abstract.
Muzik O, Beanlands RSB, Hutchins G, Manger TJ, Nguyen N, Schwaiger M. Validation of nitrogen-13-ammonia tracer kinetic model for the quantification of myocardial blood flow using PET. J Nucl Med. 1993;34:83-91.
Bol A, Melin JA, Vanoverschelde JL, Baudhuin T, Vogelaers D, De Pauw M, Michel C, Luxen A, Labar D, Cogneau M, Robert A, Heyndrickx GR, Wijns W. Direct comparison of [13N]ammonia and [15O]water estimates of perfusion with quantification of regional myocardial blood flow by microspheres. Circulation. 1993;87:512-525.
SAS/STAT User's Guide, Version 6, Fourth Edition. Cary, NC: SAS Institute Inc; 1989.
Schulz R, Joechen R, Martin C, Brodde OE, Heusch G. Development of short term myocardial hibernation: its limitation by the severity of ischemia and inotropic stimulation. Circulation. 1993;88:684-695.
Bolukoglu H, Liedtke J, Nellis S, Eggleston A, Subramanian R, Renstrom B. An animal model of chronic coronary stenosis resulting in hibernating myocardium. Am J Physiol. 1992;263(Heart Circ Physiol 32):H20-H29.
Gewirtz H, Fischman A, Abraham S, Gilson M, Strauss W, Alpert N. Positron emission tomographic measurements of absolute regional myocardial blood flow permits identification of nonviable myocardium in patients with chronic myocardial infarction. J Am Coll Cardiol. 1994;23:851-859.
Ausma J, Schaart G, Thone´ F, Shivalkar B, Flameng W, Depre´ C, Vanoverschelde JL, Ramaekers F, Borgers M. Chronic ischemic viable myocardium in man: aspects of dedifferentiation. Cardiovasc Pathol. 1995;4:29-37.
Schelbert H. Different roads to the assessment of myocardial viability: lessons from PET for SPECT. Circulation. 1995;91:1894-1895.