18F-2-Deoxyglucose Deposition and Regional Flow in Pigs With Chronically Dysfunctional Myocardium
Evidence for Transmural Variations in Chronic Hibernating Myocardium
Background Hibernating myocardium in patients with collateral-dependent myocardium is characterized by relative reductions in resting flow and increases in the uptake of 18F-2-deoxyglucose (FDG) in the fasting state. We performed the present study to examine whether these key physiological alterations could be produced in a porcine model of chronic coronary occlusion and to assess whether the adaptations consistent with hibernation varied across the myocardial wall.
Methods and Results We chronically instrumented pigs (n=18) with a fixed occluder on the proximal left anterior descending coronary artery (LAD). Three months later, ventricular function, regional myocardial perfusion, and FDG deposition (by excised tissue counting or positron emission tomography) were assessed in pigs after an overnight fast in the closed-chest anesthetized state. Total LAD occlusion with angiographic collaterals was present in the majority of animals. Left ventriculography showed severe anterior hypokinesis, and resting perfusion was significantly reduced in the hibernating LAD region in comparison with the normal remote regions (subendocardium: 0.80±0.06 versus 1.07±0.06 mL·min−1·g−1, P<.001; full-thickness: 0.87±0.04 versus 0.99±0.06 mL·min−1·g−1, P<.01). There was a twofold increase in full-thickness fasting FDG uptake in the dysfunctional LAD region (1.8±0.2 by positron emission tomography versus 1.9±0.1 by ex vivo counting). Ex vivo tissue counting revealed a pronounced transmural variation in FDG uptake in the hibernating region (LAD/normal), which averaged 2.5±0.2 in the subendocardium, 1.9±0.2 in the midmyocardium, and 1.4±0.1 in the subepicardium.
Conclusions These results demonstrate that pigs instrumented with a proximal LAD stenosis develop hibernating myocardium characterized by relative reductions in resting function and perfusion in association with increased uptake of FDG in the fasting state. The transmural variations in relative resting flow and FDG uptake suggest that myocardial adaptations consistent with hibernation are most pronounced in the subendocardial layers and vary in relation to local coronary flow reserve.
An increasing effort is being made to identify patients with viable but chronically dysfunctional myocardium. Reversible dyssynergy can arise through a number of different mechanisms, including postischemic stunning and chronic hibernation.1 2 3 Metabolic imaging with positron emission tomography (PET) has been very helpful in identifying viability in patients with chronically dysfunctional myocardium. Both retrospective4 and prospective5 studies have demonstrated that the presence of a metabolic/flow mismatch with increased 18F-2-deoxyglucose (FDG) deposition relative to resting flow is predictive of viability and associated with an improvement in myocardial function after coronary bypass surgery. The extent to which this mismatch pattern reflects the effects of reduced resting flow as opposed to increases in relative FDG uptake is difficult to determine because of the variabilities in the techniques used to stimulate FDG uptake. In addition, whether the physiological adaptations observed in hibernating myocardium vary between the subendocardial and subepicardial layers of the heart remains unanswered because of the limited spatial resolution of clinical imaging modalities as well as the difficulties in reproducing the key physiological features of hibernating myocardium in a chronic animal model.
To investigate this in more detail, we used techniques in which transmural flow (microspheres) and FDG uptake (excised tissue counting) could be assessed in myocardial tissue samples from a chronic animal model of hibernating myocardium. We used a technique to provoke coronary collateral growth in pigs that was originally described by Millard6 ; the proximal left anterior descending coronary artery (LAD) was banded to a fixed dimension. Mills et al7 recently demonstrated that a similar animal model of chronic coronary stenosis is associated with some of the features of hibernating myocardium in humans, including reductions in resting myocardial perfusion and oxygen consumption in the absence of tissue necrosis. In the present study, we sought to determine whether myocardial function was reduced in this model and whether the dysfunctional myocardium had an increase in FDG deposition in the fasting state by PET imaging, as has been recently described in patients with hibernating myocardium.8 The second objective of our study was to determine whether there are transmural variations in perfusion and FDG deposition across the wall of hibernating regions. The results demonstrate that the magnitudes of resting flow reduction and fasting FDG uptake are critically related to local coronary flow reserve and that both changes are most pronounced in the subendocardium of pigs with chronic hibernation.
The studies were conducted in farm-bred pigs. All experimental procedures and protocols conformed to institutional guidelines for the care and use of animals in research.
Juvenile pigs were fasted overnight. On the morning of surgery, they were premedicated with a mixture of Telazol (tiletamine [50 mg/mL]/zolazepam [50 mg/mL])/ketamine [100 mg/mL] (0.037 mL/kg IM) and administered prophylactic antibiotics (50 mg/kg IV cephalothin and 5 mg/kg IM gentamicin). After endotracheal intubation, they were mechanically ventilated, and a surgical plane of anesthesia was maintained with a halothane (0.5% to 2%) and oxygen (balance) mixture. A thoracotomy was performed in the fourth left intercostal space and a limited pericardiotomy used to expose the proximal LAD immediately before or after the first major diagonal branch. The proximal LAD was dissected free and instrumented with a 5-mm-long Delran occluder with a fixed internal diameter of 1.5 to 2.25 mm (average±SEM, 1.6±0.1 mm) that was similar to that previously described by Folts et al.9 To insert the artery, a longitudinal groove was machined in the occluder, and the artery was secured within the occluder with a silk ligature or umbilical tape. The chest incision was closed in layers, the intercostal nerves were infiltrated with 2% lidocaine for analgesia, and the pneumothorax was evacuated. A single postoperative dose of antibiotics was administered after the chest was closed, and intramuscular analgesics (0.025 mg/kg IM butorphanol) were administered postoperatively as required to alleviate pain. Animals were fully recovered within 24 hours. They were group housed and fed a diet of Pro-Lean (Agway Inc) ad libitum.
Animals were returned to the laboratory for studies after ≈3 to 4 months. The pigs were fasted overnight, and anesthesia was induced with a mixture of Telazol/xylazine (100 mg/mL) (0.022 mL/kg IM). After tracheal intubation, a surgical plane of anesthesia was maintained with a halothane (0.5% to 1%) and oxygen (balance) mixture supplemented with intramuscular doses of Telazol/xylazine (0.011 mL/kg IM) as required. Cutdowns were used to expose the left and right carotid arteries for catheter placement. The left carotid artery was instrumented with an 8F introducer, through which a 7F pigtail catheter was placed into the left ventricle. Arterial pressure and reference withdrawal samples for microsphere flow measurements were taken from the sideport of the introducer. In 6 animals, microspheres were injected into the left ventricle. In the remaining animals, a pigtail catheter was placed retrograde into the left atrium for microsphere injection. Catheters were placed in the jugular veins to administer fluids (normal saline) and pharmacological agents throughout the study. After completion of the instrumentation, the animals were administered heparin (10 000 units IV), and hemodynamics were allowed to equilibrate for ≥30 minutes before the protocol was started.
After the equilibration period, colored microspheres were injected into the left atrium or left ventricle to allow assessment of regional perfusion under resting conditions. After the initial flow measurement was completed, myocardial function was assessed with contrast left ventriculography. In 11 of the animals, responses to a submaximal inotropic stimulus were assessed with an intravenous infusion of epinephrine that was titrated so that heart rate increased by ≈50 bpm above resting values (0.32±0.04 μg·kg−1·min−1 IV). Once hemodynamics reached a steady state (≈10 minutes), a second microsphere flow measurement was performed, followed by a left ventriculogram. The epinephrine was stopped, and hemodynamics were allowed to return to baseline. Pharmacological vasodilation was produced to assess regional coronary flow reserve using an intravenous infusion of adenosine (0.9 mg·kg−1·min−1 IV). Because this dose of adenosine was accompanied by arterial hypotension, a simultaneous infusion of phenylephrine (5.4±0.9 μg·kg−1·min−1 IV) was started and adjusted to restore mean arterial blood pressure toward baseline levels. After completion of the last microsphere injection, hemodynamics were allowed to return to baseline, and selective coronary angiography was performed to assess the anatomic severity of the LAD stenosis as well as the angiographic extent of collateralization.
Transmural Analysis of FDG by Ex Vivo Tissue Counting
At ≈60 to 70 minutes after the last pharmacological intervention, 9 animals (as well as a sham-operated control) were heparinized (10 000 units IV), and 2 to 8 mCi of FDG was injected intravenously as a rapid bolus. Arterial blood samples (2 mL) were withdrawn at frequent intervals to characterize the arterial FDG time-activity curve. At 45 minutes after the administration of FDG, the heart was rapidly excised for tissue sampling. The left ventricle was weighed, and a 1.5-cm-thick concentric ring parallel to the atrioventricular groove and midway between the occluder and apex was used for analysis of microsphere perfusion and FDG deposition (average weight, 41.4±3.2 g). The concentric ring was radially divided into 11 to 15 full-thickness wedges, which were subsequently subdivided into three layers of approximately equal thickness. Myocardial samples were placed into tared vials, and positron emissions were quantified through direct measurement of annihilation γ radiation at 511 keV in a germanium well detector (Canberra Inc). Myocardial activity was expressed as counts per minute per gram of wet weight of tissue. Arterial blood samples were counted in a similar fashion and normalized to the sample volume (counts·min−1·mL of blood−1). All tissue and blood samples were decay corrected to the time of FDG administration. We performed Euler integration of the arterial time-activity curves to normalize tissue activity for differences in the administered FDG activity among animals. Sample activities were subsequently divided by the arterial input function to obtain the regional deposition or retention fraction, as previously described.4 10
Average values for FDG deposition in the hibernating LAD and control regions were obtained by determining the weighted mean values for all of the samples within a given region after the perfusion boundaries were determined by analyzing the circumferential distribution of myocardial perfusion during pharmacological vasodilation (see below). Border samples between the hibernating region and normal regions were excluded. Approximately half of the tissue samples were usually supplied by the LAD perfusion territory (LAD, 18.6±2.2 g; normal, 16.1±2.7 g). After tissue activity had been counted, the myocardial samples were frozen at −20°C for ≈48 hours to allow the FDG to decay to background levels. They were subsequently processed to assess regional myocardial perfusion in each tissue sample, as outlined below.
Fasting FDG Uptake by In Vivo PET
At ≈1 week (8.0±1.1 days) before the terminal physiological study, 9 animals underwent FDG imaging by PET. Animals were fasted overnight, sedated with Telazol (100 mg/mL)/xylazine (100 mg/mL) (0.022 mL/kg IM), and transported to the PET suite. Supplemental doses of Telazol/xylazine (0.011 mL/kg IM) were given as needed. The animals were positioned in a 31-slice ECAT 951/31-R PET scanner (Siemens/CTI) with a 10.8-cm axial field of view and a resolution of ≈5.9 mm3 full-width-at-half-maximum (Ramp filter 0.5 cycles/pixel cutoff). A 5-minute transmission scan was used to confirm correct positioning, and then a 20-minute transmission scan was performed with retractable 68Ge rod sources to correct for attenuation. After blood glucose was measured, 5 to 10 mCi FDG was administered via a cannulated ear vein as a slow bolus. Dynamic imaging was begun at the time of tracer injection and continued for 54 minutes (8×15 seconds, 4×30 seconds, 10×300 seconds).
Regions of interest (n=4 for each territory) were drawn on single-frame transaxial images (summed final four dynamic frames, Hann filter, 0.3 cycles/pixel cutoff) in the midanteroapical wall corresponding to the LAD perfusion territory, with similar regions drawn in the posterolateral and inferior walls corresponding to the normal perfusion territory. Four additional regions were drawn in the left atrium and left ventricle for determination of arterial activity. Spillover into the blood pool was minimized by drawing small regions of interest approximately two full-widths-at-half-maximum away from the cardiac walls and using high spatial resolution images as the data source. The final quantitative time-activity curves for tissue regions and blood pool were obtained by copying each of the regions of interest on the full set of dynamic images (Ramp filter, 0.5 cycles/pixel cutoff) and extracting regions of interest values (μCi/mL) for each frame of data. The regions of interest for each territory (LAD, normal, blood pool) were averaged for subsequent analysis.
Decay-corrected tissue and blood time-activity curves were automatically analyzed using the technique of Patlak and Blasberg.11 The slope of the linear portion was determined by least-squares singular value decomposition fitting to the final 10 frames of data and is equal to the FDG utilization constant (K). The regional rate of metabolic glucose utilization (rMGU) was determined with the following equation: rMGU=K·Pglu/L, where Pglu is the plasma glucose and L is the lumped constant that corrects FDG kinetics for the metabolism of glucose in myocardium. The lumped constant was assumed to be 0.67.12 In seven of the animals, terminal physiological studies were performed as outlined above.
Left Ventriculography and Coronary Angiography
Myocardial function was assessed with contrast left ventriculography performed in the left lateral projection. This approximates a view that is similar to the shallow right anterior oblique projection used for left ventriculography in humans. Approximately 10 to 20 mL of warmed nonionic contrast (iohexol 365 mg iodine/mL, Winthrop Pharmaceuticals) was hand injected over 3 to 6 seconds to visualize the left ventricle. Adequate images could not be recorded for two ventriculograms due to failure of the fluoroscopy unit. Fluoroscopic images were recorded and stored on Super VHS tape for later playback and off-line analysis. End-systolic and end-diastolic images of the left ventricle were traced by two observers. Global ejection fraction was calculated from digitized tracings using the area-length method,13 and the measurements from each observer were averaged. Each observer also graded regional wall motion in the anterior wall using the following scoring system (3, normal; 2, mild hypokinesis; 1, severe hypokinesis; and 0, akinesis). Dyskinesis was not present under any condition.
Selective left and right coronary angiography was performed using a 7F Sones or a 7F Judkins right (4R) coronary catheter and hand injections of Renografin (Squibb Diagnostics). Images were recorded on Super VHS tape for later review. In cases in which antegrade filling of the LAD through the native vessels versus collaterals was in question, subselective left coronary injections were performed into the left circumflex and the LAD. Each series of angiograms was interpreted by two observers regarding the extent of collaterals (0, none; 1, faint opacification of the distal LAD; 2, delayed but complete opacification of the LAD; and 3, rapid, complete opacification of the LAD). Caliper measurements of magnified coronary angiograms were used to quantify the percent diameter reduction of the LAD stenosis in relation to the proximal and distal native vessels. The right coronary artery could not be engaged in 1 animal with a patent but stenotic LAD and in 1 animal with an occluded LAD that had grade 3 left-to-left collaterals.
Microsphere Flow Measurements
Regional perfusion was assessed with colored microspheres (15-μm diameter, Dye-Trak, Triton Inc) according to previously published techniques.14 Briefly, microspheres suspended in saline with thimerosal (0.01%) and Tween 80 (0.01%) were sonicated and vortex agitated before injection. Approximately 3 to 5 million microspheres labeled with one of four differently colored dyes (yellow, red, white, and blue) were administered as a bolus through the left atrial or ventricular catheter after confirmation of the position of the catheter through the use of pressure tracings and fluoroscopy. Immediately before injection, an arterial reference sample was started from the left carotid artery at a withdrawal rate of 6 mL/min and continued for 90 seconds.
After activity in tissue samples had decayed to background levels, the myocardial tissue samples were thawed and digested in 4 mol/L KOH with 2% Tween 80 and processed according to previously published techniques.14 The color dyes were eluted from the microspheres using a measured volume of dimethylformamide and aliquots placed in a multiple wavelength spectrophotometer (model U-2000, Hitachi Ltd). Absorbance was measured at the principal absorbance peak of each pure color dye. Corrections for the absorbance from overlapping spectra were performed using a matrix inversion technique.14 15 With the use of the absorbance and flow rate of the arterial reference sample and myocardial absorbance per unit sample weight, regional myocardial perfusion was calculated as follows14 15 : Qsample=Abssample*Qreference/Absreference, where Qsample is the flow (mL·min−1·g−1) in the tissue sample, Abssample is the absorbance of a given dye eluted from the tissue sample, Qreference is the reference blood sample withdrawal rate (mL/min), and Absreference is the absorbance of a given dye eluted from the blood reference.
A circumferential myocardial sample immediately basal to the sample taken for FDG and flow analysis was immersed in Z-Fix (Anatech Ltd), a buffered fixative containing 3.7% formaldehyde and ionized zinc, for histology (n=9). Thin sections from the anterior and posterior regions were stained with Masson’s trichrome stain to delineate connective tissue and fibroblasts from normal myocytes. Connective tissue staining was quantified with standard point-counting techniques using a 121-point grid at a magnification of ×100 as the grid was progressively moved across the myocardial wall from the subendocardium to the subepicardium.16 Any epicardial fat or connective tissue beyond the visceral pericardium was excluded from analysis. Two full-thickness segments were analyzed in the anterior and posterior regions from each heart. They were selected to represent the area of greatest and least connective tissue staining in each of the two regions. Connective tissue staining was expressed as a percentage of the total across the myocardial wall. Perivascular and endocardial connective tissue staining is included in the reported results. The results represent the average of ≥30 fields from each region of a given heart.
All data are presented as mean±SEM. Measurements of perfusion, FDG, and connective tissue staining in the LAD region were compared with corresponding measurements in the normally perfused control region with the use of paired t tests. Differences in flow and hemodynamics during pharmacological interventions were initially assessed with the use of an ANOVA and posthoc paired t tests with the Bonferroni correction for multiple comparisons. The P<.05 level was taken as significant.
All animals were in good health at the time of the study. Body weight increased from 7.9±0.6 kg at the time of instrumentation to 75±4 kg at the time of study (average growth interval, 107±5 days). Arterial blood gases immediately before resting measurements of myocardial perfusion were pH 7.41±0.01, Pco2 37±1 mm Hg, and Po2 453±27 mm Hg. Hematocrit levels averaged 0.34±0.01.
At the time of study, 11 animals had total LAD occlusion with angiographic collaterals, and 5 had severe proximal LAD stenoses (mean diameter stenosis severity of the group, 97±2%). In 1 animal (not included in the subsequent data), the occluder was not secured, and it migrated off of the LAD. There was no significant stenosis, and it served as a sham. Representative left coronary angiograms and the end-systolic and end-diastolic tracings of the resting left lateral ventriculograms from an experimental animal and the sham pig are shown in Fig 1⇓. Sources of angiographic collaterals to the LAD included bridging vessels across the stenosis and the circumflex artery (mean grade, 2.0±0.3, usually to the mid-LAD) and the right coronary artery (mean grade, 2.2±0.2, usually to the distal LAD). There were no angiographic collaterals visible in the sham animal or in pigs with a patent LAD. Analysis of the left ventriculograms revealed a resting ejection fraction of 49±3%. Regional anterior wall motion was severely depressed and usually exhibited akinesis or severe hypokinesis (wall motion score, 0.7±0.2; normal wall motion, 3). Histological sections taken from the anterior and posterior regions did not show light microscopic evidence of myocardial necrosis. Only occasional small islands of patchy fibrosis were seen. A more consistent feature was a subtle, but generalized, increase in the usual connective tissue staining in the LAD region. The extent of connective tissue staining (including perivascular tissue) assessed by point counting averaged 6.2±0.9% in the hibernating regions and 2.9±0.3% in the control region (P<.01).
The distribution of perfusion at rest in the LAD and normal remote region is shown in Fig 2⇓. Corresponding measurements of hemodynamics and myocardial perfusion are summarized in Tables 1⇓ and 2⇓. Under resting conditions, subendocardial perfusion in the LAD region of hibernating animals was reduced by 24% compared with corresponding control region measurements (0.80±0.06 versus 1.07±0.06 mL·min−1·g−1, P<.001; Fig 2⇓). Resting full-thickness flow was also lower in the LAD region of hibernating animals (0.87±0.04 versus 0.99±0.06 mL·min−1·g−1 in the control region, P<.01).
Fig 3⇓ summarizes the effects of epinephrine and adenosine on flow in hibernating and normal regions. During submaximal inotropic stimulation with epinephrine, heart rate increased from 85 to 137 bpm. There was no significant increase in left ventricular ejection fraction (50±4% at rest to 56±6%, n=10, P=NS), but the anteroapical wall motion score improved (0.6±0.2 at rest to 1.2±0.2, P<.01). Pharmacological vasodilation with adenosine produced no significant change in heart rate or systolic blood pressure, but there was a 10% decrease in mean arterial pressure (115±3 to 104±4 mm Hg, P=.01) compared with resting values. In the normally perfused region, full-thickness flow increased an average of five times over the resting values as well as increasing significantly in each of the individual transmural layers. In contrast, there was an attenuation of the flow increase in the hibernating LAD region. Full-thickness flow during vasodilation was less than one third of that in the corresponding normal region (1.51±0.13 versus 4.90±0.23 mL·min−1·g−1, P<.001). Subendocardial flow reserve was markedly attenuated, and perfusion did not increase above the corresponding resting value (0.80±0.06 mL·min−1·g−1 at rest versus 1.05±0.15 mL·min−1·g−1 during adenosine, P=NS).
Transmural Variations in FDG Deposition
The regional activity of FDG was systematically elevated in the hibernating LAD region compared with the normally perfused posterior regions. Fig 4⇓ shows a representative circumferential profile of FDG activity in subendocardial, midmyocardial, and subepicardial layers. The hibernating LAD region that encompasses the samples in the middle of this unrolled ring demonstrates an increase in FDG deposition that is most marked in the subendocardial samples and decreases toward the outer layers of the heart. Although the magnitude of the increase in FDG varied, a similar transmural pattern of FDG deposition was observed in each of the animals with a chronic LAD stenosis. In full-thickness samples, fasting FDG deposition averaged 24.3±3.8×103 cpm/g in hibernating LAD regions versus 12.5±1.4×103 cpm/g in normally perfused posterior regions (P<.01). There was a more pronounced increase in FDG uptake in the subendocardium, which averaged 36.9±6.8×103 cpm/g in hibernating versus 15.0±2.2×103 cpm/g in normally perfused remote regions (P<.02). There was a uniform circumferential pattern in the sham animal, whereas there was a mismatch between FDG and flow in the LAD region of the hibernating animal.
Fig 5⇓ shows the average data from all 9 animals. Data are expressed as the deposition [extraction (E)*flow (F)]. This is equivalent to myocardial FDG activity in each region divided by the integral of the arterial concentration-time curve. Deposition of FDG was significantly increased in each of the three transmural layers of the hibernating region compared with the corresponding normally perfused region (P<.02) and was greatest in the subendocardium. When expressed as a weighted average of FDG activity across the entire myocardial wall, there was an ≈1.9-fold increase in FDG in hibernating versus normally perfused myocardium. Fig 6⇓ summarizes the relative increase (LAD/normal remote region) in FDG deposition in relation to the relative reduction in flow in each transmural layer for all of the animals. The data were obtained by pooling samples in the hibernating LAD region and dividing them by the corresponding values from the normally perfused region. The deposition of FDG was inversely related to the degree that resting perfusion was reduced and greatest in the subendocardium.
FDG Uptake by PET
FDG uptake was assessed after an overnight fast by dynamic imaging with PET to exclude the possibility that prior pharmacological stimulation (ie, epinephrine, adenosine, or phenylephrine) may have produced transient ischemia and the regional variations in FDG deposition we found through ex vivo counting. A representative reconstructed short-axis PET image is shown in Fig 7⇓. Regional FDG uptake in the anterior wall of the left ventricle, corresponding to the LAD perfusion territory, was markedly increased compared with the normally perfused inferoposterior region. After an overnight fast, plasma glucose averaged 6.8±0.9 mmol/L. Fasting FDG uptake without antecedent pharmacological stimulation was increased in the LAD region of each animal, with an average 1.8±0.2-fold increase in FDG uptake in comparison with normally perfused remote regions (LAD, 0.54±0.19 μCi/mL; normal regions, 0.30±0.09 μCi/mL; P<.05). The relative increase in FDG uptake in hibernating versus normal regions by in vivo imaging was in remarkable agreement with the full-thickness values obtained from ex vivo tissue counting 1 hour after pharmacological stimulation (Fig 8⇓). The FDG rate constant (K*100, 2.5±0.9 versus 1.3±0.4 min−1, P<.05) and the calculated rate of metabolic glucose utilization (rMGU, 21±4 versus 11±3 μmol · min−1 · 100 g−1, P<.05) were also higher in hibernating myocardium.
There are several important new findings from our study. First, pigs instrumented with a fixed proximal LAD occluder for 3 months develop a critical coronary stenosis or total occlusion with collateral-dependent myocardium that has regional dysfunction and is viable as assessed by histology. This model exhibits the key salient physiological features of hibernating myocardium found in humans. These include regional reductions in resting myocardial perfusion and an approximately twofold increase in FDG deposition in the fasting state as assessed by in vivo imaging with PET or ex vivo tissue counting. Second, although reductions in resting flow and increases in FDG deposition occurred in full-thickness myocardial samples, they were most pronounced in the subendocardial layers. This suggests that the propensity of a myocardial region to develop the intrinsic adaptation of hibernation may be critically related to its regional coronary flow reserve.
Clinical and Experimental Studies of Dysfunctional Collateral-Dependent Myocardium
With the advent of clinical techniques to quantify myocardial perfusion, a number of groups have demonstrated that relative and/or absolute reductions in resting flow occur in patients who have a total occlusion and collateral-dependent myocardium with regional dysfunction at rest. The reductions in resting perfusion have been difficult to attribute to tissue necrosis or acute ischemia, and function has been shown to improve in patients who have undergone surgical revascularization.4 8 17 These physiological features have led to the description of this state as hibernating myocardium. Arani et al18 found an absolute reduction in resting flow using inert gas washout that averaged 0.51 mL·min−1·g−1 in collateral-dependent regions. This was not associated with metabolic evidence of ischemia as reflected by selective coronary venous sampling and was lower than the 95% confidence interval for absolute flow in a normal population at a similar level of double product. In more recent studies, flow in patients with dysfunctional collateral-dependent myocardium has been examined with the use of PET and resting flow has been reported to be significantly lower than remote, normally perfused regions in the same patient. The study by Vanoverschelde et al17 reported resting flow to average 0.77 mL·min−1·g−1 in dysfunctional, collateral-dependent regions versus 0.95 mL·min−1·g−1 in remote regions. Mäki et al8 found resting flow to average 0.81 mL·min−1·g−1 in dysfunctional regions versus 1.02 mL·min−1·g−1 in remote regions. In our porcine model, differences in full-thickness microsphere flow measurements were very similar to the results in patients with chronic coronary occlusion and averaged 0.87 mL·min−1·g−1 in dysfunctional regions versus 0.99 mL·min−1·g−1 in remote, normally perfused regions. By using microspheres, we were also able to identify significant transmural variations in the distribution of resting flow that are beyond currently available clinical techniques to image myocardial perfusion. In dysfunctional regions of pigs, reductions in resting flow were most pronounced in the subendocardial layers, averaging 0.80 mL·min−1·g−1 versus 1.07 mL·min−1·g−1 in normal remote subendocardial regions. Thus, although speculative, it seems likely that important transmural variations in flow may also occur in humans with hibernating myocardium.
There has been limited success in reproducing the clinical findings of hibernating myocardium in previous chronic animal models of collateral-dependent myocardium. In dogs, this appears to relate to the rapid maturation of the collateral circulation, which usually produces no change in resting flow or function19 20 and only a modest impairment in coronary flow reserve. Such a limitation is frequently manifested only during heavy exercise.21 In a previous study, we attempted to circumvent this problem by surgically ligating potential sources of epicardial collaterals in a chronic dog ameroid model that resulted in relative reductions in flow and function for several weeks that were consistent with hibernation.22 Other researchers have used pigs to circumvent the rapid and extensive collateralization that occurs in dogs. Unfortunately, ameroid occlusion in mature pigs usually results in some degree of tissue necrosis, making it difficult to interpret relative changes in flow and/or function without correction for the amount of infarcted tissue in the collateral-dependent region.23 24 25 26 Only one study has been able to demonstrate a reduction in regional myocardial function during ameroid occlusion in the absence of potentially confounding tissue necrosis.27 In this study, serial measurements were used to demonstrate that persistent reductions in wall thickening could be produced for a period of ≈3 weeks. Unlike clinical studies of hibernating myocardium, the dysfunction in this model was not associated with significant reductions in resting flow. This contrasts with the present findings in which resting flow in pigs was systematically reduced. The differences may be due to the longer time interval over which collateral-dependent myocardium is present in our model, which would also be consistent with the results of our previous study, in which there was a transition from normal to reduced flow in dysfunctional collateral-dependent regions in dogs.
FDG Uptake in Dysfunctional Collateral-Dependent Myocardium
Tillisch et al4 demonstrated that a mismatch pattern with reduced flow and increased FDG uptake was predictive of functional recovery after surgical revascularization. Interestingly, determining the extent to which the mismatch is related to reduced flow or increased FDG uptake has been complicated by the considerable variability in the specific protocols that have been used to assess FDG uptake. Enhancement of myocardial glucose (and FDG) uptake through stimulation of the externalization of insulin-responsive myocardial glucose transporters optimizes myocardial image characteristics to identify viability. However, there may be pathophysiological alterations unique to hibernating myocardium that can be more optimally identified by examining FDG uptake in the unstimulated, fasting state. In this regard, Mäki et al8 found that the rMGU in the fasting state was 15 μmol·100 g−1·min−1 in dysfunctional, collateral-dependent regions versus 11 μmol·100 g−1·min−1 in normal regions. Likewise, a preliminary study by Gerber et al28 found values of 14 μmol·100 g−1·min−1 in dysfunctional collateral-dependent regions versus 7 μmol·100 g−1·min−1 in normal remote regions. In both studies, stimulation of FDG uptake with glucose and insulin increased rMGU but caused relative FDG uptake to be lower in hibernating versus normally perfused remote regions. Our findings in pigs in which FDG was also administered in the fasting state are consistent with the observations in fasting humans. In full-thickness dysfunctional regions, FDG uptake was increased by 1.9-fold through ex vivo counting and 1.8-fold through in vivo PET imaging. Kinetic analysis of our PET data found the rMGU to average 21 μmol·100 g−1·min−1 in dysfunctional regions versus 11 μmol·100 g−1·min−1 in remote regions.
We were also able to define the transmural variation in the distribution of FDG deposition through ex vivo tissue counting. In normal regions, the endocardium-to-epicardium ratio for FDG deposition was 1.35 and similar to the transmural gradient in resting perfusion, which was 1.20. This basal distribution favoring the subendocardium is similar to results in rat hearts29 as well as in anesthetized open-chest dogs.30 In contrast, there was a more pronounced variation across the myocardial wall of dysfunctional regions in the present study, and the endocardium-to-epicardium ratio for FDG increased to 2.4. Although there are no previous studies regarding transmural variations in FDG uptake in chronically dysfunctional myocardium, the pattern is similar to that recently reported by Zhang et al31 using in vivo NMR spectroscopy in dogs with severe left ventricular hypertrophy. They found that 2-deoxyglucose accumulation averaged 7.9 μmol/g in the subendocardium versus 3.3 μmol/g in the subepicardium (endocardium-to-epicardium ratio, 2.4). Whether these changes reflect intrinsic protective adaptations that are common to hibernating myocardium and hypertrophied hearts or are a reflection of chronic repetitive myocardial ischemia will require further study.
Relation Among Coronary Flow Reserve and Chronic Hibernation
We and others have speculated that there may be a key link between the severity to which coronary flow reserve is reduced and the induction of adaptations consistent with chronic hibernation.3 17 22 In humans with collateral-dependent myocardium, there was an inverse relation between the extent of myocardial dysfunction and regional coronary flow reserve.17 Patients with regional dysfunction and relative reductions in flow consistent with hibernating myocardium had values of coronary flow reserve that averaged 1.4. In contrast, in a subgroup of patients with well-developed collaterals who exhibited near-normal function, the values of resting flow and coronary flow reserve were the same in collateral-dependent and remote regions. Because we found significant variations in flow and FDG deposition across the myocardial wall of the hibernating regions in our model, we evaluated whether heterogeneity in local FDG uptake varied with respect to coronary flow reserve. Fig 9⇓ shows the relation between coronary flow reserve as assessed through microsphere flow measurements in individual myocardial examples and regional FDG as assessed through ex vivo tissue counting. There was an inverse relation between the magnitude of FDG deposition in the fasting state and the reduction in coronary flow reserve. Although speculative, these data suggest that the propensity of a region to exhibit physiological adaptations consistent with hibernation may be critically dependent on its local flow reserve.
Although the findings in our animal model are similar to findings in patients with hibernating myocardium, a limitation that must be acknowledged is that collateral growth is stimulated by a fixed stenosis placed on the LAD of growing pigs, and this may not be identical to an acquired stenosis in adult animals. Nevertheless, the gradual increase in stenosis severity in this model as opposed to ameroid models, in which the ameroid stenosis reaches a total occlusion over only 2 to 3 weeks,22 may be a key reason why we were successful in producing hibernating myocardium in the absence of infarction.
Although we did not demonstrate that function could improve after surgical revascularization, all of our physiological findings are consistent with observations in humans that are predictive of functional recovery and submaximal inotropic stimulation improved anterior wall motion. We found a small increase in connective tissue staining by point counting, but the quantitative changes were actually somewhat less than those found in patients with clinically defined hibernating myocardium in whom function improved after revascularization.17 32 33
Finally, acute ischemia34 and pharmacological stimuli35 have been shown to rapidly alter myocardial glucose transport. Although glucose transport has been reported to be unchanged,36 decreased,36 37 or increased10 38 after ischemia, the similarity of FDG uptake obtained through in vivo imaging with PET (without antecedent pharmacological stimuli) and ex vivo counting suggests that a 1-hour time interval is adequate for glucose transport to return to baseline in our model.
Both the preservation of FDG uptake and a flow-metabolism mismatch pattern are predictive of reversible dyssynergy in chronically dysfunctional myocardium. Our results suggest that alterations in regional FDG uptake in the fasting state may be useful in identifying the pathophysiological basis for regional dysfunction. This may be of particular clinical importance when using imaging modalities that cannot resolve transmural variations in tracer uptake and trying to distinguish nontransmural scar from reversibly dysfunctional myocardium.39 The results of our study also suggest that the physiological adaptations associated with hibernating myocardium, including alterations in resting flow, flow reserve, and FDG uptake, vary in magnitude across the myocardial wall in a way that may not be detectable with currently available physiological imaging modalities. We have preliminary evidence that there are also transmural variations in candidate genes that may be responsible for some of the physiological responses at the molecular level.40 Whether hibernating myocardium is an entity that can be distinguished from chronic stunning will require further comparative studies in which physiological and molecular responses can be examined in chronic animal models.
This work was supported by an Affiliate Grant-in-Aid from the American Heart Association; National Heart, Lung, and Blood Institute grant HL-55324; a Merit Review Award from the Department of Veterans Affairs; and the Albert and Elizabeth Rekate Fund. Dr Fallavollita is the recipient of a Clinician Scientist Award from the American Heart Association. We would like to thank all of the staff at the Center for Positron Emission Tomography and, in particular, Brian Murphy and David Wack for help with the kinetic analysis. Deana Gretka, Amy Johnson, Susan Fopeano, Felicia Bosinski, and Christopher Trojan provided technical assistance throughout this study. We would also like to thank Nina Goss and Karin Brady for their help in preparing the manuscript.
- Received August 27, 1996.
- Revision received November 4, 1996.
- Accepted November 19, 1996.
- Copyright © 1997 by American Heart Association
Ross J Jr. Myocardial perfusion-contraction matching: implications for coronary heart disease and hibernation. Circulation. 1991;83:1076-1083.
Bolli R. Myocardial ‘stunning’ in man. Circulation. 1992;86:1671-1691.
vom Dahl J, Eitzman DT, Al-Aouar AR, Kanter HL, Hicks RJ, Deeb GM, Kirsh MM, Schwaiger M. Relation of regional function, perfusion, and metabolism in patients with advanced coronary artery disease undergoing surgical revascularization. Circulation. 1994;90:2356-2366.
Mills I, Fallon JT, Wrenn D, Sasken H, Gray W, Bier J, Levine D, Berman S, Gilson M, Gewirtz H. Adaptive responses of coronary circulation and myocardium to chronic reduction in perfusion pressure and flow. Am J Physiol. 1994;266:H447-H457.
Mäki M, Luotolahti M, Nuutila P, Iida H, Voipio-Pulkki L-M, Ruotsalainen U, Haaparanta M, Solin O, Hartiala J, Harkonen R, Knuuti J. Glucose uptake in the chronically dysfunctional but viable myocardium. Circulation. 1996;93:1658-1666.
Folts JD, Crowell EB Jr, Rowe GG. Platelet aggregation in partially obstructed vessels and its elimination with aspirin. Circulation. 1976;54:365-370.
Ratib O, Phelps ME, Huang S-C, Henze E, Selin CE, Schelbert HR. Positron tomography with deoxyglucose for estimating local myocardial glucose metabolism. J Nucl Med. 1982;23:577-586.
Kowallik P, Schulz R, Guth BD, Schade A, Paffhausen W, Gross R, Heusch G. Measurement of regional myocardial blood flow with multiple colored microspheres. Circ Res. 1991;83:974-982.
Elias H, Henning A, Schwartz DE. Stereology: applications to biomedical research. Physiol Rev. 1971;51:158-200.
Vanoverschelde J-L, Wijns W, Depré C, Essamri B, Heyndrickx GR, Borgers M, Bol A, Melin JA. Mechanisms of chronic regional postischemic dysfunction in humans: new insights from the study of noninfarcted collateral-dependent myocardium. Circulation. 1993;87:1513-1523.
Hill RC, Kleinman LH, Tiller WH Jr, Chitwood WR Jr, Rembert JC, Greenfield JC Jr, Wechsler AS. Myocardial blood flow and function during gradual coronary occlusion in awake dog. Am J Physiol. 1983;244:H60-H67.
Yamamoto H, Tomoike H, Shimokawa H, Nabeyama S, Nakamura M. Development of collateral function with repetitive coronary occlusion in a canine model reduces myocardial reactive hyperemia in the absence of significant coronary stenosis. Circ Res. 1984;55:623-632.
Kumada T, Gallagher KP, Battler A, White F, Kemper WS, Ross J Jr. Comparison of postpacing and exercise-induced myocardial dysfunction during collateral development in conscious dogs. Circulation. 1982;6:1178-1185.
Canty JM Jr, Klocke FJ. Reductions in regional myocardial function at rest in conscious dogs with chronically reduced regional coronary artery pressure. Circ Res. 1987;61(suppl II):II-107-II-116.
Roth DM, Maruoka Y, Rogers J, White FC, Longhurst JC, Bloor CM. Development of coronary collateral circulation in left circumflex ameroid-occluded swine myocardium. Am J Physiol. 1987;253:H1279-H1288.
Roth DM, White FC, Nichols ML, Dobbs SL, Longhurst JC, Bloor CM. Effect of long-term exercise on regional myocardial function and coronary collateral development after gradual coronary artery occlusion in pigs. Circulation. 1990;82:1778-1789.
Symons JD, Pitsillides KF, Longhurst JC. Chronic reduction of myocardial ischemia does not attenuate coronary collateral development in miniswine. Circulation. 1992;86:660-671.
Shen Y-T, Vatner SF. Mechanism of impaired myocardial function during progressive coronary stenosis in conscious pigs: hibernation versus stunning? Circ Res. 1995;76:479-488.
Gerber B, Melin JA, Bol A, Vanoverschelde J-L. Attenuated response of myocardial glucose utilization to insulin stimulation in hibernating myocardium. Circulation. 1995;92:I-313. Abstract.
Takala TES, Hassinen IE. Effect of mechanical work load on the transmural distribution of glucose uptake in the isolated perfused rat heart, studied by regional deoxyglucose trapping. Circ Res. 1981;49:62-69.
Zhang J, Duncker DJ, Ya X, Zhang Y, Pavek T, Wei H, Merkle H, Ugurbil K, From AHL, Bache RJ. Effect of left ventricular hypertrophy secondary to chronic pressure overload on transmural myocardial 2-deoxyglucose uptake: a 31P NMR spectroscopic study. Circulation. 1995;92:1274-1283.
Maes A, Flameng W, Muyts J, Borgers M, Shivalkar B, Ausma J, Bormans G, Schiepers C, DeRoo M, Mortelmans L. Histological alterations in chronically hypoperfused myocardium: correlation with PET findings. Circulation. 1994;90:735-745.
Shivalkar B, Maes A, Borgers M, Ausma J, Scheys I, Nuyts J, Mortelmans L, Flameng W. Only hibernating myocardium invariably shows early recovery after coronary revascularization. Circulation. 1996;94:308-315.
Kalff V, Schwaiger M, Nguyen N, McClanahan TB, Gallagher KP. The relationship between myocardial blood flow and glucose uptake in ischemic canine myocardium determined with fluorine-18-deoxyglucose. J Nucl Med. 1992;33:1346-1353.
Hariharan R, Bray M, Ganim R, Doenst T, Goodwin GW, Taegtmeyer H. Fundamental limitations of [18F]2-deoxy-2-fluoro-d-glucose for assessing myocardial glucose uptake. Circulation. 1995;91:2435-2444.
Schneider CA, Nguyêñ VTB, Taegtmeyer H. Feeding and fasting determine postischemic glucose utilization in isolated working rat hearts. Am J Physiol. 1991;260:H542-H548.
McFalls EO, Ward H, Fashingbauer P, Gimmestad G, Palmer B. Myocardial blood flow and FDG retention in acutely stunned porcine myocardium. J Nucl Med. 1995;36:637-643.
Camici P, Araujo LI, Spinks T, Lammertsma AA, Kaski JC, Shea MJ, Selwyn AP, Jones T, Maseri A. Increased uptake of 18F-fluorodeoxyglucose in postischemic myocardium of patients with exercise-induced angina. Circulation. 1986;74:81-88.
Canty JM Jr, Fallavollita JA, Perry BJ, Young RS. Transmural induction of mRNA for HSP-70 in hibernating myocardium. Circulation. 1995;92(suppl I):I-654. Abstract.