Adenosine-Recruitable Flow Reserve Is Absent During Myocardial Ischemia in Unanesthetized Dogs Studied in the Basal State
Abstract We conducted the present study to determine if pharmacologically recruitable flow reserve was present during ischemia in unanesthetized dogs studied in the absence of adrenergic activation. Microsphere measurements of regional perfusion at reduced coronary pressures in a distal circumflex region subjected to intracoronary adenosine infusion (0.265 mg/min IC) were compared with measurements in a proximal circumflex region where pressure was reduced but flow was being autoregulated normally. Ischemia began when coronary pressure was reduced below 40 mm Hg. At a coronary pressure of 28±1 (mean±SEM) mm Hg, subendocardial flow with autoregulation intact was 0.59±0.05 mL · min · g−1 and similar to that in the region receiving adenosine, which averaged 0.61±0.05 mL · min−1 · g−1 (P=NS). Although adenosine increased subepicardial flow from 0.81±0.05 to 1.16±0.09 mL · min−1 · g−1 (P<.001), the magnitude of the increase was minimal when coronary pressure fell to a level that caused subepicardial flow during autoregulation to be reduced below resting values. These results demonstrate that the substantially lower autoregulatory break point found in unanesthetized dogs studied in the basal state reflects the ability of intrinsic autoregulatory mechanisms to match local vasodilator reserve to that recruitable pharmacologically. This contrasts with the presence of pharmacologically recruitable subendocardial flow reserve in anesthetized animals and is most likely related to a low level of circulating catecholamines and lack of sympathetic activation in unanesthetized animals studied under basal conditions.
A number of studies in anesthetized animals have demonstrated that intrinsic autoregulatory mechanisms are unable to match local coronary flow to local vasodilator reserve recruited pharmacologically in the setting of ischemia.1 2 3 4 Similar reductions in myocardial perfusion with pharmacologically recruitable flow reserve have also been demonstrated to occur during ischemia following activation of adrenergic tone in unanesthetized animals.5 6 7 8 Collectively, the results of these studies support the notion that there is a continuous competition between adrenergic constrictor tone and metabolic vasodilation that can accentuate the severity of ischemia at any given level of reduced coronary pressure.
The potential relevance of these findings to coronary flow regulation in the basal unanesthetized state has not been established. During autoregulation in unanesthetized animals, subendocardial flow remains constant until coronary pressure falls below 40 mm Hg9 compared with lower autoregulatory pressure limits of 60 mm Hg and higher in open-chest anesthetized animals studied under comparable systemic hemodynamic conditions.10 11 This large difference cannot be explained by hemodynamic factors and may reflect the ability of the unanesthetized animal to match local flow to local vasodilator reserve more precisely. It may in part relate to the fact that circulating catecholamine levels are not elevated in conscious dogs at rest12 or during moderate ischemia13 and to the lack of reflex sympathetic nerve activation until severe levels of myocardial ischemia are produced.14 Although indirect, the shift in the autoregulatory pressure limit to lower pressures in unanesthetized animals suggests that pharmacologically recruitable flow reserve may not be present during subendocardial ischemia under basal conditions.
The present study tested the hypothesis that reductions in subendocardial perfusion are not associated with pharmacologically recruitable coronary flow reserve in unanesthetized animals when adrenergic activation is absent. Experiments were conducted in chronically instrumented dogs studied at rest, with radioactive microspheres used to assess regional perfusion. We used a paired experimental design to simultaneously compare regional perfusion in myocardial samples from a proximal left circumflex (LC) region that was autoregulating flow normally to measurements in a distal LC region that was subjected to adenosine vasodilation delivered through an intracoronary catheter. Since both regions were subjected to the same systemic and coronary hemodynamics, this experimental approach circumvented the confounding effects of changes in flow due to small variations in hemodynamics when sequential microsphere flow measurements to the same region are compared. The results contrast with previous studies examining autoregulation in open-chest anesthetized animals and demonstrate that adenosine-recruitable flow reserve is absent during subendocardial ischemia.
Materials and Methods
The studies were conducted in 15 chronically instrumented mongrel dogs of either sex weighing 28±1.3 (mean±SEM) kg. All procedures were performed in accordance with institutional guidelines for the care and use of animals in research.
The details of the experimental preparation have been previously published.9 Briefly, animals were fasted overnight and premedicated with Innovar-Vet (0.4 mg/mL fentanyl and 20 mg/mL droperidol, 1 to 3 mL IM). A surgical plane of anesthesia was induced with sodium thiamylal (20 mg/kg IV). After endotracheal intubation, the dogs were mechanically ventilated, and anesthesia was maintained with a halothane (1% to 2%)/nitrous oxide (≈60%)/oxygen (balance) mixture. Under sterile conditions, a thoracotomy was performed in the fourth or fifth left intercostal space.
The experimental preparation is shown in Fig 1⇓. After suspending the heart in a pericardial cradle, Tygon catheters were placed into the left atrium and descending thoracic aorta for microsphere injection and reference blood flow sampling, respectively. A micromanometer (model P6.5, Konigsberg Instruments Inc) was inserted into the left ventricular apex through a stab wound. A short segment of the proximal LC artery before the first major marginal branch was dissected free and instrumented with a hydraulic occluder. A small polytetrafluoroethylene (Teflon) catheter was placed into the midportion of the LC artery so that its tip was beyond the first major marginal branch. The coronary catheter was used for pressure measurement as well as to infuse adenosine (53.5 mg/100 mL saline at 0.5 mL/min IC) into the distal segment of the LC artery. By selectively infusing adenosine into the distal half of the LC region, it was possible to simultaneously assess perfusion in tissue samples from a proximal LC region that was autoregulating flow normally and a distal LC region subjected to vasodilation with intracoronary adenosine by using radioactive microspheres (see Fig 1⇓). This approach also ensured that both regions of the heart were subjected to identical hemodynamic conditions during the microsphere flow measurements and circumvented potential difficulties in assessing small flow differences with sequential microsphere flow measurements. The coronary catheter was constructed by modifying a 22-gauge angiocatheter (model 38-2862-1, Deseret Inc). Briefly, a 1-in length of polyethylene tubing (PE-60, Clay Adams) was placed over the original angiocatheter needle. The Teflon angiocatheter sheath was removed and cut to a length of 1 in from its tip, and the remaining sheath and luer connector were discarded. The Teflon tip was gently advanced over the needle and inserted into the polyethylene tubing until the bevel of the needle was visible. A small 0.25-in silastic cuff (inner diameter, 0.03 in; outer diameter, 0.065 in; model 602-175, Dow Corning) was placed over both the Teflon and polyethylene tubing. This was used to bind the two tubes together and attach sutures and needles that were used to anchor the catheter after it was placed in the vessel. The original needle and modified Teflon sheath were inserted through the vessel wall at an angle of 45°. The catheter was carefully advanced over the needle and into the vessel lumen. After adequate blood return was confirmed, the needle was withdrawn, and the catheter was connected to an 18-in length of Tygon tubing (inner diameter, 0.020 in; outer diameter, 0.060 in; Norton Plastics) by using a 1-in length of 21-gauge stainless steel hypodermic tubing. A similar Teflon catheter was placed into the ascending aorta for pressure measurement (not shown in Fig 1⇓).
After placing all of the instruments, the chest was closed and the pneumothorax was evacuated. The catheters and wires were exteriorized and placed into a jacket that the animal was trained to wear. Parenteral analgesics (meperidine, 50 mg IM as needed) were given during the early postoperative period as required. Each animal was given prophylactic antibiotics for 3 to 5 days after surgery (procaine penicillin G, 300 000 U; streptomycin, 300 mg IM QD). All catheters were flushed and filled with heparinized saline (1000 U/mL) at 2- to 3-day intervals. They were allowed to recover for at least 10 days before the studies were conducted with the animals in the unanesthetized state (19±2 days after instrumentation).
On the day of the study, the dogs were lightly sedated with Innovar-Vet (1 to 3 mL) and studied while they were lying in the right lateral decubitus position. The experimental protocol was begun after allowing at least 30 minutes for the animals to adjust to the laboratory. Regional myocardial perfusion was assessed with radioactive microspheres (15 μm in diameter); the reference withdrawal technique was used as previously described.9 Briefly, microspheres were suspended in 10% dextran and 0.01% Tween 80. They were ultrasonicated for at least 15 minutes and vortex-agitated immediately before they were withdrawn. Approximately 3×106 microspheres were injected into the left atrium over 15 to 20 seconds after beginning an arterial reference withdrawal sample (4 mL/min) that was continued for 2 minutes. A combination of the following radionuclides was used: 54Co, 141Ce, 114In, 113Sn, 104Ru, 85Sr, 75Nb, 46Sc, 51Cr, and 54Mn. Control microsphere flow measurements were obtained at the spontaneous coronary pressure under resting conditions. After this, adenosine was infused into the distal coronary artery at a rate of 0.5 mL/min, and coronary pressure was gradually reduced to a level below 40 mm Hg. We have previously shown that this degree of pressure reduction is associated with the onset of subendocardial ischemia under resting conditions in unanesthetized dogs.9 After allowing at least 5 minutes for coronary pressure and systemic hemodynamics to reach a new steady state, a microsphere flow measurement was repeated. Additional flow measurements were performed at up to three levels of steady state coronary pressure reduction in each of the animals. In 10 of the animals, microsphere flow measurements were obtained in the awake state during intracoronary adenosine infusion at normal coronary pressures in the absence of a stenosis and during either a total LC occlusion or severe ischemia in the absence of adenosine infusion. These measurements were used to demarcate the areas subjected to adenosine infusion and the LC risk area distal to the occluder that was subjected to reduced coronary pressure. In the remaining five animals, measurements during adenosine infusion at normal pressures and during coronary occlusion in the absence of adenosine were obtained during anesthesia with pentobarbital (30 mg/kg IV) as part of a separate study protocol. Although the latter measurements were not included in the present study, they were used to demarcate the perfusion territory as outlined below. At the end of the experiments, the animals were anesthetized with sodium pentobarbital and euthanatized with an injection of potassium chloride. The hearts were removed and placed in 10% formalin for fixation.
Microsphere Flow Analysis and Heart Sectioning
Several days later, the left ventricle was sectioned into four concentric rings, and the apex was discarded. After removing epicardial fat and connective tissue, we discarded the septum and sectioned the free wall of the left ventricle from each of the concentric rings into 10 to 20 small wedges that weighed ≈1 to 2 g. Each of the wedges was subsequently sectioned into four transmural layers of approximately equal thickness. The myocardial samples were weighed and counted on a germanium well detector (Canberra, Inc). The principal photon peaks of gamma-emitting radionuclides analyzed with this instrument have a full-width half-maximum of ≈2 to 4 keV as opposed to the 20- to 40-keV photopeak widths obtained on conventional sodium iodide detectors.15 As a result, up to 12 radionuclides with nonoverlapping photopeaks can be analyzed in the same sample by simply subtracting the local background energy from the area under the principle photopeaks by using standard spectral analysis software (Spectran). Flow measurements in each tissue sample were calculated from the integral counts under each photopeak in relation to the reference blood samples and expressed as flow per gram of tissue, as previously described.9
Flow distributions under control conditions, during coronary occlusion in the absence of adenosine, and during adenosine administration at normal coronary pressures were used to define the myocardial tissue samples in the proximal LC region subjected to reduced LC pressure during autoregulation and the samples in the distal region subjected to adenosine vasodilation. The LC area that was subjected to coronary pressure reduction distal to the occluder (or risk area) was determined during coronary occlusion in the absence of adenosine. By examining the circumferential flow pattern, border samples between the LC free wall and normally perfused left anterior descending (LAD) or septal regions could be identified from their intermediate flows due to overlapping perfusion from normal and ischemic regions and thus eliminated from further analysis. Once the LC samples subjected to coronary pressure reduction were identified, we compared the flow measurements under resting conditions with those obtained during an intracoronary adenosine infusion at the normal coronary pressure. The proximal autoregulating LC region was identified as wedges that had flows during infusion of adenosine into the distal coronary artery that were within 50% of their corresponding values at rest. Distal regions subjected to adenosine were identified as those samples having flows that were at least 300% higher than the corresponding resting flow value. Samples between the proximal and distal regions that had intermediate flow levels between these two values were not used for analysis since they represented a “border region” having an admixture of maximally vasodilated and autoregulating LC tissue. Once the proximal autoregulating and distal adenosine samples were identified, we obtained weighted averages of the flow measurements by using the same aggregate samples in each region for each experimental condition. These aggregate regions were felt to reflect core samples that were free of any overlapping perfusion. Two additional animals were excluded from analysis because the flow distribution demonstrated that the distal coronary infusion catheter had entered a proximal marginal branch as opposed to the main LC artery. A third animal was excluded from analysis because of anemia and a hematocrit of <30%.
Effect of Innovar-Vet on the Phenylephrine Dose-Response Curve
Four additional animals were instrumented to assess the magnitude of adrenergic blockade produced by Innovar-Vet. Mean arterial pressure, heart rate, and LC coronary flow (Transonics flowmeter) were assessed with the animals in the conscious state and repeated 30 and 90 minutes (n=3) after sedation with an injection of Innovar-Vet (2 mL IM). The extent of α-blockade was assessed by examining the pressor response following steady state infusions of phenylephrine at doses of 1, 3, 10, and 30 μg · kg−1 · min−1 IV. The phenylephrine infusion was stopped once mean arterial pressure increased by 60 mm Hg above baseline. Heart rate was kept constant by pacing at a rate approximately equal to that obtained in the resting unsedated state.
All experimental data were digitized for 15 seconds at a sampling rate of 200 Hz with a DT-280lA analog-to-digital convertor (Data Translation) interfaced to an IBM PC AT computer. Measurements of regional perfusion during ischemia were pooled and also categorized on the basis of the degree to which subendocardial perfusion in the proximal autoregulating region was reduced compared with simultaneous measurements in the normally perfused LAD region. All hemodynamic parameters and microsphere flow measurements were subjected to a one-way ANOVA. Significant differences between each category and the corresponding control values were assessed by paired t tests and the Bonferroni correction for multiple comparisons when appropriate. Differences were determined to be significant at P≤.05.
Summary of Baseline Hemodynamics and Regional Perfusion at Rest
Under resting conditions, heart rate averaged 91±4 beats per minute (mean±SEM, n=15), left ventricular systolic pressure averaged 120±2 mm Hg, and left ventricular end-diastolic pressure averaged 9.5±0.9 mm Hg. Peak positive left ventricular dP/dt was 3800±103 mm Hg · s−1, and peak negative left ventricular dP/dt was −2677±143 mm Hg · s−1. Full-thickness myocardial perfusion at rest averaged 1.10±0.06 mL · min−1 · g−1 in the distal LC region, 1.09±0.05 mL · min−1 · g−1 in the proximal LC region, and 1.10±0.06 mL · min−1 · g−1 in the LAD region (P=NS). Thus, regional perfusion under resting conditions was not systematically different in the three core regions of the left ventricular free wall.
Effect of Coronary Occlusion and Intracoronary Adenosine on Regional Perfusion at Normal Coronary Pressures
Fig 2⇓ shows the effect of a coronary occlusion and intracoronary adenosine infused at a normal coronary pressure on the transmural distribution of perfusion. Measurements in the distal LC area subjected to the intracoronary drug infusion are compared with the proximal LC and LAD regions, which were autoregulating blood flow normally. Intracoronary adenosine increased full-thickness flow in the distal LC region fivefold from 1.18±0.08 to 6.40±0.47 mL · min−1 · g−1 (P<.001). In contrast to the marked increase in flow to the distal LC region, flow in both the proximal LC region (1.20±0.08 versus 1.34±0.07 mL · min−1 · g−1) and the normally perfused LAD free wall (1.18±0.08 versus 1.30±0.10 mL · min−1 · g−1) did not significantly change during intracoronary adenosine. Fig 2⇓ also shows the effects of a total coronary occlusion on flow in the LC and LAD regions. During a total coronary occlusion, flow fell to a similar extent in the proximal LC (0.26±0.05 mL · min−1 · g−1) and distal LC (0.29±0.05 mL · min−1 · g−1) regions (P=NS). During a total occlusion, LAD flow tended to increase over that under resting conditions (from 1.15±0.08 to 1.25±0.06 mL · min−1 · g−1), but the difference did not reach statistical significance. These results demonstrate that myocardial perfusion at rest and during a total occlusion of the LC were similar in the proximal and distal LC regions. Intracoronary adenosine produced selective vasodilation in the LC region distal to the coronary catheter but did not significantly affect flow in the proximal LC region that was autoregulating flow normally.
Effect of Intracoronary Adenosine on Regional Perfusion During Ischemia
The effects of adenosine vasodilation on transmural perfusion are summarized for all ischemic flow measurements (n=31) in Fig 3⇓. At an average distal coronary pressure of 28±1 mm Hg, there was no effect of adenosine on subendocardial perfusion (0.59±0.05 mL · min−1 · g−1 in the proximal autoregulating region and 0.61±0.05 mL · min−1 · g−1 in the distal region receiving adenosine, P=NS). Nevertheless, flow to the outer myocardial layers showed small but significant increases. These were most pronounced in the subepicardial layers, where flow increased from 0.81±0.05 mL · min−1 · g−1 in the proximal autoregulating region to 1.16±0.09 mL · min−1 · g−1 in the distal subepicardial region receiving adenosine (P<.001).
We examined the possibility that the increase in flow to outer myocardial layers may have reflected the fact that coronary pressure was still above the lower autoregulatory pressure limit, which for the subepicardium is 25 mm Hg in unanesthetized dogs.9 To accomplish this, the effects of adenosine were examined as a function of the level of ischemia. Hemodynamics and flows categorized according to the magnitude of subendocardial flow reduction (expressed as a percentage of control flow) are summarized in Tables 1⇓ and 2⇓ and Fig 4⇓. There was no significant effect of adenosine on subendocardial perfusion at any level of ischemia (Table 2⇓ and Fig 4⇓). In contrast, subepicardial flow reserve was present at all levels of ischemia. Mild and moderate levels of subendocardial ischemia were not associated with significant reductions in subepicardial flow in the proximal autoregulating region compared with corresponding measurements under resting conditions (Fig 4⇓), demonstrating the absence of subepicardial ischemia despite significant reductions in coronary pressure. Subepicardial flow became significantly reduced during severe ischemia (25±2.5 mm Hg) but, at this time, increased minimally during adenosine administration (0.65±0.07 mL · min−1 · g−1 in the autoregulating region and 0.76±0.06 mL · min−1 · g−1 in the distal adenosine region, P<.05). Thus, the presence of adenosine-recruitable flow reserve in the outer layers in the face of mild and moderate subendocardial ischemia reflected the fact that despite significant reductions in coronary pressure, the subepicardium was not ischemic. When subepicardial flow fell below the resting value, the magnitude of adenosine-recruitable reserve was small.
Table 3⇓ summarizes the number of microspheres in each myocardial region during ischemia. On average, all regions had more than 400 microspheres, which would result in a relative error16 in flow measurement of <5%. For the most severely ischemic subendocardial sample with the lowest microsphere flow number (127 microspheres), the relative error would have been 8.9%. The total weight of samples from the normally perfused LAD free wall (that was selected by visual inspection of the heart) averaged 7.8±0.7 g. The sizes of the proximal LC autoregulating region and distal LC adenosine region were roughly similar and averaged 4.8±0.7 and 7.2±0.7 g, respectively.
Effect of Innovar-Vet on the Phenylephrine Dose-Response Curve
Four additional dogs were studied to quantify the α-adrenergic blocking effects of Innovar-Vet. Hemodynamic responses at 30 and 90 minutes after an intramuscular injection were similar and therefore averaged to compare with measurements in the conscious state. Resting hemodynamic measurements with animals in the conscious state were not significantly different from those after sedation with intramuscular Innovar-Vet (mean aortic pressure, 99±8 versus 94±6 mm Hg; heart rate, 109±12 versus 105±5 beats per minute; and mean LC flow, 26±6 versus 26±7 mL/min [n=3]). The extent of α-blockade was assessed by examining the change in aortic pressure after graded infusions of phenylephrine and is shown in Fig 5⇓. Innovar-Vet produced a slight attenuation of the pressor response to pharmacological doses of the α-agonist that was significant at infusion rates of 3 and 10 μg · kg−1 · min−1 (P<.01). This weak α-blocking effect produced an approximately half log shift in the phenylephrine dose-response curve.
In contrast to the results of previous studies demonstrating pharmacologically recruitable subendocardial flow reserve during ischemia in anesthetized animals, subendocardial flow in unanesthetized animals studied in the basal state is precisely autoregulated to match local vasodilator reserve, and adenosine has no effect on subendocardial flow during ischemia. Although flow to outer layers of the left ventricular free wall increased during adenosine administration, the magnitude of this effect was small and largely abolished during severe ischemia, when coronary pressure fell below the subepicardial autoregulatory pressure limit. Thus, in the absence of sympathetic activation, regional perfusion during ischemia is closely matched to local vasodilator reserve by intrinsic autoregulatory mechanisms. Before discussing our findings, it is important to discuss the potential methodological limitations of the present study.
We were able to examine LC flow during autoregulation and adenosine under identical hemodynamic conditions by comparing microsphere perfusion in two separate regions subjected to an identical level of coronary pressure reduction. A similar paired approach was previously used by Chilian and Ackell13 in a study examining the effects of regional denervation on transmural perfusion during exercise distal to a coronary stenosis. Like their study, our interpretation also assumes that perfusion in each of the two LC regions was similar. In support of this, we found that there was no difference in perfusion between the proximal and distal LC regions under resting conditions. Similarly, during a total LC occlusion in the absence of adenosine, flow fell to the same extent in each of the two regions. The latter finding indicates that during ischemia perfusion was distributed uniformly in each of the aggregate core areas and that these areas did not include a variable admixture of tissue from overlapping border regions.
A second assumption important for interpreting our results is that the intracoronary adenosine produced maximum vasodilation and was delivered selectively to the distal LC region. We found that intracoronary adenosine infused at normal coronary pressures increased flow to the distal region by over five times that found under resting conditions. Although we did not assess the reactive hyperemic response since coronary flow was measured with microspheres, infusion of a twofold higher dose of adenosine into the entire LC region (ie, both proximal and distal LC regions) in a previous study abolished reactive hyperemia.17 In addition, estimates of plasma adenosine concentration based on the present infusion rate and flow during vasodilation of 180 mL/min17 result in a calculated plasma concentration exceeding 9 μmol/L. As coronary flow is reduced to resting values in the face of a constant infusion, plasma adenosine concentrations would exceed 45 μmol/L. Both of these values are one to two orders of magnitude greater than the ED50 for adenosine of 0.57 μmol/L demonstrated by Olsson et al18 in conscious dogs. Flow measurements from myocardial samples in the proximal autoregulating region and in the normally perfused LAD region did not change, indicating that adenosine was metabolized to levels that did not alter coronary vascular resistance before it recirculated to the systemic arterial circulation. In a previous study, we found that infusion of a twofold higher dose of adenosine into the pulmonary artery also had no effect on coronary flow.17 Evidence that the intracoronary administration of adenosine remained selective at reduced coronary pressures is less direct. Although adenosine had no effect on subendocardial perfusion during ischemia, there were small increases in subepicardial blood flow at coronary pressures that, although reduced, were still above the lower pressure limit of subepicardial autoregulation. Thus, if adenosine refluxed into the proximal region at reduced coronary pressure, subepicardial flow in the proximal autoregulating region should have also increased in a fashion similar to that in the distal region. Since subepicardial flow in the proximal region did not increase, we feel confident that the selective infusion of adenosine into the distal LC region at reduced pressures was comparable to the infusion that we demonstrated at normal coronary pressures.
An additional methodological concern relates to the variability of flow due to the microsphere technique. The accuracy of regional perfusion measurements is dependent on the number of microspheres in each sample.16 This varies during ischemia and is also dependent on the size of the subregion that is being examined. On average, our core regions had >400 microspheres (Table 3⇑), and the relative error would be expected to be <5%. The lowest subendocardial flow associated with severe ischemia still had >100 microspheres per sample and would have a relative error of <10%. These errors would limit our ability to detect a 5% to 10% difference in flow after adenosine administration during severe ischemia. Flow heterogeneity among small myocardial regions of interest could have also influenced our results. Coggins et al19 have shown that pharmacologically recruitable reserve during ischemia in open-chest anesthetized animals can sometimes arise from spatial perfusion heterogeneity in both resting and maximal flow to small ventricular regions. They found that the presence of adenosine-recruitable reserve during ischemia in large myocardial regions in their preparations could be explained by the presence of an admixture of subregions that had exhausted flow reserve (with no effect of adenosine) and nonischemic subregions that were still autoregulating flow normally and in which vasodilator reserve was still present. The effects of flow heterogeneity in their study were dependent on coronary pressure. They were most pronounced at higher coronary pressures and completely abolished in the subendocardium when pressure reached 30 mm Hg. In contrast to their results, we found that adenosine-recruitable subendocardial reserve was absent at pressures averaging as high as 33 mm Hg (Table 2⇑). This continued to be true at all levels of coronary pressure and subendocardial flow studied. Thus, although flow heterogeneity within each aggregate core region was undoubtedly present, we were unable to demonstrate pharmacologically recruitable increases in subendocardial flow. We believe that this discrepancy can be resolved by differences in the magnitude of perfusion heterogeneity found in anesthetized as opposed to unanesthetized animals. In the study of Austin et al,16 the relative dispersion of flow at normal coronary pressures was larger (coefficient of variation [CV], 24.3% at rest) compared with values that we have previously reported in unanesthetized animals (CV, 19.5% at rest).20 There are probably several explanations for differences in perfusion heterogeneity in resting flow between anesthetized and unanesthetized animals, including more heterogeneous oxygen extraction and a more limited ability to maintain flow constant as pressure is reduced during autoregulation in anesthetized compared with unanesthetized animals. An additional possibility explaining the differences is that maximally vasodilated flows in anesthetized animals appear to be more heterogeneous than those at rest19 (CV, 30.4% during vasodilation). We have recently reported that within individual transmural layers, the CV for flow actually decreases in unanesthetized animals during vasodilation.20 Thus, based on the results of the present study, spatial perfusion heterogeneity does not invariably result in “apparent” recruitable subendocardial flow reserve during ischemia.
Lack of Adenosine-Recruitable Subendocardial Flow Reserve Compared With Previous Studies
The failure of adenosine to increase subendocardial perfusion at reduced coronary pressure contrasts with the results of previous studies conducted during autoregulation in open-chest anesthetized animals from our own laboratory1 as well as others.2 3 4 Fig 6⇓ compares measurements of subendocardial perfusion during autoregulation and adenosine infusion at matched coronary pressures from the present study (solid triangles) with those from the four previously published studies.1 2 3 4 The solid line shows a perfect match between autoregulated flow and subendocardial vasodilator reserve. Collectively, the experiments conducted in open-chest anesthetized animals at controlled coronary pressure found that after intracoronary vasodilation, subendocardial flow in the setting of ischemia could increase above the corresponding flow during autoregulation by an average of 80%. In contrast, the data from unanesthetized animals in the present study demonstrate a lack of adenosine-recruitable subendocardial flow reserve. This could not be explained by the severity of ischemia, since there was no effect of adenosine on subendocardial flow during mild, moderate, or severe reductions in circumflex pressure (Table 2⇑ and Fig 4⇑). Thus, in the absence of anesthesia and/or acute surgical instrumentation, coronary autoregulation in unanesthetized animals is able to match subendocardial perfusion to the local vasodilator reserve recruitable by pharmacological agents. The ability of intrinsic autoregulatory mechanisms to match flow to local vasodilator reserve probably explains our previous findings that revealed a substantially lower autoregulatory pressure limit of 40 mm Hg in the basal unanesthetized state9 that increased to only 60 mm Hg during pacing at a rate of 200 beats per minute.21 Additionally, the levels of flow during adenosine vasodilation are similar to those demonstrated by Bache and Schwartz22 at comparable levels of reduced coronary pressure in anesthetized animals.
The lack of pharmacologically recruitable subendocardial flow reserve in the basal state also contrasts with findings during regional ischemia and sympathetic activation in anesthetized as well as unanesthetized animals. Stellate ganglion stimulation23 and intracoronary norepinephrine infusion24 reduce perfusion in the setting of regional myocardial ischemia. Similarly, physiological activation of adrenergic constrictor mechanisms following exercise in unanesthetized dogs also modulates myocardial perfusion in the setting of ischemia distal to a flow-limiting coronary artery stenosis. Seitelberger et al6 and Laxson et al7 found that α-adrenergic blockade improved myocardial perfusion during ischemia, although the specific adrenergic subtype responsible for this improvement was controversial. Laxson et al8 subsequently showed that subendocardial perfusion during exercise at a constant coronary pressure could also be increased by pharmacological vasodilation with intracoronary adenosine. These investigators were careful to prevent confounding effects of a transmural steal during adenosine by keeping pressure distal to the stenosis constant during exercise.
Although previous studies clearly show that sympathetic activation can modulate subendocardial flow in unanesthetized animals, the level of activation required to overcome intrinsic autoregulatory mechanisms in the subendocardium does not appear to be present when unanesthetized animals are studied in the basal state. Our observations are consistent with the finding that resting coronary flow is not altered by adrenergic denervation in animals studied in the basal state.12 In further support of this are the observations of Chilian and Ackell,13 who found no effect of denervation on myocardial perfusion at moderate levels of ischemia until sympathetic activation was produced by submaximal exercise. In addition, Minisi and Thames14 have shown that reflex sympathetic activation does not occur until severe transmural ischemia is produced and that this primarily affects the subepicardial layers of the left ventricular wall. Thus, even though the droperidol, which was used to sedate the animal in this study, has weak α-blocking properties in vivo25 (Fig 5⇑), it would not have affected our interpretation of the results during mild to moderate ischemia due to the lack of sympathetic activation and/or elevations in circulating catecholamine levels. During severe ischemia, it could have affected the magnitude of adenosine-recruitable subepicardial reserve (see below).
Significance of Adenosine-Recruitable Subepicardial Reserve
Although adenosine did not affect perfusion to the subendocardium, flow to the outer layers increased to a small but significant amount. Much of the increase could be explained by the fact that even though subendocardial flow was reduced below control, corresponding subepicardial flow measurements were frequently not reduced because coronary pressure continued to exceed the lower limit of autoregulation, which we have previously demonstrated to be near 25 mm Hg for the subepicardium.9 Although this appears to account for most of the recruitable flow reserve in outer layers, there continued to be a small further increase in subepicardial flow during severe ischemia, when subepicardial flow during autoregulation was reduced below control values. The possibility that this small difference was due to flow heterogeneity is difficult to exclude. Nevertheless, we have previously reported that the magnitude of flow heterogeneity under resting conditions is similar within different layers of the left ventricle in unanesthetized dogs,20 and it would be difficult to explain why flow increases in the subepicardium and not the subendocardium on this basis. It is also probably not secondary to transmural differences in the ability of individual myocardial layers to autoregulate flow, since we have previously demonstrated that closed-loop autoregulatory gain is similar in subendocardial and subepicardial layers of the left ventricular free wall in unanesthetized dogs.9 21 Although Chilian et al12 showed that there was no significant difference in resting perfusion to denervated as opposed to innervated myocardial regions (demonstrating the lack of an adrenergic constrictor tone in unanesthetized dogs at rest), a small degree of reflex sympathetic activation could have occurred in response to severe regional myocardial ischemia when we reduced coronary pressure. This is supported by the fact that heart rate increased significantly during the most severe levels of coronary pressure reduction (Table 2⇑). Such a mechanism could serve to preferentially direct flow to the subendocardium in the face of transmural ischemia.13 26 In addition, the degree of flow reduction during severe ischemia is close to that required to cause increases in cardiac sympathetic nerve activity during regional ischemia (subendocardial flow, 26% of control; subepicardial flow, 54% of control).14 As shown in Fig 5⇑, the weak α-adrenergic blocking properties of droperidol may have caused us to underestimate the magnitude of recruitable subepicardial flow reserve during severe ischemia.
In summary, the present study demonstrates that pharmacologically recruitable flow reserve at reduced coronary pressures during autoregulation in unanesthetized animals is minimal. During ischemia, there is no effect of adenosine on subendocardial flow, and the small increases in flow occurring to the outer portions of the left ventricular free wall are largely due to the fact that coronary pressure continues to be above the lower autoregulatory pressure limit for these layers. Although speculative, both the shift in the autoregulatory limit observed in anesthetized open-chest animals compared with unanesthetized animals and the presence of pharmacologically recruitable subendocardial vasodilator reserve during ischemia probably reflect effects of abnormally high circulating catecholamine levels and/or adrenergic constrictor tone competing with, and overcoming, intrinsic stimuli for vasodilation. This is supported by the finding that α1-adrenergic blockade, like adenosine, increases flow at reduced coronary pressures during autoregulation in anesthetized open-chest dogs.27 Based on our findings, studies designed to examine the presence of pharmacologically recruitable flow reserve during autoregulation distal to a coronary stenosis need to consider the possibility that observations in anesthetized open-chest animals may be modulated by abnormal levels of sympathetic activation and/or circulating catecholamines. This may influence intrinsic autoregulatory mechanisms in a fashion that may not be relevant to the control of coronary flow in the basal unanesthetized state.
This study was supported by National Heart, Lung, and Blood Institute grant HL-15194, American Heart Association Grant-in-Aid 93-966, and a merit review award from the Veterans Administration. This study could not have been completed without the expert technical assistance of Kathleen Harris, Kathleen Weibel, and Amy Johnson. The secretarial assistance of Nina Goss is greatly appreciated.
Reprint requests to John M. Canty, Jr, MD, SUNY/B Clinical Center, Room CC-172, 462 Grider St, Buffalo, NY 14215-3012.
- Received January 31, 1994.
- Accepted February 28, 1995.
- © 1995 American Heart Association, Inc.
Canty JM Jr, Klocke FJ. Reduced regional myocardial perfusion in the presence of pharmacologic vasodilator reserve. Circulation. 1985;71:370-377.
Aversano T, Becker LC. Persistence of coronary vasodilator reserve despite functionally significant flow reduction. Am J Physiol. 1985;248:H403-H411.
Pantely GA, Bristow JD, Swenson LJ, Ladley HD, Johnson WB, Anselone CG. Incomplete coronary vasodilation during myocardial ischemia in swine. Am J Physiol. 1985;249:H638-H647.
Grattan MT, Hanley FL, Stevens MB, Hoffman JIE. Transmural coronary flow reserve patterns in dogs. Am J Physiol. 1986;250:H276-H283.
Heusch G, Guth BD, Seitelberger R, Ross J Jr. Attenuation of ischemia in dogs with recruitment of coronary vasodilator reserve by nifedipine. Circulation. 1987;75:482-490.
Seitelberger R, Guth BD, Heusch G, Lee JD, Katayama K, Ross J Jr. Intracoronary α2-adrenergic receptor blockade attenuates ischemia in conscious dogs during exercise. Circ Res. 1988;62:436-442.
Laxson DD, Dai XZ, Homans DC, Bache RJ. The role of α1 and α2-adrenergic receptors in mediation of coronary vasoconstriction in hypoperfused ischemic myocardium during exercise. Circ Res. 1989;65:1688-1697.
Laxson DD, Dai X-Z, Homans DC, Bache RJ. Coronary vasodilator reserve in ischemic myocardium of the exercising dog. Circulation. 1992;85:313-322.
Canty JM Jr. Coronary pressure-function and steady-state pressure-flow relations during autoregulation in the unanesthetized dog. Circ Res. 1988;63:821-836.
Gallagher KP, Kumada T, Kozioc JA, McKown MD, Kemper WS, Ross J Jr. Significance of regional wall thickening abnormalities relative to transmural myocardial perfusion in anesthetized dogs. Circulation. 1980;62:1266-1274.
Chilian WM, Boatwright RB, Shoji T, Griggs DM Jr. Evidence against significant resting sympathetic coronary vasoconstrictor tone in the conscious dog. Circ Res. 1981;49:866-876.
Chilian WM, Ackell PH. Transmural differences in sympathetic coronary constriction during exercise in the presence of coronary stenosis. Circ Res. 1988;62:216-225.
Minisi AJ, Thames MD. Activation of cardiac sympathetic afferents during coronary occlusion: evidence for reflex activation of sympathetic nervous system during transmural myocardial ischemia in the dog. Circulation. 1991;84:357-367.
Austin RE Jr, Aldea GS, Coggins DL, Flynn AE, Hoffman JIE. Profound spatial heterogeneity of coronary reserve: discordance between patterns of resting and maximal myocardial blood flow. Circ Res. 1990;67:319-331.
Canty JM Jr, Schwartz JS. Nitric oxide mediates flow-dependent epicardial coronary vasodilation to changes in pulse frequency but not mean flow in conscious dogs. Circulation. 1994;89:375-384.
Olsson RA, Khouri EM, Bedynek JL Jr, McLean J. Coronary vasoactivity of adenosine in the conscious dog. Circ Res. 1979;45:468-478.
Coggins DL, Flynn AE, Austin RE Jr, Aldea GS, Muehrcke D, Goto M, Hoffman JIE. Nonuniform loss of regional flow reserve during myocardial ischemia in dogs. Circ Res. 1990;67:253-264.
Canty JM Jr, Weibel K. Disparate effects of ischemic and pharmacologic vasodilation on myocardial flow heterogeneity in conscious dogs. Circulation. 1989;80(suppl II):II-191. Abstract.
Canty JM Jr, Giglia J, Kandath D. Effect of tachycardia on regional function and transmural myocardial perfusion during graded coronary pressure reduction in conscious dogs. Circulation. 1990;82:1815-1825.
Bache RJ, Schwartz JS. Effect of perfusion pressure distal to a coronary stenosis on transmural myocardial blood flow. Circulation. 1982;65:928-935.
Heusch G, Deussen A. The effects of cardiac sympathetic nerve stimulation on perfusion of stenotic coronary arteries in the dog. Circ Res. 1983;53:8-15.
Buffington CW, Feigl EO. Effect of coronary artery pressure on transmural distribution of adrenergic coronary vasoconstriction in the dog. Circ Res. 1983;53:613-621.
Nathan HJ, Feigl EO. Adrenergic vasoconstriction lessens transmural steal during coronary hypoperfusion. Am J Physiol. 1986;250:H645-H653.
Liang IYS, Jones CE. α1-Adrenergic blockade increases coronary blood flow during coronary hypoperfusion. Am J Physiol. 1985;249:H1070-H1077.