Pathophysiology of Chronic Left Ventricular Dysfunction
New Insights From the Measurement of Absolute Myocardial Blood Flow and Glucose Utilization
Background Chronically dysfunctional myocardium may improve after coronary revascularization. This condition was thought to be due to a chronically reduced myocardial blood flow (MBF). Recently, however, it has been shown that in patients without previous infarction but with chronic left ventricular dysfunction, baseline MBF was normal.
Methods and Results To study the pathophysiology of chronic left ventricular dysfunction in patients with previous infarction, regional MBF (milliliter per minute per gram of water-perfusable tissue) and glucose utilization (MRG; micromoles per minute per gram) during hyperinsulinemic euglycemic clamp were measured with positron emission tomography in 30 patients before bypass. At baseline, 133 myocardial segments were normal, and 107 were dysfunctional. After revascularization, 59 of 107 segments improved, while 48 of 107 were unchanged. MBF was 0.92±0.25 mL · min−1 · g−1 in normal segments, 0.87±0.31 mL · min−1 · g−1 in improved segments (P=NS versus normal), and 0.82±0.40 mL · min−1 · g−1 in unchanged segments (P<.05 versus normal). In 90% of the dysfunctional segments, MBF was >0.42 mL · min−1 · g−1, a cutoff value corresponding to the mean MBF minus 2 SD in normal segments. The MRG was 0.71±0.14 μmol · min−1 · g−1 in 9 age-matched normal subjects, 0.45±0.19 μmol · min−1 · g−1 (P<.01) in normal segments, 0.44±0.14 μmol · min−1 · g−1 in improved segments (P=NS versus normal), and 0.34±0.17 μmol · min−1 · g−1 in unchanged segments (P<.01 versus normal and improved).
Conclusions The results suggest that resting MBF measured with 15O-labeled water in chronically dysfunctional segments is not reduced and that the myocardium of these patients is less sensitive to insulin than that of normal subjects.
It is now well established that myocardial regions with chronic wall motion abnormalities subtended by stenotic arteries may have improved function after coronary revascularization. To describe this condition, the term “hibernating myocardium” was introduced by Rahimtoola.1 On the basis of a retrospective analysis of three studies, which included a large number of patients with coronary artery disease randomized to receive either medical or surgical treatment, he concluded that “there is a prolonged subacute or chronic stage of myocardial ischemia that is frequently not accompanied by pain and in which myocardial contractility and metabolism are reduced to match the reduced blood supply.”1
This hypothesis was challenged recently by Vanoverschelde et al,2 who demonstrated that baseline MBF was normal in chronically dysfunctional myocardium subtended by occluded coronary arteries in patients with angina but without previous infarction. These authors suggested that repeated episodes of ischemia rather than chronic hypoperfusion might form the basis of chronic LV dysfunction in these patients.
In most cases, however, chronic myocardial dysfunction is detected in patients with previous infarction. In these patients, wall motion abnormalities may be the result of permanent anatomic damage and/or a state of hibernation. It has been demonstrated that the simultaneous assessment of myocardial blood flow and metabolism with PET allows differentiation between scarred and viable myocardium in patients with chronic LV dysfunction who are potential candidates for coronary revascularization.3 4 5 6 In this context, however, PET has been used primarily to provide qualitative or semiquantitative information rather than to exploit its capability to give absolute measurements. In addition, no effort has been made to standardize the dietary state of patients undergoing the PET studies, thereby precluding a reliable quantification of the metabolic data.7 The latter point is particularly relevant given the high prevalence of insulin resistance among patients with coronary artery disease.
The aim of the present study was to investigate the pathophysiology of chronic LV dysfunction in patients with previous infarction before coronary artery bypass graft. With PET, absolute regional MBF and MRG during hyperinsulinemic euglycemic clamp8 were measured.
The patient population consisted of 30 patients (2 women, 28 men; mean [±SD] age, 56±11 years; range, 34 to 72 years) with coronary artery disease and at least one chronically dysfunctional LV region subtended by a diseased coronary artery amenable to revascularization (Table 1⇓). All patients had suffered a previous myocardial infarction at least 6 months before the study; 5 patients were diabetic, and 4 had arterial hypertension.
Normal Control Subjects
A group of 25 normal volunteers (9 women, 16 men; mean age, 56±12 years; range, 33 to 77 years; P=NS versus patients) served as control subjects for the MBF measurements. A second group of 9 normal volunteers (all men; mean age, 47±7 years; range, 31 to 56 years; P=NS versus patients) served as control subjects for measurement of the myocardial MRG during hyperinsulinemic euglycemic clamp. The normal volunteers were selected on the basis of their clinical histories and physical examinations, which indicated a low risk of coronary artery disease. All had normal resting ECGs and negative exercise tests in response to a high workload.
Selective arteriography of the right and left coronary arteries in multiple views was performed in all patients by use of the Judkins technique.9 The percent reduction of the internal luminal diameter in the projection with maximal severity was assessed visually by one of the investigators (P.G.C.).
Coronary Bypass Graft
All patients except 1 were operated on by one of the investigators (B.E.K.) using St Thomas’ crystalloid cardioplegia supplemented with iced topical for myocardial protection. The median number of bypass grafts was four per patient; all but 1 received a left internal mammary artery graft to the left anterior descending coronary artery.
All patients underwent radionuclide ventriculography10 before and 4 to 6 months after coronary bypass. Briefly, after red blood cells had been labeled in vivo with 740 MBq of technetium-99m sodium pertechnetate, the intracardiac blood pool was imaged with a gamma camera (GE 400 XCT) equipped with a low-energy all-purpose collimator and interfaced with a dedicated computer system (S3000). Data for each cardiac cycle, synchronized to the R wave on the ECG, were divided into 21 frames. Six million counts per view were acquired. Planar blood pool imaging was performed in two projections: the left anterior oblique projection (best septal separation) with a 20° caudal tilt and the left posterior oblique projection.11 LV ejection fraction was obtained from the left anterior oblique projection. Regional wall motion was assessed qualitatively from an endless-loop cine format by two observers (N.V.S.M. and D.C.C.) who were unaware of the patient’s diagnosis. The wall motion was graded as 0 (normal), 1 (hypokinetic), 2 (akinetic), or 3 (dyskinetic). The regional wall motion score was calculated by adding the values of all segments.
Hypersinulinemic Euglycemic Clamp
Before PET scanning, which was carried out in all patients between 11 am and 1 pm after a light breakfast with the patient lying on the scanner bed, a 20-gauge polyethylene cannula was inserted in a superficial forearm vein for infusion of glucose and insulin as described by De Fronzo et al.8 A second cannula was inserted retrogradely into a superficial vein of the wrist or hand that had been arterialized with a commercially available heating pad set at 50°C. The degree of arterialization was checked by measurement of respiratory gases on a blood sample drawn after a 30-minute heating period. At time zero, the insulin infusion was started. Insulin was given at four times the final constant rate (calculated according to De Fronzo et al8 ) for the first 4 minutes, then at two times the final constant rate for the following 3 minutes, and then at a constant rate for the remainder of the study. At 4 minutes, exogenous glucose infusion was started at an initial rate of 1.5 mg · min−1 · kg−1 body wt. The blood glucose concentration from the arterialized vein was measured at baseline (>30 minutes after the insertion of the cannula) and every 5 minutes during the clamp. The glucose infusion rate was adjusted according to the change in plasma glucose over the preceding 5 minutes. Samples for insulin assay were taken from the arterialized line immediately before insulin infusion was begun and 60 minutes into the clamp.
Positron Emission Tomography
The PET study for the measurement of MBF and MRG was carried out within a week of radionuclide ventriculography. All PET scans were performed with an ECAT 931-08/12 scanner (CTI Inc), which consisted of eight rings of bismuth germanate crystal detectors. This scanner enables the acquisition of 15 planes of data over a 10.5-cm axial field of view, thus allowing the whole heart to be imaged. All emission and transmission sinograms were reconstructed with a Hanning filter with a cutoff frequency of 0.5 maximum. This resulted in a spatial resolution of 8.4-mm FWHM for the emission and 7.7-mm FWHM for the transmission data at the center of the field of view, with a slice thickness of 6.6-mm FWHM.12 All subjects lay supine on the scanner bed. The optimal imaging position was determined by use of a 5-minute rectilinear scan after exposure of the external 68Ge ring source. A 20-minute transmission scan was then performed. These data were used to correct subsequent emission scans for tissue attenuation of the annihilation gamma photons.
After the transmission scan, the blood pool was imaged by inhalation of tracer amounts of C15O, which labels erythrocytes through the formation of carboxyhemoglobin. C15O was administered for 4 minutes at a concentration of 3 MBq/mL and a flow rate of 500 mL/min. A 6-minute single-frame emission scan was initiated 1 minute after the end of C15O inhalation to allow equilibration.13 Venous blood samples were taken every minute during the scan, and the C15O concentration in whole blood was measured with an NaI well counter cross-calibrated with the scanner.
After a 15-minute period for decay of 15O radioactivity to background levels, MBF was measured with inhaled C15O2, which is rapidly converted to H215O by carbonic anhydrase in the lungs.13 C15O2 was inhaled for 3.5 minutes (4 MBq/mL at 500 mL/min). A 25-frame dynamic PET scan (frame durations, 1×30 [background], 6×5, 6×10, 6×20, and 6×30 seconds) covering a period of 7 minutes was started 30 seconds before C15O2 inhalation.
MRG during hyperinsulinemic euglycemic clamp was measured with the glucose analogue 18FDG. 18FDG (185 MBq) was infused intravenously over 2 minutes with a pump beginning 30 seconds after the beginning of the scan. A 36-frame dynamic PET scan with progressive increases in frame duration (1×30 [background], 12×10, 3×20, 4×30, 5×60, 4×150, 5×300, and 2×600 seconds) was performed over a total period of 65 minutes. Arterialized whole blood was withdrawn continuously at 5 mL/min for the first 10 minutes and 2.5 mL/min thereafter, and an on-line detection system, cross-calibrated against the PET scanner, was used to measure radioactivity in blood as described previously.14 At set times (5, 10, 20, 45, and 60 minutes after the start of the 18FDG infusion), continuous blood withdrawal was interrupted briefly for the collection of blood samples, which were used to estimate plasma-to-whole-blood ratios of radioactivity. After each sample, the line was flushed with heparinized saline.
The study protocol was approved by the Research Ethics Committee of Hammersmith Hospital, and radiation exposure was licensed by the UK Administration of Radioactive Substances Advisory Committee. All patients gave written informed consent before the study.
PET Data Analysis
All sinograms were corrected for tissue attenuation and reconstructed on a MicroVax II computer (Digital Equipment Corp) with dedicated array processors and standard reconstruction algorithms. Images were transferred to SPARC 2 workstations (Sun Mycrosystems) for further analysis. Image manipulation and data handling were performed with the ANALYZE (version 3.0, Biodynamics Research Unit, Mayo Foundation)15 and MATLAB (The MathWorks Inc) software packages. All images were resliced in the short-axis view.
Initially, a myocardial blood volume image was generated by dividing the C15O image by the average concentration of the blood samples on a voxel-by-voxel basis and the density of whole blood (1.06 g/mL). Appropriate corrections for decay were applied to both the C15O image and the blood samples. These images were used to position two to four regions of interest, each with an average size of 1.5 mL, in the left atrial chamber so that myocardial blood volume and thus the recovery of counts were >90%. These regions of interest were then projected onto the dynamic H215O images to generate arterial time-activity curves. The average of these atrial curves was used as the arterial input function for the subsequent kinetic MBF analysis.13 The above-mentioned atrial regions of interest were not used to generate the input function for the kinetic 18FDG analysis. Because of the high tissue-to-blood ratio at late times, even a limited amount of spillover would significantly affect the tail of the blood curve. Instead, the continuously monitored arterialized venous whole-blood curve was used. This curve was multiplied by the average plasma-to-whole-blood ratio obtained from the discrete samples to generate the plasma input function. A correction for the delay of this curve (arm and tubing) was made by shifting it so that the initial rise coincided with that from the atrial regions of interest.
Four equally spaced sectors corresponding to the anterior, septal, lateral, and inferior myocardium were defined separately on each plane of the last 18FDG frame. Within each sector, three to four elliptical regions of interest (1 mL each) were drawn. These regions of interest also were projected onto the entire dynamic 18FDG and H215O data sets, and tissue time-activity curves were generated for each region of interest.
The tissue H215O time-activity curves were fitted for MBF and TF (the fraction of tissue within a region of interest that exchanges water rapidly) with standard nonlinear regression techniques and a tracer kinetic model described previously.13 Unlike other implementations, this model provides MBF values per 1 mL perfusable tissue (not per 1 mL of region of interest), based on the assumption that the uptake of H215O in scar tissue is negligible compared with that in normal myocardium. Therefore, in a myocardial region consisting of an admixture of normal and necrotic tissue, this model predominantly measures flow to the residual normal myocardium.16 At variance, the flow measured with other methods, eg, 13N-labeled ammonia, represents an average flow per unit mass of tissue as with the microsphere technique.17 Therefore, to allow comparison with published data, flow per unit mass of tissue was also computed as MBF times TF.
Because basal MBF is closely related to the RPP,18 basal flow data also were corrected for the RPP, an index of myocardial oxygen consumption, by the following equation: RPP−Corrected Basal Flow=Basal Flow×(Mean Patient RPP÷Individual RPP). The mean patient RPP also was used for correction to compute the corrected basal flow in the 25 normal subjects to normalize the entire study population to the same workload.
Tissue 18FDG time-activity curves were analyzed by use of the linearized approach proposed by Patlak et al19 for irreversible processes. The ratio of tissue concentration to plasma concentration was plotted against the ratio of the integral of the plasma concentration to the plasma concentration, and a linear regression was performed for all data points corresponding to times >10 minutes after injection. The slope of this line provides the net influx rate of 18FDG. The myocardial MRG was then obtained by multiplying these regional influx rates by the plasma concentration of stable glucose, assuming a lumped constant of one, and by dividing the product by the corresponding TF. This last step was performed to correct for partial volume effects, thereby allowing comparison with corresponding MBF values. A conversion from milliliters to grams of perfusable tissue was made by dividing the flow and metabolic data by tissue density (1.04 g/mL). Thus, the MBF values are expressed as milliliter per minute per gram and the MRG values as micromole per minute per gram of perfusable tissue.
All data are presented as mean±SD. Student’s t test was used to compare any pair of mean group values. Simultaneous comparison of more than two mean values was performed with one-way ANOVA for repeated measures; Fisher’s least significant difference method was subsequently applied to localize the source of the difference.20 Regression analysis was performed according to standard techniques. A value of P<.05 was considered significant.
Ventricular Function and Metabolic Profile
Before surgery, global LV ejection fraction was 35±11% (range, 12% to 55%). A total of 247 LV segments were analyzed in the 30 patients. Of these, 114 (46%) were dysfunctional and 133 (54%) had a normal function. Regional wall motion score was 5.3±3.1 (range, 1 to 11). Blood glucose concentration, measured before the beginning of the clamp, was 7.1±3.4 mmol/L (range, 4 to 22 mmol/L). Five patients were known diabetics (3 had insulin-dependent and 2 had non–insulin-dependent diabetes). Baseline plasma insulin concentration was 16±12 mU/L (range, 2 to 40 mU/L). During clamp, euglycemia was achieved in all but 1 patient, and blood glucose concentration decreased to 5.8±1.3 mmol/L (range, 4 to 11 mmol/L; P<.01 versus baseline). Plasma insulin during clamp was 88±17 mU/L (range, 68 to 133 mU/L; P<.0001 versus baseline). In the 9 normal subjects, baseline blood glucose was 5.3±0.9 mmol/L (range, 4.1 to 6.7 mmol/L; P=NS versus patients), and plasma insulin was 19±1 mU/L (range, 17 to 21 mU/L; P=NS versus patients). During clamp, blood glucose was 6.1±0.9 (range, 4.8 to 7.6 mmol/L; P=NS versus patients), and plasma insulin was 78±6 mU/L (range, 68 to 87 mU/L; P=NS versus patients).
Global LV ejection fraction after coronary bypass was 36±12% (range, 12% to 60%; P=NS versus baseline). In contrast, regional wall motion score improved significantly to 3.4±3.0 (range, 0 to 11; P<.001 versus corresponding baseline value), suggesting a change in the regional contribution to the global LV ejection fraction.
Ten segments (3 normal and 7 dysfunctional) worsened after surgery and were excluded from the PET data analysis, leaving a total of 107 dysfunctional and 130 normal segments. Of the 107 dysfunctional segments, 59 (55%) had a functional improvement after surgery, and 48 (45%) were unchanged.
Myocardial Blood Flow
In the 25 normal volunteers, MBF was homogeneously distributed in the different ventricular regions, and mean LV MBF was 1.02±0.25 mL · min−1 · g−1 (range, 0.69 to 1.52 mL · min−1 · g−1). After correction for the resting RPP, mean LV MBF was 1.02±0.40 mL · min−1 · g−1 (range, 0.56 to 1.79 mL · min−1 · g−1). In the patients, MBF in the 130 segments with normal function was 0.92±0.25 mL · min−1 · g−1 (range, 0.43 to 1.56 mL · min−1 · g−1; P=.001 versus normal subjects) and 1.00±0.33 mL · min−1 · g−1 (range, 0.37 to 2.45 mL · min−1 · g−1) after correction for RPP (P=NS versus normal subjects). MBF was 0.87±0.31 mL · min−1 · g−1 (range, 0.15 to 1.64 mL · min−1 · g−1) in the 59 segments that improved after bypass (P=NS versus normal segments) and 0.82±0.40 mL · min−1 · g−1 (range, 0.15 to 2.74 mL · min−1 · g−1) in the 48 segments that were unchanged after surgery (P<.05 versus normal segments). To identify dysfunctional segments with abnormally low MBF, a cutoff value of 0.42 mL · min−1 · g−1, which corresponds to the mean MBF (corrected for RPP) minus 2 SD in normal segments, was used. In 11 of 107 dysfunctional segments (10%), MBF was <0.42 mL · min−1 · g−1. After surgery, regional wall motion improved in 6 of 11, whereas it was unchanged in 5. In 96 of 107 dysfunctional segments (90%), MBF was >0.42 mL · min−1 · g−1. After surgery, regional wall motion improved in 53 of 96 (55%) and was unchanged in 43 despite comparable MBF values (0.93±0.26 versus 0.88±0.38 mL · min−1 · g−1; Fig 1⇓).
Metabolic Rate of Glucose
In the 9 normal volunteers, MRG was 0.71±0.14 μmol · min−1 · g−1 (range, 0.46 to 1.02 μmol · min−1 · g−1). In the patients, the MRG in the 130 segments with normal function was 0.45±0.19 μmol · min−1 · g−1 (range, 0.10 to 1.02 μmol · min−1 · g−1; P<.01 versus normal subjects; Fig 2⇓). The difference between normal subjects and patients wasstill significant after exclusion of the 5 patients with diabetes and the 4 with arterial hypertension (0.46±0.18 μmol · min−1 · g−1, P<.01).
The MRG in the 59 segments that improved after bypass was 0.44±0.14 μmol · min−1 · g−1 (range, 0.04 to 0.71 μmol · min−1 · g−1; P=NS versus normal segments). In the 48 segments that were unchanged after surgery, MRG was 0.34±0.17 μmol · min−1 · g−1 (range, 0.09 to 0.78 μmol · min−1 · g−1; P<.01 versus both normal and improved segments; Fig 3⇓). MRG in the 11 of 107 dysfunctional segments with MBF <0.42 mL · min−1 · g−1 was 0.26±0.17 μmol · min−1 · g−1 (range, 0.04 to 0.57 μmol · min−1 · g−1). In the 96 of 107 dysfunctional segments with MBF >0.42 mL · min−1 · g−1, MRG was 0.40±0.15 μmol · min−1 · g−1 (range, 0.09 to 0.75 μmol · min−1 · g−1; P<.01 versus segments with MBF <0.42 mL · min−1 · g−1).
Blood Flow in Chronically Dysfunctional Myocardium
The present study demonstrates that most chronically dysfunctional segments in a series of patients with previous infarction have MBF values measured with H215O and PET within the normal range (Fig 1⇑). Because the definition of low MBF (ie, mean MBF minus 2 SD in normally contracting myocardium) is somewhat arbitrary, a less conservative cutoff value of 0.67 mL · min−1 · g−1 (mean MBF minus 1 SD in normally contracting myocardium) also was considered. Even in the latter case, >80% of the dysfunctional segments had MBF values >0.67 mL · min−1 · g−1. The latter cutoff value is identical to that proposed by Bergmann et al21 and very similar to that (0.70 mL · min−1 · g−1) proposed by Perrone-Filardi et al,22 both obtained in normal subjects by use of H215O and PET.
The present finding indicates that chronic ventricular dysfunction is not accompanied by a chronically reduced MBF. It is worth noting, however, that the kinetic model used in the present study to quantify MBF with H215O provides values of flow per 1 g perfusable tissue (not per 1 g region of interest).13 Because the uptake of H215O in scar tissue is negligible compared with normal myocardium, in a myocardial region consisting of an admixture of viable and necrotic tissue, this model predominantly measures flow to the residual normal myocardium.16 At variance, the flow measured with other tracers, eg, 13N-labeled ammonia (13NH3), represents an average flow per unit mass of tissue as with the microsphere technique.17 In other words, dysfunctional regions in which perfusion is measured as milliliter per minute per gram of tissue, including fibrotic and viable myocardium (eg, 13NH3), would have low perfusion and therefore would be defined as hibernating. However, the same dysfunctional segments in the same heart in which myocardial perfusion is measured as milliliter per minute per gram of viable tissue (water-diffusible space), excluding fibrotic tissue, might have normal flow and therefore would be defined as stunned.2 To be able to compare our data with those obtained with 13NH3, our MBF values, expressed as flow per 1 g perfusable tissue, were multiplied by TF to yield values of 1 mL per unit mass of total tissue (Table 2⇓). It should be noted that even in control subjects MBF expressed as milliliter per minute per gram of total tissue is lower than that expressed as milliliter per minute per gram of water-perfusable tissue. The reason is that the latter measurement includes an intrinsic correction for partial volume effect.13 23
In several other studies, MBF has been assessed by use of PET with 13NH3 in patients with chronic myocardial infarction. In 15 patients with reperfused anterior infarction who were studied 42±25 days after the acute event, transmural flow averaged 0.8±0.1, 0.7±0.2, and 0.5±0.1 mL · min−1 · g−1 in remote segments, adjacent segments, and segments at the center of the infarcted area.24 Similar findings were reported recently in 26 patients with chronic healed infarction studied 44±65 months after the event.25 In some of these studies, MBF measured with 13NH3 in the core of the infarcted segments was lower than the flow values obtained by multiplying MBF measured with H215O by TF. This could be explained, at least in part, by the size and location of the regions of interest used in the different studies. In the present study, rather large fixed regions of interest were used, which may have led to overestimation of MBF in the infarcted areas.
Potential Mechanisms of Chronic LV Dysfunction
In our series of patients with previous myocardial infarction, both fibrotic and hibernating tissue may contribute in different proportions to the maintenance of chronic LV dysfunction. Using H215O, we have selectively measured MBF in water-perfusable tissue, excluding scar. Data showed that MBF was within the normal range in 90% of dysfunctional segments. This is consistent with the finding of Vanoverschelde et al,2 who found a preserved baseline MBF measured with PET and 13NH3 in a highly selected group of patients with coronary artery disease and chronic LV dysfunction but without previous infarction. This indicates that chronic LV dysfunction is not a consequence of a chronic reduction of baseline flow.
Our study does not provide any clues to the mechanisms responsible for chronic dysfunction. It has been suggested that dysfunction may be secondary to repeated episodes of ischemia and stunning.2 This hypothesis is in line with recent experimental work in conscious animals26 that demonstrated that “the reduced function during ameroid-induced coronary stenosis reflected cumulative myocardial stunning rather than a primary deficit in coronary blood flow… .” In patients with coronary artery disease, resting MBF remains normal despite increasing stenosis severity, whereas coronary vasodilator reserve is progressively reduced27 and myocardial ischemia may develop as a consequence of small increases in oxygen demand.28 On the other hand, it is also possible that recurrent myocardial injury can occur secondary to transient primary reductions in blood flow. In this case, plaque fissuring with the associated platelet-induced vasoconstriction and thrombosis could be the triggering events.29 In all cases, MBF measured outside the episodes would be normal. In light of this, coronary revascularization, by restoring coronary vasodilator reserve, would be of value in patients in whom stunning is secondary to ischemia owing to increased demand. Conversely, revascularization would be of limited value in those patients with recurrent plaque instability in whom alternative therapeutic approaches aiming at plaque stabilization should be evaluated. This could explain, at least in part, the variable functional response to revascularization observed in these patients.
MRG and Insulin Resistance
Traditionally, the assessment of myocardial viability with PET has involved the simultaneous semiquantitative evaluation of regional myocardial perfusion with 13NH3 and exogenous glucose uptake with 18FDG. Maintained 18FDG uptake in an area of reduced perfusion should identify viable tissue, whereas simultaneously reduced 13NH3 and 18FDG uptake should define irreversibly scarred tissue.3 Patients are generally given an oral glucose load before the PET study to promote the secretion of endogenous insulin, which in turn will stimulate glucose uptake by insulin-sensitive tissues.3 However, many patients undergoing viability studies may have other conditions known to be associated with insulin resistance.30 31 32 To circumvent this problem, 18FDG myocardial uptake can be measured during hyperinsulinemic euglycemic clamp.33 34 35 36 This involves the simultaneous infusion of insulin and glucose, acting as a metabolic stressor by promoting glucose uptake in insulin-sensitive tissues.8
With this approach, the second major finding of the present study was that MRG in the normally contracting segments in patients was 35% lower than that measured in the myocardium of normal subjects (Fig 2⇑). This was observed despite comparable values of circulating glucose and insulin achieved in patients and control subjects during clamp. This seems to be true even in the absence of other conditions known to be associated with insulin resistance. In fact, the difference between normal control subjects and patients was still highly significant after removal of the data from those patients who had diabetes or arterial hypertension, conditions known to be associated with insulin resistance. Insulin resistance has been demonstrated in the principal insulin-sensitive tissues (skeletal muscle and adipose tissue) in patients with different diseases, including diabetes,30 arterial hypertension,31 and coronary artery disease,32 but it is not clear whether myocardial tissue is equally resistant to the action of this hormone. In two different studies performed during hyperinsulinemic euglycemic clamp, myocardial MRG measured with PET in patients with type 1 diabetes was found to be similar to that in nondiabetic control subjects.33 34 In contrast, both MRG in skeletal muscle and whole body glucose uptake were significantly reduced in diabetic patients compared with normal control subjects.33 A similar study performed in patients with type 2 diabetes demonstrated that during clamp myocardial MRG was reduced by 39% compared with values in normal control subjects.35 Some differences in the absolute values of myocardial MRG exist between the above reports and the present study. It should be mentioned that in the present study a “lumped constant” value of one was used compared with a value of 0.67 in the above-mentioned reports. The value of 0.67 was obtained from studies in control animals,37 and it is not known whether this value is applicable to human myocardium. In addition, it is unknown whether the same lumped constant value can be used for normal and ischemic myocardium. Although the observation of a reduced MRG in normal myocardium of patients with coronary artery disease is original, a similar conclusion could be inferred from two separate articles from the same Finnish group.35 36 The MRG measured in a group of normal volunteers during euglycemic hyperinsulinemic clamp35 was 0.97 μmol · min−1 · g−1 using a lumped constant of 0.67, which translates to 0.65 μmol · min−1 · g−1 if the lumped constant is set to one as in the present study. This value is comparable to that measured in normal volunteers in the present study (0.71 μmol · min−1 · g−1). In the second article from the same group,36 MRG during euglycemic hyperinsulinemic clamp was measured in the normal myocardial segments of a group of patients with coronary artery disease. After conversion to a lumped constant of one, MRG was on average 0.49 μmol · min−1 · g−1, which compares well with the present value of 0.45 μmol · min−1 · g−1 measured in normal myocardial regions remote from dysfunctional segments.
Differences in the dietary state of the study population could account for differences in insulin sensitivities even during hyperinsulinemic euglycemic clamp.38 However, as stated in the “Methods” section, all patients and normal volunteers were scanned between 11 am and 1 pm after a light breakfast, providing a reasonable standardization of the dietary state.
The results of the present study indicate that on average the MRG measured during hyperinsulinemic euglycemic clamp is significantly higher in the segments that will improve after surgery compared with those that will not (Fig 3⇑). It should be noted that these differences in MRG could be explained, at least in part, by changes in the lumped constant caused by myocardial ischemia. However, this explanation probably is not applicable to the lower MRG found in the normal myocardial regions remote from dysfunctional segments.
It must be emphasized that the large range of MRG values measured in the dysfunctional territories makes it difficult to discriminate between individual segments that will or will not improve after revascularization. Part of this scatter might be due to the heterogeneity (eg, transmural differences of MBF and MRG) within the regions of interest. In this small series of patients, the quantitative analysis of MBF and MRG scans does not seem to provide additional diagnostic information relative to more qualitative assessments. However, it should be noted that the quantitative approach was essential in acquiring the new pathophysiological information.
Coronary Revascularization and LV Function
The data of the present study indicate that coronary revascularization improved regional wall motion in ≈50% of all dysfunctional segments, whereas baseline MBF measured with H215O was within normal limits in 90%. This may lead one to think that recovery of function has little to do with myocardial perfusion and even question the rationale for mechanical revascularization procedures. As previously outlined, the MBF measurements in the present study reflect flow in the viable tissue only. In a dysfunctional segment, recovery after revascularization will depend on the ratio between fibrotic and hypocontractile but viable tissue as demonstrated by De Silva et al.39 This was confirmed recently by Depré et al,40 who compared tissue morphology (intraoperative biopsies) with PET measurements of perfusion and metabolism using 13NH3 and 18FDG. They found that the amount of tissue fibrosis was the strongest determinant of recovery after revascularization and was inversely related to the uptake of 13NH3 and 18FDG. Indeed, the values of perfusion and metabolism obtained with these two tracers are expressed per unit mass of tissue and are sensitive to the amount of fibrosis within the region of interest. Accordingly, in the present study, the measurement of 18FDG was significantly lower in the dysfunctional regions that did not recover after revascularization.
The results of the present investigation have provided evidence that chronic LV dysfunction in patients with previous infarction is not associated with a reduced baseline MBF in most cases and that the myocardium of patients with coronary artery disease is less sensitive to insulin than the myocardial tissue of normal subjects. In agreement with previous studies,35 the use of the hyperinsulinemic euglycemic clamp allowed for high-quality 18FDG images in all subjects; this is particularly important in view of the high prevalence of insulin resistance demonstrated in the myocardium of these patients.
Selected Abbreviations and Acronyms
|C15O||=||15O-labeled carbon monoxide|
|C15O2||=||15O-labeled carbon dioxide|
|FWHM||=||full width at half-maximum|
|MBF||=||myocardial blood flow|
|MRG||=||metabolic rate of glucose|
|PET||=||positron emission tomography|
|TF||=||fraction of water-perfusable tissue|
This work was carried out in the framework of the European Economic Community Concerted Action on PET Investigation of Cellular Regeneration and Degeneration. We would like to express our gratitude to the radiochemists of the Medical Research Council Cyclotron Unit for the preparation of radiotracers and to Andrew Blythe and Andreanna Williams for their help in the acquisition of the positron emission tomography data. Special thanks go to Heather Boyd for her help and expertise in the data analysis and to Dr William Wijns and Prof Terry Jones for their support and critical review of the manuscript.
- Received August 7, 1995.
- Revision received September 26, 1995.
- Accepted October 4, 1995.
- Copyright © 1996 by American Heart Association
Rahimtoola SH. A perspective on the three large multicenter randomized clinical trials of coronary bypass surgery for chronic stable angina. Circulation. 1985;72(suppl V):V-123-V-135.
Vanoverschelde J-LJ, Wijns W, Depre C, Essamri B, Heyndrickx GR, Borgers M, Bol A, Melin J. Mechanisms of chronic regional postischemic dysfunction in humans: new insights from the study of noninfarcted collateral-dependent myocardium. Circulation. 1993;87:1513-1523.
Marwitz TH, McIntyre WJ, Lafont A, Nemec JJ, Salced EE. Metabolic response of hibernating and infarcted myocardium to revascularization: a follow-up study of regional perfusion, function, and metabolism. Circulation. 1992;85:1347-1353.
Gropler RJ, Getman EM, Sampathkumaran K, Perez JE, Schechtman KB, Conversano A, Sobel BE, Bergmann SR, Siegel BA. Comparison of carbon-11-acetate with fluorine-18-fluorodeoxyglucose for delineating viable myocardium by positron emission tomography. J Am Coll Cardiol. 1993;22:1587-1597.
De Fronzo RA, Jordan DT, Reubin A. Glucose clamp technique: a method for quantifying insulin secretion and resistance. Am J Physiol. 1979;237:E214-E223.
Braunwald E, ed. Heart Disease: A Textbook of Cardiovascular Medicine. Philadelphia, Pa: WB Saunders Co; 1988:268-270.
Spinks TJ, Jones T, Gilardi MC, Heather JD. Physical performance of the latest generation of commercial positron scanner. IEEE Trans Nucl Sci. 1988;NS-35:721-725.
Araujo LI, Lammertsma AA, Rhodes CG, McFallas EO, Iida H, Rechavia E, Gallassi AR, De Silva R, Jones T, Maseri A. Noninvasive quantification of regional myocardial blood flow in normal volunteers using oxygen-15 labeled carbon dioxide and positron emission tomography. Circulation. 1991;83:875-885.
Lammertsma AA, De Silva R, Araujo LI, Jones T. Measurement of regional myocardial blood flow using C15O2 and positron emission tomography: comparison of tracer models. Clin Phys Physiol Meas. 1992;13:1-20.
Bellina BC, Parodi O, Camici P, Salvadori PA, Taddei L, Fusani L, Guzzardi R, Klassen GA, L’Abbate A, Donato L. Simultaneous in vitro and in vivo validation of nitrogen-13 for the assessment of regional myocardial blood flow. J Nucl Med. 1990;31:1335-1343.
Czenin J, Muller P, Chan S, Brunken RC, Porenta G, Krivokapich J, Chen K, Chan A, Phelps ME, Schelbert HR. Influence of age and hemodynamics on myocardial blood flow and flow reserve. Circulation. 1993;88:62-69.
Vanoverschelde J-LJ, Melin JA, Bol A, Vanbutsele R, Cogneau M, Labar D, Robert A, Michel C, Wijns W. Regional oxidative metabolism in patients after recovery from reperfused anterior myocardial infarction: relation to regional blood flow and glucose uptake. Circulation. 1992;85:9-21.
Gewirtz H, Fischman AJ, Abraham S, Gilson M, Strauss HW, Alpert NM. Positron emission tomographic measurements of absolute regional myocardial blood flow permits identification of nonviable myocardium in patients with chronic myocardial infarction. J Am Coll Cardiol. 1994;23:851-859.
Shen YT, Vatner SF. Mechanism of impaired myocardial function during progressive coronary stenosis in conscious pigs: hibernation versus stunning? Circ Res. 1995;76;479-488.
Nuutila P, Knuuti J, Ruotsalainen U, Koivisto VA, Eronen E, Teras M, Bergman J, Haaparanta M, Voipio-Pulkki L-M, Viikari J, Ronnemaa T, Wegelius U, Yki-Jarvinen H. Insulin resistance is localized to skeletal but not heart muscle in type 1 diabetes. Am J Physiol. 1993;264:E756-E762.
Vom Dahl J, Herman WH, Hicks RJ, Ortiz-Alonso FJ, Lee KC, Allman KC, Wolfe ER, Kalff V, Schwaiger M. Myocardial glucose uptake in patients with insulin-dependent diabetes mellitus assessed quantitatively by dynamic positron emission tomography. Circulation. 1993;88:395-404.
Voipio-Pulkki L-M, Nuutila P, Knuuti J, Ruotsalainen U, Haaparanta M, Teras M, Wegelius U, Koivisto VA. Heart and skeletal muscle glucose disposal in type 2 diabetic patients as determined by positron emission tomography. J Nucl Med. 1993;34:2064-2067.
Knuuti MJ, Nuutila P, Ruotsalainen U, Saraste M, Harkonen R, Ahonen A, Teras M, Haaparanta M, Wegelius U, Haapanen A, Hartiala J, Voipio-Pulkki LM. Euglycemic hyperinsulinemic clamp and oral glucose load in stimulating myocardial glucose utilization during positron emission tomography. J Nucl Med. 1992;33:1255-1262.
Russel RR, Nguyen VTB, Mrus JM, Taegtmeyer H. Fasting and lactate unmask responsiveness in the isolated working rat heart. Am J Physiol. 1992;263:E556-E561.
De Silva R, Yamamoto Y, Rhodes CG, Iida H, Nihoyannopoulos P, Davies GJ, Lammertsma AA, Jones T, Maseri A. Preoperative prediction of the outcome of coronary revascularization using positron emission tomography. Circulation. 1192;86:1738-1742.
Depré C, Vanoverschelde JLJ, Melin JA, Borgers M, Bol A, Ausma J, Dion R, Wijns W. Structural and metabolic correlates of the reversibility of chronic left ventricular ischemic dysfunction in humans. Am J Physiol. 1995;268:H1265-H1275.