Myocardial Blood Flow Response to Pacing Tachycardia and to Dipyridamole Infusion in Patients With Dilated Cardiomyopathy Without Overt Heart Failure
A Quantitative Assessment by Positron Emission Tomography
Background Myocardial blood flow (MBF) impairment has been documented in advanced dilated cardiomyopathy (DCM) in which hemodynamic factors, secondary to severe ventricular dysfunction, may limit myocardial perfusion. To assess whether MBF impairment in DCM may also be present independent of hemodynamic factors, the present study was designed to quantify myocardial perfusion in patients with mild disease without overt heart failure.
Methods and Results Absolute regional MBF (milliliters per minute per gram) was measured by positron emission tomography and 13N-ammonia in resting conditions, during pacing-induced tachycardia, and after dipyridamole infusion (0.56 mg/kg over 4 minutes) in 22 DCM patients and in 13 healthy subjects. Patients were in New York Heart Association functional class I-II and showed depressed left ventricular (LV) ejection fraction by radionuclide angiography (35±8%; range, 21% to 48%), normal coronary angiography, and normal or moderately increased LV end-diastolic pressure (9.2±5.5 mm Hg; range, 2 to 20 mm Hg). There were no differences in arterial blood pressure, heart rate, and rate-pressure product between patients and control subjects in the three study conditions. Compared with control subjects, DCM patients had lower mean MBF at rest (0.80±0.25 versus 1.08±0.20 mL · min−1 · g−1, P<.01), during atrial pacing tachycardia (1.21±0.59 versus 2.03±0.64 mL · min−1 · g−1, P<.01), and after dipyridamole infusion (1.91±0.76 versus 3.78±0.86 mL · min−1 · g−1, P<.01). LV MBF values were related to baseline LV end-diastolic pressure at rest (r=−.57, P<.01) and during pacing (r=−.67, P<.01) but not after dipyridamole infusion (r=.19, P=.40). Five patients had LV end-diastolic pressure >12 mm Hg; in 4, myocardial perfusion was severely depressed both at baseline and in response to stress.
Conclusions In patients with DCM without overt heart failure, myocardial perfusion is impaired both at rest and in response to vasodilating stimuli. The abnormalities in vasodilating capability can be present despite normal hemodynamics; progression of the disease is associated with more depressed myocardial perfusion.
Myocardial blood flow (MBF) abnormalities, despite the presence of angiographically normal coronary arteries, were documented in patients with heart failure caused by idiopathic dilated cardiomyopathy (DCM). In particular, a reduction in MBF at rest1 2 3 or in response to either metabolic2 or pharmacological vasodilating stimuli4 5 6 7 8 9 was reported. The proposed pathogenic mechanisms include factors secondary to the ventricular dysfunction or alterations of the small coronary vessels associated with the disease process.
Studies so far included patients with advanced disease characterized by severe ventricular dysfunction, usually associated with overt heart failure. In these patients, impaired ventricular hemodynamics and major cardiovascular changes secondary to heart failure might have had a predominant role in affecting myocardial perfusion, overwhelming possible abnormalities at the microvascular level. Evaluation of MBF at an early stage of the disease might have provided a definite clue to an understanding of the relation between flow and function in DCM, but the difficulty of an early diagnosis has so far prevented this. Another major limitation of previous studies was the assessment of MBF by invasive techniques (coronary sinus thermodilution, xenon washout curves, and intracoronary Doppler catheter), which are more able to evaluate relative flow changes during interventions than to measure absolute differences in regional MBF values among different patients.
The present study was designed to assess the presence of MBF impairment in a selected population of DCM patients without heart failure in whom the possible influence of secondary hemodynamic factors in limiting myocardial perfusion was minimized. All patients had mild ventricular dysfunction without clinical and hemodynamic evidence of heart failure. The study population was selected by means of heart catheterization after a careful noninvasive screening of patients with suspected myocardial disease,10 presenting with ventricular arrhythmias, conduction defects, or chest pain, after exclusion of other cardiac or systemic disorders.
Absolute MBF was measured by dynamic positron emission tomography (PET) and 13N-ammonia as a flow tracer. This method provides a quantification of regional specific MBF (milliliters per minute per gram) and allows direct comparison of flow values of different patients in different conditions. The protocol included quantification of MBF at baseline, during pacing-induced tachycardia, and after dipyridamole infusion in DCM patients and in a control population of healthy subjects.
From January 1989 to September 1993, at the Cardiological Department of the Institute of Clinical Physiology, National Council of Research, 55 of 1465 patients undergoing coronary angiography had a diagnosis of DCM. They were diagnosed from 362 patients with angiographically normal coronary arteries on the basis of a demonstration of abnormal left ventricular (LV) function by means of contrast ventriculography. The diagnosis was made after exclusion of other systemic and cardiological disorders, including vasospastic ischemia (n=73), valvular or hypertensive heart disease (n=106), hypertrophic cardiomyopathy (n=13), acute (histologically proven) myocarditis (n=3), and acute pericarditis (n=2). Of the 55 DCM patients, 24 had been referred because of symptoms of heart failure (New York Heart Association [NYHA] functional class III); the remaining 31 patients (NYHA class I-II) had been referred for heart catheterization because of myocardial disease suspected from the presence of ECG abnormalities or atypical chest pain associated with abnormal LV function at preliminary echocardiographic evaluation.
Of the 31 patients with DCM and no heart failure, 22 were enrolled in this study, which included a new echocardiogram, baseline functional evaluation by equilibrium radionuclide angiography, and MBF assessment by PET. These patients satisfied the following inclusion criteria: (1) ECG abnormalities (complex ventricular arrhythmias, left bundle-branch block) or chest pain, (2) sinus rhythm (to permit equilibrium radionuclide angiography and rapid atrial pacing), (3) abnormal LV function by contrast ventriculography (regional dyssynergies or dilation), (4) reduced LV ejection fraction (<50%) by equilibrium radionuclide angiography, (5) exclusion of other cardiac and systemic disorders, and (6) informed written consent. Of the 9 patients excluded, 2 had neoplastic disease, 1 had a history of recent stroke, and 4 did not give their consent; in 2 others, sustained elevated pressure values were first recorded shortly before the PET study.
Data obtained in this population were compared with those of 13 subjects (5 men, 8 women; mean age, 48 years; range, 37 to 57 years) with atypical chest pain, angiographically normal epicardial coronary arteries, normal LV function by contrast ventriculography, and no other detectable heart or systemic disease. They also gave informed written consent to the study, which was approved by the local Ethics Committee for Human Research.
Echocardiography and Equilibrium Radionuclide Angiography
Seventeen patients underwent a new echocardiographic evaluation within 7 days of heart catheterization. Echocardiograms were recorded (Hewlett Packard, Sonos 1500, Andover) by a 2.5-MHz probe during noninvasive blood pressure monitoring by cuff manometer. LV ID, wall thickness, and percent fractional shortening were obtained from M-mode tracings in the plane of the mitral valve under two-dimensional control according to standardized protocol.11 LV systolic wall stress was determined with the equation ς=[0.334 P (LVID)]/[PWT (1+ PWT/LVID)], where P is the mean of duplicate cuff arterial systolic pressures, LVID is the LV systolic (smallest) ID, and PWT is the systolic posterior wall thickness at the time of the smallest LVID.12 Data were analyzed as the mean of three to five cardiac cycles.
Equilibrium radionuclide angiography was performed on all patients within 7 days of heart catheterization. Gated blood pool images were obtained by in vivo labeling of red blood cells with 0.03 mg/kg stannous agent (stannous chloride, Amersham) followed, 30 minutes later, by 925 MBq (25 mCi) of technetium-99m pertechnetate. The scintigraphic data were collected by a small-field mobile gamma camera (Apex 410M, Elscint) equipped with a high-resolution parallel hole collimator. Data acquisition was performed at baseline in the left anterior oblique “best septal” view; a minimum of 100 000 counts per frame was collected. Regional wall motion was assessed by consensus of two observers for five LV regions (proximal and distal septal walls, inferoapical wall, and proximal and distal posterolateral walls) and two right ventricular (RV) regions (anterolateral wall and inferoapical wall). Normal wall motion, hypokinesis, marked hypokinesis, akinesis, and dyskinesis were scored 0, 1, 2, 3, and 4, respectively. The scores of all segments were then summed to obtain a wall motion abnormality index for each ventricle. LV and RV ejection fraction values were obtained according to methods previously described.13 14 According to the results obtained in our laboratory in a large group of healthy subjects, the cutoff values for abnormal LV and RV function were as follows: wall motion abnormality index >1, LV ejection fraction <50%, and RV ejection fraction <35%. LV volumes were computed by a count-based method.15
Contrast LV Angiography and Myocardial Biopsy
Patients underwent catheterization of the right and left sides of the heart, left ventriculography, and coronary angiography; measurements of pressures in the right and left heart chambers and cardiac index were obtained by thermodilution technique. All patients had angiographically normal coronary arteries and regional or global LV dysfunction by ventriculography.
LV endomyocardial biopsy is not performed on a routine basis in our laboratory on patients with suspected DCM. However, to rule out myocarditis, 12 patients of this population underwent the procedure because they were younger than 40 years of age or experienced very recent onset of symptoms or ECG changes. The tip of a 6F King bioptome was positioned, under fluoroscopic control, on the LV septum close to the inferior wall. Four or five samples, 3 mm in diameter, were obtained from each patient and fixed in phosphate-buffered 10% formalin. From the paraffin-embedded tissue, four or five 4- to 5-μm-thick slices were obtained and stained by Masson’s trichrome, hematoxylin and eosin, and van Gieson’s stains. Two experienced cardiac pathologists examined all microscopic sections for the presence of myocyte hypertrophy and interstitial or microfocal fibrosis, graded as mild, moderate, or severe. Myocarditis was excluded according to the Dallas criteria.16
Characteristics of the Study Population
Of the 22 patients studied, 16 were men and 6 were women with a mean age of 47 years. According to the inclusion criteria, no patient had overt heart failure (NYHA class I-II). The main clinical and ECG features included complex ventricular arrhythmias documented at 24-hour Holter monitoring (n=16), persistent or rate-related left bundle-branch block (n=10), chest pain (n=13), and syncope (n=4). All subjects had shown regional or global LV dysfunction at preliminary echocardiographic evaluation. Only 2 patients had no ventricular arrhythmias or conduction defects: a 40-year-old man referred for chest pain and abnormal LV wall motion at echocardiography and a 38-year-old woman in whom echocardiographic diagnosis of LV dysfunction had been made 3 months after her third delivery. None described alcohol abuse. Routine blood chemistry screening, hematologic profile, rheumatologic tests, thyroid function tests, serum iron studies, plasma, and urinary electrolytes were found to be in the normal range in all patients. In particular, no patient was hypoxic, polycytemic, or anemic. Repeated blood pressure measurements were obtained during hospitalization; in no case were pressure values higher than 150/90 mm Hg or had a history of hypertension been reported. Twelve patients had never been on active cardiac medication; 7 were on antiarrhythmic therapy; 3 were on digitalis, diuretics, and angiotensin-converting-enzyme inhibitors at the time of referral. No patient was taking amiodarone or β-blockers. All cardioactive drugs were withdrawn at least 7 days before cardiac catheterization and echocardiographic and scintigraphic examination.
The new echocardiographic evaluation confirmed regional and/or global LV dysfunction in all patients. LV fractional shortening was 24±8% (range, 11% to 35%), and LV meridional systolic wall stress was 115±35×103 dynes/cm2 (range, 45 to 176×103 dynes/cm2).
Equilibrium radionuclide angiography showed LV dyssynergies and abnormal LV ejection fraction in all patients and RV dyssynergies and abnormal RV ejection fraction in 4 and 2 patients, respectively. Mean LV wall motion abnormality index was 7.0±3.5 U (range, 2 to 15 U), LV ejection fraction was 35±8% (range, 21% to 48%), LV end-diastolic volume was 178±60 mL (range, 90 to 302 mL). Mean RV wall motion abnormality index and ejection fraction were 0.8±0.9 U and 44±9%, respectively.
Catheterization of the right side of the heart showed abnormal ventricular and pulmonary artery pressures in 2 patients and reduced cardiac index in 1 patient. Catheterization of the left side of the heart demonstrated normal aortic pressure in all patients and increased LV end-diastolic pressure (>12 mm Hg) in 5 patients. LV endomyocardial biopsy, performed on 12 patients, showed moderate to severe myocyte hypertrophy or interstitial and subendocardial fibrosis in 9. Mononuclear cell infiltrates were detected in 5 cases, but none had a diagnosis of active myocarditis.16 The arteriolar vessels (70- to 20-μm diameter) were properly visualized in 4 cases and showed no increase in the wall-lumen ratio.
Table 1⇓ summarizes the individual clinical, scintigraphic, hemodynamic, and bioptic data in DCM patients.
Positron Emission Tomography
All patients and control subjects were studied in the absence of any medical therapy after an overnight fasting period; caffeine, theophylline, and theophylline derivatives were withdrawn 24 hours before imaging. A bipolar pacing catheter was percutaneously advanced into the right atrium under fluoroscopy and continuous electrocardiographic monitoring; patients were then transferred to the positron emission tomography room and positioned on the bed of a 2-ring ECAT III positron tomograph (CTI Inc) that provides three simultaneous cross-sectional planes. Transmission images were acquired up to the collection of 60 million counts with a 68Ge source and subsequently used to generate attenuation correction factors. Correct positioning was maintained throughout the study with the use of the light beam and marks on the subject’s torso. Thereafter, 7.4 MBq/kg body wt (0.2 mCi/kg) of 13N-ammonia (physical half-life, 9.9 minutes) was infused over a 10- to 20-second period into the left antecubital vein. Dynamic acquisition was started simultaneously with tracer injection; 28 frames were acquired over 8 minutes (16 frames×3 seconds, 11×12 seconds, and 1×300 seconds). Fifty minutes after the baseline study, heart rate was increased by use of an external pacemaker connected to the bipolar catheter, starting from 10 beats per minute (bpm) over the patient’s heart rate, with 20-beat increments every minute to a maximum of 150 bpm or until the Wenckebach point was reached. Intravenous atropine was not administered to avoid effects of this drug on coronary blood flow. At this stage, heart rate was kept constant, 13N-ammonia was injected, and dynamic acquisition was started according to the same protocol as for the study at rest; 3 minutes after the injection, the heart rate was lowered, and the pacemaker was switched off within 2 minutes. Fifty minutes later, dipyridamole (0.56 mg/kg body weight) was infused intravenously over 4 minutes; 3 minutes after the end of the infusion, 13N-ammonia was injected. Aminophylline (120 to 240 mg) was always injected intravenously ≥3 minutes after 13N-ammonia injection to antagonize the effects of dipyridamole.
A three-lead ECG was continuously monitored, while a nine-lead ECG and arterial blood pressure by cuff manometer were recorded every minute during PET acquisitions. The PET study during pacing was performed in 19 of 22 patients and in 8 of 13 control subjects because of technical difficulties in positioning the stimulating catheter in 3 cases and the refusal of the invasive procedure in the remaining subjects. One patient, who had shown intolerance to dipyridamole, did not undergo the pharmacological study. No side effects or complications occurred during either atrial pacing or the dipyridamole test in any patient.
One experienced cardiologist unaware of the clinical findings computed the regional MBF according to a previously validated method.17 Briefly, a small region of interest was drawn within the LV cavity in the last 300-second equilibrium image, and the time activity curve of 13N in the arterial blood was computed. Data values were corrected for decay and dead-time loss. In the last “equilibrium” frame, six LV circular regions of interest (size, 13 to 22 pixels) were manually drawn (two in the posterolateral wall, two in the anterior wall, and two in the septal wall), and a dedicated program was used to perform automatic edge detection of the LV wall to measure myocardial thickness and to correct for the partial volume effect. Blood flow times 13N-ammonia extraction was then calculated in each region of interest as MBFe=Cm×60/∫Cb(t)dt, where Cm and Cb are 13N activity concentrations (counts per second per voxel) in the myocardium (obtained in the last scan from 3 to 8 minutes) and in the arterial blood at time t, respectively. The Cb(t) curve was fitted by a γ-variate function for integration. Regional MBFe values were corrected for the recovery coefficient and divided by tissue gravity (1.08 g/mL).
Correction for the recovery coefficient was performed with 13N activity concentrations in the myocardium and myocardial thickness data simultaneously measured in the last equilibrium scan (from 3 to 8 minutes, when the pacing rate was reduced and dipyridamole was antagonized by aminophylline) to avoid the effects of tachycardia or dipyridamole infusion on LV wall thickness and make MBF values comparable in different conditions. To compensate for the decrease in ammonia extraction with increase in flow rates, MBFe data were corrected on the basis of the experimental relation between ammonia uptake and microsphere-determined MBF as obtained in animal studies17 : MBF=exp[(MBFe+0.04)/1.45]−1.
Regional MBF values were averaged for each of the three LV walls; the average of values in all six original regions of interest gave the mean MBF in the left ventricle. Regional perfusion defects were defined by comparison of flow ratios in different LV myocardial regions in control subjects and patients. In particular, the septal-posterolateral flow ratio in controls showed the wider range either at baseline (1.16±0.14; range, 0.93 to 1.38), during pacing (1.14±0.18; range, 0.90 to 1.51), and after dipyridamole infusion (1.29±0.27; range, 0.69 to 1.71) and was taken as reference. Accordingly, a flow defect was detected in patients when the regional flow ratios were below or above the mean septal/posterolateral flow ratio in control subjects ±2 SD. Coronary flow reserve was calculated for both the three LV walls and the whole left ventricle as the ratio between MBF values after dipyridamole infusion and MBF values under resting conditions.
All data are presented as mean±SD. Two-tailed ANOVA, followed by the Scheffé F test, was used for multiple comparisons among groups. Student’s t test for paired samples was used to compare flow values in the same patients in different study conditions. Simple regression analysis was used to correlate MBF values with determinants of ventricular function. Values of P<.05 were considered significant.
During the PET study, ST-T changes or chest pain was documented in 7 of 19 patients during pacing and in 8 of 21 patients during dipyridamole infusion. Subjects in the control group had neither ECG changes nor chest pain during both stresses.
In each study condition, no significant difference in mean arterial pressure, heart rate, and rate-pressure product was found between control subjects and DCM patients (Table 2⇓).
Baseline MBF and Response to Stress
In the control group, mean LV MBF was 1.08±0.20 mL · min−1 · g−1 at baseline and significantly increased during pacing (to 2.03±0.64 mL · min−1 · g−1, P<.01) and after dipyridamole infusion (to 3.78±0.86 mL · min−1 · g−1, P<.01). In the DCM group, mean LV MBF was 0.80±0.25 mL · min−1 · g−1 at baseline and significantly increased during pacing (to 1.21±0.59 mL · min−1 · g−1, P<.01) and after dipyridamole infusion (to 1.91±0.76 mL · min−1 · g−1, P<.01). In all conditions, flow values were significantly reduced in DCM patients compared with control subjects (P<.01) (Fig 1⇓). Mean coronary flow reserve was significantly lower in DCM patients than in control subjects (2.45±0.71 versus 3.59±0.95, P<.01).
In DCM patients, mean LV MBF was correlated with LV end-diastolic pressure at baseline (r=−.57, P<.01) and during pacing (r=−.67, P<.01) but not after dipyridamole infusion (r=.19, P=.40). There was no correlation between MBF values and LV ejection fraction, LV systolic wall stress, or rate pressure products (Table 3⇓).
Regional MBF and regional coronary flow reserve were significantly reduced in DCM patients compared with control subjects in all the myocardial walls explored (Table 4⇓).
Behavior of MBF in Individual Patients
Table 2⇑ lists the individual mean MBF values in each study condition for DCM patients and control subjects. The MBF values in the control group ranged from 0.81 to 1.47 mL · min−1 · g−1 at rest, from 1.53 to 3.41 mL · min−1 · g−1 during pacing and from 2.34 to 5.66 mL · min−1 · g−1 after dipyridamole infusion. In the DCM group, MBF values were depressed below the range of control subjects at rest in 12 of 22 patients (55%), during pacing in 14 of 19 patients (74%), and after dipyridamole infusion in 15 of 21 patients (71%). Mean MBF was depressed in all the study conditions in 10 patients and was completely normal in 4 cases. Of 22 DCM patients, 18 (82%) had depressed MBF in at least one study condition. Fig 2⇓ shows the individual MBF values at baseline and in response to vasodilatory stimuli for control subjects and patients.
The individual flow behavior was compared with functional and histological data. Among the 17 patients with normal LV end-diastolic pressure, 13 had a depressed MBF in at least one study condition. Of the 5 patients with LV end-diastolic pressure >12 mm Hg, 4 had severely depressed perfusion both at baseline and in response to stress. Among the 7 patients with mild fibrosis at endomyocardial biopsy, 5 had a depressed MBF in at least one study condition. A severely depressed myocardial perfusion both at rest and in response to stress was documented in 4 of the 5 patients with moderate to severe fibrosis.
Regional flow defects, as evaluated by regional flow ratios, were present in 7 of 22 patients (32%) at baseline, in 5 of 19 patients (26%) during pacing, and in 4 of 21 patients (19%) after dipyridamole infusion. Overall, 10 of 22 patients (45%) showed a regional perfusion defect at least in one study condition. Interestingly, 2 of the 4 patients with completely normal mean MBF values showed stress-induced regional perfusion defects.
Fig 3⇓ shows representative PET scans of 2 patients with different degrees of flow impairment.
The major finding of the present study was an obvious LV MBF impairment in patients with DCM before clinical and hemodynamic evidence of heart failure. MBF values were reduced at baseline or in response to stress in most of the patients studied (82%) and even in the presence of completely normal central hemodynamics. However, the extent of MBF impairment in this population also appeared to be related to factors correlated with the severity of the disease. Higher LV end-diastolic pressure values and moderate to severe fibrosis at endomyocardial biopsy were found in those subjects with more severely depressed perfusion, although no correlation between coronary flow and other determinants of ventricular function was found.
These results demonstrate that myocardial perfusion may be abnormal in DCM patients even before the occurrence of hemodynamic changes that can further limit blood flow delivery to the myocardium; moreover, the noninvasive quantification of flow distribution in these patients may give valuable information about the severity of the underlying disease process.
Myocardial Perfusion in DCM
Previous studies on myocardial perfusion in DCM clearly showed a limitation in coronary flow reserve. A reduced vasodilatory response to dipyridamole infusion4 5 6 and, more recently, a blunted flow reserve in response to acetylcholine7 9 and to papaverine8 were reported, while the response to pacing tachycardia was more controversial.1 2 Importantly, all previous studies focused on patients with advanced disease with clinical and hemodynamic evidence of heart failure and LV enlargement. At this stage, resting myocardial perfusion and flow response to stress might be blunted as a consequence of multiple factors, including hemodynamic impairment with increased extravascular component of coronary resistance and reduced driving pressure4 8 and structural or functional abnormalities of the coronary small vessels associated with the disease process.6 7 9
The present study demonstrated an obvious myocardial hypoperfusion both at rest and in response to stress in a population of patients with mild DCM. Previous studies in advanced DCM stressed the role of impaired ventricular function in causing a reduction of coronary flow reserve as a result of extrinsic compression of the coronary bed by stressed myocardium.4 8 In a recent study in patients with end-stage heart failure undergoing heart transplantation, however, the actual role of increased extravascular resistance in limiting coronary flow was questioned.3 A preserved physiological transmural distribution of MBF was demonstrated in the explanted hearts in the presence of a severely depressed baseline perfusion; if increased LV end-diastolic pressure were to have a major role in limiting flow, a pronounced decrease in subendocardial to subepicardial flow ratio would have been expected.18 19 The above-mentioned mechanism, secondary to severe ventricular dysfunction, does not seem to be a relevant cause of perfusion abnormalities in the patients with mild DCM in the present study. In this population, no correlation was found between coronary flow values and major determinants of ventricular function such as LV ejection fraction, LV wall stress, and rate-pressure products. Similarly, MBF values were not related to LV end-diastolic pressure during pharmacological vasodilation by dipyridamole infusion, when the effects of an extravascular compression of the coronary bed should have been evident.20
In dilated and hypertrophic cardiomyopathy, Weiss et al1 reported a decrease in resting MBF per unit mass that was apparently related to determinants of myocardial metabolic demand such as heart rate, systolic wall stress, and LV performance. In the present study, MBF values at baseline and during pacing were significantly correlated with LV end-diastolic pressure, suggesting a close match between flow and function at rest and during metabolic stimulation, ie, in the presence of autoregulation. The relative role of a depressed inotropic state and lower oxygen demand in regulating MBF supply21 or, conversely, a decreased coronary blood flow in causing myocardial dysfunction22 could not be investigated in these patients with DCM. The lack of an accurate assessment of myocardial contractility and of a direct evaluation of oxygen consumption made it difficult to establish a reliable correlation between flow delivery and metabolic demand. Contractile differences might have affected oxygen consumption variably so that a down-resetting of MBF owing to decreased metabolic demand cannot be ruled out in some patients. As a matter of fact, however, clearly different flow patterns were not mirrored by differences in rate-pressure products.
Aging may cause a gradual decline of MBF reserve because of an age-related increase of baseline myocardial work and blood flow.23 In this study, the age range in the DCM patients was quite wide (29 to 73 years), but age-related changes in flow are unlikely to have influenced the results. PET permitted a quantitative assessment of MBF, demonstrating that coronary flow reserve was reduced in DCM patients because of a concurrent reduction of both baseline and maximal flow values. Moreover, there was no correlation between MBF values and the age of the patients or rate-pressure product values in each condition.
Myocardial structural abnormalities, already present in patients with subclinical myocardial disease,24 25 might involve the microvascular bed in causing a limitation of MBF even in this population. In the present study, consistent with most bioptic and pathological studies in DCM,3 4 26 27 28 biopsy showed myocyte hypertrophy and interstitial or microfocal fibrosis of variable extent in patients with mild disease. Because 4 of the 6 patients with moderate to severe fibrosis at biopsy had a severely depressed myocardial perfusion, it is conceivable that, at least in some cases, increased fibrosis per gram of tissue sampled by PET might have contributed to a decrease in resting and maximal absolute MBF values. Conversely, several patients with nearly normal histology or mildly impaired LV function still had depressed MBF values, suggesting that structural abnormalities associated with the disease process are not the unique factor involved in the pathogenesis of abnormal myocardial perfusion. In explanted hearts from patients with end-stage disease, no correlation was found between the degree of MBF impairment and the extent of myocardial fibrosis as assessed by both histology and biochemistry.3 Furthermore, according to pathological studies in DCM patients, endomyocardial fibrosis from bioptic samples may overestimate the actual extent of fibrosis that does not exceed 20% of the whole myocardium and involves mainly the subendocardial layers.26 27 28
It has been suggested that functional small-vessel abnormalities caused by increased “tone” or endothelial dysfunction might account for the reduced coronary flow reserve in DCM.6 7 29 The results of the present study, in particular the demonstration of decreased MBF response to dipyridamole infusion independent of hemodynamic impairment, support the hypothesis that a functional alteration of the coronary small vessels may be one of the pathogenetic mechanisms leading to increased minimal coronary resistance and impaired myocardial perfusion in this disease.
Usefulness of MBF Measurement in DCM
The present study might have clinical implications for the diagnostic approach to patients with DCM. Globally depressed ventricular function and angiographically normal epicardial coronary arteries allow identification of patients with DCM. In some patients with electric abnormalities or chest pain, however, in the absence of symptoms and signs of heart failure, diagnosis of DCM might be doubtful, particularly when the ventricular function is only slightly impaired.10 24 25 30 The present results demonstrate that as many as 87% of patients with mild DCM without overt heart failure, show a clear reduction in absolute MBF either at rest or in response to stress when evaluated by PET. Because a globally depressed myocardial perfusion and a blunted flow response to stress is part of the disease process in advanced DCM,1 2 3 4 5 6 7 8 9 similar findings in patients with a doubtful diagnosis would suggest the presence of an underlying myocardial disease. The PET technique is best suited to recognizing these abnormalities, given its unique ability to quantify absolute MBF in different conditions. Moreover, because the extent of MBF impairment appeared to be related to factors correlated with the degree of myocardial functional and structural involvement, noninvasive quantification of flow impairment in these patients may give valuable information about the severity of the underlying disease process.
Although there is a high incidence of globally depressed MBF, regional perfusion defects are also present in 45% of patients with DCM without overt heart failure. It is well known that greater homogeneity of MBF distribution is a distinctive feature of DCM with respect to ischemic cardiomyopathy,31 although segmental flow defects were also described in DCM patients.3 32 33 The present findings show that DCM may also be associated with regional perfusion defects in patients with less advanced disease. It is conceivable that regional myocardial perfusion abnormalities may precede a diffuse flow impairment. As a matter of fact, 2 of the 4 patients of the present series with completely normal mean MBF values in all the study conditions showed regional perfusion defects during stress.
A thorough evaluation of the role of extravascular variables in limiting flow response to stress would have required continuous invasive hemodynamic monitoring during the PET study. This approach was not feasible in the present study for both ethical and organizational reasons. However, baseline LV end-diastolic pressure was normal in most patients and, in the absence of ischemia, should not change during dipyridamole infusion,34 although a mild hemodynamic impairment during atrial pacing tachycardia cannot be completely ruled out.
Some limitations of the PET method should also be mentioned. First, the uptake of 13N-ammonia in the myocardium is dependent on both coronary flow and metabolism; metabolic changes, possibly occurring in DCM patients, might have limited 13N-ammonia uptake independently of myocardial perfusion. However, such a possibility was observed only under severe ischemia and at very low intracellular pH,35 36 which was not the case in the present study. Second, a small difference in MBF was observed among the LV walls in healthy subjects both at rest and during stress, with the highest flow values in the septum, which is consistent with previous studies.37 In the present study, the MBF values for the same LV region were compared among different groups, and the evaluation of the homogeneity of myocardial perfusion was based on the comparison of regional flow ratios between patients and control subjects. The decrease in 13N-ammonia extraction with increasing flow rates requires a correction to obtain a reliable quantification of MBF during dipyridamole infusion. Such correction was performed according to previous results obtained experimentally in dogs.17 Flow differences occurring at baseline were affected little by this phenomenon, whereas flow differences during pacing or dipyridamole infusion were still obvious and significant even when flow time extraction values were considered.
This is the first study in which an obvious decrease of MBF at rest and in response to vasodilating stimuli has been documented in patients with mild LV dysfunction caused by DCM. The demonstration of globally or regionally altered myocardial perfusion in this population suggests that a careful evaluation and a prolonged follow-up should be carried out in these patients. The population of the present study is undergoing a follow-up that is still ongoing. At present, after a mean of 24±14 months from the PET evaluation, 2 patients died suddenly (patients 2 and 16), and 2 other patients developed overt heart failure and were given transplants (patients 7 and 8).
The early recognition of patients who are at risk of developing heart failure and the development of therapeutic strategies targeting specific pathophysiological processes might have considerable impact on the prognosis of DCM. Prompt therapeutic intervention is probably warranted when signs of LV functional impairment appear. Moreover, it is conceivable that early treatment with drugs that can influence microvascular vasodilating properties, along with conventional treatment, would be especially useful in selected patients with documented MBF impairment. Although PET appears to be the most suitable tool for identifying depressed MBF in these patients, the cost-benefit ratio of this approach should be evaluated further and balanced against the social cost of advanced heart failure to which the actual contribution of unrecognized forms of myocardial diseases is currently unknown.
This study was supported in part by the Prevention and Control of Disease Factors project and the Control of Cardiovascular Disease subproject, the National Research Council, Rome, Italy. We thank Nicola Nista, Oreste Sorace, and Monica Bartoli for their technical assistance. We are indebted to Ilaria Citti for expert secretarial work and to Antonio Caselli for his help in the preparation of this manuscript.
Reprint requests to Dr Oberdan Parodi, CNR Institute of Clinical Physiology, Via P. Savi, 8-56100 Pisa, Italy.
This study was presented in part at the 43rd Annual Scientific Session of the American College of Cardiology, Atlanta, Ga, March 13-17, 1994.
- Received January 5, 1995.
- Revision received February 15, 1995.
- Accepted February 21, 1995.
- Copyright © 1995 by American Heart Association
Weiss MB, Ellis K, Sciacca RR, Johnson LL, Schmidt DH, Cannon PJ. Myocardial blood flow in congestive and hypertrophic cardiomyopathy: relationship to peak wall stress and mean velocity of circumferential fiber shortening. Circulation. 1976;54:484-493.
Pasternac A, Noble J, Streulens Y, Elie R, Henschke C, Bourassa MG. Pathophysiology of chest pain in patients with cardiomyopathies and normal coronary arteries. Circulation. 1982;65:778-789.
Parodi O, De Maria R, Oltrona L, Testa R, Sambuceti G, Roghi A, Merli M, Belingheri L, Accinni R, Spinelli F, Pellegrini A, Baroldi G. Myocardial blood flow distribution in patients with ischemic heart disease or dilated cardiomyopathy undergoing heart transplantation. Circulation. 1993;88:509-522.
Treasure CB, Vita JA, Cox DA, Fish RD, Gordon JB, Mudge GH, Colucci WS, Sutton MG, Selwyn AP, Alexander RW, Ganz P. Endothelium-dependent dilation of the coronary microvascolature is impaired in dilated cardiomyopathy. Circulation. 1990;81:772-779.
Keren A, Popp RL. Assignment of patients into the classification of cardiomyopathies. Circulation. 1992;86:1622-1633.
Sahn DJ, DeMaria A, Kisslo J, Weyman A, for the Committee on M-Mode Standardization of the American Society of Echocardiography. Recommendations regarding quantification in M-mode echocardiography: results of a survey of echocardiographic measurements. Circulation. 1978;58:1072-1083.
Reichek N, Wilson J, Sutton MJ, Plappert TA, Goldberg S, Hirshfeld JW. Noninvasive determination of left ventricular end-systolic stress: validation of the method and initial application. Circulation. 1982;65:99-108.
Schelbert HR, Verba JW, Johnson AD, Brock GW, Alazraki NP, Rose FJ, Ashburn WL. Nontraumatic determination of left ventricular ejection fraction by radionuclide angiocardiography. Circulation. 1975;51:902-909.
Maddahi J, Berman DS, Matsuoka DT, Waxman AD, Stankus KE, Forrester JS, Swan HJC. A new technique for assessing right ventricular ejection fraction using rapid multiple-gated equilibrium cardiac blood pool scintigraphy. Circulation. 1979;60:581-589.
Aretz HT, Billingham ME, Edwards WD, Factor SM, Fallon JT, Fenoglio JJ, Olsen EG, Schoen FJ. Myocarditis: a histopathologic definition and classification. Am J Cardiovasc Pathol. 1986;1:3-14.
Bellina RC, 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-ammonia for the assessment of regional myocardial blood flow. J Nucl Med. 1990;31:1335-1343.
Rouleau J, Boerboom LE, Surjadhana A, Hoffman JIE. The role of autoregulation and tissue diastolic pressures in the transmural distribution of left ventricular blood flow in anesthetized dogs. Circ Res. 1979;45:804-815.
Chilian WM, Layne SM, Klausner EC, Eastham CL, Marcus ML. Redistribution of coronary microvascular resistance produced by dipyridamole. Am J Physiol. 1989;256:H383-H390.
Ross J Jr. Myocardial perfusion-contraction matching: implications for coronary heart disease and hibernation. Circulation. 1991;83:1076-1083.
Czernin J, Müller P, Chan S, Brunken RC, Porenta G, Kripokapich 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.
Strain JE, Grose RM, Stephen MF, Fisher JD. Results of endomyocardial biopsy in patients with spontaneous ventricular tachycardia but without apparent structural heart disease. Circulation. 1983;68:1171-1181.
Olsen EGJ. Pathology of primary cardiomyopathies. Postgrad Med J. 1972;48:732-737.
Unverferth DV, Baker PB, Swift SE, Chaffee R, Fetters JK, Uretsky BF, Thompson ME, Leier CV. Extent of myocardial fibrosis and cellular hypertrophy in dilated cardiomyopathy. Am J Cardiol.1986;57:816-820.
Yonesaka S, Becker AE. Dilated cardiomyopathy: diagnostic accuracy of endomyocardial biopsy. Br Heart J. 1987;58:156-161.
Factor SM, Minase T, Cho S, Dominitz R, Sonnenblick EH. Microvascular spasm in the cardiomyopathic Syrian hamster: a preventable cause of focal myocardial necrosis. Circulation. 1982;66:342-354.
Rauch B, Helus F, Grunze M, Braunwell E, Mall G, Hasselbach W, Kubler W. Kinetics of 13N-ammonia uptake in myocardial single cells indicating potential limitations in its applicability as a marker of myocardial blood flow. Circulation. 1985;71:387-393.
Bergmann SR, Hack S, Tewson T, Welch MJ, Sobel BE. The dependence of accumulation of 13NH3 by myocardium on metabolic factors and its implications for quantitative assessment of perfusion. Circulation. 1980;61:34-43.
Berry JJ, Baker JA, Pieper KS, Hanson MW, Hoffman JM, Coleman RE. The effect of metabolic milieu on cardiac PET imaging using fluorine-18-deoxyglucose and nitrogen-13-ammonia in normal volunteers. J Nucl Med. 1991;32:1518-1525.