Keywords
INTRODUCTION
The dilatory capacity of the coronary microvasculature or coronary flow reserve (CFR) is altered in patients with significant coronary lesions.1 This can also change in the presence of coronary risk factors,2,3 with arterial hypertension,4 in hypertrophic5 or dilated6,7 myocardiopathy, and in hypertrophy secondary to valve lesions.8
CFR has been found to be altered in experimental heart failure.9 Heart failure is associated with neurohumoral activation and changes in peripheral circulation.10 It is possible that the increase in cytokines and the reduction in flow influence the development of endothelial dysfunction in these patients.10,11 Endothelial dysfunction may contribute to an increase in peripheral vasomotor tone during exercise12 and to abnormal control of brachial blood flow.13 One experimental study9 showed that coronary endothelial dysfunction with reduction of CFR appears before heart failure. It is possible that the vasoconstriction caused by the increase in enothelin-1 has a relationship with these changes, as carvedilol improved heart failure at the same time as it reduced concentrations of this substance.14
CFR can be measured by different methods; nevertheless, positron emission tomography (PET) is the only non-invasive technique that provides absolute measurement of overall and regional myocardial blood flow in mL/min/g.15
The aim of this study is to verify with PET whether patients in heart failure of non-ischemic origin have a functional alteration in coronary microvasculature, and if this change depends on the severity of the heart failure at the time of the study.
PATIENTS AND METHODS
We studied 40 patients with cardiopathy of non-ischemic origin who were in heart failure, New York Heart Association (NYHA) functional class III-IV, at the time of the study or prior to the study. A total of 41 PET studies were performed to measure the CFR. The study was approved by the ethics committee of our institution and all patients signed an informed consent form. We established the following groups according to the functional level of the patient at the time the study was performed: 1. Ten patients in NYHA class III-IV; 2. Ten patients in NYHA class II who were not receiving beta-blockers; 3. Eleven patients in NYHA functional class II, treated with carvedilol, and 4. Ten patients who had previously had heart failure with an ejection fraction <0.45 and at the time of the study, in addition to being in NYHA functional class I, had on ejection fraction of ≥0.45. Only 1 patient was studied twice: the first time in group 2 and the second time in group 3 (after 1 year of treatment with carvedilol). We obtained a clinical history, physical examination, 12-lead electrocardiogram, echocardiogram, usual blood work, and coronary angiography in all patients. Only patients with significant coronary lesions were included.
Table 1 shows the principal baseline characteristics of each group. Table 2 shows the diagnosis, ejection fraction, and mass and volume index of the left ventricle.
The patients in groups 2, 3, and 4 had previously been in NYHA functional class III-IV and improved with drugs or surgery (in the case of valve lesions). The 4 groups received digitalis, diuretics, and angiotensin converting enzyme inhibitors alone or in combination. In addition to these drugs, all patients in group 3 received carvedilol (the only beta blocker used), at a dose of between 6.25 mg/day and 50 mg/day. Treatment was established according to the criteria of the attending physician for each patient, who in general adjusted to the recommendations based on the evidence.16 Only 1 patient in group 2 was able to be treated with carvedilol, and was then placed in group 3. The remaining patients in group 2 had some type of contraindication for treatment with betablockers.
All patients abstained from ingesting caffeine for at least 24 hours prior to the PET study. The patients who smoked stopped smoking for at least 1 week prior to the study. The morning of the study the patients took their usual medications and the PET study was performed in the afternoon.
Positron emission tomography
The transmission and emission images were obtaied with a Siemens Ecat Exact HR+ tomograph. This equipment consisted of a system of 32 crystal ring detectors that allowed acquisition, with 32 direct planes and 31 crossed planes, of 63 simultaneous transaxial images that covered a 15.5 cm field, with a 2-dimensional resolution of 4.5 mm in the transaxial plane and 4.5 mm in the axial plane in the center of the viewing field.
Myocardial blood flow (MBF) was measured at baseline and during hyperemia induced by adenosine triphosphate (ATP) using N-13 ammonia and dynamic PET acquisition. First, an transmission image was obtained for 15 minutes to correct photon attenuation. After the first intravenous injection of N-13 ammonia (9.25 MBq/kg, up to a maximum of 740 MBq), serial images were obtained at rest, with a variable length dynamic sequence (12 imagesx10 seconds, 4 imagex15 seconds, 4 imagesx30 seconds, 3 imagesx300 seconds). The protocol used for PET data acquisition has been described by other authors.17 After acquisition of the baseline study, we waited 50 minutes to allow N-13 ammonia radioactive fall-out (T1/2=9.9 minutes). ATP was infused for 6 minutes at a dose of 0.160 mg/kg/min. During the infusion cardiac frequency, arterial pressure, and a 12-lead electrocardiogram were monitored continuously. At the 4th minute of the ATP infusion, the second injection of N-13 ammonia was administered. The acquisition of the stress images began from the moment of the injection and followed the same protocol as for the images obtained with the patient at rest.
Image processing. The images were reconstructed using a Hann filter with an 0.4 slice frequency, providing an effective resolution for a 7 mm plane. The transaxial mages were reoriented on the short axis and on the long vertical and horizontal axes. The angles of the long horizontal and vertical axes were defined by using the last 3 images from the dynamic sequence, and then were used for the reorientation of the 23-image sequence. For quantitative analysis we used 6 continuous sections of the short axis corresponding to the middle of the left ventricle.
MBF measurement. Regional MBF was calculated in accordance with a 3-compartment model,16 which represented vascular and extravascular N-13 ammonia, and the N-13 ammonia metabolically trapped in the form of glutamine, which allowed estimation of the constant K1 that represented MBF in mL/g/min.
The informatics program used to calculate the regional MBF was developed by Muzic et al.18 To determine the radioactivity input function, an area of interest in the most basal planes of the left ventricular cavity on the short axis was delimited. Twelve areas of interest were defined per plane on the 6 planes in the last image of the dynamic sequence. A sample of the collection of dynamic images was taken and 72 activity-time curves were obtained. Regional MBF was analyzed in 4 areas of the left ventricle: anterior, septal, inferoposterior, and lateral.
Statistical analysis
Descriptive statistics were expressed as mean±standard deviation. Quantitative measurements were compared via ANOVA, followed by the Tukey test. The differences in percentages between groups were compared with Fisher exact test (PEPI statistical packet, J.H. Abramson and P.M. Gahlinger, 1993-2000). In order to establish a possible relationship between variables, we used the Pearson product moment correlation test. We calculated the coefficient of variation of the regional MBF at rest in order to evaluate the spatial heterogeneity of myocardial perfusion,19 determining in each subject the quotient of the standard deviation and the mean regional MBF in 4 areas of the myocardium.
RESULTS
Baseline characteristics
We did not find any significant differences between the 4 patient groups with regard to age, sex, body mass index, smoking habits, diabetes mellitus, arterial hypertension, or the total cholesterol/high density lipoprotein cholesterol quotient (Table 1). Given that CFR tends to be reduced after the age of 70 years,20 we found that 12 patients were older than 69 years of age: 3 in group 1, 2 in group 2, 2 in group 3, and 5 in group 4 (differences were without statistical significance). We also found no significant differences in the proportion of the various types of cardiopathy among the 4 groups (Table 2). All valvulopathy was rheumatic, mitroaortic or degenerative in nature, with different degrees of stenosis and insufficiency. Three of these patients also had serious tricuspid insufficiency. Only 1 patient had restrictive myocardiopathy of unknown cause. Eight of the patients in group 4 were diagnosed with dilated myocardiopathy (hypertensive in origin in 6 patients). Ten patients in this group, who previously had a reduced ejection fraction and were in NYHA functional class III-IV, improved with treatment. At the time of the study they were in NYHA functional class I with a normal ejection fraction.
The ejection fraction was significantly higher in the patients in group 4 than those in groups 1 and 3. We did not observe a significant difference between groups 2 and 4. This was due to the fact that some patients in group 2, with valve lesions, had normal ejection fractions. There was no significant difference between patient groups with regard to left ventricle mass and volume indices (Table 2).
Hemodynamic findings
Table 3 shows that there was no significant difference in baseline arterial systolic pressure among the 4 patient groups. The double product (cardiac frequency xarterial systolic pressure), indicative of the consumption of myocardial oxygen, increased slightly but significantly (P=.001) with ATP. There were no significant differences in this parameter among the 4 groups either at rest or during the infusion of ATP.
Myocardial flow
There was no significant spatial heterogeneity (P=.798) in baseline regional myocardial blood flow (BMFb) between the 4 myocardial regions studied: anterior, septal, inferoposterior, and lateral. For this reason, we only analyzed the overall BMFb.
MBF, coronary resistance and CFR
Table 3 shows that there was no significant difference in MBF among the 4 groups. As BMFb increases when the double produce is raised,21 we calculated the normalized BMFb (quotient of BMFbx10 000/double product). We did not find significant differences between the 4 groups, although we did note higher values in group 4 and lower values in group 1. MBFATP was significantly higher in group 4 than in group 1.
Given than ATP disrupted the relationship between MBF and cardiac load, hyperemic flow was not corrected for the double product. The CFR (uncorrected BMFATP/BMFb) was significantly higher in group 4 than in group 1. No significant differences were observed among the 4 groups with regard to baseline coronary resistance (mean arterial pressure/BMFb) or minimal coronary resistance (with ATP), but there was more of a tendency toward less coronary resistance with ATP in group 4 than in the other groups. Figure 1 shows the uncorrected BMFb the BMFATP, and the CFR in the 4 patient groups. Compared with reported CFR values21 (3.01±0.73) in healthy volunteers of 64 years of age±9 years of age, the CFR was clearly reduced in the 4 patient groups, especially in group 1. Although the CFR was slightly higher in groups 2 and 3 than in group 1, only group 4 had values that were significantly higher than group 1. The patient who underwent 2 PET studies had a CFR of 1.63 when in group 2 and of 2.99 when included in group 3, after a year of treatment with carvedilol. The ejection fraction was 0.25 during the first study and 0.35 during the second study.
Fig. 1. Graph showing myocardial blood flow (MBF) in each group: at baseline (BMFb) and with adenosine triphosphate (MBFATP). Coronary flow reserve (CFR) represents the quotient of MBFATP/BMFb. Only group 4 had MBFATP and CFR values that were significantly higher than those of group 1.
There was no significant correlation between CFR and: age (r=0.171; P=.143), systolic arterial pressure (r=-0.022; P=.445), ejection fraction (r=0.096; P=.275), left ventricle mass index (r=0.047; P=.388), left ventricular telediastolic volume index (r=0.025; P=.439), or left ventricular telesystolic volume index (r=0.056; P=.368).
None of the following co-variables significantly changed the CFR: age, body mass index, ejection fraction, total cholesterol/HDL cholesterol quotient, or left ventricular mass or volume indices. Sex, smoking habits, diabetes, and arterial hypertension likewise did not affect the differences between the patient groups (data not expressed).
DISCUSSION
CFR in heart failure
Our study shows for the first time with PET that patients with heart failure that is non-ischemic in origin, whatever the etiology, have depressed CFR, which is related to their functional class. The CFR was 1.46±0.48 in patients in NYHA functional class III-IV; 1.73±0.72 and 1.89±0.75 in patients in NYHA functional class II (with and without carvedilol, respectively), and 2.40±0.95 in patients in NYHA functional class I. These differences cannot be attributed to age, which was similar for patients in all 4 groups. Actually, it is not clear whether CFR diminished in the elderly (patients >69 years of age) due to deficient dilation of coronary microvasculature20 or due to an increase in cardiac load (double product), which elevated the BMFb20 CFR measured in healthy volunteers of a similar age to that of our patients was 3.01±0.73.21 In our study, the difference in CFR between the patients in groups 1 and 4 was significant. Theoretically, ventricular dysfunction can influence the diminishment of CFR. Nevertheless, in our patients, neither the ejection fraction nor the ventricular volume indices, taken as co-variables, significantly changed the results. It is probable that the functional class was more a determinant of CFR than the underlying cardiac disease.
Previous studies
PET and intracoronary echo-Doppler studies22 of patients with dilated myocardiopathy and heart failure have shown a reduction in CFR. Intracoronary Doppler reveals endothelial dysfunction in patients with both microvascular and epicardial dilated myocardiopathy.23 Peripheral resistance is elevated in heart failure12 and brachial hyperemic flow is reduced due to endothelial dysfunction.13
Our study was performed with ATP. This drug, similar to adenosine,24 had been shown to be useful in myocardial perfusion studies,25 and some studies26 report that its vasodilator effect is dependent on the endothelium.
Cardiac hypertrophy
The majority of our patients had dilated myocardiopathy or mitroaortic valve disease. Dilated myocardiopathy is associated with reduced CFR in the absence of heart failure, probably due to vasodilation anomalies.7 The progression of this disease is associated with greater depression of myocardial perfusion, both at rest18 and after the administration of dipiridamol.7,18 This depression cannot be totally explained by the elevation of left ventricular diastolic pressure or by reduced coronary perfusion pressure.27 Arterial hypertension is another cause of the reduction of CFR,4 partly due to structural arterial changes. Nevertheless, CFR in hypertrophy secondary to valvulopathy has been studied less frequently. One study8 showed less change in the CFR in ventricular hypertrophy secondary to aortic stenosis than in arterial hypertension. Our patients had a decrease in CFR related to the functional class of their heart failure, and not to the type of underlying heart disease. These findings suggest that heart failure, on its own, may change coronary microvascular dilation, independent of its cause. This change was in addition to that produced by their baseline cardiopathy.
Neurohumoral factors
In heart failure there is an increase in different neurohumoral and inflammatory factors28-31 that can induce not only peripheral vasoconstriction, but also coronary vasoconstriction. It is possible to diminish the vasoconstriction when the heart failure is controlled and these factors are reduced. We believe that this is the most likely explanation for our findings, although the study itself does not allow confirmation of this. It is also unknown how much time is needed for the CFR to normalize when heart failure improves clinically. The patients in group 4, in NYHA functional class I, are those who had the highest CFR. Although their ejection fraction had normalized by the time the study was performed, this variable had no significant relation to the CFR. This fact may indicate that the diminishment of CFR was not due to a decrease in the need for oxygen in the face of a reduced ejection fraction, but to a microcirculation problem.
Carvedilol
The patients in group 3, who were being treated with carvedilol, had a tendency toward a higher CFR than those in group 2, who were in the same functional class, although their ejection fraction was somewhat higher (but without statistical significance) in group 3. This may be attributable to the reduced BMFb in group 3 (Table 3). Nevertheless, the patients in group 1 had a lower CFR (but without statistical significance) than those in group 3, in spite of having a lower BMFb (not statistically significant). The patient who underwent 2 different studies (with and without carvedilol) had an increase in CFR with carvedilol that was disproportionate to the increase in the ejection fraction. Carvedilol is a beta and adrenergic alpha-1 antagonist that also has antioxidant32 and antiendothelin-1 properties.33 These properties may explain, at least in part, the slightly elevated CFR in group 3. A previous study14 showed that the plasma changes in endothelin-1 reflect the clinical response to carvedilol in patients with heart failure. Several patients in group 4, who were those patients who had the highest CFR, also received carvedilol.
Study limitations
1. We did not have a fifth group of healthy volunteers of similar ages to compare with the 4 patient groups, but we believe that they can be compared with the healthy volunteers of a similar age in the study by Czernin et al21 as these authors used methods similar to ours and obtained results similar to ours (unpublished data) in young volunteers.
2. It would have been ideal to have studied the same patients in 4 different situations during the course of their illness, but this was not possible. The majority of the patients in group 1 received a heart transplant or died. In contrast, the patients in group 2 had some contraindication for treatment with beta blockers.
3. Mixing patients with valve disease with patients with myocardiopathy could be considered a limitation. Nevertheless, CFR is changed with both pathological processes. In figure 1 it can be seen that there is no appreciable difference in the standard deviation among the patient groups; therefore, from the statistical point of view, there is no heterogeneity amongst them. There was only a significant difference in the ejection fraction, which was higher in group 4, as it was a selection criterion for this group. If group 4 had been made up of healthy volunteers, their greater CFR would have to be attributed to the absence of cardiopathy. Nevertheless, the group was made up of patients with cardiopathy whose heart failure had improved; therefore, the difference between group 4 and the rest of the groups must be attributed to their distinct functional class, once the existence of a correlation between hemodynamic variables and CFR was discarded.
CLINICAL IMPLICATIONS
CFR measured with PET provides greater knowledge of the coronary microvascular physiopathology in non-ischemic heart failure. Patients who are in a better functional class are those who present with higher CFR levels. If low CFR is due to a primary microcirculatory problem, as it seems to be, microcirculation must be a therapeutic objective. Other studies are needed with evolutionary followup to establish a possible relationship between CFR, neurohumoral and inflammatory factors, and the prognosis of patients in heart failure.
ACKNOWLEDGEMENTS
The authors are grateful to Iván Peñuelas for the preparation of the radiotracers during the PET studies and to Javier Díez for his scientific assessment.
This study was financed by a FIS grant.
Correspondence: Dra. Isabel Coma-Canella.
Departamento de Cardiología. Clínica Universitaria de Navarra.
Avda. de Pío XII, 36. 31008 Pamplona. España.
E-mail: icoma@unav.es