Long-Term Right Ventricular Volume Overload Increases Myocardial Fluorodeoxyglucose Uptake in the Interventricular Septum in Patients With Atrial Septal Defect
Background—Several studies have shown that long-term right ventricular (RV) overload in animal models alters myocardial energy substrate metabolism. However, whether long-term RV volume overload alters this metabolism in the human is unclear.
Methods and Results—We performed positron emission tomography with [18F]fluorodeoxyglucose (FDG) and single-photon emission tomography (SPECT) with [201Tl]TlCl (Tl) and [123I]15-(p-iodophenyl)-3-R,S-methylpentadecanoic acid (BMIPP) in 11 patients with atrial septal defect (ASD) and 11 control subjects. In the FDG study, we calculated myocardial metabolic rate of glucose (MMR) in interventricular septum (IVS) and left ventricular (LV) free wall. MMR was significantly increased in IVS compared with LV free wall in the ASD patients (420±35 versus 333±32 mol · kg−1 · min−1; P<0.05) but not in the control group (347±27 versus 357±25 mol · kg−1 · min−1). In both ASD and control groups, SPECT count was not significantly different between IVS and LV free wall in Tl (ASD, 160±11 versus 177±12; control, 141±12 versus 157±14 counts per 15 minutes) and BMIPP studies (ASD, 203±14 versus 212±18; control, 162±16 versus 176±16 counts per 15 minutes). MMR in the IVS/LV free wall ratio in the ASD group significantly correlated with indices related to RV volume overload.
Conclusions—Given the assumption that long-term RV volume overload did not affect the lumped constant, the present study suggests that, unlike myocardial perfusion or fatty acid analogue uptake, myocardial glucose utilization in IVS relative to LV free wall is increased in relation to long-term RV volume overload in patients with ASD.
Long-term ventricular pressure or volume overload has been shown to alter myocardial energy substrate metabolism.1 2 3 4 Our previous study and others demonstrated that glucose utilization was increased whereas fatty acid analogue uptake was decreased in long-term pressure-overloaded left ventricle (LV) in rats with aortic constriction,3 hypertensive rats,1 and rabbits.2 Other studies showed decreased fatty acid oxidation in long-term volume-overloaded LV in rats with aortocaval fistula.4 In right ventricular (RV) overload, Schwartz et al5 reported increased glucose utilization in the RV free wall in pigs with pulmonary artery constriction. Recently, we demonstrated that long-term RV pressure overload not only increased myocardial glucose utilization in the RV free wall but also altered regional profiles of substrate utilization in the interventricular septum (IVS) and LV free wall in rats with pulmonary artery constriction.6 However, few studies have been published regarding myocardial energy substrate metabolism in RV overload in the human. In patients with atrial septal defect (ASD), left-to-right shunt imposes a RV volume overload and causes RV hypertrophy. Furthermore, LV function and volume are also altered.7 Thus, in patients with ASD, myocardial energy substrate metabolism in patients with ASD may be different from that in normal subjects.
The present study was undertaken to examine whether long-term RV volume overload alters myocardial energy substrate metabolism in patients with ASD. To accomplish this, we performed positron emission tomography (PET) with [18F]fluorodeoxyglucose (FDG) and single-photon emission tomography (SPECT) with [201Tl]TlCl (Tl) and [123I]15-(p-iodophenyl)-3-R,S-methylpentadecanoic acid (BMIPP), a fatty acid analogue, in patients with ASD.
Eleven patients with ASD and 11 control subjects were studied. All patients and control subjects underwent echocardiography. Diagnosis of ASD was based on cardiac catheterization in 10 patients and on echocardiographic findings in 1. None of the patients had heart disease other than ASD. All patients were New York Heart Association functional class 1 or 2. Five normal volunteers and 6 patients with atypical chest pain served as a control group. These control subjects had no echocardiographic evidence of heart disease. All patients with atypical chest pain had normal exercise ECGs and angiographically normal coronary arteries. Provocation tests with ergonovine during coronary angiography were negative in all patients with atypical chest pain. No subjects studied had hypertension, diabetes mellitus, or hyperlipidemia. The purpose and nature of the present study were approved by the Committee for the Administration of Radioactive Substances of the Tohoku University School of Medicine. Written consent was obtained from all subjects before each study.
Positron Emission Tomography
All patients with ASD and 8 control subjects underwent PET studies with FDG to assess myocardial glucose utilization. PET studies were performed with a PT 931/04 PET scanner (CTI). Seven 7.15-mm-wide slices were simultaneously acquired with a 50-mm axial field of view and full width at half maximum (7.1 mm). Each subject fasted for ≥6 hours, was administered 50 g of glucose orally, and was positioned in the scanner after we examined a chest X-ray film obtained previously of the subject in the supine position.. Before we filmed the chest roentgenogram, we placed 6 fine steel wires on each subject’s chest wall at intervals of 2 cm and used them as markers for positioning of the subject. Sixty minutes after glucose loading, 148 to 296 MBq of FDG was injected intravenously. Dynamic scanning was started simultaneously and continued for 45 minutes (9×300 s). Serial arterial blood samples for measurement of plasma FDG radioactivity were withdrawn until the end of scanning (at 20-s intervals to 180 seconds and then at 4, 5, 7, 10, 15, 25, 30, 35, 40, and 45 minutes after FDG injection). Arterial plasma concentrations of glucose, insulin, and FFA were measured 3 times during the PET study (0, 20, and 45 minutes after FDG injection).
All data were corrected for dead time, decay, and measured photon attenuation. A Hanning filter (cutoff frequency, 1.6 cm−1) was used to reconstruct images. The transaxial slice with the largest LV cavity area and that also included the RV free wall was selected for analysis. Two to 5 elliptical regions of interest (ROIs; 82 mm2/ROI, 17 pixels/ROI) were placed on the RV free wall, IVS, and LV free wall. Plasma and tissue time-activity curves were analyzed graphically to quantify fractional rate of tracer transport and phosphorylation, Ki, described by Patlak et al.8 MMR was obtained as follows: MMR=([Glc]p/LC · KI,
where [Glc]p is plasma glucose concentration. We used the LC value of 1.0 reported by Ng et al,9 because we performed the present study with subjects in a mildly hyperinsulinemic state after oral glucose loading. The LC in that condition was assumed to be close to that obtained by Ng et al from the pooled data from subjects in both fasted and hyperinsulinemic conditions.
Single-Photon Emission Tomography
Eight patients with ASD and 6 control subjects underwent SPECT studies with Tl (Nihon Mediphysics) and a fatty-acid analogue, BMIPP (Nihon Mediphysics). SPECT study was performed with a MULTISPECT 3, a triple-headed gamma camera equipped with a low-energy, high-resolution collimator (Siemens; full width at half maximum, 9.7 mm). Butterworth and ramp filters were used to reconstruct images in the transaxial slice. After each subject fasted overnight, 111 MBq of Tl or BMIPP was injected intravenously on different days and SPECT data were acquired. Two to 4 square ROIs (96 mm2 per ROI; 4 pixels per ROI) were placed on the RV free wall, IVS, and LV free wall for the transaxial tomograms of Tl and BMIPP. We then averaged mean counts of Tl and BMIPP in each wall segment.
Magnetic Resonance Imaging
To determine precise ventricular wall thickness in the transaxial slice that was almost identical to the PET and SPECT images, 7 patients with ASD and 5 age-matched normal volunteers who were free from clinical evidence of heart disease underwent MRI (Magnetom Vision, Siemens). Cine scan was performed at intervals of 10 mm and encompassed the heart with a fast, low-angle shot sequence. We visually selected transaxial slices that corresponded to PET and SPECT slices based on the tomographic shape of the RV and LV and measured wall thickness of the RV free wall, IVS, and LV free wall at end-diastole and end-systole.
All values are expressed as mean±SEM. Statistical analysis of differences between groups was done by Student’s unpaired t test or by ANOVA for repeated measures and Fisher’s protected least significant difference test for post hoc analyses when appropriate. Statistical analysis of differences among myocardial regions was done by ANOVA. Correlations between the 2 parameters were determined by simple linear regression analysis. Statistical significance was accepted at a level of P<0.05.
Clinical Characteristics and Echocardiographic Data
Table 1⇓ shows clinical characteristics and echocardiographic data from control and ASD groups. Because separate analysis revealed that these data were not significantly different between normal volunteers and patients with atypical chest pain, we treated both as the control group. Although IVS wall thickness was significantly decreased in the ASD group compared with the control group, the IVS/LV posterior-wall thickness ratio was not significantly different between groups.
PET and SPECT Data
Figure 1⇓ shows transaxial tomograms of FDG, Tl, and BMIPP from a control subject and a patient with ASD. Myocardial metabolic rate of glucose in the IVS (MMRIVS) was significantly increased compared that in the LV free wall (MMRLV) in the ASD group but not the control group (Table 2⇓). MMRIVS and MMRLV were not significantly different between the ASD and control groups. MMRIVS/MMRLV ratio was also significantly increased in the ASD group versus control group (1.28±0.03 versus 0.97±0.06; P<0.01; Figure 2A⇓). IVS/LV count ratios were not significantly different between ASD and control groups for Tl (0.91±0.03 versus 0.91±0.01, respectively; Figure 2B⇓) and BMIPP studies (0.97±0.02 versus 0.92±0.02, respectively; Figure 2C⇓). When we analyzed data from 8 patients with ASD who underwent all 3 tomographic studies, MMRIVS/MMRLV ratio was also significantly increased in patients with ASD versus the control group (1.30±0.04 versus 0.97±0.06; P<0.01). MMRIVS/MMRLV ratio positively correlated with the left-to-right shunt ratio determined from oximetry data (Figure 3A⇓) and negatively correlated with systolic excursion of the IVS (Figure 3B⇓), LV end-diastolic diameter (Figure 3C⇓) determined by echocardiography, and LV end-diastolic volume index determined by left ventriculography (Figure 3D⇓) in the ASD group.
MMR in the RV free wall (MMRRV) tended to be increased in the ASD group compared with the control group (198±19 versus 130±29; P=0.06; Table 2⇑). MMRRV/MMRLV ratio was significantly increased in the ASD group compared with the control group (0.63±0.06 versus 0.36±0.08; P<0.05; Figure 2A⇑). RV/LV count ratio was significantly increased in the ASD group compared with the control group in the Tl (0.57±0.04 versus 0.48±0.01; P<0.05; Figure 2B⇑) and BMIPP studies (0.60±0.03 versus 0.54±0.01; P<0.05; Figure 2C⇑). MMRRV/MMRLV ratio negatively correlated with systolic excursion of the IVS (r=−0.75; P<0.01) in the ASD group.
Insulin, Glucose, and Free Fatty Acid Concentrations
Plasma concentrations of insulin, glucose, and free fatty acid (FFA) during the PET study were not significantly different between ASD and control groups (Table 3⇓).
Cardiac Catheterization Data
Cardiac catheterization data from the ASD group are shown in Table 4⇓. Coronary angiography showed no significant percentage diameter stenosis >50% in any patient. All patients with atypical chest pain underwent cardiac catheterization, and their data, described below, did not differ significantly from those of the ASD group: mean pulmonary artery pressure, 13±2 mm Hg; RV systolic pressure, 25±4 mm Hg; RV end-diastolic pressure, 6±2 mm Hg; LV systolic pressure, 145±12 mm Hg; LV end-diastolic pressure, 10±2 mm Hg; LV end-diastolic volume index 75±11 mL/m2; and LV ejection fraction, 70±3%.
RV wall thickness was significantly increased in patients with ASD compared with normal volunteers at end-diastole (7.4±1.1 versus 4.4±0.2 mm; P<0.05) but not at end-systole (11.1±1.0 versus 9.0±0.5 mm, respectively). IVS thickness (10.7±0.7 versus 9.8±0.4 mm at end-diastole; 13.7±0.7 versus 14.0±1.0 mm at end-systole) and LV free wall thickness (10.6±0.6 versus 9.8±0.2 mm at end-diastole; 13.7±0.4 versus 14.4±1.0 mm at end-systole) were similar in patients with ASD and normal volunteers, respectively. IVS/LV free wall thickness ratio was similar in the patients with ASD and normal volunteers at both end-diastole (0.95±0.03 versus 1.00±0.03, respectively) and end-systole (1.00±0.04 versus 0.97±0.02).
Increased FDG Uptake in IVS Relative to LV Free Wall
The mechanism for increased FDG uptake in the IVS relative to the LV free wall in patients with ASD is unclear. Decreased coronary blood flow affects myocardial FDG uptake. However, this would not be responsible for the relatively higher FDG uptake in the IVS in the ASD group, because all patients in the ASD group who underwent selective coronary angiography had normal coronary arteries.
Decreased count recovery secondary to partial-volume effects is a well-recognized limitation of cardiac PET and SPECT imaging.10 In our echocardiographic and MRI studies, IVS/LV wall-thickness ratios in the ASD group did not significantly differ from the control subjects. Therefore, the difference in wall thickness between the IVS and LV free wall is unlikely to account for the relatively higher FDG uptake in the IVS in the ASD group.
Because we did not perform the spillover correction in the present study, we may have overestimated tracer uptake in the myocardium, especially in the IVS, in which recovered counts are affected by spillover from both chambers. Furthermore, the radioactivity in the chambers might remain longer due to a shunt in the ASD group. However, when we plotted the radioactivity in the RV and LV chambers at each time point normalized by that in the LV chamber, which was measured during the first 5 minutes after the FDG injection, time-activity curves for the RV and LV chambers in the two groups could easily be superimposed (data not shown). Therefore, we believe that the IVS in the 2 groups was affected similarly by spillover.
Several studies have demonstrated that myocardial glucose utilization is increased in hypertrophied myocardium.1 3 6 Dong et al11 reported that IVS thickness was increased in patients with long-term RV pressure overload. Sekine et al12 reported that myocyte hypertrophy was present in the IVS in addition to the RV free wall but was not in the LV free wall in patients with long-term RV pressure overload. Thus, hypertrophic response in RV overload may be different between the IVS and the LV free wall. Furthermore, Booth et al7 reported that LV mass was increased in patients with ASD. In the present study, although our data from echocardiography and MRI showed no evidence of increased wall thickness in the IVS, we cannot exclude the possibility of myocyte hypertrophy in the IVS because we did not perform histological analysis.
Changes in myocardial contractile work may also account for changes in myocardial FDG uptake. Gerz et al13 reported that, in the human, myocardial glucose utilization increases with exercise. Although systemic blood flow was not significantly different between the ASD and control groups (3.20±0.26 versus 3.27±0.28 L/min, respectively; P=NS), pulmonary blood flow would be increased in the ASD group compared with the control group because of the shunt. Because the IVS contributes to RV cardiac output, an increase in cardiac work in the RV may account for relatively higher FDG uptake in the IVS in the ASD group. Furthermore, RV volume overload causes a shift of the IVS toward the LV cavity at end-diastole and a paradoxical rightward excursion of the IVS during systolic contraction.14 This shift of the IVS increases the radius of the curvature of the IVS and may increase the circumferential wall stress in the IVS. Because the MMRIVS/MMRLV ratio negatively correlated with the systolic excursion of the IVS in the ASD group, the regional increase in myocardial wall stress and workload in the IVS possibly accounts for increased FDG uptake in the IVS relative to LV free wall in the ASD group.
BMIPP Uptake in IVS and LV Free Wall
In contrast to the FDG PET study, the IVS/LV count ratios of BMIPP were not significantly different between the ASD and control groups. Myocardial uptake of BMIPP was reported to be decreased compared with distribution of flow tracers in patients with coronary artery disease15 and hypertrophic cardiomyopathy.16 On the other hand, we have shown that the myocardial uptake of β-methyl[1-14C]-heptadecanoic acid, another branched-chain fatty acid analogue, is unchanged in a rat model of long-term RV pressure overload, although the myocardial uptake of 2-deoxyglucose is accelerated.6 It is therefore consistent with the results of our previous study that the IVS/LV count ratio of BMIPP was unchanged despite the increased MMRIVS/MMRLV ratio in the ASD group. BMIPP is mainly incorporated into the endogenous lipid pool in the myocardium due to its β-methyl branched structure. Therefore, its uptake may not directly reflect β-oxidation of fatty acids. However, because the myocardial uptake of BMIPP is reported to be related to intracellular ATP levels,17 it may reflect in part myocardial fatty acid utilization and energy production. Further study with a non–branched-chain fatty acid analogue may be helpful to assess myocardial β-oxidation of FFAs in patients with ASD.
Increase in Measured Radioactivity of Tracers in RV Free Wall
RV wall thickness measured by echocardiography tended to be increased in the ASD group compared with the control group, and that measured by MRI was significantly increased in the ASD group compared with the normal volunteers. To examine the partial volume effect10 on the recovered count in the RV, we performed phantom studies with a double-walled cylindrical phantom and determined the recovery coefficients. Recovery coefficients for the RV free wall in normal subjects in the FDG study, for instance, changed greatly, from 0.260.03 at end-diastole (wall thickness, 4.4±0.2 mm) to 0.63±0.03 at end-systole (wall thickness, 9.0±0.5 mm). Because we performed PET and SPECT studies without ECG-gated data acquisition, we could not accurately correct recovered counts with these recovery coefficients in the present study. The increased RV wall thickness may explain the increases in measured radioactivity of FDG, Tl, and BMIPP in the RV free wall in the ASD group because the object size affects the recovered count by the partial volume effect. It is also possible that accelerated FDG uptake due to myocardial hypertrophy1 3 6 contributed to the increased MMRRV/MMRLV ratio in the ASD group.
Spillover may affect the quantitative analysis more in the RV free wall because of its relatively fewer counts. With regard to spillover in the FDG study, normalized time-activity curves for the RV chamber in the 2 groups could be superimposed, as mentioned in the previous section. Thus, RV free walls in the 2 groups should have been affected similarly by the spillover, although the absolute values may not be accurate.
Several limitations exist when data are compared from different imaging modalities. First, the spatial resolution is different between PET and SPECT studies. Second, the count ratios calculated from the SPECT studies are not truly quantitative because attenuation correction was not performed in the SPECT studies. Third, slices obtained from different imaging modalities in each patient may be slightly different. However, these factors should have affected the results equally in both ASD and control groups.
Regarding the quantitative assessment of MMR, several studies have shown that the lumped constant (LC; the correction factor used to equate FDG to glucose uptake) changes with the changes in the insulin concentration and substrate composition in animal and human studies.9 18 19 20 21 In the present study, the FDG PET study was performed in the same glucose-loaded condition in both groups. During the present study, plasma concentrations of insulin, glucose, and FFA were not significantly different between groups, although that of FFA decreased from 20 to 45 minutes after FDG injection in both groups. Therefore, the change in the LC due to changes in insulin and substrate concentrations should have affected calculated values for MMR equally in the 2 groups. The LC may also differ between IVS and LV free wall, because myocyte hypertrophy may exist in the IVS in patients with ASD. Further study is necessary to determine whether the LC differs between IVS and the LV free wall in RV volume overload.
If it is assumed that the effects of volume overload on the RV did not affect the LC, the MMR probably reflects myocardial glucose utilization in patients with ASD. Therefore, the present study suggests that, unlike myocardial perfusion or fatty acid analogue uptake, myocardial glucose utilization in the IVS relative to the LV free wall is increased in relation to long-term RV volume overload in patients with ASD.
We are grateful to Shoichi Watanuki, Shinya Seo, Masayasu Miyake, and Youetsu Abe for their support with the tomographic studies and Yuko Itoh, MD, and Masato Endo, MD, for their support with the echocardiographic study.
- Received July 12, 1999.
- Revision received October 25, 1999.
- Accepted November 5, 1999.
- Copyright © 2000 by American Heart Association
Yonekura Y, Brill AB, Som P, Yamamoto K, Srivastava SC, Iwai J, Imaleh KR, Livni E, Strauss AW, Goodman MM, Knopf F. Regional myocardial substrate uptake in hypertensive rats: a quantitative autoradiographic measurement. Science. 1985;227:1494–1496.
Taegtmeyer H, Overturf ML. Effects of moderate hypertension on cardiac function and metabolism in the rabbit. Hypertension. 1988;11:416–426.
Kagaya Y, Kanno Y, Takeyama D, Ishide N, Maruyama Y, Takahashi T, Ido T, Takishima T. Effects of long-term pressure overload on regional myocardial glucose and free fatty acid uptake in rats: a quantitative autoradiographic study. Circulation. 1990;81:1353–1361.
Cheikh RD, Guendouz A, Moravec J. Control of oxidative metabolism in volume overloaded rat hearts: effects of different lipid substrates. Am J Physiol. 1994;266:H2090–H2097.
Schwartz GG, Steinman S, Garcia J, Greyson C, Massie B, Weiner MW. Energetics of acute pressure overload of the porcine right ventricle: in vivo 31P nuclear magnetic resonance. J Clin Invest. 1992;89:909–918.
Ng CK, Soufer R, McNulty PH. Effect of hyperinsulinemia on myocardial fluorine-18-FDG uptake. J Nucl Med. 1998;39:379–383.
Dong S, Crawley AP, MacGregor JH, Petrank YF, Bergman DW, Blenkie I, Smith ER, Tyberg JV, Beyar R. Regional left ventricular systolic function in relation to the cavity geometry in patients with chronic right ventricular pressure overload: a three-dimensional tagged magnetic resonance imaging study. Circulation. 1995;91:2359–2370.
Gertz EW, Wisneski JA, Stanley WC, Neese RA. Myocardial substrate utilization during exercise in humans. J Clin Invest. 1988;82:2017–2025.
Louie EK, Lin SS, Reynertson SI, Brundage BH, Levitsky S, Rich S. Pressure and volume loading of the right ventricle have opposite effect on the left ventricular ejection fraction. Circulation. 1995;92:819–824.
Tamaki N, Kawamoto M, Yonekura Y, Fujibayashi Y, Takahashi N, Konishi J, Nohara R, Kambara H, Kawai C, Ikebukuro K, Kato H. Regional metabolic abnormality in relation to perfusion and wall motion in patients with myocardial infarction: assessment with emission tomography using an iodinated branched chain fatty acid analogue. J Nucl Med. 1992;33:659–667.
Kurata C, Tawarahara K, Taguchi T, Aoshima S, Kobayashi A, Yamazaki N, Kawai H, Kaneko M. Myocardial emission computed tomography with iodine-123-labeled beta-methyl-branched fatty acid in patients with hypertrophic cardiomyopathy. J Nucl Med. 1992;33:6–13.
Fujibayashi Y, Yonekura Y, Takemura Y, Wada K, Matsumoto K, Tamaki N, Yamamoto K, Konishi J, Yokoyama A. Myocardial accumulation of iodinated beta-methyl-branched fatty acid analogue, iodine-125-15-(p-iodophenyl)-3(R, S) methyl pentadecanoic acid (BMIPP), in relation to ATP concentration. J Nucl Med. 1990;31:1818–1822.
Ng CK, Holden JE, DeGrado T, Raffel DM, Kornguth ML, Gatley SJ. Sensitivity of myocardial fluorodeoxyglucose lumped constant to glucose and insulin. Am J Physiol. 1991;260:H593–H603.
Russell RR III, Mrus JM, Mommessin JI, Taegtmeyer H. Compartmentation of hexokinase in rat heart: a critical factor for tracer kinetic analysis of myocardial glucose metabolism. J Clin Invest. 1992;90:1972–9977.
Hariharan R, Bray M, Ganim R, Doenst T, Goodwin GW, Taegtmeyer H. Fundamental limitations of [18F]2-deoxy-2-fluoro-d-glucose for assessing myocardial glucose uptake. Circulation. 1995;91:2435–2444.
Botker HE, Bottcher M, Schmitz O, Gee A, Hansen SB, Cold GE, Nielsen TT, Gjedde A. Glucose uptake and lumped constant variability in normal human hearts determined with [18F] fluorodeoxyglucose. J Nucl Cardiol. 1997;4(pt 1):125–132.