ORIGINAL ARTICLE 
Annals of Nuclear Medicine Vol. 14, No. 3, 165-172, 2000 
Impaired coronary microvascular function in diabetics 
Go TSUJIMOTO 
Department of Internal Medicine, Kobe Rosai Hospital 
Global and regional myocardial uptake was determined with technetium-99m tetrofosmin and a 4 hour exercise (370 MBq iv) and rest (740 MBq iv) protocol, in 24 patients with non-insulin dependent diabetes mellitus and in 22 control subjects. The purpose of this study was to evaluate impaired coronary microvascular function in diabetics by measurement of % uptake increase in myocardial counts. The parameter of % uptake increase (DMTU) was calculated as the ratio of exercise counts to rest myocardial counts with correction of myocardial uptake for dose administered and physical decay between the exercise study and the rest study. Global DMTU was significantly lower in the diabetics than in control subjects ( 14.4 +- 5.4% vs. 21.7 +- 8.5%, p < 0.01 ). Regional DMTU in each of 4 left ventricular regions (anterior, septal, inferior, posterolateral) was significantly lower in the diabetic group than in the control group (p < 0.01 ) respectively, but there were no significant differences between DMTU in the 4 left ventricular regions in the same group. DMTU was useful as a non-invasive means of evaluating impaired coronary microvascular function in diabetics. 
Key words: diabetes mellitus, coronary flow reserve, tetrofosmin, coronary microcirculation, SPECT 
INTRODUCTION 
MICROANEURYSMS,1 hyalinization2 or wall thickening of intramural arterioles,2-4 and reduced density of capillary vessels5 have been described as morphological changes in the heart of diabetics. Functionally, the maximum pharmacological reserve of coronary blood flow has been found to be depressed, and the mechanism of regulation of coronary vascular resistance during atrial tachypacing has been found to be impaired in diabetics compared with nondiabetic patients, suggesting impairment of coronary microcirculation.6-14 Reduced coronary flow reserve has been reported not only in diabetes mellitus but in various other diseases such as coronary stenosis,15 hypertrophied heart16,17 and syndrome X.18,19 
 An index of 99mTc tetrofosmin (TF) or 99mTc methoxybutylisonitrile (MIBI) myocardial single-photon emission computed tomography (SPECT) calculated from the 
Received October 18, 1999, revision accepted February 21, 2000. For reprint contact: Go Tsujimoto, M.D., Department of Internal Medicine, Kobe Rosai Hospital, 4-1-23 Kagoike-tori, Chyuo-ku, Kobe 651-0053, JAPAN. 
ratio of exercise counts to rest myocardial counts,20-24 has shown promise as a tool to assess coronary flow reserve noninvasively. The index provides functional and quantitative information for use in combination with conventional inspection of myocardial SPECT images. 
 The purpose of this study was to evaluate impaired coronary microvascular function based on noninvasive measurement of coronary flow reserve in diabetics by a radioisotope technique. 
MATERIALS AND METHODS 
Patient Characteristics 
Forty-six subjects without significant (> 75%) coronary artery stenosis on coronary angiography or left ventricular wall motion abnormality were studied. They were divided into a control group and a diabetic group.
 The 22 control subjects did not have mellitus on the basis of their history and blood chemistry data.
 The 24 diabetics had non-insulin dependent diabetes mellitus, and 3 of them (12.5%) were on insulin therapy. The average interval since diagnosis was 11.0 +- 10.9 years. The average serum HbA1c was 7.4 +- 2.0%, and their serum immuno-reactive insulin was 5 .7 +- 4.3 uU/ml. Eighteen of them (75%) had a history of proteinuria, and 3 (12.5%) had a history of diabetic retinopathy. No patients had a history of diabetic neuropathy. The diagnosis of diabetes mellitus was made on the basis of presentation with classic symptoms and a random blood glucose test value of 200 mg per deciliter or greater. If diabetes was suspected but not confirmed by the random glucose determination, the diagnosis of diabetes mellitus was established by a blood glucose value after an overnight fasting of 140 mg or more per deciliter on at least two separate occasions.
 No significant difference was found between the two groups with regard to sex distribution or average age. The incidence of a history of hypertension was significantly higher in the diabetic patients (29.1%) than in the control patients (O%) (p < 0.05). Fasting blood sugar was significantly higher in the diabetic patients than in the control patients (p < 0.05), but there were no significant differences in the serum total cholesterol or serum triglyceride values. The body mass index was similar in both groups (Table 1). 
Exercise and Rest TF Myocardial SPECT 
We performed exercise stress testing with a bicycle ergometer in a symptom-limited manner. Exercise was started at 25 watts, and the load was increased by 25 watts every 3 minutes until the exercise endpoint was reached. In each group we selected only patients who reached a rate pressure product greater than 20000 at the exercise endpoint. 
SPECT Data Acquisition and Processing 
A 370 MBq dose of TF was injected during peak exercise, and a 740 MBq dose of TF was re-injected 4 hours after the completion of exercise. After each injection, we flushed the TF remaining in the vein with 5 ml of normal saline. The images were acquired at 30 minutes after the first and the second injection, by means of a rotating large-field-of-view gamma camera (SNC5100R, Shimadzu, Kyoto, Japan) equipped with a low-energy high-resolution parallel-hole collimator centered on the 140 keV photopeak for Tc-99m, with a 20% window. Thirty-two views were collected with a 64 x 64 acquisition matrix for 30 seconds per projection over 180 degrees. A series of transmural slices were reconstructed from the raw scintigraphic data by a backprojection technique with a Ramp filter. Short-axis and vertical and horizontal long axis tomograms were reconstructed with a data processor (Scintipac 7000, Shimadzu, Kyoto, Japan). 
Quantitative Analysis and Display 
To quantitatively analyze the tracer distributions, a circumferential profile curve was generated from apical to basal short-axis slices to create a bull's-eye polar map normalized with 100% as the maximum count, and the left ventricle was divided into 24 segments on the bull's-eye polar map (Fig. 1). 
Ratio of Myocardial Tetrofosmin Uptake During Exercise to Uptake at Rest 
Myocardial counts of initial and rest images were abbreviated as C1 and C2. The remaining counts from the examination after the exercise, reduced by actual physical decay, are represented by C1'. The injected doses of TF at peak exercise and at rest are referred to as D1 and D2 (Fig. 2). The ratio of myocardial TF uptake (MTU) during exercise to uptake at rest was calculated by the formula. 
DMTU = {[C1/(C2 - C1')] x R - 1 } x 100 (%) (where R = D2/D1) 
 DMTU was calculated in each of the 24 left ventricular segments, and the mean value of all segments in the entire left ventricle is represented by DMTU in the whole left ventricle. The regional differences in DMTU in 4 left ventricular regions (anterior, septal, inferior and posterolateral segments) were also analyzed. 
Statistics 
Values are expressed as means +- SD, and the data were analyzed by the chi-squared test and an unpaired t-test. Differences between multiple data were evaluated by ANOVA with Scheffe's test for continuous variables. The relationship between the variables was assessed by linear regression analysis. Statistical significance was defined as a p value of 0.05 or less. 
RESULTS 
Visual Inspection of Exercise and Rest TF Myocardial SPECT Images: There were no perfusion abnormalities in any of the patients. 
Hemodynamic Change Before and After Exercise: As shown in Table 2, the rate pressure product at the exercise endpoint and the maximum heart rate were similar in the two groups. 
Ratio of Myocardial Tetrofosmin Uptake During Exercise to Uptake at Rest: DMTU in the entire left ventricle was 21.7 +- 8.5% in the control patients, and 14.4 +- 5.4% in the diabetic patients. The DMTU value of the diabetic group was significantly lower than that of the control group (p < 0.01) (Fig. 3). 
Regional Differences in DMTU: We analyzed regional difference in DMTU in the control group and the diabetic group, and the DMTU in each of the left ventricular regions in the diabetic group was significantly lower than in the control group (p < 0.01), but no significant differences were found between the DMTU value for the 4 regions in the control group or the diabetic group (Fig. 4). 
Correlation between DMTU and Rate Pressure Product (RPP) in Each Group: There was a significant positive correlation between DMTU and RPP at the exercise endpoint in the diabetic patients, but not in the control group. In the diabetic group, DMTU was positively correlated with heart rate during exercise, but in the control group no significant correlation between the two was found (Fig. 5). 
Correlation between DMTU and Parameters of Diabetic Control: There were no significant correlations between DMTU and serum HbA1c levels, fasting blood sugar levels or insulinogenic index (DIRI/DBS) (Fig. 6).
 In the diabetic patients, DMTU was the same in patients with hypertension and without hypertension (12.7 +- 4.5% vs. 15.1 +- 5.7%, n.s.). 
Electrocardiographic changes during exercise 
There were no patients with horizontal or downsloping ST-segment depression over 1 mm during exercise in any of the groups. ST segment deviation from the resting level in lead V5 during exercise was -0.46 +- 0.50 mm in the control group and -0.48 +- 0.70 mm in the diabetic group (n.s.). 
Reproducibility of the DMTU measurements 
The measurements of the DMTU in each subject were made by two different observers. There was a significant correlation between the DMTU value measured by observer A and by observer B (r = 0.869) (Fig.7), and thus the DMTU measurement was found to be reproducible. 
Validation of DMTU with a Doppler flow wire 
We performed a preliminary study to determine if DMTU is a reliable parameter to detect change in myocardial blood flow during stress, a Doppler flow wire was inserted into the anterior descending artery (LAD) of 12 subjects with normal coronary arteries, and the peak-to-resting coronary flow ratio (coronary flow reserve) was measured during intravenous infusion of adenosine triphosphate (ATP). One day after the Doppler flow wire study, the DMTU of the area perfused by the LAD (LAD area) was detemined by ATP stress TF SPECT, and a significant positive correlation was found between the DMTU of the LAD area and the CFR (r = 0.766) (Fig. 8). 
DISCUSSION 
If the myocardial extraction rate of TF during exercise is assumed to be the same as that at rest in the present study, DMTU would seem to reflect the relative increase in the coronary blood flow/cardiac output ratio during exercise.20-24 On the other hand, our preliminary study showed that the DMTU of the LAD area was correlated positively with peak-to-resting coronary flow ratio determined with a Doppler flow wire. DMTU may therefore reflect the changes in coronary blood flow during stress. 
 Since the rate pressure product during exercise and the maximum heart rate in the two groups were similar, exercise tolerance appears to have been similar in both groups. As the rate pressure product is positively correlated with myocardial oxygen consumption,25 the workload in the diabetic group was similar to that in the control group. Our data showed that the DMTU in the diabetic group was significantly lower than in the control group. Global DMTU was 14.4 +- 5.4% in the diabetic group and 21.7 +- 8.5% in the control group. This lower DMTU value in diabetic patients may mean an attenuated coronary blood flow response to increased workload.
 The DMTU of the diabetic patients was significantly lower than that of the control patients, not only in the entire left ventricle but in each of the left ventricular regions, and there were no significant regional differences between the DMTU values in any of the 4 left ventricular regions: anterior, septum, inferior or posterolateral. The low DMTU value in the diabetic group suggests diffuse impairment of myocardial blood flow, because there was no significant stenosis in any of the epicardial coronary arteries. 
 Some reports of animal experiments have suggested that coronary microcirculation is impaired in diabetes mellitus. The increase in coronary blood flow in isolated and perfused diabetic rat heart was found to be significantly lower than in the control rats during tachycardia elicited by electrical pacing.26 The coronary dilator action of adenosine has been found to be lower in the heart of diabetic lambs than in the heart of control lambs.27 A decrease in PGI2 (a potent coronary vasodilator) synthesis in response to adrenergic stimulation has been reported in the coronary arteries of diabetic dogs.28 This experimental evidence is consistent with our clinical findings. Three categories of factors influence coronary blood flow: physical factors, neural and neurohumoral factors, and metabolic factors.29 Since aortic pressure, and workload during exercise of the two groups were similar, the physical and metabolic factors in the diabetic and control groups appear to have been similar, but the difference between the neural and neurohumoral factors of the two groups in our study remained to be determined.
 There was significant positive correlation between DMTU and the hemodynamic parameters (RPP and maximum heart rate) in diabetic patients, but not in the control group, and this can be explained as follows. Because of the ample capacity for coronary dilation in control patients, high coronary blood flow was achieved during exercise. At high coronary blood flow, however, myocardial uptake of TF reaches a plateau,30-32 so that there was no significant correlation between DMTU and the hemodynamic parameters in the control group. In the diabetic group, on the other hand, the capacity for coronary dilation is limited, and myocardial uptake of TF does not reach a plateau during exercise. As a result, there was a close correlation between DMTU and the hemodynamic parameters. 
 There was no significant correlation between DMTU in the diabetic group and the clinical indicators of diabetic control (fasting blood sugar, serum HbA1c and the insulinogenic index) or the presence of diabetic complications. DMTU may therefore be an independent parameter that reflects impairment of coronary microcirculation in diabetic patients. 
 Nitenburg et al. reported, by using a Doppler guidewire in 11 diabetic patients, that maximum coronary flow reserve was reduced,7 but 10 of the 11 patients had systemic hypertension, which affects coronary microcirculation, so that, they could not rule out the possibility of hypertensive modifications. Although the incidence of systemic hypertension was higher in the diabetic patients than in the control patients in our study, there was no significant difference in DMTU according to whether the diabetic patients had hypertension or not. Therefore, the low DMTU value in the diabetic group was not attributable to hypertension, but to the diabetes itself. 
 In addition, there were no significant differences between the diabetic group and the control group in the serum lipid levels, body mass index or the frequency of obesity. 
 There were no ischemic changes in the ECG of the diabetic group or in the control group either at rest or during exercise. Although there was no significant coronary stenosis in either group, this evidence does not rule out myocardial ischemia in the diabetic patients, because they may have had slight myocardial ischemia that was not detected by the ECG.33
 The pathogenesis of the impaired microvascular function in diabetics remains uncertain. Nevertheless, there is evidence that suggests impaired coronary blood flow in diabetes mellitus because of both structural and functional changes in coronary microvessels,1-7,13,14,26-28 and recent reports have suggested coronary endothelial dysfunction in diabetics.7,34,35 In this study, visual and quantitative analysis of the RI examination non-invasively revealed changes in coronary microcirculation in diabetes mellitus. 
 A recent study by means of positron emission tomography showed that hyperglycemia is related to coronary flow reserve in non-insulin dependent diabetes mellitus.13 It would therefore be important to control blood glucose levels to prevent deterioration in the function of coronary microcirculation in diabetic patients. 
Limitations 
When the presence of coronary epicardial stenosis is unknown, it is impossible to evaluate the coronary microvascular function in diabetic patients with DMTU, and the parameter may be influenced by the presence of coronary microvascular abnormalities due to other diseases such as hypertrophied heart and syndrome X. 
 DMTU values showed a substantial overlap between the normal group and the diabetic group, and this may be explained by the fact that the DMTU value is influenced by biological tracer washout, the extraction fraction of the tracer, gamma ray absorption by the body and cardiac output. 
Clinical implications 
With DMTU it was possible to evaluate impaired coronary microvascular function in diabetics. This parameter may allow assessment of impaired coronary microvascular function in other diseases such as hypertrophied heart and syndrome X. 
 In spite of the limitations as mentioned above, DMTU can be easily obtained without using invasive techniques. We selected exercise instead of pharmacologic stress in the present study because we think that it is clinically important to assess the response of the coronary microcirculation to exercise in diabetics. In this way, DMTU may facilitate early detection of a subset of patients with abnormalities in coronary microcirculation who are prone to myocardial ischemia and left ventricular dysfunction. DMTU will also make it possible to evaluate the effects of drugs on coronary microvascular function non-invasively . 
CONCLUSIONS 
The increase in myocardial tetrofosmin uptake during exercise was quantified to evaluate impaired coronary microvascular function in diabetics. With this parameter, it was possible to non-invasively assess impaired coronary microvascular function on exercise in diabetic patients. 
ACKNOWLEDGMENTS 
We thank Professor Tadaaki Iwasaki, First Department of Internal Medicine, Hyogo College of Medicine, for his valuable and helpful suggestions. 
REFERENCES 
1. Factor SM, Okun EM, Minase T. Capillary microaneurysms in the human diabetic heart. N Engl J Med 302: 384-388, 1980 
2. Sutherland CG, Fisher BM, Frier BM, Dargie HJ, Lindop GBM. Endomyocardial biopsy pathology in insulin-dependent patients with abnormal ventricular function. Histopathology 14: 593-602, 1988. 
3. Fein FS, Sonnenblick EH. Diabetic cardiomyopathy. Prog Cardiovasc Dis 27: 255-270, 1985. 
4. Zarich SW, Nesto RW. Diabetic cardiomyopathy. Am Heart J 118: 1000-1012, 1989. 
5. Yarom R, Zirkin H, Stammler G, Rose AG. Human coronary microvessels in diabetes and ischaemia. Morphometric study of autopsy material. J Pathol 166: 265-270, 1992. 
6. Nahser PJ Jr, Brown RE, Oskarsson H, Winniford MD, Rossen JD. Maximal coronary flow reserve and metabolic coronary vasodilation in patients with diabetes mellitus. Circulation 91: 635-640, 1995. 
7. Nitenberg A, Valensi P, Sachs R, Dali M, Aptecar E, Attali JR. Impairment of coronary vascular reserve and Ach-induced coronary vasodilatation in diabetic patients with angiographically normal coronary arteries and normal left ventricular systolic function. Diabetes 42: 1017-1025, 1993. 
8. Marcus M, Wright C, Doty D, Eastham C, Laughlin D, Krumm P, et al. Measurements of coronary velocity and reactive hyperemia in the coronary circulation of humans. Circ Res 49: 877-891, 1981. 
9. Wilson RF, Laughlin DE, Ackel PH, Chilian WM, Holida MD, Hartley CJ, et al. Transluminal, subselective measurement of coronary artery blood flow velocity and vasodilator reserve in man. Circulation 72: 82-92, 1985. 
10. Doucett JW, Corl PD, Payne HM, Flynn AE, Goto M, Nassi M, et al. Validation of a Doppler guide wire for intravascular measurement of coronary artery flow velocity. Circulation 85: 1899-1911, 1992. 
11. Ofili EO, Labovitz AJ, Kern MJ. Coronary flow velocity dynamics in normal and diseased arteries Am J Cadiol 71: 3D-9D, 1993. 
12. White CW. Clinical applications of Doppler coronary flow reserve measurements. Am J Cardiol 71: 10D-16D, 1993. 
13. Yokoyama I, Ohtake T, Momomura S, Yonekura K, Woo-Soo S, Nishikawa J, et al. Hyperglycemia rather than insulin resistance is related to reduced coroanry flow reserve in NIDDM. Diabetes 47: 119-124, 1998. 
14. Pikanen OP, Nuutila P, Raitakari OT, Ronnemaa T, Koskinen PJ, Iida H, et al. Coronary flow reserve is reduced in young men with IDDM. Diabetes 47: 248-254, 1998. 
15. Davidson CJ, Fishman RF, Bonow RO. Cardiac catheterization. In Heart Disease, Braunwald E (ed.), 5th ed., Philadelphia, W.B. Saunders, Inc., pp, 199-200, 1997. 
16. Opherk D, Mall G, Zebe H, Scwarz F, Weihe E, Manthey J, et al. Reduction of coronary reserve: a mechanism for angina pectoris in patients with arterial hypertension and normal coronary arteries. Circulation 69 (1): 1-7, 1984. 
17. Lucarini AR, Picano E, Salvetti A. Coronary microvascular disease in hypertensives. Clin Exp Hypertens A14: 55-66, 1992. 
18. Cannon RO 3d, Watson RM, Rosing DR, Epstein SE. Angina caused by reduced vasodilator reserve of the small coronary arteries. J Am Coll Cardiol 1: 1359-1373, 1983. 
19. Legrand V, Hodgson JM, Bates ER, Aueron FM, Mancini GB, Smith JS, et al. Abnormal coronary flow reserve and abnormal radionuclide exercise test results in patients with normal coronary angiograms. J Am Coll Cardiol 6: 1245-1253, 1985. 
20. Kinoshita N, Sugihara H, Nakamura T, Ito K, Azuma A, Maeda T, et al. A calculation of an index of the coronary flow reserve using exercise and rest myocardial SPECT with technetium-99m tetrofosmin. KAKU IGAKU (Jpn J Nucl Med) 34: 45-48, 1997. 
21. Hori A, Taki J, Nakajima K, Shimizu M, Tonami N. Evaluation of ischemic heart disease using the index of relative 99mTc-tetrofosmin uptake increase at exercise. KAKU IGAKU (Jpn J Nucl Med) 34: 433-441, 1997. 
22. Takahashi K, Takeishi Y, Fujiwara S, Atsumi H, Akutsu T, Komatani A, et al. Quantitative assessment of an increase of myocardial 99mTc-MIBI accumulation during exercise -usefulness of response rate-. KAKU IGAKU (Jpn J Nucl Med) 33: 779-784, 1996. 
23. Fujiwara S, Takeishi Y, Atsumi H, Chiba J, Takahashi K, Tomoike H. Quantitative assessment of myocardial 99mTc-sestamibi uptake during exercise: usefulness of response rate for assessing severity of coronary artery disease. Jpn Circ J 62: 592-598, 1998. 
24. Buell U, Kaiser HJ, Dupont F, Uebis R, Kleinhans E, Hanrath P. Methoxyisobutylisonitrile (MIBI) Tc 99m SPECT to establish a correlate to coronary flow reserve. the perfusion reserve, from global and regional myocardial uptake after exercise and rest. Eur J Nucl Med 16: 3-9, 1990. 
25. Nelson RR, Gobel FL, Jorgensen CR, Wang K, Wang Y, Taylor HL. Hemodynamic predictors of myocardial oxygen consumption during static and dynamic exercise. Circulation 50: 1179-1189, 1974. 
26. Durante W, Sunahara FA, Sen AK. Effect of diabetes on metabolic coronary dilatation in the rat. Cardiovasc Res 23: 40-45, 1989. 
27. Downing SE, Lee JC, Weinstein EM. Coronary dilator actions of adenosine and CO2 in experimental diabetes. Am J Physiol 243: H252-H258, 1982. 
28. Koltai MZ, Roesen P, Hadhazy P, Ballagi-Podany G, Koszeghy A, Pogatsa G. Effects of hypoxia and adrenergic stimulation induced alterations in PGI2 synthesis by diabetic coronary arteries. J Diabet Complications 2: 5-7, 1988. 
29. Berne RM, Levy MN. Coronary circulation and cardiac metabolism. In Cardiovascular Physiology, Berne RM, Levy MN (eds.), 2nd ed., Saint Louis. The C.V. Mosby company, Inc., pp. 206-217, 1972. 
30. Sinusas AJ, Shi Q, Saltzberg MT, Vitols P, Jain D, Wackers FJ, et al. Technetium-99m-tetrofosmin to assess myocardial blood flow: experimental validation in an intact canine model of ischemia. J Nucl Med 35: 664-671, 1994. 
31. Platts EA, North TL, Pickett RD, Kelly JD. Mechanism of uptake of technetium-tetrofosmin. I: uptake into isolated adult rat ventricular myocytes and subcellular localization. J Nucl Cardiol 2: 317-326, 1995. 
32. Younes A, Songadele JA, Maublant J, Platts E, Pickett R, Veyre A. Mechanism of uptake of technetium-tetrofosmin. II: uptake into isolated adult rat heart mitochondria. J Nucl Cardiol 2: 327-333, 1995. 
33. Scharmroth L. Electrophysiology and electropathology. In The Electrocardiology of Coronary Artery Disease, Scharmroth L (ed.), 1st ed., Oxford London Edinburgh Melbourne, Blackwell Scientific publications, pp. 11-24, 1977. 
34. Mahta JL. Endothelium, coronary vasodilation, and organic nitrates. Am Heart J 129 (2): 382-391, 1995. 
35. Nitenberg A, Paycha F, Ledoux S, Sachs R, Attali JR, Valensi P. Coronary artery responses to physiological stimuli are improved by deferoxamine but not by L-arginine in non-insulin-dependent diabetic patients with angiographically normal coronary arteries and no other risk factors. Circulation 97: 736-743, 1998.