Maximal Coronary Flow Reserve and Metabolic Coronary Vasodilation in Patients With Diabetes Mellitus
Background Structural and functional abnormalities of the coronary microcirculation have been reported in experimental diabetes mellitus. The purpose of this study was to evaluate coronary microvascular function in human diabetes.
Methods and Results Twenty-four diabetic and 31 nondiabetic patients were studied during cardiac catheterization. A Doppler catheter or guidewire was used to measure changes in coronary blood flow velocity in a nonstenotic artery. Maximal coronary blood flow reserve was determined by using intracoronary adenosine or papaverine. Coronary dilation in response to an increase in myocardial metabolic demand was assessed by using rapid atrial pacing. Maximal vasodilator responses to papaverine and adenosine were compared in 12 diabetic patients. Maximal pharmacologic coronary flow reserve was depressed in diabetic (2.8±0.2, n=19) compared with nondiabetic (3.7±0.2, n=21, P<.001) patients. During atrial pacing, the decrease in coronary vascular resistance was attenuated in the diabetic (−14±3%) compared with the nondiabetic (−24±2%, P<.05) patients. Differences in coronary microvascular function between diabetic and nondiabetic patients were not attributable to differences in drug therapy, resting hemodynamics, or incidence of hypertension. In 12 diabetic patients the maximal coronary vasodilator responses to papaverine and adenosine were similar.
Conclusions This study demonstrates both reduced maximal coronary vasodilation and impairment in the regulation of coronary flow in response to submaximal increases in myocardial demand in patients with diabetes mellitus. These microvascular abnormalities may lead to myocardial ischemia in the absence of epicardial coronary atherosclerosis in some circumstances, and thus contribute to adverse cardiovascular events in diabetic patients.
Abnormalities of the coronary microcirculation distinct from large-vessel atherosclerosis have been reported in clinical and experimental diabetes mellitus.1 Morphological studies have described abnormalities in resistance vessels in diabetic patients and animals,2 3 4 and functional abnormalities of the coronary circulation of diabetic animals have also been reported. In isolated perfused hearts from diabetic rats, coronary vasodilation in response to norepinephrine, calcium, and paced tachycardia was impaired compared with nondiabetic rats.5 In diabetic lambs, resting coronary blood flow was reduced over a wide range of aortic pressures compared with controls.6 The peak reactive hyperemic response after brief coronary occlusion was attenuated in anesthetized diabetic dogs compared with controls.7 Basal production of thromboxane by coronary artery rings was increased8 and production of prostacyclin by coronary rings during hypoxia and α-adrenergic stimulation was impaired in diabetic rats.9 Attenuation of coronary vasodilation to adenosine in diabetic animals has been reported.7 10
Few data on coronary microvascular function in diabetic patients are available. Nitenberg et al11 reported reduced maximal coronary blood flow reserve and impaired endothelial-dependent epicardial coronary vasodilation in a study of 11 diabetic patients. However, since all but one diabetic patient had systemic hypertension in this study, attribution of the abnormalities to diabetes per se was uncertain.
The purpose of the present study was to evaluate coronary microvascular function in patients with diabetes mellitus. The coronary vasodilator response to increased myocardial oxygen demand was evaluated by using atrial pacing stress. Maximal vasodilation was assessed by measurement of pharmacologic coronary flow reserve, similar to the study of Nitenberg et al.11 Finally, we compared maximal vasodilator responses to adenosine and papaverine in diabetic patients to determine if resistance to the vasodilator effects of adenosine is present in diabetics.
Patients undergoing elective coronary arteriography for routine clinical indications were considered for study if angiography revealed a nonstenotic coronary artery that was anatomically suitable for placement of a coronary Doppler catheter, and there was no evidence of left ventricular or valvular dysfunction. Patients were considered to have diabetes mellitus if medical record review revealed fasting hyperglycemia and drug (insulin or oral hypoglycemic agents) or dietary treatment for at least 1 year before catheterization. Patients with hyperglycemia by history only or hyperglycemia for less than 1 year were not enrolled. The research protocol was approved by the University of Iowa Institutional Review Board, and written informed consent for the research protocol was obtained from each patient before cardiac catheterization.
Coronary Flow Velocity Measurement
Changes in coronary blood flow were assessed by measurement of flow velocity changes using an intracoronary Doppler catheter12 or Doppler-tipped angioplasty guidewire.13 In studies performed with the Doppler catheter, a 7F or 8F coronary angioplasty guiding catheter was positioned at the coronary ostium and a 0.014-in. coronary angioplasty guidewire was advanced into the coronary artery to be studied. A 3F 20-mHz coronary Doppler catheter (NuMed Inc) was advanced over the guidewire into the proximal vessel and positioned to obtain a high-quality phasic signal of blood flow velocity. The pulsed Doppler meter (Bioengineering Department, University of Iowa Hospitals and Clinics) was range-gated to maximize the amplitude of the mean coronary blood flow velocity signal. In studies performed with the Doppler-tipped angioplasty guidewire, a 7F coronary angioplasty guiding catheter was positioned at the coronary ostium, and the 0.018-in. 12-mHz Doppler wire (Cardiometrics, Inc) was advanced into the coronary artery to be studied. Audio signals from the Doppler guidewire were processed with a system that incorporates a real-time spectrum analyzer to provide a scrolling gray-scale spectral display as well as determination of instantaneous spectral-peak velocity and the time average of spectral-peak velocity. Phasic and mean coronary blood flow velocity signals, mean arterial pressure (in millimeters of mercury) from the guiding catheter, heart rate, and the surface ECG were continuously recorded on a multichannel recorder.
Subjects were brought to the cardiac catheterization laboratory in a fasting state. Oral hypoglycemic agents were withheld, and the insulin dose was reduced by one half in patients receiving these medications on the day of the research study. Prescribed cardiac medications were continued on the day of the study. Diazepam (5 to 10 mg IV or PO) was given for sedation. No patient received atropine premedication. The study was performed during infusion of nitroglycerin at 8 μg/min IV or 5 to 10 minutes after intracoronary administration of 200 μg nitroglycerin to prevent catheter-induced coronary artery spasm and avoid changes in coronary artery caliber that would influence the relation between changes in coronary flow velocity and volumetric coronary blood flow.
Measurement of Pharmacologic Coronary Flow Reserve
After measurement of resting coronary blood flow velocity, a bolus dose of 6 to 10 mg papaverine hydrochloride14 (2 mg/mL 0.9% saline) or 8 to 12 μg adenosine15 (4 μg/mL 0.9% saline) was injected through the guiding catheter into the coronary ostium, and the resultant increase in coronary blood flow velocity was recorded. To confirm that maximal hyperemia was produced, coronary blood flow velocity was recorded during administration of an additional larger dose of papaverine (2 to 4 mg larger than the initial dose) or adenosine (4 μg larger than the initial dose). Flow velocity was allowed to return to control levels between drug doses. Coronary flow reserve was calculated as the quotient of the peak mean flow velocity (expressed in units proportional to the Doppler kHz shift) after vasodilator and the control mean flow velocity during the 15 to 30 seconds preceding vasodilator administration. Coronary flow reserve was measured in 12 randomly selected diabetic patients sequentially with both papaverine and adenosine using the protocol described above. In these patients, the order of drug administration was also randomly selected.
Assessment of Metabolic Coronary Vasodilation
Metabolic coronary vasodilation was assessed by measurement of the coronary hemodynamic response to rapid atrial pacing. Atrial pacing was performed with a 6F bipolar pacemaker positioned at the high right atrium. The paced rate was increased over 1 to 2 minutes until a 45 to 50 beats per minute increment over the sinus rate was reached or atrioventricular block developed. Coronary blood flow velocity and mean arterial pressure were measured during sinus rhythm and when the flow signal stabilized after at least 2 minutes of pacing at the peak rate. An index of coronary vascular resistance was calculated as the quotient of mean arterial pressure (in millimeters of mercury) and coronary flow velocity expressed in units proportional to the Doppler kHz shift. Alterations in coronary vascular resistance during pacing were expressed as percent change from the control value. Despite adherence to a standardized protocol, pacing resulted in a wide range of metabolic stress as assessed by changes in the product of heart rate and mean arterial pressure (rate-pressure product). To control for this, an index of coronary vasodilation relative to magnitude of metabolic stress was calculated as the quotient of percent change in coronary vascular resistance and percent change in rate-pressure product during pacing.
Group data are reported as mean±SEM. Continuous variables were compared by using Student’s t test for paired or unpaired data as appropriate. Discrete variables were analyzed by χ2 test. In assessment of pharmacologic coronary flow reserve, multivariate stepwise regression analysis was used to examine the influence of the following variables: diabetes status, history of hypertension, gender, age, β-adrenergic antagonist use, glucose level, control heart rate, control mean arterial pressure, and control rate-pressure product. In assessment of metabolically mediated coronary vasodilation, multivariate stepwise regression analysis was used to examine the influence of the following variables: diabetes status, history of hypertension, gender, age, β-adrenergic antagonist use, glucose level, control heart rate, percent change in heart rate during pacing, control mean arterial pressure, percent change in mean arterial pressure during pacing, control rate-pressure product, and percent change in rate-pressure product during pacing. Differences were considered significant at the P≤.05 level.
The ages of the 24 diabetic and 31 nondiabetic patients were similar (52±3 versus 53±2 years; Table⇓). The blood glucose level on the day of catheterization averaged 191±19 mg percent in the diabetics and 106±4 mg percent in the nondiabetics (P<.01). Women comprised 61% of the diabetics and 42% of the nondiabetics (P=NS). A history of hypertension was present more frequently in diabetic than nondiabetic patients (70% versus 39%, P<.05), although the frequency of left ventricular hypertrophy on ECG was low and similar in diabetics and nondiabetics (9% versus 10%, P=NS). Diabetics were taking β-adrenergic antagonists at the time of the study less often than nondiabetics (9% versus 35%, P<.05). Significant coronary obstruction (diameter stenosis >50%) was present in 17% of the diabetic patients (all one vessel) and 16% of the nondiabetics (one vessel in 4, two vessels in 1). In all patients but 2 diabetics the serum creatinine was ≤2.0 mg/dL.
Effect of Diabetes Mellitus on Pharmacologic Coronary Flow Reserve
Maximal coronary flow reserve was measured using intracoronary papaverine or adenosine in 19 diabetic and 22 nondiabetic patients. The heart rate averaged 77±3 bpm in the diabetic and 69±3 bpm in the nondiabetic (P=.05) patients. Mean arterial pressure averaged 100±3 mm Hg in the diabetics and 95±3 mm Hg in the nondiabetics (P=NS). The rate-pressure product was higher in the diabetic than in the nondiabetic (7731±443 versus 6549±365 bpm×mm Hg, P<.05) patients.
Coronary flow reserve was lower in the diabetics than in the nondiabetics (2.8±0.2 versus 3.7±0.2, P<.001; Fig 1⇓).
Since β-adrenergic antagonists were taken more frequently by nondiabetic than diabetic patients and coronary flow reserve was higher in patients receiving these agents (β-antagonist, 3.9±0.4; no β-antagonist, 3.1±0.2, P<.05), the analysis was repeated after excluding patients receiving β-antagonists. After this exclusion, a reduction in coronary flow reserve in diabetic (n=17) compared with nondiabetic (n=14) patients was still observed (2.8±0.2 versus 3.5±0.2, P<.05). The diagnosis of hypertension was not associated with a reduction in coronary flow reserve in the patients studied (hypertension, 3.2±0.3; no hypertension, 3.4±0.2, P=NS). Coronary flow reserve was reduced in the diabetic patients without hypertension (2.7±0.2, n=4) compared with the nondiabetic patients without hypertension (3.6±0.2, n=12, P<.05).
Multivariate stepwise regression analysis revealed that diabetic status (P=.02), age (P=.01), and control heart rate (P=.05) were independently related to coronary flow reserve. The relation of β-antagonist therapy with coronary flow reserve was of marginal significance (P=.07).
Effect of Diabetes Mellitus on Metabolic Coronary Vasodilation
The coronary hemodynamic response to the standardized atrial pacing protocol was measured in 24 diabetic and 31 nondiabetic patients. The control heart rate averaged 79±2 bpm in the diabetic and 73±2 bpm in the nondiabetic (P=.06) patients. The control mean arterial pressure was 102±3 mm Hg in the diabetics and 96±2 mm Hg in the nondiabetics (P=.10). The control rate-pressure product was higher in the diabetic than the nondiabetic (8087±385 versus 6994±296 bpm×mm Hg, P<.05) patients.
By experimental design, the increase in heart rate during atrial pacing was similar in diabetic and nondiabetic (45±1 versus 48±2 bpm, P=NS) patients. During atrial pacing in diabetics, mean arterial pressure was unchanged from control (−0.5±1 mm Hg, P=NS), whereas in nondiabetics, mean arterial pressure increased by 4±1 mm Hg (P<.01). While the absolute increase in rate-pressure product in diabetics and nondiabetics was similar (4506±226 versus 5089±245 bpm×mm Hg, P=NS), the increase relative to the control rate-pressure product was smaller in the diabetic than nondiabetic (59±3% versus 77±5%, P<.01) patients.
The reduction in coronary vascular resistance during pacing was smaller in diabetics than in nondiabetics. Coronary vascular resistance decreased by 14±3% in the diabetic patients and by 24±2% in the nondiabetics (P<.05, Fig 2⇓). Impairment in coronary vasodilation during pacing in diabetics was observed after controlling for the variation in the magnitude of metabolic stress during pacing; the quotient of percent change in coronary vascular resistance and percent change in rate-pressure product was −0.22±0.05 in the diabetic and −0.34±0.03 in the nondiabetic (P<.05, Fig 2⇓) patients.
β-Adrenergic antagonist therapy did not influence the reduction in coronary resistance during pacing (β-antagonist, −21±3%; no β-antagonist, −19±2%, P=NS). In the subset of 21 diabetic and 19 nondiabetic patients not receiving β-antagonist therapy, the reduction in coronary resistance remained attenuated in the diabetic compared with the nondiabetic (−14±3% versus −23±3%, P<.05) group. The diagnosis of hypertension did not influence the reduction in coronary resistance during pacing (hypertension, −18±2%; no hypertension, −21±3%, P=NS). In the subset of 7 diabetic and 19 nondiabetic patients without hypertension, the reduction in coronary resistance remained blunted in the diabetic compared with the nondiabetic (−12±7% versus −24±2%, P=.05) group.
Multivariate stepwise regression analysis revealed that diabetic status (P=.06) and control rate-pressure product (P=.002) were independently related to the change in coronary resistance during pacing. Only diabetic status (P=.05) was independently related to the quotient of percent change in coronary vascular resistance and percent change in rate-pressure product.
Effect of Hypertension on Pharmacologic and Metabolic Coronary Flow Responses in Diabetic Patients
Of the 24 diabetic patients studied, 17 also had systemic hypertension. At the time of study, the control heart rate (hypertension, 78±3 bpm; normotension, 80±5 bpm, P=NS), control mean arterial pressure (hypertension, 101±4 mm Hg; normotension, 102±4, P=NS), and percent change in rate-mean pressure product during pacing (hypertension, 59±4%; normotension, 57±8%, P=NS) were similar in the diabetic hypertensive and diabetic normotensive patients. Pharmacologic coronary flow reserve was 2.8±0.3 in the diabetic hypertensive and 2.7±0.2 in the diabetic normotensive patients (P=NS), and coronary vascular resistance during pacing decreased by 15±3% in the diabetic hypertensive patients and by 12±7% in the diabetic normotensive patients (P=NS).
Maximal Coronary Flow Responses to Papaverine and Adenosine in Diabetic Patients
Coronary vasodilator responses to papaverine and adenosine as assessed by maximal coronary flow reserve were comparable in the 12 diabetic patients who received both drugs (papaverine, 2.7±0.3; adenosine, 2.8±0.3, P=NS).
Pharmacologic Flow Reserve and Metabolic Vasodilation in Diabetic Patients Receiving Oral Hypoglycemic Agents
Of the 24 patients with diabetes, 12 were receiving sulfonylurea oral hypoglycemic agents at the time of the study. Maximal pharmacologic flow reserve was 2.8±0.2 (n=10) in diabetics not receiving and 2.7±0.4 (n=9) in those receiving oral hypoglycemic agents (P=NS). The decreases in coronary vascular resistance during atrial pacing were identical (14±4%) in both diabetics receiving (n=12) and those not receiving (n=12) oral hypoglycemic agents.
This study demonstrates impairment of coronary microvascular function in patients with diabetes mellitus in the absence of obstructive coronary atherosclerosis. Both maximal pharmacologic coronary flow reserve and metabolic vasodilation were assessed, since these responses may reflect different properties of the coronary microcirculation. Coronary flow reserve assessed using pharmacologic dilation is a luxuriant response occurring without regard to myocardial substrate demand. Pharmacologic coronary flow reserve is commonly measured with drugs (eg, papaverine) that have no demonstrated relation to physiological mechanisms of coronary flow regulation. Pharmacologic coronary flow reserve is measured by using progressively increasing drug doses until a peak flow response is reached, whereas metabolic vasodilation as assessed in our protocol reflects submaximal vasodilation in response to a standardized physiological stress. Pharmacologic coronary flow reserve and metabolic coronary vasodilation were both attenuated in diabetic patients. The impairment in pharmacologic coronary flow reserve concurs with the study of Nitenberg et al11 in a smaller patient population.
The mechanisms leading to impairment in coronary vasodilation are uncertain. Histopathologic studies in diabetic patients describe abnormalities in coronary resistance vessels including arteriolar thickening, perivascular accumulations of connective tissue, and capillary microaneurysms.2 16 Capillary density is reduced in hearts of diabetic patients sustaining myocardial infarction.3 Microvascular morphological abnormalities have also been reported in diabetic animals.4 However, structural abnormalities of the coronary microvascular circulation in diabetes have not been universally observed.17 18
The reduction in metabolic coronary vasodilation during atrial pacing in diabetics may be related to dysfunction of the coronary endothelium. Impairment in endothelium-dependent vasodilation related to diabetes has been described in animals19 20 21 as well as the human forearm22 23 and corpus cavernosum.24 Epicardial coronary artery constriction to acetylcholine is augmented in diabetic patients compared with control subjects.11 Furthermore, the magnitude of coronary vasodilation during pacing stress is related to endothelium-dependent vasodilation as assessed by the flow response to intracoronary acetylcholine in nondiabetic patients with chest pain and angiographically normal coronary arteries.25 Attenuation of endothelium-dependent vasodilation related to increased production of vasoconstrictor prostaglandins has been reported in diabetic rabbit aorta,20 suggesting that enhanced vessel wall production of vasoconstrictor substances could also blunt coronary vasodilation in human diabetics. Finally, increased sensitivity to the vasoconstrictor effects of α-adrenergic activation has been reported in experimental diabetes,26 27 28 which could limit physiological29 and pharmacologic vasodilation. Additional studies in diabetic patients evaluating the relation of endothelium-dependent vasomotion and metabolic vasodilation, the influence of glycemic control on microvascular function, adrenergic influences on microvascular dilation, and the relation between coronary microvascular function and diabetic microvascular disease in other organs are needed.
Coronary vasodilation in response to exogenous adenosine is impaired in some animal models of diabetes.7 10 Given the potential role of adenosine in the regulation of coronary blood flow in response to changing oxygen demand, diminished vascular responsiveness to adenosine has been proposed as a link between impaired metabolic vasodilation and diabetes mellitus. However, we found equivalent maximal vasodilation with adenosine and papaverine, suggesting that impairment of metabolic vasodilation in human diabetes is not related to selective resistance to adenosine.
Sulfonylurea oral hypoglycemic agents are antagonists of the ATP-sensitive potassium channel. Basal coronary resistance is increased,30 and hyperemia in response to ischemia is attenuated,31 32 by the oral hypoglycemic agent glibenclamide in dogs. The effects of oral hypoglycemic agents on coronary reactivity in experimental animals and the differences in cardiovascular morbidity between patients treated with insulin and the sulfonylurea tolbutamide in a large multicenter trial33 have led to concern regarding the coronary effects of these agents in diabetic patients. In the present study, no differences in pharmacologic or metabolic coronary vasodilation were observed in diabetic patients treated with oral hypoglycemic agents compared with those receiving other therapy.
While morphological studies in rats demonstrate more marked coronary microvascular abnormalities in hypertensive diabetic animals compared with those with hypertension or diabetes alone,18 we observed no additional impairment in microvascular function in diabetic hypertensive compared with diabetic normotensive patients.
In the present study, coronary blood flow responses were assessed by using intracoronary Doppler measurements of coronary flow velocity. Although extensive animal studies have validated the accuracy of Doppler measurements in the assessment of changes in coronary flow,12 13 the technique is not capable of measuring absolute myocardial perfusion. Thus, it is uncertain if the impairment in vasodilation in diabetics is due to elevation of resting myocardial perfusion, a reduction in perfusion during the pharmacologic and metabolic stimuli, or both. However, limited data in diabetic patients indicate that resting coronary blood flow measured by coronary sinus thermodilution is similar in nondiabetic control subjects.34
In our study, the heart rate and arterial pressure in patients with diabetes tended to be higher than in nondiabetics. We35 and other investigators36 have observed that maximal coronary flow reserve is progressively reduced with increases in heart rate but is not influenced by acute increases in arterial pressure. Multivariate regression analysis in the present study revealed that the presence of diabetes was associated with reduced coronary flow reserve independent of the differences in heart rate or other differences between the diabetic and nondiabetic patients.
The diabetic patients enrolled in this study were relatively free of renal and other end-organ complications. The severity of coronary microvascular dysfunction in diabetes may parallel the development of other complications attributable to microangiopathy. Thus, patients with diabetes that is more advanced than in our subjects may have a greater impairment in coronary vasodilator responses.
Our studies provide evidence for both reduced maximal coronary vasodilation and impairment in the regulation of coronary flow in response to submaximal increases in myocardial demand in patients with diabetes mellitus. It is possible that microvascular abnormalities could lead to myocardial ischemia in the absence of epicardial coronary atherosclerosis in some circumstances, and could thus contribute to chronic left ventricular dysfunction and other adverse cardiovascular events in diabetic patients.
This study was supported in part by National Heart, Lung, and Blood Institute SCOR in Coronary and Vascular Diseases (HL 32295). We thank the personnel of the Cardiac Catheterization Laboratory of the University of Iowa Hospitals and the Iowa City Veterans Affairs Medical Center for their technical assistance and Bette Brant for secretarial assistance.
- Received July 20, 1994.
- Accepted August 31, 1994.
- Copyright © 1995 by American Heart Association
Factor SM, Segal BH, van Hoeven KH. Diabetes and coronary artery disease. Coron Artery Dis. 1992;3:4-10.
Lee JC, Downing SE. Coronary dynamics and myocardial metabolism in the diabetic newborn lamb. Am J Physiol. 1979;237:H118-H124.
Downing SE, Lee JC, Weinstein EM. Coronary dilator actions of adenosine and CO2 in experimental diabetes. Am J Physiol. 1982;243:H252-H258.
Nitenberg A, Valensi P, Sachs R, Dali M, Aptecar E, Attali JR. Impairment of coronary vascular reserve and ACh-induced coronary vasodilation in diabetic patients with angiographically normal coronary arteries and normal left ventricular systolic function. Diabetes. 1993;42:1017-1025.
Wilson RF, Laughlin DE, Ackell PH, Chilian WM, Holida MD, Hartley CJ, Armstrong ML, Marcus ML, White CW. Transluminal, subselective measurement of coronary artery blood flow velocity and vasodilator reserve in man. Circulation. 1985;72:82-92.
Doucette JW, Corl PD, Payne HM, Flynn AE, Goto M, Nassi M, Segal J. Validation of a Doppler guide wire for intravascular measurement of coronary artery flow velocity. Circulation. 1994;85:1899-1911.
Wilson RF, White CW. Intracoronary papaverine: an ideal coronary vasodilator for studies of the coronary circulation in conscious humans. Circulation. 1986;73:444-451.
Wilson RF, Wyche K, Christensen BV, Zimmer S, Laxson DD. Effects of adenosine on human coronary arterial circulation. Circulation. 1990;82:1595-1606.
Meraji S, Jayakody L, Senaratme P, Thomson ABR, Kappagoda T. Endothelium-dependent relaxation in aorta of BB rats. Diabetes. 1987;36:978-981.
Tesfamariam B, Jakubowski JA, Cohen RA. Contraction of diabetic rabbit aorta caused by endothelium-derived PGH2-TxA2. Am J Physiol. 1989;257:H1327-H1333.
Fujii K, Heistad DD, Faraci FM. Effect of diabetes mellitus on flow-mediated and endothelium-dependent dilatation of the rat basilar artery. Stroke. 1992;23:1494-1498.
Johnstone MT, Creager SJ, Scales KM, Lee BK, Creager MA. Impaired endothelium-dependent vasodilation in patient with insulin-dependent diabetes mellitus. Circulation. 1993;88:2510-2516.
Calver A, Collier J, Vallance P. Inhibition and stimulation of nitric oxide synthesis in the human forearm arterial bed of patients with insulin-dependent diabetes mellitus. J Clin Invest. 1992; 90:2548-2554.
Quyyumi AA, Cannon RO, Panza JA, Diodati JG, Epstein SE. Endothelial dysfunction in patients with chest pain and normal coronary arteries. Circulation. 1992;86:1864-1871.
Downing SE, Lee JC, Fripp RR. Enhanced sensitivity of diabetic hearts to alpha-adrenergic stimulation. Am J Physiol. 1983;245:H808-H813.
Heusch G, Deussen A. The effect of cardiac sympathetic nerve stimulation on perfusion of stenotic coronary arteries in the dog. Circ Res. 1983;53:8-15.
Samaha FF, Heineman FW, Ince C, Fleming J, Balaban RS. ATP-sensitive potassium channel is essential to maintain basal coronary vascular tone in vivo. Am J Physiol. 1992;262:C1220-C1227.
Aversano T, Ouyang P, Silverman H. Blockade of the ATP-sensitive potassium channel modulates reactive hyperemia in the canine coronary circulation. Circ Res. 1991;69:618-622.
Kanatsuka H, Sekiguchi N, Sato K, Akai K, Wang Y, Komaru T, Ashikawa K, Takishima T. Microvascular sites and mechanisms responsible for reactive hyperemia in the coronary circulation of the beating canine heart. Circ Res. 1992;71:912-922.
The University Group Diabetes Program. A study of the effects of hypoglycemic agents on vascular complications in patients with adult-onset diabetes. Diabetes. 1970;19:747-830.
Doria A, Nosadini R, Avogaro A, Fioretto P, Crepaldi G. Myocardial metabolism in type 1 diabetic patients without coronary artery disease. Diabet Med. 1991;8(suppl):S104-S107.
McGinn AL, White CW, Wilson RF. Interstudy variability of coronary flow reserve: influence of heart rate, arterial pressure, and ventricular preload. Circulation. 1990;81:1319-1330.