Chronic Hyperglycemia Impairs Endothelial Function and Insulin Sensitivity Via Different Mechanisms in Insulin-Dependent Diabetes Mellitus
Background We explored whether chronic hyperglycemia is associated with defects in endothelium-dependent vasodilatation in vivo and whether defects in the hemodynamic effects of insulin explain insulin resistance.
Methods and Results Vasodilator responses to brachial artery infusions of acetylcholine, sodium nitroprusside, and NG-monomethyl-l-arginine and, on another occasion, in vivo insulin sensitivity (euglycemic insulin clamp combined with the forearm catheterization technique) were determined in 18 patients with insulin-dependent diabetes mellitus (IDDM) and 9 normal subjects. At identical glucose and insulin levels, insulin stimulation of whole-body and forearm glucose uptake was 57% reduced in the IDDM patients compared with normal subjects (P<.001). The defect in forearm glucose uptake was attributable to a defect in glucose extraction (glucose AV difference, 1.1±0.2 versus 1.9±0.2 mmol/L, P<.001, IDDM versus normal subjects), not blood flow. Within the group of IDDM patients, hemoglobin A1c was inversely correlated with forearm blood flow during administration of acetylcholine (r=−.50, P<.02) but not sodium nitroprusside (r=.07). The ratio of endothelium-dependent to endothelium-independent blood flow was ≈40% lower in patients with poor glycemic control than in normal subjects or patients with good or moderate glycemic control.
Conclusions We conclude that chronic hyperglycemia is associated with impaired endothelium-dependent vasodilatation in vivo and with a glucose extraction defect during insulin stimulation. These data imply that chronic hyperglycemia impairs vascular function and insulin action via distinct mechanisms. The defect in endothelium-dependent vasodilatation could contribute to the increased cardiovascular risk in diabetes.
Endothelial dysfunction, measured as the vasodilatory response of forearm resistance vessels to endothelium-dependent vasoactive agents such as ACh, characterizes patients with atherosclerotic vascular disease.1 2 Recent epidemiological studies have demonstrated that chronic hyperglycemia is an independent risk factor for coronary heart disease, but the mechanism(s) underlying this association are unclear.3 Abundant experimental evidence from animal studies implies that hyperglycemia blunts endothelium-dependent vasodilatation in isolated aortas from normal4 and diabetic5 6 animals.
Impaired relaxation of corporal smooth muscle in response to stimulation of autonomic nerves and to ACh but not to SNP, an endothelium-independent vasodilator, was found in diabetic men with impotence.7 Studies addressing endothelial vasoregulatory responses to vasoactive agents in forearm resistance vessels, however, have yielded conflicting results in diabetic patients.8 9 10 11 12 Both normal8 9 10 and blunted11 12 vasodilatory responses to ACh or carbacholine have been found in forearm resistance vessels. The response to SNP was normal in four studies8 10 11 12 and impaired in one study in patients with IDDM.9 In vitro, in arterial segments isolated from gluteal fat biopsies of patients with IDDM,12 the vasodilatory responses to bradykinin and SNP were unaltered, whereas that to ACh was blunted. The reasons for these contradictory results are unclear. Careful characterization of the study groups would seem to be the first step in resolving the discrepant findings in diabetic patients. For example, it is possible that the impaired endothelium-dependent vasorelaxation to the neurotransmitter ACh found in diabetic men with symptomatic autonomic neuropathy and impotence7 was a result of neuropathy and not hyperglycemia. Also, factors such as the ambient glucose concentration during the endothelial function tests and albuminuria13 might be confounding variables in human studies. Microalbuminuria reflects endothelial damage at least in the kidney.1 It is also a predictor of coronary heart disease and has been suggested to be associated with whole-body insulin resistance independent of glycemic control in patients with IDDM.13 Whether such resistance is due to a defect in cellular glucose extraction or blood flow is unknown.
Insulin increases forearm blood flow in normal subjects via an endothelium- and NO synthesis–dependent mechanism,14 since the flow response can be abolished by use of L-NMMA, an arginine analogue that competitively inhibits NO synthesis from arginine.15 It therefore seems feasible to postulate that endothelial dysfunction might cause insulin resistance via a vascular mechanism. This hypothesis, however, is inconsistent with our previous studies in patients with IDDM, in which we have attributed insulin resistance to chronic hyperglycemia–induced defects in glucose extraction rather than delivery under physiological conditions.16 17 On the other hand, in none of the previous studies13 16 17 have direct measurements of endothelial function and insulin sensitivity been performed in the same patients.
The present studies were undertaken to determine whether (1) chronic hyperglycemia is associated with impaired ACh-induced vasodilatory responses in patients with IDDM in vivo, (2) insulin resistance is more severe in patients with endothelial dysfunction than in those without, and (3) insulin-stimulated blood flow or the ability of cells to extract glucose underlies insulin resistance in patients with microalbuminuria.
Eighteen men with IDDM and 9 normal men volunteered for the studies. The diabetic patients were recruited from the outpatient clinic according to the following criteria: (1) age 18 to 60 years; (2) age at diagnosis of diabetes <35 years; (3) undetectable fasting C-peptide concentration (<0.1 nmol/L); and (4) no proteinuria, symptomatic neuropathy, or proliferative retinopathy. The normal subjects and the diabetic patients were matched for age, body mass index, and body composition (Table 1⇓). Histories were taken and physical examinations and laboratory tests were performed in all subjects to exclude diseases other than diabetes mellitus. All patients and normal subjects had normal blood counts, serum creatinine, electrolyte concentrations, and ECGs (data not shown). Timed overnight urine collections were taken to classify the patients according to their UAER. The diabetic patients were treated with three (n=2) or four (n=15) daily injections of a combination of intermediate- and short-acting insulins. One patient was using continuous subcutaneous insulin infusion therapy. The mean insulin doses are shown in Table 1⇓. The patients did not ingest any drugs known to affect glucose metabolism. Informed written consent was obtained after the purpose, nature, and potential risks had been explained to the subjects. The experimental protocol was designed and performed according to the principles of the Declaration of Helsinki and was approved by the Ethical Committee of the Helsinki University Central Hospital.
All subjects were studied on two separate occasions with at least a 1-week interval between them. On one occasion, whole-body and forearm insulin sensitivities were measured, and on the other, in vivo endothelial function was determined by measurement of the effects of intra-arterial infusions of endothelium-dependent and -independent vasodilators on forearm blood flow. For 2 days before the studies, the subjects ingested a weight-maintaining diet containing at least 200 g carbohydrate per day with 15% to 20%, 45% to 50%, and 35% to 40% of calories from protein, carbohydrate, and fat, respectively. All studies were started at 7:30 am after a 10- to 12-hour overnight fast during which the subjects were allowed to drink only water. Studies were started after the subjects had rested supine in a quiet environment for at least 30 minutes.
Whole-body glucose uptake
Insulin sensitivity was measured by the euglycemic insulin clamp technique.18 Three 18-gauge catheters (Venflon, Viggo-Spectramed) were inserted as previously described.19 Insulin and glucose were infused through a catheter inserted into the left antecubital vein. The left hand was kept in a heated chamber (65°C), and arterialized venous blood was withdrawn from a catheter inserted retrogradely into a heated dorsal hand vein. The deep branch of the right medial cubital vein draining forearm muscles was cannulated retrogradely so that the tip of the cannula could not be palpated superficially. Insulin (Actrapid Human, Novo Nordisk) was infused in a primed continuous fashion. The rate of the insulin infusion was 1 mU·kg−1·min−1. After the insulin infusion was started, plasma glucose was allowed to reach normoglycemia in the diabetic patients. Normoglycemia was maintained thereafter by adjusting the rate of a 20% glucose infusion on the basis of plasma glucose measurements, which were performed at 5-minute intervals. Hepatic glucose output was not measured in the present study, because it will not influence blood flow or the glucose AV difference across the forearm and because it has repeatedly been shown to be completely suppressed in both normal subjects and patients with IDDM20 at insulin concentrations similar to those in the present study. Whole-body glucose uptake rates were therefore calculated from the glucose infusion rate between 60 and 120 minutes after correction for changes in the glucose pool size.18
Forearm glucose uptake
Forearm glucose uptake was calculated by multiplication of the glucose AV difference by forearm blood flow.19 Total forearm blood flow was measured every 30 minutes during the clamp with venous occlusion plethysmography with mercury-in-rubber strain gauges (EC 4 Strain Gauge Plethysmograph, Hokanson), as previously described in detail.21 The gauge was attached around the widest, most muscular segment of the forearm.21 Two minutes before blood sampling and flow measurements, circulation to the hand was interrupted by inflation of a pediatric blood pressure cuff around the wrist to above the systolic blood pressure. Venous return was then occluded by a rapid cuff inflator (E 20 Rapid Cuff Inflator, Hokanson) by inflation of a sphygmomanometer cuff around the upper arm to 40 to 50 mm Hg. An analog-to-digital convertor (McLab/4e, AD Instruments Pty Ltd) connected to a personal computer was used for recording blood flow. At least five flow curves were recorded for each flow measurement, as previously described.21 Calibration was performed by use of the built-in electronic calibration signal for a 1% volume change, the height of which is used for blood flow calculations.
In vivo endothelial function
To avoid possible confounding effects of acute hyperglycemia on endothelial function, an insulin infusion (0.1 mU·kg−1·min−1 IV) was started at 7:30 am after an overnight fast of 10 to 12 hours to normalize the plasma glucose concentration in the diabetic patients (Table 2⇓). Normoglycemia was reached within 73±17 minutes after the insulin infusion was started. Glucose was infused if necessary to maintain normoglycemia on the basis of plasma glucose measurements performed at 20-minute intervals. Insulin was infused and blood samples were drawn through an 18-gauge (Venflon, Viggo-Spectramed) catheter inserted into the right antecubital vein. Blood flow was measured in both forearms by venous occlusion strain-gauge plethysmography as described above. The occlusion pressure was 40 to 50 mm Hg and the wrist cuff occlusion pressure, 200 mm Hg. Flows were recorded for 10 seconds every 15 seconds, and the mean of the final five measurements of each recording period was used for analysis. A 27-gauge unmounted steel cannula (Coopers Needle Works) connected to an epidural catheter (Portex) was inserted into the left brachial artery. Drugs were infused with a constant-rate infusion pump (Braun AG and Harvard Apparatus model 22). Subjects rested for 30 minutes after needle placement before blood flow measurements were begun. Normal saline was infused for 12 minutes at a rate of 1 mL/min. Drugs were then infused at the same rate (1 mL/min) in the following sequence: SNP (Roche, 3 and 10 μg/min), ACh (Iolab Corp, 7.5 and 15 μg/min), and L-NMMA (Clinalfa AG, 4 μmol/min). The doses of ACh and SNP were chosen because both the lower (3 and 7.5 μg/min) and higher (10 and 15 μg/min) doses produce similar increases in blood flow in normal subjects, and these doses are associated with impaired vasodilatation in hypercholesterolemic subjects predisposed to atherosclerosis.2 The dose of L-NMMA was chosen because this dose induces maximal inhibition of basal NO synthesis–dependent vasodilatation in normal subjects.16 Each dose was infused for 6 minutes, and the infusion of each drug was separated by infusion of normal saline for 18 minutes, during which time blood flow was allowed to return to basal values (Fig 1⇓).
Fat-free mass was measured by a single-frequency bioelectrical impedance device (model BIA-101A, Bio-Electrical Impedance Analyzer System).22
Plasma glucose concentrations were measured in duplicate by the glucose oxidase method with a Beckman Glucose Analyzer II (Beckman Instruments). Serum free insulin concentrations were determined every 30 minutes during the clamp by double-antibody radioimmunoassay (Pharmacia Insulin RIA kit) after precipitation with polyethylene glycol. HbA1c was measured by high-performance liquid chromatography with the fully automated Glycosylated Hemoglobin Analyzer System (BioRad).
Data between the study groups were analyzed by ANOVA followed by pairwise comparison with Fisher's least significant difference test. Simple correlations between selected study variables were calculated by Spearman's rank correlation coefficient. Multiple linear regression analysis was used to analyze the causes of variation in parameters of insulin sensitivity and endothelial function. UAER was logarithmically transformed to normalize its distribution for multiple linear regression analysis. All calculations were made with the SYSTAT statistical package (SYSTAT Inc). All data are expressed as mean±SEM.
Fasting plasma glucose and insulin concentrations are shown in Table 2⇑. During the second hour of the 1-mU·kg−1·min−1 insulin infusion, plasma glucose and free insulin concentrations were comparable in the diabetic patients and the normal subjects (Table 2⇑). At similar levels of glycemia and insulinemia, both whole-body glucose uptake (24±2 versus 55±4 μmol·kg body wt−1·min−1, P<.001) and forearm glucose uptake (26±4 versus 61±7 μmol·kg forearm−1·min−1, P<.001) were 57% lower in the diabetic patients than in the normal subjects. This difference was due to a 44% lower glucose AV difference in the diabetic patients (1.1±0.2) than in the normal subjects (1.9±0.2 mmol/L, P<.001), whereas blood flows were virtually identical in the two groups (2.8±0.3 versus 3.2±0.3 mL·dL forearm−1·min−1, respectively, P=NS).
Impact of glycemic control and the UAER on insulin sensitivity in IDDM
Two approaches were used to determine the effect of glycemic control independent of other factors on insulin sensitivity in patients with IDDM. First, we used simple and multiple linear regression analysis. In simple linear regression analysis, HbA1c and the UAER but not factors such as body mass index, percent body fat, or age were significantly and inversely correlated with forearm glucose uptake (r=−.53 and P<.05 for both). Both HbA1c (r=−.48, P<.05) and the UAER (r=−.73, P<.001) were inversely correlated with the glucose AV difference but not with insulin-stimulated blood flow (r=.14 and r=.35, respectively, P=NS). In multiple linear regression analysis, HbA1c and the UAER together explained 65% of the variation in the glucose AV difference (multiple r=.81, P<.001). Both HbA1c (P<.05) and the UAER (P<.001) were independent determinants of glucose extraction as well as of forearm glucose uptake (multiple r=.61, P<.02, P<.05 for both HbA1c and the UAER).
To further examine the independent contributions of glycemic control and the UAER on parameters of insulin sensitivity, we subdivided the IDDM patients according to their mean HbA1c and median UAER. The median rather than mean UAER was used because UAER was not normally distributed. The subgroups differing with respect to glycemic control were matched for other clinical and laboratory parameters (Table 3⇓). The subgroups differing with respect to the UAER were also well matched except for the duration of IDDM, which was significantly longer in patients with microalbuminuria than in those with normoalbuminuria (Table 3⇓). As illustrated in Figs 2 and 3⇓⇓, patients with poor glycemic control exhibited defects in whole-body and forearm glucose uptake and in glucose extraction compared with those with good or moderate glycemic control. Similarly, patients with microalbuminuria (and longer duration of diabetes) were more insulin resistant at the level of the whole body and across forearm tissues than those with normoalbuminuria. This excessive insulin resistance was also due to a defect in glucose extraction and not blood flow (Fig 3⇓).
When the entire group of diabetic patients was compared with the normal subjects, no significant differences were observed between blood flow basally or during the ACh, SNP, and L-NMMA infusions (data not shown).
Impact of glycemic control and UAER on in vivo endothelial function in IDDM
The possible impact of glycemic control and UAER on endothelial function was analyzed as above, first by use of correlation analyses and then by examination of endothelial function in patients subdivided as in Table 3⇑. Within the group of diabetic patients, HbA1c (r=−.50, P<.02) but not UAER (r=.01, P=NS) was inversely correlated with forearm blood flow during the submaximal dose of ACh (r=−.50, P<.02). Neither HbA1c (r=.07 and r=.10) nor UAER (r=−.15 and −0.19, P=NS for submaximal and maximal doses) was significantly correlated with blood flow during SNP administration. HbA1c was also inversely correlated (r=−.51, P<.02) with the ratio of blood flow during the submaximal ACh dose (7.5 μg/min) to blood flow during the submaximal SNP dose (3 μg/min). The independence of the association between HbA1c and endothelium-dependent vasodilatation was confirmed in multiple linear regression analysis (P<.01 for HbA1c, P=NS for UAER). Fig 4⇓ shows blood flows during the entire endothelial function test in the subgroups of patients with IDDM divided according to the mean HbA1c value (Table 3⇑) compared with normal subjects. L-NMMA decreased blood flow similarly in all groups (from 2.9±0.4 to 2.4±0.3, 2.4±0.2 to 1.8±0.1, and 2.9±0.3 to 2.2±0.3 mL·dL−1·min−1 in normal subjects and in IDDM patients in good and poor control, respectively, Fig 4⇓). The ratio of endothelium-dependent to endothelium-independent blood flow during submaximal ACh and SNP doses was ≈40% lower in the IDDM patients with poor glycemic control than in those with good or moderate glycemic control and the normal subjects (Fig 5⇓).
The present data provide the first in vivo evidence for an association between chronic hyperglycemia and impaired endothelium-dependent vasodilation in patients with IDDM. Furthermore, we confirm our previous findings demonstrating that insulin resistance in patients with IDDM can be attributed to a defect in insulin stimulation of glucose extraction16 17 rather than blood flow, which can be blocked by inhibition of NO synthesis with L-NMMA.14 Together, these data imply that chronic hyperglycemia induces endothelial dysfunction and insulin resistance via different mechanisms.
Several factors may contribute to the failure to find an association between chronic hyperglycemia and endothelial dysfunction in previous studies in patients with IDDM. Normoglycemia was maintained in only one of the previous studies8 during the endothelial function tests. Since acute hyperglycemia (2 hours) prevents the normal dilatory response to graded coronary occlusion in dogs,23 it might modify vasodilatory responses. In keeping with this postulate, basal blood flow was significantly increased in the IDDM patients with a fasting glucose concentration of 11.4 mmol/L in the study by Halkin et al.24 The present data demonstrate abnormal ACh stimulation of blood flow, which could be explained by chronic but not acute hyperglycemia. Variation in glycemic control between study groups in previous studies might also provide some explanation for the variable results. In the study by Calver et al,9 the blood flow response to ACh was normal in IDDM patients who were better controlled (HbA1c, 6.7%) than in the patients characterized by impaired ACh responsiveness in the present study (HbA1c, 8.6%) or the IDDM patients studied by Johnstone et al11 (HbA1c, 11.9%), who were hyporesponsive to metacholine. In the latter study, however, no correlation was found between glycemic control and the vasodilatory response to metacholine. Inclusion of both male and female subjects might add another source of variation, since women seem to be protected, eg, against the adverse effects of hypercholesterolemia on endothelium-dependent vasodilatation compared with men.25 The sex difference will increase the overall variation in responses to ACh but not SNP. This might obscure differences in endothelial function between mixed study groups.
Microalbuminuria is an early marker of diabetic nephropathy and is also regarded as a risk factor for coronary artery disease and myocardial infarction in patients with IDDM and NIDDM.26 In patients with IDDM, microalbuminuria has also been shown to be associated with insulin resistance independent of glycemic control.13 This was confirmed in the present study. In addition, the present data demonstrate that insulin resistance in microalbuminuric patients was caused by a defect in insulin stimulation of glucose extraction and not blood flow (Fig 3⇑). It is well established that chronic hyperglycemia is an important cause of microalbuminuria.27 Given the longer duration of diabetes in the IDDM patients in the present study (Table 3⇑), it is therefore possible that the more severe insulin resistance in the patients with microalbuminuria than in those with normoalbuminuria was due to long-term effects of hyperglycemia.
Blood pressure was not significantly increased in the IDDM patients with microalbuminuria compared with those with normoalbuminuria and the normal subjects (Tables 1 and 3⇑⇑). However, recent studies have indicated that the degree of microalbuminuria may be better correlated with diurnal than with single blood pressure measurements.28 Since hypertension is also characterized by insulin resistance and a defect in cellular glucose extraction at physiological insulin concentrations,29 it is possible that the greater insulin resistance in the patients with microalbuminuria was a reflection of some as yet unidentified cellular defect linked with subclinical hypertension.
We did not observe any defect in the ability of ACh to stimulate blood flow in forearm resistance vessels in patients with microalbuminuria. This finding is consistent with data reported by Elliot et al,8 although the latter investigators did observe an impairment in the percent reduction of basal blood flow by L-NMMA in patients with microalbuminuria. In the other studies in IDDM patients in which endothelium-dependent vasodilatation was determined, albuminuria either was not quantified9 11 or was normal.12 24 30 Thus, there is at present no evidence for an association between microalbuminuria and ACh-dependent vasodilation independent of glycemic control. Interestingly, it was recently reported that long-term hyperglycemia markedly increases ecNOS activity in the rat heart and also increases ecNOS mRNA.31 If chronic hyperglycemia altered ecNOS similarly in human forearm resistance vessels, the apparent normality of endothelium-dependent vasodilatation might represent an adaptive phenomenon aimed at compensating for hyperglycemia-induced defects in endothelial function.
The finding of a glucose extraction defect in peripheral tissues as the major cause of insulin resistance in IDDM is consistent with our previous studies16 17 but in contrast to the report by Baron et al,32 who attributed insulin resistance in IDDM patients exclusively to a defect in blood flow. Several factors could contribute to the opposite results. First, 5 subjects were studied in the study by Baron et al32 and 18 in the present study. Second, the insulin concentration was three to five times higher in the former than in the present study. Physiological insulin concentrations induce only minor increases in blood flow, whereas supraphysiological insulin concentrations markedly stimulate blood flow.21 Thus, it is possible that use of a higher insulin concentration might have revealed a defect in blood flow in the present study group. Even if this were to be the case, the finding of a defect in insulin stimulation of blood flow at extremely high insulin concentrations would not explain insulin resistance under physiological conditions.
In the present study, as previously,16 we found that insulin resistance in patients with IDDM is due to a cellular (glucose extraction) rather than vascular defect in peripheral tissues. Furthermore, the magnitude of the glucose extraction defect was significantly, and independently of other factors, explained by the degree of glycemic control. These data, together with our previous study showing that a glucose extraction defect can be induced by exposing body tissues to hyperglycemia for 24 hours,17 are consistent with the idea that chronic hyperglycemia is the major cause of insulin resistance in patients with IDDM. Recent studies in patients with NIDDM33 have suggested that an increase in the activity of the hexosamine biosynthetic pathway, which maintains cellular glucose fluxes at normal levels by downregulating insulin-sensitive glucose transport or phosphorylation, may serve as a sensor of extracellular glucose concentrations.34
We conclude that chronic hyperglycemia in IDDM is associated with insulin resistance that is caused by a defect in glucose extraction, not blood flow. Chronic hyperglycemia is also associated with impaired endothelium-dependent vasodilatation in IDDM. These data suggest that insulin resistance in IDDM is of cellular origin and is a phenomenon distinct from the chronic hyperglycemia–induced alterations in vascular function. These data show that chronic hyperglycemia impairs insulin sensitivity and vascular function by distinct mechanisms in patients with IDDM. The data raise the possibility that chronic hyperglycemia promotes atherosclerosis by impairing endothelial function.
Selected Abbreviations and Acronyms
|ecNOS||=||endothelial constitutive nitric oxide synthase|
|IDDM||=||insulin-dependent diabetes mellitus|
|NIDDM||=||non–insulin-dependent diabetes mellitus|
|UAER||=||urinary albumin excretion rate|
This study was supported by grants from the American Diabetes Association, the Academy of Finland, and the Ahokas and Sigrid Juselius foundations (all to Dr Yki-Ja¨rvinen) and from the British Council (Dr Cockcroft, Dr Groop, and Dr Yki-Ja¨rvinen). We thank Sari Ha¨ma¨la¨inen for excellent technical assistance, Soile Aarnio for drawing the figures, and the volunteers for their help.
- Received December 18, 1995.
- Revision received March 27, 1996.
- Accepted March 28, 1996.
- Copyright © 1996 by American Heart Association
Kuusisto J, Mykka¨nen L, Pyo¨ra¨la¨ K, Laakso M. NIDDM and its metabolic control predict coronary heart disease in elderly subjects. Diabetes. 1995;43:960-967.
Tesfamariam B, Brown ML, Deykin D, Cohen R. Elevated glucose promotes generation of endothelium-derived vasoconstrictor prostanoids in rabbit aorta. J Clin Invest. 1990;85:929-932.
Meraji S, Jayakody L, Senaratne MP, Thomson ABR, Kappagoda T. Endothelium-dependent relaxation in aorta from diabetic rats. Diabetes. 1987;36:978-981.
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. J Clin Invest. 1992;90:2548-2554.
Smits P, Kapma J-A, Jacobs M-C, Lutterman J, Tien T. Endothelium-dependent vascular relaxation in patients with type 1 diabetes. Diabetes. 1993;42:148-153.
Johnstone MT, Craeger SJ, Scales KM, Cusco JA, Lee BK, Craeger MA. Impaired endothelium-dependent vasodilation in patients with insulin-dependent diabetes mellitus. Circulation. 1993;88:2510-2516.
Scherrer U, Randin D, Vollenweider P, Vollenweider L, Nicod P. Nitric oxide release accounts for insulin's vascular effects in humans. J Clin Invest. 1994;94:2511-2515.
Rees DD, Palmer RMJ, Hodson HF, Moncada S. A specific inhibitor of nitric oxide formation from L-arginine attenuates endothelium-dependent dilatation. Br J Pharmacol. 1989;996:418-424.
Yki-Ja¨rvinen H, Sahlin K, Ren JM, Koivisto VA. Localization of the rate-limiting defect for glucose disposal in skeletal muscle of insulin-resistant type 1 diabetic patients. Diabetes. 1990;39:157-167.
Vuorinen-Markkola H, Koivisto VA, Yki-Ja¨rvinen H. Mechanisms of hyperglycemia-induced insulin resistance in whole body and skeletal muscle of type 1 diabetic patients. Diabetes. 1992;41:571-580.
DeFronzo RA, Tobin JD, Andres R. Glucose clamp technique: a method for quantifying insulin secretion and resistance. Am J Physiol. 1979;237:E214-E223.
Yki-Ja¨rvinen H, Young AA, Lamkin C, Foley JE. Kinetics of glucose disposal in whole body and across the forearm in man. J Clin Invest. 1987;79:1713-1719.
Rising R, Swinburn B, Larson K, Ravussin E. Body composition in Pima Indians: validation of bioelectrical resistance. Am J Clin Nutr. 1991;53:594-598.
Kersten JR, Brooks LA, Dellsperger KC. Impaired microvascular response to graded coronary occlusion in diabetic and hyperglycemic dogs. Am J Physiol. 1995;268:H1667-H1674.
Deckert T, Kofoed-Enevoldsen A, Norgaard K, Borch-Johnsen K, Feldt-Rasmussen B, Jensen T. Microalbuminuria: implications for micro- and macrovascular disease. Diabetes Care. 1992;15:1181-1191.
Nielsen S, Schmitz A, Poulsen PL, Hansen KW, Mogensen CE. Albuminuria and 24-h ambulatory blood pressure in normoalbuminuric and microalbuminuric NIDDM patients. Diabetes Care. 1995;11:1434-1441.
Natali A, Santoro D, Palombo C, Cerri M, Ghione M, Ferrannini E. Impaired insulin action on skeletal muscle metabolism in essential hypertension. Hypertension. 1991;17:170-178.
Smits P, Hersbach F, Jansen T, Tien T, Lutterman J. Impaired vasodilator response to atrial natriuretic factor in IDDM. Diabetes. 1993;42:1454-1461.
Ro¨sen P, Ballhausen T, Stockklauser K, Ho¨nack C. Influence of glucose and diabetes on the endothelial NO-synthase. Diabetologia. 1995;38:A48. Abstract.
Yki-Ja¨rvinen H, Daniels MC, Virkama¨ki A, Ma¨kimattila S, DeFronzo RA, McCain D. Increased glutamine: fructose-6-phosphate amidotransferase activity in skeletal muscle of patients with NDDM. Diabetes. 1996;45:302-307.
Marshall S, Bacote V, Traxinger RR. Discovery of a metabolic pathway mediating glucose-induced desensitization of the glucose transport system: role of hexosamine biosynthesis in the induction of insulin resistance. J Biol Chem. 1991;266:4706-4712.