Objective: To examine the effects of intravenous insulin and acute hypoglycemia on arterial wall stiffness and central hemodynamic responses in adults with and without type 1 diabetes. Research Design and Methods: In 30 young male volunteers [10 nondiabetic (Group 1); 10 with type 1 diabetes, <5 yr duration (Group 2); 10 with type 1 diabetes, >15 yr duration (Group 3)], intravenous insulin was administered to provoke an acute autonomic reaction (R) to hypoglycemia. Heart rate, peripheral blood pressure, and pulse wave analysis (radial artery) were monitored. Augmentation index (AIx), a measure of arterial wall stiffness and wave reflection, and central arterial pressure were recorded. Results: At baseline, no significant differences were observed between Groups 1 and 2 in either AIx or in central arterial pressure, but in Group 3, both measures were significantly higher. All three groups exhibited similar responses to intravenous infusion of insulin and to hypoglycemia: AIx fell progressively from baseline to R, peripheral systolic blood pressure increased, and central systolic pressure decreased. Conclusion: Compared with age- and sex-matched nondiabetic controls, people who had type 1 diabetes for a long duration had increased stiffness of vessel walls. The opposing responses in peripheral and central systolic pressures during hypoglycemia may be related to the reduction in AIx, which causes diminished amplification of the systolic pressure wave. Changes in AIx are probably mediated by a direct action of insulin on arterial endothelium, or changes in heart rate. These functional changes may contribute to the increased cardiovascular morbidity that is associated with type 1 diabetes of long duration.
- arterial wall stiffness
- central arterial pressure
people with type 1 diabetes are frequently exposed to acute hypoglycemia, which, in humans, is associated with profound hemodynamic changes, including an increase in heart rate and stroke volume, increased myocardial contractility, and a rise in cardiac output (14, 15). Changes in peripheral blood pressure include an increase in systolic and a decrease in diastolic pressure, with no change in mean arterial pressure. The effects on central arterial pressure are not known. Acute hypoglycemia also has significant, albeit transient, effects on peripheral blood that encourage hemostasis and increase blood viscosity (15). These changes influence capillary blood flow and in some tissues may predispose to intravascular stasis during hypoglycemia. The hemodynamic and hematological changes that occur during acute hypoglycemia do not appear to compromise vascular perfusion in the normal healthy vasculature, but they may have potentially adverse effects in people with diabetes, in whom endothelial dysfunction and vascular disease are common, and in whom premature atherosclerosis and cardiovascular disease are often present. Indeed, acute hypoglycemia may precipitate an acute vascular event or a dysrhythmia in people who have preexisting coronary heart disease (15). Angina and acute myocardial infarction have been documented, albeit infrequently and anecdotally, as the direct consequences of acute hypoglycemia (35).
Arterial blood pressure is usually recorded noninvasively at the brachial artery by sphygmomanometry. It is assumed that the pressure is the same throughout the arterial tree, and that it represents an accurate index of aortic peak pressure and of left ventricular systolic pressure, and so of left ventricular afterload. However, blood pressure measured in the peripheral circulation is not an accurate estimate of central pressure because of amplification of the pressure pulse between central and peripheral arteries (34, 43). Normally, there is amplification of the pulse pressure between the aorta and brachial artery (24). It is the aortic, rather than the brachial, pressure that determines left ventricular workload (42, 47).
Central arterial pressure and arterial wall stiffness, which are independent predictors of cardiovascular morbidity (1, 3, 20, 25, 31), can be studied noninvasively using pulse wave analysis (30, 32, 44, 51), which requires the external application of a micromanometer-tipped probe over a peripheral artery (6, 8). Pulse wave analysis allows the measurement of central arterial pressure, and the degree of its augmentation by pulse wave reflection, to be studied reliably and noninvasively (32). The system utilizes applanation tonometry, with the external application of a micromanometer-tipped probe to record peripheral pulse waveforms. The radial artery is usually used for these measurements because of its close proximity to and the support provided by nearby bony structures (6, 8). Changes occur in the pressure contour as it travels from the aorta to more peripheral sites, so that pressures at the radial artery cannot be used directly as a surrogate to predict central pressure in the aorta. However, it is possible to estimate the central aortic pressure wave from measurements of radial artery tonometry with the use of a mathematical transformation (8). Several studies have validated the accuracy and reliability of pulse wave analysis (44, 51). Contraction of the left ventricle generates a pressure wave that travels along the major arteries until it meets sites of resistance (high-resistance arterioles), and then the wave is reflected back to the heart (29). The stiffness of the arterial wall influences the velocity of the reflected wave. In healthy young people, the reflected wave reaches the heart during diastole and, by increasing diastolic pressure, enhances coronary perfusion. However, when arterial stiffening has developed, the increased pulse wave velocity of the stiffened arteries leads to earlier reflection of the wave, so that it returns to the heart during late systole (31, 33). The increase in systolic pressure that results from this earlier reflection of the pressure wave is known as “augmentation,” with augmentation index (AIx) measured as the augmented pressure expressed as a percentage of the pulse pressure. AIx increases with age and in disease states such as hypertension (28) and hypercholesterolemia (49). It is also influenced by heart rate; a decrease in AIx is observed with an increase in heart rate (45). Central arterial pressure and waveform convey important information about cardiovascular status (21, 27, 31).
In view of the frequency of exposure to acute hypoglycemia in people with diabetes and the potential effects of acute hypoglycemia on the cardiovascular system, it is valuable to establish the effects of acute hypoglycemia on central arterial pressure and left ventricular load. The present study sought to measure arterial wall stiffness and the central hemodynamic responses to the administration of intravenous insulin and the induction of acute hypoglycemia in people with type 1 diabetes of varying duration and in nondiabetic subjects.
To avoid potential gender differences in the counterregulatory hormonal responses to hypoglycemia (10), three groups of male subjects, matched for age, body mass index (BMI), and peripheral blood pressure, were studied: 10 nondiabetic volunteers (Group 1), 10 subjects with type 1 diabetes of short duration (<5 yr) (Group 2), and 10 subjects with type 1 diabetes of longer duration (>15 yr) (Group 3). Subjects with diabetes were matched for glycated hemoglobin (HbA1c) and were recruited from the outpatient clinic at the Department of Diabetes at the Royal Infirmary of Edinburgh. All patients were receiving treatment with a basal-bolus insulin regimen consisting of a preprandial short-acting insulin analog and a long-acting insulin analog, administered once daily. Blood pressure and plasma cholesterol concentration were measured before inclusion in the study. HbA1c was measured by HPLC (Variant II Hemoglobin Testing System; Bio-Rad Diagnostics Group, Hercules, CA), with a local Diabetes Control and Complications Trial-adjusted nondiabetic reference range of 4.3–6.5%. Subjects were excluded from the study if there was a history of hypertension (blood pressure >140/80 mmHg), hypercholesterolemia (total cholesterol >5.0 mmol/l), or other chronic disease or a history of a previous head injury, alcohol or drug abuse, seizures, or blackouts. All of the subjects were nonsmokers, had no intercurrent illness, and were not taking any regular medication (other than insulin for diabetes). Subjects with diabetes were excluded if they had any evidence of microvascular disease other than background diabetic retinopathy (2 subjects in Group 2 and 5 subjects in Group 3). The presence of retinopathy was determined by direct ophthalmoscopy, peripheral neuropathy by clinical examination, and nephropathy by the presence of microalbuminuria (albumin-to-creatinine ratio >2.5 mg/mmol for women and >3.5 mg/mmol for men). All potential participants underwent standard autonomic function testing (13) before inclusion in the study, and any with autonomic dysfunction were excluded. Subjects were also excluded if they had a history of impaired awareness of hypoglycemia.
Studies were postponed if a subject had experienced hypoglycemia in the preceding 48 h, and subjects were required to measure their blood glucose frequently during this period, including bedtime tests on the evening before the study to identify biochemical hypoglycemia. Ethical permission for the study was given by the Lothian Research Ethics Committee. All subjects gave their written, informed consent.
Each study session began at 0800 after a 10- to 12-h overnight fast. The subjects with type 1 diabetes omitted their insulin on the morning of the study. A hyperinsulinemic glucose clamp (12) was used initially to maintain euglycemia (blood glucose concentration 4.0–7.0 mmol/l) for 20 min before the induction of acute hypoglycemia. Because of the variable fasting blood glucose level in people with diabetes, it took a variable time to achieve euglycemia (usually between 20 and 30 min). An intravenous cannula was inserted retrogradely into a vein on the dorsum of the nondominant hand for regular blood sampling. The hand was placed in a heated blanket to arterialize the venous blood. A second intravenous cannula was inserted into a vein in the antecubital fossa of the same arm for infusion of human soluble insulin (Humulin S; Eli Lilly) and 20% dextrose. Intradermal lignocaine (1%) was used for insertion of each cannula. Insulin was infused at a constant rate of 60 mU·m−2·min−1 using an IMED Gemini PCI pump (Alaris Medical Systems, San Diego, CA). A variable intravenous infusion of 20% dextrose was given simultaneously. The rate of dextrose infusion was varied according to the arterialized blood glucose concentration. This was measured at the bedside using the glucose oxidase method (Yellow Springs Instrument 2300 Stat, Yellow Springs, OH). Blood glucose concentration was measured every 3 min initially and then every 5 min once a stable level had been achieved.
Because the purpose of inducing acute hypoglycemia was to provoke an acute autonomic reaction with profound activation of the sympathoadrenal system, a rapid fall in blood glucose was then induced by infusing insulin intravenously at a rate of 2 mU·kg−1·min−1. This caused a controlled decline in blood glucose to a level that triggered the autonomic reaction associated with hypoglycemia. This method of inducing hypoglycemia simulates the development of hypoglycemia under everyday conditions and cannot be induced using the hyperinsulinemic glucose clamp technique. Any variability among subjects in the rate of fall in blood glucose and the depth of the hypoglycemia achieved is inconsequential, as the intention is to lower blood glucose to the blood glucose threshold at which autonomic activation is stimulated.
The autonomic reaction (designated as R) was identified by an abrupt rise in heart rate (15% above baseline), the rapid onset of autonomic symptoms, such as sweating and tremor, and a rapid rise in peripheral systolic blood pressure. When the R occurred, the clock timer was reset to zero, and all subsequent measurements were timed relative to this point. This accommodates the variable time interval that occurs among individuals in the development of the R using the insulin infusion technique (9). As soon as R had occurred, the insulin infusion was discontinued, and hypoglycemia was reversed using an intravenous infusion of 20% dextrose to restore normoglycemia. All subjects consumed a meal on completion of the study.
During the study, arterialized blood glucose was measured at the bedside (Yellow Springs Instrument 2300 Stat) every 5 min until R and every 15 min thereafter until R+30 min. The maximal hemodynamic changes in response to hypoglycemia occur in the 30 min following the onset of the R (14). Pulse wave analysis at the radial artery was therefore performed every 5 min until R and every 5 min thereafter until R+30 min. Heart rate was monitored continuously using a pulse oximetry probe. Peripheral blood pressure was recorded every 5 min using a digital automated sphygmomanometer.
The Edinburgh Hypoglycemia Scale (11), a validated subjective self-rating questionnaire, was used to document the symptoms of hypoglycemia experienced by the subjects during the study. The symptoms of hypoglycemia were classified as autonomic (hunger, palpitations, sweating, shaking), neuroglycopenic (drowsiness, confusion, inability to concentrate, speech difficulty, blurred vision), and nonspecific (malaise, nausea, headache). Each symptom was graded on a Likert scale of 1 to 7 (1 = not present, 7 = very intense). The questionnaire was applied at baseline and at 5-min intervals thereafter until the onset of the R. It was repeated at R+30 min.
A general linear model (repeated-measures ANOVA) was used to compare means within individual groups. Between groups, an ANOVA with post hoc t-tests was used. Correlation analyses for changes in AIx and insulin dose and duration of the study were performed using the Pearson correlation. A P value <0.05 was considered to be significant. All analyses were performed using SPSS version 11.0 for Windows. All results are expressed as means ± SD.
The clinical characteristics of the subjects in the study are shown in Table 1.
Changes in blood glucose concentration.
The blood glucose concentrations (means ± SD) at baseline were 4.3 ± 0.5 mmol/l in Group 1 (nondiabetic) and 8.4 ± 2.1 and 8.9 ± 3.1 mmol/l in Groups 2 and 3 (diabetic), respectively. Following application of the euglycemic clamp, the blood glucose concentrations were stabilized at 4.2 ± 0.1 mmol/l (Group 1), 4.3 ± 0.2 mmol/l (Group 2), and 4.3 ± 0.2 mmol/l (Group 3). The nadir blood glucose concentrations at R were 2.9 ± 0.3 mmol/l (Group 1), 2.8 ± 0.3 mmol/l (Group 2), and 2.6 ± 0.3 mmol/l (Group 3) (P = 0.2).
Peripheral blood pressure.
No significant differences were observed in the peripheral systolic, diastolic, and mean arterial pressures among the three groups at baseline. In all three groups of subjects, an increase in peripheral systolic blood pressure and a decrease in peripheral diastolic pressure occurred at R compared with baseline (Table 2). Systolic and diastolic pressures had returned to baseline levels by R+30 min in all three groups (Table 2).
Central arterial pressure.
Central systolic blood pressure at baseline was significantly higher in Group 3 compared with Groups 1 and 2 (Table 1). A small, but significant, reduction in central systolic pressure was observed at R in all three groups compared with the respective baseline levels (Table 2). No change in central diastolic blood pressure was observed throughout the study in any of the three groups (Table 2).
Heart rate and autonomic symptom score.
The induction of hypoglycemia and the onset of the R were associated with significant increments of heart rate and autonomic symptom score in all three groups compared with baseline levels (Table 2). In all groups, the heart rate and the autonomic symptom score had returned to baseline values at R+30 min.
At baseline, no difference in AIx was discernible between Group 1 and Group 2, whereas the AIx in Group 3 was higher than in either of these groups (Table 1).
All three groups exhibited a progressive decline in AIx between baseline and the development of hypoglycemia at R (Fig. 1). The fall in AIx from baseline to R was significantly less in Group 3 compared with the reduction in AIx from baseline to R in Groups 1 (P < 0.01) and 2 (P < 0.01). At R+30 min, AIx had risen to approach baseline levels in all three groups.
Changes in AIx, duration of insulin exposure, and insulin dose.
There was no significant difference in the duration of the euglycemic clamp procedure or the time taken to induce hypoglycemia among the three groups. The duration of the euglycemic clamp [means (SD)] measured 37.3 (11.1) min in Group 1, 44.2 (8.5) min in Group 2, and 46.9 (10.0) min in Group 3 (P = 0.3). The time taken to induce the R measured 18.1 (6.4) min in Group 1, 22.4 (5.2) min in Group 2, and 20.7 (6.9) min in Group 3 (P = 0.5). However, among individuals within the separate groups, a correlation was observed between the time taken to induce hypoglycemia and changes in AIx, with a longer duration being associated with a greater reduction in AIx (Group 1, r = 0.25, P = 0.04; Group 2, r = 0.38, P = 0.3; Group 3, r = 0.28, P = 0.04). Similarly, no significant difference was found in the amount of insulin infused during the studies among the three different groups. The total insulin dose administered was 14.6 (4.4) U in Group 1, 12.7 (3.1) U in Group 2, and 15.7 (4.8) U in Group 3 (P = 0.6). However, within the separate groups, a significant correlation was present between changes in AIx and insulin exposure. A greater decrement in AIx was associated with greater insulin exposure (Group 1, r = 0.30, P = 0.05; Group 2, r = 0.34, P = 0.04; Group 3, r = 0.3, P = 0.03).
Arterial stiffness is increased in people with type 2 diabetes (5, 7, 16, 17), in their nondiabetic relatives (19), and in people with impaired glucose tolerance (37). In people with type 1 diabetes, discrepant results have been reported. Some studies have demonstrated greater arterial stiffness compared with nondiabetic controls (4, 36, 46), whereas others have shown either no difference (23) or lower arterial stiffness (26). These differences may relate to the methodologies used and to patient selection, with participants differing in their duration of diabetes, age and gender, smoking status, peripheral blood pressure, serum cholesterol concentrations, and the presence or absence of microvascular disease.
In the present study, arterial stiffness and central arterial pressure were studied both in nondiabetic adults and in people with type 1 diabetes of differing duration, using radial artery pulse wave analysis and calculation of the AIx, and the subjects, all of whom were nonsmokers, were matched for age, gender, blood pressure, serum cholesterol, and body mass index. The results showed no difference in vessel wall stiffness or central arterial pressure between a group with well-controlled type 1 diabetes of short duration and a matched cohort of healthy nondiabetic volunteers. However, the people with type 1 diabetes of longer duration had greater stiffness of the arterial vessel wall and an elevated central arterial pressure compared with nondiabetic controls. This suggests that the duration of exposure to chronic hyperglycemia influences vessel wall stiffness rather than the presence of type 1 diabetes itself. This may result from structural changes caused by the gradual deposition of advanced glycation end products within the walls of large blood vessels (18). The endothelium (22, 50) and arterial wall smooth muscle bulk and tone (2) (the latter under the control of the endothelium) also influence the elasticity of arteries, and it is also possible, therefore, that increased vessel wall stiffness in people with type 1 diabetes of long duration is the result of functional changes that occur within the endothelium with increasing duration of diabetes.
In the present study, the rise in peripheral systolic blood pressure that is known to accompany acute insulin-induced hypoglycemia was accompanied by a reduction of central systolic blood pressure. These findings highlight the potential for divergent responses of central and peripheral arterial pressures and support the premise that the measurement of peripheral blood pressure by conventional methods does not provide an accurate measure of the central arterial pressure load on the left ventricle. Specifically, the results show how the changes of central and peripheral arterial pressures differ in response to the stress provoked by acute hypoglycemia and mediated by sympathoadrenal stimulation. It could be speculated that the reduction in central arterial pressure that was observed during acute hypoglycemia may lead to a reduction of coronary perfusion, which could potentially precipitate a vascular event in people with vascular disease.
The difference that was observed between the responses of peripheral and central systolic blood pressures to insulin-induced hypoglycemia is indicated by the observed changes in AIx; a reduction in AIx reflects diminished amplification of the systolic pressure wave and so causes a net overall reduction in central arterial systolic blood pressure. It is possible that this observed change in AIx had resulted either from changes in heart rate (45) or from the actions of vasoactive hormones that are secreted in response to hypoglycemia (38), or as a combination of both effects. An increment in heart rate of 10 beats/min causes a fall in AIx of ∼4% (45), and the infusion of norepinephrine promotes an increase in AIx (48). However, in the present study, the AIx initially was observed to fall during euglycemia, before any stimulated changes in heart rate, autonomic symptom score, or peripheral blood pressure. Insulin per se may have a direct effect on the functional integrity of large blood vessels. In the present study, no significant differences were observed in the duration of insulin exposure or in the total insulin dose administered, but it was observed that, in individuals within each study group, a significant correlation existed between changes in AIx and the duration of insulin exposure and, consequently, the amount of insulin infused. This would support the suggestion that the early changes in AIx may therefore be a consequence of the direct action of insulin on endothelial function or on smooth muscle tone of large arteries. This interpretation would be consistent with the results of other studies that have demonstrated that the intravenous administration of insulin increases the distensibility of large arteries (39, 40).
The decrement of AIx was less in the group with type 1 diabetes of longer duration compared with that of subjects with diabetes of short duration and nondiabetic subjects. This would be consistent with greater stiffness and diminished elasticity of the vessel walls of large arteries in people with type 1 diabetes of long duration. Consistent with this observation, Westerbacka et al. (41) have shown that the effect of insulin to reduce vessel wall stiffness in large arteries is diminished in people with type 1 diabetes who have no vascular complications. The mean duration of type 1 diabetes in their participants was 18 yr (41), which is similar to the duration of diabetes of Group 3 in the present study. It has to be acknowledged that the algorithm to relate peripheral and central pressures by applanation tonometry was derived in nondiabetic subjects. However, in the present study, the changes in the central hemodynamic responses to insulin infusion and the induction of acute hypoglycemia were similar in both the nondiabetic and the diabetic groups.
Type 1 diabetes is associated with an increased prevalence of cardiovascular disease, and vessel wall stiffness is an independent predictor of cardiovascular morbidity and mortality. The present study has demonstrated that a greater duration of type 1 diabetes was associated with an increase in vessel wall stiffness, before the development of overt macrovascular disease. Central systolic pressure at rest was also greater in people with type 1 diabetes of longer duration, although peripheral blood pressure did not differ among the three study groups. Changes in brachial pressure may therefore underestimate changes in aortic pulse pressure and the response of left ventricular systolic function to hemodynamic stress in people who have had type 1 diabetes for several years. These vascular changes and the attendant rise in central arterial blood pressure may contribute to the increased risk of cardiovascular morbidity in people with type 1 diabetes.
A. J. Sommerfield was supported by research funding from Eli Lilly and Company.
We wish to acknowledge the expert assistance of the staff of the Wellcome Trust Clinical Research Facility, Western General Hospital, Edinburgh, Scotland.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 by American Physiological Society