INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

ALTERATIONS IN THE MICROCIRCULATION contribute to the vascular and cutaneous complications of Type 2 diabetes mellitus (2, 17, 22, 29). The slow onset of Type 2 diabetes mellitus, with gradual development of abnormal glucose regulation over several years, often goes undiagnosed because of the lack of overt symptoms (24). Importantly, microvascular complications are frequently present at the time of diagnosis, indicating that microvascular pathology is often undetected during the period of undiagnosed hyperglycemia (11).

Abnormal glucose regulation during the development of Type 2 diabetes mellitus is characterized by hyperglycemia, which is usually most marked after meals. Thus a person with impaired glucose tolerance is likely to experience intermittent episodes of acute hyperglycemia interspersed among periods of relative normoglycemia. Furthermore, the risk of microvascular complications rises dramatically when fasting blood glucose exceeds 7 mmol/l (126 mg/dl) (8, 20, 24). Many therapies used in the treatment of Type 2 diabetes are more effective in lowering fasting than postprandial glucose levels (9, 10). In recent years, the relative importance of postprandial vs. fasting hyperglycemia in the pathogenesis of vascular complications has become a matter of debate (7).

It is presently not known whether either a single transient postprandial rise in glucose or a single episode of sustained but modest hyperglycemia alters vascular function. Some evidence suggests that such acute hyperglycemia may do so. In coronary arteries of healthy dogs (18) and in mesenteric resistance vessels of healthy rats (26), vasodilator responsiveness to acetylcholine was diminished during acute exposure to hyperglycemia. In healthy humans, brachial artery flow-mediated dilation was diminished by acute systemic hyperglycemia induced by oral glucose loading (28). Also in healthy subjects, 6 h of "local" hyperglycemia [16.7 mmol/l (300 mg/dl); glucose infusions into the brachial artery] decreased endothelium-dependent forearm vasodilation to intra-arterial methacholine (30). This impairment was reversed by inhibition of protein kinase C-, suggesting that hyperglycemia might impair vascular function by increasing the activity of this enzyme (3). Whether lower levels of hyperglycemia, in patterns seen early in diabetes, have similar effects is unknown.

Clearly, the skin is a site of substantial pathology in diabetes. However, whether the acute effects of hyperglycemia noted above are true for the cutaneous microcirculation and could thus contribute to the development of pathology in this vascular bed is not clear. In the present study, we tested whether, in humans, systemic hyperglycemia per se (i.e., in the absence of the insulin response) at levels corresponding to those shown to increase risk for vascular disease [i.e., fasting glucose of 7 mmol/l (126 mg/dl) or peak postprandial glucose of 11.1 mmol/l (200 mg/dl)] has the effects in the skin previously shown for higher levels of hyperglycemia in other species and/or circulations.

In humans, direct local warming of the skin causes substantial vasodilation of the area being warmed. The cutaneous vasodilator response to prolonged local warming is biphasic, including an initial peak and modest decrease followed by a slower second phase that attains a plateau around 30 min of warming (13, 15, 21). The slower phase of this vasodilation was recently identified to be largely (65-75%) dependent on nitric oxide (15, 21), whereas the initial peak is dependent on axon reflex activity of cutaneous sensory nerves (21). Thus this local warming paradigm can be used as a tool to assess nitric oxide-dependent and axon-reflex vasodilation in the cutaneous microvasculature (15, 21).

With this information as a background, we hypothesized that acute hyperglycemia would diminish the second, largely nitric oxide-dependent phase of this vasodilator response but have no effect on the initial, neurally mediated phase. We tested this hypothesis by observing cutaneous vasodilator responses to local warming of the skin before, during, and after elevated plasma glucose concentrations mimicking those observed after a meal in an individual with Type 2 diabetes mellitus and compared these with the cutaneous vasodilator responses to local warming during sustained hyperglycemia and euglycemia.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

At 1700 the evening before the study, subjects checked in to the Mayo Clinic General Clinical Research Center and consumed a standard mixed meal (10 kcal/kg, 55% carbohydrate, 30% fat, and 15% protein). Subjects then remained in the General Clinical Research Center and fasted overnight. At 0600, a forearm vein in the dominant arm was cannulated with an 18-gauge intravenous catheter for the systemic infusions (see below). At 0700, the brachial artery of the nondominant arm was cannulated with an 18-gauge catheter for blood sampling and measurement of arterial blood pressure. From this time until the end of the study (~1600), the subject remained supine and did not consume any food or beverage.

Procedures

Skin blood flow was measured by laser-Doppler flowmetry (LDF) (Perimed Periflux System 5000, Stockholm, Sweden) on the lateral aspect of the lower leg, approximately halfway between the knee and the ankle. The laser-Doppler probe, in the center of a 12-cm² local warming unit, was attached to the measurement site at the beginning of the study day and was not moved or removed until the end of the study day.

Protocols

View larger version (13K): [in this window] [in a new window]   Fig. 1.   Time course for plasma glucose for the 3 glucose protocols. Protocol 1 simulated the plasma glucose concentrations in individuals with glucose intolerance after carbohydrate ingestion. Protocol 2 simulated fasting hyperglycemia in early Type 2 diabetes mellitus. Protocol 3 was the euglycemia time control. Solid rectangles show time points for local warming trials.

Protocol 1 (glucose profile). This protocol was designed to mimic the plasma glucose concentrations likely to occur in individuals with glucose intolerance after carbohydrate ingestion. Starting at 0730, somatostatin was infused (60 ng · kg¹ · min¹) to inhibit endogenous insulin secretion; growth hormone (3 ng · kg¹ · min¹) and glucagon (0.65 ng · kg¹ · min¹) were also infused to maintain systemic fasting concentrations. Insulin was infused at 0.20 mU · kg¹ · min¹ to achieve high physiological concentrations for the duration of the experiment. A variable glucose infusion was begun at 0730 to maintain plasma glucose concentrations at 5 mmol/l (95 mg/dl) until 0900. At 0900 (time 0 in Fig. 1), plasma glucose concentrations were increased in such a fashion that a peak plasma glucose concentration of ~11.1 mmol/l (~200 mg/dl) was achieved at ~60 min (1000) followed by a return to baseline levels as shown in Fig. 1. Trace amounts of [3-³H]glucose were also infused as part of a separate protocol. Blood glucose was measured every 10 min by use of a Beckman glucose analyzer to facilitate clamping of glucose at target levels.

Protocol 2 (constant hyperglycemia). This protocol was designed to reproduce the modest elevations in fasting glucose that commonly occur in individuals with early Type 2 diabetes mellitus. This protocol was identical to protocol 1, except that from 0900 sufficient exogenous glucose was infused to clamp plasma glucose at ~7.0 mmol/l (126 mg/dl) for the 6 h of study. Protocols 1 and 2 were designed such that the total glycemic excursions (areas under the curve for glucose from minute 0 to minute 390) did not differ.

Protocol 3 (euglycemia). Protocol 3 was the same as the other two protocols except that sufficient glucose was infused to maintain glucose concentrations at ~5.3 mmol/l.

Local Warming of the Skin

There were four local warming trials during each study. Figure 1 shows the timing of the four local warming trials as related to plasma glucose levels during the three protocols. A baseline local warming trial was conducted at 30 min, such that all individuals were at euglycemia; in this way each individual served as his or her own control. The second trial, at 60 min, corresponded to peak plasma glucose in protocol 1. The third trial (minute 180) corresponded to the time point at which plasma glucose levels for protocols 1 and 2 were similar. The fourth trial (minute 360) corresponded to the time point at which the areas under the curves for glucose were the same between protocols 1 and 2. Importantly, no subject experienced pain at any time during any local warming trial.

Blood Samples

A separate set of samples for poststudy glucose and hormone analysis was placed on ice, centrifuged at 4°C, separated, and stored at 20°C until assay. Plasma C-peptide and glucagon concentrations were measured in the Mayo Immunochemical Core Laboratory by using reagents purchased from Linco Research (St. Louis, MO). Insulin and growth hormone were measured by chemiluminescence (Access Assay, Beckman, Chaska, MN). Glucose was measured by using a Yellow Springs glucose analyzer.

Data Analysis

To compare total glycemic excursion between protocols 1 and 2, the area under the curve for glucose was calculated for each subject by taking the sum of time-weighted averages of plasma glucose values between minute 0 and minute 390. Total glycemic excursion was compared between protocols 1 and 2 by unpaired t-test.

For the remaining statistical analysis, each subject served as his or her own control, such that data during hyperglycemia were compared with data from the baseline period. Baseline CVC and CVC response to local warming were compared across local warming trials by ANOVA with repeated measures. To test whether hyperglycemia altered the nadir after the initial peak, peak and nadir values were compared by t-test at each trial time point. Plasma glucose and insulin concentrations, as well as heart rate and blood pressure during each trial, were analyzed over each protocol by repeated-measures ANOVA. When a significant effect was detected, Tukey's post hoc test was used to identify individual differences. Statistical significance was accepted for P < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Plasma Glucose, Insulin, and C-Peptide Concentrations

View larger version (14K): [in this window] [in a new window]   Fig. 2.   Plasma glucose (A) and insulin (B) concentrations during the 4 local warming trials for each of the 3 protocols.

As designed, the total glycemic excursion during protocols 1 and 2 did not differ (2,942.3 ± 56.1 vs. 2,823.2 ± 27.3 mmol/390 min, P > 0.05). Insulin concentrations did not differ among the three protocols. Somatostatin resulted in prompt and comparable suppression of C peptide during all three protocols; glucagon and growth hormone remained constant. These values are shown in Table 2.

Maximum CVC

CVC at 33°C

Dynamic Response to Local Warming

View larger version (16K): [in this window] [in a new window]   Fig. 3.   Cutaneous vascular response to local warming during the 4 local warming trials in a representative individual from the glucose profile group (protocol 1). As in all subjects, there was no effect of acute hyperglycemia on the vasodilator response. At 360 min, the nadir after the initial peak was missing; this was true for all groups, including those maintained at euglycemia, and so did not appear to be an effect of hyperglycemia (see text). CVC, cutaneous vascular conductance; max, maximum.

Figure 4 shows average values for the initial peak phase of the vasodilator response. For all trials, the initial peak in CVC attained ~80% of maximum and was thus consistent with previously reported values for this peak (5, 21). This initial peak value was unaffected by acute hyperglycemia. There was a slight, but significant, increase in initial peak CVC during the 180- and 360-min trials; as with the above, however, this did not appear to be an effect of hyperglycemia because it was seen in all groups, including those maintained at euglycemia. As mentioned above, this slight augmentation of the initial peak may have been due to a sensitization of the skin the local heating; this possibility requires further investigation.

View larger version (29K): [in this window] [in a new window]   Fig. 4.   Initial peak vasodilation seen during the first 1-3 min of the cutaneous vasodilator response to local warming. Average values ± SE are shown for all trials and all protocols. There was a slight but significant augmentation of this initial peak at 180 and 360 min; however, because this also occurred during euglycemia, it was probably not an effect of the hyperglycemia itself.

Systemic Hemodynamics

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The main new finding of the present study is that acute changes in systemic plasma glucose concentrations of a magnitude that presumably commonly occur after eating in individuals with impaired glucose tolerance or during the night in individuals with mild Type 2 diabetes mellitus do not alter the vasodilator response to local warming of the skin in healthy humans. This is of interest because the degrees of hyperglycemia created in the present study correspond to those shown to represent significantly increased risk for vascular disease in population-based studies (8, 20). The present data indicate that up to 6 h of sustained mild hyperglycemia or 4-5 h of postprandial hyperglycemia do not modify baseline skin blood flow or the ability to vasodilate during local warming.

We do not, however, exclude the possibility that longer episodes of hyperglycemia, or repeated episodes of similar duration, could have an effect on the cutaneous microcirculation over time. In the forearm circulation, higher levels of "local" hyperglycemia caused decreased endothelium-dependent vasodilation (3, 30). It is possible that higher levels of hyperglycemia in the present study would have resulted in impairment of the cutaneous vasodilator response to local warming.

Recent studies of the cutaneous microvascular response to local warming in humans provide important information regarding the mechanisms of this response. Regarding the second, prolonged phase of the response, Kellogg et al. (15), as well as Minson et al. (21), demonstrated that local administration of the nitric oxide synthase inhibitor nitro-L-arginine methyl ester via microdialysis causes a 65-75% reduction. Although the remaining 25-35% of the vasodilation remains unexplained, the findings of these investigators make local warming of the skin a useful tool to study the endothelial nitric oxide system in the human cutaneous microcirculation. The present experimental design did not allow us to distinguish between the nitric oxide-mediated and non-nitric oxide-mediated portions of the vasodilation; however, the fact that the response was unaltered suggests that both portions were unaffected. The lack of effect of hyperglycemia on the vasodilator response as a whole also indicates, more generally, that blood vessel function "downstream" from the endothelium remained intact as well.

It is clear that the cutaneous microvascular dysfunction seen in diabetes is due in large part to endothelial dysfunction. For example, individuals with established diabetes have decreased cutaneous microvascular responses to local iontophoresis of acetylcholine (2, 17, 22, 29). This decrease in endothelium-dependent vasodilation is associated with decreased expression of endothelial nitric oxide synthase in the skin of diabetic subjects (4, 29). Epidemiological data show a clear relationship between hyperglycemia and microvascular disease, when assessed as either fasting plasma glucose or response to an oral glucose tolerance test (8, 20). In this context, we questioned whether acute hyperglycemia over several hours has deleterious effects that could contribute to the development of this pathology. We addressed this question using transient hyperglycemia such as that seen after a carbohydrate meal in an individual with impaired glucose tolerance or in an individual with impaired fasting glucose or early Type 2 diabetes. We hypothesized that acute hyperglycemia would decrease the prolonged, largely nitric oxide-dependent vasodilator response to local warming of the skin. The present results argue against our hypothesis.

Our results contrast with the conclusions of Akbari et al. (1), who measured the cutaneous microvascular vasodilator response to iontophoresis of acetylcholine in healthy subjects before and 1 h after ingestion of 75 g of glucose. They concluded that the vasodilator response was diminished during acute hyperglycemia because the percent increase from baseline was less after glucose ingestion than in the fasting state. However, their data are open to several interpretations (27). First, the baseline level of skin blood flow was increased after glucose ingestion. Therefore, even though they report the same absolute level of vasodilation between fasting and postprandial measurements, this value was necessarily a lower percent of the higher baseline in the postprandial state. Second, these investigators apparently did not control (or report measurements of) insulin levels between fasting and postprandial states. Insulin, itself a nitric oxide-dependent vasodilator (25), may have confounded the results (for example, it could have been the reason for the higher baseline blood flow in the postprandial state).

In the present study, we expressed CVC data as a percentage of the maximum CVC seen during local warming. This value represents the level of conductance during complete relaxation of vascular smooth muscle and was not different among groups (14). This allowed us to compare the vasodilator response among groups without the confounding influence of shifting baselines seen in the study of Akbari et al. (1) and Title (27). Furthermore, in the present study, insulin was controlled at the same level across all protocols and time points to minimize any potential effects it might have on the local vasodilator response.

In an attempt to minimize the effects of endogenous insulin release, Houben et al. (12) studied the effects of "local" hyperglycemia (intra-arterial glucose infusion over 24 h) on skin blood flow and total forearm blood flow, and on vasodilator responses to intra-arterial acetylcholine. These investigators concluded that acute hyperglycemia had no effect on baseline forearm or skin blood flow or on the vasodilator response to acetylcholine in either of these vascular beds. However, analysis of the skin blood flow data presented in that study (12) led us to question whether the acetylcholine, delivered intra-arterially, ever reached the skin in sufficient concentrations to cause vasodilation. The negative conclusion of those authors was based on a complete lack of vasodilation to acetylcholine in the skin, either before or during hyperglycemia. Acetylcholine delivered directly to the skin causes substantial cutaneous vasodilation via activation of cholinergic muscarinic receptors (16). Thus the lack of vasodilation in the study by Houben et al. likely means that sufficient acetylcholine did not reach the skin area being measured. To avoid any such potentially confounding factors in the present study, we utilized a stimulus specific to the skin and measurement of skin blood flow directly at the site of that stimulus.

The initial peak phase of the vasodilator response to local warming was recently shown by Minson et al. (21) to be dependent on local sensory nerves. This is most likely via an axon reflex mechanism, because the vasodilation does not require intact connection to the central nervous system (21, 23) but was abolished by local application of anesthetic (EMLA cream) (21). Axon reflex mechanisms probably also account for the indirect vasodilation in response to iontophoresis of acetylcholine; this response has been shown to be diminished in diabetes (2, 17). Our results suggest that sensory nerve-dependent vasodilation in the skin was not immediately affected by the hyperglycemia protocols of the present study.

We observed a time-dependent alteration in the vasodilator response that was not a function of hyperglycemia, because it was also seen in the euglycemia control group. That is, the initial peak was slightly augmented and the nadir was absent in the fourth trial of each protocol compared with the first three. This phenomenon may be related to sensitization of the skin or sensory nerves to local warming over several trials; however, the mechanisms for these changes are unknown.

In the present experiments, we studied the effect of transient hyperglycemia per se in healthy volunteers. Although we did not observe an effect on cutaneous microvascular function, it is important to note that individuals with impaired glucose tolerance and diabetes are exposed daily to rising and falling glucose concentrations, as well as to transient changes in insulin concentrations (absent in the present study). Potential influences of these repeated transient changes in glucose and insulin levels on microvascular control and the mechanisms involved in such effects remain unclear at this time. Additionally, the manner in which acute and chronic hyperglycemia interact with other factors (e.g., hypertension, hyperlipidemia) likely to be present in diabetics or in those at risk for diabetes is presently unknown.

In summary, neither the sensory nerve-dependent initial peak nor the largely nitric oxide-dependent second phase of the cutaneous vasodilator response to local warming was altered by either an acute increase in glucose to ~11.1 mmol/l or by 6 h of sustained modest hyperglycemia. These data indicate that a transient increase in glucose to concentrations commonly observed in people with either impaired glucose tolerance or impaired fasting glucose do not alter nitric oxide-dependent or axon reflex vasodilation in the skin. However, it remains to be determined whether repeated similar rises or less frequent more marked increases in glucose contribute to the cutaneous microvascular endothelial dysfunction typical of Type 2 diabetes mellitus.