Impaired Exercise-Induced Blood Volume in Type 2 Diabetes With or Without Peripheral Arterial Disease Measured by Continuous-Wave Near-Infrared Spectroscopy

Subjects were recruited through the vascular laboratory at the University of Pennsylvania and from community exercise programs. The study was approved by the University of Pennsylvania Institutional Review Board. After signing informed consent, a total of 74 volunteers over the age of 50 years were recruited into four groups: 1) healthy control subjects (n = 25, 60% female), 2) diabetes only (n = 17, 18% female), 3) PAD only (n = 13, 23% female), and 4) diabetes and PAD (n = 19, 16% female). The diabetes-only group was defined as those patients diagnosed with type 2 diabetes by a physician and an ABI of >0.9 and who had no claudication symptoms. The PAD-alone group was defined as those patients with an ABI of ≤0.9 and symptoms of claudication. The patients with PAD and diabetes were those with an ABI of ≤0.9 and symptoms of claudication, as well as a previous diagnosis of type 2 diabetes.

The ABI was measured after the subjects were supine on an examination table for 15 min. Subjects refrained from any significant physical activity for at least an hour before the test. Standard blood pressure cuffs were placed on both arms and both ankles, and systolic pressure was measured with a handheld continuous-wave Doppler machine (Imex Elite; Nicolet Vascular, Conshohocken, PA). The ABI is calculated as a ratio of the ankle-to-arm measurement of systolic pressure as previously published (8). The higher of the arm pressures is used as the denominator and the higher of the ankle pressures is used as the numerator for each respective right and left ABI.

The NIRS measurements were made using a continuous light source, dual-wave-length spectrophotometer (Runman) as previously published (9). The NIRS probe was placed on the belly of the medial gastrocnemious. For those with PAD, the probe was placed on the most symptomatic leg, and if bilateral claudication was present, the probe was placed on the leg with the most severely reduced ABI. After application of the probe, continuous measurements were obtained at baseline and during exercise. Subjects initially performed a 30–plantar-flexion exercise over 1 min. After resting for 10 min, subjects walked on a treadmill according to the Gardner protocol (10). The subjects continued to the point of maximum claudication or up until 10 min of exercise. Of note, all hemodynamic parameters were allowed to return to normal before the next exercise phase of the study.

The total hemoglobin/myoglobin content (correspondent to blood volume in the vascular territory of the underlying muscle) and relative oxyhemoglobin amount (compared with resting baseline) was measured and graphed continuously during the exercise protocol. The blood volume expansion was calculated as the increase in measured volume over time (slope) during treadmill walking after the initial decrease due to muscle contraction. This calculation reflects the expansion of capillary blood volume. The blood deoxygenation was calculated as a percent change from baseline of measured oxyhemoglobin/myoglobin during exercise. The oxygen recovery time (T₅₀) was calculated as the time it takes for oxygenation to return to half baseline after termination of exercise.

Linear models were used to evaluate difference among groups for each outcome of blood volume expansion, oxygen recovery time, and percent deoxygenation after plantar flexion or treadmill exercise. Since patients were not randomized to the four groups, group comparisons for NIRS measurements were adjusted with respect to potential confounding factors, including age, race, sex, BMI, smoking status, and hypertension. P values <0.05 were considered statistically significant. All analyses were carried out using SAS statistical software (version 9.1).

The demographics of the study population are listed in Table 1. Of those enrolled in the study, there were more men than women with diabetes and PAD. The mean ABI was not significantly different between the PAD-alone and PAD with diabetes group (P = 0.44).

A representative NIRS tracing is depicted in Fig. 1. The slopes of the mean values for blood volume expansion were calculated for all groups (Table 2). Group comparisons were adjusted for BMI and smoking status. Compared with healthy control subjects, the diabetes-only group did not show a significant difference in the change in blood volume during treadmill exercise. Their slope difference is −0.0002 (95% CI −0.0027 to 0.0023). The presence of PAD without diabetes was associated with a significant increase in the volume compared with control subjects (P = 0.039), and the difference between them in terms of the rate of blood volume expansion was 0.0032 (0.0002–0.0062). However, diabetic patients with PAD had an attenuated blood volume response compared with PAD-only patients (P = 0.04). The rate of blood volume expansion for diabetes with PAD patients was 0.0031 units smaller than for PAD-only patients (with 95% CI −0.0064 to −0.0001). The use of vasoactive medications (β-blockers, calcium channel blockers, ACE inhibitors, and angiotensin receptor blockers) had no significant effect on blood volume (data not shown). The blood volume was not measured during plantar-flexion due to the brief period of exercise (1 min), which does not allow adequate time for responses unconfounded by blood volume changes from onset and cessation of exercise.

The mean values for T₅₀ were calculated for all four groups to determine the relative oxygenation recovery time (Table 2). While diabetes had no effect (the ratio of T₅₀ for diabetes patients relative to healthy controls was 0.9 [95% CI 0.61–1.4]), the presence of PAD significantly increased the T₅₀. Compared with healthy control subjects, oxygen recovery time was 2.6 (1.6–4.1) and 2.4 (1.6–3.6) times longer in PAD-only patients and patients with both PAD and diabetes, respectively. After adjusting for smoking status, race, and personal history of coronary artery disease, these comparisons remained significant.

The mean values for deoxygenation were calculated in all four groups to determine the relative amount of desaturation of hemoglobin/myoglobin in the tissue (Table 2). The presence of claudication was associated with a significant increase in deoxygenation during both treadmill walking (P = 0.03) and plantar-flexion (P = 0.02) compared with healthy control subjects (P = 0.03). There was no significant increase in deoxygenation for the group with only diabetes during treadmill exercise or plantar-flexion compared with healthy control subjects (P = NS).

This is the first report of impaired blood volume expansion with exercise in patients with diabetes despite a normal ABI and relatively normal parameters of deoxygenation and recovery times. The impaired blood volume expansion likely reflects a decrease in the vasodilatation capacity of capillaries in the skeletal muscle of the lower extremity during physical activity. Patients with diabetes are known to have endothelial dysfunction (11,12), and it is likely that microvascular dysfunction contributes to the impaired blood volume expansion seen in this study. Our study results also show that patients with PAD, but without diabetes, have increased blood volume during exercise compared with healthy control subjects, likely reflecting a greater local hypoxemia, causing a vasodilatation response. The hypoxemia may cause release of vasodilating factors such as lactate, CO₂, and adenosine. The data indicate that patients with PAD and diabetes lose this ability to increase blood volume, indicating a severely blunted microvascular vasodilator response. Similar to a previously published study (6), the deoxygenation and recovery times were abnormal in patients with PAD.

Tissue oxygenation, defined as relative saturation of oxyhemoglobin and myoglobin, depends on the balance between oxygen delivery, as reflected by the product of blood flow and oxygen content and metabolic rate or oxygen consumption. The normal response during exercise is a gradual decline in tissue oxygen saturation with increasing metabolic demand, and the initial response depends on exercise intensity. Our study results show that percent deoxygenation of hemoglobin and myoglobin occurs rapidly in both normal control subjects and those with PAD alone or with diabetes. The plateau level for percent deoxygenation is lower in the control group than in the patients with PAD, whereas the plateau for patients with diabetes and no PAD was not significantly different from healthy control subjects. The T₅₀, a measure of oxygen recovery, is longer in patients with PAD compared with control patients, whereas the blood volume expansion is higher in PAD than control subjects and significantly lower in those with diabetes.

The pathologic influences resulting from the hyperglycemic state and resulting in endothelial dysfunction are multifactorial and include advanced glycation end products, increased formation of oxygen-derived free radicals, and activation of protein C (13). The presence of PAD results in reduced endothelium-derived vasodilator prostanoids and nitric oxide (4). The current study results indicate that a unique aspect of the diabetic state is impairment of blood volume during exercise that does not appear evident in patients without diabetes and hemodynamically significant lower extremity atherosclerosis. Our results are consistent with those of Kingwell et al. (14), who found that limb blood flow response to acetylcholine and exercise were significantly attenuated in patients with diabetes compared with healthy control subjects. Also, others have shown that impaired limb blood flow response to exercise may limit exercise capacity in patients with type 2 diabetes (15–17).

In conclusion, our study results indicate that the impaired blood flow response occurs early in diabetic vasculopathy and may occur before hemodynamically significant PAD. These findings may partly explain the accelerated atherosclerosis observed in patients with diabetes. Further studies are needed to determine whether impaired blood volume expansion in the diabetic state portends increased risk of developing PAD.

We thank Elizabeth Medenilla for her assistance with treadmill testing.