Increased Shear Rate Resistance and Fastest Kinetics of Erythrocyte Aggregation in Diabetes Measured With Ultrasound

Recruited individuals were nonsmoking males. They completed a questionnaire on current medications and medical history. BMI and blood pressure were measured. Nine patients with type 2 diabetes and eight healthy control subjects gave informed consent to the approved protocol. Patients with poor metabolic control were intentionally chosen. Lipid profile and inflammatory proteins (fibrinogen, Von Clauss method; haptoglobin and immunoglobin G, immunonephelometric method; and C-reactive protein, latex agglutination technique) were determined for each participant.

Six patients were on oral antidiabetic and aspirin medications, three were on insulin, five were on cholesterol-lowering therapy, and four were treated for hypertension. Those on insulin had a history of coronary artery disease, and two of them had suffered from myocardial infarction. Distal angiopathy or cutaneous trophic disorders were not present. Mean ± SD age of patients was 58.3 ± 8.8 years (range 41–70). Healthy subjects were age matched (51.1 ± 8.7 years [range 41–64]), did not take regular medications, and had no history of cardiovascular disease. None had lipid disorders or hypertension.

A Couette instrument made of two concentric cylinders and installed in an incubator at 37°C generated homogeneous shear rates via a rotating outer cylinder. Fifty milliliters of EDTA anticoagulated blood at 40% hematocrit was introduced between both cylinders. The stationary inner cylinder held an ultrasound transducer perpendicular to flow. The 32-MHz polyvinylidine flouride ultrasound transducer (Visualsonics, Toronto, Canada) had a −6 dB bandwidth of 15–45 MHz. It was pulsed with bipolar square waves (model no. AVB2-TA-C-CRIMA; Avtech, Ottawa, Canada). Received radio frequency (RF) echoes (model no. AU-3A-0120; Miteq, Hauppauge, NY) were amplified by 54 dB, filtered between 10 and 50 MHz (model no. 5900 PR; Panametrics, Waltham, MA), and digitized at 250 MHz (model no. 8500 CS; GageScope, Montreal, Canada).

The protocol consisted of imposing a 500 s⁻¹ shear rate for 120 s to disrupt aggregates. Then, aggregation kinetics were recorded for 380 s at randomly applied reduced shear rates of 0, 1, 2, 10, 50, 100, and 200 s⁻¹. One hundred RF echoes were acquired every 2 s. For each blood sample, the protocol was repeated three times for averaging. RF signals were adjusted to compensate for blood attenuation at each shear rate and integrated backscatter coefficient (IBSC) was determined as previously described (6). IBSC reflects the number of RBC per aggregate (7).

BMI, blood pressure, triglycerides, total cholesterol, HDL and LDL cholesterol, fibrinogen, and C-reactive protein were not different between groups (P > 0.05; unpaired t tests), whereas immunoglobin G (8.4 ± 1.0 vs. 11.3 ± 1.8 g/l, P < 0.001), haptoglobin (1.0 ± 0.4 vs. 1.6 ± 0.5 g/l, P < 0.05), and A1C (5.5 ± 0.8 vs. 8.8 ± 2.1%, P < 0.01) were significantly higher in diabetic patients than in control subjects. Couette flow protocol resulted in disaggregation of RBC and minimum IBSC at 500 s⁻¹, formation of aggregates and rapid increase in IBSC depending on reduced applied shear (Fig. 2 of ref. 7), and development of stable aggregate sizes and plateaus of IBSC after a few seconds for most shears to minutes at 0 s⁻¹.

Figure 1A summarizes mean raising slopes of IBSC between 2 and 8 s after shear rate reductions. In control subjects, maximum slope at 2 s⁻¹ was not different from that at 1 and 10 s⁻¹ (P > 0.89), whereas in diabetic patients, maximum slope of IBSC occurring at 2 s⁻¹ was similar to that at 1 s⁻¹ (P = 1.0). Except for that at 200 s⁻¹ (P = 0.11), kinetic slopes were always faster in diabetic patients, which is indicative of higher rates of neighboring RBC clustering.

IBSC at plateaus averaged between 180 and 380 s after shear rate reductions as presented in Fig. 1B. In control subjects, maximum IBSC at 2 s⁻¹ was similar to those at 0 and 1 s⁻¹ (P > 0.95). Similarly, IBSCs at 1 and 2 s⁻¹ were similar in diabetic patients (P > 0.95). IBSCs were statistically higher in diabetic patients between 2 and 100 s⁻¹, which reflects stronger adhesions and bigger steady-state aggregate sizes in diabetes.

Statistically significant differences in Fig. 1 were noted for shears between 2 and 100 s⁻¹, which correspond to normal flow at center streams and pathological flow stasis in recirculation zones of large systemic veins and arteries. Accordingly, EA in diabetes can be related to lower-limb artery ischemic events, microangiopathy in foot extremities, and retinopathy. Inflammation is involved in pathogenesis of type 2 diabetes and RBC aggregation, which agrees with our results (legend of Fig. 1). Subacute inflammatory state promoting RBC aggregation is also associated with obesity (8) and metabolic syndrome (9). Thus, reducing inflammation (and indirectly aggregation) with statins and A1C with antidiabetic medication and/or diet are indicated because both have known benefits to cardiovascular consequences of diabetes. We reported measurements from a laboratory instrument, but a short-term objective is sizing RBC aggregates in vivo with ultrasound (10). At 32 MHz, superficial (5–6 mm depth) vessels can be scanned. The proposed noninvasive method should be investigated further because it may have potential benefit for diagnosis and follow-up of diabetic foot complications and for monitoring therapy.

This study was supported by operating grants MOP-36467 and MOP-84358 of the Canadian Institutes of Health Research and the National Scientist award of the Fonds de la Recherche en Santé du Québec (to G.C.).

We thank Dr. F.S. Foster for providing the Couette flow system and D. Poisson for her assistance in blood sampling.