INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

MICROVASCULAR DISORDERS are prominent in systemic sclerosis. Typical microvascular abnormalities are capillary loss and dilatation, luminal narrowing of arterioles and small arteries with consequent severe Raynaud's phenomenon, and tissue ischemia. Microvascular studies (1, 14, 16) demonstrated significantly decreased skin blood flow in systemic sclerosis compared with that in healthy subjects. Flow velocity is also markedly reduced in the giant capillary loops typical of the disease (16, 20). Although most research work has focused on the abnormalities of the vessel wall, several studies also demonstrated the presence of rheological disturbances (14, 17, 18, 26, 34), possibly participating in the pathogenesis of ischemia in systemic sclerosis. These studies were focused on the measurement of plasmatic viscosity, steady blood viscosity, or thixotropy at moderate shear rates (>0.3 s¹). Measurements at still lower shear rates would be of interest because they would provide information on blood critical threshold of stress, referred to here as blood yield stress. This critical threshold is representative of the formation of a weak percolating physical gel. It could lead to extremely hindered flow situations and even to compaction states (9), depending on the cohesion of the clusters in the network of aggregates. As a matter of fact, measurement of blood yield stress is of critical importance in the pathophysiological evaluation of the microvascular ischemic diseases, but accurate measurements can hardly be obtained in the physiological range of hematocrits (25). As was experimentally evidenced, rheometric measurements at shear rates <1 s¹ with conventional Couette measuring devices are disturbed by slip and migrational effects (7, 8) and show a stress decay during shearing duration. We recently showed (27, 28) that measuring systems with different surface roughness (32 and 170 µm) are able to mitigate migrational and slip effects at low shear rates in normal blood as well as in blood samples with high fibrinogen levels.

The aim of the present work was to determine blood yield stress using low-shear-rate rheometry with a roughened Couette device in patients with systemic sclerosis compared with healthy controls. Additional data (blood biochemistry and microscopy) were obtained to preliminarily investigate the possible explanations for an abnormal red blood cell (RBC) organization mechanism in systemic sclerosis.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Subjects. The study was carried out according to International Committee for Standardization in Hemorheology guidelines (13) in 20 women: 10 patients with systemic sclerosis (age 34-71 yr, mean age 53.5 yr) were compared with 10 healthy controls (age 34-61 yr, mean age 52 yr). All were nonsmokers. Diagnostic criteria were the following: systemic sclerosis (scleroderma) patients (SSc), patients with Raynaud's phenomenon meeting the American Rheumatism Association criteria for systemic sclerosis (1a) and showing nailfold capillaries with typical scleroderma pattern; healthy controls (HC), donors from the Grenoble Blood Transfusion Center.

Preparation of blood samples. A blood sample was obtained by venipuncture on the occasion of routine biological analysis, withdrawn, and anticoagulated with EDTA. Hematocrit determination was performed automatically with a Coulter device (cell counting). Suspensions of RBC in autologous plasma were prepared from whole blood centrifuged at 3,000 rpm for 10 min. Plasma was partially removed to obtain the cell concentration wanted. Because in vivo hematocrit can vary over the range from 0 to 90-100%, according to the vessel type, size, and position in the lumen (31), two cell concentrations were studied, 45 and 60%. This choice was made for technical reasons and because these concentrations are representative of distinct RBC organization processes. At 45% hematocrit cell-protein interactions are of prime importance (3), and perturbing effects are very common (7) in low-shear-rate rheometry. At 60% hematocrit cell-cell interactions are increased, and perturbing effects are more limited (27). High concentrations of RBC are expected to be present during microcirculation stasis (9). Measurements at lower hematocrits would also have been of physiological interest because 10-20% cellular fractions are also present in the microcirculation (15), but they are not reachable at low shear rates with the rheometric devices, because of their sensitivity limits.

Rheometry. Stress measurements were made using the Contraves Low Shear 40 rotating coaxial rheometer working under controlled velocity conditions. All measurements were carried out with a homemade roughened measuring system (measuring cup radius = 7.34 mm, measuring bob radius = 6.26 mm, and bob length = 18 mm). The internal and external walls were roughened by sticking silicon carbide particles (average size 200 µm) on double-sided waterproof adhesive that was firmly stuck to the stainless steel surfaces. The device, including geometric factor calculations and surface roughness evaluation, has been described in greater detail previously (25, 27). Surface roughness was measured with a roughness tester and found to be 170 µm. For this geometry, the theoretical minimal shear stress measured was 0.56 mPa. Calibration was performed with a silicon oil of known viscosity. The temperature of the outer cylinder was regulated with a controlled water bath (25°C). The rheometer was connected to a computer for automatic operation and data acquisition.

Methodology. Measurements were made at 25°C with a volume sample of 4 ml. The internal static cylinder was centered with distilled water in the most sensitive range. The torque exerted on the internal bob was measured via a deflection system in the measuring head. Measurements at shear rate >0.3 s¹ were made by decreasing the shear rate from 100 to 3 × 10² s¹ in eight steps and reading the stationary value for each step. Particular attention was paid to data acquisition at lower shear rates using the following procedure to have the same initial conditions: 1) 15-s preshearing at 30 s¹; 2) short resting period (10 s); and 3) application of the given shear rate for a given period (summarized in Table 1). Blood was stirred between data acquisition tests. Results were analyzed in terms of total net shear deformation instead of time (net shear deformation equals shear rate times duration:  =  × t₀, which is a dimensionless parameter). Acquisition time for each shear rate was chosen to have a deformation of at least 2.5 (25, 27), except for the lowest shear rate of 10³ s¹. A 20-min acquisition time was deliberately chosen to ensure that the whole procedure time did not exceed 50 min. Values obtained at 3 × 10² s¹ by decreasing shear rate or by using the protocol described above were found to agree within 15%. Three parameters were derived from the stress deformation curves (Fig. 1): maximum stress (_m), extrapolation to zero time (₀) obtained in case of stress decay, and slope of this decay (s_D) were also given. Stress decay during shearing, when it occurred, was caused by migrational and slip effects along the rheometer walls (29).

View larger version (15K): [in this window] [in a new window]   Fig. 1.   Three parameters derived from stress deformation curve: maximum stress (m), stress extrapolated to zero time (0), slope of stress decay if present (sD in mPa per unit of deformation; when not observed, m = 0). Experimental curves were plotted in terms of shear stress against total net shear deformation instead of duration ( =  × t0, dimensionless parameter). This parameter is commonly used in rheometry and allows different plots to be directly compared because samples have seen same total net shear deformation.

1

Biochemical parameters. Fibrinogen concentration was measured by a thermocoagulation method (11), and serum protein concentrations were obtained from standard plasma protein electrophoresis.

Microscopic observations. Some blood samples were observed with an optical microscope (Axioskop, Zeiss) with PlanApo ×10, 0.25 NA, and PlanApo ×40 oil, 1.0 NA immersion objectives. The images were magnified with a ×1.6 lens and photographed with a Nikon camera. One drop of the suspension was set with a capillary tube on a microscope slide. A coverslip was carefully laid on the top and held down with a finger. In these conditions the thickness of the blood sample could not be precisely controlled, but it was estimated at 40 µm with a micrometer (Palmer) and a monolayer of cells could be observed. Photomicrographs were taken at ×160 and ×640 magnifications after 3 min of stasis. Preliminary results (not shown) of image processing on these photographs enabled us to quantify the structures formed and to evaluate the branch width of the networks. This parameter represents a mean width of the clumps; the more clustered the network, the higher the parameter.

Statistical analysis. Statistical analysis was performed with SPSS for Windows (release 7.01). Differences between SSc and HC were evaluated through the Mann-Whitney nonparametric test. A type I error probability <0.05 was considered significant. Results are expressed as means ± SE. In evaluating the relationship between blood yield stress and plasmatic proteins, we used the determination coefficient (square of the correlation coefficient) as the proportion of the total variance of yield stress statistically explained by plasmatic proteins.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Comparability of groups. Comparison of scleroderma patients (SSc) versus healthy controls (HC) is shown in Table 2.

Range of reliability. At 45% hematocrit, the shear stress versus shear curves (Fig. 2) had a mild slope for shear rates >10² s¹, but there was a sudden break for shear rates <10² s¹. Slip and migrational effects present below 10² s¹ lead to a steep decay (s_D > 1.5) and explain the fact that the maximum stress value was not representative of the true stress value, which should be much higher. Stress decay was statistically more pronounced for SSc. Figure 3 shows an example of stress-deformation curves at 10² s¹ and at 3 × 10² s¹ for an HC and SSc patient 6, who showed the strongest perturbing effects (this patient had severe systemic sclerosis and died a few weeks after the study). The mean decay slope was <0.9 mPa per unit deformation at 10² s¹ for both groups, and differences between _m and ₀ did not exceed 25%. At 3 × 10² s¹, the difference between _m and ₀ was always <10%. Therefore, at 45% hematocrit, the mean stress values given in Table 3 were considered to be reliable for shear rates >10² s¹.

View larger version (17K): [in this window] [in a new window]   Fig. 2.   Maximum shear stress as a function of shear rate for both groups at 45% (open symbols) and at 60% (closed symbols) hematocrit. Circles, healthy controls (HC); triangles, scleroderma patients (SSc). Standard errors are not represented because they were low and could not be seen on plot. Experimental points at shear rates <3 × 102 s1 are the maximum stresses derived from stress-deformation curves as indicated in Fig. 1. Results for both groups are given in Table 3.

View larger version (23K): [in this window] [in a new window]   Fig. 3.   Shear stress as a function of deformation at 45% hematocrit at 102 s1 and at 3 × 102 s1. Open symbols, HC sample 2; closed symbols, SSc patient 6, who showed most severe stress decay. Stress decay was statistically more pronounced for SSc at both shear rates. Mean of 10 samples for each group was considered reliable because stress decay never exceeded 0.8 mPa per unit deformation. Moreover, difference between 0 and m was always <25%.

3

1

3

1

3

1

View larger version (21K): [in this window] [in a new window]   Fig. 4.   Stress as a function of deformation at 60% hematocrit in shear rate range from 103 s1 to 3 × 102 s1 for an HC (A) and an SSc patient (B), who showed most perturbed rheometric behavior. For this patient, migrational and slip effects were very pronounced at shear rates of 103 s1 and 3 × 103 s1 but were mitigated for greater shear rates.

Maximum stress and yield stress at shear rates <3 × 10² s¹. Maximum stress was significantly increased for SSc over the whole range with also a higher intragroup heterogeneity. The differences between SSc and HC for the parameters _m and ₀ were more pronounced at 60% than at 45% hematocrit (Table 3).

2

1

2

1

View larger version (13K): [in this window] [in a new window]   Fig. 5.   A: hematocrit 45%. Maximum shear stress measured at 102 s1 for SSc and HC. At this hematocrit, measurements could not be considered reliable at lower shear rates because of strong perturbing effects and high slopes of decay (>1.5; see Table 3). B: hematocrit 60%. Yield stress at 60% (estimated by value of 0 at 103 s1) for SSc and HC. Each point represents 1 subject. Yield stress was significantly increased for SSc. N, no. of subjects.

3

1

3

1

3

1

3

1

3

1

Microscopic observations. Figure 6 shows the organization of RBC in four blood samples that had different rheometric behavior: an HC (Fig. 6A), SSc patient 7 (Fig. 6B), whose rheometric behavior was close to that of controls, SSc patient 6 (Fig. 6C), and SSc patient 9 (Fig. 6D). The samples in Fig. 6, C and D, had the most abnormal rheometric behavior with the most pronounced stress decay at low shear rates. For the samples in Fig. 6, A and B, linear rouleaux of different sizes were formed and were interconnected at a few contact points. Therefore, the aggregates were composed of linear rouleaux. In Fig. 6, C and D, clusterlike aggregates were formed, in which the arrangement of RBC was not linear. A circular aggregate was visible (point 1), and several connections linked the clusters to each other.

View larger version (133K): [in this window] [in a new window]   Fig. 6.   Photomicrograph of blood at ×640 magnification. A: HC no. 5 at 37%. Linear rouleaux are visible. B: SSc patient 7 at 38%. Same appearance as for A. C: SSc patient 6 at 40%. Clusters of compacted red blood cells were observed (point 1). D: SSc patient 9 at 45%. Clusterlike aggregates are visible.

Correlation between plasmatic proteins and yield stress. The correlations between plasmatic proteins (fibrinogen, albumin-to-globulin ratio) and blood yield stress at 60% are shown in Fig. 7. The best (negative) correlation was obtained for the albumin-to-globulin ratio.

View larger version (17K): [in this window] [in a new window]   Fig. 7.   Correlation between plasmatic proteins and blood yield stress (estimated by value of 0 at 103 s1) at hematocrit 60%. A: fibrinogen. B: albumin-to-globulin ratio. Best (negative) correlation was found for albumin-to-globulin ratio (R = 0.81).

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

Reliability of measurements. Measurements showed a strong time dependency at shear rates below 10² s¹ at 45% hematocrit, but they were reliable above 10² s¹. Torque decay is explained by slip and migrational effects of the aggregated structure along the rheometer walls and was limited by the use of roughened surfaces (25, 27); in a previous study (28, 29), we also showed that this phenomenon was influenced by the plasma fibrinogen concentration. Perturbing effects were greatly limited when the hematocrit was increased to 60%, even if some pathological blood samples still showed some stress decay. Thus, in most cases, yield stress could be safely estimated directly at 60% hematocrit. The inaccuracy of rheometric measurements at the 10³ s¹ shear rate is estimated at 15% for a measured stress of 6 mPa. Yield stress may thus be slightly underestimated compared with the true value at rest because a longer period of data acquisition may be required to ascertain that a steady value was reached (>20 min).

Role of plasmatic and cellular factors. The results presented in this study demonstrate that blood yield stress and stress levels at low shear rates were significantly higher for SSc than for HC. Conversely, microscopic observations showed the presence of clusterlike aggregates for some blood samples, as opposed to the rouleaux aggregates formed by normal RBC. The structure of samples shown in Fig. 6, C and D, seemed to be more cohesive, and higher stress levels are expected to be required to disperse them. These illustrative results are very preliminary, and further work has been initiated in image analysis to quantify the structural organization of the samples.

Consequences for microvascular blood flow in scleroderma patients. It has long been known that increased aggregability induces the formation of larger than normal RBC aggregates that are resistant to disaggregation by flow, particularly in the microcirculation (2a). Recently, Cabel et al. (4) showed, by direct measurements in cat skeletal muscle, that RBC aggregation was the main factor influencing venous vascular conductance and that it decreased conductance by 50% at normal flow rate (5 ml · min¹ · 100 g¹). RBC aggregation parabolically enhances blood viscosity, which may initiate a self-accelerating vicious circle leading to the formation of sludge blood (6). This is the case for scleroderma patients, for whom microvascular blood flow disorders, giant capillaries, and low flow have been evidenced (5, 12, 16, 20). The fact that capillaries are abnormal and larger than those of healthy subjects may favor aggregate formation and further flow reduction. Our results also suggest an increased RBC organization mechanism, dependent on plasmatic proteins, in addition to the vessel abnormalities. In this vulnerable microvascular condition, an elevated yield stress may be a highly contributive factor for the reduction of tissue perfusion, which leads to ischemia and which deserves attention in the therapeutic approach of such patients.

2

1

3

1