The effect of hematocrit,fibrinogen concentration and temperature on the kinetics of clot formation of whole blood

Abstract

BACKGROUND:

Dynamic mechanical analysis of blood clots can be used to detect the coagulability of blood.

OBJECTIVE:

We investigated the kinetics of clot formation by changing several blood components, and we looked into the clot “signature” at its equilibrium state by using viscoelastic and dielectric protocols.

METHODS:

Oscillating shear rheometry, ROTEM, and a dielectro-rheological device was used.

RESULTS:

In fibrinogen- spiked samples we found the classical high clotting ability: shortened onset, faster rate of clotting, and higher plateau stiffness. Electron microscopy explained the gain of stiffness. Incorporated RBCs weakened the clots. Reduction of temperature during the clotting process supported the development of high moduli by providing more time for fiber assembly. But at low HCT, clot firmness could be increased by elevating the temperature from 32 to 37°C. In contrast, when the fibrinogen concentration was modified, acceleration of clotting via temperature always reduced clot stiffness, whatever the initial fibrinogen concentration. Electrical resistance increased continuously during clotting; loss tangent (D) (relaxation frequency 249 kHz) decreased when clots became denser: fewer dipoles contributed to the relaxation process. The relaxation peak (Dmax) shifted to lower frequencies at higher platelet count.

CONCLUSION:

Increasing temperature accelerates clot formation but weakens clots. Rheometry and ROTEM correlate well.

Keywords

Clot stiffness kinetic temperature erythrocytes fibrinogen rheometry thrombelastometry dielectric test

1 Introduction

Recent trends in coagulation research, intensive care and perioperative medicine are shifting focus to whole blood viscoelastic haemostatic assays for diagnosis and treatment monitoring of haemostatic disorders [1, 2]. They are mainly used in cardiac surgery, acute trauma care or obstetrics and have been shown to decrease the use of blood products [3 –5], but efforts have been made to extend the range of application into anticoagulation monitoring [6] and diagnosis of states of hypercoagulation [7].

A large body of evidence relates the physical properties of blood clots, most importantly clot stiffness, to various pathologies: Decreased clot modulus, as indicated by thrombelastometry and increased incidence of perioperative bleeding, have been associated with hypo-fibrinogenemia [8], whereas increased plasma fibrinogen concentrations [9, 10] or fiber size [11] have been associated with increased cardiovascular risk. Whereas the gain in complexity of the architecture by the addition of fibrinogen will generate a stiffer clot [12], red blood cells (RBCs) can generate thrombin and affect clotting by several ways [13]. Maybe the most significant relates clot stability to the presence of RBC clusters inside the clot, which can locally weaken the material [14], and eventually lead to thromboembolic events.

A suitable tool to investigate the development of stiffness during clot formation, together with the material properties of fully formed clots, is oscillating shear rheometry, a method by which materials are subjected to well defined shear forces over a large order of magnitude, enabling the application of amplitude and frequency sweep tests or creep and stress relaxation tests [15 –17]. This sets rheometry apart from thrombelastometry, whose limitations are well known [18, 19]. Oscillating shear rheometry can operate at quasi-static conditions of the developing material, and thus will only minimally influence clot architecture. In demonstrating the relevance of fibrin on the shear stiffness of a clot, most studies have used fibrin clots without blood cells [20 –22]. Although the presence of RBCs in whole blood will certainly mask properties that can be detected in pure networks, blood cells being incorporated into the fiber network contribute to clot mechanical performance and reflect the physiological circumstance better than pure fibrin gels. Incorporated RBCs could increase clot stiffness by their membrane modulus, but, on the other hand, RBCs could also prevent the formation of a dense network by preventing fibers from developing at the same place. Another aspect from the clinical view is that centrifugation of a patient’s blood to gain plasma can modify platelet activity and would not perfectly show its current ability to coagulate. We therefore decided to use whole blood to investigate the influence of hematocrit (HCT) and plasma fibrinogen concentration on the stiffness of clots during quasi-static shear. In addition, we correlated rheometry and thrombelastometry to find limitations that might appear when rheometry is used to predict the coagulability of blood. Apart from the finding that RBCs weaken the clots, whereas the plasma fibrinogen concentration strengthens the networks, the time available for clot formation significantly alters their stiffness.

2 Materials and methods

2.1 Blood samples

A total of 37 healthy male volunteers (age: 23–33, BMI < 30, non-smokers, no intake of any medication for the last 7 days) were included in the study. After obtaining informed consent and with approval of the ethical review committee of the Medical University of Vienna (EK No. 1371/2015), blood was collected from the antecubital vein using a 21G butterfly needle and a Vacuette blood collection system (Greiner Bio-One GmbH, Austria), containing 3.8% sodium citrate for anticoagulation. Blood was stored at ambient temperature and tested after a maximum of 4 hours. Coagulation was started via stoichiometric re-calcification using a 0.2 M CaCl₂-solution (TEG^® Hemostasis System, Haemonetics, USA).

2.2 Study groups

In study group A (n = 8) the effect of hematocrit, and in study group B (n = 9), the effect of plasma fibrinogen on the kinetics of clot formation and clot stiffness were investigated. All sample processing was performed on the day of blood collection from the respective individual. Measurements were performed at 37°C.

For study group A, three hematocrit (HCT) levels (25%, native = 40%, 55%) were established. Whole blood was centrifuged (800 U/min, 20 min), plasma was separated and samples were generated by reconstitution of the RBC concentrate with platelet-rich and/or platelet-poor plasma, as well as with an aliquot of the buffy coat [23]. Aliquots were taken for routine hematological and blood coagulation tests. Great care was taken to maintain platelet count, leucocyte count, and plasma fibrinogen concentration at the values of the native sample.

For study group B, three fibrinogen concentrations (baseline (BL), medium (MED), high (HIGH)) were established. Fibrinogen powder (F3879; Sigma, Germany) was added (50 or 100 mg) to 9 mL of native citrated whole blood. The blood was gently mixed thereafter for 30 minutes by hand and aliquots were removed for routine laboratory tests. Even though we did not add a vehicle to our samples, a drop of platelet count could not be avoided, which could be the result of platelet aggregation after addition of fibrinogen (details can be found in [24]). In subgroup B, stock fibrinogen solutions (1 mg/mL; 2 mg/mL) were prepared by dissolving fibrinogen in saline according to the supplier’s recommendations. One mL of vehicle (LOW) or stock solution (MED, HIGH) was added to 9 mL of whole blood and the samples were gently mixed by hand.

In subgroups A and B the influence of temperature (32, 37, 42°C) was investigated in a limited number of samples (six individuals in group A, four individuals in group B). Samples were prepared as described above, divided in three portions, and tested at 32°C, 37°C, and 42°C after randomized selection of the temperature sequence. In these subgroups blood had to be withdrawn from individuals on three consecutive days to finalize the measurements on the day of blood sampling.

2.3 Viscoelastic tests

The Physica MCR 301 and 302 rheometers (Anton Paar, Austria) were used. Temperature was Peltier controlled and a tempered hood mounted the measuring system. A silicon oil filled evaporation blocker was used to prevent sample drying. The stainless steel plate-plate measuring system (50 mm diameter, gap height 1 mm) was filled with 2 mL citrated whole blood after re-calcification with 133μL of 0.2 M CaCl₂-solution (TEG^® Hemostasis System, Haemonetics, USA). Time sweeps were conducted at constant frequency (1.5 Hz) and deformation amplitude (0.001%). Frequency sweep tests (0.1–10 Hz, linear mode, logarithmic ramp) were performed when clotting was completed. In group B, we observed RBC sedimentation prior to the start of clotting due to the addition of fibrinogen. This was avoided in the tests provided here by setting a logarithmic strain ramp (100–0.01%) during the first 350 s. In group B, a 50 mm cone-plate symmetry was used and the amplitude was set to 0.01%. Storage modulus (G´), reflecting the elastic property of the sample was obtained from the stress-strain-relationship as follows: the dynamic shear modulus (G*) is calculated during the sinusoidal change of time and amplitude (G* = τ(ω, A)/γ(ω, A)). The phase shift angle (δ) shows the lag phase between the applied stress and the resulting strain, and indicates therefore the in-phase and out-of-phase components of G*. Multiplication of G* with cos (δ) determines the storage modulus (G´) [25].

ROTEM^® measurements were conducted according to the manufacturer’s specifications using the ROTEM^® delta (TEM International, Germany) in NATEM mode at 37°C for 1 hour. Samples were re-calcified with 0.2 M CaCl₂-solution (TEG^® Hemostasis System, Haemonetics, USA) using the built-in auto-pipette.

2.4 Influence of the surface profile on the shear moduli

To investigate the influence of the surface profile on the shear moduli we compared polished (PP25) against profiled (PP25/P3) stainless steel plates (both at 25 mm diameter, gap width 1mm, purchased from Anton Paar, Austria). Blood samples were withdrawn from eight volunteers, and three different fibrinogen concentrations (BL, MED, HIGH) were established as described above.

2.5 Scanning electron-microscopy (SEM)

Representative blood clots were analysed using a Quanta 250 FEG (E)SEM (Thermo Fisher Scientific, USA). After reaching the G´-plateau value samples were rinsed three times with phosphate buffered saline while still adhering to the upper measuring plate or cone, and fixed thereafter in situ in 3.5 % formaldehyde. After one hour of fixation they were carefully removed from the surfaces, rinsed three times in saline, followed by stepwise dehydration in ethanol-water solutions (30, 50, 70, 90, 98% ethanol). Samples were incubated in each ethanol-water solution for 20 minutes each, and finally stored in 98% ethanol until assayed. For SEM, they were dried in a vacuum oven for 30 minutes at 40°C and gold coated (Scancoat Six; Edwards, UK). The ESEM was operated under environmental conditions with the native sample, whereas SEM was operated under high vacuum conditions and with variable voltages between 5 and 20 kV. The secondary electron detector was used to generate the micrographs.

2.6 Dielectric tests

A pilot study was performed to show the development of the permittivity and the electrical conductivity with clot formation. The Dielectro-Rheological Device (DRD, Anton Paar, Austria) was connected to the rheometer to record rheometrical and electrical data simultaneously. Therefore, a stainless steel plate-plate measuring system with an insulated shaft (diameter: 50 mm, gap width: 1 mm) was used. Between the two plates a sinusoidal voltage of 1 V was applied and the capacitance C_p, the resistance R_p, and the loss tangent D were measured in a frequency range of 100 Hz to 2 MHz with a LCR meter (Keysight E4980A; Keysight Technologies, USA). The dielectric permittivity ɛ, that is proportional to the capacitance C_p, is a measure of the ability to store the energy of an applied electric field in the sample or medium. In case of an alternating electric field the polarization in the sample cannot follow the oscillation of the external electric field at higher frequencies; thus the permittivity is a complex value and a function of the angular frequency ω of the electric field $ɛ^{*} (ω) = ɛ^{'} (ω) - i ɛ^{″} (ω)$

ɛ^′ is the real part and ɛ^′′ is the imaginary part, that is proportional to the resistance R_p and reflecting the electric losses in the material. The dissipation factor D, also called loss tangent, $D = \tan δ = \frac{ɛ^{″}}{ɛ^{'}}$

describes the relationship of the real and imaginary part and provides information about the characteristics of the material when evaluated for several frequencies [26]. D has its maximum (relaxation peak) at the relaxation frequency f_r, that is inversely proportional to the relaxation time. A shift of the relaxation peak for various samples means that the dominant dipole of molecules or ions changes between these samples, and therefore changes the dipole moment of the bulk material.

Two sets of experiments were performed. First, fibrinogen was added to whole blood of a volunteer (male, 28 yr) as described above to generate BL, MED, and HIGH samples. Second, blood was withdrawn from another volunteer (female, 39 yr), and platelet count was lowered in blood plasma by mixing platelet rich plasma (PRP, containing 218.000 PLT mm^–3) with platelet-depleted plasma (1:4 and 1:64). After having reached the final clot stiffness, the loss tangent D was recorded in linear viscoelastic mode.

2.7 Routine laboratory tests

A routine hematological profile and coagulation profile was performed for each sample of the study groups. Routine hematology was performed with Sysmex XN-2000 (Sysmex, Japan), coagulation tests were performed with Sta R Max2 (Stago, USA). Fibrinogen was measured by the method of Clauss.

2.8 Data analysis and statistics

Rheometry data were exported from the RheoPlus (version 3.40) and Rheocompass (versions 1.17 and 1.19) software and processed in GraphPad Prism 6 (version 8.1; GraphPad, USA) for MacOS Sierra. The following parameters were extracted from the time sweep tests: maximum clot stiffness (G´ max, in Pa), time required to reach this final value (G´ max-t, in s), initial clot stiffness (G´ init, in Pa), time to reach initial stiffness (G´ init-t, in s), and normal force in the rheometer gap (F_N, in mN). Initial clot stiffness was calculated from the G´/ -time dependency by use of the first derivative’s maximum. Extrapolation of this maximum to the y-axis indicated G´ init, and extrapolation to the x-axis gave G´ init-t. We were unable to obtain a G´-plateau value in subgroup A tests. Therefore we used an algorithm provided by the GraphPad Prism 6 software that enables the calculation of the maximum (= Top) of our curves by sigmoidal interpolation of data points (Y = Bottom + (Top-Bottom))/-(1 + 10^∧((LogEC50-X) * HillSlope)). We used this method to extrapolate from our kinetic curves (X = time, Y = G´) to the plateau value when the kinetic curve clearly tended towards such a plateau. Since clot formation was more continuous, instead of gaining a plateau at some circumstance (high HCT associated with high temperature), we cannot provide a plateau value for these conditions in Table 2. In addition, to compare rheometry with ROTEM, two-tailed Pearson’s correlation coefficients were calculated. Paired two-sided t-tests were used to determine the differences between parameters and conditions (HCT, fibrinogen, temperature). One-way ANOVA was used to calculate differences in the development of G´, G˝, and F_N with time. Dielectric data were processed by MATLAB (MathWorks Inc., Version R2016a).

3 Results

3.1 Group A

Addition of RBCs lowered initial and final clot stiffness, delayed the onset and reduced the velocity of clot formation. Hematocrit also affected the normal force (F_N) in the gap during clotting. The negative F_N increased in parallel with the modulus. F_N was –394±75 mN (25% HCT), –303±85 mN (40% HCT), and –213±65 mN (55% HCT) after 30 minutes of clot formation. G´ max (rheometry) and MCF (ROTEM) correlated significantly (r = 0.64), but there was no relationship between G´ max-t and MCF-t. Tests for the clot’s resting stability revealed that G´-values were weak functions of frequency. Although G´-slope increased with HCT, all slopes remained very low (at 25% HCT: 0.026±0.007 Pa s^–1; at 40% HCT: 0.033±0.008 Pa s^–1; and at 55% HCT: 0.040±0.009 Pa s^–1). G˝ showed a minimum between 0.5 and 1 Hz, indicating optimized clot stability at this frequency range (all p < 0.001). Data are provided in Table 1 and Fig. 1.

Table 1
Clot formation parameters at 37°C obtained using time sweep in rheometry and ROTEM. Values are rounded to integers and given as mean±standard deviation. Differences in G´ max and G´ init in relation to HCT and fibrinogen are significant (p < 0.01)

Group A 25% HCT 40% HCT 55% HCT

G’ max [Pa] 414±73 252±26 182±20

G’ max -t [s] 2378±656 2307±401 2730±764

G´ init [Pa] 110±26 070±20 47±12

G´ init -t [s] 619±258 820±302 890±375

MCF [mm] 60±9 059±9 45±11

CFT [s] 186±120 177±92 418±316

Group B BL MED HIGH

G’ max [Pa] 282±27 454±83 543±98

G’ max -t [s] 2831±366 2329±267 1941±179

G´ init [Pa] 89±18 105±28 106±27

G´ init -t [s] 983±155 622±122 436±98

MCF [mm] 48±5 65±3 67±3

CFT [s] 358±61 118±24 118±110

Group A	25% HCT	40% HCT	55% HCT
G’ max [Pa]	414±73	252±26	182±20
G’ max -t [s]	2378±656	2307±401	2730±764
G´ init [Pa]	110±26	070±20	47±12
G´ init -t [s]	619±258	820±302	890±375
MCF [mm]	60±9	059±9	45±11
CFT [s]	186±120	177±92	418±316
Group B	BL	MED	HIGH
G’ max [Pa]	282±27	454±83	543±98
G’ max -t [s]	2831±366	2329±267	1941±179
G´ init [Pa]	89±18	105±28	106±27
G´ init -t [s]	983±155	622±122	436±98
MCF [mm]	48±5	65±3	67±3
CFT [s]	358±61	118±24	118±110

Fig. 1

Development of clot stiffness in whole blood samples with different HCT (a) and fibrinogen concentrations (b), smoothed using local regression. There is good match between rheometry and ROTEM (c).

Hematological values were maintained in all samples as follows: platelet count in native samples: 216±16, in samples with 25% HCT: 186±34, in samples with 55% HCT: 126±16 G/L; leucocyte count in native samples: 5.96±1.7, in samples with 25% HCT: 3.17±0.8, in samples with 55% HCT: 5.96±1.7 G/L; plasma fibrinogen concentration in native samples: 274±44, in samples with 25% HCT: 286±54, in samples with 55% HCT: 278±40 mg/dL. Prothrombin time (PT) and activated partial thromboplastin time (aPTT) were essentially unaffected as well: PT in native samples: 85±10, in samples with 25% HCT: 80±7, in samples with 55% HCT: 77±9 %; aPTT in native samples: 41±6, in samples with 25% HCT: 38±3, in samples with 55% HCT: 37±10 s.

3.2 Subgroup A

Increasing the temperature accelerated the start of clot formation. Generally, clot stiffness decreased when the temperature was raised, and when the HCT was increased - with one exception. At 25% HCT clot stiffness tended to increase with the temperature rise. Likewise, addition of RBCs up to physiological values (from 25% to 40% HCT) did not weaken the clot when the temperature remained at 32°C. All data p < 0.01. Data are provided in Table 2 and Fig. 2.

Table 2
Effect of temperature on clot stiffness. Data are given as mean±standard deviation. In subgroup A, we provide an estimation of G´ max (see section Materials and Methods) and display the mean value with the 95% confidence interval in parentheses. Since at 55% HCT and high temperatures the process was rather continuously ongoing and did not tend to reach a plateau, no G´ max value can be provided for several tests of subgroup A

Subgroup A 32°C 25% HCT 40% HCT 55% HCT

G’ max [Pa] 358 (336–380) 400 (379–421) 254 (233–274)

G´ init [Pa] 79±25 85±13 57±5

G´ init -t [s] 990±137 935±215 1148±156

Subgroup A 37°C 25% HCT 40% HCT 55% HCT

G’ max [Pa] 385 (327–443) 323 (309–338) n.a.

G´ init [Pa] 82±36 88±28 51±6

G´ init -t [s] 829±153 816±130 916±135

Subgroup A 42°C 25% HCT 40% HCT 55% HCT

G’ max [Pa] 246 (216–276) n.a. n.a.

G´ init [Pa] 102±15 61±20 45±15

G´ init -t [s] 799±95 684±128 640±113

Subgroup B 32°C LOW BL HIGH

G’ max [Pa] 285±67 292±35 399±67

G´ init [Pa] 70±31 82±18 107±25

G´ init -t [s] 657±254 808±162 695±145

Subgroup B 37°C LOW BL HIGH

G’ max [Pa] 277±44 303±33 360±46

G´ init [Pa] 67±29 75±15 95±33

G´ init -t [s] 391±134 520±138 450±58

Subgroup B 42°C LOW BL HIGH

G’ max [Pa] 193±36 276±22 377±73

G´ init [Pa] 57±22 69±46 76±19

G´ init -t [s] 318±165 507±129 447±48

Subgroup A 32°C	25% HCT	40% HCT	55% HCT
G’ max [Pa]	358 (336–380)	400 (379–421)	254 (233–274)
G´ init [Pa]	79±25	85±13	57±5
G´ init -t [s]	990±137	935±215	1148±156
Subgroup A 37°C	25% HCT	40% HCT	55% HCT
G’ max [Pa]	385 (327–443)	323 (309–338)	n.a.
G´ init [Pa]	82±36	88±28	51±6
G´ init -t [s]	829±153	816±130	916±135
Subgroup A 42°C	25% HCT	40% HCT	55% HCT
G’ max [Pa]	246 (216–276)	n.a.	n.a.
G´ init [Pa]	102±15	61±20	45±15
G´ init -t [s]	799±95	684±128	640±113
Subgroup B 32°C	LOW	BL	HIGH
G’ max [Pa]	285±67	292±35	399±67
G´ init [Pa]	70±31	82±18	107±25
G´ init -t [s]	657±254	808±162	695±145
Subgroup B 37°C	LOW	BL	HIGH
G’ max [Pa]	277±44	303±33	360±46
G´ init [Pa]	67±29	75±15	95±33
G´ init -t [s]	391±134	520±138	450±58
Subgroup B 42°C	LOW	BL	HIGH
G’ max [Pa]	193±36	276±22	377±73
G´ init [Pa]	57±22	69±46	76±19
G´ init -t [s]	318±165	507±129	447±48

3.3 Group B:

Addition of fibrinogen increased initial and final clot stiffness. The gain in fiber network density by the addition of fibrinogen that explains these higher moduli, is shown in Fig. 3. Addition of fibrinogen also shortens the time to achieve the G´-plateau and accelerates clot formation (Fig. 1). In contrast to group A, normal forces (F_N) in the gap were maintained, although shear moduli were significantly modified by the fibrinogen concentration. Normal force in the gap was triphasic. In the initial period, it increased slightly, only to fall rapidly afterwards, also forming a plateau value. Minimum achieved F_N was –51±6 mN (BL), –57±14 mN (MED), and –59±17 mN (HIGH). Fibrinogen concentrations correlated with ROTEM-MCF: r = 0.77, but G’ max was detected later than MCF. The G’ max-t versus MCF-t mean difference was 164 s. All data < 0.01. Data are provided in Table 1.

By addition of fibrinogen, the concentration of the protein in native samples (254±45 mg/dL; BL) was raised to 626±174 (MED) and 812±282 mg/dL (HIGH). Erythrocyte and leucocyte counts were unaffected, however, platelet count was reduced (from 282±84 to 62±50 G/L). Coagulation markers were unaffected by fibrinogen addition. PT: 97±18 (native), 94±18 (MED), 83±28 % (HIGH); aPTT: 37±3 (native), 39±4 (MED), 41±3 s (HIGH).

3.4 Subgroup B

Increasing the temperature accelerated the start of clot formation (see Fig. 2). A higher temperature was associated with clot weakening. Similar to group B, moduli increased with the fibrinogen concentration, and this occurred at each temperature. All data p < 0.01. Data are provided in Table 2.

Fig. 2

Temperature accelerates clot formation in HCT-adjusted (a) and fibrinogen-adjusted (b) samples. The grey-scale of the data points represent the different HCT values and fibrinogen concentrations. Lines indicate the geometric mean±95% CI.

Fig. 3

Fibrin networks differ in whole blood clots when fibrinogen is added (representative SEM pictures of WB clots from the same subject). Fibrinogen concentrations are 2.71 g L^–1 (a) and 5.78 g L^–1 (b). In the center of both pictures are platelet or platelet aggregates with fibers radiating out of them. The clot with higher fibrinogen concentration exhibits network densification.

Frequency sweeps in linear test mode revealed that, similar to group A, both shear moduli were functions of frequency with very low slopes. But in contrast to group A, the G´-slopes did not differ in subgroup B upon addition of fibrinogen (at 37°C: LOW: 0.016±0.005; BL: 0.017±0.003; HIGH: 0.028±0.007 Pa/s^–1). Again, G˝ showed a minimum between 0.5 and 1 Hz, indicating optimized clot stability in this range.

3.5 Influence of the surface profile in shear moduli

Changing from a smooth, polished plate-plate system to a system with a profiled surface significantly increased G´ max (p < 0.000) and resulted in its correlation with ROTEM MCF (r = 0.45; p < 0.05), which was otherwise not obtained by using the polished plates. Table 3 shows these data. Leucocyte and RBC counts were not relevantly affected by adding fibrinogen. Platelet count was 149±72 G/L in BL samples, and fell marginally to 117±68 G/L in MED samples, and to 119±66 G/L in HIGH samples. Plasma fibrinogen concentrations in these study groups were 232±51 mg/dL (BL), 713±221 mg/dL (MED), and 1010±97 mg/dL (HIGH).

Table 3
Influence of the profile of the measuring system. Storage modulus increased by a factor 10 when a profiled measuring surface was used. But the use of a profiled surface did not affect the development of F_N in the gap. ROTEM data are shown as MCF. Data are given as mean±standard deviation

BL MED HIGH

G´ max polished [Pa] 20.6±4.2 25.9±10.1 26.7±5.1

G´ max profiled [Pa] 223.5±60 288.12±78.8 327.9±58.9

F_N min polished [mN] 90±10 80±20 70±20

F_N min profiled [mN] 80±10 80±20 70±20

MCF [mm] 51.5±5.1 66.0±5.5 73.6±5.1

	BL	MED	HIGH
G´ max polished [Pa]	20.6±4.2	25.9±10.1	26.7±5.1
G´ max profiled [Pa]	223.5±60	288.12±78.8	327.9±58.9
F_N min polished [mN]	90±10	80±20	70±20
F_N min profiled [mN]	80±10	80±20	70±20
MCF [mm]	51.5±5.1	66.0±5.5	73.6±5.1

3.6 Electrical conductivity during clotting

The sample resistance R_p increased with clot formation. As shown in Table 4, the initial resistance increased with the fibrinogen concentration. This effect was also observed by the maximum of the loss tangent D (relaxation peak), which is inversely proportional to R_p (Fig. 4). Fibrinogen concentration did not influence the relaxation frequency f_r of 249 kHz. However, in plasma containing different platelet counts, the relaxation peak not only decreased with higher platelet count, but also shifted to lower frequencies (Table 5, Fig. 4). This change of relaxation frequency indicates a change of the electric dipole moment of the bulk sample.

Table 4
Influence of the fibrinogen concentration (BL, MED, HIGH) on the electric resistance R_p at 100 kHz during clotting

ΔR_p after 15 min. in %

BL 9.79

MED 11.63

HIGH 22.06

	ΔR_p after 15 min. in %
BL	9.79
MED	11.63
HIGH	22.06

Table 5

Influence of platelet count on the relaxation peak D_max and the relaxation frequency f_r

	D_max (–)	f_r (kHz)
Native whole blood	3.36	200
PRP	2.27	245
PRP:PDP = 1:4	2.84	270
PRP:PDP = 1:64	3.47	330

Fig. 4

(a) Dielectric spectroscopy of whole blood clots with addition of fibrinogen shows a unique loss tangent peak at 249 kHz. (b) Dielectric spectroscopy of plasma clots shows loss tangent peaks between 245 kHz and 330 kHz. The relaxation frequency shifts to lower frequencies at higher platelet counts (PRP = 218.000 platelets per mm³). In both conditions, D_max is reduced when the network becomes dense.

4 Discussion

We utilised oscillating shear rheometry and thrombelastometry to investigate the kinetics of clotting by changing blood components and environment. Additionally, we looked into the clot “signature” at its resting state by viscoelastic and dielectric tests. This information is not provided by other viscoelastic haemostatic assays.

Adding RBCs to whole blood significantly reduced clot firmness at quasi-static flow conditions. It also made the clot more susceptible to deformations applied at different frequencies. This underlines the finding of Riha et al. [27], and does not support the argument that RBCs strengthen clots by adding their membrane modulus to the overall elasticity, although RBCs are incorporated by ICAM-4 interacting with the platelet GPIIb/IIIa receptor [28]. In-vivo, hematocrit affects coagulability by several mechanisms. High hematocrit favours platelet margination and increases thrombotic risk by increasing blood viscosity [29]. Erythrocytes can induce platelet aggregation by ADP release [29], stimulate the production of cyclooxygenase- and lipoxygenase-metabolites [30], activate plasmatic prothrombinase through surface phosphatidyl-serine exposure, and activate factor IX through RBC membrane elastase [31]. Due to these prothrombotic effects it was recommended to transfuse erythrocyte concentrates at HCT values below 30% [32]. However, in-vitro tests usually show a reduction of clot stiffness when the HCT of the sample was elevated [33, 34]. This is obviously due to the fact that the volume of the plasma phase is always diminished at higher hematocrit in a sample. If the fibrinogen concentration is maintained (which was our approach), there is less fibrinogen available. Fewer fibers can be formed, which automatically lowers the modulus and explains the findings. But the most intriguing observation was that temperature had a biphasic influence on clot stiffness in our series of HCT-modified samples. A temperature increase augmented clot firmness when few RBCs were present (at 25% HCT), whereas, as soon as the HCT became physiologic or polyglobulic, acceleration of the clotting process weakened the material. In comparison, when the fibrinogen concentration was modified, acceleration of clotting via a temperature increase always reduced clot stiffness, whatever the initial fibrinogen concentration. This indicates that sufficient time must be provided to the system to ensure a fiber network with “good” stiffness and supports the argument that the fibrin fiber size development is kinetically driven [35]. The relevance of a few RBCs added to the clotting system that eventually might guide the process towards higher stiffness warrants future studies.

In fibrinogen-spiked samples we found the classical profile of hypercoagulability: shortened onset and faster rate of clotting, and higher plateau clot stiffness. This is in good agreement with previous studies involving ROTEM [36 –38] and free oscillating rheometry (FOR) [39], which demonstrated increased clot strength after fibrinogen supplementation in the context of hemodilution or thrombocytopenia. The fiber network of the clots obtained by SEM fully supports the gain of stiffness. Our clots showed network densification and thicker fibers by addition of fibrinogen. Increased fiber thickness is typically not associated with a higher fibrinogen concentration in pure fibrin gels [40]. The thicker fibers in our study must therefore result from the interaction of fibrin with other plasma components. The effect of fibronectin, FXIIIa, thrombospondin, and immunoglobulin has been shown in previous investigations [41 –44], but among all plasma proteins, albumin is present at the highest concentration and will also contribute to clot architecture. Addition of albumin to a clotting system can make the fibrin gels more clustered or even granulated [45, 46], a matter that we avoided by several steps of sample preparation prior to SEM. Although striking evidence of direct binding of albumin to fibrin has been seen [47, 48], some coating of fibers via unspecific binding of albumin might have occurred (compare with Fig. 5). This could increase the fiber size. But apart from the gain of fiber thickness as a result of fibrinogen addition in our systems, the high network density of the blood clots associated with doubling of the monomer concentration can be clearly seen in Fig. 3.

Fig. 5

SEM of a fibrin network generated in blood plasma and taken from the rheometer gap. Deposits on the fibers are clearly visible.

Analysis of the negative normal force development during clotting reveals a three-phase response, and the F_N-range was comparable, although slightly higher, than other studies investigating bulk clot forces [49]. Although we observed an inverse relationship between F_N and the shear modulus in HCT-adapted samples, we did not observe a significant relationship between F_N-range and fibrinogen (respectively) thrombocyte concentrations. This was surprising, as several other studies demonstrated those parameters to influence clot contractile force [50]. Intrigued by this finding, we also investigated platelet-rich plasma (PRP) and platelet-depleted plasma samples (PDP) from three healthy individuals and found that even though the F_N-range is higher in PRP, also in PDP samples a F_N-response is detected (> –200 mN after 1500 s). Our findings could thus indicate that bulk normal force only partly depends on thrombocyte contractile force, but on other factors as well as, which are similar to the intrinsic negative normal force occurring in biopolymer gels [51]. Terms like “platelet contractile force” are therefore imprecise terms when referring to bulk forces of blood clots.

An interesting finding was obtained by looking at the dielectric property of the clots in association with their network density. In both systems (whole blood and blood plasma), the maximum of the loss tangent D decreased when the networks became denser. A reduction of D indicates a quantitative response: fewer dipoles are contributing to the relaxation process at the given frequency spectrum. Hypothetically, they are progressively immobilized in the networks. In contrast, a shift of the relaxation frequency indicates a qualitative response, and we observed a difference between whole blood and plasma samples. Whereas the maximum relaxation frequency did not change with the fibrinogen concentration, it shifted systematically to lower frequencies when the platelet count was elevated. This shift of the dominant dipole of the bulk clot indicates a selection of molecules and ions in association with the network density. In whole blood clots such an effect might have been masked by the RBCs. However, the relaxation frequency of whole blood is much lower [52], and therefore a superposition of the RBC signal is very unlikely.

As far as we know, no other studies directly compared oscillatory shear rheometry with ROTEM, although several studies compared viscoelastic hemostatic assays with free oscillation rheometry, a related technique also providing viscoelastic moduli [53 –55]. Although the results are not directly comparable due to the use of different activators or inhibitor substances in these studies, they showed good correlation between ROTEM and free oscillation rheometry for clotting time and maximum clot stiffness parameters. This is in good agreement to our results. In both study groups we found good correlation between rheometry and ROTEM, which makes it possible to cross-validate both methods. Interestingly, maximum clot formation is detected significantly later by rheometry when compared to ROTEM, which could be interpreted as a disadvantage of rheometry since point-of-care methods should provide data as quickly as possible to treat patients at the earliest possibility. The explanation for this finding lies in the different oscillation amplitudes of the two systems: ROTEM works at much higher amplitudes that generate a higher deformation of the clot in the gap [19]. It is known that shear stress during clot formation alters clot configuration and, as such, also the viscoelastic parameters obtained [27 , 57]. Then, when the applied strain exceeds the elastic limit of the sample, structural changes in the material occur. This is the principal of dynamic mechanical analysis. Such structural changes within the clot could include any kind of network orientation, but could also include loss of cohesion of components if the deformation is high enough. Whatever happens depends on the network itself, its geometrical parameters and the way by which the RBCs are incorporated. Our experiments show that yielding of whole blood clots occurs quite early (around 1% amplitude [24], also confirmed previously [19]). When the applied amplitude is higher than the yield point of the clot, the subsequent phase of “clot lysis” after reaching MCF is highly influenced by the interaction of the clot with the shear forces. This interaction starts as soon as the network is firm enough to yield at the applied strain. The earlier appearance of maximum clot firmness in ROTEM does therefore not reflect the clot firmness in equilibrium conditions, but reflects the maximum stiffness that can be gained before yielding occurs.

We conclude that rheometry provides a suitable tool to detect clot stiffness, since it can use equilibrium conditions during clot generation. Parameters are derived from a robust theoretical framework, which can bridge the gap between basic sciences and clinical research. More tests can be applied to look into the clot “signature”, for example simple tests such as the nonlinear behaviour or the dielectric property. This of course doesn’t override the advantage of thrombelastometry, which is the automation of the starting process by the recent instruments. But it must be noted that all in-vitro tests work with a finite sample volume, whereas in-vivo clot formation gains material from a large pool as long as the vessel segment in which the clot forms is perfused. This will constantly alter the conditions for clot formation and can explain differences between in-vivo and in-vitro findings. Although RBCs can alter thrombin generation [58] and interfere with the kinetic of clot formation, incorporated RBCs weaken a clot. This is due to the fact that the fiber meshwork – even if only 1% of the total mass – determines clot strength.

Footnotes

ACKNOWLEDGMENTS

The financial support by the Österreichische Austauschdienst (OeAD) (Wissenschaftlich-technische Zusammenarbeit, contract BG 12/2017) and the Bulgarian National Science Fund (project for bilateral cooperation 2016 –Bulgaria –Austria; No AHTC 01/10): “A comparative study of the kinetic of clot (thrombus) formation” is gratefully acknowledged. Part of the work was also supported by Graz University of Technology (LEAD project 2018-2021: “Mechanics, Modeling and Simulation of Aortic Dissection”).

We also appreciate the help of the Department of Laboratory Medicine of the Medical University Vienna for performing hematology and coagulation tests of our samples. We are also indebted to Dr. Loredana Voelker-Pop for helping us with the DRD-rheological device. We also greatly thank Dr. Elizabeth Crichton for her English editing of the manuscript.

References

Kozek-Langenecker

, Ahmed

, Afshari

, et al. Management of severe perioperative bleeding: guidelines from the European Society of Anaesthesiology: First update 2016. Eur J Anaesthesiol. 2017;34:332–95.

Whiting

, DiNardo

. TEG and ROTEM: Technology and clinical applications. Am J Hematol. 89:228-32.

Serraino

, Murphy

. Routine use of viscoelastic blood tests for diagnosis and treatment of coagulopathic bleeding in cardiac surgery: updated systematic review and meta-analysis. Br J Anaesthesiol. 2017;118:823–33.

Wikkelso

, Wetterslev

, Moller

, Afshari

. Thromboelastography (TEG) or rotational thromboelastometry (ROTEM) to monitor haemostatic treatment in bleeding patients: a systematic review with meta-analysis and trial sequential analysis. Anaesthesia. 2017;72:519–31.

Fahrendorff

, Oliveri

, Johansson

. The use of viscoelastic haemostatic assays in goal-directing treatment with allogeneic blood products –A systematic review and meta-analysis. Scand. J Trauma Resusc Emerg Med. 2017;25:39.

Henskens

YMC

, Gulpen

AJW

, van Oerle

, Wetzels

, Verhezen

, Spronk

, Schalla

, Crijns

, ten Cate

. Detecting clinically relevant rivaroxaban or dabigatran levels by routine coagulation tests or thromboelastography in a cohort of patients with atrial fibrillation. Thromb J. 2018;16:3.

Lipets

, Ataullakhanov

. Global assays of hemostasis in the diagnostics of hypercoagulation and evaluation of thrombosis risk. Thromb J. 2015;13:4.

Levy

, Goodnough

. How I use fibrinogen replacement therapy in acquired bleeding. Blood. 2015;125:1387–93.

Scrutton

, Ross-Murphy

, Bennett

, Stirling

, Maede

. Changes in clot deformability –a possible explanation for the epidemiologic association between plasma fibrinogen concentration and myocardial infarction. Blood Coagulation and Fibrinolysis. 1994;5:719–23.

10.

Zhu

, Kang

, Li

, Wang

, Shang

, Luo

, Xiong

, Hu

. Chronic intermittent hypoxia vs chronic continuous hypoxia: effects on vascular endothelial function and myocardial contractility. Clin Hemorheol Microcirc. 2019, doi: 10.3233/CH-190706. [Epub ahead of print]

11.

Pretorius

, Steyn

, Engelbrecht

, Swanepoel

, Oberholzer

. Differences in fibrin fiber diameters in healthy individuals and thromboembolic ischemic stroke patients. Blood Coagul. Fibrinolysis. 2011;22:696–700.

12.

Sabra

, Lawrence

, Curtis

, Hawkins

, Williams

, Evans

. In vitro clot modelt o evaluate fibrin-thrombin effects on fractal dimension of incipient blood clot. Clin Hemorheol Microcirc. 2019;27:1–7.

13.

Litvinov

, Weisel

. Role of red blood cells in haemostasis and thrombosis. ISBT Sci Ser. 2017;12:176–83.

14.

Cines

, Lebedeva

, Nagaswami

, Hayes

, Massefski

, Litvinov

, Rauova

, Lowery

, Weisel

. Clot retraction: compression of erythrocytes into tightly packed polyhedral and redistribution of platelets and fibrin. Blood. 2014;123:1596–603.

15.

Gerth

, Wesley

, Ferry

. Rheology of fibrin clots II: Linear viscoelastic behavior in shear creep. Biophysical Chemistry. 1974;2:208–17.

16.

Janmey

, Georges

, Hvidt

. Basic rheology for biologists. Methods in Cell Biology. 2007;83:3–27.

17.

van Kempen

THS

, Donders

, van de Vosse

, Peters

GWM

. A constitutive model for developing blood clots with various compositions and their nonlinear viscoelastic behavior. Biomech. Model Mechanobiol. 2016;15:279–91.

18.

Evans

, Hawkins

, Williams

. Rheometry for blood coagulation studies. Rheology Reviews. 2006;255-91.

19.

Burghardt

, Goldstick

, Leneschmidt

, Kempka

. Nonlinear viscoelasticity and the thromboelastograph: I. Studies on bovine plasma clot. Biorheology. 1995;52(6):621–30.

20.

Lamer

, Thomas

, Curtis

, Badiei

, Williams

, Hawkins

. The application of large amplitude oscillatory stress in a study of fully formed fibrin clots. Phys. Fluids. 2017;29:121606.

21.

Piechocka

, Bacabac

, Potters

, Mackintosh

, Koenderink

. Structural hierarchy governs fibrin gel mechanics. Biophys. J. 2010;98:2281–9.

22.

Kurniawan

, Vos

, Biebricher

, Wuite

GJL

, Peterman

EJG

, Koenderink

. Fibrin networks support recurring mechanical loads by adapting their structure across multiple scales. Biophys J. 2016;111:1026–34.

23.

Lotz

, diploma thesis Medical University Vienna, Austria, 2018.

24.

Dibiasi

, diploma thesis Medical University Vienna, Austria, 2017.

25.

Mezger

, Das Rheologie Handbuch. 3rd ed. Hannover: Vincentz Network; 2010.

26.

Grimnes

, Martinsen

. Bioimpedance and Bioelectricity Basics. 3rd ed. London, Oxford: Academic Press; 2014.

27.

Riha

, Wang

, Liao

, Stoltz

. Elasticity and fracture strain of whole blood clots. Clin Hemorheol Microcirc. 1999;21:45–9.

28.

Hermand

, Gane

, Huet

, Jallu

, Kaplan

, Sonneborn

, Cartron

, Bailly

. Red cell ICAM-4 is a novel ligand for platelet-activated alpha IIbbeta 3 integrin. J Biol Chem. 2003;278:4892–8.

29.

Leone

, Sica

, Chiusolo

, Teofili

, De Stefano

. Blood cells diseases and thrombosis. Haematologica. 2001;86:1236–44.

30.

Eugster

, Reinhart

. The influence of the haematocrit on primary haemostasis in vitro. Thromb Haemost. 2005;94:1213–8.

31.

Spahn

, Rossaint

. Coagulopathy and blood component transfusion in trauma. Br J Anaesth. 2005;95:130–9.

32.

Valeri

, Crowley

, Loscalzo

. The red cell transfusion trigger: has a sin of commission now become a sin of omission? Transfusion. 1998;38:602–10.

33.

Iselin

, Willimann

. Isolated reduction of haematocrit does not compromise in vitro blood coagulation. Br J Anaesth. 2001;87:246–9.

34.

Roeloffzen

, Kluin-Nelemans

, Bosman

, DeWolf

JTM

. Effects of red blood cells on hemostasis. Transfusion. 2010;50:1536–44.

35.

Hantgan

, Hermans

. Assembly of fibrin: a light scattering study. J Biol Chem. 1979;25:11272–81.

36.

Lang

, Johanning

, Metzler

, Piepenbrock

, Solomon

, Rahe-Meyer

, Tanaka

. The effects of fibrinogen levels on thromboelastometric variables in the presence of thrombocytopenia. Anesth Analg. 2009;108:751–8.

37.

Martini

, Cabrales

, Fries

, Intaglietta

, Tsai

. Effects of fibrinogen concentrate after shock/resuscitation. Crit Care Med. 2013;41:e301–e308.

38.

Schenk

, Lindner

, Treichl

, Bachler

, Hermann

, Larsen

, Fenger-Eriksen

, Wally

, Tauber

, Velik-Salchner

, Fries

. Fibrinogen supplementation ex vivo increases clot firmness comparable to platelet transfusion in thrombocytopenia. Br J Anaesth. 2016;117:576–82.

39.

Tynngård

, Lindahl

, Ramström

, Berlin

. Effects of different blood components on clot retraction analysed by measuring elasticity with a free oscillating rheometer. Platelets. 2006;17:545–54.

40.

Ryan

, Mockros

, Weisel

, Lorand

. Structural origins of fibrin clot rheology. Biophys J. 1999;77:2813–26.

41.

Mosesson

, Amrani

. The structure and biologic activities of plasma fibronectin. Blood. 1980;56:145–58.

42.

Sakata

, Mimuro

, Aoki

. Differential binding of plasminogen to crosslinked and noncrosslinked fibrins: its significance in hemostatic defect in factor XIII deficiency. Blood. 1984;63:1393–401.

43.

Bale

, Mosher

. Effects of thrombospondin on fibrin polymerization and structure. J Biol Chem. 1986;261:862–7.

44.

Gabriel

, Smith

, Folds

, Davis

, Cancelosi

. The influence of immunoglobulin (IgG) on the assembly of fibrin gels. J Lab Clin Med. 1983;101:545–53.

45.

Marx

, Harari

. Albumin indirectly modulates fibrin and protofibrin ultrastructure. Biochemistry. 1989;28:8242–8.

46.

Carr

. Fibrin formed in plasma is composed of fibers more massive than those formed from purified fibrinogen. Thromb. Haemostasis. 1988;59:535–9.

47.

Carr

. Turbidity evaluation of the impact of albumin on the structure of thrombin-mediated fibrin gelation. Haemostasis. 1987;17:189–94.

48.

Torbet

. Fibrin assembly in human plasma and fibrinogen/albumin mixtures. Biochemistry. 1986;25:5309–14.

49.

, Li

, McCracken

, Shao

, Ward

, Fu

. A miniaturized hemoretractometer for blood clot retraction testing. SMALL. 2016;12:3926–34.

50.

Dempfle

C-E

, Kaelsch

, Elmas

, Suvajac

, Lücke

, Muench

, Borggrefe

. Impact of fibrinogen concentration in severely ill patients on mechanical properties of whole blood clots. Blood Coagul. Fibrinolysis. 2008;19:765–70.

51.

de Cagny

HCG

, Vos

, Vahabi

, Kurniawan

, Doi

, Koenderink

, MacIntosh

, Bonn

, Porosity governs normal stresses in polymer gels. Phys Rev Lett. 2016;117:217802.

52.

Wolf

, Gulicha

, Lunkenheimera

, Loidla

. Broadband dielectric spectroscopy on human blood. Biochim Biophys Acta. 2011;8:727–40.

53.

Solomon

, Schöchl

, Ranucci

, Schlimp

. Comparison of fibrin-based clot elasticity parameters measured by free oscillation rheometry (ReoRox^®) versus thromboelastometry (ROTEM^®). Scand J Clin Lab Invest. 2015;75:239–46.

54.

Thomas

, Larsson

, Tynngard

, Schoett

. Thromboelastometry versus free-oscillation rheometry and enoxaparin versus tinzaparin: an in-vitro study comparing two viscoelastic haemostatic tests’ dose-responses to two low molecular weight heparins at the time ofwithdrawing epidural catheters fromten patients aftermajor surgery. BMC Anesthesiology. 2015;15:170.

55.

Solbeck

, Windelov

, Baek

, Nielsen

, Ostrowski

, Johansson

. in-vitro comparison of free oscillation rheometry (ReoRox) and rotational thromboelastometry (ROTEM) in trauma patients upon hospital admission: Blood Coagul. Fibrinolysis. 2012;23:688–92.

56.

Badiei

, Sowedan

, Curtis

, Brown

, Lawrence

, Campbell

, Sabra

, Evans

, Weisel

, Chernysh

, Nagaswami

, Williams

, Hawkins

. Effects of unidirectional flow shear stresses on the formation, fractal microstructure and rigidity of incipient whole blood clots and fibrin gels. Clin Hemorheol Microcirc. 2015;60:451–64.

57.

Hochleitner

, Sutor

, Levett

, Leyser

, Schlimp

, Solomon

. Revisiting Hartert’s 1962 calculation of the physical constants of thrombelastography. Clin Appl Thromb. 2017;23:201–10.

58.

Horne

MK III

, Cullinane

, Merryman

, Hoddeson

. The effects of red blood cells on thrombin generation. Br J Haematol. 2006;133:403–8.

The effect of hematocrit,fibrinogen concentration and temperature on the kinetics of clot formation of whole blood

Abstract

BACKGROUND:

OBJECTIVE:

METHODS:

RESULTS:

CONCLUSION:

Keywords

1 Introduction

2 Materials and methods

2.1 Blood samples

2.2 Study groups

2.3 Viscoelastic tests

2.4 Influence of the surface profile on the shear moduli

2.5 Scanning electron-microscopy (SEM)

2.6 Dielectric tests

2.7 Routine laboratory tests

2.8 Data analysis and statistics

3 Results

3.1 Group A

3.4 Subgroup B

Table 4 Influence of the fibrinogen concentration (BL, MED, HIGH) on the electric resistance Rp at 100 kHz during clotting ΔRp after 15 min. in % BL 9.79 MED 11.63 HIGH 22.06

Footnotes

ACKNOWLEDGMENTS

References

Table 4
Influence of the fibrinogen concentration (BL, MED, HIGH) on the electric resistance R_p at 100 kHz during clotting

ΔR_p after 15 min. in %

BL 9.79

MED 11.63

HIGH 22.06