Anesthetists and critical care specialists use a variety of intravenous fluids to correct volume depletion in the vascular space to maintain normal tissue and organ perfusion.1,2 There are 2 main categories of fluids used: crystalloids or colloidal suspensions. Crystalloids are mainly recognized to have an adverse effect on coagulation, because of their purely extrinsic dilutional effect.3,4 In comparison, colloids in addition to any dilutional effect have an intrinsic effect on coagulation because of the physiochemical properties of the colloids.3–7 The intrinsic effects of the colloids can have a pronounced detrimental effect on coagulation when compared with crystalloids and are associated with increased blood loss and unfavorable outcomes.8 Despite this, colloids are still used routinely, because they produce a quicker and more pronounced physiologic effect in maintaining tissue perfusion.1,2
Previous studies have attempted to investigate the effects of dilution with different colloids on coagulation by using viscoelastic techniques including thromboelastography (Haemoscope Corporation, Niles, IL).3–5 These studies have shown that colloids have a detrimental and varying effect on coagulation. However, there is continued debate and conflicting evidence on which colloid has the most detrimental effect on coagulation and deciding on what colloid to use, when/if to use it, and in what quantities still remain controversial.8
Several studies have highlighted the importance of fibrinogen in correcting the effects of dilution on coagulation.9,10 Recent advances in our understanding of viscoelastic changes in coagulation have resulted in the development of new techniques for the analysis of clot characteristics, including clot microstructure.11 A previous study using a new functional biomarker of clot microstructure has identified how isolated dilution with a crystalloid changes clot microstructural organization and strength.11 This novel technique uses whole blood and measures the formation of the gel point (GP), which provides values of the elasticity of the clot (G′GP) and a numerical description of the clots microstructure (df).12,13 The df is a quantitative measure of clot microstructure, a decrease in df corresponding to more permeable, less-branched polymerized weaker fibrin structure.14
We hypothesize that dilution with colloid resuscitation fluids has a detrimental effect on clot microstructure formation as measured by df. In addition, we aim to investigate whether different colloids will have a lesser or greater detrimental effect on clot microstructure formation when compared against each other. In this study, we investigate 3 commonly used colloids to determine which has the most detrimental effect on clot microstructure in a purely in vitro model of fixed volume dilution. The 3 colloids used in this study include a naturally occurring colloid 4.5% albumin and 2 synthetic colloids: a hydroxyethyl starch based polymer suspension (Voluven 6%; Fresenius Kabi Ltd, Cheshire, UK) and a gelatin-based polymer suspension (Geloplasma; Fresenius Kabi Ltd).
The study was approved by local research ethics committee (REC 07/WMW/136). Written informed consent was obtained from all subjects before enrollment. Healthy volunteers were included if they met the following criteria: 18 years and older, no history of an acute or chronic condition known to affect coagulation (eg, cancer, hepatic, and/or renal dysfunction) and no personal or family history of bleeding or thromboembolic disorders, or have taken any medication known to effect coagulation with all participants having a normal full blood count (FBC) and clotting profile. Subjects were excluded if currently undergoing any antiplatelet or anticoagulation treatment. The 3 different fluids used in this study were a gelatin-based colloid (Geloplasma, anhydrous gelatin/15 g; Fresenius Kabi Ltd; BN: 13GIL272), a commonly used hydroxyethyl starch (Voluven 6%, starch 130/0.4; Fresenius Kabi Ltd; BN: 13EML053), and a naturally occurring colloid (4.5% human albumin, Zenlab 4.5%; Bio Products Laboratory Ltd, Herts, UK; BN: ADAN0425).
Before taking the sample, each volunteer was assigned to 1 of the 3 different colloidal fluids, and they were then assigned to 1 of the 4 dilutions tested in this study. The dilutions investigated for each fluid were 10% (n = 8), 20% (n = 8), 40% (n = 8), and 60% (n = 8). An additional 8 volunteers were recruited for a 0% dilution reference range. For each volunteer, a 20-mL sample of blood was obtained from the antecubital vein through an 18-gauge needle. The sample was added to a specified volume of their designated fluid and gently inverted 3 times to aid mixing. The diluted blood was divided into 2 aliquots. The first (7 mL) was immediately transferred to the measuring geometry of a Rheometer for viscoelastic analysis. The second aliquot (≈9 mL) was used to obtain a coagulation screen and FBC. The limitations of the study include the inability to run all 5 dilutions (0%, 10%, 20%, 40%, and 60%) for all 3 colloids for each volunteer at 1 sampling point. The GP technique uses whole blood limiting the number of experiments that can be run at any 1 time by the availability of machines.
Rheometric Measurements: The Gel Point
A 6.6-mL aliquot of diluted blood was immediately loaded into a TA Instruments AR-G2 controlled-stress rheometer (TA Instruments, New Castle, DE) at 37°C (± 0.1°C). All measurements were made on aliquots of blood using similar measuring geometries with identical measuring surfaces and surface preparation procedures. The measuring geometry used was a double-gap concentric cylinder (TA Instruments; see Figure 1). Once loaded, the experiment was started, the time from venipuncture to starting the experiment was recorded, and was always <60 seconds.
The GP measurement involves performing oscillatory deformations over a range of frequencies called small-amplitude oscillatory shear measurements. The measurement uses 4 different frequencies (2, 0.93, 0.43, and 0.2 Hz; see Figure 2) with a peak stress amplitude of 0.03 Pa. The small-amplitude oscillatory shear measurements provide assessment of the viscoelastic properties of the material. The small-amplitude oscillatory shear experiments measure the shear elastic modulus, G′ (a measure of elastic component of the sample); the loss modulus, G″ (a measure of viscous component of the sample); and phase angle, δ (where tan δ = G″/G′). δ represents the ratio of the viscous and elastic components of the blood and has a range of 0° to 90°, where 90° identifies a purely viscous response and 0° identifies a purely elastic response with all values in between being viscoelastic. We measure δ at 4 different frequencies sequentially (with time) to record how δ changes as the blood clots. In elasticoviscous fluids, the higher the frequency, the lower the value of δ that will be recorded at the same point in time; conversely, in viscoelastic solids, the higher the frequency, the higher the value of δ (see Figure 2). Therefore, by measuring multiple frequencies during a GP measurement, the transition of the blood from a elasticoviscous fluid to a viscoelastic solid can be identified at the point where δ is independent of frequency, where the material changes from a liquid to a solid.15 This point is called the GP, which in coagulating blood corresponds to the formation of the incipient clot.12 The incipient clot defines the point where fibrin network becomes sample spanning and establishes sufficient connectivity to confer elastic solid-like properties required for its hemostatic function.12
It is known that the viscoelastic properties at the GP are directly linked to the structural properties of the system, which can be expressed in terms of fractal dimension, df.12,16 Fractal dimension is a tool commonly used to quantify complex structures. The greater the value of df, the more compact/dense is the network structure, whereas low values of df correspond to more open/permeable networks. At the GP, fractal dimension, df, is related to the stress relaxation exponent θ which can be obtained from the value of δ at the GP, where δGP = θπ/2.The relationship between df and θ is described by the equation: df = (D + 2)(2θ − D)/2(θ − D), where D is the space dimension (D = 3 herein).11–14,16
Mass Versus df
In conjunction with the GP measurements, we also provide a previously published computational simulation.14,17,18 This simulation was used alongside the experimental data collected from the GP measurements to help illustrate how any changes in incipient clot microstructure (df) will relate to changes in the mass of the clot. Previous studies using light scattering and microscopy have established that incipient fibrin clots have fractal properties, where the mass, M, is related to df by the following power law equation (M ≈ εdf, where ε is some length scale value in the range 100 nm to 10 μm).18,19 This nonlinear relationship is presented in Figure 3.
Full Blood Count. A 4-mL aliquot of blood was collected into plastic, full-draw dipotassium EDTA Vacuettes (Greiner Bio-One, Stonehouse, UK; Ref: 454286) for FBC analysis and was analyzed using a Sysmex XE 2100 (Sysmex UK, Milton Keynes, UK) automated hematology analyzer within 2 hours of collection. Fibrinogen concentration was verified against the second International Fibrinogen Standard version 4 (NIBSC code 96-612). All reagents were obtained from Siemens (Frimley, UK). The analyzer was calibrated according to the manufacturer’s instructions.
Coagulation Screen. An additional 4.5 mL was transferred immediately into citrated siliconized glass Vacutainers (0.109M; Becton-Dickinson, Plymouth, UK; Ref: 367691) for routine coagulation studies. Prothrombin time (PT), activated partial thromboplastin time (APTT), and Clauss fibrinogen were measured using a Sysmex CA1500 analyzer (Sysmex, Milton Keynes, UK) within 2 hours of collection. All reagents were obtained from Siemens.
Statistical analysis was performed using Minitab version 16 software (Havertown, PA) and GraphPad Prism version 6.0 (GraphPad software Inc, La Jolla, CA). We powered this study for the main effect that each of the different fluids would produce a significant change in the df measurement. Using data from a previous study investigating crystalloid dilution, we observed a significant change in df of 0.06 from the baseline value (at 20% dilution).11 Using an expected change in df of 0.06 and assuming an α of 0.05 and a power of 0.90, we calculated that 8 subjects would be required for each studied dilution with a total of 4 different dilutions and 3 different fluids (with an additional 8 for a 0% dilution) giving a total of 104 participants.
All results included are reported as mean and standard deviation unless otherwise stated. Two-way analysis of variance with a multiple comparisons test was used to investigate when differences arose in the GP results caused by both the amount of dilution and the type of colloid. For the multicomparison test, because the number of comparisons being used is large, only values of P < .001 are considered significant. Pearson correlation was undertaken to explore associations between df and standard markers of coagulation. Exact P values are stated unless <.0001.
Rheometric Measurements: The Gel Point
The df at 0% dilution was 1.74 ± 0.033, a result commensurate with the values reported for healthy volunteers in previous studies.12,13 For all 3 fluids, an overall decrease in df is observed as dilution is increased (Figure 4A). Significant changes from the baseline (0% dilution) value were recorded at a dilution of 20% for albumin (1.587 ± 0.056, P < .0001), 40% for starch (1.561 ±0.084, P < .0001), and 60% for gelatin (1.584 ± 0.061, P < .0001). At 0% dilution, the value of G′GP was 0.0091 ± 0.0030 Pa. As the dilution was increased, the strength of the incipient clot, as measured by G′GP, decreases, which was consistent for all 3 fluids (see Figure 4B). The G′GP is significantly reduced at a dilution of 20% from the baseline value for all 3 fluids (P < .0001 for albumin and starch and P = .0009 for gelatin). df and G′GP were significantly correlated with each other (r = 0.733, P = 0.0005).
We used the 2-way analysis of variance with a multiple comparisons test to determine whether there were any differences in the GP results when comparing the 3 fluids against each other. Although dilution did have a significant effect (P < .0001) on G′GP, no significant difference was observed among the 3 fluids (P = .12). For the df results, we observe a significant difference not only because of dilution (P < .0001) but also because of the type of fluid (P < .0001).
Blood was obtained from (n = 104) healthy volunteers. The results of the standard and specific laboratory markers for the 3 fluids are shown in the Table. Hematocrit, fibrinogen, platelet count, and thrombin generation all show a progressive linear decrease as dilution is increased. Fibrinogen concentration and hematocrit show significant (P < .001) reductions from the baseline values occurring at a 20% dilution, or greater, for all 3 fluids. PT and APTT show a gradual increase in clotting time as dilution is increased for all 3 fluids. The PT and APTT were significantly (P < .001) prolonged at a 40% dilution or greater for all 3 fluids. When comparing the 3 fluids, we found that there was no difference between them in terms of their standard laboratory markers.
We found significant correlations between df and the standard laboratory markers: fibrinogen (r = 0.588, P = .0005), hematocrit (r = 0.652, P = .0005), platelet count (r = 0.457, P = .0005), PT (r = −0.501, P = .0005), and APTT (r = −458, P = .0005), likely as a result of the dilution effect on all parameters.
In this in vitro study of fixed volume dilution, we quantify, using a novel viscoelastic technique, how the physiochemical properties of the 3 colloids exert individual changes on clot microstructure in addition to the effects of dilution. We found that of the 3 colloids tested, gelatin caused the least changes in df compared with albumin, which caused the largest changes (P < .0001).
In a previous study, we quantified, using an isotonic crystalloid, how progressive dilution effects clot microstructure, producing a decrease in df as dilution increases.12 In this study, we identified that 3 colloidal fluids also produce a steady decrease in df (Figure 4). However, Figure 4 illustrates that the magnitude of the changes on the df, is different depending on the type of colloid being used. We identified significant changes in df occurring at a dilution of 60% for gelatin (df = 1.584 ± 0.061, P < .0001) compared with starch that occurred at a 40% dilution (1.561 ± 0.084, P < .0001) and albumin that occurred at a 20% dilution (1.587 ± 0.056, P < .0001). Analysis of the differences in df among the 3 fluids shows that there are significant differences among the 3 fluids that are independent of dilution (P < .0001). This finding suggests that the properties of the 3 colloids themselves are altering coagulation. In addition, the response of coagulation (in terms of df) to in vitro dilution appears less impaired for gelatin than either the starch or the albumin, and the most pronounced changes occurred with albumin (see Figure 4).
The intrinsic differences among the 3 colloids, reported by the GP parameters, were not reflected in the results of the standard coagulation markers (Table), and no significant differences were observed. The standard laboratory markers seem to purely measure the effect of dilution (see results of hematocrit, fibrinogen, and platelets), which all decrease by a similar magnitude with increasing dilution. The fact that the GP results identify that the colloids do affect coagulation in different ways suggests that the conventional markers cannot identify the intrinsic differences among them. Previous studies have highlighted that traditional coagulation screening tests such as PT, APTT, platelet count, and fibrinogen levels are of limited value in acute hemorrhage and do not correlate well with bleeding outcomes, which is confirmed in this study.20 This may be because of laboratory-based assays being carried out in platelet-poor plasma, omitting the vital contribution of the cellular aspects of coagulation. This is a major limitation because the intrinsic physiochemical properties of the different colloids are known to affect coagulation by interacting with platelets and other coagulation factors.21–25 Furthermore, different types of colloids will differ not only in what they affect, but also in their severity.
The individual physiochemical properties of the colloids exert different intrinsic effects on coagulation and clot development. Many studies have shown that dilution with starch has a large detrimental effect on coagulation and that this effect can be different depending on the type of starch used.26,27 Dilution with starch has been shown to reduce platelet aggregation, cause an acquired von Willebrand syndrome, and affect fibrinolysis.25,28–30 Furthermore, in vivo studies link their use to increased bleeding and poorer outcome.31 Although gelatin is known to inhibit platelet aggregation, it has been reported that its detrimental effect on coagulation is not as severe as either starch or albumin.4,24,32
In this study of all the 3 fluids investigated, albumin produced the largest alterations in df. This is a somewhat controversial result considering it has been suggested that albumin is no worse than saline, having less effect on hemostasis than other synthetic colloids.33 However, a previous in vitro study that investigated the effect of dilution with albumin compared with starch in whole blood, similar to this study, has also shown that albumin caused the greatest derangement in clotting.7 Another study has also reported that albumin has a heparin-like effect on coagulation.34 Possible explanations for this effect and the large changes observed in df may be because of the known inhibition of platelets by albumin caused by binding to the platelet-activating factor and reducing histone-mediated aggregation.22,23 Another important consideration in the change in df is that albumin has been shown to have an inhibitory property in fibrin polymerization, resulting in smaller diameter fibrils.35,36df is a measure of clot microstructure and is directly affected by fibrin polymerization.
In this study, we also utilize computer modeling to visually represent what the changes in the values of df mean in terms of clot mass and functionality. The values of df at higher dilutions, particularly for albumin, are noteworthy being very low (<1.5). The latter value has only been previously reported for blood samples anticoagulated with heparin at levels above the upper limit of the therapeutic range.12 Although values of df as low as 1.3 have been estimated for fibrin gel systems in light-scattering studies,19 the latter were formed at markedly subphysiologic levels of fibrinogen (<0.5 g/L). The mean fibrinogen concentration at 60% dilution with albumin in this study being 0.9 ± 0.2 g/L. Fractal systems for which df is <1.5 have relatively low levels of incorporated mass (Figure 3). Figure 3 shows that a clot that forms with a df of 1.50 will contain approximately 3% to 4% the mass of a clot formed with a df of 1.74 (0% dilution). The example images of what typical fractal structures at df values of 1.74, 1.60, and 1.50 are shown in Figure 3. These images clearly represent how the changes in df and hence mass result in significantly different-looking structures. The images show that the structures with a higher df have a large amount of interconnectivity creating very strong clots. Consequently, clots with lower values of df will have a low amount of interconnectivity and as a result will be mechanically weak and highly friable. These observations are supported by the very low values of G′ associated with the higher dilutions in albumin (Figure 4). These systems may prove ineffective as microstructural templates for ensuing normal clot development, where such low levels of shear elasticity may be incompatible with hemostatic functionality with the clots bordering on being physiologically viable.
One limitation of this study is that we did not perform analysis of factor XIII and its role in dilution with the different colloids. Previous studies using other viscoelastic devices, investigating the effect of dilution, have shown that factor XIII has an effect on final clot strength, where dilution with starch-based colloids nearly always produces a larger reduction than albumin.37,38 This is in contrast to the findings of this study and in a previous study using whole blood with thrombelastography.7 There is conflicting evidence in the literature regarding the ability of the other viscoelastic analyzers to detect the changes in XIII when it is added back in vitro when comparing whole blood systems with those in plasma and dilution with colloids to albumin.37,39 Furthermore, in a previous randomized study of patients undergoing hip replacement with the use of colloids and albumin, the dilution of factor XIII in these patients was in accordance with dilution, and there was no intrinsic effect of the colloids on factor XIII even at levels of 60% dilution.40 Although this study was not designed to investigate specific mechanistic effects such as that of factor XIII and other component replacement therapies, this is potentially an important factor in clot microstructure formation, which may require further investigation.
The study is the first to show how a marker of clot microstructure, in a preclinical model of dilution, can measure and differentiate between the intrinsic effects of different intravenous colloids. Although different fluids have similar and dissimilar intrinsic effects on the coagulation system, df was able to quantify the overall effect. The results from our previous and current studies suggest that measurements of the GP may have clinically relevant implications in guiding not only the timing and amount of fluid required, but also the type of fluid and any additional blood products. The next phase of this research will include investigating how particular coagulation factors and components modify the GP in hemodilution and assess the GP in various clinical settings to determine how changes in clot quality affect physiologic parameters and outcomes.
We would like to acknowledge the outstanding support provided to us by the Abertawe Bro Morgannwg University Health Board.
Name: Matthew James Lawrence, MEng, PhD.
Contribution: This author helped design the study, conduct the study, collect the data, analyze (rheology) the data, and prepare the manuscript.
Conflict of Interest: Matthew James Lawrence declares no conflicts of interest.
Name: Nick Marsden, BSc, MBBCh, MRCS.
Contribution: This author helped recruit the subjects, collect the data, design the study, and prepare the manuscript.
Conflicts of Interest: Nick Marsden declares no conflicts of interest.
Name: Jakub Kaczynski, MBChB, MRCS, MSc.
Contribution: This author helped recruit the subjects, collect the data, design the study, and prepare the manuscript.
Conflicts of Interest: Jakub Kaczynski declares no conflicts of interest.
Name: Gareth Davies, BSc, PhD.
Contribution: This author helped collect (rheology) the data, revise the article for scientific and intellectual content, and prepare the manuscript.
Conflicts of Interest: Gareth Davies declares no conflicts of interest.
Name: Nia Davies, BSc, MSc, PhD.
Contribution: This author helped collect (laboratory) the data, revise the article for scientific and intellectual content, and prepare the manuscript.
Conflicts of Interest: Nia Davies declares no conflicts of interest.
Name: Karl Hawkins, MEng, PhD.
Contribution: This author helped revise the article for scientific and intellectual content, interpret the data, and prepare the manuscript.
Conflicts of Interest: Karl Hawkins declares no conflicts of interest.
Name: Sounder Perumal, MBBS, MCEM, FCEM.
Contribution: This author helped recruit the subjects, revise the article for scientific and intellectual content, and prepare the manuscript.
Conflicts of Interest: Sounder Perumal declares no conflicts of interest.
Name: Martin Rowan Brown, MPhys, PhD.
Contribution: This author helped design the study, analyze the data, revise the article for scientific and intellectual content (computer simulation), perform computer simulation, and prepare the manuscript.
Conflicts of Interest: Martin Rowan Brown declares no conflicts of interest.
Name: Keith Morris, BSc, PhD.
Contribution: This author was the statistician for this project and helped design the study and prepare the manuscript.
Conflicts of Interest: Keith Morris declares no conflicts of interest.
Name: Simon J. Davidson, MPhil.
Contribution: This author helped collect the data (laboratory), revise the article for scientific and intellectual content, and prepare the manuscript.
Conflicts of Interest: Simon J. Davidson declares no conflicts of interest.
Name: Phylip Rhodri Williams, FIChemE, FInstP, FRSA, FLSW.
Contribution: This author helped design the study, analyze the data, revise the article for scientific and intellectual content (rheology), interpret of the rheologic findings, and prepare the manuscript.
Conflicts of Interest: Phylip Rhodri Williams has a conflict of interest arising as an unpaid nonexecutive director of the Swansea University spin out company, Haemometrics.
Name: Phillip Adrian Evans, MD, MBBS, FRCS, FFAEM.
Contribution: This author helped design the study, analyze the data, interpret the data, and final approval of the version to be submitted.
Conflicts of Interest: Phillip Adrian Evans has a conflict of interest arising as an unpaid nonexecutive director of the Swansea University spin out company, Haemometrics.
This manuscript was handled by: Roman Sniecinski, MD.
1. Harris T, Thomas GO, Brohi K. Early fluid resuscitation in severe trauma. BMJ. 2012;345:e5752.
2. Santry HP, Alam HB. Fluid resuscitation: past, present, and the future. Shock. 2010;33:22941.
3. Roche AM, James MF, Bennett-Guerrero E, Mythen MG. A head-to-head comparison of the in vitro coagulation effects of saline-based and balanced electrolyte crystalloid and colloid intravenous fluids. Anesth Analg. 2006;102:12749.
4. Ekseth K, Abildgaard L, Vegfors M, Berg-Johnsen J, Engdahl O. The in vitro effects of crystalloids and colloids on coagulation. Anaesthesia. 2002;57:11028.
5. de Jonge E, Levi M, Berends F, van der Ende AE, ten Cate JW, Stoutenbeek CP. Impaired haemostasis by intravenous administration of a gelatin-based plasma expander in human subjects. Thromb Haemost. 1998;79:28690.
6. Mittermayr M, Streif W, Haas T, et al. Hemostatic changes after crystalloid or colloid fluid administration during major orthopedic surgery: the role of fibrinogen administration. Anesth Analg. 2007;105:90517.
7. Tobias MD, Wambold D, Pilla MA, Greer F. Differential effects of serial hemodilution with hydroxyethyl starch, albumin, and 0.9% saline on whole blood coagulation. J Clin Anesth. 1998;10:36671.
8. Brummel-Ziedins K, Whelihan MF, Ziedins EG, Mann KG. The resuscitative fluid you choose may potentiate bleeding. J Trauma. 2006;61:13508.
9. Fenger-Eriksen C, Christiansen K, Laurie J, Sørensen B, Rea C. Fibrinogen concentrate and cryoprecipitate but not fresh frozen plasma correct low fibrinogen concentrations following in vitro haemodilution. Thromb Res. 2013;131:e2103.
10. Bolliger D, Szlam F, Molinaro RJ, Rahe-Meyer N, Levy JH, Tanaka KA. Finding the optimal concentration range for fibrinogen replacement after severe haemodilution: an in vitro model. Br J Anaesth. 2009;102:7939.
11. Lawrence MJ, Kumar S, Hawkins K, et al. A new structural biomarker that quantifies and predicts changes in clot strength and quality in a model of progressive haemodilution. Thromb Res. 2014;134:48894.
12. Evans PA, Hawkins K, Morris RH, et al. Gel point and fractal microstructure of incipient blood clots are significant new markers of hemostasis for healthy and anticoagulated blood. Blood. 2010;116:33416.
13. Stanford SN, Sabra A, D’Silva L, et al. The changes in clot microstructure in patients with ischaemic stroke and the effects of therapeutic intervention: a prospective observational study. BMC Neurol. 2015;15:35.
14. Curtis DJ, Williams PR, Badiei N, et al. A study of microstructural templating in fibrin–thrombin gel networks by spectral and viscoelastic analysis. Soft Matter. 2013;9:48839.
15. Winter HH, Chambon F. Analysis of linear viscoelasticity of a cross-linking polymer at the Gel-Point. J Rheol. 1986;30:36782.
16. Muthukumar M, Winter HH. Fractal dimension of a crosslinking polymer at the gel point. Macromolecules. 1986;19:12845.
17. Curtis DJ, Brown MR, Hawkins K, et al. Rheometrical and molecular dynamics simulation studies of incipient clot formation in fibrin–thrombin gels: an activation limited aggregation approach. J Nonnewton Fluid Mech. 2011;166:92832.
18. Brown MR, Curtis DJ, Rees P, et al. Fractal discrimination of random fractal aggregates and its application in biomarker analysis for blood coagulation. Chaos Solitons Fractals. 2012;45:102532.
19. Ferri F, Greco M, Arcòvito G, De Spirito M, Rocco M. Structure of fibrin gels studied by elastic light scattering techniques: dependence of fractal dimension, gel crossover length, fiber diameter, and fiber density on monomer concentration. Phys Rev E Stat Nonlin Soft Matter Phys. 2002;66:011913.
20. Kitchens CS. To bleed or not to bleed? Is that the question for the PTT? J Thromb Haemost. 2005;3:260711.
21. Caballo C, Escolar G, Diaz-Ricart M, et al. Impact of experimental haemodilution on platelet function, thrombin generation and clot firmness: effects of different coagulation factor concentrates. Blood Transfus. 2013;11:3919.
22. Grigoriadis G, Stewart AG. Albumin inhibits platelet-activating factor (PAF)-induced responses in platelets and macrophages: implications for the biologically active form of PAF. Br J Pharmacol. 1992;107:737.
23. Lam FW, Cruz MA, Leung HC, Parikh KS, Smith CW, Rumbaut RE. Histone induced platelet aggregation is inhibited by normal albumin. Thromb Res. 2013;132:6976.
24. Evans PA, Glenn JR, Heptinstall S, Madira W. Effects of gelatin-based resuscitation fluids on platelet aggregation. Br J Anaesth. 1998;81:198202.
25. Liu FC, Liao CH, Chang YW, Liou JT, Day YJ. Hydroxyethyl starch interferes with human blood ex vivo coagulation, platelet function and sedimentation. Acta Anaesthesiol Taiwan. 2009;47:718.
26. Tynngård N, Berlin G, Samuelsson A, Berg S. Low dose of hydroxyethyl starch impairs clot formation as assessed by viscoelastic devices. Scand J Clin Lab Invest. 2014;74:34450.
27. Langeron O, Doelberg M, Ang ET, Bonnet F, Capdevila X, Coriat P. Voluven, a lower substituted novel hydroxyethyl starch (HES 130/0.4), causes fewer effects on coagulation in major orthopedic surgery than HES 200/0.5. Anesth Analg. 2001;92:85562.
28. Fenger-Eriksen C, Tønnesen E, Ingerslev J, Sørensen B. Mechanisms of hydroxyethyl starch-induced dilutional coagulopathy. J Thromb Haemost. 2009;7:1099105.
29. de Jonge E, Levi M, Büller HR, Berends F, Kesecioglu J. Decreased circulating levels of von Willebrand factor after intravenous administration of a rapidly degradable hydroxyethyl starch (HES 200/0.5/6) in healthy human subjects. Intensive Care Med. 2001;27:18259.
30. Neilsen VG. Hydroxethyl starch enhances fibrinolysis in human plasma by diminishing alpha2-antiplasmin-plasmin interactions. Blood Coagul Fibrinolysis. 2007;18 64758.
31. Treib J, Haass A, Pindur G. Coagulation disorders caused by hydroxyethyl starch. Thromb Haemost. 1997;78:97483.
32. Schramko A, Suojaranta-Ylinen R, Kuitunen A, Raivio P, Kukkonen S, Niemi T. Hydroxyethylstarch and gelatin solutions impair blood coagulation after cardiac surgery: a prospective randomized trial. Br J Anaesth. 2010;104:6917.
33. Winstedt D, Hanna J, Schött U. Albumin-induced coagulopathy is less severe and more effectively reversed with fibrinogen concentrate than is synthetic colloid-induced coagulopathy. Scand J Clin Lab Invest. 2013;73:1619.
34. Jøorgensen KA, Stoffersen E. Heparin like activity of albumin. Thromb Res. 1979;16:56974.
35. Galanakis DK, Lane BP, Simon SR. Albumin modulates lateral assembly of fibrin polymers: evidence of enhanced fine fibril formation and of unique synergism with fibrinogen. Biochemistry. 1987;26:2389400.
36. Galanakis DK. Anticoagulant albumin fragments that bind to fibrinogen/fibrin: possible implications. Semin Thromb Hemost. 1992;18:4452.
37. Nielsen VG. Colloids decrease clot propagation and strength: role of factor XIII-fibrin polymer and thrombin–fibrinogen interactions. Acta Anaesthesiol Scand. 2005;49:116371.
38. Nielsen VG. Effects of PentaLyte and Voluven hemodilution on plasma coagulation kinetics in the rabbit: role of thrombin–fibrinogen and factor XIII–fibrin polymer interactions. Acta Anaesthesiol Scand. 2005;49:126371.
39. Hanna J, Winstedt D, Schött U. Fibrinogen and FXIII dose response effects on albumin-induced coagulopathy. Scand J Clin Lab Invest. 2013;73:55362.
40. Evans PA, Heptinstall S, Crowhurst EC, et al. Prospective double-blind randomized study of the effects of four intravenous fluids on platelet function and hemostasis in elective hip surgery. J Thromb Haemost. 2003;1:21408.