In Normal Controls, Both Age and Gender Affect Coagulability as Measured by Thrombelastography : Anesthesia & Analgesia

Secondary Logo

Journal Logo

Cardiovascular Anesthesiology: Research Reports

In Normal Controls, Both Age and Gender Affect Coagulability as Measured by Thrombelastography

Roeloffzen, Wilfried W. H. MD*; Kluin-Nelemans, Hanneke C. PhD*; Mulder, Andre B. PhD; Veeger, Nic J. G. M. PhD; Bosman, Lotte MD*; de Wolf, Joost Th. M. PhD*

Editor(s): Hogue, Charles W. Jr.; London, Martin J.; Levy, Jerrold H.

Author Information
Anesthesia & Analgesia 110(4):p 987-994, April 2010. | DOI: 10.1213/ANE.0b013e3181d31e91
  • Free

Abstract

Thrombelastography (TEG®) is classically used in situations in which point-of-care testing of hemostasis is desired. There is also increasing clinical interest not only in assessment of the prothrombotic tendency by TEG®, for example, in thrombophilia screening, but also in prediction of arterial or venous thrombosis in the general population.1 An advantage of TEG® over conventional tests of hemostasis is that it is performed on whole blood, considering the role of interacting blood elements such as phospholipid-bearing cells and platelets.2,3 The technique offers a rapid overview of the cumulative effect of all the individual components of hemostasis, without the need to analyze each of these components separately. Furthermore, TEG® provides information about the quality of the clot and the dynamics of its formation and its lysis. The different parts of the TEG® tracing correspond to specific deficiencies in coagulation factors or inhibitors of coagulation, use of anticoagulants, platelet count, and platelet function and fibrinogen level. Depending on the shape of the TEG® tracing, the hemostatic condition of a patient can be defined as normal, hypocoagulable, or hypercoagulable. TEG® is considered more sensitive than routine assays in detecting hypercoagulability.47 However, the interpretation of TEG® data is hampered by the lack of a validated large series of reference values, especially from elderly persons and from subjects with a physiologic hypercoagulable state, such as users of oral contraceptives (OCs) and pregnant women.8,9 The purpose of this study was both to establish reference ranges and to assess whether age, gender, and use of OCs influence hemostasis as measured by TEG® in fresh native whole blood in healthy persons. Although TEG® was originally designed as a bedside monitor for native whole blood, recalcified citrated blood is used as an alternative to perform tests within the laboratory. Therefore, we also performed TEG® in recalcified citrated native whole blood to explore whether both TEG® techniques show comparable results and are thus exchangeable.

METHODS

The IRB approved the study, and informed consent was obtained from all study participants. Sixty healthy male and 60 healthy female subjects, age distributed equally between 19 and 87 years, participated in the study. The following exclusion criteria were used: coagulation disorders, use of anticoagulants, use of OCs (except for the substudy, see below), use of acetylsalicylic acid within the past 10 days, use of nonsteroidal antiinflammatory drugs within the last 24 hours, renal diseases or plasma concentration of creatinine >120 μmol/L, and liver disease or increased plasma concentration of aspartate aminotransferase (>50 U/L) or alanine aminotransferase (>50 U/L). A history of thromboembolism was permitted. The effects of OCs on coagulation were examined by comparing TEG® variables of 29 healthy young female OC users with age-matched controls.

Blood Sampling and Assays

Blood samples were obtained simultaneously for TEG® analysis and standard laboratory and coagulation tests (i.e., complete blood count, white blood cell differentiation, creatinine, aspartate aminotransferase, alanine aminotransferase, prothrombin time [PT], activated partial thromboplastin time [aPTT], fibrinogen concentration, and antithrombin [AT]). Venous blood samples were collected by vein puncture at the antecubital fossa, using a 19-gauge butterfly needle. One experienced examiner obtained all blood samples. To minimize the effects of venous endothelial damage by using a tourniquet, the first aspirate of 10 mL blood was discarded. Blood was collected into a 20-mL polypropylene syringe to prevent contact activation of clotting by glass. Some of the collected blood, 3.5 mL, was filled into 2 Vacutainers (Greiner Bio-One, Kremsmünster, Austria) containing 0.5 mL coagulation sodium citrate 3.2% for subsequent standard coagulation tests and TEG® analysis (recalcified citrated native whole blood; C). All standard coagulation tests were performed on the STA-R coagulation analyzer (Roche, Diagnostica Stago, Asnières, France): PT with Thromborel S reagents and aPTT with Actin FS reagents (Dade Behring, Marburg, Germany), fibrinogen with excess thrombin (BioPool, Umea, Sweden) according to the Clauss method, and AT with thrombin as enzyme (STAchrome ATIII, Roche kit). Normal values for these variables in our laboratory are PT 11 to 16 seconds, aPTT 26 to 36 seconds, fibrinogen 1.7 to 3.5 g/L, and AT 75% to 125%.

Thrombelastographic assays were performed using a computerized TEG® coagulation analyzer (Model 5000, Haemoscope Corp., Niles, IL). All analyses were performed with TEG® disposable cups and pins as devised by the manufacturer. Polypropylene and polyethylene pipettes were used to handle reagents and blood. TEG® analyses were performed in native whole blood and recalcified citrated native whole blood. For TEG® analysis in native blood, 360 μL whole blood was pipetted into the prewarmed TEG® cup, and measurements were performed within 6 minutes of sampling.10 Recalcification and TEG® measurements in citrated blood were performed after storage at room temperature for 1 hour, as described previously.11 Twenty microliters of 200-mM calcium chloride was pipetted into the prewarmed TEG® cup. The citrated blood was gently inverted to ensure mixing of the sample. Next, 340 μL citrated blood was added to the TEG® cup.

The following TEG® variables were recorded: the reaction time (R time, minutes), representing the rate of initial fibrin formation; the clotting time (K time, minutes), representing the time until a fixed level of clot firmness is reached; the angle (α, degrees), which is closely related to K time and represents the rate of clot growth; the maximal amplitude (MA, mm), which is a measurement of maximal strength or stiffness of the developed clot; and the shear elastic modulus strength (SEMS or G, dynes/cm2), which is a parametric measure of clot firmness expressed in metric units calculated from MA as follows: G = (5000 × MA)/(100 − MA). Furthermore, we calculated the coagulation index (CI), which is an overall measurement of coagulation, using the following equation: CI = (−0.1227 R + 0.0092 K) + (0.1655 MA − 0.0241 α) − 5.022.

R time, K time, and α are prolonged by anticoagulants and factor deficiencies, but they can also be affected to a degree by platelet dysfunction or thrombocytopenia. MA is especially influenced by platelet count and platelet function as well as fibrinogen level.

In addition to “classical” TEG® variables, we made velocity calculations, describing thrombus generation (TG) during blood clotting, from the signature graph produced by TEG®. These variables give additional information on the kinetics of the coagulation cascade because they represent the more parametric measurements of clot propagation.12 The following TG variables were recorded: the maximal TG (MTG, dynes/cm2/s); this variable presents the first derivate of the velocity of the increase in clot strength, beginning as G begins to increase and ending after clot strength stabilizes. The information from this variable is equivalent to the information from the α angle; however, MTG provides a more parametric evaluation than the determination of α. The second variable is time to maximal rate of TG (TMG, seconds), which is the time it takes to reach MTG. Finally, we determined total TG (TTG, dynes/cm2), which is the total positive area under the velocity curve, representing the total change in elastic resistance until clot strength stabilization occurs. MTG and TTG are expressed using metric units of elastic resistance that accurately describe changes in clot strength.

Statistical Analysis

TEG® values are presented as mean and SD. All normal values of TEG® variables were calculated by the mean value ± 2 × SD. Group comparisons were made by the Student t test if normally distributed or Mann-Whitney U test if the distribution was skewed. Where appropriate, age was treated as a dichotomous variable, with a cutoff value of 50 years. Linear regression was used to quantify the associations of TEG® variables with age, sex, and relevant blood test. In these analyses, Pearson correlation coefficients (r2) were calculated, as well as the level of significance (with a null-hypothesis r = 0). Multivariable regression analyses were performed to obtain adjusted correlation coefficients of TEG® variables with age, sex, and blood tests. All variables univariately associated with the TEG® variable of interest at a P level of <0.10 were included in the multivariable regression analysis. Colinearity among independent variables was assessed before the multivariable analyses. To evaluate the level of agreement of TEG® values in noncitrated whole blood versus citrated whole blood, the method of Bland and Altman13 was used. The Bland-Altman plot is a tool for the presentation of method comparison between studies. The range in limits of agreements determines how much the new method (citrated native whole blood) differs from the old one (native whole blood). All individually reported P values must be interpreted within the concept of an explorative testing, rather than formal hypothesis testing. Therefore, a correction for multiple testing was not applied. Analyses were performed using commercially available computer software (Statistical Analysis System, version 8.0, SAS Institute, Cary, NC, and SPSS 17.0 software for Windows, SPSS, Chicago, IL).

RESULTS

We studied 120 healthy adults, 10 per age decade, mean age 50 ± 17 years, 60 women and 60 men. Another 29 healthy female controls using OCs, of which >60% used a second-generation OC, were analyzed. In Table 1, the demographics of all 3 subgroups are shown, with a comparison of male with female (non–OC users) and a comparison of OC users with age-matched non–OC users. Apart from hemoglobin and creatinine levels, there were no statistically significant differences between the male and female control group.

T1-3
Table 1:
Healthy Control Demographics

Effect of Gender on TEG® Variables

Mean values ± SD of TEG® variables, measured in native and citrated whole blood, and classical coagulation tests in the 3 subgroups are presented in Table 2. There were statistically significant differences in coagulability between male and female subjects. Except for R time measured in citrated blood (P = 0.06), all other TEG® variables in the female group were statistically significant in hypercoagulability, when compared with the male group, whereas no statistically significant differences were found between the classical coagulation tests in both sexes. In women, we observed a significantly faster and higher initial rate of fibrin formation as illustrated by a shorter R and K time. The female group also had a significantly higher rate of clot growth, represented by a wider α angle. Moreover, MA, SEMS, and CI were significantly higher in the female group. The latter suggests that both the maximal clot strength and the viscoelastic property of the formed clot were higher in women compared with men. Also, the novel TG variables were different in men compared with women. According to these variables, thrombin was generated significantly quicker (TMG) and with greater velocity (MTG) in women. Finally, the value for TTG was also significantly higher in women.

T2-3
Table 2:
Thrombelastography Variables and Classical Coagulation Tests in Male and Female Subgroups, in Native and Citrated Whole Blood

Effect of OCs on TEG® Variables

We compared TEG® variables of the group using OCs with an age-matched subgroup of women who did not take OCs (n = 22). In our study, women using OCs had significantly higher values for MA, SEMS, and TTG generated. Apart from a higher fibrinogen level in women using OCs, no other statistically significant differences were found in classical coagulation tests between groups (Table 2).

Effect of Age on TEG® Variables

Weak to moderate correlations between age and most TEG® variables were observed in both sexes. The strongest correlations emerged between age and MA (r = 0.47), SEMS (r = 0.47), and TTG (r = 0.47). The scatter diagrams and regression equations of these 3 TEG® variables versus age are presented in Figure 1, a–c. Observing the different TEG® variables per decade, a gradual change toward hypercoagulability with increasing age can be seen. Significant differences in TEG® variables were found in both sexes when subjects younger than 50 years were compared with subjects older than 50 years (Table 3). In the classical coagulation tests, the older male group also had statistically significant shorter aPTT (P = 0.03), higher fibrinogen levels (P = 0.003), and lower AT percentages (P < 0.001) compared with the younger male group. Women older than 50 years had significantly higher fibrinogen levels (P < 0.001) and demonstrated a trend toward a shorter aPTT (P = 0.055) compared with the younger female group.

F1-3
Figure 1:
Correlation and linear regression of age with maximal clot strength (maximal amplitude), maximal clot elasticity (shear elastic modulus strength), and the total amount of thrombus that is generated (total thrombus generation) and of hemoglobin with the time to reach a fixed level of clot firmness (clotting time); R Sq linear = correlation coefficient.
T3-3
Table 3:
Thrombelastography Variables and Classical Coagulation Tests for Both Sexes, Comparing Subjects Younger than 50 Years with Subjects Older than 50 Years

Association of Other Variables with TEG® Variables

In addition to the relationship of TEG® variables with age and gender, we also studied the association among TEG® variables, PT, aPTT, and blood cell counts. In a univariate analysis, TEG® variables were associated with gender, age, hemoglobin level, platelet count, aPTT, and fibrinogen level (data not shown). Remarkably, no correlation was found between the PT and any of the TEG® variables. In multivariable analysis, age remained statistically correlated with the MA and with the derived variables SEMS and TTG (Table 4). Furthermore, gender was associated with all TEG® variables except TMG. Regarding classical coagulation tests, the lack of correlation between platelet count and TEG® variables is surprising but in the context of a platelet count in the normal range. However, aPTT showed a correlation with the initial fibrin formation as demonstrated by a positive correlation with R time, K time, and the time to MTG and a negative correlation with the α angle. Fibrinogen was associated with all TEG® variables, except the R time.

T4-3
Table 4:
Significant Correlation Coefficients by Multivariate Analysis Among Variables: Age, Gender, Classical Coagulation Tests, and TEG® Variables in Native Whole Blood

Finally, a positive correlation between hemoglobin level and the K time (r = 0.24) was present. With increasing hemoglobin level, a prolongation of the time to reach a fixed degree of viscoelasticity, as a result of fibrin buildup and cross-linking, was observed (Fig. 1d). Hemoglobin level had comparable effects on TEG® variables in both sexes.

TEG® Variables in Fresh Whole Blood Versus Recalcified Citrated Whole Blood

We determined limits of agreement for the differences in native and recalcified citrated whole blood for each of the TEG® variables (Table 5 and Fig. 2). Our data show a lack of correlation between the TEG® measurements in both types of blood samples because of the large limits of agreement. As shown in Figure 2, for R time and K time, the lack of agreement increased with an increase of R time and K time. For MA, this was less pronounced. The mean difference between the K time measured in native blood and recalcified citrated whole blood was 246% (Table 5).

T5-3
Table 5:
Limits of Agreement for Differences in Native and Citrated Whole Blood
F2-3
Figure 2:
Scatter plots for the time until initial fibrin formation occurs (R), the time until a fixed level of clot firmness is reached (clotting time), the rate of clot growth (α), and maximal clot strength (maximal amplitude) of mean versus difference of the fresh and recalcified citrated whole blood thrombelastography values.

DISCUSSION

In this study, we assessed reference values for TEG® variables in a large population of healthy volunteers. Our report is unique because we are the first to study a population with a well-balanced age and gender distribution. Moreover, we describe both classical and dynamic (TG) TEG® variables and compare these with classical coagulation tests. Finally, we performed TEG® measurements in both native and recalcified citrated whole blood to determine whether both techniques produce the same results and are thus interchangeable.

In our study population, we demonstrated significant gender differences in coagulation, with a more procoagulant TEG® profile in women compared with men. In women, not only a faster rate of fibrin formation was observed (clotting time 7.7 ± 2.5 minutes [women] vs 10.0 ± 2.7 minutes [men]; P < 0.001) but also a greater ultimate clot strength with better viscoelastic properties (MA 48.7 ± 6.9 mm [women] vs 44.0 ± 6.6 mm [men]). In contrast, there were no gender differences using classical coagulation tests in these same subjects. The observed gender-related differences in coagulation remained with aging and, as a consequence, could not solely be explained by the effects of female sex hormones. Neither could pregnancy nor the use of OCs be held responsible for this difference, because both conditions were excluded from this subanalysis. Lang et al.14 found a trend toward enhanced coagulation in women compared with men. However, in their female reference population, use of OCs was not an exclusion criterion, which may have affected the outcome.

Although the exact mechanism is unknown, OC use increases the risk of venous thrombosis.8 From this study, we conclude that women using OCs have higher ultimate clot strength with better viscoelastic properties compared with age-matched controls not using OCs. In contrast, Zahn et al.15 using celite as an activator found no significant TEG® changes in women using low estrogen dose OCs. Adding coagulation activators such as celite results in shortening of the initial part of the TEG® line, making rapid interpretation of the coagulation process possible. However, this same activation may be at the cost of subtle information from this initial part of the TEG® line. Therefore, we emphasize the use of nonactivated whole blood in TEG® when studying circumstances in which coagulation differences are considered to be small.

Gorton et al.16 have reported significant gender-related differences in TEG® variables, with a significant procoagulant trend from men through nonpregnant women to pregnant women. However, these results may have been biased because half of the nonpregnant women were taking the third-generation OC, and only patients with an average age of 30 years were included. In our study, we demonstrated that the effect of gender is not limited to the younger age group but is a phenomenon that persists with aging.

Aging is associated with hypercoagulability and considered an important risk factor for venous thrombosis.17 We demonstrated a tendency toward hypercoagulability in both sexes with advancing age. Although there was no statistically significant difference per decade, there was a trend that may have been relevant with a greater number of healthy study subjects. Moreover, statistically significant differences in most TEG® variables were found comparing the age group younger than 50 years with the group older than 50 years, as shown in Table 3. TEG® variables mostly influenced by age were MA, SEMS, and TTG, all linked with the building up and cross-linking of fibrin, leading to a stronger thrombus with better viscoelastic properties. For the classical coagulation tests, aging was accompanied by a shortening in aPTT in the male subgroup and an increase in fibrinogen level in both genders. Others demonstrated that in patients undergoing orthopedic surgery, aging was associated with procoagulable TEG® variables.18 However, in contrast to our study, age was not equally distributed in that study because patients older than 80 years were overrepresented. More recently, no age-related differences in kaolin-activated TEG® variables in healthy children could be identified, and no significant differences between children and adults were observed.19 A possible explanation, given by the investigators, is that the process of activation of the blood samples could have affected the degree of age-related hemostatic differences.

TEG® with recalcified citrated blood is used as an alternative to noncitrated blood in situations in which immediate TEG® determination is not practically feasible. However, the use of a citrated blood sample generates different results than those obtained when a native sample is used.11,20 Others demonstrated that coagulation analyses using blood that has been exposed to citrate and recalcified afterward do not yield reliable depictions of the natural dynamics of the blood coagulation process.21 We studied TEG® variables in noncitrated and in recalcified blood and demonstrated an important lack of agreement between the 2 techniques. As a consequence, we consider both techniques not exchangeable. Both gender- and age-related differences in TEG® variables were more prominent when measured in noncitrated than in citrated whole blood. The latter suggests that preparation of the blood sample (native versus recalcified citrated native) is important for the detection of subtle age- and gender-related hypercoagulable states.

We demonstrated by multivariate analysis a strong reversed correlation between hemoglobin level (still within the normal range) and the speed to reach a certain level of clot strength (= K or clotting time). In other words, a more rapid fibrin buildup was observed in control persons with lower hematocrit levels. The importance of erythrocytes in hemostasis has been described. In the absence of platelets, erythrocytes can contribute to thrombin generation through exposure of procoagulant phospholipids at their outer cell membrane.22 Iselin et al.23 demonstrated that a progressive isolated reduction in hematocrit from 40% to 10% resulted in an accelerated blood coagulation profile resulting in increased clot strength as measured in celite-activated citrated blood. Others found a state of relative hypercoagulability (shortened R time) immediately after a rapid 10% loss in circulating blood volume.24 Because we examined healthy subjects with normal-range hematocrit values, we can only speculate on an optimal hematocrit favoring hemostasis.

There were a number of limitations to this study. First, we did not correct for multiple testing. Although such an adjustment of P values, e.g., by Bonferroni or Hochberg step-up Bonferroni, can be applied to our data, we a priori decided not to apply such an adjustment because the purpose of all hypothesis testing was not to identify a single significant result but rather to evaluate differences between groups on all the TEG® variables as a whole. Second, although operation procedures were strictly followed in accordance with the Thrombelastograph® Operation Manual, the (native) TEG® values in general were hypocoagulable compared with what has been historically reported by several laboratories worldwide. Because TEG® outcome is influenced by many variables (blood collection site, sample stability, repeat sampling, and also modifications of the technique such as citration or adding activators), we think that for correct interpretation of TEG® results, each center should perform analyses in a standardized way, with assessment of outcomes against (own) normal ranges derived from samples handled in the same manner. Finally, in clinical practice, where TEG® is used as a rapid point-of-care test of hemostasis, coagulation activators are often added to the blood samples. Because we used only whole blood samples without activators, the study observations may not be applicable to other methods of TEG® measurement.

In summary, we present reference values in a large random population, equally distributed for age and gender. We observed significant procoagulant effects of aging, female gender, use of OCs, and low-normal hematocrit levels on TEG® variables. We also demonstrated an important lack of agreement between TEG® measurements in native and recalcified citrated blood and considered both techniques not exchangeable. Our study underlines the sensitivity of TEG® over classical coagulation tests in detecting these subtle, probably physiological, differences in hemostasis.

REFERENCES

1. Hvitfeldt PL, Christiansen K, Sorensen B, Ingerslev J. Whole blood thrombelastographic coagulation profiles using minimal tissue factor activation can display hypercoagulation in thrombosis-prone patients. Scand J Clin Lab Invest 2006; 66:329–36
2. Hoffman M, Monroe DM III. A cell-based model of hemostasis. Thromb Haemost 2001;85:958–65
3. Schenone M, Furie BC, Furie B. The blood coagulation cascade. Curr Opin Hematol 2004;11:272–7
4. Ben-Ari Z, Panagou M, Patch D, Bates S, Osman E, Pasi J, Burroughs A. Hypercoagulability in patients with primary biliary cirrhosis and primary sclerosing cholangitis evaluated by thrombelastography. J Hepatol 1997;26:554–9
5. Francis JL, Francis DA, Gunathilagan GJ. Assessment of hypercoagulability in patients with cancer using the Sonoclot Analyzer and thromboelastography. Thromb Res 1994;74:335–46
6. Schreiber MA, Differding J, Thorborg P, Mayberry JC, Mullins RJ. Hypercoagulability is most prevalent early after injury and in female patients. J Trauma 2005;58:475–80
7. Zuckerman L, Cohen E, Vagher JP, Woodward E, Caprini JA. Comparison of thrombelastography with common coagulation tests. Thromb Haemost 1981;46:752–6
8. Rosendaal FR. Thrombosis in the young: epidemiology and risk factors. A focus on venous thrombosis. Thromb Haemost 1997;78:1–6
9. Sharma SK, Philip J, Wiley J. Thromboelastographic changes in healthy parturients and postpartum women. Anesth Analg 1997;85:94–8
10. Chandler WL. The thromboelastography and the thromboelastograph technique. Semin Thromb Hemost 1995;21(suppl 4):1–6
11. Camenzind V, Bombeli T, Seifert B, Jamnicki M, Popovic D, Pasch T, Spahn DR. Citrate storage affects thrombelastograph analysis. Anesthesiology 2000;92:1242–9
12. Ellis TC, Nielsen VG, Marques MB, Kirklin JK. Thrombelastographic measures of clot propagation: a comparison of alpha with the maximum rate of thrombus generation. Blood Coagul Fibrinolysis 2007;18:45–8
13. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1:307–10
14. Lang T, Bauters A, Braun SL, Pötzsch B, von Pape KW, Kolde HJ, Lakner M. Multi-centre investigation on reference ranges for ROTEM thromboelastometry. Blood Coagul Fibrinolysis 2005;16:301–10
15. Zahn CM, Gonzalez DI Jr, Suto C, Kennedy S, Hines JF. Low-dose oral contraceptive effects on thromboelastogram criteria and relationship to hypercoagulability. Am J Obstet Gynecol 2003;189:43–7
16. Gorton HJ, Warren ER, Simpson NA, Lyons GR, Columb MO. Thromboelastography identifies sex-related differences in coagulation. Anesth Analg 2000;91:1279–81
17. White RH. The epidemiology of venous thromboembolism. Circulation 2003;107:I4–8
18. Ng KF. Changes in thrombelastograph variables associated with aging. Anesth Analg 2004;99:449–54
19. Chan KL, Summerhayes RG, Ignjatovic V, Horton SB, Monagle PT. Reference values for kaolin-activated thromboelastography in healthy children. Anesth Analg 2007;105:1610–3
20. Zambruni A, Thalheimer U, Leandro G, Perry D, Burroughs AK. Thromboelastography with citrated blood: comparability with native blood, stability of citrate storage and effect of repeated sampling. Blood Coagul Fibrinolysis 2004;15:103–7
21. Mann KG, Whelihan MF, Butenas S, Orfeo T. Citrate anticoagulation and the dynamics of thrombin generation. J Thromb Haemost 2007;5:2055–61
22. Peyrou V, Lormeau JC, Herault JP, Gaich C, Pfliegger AM, Herbert JM. Contribution of erythrocytes to thrombin generation in whole blood. Thromb Haemost 1999;81:400–6
23. Iselin BM, Willimann PF, Seifert B, Casutt M, Bombeli T, Zalunardo MP, Pasch T, Spahn DR. Isolated reduction of haematocrit does not compromise in vitro blood coagulation. Br J Anaesth 2001;87:246–9
24. Ruttmann TG, Roche AM, Gasson J, James MF. The effects of a one unit blood donation on auto-haemodilution and coagulation. Anaesth Intensive Care 2003;31:40–3
© 2010 International Anesthesia Research Society