Evaluation of the profile of thrombin generation during the process of whole blood clotting as assessed by thrombelastography


Georges E. Rivard, Division of Hematology, Sainte-Justine Hospital, Room 2617, 3175 Cote-Ste-Catherine, Montréal, QC, H3T 1C5, Canada.
Tel.: +1 514 345 4931 ext: 6717; fax: +1 514 345 4884; e-mail: rivardge@medclin.umontreal.ca


Summary.  The objective of this study was to evaluate the possibility of linking the tracing of whole blood clotting in a thrombelastograph® (TEG®) hemostasis system with the generation of thrombin assessed by thrombin/antithrombin complex (TAT). Citrated whole blood containing corn trypsin inhibitor from volunteers was clotted in the presence of CaCl2 and tissue factor. Clotting was monitored with the eight channels of a TEG® system. At different time points, the whole blood TEG® reaction cups were kept in a cold quenching solution, centrifuged, and the supernatants were kept at −80 °C until assayed for TAT by ELISA. The total thrombus generation (TTG) was calculated from the first derivative of the TEG® waveform and was compared with thrombin generation measured by TAT. The two vector values – the TAT thrombin generation data and the corresponding TEG® TTG – were analyzed using Pearson correlation coefficients (r) and linear, non-linear and natural log (ln) transformation of TAT values for least-squares goodness-of-fit curves. The best least-squares fit is an exponential curve. Linearizing using the ln of the TAT thrombin generation variable produces the same r (0.94) as of the exponential curve. The prediction equation is y = 8.0465 + 0.0005x (P ≤ 0.0001), where y is the TAT thrombin generation variable in the ln transformation and x is the TEG® TTG variable. The high magnitude of r and the high significance of the prediction equation demonstrate the high efficacy of the prediction of TAT thrombin generation by the use of TEG® TTG.


Blood coagulation is a complex physiological process, which leads to the formation of a fibrin clot through the proteolytic action of thrombin on fibrinogen. In the early days of laboratory assisted clinical practice, the competency of blood for the clot formation used to be assessed by whole blood clotting time [1]. With the advances in automated technology, clinicians have learned to use the tests performed on platelet-poor plasma with clotting [2] or chromogenic [3] end point. With better understanding of the contribution of platelets, leukocytes and erythrocytes on the process of coagulation, interest has recently been focused on the kinetics of thrombin generation in platelet-rich plasma [4] and in whole blood [5]. The evaluation of thrombin generation during blood coagulation has become the preferred clinically relevant approach to assess the global integrity of this complex process. In order to give the full picture of the physiology of this multifactorial reaction, all cellular and soluble plasma participants in this reaction have to be present in their natural whole blood environment [6]. Whole blood thrombin generation tests are best suited for this purpose but are cumbersome to use in real time clinical practice.

Thrombelastography, a technology which can assess whole blood coagulation, was first described more than 50 years ago [7]. It was designed with the intent of evaluating in real time the competency of blood for coagulation, in order to assist in the control of bleeding during surgery. Although many important observations on the physiology of blood coagulation have been made with this technology [8], it has never been widely used for the lack of reliable and easy-to-use instruments. With the advent of microcomputer-assisted equipment, this technology is gaining more and more relevance in the clinical assessment of bleeding and thrombotic conditions. This technology has been used extensively in the monitoring of hemostasis during major surgical interventions such as liver transplantations [9,10], cardiovascular procedures [11–15], trauma [16] and neurosurgery [17,18]. It has also been used in the management of obstetrical complications [19–21] and of deep vein thrombosis [22] as well as in the monitoring of antagonists of platelet GPIIb/IIIa [23,24], platelet thromboxane A2 and ADP [25], and recombinant Factor VIIa [26,27]. Its use has resulted in reduction in the use of blood products [9–12,15] during surgery, and in the rate of re-exploration in cardiac surgery [15].

The objective of the present study was to evaluate the correlation between the waveform tracing generated by clotting of whole blood in a thrombelastograph® (TEG®) Hemostasis Analyzer and the kinetic profile of thrombin generation as measured by the reference method of generation of thrombin/antithrombin (TAT) complex.

Materials and methods

Normal subjects

Two subjects, one man and one woman, were studied three times each on three different days. Two other subjects, one man and one woman, were studied once each. The TAT complex concentrations (see below) were measured in duplicate and means of each result were used in the different statistical analyses performed. Participating subjects had given informed consent, in accordance to the declaration of Helsinki.


HEPES, l-benzamidine, bovine albumin, EDTA were purchased from Sigma Chemical Co. (St Louis, MO, USA). Innovin was purchased from Dade Behring Inc. (Marburg, Germany). Corn trypsin inhibitor was purchased from Haematologic Technologies Inc. (Essex Junction, VT, USA). d-Phenylalanyl-l-prolyl-arginine chloromethylketone was obtained from Cedar Lane Laboratories Ltd. (Hornly, ON, Canada). The TAT ELISA kits were purchased from Dade Behring Inc. The TEG® Hemostasis Analyzers equipped with research software for the determination of the first derivative were obtained from Haemoscope Corporation (Niles, IL, USA).

Testing thrombus dynamics by TEG®

Thrombelastography studies were performed basically as reported by Sorensen et al. [28], with modifications to allow the measurement of thrombin generation at different time points during the clot formation assessed by TEG®. With this technology, several parameters corresponding to the rate of development of the tensile strength of the forming clot are produced from the first derivative of the waveform generated by the TEG® (Fig. 1).

Figure 1.

A representative thrombelastograph® (TEG®) tracing obtained during the clotting process of the whole blood of a normal subject. The shaded area of the tracing represents the velocity curve and the shaded vertical axis represents the velocity axis. The parameters are: MTG, maximum rate of thrombus generation (100* mm s−1); TTG, total area under the curve, measures total thrombus generation (100* mm); TMG, time to maximum rate of thrombus generation(s).

Whole blood was obtained by a two-syringe technique and anticoagulated with 3.2% buffered citrate in a 1 citrate:9 whole blood proportion. Citrate contained 1000 μg mL−1 of corn trypsin inhibitor, in order to give a final concentration of 100 μg mL−1 of whole blood. For these studies, we had to use citrate because, for practical reasons, the TEG® cups had to be filled at different moments at different time points. Tubes were capped and kept undisturbed for 30 min at 37 °C. The TEG® reaction cups kept at 37 °C were preloaded with 20 μL of CaCl2 (200 mm) and 20 μL of innovin diluted 1/100 in PBS/albumin 4%, pH 7.4. At the time the reaction was started, 320 μL of citrated whole blood was added to the cup, and the recording was initiated. Four TEG® units (eight reaction cups) were used simultaneously. At given time points from 4 to 40 min, the recording was stopped and the whole reaction cup was quickly dropped into 2640 μL (1/10 final dilution of whole blood clotting mixture) of an ice-cold quenching solution (see below). After vortexing for 10 s, the mixture was kept on melting ice until the end of the experiment (about 40 min). Specimens were centrifuged at 15 000 g for 3 min, and the supernatants were kept frozen at −80 °C until dosage of TAT complexes. The quenching solution used was prepared as follows: 50 mm EDTA, 10 mml-benzamidine in 2 mm HEPES, NaCl 150 mm, pH 7.4. Immediately before the experiment, 10 μL of 10 mmd-Phe-Pro-Arg chloromethylketone diluted in 0.01 n HCl was added to 2630 μL of the above solution. The TAT complexes were measured with the ELISA kit, as recommended by the manufacturer.

In order to correlate total thrombus generation (TTG), that is the total area under the velocity curve, with thrombin generation using the TAT reference method, we measured the TEG® TTG parameter, that is the area under the velocity curve, for each of the various incremental segments of the developing TEG® waveform and compared with the TAT measured for that same segment. These are referred as a TEG® segment and a TAT segment, respectively, and represent a single time point.

The first derivative of the TEG® segment determines the TTG of that segment, while the corresponding TAT segment determines the total thrombin generation, using the TAT test result values. For the four volunteers – two male and two female – a total of 41 segments were generated, and their corresponding TTG and TAT values were computed. Linear and non-linear regressions were performed, using the TAT variable as the dependent variable (y) and the TTG variable as the independent variable (x), to obtain the least-squares fit and the corresponding Pearson correlation r value between the TAT and TTG variables; the test of significance was based on α ≤ 0.05.


As shown in Fig. 2, fitting linear regression through the origin for the TTG variable vs. the TAT variable gives a correlation of r = 0.85 and a prediction equation of y = 6.3319x (P ≤ 0.0001). Analyzing the same data with non-linear regression gives a correlation of r = 0.94 (Fig. 2) between the TTG variable and the TAT variable at P ≤ 0.0001 and a prediction equation of y = 3122e0.0005x. Linearizing using the ln of the TAT variable produces the same r of 0.94 and a prediction equation of y = 0.0005x + 8.0465, both indicating a strong correlation and a highly significant prediction equation with the TTG variable at P ≤ 0.0001 (Fig. 3). Figure 3 shows that the majority of the data points along the regression curve at a wide range of thrombin generation as measured by the TAT reference method (ln TAT) are within the prediction intervals. Figure 4 shows a typical example of a TAT generation curve superimposed on the TEG® tracing of the same specimen. Note that the formation of a fibrin/platelet clot precedes most of thrombin generation.

Figure 2.

Linear (P ≤ 0.0001) and non-linear (P ≤ 0.0001) regressions of the total thrombus generation, the TTG variable, vs. thrombin/antithrombin generation, the TAT variable.

Figure 3.

Linear regression of total thrombus generation, the TTG variable, vs. the natural log of thrombin/antithrombin generation, the ln TAT variable. The majority of the data points are within the prediction intervals (P ≤ 0.0001) of the regression line.

Figure 4.

A typical example of the superimposition of the thrombin/antithrombin (TAT) generation curve (—–bsl00001—– right vertical axis) over the thrombelastograph® (TEG®) tracing (—— left vertical axis) of whole blood from one of the subject in the study.


The quantitative measurement of thrombin generation has recently become a highly valued tool to assess the risk of bleeding or of thrombosis [6]. The use of a test based on whole blood seems to be desirable, in order to mimic the in vivo conditions as much as possible.

Thrombelastography has been used extensively over the last five decades to assess the kinetics of blood coagulation in whole blood, but no data are available, which correlate the waveform tracing of blood coagulation generated by this technology and the kinetics of thrombin generation. Several approaches have been used to initiate coagulation in the TEG® system. We have elected to initiate coagulation with diluted tissue factor in order to create in vitro conditions as comparable as possible to in vivo environment. To this end, we have also used corn trypsin inhibitor in order to minimize the contribution of the contact system, which does not seem to have a clinically relevant role in blood coagulation, as evidenced by lack of bleeding manifestations in subjects with no measurable factor XII activity. Contrary to the experience reported by others [28], in preliminary experiments (data not presented) we have seen a significant retardation of the clotting time with the concentration of corn trypsin inhibitor used in our system; this is consistent with the contact pathway of initiation of coagulation being blocked by corn trypsin inhibitor, and thrombin generation initiated only by the factor VIIa-tissue factor pathway. Once thrombin is generated through the initiation of coagulation by the interaction of trace amounts of activated factor VII with tissue factor, platelets are activated and support further generation of higher concentrations of thrombin, which lead to the formation of fibrin [29,30]. The end result of this complex system is the formation of a fibrin clot. Therefore, it is not surprising that the velocity profile of the TEG® system, an instrument that allows acquisition of continuous quantitative information on the developing clot, correlates with thrombin generation. Thus, the thrombin generation curve as measured by the TAT reference method and the thrombus generation curve as generated by the TEG® system are inter-related. In this study, Fig. 3 shows that the majority of the generated data points representing the two variables are within the prediction intervals of the regression line. In addition, the high magnitude of r (0.94) at P ≤ 0.0001 and the high significance of the prediction equation demonstrate the high efficacy of the prediction of thrombin generation, as measured by the TAT method and by the TEG® TTG variable.

As shown in Fig. 4, the first evidence of fibrin/platelet clot formation manifested by the deflection of TEG® tracing takes place at a time when a very small proportion of the total TAT generation has occurred. This is consistent with our previously reported observation that a clot is formed in the whole blood when <5% of the total thrombin has been generated [31].

It is interesting to note in Fig. 4 that the graphic representation of the cumulative generation of thrombin in normal blood closely parallels the TEG® tracing of the progressive increment of the clot tensile strength. It remains to be seen if this parallelism will be confirmed in different hemostatic disorders, which lead to decreased thrombin generation associated with decreased clot tensile strength demonstrated by TEG®.

In conclusion, we believe that our data demonstrate a clear and physiologically relevant correlation and highly significant least-squares fit between the velocity of thrombus tensile strength generation expressed by the TTG numerical values and the velocity of thrombin generation expressed by TAT numerical values. We believe that, for the precise studies of the kinetics of clotting process in the whole blood, direct measurement of thrombin generation assessed by the TAT complex in non-citrated blood is the best test available currently. For clinical practice where tests have to be simple and results need to be available rapidly, TTG is a reasonable proxy for the assessment of thrombin generation.


This work was supported in part by a grant from Bayer Inc. Canada.