Anticoagulation is often required perioperatively for cardiovascular procedures or for the prevention of thrombotic complications. Anticoagulation has been conventionally achieved with unfractionated heparin because of its ease of use and reversibility. Bivalirudin (Angiomax®, formerly Hirulog-1) is a polypeptide direct thrombin inhibitor (D-FPRPGGGGNGDFEEIPEEYL) with a half-life of approximately 25 min used for patients with heparin-induced thrombocytopenia and for percutaneous coronary artery interventions (1). In contrast to heparin, which promotes antithrombin catalyzed inhibition of thrombin and factor Xa, bivalirudin directly inhibits thrombin, but not factor Xa.
The activated partial thromboplastin time (aPTT) and activated clotting time (ACT) are assays commonly used to monitor anticoagulation (2,3). Both tests are terminated when fibrin gel formation occurs as the result of production of trace amounts of thrombin (2–5 nM; <5% of total thrombin generated) (2). The results of aPTT and ACT may not, therefore, reflect the extent of thrombin inhibition by anticoagulants because the majority (∼95%) of thrombin generation (i.e., propagation phase) takes place on the formed thrombus surface after the testing has stopped. Because heparin and bivalirudin have different modes of action, monitoring thrombus formation beyond clotting time with viscoelastic methods may better reflect the extent of thrombin generation compared with the simple blood clotting tests (3). We hypothesized that differences in the kinetics of thrombin generation and thrombus formation in the presence of either drug could only be detected by tests that were performed beyond the initial fibrin gel formation. Therefore, the aim of this study was to evaluate for differences in the mode of thrombin inhibition between unfractionated heparin and bivalirudin using continuous measurement of thrombin generation (4), and viscoelastic changes of clot formation (5,6).
After IRB approval and informed written consent, blood samples were obtained in 3.2% citrated glass tubes from 12 healthy volunteers not receiving any medication for the previous 2 wk. For thrombin generation studies, the samples were centrifuged at 2000g for 20 min to isolate platelet poor plasma (PPP). Unfractionated heparin and bivalirudin were obtained from Elkinns-Sinn (Cherry Hill, NJ) and the Medicines Company (Parsippany, NJ), respectively. The fluorogenic substrate (Z-GGR-AMC HCl, Bachem, Switzerland) was prepared for thrombin generation assay as a 100 mM solution in dimethyl sulfoxide (Sigma Aldrich, St. Louis, MO). Just before use, 50 μL of the solution was added to 1.75 mL of prewarmed N-2-hydroxyethyl piperazine-N′-2-ethanesulfonic acid (pH 7.35, 20 mM N-2-hydroxyethyl piperazine-N′-2-ethanesulfonic acid, 60 mg/mL bovine serum albumin, Sigma Aldrich, St. Louis, MO) buffer containing 0.2 mL of 1 M CaCl2.
Thrombin Generation in the Presence of Heparin or Bivalirudin
The Thrombinoscope™ system (Synapse BV, Maastricht, the Netherlands), measures the onset and the amount of thrombin generation based on the change in fluorescence produced by the hydrolysis of a fluorogenic peptide that acts as substrate for the active site of thrombin. The ability of thrombin inhibitors to decrease the substrate cleavage (rate and amount) is displayed as a decrease in peak height of thrombin curve, and the delay in thrombin generation shows as a prolongation of lag time.
Thrombin generation experiments were performed in plasma samples from six volunteers (n = 6). Samples were run in quadruplicate according to method of Hemker et al. (4). Briefly, to each well of a 96-well microtiter plate (Microfluor black, Thermolabsystems, Franklin, MA), we added 80 μL of PPP containing heparin (final concentration, 0–0.5 U/mL) or bivalirudin (final concentration, 0–30 μg/mL; 0–13.6 μM). PPP without the drugs served as control. Calibrator wells, in which 20 μL of thrombin calibrator (Synapse BV, Maastricht, Netherlands) was added to 80 μL of plasma samples, were run in parallel for each volunteer's plasma. Coagulation was triggered with 20 μL commercially available PPP reagent (relipidated tissue factor, Thrombinoscope BV, Maastricht, the Netherlands) or diluted Actin® FS (1:20 dilution, Dade Behring, Marburg, Germany) to simulate activation of extrinsic and intrinsic (contact) coagulation pathways. The reaction was started by adding 20 μL/well of substrate—CaCl2-subtrate buffer. The development of fluorescence was continuously monitored for 70–90 min with the fluorescence reader (Fluoroscan Ascent, 390/460 nm excitation/emission wavelengths, Thermo Labsystems, Franklin, MA). A dedicated software program (Thrombinoscope, Synapse BV, Maastricht, The Netherlands) was used to record the experiment, and for the calculation of the thrombinogram parameters (lag time, peak thrombin level, etc.).
The Measurement of Prothrombin Fragment 1.2 in the Presence of Heparin or Bivalirudin
We compared the generation of prothrombin F1.2, a sensitive marker of thrombin generation, in whole blood samples anticoagulated with heparin (1.5 U/mL, 1.0 μM) or bivalirudin (12.5 μg/mL, 5.67 μM) (n = 6 for each treatment). Because F1.2 levels reflect the extent of conversion of prothrombin to thrombin, differences in F1.2 levels between anticoagulated samples reflect the extent of the inhibition of this conversion. Briefly, citrated whole blood samples (n = 6) (300 μL) were incubated with 0.2 M calcium chloride (37.5 μL) in polystyrene tubes preloaded with kaolin. At certain intervals (0, 1, 2, 3, 5, 10, and 15 min) during incubation at 37°C, the reaction was quenched with 1.0 mL of buffer (100 mM NaCl, 50 mM Tris, 1% bovine serum albumin, and 100 mM EDTA, pH 7.4, Sigma Aldrich, St. Louis, MO), and samples were centrifuged for 15 min (2500g) to obtain plasma. The plasma samples were stored at −70°C until batch analysis.
Prothrombin F1.2 analyses were performed using an Enzygnost F1 for 2 microenzyme immunoassay kit based on the “sandwich principle” (Dade Behring, Deerfield, IL) according to the manufacturer's directions. The absorbance was measured using a microplate reader at 490 nm (Spectramax 340, Molecular Devices, Sunnyvale, CA).
The Sonoclot Analyzer (Sienco, Denver, CO) detects the changes of viscosity in coagulating blood with an oscillating probe, displaying viscosity in the normalized scale (clot signal unit) over time. Standard variables that can be calculated include ACT in minutes and clot rate (clot signal unit per minute) (5). Sonoclot ACT is comparable to the ACT commonly used for monitoring high-dose heparin (5,7).
For analyses, the whole blood samples (n = 7) were incubated with normal saline (control), unfractionated heparin (1.5 or 2.5 U/mL, final concentration), or bivalirudin (12.5 or 25.0 μg/mL, final concentration) for 5 min at 37°C. For the Sonoclot measurement, 10 μL of 0.4 M CaCl2 was added to kaolin cuvette (kACT kit, Sienco, Denver, CO), followed by 360 μL of a whole blood sample containing saline, or the drug of interest, and the test was run according to the manufacturer's instructions.
Thrombelastogram (TEG®) Analyses
TEG (Model 5000, Hemoscope Corporation, Niles, IL) measures the viscoelastic property of clot development by fibrin formation and platelet activation during coagulation (6). The width of the tracing reflects the tensile strength of clot [maximum amplitude (MA) in millimeters]. For TEG analyses, whole blood samples were activated with kaolin in the presence of saline (control), bivalirudin, or heparin (n = 7 each). The analysis was conducted with 360 μL of whole blood and 10 μL of 0.4 M CaCl2, and the test was allowed to proceed for 70 min. Collected data included reaction time (R, min), α angle (degree), and MA (millimeter) (6).
Based on previous studies with Thrombinoscope and TEG, a sample size of 6 was needed to detect a 20% change in peak thrombin or TEG amplitude with a β ≥ 0.8 and an α < 0.05 (3,8). Statistical analyses were conducted with the paired t test or the repeated-measures analysis of variance, followed by the paired t test with the Bonferroni correction using with SPSS 15.0 (SPSS, Chicago, IL). All data were expressed as mean ± sd. P ≤ 0.05 was considered significant.
Thrombin Generation (Thrombinoscope) in PPP
Based on Thrombinoscope measurements, heparin (0–0.5 U/mL) dose-dependently reduced peak thrombin generation, resulting in an 85.0% decrease in thrombin peak when Actin was used to trigger the reaction (intrinsic activation, Fig. 1A). Essentially no thrombin was formed at a heparin concentration of 0.5 U/mL using tissue factor as trigger (extrinsic activation) (Fig. 1B). Heparin dose-dependently increased the lag time of thrombin generation when Actin was used as an activator (from 4.7 ± 1.5 min for the control to 14 ± 2.1 min at heparin 0.5 U/mL), but there was no lag time effect with tissue factor activation. Increasing concentrations of bivalirudin progressively delayed thrombin formation regardless of the activator (Figs. 2A and B). Thrombin peak levels were progressively decreased with Actin activation (21.5% ± 9.2% at 1.5 μg/mL to 69.9% ± 12.3% at 30.0 μg/mL), but with tissue factor as a trigger the decrease was more gradual; therefore, the peak level of thrombin was only reduced by 29.4% ± 6.2% at the highest bivalirudin concentration (Fig. 2B).
Prothrombin F1.2 Generation
In non-anticoagulated (control) samples F1.2 levels started to increase after the first minute (Fig. 3). Both heparin and bivalirudin delayed the formation of prothrombin F1.2 for 3 min. The levels F1.2 increased more rapidly thereafter with bivalirudin than with heparin. By 10 min, the F1.2 level in bivalirudin samples was not significantly different from control, non-anticoagulated samples (206 ± 28.2 vs 182 ± 23.9 nmol/L, P = 0.09). F1.2 levels also increased in the heparin samples, but peak levels (75.7 ± 29.8 nmol/L, P < 0.05) remained significantly below bivalirudin samples (Fig. 3).
Both heparin and bivalirudin dose-dependently decreased Sonoclot ACT values, but there was no significant difference between samples with heparin 1.5 U/mL and bivalirudin 12.5 μg/mL, or with heparin 2.5 U/mL and bivalirudin 25 μg/mL (Table 1). Clot rate, which provides information on the rate of fibrin polymerization, was more profoundly inhibited with heparin than with bivalirudin (Table 1 and Fig. 4A). With heparin, clot rate decreased ∼67% at 1.5 U/mL to ∼74% at 2.5 U/mL, whereas bivalirudin caused a ∼43% decrease at 12.5 μg/mL and ∼54% decrease at 25 μg/mL.
The R (min) representing the lag time for clot formation was similarly increased for heparin and bivalirudin samples (Table 1). The rate of clot development as shown by the α angle decreased in a concentration-dependent manner in heparin samples (>90%), but not in bivalirudin samples (∼40% decrease at 12.5 and 25.0 μg/mL) (Table 1). Likewise, the overall tensile strength of the clot, as reflected by the MA values, decreased in a concentration-dependent manner with heparin, but not with bivalirudin. Even at high heparin concentrations (25 μg/mL, Table 1 and Fig. 4B), the MA only decreased by ∼12.5% vs ∼99% for heparin at 2.5 U/mL.
In the current in vitro study, we have shown that thrombin generation kinetics is distinctively different for heparin and bivalirudin. In the thrombin generation experiments using Thrombinoscope methodology, heparin reduced peak thrombin levels regardless of the activation trigger (Figs. 1A and B). Peak thrombin levels, however, remained unchanged with bivalirudin unless high plasma concentrations were reached (>15 μg/mL, Figs. 2A and B).
The more pronounced lag time prolongation with heparin and bivalirudin underlies the intrinsic pathway's dependence on thrombin-mediated feedback activation of factors V, VIII, and XI (9–11). In plasma-activated via extrinsic activation using 5 pM of tissue factor with phospholipids (4), thrombin generation was less dependent on the feedback involvement of the intrinsic pathway, and the lag time was less affected by heparin and bivalirudin. The present data show that heparin and bivalirudin similarly delayed the onset of thrombin generation by inhibiting thrombin's initial autoactivation (propagation), but thrombin's catalytic activity was almost fully recovered after the decay of bivalirudin. Our observations are in agreement with Prasa et al. who showed that the peak thrombin is reduced by active site blockade with FPR-CH2Cl (PPACK), but not with hirulog-1 (former name of bivalirudin) (12), and closely follows the simulated thrombin generation in the presence of direct thrombin inhibitors as shown by Adams et al. (13).
Prothrombin F1.2 is released during the conversion of prothrombin to thrombin; its measurement quantitatively determines the extent of thrombin generation. In kaolin-activated whole blood, the burst increases in F1.2 levels were observed early (1 min) in non-anticoagulated samples, but were delayed with either heparin (1.5 U/mL) or bivalirudin (12.5 μg/mL). After 8 min delay, there was a late surge of F1.2 level observed with bivalirudin only, reaching the level obtained in control (non-anticoagulated) samples by 10 min. The concentrations of the bivalirudin used in our experiments (>15 μg/mL; 6.8 μM) were well above the plasma levels of prothrombin, 1.4 μM. Therefore, apparent recovery of thrombin catalytic activity shown on Thrombinoscope and F1.2 assay is attributable to thrombin-mediated N-terminal cleavage (14) and fast dissociation of bivalirudin from thrombin (15). Once thrombin regains its catalytic function from bivalirudin, feedback activation of prothrombinase and subsequent explosive thrombin generation take place (12). Conversely, heparin–antithrombin-mediated thrombin inhibition involves irreversible conformational changes in the active site of thrombin (16). Therefore, the differences in the kinetics of clot formation between heparin and bivalirudin can be explained by recovery of thrombin's catalytic activity with the latter drug.
Our results from the thrombin generation assay using Thrombinoscope and the F1.2 assay does not reflect the extent of interaction between thrombin with its natural substrates, e.g., fibrinogen. Therefore, viscoelastic coagulation monitors were used to assess the differences in thrombus formation between heparin and bivalirudin using fibrin polymerization as an end-point. Despite similar delays in Sonoclot ACT and TEG R time (Table 1), the rates of thrombus development reflected by the clot rate of Sonoclot and α angle of TEG were inhibited more with heparin than with bivalirudin (Table 1). Further, clot strength reflected by the TEG MA variable was significantly decreased in the heparin samples, but not in bivalirudin samples, when compared with non-anticoagulated controls. These findings agree with the results of prothrombin F1.2 measurements (Fig. 3). Clot detection with Sonoclot and TEG occurred later than the increases of F1.2 with either heparin or bivalirudin, and the rate of F1.2 formation with bivalirudin subsequently surpassed the level obtained with heparin (Table 1, Fig. 3). It has been reported that heparin may directly activate or inhibit platelet function (17,18). However, subtle changes of platelets per se are unlikely to reduce thrombus formation because platelet function can be evaluated by polymerized fibrin interaction with activated platelets on TEG under heparin anticoagulation (19). The propagation of thrombus formation in the presence of bivalirudin is consistent with the functional recovery of thrombin. Our interpretation of TEG findings are in agreement with Nielsen et al.'s data, which show that thrombus formation was suppressed with heparin (>3 U/mL) or lepirudin (>10 μg/mL), but not with short-acting bivalirudin and argatroban (3).
The concentrations of heparin and bivalirudin used in our study are within target therapeutic levels during cardiovascular surgery: 2.5 U/mL and 20 μg/mL (∼10 μM), respectively (20,21). However, the in vitro recovery of thrombin activity under bivalirudin anticoagulation cannot be explained simply by the reduced anticoagulant efficacy of this drug, because our experiments have been conducted with minimal shear and without any endothelial component. When the blood is circulating in vivo, procoagulant activities of thrombin may be efficiently inhibited with continuously infused bivalirudin. Additionally, systemic endothelial cell surface holds immense anticoagulant capacity (e.g., heparan sulfate, nitric oxide, prostacyclin, plasminogen activator, thrombomodulin). Nevertheless, our in vitro results indicate that, in the static environment, thrombin activity can be recovered from transient bivalirudin inhibition. In the absence of heparin, antithrombin-mediated inhibition of thrombin is slow, and thrombin activity may not be well regulated after bivalirudin anticoagulation “wears off” in tissue factor-rich pleural cavity (22). Increased circulating F1.2 levels have been found after reinfusion of salvaged blood during bivalirudin anticoagulation (23).
In our in vitro study, we focused primarily on the intrinsic activation of clot formation for several reasons. First, aPTT (24) and ACT (25) are used for monitoring heparin and bivalirudin because the intrinsic pathway is very sensitive to inhibition of thrombin feedback activation of prothrombinase (9,10). Second, intrinsic activation is an important trigger of pathological coagulation during vessel injury or in the presence of the thrombogenic surface (26,27). Sonoclot and TEG with intrinsic activation provide advantages over conventional clotting tests, which terminate when the trace amount (2–5 nM) of thrombin is formed (28). These devices depict the stability of a clot when covalent crosslinking of fibrin polymers formed from fibrin monomers has occurred. Because the crosslinking of fibrin occurs after the end-point of ACT, prothrombin time or aPTT assays, they are less dependent on the activity of factor XIII, and do not reflect the clot strength. Further, some studies suggest better predictive values of Sonoclot and TEG testing (29,30) than prothrombin time or aPTT (31) and, therefore, bedside application of viscoelastic monitoring may allow a better individualized monitoring of antithrombotic therapy (3,5,7,32,33).
In summary, the current study shows that the similar inhibition of the onset of clotting with heparin and bivalirudin does not necessarily reflect a similar extent of thrombin formation or clot formation kinetics. The optimal anticoagulation with bivalirudin during cardiac surgery with cardiopulmonary bypass may require a continuous maintenance of plasma levels and avoidance of potent procoagulant stimuli and blood stasis. The viscoelastic devices can be useful in monitoring the therapeutic levels of the drug, and the recovery from anticoagulation during bivalirudin infusion.
1. Kelly AB, Maraganore JM, Bourdon P, Hanson SR, Harker LA. Antithrombotic effects of synthetic peptides targeting various functional domains of thrombin. Proc Natl Acad Sci USA 1992;89:6040–4
2. Rand MD, Lock JB, van't Veer C, Gaffney DP, Mann KG. Blood clotting in minimally altered whole blood. Blood 1996;88:3432–45
3. Nielsen VG, Steenwyk BL, Gurley WQ, Pereira SJ, Lell WA, Kirklin JK. Argatroban, bivalirudin, and lepirudin do not decrease clot propagation and strength as effectively as heparin-activated antithrombin in vitro. J Heart Lung Transplant 2006;25:653–63
4. Hemker HC, Giesen P, Al Dieri R, Regnault V, de Smedt E, Wagenvoord R, Lecompte T, Beguin S. Calibrated automated thrombin generation measurement in clotting plasma. Pathophysiol Haemost Thromb 2003;33:4–15
5. Shibata T, Sasaki Y, Hattori K, Hirai H, Hosono M, Fujii H, Suehiro S. Sonoclot analysis in cardiac surgery in dialysis-dependent patients. Ann Thorac Surg 2004;77:220–5
6. Kawasaki J, Katori N, Kodaka M, Miyao H, Tanaka KA. Electron microscopic evaluations of clot morphology during thrombelastography. Anesth Analg 2004;99:1440–4
7. Dalbert S, Ganter MT, Furrer L, Klaghofer R, Zollinger A, Hofer CK. Effects of heparin, haemodilution and aprotinin on kaolin-based activated clotting time: in vitro comparison of two different point of care devices. Acta Anaesthesiol Scand 2006;50:461–8
8. Tanaka KA, Szlam F, Katori N, Sato N, Vega JD, Levy JH. The effects of argatroban on thrombin generation and hemostatic activation in vitro. Anesth Analg 2004;99:1283–9
9. Ofosu FA, Sie P, Modi GJ, Fernandez F, Buchanan MR, Blajchman MA, Boneu B, Hirsh J. The inhibition of thrombin-dependent positive-feedback reactions is critical to the expression of the anticoagulant effect of heparin. Biochem J 1987;243:579–88
10. Beguin S, Lindhout T, Hemker HC. The mode of action of heparin in plasma. Thromb Haemost 1988;60:457–62
11. Ofosu FA, Fenton JW II, Maraganore J, Blajchman MA, Yang X, Smith L, Anvari N, Buchanan MR, Hirsh J. Inhibition of the amplification reactions of blood coagulation by site-specific inhibitors of alpha-thrombin. Biochem J 1992;283:893–7
12. Prasa D, Svendsen L, Sturzebecher J. The ability of thrombin inhibitors to reduce the thrombin activity generated in plasma on extrinsic and intrinsic activation. Thromb Haemost 1997;77: 498–503
13. Adams TE, Everse SJ, Mann KG. Predicting the pharmacology of thrombin inhibitors. J Thromb Haemost 2003;1:1024–7
14. Parry MA, Maraganore JM, Stone SR. Kinetic mechanism for the interaction of Hirulog with thrombin. Biochemistry 1994;33: 14807–14
15. Elg M, Gustafsson D, Deinum J. The importance of enzyme inhibition kinetics for the effect of thrombin inhibitors in a rat model of arterial thrombosis. Thromb Haemost 1997;78:1286–92
16. Li W, Johnson DJD, Esmon CT, Huntington JA. Structure of the antithrombin-thrombin-heparin ternary complex reveals the antithrombotic mechanism of heparin. Nat Struct Mol Biol 2004;11:857–62
17. Matsuo T, Nakao K, Yamada T, Matsuo O. Effect of a new anticoagulant (MD 805) on platelet activation in the hemodialysis circuit. Thromb Res 1986;41:33–41
18. Carr ME Jr, Carr SL, Roa V, McCardell KA, Greilich PE. Aprotinin counteracts heparin-induced inhibition of platelet contractile force. Thromb Res 2002;108:161–8
19. Craft RM, Chavez JJ, Bresee SJ, Wortham DC, Cohen E, Carroll RC. A novel modification of the Thrombelastograph assay, isolating platelet function, correlates with optical platelet aggregation. J Lab Clin Med 2004;143:301–9
20. Despotis GJ, Joist JH, Hogue CW Jr, Alsoufiev A, Joiner-Maier D, Santoro SA, Spitznagel E, Weitz JI, Goodnough LT. More effective suppression of hemostatic system activation in patients undergoing cardiac surgery by heparin dosing based on heparin blood concentrations rather than ACT. Thromb Haemost 1996;76:902–8
21. Koster A, Fischer T. Management of patients with heparin-induced thrombocytopenia during cardiac surgery. Semin Cardiothorac Vasc Anesth 2003;7:411–16
22. Burman JF, Westlake AS, Davidson SJ, Rutherford LC, Rayner AS, Wright AM, Morgan CJ, Pepper JR. Study of five cell salvage machines in coronary artery surgery. Transfus Med 2002;12:173–9
23. Koster A, Yeter R, Buz S, Kuppe H, Hetzer R, Lincoff AM, Dyke CM, Smedira NG, Spiess B. Assessment of hemostatic activation during cardiopulmonary bypass for coronary artery bypass grafting with bivalirudin: Results of a pilot study. J Thorac Cardiovasc Surg 2005;129:1391–4
24. Warkentin TE, Greinacher A, Craven S, Dewar L, Sheppard JI, Ofosu FA. Differences in the clinically effective molar concentrations of four direct thrombin inhibitors explain their variable prothrombin time prolongation. Thromb Haemost 2005;94:958–64
25. Merry AF, Raudkivi PJ, Middleton NG, McDougall JM, Nand P, Mills BP, Webber BJ, Frampton CM, White HD. Bivalirudin versus heparin and protamine in off-pump coronary artery bypass surgery. Ann Thorac Surg 2004;77:925–31
26. Rosen ED, Gailani D, Castellino FJ. FXI is essential for thrombus formation following FeCl3-induced injury of the carotid artery in the mouse. Thromb Haemost 2002;87:774–6
27. Gruber A, Hanson SR. Factor XI-dependence of surface- and tissue factor-initiated thrombus propagation in primates. Blood 2003;102:953–5
28. Brummel KE, Paradis SG, Butenas S, Mann KG. Thrombin functions during tissue factor-induced blood coagulation. Blood 2002;100:148–52
29. Nuttall GA, Oliver WC, Ereth MH, Santrach PJ. Coagulation tests predict bleeding after cardiopulmonary bypass. J Cardiothorac Vasc Anesth 1997;11:815–23
30. McCrath DJ, Cerboni E, Frumento RJ, Hirsh AL, Bennett-Guerrero E. Thromboelastography maximum amplitude predicts postoperative thrombotic complications including myocardial infarction. Anesth Analg 2005;100:1576–83
31. Murray D, Pennell B, Olson J. Variability of prothrombin time and activated partial thromboplastin time in the diagnosis of increased surgical bleeding. Transfusion 1999;39:56–62
32. Katori N, Tanaka KA, Szlam F, Levy JH. The effects of platelet count on clot retraction and tissue plasminogen activator-induced fibrinolysis on thrombelastography. Anesth Analg 2005;100:1781–5
33. Carroll RC, Chavez JJ, Simmons JW, Snider CC, Wortham DC, Bresee SJ, Cohen E. Measurement of patients' bivalirudin plasma levels by a thrombelastograph ecarin clotting time assay: a comparison to a standard activated clotting time. Anesth Analg 2006;102:1316–19