To the Editor:—
Bivalirudin is a bivalent direct thrombin inhibitor (DTI) with a plasma half-life of approximately 25 min. It is increasingly used as a heparin alternative in percutaneous coronary intervention procedures,1and in cardiac surgical patients with heparin-induced thrombocytopenia.2,3The anticoagulant effect of bivalirudin usually is monitored with activated clotting time, but the utility of viscoelastic monitors, including Thrombelastograph® or thromboelastometry (e.g. , ROTEM®; Pentapharm, Munich, Germany), has been recently demonstrated by us and the other group.4,5We herein describe a case in which bivalirudin monitoring with ROTEM® was found useful. A 68-yr-old, 110-kg female was diagnosed with acute myocardial infarction. The cardiac catheterization showed ostial occlusion of left anterior descending artery, 90% stenosis of right coronary artery, and 40% stenosis of circumflex artery. Past medical history was notable for hypertension, hypercholesterolemia, type II diabetes, chronic anemia, and mild cirrhosis. Given her previous history of heparin-induced thrombocytopenia and persistent antibody titer on admission, she was treated with argatroban infusion at 0.5 μg · kg · min until 4 h before the scheduled off-pump coronary bypass graft surgery. Baseline laboratory results showed hematocrit 33.3%, platelet 111 × 103/mm3, fibrinogen 432 mg/dl, partial thromboplastin time 73.8 s, and celite-activated clotting time 188 s. For anticoagulation, bivalirudin was given at 0.75 mg/kg, followed by infusion at 0.75 mg · kg · h, according to her ideal body weight (70 kg). This regimen maintained activated clotting time above 400 s, and delayed tissue factor induced thrombus formation on ROTEM® (fig. 1A). Bivalirudin infusion was stopped at the beginning of proximal anastomoses, and laboratory values were hematocrit 21%, platelet 103 × 103/mm3, and activated clotting time 278 s at the end of three-vessel bypass procedure. Two units of packed red blood cells were administered, and infusion of tranexamic acid 2 mg · kg · h was started. Notably, the plasma fibrinogen level using the modified Clauss method (BCS®, Dade Behring, Deerfield, IL) was reported at that time as less than 60 mg/dl, but the ROTEM®-based fibrinogen assay (maximal clot formation of FibTEM®[Pentapharm, Munich, Germany]) indicated functional fibrinogen level6throughout the bivalirudin infusion (fig. 1A). The postoperative chest tube drainage was 180 ml over 12 h.
Fig. 1. A Thromboelastometry (ROTEM®; Pentapharm, Munich, Germany) tracings during bivalirudin anticoagulation. Native whole blood samples (300 μl) were triggered with 20 μl of the commercial tissue factor reagent EXTEM (Pentapham). FibTEM® assays were performed similarly with the addition of cytochalasin D to exclude platelet interactions with polymerized fibrin. It is notable that blood clotting ( i.e. , fibrin polymerization) occurs even after the bolus bivalirudin injection after a delay. B Thromboelastometry (ROTEM®) tracings in pooled plasma spiked with argatroban or bivalirudin. Plasma samples (300 μl) were triggered with 20 μl of EXTEM reagent and CaCl2. The onset of fibrin polymerization was more delayed with argatroban at 10 μg/ml than bivalirudin at 20 μg/ml due to lesser molar concentration of the latter (20 μm vs. 10 μm). The maximal clot formation reaches the same level over time.
Fig. 1. A Thromboelastometry (ROTEM®; Pentapharm, Munich, Germany) tracings during bivalirudin anticoagulation. Native whole blood samples (300 μl) were triggered with 20 μl of the commercial tissue factor reagent EXTEM (Pentapham). FibTEM® assays were performed similarly with the addition of cytochalasin D to exclude platelet interactions with polymerized fibrin. It is notable that blood clotting ( i.e. , fibrin polymerization) occurs even after the bolus bivalirudin injection after a delay. B Thromboelastometry (ROTEM®) tracings in pooled plasma spiked with argatroban or bivalirudin. Plasma samples (300 μl) were triggered with 20 μl of EXTEM reagent and CaCl2. The onset of fibrin polymerization was more delayed with argatroban at 10 μg/ml than bivalirudin at 20 μg/ml due to lesser molar concentration of the latter (20 μm vs. 10 μm). The maximal clot formation reaches the same level over time.
To evaluate the effect of direct thrombin inhibitors (DTIs) on the fibrinogen measurements using the modified Clauss method, argatroban, bivalirudin, lepirudin, and heparin at various concentrations were added in vitro to the pooled human plasma, (CRYOcheck lot No. A1044; PrecisionBioLogic, Dartmouth, Nova Scotia). The concentrations of DTIs and heparin used in our study span from therapeutic to supratherapeutic levels according to the reported plasma (molar) concentrations; argatroban 1–5 μg/ml (2–10 μm), bivalirudin 2–20 μg/ml (1–10 μm), lepirudin 1–5 μg/ml (0.15–0.75 μm), and heparin 0.4–4 U/ml.7–9The measured fibrinogen level in the presence of three DTIs and heparin was progressively lowered in the BCS® system relative to the baseline measurement (311 ± 7.6 mg/dl), and it also was affected to a lesser extent in the STA-R Evolution® system (Diagnostica-Stago, Parsippany, NJ) (table 1). The fibrinogen measurements using the maximal clot formation parameter of ROTEM® were unchanged when the spiked plasma samples were recalcified and activated with tissue factor, although DTIs delayed the clotting process (fig. 1B). Because bovine thrombin is used as a reagent for these assays based on the modified Clauss method,10test results are susceptible to antithrombin effects of DTIs and heparin. This effect was more evident at therapeutic concentrations of bivalirudin tested when using the BCS® system compared with STA-R® (table 1). The difference may be attributed to the clot detection mechanisms, or different amounts of exogenous thrombin added for clotting. The BCS® system derives fibrinogen levels using photo-optical turbidity changes, whereas STA-R® magnetically senses an increased viscosity due to clotting using the pendulum motion of a steel ball. The rate of change in turbidity (i.e. , fibrin polymerization) is clearly affected by DTIs, particularly argatroban (table 1). The STA-R® and ROTEM® maximal clot formation seem to be less susceptible to DTIs, due to their viscosity detection mechanisms, but α-angle of ROTEM®, which reflects the rate of fibrin polymerization, is reduced by DTIs (fig. 1B).4,5,11Our present data demonstrate that fibrinogen measurements are affected in the order of argatroban > bivalirudin > lepirudin for the BCS® method. Because the molecular weight of argatroban (527 Da) is much smaller than that of bivalirudin and lepirudin (2,180 Da and 6,980 Da, respectively), exogenously added thrombin for the fibrinogen assay is more inhibited in the molar excess of argatroban. Similarly, Warkentin et al. 12previously demonstrated that DTIs affect prothrombin time in a molar concentration dependent manner, thus therapeutic concentrations of argatroban prolong prothrombin time more than bivalirudin and lepirudin. Heparin also affects fibrinogen measurements, but less extensively than lepirudin (table 1). Presently, there is no antidote for anticoagulation with DTIs. Low fibrinogen levels during DTI therapy may worsen bleeding because all DTIs compete with fibrinogen for thrombin. The maintenance of fibrinogen levels is critical in achieving hemostasis after cardiac surgery.13,14
In summary, we demonstrate the turbidmetric fibrinogen assay is particularly affected by DTIs, and falsely low fibrinogen levels are reported during bivalirudin anticoagulation. The modified thromboelastometry using abciximab or cytochalasin D can be used to assess functional fibrin polymerization, and it may be useful for evaluating hemostatic function and recovery from bivalirudin therapy.5,6,15
*Emory University School of Medicine, Atlanta, Georgia. kenichi.tanaka@emoryhealthcare.org
