Heparin-induced thrombocytopenia (HIT) is a severe

Transcription

1 Bivalirudin During Cardiopulmonary Bypass in Patients With Previous or Acute Heparin-Induced Thrombocytopenia and Heparin Antibodies: Results of the CHOOSE-ON Trial Andreas Koster, MD, Cornelius M. Dyke, MD, Gabriel Aldea, MD, Nicholas G. Smedira, MD, Harry L. McCarthy II, CCP, Solomon Aronson, MD, Roland Hetzer, MD, Edwin Avery, MD, Bruce Spiess, MD, and A. Michael Lincoff, MD Departments of Anesthesia and Department of Cardiothoracic and Vascular Surgery, Deutsches Herzzentrum Berlin, Germany; Department of Cardiovascular and Thoracic Surgery, Gaston Memorial Hospital, Gastonia, North Carolina; Department of Cardiac Surgery, Washington State University, Seattle, Washington; Departments of Cardiothoracic Surgery and Cardiovascular Medicine, The Cleveland Clinic Foundation, Cleveland, Ohio; Departments of Cardiac Surgery and Cardiothoracic Anesthesia, Virginia Commonwealth University, Richmond, Virginia; Department of Anesthesia, Duke University Medical Center, Durham, North Carolina; Department of Cardiac Anesthesia, Massachusetts General Hospital, Boston, Massachusetts Background. The coronary artery bypass grafting (CABG) heparin-induced thrombocytopenia thrombosis syndrome (HITTS) on- and off-pump safety and efficacy (CHOOSE-ON) trial was designed as a safety and efficacy trial of bivalirudin for use in anticoagulation during cardiopulmonary bypass (CPB) in patients with confirmed or suspected HIT and (or) antiplatelet factor 4/heparin (anti-pf4/h) antibodies. Methods. In an open-label, multicenter trial, 50 patients were enrolled prospectively. The primary study endpoint was in-hospital acute procedural success, defined as the absence of death, Q-wave myocardial infarction (MI), repeat operation for coronary revascularization, and stroke at day seven after surgery or hospital discharge, whichever occurred first. The secondary study endpoints were procedural success, defined as the absence of death, Q-wave MI, repeat operation for coronary revascularization, and stroke, at 30 days and 12 weeks after surgery. Perioperative blood loss, transfusions, and the incidence of major bleeding events were also captured. Results. There were 49 patients treated with bivalirudin of which 43 had acute HIT and thrombosis syndrome (HITTS) with antibodies at time of surgery. Procedural success in-hospital or at 7 days was achieved in 46 (94%) patients. At day 30 procedural success was achieved in 42 (86%) patients, and after 12 weeks in 40 (82%) patients. Mean intraoperative blood loss was ml, and mean 24-hour postoperative blood loss was ml. Forty-one (84%) patients received transfusions before day 7 or discharge with a mean of units of red blood cells, units of platelets, and units of fresh frozen plasma. No differences in outcome among bivalirudin-treated patients were observed between those in the overall group and those with moderately impaired renal function (n 10). Conclusions. The current investigation expands the experience of safe and effective anticoagulation with bivalirudin during CPB to patients with confirmed or suspected HIT and anti-pf4/h antibodies, including in the setting of impaired renal function. (Ann Thorac Surg 2007;83:572 7) 2007 by The Society of Thoracic Surgeons Heparin-induced thrombocytopenia (HIT) is a severe immune-mediated disease associated with devastating thromboembolic complications such as limb ischemia, myocardial infarction (MI), stroke, pulmonary embolism, and death [1]. In patients diagnosed with HIT further heparin administration should be avoided. For patients requiring medical management of HIT and nonsurgical preoperative and postoperative anticoagulation, approved therapies in the United States are limited to argatroban and lepirudin, and in non-us jurisdictions additionally to the heparinoid danaproid sodium. At this time, however, for HIT patients requiring cardiac surgery, especially if use of cardiopulmonary bypass (CPB) is necessary, no approved alternative to heparin for intraoperative anticoagulation exists. Moreover, current strategies are associated with significant limitations, particularly with regard to their safety profiles [1]. Accepted for publication Sept 11, Address correspondence to Dr Koster, Deutsches Herzzentrum Berlin, Augustenburger Platz 1, D Berlin, Germany; koster@ dhzb.de. Drs Lincoff, Koster, Dyke, Aronson, Spiess, and Mr McCarthy disclose that they have a financial relationship with The Medicines Company by The Society of Thoracic Surgeons /07/$32.00 Published by Elsevier Inc doi: /j.athoracsur

2 Ann Thorac Surg KOSTER ET AL 2007;83:572 7 THE CHOOSE-ON TRIAL 573 Bivalirudin (Angiomax; The Medicines Company, Parsippany, NJ) is a short-acting, bivalent, reversible direct thrombin inhibitor. Its pharmacokinetics is characterized by a rapid onset of effect and a short half-life of approximately 25 minutes. The drug s elimination is predominately achieved by proteolytic cleavage and, to a minor extent, by renal excretion [2]. Bivalirudin has been safely studied in large studies enrolling more than 40,000 patients and is approved by the Food and Drug Administration in patients undergoing percutaneous coronary intervention, including those with or at risk for HIT. A rapid onset of action, along with rapid elimination essentially independent of specific organ involvement, renders bivalirudin a potentially valuable alternative to heparinprotamine for high-dose anticoagulation during cardiovascular surgery and CPB. In pilot investigations and a recent multicenter study comparing bivalirudin and heparin-protamine in non-hit patients, protocols for CPB anticoagulation with bivalirudin have been established and successfully used for anticoagulation during CPB [3 5]. The coronary artery bypass (CABG) HIT thrombosis syndrome (TS) on- and off-pump safety and efficacy (CHOOSE-ON) trial was designed to assess the safety and efficacy of bivalirudin anticoagulation during CPB in patients with confirmed or suspected HIT and (or) antiplatelet factor 4/heparin (anti-pf4/h) antibodies. Patients and Methods Trial Design The CHOOSE-ON trial was an open-label, multicenter trial assessing the safety and efficacy of bivalirudin anticoagulation during CPB in patients with confirmed or suspected HIT and (or) anti-pf4/h antibodies. Because of the lack of an approved alternative, no active comparator arm could be planned. Bivalirudin safety and efficacy was therefore intended to be compared with a historical control group of patients with confirmed or suspected HIT who had undergone cardiac surgery with CPB using alternative anticoagulation strategies at the enrolling centers within a period of five years before initiation of the trial. Due to the inherent difficulties in identifying suitable historical controls that would allow matched propensity scoring with the prospectively enrolled patients, this first analysis of the trial exclusively presents the prospective bivalirudin-treated patients. The study followed the Declaration of Helsinki on Ethical Principles for Medical Research Involving Human Subjects. All centers had approval of the local ethics committee and all prospectively enrolled patients signed informed consent prior to participation in the study. Patients were at least 18 years of age, scheduled for first time or redo-cabg, single valve surgery, or CABG plus single-valve surgery, and had a diagnosis of confirmed or suspected HIT. Patients who had severe renal dysfunction (creatinine clearance [CrCl] 30 ml/minute), were dependent on dialysis, had a ventricular ejection fraction of less than 0.30, or required surgery on more than one heart valve were excluded from the trial, due to the potential effect of multiple morbidities in confounding the evaluation of drug-dependent complications. Patients with a recent stroke (within six months of potential enrollment), or with a residual neurologic deficit were also excluded from participation in the clinical program. This exclusion was implemented to allow an accurate safety evaluation in relation to the incidence of neurologic events in the postoperative period. Patients with confirmed HIT were those with serologic and clinical evidence of the disease. Those with suspected HIT had a past diagnosis of HIT, a drop in platelets greater than 50% from baseline in conjunction with heparin therapy, or serologic evidence of anti-pf4/h antibodies. As HIT antibodies are transient and may persist for 40 to 100 days (essentially dependent on the type of assay performed to identify antibodies-hit) after termination of heparin therapy, patients were considered to be antibody positive at time of surgery when the time point of diagnosis was no longer than 40 days prior to surgery [6, 7]. The primary study endpoint was in-hospital acute procedural success, defined as the absence of death, Q-wave MI, repeat operation for coronary revascularization, or stroke, at day seven after surgery or hospital discharge, whichever occurred first. The secondary study endpoints included procedural success, defined as the absence of death, Q-wave MI, repeat operation for coronary revascularization, or stroke, at 30 days and 12 weeks after surgery. Perioperative blood loss, transfusions, and the incidence of major bleeding events were also captured through day seven or hospital discharge, whichever occurred first. As there was no predefined algorithm for triggers of transfusions, donor blood components were transfused according to institutional standards. Criteria for diagnosis of immediate postoperative non- Q-wave MI were a creatine kinase-myocardial band isoenzyme (CK-MB) elevation to 10 times or more of the upper limit of normal within 24 hours after surgery. After 24 hours, non-q-wave MI was diagnosed if the CK-MB was elevated to two times the upper limit of normal on a single sample, or to greater than the upper limit of normal in two successive samples and either clinical symptoms or electrocardiographic changes indicating ischemia were present. Myocardial infarctions were adjudicated by an independent and blinded clinical events committee prior to statistical analysis. Major bleeding was defined as any intracranial or intraocular hemorrhage, retroperitoneal or gastrointestinal bleeding, or persistent postoperative hemorrhage requiring surgical reexploration. Anticoagulation Protocols Bivalirudin anticoagulation during CPB was performed as reported in previous studies [2, 4, 5]. In addition to a bolus of 50 mg added to the priming solution of the CPB, a bolus of 1 mg/kg was intravenously administered to the patient followed by a continuous infusion of 2.5 mg/kg per hour until approximately 15 minutes before the planned end of CPB. Anticoagulation was regarded as

3 574 KOSTER ET AL Ann Thorac Surg THE CHOOSE-ON TRIAL 2007;83:572 7 adequate if a 2.5-fold or greater prolongation of the baseline value of the institutionally used activated clotting time (ACT) system was achieved. Additional boluses of 0.1 to 0.5 mg/kg could be administered at the discretion of the treating team. Due to the fact that bivalirudin can be eliminated by hemofiltration, the use of ultrafiltrationhemoconcentration was discouraged during CPB. However, modified ultrafiltration was permitted after CPB to facilitate bivalirudin elimination [3]. Due to the unique pharmacology of bivalirudin, stasis in the CPB circuit was avoided-minimized. This was achieved by the following strategies: the use of closed systems whenever possible; the creation of shunting lines from the arterial filter to the cardiotomy reservoir; intermittent compression of the collapse venous reservoir with flushing back in the hard shell cardiotomy reservoir to provide flow of systemic blood in the cardiotomy reservoir and to maintain bivalirudin levels; storage of excessive blood in citrated bags instead of the hard shell cardiotomy reservoir; or processing excessive blood with the use of cell savers. The use of cardiotomy suction was minimized whenever possible and replaced by the use of a cell saver to avoid aspiration and systemic infusion of activated blood from the operative field [4]. After cessation of CPB, the venous line was infused into the patient, the system was refilled with saline, the arterial and venous line reconnected, and circulation of the closed system started to avoid stasis and thrombosis of the system. With the beginning of recirculation, a bolus of 50 mg, followed by a continuous infusion of 50 mg/hour, was added to provide adequate bivalirudin concentrations. When it was deemed that CPB definitely did not have to be reestablished, this volume was processed with a cell saver. Regarding surgical practice, in cases of CABG surgery, assessments of grafts for patency and leakage were performed with saline or, if bivalirudin-containing blood was used, grafts were thereafter flushed with saline and bulldogged while applying pressure on the saline Table 1. Criteria for the Diagnosis of HIT Acute HIT/TS (diagnosed prior to surgery a ) n 43 Positive serologic assay 35 Positive functional assay 7 Only thrombocytopenia ( 50%) with or 1 without thromboembolism History of HITTS (without documentation n 6 of anti-pf4/h antibodies or thrombocytopenia prior to surgery a ) Positive serologic assay 3 Positive functional assay 1 Only thrombocytopenia ( 50%) with or 2 without thromboembolism a Due to the assumed persistence of HIT antibodies of approximately 40 to 100 days, patients were evaluated as having HIT antibodies at the time of surgery if the serologic-clinical diagnosis of HIT was less than 40 days before date of surgery. anti-pf4/h anti-platelet factor 4/heparin; HIT heparin-induced thrombocytopenia; TS thrombosis syndrome. Table 2. Demographic Data, Type, and Duration of Surgery Age: 65 years 21 (42.9%) 65 years 28 (57.1%) Sex: Female 21 (42.9%) Male 28 (57.1%) Weight (kg) Surgery: CABG 22 (44.9%) CABG valve 13 (26.5%) CABG other 2 (4.1%) Valve 9 (18.4%) Valve other 1 (2.0%) Other 2 (4.1%) Repeat 8 (16.3%) Duration of surgery minutes CABG coronary artery bypass grafting. syringe. If a left or right internal thoracic artery was used for grafting the vessel was transected shortly before grafting was performed, in order to avoid stasis and the potential risk of thrombus formation in the graft. Statistical Analysis Statistical analyses were performed on the protocolspecified safety population, which was defined as patients enrolled for administration of bivalirudin who actually received the drug. The 2 tests were used for comparison of categoric variables and t tests for comparison of continuous variables. Subgroup analyses were performed for patients based on the criteria for diagnosis of heparin-induced thrombocytopenia and thrombosis Table 3. Primary and Secondary Endpoints Endpoints n 49 Primary endpoints (day 7/discharge) Death 1 (2.0%) Q-wave MI 0 (0.0%) Revascularization 1 (2.0%) Stroke 1 (2.0%) Secondary endpoints (day 30) Death 3 (6.1%) Q-wave MI 0 (0.0%) Revascularization 1 (2.1%) Stroke 1 (2.1%) Hemorrhage and reexploration 2 (4.1%) Secondary endpoints (12 weeks) Death 4 (8.2%) Q-wave MI 0 (0.0%) Revascularization 1 (2.2%) Stroke 1 (2.2%) Hemorrhage and reexploration 2 (4.1%) MI myocardial infarction.

4 Ann Thorac Surg KOSTER ET AL 2007;83:572 7 THE CHOOSE-ON TRIAL 575 Table 4. Bleeding and Transfusion Data Data All (n 49) Renally Impaired (CrCl ml/min) (n 10) Intraoperative blood loss in ml (mean SD [range]) (50 1,900) ( ) 24-hour blood loss in ml (mean SD [range]) (100 3,200) (370 2,160) Any transfusion at day 7 41 (83.7%) 10 (100.0%) Red blood cells 38 (77.6%) 9 (90.0%) Platelets 17 (34.7%) 4 (40.0%) Fresh frozen plasma 24 (49.0%) 6 (60.0%) Red blood cells (mean SD [range]) (1 16) (1 14) Platelets (mean SD [range]) (1 23) (5 23) Fresh frozen plasma (mean SD [range]) (1 20) (1 20) CrCl creatinine clearance; SD standard deviation. syndrome (serologic versus clinical) and baseline renal function. There were 50 patients prospectively enrolled at 18 centers in the United States and Germany. In 49 of these patients, anticoagulation during CPB was performed with bivalirudin. One patient died prior to surgery. Of the 49 bivalirudin-treated patients, 42 patients were diagnosed shortly before their operation with positive serologic or functional assays. Of the remaining patients, six had a history of HITTS without documented anti- PF4/H antibodies or thrombocytopenia at the time of surgery. In three patients, antibody test results were not available. However, all patients had a decrease of the platelet count of more than 50% from baseline during heparin therapy and one experienced a thromboembolic event (Table 1). Baseline characteristics, type of surgery, and duration of CPB are presented in Table 2. Clinical endpoints are summarized in Table 3. Ten patients had moderately impaired renal function (CrCl 30 to 60 ml/minute). Procedural success was achieved in 46 (94%) patients. At 30 days, procedural success was achieved in 42 (86%) patients, and after 12 weeks in 40 (82%) patients. Explanations for deaths were as follows: an 83-year-old patient due to an irreparable tear to the right atrium intraoperatively; a 79-year-old patient on postoperative day 54 due to septicemia; a 74-year-old patient on postoperative day 26 due to respiratory failure; and a 66-year-old patient on postoperative day 11 because of ventricular arrhythmia. Perioperative blood loss data and transfusion requirements are presented in Table 4. Mean intraoperative blood loss was ml, and mean 24-hour postoperative blood loss was ml. Forty-one (84%) patients received transfusions before day sevendischarge with a mean of units of red blood cells, platelet concentrates, and units of fresh frozen plasma. The remaining parameters revealed comparable results. There were no significant differences regarding bleeding and(or) transfusions between bivalirudin-treated patients with normal renal function compared with bivalirudin-treated patients with impaired renal function. Results Comment The management of HIT patients requiring cardiac surgery with CPB is a challenging situation for the surgical team, as no alternative anticoagulant to heparin is approved for this indication. Additionally, all currently available alternative anticoagulation strategies are associated with significant limitations. These include a prolonged effect due to agent pharmacokinetics as well as difficulty monitoring the anticoagulant effect. Currently used agents and techniques lack favorable safety profiles, especially related to bleeding complications. The CHOOSE-ON study is the first investigation targeting regulatory approval of a drug for anticoagulation during surgery requiring CPB in patients with clinically apparent HIT. The first results of this investigation provide evidence that bivalirudin can be safely and effectively used for this indication. There are three alternative anticoagulation protocols for cardiac surgery with CPB in patients diagnosed with HIT that are currently most often used and they are well-described in recent literature. They are the following: (1) to avoid any heparin exposure and use the direct thrombin inhibitor lepirudin for anticoagulation; (2) to use a combination of heparin with a potent, short-acting antiplatelet agent such as a platelet glycoprotein IIb/IIIa antagonist, or use a prostaglandin to attenuate HITinduced platelet activation and the consequent promotion of procoagulant activity; or (3) to wait until the transient HIT antibodies are no longer detectable, perform CPB with heparin, and establish an alternative anticoagulation in the postoperative period [1, 8 11]. Use of lepirudin during CPB is limited because of the risk of anaphylactic reaction after reexposure to the drug with a high systemic bolus, difficulty in monitoring anticoagulation effect with a point-of-care test since the point-of-care ecarin clotting time assay is no longer commercially available, the relatively long half-life of approximately 60 to 80 minutes, the lack of an antidote, and its exclusive renal elimination. The final limitation is

5 576 KOSTER ET AL Ann Thorac Surg THE CHOOSE-ON TRIAL 2007;83:572 7 associated with significant bleeding complications in patients with renal impairment [10]. Limitations of the use of heparin with a short-acting antiplatelet agent are the potential for severe arterial hypotension with prostaglandin and bleeding complications with tirofiban. The latter is especially dangerous if renal function is markedly impaired [1]. In addition, there is the concern that with recovery of platelet function, circulating anti-pf4/h antibodies may trigger severe thromboembolic complications. Due to the transient nature of anti-pf4/h antibodies, surgery may be delayed until after the disappearance of antibodies and then performed with heparin for anticoagulation. This strategy was first published in 2000, is recommended in the ACCP guidelines, and is widely used [11, 12]. In our patients, this strategy would have been applicable for only six patients. However, the data supporting this strategy in the specialized field of cardiac surgery employing CPB are only based on a small number of case reports. Therefore, we believe that further data from a large controlled trial are needed to confirm this procedure as a safe first-line strategy. However, even if this strategy is the first-line option in patients with a history of HIT who require elective surgery, the need for an alternative anticoagulation still exists, because a large number of procedures are urgent and cannot be postponed for 40 to 100 days until antibodies subside. With the use of bivalirudin, although no reversal agent is available, the short half-life of the agent leads to quick cessation of effect after the drug is discontinued before termination of CPB. Additionally, with bivalirudin the limitations of the other agents outlined above appear to be overcome: the use of heparin is avoided; bivalirudin does not cause hemodynamic side effects or anaphylaxis; elimination is rapid and largely independent of specific organ function; and point-of-care monitoring can be performed with widely available ACT tests [2, 13]. Using the currently described drug dosing protocol, bivalirudin has been successfully employed for anticoagulation during CPB in non-hit patients [3 5]. Many of the same institutions participated in the CHOOSE-ON and evaluation of patients during coronary artery bypass graft operation (EVOLUTION-ON; a safety study which compared bivalirudin anticoagulation and heparin-protamine in non-hit patients) trials, so that despite the relatively small number of patients in these two trials, a comparison is appropriate. The data for primary and secondary endpoints, as well as for bleeding and transfusion requirements in CHOOSE-ON patients, are comparable with the results achieved in the EVOLUTION-ON trial [5]. This is even more noteworthy in light of the generally accepted notion that patients with or at risk for HIT/TS are at higher risk for adverse events. Additionally, in contrast to the experience with the use of lepirudin in patients with renal impairment, moderate impairment of kidney function did not appear to affect the safety profile of bivalirudin, particularly with respect to hemorrhagic complications. This confirms results of previous trials in non-hit patients and reinforces the safety of bivalirudin in cardiac surgery in this higher risk patient population [3]. There are limitations in the design of the trial due to the inherent difficulty in identifying and enrolling patients with HIT, as well as the lack of a suitable approved comparator. The diagnosis of HIT is not always based on the result of a serologic or functional assay but also on observation of a falling platelet count in the setting of heparin with or without concomitant thromboembolism. This reflects the situation of daily practice when HIT may be suspected and specific and rapid assays or time to wait for a result are not viable options. This may prompt the decision to use alternative anticoagulation regimens preemptively [14]. The techniques of cardiac surgery and perfusion are not identical to standard practice when bivalirudin is used as a replacement for heparin. Bivalirudin undergoes proteolytic cleavage by thrombin and bivalirudin levels may diminish in areas of stagnant blood leading to the potential problem of thrombus formation in areas of stasis. Moreover, aspiration of highly activated blood by cardiotomy suction may trigger the systemic coagulationinflammation system [4, 15]. Surgical and perfusion practice must be carefully adjusted to the unique pharmacology of the drug. However, with increasing experience with the use of bivalirudin in the field of cardiac surgery, procedures will be optimized and handling will become more routine. The results of the current investigation, which is to date the first and only large prospective study for the use of an alternative anticoagulant during CPB in HIT patients, expand those of pilot investigations of the use of bivalirudin during CPB, as well as the results of the EVOLUTION-ON trial. Despite the limitations of the current study design and potential special considerations when bivalirudin is used in the setting of CPB, we conclude that bivalirudin is a safe and effective strategy for performing cardiac surgery with CPB in patients with anti-pf4/h antibodies and HITTS. Based on currently available data, we believe that bivalirudin has the potential to become the first-line strategy in HITTS patients requiring the use of an alternative anticoagulant during CPB. This study was supported by The Medicines Company, Parsippany, New Jersey. Drs. Koster, Dyke, Aronson, Mr. McCarthy, and Dr. Spiess have a consultancy agreement with The Medicines Company and from the beginning were closely involved in the development of the entire program. No authors remain undisclosed. The entire group of authors had complete freedom in evaluation of the data and the writing process of this manuscript. This work was attributed to the Department of Anesthesia, Deutsches Herzzentrum Berlin, Germany. We thank Anne Gale, medical editor, for editorial assistance. References 1. Warkentin TE, Greinacher A. Heparin-induced thrombocytopenia and cardiac surgery. Ann Thorac Surg 2003;76: Warkentin TE, Koster A. Bivalirudin: a review. Expert Opin Pharmacother 2005;6:

6 Ann Thorac Surg KOSTER ET AL 2007;83:572 7 THE CHOOSE-ON TRIAL Koster A, Spiess B, Chew DP, et al. Effectiveness of bivalirudin as a replacement for heparin during cardiopulmonary bypass in patients undergoing coronary artery bypass grafting. Am J Cardiol 2004;93: Koster A, Yeter R, Buz S, et al. 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: Dyke CM, Smedira NG, Koster A, et al. A comparison of bivalirudin to heparin with protamine reversal in patients undergoing cardiac surgery with cardiopulmonary bypass: the EVOLUTION-ON study. J Thorac Cardiovasc Surg 2006; 131: Harenberg J, Wang L, Hoffmann U, Huhle G, Feuring M. Improved laboratory confirmation of heparin-induced thrombocytopenia type II. Time course of antibodies and combination of antigen and biologic assay. Am J Clin Pathol 2001;115: Warkentin TE, Kelton JG. Temporal aspects of heparininduced thrombocytopenia. N Engl J Med 2001;344: Koster A, Meyer O, Fischer T, et al. One-year experience with the platelet glycoprotein IIb/IIIa antagonist tirofiban and heparin during cardiopulmonary bypass in patients with heparin-induced thrombocytopenia type II. J Thorac Cardiovasc Surg 2001;122; Palatianos GM, Foroulis CN, Vassili MI, et al. Preoperative detection and management of immune heparin-induced thrombocytopenia in patients undergoing heart surgery with iloprost. J Thorac Cardiovasc Surg 2004;127: Koster A, Hansen R, Kuppe H, Hetzer R, Crystal GJ, Mertzlufft F. Recombinant hirudin as an alternative for anticoagulation during cardiopulmonary bypass in patients with heparin-induced thrombocytopenia type II: a 1-year experience in 57 patients. J Cardiothorac Vasc Anesth 2000;14: Poetzsch B, Klövekorn WP, Madlener K. Use of heparin during cardiopulmonary bypass in patients with a history of heparin-induced thrombocytopenia [letter]. N Engl J Med 2000;343: Warkentin TE, Greinacher A. Heparin-induced thrombocytopenia: recognition, treatment, and prevention. The seventh ACCP conference on antithrombotic and thrombolytic therapy. Chest 2004;126:311S 37S. 13. Zucker ML, Koster A, Prats J, Laduca FM. Sensitivity of a modified ACT test to levels of bivalirudin used during cardiac surgery. J Extra Corpor Technol 2005;37: Dyke CM, Koster A, Veale JJ, Maier GW, McNiff T, Levy JH. Preemptive use of bivalirudin for urgent on-pump coronary artery bypass grafting in patients with potential heparininduced thrombocytopenia. Ann Thorac Surg 2005;80: Veale JJ, McCarthy HM, Palmer G, Dyke CM. Use of bivalirudin as an anticoagulant during cardiopulmonary bypass. J Extra Corpor Technol 2005;37: