HEPARIN-INDUCED thrombocytopenia (HIT) is an immunologically-mediated reaction to unfractionated heparin that can result in life-threatening thrombotic complications. When patients with a history of HIT present for cardiac surgery, anticoagulation for cardiopulmonary bypass (CPB) can be challenging. Recombinant hirudin (r-hirudin) has been reported as a feasible alternative for anticoagulation. 1–8R-hirudin is a potent direct thrombin inhibitor that works independently of antithrombin-III, offers a rapid onset of anticoagulation, and has a relatively short half-life (60–90 min) in patients with normal renal function. 9
The ability to rapidly and reliably monitor intraoperative levels of r-hirudin anticoagulation, however, has been problematic. The levels of r-hirudin recommended for anticoagulation during CPB (3.5–4.5 μg/ml) 10exceed those that can be effectively monitored with either the partial thromboplastin time or activated clotting time. 11The authors described a plasma-modified kaolin-activated clotting time (PM-ACT) assay as a new point-of-care monitoring tool for r-hirudin anticoagulation has been previously described. 12This assay, accomplished by mixing patient whole blood samples with an equal volume of citrated commercial normal plasma, extends the ACT monitoring range by correcting for hemodilution of coagulation factors and diluting r-hirudin plasma concentrations. The authors present two cases where the PM-ACT was used to monitor r-hirudin anticoagulation during cardiac surgery, and describe the potential for significant perioperative bleeding complications.
Case Reports
Case 1
An 80-yr-old, 58-kg woman presented for repeat coronary artery bypass graft surgery and closure of patent foramen ovale. After her coronary artery bypass grafting 3 yr earlier, HIT was suspected as the cause of two bypass graft occlusions on postoperative day 6 and subsequent inferior vena cava thrombosis, pulmonary embolism, and paradoxical cerebral embolism. She was treated initially with low molecular weight heparin and then, when platelet count continued to remain low, with danaparoid. Heparin- and low molecular weight heparin-reactive platelet antibodies were detected by an ELISA assay. Preoperatively, she presented with an estimated ejection fraction of 25%, atrial fibrillation, and adult-onset diabetes mellitus but with normal renal function (creatinine 0.9 mg/dl).
Upon presentation, the patient no longer had detectable HIT-antibodies by ELISA. Given her previously significant thrombotic complications, the decision was made by consensus of the hematology, anesthesiology, and surgical services to avoid heparin administration. R-hirudin was selected as the alternative anticoagulant for CPB.
On the day of surgery, intravenous and arterial lines were placed, along with a nonheparin-coated pulmonary artery catheter, using saline flushes only. Anesthesia was fentanyl-based, with midazolam, isoflurane, and pancuronium supplementation. Aprotinin was administered using a weight-adjusted modification of the full-Hammersmith regimen, based on evidence that weight adjustment results in less variability in plasma aprotinin concentrations. 13The weight-adjusted full-dose Hammersmith regimen consisted of 2 million units (or 280 mg) for a patient with an ideal body weight of 80 kg, and for patients whose ideal body weight was more or less than 80 kg, the dose was increased or decreased using the following formula: 280 mg × ideal body weight in kilograms (kg) per 80 kg for bolus and pump prime, and 70 mg × ideal body weight in kg per 80 kg. Adjunctive platelet inhibition therapy was employed with oral aspirin and dipyramidole preoperatively and prostaglandin E1infusion (0.01–1.0 μg·kg1·min1) intraoperatively. The r-hirudin anticoagulation strategy was adapted from previously published recommendations and consisted of 0.25 mg/kg intravenous bolus plus 0.25 mg/kg to CPB circuit prime followed by 0.5 mg/min intravenous infusion to continue until approximately 30 min before discontinuing CPB. 10Using the previously established in vitro relationship between PM-ACT and hirudin concentrations (fig. 1), 12a target PM-ACT value of 212 s would correspond to a r-hirudin concentration of 4 μg/ml. Therefore, in following the patient's PM-ACT every 30 min, an additional four 5-mg boluses were given during CPB to maintain this PM-ACT minimum of 212 s. The r-hirudin infusion lasted 2.5 h. In an effort to later validate the PM-ACT in vitro relationship shown in figure 1, additional patient whole blood samples were obtained with each PM-ACT sample for measurements of ex-vivo r-hirudin concentrations using a chromogenic method (fig. 2). Time on CPB was 3 h, 32 min with an aortic cross-clamp time of 1 h, 56 min. Eight minutes after initial CPB discontinuation, CPB was urgently reinitiated for significant hemodynamic instability, and an additional 5 mg r-hirudin was given intravenously to ensure adequate anticoagulation. An intraaortic balloon pump was placed, and the patient was weaned from CPB with significant inotropic support. Major bleeding further complicated the post-CPB period. Initially managed by retransfusing salvaged blood via the CPB circuit, this was discontinued when right atrial thrombus was detected via transesophageal echocardiography, necessitating right atrial thrombectomy. In addition to the 8 units packed red blood cells and 3 units of fresh frozen plasma that were given during CPB, the patient was transfused with 5 units packed red blood cells, 10 units of fresh frozen plasma, 3 six-packs of platelets, and 20 units of cryoprecipitate plus 24 μg l-deamino-8D-arginine vasopressin (DDAVP). Subsequent laboratory evaluation showed a hematocrit 24.1%, platelet count 281,000/ml, International Normalized Ratio 1.4, and activated partial thromboplastin time 117.4 s. Good urine output was maintained throughout the intraoperative course (average 4.5 ml ·kg−1·h−1). The patient was taken directly to the intensive care unit with the sternum open and sterilely covered.
The massive blood loss continued in the immediate postoperative period (fig. 3). Within the first 6 h, the patient lost approximately 18 l of blood through her chest tubes. This blood loss required the transfusion of an additional 18 units packed red blood cells, 11 units fresh frozen plasma, 8 six-packs platelets, and 30 units cryoprecipitate. The international normalized ratio rose as high as 3.01 and activated partial thromboplastin time > 150. Because of persistent severe coagulopathy, Factor IX concentrate (6,068 international units) was administered 4 h postoperatively (chosen primarily because it was the only factor concentrate immediately available) resulting in a considerable decrease in bleeding (fig. 3). Two hours later, the following laboratory values were noted: hematocrit 28%; platelet count 158,000; international normalized ratio 1.29; activated partial thromboplastin time 52.6; Creatinine 1.4 mg/dl. The patient remained in the intensive care unit for approximately 3 weeks and was discharged from the hospital on postoperative day 29.
Case 2
A 58-yr-old, 100-kg man was admitted for a left carotid endarterectomy and single-vessel coronary artery bypass grafting without CPB. His past medical history was significant for left ventricular ejection fraction 23%, atrial fibrillation, adult-onset diabetes mellitus, chronic renal insufficiency (Cr 1.7 mg/dl), and history of right femoral-popliteal bypass 15 yr earlier complicated by HIT during unfractionated heparin treatment of postoperative deep venous thrombosis and pulmonary embolism. During this episode, he suffered three cerebral vascular embolic infarcts; diagnosis of HIT was confirmed with serotonin-release assay.
Upon presentation, his HIT-antibodies were negative by ELISA. Given his previously significant thrombotic complications, the decision was made by consensus of the hematology, anesthesiology, and surgical services to avoid heparin administration. R-hirudin was selected as the alternative anticoagulant.
On the day of surgery, intravenous and arterial lines were placed, along with a nonheparin-coated pulmonary artery catheter, using saline flushes only. Anesthesia was fentanyl-based, with midazolam, isoflurane, and rocuronium supplementation. Aprotinin was not administered to this patient. Because of his preexisting renal insufficiency, mannitol, renal-dose dopamine (2 μg·kg−1·min−1) and furosemide (0.5 mg·kg−1·h−1) infusions were administered to maintain adequate clearance of r-hirudin through forced diuresis. Hirudin dosing and adjunctive therapies were identical to those administered to patient 1, except that the infusion was stopped just before completing the last anastamosis (total 2.25 h). Additional 5 mg boluses were given on two occasions after the surgeon noted clots on the surgical field. Therefore, although the target r-hirudin plasma concentration was 3 μg/ml, 10which, based on in vitro data, would correspond to an estimated PM-ACT value of 200 (fig. 1), the additional doses resulted in a PM-ACT value of 225, or a projected concentration of greater than 4 μg/ml. In an effort to later validate the PM-ACT in vitro relationship shown in figure 1, additional patient whole blood samples were obtained with each PM-ACT sample for measurements of ex-vivo r-hirudin concentrations (fig. 2). The total amount of fluid administered was 1 l of 6% hetastarch, 9 l crystalloid, 2 units of packed red blood cells, and 400 ml cell saver blood to correct for intraoperative blood loss and forced diuresis (urine output of 6.5 ml·kg−1·h−1).
Upon arrival to the intensive care unit, laboratory values were: hematocrit 33.3%, platelet count 126,000/ml, prothrombin time 16.2 s, and activated partial thromboplastin time 68.2 s. The patient was extubated in stable condition 5 h later. During the first 12 h postoperatively, the patient received 4 more units of fresh frozen plasma for a transient chest tube output of 300 ml in 1 h; cumulative chest tube output was 1,300 ml over 18 h. He was transferred to the telemetry floor on postoperative day 1 and discharged a few days later.
Discussion
The optimal strategy for managing patients with current or past history of HIT during cardiac surgery is a matter of controversy and continued investigation. This report describes two cardiac surgical scenarios where r-hirudin was used for anticoagulation. They demonstrate that (1) r-hirudin can be used as an alternative to unfractionated heparin, (2) a PM-ACT system can be used to monitor the concentration of anticoagulation, and (3) bleeding may or may not be a serious complication in certain high-risk patients (e.g. , repeat cardiac surgical procedures, patients who require reinitiation of CPB, patients with significant renal insufficiency).
Anticoagulation for cardiac surgery was successfully performed in both patients. The r-hirudin dosing schedule was derived from previously published experiences using r-hirudin for cardiac surgery, in an effort to achieve the recommended plasma concentrations of 4 μg/ml for CPB, 3 μg/ml for coronary artery bypass grafting without CPB, and 2 μg/ml for carotid endarterectomy. 4,10Using previously obtained in vitro correlation between PM-ACT values and ex-vivo r-hirudin concentrations, 12additional doses were administered to these patients in an effort to maintain these plasma concentrations. In contrast to unfractionated heparin, r-hirudin does not inhibit platelet activity, therefore prostaglandin E1,aspirin, and dipyridamole were administered. The right atrial clot formation that occurred in the first patient may have been related to inadequate anticoagulation of the retransfused blood from the CPB circuit. In the second case, the clots noted on the field were found in static pools of blood outside the patient's body; therefore, the clinical significance of this finding is uncertain. Although the clots were not observed after higher concentrations of r-hirudin were achieved with additional doses, this does not necessarily imply that higher concentrations are needed for cardiac surgery without CPB.
Previous studies have described the ecarin clotting time assay as a reliable method for monitoring r-hirudin doses during CPB. 8,11However, the ecarin reagents and test system are currently not approved for clinical use in the United States. PM-ACT system uses a currently approved ACT method (Hepcon instrument, Medtronic Perfusion Systems, Minneapolis, MN). 12By supplementing patient blood specimens with normal plasma in a 1:1 ratio, the PM-ACT corrects for the prolongation in ACT measurements caused by CPB-induced hemodilution of coagulation proteins. 12In addition, plasma supplementation leads to dilution of circulating plasma r-hirudin and platelet concentrations thus extending the ACT monitoring range, reducing the effect of platelets on the ACT, and facilitating the monitoring of higher hirudin concentrations.
These initial cases have confirmed a linear relationship between r-hirudin plasma concentrations and PM-ACT values (fig. 4). However, in at least these two patients, the PM-ACT values seem to underestimate the ex-vivo hirudin plasma concentration as detected by the chromogenic method. This result may be caused by interpatient variability, a limitation of the plasma-supplemented ACT method, or perhaps inaccuracies in measuring hirudin concentrations using the chromogenic method. No standardized method has been currently accepted for measurement of hirudin concentrations, and in the in vitro study the concentrations were calculated by dose and not actually measured. Nevertheless, it seems that one can conclude that a reasonable (r2= 0.74) relationship exists between these two methods, although the relationship (slope and y-intercept) may need adjustment based on this new ex-vivo information. Based on the revised current linear relationship (y-intercept at 4 μg/ml), a PM-ACT of 187 s should be used to maintain hirudin concentraions at 4 μg/ml). Further studies are needed to evaluate the reproducibility of the PM-ACT technique in monitoring r-hirudin concentrations and to determine the optimal therapeutic range of r-hirudin in cardiac surgery patients.
The first patient had significant, life-threatening blood loss in the postoperative period. It is likely that a number of factors contributed to this bleeding, including the repeat operative status and the urgent reinitiation of CPB that resulted in more extensive hemodilution of coagulation factors and platelets. It is unknown whether the residual concentration of r-hirudin contributed to the bleeding. The decline in r-hirudin concentration after cessation of CPB (fig. 2) demonstrates its expectedly short half-life in a patient with normal renal function (approximately 60 min). It is possible, however, that its duration of action, or context sensitive half-time, exceeded the pharmacokinetic half-time and thus contributed to the bleeding. This experience prompted the use of forced diuresis in the second patient who had mild renal insufficiency, on the basis that the risk of bleeding increases as renal function decreases. 9
The severe bleeding of patient 1 must be recognized as a potential problem for all patients when using r-hirudin as an alternative coagulant, given the lack of an antidote for this agent. While minimal blood loss has been described in several relatively small series of cardiac surgical patients treated with r-hirudin during CPB, 4other reports also indicate that blood loss may be considerable. None have described blood loss as severe as that described with our patient. 6,8Other newer therapies, like bivalirudin, which have a shorter biologic half-life (i.e. , 30 min) may be advantageous with respect to minimizing bleeding in this clinical setting. On the other hand, inadequate anticoagulation in cardiac surgical patients could lead to potentially fatal thrombosis. These potential complications highlight the critical importance of point-of-care monitoring, such as with PM-ACT or ecarin clotting time testing, to achieve and maintain therapeutic r-hirudin concentrations and yet avoid overdose.
The authors thank Charles W. Hogue, Jr., M.D. (Associate Professor, Department of Anesthesiology Washington University School of Medicine St. Louis, Missouri), Ralph Damiano, M.D. (Professor and Chief of Cardiothoracic Surgery, Washington University School of Medicine, St. Louis, Missouri), and Hendrick Barner, M.D., (Professor of Cardiothoracic Surgery, Washington University School of Medicine, St. Louis, Missouri).