Comparison of the effects of aprotinin and tranexamic acid on blood loss and related variables after cardiopulmonary bypass

Transcription

1 Comparison of the effects of aprotinin and tranexamic acid on blood loss and related variables after cardiopulmonary bypass Aprotinin reduces blood loss after cardiopulmonary bypass, but may sensitize recipients and is expensive. Tranexamic acid, a synthetic antifibrinolytic, has less disadvantages, but opinions differ regarding its efficacy. We studied three groups of patients undergoing cardiopulmonary bypass for coronary disease: recipients of aprotinin (total dose 4.2 X 10 6 kallikrein inhibiting units, n = 14), recipients of tranexamic acid (total dose 20 mg/kg body weight, n = 15), and nonmedicated controls (n = 14) during 24 hours after cardiopulmonary bypass. Compared with controls, aprotinin reduced blood loss, the number of patients requiring transfusions, and the mean number of transfused red cell units (all with p < 0.05), whereas the recipients of tranexamic acid did not differ either from aprotinin recipients or from controls. Aprotinin and tranexamic acid both mitigated the early postoperative reduction of adenosine diphosphate-induced platelet aggregation seen in the controls (p < 0.05). Postoperative increases of plasma concentrations of the prothrombin activation fragment F1+2 and the thrombinantithrombin ill complex showed an activation of intravascular coagulation, without any intergroup differences. The balance between concentrations of tissue plasminogen activator and the type 1 plasminogen activator inhibitor disclosed an activation of fibrinolysis, without differences between the groups. The concentrations of D-dimer, a breakdown product of cross-linked fibrin, remained at baseline in the recipients of aprotinin and tranexamic acid but tripled in the controls (p < 0.05). By contrast, the plasma antiplasmin activity was equally depressed in the tranexamic acid and the control groups but decreased less in the recipients of aprotinin (p < 0.05). This discrepancy may reflect the different modes of action of the two agents, which may make aprotinin more efficacious than tranexamic acid in the "nonfibrinolytic" act of protecting platelet function against attack by plasmin during cardiopulmonary bypass. (J THoRAc CARDIOVASC SURG 1994;108: ) Barbara Blauhut, MD,a Wolfgang Harringer, MD,b Peter Bettelheim, MD,c Jan Eva Doran, PhD,d Peter Spath, PhD,d and Per Lundsgaard-Hansen, MD,e Linz, Austria, and Berne, Switzerland Wth few exceptions,i-4 the significant reduction of blood loss and needs for homologous red cell transfusions From the Red Cross Blood Transfusion Service; the Department of Surgery,b and the Central Laboratory,C General Hospital, Linz, Austria; the Central Laboratory of the Swiss Red Cross Blood Transfusion Service,d and the University Department of Experimental Surgery,< Berne, Switzerland. Received for publication Feb. 9,1994. Accepted for publication May 25, Address for reprints: Barbara Blauhut, MD, Blutspendedienst des Roten Kreuzes fur Oberosterreich, Krankenhausstrasse 9, A-4020 Linz, Austria. Copyright 1994 by Mosby-Year Book, Inc /94 $ /1/57783 brought about by aprotinin after coronary bypass proceduress has now been confirmed by numerous studies Aprotinin is a broad-spectrum antiprotease isolated from bovine tissues and has a molecular weight of approximately 6500 d. Anaphylactic reactions possibly associated with an activation of the complement system are rare «1 %) in patients on first-time exposure. 6, 8,13,14 Subsequent exposures may, however, cause severe incidents in 5% to 6% of patients undergoing surgery of the thoracic aorta. IS In many centers, the cost of aprotinin moreover restricts its use to patients at particularly high risks of postoperative bleeding, which means that most patients are denied the potential benefits of a reduced exposure to homologous transfusions. For these reasons, other agents 1083

2 I 08 4 Blauhut et al. The Journal of Thoracic and December 1994 capable of decreasing blood loss and transfusion requirements after cardiopulmonary bypass (CPB) continue to be of interest. One group of such agents includes the synthetic, lowmolecular antifibrinolytic drugs, which seem to carry less risk of sensitization, 16 and which are much less expensive than aprotinin. The first of these compounds was E-aminocaproic acid (EACA), which was tested without 17 or with effect (usually modest)18-20 in patients undergoing CPB. Its successor tranexamic acid (AMCA; trans-4- (aminomethyl)-cyclohexane-carbonic acid) is 5 to 10 times more potent than EACA 1 6 However, positive reports as well as negative findings24 relative to nonmedicated controls have been published. In one study comparing aprotinin and E-aminocaproic acid, both agents were equally effective,25 but the study did not include a nonmedicated control group. We therefore undertook a comparison of the effects of aprotinin, tranexamic acid, and no treatment on blood loss and related variables during the first 24 hours after CPB. Patients and methods The study protocol was approved by the institutional ethics committee, and informed consent of participants was obtained. Criteria for nonadmission to the study were intake of aspirin, other nonsteroidal antirheumatics, or beta-lactam antibiotics; treatment with heparin, fibrinolytic agents, or oral anticoagulants; a condition requiring emergency surgery or reoperation; and liver or kidney disease. Forty-five patients (37 male and 8 female) were allocated at random to three treatment groups intended to comprise 15 patients each: aprotinin (n = 15; two female), tranexamic acid (n = 16; three female) and nonmedicated controls (n = 14; three female). One male patient each in the aprotinin and the tranexamic acid groups was subsequently excluded because of postoperative bleeding of surgical origin necessitating immediate revision. All patients were cared for by the same team, and no other differences existed in their management, which, as a basic principle, was open and determined by the clinical condition of the individual patients. 8 The only therapeutic decision directly pertinent to the studywhether to transfuse homologous red cells-was based on a hematocrit value of less than 30% versus 30% or greater determined after 1 hour in the recovery area of the intensive care unit. CPB. The details of premedication, anesthesia, and CPB including cardioplegia as well as administration of heparin and its neutralization with protamine were as previously specified,8 except that a Bard HF-5000 membrane oxygenator (Bard Inc, Billeria, Mass.) with a priming volume of 560 ml was used in the present study. Aprotinin (Trasylol; Bayer AG, Leverkusen, Germany) was administered as recommended by Royston and associates,s that is a loading dose of 2 X 10 6 kallikrein inhibiting units (KID) plus a maintenance dose of 0.5 X 10 6 KIU /hr until the patient was transferred to the recovery area of the intensive care unit. In addition, 10 6 KIU was added to the oxygenator priming fluid, giving an average total dose of 4.2 X 10 6 KIU as previously reported. 8 Tranexamic acid (Cyklokapron; Kabi AB, Stockholm, Swe- den) was administered intravenously at a dose of 10 mg/kg body weight, beginning 30 minutes before incision of the skin and followed by 1 mg/kg per hour for 10 hours after the beginning of the surgical procedures. 21,22 Measurements of samples for the study were obtained at four times: 1 = baseline, immediately after induction of anesthesia; 2 = end of bypass, before administration of protamine; 3 = 1 hour after the patient's transfer to the intensive care unit; and 4 = next morning in the intensive care unit (approximately 7 AM). Blood loss was quantitated by the volume of drainage fluid accumulating during the first 24 postoperative hours. Laboratory methods. We specify only those methods that are not routine in every clinical chemistry or hematology laboratory. Reference ranges (mean ± 2 standard deviations) in percentage of standard (100%) or as direct units of measurement are given in parentheses and are those stated by the manufacturers of the equipment or test reagents used. Platelet aggregation induced by adenosine diphosphate (ADP)26 was measured with Cluster ADP reagent (Baxter Healthcare Corp., Dade Division, Miami, Fla.) in a final ADP test concentration of 2,umol/L. The change of the transmission of light in percentage of the attainable maximum was read with an Aggrecorder II PA 3220 aggregometer (DIC, Kyoto, Japan). The normal range is 77% to 94%. Fibrinogen (1 to 4 gm/l) was measured according to Clauss.27 Factor V (70% to 130%) was assayed with a thromboplastin reagent (Thromborel; Behringwerke, Marburg, Germany) and factor V-deficient plasma. Chromogenic peptide substrates and test specifications from Kabi AB, M61ndal, Sweden, were used for the following assays: factor VII (CPS S-2765; 80% to 1300/0); factor VIII:c (S-2222; 500/0 to 120%); factor X (S-2337; 60% to 120%); antithrombin III (S-2238; 80% to 120%); Protein C (S-2366; 70% to 140%); plasminogen (S-2251; 75% to 125%), and antiplasmin (S-2251; 80% to 120%). Enzyme-linked immunosorbent assay (ELISA) methods (Asserachrom) from Boehringer, Mannheim, Germany were used to measure von Willebrand Factor-related antigen (60% to 150%), factor IX antigen (60% to 150%), and D-dimers ( <500 ng/mi). Similarly, ELISA methods (Enzygnost) from Behringwerke were used for the assays of prothrombin activation fragment ("Fl+2micro", 0.4 to 1.1 nmol/l) and the thrombin-antithrombin III complex ("TAT micro", 1 to 4.1,ug/L). Finally, ELISA kits (Coaliza) from Chromogenix/ Kabi were used to measure tissue plasminogen activator (tp A) (1 to 12 ng/m!) and the type 1 plasminogen activator inhibitor (10 to 70 ng/mi). Statistical analysis. Differences among the three patient groups with respect to patient characteristics, time on bypass, intervals between measurements and drainage losses, as well as transfusion requirements, were assessed by analysis of variance, with the use of Duncan's multiple range statistic to determine significant differences. A repeated-measures multivariate analysis of variance procedure was used to screen for differences of other parameters between groups and sampling time points, as well as interactions. Differences between groups at specific times were then inspected by a one-way analysis of variance with Duncan's multiple range test. For some variables, the statistical analysis was performed with nonparametric statistic procedures; differences between individual groups were then determined with the Mann-Whitney U test. For those variables where means were calculated, means ± standard error of the mean (SEM) are provided in the text. In all cases, a two-tailed p < 0.05 is considered significant.

3 The Journal of Thoracic and Volume 108, Number 6 Blauhut et al Table I. Patient characteristics, time on bypass, and intervals between measurements in the three treatment groups Aprotinin No. of 14 patients Age (yr) 64.1 ± 2.2 Body size 169 ± 2.4 (cm) Weight (kg) 78.9 ± 2.8 Body surface 1.89 ± 0.04 area (m 2 ) Time on bypass 75.1 ± 5.6 (min) Intervals between measurements/samples (min) 1-2' 141± ± ± 27 Tranexamic acid ± ± ± ± ± ± ± ± 22 Control ± ± ± ± ± ±8 137 ± ± 17 Values are expressed as means ± SEM; no significant differences were found among the groups, Results Patient characteristics. The patient characteristics and the intervals between the measuring or sampling points 1,2, 3, and 4 are specified in Table I. No differences in patient characteristics were found among the three treatment groups. One patient allocated to the aprotinin group died of a massive cerebrovascular insult with cardiac arrest on the third postoperative day. None of the data obtained from this patient during the first 24 hours (which were included in the analysis) were remarkable. Autopsy showed an apical area of moderately recent, left ventricular subendocardial necrosis involving the anterior and, to a lesser extent, posterior wall together with the septum. Parietal thrombi were conspicuous in this area. Embolic occlusions of branches of the carotid artery had caused an extensive, parieto-occipital infarction of the left cerebral hemisphere, with swelling and uncal herniation. All branches of the triple bypass created 3 days earlier were patent and contained no thrombotic material. Blood loss, hematocrit levels, and red cell transfusions. The drainage losses during the first 24 postoperative hours are depicted in Fig. 1. Analysis of variance showed significant differences between the groups (p = 0.03). Drainage in the aprotinin recipients (269 ± 38 ml) and the control group (453 ± 52 ml) was significantly different (p < 0.05 by Duncan's range test), whereas the tranexamic acid group (403 ± 52 ml) was not significantly different from either the aprotinin group or the nonmedicated controls. Blood Loss in mix Hours after CPB ""Q Y APRO TRAN CTRL Fig. 1. Blood losses (ml X 10 2 ) until 24 hours after CPB in the aprotinin (APRO, n = 14), tranexamic acid (TRAN, n = 15) and control (CRTL, n = 14) groups. The means of the three groups are indicated by horizontal bars. The medians of the drainage fluid volumes were 275, 350, and 465 ml in the aprotinin, tranexamic acid, and control groups, respectively. The significance (and nonsignificance) of the differences between the groups assessed by the Mann-Whitney U test were the same as those found for the means. The indication for red cell transfusions during the first 24 hours after surgical procedures was a hematocrit level of less than 30% at time point 3, that is, after 1 hour in the intensive care unit. By this criterion, 19 of 43 (44%) of the patients with a mean hematocrit level of 28.5% ± 0.5% (standard error of the mean) received 1 or more red cell units. The remaining 24 patients with a mean hematocrit level of 33.8% ± 0.5% were not transfused. The proportions of patients with versus those without transfusions in the three treatment groups during the period of study were 3 versus 11 aprotinin recipients, 7 versus 8 recipients of tranexamic acid, and 9 versus 5 in the control group. These proportions differed significantly with p = Expressed differently, the three groups required statistically different, mean numbers of red cell units to be transfused within 24 hours of surgical procedures (analysis ofvariance,p = 0.026). The means of the aprotinin and the control groups (0.36 ± 0.20 and 1.57 ± 0.40 red cell units, respectively) differed with p < 0.05 by Duncan's range test, whereas the tranexamic acid group, with 0.80 ± 0.28 red cell units, did not differ from either aprotinin recipients or controls. The num-

4 Blauhut et al. The Journal of Thoracic and December 1994 Table II. No. of patients in the three treatment groups requiring between zero and four red cell units during the first 24 postoperative hours No. of patients Tranexamic No. of Aprotinin acid Control red cell units (n = 14) (n= 15) (n = 14) Table IV. Mean ADP-induced platelet aggregation in the three treatment groups at time points 1 to 4 Tranexamic Time points Aprotinin acid Control 89.3 ± ± ± ± ± l.l 76.2 ± 3.5t ± ± ± ± ± ± 1.4 Values are expressed as the change of the transmission of light in mean percentage of the attainable maximum (±SEM); normal range is 77% to 94%. 'Time points are described in Table III. tcontrol group differs from aprotinin and tranexamic acid groups with p < :j:control group differs from tranexamic acid group with p < Table III. Mean hematocrit levels in the three treatment groups at time points 1 to 4 Tranexamic Time points Aprotinin acid Control ± ± ± ± ± ± l.l ± O.5t 32.1 ± ± ± ± ± 0.9 Values expressed as mean percentages ± SEM. *1, Immediately after induction of anesthesia (baseline); 2, end of CPB, before administration of protamine; 3, 1 hour after patient's transfer to the recovery area in the intensive care unit: 4, in the intensive care unit, 6 to 7 AM on the morning after the operation. t Aprotinin group differs from tranexamic acid and control groups with p < :j:aprotinin differs from control group with p < bers of patients requiring between 0 and 4 red cell units are summarized in Table II. In Table II, the distributions of patients in the aprotinin and the control groups differ with p = by the Mann-Whitney U test, whereas the recipients of tranexamic acid can not be distinguished from either of the other groups. From Table II, it is also seen that the total numbers of transfused red cell units were 5, 12, and 22, respectively, in the groups receiving aprotinin, tranexamic acid, and no medication. In addition to 3 and 4 red cell units, respectively, two control patients required one pooled, 6-donor platelet concentrate each. The mean hematocrit levels of the treatment groups at the four measurement points are shown in Table III. Platelet function. The platelet counts did not differ among the three treatment groups, the overall means (in 10 9 per liter ± SEM) at the four time points, being 210 ± 8, 143 ± 8, 130 ± 7, and 151 ± 7. The data for the ADP-induced platelet aggregation are summarized in Table IV. The means for the control group were significantly lower than those of the aprotinin and tranexamic acid groups at time points 2 (end of CPB) and 3 (1 hour in the intensive care unit). At time point 3, the difference between the two medicated groups was of borderline significance with p = Plasmatic coagulation system. The overall averages of fibrinogen levels in grams per liter at the time points 1, 2, 3, and 4 were, respectively, 3.29 ± 0.12 (SEM), 1.90 ± 0.08, 2.39 ± 0.12, and 3.36 ± 0.12 (i.e., within the normal range of 1 to 4 gmjl). Similarly, the mean activities of the von Willebrand factor were 106% ± 9%, 72% ± 5%, 107% ± 7%, and 168% ± 10% of standard (i.e., they remained within the reference range (60% to 150%). Factor V, VII, VIllc, IX, and X were not depleted to a critical degree after CPB (data not shown), and none of these variables differed significantly between the three treatment groups. The variables reflecting activation and inhibition of intravascular coagulation are shown in Table V. The plasma concentrations of the prothrombin fragment and the thrombin-antithrombin complex as "molecular markers" of intravascular coagulation peaked at time point 2 (end of CPB) (Le., when the patients were still fully heparinized), and the thrombin-antithrombin complex remained above normal the next morning in the intensive care unit. The coagulation inhibitors antithrombin III and protein C mirrored these profiles, but none of the variables in Table V differed between treatments. Fibrinolytic system. The data reflecting tissue-derived activation and inhibition of the fibrinolytic system are summarized in Table VI. At the end of CPB and after 1 hour in the intensive care unit (time points 2 and 3), the activity increase of the activator tpa and the simultaneous activity decrease of the type 1 plasminogen activator inhibitor shifted the balance expressed by the ratio of type 1 plasminogen activator inhibitorjtpa toward a stimulation of fibrinolysis. At time point 4 (the next morning in the intensive care unit), the baseline situation was essentially restored. At this moment, the activator and inhibitor activities were

5 The Journal of Thoracic and Volume 108, Number 6 Blauhut et al Table V. Variables related to intravascular coagulation at time points 1 to 4 Fl+2 TAT AT 1II Protein C Time points* «1.2 nmolll) «4 JLgIL) (80%-120%) (70%-140%) ± ± ± ± ± ± ± 2 65 ± ± ± ± 2 73 ± ± ± ± 2 82 ± 2 Values are expressed as means ± SEM, with normal ranges in parentheses; no differences were found among the three treatment groups. Fl+lo Prothrombin activation fragment; TAT, thrombin-antithrombin; AT III, antithrombin III. 'Time points are described in Table III. Table VI. Tissue-associated activation or inhibition of the fibrinolytic system for the three treatment groups at time points 1 to 4 t-pa (1-12 nglml) PAI-1 (10-70 nglml) PAI-1It-PA Time Tranexamic Tranexamic Tranexamic points* Aprotinin acid Control Aprotinin acid Control Aprotinin acid. Control I 8.0 ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± 0.8t 10.6 ± ± ± 32.7t ± ± ± ± ± 1.6 Values are expressed as mean ± SEM, with normal ranges in parentheses. PAI-l, Type 1 plasminogen activator inhibitor. 'Time points are described in Table III. t Aprotinin group differs from control group with p < both significantly higher in the recipients of aprotinin than in the controls, but the inhibitor/activator ratio was indistinguishable from those in the other two groups. No other intergroup differences were found, As previously noted,8 plasminogen cannot be assayed in plasma samples containing aprotinin. No differences of plasminogen were found between the tranexamic acid and the control group, the means of both groups at time points 1 t04, respectively, being 125% ± 2%,71% ± 3%, 88% ± 4%, and 95% ± 3% (reference range 75% to 125%) of standard. The variables illustrating the effects of the tissue factors plus the antifibrinolytic agents administered in this study are shown in Table VII. Here, a difference between the two antifibrinolytic agents was apparent. As judged by the plasma concentrations of D-dimers, that is, the breakdown products of cross-linked fibrin, aprotinin and tranexamic acid inhibited fibrinolysis markedly and to a virtually identical degree as compared with the controls. By contrast, the antiplasmin activities decreased to essentially the same extent in the tranexamic acid and the control groups, remaining significantly higher in the recipients of aprotinin. This finding may be partly due to the antiplasmin activity of aprotinin because the assay used in our study (see Patients and methods) is not specific for the natural inhibitor a2-antiplasmin in the presence of an exogenous antiplasmin. Discussion The dose of aprotinin that we used was the same as that used in the pioneering study,s and our results agreed with those subsequently reported on the effects of high-dose aprotinin therapy during CPB With tranexamic acid, we used the same dose as that recommended by Horrow and associates 21,22-a loading dose of 10 mg/kg body weight after induction of anesthesia but preceding incision of the skin, plus 1 mg/kg per hour for 10 hours-yet we failed to reproduce the positive results communicated by that author. The question arises if this discrepancy may be due to an insufficient number of patients in our study. Power analysis shows that to have an 80% chance of attaining a significance level of p = 0.05 for the difference of the observed means of the tranexamic acid and the control group, as depicted in Fig. 1, a study would need to comprise 180 patients in each treatment arm. In other words, our data on postoperative blood loss are suggestive but not strong enough to prove definitively that tranexamic acid is without effects. In any event, tranexamic acid is less effective if its administration starts only after the neutralization ofhep-

6 Blauhut et al. The Journal of Thoracic and December 1994 Table VII. Effects oj activated or inhibited fibrinolysis: plasma concentration oj D-dimer and activity oj antiplasminjor the three treatment groups at time points 1 to 4 D-dimer «500 ngfml) Antiplasmin (80 % -120 % ) Time Tranexamic Tranexamic points* Aprotinin acid Control Aprotinin acid Control ± ± ± ± ± ± ± ± ± 325t 106 ± ± 3 70 ± ± ± ± 303t 113 ± ± 3 88 ± ± ± ± 266t 136 ± ± ± 3 Values are expressed as mean ± SEM, with normal ranges in parentheses. 'Time points are described in Table III. tcontrol group differs from aprotinin and tranexamic acid groups with p < :j:aprotinin group differs from tranexamic acid and control groups with p < Control group differs from aprotinin group with p < arin by protamine. 23 In this case, a single intravenous dose of 40 mg/kg body weight did not sustain postoperative hemoglobin levels any better than a consistently implemented, non pharmacologic blood conservation protocol. 24 In the study comparing high-dose aprotinin with E-aminocaproic acid,25 neither postoperative blood losses nor transfusion requirements differed significantly between the groups. Without inclusion of a nonmedicated control group, the authors concluded that both agents were equally effective. Altogether, however, the effects of tranexamic acid on postoperative blood loss and transfusion requirements after CPB seem to be less clear-cut than those of aprotinin. The data summarized in Table IV and showingthatthe impact of CPB on platelet function 28,29 is mitigated by aprotinin as well as tranexamic acid essentially agree with published reports. Apart from its effect on ADP = and epinephrine-induced aggregation,2 aprotinin is protective with respect to platelet aggregation on extracellular matrix,30,31 ristocetin-induced agglutination,32 and the membrane glycoproteins Ib and IIa/IIIb. 33 Similarly, tranexamic acid has been reported to mitigate plasmininduced platelet activation and to sustain the ADP content of dense granules,23 whereas E-aminocaproic acid counteracts the reduction of ristocetin-induced agglutination. 34 These positive in vitro findings do not, however, provide a quantitative relationship between platelet function and postoperative blood losses and transfusion requirements, nor do they define the relative in vivo efficacy of aprotinin and tranexamic acid. Nonuniformityof in vitro methods, as well as variations of in vivo perfusion techniques, complicate the interpretation of such in vitro data. In agreement with recent studies,35, 36 the time profiles of the prothrombin activation fragment and the thrombin-antithrombin complex, summarized in Table V, show a significant activation of intravascular coagulation at the end of CPB (time point 2), that is, when the patients were still fully heparinized. Next morning in the intensive care unit (time point 4), this activation had not disappeared completely. The means of prothrombin activation fragment and thrombin-antithrombin complex in Table V are comparable with those reported in conditions as diverse as acute leukemia, fulminant hepatic failure, postoperative thrombosis, shock, sepsis, and investigate infusion of tumor necrosis factor in human beings 37-41; however, the accompanying activities of antithrombin III and protein C do not suggest the presence of disseminated intravascular coagulation in the sense of a consumptive coagulopathy with manifestly abnormal bleeding. Apart from contact activation of the coagulation cascade by foreign surfaces in the extracorporeal circuit, the mean antithrombin activity at time point 2 (57% ± 2%) is sufficiently low to activate intravascular coagulation, as reflected by a significant increase of prothrombin activation fragment levels in individuals with a hereditary deficiency and plasma activities of the inhibitor below 70%.42 At any given time point, the individual thrombin-antithrombin complex concentrations in our patients did not correlate with total antithrombin activities. This finding makes sense when the molar concentrations of prothrombin, antithrombin, and the complex in plasma are considered: thrombin-antithrombin complex concentrations as high as 200 /lg/l are equivalent to no more than 0.1 % to 0.2% of the constituents of the complex, and variations of this magnitude cannot be picked up by total activity assays.43 The concentrations of prothrombin activation fragment and thrombin-antithrombin complex in our patients did not differ among the three treatment groups. We thus cannot confirm the observation 44 that aprotinin lowered thrombin-antithrombin complex concentrations relative

7 The Journal of Thoracic and Volume 108, Number 6 Blauhut et al to untreated controls during and after CPB, but the concentrations were not higher. This finding is of some interest in view of recent concerns that the use of antifibrinolytic agents might promote occlusions of coronary bypass grafts and other vessels of similar or smaller caliber. 45-5o In the patient who died of a cerebral infarction on the third postoperative day (see Patient characteristics), the concentrations of prothrombin activation fragment and thrombin-antithrombin complex were somewhat above average but by no means the highest recorded during the study period. The autopsy findings were equally inconclusive with respect to a possible role of aprotinin in the fatal sequence of thromboembolic events. Our data on the activities of tpa and type 1 plasminogen activator inhibitor, both of which are released by vascular endothelium, were displayed in Table VI. The transient activity increase of tpa in our patients has also been noted by other authors after CPB and especially during the anhepatic stage of orthotopic liver transplantation. JO,51-55 Some authors have found that the use of aprotinin blunted this increase of tpa activity. 51, 52, 54 Our data agree with those of other studies, 12,56 showing no such effects of aprotinin (or tranexamic acid). In fact, the activity of tpa was higher in the aprotinin group than in the control group at time point 4. The time profile of the type 1 plasminogen activator inhibitor mirrored that of tpa, showing a nadir at the end of CPB and a secondary increase comparable with that observed during the post-anhepatic stage of liver transplantation. 53 Like tpa, type 1 plasminogen activator inhibitor was higher in the recipients of aprotinin than in the controls next morning in the intensive care unit, but no other differences were found between the groups. As a result of these variations with time, the mean activity ratio of type 1 plasminogen activator inhibitor ItPA, calculated from individual values and giving an impression of the fibrinolytic inhibitor lactivator balance, pointed to an acceleration of fibrinolysis at time point 2 and 3. The baseline conditions were restored next morning in the intensive care unit. In the balance, however, neither aprotinin nor tranexamic acid modified the release of the tissue factors affecting fibrinolysis as compared with nonmedicated controls. By contrast, the variables shown in Table VII and illustrating the effects of the tissue factors plus the agents administered in this study showed a discrepancy between aprotinin and tranexamic acid. The time profiles of D-dimers in our patients fully agree with published reports on the effects of aprotinin versus nonmedicated controls. 8, 12,33,44,57 As judged by this variable, the antifibrinolytic effect of tranexamic acid is just as good as that of aprotinin. However, the pattern of antiplasmin activ- ities in plasma conveys a different impression, in that the tranexamic acid and the control groups are virtually indistinguishable, whereas antiplasmin activity is significantly better sustained in the recipients of aprotinin. It is tempting to speculate that this discrepancy is related to the different modes of action of these two antifibrinolytic agents. Aprotinin inactivates plasmin by the rapid and virtually irreversible formation of an antiprotease-protease complex. By contrast, tranexamic acid inhibits fibrinolysis specifically by blocking the access of plasmin to its binding sites on fibrinogen and fibrin monomers.16 Before it can be definitively neutralized by the relatively low potential of its natural plasmatic antagonist arantiplasmid, which is consumed in the process, plasmin might still be able to damage the platelet membrane Ib glycoproteins, that is, it might interfere with platelet adhesion, which is probably at.least as important for hemostasis as the platelet aggregation which we were able to measure. 33, A study comparing the interaction of aprotinin and tranexamic acid with the platelet membrane Ib glycoproteins might contribute to the evaluation of the relative hemostatic efficacy of aprotinin and tranexamic acid under the conditions of CPB. In summary, our study reconfirmed the efficacy of aprotinin in reducing blood loss and transfusion requirements after CPB but failed to show a significant effect of tranexamic acid. Both agents protected platelet aggregation induced by ADP; neither modified post-cpb activation of intravascular coagulation or the release of plasminogen activator tpa and its type 1 plasminogen activator inhibitor as compared with non medicated controls. Plasma concentrations of D-dimer, a breakdown product of cross-linked fibrin, remained at baseline in the recipients of both agents but tripled in the controls. By contrast, plasma antiplasmin activities decreased to the same extent in the tranexamic acid and the control groups, remaining significantly higher in the recipients of aprotinino We submit that this discrepancy reflects the different modes of action of the two agents, which may make aprotinin more effective than tranexamic acid in the "nonfibrinolytic" act of protecting the membrane glycoproteins mediating platelet adhesion against an attack by plasmin generated during and after CPB. We sincerely appreciate the expert technical assistance of F. Etzlsdorfer and the preparation of the manuscript by S. Bussard and Y. Mermod. REFERENCES 1. Boldt J, Knothe C, Zickmann B, Wege N, Dapper F, Hempelmann G. Comparison of two aprotinin dosage regimens in pediatric patients having cardiac operations:

8 1 090 Blauhut et al. The Journal of Thoracic and December 1994 influence on platelet function and blood loss. J THORAC CARDIOVASC SURG 1993;105:705-1l. 2. Boldt J, Knothe C, Zickmann B, Bill S, Dapper F, Hempelmann G. Platelet function in cardiac surgery: influence of temperature and aprotinin. Ann Thorac Surg 1993;55: Boldt J, Knothe C, Zickmann B, Wege N, Dapper F, Hempelmann G. Aprotinin in pediatric cardiac operations: platelet function, blood loss, and use of homologous blood. Ann Thorac Surg 1993;55: Hardy JF, Desroches J, Belisle S, Perrault J, Carrier M, Robitaille D. Low-dose aprotinin infusion is not clinically useful to reduce bleeding and transfusion of homologous blood products in high-risk cardiac surgical patients. Can J Anaesth 1993;40:625-3l. 5. Royston D, Bidstrup BP, Taylor KM, Sapsford RN. Effect of aprotinin on need for blood transfusion after repeat open-heart surgery. Lancet 1987;2:1289-9l. 6. Bidstrup BP, Harrison J, Royston D, Taylor KM, Treasure T. Aprotinin therapy in cardiac operations: a report on use in 41 cardiac centers in the United Kingdom. Ann Thorac Surg 1993;55: Bidstrup BP, Royston D, Sapsford RN, Taylor KM. Reduction in blood loss and blood use after cardiopulmonary bypass with high dose aprotinin (Trasylol). J THORAC CARDIOVASC SURG 1989;97: Blauhut B, Gross C, Necek S, Doran le, Spath P, Lundsgaard-Hansen P. Effects of high-dose aprotinin on blood loss, platelet function, fibrinolysis, complement and renal function after cardiopulmonary bypass. J THORAC CAR D10VASC SURG 1991;101: Hardy JF, Desroches J. Review article: natural and synthetic antifibrinolytics in cardiac surgery. Can J Anaesth 1992;39: Hunt BJ. Modifying perioperative blood loss. Blood Reviews 1991;5: Royston D. Review article. High-dose aprotinin therapy: a review of the first five years' experience. J Cardiothor Vasc Anesth 1992;6: van Oeveren W, Jansen NJG, Bidstrup BP, et al. Effects of aprotinin on hemostatic mechanisms during cardiopulmonary bypass. Ann Thorac Surg 1987;44: Bohrer H, Bach A, Fleischer F, Lang J. Adverse haemodynamic effects of high-dose aprotinin in a paediatric cardiac surgical patient. Anaesthesia 1990;45: Wuthrich B, Schmid P, Schmid ER, Tornic M. IgE-mediated anaphylactic reaction to aprotinin during anaesthesia. Lancet 1992;340: Schulze K, Graeter T, Schaps D, Hansen B. Severe anaphylactic shock due to repeated application of aprotinin in patients following intrathoracic aortic replacement. Eur J Cardiothorac Surg 1993;7: Verstraete M. Clinical application of inhibitors of fibrinolysis. Drugs 1985;29:236-6l. 17. Gomes MMR, McGoon DC. Bleeding patterns after open-heart surgery. J THORAC CARDIOVASC SURG 1970;60: MidellAI,GradyLH, BloodwellRD,BeallAC, Yashar JJ, Cooley DA. Epsilon aminocaproic acid for bleeding after cardiopulmonary bypass. Ann Thorac Surg 1971;11: Sterns LP, Lillehei CWo Effect of epsilon aminocaproic acid upon blood loss following open-heart surgery: an analysis of 340 patients. Can J Surg 1967;10: Vander Salm TJ, Ansell JE, Okike ON, et al. The role of epsilon aminocaproic acid in reducing bleeding after cardiac operation: a double-blind randomized study. J THORAC CARDIOVASC SURG 1988;95: l. Horrow JC, Hlavacek J, Strong MD, et al. Prophylactic tranexamic acid decreases bleeding after cardiac operations. J THORAC CARDIOVASC SURG 1990;99: Horrow JC, van Riper DF, Strong MD, Brodsky I, Parmet JL. The hemostatic effects of tranexamic acid and desmopressin during cardiac surgery. Circulation 1991;84: Soslau G, Horrow J, Brodsky I. Effect of tranexamic acid on platelet ADP during extracorporeal circulation. Arrl J HematoI1991;38: <pvrum E, Holen EA, Abdelnoor M, <pystese R, Ringdal ML. Tranexamic acid (Cyclokapron) is not necessary to reduce blood loss after coronary artery bypass operations. J THORAC CARDIOVASC SURG 1993;105: Trinh-Duc P. Wintrebert P, Boulfroy D, Albat B, Thevenet A, Roquefeuil B. Comparaison des effets de l'acide f-aminocaproique et de l'aprotinine sur Ie saignement per- et post-operatoire en chirurgie cardiaque. Ann Chir Thorac Cardiovasc 1992;46: Born GVR. Aggregation of blood platelets by adenosine diphosphate and its reversal. Nature 1962;194: Clauss A. Gerinnungsphysiologische Schnellmethode zur Bestimmung des Fibrinogens. Acta Haemat 1957;17: Edmunds LH Jr., Ellison N, Colman RW, et al. Platelet function during cardiac operation: comparison of membrane and bubble oxygenators. J THORAC CARDIOV ASC SURG 1982;83: McKenna R, Bachmann F, Whittacker B, Gilson JR, Weinberg M Jr. The hemostatic mechanism after openheart surgery. II. Frequency of abnormal platelet functions during and after extracorporeal circulation. J THORAC CARDIOVASC SURG 1975;70: Lavee J, Savion N, Smolinsky A, Goor DA, Mohr R. Platelet protection by aprotinin in cardiopulmonary bypass: electron microscopic study. Ann Thorac Surg 1992;53:477-8l. 3l. Mohr R, Goor DA, Lusky A, Lavee J. Aprotinin prevents cardiopulmonary bypass-induced platelet dysfunction: a scanning electron microscope study. Circulation 1992;86[Suppl]:II Huang H, Ding W, Su Z, Zhang W. Mechanism of the preserving effect of aprotinin on platelet function and its use in cardiac surgery. J THORAC CARDIOV ASC SURG 1993; 106: van Oeveren W, Harder MP, Roozendaal KJ, Eijsman L,

9 The Journal of Thoracic and Volume 108. Number 6 Blauhut et al Wildevuur CRH. Aprotinin protects platelets against the initial effect of cardiopulmonary bypass. J THORAC CAR OIOVASC SURG 1990;99: Adelman B, Michelson AD, Greenberg J, Handin RI. Proteolysis of platelet glycoprotein Ib by plasmin is facilitated by plasmin lysine-binding regions. Blood 1986;68: Boisclair MD, Lane DA, Philippou H, Sheikh S, Hunt B. Thrombin production, inactivation and expression during open heart surgery measured by assays for activation fragments including a new ELISA for prothrombin fragment F Thromb Haemost 1993;70: Brister SA, Ofosu FA, Buchanan MR. Thrombin generation during cardiac surgery: Is heparin the ideal anticoagulant? Thromb Haemost 1993;70: Bauer KA, Rosenberg RD. Thrombin generation in acute promyelocytic leukemia. Blood 1984;64: Bauer KA, ten Cate H, Barzegar S, Spriggs DR, Sherman ML, Rosenberg RD. Tumor necrosis factor infusions have a procoagulant effect on the hemostatic mechanism of humans. Blood 1989;74: J~rgensen LN, Lind B, Hauch 0, Leffers A, Albrecht-Beste E, Konradsen LAG. Thrombin-antithrombin III-complex and fibrin degradation products in plasma: surgery and postoperative deep venous thrombosis. Thromb Res 1990;59: Langley PG, Forbes A, Hughes RD, Williams R. Thrombin-antithrombin III complex in fulminant hepatic failure: evidence for disseminated intravascular coagulation and relationship to outcome. Eur J Clin Invest 1990;20: Okamoto K, Takaki A, Takeda S, Katoh H, Ohsato K. Coagulopathy in disseminated intravascular coagulation due to abdominal sepsis: determination of prothrombin fragment I + 2 and other markers. Haemostasis 1992; 22: Bauer KA, Goodman TL, Kass BL, Rosenberg RD. Elevated factor Xa activity in the blood of asymptomatic patients with congenital antithrombin deficiency. J Clin Invest 1985;76: Hoek JA, Sturk A, ten Cate JW, Lamping RJ, Berends F, Borm JJJ. Laboratory and clinical evaluation of an assay of thrombin-antithrombin III complexes in plasma. Clin Chern 1988;34: Dietrich W, Spannagl M, Jochum M, et al. Influence of high-dose aprotinin treatment on blood loss and coagulation patterns in patients undergoing myocardial revascularization. Anesthesiology 1990;73: Bidstrup BP, Underwood SR, Sapsford RN. Effect of aprotinin (trasylol) on aorta-coronary bypass graft patency. J THORAC CARDIOVASC SURG 1993;105: Cosgrove DM III, Heric B, Lytle BW, et al. Aprotinin therapy for reoperative myocardial revascularization: A placebo-controlled study. Ann Thorac Surg 1992;54: Hardy JF, Belisle S. Aprotinin therapy for reoperative myocardial revascularization: an opposing view. Ann Thorac Surg 1993;56: Jegaden 0, Vedrinne C, Rossi R. Aprotinin does not compromise arterial graft patency in coronary bypass operations [Letters]. J THORAC CAROIOVASC SURG 1993; 106: Saffitz JE, Stahl DJ, Sundt TM, Wareing TH, Kouchoukos NT. Disseminated intravascular coagulation after administration of aprotinin in combination with deep hypothermic circulatory arrest. Am J Cardiol 1993; 72: Sundt TM III, Kouchoukos NT, Saffitz JE, Murphy SF, Wareing TH, Stahl DJ. Renal dysfunction and intravascular coagulation with aprotinin and hypothermic circulatory arrest. Ann Thorac Surg 1993;55: Cottam S, Hunt B, Segal H, Ginsburg R, Potter D. Aprotinin inhibits tissue plasminogen activator-induced fibrinolysis during orthotopic liver transplantation. Transplant Proc 1991;23: Grosse H, Lobbes W, Frambach M, von Broen 0, Ringe B, Barthels M. The use of high dose aprotinin in liver transplantation: the influence on fibrinolysis and blood loss. Thromb Res 1991;63: Himmelreich G, Kierzek B, Neuhaus P, Siamer KJ, Riess H. Fibrinolytic changes and the influence of the early perfusate in orthotopic liver transplantation with intraoperative aprotinin treatment. Transplantation Proc 1991; 23: Hunt BJ, Cottam S, Segal H, Ginsburg R, Potter D. Inhibition by aprotinin of tpa-mediated fibrinolysis during orthotopic liver transplantation. Lancet 1990;2: van Oeveren W, van Oeveren B, Wildevuur CRH. Anticoagulation policy during use of aprotinin in cardiopulmonary bypass. J THORAC CARDIOVASC SURG 1992;104: Vandenvelde C, Fondu P, Dubois-Primo J. Low-dose aprotinin for reduction of blood loss after cardiopulmonary bypass. Lancet 1991;337: Havel M, Teufelsbauer H, Knobl P, Dalmatiner R, Jaksch P, Zwoifer W. Effect of intraoperative aprotinin administration on postoperative bleeding in patients undergoing cardiopulmonary bypass operation. J THORAC CARDlO VASC SURG 1991;101 : Edmunds LH. Blood platelets and bypass [Letter]. J THO RAC CARDIOVASC SURG 1989;97: Wenger RK, Lukasiewicz H, Mikuta BS, Niewiarowski S, Edmunds LH. Loss of platelet fibrinogen receptors during clinical cardiopulmonary bypass. J THORAC CARDIOV ASC SURG 1989;97: Zilla P, Fasol R, Groscurth P, Klepetko W, Reichenspurner H, Wolner E. Blood platelets in cardiopulmonary bypass operations: recovery occurs after initial stimulation, rather than continual activation. J THORAC CARDIOVASC SURG 1989;97: