Does Hyperfibrinolytic Activity Occur During Cardiopulmonary Bypass from Blood Exposed to Pleural Surfaces?

Transcription

1 Original Article Does Hyperfibrinolytic Activity Occr Dring Cardioplmonary Bypass from Blood Exposed to Pleral Srfaces? Michelle L. Pierce, BS; Alfred H. Stammers MSA, CCP; Melinda S. Dickes, BS; Kimberly J. Taft, BS; Daniel J. Beck, BS; Clinton C. Jones, BS Division of Clinical Perfsion Edcation, School of Allied Health Professions, University of Nebraska Medical Center, Omaha, Nebraska Keywords: cardioplmonary bypass, fibrinolysis, mesothelim, pleral serosa Presented at AmSECT Region VI and Region VII Meetings, October, 1997, and the AmSECT 36th International Conference, March 12-14, 1998, Philadelphia, Pennsylvania ABSTRACT Dring cardiac srgery with cardioplmonary bypass (CPB), fibrinolytic activity may be stimlated when blood exposed to pleral srfaces is sctioned into the extracorporeal circit (ECC). The prpose of this stdy was to determine the effect of reinfsed blood exposed to pleral srfaces on systemic fibrinolytic activity. Following Instittional Animal Care Utilization Committee approval, 12 ml of blood was drawn from the femoral artery of 4 pigs and placed in both pleral cavities, where it remained for 12 min dring CPB. After this time, the exposed blood was sctioned back into the ECC. Blood samples were drawn at the following times: 4 min prior to median sternotomy, 3 and 9 min dring CPB, and 3 min post-sction. Tests performed on the samples inclded thromboelastography (TEG), D-dimer (DD), fibrin degradation prodcts (FDP), fibrinogen concentration, activated clotting time (ACT), hematocrit, and platelet cont. TEG index decreased significantly in the circit following sction (5.28 ±.45 vs..98 ± 1.86, p <.7), while fibrinolytic activity increased (6.25 ± 1.5%) in the ECC when compared to pleral blood (2.17 ± 1.4%, p <.1). The DD and FDP were both elevated in the systemic circlation following sction of the pleral blood, althogh statistical significance was not achieved. The ACT was significantly elevated in the pleral flid dring CPB (77 ± 213) compared with the ECC (378 ± 32, p <.3), which may indicate an accelerated consmption of coaglation factors. In conclsion, blood exposed to pleral srfaces may have increased fibrinolytic activity, bt systemic hyperfibrinolysis was not seen. Address correspondence to: Michelle L. Pierce Division of Clinical Perfsion Edcation University of Nebraska Medical Center 6 Soth 42nd Street Omaha, NE

2 INTRODUCTION Postoperative hemorrhage following cardiac srgery with cardioplmonary bypass (CPB) contines to be one of the most important contribtors to patient morbidity and mortality despite increased efforts to redce hemostatic alterations and blood loss. It is well docmented that CPB impairs hemostatic mechanisms (l) and stimlates the fibrinolytic system (2), leading to inadeqate clot formation and ncontrollable bleeding reqiring homologos blood transfsion (3). Crrent methods sed by clinicians to redce postoperative bleeding and blood transfsion inclde heparin-coated circitry (4), anti-fibrinolytic pharmacological agents (5), atologos plasmapheresis (6,7), and atotransfsion with or withot cell-saving techniqes (8). Intraoperative whole blood atotransfsion is commonly sed in cardiac srgery, as blood collected in the thoracic cavity is sctioned into the extracorporeal circit (ECC) in an effort to redce blood loss dring the procedre. The pericardia! and pleral srfaces contain fibrinolytic activators that indce fibrinolysis in blood that is collected in these cavities (9-12). The introdction of these activators and by-prodcts of fibrinolysis into the systemic circlation may have an effect on systemic hyperfibrinolytic activity dring CPB ( 13-17), althogh controversy exists as to the degree of stimlation and its clinical importance (8,18,19). The prpose of this stdy was to determine the contribtion of blood exposed to pleral srfaces to hyperfihrinolytic activity in cardiac srgery. This stdy will provide a fondation on which clinicians can better evalate the sage of intraoperative whole blood atotransfsion as a safe and beneficial means of blood conservation dring cardiac srgery. MATERIALS AND METHODS ANIMAL PROTOCOL This research protocol was approved by the Instittional Animal Care Utilization Committee at the University of Nebraska Medical Center. All animals sed in this stdy received hmane care in compliance with the Gide for the Care and Use of Laboratory Animals as pblished by the National Instittes of Health (NIH Pblication No , revised 1985). A total of 4 porcine models (4-5 kg) were sed as experimental sbjects in this stdy. All procedres were condcted as acte stdies with the animal anesthetized with ketamine (2 mg/kg) and xylazine (2 mg/kg) given intramsclarly. Once adeqate anesthesia was obtained, the animal was intbated with a 6.5 F endotracheal tbe, and ventilated with a tidal volme of 2-3 ml/kg body weight at a rate of 15-2 breaths per minte. Electrocardiogram leads were placed and the heart rate continosly monitored. The femoral artery and vein were cannlated with 14 gage needles for hemodynamic monitoring, medication infsion, and blood sampling. Once these access lines had been established, the ani- mal was given intermittent doses of pentothal to maintain anesthesia. A median sternotomy was performed and the great vessels dissected free in preparation for cannlation. Prior to cannlation, the animal received a bols dose of 3 IU/kg of bovine lng heparin and adeqate anticoaglation was assred by maintaining an activated clotting time (ACT) of greater than 48 sec. One prse string stre was placed in the aorta for arterial cannlation, and a second was placed in the right atrial appendage to prepare for venos retrn. At the termination of each experiment, the animal was ethanized by the concrrent administration of barbitrate and potassim chloride directly into the aortic root. CARDIOPULMONARY BYPASS The CPB circit consisted of a hollow fiber membrane oxygenator", a cardiotomy/venos reservoirb, an arterial line filter', polyvinyl chloride tbing, and a twin head positive displacement roller pmp. The circit was primed with 12 ml of a balanced electrolyte soltion, 5 ml of 8.4% sodim bicarbonate (5 meq), and 3 IU of bovine lng heparin. Initial heparinization was accomplished with 3 IU/kg of bovine lng heparin and spplemented as needed to maintain an ACT of 48 sec or greater. Moderate hypothermic CPB was condcted with the perfsate being allowed to drift downward and then maintained at 32T. Alpha stat blood gas management was employed throghot the CPB procedre, and lab vales were maintained within normal physiologic ranges of ph ( ), pc 2 (35-45 mmhg), p 2 (1-2 mmhg dring CPB), and base excess ( ± 2). Venos satrations were maintained at a vale greater than 75%. The mean arterial blood pressre was maintained between 5 and 8 mmhg with the se of either phenylephrine or isoflrane as necessary. After 12 min of CPB, the sbjects were actively warmed to a core temperatre of 3TC. EXPERIMENTAL PROTOCOL I. Pre-median sternotomy: A 15 ml sample of blood was drawn into a syringe from the femoral artery catheter and injected into the appropriate tbes for laboratory evalation of baseline systemic fibrinolytic activity (description of tests to follow). 2. Post-median sternotomy: Two milliliters of flid was drawn from the pericardia! space and mixed with 6 ml of blood from the femoral artery, and evalated for baseline pericardia! fibrinolytic activity. As a control, 2 ml of normal saline was mixed with 6 ml of blood from the femoral artery and evalated for fibrinolytic activity as well. 3. Pre-cardioplmonary bypass: 12 ml of blood was drawn from the femoral artery into a heparin-coated syringe, and placed in the pleral cavities (6 ml each). The blood was allowed to pool in these cavities for 12 min. a b c Univox Gold, Bentley Laboratories, Baxter Healthcare, Irvine, CA HSVR-CR, Bentley Laboratories, Baxter Healthcare Irvine CA ALF-4, Bentley Laboratories, Baxter Healthcare, line, CA 121

3 4. Cardioplmonary bypass procedre: Two samples of 15 ml each were drawn from the pleral cavities dring the CPB procedre (3 and 9 min), to determine fibrinolytic activity. Two samples of 15 ml each were also drawn from the ECC at these times and evalated for systemic fibrinolytic activity. After 12 min of CPB, the blood from the pleral cavities was retrned to the ECC via cardiotomy sction. After 15 min of CPB (3 min post infsion of pleral blood into the CPB circit), one 15 ml sample of blood was drawn from the circit to determine the extent of systemic fibrinolytic activity. pared to baseline (p <. I), and FIB was significantly higher in the pleral cavity at 3 min as compared with the circit at 3 min (p <.2), and following sction of the pleral blood (p <.2) (Figre 2). The ACT was significantly elevated in the pleral flid after 3 min of CPB compared with the ECC at the same time (p <.3), bt decreased significantly in the pleral cavities at 9 mintes of CPB (p <.2), and in the ECC after sctioning the pleral blood (p <.8) (Figre 3). LABORATORY EVALUATION A panel of 4 tests was sed to determine fibrinolytic activity: fibrinogen concentration (FIB), fibrin degradation prodct concentration (FDP), D-dimer concentration (DD), and thromboelastography (TEG). Platelet cont (PL T) was also tested with the fibrinolytic indicators. Hematocrit (HCT) and ACT were monitored throghot this experiment to ensre adeqate anticoaglation and oxygen carrying capacity (Figre 1). Figre 1: Sampling timeline A poststernotomy B sample a lbaselinel sample b..,... presternotomy c 1 D 3 CPB E 9 F G sction!5 STATISTICAL ANALYSIS Parametric data was analyzed sing oneway analysis of variance. Fisher's least significant difference was sed when significant f ratios were reached. Nonparametric data were analyzed sing the chi sqare test. Statistical significance was accepted at the p <.5 level. All data is presented as mean ± standard deviation. RESULTS There was a significant redction in FIB observed across all sample points when com- A- TEG, Flli, FDP, DD, PLT, ACT, HCT B - sample a: 6cc blood mixed with 2cc saline sample b: 6cc blood mixed with 2cc pericardia! flid C - 6cc heparinized blood placed into each pleral cavity D- pleral: TEG, Flli, FDP, DD, PLT, ACT, HCT circit: TEG, Flli, FDP, DD, PLT, ACT, HCT E- pleral: TEG, ACT, HCT circit: TEG, ACT, HCT F -blood in pleral cavities sctioned into circit G- circit: TEG, Flli, FDP, DD, PLT, ACT, HCT TEG, ACT, HCT TEG,ACT,HCT Figre 2: Fibrinogen concentration Figre 3: Platelet cont 35-3 j 25 * p <. I vs baseline + p <.2 vs pleral 3 mintes 6 j p <.3 vs baseline -l 2!5 1 5 J Baseline Pleral 3 mintes Circit 3 mintes Circit postsction Baseline Plera1 3 mintes Circit 3 mintes Circit postsction 122

4 A significant redction in PL T was observed across all sample points compared with baseline (p <.3) (Figre 4). However, it is important to note that only one data point was obtained from the pleral cavity at 3 min de to clotting of the specimens. The TEG index is a qantitative smmary of all the variables measred by the TEG, with a higher TEG index indicative of increased clotting activity. The TEG index decreased significantly in the ECC from 3 min as compared with the 9 min time point (p <.2), and in the ECC post-sction (p <.7) (Figre 5). Fibrinolytic activity as measred by TEG was significantly higher in the ECC after 3 min of CPB (p <.4) and in the ECC post-sction (p <.1) when compared with pleral blood (Figre 6). The FDP and DD were both elevated in the pleral cavities, althogh statistical significance was not achieved. Both were also elevated in the systemic circlation following sction of the pleral blood althogh the difference was not statistically significant (Figres 7 and 8). There was a significant decrease (p <.3) in HCT pon Figre 4: Activated clotting time 1 * p <.1 vs baseline, FA/Saline, FA!Pencard!al p <.3 VS pleraljq FA = femoral artery.5".5" ;e M 'R -;;; :;;: ti il :;;: "" -;;; ;; i; 1l " 1l '5. ".5" -.2 " ".. :: = il :; :: ;;: "" Figre 5: Thrombelastograph index 9 * p <.9 vs circit post-sction +p <.2 vscircit 9 + Figre 6: Percent fibrinolysis at 3 min as assessed by thrombelastograph p <.4 vs pleral 9 mintes.5 g ] - < ' ' FA = femoral artery ;;; ;;; v "i! ' :J. '2! "i! '5 '2! ' ' ' :: - "" c - g v -;; " 'i3 v E ' v '2! e a; v '2! '5 a - c. - - Figre 7: Fibrin degradation prodct concentration Figre 8: D-dimer concentration >1-<4 -..l e < e Baseline Pleral Circit Circit 3 mintes 3 mintes post-sction..l e > <25 Basehne -- Pleral 3 mintes C1rcU1t 3 mintes - Ctrcmt post-sction 123

5 initiation of CPB. The average amont of blood sctioned into the ECC from the pleral cavities was 35 ± 155 mi. DISCUSSION Postoperative blood loss is a common complication following cardiac srgery with CPB. The se of CPB involves hemodiltion, hypothermia, and activation of the clotting cascade by the ECC. All of these factors indce a host of hemostatic responses that reslt in an increased potential for bleeding postoperatively. These responses inclde, bt certainly are not limited to: diltional coaglopathies, qalitative and/or qantitative platelet and coaglation disorders, and excessive fibrinolytic activity. Activation of the clotting system throgh the intrinsic pathway de to contact activation has been an isse throghot the development of the ECC ( 4 ). However, even with improvements in circit biocompatibility, postoperative blood loss remains a problem (2). Perhaps we need to consider an alternate rote of blood activation dring CPB, sch as excessive fibrinolysis cased by activators other than artificial srfaces. FIBRINOLYSIS There are two primary components of the fibrinolytic pathway, plasminogen and plasmin. Plasminogen, a cleaving enzyme, is the circlating precrsor of plasmin. Once plasminogen is activated, it is cleaved from a single chained molecle to a two-chained molecle called plasmin, a proteolytic enzyme that binds fibrin, reslting in solbilization of clot (7,21 ). As stated previosly, plasminogen mst first be activated in order to form plasmin and initiate the fibrinolytic pathway. Plasminogen is activated by factor Xlla, prekallikrein, high moleclar weight kininogen, exogenos factors, and tisse type plasminogen activator (t-pa) (21). The endothelim prodces t-pa, which is the fibrinolytic activator primarily focsed pon in this stdy, as its activity is abndant in the pericardia! and pleral mesothelial tisses (9-12). The resltant prodcts of fibrinolysis are called fibrin degradation or split prodcts. There are 5 primary degradation prodcts: fragments X, D, E, Y, and DD. Frther tilization of anticoaglation occrs de to the ability of FDP to bind platelet srface membrane receptors, reslting in platelet inhibition (21). FIBRINOLYTIC NATURE OF PERICARDIUM AND PLEURAL SEROSA Many stdies have been performed to determine the hemostatic natre of mesothelial tisses sch as the pericardim and pleral serosa. A nmber of researchers have identified strong clotting and fibrinolytic systems that are activated by tisse factor and t-pa in hman and animal mesothelial tisses (9-13,22). The sample of blood mixed with pericardia! flid in this stdy did not demonstrate greater clotting activity or fibrinolytic activity as measred by TEG when compared with the control sample mixed with saline. There was also no significant difference ob- served between these 2 samples in ACT vales. However, we did note a trend toward increased clotting and fibrinolytic activity from the baseline femoral artery sample to the pericardia! flid sample. Stdies have shown that tisse factor derived from mesothelial srfaces stimlates thrombin generation more strongly throgh the extrinsic clotting pathway than does the ECC throgh the intrinsic pathway (23,24). In or stdy, FIB was significantly greater in the pleral cavities than in the ECC, althogh the difference was probably de to hemodiltion rather than increased clot formation. The blood in the pleral cavities had been drawn directly from the femoral artery prior to the initiation of CPB, therefore the diltional factor was not as great in this sample. The ACT vales were significantly higher in the pleral cavities at 3 min than in the ECC at this time point. There was no significant difference in TEG index vales between the pleral cavities and ECC at the 3 min sample point. However, after 9 min of CPB, the pleral TEG index was significantly greater than in the ECC. Or reslts indicate that the ECC generates clotting activity more so than the pleral cavity immediately pon the initiation of CPB. However, over time, thrombin generation is stimlated to a greater extent in the pleral cavity than in the ECC. A nmber of researchers have demonstrated flminant fibrinolysis in flids of the pericardia! and pleral cavities de to the activity of t-pa and rokinase (11,13,25). Along with the identification ofthese activators, high concentrations of FDP and DD were observed in pleral and mediastinal drainage samples, indicating active fibrinolysis occrring within mesothelial spaces (II). Or reslts also exhibited high concentrations of FDP and DD in the pleral cavities indicating active fibrinolysis. However, there was no statistical significance between the pleral blood and the ECC at the same time period. REINFUSION OF WHOLE BLOOD COLLECTED FROM MESOTHELIAL SPACES The effect of retransfsing whole blood collected from the pleral and mediastinal cavities may impair systemic hemostasis. Tabchi and associates performed a stdy in which it was hypothesized that local fibrinolytic activation by the pericardia! cavity may have conseqences for the systemic activation processes dring CPB. To test their hypothesis, they interrpted blood sction from the pericardia! cavity dring CPB in coronary artery bypass operations. The blood in the pericardia! cavity exhibited levels of thrombin-antithrombin III complex, t-pa antigen, and FDP that were significantly higher than the systemic blood. When the blood was retrned to the systemic circlation an immediate significant increase in these vales was observed ( 13). Similar stdies by other researchers have conclded that reinfsion of highly activated sctioned blood exacerbates wond bleeding, and they noted that high concentrations offdp in this blood may play a role in the observed impaired hemostasis (14,16,17,26-3). In contrast, other stdies have shown that an increased level of fibrinolytic activator and FDP in whole 124

6 blood reinfsions has no clinically significant effect on the stimlation of systemic fibrinolytic activity. These stdies state that atotransfsion of nwashed, filtered blood is a safe and efficacios alternative to homologos blood replacement in these patients (19,31-35). In this stdy, both FDP and DD increased in the ECC after infsion of the pleral blood, althogh not significantly. Fibrinolytic activity as measred by TEG increased slightly in the ECC post-infsion of pleral blood, bt again, statistical significance was not achieved. In smmary, we did not observe significantly greater clotting or fibrinolytic activity in the pericardia! cavity as compared with the systemic circlation prior to CPB. The ECC generated clotting activity moreso than the pleral cavity immediately pon initiation of CPB. However, over time, thrombin generation was stimlated to a greater extent in the pleral cavity than in the ECC. High concentrations of FDP and DD were observed in the pleral cavities, althogh there was no significant difference in the degree offibrinolytic activation when compared to the ECC prior to infsion of the pleral blood. A trend toward decreased fibrinolytic activity was noted in the pleral blood as well as in the ECC over time. After sctioning the pleral blood, both FDP and DD concentrations increased in the ECC, althogh not significantly. The fibrinolytic activity as assessed by TEG did increase in the ECC, bt again, statistical significance was not achieved. There was, however, a significant increase in the ECC as compared with the pleral cavity at 9 min. Perhaps the mere presence of fibrinolytic by-prodcts in the pooled blood from the initial fibrinolytic episode seen at 3 min was sfficient to have an additive effect on the systemic fibrinolysis already occrring in the systemic circlation. Other investigators have felt that impairment of systemic hemostasis only occrs when very large volmes of mediastinal shed blood are reinfsed (15,32,34). Perhaps the small volmes sctioned into the ECC in or stdy were not sfficient to significantly affect the hemostatic mechanisms taking place in the systemic circlation. Another deficiency of this stdy is that we were nable to follow these sbjects throgh a postoperative corse and see if there was in fact a significant clinical effect de to this practice. In conclsion, this stdy fond that blood exposed to pleral srfaces may have increased fibrinolytic activity, bt systemic hyperfibrinolysis was not observed. Clinical stdies which have been performed on this topic have reslted in differing conclsions, and therefore more research in this area is indicated. REFERENCES 1. Royston D. Systemic inflammatory responses to srgery with cardioplmonary bypass. Perfsion. 1996; II: Bick RL, Arbegast N, Crawford L, Holterman M, Adams T, Schmalhorst W. Haemostatic defects indced by cardioplmonary bypass. Vase Srg. 1975; 9: Li B, Belbol A, Larsson S, Roberts D. Factors inflencing haemostasis and blood transfsion in cardiac srgery. Perfsion. 1996; II: Hs LC. Principles of heparin-coating techniqes. Perfsion ;6: Karski JM, Teasdale SJ, Norman PH, Carroll JA, Weisel RD, Glynn MFX. Prevention of postbypass bleeding with tranexamic acid and e-aminocaproic acid. J Cardiothorac Vase Anesth. 1993;7: Stammers AH, Kratz J, Johnson T, Crmbley 1, Merrill 1. Hematological assessment of patients ndergoing plasmapheresis dring cardiac srgery. 1 Extra-Corpor Techno!. 1993; 25: Stammers AH, Rasmssen C, Kratz 1M. The effects of platelet-rich-plasma on post-cardioplmonary bypass fibrinolysis. J Extra-Corpor Techno!. 1993;25: Ley S1. Intraoperative and postoperative blood salvage. AACN Clinical Isses. 1996; 7: Whitaker D, Papadimitrio 1M, Walters MN-I. The mesothelim: its fibrinolytic properties. 1 Path. 1982; 136: Nkere UU, Whawell A, Thompson EM, Thompson JN, Taylor KM. Changes in pericardia! morphology and fibrinolytic activity dring cardioplmonary bypass. 1 Thorac Cardiovasc Srg. 1993; 16: II. Porter 1M, Ball AP, Silver D. Mesothelial fibrinolysis. 1 Thorac Cardiovasc Srg ;62: Van Hinsbergh VWM, Kooistra T, Scheffer MA, van Boeke! 1H, van Mijen GNP. Characterization and fibrinolytic properties of hman omental tisse mesothelial cells. Comparison with endothelial cells. Blood. 199;75: Tabchi N, de Haan 1, Boonstra PW, van Oeveren W. Activation of fibrinolysis in the pericardia! cavity dring cardioplmonary bypass. J Thorac Cardiovasc Srg. 1993; I 6: De Haan J, Boonstra PW, Monnink SH1, Ebels T, van Oeveren W. Retransfsion of sctioned blood dring cardioplmonary bypass impairs hemostasis. Ann Thorac Srg. 1995;59: Okies 1E, Goodnight SH, Litchford B, Connell RS, Starr A. Effects of infsion of cardiotomy sction blood dring extracorporeal circlation for coronary artery bypass srgery. 1 Thorac Cardiovasc Srg. 1977;74: Bartels C, Bechtel M, Winkler C, Horsch S. Qality of retransfsed blood: whole blood verss cell separation (letter to the editor). Ann Thorac Srg. 1996;61 : Bartels C, Matthias Bechtel 1V, Winkler C, Horsch S. Intraoperative atotransfsion in aortic srgery: Comparison of whole blood atotransfsion verss cell separation. 1 Vase Srg. 1996;24: I Bartels C, Claeys L, Ktenidis K, Nigr H, Horsch S. Intraoperative whole blood atotransfsion dring venos thrombectomy. 1 Cardiovasc Srg. 1994;35:

7 19. Oriel K, Shortell CK, Green RM, DeWeese JA. Intraoperative atotransfsion in aortic srgery. J Vase Sr g. 1993;18: Bafreton C, Jansen PGM, Le Besnerais P, eta!. Heparin coating with aprotinin redces blood activation dring coronary artery operations. Ann Thorac Srg. 1997;63: Horrow JC. Management of coaglopathy associated with cardioplmonary bypass. In: Gravlee GP, Davis RF, Utley JR, eds. Cardioplmonary Bypass Principles and Practice. Baltimore: Williams & Wilkins; 1993: Kmar A, Koenig KB, Johnson AR, Idell S. Expression and assembly of procoaglant complexes by hman pleral mesothelial cells. Thromb Haem. 1994;71: Marlar RA, Kleiss AJ, Griffin JH. An alternative extrinsic pathway of hman blood coaglation. Blood. 1982;6: Nemerson Y. Tisse factor and hemostasis. Blood. 1988;71: Idell S, Girard W, Koenig KB, McLarty J, Fair DS. Abnormalities of pathways of fibrin trnover in the hman pleral space. Am Rev Respir Dis. 1991;144: DeHaan J, Schonberger J, Haan J, van Oeveren W, Eijgelaar A. Tisse-type plasminogen activator and fibrin monomers synergistically case platelet dysfnction dring retransfsion of shed blood after cardioplmonary bypass. J Thorac Cardiovasc Srg. 1993;16: Schonberger JPAM, Everts PAM, Bredee JJ, eta!. The effect of postoperative normovolaemic anaemia and atotransfsion on blood saving after internal mammary artery bypass srgery. Perfsion. 1992;7 : Vertrees RA, Conti VR, Lick SD, Zwischenberger JB, McDaniel LB, Shlman G. Adverse effects of postoperative infsion of shed mediastinal blood. Ann Thorac Srg. 1996;62: Wheeler TJ, Tobias JD. Complications of atotransfsion with salvaged blood. J Post Anesth Nrs. 1994;9: Griffith LD, Billman GF, Daily PO, Lane T A. Apparent coaglopathy cased by infsion of shed mediastinal blood and its prevention by washing of the infsate. Ann Tho rae Srg. 1989;47: Stibbe J, Klft C, Brommer EJP, Gomes M, DeJong DS, Nata J. Enhanced fibrinolytic activity dring cardioplmonary bypass in open-heart srgery in man is cased by extrinsic (tisse-type) plasminogen activator. Er J Clin Invest. 1984; 14: Axford TC, Dearani JA, Ragno G, eta!. Safety and therapetic effectiveness of reinfsed shed blood after open heart srgery. Ann Thorac Srg. 1994;57: Hartz RS, Smith JA, Green D. Atotransfsion after cardiac operation. J Thorac Cardiovasc Srg. 1988;96: Kongsgaard UE, Tollofsrd S, Brosstad F, Ovrm E, Bjornska L. Atotransfsion after open heart srgery: characteristics of shed mediastinal blood and its inflence on the plasma proteases in circlating blood. Acta Anaesthesiol Scand. 1991;35: Ward HB, Smith RRA, Landis KP, Nemzek TG, Dalmasso AP, Swaim WR. Prospective, randomized trial of atotransfsion after rotine cardiac operations. Ann Thorac Srg. 1993;56: