Pericardial Blood Activates the Extrinsic Coagulation Pathway During Clinical Cardiopulmonary Bypass
Background Coagulation during cardiopulmonary bypass (CPB) traditionally has been attributed to activation of the contact system of plasma proteins and the intrinsic coagulation pathway by blood contact with negatively charged surfaces not lined by endothelium. Recent studies have focused on the possible role of the extrinsic coagulation pathway during cardiac surgery. We postulated that the wound activates the extrinsic coagulation pathway during CPB by producing procoagulant cells and enzymes that enter the general circulation.
Methods and Results Blood samples taken from 20 consenting patients who had elective cardiac surgery were assayed for peripheral blood mononuclear cell tissue factor (TF) expression, plasma F1.2, and factor VII and VIIa concentrations. Peripheral blood mononuclear cell TF expression increased in the perfusate after the surgical incision and after CPB was started and in monocytes that adhered to the perfusion circuit. TF on circulating monocytes, however, did not continue to rise during CPB. Peripheral blood mononuclear cell TF was elevated in cells isolated directly from blood in the pericardial cavity and was twice that detected in simultaneous samples from the perfusate (P<.05). F1.2 levels were highest in pericardial blood and increased progressively during CPB. Plasma factor VIIa concentrations, corrected for hemodilution, and ratios of factor VIIa to factor VII were highest in pericardial samples (P<.05) and increased progressively during and immediately after CPB. Pericardial biopsies obtained before and after CPB in 7 patients did not show TF expression by mesothelial cells.
Conclusions These data provide direct evidence of TF expression, activation of the extrinsic coagulation pathway, and thrombin formation in the surgical wound. Addition of pericardial blood to the perfusate and expression of TF by both circulating and adherent monocytes strongly promote thrombus formation during open heart surgery.
CPB produces a massive thrombotic stimulus,1 2 which is suppressed only partially by heparin doses that are two to three times those used to treat other clotting disorders. During CPB, the contact system of plasma proteins3 and the intrinsic coagulation pathway are stimulated by contact with large areas of nonendothelialized biomaterials in the extracorporeal circuit. Previously, this contact was considered the primary thrombotic stimulus during CPB4 ; the wound was considered a secondary stimulus. Recent reports have reevaluated the relative roles of the intrinsic and extrinsic coagulation pathways during clinical cardiac surgery5 6 and have provided indirect evidence that the extrinsic coagulation pathway is activated during open heart surgery.
Evidence for stimulation of the extrinsic coagulation pathway during CPB is largely circumstantial. Boisclair and colleagues5 demonstrated a progressive increase in thrombin formation during clinical cardiac surgery with moderate correlation with an increase in a marker of factor IX activation and poor correlation with factor XII activity. Burman et al6 found a sharp increase in thrombin formation in a patient with factor XII deficiency during closure of an atrioseptal defect without a significant change in factor IX activity. Tabuchi and colleagues7 observed increased fibrinolytic activity in blood aspirated from the pericardial cavity during clinical cardiac surgery and suggested that raising pericardial blood heparin concentrations may reduce blood activation during CPB. Kappelmeyer et al8 found that monocytes expressed TF and procoagulant activity between 2 and 4 hours of blood recirculation in an in vitro pump-oxygenator perfusion circuit.
This study was designed to determine whether blood contact with the surgical wound activated monocytes to express TF and stimulated the extrinsic coagulation system to produce thrombin during clinical cardiac surgery.
Over a 7-month period, 20 patients who had elective cardiac surgery with CPB during weekday hours were studied. The study was approved by the Human Research Committee at the Hospital of the University of Pennsylvania.
Patients ranged in age from 22 to 80 years. All patients received forane anesthesia and pancuronium bromide and had Swan-Ganz and arterial catheters placed. After sternotomy, each patient received 300 U/kg bovine lung heparin IV before cannulation. Two patients also received aprotinin. Activated clotting times were measured (Hemochron model 801, International Technidyne Corp) at baseline, after heparinization, and every 30 minutes during CPB. If necessary, additional boluses of heparin were administered to maintain a clotting time >400 seconds. Patients were cooled to nasopharygeal temperatures between 28°C and 34°C. After aortic cross clamping, hearts were arrested with topical cold saline and intermittent, cold blood, potassium cardioplegia given antegrade and retrograde.
Perfusion circuits included a soft-shell reservoir (Medtronic Cardiopulmonary), a centrifugal blood pump (Medtronic Bio-Medicus), and a hollow-fiber membrane oxygenator (Maxima, Medtronic Cardiopulmonary). The pump and table pack included two lines for suction and one for active venting of the left side of the heart (Sarns 3M Health Care). Blood from these sources was returned to the systemic circulation through a filtered cardiotomy reservoir (Gish Biomedical). An autologous cell saver system (Haemonetics) also was used. Blood aspirated by this system was washed with 1 L normal saline before reinfusion as packed red cells.
Five (18-mL) blood samples were obtained from the arterial catheter or oxygenator before incision, 5 minutes after heparin infusion before CPB, 30 to 45 minutes after the start of CPB, 5 minutes after the end of CPB, and 15 minutes after the administration of protamine. An additional blood sample (18 mL) was obtained directly from the pericardial cavity by the surgeon after 30 to 45 minutes of CPB at the same moment that the perfusionist obtained a sample from the oxygenator.
Blood samples were collected in 3.8% sodium citrate (9:1 vol/vol) and spun at 900g for 20 minutes at 23°C to isolate plasma. The pellet was resuspended in an equal volume of 25 mmol/L HEPES in Hanks’ balanced salt solution, pH 7.4 (GIBCO Laboratories) and layered onto a histopaque gradient (Sigma Chemical Co) for mononuclear cell isolation by the technique of Boyum.9 Mononuclear cells isolated from the initial blood sample were maximally stimulated to express TF by incubation with 10 μg/mL LPS from Escherichia coli serotype 026:B6 (LPS, Sigma) in 20% human plasma for 2 hours at 37°C.
Adherent mononuclear cells were isolated from the oxygenator and filter by washing with 500 mL of 10 mmol/L EDTA, 150 mmol/L NaCl, and 10 mmol/L HEPES. Cells were concentrated from the resulting mixture by centrifuge at 900g for 20 minutes. From the resulting pellet, mononuclear cells were isolated by the above procedure.
PCA of lysed mononuclear cells was measured by a one step recalcification time. Cells were lysed by the addition of 15 mmol/L octyl-β-D-glucopyranoside (Calbiochem) and 25 mmol/L HEPES in Hanks’ balanced salt solution at pH 7.4. The resulting suspension (100 μL) was incubated with 100 μL pooled normal plasma (George King, Kansas City, Mo) at 37°C. After 3 minutes, 100 μL of 25 mmol/L CaCl2 was added, and the clotting time was determined by manual detection of clot formation. Each sample was run in duplicate. To quantify TF, serial dilutions of recombinant native human TF (T.S. Edgington, Scripps, La Jolla, Calif) were used to generate a standard curve. PCA was calculated in picograms TF and expressed as percent in the standard curve. Previous studies have shown that PCA measured with this assay during simulated extracorporeal circulation was completely neutralized by anti-human TF antibody.8
Factor VII levels in plasma were measured with a modification of the Quick prothrombin time. To a plastic fibrometer cup (Thomas Scientific), 100 μL sample plasma, previously treated with heparin adsorbent (Sigma), 100 μL rabbit brain thromboplastin (Sigma), and 100 μL factor VII–deficient plasma (George King) were added and incubated for 3 minutes. Clotting was then initiated with 100 μL of 25 mmol/L CaCl2, and the time for clot formation was measured with a Fibrometer (BBL Fibrosystems). A standard curve was obtained by use of pooled normal plasma (George King).
Factor VIIa levels in plasma were measured with the technique of Wildgoose et al.10 Briefly, 50 μL sample plasma, previously treated with heparin adsorbent (Sigma), was incubated at 37°C with 50 μL factor VII–deficient plasma (George King), 100 μL cephalin (Ortho Diagnostics Systems, Inc), and 100 μL soluble TF (10 nmol/L).11 After 5 minutes, clotting was initiated by the addition of 100 μL of 25 mmol/L CaCl2, and the time for clot formation was measured with a Fibrometer (BBL Fibrosystems). A standard curve for each assay was obtained using pooled normal plasma (George King). The peak concentration of factor VIIa in pooled normal plasma was calculated by use of purified factor VIIa– plus factor VII–deficient plasma (supplied by Dr James Morrissey, Oklahoma City, Okla).
F1.2 concentrations were measured by an ELISA from Behring Diagnostics, Inc.12
Fresh pericardial tissue samples (approximately 50×100 mm), one taken just after the chest was opened and a second taken at the end of CPB, were removed from 7 of the 20 patients. Sterile samples were removed, placed in warm saline, rapidly trimmed, and placed in disposable vinyl specimen molds (Cryomold, Miles Inc) with the mesothelial surface facing of the heart up.13 Specimens were covered with OCT embedding medium (Miles Inc) and immersed in liquid nitrogen. The resulting molds were stored at −70°C and shipped in dry ice to Dr Drake.
Immunohistochemistry was performed on cryostat sections (8 μm thick) as described previously.13 Freshly cut sections were fixed in 100% methanol for 2 minutes at −20°C, air-dried, and used immediately or stored at −70°C for use within 3 weeks. Primary anti-TF antibodies were a mixture of two murine monoclonal antibodies (5 μg/mL each) that recognize distinct TF epitopes, one functional and one nonfunctional (TF9-9C3 and TF9-10H10). An irrelevant murine monoclonal antibody of the same IgG class was used as a negative control (TIB-115) at 10 μg/mL. Primary antibodies were incubated with sections overnight at 4°C; bound antibodies were detected with an avidin-biotin-peroxidase technique (Vector ABC Elite system, Vector Inc) according to the manufacturer’s instructions, with 3-amino-5-ethylcarbazole as the chromogen and Mayer’s hematoxylin as a counterstain.
Data are expressed as mean±SEM. Simultaneous samples of pericardial blood and perfusate are compared by the t statistic for paired samples (SYSTAT for Windows, version 5). This statistic also is used to compare samples during and after CPB with the sample taken before CPB and after heparin administration.
Table 1⇓ gives descriptive data for the patients, operations, and perfusion. No patients developed bleeding complications; no patients died.
Table 2⇓ summarizes the measured procoagulant factors (mean±SE), including mononuclear cell TF and plasma levels (corrected for hemodilution) of factor VII, factor VIIa, and the ratios of factor VIIa to VII. Circulating mononuclear cells do not constitutively express TF14 ; we previously measured 13 pg TF per 1×106 intact monocytes and 11 pg TF per 1×106 lysed monocytes8 ; in this study, we measured <25 pg TF per 1×106 lysed monocytes or 0.1% to 0.2% of LPS stimulation (13 300 pg per 1×106 cells) in two patients before incision. There was an initial appearance of mononuclear cell TF expression (Fig 1⇓) as indicated by the presence of TF (10.76±2.77%) on cells isolated after heparin infusion. After 30 to 45 minutes of CPB, TF expression doubled (21.32±3.30%) but did not increase further during CPB (18.84±3.16%) or after protamine infusion (22.18±4.87%). When the 30- and 45-minute simultaneous pericardial and perfusate samples were compared (Fig 1⇓), the percentage of mononuclear cells from pericardial blood that express TF (40.39±6.80%) was twice that of monocytes from the perfusate. Adherent cells obtained from the oxygenator at the end of CPB also showed elevated TF expression (43.42±8.84%).
Plasma factor VII levels (corrected for hemodilution) did not vary significantly during the procedure (Fig 2⇓). In contrast, corrected factor VIIa levels did not increase in the initial perfusate sample but were significantly higher in pericardial blood (0.64±0.10 ng/mL, P=.001), after CPB (P=.041), and after protamine infusion (P=.026) compared with factor VIIa levels after heparin administration (0.38±0.04; Fig 2⇓). A direct comparison of factor VIIa levels in the pericardial cavity versus the simultaneous perfusate sample (Fig 3⇓) showed significantly higher factor VIIa levels in the pericardial blood (P=.003) than in the perfusate (0.39±0.04 ng/mL).
The ratio of factor VIIa to factor VII also increased significantly in the pericardial sample compared with the simultaneous perfusate sample (Fig 3⇑) and in all samples taken after CPB and after protamine infusion compared with the ratio after heparin was given (Table 2⇑).
F1.2, a fragment produced during the conversion of prothrombin to thrombin, increased progressively during CPB but not after protamine. However, the highest concentration of F1.2 was found in pericardial blood samples (Fig 4⇓), and this was significantly greater (P=.001) than F1.2 levels in simultaneously obtained perfusate samples.
Immunohistochemical assessment of early and late operative pericardial biopsy samples from seven patients showed no TF expression by mesothelial cells. Underlying connective tissue elements showed weak to moderate TF expression; adipocytes had consistently positive but weak expression, and perivascular (adventitial) cells and nerves were moderately positive.
The present study demonstrates that activation of the extrinsic coagulation pathway occurs during clinical CPB and that PBMC TF contributes to this process. TF is a cell surface transmembrane protein that, in conjunction with factor VII, causes thrombus formation through the extrinsic coagulation pathway. TF is expressed constitutively on subendothelium and organ capsules and acts as a hemostatic envelope when injury occurs to the physical barrier that normally separates factor VII from TF.14 TF usually is not present on cells in the blood or on cells that come in direct contact with blood. Specific tissues known to constitutively express TF include myocardial myocytes.13
In addition to constitutive TF expression, certain cells such as monocytes and endothelial cells can be induced to express TF in response to cytokines such as interleukin-1 and tumor necrosis factor. Cytokines are short-acting mediators of inflammation released in response to infection or tissue injury. Endotoxin (LPS), found in the cell membranes of gram-negative bacteria and thought to be an important mediator of sepsis, also induces expression of TF in monocytes.15 LPS or complexes of LPS and LPS-binding protein bind to the glycosyl-phosphatidylinositol–anchored or integral membrane forms of CD14 and mediate NF-κB activation,15 a transcriptional factor for TF.
The human gene for TF is found on chromosome 1, and the mature protein consists of 263 residues with a molecular weight of 29 593.14 This protein contains three distinct domains: a 21-residue cytoplasmic domain; a 23-residue hydrophobic, transmembrane domain; and a 219-residue hydrophilic, extracellular domain.14 The protein in its entirety enhances activation of factor VII to VIIa, possibly by factor Xa, and acts as a cofactor for factor VIIa–catalyzed activation of factors X and IX.14
A soluble, mutant form of TF consisting of only the extracellular domain (sTF1-219) has been shown not to support the autoactivation of factor VII11 but to support factors IX and X activation. This reagent is integral in a novel clotting assay for the determination of factor VIIa levels in plasma.10 16
This factor VIIa–specific assay was used in the present study to determine the extent of extrinsic coagulation pathway activation during clinical CPB. We found that both factor VIIa and the ratio of factor VIIa to factor VII were elevated at the end of CPB and after protamine infusion compared with levels after heparin infusion. These data indicate that the extrinsic coagulation pathway was activated and provides an alternative to the intrinsic system for thrombin formation during clinical CPB. The increase in F1.2 indicates that prothrombin has been cleaved by factor Xa to form thrombin and occurs at the same sampling points as increases in TF and factor VIIa. Studies by Boisclair et al5 and Burman et al6 did not show activation of the contact system during CPB. In the Boisclair et al5 study, factor XIIa activity did not correlate with generation of F1.2 during CPB. In the Burman et al6 study, a 12-year-old girl with severe factor XII deficiency who had CPB for the repair of an atrial septal defect generated as much thrombin as normal control subjects; a corresponding rise in activation products of the intrinsic coagulation system was not detected. A comprehensive study of contact, extrinsic, and intrinsic system activation with correlation of factor X and thrombin formation is needed to fully understand the mechanisms of coagulation during clinical CPB.
Our study is the first to provide direct evidence that the extrinsic coagulation pathway is activated to a greater extent in the wound than in the systemic circulation; levels of both factor VIIa and the ratio of factor VIIa to factor VII are significantly greater in the pericardial blood than in the perfusate. Because thrombin formation leads to secondary fibrinolysis, this observation supports the conclusion of a previous study by Tabuchi et al,7 who found that blood collected from the pericardial cavity of patients undergoing CPB had evidence of increased fibrinolysis compared with systemic blood.
Potential sources of TF for activation of the extrinsic coagulation pathway are abundant during clinical CPB and include the chest wound, monocytes adherent to the heart-lung machine, myocardium, and pericardium. The median sternotomy produces tissue damage and causes the release of cytokines. Cytokines are also released as the result of blood contact with the heart-lung machine.17 18 Myocardial myocytes, as previously mentioned, are known to constitutively express TF.13 The pericardial surface is covered with mesothelial cells,19 and until this study, it was not known whether these cells express TF. They do not, but other cells that are in contact with field blood, including adventitia of the great vessels, suffice to activate the extrinsic coagulation pathway.
In the present study, monocytes express TF at a concentration of 10.8% of the maximal levels observed after stimulation by LPS in the samples obtained after heparin infusion before CPB. Freshly isolated monocytes from peripheral blood have no detectable TF activity20 21 ; therefore, our findings suggest that the trauma of surgery, possibly as a result of the release of cytokines, is responsible, at least in part, for the observed increase in monocyte TF expression. However, because the induction of TF in monocytes occurs in parallel with the initial induction of tissue necrosis factor–α and interleukin-1 and because induction is observed within 1 hour, transcriptional activation by a signaling agonist also may contribute. With CPB, the percentage of monocytes that express TF doubles. No further rise in PBMC TF expression occurs, although mononuclear cells eluted from the oxygenator at the end of CPB are strongly positive for TF. In a previous study, Kappelmeyer and colleagues8 observed a 2- to 4-hour delay in circulating monocyte TF expression during in vitro recirculation of heparinized human blood in an extracorporeal perfusion circuit but found maximal expression of TF on adherent monocytes removed after 6 hours of recirculation. In this study of cardiac surgical patients, we found strong expression of TF by adherent monocytes even in patients who had <1 hour of CPB. Thus, during short-term CPB for cardiac surgery, the surgical wound made before cannulation and CPB, monocytes attached to the surfaces of the heart-lung machine, and pericardial blood aspirated from the wound are major stimuli for thrombin formation through the extrinsic coagulation pathway.
Tabuchi et al7 recommended that whenever possible, pericardial blood not be directly added to the circulating perfusate. The need to conserve red cells and minimize the transfusion of blood products has led to the use of blood aspirated from the surgical field.22 In most cardiac procedures, however, blood aspirated from the field can be washed, concentrated, and returned to the perfusate as packed cells except when exceptionally rapid and heavy bleeding is encountered. An alternative solution may be to add TF antibodies or TF pathway inhibitor directly to pericardial blood during operation.
Selected Abbreviations and Acronyms
|PBMC||=||peripheral blood mononuclear cell|
This work was supported by the American Heart Association and NHLBI grants HL-47186 and PO1-HL-16411, NIH, Bethesda, Md. We thank Drs T.S. Edgington for generously supplying recombinant native and soluble TF and for his valuable critique of this manuscript and Yale Nemerson for initially providing soluble TF for the factor VIIa assay. We also thank Michelle Money for her willing help and interest.
- Received August 28, 1995.
- Revision received November 9, 1995.
- Accepted November 15, 1995.
- Copyright © 1996 by American Heart Association
Dela Cadena RA, Wachtfogel YT, Colman RW. Contact activation pathway: inflammation and coagulation. In: Colman RW, Hirsh J, Marder VJ, Salzman EW, eds. Hemostasis and Thrombosis: Basic Principles and Clinical Practice. 3rd ed. Philadelphia, Pa: JB Lippincott Co; 1994:219-240.
Wachtfogel WT, Harpel PC, Edmunds LH Jr, Colman RW. Formation of C1s-C1-inhibitor, and plasmin-α2-plasmin inhibitor complexes during cardiopulmonary bypass. Blood. 1989;73:468-471.
Kappelmeyer J, Bernabei A, Edmunds LH Jr, Edgington TS, Colman RW. Tissue factor is expressed on monocytes during simulated extracorporeal circulation. Circ Res. 1993;72:1075-1081.
Wildgoose P, Nemerson Y, Hansen LL, Nielson FE, Glazer S, Hedner U. Measurement of basal levels of factor VIIa in hemophilia A and B patients. Blood. 1992;80:25-28.
Neuenschwander PF, Morrissey JH. Deletion of the membrane anchoring region of tissue factor abolishes autoactivation of factor VII but not cofactor function. J Biol Chem. 1992;267:14477-14482.
Lee JD, Kravchenko V, Kirkland TN, Han J, Mackman N, Moriarty A, Leturcq D, Tobias PS, Ulevitch RJ. Glycosyl-phosphatidylinositol anchored or integral membrane forms of CD14 mediate identical cellular responses to endotoxin. Proc Natl Acad Sci U S A. 1993;90:9930-9934.
Morrissey JH, Macik BG, Neuenschwander PF, Comp PC. Quantitation of activated factor VII levels in plasma using a tissue factor mutant selectively deficient in promoting factor VII activation. Blood. 1991;81:734-744.
Woodman RC, Harker LA. Bleeding complications associated with cardiopulmonary bypass. Blood. 1990;76:1680-1697.