- Email: [email protected]
Changes in pericardial morphology and fibrinolytic activity during cardiopulmonary bypass
Changes in pericardial morphology and fibrinolytic activity during cardiopulmonary bypass The presence of pericardial adhesions at resternotomy not only increases the operation time but also increases the risk of serious damage to the heart, great vessels, and extracardiac grafts. The reported prevalence of damage is 2 % to 6 %. The fibrinolytic activity of pericardial tissue may be a crucial factor in determining the extent of adhesion formation following primary operation. Ten patients undergoing cardiac operations were studied to assess the plasminogen activating activity of homogenates of pericardial tissue samples. Samples were taken at three times during the operation and the plasminogen activating activity was measured by means of a standard fibrin plate technique. Tissuetype plasminogen activator, urokinase-type plasminogen activator, plasminogen activator inhibitor-I, and plasminogen activator inhibitor-2 were also measured by means of enzyme-linked immunosorbent assays. Compared with its initial levels (median 2.06 IU/cm2, range 1.28 to 6.48 IU/cm2~ the plasminogen activating activity of pericardial biopsy tissue was significantly reduced at 75 minutes (median 0.64 IU/cm2, range 0.12 to 2.44 IU/cm2, p < 0.01) and at 135 minutes (median 1.45 IU/cm2, range 0.12 to 4.39 IU/cm2, p < 0.05). The major plasminogen activator present was tissue-type plasminogen activator. Compared with its initial levels (median 2.34 ng/mI, range 1.03 to 6.42 ng/ml), subsequent tissue-type plasminogen activator values were also significantly reduced at 75 minutes (median 0.83 ng/ml, range 0.75 to 5.13 ng/mI, p < 0.005) and at 135 minutes (median 1.24 ng/ml, range 0.75 to 6.67 ng/mI, p < 0.05). Low levels of urokinase-type plasminogen activator were found in 5 of 10 patients. However, neither plasminogen activator inhibitor-l nor plasminogen activator inhibitor-2 was detected. Examination with a light microscope showed both increasing pericardial mesothelial damage and increasing features of acute inflammatory changes with time. This study shows that plasminogen activating activity is present in pericardial tissue and that tissue-type plasminogen activator is the major plasminogen activator. The observed inflammatory changes and concomitant damage to the pericardial mesothelium, and the significant reductions in pericardial tissue-type plasminogen activator and plasminogen activating activity seen during cardiac operations, may be important factors contributing to the early development of pericardial adhesions. (J THoRAe CARDIDVASe SURG 1993;106:339-45)
U. U. Nkere, FRCSE, S. A. Whawell, PhD, E. M. Thompson, MRCP, MRCPath,
J. N. Thompson, MChir, FRCS, and K. M. Taylor, MD, FRCS, FRCSE, FSA, London, England
"Wh the expanding volume of cardiac surgery throughout the world, reoperations have become more From the Royal Postgraduate Medical School, Hammersmith Hospital, London, England. Received for publication Feb. 5, 1992. Accepted for publication July 27,1992. Address for reprints: Udim U. Nkere, FRCSE, Department of Cardiothoracic Surgery, Second Floor, B Block, Hammersmith Hospital, Du Cane Road, London WI2 OHS, England. Copyright
1993 by Mosby-Year Book, Inc.
0022-5223/93 $1.00 +.10
12/1/41365
common. In some centers they constitute 10% to 20% of the annual caseload.' The presence of pericardial adhesions at resternotomy not only increases the operation time but also increases the risk of life-threatening damage to the heart, great vessels,or extracardiac grafts.? The reported prevalence of damage is 2% to 6%.3 Furthermore, there is evidence that adhesions have a deleterious effect on cardiac function" and that postoperative pericarditis and subsequent adhesion formation may lead to decreased graft patency.f 6 Various strategies have been used to overcome or prevent the problems associated with adhesions, such as modified pericardial closure.I: 8 peri-
339
The Journal of Thoracic arid
340
Nkere et al.
Cardiovascular Surgery August 1993
Tissuetype PA
Plasminogen Activator (PA) Urokinase type PA
Plasrnfnoqen Act ...."
'Ohl,,,,--It
(pAI-1 and PAI-2)
PLASMINOGEN ~ PLASMIN
t. FIBRIN~ '\
FIBRINOGEN
FIBRIN DEGRADATION PRODUCTS
THROMBIN
Fig. 1. Fibrinolytic pathway. cardia I substitutes," dextran pericardial washout, 10 pharmacologic agents, II and the left thoracotomy approach. 12 Nevertheless, although successful in an experimental setting, most of these methods have been disappointing clin-
ically.' Pericardial tissue injury and subsequent inflammation may give rise to an outpouring of fibrin-rich exudate. This fibrin, together with that of spilled pericardial blood, gives rise to fibrinous adhesions. This fibrin may then either undergo organization by fibroblast infiltration and collagen deposition to form dense adhesions or, as a result of fibrinolysis, be resolved. Damage to pericardial tissue and a reduction in the fibrinolytic capacity of pericardial tissue, as reflected in the plasminogen activating activity (PAA), may be crucial in allowing these pericardial adhesions to form. The fibrinolytic system 13 is a complex pathway, simplified in Fig. 1, involving enzymes and their inhibitors that enable precise control of fibrin deposition and removal.!" There are two principal plasminogen activators (PA), tissue-type PA (t-PA), which is present in plasmal" and many tissues and body fluids.!" and urokinase-type PA (u-PA), which is present in urine!" and some tissue mucosae. I 8 These enzymes are inhibited by PA inhibitor-l (PAl-I) and PA inhibitor-2 (PAI-2). Free PA acts on plasminogen to produce the fibrindegrading enzyme plasmin. The aims of this study were to determine whether pericardial tissue possesses PAA and, if so, to identify the principal fibrinolytic activators and inhibitors, to measure quantitative variation in the activity of these enzymes during cardiopulmonary bypass (CPB), and to document simultaneous changes in pericardial morphologic characteristics.
Patients and methods Patients. Ten patients undergoing primary elective cardiac operations were studied. Eight patients had coronary artery bypass grafts, one patient had mitral valvereplacement, and one patient had aortic valvereplacement. There were sevenmen and
three women. The median age was 62.5 years (range 42 to 74 years). Myocardial protection was maintained with cardioplegia (St. Thomas' Hospital solution), moderate systemic hypothermia (28 0 C, nasopharyngeal), and topical hypothermia. The median crossclamp time was 55.5 minutes (range 29 to 90 minutes). The median perfusion time was 99.5 minutes (range 49 to 124 minutes). Pericardial biopsy technique. The pericardium was opened centrally in the usual manner. Pericardial biopsy specimens were taken with a disposable 6 mm diameter biopsy punch (Stiefel Laboratories [U.K.) Ltd., Wooburn Green, United Kingdom). The samples were obtained from the nondiaphragmatic sides of the open pericardium, all within 3 em of its cut edges. Three samples were taken at times 0, 75, and 135 minutes from the time of openingof the pericardium. Two werethen wrapped in aluminum foil and snap frozen before transfer to a storage freezer at -20 0 C. The other sample was kept in 10% buffered formalin for subsequent histologic analysis. Sixty samples were therefore available for biochemical analysis and 30 samples for histologic examination. Preparation of pericardial homogenates. All sampleswere assayed within 2 weeks. The pericardial biopsyspecimens were thawed at room temperature and weighed. Each sample was washed with 0.5 ml rinsing solution (sodium dihydrogen phosphate 5 mrnol/L, sodium chloride 0.15 mol/L, pH 7.4) and placed into a small plastic test tube, on ice, containing I ml homogenizing solution (sodium dihydrogen phosphate 2.5 mmol/L, sodium chloride 0.075 mol/L, 0.25% Triton XlOO [Sigma Chemical Co. Ltd., London, U.K.), pH 7.8). The tissues were then homogenizedwith an Ultra- Turrax (Janke & Kunkel GmbH & Co., KG, Staufen, Germany) homogenizer for 30 seconds.The homogenates were centrifuged at 12,000g for 20 minutes at 4 0 C. Aliquots of 0.25 ml of supernatant werestored at -20 0 C until assay. Assay for P AA. PAA in the pericardial tissue extracts was assayed by a standard fibrin plate technique.19 A layer of fibrin was produced in the base of 5 cm diameter plastic Petri dishes (Sterlin Ltd., London, U.K.) by pouring in a solution containing 0.9 ml buffer (sodium acetate 0.03 mol/L, sodium barbitone 0.03 mol/L, hydrochloric acid 0.02 mol/L, sodium chloride 1.2 mrnol/L, pH 7.4), 3 ml of 1% human fibrinogen (KabiVitrum Ltd., Uxbridge, U.K.), 0.3 ml of human plasminogen (KabiVitrum Ltd., 2.5 units/rnl), and 0.2 ml bovinethrombin (Armour Pharmaceutical Co., Kankakee, Ill., 20 units/rnl). After the fibrinsolution had set, a 6 mm diameter wellwas cut in the center of the plate. The pericardial tissue homogenates were
The Journal of Thoracic and Cardiovascular Surgery Volume 106, Number 2
Nkere et al.
34 I
Fig. 2. Histologic features of parietal pericardium. A, Section through the fibrosal and outer epipericardiallayers shows vascular congestion (VC). B, Section shows margination of neutrophils (M).
brought to room temperature and 20 ILl of each specimen was placed in the central well of a fibrin plate. Standard solutions of human t-PA (Second International Standard, National Institute of Biological Standards and Controls, Mill Hill, U.K.) reconstituted in sterile water were placed on fibrin plates at the time of each batch of assays. All plates were then incubated at 37° C for 24 hours and the diameter of the zone of lysis around the central well was measured directly by placing the plates on a measuring scale. The maximum and minimum diameters were recorded and the mean value used. A standard plot of diameter of fibrin plate lysis against log t-PA was constructed for each assay and the line of best fit was calculated by means of least squares regression analysis. The mean PAA level for the paired samples was calculated. Homogenate supernatant was also applied to fibrin plate that had been previously heated to 80° C for 20 minutes (thereby destroying the plasminogen) to establish whether fibrinolysiswas plasminogen mediated. The fact that lysis did not occur in those plates so treated confirmed that fibrinolysis was mediated through plasminogen.P The PAA was expressed in international units per square centimeter of pericardium. The lower limit of sensitivity of the assay was 0.07 IU j cm-, Enzyme-linked immunosorbent assays. All the homogenates were assayed for the following fibrinolytic components by means of commercially available enzyme-linked immunosorbent assays (Tintylse, Porton International Group, London, U.K.): t-PA, u-PA, PAl-I, and PAI-2. The lower limits of detection were 0.75 ngjml for t-PA, 0.1 ngjml for u-PA, 2.5 ngjml for PAl-I, and 6 ngjml for PAI-2. Histologic evaluation. Formalin-fixed paraffin-embedded
sections of pericardial tissue were stained with hematoxylin and eosin and assessed by a histopathologist who had no knowledge of the patient or of the time of biopsy. The following features were assessed as evidence of progressive inflammatory change:
I. Vascular congestion (Fig. 2, A) 2. Margination of neutrophils (Fig. 2, B) 3. Frank inflammation (cellular infiltration of the connective tissue) In addition, progressive mesothelial cell damage was evidenced by rounding of cells (Fig. 3, A) and then cell loss (Fig. 3, 8). Statistical methods. Friedman's nonparametric two-way analysis of variance was used to investigate whether there was any overall discernible reduction in the PAA and t-PA levels with time. The Wilcoxon matched-pairs test was then used to look more closely at where the discernible reduction lay. The x 2 test was used to analyze the significance of the histologic changes, and Spearman's rank correlation coefficient was used to detect the linear association between PAA and t-PA and between PAA or t-PA and the mean sample weight of the pericardial biopsy specimens.
Results Examination of the data by means of Friedman's nonparametric two-way analysis of variance suggested that there was an overall discernible reduction in the P AA (p = 0.05) and t-PA (p = 0.004) levels with time. This possibility was then investigated more closely, by means
The Journal of Thoracic and Cardiovascular Surgery August 1993
3 4 2 Nkere et al.
Fig. 3. Histologic features of parietal pericardium. A, Section shows rounded mesothelial cells (RC) becoming detached from the fibrosallayer. B, Section shows detached (D) and free-floating cells (FF).
... Spearman rank Corr.Coe1._ 0.73
... I
...
PM (IUterl) 4
P.O.OOOl
... <0.07
.j2:::;::::!~:::;::::::::;;:::;::::;::::;:::::;:::::;::::;:::;::"::1 <0.751
2
3
4
tPA(nglml)
Fig. 4. PAA ofallpericardial biopsy specimens plotted against the corresponding topA level. of the Wilcoxon matched-pairs test, to see where the difference lies. The results are explained in the following paragraphs. For each of the 10 patients, the PAA ofthe pericardial biopsy specimens was measured in all the homogenate samples. Compared with initial levels (median 2.06 IV/cm 2, range 1.28 to 6.48 IV/cm 2) , the PAA of pericardial biopsy specimens was significantly reduced at 75 minutes (median 0.64 IV/cm 2, range 0.12 to 2.44 IV/cm 2, p < 0.01) and at 135 minutes (median 1.45 IV/cm2, range 0.12 to 4.39 IV/cm 2, p < 0.05). There
was no significant difference between the PAA levels at 75 and 135 minutes. The major PA present was t-PA. Compared with the initial levels (median 2.34 ng/rnl, range 1.03 to 6.42 ng/rnl), subsequent t-PA values were also significantly reduced at 75 minutes (median 0.83 ng/rnl, range 0.75to 5.13 ng/rnl, p < 0.01) and at 135 minutes (median 1.24 ng/rnl, range 0.75 to 6.67 ng/rnl.p < 0.05). Again, there was no significantdifferencebetweenthe t-PA levels at 75 and 135 minutes. Low levels of u-PA were detected in 5 of 10 patients (median 0.125 ng/rnl, range 0.1 to OJ ng/rnl, lower limit of detection 0.1 ng/rnl). However, neither PAI-! nor PAI-2 was detected (lower limit of detection ::;2.5 ng/rnl and ::;6 ng/rnl, respectively). There wasa significantpositive correlationbetweenthe PAA and t-PA levels (Fig. 4). Together with the fact that t-PA levels were approximately 10 times higher than u-PA levels, this suggests that t-PA rather than u-PA is responsible for the PAA exhibited by the pericardium. The histologicfindingsof mesothelialdamage and the extent of pericardial inflammation are shown in Figs. 5 and 6, respectively. These showclear evidenceof increasing mesothelial damage and progressive inflammatory changes with time. The extent of mesothelial damage at times 75 and 135 minutes compared with that at time 0 show a significant (p < 0.005) preponderanceof pericar-
The Journal of Thoracic and Cardiovascular Surgery Volume 106, Number 2
Nkere et al.
34 3
Fig. 5. Mesothelial damage with time. Extent of mesothelial damage at times 75 and 135 minutes compared with that at time a shows a significant (p < 0.005) preponderance of tissue either with rounded mesothelial cells or with
mesothelial cell loss.
Fig. 6. Inflammatory changes with time. Pericardial samples at times 75 and 135 minutes compared with those at time a show a significant (p < 0.005) preponderance of tissue with margination and frank inflammation.
dial tissue either with rounded mesothelial cells or with mesothelial cell loss. Similarly, a significant (p < 0.005) preponderance of tissuewith features of margination and frank inflammation, rather than normal or congestedtissue, was noted at times 75 and 135 minutes compared with that at time O. Discussion
This study, in keeping with previous work by Porter, Ball, and Silver" and again by Gervin and associates.F has shown that mesothelial pericardial tissue expresses fibrinolytic activity. Pericardial tissueshould therefore be capable of removing fibrinand hence,at least potentially, beable to prevent or minimize adhesion formation after cardiac operations. The fibrinolytic activityis reflectedby the PAA of pericardial tissue; this activity, as suggested by the positive correlation between PAA and t-PA, appears to be primarily due to the presenceof t-PA rath-
er than u-PA. In comparisonwith initial values,the PAA and the levels of t-PA were significantly lower at later stages of cardiac surgical procedures. The reduction in pericardial fibrinolytic activity,seen during cardiac operations,suggests a reducedcapacity of pericardial tissueto prevent adhesion formation. As appears to be the case in the peritoneum,wherethe reduction in peritoneal PAA is stronglyimplicated as a causal factor in peritoneal adhesionformation,23-25 this reduced pericardial PAA may be a significant factor in the pathophysiology of pericardial adhesions. The hemodilutionthat occurs as a result of the institution of CPB is known to decrease the blood protein concentrations. Hence it could be argued that the observed reductionin PAA and t-PA are, in part, a reflection of this hemodilution. However, it has been observedthat a similar perioperative reduction in PAA occurs in patients undergoingelectivelaparotomy" in whom hemodilution
344
The Journal of Thoracic and Cardiovascular Surgery August 1993
Nkere et al.
Spearman rank Corr. Coet.• 0.'92
I
P.0.30
., .
...,.. ... '
.
10 20 Sampleweight (mg)
30
Fig. 7. PAA ofall pericardialbiopsyspecimens plottedagainst the corresponding mean sample weight.
does not playas major a role. Nevertheless, the role of hemodilution will be addressed in further studies. Although thePAAand t-PA levels at times 75 minutes and 135 minutes were not statistically different, inspection of the data suggest that these values are on the increase at 135 minutes. Furthermore, the statistical difference between the PAA and t-PA levels at times 0 and 75 minutes are greater than at 0 and 135 minutes. All patients were rewarmed and weaned from CPB by 135 minutes; therefore, the apparent recovery in fibrinolytic activity may have some relation to the effects of CPB and temperature. The extent of this recovery and the degree to which it is impaired by the mesothelial damage and the simultaneous inflammatory response has not been answered by this study. PAI-l and PAI-2 were not detected in the pericardial tissue within the duration of CPB, which suggests that they play no part in the observed early decrease in PAA. However, in peritoneal studies." normal peritoneum likewise has no detectable PAl-lor PAI-2 whereas inflamed peritoneum has reduced PAA and detectable PAl-I andPAI-2. It may be that, asin peritoneum, PAI-I and PAI-2 appear in pericardial tissue at a later stage than after the initial insult. The lack of correlation between the PAA, or t-PA levels, and the mean sample weight of the pericardial biopsy specimens (Fig. 7) provides indirect evidence for the site of production of PA. The use of a 6 mm diameter punch biopsy standardizes the area of mesothelium sampled. The different weights of the biopsy specimens thus for the most part reflect the different amounts of underlying tissues sampled. If submesothelial tissues were the major site of plasminogen activity, a positive correlation between PAA and mean sample weight might be expected. The fact that several low-weight samples exhibited high PAA values and vice versa supports the theory that the surface mesothelial layer is responsible for most of the pericardial
fibrinolytic activity.P' However, the precise identification of the exact cellular origin and tissue localization of pericardial PAA requires further study by means of techniques such as in situ messenger ribonucleic acid hybridization. Previous studies not only have shown the presence of PAA in pericardium and other mesothelial structures-": 22 but also have indicated that injury/? (e.g., physical trauma, drying, diathermy, and ischemia) decreases this activity. 30, 31 It is possible that mesothelial damage during surgical procedures, caused by drying or manipulation, may be responsible for the decrease in PAA and t-PA observed in this study. However, it is not yet clear whether the diminished PAA demonstrated in the pericardium is a causative factor in the process of formation of the dense adhesions seen at reoperation. Nevertheless, it has been shown by Dorr and associates 32 and Menzies and Ellis 33 that, after mesothelial injury, enhancing PAA or fibrinolytic activity, in the form of topical t-PA, may prevent adhesion formation in the peritoneum. The present study also confirms that there is an early inflammatory reaction in pericardial tissue in response to the insult of cardiac surgery. The response of mesothelium to injury has been described by previous investigators using bacterial toxins.l" silica,35 and various chemical and mechanical." stimuli. In general, this cellular reaction is characterized by a change in cell shape from a normal flat appearance'? to cuboidal, a separation of neighboring cells, and a shedding or detachment of the mesothelium from the basal lamina.P Further studies are planned to determine the time scale and degree of pericardial tissue regeneration in the postoperative period and to investigate the relation, if any, to recovery of pericardial fibrinolytic activity. Ryan, Grobety, and Majno,29 in assessing both morphologic and fibrinolytic parameters, suggests that there is evidence for some degree of mesothelial recovery over 3 to 21 days after the operation.
I. 2. 3. 4. 5.
REFERENCES Gabbay S. The need for intensive study of pericardialsubstitutionafter openheart surgery[editorial].ASArO Trans 1990;36:789-91. Loop FO, Cosgrove OM, Kramer JR. Late clinical and arteriographic results in 500 coronary artery reoperations. J THORAC CARD10VASC SURG 1981 ;81 :675-84. LoopFO. Catastrophic hemorrhageduring sternalreentry. Ann Thorac Surg 1984;37:271-2. BaileyLL, Li Z-j, Schulz E, Roost H, Yahiku P. A cause of right ventricular dysfunction after cardiac operations. J THORAC CARDIOVASC SURG 1984;87:539-42. Urschel HC Jr, Razzuk MA, Gardner M. Coronaryartery
The Journal of Thoracic and Cardiovascular Surgery Volume 106, Number 2
bypass occlusion secondary to postcardiotomy syndrome. Ann Thorac Surg 1976;22:528-31. 6. Shapira N, Gordon CI, Lemole GM. Occlusion of aortocoronary veingrafts in association with bovinepericardium. Am J Cardiovasc PathoI1989;3:87-90. 7. Zapolanski A, Fishman NH, Bronstein MN, Ellertson DG, O'Connell TJ, Siegel S. Modified pericardial closure to protect cardiovascular structures during sternal reentry. Ann Thorac Surg 1990;50:665-6. 8. Milgalter E, Uretzky G, Siberman S, et al. Pericardial meshing: an effective method for prevention of pericardial adhesions and epicardial reaction after cardiac operations. J THORAC CARDIOVASC SURG 1985;90:281-6. 9. Gabbay S, Guindy AM, Andrews JF, Amato JJ, Seaver P, Khan Y. New outlook on pericardial substitution after open heart operations. Ann Thorac Surg 1989;48:803-12. 10. Reikeras 0, Nordstrand K, Serlie D. Use of dextran to prevent pericardial adhesions caused by maize starch powder. Eur Surg Res 1987;19:62-4. 11. Vander Salm TJ, Okike ON, Marsicano TH, Compton C, Espinoza E. Prevention of postoperative pericardial adhesions. Arch Surg 1986;121:462-7. 12. Gandjbakhch I, Acar C, Cabrol C. Left thoracotomy approach for coronary artery bypass grafting in patients with pericardial adhesions. Ann Thorac Surg 1989;48: 871-3. 13. Paques EP. Recent advances in the biochemistry of the fibrinolytic system. Behring Inst Mitt 1988;82:68-81. 14. Collen D. On the regulation and control of fibrinolysis. Thromb Haemost 1980;43:77-89. 15. Rijken DC, Wijngaards G, Welbergen J. Relationship between t-PA and the activators in blood and vascular wall. Thromb Res 1980;18:815-30. 16. Rijken DC, Wijngaards G, Welbergen J. Immunological characterisation of plasminogen activator activities in human tissues and body fluids. J Lab Clin Med 1981; 94:477-86. 17. Williams JRB. The fibrinolytic activity of urine. Br J Exp PathoI1951;32:530-7. 18. Larsson LI, Skriver L, Nielsen LS, Grondahl-Hansen J, Kristensen P, Dano K. Distribution of urokinase-type plasminogen activator immunoreactivity in the mouse. J Cell Bioi 1984;98:894-903. 19. Astrup T, Mullertz S. The fibrin plate method for estimating fibrinolytic activity. Arch Biochem Biophys 1952; 40:346-51. 20. Lassen M. Heat denaturation of plasminogen in the fibrin plate method. Acta Physiol Scand 1952;27:371-6. 21. Porter JM, Ball AI', Silver D. Mesothelial fibrinolysis. J THORAC CARDIOVASC SURG 1971;62:725-30. 22. Gervin AS, Jacobs G, Hufnagel HV, Mason KG. Surgical trauma and pericardial fibrinolytic activity. Am Surg 1975;41 :225-9.
Nkere et al. 3 4 5
23. Thompson IN, Paterson-Brown S, Harbourne T, Whawell SA, Kalodiki E, Dudley HAF. Reduced human peritoneal plasminogen activating activity: possible mechanism of adhesion formation. Br J Surg 1989;76:382-4. 24. Gervin SA, Puckett CL, Silver D. Serosal hypofibrinolysis: a cause of postoperative adhesions. Am J Surg 1973; 125:80-7. 25. Whitaker D, Papadimitriou JM, Walters MNI. The mesothelium: its fibrinolytic properties. J Pathol 1982; 136:291-9. 26. Scott-Coombes DM, Whawell SA, Thompson IN. Per-operative changes in peritoneal fibrinolytic activity. Fibrinolysis 1992;6(suppl 2):61. 27. Vipond MN, Whawell SA, Thompson IN, Dudley HAF. Peritoneal fibrinolytic activity and intraabdominal adhesions. Lancet 1990;335:1120-2. 28. Merlo G, Fausone G, Barbero C, Castagna B. Fibrinolytic activity of the human peritoneum. Eur Surg Res 1980; 12:433-8. 29. Ryan GB, Grobety J, Majno G. Mesothelial injury and recovery. Am J Pathol 1973;71:93-102. 30. Buckman RF, Woods M, Sargent L, Gervin AS. A unifying pathogenetic mechanisms in the etiology of intraperitoneal adhesions. J Surg Res 1976;20:1-5. 31. Raftery AT. Regeneration of peritoneum: a fibrinolytic study. J Anat 1979;129:659-64. 32. Dorr PJ, Verner HM, Brommer EJ, Willemsen WN, Veldhuizen RW, Rolland R. Prevention of postoperative adhesions by tissue-type plasminogen activator (t-PA) in the rabbit. Eur J Obstet Gynecol Reprod Bioi 1990;37:28791. 33. Menzies D, Ellis H. Intra-abdominal adhesions and their prevention by topical t-PA. J Royal Soc Med 1989;82: 534-5. 34. Onderdonk AB, Moon NE, Kasper DL, Bertlett JG. Adherence of Baeteriodes fragilis in vivo. Infect Immun 1978;19:1083-7. 35. Shade DS, Williamson JR. The pathogenesis of peritoneal adhesions: an ultrastructural study. Ann Surg 1968; 167:500-10. 36. Cliff WJ, Grobety J, Ryan GB. postoperative pericardial adhesions: the role of mild serosal injury and spilled blood. J THORAC CARDIOVASC SURG 1973;65:744-50. 37. Ishihara T, Ferrans VJ, Jones M, Boyce SW, Kawanami 0, Roberts WC. Histologic and ultrastructural features of normal human parietal pericardium. Am J Cardiol 1980; 46:744-53. 38. Leak LV, Ferrans VJ, Cohen SR, Eidbo EE, Jones M. Animal model of acute pericarditis and its progression to pericardial fibrosis and adhesions: ultrastructural studies. Am J Anat 1987;180:373-90.
U. U. Nkere, FRCSE, S. A. Whawell, PhD, E. M. Thompson, MRCP, MRCPath,
J. N. Thompson, MChir, FRCS, and K. M. Taylor, MD, FRCS, FRCSE, FSA, London, England
"Wh the expanding volume of cardiac surgery throughout the world, reoperations have become more From the Royal Postgraduate Medical School, Hammersmith Hospital, London, England. Received for publication Feb. 5, 1992. Accepted for publication July 27,1992. Address for reprints: Udim U. Nkere, FRCSE, Department of Cardiothoracic Surgery, Second Floor, B Block, Hammersmith Hospital, Du Cane Road, London WI2 OHS, England. Copyright
1993 by Mosby-Year Book, Inc.
0022-5223/93 $1.00 +.10
12/1/41365
common. In some centers they constitute 10% to 20% of the annual caseload.' The presence of pericardial adhesions at resternotomy not only increases the operation time but also increases the risk of life-threatening damage to the heart, great vessels,or extracardiac grafts.? The reported prevalence of damage is 2% to 6%.3 Furthermore, there is evidence that adhesions have a deleterious effect on cardiac function" and that postoperative pericarditis and subsequent adhesion formation may lead to decreased graft patency.f 6 Various strategies have been used to overcome or prevent the problems associated with adhesions, such as modified pericardial closure.I: 8 peri-
339
The Journal of Thoracic arid
340
Nkere et al.
Cardiovascular Surgery August 1993
Tissuetype PA
Plasminogen Activator (PA) Urokinase type PA
Plasrnfnoqen Act ...."
'Ohl,,,,--It
(pAI-1 and PAI-2)
PLASMINOGEN ~ PLASMIN
t. FIBRIN~ '\
FIBRINOGEN
FIBRIN DEGRADATION PRODUCTS
THROMBIN
Fig. 1. Fibrinolytic pathway. cardia I substitutes," dextran pericardial washout, 10 pharmacologic agents, II and the left thoracotomy approach. 12 Nevertheless, although successful in an experimental setting, most of these methods have been disappointing clin-
ically.' Pericardial tissue injury and subsequent inflammation may give rise to an outpouring of fibrin-rich exudate. This fibrin, together with that of spilled pericardial blood, gives rise to fibrinous adhesions. This fibrin may then either undergo organization by fibroblast infiltration and collagen deposition to form dense adhesions or, as a result of fibrinolysis, be resolved. Damage to pericardial tissue and a reduction in the fibrinolytic capacity of pericardial tissue, as reflected in the plasminogen activating activity (PAA), may be crucial in allowing these pericardial adhesions to form. The fibrinolytic system 13 is a complex pathway, simplified in Fig. 1, involving enzymes and their inhibitors that enable precise control of fibrin deposition and removal.!" There are two principal plasminogen activators (PA), tissue-type PA (t-PA), which is present in plasmal" and many tissues and body fluids.!" and urokinase-type PA (u-PA), which is present in urine!" and some tissue mucosae. I 8 These enzymes are inhibited by PA inhibitor-l (PAl-I) and PA inhibitor-2 (PAI-2). Free PA acts on plasminogen to produce the fibrindegrading enzyme plasmin. The aims of this study were to determine whether pericardial tissue possesses PAA and, if so, to identify the principal fibrinolytic activators and inhibitors, to measure quantitative variation in the activity of these enzymes during cardiopulmonary bypass (CPB), and to document simultaneous changes in pericardial morphologic characteristics.
Patients and methods Patients. Ten patients undergoing primary elective cardiac operations were studied. Eight patients had coronary artery bypass grafts, one patient had mitral valvereplacement, and one patient had aortic valvereplacement. There were sevenmen and
three women. The median age was 62.5 years (range 42 to 74 years). Myocardial protection was maintained with cardioplegia (St. Thomas' Hospital solution), moderate systemic hypothermia (28 0 C, nasopharyngeal), and topical hypothermia. The median crossclamp time was 55.5 minutes (range 29 to 90 minutes). The median perfusion time was 99.5 minutes (range 49 to 124 minutes). Pericardial biopsy technique. The pericardium was opened centrally in the usual manner. Pericardial biopsy specimens were taken with a disposable 6 mm diameter biopsy punch (Stiefel Laboratories [U.K.) Ltd., Wooburn Green, United Kingdom). The samples were obtained from the nondiaphragmatic sides of the open pericardium, all within 3 em of its cut edges. Three samples were taken at times 0, 75, and 135 minutes from the time of openingof the pericardium. Two werethen wrapped in aluminum foil and snap frozen before transfer to a storage freezer at -20 0 C. The other sample was kept in 10% buffered formalin for subsequent histologic analysis. Sixty samples were therefore available for biochemical analysis and 30 samples for histologic examination. Preparation of pericardial homogenates. All sampleswere assayed within 2 weeks. The pericardial biopsyspecimens were thawed at room temperature and weighed. Each sample was washed with 0.5 ml rinsing solution (sodium dihydrogen phosphate 5 mrnol/L, sodium chloride 0.15 mol/L, pH 7.4) and placed into a small plastic test tube, on ice, containing I ml homogenizing solution (sodium dihydrogen phosphate 2.5 mmol/L, sodium chloride 0.075 mol/L, 0.25% Triton XlOO [Sigma Chemical Co. Ltd., London, U.K.), pH 7.8). The tissues were then homogenizedwith an Ultra- Turrax (Janke & Kunkel GmbH & Co., KG, Staufen, Germany) homogenizer for 30 seconds.The homogenates were centrifuged at 12,000g for 20 minutes at 4 0 C. Aliquots of 0.25 ml of supernatant werestored at -20 0 C until assay. Assay for P AA. PAA in the pericardial tissue extracts was assayed by a standard fibrin plate technique.19 A layer of fibrin was produced in the base of 5 cm diameter plastic Petri dishes (Sterlin Ltd., London, U.K.) by pouring in a solution containing 0.9 ml buffer (sodium acetate 0.03 mol/L, sodium barbitone 0.03 mol/L, hydrochloric acid 0.02 mol/L, sodium chloride 1.2 mrnol/L, pH 7.4), 3 ml of 1% human fibrinogen (KabiVitrum Ltd., Uxbridge, U.K.), 0.3 ml of human plasminogen (KabiVitrum Ltd., 2.5 units/rnl), and 0.2 ml bovinethrombin (Armour Pharmaceutical Co., Kankakee, Ill., 20 units/rnl). After the fibrinsolution had set, a 6 mm diameter wellwas cut in the center of the plate. The pericardial tissue homogenates were
The Journal of Thoracic and Cardiovascular Surgery Volume 106, Number 2
Nkere et al.
34 I
Fig. 2. Histologic features of parietal pericardium. A, Section through the fibrosal and outer epipericardiallayers shows vascular congestion (VC). B, Section shows margination of neutrophils (M).
brought to room temperature and 20 ILl of each specimen was placed in the central well of a fibrin plate. Standard solutions of human t-PA (Second International Standard, National Institute of Biological Standards and Controls, Mill Hill, U.K.) reconstituted in sterile water were placed on fibrin plates at the time of each batch of assays. All plates were then incubated at 37° C for 24 hours and the diameter of the zone of lysis around the central well was measured directly by placing the plates on a measuring scale. The maximum and minimum diameters were recorded and the mean value used. A standard plot of diameter of fibrin plate lysis against log t-PA was constructed for each assay and the line of best fit was calculated by means of least squares regression analysis. The mean PAA level for the paired samples was calculated. Homogenate supernatant was also applied to fibrin plate that had been previously heated to 80° C for 20 minutes (thereby destroying the plasminogen) to establish whether fibrinolysiswas plasminogen mediated. The fact that lysis did not occur in those plates so treated confirmed that fibrinolysis was mediated through plasminogen.P The PAA was expressed in international units per square centimeter of pericardium. The lower limit of sensitivity of the assay was 0.07 IU j cm-, Enzyme-linked immunosorbent assays. All the homogenates were assayed for the following fibrinolytic components by means of commercially available enzyme-linked immunosorbent assays (Tintylse, Porton International Group, London, U.K.): t-PA, u-PA, PAl-I, and PAI-2. The lower limits of detection were 0.75 ngjml for t-PA, 0.1 ngjml for u-PA, 2.5 ngjml for PAl-I, and 6 ngjml for PAI-2. Histologic evaluation. Formalin-fixed paraffin-embedded
sections of pericardial tissue were stained with hematoxylin and eosin and assessed by a histopathologist who had no knowledge of the patient or of the time of biopsy. The following features were assessed as evidence of progressive inflammatory change:
I. Vascular congestion (Fig. 2, A) 2. Margination of neutrophils (Fig. 2, B) 3. Frank inflammation (cellular infiltration of the connective tissue) In addition, progressive mesothelial cell damage was evidenced by rounding of cells (Fig. 3, A) and then cell loss (Fig. 3, 8). Statistical methods. Friedman's nonparametric two-way analysis of variance was used to investigate whether there was any overall discernible reduction in the PAA and t-PA levels with time. The Wilcoxon matched-pairs test was then used to look more closely at where the discernible reduction lay. The x 2 test was used to analyze the significance of the histologic changes, and Spearman's rank correlation coefficient was used to detect the linear association between PAA and t-PA and between PAA or t-PA and the mean sample weight of the pericardial biopsy specimens.
Results Examination of the data by means of Friedman's nonparametric two-way analysis of variance suggested that there was an overall discernible reduction in the P AA (p = 0.05) and t-PA (p = 0.004) levels with time. This possibility was then investigated more closely, by means
The Journal of Thoracic and Cardiovascular Surgery August 1993
3 4 2 Nkere et al.
Fig. 3. Histologic features of parietal pericardium. A, Section shows rounded mesothelial cells (RC) becoming detached from the fibrosallayer. B, Section shows detached (D) and free-floating cells (FF).
... Spearman rank Corr.Coe1._ 0.73
... I
...
PM (IUterl) 4
P.O.OOOl
... <0.07
.j2:::;::::!~:::;::::::::;;:::;::::;::::;:::::;:::::;::::;:::;::"::1 <0.751
2
3
4
tPA(nglml)
Fig. 4. PAA ofallpericardial biopsy specimens plotted against the corresponding topA level. of the Wilcoxon matched-pairs test, to see where the difference lies. The results are explained in the following paragraphs. For each of the 10 patients, the PAA ofthe pericardial biopsy specimens was measured in all the homogenate samples. Compared with initial levels (median 2.06 IV/cm 2, range 1.28 to 6.48 IV/cm 2) , the PAA of pericardial biopsy specimens was significantly reduced at 75 minutes (median 0.64 IV/cm 2, range 0.12 to 2.44 IV/cm 2, p < 0.01) and at 135 minutes (median 1.45 IV/cm2, range 0.12 to 4.39 IV/cm 2, p < 0.05). There
was no significant difference between the PAA levels at 75 and 135 minutes. The major PA present was t-PA. Compared with the initial levels (median 2.34 ng/rnl, range 1.03 to 6.42 ng/rnl), subsequent t-PA values were also significantly reduced at 75 minutes (median 0.83 ng/rnl, range 0.75to 5.13 ng/rnl, p < 0.01) and at 135 minutes (median 1.24 ng/rnl, range 0.75 to 6.67 ng/rnl.p < 0.05). Again, there was no significantdifferencebetweenthe t-PA levels at 75 and 135 minutes. Low levels of u-PA were detected in 5 of 10 patients (median 0.125 ng/rnl, range 0.1 to OJ ng/rnl, lower limit of detection 0.1 ng/rnl). However, neither PAI-! nor PAI-2 was detected (lower limit of detection ::;2.5 ng/rnl and ::;6 ng/rnl, respectively). There wasa significantpositive correlationbetweenthe PAA and t-PA levels (Fig. 4). Together with the fact that t-PA levels were approximately 10 times higher than u-PA levels, this suggests that t-PA rather than u-PA is responsible for the PAA exhibited by the pericardium. The histologicfindingsof mesothelialdamage and the extent of pericardial inflammation are shown in Figs. 5 and 6, respectively. These showclear evidenceof increasing mesothelial damage and progressive inflammatory changes with time. The extent of mesothelial damage at times 75 and 135 minutes compared with that at time 0 show a significant (p < 0.005) preponderanceof pericar-
The Journal of Thoracic and Cardiovascular Surgery Volume 106, Number 2
Nkere et al.
34 3
Fig. 5. Mesothelial damage with time. Extent of mesothelial damage at times 75 and 135 minutes compared with that at time a shows a significant (p < 0.005) preponderance of tissue either with rounded mesothelial cells or with
mesothelial cell loss.
Fig. 6. Inflammatory changes with time. Pericardial samples at times 75 and 135 minutes compared with those at time a show a significant (p < 0.005) preponderance of tissue with margination and frank inflammation.
dial tissue either with rounded mesothelial cells or with mesothelial cell loss. Similarly, a significant (p < 0.005) preponderance of tissuewith features of margination and frank inflammation, rather than normal or congestedtissue, was noted at times 75 and 135 minutes compared with that at time O. Discussion
This study, in keeping with previous work by Porter, Ball, and Silver" and again by Gervin and associates.F has shown that mesothelial pericardial tissue expresses fibrinolytic activity. Pericardial tissueshould therefore be capable of removing fibrinand hence,at least potentially, beable to prevent or minimize adhesion formation after cardiac operations. The fibrinolytic activityis reflectedby the PAA of pericardial tissue; this activity, as suggested by the positive correlation between PAA and t-PA, appears to be primarily due to the presenceof t-PA rath-
er than u-PA. In comparisonwith initial values,the PAA and the levels of t-PA were significantly lower at later stages of cardiac surgical procedures. The reduction in pericardial fibrinolytic activity,seen during cardiac operations,suggests a reducedcapacity of pericardial tissueto prevent adhesion formation. As appears to be the case in the peritoneum,wherethe reduction in peritoneal PAA is stronglyimplicated as a causal factor in peritoneal adhesionformation,23-25 this reduced pericardial PAA may be a significant factor in the pathophysiology of pericardial adhesions. The hemodilutionthat occurs as a result of the institution of CPB is known to decrease the blood protein concentrations. Hence it could be argued that the observed reductionin PAA and t-PA are, in part, a reflection of this hemodilution. However, it has been observedthat a similar perioperative reduction in PAA occurs in patients undergoingelectivelaparotomy" in whom hemodilution
344
The Journal of Thoracic and Cardiovascular Surgery August 1993
Nkere et al.
Spearman rank Corr. Coet.• 0.'92
I
P.0.30
., .
...,.. ... '
.
10 20 Sampleweight (mg)
30
Fig. 7. PAA ofall pericardialbiopsyspecimens plottedagainst the corresponding mean sample weight.
does not playas major a role. Nevertheless, the role of hemodilution will be addressed in further studies. Although thePAAand t-PA levels at times 75 minutes and 135 minutes were not statistically different, inspection of the data suggest that these values are on the increase at 135 minutes. Furthermore, the statistical difference between the PAA and t-PA levels at times 0 and 75 minutes are greater than at 0 and 135 minutes. All patients were rewarmed and weaned from CPB by 135 minutes; therefore, the apparent recovery in fibrinolytic activity may have some relation to the effects of CPB and temperature. The extent of this recovery and the degree to which it is impaired by the mesothelial damage and the simultaneous inflammatory response has not been answered by this study. PAI-l and PAI-2 were not detected in the pericardial tissue within the duration of CPB, which suggests that they play no part in the observed early decrease in PAA. However, in peritoneal studies." normal peritoneum likewise has no detectable PAl-lor PAI-2 whereas inflamed peritoneum has reduced PAA and detectable PAl-I andPAI-2. It may be that, asin peritoneum, PAI-I and PAI-2 appear in pericardial tissue at a later stage than after the initial insult. The lack of correlation between the PAA, or t-PA levels, and the mean sample weight of the pericardial biopsy specimens (Fig. 7) provides indirect evidence for the site of production of PA. The use of a 6 mm diameter punch biopsy standardizes the area of mesothelium sampled. The different weights of the biopsy specimens thus for the most part reflect the different amounts of underlying tissues sampled. If submesothelial tissues were the major site of plasminogen activity, a positive correlation between PAA and mean sample weight might be expected. The fact that several low-weight samples exhibited high PAA values and vice versa supports the theory that the surface mesothelial layer is responsible for most of the pericardial
fibrinolytic activity.P' However, the precise identification of the exact cellular origin and tissue localization of pericardial PAA requires further study by means of techniques such as in situ messenger ribonucleic acid hybridization. Previous studies not only have shown the presence of PAA in pericardium and other mesothelial structures-": 22 but also have indicated that injury/? (e.g., physical trauma, drying, diathermy, and ischemia) decreases this activity. 30, 31 It is possible that mesothelial damage during surgical procedures, caused by drying or manipulation, may be responsible for the decrease in PAA and t-PA observed in this study. However, it is not yet clear whether the diminished PAA demonstrated in the pericardium is a causative factor in the process of formation of the dense adhesions seen at reoperation. Nevertheless, it has been shown by Dorr and associates 32 and Menzies and Ellis 33 that, after mesothelial injury, enhancing PAA or fibrinolytic activity, in the form of topical t-PA, may prevent adhesion formation in the peritoneum. The present study also confirms that there is an early inflammatory reaction in pericardial tissue in response to the insult of cardiac surgery. The response of mesothelium to injury has been described by previous investigators using bacterial toxins.l" silica,35 and various chemical and mechanical." stimuli. In general, this cellular reaction is characterized by a change in cell shape from a normal flat appearance'? to cuboidal, a separation of neighboring cells, and a shedding or detachment of the mesothelium from the basal lamina.P Further studies are planned to determine the time scale and degree of pericardial tissue regeneration in the postoperative period and to investigate the relation, if any, to recovery of pericardial fibrinolytic activity. Ryan, Grobety, and Majno,29 in assessing both morphologic and fibrinolytic parameters, suggests that there is evidence for some degree of mesothelial recovery over 3 to 21 days after the operation.
I. 2. 3. 4. 5.
REFERENCES Gabbay S. The need for intensive study of pericardialsubstitutionafter openheart surgery[editorial].ASArO Trans 1990;36:789-91. Loop FO, Cosgrove OM, Kramer JR. Late clinical and arteriographic results in 500 coronary artery reoperations. J THORAC CARD10VASC SURG 1981 ;81 :675-84. LoopFO. Catastrophic hemorrhageduring sternalreentry. Ann Thorac Surg 1984;37:271-2. BaileyLL, Li Z-j, Schulz E, Roost H, Yahiku P. A cause of right ventricular dysfunction after cardiac operations. J THORAC CARDIOVASC SURG 1984;87:539-42. Urschel HC Jr, Razzuk MA, Gardner M. Coronaryartery
The Journal of Thoracic and Cardiovascular Surgery Volume 106, Number 2
bypass occlusion secondary to postcardiotomy syndrome. Ann Thorac Surg 1976;22:528-31. 6. Shapira N, Gordon CI, Lemole GM. Occlusion of aortocoronary veingrafts in association with bovinepericardium. Am J Cardiovasc PathoI1989;3:87-90. 7. Zapolanski A, Fishman NH, Bronstein MN, Ellertson DG, O'Connell TJ, Siegel S. Modified pericardial closure to protect cardiovascular structures during sternal reentry. Ann Thorac Surg 1990;50:665-6. 8. Milgalter E, Uretzky G, Siberman S, et al. Pericardial meshing: an effective method for prevention of pericardial adhesions and epicardial reaction after cardiac operations. J THORAC CARDIOVASC SURG 1985;90:281-6. 9. Gabbay S, Guindy AM, Andrews JF, Amato JJ, Seaver P, Khan Y. New outlook on pericardial substitution after open heart operations. Ann Thorac Surg 1989;48:803-12. 10. Reikeras 0, Nordstrand K, Serlie D. Use of dextran to prevent pericardial adhesions caused by maize starch powder. Eur Surg Res 1987;19:62-4. 11. Vander Salm TJ, Okike ON, Marsicano TH, Compton C, Espinoza E. Prevention of postoperative pericardial adhesions. Arch Surg 1986;121:462-7. 12. Gandjbakhch I, Acar C, Cabrol C. Left thoracotomy approach for coronary artery bypass grafting in patients with pericardial adhesions. Ann Thorac Surg 1989;48: 871-3. 13. Paques EP. Recent advances in the biochemistry of the fibrinolytic system. Behring Inst Mitt 1988;82:68-81. 14. Collen D. On the regulation and control of fibrinolysis. Thromb Haemost 1980;43:77-89. 15. Rijken DC, Wijngaards G, Welbergen J. Relationship between t-PA and the activators in blood and vascular wall. Thromb Res 1980;18:815-30. 16. Rijken DC, Wijngaards G, Welbergen J. Immunological characterisation of plasminogen activator activities in human tissues and body fluids. J Lab Clin Med 1981; 94:477-86. 17. Williams JRB. The fibrinolytic activity of urine. Br J Exp PathoI1951;32:530-7. 18. Larsson LI, Skriver L, Nielsen LS, Grondahl-Hansen J, Kristensen P, Dano K. Distribution of urokinase-type plasminogen activator immunoreactivity in the mouse. J Cell Bioi 1984;98:894-903. 19. Astrup T, Mullertz S. The fibrin plate method for estimating fibrinolytic activity. Arch Biochem Biophys 1952; 40:346-51. 20. Lassen M. Heat denaturation of plasminogen in the fibrin plate method. Acta Physiol Scand 1952;27:371-6. 21. Porter JM, Ball AI', Silver D. Mesothelial fibrinolysis. J THORAC CARDIOVASC SURG 1971;62:725-30. 22. Gervin AS, Jacobs G, Hufnagel HV, Mason KG. Surgical trauma and pericardial fibrinolytic activity. Am Surg 1975;41 :225-9.
Nkere et al. 3 4 5
23. Thompson IN, Paterson-Brown S, Harbourne T, Whawell SA, Kalodiki E, Dudley HAF. Reduced human peritoneal plasminogen activating activity: possible mechanism of adhesion formation. Br J Surg 1989;76:382-4. 24. Gervin SA, Puckett CL, Silver D. Serosal hypofibrinolysis: a cause of postoperative adhesions. Am J Surg 1973; 125:80-7. 25. Whitaker D, Papadimitriou JM, Walters MNI. The mesothelium: its fibrinolytic properties. J Pathol 1982; 136:291-9. 26. Scott-Coombes DM, Whawell SA, Thompson IN. Per-operative changes in peritoneal fibrinolytic activity. Fibrinolysis 1992;6(suppl 2):61. 27. Vipond MN, Whawell SA, Thompson IN, Dudley HAF. Peritoneal fibrinolytic activity and intraabdominal adhesions. Lancet 1990;335:1120-2. 28. Merlo G, Fausone G, Barbero C, Castagna B. Fibrinolytic activity of the human peritoneum. Eur Surg Res 1980; 12:433-8. 29. Ryan GB, Grobety J, Majno G. Mesothelial injury and recovery. Am J Pathol 1973;71:93-102. 30. Buckman RF, Woods M, Sargent L, Gervin AS. A unifying pathogenetic mechanisms in the etiology of intraperitoneal adhesions. J Surg Res 1976;20:1-5. 31. Raftery AT. Regeneration of peritoneum: a fibrinolytic study. J Anat 1979;129:659-64. 32. Dorr PJ, Verner HM, Brommer EJ, Willemsen WN, Veldhuizen RW, Rolland R. Prevention of postoperative adhesions by tissue-type plasminogen activator (t-PA) in the rabbit. Eur J Obstet Gynecol Reprod Bioi 1990;37:28791. 33. Menzies D, Ellis H. Intra-abdominal adhesions and their prevention by topical t-PA. J Royal Soc Med 1989;82: 534-5. 34. Onderdonk AB, Moon NE, Kasper DL, Bertlett JG. Adherence of Baeteriodes fragilis in vivo. Infect Immun 1978;19:1083-7. 35. Shade DS, Williamson JR. The pathogenesis of peritoneal adhesions: an ultrastructural study. Ann Surg 1968; 167:500-10. 36. Cliff WJ, Grobety J, Ryan GB. postoperative pericardial adhesions: the role of mild serosal injury and spilled blood. J THORAC CARDIOVASC SURG 1973;65:744-50. 37. Ishihara T, Ferrans VJ, Jones M, Boyce SW, Kawanami 0, Roberts WC. Histologic and ultrastructural features of normal human parietal pericardium. Am J Cardiol 1980; 46:744-53. 38. Leak LV, Ferrans VJ, Cohen SR, Eidbo EE, Jones M. Animal model of acute pericarditis and its progression to pericardial fibrosis and adhesions: ultrastructural studies. Am J Anat 1987;180:373-90.