ascites fluid as a possible origin for hyperfibrinolysis in advanced liver disease

Ascites Fluid as a Possible Origin forHyperfibrinolysis in Advanced Liver DiseaseSwati Agarwal, M.D.,* Kelly A. Joyner, Jr., M.S., M.T. (A.S.C.P.)S.H., and Mark W. Swaim, M.D., Ph.D.†School of Medicine, Clinical Coagulation Laboratory, and Liver Center, Division of Gastroenterology,Department of Medicine, Duke University Medical Center, Durham, North Carolina

OBJECTIVES: Advanced liver disease is associated with bothexaggerated fibrinolysis and with ascites. This study wasundertaken to determine whether fibrinolytic activity existsin the ascites fluid of patients with liver disease and to seewhether such activity is associated with evidence of plasmafibrinolysis.

METHODS: Both the ascites fluid and plasma from 15 pa-tients with cirrhotic ascites (group A) were evaluated formarkers of fibrinolysis: fragmentD-dimer, plasminogen, fi-brinogen, and fibrin split products. In addition, the euglob-ulin lysis time, a test highly specific for fibrinolysis, wasevaluated in the ascites fluid samples. As a control group,the plasma from 15 cirrhotic patients without ascites (groupB) was evaluated for markers of fibrinolysis.

RESULTS: In group A, elevated fragmentD-dimer and fibrinsplit products were uniformly found in ascites fluid in con-centrations that would be considered pathologically elevatedif in plasma. Ascites fluid was also depleted, compared withplasma, of both plasminogen and fibrinogen. These results,along with the short euglobulin lysis time in 83% of thepatients, suggest that increased fibrinolytic activity ispresent in ascites fluid. In 93% of these patients, plasmaD-dimer was elevated. The mean plasma plasminogen wasalso low in these patients. In group B, only 33% of patientshad elevated plasmaD-dimer.

CONCLUSIONS: Ascites fluid has fibrinolytic activity. Be-cause ascites fluid reenters the systemic circulation via thethoracic duct, via a natural peritoneovenous shunt, ascitesfluid warrants serious consideration as a pathological fluidthat contributes to the systemic fibrinolytic state found in themajority of our patients with ascites. (Am J Gastroenterol2000;95:3218–3224. © 2000 by Am. Coll. of Gastroenter-ology)

INTRODUCTION

Patients with cirrhosis have a bleeding tendency that is oftennot evident from routine clotting studies, such as the pro-

thrombin time (PT). Many such patients have increasedfibrinolysis that may contribute to their susceptibility toserious bleeding events, such as variceal hemorrhage (1–3).Even in 1914, Goodpasture observed that clotted blood fromcirrhotic patients undergoesex vivopremature autolysis (4).

Hemostasis is a dynamic balance between procoagulantand anticoagulant forces. The anticoagulant forces includetwo predominant classes of proteins, antithrombotic andfibrinolytic (reviewed in Table 1). Whereas the antithrom-botic proteins prevent fibrin clot formation, the fibrinolyticproteins lead to digestion of fibrin(ogen). In cirrhosis, wheremany patients are in chronic accelerated fibrinolysis (1–3),fibrin clot may be degraded as quickly as it is formed,severely compromising hemostasis. As Table 1 implies,normal PT and activated partial thromboplastin times(aPTT) may not reveal a chronic fibrinolytic state becausethe procoagulant systems may not be compromised.

Why patients with liver disease should undergo exagger-ated fibrinolysis is not known. Because advanced liver dis-ease is associated with both exaggerated fibrinolysis andwith ascites, we have hypothesized that ascites plays a rolein causing fibrinolysis. Ascites fluid is essentially an ultra-filtrate of plasma in patients with portal hypertension, oftenpresent in volumes that exceed total blood volume. Beingderived from plasma, the fluid contains coagulation-relevantproteins, but in an environment where their actions cannotbe well-regulated.

Prior observations support ascites as a cause of systemicfibrinolysis. Serious derangements in coagulation followperitoneovenous (LeVeen) shunting (5–10). Such shunting,which is venous reinfusion of ascites, causes laboratoryabnormalities consistent with severe induced fibrinolysis(5–10). If the ascites compartment is a site at which fibrino-lytic activity arises, then ascites may contribute to the ex-aggerated fibrinolysis and bleeding tendency typical in ad-vanced liver disease.

Our present study has three objectives: 1) to determinewhether ascites fluid from patients with cirrhosis has fi-brinolytic activity; 2) to determine whether cirrhotic patientswith ascites are more likely than those without ascites tohave increased plasma fibrinolysis; and 3) to determinewhether ascites fluid in patients without cirrhosis has fi-

* Swati Agarwal, M.D., is currently Intern, Department of Pediatrics, StanfordUniversity, Palo Alto, California.

† Mark W. Swaim, M.D., Ph.D., is currently Assistant Professor, M.D. AndersonCancer Center, University of Texas, Houston, Texas.

THE AMERICAN JOURNAL OF GASTROENTEROLOGY Vol. 95, No. 11, 2000© 2000 by Am. Coll. of Gastroenterology ISSN 0002-9270/00/$20.00Published by Elsevier Science Inc. PII S0002-9270(00)02092-X

brinolytic activity and whether plasma from these patientshas fibrinolytic activity.

MATERIAL AND METHODS

Cirrhotic Patients With Ascites (Group A)Of 24 patients with ascites and cirrhosis that underwentparacentesis on the Duke University Medical Center LiverService between August 1998 and April 2000, 15 patientswith biopsy-proven cirrhosis met study criteria (Table 2).Five were excluded because of spontaneous bacterial peri-tonitis, and one was excluded for each of the followingconditions: nephrotic syndrome, renal failure, pancreatitis,and GI bleeding.

Of the 15 that met study criteria, there were 11 men and4 women aged 38 to 73 yr old (mean age, 56 yr). Theunderlying disease was Laennec cirrhosis in seven patients,hepatitis C cirrhosis in one patient, hepatitis C/Laenneccirrhosis in two patients, hepatitis B cirrhosis in one patient,and cryptogenic cirrhosis in four patients.

Noncirrhotic Patients Without Ascites (Group B)The plasma of 15 Duke Liver Center patients with biopsy-proven cirrhosis and recent (during the study period) radio-graphic confirmation of no ascites was collected during thestudy period. These patients met the study criteria in Table3.

Seven men and eight women aged 34 to 66 (mean age, 51yr) were studied. The underlying diagnosis was Laenneccirrhosis in one patient, hepatitis C in four patients, hepatitisC and Laennec cirrhosis in three patients, cryptogenic cir-rhosis in one patient, hepatitis C and hemochromatosis inone patient, autoimmune hepatitis with cirrhosis in threepatients, and cirrhosis from nonalcoholic steatohepatitis intwo patients.

Noncirrhotic Patients With Ascites (Group C)During the study period, two patients were identified withascites that met the criteria for inclusion set forth in Table 2,except that these patients were proven by liver biopsy not tohave cirrhosis. The ascites and plasma of these patients wereobtained for study. The first patient was a 51-yr-old manwith constrictive pericarditis. The second patient was a76-yr-old woman with postoperative malnutrition.

Table 1. Hemostasis: A Balance Between Procoagulant and Anticoagulant Forces

Procoagulant Anticoagulant

Prothrombotic Antifibrinolytic Antithrombotic Fibrinolytic

Proteinases Proteinase inhibitors Proteinase Inhibitors ProteinasesFactor IIa (thrombin) PAI Anti-thrombin III Pg activatorsFactor IXa a2PI TFPI PlasminFactor Xa Heparin cofactor IIFactor VIIa a2-macroglobulinFactor XIa

Proteinase cofactors Proteinase cofactorFactor V (cofactor for Factor Xa) Protein S (cofactor for protein C)Factor VIII (cofactor for Factor IXa) Thrombomodulin (cofactor fo FIIa)

Proteinase inhibitor ProteinaseProtein C inhibitor Protein C

PAI 5 plasminogen activator inhibitor;a2PI 5 alpha-2-plasmin inhibitor; TFPI5 tissue factor pathway inhibitor; Pg activators5 plasminogen activators.

Table 2. Patient Selection Criteria for Patients With CirrhosisWith Ascites (Group A)

Inclusion Criteria Exclusion Criteria

1) Biopsy-proven cirrhosis Malignant ascites2) SAAG $ 1.1 Gastrointestinal hemorrhage in

previous 72 hFunctioning transjugular

intrahepatic portosystemic shuntPrior liver transplant without

recurrent cirrhosisRenal failure patients on

hemodialysisNephrotic syndromeSpontaneous bacterial peritonitis

(granulocyte count. 250/mm3)PancreatitisSepsisEvidence of thromboembolic eventChylous ascites

SAAG 5 serum-ascites albumin gradient.

Table 3. Patient Selection Criteria for Patients With CirrhosisWithout Ascites (Group B)

Inclusion Criteria Exclusion Criteria

Biopsy-proven cirrhosis Gastrointestinal hemorrhage inprevious 72 hRadiographic evidence of

no ascites Functioning TIPSSPrior liver transplant without

recurrent cirrhosisRenal failure patients on

hemodialysisNephrotic syndromePancreatitisSepsisEvidence of thromboembolic event

TIPSS5 transjugular intrahepatic portosystemic shunt.

3219AJG – November, 2000 Ascites: Origin for Hyperfibrinolysis?

Sample AssaysAscites fluid was evaluated for the following: granulocytecount, albumin, fibrinogen, fragmentD-dimer, functionalplasminogen, fibrin split products (FSPs), and ascites eu-globulin lysis time (AELT). The specimens were collectedduring routine diagnostic or therapeutic paracentesis andplaced immediately on ice. In addition, patient plasma wasevaluated for albumin, fragmentD-dimer, plasminogen, andfibrinogen.

All blood or ascites fluid for the coagulation studies werecollected in either 3.2% (0.105 mol/L) sodium citrate orsterile interior-evacuated tubes (Becton Dickinson, FranklinLakes, NJ) and kept at 4°C until assay. All samples were

centrifuged at 3500 rpm for 15 min to harvest platelet-poorplasma or serum for the studies.

Plasminogen (functional) was performed on the MDA180 coagulometer (Organon Teknika, Durham, NC). Thezymogen plasminogen was activated by streptokinase, andthe resulting plasminogen-streptokinase complex (plasmin-like activity) was directed toward the chromogenic substratesupplied. Release of the chromophorep-nitroaniline is di-rectly related to plasminogen concentration in the patientplasma samples. Plasminogen is reported as percentage ofnormal activity.

Fibrinogen was measured using the method of Clauss onthe MDA 180 coagulometer and reported in milligrams perdeciliter. FragmentD-dimer was performed by latex agglu-tination (American Bioproducts, Parsippany, NJ) on ascitesfluid and plasma samples. The latex agglutination uses amonoclonal antibody specific for the fragmentD-dimer frombreakdown of cross-linked fibrin clot. Titration to an end-point was performed on all samples; results are reported inmicrograms per milliliter.

A semiquantitative measurement of fibrinolysis in ascitesfluid was performed by a method analogous to that used forplasma euglobulin lysis time determinations (11). An ascitesprotein precipitate was obtained by combining 0.2 ml ofacetic acid (1% v/v), 8 ml H2O, and 0.5 ml of ascites. Thismixture was kept at 4°C for 30 min. A precipitate wasobtained by centrifugation (2500 rpm for 15 min). Pelletswere resuspended in 0.5 ml borate buffer, followed byaddition of calcium chloride (0.5 ml of 0.25 mol/L solution).After 10 min at 4°C, samples were incubated at 37°C.Observations were made every 15 min until complete pre-cipitate lysis had occurred. This time measurement wasrecorded as the AELT.

Fibrin and FSPs were measured by using an FSP kit(Organon Teknika, Durham, NC). The test latex particles

Table 4. Ascites and Plasma Data for Patients with Cirrhosisand Ascites (Group A)

Patient

AscitesGranulocyte

Count(per mm3)

AscitesD-dimer(mg/ml)

PlasmaD-dimer(mg/ml)

AscitesFibrinSplit

Products(mg/ml)

AscitesEuglobulin

LysisTime(min)

1A 11 .1.25 .1.25 .20 ,1202A 9 .1.25 .25–1.25 .20 ,1203A 2 .1.25 .1.25 .20 ,1204A 11 .1.25 .1.25 .20 ,1205A 8 .1.25 .1.25 .20 ,1206A 16 .1.25 .25–1.25 .20 .1207A 5 .1.25 .1.25 .20 ,1208A 10 .1.25 .1.25 .20 ,1209A 0 .1.25 .1.25 .20 ,12010A 0 .1.25 .20–1.25 .20 .12011A 32 .1.25 .1.25 .20 ,12012A 8 .1.25 ,1.25 .20 .12013A 4 .1.25 .0.25 .20 ,12014A 14 .1.25 .1.25 .20 ,12015A 14 .1.25 .1.25 .20 ,120

Normal plasma values: granulocyte count,,250/mm3; D-dimer,,0.25mg/ml; fibrinsplit products,,5 mg/ml; ascites euglobulin lysis time,.120 min.

Table 5. Ascites and Plasma Data for Patients With Cirrhosis and Ascites (Group A)

Patient

AscitesPlasminogen

(%)

PlasmaPlasminogen

(%)

AscitesFibrinogen

(%)

PlasmaFibrinogen

(mg/dl)

Plasma-AscitesPlasminogen

Ratio SAAG

1A 10.7 70.5 ,10 333.7 7 .1.32A 12.6 86.1 ,10 187.2 7 2.33A 7.5 49.4 ,10 120.6 7 .1.44A 11.6 68.0 ,10 68.7 6 .1.85A 24.5 124.2 ,10 276.2 5 2.36A 21.6 66.7 ,10 305.8 3 .2.07A 3.9 54.6 ,10 164.1 14 .1.68A 7.2 69.7 ,10 270.3 10 .1.59A 7.4 77.3 ,10 148.4 10 .1.310A 17.0 76.7 ,10 142.0 5 .1.211A 3.7 33.8 ,10 166.3 9 .2.312A 15.1 61.6 ,10 109.1 4 .1.713A 19.7 62.5 ,10 194.8 3 1.914A 5.4 62.2 ,10 178.8 12 .1.915A ,3.0 83.7 ,10 287.5 .28 .1.4Mean #11.4 69.8 ,10 196.9 $9 $1.7

SAAG 5 Serum-ascites albumin gradient.Normal plasma values: plasminogen, 103.9–147.4%; fibrinogen, 183.0–434.0 mg/dl; SAAG$ 1.1 5 transudative ascites.

3220 Agarwal et al. AJG – Vol. 95, No. 11, 2000

are coated with mouse monoclonal anti-FSP antibody.When mixed with a sample containing corresponding anti-gens, latex particles form macroscopic clumps. Titration toan endpoint was performed on all samples; results are re-ported in microcentigrams per milliliter.

Granulocyte count was performed on a standardNeubauer counting chamber using light microscopy and wasreported in cells per cubic millimeter.

Albumin was performed by standard methodology on aVitros analyzer (Johnson and Johnson, Tarrytown, NY).

RESULTS

The results in general are noted in Tables 4–7. Fibrinolyticstates, such as those present in cirrhosis or advanced ma-lignancy, resemble disseminated intravascular coagulation(DIC) and are difficult to distinguish from it. DIC, as de-scribed in many studies of cirrhosis, likely represents afibrinolytic state. In both, PT and aPTT may be elevated,with detectable fragmentD-dimer (12) and consumption offibrinogen.

Fibrinolysis would be quantified ideally by measuringplasmin activity, but this is not feasible with available meth-ods because of substrate nonspecificity. Hence, for patientsin groups A and C, AELT of less than 120 min (13) anddepletion of activatable plasminogen were used as specific(though insensitive) laboratory markers for fibrinolysis.

FragmentD-dimer and plasminogen levels are footprintsof fibrinolytic activity. Fragment D is cleaved from fibrin bythe proteolytic action of plasmin. Because fibrin monomer,generated by thrombin (factor IIa), is cross-linked by factorXIII (a transglutaminase), fragmentD-dimer represents theaction of plasmin upon cross-linked fibrin. Elevated plasmaD-dimer is seen only when significant amounts of fibrinundergo lysis. WhereasD-dimer rises in fibrinolysis, plas-minogen, the zymogen from which plasmin arises, fallsbecause of activation to plasmin. Plasminogen assays, as inour study (see Methods), do not reflect preexisting plasminactivity (14).

Our results show that ascites fluid and plasma of cir-rhotic patients (group A) had fibrinolytic activity (Figs. 1and 2). In all of the ascites samples studied, fragmentD-dimer and FSPs were elevated to levels that, forplasma, would be considered quite high. The ascitesfibrinogen and plasminogen in all of these patients weredepleted to levels that, for plasma, would be consideredquite low. In 12 of 15 (80%) of the group A patients, theascites fluid had a short AELT, providing compellingevidence for fibrinolysis (13).

Additionally, in 14 of 15 (93%) of the patients in groupA, plasmaD-dimer was elevated to levels greater than 0.25mg/ml. The mean plasma plasminogen concentration wasalso low, indicating activation to plasmin.

In the 15 group B patients (Table 6), plasmaD-dimer waselevated to levels greater than 0.25mg/ml in only 33%.Plasminogen activity and fibrinogen were less often de-pleted in this group (Fig. 3).

The ascites fluid and plasma results of the two Group Cpatients is found in Table 7. The ascites fluid had parametersconsistent with fibrinolysis, whereas the plasma did not.

DISCUSSION

Physicians have traditionally regarded ascites as an inertfluid, to be intervened upon primarily because of infectionor discomfort. However, in patients with decompensatedcirrhosis, up to 20 L of ascites is resorbed daily into thecentral circulation via the thoracic duct (15). Our data dem-onstrates that ascites fluid has fibrinolytic activity. Becauseascites reenters the systemic circulation, cirrhotic ascitesmay be a pathological fluid that contributes to a systemicfibrinolytic state.

Ascites fluid, a derivative of plasma, accumulates withinthe abdominal cavity from transudative leakage out of cir-

Table 6. Plasma Data for Patients With Cirrhosis WithoutAscites (Group B)

Patient


PlasmaPlasminogen

(%)

PlasmaFibrinogen

(mg/dl)

1B .25–1.25 109.7 187.02B ,0.25 92.2 293.13B ,0.25 104.6 363.04B ,0.25 86.7 250.55B 0.25–1.25 95.9 205.46B ,0.25 59.0 236.87B ,0.25 78.9 163.38B .1.25 73.7 255.39B ,0.25 144.4 251.010B 0.25–1.25 120.6 208.011B ,0.25 73.4 159.712B ,0.25 120.2 324.913B ,0.25 119.0 332.514B ,0.25 88.2 230.815B .25–1.25 106.6 289.8

Normal plasma values:D-dimer,,0.25 mcg/ml; plasminogen, 103.9–147.4%; fibrin-ogen, 183.0–434.0 mg/dl.

Table 7. Plasma and Ascites Data for Patients With Ascites Without Cirrhosis (Group C)

Patient


PlasmaPg (%)

PlasmaFibrinogen

(mg/dl)

AscitesD-dimer(mg/ml)

AscitesPg (%)

AscitesFibrinogen

(mg/dl)AELT(min)

1C ,0.25 115.0 414.9 .1.25 78.5 ,10 ,1202C ,0.25 114.0 190.5 .1.25 46.1 10.9 ,120

Normal plasma values:D-dimer,,0.25 mg/ml; plasminogen, 103.9–147.4%; fibrinogen, 183.0–434.0 mg/dl; AELT. 120 min; SAAG. 1.1 in both patients.


rhotic liver. This fluid contains plasma proteins (includingthose that participate in coagulation and fibrinolysis) in anonphysiological environment that is in continuous turnoverwith plasma, relatively acellular, and without exposure toendothelium. In an underappreciated study of relationshipsbetween protein compositions of plasma and ascites fluid,Hendersonet al. showed that ascites contains all proteins,over a broad molecular weight range, present in plasma (16).The extent to which these proteins traverse the barrier be-tween hepatic sinusoids and the peritoneal space is relatedinversely to protein molecular weight (16).

Because ascites fluid contains all plasma proteins, it iscapable of forming clot but rarely does soex vivo. This issurprising because ascites is in contact with tissue factorneeded to trigger the tissue factor pathway (Fig. 4), con-firmed by the fact that ascites fluid has evidence of fibrinformation (17, 18). Our data showD-dimer and FSPs inascites.D-dimer, as discussed in Results, is evidence ofplasmin action on cross-linked fibrin.

We conjecture that clot does not form in ascites becauseincreased fibrinolytic activity emulates anticoagulation, asour results suggest. Elevated fragmentD-dimer and FSPs areuniformly present in ascites fluid in concentrations thatwould be considered pathologically elevated if in plasma.Ascites fluid is also uniformly depleted, compared withplasma, of plasminogen and fibrinogen. These results, alongwith the short AELT in 83% of the patients, indicate in-creased fibrinolytic activity in ascites fluid.

Interestingly, Hendersonet al. noted that two proteins,plasminogen and fibrinogen, did not fit their model formolecular weight as a predictor of plasma protein concen-tration in ascites (16). On the basis of their model, themedian plasma to ascites plasminogen concentration ratio(PAPR) should be approximately 4. The actual medianvalue, however, was 7. Our median PAPR is also approxi-mately 7. PAPR of#7 indicates even further depletion ofplasminogen in ascites (Table 5, Fig. 1).This exaggerateddepletion of plasminogen in ascites beyond that predicted onthe basis of molecular weight can be explained by activationto plasmin in the ascites compartment. On activation, 92-kilodalton plasminogen is cleaved to form;70-kD plasmin(12). A similar argument applies for the exaggerated fibrin-ogen depletion we observed.

Although group B patients had cirrhosis but no ascites,we also identified two patients with nonmalignant ascitesbut no cirrhosis (group C). These patients had normalliver histology, no active neoplastic disease, and no ev-idence of infection. Their ascites was ascribed, respec-tively, to constrictive pericarditis and to malnutritionfrom prolonged illness. In both cases, the ascites laterresolved.

The ascites fluid from the group C patients had parame-ters consistent with a fibrinolytic state, suggesting that con-trol ascites, in the absence of liver disease, inherently isfibrinolytic. That these patients were without elevatedplasma D-dimer leads us to speculate that normal liverfunction compensates to maintain normal hemostasis pa-rameters. However, because neither case of ascites wascaused by portal hypertension, there is also no reason tobelieve that such fluid in these patients readily reenterscirculation. Therefore, ascites fluid only in the setting ofliver disease may predispose patients to fibrinolytic activityin their plasma.

Ninety-three percent of the patients with cirrhosis andascites (group A) were demonstrated to have fibrinolyticactivity in their plasma by elevated levels of fragmentD-dimer. This elevation is not incidental. In the absence ofinfection or bleeding, as with our patients,D-dimer is ahighly specific and sensitive test for the presence of fibri-nolysis (19, 20).

Figure 1. Fibrinolytic characteristics of ascites in cirrhotic patients(group A).

Figure 2. Fibrinolytic characteristics of plasma in cirrhotic patientswith ascites (group A).

Figure 3. Fibrinolytic characteristics of plasma in cirrhotic patientswithout ascites (group B).


What distinguished group B patients from group A pa-tients was the absence of ascites in the setting of cirrhosis.These patients were studied as an attempt to separate theeffect of ascites from cirrhosis on plasma coagulation pa-rameters. Overall, group B patients were significantly lesslikely to have plasma parameters resembling those of fibri-nolysis.

Fibrinolysis can be either a primary process or a second-ary process that is in response to certain clinical situations.Secondary elevations inD-dimer are seen after major sur-gery or recent trauma, during sepsis, and as a manifestationof thromboembolic events (12, 19–21). Our study excludedsuch patients, and none of the patients we studied wereacutely ill at time of evaluation. In the absence of secondarycauses for fibrinolysis, as with our patients, elevatedD-dimer is evidence for primary fibrinolysis in the setting ofcirrhosis and ascites.

Our data suggest that ascites contributes to the exagger-ated fibrinolysis typical in advanced liver disease and maythus be partly responsible for the bleeding complicationstypical in cirrhosis. One observation included in the discus-sion above suggests that although ascites fluid has fibrino-lytic activity, it does not cause systemic fibrinolysis in theabsence of liver disease. If ascites leads to a systemicfibrinolytic state, it may happen in situations in which com-promised liver function is permissive. Whether cirrhosisitself, in the absence of ascites, leads to a systemic fibrino-lytic state is unknown. The extent to which hyperfibrinolysiscontributes to mortality from hemorrhage in cirrhotic pa-tients also needs to be determined. Cirrhotic patients, espe-cially those with ascites, might be candidates for treatmentwith antifibrinolytic pharmacotherapy for prevention ofvariceal hemorrhage and for treatment of acute varicealbleeding. This therapeutic option has largely been unexam-ined.

ACKNOWLEDGMENT

The authors gratefully acknowledge the assistance and ad-vice of Dr. Paul Killenberg during this study.

Reprint requests and correspondence:Mark W. Swaim, M.D.,Ph.D., Box 78, University of Texas M.D. Anderson Cancer Center,Department of Gastroenterology, 1515 Halcombe Boulevard,Houston, TX 77030.

Received Apr. 9, 2000; accepted June 29, 2000.

REFERENCES

1. Violli F, Ferro D, Basili S, et al. Hyperfibrinolysis increasesthe risk of gastro-intestinal hemorrhage in patients with ad-vanced cirrhosis. Hepatology 1992;15:672–6.

2. Tytgat G, Collen R, De Vreker R, et al. Investigations on thefibrinolytic system in liver cirrhosis. Acta Haematol 1968;40:265–74.

3. Paramo JA, Rocha E. Hemostasis in advanced liver disease.Semin Thromb Hemost 1993;19(3):184–90.

4. Goodpasture EW. Fibrinolysis in chronic hepatic insuffi-ciency. Johns Hopkins Hosp Bull 1914;25:330–6.

5. Matseshe JW, Beart RW Jr., Bartholomew LG, et al. Fataldisseminated intravascular coagulation after peritone-ovenous shunt for intractable ascites. Mayo Clin Proc 1978;53:526 – 8.

6. Stein SF, Fulenwider T, Ansley JD, et al. Accelerated fibrin-ogen and platelet destruction after peritoneovenous shunting.Arch Int Med 1981;141:1149–51.

7. Harmon DC, Demirjian Z, Ellman L, et al. Disseminatedintravascular coagulation with peritoneovenous shunt. Ann IntMed 1979;90:774–6.

8. Grossi CE, Rousselot LM, Panke WF. Coagulation defects inpatients with cirrhosis of the liver undergoing portasystemicshunts. Am J Surg 1962;104:512–26.

9. Lerner RG, Nelson JC, Corines P, et al. Disseminated intra-vascular coagulation: Complication of LeVeen peritone-ovenous shunts. J Am Med Assoc 1978;240:2064–6.

10. Schwartz ML, Swaim WR, Vogel SB. Coagulopathy follow-ing peritoneovenous shunting. Surgery 1979;85:671–6.

Figure 4. Coagulation pathways.


11. Kowalski E, Kopec M, Niewiarowski S. An evaluation of theeuglobulin method for the determination of fibrinolysis. J ClinPathol 1959;12:215–218.

12. Collen D, Lijnen HR, Verstraete M. Fibrinolytic system and itsdisorders. In: Handin RI, Lux SE, Stossel TP, eds. Blood:Principles and practice of hematology. Philadelphia, PA: J.B.Lippincott, 1995:1261–88.

13. Downie GR, Clifton EE. Method for assay of serum proteo-lytic activity. Proc Soc Exp Biol Med 1949;71:138–41.

14. Reddy KNN, Markus G. Mechanism of activation of humanplasminogen by streptokinase: Presence of active center instreptokinase-plasminogen complex. J Biol Chem 1972;247:1683–91.

15. Dumont AE, Mulholland JH. Flow rate and composition ofthoracic-duct lymph in patients with cirrhosis. N Engl J Med1960;263:471–4.

16. Henderson JM, Stein SF, Kutner M, et al. Analysis of twenty-three plasma proteins in ascites. Ann Surg 1980;192:738–42.

17. Osterud B. Cellular interactions in tissue factor expression by

blood monocytes. Blood Coagul Fibrinolysis 1995;6(suppl1):S20–5 (review).

18. Osterud B, Bajaj MS, Bajaj SP. Sites of tissue factor pathwayinhibitor (TFPI) and tissue factor expression under physiologicand pathologic conditions. On behalf of the Subcommittee onTissue factor Pathway Inhibitor (TFPI) of the Scientific andStandardization Committee of the ISTH. Thromb Haemost1995;73:873–5 (review).

19. Kroneman R, Nieuwenhuizen W, Knot EAR. Monoclonalantibody-based plasma assays for fibrin(ogen) derivatives andtheir clinical relevance. Blood Coagul Fibrinolysis 1990;1:91–111.

20. Charles LA, Edwards T, Macik BG. Evaluation of sensitivityand specificity of sixD-dimer latex assays. Arch Pathol LabMed 1994;118:1102–5.

21. Greenberg CS, Devine DV, McCrae KM. Measurement ofplasma fibrinD-dimer levels with the use of monoclonal an-tibody coupled to latex beads. Am J Clin Pathol 1987;87:94–100.


ascites fluid as a possible origin for hyperfibrinolysis in advanced liver disease

Documents