- Email: [email protected]
Coagulation disturbances following ascites recirculation
Coagulation disturbances following ascites recirculation
To investigate the effects of extra-corporeal ascites recirculaticm on coagulation, several coagulation variables were measured in ascitic return fluid and plasma before, during and after this profedme in 16 patients with diuretic resistant titer. Small but significant reductions in pkoma fibrinogen levels and platelet munts and inneases in plasma X-cdigomer were observed during and after the procedure compared with before. These findings are mnlaent with the :iew that as&es recirculation induces dissemioated iotravascular coagulation although this would appear to be ooly mild and of no cliiical sigoifica”ce in the majority of patients. Althou~ increased levels of activated factm VII were observed in asciticfluid indicating activation of the exbinsic p&way of coagulation, a significant increase io plana activated factor VII during the procedure was not demoostrated. Increased fibrinolytic activity was observed in axitic fluid doe to the preseoee of tissue plastioge” activator. Increased X-aligomer levels were observed in ascitic fluid indicating that lysis of cross-linked fibrin is also an active pmfess within aszites.
Diuretic-resistant ascites se”& a major therapeutic
in portal problem.
hypenensmn prePeritoneo-venous
shunting and extra-corporeal ascites rtxirculatio” are two well ncopised forms of treahned, but both have been reported to cause significant coagulation disturbances (l-3). The mechaniwn of ihese disturbances is unclear. Both primary systemic fibrioolysis and disseminated tiavascular coagulation (DIC) have been implicated.
in-
Recent advances have made it possible to measure individual cmnpweots and metabolites of the fibrinolytic pathway and thereby distinguish between the processes of primary fibrbmlysis and DIC. We hwe measwed a range of magtdation and fibrinolytic variables before, during and after zscites recirctdado” in or&r to s&tdy the mechanism of the coagulopathy aswciated with this procedure. The technique of ascites recirculation provides a anwenient mews of assessing the wagulopathy as the changes produced are only trasient and the procedure is EOshort that significant changes in the patient’s overall condition are unlikely to occur.
Materids and Metheds Sixteen patients with diuretic resistant ascites undergoing ascites recirctdatio” were studied. Diuretic rcsistance was defined as patients who failed to respond to a minimum of ZCMI mg spirowlactone or who developed
cmnplicatians
and 80 mg frusmnide, on these or smaller daosrs
of diuretic. Such mmplicztions wep_ defined as diuretic related awtaemia or eneephalopathy cn diuretic induced eiectrolyte imbalace (hypamtmemia a hyperkaiaemia). Ten of tbe 16 patients had portal hyvrtensio” xconday to alcoholic cirdmsis. The aeti&Sy of the remainder was cryptogenic cirrhosis (n = 3), primary biliary cirrhosis (one patient),
scleroring cholangitis (one) and pa-
tal vencw obstruction caused by carfinoma of the pancreas (one). The latter patient had no evidence of carcinomatous spread to the peritoneal cavity as the cause of the as&es. The technique of ascites recirculation and reinfusion was the same as that described in detail elsewhere. The
218 principle is summarked as follows. A peritoneal dialysis catheter is inserted into the abdominal cavity and ascites extracted by means of a roller pump. The fluid is then passedthrough a semi-permeabk dialysis membranewhich removes water and electrolytes. The reselling proteinrich concentrate (return fluid) is then reinfused into the
were collected into one tenth volume 0.108 M sodium citrate and placed in melting ice. All citraled blood and ascites samples were centrifuged at leD0 x g at 4 “C for 15 min and aliquots were snap-frozen and stored at -70 “C for later analysis. Blood cell counts were meattred on a Tezhnicon H@XII
patient through an ante-cubital vein. The rate of retom of fluid cao be varied but Is usually kept between 600 and 800 ml per h. Tbe procedure is continued until no more fluid can be extracted from the abdominal cavity. The length of the procedure is variable depending upon the rate of fluid removal and the amomtt ef axihs bgt xsw!!y !asts between eight and 24 h. The median amotmt of fluid removed was 6.9 I (range 2.5-14 I). All patients were row tinely screened for cytological and bacteriological evidence of peritonitis before the procedure (4). Subjects with infection orgastm-ittleslbml haemorrhage within the preceding2 weeks were excluded. Venous blood was obtained with minimum stasis,placed
anslyser. ‘fbe prothrombin time (FT) and ~aolin_Fephalio clotting lime (KCCT) were performed by standard techniques (5). The prothrombin and lhrombin times were expressed es ratios of the control value. Fibrinogen was as sayed by the method of Claws. Serum fibrinogen/fibrin degradation products (FDI’s) were measured semiquattii-
into one tenth volume 0.108 M sodium citrate and trans-
vetor activity (t-PA) was measured
fcrred into melting ice. A sample of blood was placed into EDTA. A serum sample was prepared by placing venous blood iato a Iltrombo-Wellcotest F.D.P. tohe (Wellcome Diagnostics, U.K.). Blood samples were taken before ascites recirculation, approximately 5 h into the procedure (i.e., half way through the procedure) and on the morning after recirculation. Samples of both ascilic fluid and concentrated ascites (return fluid) were obtained immediately before recircui&ion and between 4 and 5 h into the Procedure. Samples
tions of plasma by the method of Verheijen (7). Lysis of bovine fibrin was measured according to the method of Brakman (8). Using this method, mtfractionated ascites was tested in the presence or absence of plyclond antibody lo t-PA (Omanon Teknika). Activated factor VII (FViIa) levels.& measured in’plasma and es&a re-
tatively using the ThrombaWellcotest latex agglutination assay (Wellcome Diagnostics, U.K.). X-oligomershigher molecular weight degradation products of cross-linked & brbt (XDPs) - were quantitated in both plasma sod ascites by the ELISA method develowd hv Gaffoev foormal value: co.2 &ml) (6).‘Tis& plasminogen acti-
pimm
in the euglobtdie
frac-
tom fluid according to the method of Kitchen (9). In this method factor VII activity is determined wing bovine and rabbit thromboplastin and the result expressed as the ratio factor VII (bovine thromboplastin): factor VII (rabbit
SeFDPl~~giml)
P”
dur.
part
16
64
126
chrombaplastin) (normal ratio: C1.25). Increased ratios indicate the presence of activated factor VII. To allow for plasma dilutional effects and methodological error, it was decided that a change of at least 20% in fibrinogen and platelet values was required for the change to be deemed genuine. Coagulation variables during and after ascites recirculation were compared with those before the procedure using the Wilcoxon sum rank method of statistical analysis.
The plasma PT ratios, KCCT, fibrinogen. platelet and serum FDP levels before, during and after ascites recirculation ere shown in Table 1. There was no significant difference in the PT ratio during recirculation compared with before, whereas there was a slight but significattt increase
in the ratio after the
procedure @ .S 0.05). llte KCCT was slightly but significantly prolonged dtuing recirwlation compared with before the procedure @ < 0.05). Although most of the patients showed only a modest change in the KCCI the prolongation probably reached statistical sigttiftcance on account of the marked increase in four of the patients. In the majority of patients, the KCXT showed little change after recir~tdation com,,ered with during the procedure. Two of the four patie& who showed irked KCCT prolongation during the txocedure showed a notable shortening of KC‘T aftertheprocedureof 3Oand9s. respectively. There was a significant reduction in plasma fibrinogen levels @ < 0.05) and platelet co”“ts @ < 0.02) both dttring and after recirculation compared with before the procedure. A reduction in plasma fibrinogen levels of at least 20% was observed in eight patients during the procedure and in a fttnber camped
four patients
following
the procedure
with before. There was a reduction in platelet
count of at least 20% in seven patients during the prow dttre and in a further three patients after the procedure compared with before. There we8 a significant
increase
both during and aher recirculation the pmcedttre @ < 0.05).
in serum PDP levels compared
with before
The packed cell volume (PCV) was determined both before and after recirculation to assess the infra-vascular dilutional effect of the return fluid infusion. The mean increase in the plasma v&me following the procedure was 4.31(range-1.0% to +g.o%,n = 11). Activated factor VU level ratios in ascitic return fluid during recirculation and plasma ratios both before and
during the procedure in the five patients studied are shown in Table 2. Increased ratios indicative of activated factor VII were observed in all return fluid samples. Four patients had a nmmal ratio before recirculation. IO only one of these did the ratio increase during the procedure. In this patient there was minimal change in the plasma clotting times altbottgh the fibrinogen level fell from 3.5 g/l to 2.9gIl and theplateletcount bom72to53~1@‘ll. The remaining patient had an elevated activated factor VII level before recirculation. This did not change signi&atttly during the procedure. The fibrin plate Iysis diameters of the ten samples of ascitic fluid analyzed before. recirculation with and without anti-t-PA antibody are shown in Pig. 1. The median dkutt-
eter of the untreated sampleswas 11.4 mm (range, O-15.5 mm). In seven of the eight samples which induced lysis, the lysis wes completely abolished following addition of anti-t-PA antibody. In the remaining sample, IFis was reduced from 14.2 mm tc 2 mm. Fibrinogen was not detectable in any of the asciticfluid samples. There was a marked variation in plasma t-PA levels before the procedure (nxdian level 0.035 U/ml; range, O-12.1 U/ml: n = l3). Ihex was no significant change in levels during or following the procedure (median levels 0.12 U/ml: range, O-4.5: n = 12 and 0.17 U/ml: range, O-1.9: n = 7, respectively). X-oligomer levels in ascitesand return fluid at the start
during recirculation and increasedfurther atier the procedure (median levels 5.7/rgM and ll.Z&ml, respectively). There was no relationship between the total volume of ascitesremoved and subsequently reinhrsed into the patient and the degree of of any of the coagule-
abnormality
tion variables analysed.
Discussiw We have cont%nted that recirculation of escitic fluid in patients with portal hypertension is frequently associated
of and during recirculation and in plasma before, during and after the procedure in the five patients analysed are shown in Fig. 2. Before infusion, the median X-oligomer concentration in the ascitic fluid samples was 7.0 &ml (range, 0.2-17.7yglml). During the procedure, levels increased in four of the five patients to the median level of
with the development of coagulation abnormalities. The cause of the coagulation disturbance is open to debate. Both disseminated intravascular coagulation (DE) and primary fibrinolysis have been implicated. The pathophysiological events occurring in DIC and
13.7&ml (range, 9.6-17.2pglml). At the Stan of recirculation, the median X-oligomer level in the return fluid in the four samples analysed was 6.5 &ml (range, O.l17.0 &ml). Levels increased in all four patients during the txoce&re (median level 14.4 ue/ml. ranne 6.7-20.0 &&. PlasmaX-oligomer levels were elevated in all patients before recirculation with a median value of 1.7 &ml (range, 0.7-10.3). Levels increased in all patients
stimulus activatesthe coagulation systemwithin the circulation and fibrinogen is converted to fibrin by thrombin. Fibrin moleculespolymer& forming micmthrombi. Within the microthrombi fibrin molecules are cross-linked by the action of factor XIII to stabilise the thmntbus. As a ICsuit of intravascular coagulation, platelets are also activated and become incorporated into microthmmbi. The consumption of clotting factors, fibrinogen end platelets
._
I
primary fibrinolysis ere es follows. In DIC, the ttigger
221 during DIG results in B prolongatim of plasma clotting times and a fall in both fibrinogen. levels and ollelet couttts. As a consequence of coagulation activation. the fibrinolytic system is activated and plasmin induced lysis of both fibrinogen and cross-linked fibrin microthrombi ensues resulting in an increase in both fibrinogen and cross-linked fibrin degradation products (FDR and XDPs. respectively). In primary fibrinolysis, the fibrinolytic system is activated directly and plasmin is generated in a free form within the circulation. Plasmin induces Iv& of circtitin~ fibrinogen. ‘Ibis results in a fall in the circulating fibrino. gen level associated with a possible prolongation of the prothrombin, thrombin and kaolin cephalin times &pendent upon the degree of hypofibrbtogenaemia. Activated factor VII is not generated in the circolation and platelet levels are maintained as platelets are not involved in the pathological process. As a consequence of fibiinogenolysis circulating fibrinogen degradation product levels rise, but cro+s-linked fibrin degradation product levels remain normal. Table 3 shows the expected changes io the coagolation variables in DIC campared to primary fibrinolysis. A fall in both fibrinogen levels and platelet counts in association with a rise in FDP levels following inseztion of Le Veen shunts has led several workers to conclu& that intx&ction of ascitic fluid into the systemic circulation induces DIG (W-13). Constitoents of ascitfc fluid which have been implicated in the initiation of DIG include bacterial endoioxin. activated cloning factors and thrombin. It would appear, however, that the major component of procoagulant activity of ascitic fluid derives from the presence of tluomboplastin-like substances which are able to activate the extrinsic coagulation system when introduced into the ckolation
(11,13). It has been suggested
that
these thmmboplastins are released from macrophages attd/or peritoneal cells preseat in ascitic fluid, possibly by the action of bacterial endototin
(13).
Other workers have proposed that the coagulation abnormalities induced by reclrculatioo of ascitic fluid are a result of primary fibrinolysis (14,15). In the study of kho!merick et al. (15). 17 patients undergoing either extncorporeal or paiioneo-venous shunting of ascitic fluid
were studied. Seven patients developed coagulation abnormalities following commencement of the shunt proadue. In these patients there was a signilicam reduction in fibrinogen levels and a significant elevation in serum FDP levels, but both platelet and factor V levels did not alter significantly. There was a signScant reduction in plasminogen and a,antiplasmin kveis in the a&tic fluid of the patients who developed a coagolopathy compared to those who did not. It was concluded from these findigs that in some patients there were significant amounts of fibrinolytic system activators present within titic fluid and that the transfer of such activators into the circulation induced systemic plasminogen activetion and hence. syr temic fibrinolysis. Patrassi et al. also demonstrated increased fihrinolytic activity within the as&c fluid of patients with cirrhosis and proposed that tbk was doe to the presence of tissue plaminogen activators (14). As tissue plasminogen activator (t-PA) had previously been demonstrated within par&l peritoneal cells (16). Patrassi concluded that these cells were the source of the plasminogen activators. Our study has also demonstrated increessd fibrinolytic activity within the titic tloid of patients with portal hypertension as witnessed by the ly!ic effects of ascitic fluid on fibrin ok&es. Tlte auenchine of the lwic effect hv the addition of antibody to t-PA indicates that the fibrbtolytic activity results from the presence of t-PA. The presence of increased levels of activated factor VU in asxltic fluid indicates activation of the extrinsic pathway of coagulation. Flutbumo re. the presence of si@ficam amoonts of X-aligomer within the sscitic fluid analysed before recirculation suggests that IyGs of croxs-linked fibrin is also an active process within aszites. Whether the cross-linked fibrin is fomted as a result of coagulation activation 01 whether
it derives from lysis of a physiologkal
pool of fibrin within the abdominal determined. Interestingly,
during recirculation,
cavity remains to be the tibrinolytic
ac-
tivity of the &tic fluid appeared to increaseas witnessed by the increase in X-oligomer levels in both the ascites and the retom fluid. If is possible that the removal of ascitic fluid from the abdominal cavity acts as a stimulus for increased release of t-PA from peritoneal cells. The reduction
in plasma fibrinogen
levels and platelet
counts associated with the increase in FDP and X-oligomer levelsobserved in mattyofourpatients during and atier ascites recirculation is consistent with the view that intravascular infusion of ascites does induce a degree of DIG. As the PT ratios and KCCTs were only slightly prolonged in the majority of patients. this would suggest that the DIC is usually only mild and of no clinical significance. Furtbemtore. although activated factor VII levels were
222
LT. WILDE ct al. ton as soon as it entered the bload stream. In our experience, recirculation of ascites in patients with portal hypertension results in only a mild disturbance of coagulation in the majority of patients although some patients exhibited a marked derangement of clotting tests
markedly elevated in the ascitic return fluid of the five patients analyzed, only one patient showed a significant increase in plasma levels during the proadure. This again would suggest that there is no significant degree of coagulation pathway activation within the circulation during ascites recirculation. Presumably the activated factor VII within the sscitic ~ethm fluid is rapidly cleared from the circulation by inhibitor proteins and the reticulo-endothelial system. If is likely that a large proportion of the rise in circulatory FDP and X-oligomer observed in our patients may have resulted from passive transfer of these from the
both during and after the procedure. Even so, none of the patients developed an overt bleeding tendency. The cause of the coagulation abnormality woqld appear to be a mild degree of DIC presumably precipitated by the transfer of thmmboplastic substances within the ascitic return fluid into the systemic circulation. Although it is also possible
peritoneal cavity within the ascitic return fluid. The observed pattern of coagulation disturbances in our patients does not suggest that ascites recirculation resulted in a significant primary fibrinolytic response as a fall in fibrinogen level associated with a rise in FDPs did not occur in isolation. Furthermore, there was no significant change in plasma t-PA levels as a consequence of the
that direct transfer of gram-negative bacteria or bacterial endotoxin from the peritoneal cavity into the circulation may have induced DIG, there was no evidence clinically of such an occurrence. Although we have demonstrated fibtiolytic activity within ascitic fluid due to the presence of t-PA which increases during removal of the ascites from the abdominal cavity, it does not appear that there is a sig-
recirculation suggesting that free t-PA within the a&tic return fluid was rapidly neutralized by plasma t-PA inhibi-
nificant transfer of this activity into the systemic circulation resulting in primary Rbrinolysis.
5 Pitney WR, Bmzovic M. lnvertigatio” of the hacmostatic mecha&“. In: Dacie JV. Lewis SM. &is. Practical Haematology. 6th ed. Edinburgh: Churchill-Livi~grtone, 1984:224-247. 6 Gaffney A. Creighton U, Perry MI, Callus M, Thorpe R, Spitz M. Manorlonal antibadies to cmnlinked flbri” degradation emducts (XL-FDP) I. Characterisation and preliminary cvaiuationiopkas”la. Br, H?.emato, 1988;68:83-W. 7 VcrheijenJH. MuUaat E, ChangGT,CluftC, WijngaardsG. A simple, sensitive rpeclm*hota”etdc assay for extrinsic (liS”e. type) plasminoge” activator applicable to measureme”* in plar ma. ThrombHsemostas 1982;48: 266-270. 8 Brakma” P. Fibrin lysir. A standard fibrin plate me,had and a fibrin0lyticaw.y Ofplarminow”. PhDTksrir. Amsterdam: gcheltemaand ffolkcma, 1967. 9 Kitchen 3. Malia RG, GreavesM, Presto” FE. A method f”orthe
To investigate the effects of extra-corporeal ascites recirculaticm on coagulation, several coagulation variables were measured in ascitic return fluid and plasma before, during and after this profedme in 16 patients with diuretic resistant titer. Small but significant reductions in pkoma fibrinogen levels and platelet munts and inneases in plasma X-cdigomer were observed during and after the procedure compared with before. These findings are mnlaent with the :iew that as&es recirculation induces dissemioated iotravascular coagulation although this would appear to be ooly mild and of no cliiical sigoifica”ce in the majority of patients. Althou~ increased levels of activated factm VII were observed in asciticfluid indicating activation of the exbinsic p&way of coagulation, a significant increase io plana activated factor VII during the procedure was not demoostrated. Increased fibrinolytic activity was observed in axitic fluid doe to the preseoee of tissue plastioge” activator. Increased X-aligomer levels were observed in ascitic fluid indicating that lysis of cross-linked fibrin is also an active pmfess within aszites.
Diuretic-resistant ascites se”& a major therapeutic
in portal problem.
hypenensmn prePeritoneo-venous
shunting and extra-corporeal ascites rtxirculatio” are two well ncopised forms of treahned, but both have been reported to cause significant coagulation disturbances (l-3). The mechaniwn of ihese disturbances is unclear. Both primary systemic fibrioolysis and disseminated tiavascular coagulation (DIC) have been implicated.
in-
Recent advances have made it possible to measure individual cmnpweots and metabolites of the fibrinolytic pathway and thereby distinguish between the processes of primary fibrbmlysis and DIC. We hwe measwed a range of magtdation and fibrinolytic variables before, during and after zscites recirctdado” in or&r to s&tdy the mechanism of the coagulopathy aswciated with this procedure. The technique of ascites recirculation provides a anwenient mews of assessing the wagulopathy as the changes produced are only trasient and the procedure is EOshort that significant changes in the patient’s overall condition are unlikely to occur.
Materids and Metheds Sixteen patients with diuretic resistant ascites undergoing ascites recirctdatio” were studied. Diuretic rcsistance was defined as patients who failed to respond to a minimum of ZCMI mg spirowlactone or who developed
cmnplicatians
and 80 mg frusmnide, on these or smaller daosrs
of diuretic. Such mmplicztions wep_ defined as diuretic related awtaemia or eneephalopathy cn diuretic induced eiectrolyte imbalace (hypamtmemia a hyperkaiaemia). Ten of tbe 16 patients had portal hyvrtensio” xconday to alcoholic cirdmsis. The aeti&Sy of the remainder was cryptogenic cirrhosis (n = 3), primary biliary cirrhosis (one patient),
scleroring cholangitis (one) and pa-
tal vencw obstruction caused by carfinoma of the pancreas (one). The latter patient had no evidence of carcinomatous spread to the peritoneal cavity as the cause of the as&es. The technique of ascites recirculation and reinfusion was the same as that described in detail elsewhere. The
218 principle is summarked as follows. A peritoneal dialysis catheter is inserted into the abdominal cavity and ascites extracted by means of a roller pump. The fluid is then passedthrough a semi-permeabk dialysis membranewhich removes water and electrolytes. The reselling proteinrich concentrate (return fluid) is then reinfused into the
were collected into one tenth volume 0.108 M sodium citrate and placed in melting ice. All citraled blood and ascites samples were centrifuged at leD0 x g at 4 “C for 15 min and aliquots were snap-frozen and stored at -70 “C for later analysis. Blood cell counts were meattred on a Tezhnicon H@XII
patient through an ante-cubital vein. The rate of retom of fluid cao be varied but Is usually kept between 600 and 800 ml per h. Tbe procedure is continued until no more fluid can be extracted from the abdominal cavity. The length of the procedure is variable depending upon the rate of fluid removal and the amomtt ef axihs bgt xsw!!y !asts between eight and 24 h. The median amotmt of fluid removed was 6.9 I (range 2.5-14 I). All patients were row tinely screened for cytological and bacteriological evidence of peritonitis before the procedure (4). Subjects with infection orgastm-ittleslbml haemorrhage within the preceding2 weeks were excluded. Venous blood was obtained with minimum stasis,placed
anslyser. ‘fbe prothrombin time (FT) and ~aolin_Fephalio clotting lime (KCCT) were performed by standard techniques (5). The prothrombin and lhrombin times were expressed es ratios of the control value. Fibrinogen was as sayed by the method of Claws. Serum fibrinogen/fibrin degradation products (FDI’s) were measured semiquattii-
into one tenth volume 0.108 M sodium citrate and trans-
vetor activity (t-PA) was measured
fcrred into melting ice. A sample of blood was placed into EDTA. A serum sample was prepared by placing venous blood iato a Iltrombo-Wellcotest F.D.P. tohe (Wellcome Diagnostics, U.K.). Blood samples were taken before ascites recirculation, approximately 5 h into the procedure (i.e., half way through the procedure) and on the morning after recirculation. Samples of both ascilic fluid and concentrated ascites (return fluid) were obtained immediately before recircui&ion and between 4 and 5 h into the Procedure. Samples
tions of plasma by the method of Verheijen (7). Lysis of bovine fibrin was measured according to the method of Brakman (8). Using this method, mtfractionated ascites was tested in the presence or absence of plyclond antibody lo t-PA (Omanon Teknika). Activated factor VII (FViIa) levels.& measured in’plasma and es&a re-
tatively using the ThrombaWellcotest latex agglutination assay (Wellcome Diagnostics, U.K.). X-oligomershigher molecular weight degradation products of cross-linked & brbt (XDPs) - were quantitated in both plasma sod ascites by the ELISA method develowd hv Gaffoev foormal value: co.2 &ml) (6).‘Tis& plasminogen acti-
pimm
in the euglobtdie
frac-
tom fluid according to the method of Kitchen (9). In this method factor VII activity is determined wing bovine and rabbit thromboplastin and the result expressed as the ratio factor VII (bovine thromboplastin): factor VII (rabbit
SeFDPl~~giml)
P”
dur.
part
16
64
126
chrombaplastin) (normal ratio: C1.25). Increased ratios indicate the presence of activated factor VII. To allow for plasma dilutional effects and methodological error, it was decided that a change of at least 20% in fibrinogen and platelet values was required for the change to be deemed genuine. Coagulation variables during and after ascites recirculation were compared with those before the procedure using the Wilcoxon sum rank method of statistical analysis.
The plasma PT ratios, KCCT, fibrinogen. platelet and serum FDP levels before, during and after ascites recirculation ere shown in Table 1. There was no significant difference in the PT ratio during recirculation compared with before, whereas there was a slight but significattt increase
in the ratio after the
procedure @ .S 0.05). llte KCCT was slightly but significantly prolonged dtuing recirwlation compared with before the procedure @ < 0.05). Although most of the patients showed only a modest change in the KCCI the prolongation probably reached statistical sigttiftcance on account of the marked increase in four of the patients. In the majority of patients, the KCXT showed little change after recir~tdation com,,ered with during the procedure. Two of the four patie& who showed irked KCCT prolongation during the txocedure showed a notable shortening of KC‘T aftertheprocedureof 3Oand9s. respectively. There was a significant reduction in plasma fibrinogen levels @ < 0.05) and platelet co”“ts @ < 0.02) both dttring and after recirculation compared with before the procedure. A reduction in plasma fibrinogen levels of at least 20% was observed in eight patients during the procedure and in a fttnber camped
four patients
following
the procedure
with before. There was a reduction in platelet
count of at least 20% in seven patients during the prow dttre and in a further three patients after the procedure compared with before. There we8 a significant
increase
both during and aher recirculation the pmcedttre @ < 0.05).
in serum PDP levels compared
with before
The packed cell volume (PCV) was determined both before and after recirculation to assess the infra-vascular dilutional effect of the return fluid infusion. The mean increase in the plasma v&me following the procedure was 4.31(range-1.0% to +g.o%,n = 11). Activated factor VU level ratios in ascitic return fluid during recirculation and plasma ratios both before and
during the procedure in the five patients studied are shown in Table 2. Increased ratios indicative of activated factor VII were observed in all return fluid samples. Four patients had a nmmal ratio before recirculation. IO only one of these did the ratio increase during the procedure. In this patient there was minimal change in the plasma clotting times altbottgh the fibrinogen level fell from 3.5 g/l to 2.9gIl and theplateletcount bom72to53~1@‘ll. The remaining patient had an elevated activated factor VII level before recirculation. This did not change signi&atttly during the procedure. The fibrin plate Iysis diameters of the ten samples of ascitic fluid analyzed before. recirculation with and without anti-t-PA antibody are shown in Pig. 1. The median dkutt-
eter of the untreated sampleswas 11.4 mm (range, O-15.5 mm). In seven of the eight samples which induced lysis, the lysis wes completely abolished following addition of anti-t-PA antibody. In the remaining sample, IFis was reduced from 14.2 mm tc 2 mm. Fibrinogen was not detectable in any of the asciticfluid samples. There was a marked variation in plasma t-PA levels before the procedure (nxdian level 0.035 U/ml; range, O-12.1 U/ml: n = l3). Ihex was no significant change in levels during or following the procedure (median levels 0.12 U/ml: range, O-4.5: n = 12 and 0.17 U/ml: range, O-1.9: n = 7, respectively). X-oligomer levels in ascitesand return fluid at the start
during recirculation and increasedfurther atier the procedure (median levels 5.7/rgM and ll.Z&ml, respectively). There was no relationship between the total volume of ascitesremoved and subsequently reinhrsed into the patient and the degree of of any of the coagule-
abnormality
tion variables analysed.
Discussiw We have cont%nted that recirculation of escitic fluid in patients with portal hypertension is frequently associated
of and during recirculation and in plasma before, during and after the procedure in the five patients analysed are shown in Fig. 2. Before infusion, the median X-oligomer concentration in the ascitic fluid samples was 7.0 &ml (range, 0.2-17.7yglml). During the procedure, levels increased in four of the five patients to the median level of
with the development of coagulation abnormalities. The cause of the coagulation disturbance is open to debate. Both disseminated intravascular coagulation (DE) and primary fibrinolysis have been implicated. The pathophysiological events occurring in DIC and
13.7&ml (range, 9.6-17.2pglml). At the Stan of recirculation, the median X-oligomer level in the return fluid in the four samples analysed was 6.5 &ml (range, O.l17.0 &ml). Levels increased in all four patients during the txoce&re (median level 14.4 ue/ml. ranne 6.7-20.0 &&. PlasmaX-oligomer levels were elevated in all patients before recirculation with a median value of 1.7 &ml (range, 0.7-10.3). Levels increased in all patients
stimulus activatesthe coagulation systemwithin the circulation and fibrinogen is converted to fibrin by thrombin. Fibrin moleculespolymer& forming micmthrombi. Within the microthrombi fibrin molecules are cross-linked by the action of factor XIII to stabilise the thmntbus. As a ICsuit of intravascular coagulation, platelets are also activated and become incorporated into microthmmbi. The consumption of clotting factors, fibrinogen end platelets
._
I
primary fibrinolysis ere es follows. In DIC, the ttigger
221 during DIG results in B prolongatim of plasma clotting times and a fall in both fibrinogen. levels and ollelet couttts. As a consequence of coagulation activation. the fibrinolytic system is activated and plasmin induced lysis of both fibrinogen and cross-linked fibrin microthrombi ensues resulting in an increase in both fibrinogen and cross-linked fibrin degradation products (FDR and XDPs. respectively). In primary fibrinolysis, the fibrinolytic system is activated directly and plasmin is generated in a free form within the circulation. Plasmin induces Iv& of circtitin~ fibrinogen. ‘Ibis results in a fall in the circulating fibrino. gen level associated with a possible prolongation of the prothrombin, thrombin and kaolin cephalin times &pendent upon the degree of hypofibrbtogenaemia. Activated factor VII is not generated in the circolation and platelet levels are maintained as platelets are not involved in the pathological process. As a consequence of fibiinogenolysis circulating fibrinogen degradation product levels rise, but cro+s-linked fibrin degradation product levels remain normal. Table 3 shows the expected changes io the coagolation variables in DIC campared to primary fibrinolysis. A fall in both fibrinogen levels and platelet counts in association with a rise in FDP levels following inseztion of Le Veen shunts has led several workers to conclu& that intx&ction of ascitic fluid into the systemic circulation induces DIG (W-13). Constitoents of ascitfc fluid which have been implicated in the initiation of DIG include bacterial endoioxin. activated cloning factors and thrombin. It would appear, however, that the major component of procoagulant activity of ascitic fluid derives from the presence of tluomboplastin-like substances which are able to activate the extrinsic coagulation system when introduced into the ckolation
(11,13). It has been suggested
that
these thmmboplastins are released from macrophages attd/or peritoneal cells preseat in ascitic fluid, possibly by the action of bacterial endototin
(13).
Other workers have proposed that the coagulation abnormalities induced by reclrculatioo of ascitic fluid are a result of primary fibrinolysis (14,15). In the study of kho!merick et al. (15). 17 patients undergoing either extncorporeal or paiioneo-venous shunting of ascitic fluid
were studied. Seven patients developed coagulation abnormalities following commencement of the shunt proadue. In these patients there was a signilicam reduction in fibrinogen levels and a significant elevation in serum FDP levels, but both platelet and factor V levels did not alter significantly. There was a signScant reduction in plasminogen and a,antiplasmin kveis in the a&tic fluid of the patients who developed a coagolopathy compared to those who did not. It was concluded from these findigs that in some patients there were significant amounts of fibrinolytic system activators present within titic fluid and that the transfer of such activators into the circulation induced systemic plasminogen activetion and hence. syr temic fibrinolysis. Patrassi et al. also demonstrated increased fihrinolytic activity within the as&c fluid of patients with cirrhosis and proposed that tbk was doe to the presence of tissue plaminogen activators (14). As tissue plasminogen activator (t-PA) had previously been demonstrated within par&l peritoneal cells (16). Patrassi concluded that these cells were the source of the plasminogen activators. Our study has also demonstrated increessd fibrinolytic activity within the titic tloid of patients with portal hypertension as witnessed by the ly!ic effects of ascitic fluid on fibrin ok&es. Tlte auenchine of the lwic effect hv the addition of antibody to t-PA indicates that the fibrbtolytic activity results from the presence of t-PA. The presence of increased levels of activated factor VU in asxltic fluid indicates activation of the extrinsic pathway of coagulation. Flutbumo re. the presence of si@ficam amoonts of X-aligomer within the sscitic fluid analysed before recirculation suggests that IyGs of croxs-linked fibrin is also an active process within aszites. Whether the cross-linked fibrin is fomted as a result of coagulation activation 01 whether
it derives from lysis of a physiologkal
pool of fibrin within the abdominal determined. Interestingly,
during recirculation,
cavity remains to be the tibrinolytic
ac-
tivity of the &tic fluid appeared to increaseas witnessed by the increase in X-oligomer levels in both the ascites and the retom fluid. If is possible that the removal of ascitic fluid from the abdominal cavity acts as a stimulus for increased release of t-PA from peritoneal cells. The reduction
in plasma fibrinogen
levels and platelet
counts associated with the increase in FDP and X-oligomer levelsobserved in mattyofourpatients during and atier ascites recirculation is consistent with the view that intravascular infusion of ascites does induce a degree of DIG. As the PT ratios and KCCTs were only slightly prolonged in the majority of patients. this would suggest that the DIC is usually only mild and of no clinical significance. Furtbemtore. although activated factor VII levels were
222
LT. WILDE ct al. ton as soon as it entered the bload stream. In our experience, recirculation of ascites in patients with portal hypertension results in only a mild disturbance of coagulation in the majority of patients although some patients exhibited a marked derangement of clotting tests
markedly elevated in the ascitic return fluid of the five patients analyzed, only one patient showed a significant increase in plasma levels during the proadure. This again would suggest that there is no significant degree of coagulation pathway activation within the circulation during ascites recirculation. Presumably the activated factor VII within the sscitic ~ethm fluid is rapidly cleared from the circulation by inhibitor proteins and the reticulo-endothelial system. If is likely that a large proportion of the rise in circulatory FDP and X-oligomer observed in our patients may have resulted from passive transfer of these from the
both during and after the procedure. Even so, none of the patients developed an overt bleeding tendency. The cause of the coagulation abnormality woqld appear to be a mild degree of DIC presumably precipitated by the transfer of thmmboplastic substances within the ascitic return fluid into the systemic circulation. Although it is also possible
peritoneal cavity within the ascitic return fluid. The observed pattern of coagulation disturbances in our patients does not suggest that ascites recirculation resulted in a significant primary fibrinolytic response as a fall in fibrinogen level associated with a rise in FDPs did not occur in isolation. Furthermore, there was no significant change in plasma t-PA levels as a consequence of the
that direct transfer of gram-negative bacteria or bacterial endotoxin from the peritoneal cavity into the circulation may have induced DIG, there was no evidence clinically of such an occurrence. Although we have demonstrated fibtiolytic activity within ascitic fluid due to the presence of t-PA which increases during removal of the ascites from the abdominal cavity, it does not appear that there is a sig-
recirculation suggesting that free t-PA within the a&tic return fluid was rapidly neutralized by plasma t-PA inhibi-
nificant transfer of this activity into the systemic circulation resulting in primary Rbrinolysis.
5 Pitney WR, Bmzovic M. lnvertigatio” of the hacmostatic mecha&“. In: Dacie JV. Lewis SM. &is. Practical Haematology. 6th ed. Edinburgh: Churchill-Livi~grtone, 1984:224-247. 6 Gaffney A. Creighton U, Perry MI, Callus M, Thorpe R, Spitz M. Manorlonal antibadies to cmnlinked flbri” degradation emducts (XL-FDP) I. Characterisation and preliminary cvaiuationiopkas”la. Br, H?.emato, 1988;68:83-W. 7 VcrheijenJH. MuUaat E, ChangGT,CluftC, WijngaardsG. A simple, sensitive rpeclm*hota”etdc assay for extrinsic (liS”e. type) plasminoge” activator applicable to measureme”* in plar ma. ThrombHsemostas 1982;48: 266-270. 8 Brakma” P. Fibrin lysir. A standard fibrin plate me,had and a fibrin0lyticaw.y Ofplarminow”. PhDTksrir. Amsterdam: gcheltemaand ffolkcma, 1967. 9 Kitchen 3. Malia RG, GreavesM, Presto” FE. A method f”orthe