BritishJournal .fHaematology, 1976,34, 427.
Abnormal Fibrin Polymerization in Liver Disease G. GREEN, JEAN M. THOMSON, I. W. DYMOCK* AND L. POLLER Departments of Medicine and Haematology , University Hospital of South Manchester (Received 13 October 1975 ; acceptedfor publication 7January 1976)
SUMMARY.Although there have been isolated reports of an acquired abnormal fibrinogen in patients with liver disease, its frequency and clinical significance is not known. In this study 121 consecutive patients with a wide spectrum of hepatic disorders were screened for abnormal fibrin polymerization. A simple colorimetric method using Reptilase was employed. Of 32 patients with proven cirrhosis, 16 (50%) showed abnormal fibrin polymerization. The incidence in decompensated alcoholic cirrhosis was particularly high. The abnormality was also detected in all patients with acute liver failure and seven of 15 with chronic active liver disease. Clinical improvement often correlated with its disappearance. Two patients with primary liver cell tumours demonstrated the abnormal polymerization. In patients with bleeding oesophageal varices the detection of abnormal fibrin polymerization was associated with a poor prognosis. None of the patients with surgical obstructive jaundice (26) or miscellaneous liver disorders (3 7) had abnormal fibrin polymerization. The occurrence of abnormal fibrin polymerization in liver disease is more frequent than previously suspected and usually signifies severe primary hepatocellular dysfunction. Evidence is presented to support the presence of a primary abnormality of fibrinogen as the cause of impaired fibrin monomer polymerization. The disturbance of blood coagulation in liver disease is complex and not fully understood. One puzzling feature is the common occurrence of a prolonged thrombin time in these patients. In acute and chronic disturbance of hepatocellular function the fibrinogen content of the blood is usually normal (Hallen & Nilsson, 1964; Donaldson et al, 1969). Although fibrinogen breakdown products are increased in some patients because of disseminated intravascular coagulation (Tytgat et a!, 1968; Horder, 1969;Rake et al, 1970)this is relatively uncommon and does not provide an adequate explanation. The possibility that a qualitative rather than a quantitative abnormality of fibrinogen might be the cause of the prolonged thrombin time has therefore been investigated. The polymerization of fibrin has been studied in 121 consecutive patients presenting with clinical or laboratory evidence of liver dysfunction. PATIENTS AND METHODS Venous blood ( 5 ml) was collected into 0.5 ml of.3.13% trisodium citrate solution from 34 healthy male and female hospital staff and from 121 in-patients at the same time as venePresent address: Stepping Hill Hospital, Stockport, Cheshire. Correspondence:Dr L. Poller, WHO Collaborating Centre for Anticoagulant Control Reagents, Withington Hospital, West Didsbury, Manchester Mzo 8LR.
427
G. Green et af TABLE I. Clinical groups showing numbers of patients and controls studied and incidence of abnormal fibrin polymerization
Group I 2
3 4
S 6 7 Total
Diagnosis
No. studied
Controls Cirrhosis Chronic active liver disease Obstructive jaundice Primary liver tumour Acute hepatic failure Miscellaneous liver disorders
34
Incidence of abnormal FMP*
32
IS
26 3 8 37 ISS
* FMP = fibrin monomer poIymerization.
TABU 11. Diagnosis and number of patients comprising ‘miscellaneous liver disorder’ (Group 7)
Diagnosis
Secondary malignancy Gilberts syndrome Congested liver due to heart failure Rheumatoid arthritis Non-cirrhotic alcoholic liver damage Septicaemia Miliary tuberculosis Brucellosis
No. of patients 14
5 3 4 7 2
I I
puncture for routine liver function tests. In all patients the tests were performed before a clinical diagnosis had been established. The final diagnoses reached in the 121 patients studied are shown in Table I. Thirty-two patients had cirrhosis, 15 chronic active liver disease, 26 surgical obstructive jaundice, three primary liver tumours and eight acute liver failure. The remaining 37 had disturbance of liver function secondary to systemic disorders such as heart failure, septicaemia, reticulosis and metastatic liver disease (Table 11). To test for the presence of delayed fibrin polymerization the following investigations were performed, and means+ SDs are quoted unless otherwise stated: Thrombin-fibrinogen time (observed normal range I 1-13 s) using bovine thrombin (Parke Davis), diluted I in 10with normal saline. Reptilase time (Funk et al, 1971), (observed normal range 11-14s). Clottablefibrinogen (Ingram & Matchett, 1960) a rapid semi-quantitative method (observed normal range 1.5-3.0 g/l.).
Fibrin Polymerization in Liver Disease
429
Quantitative fibrinogen (Ratnoff & Menzie, 195I) modified by leaving the plasma to clot overnight following the addition of thrombin. Fibrin polymerization was measured by a colorimetric technique (Soria et al, 1969) modified as follows: plasma was diluted in 0.9% saline, according to the fibrinogen level measured by the Ingram & Matchett technique, to give a fibrinogen concentration of 0.6 g/l. Reptilase (0.1ml) was added to 2.5 nll of the diluted plasma in a colorimeter. The initial optical density reading was adjusted for each patient to compensate for any colour or turbidity differences, e.g. bilirubinaemia. RESULTS Normal Controls (Group I) In all 32 controls the maximum rise in optical density in the fibrin polymerization test had occurred by 10 min. Fig I shows the typical appearance of the curve obtained by plotting optical density against time in a normal control. The optical density of the clot at 10min was defined as the H'O value. For statistical analysis this was the value taken to compare fibrin polymerization in the controls and patients. As the initial concentration of fibrinogen was made constant by our technique the height of the curve at this time reflects the degree of aggregation of the fibrin. The calculated normal range for the H'O value was based on twice the standard deviation from the mean of the 34 controls. Abnormal fibrin polymerization was taken as a value below the lower limit ofthe normal range. Fig 2 demonstrates an example of impaired fibrin polymerization in a patient, shown by such a reduction in the H'O level. Fig 3 illustrates the H'O values observed in all subjects studied. The mean value for the controls was 0.66 u with a standard deviation of k0.15. An H'O value in a patient of less than 0.36 u was considered abnormal. Cirrhosis (Group 2 ) This group was composed of 32 patients with histologically proven cirrhosis, in 15 of whom there was an alcoholic aetiology.
~
Abnormal range
0.6
.-> f
d
.-c0
04
0" 0.2
0
2
4
6 Time ( min 1
8
10
FIG I. Fibrin polymerization curve of a healthy control.
12
G. Green et a1
430
Abnormal range
.-f In
04
Q
.Y L
0”
0-2
2
0
4
6 Time ( min)
8
10
12
FIG 2. Abnormal fibrin polymerization curve of a patient.
.. 3
PI
0.4
6
..
.
F
.f . . ....
-
F
f
F
0.2
0 Group Group Group 1 2 3
Group Group Group Group 4 5 6 7
FIG 3. Hl0 values observed in the controls and patients of each diagnostic group (groups defined in Table I).
Fibrin Polymerization in Liver Disease
43 1
Thrombin-fibrinogen time. There was a significant prolongation in patients with cirrhosis, 28 of whom demonstrated an abnormal result (mean 18.0k5.0 s, P< 0.001). Reptilase time. This was significantly prolonged, 26 patients showing an abnormality (mean 17.5 3.0 s, P< 0.001). Quantitative fibrinogen. No significant difference from controls was observed. Fibrin polymerization. An abnormality was demonstrated in 16 patients of this group. Nine suffered from deconipensated alcoholic cirrhosis. In all cases a significant prolongation of both the thrombin and Reptilase times was observed. The overall incidence of abnormal fibrin polymerization was 50%, its frequency in the alcoholic group being even greater (60%); see Table I. The mean HIo value for the 32 patients (0.32k0.23 u) was significantly lower than the controls (P< 0.001).Thirteen patients in the cirrhotic group were bleeding from gastro-oesophageal varices. In six of these patients an abnormality of fibrin polymerization was detected prior to any therapy being given. Of this latter group five patients died from continuous haemorrhage despite medical and/or surgical intervention, whereas the remaining seven patients, in whom fibrin polymerization was normal, survived (Fig 4). However, no difference was observed in the mean H'O values of patients with or without bleeding varices. Chronic Active Liver Diseuse (Group 3) Thrombin-fibrinogen time. Fourteen of the I 5 patients exhibited a prolongation (mean 16.0+ 3.0 s).
R e p t i h e time. This was abnormally prolonged in 11 patients (mean 16.5-t 3.0 s). Quantitative fibrinogen. No significant difference from controls was observed. Fibrin polymerization. Of the 15 patients studied seven (47%) had HO ' values less than the ' of the group was 0.45k0.32 u. accepted lower limit of the normal range. The mean HO
Abnormal range
%I- t 04
t
0.2
0
Controls
Cirrhotics Without bleeding
With bleeding
varices
mrices
FIG 4.Hl0 values in controls and cirrhotic patients with and without bleeding varices.
G. Green et al 432 This value was significantly lower than the control mean (P-co.001). One patient had a relatively high value of 1.3 u (see Discussion). Obstructive Jaundice (Group 4) Twenty-six patients were found to have had ‘surgical’ obstructive jaundice for periods varying from I week to I year. Thrombin-fibrinogen time. Twenty of the patients in this group had abnormal prolongation (mean 15.of2.5 s). Reptilase time. This was lengthened in 17patients (mean 15.5f 2.5 s). Quantitative fibrinogen. The mean value was significantly higher than the control group (P
Abnormal Fibrin Polymerization in Liver Disease G. GREEN, JEAN M. THOMSON, I. W. DYMOCK* AND L. POLLER Departments of Medicine and Haematology , University Hospital of South Manchester (Received 13 October 1975 ; acceptedfor publication 7January 1976)
SUMMARY.Although there have been isolated reports of an acquired abnormal fibrinogen in patients with liver disease, its frequency and clinical significance is not known. In this study 121 consecutive patients with a wide spectrum of hepatic disorders were screened for abnormal fibrin polymerization. A simple colorimetric method using Reptilase was employed. Of 32 patients with proven cirrhosis, 16 (50%) showed abnormal fibrin polymerization. The incidence in decompensated alcoholic cirrhosis was particularly high. The abnormality was also detected in all patients with acute liver failure and seven of 15 with chronic active liver disease. Clinical improvement often correlated with its disappearance. Two patients with primary liver cell tumours demonstrated the abnormal polymerization. In patients with bleeding oesophageal varices the detection of abnormal fibrin polymerization was associated with a poor prognosis. None of the patients with surgical obstructive jaundice (26) or miscellaneous liver disorders (3 7) had abnormal fibrin polymerization. The occurrence of abnormal fibrin polymerization in liver disease is more frequent than previously suspected and usually signifies severe primary hepatocellular dysfunction. Evidence is presented to support the presence of a primary abnormality of fibrinogen as the cause of impaired fibrin monomer polymerization. The disturbance of blood coagulation in liver disease is complex and not fully understood. One puzzling feature is the common occurrence of a prolonged thrombin time in these patients. In acute and chronic disturbance of hepatocellular function the fibrinogen content of the blood is usually normal (Hallen & Nilsson, 1964; Donaldson et al, 1969). Although fibrinogen breakdown products are increased in some patients because of disseminated intravascular coagulation (Tytgat et a!, 1968; Horder, 1969;Rake et al, 1970)this is relatively uncommon and does not provide an adequate explanation. The possibility that a qualitative rather than a quantitative abnormality of fibrinogen might be the cause of the prolonged thrombin time has therefore been investigated. The polymerization of fibrin has been studied in 121 consecutive patients presenting with clinical or laboratory evidence of liver dysfunction. PATIENTS AND METHODS Venous blood ( 5 ml) was collected into 0.5 ml of.3.13% trisodium citrate solution from 34 healthy male and female hospital staff and from 121 in-patients at the same time as venePresent address: Stepping Hill Hospital, Stockport, Cheshire. Correspondence:Dr L. Poller, WHO Collaborating Centre for Anticoagulant Control Reagents, Withington Hospital, West Didsbury, Manchester Mzo 8LR.
427
G. Green et af TABLE I. Clinical groups showing numbers of patients and controls studied and incidence of abnormal fibrin polymerization
Group I 2
3 4
S 6 7 Total
Diagnosis
No. studied
Controls Cirrhosis Chronic active liver disease Obstructive jaundice Primary liver tumour Acute hepatic failure Miscellaneous liver disorders
34
Incidence of abnormal FMP*
32
IS
26 3 8 37 ISS
* FMP = fibrin monomer poIymerization.
TABU 11. Diagnosis and number of patients comprising ‘miscellaneous liver disorder’ (Group 7)
Diagnosis
Secondary malignancy Gilberts syndrome Congested liver due to heart failure Rheumatoid arthritis Non-cirrhotic alcoholic liver damage Septicaemia Miliary tuberculosis Brucellosis
No. of patients 14
5 3 4 7 2
I I
puncture for routine liver function tests. In all patients the tests were performed before a clinical diagnosis had been established. The final diagnoses reached in the 121 patients studied are shown in Table I. Thirty-two patients had cirrhosis, 15 chronic active liver disease, 26 surgical obstructive jaundice, three primary liver tumours and eight acute liver failure. The remaining 37 had disturbance of liver function secondary to systemic disorders such as heart failure, septicaemia, reticulosis and metastatic liver disease (Table 11). To test for the presence of delayed fibrin polymerization the following investigations were performed, and means+ SDs are quoted unless otherwise stated: Thrombin-fibrinogen time (observed normal range I 1-13 s) using bovine thrombin (Parke Davis), diluted I in 10with normal saline. Reptilase time (Funk et al, 1971), (observed normal range 11-14s). Clottablefibrinogen (Ingram & Matchett, 1960) a rapid semi-quantitative method (observed normal range 1.5-3.0 g/l.).
Fibrin Polymerization in Liver Disease
429
Quantitative fibrinogen (Ratnoff & Menzie, 195I) modified by leaving the plasma to clot overnight following the addition of thrombin. Fibrin polymerization was measured by a colorimetric technique (Soria et al, 1969) modified as follows: plasma was diluted in 0.9% saline, according to the fibrinogen level measured by the Ingram & Matchett technique, to give a fibrinogen concentration of 0.6 g/l. Reptilase (0.1ml) was added to 2.5 nll of the diluted plasma in a colorimeter. The initial optical density reading was adjusted for each patient to compensate for any colour or turbidity differences, e.g. bilirubinaemia. RESULTS Normal Controls (Group I) In all 32 controls the maximum rise in optical density in the fibrin polymerization test had occurred by 10 min. Fig I shows the typical appearance of the curve obtained by plotting optical density against time in a normal control. The optical density of the clot at 10min was defined as the H'O value. For statistical analysis this was the value taken to compare fibrin polymerization in the controls and patients. As the initial concentration of fibrinogen was made constant by our technique the height of the curve at this time reflects the degree of aggregation of the fibrin. The calculated normal range for the H'O value was based on twice the standard deviation from the mean of the 34 controls. Abnormal fibrin polymerization was taken as a value below the lower limit ofthe normal range. Fig 2 demonstrates an example of impaired fibrin polymerization in a patient, shown by such a reduction in the H'O level. Fig 3 illustrates the H'O values observed in all subjects studied. The mean value for the controls was 0.66 u with a standard deviation of k0.15. An H'O value in a patient of less than 0.36 u was considered abnormal. Cirrhosis (Group 2 ) This group was composed of 32 patients with histologically proven cirrhosis, in 15 of whom there was an alcoholic aetiology.
~
Abnormal range
0.6
.-> f
d
.-c0
04
0" 0.2
0
2
4
6 Time ( min 1
8
10
FIG I. Fibrin polymerization curve of a healthy control.
12
G. Green et a1
430
Abnormal range
.-f In
04
Q
.Y L
0”
0-2
2
0
4
6 Time ( min)
8
10
12
FIG 2. Abnormal fibrin polymerization curve of a patient.
.. 3
PI
0.4
6
..
.
F
.f . . ....
-
F
f
F
0.2
0 Group Group Group 1 2 3
Group Group Group Group 4 5 6 7
FIG 3. Hl0 values observed in the controls and patients of each diagnostic group (groups defined in Table I).
Fibrin Polymerization in Liver Disease
43 1
Thrombin-fibrinogen time. There was a significant prolongation in patients with cirrhosis, 28 of whom demonstrated an abnormal result (mean 18.0k5.0 s, P< 0.001). Reptilase time. This was significantly prolonged, 26 patients showing an abnormality (mean 17.5 3.0 s, P< 0.001). Quantitative fibrinogen. No significant difference from controls was observed. Fibrin polymerization. An abnormality was demonstrated in 16 patients of this group. Nine suffered from deconipensated alcoholic cirrhosis. In all cases a significant prolongation of both the thrombin and Reptilase times was observed. The overall incidence of abnormal fibrin polymerization was 50%, its frequency in the alcoholic group being even greater (60%); see Table I. The mean HIo value for the 32 patients (0.32k0.23 u) was significantly lower than the controls (P< 0.001).Thirteen patients in the cirrhotic group were bleeding from gastro-oesophageal varices. In six of these patients an abnormality of fibrin polymerization was detected prior to any therapy being given. Of this latter group five patients died from continuous haemorrhage despite medical and/or surgical intervention, whereas the remaining seven patients, in whom fibrin polymerization was normal, survived (Fig 4). However, no difference was observed in the mean H'O values of patients with or without bleeding varices. Chronic Active Liver Diseuse (Group 3) Thrombin-fibrinogen time. Fourteen of the I 5 patients exhibited a prolongation (mean 16.0+ 3.0 s).
R e p t i h e time. This was abnormally prolonged in 11 patients (mean 16.5-t 3.0 s). Quantitative fibrinogen. No significant difference from controls was observed. Fibrin polymerization. Of the 15 patients studied seven (47%) had HO ' values less than the ' of the group was 0.45k0.32 u. accepted lower limit of the normal range. The mean HO
Abnormal range
%I- t 04
t
0.2
0
Controls
Cirrhotics Without bleeding
With bleeding
varices
mrices
FIG 4.Hl0 values in controls and cirrhotic patients with and without bleeding varices.
G. Green et al 432 This value was significantly lower than the control mean (P-co.001). One patient had a relatively high value of 1.3 u (see Discussion). Obstructive Jaundice (Group 4) Twenty-six patients were found to have had ‘surgical’ obstructive jaundice for periods varying from I week to I year. Thrombin-fibrinogen time. Twenty of the patients in this group had abnormal prolongation (mean 15.of2.5 s). Reptilase time. This was lengthened in 17patients (mean 15.5f 2.5 s). Quantitative fibrinogen. The mean value was significantly higher than the control group (P
