Britishjoumal cf Haemafology, 1976,34, 655.

Relationship of Factor VIII-like Antigen (VIII AGN) and Clot Promoting Activity (VIII AHF) as Measured by Oneand Two-Stage Assays in Patients with Liver Disease J. S. ROGERS* AND M. ELAINEEYSTER Division of Hematology, Department of Medicine, The Pennsylvania State University School of Medicine, Hershey, Pennsylvania (Received 29 October 1975; accepted-for publication 8 April 1976) SUMMARY. W e recently observed an increase in factor-VIII clot promoting activity as measured by a one-stage assay (VIII AHF,) in a haemophiliac with hepatitis. However, VIII AHF as measured by a two-stage assay (VIII AHF2) was 0.013 u/ml a t a time when VIII AHF, measured 0.38 ujml.

We then studied seven non-haemophiliacs with liver disease, and attempted to correlate the levels of VIII AHF, and VIII AHF, with factor VIII-like antigen (VIII AGN) as measured by quantitative immunoelectrophoresis. In four of the seven patients, disproportionate elevations of VIII AHF, compared to VIII AHF, were found. Furthermore, VIII AHF, values correlated well with VIII AGN values. No such discrepancy was apparent in four normal control subjects. These findings emphasize the necessity for performing two-stage assays in haemophiliacs as well as non-haemophiliacs with liver disease to assess factor-VIII levels. In addition, they suggest that confirmation of the diagnosis of haemophilia may not be possible in the haemophiliac with hepatitis unless VIII AHF, determinations are performed. The reason for the disparity between VIII AHF, and VIII AHF2 levels is not apparent. However, the correlation of VIII AGN and VIII AHF, levels in the non-haemophiliacs with liver disease provides further support for the concept that VIII AGN and VIII AHF are closely related or identical molecular entities. Factor-VIII levels have been shown to be increased in vivo following exercise (Rizza, 1961), administration of adrenalin (Prentice et al, 1972), during pregnancy (Jorpes & Ramgren, 1962) and in patients with hepatitis and cirrhosis (Baele et a!, 1973; Green & Ratnoff, 1974; Herold & Straub, 1954; Koller, 1973; Meili & Straub, 1970; Van Outryve et al, 1973). We recently observed an apparent increase in factor-VIII clot promoting activity (VIII AHF) as measured by a one-stage assay (AHF,) in a haemophiliac with hepatitis. However, factor VIII as measured by a two-stage assay (VIII AHF,) remained low. W e then measured Present address: Ohio Valley Medical Center Inc., Department of Medicine, Wheeling, West Virginia 26003, USA. Correspondence: Dr M. Elaine Eyster, The Milton S. Hershey Medical Center, Hershey, Pennsylvania 17033, U.S.A.

656

1.S . Rogers and M. Elaine Eyster

VIII AHF, , VIII AHF,, and factor VIII-like antigen (VIII AGN) in a group of seven nonhaemophiliacs with livcr disease. The purpose of this report is to emphasize that discrepancies between VIII AHFl, VIII AHF,, and VIII AGN exist in patients with liver disease and that the diagnosis of classic haeniophilia may be obscured during the course of viral hepatitis unless VIII AHF is measured by a two-stagc assay. METHODS Blood for coagulation studies was collected by clean venipuncture in plastic tubes containing 1/9th volume ofbalanced citrate (2 vol. 0.1 M citric acid, 3 vol. 0.1 M sodium citrate). Plateletpoor plasma was prepared at 4°C by centrifugation at 9000 revlmin for 20 min. Coagulation assays were performed on fresh plasma as well as on plasma stored at - 70°C. V I I A H F , assays were performed by a one-stage method based on the activated partial thromboplastin time (PTT) (Biggs, 1972). One-tenth ml diluted plasma, 0.1 ml VIII AHF deficient substrate plasma (George King Biomedical Inc., Salem, N.H.), and 0.1 nil activated platelin (General Diagnostics, Morris Plains, N.J.) were incubated for 4 min at 37°C. Onetenth in1 of 0.030 M CaCl, was then added, and thc interval until clot formation measured. A standard curve was plotted on double log paper with time in seconds on the ordinate and VIII AHF units per ml on the abscissa, with a I :5 dilution cqual to 1.0unit. At lcast six serial dilutions of reference plasma and three dilutions of test plasma (usually I :s, I :IO and I 20) were used. All tests were performed in duplicate and the clotting times converted to AHF units by interpolation from the curve. V I I I A H F , assays were performed by a two-stage method based on the thromboplastin generation test as described by Biggs (197z), subsampling incubated mixtures of test plasma adsorbed with Al(OH),, normal human serum, bovine factor V, phospholipid and calcium into normal substrate plasma when a plateau of prothrombin activator had been reached. A straight line was obtained for values plotted on double log paper, and parallelism was observed between the dose-response curves of reference and test plasmas. Bovine factor V was prepared by (NH,), SO4 precipitation following BaSO, absorption of bovine plasma (Biggs, 1972). VIII AGN determinations were performed by a modification of the Laurell technique of quantitative immunoelectrophoresis on whole plasma as previously described (Eyster et al, 1976; Laurell, 1972). Case Report

A 52-year-old white male with a lifelong history ofjoint bleeding was reported to have an VIII AHF level of 5 % (0.05 u/ml). Past history included numerous transfusions of blood and plasma for joint bleeding, haematemesis, bleeding following herniorrhaphy and dental extractions. Family history revealed that seven out of 10 maternal uncles and one grandson by his daughter were haemophiliacs. Three weeks prior to admission he was treated with lyophilized factor VIII concentrate for the first time for a severe haemarthrosis of the left knee. Because of persistent bleeding,

Factor VIII in Liver Disease

657

he was admitted to hospital on 21 October 1973 for further evaluation and treatment. He denied any history ofjaundice or hepatitis. Physical examination revealed a well-developed, well-nourished white male in no distress. The liver edge was palpated 3 cm below the right costal margin. Orthopedic abnormalities included mild bilateral quadriceps atrophy and moderate tenderness and effusion of the left knee, with limited range of movement in both knees. RESULTS Routine VIII AHF, determinations performed to confirm the diagnosis of classic haemophilia unexpectedly revealed values of 0.32, 0.35, and 0.33 u/ml on three separate plasma samples (Table I). Over the course of the next week the patient becamejaundiced and the diagnosis of hepatitis was made. An AHF, assay was then performed on the initial frozen specimen and found to be 0.013 u/ml compared to a value of 0.38 u/ml for the VIII AHF,, performed simultaneously on the frozen-thawed specimen. VIII AGN was 2.95 u/nil. With resolution of the hepatitis, VIII AHF, values of 0.06u/inl and 0.03 u/ml were obtained. An VIII AGN determination on the latter specimen was 2.90 u/ml. Six white non-haemophiliacs with hepatitis and one white woman with breast cancer metastic to the liver were then studied (Table 11). In four patients with the highest levels of VIII AHF and VIII AGN (patients I-4),VIII AHF, was disproportionately increased over VIII AHF, . Moreover, VIII AGN values correlated with AHF, in these patients. In the remaining three patients, VIII AHF, and VIII AHF2 levels were proportionate and in agreement with VIII AGN levels. Four normal volunteers who served as controls showed VIII AHF levels between 0.93 and 1.31 u/ml, VIII AHF, levels between IOO and 1.40u/ml, and VIII AGN levels between 1.00and 1.10 u/ml. DISCUSSION The complicated nature of the reagents and the longer incubation times required for the performance of the two-stage assay for VIII AHF (AHF2) have necessitated the adoption of the simpler one-stage assay (AHF,) in most clinical coagulation laboratories. The AHF, assay is based on a modified PTT, with factor-VIII deficient substrate plasma being added to dilutions of test plasma. This ensures excess of all other clotting factors, thus making the rate of clotting proportional to the concentration of factor VIII in the test plasma. This assumption is not always true, however. False elevation of the AHF, assay may occur with release of tissue factor following traumatic venipunctures, or in association with intravascular coagulation where active intermediates such as thrombin or activated factor X act subsequent to the point at which AHF participates in clotting (Denson, 1971; Niemetz & Nossel, 1969).False depressions of the AHF, assay may be seen in the presence of circulating anticoagulants which could lengthen the clotting time (Niemetz & Nossel, 1969). In the AHF, assay, test plasma adsorbed with Al(OH), is incubated with normal human serum, bovine factor V, phospholipid and calcium. Both serum and adsorbed plasma contain some prothrombin, but in markedly decreased amounts. With the exception of factor VIII, all the coagulation factors which are involved in prothrombin activation are present in

~

~

10-50

40-110 ~~

1740 139 29s

500 128

9-1

0.1-1.2

0.5

1-3

-

-

-

191 556

SCOT (i.11.)

0.7

(mgldl)

Alkaline phosphatase (iUl0

11.7

neglpos neglpos negl-

10-13

-

-

11.7

(5)

PT

-

neg/-

HbAglAb

25-45

-

72

54

-

56

(5)

PTT

~

~~

0.03 1.08 +0.32*

0.06

0.3 8 -

0.32,0-35. 0.33

(44

VIII AHFl

~~

0.5-z.ot

-

0.013 -

-

(ul4

VIII AHF2

2.90 1.18+0.38*

-

2.95

(44

VIII A G N

Abbreviations: SGOT, serum glutamic oxalacetic acid transaminase; HbAg, hepatitis-B antigen by counter electrophoresis; Ab, antibody to hepatitis-B antigen by counter electrophoresis; PT, prothrombin time; PTT, partial thromboplastin time; VIII AHF1. one-stage assay for factor VIII; VIII A m 2 , two-stage assay for factor VIII; VIII AGN, factor VIII-like antigen. * Mean+ SD. t Normal range from literature. - Test not performed.

22

October 1973 Frozen-thawed sample 22 October 1973 30 October 1973 12December 1973 21 May I974 Normal ranges

Date

Bilirubin

TABLE I. A haemophiliac with hepatitis: sequential liver function and coagulation studies

3

fr

Diagnosis

Breast cancer metastatic to liver. 27, F Hepatitis-resolving. Granulomatous inflammation on biopsy. 41, M Hepatitis following transfusion, resolving. Antitrypsin deficiency with emphysema. 58, M Hepatitis following transfusion, resolving. Acute myelomonocytic leukaemia in remission. 53, F Alcoholic hepatitis and fibrosis on biopsy. 31, F Hepatitis, resolving. Cholestatic hepatitis on biopsy.

53, F

51, M Alcoholic hepatitis with ascites and asterixis. Expired.

Sex

Age,

Normal ranges

Patient

0.1-1.2

10-50

neg

Abbreviations as for Table I. * Normal range from literature.

40-110

63

IS0

2.7

10-13

13.4

11.4

1.08k0.32

2.71

25-45

1.70

3s

neg

136

210

1.1

2.51

-

neg

I37

I20

2.5

1.15

-

13.0

neg

87

77s

0.8

1.93

31

I47

I .2

-

142 11.9

2.58

(ulml)

VIII AHFl

neg

(4

(4

39.7

DTT

PT

11.1

HbAg

-

SCOT (iull.)

550

Alkaline phosphatase (iull.)

1.1

Bilirubin (mgldl)

TABLE 11. Seven non-haemophiliacs with liver disease: liver function and coagulation studies

0.5 -2.0*

2.75

1-55

3.00

2.70

4.90

8.00

VIIZ AH& (uiml)

1.18k0.38

2.95

2.29

3.25

3.60

4.00

6.17

VIII AGN (ulml)

660

J. S. Rogers and M. Elaine Eyster

excess. The incubation is then subsampled into normal substrate plasma, where the recalcified clotting time is proportional to the amount of prothrombin activator found in the incubation mixture. Since all other factors except VIII AHF are present in excess, the clotting times obtained will be proportional to the amount of VIII AHF in the test plasma. The presence of active intermediates have minimal effect on the clotting times obtained in this method which is felt to be a more accurate reflection of VIII AHF levels (Demon, 1971; Biggs, 1972). Elevated levels of VIII AHF and VIII AGN have been described previously in hepatitis and alcoholic cirrhosis (Baele ct al, 1973 ; Green & Ratnoff, 1974; Herold & Straub, 1954; Koller, 1973; Mcili & Straub, 1970; Van Outryve ct al, 1973). No relationship between VIII AHF, chronicity of disease, or severity of hepatic damage as measured by tests of liver function was apparent (Green & Ratnoff, 1974; Van Outryve et a f , 1973). Green & Ratnoff (1974) described elevated VIII AHF, in 22 patients with advanced hepatic cirrhosis. Thirteen had VIII AGN significantly higher than VIII AHF, , but AHF2 assays were not performed. The possibility that disseminated intravascular coagulation influenced these results was excluded by negative protamine sulphate and fibrinogen-related antigen tests, and it was postulated that the elevation of factor VIII was caused by either increased production, release of factor VIII from hepatocellular damage, or impaired catabolism. In the present report, we describe a haemophiliac whose VIII AHF, assay remained low during an cpisodc of icteric hepatitis when his VIII AHF, increased. Unfortunately, sufficient plasma was not available to test for the presence of active intermediates which could have caused a false elevation of VIII AHF, . In contrast, the non-haemophiliacs showed the same pattern found by Green & Ratnoff,with VIII AHF, levels being disproportionately increased over VIII AHF, in four of seven patients. Furtliermore, VIII AGN measurements correlated well with VIII AHF, levels. The reason for the disparity between VIlI AHF, and VIII AHF, determinations in nonhaemophiliacs with liver disease is not apparent. It is theoretically possible that excessive amounts of precursor substrate factor VIII having antigenic activity could exhaust the supply of a serum factor which was required for the production of VlII AHF. Since the two-stage assay entails the incubation of test plasma with normal serum during the first stage, it is conceivable that sufficient serum factor could be added in vitro during the first stage to convert additional quantities of substrate precursor material to VIII AHF. This would not occur during the performance of thc one-stage assay with a shorter incubation time if both test plasma and haemophilic substrate plasma were deficient in serum factor. However, we have no data to substantiate this hypothesis. These findings emphasize the necessity for performing two-stage assays in haeniophiliacs as well as non-haemophiliacs with liver disease to assess VIII AHF levels. Furthermore, they suggest that confirmation of the diagnosis of haemophilia may not be possible in the haemophilic with hepatitis unless VIII AHF, determinations are performed. Finally, the observation that VIII AGN levels correlate with VIII AHFz levels in non-haemophiliacs with liver diseases provides further support for the concept that VIII AGN and VIII AHF are closely related or identical molecular entities. Aaaelzaum Since this paper was submitted for publication, Gazzard t t uI(1975) have described an in-

Factor VIII in Liver Disease

66 I

creased VIII AHFl with a disproportionately increased VIII AGN in a haemophiliac with hepatitis. In the early stages of the illness there was evidence of disseminated intravascular coagulation, but the factor-VIII levels remained elevated beyond this period. ACKNOWLEDGMENTS

The authors are indebted to Dr Ralph L. Nachman for review of this manuscript. This study was made possible by the Hemophilia Program of The Pennsylvania Department of Health and was supported in part by The Central Pennsylvania Chapter of The National Hemophilia Foundation. REFERENCES P., DEMEULENAERE, L. & BAELE,G., VERMEIRE, BARBIER, F. (1973) Elevation of factor VIII in acute liver necrosis: influence of plasma pheresis. Digestion, 8, 360. BIGGS, R. (1972) Human Blood Coagulation, Haemostasis and Thrombosir, pp 614 and 616. Blackwell Scientific Publications, Oxford. DENSON,K.W.E. (1971) Standardization of methods for the determination of factor VIII. Thrombosis et Diathesis Haemorrhagica, Supplement 43, 99. EYSTER, M.E., JONES,M.B., DELLI-BOVI, L. & MOORE, T . (1976) Camer detection in classic hemophilia by combined measurement of immunologic (VIII AGN) and procoagulant (VIII AHF) activities. American Journal of Clinical Pathology, 65, 975. GAZZARD, B.G., CLARK, R., FLUTEP.T. & WILLIAMS, R. (197s) Factor VIII levels during the course of acute hepatitis in a haemophiliac. Journal of Clinical Pathology, 28, 972. GREEN,A.J. & RATNOFF, O.D. (1974) Elevated antihemophilic factor (AHF, factor VIII) procoagulant activity and AHF-like antigen in alcoholic cirrhosis of the liver. Journal ofLaboratory and Clinical Medicine, 83, 189. HEROLD, R. & STRAUB,P.W. (1954) Acute hepatic necrosis of hepatitis and mushroom poisoning. (The value of coagulation tests in their differentiation, prognostic assessment and pathogenesis). Helvetica Medica Acta, 37, 5.

JORPES,E & RAMGREN, 0. (1962) The haemophilia situation in Sweden. Acta Medica Scandinavica, Supplement 379, 22. KOLLER, F. (1973) Theory and experience behind the use of coagulation tests in diagnosis and prognosis of liver disease. ScandinavianJournal of Gastroenterology, Supplement 19, 51. LAURELL, C.-B. (1972) Electroimmunoassay. Scandinavian Journal of Clinical and Laboratory Investigation, 29, Supplement 124, 2. MEILI, E.O. & STRAUB,P.W. (1970) Elevation of factor VIII in acute fatal liver necrosis. Thrombosis et Diathesis Haemorrhagica, 24, 161. NIEMETZ, J. & NOSSEL, H.L. (1969) Activated coagulation factors: in-vivo and in-vitro studies. British Journal ofHaematology, 16, 337. PRENTICE, C.R.M., FORBES, C.D.& SMITH,S.M. (1972) Rise of factor VIII after exercise and adrenalin infusion measured by immunological and biological techniques. Thrombosis Research, I, 496. RIZZA,C.R. (1961) Effect of exercise on the level of antihaemophilic globulin in human blood. Journal ofPhysiology, 156, 128. VANOUTRYVE, M., BAELE,G., DE WEERDT,G.A. & BARBIER,F. (1973) Antihemophilic factor A (F VIII) and serum fibrin-fibrinogen degradation products in hepatic cirrhosis. Scandinavian Journal of Haematology, 11, 148.