PIP1PICATION

PORE IgG

CONTROL OF SPECIFICITY

111CR PURITY PEROXIDASE

IAT ACTIVATION

COUPLiNG

PEPSIN DIGESTION

GEL FILTRATION ON SEPHACRYL S 200

F(b’)2 FRAGMENTS LPEROXIDASE ANTIBODY CONJiJ

CLIN. CHEM. 30/9, 1512-1516 (1984)

1512 CLINICALCHEMISTRY, Vol. 30, No. 9, 1984

Applicationof Enzyme Immunoassaysto CoagulationTestingJean Amiral, Benedicte Adalbert, and Mariette Adam

Enzyme immunoassaysare very useful for the detectionoflow concentrationsof coagulationproteinsand pathologicalmarkers in plasma. Analytes in the ng/mL range are measur-able with good reproducibility with intra- and interassay CVsof less than 5% to 10%. “Sandwich” methods have beendeveloped for von Willebrand factor (plasma concentrationabout 8 .&.g/mL,Factor IX (5 p.g/mL),proteinC (4 g/mL),and Factor X (10 g/mL). However, this technique is onlysuitable for macromolecules; for low-molecular-mass pep-tides such as fibrinopeptide A a competitive method is used.Normal concentrationsoffibrinopeptideA arebelow 3 ng/mL,with greater values suggesting in vivo generation of throm-bin; thus this test is quite useful in detecting thrombosis.Reagents for both the sandwich and competitive methodsare commerciallyavailable and cost effective, and have alongershelf-lifethan those for radioimmunoassays.

AddItIonal Keyphrases: “sandwich” immunoassay thrombosisvon Willebrand factor Factor IX . Factor X . protein Cfibrinopeptide A . peptides . thrombin

Maintenance of blood in a fluid state and prevention ofhemorrhage requires a balance of activators and inhibitorsof coagulation and fibrinolysis. Although clotting methodsare useful for investigating most hemorrhagic diathesis and

for monitoring procoagulant or hypocoagulant therapy, theyare ineffective in early diagnosis of prethrombotic states andare limited for studying the subclinical stages of hemostaticdisorders. A dramatic expansion in coagulation testing iscurrently underway. Highly sophisticated, specific, and ul-trasensitive methods have been introduced for determina-tion of low concentrations of coagulation and related pro-teins in plasma that can be used as pathological markers.Purification and characterization of coagulation factors andrelated peptides have led to the development of immuno-technology for these analytes. Besides their activity, theabsolute concentrations of these analytes can be measured(1). Moreover, measurement of analytes that have no specif-ic activity, such as degradation products or peptides result-ing from activation processes, has become possible (2).

The relative amounts in which coagulation factors andrelated proteins or peptides are present in blood vary widely,ranging from nanograms to milligrams per milliliter-quantities that can be easily measured by using immunoas-says with labels (radioimmunoassay, enzyme immunoassay,fluoroimmunoassay) (3). Immunoassays for coagulation fac-tors can provide useful information in the study of function-al deficiencies, inactive complexes, abnormal proteins in-duced in various disease states or subsequent to therapy, thepresence of pathological inhibitors, an increased catabolism,or abnormal degradation. Other new applications includethe measurement of activation peptides (4), which are ofhigh predictive value in early diagnosis of hypercoagulabil-ity or fibrinolysis.

Enzyme immunoassays have been introduced for von

Lab. Diagnostica Stago, Immunology Department, Franconville,France.

Received April 12, 1984; accepted June 22, 1984.

Willebrand factor (5-8), Factor IX (9, 10), Factor X (11),protein C (12), and fibrinopeptide A (13), with new applica-tions under development. Our laboratory has developedvarious enzyme immunoassays for plasma coagulation pro-teins and molecular markers of various pathologic states.Depending on the molecular mass of the analyte, we use twodifferent analytical principles (14, 15): an ELISA (enzyme-linked immunosorbent “sandwich” assay) method or a com-petitive method. Here we explain these in detail as theyhave been adapted for use in the coagulation laboratory.

Materials and Methods

Reagents

All the reagents are analytical grade and purchased from

Prolabo, Paris, France, except when specified otherwise.Preparation of antisera: Ultra-specific antibodies with the

high affinity required for these analytes are obtained from

hyperimmune animals, immunized with the purified anti-gen (16). Rabbit or goat sera are used. When necessary,trace amounts of nonspecific antibodies are removed byadsorption to solid phase. Specificity of antisera is tested byconventional methods (Ouchterlony, immunoelectrophore-sis) and by two-dimensional electrophoresis with a highconcentration of antiserum in the second migration step.Specificity is checked again in the more-sensitive enzymeimmunoassay. Antibodies are purified either by ion-ex-change chromatography on DEAE-Trisacryl (IBF, Ville-neuve la Garenne, France) or affinity chromatography oncolumns of agarose-linked specific antigens.

Preparation of enzyme immunoassay reagents: Figure 1illustrates the preparation of reagents for the ELISA sand-wich method. Peroxidase (type VI; Sigma Chemical Co., St.Louis, MO) was chosen to label antibodies because of its

I4ONOSPECIFIC ANTISERUM

CONTROL OF SPECIFiCITY

E.I.A. REA

Fig. 1. Flowsheet for the preparation of reagents for EUSA sandwichmethod

Congenitallydeficientplasmafrompatientswithsevere diseases (activityoffactor<lint.unit/i..)is usefultotestthespecificityof reagents and efficacyof the ELISA

technique

ready availability, low cost, high activity, and chromogenic mmol/L; citrate/blood, 1/9 by vol). Centrifuge, then separatecharacteristics (17). In our hands, the reaction of o-phenyl- the plasma and freeze at -80 #{176}Cuntil used. Calibrate withenediamine with hydrogen peroxide produces high respons- small aliquots of plasma pooled from normal individuals (15,es and good sensitivity. Labeling antibodies with alkaline in our studies) that has been kept frozen (-80 #{176}C).phosphatase or 3-galactosidase is also potentially useful. To determine various proteins (Factor IX, von Wille-The respective substrates with these methods would be p- brand’s factor, and protein C) by electroimmunodiffusion, wenitrophenyl phosphate and o-nitrophenyl 3-galactoside (18). used Assera#{174}-Plates (Diagnostica Stago) and the method ofAntibody-peroxidase conjugates are prepared according to Laurell.Nakane and Kawaoi (19). Unreacted proteins are removedby repeated salt fractionation with ammonium sulfate. Results

F(ab’)2 fragments are obtained by digesting the specificimmunoglobulin (20) with pepsin (Sigma). To check thespecificity of the reagents [F(ab’)2 fragments and antibody-enzyme conjugates] in the routine enzyme immunoassayprocedure, we use plasmas from patients homozygous forsevere deficiencies in specific factors. Values for thesesamples must be similar to those obtained for buffer alone,

Dilution buffer: per liter, 50 mmol of phosphate, 150 mmolof sodium chloride, 1 mL of Tween 20, and 1 g of bovineserum albumin; final pH is 7.50.

Washing solution: per liter, 20 mmol of phosphate, 150mmol of sodium chloride, and 1 mL of Tween 20, adjusted topH 7.00.

Substrate: o-phenylenediarnine, 0.4 g/L of phosphate/ci-trate buffer (50 mmol of phosphate and 25 mmol of citrateper liter, pH 5.00). Just before use, add 40 L of a 30%hydrogen peroxide solution per 100 mL of substrate solu-tion.

Solid-phase support: Microplates were purchased fromseveral manufacturers: Nunc, Roskilde, Denmark; Costar,OSI, Paris, France; or Dynatech, South Windham, ME;Maxisorp tubes for ELISA were purchased from Nunc.

Table 1 summarizes the performance characteristics ofthe Asserachrom kits. Inter- and intra-assay reproducibili-ties are good, and results depend on the concentrations ofthe analytes. Detection limits are in the nanogram rangeand compare well with those obtained with radioimmunoas-says for similar applications. Determinations take 3 to 5 h,depending on the incubation time (1 to 2 h) chosen for eachstep. These assay characteristics are well adapted to theconcentrations and variations observed clinically with theseanalytes (Table 2).

Von Willebrand factor: Determination of this coagulationfactor (also known as VIII R:Ag) is useful in the diagnosis ofvon Willebrand’s disease. Although several types of immu-noassays for it have been proposed (5, 8, 23, 24), thesandwich method appears to be the more practical becauseno pure antigen is needed and specific antibodies areavailable. This method has been fully applied in hemophiliaA, in von Willebrand’s disease (particularly in severe defi-ciencies, where very low quantities are present), in theidentification and classification of carriers of hemophilia A,and in evaluating endothelial cell damage (25).

By the EUSA method for von Willebrand factor, the mean

ProceduresELISA ‘sandwich” method: This method is used for von

Willebrand factor, Factors IX and X, and coagulation pro-

percentage of the factor in plasma from 30 normal individ-

tein C. Coat a solid-phase support (microplate or tube) withexcess specific F(ab’)2 fragments; remove unbound materialby repeated washes with washing buffer. Add standard ortest sample (antigen binds to the surface-linked antibodies),wash once again, then add the immunoconjugate, whichbinds to free antigenic sites. Remove excess conjugate bywashing again, then measure peroxide activity by its oxida-tion of o-phenylenodiamine. Stop the development of colorby adding sulfuric acid (3 molIL), then measure absorbanceat 492 nm. F(ab’)2 fragments, instead of whole immunoglob-ulin, are used to avoid possible interference from rheuma-toid factor (21). The enzyme immunoassays we evaluatedand report on here are from the Asserachrom#{174}line (Diag-nostica Stago, Asni#{233}res,France).

Competitive method for fibrinopeptid.e A: The competitivemethod for enzyme immunoassays, which has been reportedin detail by Soria et a!. (13), involves competition for alimited amount of rabbit antibodies to fibrinopeptide A(FPA) between the test FPA (free in the sample) andimmobilized FPA (coated on a solid-phase support). Concen-trations of bound antibodies, which are inversely related tothe concentrations of test FPA, are measured with sheepimmunoglobuirns (anti-rabbit IgG) labeled with peroxidase.Subsequent steps are identical to those described for theELISA “sandwich” method. Specificity of the determination isachieved by collecting samples into a special anticoagulantsupplemented with inhibitors and by using bentonite to

Analyte PA, Normal range Abnormal range600-1500 <600 >2000

(5-12)Factor IX 55 000 600-1400 <600

(3-5)Factor X 65000 700-1300 <600

remove cross-reacting fibrinogen before analysis (22). Wecompared the results by a commercially available test(Asserachrom FPA, Diagnostica Stago) with those by aradioimmunoassay kit (Mallinckrodt, St. Louis, MO).

(7-12)Protein C 62 000 600-1400 <600

FPA 1536

Plasma collection: Collect blood in sodium citrate (110

Table 1. Some Characteristics of AsserachromKits for Measuring Coagulation Proteins in

Plasma

LImIt ofdetectIon,mt. unlts/L

(ng/mL)

1/50 to 1/5000 0.2 (1.5)

CV, %

Intra- Inter-assay assay

(n=20) (n=20)

2-5 4-8

4-7 5-104-7 5-103-6 4-9

0.2 (0.75)0.1 (1)0.2 (0.75)

am terms of dilution of normal pooledplasma. 6ng/mL.(0.5) 4-7 5-10

Table 2. Characteristics of Various AnalytesMeasured by Enzyme Immunoassay

von Willebrands factor >1 x 106

6ng/mL.

CLINICAL CHEMISTRY, Vol.30,No.9,1984 1513

Analyte WorkIng range

Sandwich” ELISA

von Willebrandfactor

FactorIX 1/50 to 1/5000FactorX 1/50to1/5000ProteinC 1/50 to 1/5000Competitive enzyme immunoassayFPA 0.5to5Ot

Concns, mt. unlts/L(and tg/mL)

Table 3. Factors IX, X, and Protein C Measured byELISA Sandwich for Different Groups of Subjects

% (normal human pooledplasma = 100%)

Factor IX Factor X ProteIn C

SD Mean SD Mean SD

19.5 98.6 16.2 97.4 18.1

16.2 44.1 14.2 43.8 14.719.2 37.5 26.5 19.3 5.2

23.18.10.27

1514 CLINICAL CHEMISTRY, Vol.30,No. 9,1984

uals and from 10 patients suffering from hemophilia A was101.3 (SD 24.8)% and 103.2 (SD 22.8)%, respectively (100%= the value in pooled normal human plasma). Results ofsingle determinations on five patients with von Wille-brand’s disease were 0.1, 6, 7, 18, and 30%. Correlationbetween these results and those by the Laurell electroim-munodiffusion method was good for normals, patients withhemophilia A, and patients with classical von Willebrand’sdisease (results not shown). Moreover, the presence of anabnormal protein (as in von Willebrand’s disease type II)changes the results obtained with both the ELISA and theLaurell methods, indicating that a different electrophoreticmobility and a loss of immunoreactivity to antibodies ischaracteristic of the abnormal protein (6).

Factors IX, X, and protein C: These vitamin K-dependentfactors, synthetized in the liver in low concentrations, are ofdiagnostic value in various pathological states. Deficienciesor abnormalities of Factor IX result in one of the manyheterogeneous types of hemophilia B. Measurement of thisfactor allows classification of hemophilia B into subclasses(Bk, BR, or B) (26), and may help to detect carriers ofhemophilia B and make prenatal diagnoses of these defi-ciencies. Protein C has a major role in regulating in vivoactivation of blood, by inactivation of thrombin-modifiedFactors Vffl:C and V. Congenital or acquired deficiencies ofprotein C result in thrombotic tendencies (27), and itsroutine measurement in coagulation laboratories can aid inthe diagnosis of thrombosis. Concentrations of Factor X,Factor IX, and protein C are decreased in liver cirrhosis andin patients treated with oral anticoagulant drugs, e.g.,warfarin. Table 3 shows several applications of the measure-ment of these three factors in various groups of patients. Inaddition, patients with various subclasses of hemophilia Bwere tested for Factor IX, for which normal (Bk), low (BR),and very low (B) values were consistently obtained for theconcentration of antigen. For these several groups of pa-tients, there was a good correlation between the ELISA

method and the Laurell electroimmunodiffusion method(results not shown). We observed no discrepancy betweenresults by each, and mean values and standard deviationsby both methods did not significantly differ.

In warfarin-treated patients, the concentrations of Fac-tors IX and X and protein C, measured immunologically, aredecreased, but usually less than as measured with clot-based functional assays. Although both normal and abnor-mal proteins are synthesized in the liver during therapywith anticoagulant, the total functional concentration ofthese in plasma is lowered. However, the antibodies used inthese tests react with normal molecules as well as with theabnormal forms (hypo or acarboxylated) induced by antico-agulant therapy. Therefore the concentrations of antigenicand functional proteins do not always coincide. Protein Cand Factor X are decreased to a comparable extent, withsimilar variations. Factor 1X is less decreased in the warfa-rin-treated group. In patients suffering from cirrhosis of theliver, protein C is the factor decreased the most, reachingvery low concentrations in some patients. Factor IX andFactor X present much greater variations than other fac-tors, but their total decrease is less severe.

Fibrinopeptide A (FPA): This is a marker of high predic-tive value for the early diagnosis of hypercoagulability (28),and can be used to monitor procoagulant and hypocoagulanttherapy, and to survey post-surgical states, heart diseases,progression of cancer, and risk of thrombosis. Above-normalconcentrations, >3 ng/mL, indicate thrombin generation invivo. Mean FPA values for 20 normal subjects and 10patients with active thrombosis were 1.2 (SD 0.6) and 22.8(15.4) ng/mL, respectively, with the Asserachrome FPA kit,

Subjects Mean

Normal (n = 40) 102.3Dicoumarol-treated

(n = 25) 54.6Livercirrhosis(n = 12) 54.3Hemophiliacs(n = 5 each)

BBRB

8222

0.48

and 1.1 (SD 0.8) and 23.6 (SD 14.8) with the competitiveradioimmunoassay kit (Mallinckrodt).

DiscussionAmounts in which coagulation factors and related pro-

teins or peptides are present in blood vary widely, rangingfrom nanograms to milligrams per milliliter. Thus a widevariety of immunological methods, with various detectionsensitivities, can be used. Conventional methods (radioim-munodiffusion, electroimmunodiffusion, immunonephelom-etry, or agglutination) can be used for proteins present inmedium or high concentrations (above 10 .i.g/mL). Smallquantities such as micrograms or nanograms per millilitercan be easi)y measured by using labeled immunoassay orfluoroimmunoassay techniques. For protein concentrationsin the range 1-20 ig/mL, both types of methods may beconsidered, depending on the accuracy and sensitivity need-ed. Figure 2 summarizes plasma concentrations of the maincoagulation proteins and related peptides, and shows theworking range of the main immunological methods used fortheir assay.

For some coagulation factors, immunoassays provide use-ful information. In physiological states, there is usually astrong correlation between the quantity of a specific proteinand its associated activities, although there are discrepan-cies in various circumstances. Immunoassays based onpolyclonal antibodies measure the entire protein. If theantiserum has been obtained from hyperimmune animals,there is a complete recognition of the structure of themolecule, and normal and abnormal proteins are deter-mined to the same extent. The development of monoclonalantibodies, however, introduces quite a different kind ofanalysis. Monoclonal antibodies can be selected to reactspecifically with a single epitope and can provide a betterknowledge of protein structure and intermolecular interac-tions.

Obtaining a good surface coating is of crucial importanceto the performance of these assays. The solid phase to becoated must be carefully chosen. It must offer high bindingcapacities and stable biological and immunological reactiv-ity of linked material, as well as good reproducibility withinbatch and from batch to batch. The amount of coatedmaterial determines the performance of the assay; use of toolittle coating severely reduces the rate of color development.We successfully used the three microplates (Nunc Type 1,Dynatech M129B, or Costar 3590) and the Nunc Maxisorptubes for the enzyme immunoassays reported here. In theseassays the reaction takes place on a solid surface, so thatimmunological equilibrium is reached slowly, taking atleast 2 h at room temperature. However, one can reduceincubation times to 1 h (or even 30 mm) without affecting

#{149}6 Keto PCFJ0

#{149}Thr’omboxane B2 #{149}pF4

LABELI.ED METhODS

RADJOIMMUNOASSA Y

FLUOROIMMUNOASS.4 Y

ENZYME IMMUNOASSA Y

1.0

0.5

0.2

0.1

fbi

0

2.5 5 10 25 50 100 t of Protein C1:5000 1:2000 1:1000 1:500 1:200 1:100 1:50 Plasma dilution

I I I I I -

0.78 1.56 3. 12 6.25 12.5 25.0 ng FPA

I I

CLINICALCHEMISTRY, Vol. 30, No. 9, 1984 1515

I ATC #{149}F VII F XI U HMWK . FibrinogenU F V i a2Anti-Plasmin #{149}cz2Macroglobuiin

‘F VIl1:C ‘FX U I1JAntitrypsineF VIIIR:Ag #{149}Fibronectin

IFIX ‘AT IIIU tPA ‘Protein C ‘ Prothrornbin

i FPA U F.D.P.

‘ R-l5-42 . F’ XIII#{149}APCI

PROTEINI I I CONCENTRATIONS

JO 100 1 10 100 1 10 100 1 10U

pg/mt ng/mL pg/mt mg/ WORKING RANGES P01- _____________________________________________ _____________________________ VARIOUS METHODS

OUCHTERLONY’S DOUBLE DIFFUSION

IMI4iWOELECTROPHORESIS

RADIAL IMMUNODIFFUSION

TURBIDIME TRY

LASER IMMUNONEPIIELOMETW

ELECTROIMMUNODIFFUSIOV

AGGLUTINATION

I I I I

Ag. 2. Usual plasma concentrations(log)for coagulation proteins in plasma and molecular markers ofhemostasis, along with the working range ofvarious immunological methods used for analysisF,Factor;BTG, p-thromboglobulmn;HMWK, high-molecular-masskininogen;PF4, plateletfactor4; AT Ill, anlithrombinIll; I-PA, tissue plasminogenactivator;FDP,fibrinogen degradation products; APCI. activated protein C inhibitor

the results, provided the incubation times are exactly thesame for the standard and test dilutions. With standardworking conditions, color develops in 2 to 3 mm; shorteningthe incubation times increases the time for this develop-ment, usually to 5 miii. Absorbance at 492 nm is between 1and 1.5 for the highest point of the calibration curve (Figure3).

Enzyme immunoassays are convenient methods for rou-tine use in clinical laboratories. Numerous possibilities forautomation exist (29). A multi-channel pipette for distribu-tion, a washing machine, and a reader (allowing goodlinearity up to an absorbance of 2.0) are useful for obtainingreproducible results. Washing is critical because nonboundmaterial must be removed completely to minimize nonspe-cific results.

Introduction of enzyme immunoassays in the field ofhemostasis allows determination of low-concentration coag-ulation proteins in plasma and trace amounts of molecularmarkers of pathologic states. Use of commercial kits forthese determinations standardizes performance characteris-tics and makes it possible to include these tests in theroutine coagulation laboratories. Simplified procedures andspecific and automatic equipment allow for analysis of largeseries of samples. This breakthrough in the practice ofcoagulation testing is due in part to the introduction ofadvanced imniunotechnology.

Sensitivities in the nanogram range are available in mostspecialized laboratories. Important developments to be ex-pected in the future will include the evaluation of molecularmarkers, which should improve the detection of prethrom-

492

1.0 -

0.5 -

0

FIg. 3. Standard calibration curves for(tell) an EUSAsandwich method (Asserachrom)forproteinC and (right) a competitive enzyme immunoassay(Asseracitrom) for FPA

1516 CLINICAl..CHEMISTRY, Vol.30,No. 9, 1984

botic states and prevention of thrombosis (4). Determinationof FPA is, in this way, a very important test for thissensitive marker of thrombin generation in vivo. In comingyears development of hybridoma technology will introducemore standardized tests with well-characterized monoclonalantibodies. In addition, studies of molecular interactionsthat result from minor but pathological enzymmc activitieswill become available.

We thank Drs. D. Meyer and C. Boyer (Le Kremlin Bic#{234}tre,France) and C. Mazurier and A. Parquet-Gernex (CRTS, Lille,France) for help in the clinical evaluation and providing plasmasamples; we also thank Dr. J. Fareed and Ms. J. Walenga (LoyolaUniversity, Chicago, IL) for their helpful advice in the preparationof this manuscript.

References1. Hassouna HI, Penner JA. Antibody technics and blood coagula-tion. Sem Thromb Haemosto.s 7, 61-ill (1981).2. Hirch J. Blood tests for the diagnosis of venous and arterialthrombosis. Blood 57, 1-S (1981).3. Fareed J, Messmore IlL, Bermes EW. New perspectives incoagulation testing. Clin Chem 26, 1380-1391 (1980). Review.4. Fareed J, Bick RL, Squillaci G, et al. Molecular markers ofhemostatic disorders: Implications in the diagnosis and therapeuticmanagement of thrombotic and bleeding disorders. Clin Chem 29,1641-1658 (1983).5. Bartlett A, Dormandy KM, Hawkey CM, et al. Factor Vifi-related antigen: Measurement by enzyme immunoassay. Br Med Ji, 994-996 (1976).6. Mazurier C, Parquet-Gernez A, Goudemand M. Dosage del’antigene lie au facteur VIII par Ia technique ELISA. Int#{233}r#{233}tdensl’#{233}tudede la maladie de Willebrand. Pathol Biol 25 (suppl), 18-24(1977).7. Yorde LD, Hussey CV, Yorde DE, et al. Competitive enzyme-linked immunoassay for factor Vifiantigen. Clin Chem 25, 1924-1927 (1979).8. Ness PM, Perkins HA. A simple enzyme-immunoassay (EIA)teatforfactor Vifi-related antigen(VIIIAG N).Thromb Haemostas42,848-854 (1979).9. Parquet-Gernez A, Mazurier C, Amiral J. Dosage de l’antig#{232}nefacteur IX par une technique immuno-enzymatique. Nouv Rev FrHemo.tol 24, 114-115 (1982).10. Takainatsu J, Karniya T, Ogata K, et al. Sensitive solid phaseenzyme immunoassay for factor IX antigen and classification ofhemophilia B. Haemostatis 13, 9-16 (1983).11. Berthier AM, Pommereuil M, Amiral J, et al. Comparison ofimmunological (ELIsA) and biological determination of factor X in

oral anticoagulant therapy. Haemostasis 12, 142 (1982).12. Boyer C, Rothschild C, Wolf M, et al. A new ELISA for theestimation of protein C. Submitted to Thromb Res.13. Soria J, Soria C, Rickewaert JJ. A solid-phase immunoenzymo-logical assay for the measurement of human fibrinopeptide A.Thromb Res 20, 425-435 (1980).14. Bartlett A, Bidwell DE, Voller A. Practical methods for thedetection of antigens by ELISA. Protides Biol Fluids Proc Colloq 24,767-770 (1976).15. Voller A, Bidwell DE, Bartlett A. Enzyme immuno assays indiagnostic medicine: Theory and practice. BuU WHO 53, 55-56(1976).16. Peace AJ, Ford DJ, Gaizutis MA. Qualitative and quantitativeaspects of immunoassays. Scand J immunol 8 (suppl), 1-6 (1978).17. Avrameas S, Guilbert B. Enzyme-immunoassay for the mea-surement of antigens using peroxidase conjugates. Biochimie 54,837-842(1972).18. Avrameas S, Ternynck T, Guesoon JL. Coupling of enzymes toantibodies and antigens. Scand J Immunol 8 (suppl),7-23 (1978),

19. Nakane PK, Kawaoi A. Peroxidase labelled antibody: A newmethod of conjugation. J Histochem Cytochem 22, 1084-1091(1974).20. Nisonoff A, Wissler FC, Lipman LN. Properties of a majorcomponent of a peptic digest of rabbit antibody. Science 132, 1770-1771 (1960).21. Hagenaars AM, Kuipers All, Nagel J. Elimination of interfer-ence of rheumatoid factor in ELISA by peptic digestion of antibodies.immuno Enzym Assay Tech 1, 209-215 (1980).22. Kockum C, Frebelius S. Rapid radioimmuno assay of humanfibrinopeptide A. Removal of cross reacting fibrinogen with benton-ite. Thromb Res 19, 589-598 (1980).

23. Fishinan DS, Jones PK, Menitove JE, et al. Detection of thecarrier state for classic hemophilia using an enzyme linked ire-munosorbent assay. Blood 59, 1163-1168 (1983).24. Cejka J. Enzyme immunoassay for factor VilI-related antigen.Clin Chem 28, 1356-1358 (1982).25. Hoyer LW. The factor VIII complex: Structure and function.Blood 58,1-13 (1981).

26. PechetL,TiarksCY, StevensJ,etal.RelationshipsoffactorIX:Ag and coagulant in hemophiliaB patientsand carriers. ThrombHaemostas 40, 465-477 (1978).27. Esmon CT. Protein C. Biochemistry, physiology and clinicalimplications. Blood 62, 1155-1158 (1983).28. Nossel HL, Yudelman I, Canfield RE, et al. Measurement offibrinopeptide A in human blood. J Clin invest 54, 43-53 (1974).29. Martinoli JL, Amiral J. The impact of automation in thedevelopment of reagents and kits for automated methods in coagu-lation testing. Sem Thromb Haemostas 9, 194-205 (1983).