the thromboplastin generation

J. clin. Path. (1953), 6, 23.

THE THROMBOPLASTIN GENERATION TESTBY

ROSEMARY BIGGS AND A. S. DOUGLAS

From the Department of Pathology, Radelifle Infirmary, Oxford

(RECEIVED FOR PUBLICATION DECEMBER 24, 1952)

Physiological coagulation is dependent on theformation of a powerful thromboplastin within theblood. It is well recognized that coagulationdefects, such as that in haemophilia, are due to a

failure of the normal thromboplastin mechanism,but lack of knowledge has hitherto prevented any

precise or detailed study of these abnormalities.Biggs, Douglas, and Macfarlane (1953a) haveshown that three components, normal plasmatreated with AI(OH)3, platelets, and normal serum,

react to form thromboplastin. Further analysishas shown that two essential substances, anti-haemophilic globulin and factor V, occur in theAl(OH)3-treated plasma, while serum containstwo factors, factor VII and the Christmas factor,both of which are required for thromboplastinformation (Biggs, Douglas, and Macfarlane,1953b). These five components-platelets, anti-haemophilic globulin, the Christmas factor, andfactors V and VII-react together in the presence

of CaCl2 to form a labile thromboplastin as

powerful as any so far described. A lack of any

one of the five produces a coagulation defectassociated with abnormal thromboplastin form-ation.

Factors V and VII are usually thought of as

" accelerators " of prothrombin conversion becausethey are necessary for the rapid conversion of pro-

thrombin to thrombin in the presence of brainthromboplastin. This conception is valid becausebrain extract does not contain a complete thrombo-plastin ; it replaces part of the normal intrinsicthromboplastin system, but is incomplete because itlacks factors V and VII (Biggs, Douglas, andMacfarlane, 1953b). When brain extracts are addedto plasma they first react with factors V and VII toform a complete thromboplastin which subse-quently reacts with prothrombin. Thus in theabsence of factor V or factor VII the conversion ofprothrombin to thrombin with brain thrombo-plastin is delayed, and deficiency of either of thesefactors can conveniently be demonstrated by theclassical one-stage " prothrombin " test.

Factors necessary for blood thromboplastinformation which do not lengthen the one-stage"'prothrombin " time have hitherto been studiedby such empirical procedures as the whole-bloodclotting time, the calcium clotting time, and theprothrombin-consumption test. These techniquescan be used to show that some abnormality inthromboplastin formation exists, but they areneither specific nor very sensitive. The whole-blood clotting time, for example, is often a goodtest for haemophilia (antihaemophilic globulindeficiency), but it is not a sensitive measure of thehaemophilic defect. Some haemophilic patients,who suffer severe haemorrhagic episodes, have anormal whole-blood clotting time (Merskey, 1950,1951), and after transfusion with normal bloodthe clotting time of haemophilic patients may berestored to normal while the defect must remainuncorrected because the haemorrhage continues.

In spite of the number of factors involved andthe apparent complications of the reactions whichproduce thromboplastin, the thromboplastin gener-ation test (Biggs, Douglas, and Macfarlane, 1953a)is simple to carry out and to interpret. Using thistest together with the one-stage "prothrombin "

time, it is possible readily to distinguish betweendeficiencies in the various factors which react toform thromboplastin. It is the purpose of thiscommunication to describe the application of thethromboplastin generation test to the study ofvarious coagulation abnormalities. Factor V waspresent in normal amounts in the blood of all ofthe patients studied. Factor V deficiency is notconsidered in this investigation.

TechniqueCollection of Blood.-Venous blood is collected

from a normal subject and the patient under investi-gation. Part of each sample is citrated by adding1 part of 3.8% sodium citrate to 9 parts of blood.The plasma is separated after centrifuging at 2,000r.p.m. for 15 minutes. Five millilitres of blood willsuffice for the thromboplastin-generation test.

copyright. on D

ecember 29, 2021 by guest. P

rotected byhttp://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.6.1.23 on 1 February 1953. D

ownloaded from

http://jcp.bmj.com/

ROSEMARY BIGGS and A. S. DOUGLAS

A further 3 ml. of the whole venous blood is placedin a tube with three glass beads. The tube is stop-pered and inverted repeatedly until clotting occurs.This process encourages the conversion of prothrom-bin to thrombin during coagulation. When coagula-tion is complete the blood is allowed to stand at370 C. for two hours or more for the completeneutralization of thrombin, maximum utilization ofprothrombin, and disappearance of active thrombo-plastin. The serum is then separated after centri-fuging and is diluted 1 in 10 with 0.85% saline foruse in the thromboplastin-generation test. The"serum" factors are relatively stable on storage,and serum specimens may be used for some daysafter collection. Normal serum contains both factorVII and the Christmas factor.

Aluminium Hydroxide Ca.-This is prepared bythe method of Bertho and Grassmann (1938) whichis described by Biggs and Macfarlane (1953). It isreferred to in this communication as "alumina."

Alumina Plasa.-To 1 ml. of citrate plasma isadded 0.1 ml. of aluminium hydroxide suspension.The mixture is incubated at 37' C. for three minutesand the alumina is separated by centrifuging. Thesupematant plasma is tested by the one-stage "pro-thrombin" test. By this test the clotting time shouldlie between one and four minutes. The clotting timeshould not exceed four minutes. This treated plasmacontains the antihaemophilic globulin and factor V,but lacks significant amounts of prothrombin and the".serum" factors necessary for thromboplastin for-mation. For use in the thromboplastin generationtest the treated plasma is diluted 1 in 5 with 0.85%saline. (It will be referred to hereafter as " aluminaplasma.")

Platelets and Substrate.-For the thromboplastingeneration test 20 ml. of whole normal blood iscollected into two silicone-treated 10 ml. graduatedcentrifuge tubes each containing 1 ml. of 3.8% sodiumcitrate. The blood is centrifuged at 1,500 r.p.m. for10 minutes and the platelet-containing plasma is sepa-rated and put into silicone-treated tubes and centri-fuged again for 15 minutes at 3,000 r.p.m., or at 15,000r.p.m. for five minutes. The platelets are depositedand the clear supernatant plasma is separated andreserved as substrate for the test. The platelets arewashed twice with saline, the platelet mass beingfragmented by a wooden applicator stick and re-deposited on each occasion by centrifuging. Afterthe second washing the platelets are resuspended ina volume of 0.85% saline equal to one-third of thevolume of original plasma from which they werederived. This suspension of the platelets is facilitatedby fragmentation of the platelet mass with a woodenapplicator stick. The suspension is used undilutedfor the test and may be used on the day afterpreparation.

Thromboplastin Generation Test.-Immediatelybefore the test is started 0.1 ml. of substrate plasmais pipetted into each of six small tubes of uniform

diameter which are placed in a water-bath at 37' C.In a further tube in the water-bath at 37' C. isplaced 0.3 ml. of alumina plasma diluted 1 in 5,0.3 ml. of platelet suspension, and 0.3 ml. of serumdiluted 1 in 10. To this is added 0.3 ml. of M-40 CaCI2and a stop-watch started. At intervals of one minute0.1 ml. of the mixture is withdrawn into a graduatedPasteur pipette and, using the other hand, 0.1 ml. ofM-40 CaC12 is withdrawn into a second pipette. Thecontents of the two pipettes are then discharged simul-taneously into one of the tubes containing 0.1 ml. ofsubstrate. The clotting times of the substrate samplesare recorded. It is usually not necessary to continuethe test for more than six minutes. The clottingtimes of the substrate give a measure of thrombo-plastin concentration in the incubation mixture andmay be expressed in terms of thromboplastin con-centration using a thromboplastin-dilution curve.

Thromboplastin Dilution Curve.-A potent prepara-tion of blood thromboplastin, as made in the thrombo-plastin generation test described above, will usuallycause clotting of normal plasma in eight to tenseconds. When made deterioration of the thrombo-plastin can be delayed by placing the tube in meltingice. The preparation can then be diluted 1 in 2,1 in 4, 1 in 8, 1 in 16, and 1 in 32 with 0.85% saline.The dilutions are also kept at the temperature ofmelting ice and tested with normal plasma substrateas described above. A curve relating clotting timeand thromboplastin concentration can then be drawn(Fig. 1).

Experimental ResultsNomal Vartion-When platelets, alumina

plasma, and serum prepared from a normal personare incubated with CaCl2 a very powerful thrombo-plastin is formed. In Fig. 2 is shown the averageresult of carrying out this test on 40 different sets

120

110100

90

9 80

S 708 60

50

tJ 40 _\

30-

20

10

10 20 30 40 50 60 70 80 90 100 110 120 130 140Thromboplastin %

FIG. I -Thromboplastin dilution curve: 0.1 ml. of various dilutionsof blood thromboplastin and 0.1 ml. of MA40 CaCI, were addedto 0.1 ml. of normal plasma and the clotting times were recorded.The curve shows the relation between clotting time and thrombo-plastin concentration where a clotting time of 10 seconds istaken to represent 100%° thromboplastin.

s A-

24

- - - - - - -. - I

copyright. on D



j.com/

J Clin P


ownloaded from

http://jcp.bmj.com/

THE THROMBOPLASTIN GENERATION TEST

The observed range in two experiments is shownin Fig. 3. In each of these experiments fivedifferent samples of alumina plasma were testedwith the same preparation of serum and platelets.It will be seen that although the results in twoexperiments are different the range of observationsin one experiment is not wide. To assess, forexample, whether or not a preparation of anti-haemophilic globulin is abnormal the aluminaplasma is prepared simultaneously from normaland from the patient's p'asma and tested with thesame preparations of platelets and serum. Whenexpressed graphically the difference betweennormal and abnormal should be wider than thetotal normal range shown in the shaded area ofFig. 2. If the normal, for example, correspondsto the upper limit, curve C in Fig. 3, then the

FIG. 2.-The normal range of the thromboplastin generation test.The central curve represents the average of 40 different tests.The extreme limits indicate the range of variation obtained whenall of the thromboplastin components are varied simultaneously.The shaded area shows the range of variation obtained in 38 of40 observations in which the alumina plasma was obtained fromdifferent normal subjects, but serum and platelets were fromone subject.

of reagents. It will be seen that a level of 100%of thromboplastin, indicating a plasma clottingtime of 10 seconds, is achieved on an average afteran incubation time of five minutes. This very rapidclotting time cannot be due to the transfer ofthrombin from the incubation mixture because no

significant amount of thrombin is transferred. Inthis test three components are used-platelets,alumina plasma, and normal serum; each of thesemay affect the amount of thromboplastin formedor the speed of its formation. There is thereforea wide range of variability between tests made ondifferent days in which all three reagents are

different (Fig. 2). It might be thought that thislarge variation greatly reduced the usefulness ofthe method, but in practice the test is not used tomeasure differences between two sets of the threecomponents. In any particular case interest iscentred on one component; on antihaemophilicglobulin in haemophilia, on a serum factor inChristmas disease, or on platelets in thrombo-asthenia. We have found that for this limitedobjective the normal range is less.

In tests on 40 samples in which the platelets andserum factor were constant, but the aluminaplasma was prepared from different normalsamples, the observations exceeded the rangeshown by shading in Fig. 2 on only two occasions.Similar results were obtained in 15 samples inwhich the alumina plasma and platelet preparationswere constant but the source of serum was varied.

Incubation time in mnutes

FIG. 3.-The curves show the range of variation in thromboplastinformation in two experiments. In each five samples of normalalumina plasma were tested with the same preparations ofplatelets and serum.

abnormal should lie rather be'ow curve D beforea deficiency in antihaemophilic globulin couldbe deduced. No exact quantitative measure ofdeficiency in any factor can be made by thismethod, but a distinction between normal andabnormal is possible. A rough quantitativemeasure can often be achieved by comparing therelative effects on thromboplastin formation ofdilutions of normal and abnormal samples. Forexample, if a serum factor is thought to bedeficient an approximate measure of the extent ofthe deficiency can be obtained by carrying out thethromboplastin generation test on a 1 in 5 or 1 in10 dilution of the patient's serum and comparingthe resulting curve with a series of curves preparedusing dilutions of normal serum. If the 1 in 10dilution of the patient's serum gives a curve similarto that of the 1 in 100 dilution of normal serum,the patient's serum may be said to have 10% of

5

E

25

9-sI

E9 copyright.

on Decem

ber 29, 2021 by guest. Protected by

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.6.1.23 on 1 F

ebruary 1953. Dow

nloaded from

http://jcp.bmj.com/


the normal thromboplastin-forming capacity. Themethod is not exact because the second serumfactor will be diluted in the normal serum to agreater extent than in the patient's serum. In mostof the conditions studied the difference betweennormal and abnormal was very large, and for diag-nostic purposes finer criteria of distinction wereunnecessary.

Antihaemophilic Globulin Deficiency (Haemo-philia).-When haemophilic plasma is treated withAl(OH)3 and supernatant alumina plasma used toreplace normal alumina plasma in the thrombo-plastin generation test thromboplastin formationis greatly reduced. In Fig. 4 are shown the resultsof carrying out this test with plasma samples fromeight haemophiliacs. In each of these experimentsthe alumina plasmas from the normal and the

140.

120-

100-

80

60

Incubation time ir, minutes

F1G. 4.-Thromboplastin formation in eight cases of haemophilia.The shaded area shows the total range of thromboplastin forma-tion using normal alumina plasma samples tested in parallelwith the samples from haemophilic subjects. The curves shovthe results of replacing the normal alumina plasma by aluminaplasma from haemophilic blood

haemophilic subjects were compared in theirability to form blood thromboplastin using prepar-ations of platelets and serum from normal sub-jects. In these experiments no fine distinctionbetween normal and abnormal was required andthe total observed range of normal variation isshown.The test is more sensitive than other laboratory

tests for haemophilia. Three patients whose whole-blood clotting times and prothrombin-consumptionindices were normal showed grossly reduced abilityto form blood thromboplastin.The sensitivity of the thromboplastin generation

test to the haemophilic defect is well illustrated byexperiments on mixtures of normal and haemo-

12 3 4 5 6 7 8Incubation time in minutes

FIG. 5-The curves show the progress of thromboplastin fornmationusing normal and haemophilic alumina plasma (upper and lowercurves) and using mixtures of the two. The figures are theaverage results of four similar experiments.

philic plasma. When one part of normal is mixedwith nine parts of haemophilic plasma the clottingtime and the consumption of prothrombin duringclotting are usually normal in the mixtures. Onthe other hand, when mixtures of normal andhaemophilic plasma treated with A1(OH):, aretested by the thromboplastin generation test theresults are not normal (Fig. 5). Even 50% ofAl(OH)3-treated normal plasma will not eliminatethe haemophilic defect.When a haemophilic patient is transfused

haemorrhage may continue despite the normalresults of whole-blood clotting time and pro-thrombin consumption tests. On one occasion itwas necessary to remove two teeth from a childaged 61, weighing 40 lb. He was given one pintof fresh plasma and two-thirds of a pint of freshwhole blood before the extractions. His whole-blood clotting time and prothrombin consump-tion index were normal, but he bled rather morethan normal and his thromboplastin generationremained abnormal (Fig. 6). The thromboplastingeneration test may prove to be the most reliableguide to the efficiency of treatment in haemophilia.

Seven known female carriers of the haemophilictrait were tested in the hope that this more

sensitive test might reveal some difference fromnormal, but in no instance was the amount ofantihaemophilic globulin significantly reduced as

judged by the narrower criterion of significancegiven in Fig. 2.

In addition to its sensitivity to the haemophilicdefect the thromboplastin generation test gives a

more specific diagnosis of haemophilia than othertests. The test makes possible a distinctionbetween haemophilia and Christmas disease, whichis not easy using other methods.

26

E

copyright. on D



j.com/

J Clin P


ownloaded from

http://jcp.bmj.com/


Incubation time in minutes

FIG. 6.-The lower curve represents the progress of thromboplastinformation using the alumina plasma collected immediately afterthe transfusion to a haemophilic child of about 600 ml. of freshnormal plasma. The upper curve represents thromboplastinformation using normal alumina plasma.

Deficiency of Serum FactorsChristmas Disease (Biggs et al., 1952).-Christ-

mas disease is a recently recognized conditionclosely resembling haemophilia. It differs fromhaemophilia in that the antihaemophilic globulinis present in normal amounts as tested by thethromboplastin generation test and in that theplasma of patients with Christmas disease willcorrect the abnormality of haemophilic plasma.Thromboplastin formation is deficient in Christ-mas disease due to a reduction in the blood of afactor which differs from the antihaemophilicglobulin and is present in normal and haemophilicserum. The distinction between Christmas diseaseand haemophilia can readily be made using thethromboplastin generation test. When aluminaplasma and platelets are made from normal orfrom the patient's plasma and the diluted serumfrom a patient with Christmas disease is added theamount of thromboplastin formed is reduced (Fig.7). If haemophilic or normal serum is used thegeneration of thromboplastin is normal. Thepatient's blood therefore contains normal amountsof antihaemophilic globulin, but lacks a factorpresent in normal and haemophilic serum.

" Tromexan " Therapy.-The administration of"tromexan " reduces a substance in the bloodcalled factor VII (Koller et al., 1951). Factor VIIis necessary for the action of brain thrombo-plastin; it is obtained from the same fraction ofthe serum as is the Christmas factor. The reduc-tion of factor VII in the blood of patients receivingtromexan is the main factor which controls theone-stage prothrombin time (Douglas, 1953). Ifthe serum of patients treated with tromexan istested with alumina plasma and platelets prepared

Inctibation time in mintilesFIG. 7.-Thromboplastin formation in five cases of Christmas disease.

rhe shaded area shows the total range of thromboplastin forma-tion using normal alumina plasma and serum samples whichwere tested in parallel with the samples from the subjects withChristmas disease. The curves show the results of replacingthe normal serum by the patients' serum.

from normal plasma it is found that the ability toform thromboplastin is reduced parallel to thereduction of factor VII (Fig. 8).The thromboplastin generation test shows that

in the blood of patients with Christmas diseaseand in the blood of patients receiving tromexanthere is a deficiency of a factor necessary forthrombop'astin formation. Christmas disease andthe defect in tromexan therapy are obviously verydifferent. In Christmas disease there is a normalone-stage " prothrombin time" and characteristic-ally a long whole-blood clotting time: in tromexantherapy the one-stage " prothrombin " time is longand the whole-blood clotting time is usuallynormal. It is now clear that these two conditions

too-

80-

60-

40-

20.1

2 4' 6' 8' 10 12 14Time (days'

FIG 8.-The upper curve shows the ability ofthe serum from patientstreated with tromexan to form blood thromboplastin from normalplatelets and alumina plasma. The lower curve shows the factorVII content of plasma samples collected at the same timemeasured by the one-stage technique (Douglas, 1953). Thefigures are expressed as percentages of normal and represent theaverage results obtained with six cases.

T r

.7

c-

0.0E9-C

copyright. on D



j.com/

J Clin P


ownloaded from

http://jcp.bmj.com/


lack different factors, both of which are necessaryfor thromboplastin formation (Biggs, Douglas, andMacfarlane, 1953b; Douglas, 1953).

Functional Platelet Deficiency (Thromboasthenia)The concept of functional platelet deficiency was

put forward by Glanzmann in 1918. Originally,patients who would now be said to havethrombocytopenic purpura or von Willebrand'sdisease were included in the category "thrombo-asthenia." The name " thromboasthenia " shouldbe reserved for the few patients described whohave functionally and morphologically abnormalplatelets. These cases are reviewed by Biggs andMacfarlane (1953).We had the opportunity to examine a patient

whose blood contained giant platelets resemblingthose described in cases of thromboasthenia;these platelets behaved abnormally in the thrombo-plastin generation test (Fig. 9). In this experimentalumina plasma and serum were prepared fromnormal blood and the normal and abnormal plate-lets were compared. In Fig. 9 the normal andabnormal platelet preparations contain approxi-mately equal numbers of platelets. From experi-ments with dilutions of platelet suspensions itappeared that the normal platelets were about tentimes as effective as those from the patient.

Thromboplstin InhibitorsThere are now a number of reports of patients

who acquire a disease closely resembling haemo-philia in adult life. We have found records ofabout 15 patients with this type of abnormality.These patients may have in their blood an anti-

Incubation time in minutes

FIG. 10.-The curves show the thromboplastin generated from amixture of normal alumina plasma, serum, and platelets, andfrom the same mixture to which was added various concentra-tions of alumina plasma containing an inhibitor.

coagulant which interferes with intrinsic thrombo-plastin formation. In addition, true haemophilicpatients may develop a very similar anticoagulant,possibly as a result of immunization followingtransfusion. There are records of at least 12 suchpatients.The thromboplastin generation test has proved

very useful for defining the site of action of theanticoagulant. We have now examined the bloodof four patients; in three the disease arose in adultlife, with no previous history of a haemorrhagicdiathesis; the other patient had haemophilia. Inall cases it was found that the anticoagulantinhibited the formation of active thromboplastinbut did not interfere with the ability of thisthromboplastin to form thrombin from prothrom-bin. The result of an experiment on thrombo-plastin generation is shown in Fig. 10. In thisexperiment alumina plasma, platelets, and serumwere prepared from normal blood. In three testsfinal dilutions of 1 in 100, 1 in 250, and 1 in 500of the patient's alumina plasma were added to themixtures forming thromboplastin. It will be seenthat at a concentration of 1 in 250 the patient'splasma inhibited the formation of plasma throm-boplastin. The patient's serum was similarlyinhibitory.

In these three cases the inhibitor apparentlyinterfered with the activity of both antihaemophilicglobulin and the Christmas factor, because neitherantihaemophilic globulin activity nor serum acti-vity could be demonstrated in the patient's blood.In the fourth case, a woman aged 46, the serumactivity was normal but the patient's plasma

12 3 4 5 6 7 8

Incubation time in minutesFIG. 9.-The curves show the thromboplastin generated from normal

alumina plasma and serum using normal platelets (upper curve)and the abnormal platelets from a patient with thromboasthenia(lower curve).

28

copyright. on D



j.com/

J Clin P


ownloaded from

http://jcp.bmj.com/


TABLE IRESULTS OF THE ONE-STAGE " PROTHROMBIN " TIME AND THE THROMBOPLASTIN GENERATION TEST ON

PATIENTS WITH VARIOUS COAGULATION DEFECTS

Type of Coagulation Defect

Technique ~~~~~~~~~~Source of Patient withTechnique Thromboplastin Components Haemo- Christmas Factor V Factor VII Thrombo- Thrombo-

philia Disease Deficiency Deficiency asthenia plastinInhibitor

One-stage " pro- Ithrombin" time - Normal Normal Prolonged Pro'onged Normal Normal

Thromboplastin Alumina Serum Platelets - _-generation test plasma

Patient Normal Normal Abnormal ,, Abnormal Normal Abnormal

Normal Patient ,, Normal Abnormal Normal Abnormal ,, May be_ __________________________________________________________ ___________________________ abnormal

r,,No Normal Patient N,,ormal ,, Normal j Abnormal Normal

lacked antihaemophilic globulin. In this case itappears that the inhibitor specifically neutralizedantihaemophilic globulin. Factors V and VII wereunaffected by the inhibitor in all these casesbecause the one-stage " prothrombin " time usingbrain thromboplastin was normal.

DiscussionThe thromboplastin generation test is technic-

ally simple, the final test system being essentiallya series of one-stage " prothrombin " times. Thepreparation of the platelet suspension is time-consuming, but not more so than the preliminarystages in, say, the red-cell fragility test. Since thethromboplastin generation test is not required veryfrequently in ordinary routine diagnosis this dis-advantage is not important. Used in appropriatecases, the test has great advantages over any pre-viously described technique. Using this test, it ispossible to differentiate between the abnormalitiesin which there is a reduction in the formation ofintrinsic thromboplastin. Moreover, the test maybe a much more sensitive indicator of abnormalitythan any other test.

Using the thromboplastin generation test, it ispossible to make a certain diagnosis of haemophiliaeven in mildly affected individuals, and it maybe the most reliable index of the effectiveness oftreatment in haemophilia. The thromboplastingeneration test makes possible a definite differen-tiation between haemophilia and Christmas disease,and this distinction is of practical importancebecause the treatment of the two conditions isdifferent. In addition, the thromboplastin gener-ation test can be used to detect abnormal plateletfunction and to define the site of action ofanticoagulants which depress thromboplastinformation. For these reasons we believe that thethromboplastin generation test may prove an un-usually valuable test of clotting function. The

results of carrying out the test together with theone-stage "prothrombin" time on the blood ofpatients with various clotting abnormalities aresummarized in Table 1.

SummaryThe thromboplastin generation test is described

in detail. This test is the most reliable method forthe diagnosis of haemophilia because, using this test,cases which are mildly affected can be diagnosedwith certainty and because it permits a differen-tiation between haemophilia and the closely relatedcondition, Christmas disease. Being the mostsensitive test for deficiency of antihaemophilicglobulin, the thromboplastin generation test givesthe best indication of the effectiveness of treat-ment in haemophilia.The thromboplastin generation test has also

proved valuable in the study and diagnosis ofpatients with Christmas disease, with thrombo-asthenia, and with circulating anticoagulants whichinhibit the normal intrinsic thromboplastin system.

We should like to thank Miss G. Richards fortechnical assistance, Dr. R. G. Macfarlane for hisadvice and interest in this work, and Dr. A. H. T.Robb-Smith, who kindly read the manuscript. Oneof us (A. S. D.) wishes to thank the Medical ResearchCouncil for a fellowship in clinical research enablinghim to carry out this work.

REFERENCESBertho, A., and Grassmann, W. (1938). Laboratory Methods of

Biochemistry. Macmillan, London.Biggs, R.,Douglas, A. S., and Macfarlane, R. G. (1953a).-. Physiol.

In the press.(1953b). Unpublished observations.Dacie. J. V., Pitney, W. R., Merskey, C., and O'Brien.

J. R. (1952). Brit. med. J., 2, 1378.and Macfarlane, R. G. (1953). Human Blood Coagulation and

its Disorders. Blackwell, Oxford. To be pub:ished in February.Douglas, A. S. (1953). Unpublished observations.Glanzmann, E. (1918). Jb Kinderheilk, 88, 1 13.Koller,F.,Loeliger,A.,andDuckert, F.(1951). ActaHaemat., Basel,

6, 1.Merskey, C. (1950). Journal of Clinical Pathology, 3, 301.

(1951). Brit. med. J., 1, 906.

29

copyright. on D



j.com/

J Clin P


ownloaded from

http://jcp.bmj.com/

the thromboplastin generation

Documents