Reappraisal of Thromboplastin

Transcription

1 ANNALS O F CLINICAL AND LABORATORY SCIEN CE, Vol. 11, No. 3 Copyright 1981, Institute for Clinical Science, Inc. Reappraisal of Thromboplastin F. D. ZIEGLER, Ph.D.,* S. A. RICH, Ph.D.,* M. J. FASCO,* CAMERON RUSSELL, F.I.M.L.S.,1 and J. H. KELLY, M.D.f *Division o f Laboratories and Research, New York State Department o f Health, Albany, NY and \Bender Hygienic Laboratory, Albany, NY ABSTRACT Proficiency testing surveys in the state of New York indicate that despite increased sophistication in instrumentation, there has been no real improvement in interlaboratory reproducibility in prothrombin-time determinations over the last 10 years. This lack of improvement is most pronounced in the therapeutic range of 20 to 30 sec. One reason may be that thromboplastins are variable from one manufacturer to another and even between types produced by the same manufacturer. While it has been possible in several laboratories to synthesize experimentally a thromboplastin with known content of lipid and active protein, no efforts have been made to make such a product commercially available. Blood levels of warfarin were measured but cannot be reliably used to monitor anticoagulation. In a preliminary study, factor Xa activity was measured using chromogenic substrate S2222. Factor Xa activity gave a positive correlation with prothrombin times of patients receiving warfarin therapy. Chromogenic substrate factor assays may represent a future method of choice for controlling anticoagulant therapy. Introduction ticoagulant therapy, at an estimated cost of Throm boplastins are tissue extracts over $100 million. This estimate is based which, in the presence of calcium, initiate solely on the population and frequency of the clotting activity of plasma. In the laboratory, thromboplastin is important for its ups for liver disease or bleeding disorders testing; taking account of routine work use in the prothrom bin-tim e te st to would vastly increase the total. monitor or control anticoagulant therapy Warfarin has been widely used for over and as a tool to investigate bleeding disorders. Extrapolating from a European re myocardial infarction, and atherosclero three decades to treat thromboembolism, port,8 it is estimated that, in the United sis. Although doubts have been cast on its States alone, more than 45 million tests efficacy,11,26 statistical evidence still supports its use.6,12,32 Intensive research dur are performed each year to control an /81/ $01.50 Institute for Clinical Science, Inc.

2 REAPPRAISAL OF THROMBOPLASTIN 203 ing the past 10 years has established the biochemistry of warfarin suppression of vitamin K-dependent coagulation factors, but prothrombin-time control of therapy remains unchanged. Accurate and reproducible prothrom bin-tim e tests are needed especially to avoid complications in the use of warfarin and phenindione in anticoagulant therapy. Among hospital drugs used in the United Kingdom, these cause the second and third highest numbers of treatment-related deaths.13 The prothrombin-time determination, devised in its present form by Quick in the mid-1930s,2s is a technically easy procedure. Thromboplastin (an aqueous suspension of acetone- or saline-extracted tissue, usually brain and/or lung) is added to plasma together with calcium. The end p o in t is a solid fibrin clot. The p ro thrombin-time determ ination is dependent on the extrinsic coagulation system2,3,34 and involves three of the four vitamin K -dependent clotting factors. Factor VII is present in the least amount, and its turnover is the fastest. It is, therefore, rate-lim iting w hen therapy is initiated. Thromboplastin activates factor VII to factor Vila, which then activates factor X to factor Xa. Factor Xa joins with factor V and Ca++ to convert prothrombin to thrombin. Thrombin hydrolyzes four small peptides from fibrinogen to form fibrin, which aggregates to form the clot. Sufficient fibrinogen and factor V must be present to produce a detectable clot, and no interfering substances may be present. When these conditions are met, the prothrombin time reflects the activity level of three vitamin K-dependent factors: VII, X, and II. Under a New York State-mandated proficiency testing program, the performance of prothrombin-time tests by most clinical laboratories in New York State was surveyed by us. In these surveys, commercial plasma samples with normal and abnormal prothrombin-time target values are supplied to participants. An early survey (1969) showed that the results were quite variable,33 with coefficients of variation of 10 percent for normal-range target values of 10 to 12 sec and 11 percent for a therapeutic-range target value of about 30 sec. In the survey of 1979 reported here, the coefficient of variation had decreased to 8 percent for the normal range but had increased to 14 percent for the therapeutic range. Apart from poor technique, this problem in achieving uniformity of prothrombin time may be due either to instrumentation and methods or to the source of thromboplastin. Instrumentation Advances in the technique for measuring prothrombin time have been primarily in the development of automated devices to detect fibrin strands. The number of automated instruments in common use has tripled over the last 10 years. There are now 10, six mechanical and four photo-optical. The most recent instruments are photo-optical.4,30 A survey in 1978 by the College of American Pathologists (CAP) reported analyses with six automated instruments and nine thromboplastins. Of the 54 possible instrument-thromboplastin combinations, only 24 were used frequently enough to be treated statistically. Over 3000 laboratories, about 50 percent of the total in the CAP survey, used combinations com mon to 18 or few er other laboratories. Sim ilarly, a survey of laboratories in New York State showed that only a few method-thromboplastin combinations were used in sufficient numbers to be treated individually (table I). Only the data for major categories, such as the manual method with brain-lung thromboplastins, could be analyzed statistically with such an approach. The effects of thromboplastin source and method on prothrombin time in a recent (1979) New York State survey are shown in figure 1. Standard plasma samples with a target value of 30 sec, all from the same lot (purchased from General Diagnostic Division, Warner Lambert &

3 204 ZIEGLER, RICH, FASCO, RUSSELL, AND KELLY TABLE I Thromboplastin-Instrument Combinations Reported by 303 New York State Laboratories, Survey III, 1979 Method and Instrument Hyland* Dried Simplastinf A Simplastint Automated Simplastini Dadei Activated Dadeî C OrthoH Pure Brain Manual 4 _ Automated-Mechanical Fibrometer 1I1I Clotek* Clot timer Not specified Automat e d-photo-opt ica1 Coagulation profiler** Coag-A-Matet _ Coagulizer// Electra+t Not specified *Hyland Diagnostic Division, Travenol Laboratories tgeneral Diagnostics Division, Warner-Lambert Dade Diagnostics KOrtho Diagnostics //Sherwood **BioData t+medical Laboratory Automation Mechrolab 1ÍHBaltimore Biological Laboratory Co.), were distributed to 312 laboratories. W hen various throm boplastins w ere compared regardless of method, one (Hyland, dried brain),* which had relatively few users, seemed to be least variable. This thromboplastin gave a mean value of 25 seconds, lower than the other products tested. Several thromboplastins, which were individually less popular (Thromboplastin 1 in figure 1A), showed the greatest variability. The most popular product (Ortho Pure Brain) f also showed a wide range. When prothrombin time was graphed by method, regardless of thromboplastin source, all methods appeared variable (figure IB). The manual method was actually slightly less variable than the mechanical or optical detection methods. Similar results have led the British to recommend a return to manual methods.24 Thromboplastin The throm boplastin used in the prothrombin-time test is an acetone or * H yland D iag n o stics D iv isio n, T ravenol Laboratories, D eerfield, IL f Ortho Diagnostics, Raritan, NJ saline extract of animal or human tissues; the highest yields are from brain, placenta, or lung. The extract is a mixture of phospholipid and protein associated with the microsomal membrane fraction. The preferred thromboplastin composition and source, whether animal or human, varies from one country to another and from one commercial supplier to the next. One manufacturer! in the United States supplies three rabbit brain-lung thromboplastins; the brain components are extracted with acetone, the lung components with saline. Another^ supplies several pure-brain acetone-extracted preparations. These products have been compounded to meet various needs, most recently the demands of photo-optical instruments.30 Rabbit brain and brain-lung combinations are used in the United States and Canada, but there is no consensus either nationally or internationally on thromboplastin source. R eference throm boplastins are used by the Netherlands, U n ited K ingdom, and A ustralia to t G eneral Diagnostics, Morris Plains, NJ Dade, Miami, FL

4 REAPPRAISAL O F TH ROM BOPLA STIN 205 SECONDS F ig u r e 1. D istribution of prothrom bin times from 312 New York State Laboratories in proficiency testing survey IV, Plasma samples (target value of 30 sec) were purchased from General Diagnostics and d is trib u te d as unknow ns to N ew York L aboratories perform ing prothrom bin tim es. A: comparison by throm boplastin source. Thromboplastin 1 was the com bined results from thromboplastins used by too few laboratories to evaluate alone; Thromboplastin 2, Hyland D ried Thromboplastin; 3, Ortho Pure Brain Throm boplastin; 4, D ade Activated Thromboplastin; 5, General Diagnostics Simplastin. B: comparison by method. equalize interlaboratory differences. The N etherlands Thrombosis Service uses predom inantly T hrom botest,21 w ith whole blood rather than plasma as a standard.8 The British correctly m aintain that throm boplastin is speciesspecific.23,24 T hey feel that throm boplastins derived from rabbit or bovine tissue lack human specificity, and strongly support a human brain product as a standard, using ratios of normal to abnormal plasma prothrombin times and activity coefficients to compare results between laboratories.24 Australia has followed the exam ple of B ritain and provides a phenol-stabilized saline suspension of human brain as the Australian Reference Thromboplastin.1 In the United States, a standard throm boplastin has b een rejected as too complex for agreement by com m ercial sources, and a standard plasma is used instead.17,19 Canada follows the practices of the United States, while Scandinavian countries use a system based on Thrombotest, Normotest, or the P and P test. These reagents contain bovine or rabbit brain thromboplastin supplem ented with bovine factor V. Our observations of widely variable results in prothrombin-time determinations in 1969 prompted us to fractionate several commercial thromboplastins. At that time great emphasis was placed on the phospholipid composition.15,27 It was shown33 that the phospholipid classes of the most popular com m ercial throm boplastins were the same and that those with the highest percentage of phosphotidyl ethanolamine were the most active. Phosphotidyl ethanolamine then was considered the most important phospholipid com ponent of throm boplastin.16,27 Recently, however, Wijngaards et a l31 demonstrated that the charge on the phospholipid must be about 7 mv. These workers found that several phospholipid combinations, each with a net charge of about - 7 mv, were all highly active. It may be concluded that contamination of supposedly pure phosphotidyl ethanolam ine fractions w ith other phospholipids was responsible for the activity reported by us for this fraction. The assumption that a protein fraction is necessary for thromboplastin activity has recently been further investigated and refined, largely by Nemerson s group.16,20,22 Using differential centrifugation, it was dem onstrated that tissue factor is con-

5 206 Z IE G L E R, R IC H, FA SC O, R U S S E L L, AND K E LL Y Commercial Thromboplastin 3000 X g - 30' Supernatant Ppt. discard 105,000 X 1 hr Super discard Ppt. Wash 3 X in PBS Super Ppt percent DOC 105,000 X g - 4 Super Dialyze vs Tris buffered NaCl Ppt. discard Ppt. with TCA, SDS, PAGE F ig u r e 2. F lo w diagram o f p roced u re for iso la tio n o f p ro tein from co m m e rcia l th rom b op lastin. DOC: so d iu m d eo x y ch o la te; PAGE: p o ly a cry la m id e g el-electro p h o r esis; SDS: so d iu m d o d e c y l sulfate; TCA: trich lo ro a cetic acid. tained in the microsomal membrane frac tion of many tissues, including brain and lung. It is also present in blood vessel BL PB F ig u r e 3. Electrophoretogram s of protein frac tion from commercial throm boplastin. BL, brainlu n g c o m b in a tio n (G e n e ra l D ia g n o stic s, Sim plastin), PB, pure brain (Ortho, Pure Brain Thromboplastin). intima, both in endothelial cells and in the atherosclerotic plaque. Isolation of the protein fraction of rabbit brain and brain-lung thromboplastins was achieved by us following the procedure shown in figure 2. The protein was sol ubilized in sodium dodecyl sulfate and electrophoresed on poly-acrylamide gel according to published procedures.28 The electrophoretogram (figure 3) shows that a major band and one minor band are com mon to both types of throm boplastin. Other minor bands are different between thromboplastins. The two major bands may co rresp o n d to those in bovine throm boplastin which have m olecular weights of 220,000 and 330, This protein is specific and necessary for thromboplastin activity The corre sponding human brain apoprotein is of lower solubility and molecular weight (55,000).5 The protein components vary from species to species,5and often display differing activity. This is in part the rationale of the British group, who elect to use hum an brain as a source for thromboplastin. It can now be stated with certainty that active throm boplastin is a lipoprotein from the microsomal fraction of many tis sues, in c lu d in g v essel w alls. Both phospholipid and protein are necessary for activity; neither purified component is active alone. It has been suggested7' 14that hydrophobic lipid-protein interactions are the basis for formation of the active complex. In this scheme (figure 4), the phospholipid forms a vesicle, with the charged portion orien ted toward the aqueous phase (plasma). This vesicle without protein is inactive. However, the protein has considerable hydrophobicity,3 and the hydrophobic portion can be in serted into the nonpolar portion of the lipid vesicle. This arrangement provides a specific surface for binding factor VII in the proper conformation to cleave an ArgIle bond, releasing a peptide and leaving behind the active factor. It may also help to align factor Xa, factor V, and prothrom bin in the correct spatial conformation to

6 REAPPRAISAL O F THROM BOPLA STIN 207 Phospholipid Vesicle Protein Lipoprotein F ig u r e 4. Schematic diagram of combination of lipid and protein components of thromboplastin to form a surface for the interaction of coagulation factors. Diagram based on data from Baugh, R. F. and Houghie, C. Clinics Hemat. 81 :3-30, F i g u r e 5. H igh-perform ance liq u id chro m atogram s of (A) n o r m al serum spiked w ith a warfarin standard and (B) serum from a patient on w arfarin th erap y. Peaks are detected by absorbance at 313 nm. Retention time in seconds is given beside the peak(s) of interest. Warfarin is the peak with retention time of 666 sec.

7 208 Z IE G L E R, R IC H, FA SC O, R U SSE L L, AND K ELLY and recombine the specific protein and the active phospholipids of thromboplastin. It should, therefore, be possible to produce a standard, specific, sensitive, thromboplastin. So far this has not been attempted on a large scale. PROTHROMBIN TIME (sec) F ig u r e 6. C orrelation o f prothrom bin tim e and warfarin concentration in plasm a from patients rec e iv in g therapy. release active thrombin. Factors VII, X, and II are bound to the lipoprotein surfaced by Ca++ bridges between carboxyglutamic acid residues of the protein and phospholipid.34 These studies and other WOrk16'20 22,31 demonstrate that it is possible to isolate POOLED NORMAL PLASMA (p\) F ig u r e 7. Factor Xa activity of serial dilutions of pooled normal plasma m easured with substrate S2222. Factor X was activated with Russell s viper venom and activity was measured as described in the text. Plasma Warfarin The diversity of methods and thromboplastins for prothrombin-time tests has led to a search for alternatives to thrombop lastin for m onitoring anticoagulant therapy. On the rationale that one can measure therapeutic blood levels of a drug as is done with digoxin, theophylline, and D ilantin high performance liquid chromatography was used to measure warfarin levels in sera from normal subjects and from patients receiving warfarin. Normal serum spiked with a warfarin standard and serum from a patient receiving warfarin therapy are compared in figure 5. While it is possible to quantitate the warfarin in the patient sample, a number of other peaks, possibly due to other drugs or caffeine, are also present. Even though warfarin in serum was measurable, there was no apparent correlation betw een warfarin level and anticoagulation effect (figure 6). This might be expected because of variations in the human response to and rate of metabolism of warfarin. Also, warfarin binds strongly to albumin and only the small unbound portion is active. Drugs which displace warfarin from albumin enhance its anticoagulant effect w hile drugs which stimulate hepatic enzymes enhance degradation of warfarin and, hence, reduce its anticoagulant effect. A wide variety of xenobiotic compounds affect warfarin pharmacokinetics and pharmacodynamics,10 so it appears that measurements of warfarin in the blood can make no contribution to management of warfarin therapy. This position is also held by Duffield et al.9 Chromogenic Substrates A more hopeful alternative may be the use of synthetic substrates for coagulation

8 REAPPRAISAL O F THROM BOPLA STIN 209 factors.29 Each has an amino acid sequence which makes it highly specific for one of the active coagulation factors. The feasibility was investigated by using one of these commercially available substrates, S 2222,* which is specific for factor Xa. Plasma was collected from 15 patients receiving warfarin therapy. A pool of plasma from 10 normal subjects was prepared and used as a control. All plasmas were collected in citrate and frozen at -20 until tested. The normal plasma pool suffered no loss of activity after freezing and thawing. For testing, factor X in buffered plasma was first activated by incubation with Russell viper venom (1:10,000) and CaCl2 at 37 for five min. Substrate S 2222 was added, and the 37 incubation was continued for an additional 5 min. The reaction was arrested with excess acetic acid, and the 405 nm absorbance was read in a spectrophotom eter.f A ctivity varied linearly with serial dilutions of the pooled normal plasma (figure 7) and was thus apparently Xa-dependent in agreement with the findings of Lammle et al.18 Patients receiving warfarin had much less factor Xa activity than the normal plasma pool (figure 8). Interestingly, the patients with the longest prothrombin time had the least factor X activity. The correlation coefficient of the regression line was 0.84, equivalent to that observed by Lammle et al.18 While the number of patients was small, enough correlation was observed to warrant further investigation. It must be remembered, however, that the method is sensitive only to factor Xa while warfarin reduces the activity of factors II, VII, and IX as well as factor X. The prothrombin time test, in use in its various forms for over 40 years, is sensitive to three of the four factors suppressed by warfarin and had the advantages of reasonable cost and ease of performance. * M anufactured by AB-Kabi, Stockholm, Sweden and distributed in the U nited States by Ortho Diagnostics, Raritan, NJ f Beckman Model 25. Beckman Instrum ents Co., Fullerton, CA FACTOR Xb ACTIVITY (% pooled normal plasma) F ig u r e 8. Correlation of factor Xa activity and prothrombin tim e factor Xa activity was m easured as in figure 7. Pooled normal plasm a = patient plasmas =». Although it has been used for years, interlaboratory standardization, both nationally and internationally, has proved an intractable problem. Chromogenic substrates are now, and for the foreseeable future will be, more costly and their use more technically dem anding than thromboplastins and prothrombin time determinations. Still, the fact that they are based on a specific enzyme reaction may gain them acceptance in the search for standard, reliable, and reproducible laboratory m ethods for th e control of anticoagulant therapy. References 1. Ba in, B., F o r s t e r, T., and Sl e ig h, B.: Control of oral anticoagulants by prothrombin time: A plea for uniformity. Med. J. Australia 2 : , Ba u g h, R. F. and H o u g ie, C.: Biochemistry of blood coagulation. R ecent A dvances in Blood C oagulation. Poller, L., ed. London, Churchill-Livingstone, 1977, pp Ba u g h, R. F. and H o u g ie, C. : The chemistry of blood coagulation. Clinics Hemat. 81 :3-30, B e c k a l a, H. R., L e a v e l l e, D. E., and Di- DISHEIM, P.: A comparison of five manually operated coagulation instrum ents. Amer. J. Clin. Path. 70:71-75, Bjo r k l id, E., S t r o m, E., and P r y d z, H.: The protein com ponent of hum an brain thrombop lastin. B iochem. B iophys. R es. C om m. 55: , C h a n d le r, A. B., C h a p m a n, I., E r h a r d t, L. R., R o b e r t s, W. C., S c h w a r t s, C. J., S m apius,

9 210 Z IE G L E R, R IC H, FA SCO, R U SSELL, AND KELLY D., S p a in, D. M., Sh e r r y, S., N e s s, P. M., and SIMON, T. L.: Coronary thrombosis in myocardial infarction. Amer. J. Cardiol. 34: , D a v ie, E. W. and F u jik a w a, K.: Basic m echanisms in blood coagulation. Ann. Rev. Biochem. 44: , D ij k -W i e r d a, C. A. v a n, H e r m a n s, J., L o e l ig e r, E. A., and ROOS, J.: Inter-laboratory oral anticoagulant quality a ssessm en t b y the N etherlands F ederation o f T hrom bosis Services. Throm. H aem ostas. 37: , D u f f i e l d, P., D u f f ie l d, A. M., Ke n n e d y, M., B e r k e t t, D. J., and W a d e, D. N.: Warfarin and warfarin-alcohol levels in anticoagulated patients. Aust. New Zeal. J. Med. 9: , F a s c o, M. J., C a s h in, M. J., and Ka m in sk y, L. S. : A novel method for the quantitation of warfarin and its metabolites in plasma. J. L iquid Chromatog. 2: , F l e t c h e r, A. P., A lk ja er sig, N. K., G h a n i, F. M., T u l e v s k i, V., and O w e n s, O.: Blood coagula tio n system p ath o physiology in acute m yocardial infarction: The influence of anticoagulant treatm ent on laboratory findings. J. Lab. Clin. Med. 93: , F r ie d b e r G, C. K.: Diseases of the Heart, 3rd ed. Philadelphia, W. B. Saunders Co., 1966, p G u m p e l, J. M.: Deaths associated w ith gold treatm ent, a reassessment. Brit. Med. J. 1 : , H a s e g a w a, H., H a g t a, H., and M u r a o, M.: Studies on the tissue thromboplastin during th e coagulation-fibrinolytic process ultrastru ctu ral changes. T hrom b. H aem ostas. 37: , H e l g e l a n d, L.: The submicrosomal site for the conversion of prothrom bin precursor to biologically active prothrombin in rat liver. Biochim. Biophys. Acta 499: , Je s t y, J., M a y n a r d, J. R., Ra d c l if f e, R. D., Sil v e r g e r g, S. A., Pit l ic k, F. S., and N e m e r - SON, Y.: Proteases and Biological Control. Cold Spring, NY, Cold Spring Harbor Laboratory, , p p Ko e p k e, J. A., Gil m e r, P. R., T r ip l e t t, D. A., and O S u l l iv a n, M. B.: T h e prediction of prothrom bin tim e system performance using secondary standards. Amer. J. Clin. Path. 68: , L am m le, B., E ic h l is b e r g e r, R., H a n n i, L., B o u n a m e a u x, H., M a r b e t, G. A., and Duc- KERT, F.: Kontrolle d er oralen Antikoagulation: V ergleich zw isch en Q uick- und kolorim etrischer Faktor-X -B estim m ung b ei 107 Patienten. Schweiz. M ed. Wschr. 109: , Mia l e, J. B. and La F o n d, D.: Prothrombin tim e standardization. Proposal of the Standards Committee, College of American Pathologist, Subcommittee on Coagulation Reagents. Amer. J. Clin. Path. 52: , N e m e r s o n, Y and PiTLICH, F. S.: Purification and characterization of the protein component of tissue factor. Biochem istry 9: , ,, 21. O w r e n, P. A.: T he history of throm botest. Human Blood Coagulation. Hem ker, H. D., Loeliger, E. A., and Veltkamp, J. J., eds. New York, Springer, 1969, pp P it l ic k, F. A. and N e m e r s o n, Y.: Binding of the protein com ponent of tissue factor to phospholipids. Biochemistry 9: , P o l l e r, L. and T h o m s o n, J.: The M anchester comparative reagent: A national standard for anticoagulant therapy. Human Blood Coagulation. Hemker, H. C., Loeliger, E. A., and Veltkamp, J. J., eds. New York, Springer, 1969, pp P o l l e r, L., T h o m s o n, J. M., and Ye e, K. F.: Quality control trials of prothrombin time. J. Clin. Path. 32: , QUICK, A. J.: The prothrombin in hem ophilia and in obstructive jaundice. J. Biol. Chem. J09:LXXIII, Ro b e r t s, W. C. and Bu ja, L. M.: The frequency and significance of coronary arterial thrombi and other observations in fatal acute myocardial infarction. A study of the 107 necropsy patients. Amer. J. Med. 52: , R o u s e r, G., W h it e, S., and Sc h l o r e d t, O.: Phospholipid structure and thromboplastin activity. I. The phospholipid fraction active in recalcified normal human plasma. Biochim. Biophys. Acta 28:71-87, Spa r k s, C. E., D e h o f f, J. L., Ca p u z z i, D. M., P ie t r a, G., and Ma r sh, J. B.: Proteolysis of very low density lipoprotein in perfused lung. Biochim. Biophys. Acta 529: , Van Wijk, E. M., Ka h l e, L. H., and T e n Ca t e, J. W.: The determination of factor X using an autom ated am idolytic technique. T hrom b. Haemostas. 42:55, W EISBRO T, I. M. and W a l d n e r, D. K.: Evaluation of a single-channel photo-optical clottim in g in stru m en t. A m er. J. C lin. Path. 68: , W ijn g a a r d s, G., Va n D e e n e n, L. L. M., and H em k er, H. D.: Reconstitution and lipid requirem ents of porcine tissue thromboplastin. Biochim. Biophys. A cta488: , W r ig h t, I. S., M a r p l e, C. D., and Be c k, E. F.: Myocardial Infarction. Its Clinical Manifestations and Treatm ent W ith Anticoagulants. New York, Grune and Stratton, Inc., Z ie g l e r, F. D. and Ke l l y, J. H.: The critical evaluation of thromboplastin. Ann. Clin. Lab. Sci. 2:16-25, Z w a a l, R. F. A.: M embrane and lipid involvem ent in blood coagulation. Biochim. Biophys. Acta 515: , 1978.