Thromboplastin--sensitivity, precision and other characteristics

A more sensitive or higher concentration of rabbit brain thromboplastin does not result in greater accuracy and precision of results in oral anticoagulant therapy and is unable to mimic the PIVKA sensitivity of human brain. In terms of International Normalized Ratios the British Comparative Thromboplastin and Manchester Comparative Reagent (both now discontinued) and the Manchester Reagent had the poorest sensitivity to factor VII of all the reagents studied. It is not possible accurately to calibrate rabbit brain against human brain thromboplastin in the upper therapeutic range and beyond.     

Comparative study of a portable monitor for prothrombin time determination, Coaguchek, with three systems for control of oral anticoagulant treatment.

Self-testing of oral anticoagulation is a new possibility related to the development of portable capillary whole blood prothrombin time monitors. The aim of this study was to evaluate one of this monitors, Coaguchek, with respect to its comparability with our routine prothrombin time determination system, as well as with the reference manual technique and two thromboplastins of high sensitivity, Manchester Reagent and one manufactured in our center, Thromboplastin Bilbao, in a group of patients on oral anticoagulant treatment. Although a correlation of r = 0.9271 was found between international normalized ratio (INR) values of Coaguchek and our routine method, Neoplastine/STA analyzer, the difference of the INR scatter increased with the magnitude of measurement, being lowest for INR between the portable monitor and Manchester Reagent and Thromboplastin Bilbao, with a similar coefficient of correlation, r = 0.8948 and r = 0.8905, respectively. A test was performed showing a 65.6% agreement with the INR values of the STA analyzer, 66.4% with Manchester Reagent and 73.4% with Thromboplastin Bilbao. On the basis of this correspondence with laboratory prothrombin time results Coaguchek may be considered as a possible option for monitoring anticoagulated patients even though patients should be given instructions and advice as regards the management and interpretation of the results.     

The Interpretation of Prothrombin Results: A National Survey

The practical value of a formula for the interpretation of prothrombin time results has been studied in a national survey incorporating 132 of the hospital centres in our reference thromboplastin scheme. In this formula the 'equivalent ratio' is used to calculate the 'Index of Anticoagulation' in an attempt to provide a uniform scale of measurement for all hospitals.
In our survey the calculated results using the formula were compared with the actual observed values. Only 15 hospitals were able to calculate the 'Index of Anticoagulation'; of these only nine sent a complete set of results. Incomplete schedules were interpreted by us.
The survey showed that the value of the formula was least with commercial reagents, but even with human brain extracts the difference between calculated and observed results was considerable. From this study it is concluded that the term 'equivalent ratio' is of considerable value in measuring the sensitivity of a tissue extract thromboplastin, but nothing is to be gained from the vise of the formula for the 'Index of Anticoagulation'in the Manchester scheme for providing uniform anticoagulant treatment.     

The Variability of Measurements of the Prothrombin Time Ratio in the National Quality Control Trials: a Follow-up Study

Summary . British hospitals which use the one-stage prothrombin time tat my participatc in one of two schemes. Those in the Routine Supply Scheme use a standardized tissue thromboplastin, the Manchester Comparative Reagent (MCR). Those in the National Referencc Scheme use other reagents routinely, but calibrate thcm against the British Comparative Thromboplastin (BCT) which consists of batches of thc MCR which have been checked for accuracy by a group of expert Monitoring Centrcs selected from the hospitals in each scheme. The performance of hospitals in both schemes is monitored by trials in which they are asked to measure prothrombin time ratios, using BCT, on lyophilized plasma preparations which are sent to them. Four trials are reported here. In one, a single plasma preparation was used. In each of thc other three, samples of three plasmas were sent. There was considerable variation in the ratios reported by different hospitals for each plasma. In the trials involving three plasma samples, a marked tendency was demonstrated for hospitals which obtained a particularly high or low reading for one plasma to deviate in the same direction when measuring other plasmas in the same trial, but this was not found when readings from different trials (separated by several months) were compared. A breakdown of the results by type of hospital suggested that those obtained by the participants in the Routine Supply Scheme, and especially by the Monitoring Centres m this group, were more accurate than those of the National Reference Scheme hospitals, but within each group no persistent differences in accuracy between hospitals and no general improvement between the earlier and later trials could be demonstrated. Confidence limits estimated com the results of all mals suggest that even in Routine Supply Scheme hospitals, the chances of having a true prothrombin time ratio outside what is usually quoted as the therapeutic range (1.8-3.0) should be regarded as higher than I in 20 for every patient whose reported ratio @ad on duplicate readings) is below 2.1 or above 2.45. It is concluded that the National Reference Scheme should be phased Out in favour of the Routine supply Scheme, and that even among most of the hosPimls in the latter scheme there is scope for greater accuracy, which the experience Of the Correspondence: Dr L. Poller, National Reference Laboratory for Anticoagulant Control Reagents, Withington Hospital. Manchester M20 8LR.     

Quality Control Trials in the National Reference Thromboplastin Scheme

S ummary . In British hospitals routine measurements of prothrombin times are made either with a standardized thromboplastin reagent (the Manchester Comparative Reagent–MCR) or with the hospitals' own reagents, which have first been calibrated in terms of the national reference thromboplastin (the British Comparative Thromboplastin-BCT). The fundamental assumption in the National Reference Scheme is that hospitals will obtain standardized results with the uniform technique and the BCT. Four collaborative studies have been conducted to evaluate this assumption. Instead of assessing their own local reagents hospitals were asked to calibrate given suspensions of unknown thromboplastin in terms of the BCT. In the last trial they were also asked to test three lyophilized plasma preparations. Since aliquots of given thromboplastin suspensions and dried plasma samples were distributed, all hospitals should have obtained similar results. There was, however, a wide variation, both in the ratios with the unknown thromboplastins that were equated with particular ratios with the BCT, and in the readings for the plasma samples. This variation was partly due to a tendency for some hospitals to obtain consistently higher readings than others. The situation might be improved by extending technical training in the prothrombin time test. If the differences between hospitals could be shown to be stable over long periods, a new system of quality control, involving the routine use of lyophilized plasma in addition to the BCT, might also be justified.     

The preparation of a sensitive partial thromboplastin reagent from bovine brain.

Commercial partial thromboplastin reagents markedly differ in their sensitivity to factor deficiency, heparin, or the lupus anticoagulant. These differences may be partly due to the variable phospholipid content of different commercially available reagents. For over 15 years, we have routinely used a partial thromboplastin prepared from human brain. In the past four years, we have been using a similarly prepared bovine partial thromboplastin reagent. This report describes the preparation of our partial thromboplastin reagent, as well as an analysis of the phospholipid composition of both the human and bovine thromboplastin reagents. Four separate brain preparations produced consistent percentages of the anionic phospholipids, phosphatidylserine, and phosphatidylinositol. The bovine reagent was also compared with commercial partial thromboplastin reagents in the detection of coagulation factor deficiency, heparin, and the lupus anticoagulant.     

Relation of simple clotting tests to clotting factor levels in liver disease

The prothrombin time with Manchester and ox thromboplastins, the ‘P & P’ test of Owren and Aas, the partial thromboplastin time, the thrombin time and assays for factors II, V, VII, VIII:C, IX and X were performed by one observer in 18 patients with liver disease and 27 normal subjects; the prothrombin time and partial thromboplastin time were also carried out on another 28 similar patients by various observers. Routine liver function tests were measured in all patients. The prothrombin time with Manchester thromboplastin was well correlated with other clotting tests, and performing the other tests did not add to the information. Discriminant function analysis confirmed that clotting tests did not distinguish between different types of liver disease. Correlations between clotting tests and liver function tests reflected liver cell damage but were also influenced by acute phase reactions.     

No deterioration after 13 years in a stability study of a rabbit brain, plain, thromboplastin, RBT 1010, in rubber-stoppered ampoules

RBT 1010, a rabbit brain thromboplastin, plain, was prepared at the Thrombosis Reference Centre, Withington Hospital, Manchester, in 1989. The batch has been stored at -20 degrees C, in rubber-stoppered ampoules, for 13 years. The material was unchanged after this time. This was confirmed in a stability study, in which the reagent was used to test the prothrombin times of a panel of plasmas stored in liquid nitrogen, one from a normal volunteer and two from stable anticoagulated patients. RBT 1010 was also tested in calibrations, according to World Health Organization recommendations, against the reference thromboplastin preparations CRM 149S and RBT 90.     

Factor VII activity and antigen in a patient with abnormal factor VII

A patient with abnormal factor VII, which showed different activity when thromboplastin from different sources was used, is reported. The propositus, who was first seen at routine health examination 1 month after delivery, was a girl with no complaints of bleeding. The coagulation pattern was characterized by an abnormal clotting time using Normotest reagent which did not respond to the administration of vitamin K. Factor VII activity was decreased when measured using rabbit brain and lung thromboplastin, mildly decreased using human brain or human placenta thromboplastin and within the normal range using ox brain thromboplastin. The level of factor VII antigen was normal and revealed a normal mobility on immunoelectrophoresis. The molecular weight of this factor VII was not different from normal factor VII when analysed using autoradiography after sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE). It seems probable that the propositus had an abnormality similar to the factor VII Padua 1 described previously.     

Ox brain versus rabbit brain thromboplastin assays are the best tool for a preliminary diagnosis of the Arg304Gln factor VII defect (FVII Padua)

Factor VII (FVII) deficiency, the most frequent defect among the rare bleeding disorders, is commonly divided into type I and type II. In the former, there is a concomitant decrease in FVII activity and antigen. In the latter, there is a clear discrepancy between activity which is low and antigen which is normal or nearly normal. FVII Padua (Arg304Gln) is characterized by different reactivity towards different tissue thromboplastins. FVII levels were assayed by the use of different tissue thromboplastins, namely rabbit brain, human placenta, human recombinant and ox brain thromboplastin, in 6 homozygous patients. Cases reported in the literature were also evaluated. Ox brain thromboplastins yielded normal values, whereas human tissue or recombinant human thromboplastins yielded only slightly higher levels of activity than those obtained with rabbit brain reagents. The ox brain versus rabbit brain ratio was about 22, whereas the ratio for human placenta or human recombinant versus rabbit brain thromboplastin was only about 5. The FVII antigen versus rabbit brain, human tissue and ox brain activity ratios were 24.8, 4.3 and 1.1, respectively. These results indicate that the ox brain versus the rabbit brain thromboplastin ratio supplies a wider difference than the one between human tissue and rabbit brain. The antigen/ox brain activity ratio of 1.1 fully confirms this assertion.     

Comparison of the behavior of normal factor IX and the factor IX BM variant hilo in the prothrombin time test using tissue factors from bovine,human, and rabbit sources

A subset of hemophilia B patients have a prolonged bovine-brain prothrombin time. These CRM⁺ patients are classified as having hemophilia Bm. The prolongation of the prothrombin time has been reported only with bovine brain (referred to as ox brain in some literature) as the source of thromboplastin; prothrombin times determined with thromboplastin from rabbit brain or human brain are not reported to be prolonged. Factor IX from a hemophilia Bm patient (factor IX Hilo) was isolated. The activity of factor IX Hilo was compared to that of normal factor IX in prothrombin time assays when the thromboplastin source was of bovine, rabbit, or human origin. Factor IX, either normal or Hilo, prolonged a prothrombin time regardless of the tissue factor source. However, unless thromboplastin was from a bovine source, this prolongation required high concentrations of factor IX. Further, factor IX normal was as effective as factor IX Hilo in prolonging the prothrombin time when rabbit or human thromboplastin was used. With bovine thromboplastin, factor IX Hilo was significantly better than factor IX normal at prolonging the prothrombin time. The amount of prolongation was dependent on the amount of factor IX Hilo added. In addition, the prolongation was dependent on the concentration of factor × present in the sample. The prothrombin time changed as much as 20 seconds when the factor × concentration was varied from 50% to 150% to normal (fixed concentration of factor IX Hilo). These results demonstrate the difficulty of classifying the severity of a hemophilia Bm patient based on the bovine brain prothrombin time unless both the factor IX and factor × concentrations are known.     

An Inhibitor of Fibrin Formation in Thromboplastins Prepared by Saline Extraction of Human Brain

Human brain is a common source of thromboplastin for the prothrombin time, where the end point is the conversion of fibrinogen to fibrin. Experiments showed that human brain also contains a proteolipid which inhibits the conversion of fibrinogen to fibrin. The proteolipid is removed when brain tissue is washed with acetone, but remains as a contaminant when brain is extracted with saline. For this reason prothrombin times on the same plasma are longer when saline extracts, rather than acetone dried preparations, are the source of thromboplastin. The proteolipid explains why the prothrombin time becomes shorter when saline extracts are diluted to standardize their activity against the British comparative thromboplastin.     

Discrepant INR values: a comparison between Manchester and Thrombotest reagents using capillary and venous samples

The International Sensitivity Index (ISI) for different thromboplastin reagents is obtained by calibration against WHO reference preparations. It is hoped that calculation of the International Normalized Ratio (INR) from the ISI will permit accuracy and conformity in reporting laboratory assays of warfarin effect even across a range of different techniques. We have examined the INR values of 128 warfarin patients obtained by four different techniques in common use, namely venous and capillary Thrombotest and venous and capillary Manchester reagent. Discrepant INR values were obtained. The mean Manchester venous INR values were lower than those obtained by the other three methods (P less than 0.0001). This suggests that patients dosed by reference to Manchester venous INR are liable to receive more warfarin than those dosed by the other methods.     

Reliance on prothrombin time ratios causes significant errors in anticoagulation therapy.

BACKGROUND--The intensity of warfarin anticoagulation in the United States may be inappropriate if the international normalized ratio (INR) is not used, or if the international sensitivity index (ISI) of the thromboplastin is outside the range of 2.2 to 2.6. METHODS--Fifty-three hospital laboratories provided data on the sensitivity of their thromboplastin and whether they reported INR values. Additional data on thromboplastin sensitivity were obtained from 140 laboratories involved in the Stroke Prevention in Atrial Fibrillation study. The three major manufacturers of thromboplastin confirmed the range of thromboplastin sensitivity reported by the laboratories. RESULTS--Of 53 laboratories surveyed, 16 (30%) could not provide ISI data and only 11 (21%) reported INR results. Unlabeled thromboplastin was being used by 20% to 24% of laboratories, and only 8% to 20% were using thromboplastins with an ISI of 2.2 to 2.6. At the time the three manufacturers were contacted, they reported marketing thromboplastins with ISI values from 1.2 to 2.8, but none of the thromboplastins at that time had ISI values between 2.2 and 2.6. CONCLUSION--Warfarin therapy in the United States is managed inappropriately because most laboratories do not report INRs and the variability in thromboplastin sensitivity produces misleading prothrombin time ratio results. Additionally, recent research may require reexamination if INR or ISI data were not provided.     

Monitoring Successive Batches of British Comparative Thromboplastin

S ummary . Procedures used in the independent monitoring of British Comparative Thromboplastin are described. Recent results from batches of similar thromboplastin have been analysed for differences in their sensitivity to the anticoagulant defect. On the basis of observed variability, the minimum number of tests required to achieve a given level of discrimination between standardized and modified thromboplastins has been calculated to determine the number of readings which should be obtained to give a reasonable certainty of detecting unsatisfactory batches.     

Thrombomodulin activity is found in tissue thromboplastin preparations from placenta and from lung but not from brain

Substantial thrombomodulin activity could be detected in tissue thromboplastin preparations from placenta or from lung but not from brain. When the amount of these preparations was adjusted to contain 1 unit of tissue factor activity, up to 0.85 units of thrombomodulin activity could be measured, corresponding to the generation of 17 pmol/ml/min of activated protein C when 1.5 microM human protein C was activated by 20 nM human alpha-thrombin in the presence of 5 mM CaCl2. After treatment by phospholipase C, thrombomodulin activity was reduced in these samples. Addition of mixed brain procoagulant phospholipids partially restored thrombomodulin activity in the phospholipase C-treated samples. These results emphasize the role of phospholipids in the expression of optimal thrombomodulin activity in tissue thromboplastin preparations from placenta or from lung.     

An alternative extrinsic pathway of human blood coagulation

To study the interrelationships of the major human coagulation pathways, factor X activation in normal and various deficient human plasmas was evaluated when clotting was triggered by dilute rabbit or human thromboplastin. Various dilutions of thromboplastin were added to plasma samples containing 3H-labeled factor X, and the time course of factor X activation was determined. At a 1/250 dilution of rabbit brain thromboplastin the rate of factor X activation in factor VIII or factor IX deficient plasma was only 10% of the activation rate seen for normal or factor XI deficient plasma. Reconstitution of the deficient plasmas with factors VIII or IX, respectively, restored normal factor X activation. Similar results were obtained when various dilutions of human thromboplastin replaced the rabbit thromboplastin. From these experiments, it is inferred that normal activation of factor X in plasma due to dilute thromboplastin requires factors VII, IX and VIII. An alternative extrinsic pathway that involves factors VII, IX, and VIII may be a major physiologic extrinsic pathway, and this pathway may help to explain the clinical observations of bleeding diatheses in patients deficient in factors IX or VIII.     

The British Comparative Thromboplastin: the Relationship between Lipid Class Composition and Procoagulant Activity

S ummary . Twenty-eight consecutive batches of the reference reagent British Comparative Thromboplastin (BCT) were produced in the National (UK) Reference Laboratory for Anticoagulant Reagents and Control (NRLARC) between 1969 and 1977. The relationship between procoagulant activity and lipid class composition in these batches at various stages of age deterioration on storage has been studied by a modification of the method of high pressure chromatography which allows better definition of the individual lipids. The free fatty acid concentration rose markedly while concentrations of phosphatidyl choline, phosphatidyl ethanolamine and phosphatidyl serine were reduced. An oxidative process is suggested and supported by the increase in malonaldehyde levels. The method of production of the BCT involves maceration of human brain which causes the breakdown of cell membranes and the release of many potentially oxidative materials from subcellular particles. The loss of procoagulant activity of BCT on storage may be due to the loss of phospholipid or to the increase of inhibitory degradation products, i.e. free fatty acid and malonaldehyde. The examination of the lipid class composition of tissue thromboplastin extracts appears useful, therefore, not only in determining the phospholipids necessary for procoagulant activity, but also in monitoring the deterioration of tissue thromboplastins on storage.     

Dosage and control of oral anticoagulants: an international collaborative survey

An international survey of oral anticoagulant dosage has been carried out comparing the mean dosage prescribed in hospitals in 23 countries. In addition, participants using the Quick prothrombin time test were asked to assess the adequacy of dosage of a lyophilized test plasma which was mid-therapeutic using the British Comparative Thromboplastin (BCT).
The overall mean dosage proved similar for the groups of laboratories using the Quick test and human brain thromboplastin and Thrombotest although wide differences existed between individual centres. The survey indicated that these discrepancies were due partly to the adoption of different intensities of anticoagulation. In addition, local differences in patients'response to anticoagulants were apparent, e.g. North American centres prescribed a higher mean dose with a more intense therapeutic range than Europeans. Hong Kong physicians appear to prescribe a much lower dose than the rest of the world although the intensity of their treatment is comparable, whereas South African hospitals give moderate doses of warfarin despite a conservative therapeutic range. Such geographical variation in response would invalidate standardization of anticoagulant treatment based on the mean dosage approach.     

Factor VII Padua 2: another factor VII abnormality with defective ox brain thromboplastin activation and a complex hereditary pattern.

A new factor VII abnormality is presented. The propositus was a 9-yr-old child who presented a mild bleeding tendency characterized by epistaxis and easy bruising. The parents were not consanguineous, but they came from the same area. The laboratory features were mild prolongation of prothrombin time and P.P. test and normal partial thromboplastin and Stypven cephalin clotting times. The Thrombotest was moderately prolonged. Factor VII was 40%-50% of normal using rabbit or human brain thromboplastin, but only 13%-24% using ox brain thromboplastin. Factor VII cross-reacting material (CRM) was about 50% of normal. The father, a paternal aunt, and a paternal cousin showed similar clinical and laboratory findings. The brother of the propositus, the mother, and other members of her family showed about 50% factor VII activity and CRM and were considered to be heterozygotes for true factor VII deficiency. Similar findings were also present in the father and in the brother of the affected cousin. The defect in the propositus seems to consist of a double heterozygosity between abnormal factor VII and heterozygous factor VII true deficiency. The factor VII abnormality appears to consist of abnormal reactivity toward ox brain tissue thromboplastins and appears to be different from previously described factor VII abnormalities. The name factor VII Paudua2 is proposed for this condition.