SCIENTIFIC SUBCOMMITTEE SESSION
6 July 2007 Palexpo, Geneva, Switzerland
Fibrinolysis
Chair: C. Longstaff (UK)
Co-Chairs: C. Dempfle (Germany), D. Hendriks (Belgium), O. Matsuo (Japan), M. Nesheim (Canada)
TAFI/CPU Chaired by D Hendriks
M Nesheim presented work on a new assay for TAFIa in plasma based on modifed plasminogen with a fluorescently labelled active site and a chemically-derivatized FDP containing a quencher molecule. Removal of C-terminal lysines by TAFIa is measured by increasing fluorescence. The assay was sensitive down to 10-15 pM TAFIa with good results for intra- and inter assay variabliltiy of 6.3 and 8.3%. There was no interference from normal plasma plasminogen in the normal physioligical range or the TAFI polymorphism at position 325. Several normal plasma samples from volunteers were tested and an average of 20 pM TAFIa determined (range 4-32). Data were presented from animal experiments following injection of procoagulant FXa and PCPS demonstrating a dramatic, transient increase in TAFIa levels with the expected half-life. Sub-lethal doses of E.coli injected into a baboon as a model of sepsis were able to activate up to 30% of available TAFI zymogen. Conditions suitable for collection of plasma samples to optimise TAFIa stability were discussed.
J Willemse discussed approaches for the measurement of proCPU and CPU using specific small chemical substrates with the structure Bz-Xaa-Arg (where Xaa represents a naturally occurring amino acid). These assays are complicated by the presence of constitutively active CPN, the activity of which must be subtracted to determine CPU activity. A total of 15 synthetic substrates were screened with CPU and CPN to determine which amino acids optimised the ratio of activity (expressed as kcat/Km) for CPU/CPN. Aromatic residues or chemically modified aromatic residues were optimum for maximising the ratio of CPU/CPN activity. A summary of results from a pilot study measuring CPU levels generated during thrombolytic therapy for ischaemic stroke was presented demonstrating significant generation of CPU activity. Conditions for the minimisation of ex-vivo proCPU activation were discussed.
A Gils discussed the application of 3 different ELISA approaches to investigate the extent of TAFI activation in clinical studies. It was hypothesized that not the total amount of TAFI protein but the amount of activated TAFI may play a critical role in the interference with fibrinolysis. Therefore, two ELISAs were developed measuring either the activation peptide or activated TAFI (TAFIa). Intact TAFI and TAFI fragments were determined in three different groups of patients i.e. patients with hyperlipidemia, patients with stroke and patients with sepsis.
From the data obtained it was concluded that the ELISAs that measure the extent of TAFI activation are more sensitive markers in studies on the relationships between TAFI and cardiovascular diseases, but assessment of all possible markers may be needed and should be assessed on a case by case basis. Standardisation of the assay was discussed and it was confirmed that recombinant active peptide did not react in the same way as native peptide found in plasma and could not be used as a standard, however it may be possible to express results in pM of peptide with further work.
T Lisman discussed 3 epidemiological studies on venous and arterial thrombosis in which thrombosis risk associated with hypofibrinolysis was investigated. Hypofibrinolysis was assessed with a plasma-based global fibrinolysis assay where plasma is clotted with tissue factor in the presence of calcium and phospholipid vesicles and fibrinolytic potential estimated from clot lysis curves. Three studies were discussed. LETS ( leiden thrombophilia study), a case/control study on venous thrombosis (421 pts/469 controls) where hypofibrinolysis was associated with a 2-fold increased risk for a first venous thrombosis. MEGA (multiple genetic and environmental assessment of venous thrombosis), a case/control study on venous thrombosis (2913 pts/2129 controls), where increased thrombosis risk was associated with hypofibrinolysis, and showed an interaction of hypofibrinolysis with factor V Leiden. SMILE (study of myocardial infarction leiden), a case/control study on myocardial infarction in men (426 pts/646 controls), where hypofibrinolysis was associated with an increased risk of MI only in men below the age of 50. The SMILE study also showed elevated TAFI levels (measured by activity assay, Pentapharm) to be protective against MI.
During the general discussion D Hendriks proposed that a number of common samples should be compared using the different approaches described for measuring active CPU/TAFIa.
D-Dimer Chaired by C-E Demple
I Jennings reported on two separate UK NEQAS for Blood Coagulation exercises to explore the degree of precision amongst laboratory D-Dimer measurements, and the degree by which inter-method agreement could be improved using a calibration curve model. The first exercise demonstrated generally good within-centre precision, with 82% centres reporting results for two identical but differently coded samples within 10% of each other. However, 6 centres reported results which would have excluded DVT for one sample but failed to exclude DVT for the other, identical sample. In the second exercise, overall between-method precision of D-Dimer results for two samples was shown to improve markedly when a calibration model was applied, using the consensus median values obtained by all participants for three “calibration plasmas” to recalculate D-Dimer values. For centres reporting results in fibrinogen equivalent units (FEUs), between-centre coefficients of variation (CVs) fell from 25.9% to 11.6% and 22.4% to 7.7% respectively for the two samples. For centres reporting in ng/ml, CVs fell from 45.3-21.6% and 40.8-11.6% respectively. Improved harmonisation of D-Dimer results by different methods may be achieved by a calibration model and common calibrant plasmas.
P Meijer reported results from the latest ECAT study involving 600 participants and covering 25 methods, although 8 methods account for 90% results. More participants report results in FEU rather than ng, in contrast to UK NEQAS studies. Issues of repeatability and the importance of repeat testing for results around the cut-off level were discussed as an approach to minimise false positives and negatives. Sources of variability in assay results were discussed which include the usual reasons of differences between antibodies and reference standards used in different methods but also additional problems from different lots of reagents in the same kits.
Both NEQAS and ECAT quality control studies highlight the large discrepancies of numerical D-dimer values reported, which add to the complications surrounding interpretation of assay results in the clinical context of VTE exclusion. The main cause of discrepancy is the different calibrations used by the manufacturers of the assays, but problems for clinical decisions arise from choice of cutoff values. These may be taken from package inserts but may be in-house evaluations. Even for individual assays, the cutoff values used may vary considerably between laboratories and VTE exclusion cutoff values have been validated in appropriate clinical trials only for a minority of D-dimer assays. Assay results in the high concentration range are also highly discrepant, which makes it difficult to use D-dimer assays in scoring systems, such as the ISTH-DIC score, or establishing cutoff values for other indications apart from VTE exclusions.
P Meijer also discussed harmonisation of immuno-assay in general with particular relevance to possible approaches to measurement of plasma tPA and PAI-1 antigen. Ideally a standardisation hierarchy should be adopted with a SI units and a primary reference methods at the head, which also requires clear definition of the entity to be measured. This is often not possible for complex biological mixtures (“soups”) and approaches to harmonisation may propose the use of consensus values, but it should be recognised that this approach is “unstable”. A number of proposals were made to improve standardisation and harmonisation of immuno-assays including the organisation of collaborative studies with samples having a range of values to be measured, inclusion of all available methods and assessment of clinical samples where possible and clear understanding of the entity being measured.
C-E Demplfe discussed applications of D-dimer assays for purposes other than VTE exclusion including DIC, aortic aneurysm exclusion, monitoring of anticoagulant therapy (including detection of heparin-induced thrombocytopenia), and monitoring of intensive care patients and risk stratification after stopping anticoagulant therapy. Reporting on a common scale is particularly important in these cases as is the ability to measure a wide range of D-dimer values.
A summary of a consensus statement issued at a meeting on D-dimer hosted by the ECAT was presented which included the following points:
It was suggested, that pooled plasma from patients with high D-dimer antigen concentration be used for harmonization, applying a series of dilutions of the pooled plasma as used in NEQAS and ECAT quality control
C-E Dempfle showed results from the Fibrin Assay Comparison Trials (FACT) parts 4 and 5. FACT4 compared assay reactivity of fibrin fragment D-dimer and pooled plasma from patients with DIC, using serial dilutions in plasma from healthy blood donors, and in buffer. Some D-dimer assays displayed identical dose-response with fibrin fragment D-dimer and the pooled plasma samples, some assay reacted considerably better with the low molecular weight fibrin degradation product than with the predominantly higher molecular weight fibrin contained in the plasma, and other assays responded poorly to fibrin fragment D-dimer. For the pooled plasma, assay reactivity was similar for dilutions with plasma and buffer. It was concluded that a single pooled plasma with high concentration of D-dimer antigen would be sufficient as common calibrator, with no need to prepare sets of dilutions, or pooled plasma with different levels of D-dimer antigen.
In the recently completed FACT5 trial, reference laboratories of assay manufacturers received a pooled plasma with high concentration of D-dimer antigen for preparation of serial dilutions with assay-specific diluents, and a set of 50 pooled plasma samples with different levels of D-dimer antigen. The correlation of assay results for all 30 assays included was excellent, with a mean regression coefficient of 0.946±0.054 (range 0.703 – 0.999). Common calibration with the pooled plasma reduced the coefficients of variation from nearly 60% to approximately 20%.
Since the procedure for harmonization based on consensus values produced variable results depending on the set of assays included, a new procedure was suggested by C-E Dempfle to generate D-dimer antigen values on a common scale for all D-dimer assays. This procedure is based on the distribution of a pooled reference plasma with high concentration of D-dimer and an assigned D-dimer level to the assay manufacturers. The D-dimer concentration is assigned by a procedure involving ‘homogenization’ of the D-dimer antigen by extensive plasmin digestion of the reference plasma and quantitation of the amount of D-dimer generated with a calibrator consisting of terminal plasmin digest of a cross linked fibrin clot prepared from a known amount of fibrinogen. By the plasmin digestion, all D-dimer antigen present in higher molecular weight form is transformed to fibrin fragment D-dimer, which is a homogeneous analyte. This allows the preparation of successive plasma pools with constant levels of D-dimer antigen.
Future activities will include quality control issues of the procedure used for assigning D-dimer concentration values, and investigations on the effect of lyophilization of the plasma on the results as part of the process of making long term stable standards. The effect of common calibration on future quality control exercises, and the performance of diagnostic algorithms involving D-dimer antigen will be other topics.
Plasminogen Activators and Plasmin chaired by O Matsuo
C Longstaff reported the conclusions of a collaborative study to determine tPA antigen in 4 samples: (1) SSC/ISTH secondary coagulation standard lot 2; (2) SSC/ISTH secondary coagulation standard lot 3; (3) NIBSC Preparation 94/730; (4) NIBSC Preparation 86/670. In total 14 sets of results comprising 48 independent assays were analysed using 8 different methods: 6 commercial kits and 2 in-house methods. Results for the 2 SSC/ISTH plasma samples were similar and within the expected range at 2.9 and 3.0 ng/ml for lot 2 and 3, respectively. The overall mean antigen value for 94/730 was close to 25 ng/ml, the expected value based on the formulation of this preparation and on past studies. Data were also analyzed using local standards and a common standard for all assays: 94/730 with an assigned value of 25 ng/ml. In this analysis the mean antigen values for the SSC plasmas were not changed from the analysis using local standards but there was a modest reduction of up to 7 % in inter-laboratory gcv. Analysing data according to method, grouping different methods or kits, highlighted significant differences between methods. However, it was possible to correct for these differences and harmonise results for normal plasma pools using data from the SSC plasma samples which produced significant reductions in % gcv, and left the antigen values unchanged. Sample 94/730 (recombinant tPA in plasma) would make a satisfactory reference preparation for tPA antigen determinations in plasma with a consensus value of 25 ng/ml. It was recommended that 94/730 be proposed as the WHO 1 st International Standard for tPA antigen in plasma. SSC coagulation plasma lot 3 can be assigned a consensus value of 3.0 ng/ml tPA antigen. The process of approval developed by SSC/ISTH was followed to prepare a report for the WHO Expert Committee on Biological Standards to recommend establishment of 94/730 as an International Standard for tPA Antigen. Approval statistics and comments from the collaborative study participants (10/13), a panel of experts with a background in fibrinolysis standardisation issues (9/13) and Fibrinolysis Subcommittee co-chairs were summarised. Among the responders no one disagreed with the proposals that 94/730 should be established as the WHO 1 st IS for tPA antigen in plasma with a value of 25 ng/ampoule or the ISTH/SSC coagulation plasma lot 3 should be calibrated at 3.0 ng/vial. The only comments received suggested possible improvements to the calibration process which would include the use of expert laboratories rather than collaborative studies and consensus values Another respondent pointed out a possible source of variability in the data obtained not identified in the report which could be due to matrix effects and “cryptic” tPA. No further comments or objections were received from the meeting.
C Longstaff reported results from a collaborative study to measure PAI-1 antigen in plasma recently completed with an aim to investigate the possibility of harmonising results obtained for PAI-1 measurements using different methods. Participants were provided with 5 different samples comprising 3 freshly collected frozen small plasma pools containing Low (L), Medium (M) and High (H) levels of PAI-1 and 2 lyophilised plasma preparations from pools of donors, which were the SSC/ISTH secondary coagulation standard lot 3 and a sample prepared at NIBSC coded 06/053. Twelve sets of data were returned comprising 7 methods, designated A-G. As expected, results for the 5 samples were highly variable between methods. The 5 samples were assigned a consensus value for PAI-1 antigen content as the arithmetic mean value from the 12 sets of data. A regression equation was then calculated from plots of each laboratory’s results for the 5 samples versus the consensus value. Conversion factors for slope and intercept were calculated for each method to make it possible to convert results from each method into values on the consensus scale. The harmonisation process worked well for most methods except one, which determined a different ranking for the PAI-1 content of the 5 samples from all other methods. The harmonisation procedure allows expression of results on a common scale so results can be compared. However, the consensus approach does not allow determination of PAI-1 antigen content in absolute units of real ng/ml and further work is required to achieve this. The results obtained raised questions about the normal range of circulating PAI-1 antigen in plasma.
C Thelwell reported on the possible usefulness of a standard of 4-nitrophenol for the standardising active site titrations using NPGB. Traditionally IS have been calibrated IU following an international collaborative study using laboratory’s own in-house methods. There has been a recent movement towards introducing SI units for standard preparations. Such a standard might be useful in conjunction with active-site titrants available for a range of proteases including plasmin, thrombin, urokinase and factor Xa, and associated inhibitors. NPGB, a suitable active-site titrant for trypsin, and an example of a collaborative study to establish the 1st IS for Alpha-1 antitrypsin (proteinase inhibitor) in units of moles of active inhibitor was brieflydescribed. Trypsin cleaves NPGB to release 4-nitrophenol, which can be measured by absorbance and converted into moles of active enzyme based on a 4-nitrophenol calibration curve. This approach relies on the accuracy of generating the calibration curve. This accuracy could be improved if a standard for 4-nitrophenol was available to eliminate variation introduced in the preparation of stock solutions. Fluorimetric active-site titrants also exist, such as MUGB, which offer greater sensitivity. A standard for 4-methylumbelliferone could be used to calibrate titrations with MUGB. This approach would allow new and replacement protease (and associated inhibitor) standards to be calibrated in molar concentrations as well as assigning SI units to existing IS. This was recommended as a possible approach for standardising plasmin as it will soon be necessary to work on replacing the existing 3 rd IS.
P Vandeberg presented work on the use of active site titration for the value assignment of a reference standard for plasmin, used during development of therapeutic plasmin which is being investigated as a direct acting thrombolytic. This standardization was done for the purpose of maintaining consistent dosing. The method used is essentially the same as what was published by Chase and Shaw (Biochem Biophys Res Commun. 1967 Nov 30: 29(4): 508-14.), but has been adapted for use in 96 well plates for manual or automated execution. Work involving crossover testing with the International Reference Preparation for Plasmin (97/536) using an amidolytic assay (chromogenic substrate S-2403) was also covered, although concerns were raised over the long term stability of the WHO IS for Plasmin which mitigated against over-reliance on this preparation.
Fibrinlysis Subcommittee Announcements by C Longstaff
C Longstaff closed the meeting with several short announcements related to Subcommittee activities. These included a request from the European Pharmacopoeia for a standard for plasmin inhibitor (alpha-2-antiplasmin) which is needed to measure the remaining inhibitor activity in virus inactivated human plasma used therapeutically. The feasibility of making such a standard will be explored and a request was made for groups interested in taking part in a collaborative study should contact C Longstaff.
An update on the modification of the Instructions for Use accompanying the 3 rd IS for Streptokinase 00/464 was provided. It is proposed to recommend that the IS only be used for native streptokinase or for recombinant streptokinase that has been checked for suitable activity in fibrin and non-fibrin-based assays. The presence/absence of fibrin can lead to discrepant results using the 3 rd IS for Streptokinase with some recombinant products and this is potentially dangerous. Where discrepant results are obtained there is currently no way of assigning a potency with the 3 rd IS for Streptokinase.
An update was provided on the status of the WHO 1 st IS for Streptodornase. This IS will now likely need to be replaced following earlier consideration that it might be discontinued.
There was one Fibrinolysis SSC publication in the past year briefly reporting the outcome of a study on fibrinolysis methods for potency determinations of streptokinase, tPA and urokinase, C Longstaff et al, J Thromb Haemost 5(2) 412-4: 2007.