SCIENTIFIC SUBCOMMITTEE SESSION
7 July 2007 Palexpo, Geneva, Switzerland
Control of Anticoagulation
Chair: S. Schulman (Canada)
Co-Chairs: W. Ageno (Italy), T. Baglin (UK), J. Harenberg (Germany), C Kearon (Canada), A. Lubetsky (Israel), J. Olson (USA), G. Palareti (Italy), A.M.H.P. van den Besselaar (The Netherlands)
Chairmen: S. Schulman (Canada) and A.M.H.P. van den Besselaar (The Netherlands)
S. Schulman opened the meeting of the committee and briefed on the activities over the past year. He also instructed ion the importance of publishing work from within the SSC as official SSC publications.
Jörgen Jespersen (Denmark) and Leon Poller (UK) presented first the design of this multicentre study.
Previous studies claiming benefit for computer-dosagehave depended solely on laboratory results and have not been sufficiently large to determine whether observed improvement in international normalised ratio (INR) control resulted in clinical benefit or whether computer-dosage was as safe clinically as that by experienced medical staff.
The aim of this study was to compare dose control with computer-assistance versus dosage by experienced medical staff at establishments with considerable experience of administration of oral anticoagulants. Forty centres were invited to provide data from 400 patient-years thus to achieve a 16,000 patient-years’ target. Patients at participant centres were to be randomised to manual (medical staff) dosage or to one of two commercial computer programs, either a new version of the PARMA program [PARMA 5], or to the established DAWN AC program. The use of two different commercial computer programs aimed to preserve the study’s independence from industry.
New definitions of major and minor bleeding have been introduced in a standardised Clinical Events Assessment Form designed by the Project Adjudication Committee. Clinical Events forms were returned as the events occurred to the EAA Central Facility by participant centres and were assessed “blind” by the Project Adjudication Committee. The study was monitored by the Project Steering Group and by the Project Safety Committee.
A continuous programme of international sensitivity index (ISI) calibration and external quality control of INR testing was conducted at participant centres.
The first 6-12 months were devoted to recruitment and training of staff from centres, to familiarisation with the computer programme and to enlisting patients.
Gordon Lowe (UK), also on behalf of M. Moia (Italy), A. Turpie (Canada) described the process of the EAA Adjudication Committee, then the definitions and total numbers of clinical events, followed by a brief comparison with the literature.
Jörgen Jespersen (Denmark) and Leon Poller (UK) continued by reporting clinical results from the randomized study by the European Action on Anticoagulation (EAA). 32 centres with a special interest in oral anticoagulation participated and incorporated 13,219 patients, which provided 18,617 patient-years. One computer program was used much more than the other program in this study. INR tests numbered 193,890 with manual dosage and 193,424 with computer-assistance giving “time-in-range” (Rosendaal) of 64.7% and 65.9% respectively. The overall number of confirmed adjudicated clinical events although 7.6% lower in all clinical groups with computer-dosage was not significantly different but the number of clinical events in the 3208 patients with deep vein thrombosis was significantly lower (115) with computer-assisted dosage compared to 152 in the medically dosed arm of the study (p<0.01).
The study confirms the clinical safety and effectiveness of computer-assisted dosage using two different programs (PARMA 5 and DAWN AC) compared to the standards of the 32 participant centres with a special interest in oral anticoagulation.
R. Maclean on behalf also of I. Jennings, S. Kitchen and I.Walker (all from UK) reported on a survey on computer-assisted dosing resulted in particular cases. They examined 3 hypothetical scenarios, provided clinical details, past INRs and warfarin dosages. 955 UK NEQAS participants completed the exercise. They had to recommend a continued dose and the time to next test. The participants were subgrouped as 250 used CDSS, 129 dosed manually, Blood coagulation programme 71 vs 210. Dawn most common followed by RAID. Huge variability in recommended recall times was observed. For patients with a recent historic change in INR there was also a huge variability in recommended new dose. Clinicians tend to override computer doses in 30%.
Michelle Keown (UK) presented data from an international collaborative study at 77 centres which compared local INR correction using the two alternative methods recommended in the Scientific and Standardisation Committee of the International Society on Thrombosis and Haemostasis guidelines: Local ISI calibration and “Direct INR”.
Success of INR correction by local ISI calibration and with Direct INR was assessed with a set of 27 certified plasmas (20 from warfarin patients and 7 from normals).
At 49 centres using human thromboplastins, 3.0% initial mean local INR deviation from certified INR was reduced by local ISI calibration to 0.7% and at 25 centres using rabbit reagents, from 15.9% to 7.5%. In contrast with “Direct INR” mean deviation using human thromboplastins increased from 3.0% to 6.6% but there was some reduction with rabbit reagents from 15.9% to 10%.
Local ISI calibration gave INR correction for the majority of PT systems but failed at the small number using combined rabbit reagents suggesting a need for a combined reference thromboplastin. Direct INR gave correction overall with rabbit but not with human reagents. INR correction was better than local ISI calibration with combined rabbit reagents.
A.M.H.P. van den Besselaar presented this project also on behalf of V. Chantarangkul, A. Tripodi.
ISI calibration is performed with fresh samples from many individuals, both normals and patients treated with oral anticoagulants. According to 1999 WHO guidelines, samples outside the 1.5-4.5 INR range should be excluded followed by exclusion of data points with greater distance than 3 residual standard deviations about the regression line. Dr van den Besselaar et al analyzed three multicentre ISI calibration studies of international thromboplastin standards (performed in 1990, 1995, and 2005), to determine the effects of INR and outlier exclusion. Obviously, the within-laboratory variation was reduced by outlier exclusion, but there was also a tendency to reduced between-laboratory variation. The effect on the mean ISI was only minor: in most cases <1% and in all cases <3%. It is recommended to follow the 1999 WHO guidelines for outlier exclusion.
Winfried Plesch, Roche Diagnostics, Germany, presented data on the quality control of the CoaguChek XS strip. The CoaguChek® XS system is based on the amperometric measurement of the thrombin activity initiated by starting the coagulation cascade using a human recombinant thromboplastin. The concept of the CoaguChek XS test strip integrated QC function is to check the performance of the system at the spot of the coagulation testing. This could be done by adding the blue indicator Resazurin and a mixture of redox partners for this molecule to the PT test formulation. This means that the integrated control testing is performed exactly within the physical and chemical environment of the thromboplastin, which activates the clotting cascade in the blood sample. Under conditions of mishandling of the test strip, such as exposure to high humidity or longer exposure to daylight, the Resazurin is transformed to the purple by-product Resorufin. After blood application both molecules are quantified electrochemically using the same electrodes and the same meter functions as used for the clot detection. The reliability of detecting mishandling of test strips was validated under conditions reflecting the different climate zones in the world. Even under extreme conditions the integrated QC performed reliably.
In conclusion, incorrect handling of test strips, e.g. exposure to high humidity, will lead to deterioration of the thromboplastin in the reagent mixture, which in turn would result in unreliable INR values. Such mishandling is reliably detected by the integrated QC of the CoaguChek XS system, which ensures that only correct INR results are displayed to the user. Therefore, liquid quality control samples are no longer needed to check the CoaguChek XS system performance.
Steve Kitchen also on behalf of D. Kitchen, I. Jennings, T. Woods, I. Walker (all from UK) reported on this quality assessment. UK NEQAS Blood Coagulation has successfully established an External Quality Assessment (EQA) programme for INR testing using 2 new Point of Care devices - the CoaguChek XS (CUC XS) and the CoaguChek XS Plus (CUC XS Plus) during 2006. Their end point detection is based on the amperometric measurement of the thrombin activity initiated using a human recombinant thromboplastin, and a new formulation of EQA material was required. 2 lyophilised plasmas and diluents for reconstitution and recalcification were developed and tested by centres routinely using these devices. Results from pilot exercises (n=23 participants) showed good precision: sample 1 CUC XS median INR = 3.0, CV 8%, CUC XS Plus median INR =3.2, CV 6%; sample 2 CUC XS median INR = 3.45, CV 4%, CUC XS Plus median INR = 3.4, CV 6%, but included a number of test error messages from devices. Sample modifications were made and after further pilots a full programme has been launched in which the reproducibility of results has been better than that seen with other NPT devices in UK NEQAS programmes, and with very few testing errors. For the CUC S programme analysis of EQA data from lyophilised plasma samples has on rare occasions identified test strip lot numbers giving discrepant INRs (compared to other lot numbers) , but these same lot numbers have had smaller or absent differences when native whole blood is tested. This means that findings from plasma testing can only safely be considered genuine if there are supporting data based on whole blood analysis since the devices are calibrated only for whole blood testing.
PART 2.
Chairs: S. Schulman ( Canada) and J. Harenberg ( Germany)
Valerie Eschwège on behalf of also L Bellest, R Poupon, O Chazouillères and A Robert from Unité d’hémostase et service d’hépatologie, Hôpital Saint-Antoine, Paris, France discussed the use of INR in liver disease. International Normalized Ratio (INR)/International Sensitivity Index (ISI) system developed to standardize PT reporting during oral anticoagulation (OA) has been extended in liver disease (LD) and included in prognostic models such as the Model for End Stage Liver Disease (MELD) prioritizing liver transplant. Dr. Eschwège and colleagues have previously reported that, in LD, INR fails to yield a PT reporting independent of the thromboplastin used. A new standardization of PT in liver disease is evaluated using the thromboplastin calibration model proposed by the World Health Organization (WHO) using plasmas from LD instead of OA patients and leading to a new INR LD /ISI LD system specific for LD.
1) The ISI LD of 5 thromboplastins (#1, 2, 3:rabbit, 4:human, 5:human recombinant) were determined by calibration, following WHO guidelines, against the reference preparation rTF/95 using 60 plasmas of patients covering the whole range of LD severity. 2) In 34 other patients, the differences between mean PT reported as seconds, ratio, INR and INR LD across the 5 thromboplastins were analysed by ANOVA and multiple comparison test.
The ISI LD/ISI were 0.98/1.67, 0.94/1.54, 0.70/1.05, 0.84/1.03 0.85/0.83 for thromboplastins 1 to 5 respectively, demonstrating the difference in reagents sensitivities to defects induced by LD and by OA. For the 2 most different reagents (1 and 5) and for the 5 patients with INR>3, the discrepancy between INR was >60% of the mean INR and consequently the difference between MELD scores ( scale from 0 to 40) calculated with these INR was ³ 7. Mean PT differed significantly for most pairs of reagents in all reporting modes except INR LD that eliminated all results variability.
Adoption of INR LD instead of INR as an international scale of PT reporting appears as an important goal in hepatology.
Armando Tripodi discussed the use of prothrombin time (PT) for staging severity of liver disease. The model for end-stage-liver-disease (MELD) is a mathematical score used to prioritize patients for liver transplantation and includes results for creatinine, bilirubin and PT expressed as international normalized ratio (INR). The rational of using the MELD rests on the assumption that the score would be the same across the country if the methods used to measure the variables yield the same numerical results regardless of the testing laboratory. Evidence was provided that specific methodologies may influence the MELD and the PT-INR was identified as the most important. This study was designed to provide information on the between-thromboplastin variability and to explore alternatives to obviate such variability. Fifty-seven cirrhotics were selected and their PTs were measured with 7 thromboplastins. The thromboplastins were previously calibrated by testing plasmas from patients on vitamin-K-antagonists (VKA) and healthy subjects to assign the international sensitivity index (ISI vka) needed to convert PT into INR. Each of the thromboplastins was also assigned an ISI liver by substituting in the calibration the plasmas from VKA-patients with plasmas from cirrhotics. INR and MELD values for individual patients were calculated by using the ISI vka or the ISI liver. The mean INR vka obtained with the 7 thromboplastins were significantly different (p<0.001). Conversely, the mean INR liver were not. Similarly, the mean MELD vka were significantly different (p<0.001), but those differences were abrogated for the MELD liver. In conclusion, the alternative thromboplastin calibration using plasmas from cirrhotics instead of from VKA-patients is feasible and may resolve the variability of the MELD to prioritize patients for transplantation.
Jawed Fareed (USA) discussed the issue that several commercially available low molecular weight heparins (LMWHs) are now widely used in the management of thrombotic and cardiovascular disorders. Although derived from porcine mucosal heparin, these drugs are manufactured by distinct chemical and enzymatic methods with products, which can be differentiated in chemical and biologic assays. The preclinical pharmacological profile of these products is also distinct and profoundly impact on their therapeutic actions. More recently several generic versions of enoxaparin and dalteparin have been introduced in some Asian and South American countries. In addition, numerous generic suppliers have applied for the approval to sell the generic versions of enoxaparin and dalteparin in the US and European union. Neither the US FDA nor EMEA have any guidelines for the generic interchange of branded products at this time. The current pharmacopeial guidelines are inadequate to accept the generic version of the branded LMWHs since these apply the older guidelines. The LMWHs represent a hybrid of the biologic and chemical manufacturing processes. The European Pharmacopeial description of each of the individual LMWHs is incomplete and US Pharmacopeia is working towards developing monographs for such drugs as enoxaparin and dalteparin. Considering the complexities related to chemical and biologic profiles of LMWHs, additional guidelines for the therapeutic and generic interchangeability are warranted. Product characterization and structural equivalence are not adequate to validate the generic versions of branded LMWHs. Data obtained on the currently available generic versions of LMWHs show that while these products exhibit similar molecular and pharmacopeial profile, marked differences in their in vivo pharmacology are noted. Thus, animal studies and qualified clinical trials may be needed as acceptance criterion for the generic LMWHs. Moreover, process controls including starting material and pharmacodynamic characterization in valid models may be needed. The ISTH/SSC has played a role in the biologic standardization and characterization of LMWHs, it is proposed that ISTH/SSC on the Control on Anticoagulation, consider the development of specific guidelines for the requirements to accept a generic version of the branded product.
Gualtiero Palaret and colleagues studied a pregnant woman with antithrombin and protein S deficiencies who had had two previous DVT episodes. The latter episode occurred at the beginning of a previous pregnancy that ended in intra-uterine death, even though the patient was receiving antithrombin concentrate together with therapeutic LMWH doses. Reporting to their clinic years later on occasion of a new pregnancy, the warfarin therapy was stopped and treatment started with antithrombin concentrate (twice a week) and therapeutic LMWH doses b.i.d. When anti-Xa measurement was performed with different methods the levels varied according to the presence or absence of antithrombin supplementation in the assay reagents. Anti-Xa levels were within the therapeutic range using reagents that included antithrombin, but were clearly below the therapeutic levels after exclusion of antithrombin from the reagent, except when the antithrombin level in the patient was normal thanks to supplementation. This prompted an increase of antithrombin concentrate supplementation and LMWH doses in order to maintain adequate anti-Xa levels (excluding antithrombin from the reagent) throughout the pregnancy. The pregnancy ended well, with no complications for either mother or foetus. Dr Palareti is following two similar cases in collaboration with colleagues from another Institution.
He would like to warn against the use of anti-Xa assays containing antithrombin in the reagents to regulate a LMWH anticoagulation in subjects with antithrombin deficiency/reduction.
Sam Schulman reported that the recruitment to this registry has been very slow in spite of announcements at several conferences, mass distribution e-mail via the ISTH roster and promotion by the sponsor of the unrestricted grant, Leo. Efforts will continue to encourage reporting.
Francesco Dentali and Walter Ageno from Italy presented a suggestion for a new registry. Splanchnic vein thrombosis is an uncommon, but potentially life-threatening disease. Symptoms are non-specific, and clinical presentations are variable. Advances in imaging techniques have facilitated its early diagnosis, but information on how these techniques are applied in clinical practice is scant. Several etiologic factors have been reported, but their true prevalence is uncertain.
The treatment of splanchnic vein thrombosis may involve anticoagulation alone or in combination with surgery. The timing and intensity of antithrombotic treatment in this setting are unknown, as well as the natural history of the disease. Because information on risk factors, clinical features, diagnosis, treatment and outcome of splanchnic vein thrombosis is mainly based on small, uncontrolled retrospective series of patients, the aim of the proposed prospective registry is to improve the knowledge on this important disorder. Information on most common clinical presentations, most used diagnostic approaches, most common risk factors, and on the natural history of splanchnic vein thrombosis in a large prospective cohort of patients will be collected. Use of anticoagulation, duration of anticoagulation, thromboembolic and major bleeding complications during anticoagulant therapy and when anticoagulant therapy is stopped will also be registered.