John Sioufi

Identification of factor inhibitors by diagnostic haemostasis laboratories. A large multi-centre evaluation

We have assessed the proficiency of diagnostic haemostasis facilities to correctly identify coagulation factor abnormalities and inhibitors. Forty-two laboratories participating in the external Quality Assurance Program (QAP) conducted by the RCPA agreed to participate and were each sent a set of eight samples (each 3 x 1 ml) for evaluation. They were asked to blind test these samples for the presence or absence of inhibitors, and where identified, to perform further analysis (including specific inhibitor analysis). In order to make the exercise more challenging, in addition to true factor inhibitors, samples were provided that reflected potential pre-analytical variables that might arise and complicate inhibitor detection or lead to false inhibitor identification. In brief, the sample set comprised a true high level factor (F) V inhibitor, a true moderate level FVIII inhibitor (but sample was defibrinogenated), a true lupus anticoagulant (LA), a normal (but slightly aged) plasma sample, a normal serum sample, a normal EDTA sample, an oral anticoagulant/vitamin K deficiency sample, and a gross heparin ( approximately 10 U/ml) contaminated sample. Sixty-three percent of participants correctly identified the true FV inhibitor as such, although the reported range varied greatly [10 to >250 Bethesda units (BU/ml)] and 46% correctly identified the true FVIII inhibitor, despite the complication of the sample presentation, although the reported range also varied (7 to 64 BU/ml). Some laboratories either failed to identify the inhibitor present, or misidentified the inhibitor type. The LA, the oral anticoagulant/vitamin K deficiency, the normal serum sample, and the normal (aged) sample were also correctly identified by most laboratories, as was the absence of specific factor inhibitors in these samples. However, a small subset of laboratories incorrectly identified the presence of specific factor inhibitors in some of these samples. The heparin sample was also correctly identified by most (68%) laboratories. In contrast, the normal EDTA sample was misidentified as a FV and/or FVIII inhibitor by most (68%) laboratories, and only one laboratory correctly identified this as an EDTA sample. Thus, we conclude that although laboratories are able, in most cases, to identify the presence of true factor inhibitors, there is a large variation in identified inhibitor levels and there are also some significant errors in identification (i.e. false negatives and misidentifications). In addition, there is a significant false positive error rate where some laboratories will identify the presence of specific factor inhibitors where no such inhibitor exists (i.e. false positives).

Laboratory diagnosis of von Willebrand's disorder: quality and diagnostic improvements driven by peer review in a multilaboratory test process.

Summary.  Regular multilaboratory surveys of laboratories derived primarily from Australia, New Zealand and Southeast Asia have been conducted over the past 7 years to evaluate testing proficiency in the diagnosis of von Willebrands disorder (VWD) and to assess changes to test practice. Participating laboratories (currently 45) are asked to perform their usual panel of tests for VWD, and then to self‐interpret test results as to the likelihood (or not) of VWD, as well as to the potential subtype identified. Samples provided in the past two survey distributions (both conducted in 2003) were as follows. Survey part A/distribution 1: Normal donor plasma, plasma with borderline normal/reduced levels of VWF (×2) and plasma from an individual with type 2 A VWD. Survey part B/distribution 2 (family VWD study): Plasma from a father, mother and son with borderline normal/reduced von Willebrand factor (VWF), and a daughter with type 3 VWD. In line with previously published survey results, the interassay and within method coefficients of variation (CV) were similar for all assays (around 15–25%), although tending to be slightly higher for VWF:RCo and VWF:CB than VWF:Ag and FVIII:C. Most laboratories reported test values consistent with expected findings, and made correct interpretations or predictions regarding the nature of the samples, although discrepant assay results or interpretations are still seen in approximately 5–10% of responses (typically from laboratories using a more limited test panel or not performing the VWF:CB). Overall, problems with the non‐identification of functional VWF discordance in type 2 VWD, the misidentification of functional VWF discordance in type 1 VWD, and difficulties in discriminating types 1 and 3 VWD appear to predominate. In comparison with previous surveys, performance of electro‐immuno diffusion (EID) (or Laurel gel) procedures has now ceased, and a reduction in VWF:RCo and VWF:Multimer testing and an increase in latex immunoassay (LIA) testing is sustained. We conclude that laboratories are generally proficient in tests for VWD, and that diagnostic error rates are reduced when test panels are more comprehensive and include the VWF:CB.

The effect of dabigatran on haemostasis tests: a comprehensive assessment using in vitro and ex vivo samples

Summary The new direct oral anticoagulants (DOACS) dabigatran, rivaroxaban, apixaban and edoxaban provide alternatives to warfarin for treatment and prevention of atrial fibrillation and venous thromboembolic disease in various settings. These have been developed as not requiring laboratory monitoring; however, under certain clinical situations, including recent haemorrhage/thrombosis, emergency surgical procedures, testing may be indicated. The aim of this study was to assess findings of haemostasis laboratory tests for one of the DOACs, dabigatran (Pradaxa), tested across a wide range of laboratory assays. Laboratories (n = 72) enrolled in the Royal College of Pathologists of Australasia Quality Assurance Programs (RCPAQAP) Haematology program were sent set(s) of seven dabigatran spiked plasma samples covering the concentration 0–800 ng/mL. Also, 30 ex vivo patient samples under therapy with dabigatran were assessed. Prothrombin time and activated partial thromboplastin time assays showed some sensitivity to dabigatran; however, a normal result could not inform on drug exclusion. The thrombin time (TT) was very sensitive to dabigatran, and a normal TT could generally be used for drug exclusion. More specialised assays such as the Hemoclot, a direct thrombin inhibition assay, and in-house dilute TT methods, showed good reproducibility and concordance with expected drug levels assessed by mass spectrometry and were effective to quantify drug levels. Dabigatran also affected factor assays, lupus anticoagulant and factor inhibitor measurement, leading to potential misinterpretation of test results. Ex vivo sample testing provided similar and extended information. Dabigatran affects many haemostasis tests. Some can be used to predict the presence, absence or quantity of dabigatran in patient plasma. For others, interference may lead to false conclusions regarding patients’ haemostatic status.

Investigations from External Quality Assurance Programs Reveal a High Degree of Variation in the Laboratory Identification of Coagulation Factor Inhibitors

The laboratory has a key role in the initial detection of factor inhibitors and an ongoing role in the measurement of inhibitor titers during the course of inhibitor eradication therapy. The most commonly seen factor inhibitors are those directed against factor VIII (FVIII), usually detected either with the original or the Nijmegen-modified Bethesda assay. In addition, several circumstances can arise in which the laboratory may test samples that potentially reflect false identification of factor inhibitors. These include lupus anticoagulants and other events generally related to preanalytical variables, including incorrect sample presentations. This article reviews each of these elements, largely from the perspective of cross-laboratory studies undertaken within the framework of external quality assurance (EQA), a peer-laboratory process that aims to assess the ongoing performance of groups of similar laboratories. This review details the experience of the Royal College of Pathologists of Australasia Haematology Quality Assurance Program, and it also reflects on the experience of other EQA organizations. Our analysis reveals a wide variety of test practice among inhibitor testing laboratories, a wide variation in detected inhibitor levels in cross-tested samples, and substantial evidence of false-positive and false-negative detection of factor inhibitors. These findings hold some significance for the clinical management of patients affected by these inhibitors. There is still much need for standardization and improvement in factor inhibitor detection, and we hope that this report provides a basis for future improvements in this area.

Laboratory diagnosis of von Willebrand disorder. Current practice in the southern hemisphere.

A survey of 44 laboratories was conducted to evaluate current testing proficiency in the diagnosis of von Willebrand disorder (vWD) and to assess recent changes in test practices. Laboratories performed their usual panel of tests for vWD and interpreted results for the likelihood of vWD and potential subtype. Samples were as follows: normal plasma; borderline normal or abnormal levels of von Willebrand factor (vWF) and factor VIII; type 3 vWD; type 2A vWD; and 2 samples from a healthy person, processed after handling at 22°C and 4°C, respectively. Interassay and within-method coefficients of variation were similar for all assays (approximately 15%–25%). Most laboratories reported test values consistent with expected findings and made correct interpretations, although discrepant results for 5% to 10% of responses are of concern. For the sample stored at 4°C, all laboratories detected low or borderline levels of vWF and factor VIII coagulant, and no laboratory identified this sample as from a healthy person. In contrast, for the sample stored at 22°C, most laboratories reported normal results. Compared with previous results, performance of some assays has declined while that of others has increased. Laboratories generally are proficient in tests for vWD, and transport of samples at 4°C before processing may lead to false identification of vWD, suggesting that NCCLS guidelines should be reviewed.

Type 2M von Willebrand disease - more often misidentified than correctly identified.

Appropriate diagnosis of von Willebrand disease (VWD), including differential identification of qualitative vs. quantitative von Willebrand factor (VWF) defects has important management implications, but remains problematic.

Laboratory diagnosis of von Willebrand disorder: use of multiple functional assays reduces diagnostic error rates.

Regular multilaboratory surveys of laboratories primarily in Australia, New Zealand, and Southeast Asia have been conducted over the past 8 years to evaluate testing proficiency in the diagnosis of von Willebrand disorder (VWD). We have reassessed the findings of these surveys with a particular emphasis on the diagnostic errors and error rates associated with particular tests or test panel limitations. The 37 plasma samples dispatched to survey participants include 9 normal samples, 4 type 1 VWD samples, 8 type 2 VWD samples (2A x 3, 2B x 3, 2M x 1, and 2N x 1), and 4 type 3 VWD samples. In addition to providing numerical test results, participant laboratories (average, n = 35) were asked to provide diagnostic interpretations of their test results regarding whether VWD was evident and, if so, the probable subtype. Although laboratories usually provided correct interpretative responses, diagnostic errors occurred in a substantial number of cases. On average, type 1 VWD plasma was misidentified as type 2 VWD plasma in 11% of cases, and laboratories that performed the ristocetin cofactor assay for von Willebrand factor (VWF:RCo) without performing the collagen-binding activity assay for VWF (VWF:CB) were 6 times more likely to make such an error than those that did perform the VWF:CB. Similarly, type 2 VWD plasma samples were misidentified as type 1 or type 3 VWD in an average of 20% of cases, and laboratories that performed the VWF:RCo without the VWF:CB were 3 times more likely to make such an error than those that performed the VWF:CB. Finally, normal plasma was misidentified as VWD plasma in an average of 5% of cases, and laboratories that performed the VWF:RCo without the VWF:CB were 10 times more likely to make such an error than those that performed the VWF:CB. We conclude that laboratories are generally proficient in their testing for VWD and that diagnostic error rates are substantially reduced when test panels are more comprehensive and include the VWF:CB.

The diagnostic dilemma: dual presentations of clinical mucosal bleeding and venous thrombosis associated with the presence of thrombophilia markers and mild reduction in von Willebrand factor.

A prothrombotic and hemorrhagic state can separately manifest in one patient and can potentially cause several diagnostic problems. We report an intriguing case as an example of a potential hemostasis-based diagnostic dilemma. A 29-year-old female patient presented with a personal history of menorrhagia and other mucosal bleeding and renal ovarian thrombosis. Previous investigations had uncovered several diagnostic anomalies, including von Willebrand disease (VWD), factor V Leiden (FVL), antiphospholipid syndrome, and thrombocytopaenia. Previous therapy in this patient included heparin and warfarin for the thrombosis and desmopressin acetate (DDAVP) and antifibrinolytic therapy for surgical management. Subsequent laboratory testing with fresh samples consistently confirmed an equivocal (borderline normal/abnormal) level of von Willebrand factor (VWF) and FVL with activated protein C resistance (APCR). A patient sample, differentially labeled according to the tests being performed, was later distributed for blind testing to participants within several modules of the RCPA Quality Assurance Program (QAP). Most participants reported a low level of VWF consistent with possible mild Type 1 VWD, and most (but not all) reported a positive finding for APCR. All participants correctly reported the sample as heterozygous for the FVL mutation, negative for the Prothrombin gene mutation G20210A, and heterozygous for the methylenetetrahydrofolate reductase (MTHFR) mutation C677T. Interestingly, a significant number of laboratories performing Protein S testing using clot-based procedures also identified a false Protein S deficiency. In conclusion, this exercise showed how, either depending on the clinical review and specific laboratory investigation and tests performed, a pro-bleeding diagnosis (of either VWD or thrombocytopenia) or pro-thrombophilia risk (Antiphospholipid Syndrome or FVL/APCR or false Protein S deficiency) could potentially and differentially arise in the one patient.

Future developments in the RCPA haematology QAP

The Royal College of Pathologists of Australasia Quality Assurance Program (RCPA QAP) for Haematology Morphology was introduced in 1965. In these surveys participants submitted morphological findings and diagnoses on unknown blood and bone marrow films as free text. QAP staff then reviewed each response for key words or themes and recorded them manually. Responses were graded as acceptable or unacceptable based on the advice of an expert morphology referee panel. Issues to assess the laboratory performance were raised and our aim was to develop software to assess the laboratory performance. Software was developed to facilitate an online data entry facility, which included a comprehensive code list, where participants were required to submit the relevant code for the important morphological features and diagnosis, either online or as hard copy. An expert advisory committee allocated a score for each morphological feature and diagnosis according to written criteria. The reports included histograms and tables with details of the individual participant’s scores as well as information and ranking of all participants’ responses and scores. This reporting system is now used for our Morphology and Malarial Parasite Program. An end of cycle report is currently being developed to rank subscriber performance, which will be discussed, as well as changes made to the interim report. Virtual microscopy has also become an important adjunct to our Morphology Program for the purpose of education and a DVD is now available. A review of the DVD will be illustrated in this presentation as well as other possible future developments for virtual microscopy.