Towards improved diagnosis of von Willebrand disease: comparative evaluations of several automated von Willebrand factor antigen and activity assays.

TR-05686; No of Pages 9 Thrombosis Research xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

Thrombosis Research journal homepage: www.elsevier.com/locate/thromres

Regular Article

Towards improved diagnosis of von Willebrand disease: Comparative evaluations of several automated von Willebrand factor antigen and activity assays Emmanuel J. Favaloro ⁎, Soma Mohammed Department of Haematology, Institute of Clinical Pathology and Medical Research (ICPMR), Pathology West, Westmead Hospital, Westmead, NSW, Australia

a r t i c l e

i n f o

Article history: Received 18 July 2014 Received in revised form 1 September 2014 Accepted 20 September 2014 Available online xxxx Keywords: von Willebrand disease von Willebrand factor Laboratory testing Collagen binding Ristocetin cofactor INNOVANCE VWF activity

a b s t r a c t Introduction: von Willebrand disease (VWD) is reportedly the most common bleeding disorder and arises from deficiency and/or defects of von Willebrand factor (VWF). Laboratory diagnosis and typing has important management implications and requires a wide range of tests, including VWF activity and antigen, and involves differential identification of qualitative vs quantitative defects. Methods: We have assessed several VWF antigen and activity assays (collagen binding [VWF:CB], ristocetin cofactor [VWF:RCo] and the new Siemens INNOVANCE assay [VWF:Ac], employing latex particles and gain of function recombinant glycoprotein Ib to facilitate VWF binding and agglutination without need for ristocetin) using different instrumentation, including the new Sysmex CS-5100, with a large sample test set (n = 600). We included retrospective plus prospective study designs, and also evaluated desmopressin responsiveness plus differential sensitivity to high molecular weight VWF. Results: VWF:Ag and VWF:RCo results from different methods were respectively largely comparable, although some notable differences were evident, including one high false normal VWF:Ag value (105 U/dL) on a type 3 VWD sample, possibly due to heterophile antibody interference in the latex-based CS-5100 methodology. VWF:Ac was largely comparable to VWF:RCo, but VWF:CB showed discrepant findings to both VWF:RCo and VWF:Ac with some patients, most notably patients with type 2M VWD. Conclusions: (a) VWF:Ag on different platforms are largely interchangeable, as are VWF:RCo on different platforms, except for occasional (some potentially important) differences, and manufacturer recommended methods may otherwise require some assay optimization; (b) VWF:RCo and VWF:Ac are largely interchangeable, except for occasional differences that may also relate to assay design (differing optimizations); (c) VWF:CB provides an additional activity to supplement VWF:RCo or VWF:Ac activity assays, and is not interchangeable with either. Crown Copyright © 2014 Published by Elsevier Ltd. All rights reserved.

Introduction von Willebrand disease (VWD) is reportedly the most common congenital bleeding disorder and arises from deficiency and/or defects of von Willebrand factor (VWF), an adhesive plasma protein essential for effective primary haemostasis [1,2]. Clinical identification, diagnosis and sub-typing of VWD are aided by laboratory testing but remain problematic for a variety of reasons. First, VWD is very heterogeneous,

Abbreviations: AVWS, acquired von Willebrand syndrome; ELISA, enzyme linked immunosorbent assay; EQA, external quality assessment; FVIII:C, FVIII coagulant; GPIb, glycoprotein Ib; HMW, high molecular weight (VWF); LIA, latex-particle immuno-assay; VWD, von Willebrand disease; VWF, von Willebrand factor; VWF:Ac, VWF INNOVANCE activity (assay); VWF:Ag, VWF antigen (assay); VWF:CB, VWF collagen binding (assay); VWF: RCo, VWF ristocetin cofactor (assay). ⁎ Corresponding author at: Department of Haematology, Institute of Clinical Pathology and Medical Research (ICPMR), Pathology West, Westmead Hospital, WESTMEAD, NSW, 2145, Australia. Tel.: +61 612 9845 6618; fax: +61 612 9689 2331. E-mail address: [email protected] (E.J. Favaloro).

as mirrored by the fact that VWF possesses many functional properties, including binding to platelets via several receptors, most notable glycoprotein Ib (GPIb), binding to sub-endothelial matrix components (most notably collagen), and binding and protection of factor VIII (FVIII) function [3]. Thus, defects may occur anywhere within the VWF gene, and lead to a wide variety of phenotypes on a case-by-case basis. Second, the laboratory tests used to assist identification, diagnosis and sub-typing of VWD are themselves heterogeneous and imperfect, being reflective of varied sensitivity to VWF level and activity, as well as often processing considerable inherent variability [4–7]. The most recent classification scheme from the International Society on Thrombosis and Haemostasis (ISTH) recognizes six different types of VWD [8]. Type 1 represents a partial quantitative VWF deficiency, with VWF essentially functionally normal, but produced in lowered quantity. Type 3 VWD represents ‘complete’ deficiency of VWF. Type 2 VWD represents a heterogeneous group of qualitative VWF defects that comprise (i) 2A VWD (loss of high molecular weight (HMW) VWF), (ii) 2B VWD (enhanced functional binding of VWF that leads to loss of HMW VWF

http://dx.doi.org/10.1016/j.thromres.2014.09.024 0049-3848/Crown Copyright © 2014 Published by Elsevier Ltd. All rights reserved.

Please cite this article as: Favaloro EJ, Mohammed S, Towards improved diagnosis of von Willebrand disease: Comparative evaluations of several automated von Willebrand factor antigen and activity assays, Thromb Res (2014), http://dx.doi.org/10.1016/j.thromres.2014.09.024

2

E.J. Favaloro, S. Mohammed / Thrombosis Research xxx (2014) xxx–xxx

and typically mild thrombocytopenia), (iii) 2N VWD (loss of VWF-FVIII binding), and (iv) 2M VWF (VWF dysfunction not associated with loss of HMW VWF). For cases reflecting a minor quantitative deficiency of VWF but without a formal diagnosis of VWD, the concept of ‘low VWF’ as a risk factor for bleeding has alternatively been proposed [9]. The proper identification of VWD and differentiating its type is important for therapeutic management [6,10]. In normal practice, VWD and VWD type can be determined by laboratory testing that encompasses a broad panel of different tests [1,2,4–9]. Virtually all laboratories perform VWF antigen (VWF:Ag) and FVIII coagulant (FVIII:C) [1,5], respectively measuring the level of VWF protein and FVIII activity. VWF: Ag is most usually assessed using either ELISA (enzyme linked immunosorbent assay) or LIA (latex-immuno-assay) technologies. The most commonly performed activity based test is VWF ristocetin cofactor (VWF:RCo) [1,2,4–9,11], usually performed as a platelet agglutination assay using aggregometry or automated methods with standard coagulation instruments. VWF collagen binding (VWF:CB) is an additional VWF activity assay typically performed by a smaller proportion of laboratories, typically by ELISA [1,2,4–9,11]. Ostensibly, VWF:RCo and VWF: CB represent surrogate laboratory markers for two essential in vivo primary haemostasis VWF functions (notably, platelet GPIb and collagen binding respectively). As both assays also have a similar preference for high molecular weight (HMW) VWF, reflective of the most adhesive/ functional form of VWF, both can also be used as surrogate markers for HMW VWF [12,13], which may otherwise be assessed using gelbased assays of VWF multimer analysis to determine loss of HMW VWF as well as structural abnormalities [14]. A variety of additional or alternative VWF ‘activity’ assays may be employed by laboratories, either in addition to, or instead of, VWF:RCo and VWF:CB [1,7]. These assays include newly released commercial assays, many based on LIA technology, but generally representing distinct assays, using distinct reagents, and detecting VWF differently. Some of these assays do not use ristocetin, so cannot truly claim to be VWF:RCo assays, although many laboratories may use these instead of VWF:RCo. Indeed, one of these new assays, the Siemens INNOVANCE VWF activity (VWF:Ac) assay uses recombinant ‘mutant/gain of function’ GPIb that binds VWF in the absence of ristocetin, and has been proposed to represent a possible replacement to VWF:RCo. Importantly, the perceived problems with classical VWF:RCo have lead to a high proportion of laboratories changing over to the VWF:Ac assay [5], despite the paucity of comparative published data. Nevertheless, the theoretical benefits of the VWF:Ac compared to VWF:RCo include easier performance (including easier automation), better assay precision and accuracy, lack of sensitivity to VWF polymorphisms affecting ristocetin binding, as well as improved detection of VWF at low levels [7]. Accordingly, we have assessed the comparative utility of different VWF activity assays, namely VWF:CB, VWF:RCo as well as the new Siemens INNOVANCE VWF:Ac assay, using a large sample set, and both a retrospective and prospective study design. Methods Assays and instrumentation FVIII:C was assessed for patient samples as part of their original analysis within our standard VWD diagnostic test panel, using a one stage clot-based assay on a Behring BCS analyser (Siemens Healthcare, Marburg, Germany) and Siemens reagents, but has not otherwise been formally assessed in the current report. Our standard VWF:Ag was performed as an in house sandwich ELISA assay, essentially as previously reported [15,16]. The assay is currently performed on a BioKit BEST 2000 ELISA workstation (Werfen, Barcelona, Spain), using polyclonal antibodies from Dako (Glostrup, Denmark; rabbit anti-human VWF; catalogue no. A0082) for coating 96-well plates (Linbro Titertek EIA plate; ICN Biomedicals, Aurora, OH, USA) and Dako horse radish peroxidase labeled rabbit anti-human VWF (catalogue no. P0226), as

the detection system. Our current standard VWF:CB was performed in parallel with VWF:Ag on the same ELISA workstation as an in house sandwich ELISA assay, essentially as previously reported [15,16], using bovine collagen from ICN Biomedicals (catalogue no. 193492) for coating 96-well plates (Pierce Maleic Anhydride activated plates; catalogue no. 15110; Thermo Scientific, Rockford, USA) and Dako horse radish peroxidase labeled rabbit anti-human VWF (catalogue no. P0226), as the detection system. Our standard VWF:RCo was performed as an agglutination assay, essentially as previously reported [15–17], on a Behring BCS analyser using Siemens reagents (BC VWF reagent; catalogue no. 10446425). For the purpose of this study, these three assays define the assay ‘reference’ methods, and they have been thoroughly validated for use in the identification and preliminary typing of VWD cases, as previously extensively reported by our laboratory [7,15–25]. Siemens Standard Human Plasma (catalogue no. 10446238) was used as the calibrator for all three ‘reference’ assays. The test systems evaluated in this study comprised additional assays performed on a Siemens CS-5100 instrument, and namely: (a) VWF:Ag using a LIA based Siemens assay (catalogue no. 10445967), (b) VWF: RCo using the Siemens agglutination assay (BC VWF reagent; catalogue no. 10446425; same reagent methodology as employed on the BCS); (c) VWF:Ac using the Siemens INNOVANCE VWF activity assay (catalogue no. 10487040). Each of these tests were performed using standard manufacturer provided protocols for the CS-5100, except for VWF:RCo which was performed using the manufacturer protocol plus an additional (‘low curve’) protocol based on a published validated method for the BCS instrument [17]. The CS-5100 is the latest addition to the CS family of systems from Sysmex Corporation (Kobe, Japan), and is distributed by Siemens in Australia. The CS-5100 instrument represents advanced instrumentation using a random access multi–wavelength scanning of clotting reactions. Further information is available via the Siemens website (http:// www.healthcare.siemens.com/hemostasis/systems/sysmex-cs-5100system). Of importance to the current report is that the system is optimised for use of Siemens INNOVANCE reagents, including the VWF:Ac assay, and also incorporates active stirring for VWF:RCo agglutination reactions. The tests evaluated in this study are summarised in Table 1. Retrospective study and VWD cases The retrospective study comprised assessment of a large set of samples (n = 97) from patients previously identified to have VWD or possible VWD according to clinical bleeding histories and subsequent assessment with our standard (‘reference’) test panels, selectively supplemented as required by additional evaluation using desmopressin trials, genetic testing and multimer analysis [5,25–27].

Table 1 Summary of VWF test methods comparatively evaluated in this study. VWF assay

Description

VWF: Ag

Assessment of VWF protein level using an 'antigen' assay. Performed in this study by automated ELISA using an ELISA workstation (Best 2000) (as ‘reference’) and by automated LIA using a CS-5100 (as ‘comparator’) Assessment of VWF activity level utilising ristocetin and an 'agglutination' assay. Performed in this study by automated assays using BCS (‘reference’) and CS-5100 (‘comparator’) instruments. Assessment of VWF activity level utilising collagen. Performed in this study by automated ELISA using an ELISA workstation (Best 2000). Siemens INNOVANCE 'activity' assay. Assessment of VWF activity level utilising a Glycoprotein Ib binding method. The system employs two gain of function Glycoprotein Ib mutations within a recombinant molecule that facilitates VWF binding. Performed in this study by automated LIA using a CS-5100.

VWF: RCo VWF: CB VWF: Ac

Abbreviations: ELISA, enzyme linked immunosorbant assay; LIA, latex-particle immunoassay; VWF, von Willebrand factor.



For the purpose of this report, VWD cases were identified as type 1 VWD (‘VWD-1’) when levels of VWF fell below 36 U/dL, but VWF activity assays showed concordance with VWF:Ag (i.e., VWF:CB/VWF:Ag [CB/Ag] and VWF:RCo/VWF:Ag [RCo/Ag] ratios were both above 0.7). These cases were sub-classified as ‘severe’ type 1 (‘VWD-1s’) when VWF:Ag was below 16U/dL, and ‘moderate’ type 1 (‘VWD-1m’) when VWF:Ag levels were between 16-35U/dL. A separate group of patients with VWF:Ag results close to the normal range cut-off value were identified as plausibly suffering from ‘low VWF’ [9] (as a risk factor for bleeding) when VWF:Ag levels were between 36-60U/dL (‘VWD-1p’), together with VWF:CB and VWF:RCo concordance. VWD cases were identified as type 3 (‘VWD-3’) when VWF:Ag levels were below 5U/dL. VWD cases were defined as type 2 VWD when there was evident discordance between VWF:Ag and functional VWF assays. Type 2N VWD cases were not included in this study since the investigated VWF activity assays do not substantially contribute to their diagnosis (which requires either a VWF:FVIII binding assay and/or genetic analysis). Type 2B VWD cases (‘VWD-2B’) were identified by heightened ristocetin induced platelet aggregation (RIPA) and this was confirmed where possible by genetic analysis. Most, but not all, cases of 2B VWD show functional VWF discordance with CB/Ag and RCo/Ag below 0.7. Type 2A VWD cases (‘VWD-2A’) were primarily characterized by functional VWF discordance with both CB/Ag and RCo/Ag below 0.7, confirmed by repeat testing, supplemented if available with data from desmopressin responsiveness and multimer analysis, and confirmed where possible by genetic analysis [5,25–29]. Where confirmation was not possible, the term provisional 2A VWD is used (‘pVWD-2A’). Platelet binding defect type 2M VWD cases (‘VWD-2M’) were primarily characterized by functional VWF discordance with RCo/Ag below 0.7, but normal CB/Ag (N0.7), confirmed by repeat testing, supplemented if available with data from desmopressin responsiveness and multimer analysis, and confirmed where possible by genetic analysis [5,25–29]. Where confirmation was not possible, the term provisional 2M VWD is used (‘pVWD-2M’). This is summarised in Table 2, which also details test sample numbers.

Historical results For the retrospective study, some comparisons were also performed for new data vs historical data, the latter reflecting the original test data obtained on the samples prior to their storage and subsequent retrieval

Table 2 Summary of samples evaluated in this study. Sample type Individual normal samples ‘Low VWF’ Type 1 VWD

Type 3 VWD Type 2 VWD

Platelet-type VWD Non-VWD cases⁎ Other⁎⁎ Totals

Category

‘VWD-1p’ ‘Moderate’ (‘VWD-1m’) ‘Severe’ (‘VWD-1s’) ‘VWD-3’ ‘VWD-2A’ + ‘pVWD-2A’ ‘VWD-2M’ + ‘pVWD-2M’ ‘VWD-2B’ ‘PT-VWD’

Retrospective study

Prospective study

Total

56

56

30 8

30 3

60 12

3 4 21

0 2 7

3 6 28

18

6

24

11 2

0 0 242

97

346

11 2 242 156 600

⁎ Cases that were assessed for VWF/VWD in the prospective study that were found not to have VWD following laboratory testing ⁎⁎ all other samples, including VWD post desmopressin or post concentrate samples, purpose made HMW deficient test samples, and other miscellaneous test samples.

3

for the current retrospective study. Historical test methods were essentially the same as the current ’reference’ methods identified above. Prospective study The prospective study involved comparative testing of all patient samples submitted for VWF testing to our laboratory for a period of 8 months, and using all of the laboratory assays noted previously, to assess for any differential findings. In total, 290 patient samples were assessed in this part of the study, although as typical of routine VWF testing, these mostly comprised samples that were later defined to not be VWD cases. In addition, a set of 56 normal individual samples was cross-tested with all assays, as was a small set of four samples prepared to yield step-wise loss of HMW VWF, similar to those recently reported by our laboratory [30]. Desmopressin response study We also assessed several sets of desmopressin trial samples with all assays, to again determine comparability. These were mostly retrospectively collected and stored samples, and comprised type 1, type 2A and 2M VWD cases. Data analysis Data in this report is analyzed largely using descriptive statistics, and comparisons are primarily made using regression analysis. In addition, testing for statistical equivalence of test data for different assays was performed using GraphPad Prism software (La Jolla, CA, USA) and non-parametric analysis. Results Comparability of assay data VWF:Ag Data for VWF:Ag performed using the ‘reference’ ELISA method vs the LIA method on the CS-5100 is shown in Fig. 1A. Results were in general highly correlated (r = 0.952), and indeed largely comparable to historical VWF:Ag results (r = 0.946; see Table 3 for summary data). As shown in Fig. 1A, most data points were scattered around the line of equivalence, although a few points fell outside expected limits. As this occurred for both comparisons, this would most likely be due to random error events in the majority of cases. However, there was a single high outlier value (circled in Fig. 1A) that was inconsistent with such an event, and where repeated testing showed identical findings (~100U/dL VWF:Ag by LIA, but b 5U/dL VWF:Ag by ELISA). This sample was derived from a patient historically characterized as a type 3 VWD, who also displayed VWF:CB, VWF:RCo and VWF:Ac values of b 5U/dL. The patient sample was also tested in another LIA based VWF:Ag assay (Stago instrument and Stago reagent) and similarly yielded a (false) normal VWF:Ag level of 74U/dL. This sample therefore most likely represents an analytical interference in the VWF:Ag LIA assay, potentially due to heterophile antibodies since rheumatoid factor was not detected in the sample. The ELISA VWF:Ag test data was otherwise statistically ‘equivalent’ to VWF:Ag performed on the CS-5100, but not to the data derived from other VWF assays (Table 3). VWF:RCo Data for VWF:RCo performed using the ‘reference’ method (BCS analyzer) vs the test method (CS-5100 analyser) is shown in Fig. 1B. Results were again in general highly correlated (r = 0.962), and largely comparable to historical VWF:RCo results (r = 0.968; see Table 3 for summary data). As shown in Fig. 1B, and similar to VWF:Ag comparisons, most data points were scattered around the line of equivalence, although a few points fell outside expected limits. Again, as this occurred


4


Fig. 1. Comparative data for various assays for all samples assessed in this study (n = 600), and shown by regression analysis, using ‘reference’ method on x-axis, and comparative test data on y-axis. All data shown in U/dL in Figure A-C. The line of equivalence identifies ‘reference method’ vs ‘reference method’ data. The dotted lines in this figure reflect the linear regression lines for each comparator method. Dotted lines close to the line of equivalence reflect minimal bias of data. Figure A: VWF:Ag ‘reference method’ data (automated ELISA) vs historical VWF: Ag ELISA data (Ag-E-O) and vs CS-5100 LIA VWF:Ag (C5 Ag L). Most data followed the line of equivalence with acceptable variation around this line, except for occasional data; one such data point (circled) identified a high false value for VWF:Ag by the LIA method in a type 3 VWD patient. Figure B: VWF:RCo ‘reference’ method data (automated agglutination by BCS) vs historical VWF:RCo by same method (RCo-O) and vs CS-5100 automated VWF:RCo agglutination method using the same reagent (C5 RCo). Most data followed the line of equivalence with acceptable variation around this line, except for occasional data, which is largely expected to be due to the relative high assay variation with this assay methodology. Figure C: VWF:RCo ‘reference method’ data (automated agglutination by BCS) vs VWF:Ac on CS-5100 (VWF:Ac) or vs VWF:CB (CB-R; automated ELISA on Best 2000). Most data followed the line of equivalence with acceptable variation around this line, except for occasional data, which is due to a combination of assay variation plus differential activities identified. Figure D: VWF:RCo/VWF: Ag ratios using ‘reference’ methods vs other VWF Activity / VWF:Ag ratios (VWF:Ac/VWF:Ag by ELISA = Ac/Ag E; VWF:Ac/VWF:Ag by LIA = Ac/E L; VWF:CB/VWF:Ag by ‘reference’ ELISA methods = CB/Ag-R; VWF:RCo (CS-5100) / VWF:Ag ELISA = C5 RCo/Ag E; VWF:RCo (CS-5100) / VWF:Ag LIA = C5 RCo/Ag L). Data is scattered around the line of equivalence based on multiplied effects of assay variation as well as differential activities identified.

for comparisons of VWF:RCo BCS vs CS-5100 and VWF:RCo BCS vs historical VWF:RCo BCS data, these would most likely be due to random error (assay variability) events. Data comparability for VWF:RCo (‘reference’ BCS method) vs VWF: Ac (Siemens INNOVANCE on CS-5100) and vs VWF:CB is shown in Fig. 1C. Results were again in general highly correlated (r = 0.958 and 0.908, respectively). That the correlation is better for the former comparison is consistent with the nature of the activities detected (GPIb binding related for both VWF:RCo and VWF:Ac), and is explored in more detail later in this report. The ‘reference’ BCS VWF:RCo test data was statistically ‘equivalent’ to both VWF:RCo and VWF:Ac performed on the CS-5100, but not to the data derived from other VWF assays (Table 3).

CB/Ag. Again, the likely explanation for this slightly reduced comparability is the differential activities expressed by collagen (VWF:CB) vs GPIb (VWF:RCo, VWF:Ac) binding. To further explore this, data was sub-analysed for a subset of samples comprising VWD and normal samples. As shown in Fig. 2, there was a clear set of samples that showed discrepancy in VWF activity/antigen ratios, and these were determined to be primarily type 2M VWD samples. The ‘reference’ RCo/Ag ratio test data was statistically ‘equivalent’ to both RCo/Ag (CS-5100/ELISA) and Ac/Ag (CS-5100/ELISA), but not to the data derived from other assay combinations (Table 3), including RCo/Ag (CS-5100/CS-5100) and Ac/Ag (CS-5100/CS-5100). Additional correlation data is presented in the summary Table 3. Desmopressin test data

Activity to Antigen ratios Data for RCo/Ag performed using the ‘reference’ methods (BCS analyzer/ELISA) vs other VWF activity/antigen ratios is shown in Fig. 1D. As these ratios each derive from two separate assays, the additive assay variations lead to amplified ‘signal noise’, and thus reduced correlation of data. Nevertheless, all comparisons showed reasonable correlations, and correlation was strongest (and similarly so) for RCo/Ag and Ac/Ag ratios in all possible combinations (r values = 0.750 to 0.856; see also Table 3 summary data). A slightly reduced correlation was evident for

Data for select patients is shown in Fig. 3. Type 1 VWD and ‘low VWF’ (‘VWD-1p’) patients showed similar rises in all test parameters with VWF activity/antigen ratios above 0.5 at all time points. Patients with type 2A VWD showed good increments in VWF:Ag, but minimal increments in VWF activity assays, and VWF activity/antigen ratios remained below 0.5 at all time points. Patients with platelet GPIb binding defect type 2M showed good increments in VWF:Ag, reasonable increments in VWF:CB, but little increments in VWF:RCo and VWF:Ac; thus, CB/Ag


E.J. Favaloro, S. Mohammed / Thrombosis Research xxx (2014) xxx–xxx Table 3 Summary of linear regression r-values and statistical ‘equivalence’ test for different comparisons. VWF assay ‘reference’

VWF comparator assay

Regression r-value*

Statistical equivalence?

VWF:Ag by ELISA

VWF:Ag by ELISA (historical data)

0.946

NA**

VWF:Ag by LIA on CS-5100 VWF:RCo on BCS VWF:RCo on CS-5100 VWF:Ac on CS-5100 VWF:CB by ELISA VWF:CB by ELISA (historical data)

0.952 0.845 0.844 0.842 0.891 0.919

Yes No No No No NA**

VWF:RCo on BCS (historical data)

0.968

NA**

VWF:RCo on CS-5100 VWF:Ac on CS-5100 VWF:CB by ELISA RCo/Ag (‘reference’ methods; historical data)

0.962 0.958 0.908 0.757

Yes Yes No NA**

RCo/Ag (CS-5100/CS-5100) RCo/Ag (CS-5100/ELISA) Ac/Ag (CS-5100/CS-5100) Ac/Ag (CS-5100/ELISA) CB/Ag

0.758 0.853 0.750 0.856 0.712

No Yes No Yes No

VWF:CB by ELISA VWF:RCo on BCS

RCo/Ag (‘reference’ methods)

Abbreviations: ELISA, enzyme linked immunosorbant assay; LIA, latex-particle immunoassay; VWF, von Willebrand factor. Note: *p-value b0.001 for all comparisons. **NA, not applicable, as different numbers of samples within each data set.

ratios remained normal (N0.5), whereas RCo/Ag and Ac/Ag ratios remained abnormal (b0.5). Differential sensitivity to loss of HMW VWF Using a set of four samples with step-wise reduction in HMW VWF, VWF:Ag and FVIII:C values were determined to remain within normal levels, whereas all VWF activity assays showed sequential falls in values (Fig. 4). There was a suggestion that the VWF:RCo (CS-5100) and thus RCo/Ag (CS-5100) ratios were least responsive in this analysis. Differentials in VWD identification/diagnosis An assessment was undertaken of the differential in identification and preliminary sub-classification of VWD type (should VWD have

Fig. 2. VWF:RCo/VWF:Ag ratios using ‘reference methods’ (RCo/Ag) vs other VWF Activity / VWF:Ag ratios for selected data (normal samples [Nor] plus VWD cases [types 1, 2A, 2B and 2M, as respectively identified]). VWF:Ac/VWF:Ag shown for CS-5100 = Ac/Ag, and VWF:CB/VWF:Ag by ‘reference’ ELISA methods = CB/Ag. Deviation of data is greatest for CB/Ag, and especially evident for type 2M VWD cases. Cross broken lines at 0.65 represent a nominal cut-off value for functional VWF activity to antigen concordance/discordance, the latter reflecting a type 2 VWD-like pattern.

5

been suggestive of preliminary testing). Except for previously noted cases, there were no major discrepancies in these between the original (historical) findings (using the ’reference’ assays and test panel), the repeat (‘retrospective study’) test findings using the same ’reference’ assays and test panel, and the evaluation tests and various test panels. There were, however, some minor discrepancies in all evaluations, proposed to be due primarily to test result variation (random error events), and summarised as follows: (i) normal samples with values close to normal/abnormal cut-off values gave values within normal range on occasion and values just below the normal range on different occasions, and thus samples would be identified as normal on some occasions, and possible ‘low VWF’ (‘VWD-1p’) on other occasions; (ii) samples given a provisional interpretation of ‘low VWF’ (‘VWD-1p’, moderate type 1 VWD (‘VWD-1m’), or severe type 1 VWD (‘VWD’1s’), on one occasion might be given another provisional interpretation one step to either side of these categories (based on movements of test values around the respective cut-off values defined in the Methods section); (iii) samples given a provisional interpretation as possible type 2M (‘pVWD-2M’) or 2A VWD (‘pVWD-2A’) based on activity/antigen discordance using one class (RCo/Ag or Ac/Ag) or two classes (RCo/Ag and/or Ac/Ag plus CB/Ag) of activity assays respectively, were occasionally given the alternate provisional interpretation based on similar (usually small) movements in test values. In all the above cases, standard operating procedures, including repeat testing using a fresh sample on another occasion, or confirmatory and extended testing using additional test procedures, would have provided clarity of the provisional diagnosis. Discussion In this study, a large number of overall test samples (n = 600) of diverse nature have been cross-tested using a variety of VWF assays. There was a high level of concordance between VWF:Ag assays performed using our standard (‘reference’) ELISA method and the automated LIA based assay on the CS-5100 (Fig. 1A). Most outliers are likely to reflect assay variation (random error) events, given that similar occasional outliers were seen using repeat testing (ELISA) vs original (‘historical’) test (ELISA) data comparisons. Occasional outliers reflect an alternate explanation. One significant outlier (circled in Fig. 1A) reflected a false ‘normal’ VWF:Ag result by the LIA method in a type 3 VWD case. This sample also gave a false ‘normal’ VWF:Ag result using another LIA method (Stago); rheumatoid factor (known to reflect a potential false elevation in test results by LIA assays [31]) was normal in this sample, and so this result may reflect the possibility of another interference such as heterophile antibodies. This has not been reported for this assay to our knowledge, and could not be directly evaluated in this case due to insufficient remaining sample, but has been reported for other LIA based assays [32]. Accordingly, some of the other significant outliers (Fig. 1A) could reflect similar events. The assay manufacturer (Siemens) advises in the product information that ‘anti-bovine albumin’, ‘anti-rabbit-antibodies’ and ‘rheumatoid factor’ may lead to an over-estimation of VWF:Ag, and this case may reflect one such event, although rheumatoid factor was excluded; anti-rabbit-antibodies are also unlikely as rabbit-antibodies are also used in the ELISA assay. It is also important to reflect that the ELISA vs LIA test procedures reflect different test procedures that will lead to additional discrepancies on a case-by-case test basis. This also seems reflected by the differential assay sensitivities to the stepwise HMW deficiency samples (Fig. 4) and in desmopressin data (Fig. 3). In summary, VWF:Ag assessment by ELISA and LIA both measure VWF:Ag, and in most cases will provide similar data, but occasional discrepancies may be evident. In summary, however, the overall test data derived from each assay was statistically ‘equivalent’ (Table 3). There was also a high level of concordance between VWF:RCo assays performed using our standard (‘reference’) automated BCS method and the automated CS-5100 based assay (Fig. 1B). Most outliers are again (like the case for VWF:Ag above) likely to reflect assay variation


6


Fig. 3. Data for select VWD patients for desmopressin responses using different assays evaluated in this study. X-axis in each case reflects time-point of sample collection pre (‘0’ time) and post (in hours) desmopressin; left y-axis in each case identifies VWF level in U/dL, and right y-axis various VWF activity to antigen ratios. Panel A: patients with VWD-1p (‘low VWF’); panels B & C: patients with VWD-1m (‘moderate type 1 VWD’); panels D & E: patients with VWD-2A; panel F: patient with VWD-2M. Abbreviations: Ag ref = VWF:Ag (‘reference’ ELISA); Ag CS = VWF:Ag using automated LIA on CS-5100; Ac = VWF:Ac using CS-5100; CB = VWF:CB (‘reference’ ELISA); RCo ref (VWF:RCo using automated agglutination on BCS as reference method); RCo CS (VWF:RCo using automated agglutination on CS-5100); CB/Ag = VWF:CB/VWF:Ag using ‘reference methods’; Ac/Ag CS = VWF:Ac/VWF:Ag using CS-5100; RCo/Ag ref = VWF:RCo/VWF:Ag using ‘reference methods’; RCo/Ag CS = VWF:RCo/VWF:Ag using CS-5100.

(random error) events, given that similar occasional outliers were seen using repeat testing (BCS) vs original (‘historical’) test (BCS) data comparisons (Fig. 1B). It should also be noted that the VWF:RCo assay in itself reflects a highly variable assay [5], and this provides a major impetus for laboratories to move to alternative VWF activity assays; thus, such discrepancies may be expected in normal test practice. It is also worth noting that the ‘reference’ BCS VWF:RCo method has undergone considerable optimization in our laboratory, including low level VWF sensitivity to below 5U/dL [17]. The CS-5100 VWF:RCo assay, performed according to manufacturer recommendations, had a lower limit of detection of only 9U/dL or so. This could be further improved in our study using a modification in line with that previously published [17] and used on the BCS, to generate an assay with a limit of detection of b5U/dL. However, it is feasible that the level of optimization we have achieved on our BCS [17] is still better than that we have currently achieved on the CS-5100. Additional subtle differences between test procedure results were also evident in test data from the stepwise HMW deficiency samples (Fig. 4) and in desmopressin data (Fig. 3). In summary, VWF:RCo assessment by BCS and CS-5100 both measure VWF:RCo, and in most cases will provide similar data, but occasional discrepancies may still be evident. Overall test data derived from each assay was statistically ‘equivalent’ (Table 3). Further optimization of the VWF:RCo on the CS-5100 may be required to achieve the level of optimization reached on the BCS. There was also a high level of concordance between VWF:RCo assays and VWF:Ac and VWF:CB test data (Fig. 1C), although the former comparison showed a higher level of concordance. This is not unexpected, given that VWF:RCo and VWF:Ac both reflect platelet GPIb binding assays, whereas VWF:CB reflects another VWF activity marker, viz collagen binding. The bulk of test samples included in this study as shown in Fig. 1C reflected either non-VWD cases, or samples showing a range of VWF deficiencies. The non-concordance of data between VWF:RCo

and VWF:CB essentially reflects a combination of inherent assay variabilities plus different sensitivities to functional VWF-GPIb binding defects as reflected by type 2M VWD. This is further shown in Figs. 1D and 2. Thus, most outliers in these data comparisons reflect either assay variation (random error) events or important distinctions in functional defects. Again, additional subtle differences between test procedure results were evident in test data from the stepwise HMW deficiency samples (Fig. 4) and in desmopressin data (Fig. 3). Importantly, the VWF:Ac assay seemed equally sensitive to HMW deficiency as compared to both VWF:RCo and VWF:CB (Fig. 4). Test data for VWF: RCo, VWF:CB and VWF:Ac were also comparable during desmopressin evaluations, except for platelet GPIb binding defect type 2M VWD, where VWF:CB may be normal, both pre and post desmopressin, whereas the VWF:RCo and VWF:Ac assays are low and remain low post desmopressin (Fig. 3). In summary, VWF:RCo and VWF:Ac assessments will in most cases provide similar data (Figs. 1C, 3, 4), but occasional discrepancies may be evident; here, discrepancies are likely to be mostly due to assay variation (random error) events and more subtle effects of assay design/optimization. Similarly, VWF:RCo and VWF:Ac assessments will in most cases provide similar data to VWF:CB (Figs. 1C, 3, 4), but occasional discrepancies may be evident; here, discrepancies will be due to assay variation (random error) events as well as differential sensitivity to platelet VWF-GPIb binding defects (type 2M VWD; Figs. 1C, 2 and 3). Moreover, overall test data derived from the VWF: RCo and VWF:Ac assays were statistically ‘equivalent’ to each other, but not to VWF:CB (Table 3). There was also a high level of concordance between different VWF activity (VWF:RCo, VWF:Ac and VWF:CB) vs VWF:Ag ratios (Fig. 1D). Here, concordance overall was less than that for individual assays, due to multiplication of assay variation events because of inclusion of data from multiple assays to generate ratio values. Again, there was greater concordance between VWF:RCo/VWF:Ag and VWF:Ac/VWF:Ag ratios,



Fig. 4. Assay data using a set of four samples showing step-wise reduction in high molecular weight (HMW) VWF. VWF ‘1’ is a normal pool plasma sample, then each subsequent sample (‘2’, ‘3’, ‘4’) has a reduction of ~ 25%, 50% and 75% of the HMW VWF multimers. Figure A: Assay data, with FVIII or VWF values shown on y-axis in U/dL. Figure B: VWF activity / VWF:Ag assay ratio data. Abbreviations: FVIII:C = factor VIII coagulant; VWF:Ag E = ‘reference’ (ELISA); VWF:Ag L = automated LIA on CS-5100; VWF: Ac = using CS-5100; VWF:CB = ‘reference’ (ELISA); VWF:RCo BCS = automated agglutination on BCS as ‘reference method’; VWF:RCo CS-5100 = automated agglutination on CS-5100; CB/Ag (‘reference’) = VWF:CB/VWF:Ag using ‘reference’ methods; Ac/Ag L = VWF:Ac/VWF:Ag using CS-5100; RCo/Ag (‘reference’) = VWF:RCo/VWF:Ag using ‘reference’ methods; RCo/Ag (CS-5100) = VWF:RCo/VWF:Ag using CS-5100.

7

Moreover, data for the INNOVANCE VWF:Ac assay has only so far been independent reported in three very recently published studies [36–38], despite the assay being reported in many earlier meeting abstracts from the manufacturer and collaborators [39–43]. Lawrie et al. [36] assessed the VWF:Ac assay on a Siemens CS-2100i instrument in comparison with a VWF:RCo assay performed on a BCS, and using a smaller number of test samples (total n = 180). They reported a high correlation between the methods, with some largely minor discrepancies. The study did not perform any comparative evaluation using a VWF:CB assay. Graf et al. [37] also assessed the VWF:Ac assay on a Siemens CS-2100i instrument as well as a Stago STA-R Evolution in comparison with a VWF:RCo assay performed on an aggregometer; they used a larger number of overall test samples (total n = 362), and also reported a high correlation between the methods, with several noted discrepancies that they felt primarily due to higher variation in VWF:RCo and better sensitivity of their VWF:Ac to lower levels of VWF. Thus, in their study, the Ac/Ag ratio detected more discordance (Type 2 VWD-like) test patterns than did the RCo/Ag ratio; however, it is unclear if this actually reflected ‘better’ detection of type 2 VWD patterns by Ac/Ag, or alternatively reflected poorer detection of type 2 VWD patterns by their RCo/Ag ratio. Their findings may therefore potentially be explained by higher assay variation and poorer low VWF level sensitivity using the VWF:RCo. The study by Graf et al. [37] did not perform any comparative evaluation using a VWF:CB assay. Finally, a recent paper by Geisen et al. [38] evaluated the VWF:Ac against the VWF:RCo on the BCS instrument. These authors utilized the largest number of samples overall (n = 942), and also included a comparative evaluation using a VWF:CB. Their study was primarily focused on evaluation of acquired von Willebrand syndrome (AVWS); they reported that (i) the results obtained with the VWF:Ac assay were comparable to that of the VWF:RCo assay, (ii) the ratio Ac/Ag could detect AVWS more reliably than their standard RCo/Ag, and (iii) according to their data, VWF:Ac could replace VWF:RCo in routine screening. Although data for VWF:CB and CB/Ag was provided, the authors did not provide much comparative data between this data and that for VWF:RCo, RCo/ Ag, VWF:Ac and Ac/Ag, instead focusing primarily on comparisons between the latter. Again, whether their data suggested ‘superiority’ of VWF:Ac or alternatively ‘inferiority’ of VWF:RCo can also be debated. The current study therefore confirms and extends the previous reports, contemporarily important given the increasing trend for laboratories to incorporate the Siemens INNOVANCE VWF:Ac into test practice. An analysis of such uptake using data from one external quality assessment (EQA) program is shown in Fig. 5, and updates recently reported data [5]. Information provided by other EQA organizations similarly informs on the increasing usage of this assay in normal test practice. Summary and Conclusion

than with VWF:CB/VWF:Ag ratios, for reasons already mentioned above. Moreover, the overall ratio test data derived from the ‘reference’ RCo/Ag methods were statistically ‘equivalent’ to RCo/Ag (CS-5100/ ELISA) and Ac/Ag (CS-5100/ELISA). To our knowledge, an evaluation of the VWF:RCo and VWF:Ac assays performed on the CS-5100 have not before been reported. Indeed, a Medline search for ‘CS-5100’ only yielded three papers [33–35], with the most comprehensive [33] including data for a variety of assays (but excluding VWF activity assays), and the other two selectively evaluating APTT factor sensitivity [34] and changes in FXII levels during pregnancy [35]. In the single study evaluating VWF:Ag on the CS-5100, the data obtained using the Siemens LIA test reagent was essentially equivalent to the comparator assay, in that study the Stago LIA test reagent and a Stago instrument (STAR-Evolution). This comparison was therefore of a like-like method (both LIA based), albeit on two different instruments; also, the samples included in that study were not identified, and possibly did not include VWD patient samples.

VWF:Ag by the two methods compared in this study were essentially interchangeable, barring some subtle differences likely reflective of respective assay methodology (ELISA vs LIA) and design, and some less subtle differences likely to be reflective of assay interferences occasionally evident in LIA based assays. Should cases yield unexpectedly high VWF:Ag by LIA, discordant to activity assays, or to previous findings by an alternate method such as ELISA, then evaluation of rheumatoid factor [31] or heterophile antibodies [32], or repeat testing using an alternate method such as ELISA, may be required for clarification. VWF:RCo by the two methods compared in this study were also essentially interchangeable, barring some subtle differences most likely reflective of assay variability and respective assay design/optimization. The VWF:Ac assay was also essentially comparable to VWF:RCo in this study, and thus may reflect a feasible VWF:RCo alternative assay. We could not identify any major discrepancies in VWD identification or provisional VWD type between use of VWF:RCo vs VWF:Ac. However, although VWF:RCo and VWF:Ac assays were found to be essentially


8


Fig. 5. Data from the RCPA QAP showing evolution in VWF activity testing over recent years. Figure A: data from the start of the program (1998) until 2014, and showing percent of participants performing VWF:RCo, VWF:CB and/or other VWF ‘activity’ assays. There has been a steady decline in VWF:RCo performance by participants from a high of over 90% to current levels of just over 50%. VWF:CB over this time has remained reasonably stead at ~50% of laboratories. Other VWF activity assays were always performed at lower levels, but there has been a recent rise to nearly 40% in 2014. Figure B: data for the these ‘other’ VWF activity assays for the past four years showing the decline in the IL Werfen assay and the surge in Siemens INNOVANCE usage.

interchangeable, the same cannot be said for VWF:RCo and VWF:Ac vs VWF:CB. Thus, VWF:CB provides additional value to use of either VWF:RCo or VWF:Ac, in particular by helping to define type 2M VWD. Conflict of interest statement The authors have no conflicts of interest in relation to this work. Acknowledgements The authors would like to thank Jane McDonald, Ella Grezchnik and Monica Ahuja for technical assistance related to routine sample testing in our laboratory, and Roslyn Bonar for the RCPA QAP data reported in this paper. We also thank Geoff Kershaw for sharing some of his VWD patient samples with us. References [1] Favaloro EJ. Von Willebrand disease: local diagnosis and management of a globally distributed bleeding disorder. Semin Thromb Hemost 2011;37:440–55. [2] Flood VH. Perils, problems, and progress in laboratory diagnosis of von Willebrand disease. Semin Thromb Hemost 2014;40:41–8. [3] Yee A, Kretz CA. von Willebrand Factor: Form for Function. Semin Thromb Hemost 2014;40:17–27. [4] Favaloro EJ. Rethinking the diagnosis of von Willebrand disease. Thromb Res 2011; 127(Suppl. 2):s17–21. [5] Favaloro EJ, Bonar RA, Meiring M, Duncan E, Mohammed S, Sioufi J, et al. Evaluating errors in the laboratory identification of von Willebrand disease in the real world. Thromb Res 2014;134:393–403. [6] Favaloro EJ, Bodó I, Israels SJ, Brown SA. von Willebrand disease and platelet disorders. Haemophilia 2014;20(Suppl. 4):59–64. [7] Favaloro EJ. Diagnosing von Willebrand Disease: A Short History of Laboratory Milestones and Innovations, Plus Current Status, Challenges, and Solutions. Semin Thromb Hemost 2014;40:551–70. [8] Sadler JE, Budde U, Eikenboom JC, Favaloro EJ, Hill FG, Holmberg L, et al. Working Party on von Willebrand Disease Classification. Update on the pathophysiology and classification of von Willebrand disease: a report of the Subcommittee on von Willebrand Factor. J Thromb Haemost 2006;4:2103–14. [9] Nichols WL, Hultin MB, James AH, Manco-Johnson MJ, Montgomery RR, Ortel TL, et al. Willebrand disease (VWD): evidence-based diagnosis and management guidelines, the National Heart, Lung, and Blood Institute (NHLBI) Expert Panel report (USA). Haemophilia 2008;14:171–232. [10] Favaloro EJ, Franchini M, Lippi G. Biological therapies for von Willebrand Disease. Expert Opin Biol Ther 2012;12:551–64. [11] Favaloro EJ, Plebani M, Lippi G. Regulation in hemostasis and thrombosis: Part I-in vitro diagnostics. Semin Thromb Hemost 2013;39:235–49. [12] Adcock DM, Bethel M, Valcour A. Diagnosing von Willebrand disease: a large reference laboratory's perspective. Semin Thromb Hemost 2006;32:472–9.

[13] Flood VH, Gill JC, Friedman KD, Christopherson PA, Jacobi PM, Hoffmann RG, et al. Collagen binding provides a sensitive screen for variant von Willebrand disease. Clin Chem 2013;59:684–91. [14] Budde U, Pieconka A, Will K, Schneppenheim R. Laboratory testing for von Willebrand disease: contribution of multimer analysis to diagnosis and classification. Semin Thromb Hemost 2006;32:514–21. [15] Favaloro EJ, Grispo L, Exner T, Koutts J. Development of a simple collagen binding assay aids in the diagnosis of, and permits sensitive discrimination between, Type I and Type II von Willebrand's disease. Blood Coagul Fibrinolysis 1991;2:285–91. [16] Favaloro EJ. Evaluation of commercial von Willebrand factor collagen binding assays to assist the discrimination of types 1 and 2 von Willebrand disease. Thromb Haemost 2010;104:1009–21. [17] Favaloro EJ, Mohammed S, McDonald J. Validation of improved performance characteristics for the automated von Willebrand factor ristocetin cofactor activity assay. J Thromb Haemost 2010;8:2842–4. [18] Favaloro EJ, Grispo L, Exner T, Koutts J. Development of a simple collagen binding assay aids in the diagnosis of, and permits sensitive discrimination between, Type I and Type II von Willebrand's disease. Blood Coagul Fibrinolysis 1991;2:285–91. [19] Favaloro EJ, Grispo L, Dinale A, Berndt M, Koutts J. von Willebrand's Disease: Laboratory investigation using an improved functional assay for von Willebrand Factor. Pathology 1993;25:152–8. [20] Favaloro EJ, Facey D, Grispo L. Laboratory assessment of von Willebrand Factor: Use of different assays can influence the diagnosis of von Willebrand's Disease, depending on differing sensitivity to sample preparation and differential recognition of high molecular weight VWF forms. Am J Clin Pathol 1995;104:264–71. [21] Favaloro EJ, Koutts J. Laboratory assays for von Willebrand Factor: Relative contribution to the diagnosis of von Willebrand's disease. Pathology 1997;29:385–91. [22] Favaloro EJ. Collagen binding assay for von Willebrand Factor (VWF:CBA): Detection of von Willebrands Disease (VWD), and discrimination of VWD subtypes, depends on collagen source. Thromb Haemost 2000;83:127–35. [23] Favaloro EJ, Henniker A, Facey D, Hertzberg M. Discrimination of von Willebrands Disease (VWD) subtypes: Direct comparison of von Willebrand Factor:Collagen binding activity/assay (VWF:CBA) with Monoclonal Antibody (MAB) based ELISA VWF-detection systems. Thromb Haemost 2000;84:541–7. [24] Favaloro EJ. Detection of von Willebrand Disorder (VWD) and identification of qualitative von Willebrand Factor (VWF) defects: Direct comparison of commercial ELISA-based ‘VWF:Activity’ options. Am J Clin Pathol 2000;114:608–18. [25] Favaloro EJ, Krigstein M, Koutts J, Brighton T, Lindeman R. Genetic testing for the diagnosis of von Willebrand Disease: benefits and limitations. J Coagul Disord 2010;2:37–47. [26] Favaloro EJ, Dean M, Grispo L, Exner T, Koutts J. von Willebrand's Disease: Use of Collagen Binding Assay provides potential improvement to laboratory monitoring of desmopressin (DDAVP) therapy. Am J Hematol 1994;45:205–11. [27] Favaloro EJ, Thom J, Patterson D, Just S, Dixon T, Koutts J, et al. Desmopressin therapy to assist the functional identification and characterisation of von Willebrand disease: differential utility from combining two (VWF:CB and VWF:RCo) von Willebrand factor activity assays? Thromb Res 2009;123:862–8. [28] Favaloro EJ. Diagnosis of type 1 versus 2A and 2M von Willebrand disease. Haemophilia 2012;18:e9-11. [29] Favaloro EJ, Forsyth C, Koutts J. Distinguishing types 1 and 2 M von Willebrand disease. Int J Lab Hematol 2012;34:102–5. [30] Favaloro EJ, Bonar R, Chapman K, Meiring M, Adcock DF. Differential sensitivity of von Willebrand factor 'activity' assays to large and small VWF molecular weight forms: a cross-laboratory study comparing ristocetin cofactor, collagen binding and monoclonal antibody based assays. J Thromb Haemost 2012;10:1043–54.


E.J. Favaloro, S. Mohammed / Thrombosis Research xxx (2014) xxx–xxx [31] Favaloro EJ, Aboud M, Arthur C. Possibility of potential VWD misdiagnosis or misclassification using LIA technology and due to presence of rheumatoid factor. Am J Hematol 2001;66:53–6. [32] Lippi G, Ippolito L, Tondelli MT, Favaloro EJ. Interference from heterophilic antibodies in D-dimer assessment. A case report. Blood Coagul Fibrinolysis 2014;25:277–9. [33] Ratzinger F, Schmetterer KG, Haslacher H, Perkmann T, Belik S, Quehenberger P. Evaluation of the automated coagulation analyzer CS-5100 and its utility in high throughput laboratories. Clin Chem Lab Med 2014;52:1193–202. [34] Lawrie AS, Kitchen S, Efthymiou M, Mackie IJ, Machin SJ. Determination of APTT factor sensitivity–the misguiding guideline. Int J Lab Hematol 2013;35: 652–7. [35] Sharief LT, Lawrie AS, Mackie IJ, Smith C, Peyvandi F, Kadir RA. Changes in factor XIII level during pregnancy. Haemophilia 2014;20:e144–8. [36] Lawrie AS, Stufano F, Canciani MT, Mackie IJ, Machin SJ, Peyvandi F. A comparative evaluation of a new automated assay for von Willebrand factor activity. Haemophilia 2013;19:338–42. [37] Graf L, Moffat KA, Carlino SA, Chan AK, Iorio A, Giulivi A, et al. Evaluation of an automated method for measuring von Willebrand factor activity in clinical samples without ristocetin. Int J Lab Hematol 2014;36:341–51.

9

[38] Geisen U, Zieger B, Nakamura L, Weis A, Heinz J, Michiels JJ, et al. Comparison of Von Willebrand factor (VWF) activity VWF:Ac with VWF ristocetin cofactor activity VWF:RCo. Thromb Res 2014;134:246–50. [39] Schneppenheim R, Obser T, Budde U, Patzke J. Development of a new functional assay for von Willebrand factor binding to platelet GPIba that does not require ristocetin. Hamostaseologie 2010;30:A28 [Abstract FC4-05 of the 54th Annual Meeting of the GTH]. [40] Patzke J, Muth H, Wilkens M, Althaus H, Budde U, Schneppenheim R. Development of a gain-of-function GPIBA binding assay as an alternative to the ristocetin cofactor assay (7th Bari International Conference; Abstract PO004). Blood Transfus 2011;9(S 3):s7. [41] Huber AR, Budde U, Patzke J, Muth H. Comparative evaluation of a new von Willebrand factor activity assay. J Thromb Haemost 2011;9:919 [Abstract P-TH479 of the XXII ISTH Congress]. [42] Budde U, Schneppenheim R, Mendez A, Patzke J, Muth H. Clinical evaluation of a new von Willebrand Factor activity assay. J Thromb Haemost 2011;9:913 [Abstract P-TH-461 of the XXII ISTH Congress]. [43] Patzke J, Muth H, Wilkens M, Schneppenheim R. Performance data of the new automated von Willebrand factor activity assay based on GPIbα binding. Hamostaseologie 2011 [Abstract No. P-12-5 of the 54th Annual Meeting of the GTH].


Comparison of automated von Willebrand factor activity assays.

Acquired von Willebrand syndrome: von Willebrand factor propeptide to von Willebrand factor antigen ratio predicts remission status.

Predictors of von Willebrand disease diagnosis in individuals with borderline von Willebrand factor plasma levels.

Von Willebrand Disease Lab Diagnosis.

Current management of von Willebrand disease and von Willebrand syndrome.

von Willebrand factor protein. Galactose a cryptic determinant of von Willebrand factor activity.

Laboratory diagnosis of von Willebrand disease.

[Von Willebrand disease. Molecular biology and diagnosis].

von Willebrand factor propeptide and the phenotypic classification of von Willebrand disease.

Factor VIII (Willebrand) antigen and ristocetin-Willebrand factor in pigs with von Willebrand's disease.

Inhibition of Willebrand factor in von Willebrand disease.

Diagnostic Value of Measuring Platelet Von Willebrand Factor in Von Willebrand Disease.

Von Willebrand factor and aging.

First report of inhibitory von Willebrand factor alloantibodies in type 2B von Willebrand disease.

Von Willebrand disease.

Willebrand factor in von Willebrand's disease.

Subcellular platelet factor VIII antigen and von Willebrand factor.

Performance of two new automated assays for measuring von Willebrand activity: HemosIL AcuStar and Innovance.

Raised concentrations of von Willebrand factor antigen in systemic sclerosis.

Further evidence for recessive inheritance of von Willebrand disease with abnormal binding of von Willebrand factor to factor VIII.

FVIII concentrates: improved performance with a new modified automated method.

hemophilia A.

von Willebrand factor antigen in giant cell arteritis.

Cell biology of von Willebrand factor.