Evaluating errors in the laboratory identification of von Willebrand disease in the real world

Introduction

von Willebrand disease (VWD), reportedly the most common bleeding disorder, arises from deficiency and/or defects of von Willebrand factor (VWF). Assessment requires a wide range of tests, including VWF activity and antigen. Appropriate diagnosis including differential identification of qualitative vs quantitative defects has important management implications, but remains problematic for many laboratories and clinicians.

Methods

Data using a large set (n = 29) of varied plasma samples comprising both ‘quantitative’ VWF deficiency (‘Type 1 and 3’ VWD) vs ‘qualitative’ defects (‘Type 2 VWD’) tested in a cross-laboratory setting has been evaluated to assess the ability of real world laboratories to differentially identify these sample types.

Results

Different VWF assays and activity/antigen ratios show different utility in VWD and type identification. VWD identification errors were often linked to high inter-laboratory test variation and result misinterpretation (i.e., laboratories failed to correctly interpret their own test panel findings). Thus, moderate quantitative VWF deficient samples were misinterpreted as qualitative defects on 30/334 occasions (9% error rate); 17% of these errors were due to laboratories misinterpreting their own data, which was instead consistent with quantitative deficiencies. Conversely, whilst qualitative VWF defects were misinterpreted as quantitative deficiencies at a similar error rate (~ 9%), this was more often due to laboratories misinterpreting their data (~ 50% of errors). For test-associated errors, ristocetin cofactor was associated with the highest variability and error rate, which was at least twice that using collagen binding.

Conclusion

These findings in part explain the high rate of errors associated with VWD diagnosis.     

Flow-based measurements of von Willebrand factor (VWF) function: Binding to collagen and platelet adhesion under physiological shear rate

Introduction

VWF circulates in plasma as a series of heterogeneous multimers, mediating platelet tethering, translocation and finally adhesion to areas of injured endothelium under physiological high arterial blood flow. VWF-platelet binding requires conformational changes in VWF, which are induced by immobilization and shear. Because of unavailability of a simple flow-based measurement system, VWF activity assays are generally performed under static conditions. We describe an easily reproducible in vitro flow-chamber model using commercially available flow devices to examine VWF-collagen binding and VWF-mediated platelet adhesion under physiological flow conditions.

Methods

The collagen surface of the flow-chamber was analyzed by atomic force microscopy. Collagen-bound VWF was characterized by multimer analysis and multi labelling immunofluorescence detection of exposed GPIb binding domains. Platelet adhesion was captured by time-lapse microscopy.

Results

The described flow-chamber system facilitates multimer analysis of collagen-bound VWF, whereas all VWF multimers bound to collagen under physiological low to high shear rates. Multi labelling immunofluorescence detection exhibited exposed GPIb binding domains co-localized with VWF molecules. VWF-dependent platelet adhesion using time-lapse microscopy showed values comparable to experiments done with whole blood, and platelet adhesion was dependent on the VWF concentration.

Conclusions

The established flow-chamber model represents an easy-to-set-up and customized tool for the characterization of VWF-binding to collagen as well as the determination of VWF-dependent platelet adhesion under defined flow conditions in real-time.     

Plasma von Willebrand factor multimer quantitative analysis by in-gel immunostaining and infrared fluorescent imaging

Introduction

Electrophoretic analysis of plasma von Willebrand factor (VWF) multimer distribution and infrastructure is essential for subtyping von Willebrand disease. To improve the sensitivity, precision and efficiency of this assay, we developed and validated a new in-gel infrared fluorescent VWF multimer imaging method to visualize and quantify VWF multimers directly in the agarose gel, thus eliminating electroblotting or autoradiographic steps.

Materials/Methods

VWF multimer analyses of plasma samples from 34 patients with known von Willebrand disease or acquired von Willebrand syndrome, 9 patients with acquired VWF abnormalities, 26 normal volunteer donors and 49 patient samples referred for von Willebrand factor multimer analysis were performed by both traditional autoradiographic and the new infra-red imaging methods and compared. VWF multimer image data were electronically acquired, archived and analyzed.

Results

The in-gel infrared method has a sensitivity of detecting VWF antigen as low as approximately 1.6 IU/dL, a reliable fluorescent intensity with intra- and inter-day variability (CV) of 5% and 6% respectively, and provides superior imaging resolution and shortened test turnaround time. Using intermediate resolution agarose gel electrophoresis, the infra-red method sensitively detects subtle loss of highest molecular weight von Willebrand factor multimers in plasmas with acquired VWF abnormalities and in commercial normal reference plasmas, and provides evidence of increased proteolysis of ultralarge multimers in some type 2 VWD plasmas.

Conclusions

The in-gel infrared fluorescent VWF multimer imaging method provides a sensitive, reliable, efficient and robust system to improve laboratory testing for von Willebrand disease classification.     

Comparison of Von Willebrand factor (VWF) activity VWF:Ac with VWF ristocetin cofactor activity VWF:RCo

Introduction

Ristocetin cofactor activity of Von Willebrand factor (VWF:RCo) and the ratio VWF:RCo to its antigen VWF:Ag are used as routine screening to estimate VWF function and to detect types of Von Willebrand disease (VWD) caused by loss of high molecular weight multimers. However, the VWF:RCo test is prone to analytic imprecisions due to various reasons. We compared an assay for VWF activity (VWF:Ac) with VWF:RCo putting emphasis on the ratios to VWF:Ag.

Materials and Methods

We evaluated 942 samples from 432 patients and evaluated three groups in detail: normal patients (normal multimers, VWF:Ag and VWF:RCo > 0.5 U/ml, VWD type 1 excluded; n = 258), VWD type 1 (n = 76) and acquired Von Willebrand syndrome (AVWS, n = 326). In addition, 30 healthy subjects were analysed.

Results

VWF:Ac and VWF:RCo correlated well (Pearson´s r = 0.96, p < 0.01), so did their ratios to VWF:Ag (Pearson´s r = 0.82, p < 0.01). We calculated the normal range of VWF:Ac/VWF:Ag for healthy subjects as 0.8-1.16. In comparison, the reference range (mean ± 2std) derived from normal patient samples was 0.73-1.14. The corresponding ranges for VWF:RCo/VWF:Ag came to 0.74-1.23 (healthy) and 0.62-1.25 (normal patients). The ratios showed similar results regarding VWD type 1. The sensitivity for AVWS was higher with VWF:Ac/VWF:Ag than with VWF:RCo/VWF:Ag (97.5% versus 84.7%).

Conclusions

The data suggest that the results obtained with the VWF:Ac assay are comparable to that of the VWF:RCo assay. An AVWS was more reliably detected by VWF:Ac/VWF:Ag. We assume that the VWF:Ac assay could replace VWF:RCo for routine screening for AVWS, especially when prompt evaluation is required.     

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?

We performed a retrospective audit of desmopressin (DDAVP) usage to assist in the functional characterisation of von Willebrand disease (VWD). Data was evaluated for 208 patients, comprising those with VWD (Type 1 [n = 160], Type 2A [n = 19], Type 2M [n = 10]), plus 19 individuals with haemophilia or carriers of haemophilia. Laboratory testing comprised pre- and post-DDAVP evaluation of factor VIII (FVIII:C), von Willebrand factor (VWF) antigen (VWF:Ag), VWF ristocetin cofactor (VWF:RCo) activity, VWF collagen binding (VWF:CB) activity, and in one laboratory an alternate VWF activity assay. In brief, combined usage of VWF:RCo and VWF:CB appears to provide improved functional characterisation and/or ‘classification’ of VWD types, in particular better differentiation of Type 2A and 2M VWD, and clearer validation of a Type 1 VWD diagnosis. Thus, (i) Type 1 VWD displayed generally good absolute and relative rises in all test parameters, although relative rises were greatest for FVIII:C and VWF:CB, and CB/Ag ratio increases overshadowed those for RCo/Ag; (ii) Type 2A VWD patients showed good absolute and relative rises in both FVIII:C and VWF:Ag, but poor absolute rises in both VWF:CB and VWF:RCo; although small rises in both CB/Ag and RCo/Ag were also observed, both ratios tended to remain below 0.7; (iii) finally, Type 2 M VWD patients generally showed good absolute and relative rises in FVIII:C, VWF:Ag and VWF:CB, but a poor absolute and relative rise in VWF:RCo; thus, there were good rises in CB/Ag ratios but little change in RCo/Ag, which tended to remain below 0.7. Future multi-centre prospective investigations are warranted to validate these findings and to investigate their therapeutic implications.     

A new method measuring the interaction between von Willebrand factor and coagulation factor VIII

Introduction

There is a need for more reliable methods measuring the binding of coagulation factor VIII (FVIII) to von Willebrand factor (VWF) in plasma samples, for use in the clinical routine. We have developed such a method measuring FVIII binding in plasma, utilizing an ELISA system.

Materials and Methods

Microtiter plates were coated with a monoclonal antibody (ESH-8), reacting with the C2 domain in FVIII. Thereafter the wells were treated with recombinant FVIII (Kogenate Bayer®). After washing, diluted plasma samples were added and incubated for 1 h. Then HRP-conjugated antibodies against VWF were added and used for quantification of bound VWF.

Results

A strong signal to VWF concentration response was obtained. Plasma from patients with different types of von Willebrand disease gave frequently diminished responses. However, after correction for the VWF antigen levels, by calculation of FVIII binding/VWF antigen ratio, only the patients with known von Willebrand disease type 2 N (n = 4) had clearly abnormal results. The FVIII binding in 40 healthy individuals was determined as 1.08 ± 0.48 U/mL (SD). After correction for the VWF antigen levels the result was 0.94 ± 0.15. Thus, the SD declined substantially by this correction. The within-series CV and between-series CV were determined as 6.8 and 11.3%, respectively.

Conclusions

We have established a simple and reliable method to detect decreased binding of FVIII to von Willebrand factor in plasma samples. The method can conveniently be used to study large populations, as well as finding minor binding defects in patients.     

A two-step approach (Enzyme-linked immunosorbent assay and confirmation assay) to detect antibodies against von Willebrand factor in patients with Acquired von Willebrand Syndrome

Background

Acquired von Willebrand Syndrome is a rare bleeding disorder, which arises in individuals with no personal or family history of bleeding, associated with lymphoproliferative and myeloproliferative disorders or other diseases.

Aim

To develop a two-step approach assay to detect autoantibodies against VWF and to verify their prevalence in AVWS.

Methods

AVWS definition: negative personal or family history of bleeding diathesis, VWF below normal range and recent history of bleeding manifestations. Twenty-three consecutive patients affected by AVWS were enrolled. An ELISA assay (first step) using recombinant VWF protein immobilized on plates and sheep/goat polyclonal anti-human IgG or IgM labelled with peroxidase was developed. A group of 40 healthy subjects was tested to calculate the floating cut point value. A confirmation assay (with addition of purified VWF vs buffer) was performed (second step).

Results

Twenty–one patients (93%) had an associated disease, two patients had idiopathic AVWS. Anti-VWF autoantibodies were detected in 9 patients (39%). Of these, eight (89%) had VWF:RCo levels < 10%, but none of them resulted positive using Bethesda assay (neutralizing antibodies). The confirmation test confirmed the positive results obtained with ELISA and resulted negative in those patients with negative results and in the controls.

Conclusion

With the present two-step approach assay nine out of 23 (39%) patients affected with AVWS resulted positive for anti-VWF autoantibodies. This ELISA assay might be used as an additional confirmation tool in the diagnostic procedure in patients affected by AVWS or in the follow-up of congenital and acquired patients exposed to replacement therapy.     

Sources of variation in factor VIII, von Willebrand factor and fibrinogen measurements: Implications for detecting deficiencies and increased plasma levels

Objectives

Differences in pre-analytical and assay conditions, inappropriate reference ranges, or inflammation may have the potential to impair clinical decisions based on measurements of factor VIII (FVIII), von Willebrand factor (VWF) and fibrinogen (Fg). This study examined the impact on FVIII, VWF and Fg in plasma of freezing and thawing, different citrate anticoagulant concentrations, and inflammation, as determined by high-sensitivity C-reactive protein (hsCRP).

Materials and Methods

FVIII was determined prior to freezing and after thawing using a one-stage clotting assay (FVIII:C), an amidolytic assay (FVIII:AM) and an enzyme immunoassay (FVIII:Ag). Samples were anticoagulated with 106 or 129 mmol/L of citrate. FVIII, VWF and Fg were quantified in 300 individuals to establish reference ranges and to investigate associations with hsCRP.

Results

Freezing and thawing reduced FVIII:C and FVIII:AM markedly. FVIII coagulant activities were not significantly different between samples anticoagulated with 106 or 129 mmol/L of citrate, respectively. FVIII, VWF and Fg were significantly associated with hsCRP. FVIII:C was greater than FVIII:AM and FVIII:Ag in all experiments, indicating that the presence of activated FVIII may lead to overestimation of FVIII:C.

Conclusions

Standardized freezing and thawing of plasma samples appears to be indispensable if reliable FVIII results are to be obtained. Because inflammation can potentially mask deficiency states or mimic an increased risk of thrombosis, FVIII, VWF and Fg determinations should be supplemented by measurements of hsCRP.     

Coating conditions matter to collagen matrix formation regarding von Willebrand factor and platelet binding

Introduction

Von Willebrand factor (VWF) and platelet binding needs a uniform collagen matrix therefore we aimed to find an optimal condition for the preparation of human type-I and type-III collagen matrices.

Method

The effects of pH, salt and ligand concentration and binding time were tested when collagen matrices were prepared by adsorption. Surface-bound collagen and collagen-bound VWF measured by specific antibodies. Platelet adhesion was tested under flow conditions at a shear rate of 1800 s^− 1 for 2 min. Matrices and platelets were visualized by atomic force and scanning electron microscope.

Results

The extent of human collagens type-I and III binding to the surface was 10 and 4 times greater and binding was maximal under 8-16 hours, when coated from physiological buffer solution versus acid solution. Collagen fibrils were more developed and platelet adhesion was higher, with more organized and denser aggregates. VWF binding was parallel to the surface bound collagen in both collagen types.

Conclusion

Collagen coating of surfaces for VWF binding and platelet adhesion studies is very variable from acid solution. Our experiments provide evidences that neutralizing the acid and adding NaCl in physiological concentration, thereby facilitating formation of collagen fibril molecules in solution, results in efficient coating of human type-I and type III collagens, which then bind normal VWF equally well.     

Assessment of a cohort of primarily pediatric patients with a presumptive diagnosis of type 1 von Willebrand disease with a novel high shear rate, non-citrated blood flow device

Background

A precise approach to the diagnosis of von Willebrand disease (vWD) remains elusive. One important reason is that vWD is a blood flow‐related disorder: a vW Factor‐platelet GPIb binding defect exists in this condition under the high shear‐rate (> 1000 sec‐1 in whole blood; > 3000 sec‐1 in PRP) conditions of physiologic blood flow which exist in the arterioles of mucous membranes, from which most bleeding in vWD occurs.

Methods

We therefore studied 28 patients (mean 18.9 yrs) with vWD, diagnosed according to the 2007 NHLBI clinical guidelines, and 26 healthy controls (mean 17.5 yrs). Blood was collected into a plastic tube containing 4 U/ml FC dalteparin, 1.75 μg/ml of the Tab (anti‐CD41) monoclonal antibody directed against platelet GPIIb, and 1.0 μg/ml of an ALEXA 555‐conjugated rabbit anti‐mouse second antibody. Within 30-90 min, the blood was then withdrawn at 667 and 1330 sec^− 1 through a special flow chamber allowing for real-time epifluorescence digital videomicroscopy of platelets interacting with a microfibrillar collagen substrate. With MetaMorph software (Universal Imaging) we quantified the percent area (PA) covered by and total volume (TV) of adherent platelet aggregates within a 435 μm × 580 μm field of view.

Results

At 667 sec^− 1 after 1 min PA and TV were similar for patients and controls, but at 1330 sec^− 1 PA was 9.32 ± 4.21 (mean ± SD) for patients, a value lower (p < 0.001) than the 12.8 ± 3.39 for controls. TV was (1.43 ± 0.91)x10⁴ for patients, a value also lower (p < 0.001) than the (2.22 ± 0.77)x10⁴ for controls. PA or TV was below the 2.5th percentile for controls in 10 patients (36%) and both PA and TV were below the 2.5th percentile in eight.

Conclusions

The novel flow device found that PA and TV were significantly reduced under high shear stress in vWD patients compared to normal controls. However, there was some overlap between the vWD and the control group, suggesting that some vWD patients had normal platelet adhesion/aggregation under the conditions studied. Further study with a higher shear rate appears indicated.     

Increased basal platelet activity, plasma adiponectin levels, and diabetes mellitus are associated with poor platelet responsiveness to in vitro effect of aspirin

INTRODUCTION: Aspirin is one of the most effective antiplatelet agents and is now commonly used to prevent vascular events. In some patients, however, recurrent vascular events have been demonstrated despite aspirin therapy. Our objective was to characterize individuals showing poor response to in vitro effect of aspirin, using PFA-100. METHODS: One hundred sixty-eight healthy male subjects were analyzed. We assessed platelet function tests, including PFA-100, whole blood aggregation, and optical platelet aggregation. Also measured were hemostatic and other parameters including von Willebrand factor (VWF:Ag), VWF ristocetin cofactor activity (VWF:RCo), soluble vascular adhesion molecule-1 (sVCAM-1), high sensitive C-reactive protein (hs-CRP), and adiponectin. Poor responders were defined as having a collagen/epinephrine-induced closure time (CEPI-CT) under 250 s with PFA-100 when incubated with 10 microM aspirin, whereas good responders were defined as having a CEPI-CT of more than 250 s. RESULTS AND CONCLUSIONS: PFA-100 tests revealed that 40 subjects (24%) were poor responders (PR) and 128 (76%) were good responders (GR). Poor responsiveness was significantly associated with (1) higher basal platelet activities in PFA-100, as well as in whole blood aggregation and aggregometer;(2) increased level of adiponectin (8.8+/-4.1 micro g/mL [PR] vs 7.3+/-2.9 micro g/mL [GR], p=0.010);and (3) the presence of diabetes mellitus (17.5% [PR] vs 4.7% [GR], p=0.009). Importantly, whereas 24% of the subjects showed insufficient inhibition in PFA-100 when incubated with 10 microM aspirin, almost all subjects showed maximum inhibition with 30 microM aspirin. These observations suggest that higher doses of aspirin might overcome aspirin resistance.