|Values are valid only on day of printing.|
Click CC to turn on closed captioning.
Published: June 2011Print Record of Viewing
Von Willebrand disease (VWD) is the most common of the inherited bleeding disorders. Plasma von Willebrand factor (VWF) activity is used in conjunction with VWF antigen and factor VIII coagulant activity to diagnosis VWD, differentiate VWD subtypes, and differentiate VWD from hemophilia A.
This is the third in a four-part series on von Willebrand disease, and is also included in the series on test utilization.
Presenter: Dong Chen, MD, PhD
Welcome to Mayo Medical Laboratories' Hot Topics. These presentations provide short discussions of current topics and may be helpful to you in your practice.
Our speaker for this program is Dong Chen, MD, PhD, Co-director of the Special Coagulation Laboratory at Mayo Clinic in Rochester, Minnesota. Dr. Chen will describe the newly implemented von Willebrand factor activity assay and its use in the laboratory diagnosis of von Willebrand disease.
Thank you for tuning in for this series regarding von Willebrand Disease laboratory testing. Today, I will discuss our newly implemented automated latex particle-enhanced immunoturbidimetric von Willebrand factor activity assay, or VWF latex activity in short.
In this lecture, I will show you our validation data and explain the rationale behind our new VWD profile testing algorithm.
VWF activity assay is required to diagnose and correctly classify von Willebrand disease or acquired von Willebrand syndrome. VWF ristocetin cofactor activity assay is considered a reference method to assess VWF-platelet-binding activity. In 1971, Dr. Howard and Firkin first observed ristocetin-induced platelet aggregation, which was apparently defective in von Willebrand disease patients. 2 years later, Dr. Weiss developed the first VWF ristocetin cofactor activity assay. We validated this assay in 1975 and have performed this assay as part of our VWD panel for the past 35 years. In the last year, we performed about 10 thousand VWF ristocetin cofactor activity assays with the majority of samples from Mayo Medical Laboratories.
Despite its vibrant history and the status of being considered as the reference method, VWF ristocetin cofactor activity assay is problematic. It is imprecise with an average coefficient variation of 20% to 30%, possibly due to diversity of the reagent donor platelets and complexity of platelet aggregation kinetics. It has suboptimal lower and higher detection limit. Currently, the lower detection limit of VWF ristocetin cofactor activity is 12.5 IU/dL. In addition, VWF ristocetin cofactor activity is very imprecise when VWF antigen is lower than 30 IU/dL or higher than 200 IU/dL. And finally, VWF ristocetin cofactor activity over antigen ratio at a cut off of 0.7 is generally insensitive for acquired VWF abnormality when there are subtle losses of the highest molecular weight multimers.
These problems are invincible in multiple US and international external quality assurance programs. As summarized in this table, this VWF ristocetin cofactor activity assay has poor precision among participating laboratories in US, UK, Australia and non-Western countries.
Consequently, more than half of the misdiagnosis and misclassification of VWD are attributed to erroneous VWF ristocetin cofactor activity results as demonstrated by Dr. Favaloro in 2006.
Now let us look at acquired VWF abnormalities. Here are 2 patients; patient A was status post left ventricular assist device implantation, patient B had polycythemia vera with marked thrombocytosis. Both patients had abnormal VWF multimer profile with subtle losses of the highest molecular weight multimers. Both patients had normal VWF antigen, VWF ristocetin cofactor activity and normal VWF activity over antigen ratios as listed in the table. Therefore, without a multimer analysis, the acquired VWF abnormality could easily be missed.
We studied 23 patients who received left ventricular assist device implantation and 1 patient with polycythemia vera and marked thrombocytosis. All patients had subtle losses of the highest molecular weight VWF multimers. By using the receiver operating characteristic curve analysis, we studied the sensitivity and specificity of VWF ristocetin cofactor activity over antigen ratio. The area under the curve (AUC) is 0.75. When a 0.7 is set as the cut-off, its sensitivity of detecting such acquired VWF abnormality is less than 50%.
Because of all these shortcomings of VWF ristocetin cofactor activity assay, last year we decided to explore and validate a new automated VWF activity assay by latex- immunoturbidimetric method.
This new assay employs a previously characterized monoclonal antibody that recognizes platelet glycoprotein 1b-binding epitope on the VWF A1 domain. Here is a 3 dimensional presentation of the botrocetin, VWF A1 domain and GPIb alpha. The arrow indicates the functional epitope that is specifically recognized by this antibody. Any conformation change at or around this epitope, particularly in type 2 von Willebrand disease, could affect antibody binding to this site.
The key laboratory characteristics of the new VWF activity assay, are as follows: The lower detection limit is 3 U/dL. Linearity is between 3 to 350 U/dL. The inter- and intra-precisions are all around 5%. Citrated plasma samples are stable for 8 hours at room temperature. This assay is fully automated on our TOP coagulation analyzers which could significantly improve test turn-around time.
We included a total of 492 samples for this validation study. 168 samples were from Mayo Clinic patients and the rest were consultation samples from Mayo Medical Laboratories. We compared VWF latex activity method with the reference VWF ristocetin cofactor activity assay by light transmission platelet aggregation. VWF latex activity assay has very good linear correlation with VWF ristocetin cofactor activity assay with R square at 0.93.
And the new method showed no significant bias from VWF ristocetin cofactor activity assay by Bland-Altman analysis.
We then categorized samples into normal, type 1, type 2 VWD, severe type 1 vs 2M VWD and acquired VWF abnormality (AVWA in abbreviation). This categorization is for method comparison. Most of them are self-explanatory. The severe type 1 versus 2M VWD categorize samples whose VWF ristocetin cofactor activities were below lower detection limit and their VWF multimer analysis showed no significant loss of high molecular weight multimers. Therefore, they can not be confidently distinguished between a severe type 1 or 2M von Willebrand disease. Acquired VWF abnormality (AVWA) categorizes samples with normal VWF antigen and ristocetin cofactor activity but have very subtle losses of highest molecular weight multimers.
Let us look at the distribution of the VWF latex activity over antigen ratio in different sample categories. The normal cut off is 0.7 as indicated by the red line. A vast majority of the normal samples had normal VWF latex over antigen ratios with rare outliers. All 57 type 1 von Willebrand disease samples had normal VWF latex activity ratios.
Twenty-eight of the thirty-one type 2A or 2B samples had abnormal ratios with only 3 outliers. All 4 2M VWD samples had abnormal ratios. We can not really compare the 2 methods in the group of severe types 1 vs 2M when VWF ristocetin cofactor activity is below lower detection limit. Finally, and interestingly, of the 24 samples with subtle losses of the highest molecular weight multimers, VWF latex activity ratios were significantly lower than those of the VWF ristocetin cofactor activity, and subsequently more sensitive for detecting such an acquired VWF abnormality. Occasional discrepant VWF ristocetin cofactor activity over antigen ratios were also observed.
Now let us examine the outliers in detail. 2 of the 3 samples of the normal group and type 2A or 2B group were lipemic. As we know that lipemia could potentially interfere with the optical property of a turbidimetric assay. However, the discrepancy of the remaining 1 sample from the normal and the 1 sample from the type 2A or 2B group can not be clinically verified since they were both Mayo Medical Laboratories samples. This underscores a potential limitation of this assay that it is still uncertain if the new assay will detect all type 2 von Willebrand disease variants, or conversely, if the new assay will pick up new type 2 von Willebrand variants that would have been missed by the VWF ristocetin cofactor activity assay. As for the group of severe types 1 or 2M, we classify the 8 samples by using a highly sensitive flow cytometric method.
Let us look at this group in detail. We used VWF ristocetin cofactor activity by flow cytometry and antigen ratios to classify these samples. At a cut off of 0.5, 6 samples were classified as 2M and 2 samples as severe type 1. Interestingly, 7 of the 8 samples showed concordant VWF latex activity over antigen ratios, with exception of one lipemic sample. 4 samples in the 2M VWD group were from 2 families whose genotyping studies showed a cytosine to thymine mutation at nucleotide 4120. This variant is now classified as 2M and their VWF multimer usually show no loss of high molecular weight multimer, but instead frequently demonstrate ultra-large multimers and smeary multimer pattern.
Lastly, at a cut of 0.7, VWF latex activity over antigen ratio only miss 9 of the 24 samples with subtle loss of highest molecular weight multimer in comparison to 19 samples by VWF ristocetin cofactor activity assay.
When we look again at our previously described 2 samples, in contrast to VWF ristocetin cofactor activity ratios, highlighted in blue, both VWF latex ratios, highlighted in red, were abnormal.
We also performed the ROC curve analysis of VWF latex activity over antigen ratio to determine its optimal cutoff. It has area under the curve AUC of 0.97 which is significantly better that of the VWF ristocetin cofactor activity ratios. When sensitivity and specificity are equally weighted, the optimal cutoff of VWF latex activity over antigen ratio for detecting such an acquired VWF abnormality is 0.8, where both sensitivity and specificity are close to 90%.
When the VWF latex activity ratio cutoff is set at 0.8 as indicated as the blue line, majority of the samples of acquired VWF abnormalities are now detected.
We then adopted a screening scheme to increase the accuracy and efficiency of VWD testing and at the same time decreased the testing volume of the laborious VWF ristocetin cofactor activity assay. Any sample that has less than 55 U/dL VWF antigen or latex activity, or less than 0.8 VWF latex activity over antigen ratio will be reflected for further VWF testing which includes VWF ristocetin cofactor activity and multimer analysis.
By using this criteria, we screened the same 492 samples; we can catch 100% congential VWD samples and 83% of acquired VWF abnormalities due to subtle losses of highest molecular weight multimers. The specificity is 92% which can be translated into a 92% reduction of VWF ristocetin cofactor activity testing volume of the potentially normal samples.
Based on our validation data, we developed this new VWD profile testing algorithm, as illustrated here.
Let us focus on the VWD testing part. In this panel, we will first test for factor VIII coagulant activity, VWF antigen and VWF latex activity.
If VWF antigen and latex activity are equal to or higher than 55 U/dL, and the activity over antigen ratio is above 0.8, no further testing will be performed as highlighted by the green lines and arrows. But keep in mind that the sensitivity of detecting an acquired VWF abnormality due to subtle losses of highest molecular weight multimers is 90%.
When a sample has less than 55 U/dl of VWF antigen, latex activity, or VWF activity ratio is less than 0.8, this sample, as indicated be the red lines and arrows, will be further tested by VWF ristocetin cofactor activity assay and multimer analysis.
Finally, what are the limitations of this new VWF activity assay? First: it is not really a biological function assay; therefore it is uncertain if it will pick up all type 2 VWD variants. Finally: whether this new assay can be used to monitor VWF replacement therapy is still under investigation.
In summary: We think that these new VWF latex activity assay has adequate sensitivity and specificity to be used as part of an initial screening panel. It has significantly improved the precision and lower detection limit. It is more sensitive than VWF ristocetin cofactor activity assay to detect acquired VWF abnormalities due to subtle losses of the highest molecular weight multimers. We also discussed some limitations of this new assay and explained our new testing algorithm. We hope that by implementing this new VWF activity assay and VWD testing algorithm we can significantly improve the accuracy and efficiency of VWD laboratory testing.
Here is the list of the references that I used in my talk.
Please let us know your comments and suggestions regarding to our new VWF testing algorithm. Please also check Hot Topic lectures by Drs. William Nichols and Rajiv Pruthi regarding other aspects of the VWD laboratory testing. This concludes my talk, and I thank you for your attention.