Diagnostic Fidelity of Historically Diagnosed Patients with VWD Enrolled from 27 Centers in the ZPMCB-VWD

Abstract 381^FN2

The TS Zimmerman Program for the Molecular and Clinical Biology of von Willebrand Disease (ZPMCB-VWD) is a large NIH PPG to study existing subjects with von Willebrand Disease in the United States and to contrast these with prior and ongoing studies in Canada and the UK. 569 index cases (ICs) and 1732 family members were recruited from 8 Primary Clinical Centers and 19 Secondary Clinical Center. The inclusion criteria were that the subjects had a historical diagnosis of VWD and were registered as ongoing patients in the local HTC. 247 normal controls (NCs) were studied for comparison. Data included pre-existing diagnosis, historical diagnostic VWD testing, detailed bleeding history using a modified MCMDM-1VWD QBS, and subjects had plasma and DNA drawn for studies at a central laboratory for VWD, VWF phenotyping, and full length VWF exon sequencing. ICs included 391 type 1, 105 type 2, 43 type 3, and 30 unclassified. The recent HLBI Guidelines suggested that the diagnosis of VWD be based on a VWF level of <30 IU/dL (either VWF:Ag or VWF:RCo).

Using the historical data, only 39.4% of ICs had either VWF:Ag or VWF:RCo of <30. Since historic data were studied using various lower limits of normal, we found 80% of subjects had either an Ag or RCO of <60. Central Laboratory testing of NCs and ICs was performed. While 7 of 247 NCs had reduced VWF:Ag or VWF:RCo, none of the NCs had <30 by either assay. Of the NCs with low VWF, 6 of 7 were blood type O and none had an abnormal bleeding score. While 372 of 569 ICs (65.4%) had reduced VWF assays by either method, only 201 (35.3%) had VWF assays that were below the NHLBI Guidelines of <30. Among ICs the correlation between current and historic VWF levels was an r²=0.199 for VWF:RCo and r²=0.268 for VWF:Ag. The current VWF:RCo correlated with the current VWF:Ag with an r²=0.827. The VWF:Ag and VWF:RCo were also restudied in a second laboratory with an r²=0.849 for VWF:Ag and r²=0.854 for VWF:RCo.

The MCMDM-1VWD QBS was normal in all normal controls. In contrast, 350 of the ICs born prior to 1996, 247 (70.5%) had an abnormal QBS. Evaluated by pre-existing diagnosis, 65% of type 1 VWD, 78.6% of type 2 VWD and 100% of type 3 VWD had abnormal QBS. Interestingly, of the 351 VWD subjects born before 1996, the QBS was abnormal in 83.5% of males and 67.7% of females. In this age group, there were 248 females and 103 males followed with the diagnosis of VWD.

VWF sequencing was carried out on all ICs and compared to the VWF Database maintained at the University of Sheffield. Sequence variations were present in 100% of the 391 type 1 VWD subjects with VWF levels of <10, 96% with VWF 11–20, 77% with VWF 21–30, 70% with VWF 31–40, 39% with VWF 41–50, and 42% of those with VWF 51–60. The decrease in VWF sequence variation between those <40 and those >40 seemed to be striking. The mean VWF:Ag levels of type 1 ICs with a VWF sequence variation was significantly lower (p<0.0001) than those with normal VWF sequence.

More than 20 years ago, we demonstrated that besides blood type, age had the most significant effect on level of VWF in normal blood donors. We evaluated the VWF:Ag levels in our 247 NCs and VWF:Ag levels rose about 5% for each 10 years of age from age 20 to 60. In VWD it is not known if a similar process results in reduced bleeding symptoms with age or if this increase is related to the aging process or progressive vascular pathology. Stress is a significant confounding variable in the hemostatic evaluation of children – particularly for the diagnosis of VWD, but the effect of aging has not been defined.

We therefore have demonstrated that the diagnosis of VWD was not consistently substantiated in a large group of patients diagnosed with VWD – particularly if using the VWF levels recommended by the NHLBI Guidelines. Historic VWF levels do not correlate with current levels of VWF. Moreover, 35% of these VWD subjects do not have evidence even of reduced VWF levels upon retesting. A number of questions remain. Are the NHLBI recommendations too strict? Is there an effect of aging on the normal level of VWF that affects the correct diagnosis of VWD? Will more rigid initial diagnostic testing improve the fidelity of the diagnosis of VWD? Is the diagnosis of VWD applied to a group of individuals with clinical bleeding with only some of these being associated with low VWF? Further longitudinal study of these VWD subjects is important. In those with significant clinical bleeding, other hemostatic abnormalities must be sought. The fidelity of the diagnosis of VWD needs improvement.