Clinical and laboratory variability in a cohort of patients diagnosed with type 1 VWD in the United States

Abstract

von Willebrand disease (VWD) is the most common inherited bleeding disorder, and type 1 VWD is the most common VWD variant. Despite its frequency, diagnosis of type 1 VWD remains the subject of debate. In order to study the spectrum of type 1 VWD in the United States, the Zimmerman Program enrolled 482 subjects with a previous diagnosis of type 1 VWD without stringent laboratory diagnostic criteria. von Willebrand factor (VWF) laboratory testing and full-length VWF gene sequencing was performed for all index cases and healthy control subjects in a central laboratory. Bleeding phenotype was characterized using the International Society on Thrombosis and Haemostasis bleeding assessment tool. At study entry, 64% of subjects had VWF antigen (VWF:Ag) or VWF ristocetin cofactor activity below the lower limit of normal, whereas 36% had normal VWF levels. VWF sequence variations were most frequent in subjects with VWF:Ag <30 IU/dL (82%), whereas subjects with type 1 VWD and VWF:Ag ≥30 IU/dL had an intermediate frequency of variants (44%). Subjects whose VWF testing was normal at study entry had a similar rate of sequence variations as the healthy controls (14%). All subjects with severe type 1 VWD and VWF:Ag ≤5 IU/dL had an abnormal bleeding score (BS), but otherwise BS did not correlate with VWF:Ag. Subjects with a historical diagnosis of type 1 VWD had similar rates of abnormal BS compared with subjects with low VWF levels at study entry. Type 1 VWD in the United States is highly variable, and bleeding symptoms are frequent in this population.

Figure 1

Sequence variations in VWD are most common in subjects with VWF:Ag <30 IU/dL. This graph shows the number of subjects with sequence variations (either point mutations, or insertions or deletions) in the VWF coding sequence (dark gray) as compared with those without sequence variations in the VWF coding sequence (light gray) for the entire type 1 VWD cohort by VWF:Ag as compared with the healthy controls. The percent of each group with sequence variations is shown at the top of each column. Sequence variations were most common in those with VWF:Ag <30.

Figure 2

No significant difference in BS for type 1 VWD subjects regardless of VWF:Ag level. This graph shows the number of subjects with abnormal BS (defined as >2 in children <18 years of age, >3 in adult males, and >5 in adult females) in dark gray as compared with those with normal BS (light gray) for the entire type 1 VWD cohort by VWF:Ag. The percent of each group with abnormal BS is shown at the top of each column. BS were similar for type 1 subjects regardless of VWF:Ag.

Figure 3

Correlation of sequence variations with BS. This box and whisker plot compares BS using the ISTH BAT for adult subjects (≥18 years of age) with type 1 VWD (VWF:Ag and/or VWF:RCo below the lower limit of normal at study entry) in the first pair of columns (“VWD 1”), those with a historical diagnosis of type 1 VWD but normal laboratory findings at study entry in the second pair of columns (“VWD 1 Hist”), and a comparison group of healthy control subjects in the third pair of column (“Controls”). Those subjects with a sequence variation are shown in dark gray, whereas those without a sequence variation are shown in light gray. There was no significant difference in BS between those with and those without a sequence variation for the historical type 1 cohort, and a borderline significant difference for the type 1 VWD cohort. NS, not significant.

Figure 4

BS vary across blood groups in type 1 VWD subjects. This graph shows BS for subjects with blood group A, AB, B, and O. Median BS are shown at the top of the graph. No significant difference was seen between blood group O and blood group B or AB. A borderline significant difference was seen comparing blood group O and blood group A (P < .015).

Figure 5

Variation in historical VWF testing for Zimmerman Program subjects. (A-B) Comparison of the historical (A) and study entry (B) VWF:Ag (circles) and VWF:RCo (triangles) for all subjects enrolled with a diagnosis of type 1 VWD. The insets show the comparison of VWF:Ag on the x-axis and VWF:RCo on the y-axis for historical laboratory values (A) and study entry laboratory values (B). The correlation is much lower for historical values and improved for study entry values, as expected, given that all study testing was performed in the same laboratory and all testing was performed on the same sample for each subject. However, there still remain issues with the lower limit of the ristocetin cofactor assay, as seen by the number of VWF:RCo values at or below the lower limit of detection. hist, historical.