The diagnosis of type 1 von Willebrand disease (VWD), the most common inherited bleeding disorder, presents a diagnostic challenge in children. In the absence of a prior hemostatic challenge in children, a VWD diagnosis may be missed by the Tosetto bleeding score, as it is based, in part, on bleeding symptoms, i.e. with surgery, menses, or childbirth, that may not occur until after childhood. In fact, 25% or more of children with VWD may be diagnosed only after they experience postoperative bleeding. Other bleeding scores, such as the James score, which assesses early life bleeding, e.g. cephalohematoma, macroscopic hematuria, and umbilical stump, post-circumcision, post-venipuncture bleeding, has been validated in children with VWD, but early life bleeding events are uncommon in VWD. We previously described a 4-variable Composite Score that has 92.5% sensitivity and 95% specificity for diagnosing VWD in children < 11 years of age, when at least two of four criteria are positive: 1) Tosetto bleeding score ≥ 1; 2) family history of VWD or bleeding; 3) personal history of iron deficiency anemia; and/or 4) positive James early bleeding score. The purpose of this study was to validate a Composite Score of ≥ 2 (i.e. at least 2 of 4 variables positive) for identifying children with VWD.
We designed and conducted a prospective validation study to determine the accuracy of the Composite Score for identification of children (<11 years of age) with VWD. Following parental consent, children without a prior history of VWD undergoing VWD testing were enrolled, and survey data were collected including personal bleeding history (including Tosetto and James scores), history of iron deficiency anemia, and family history of VWD or bleeding. Results of VWD testing were collected, including VWF:RCo, VWF:Ag, and VIII:C in addition to VWF multimers, if performed. Survey data and bleeding scores were compared between those with and without VWD, defined by VWF:RCo <30 (NIH 2008 criteria), or <40, or <50 IU/dL. Prior to study, it was determined that a Composite Score ≥ 2 would have 90% power to identify VWD defined by VWF:RCo <0.30 IU/dL, and 80% power to identify VWD by VWF:RCo <40 IU/dL. Categorical variables were compared using chi-square or Fischer’s exact test. Continuous variables were compared using two-sample t-test or Wilcoxon’s rank sum test. Sensitivity, specificity, positive and negative predictive value of the Composite Score of ≥ 2 to diagnose VWD were determined.
A total of 195 subjects were enrolled from 12 participating centers from 2010-2013, of whom 193 patients with no missing data were included in the analysis. A total of 81 (41.9%) children had type 1 VWD, including 11 (5.7%) with VWF:RCo <30 IU/dL (Group A), 24 (12.4%) with VWF:RCo <40 IU/dL (Group B), and 46 (23.8%) with VWF:RCo <50 IU/dL (Group C); the remainder with VWF:RCo ≥50 IU/dL were unaffected. There were no significant demographic differences between groups: overall 54.9% were male, 82.4% were Caucasian, and the mean age was 4.7 years. A composite score of ≥ 2 had 63.6%, 75%, and 67.4% sensitivity in diagnosing type 1 VWD in Groups A, B, and C respectively; 33.5%, 34.9%, and 34% specificity in diagnosing type 1 VWD, respectively; and 93.8%, 90.8%, and 76.9% negative predictive value, respectively. Based on the performance of the composite, with VWD being defined by a VWF:RCo of <30, <40, or <50 IU/dL, 31.6%, 30.6%, and 25.9% of subjects, respectively, had a negative composite score and therefore would not have required VWD testing.
This is the first study to prospectively validate the Composite Score to identify children with type 1 VWD. Although the sensitivity and specificity of the Composite Score were lower than in the original study (above), the latter included only VWD subjects, while this validation study included all, VWD and non-VWD, undergoing testing. The negative predictive value of the Composite Score was robust, especially at lower VWF:RCo, and indicates that VWD testing could be eliminated in nearly a third of children referred for VWD testing. Given the difficulty of diagnosing VWD in children, the Composite Score may be an important screening tool for diagnosis of type 1 VWD in children.
Kerlin:Bayer Healthcare US: Consultancy, Honoraria, Membership on an entity’s Board of Directors or advisory committees; NovoNordisk A/S: Consultancy, Honoraria.
Asterisk with author names denotes non-ASH members.