Von Willebrand disease (VWD) is the single most common congenital bleeding disorder. The Tosetto Bleeding Score (BS), a quantitative measure of bleeding, has improved specificity of diagnosis, but symptoms are variable, and, among children a diagnosis of VWD may be missed as they may not have lived long enough to experience symptoms on which the score is based. As many as 25% of affected children come to clinical attention after postoperative bleeding: thus, a better diagnostic tool is needed in children. A Tosetto BS modified for children by one additional question on early life bleeding, known as the James BS, has been validated for a small cohort of VWD children, but its utility in type 1 VWD diagnosed by the 2008 NHLBI-defined VWF:RCo<30 or in children <11 years of age remains unknown. Further, the role of family bleeding history or personal history of anemia in VWD diagnosis has not been studied.
We conducted a case-control study to determine predictors of VWD in children younger than 11 years of age. Cases included 40 children with VWF:RCo<30 IU/dL (Group A) and 39 children with VWF:RCo=30-50 IU/dL (Group B). Controls included 80 children without VWD, age-, sex-, race-matched to CASES<30. Tosetto and James bleeding scores, family history of VWD or bleeding, and personal history of anemia were determined by prospective data extraction from charts of cases at the time of VWD diagnosis, and by parental interview of controls.
Group A included 52% male, 90% Caucasian, and age 7.5 years (median), not different from Group B, p=0.91, p=0.67, and p=0.26; or controls, p=0.90; p=1.00, and p=0.78, respectively. The median Tosetto BS in Group A was 2, greater than in controls, median 0, p<0.0001, but similar to Group B, median 1, p=0.10. The median James BS in Group A was 2, greater than in controls, median 0, p<0.0001, but not different from Group B, median 1, p=0.06. The majority of Group A had a family history of VWD or bleeding, 97%, similar to Group B, 85.0%, p=0.06; but higher than in controls, 0%, p<0.0001. In Group A, 47.5% had a personal history of iron deficiency anemia (by history and/or MCV<80), higher than Group B, 18%, p=0.005, and higher than controls, 1.3%, p<0.0001. The most common bleeding symptom in Group A was cutaneous bleeding, in 60%, similar to Group B, 44%, p=0.14, but higher than controls, 2.5%, p<0.0001. Comparing Group A with controls, a composite score of four variables, 1) Tosetto BS>1, 2) a family history of VWD or bleeding, 3) a personal history of iron deficiency (by history and/or MCV<80), and/or 4) a James ‘other' early event bleeding, was predictive of VWD, p<0.0001, with 100% specificity and 75% sensitivity if at least 2 variables were present; similarly, a composite of three variables, excluding Tosetto BS>1, was also predictive of VWD, p<0.0001, with 98.8% specificity and 100% sensitivity if at least one variable was present. Comparing Groups A and B (all cases) with controls, a composite score of all four variables, was predictive of VWD, p<0.0001, with 100% specificity and 58.2% sensitivity if at least 2 variables were present; similarly, a composite of three variables, excluding Tosetto BS >1, was predictive of VWD, p<0.0001, with 98.8% specificity and 92.4% sensitivity if at least 1 variable was present. The area under the curve for diagnosing VWD in Group A by the Tosetto BS was AUC=0.872, Wald 95% CI (0.804−0.940), and by the James BS was AUC=0.885 (Wald 95% CI 0.820–0.950), and not different from the area under the curve for diagnosing VWD in Group B, AUC=0.801, Wald 95% CI (0.721−0.881), for each score.
The AUC for the Tosetto bleeding score for diagnosing VWD in children < 11 years of age, exceeds 0.87, and the James exceeds 0.80, and both appear to be independent of VWF level <30 vs. 30–50. Combining a family history of VWD or bleeding, a personal history of iron deficiency anemia (by history or MCV<80) improves the Tosetto BS in children, independent of VWF level. These findings should be validated in a larger group of children.
Kulkarni:Novo Nordisk: Membership on an entity's Board of Directors or advisory committees, Participate in clinical trials. Yee:Novo Nordisk: Consultant. Ragni:Novo Nordisk, Inc.: Research Support for this study.
Asterisk with author names denotes non-ASH members.