Diagnosis of type 1 VWD: Can the clinical history trump laboratory findings?

The diagnosis of von Willebrand disease (VWD) is often problematic and at best, challenging, especially among children. VWD is defined by a positive personal and family bleeding history as well as by low von Willebrand factor (VWF) levels. Additional focus on the diagnosis has been directed by the International Society of Thrombosis and Haemostasis (ISTH) 1 and the Hospital for Sick Children (HSC) 2. The prevalence of VWD ranges from population studies (0.1–1%) 3, 4 to those assessed based upon referral for symptomatic treatment (0.01%) 5. Due to this range of disease manifestation, one might argue, therefore, which aspect of VWD diagnosis is most important, the bleeding history or determination of low VWF? And should children with VWD be assessed differently than adults? In this issue of PBC, Hyatt and colleagues sought to apply the ISTH and HSC diagnostic criteria to a large population of patients from Ohio, previously diagnosed with type 1 VWD. They theorized that a large percentage of their patients might not meet the more narrowly focused criteria put forth by the ISTH subcommittee on VWF. Indeed, this lack of correlation in children with VWD has been suggested by investigators from the HSC 2. The study population of Hyatt and colleagues included 201 patients identified from medical records at the Nationwide Children's Hospital, who were previously diagnosed with type 1 VWD based upon the treating hematologist. Patients were excluded if there were no identifiable laboratory or clinical history. Low laboratory levels for VWF were defined as <2 SD below the mean and standard laboratory methodology for VWF detection was employed. No attempt was made, however, to re-define the diagnosis (based upon history or VWF testing) as originally rendered by the treating hematologist. Almost the entire population was comprised of children and adolescents, with a median age of 8 years. Bruising (59%) and epistaxis (52%) were the most prevalent findings on history. Seventy-four percent of the patients had a first degree relative with a positive bleeding history and 35% actually had a relative with a diagnosis of type 1 VWD. Almost all patients (94%) had two laboratory assessments for VWD. Their patient population was categorized as having possible to definite type 1 VWD in 20% (ISTH criteria) or 90% (HSC criteria) of cases, respectively. Details of these criteria may be found in the manuscript text and figure. The authors conclude that the ISTH criteria for diagnosis of type 1 VWD falls short of identifying children who present with significant mucocutaneous bleeding and that development of pediatric focused diagnostic criteria for VWD would be helpful. This study presents a few challenges, including its retrospective nature. First, it did not employ ABO specific VWF normal ranges, a known modulator of VWF. Second, other conditions which might cause bleeding such as collagen defects (e.g., Ehlers Danlos) or platelet function defects where not necessarily considered in rendering a presumptive diagnosis of VWD. Third, the treating physician assigned a diagnosis of VWD and in this assessment, there is little corroborative evidence that a comprehensive bleeding history was obtained. Additionally, for those patients who had VWF testing in the normal range (33–39% of cases), was there an assumption by the treating physician that these patients “must” have had type 1 disease? This supposition could have been reinforced by the knowledge that some children undergoing phlebotomy manifest elevated VWF levels, or by being aware that another family member had known bleeding symptoms or type 1 VWD. Fourth, since all of the historical bleeding information was not necessarily evident from the chart review, comparison with the ISTH and HSC criteria might have been problematic. Finally, despite the public vetting of both the ISTH and HSC criteria, these are not bleeding scores per say and for that reason, may not accurately characterize the mucocutaneous symptoms of VWD from a diagnostic standpoint. Collecting the bleeding history and quantifying that history (the bleeding score) are important tools in the evaluation of a patient with possible type 1 VWD. However, sorting this information out from normal controls can be challenging, as 25% of the normal population may indeed have “bleeding symptoms,” particularly nose bleeds, gum bleeding, bruising, and menorrhagia 5. Accordingly, the scope and details surrounding taking a mucocutaneous bleeding history are extremely important. Tosetto et al. 6 have appropriately utilized a bleeding history and developed a score in the diagnosis of type 1 VWD, but their study comprised very few children. Bleeding symptoms among children appear different than adults, due to fewer hemostatic challenges (surgery, tooth extraction, and menstruation) and lack of proper determination of the importance of bruising, nose/mouth bleeding and bleeding with cuts/abrasions in defining a bleeding disorder in the young age group. The diagnosis of type 1 VWD cannot be established without having a reduced VWF. How reduced, however, is the subject of debate. Most laboratories utilize a lower limit of normal VWF level of 50 IU/dl (<2 SD below the mean) but others suggest lower levels should be employed to define the disease: <20 1, <30 7, or <40 IU/dl 8. However, VWF detection and its importance to the diagnosis of VWD becomes even trickier when one considers the effect of certain modifiers (blood type O and its 25% lower VWF levels) and other co-existing conditions which are associated with higher VWF levels: hyperthyroidism, advancing age, acute and chronic inflammation, diabetes, malignancy, pregnancy, the menstrual cycle, oral contraception, stress and exercise. In light of the potential difficulties of making a diagnosis of type 1 VWD given the irregularities of the bleeding history and VWF determination, a recent study by Tosetto and colleagues employed a statistical method known as the Bayes theorem to investigate the relative contribution of both tools in making an accurate diagnosis of VWD. Using this approach, they established the relative importance (higher likelihood ratio; LR) of the family history and VWF level over the bleeding score in rendering a diagnosis 8. Their work also emphasized that at least 2/3 of the ISTH criteria (personal bleeding history, family history and laboratory determination) were needed to make a diagnosis of type 1 VWD. Additionally, along with the lack of absolute clarity in the history and laboratory determination of VWD, there also remains the difficulty of completely linking type 1 VWD with the VWF gene. Linkage occurs in approximately two-thirds of cases and usually in those who have a VWF level <30 IU/dl 9. Lack of linkage and the known association of clinical VWD with VWF levels >30 IU/dl may be due to various modifying genetic factors preventing complete expression of type 1 VWD such as: blood type O, variable penetrance of the VWF gene, associated platelet receptors/genes 10 and unknown modifying genes or other complex factors outside the VWF locus. So, why bother making a diagnosis of type 1 VWD? Similarly, some argue that because of possible inappropriate therapy, limitation of activities and insurance denial which might accompany the diagnosis of VWD, these realities by themselves should be pause for concern in rendering an inappropriate diagnosis. Conversely, what should be the approach to trauma and surgery for those who might have a diagnosis of VWD? Uncertainty is almost always magnified in the clinical setting. Hyatt and colleagues have made a case that the current bleeding criteria as proposed by the ISTH may be inadequate in identifying those with mucocutaneous bleeding symptoms leading to a diagnosis of type 1 VWD. These criteria and the recently published bleeding score 6 might be less than adequate in their application to children. Indeed, this observation has been recognized internationally and currently there are a group of investigators re-visiting the bleeding score and its application to various age groups and bleeding conditions. Given the uncertainty of VWF levels and genetics, validation of this expanded bleeding score may certainly provide a greater role for the clinical history in establishing a diagnosis of type 1 VWD. Indeed, the history may trump determination of VWF levels!