Pediatr Blood Cancer 2009;52:102107

Applying Diagnostic Criteria for Type 1 von Willebrand Disease to a Pediatric Population
Sarah A. Hyatt,
BS,
1

Wei Wang,

2 MS, MAS,

Bryce A. Kerlin,

MD,

2,3

and Sarah H. OBrien,

MD, MSc

2,3

Background. Denitive diagnosis of type 1 von Willebrand disease (VWD) remains elusive in clinical practice. Both the International Society on Thrombosis and Haemostasis (ISTH) and the Hospital for Sick Children (HSC) have proposed diagnostic criteria for type 1 VWD, which include abnormal laboratory values, signicant mucocutaneous bleeding history, and/or positive family history. The objective of this study was to apply ISTH and HSC criteria to a large pediatric population previously diagnosed with type 1 VWD. We hypothesized that a substantial number of patients would not meet these diagnostic criteria. Procedure. We performed a retrospective medical record review of all type 1 VWD patients at a single childrens hospital. Frequencies of t for ISTH and HSC diagnostic criteria were calculated. Results. Of 201 patients, 4.5%

and 34% met the ISTH and HSC denitions, respectively. An additional 15% (ISTH) and 56% (HSC) met the denition of possible disease. For each denition, criteria for signicant bleeding were most likely to be met, while criteria for abnormal laboratory values were least likely. When dened as simply VWF:Ag 520 IU/ml, as recently proposed by the ISTH, only 9% of patients met the denition of type 1 VWD. Conclusions. ISTH criteria failed to identify the majority of children and adolescents who presented to medical attention with signicant mucocutaneous bleeding. Further work is needed to develop diagnostic criteria with improved clinical validity in the pediatric setting. Pediatr Blood Cancer 2009;52:102107.
2008 Wiley-Liss, Inc.

Key words:

pediatric; type 1 von Willebrand disease

INTRODUCTION
Von Willebrand factor (VWF) is a multimeric glycoprotein which mediates platelet adhesion and serves as a carrier for Factor VIII [1]. Type 1 von Willebrand disease (VWD), a partial, quantitative VWF deciency, is reported to be the most common inherited bleeding disorder, with an estimated prevalence of approximately 1% [2]. However, population studies likely overestimate the prevalence of medically signicant disease by identifying chance associations between mild non-specic bleeding and mildly depressed VWF levels [3]. Making a denitive diagnosis of type 1 VWD remains a challenge for hematologists, particularly in the pediatric setting. Bleeding manifestations may vary between affected family members, and VWF levels can be affected by outside factors such as stress, illness, and estrogen exposure [4]. In 1995, a working party of the International Society on Thrombosis and Haemostasis (ISTH) proposed consensus criteria for the diagnosis of type 1 VWD, which included signicant mucocutaneous bleeding history, positive family history, and abnormal laboratory values (Tables IIII) [5]. These requirements are often difcult to satisfy in clinical situations. Bleeding manifestations may be more subtle in children, who have had fewer challenges to the hemostatic system, such as surgeries, menstruation, and childbirth. Family history can also be problematic, as the index patient may have young siblings who have also had few hemostatic challenges. Recognizing these difculties, the ISTH also developed criteria for possible type 1 VWD. In 2000, the Hospital for Sick Children (HSC) in Toronto developed denitions of possible and denite type 1 VWD, designed with less stringent requirements (Tables IIII). When the ISTH and HSC criteria were applied to 41 previously diagnosed type 1 VWD patients at the HSC, only 24% met ISTH criteria for possible or denite type 1 VWD, while 93% met HSC criteria for possible or denite disease [6]. The authors concluded that ISTH criteria were problematic for the pediatric population due to: (1) requiring parents to recall the exact duration of bleeding episodes, and (2) the denition of positive family history as signicant bleeding and abnormal labs in a rst degree relative.

In 2006, the ISTH published an update on the pathophysiology and classication of VWD, stating that the diagnosis of type 1 VWD seems justied only in patients with very low VWF:Ag levels (520 IU/ml) [7]. Our objective was to apply the original ISTH (ISTH-O) and HSC criteria, as well as the updated ISTH recommendation (ISTH-U) to a larger population of pediatric patients currently labeled with a diagnosis of type 1 VWD. We hypothesized that a substantial number of patients would not meet either ISTH or HSC diagnostic criteria.

METHODS
Our study population included patients previously diagnosed with type 1 VWD from 1989 to 2007 at the Hemostasis and Thrombosis Center of Nationwide Childrens Hospital (Columbus, OH). Our institution does not have specied historical or laboratory requirements for the diagnosis of type 1 VWD in our patient population. Patients were identied using a LAB TRACKER database (Ground Zero Software Inc., San Francisco, CA) which includes all patients diagnosed with a bleeding disorder at our center. The medical records of these patients were reviewed and the following information collected: age, gender, documented bleeding symptoms, family history, and laboratory values (VWF:Ag,
College of Medicine, The Ohio State University, Columbus, Ohio; The Research Institute at Nationwide Childrens Hospital, Columbus, Ohio; 3Division of Pediatric Hematology/Oncology, Nationwide Childrens Hospital/Ohio State University, Columbus, Ohio
2

Sarah A. Hyatt and Sarah H. OBrien contributed equally to this work. This study was presented in abstract form at the 2007 American Society of Hematology Annual Meeting in Atlanta, GA. Grant sponsor: Landacre Honor Society of the Ohio State University College of Medicine. *Correspondence to: Sarah H. OBrien, The Research Institute at Nationwide Childrens Hospital, 700 Childrens Drive, Columbus, OH 43205. E-mail: sarah.obrien@nationwidechildrens.org Received 10 January 2008; Accepted 7 July 2008

2008 Wiley-Liss, Inc. DOI 10.1002/pbc.21755 Published online 24 September 2008 in Wiley InterScience (www.interscience.wiley.com)

Applying Criteria for VWD to Pediatric Population

TABLE I. Signicant Mucocutaneous Bleeding Symptoms

103

I. ISTH provisional consensus criteria A signicant mucocutaneous bleeding history requires two or more of the following symptoms in the absence of a blood transfusion history, or one symptom requiring treatment with blood transfusion, or one symptom recurring on at least 3 distinct occasions Nose bleeding, 2 episodes without a history of trauma not stopped by short compression of <10 min, or 1 episode requiring blood transfusion. Cutaneous hemorrhage and bruisability with minimal or no apparent trauma Prolonged bleeding from trivial wounds, lasting 15 min or more or recurring spontaneously during the 7 days after injury Oral cavity bleeding that requires medical attention, such as gingival bleeding, or bleeding with tooth eruption or bites to lips or tongue Spontaneous gastrointestinal bleeding requiring medical attention, or resulting in acute or chronic anemia, unexplained by ulceration or portal hypertension Heavy, prolonged, or recurrent bleeding after tooth extraction or other oral surgery such as tonsillectomy and adenoidectomy, requiring medical attention Menorrhagia resulting in acute or chronic anemia, or requiring medical treatment, not associated with structural lesions of the uterus Bleeding from other skin or mucous membrane surfaces requiring medical treatment (e.g., eye, ear, respiratory tract, genitourinary tract other than uterus) II. HSC working denition A signicant mucocutaneous bleeding history requires any one of the following Recurrent nose bleeds requiring medical treatment (packing, cautery, DDAVP, etc.) or leading to anemia Oral cavity bleeding lasting for at least half an hour, or restarting over the next 7 days or requiring medical treatment Skin laceration bleeding lasting for at least half an hour, restarting over the next 7 days or requiring medical treatment Menorrhagia requiring medical attention or leading to anemia Prolonged bleeding associated with dental extraction or other oral surgery Spontaneous gastrointestinal hemorrhage requiring medical attention or leading to anemia, unexplained by local causes, for example, ulceration or portal hypertension Prolonged bleeding from other skin or mucous membrane surfaces requiring medical treatment (e.g., eye, ear, respiratory tract, genitourinary)

VWF:RCo, and VWF multimer studies). Patients were excluded if bleeding histories and/or laboratory values were not documented in the medical record.

2 second degree relatives have laboratory tests compatible with type 1 VWD and a personal history of signicant mucocutaneous bleeding.

Signicant Mucocutaneous Bleeding History

All charts were individually assessed by one investigator (SAH) for the documented history of mucocutaneous bleeding symptoms. Bleeding histories were categorized as signicant or non-signicant by both HSC and ISTH-O criteria. The rst 10% of study subjects were independently reviewed by the principal investigator (SHO) to ensure concordance between investigators.

Laboratory Values
Much of the genetic variation in VWF levels is due to blood type, in that patients with Type O have levels 25% lower than non-Type O [8]. While both ISTH and HSC criteria specify ABO-adjusted reference ranges, our laboratory does not provide ABO-specic reference ranges, so we dened an abnormal VWF:Ag or VWF:RCo as greater than two standard deviations below the mean. From 1987 to 2004, VWF:Ag levels were determined by a routine Laurell rocket immunoelectrophoresis method (Helena Laboratories Corp., Beaumont, TX) and VWF:RCo was measured using an optical density-based platelet aggregometer, using ristocetin as the aggregating agent of a formalin-xed platelet suspension (Bio/Data Corporation, Horsham, PA). Since 2005, VWF:Ag is determined by an immunoturbidimetric assay and VWF:RCo is still measured by an optical density-based aggregometer (both from Diagnostica Stago,

Family History
Subjects were classied as having a positive or negative family history for type 1 VWD according to HSC and ISTH-O criteria. HSC criteria require that at least one rst degree relative meets the HSC laboratory denitions of VWD. By ISTH criteria, a family history is considered positive if at least one rst degree relative or at least

TABLE II. Laboratory Criteria for Type 1 VWD I. ISTH consensus criteria Levels of both VWF:Ag and VWF:RCo are >2 SD below the population mean and ABO blood type adjusted mean on two or more determinations II. HSC working denition Levels of VWF:Ag are below 95% lower condence limit for appropriate ABO blood type on two or more determinations OR Levels of VWF:RCo are below 95% lower condence limit for appropriate ABO blood type on two or more determinations, with a VWF:Ag level below the ABO-adjusted 95% lower condence limit on 1 or more determination III. New ISTH criteria Level of VWF:Ag is 520 IU/dl
Pediatr Blood Cancer DOI 10.1002/pbc

104

Hyatt et al.

TABLE III. Diagnostic Criteria for Type 1 VWD I. ISTH consensus criteria Type 1 VWD All of the following criteria must be satised Signicant mucocutaneous bleeding history Laboratory tests compatible with type 1 VWD Positive family history for type 1 VWD or an appropriate VWF mutation Possible type 1 VWD Laboratory tests compatible with type 1 VWD and Signicant mucocutaneous bleeding or positive family history of type 1 VWD If the person is asymptomatic with low VWF, they must have a positive family history dened by at least two relatives with denite type 1 VWD II. HSC working denition Type 1 VWD Laboratory tests compatible with type 1 VWD and Signicant mucocutaneous bleeding or positive family history of type 1 VWD Possible type 1 VWD Subjects with any measurement of VWF:Ag or VWF:RCo below the ABO-adjusted 95% lower condence limit

Inc., Parsippany, NJ). Reference ranges for VWF:Ag and VWF:RCo have not changed throughout the time period of the study. Our institution has developed a standardized initial laboratory assessment for the evaluation of bleeding symptoms. This prole includes a prothrombin time (PT), activated partial thromboplastin time (APTT), brinogen, thrombin time, FVIII, VWF:Ag, VWF:RCo, and platelet function analysis (PFA-100). Prior to the availability of the PFA-100 in 2003, bleeding times were performed. Additionally, a complete blood count is routinely obtained to rule out thrombocytopenia.

positive rst degree family history of bleeding symptoms and 35% had a positive rst degree family history of type 1 VWD. Ninety-four percent of the patients underwent VWD testing at least two times, and the lowest values were used for analysis. Mean VWF:Ag and VWF:RCo were 43.9 (18.7) and 43.2 (20.3) IU/ml, respectively. In our laboratory, two standard deviations below the mean is 50 IU/ml for VWF:Ag and 46 IU/ml for VWF:RCo. Of the 197 patients (97.5%) who underwent multimer testing, all had normal multimer distribution. Among the 41 subjects tested for blood type, 80% were Type O, and the remaining were Type A.
TABLE IV. Clinical and Laboratory Characteristics of 201 Patients Diagnosed With Type 1 VWD at Nationwide Childrens Hospital Percentage (%) Age 06 years 712 years 13 years Gender Male Female Presence of bleeding symptoms Easy bruising Epistaxis Menorrhagia Oral cavity bleeding Prolonged bleeding from minor lacerations Gastrointestinal bleeding Bleeding after tooth extractions or oral surgery Lowest recorded VWF:Ag 520 IU/ml 2149 IU/ml 50 IU/ml (normal values) Lowest recorded VWF:RCo 520 IU/ml 2145 IU/ml 46 IU/ml (normal values) 14 34 52 47 53 59 52 25 23 16 14 12 9 58 33 12 49 39

Data Analysis
Data was analyzed using SAS 9.1. Frequencies of age, gender and t of diagnostic criteria were calculated. We compared three sets of diagnostic denitions: ISTH, HSC and updated ISTH [57]. Based on the criteria met for bleeding symptoms, family history and laboratory values, subjects were classied as denite type 1 VWD, possible type 1 VWD, or normal according to HSC and ISTH-O denitions (Table III). Subjects were classied as simply denite type 1 VWD or normal based on the ISTH-U laboratory denition of VWF:Ag 520 IU/ml [7].

RESULTS
We identied 249 patients diagnosed with type 1 VWD from 1989 to 2007. Forty-three patient charts were no longer available for review, and an additional 5 patients were excluded due to incomplete laboratory values or bleeding history in the medical record. Our analysis is based on the remaining 201 patients. Age at diagnosis ranged from 3 weeks to 29 years, with a median of 8 years (Table IV). Only three patients were >18 years at the time of diagnosis. Most patients (60%) were referred for evaluation of bleeding symptoms, with family history of VWD (20%) and abnormal laboratory values (17%) being other common reasons for referral. The most common bleeding symptoms documented were easy bruising (59%) and epistaxis (52%). Other symptoms included menorrhagia, oral cavity bleeding, prolonged bleeding from minor lacerations, gastrointestinal bleeding, and bleeding after tooth extractions or oral surgery. Seventy-four percent of subjects had a
Pediatr Blood Cancer DOI 10.1002/pbc

Diagnosis was based on the opinion of the treating hematologist, and was not dependent on any specied historical or laboratory criteria.

Applying Criteria for VWD to Pediatric Population

In our patient population, 90% of patients met HSC criteria for having either denitive or possible type 1 VWD, with 34% classied as denite, and 56% classied as possible (Fig. 1). When applying the more stringent ISTH denition, the number of patients meeting denite or possible criteria decreased dramatically, with 80% of patients classied as normal. Finally, we analyzed each component of the HSC and ISTH denitions separately (bleeding, family history, laboratory values) to assess the proportion of patients who did and did not meet each component (Table V). Patients classied as normal according to these denitions were most likely to meet the bleeding history component and least likely to meet the criteria for abnormal laboratory values.

105

DISCUSSION
From 1995 to 1996, the ISTH Subcommittee on VWF took the rst steps towards developing a formal set of diagnostic criteria for type 1 von Willebrand Disease. The difculties of diagnosis in young children lacking a convincing personal or family history of bleeding were recognized from the start, and therefore the committee made a distinction between denite and possible type 1 VWD [5]. This concern was validated by Dean et al. when they demonstrated that only 10 of 41 (24%) previously diagnosed type 1 VWD patients at the Hospital for Sick Children met ISTH criteria for denitive or possible disease [6]. These proportions are similar to our own ndings in a much larger population (Fig. 1). However, our results are likely overestimates, as it is quite possible that some of our Type O patients would no longer have met laboratory criteria if our laboratory provided ABO-specic normal ranges. The importance of characterizing blood type in the diagnosis of type 1 VWD continues to be debated. Recently published guidelines from the National Heart, Lung, and Blood Institute (NHLBI) state that although it has been recommended to stratify reference ranges with respect to blood group, evolving limited information supports that the major determinant of bleeding symptoms is low VWF. Therefore, referencing VWF testing results to the population reference range may be more useful clinically [9]. When considering the diagnosis of VWD in the pediatric setting, attention is most often paid to the difculties with bleeding history

(children have not experienced the same number of hemostatic challenges) and family history (young siblings may not yet have come to medical attention). However, in our population of patients who did not meet ISTH or HSC criteria, it was more often due to their laboratory values, the most objective component of the evaluation. Due to uctuations in VWF levels, repeated testing is sometimes needed to identify low levels of VWF [9]. Not meeting laboratory criteria may be partly due to a hesitancy to expose children to multiple rounds of testing, yet the majority of our population had at least two separate measurements of VWF levels. Venipuncture performed under stressful circumstances, including crying and anxiety, common occurrences in the pediatric setting, may result in increased VWF levels [10]. It is possible that our physicians took this information into account when diagnosing children with borderline or even normal levels with type 1 VWD. Physicians may also have taken family history into account, as 16 of the 20 patients with normal laboratory studies had either a family history of VWD (n 12) or a family history of abnormal bleeding (n 4). Most persons diagnosed with mild type 1 VWD likely have a coincidental combination of mild bleeding symptoms (very common in the general population) and a moderately low VWF level [3]. The relative contribution of VWF mutations appears to be more important in severely affected patients and are highly heritable in this population, while the contribution of other factors such as ABO blood type, platelet receptor expression variability, and other unknown genetic modiers appear to be more important in milder cases [1113]. As only very low VWF:Ag levels have been consistently associated with signicant bleeding and VWF mutations, various cutoffs of VWF:Ag have been proposed for the denitive diagnosis of type 1 VWD. In an update on the pathophysiology and classication of VWD by the ISTH, a VWF:Ag level of 520 IU/ml was proposed [7]. The appropriate VWF:Ag cutoff for what should be considered denitive disease remains unclear, however. Tosetto et al. recently reported that using a higher cutoff of <40 IU/ml still predicted a high likelihood of type 1 VWD [14]. A cutoff of <30 IU/ml has been proposed by the National Heart, Lung, and Blood Institute (NHLBI) [9]. Using these cutoffs in our population identied 9% (ISTH), 46%

Fig. 1. Classication of 201 previously diagnosed type 1 VWD patients as denite disease, possible disease, or normal once HSC or ISTH criteria are applied. The bottom half of the gure identies what percentage of patients would meet each of three recently proposed laboratory denitions of type 1 VWD: (1) VWF:Ag 520 IU/ml, as proposed by the ISTH; (2) VWF:Ag <30 IU/ml, as proposed by the National Heart, Lung, and Blood Institute; and (3) VWF:Ag <40, as proposed by Tosetto et al..
Pediatr Blood Cancer DOI 10.1002/pbc

106

Hyatt et al.
TABLE V. A Comparison of the Diagnostic Impact of Bleeding History, Family History, and Laboratory Values When Using the HSC and ISTH Denitions of Denite Type 1 VWD Subject nal diagnosis by the two denitions HSC classication Denite type 1 VWD Possible type 1 VWD Normal ISTH classication Denite type 1 VWD Possible type 1 VWD Normal Positive bleeding history 63 (93%) 75 (66%) 11 (55%) 9 (100%) 17 (55%) 87 (54%) Positive family history 24 (35%) 35 (31%) 10 (50%) 9 (100%) 15 (48%) 47 (29%) Positive laboratory values 68 (100%) 10 (9%) 0 (0%) 9 (100%) 17 (55%) 5 (3%)

Total n 68 113 20 9 31 161

(Tosetto), and 22% (NHLBI) of our patients as having type 1 VWD (Fig. 1). How important is it to distinguish between true type 1 VWD and low VWF in clinical practice? This is perhaps the source of greatest controversy for practicing hematologists. Patients misdiagnosed with type 1 VWD may receive unnecessary desmopressin or factor products for minor procedures, may limit their activities due to risk of bleeding, may have unfounded fears about passing a genetic disease on to their offspring, and have even been denied insurance coverage [3]. On the other hand, will insurance companies cover desmopressin, an expensive medication, for children only labeled as low VWF? Will emergency providers take the appropriate precautions if a trauma patient does not have a denitive diagnosis of VWD? Several limitations of our study deserve mention. First, ABO type was identied in only 20% of our study population. Second, we cannot be certain that a comprehensive bleeding history was obtained in each case, as our retrospective review relied on the data physicians chose to document in the medical record. Finally, it is possible that some of our patients may have had a bleeding disorder other than type 1 VWD. Rare causes of bleeding such as FXIII or antiplasmin deciency may have been missed, although these entities are more likely to present with severe bleeding symptoms similar to hemophilia. Since our institution performs a standard laboratory panel for the evaluation of bleeding symptoms (see Methods Section), patients with more common bleeding disorders should have been captured. Platelet aggregation studies are not included in this panel, so patients with platelet dysfunction or collagen defects, which do typically present with mild bleeding symptoms, may have been missed if the initial platelet function analysis (PFA-100) was normal. In summary, we found that the majority of our patients with type 1 VWD did not meet the ISTH denition of denite or even possible type 1 VWD, thus conrming in a larger population the ndings of HSC investigators [6]. In addition, we found that <10% of our patients met the newly proposed ISTH denition of a VWF:Ag level between 5 and 20 IU/ml. Therefore, these criteria failed to identify a substantial number of children and adolescents who presented to medical attention with signicant mucocutaneous bleeding. It is also possible, and in fact likely, that type 1 VWD was over-diagnosed in our patient population and some of these patients do not truly have VWD. The ISTH criteria may be an excellent scientic tool for identifying a narrow, severely affected population of patients likely to have autosomal dominant VWD mutations. However, further work is needed to develop a set of
Pediatr Blood Cancer DOI 10.1002/pbc

diagnostic criteria with improved clinical validity in the pediatric setting.

ACKNOWLEDGMENT
SAH received funding from the Landacre Honor Society of the Ohio State University College of Medicine. Special thanks to Dr. Kelly Kelleher for his thoughtful review of our manuscript, and to Angel Hateld, Claudia Lupia, Charmaine Biega, and Linda Grooms for their administrative and nursing support.

REFERENCES
1. Ruggeri ZM, Ware J. von Willebrand factor. FASEB J 1993;7:308 316. 2. Rodeghiero F, Castaman G, Dini E. Epidemiological investigation of the prevalence of von Willebrands disease. Blood 1987;69: 454459. 3. Sadler JE. Von Willebrand disease type 1: A diagnosis in search of a disease. Blood 2003;101:20892093. 4. Nichols WC, Ginsburg D. von Willebrand disease. Medicine 1997; 76:120. 5. Sadler JE, Rodeghiero F. Provisional criteria for the diagnosis of VWD type 1. J Thromb Haemost 2005;3:775777. 6. Dean JA, Blanchette VS, Carcao MD, et al. von Willebrand disease in a pediatric-based populationcomparison of type 1 diagnostic criteria and use of the PFA-100 and a von Willebrand factor/collagen-binding assay. Thromb Haemost 2000;84:401 409. 7. Sadler JE, Budde U, Eikenboom JC, et al. Update on the pathophysiology and classication of von Willebrand disease: A report of the Subcommittee on von Willebrand Factor. J Thromb Haemost 2006;4:21032114. 8. Gill JC, Endres-Brooks J, Bauer PJ, et al. The effect of ABO blood group on the diagnosis of von Willebrand disease. Blood 1987;69: 16911695. 9. Nichols WL, Hultin MB, James AH, et al. von Willebrand disease (VWD): Evidence-based diagnosis and management guidelines, the National Heart, Lung, and Blood Institute (NHLBI) Expert Panel report (USA). Haemophilia 2008;14:171 232. 10. Laffan M, Brown SA, Collins PW, et al. The diagnosis of von Willebrand disease: A guideline from the UK Haemophilia Centre Doctors Organization. Haemophilia 2004;10:199 217. 11. Eikenboom J, Van Marion V, Putter H, et al. Linkage analysis in families diagnosed with type 1 von Willebrand disease in the European study, molecular and clinical markers for the diagnosis

Applying Criteria for VWD to Pediatric Population

and management of type 1 VWD. J Thromb Haemost 2006;4:774 782. 12. James PD, Notley C, Hegadorn C, et al. The mutational spectrum of type 1 von Willebrand disease: Results from a Canadian cohort study. Blood 2007;109:145154. 13. Di Paola J, Federici AB, Mannucci PM, et al. Low platelet alpha2beta1 levels in type I von Willebrand disease correlate with

107

impaired platelet function in a high shear stress system. Blood 1999;93:35783582. 14. Tosetto A, Rodeghiero F, Castaman G, et al. Impact of plasma von Willebrand factor levels in the diagnosis of type 1 von Willebrand disease: Results from a multicenter European Study (MCMDM-1VWD). J Thromb Haemost 2007;5:715 721.

Pediatr Blood Cancer DOI 10.1002/pbc