SEARCH

SEARCH BY CITATION

Keywords:

  • blood coagulation disorders;
  • diagnosis;
  • epidemiology

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure of Conflict of Interests
  9. References

Summary. Background: Quantitative bleeding assessment tools (BATs) have been used to describe the severity of the bleeding phenotype in patients with von Willebrand disease. Objectives: To evaluate the clinical usefulness of a BAT for the diagnosis of mild bleeding disorders (MBDs) in previously undiagnosed patients. Methods: We prospectively assessed 215 patients who were consecutively referred for evaluation of bleeding symptoms (n = 71), abnormal laboratory clotting test results (n = 105) or family investigation (n = 39) at two second-level centers. The bleeding history was assessed by a young investigator who administered the BAT instrument, and also by a senior physician who independently evaluated the patient and made the final diagnoses. Sensitivity, specificity, negative predictive value (NPV) and positive predictive value (PPV) were computed for a predefined bleeding score (BS) cut-off (BS of > 3). Receiver operating characteristic curves were used to establish a diagnostic prediction rule. Results: Assuming the prevalence of MBD in the general population to be ∼ 1%, a normal BS (≤ 3) had a very high NPV (99.2%). The PPVs in patients referred for hemostatic or family evaluation at second-level clinics were estimated to be 71.0% and 77.5% (assuming MDB prevalences of 20% and 50%, respectively, in these settings). Measurement of BS in addition to activated partial thromboplastin time significantly increased the diagnostic efficiency of the BAT instrument (NPV of 99.6%). Conclusions: BAT use improves the evaluation of patients with suspected MBD, and we propose its use in a clinical prediction guide based on BAT and activated partial thromboplastin time for the exclusion of patients with suspected MBD in a low-prevalence setting.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure of Conflict of Interests
  9. References

Patients with mild bleeding disorders (MBDs) usually exhibit a mild to moderate bleeding tendency that is most often characterized by mucocutaneous bleeding after trauma or invasive procedures [1]. In contrast to patients with severe bleeding disorders, MBD patients neither experience major bleeding nor have increased mortality. Nonetheless, their quality of life can be impaired [2], and as they can bleed, particularly after surgery, they may require blood-derived products to control this.

In the general population, the most frequent MBDs are von Willebrand disease (VWD) and platelet function disorders, each of which has an estimated frequency of up to 1% [3,4]. For these reasons, the diagnosis of MBD in patients presenting with minor bleeding symptoms is frequently pursued, at least in affluent societies. Physicians are often asked to exclude an MBD, particularly in patients requiring invasive procedures or elective surgery who have abnormal clotting screening test results.

Unfortunately, most patients with an MBD do not show a definitive bleeding history, and are difficult to distinguish from normal subjects [5–7]. For this reason, more quantitative bleeding assessment tools (BATs) have been proposed to reduce interobserver variability and improve the diagnostic criteria for MBD [8]. The use of BATs has been mainly limited to the diagnosis of type 1 VWD, which is a useful model for mild to moderate bleeding disorders [9], but they could also be potentially useful for the diagnosis of platelet function disorders [10]. The clinical usefulness of a BAT in daily clinical practice is, however, unknown, as very few data are available on the use of a BAT in consecutive patients referred for the evaluation of bleeding disorders.

In this study, we evaluated the diagnostic utility of a standardized BAT in a prospective study on patients referred for the evaluation of abnormal bleeding or laboratory test results at two second-level hemostasis centers.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure of Conflict of Interests
  9. References

Patients

In this cross-sectional study, we consecutively enrolled all patients referred for hemostatic evaluation at two hemostasis services (Vicenza General Hospital, Italy, and Leiden University Medical Center, The Netherlands) over a 27-month period. Patients were considered to be eligible if they were referred by their general practitioner or another caring physician for evaluation of: (i) bleeding symptoms; (ii) abnormal laboratory test results (prothrombin time [PT] or activated partial thromboplastin time [APTT] prolongation); or (iii) family study, the patient being the first-degree relative of a patient with a known bleeding disorder. Patients were considered to be not eligible if they were older than 80 years, were taking vitamin K antagonists or receiving antiplatelet therapy, were pregnant, or had thrombocytopenia (< 1011 platelets L−1) or lupus anticoagulant. All patients, or their parents, gave informed consent.

BAT

For the present investigation, we used a condensed version of the BAT developed for the European MCMDM-1 VWD study [11], retaining in the questionnaire only those questions relevant for computation of the MCMDM-1 VWD bleeding score, as previously described by Bowman [12]. The BAT used in the study is presented in Fig. 1. The bleeding score (BS) was computed according to the MCMDM-1 VWD study criteria [11].

image

Figure 1.  Condensed bleeding assessment tool used in the study. CNS, central nervous system; GI, gastrointestinal.

Download figure to PowerPoint

Study protocol

All eligible patients underwent a two-step evaluation. In the first step, the patient was interviewed by a young investigator, who administered the condensed questionnaire and computed the BS. The young investigator was blinded to the reason for referral of the patient; a parent was interviewed, or helped in the interview, if the patient age was <18 years of age. In the second step, a senior investigator (G.C., J.E., A.T.), who was expert in the diagnosis of bleeding disorders and blinded to the results of the first step, evaluated the patient, ordered additional laboratory investigations if deemed necessary, and offered a final diagnosis on the basis of their clinical expertise and judgement. This latter diagnosis was considered to be the ‘true’ (reference) diagnosis, and was not disputed on the basis of the results of the first step. To avoid any possible influence on the patient of the interview with the senior investigator, the young investigator always collected the data before the senior investigator. On the basis of a preliminary chart review of patients referred at the Vicenza Hemophilia and Thrombosis Center, we estimated that ∼ 25% of patients had been diagnosed as having an MBD. We therefore aimed at consecutively enrolling at least 200 patients to identify a sufficient number of affected patients (n > 50).

Laboratory investigations

Per protocol, the minimal screening panel in patients considered by the senior investigator as worthy of further investigations was as follows: PT, APTT, fibrinogen, and platelet count. These investigations were always performed in patients referred for evaluation of abnormal laboratory test results. Other specific tests (von Willebrand factor antigen or ristocetin cofactor, Platelet Function Analyzer [PFA-100, Siemens Healthcare Diagnostics Products, Marburg, Germany] closure time or bleeding time; measurement of specific clotting factors; impedance aggregometry) were performed by each local laboratory according to the local standards and upon request of the senior physician. All reported laboratory values were measured at the time of evaluation, in the investigators’ laboratories.

Criteria for an abnormal bleeding history

Per protocol, and based on the findings of the European MCMDM-1 VWD Study, we prespecified that all subjects with BS of > 3 had an abnormal bleeding history [11]. As a secondary analysis, we evaluated the sensitivity and specificity of BS of > 2 [13].

Diagnoses

Per protocol, senior physicians classified subjects either as having no disease, an MBD (i.e. platelet function disease, VWD, mild hemophilia, or factor XI deficiency), or an alternative diagnosis (senile purpura, asymptomatic factor deficiency [e.g. FXII deficiency], or Rendu–Weber–Osler disease), on the basis of their clinical findings and the local laboratory results. As patients classified as having an ‘alternative diagnosis’ do not have an appreciable risk of surgical bleeding (the reason why subjects were investigated), they were pooled in the normal group for subsequent analyses.

Statistics

Differences between subgroups were tested with the exact Fisher or median test, for qualitative or quantitative variables, respectively. Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and area under the receiver operating characteristic (ROC) curve were computed according to Pepe [14]. PPV and NPV were computed under the hypothesis of a different prevalence, to simulate different clinical scenarios. The following prevalences were chosen: 1%, to account for screening in a general-population setting (assuming such a prevalence for VWD and platelet function disorders [3,4]); 20% for prevalence of MBD in a second-level clinic; and 50% for expected prevalence of an autosomal dominant MBD within a family [15]. Multivariate logistic regression was used to identify predictors of the presence of MBD in enrolled patients, by contrasting patients with a diagnosis of MBD with subjects receiving another diagnosis. We finally used the c-statistic to test whether addition of BS could actually improve the identification of subjects with MBD and guide the clinician towards the request for further investigations. The c-statistic tests the null hypothesis that the areas under the ROC curves of two diagnostic tests are the same. All computations were performed with the Stata software package (version 11; Statacorp, College Station, TX, USA) [16].

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure of Conflict of Interests
  9. References

Patients

Two hundred and fifteen patients were enrolled in the study, 71 being referred for evaluation of abnormal clotting test results, 105 for the presence of bleeding symptoms, and 39 for family investigation (Table 1). There were 37 pediatric (<14 years) patients; the mean BS in pediatric cases was lower than, but non-statistically different from, that in the adult cases (1.37 vs. 1.78, P = 0.96 by median test). Subjects referred for evaluation of an abnormal clotting test result were younger and had a longer APTT ratio than subjects referred for bleeding symptoms or family investigation. After exclusion of patients with Rendu–Weber–Osler disease (in whom the diagnosis was based on physical examination alone), a full laboratory evaluation (including von Willebrand factor assay and platelet aggregation) was ordered by the senior investigator in 89/99 (90%) of cases referred for the evaluation of bleeding symptoms, and in 82/85 (96%) of those with BS of > 0.

Table 1.   Characteristics of evaluated subjects, by type of referral
 Reason for referralP*
Abnormal clotting test result (n = 71)Bleeding symptoms (n = 105)Family investigation (n = 39)
  1. APTT, activated partial thromboplastin time; PT, prothrombin time. *Test differences between referral subgroups. †P = 0.61 when comparing patients referred for abnormal clotting tests and those with bleeding symptoms. ‡PT and APTT ratios as measured at the time of evaluation, in the investigators’ laboratories.

Male/female ratio35/3637/6817/220.17
Age (years), median (range)34 (5–69)38 (5–79)45 (12–72)0.08
Pediatric (< 14 years) (%)18/53 (25)18/87 (17)1/38 (3)0.01
Blood group O/non-O (%)20/51 (28)26/79 (25)2/37 (5)0.01†
PT ratio, median (range)‡1.08 (0.83–1.49)1.05 (0.9–1.28)1.04 (0.95–1.21)0.14
APTT ratio, median (range)‡1.2 (0.7–5.0)1.07 (0.74–1.54)1.1 (0.97–1.49)0.001

Diagnoses

Table 2 reports the frequency of specific diagnoses by type of referral. In total, 56 patients were classified as having MBD and 159 as not having MBD. MBD was more frequently (44%) diagnosed in patients referred for family investigation, but there was no difference in MBD prevalence for patients referred for bleeding symptoms or abnormal clotting test results (26% vs. 17%, respectively, P = 0.198).

Table 2.   Diagnoses given by the senior physicians, by type of referral
 Reason for referral
Abnormal clotting test (n = 71)Bleeding symptoms (n = 105)Family investigation (n = 39)
  1. MBD, mild bleeding disorder; VWD, von Willebrand disease.

VWD6111
Mild hemophilia141
Platelet function defect1112
FXI deficiency4113
Total number of patients with MBD (%)12 (17)27 (26)17 (44)
Senile purpura2
Asymptomatic clotting defect1721
Rendu–Weber–Osler disease6
Normal subjects426821
Total number of patients without bleeding disorder (%)59 (83)78 (74)22 (56)

Diagnostic utility of a quantitative BS

On the basis of the presence or absence of MBD, the sensitivity, specificity and PPV or NPV were computed (Table 3). In a low-prevalence scenario (MBD prevalence of 1%), a normal BS (≤ 3) showed a very high NPV (above 99%), essentially ruling out MBD in patients with a normal BS from the general population. Conversely, the PPV was acceptable in patients referred for hemostatic evaluation at secondary-level clinics or for family evaluation (71.0% and 77.5%, for MDB prevalences of 20% and 50%, respectively). The area under the ROC curve was significantly greater with a cut-off for BS of > 2 than with a cut-off for BS of > 3 (0.69 vs. 0.63, P = 0.04).

Table 3.   Sensitivity and specificity of an abnormal bleeding score (BS) for the diagnosis of a mild bleeding disorder
Type of referralSensitivitySpecificityPPV% PrevalenceNPV% Prevalence
1205012050
  1. NPV, negative predictive value; PPV, positive predictive value. PPVs and NPVs are reported on the assumption of different prevalences (depending on clinical setting); for subjects referred for family investigation, PPVs and NPVs are reported only for a prevalence of 50% (assumed on the basis of autosomal dominant model). *99.4, based on a cut-off for abnormal BS of > 2 (instead of > 3); †87.6, based on a cut-off for abnormal BS of > 2 (instead of > 3).

Abnormal clotting test results0.250.9813.071.199.282.5
Bleeding symptoms0.410.812.134.699.384.5
Abnormal clotting test results or bleeding symptoms0.330.882.841.699.2*84.1†
Family investigation0.470.8677.562.0

We finally evaluated the additive value of including the BS in a diagnostic pathway directed at identifying MBD. In a logistic regression model that included age, gender, blood group, PT, APTT, and BS, only an abnormal APTT ratio (defined as an APTT ratio of > 1.20) and an abnormal BS (> 3) were significantly and independently associated with the presence of MBD (odds ratio [OR] 4.3 and OR 6.6, respectively; P < 0.0001 for both variables). These results did not materially change after exclusion of patients with FXI deficiency (OR 4.3 and OR 4.6, respectively; P < 0.001 for both variables). Addition of the BS significantly increased the area under the ROC curve of a prediction model based on the presence of an abnormal APTT alone (from 0.63 to 0.74, P = 0.013). Following this finding, we tested the diagnostic efficiency of a hypothetical diagnostic scenario in which patients referred for evaluation of bleeding symptoms or laboratory abnormalities were referred for further clinical assessment if they had a BS of > 3 or an APTT ratio of > 1.2, despite having a normal BS (Fig. 2). In this scenario, a ‘negative’ screening had an NPV of 99.6% under low-prevalence conditions (prevalence 1%), which increased to 99.8% when a BS of > 2 was considered as a cut-off for abnormal bleeding.

image

Figure 2.  Derivation of a clinical prediction guide in patients referred for evaluation of bleeding symptoms or abnormal laboratory test results. APTT, activated partial thromboplastin time; BS, bleeding score.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure of Conflict of Interests
  9. References

In this study, we assessed the clinical utility of a quantitative BAT in outpatients referred to a second-level clinic for the possible presence of an MBD. Few studies have prospectively addressed the usefulness of the bleeding history in the identification of MBD [6], an otherwise relevant topic, as the usual screening tests (PT, APTT, and bleeding time) have a very low predictive power for screening purposes [17,18]. The main strengths of the study are the use of a well-described BAT that has been extensively used in patients with type 1 VWD [11], the prospective enrolment of consecutive patients at two hemostasis centers, and the design of the study, which allows a blinded evaluation of enrolled patients.

As expected, we observed that the prevalence of MBD was strongly dependent on the type of referral of patients. Among subjects not referred for family study, 39/176 subjects were diagnosed as having an MBD (22%), most of them (22/39, 56%) having a disorder of primary hemostasis (platelet function disorder or type 1 VWD). In contrast, in subjects referred for family investigation, the prevalence of MBD was close to that expected for autosomal dominant disorders (17/39, 44%) [15]. These wide differences in the prevalence of MBD are important when evaluating the diagnostic efficiency of the BAT used in the present study. In clinical practice, a BAT could be used both to exclude an MBD in unselected subjects (hence negating further clinical or laboratory testing) and to identify patients who are likely to have an MBD (who may benefit from specific treatment). To attain the first goal, the BAT should have a high sensitivity (and NPV), whereas a high specificity (and PPV) is required for the second goal.

For the exclusion of MBD in an unselected population, it could be estimated that about one patient in every 125 (NPV of 99.2%) could be falsely considered as being normal when his or her BS is below or equal to 3. With a less stringent criterion (and hence increasing sensitivity), i.e. by considering as abnormal those subjects with a BS of > 2 (instead of 3, as prespecified in our study protocol), only one patient in every 167 (NPV of 99.4%) would have been misclassified as normal when in fact having an MBD (Table 2).

We subsequently tested the hypothesis that the BAT could be incorporated into a clinical prediction guide (CPG) including additional tests, and for this purpose we examined, using logistic regression and ROC analysis, whether additional tests could be useful to identify patients with MBD. Only an abnormal APTT ratio was found to be associated with MBD, and the observed marked increase in the c-statistic supports the combination of these two tests as a screening strategy for MBD. As reported in Fig. 2, a CPG considering those individuals with a BS of > 3 or an abnormal APTT ratio as possible MBD patients would misclassify one patient in every 250 (NPV of 99.6%); alternatively, if based on a BS cut-off of > 2, it would misclassify one patient in every 500 (NPV of 99.8%). These and the preceding values could, however, be considered to represent a worst-case scenario, if we assuming the MBD prevalence to be as high as 1%. In an unselected population (e.g. before surgical procedures), the NPV could actually be higher, assuming a lower prevalence for more severe forms of MBD. Therefore, although the present findings need to be supported by additional studies, a CPG based on two rapid, inexpensive and standardized investigations (BAT and APTT) could be cost-effective and potentially negate the need for additional testing or referral to second-level hemostasis centers.

For a positive diagnosis of MBD, this study suggests that about one-fourth of patients referred to a second-level clinic for family investigation or abnormal hemostatic screening test results (PT, APTT, fibrinogen) may have MBD. In these patients, a complete laboratory investigation should be considered to be mandatory in the presence of an abnormal BS, given the high PPV in such a setting (71.0% and 77.5%, respectively, for patients referred for abnormal bleeding test results and family investigation).

Our study has some limitations. First, as there is no reference standard for the diagnosis of MBD, we relied on the diagnoses given by senior investigators working in experienced centers, following consensus guidelines when available [19]. Notably, these expert clinicians classified 11 patients who were referred for evaluation of bleeding symptoms or with an abnormal BS (> 3) but with no laboratory abnormalities as ‘normal subjects’. These patients have been previously identified as patients with unexplained mucocutaneous bleeding [20], and were conservatively classified as normal subjects, given the uncertain clinical relevance of these subjects. Although a predefined set of investigations was not prespecified in our study protocol, the expert clinicians requested a complete hemostatic evaluation (including von Willebrand factor and platelet function assays) in 82/85 (96%) of those referred for evaluation of bleeding symptoms who had a ‘blinded’ BS of > 0. As a specific diagnosis was reached in 26% of patients referred for evaluation of bleeding symptoms (Table 2), the expected misclassification rate would be ([85−82] × 0.26)/85 = 0.91% or lower. This would change our specificity estimates from 0.81 to 0.80 in the group referred for evaluation, and so would not materially affect our conclusions. A classification bias could be excluded in the groups referred for evaluation of abnormal bleeding test results or for family investigation, because, in these patients, the expert clinicians always sought for the diagnosis of a specific defect.

Second, the intraindividual and interindividual variability of diagnoses made by expert clinicians has never been formally assessed, and this may have introduced some random (but non-differential) classification bias. Third, the lack of a central core laboratory may have introduced a further source of classification bias. Fourth, we could blind only the young, and not the expert, clinician as to the reason for patient referral.

Finally, the inclusion in this study of young subjects, with a shorter exposure to hemostatic challenges and, consequently, a lower BS, may have reduced the sensitivity and NPV of an abnormal BS in this setting, as has been recently reported [21]. Although this conservatively affects our main study findings, it should be kept in mind that a detailed laboratory investigation is always required to exclude with confidence the presence of an MBD in a very young subject. On the other hand, a retrospective study on patients diagnosed with type 1 VWD suggests that the bleeding risk in patients with an MBD and a negative BS could possibly be low [11], and so the need for a detailed investigation in subjects with a normal BS should be balanced against the procedural bleeding risk.

To summarize, we found that the addition of a standardized BAT is valuable for assessment of MBD based on laboratory screening alone. A normal BS could reasonably exclude the presence of an MBD in an unselected population, especially when coupled with a normal APTT. In contrast, an abnormal BS always calls for a more detailed hemostatic evaluation, particularly in patients referred for abnormal laboratory or family investigation. On the basis of these findings and the recent availability of a consensus ISTH BAT for inherited bleeding disorders (largely based on the BAT used for this work) [22], we encourage the use of a standardized BAT in the prospective evaluation of patients referred for hemostatic evaluation.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure of Conflict of Interests
  9. References

We thank F. Scognamiglio and A. Cappelletti for their valuable work in collecting clinical data of enrolled patients.

Disclosure of Conflict of Interests

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure of Conflict of Interests
  9. References

The authors state that they have no conflict of interest.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure of Conflict of Interests
  9. References
  • 1
    Rodeghiero F, Tosetto A, Castaman G. How to estimate bleeding risk in mild bleeding disorders. J Thromb Haemost 2007; 5 (Suppl. 1): 15766.
  • 2
    de Wee EM, Mauser-Bunschoten EP, van der BoJ, Degenaar-Dujardin ME, Eikenboom JC, Fijnvandraat K, de Goede-Bolder A, Laros-van Gorkom B, Meijer K, Raat H, Leebeek FW, Win Study Group. Health related quality of life among adult patients with moderate and severe von Willebrand disease. J Thromb Haemost 2010; 8:14929.
  • 3
    Nieuwenhuis HK, Akkerman JW, Sixma JJ. Patients with a prolonged bleeding time and normal aggregation tests may have storage pool deficiency: studies on one hundred six patients. Blood 1987; 70: 6203.
  • 4
    Rodeghiero F, Castaman G, Dini E. Epidemiological investigation of the prevalence of von Willebrand’s disease. Blood 1987; 69: 4549.
  • 5
    Silwer J. von Willebrand’s disease in Sweden. Acta Paediatr Scand Suppl 1973; 238: 1159.
  • 6
    Sramek A, Eikenboom JC, Briet E, Vandenbroucke JP, Rosendaal FR. Usefulness of patient interview in bleeding disorders. Arch Intern Med 1995; 155: 140915.
  • 7
    Wahlberg T, Blomback M, Hall P, Axelsson G. Application of indicators, predictors and diagnostic indices in coagulation disorders. I. Evaluation of a self-administered questionnaire with binary questions. Methods Inf Med 1980; 19: 194200.
  • 8
    Tosetto A, Castaman G, Rodeghiero F. Bleeding scores in inherited bleeding disorders: clinical or research tools? Haemophilia 2008; 14: 41522.
  • 9
    Goodeve A, Eikenboom J, Castaman G, Rodeghiero F, Federici AB, Batlle J, Meyer D, Mazurier C, Goudemand J, Schneppenheim R, Budde U, Ingerslev J, Habart D, Vorlova Z, Holmberg L, Lethagen S, Pasi J, Hill F. Phenotype and genotype of a cohort of families historically diagnosed with type 1 von Willebrand disease in the European study, Molecular and Clinical Markers for the Diagnosis and Management of Type 1 von Willebrand Disease (MCMDM-1VWD). Blood 2007; 109: 11221.
  • 10
    Biss TT, Blanchette VS, Clark DS, Bowman M, Wakefield CD, Silva M, Lillicrap D, James PD, Rand ML. Quantitation of bleeding symptoms in children with von Willebrand disease: use of a standardized pediatric bleeding questionnaire. J Thromb Haemost 2010; 8: 9506.
  • 11
    Tosetto A, Rodeghiero F, Castaman G, Goodeve A, Federici AB, Battle J, Meyer D, Fressinaud E, Mazurier C, Goudemand J, Eikenboom JC, Schneppenheim R, Budde U, Ingerslev J, Vorlova Z, Habart D, Holmberg L. A quantitative analysis of bleeding symptoms in type 1 von Willebrand disease: results from a multicenter European study (MCMDM-1 VWD). J Thromb Haemost 2006; 4: 76673.
  • 12
    Bowman M, Mundell G, Grabell J, Hopman WM, Rapson D, Lillicrap D, James P. Generation and validation of the Condensed MCMDM-1VWD Bleeding Questionnaire for von Willebrand disease. J Thromb Haemost 2008; 6: 20626.
  • 13
    Rodeghiero F, Castaman G, Tosetto A, Battle J, Baudo F, Cappelletti A, Casana P, De Bosch N, Eikenboom JC, Federici AB, Lethagen S, Linari S, Srivastava A. The discriminant power of bleeding history for the diagnosis of von Willebrand disease type 1: an international, multicenter study. J Thromb Haemost 2005; 3: 261926.
  • 14
    Pepe M. The Statistical Evaluation of Medical Tests for Classification and Prediction. Oxford: Oxford University Press, 2003.
  • 15
    Weiss KM. Genetic Variation and Human Disease. Cambridge: Cambridge University Press, 1993.
  • 16
    StataCorp. Stata Statistical Software: Release 11.0. College Station, TX: Stata Corporation, 2010.
  • 17
    Shaw PH, Reynolds S, Gunawardena S, Krishnamurti L, Ritchey AK. The prevalence of bleeding disorders among healthy pediatric patients with abnormal preprocedural coagulation studies. J Pediatr Hematol Oncol 2008; 30: 13541.
  • 18
    Watson HG, Greaves M. Can we predict bleeding? Semin Thromb Hemost 2008; 34: 97103.
  • 19
    Sadler J, Rodeghiero F. Provisional criteria for the diagnosis of VWD type 1; on behalf of the SSC Subcommittee on von Willebrand factor. J Thromb Haemost 2005; 3: 7757.
  • 20
    Quiroga T, Goycoolea M, Panes O, Aranda E, Martinez C, Belmont S, Munoz B, Zuniga P, Pereira J, Mezzano D. High prevalence of bleeders of unknown cause among patients with inherited mucocutaneous bleeding. A prospective study of 280 patients and 299 controls. Haematologica 2007; 92: 35765.
  • 21
    Marcus PD, Nire KG, Grooms L, Klima J, O’Brien SH. The power of a standardized bleeding score in diagnosing paediatric type 1 von Willebrand’s disease and platelet function defects. Haemophilia 2011;17: 2237.
  • 22
    Rodeghiero F, Tosetto A, Abshire T, Arnold DM, Coller B, James P, Neunert C, Lillicrap D; Group ft. ISTH/SSC bleeding assessment tool: a standardized questionnaire and a proposal for a new bleeding score for inherited bleeding disorders. J Thromb Haemost 2010; 8:20635.