The Egyptian Journal of Haematology

: 2016  |  Volume : 41  |  Issue : 1  |  Page : 1--8

Platelet-dependent von Willebrand factor activity in acute myeloid leukemia patients: role in hemostatic alterations

Hala M.H. Abaza, Botheina A.T. Farweez, Sylvia F Samaan 
 Haematology Unit, Department of Clinical Pathology, Faculty of Medicine, Ain Shams University, Cairo, Egypt

Correspondence Address:
Botheina A.T. Farweez
Haematology Unit, Department of Clinical Pathology, Faculty of Medicine, Ain Shams University, Cairo, 11769


Background Patients with acute myeloid leukemia (AML) may suffer from bleeding tendency, which can aggravate their pre-existing condition. Bleeding in these patients occurs mainly because of thrombocytopenia, disseminated intravascular coagulopathy, and platelet dysfunction. However, other underlying mechanisms are complex and are not completely understood. Von Willebrand factor (vWF) defects have been implicated as contributors to bleeding complication. Aim The study aimed to investigate the pattern of vWF-ristocetin cofactor (RCO) activity in newly diagnosed AML patients and its relation to therapy, to explore its potential role in the prevalent phenomenon of bleeding in AML patients, in order to correlate between vWF-RCO activity and other hemostatic parameters and known prognostic factors in AML. Patients and methods Thirty newly diagnosed patients with AML and 20 healthy age-matched and sex-matched individuals (the control group) were studied. The vWF-RCO activity using the platelet agglutination method was determined on plasma samples of both patients and controls. Results Our results showed significantly reduced vWF-RCO activity in AML patients at diagnosis, which significantly increased 2 weeks after treatment. Our study proposes the following cutoffs for vWF-RCO: less than or equal to 85.2% for predicting AML patients prone to bleeding manifestations and less than or equal to 82.6% for the prediction of bad outcome. Conclusion In AML, vWF-RCO activity at diagnosis represents a valuable prognostic marker for predicting bleeding complication and bad outcome, and the provided cutoffs for vWF-RCO activity represent the first steps for its use as a bleeding predictor and prognostic marker, although future studies with larger series of patients may be needed before it can be incorporated into routine use.

How to cite this article:
Abaza HM, Farweez BA, Samaan SF. Platelet-dependent von Willebrand factor activity in acute myeloid leukemia patients: role in hemostatic alterations.Egypt J Haematol 2016;41:1-8

How to cite this URL:
Abaza HM, Farweez BA, Samaan SF. Platelet-dependent von Willebrand factor activity in acute myeloid leukemia patients: role in hemostatic alterations. Egypt J Haematol [serial online] 2016 [cited 2020 Jan 22 ];41:1-8
Available from:

Full Text


Patients with acute myeloid leukemia (AML) are at increased risk for bleeding, which occurs in 40-70% of patients at presentation. The coagulopathy in AML has been attributed to disease-related thrombocytopenia, platelet dysfunction, disseminated intravascular coagulation (DIC), excessive fibrinolysis, and the action of nonspecific proteases [1] .

Von Willebrand factor (vWF) is a multimeric glycoprotein essential for normal hemostasis. It carries and protects coagulation factor VIII in the circulation, and it recruits platelets to the site of clot formation during primary hemostasis. The vWF cross-links platelets with exposed collagen at a site of vessel damage, and together with fibrinogen it cross-links platelets during subsequent platelet aggregation. The efficacy of vWF in primary hemostasis depends on its level and function [2] .

Several vWF defects have been described in hematologic malignancies, such as proteolytic degradation of the vWF by proteinases released into the circulation by leukemic cells [3] . A vWF multimeric pattern resembling type 2A has also been reported in chronic myeloid leukemia [4] . In addition, an abnormal vWF multimeric pattern has been reported in patients with myeloproliferative disorder [5] .

This bleeding hazard has a significant negative impact on the morbidity and mortality of patients with AML. Recognition of the underlying causes is important in the diagnosis and management of acute leukemia [6] .

Therefore, this work aimed to investigate the vWF-ristocetin cofactor (RCO) activity pattern in newly diagnosed AML patients before the start of induction treatment and 2 weeks after induction therapy. In addition, the study aimed to correlate vWF-RCO activity with bleeding complication, to explore the prognostic impact of vWF-RCO activity measurement at AML diagnosis, and to correlate vWF-RCO activity with other hemostatic parameters and known prognostic factors in AML.

 Patients and methods

This study included 30 newly diagnosed patients with AML, attending the Hematology Oncology Unit, Ain Shams University Hospitals, before the start of induction treatment and 2 weeks after induction. The patients were included in the study on the basis of the standard morphological, immunophenotypic, and cytogenetic criteria for diagnosis of AML [7] . The study was performed during 2013-2014. Twenty healthy age-matched and sex-matched individuals were studied as a control group.

All patients were subjected to the following: full history taking (laying stress on bleeding manifestations) and thorough clinical examination (laying stress on pallor, menorrhagia, purpura, ecchymosis, bleeding from mucous membranes, organomegaly, or lymphadenopathy). The severity of bleeding was assessed according to the WHO standardized grading scale [8] . Laboratory investigations at presentation included automated complete blood count (Coulter LH750; Beckman Coulter, Miami, Florida, USA), with examination of Leishman-stained peripheral blood (PB) and bone marrow (BM) smears and cytochemical staining of BM smears with myeloperoxidase. Coagulation studies included prothrombin time, activated partial thromboplastin time (APTT), and fibrin-degradation products (FDPs) and/or D-dimer. Conventional cytogenetic analysis and/or fluorescence in-situ hybridization was performed on PB or BM samples to detect significant chromosomal abnormalities of probable diagnostic or prognostic value in addition to routine immunophenotyping by EPICS XL Flow Cytometer, (Beckman Coulter, Miami, Florida, USA) on PB and/or BM samples. The patients were followed up for 3 months after diagnosis to assess response to treatment according to guidelines of the European Society for Medical Oncology (ESMO) Working Group for diagnosis, treatment, and follow-up of AML in adult patients [9] .

The vWF-RCO assay (a platelet agglutination method) measures the ability of a patient's plasma to agglutinate platelets in the presence of the antibiotic ristocetin [10] . The rate of ristocetin-induced agglutination relates to the concentration and functional activity of the plasma vWF. It was performed on plasma samples of AML patients (at diagnosis and 2 weeks after induction treatment) and controls according to the manufacturer's instructions (Siemens Healthcare Diagnostics Products, Marburg, Germany). The study was conducted in accordance with the stipulations of the local ethical and scientific committees of Ain Shams University, and the procedures respected the ethical standards of the Helsinki Declaration of 1964.

Statistical analysis

Statistical analysis was performed using Statistical Package for the Social Sciences (SPSS) (version 17; SPSS Inc., Chicago, Illinois, USA). Quantitative data were presented as mean and SD for parametric data. Qualitative data were presented as number and percentage. Comparisons of quantitative variables were made between groups using the Student t-test for parametric data, and comparisons between more than two groups with parametric distribution were made using one-way analysis of variance. Correlations between quantitative variables within groups were determined using the Pearson correlation coefficient. A P value less than 0.05 was considered statistically significant and that less than 0.001 was considered highly statistically significant. A receiver-operator characteristic curve was constructed to establish clinically relevant cutoffs for the studied parameters, with calculation of sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and prognostic accuracy.


At diagnosis characteristics

The studied patients included nine (30%) men and 21 (70%) women, with a male to female ratio of 1: 2.3. Their ages ranged from 18 to 72 years, with a mean of 33.6 ± 14.3 years. The control group included six (30%) men and 14 (70%) women with a male to female ratio of 1: 2.3. Their ages ranged from 17 to 53 years, with a mean of 36.9 ± 11.1 years. Clinical and laboratory data of the studied AML patients at diagnosis are shown in [Table 1].{Table 1}

According to the bleeding score, 13 (43%) studied AML patients had no bleeding (grade 0) versus 17 (57%) with bleeding manifestations. The latter were classified as follows: seven (23%) patients had grade 1 bleeding (petechial bleeding), five (16.6%) had grade 2 bleeding (mild blood loss but clinically significant), and five (16.6%) patients had grade 3 bleeding (gross blood loss, which required transfusion); none of the studied patients had grade 4 bleeding (debilitating blood loss, retinal or cerebral, associated with fatality).

According to French American British (FAB) morphological classification [11] , cytochemistry [12] , and immunophenotyping [13] , five (16.6%) patients were classified as M0, four (13.3%) patients as M1, seven (23.3%) patients as M2, four (13.3%) patients as M3, five (16.6%) patients as M4, five (16.6%) patients as M5, and no patient was diagnosed as M6 or M7.

Cytogenetic studies of AML patients revealed that 23 (76.6%) patients were positive for cytogenetic abnormalities ([Table 2]) versus seven (23.3%) patients with normal karyotype. The patients were classified into cytogenetic prognostic groups as follows [14] : 12 (40%) patients had good prognosis, nine (30%) patients had intermediate prognosis, and nine (30%) patients had poor prognosis.{Table 2}

Comparative studies

The von Willebrand factor-ristocetin cofactor activity in acute myeloid leukemia versus the control group

The vWF-RCO activity among AML patients ranged from 52.8 to 100.6%, with a mean of 80.5 ± 12.6%. Meanwhile, the range in the control group was from 90.8 to 126.4%, with a mean of 99.5 ± 10.5%. The vWF-RCO activity fell within the established reference range (55-200%) [15] , except in the case of two patients. A statistically highly significant lower mean of vWF-RCO activity was documented among the patients when compared with controls (P < 0.001) ([Table 3]).{Table 3}

Von Willebrand factor-ristocetin cofactor activity 2 weeks after induction

In AML patients, vWF-RCO activity 2 weeks after chemotherapy ranged from 58.0 to 126.4%, with a mean of 87.7 ± 14.2%, which was higher than the pretreatment mean vWF-RCO activity. The difference was highly statistically significant (P < 0.001) ([Table 4]).{Table 4}

Relation between von Willebrand factor - ristocetin cofactor activity and different parameters

The vWF-RCO activity showed a statistically significant lower mean vWF-RCO activity among patients presenting with bleeding manifestations (76.0 ± 11.7%) in relation to those without bleeding (grade 0: 86.5 ± 11.4%) (P = 0.02). In addition, on grading the bleeding manifestation, patients with grade 3 had the lowest mean value of vWF-RCO activity (61.1 ± 5.2%), and showed a statistically significant difference compared with patients with grade 1 and grade 2 (P = 0.004 and 0.005, respectively) ([Table 5]). Moreover, the mean value of vWF-RCO activity was different among different FAB subgroups where M2 and M3 patients had a lower mean value of vWF-RCO activity, but the difference was not statistically significant (P = 0.06) ([Table 6]).{Table 5}{Table 6}

The vWF-RCO activity showed a significant negative correlation with international normalized ratio (INR) (r = −0.365, P = 0.047) and a statistically significant difference between FDP subgroups (>10, 10-40, and >40 mg/l). However, there was no statistically significant difference between vWF-RCO activity and demographic data (age and sex), clinical data (organomegaly and lymphadenopathy), or other studied laboratory data with known prognostic significance, including total leukocyte count, hemoglobin level, platelet count, BM blast percentage, prothrombin time, APTT, and cytogenetic subgroups ([Table 7]).{Table 7}

Etiology of bleeding in studied acute myeloid leukemia patients

Bleeding manifestation (and its grades) showed a significant relation with vWF-RCO activity (P < 0.001) ([Table 5]), but there was no relation to platelet count (P = 0.4) or FDP subgroups (P = 0.1). It is noteworthy that none of the five patients with major bleeding had FDPs more than 40 mg/l (no frank DIC) ([Table 8]).{Table 8}

Relation between FAB subgroups of acute myeloid leukemia patients, bleeding, and its etiology

There was no significant relation between different FAB subgroups and bleeding grades, vWF-RCO activity, and platelet count. However, it was noted that FAB M2 patients had lower mean vWF-RCO activity and a higher percentage of major bleeding. Meanwhile, there was a significant relation between different FAB subgroups and different FDP subgroups (P = 0.02) ([Table 6]).

Patient outcome

Regarding the clinical outcome, eight (26.6%) of 30 newly diagnosed AML patients achieved complete remission over 3 months of follow-up and were considered to have good outcome. Meanwhile, five (16.6%) patients did not achieve complete remission and 17 (26.6%) patients died shortly after diagnosis and they were considered to have poor outcome.

Relation between von Willebrand factor-ristocetin cofactor activity and outcome

There was a statistically significant relation between vWF-RCO activity and patients' outcome (P = 0.002) ([Table 9]).{Table 9}

Performance characteristics of von Willebrand factor-ristocetin cofactor activity

Using the receiver-operator characteristic curve analysis, a cutoff of less than or equal to 85.2% for vWF-RCO activity was able to predict patients with bleeding manifestations. This cutoff yielded a sensitivity of 94.1%, specificity of 69.2%, PPV of 80.0%, NPV of 90.0%, and prognostic accuracy of 0.790 ([Table 10] and [Figure 1]), whereas a cutoff of less than or equal to 82.6% for vWF-RCO activity was able to predict bad outcome. This cutoff yielded a sensitivity of 77.3%, specificity of 87.5%, PPV of 94.4%, NPV of 58.3%, and prognostic accuracy of 0.86 ([Table 11] and [Figure 2]).{Figure 1}{Figure 2}{Table 10}{Table 11}


In the current work, there was significantly reduced vWF-RCO activity among the studied AML patients at diagnosis, compared with the control group. This finding agreed with those of Shen [16] , who studied the vWF on the basis of its quantity and quality in acute leukemia patients, and revealed that the level of plasma vWF: Ag was significantly higher, whereas the vWF-RCO activity was reduced (or even absent) compared with normal controls. The investigator reported a relatively increased amount of small multimers in leukemic patients' plasma that did not have the active function of vWF [16] . Similarly, Federici and D'Amico [3] reported that proteolytic degradation of vWF as a part of the massive proteolytic state, triggered by procoagulant substances, plasminogen activators, and proteinases released into circulation from leukemic cells, might explain the decreased vWF-RCO activity at diagnosis, and add to the increased bleeding tendency in FAB-M3 patients. In addition, they reported a parallel reduction in the relative proportion of high-molecular-weight multimers in the FAB-M3 patients' plasma.

Regarding bleeding manifestations, the present work showed that low vWF-RCO activity was statistically significantly more frequent in AML patients with bleeding symptoms than in those without bleeding. This finding agreed with that of Lethagen et al. [17] , who documented the same findings on studying the vWF-RCO activity in young women with and without bleeding symptoms, taking into consideration the fact that no case fulfilling the strict diagnostic criteria for von Willebrand disease was included in their study. Moreover, Gudmundsdottir et al. [18] studied mild primary hemostatic defects (low vWF and mild platelet dysfunction) as a cause of increased bleeding symptoms in healthy teenagers, and revealed that vWF-RCO was significantly lower in those with excessive bleeding than in controls.

On grading the bleeding manifestation, the presently studied AML patients with grade 3 bleeding (gross blood loss requiring transfusion) showed a statistically significantly lower vWF-RCO activity compared with patients with grade 1 and grade 2 bleeding ([Table 5]). Thrombocytopenia, DIC, and the FAB subgroup did not show a relation to bleeding manifestation ([Table 6] and [Table 8]), as 83% of patients had FDPs less than 10 mg/l ([Table 1]) (excluding DIC as a cause) and bleeding manifestation was present in 57% of the patients, although all patients were thrombocytopenic (the mean platelet count in AML patients with bleeding was 39 × 10 3 vs. 49 × 10 3 in nonbleeding patients). Lower activity of vWF-RCO activity was noted in association with FAB-M2 and M3 subgroups, and the highest percentage of major bleeding was present in the FAB-M2 subgroup. Thus, the vWF-RCO activity may be implicated in the pathogenesis of bleeding in those groups of patients, and may represent a valuable prognostic marker for predicting bleeding complication; moreover, its prognostic impact may be independent from other established prognostic markers, as it did not show significant correlation with other laboratory data with prognostic significance ([Table 7]).

In the present study, a statistically highly significant difference in vWF-RCO activity was found after 2 weeks of therapy among the studied AML patients. The vWF-RCO activity increased and almost returned to normal after treatment. Likewise, Federici and D'Amico [3] reported that plasma vWF activity in FAB-M3 patients improved after treatment, and at day 15 it did not differ from those of normal controls. The same investigators related the post-treatment improvement in vWF-RCO activity to an improvement in the defect of the multimeric structure of vWF found at diagnosis. Not only the largest multimers found lacking at the time of onset of FAB-M3 reappeared, but also a novel set of ultralarge multimers appeared. This improvement in vWF integrity may be related to smaller amounts of proteolytic agents circulating in the plasma following myelocyte differentiation, or due to the direct effect of treatment on processing and secretion of vWF from endothelial cells, thus explaining the appearance of these ultralarge multimers in FAB-M3 patients' plasma.

In contrast, Nilsson et al. [19] studied coagulation activation and changes before and during treatment of AML. They measured vWF: Ag and vWF-RCO before induction of treatment and during the following 2 weeks of therapy. They documented that vWF: Ag and vWF-RCO activity levels were most elevated before induction of treatment and they attributed their findings to an initial ongoing increased thrombin generation that occurs before consumption of clotting factors, which is apparent in global hemostatic tests (e.g. prolonged APTT). The vWF: Ag and vWF-RCO activity levels dipped somewhat during induction therapy on days 3-5, then showed gradual improvement in activity, which peaked during days 13-14 after treatment. This discrepancy may be due to the different number of patients included in the studies (30 vs. 12) and/or due to the different analytical methods used for measuring vWF-RCO activity in both studies [manual platelet agglutination (vWF reagent for the determination of RCO activity; Siemens Healthcare Diagnostics Products) vs. rapid automated assay for analysis of vWF-RCO] [20] . The diversity of the reagent donor platelets and complexity of platelet aggregation kinetics [10],[15] or the difference in ethnic background (Egypt vs. Sweden) may also play a role.

On assessing the relation of vWF-RCO activity with outcome of AML patients, the present study revealed a statistically significantly lower vWF-RCO activity in AML patients with bad outcome (incomplete remission or death) at diagnosis. This further confirms the prognostic value of vWF-RCO activity at diagnosis in the prediction of outcome in AML patients.

Finally, the present study documented a significantly reduced vWF-RCO activity in AML patients at diagnosis, which significantly increased 2 weeks after induction therapy. The vWF-RCO activity, at diagnosis, represented a valuable prognostic marker for predicting bleeding complication and bad outcome (incomplete remission or death). The study proposed measuring the vWF-RCO activity using the platelet agglutination method with a cutoff of less than or equal to 85.2% for predicting the AML patients prone to bleeding manifestations and less than or equal to 82.6% for the prediction of bad outcome in AML patients. Moreover, the present study indicated the possible prognostic impact of reduced vWF-RCO activity as an independent prognostic marker in AML.

To confirm the documented findings, the present study recommends further evaluation of larger series of AML patients and study vWF-RCO activity and other coagulation parameters such as vWF : Ag, vWF : collagen binding, multimeric structure of vWF, protein C and S, and antithrombin III in AML to detect the correlations between them, and their prognostic importance. The measurement of vWF-RCO activity should be included in the routine laboratory screening of AML patients, especially in those who bleed at presentation. In addition, the therapeutic potential of vWF/factor VIII concentrate before or during induction therapy needs further evaluation.

Financial support and sponsorship


Conflicts of interest

There are no conflicts of interest.


1Langer F, Amirkhosravi A, Loges S, Meyer T, Eifrig B, Hossfeld D, et al. An in-vitro study on the mechanisms of coagulation activation in acute myelogenous leukemia (AML): role of tissue factor regulation by cytotoxic drugs and GM-CSF. Thromb Haemost 2004; 92 :1136-1146.
2 Vischer U. Von Willebrand factor, endothelial dysfunction, and cardiovascular disease. J Thromb Haemost 2006; 4 :1186-1193.
3 Federici A, D′Amico E. The role of von Willebrand factor in the hemostatic defect of acute promyelocytic leukemia. Leuk Lymphoma 1998; 31 :491-499.
4 Mohri H, Tanabe J, Yamazaki E, Yoshida M, Harano H, Matsuzaki M, et al. Acquired type 2A von Willebrand disease in chronic myelocytic leukemia. Hematopathol Mol Hematol 1996; 10 :123-133.
5 Budde U, Scharf R, Franke P, Hartmann-Budde K, Dent J, Ruggeri Z. Elevated platelet count as a cause of abnormal von Willebrand factor multimer distribution in plasma. Blood 1993; 82 :1749-1757.
6 Kwaan H, Huyck T. Thromboembolic and bleeding complication in acute leukemia. Expert Rev Hematol 2010; 3 :719-730.
7 Swerdlow S, Campo E, Harris N, Jaffe E, Pileri S, Thiele J et al. editors. World Health Organization (WHO) classification of tumors: pathology and genetics of haematopoietic and lymphatic tissues. 4th ed. Lyon, France: IARC Press; 2008.
8 Webert K, Cook R, Sigouin C, Rebulla P, Heddle N. The risk of bleeding in thrombocytopenic patients with acute myeloid leukemia. Haematologica 2006; 91 :1530-1537.
9 Fey M, Buske C. On behalf of the ESMO Guidelines Working Group. Acute myeloblastic leukaemias in adult patients: ESMO clinical practice guidelines for diagnosis, treatment and follow-up. Ann Oncol 2013; 24 :vi138-vi143.
10Flood V, Friedman K, Morateck P, Wren J, Scott J, Robert R. Limitations of the ristocetin cofactor assay in measurement of von Willebrand factor function. J Thromb Haemost 2009; 7 :1832-1839.
11Bennett J, Catovsky D, Daniel M, Flandrin G, Galton D, Gralnick H et al. Proposals for the classification of the acute leukaemias: French-American-British Cooperative Group. Br J Haematol 1976; 33 :451-458.
12Miller K, Daoust P. Clinical manifestations of acute myeloid leukemia. In: Hoffman R, Benz E, Shatti SJ, Furie B, Cohen H, Silberstain L, McGlave P, editors. Hematology basic principles and practice. 3rd ed. New York: Churchill Livingstone (a Division of Harcourt Brace Company); 2000.
13Matutes E, Morilla R, Morilla A. Immunophenotyping. Immunological classification of acute leukaemias. In: Bain B, Bates I, Laffan M, Lewis S, editors. Practical haematology. 11th ed. chapter 16, UK: Elsevier, Churchill Livingstone El-Sevier; 2012. p. 362-364.
14Betz B, Hess J. Acute myeloid leukemia diagnosis in the 21st century. Arch Pathol Lab Med 2010; 134 :1427-1433.
15Blackshear J, Schaff H, Ommen S, Chen D, Nichols W. Hypertrophic obstructive cardiomyopathy, bleeding history, and acquired von Willebrand syndrome: response to septal myectomy. Mayo Clin Proc 2011; 86 :219-224.
16Shen J. Von Willebrand factor in acute leukemia. Zhonghua Yi XueZa Zhi 1990; 70 :26-28.
17Lethagen S, Hillarp A, Ekholm C, Mattson E, Halldén C, Friberg B. Distribution of von Willebrand factor levels in young women with and without bleeding symptoms: influence of ABO blood group and promoter haplotypes. Thromb Haemost 2008; 99 :1013-1018.
18Gudmundsdottir B, Marder V, Onundarson P. Risk of excessive bleeding associated with marginally low von Willebrand factor and mild platelet dysfunction. J Thromb Haemost 2007; 5 :274-281.
19Nilsson G, Astermark J, Strandberg K, Wichert S, Bengt S, Erik B. Coagulation activation and changes during treatment of acute myeloid leukemia: a prospective cohort study. J Hematol Malig 2012; 2 :11-17.
20Strandberg K, Lethagen S, Andersson K, Carlsson M, Hillarp A. Evaluation of a rapid automated assay for analysis of von Willebrand ristocetin cofactor activity. Clin Appl Thromb Hemost 2006; 12:61-67.