Genotypic and Phenotypic Analysis of Adolescents with Heavy Menstrual Bleeding and Low Von Willebrand Activity - Interim Report of a Multi-Center Study

Introduction:

Low Von Willebrand factor (VWF) activity, considered a bleeding risk factor, is prevalent in some adolescents with heavy menstrual bleeding (HMB), wherein upfront hemostatic therapy may not be readily considered. There is a need to better define this patient subset phenotypically and genotypically, to stratify their risk to bleed, and to tailor their therapy in hopes of preventing complications. In adolescents with HMB and low VWF, we hypothesized that a significant proportion will have disease-causing sequence variations in the VWF gene and/or modifier genes that affect hemostasis, thrombosis or vascular biology, and these will correlate with their bleeding phenotype.

Methods: The objectives of this multi-center, single arm, observational cohort study were to 1) study the genotype of adolescent females with HMB and low VWF (≥ 30 and ≤ 50 IU/dL), 2) correlate genotype with bleeding phenotype by Pictorial Blood Assessment Chart (PBAC) score and ISTH bleeding assessment tool (BAT) score. Post-menarchal females < 21 years, with HMB (defined as PBAC score >100) and low VWF were eligible for the study. Patients who did not meet these criteria or diagnosed with other bleeding disorders were ineligible. Members of the Foundation for Women and Girls with Blood Disorders are participating centers in the study. Clinical phenotype data including HMB characteristics, PBAC, BAT, response to desmopressin challenge, management details, clinical outcomes, and laboratory values were obtained. Blood samples were collected for analysis of a 142 gene array that includes VWF, genes involved in hemostasis, thrombosis and vascular biology. DNA sequencing of all exons and intron/exon boundaries was performed; variants were called presumably pathogenic if categorized as damaging by the pipeline with allele frequency <1% in databases of human genetic variation (1,000 genomes and ExAC). The prevalence of sequence variations in the VWF and other modifier genes were estimated with an exact, 95% Binomial confidence. PBAC and BAT scores were compared between groups (mutation vs. wild-type) using the Wilcoxon rank sum test.

Results: 63 subjects were enrolled to date; 1 subject was later found to be ineligible due to detection of another bleeding disorder. The mean age was 16 years (range 11.5-20.2). The median BAT score was 5.0 (N=61; 2-20), median PBAC score was 481 (N=62; 114-8150). 11/48 (23%) had hemoglobin <12 gm/dl and 31/47 (66%) had ferritin <20 ng/ml at HMB diagnosis. 5/52 (10%) were hospitalized for HMB and 8/51 (16%) received PRBC transfusions. 57/62 (92%) underwent DDAVP challenge (intranasal, N=52; intravenous, N=5), with a mean rise of VWF activity at 1 hour of 3.0-fold (SD=1.3) and at 3-4 hours of 2.4-fold (SD=1.2), post-DDAVP. Genetic analysis in 51 subjects showed pathogenic gene variants: VWF - 9 in 7 subjects (14%; 95% CI: 6, 26), platelet (PLT) genes - 19 in 16 subjects (31%; 95% CI: 19, 46), coagulation factor (CF) genes - 10 in 7 subjects (14%; 95% CI: 6%, 26%). Two subjects had 2 VWF variants each, 3 had 2 PLT gene variants each and 1 had 4 CF gene variants; 1 had all 3 gene variants. Rest with VWF variants did not have co-existing PLT or CF gene variants. The median PBAC and BAT scores in subjects with/without VWF variants were 268/675 (P=0.06) and 4.5/5.5 (P=0.39), respectively. The median PBAC and BAT scores in subjects with/without PLT or CF genes and any variant were 561/525 (P=0.79) and 5.5/5.0 (P=0.77), and 380/777 (P=0.19) and 5/5 (P=0.94) respectively.

Conclusion: Our study confirms the feasibility of a multi-center study in adolescent females with HMB and low VWF. All subjects had significant bleeding phenotype with elevated BAT and PBAC scores, with complications including anemia, iron deficiency, transfusion requirement, and hospitalization. Response to DDAVP challenge in subjects tested was good and sustained. Potential pathogenic gene variants, not only in VWF, but also in CF and PLT genes were found in 51% of the subjects, which may account for their bleeding phenotype. The separate or the combined presence of VWF, PLT and CF damaging gene variants does not appear to correlate with the subjects' bleeding severity in this interim analysis. A larger sample size and further analysis of gene variants may provide more information regarding the phenotype/genotype correlation in this patient population.

Study supported by an investigator-initiated research grant from Baxalta US Inc., now part of Shire