Background: In patients with bleeding, von Willebrand disease (VWD) is diagnosed at von Willebrand factor (VWF) levels of ≤50 IU/dL ( James et al. Blood Adv. 1:280-300. 2021). Nevertheless, the majority of patients with a mild to moderate bleeding disorder (MBD) is diagnosed with bleeding disorder of unknown cause (BDUC), upon exclusion of VWD, platelet function defects (PFD), and coagulation factor deficiencies (CFD). Bleeding severity in patients with VWD does not correlate well with VWF levels in type 1 VWD, suggesting other modifiers of the bleeding tendency, when patients with severe VWD are excluded. Blood group O (BG O) is associated with lower VWF levels and has been found to be overrepresented in VWD. In BDUC patients, BGO is also overrepresented and recognized as an independent risk factor for bleeding severity and recurrent bleeding ( Mehic et al. Blood Adv.20: 5157-5164. 2020; Mehic et al. JTH. 7: 1757-1768. 2023). Aim: This study aimed to comprehensively investigate the clinical phenotype of patients with VWD or BDUC from the Vienna Bleeding Biobank (VIBB) and evaluate the direct effect of ABO BG on bleeding severity beyond its effect on VWF levels. Methods: Patients with VWD or BDUC from the VIBB, a single-center cohort study (EC 603/2009), were categorized by lowest VWF, as measured by VWF:antigen (VWF:Ag) or VWF ristocetin co-factor activity (VWF:RCo), into four groups: <30 IU/dL (very low VWF), 30-50 IU/dL (low VWF), 51-80 IU/dL, and >80 IU/dL. Patients diagnosed with PFD (n=221), CFD (n=28), or other diagnoses (n=4) were excluded. To study the association of BG O with bleeding severity irrespective of VWF levels, we fitted linear regression models adjusting for age, sex, and lowest VWF. We allowed for non-linearity by modelling age and sex with restricted cubic splines. Results: A total of 637 patients, 18 (2.8%) with very low VWF, 55 (8.6%) with low VWF, 237 (37.2%) with VWF levels ranging from 51 to 80 IU/dL and 327 (51.3%) with VWF levels >80 IU/dl, were included in this analysis, as summarized in Table 1. BG O was more prevalent in patients with very low VWF levels (61%), low VWF levels (71%), and VWF levels 51-80 IU/dL (66%) than in patients with VWF levels >80 IU/dL (30%). Patients with low VWF were youngest, followed by very low VWF and those with VWF 51-80 IU/dL. Age was highest in the patient group with VWF levels >80 IU/dL at clinical presentation and might influence VWF levels and bleeding scores. With regard to bleeding severity, patients with very low VWF levels had highest median bleeding scores or number of bleeding manifestations, while the bleeding severity was similar amongst the other three categories (Table 1). VWF levels did not correlate with bleeding severity scores (Vicenza score: Spearman rho=-0.03, p=0.53; ISTH BAT: rho=-0.02, p=0.75; number of bleeding manifestations: rho=-0.02, p=0.66). In linear regression, holding age, sex, and lowest VWF levels constant, predicted mean bleeding scores and number of bleeding manifestations were higher with BG O (Vicenza score mean difference [MD], 0.64; 95% confidence interval [95%CI], 0.12 to 1.25; p=0.015; ISTH BAT MD 0.52; 95%CI, -0.25 to 1.28; p=0.18; number of bleeding manifestations MD, 0.24; 95%CI, -0.06 to 0.54; p=0.12) (Figure 1). Conclusion: Our findings demonstrate an overlap between VWD and BDUC with a similar clinical phenotype and bleeding phenotype, but increased severity in patients with VWF levels <30 IU/dL. Patients with BG O had increased bleeding scores compared to those with BG non-O. BG O appears to be an independent factor in modifying bleeding severity in both patients with VWD or BDUC across all VWF ranges.