Genetic and laboratory predictors of bleeding severity in FVII deficiency: insights from the PRO-RBDD and EN-RBD databases

Authors:
Publisher website
Google Scholar
Abstract
Factor VII deficiency (FVIID) is among the most common rare bleeding disorders. The ability of FVII coagulant activity (FVII:C) levels and genetic variants to predict clinical outcomes remain limited. We evaluated the discriminative performance of FVII:C levels, genotype-phenotype associations, and the accuracy of in silico tools in identifying pathogenic variants. In total, 422 patients from 3 international registries were analyzed. Overall, 61% experienced bleeding, predominantly spontaneous (73%); 20% had grade I, 27% had grade II, and 14% had grade III bleeding. Lower FVII:C levels showed a moderate inverse correlation with bleeding scores and were more common in individuals with spontaneous bleeding (r = −0.49; P< .0001). FVII:C levels showed moderate discrimination for asymptomatic status (area under the curve [AUC], 0.72) and strong discrimination for grade III bleeding (AUC, 0.84). Cutoffs of ≥19 IU/dLand <11 IU/dL identified asymptomatic individuals and those with grade III bleeding. We identified 90 variants, including 10 novel; 60% were missense, and 63% localized to the serine protease domain. According to the American College of Medical Genetics and Genomics criteria, 86% were pathogenic/likely pathogenic. The sensitivity of in silico tools was 96%, 85%, and 60% for REVEL, CADD, and AlphaMissense. Severe deficiency was more common in homozygous individuals (90%), whereas heterozygous individuals predominantly exhibited mild/moderate deficiency (92%). Consistently, bleeding phenotypes were frequent in individuals with homozygous/compound heterozygous variants, whereas most heterozygous variants were asymptomatic (71%). Overall, FVII:C levels moderately stratify bleeding severity. Homozygous/compound heterozygous variants are associated with lower FVII:C levels and more severe bleeding, whereas heterozygous variants are mostly asymptomatic. Among in silico tools, REVEL demonstrated the highest sensitivity.

This publication has 33 references indexed in Scilit: