Presenting Author:

Tanima De, Ph.D.

Principal Investigator:

Minoli Perera, Pharm.D.

Department:

Pharmacology

Keywords:

warfarin, bleeding, gene, single nucleotide polymorphism, African American

Location:

Third Floor, Feinberg Pavilion, Northwestern Memorial Hospital

B155 - Basic Science

Novel genomic predictor of warfarin related bleeding in African Americans

Hemorrhage is a major concern for patients treated with warfarin. The relative risk of major bleeding from warfarin is 44% higher in patients of African descent than of other ethnicities. The commonly studied genetic variants associated with bleeding explain substantially less variability in African Americans (AAs). Uncovering the causal variants is critical for safer use of the drug and drug selection. We conducted the first genome-wide association study in 215 AAs to identify the genetic predictors of warfarin related bleeding requiring hospitalization at International Normalized Ratio (INR) <4, a range in which most physicians do not consider warfarin reversal. To determine if clinical factors could further help to predict bleeding risk, the predictive performance of HAS-BLED (hypertension, abnormal renal/liver function, stroke, bleeding history, labile INR, elderly, drugs/alcohol) scheme was evaluated by Cox proportional hazards analysis and C-statistics. We identified a variant (rs78132896) on chromosome 6 which increased bleeding risk by 8.3 folds (p < 8.89E-07). It is a regulatory variant, present on the promoter region of EPHA7 gene which belongs to the family of ephrin receptor tyrosin kinases. These kinases support thrombus growth and stability by regulating integrin outside-in signaling in platelets. rs78132896 has been reported only in people of African ancestry and is absent in other populations. Validation of our finding in 188 AAs confirmed this association (p=0.002). A HAS-BLED score of ≥3 was associated with 2-fold risk of major bleeding (p=0.02). rs78132896 inclusion to HAS-BLED improved major bleeding risk prediction by correctly reclassifying 12.5% of major bleeding, thereby increasing sensitivity of prediction from 77.5% to 90% (C-indices: 0.65 vs. 0.70 for HAS-BLED vs. HAS-BLED+rs78132896 respectively). The number needed to genotype to avoid misclassifying patients as bleeder or non-bleeder from warfarin therapy was calculated to be 9. CYP2C9 star alleles showed no significant association with warfarin related bleeds. Our findings may help to predict clinically avoidable bleeding events and improve the safety of anticoagulation therapy in AAs.