Pharmacogenomics of Rivaroxaban: Association of CYP3A4, CYP3A5, CYP2J2, ABCB1, and ABCG2 Variants with Bleeding and Thrombotic Outcomes in Real-World Clinical Practice.
To evaluate associations between polymorphisms in CYP3A4 (*1B, *22), CYP3A5 (*3), CYP2J2 (*7, rs11572325), ABCB1 (c.1236C>T, c.2677G>T/A, c.3435C>T, rs4148738) and ABCG2 (c.421C>A) and the occurrence of bleeding or occlusive events in patients receiving rivaroxaban in real-world clinical practice.
A nested case-control study, divided into two substudies (bleeding and thromboembolic events), was conducted within a prospective cohort of 385 adults receiving rivaroxaban at University Hospital Centre Zagreb (September 2021-September 2024). Bleeding events were classified per ISTH criteria, and genotyping was performed using TaqMan real-time PCR. Cases and controls were balanced using entropy balancing, and associations were estimated with Bayesian logistic regression under a skeptical prior N(0, 0.355); LASSO regression was used to identify clinical and genetic predictors of outcomes.
In total, 71 patients (18.4%) experienced bleeding events, most frequently gastrointestinal (47.9%), while 314 patients served as controls. No pharmacogenomic variant showed a clear association with bleeding risk (raw and balanced odds ratios 0.80-1.35; 95% credible intervals crossing 1.0). LASSO regression identified age (OR 2.00 per decade), gastrointestinal comorbidity (OR 8.77), and eGFR as the dominant predictors of bleeding. Twenty-one patients experienced occlusive events (15 venous, 6 arterial); however, the low event count precluded meaningful pharmacogenomic analysis.
Individual pharmacogenomic variants in CYP3A4, CYP3A5, CYP2J2, ABCB1, and ABCG2 together with pharmacogenetic-based phenotypes were not associated with clinically relevant bleeding in rivaroxaban-treated patients. Traditional clinical risk factors, particularly advanced age and gastrointestinal comorbidity, remain the dominant determinants of adverse outcomes. Routine pharmacogenomic testing to guide rivaroxaban dosing is not currently supported.
A nested case-control study, divided into two substudies (bleeding and thromboembolic events), was conducted within a prospective cohort of 385 adults receiving rivaroxaban at University Hospital Centre Zagreb (September 2021-September 2024). Bleeding events were classified per ISTH criteria, and genotyping was performed using TaqMan real-time PCR. Cases and controls were balanced using entropy balancing, and associations were estimated with Bayesian logistic regression under a skeptical prior N(0, 0.355); LASSO regression was used to identify clinical and genetic predictors of outcomes.
In total, 71 patients (18.4%) experienced bleeding events, most frequently gastrointestinal (47.9%), while 314 patients served as controls. No pharmacogenomic variant showed a clear association with bleeding risk (raw and balanced odds ratios 0.80-1.35; 95% credible intervals crossing 1.0). LASSO regression identified age (OR 2.00 per decade), gastrointestinal comorbidity (OR 8.77), and eGFR as the dominant predictors of bleeding. Twenty-one patients experienced occlusive events (15 venous, 6 arterial); however, the low event count precluded meaningful pharmacogenomic analysis.
Individual pharmacogenomic variants in CYP3A4, CYP3A5, CYP2J2, ABCB1, and ABCG2 together with pharmacogenetic-based phenotypes were not associated with clinically relevant bleeding in rivaroxaban-treated patients. Traditional clinical risk factors, particularly advanced age and gastrointestinal comorbidity, remain the dominant determinants of adverse outcomes. Routine pharmacogenomic testing to guide rivaroxaban dosing is not currently supported.
Authors
Slišković Slišković, Trkulja Trkulja, Ganoci Ganoci, Božina Božina, Pašara Pašara, Vrkić Kirhmajer Vrkić Kirhmajer, Palić Palić, Strikić Strikić, Narančić Narančić, Sopek Merkaš Sopek Merkaš, Merćep Merćep, Bulum Bulum, Šimičević Šimičević
View on Pubmed