Anticoagulants
Corey Diamond, PharmD
A May 2025 analysis in JAMA Internal Medicine directly compares apixaban, rivaroxaban, and warfarin across three major U.S. insurance databases and finds that apixaban may provide the most favorable balance of efficacy and safety in routine practice for patients with venous thromboembolism (VTE).
Direct oral anticoagulants (DOACs) have become the primary treatment choice for VTE, which encompasses deep vein thrombosis and pulmonary embolism. Clinicians now use apixaban and rivaroxaban most frequently, yet questions remain regarding their comparative safety and effectiveness.
According to the findings, apixaban consistently reduced both recurrent VTE and major bleeding compared with the other two agents.
The consistency across databases and subgroups also adds weight to the findings.
The investigators set hospitalization for recurrent VTE as the primary efficacy outcome, and hospitalization for major bleeding as the primary safety outcome. They also measured all-cause mortality and a composite outcome consisting of hospitalization for recurrent VTE, hospitalization for major bleeding, and all-cause mortality.
Results
The cohort included 163,593 patients with an average age of 71 years, of whom 57% were women. Apixaban accounted for 58.5% of initiations, rivaroxaban for 25.7%, and warfarin for 15.8%. Median follow up lasted 169 days. During this time, investigators recorded 3,270 hospitalizations for recurrent VTE and 4,229 for major bleeding.
Apixaban produced a 33% statistically significant reduction in recurrent VTE compared with warfarin. Compared with rivaroxaban, apixaban showed a 13% significant reduction in recurrence. Rivaroxaban lowered recurrence by 23% compared with warfarin.
For bleeding, apixaban led to a 30% significant reduction compared with warfarin, and a 31% significant reduction compared with rivaroxaban. Rivaroxaban and warfarin showed no significant difference in bleeding risk.
All-cause mortality appeared similar across the three groups, with no significant differences. The composite outcome of recurrent VTE, major bleeding, and all-cause mortality showed a 13% significant reduction for apixaban compared with warfarin and a 9% significant reduction compared with rivaroxaban.
Study design
The research team designed a new-user, parallel-group cohort study using Medicare, MarketScan, and Optum Clinformatics data from 2016 through 2024. They identified adults who filled a prescription for apixaban, rivaroxaban, or warfarin within 30 days after a VTE hospitalization. Each patient had at least 1 year of continuous insurance enrollment before the index prescription. The team excluded patients with prior oral anticoagulant use, end-stage kidney disease, hospice care, or those who initiated multiple anticoagulants on the same day.
The researchers defined exposure by the first filled prescription and required ongoing therapy, allowing a refill gap of up to 14 days. They tracked outcomes starting the day after initiation and continued until patients experienced an outcome, switched or discontinued therapy, unenrolled, died, or reached the end of available data.
The research team also examined differences in clinically important subgroups. Among younger adults, rivaroxaban showed lower bleeding risk compared with warfarin, though this pattern did not extend to older patients. In individuals with cancer, both apixaban and rivaroxaban were linked with higher mortality than warfarin, but the authors cautioned that this result may reflect residual confounding because administrative data lacked information on cancer stage and severity.
CKD and prior bleeding history did not alter the main conclusions, and across frailty categories, apixaban consistently reduced both recurrent VTE and major bleeding. These subgroup findings suggest that while apixaban offers the broadest benefit, treatment decisions still require tailoring to age, cancer status, and comorbidity profiles.
The authors highlight that “apixaban was associated with a lower risk of hospitalizations for recurrent VTE and major bleeding events than rivaroxaban or warfarin,” and they emphasize that this finding persisted across all analyses. They also acknowledged potential residual confounding due to incomplete data on cancer stage, adherence, and BMI.
The authors said that these results “provide evidence to guide the selection of appropriate initial oral anticoagulant regimens for adult patients with VTE.” ■