AUTHORS
Amir Mahmoud, MBBCh, MSc, Amber Afzal, MD, MSCI2, Suhong Luo3, Kristen Marie Sanfilippo, MD, MPHS4
AFFILIATIONS
1 Washington University In Saint Louis, Saint Louis, Missouri, United States
2 Washington University In Saint Louis, Saint Louis, Missouri, United States
3 Washington University In Saint Louis, Saint Louis, Missouri, United States
4 Washington University In Saint Louis, Saint Louis, Missouri, United States
ABSTRACT
Background: Persons with cancer are at increased risk of both venous thromboembolism (VTE) as well as atrial fibrillation (AF), two common indications for anticoagulant (AC) therapy. Guideline recommendations favor direct oral anticoagulant therapy (DOAC) as the treatment of choice for most persons with cancer and VTE or AF. Apixaban and rivaroxaban are frequently prescribed, and both have randomized trial evidence supporting their use in cancer-associated VTE. Recently, the COBRRA trial compared the safety of apixaban versus rivaroxaban in persons with newly diagnosed VTE; however, cancer-specific head-to-head safety evidence remains limited and inconsistent across outcome definitions and follow-up periods. We aimed to assess the safety of apixaban versus rivaroxaban in persons with newly diagnosed cancer and recent initiation of anticoagulant therapy in a nationwide cohort.
Methods: We conducted a retrospective study using a nationwide cohort of United States Veterans with newly diagnosed solid tumors between 2012 and 2021 using ICD-9/10 codes. Patients were eligible if they had at least one new prescription (≥30 days) for AC therapy with apixaban or rivaroxaban within 30 days before or after cancer diagnosis. The primary outcome was clinically significant bleeding which was defined as bleeding events resulting in hospitalization. These events were ascertained using a previously validated ICD-code based algorithm. The association between anticoagulant type and clinically significant bleeding was measured using Fine and Gray competing risk analysis. The analyses were controlled for additional clinical factors associated with the propensity to bleeding including age, cancer type, metastatic disease, chemotherapy exposure, alcohol use, anemia (hemoglobin <10g/dL), prior clinically significant bleeding, renal function (eGFR), uncontrolled hypertension, liver disease, and prior stroke. SAS 9.4 was used; significance was P <0.05. Prior to any research activities, the study was approved by the Washington University and St. Louis Veterans Administration Medical Center IRBs.
Results: Among 6,284 patients with cancer and new prescriptions for AC therapy, 1,914 patients received apixaban [1197 (62.5%)] or rivaroxaban [717 (37.5%)] and were included in our analysis. Compared with rivaroxaban recipients, patients receiving apixaban were older (median age, 74.5 versus 71.6 years) and more frequently had uncontrolled hypertension (25.7% versus 17.3%) and severe renal impairment (eGFR <30) (3.2% versus 0.8%). Other baseline comorbidities, cancer characteristics, and prior bleeding were similar between groups. Median observed follow-up was 40 months. Overall, 127 bleeding-related hospitalizations occurred within 12 months, 82 (6.9%) in the apixaban group vs 45 (6.3%) in the rivaroxaban group. Median time to bleeding was 50.5 days (apixaban) versus 46 days (rivaroxaban). After adjustment, apixaban was not associated with a different hazard of 12-months bleeding-related hospitalization than rivaroxaban (sHR, 0.985; 95% CI, 0.678-1.432; P=.9375). Other significant predictors included metastatic disease (sHR, 1.897; 95% CI, 1.168-3.081), genitourinary cancer (sHR, 2.753; 95% CI, 1.732-4.376), and prior bleeding (sHR, 2.442; 95% CI, 1.364-4.373). At 30 days, there was also no statistically significant difference between both agents on adjusted analysis (sHR, 1.008; 95% CI, 0.697-1.459).
Conclusions: In this national VHA cohort of patients with cancer, the risk of bleeding-related hospitalization did not differ significantly between apixaban and rivaroxaban at either 30 days or 12 months. These findings did not identify a clear safety advantage for either agent; however, larger prospective head-to-head studies are needed to more precisely inform DOAC agent selection in patients with cancer.
DISCLOSURES
Amir Mahmoud, MBBCh, MSc: No relevant conflicts of interest to declare.