CASSINI Trial Assesses Rivaroxaban for Thromboprophylaxis in Cancer Patients

In a phase IIIb trial, researchers evaluated the efficacy of an oral anticoagulant in cancer patients at high risk for venous thromboembolism.

In the multicenter, double-blind, randomized, phase IIIb CASSINI trial, rivaroxaban, an oral anticoagulant, failed to show efficacy as a thromboprophylaxis agent in cancer patients at high risk for venous thromboembolism. The findings were recently published in the New England Journal of Medicine.

Despite the negative findings, Jacob Mandel, MD, an assistant professor of neurology and member of the Dan L Duncan Comprehensive Cancer Center at Baylor College of Medicine, told Cancer Network that the study still has merit. “Even though [the trial] didn’t hit its primary endpoint, it did seem to definitely trend in that direction, with patients on the drug having less events than the placebo group,” he said.

The trial included adult patients with a solid tumor or lymphoma who were at high risk for venous thromboembolism, as determined by a Khorana risk score of 2 or higher. Patients with a brain tumor or brain metastases were not eligible for enrollment, and patients were screened for deep vein thrombosis (DVT) and excluded if one was found.

A total of 841 patients were randomly assigned in a 1:1 fashion to preventative treatment with rivaroxaban 10 mg or placebo once daily for up to 180 days. The primary efficacy endpoint was a composite measure of symptomatic or asymptomatic proximal DVT in a lower limb, symptomatic DVT in an upper limb or distal DVT in a lower limb, symptomatic or incidental pulmonary embolism, and death from venous thromboembolism. The primary safety endpoint was incidence of major bleeding.

The most common cancer in patients was pancreatic cancer (32.6%), and more than half of patients (54.5%) had a solid tumor with metastatic disease. The dropout rate was high for both arms, with 43.7% of patients on the rivaroxaban arm and 50.2% on the placebo arm dropping out while on trial.

In all, 25 of 420 patients (6.0%) on the rivaroxaban arm had an event compared with 37 of 421 patients (8.8%) on the placebo arm. The result translated to patients on the rivaroxaban arm having a 34% lower risk of DVT or death from DVT, but the difference lacked statistical significance (hazard ratio [HR], 0.66; 95% CI, 0.40–1.09; P = .10).

As for the primary safety endpoint, 8 of 405 patients (2.0%) on the rivaroxaban arm had major bleeding compared with 4 of 404 (1.0%) on the placebo arm (HR, 1.96; 95% CI, 0.59–6.49). Although the incidence of major bleeding was higher on the rivaroxaban arm, Mandel said the difference was within a “reasonable range” to consider using rivaroxaban.

Overall, the CASSINI trial missing the primary efficacy endpoint highlights the difficulty of conducting trials for interventions that address complications to cancer, not the cancer itself.

“When you’re looking at trials that are not necessarily meant to treat the cancer but to prevent a complication, a lot of times patients don’t have the same commitment to doing those trials as trials with interventions,” Mandel said. The high dropout rate in both arms, as well as the heterogeneity of the patient population, he said, could have affected the results. He explained that trials like CASSINI may be designed to include patients with different cancer types, stages of disease, and chemotherapy regimens so that enough patients will enroll on trial, but the tradeoff is having a heterogeneous population.

As for what this trial means for clinical practice, Mandel said the decision to recommend an anticoagulant like rivaroxaban still has to be “individualized” and based on the patient’s cancer type, stage of disease, and chemotherapy regimen.