The CASSINI trial found that treatment with rivaroxaban did not significantly reduce the incidence of thromboembolism or death caused by thromboembolism in high-risk ambulatory patients with cancer during the 180-day study period. However, during the study’s intervention period, there was a significantly lower incidence of venous thromboembolism and death due to thromboembolism among patients treated with rivaroxaban, and there was a low incidence of major bleeding. The results of the trial were reported by Alok A Khorana (Department of Hematology and Medical Oncology, Cleveland Clinic, Cleveland, USA) and colleagues in the New England Journal of Medicine (NEJM).
It is known that cancer and cancer treatment put patients at risk of venous thromboembolism, but guidelines do not recommend routine thromboprophylaxis for patients with cancer even though it is possible that such treatment could improve their outcomes. The study investigators therefore sought to “assess the efficacy and safety of rivaroxaban thromboprophylaxis in patients with a solid tumour or lymphoma who had a Khorana score of two or higher and were initiating in a new systemic cancer regimen”.
The CASSINI trial was a double-blind, multicentre, randomised trial, which randomised high-risk ambulatory patients with cancer to receive either 10mg rivaroxaban or placebo daily for a period of up to 180 days.
The primary efficacy endpoint—a composite of deep vein thrombosis, pulmonary embolism or death from venous thromboembolism—was assessed for up to 180 days. The same endpoint was similarly assessed during the intervention period, which was the time from first receiving treatment or placebo to the last treatment plus two days. The primary safety endpoint was major bleeding.
Of 1,080 patients originally enrolled in the trial, 49 (4.5%) were found to have thrombosis during screening and 190 had other reasons to be excluded. A total of 841 patients who were all assessed as having a higher risk of venous thromboembolism (Khorana score ≥2 at baseline) and an expected survival of more than six months were included in the randomisation: 420 patients were assigned to the rivaroxaban group and 421 to the placebo group.
The mean intervention period—the time when patients actually received treatment—was 4.3 months, with 43.7% of patients in the rivaroxaban group and 50.2% of those in the placebo group stopping treatment before the end of the trial.
During the study period up to day 180, the primary endpoint occurred in 25 patients receiving rivaroxaban (6%) and 37 patients receiving placebo (8.8%; hazard ratio=0.66; 95% confidence interval [CI]=0.40–1.09; p=0.10). During the intervention period, the primary endpoint occurred in 11 patients receiving rivaroxaban (2.6%) and 27 patients receiving placebo (6.4%; hazard ratio=0.40; 95% CI=0.20–0.80). Major bleeding occurred in eight patients receiving rivaroxaban (1.9%) and four patients receiving placebo (1%; hazard ratio=1.96; 95% CI=0.59–6.49).
Khorana et al comment: “Although the primary endpoint occurred in a lower percentage of patients who had been randomly assigned to the rivaroxaban group in this analysis, the difference was not significant”. However, they note that “in a prespecified supportive analysis involving the same population but assessing the more conventional period of during the intervention, we found a difference of four percentage points in favour of rivaroxaban over placebo with regard to the primary composite endpoint of venous thromboembolism and venous thromboembolism-related death.”
Arterial and isolated distal thromboembolism, which are common in patients with cancer, were found in lower rates among the rivaroxaban-treated patients than in the patients receiving placebo. According to Khorana et al, this “could further increase the net benefit of prophylaxis for patients”.
A limitation of this study is the large percentage of patients who stopped the trial regimen prematurely, but this is not surprising in patients who mostly have advanced cancer.
The investigators conclude: “The results of the CASSINI trial provide important information regarding the baseline prevalence and incidence of thromboembolism among high-risk ambulatory patients with cancer”. However, they were unable to confirm the benefits of treating these patients with rivaroxaban in their trial “because the between-group difference in the prespecified primary efficacy endpoint up to day 180 was not significant”.
Khorana noted to Interventional News that there was a second confirmatory trial in the same issue of the NEJM by Marc Carrier (Ottawa, Canada) et al investigating the same patient population as observed in the AVERT trial. In an accompanying editorial addressing both studies, Giancarlo Agnelli (Perugia, Italy) stated that both trials together “showed a significant benefit of direct oral anticoagulants for the prevention of venous thromboembolism, with a low incidence of major bleeding” and the data are “quite compelling”. Khorana agrees with this editorial leaning.