The use of direct oral anticoagulants for the prevention of venous thromboembolism (VTE) in high-risk ambulatory cancer patients showed promising results in two clinical trials published online in the New England Journal of Medicine February 21.
But should the preventive therapy now be used among high-risk patients who are out and about in the world?
In an accompanying editorial, Giancarlo Agnelli, MD, Internal Vascular and Emergency Medicine–Stroke Unit, University of Perugia, Italy, has some doubts, even though there was a low incidence of major bleeding, a concerning adverse event.
Agnelli believes the evidence from these two trials is “quite compelling,” but acknowledges that some physicians may still be reluctant to change their practice. That’s because there are gaps in these trials that may limit their generalizability, he says.
Trial of Xarelto
The CASSINI trial assessed the efficacy and safety of rivaroxaban (Xarelto, Janssen) thromboprophylaxis in patients with a solid tumor or lymphoma, who had a Khorana score of 2 or higher, and were beginning outpatient chemotherapy.
A total of 841 patients without deep vein thrombosis (DVT) at screening were randomly assigned to receive rivaroxaban (at a dose of 10 mg) or placebo daily for up to 180 days, with screening every 8 weeks. The primary efficacy endpoint was a composite of objectively confirmed symptomatic or asymptomatic lower-extremity proximal DVT, symptomatic upper- or lower-extremity distal DVT, symptomatic or incidental pulmonary embolism, and VTE-related death.
At 180 days, the rate of thrombotic events was not statistically significantly different between the patients who received rivaroxaban (5.95%) and those who received placebo (~8.79%). But this was primarily due to the large proportion of patients who had stopped taking the medication before the end of the study period, explained lead author Alok A. Khorana, MD, professor of medicine, Cleveland Clinic Lerner College of Medicine in Ohio.
More than half of patients who were randomly assigned to receive placebo and almost 44% of those taking rivaroxaban stopped the regimen before 180 days, and more than one third of clotting events occurred after participants had discontinued their assigned regimen.
However, in a prespecified analysis that only evaluated the on-treatment period, the difference between the two groups was statistically significant. In the rivaroxaban group, 2.6% of patients experienced an event, as compared with 6.4% of patients who received placebo (P = .007). In addition, rivaroxaban treatment was associated with a low incidence of major bleeding: 8 of 405 patients (2%) in the rivaroxaban group and 4 of 404 (1%) in the placebo group (hazard ratio [HR], 1.96, 95% confidence interval [CI], 0.59 – 6.49).
Trial of Apixaban
In the AVERT trial, the efficacy and safety of another direct oral anticoagulant was evaluated in a similar patient population. In this study, 574 patients were randomly assigned to apixaban (Eliquis, Bristol-Myers Squibb) given at 2.5 mg twice daily or placebo. All patients were in the ambulatory setting, were at intermediate-to-high risk for venous thromboembolism (Khorana score 2 or higher), and were initiating chemotherapy.
The primary efficacy outcome was objectively documented venous thromboembolism over a follow-up period of 180 days, while the main safety outcome was a major bleeding episode.
Of the 563 patients who were included in the modified intention-to-treat analysis, VTEs were reported in 12 (4.2%) of 288 patients who received apixaban vs 28 (10.2%) of 275 patients in the placebo group (HR, 0.41; 95% CI, 0.26 – 0.65). The study met its primary endpoint as this was a significantly lower rate of venous thromboembolic complications as compared with placebo (P < .001).
In the modified intention-to-treat analysis, major bleeding occurred in 10 patients (3.5%) in the apixaban group and 5 patients (1.8%) in the placebo group (hazard ratio, 2.00; 95% CI, 1.01 – 3.95; P = .046). During the treatment period, major bleeding occurred in 6 patients (2.1%) in the apixaban group and in 3 patients (1.1%) in the placebo group (HR, 1.89; 95% CI, 0.39 to 9.24).
When looking at the safety endpoint, the researchers found that the rate of major bleeding was significantly higher with apixaban vs placebo (3.5% and 1.8%, respectively; HR, 2.00), but the rate was not significantly higher with apixaban in the analysis of outcomes made during treatment period (2.1% apixaban group, 1.1% placebo group; HR, 1.89). The differences in major bleeding complications between groups was due primarily to higher rates of gastrointestinal bleeding, hematuria, and gynecologic bleeding associated with apixaban.
Definitive Evidence Needed
In his editorial, Agnelli points out that several trials have evaluated the use of low-molecular-weight heparins (LMWH) in ambulatory patients with different types of metastatic or locally advanced solid cancer being treated with chemotherapy. Patients treated with LMWHs had an approximately 50% lower risk of symptomatic VTE as those who received placebo.
However, the incidence of symptomatic VTE in the placebo group as well as the absolute difference in risk between groups were considered too low to recommend anti-thrombotic prophylaxis. “Several international guidelines suggest that antithrombotic prophylaxis be considered only in high-risk patients,” he notes.
Several strategies have been proposed to identify cancer patients at high risk of VTEs, including the use of the Khorana score (a risk-assessment algorithm incorporating cancer type developed in 2008 by the CASSINI trial lead author), pretreatment hematologic factors, and body-mass index to quantify risk.
As for the two recent studies, Agnelli writes that the findings are “quite reassuring,” but the combined data did not show a significant difference in mortality between patients who received a direct oral anticoagulant and those on placebo.
As for the effect on clinical practice, “Some clinicians could still be reluctant to change their practice,” Agnelli writes, as “some of the most common cancers, such as colorectal, breast, and prostate cancers, were underrepresented in the two trials.”
The Khorana score, which was the “cornerstone of the two trials,” is not consistent in all cancer types, and performs poorly in some, including lung cancer. In addition, the type of chemotherapy administered is not accounted for in the Khorana score.
“All these considerations may limit the generalizability of the AVERT and CASSINI trials, and some clinicians may consider that data on individual cancer types or individual chemotherapy agents are required before prophylaxis can be generally accepted,” writes Agnelli.
He also notes that the percentage of patients who continued on treatment for the entire study period was relatively low, and this could not be explained by the expected deaths in this population. “This finding confirms the complexity of treating patients with cancer, regardless of the route of administration.” Agnelli writes.
Overall, the AVERT and CASSINI trials demonstrated that thromboprophylaxis with these agents was safe and effective. But “trials involving patients with individual types of cancer would provide the definitive evidence about the clinical benefit associated with prophylaxis with direct oral anticoagulants in ambulatory patients with cancer,” Agnelli concludes.
The CASSINI trial is supported by Janssen; Bayer; the Sondra and Stephen Hardis Chair in Oncology Research; the National Heart, Lung, and Blood Institute; the Cleveland Clinic Center of Excellence for Cancer-Associated Thrombosis; and the Porter Family Fund. The AVERT trial was funded by the Canadian Institutes of Health Research and Bristol-Myers Squibb–Pfizer Alliance. A full list of disclosures from the CASSINI trial and AVERT trial coauthors is available with the full text of each article. Agnelli reports personal fees from Pfizer outside the submitted work.