乌鸦传媒

SAN DIEGO -- A closely watched trial of venous thromboembolism (VTE) prevention with an oral anticoagulant in cancer patients fell short of a clear win, but not because the drug didn't work, according to a report here.

The placebo-controlled of rivaroxaban (Xarelto) in the outpatient setting failed to significantly reduce 180-day VTE rates, the study's primary endpoint. However, 44% of patients assigned to oral rivaroxaban stopped taking the drug before the 180-day cutoff, as did more than half of the patients in the placebo arm. An analysis of VTE incidence during actual on-treatment time did produce a statistically significant 60% reduction in VTE with rivaroxaban vs placebo, as reported at the meeting.

"It's not unexpected, because we know that if patients don't take the drug, obviously it's not going to prevent a clot," said Alok Khorana, MD, of the Cleveland Clinic.

On average, patients were on treatment for about 4.5 months of the 6-month primary outcome period.

"If you just look at the results out to month 4, you see a much greater suppression of VTE, but then the risk goes up again," he added.

'A Pretty Big Deal'

Despite missing the primary endpoint, the CASSINI trial is still "a pretty big deal," said Mark Crowther, MD, of McMaster University in Hamilton, Ontario.

Most patients with cancer receive injectable low molecular weight heparins (LMWH) for VTE prevention, but an oral treatment would be preferable. "You're really avoiding the complexity but also an event that is very unpalatable to patients undergoing cancer therapy, Crowther said. "It represents a major advance."

"VTE is very common in patients undergoing treatment for cancer, and in an ideal world, would be completely or largely preventable," Crowther explained. "We've never had good quality evidence that any intervention that is easy to use, relatively inexpensive compared to the cost of cancer therapy, and tolerable to patients can actually reduce the risk of VTE."

Two large randomized trials of cancer outpatients showed that thromboprophylaxis with LMWH reduced the rate of VTE. However, both trials had low event rates, leading to relatively high number-needed-to-treat (NNT) values of and to prevent one VTE episode. Use of a validated scoring system to risk-stratify patients could help target prophylaxis to patients who are most likely to benefit from the treatment, a strategy that has been evaluated in several studies, Khorana said.

In the CASSINI trial, investigators from the U.S., Canada, and Germany evaluated rivaroxaban prophylaxis in 841 ambulatory adult patients initiating systemic therapy for various types of cancer. Eligible had no evidence of deep-vein thrombosis (DVT) at enrollment screening but had an increased risk of VTE, defined as a ≥2. Patients were randomized to rivaroxaban or placebo, and treatment continued for a maximum of 180 days, followed by an additional 30-day follow-up. Patients had lower-extremity ultrasound assessments for DVT every 8 weeks.

The trial had a composite primary endpoint consisting of symptomatic or asymptomatic proximal DVT, symptomatic upper- or lower-extremity distal DVT, symptomatic or incidental pulmonary embolism, and VTE-related death. The primary safety endpoint was major bleeding episodes, as defined by International Society on Thrombosis and Haemostasis criteria. A blinded independent review committee adjudicated all endpoints.

Trial Results

The primary data analysis was the intent-to-treat (ITT) population followed for up to 180 days. A supplemental analysis focused on VTE in patients on treatment.

The ITT analysis showed VTE rates of 8.79% in the placebo group and 5.95% in the rivaroxaban group. The difference represented a 34% risk reduction in favor of rivaroxaban, but fell short of statistical significance (RR 0.66, 95% CI 0.40-1.09, P=0.101).

Khorana said that almost 40% of all VTEs occurred in patients who stopped taking their assigned therapy. The supplemental analysis of patients on treatment showed the VTE rate was more than twice as high in the placebo group, 6.41% vs 2.62%, a difference that did achieve statistical significance (95% CI 0.20-0.80, P=0.007).

Analysis of a secondary outcome comprising the composite efficacy endpoint plus all-cause mortality showed a significant advantage in favor of rivaroxaban (23.1% vs 29.5%, P=0.03). The safety analysis showed no significant difference in the frequency of major bleeding (0.99% with placebo, 1.98% with rivaroxaban, P=0.265) or clinically relevant nonmajor bleeding (1.98% vs 2.72%, P=0.532).

A calculation of NNT and number needed to harm (NNH), a reflection of the risks and benefits of rivaroxaban prophylaxis, yielded an NNT of 35 for the 180-day primary efficacy analysis. The NNT declined to 26 during the period patients remained on treatment. After combining the primary efficacy endpoint, the NNT declined to 20 in treated patients.

The NNH for patients who remained on treatment was 101 for one major bleeding event, the primary safety endpoint. The NNH for clinically relevant nonmajor bleeding was 135.

"On balance, it looks like there would be a benefit for patients from this primary prevention approach," said Khorana.

A Khorana score ≥2 did identify a group of patients with an increased risk of thrombosis. Considering patients who were disqualified at baseline because of an existing clot, on-study VTE, secondary VTE-related endpoints, and arterial clots, the total thrombosis rate reached 17% in the placebo group.

"Because patients with VTE at baseline accounted for about a third of the total VTE events, baseline screening of cancer outpatients might be worth considering," Khorana concluded. A phase III trial evaluating a risk-adjusted strategy for VTE prevention is ongoing.

Comment