NEW ORLEANS – In routine practice, rivaroxaban was superior to low-molecular-weight heparin for venous thromboembolism prevention in older adults undergoing hip or knee arthroplasty, without an increase in bleeding risk.
Among 24,321 patients, aged 66 years or older, the 30-day VTE event rate was 0.47% for rivaroxaban (Xarelto) (61 events) and 0.81% for low-molecular-weight heparin (LMWH) (93 events).
These findings resulted in an unadjusted relative risk of 0.58, which was statistically significant (P = .001) and did not change after adjustment for significant covariates, Dr. Alejandro Lazo-Langner said during an antithrombotic therapy session at the annual meeting of the American Society of Hematology.
There were 23 major bleeding events in both the rivaroxaban (0.18%) and LMWH (0.20%) groups. The unadjusted relative risk was 0.89 (P = .700) and did not change after adjustment in the population-based, retrospective cohort analysis.
Rivaroxaban has been the subject of numerous randomized controlled trials, but "We don’t have much real-life data in these patients," said Dr. Lazo-Langner, a hematologist specializing in thromboembolic diseases at Western University, London, Ontario, Canada.
In a meta-analysis of eight randomized rivaroxaban trials, the factor Xa inhibitor was associated with a significant 52% reduction in thrombosis after total hip or knee replacement, compared with enoxaparin (Lovenox) in roughly 14,000 patients (BMJ 2012;344:e3675 [doi:10.1136/bmj.e3675]). This came at a cost, however, of a slightly increased risk of clinically significant bleeding (relative risk, 1.25), he noted.
For the current analysis,the investigators used linked health care databases in Ontario to identify 24,321 patients who received an outpatient prescription for rivaroxaban or subcutaneous LMWH including dalteparin (Fragmin), tinzaparin (Innohep), or enoxaparin on discharge after total hip or knee arthroplasty between 2002 and 2012 across 121 hospitals.
Their average age was 74 years, 59% were women, and 12,850 received rivaroxaban. The anticoagulants were prescribed for a median of 14 days. Patients were excluded if they had other indications for anticoagulation, prosthetic mechanical heart valves, required dialysis, or lived in a long-term-care facility.
At 90 days, the VTE event rate was significantly lower in the rivaroxaban group than in the LMWH group (0.71% vs. 1.20%; adjusted RR, 0.59; P = .001). Once again, major bleeding events were similar (0.24% vs. 0.27%; adjusted RR, 0.63; P = .138), Dr. Lazo-Langner reported.
No differences were observed between the two groups at 30 or 90 days for hospitalization with endoscopy or hospitalization with major bleeding or endoscopy.
All-cause mortality was not estimable at 30 days but was lower with rivaroxaban at 90 days (16 deaths vs. 25 deaths; adjusted RR, 0.52; P = .058).
Additional analyses were conducted to test the robustness of the findings and no differences in rates of thrombosis were found when the analysis was restricted to 2009 to 2012, by type of joint replacement, or different low-molecular-weight heparins, he said.
Finally, a cost analysis was performed that showed a modest, but significant increase in direct drug costs to patients prescribed LMWH rather than rivaroxaban (Canadian $242 vs. $228; P less than .001) and home-care costs, likely from increased nursing ($1,082 vs. $959; P less than .001), Dr. Lazo-Langner said.
During a discussion of the results, he said there was no difference in novel anticoagulant use across surgeons or hospital settings, which was academic for 21% of LMWH patients and 15% of rivaroxaban patients.
Session comoderator Dr. Elaine Hylek, professor of medicine at Boston University, called this reassuring, but also urged caution in extrapolating conclusions on treatment effect outside a randomized trial.
Dr. Lazo-Langner reported research funding from Alexion, serving as a speaker for Pfizer, and honoraria from Pfizer, Leo Pharma, and Boehringer Ingelheim.