Real-World Study Largely Confirms Effects of Apixaban, Rivaroxaban for A-fib

An analysis of claims data shows that both rivaroxaban and apixaban carry lower risks of intracranial hemorrhage (ICH) versus warfarin when used in everyday practice for patients with nonvalvular A-fib. Only rivaroxaban was associated with a reduced risk of the composite of ICH and ischemic stroke, however; in fact, the rate of ischemic stroke was numerically—though not significantly—higher with apixaban than with warfarin.

Care must be taken when interpreting results derived from claims data, but the advantage in terms of ICH for the two factor Xa inhibitors is reassuring because it’s consistent with what was seen in the phase III trials, according to Craig Coleman, PharmD (University of Connecticut, Storrs, CT), who presented the findings on April 17, 2016, at the European Cardiac Arrhythmia Society Congress in Paris, France.

“In my personal opinion, one of the major reasons that [novel oral anticoagulants] have become frequently used in nonvalvular atrial fibrillation patients is this reduction in risk of intracranial hemorrhage,” he told TCTMD.

The ischemic stroke findings showed concordance between the real world and the pivotal ROCKET AF trial for rivaroxaban (Xarelto; Janssen Pharmaceuticals), but the suggestion of a slightly higher rate of stroke with apixaban (Eliquis; Bristol-Myers Squibb) relative to warfarin was not consistent with phase III results and was a bit surprising, Coleman said.

He said the explanation for the difference is unclear, but speculated that it could have to do with greater use of reduced-dose apixaban in everyday practice compared with what was seen in the ARISTOTLE trial or poorer adherence related to twice-a-day dosing. By contrast, use of a reduced dose in real-world practice with rivaroxaban was roughly equivalent to that used in the registration trial.

He cautioned, however, that the claims analysis included only about 4,000 apixaban patients, less than the roughly 9,000 included in ARISTOTLE.

Commenting on the study for TCTMD, Matthew Reynolds, MD (Lahey Hospital & Medical Center, Burlington, MA), warned that “you have to be really careful how you interpret” the ischemic stroke finding with apixaban.

He pointed out that the absolute rates of ischemic stroke were nearly identical with rivaroxaban (0.54% per year) and apixaban (0.56% per year), but that the rate among warfarin patients was much higher in the analysis of rivaroxaban (0.83% vs 0.51%), providing a more favorable comparison. That discrepancy is likely due to chance, he added.

“There is no suggestion of harm here,” Reynolds said. “I think that would be a very dangerous conclusion to reach from these data.”

The main messages to take away from the study, he said, are that the findings are relatively consistent with what was seen in the pivotal trials and that event rates overall are very low. “I think it’s confirmatory data. I think it’s reassuring,” Reynolds noted.

For the REVISIT-US study, Coleman and colleagues examined US MarketScan claims data for patients newly started on rivaroxaban, apixaban, or warfarin for nonvalvular A-fib between January 2012 and October 2014. After exclusion of patients with a prior history of stroke, systemic embolism, or ICH and propensity-score matching, there were 11,411 patient pairs for the comparison of rivaroxaban and warfarin and 4,083 pairs for the comparison of apixaban and warfarin.

The researchers focused on ICH and ischemic stroke because those are both events that can result in prolonged neurological deficit or death, Coleman said.

Compared with warfarin, rivaroxaban was associated with a reduction in ICH (0.49% vs 0.96% per year; HR 0.53; 95% CI 0.35-0.79), a nonsignificant trend toward a reduction in ischemic stroke (0.54% vs 0.83% per year; HR 0.71; 95% CI 0.47-1.07), and a lower risk of the composite of ICH or ischemic stroke (0.95% vs 1.60% per year; HR 0.61; 95% CI 0.45-0.82).

Apixaban carried a similarly lower rate of ICH compared with warfarin (0.38% vs 0.97% per year; HR 0.38; 95% CI 0.17-0.88), with nonsignificant trends toward a higher rate of ischemic strokes (0.56% vs 0.51% per year; HR 1.13; 95% CI 0.49-2.63) and a lower rate of the composite endpoint (0.89% vs 1.44%; HR 0.63; 95% CI 0.35-1.12).

Coleman noted some caveats that need to be considered when interpreting studies based on claims data, including the little amount of information available on vital signs or laboratory results, the lag in when data are available, potential misclassification bias, and the evolution of the benefit-risk calculation over time as experience with these newer agents grows.

Nevertheless, Reynolds said, what can be gleaned from this and prior studies is that “these are good drugs, and I think that these kinds of data should reassure people that out in practice the results are as good if not better than what happened in the trials.”

Related Stories:

• Comparing the Novel Oral Anticoagulants: Effectiveness Similar, but Apixaban May Come With Less Bleeding
• Understanding Differences Among New Oral Anticoagulants Should Increase Compliance
• Meta-analysis Takes an Indirect Look at Differences Among Novel Anticoagulants