NOACs Perform Worse Than Warfarin in Real-World Study, Especially When Adherence Is Low

BOSTON, MA—In pivotal randomized trials and most observational data sets, the non-vitamin K antagonist oral anticoagulants (NOACs) have fared well against warfarin, providing stroke protection at least as good if not superior to that of the older drug in patients with nonvalvular A-fib. Now, a new, real-world study based on administrative data has shown just the opposite.

Overall, patients taking one of the NOACs had significantly higher rates of thromboembolic events and any stroke compared with those taking warfarin, Dhanunjaya Lakkireddy, MBBS (Kansas City Heart Rhythm Institute, Overland Park, KS), reported at the Heart Rhythm Society (HRS) 2018 Scientific Sessions here. Hemorrhagic stroke and bleeding complications, however, were lower with the NOACs.

Of note, whether patients were adherent to their treatment regimens seemed to influence the findings. When adherence was high, the differences between the NOACs and warfarin were relatively small. On the other hand, when adherence was low, the gaps in thromboembolic events and strokes favoring warfarin widened and the differences in hemorrhagic stroke and bleeding grew smaller.

“This really makes an important point that when people don’t take their medications, are these benefits really translated in the same way?” Lakkireddy said during his presentation, alluding to the benefits of NOACs seen in clinical trials.

Acknowledging the limitations, he also said the study “really exposes some inherent challenges that we have in the processes related to OAC [oral anticoagulant] use, and it highlights the importance of improving adherence to oral anticoagulation.” He added, too, that alternative therapies like left atrial appendage occlusion devices should be studied.

Explaining the Findings

Experts looking at the results tried to make sense of the findings, which contradict most prior research comparing the efficacy and safety of NOACs and warfarin.

Serving as a discussant after Lakkireddy’s presentation, Hein Heidbüchel, MD, PhD (University of Antwerp, Belgium), raised several methodological concerns with the study, including the lack of propensity matching; the lack of more detailed clinical information on the patients with low adherence, including whether they had prior bleeding or were also taking antiplatelets; the exclusion of patients with temporary A-fib or temporary anticoagulation; the lack of adjudication of outcomes; and the lack of information on potential differences between the various NOACs.

He said that many questions remain, making it hard to draw firm conclusions, and that further analysis is required to explain the findings.

Serving as a moderator at a press conference, Andrew Krahn, MD (University of British Columbia, Vancouver, Canada), also highlighted some of the limitations of these types of analyses and said that’s why “we don’t try to make a lot of therapeutic recommendations from them.”

But what is clear, according to Heidbüchel, is that “the adherence issue is something we have to think of and we have to address. And that starts by knowing why patients are nonadherent. There may be multiple reasons, and we can only correct it when we know why, because we have to change the behavior of the patient and that [involves] a multipronged approach.”

Adherence Key to Effectiveness

Lakkireddy noted that adherence is critical when it comes to the effectiveness of chronic medical therapy. This factor is an issue both for warfarin, with reported time in therapeutic range of 40% to 60% and a discontinuation rate of 40%, and the NOACs, with reported discontinuation rates of 25% to 50%.

Heidbüchel pointed out that adherence is related to efficacy, with prior studies showing that nonadherence is associated with more thromboembolic events and mortality.

In the current study, Lakkireddy and his colleagues aimed to evaluate the impact of adherence on the relative safety and efficacy of the NOACs versus warfarin. They examined administrative data regarding inpatient, outpatient, and pharmaceutical claims from 2012 to 2016 using the IBM Watson Health MarketScan database. The analysis included 83,168 patients with A-fib and a CHA2DS2-VASc score of at least 2 who had a prescription for oral anticoagulation in 2014 or 2015; around half (51.3%) were taking warfarin and the rest a NOAC.

The researchers further divided patients into groups by adherence. Lower adherence, defined as having prescriptions covering 40% to 80% of days, was seen in 36.9% of warfarin-treated patients and 26.5% of NOAC-treated patients.

Kaplan-Meier survival curves in the overall cohort and in the high- and low-adherence groups showed that risks of thromboembolic events and any stroke were higher in patients taking NOACs versus warfarin, although hemorrhagic stroke and bleeding favored the NOACs.

What Does It Mean?

When asked about the potential clinical implications of the findings, Andrea Russo, MD (Cooper University Health Care, Camden, NJ), first vice president of the HRS, said, “I think we need to have more investigations that are prospective to examine what’s happening out in real-life practice.”

She said administrative data sets often lack important clinical information that could be used to provide more clarity to analyses like this.

“This is just a quick overview,” Russo told TCTMD. “It’s not really giving us the granularity in terms of INR management, what are [patients’] INRs doing, the time in therapeutic range, we don’t have that. This is not that kind of study.”

During a discussion at the press conference, Russo also raised the possibility that there could be selection bias that explains the findings. In practice, if patients have been well controlled on warfarin for several years, they will likely continue taking that drug instead of switching to a NOAC, she indicated. Thus, real-world analyses may be looking at warfarin-treated cohorts that are inherently compliant.

Addressing some of the criticisms, Lakkireddy said that his team is going to perform a propensity-matched analysis and will take a deeper dive into potential differences between the NOACs.