Off-label DOAC Dosing Tied to Worse Adherence, More Discontinuation

Patients with atrial fibrillation (AF) who use a direct oral anticoagulant (DOAC) dose that’s inconsistent with US Food and Drug Administration labeling—a practice believed, in many cases, to be related to fears of bleeding—are less likely to stick with therapy, a retrospective study shows.

Compared with an appropriate dose, use of inappropriate lower or higher doses was associated with a reduced likelihood of adherence and a greater likelihood of discontinuation through the first year of treatment, according to researchers led by Jennifer Rymer, MD, MBA (Duke Clinical Research Institute, Durham, NC).

Use of nonrecommended doses was particularly prevalent among patients with renal dysfunction, and for every 10-unit decrease in creatinine clearance, there was a 21% drop in the likelihood of getting an appropriate DOAC dose.

Patients with renal disease are at especially high risks of both bleeding and ischemic outcomes, Rymer noted. “So it’s really important to try to get adequate dosing in these patients.”

Some prior studies have found off-label DOAC dosing to be associated with worse clinical outcomes, including one showing that use of inappropriate higher or lower doses was associated with greater risks of mortality and CV hospitalization, respectively. Nonrecommended dosing also was tied to a higher risk of 2-year mortality in the GARFIELD-AF registry.

In the current study, however, off-label dosing was not associated with either stroke or hospitalization involving bleeding.

Nevertheless, “I would still say . . . based on all the other data that we have that when we know what the dose should be according to FDA labeling, when at all possible, the patient should be on that dose to . . . decrease their risk of stroke,” Rymer said.

For the study, published online recently in JAMA Network Open, Rymer et al examined medical and prescription claims from the Symphony Health data set, which contains information on about 280 million patients and 1.8 million prescribers in the United States.

The analysis included 86,919 patients (median age 74 years; 50.3% men) who had at least two claims for AF between 2015 and 2017, had a CHA₂DS₂-VASc score of 2 or higher, and received a prescription for a DOAC. Overall, 8.4% of patients received a reduced DOAC dose after meeting label-specified criteria and another 12.6% were underdosed inappropriately. Another 5.0% of patients received an off-label excess dose.

Compared with patients who were appropriately treated, those who received a nonrecommended DOAC dose tended to be older (median age 79 vs 73) and to have a higher CHA₂DS₂-VASc score (median 5 vs 4).

Receipt of an off-label lower dose was associated with a lower likelihood of adherence (adjusted OR 0.88; 95% CI 0.83-0.94) and higher odds of anticoagulation discontinuation (adjusted OR 1.20; 95% CI 1.13-1.28) through 1 year of follow-up compared with appropriate dosing, with similar findings observed for excess dosing.

Adherence and persistence were not adversely affected among patients who met labeled criteria for dose reduction, however.

As for why off-label dosing would be associated with worse adherence and persistence, Rymer said that it could be a marker of other factors. For instance, patients with multiple comorbidities, who may be the types of patients physicians worry about when it comes to dosing DOACs, may be taking a lot of other medications, which in turn makes it difficult to stick with various drug regimens.

In terms of clinical outcomes, off-label underdosing was not associated with differences in risks of hospitalization involving bleeding (adjusted HR 0.95; 95% CI 0.84-1.08) or stroke (adjusted HR 0.99; 95% CI 0.90-1.09) compared with appropriate dosing.

Of note, however, those with an appropriate reduction in dose carried a greater risk of hospitalization involving bleeding compared with those who received an appropriate full dose (adjusted HR 1.21; 95% CI 1.01-1.44), “suggesting that the developed dose criteria are appropriately discriminating patients who are at higher risk of bleeding,” according to the authors.

The study, they say, “suggests that opportunities remain to improve DOAC dosing through quality improvement initiatives and potentially through best practice alerts in the electronic health record that can detect off-label DOAC dosing.”

Using a more algorithmic approach will be especially important in patients with renal dysfunction, Rymer said, “because these patients are often the ones that I think we worry quite a bit about their bleeding but we don’t worry enough about their ischemic risk.”