Higher risk of all-cause mortality in AF with nonrecommended dosing of NOACs
Mortality in Patients With Atrial Fibrillation Receiving Nonrecommended Doses of Direct Oral Anticoagulants
Introduction and methods
Novel oral anticoagulants (NOACs) are approved in most countries in the world, but differences exist with regard to regulatory authority-specific guidance that takes into account baseline characteristics of the patients, such as kidney function, low body weight, age, bleeding risk and drug-drug interactions. Country-specific guidelines vary for the different NOACs [1-3] and confusion about correct prescription may lead to inappropriate dosing (both underdosing and overdosing).
This study examined the patterns of NOAC prescription with regard to dosing, the impact of NOAC dosing on rate of events at 2-year follow-up and predictors of underdosing in newly diagnosed AF patients in the GARFIELD-AF registry.
The GARFIELD-AF registry enrolled patients with AF diagnosed within the previous 6 weeks, with at least 1 risk factor for stroke and no valvular disease. For this analysis patients from cohort 3 to 5 were enrolled (n=34926). Those not on any NOAC, with more than 1 NOAC type simultaneously, or no information on starting date, no CKD stage information, unavailable dose information were excluded, leaving 10426 patients. Rules of prescription for the NOACs rivaroxaban, apixaban, edoxaban and dabigatran were reviewed for definition of recommended dosing, underdosing or overdosing. 4491 (43.1%) Patients received rivaroxaban, 3290 (31.6%) apixaban, 2359 (22.6%) dabigatran and 286 (2.7%) edoxaban.
- Majority of patients received the recommended dosing (n=7603, 72.9%), 2423 (23.2%) were underdosed and 400 (3.8%) were overdosed.
- Recommended dosing was between 70.1% and 81.9% for rivaroxaban, dabigatran and apixaban, but lower (40.6%) for edoxaban. Underdosing occurred between 15.8% to 28.7% of patients taking rivaroxaban, dabigatran and apixaban, and was more common in those on edoxaban (55.9%). Overdosing rates were low, from 1.3% in those on dabigatran to 6.5% with rivaroxaban. In those with overdosing, 67.5% had moderate-to-severe CKD.
- Highly variable pattern for nonrecommended dosing was observed worldwide.
- Risk of all-cause mortality was higher in patients with nonrecommended doses (HR 1.24, 95%CI: 1.04-1.48) and in underdosed patients (HR 1.25, 95%CI: 1.04-1.50) compared to patients treated with recommended doses.
- Risk of major bleeding in underdosed patients was lower than in those with recommended dosing (HR 0.50, 95%CI:0.28-0.88).
- Increase in all-cause mortality was mainly due to higher rates of CV death (heart failure and MI) in underdosed patients compared to those with recommended doses.
- No difference was observed for stroke-related death and bleeding in those with nonrecommended dosing compared to recommended dosing.
- Most potent predictors of underdosing were female sex, non-Caucasian ethnicity, acute coronary syndrome, vascular disease, prior stroke, diabetes, concomitant AP therapy. After removal of the variables that were used to define underdosing and were strong predictors (age, weight, CKD) congestive HF, management in office or anticoagulant clinics and diastolic blood pressure emerged as predictors.
Most AF patients who were enrolled in the GARFIELD-AF registry received appropriate dosing of NOACs. Underdosing was not uncommon though, but overdosing rate was low. Nonrecommended doses and underdosage were associated with increased risk of mortality, but not with increased risk of stroke-related death. Nonrecommended dosing was not associated with bleeding compared to those who received recommended doses, but those who received underdosing had reduced bleeding compared to those with recommended doses.