The nationwide ANATOLIA-AF study provides important real-world data on the prevalence and associated factors of inappropriate direct oral anticoagulant dosing in patients with atrial fibrillation. The principal findings of this study are as follows: 1) more than one-fourth of the patients with atrial fibrillation receiving direct oral anticoagulant therapy received a reduced dose of direct oral anticoagulants; 2) according to the European Heart Rhythm Association recommendations, nearly one-fourth of the patients with atrial fibrillation received inappropriate doses of direct oral anticoagulant in real-life settings; 3) in contrast to previously published registries and studies, the rate of inappropriate high-dose prescription of direct oral anticoagulants was similar to the rate of inappropriate low-dose prescription; and 4) advanced age, chronic kidney disease, presence of permanent atrial fibrillation, prescription of edoxaban, reduced dose of direct oral anticoagulants, concomitant usage of amiodarone, and non-use of statin treatment were independently associated with inappropriate direct oral anticoagulant dosing.
The efficacy and safety of reduced doses of direct oral anticoagulants, especially rivaroxaban and apixaban, have been tested in a small proportion of patients in pivotal randomized controlled trials [3–6]. In the ROCKET-AF trial [4], 21% of patients received a reduced dose of rivaroxaban, and in the ARISTOTLE trial [5], only 4.7% of study participants received a reduced dose of apixaban. However, national registries and observational studies have demonstrated that the frequency of reduced doses of direct oral anticoagulant prescription varies between 29–56% [10, 19, 20]. In the ANATOLIA-AF study, the percentage of individuals that were prescribed reduced doses of direct oral anticoagulants was 27.8%, which is higher than that in pivotal trials and similar to that in national registries and observational studies [4, 5, 10, 19, 20]. Compared with the patients enrolled in phase 3 randomized controlled trials, patients in routine clinical practice are older, more frail, more likely to have chronic kidney disease, have a history of major and/or clinically relevant non-major bleeding and/or gastrointestinal bleeding, and more likely to receive concomitant antiplatelet therapy and/or interacting drugs [8]. As a result, the prescription of low doses of direct oral anticoagulants (appropriate or inappropriate) is more common in real-life settings [9, 10, 12, 14–16]. This considerable difference in the prescription of reduced doses of direct oral anticoagulants between randomized controlled trials and real-world studies may be due to numerous reasons, including patient-related factors such as advanced age, frailty, and the presence of multiple comorbidities, or physician-related factors such as lack of awareness about the recommendations of the guidelines and/or fear of bleeding and other adverse events.
Inappropriate dosing of direct oral anticoagulants is associated with major adverse cardiovascular events [8, 12, 13, 15]. The prescription of inappropriately low doses of direct oral anticoagulants may increase the risk of stroke and/or systemic embolism and cardiovascular hospitalization [8, 13, 21], and the prescription of inappropriately high doses of direct oral anticoagulants may increase the risk of major bleeding and all-cause mortality [8, 21]. Previously published national registries and studies have reported a wide range of inappropriate direct oral anticoagulant dosing in real-life settings [9, 10, 12–16]. The FANTASIIA Registry from Spain revealed that inappropriate doses of direct oral anticoagulants were prescribed to 32% of study participants, and the SAKURA AF Registry from Japan reported that 26.2% of patients with atrial fibrillation received inappropriate doses of direct oral anticoagulants [10, 12]. The prevalence of inappropriate direct oral anticoagulant dosing reaches 39% in the Middle East region [15]. In contrast, the ORBIT-AF II Registry, showed that an inappropriate dose of a direct oral anticoagulant was prescribed to only 12.8% of the patients [13]. A recently published epidemiological meta-analysis comprising 23 real-world studies and 162,474 patients with atrial fibrillation reported that the overall prevalence of inappropriate direct oral anticoagulant dosing was 24% [11]. In our study, we found that the prevalence of inappropriate direct oral anticoagulant dosing was 24.9%. The wide range in the prevalence of inappropriate dosing of direct oral anticoagulants between real-world observational studies might be related to the differences in the criteria for establishing appropriate doses, geographic and clinical variations, patient and/or center selection, and physicians’ knowledge.
Data from national registries and observational studies suggest that while the proportion of patients receiving inappropriately low doses of direct oral anticoagulants vary between 19.3% and 39% [12, 19, 15, 22, 23], the prevalence of inappropriate high-dose direct oral anticoagulant prescription varies between 1.3% and 4% [12, 13, 19, 22]. A recently published epidemiological meta-analysis reported that the estimated global prevalence of inappropriate low-dose direct oral anticoagulant prescriptions is 20%, and the estimated global prevalence of inappropriate high-dose direct oral anticoagulant prescriptions is relatively low (4–6%) [11]. In contrast, the prevalence of inappropriate high-dose prescriptions of direct oral anticoagulants was similar to the prevalence of inappropriate low-dose prescriptions in our study (12.8% versus 12%, respectively). These inconsistent results might be related to differences in the criteria for establishing an appropriate dose of direct oral anticoagulants. Most observational studies have assessed the appropriateness of direct oral anticoagulant dosing according to the summary of product characteristics written by pharmaceutical companies. In contrast, we used the criteria proposed by the European Heart Rhythm Association for the use of direct oral anticoagulants in patients with atrial fibrillation in the present study. The European Heart Rhythm Association Practical Guide recommends a systematic assessment algorithm for dose reduction for each direct oral anticoagulant. These recommendations comprise several patient-related risk factors, such as older age, impaired kidney function, lower body weight, high bleeding risk, and concomitant use of antiplatelet therapy, verapamil, diltiazem, amiodarone, or strong P-glycoprotein inhibitors. Some of these criteria are called yellow dose reduction criteria in the European Heart Rhythm Association Practical Guide. This document recommends dose reduction if two or more yellow criteria are present, which are not completely present in the summary of product characteristics [18]. For example, the summary of product characteristics of rivaroxaban reported only one dose reduction criterion, which recommends a dose reduction if the patient has a creatinine clearance below 50 mL/min. However, the European Heart Rhythm Association Practical Guide took into account the patients’ age (≥ 75 years), bleeding risk (HAS-BLED score ≥ 3), low body weight (≤ 60 kg), and concomitant usage of antiplatelet therapy or amiodarone as yellow criteria for rivaroxaban dose reduction. The FANTASIIA Registry, which determined the appropriateness of direct oral anticoagulant dosing according to the European Heart Rhythm Association Practical Guide recommendations, reported results similar to those of the present study. In this registry, the prevalence of inappropriately high-dose direct oral anticoagulant prescriptions was 15%, which is higher than previously published registries, and was approximately similar to the prevalence of inappropriately low-dose direct oral anticoagulant prescriptions (17%) [10]. Another important real-life study that analyzed the appropriateness of direct oral anticoagulant dosing according to both the summary of product characteristics and European Heart Rhythm Association Practical Guide revealed that the rates of inappropriate direct oral anticoagulant dosing were 18.3% and 23.4%, respectively. According to the summary of product characteristics, the rate of inappropriately low-dose direct oral anticoagulant prescriptions (9.8%) was higher than the rate of inappropriate high-dose direct oral anticoagulant prescriptions (7.8%). In contrast to the summary of product characteristics criteria, the inappropriately high-dose direct oral anticoagulant prescription was more frequent than the inappropriately low-dose direct oral anticoagulant prescription according to the European Heart Rhythm Association Practical Guide recommendations in the same study population (15.0% versus 7.6%, respectively). The authors concluded that ‘nearly 10% of dosing recommendations by the summary of product characteristics and the European Heart Rhythm Association Practical Guide were inconsistent, with the summary of product characteristics recommending the use of the standard direct oral anticoagulant dose while the European Heart Rhythm Association Practical Guide recommending dose reduction’ [9]. It is important to highlight that several risk factors were not considered by the summary of product characteristics of direct oral anticoagulants. Although advanced age is a well-known risk factor for both stroke and bleeding [9, 24, 25], neither the summary of product characteristics of rivaroxaban nor that of edoxaban consider age as a risk factor for stroke or bleeding. According to the European Heart Rhythm Association Practical Guide, age ≥ 75 years is a yellow criterion for dose reduction for all direct oral anticoagulants [18]. Low body weight is another important risk factor for adverse events, especially major and/or clinically relevant non-major bleeding, in patients with atrial fibrillation who are receiving direct oral anticoagulants [1]. However, only the summary of product characteristics of apixaban and edoxaban included body weight as a dose reduction criterion. The European Heart Rhythm Association Practical Guide includes body weight ≤ 60 kg as one of the yellow criteria for dose reduction of dabigatran and rivaroxaban [18]. Concomitant use of antiplatelet drugs, non-steroidal anti-inflammatory drugs, or amiodarone increases bleeding risk and/or affects drug plasma levels in patients receiving direct oral anticoagulant treatment [1, 7, 8, 11]. Although the summary of product characteristics of each direct oral anticoagulant does not have a specific recommendation, the European Heart Rhythm Association Practical Guide recommends direct oral anticoagulant dose reduction in those patients [18]. In the context of inconsistent direct oral anticoagulant dosing recommendations by the summary of product characteristics and the European Heart Rhythm Association Practical Guide, prescribing the appropriate direct oral anticoagulant dose can be challenging in clinical practice. Thus, there is a need for more research on the clinical importance of risk factors causing inconsistencies in dosing recommendations between the European Heart Rhythm Association Practical Guide and the summary of product characteristics of direct oral anticoagulants.
Previously published registries and meta-analyses have reported an association between older age and inappropriate direct oral anticoagulant dosing [8, 11, 16, 26, 27]. Similarly, we observed a strong association between age and inappropriate direct oral anticoagulant dosing. In our study, patients who received inappropriate low or high doses of direct oral anticoagulants were significantly older than those who received an appropriate dose of direct oral anticoagulants. Moreover, we found that older age was an independent predictor of inappropriate direct oral anticoagulant dosing according to the logistic regression analysis. Future directions and efforts targeting gaps in the elderly patient population may lead to improved adherence to the recommendations of the European Heart Rhythm Association Practical Guide.
Chronic kidney disease has been associated with inappropriate direct oral anticoagulant dosing in previous studies [10, 13, 26, 28]. In the ORBIT-AF II Registry, creatinine clearance was significantly lower in patients with inappropriate direct oral anticoagulant dosing, and mild-to-moderate renal impairment had the highest rates of inappropriate dosing [13]. An analysis of a large administrative database including 14,865 patients with atrial fibrillation showed that 43% of the patients with a renal indication for dose reduction received inappropriate high-dose direct oral anticoagulants, which was associated with a higher risk of major bleeding [29]. In our study, the estimated glomerular filtration rate, which was calculated using the Cockcroft-Gault equation, was significantly lower in patients with inappropriate direct oral anticoagulant dosing. According to the logistic regression analysis, mild-to-moderate renal impairment (estimated glomerular filtration rate 30–50 mL/min/1.73 m2) is associated with prescription of inappropriate direct oral anticoagulant dosing. This result suggests that many clinicians may not adjust the direct oral anticoagulant dose according to the renal function and estimated glomerular filtration rate. There may be several reasons for this, including a lack of awareness about dose reduction criteria, physicians’ knowledge, and the use of different calculations to estimate renal function (e.g., Modification of Diet in Renal Disease formula). Physician awareness of renal function in appropriate direct oral anticoagulant dosing should be increased, and physicians should be encouraged to use the Cockcroft-Gault equation, a method used in pivotal trials of direct oral anticoagulants to calculate estimated glomerular filtration rate in patients with atrial fibrillation and direct oral anticoagulant treatment.
The prescription of a reduced dose of direct oral anticoagulants without dose reduction criteria is associated with an increased risk of stroke, systemic embolism, cardiovascular hospitalization, and death [8, 12, 13, 15, 16]. Jacobs et al. [30] reported that receiving a reduced dose of direct oral anticoagulants was associated with a 2.7-fold increased risk of inappropriate dosing. Similarly, we found that prescription of a reduced dose of direct oral anticoagulants was independently associated with inappropriate direct oral anticoagulant dosing. According to age-adjusted logistic regression analysis, receiving a reduced dose of direct oral anticoagulants was associated with a 2.5-fold increased risk of inappropriate direct oral anticoagulant dosing in our study population. Thus, to ensure the safety and efficacy of direct oral anticoagulants, a reduced dose of direct oral anticoagulants should be prescribed in accordance with the recommendations of the guidelines in real-life clinical practice [11].
A limited number of real-life observational studies are available on inappropriate dosing of edoxaban [8]. An epidemiological meta-analysis comprising of 1213 edoxaban-treated patients reported that the highest prevalence of inappropriate direct oral anticoagulant dosing was found in rivaroxaban and edoxaban [11]. The prevalence of inappropriate high-dose prescriptions of edoxaban was 9%, while the pooled prevalence of inappropriate high-dose prescriptions for all direct oral anticoagulants was 5% [11]. Similar to this meta-analysis, the highest prevalence of inappropriate high-dose prescriptions of direct oral anticoagulant was found in edoxaban (20.5%) in the present study. Moreover, the age-adjusted logistic regression model demonstrated that edoxaban treatment was independently associated with inappropriate direct oral anticoagulant dosing. Although both regimens of edoxaban were non-inferior to warfarin with respect to the prevention of stroke or systemic embolism in the ENGAGE AF-TIMI 48 trial, the rate of ischemic stroke was higher in the low-dose edoxaban group than in the active comparator group [6]. Thus, physicians may have concerns about the efficacy of low-dose edoxaban in the prevention of ischemic stroke. These concerns may lead to the prescription of high-dose edoxaban despite the presence of impaired renal function, low body weight, or concomitant use of strong P-gp inhibitors.
Concomitant use of amiodarone is another important risk factor affecting direct oral anticoagulant plasma levels and/or bleeding risk. Amiodarone is the most frequently used pharmacokinetic interacting drug in direct oral anticoagulant-treated patients [9, 29]. Therefore, the European Heart Rhythm Association Practical Guide recommends dose reduction in patients receiving direct oral anticoagulant and amiodarone treatment in the presence of at least one other yellow criterion, while the summary of product characteristics does not. The FANTASIIA Registry reported that the prescription of concomitant amiodarone therapy was significantly higher in patients receiving inappropriately high doses of direct oral anticoagulants [10]. Similarly, we found that the prescription of amiodarone was significantly higher in patients who received an inappropriately low or high dose of direct oral anticoagulants compared to patients receiving an appropriate dose of direct oral anticoagulants. More importantly, age-adjusted logistic regression analysis demonstrated that the concomitant use of amiodarone was significantly associated with potentially inappropriate dosing of direct oral anticoagulants. These results suggest that physicians should consider pharmacokinetic drug-drug interactions, especially between amiodarone and direct oral anticoagulants, to determine the appropriate dose of direct oral anticoagulants in real-life settings.
The present study has some limitations. First, the most important limitation of the ANATOLIA-AF study was its observational design, which may have led to patient evaluation and/or patient selection bias. Second, the study population was enrolled from cardiology outpatient clinics, and this population did not include those presenting at family medicine and internal medicine outpatient clinics. Finally, the present study data were based on the documentation of demographics, medical history, and treatments during the first outpatient clinic visit; follow-up data were not obtained. Thus, we were unable to report the stroke, systemic embolism, bleeding, and mortality rates in the study population. We will assess the association between inappropriate direct oral anticoagulant dosing and major adverse cardiac events when our follow-up data become available. Due to these limitations, the results of this study should be interpreted carefully.