The Persistence of Non-Vitamin K Antagonist Oral Anticoagulants in Korean Patients with Non- Valvular Atrial Fibrillation

Received: November 14, 2016 Revision Received: December 19, 2016 Accepted: December 23, 2016 Correspondence: Young Keun On, MD Division of Cardiology, Department of Internal Medicine, Heart Vascular Stroke Institute, Samsung Medical Center, Sungkyunkwan University School of Medicine, 81, Irwon-ro, Gangnam-gu, Seoul 06351, Republic of Korea Tel: +82-2-3410-3420 Fax: +82-2-3410-3849 E-mail: yk.on@samsung.com


Introduction
Atrial fibrillation (AF) is the most common cardiac arrhythmia in clinical practice. AF is a global health-care problem with an increasing prevalence and incidence. 1 Patients with AF are at increased risk of death, heart failure, hospitalization, and thromboembolic events. 2 AF-associated strokes are often severe, resulting in disability or death. 3 Thus, an oral anticoagulation therapy such as vitamin K-antagonist anticoagulants (VKAs) is important for the prevention of stroke in patients at risk of AF.
Recently, several non-vitamin K-antagonist oral anticoagulants (NOACs) have been introduced. Randomized clinical trials (RCTs) showed that NOACs were superior to VKAs in efficacy and safety of preventing stroke and systemic embolisms in patients with non-valvular AF (NVAF). [4][5][6] Current AF guidelines recommend risk stratification and use of NOACs. 7 NOACs provide lifestyle advantages, and efficacy and safety in stroke prevention. The use of NOACs is gradually increasing worldwide. The aim of this study was to evaluate the medication persistence of NOACs in comparison with that of warfarin.

Study population and study design
This is a retrospective study of patients with NVAF who were indicated for AF treatment. We included 1,527 patients with NVAF who had newly started NOAC therapy between January Clinical data were collected from the patients' medical records to identify patients to whom NOACs or warfarin was prescribed, and those diagnosed with AF. We excluded patients with valvular heart disease, such as patients with moderate or severe mitral stenosis, mild rheumatic mitral stenosis, or a history of valvular surgery.

Clinical characteristics and definition of variables
Information on baseline characteristics such as patient age, sex, risk factor, and underlying disease were collected from medical records and laboratory data. Hypertension was defined as a systolic blood pressure of ≥140 mmHg, a diastolic blood pressure of ≥90 mmHg, or a history of treatment for hypertension.
Diabetes mellitus (DM) previously diagnosed by a physician was defined as treatment with hypoglycemic agents or indicated by poor glycemic control (defined as a glycohemoglobin A1c level of ≥6.5%).
Heart failure was defined in accordance with the American College of Cardiology/American Heart Association criteria.

Follow-up
Information on drug cessation, major cardiovascular events, and minor adverse events were obtained at routine or additional visits at our hospital. The patients were followed up until the end of the follow-up period or until discontinuation of the drug, loss to follow-up, or death. The follow-up period was until September 2016. Information concerning deceased patients was obtained from medical records, family members, the patients' general practitioners, and the hospitals to which they had been admitted.
In the patients who started warfarin, we also collected prothrombin time (PT)-INR data.

Outcomes
Drug cessation was defined as discontinuation of the prescribed drug and physician's mention of drug cessation in the medical record. All cases of drug cessation were reviewed from the medical records, and the reasons for cessation were also obtained.
Temporary discontinuation for specific reasons such as surgery was not considered as drug cessation. Major cardiovascular events are defined as thromboembolism and major bleeding. The occurrence of thromboembolic and bleeding events was evaluated through the medical records reviewed by the 2 investigators.

Statistical analysis
Summary data are presented either as mean and standard deviation (SD) or as number of patients. Differences in baseline characteristics between NOACs and warfarin were tested with chi-square analysis for categorical variables and analysis of variance for continuous variables. The incidence rates of events were compared between the individual groups by using the chi-square test. The Kaplan-Meier curves for the outcomes of major

Patient characteristics
The baseline characteristics of the patients in the 3 NOAC groups and the warfarin group are shown in Table 1

Discontinuation and persistence of NOACs and warfarin
The mean follow-up duration was 532 days (median, 404 days).
The discontinuation rates and cause of cessation in the 3 NOAC groups and warfarin groups are shown in Table 2 Figure 2. The discontinuation rate in the warfarin group was lower than that in the NOAC groups.
We excluded maintenance of sinus rhythm to calculate the persistence rate because many patients discontinued anticoagulation because of maintenance of sinus rhythm by operation or procedure such as radiofrequency ablation (RFCA) or total thoracoscopic ablation (TTA; Table 1). The Kaplan-Meier curves for persistence of NOACs and warfarin without maintenance of sinus rhythm are shown in Figure 3. The discontinuation rate was higher in all the NOAC groups than in the warfarin group. The incidence rate of adverse events in the warfarin group was lower than those in the 3 NOAC groups within the first year of treatment.

Major cardiovascular events
The warfarin group had lower mean CHA2DS2-VASc and HAS-BLED scores than the 3 NOAC groups. No significant differences in the incidence of thromboembolic and major bleeding events were found between the NOAC groups and the warfarin groups (Table 3). The warfarin group tended to have a higher incidence of major bleeding than the 3 NOAC groups, but the difference was not statistically significant.

Antithrombotic treatment after discontinuation
Antithrombotic treatments after discontinuation are shown in Table 4. The patients who discontinued NOACs mostly received another NOAC or an antiplatelet drug except for the apixaban group. Apixaban group tended not to receive any antithrombotic treatment. The patients who discontinued warfarin also received a NOAC or an antiplatelet drug.

The persistence of NOAC and warfarin
This study showed the persistence and adverse events related to    The values are presented as mean±SD or number of patients (%). NOACs, non-vitamin K antagonist oral anticoagulants *Antiplatelet agent = aspirin or P 2 Y 12 receptor antagonist International Journal of Arrhythmia 2016;17 (4):190-199 years. 4 The clinical trial of apixaban in comparison with warfarin showed that fewer patients (25.3%) in the apixaban group than in the warfarin group (27.5%) discontinued a study drug before the end of the study. 5 The clinical trial of rivaroxaban in comparison with warfarin showed that the proportion of patients who discontinued their assigned drug was 23.7% in the rivaroxaban group and 22.2% in the warfarin group. 6 The discontinuation rates for the NOACs in the major clinical trials ranged from 21-25%. The major clinical trials [4][5][6] showed that the persistence rates in the dabigatran and rivaroxaban groups were lower than that in the warfarin group, and that the persistence rate in the apixaban group was similar to that of warfarin group. These findings are similar to those of our study.

Reasons of cessation and Major complication
In this study, the major causes of discontinuation were the adverse events or patient's request. Patient's request in major clinical trials accounted for 8-10% of the total treated population.
This finding was similar in our study. The adverse events were not severe, and the minor bleeding or gastrointestinal symptoms were the major reasons. Minor bleeding does not lead to severe major bleeding. Compared with the other NOAC groups and the warfarin group, the dabigatran group showed a higher proportion of patients with gastrointestinal symptoms than those with bleeding due to adverse events. The incidence rates of thromboembolic and major bleeding events were not significantly different between the NOAC groups and the warfarin group in clinical practice. These results were lower than those of the major clinical trials. [4][5][6] However, these findings should be explained cautiously because this study was retrospective.

Difference of the drug approval time
The drug approval times of NOACs differ in South Korea.
Dabigatran was first approved, which was followed by the approval of rivaroxaban and then apixaban. At the time NOACs were approved, its safety and effectiveness in the clinical real world was not clarified. Concerns have been raised regarding the potential risk of serious adverse events induced by NOACs in comparison with warfarin. Therefore, the discontinuation rates of dabigatran might be higher than that of rivaroxaban or apixaban.
As apixaban is the most recently approved drug, the number of patients that can be evaluated is smaller than that of other NOACs.

Persistence to NOAC treatment
The effectiveness and safety of NOACs were identified continuously by recent studies. The clinical outcomes of AF, such as mortality, stroke, and cardiovascular events, are strongly dependent on the quality of anticoagulation. Although NOACs were more convenient than warfarin as it does not require regular laboratory examination such as PT-INR and lifestyle advantage, it could decrease patient's attention to disease. The patient-physician relationship is important for better adherence. Physicians should provide sufficient information and explain the risk of discontinuation, and the necessity of consultation before drug discontinuation.

Limitations
This study has several limitations. This study was a retrospective observational study in a single center. We could not detect all minor events due to dependence on medical records. The results of this study might not reflect the situation in the average patient population in South Korea. The follow-up duration was relatively short. The starting points and follow-up period differed between the NOAC and warfarin groups because of the differences in drug-approved times.

Conclusion
In this study of AF patients, NOACs showed lower persistence rate than warfarin.

Conflict of interest
The authors report no relationships that could be construed as a conflict of interest.