Bleeding Associated With Antiarrhythmic Drugs in Patients With Atrial Fibrillation Using Direct Oral Anticoagulants: A Nationwide Population Cohort Study

Abstract

Background

This study investigated drug–drug interactions in patients with atrial fibrillation taking both a direct oral anticoagulant (DOAC) and an antiarrhythmic drug.

Methods and Results

Using data from the National Health Insurance database (2012–2018), we identified 78 805 patients with atrial fibrillation on DOACs, with 24 142 taking amiodarone, 8631 taking propafenone, 2784 taking dronedarone, 297 taking flecainide, 177 taking sotalol, and 42 772 on DOACs alone. Patients with bradycardia, heart block, heart failure, mitral stenosis, prosthetic valves, or incomplete data were excluded. Propensity score matching compared those taking both DOACs and antiarrhythmic drugs with those on DOACs alone. There was an increased risk of major bleeding in patients concomitantly taking DOACs with amiodarone when compared with matched patients taking DOACs alone (hazard ratio [HR],1.13 [95% CI, 1.04–1.23]; P=0.0044), particularly in patients taking dabigatran (HR, 1.19 [95% CI, 1.03–1.38]; P=0.0175). No significant difference in bleeding risk was found for propafenone, dronedarone, flecainide, or sotalol. The small sample sizes in the flecainide and sotalol groups limit interpretation. Notably, intracranial bleeding risk was higher in patients on DOACs and amiodarone, regardless of age. Additionally, patients <80 years old taking dabigatran with amiodarone or propafenone had a higher risk of gastrointestinal bleeding.

Conclusions

Concomitant use of DOACs with amiodarone, but not dronedarone or propafenone, increases the risk of major bleeding, particularly intracranial bleeding. This study provides new evidence to guide clinicians to tailor concomitant anticoagulation and antiarrhythmic therapy for patients with atrial fibrillation.

Nonstandard Abbreviations and Acronyms

AAD

antiarrhythmic drug

DDI

drug–drug interaction

DOAC

direct oral anticoagulant

NHI

National Health Insurance

PSM

propensity score matching

Clinical Perspective.

What Is New?

• The retrospective database analysis showed that in patients with atrial fibrillation, the concomitant use of direct oral anticoagulants (DOACs) with amiodarone, but not propafenone or dronedarone, was associated with an increased risk of major bleeding when compared with the use of DOACs alone.

• The concomitant use of some DOACs with amiodarone was associated with an increased risk of intracranial bleeding when compared with DOACs alone in patients above or below 80 years old.

• The DOAC doses were similar between patients concomitantly taking DOACs with antiarrhythmic drugs and patients taking DOACs alone in a real‐world setting, regardless of whether there are known drug–drug interactions between the DOAC and antiarrhythmic drug.

What Are the Clinical Implications?

• The choice and dose of a DOAC and antiarrhythmic drug, particularly amiodarone, need to be carefully considered for patients with atrial fibrillation who require the use of DOACs and who have a high bleeding risk.

Atrial fibrillation (AF) is the most common sustained arrhythmia and increases the risk of embolic stroke. ¹ , ² , ³ , ⁴ The Atrial Fibrillation Better Care pathway, known as the ABC pathway, entails the integrated care of patients with AF, with A referring to anticoagulation/avoid stroke, B referring to better symptom control, and C referring to cardiovascular and comorbidity optimization. The component of A requires the use of anticoagulants such as direct oral anticoagulants (DOACs) and vitamin K antagonists to decrease the risk of thromboembolism. The component of B suggests the use of rate control drugs to reduce heart rate (rate control) and antiarrhythmic drugs (AADs) to restore and maintain sinus rhythm (rhythm control). The rhythm control strategy may also involve catheter ablation and cardioversion. The role of rhythm control for newly diagnosed patients has become increasingly important in the past few years due to the Early Treatment of Atrial Fibrillation for Stroke Prevention Trial–Atrial Fibrillation Network (EAST‐AFNET) 4 trial. ⁵ As such, the 8th Atrial Fibrillation Network/European Heart Rhythm Association consensus document published in 2022 has suggested to change the B component to better rhythm management. ⁶ The key priority when choosing an AAD for long‐term rhythm control is to minimize proarrhythmia risk and organ toxicity, and each of the AADs has specific risks. Flecainide and propafenone may promote 1:1 atrioventricular conduction and increase ventricular rate. Amiodarone is the most effective AAD but can adversely affect multiple organ systems. Dronedarone is contraindicated in patients with recently decompensated heart failure or permanent AF. Nonetheless, it has the most solid safety data ⁷ and may thus be a preferable first choice according to the 2020 European Society of Cardiology guidelines. ¹

Male patients with a CHA₂DS₂‐VASc score ≥2 and female patients with a score ≥3 are recommended to take oral anticoagulant therapy to reduce the risk of thromboembolic stroke. ¹ The 4 DOACs, dabigatran, rivaroxaban, apixaban, and edoxaban, are increasingly preferred over warfarin to prevent stroke and systemic embolism in patients with nonvalvular AF. ⁸ All of the DOACs were at least as effective as warfarin in preventing stroke or systemic embolism and were associated with lower rates of hemorrhagic stroke or intracranial bleeding. ⁹ , ¹⁰ , ¹¹ , ¹² Furthermore, the use of a DOAC does not require routine international normalized ratio monitoring or adherence to specific diets. Nonetheless, bleeding is still a concern for patients with AF using DOACs, particularly when used concomitantly with an AAD. Given the increasing role of rhythm control, the effect of drug–drug interactions (DDIs) between DOACs and AADs on bleeding events, has attracted more attention than before. The 2021 European Heart Rhythm Association Practical Guide has laid out the recommendations for the concomitant use of DOACs and AADs, but these recommendations were supported by limited studies. ¹³ , ¹⁴ , ¹⁵ , ¹⁶ Therefore, in this study, we aimed to investigate the effect of AADs on bleeding events in patients with AF taking DOACs.

METHODS

Data Source

The data that support the findings of this study are available from the corresponding author upon reasonable request. The data of this study were obtained from the universal health insurance claims database provided by the National Health Insurance (NHI) Administration and managed by the Health and Welfare Data Science Center, Ministry of Health and Welfare of Taiwan. Taiwan's NHI program started in 1995 and provides >99.5% coverage for the 23 million residents. The NHI Research Database is a claims‐based administration data set, which provides inpatient, outpatient, and emergency department services, diagnoses, prescriptions, examinations, operations, and expenditures. The following measures were implemented to protect the privacy of patients. The identifier numbers are encrypted in the NHI Research Database's claims database of the Health and Welfare Data Science Center. Analyses were conducted on‐site, and only the result reports were allowed to be taken out. Due to these stringent privacy protection measures, informed consent was waived for this study. The institutional review board of Chang Gung Memorial Hospital Linkou Branch approved this study (institutional review board number 202000065B0C501).

Study Patients and Outcomes

We retrieved data on patients with a discharge diagnosis of AF undergoing anticoagulation therapy by searching claims of medical records between January 1, 2012 and December 31, 2018 in the database. We excluded patients with bradycardia, heart block, history of admission for heart failure, mitral stenosis, prosthetic heart valve, incomplete demographic data, index date after September 30, 2018 (follow‐up <3 months), and those who used warfarin after DOACs (Figure 1).

Figure 1. Study design and flowchart of patient enrollment.

*For each antiarrhythmic drug user, we performed propensity score matching to find 1 comparable nonuser. AF indicates atrial fibrillation; and DOAC, direct oral anticoagulant.

The primary outcome was time to first occurrence of major bleeding, which was defined as life‐threatening bleeding, vital organ hemorrhage, or blood transfusion >2 U upon admission, or emergency department visit, with a principal or secondary diagnosis of major bleeding. Major bleeding included both intracranial bleeding and gastrointestinal bleeding. The outcome was defined using International Classification of Diseases, Ninth Revision and Tenth Revision, Clinical Modification (ICD‐9‐CM and ICD‐10‐CM) diagnostic codes (Table S1).

Covariates

Patient demographics, comorbidities, and nonstudy medications were identified as covariates (Table 1). Patient demographics included age and sex. Comorbidities included hypertension, myocardial infarction, congestive heart failure, peripheral vascular disease, cerebrovascular disease, diabetes, chronic pulmonary disease, peptic ulcer disease, chronic liver disease, ischemic stroke, chronic kidney disease, anemia, rheumatic disease, and malignancy. Patients were considered to have a comorbidity when they had at least 2 outpatient diagnoses or 1 inpatient diagnosis in the previous year. The components of the CHA₂DS₂‐VASc score, HAS‐BLED score, and the Charlson score were assessed. The use of medications, which included antiplatelet agents, rate control agents, lipid‐lowering agents, insulin, antihypertensives, glucocorticoids, nonsteroidal anti‐inflammatory drugs, and warfarin, was retrieved based on claims data within 6 months before and after the index date. Patients were followed from January 1, 2012 until December 31, 2018 with at least 3 months of follow‐up.

Table 1.

Baseline Characteristics of the Study Patients

Characteristic	DOAC alone		Amiodarone + DOAC		Propafenone + DOAC		Dronedarone + DOAC		Flecainide + DOAC		Sotalol + DOAC
	n=42 772		n=24 142		n=8631		n=2784		n=297		n=179
Men (n, %)	25 647	59.96%	13 523	56.01%	4474	51.84%	1399	50.25%	163	54.88%	86	48.08%
Age, y, mean±SD	75±10		75±10		72±10		76±8		71±10		72±10
Comorbidities, (n, %)
Hypertension	35 922	83.98%	20 551	85.13%	7069	81.90%	2374	85.27%	237	79.80%	142	79.33%
Myocardial infarction	1611	3.77%	1451	6.01%	179	2.07%	98	3.52%	7	2.36%	7	3.91%
Congestive heart failure	16 213	37.91%	9585	39.70%	2181	25.27%	736	26.44%	67	22.56%	62	34.64%
Peripheral vascular disease	3792	8.87%	2437	10.09%	707	8.19%	258	9.27%	24	8.08%	9	5.03%
Cerebrovascular disease	19 537	45.68%	10 560	43.74%	2998	34.74%	987	35.45%	99	33.33%	66	36.87%
Diabetes	15 652	36.59%	9269	38.39%	2802	32.46%	966	34.70%	91	30.64%	60	33.52%
Chronic pulmonary disease	14 366	33.59%	8864	36.72%	2829	32.78%	944	33.91%	105	35.35%	56	31.28%
Peptic ulcer disease	15 935	37.26%	10 426	43.19%	3825	44.32%	1353	48.60%	136	45.79%	73	40.78%
Chronic liver disease	8617	20.15%	5562	23.04%	2145	24.85%	650	23.35%	86	28.96%	38	21.23%
Ischemic stroke	15 067	35.23%	7744	32.08%	1949	22.58%	614	22.05%	54	18.18%	39	21.79%
Chronic kidney disease	9165	21.43%	6081	25.19%	197	2.28%	612	21.98%	69	23.23%	32	17.88%
Anemia	4492	10.50%	3126	12.95%	912	10.57%	374	13.43%	27	9.09%	21	11.73%
Rheumatic disease	1819	4.25%	1249	5.17%	506	5.86%	168	6.03%	18	6.06%	9	5.03%
Malignancy	3782	8.84%	2516	10.42%	828	9.59%	293	10.52%	34	11.45%	12	6.70%
CHA2DS2‐VASc, mean±SD	4.29±1.71		4.35±1.81		3.79±1.69		4.17±1.62		3.51±1.69		3.94±1.6
HAS‐BLED, mean±SD	3.02±1.17		3.03±1.19		2.76±1.14		2.99±1.08		2.63±1.2		2.67±1.08
Charlson score, mean±SD	3.25±2.20		3.57±2.36		2.91±2.15		3.14±2.2		2.95±2.14		2.79±2.03
Medications (n, %)
Aspirin	19 392	45.34%	9992	41.39%	3222	37.33%	1043	37.46%	82	27.61%	51	28.49%
Clopidogrel	4426	10.35%	3207	13.28%	648	7.51%	364	13.07%	23	7.74%	19	10.61%
Ticlopidine	901	2.11%	578	2.39%	235	2.72%	72	2.59%	8	2.69%	10	5.59%
Ticagrelor	264	0.62%	505	2.09%	36	0.42%	19	0.68%	≤3	≤1.01%	≤3	≤1.68%
Bisoprolol	15 843	37.04%	9707	40.21%	2854	33.07%	803	28.84%	113	38.05%	46	25.70%
Digoxin	7617	17.81%	3984	16.50%	700	8.11%	152	5.46%	20	6.73%	15	8.38%
Diltiazem	7193	16.82%	6584	27.27%	1693	19.62%	520	18.68%	60	20.20%	33	18.44%
Propranolol	4177	9.77%	347	1.44%	126	1.46%	44	1.58%	7	2.36%	≤3	≤1.68%
Erythromycin	1068	2.50%	1073	4.44%	163	1.89%	43	1.54%	≤3	≤1.01%	≤3	≤1.68%
Atorvastatin	6557	15.33%	3893	16.13%	1906	22.08%	470	16.88%	57	19.19%	23	12.85%
Pitavastatin	1024	2.39%	653	2.70%	201	2.33%	72	2.59%	17	5.72%	5	2.79%
Ezetimib	922	2.16%	661	2.74%	239	2.77%	73	2.62%	7	2.36%	5	2.79%
Insulin	2991	6.99%	3404	14.10%	371	4.30%	138	4.96%	11	3.70%	6	3.35%
Irbesartan	2373	5.55%	1222	5.06%	492	5.70%	185	6.65%	22	7.41%	7	3.91%
Losartan	4586	10.72%	2506	10.38%	852	9.87%	258	9.27%	46	15.49%	9	5.03%
Olmesartan	3172	7.42%	1902	7.88%	603	6.99%	187	6.72%	20	6.73%	17	9.50%
Glucocorticoids	3619	8.46%	2925	12.12%	626	7.25%	206	7.40%	16	5.39%	13	7.26%
NSAID	9727	22.74%	6606	27.36%	2133	24.71%	630	22.63%	79	26.60%	40	22.35%
DOAC
Dabigatran	15 121	35.35%	8135	33.70%	3234	37.47%	459	16.49%	67	22.56%	56	31.28%
Rivaroxaban	17 976	42.03%	10 639	44.07%	3323	38.50%	1445	51.90%	132	44.44%	75	41.90%
Apixaban	5799	13.56%	3457	14.32%	1294	14.99%	485	17.42%	44	14.81%	36	20.11%
Edoxaban	3876	9.06%	1911	7.92%	780	9.04%	395	14.19%	54	18.18%	12	6.70%
Follow‐up, y, mean±SD	2.56±1.73		2.26±1.69		2.22±1.69		2.07±1.61		1.83±1.44		1.82±1.68

DOAC indicates direct oral anticoagulant.

Statistical Analysis

To reduce the influence of confounders when comparing outcomes, propensity score matching (PSM) was performed using the baseline characteristics listed in Table 1 as covariates. The propensity scores were estimated for both groups of each comparative analysis using a logistic regression model (Figure 1). The matching was processed using a greedy nearest neighbor algorithm with a caliper of 0.1 times the SD of the logit of the propensity score, and a matching ratio of 1.

Patients concomitantly taking DOACs with amiodarone, propafenone, dronedarone, flecainide, or sotalol were compared against matched patients taking DOACs alone. In a secondary analysis, patients concomitantly taking each DOAC with an AAD were also compared against patients taking each DOAC alone. Furthermore, patients were stratified to <80 years old and ≥80 years old and were assessed on their risks of intracranial bleeding and gastrointestinal bleeding. The Kaplan‐Meier method and Cox proportional hazards model were used to compare the study outcome. Patients who switched AADs or DOACs during the treatment of AF were censored. Events that occurred within 7 days after the switch were considered to be related to the drug before the switch. Statistical significance was set at P<0.05. All statistical operations were performed using SAS version 9.4.

RESULTS

Study Population

In total, 142 255 patients with AF took DOACs from 2012 to 2018. After applying exclusion criteria, 78 805 patients with AF taking DOACs were studied. Among these patients, there were 24 142 patients taking amiodarone, 8631 patients taking propafenone, 2784 patients taking dronedarone, 297 patients taking flecainide, 177 patients taking sotalol, and 42 772 patients not taking any AAD, that is, on DOACs alone. Table 1 shows the baseline characteristics of these patients, which include demographics, comorbidities, and nonstudy medications. For each patient taking both DOACs and AADs concomitantly, we performed PSM to find 1 comparable patient taking DOACs alone (Figure 1). The quality of PSM is shown in Figure S1. The baseline characteristics of patients taking both DOACs and AADs and of matched patients taking DOACs alone are shown in Table S2 (amiodarone + DOACs versus DOACs alone), Table S3 (propafenone + DOACs versus DOACs alone), Table S4 (dronedarone + DOACs versus DOACs alone), Table S5 (flecainide + DOACs versus DOACs alone), and Table S6 (sotalol + DOACs versus DOACs alone).

Bleeding Associated With DOAC–AAD Interactions

As shown in Figure 2, after PSM, there was an increased risk of major bleeding for patients concomitantly taking DOACs and amiodarone compared with patients taking DOACs alone (hazard ratio [HR], 1.13 [95% CI, 1.04–1.23]; P=0.0044). When compared with patients taking DOACs alone, there was no significant difference in the risk of major bleeding for patients concomitantly taking DOACs with propafenone (HR, 1.03 [95% CI, 0.88–1.20]; P=0.7243), with dronedarone (HR, 0.87 [95% CI, 0.68–1.10]; P=0.2471), with flecainide (HR, 0.59 [95% CI, 0.23–1.56]; P=0.2893), or with sotalol (HR, 0.70 [95% CI, 0.21–2.41]; P=0.5770). However, the low patient number in the flecainide and sotalol groups limited data interpretation. Furthermore, recurrent event analysis showed that the concomitant use of amiodarone with DOACs was associated with a higher risk of recurrent major bleeding event when compared with DOACs alone (HR, 1.14 [95% CI, 1.05–1.25]; P=0.0032) (Table S7). No significant difference was observed with the concomitant use of DOACs with propafenone or dronedarone after PSM.

Figure 2. Cox proportional hazard ratio of the event of major bleeding in patients with atrial fibrillation on DOACs concomitantly taking dronedarone, amiodarone, propafenone, flecainide, or sotalol compared with propensity‐matched patients on DOACs alone (A).

The major bleeding events are further separated into intracranial bleeding (B) and gastrointestinal bleeding (C). AAD indicates antiarrhythmic drug; DOACs, direct oral anticoagulants; and HR, hazard ratio.

In regard to intracranial bleeding, a significantly higher risk when treated with amiodarone in combination with DOACs (HR, 1.36 [95% CI, 1.19–1.54]; P<0.0001) compared with those on DOACs alone was found. In contrast, the risks associated with propafenone, dronedarone, flecainide, or sotalol were not significantly different from the DOACs‐alone group, with hazard ratios of 1.14 (95% CI, 0.91–1.44; P=0.2609), 0.83 (95% CI, 0.56–1.24; P=0.3597), 0.38 (95% CI, 0.08–1.79; P=0.2233), and 2.19 (95% CI, 0.32–15.05; P=0.4248), respectively. In regard to gastrointestinal bleeding, no significant differences were observed between patients taking any of the AADs in combination with DOACs and those on DOACs alone. The hazard ratios were 0.98 (95% CI, 0.88–1.09; P=0.6917) for amiodarone, 0.95 (95% CI, 0.77–1.18; P=0.6478) for propafenone, 0.89 (95% CI, 0.66–1.22; P=0.4745) for dronedarone, 0.82 (95% CI, 0.23–2.94; P=0.7613) for flecainide, and 0.23 (95% CI, 0.03–1.81; P=0.1628) for sotalol.

The analysis indicates that amiodarone, when coadministered with DOACs, is associated with a significantly increased risk of major bleeding, particularly due to the elevated risk of intracranial bleeding. Other AADs, including propafenone, dronedarone, flecainide, and sotalol, did not significantly increase the risk of either major, intracranial, or gastrointestinal bleeding when used with DOACs.

Patients concomitantly taking each DOAC and AAD were also compared against patients taking each DOAC alone. As shown in Figure S2, there was a difference in the risk of major bleeding associated with the use of amiodarone and dabigatran when compared with dabigatran alone (HR, 1.19 [95% CI, 1.03–1.38]; P=0.0175). However, there was no significant difference in the risk of major bleeding associated with the use of amiodarone and rivaroxaban (HR, 1.10 [95% CI, 0.98–1.25]; P=0.1102), apixaban (HR, 1.14 [95% CI, 0.90–1.43]; P=0.2794), or edoxaban (HR, 1.03 [95% CI, 0.73–1.45]; P=0.8706) when compared with each DOAC alone. As shown in Figure S3, there was also a difference in the risk of major bleeding associated with the use of propafenone and dabigatran when compared with dabigatran alone (HR, 1.38 [95% CI, 1.06–1.80]; P=0.0170). However, there was no significant difference in the risk of major bleeding associated with the use of propafenone and rivaroxaban (HR, 0.93 [95% CI, 0.73–1.20]; P=0.5966), apixaban (HR, 0.80 [95% CI, 0.52–1.22]; P=0.2953), or edoxaban (HR, 0.73 [95% CI, 0.42–1.28]; P=0.2705) when compared with each DOAC alone. As shown in Figure S4, there was no significant difference in the risk of major bleeding associated with the use of dronedarone and dabigatran (HR, 1.19 [95% CI, 0.69–2.07]; P=0.5314), rivaroxaban (HR, 0.82 [95% CI, 0.58–1.14]; P=0.2296), apixaban (HR, 1.06 [95% CI, 0.56–2.01]; P=0.8511), or edoxaban (HR, 0.48 [95% CI, 0.22–1.06]; P=0.0684) when compared with each DOAC alone. However, the low patient number or event rate in many groups may limit data interpretation.

The event free rate of major bleeding was also analyzed in these patients. There was a significant difference in the risk of bleeding associated with patients concomitantly taking DOACs and amiodarone when compared with patients on DOACs alone (P=0.0034). However, there was no significant difference associated with patients concomitantly taking DOACs and dronedarone, propafenone, flecainide, or sotalol when compared with DOACs alone (Figure 3). The low patient number in the flecainide and sotalol groups may limit data interpretation.

Figure 3. Kaplan‐Meier curve for event‐free rate of major bleeding in patients with atrial fibrillation on DOACs concomitantly taking amiodarone, propafenone, dronedarone, flecainide, or sotalol propensity‐matched to patients with DOACs alone.

DOACs indicates direct oral anticoagulants.

Patients were also stratified to <80 and ≥80 years old and assessed on their risks of intracranial bleeding and gastrointestinal bleeding. The incidence rates of intracranial bleeding and gastrointestinal bleeding appear to be higher in patients ≥80 years old either taking DOACs alone or taking DOACs concomitantly with AADs (Figures S5 through S8). In patients <80 years old, the concomitant use of amiodarone with dabigatran (HR, 1.35 [95% CI, 1.03–1.76]; P=0.0289), rivaroxaban (HR, 1.36 [95% CI, 1.09–1.69]; P=0.0060), or apixaban (HR, 1.67 [95% CI, 1.09–2.58]; P=0.0189) was associated with an increased risk of intracranial bleeding (Figure S5). In patients ≥80 years old, the concomitant use of amiodarone with dabigatran was associated with an increased risk of intracranial bleeding (HR, 1.76 [95% CI, 1.15–2.69]; P=0.0096) (Figure S6). There was no significant difference in intracranial bleeding associated with the concomitant use of DOACs with propafenone or dronedarone in both age groups.

In patients <80 years old, the concomitant use of dabigatran with amiodarone (HR, 1.41 [95% CI, 1.09–1.83]; P=0.0087) or propafenone (HR, 1.77 [95% CI, 1.10–2.85]; P=0.0189) was associated with an increased risk of gastrointestinal bleeding (Figure S7). In patients ≥80 years old, the concomitant use of dabigatran with amiodarone was associated with a lower risk of gastrointestinal bleeding (HR, 0.69 [95% CI, 0.52–0.93]; P=0.0151) (Figure S8). There was no significant difference in gastrointestinal bleeding associated with the concomitant use of DOACs with dronedarone in both age groups. The low patient number of the propafenone and dronedarone groups may limit the data interpretation.

Doses of DOACs and Interaction With Bleeding Outcome

After PSM, the starting dose (205.71±52.70–228.5±38.79 mg) and average dose during follow‐up (205.39±50.57–226.4±41.42 mg) of dabigatran in patients taking amiodarone, propafenone, dronedarone, flecainide, and sotalol were comparable. The starting dose (13.24±3.69–14.08±4.17 mg) and average dose during follow‐up (12.95±3.21–13.56±3.06 mg) of rivaroxaban in patients taking amiodarone, propafenone, dronedarone, flecainide, and sotalol were comparable. The starting dose (9.65±1.27–10.00±0.00 mg) and average dose during follow‐up (9.68±1.10–9.88±0.61 mg) of apixaban in patients taking amiodarone, propafenone, dronedarone, flecainide, and sotalol were comparable. The starting dose (42.94±15.01–54.55±12.14 mg) and average dose during follow‐up (41.51±13.68–51.03±11.83 mg) of edoxaban in patients taking amiodarone, propafenone, dronedarone, flecainide, and sotalol had a wider variation (Tables 2 and 3). Overall, the doses of DOACs were comparable among groups.

Table 2.

DOAC Starting Daily Doses at Index

Treatment Groups	Before PSM				After PSM
	Dabigatran	Rivaroxaban	Apixaban	Edoxaban	Dabigatran	Rivaroxaban	Apixaban	Edoxaban
DOACs alone vs amiodarone + DOACs
DOACs alone	218.4±40.37	13.96±3.64	9.71±1.18	49.11±14.6	218.13±40.64	13.93±3.78	9.71±1.17	48.94±14.61
Amiodarone + DOACs	216.33±42.82	13.87±3.63	9.75±1.09	48.09±14.77	216.31±43.00	13.87±3.60	9.76±1.08	48.11±14.76
DOACs alone vs propafenone + DOACs
DOACs alone	218.4±40.37	13.96±3.64	9.71±1.18	49.11±14.6	220.43±40.00	14.07±3.82	9.72±1.15	49.12±14.50
Propafenone + DOACs	219.44±38.08	13.82±3.54	9.74±1.11	49.34±14.37	219.44±38.08	13.82±3.54	9.74±1.11	49.34±14.37
DOACs alone vs dronedarone + DOACs
DOACs alone	218.40±40.37	13.96±3.64	9.71±1.18	49.11±14.60	218.11±40.68	13.88±3.82	9.62±1.33	47.49±14.81
Dronedarone + DOACs	205.71±52.70	13.24±3.69	9.65±1.27	43.99±15.04	205.71±52.70	13.24±3.69	9.65±1.27	43.99±15.04
DOACs alone vs flecainide + DOACs
DOACs alone	218.4±40.37	13.96±3.64	9.71±1.18	49.11±14.6	221.9±32.68	13.76±3.93	10.00±0.00	48.38±15.04
Flecainide + DOACs	228.5±38.79	13.41±2.73	10.00±0.00	42.94±15.01	228.5±38.79	13.41±2.73	10.00±0.00	42.94±15.01
DOACs alone vs sotalol + DOACs
DOACs alone	218.4±40.37	13.96±3.64	9.71±1.18	49.11±14.6	221.67±45.63	13.58±3.56	9.67±1.27	43.33±15.81
Sotalol + DOACs	220.38±28.89	14.08±4.17	9.85±0.86	54.55±12.14	220.38±28.89	14.08±4.17	9.85±0.86	54.55±12.14

Data are presented as mean±SD. DOAC indicates direct oral anticoagulant; and PSM, propensity score matching.

Table 3.

DOAC Average Daily Doses During Follow‐Up

Treatment Groups	Before PSM				After PSM
	Dabigatran	Rivaroxaban	Apixaban	Edoxaban	Dabigatran	Rivaroxaban	Apixaban	Edoxaban
DOACs alone vs amiodarone + DOACs
DOACs alone	218.38±37.98	13.71±3.17	9.69±1.07	48.61±14.07	217.98±38.34	13.69±3.27	9.7±1.05	48.31±14.1
Amiodarone + DOACs	215.51±39.62	13.55±3.07	9.73±0.99	47.45±14.20	215.47±39.84	13.56±3.06	9.74±0.98	47.47±14.18
DOACs alone vs propafenone + DOACs
DOACs alone	218.38±37.98	13.71±3.17	9.69±1.07	48.61±14.07	220.42±37.75	13.79±3.25	9.67±1.09	48.89±13.86
Propafenone + DOACs	219.63±35.67	13.53±3.03	9.74±0.99	48.79±13.86	219.63±35.67	13.53±3.03	9.74±0.99	48.79±13.86
DOACs alone vs dronedarone + DOACs
DOACs alone	218.38±37.98	13.71±3.17	9.69±1.07	48.61±14.07	215.95±37.24	13.54±3.23	9.64±1.14	46.81±14.32
Dronedarone + DOACs	205.39±50.57	12.95±3.21	9.68±1.10	43.05±14.35	205.39±50.57	12.95±3.21	9.68±1.10	43.05±14.35
DOACs alone vs flecainide + DOACs
DOACs alone	218.38±37.98	13.71±3.17	9.69±1.07	48.61±14.07	223.26±25.13	13.37±3.13	9.83±0.52	49.08±14.09
Flecainide + DOACs	226.44±41.42	13.45±2.55	9.88±0.61	41.51±13.68	226.44±41.42	13.45±2.55	9.88±0.61	41.51±13.68
DOACs alone vs sotalol + DOACs
DOACs alone	218.38±37.98	13.71±3.17	9.69±1.07	48.61±14.07	222.47±45.01	13.55±3.42	9.76±0.99	43.33±15.81
Sotalol + DOACs	218.58±27.88	13.30±3.54	9.86±0.85	51.03±11.83	218.58±27.88	13.30±3.54	9.86±0.85	51.03±11.83

Data are presented as mean±SD. DOAC indicates direct oral anticoagulant; and PSM, propensity score matching.

After PSM, the starting daily doses and average daily doses during follow‐up of dabigatran, rivaroxaban, and edoxaban were lower in patients ≥80 years old than in patients <80 years old (Tables S8 and S9). The doses of apixaban seemed to be comparable between both age groups. The doses of DOACs were comparable between comparative groups of amiodarone (eg, amiodarone + DOACs versus DOACs alone) and dronedarone, and thus should not have affected the outcome. On the other hand, the doses of DOACs were reduced when used concomitantly with propafenone in both age groups.

We also performed a comprehensive evaluation of major bleeding risks associated with specific doses of DOACs when used concomitantly with amiodarone, propafenone, and dronedarone (Figure S9).

For amiodarone, the data indicate a significant increase in bleeding risk when combined with the higher dose of dabigatran (150 mg twice daily), with an HR of 1.77 (95% CI, 1.08–2.90; P=0.0244). This elevated risk was not observed with the lower dose of dabigatran (110 mg twice daily) or with other DOACs, suggesting that dose adjustments are critical. Similarly, rivaroxaban, apixaban, and edoxaban did not show significant increases in bleeding risk across their dosing ranges when combined with amiodarone, indicating a relatively safer profile.

For propafenone, the combination with 110 mg dabigatran showed a significant bleeding risk (HR, 1.41 [95% CI, 1.07–1.85]; P=0.0147), whereas the higher dose of dabigatran did not exhibit the same risk. Other DOACs, including rivaroxaban, apixaban, and edoxaban, generally showed no significant increase in bleeding risk when used with propafenone, highlighting their safer profiles. Notably, propafenone combined with rivaroxaban across different doses (10, 15, 20 mg), apixaban (2.5, 5 mg), and edoxaban (30, 60 mg) did not significantly elevate bleeding risks, suggesting these combinations may be preferable in clinical practice.

Dronedarone presents a relatively safer profile when combined with DOACs. Neither dose of dabigatran (110 or 150 mg) showed a significant increase in bleeding risk when used with dronedarone. Similarly, rivaroxaban, apixaban, and edoxaban exhibited no significant bleeding risks across their respective doses. Interestingly, the combination of dronedarone with 30 mg edoxaban showed a significantly lower bleeding risk when compared with 30 mg edoxaban alone (HR, 0.31 [95% CI, 0.11–0.88]; P=0.0281). These findings indicate that dronedarone might be a safer alternative for patients requiring AAD therapy alongside anticoagulation, particularly where bleeding risks are a concern. This is especially relevant given the trend toward increased use of rhythm control strategies in AF management. Furthermore, interaction tests showed that the dose of all DOACs did not affect the major bleeding outcome. Nonetheless, it is important to note that the low number of patients in many comparative groups may limit data interpretation.

DOAC Switching

After PSM, there were 48.91%, 47.12%, 62.93%, 53.73%, and 55.74% of patients taking dabigatran who switched to another DOAC when taking amiodarone, propafenone, dronedarone, flecainide, or sotalol, respectively, during follow‐up. There were 30.99%, 32.99%, 38.89%, 31.06%, and 47.06% of patients taking rivaroxaban who switched to another DOAC when taking amiodarone, propafenone, dronedarone, flecainide, or sotalol, respectively. There were 30.45%, 30.36%, 32.16%, 29.55%, and 35.14% of patients taking apixaban who switched to another DOAC when taking amiodarone, propafenone, dronedarone, flecainide, or sotalol, respectively. There were 24.21%, 24.23%, 23.29%, 25.93%, and 35.71% of patients taking edoxaban who switched to another DOAC when taking amiodarone, propafenone, dronedarone, flecainide, or sotalol, respectively (Table S10).

Initial and Average Daily Doses of AADs

After PSM, the initial dose of amiodarone was 258.50±112.64 mg, and the average dose of amiodarone during follow‐up was 237.47±83.85 mg. The initial dose of propafenone was 309.12±84.47 mg, and the average dose of propafenone during follow‐up was 306.22±79.24 mg. The initial dose of dronedarone was 728.85±153.32 mg, and the average dose of dronedarone during follow‐up was 722.97±148.85 mg. The initial dose of flecainide was 192.52±36.43 mg, and the average dose of flecainide during follow‐up was 188.25±39.45 mg. The initial dose of sotalol was 279.77±78.41 mg, and the average dose of sotalol during follow‐up was 275.59±78.34 mg (Table S11).

After patients were stratified to <80 and ≥80 years old, we found that there was no difference in the starting daily doses and average daily doses during follow‐up of AADs between the 2 age groups, except that the dose of dronedarone, when used with dabigatran, was reduced in patients ≥80 years old (Tables S12 and S13).

Dose Change and Drug Switching of AAD

After PSM, there were 29.01% of patients taking amiodarone who changed doses, and there were 16.59% of patients taking amiodarone who switched to another AAD. There were 21.85% of patients taking propafenone who changed doses, and there were 40.27% of patients taking propafenone who switched to another AAD. There were 10.48% of patients taking dronedarone who changed doses, and there were 44.18% of patients taking dronedarone who switched to another AAD. There were 12.99% of patients taking flecainide who changed doses, and there were 56.57% of patients taking flecainide who switched to another AAD. There were 10.40% of patients taking sotalol who changed doses, and there were 73.18% of patients taking sotalol who switched to another AAD (Table S14).

DISCUSSION

This is the first Taiwan national cohort study that investigates the real‐world risk of major bleeding in patients with AF taking DOACs and AADs concomitantly, and therefore presumably arising from DDIs, and the first such analysis that shows the dosing of DOACs when used with AADs. Our findings showed that when used with DOACs, amiodarone was the only AAD associated with increased major bleeding events, whereas dronedarone and propafenone were not significantly associated with increased bleeding events. Furthermore, the secondary analysis of the concomitant use of AAD with each DOAC showed that the concomitant use of amiodarone with dabigatran was associated with an increased risk of major bleeding when compared with dabigatran alone. In the analysis where DOACs were pooled together, the distribution and doses of individual DOACs were similar between groups, and the interaction tests showed that the doses of DOACs did not significantly affect the major bleeding risk. The doses of DOACs when used with different AADs were also comparable regardless of known interactions.

As shown in Figure 2, PSM revealed that patients on DOACs with amiodarone have an increased risk of major bleeding compared with those on DOACs alone (HR, 1.13 [95% CI, 1.04–1.23]; P=0.0044), with a similar trend in recurrent major bleeding events (HR, 1.14 [95% CI, 1.05–1.25]; P=0.0032). In contrast, the use of other antiarrhythmic drugs, such as propafenone, dronedarone, flecainide, and sotalol, did not significantly alter bleeding risks, although patient numbers for flecainide and sotalol were low. Notably, amiodarone was particularly associated with an increased risk of intracranial bleeding (HR, 1.36 [95% CI, 1.19–1.54]; P<0.0001). These findings highlight the need for careful consideration of amiodarone in patients at high bleeding risk and underscore its significant impact on bleeding outcomes compared with other AADs when used with DOACs.

In a small cohort study in Sweden, 33 patients concomitantly treated with 110 mg dabigatran twice daily and 400 mg dronedarone twice daily had a median trough plasma concentration similar to that of 150 mg dabigatran twice daily in earlier studies. ¹⁷ These results suggest that the concomitant use of dronedarone and dabigatran at a reduced dose may have a similar bleeding risk as dabigatran at a standard dose. A Taiwan national cohort study investigated the bleeding risk associated with the concomitant use of DOACs and commonly prescribed medications. ¹⁸ The study showed that when compared with the use of DOACs alone, there was no significant difference in the risk of major bleeding for the concomitant use of DOACs with dronedarone (adjusted rate ratio [aRR], 0.89 [95% CI, 0.71–1.13]), but there was a higher bleeding risk associated with the concomitant use of DOACs with amiodarone (aRR, 1.37 [95% CI, 1.25–1.50]). A Swedish nationwide health registry study investigated the risk of bleeding in patients who used dronedarone in combination with apixaban or warfarin. ¹⁹ The bleeding risk was lower when dronedarone was used concomitantly with apixaban than with warfarin (HR, 0.66 [95% CI, 0.35–1.23]; P=0.121). Another small Taiwanese retrospective study showed that the concomitant use of rivaroxaban with dronedarone, amiodarone, or propafenone and the use of rivaroxaban alone had a similar incidence of safety end point, which was a composite of major and minor bleeding (P=0.892). ²⁰ A US retrospective study showed that when compared with apixaban alone, the concomitant use of dronedarone and apixaban was not associated with an increased risk of major bleeding (adjusted HR [aHR], 0.69 [95% CI, 0.40–1.17]; P=0.16). ²¹ However, the concomitant use of dronedarone and dabigatran was associated with an increased risk of gastrointestinal bleeding but not overall bleeding (aHR bleeding, 1.18 [95% CI, 0.89–1.56]; P=0.26; aHR gastrointestinal bleeding, 1.40 [95% CI, 1.01–1.93]; P=0.04). The concomitant use of dronedarone and rivaroxaban was also associated with an increased risk of overall bleeding driven by gastrointestinal bleeding (aHR bleeding, 1.31 [95% CI, 1.01–1.69]; P=0.04; aHR gastrointestinal bleeding, 1.39 [95% CI, 0.98–1.95]; P=0.06).

Our study was consistent with the previous Taiwan national cohort study, which showed that the concomitant use of DOACs with amiodarone, but not with dronedarone, was associated with an increased risk of major bleeding. ¹⁸ The previous study also showed that there was an increased risk of major bleeding associated with the concomitant use of amiodarone and dabigatran (aRR, 1.36 [99% CI, 1.17–1.59]), rivaroxaban (aRR, 1.38 [99% CI, 1.21–1.58]), or possibly apixaban (aRR, 1.30 [99% CI, 0.98–1.72]) when compared with each DOAC alone. Our study only showed an increased major bleeding associated with the concomitant use of amiodarone and dabigatran (HR, 1.19 [95% CI, 1.03–1.38]), but not rivaroxaban or apixaban, when compared with each DOAC alone. The discrepancy may be caused by the use of different statistical methods. Another reason could be that the drug doses were not known in the previous study and thus may be different between groups, whereas in our study, the drug doses were comparable between different groups. We also performed interaction tests and found that the DOAC doses did not significantly affect the outcome. Another Taiwanese retrospective study also showed that there was no difference in the bleeding end point among patients on rivaroxaban concomitantly taking or not taking AADs. ²⁰ In contrast, our study may differ from the US retrospective study, which showed that the concomitant use of dronedarone with dabigatran or rivaroxaban was associated with an increased risk of gastrointestinal bleeding. ²¹ There were a few possible reasons; first, racial or ethnic differences; second, our analysis of the concomitant use of dronedarone with each DOAC may be underpowered; third, dabigatran or rivaroxaban were used at reduced doses in our study, whereas the doses used in the US study were not known and were likely higher.

Our study showed that patients ≥80 years old appeared to have a higher incidence of intracranial bleeding and gastrointestinal bleeding than patients <80 years old. This is an expected observation, because aging is known to be associated with increased bleeding complications with oral anticoagulant use. ²² , ²³ , ²⁴ In the age groups <80 and ≥80 years old, the concomitant use of amiodarone with some DOACs was associated with increased risk of intracranial bleeding. There was no significant difference in intracranial bleeding associated with the concomitant use of DOACs with propafenone or dronedarone. It thus appears that the drug–drug interaction between amiodarone and DOACs may have increased the risk of intracranial bleeding. Intracranial bleeding is a particularly devastating complication due to its high mortality and long‐term disability rates, which makes this finding critically important. Existing risk scores, such as CHA₂DS₂‐VASc and HAS‐BLED, do not adequately predict the risk of intracranial bleeding in this context, highlighting the need for additional caution and possibly new predictive tools. ²⁵ Furthermore, although DOACs are generally preferred over vitamin K antagonists because of their lower incidence of intracranial hemorrhage, this advantage could be reduced when these drugs are used concomitantly with amiodarone. Consequently, the primary advantage of DOACs over vitamin K antagonists in terms of reducing cerebral hemorrhage is compromised, altering the risk–benefit analysis for clinicians.

The addition of amiodarone or propafenone to dabigatran was also associated with increased risk of gastrointestinal bleeding. However, it is interesting that the addition of amiodarone to DOACs did not increase the risk of gastrointestinal bleeding in patients ≥80 years old. We noted that there was already a high incidence rate of gastrointestinal bleeding among the older patients on DOACs alone. A previous study showed that in patients ≥80 years old, the use of DOACs was associated with lower risk of intracranial bleeding but higher risk of gastrointestinal bleeding than vitamin K antagonists. ²⁶ It has also been previously reported that DOACs may increase the risk of gastrointestinal bleeding, probably because they are direct anticoagulants that are immediately active when absorbed by the gut. ²⁷

Certain combinations of AADs and DOACs were associated with a decreased risk of bleeding. In patients <80 years old, propafenone combined with apixaban was associated with a reduced risk of gastrointestinal bleeding, possibly due to a pharmacodynamic interaction that influences apixaban's anticoagulant effects (Figure S7). Similarly, in patients ≥80 years old, amiodarone combined with dabigatran was associated with a lower risk of gastrointestinal bleeding, potentially through altering dabigatran's pharmacokinetics, although it is important to consider the increased risk of intracerebral bleeding with this combination (Figure S8). Additionally, the combination of dronedarone and a low dose of edoxaban showed a decreased risk of major bleeding, likely due to modified pharmacokinetics that reduced edoxaban's systemic exposure (Figure S9). These findings suggest that interactions between certain AADs and DOACs might be associated with a lower bleeding risk by altering drug absorption, metabolism, or elimination, emphasizing the need for individualized therapy, especially in older populations. Further studies are essential to confirm these interactions and refine treatment approaches to balance efficacy and safety.

Amiodarone was associated with the highest rate of dose change among all AADs in our study, possibly due to down‐titration of amiodarone after initiation, or because the physicians perceived that the concomitant use of amiodarone and DOAC was associated with bleeding risk. On the other hand, amiodarone was least likely to be switched to another AAD, possibly because the physicians were familiar with this commonly prescribed drug.

Our study revealed that the DOAC doses were similar between patients concomitantly taking DOACs with AADs and patients taking DOACs alone, regardless of whether there are known DDIs (as in the case for amiodarone and dronedarone, where P‐glycoprotein competition and CYP3A4 inhibition underlie the DDIs ¹³ , ¹⁴ , ¹⁵ , ¹⁶ ) or no known DDIs (as in the case for propafenone, flecainide, and sotalol). Additionally, when used with different AADs, the doses of DOACs were also comparable regardless of known interactions. Previous studies at this scale lacked the dosage data. In this study, the mean daily dose of dabigatran and rivaroxaban during follow‐up were 205 to 226 mg and 13.0 to 13.8 mg, respectively. These doses were similar to, or lower than, the reduced dose recommended in the 2020 European Society of Cardiology guidelines and 2021 European Heart Rhythm Association Practical Guide. ¹ , ¹³ The mean daily dose of edoxaban was 43.1 to 51.0 mg, which was in between the standard and lower dose recommended in the 2020 European Society of Cardiology guidelines. In contrast, apixaban was used closer to the standard dose, at a mean daily dose of 9.64 to 9.88 mg. Physicians usually prescribe lower doses of anticoagulants for patients with AF due to the preventative nature of such prescriptions. This is especially prevalent in Asian countries, because Asian patients have a smaller body size and are presented with more risk factors such as old age and chronic kidney diseases. ²⁸ , ²⁹ Furthermore, the Japanese Rivaroxaban Once‐daily Oral Direct Factor Xa Inhibition Compared with Warfarin in Patients with Non‐Valvular Atrial Fibrillation (J‐ROCKET AF) study found that Japanese patients treated with 15 mg of rivaroxaban had a similar pharmacokinetic profile as in White patients treated with 20 mg of rivaroxaban. ³⁰ A subgroup analysis of the Randomized Evaluation of Long‐term Anticoagulant Therapy with Dabigatran Etexilate (RE‐LY) study found that low‐dose dabigatran was associated with a decreased risk of major bleeding compared with warfarin among Asian patients. ³¹ These studies may have prompted the physicians to use a lower dose of rivaroxaban or dabigatran in Asian patients. In contrast, a subgroup analysis of the Apixaban for the Reduction in Stroke and Other Thromboembolic Events in Atrial Fibrillation (ARISTOTLE) study showed that the use of a standard dose of apixaban was associated with a significantly lower risk of major bleeding compared with warfarin among Asian patients. ³² This study may have prompted the physicians to use a standard dose of apixaban in Taiwan. This was later supported by a retrospective study in Taiwan, which showed that the use of a standard dose of apixaban was associated with a lower risk of ischemic stroke/systemic embolism, major bleeding, or mortality when compared with warfarin. ³³ The risk of thromboembolism varies among different ethnic groups, in part due to genetic polymorphisms within the coagulation system. ³⁴ , ³⁵ This variability underscores the necessity of optimizing anticoagulant doses across ethnicities to balance the risks of stroke and bleeding effectively.

For patients on amiodarone, particularly those prescribed dabigatran, clinicians should consider using the lower dose of dabigatran to mitigate bleeding risks. The combination of high‐dose dabigatran with amiodarone poses substantial risks and should be approached with caution. This necessitates a more tailored approach in prescribing anticoagulation therapy, considering both the AAD and DOAC doses to optimize patient safety. The data also suggest that rivaroxaban, apixaban, and edoxaban may be safer alternatives for patients on amiodarone, because they do not significantly increase bleeding risks across their dosing ranges. Propafenone's interaction with dabigatran also warrants careful consideration. However, the safer profile observed with rivaroxaban, apixaban, and edoxaban suggests these might be preferable options for patients requiring propafenone. This highlights the necessity of individualized treatment plans that consider the specific DOAC and its dose when used with propafenone. Dronedarone's safer profile across various DOACs is reassuring, indicating it may be a preferable choice for patients at higher risk of bleeding. The lack of significant bleeding risks with dabigatran, rivaroxaban, apixaban, and edoxaban suggests that dronedarone can be safely combined with these anticoagulants. This is particularly relevant in the context of increasing rhythm control strategies for managing AF. Overall, these detailed dose‐specific analyses provide critical insights that can guide clinicians in optimizing anticoagulation therapy, ensuring both efficacy and safety for patients with AF.

The 2021 European Heart Rhythm Association Practical Guide has laid out the recommendations for the concomitant use of AADs and DOACs largely based on pharmacokinetic studies. ³⁶ , ³⁷ , ³⁸ Amiodarone increased the plasma level of dabigatran and edoxaban by 12% to 60% and 40%, respectively, and thus, these 2 combinations were given a caution recommendation. There were no data available on the concomitant use of amiodarone with apixaban or rivaroxaban, but a caution recommendation was also given. On the other hand, dronedarone increased the plasma level of dabigatran by 70% to 100%, and this combination was contraindicated. The combination of dronedarone and rivaroxaban was also contraindicated, although no data were available. Caution was advised when dronedarone was used with apixaban without any supporting evidence. Dose reduction of edoxaban was recommended when used concomitantly with dronedarone. However, we did not observe an increased bleeding risk, including intracranial bleeding and gastrointestinal bleeding, associated with the concomitant use of dronedarone and DOACs in our study, probably because the doses of DOACs were already reduced, except for apixaban. It is also possible that the analysis of dronedarone group was underpowered, given the low patient number when compared with the amiodarone group. In a recent study, when used concomitantly with DOACs, dronedarone was associated with lower risks of ischemic stroke/systemic embolism (HR, 0.62 [95% CI, 0.49–0.8]; P=0.0002), intracranial bleeding (HR, 0.58 [95% CI, 0.37–0.92]; P=0.0193), cardiovascular death (HR, 0.15 [95% CI, 0.10–0.22]; P<0.0001), all‐cause mortality (HR, 0.20 [95% CI, 0.16–0.24]; P<0.0001), or Major Adverse Cardiovascular Events (HR, 0.38 [95% CI, 0.31–0.47]; P<0.0001), when compared with amiodarone. ³⁹ In that study, there was no significant difference in major bleeding between dronedarone and amiodarone when used concomitantly with DOACs. However, there was a trend in favor of dronedarone (HR, 0.78 [95% CI, 0.59–1.04]; P=0.0918), which appears to be consistent with our study. Another previous study also showed that amiodarone used concomitantly with rivaroxaban or apixaban was associated with increased risk of bleeding‐related hospitalizations than flecainide or sotalol. ⁴⁰ These observations are important, because amiodarone is still a commonly prescribed AAD in the DOAC era. Crucially, our study showed that the concomitant use of amiodarone with some DOACs was associated with increased intracranial bleeding in both age groups of <80 and ≥80 years old. It is important that the choice and dose of DOACs and AADs are carefully considered for patients with AF who require the use of DOACs and who have a high bleeding risk.

Limitations

There are several limitations in this study. First, patient screening using ICD‐9‐CM and ICD‐10‐CM codes may contain missing cases when the conditions or diagnoses were not coded correctly. Second, using ICD‐9‐CM codes, AF was not coded in terms of paroxysmal, persistent, or permanent as in using ICD‐10‐CM codes. Nonetheless, it is assumed that physicians did not prescribe dronedarone for permanent AF. Third, the observational nature of the study made the causality of medications and outcome events not definitively established. Fourth, the time of exposure to DOACs was not matched between groups, which could introduce immortal time bias. Fifth, this study was conducted within a homogenous ethnic population and may not be applicable to other populations. Sixth, the timing of drug intake for AADs and DOACs was not known, which may affect the DOAC exposure according to a previous study of dronedarone. ⁴¹ Seventh, the covariates used in PSM do not contain information about the severity of the underlying diseases. Eighth, the study did not capture other confounders, such as alcohol use, which is independently associated with bleeding risk. Ninth, for the patient number or event rates of the flecainide and sotalol groups, many combinations of individual AADs and individual DOACs in the secondary analysis, and combinations of individual AADs with specific doses of DOACs, were too low for meaningful interpretation. Tenth, there were moderate imbalances in the flecainide and sotalol groups, which may limit the interpretation of these data.

CONCLUSIONS

In patients with paroxysmal or persistent AF, the concomitant use of DOACs with amiodarone, but not propafenone or dronedarone, was associated with an increased bleeding risk, when compared with DOACs alone. In particular, the concomitant use of some DOACs with amiodarone was associated with an increased risk of intracranial bleeding. This study provides new evidence to guide clinicians to tailor concomitant anticoagulation and antiarrhythmic therapies in AF. These results contribute new insights to the existing literature, offering guidance on safer medication practices for managing AF.

Sources of Funding

This investigator‐sponsored study received funding from Sanofi.

Disclosures

Drs Wu, Wang, and Chang received grants and support from Sanofi. The remaining authors have no disclosures to report.

Supporting information

Data S1