Summary
Objectives:
To compare the effects on cerebral health in atrial fibrillation patients receiving either warfarin or direct oral anticoagulants (DOACs). Atrial fibrillation is linked to a higher likelihood of both stroke and cognitive impairment. While warfarin has historically served as the leading anticoagulant treatment, DOACs are now more frequently viewed as alternatives with an improved safety profile.
Methods:
The methodology involved an extensive search of multiple databases, completed up to June 2024. This search yielded randomized controlled trials, which together enrolled over 60,000 participants between the ages of 70 and 84. The DOACs studied were dabigatran, rivaroxaban, apixaban, and edoxaban, given in various dosages. Warfarin was administered with dose adjustments to keep the international normalized ratio within a target range of 2.0 to 3.0.
Results:
With a relative risk of 0.81 (95% CI, p < 0.00001; I2 = 19%), the results showed that 20 high-dose DOACs considerably decreased the risk of stroke when compared to warfarin. However, there were no discernible variations for major bleeding (relative risk 0.89), moderate bleeding (relative risk 0.92), or cardiovascular events (relative risk 0.96).
Conclusion:
The DOACs are frequently more efficacious than warfarin in avoiding stroke and systemic embolism in patients with non-valvular atrial fibrillation. A significant decline in hemorrhagic stroke was the main cause of the pooled data's 19% relative risk reduction in stroke or systemic embolism. While bleeding outcomes varied among the individual DOACs, the overall findings support their favorable risk-benefit profile. Compared to warfarin, DOACs consistently showed fewer intracranial hemorrhages and a more favorable bleeding profile. These results support prioritizing DOACs in clinical practice, given their enhanced safety and reduced mortality rates. However, due to differences in bleeding risks among DOACs, treatment decisions should be tailored to individual patient factors.
Keywords: Stroke, Warfarin, Apixaban, Dabigatran, Bleeding
Introduction
Atrial fibrillation (AF) represents the world's most prevalent persistent arrhythmia, affecting more than 33 million people globally [1]. This condition markedly elevates the risk of ischemic stroke, making patients approximately 5 times more likely to experience this event compared to those without AF; consequently, stroke is considered the most serious cerebral complication of the arrhythmia [2]. The consequences extend beyond the acute phase, contributing to significant long-term disability, cognitive decline, and a substantial burden on healthcare systems [3,4].
To guide stroke prevention strategies in non-valvular AF, the CHA2DS2-VASc score is extensively used for thromboembolic risk stratification. This scoring system provides a straightforward and validated tool for identifying patients most likely to benefit from anticoagulation therapy [5]. The guidelines strongly recommend initiating anticoagulant treatment for patients with moderate-to-high risk scores, defined as ≥2 for males and ≥3 for females, to effectively prevent ischemic events [6]. Managing AF involves three main goals: preventing blood clots with anticoagulants, controlling the heart rate, and maintaining a normal heart rhythm. For many years, warfarin was the standard treatment for stroke prevention [7]. However, the advent of direct oral anticoagulants (DOACs), including rivaroxaban and dabigatran, has offered alternatives with more predictable pharmacokinetics, fewer interactions, and decreased intracranial hemorrhage (ICH) risk [8,9].
Despite the protective role of anticoagulation, AF patients remain at risk for various cerebral outcomes, including ischemic stroke, intracerebral hemorrhage, transient ischemic attacks, and vascular dementia [10,11]. These complications profoundly impact patient survival and their quality of life (QoL). Furthermore, patients with anticoagulated AF are more likely to experience hemorrhage and thromboembolic consequences, which raises the overall mortality rate, particularly in older persons with comorbidities [10,12]. Major bleeding, as defined by the International Society on Thrombosis and Haemostasis (ISTH), encompasses clinically apparent bleeding that meets one or more of the following criteria: it is fatal, occurs in a critical organ (such as intracranial, intraspinal, intraocular, retroperitoneal, intra-articular, or pericardial), requires reoperation or surgical intervention to control, leads to a hemoglobin decrease of at least 2 g/dL, necessitates a transfusion of at least 2 units of packed red blood cells, or presents as unexpected bleeding at a surgical site that demands clinical intervention [13]. Given the increased use of DOACs, it is critical to investigate how these medications compare to warfarin in terms of cerebral outcomes. This review aims to compare cerebral outcomes, with emphasis on stroke, in AF patients receiving warfarin versus DOAC therapy.
Method
Search strategy
This systematic review and meta-analysis was prospectively registered on the International Platform of Registered Systematic Review and Meta-analysis Protocols (INPLASY) under the unique registration number INPLASY202550083 [14].
The systematic review was conducted in accordance with the PRISMA guidelines [15]. To find all pertinent papers published up until June 2024, a thorough literature search was conducted across several electronic databases, including PubMed, Google Scholar, the Cochrane Library, and Embase. The reference lists of all chosen publications were also thoroughly examined to reduce the possibility of missing important research. The search strategy utilized a combination of keywords such as “Warfarin”, “DOAC”, “Acute Ischemic Stroke’, “Atrial Fibrillation”, and “Randomized Controlled Trials”, ensuring a comprehensive retrieval of studies.
Eligibility Criteria: Studies that were: (1) published in English, (2) designed as randomized controlled trials (RCTs), (3) involved adult participants (≥18 years) with confirmed AF, (4) compared warfarin with DOACs, and (5) reported at least one clinical outcome of interest (stroke incidence, systemic embolism, or bleeding events) were included. Studies were excluded if they were non-RCTs, involved pediatric populations, lacked relevant treatment comparisons, did not report necessary outcomes, had a high risk of bias or poor methodological quality.
Data Extraction: The search results were imported into Rayyan software to streamline the screening process, eliminate duplicate records, and facilitate study selection [16]. Two independent reviewers examined the titles and abstracts of all identified records for eligibility during the research selection procedure. The same reviewers then evaluated those that were thought to be possibly relevant in full. In order to obtain a final consensus, any differences that emerged during any step were settled by debate or consultation with a third reviewer.
Extracted data included demographic details (age, sex, body mass index (BMI)), clinical characteristics (renal function, CHA2DS2-VASc score, mRS score), comorbidities (diabetes, hypertension, prior stroke), treatment specifics (anticoagulant type and dosage), and clinical outcomes (stroke, embolism, bleeding events, mortality).
Quality assessment
Risk of bias assessment was conducted using the Cochrane Risk of Bias Tool (RoB 2), which evaluates bias across 5 critical domains [17]. Based on predetermined criteria, each study was categorized as having a low risk, considerable concerns, or high risk of bias. The evaluation was carried out independently by 2 reviewers, and differences were settled by consensus.
Statistical analysis
For meta-analysis, RevMan 5.3 was used. Mean differences (MD) and risk ratios (RR) with 95% confidence intervals (CI) were used to evaluate continuous and dichotomous outcomes, respectively. The I2 statistic was used to assess heterogeneity; values ≥75% indicated significant heterogeneity, supporting a random-effects model. A fixed-effects model was used for smaller heterogeneity (I2 < 75%). This method guaranteed a solid and trustworthy synthesis of the data.
Results
Search and screening results
The initial literature search identified 855 records from multiple databases, including 147 from PubMed, 293 from Google Scholar, 334 from Embase, and 81 from the Cochrane Library. Following the removal of duplicate records, 632 unique studies remained for initial screening. Based on a review of titles and abstracts, 619 of these were excluded. The remaining 13 full-text articles underwent a detailed eligibility assessment. After this full-text review, 7 studies were excluded due to outcomes and data that were not relevant to the specific objectives of this review, based on the predefined inclusion and exclusion criteria. As a result, 6 studies in all satisfied all requirements and were incorporated into the meta-analysis and qualitative synthesis (Fig. 1).
Fig. 1.
PRISMA chart.
This study comprises 6 randomized controlled trials (RCTs) that were published between 2009 and 2024.18–23. They were all included in the qualitative synthesis and then combined for a meta-analysis. Dabigatran, rivaroxaban, apixaban, and edoxaban were the direct oral anticoagulants (DOACs) assessed in these trials at different dosages. Conversely, in order to maintain a target international normalized ratio (INR) between 2.0 and 3.0, warfarin was dose-adjusted in all pertinent research arms (Table 1).
Table 1.
Characteristics of included studies.
| Author | Year | Country | Study Design |
|---|---|---|---|
| Hong et al. [18] | 2017 | South Korea | Randomized Controlled Trial |
| Connolly et al. [19] | 2009 | Canada | Randomized Controlled Trial |
| Patel et al. [20] | 2011 | United States | Randomized Controlled Trial |
| Joosten et al. [21] | 2024 | Netherlands | Randomized Controlled Trial |
| Granger et al. [22] | 2011 | United States | Randomized Controlled Trial |
| Giugliano et al. [23] | 2013 | United States | Randomized Controlled Trial |
Patient demographics
Across these studies, a total of over 60,000 participants were enrolled, with a median age ranged from 70 to 83 years, and most trials had a male majority, with some reporting a near-balanced sex distribution [19,20]. The included populations were consistently at moderate to high risk for stroke, supported by reported CHA2DS2-VASc scores and trial eligibility criteria.
The modified Rankin Scale (mRS) score at 4 weeks was reported in one study, with the majority of patients having scores of 0–1, indicating minimal functional disability: 84% in the DOAC group and 74.4% in the Warfarin group [18] This suggests the enrolled populations had good baseline functional status, a factor that may influence outcomes and tolerability.
Renal function reporting varied by study: serum creatinine was reported in Hong et al., 2017 (mean 0.89 mg/dL)[18] while others reported creatinine clearance or Estimated Glomerular Filtration Rate (eGFR) (e.g., Patel et al., 2011: median clearance 67 mL/min; Joosten et al., 2024: eGFR 62.5 mL/min) [20,21]. Granger et al. provided a granular breakdown of renal impairment [22]. BMI was reported in two trials and ranged from 27.4 to 28.3 kg/m2 [20,21].
Prior use of antiplatelet or anticoagulant therapy was reported in three of the included studies. In Hong et al., prior use was noted in 41.1% of the DOAC group and 40.9% of the Warfarin group [18]. Similarly, in Granger et al. prior VKA use was documented in 57.1% of the Apixaban (DOAC) group and 57.2% in the Warfarin group, while Giugliano et al. reported nearly identical prior VKA exposure in 58.8% of both the Edoxaban and Warfarin groups.22,23 This balance in prior exposure to antithrombotic agents helps to mitigate confounding and supports the comparability between the groups.
The concomitant use of antiplatelet therapy was a common practice in the included studies, with aspirin being the most frequently administered agent. Additional antiplatelet medications, such as clopidogrel or other thienopyridines, were also used. All trials employed standardized dosing regimens for the direct oral anticoagulants (DOACs). These included dabigatran (110 mg or 150 mg twice daily), rivaroxaban (20 mg once daily), apixaban (5 mg twice daily), and edoxaban (60 mg once daily or a reduced dose of 30 mg once daily) [18,19,20,21,22,23]. Follow-up duration ranged from 5 weeks in Hong et al., 2017 to 2.8 years in Giugliano et al., 2013 [18,23]. Intermediate durations included 344 days in Joosten et al., 2024, 1.8 years in Granger et al., and nearly 2 years in Connolly et al. and Patel et al [19,21,22]. These varying durations impacted the extent to which long-term outcomes, particularly bleeding, cardiovascular, and cerebrovascular events, could be captured and compared (Table 2).
Table 2.
Patient demographics and data.
| Author | Drug Class | Drug Name | Dose | Total Number Patients, N | Sex Male:Female, N | Age in Years | BMI | Renal Function | CHAD – VAS score | mRS score 0-1 at 4 Weeks, N | Prior Anti-PLT/ coagulant use, N | Concomitant Anti-PLT/ coagulant use, N | Follow up |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Hong et al., 2017 [18] | DOAC | Rivaroxaban | 10mg/day for 5 days then 20mg/day | 95 | 55:40 | Mean (SD): 70.2 (10.1) | NR | Serum Cr Mean: 0.89 mg/dl | Mean (SD): 2.7 (1.7) | 79 | 39 | Total use: 34 Aspirin: 25 Clopidogrel: 2 | 5 weeks |
| VKA | Warfarin | Target INR 2.0 – 3.0 | 88 | 52:36 | Mean (SD): 70.6 (10.9) | NR | Serum Cr Mean: 0.89 mg/dl | Mean (SD): 2.3 (1.4) | 64 | 36 | Total use: 23 Aspirin: 11 Clopidogrel: 7 | 5 weeks | |
| Connolly et al., 2009 [19] | DOAC | Dabigatran | 110mg | 6015 | 3865:2150 | Mean (SD): 71.4 (8.6) | NR | NR | Mean (SD): 2.1 (1.1) | NR | NR | Aspirin: 2404 VKA: 3011 | 2 years |
| 150mg | 6076 | 3840:2236 | Mean (SD): 71.5 (8.8) | Mean (SD): 2.2 (1.2) | Aspirin: 2352 VKA: 3049 | ||||||||
| VKA | Warfarin | Target INR 2.0 – 3.0 | 6022 | 3809:2213 | Mean (SD): 71.6 (8.6) | NR | NR | Mean (SD): 2.1 (1.1) | NR | NR | Aspirin: 2442 VKA: 2929 | 2 years | |
| Patel et al., 2011 [20] | DOAC | Rivaroxaban | 20mg | 7131 | 4300:2831 | Median: 73 | Median: 28.3 | Cr Clearance Median: 67 ml/min | Mean (SD): 3.48 (0.94) | NR | NR | Aspirin: 2586 VKA: 4443 | 1 year and 11 months |
| VKA | Warfarin | Target INR 2.0 – 3.0 | 7133 | 4301:2832 | Median: 73 | Median: 28.1 | Cr Clearance Median: 67 ml/min | Mean (SD): 3.46 (0.95) | NR | NR | Aspirin: 2619 VKA: 4461 | 1 year and 11 months | |
| Joosten et al., 2024 [21] | DOAC | Dabigatran | NR | 662 Dabigatran: 57 Rivaroxaban: 332 Apixaban: 115 Edoxaban: 109 | 388:274 | Mean (SD): 83 (5.1) | Mean (SD): 27.4 (6.0) | eGFR Mean (SD): 62.5 (015.8) | Median (IQR): 4.0 (3-5) | NR | NR | 16 | 344 days |
| Rivaroxaban | |||||||||||||
| Apixaban | |||||||||||||
| Edoxapan | |||||||||||||
| VKA | Acenocoumarol | 1mg | 661 | 422:239 | Mean (SD): 82.8 (5.1) | Mean (SD): 27.4 (11.7) | eGFR Mean (SD): 62.7 (15.6) | Median (IQR): 4.0 (3-5) | NR | NR | 13 | 344 days | |
| Phenoprocoumon | 3mg | ||||||||||||
| Granger et al., 2011 [22] | DOAC | Apixaban | 5mg | 9120 | 5886:3234 | Median: 70 | NR | Cr Clearance: N: Normal: 3761 Mild impairment: 3817 Moderate: 1365 Severe: 137 Not reported: 40 | Mean (SD): 2.1 (1.1) | NR | 5208 | Aspirin: 2859 Clopidogrel: 170 | 1.8 years |
| VKA | Warfarin | Target INR 2.0 – 3.0 | 9081 | 5899:3182 | Median: 70 | NR | Cr Clearance: N: Normal: 3757 Mild impairment: 3770 Moderate: 1382 Severe: 133 Not reported: 39 | Mean (SD): 2.1 (1.1) | NR | 5193 | Aspirin: 2773 Clopidogrel: 168 | 1.8 years | |
| Giugliano et al., 2013 [23] | DOAC | Edoxapan | 60mg | 7035 | 4366:2669 | Median: 72 | NR | NR | Mean (SD): 2.8 (1.0) | NR | 4140 | Aspirin: 2070 Thienopyridine: 174 | 2.8 years |
| 30mg | 7034 | 4304:2730 | Median: 72 | Mean (SD): 2.8 (1.0) | 4163 | Aspirin: 2018 Thienopyridine: 149 | |||||||
| VKA | Warfarin | Target INR 2.0 – 3.0 | 7036 | 4395:2641 | Median: 72 | NR | NR | Mean (SD): 2.8 (1.0) | NR | 4138 | Aspirin: 2092 Thienopyridine: 164 | 2.8 years |
NR: Not reported; N: Number; SD: Standard Deviation; IQR: Interquartile Range; DOAC: Direct Oral Anticoagulants; VKA: Vitamin K Antagonists; mg; Milligram; mg/dl: milligram per deciliter; INR: Internationl Normalized Ratio; BMI: Body mass index; Cr: Creatinine; eGFR: Estimated glomerular filtration rate; mRS: Modified Rankin Scale; PLT: Platelet.
Clinical outcomes
The DOACs were consistently associated with either superior or non-inferior stroke prevention compared to VKAs. Connolly et al. (2009) reported fewer strokes with dabigatran 150 mg (122 events) than with warfarin (185 events) [19]. Granger et al. (2011) found a similar reduction in stroke events with apixaban (401) compared to warfarin (503) [22]. Giugliano et al. (2013) showed fewer strokes with high-dose edoxaban (281 vs. 317 with warfarin), while low-dose edoxaban had slightly more events (360) [23]. Patel et al. (2011) observed a moderate reduction in stroke events with rivaroxaban (184) versus warfarin (221), although embolic events were unexpectedly lower in the warfarin group (22 vs. 5) [20] Joosten et al. (2024) reported a lower percentage of cerebrovascular events in the DOAC group (1.8%) compared to warfarin (2.9%) [21]. Hong et al. (2017) reported nearly equivalent stroke events in both arms (48 in the DOAC group vs. 49 in the warfarin group), with no embolic or cerebrovascular events noted [18].
Most studies demonstrated reduced major bleeding rates with DOACs. Granger et al. (2011) reported 327 major bleeding events with apixaban compared to 462 with warfarin [22]. Similarly, Giugliano et al. (2013) found fewer events with both low-dose (254) and high-dose edoxaban (418) than with warfarin (524) [23]. In Connolly et al. (2009), dabigatran 110 mg had fewer major bleeds (322) than warfarin (397), while the 150 mg dose had slightly more (375) [19]. Patel et al. (2011) reported similar rates between DOAC (395) and warfarin (386) [20]. Joosten et al. (2024) showed a slightly higher rate of major bleeding with DOACs (3.6%) compared to warfarin (2.4%) [21]. Hong et al. (2017) reported only one major bleeding event in the DOAC group (1.1%) and none in the warfarin group. Minor bleeding was not reported in this study [18].
The mortality outcomes generally favored DOACs. Connolly et al. (2009) found fewer deaths with dabigatran 150 mg (274) compared to warfarin (317) [19]. Granger et al. (2011) reported 603 deaths in the apixaban group versus 669 in the warfarin group [22]. Giugliano et al. (2013) showed a similar benefit with high-dose edoxaban (530 vs. 611 with warfarin) [23]. Patel et al. (2011) also demonstrated slightly lower mortality with rivaroxaban (170) versus warfarin (193) [20]. Joosten et al. (2024) reported nearly identical mortality rates between DOAC (44 deaths) and warfarin (46 deaths) [21]. Hong et al. (2017) reported no deaths in either group during the 5-week follow-up period[18] (Table 3).
Table 3.
Clinical outcome parameters.
| Author | Drug Name | Dose | Stroke Occurrence, N | Stroke Embolism, N | Cerebrovascular events, N | Minor Bleeding, N | Major Bleeding, N | Death/Mortality, N |
|---|---|---|---|---|---|---|---|---|
| Hong et al., 2017 [18] | Rivaroxaban | 10mg/day for 5 days then 20mg/day | 48 (50.5) | 0 | 0 | NR | 1 (1.1) | 0 |
| Warfarin | Target INR 2.0 – 3.0 | 49 (55.7) | 0 | 0 | NR | 0 | 0 | |
| Connolly et al., 2009 [19] | Dabigatran | 110mg | 171 | 11 | 86 | 1566 | 322 | 289 |
| 150mg | 122 | 12 | 89 | 1787 | 375 | 274 | ||
| Warfarin | Target INR 2.0 – 3.0 | 185 | 14 | 63 | 1931 | 397 | 317 | |
| Patel et al., 2011 [20] | Rivaroxaban | 20mg | 184 (2.61) | 5 (0.7) | 101 | NR | 395 (5.6) | 170 (2.4) |
| Warfarin | Target INR 2.0 – 3.0 | 221 (3.12) | 22 (0.3) | 126 | NR | 386 (5.4) | 193 (2.7) | |
| Joosten et al., 2024 [21] | Dabigatran | NR | 14 (2.1) | 16 (2.4) | 12 (1.8) | 84 (12.7) | 24 (3.6) | 44 (6.7) |
| Rivaroxaban | ||||||||
| Apixaban | ||||||||
| Edoxapan | ||||||||
| Acenocoumarol | 1mg | 11 (1.7) | 13 (2) | 19 (2.9) | 49 (7.4) | 16 (2.4) | 46 (7) | |
| Phenoprocoumon | 3mg | |||||||
| Granger et al., 2011 [22] | Apixaban | 5mg | 401 | 15 | Myocardial Infarction: 90 | 613 | 327 | 603 |
| Warfarin | Target INR 2.0 – 3.0 | 503 | 17 | Myocardial Infarction: 102 | 877 | 462 | 669 | |
| Giugliano et al., 2013 [23] | Edoxapan | 60mg | 281 | 15 | 133 | 604 | 418 | 530 |
| 30mg | 360 | 29 | 169 | 533 | 254 | 527 | ||
| Warfarin | Target INR 2.0 – 3.0 | 317 | 23 | 141 | 714 | 524 | 611 |
NR: Not reported; N: Number; DOAC: Direct Oral Anticoagulants; mg: Milligram.
Comorbidities
Comorbid conditions were commonly documented across all included studies, with hypertension and diabetes mellitus being the most prevalent. Heart failure, past myocardial infarction, and prior stroke were other disorders that were commonly reported. These conditions are consistent with the high cardiovascular risk profile that is characteristic of individuals with AF. Special populations, such as those with liver cirrhosis, chronic obstructive pulmonary disease (COPD), and active malignancy, were also included in several trials. Furthermore, several studies stratified participants by age, with a significant proportion being aged 75 years or older (Table 4).
Table 4.
Comorbidities as reported among the included studies.
| Author | Drug Name | Dose | Hypertension, N | Diabetes mellites, N | Hyperlipidemia, N | Myocardial infarction, N | Heart failure, N | Other coronary artery diseases, N | Peripheral Vascular, N | Transient ischemic attack, N | Prior stroke, N | Other cerebrovascular, N | Systemic Embolus, N | Others, N |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Hong et al., 2017 [18] | Rivaroxaban | 10mg/day for 5 days then 20mg/day | 65 (68.4) | 24 (25.3) | 16 (16.8) | 0 | 0 | 13 (13.7) | 0 | 0 | 0 | 0 | 0 | 0 |
| Warfarin | Target INR 2.0 – 3.0 | 52 (59.1) | 10 (11.4) | 18 (20.5) | 0 | 0 | 12 (13.6) | 0 | 0 | 0 | 0 | 0 | 0 | |
| Connolly et al., 2009 [19] | Dabigatran | 110mg | 4738 (78.8) | 1409 (23.4) | 0 | 0 | 1937 (32.2) | 1008 (16.8) | 0 | 0 | 0 | 0 | 0 | 0 |
| 150mg | 4795 (78.9) | 1402 (23.1) | 0 | 0 | 1934 (31.8) | 1029 (16.9) | 0 | 0 | 0 | 0 | 0 | 0 | ||
| Warfarin | Target INR 2.0 – 3.0 | 4750 (78.9) | 1410 (23.4) | 0 | 0 | 1922 (31.9) | 968 (16.1) | 0 | 0 | 0 | 0 | 0 | 0 | |
| Patel et al., 2011 [20] | Rivaroxaban | 20mg | 6436 (90.3) | 2878 (40.4) | 0 | 1182 (16.6) | 4467 (62.6) | 0 | 401 (5.6) | 0 | 0 | 3916 (54.9) | 0 | Chronic Obstructive Pulmonary Disease: 754 (10.6) |
| Warfarin | Target INR 2.0 – 3.0 | 6474 (90.8) | 2817 (39.5) | 0 | 1286 (18) | 4441 (62.3) | 0 | 438 (6.1) | 0 | 0 | 3895 (54.6) | 0 | Chronic Obstructive Pulmonary Disease: 743 (10.4) | |
| Joosten et al., 2024 [21] | Dabigatran | NR | 365 (55.1) | 140 (21.1) | 0 | 0 | 129 (19.5) | 0 | 0 | 0 | 0 | 0 | 139 (21) | Active cancer: 44 (6.6) Liver Cirrhosis: 3 (0.5) |
| Rivaroxaban | ||||||||||||||
| Apixaban | ||||||||||||||
| Edoxapan | ||||||||||||||
| Acenocoumarol | 1mg | 336 (50.8) | 140 (21.2) | 0 | 0 | 150 (22.7) | 0 | 0 | 0 | 0 | 0 | 117 (17.7) | Active cancer: 35 (5.3) Liver Cirrhosis: 5 (0.8) | |
| Phenopr- ocoumon | 3mg | |||||||||||||
| Granger et al., 2011 [22] | Apixaban | 5mg | 7962 (87.3) | 2284 (25) | 0 | 0 | 3235 (35.5) | 0 | 0 | 0 | 0 | 1748 (19.2) | 0 | 0 |
| Warfarin | Target INR 2.0 – 3.0 | 7954 (87.6) | 2263 (24.9) | 0 | 0 | 3216 (35.4) | 0 | 0 | 0 | 0 | 1790 (19.7) | 0 | 0 | |
| Giugliano et al., 2013 [23] | Edoxapan | 60mg | 6591 (93.7) | 2559 (36.4) | 0 | 0 | 4097 (58.2) | 0 | 0 | 0 | 0 | 1976 (28.1) | 0 | 0 |
| 30mg | 6575 (93.5) | 2544 (36.2) | 0 | 0 | 3979 (56.6) | 0 | 0 | 0 | 0 | 2006 (28.5) | 0 | 0 | ||
| Warfarin | Target INR 2.0 – 3.0 | 6588 (93.6) | 2521 (35.8) | 0 | 0 | 4048 (57.5) | 0 | 0 | 0 | 0 | 1991 (28.3) | 0 | 0 |
N: Number; DOAC: Direct Oral Anticoagulants; mg: Milligram.
Meta-analysis outcomes
Patients on high-dose direct oral anticoagulants (DOACs) had a statistically significant lower risk of stroke when compared to warfarin in randomized controlled studies. DOACs were favored by the pooled risk ratio of 0.81 (95% CI: 0.74 to 0.89; p<0.00001). It is evident that DOACs are beneficial in lowering cerebrovascular incidents because the confidence interval does not cross 1. Heterogeneity was low (I2=19%), reflecting consistent results across the included studies and supporting the robustness and generalizability of the findings (Fig. 2).
Fig. 2.
Forest plot of stroke occurrence in included studies (high dose groups).
However, high-dose DOACs were not associated with a statistically significant reduction in cardiovascular outcomes, including myocardial infarction and systemic embolism. The pooled risk estimate (RR = 0.96, 95% CI: 0.83 to 1.11; p = 0.58) was close to unity, and the confidence interval spanned 1, indicating no clear superiority over Warfarin. Heterogeneity was moderate (I2=65%), suggesting some variability across trials and warranting cautious interpretation of the findings (Fig. 3).
Fig. 3.
Forest plot of cardiovascular events in included studies (high dose groups).
The high-dose DOACs demonstrated a non-significant reduction in major bleeding risk compared to warfarin, with a pooled risk ratio of 0.89 (95% CI: 0.75–1.04; p=0.14). The lack of statistical significance suggests comparable bleeding risks between the two treatments. However, substantial heterogeneity was observed (I2=76%), possibly due to variations in patient demographics, trial methodologies, or differing definitions of major bleeding across studies (Fig. 4).
Fig. 4.
Forest plot of major bleeding in included studies (high dose groups).
Risk of bias
No significant difference was found in minor bleeding rates between high-dose DOACs and warfarin (RR: 0.92, 95% CI: 0.76–1.11; p=0.37). The CI crossing 1 indicates no clear advantage of either treatment. Notably, the high heterogeneity (I2 = 92%) underscores considerable variability among studies, which may stem from differences in bleeding assessment criteria or patient populations, thereby limiting definitive conclusions (Fig. 5).
Fig. 5.
Forest plot of minor bleeding in included studies (high dose groups).
Among the six trials, three were deemed to have a low risk of bias across all domains [20,22,23]. The remaining three studies raised minor concerns, primarily related to randomization or protocol deviations, but none were classified as having a high overall risk of bias [18,19,21]. These findings suggest that the included studies generally maintained acceptable methodological rigor, with half providing high-quality evidence and the others presenting only minor limitations (Table 5).
Table 5.
Risk of bias assessment of included studies.
| Randomization process | Deviations from intended intervention | Missing outcome data | Measurement of the outcome | Selection of the reported result | Overall Risk of Bias | |
|---|---|---|---|---|---|---|
| Hong et al., 2017 [21] | Some concerns | Low risk | Low risk | Low risk | Low risk | Some concerns |
| Connolly et al., 2009 [18] | Low risk | Some concerns | Low risk | Low risk | Low risk | Some concerns |
| Patel et al., 2011 [19] | Low risk | Low risk | Low risk | Low risk | Low risk | Low risk |
| Joosten et al., 2023 [22] | Low risk | Some concerns | Low risk | Low risk | Low risk | Some concerns |
| Granger et al., 2011 [23] | Low risk | Low risk | Low risk | Low risk | Low risk | Low risk |
| Giugliano et al, 2013 [20] | Low risk | Low risk | Low risk | Low risk | Low risk | Low risk |
Discussion
High-dose direct oral anticoagulants (DOACs) were linked to a 19% relative risk decrease in stroke or systemic embolism when compared to warfarin (RR ≈ 0.81, 95% CI 0.74–0.89; p<0.00001), with low heterogeneity (I2 = 19%), according to the results of the pooled meta-analysis. The efficacy of DOACs across trials is highlighted by this uniformity. The main factor contributing to the overall benefit was a notable decrease in hemorrhagic stroke. For example, in the ARISTOTLE trial, the incidence of hemorrhagic stroke was halved with apixaban versus warfarin (0.24% vs. 0.47%; HR 0.51) [22]. No significant heterogeneity in this benefit was observed across different patient subgroups.
While DOACs consistently demonstrated reductions in hemorrhagic stroke and intracranial bleeding, their major bleeding outcomes varied by specific agent. A meta-analysis of major bleeding revealed substantial heterogeneity (I2 ≈ 76%), reflecting the different safety profiles among DOACs. When compared to warfarin, apixaban significantly reduced major bleeding in the ARISTOTLE trial (2.13% vs. 3.09%; HR 0.69, 95% CI 0.60–0.80) [22]. Similarly, edoxaban decreased major bleeding by approximately 20% (2.75% vs. 3.43%; HR 0.80, 95% CI 0.71–0.91), and dabigatran 110 mg twice daily was associated with reduced bleeding in the RE-LY trial (2.71% vs. 3.36%; p=0.003) [19,23]. In contrast, warfarin and rivaroxaban demonstrated comparable rates of major bleeding (14.9% vs. 14.5%; HR 1.03, p=0.44) [20].
Even when total major bleeding risk was comparable, DOACs consistently showed fewer intracranial hemorrhages. For instance, rivaroxaban users had fewer intracranial bleeds despite similar overall bleeding, and both doses of dabigatran and edoxaban halve intracranial bleeding versus warfarin. This suggests DOACs bleeds are less catastrophic. Minor bleeding was common across all arms, but rates were highly heterogeneous (I2 > 90%) due to differing definitions and reporting practices.
Overall, DOACs show a safer bleeding profile, especially in reducing life-threatening hemorrhages with no increase in average major bleeding risk. This safety contributes to improved clinical outcomes.
The trials and pooled data suggest that the improvements in stroke prevention and bleeding translate into a modest but significant reduction in mortality with DOACs. Apixaban treatment was linked to a lower annual all-cause mortality rate in ARISTOTLE than warfarin (3.52% vs. 3.94%; HR 0.89, 95% CI 0.80–0.99) [22]. Additionally, there was a trend toward lower mortality with dabigatran 150 mg (3.64% vs. 4.13% annually, p=0.051) [19]. The edoxaban trial reported significantly lower rates of cardiovascular death with both high- and low-dose edoxaban relative to warfarin [23]. This survival benefit likely arises from fewer fatal strokes and intracranial hemorrhages, as well as overall improved safety.
The findings of our meta-analysis align closely with previous systematic reviews. Pan et al. demonstrated that DOACs offered superior stroke prevention and reduced bleeding risks in valvular AF and non-valvular AF [24]. These outcomes agree with our findings, which also support the benefit of high-dose DOACs in reducing cerebrovascular events. However, while Pan et al. observed significant reductions in major bleeding, our analysis did not find a statistically significant difference, likely due to variability across trials. Notably, both studies identified rivaroxaban as less favorable regarding safety outcomes, highlighting the importance of drug-specific considerations in clinical practice.
Similarly, DOACs significantly decreased stroke, major bleeding, cerebral hemorrhage, and all-cause mortality when compared to warfarin, according to a meta-analysis by Fong et al. on patients with AF at low to moderate stroke risk [25]. This suggests that the benefits of DOACs extend across different stroke risk categories in AF populations, not only in high-risk patients typically represented in clinical trials.
Hicks et al. [26] also reported that DOACs were associated with improved outcomes across several domains, including stroke prevention, intracranial bleeding, and all-cause mortality. Our analysis observed a consistent trend, although the reduction in major bleeding was not statistically conclusive. This may reflect differences in study design, patient populations, or the specific DOACs evaluated.
Our findings are also consistent with the large-scale meta-analysis by Ruff et al., which confirmed the overall superiority of DOACs over warfarin for stroke prevention, intracranial safety, and mortality outcomes [27]. In contrast, gastrointestinal bleeding remained a potential limitation with certain DOACs, particularly rivaroxaban.
Taken together, these studies reinforce the robustness of our findings and support the evolving consensus that DOACs—especially at appropriate doses—are a safe and effective alternative to warfarin for stroke prevention in AF.
While this meta-analysis provides valuable insights into the comparative safety and efficacy of high-dose DOACs versus warfarin, certain limitations should be considered. Some variability was observed across the included studies, particularly in bleeding outcomes. This may reflect differences in trial design, patient characteristics, and definitions of endpoints. Additionally, the analysis combined results from different DOACs agents and dosing strategies. Although all DOACs share a similar mechanism of action, subtle pharmacological differences could influence outcomes in specific patient groups. Despite these factors, the consistency of findings across large, high-quality trials supports the reliability of the results.
In conclusion, for patients with non-valvular AF, DOACs demonstrate better efficacy than warfarin in stroke and systemic embolism prevention. Meta-analyses revealed a 19% relative risk reduction for these outcomes with DOACs. Although bleeding profiles varied between different DOACs, their collective advantage over warfarin remains evident in comprehensive analyses. Findings support the preferential use of DOACs in clinical practice, especially given their improved safety and mortality profiles. Nevertheless, due to differences in bleeding results and variations among DOAC agents, the treatment decision should be personalized. Future studies should focus on the long-term safety, cost-effectiveness, and real-world outcomes of single DOACs, especially in diverse populations with different stroke and bleeding risk profiles. Personalized anticoagulation strategies remain essential.
Acknowledgment
We would like to thank Sofia Fields editing service (www.sofiafields.com) for English Language editing and plagiarism check.
Disclosure
Authors have no conflict of interests, and the work was not supported or funded by any drug company.
Contributor Information
Abeer A. Alomairi, Email: Abeeromairi5@gmail.com.
Rana Alduhaysh, Email: Rnaa.alduhaish@gmail.com.
Alaa Samandar, Email: Alaa.samandar@outlook.com.
Abeer Alshehri, Email: Abeer.alsh003@gmail.com.
Ghada AlAbdullah, Email: Ghada.pm128@gmail.com.
Saud Almutairi, Email: Sauiali2902@gmai.com.
Faisal Althekair, Email: thekairF@mngha.med.sa.
References
- [1].Chugh SS, Havmoeller R, Narayanan K, Singh D, Rienstra M, Benjamin EJ, et al. , Worldwide epidemiology of atrial fibrillation: a Global Burden of Disease 2010 Study. Circulation. 2014;129(8):837–847. 10.1161/CIRCULATIONAHA.113.005119 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [2].Son MK, Lim NK, Kim HW, Park HY. Risk of ischemic stroke after atrial fibrillation diagnosis: A national sample cohort. PLOS ONE. 2017;12(6):e0179687. 10.1371/journal.pone.0179687 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [3].Strilciuc S, Grad DA, Radu C, Chira D, Stan A, Ungureanu M, et al. , The economic burden of stroke: a systematic review of cost of illness studies. J Med Life. 2021;14(5):606–619. 10.25122/jml-2021-0361 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [4].Chaurasia RN, Sharma J, Pathak A, Mishra VN, Joshi D. Poststroke cognitive decline: a longitudinal study from a tertiary care center. J Neurosci Rural Pract. 2019;10(3):459–464. 10.1055/s-0039-1697872 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [5].Philippart R, Brunet-Bernard A, Clementy N, Bourguignon T, Mirza A, Babuty D, et al. , Prognostic value of CHA2DS2-VASc score in patients with non-valvular atrial fibrillation and valvular heart disease: the Loire Valley Atrial Fibrillation Project. Eur Heart J. 2015;36(28):1822–1830. 10.1093/eurheartj/ehv163 [DOI] [PubMed] [Google Scholar]
- [6].Mason PK, Lake DE, DiMarco JP, Ferguson JD, Mangrum JM, Bilchick K, et al. , Impact of the CHA2DS2-VASc score on anticoagulation recommendations for atrial fibrillation. Am J Med. 2012;125(6):603.e1–603.e6. 10.1016/j.amjmed.2011.09.030 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [7].Hindricks G, Potpara T, Dagres N, Arbelo E, Bax JJ, Blomström-Lundqvist C, et al. , 2020 ESC Guidelines for the diagnosis and management of atrial fibrillation developed in collaboration with the European Association for Cardio-Thoracic Surgery (EACTS). Eur Heart J. 2021;42(5):373–498. 10.1093/eurheartj/ehaa612 [DOI] [PubMed] [Google Scholar]
- [8].Bauer KA. Pros and cons of new oral anticoagulants. Hematology Am Soc Hematol Educ Program. 2013;2013(1):464–470. 10.1182/asheducation-2013.1.464 [DOI] [PubMed] [Google Scholar]
- [9].Gelosa P, Castiglioni L, Tenconi M, Baldessin L, Racagni G, Corsini A, et al. , Pharmacokinetic drug interactions of the non-vitamin K antagonist oral anticoagulants (NOACs). Pharmacol Res. 2018;135:60–79. 10.1016/j.phrs.2018.07.016 [DOI] [PubMed] [Google Scholar]
- [10].Kwok CS, Loke YK, Hale R, Potter JF, Myint PK. Atrial fibrillation and incidence of dementia: a systematic review and meta-analysis. Neurology. 2011;76(10):914–922. 10.1212/WNL.0b013e31820f2e38 [DOI] [PubMed] [Google Scholar]
- [11].Perera KS, Vanassche T, Bosch J, Swaminathan B, Mundl H, Giruparajah M, et al. , Global survey of the frequency of atrial fibrillation-associated stroke: Embolic Stroke of Undetermined Source Global Registry. Stroke. 2016;47(9):2197–2202. 10.1161/STROKEAHA.116.013378 [DOI] [PubMed] [Google Scholar]
- [12].Rohla M, Weiss TW, Pecen L, Patti G, Siller-Matula JM, Schnabel RB, et al. , Risk factors for thromboembolic and bleeding events in anticoagulated patients with atrial fibrillation: the prospective, multicentre observational PREvention oF thromboembolic events–European Registry in Atrial Fibrillation (PREFER in AF). BMJ Open. 2019;9(3):e022478. 10.1136/bmjopen-2018-022478 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [13].Schulman S, Angeras U, Bergqvist D, Eriksson B, Lassen MR, Fisher W, et al. , Definition of major bleeding in clinical investigations of antihemostatic medicinal products in surgical patients. J Thromb Haemost. 2010;8(1):202–204. 10.1111/j.1538-7836.2009.03678.x [DOI] [PubMed] [Google Scholar]
- [14].INPLASY [Internet]. INPLASY. [cited 2025. May 28]. Available from: https://inplasy.com/. [Google Scholar]
- [15].Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, et al. , The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. Syst Rev. 2021;10(1):89. 10.1186/s13643-021-01626-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [16].Rayyan. Rayyan: AI-powered systematic review management platform [Internet]. Rayyan.ai. [cited 2025. May 28]. Available from: https://www.rayyan.ai/. [Google Scholar]
- [17].RoB 2: A revised Cochrane risk-of-bias tool for randomized trials [Internet]. Cochrane.org. [cited 2025. May 28]. Available from: https://methods.cochrane.org/bias/resources/rob-2-revised-cochrane-risk-bias-tool-randomized-trials. [Google Scholar]
- [18].Hong K-S, Kwon SU, Lee SH, Lee JS, Kim Y-J, Song T-J, et al. , Rivaroxaban vs warfarin sodium in the ultra-early period after atrial fibrillation-related mild ischemic stroke: A randomized clinical trial. JAMA Neurol. 2017;74(10):1206–1215. 10.1001/jamaneurol.2017.2161 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [19].Connolly SJ, Ezekowitz MD, Yusuf S, Eikelboom J, Oldgren J, Parekh A, et al. , Dabigatran versus warfarin in patients with atrial fibrillation. N Engl J Med. 2009;361(12):1139–1151. 10.1056/NEJMoa0905561 [DOI] [PubMed] [Google Scholar]
- [20].Patel MR, Mahaffey KW, Garg J, Pan G, Singer DE, Hacke W, et al. , Rivaroxaban versus warfarin in nonvalvular atrial fibrillation. N Engl J Med. 2011;365(10):883–891. 10.1056/NEJMoa1009638 [DOI] [PubMed] [Google Scholar]
- [21].Joosten LPT, van Doorn S, van de Ven PM, Köhlen BTG, Nierman MC, Koek HL, et al. , Safety of switching from a vitamin K antagonist to a non-vitamin K antagonist oral anticoagulant in frail older patients with atrial fibrillation: Results of the FRAIL-AF randomized controlled trial. Circulation. 2024;149(4):279–289. [DOI] [PubMed] [Google Scholar]
- [22].Granger CB, Alexander JH, McMurray JJV, Lopes RD, Hylek EM, Hanna M, et al. , Apixaban versus warfarin in patients with atrial fibrillation. N Engl J Med 2011;365(11):981–992. 10.1056/NEJMoa1107039 [DOI] [PubMed] [Google Scholar]
- [23].Giugliano RP, Ruff CT, Braunwald E, Murphy SA, Wiviott SD, Halperin JL, et al. , Edoxaban versus warfarin in patients with atrial fibrillation. N Engl J Med. 2013;369(22):2093–2104. 10.1056/NEJMoa1310907 [DOI] [PubMed] [Google Scholar]
- [24].Pan KL, Singer DE, Ovbiagele B, Wu YL, Ahmed MA, Lee M, et al. , Effects of non-vitamin K antagonist oral anticoagulants versus warfarin in patients with atrial fibrillation and valvular heart disease: a systematic review and meta-analysis. J Am Heart Assoc. 2017;6(7):e005835. 10.1161/JAHA.117.005835 [DOI] [PMC free article] [PubMed] [Google Scholar]
- [25].Fong KY, de Silva HA, Milne N, Duffull SB, Goyal VK, Harrison J, et al. , Systematic review and meta-analysis of direct oral anticoagulants versus warfarin in atrial fibrillation with low stroke risk. Clin Appl Thromb Hemost. 2022;28:10760296221082879. [Google Scholar]
- [26].Hicks T, Stewart F, Eisinga A, Noorani A, Clough A, Goodacre S. NOACs versus warfarin for stroke prevention in patients with atrial fibrillation: a systematic review and meta-analysis. BMJ Open. 2016;6(11):e011115. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [27].Ruff CT, Giugliano RP, Braunwald E, Hoffman EB, Deenadayalu N, Ezekowitz MD, et al. , Comparison of the efficacy and safety of new oral anticoagulants with warfarin in patients with atrial fibrillation: a meta-analysis of randomised trials. Lancet. 2014;383(9921):955–962. 10.1016/S0140-6736(13)62343-0 [DOI] [PubMed] [Google Scholar]