Patients with atrial fibrillation (AF) who take direct oral anticoagulants (DOACs) face the risk of intracranial hemorrhage (ICH), which can be serious and even life threatening, but the risk of ICH of anticoagulants is still controversial. In this meta-analysis, we compared the risk of ICH between vitamin K antagonists (VKAs) and DOACs. Furthermore, we also compared the risk of ICH in different DOACs. PubMed, Embase, Web of Science, and the Cochrane Library were searched for relevant randomized controlled trials. The outcome was ICH, shown as the odds ratio (OR) with a 95% confidence interval (CI). DOACs were ranked by calculating the surface under the cumulative ranking curve (SUCRA). We included a total of 82,404 patients with AF. DOACs reduced the ICH risk by nearly half compared with VKAs (OR 0.47, 95% CI 0.40 to 0.54, p <0.001). VKAs were the least safe among all oral anticoagulants (SUCRA 1.7). Dabigatran 110 mg was the safest DOAC (SUCRA 87.3) for ICH risk, whereas rivaroxaban 20 mg was a relatively unsafe DOAC (SUCRA 27.5). Compared with rivaroxaban 20 mg, dabigatran 110 mg presented 53% (OR 0.47, 95% CI 0.27 to 0.82) lower relative risk for ICH. In conclusion, DOACs present less ICH risk than VKAs in patients with AF. For patients with AF who are at high risk of ICH, dabigatran 110 mg may be the safest choice among the DOACs.
Atrial fibrillation (AF) is the most common arrhythmia, which can lead to a series of complications, in which thromboembolism is the most serious, because it can lead to stroke and further lead to disability and even death. , Anticoagulant therapy is one of the important means of reducing the risk of thromboembolism in AF. For more than 60 years, warfarin has been the main oral anticoagulant drug for preventing stroke in patients with AF. However, warfarin needs frequent monitoring and dosage adjustment during the administration period and is greatly affected by food and drugs, which greatly limits its use. In recent years, direct oral anticoagulants (DOACs) have been developed and put into use and include dabigatran, apixaban, rivaroxaban, and edoxaban. DOACs have good safety and compliance, do not require routine monitoring, and provide a new option for clinical anticoagulant treatment of AF. Several large-scale clinical trials have shown that DOACs are as effective as warfarin in preventing stroke in patients with nonvalvular AF, but DOACs have the potential to increase bleeding, and the greatest safety hazard is intracranial hemorrhage (ICH). , This study aims to compare the risk of ICH with different oral anticoagulants in patients with AF.
This study was performed according to the recommendations of the Preferred Reporting Items of Systematic Reviews and Meta-Analyses statement.
We conducted a comprehensive literature search of the PubMed, Embase, Web of Science, and Cochrane Library databases. We searched for relevant research published on November 31, 2020 and earlier. The search keywords were as follows: (1) Dabigatran OR Rivaroxaban OR Apixaban OR Edoxaban OR Factor IIa inhibitors OR Factor Xa Inhibitors OR Direct thrombin inhibitors OR Non-vitamin K antagonist oral anticoagulants OR NOACs OR Direct oral anticoagulants OR DOACs OR Novel/New oral anticoagulants; (2) Atrial fibrillation OR non-valvular atrial fibrillation OR AF OR NVAF; and (3) Randomized controlled trial. The detailed search strategy for each database is shown in the Supplementary Material. Two researchers performed the literature search and screening independently. For comprehensiveness, we also examined the reference lists of the studies found in the database searches.
Studies were eligible for inclusion if they met the following criteria: (1) the study design was an randomized controlled trial; (2) the study population was patients with nonvalvular AF; (3) compared DOACs (dabigatran, rivaroxaban, apixaban, or edoxaban) and vitamin K antagonists (VKAs); (4) ICH data reported in 2 groups; and (5) for trials comparing various doses of DOACs with a control, we included only the regular-dose arms in our study.
The exclusion criteria were (1) DOACs in combination with antiplatelet drugs; (2) studies currently underway or with insufficient data; and (3) duplicate studies.
The data extracted from the selected articles included study information (author, publication year), study characteristics (study population, sample size, follow-up time), participant characteristics (age, gender), drug dose information, and outcome indicators (ICH). Two researchers extracted the data independently and cross-referenced them to avoid potential data extraction errors. Any disputes were discussed with a third researcher to reach an agreement.
Two independent researchers used the Cochrane Collaboration risk of bias tool to evaluate the quality of the selected literature. If a dispute arose during the evaluation, another researcher evaluated it to help resolve it.
All statistical analyses were performed using RevMan 5.3 (The Cochrane Collaboration, London, United Kingdom) and Stata 14.0 (Stata Corporation, College Station, Texas). Data pairing and preprocessing were performed using Stata. A network diagram was constructed by showing the sample size between every 2 interventions and the counts of direct comparison studies. In network meta-analysis (NMA), the inconsistent test is important because it estimates the difference between direct and indirect evidence. We used the inconsistency factor (IF) with a 95% confidence interval (CI) to evaluate the consistency of each closed loop. When the 95% CI of the IF value includes a zero, the consistency is considered good. The direct and indirect evidence are consistent, and then the consistency model is used for analysis; otherwise, the inconsistency model is used. In addition, we identified publication bias using funnel plots and Egger’s tests.
To evaluate the effect of DOACs on ICH risk, we used the odds ratio (OR) with 95% CI as the statistics of interest; p <0.05 was considered statistically significant. The heterogeneity in the included studies was analyzed by the Cochran’s Q test and I 2 statistic; if p ≥0.1 and I 2 ≤50%, a fixed-effect model was used for meta-analysis, whereas, conversely, a random-effect model was used for I 2 >50%. The results were ranked based on the surface under the cumulative ranking curve (SUCRA) to determine the safety of the DOACs for patients with AF. A larger SUCRA indicated lower ICH risk.
Using the previously mentioned search strategy, we retrieved a total of 8,837 studies. After screening, 19 RCTs were finally included in the present study and involved a total of 82,404 patients, 48,329 in the intervention group and 34,075 in the control group. , The literature screening process and results are shown in Figure 1 .
The baseline characteristics of the included studies are listed in Table 1 . Rivaroxaban 15 or 20 mg once a day, apixaban 5 mg twice a day, edoxaban 30 or 60 mg once a day, or dabigatran 110 or 150 mg twice a day were assessed for ICH risk presented in the patients. For patients with atrial fibrillation, no ICH data were reported for apixaban 2.5 mg twice a day. All control patients received VKAs.
| Study | Clinical conditions | DOACs | Control | Number | Age (mean or median) | Gender, female (%) | ICH (N) | Follow-up | ||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| DOACs | Control | DOACs | Control | DOACs | Control | DOACs | Control | |||||
| Ezekowitz et al, 2007 | AF | Dabigatran: 150 mg bid | VKAs: target INR of 2.0–3.0 | 100 | 70 | 70 (8.1) | 69 (8.3) | 18.7 | 15.7 | 0 | 0 | 2 weeks |
| Connolly et al, 2009 | AF | Dabigatran: 110 mg bid or 150 mg bid | VKAs: target INR of 2.0–3.0 | 12,091 | 6,022 | 110 mg bid: 71.4 (8.6) 150 mg bid: 71.5 (8.8) | 71.6 (8.6) | 36.3 | 36.7 | 63 | 87 | 2.0 years |
| Weitz et al, 2010 | AF | Edoxaban: 30 mg qd or 60 mg qd | VKAs: target INR of 2.0–3.0 | 469 | 250 | 30 mg qd: 65.2 (8.3) 60 mg qd: 64.9 (8.8) | 66.0 (8.5) | 37.4 | 39.6 | 0 | 0 | 12 weeks |
| Chung et al, 2011 | AF | Edoxaban: 30 mg qd or 60 mg qd | VKAs: target INR of 2.0–3.0 | 159 | 75 | 30 mg qd: 64.9 (9.1) 60 mg qd: 65.9 (7.7) | 64.5 (9.5) | 33.3 | 37.3 | 0 | 0 | 2 months |
| Granger et al, 2011 | AF | Apixaban: 5 mg bid | VKAs: target INR of 2.0–3.0 | 9,088 | 9,052 | 70 (63-76) | 70 (63-76) | 35.5 | 35.0 | 52 | 122 | 1.8 years |
| Ogawa et al, 2011 | AF | Apixaban: 5 mg bid | VKAs: INR of 2.0–3.0 (aged ≤70 years); INR of 2.0–2.6 (aged >70 years) | 71 | 75 | 70 | 71.7 | 17.6 | 18.9 | 0 | 1 | 16 weeks |
| Patel et al, 2011 | AF | Rivaroxaban 20 mg qd (15 mg qd under special clinical conditions) | VKAs: target INR of 2.0–3.0 | 7,111 | 7,125 | 73 (65–68) | 73 (65–68) | 39.7 | 39.7 | 55 | 84 | 590 days |
| Hori et al, 2012 | AF | Rivaroxaban 15 mg qd (10 mg qd under special clinical conditions) | VKAs: INR of 2.0–3.0 (aged <70 years); INR of 1.6–2.6 (aged ≥70 years) | 639 | 639 | 71.0 (34–89) | 71.2 (43–90) | 17.1 | 21.8 | 5 | 10 | 30 months |
| Yamashita et al, 2012 | AF | Edoxaban: 30 mg qd or 60 mg qd | VKAs: INR of 2.0–3.0 (aged <70 years); INR of 1.6–2.6 (aged ≥70 years) | 260 | 125 | 30 mg qd: 69.4 60 mg qd: 68.4 | 68.8 | 17.8 | 17.1 | 1 | 0 | 20 weeks |
| Giugliano et al, 2013 | AF | Edoxaban: 30 mg qd or 60 mg qd | VKAs: target INR of 2.0–3.0 | 14,014 | 7,012 | 30 mg qd: 72 (64-78) 60 mg qd: 72 (64-78) | 72 (64-78) | 38.4 | 37.5 | 102 | 132 | 2.8 years |
| Cappato et al, 2014 | AF and cardioversion | Rivaroxaban: 20 mg qd (15 mg qd under special clinical conditions) | VKAs: target INR of 2.0–3.0 | 988 | 499 | 64.9 (10.6) | 64.7 (10.5) | 27.4 | 26.9 | 2 | 1 | 30 days |
| Mao et al, 2014 | AF | Rivaroxaban: 20 mg qd (15 mg qd under special clinical conditions) | VKAs: target INR of 2.0–3.0 | 177 | 176 | 75 (68–79) | 75 (68–79) | 39.0 | 37.5 | 1 | 3 | 1.5 years |
| Goette et al, 2016 | AF and cardioversion | Edoxaban 60 mg qd (30 mg qd under special clinical conditions) | VKAs: target INR of 2.0–3.0 | 1,067 | 1,082 | 64.3 (10.3) | 64.2 (10.8) | 34.0 | 35.0 | 0 | 0 | 58 days |
| Kuwahara et al, 2016 | AF and ablation | Apixaban 5 mg bid (2.5 mg bid under special clinical conditions) | VKAs: target INR of 2.0–3.0 | 100 | 100 | 65 (9) | 66 (8) | 25.0 | 28.0 | 0 | 0 | 1-3 months |
| Calkins et al, 2017 | AF and ablation | Dabigatran: 150 mg bid | VKAs: target INR of 2.0–3.0 | 317 | 318 | 59.1 (10.4) | 59.3 (10.3) | 27.4 | 23.0 | 0 | 2 | 56 days |
| Ezekowitz et al, 2018 | AF and cardioversion | Apixaban 5 mg bid (2.5 mg bid under special clinical conditions) | VKAs: target INR of 2.0–3.0 | 735 | 721 | 64.7 (12.2) | 64.5 (12.8) | 32.9 | 33.5 | 0 | 1 | 30 days |
| Kirchhof et al, 2018 | AF and ablation | Apixaban: 5 mg bid (2.5 mg bid under special clinical conditions) | VKAs: target INR of 2.0–3.0 | 318 | 315 | 64 (57-70) | 64 (58-70) | 31.0 | 35.0 | 0 | 1 | 3 months |
| Hohnloser et al, 2019 | AF and ablation | Edoxaban: 60 mg qd (30 mg qd under special clinical conditions) | VKAs: target INR of 2.0–3.0 | 405 | 197 | 60.0 (53–67) | 61.0 (52–67) | 29.4 | 26.6 | 1 | 0 | 90 days |
| Nogami et al, 2019 | AF and ablation | Dabigatran 150 mg bid (110 mg bid under special clinical conditions) | VKAs: INR of 2.0–3.0 (aged <70 years); INR of 1.6–2.6 (aged ≥70 years) | 220 | 222 | 65.0 (59.0-71.0) | 66.0 (59.0-71.0) | 22.3 | 27.9 | 0 | 0 | 12 months |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
