Abstract
Background and objective
Patients with atrial fibrillation (AF) are prescribed oral anticoagulants for stroke prevention; however, no evidence indicates that the use of direct oral anticoagulants (DOACs) in the first few days after ischemic stroke (IS) would result in favorable outcomes. This study evaluated the association between the timing of using DOACs after IS and their effectiveness and safety to determine the optimal timing.
Methods
In this retrospective cohort study, we reviewed the electronic medical records of Taipei Veterans General Hospital. The 1-year outcomes of patients after DOAC initiation were evaluated. Different initiation time windows were compared (initiation time ≤3 days and >3 days in primary analysis). The primary composite outcome was stroke, transient ischemic attack, systemic embolism, or death due to IS. The primary safety outcome was major bleeding or clinically relevant nonmajor bleeding. The secondary composite outcome was all-cause mortality, thromboembolic event, or acute myocardial infarction/hemorrhagic events.
Results
This study included 570 patients. The median initiation time of DOACs after IS in the patients with AF was 14 days. Compared the patients in whom DOACs were initiated after >3 days with those DOACs were initiated after ≤3 days, the adjusted hazard ratios (aHRs) of the primary composite outcome was 0.73 (95% confidence interval [CI]: 0.23–1.79), the aHR of primary safety outcome was 0.87 (95% CI: 0.34–1.90), and the aHR of secondary composite outcome was 0.65 (95% CI: 0.32–1.19). All the results were not statistically significant. In secondary analysis, we tested multiple time points of initiating DOACs. Compared with DOAC initiation after >14 days, the primary composite outcomes in the patients in whom DOACs were initiated ≤3, 4–7, and 8–14 days after IS were the same as the findings of the main analysis. After separating patients into different stroke severity groups, the results were similar to those in the main analysis.
Conclusion
No significant association was observed between the timing of using DOACs and ischemic or hemorrhagic outcomes. The findings did not differ among different time points. Although we do not recommend avoiding the initiation of DOACs in the first few days after IS, we should consider that the early initiation of DOACs (≤3 days) would be appropriate only for patients who tend to experience thromboembolic events and have a low risk of bleeding. The optimal timing of initiation still must be confirmed by randomized controlled trials.
1. Introduction
In patients with atrial fibrillation (AF), secondary stroke prevention by using oral anticoagulants (OAC) is necessary; however, the optimal timing of using OACs after ischemic stroke (IS) remains controversial. Moreover, no sufficient evidence is available regarding the optimal timing of using OACs in the East Asian population. Although the initiation of OACs few days after IS can reduce its early recurrence, OACs can increase the risk of intracranial hemorrhage (ICH) including that of hemorrhagic transformation (HT) [[1], [2], [3]]. Thus, the risks of thromboembolism and hemorrhage resulting from the use of OACs for secondary stroke prevention can lead to uncertainties in determining the optimal timing for initiating OACs. In addition, physicians might encounter various clinical challenges while treating Asian patients. Studies have reported that compared with Caucasian patients, Asian patients with AF who were prescribed with either warfarin or direct OACs (DOACs) more frequently developed hemorrhagic events [4,5]. DOACs are a newer class of OACs that cause fewer hemorrhagic events than vitamin K antagonists (VKAs) do, thus being preferred by physicians. However, randomized controlled trials on DOACs that excluded patients with acute IS have observed an administration latency when DOACs were administered between 7 and 30 days before the enrolment of patients into the trials [[6], [7], [8], [9]]. Although these time gaps may help physicians understand therapeutic efficacy, they may also restrict them from making clinical decisions. A pooled analysis reported that the risk of ICH was lower when using DOACs (hazard ratio [HR]: 0.42; 95% confidence interval [CI]: 0.24–0.71) than warfarin for secondary prevention without the limitation of the timing of initiation [10]. The timing of administering OACs after IS has been variable in previous observational studies. The findings of these studies cannot be generalized to other populations because the studied drugs were mixed with different types of OACs (VKAs and DOACs) that have different pharmacokinetic properties. In addition, DOAC initiation time points and infarction size varied among different studies [[11], [12], [13], [14]]. Although current international guidelines recommend starting OACs between 4 and 14 days after IS or following the 1-3-6-12 day rule stated by expert consensus, universal guidelines regarding the use of OACs still do not exist [[15], [16], [17]]. The present study investigated the association between the different time points of using DOACs after IS and their effectiveness and safety to determine the optimal timing in East Asian.
2. Methods
2.1. Data source
The electronic medical records of Taipei Veterans General Hospital (TVGH) were used as the data source in this retrospective cohort study. TVGH contains approximately 2800 beds and provides approximately 10,000 outpatient prescriptions per day. The TVGH database includes information regarding patients’ diagnoses, medications, laboratory findings, smoking and drinking status, and imaging reports for the outpatient, inpatient, and emergency departments. In addition, the TVGH database contains a stroke registration system in which the details of patients with any type of stroke are included. All patients were followed up for at least 1 year after stroke, and their hospitalization or discharge follow-up records are available including information regarding the timing of the initiation of OACs and major adverse events occurring in other hospitals. All follow-ups were conducted by professional physicians through telephonic or face-to-face interviews at fixed time intervals after the registered stroke: 1, 3, 6, and 12 months. In addition, the interviews were conducted at each return visit in TVGH. This study was approved by the Institutional Review Board of Taipei Veterans General Hospital (IRB number: 2020-10-008BC), and the requirement of informed consent was waived by the review board due to data collection occurred as part of routine quality of care assessment.
2.2. Patients
The present study included patients with AF who experienced IS and were admitted in TVGH for management between September 1, 2012, and March 31, 2019. All the included patients were aged ≥20 years and considered to be suitable for receiving DOACs after their IS by their treating physician.
IS was defined based on the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM codes 433.x1, 434.xx, and 436) or International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM codes I63, and I64) codes at hospital discharge with validity in Taiwan's claims database [18,19]. AF could be either diagnosed previously or newly diagnosed during stroke hospitalization. To ensure the diagnostic accuracy, we considered AF only when it was diagnosed at least two times in the outpatient department or once during hospital discharge based on ICD-9-CM code 427.31 or ICD-10-CM codes I48.0, I48.1, I48.2, and I48.9 [20,21]; besides, we reviewed medical charts to ensure each diagnosis of AF was also confirmed by 12-lead electrocardiography (ECG) and at least once 24-h Holter monitoring ECG during index hospitalization or previous medical imaging examination. DOACs administered by physicians were dabigatran, rivaroxaban, apixaban, and edoxaban (Anatomical Therapeutic Chemical Index 2021: B01AE07, B01AF01, B01AF02, and B01AF03, respectively), and all the studied DOACs must be used at least 2 days and within 90 days after IS to minimize the effect of confounding by indication of the studied population, i.e. restricting the studied population in the similar disease stage and similar clinical consideration for DOAC prescription to prevent the effects of severity [22]. Patients who had missing National Institute of Health Stroke Scale (NIHSS) scores on admission, unknown initiation time, used VKAs first after IS, and received DOACs >90 days after IS were excluded.
In line with the management guideline in TVGH, all patients underwent computed tomography (CT) or magnetic resonance imaging (MRI) on admission for the evaluation of brain infarction size and HT by neurologists. The repeated cerebral images during hospitalization were obtained within 24 h after reperfusion therapy, before initiating DOACs or when patients exhibited a new onset of a neurological deficit [16]. Brain infarction size and HT were defined as described previously. Small: lesion in the anterior or posterior circulation <1.5 cm; large: lesion involving complete territory of middle cerebral artery (MCA), posterior cerebral artery (PCA) or anterior cerebral artery (ACA), or lesion involving 2 cortical superficial branches of MCA, or lesion involving a cortical superficial branch of MCA associated to the MCA deep branch, or lesion involving more than 1 artery territory, or lesion in brain stem or cerebellum >1.5 cm; medium: others [23].
2.3. Exposure measurement
The primary and secondary analyses were conducted based on the different time points of using DOACs and by comparing groups receiving DOAC at different time points. In the primary analysis, patients were divided into two groups based on the timing of the initiation of DOACs after IS; according to the current guideline, the third day was considered as the earliest cutoff point (early: initiation time ≤3 days; late: initiation time >3 days) [24]. In the secondary analysis, patients were divided into four groups based on the multiple cutoff points of the initiation timing (≤3, 4–7, 8–14, and >14 days).
2.4. Study outcomes and follow-up periods
Study outcomes were evaluated by reviewing electronic medical records at each hospitalization, neurology clinic visit records, and TVGH stroke registration records and by verifying ICD-9-CM or ICD-10-CM codes (Table SI in appendix). The primary composite outcomes included stroke, transient ischemic attack (TIA), systemic embolism (SE) or death from IS, whichever occurred earlier. The primary safety outcomes included major bleeding or clinically relevant nonmajor bleeding (CRNMB) according to the definition provided by the International Society on Thrombosis and Hemostasis (ISTH), whichever occurred earlier [25,26]. In our study, intracranial major bleeding included any nontraumatic ICH and new HT after the initiation of DOACs. The secondary composite outcomes included all-cause mortality, thromboembolic event, acute myocardial infarction, or hemorrhagic event, whichever occurred first. The clinical definition and validity of outcomes were described previously [12,27,28].
The cohort entry date was defined as the DOAC initiation date, and the closest IS before DOAC initiation was considered the index stroke. Patients were followed up from the cohort entry date until the occurrence of any outcome, DOAC discontinuation, change to VKAs or low-molecular-weight heparin (LMWH), loss to follow-up or the end of the study (1 year after DOAC initiation). The permissible gap of DOAC discontinuation was defined as 30 days; thus, the discontinuation date was the last supply day of the last prescription plus 30 days [29]. To account for differences in follow-up periods, we calculated the number of patient-years (number of patients multiplied by follow-up time in years).
2.5. Covariates
The baseline period was defined as 12 months before the cohort entry date. We collected baseline characteristics by reviewing electronic medical records and TVGH stroke registration records during hospitalization for the index stroke or by searching ICD-9-CM or ICD-10-CM codes. Baseline characteristics were sex, age, creatinine clearance (CrCl, min/mL) calculated using the Cockcroft–Gault formula at the DOAC initiation date, NIHSS score on admission, brain infarction size (small, medium, and large) on admission, HT before the DOAC initiation date, CHA2DS2-VASC score after the index stroke [30], HAS-BLED score after the index stroke [31], modified Rankin scale (mRS) score on admission, type of reperfusion therapy received (thrombolytic agents or mechanical thrombectomy), medical history (hypertension, type II diabetes mellitus, dyslipidemia, cancer, and prior stroke), current status of smoking and drinking, statin use after the index stroke, antithrombotic agent use before the index stroke, use of antiplatelet agents and LMWH before the DOAC initiation, and National Health Insurance (NIH) payment status of DOACs in Taiwan (Fig. 1). Medical history records were mainly obtained from admission records in the stroke registry for all included patients. If included patients regularly visited our hospital, we examined their medical history based on ICD codes that were confirmed at least two times in the outpatient department or once during hospital discharge 12 months before DOAC initiation [32].
Fig. 1.
Summary of study design.
2.6. Statistical analyses
Continuous and categorical variables are expressed as the median (interquartile range [IQR]) and number (percentage), respectively. We compared the baseline characteristics among different groups by using the Mann–Whitney U test or Kruskal–Wallis test for continuous variables and Fisher's exact test for categorical variables. We calculated the annualized rate of outcome events (total number of events/patient-years of follow-up). The cumulative incidence probability was estimated by plotting the Kaplan–Meier curve by using the cumulative distribution function and compared among the groups by using the log-rank test. We constructed a Cox proportional hazard model by using Firth's penalized likelihood approach and the profile-likelihood CI to compute the hazard ratio (HR) and reduce the effect of biased parameter estimates considering the small number of events [33]. For the multivariable model, we computed the adjusted HR (aHR) after adjusting prespecified confounders, namely sex, age, NIHSS score on admission, brain infarction size (small, medium and large) on admission, CHA2DS2-VASC score after the index stroke, HAS-BLED score after the index stroke, mRS score on admission, and reperfusion therapy. We assessed the proportional hazards assumption by evaluating the global P value obtained using the scaled Schoenfeld residuals method. A two-sided P value of <0.05 was considered to be statistically significant. If comparisons were involved in multiple testing, Bonferroni correction was applied. Statistical analyses were performed using SAS software, version 9.4 (SAS Institute Inc., Cary, NC, USA) or R 4.0.3 (R Foundation for Statistical Computing, Vienna, Austria).
Subgroup analyses were performed to explore the effect modification of the following prespecified groups: (1) renal function (CrCl <30 or ≥ 30 mL/min); (2) prior stroke history (number ≥1); (3) infarction size (large, small, and medium); (4) severe IS (NIHSS score <16 or ≥ 16 on admission); (5) NIH payment status (insurance payment or not); and (6) recommended dose based on the DOAC package insert in Taiwan (matching recommended dose or not). Each prespecified variable for subgroup analysis was included in a separate model to test the interaction effect, and we added the covariates of the main analysis into each separate model for adjustment. We converted the NIHSS score on admission from a continuous variable to a dichotomized variable in the “severe stroke” subgroup to avoid repeated adjustment.
2.7. Sensitivity analyses
Sensitivity analysis included two parts. First, we tested a single time point of DOAC initiation, namely 7th and 14th day after the index stroke. Second, we divided patients into the mild stroke group (NIHSS score <8) and moderate-to-severe stroke group (NIHSS score ≥8). Subsequently, we separated each group into early and late initiation groups based on their own earliest timing from the present evidence [15]. Earliest time point of 3rd and 7th day after the index stroke were considered for the mild and moderate-to-severe stroke groups, respectively. In addition, we compared outcomes among different initiation time points in the mild and moderate-to-severe stroke groups.
3. Results
3.1. Baseline characteristics
A total of 570 patients were included in this study (Fig. 2). The median age of the patients was 80 (IQR: 72–87) years, and 313 (54.9%) of the patients were men. The stroke severity of all the patients was mild to moderate considering that their median NIHSS score on admission was 7 (IQR: 3–17). A total of 230 (40.4%), 159 (27.9%), and 181 (31.8%) patients had small, medium, and large infarctions, respectively. The median scores of CHA2DS2-VASC (6; IQR: 5–6) and HAS-BLED (4; IQR: 3–4) after the index stroke indicated that all the patients had high risks of thromboembolism and bleeding events. Overall, 326 patients, 57% of the patients, received DOACs within 14 days from their IS, and the median period between the index stroke and DOAC initiation was 14 (IQR: 7–18) days. DOACs prescribed after IS were dabigatran (n = 152, 26.7%), rivaroxaban (n = 213, 37.4%), apixaban (n = 159, 27.9%), and edoxaban (n = 46, 8.1%; Table 1).
Fig. 2.
Study flowchart.
Table 1.
Baseline characteristics of study population.
| Variable | Total n = 570 | Initiated ≤3 days n = 70 | Initiated >3 days n = 500 | P value† |
|---|---|---|---|---|
| Age (y) | 80 (72–87) | 78 (70–86) | 80 (72–88) | 0.33 |
| Male | 313 (54.9) | 41 (58.6) | 272 (54.4) | 0.52 |
| Creatinine clearance (ml/min) | 48.8 (36.7–63.1) | 48.1 (31.2–59.4) | 48.8 (37.6–63.3) | 0.21 |
| NIHSS on admission | 7 (3–17) | 3 (1–8) | 8 (3–17.5) | <0.01 |
| Infarction size | ||||
| Large | 181 (31.8) | 5 (7.1) | 176 (35.2) | <0.01 |
| Medium | 159 (27.9) | 16 (22.9) | 143 (28.6) | 0.39 |
| Small | 230 (40.4) | 49 (70.0) | 181 (36.2) | <0.01 |
| Hemorrhagic transformation | 103 (18.1) | 2 (2.9) | 101 (20.2) | <0.01 |
| CHA2DS2VASC | 6 (5–6) | 6 (5–6) | 6 (5–6) | 0.94 |
| HAS-BLED | 4 (3–4) | 4 (3–4) | 4 (4–4) | <0.01 |
| mRS on admission | 4 (4–5) | 4 (3–4) | 4 (4–5) | <0.01 |
| Thrombolytic agents | 73 (12.8) | 12 (17.1) | 61 (12.2) | 0.25 |
| Mechanical thrombectomy | 52 (9.1) | 4 (5.7) | 48 (9.6) | 0.38 |
| Stroke history on admission | 113 (19.8) | 20 (28.6) | 93 (18.6) | 0.06 |
| ≥2 times | 6 (1.1) | 1 (1.4) | 7 (1.4) | 1.00 |
| ICH | 8 (1.4) | 0 (0) | 8 (1.6) | 0.60 |
| Current smoker | 52 (9.1) | 5 (7.1) | 47 (9.4) | 0.66 |
| Current alcohol drinker | 66 (11.6) | 9 (12.9) | 57 (11.4) | 0.69 |
| New onset AF | 259 (45.4) | 13 (18.6) | 246 (49.2) | <0.01 |
| Hypertension | 441 (77.4) | 57 (81.4) | 384 (76.8) | 0.45 |
| Diabetes mellitus | 157 (27.5) | 13 (18.6) | 144 (28.8) | 0.09 |
| Dyslipidemia | 207 (36.3) | 28 (40.0) | 179 (35.8) | 0.51 |
| Cancer | 85 (14.9) | 7 (10.0) | 78 (15.6) | 0.28 |
| Statin | 329 (57.7) | 38 (54.3) | 291 (58.2) | 0.61 |
| Antithrombotic agents before index stroke | ||||
| Yes | 273 (47.9) | 50 (71.4) | 223 (44.6) | <0.01 |
| Antiplatelet agents | 128 (22.5) | 12 (17.14) | 116 (23.2) | |
| Dual antiplatelet agents | 11 (1.93) | 1 (1.4) | 10 (2.0) | |
| Oral anticoagulants (OAC) | 115 (20.2) | 30 (42.9) | 85 (17.0) | |
| Antiplatelet + OAC | 19 (3.3) | 7 (10.0) | 12 (2.4) | |
| None | 297 (52.1) | 20 (28.6) | 277 (55.4) | |
| Antithrombotic agents before DOAC | ||||
| Antiplatelet agents | 517 (90.7) | 49 (70.0) | 468 (93.6) | <0.01 |
| LMWH | 36 (6.3) | 4 (5.7) | 32 (6.4) | 1.00 |
| DOAC after index stroke | ||||
| Dabigatran | 152 (26.7) | 23 (32.9) | 129 (25.8) | 0.25 |
| Rivaroxaban | 213 (37.4) | 23 (32.9) | 190 (38.0) | 0.43 |
| Apixaban | 159 (27.9) | 21 (30.0) | 138 (27.6) | 0.67 |
| Edoxaban | 46 (8.1) | 3 (4.3) | 43 (8.6) | 0.35 |
| Time between index stroke and DOAC use (days) | 14 (7–18) | 2 (1–3) | 14 (9–19) | <0.01 |
| NHI payment status | 482 (84.6) | 51 (72.9) | 431 (86.2) | <0.01 |
Categorical variables were showed as n (%) and continuous variables were showed as median (IQR).
†Significant level at P < 0.05.
ICH: intracerebral hemorrhage; LMWH: Low molecular weight heparin; mRS: modified Rankin Scale; NHI: national health insurance; NIHSS: National Institutes of Health Stroke Scale; CHA2DS2VASC: chronic heart failure, hypertension, age, diabetes mellitus, ischemic stroke/TIA/thromboembolic events, vascular disease, sex; HAS-BLED: hypertension, abnormal liver and renal function, stroke, bleeding, age, drug and alcohol.
A total of 70 and 500 patients were respectively included in the early and late initiation groups (DOAC initiation time of ≤3 and >3 days, respectively). Compared with the late initiation group (>3 days), the early initiation group (≤3 days) had lower median NIHSS scores on admission (3 vs. 8), higher proportions of small infarction size (70% vs. 36.2%), lower frequencies of HT before DOAC initiation (2.9% vs. 20.2%), fewer new diagnoses of AF (18.6% vs. 49.2%), and more patients using antithrombotic agents prior to the index stroke (30 patients, 42.9% vs. 85 patients, 17.0%). The median period between the index stroke and DOAC initiation date was 2 (IQR: 1–3) days in the early initiation group and 14 (IQR: 9–19) days in the late initiation group (Table 1).
3.2. Results of main analysis
In terms of the primary composite outcomes (stroke/SE/TIA/death due to stroke), the total follow-up time was 471 patient-years, and 54 composite events (annualized rate: 11.46%/patient-year), 20 recurrent ISs (4.25%/patient-year), and 22 ISs or SEs (4.67%/patient-year) were noted. In terms of the primary safety outcomes (major bleeding/CRNMB), the total follow-up time was 452 patient-years, and 69 composite events (15.25%/patient-year) and 5 intracranial hemorrhages (1.11%/patient-year) were observed. In terms of the secondary composite outcomes, 140 events were noted, with a total follow-up period of 436 patient-years (32.08%/patient-year).
The annualized rates of the primary composite outcome (stroke/SE/TIA/death due to stroke) and primary safety outcome (major bleeding/CRNMB) for initiation time ≤3 days were 6.4% and 9.66% per person-year; for >3 days were 12.22% and 16.07% per person-year, respectively. When we compared outcomes between the early and late initiation groups, unadjusted Kaplan–Meier curves indicated no significant difference (P = 0.22 and P = 0.31, respectively) except for the secondary composite outcome (P = 0.04; Fig. 3A–C). In univariable analysis, the risk of the secondary composite outcome decreased by approximately 50% in the early initiation group (HR: 0.54, 95% CI: 0.27–0.96; P = 0.053). However, the primary or secondary outcomes did not significantly differ between the early and late initiation groups in multivariable adjusted models (Table 2). No interactions were observed between the prespecified subgroups and the primary composite outcome or primary safety outcome (Fig. 4A and B).
Fig. 3.
Kaplan-Meier curves for primary composite outcome (A), primary safety outcome (B), secondary composite outcome (C). AMI: acute myocardial infarction, CRNMB: clinically relevant non-major bleeding, SE: systemic embolism, TIA: transient ischemic attack.
Table 2.
Primary and secondary outcomes at initiation cut point equal to 3 days.
| Events | ≤3 days |
>3 days |
≤3 days |
>3 days |
|---|---|---|---|---|
| Crude HR (95% CI)# | Adjusted HR (95% CI)# | |||
| Primary composite outcome† | 0.595 (0.19–1.40) | Reference | 0.73 (0.23–1.79) | Reference |
| Primary safety outcome‡ | 0.697 (0.28–1.45) | Reference | 0.87 (0.34–1.90) | Reference |
| Secondary composite outcome* | 0.54 (0.27–0.96)♦ | Reference | 0.65 (0.32–1.19) | Reference |
HR: hazard ratio, SE: systemic embolism, TIA: transient ischemic attack.
†Stroke/SE/TIA/death from ischemic stroke.
‡Major bleeding/clinically relevant non-major bleeding.
∗All-cause mortality/ischemic events/acute myocardial infarction/hemorrhagic events.
♦P = 0.053.
#All hazard ratios were introduced by Firth correction. Adjusted model included age, sex, NIHSS on admission, CHA2DS2VASC on admission, HAS-BLED on admission, mRS on admission, reperfusion therapy and infarction size.
Fig. 4.
Subgroup analyses of primary composite outcome (A) and primary safety outcome (B). aHR: adjusted hazard ratio, CRNMB: clinically relevant non-major bleeding, CrCl: creatinine clearance, SE: systemic embolism, TIA: transient ischemic attack. All hazard ratios were introduced by Firth correction. Adjusted model in each subgroup included age, sex, NIHSS on admission, CHA2DS2VASC on admission, HAS-BLED on admission, mRS on admission, reperfusion therapy, infarction size and subgroup variable. Severe stroke subgroup replaced NIHSS continuous variable to dichotomized variable.
3.3. Result of secondary analysis
In secondary analysis, we compared the outcomes among multiple time points of DOAC initiation after the index stroke, namely ≤3, 4–7, 8–14, and >14 days. The annualized rates of the primary composite outcome (stroke/SE/TIA/death due to stroke) tended to be higher from ≤3 to >14 days (≤3, 4–7, 8–14, and >14 days: 6.40%, 7.83%, 11.46%, and 14.40% per patient-year, respectively), and the annualized rates of the primary safety outcome (major bleeding/CRNMB) within 14 days were higher than those of the primary composite outcome (Table 3). In multivariable adjusted models, compared with the reference group (initiation timing >14 days), no significant difference was observed among the other groups. However, the primary composite outcome less likely occurred when DOACs were initiated earlier after IS (results before Bonferroni: ≤3 days, aHR: 0.64, 95% CI: 0.19–1.70; 4–7 days, aHR: 0.76, 95% CI: 0.28–1.78; 8–14 days, aHR: 0.91, 95% CI: 0.47–1.69). The risks of the primary safety outcome were not significantly higher in all the groups within 14 days compared with after 14 days after Bonferroni correction (Table 3).
Table 3.
Primary and secondary outcomes at multiple initiation cut points.
| Events | ≤3 d |
4–7 d |
8–14 d |
>14 d |
≤3 d |
4–7 d |
8–14 d |
|---|---|---|---|---|---|---|---|
| Annualized rate, %/year (n) | Adjusted HR (95% CI)#, ** | ||||||
| Primary composite outcome† | 6.40 (4) | 7.83 (6) | 11.46 (15) | 14.40 (29) | 0.64 (0.19–1.70) | 0.76 (0.28–1.78) | 0.91 (0.47–1.69) |
| Ischemic stroke | 1.60 (1) | 3.92 (3) | 5.35 (7) | 4.47 (9) | N/A | N/A | N/A |
| Ischemic events♦ | 3.20 (2) | 3.92 (3) | 6.11 (8) | 5.96 (12) | N/A | N/A | N/A |
| Primary safety outcome‡ | 9.96 (6) | 17.94 (13) | 21.11 (26) | 12.22 (24) | 1.38 (0.51–3.34) | 2.12 (1.00–4.33) | 2.01 (1.13–3.57) |
| Intracranial hemorrhage | 1.66 (1) | 0 | 2.44 (3) | 0.51 (1) | N/A | N/A | N/A |
| Major bleeding | 4.98 (3) | 0 | 5.68 (7) | 3.05 (6) | N/A | N/A | N/A |
| CRNMB | 4.98 (3) | 17.94 (13) | 15.43 (19) | 9.16 (18) | N/A | N/A | N/A |
| Secondary composite outcome* | 17.11 (10) | 32.56 (23) | 37.81 (45) | 32.93 (62) | 0.77 (0.36–1.48) | 1.35 (0.80–2.23) | 1.31 (0.88–1.94) |
Annualized rate = events/total follow-up person-year.
CRNMB: clinically relevant non-major bleeding, d: day, HR: hazard ratio, IS: ischemic stroke, N/A: not applicable, SE: systemic embolism, TIA: transient ischemic attack.
†Stroke/SE/TIA/death from ischemic stroke.
‡Major bleeding/clinically relevant non-major bleeding.
*All-cause mortality/ischemic events/acute myocardial infarction/hemorrhagic events.
♦Ischemic stroke/TIA/SE.
#Initiated >14 days as reference group; all hazard ratios were introduced by Firth correction. Adjusted model included age, sex, NIHSS on admission, CHA2DS2VASC on admission, HAS-BLED on admission, mRS on admission, reperfusion therapy and infarction size.
**Significant level corrected by Bonferroni post hoc test; therefore, no result reached statistically significant.
3.4. Results of sensitivity analysis
When we changed the initiation time point to 7th or 14th day after IS, all the results were the same as those of the main analysis except that the group in which DOACs were initiated ≤14 days had a higher risk of the primary safety outcome (major bleeding/CRNMB; aHR: 1.91, 95% CI: 1.13–3.29, P = 0.018). In the second part of sensitivity analysis, the patients were stratified by stroke severity. The results obtained at either a single time point or multiple time points did not significantly differ between the mild and moderate-to-severe stroke groups and were similar to those of the main and secondary analyses.
4. Discussion
In the present study, we retrospectively included 570 patients with AF in whom DOACs were initiated or restarted after they experienced IS (patients with TIA were excluded) and analyzed the association between the timing of using DOACs and their effectiveness or safety within 1 year after DOAC initiation. Our findings may slightly differ from those of other studies. Of the 570 patients, only 57% (326) were prescribed DOACs within 14 days after IS, and the majority of the patients (15.3%, 50/326) were prescribed DOACs on the 14th day; this finding is in contrast to the previous RAF-NOACs study in which at least 80% of the patients were prescribed DOACs within 14 days [12,13]. Considering that our patients were slightly older (median age: 80 years) and had higher risks of ischemic and hemorrhagic events (median CHA2DS2VASC score: 6 and median HAS-BLED score: 4), both physicians and patients may hesitate to use anticoagulants early according to the 2016 ESC AF guideline [15]. In addition, more than half of our patients had a medium- or large-sized infarction (n = 340, 59.7%), which may increase the risk of HT and result in physicians delaying the initiation of DOACs [23,34]. Furthermore, because of the presence of a universal insurance system in Taiwan, the insurance payment status of medication should be considered; according to NHI payment guidelines, insurance-paid DOAC prescriptions after stroke should be initiated beyond the 14th day. The insurance payment conditions may reduce the willingness of some patients to start DOACs earlier; however, 240 (49.8%) of the 482 patients still started DOACs within 14 days after IS through insurance payment. Neurologists in other countries indicated that stroke severity, infarction size, and symptomatic HT are crucial factors for deciding the optimal timing of initiating DOACs [35]. The majority (79%) of the patients with HT should be prescribed OACs at least after 14 days, and 64% of the neurologists preferred DOACs as the first choice [36].
In our study, the late initiation group had higher stroke severity, more new diagnosis of AF, and fewer patients using antithrombotic agents prior to the index stroke; hence, additional time would be required to confirm AF diagnosis and decide the optimal timing for the administration of DOACs. No study has provided robust evidence indicating the benefits of the early initiation of DOACs.
In terms of the outcomes, we observed that the annualized rates of composite bleeding events were considerably higher than those of the composite events of stroke, SE, TIA, and death due to ischemic stroke in the intervals of ≤3, 4–7, and 8–14 days. However, the timing of using DOACs was not significantly associated with thromboembolic or hemorrhagic events in the adjusted models. This finding is similar to that of a Japanese SAMURAI-NVAF study (≤3 vs. >3 days) and the UK CROMIS-2 study (≤4 vs. >4 days) [13,14] that have reported that the early initiation of OACs did not result in unfavorable outcomes and that the outcomes may not be affected by different races or components in combined outcomes. Although the composite outcome of all ischemic and hemorrhagic events in some studies may present a clinical net benefit [12,13], we still interpreted them with caution because the opposite effects of ischemia and hemorrhage may offset each other, possibly resulting in the exclusion of patients with a higher risk of bleeding immediately after their recent stroke [37].
We evaluated the association between different initiation timings and studied outcomes and observed that the risk of the primary composite outcome (stroke/SE/TIA/death due to stroke) was the lowest while initiating DOACs ≤3 days (vs. >14 days; aHR: 0.64, 95% CI: 0.19–1.70). However, the initiation of DOACs within 14 days after IS could not significantly reduce the risks of both thromboembolic and bleeding events. Yaghi et al. [38] reported that initiating OACs during 4–14 days after IS did not reduce ischemic events (vs. >14 days; odds ratio [OR]: 0.76, 95% CI: 0.36–1.62) or hemorrhagic events (vs. 0–3 days; OR: 1.49, 95% CI: 0.50–4.43). Although the reference timing and statistical method were not exactly identical, our findings are still comparable to those reported by Yaghi et al. However, the inclusion of different initiation time points and different adjusted covariables (selection method) [39] in each study would generate some diverse interpretations, which should be examined in future larger clinical trials. The findings in recent real-world data meta-analysis implied that apixaban and edoxaban may be the better choices for patients with high risk of bleeding to prevent thromboembolic events, and ESC AF guideline and EHRA practical guide both recommended individualized regimen of DOAC, e.g. drug-drug interaction, renal function or gastrointestinal disorders [17,[40], [41], [42]]. Decision of proper DOAC and its initiation time after IS are both important and challenging in clinical practice, however, considering smaller sample size and small or null events in each subgroup, we were not able to test effect modification for different DOACs as the pre-specified group for subgroup analysis.
Our sensitivity analysis findings revealed that the time point of DOAC initiation should be chosen carefully. In clinical practice, the use of multiple time points was more practical and less biased than that of a single time point. Moreover, to reduce the effect of confounding by indication and to determine which stroke severity would be appropriate for the early initiation of DOACs, we divided the patients into mild and moderate-to-severe stroke groups and slightly altered time points in accordance with the severity. No differences in outcomes were observed between the main and sensitivity analyses when we applied an NIHSS score of 8 as the severity cut point. The RELAXED study reported that the initiation of DOACs within 3 days would be suitable for patients with small- to medium-sized infarcts because the findings of patients with an infarct size of <4.0 cm3 were similar to those of overall studied patients (i.e., the earlier initiation of rivaroxaban resulted in a lower incidence of IS) [43]. Thus, the early initiation of DOACs would be more suitable for patients with mild stroke severity.
5. Strengths and limitations
This study has several limitations. First, because we used data only from a single tertiary medical center, this study included a smaller sample size and studied fewer events; had insufficient generalizability; and had missing information during the data collection process (i.e., records from other outpatient visits or outside hospitals may be lost and loss to follow-up may cause bias in estimations). However, our database includes the stroke registry that contains not only medical records but also discharge follow-up records; hence, the use of this registry could reduce the probability of underestimating major outcomes. Second, because of the observational and retrospective nature of this study, although we adjusted some crucial confounders such as brain infarction size and CHA2DS2-VASc and HAS-BLED scores, we did not consider other factors such as the use of antithrombotic agents before IS and could not adjust unmeasured factors, which may have affected results such as the timing of DOAC initiation judged by physicians. Finally, unequal group sizes would lead to reduce statistical power, i.e. it would be difficult to detect the true effect; therefore, the results of our adjusted model in main analysis should be conservatively interpreted. Nevertheless, most of our results are still in line with those of previous studies.
To the best of our knowledge, this is the first study to investigate the timing of initiating DOACs by using complete stroke registry data and image interpretation in Taiwan. Furthermore, this is also the first study to perform a detailed analysis by assessing patients with different stroke severity. Our findings provide comprehensive information regarding East Asian and Chinese populations for physicians to decide the regimens. To minimize the effect of immortal time bias and confounding by indication, we shifted the cohort entry date to the DOAC initiation date and separated patients by stroke severity, respectively.
6. Conclusion
In the present study, no significant association was observed between the timing of using DOACs and ischemic or hemorrhagic outcomes. The findings did not differ among different time points. Although we do not recommend avoiding the initiation of DOACs in the first few days after IS, we should consider that the early initiation of DOACs (≤3 days) would be appropriate only for patients who tend to experience thromboembolic events and have a low risk of bleeding. The optimal timing of initiation still must be confirmed by randomized controlled trials.
Author contribution statement
Hui-Tzu Yu: Conceived and designed the experiments; Performed the experiments; Analyzed and interpreted the data; Wrote the paper.
Kuan-Hsuan Chen: Conceived and designed the experiments; Performed the experiments; Analyzed and interpreted the data; Contributed reagents, materials, analysis tools or data; Wrote the paper.
Chun-Jen Lin and Chia-Chen Hsu: Contributed reagents, materials, analysis tools or data.
Yuh-Lih Chang: Conceived and designed the experiments, Contributed materials and analysis tools or data; Wrote the paper.
Funding statement
This study was supported by research grants from the Taipei Veterans General Hospital (V112C-228).
Data availability statement
The data that has been used is confidential.
Additional information
No additional information is available for this paper.
Table I.
ICD-CM of outcomes
| Outcome | ICD-9-CM | ICD-10-CM |
|---|---|---|
| Stroke | 430, 431, 433.01, 433.11, 433.21, 433.31, 433.81, 433.91, 434.x, 436 | I60, I61, I63, I64 |
| Transient ischemic attack | 435.x | G45 (exclude G45.3, G45.4) |
| Systemic embolism | 444.xx | I74 |
| Thromboembolic events | 433.01, 433.11, 433.21, 433.31, 433.81, 433.91, 434.xx, 436, 435.x, 443.9, 444.xx | I63, I64, G45 (exclude G45.3, G45.4), I73.9, I74 |
| Acute myocardial infarction | 410 | I21 |
| Intracranial hemorrhage | 430, 431, 432.x | I60, I61, I62 |
| Gastrointestinal bleeding | 456.0, 456.1, 531.0, 531.2, 531.4, 531.6, 532.0, 532.2, 532.4, 532.6, 533.0, 533.2, 533.4, 533.6 | I85.0, K25.0, K25.2, K25.4, K25.6, K26.0, K26.2, K26.4, K26.6, K27.0, K27.2, K27.4, K27.6, K28.0, K28.2, K28.4, K28.6, K62.5, K92.2 |
| Other site bleeding | 362.81, 379.23, 423.0, 511.1, 511.8 | H35.6x, H43.1x, I31.2, J94.2 |
| Adverse effect of anticoagulants | 964.2, 995.2 | T45.51, T45.52 |
| Hemorrhagic events | 430, 431, 432.x, 456.0, 456.1, 531.0, 531.2, 531.4, 531.6, 532.0, 532.2, 532.4, 532.6, 533.0, 533.2, 533.4, 533.6, 362.81, 379.23, 423.0, 511.1, 511.8, 964.2, 995.2 | I60, I61, I62, I85.0, K25.0, K25.2, K25.4, K25.6, K26.0, K26.2, K26.4, K26.6, K27.0, K27.2, K27.4, K27.6, K28.0, K28.2, K28.4, K28.6, K62.5, K92.2, H35.6x, H43.1x, I31.2, J94.2, T45.51, T45.52 |
Declaration of interest's statement
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
References
- 1.Álvarez-Sabín J., Maisterra O., Santamarina E., Kase C.S. Factors influencing haemorrhagic transformation in ischaemic stroke. Lancet Neurol. 2013;12:689–705. doi: 10.1016/s1474-4422(13)70055-3. [DOI] [PubMed] [Google Scholar]
- 2.Hart R.G., Coull B.M., Hart D. Early recurrent embolism associated with nonvalvular atrial fibrillation: a retrospective study. Stroke. 1983;14:688–693. doi: 10.1161/01.str.14.5.688. [DOI] [PubMed] [Google Scholar]
- 3.Paciaroni M., Bandini F., Agnelli G., et al. Hemorrhagic transformation in patients with acute ischemic stroke and atrial fibrillation: time to initiation of oral anticoagulant therapy and outcomes. J. Am. Heart Assoc. 2018;7 doi: 10.1161/JAHA.118.010133. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Shen A.Y., Yao J.F., Brar S.S., Jorgensen M.B., Chen W. Racial/ethnic differences in the risk of intracranial hemorrhage among patients with atrial fibrillation. J. Am. Coll. Cardiol. 2007;50:309–315. doi: 10.1016/j.jacc.2007.01.098. [DOI] [PubMed] [Google Scholar]
- 5.Li Y.G., Lee S.R., Choi E.K., Lip G.Y. Stroke prevention in atrial fibrillation: focus on Asian patients. Korean Circ. J. 2018;48:665–684. doi: 10.4070/kcj.2018.0190. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Connolly S.J., Ezekowitz M.D., Yusuf S., et al. Dabigatran versus warfarin in patients with atrial fibrillation. N. Engl. J. Med. 2009;361:1139–1151. doi: 10.1056/NEJMoa0905561. [DOI] [PubMed] [Google Scholar]
- 7.Granger C.B., Alexander J.H., McMurray J.J.V., et al. Apixaban versus warfarin in patients with atrial fibrillation. N. Engl. J. Med. 2011;365:981–992. doi: 10.1056/NEJMoa1107039. [DOI] [PubMed] [Google Scholar]
- 8.Patel M.R., Mahaffey K.W., Garg J., et al. Rivaroxaban versus warfarin in nonvalvular atrial fibrillation. N. Engl. J. Med. 2011;365:883–891. doi: 10.1056/NEJMoa1009638. [DOI] [PubMed] [Google Scholar]
- 9.Giugliano R.P., Ruff C.T., Braunwald E., et al. Edoxaban versus warfarin in patients with atrial fibrillation. N. Engl. J. Med. 2013;369:2093–2104. doi: 10.1056/NEJMoa1310907. [DOI] [PubMed] [Google Scholar]
- 10.Seiffge D.J., Paciaroni M., Wilson D., et al. Direct oral anticoagulants versus vitamin K antagonists after recent ischemic stroke in patients with atrial fibrillation. Ann. Neurol. 2019;85:823–834. doi: 10.1002/ana.25489. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Seiffge D.J., Traenka C., Polymeris A., et al. Early start of DOAC after ischemic stroke: risk of intracranial hemorrhage and recurrent events. Neurology. 2016;87:1856–1862. doi: 10.1212/wnl.0000000000003283. [DOI] [PubMed] [Google Scholar]
- 12.Paciaroni M., Agnelli G., Falocci N., et al. Early recurrence and major bleeding in patients with acute ischemic stroke and atrial fibrillation treated with non–vitamin‐K oral anticoagulants (RAF‐NOACs) study. J. Am. Heart Assoc. 2017;6 doi: 10.1161/jaha.117.007034. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Wilson D., Ambler G., Banerjee G., et al. Early versus late anticoagulation for ischaemic stroke associated with atrial fibrillation: multicentre cohort study. J. Neurol. Neurosurg. Psychiatry. 2019;90:320–325. doi: 10.1136/jnnp-2018-318890. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Mizoguchi T., Tanaka K., Toyoda K., et al. Early initiation of direct oral anticoagulants after onset of stroke and short- and long-term outcomes of patients with nonvalvular atrial fibrillation. Stroke. 2020;51:883–891. doi: 10.1161/STROKEAHA.119.028118. [DOI] [PubMed] [Google Scholar]
- 15.Kirchhof P., Benussi S., Kotecha D., et al. 2016 ESC Guidelines for the management of atrial fibrillation developed in collaboration with EACTS. Eur. Heart J. 2016;37:2893–2962. doi: 10.1093/eurheartj/ehw210. [DOI] [PubMed] [Google Scholar]
- 16.Powers W.J., Rabinstein A.A., Ackerson T., et al. Guidelines for the early management of patients with acute ischemic stroke: 2019 update to the 2018 guidelines for the early management of acute ischemic stroke: a guideline for healthcare professionals from the American heart association/American stroke association. Stroke. 2019;50:e344–e418. doi: 10.1161/STR.0000000000000211. [DOI] [PubMed] [Google Scholar]
- 17.Hindricks G., Potpara T., Dagres N., 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;42:373–498. doi: 10.1093/eurheartj/ehaa612. [DOI] [PubMed] [Google Scholar]
- 18.Hsieh C.Y., Chen C.H., Li C.Y., Lai M.L. Validating the diagnosis of acute ischemic stroke in a National Health Insurance claims database. J. Formos. Med. Assoc. 2015;114:254–259. doi: 10.1016/j.jfma.2013.09.009. [DOI] [PubMed] [Google Scholar]
- 19.Andrade S.E., Harrold L.R., Tjia J., et al. A systematic review of validated methods for identifying cerebrovascular accident or transient ischemic attack using administrative data. Pharmacoepidemiol. Drug Saf. 2012;21(Suppl. 1):100–128. doi: 10.1002/pds.2312. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Sung S.F., Hsieh C.Y., Lin H.J., Chen Y.W., Yang Y.H., Li C.Y. Validation of algorithms to identify stroke risk factors in patients with acute ischemic stroke, transient ischemic attack, or intracerebral hemorrhage in an administrative claims database. Int. J. Cardiol. 2016;215:277–282. doi: 10.1016/j.ijcard.2016.04.069. [DOI] [PubMed] [Google Scholar]
- 21.Chao T.F., Liu C.J., Wang K.L., et al. Using the CHA2DS2-VASc score for refining stroke risk stratification in 'low-risk' Asian patients with atrial fibrillation. J. Am. Coll. Cardiol. 2014;64:1658–1665. doi: 10.1016/j.jacc.2014.06.1203. [DOI] [PubMed] [Google Scholar]
- 22.Lund J.L., Richardson D.B., Stürmer T. The active comparator, new user study design in pharmacoepidemiology: historical foundations and contemporary application. Curr. Epidemiol. Rep. 2015;2:221–228. doi: 10.1007/s40471-015-0053-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Paciaroni M., Agnelli G., Corea F., et al. Early hemorrhagic transformation of brain infarction: rate, predictive factors, and influence on clinical outcome. Stroke. 2008;39:2249–2256. doi: 10.1161/STROKEAHA.107.510321. [DOI] [PubMed] [Google Scholar]
- 24.Steffel J., Verhamme P., Potpara T.S., et al. The 2018 European Heart Rhythm Association Practical Guide on the use of non-vitamin K antagonist oral anticoagulants in patients with atrial fibrillation. Eur. Heart J. 2018;39:1330–1393. doi: 10.1093/eurheartj/ehy136. [DOI] [PubMed] [Google Scholar]
- 25.Schulman S., Kearon C. Definition of major bleeding in clinical investigations of antihemostatic medicinal products in non-surgical patients. J. Thromb. Haemostasis. 2005;3:692–694. doi: 10.1111/j.1538-7836.2005.01204.x. [DOI] [PubMed] [Google Scholar]
- 26.Kaatz S., Ahmad D., Spyropoulos A.C., Schulman S. Subcommittee on Control of Anticoagulation, Definition of clinically relevant non-major bleeding in studies of anticoagulants in atrial fibrillation and venous thromboembolic disease in non-surgical patients: communication from the SSC of the ISTH. J. Thromb. Haemostasis. 2015;13:2119–2126. doi: 10.1111/jth.13140. [DOI] [PubMed] [Google Scholar]
- 27.Birman-Deych E., Waterman A.D., Yan Y., Nilasena D.S., Radford M.J., Gage B.F. Accuracy of ICD-9-CM codes for identifying cardiovascular and stroke risk factors. Med. Care. 2005;43:480–485. doi: 10.1097/01.mlr.0000160417.39497.a9. [DOI] [PubMed] [Google Scholar]
- 28.Friberg L., Skeppholm M. Usefulness of Health Registers for detection of bleeding events in outcome studies. Thromb. Haemostasis. 2016;116:1131–1139. doi: 10.1160/th16-05-0400. [DOI] [PubMed] [Google Scholar]
- 29.Ozaki A.F., Choi A.S., Le Q.T., et al. Real-world Adherence and persistence to direct oral anticoagulants in patients with atrial fibrillation: a systematic review and meta-analysis. Circ. Cardiovasc. Qual. Outcomes. 2020;13 doi: 10.1161/CIRCOUTCOMES.119.005969. [DOI] [PubMed] [Google Scholar]
- 30.Lip G.Y., Nieuwlaat R., Pisters R., Lane D.A., Crijns H.J. Refining clinical risk stratification for predicting stroke and thromboembolism in atrial fibrillation using a novel risk factor-based approach: the euro heart survey on atrial fibrillation. Chest. 2010;137:263–272. doi: 10.1378/chest.09-1584. [DOI] [PubMed] [Google Scholar]
- 31.Pisters R., Lane D.A., Nieuwlaat R., de Vos C.B., Crijns H.J., Lip G.Y. A novel user-friendly score (HAS-BLED) to assess 1-year risk of major bleeding in patients with atrial fibrillation: the Euro Heart Survey. Chest. 2010;138:1093–1100. doi: 10.1378/chest.10-0134. [DOI] [PubMed] [Google Scholar]
- 32.Cheng C.L., Kao Y.H., Lin S.J., Lee C.H., Lai M.L. Validation of the national health insurance research database with ischemic stroke cases in Taiwan. Pharmacoepidemiol. Drug Saf. 2011;20:236–242. doi: 10.1002/pds.2087. [DOI] [PubMed] [Google Scholar]
- 33.Heinze G., Schemper M. A solution to the problem of monotone likelihood in Cox regression. Biometrics. 2001;57:114–119. doi: 10.1111/j.0006-341X.2001.00114.x. [DOI] [PubMed] [Google Scholar]
- 34.Paciaroni M., Agnelli G., Falocci N., et al. Early recurrence and cerebral bleeding in patients with acute ischemic stroke and atrial fibrillation: effect of anticoagulation and its timing: the RAF study. Stroke. 2015;46:2175–2182. doi: 10.1161/strokeaha.115.008891. [DOI] [PubMed] [Google Scholar]
- 35.Escudero-Martinez I., Mazya M., Teutsch C., et al. Dabigatran initiation in patients with non-valvular AF and first acute ischaemic stroke: a retrospective observational study from the SITS registry. BMJ Open. 2020;10 doi: 10.1136/bmjopen-2020-037234. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Rybinnik I., Wong S., Mehta D., et al. Anticoagulation choice and timing in stroke due to atrial fibrillation: a survey of US stroke specialists (ACT-SAFe) J. Stroke Cerebrovasc. Dis. 2020;29 doi: 10.1016/j.jstrokecerebrovasdis.2020.105169. [DOI] [PubMed] [Google Scholar]
- 37.Cordoba G., Schwartz L., Woloshin S., Bae H., Gøtzsche P.C. Definition, reporting, and interpretation of composite outcomes in clinical trials: systematic review. BMJ. 2010;341:c3920. doi: 10.1136/bmj.c3920. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Yaghi S., Trivedi T., Henninger N., et al. Anticoagulation timing in Cardioembolic stroke and recurrent event risk. Ann. Neurol. 2020;88:807–816. doi: 10.1002/ana.25844. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Heinze G., Wallisch C., Dunkler D. Variable selection - a review and recommendations for the practicing statistician. Biom. J. 2018;60:431–449. doi: 10.1002/bimj.201700067. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Menichelli D., Del Sole F., Di Rocco A., et al. Real-world safety and efficacy of direct oral anticoagulants in atrial fibrillation: a systematic review and meta-analysis of 605 771 patients. Eur. Heart J. Cardiovasc. Pharmacother. 2021;7:f11–f19. doi: 10.1093/ehjcvp/pvab002. [DOI] [PubMed] [Google Scholar]
- 41.Kim S.M., Jeon E.T., Jung J.M., Lee J.S. Real-world oral anticoagulants for Asian patients with non-valvular atrial fibrillation: a PRISMA-compliant article. Medicine. 2021;100 doi: 10.1097/md.0000000000026883. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Steffel J., Collins R., Antz M., et al. 2021 European heart rhythm association practical guide on the use of non-vitamin K antagonist oral anticoagulants in patients with atrial fibrillation. Europace. 2021;23:1612–1676. doi: 10.1093/europace/euab065. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Yasaka M., Minematsu K., Toyoda K., et al. Rivaroxaban administration after acute ischemic stroke: the RELAXED study. PLoS One. 2019;14 doi: 10.1371/journal.pone.0212354. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The data that has been used is confidential.