Introduction
Atrial fibrillation (AF) is a major cause of ischemic stroke, and oral anticoagulation remains the cornerstone of secondary prevention [1]. Direct oral anticoagulants (DOACs) are generally preferred over warfarin due to their superior safety profile and ease of administration [2]. Historically, clinical guidelines recommended delaying anticoagulation, particularly in patients with acute large infarcts, to reduce the risk of hemorrhagic transformation [3].
However, recent randomized controlled trials have questioned this conventional approach. The Early versus Late initiation of direct oral Anticoagulants in post-ischaemic stroke patients with atrial fibrillatioN (ELAN) trial [4] demonstrated that early DOAC initiation was associated with a numerically lower incidence of the 30-day composite outcome (2.9% vs. 4.1%; risk difference, -1.18%; 95% confidence interval [CI], -2.84 to 0.47) without an increased risk of symptomatic intracranial hemorrhage (SICH). Similarly, the TIMING of Oral Anticoagulant Therapy in Acute Ischemic Stroke With Atrial Fibrillation (TIMING) trial [5] confirmed the non-inferiority of early DOAC initiation (within 4 days), reporting no cases of SICH in either study arm. The Optimal Timing of Anticoagulation After Acute Ischaemic Stroke (OPTIMAS) trial [6], which enrolled more than 3,600 patients, also showed non-inferiority, with identical rates of the primary composite outcome between early (≤4 days) and delayed (7-14 days) initiation (3.3% vs. 3.3%, respectively). Recently, the Collaboration on the Optimal Timing of Anticoagulation After Ischaemic Stroke and Atrial Fibrillation: Prospective Individual Participant Data Meta-Analysis of Randomised Controlled Trials (CATALYST) collaboration conducted an individual participant data meta-analysis involving 5,441 patients across these four trials. Early DOAC initiation (≤4 days) significantly reduced the 30-day risk of recurrent ischemic stroke or intracranial hemorrhage (odds ratio [OR], 0.70; 95% CI, 0.50-0.98), without increasing the incidence of SICH, thereby reinforcing the safety and efficacy of early anticoagulation after acute ischemic stroke (AIS) [7].
The risk of intracerebral hemorrhage is notably higher in Asian populations than in Western populations [8]. Asian patients with AF display a distinct clinical phenotype characterized by elevated baseline risk of both ischemic stroke [9] and intracerebral hemorrhage [10]. Large-scale trials such as the Randomized Evaluation of Long-Term Anticoagulation Therapy (RE-LY) [11] and the Effective Anticoagulation with Factor Xa Next Generation in Atrial Fibrillation-Thrombolysis in Myocardial Infarction 48 (ENGAGE AF-TIMI 48) [12] have consistently demonstrated that warfarintreated Asian participants experience two to three times higher incidence of intracranial hemorrhage than their non-Asian counterparts (RE-LY: 0.75%/yr vs. 0.32%/yr, respectively; ENGAGE AF-TIMI 48: 1.92%/yr vs. 0.74%/yr, respectively). Observational findings from the International, Multicentre, Case-control Study of Risk Factors for Stroke (INTERSTROKE) study were consistent with these results, reporting that intracerebral hemorrhage accounted for 27.6% of all strokes in China and 27.5%-33.3% in other Asian countries, compared with only 6.7%-8.4% in North America and Europe [8]. These disparities likely reflect a combination of genetic predisposition, a high prevalence of hypertension, and differential responses to antithrombotic therapy [1].
Given the heightened bleeding risk, clinicians in Asia have traditionally adopted a conservative approach to anticoagulation initiation following AIS. However, the persistent clinical uncertainty regarding potential racial or ethnic differences in optimal DOAC initiation timing remains a key barrier to standardizing post-stroke care in diverse populations. This substudy, therefore, aimed to evaluate whether the efficacy and safety of early versus late DOAC initiation differ between Asian and non-Asian patients, focusing on ischemic events, bleeding complications, and mortality. Additionally, it sought to determine whether region-specific considerations should inform anticoagulation timing in AIS.
Methods
Study design and settings
We conducted a post hoc subgroup analysis of the ELAN randomized trial, which enrolled 2,013 patients with AIS and AF across 15 countries. Participants were randomized to receive either early (within 48 hours after minor or moderate ischemic stroke and 6 or 7 days for major ischemic stroke) or late (3 or 4 days for minor ischemic stroke; 6 or 7 days for moderate ischemic stroke; 12, 13, or 14 days for major ischemic stroke) DOAC initiation [4,13-17]. Patients were categorized as “Asian” or “non-Asian” based on geographical region; specifically, those enrolled in Japan and India were classified as Asian, whereas patients from all other countries were considered non-Asian. Self-reported race or ethnicity information was not collected in the trial database. No additional exclusion criteria were employed for this subgroup analysis. Details of the trial protocol and data collection methods have been published previously [18].
Outcomes
The endpoints for this subgroup analysis were consistent with those of the main ELAN trial. The primary outcome was a composite of major extracranial bleeding, SICH, recurrent ischemic stroke, systemic embolism, vascular death within 30 days after randomization. Secondary outcomes included the same composite outcome assessed at 90 days; the individual components of the composite at both 30 and 90 days; and functional outcome measured by the modified Rankin Scale (mRS) at 30 and 90 days. Systemic embolism was defined, according to the trial protocol, as an abrupt vascular insufficiency associated with clinical and/or radiological evidence of arterial occlusion of an extremity or organ, in the absence of another likely mechanism such as atherosclerosis, instrumentation, or trauma [19].
The mRS ranges from 0 (no symptoms) to 6 (death), with lower scores reflecting better neurological function.
Events were identified based on clinical symptoms and relevant diagnostic investigations rather than protocol-mandated imaging and were adjudicated centrally by an independent clinical event committee according to predefined criteria. Diagnostic and adjudication procedures were standardized across all participating regions.
Statistical analysis
Baseline demographics and clinical characteristics were summarized for four subgroups, based on randomization groups and Asian or non-Asian groups, for patients with available primary outcome. Continuous variables were reported as medians and interquartile ranges (IQR), and categorical variables as counts and percentages. Asian and non-Asian groups were compared using Wilcoxon’s rank-sum tests for continuous variables and chi-square or Fisher’s exact tests for categorical variables, as appropriate.
For outcome analysis, we first tabulated the number and proportion of events for each endpoint by subgroup. The association between geographic region (Asian vs. non-Asian) and each primary and secondary outcome was assessed. For the composite outcomes within 30 and then 90 days, penalized logistic regression models were used to estimate ORs and 95% CIs for Asians versus non-Asians. Crude (unadjusted) and adjusted ORs were reported. To prevent overfitting due to the limited number of events, the adjusted models were restricted to age, sex, and National Institutes of Health Stroke Scale (NIHSS) score at admission to stabilize the model. We evaluated similar models for major extracranial bleeding and vascular death. The other individual components of the composite outcomes were not analyzed beyond descriptive statistics.
In addition, to formally assess whether the effect of early versus late DOAC initiation on functional outcome differed by region, we fitted ordinal logistic regression models for mRS at 30 and 90 days including region (Asian vs. non-Asian), randomization group (early vs. late), and their interaction term (region× treatment timing).
Ethical approval and informed consent
The study protocol was initially approved by the Swiss Association of Research Ethics in Switzerland (swissethics; approval no. 2017-00588). In addition, the protocol was approved by the relevant local ethics committees or institutional review boards in all 15 participating countries, and informed consent was obtained as required. The trial followed ICH-E6 good clinical practice and the Declaration of Helsinki.
Results
Among 247 Asian and 1,766 non-Asian patients enrolled in the ELAN trial, 1,006 were randomized to early DOAC initiation and 1,007 to delayed initiation, with balanced allocation across Asian and non-Asian subgroups. Primary outcome data were available for 1,975 patients (245 Asians and 1,730 non-Asians) (Figure 1). Asian patients exhibited a lower incidence of minor strokes (26.1% vs. 39.2%) and a higher incidence of major strokes (33.5% vs. 21.3%) compared with non-Asian patients. Stroke severity was consistently greater among Asian patients, as indicated by significantly higher median NIHSS scores both at admission (10 [early vs. late DOAC initiation: 13 vs. 9] vs. 5 [5 vs. 5], P<0.01 for Asians vs. non-Asians) and prior to randomization (6 [7 vs. 6] vs. 3 [2 vs. 3], P<0.01) (Table 1). The distribution of DOAC type and dosing differed significantly between Asian and non-Asian patients (Supplementary Table 1). Overall, DOAC dose reduction was more frequent among Asian patients than non-Asian patients (43.3% vs. 15.0%, P<0.01). The overall median time from stroke onset to DOAC initiation was 3 days (IQR, 1-6). The time was significantly longer in Asian patients than in non-Asian patients (5 days [IQR, 1-6] vs. 3 days [IQR, 1-6], P<0.01). When stratified by stroke severity, the time increased progressively: 2 days (IQR, 1-2) in patients with minor stroke, 3 days (IQR, 1-5) in those with moderate stroke, and 11 days (IQR, 6-11) in those with major stroke (P<0.01). Information on missing data is summarized in Supplementary Tables 2 and 3.
Baseline characteristics of the Japanese and Indian subgroups are shown in Supplementary Table 4. In these supplemental analyses, compared with Indian patients, Japanese patients were significantly older (80 vs. 70 years, P<0.01), had a higher prevalence of prior heart failure classified according to the New York Heart Association functional class (20.5% vs. 7.7%, P=0.04), had a lower prevalence of diabetes mellitus (20.0% vs. 34.0%, P=0.04), and underwent thrombectomy more frequently (26.6% vs. 13.2%, P<0.05). Outcomes between Japanese and Indian patients are presented in Supplementary Table 5.
At 30 days, the primary composite outcome occurred in 6.5% of Asian patients (early vs. late DOAC initiation: 4.8% vs. 8.3%) and in 3.1% of non-Asian patients (2.7% vs. 3.6%; P=0.01 for Asians vs. non-Asians). This difference was primarily driven by higher rates of recurrent ischemic stroke (Asians vs. non-Asians: 4.1% [early vs. late: 2.4% vs. 5.8%] vs. 1.7% [1.3% vs. 2.1%]) and systemic embolism (2.0% [0.8% vs. 3.3%] vs. 0.5% [0.4% vs. 0.6%]), rather than major extracranial bleeding (0.4% [0% vs. 0.8%] vs. 0.4% [0.4% vs. 0.5%]), SICH (0% [0% vs. 0%] vs. 0.2% [0.2% vs. 0.2%]), or vascular death (2.0% [2.4% vs. 1.7%] vs. 0.9% [0.9% vs. 0.9%]) (Table 2). Outcomes at 30 days in Asian and non-Asian patients are shown in Figure 2. At 90 days, the composite outcome occurred in 7.5% of Asian patients (early vs. late: 4.9% vs. 10.2%) and 4.3% of non-Asian patients (3.5% vs. 5.0%; P=0.03). This difference was again mainly attributable to higher rates of systemic embolism among Asian patients (2.1% [0.8% vs. 3.4%] vs. 0.5% [0.4% vs. 0.7%]). Clinical outcomes according to region and timing of DOAC initiation are summarized in Table 2.
ORs for early versus late DOAC initiation stratified by region are summarized in Table 3. For the composite outcome, early initiation was associated with numerically lower event rates at both 30 days (6.5% in Asians [adjusted OR, 0.53; 95% CI, 0.19- 1.47] and 3.1% in non-Asians [0.79, 0.46-1.37]) and 90 days (7.5% in Asians [0.42, 0.16-1.14] vs. 4.3% in non-Asians [0.74, 0.46-1.20]), although statistical significance was not reached.
No significant treatment-by-region interactions were observed for the primary composite outcome at either 30 days (P for interaction=0.49) or 90 days (P for interaction=0.32), indicating consistent relative effects of early versus late DOAC initiation across Asian and non-Asian patients. Event rates for major extracranial bleeding and vascular death were low in both groups at 30 and 90 days. No significant interaction effects were ob-served for these outcomes (all P>0.05), suggesting consistent effects of early versus late DOAC initiation across ethnic subgroups.
At 30 days, 44.1% of Asian patients achieved an of 0-2 versus 66.1% of non-Asian patients (P<0.01 for Asians vs. non-Asians); at 90 days, these proportions were 48.8% and 69.6%, respectively (P<0.01). Distribution of mRS categories according to treatment arm at 90 days for Asian and non-Asian participants is shown in Supplementary Figure 1. In ordinal logistic regression analyses of functional outcome, Asian region remained significantly associated with worse mRS scores at both 30 and 90 days (Supplementary Table 6). Significant interaction terms between region and treatment timing were also observed, indicating differential shifts in functional outcome distributions.
After additional adjustment for DOAC dose reduction, the regional differences in outcomes were not materially attenuated (Supplementary Table 7).
Discussion
In this post hoc subgroup analysis of the ELAN trial, early initiation of DOACs after AIS was associated with similar relative effects on the primary composite outcome in Asian and non-Asian patients, with no evidence of treatment-by-region het-erogeneity. Despite substantial differences in baseline characteristics, the effects of early versus later DOAC initiation on ischemic and bleeding events were consistent across regions. Importantly, early DOAC initiation in Asian patients—who presented with more severe strokes and a higher baseline risk profile—was not associated with an increased risk of hemorrhagic complications, including SICH. Although Asian patients experienced worse overall functional outcomes, these differences appear to reflect greater baseline stroke severity and post-stroke recovery processes rather than intrinsic ethnic differences in response to anticoagulation therapy.
Our findings provide three key insights. First, the direction of effect favored early initiation in both Asians and non-Asians; however, wide CIs precluded statistical significance within subgroups. Second, despite longstanding concerns regarding elevated bleeding risks among Asian patients, the safety profile ob-served was reassuring. Notably, no cases of SICH were reported in the Asian group, although one-third of these patients met the ELAN criteria for major strokes. Third, although baseline severity was substantially higher in Asians (NIHSS score ≥10 in 41% vs. 14% in non-Asians), early initiation was not linked to adverse outcomes. These findings align with an infarct size- based ELAN sub-analysis showing no differential treatment effect across stroke severities (minor, moderate, and major) [14]. These results support the conclusion that early anticoagulation can be implemented safely in Asian patients. This is particularly important given that East Asians have a higher baseline risk of intracerebral hemorrhage, especially in the setting of anticoagulation [20-22]. The absence of SICH in this high-risk group reinforces the hypothesis that early DOAC administration should not be deferred solely based on ethnicity or geographic origin.
The higher stroke severity observed in Asians may, in part, reflect enrollment practices. Although thrombectomy rates were similar (23% vs. 22%), individuals in the non-Asian cohort reported fewer severe strokes, raising the possibility that Asian centers enrolled proportionally more severe cases. Thus, comparable thrombectomy rates do not imply equivalent baseline severity across regions.
Additional regional differences were noted: Asians had higher rates of prior stroke and diabetes but lower dyslipidemia; however, whether these patterns reflect genetic predisposition or diagnostic practices remains uncertain. The higher prevalence of aspirin use (>50%) in the non-Asian group may indicate regional prevention strategies—particularly greater primary-prevention use in Western settings—amid differing perceptions of ICH risk [23]. Beyond these baseline differences, Asian patients experienced worse short-term outcomes than non-Asian patients. Notably, Asian patients presented with higher admission NIHSS scores, initiated DOAC therapy later, and more frequently received dosereduced DOACs, suggesting that baseline stroke severity likely contributed substantially to the outcome differences observed between regions. Despite these outcome differences, no treatment-by-region interaction was observed for the primary composite outcome at either 30 days or 90 days (P=0.49 and P=0.32, respectively) (Table 3). Consistent with this, the exploratory ordinal logistic regression analysis of mRS adjusted for NIHSS score at admission showed no significant Asian×early interaction at either time point (30 days: OR, 1.43, 95% CI [0.89-2.30], P=0.14; 90 days: OR, 1.31, 95% CI [0.82-2.09], P=0.26) (Supplementary Table 6). These findings suggest that the effect of early DOAC initiation did not differ meaningfully by region, and that the observed outcome differences between Asian and non-Asian patients should not be attributed to a differential treatment response. Given the exploratory, post hoc nature of the mRS analysis and the limited sample size, these results should be interpreted cautiously.
Certain limitations of this subgroup analysis should be acknowledged. First, as a post hoc analysis, the findings are exploratory, which precludes definitive causal inferences. This subgroup analysis was not pre-specified in the original ELAN trial protocol, and the study was not powered to detect differences based on race. Consequently, the statistical power was limited. Second, the small number of events in each subgroup, particularly for individual components of the endpoints, may have resulted in imprecise or and unstable estimates. Therefore, the findings of this analysis should be considered exploratory and interpreted with caution. Third, the racial classification was based on the country of enrollment rather than self-reported ethnicity, which may not accurately reflect an individual’s racial or eth-nic background. Finally, potential regional differences in DOAC selection and post-stroke rehabilitation practices were not accounted for and may have influenced the outcomes. As a limitation, although Japanese and Indian patients were analyzed together as an Asian cohort for statistical reasons, supplemental analyses revealed substantial differences in baseline characteristics between these populations, highlighting heterogeneity within Asian patients and warranting cautious interpretation of region-based subgroup analyses. Furthermore, the lower rate of baseline aspirin use among Asian patients compared to non-Asian patients (26.5% vs. 53.1%) represents a potential confounder. Given that prior antiplatelet therapy influences the risk of early stroke recurrence, this imbalance might have partially contributed to the higher event rates observed in the Asian cohort.
Conclusions
Early DOAC initiation in Asian patients with AF and AIS appears safe and potentially efficacious, without a need for ethnicityspecific modifications to treatment timing. Anticoagulation should not be delayed solely on the basis of ethnicity or initial stroke severity during acute management.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
Supplement1
Print
