Early anticoagulation after ischemic stroke in atrial fibrillation: Moving beyond habitual delay
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
For decades, the timing of anticoagulation after ischemic stroke in atrial fibrillation (AF) has been guided more by caution than by evidence. Early initiation was widely assumed to carry an unacceptable risk of hemorrhagic transformation, and this assumption shaped practice despite limited empirical support. Major guidelines reflected this uncertainty. The 2021 American Heart Association/American Stroke Association (AHA) secondary prevention guideline recommended initiating anticoagulation between 2 and 14 days after stroke, with timing individualized by infarct size and perceived bleeding risk.[1] CHEST guidance adopted a similar stance,[2] while more recent AHA scientific statements offered broad, non-directive advice.[3] This conservative approach was historically understandable. Hemorrhagic transformation is visually dramatic and clinically alarming, whereas early recurrent embolic events are often overshadowed by the index infarct, despite being clinically and radiographically detectable and prognostically relevant. Yet observational data have long shown that the risk of recurrent embolism after AF-related stroke is concentrated in the first 1–2 weeks,[4] and that early recurrent ischemia frequently exceeds the risk of symptomatic intracranial hemorrhage. In the absence of randomized trials, however, delay became standard practice. Early Signals Challenging the Safety Assumption That evidentiary gap has now been substantially narrowed. Early challenges to routine delay emerged from real-world data. In a pooled analysis of more than 4,900 patients with AF-related ischemic stroke, initiation of direct oral anticoagulants (DOACs) at a median of 5 days was associated with lower intracranial hemorrhage risk than Vitamin K antagonists, without an increase in recurrent ischemic events.[5] These findings set the stage for randomized trials directly addressing whether early anticoagulation after AF-related ischemic stroke truly confers excess hemorrhagic risk when patients are selected based on imaging stability [Table 1]. Together, these studies have substantially altered the evidentiary landscape and challenged the rationale for routine, time-based delay.Table 1: Key randomized trials and individual patient data analyses informing the timing of direct oral anticoagulant initiation after atrial fibrillation–related ischemic strokeThe central question is therefore no longer whether early anticoagulation is safe in selected patients, but whether routine delay can still be justified in the absence of concerning imaging findings. This commentary argues that, for the majority of patients with AF-related ischemic stroke and stable follow-up imaging, early initiation of DOACs should be considered the default approach. The following discussion synthesizes contemporary randomized and real-world evidence to support this position and to clarify how emerging data should inform future practice. Randomized Evidence on Anticoagulation Timing TIMING was the first modern randomized trial to compare early DOAC initiation (≤ 4 days) with later initiation (5–10 days).[6] Early treatment met non-inferiority criteria for the composite outcome of recurrent ischemic stroke, symptomatic intracerebral hemorrhage, or death. No symptomatic intracerebral hemorrhages occurred in either group, providing clear reassurance regarding short-term safety, although the trial was not powered to demonstrate superiority. ELAN refined the question by anchoring treatment timing to imaging-defined infarct size rather than clinical severity.[7] Early initiation occurred within 48 h for minor and moderate infarcts, with later initiation for large strokes. Importantly, patients with minor hemorrhagic transformation (HI-1 or HI-2) were included if imaging was stable. Early initiation was not associated with excess symptomatic intracranial hemorrhage, and no protective effect of delay was observed. ELAN provided critical reassurance in a population previously considered high risk, while still excluding patients with unstable imaging or major parenchymal hematomas. OPTIMAS provided the largest randomized safety dataset to date.[8] Among 3,621 patients randomized to early (≤ 4 days) or delayed (7–14 days) DOAC initiation, the primary composite outcome occurred in 3.3% of patients in both groups. Symptomatic intracranial hemorrhage occurred in fewer than 1% of patients overall. No subgroup, including those treated with intravenous thrombolysis or mechanical thrombectomy, demonstrated a safety advantage with delayed initiation, provided follow-up imaging at 24–48 h was stable. These results apply primarily to anticoagulant-naïve patients without severe brain edema, symptomatic parenchymal hematoma, or evolving radiographic injury. START, though smaller and adaptive in design, similarly failed to identify any safety benefit of later initiation,[9] reinforcing the consistency of randomized evidence. The CATALYST individual patient data meta-analysis pooled 5,441 participants from TIMING, ELAN, OPTIMAS, and START.[10] Early anticoagulation was associated with a modest reduction in recurrent ischemic stroke at 30 days, without an increase in symptomatic intracranial hemorrhage. Importantly, safety signals were consistent across infarct sizes, stroke severities, and reperfusion modalities, within the constraints of trial inclusion criteria. The major randomized trials and individual patient data meta-analysis are summarized in Table 1. What the Trials Tell Us – and What They Do Not Importantly, not all meta-analyses demonstrate a statistically significant reduction in early ischemic recurrence with early initiation. Recent systematic reviews aggregating randomized and prospective cohort data confirm the safety of early DOAC use but report neutral effects on ischemic outcomes, likely reflecting low absolute event rates, heterogeneous timing definitions, and limited statistical power.[11] These findings do not suggest a protective effect of delay; rather, they indicate that early initiation does not incur excess harm while maintaining anticoagulant protection during a period of heightened embolic risk. Hemorrhagic Transformation: Reframing Risk Hemorrhagic transformation remains a major driver of delayed treatment. However, most hemorrhagic transformations, particularly HI-type petechial lesions frequently observed after reperfusion, are clinically benign. A small, randomized study evaluating early rivaroxaban initiation in patients with HI-1 hemorrhagic transformation reported no safety signal and suggested improved functional outcomes compared with delayed treatment,[12] although sample size and external validity are limited. Observational cohorts similarly indicate that HI-1 and HI-2 lesions rarely progress or become symptomatic after DOAC initiation.[13] In contrast, symptomatic or expanding parenchymal hematomas (PH-1 and PH-2) remain clear contraindications to early anticoagulation and were largely excluded from randomized trials.[14] Reperfusion Therapy and Anticoagulation Timing Concerns related to reperfusion therapy have also diminished. Across randomized and observational analyses, thrombolysis or thrombectomy per se has not emerged as the primary determinant of bleeding risk.[7,8,15] Imaging stability, rather than the use of reperfusion therapy, appears to be the critical factor in guiding safe initiation. Real-world Evidence and Evolving Practice Real-world data support earlier treatment. A large nationwide Chinese registry demonstrated better functional outcomes when DOACs were initiated earlier within severity-stratified windows, compared with later initiation.[16] Earlier heuristic rules, such as the “1–3–6–12-day” approach, were pragmatic tools derived from non-randomized cohorts.[17] In the context of contemporary randomized evidence, they no longer represent the best available standard. Microbleeds and Individualized Risk Assessment Additional observational data add nuance rather than contradiction. Cerebral microbleeds are associated with an increased relative risk of anticoagulation-related intracranial hemorrhage. In the CROMIS-2 cohort, microbleed presence independently predicted hemorrhage risk in anticoagulated patients with AF after ischemic stroke or TIA.[18] Nevertheless, absolute rates of recurrent ischemic stroke exceeded those of intracranial hemorrhage even among patients with microbleeds, and no threshold was identified at which anticoagulation conferred clear net harm. These findings support individualized risk assessment rather than categorical delay, particularly in the absence of extensive microbleed burden or additional hemorrhage-prone features, like cortical superficial siderosis. Biological Rationale for Early Anticoagulation The biological rationale aligns with these clinical observations. AF-related embolism reflects persistent atrial thrombogenesis and blood stasis that continue after stroke, and antiplatelet therapy does not adequately mitigate this risk.[19] DOACs, with rapid onset and a favorable safety profile compared with warfarin, are well-suited to early use.[20] Limits of the Evidence Important limitations remain. Patients with malignant brain edema, large infarcts with mass effect, symptomatic parenchymal hematomas, unstable imaging, or advanced hemorrhage-prone small-vessel disease were underrepresented in trials and warrant caution. These scenarios remain appropriate reasons to delay therapy. However, they represent a minority of patients with AF-related ischemic stroke. Implications for Clinical Practice Current stroke and cardiology guidelines were developed before the availability of contemporary randomized trials specifically addressing anticoagulation timing after AF-related ischemic stroke. As a result, they emphasize time-based delay rather than imaging-defined risk stratification. Until guideline updates incorporate these data, clinicians must reconcile evolving evidence with existing recommendations. Three conclusions emerge from contemporary evidence. First, early initiation of DOACs after AF–related ischemic stroke is safe in patients with stable imaging and without parenchymal hematoma. Second, infarct size and imaging stability are more relevant determinants of bleeding risk than arbitrary time thresholds. Third, routine delay should no longer be the default strategy but a targeted response to clearly defined high-risk scenarios. As randomized evidence continues to mature, clinical practice and future guidelines should increasingly favor imaging-guided rather than time-based decision making. Ethical policy and institutional review board statement Not applicable. Data availability statement Data sharing is not applicable to this article as no datasets were generated and/or analyzed. Financial support and sponsorship Nil. Conflicts of interest There are no conflicts of interest.
Fields of Study & Subjects
Atrial Fibrillation Management and Outcomes
Acute Ischemic Stroke Management
Venous Thromboembolism Diagnosis and Management
Publication details
- Journal / source
- Brain Circulation
- Volume
- 12
- Issue
- 2
- Pages
- 179–182
- Publication date
- 30 March 2026
- Language
- EN