Few stroke-unit questions get asked more often than “when do we start the blood thinner?” The honest answer is that the timing is a balance, not a fixed date: too early risks symptomatic hemorrhagic transformation of fresh infarct or expansion of a fresh bleed; too late leaves a high-risk patient exposed to early recurrent embolism. This page summarizes a pragmatic, imaging-anchored approach for two distinct scenarios — initiating a direct oral anticoagulant (DOAC) after an ischemic stroke in atrial fibrillation (AFib), and resuming anticoagulation after intracerebral hemorrhage (ICH). It is educational and not a substitute for individualized judgment; guidelines and institutional protocols vary, and the numbers below are common frameworks rather than mandates.

Starting a DOAC after ischemic stroke in AFib: the imaging-guided ladder

The most widely taught approach scales the delay to infarct burden, because infarct size is the dominant driver of hemorrhagic-transformation risk. A common bedside frame:

  • TIA or a very small infarct → start day 1–2.
  • Mild infarct → start day 3–4.
  • Moderate infarct → start day 6–7.
  • Large infarct → defer to roughly day 12–14, with repeat imaging first to exclude hemorrhagic transformation.

This is the spirit of the older “1-3-6-12 day” rule of thumb, but read it as a starting point that newer trial data have shifted earlier, not as a hard schedule. Lesion size is best judged on the relevant modality (DWI volume, or CT/CT-perfusion correlate), and reperfusion therapy, large vessel territory, and uncontrolled hypertension all push you toward the more cautious end.

Bedside pearl

Anchor the decision to the infarct you can see, not the day number. A small DWI lesion in a high-CHA₂DS₂-VASc patient rarely needs a two-week wait; a large MCA territory infarct earns repeat imaging before any anticoagulant, regardless of how good the patient looks.

What ELAN and OPTIMAS actually showed

Two randomized trials reshaped this conversation. ELAN (NEJM 2023) randomized patients with AFib-associated ischemic stroke to early versus later DOAC initiation, with timing categorized by infarct size; early starts were broadly within the first few days and later starts followed the traditional schedule. Early initiation was associated with a low rate of the primary composite of recurrent ischemia, systemic embolism, major bleeding, and vascular death, and there was no excess of symptomatic intracranial hemorrhage with the earlier strategy.1 OPTIMAS, a larger phase 4 UK trial, tested early (≤ 4 days) versus delayed (day 7–14) DOAC initiation and met its prespecified non-inferiority margin, supporting earlier-than-historical initiation after an imaging review to exclude hemorrhage.2

The combined message is reassuring but bounded: across the populations studied, starting earlier did not produce the symptomatic hemorrhage signal clinicians feared, and it may reduce early recurrent strokes. Neither trial licenses anticoagulating a massive infarct on day 1, and very large infarcts were under-represented — so the large-infarct, day-12–14-with-repeat-imaging arm of the ladder remains a reasonable, cautious default.

Pitfall

Do not read “early is non-inferior” as “early in everyone.” Asymptomatic hemorrhagic transformation, a large infarct, ongoing uncontrolled blood pressure, or a planned invasive procedure are all reasons to hold. Always re-image before anticoagulating a large infarct, and reconcile timing with any recent thrombolysis.

Resuming anticoagulation after intracerebral hemorrhage

Restarting anticoagulation after ICH is a higher-stakes, lower-evidence decision, and the bleed phenotype matters more than any calendar. The first question is the suspected mechanism:

  • Lobar ICH with suspected cerebral amyloid angiopathy (CAA): high recurrent-hemorrhage risk. Anticoagulation is frequently deferred or avoided, and a left atrial appendage occlusion strategy is often considered instead when AFib is the indication.
  • Deep, hypertensive ICH: generally a lower recurrence risk once blood pressure is well controlled. Resumption is typically reconsidered at roughly 4–8 weeks, conditional on durable BP control and a stable follow-up scan.

These windows are conventions, not trial-proven thresholds; the optimal timing and the choice of agent after ICH remain areas of genuine uncertainty and active study. Reversibility of the underlying bleeding risk (treatable hypertension vs. fixed amyloid microangiopathy), MRI markers such as lobar microbleeds and cortical superficial siderosis, the strength of the anticoagulation indication, and patient preference should all feed the decision.

Weighing the two risks and documenting it

Whichever scenario you are in, the core task is the same: quantify the thrombotic risk you are trying to prevent and the bleeding risk you are accepting. For AFib, CHA₂DS₂-VASc estimates the annual ischemic-stroke risk that justifies anticoagulation, and guidelines from the AHA/ASA and the European Stroke Organisation both endorse anticoagulation for secondary prevention in AFib while leaving timing to clinical and imaging judgment.34 For the post-ICH patient, the competing number is the recurrence risk of the bleed, which is highest in untreated CAA and in poorly controlled hypertension.

Because the evidence does not dictate a single right answer, this is a textbook setting for shared decision-making. Record the infarct or hematoma characteristics, the relevant scores, the imaging you reviewed, the agent and intended start date, and the explicit conversation with the patient or surrogate about competing risks. Good documentation is both clinically protective and, in this YMYL space, a marker of careful practice.