Objective:

To perform a rare-event meta-analysis of symptomatic intracranial haemorrhage (ICH) from recent randomized controlled trials (RCTs) comparing early versus delayed direct oral anticoagulant (DOAC) initiation.

Background:

The most favourable timing of DOAC initiation remains uncertain after acute ischemic stroke in patients with atrial fibrillation. Although early anticoagulation may reduce recurrent ischemic events, concerns persist regarding the risk of symptomatic ICH. RCTs have reported low event rates, making interpretation difficult.

Design/Methods:

We systematically identified RCTs enrolling patients with ischemic stroke and atrial fibrillation that compared early versus delayed DOAC initiation and reported symptomatic ICH. Data were extracted from TIMING, ELAN, OPTIMAS, and START trials. The primary outcome was symptomatic ICH within 30–90 days. Pooled odds ratios (ORs) with 95% confidence intervals (CIs) were estimated using Mantel–Haenszel fixed-effects with continuity correction, DerSimonian–Laird random-effects, and Peto’s method. Sensitivity analyses excluded double-zero trials and compared fixed- versus random-effects models.

Results:

Across four RCTs, 5,441 patients were included (early, 2,691; delayed, 2,750). Symptomatic ICH occurred in 12 patients in the early group (0.4%) versus 13 in the delayed group (0.4%). Pooled analyses demonstrated no difference in risk: fixed-effect OR 0.98 (95% CI, 0.44–2.17); random-effects OR 0.98 (95% CI, 0.44–2.18); Peto OR 0.99 (95% CI, 0.45–2.20). Heterogeneity was absent (I² = 0%). Sensitivity analyses confirmed robustness of findings.

Conclusions:

Symptomatic ICH following DOAC initiation after ischemic stroke is exceedingly rare, with no difference between early and delayed strategies. Rare-event modelling supports the safety of early anticoagulation, which may help guide clinical decision-making and future guideline development.