Early Versus Delayed Anticoagulation in Acute Ischaemic Stroke with Atrial Fibrillation According to Infarct Volume and Location: a Prespecified Subgroup Analysis of the OPTIMAS Randomised Controlled Trial.
Nash PS., Best J., Lyon J., Ruffle JK., Amornpojnimman T., Mendel R., Foulon C., Dehbi H-M., Ahmed N., Arram L., Balogun M., Bennett K., Bordea E., Caverly E., Chau M., Cohen H., Cullen M., Doré CJ., Engelter S., Fenner R., Ford GA., Gill A., Hunter R., James MA., Jayanthi A., Lip GYH., Massingham S., Murray ML., Mazurczak I., Ndoutoumou A., Norrving B., Philip J., Sims H., Sprigg N., Vanniyasingam T., Nachev P., Freemantle N., Doig D., Werring D.
BackgroundRandomised trials have demonstrated that early anticoagulation after acute atrial fibrillation-associated ischaemic stroke is safe and non-inferior to delayed initiation. Whether anticoagulation should be delayed in people with larger infarcts is uncertain.AimsTo investigate whether ischaemic stroke infarct volume, measured precisely by segmentation, modifies the treatment effect of early anticoagulation with a direct oral anticoagulant (DOAC).MethodsWe did a prespecified secondary analysis of OPTIMAS (NCT: 03759938), a randomised, parallel group, open-label trial with blinded outcome assessment which randomised people with acute ischaemic stroke and atrial fibrillation to early initiation of any licensed DOAC, within four days of onset, or delayed initiation 7-14 days from onset. The primary outcome was a composite of recurrent ischaemic stroke, symptomatic intracranial haemorrhage (ICH) and systemic arterial embolism within 90 days. A central neuroimaging laboratory determined infarct volume using diffusion-weighted MRI using a validated deep learning segmentation model; on CT, infarcts were segmented manually. We modelled infarct volume as a continuous variable using restricted cubic splines, and tested for an interaction with treatment allocation in mixed effects logistic regression.ResultsWe included 3572 participants (mean age 78±10 years, 45% female), 98·6% of the main trial population. The effect of early versus delayed anticoagulation did not vary with infarct volume (pinteraction=0.18). Rates of the primary outcome were 17/568 (3.0%) and 12/599 (2.0%) for early vs. delayed initiation with infarcts of 0-5mls; 6/220 (2.7%) and 11/229 (4.8%) with infarcts of 5-10mls; 13/258 (4.6%) and 10/283 (3.5%) with infarcts of 10-25mls; 6/145 (4.1%) and 8/145 (5.5%) with infarcts of 25-50mls; 1/93 (1.1%) and 7/94 (7.4%) with infarcts of >50mls; and 14/481 (2.9%) and 10/430 (2.2%) in participants with no infarct visible on clinically-acquired brain imaging. Corresponding odds ratios and 95% confidence intervals were 1·52 (0·71-3.20), 0·55 (0·20-1·51), 1·29 (0·55-3·00), 0·74 (0·25-2·21), 0.13 (0·02-1·11), and 1.25 (0.55-2.86) respectively. There were no increased rates of symptomatic ICH with respect to anticoagulation timing for those with very large infarcts (>25mls); there were 3/238 (1.3%) events in the early group and 5/239 (2.1%) in the delayed group.ConclusionsThe treatment effect of early anticoagulation with a DOAC in acute ischaemic stroke associated with atrial fibrillation was not modified by infarct volume. Adverse outcomes were not increased with early anticoagulation in people with larger infarcts. Our results provide no evidence that anticoagulation initiation should be delayed beyond four days on the basis of infarct size.