Abstract Number: Esoc2026lt38 Lt38 - Avert Dose: A Phase Iii, Multi-Arm, Adaptive Randomised Trial to Determine Optimal Early (48 Hr Start) Mobility Training in Patients With Mild & Moderate Ischaemic Stroke

Abstract Background and aims Mobility-focused rehabilitation is a relatively simple, broadly applicable intervention, that typically starts in the acute stroke phase. Following the therapeutic uncertainty created by the findings of the AVERT trial, the AVERT-DOSE trial aimed to determine the optimal safe and effective therapy regimens in patients with mild and moderate ischaemic stroke. Methods We designed a Multi-Arm Multi-Stage Covariate-Adjusted Response-Adaptive randomised trial of four different regimens of mobility training commenced 48 hours of stroke in each of two strata: mild (NIHSS7) and moderate (NIHSS 8-16) ischaemic stroke. Analysis of blinded outcomes is at 3 (primary) and 6 months. Eligibility criteria were broad, while excluding premorbid disability (mRS 2). All mobility training (functional, task-specific, upright) was protocolised across four dose arms (with prespecified reference arms) for each stratum, and was provided by trained physiotherapists and nurses. Favourable outcome (mRS 0-2) at 3 months was the primary outcome. Results Trial stopped early due to funding constraints. Final participants were (total n=1000: mild n=637, moderate n=363) recruited November 2025, data lock with 3-month outcomes expected end of February 2026. Fifty hospitals in seven countries (Australia, Brazil, India, Ireland, Malaysia, Singapore, United Kingdom) participated. Outcomes will be presented separately for mild and moderate strata using re-randomization tests given the adaptive trial design, together with frequentist confidence of the tested regimens being clinically meaningfully better than the reference regimen. Conclusions While closed early, our adaptive trial will allow an efficient evaluation of the safety and effectiveness of early mobility interventions for patients with mild and moderate ischaemic stroke. Conflict of interest