More than half of patients with wake-up stroke have slept through their stroke onset

Conventional time-from-onset decision-making for use of reperfusion therapies in acute ischemic stroke is challenged by patients with wake-up strokes (WUS) in whom onset time is uncertain. Management of these patients, including use of advanced imaging to make "tissue-clock" rather than "time-clock" treatment decisions, and development of wearable technologies to detect stroke onset during sleep, would be aided by delineation of how often WUS patients had onset long before or shortly before awakening.

We performed a meta-analysis of studies comparing diffusion weighted imaging-fluid-attenuated inversion recovery (DWI-FLAIR) mismatch frequencies in patients with WUS compared with patients with clear onset time stroke (COS). Differences in FLAIR-DWI mismatch frequency between WUS and COS patients in matched time intervals from first symptom detection were analyzed in a novel manner to derive the timepoints during sleep of stroke onset in WUS patients.

Two studies of WUS and COS patients met selection criteria, collectively enrolling 443 patients (201 WUS, 242 COS). For COS patients, mismatch frequency declined gradually as time from first symptom detection increased: <2 h 60 %; 2-3 h 53 %; 3-4 h 41 %; 4-5 h 37 %; 5-6 h 11 %. In contrast, for WUS patients, <2 h patients had similar mismatch rates to <2 h COS patients but mismatch then declined precipitously in all later time windows: <2 h 50 %; 2-3h13 %; 3-4 h 13 %; 4-5 h 6 %; 5-6 h 12 %. The lower mismatch rate indicated that 56.5 % of WUS patients had stroke onset long (2-8 h) before awakening and 43.5 % had stroke onset shortly (0-2 h) before awakening.

A majority of wake-up stroke patients stayed asleep for 2-8 hours after stroke onset, while a minority awoke at the time of or soon after onset. These findings can inform multimodal imaging strategies and wireless stroke detection technology development programs for patients with wake-up stroke.