A Wake-Up Stroke and the Limits of Extending Thrombolysis Past 4.5 Hours
Perfusion imaging can make a strong case for treating a stroke of unknown onset time — but the newest trial testing that logic with tenecteplase specifically, rather than alteplase, didn't confirm it, and the case for extending the newer drug's own use into this window is weaker than it looks at first read.
M.O., a 66-year-old retired postal carrier, went to bed at his usual 10:00pm after watching the news, entirely normal by his wife's account — no complaints, no slurred words, nothing she noticed at all. He has hypertension and atrial fibrillation, for which he takes metoprolol and apixaban, and a remote fifteen-pack-year smoking history he quit twelve years ago. His wife woke at 6:40am and found him unable to sit up on his own, his left arm hanging useless, his face drooping on the left, and his speech reduced to a few slurred words repeated over and over. She called 911 immediately; EMS reached the house by 7:02am and had him at the emergency department by 7:31am. Because no one witnessed onset, his last known well is 10:00pm the night before — nine and a half hours earlier, well outside the standard 4.5-hour window by clock time alone.
His NIHSS on arrival is 11, and noncontrast CT shows no hemorrhage and no established infarct on visual inspection. Because his actual time of onset is unknown but likely recent — strokes tend to cluster in the early morning hours rather than uniformly across sleep — the team obtains MRI with DWI-FLAIR sequences rather than writing him off as untreatable by the clock. DWI shows a moderate-sized right MCA territory restriction; FLAIR shows no corresponding signal change yet, a mismatch pattern that the WAKE-UP trial used to identify patients whose actual stroke onset was very likely under 4.5 hours before the scan, despite an unwitnessed nine-hour window by clock time. Perfusion imaging adds a second layer: a region of delayed perfusion substantially larger than the diffusion-restricted core, the mismatch pattern EXTEND used to extend alteplase's benefit out to nine hours in appropriately selected patients. His apixaban is the reason he nearly wasn't a candidate at all. Any factor Xa inhibitor taken within the preceding 48 hours is a formal contraindication to thrombolysis in both the AHA/ASA and European Stroke Organisation guidance, and a normal PT or aPTT does not clear it — standard coagulation studies are insensitive to apixaban at exactly the levels that matter. What makes him eligible is a lapsed prescription: his wife confirms the bottle ran out and he last took a dose three days ago, and an apixaban-calibrated anti-factor Xa level drawn on arrival comes back below the threshold at which thrombolysis is withheld. The same lapse is the most plausible reason he is having a cardioembolic stroke in the first place. The imaging makes a genuine, trial-supported case for thrombolysis despite the unwitnessed onset — the live question the team is working through is which fibrinolytic that imaging case was actually built for.
In the MRI reading room, with a mismatch on the screen
Alteplase, using the WAKE-UP protocol he already qualifies for on imaging. Both WAKE-UP and EXTEND are positive trials, on their primary functional-outcome endpoint, for exactly this scenario — unknown onset, mismatch-selected. That's not a mechanistic argument, it's the actual outcome result this exact clinical situation has been tested against, twice, by two different groups.
I'm not against tenecteplase generally — in the standard window it's a reasonable, well-supported alternative. But "reasonable in one setting" doesn't automatically transfer to a different, higher-risk selection scenario just because it's the newer drug.
Worth naming what TIMELESS actually found, though, before we rule tenecteplase out here: it used essentially the same perfusion-imaging selection logic, out to 24 hours, and it showed a real reperfusion benefit over placebo. The imaging criteria are doing the work of identifying who has salvageable tissue regardless of which fibrinolytic dissolves the clot once you've found it.
If the selection logic is sound — and WAKE-UP and EXTEND both validate that it is — there's a real mechanistic argument that any effective fibrinolytic should benefit a properly-selected patient, not just the one two specific trials happened to test.
You're right that TIMELESS found a real reperfusion signal — I'm not disputing that finding. But TIMELESS's own PRIMARY endpoint was the 90-day functional-outcome shift, and that result was neutral, not positive. That's not a smaller version of WAKE-UP and EXTEND's result, it's a different result: a real physiological effect that didn't translate into the clinical outcome the trial was actually designed to detect.
"The selection logic is doing the work" is exactly the assumption TIMELESS was built to test for this specific drug, and it's the assumption the trial didn't confirm. For a man whose actual onset time we don't know and whose risk of symptomatic hemorrhage is already higher in an extended-window population, I'd rather use the drug with two positive outcome trials behind it than the one whose own dedicated trial came back neutral on the endpoint that matters.
Agreed: alteplase given per the WAKE-UP mismatch protocol, weight-based bolus and infusion, with apixaban held and no reversal needed given the 72-hour dosing interval and a calibrated anti-Xa level below the treatment threshold. He was admitted to the stroke unit for standard post-thrombolysis monitoring.
Not agreed: whether tenecteplase deserves a place in this specific extended-window, unknown-onset scenario as more mature data accumulate. The neuroradiologist still reads TIMELESS's reperfusion finding as a real, promising signal worth revisiting once a dedicated positive-outcome trial exists; the intensivist's position is narrower than a rejection of the drug — only that the evidence doesn't support it here yet, and the two are watching the same literature for different reasons.