The expanded window for mechanical thrombectomy reaches up to 24 hours from the time a patient was last known well — but only for a narrow group. Patients with a large vessel blockage in the front of the brain and evidence of salvageable tissue on advanced imaging can be treated between 6 and 24 hours; outside that imaging-confirmed group, the practical cutoff is 6 hours. The clock starts at "last known well," not at symptom discovery, and the eligibility expansion rests on two 2018 trials, DAWN and DEFUSE 3, not on a blanket rule for every stroke.
Below, the number, the criteria behind it, and the cases where the 24-hour figure misleads.
The 24-Hour Ceiling and the Two Trials That Set It
The 24-hour figure comes from two randomized trials published in 2018 in the New England Journal of Medicine. DAWN enrolled patients 6 to 24 hours after last known well who had a mismatch between clinical severity and infarct size on imaging. DEFUSE 3 enrolled patients 6 to 16 hours after last known well using perfusion imaging criteria. Both showed benefit for thrombectomy over medical management alone, and both are the named basis for late-window eligibility.
The regulatory piece matters too. In 2018 the U.S. Food and Drug Administration expanded the indication for certain stent retriever devices to 24 hours, following the trial evidence. That cleared the pathway for late-window treatment in routine practice, though the device labels and the trial criteria are not identical — a distinction that catches clinicians out.
The time is measured from last known well, meaning the last moment the patient was seen acting normally. A patient found at 7 a.m. who was last seen well at 11 p.m. the night before is at 8 hours, not 0. Wake-up strokes default to the last-seen-well time, which is why imaging does so much work in this window.
What the trials did not do is open the door to everyone at 24 hours. DAWN's imaging criteria were specific — a small core relative to a large clinical deficit — and DEFUSE 3 required a target mismatch profile on CT perfusion or MR perfusion. A large completed infarct at 20 hours fails those criteria regardless of the clock.
| Window | Trial basis | Imaging requirement | Typical candidate |
|---|---|---|---|
| 0–6 hours | Multiple trials since 2015 (HERMES pooled analysis) | Non-contrast CT to exclude hemorrhage; CTA to confirm large vessel occlusion | Most large vessel occlusion strokes |
| 6–16 hours | DEFUSE 3 (New England Journal of Medicine, 2018) | Perfusion imaging target mismatch | Last known well 6–16 hours, salvageable tissue |
| 6–24 hours | DAWN (New England Journal of Medicine, 2018) | Clinical–imaging mismatch (small core, significant deficit) | Last known well 6–24 hours with favorable profile |
The table's three rows are not interchangeable. A patient at 14 hours may qualify under either late trial's criteria depending on which profile their imaging shows, and centers differ in which protocol they apply. Confirm the local protocol — that is the controlling document, not the trial paper.
Why "Expanded" Is the Wrong Word for Most Stroke Patients
The expansion is real but bounded. Most thrombectomy patients are still treated inside 6 hours, because most large vessel occlusions present early and because the 6-hour group has the strongest evidence base, going back to the 2015 trials pooled in the HERMES analysis. Roughly a quarter to a third of late-presenting patients screen into the extended window after imaging, depending on the series — the exact proportion varies by center and population, so treat the figure as a planning estimate rather than a fixed rate.
Two conditions shrink the group further. First, the occlusion has to be in a proximal anterior circulation vessel — the internal carotid artery or the first segment of the middle cerebral artery, with some trials including the second segment. Second, the imaging has to show tissue worth saving. Posterior circulation strokes have their own evidence base and are not covered by the DAWN and DEFUSE 3 criteria.
There is also a geography problem. Late-window thrombectomy requires CT perfusion or MR perfusion, and not every hospital runs those studies around the clock. A patient who presents at 10 hours to a center without perfusion imaging may be transferred or may miss the window entirely. The 24-hour ceiling is a capability statement, not a universal entitlement.
The number to hold onto: 24 hours is the outer limit for a selected subset, and 6 hours is where most treated patients sit. Anyone quoting 24 hours as the standard window for all strokes has compressed a conditional rule into a slogan.
The Imaging Criteria That Decide Late-Window Cases
DAWN's mismatch criterion was defined by infarct volume against clinical deficit, measured with the NIH Stroke Scale. Patients with a small core and a substantial deficit — the trial used thresholds combining core volume under 21 mL with a score of 10 or more, among other strata — qualified between 6 and 24 hours. DEFUSE 3 required a core under 70 mL, a mismatch ratio of at least 1.8, and a mismatch volume of at least 15 mL, with enrollment capped at 16 hours.
Those exact numbers belong to trial protocols, not to every hospital's order set. Many centers apply modified criteria, and some use CT perfusion thresholds that differ from the trial definitions. A patient who fails one center's screen may pass another's. This is one reason transfer decisions in late-window stroke are genuinely difficult, and why the receiving center's neurointerventionalist should be involved before imaging is repeated.
Non-contrast CT alone cannot support late-window selection. It rules out hemorrhage and gives a rough sense of established infarct, but the mismatch logic needs perfusion data or MRI. Advanced imaging adds time — typically 10 to 20 minutes on the table, plus interpretation — which is a real cost inside a window measured in hours.
If a center cannot obtain perfusion imaging promptly, the safer path is early transfer to a center that can, not a guess based on non-contrast CT. That is the failure case worth naming: a patient at 9 hours with a favorable-looking non-contrast CT and no perfusion study, treated on optimism rather than criteria, has not received evidence-based late-window care.
How the Clock Is Measured, and the Wake-Up Stroke Problem
Last known well is the anchor. For a witnessed onset, it is the moment symptoms appeared. For an unwitnessed or wake-up stroke, it is the last time the patient was observed normal — usually bedtime. A patient who goes to sleep at 10:30 p.m. and wakes at 6:30 a.m. with a deficit is at 8 hours, and if they arrive at hospital by 7:30 a.m., roughly 9 hours, they sit inside the extended window.
This measurement choice has a consequence people miss: the extended window is partly a gift to unwitnessed strokes. Patients who wake with symptoms have no earlier anchor, so the last-seen-well rule pushes them into the 6-to-24-hour band where imaging decides. That is precisely the population DAWN and DEFUSE 3 were designed to serve.
Documentation matters clinically and legally. If the last known well time is uncertain, the earlier estimate governs eligibility decisions, and clinicians should record the uncertainty explicitly. A vague "symptoms noted in the morning" can cost a patient a treatment option.
If the last known well cannot be established at all, the case falls back on imaging and local protocol. Do not assume a generous window; assume the earliest plausible time and let the imaging and the neurointerventionalist make the call.
How the Window Shifts When Imaging Is Unavailable
A hospital without 24/7 perfusion imaging operates on a shorter effective window. Non-contrast CT and CT angiography can confirm a large vessel occlusion, but without perfusion data the late-window selection logic collapses. Many such centers treat to 6 hours and transfer beyond that, and the transfer itself consumes time.
The math is unforgiving. A 45-minute ground transfer plus a 30-minute re-imaging workup plus team assembly can add 90 minutes or more. A patient at 5 hours at the spoke hospital arrives at the hub at 6.5 hours — still eligible, but with less salvageable tissue than the first scan suggested. The window is not static; it is consumed by logistics.
Telestroke changes the arithmetic. A remote neurologist can review the spoke hospital's CT angiography, confirm the occlusion, and start the transfer before the patient leaves the scanner. Programs that run this way compress door-in-door-out times substantially, and the published targets for transferred thrombectomy patients sit well under two hours at high-performing centers.
The practical rule for patients and families: if a suspected stroke occurs far from a comprehensive stroke center, call emergency services and say when the person was last known well. That single fact drives every downstream decision, including whether a helicopter is worth launching.
Thrombectomy vs. Intravenous Thrombolysis Inside and Outside the Window
The two treatments are not substitutes, and their windows differ. Intravenous thrombolysis with alteplase carries a standard window of 4.5 hours from last known well, with selected patients treated beyond that under specific imaging criteria in some jurisdictions — check local protocol, because practice varies by country. Tenecteplase has become an alternative in many centers. Thrombectomy's 24-hour outer limit applies only to the mechanical procedure.
| Treatment | Standard window | Extension basis | Imaging needed for extension |
|---|---|---|---|
| Intravenous thrombolysis (alteplase) | 0–4.5 hours | Selected cases beyond 4.5 hours under local protocol | Perfusion or MRI in some protocols |
| Intravenous tenecteplase | 0–4.5 hours (center-dependent) | Center protocol | Center protocol |
| Mechanical thrombectomy | 0–6 hours standard | 6–24 hours per DAWN and DEFUSE 3 | CT perfusion or MR perfusion |
A late-presenting patient at 12 hours is generally outside the thrombolysis window but may still qualify for thrombectomy. A patient at 3 hours may receive both — thrombolysis first, then thrombectomy if a large vessel occlusion is confirmed. The bridging decision belongs to the treating team and depends on contraindications, not on the clock alone.
One misconception to kill: the extended thrombectomy window does not extend thrombolysis. They are separate decisions with separate evidence, and conflating them is a common source of confusion in family conversations at the bedside.
What the Numbers Look Like in Real Cohorts
The HERMES pooled analysis of five trials, published in The Lancet in 2016, established the 0-to-6-hour benefit and reported a number needed to treat that clinicians still quote in consent conversations. The late-window trials reported their own effect sizes, and the pattern is consistent: benefit shrinks as time passes and as core volume grows, but it does not vanish at 16 or 24 hours in well-selected patients.
A useful way to think about it: the treatment effect is a function of salvageable tissue, and time is one proxy for how much tissue remains. A 70-year-old at 18 hours with a 10 mL core and a large deficit may have more to gain than a 50-year-old at 5 hours with a completed infarct. Age and clock both matter, but neither decides alone.
Cost and resource use follow the same logic. Thrombectomy is a high-cost procedure — device, angio suite, interventional team, intensive care bed — and the late-window group consumes more imaging before the decision. Health economic analyses of the late-window trials generally found the procedure cost-effective in selected patients, though the figures are jurisdiction-specific and should be checked against local reimbursement rules rather than imported from a paper.
What the cohort data cannot tell you is what will happen to a specific patient. Outcomes in the trials are averages across selected populations; individual prognosis depends on collateral circulation, time to reperfusion, and complications, none of which is fully predictable at the door.
Where the 24-Hour Figure Breaks Down
Three failure scenarios deserve naming. First, the large-core patient: a completed infarct involving most of the middle cerebral artery territory at 14 hours will not benefit, and the procedure carries risk without offsetting gain. Core volume thresholds in the trials were not arbitrary. Second, the posterior circulation stroke: basilar artery occlusion has its own literature and its own time logic, and the anterior circulation criteria do not transfer. Third, the patient with a contraindication to contrast or to the procedure itself, where the imaging gate cannot even be passed.
There is also a documentation failure mode. If last known well is recorded as "unknown" when it was actually obtainable from a family member, the patient may be screened out of the late window on a clerical gap. Emergency departments that ask the last-seen-well question explicitly, and record the answer with a time, catch cases that would otherwise default to the conservative path.
And a systems failure: the 24-hour window is only meaningful if the receiving center can act on it. A hospital that has the imaging but not the neurointerventional coverage at 2 a.m. has an effective window shorter than its protocol states.
For anyone advising a family member, the questions to ask are concrete: What was the last known well time? Has CT angiography confirmed a large vessel occlusion? Has perfusion imaging been done or is it available? Is this a comprehensive stroke center with thrombectomy coverage now? Those four answers determine more than any general statement about 24 hours.
FAQ
Does the 24-hour window apply to all stroke types?
No. It applies to anterior circulation large vessel occlusions with favorable imaging, per DAWN and DEFUSE 3. Posterior circulation strokes, small vessel strokes, and hemorrhagic strokes follow different pathways. A patient with a basilar occlusion may still be treated late, but on different evidence, and the treating team makes that call.
Is 24 hours measured from when symptoms started or when the patient was found?
From last known well — the last moment the patient was observed normal. For wake-up strokes, that is usually bedtime, not waking time. This is why a patient who wakes with symptoms at 6 a.m. after being seen well at 11 p.m. is at 7 hours, comfortably inside the extended window.
Can a patient get thrombectomy at 20 hours if the CT scan looks normal?
A normal non-contrast CT is not enough to qualify. Late-window selection requires perfusion imaging or MRI showing salvageable tissue, plus confirmation of a large vessel occlusion on CT angiography. A "normal" plain CT in a patient with a large deficit may actually signal a small core, but the perfusion study is what documents it.
What if the hospital doesn't have perfusion imaging?
The effective window at that facility is usually 6 hours, with transfer for later cases. Ask whether the center can perform CT perfusion and whether it has thrombectomy coverage overnight. If not, early transfer to a comprehensive stroke center is the priority, and the last known well time should be communicated to the receiving team.
How much time does the imaging add?
CT perfusion typically adds 10 to 20 minutes on the table plus interpretation time, and MRI adds more. In a late-window case that time is usually worth spending because it determines eligibility. In an early-window case, imaging should not delay thrombolysis or transfer.
Does the expanded window also extend the clot-busting drug window?
No. Intravenous thrombolysis generally remains a 4.5-hour treatment, with selected exceptions under local protocol. The 24-hour figure belongs to mechanical thrombectomy only. Families sometimes arrive expecting "the 24-hour drug" and need that corrected early.
What is the single most important fact to give the emergency team?
The last known well time, stated as a clock time with the source — "she was fine on the phone at 10:15 p.m." That one detail determines which windows apply, which imaging is ordered, and whether a transfer is worth starting. Vague timing forces the conservative path.
This article is general educational information about stroke treatment timelines and is not a diagnosis, treatment plan, or medical advice. Eligibility for thrombectomy depends on individual imaging, clinical findings, and local protocols, and only a qualified stroke team can make that determination. If you suspect a stroke, call your local emergency number immediately; for anyone in crisis or at risk of self-harm, contact local emergency or crisis services without delay.