Thrombectomy for large-core stroke: what the recent trials actually show
For years a large stroke core ruled out thrombectomy. A wave of randomized trials overturned that. Here is what the evidence shows, and what it does not.
By Manouchehr Hessabi, MD, MPH
For most of the last decade, a large established stroke was treated as a reason to do less, not more. If brain imaging showed that a big region of tissue had already died, the standard judgment was that opening the blocked artery would do little good and might cause harm. Between 2022 and 2024, a coordinated set of randomized trials tested that assumption head-on and overturned it. The result is a genuine shift in the evidence, and it is a useful case for seeing how a practice-changing finding is actually built, read, and bounded.
This is an explainer about method and evidence, in the same careful spirit as the broader stroke outcomes research it draws on. It is educational and not a substitute for personal medical advice. The aim is to separate what these trials established from what they did not.
The old rule, and why it existed
An ischemic stroke happens when a clot blocks an artery feeding the brain, starving downstream tissue of blood. Two regions form quickly. The infarct core is the tissue that has already died and cannot be recovered. The penumbra is the surrounding tissue that is starved but still alive, and still salvageable if blood flow returns soon enough. The entire logic of acute stroke treatment is a race to save the penumbra before it becomes core.
Mechanical thrombectomy is the procedure that removes the clot directly: a catheter is threaded up to the blockage and the clot is physically retrieved, restoring flow. By the late 2010s it was firmly established for patients with a small core and a large penumbra, the people with the most living tissue left to rescue.
A large established core was treated differently, and the caution was reasonable. If most of the at-risk tissue was already dead, the potential upside shrank. Worse, restoring blood flow to badly damaged tissue carries its own risks, including bleeding into the injured area. So a large core became, in practice, a reason to withhold the procedure. The fear was reperfusing tissue that was already lost and harming the patient in the process.
What the recent trials actually tested
The open question was specific and answerable: when the core is already large, does thrombectomy still help on balance, or does the harm win? Several independent groups asked it at once. The trials were RESCUE-Japan LIMIT, ANGEL-ASPECT, SELECT2, TENSION, and LASTE. Different countries, different teams, the same core question.
What makes their answer trustworthy is the design. Each was a randomized controlled trial, meaning eligible patients were assigned by chance to one of two paths: thrombectomy plus medical care, or medical care alone. Randomization is the workhorse of clinical evidence because, when the groups are large enough, it tends to balance not just the factors researchers can measure but the ones they cannot. If the thrombectomy group does better, randomization is what lets researchers attribute the difference to the procedure rather than to some hidden way the two groups differed from the start.
That design is also why these trials could answer a question that observation alone never could. Simply watching which large-core patients happened to get the procedure would mix the treatment effect with every reason a clinician chose to treat one patient and not another. Assigning the treatment by chance breaks that tangle.
How researchers measure stroke recovery
A trial is only as meaningful as the outcome it measures, and in stroke that outcome is rarely just survival. The standard instrument is the modified Rankin Scale, abbreviated mRS, a 7-point scale of disability that runs from 0 to 6. A score of 0 means no symptoms at all. A score of 6 means death. The numbers in between mark increasing dependence, from a slight disability that does not interfere with daily activities, up through needing help to walk or to attend to one's own bodily needs.
The reason this scale matters is that it captures the thing patients and families actually care about: function. A treatment that helps a person live independently rather than bedbound is a success even if it does not change whether they survive. Because of that, these trials looked at functional outcome at 90 days, not just at whether the artery reopened on a scan or whether the patient lived. A shift of even one point on the mRS, the difference between needing constant care and managing daily life with some help, is a large difference in a human life.
What the evidence shows, and how strong it is
Across these independent trials, the direction of the result was consistent: in patients with large ischemic cores, thrombectomy added to medical care improved functional outcomes compared with medical care alone. Major bodies have since folded large-core thrombectomy into their guidance, and the American Heart Association summarized the evidence in a 2024 science advisory on endovascular treatment for large-core stroke.
Convergence is what gives this weight. When several trials run by different teams in different health systems point the same way, it becomes much harder to explain the result as a quirk of one center, one population, or one definition of "large." A single positive trial invites the question of whether it would replicate. A cluster of them, asking the same question with the same basic design, largely answers it.
The follow-up data strengthened the case rather than softening it. Early benefit in stroke trials can fade, so the durability of an effect is a real question. Here it held. The 1-year outcomes of the SELECT2 trial, published in The Lancet in 2024, reported a sustained functional benefit of thrombectomy plus medical care over medical care alone at 12 months. The 12-month outcomes of TENSION, published in The Lancet Neurology in 2024, likewise extended the picture beyond the early window. The benefit looked like a durable shift, not a short-term blip.
What the trials do not settle
A practice-changing result is easy to overread, and honesty requires naming the limits. The most important one concerns who was actually enrolled. "Large core" is a range, not a single state, and the very largest cores were not well represented. Above roughly 150 milliliters of dead tissue, relatively few patients were enrolled across these trials, so the data thin out considerably at that extreme. The evidence is strongest in the middle of the large-core range and weakest at the edge, so the confident conclusion for a moderately large core does not automatically extend to the very largest.
Other limits matter too. These trials enrolled within specific time windows from symptom onset and selected patients using particular imaging criteria, so the findings speak to patients who resemble those who were studied, not necessarily to everyone with a large stroke. And every trial reports an average effect across a group. An average benefit tells you what tends to happen across many patients; it does not predict the result for any single person, whose age, other illnesses, and exact anatomy all bear on the outcome.
Reading a practice-changing result without overreading it
The clean way to hold this evidence is to keep two ideas in mind at once. The finding is real and durable: across multiple independent randomized trials, thrombectomy improved functional outcomes for many patients with large ischemic cores, and the benefit persisted at a year. And the finding is bounded: it is an average benefit in a defined population, thinner at the largest cores, framed by time and imaging criteria, and not a personal guarantee.
That pairing, a strong average effect alongside honest uncertainty about its edges, is what a mature body of evidence usually looks like. It is also why the language matters. These trials show that thrombectomy improved outcomes in the populations studied. They do not show that it helps every individual, and they were never designed to.
This article is educational and is not a substitute for personal medical advice. Decisions in acute stroke are time-critical and depend on details no general explainer can hold, and they belong to a patient and their stroke team. For readers who want to follow the evidence into the primary literature, the peer-reviewed work on stroke and outcomes research is the place to continue.