Clot removal days after a stroke helps recovery but carries risk

Removing a blood clot from a blocked brain artery can still help some stroke patients one to three days after symptoms begin, a Chinese trial has found. But the benefit came at a price: more bleeding in the brain and more deaths.

In LATE-MT, patients who had thrombectomy between 24 and 72 hours after a large-vessel stroke were more likely to recover independence than those given medicines alone. About 49% could look after themselves at 90 days, against 35% without the procedure, according to Medical Xpress. Symptomatic brain bleeding occurred in 4.8% of treated patients and in none of the control group, and 6.0% died within 90 days, against 1.8%.

The results were published in the New England Journal of Medicine on 23 October 2026, with Pengfei Yang of Changhai Hospital, Naval Medical University, Shanghai, as first author and Jianmin Liu as senior author. They were first presented at the European Stroke Organisation Conference in Maastricht in May 2026.

Why time matters in stroke

Most strokes are ischaemic, caused by a clot blocking an artery that supplies the brain. When a large artery is blocked, such as the internal carotid or the middle cerebral artery, a big area of brain is at risk.

Endovascular thrombectomy is a procedure in which a doctor threads a thin catheter from an artery in the groin or wrist up to the brain and pulls out or sucks out the clot. In 2015, several trials showed it greatly improved recovery when done within six hours. In 2018, the DAWN and DEFUSE 3 trials extended that window to 24 hours for patients whose scans showed brain tissue that could still be saved.

Beyond 24 hours, doctors have had little evidence. Some people are "slow progressors", whose brain tissue survives for longer because of good collateral blood supply from other vessels. Observational studies suggested they might still benefit. Many patients also arrive late, for example after a stroke during sleep, in remote areas, or when symptoms are mild at first.

The LATE-MT trial

LATE-MT was an open-label randomised trial with blinded assessment of outcomes, run at specialist stroke centres in China. Medical Xpress reports 35 centres took part.

Patients had an ischaemic stroke from a blocked large artery in the front part of the brain circulation, the anterior circulation. They had to be 24 to 72 hours from symptom onset, or from when they were last known to be well, and brain imaging had to show salvageable tissue.

They were assigned to thrombectomy plus standard medical care, or standard care alone. According to Medical Xpress, standard care included antiplatelet drugs, treatment for brain swelling and routine hospital support. Doctors aimed to start the procedure within 60 minutes of randomisation.

The intention-to-treat population had 336 patients, 168 in each group. The median time from onset or last known well to randomisation was 38.5 hours. NeuroNews International reported 341 randomised patients at the ESOC presentation; the published analysis uses 336.

The main measure was the modified Rankin scale at 90 days. This 0 to 6 scale runs from no symptoms (0) through increasing disability to death (6). The trial compared the whole spread of scores between the groups.

Better recovery

Thrombectomy shifted the spread of Rankin scores towards better recovery. The common odds ratio for a better outcome was 1.74 (95% confidence interval 1.19 to 2.55), according to the NEJM abstract.

Put simply, at any point on the scale, patients who had the procedure had higher odds of being in a better category than those who did not.

According to Medical Xpress:

• About 49% of treated patients were functionally independent at 90 days, against 35% in the control group.
• 26.9% had no significant disability despite some symptoms, against 15.6%.
• Treated patients also reported better quality of life and daily functioning.

At ESOC, NeuroNews International reported improvements in the utility-weighted Rankin scale, the Barthel Index of daily activities and health-related quality of life. Infarct volumes at seven days were similar between groups. Benefits appeared larger in patients with blockages in the internal carotid artery.

More bleeding and more deaths

The safety results were clearly worse with thrombectomy.

At ESOC, adverse events of special interest, such as procedure complications, neurological worsening, brain swelling, re-blockage of the treated vessel, pneumonia and recurrent stroke, were reported in 15.5% against 8.9%.

A trade-off

The key question is whether the gain in independence outweighs the extra risk of death and bleeding. LATE-MT suggests it can, in carefully selected patients.

About 14 more patients out of every 100 recovered independence with thrombectomy, according to the Medical Xpress figures, while about four more died. The ordinal analysis, which weighs death as the worst outcome, still favoured thrombectomy.

Co-principal investigator Craig Anderson, of the George Institute for Global Health in Sydney, described the results at ESOC as a "trade-off" between preventing disability and introducing a "small risk" of further complications, NeuroNews International reported. Liu concluded that "appropriate targeting" of thrombectomy could be considered effective in this late window.

The NEJM abstract states it plainly: the procedure led to better functional outcomes but a higher risk of intracranial haemorrhage and death. An accompanying NEJM editorial is titled "Expanding the Evidence and Indications for Thrombectomy in Ischemic Stroke".

Who might benefit

The trial only included people whose scans showed salvageable brain tissue. That selection is probably the most important factor in the results. Patients whose damage is already complete are unlikely to gain anything and may only face the risks.

Imaging methods such as CT perfusion and MRI can estimate how much tissue is dead and how much is at risk. Stroke centres that might offer late thrombectomy need access to this imaging and to teams able to interpret it quickly.

Limits of the evidence

• The trial was open label. Patients and treating doctors knew who had the procedure, although outcomes were assessed blind.
• All centres were in China, where patient characteristics, causes of stroke and systems of care may differ from Europe.
• With 336 patients, estimates of death and bleeding have wide uncertainty.
• Follow-up was 90 days. Longer-term outcomes are not yet known.
• Only anterior circulation strokes were included.

What it means for patients and services

For people who reach hospital late after a major stroke, LATE-MT is the first randomised evidence that clot removal can still help. Guidelines currently focus on the first 24 hours. Guideline groups are likely to review LATE-MT, but a single trial with a clear safety signal may lead to cautious, case-by-case recommendations.

For the public, the message is unchanged. Stroke is an emergency. In Spain call 112, in the UK call 999. The sooner treatment starts, the better the chances, and the lower the risks.

This article explains published research and isn't medical advice. Talk to your own doctor or specialist team before changing any treatment.