The Gale readResearch

A Stroke Drug Beat Placebo, 69.7% to 56.3%; An Earlier Trial Did Not

What the LAIS trial measured, the earlier trial of the same drug that found no significant difference, and the questions its published abstract leaves open.

By Gale Staff · November 21, 2026 · JAMA

Not reviewed by a clinician.

Illustration: Heart, diabetes & metabolism

The short answer

A large trial in China reported that an investigational intravenous drug, loberamisal, left more stroke patients with little or no disability three months later: 69.7% reached the best two grades on the standard disability scale, against 56.3% given a placebo, a difference of 13.28 percentage points. That rests on 997 patients analysed at 32 hospitals, all in China, treated within 48 hours of their symptoms starting. The caveat is the trial before it, which tested the same drug on the same measure and found no statistically significant difference, and whose authors wrote that the drug's benefit and its best dose still needed proper testing.

Who to see

Primary care clinician

A primary care clinician is the usual first stop for a question like this one.

Find care →

A headline read in a hospital corridor

Someone is sitting outside a stroke unit with a phone in their hand. A parent was brought in this morning, the doctors have been careful and unspecific, and the search that gets typed is the one anybody would type: is there a new drug for stroke. What comes back is a number big enough to reorganise an afternoon's hopes.

The number is real, and it is worth knowing precisely, because this drug has been through two trials that pointed different ways. The one that made the news is the second.

What the trial did

The trial, named LAIS and published in JAMA in 2026, was what researchers call a phase 3 trial: the large, late-stage test meant to settle whether a treatment works. It was run at 32 hospitals in China, and it was double-blind and placebo-controlled, meaning neither the patients nor their doctors knew who was getting the drug and the comparison group got an inactive infusion. Enrolment ran from July to December 2024, with the last follow-up in April 2025.

Adults aged 18 to 80 could join if they arrived within 48 hours of their symptoms starting, scored 7 to 20 on the standard stroke severity scale, and had no disability before the stroke. They were assigned, half and half, to 40 mg of loberamisal into a vein once a day for ten days or to a matching placebo, on top of ordinary stroke care. The drug is described in the report as a small molecule that acts on a pathway involving a protein called postsynaptic density 95, the idea being to protect brain tissue rather than to clear the clot.

Of 998 patients assigned, 997 got at least one dose and were counted in the main analysis, 502 on the drug and 495 on placebo. Their median age was 64, a third were women, and the median severity score was 8, which puts this group in the middle of the range rather than the worst of it.

What it found, and what the abstract leaves out

Ninety days on, 350 patients given the drug, 69.7%, scored 0 or 1 on the modified Rankin Scale, the measure clinicians use for disability after a stroke, where 0 means no symptoms at all and 1 means no meaningful limit on daily life. On placebo it was 279 patients, 56.3%. Expressed as a ratio that is 1.24, with a range of 1.12 to 1.36 either side of it, and as a plain gap it is 13.28 percentage points, with a range of 7.24 to 19.32.

Harms looked no worse on the drug. Side effects of some kind were reported in 87.8% of the drug group and 88.7% of the placebo group, serious ones in 8.6% against 10.7%, and deaths in 1.2% against 2.0%.

Two things the published summary does not contain are worth naming, because a reader cannot weigh what is not shown. There is no p-value for the main result, only the statement that the difference was statistically significant. And there is no limitations section; the one forward-looking sentence is that further studies are needed to confirm the findings. The trial's own protocol, published separately, had also listed a second measure of daily function at 90 days among its secondary outcomes, and no result for it appears in the abstract.

The earlier trial of the same drug found nothing significant

Before this, the same drug was tested in a phase 2 trial, the middle-stage study that looks for a signal and a dose. Published in Stroke and Vascular Neurology, it randomised 240 patients in 2023 to one of three doses, 20, 40 or 60 mg, or to placebo. Its main measure was identical to the later trial's: the share of patients reaching those same best two disability grades at 90 days.

The results came out at 76.7% on the middle dose, 70.0% on the high dose, 67.8% on the low dose and 60.7% on placebo, and the p-value across those four groups was 0.164. A p-value is the probability of seeing a spread this large if the drug did nothing, and 0.164 is far above the 0.05 that researchers conventionally treat as significant. So that trial did not show the drug worked. Its authors concluded it was well tolerated, and wrote that its efficacy and optimal dosage needed prospective validation, meaning a further trial built to test them.

This does not make the later result wrong, and it is important not to read it that way. A phase 2 trial of 240 people split four ways is not built to prove a benefit; spotting something a bigger trial can confirm is its purpose, and the phase 3 then used 40 mg, the dose that had performed best. What it does mean is that the headline result is the first statistically significant finding for this drug, not the confirmation of an earlier one.

What the whole published record amounts to

A search of the European PMC literature index for the drug's name returns six records, four of them about it: the phase 2 result, the phase 3 result, the phase 3 protocol, and a journal summary of the trial. Both trials were conducted in China, and no replication in another country or health system appears in that record.

That is a thin base for a treatment given into a vein for ten days inside the 48-hour window after a stroke, and thin in a particular way: one positive late-stage trial in a single country, with its own predecessor having missed significance on the very same measure. Whether the effect shows up again elsewhere is the open question, and it is the question the trial's own authors point at.

What this study can't tell you

  • Whether the drug can be obtained, or has been approved, anywhere. Nothing in the trial reports read here describes its regulatory status or availability in any country, so no claim about either can be made from this evidence.
  • How it fits with clot removal. The published phase 3 abstract does not mention mechanical thrombectomy or any endovascular treatment, as either a requirement or a bar to joining, so whether those patients were included cannot be established from it.
  • Whether the benefit appears outside China. Both trials were run there, and the published record contains no independent replication in another population or health system.
  • What happened to the second function measure. The protocol lists a Barthel Index of 95 or above at 90 days among the secondary outcomes, and no result for it appears in the published abstract.
  • Exactly how strong the main result is. The abstract gives a ratio and a percentage-point gap with ranges, but no p-value, reporting only that the difference reached statistical significance.
  • Whether people with milder or far worse strokes would see the same thing. Joining required a severity score of 7 to 20 and no disability beforehand, and the median score was 8, so the trial describes a middle-severity group.
  • Why the earlier trial's numbers do not add up cleanly. It reports 240 patients randomised and 224 who received treatment, while its four outcome groups sum to 240, and the abstract does not explain the difference.

The Gale read

This is a good result and it is one result, and the space between those two statements is where a family in a corridor can be led astray. A 13-percentage-point difference in how many people reach three months with little or no disability is not a technicality; if it holds up it matters enormously. The reason for care is not that the trial looks weak, because it does not, and it is not that the researchers overclaimed, because the paper's own closing line asks for more work. It is that the middle-stage trial of this same drug, asking the same question on the same measure, found no significant difference, and that both trials were run in one country with nothing yet repeating them elsewhere. The published summary also withholds things a careful reader would want: no p-value, no limitations section, and no result for a secondary measure its own protocol specified. The accurate description of loberamisal today is an investigational drug with one strong late-stage trial behind it and the confirmation its own authors asked for still outstanding. That is a genuinely hopeful position, and it is not the same thing as a treatment anyone can ask a doctor for this week.

Common questions

Is there a new drug for stroke recovery?

One investigational drug given into a vein, loberamisal, has a positive late-stage trial published in JAMA in 2026. Given within 48 hours of symptoms starting, at 40 mg daily for ten days, it left 69.7% of patients with little or no disability at 90 days against 56.3% on placebo, across 997 patients analysed at 32 hospitals in China. Nothing in that report describes whether it is approved or available in any country, and an earlier trial of the same drug found no significant difference on the same measure.

What did the loberamisal stroke trial show?

At 90 days, 69.7% of patients on the drug scored 0 or 1 on the modified Rankin Scale, meaning no symptoms or no meaningful limit on daily life, against 56.3% on placebo: a ratio of 1.24 and a gap of 13.28 percentage points. Serious side effects were 8.6% against 10.7% and deaths 1.2% against 2.0%, with overall side effects almost identical in both groups. The published abstract gives no p-value and has no limitations section.

Is this a new stroke treatment for 2026?

The trial was published in 2026, and it is the first statistically significant result for this drug rather than a confirmation of an earlier one. The middle-stage trial, published in Stroke and Vascular Neurology, randomised 240 patients to three doses or placebo and found 76.7%, 70.0%, 67.8% and 60.7% reaching the best two disability grades, with a p-value of 0.164 across the four groups, which does not meet the conventional threshold for significance. Its authors wrote that the drug's benefit and best dose still needed proper testing.

Sources

  1. 1.Li S, Feng B, He D, et al; LAIS investigators. Loberamisal for Acute Ischemic Stroke: The LAIS Randomized Clinical Trial. JAMA. 2026;336(18):1612-1620. doi:10.1001/jama.2026.16557 link
  2. 2.Feng B, Li H, Xu S, Li J, Wang Y, Zhao X, Wei Y, Xiao X, Wang S, Wang Y, Li S. Loberamisal injection for the treatment of acute ischaemic stroke: a multicentre, randomised, double-blind, placebo-controlled phase II clinical trial. Stroke and Vascular Neurology. 2026;11(4):465-472. doi:10.1136/svn-2025-004581 link
  3. 3.Loberamisal for Acute Ischaemic Stroke (LAIS): a multicentre, randomised, double-blind, parallel, placebo-controlled phase III clinical trial. Stroke and Vascular Neurology. 2026;11(4):504-509. doi:10.1136/svn-2025-004582 — the trial protocol, which lists the secondary outcomes and the registration NCT06517173. link

Read the study →

3 sources, numbered by first appearance. General health information, not medical advice. AI-assisted editorial content — every citation independently verified. Editorial policy

More from Gale Staff

The Latest Health Research →

Published November 21, 2026 · Gale Staff

All Gale Staff analyses →