This week in Neurology
Nov 18, 2026
Early middle meningeal artery embolisation plus standard treatment versus standard treatment alone for chronic subdural haematoma in France (OTEMACS): a randomised, open-label, controlled trial.
Early adjunctive middle meningeal artery embolisation cut the 90-day composite of reoperation, residual haematoma or death without more deaths, but the early-stopped trial is not confirmatory.
The Lancet Neurology · 2026 · PubMed
This week’s highlights
Read the weekly briefing
This AudioScholar briefing is generated by artificial intelligence from PubMed abstracts for healthcare professionals and is not medical advice.
This week: early embolisation after subdural haematoma surgery cuts reoperation, an oral edaravone trial in ALS comes up empty, and the new McDonald criteria trade some specificity for earlier diagnosis after optic neuritis.
Welcome to This Week in Neurology. We're covering 10 papers this week: 3 in depth, 4 in brief, 3 quick signals.
From The Lancet Neurology, Ng and colleagues report OTEMACS, a multicentre, open-label randomised trial at six hospitals in France [1]. It tested early middle meningeal artery embolisation with the Onyx liquid embolic system, given within five days of surgery when surgery was done, added to standard treatment for symptomatic chronic subdural haematoma. Of 247 patients randomised, 214 entered the modified intention-to-treat analysis, and outcome assessors were masked. The primary composite was reoperation or surgical rescue, residual haematoma thicker than ten millimetres, or death at 90 days. It occurred in about a third of patients in the embolisation group and about half of controls, an absolute reduction of roughly 16 percentage points. Reoperation fell from about 18 percent to about 7 percent. Deaths at 90 days were similar, five versus four, procedure-related serious adverse events were similar, and there were no embolisation-related deaths or ischaemic strokes. The main limitation is that enrolment stopped early on the monitoring board's advice without crossing the prespecified efficacy boundary, and patients without an assessable outcome were excluded. The authors therefore call the result non-confirmatory and the benefit possibly overestimated. It adds randomised evidence for early postoperative embolisation that the authors say needs confirmation before routine adoption.
Also in The Lancet Neurology, van Eijk and colleagues report ADORE, a double-blind phase 3 trial of FNP122, a once-daily oral formulation of edaravone at 100 milligrams a day, across 37 hospitals in twelve European countries [2]. Adults with ALS within two years of symptom onset, a vital capacity of at least 70 percent and a defined rate of functional decline were randomised two to one to edaravone or placebo. After nine patients were excluded from a site that breached Good Clinical Practice, 302 patients were treated and analysed. At 48 weeks, decline on the ALS Functional Rating Scale Revised was about 12.6 points with edaravone and 13.4 points with placebo, a difference of under one point that was not statistically significant. Adverse event rates were similar, and no deaths were attributed to treatment. The main limitation for generalising is that one dose was tested in a population that was more than nine in ten White. The authors conclude the results do not support this oral dose for modifying progression in the general ALS population. It adds a large, double-blind negative trial over a longer period and in a broader population than the earlier edaravone evidence.
Klyscz and colleagues, in The Lancet Neurology, applied the 2024 McDonald criteria to the Acute Optic Neuritis Network, a prospective cohort from 25 centres in 17 countries [3]. This cross-sectional analysis covered 310 patients with a first episode of optic neuritis. The 2024 criteria add the optic nerve as a fifth location for dissemination in space. Every patient with multiple sclerosis under the 2017 criteria still qualified, and about a third of patients with clinically isolated syndrome under the 2017 criteria now met criteria for multiple sclerosis. Sensitivity rose by about sixteen points, to roughly two thirds. Specificity fell from complete to about 89 percent: three of 22 patients with aquaporin-4 antibody-positive neuromyelitis optica spectrum disorder and six of 59 patients with MOGAD met the new criteria for multiple sclerosis. Overall accuracy rose modestly. The main limitation is that specificity rests on small antibody-defined groups and the reference standard was the older criteria themselves. It adds the first multinational estimate of how the 2024 criteria perform in optic neuritis, and the authors call for careful exclusion of alternative diagnoses and real-world validation.
Now, four in brief.
In JAMA, Albers and Heit review acute ischaemic stroke treatment against the 2026 American Heart Association and American Stroke Association guidelines [4]. Intravenous thrombolysis with tenecteplase or alteplase is first line within 4.5 hours for disabling symptoms, extendable to 9 hours, and up to 24 hours in patients with large vessel occlusion who cannot receive thrombectomy, when perfusion imaging shows salvageable tissue. Thrombectomy carries a class 1 recommendation within 6 hours, and in selected patients up to 24 hours, after thrombolysis unless contraindicated. Pooled trial data give about a ten point absolute gain in good outcome with thrombolysis and about twenty points in functional independence with thrombectomy. As a narrative review, it summarises guideline-level evidence without new data.
Staying with bleeding on anticoagulants, Sharma and colleagues in Stroke pooled 741 andexanet-treated participants from ANNEXA-I and ANNEXA-4, about four in five with intracranial haemorrhage [5]. About one in ten participants had a thrombotic event, and about one in five participants died. No baseline factor predicted thrombotic events overall, but arterial events were linked to prior myocardial infarction, eligibility for the high dose, female sex and older age. An arterial event was associated with more than double the hazard of death, and death was also more likely with prior heart failure and with intracerebral or intraventricular haemorrhage. Restarting anticoagulation was associated with fewer thrombotic events. This is a secondary analysis of treated patients only, so these are associations.
Two papers this week address plasma p-tau217 on the Lumipulse platform. In Neurology, Bayart and colleagues tested whether thresholds transfer across centres, analysers and reagent lots, in 411 participants from Brussels with validation cohorts in Paris, Bordeaux and monogenic frontotemporal lobar degeneration [6]. A cutoff of 0.193 picograms per millilitre gave about 92 percent sensitivity and specificity for amyloid positivity, and performance held in Paris and Bordeaux. Specificity was lower in monogenic frontotemporal degeneration, with false positives concentrated in cases with ALS. With dual thresholds, between a sixth and a quarter of Paris patients fell into a grey zone, and CSF testing resolved most of those cases. The data are cross-sectional and from specialised Western European centres.
Algeciras-Schimnich and colleagues in JAMA Neurology took the same assay into cognitively unimpaired adults aged 60 to 90 in the population-based Mayo Clinic Study of Aging and in ADNI, about 2,200 people combined [7]. Comparing the 90th with the 10th percentile of p-tau217, the hazard of clinical progression was roughly two to four times higher across sex and APOE epsilon 4 groups in both cohorts. At age 75, the modelled ten-year risk of mild cognitive impairment more than doubled across that range, to about 35 percent. At the higher concentration, ten-year risk reached about two thirds by age 85. These are modelled estimates from observational cohorts, proposed for prognosis and trial enrichment.
Three quick signals.
Scott and colleagues in JAMA Neurology followed nearly four and a half million Vietnam War era veterans in the Veterans Affairs system [8]. Self-reported Agent Orange exposure was associated with about 44 percent higher risk of Parkinson disease and higher risks of REM sleep behaviour disorder and dementia with Lewy bodies, more so with direct occupational exposure. Exposure was self-reported in an observational design.
Weber and colleagues in Neurology ran a CSF alpha-synuclein seed amplification assay that distinguishes two seed types in 1,059 consecutive movement disorder clinic patients [9]. Seeds appeared in just over half of patients, mostly type 1. Type 2 seeds dominated in multiple system atrophy and appeared in under three percent of other patients. This is single-clinic, cross-sectional reference data.
Szmulewicz and colleagues in Neurology emulated a target trial in more than 900,000 men aged 60 or older in Veterans Affairs records, without cardiovascular disease or diabetes [10]. Ten-year dementia risk was about nine and a half percent with or without statin initiation, with no meaningful difference. The authors rate this Class III evidence, limited to men.
Stepping back across this week's papers, the two randomised trials gave answers in opposite directions with different degrees of certainty: the double-blind trial returned a clear null, while the trial suggesting benefit stopped early outside its own efficacy boundary. Much of the rest rests on cross-sectional or registry data, where biomarker thresholds and diagnostic rules were tested against reference standards that are themselves imperfect, and where large observational cohorts describe associations, or their absence, without settling causation.
If you only have time for one paper this week, make it OTEMACS [1]. It reopens the question of whether embolisation belongs early after surgery for chronic subdural haematoma, with a fall in reoperation that the authors themselves say still awaits confirmation.
That's your roundup for This Week in Neurology. The full transcript and references are available on the episode page. See you next week.
If this weekly briefing is useful, follow the show in your podcast app so new episodes arrive automatically. And think of one colleague, in any specialty, who never has time to keep up with the literature. Tell them about AudioScholar: a ten-minute weekly for every specialty, to listen to in any podcast app, or to read at audioscholar dot C C.
This is an automated summary generated by artificial intelligence, which can make mistakes. Use it to choose which papers to read in full.
Get Neurology every week
By email, Spotify, Apple or any podcast app.
Spot something worth flagging?