This Week in Neurology — Sep 23, 2026

Welcome to This Week in Neurology. This week we're covering 10 notable papers spanning acute stroke reperfusion and imaging, the long-term cognitive and epidemiologic fallout of cerebrovascular disease, and a cluster of papers on epilepsy, neuroimmunology and Parkinson's prevention. Let's dive in.

We start with acute stroke, where two trials pull in opposite directions about who benefits from what. In JAMA, Li and colleagues report ATTENTION LATE, an open-label, blinded-endpoint trial at 40 stroke centres in China that randomised 330 patients with moderate to severe basilar artery occlusion, all with National Institutes of Health Stroke Scale scores of at least 10, presenting between four and a half and 24 hours from onset directly to thrombectomy-capable centres [1]. Patients received either intravenous tenecteplase at 0.25 milligrams per kilogram, capped at 25 milligrams, before thrombectomy, or thrombectomy alone. At 90 days, exactly 30 percent OF PATIENTS in each arm were functionally independent — no signal of benefit whatsoever. Symptomatic intracranial haemorrhage was uncommon and similar, around 5 percent in the tenecteplase arm and 4 percent with thrombectomy alone, and mortality was high in both groups, around 40 percent. This is a clean negative result, and the message for practice is straightforward: for the late-window basilar patient who is already in your angiography suite, adding a lytic before the procedure does not improve outcomes, and the trial gives no support for doing it routinely.

Contrast that with a post hoc analysis of the LASTE trial in Annals of Neurology, where Seners and colleagues asked a logistics question that comes up almost daily [3]. LASTE randomised anterior circulation large-vessel occlusion patients with a large ischaemic core — an ASPECTS score of 0 to 5 — treated within six and a half hours. This analysis took the 157 patients whose baseline imaging was done at a primary stroke centre and already showed that large core before the decision to transfer. Median age was 73, median stroke severity score 20, and more than half OF THE PATIENTS had an ASPECTS of just 0 to 2. Thrombectomy still shifted the 90-day disability distribution significantly in its favour, with no heterogeneity across ASPECTS strata. Mortality was numerically lower with thrombectomy and symptomatic haemorrhage numerically more frequent, neither difference conclusive. The practical implication is that a devastated-looking core on the initial non-contrast scan at a primary stroke centre should not, by itself, cancel the transfer — provided the patient can reach the comprehensive centre inside the early window.

Still on imaging, Broocks and colleagues, also in Annals of Neurology, compared quantitative net water uptake on non-contrast CT with relative FLAIR hyperintensity in 53 hyperacute stroke patients who were all diffusion-positive [8]. Net water uptake rose steadily and near-linearly over the first four and a half hours and separated early, intermediate and late edema stages cleanly, whereas FLAIR signal only jumped at the latest stage, once sulcal effacement appeared. The two measures were essentially uncorrelated. For a CT-first pathway, this suggests quantitative water uptake could serve as a practical tissue clock and edema-risk marker, though it is a small single-cohort study awaiting prospective validation.

The second theme is what happens to these patients before and long after the acute event. In Neurology, Joundi and colleagues used province-wide administrative data from Ontario to follow more than 113,000 adults after a first transient ischaemic attack, matched to controls on age, sex, deprivation, rurality and vascular comorbidity [2]. Close to one in five OF THE TIA PATIENTS developed dementia over a mean of seven years. The hazard was highest in the first year, roughly 75 percent above matched controls, and remained about a third higher out to twenty years. Compared with stroke patients, dementia risk after TIA was lower early on but converged by five years. Critically, dementia occurred far more often than recurrent stroke, and adjusting for interval stroke barely attenuated the association. That reframes the TIA clinic visit: secondary prevention is necessary but not sufficient, and cognition deserves to be on the follow-up agenda.

Also in Neurology, Fisher and colleagues report population-based stroke surveillance from a five-county region across six study periods between 1993 and 2020, standardised to United States Census data [7]. Among adults aged 20 to 54, stroke incidence rose from about 34 to about 62 cases per hundred thousand person-years, while incidence in older adults fell. The rise was driven by ischaemic stroke, which nearly doubled. Hypertension, diabetes, atrial fibrillation and substance use all became more common in these young patients over time. Thirty-day case fatality did decline modestly, from around 12 percent to around 9 percent, driven mostly by the haemorrhagic subtypes. So we are getting better at keeping young stroke patients alive while failing to stop them having strokes in the first place.

And on the front end of that pathway, JAMA Neurology publishes a review by Sars and colleagues on when to use the HINTS examination [6]. Their argument is that HINTS is being applied to the wrong patients. It is validated for continuous dizziness with spontaneous nystagmus at rest — acute vestibular syndrome — and performs poorly outside that setting. Common errors include applying it to patients with no nystagmus at all, and skipping the step of screening for central features first. They also argue that recent emergency department guidance has introduced ambiguity rather than resolving it. The takeaway for anyone teaching or supervising in the emergency department: screen for central signs first, confirm there is nystagmus at rest, and only then perform HINTS.

Turning to epilepsy, two papers address different ends of the care pathway. In Neurology, Repplinger and colleagues report the ICTAL registry — a prospective cohort of 1,123 adults admitted to 26 French intensive care units with status epilepticus, all undergoing a predefined diagnostic protocol [4]. The yield of any given test depended heavily on pretest probability. Toxicology screening confirmed the suspected cause in about 45 percent of alcohol, toxic or iatrogenic presentations, antiseizure medication assays in a similar proportion where a precipitating factor was suspected, and paraneoplastic and autoimmune workup in about 40 percent of suspected non-infectious encephalitis. But the more striking finding is reclassification: brain MRI changed the etiologic diagnosis in roughly a quarter to a third of cases across several strata, and management changed after MRI in up to about 58 percent of patients, with CT close behind and lumbar puncture changing management in around 40 percent. This supports a stratified rather than shotgun workup — and it makes a strong case for getting the MRI done.

Then in Epilepsia, Sferruzzi and colleagues asked whether pretreatment EEG can predict who will not tolerate levetiracetam [10]. Across three Italian centres they took 31 drug-naive focal epilepsy patients who developed clinically significant neuropsychiatric adverse effects within a month of starting levetiracetam, and matched them to 31 who did not. Spectral power showed nothing, but theta-band graph-theoretical connectivity measures differed at baseline, and a penalised regression model discriminated the two groups with an accuracy of about 0.71 and an area under the curve of about 0.71, performing better than chance on permutation testing. This is a small, retrospective, matched study, so it is a proof of concept rather than a clinical tool — but the idea that a pretreatment resting EEG might flag a tolerability phenotype is worth watching.

Finally, two papers on diagnosis and prevention elsewhere in neurology. In Neurology, Sukhonpanich and colleagues reviewed 331 spinal cord MRI scans from 217 patients with aquaporin-4 antibody positive neuromyelitis optica spectrum disorder [5]. Restricting analysis to first-ever myelitis, bright spotty lesions were seen in about two thirds OF THOSE SCANS, sagittal ring enhancement in about 60 percent, and a novel double-contoured longitudinally extensive lesion sign in just under half — but all three were highly specific, in the mid-to-upper nineties, against other causes of myelopathy. Combining signs pushed specificity higher but dropped sensitivity to around 30 percent. The authors are candid that their comparison cohort was enriched with spinal cord infarction and other non-inflammatory myelopathies, which limits generalisability. Notably, spinal cord atrophy was present in well over half OF THE REMISSION SCANS, a reminder of how much structural damage accrues.

And in Movement Disorders, Ju and colleagues used United Kingdom primary care records to emulate a target trial of ursodeoxycholic acid and Parkinson's disease risk, matching over 17,000 initiators with hepatobiliary disease to roughly 52,000 non-initiators [9]. Fifteen-year Parkinson's risk was 1.70 percent with the drug versus 2.16 percent without — an absolute difference of under half a percentage point, or about a fifth lower in relative terms, with consistent direction across onset-anchored and dose-response analyses. Hypothesis-generating, confounding by indication is hard to exclude entirely, and the authors themselves call for randomised confirmation.

If you only have time for one paper this week, make it ATTENTION LATE in JAMA [1]. It is a properly powered randomised answer to a question many of us have been improvising around, and a clear negative result in the late-window basilar patient should change what happens in your angiography suite tomorrow.

Here are the key takeaways from this week in Neurology. Intravenous tenecteplase before thrombectomy in late-presenting basilar occlusion offered no functional benefit and should not be routine. A large ischaemic core seen at a primary stroke centre is not by itself a reason to cancel transfer for thrombectomy within the early window. TIA carries a large, sustained dementia risk that approaches that of stroke over time — build cognition into long-term follow-up. Stroke incidence in adults under 55 has roughly doubled over three decades alongside rising vascular risk factors and substance use, even as case fatality falls. In status epilepticus, let pretest probability drive the workup, and prioritise MRI, which reclassified the etiology and changed management in a substantial minority of intensive care patients. And use HINTS only in patients with continuous dizziness and spontaneous nystagmus at rest, after screening for central features.

That's your roundup for This Week in Neurology. The full transcript and references are available on the episode page in your AudioScholar library. This is an AI-curated summary — for clinical decisions, always consult primary sources and current guidelines. 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 free 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. Always review the original source materials.