This Week in Neurology — Sep 9, 2026

Welcome to This Week in Neurology. This week we're covering 10 notable papers spanning the biomarker-driven remaking of Alzheimer's disease care, two negative or null trials that clear the decks in amyotrophic lateral sclerosis and in seizure rescue, and new prognostic signals in Parkinson disease — plus a small but striking mortality result in severe traumatic brain injury. Let's dive in.

We start with Alzheimer's disease, where The Lancet Neurology has published a cluster of papers that together describe a specialty in transition from clinical to biological diagnosis. The most immediately usable is a pooled analysis by Xu and colleagues of seven multi-ethnic prospective cohorts from Canada, the Dominican Republic, and the United States, including more than eight and a half thousand older adults with plasma phosphorylated tau 217 measured alongside APOE genotype [3]. Higher p-tau 217 was associated with both prevalent and incident cognitive impairment, as expected — but the key point is that APOE modified that relationship. Among APOE-ε4 carriers, each standard-deviation rise in p-tau 217 shortened the time to cognitive impairment by about a quarter; among non-carriers, the same rise shortened it by roughly half as much. Survival curves separated only three to four years after the blood draw, and before symptom onset. Practically, that means a raised p-tau 217 in an ε4 carrier carries a shorter runway than the identical value in a non-carrier, and genotype adds prognostic information the biomarker alone does not give you. It also means the honest answer to "when" is measured in a few years, not months. Two companion reviews frame where this fits. Teunissen and colleagues argue that no single fluid marker will suffice: neurofilament light is already in practice for frontotemporal dementia and amyotrophic lateral sclerosis, p-tau 217 for Alzheimer's diagnosis, but capturing vascular, alpha-synuclein and TDP-43 copathology — and the glial dysregulation that may explain why treatment courses differ between patients — will require panels, point-of-care platforms, and remote sampling [4]. Benzinger and colleagues make the parallel case for imaging, with amyloid PET central to establishing diagnosis, confirming eligibility for amyloid-targeting therapy and quantifying amyloid clearance, tau PET informing biological stage and prognosis where it is available, and MRI remaining indispensable for differential diagnosis, comorbid vascular burden, and surveillance for amyloid-related imaging abnormalities [6]. The most cautionary of the group is a clinical-pathways paper from Graff-Radford and colleagues on cognitively unimpaired people who have biomarker evidence of Alzheimer's disease neuropathological change [5]. Prevention trials are ongoing, and the trial data in symptomatic disease hint that benefit may be greatest at low baseline pathology — but the authors are blunt about what testing asymptomatic people creates: false positives when prevalence in the tested population is low, intermediate results that reliably indicate nothing, and downstream consequences for patients, insurance, and health systems that we are not yet equipped to manage. Read together with the p-tau 217 data, the message for the clinic is that we now have tests capable of generating a prognosis years in advance, and an obligation not to deploy them casually outside the enriched populations in which they were validated.

Turning to Parkinson disease, two papers this week ask what actually drives progression — and one of them defuses a long-running worry. In JAMA Neurology, Schumacher and colleagues examined common loss-of-function variants in the melanocortin 1 receptor gene, which regulates pigmentation and oxidative stress and is carried by more than 60 percent of people of European descent [9]. Among just over 380 participants with sporadic Parkinson disease in the Parkinson Progression Markers Initiative, carriers declined on the motor rating scale about 30 percent faster than non-carriers, and in an independent replication cohort assembled from three United States trials the gap was about 50 percent. In a small prodromal group of 53 participants, carriers had more than a fourfold higher risk of phenoconversion to manifest Parkinson disease — a finding to treat as preliminary given the numbers. This is not yet a test to order in clinic, but it identifies a very large, easily genotyped subgroup that could be used to enrich progression trials. Meanwhile in Neurology, Barbosa and colleagues addressed the recurring claim that levodopa exposure itself contributes to freezing of gait, using more than twenty-five thousand patients from the French NS-Park cohort plus around fourteen hundred from the Parkinson Progression Markers Initiative [10]. In unadjusted analyses levodopa did look associated with more frequent and more severe freezing in the larger cohort — but once disease duration, Hoehn and Yahr stage, and motor scores were accounted for, the association became non-significant and inconsistent across cohorts, and no relationship was found between future levodopa exposure and new-onset freezing. What did predict freezing was disease duration, stage, and motor severity. So the practical implication is reassurance: withholding or rationing levodopa out of fear of provoking freezing is not supported by these data.

Our third theme is acute and interventional neurology, including two trials that returned null results and one small trial with a surprising positive. In The Lancet Neurology, Cudkowicz and colleagues report HIMALAYA, a phase 2 trial of the oral, central-nervous-system-penetrant RIPK1 inhibitor SAR443820 across 63 sites in 13 countries, randomising 305 adults with amyotrophic lateral sclerosis two-to-one to drug or placebo for 24 weeks [1]. Functional decline was essentially identical in the two arms — a fall of a little under seven points on the revised functional rating scale in both — with no significant difference. Adverse events and discontinuations were more common on active drug, driven mainly by elevated hepatic enzymes, with roughly one in seven patients stopping treatment compared with one in twenty on placebo. The trial was terminated early and the authors conclude further development in amyotrophic lateral sclerosis is not warranted. It is a clean negative result for the neuroinflammatory cell-death hypothesis as targeted by this molecule. In Epilepsia, Fatima and colleagues ran a single-centre randomised trial in an epilepsy monitoring unit comparing intranasal midazolam with mouth-dissolving clobazam in 95 patients with drug-resistant epilepsy who had seizures lasting beyond two minutes [8]. There was no statistically significant difference in time to clinical or electrographic seizure termination, no significant adverse effects in either arm, and comparable treatment satisfaction. The authors position this as a signal that mouth-dissolving clobazam deserves further study as a cheaper, more convenient rescue option — but note this was an underpowered single-centre study in a monitored setting, not a prehospital efficacy trial. The positive outlier comes from Brain, where Cederberg and colleagues tested antisecretory factor, delivered as an enriched freeze-dried egg-yolk preparation via nasogastric tube, in 100 adults with isolated severe traumatic brain injury and intracranial pressure monitoring at a tertiary trauma centre in Cape Town [7]. Thirty-day mortality was 20 percent with treatment versus 39 percent with placebo — roughly a halving of risk — with no difference in intracranial pressure or therapy intensity level and no significant adverse events. That is a large effect from a phase 2 trial of 100 patients at a single site, with an unexplained mechanism given that pressure control was unchanged, so it should be read as hypothesis-generating and a strong case for multicentre replication rather than a change in practice. Finally, in Annals of Neurology, Tanaka and colleagues performed a post hoc analysis of 434 patients in the ANNEXa-I trial of andexanet for factor Xa inhibitor-associated intracerebral haemorrhage [2]. At least one non-contrast CT marker of instability was present in just over seven in ten patients, and haematoma expansion occurred in roughly a third; heterogeneous density, hypodensity, and the black hole sign were each independently associated with expansion. Andexanet's effect was broadly consistent across marker subgroups — so these signs help you identify who is at high risk on the scan you already have, rather than telling you who will respond to reversal.

If you only have time for one paper this week, make it the pooled p-tau 217 and APOE analysis in The Lancet Neurology [3]. It is the paper most likely to change how you counsel a real patient in the next month, because it turns a blood biomarker from a yes-or-no answer into a time-based prognosis that depends on genotype.

Here are the key takeaways from this week in Neurology. First, plasma p-tau 217 and APOE genotype together stratify both the risk and the timing of cognitive impairment, with the sharpest trajectories in ε4 carriers and meaningful separation emerging only after three to four years. Second, biomarker testing in cognitively unimpaired people is not ready for routine use — false positives and uninterpretable intermediate values are real and consequential. Third, RIPK1 inhibition with SAR443820 failed in amyotrophic lateral sclerosis with added hepatic toxicity, and development has stopped. Fourth, levodopa does not appear to cause freezing of gait once disease duration and severity are accounted for, so treat motor symptoms adequately. Fifth, loss-of-function melanocortin 1 receptor variants mark a common subgroup with faster motor decline — relevant now for trial enrichment, not yet for clinic. And sixth, non-contrast CT markers such as heterogeneous density and the black hole sign flag high expansion risk in anticoagulant-associated intracerebral haemorrhage without predicting reversal response.

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 if you have a paper you have been meaning to read, upload the PDF, or paste any link, at audioscholar dot C C. We will turn it into audio like this one, in any of thirty-one languages.

This is an automated summary generated by artificial intelligence, which can make mistakes. Always review the original source materials.