This Week in Neurology — Aug 12, 2026

Welcome to This Week in Neurology. This week we're covering 10 notable papers spanning blood pressure and vascular intervention in stroke, drug safety decisions in pregnancy and in dementia, risk stratification in early cognitive decline, and diagnostics and therapeutics in rare neurological disease. Let's dive in.

We start with blood pressure, where two papers pull in opposite directions on where the benefit actually lies. In Neurology, the DETERMINE trial randomised 433 patients with anterior large vessel occlusion at eight French comprehensive stroke centres to either individualised blood pressure control during thrombectomy, keeping mean arterial pressure within ten percent of the patient's own pre-procedure value, or to a standard target of systolic pressure between 140 and 180 millimetres of mercury [1]. The individualised strategy did not improve outcomes: a favourable modified Rankin score of zero to two at 90 days was achieved in about 44 percent of the individualised group versus roughly 49 percent of controls, a non-significant difference that numerically favoured standard care, with no difference in mortality or symptomatic intracranial haemorrhage. Importantly, the trial achieved no separation in actual mean pressures or pressure variability between arms, so this is as much a story about the difficulty of delivering personalised haemodynamic targets in the angiography suite as it is about the physiology. The practical message is that there is no reason to abandon the current one-size-fits-all peri-procedural target. Contrast that with RECAP-ICH in The Lancet Neurology, an individual participant data meta-analysis of four randomised trials and 2,944 patients with previous spontaneous intracerebral haemorrhage, three-quarters of them Asian, followed for a median of three and a half years [3]. Here, a sustained systolic reduction of roughly 11 millimetres of mercury cut first recurrent stroke of any type by about 38 percent — 6.5 percent of treated patients versus 10.4 percent of controls — and that was driven overwhelmingly by prevention of recurrent intracerebral haemorrhage, which fell by about 60 percent. Serious adverse events were, if anything, slightly less common in the intensive arm. Taken together with DETERMINE, the pattern is that acute, minute-to-minute pressure manipulation around reperfusion yields little, whereas durable long-term lowering after haemorrhagic stroke is one of the highest-yield things you can do in a secondary prevention clinic.

Staying with vascular lesions, Neurology also published a prospective, population-based Scottish cohort of 306 patients newly diagnosed with cerebral cavernous malformation, 190 of them symptomatic, followed for a median of 16 years [8]. Thirty-seven underwent intervention, almost all microsurgical resection. Using intervention as a time-dependent covariate, the investigators found no association with the primary threshold of persistent impairment at an Oxford Handicap Scale of two or more, but intervention roughly tripled the hazard of the more severe threshold of dependence or death, and was associated with a more than threefold increase in haemorrhage or new focal deficit definitely related to the malformation or to its treatment. There was no signal of benefit for seizures. This is observational, the treated patients were younger and more often presented with haemorrhage, and the authors are explicit that a randomised trial is needed — but it strengthens the case for conservative management as the default in symptomatic cavernous malformation, and for a frank conversation about procedural risk.

Our second theme is drug safety, where two papers ask when to withhold a treatment and when the fear is misplaced. In The Lancet Neurology, the EURAP international registry reports on nearly 16,000 singleton livebirths to women with epilepsy taking antiseizure medication, examining fetal growth [2]. Polytherapy carried roughly a 50 percent higher risk of small for gestational age, severe small for gestational age, and low birthweight compared with monotherapy, and birthweight centile fell progressively as the number of concomitant drugs rose. Among monotherapies, and using lamotrigine as the comparator, the largest reductions in birthweight centile were with topiramate — about 12 centiles lower — followed by phenobarbital, oxcarbazepine, then smaller effects for carbamazepine, valproic acid and levetiracetam. Notably, lamotrigine combined with either levetiracetam or valproic acid showed no growth deficit relative to lamotrigine alone, whereas carbamazepine plus levetiracetam did. So poor fetal growth now joins malformations and neurodevelopment as a third dimension of pre-conception counselling, and the data support minimising drug number and being cautious with topiramate in particular. The mirror image comes from a Lancet Neurology review on cholinesterase inhibitors in dementia with Lewy bodies complicated by cardiac conduction abnormalities [5]. Bradycardia is common in this population because of autonomic degeneration, and cholinesterase inhibitors are frequently withheld on cardiac grounds — yet these drugs remain the most effective treatment for the cognitive and neuropsychiatric burden, and randomised trial and meta-analytic data support a favourable cardiac safety profile. The authors argue that with baseline cardiac screening and monitoring, patients with conduction abnormalities can generally be treated without an excess of major cardiovascular events. Reflexive avoidance is depriving patients of the one therapy that works.

Our third theme is risk stratification in the memory clinic. Neurology reports a cohort study of menopausal hormone therapy using two independent datasets, the National Alzheimer's Coordinating Center autopsy series and the Alzheimer's Disease Neuroimaging Initiative, comparing women aged 50 and over who reported oestrogen-only hormone therapy with non-users [7]. Users had about 35 percent lower odds of increased Alzheimer pathology at autopsy, along with lower plasma and cerebrospinal fluid amyloid burden, and roughly 40 percent lower odds of a clinical dementia diagnosis. The effect sizes are described by the authors as small, the exposure is self-reported, and this is retrospective and non-causal — so it is not a prescribing indication, but it does add biological weight to the oestrogen hypothesis. Alongside it, a retrospective study from the Amsterdam Dementia Cohort and the SCIENCe project tested brain-predicted age difference, the gap between MRI-estimated and chronological age, in 799 patients with subjective cognitive decline or mild cognitive impairment, of whom 29 percent progressed to dementia over a median of 3.2 years [10]. Each additional year of brain-PAD raised progression risk by about four percent overall, but the effect was concentrated in subjective cognitive decline, where it was significant and added to visual rating scales, and absent in mild cognitive impairment, where it contributed nothing. A cutoff of minus 2.6 years gave a high negative predictive value, above 0.9, over two to ten years. The clinical read is that brain age may help reassure the worried patient with subjective complaints, but adds little once mild cognitive impairment is established — and it improved model fit without improving discrimination beyond amyloid status.

Finally, three papers on diagnostics and therapeutics in rarer disease. In Neurology, the ADAM study prospectively tested 327 patients with suspected myasthenia gravis across four Italian centres, comparing an in-house live cell-based assay, a commercial fixed cell-based assay, and indirect ELISA [4]. For acetylcholine receptor antibodies, sensitivities were broadly similar at roughly 64 to 72 percent, but specificity was the dividing line: about 98 percent for both cell-based assays versus only 69 percent for ELISA, which generated false positives in nearly a third of non-myasthenic patients. The live assay held a modest edge over the fixed assay, but fixed assays are a practical near-equivalent for routine labs. If your lab reports acetylcholine receptor antibodies by ELISA, treat a weakly positive result with real scepticism. In Annals of Neurology, an analysis of blood superoxide dismutase 1 activity across discovery and validation cohorts totalling more than 300 individuals found that reduced enzyme activity is specific to SOD1-related amyotrophic lateral sclerosis, present already in asymptomatic carriers and dependent on variant type [6]. Crucially, in 18 patients on tofersen, neurofilament and SOD1 protein levels fell while enzyme activity remained stable, arguing that the benefit of tofersen is not achieved at the cost of further loss of enzyme function — a reassuring safety signal, though cerebrospinal fluid activity remains to be studied. And in The Lancet Neurology, a single-arm phase 1/2 trial of guanabenz in 33 children with early-childhood-onset vanishing white matter, compared with 66 matched untreated historical controls, found a significantly lower risk of losing the ability to walk with support [9]. The tolerability picture is demanding: 63 serious adverse events occurred, hallucinations affected 55 percent of patients, mostly in the first four months and mostly self-limiting, but no patient discontinued for side-effects and there were no deaths. Historical controls limit certainty, yet in a disease with no therapy this is a meaningful signal.

If you only have time for one paper this week, make it the RECAP-ICH individual participant data meta-analysis in The Lancet Neurology [3]. It gives you a clear, immediately actionable number for a group of patients clinicians are often nervous about treating, and it makes long-term blood pressure lowering after intracerebral haemorrhage a firm expectation rather than a cautious option.

Here are the key takeaways from this week in Neurology. Individualised blood pressure targets during thrombectomy did not improve 90-day function, so keep to standard peri-procedural targets. After intracerebral haemorrhage, sustained lowering of systolic pressure by around 11 millimetres of mercury cuts recurrent stroke by nearly 40 percent, mainly by preventing further haemorrhage. Intervention for symptomatic cavernous malformation was associated with worse long-term outcomes in population-based data, supporting conservatism pending a trial. In pregnancy, antiseizure polytherapy and topiramate in particular are linked to poorer fetal growth, so minimise drug number where seizure control allows. And do not reflexively withhold cholinesterase inhibitors from patients with dementia with Lewy bodies who have cardiac conduction abnormalities; screen, monitor, and treat.

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.