Cerebrospinal fluid levels of TDP-43, the protein that defines limbic-predominant age-related TDP-43 encephalopathy (LATE), were highest in people diagnosed with probable LATE, above those measured in amyotrophic lateral sclerosis and Alzheimer's disease, according to a preprint posted to medRxiv on May 10, 2026 by Anna-Lisa Fischer, Peter Hermann, Inga Zerr and colleagues. The work has not been peer reviewed. The researchers used an ELISA to measure TDP-43 in single spinal fluid samples from 90 people: 25 controls, 32 with ALS, 9 with probable LATE and 24 with Alzheimer's disease. Controls were lowest; ALS and Alzheimer's showed intermediate and comparable increases. Standard Alzheimer's markers (total tau, phosphorylated tau and amyloid beta) followed the expected Alzheimer's pattern but were largely unchanged in the probable LATE group, which the authors describe as a dissociation between TDP-43 and Alzheimer's-type pathology.
The result matters because LATE has no molecule-specific test in living people. The 2019 consensus working group report in Brain that named the condition stated flatly that no such biomarker existed, and described LATE neuropathologic change as present in more than 20 per cent, and up to 50 per cent, of people past the age of 80 in large community-based autopsy series, usually producing an amnestic dementia easily mistaken for Alzheimer's disease. Clinical criteria published in Alzheimer's and Dementia in January 2025 by David Wolk, Peter Nelson and co-authors define probable LATE in life by progressive memory loss with substantial hippocampal atrophy and negative amyloid biomarkers, which is how the nine cases in this preprint were classified: by clinical rule, not by pathology.
That is also the study's weakest joint. None of the participants had autopsy confirmation, so the group labelled LATE is a clinical category that autopsy could revise, and nine people is a very small group on which to rank one disease above another. The finding also poses a mechanistic puzzle the authors do not resolve: it is not obvious why a proteinopathy largely confined to the amygdala and hippocampus should release more TDP-43 into spinal fluid than a disease of the motor system. Elevated cerebrospinal fluid TDP-43 in ALS was first reported in Acta Neuropathologica in 2008, and total-protein assays of this kind measure TDP-43 in bulk; they do not distinguish misfolded, aggregated protein from the normal form the protein takes in healthy cells and in blood.
Other groups have approached the same protein differently. A digital seed amplification assay published in Alzheimer's and Dementia in May 2026 by Ella Borberg and colleagues counts individual TDP-43 aggregates in 40 cerebrospinal fluid samples, mostly from frontotemporal lobar degeneration with TDP-43 pathology, and a seed detection study published in the same journal in December 2025 tested 73 people with genetic FTD or ALS and presymptomatic mutation carriers. Both were pitched at the FTLD and ALS end of the spectrum. The Fischer preprint's claim to novelty is its aim at the limbic form, which is the commoner problem in memory clinics and the one clinicians currently cannot confirm.