Reviews/Alzheimers Disease Treatment/Plasma phosphorylated tau 217 detects Alzheimer's pathology
Biomarkers · AD-BIOM-01 · v2
This corpus shows that a blood test for a tau fragment called p-tau217 agrees closely with the established ways of detecting Alzheimer's brain changes, such as PET scans and spinal-fluid tests, across many separate groups of patients, different laboratory methods, and both specialist and family-doctor settings [31][33][35][30][38][39][40][42]. In one group where brains were examined after death, the blood-based result lined up well with the actual confirmed disease [13]. The panel judged the agreement strong and repeated enough to carry the claim. It also noted real limits: the test works less well in people with a history of head injury [21], in some older and demographic groups [28][31], and its answer depends on which cutoff is chosen [31][42]. So the overall picture is a reliable test with clearly marked situations where it should be read with caution.
Many independent cohorts and assay platforms report that plasma p-tau217 separates people with and without Alzheimer's pathology with area-under-the-curve values around 0.9 to 0.97 against amyloid-PET, tau-PET and CSF references [33][35][30][38][39][41][42]. A fully automated test validated across primary and secondary care detected AD pathology with AUCs of 0.93 to 0.96 [31], and a memory-clinic implementation reported an AUC of 0.90 with 81% sensitivity and 91% specificity [40]. In a neuropathology subsample, a plasma staging model was strongly aligned with autopsy-confirmed AD pathology (AUC 0.96) [13]. Performance held across several commercial and mass-spectrometry assays, with mass-spectrometry %p-tau217 performing best [35][39]. The contested point: The division is over the word 'accurately' being unqualified: the reference standard in most studies is another biomarker (amyloid-PET, tau-PET or CSF) rather than neuropathology, with direct autopsy anchoring in only one subsample [13], and accuracy is demonstrably lower in specific contexts such as prior TBI [21], age ≥80 [31] and some demographic subgroups [28]. The panel treated these as bounding the claim's scope rather than refuting its core, so the disagreement is about breadth of the word 'accurately', not about whether the signal is real.
All 42 papers behind this claim with their DOIs and what each does to it, the panel’s arbitration, the evidence gap map and the open research questions. Free, and sent to your inbox.
We’ll also send you other claims from Alzheimers Disease Treatment, one at a time, and tell you when a release or a pivotal trial changes a grade. Unsubscribe by replying.
One reviewer's only job is to attack the finding. The challenges below were weighed before the grade was set.
The reference standard in most included studies is amyloid-PET, tau-PET or CSF (themselves biomarkers of pathology), not neuropathology; direct autopsy validation exists in only one neuropathology subsample [13], so 'detects AD pathology' rests partly on surrogate-against-surrogate comparison.
Accuracy is context-dependent, not universal: p-tau217/Aβ42 'may miss over half of amyloid-PET-positive cases' in individuals with TBI history [21], accuracy falls at age ≥80 [31], and p-tau217 was 'not associated with incident MCI in Black women' [28] — the unqualified word 'accurately' overstates performance in these subgroups.
Where it remains uncertain. Accuracy is not universal: it falls with prior TBI (to 63%) [21], at age ≥80 (83%) [31], and p-tau217 was not associated with incident MCI in Black women [28], while most validation rests on surrogate biomarker rather than neuropathology references.
Aleatory
Two-cutoff workflows leave 10-17% of cases in an intermediate result band [31], an irreducible grey zone for a share of individuals.
Epistemic
Direct neuropathological confirmation appears in only one subsample (AUC 0.96) [13]; broader autopsy validation is not in the corpus.
Inferential
Most studies validate against amyloid-PET, tau-PET or CSF, which are themselves biomarkers of pathology [31][33][35], so 'detects pathology' partly infers from surrogate to surrogate.
The load-bearing papers from a corpus of 42, each classified by what it does to the claim. Every DOI is taken from the corpus the claim was graded on.
Rare gold-standard autopsy anchoring, mitigating surrogate-endpoint concern
“In the neuropathology subsample, plasma staging was strongly aligned with autopsy-confirmed AD pathology using the Alzheimer Disease Neuropathologic Change scale (area under the curve, 0.96; 95% CI, 0.91-1.00).”
Panel auditInternal validity: high · scope of inference: CU to dementia across AD and non-AD, biological staging detection.Key limitations: %p-tau217 combined with eMTBR-tau243, not p-tau217 alone · small neuropathology subsample
Discrimination from non-AD neurodegenerative disease, not just controls
“Lumipulse and SIMOA showed similar diagnostic accuracy for differentiating Alzheimer's disease from other neurodegenerative diseases [area under the curve (AUC) 0.952, 95% confidence interval (CI) 0.927-0.978 versus 0.955, 95% CI 0.928-0.982, respectively]”
Panel auditInternal validity: high · scope of inference: AD vs other neurodegenerative disease vs controls, clinical diagnosis.Key limitations: CSF p-tau181 reference partly circular · single-center
Large multi-cohort real-world validation including primary care
“Plasma p-tau217 detected AD pathology with areas under the receiver operating characteristic curves of 0.93-0.96. In secondary care, the accuracies were 89-91%”
Panel auditInternal validity: high · scope of inference: Symptomatic patients in primary and secondary care, AD pathology detection.Key limitations: lower accuracy at age ≥80 · CSF reference not PET/autopsy
Bounds the claim: accuracy degrades with TBI history
“the p-tau217/Aβ42 ratio test may miss over half of amyloid-PET-positive cases”
Panel auditInternal validity: medium · scope of inference: Vietnam-era veterans with/without TBI, amyloid detection accuracy.Key limitations: TBI-specific degradation · 99% male veteran sample · ratio and p-tau217-alone both affected
4 of 42 papers shown. The full review lists all of them, with the panel’s arbitration →
Claims are graded at each quarterly release, and re-graded between releases only when a pivotal trial lands. Every change is a new version of the claim.
Convergent, independent, high-accuracy diagnostic evidence including autopsy concordance carries the claim at the Established tier, with context-specific caveats preventing a ceiling score.
Release 2026-q4 re-graded every claim from its corpus under protocol v0.19. An earlier release's grade is shown for the record; it came from an earlier protocol and is not a movement.
Assumptions that carry weight but are not themselves established.
The panel records the gap that would most move the claim, and the study that would close it.
Autopsy-standard validation at scale. Most accuracy figures rest on surrogate biomarker references rather than neuropathology [31][33][35], with autopsy anchoring in only one subsample [13]. The panel rates this the highest-importance open gap for the claim, and names the study design that would close it: Large neuropathology-confirmed cohort measuring antemortem plasma p-tau217 against postmortem AD neuropathologic change.
Open question: How closely does plasma p-tau217 match autopsy-confirmed AD neuropathology outside a single subsample? Why it is still open: Only [13] anchors to neuropathology (AUC 0.96); the remaining validation uses imaging or CSF surrogates [31][33][35].
Priors is a sceptical first reader, not peer review, and not medical advice.
Release 2026-q4 · graded 21 Oct 2026 · protocol v0.19