Blood Test for Alzheimer’s Reshapes Dementia Diagnosis in Real-World Clinics

A large real-world study of the Alzheimer’s disease blood biomarker p-tau217 has found that the test behaves distinctly across different clinical presentations, and that its adoption is already changing how memory clinics work. Between January and December 2025, researchers analyzed 1,368 consecutive patients undergoing plasma p-tau217 testing as part of routine evaluation for cognitive disorders across 24 participating centers in secondary and tertiary care. The results, published in Alzheimer’s Research & Therapy, offer one of the first detailed pictures of how a blood-based Alzheimer’s biomarker performs outside carefully controlled research cohorts.

The patients fell into three groups based on their referring clinicians’ diagnostic hypotheses: 997 had suspected Alzheimer’s disease, 180 had suspected non-Alzheimer cognitive disorders, and 191 had cognitive impairment without a specific etiological hypothesis. Plasma p-tau217 was categorized as negative, equivocal, or positive using predefined cut-offs. Positivity was markedly higher in the suspected Alzheimer’s group than in the suspected non-Alzheimer group, at 66.1% versus 17.8%, while patients without a specific hypothesis showed an intermediate distribution. That gradient suggests the biomarker tracks clinicians’ pre-test suspicions, but also that a substantial share of positive results can appear outside classic Alzheimer’s presentations, underscoring the need to interpret results alongside the clinical context.

Notably, equivocal results were common: 19.9% of all tests fell into this gray zone, and they were more frequent in older individuals. The study also mapped biological determinants of the marker. Plasma p-tau217 concentrations increased with age, were higher in women, were inversely correlated with MMSE scores, and were inversely correlated with renal function. These associations matter because they mean the same numerical result can carry different meaning in an 85-year-old with reduced kidney function than in a younger patient, and the authors explicitly recommend integrating patient-specific factors such as age and renal function, together with pre-test probability, when interpreting results.

To validate the blood test against established tools, the team retrospectively analyzed paired cerebrospinal fluid data where available. In 117 patients with both plasma and CSF measurements, plasma p-tau217 correlated with the CSF p-tau181/Aβ42 ratio, with a Spearman correlation coefficient of 0.640, and showed good discrimination for a positive ratio above 0.10, with an area under the curve of 0.870. These figures, drawn from a subset rather than the full cohort, support the plasma test as a reasonably accurate surrogate for CSF-based Alzheimer’s pathology assessment in routine practice, though they also leave room for discordance in individual cases.

The most striking findings may be the downstream effects on diagnostic pathways. In a survey of 53 prescribers, 71% reported increased confidence in plasma p-tau217 since starting to use it, 80% reported a shorter diagnostic delay, and 63% reported reduced use of brain fluorodeoxyglucose positron emission tomography. Objective activity data backed this up: cerebrospinal fluid biomarker requests decreased by approximately 46% between January–June 2024 and July–December 2025.

The study’s conclusions are measured rather than triumphant. The authors emphasize that plasma p-tau217 showed distinct distributions across clinical phenotypes and was associated with measurable changes in diagnostic pathways, but they caution that interpretation should integrate patient-specific factors, including age and renal function, together with the clinical context and pre-test probability. Because the analysis was observational and cross-sectional, it documents associations between the biomarker and clinical categories, not proof that the test improves diagnostic accuracy or long-term patient outcomes. The prescriber survey reflects subjective reported confidence and practice change, and the paired CSF validation covered only a small subset of patients.

Even with those limitations, the scale of the dataset is unusual for biomarker research. Most p-tau217 evidence to date has come from specialist cohorts and controlled validation studies; here, the test was applied to an unselected stream of consecutive patients across two dozen centers, capturing the messier reality of routine care, including older patients, comorbidities, and diagnostically ambiguous referrals. The high rate of equivocal results in this setting is itself an important practical finding, since it signals that a meaningful fraction of real-world tests will not yield a clean binary answer and will require clinical judgment to resolve.

For clinicians, the study offers a working framework: treat plasma p-tau217 as one input whose positive and negative predictive values depend on who is being tested. A positive result in a patient already suspected of having Alzheimer’s disease carries different weight than the same result in someone with an atypical presentation, and age- and kidney-related shifts in concentration can move individual values without changing underlying disease status.

As blood-based biomarkers move from laboratories into everyday memory clinics, this study provides a template for monitoring them in the wild: track distributions across diagnostic categories, quantify the equivocal zone, validate against established fluids where possible, and measure what actually changes in the diagnostic pathway. On that scorecard, p-tau217’s first year of routine use appears to have passed a meaningful real-world test.

Subject of Research: Real-world clinical performance of the plasma p-tau217 blood biomarker for Alzheimer's disease

Article Title: Plasma p-tau217 in real-world clinical practice: distribution across clinical diagnostic categories, clinician confidence, and impact on care pathways

Article References: Bayart, J.-L., Colmant, L., Boyer, E., Pirson, F. B., Picard, G., Pieret, F., Mormont, E., Segers, K., Barglazan, D., Ivanoiu, A., Douxfils, J., & Hanseeuw, B. J. (2026). Plasma p-tau217 in real-world clinical practice: distribution across clinical diagnostic categories, clinician confidence, and impact on care pathways. Alzheimer's Research & Therapy. https://doi.org/10.1186/s13195-026-02244-9

Image Credits: AI Generated

DOI: 10.1186/s13195-026-02244-9

Keywords: plasma p-tau217, Alzheimer's disease, blood biomarker, cerebrospinal fluid biomarkers, cognitive disorders, diagnostic pathways, memory clinics, renal function, equivocal results, clinical practice, Plasma, p-tau217