What are CSF biomarkers, and how do they fit alongside plasma and PET?

What they are

CSF biomarkers are measurements made on a sample of cerebrospinal fluid - the clear fluid that surrounds the brain and spinal cord. The sample is obtained by lumbar puncture, sometimes called a spinal tap, and the most clinically useful measurements in Alzheimer's disease are:

• Aβ42/40 ratio. The ratio of two forms of amyloid-beta protein in CSF. A low Aβ42/40 ratio is highly correlated with the presence of brain amyloid pathology - the same biology amyloid PET visualises directly.
• Phosphorylated tau (p-tau). Multiple phosphorylated forms of the tau protein can be measured in CSF. Elevated p-tau is associated with the tau pathology that develops downstream of amyloid in Alzheimer's disease.
• Total tau. A measure of overall neuronal injury, less specific than p-tau but informative in combination.

Combined into a profile, these measurements provide an indirect but well-validated readout of the underlying Alzheimer's pathology in the brain.

How they compare to the alternatives

Three diagnostic tools now address the same underlying question - whether Alzheimer's pathology is present - at different points on the cost, accessibility, and procedural-friction curve.

Plasma biomarkers, covered in our explainer on plasma biomarkers, are a blood draw. Lowest friction, lowest cost, increasingly accessible in routine clinical practice. The leading marker, plasma p-tau217, correlates well with both amyloid burden and downstream tau pathology.

CSF biomarkers are a lumbar puncture. Higher friction - the procedure takes a clinic visit, requires a clinician trained in the technique, and produces post-procedure side effects in a minority of patients (most commonly headache). The biomarker performance is excellent and well-validated against autopsy findings.

Amyloid PET, covered in our explainer on amyloid PET, is a brain scan that visualises amyloid pathology directly. High clinical confidence, higher cost than CSF, gated by scanner availability and reimbursement.

In practice, the three tools are complementary rather than substitutable. Each one answers a slightly different question with a different access profile.

Why CSF testing is still done

The shift toward plasma biomarkers and amyloid PET has reduced the front-line use of CSF biomarkers in routine Alzheimer's diagnosis, but CSF testing remains the right tool in several situations:

• Atypical presentations. When the clinical picture is unusual - younger-onset symptoms, primarily non-memory presentations, rapidly progressive courses - the broader CSF panel can help differentiate Alzheimer's from other neurodegenerative or inflammatory causes that have their own CSF signatures.
• Complex differentials. Where the question is not simply "is amyloid pathology present" but "which of several plausible neurological diagnoses is this," CSF testing can be more informative than amyloid PET alone.
• PET unavailable. In settings where amyloid PET is not accessible - by geography, scheduling, or coverage - CSF biomarkers remain the best available confirmatory test for Alzheimer's pathology.
• Equivocal plasma results. When plasma biomarker results are ambiguous and the diagnostic question is consequential enough to warrant confirmation, CSF testing is one option alongside amyloid PET.
• Research and trial enrolment. Many clinical trials still use CSF biomarker confirmation as part of their eligibility workup, sometimes alongside or instead of PET.

What the procedure involves

A lumbar puncture is performed in an outpatient clinic or hospital setting, typically by a neurologist, anaesthesiologist, or trained interventional radiologist. After local anaesthetic, a thin needle is inserted between vertebrae in the lower back to access the cerebrospinal fluid space. A small sample of CSF is collected - usually a few millilitres - and the needle is removed.

The procedure typically takes 20-45 minutes total, with the actual sampling taking only a few minutes. Most patients tolerate it well. The most common post-procedure side effect is a headache that develops within a day or two and usually resolves with rest, fluids, and time. Less common complications are well-characterised and rare in experienced hands.

What the results can and cannot tell you

CSF biomarkers can tell you:

• Whether the underlying biology of Alzheimer's disease is present, with high accuracy when interpreted using validated assay-specific cutoffs.
• The presence and rough degree of tau pathology, through the p-tau measurements.
• Whether features more consistent with other neurological diagnoses are present, when the broader CSF panel is informative for the differential.

They cannot tell you:

• Whether someone has Alzheimer's disease as a clinical condition independent of the rest of the picture. CSF biomarker positivity can occur in people without symptoms, and the clinical reading still depends on the whole picture.
• The exact future trajectory for an individual patient.
• The right answer to most of the personal questions families ask after the diagnostic workup.

Where this fits in the broader diagnostic pathway

The contemporary Alzheimer's diagnostic pathway in the US is increasingly tiered: clinical evaluation first, plasma biomarker triage second, and confirmatory testing - by amyloid PET or CSF - when treatment decisions or differential complexity warrant it. CSF testing is no longer the routine first-line confirmatory step in most settings, but it remains a well-validated and clinically useful option. The full landscape is in our snapshot of diagnostic pathways for 2026, and the broader shift toward less-invasive testing is the subject of our Insight on how blood biomarkers will reshape diagnosis.

This page is a plain-language primer. It is not medical advice. Decisions about diagnostic testing belong with the patient and their clinical team.