CSF Amyloid Beta 42 Unit Converter

Convert CSF Aβ42 between pg/mL, ng/L, µg/L, pmol/L and nmol/L — with the cut-off for each assay platform printed beside the platform, because they differ by more than half.

Last updated 8 Oct 2026

CSF Amyloid Beta 42 converter

Mass ⇄ molar

CSF amyloid beta 42

pg/mL and ng/L are the same number, so a converter that offered only those two would do nothing. The useful conversion is to pmol/L, and it needs the peptide’s exact mass.

184.75pmol/LExample

CSF Aβ42 of 834 pg/mL, which is Mayo’s amyloid-negative cut-off on Roche Elecsys

The conversion, and the mass behind it

pmol/L = pg/mL × 0.221528
pg/mL = pmol/L ÷ 0.221528
ng/L = pg/mL × 1 — the same quantity written two ways
µg/L = pg/mL ÷ 1000

4,514.1: the molar mass of Aβ42, C203H311N55O60S, a 42-residue peptide with an exact composition. PubChem CID 57339251; the formula sum from standard atomic weights gives 4,514.106 and the vendor datasheets give 4,514.08
pg/mL = ng/L: identical in magnitude, which is why a page offering only those two would be a converter that converts nothing. The molar column is the one that earns the page
not tau: the same arithmetic cannot be done for total tau. Six CNS isoforms span 36.8 to 45.9 kDa, the UniProt canonical entry is 78.9 kDa, and the assays measure mid-domain fragments of all of them, so there is no mass to divide by

Worked example

CSF Aβ42 of 834 pg/mL, which is Mayo's amyloid-negative cut-off on Roche Elecsys
834 × 0.221528 = 184.75 pmol/L
834 pg/mL = 834 ng/L = 0.8340 µg/L
The same 834 pg/mL means amyloid-negative on Elecsys and comfortably amyloid-negative on Lumipulse, whose own cut-off is 526 ng/L — a 58% difference in the threshold for one molecule
Which is why the ratio to Aβ40 rather than the absolute value is the Core 1 biomarker

Aβ42 cut-offs, each beside the platform it belongs to

A 58% difference between two CE-marked and FDA-cleared assays measuring the same peptide, and a further 31.5% step inside one of them when it was recalibrated to certified reference material in December 2018 — the same sample, the same instrument, a third less amyloid. An absolute Aβ42 figure is uninterpretable without the assay name and the calibration generation. CSF amyloid and tau cut-offs are platform-specific. Elecsys, Lumipulse, INNOTEST and Simoa do not agree on absolute concentrations, so a threshold is only meaningful beside the name of the assay that produced it.
Platform	Amyloid-negative when Aβ42 is	In pmol/L	Source
Roche Elecsys (cobas ECLIA)	above 834 pg/mL	above 184.75	Mayo ADEVL
Fujirebio Lumipulse G	above 526 ng/L	above 116.52	Clin Chem Lab Med 2021
Lumipulse, before the Dec 2018 recalibration	about 31.5% higher than the figure above	—	Clin Chem Lab Med 2021
Any platform, absolute value	not transferable	—	—

The rest of the panel, and which figures are ratios

Two of the four analytes can be converted to molar units and two cannot, and the division is not arbitrary: the amyloid peptides have exact compositions and the tau species do not. The two ratios need no conversion at all, which is part of why the 2024 Alzheimer’s Association criteria are written around ratios rather than absolute concentrations.
Biomarker	Normal direction (Elecsys, Mayo ADEVL)	Convertible to molar?
Aβ42	above 834 pg/mL	Yes — exact 42-residue composition
Aβ40	no interval published	Yes — exact 40-residue composition
Total tau	at or below 238 pg/mL	No — isoforms and fragments
p-tau181	at or below 21.6 pg/mL	No — same reason
p-tau181/Aβ42	at or below 0.028	unitless, so nothing to convert
Aβ42/Aβ40	see the ratio calculator	unitless, but it is a mass ratio

Converting Aβ42, and the reason the absolute value is the weak form

Aβ42 is a 42-residue cleavage product of amyloid precursor protein, and it is the species that aggregates into plaque. The CSF concentration falls as plaque forms, because the peptide is sequestered into the deposits rather than remaining soluble, so a low CSF Aβ42 is the abnormal result — the opposite direction from almost every other biomarker a reader will have met.

The conversion itself is well founded. Unlike tau, Aβ42 has an exact molecular formula, C203H311N55O60S, giving 4,514.1 Da, so pg/mL to pmol/L is a real mass-to-amount conversion and not a calibration. It is worth doing precisely because pg/mL and ng/L are the same number: a report in ng/L needs no arithmetic, and the only conversion that tells a reader anything is the molar one.

What the conversion cannot fix is that the absolute concentration is not comparable between assays. Mayo’s Elecsys-based evaluation puts the amyloid-negative boundary above 834 pg/mL; the published Lumipulse validation puts it above 526 ng/L, which is the same unit and a 58% lower number. Within Lumipulse alone, the December 2018 recalibration to certified reference material lowered reported values by about 31.5%, so a result from 2017 and a result from 2019 on the same instrument are not the same quantity. Pre-analytical handling adds to this: Aβ42 adsorbs to tube walls, so tube type and fill volume move the number too.

That is why the 2024 Alzheimer’s Association criteria are built on the Aβ42/Aβ40 ratio rather than on Aβ42 alone. Dividing by Aβ40 cancels most of what varies — total amyloid production, which differs severalfold between individuals, and adsorption, which affects both peptides similarly. The absolute Aβ42 remains useful for following one patient on one platform. It is the wrong number to carry between laboratories, and converting it to pmol/L does not make it transferable; it only makes it unambiguous.

Frequently asked questions

How do I convert CSF Aβ42 from pg/mL to pmol/L?

Multiply by 0.221528. The factor comes from Aβ42’s molar mass of 4,514.1 g/mol, derived from the formula C203H311N55O60S. An Aβ42 of 834 pg/mL is 184.75 pmol/L.

Are pg/mL and ng/L the same for Aβ42?

Yes — numerically identical, so a result of 834 pg/mL and one of 834 ng/L mean exactly the same thing. US laboratories tend to print pg/mL and European ones ng/L. This is why a converter offering only those two units would be doing nothing at all, and why the molar column is the point of this page.

What Aβ42 level indicates Alzheimer’s disease?

There is no platform-independent answer, which is the honest one. Mayo’s Elecsys-based evaluation treats above 834 pg/mL as amyloid-negative; the published Lumipulse validation uses above 526 ng/L — the same unit, a 58% lower threshold. Use your own laboratory’s cut-off for your own laboratory’s assay, and prefer the Aβ42/Aβ40 ratio, which is far less platform-dependent.

Why can Aβ42 be converted to molar units but total tau cannot?

Because Aβ42 is one peptide with one composition and tau is not. Tau exists as six CNS isoforms spanning 352 to 441 residues (36.8 to 45.9 kDa), the UniProt canonical entry is 758 residues (78.9 kDa), and the t-tau immunoassays use mid-domain antibody pairs that capture truncated fragments rather than any one intact species. Dividing a pg/mL figure by 45,900 would produce a number with a unit on it and nothing behind it.

Does a low Aβ42 mean the plaque is in the brain?

A low CSF Aβ42 reflects amyloid being sequestered into deposits, and in the Lumipulse ratio validation 97% of ratio-positive samples had amyloid plaque on PET. It is a biomarker of amyloid pathology, not of dementia: amyloid is present in cognitively normal people, and the 2024 criteria stage the disease on symptoms separately from the biomarkers.

Related calculators

References

Mayo Clinic Laboratories. Alzheimer disease evaluation, spinal fluid (ADEVL) — reference values for beta-amyloid (1-42), total tau, p-tau181 and the p-tau/Abeta42 ratio, by ECLIA on Roche cobas. Test catalogue, test ID 607273; 2026.
Validation of the Lumipulse automated immunoassay for the measurement of core AD biomarkers in cerebrospinal fluid. Clin Chem Lab Med. 2021; doi:10.1515/cclm-2021-0651.
PubChem. Human beta-amyloid peptide (1-42), CID 57339251, C203H311N55O60S. National Library of Medicine; accessed October 2026.
R&D Systems (Tocris) 1428. Amyloid beta-peptide (1-42), human — C203H311N55O60S, 4514.08 g/mol, CAS 107761-42-2. Product datasheet; accessed October 2026.
Jack CR Jr, Andrews JS, Beach TG, et al. Revised criteria for diagnosis and staging of Alzheimer’s disease: Alzheimer’s Association Workgroup. Alzheimers Dement. 2024. doi:10.1002/alz.13859.

Not medical advice. For healthcare professionals and education. Reference intervals vary by laboratory and assay — always use your own laboratory's. Never base a dose or a treatment decision on this page alone. Full disclaimer at calcengines.com/disclaimer/