Recei ed: 10 July 2025 Re ised: 18 Augus 2025 Accep ed: 25 Augus 2025
DOI: 10.1002/alz.70717
RESEARCH ARTICLE
Plasma p- au217, N L, GFAP diagnos ic pe o mance and
bioma ke p o iles in Alzheime ’s disease, on o empo al
demen ia, and psychia ic diso de s, in a p ospec i e
unselec ed neu opsychia y memo y clinic
Dhamidhu E a ne1,2Ma hew Kang1,2Cha les B Malpas3,4Ch is a Dang5,6
Cou ney Lewis7Oneil G Bhalala8,9Qiao-Xin Li10 S e en Collins10
Colin L Mas e s10 Saman ha M Loi1,2Alexande F San illo11
Kaj Blennow12,13,14 Hen ik Ze e be g13,15,16,17 Dennis Velakoulis1,2And The
MiND S udy G oup
1Neu opsychia y Cen e, Royal Melbou ne Hospi al, Melbou ne, Vic o ia, Aus alia
2Depa men o Psychia y, Uni e si y o Melbou ne, Melbou ne, Vic o ia, Aus alia
3Depa men o Medicine (Royal Melbou ne Hospi al), Uni e si y o Melbou ne, Melbou ne, Vic o ia, Aus alia
4Melbou ne School o Psychological Sciences, Uni e si y o Melbou ne, Melbou ne, Vic o ia, Aus alia
5Na ional Ageing Resea ch Ins i u e, Melbou ne, Vic o ia, Aus alia
6Depa men o Gene al P ac ice, Uni e si y o Melbou ne Vic o ia, Melbou ne, Vic o ia, Aus alia
7Ins i u e o Heal h and Wellbeing, Fede a ion Uni e si y, Melbou ne, Vic o ia, Aus alia
8Depa men o Medicine, Royal Melbou ne Hospi al, Melbou ne, Vic o ia, Aus alia
9Gene ics and Gene Regula ion Di ision, The Wal e and Eliza Hall Ins i u e o Medical Resea ch, Melbou ne, Vic o ia, Aus alia
10Na ional Demen ia Diagnos ics Labo a o y, The Flo ey, Melbou ne, Vic o ia, Aus alia
11Clinical Memo y Resea ch Uni , Depa men o Clinical Sciences, Facul y o Medicine, Lund Uni e si y, Lund, Sweden
12Ins i u e. o Neu oscience and Physiology, Uni e si y o Go henbu g, Mölndal, Sweden
13Clinical Neu ochemis y Labo a o y, Sahlg enska Uni e si y Hospi al, Mölndal, Sweden
14Neu odegene a i e Diso de Resea ch Cen e , Di ision o Li e Sciences and Medicine, and Depa men o Neu ology, Ins i u e on Aging and B ain Diso de s,
Uni e si y o Science and Technology o China and Fi s A ilia ed Hospi al o USTC, He ei, Anhui, P.R. China
15Depa men o Psychia y and Neu ochemis y, Ins i u e o Neu oscience and Physiology, he Sahlg enska Academy a he Uni e si y o Go henbu g, Mölndal,
Sweden
16Depa men o Pa hology and Labo a o y Medicine, Uni e si y o Wisconsin School o Medicine and Public Heal h, Uni e si y o Wisconsin-Madison, Madison,
Wisconsin, USA
17Wisconsin Alzheime ’s Disease Resea ch Cen e , Uni e si y o Wisconsin School o Medicine and Public Heal h, Uni e si y o Wisconsin-Madison, Madison,
Wisconsin, USA
Co espondence
Dhamidhu E a ne, Neu opsychia y Cen e,
The Royal Melbou ne Hospi al, 300 G a an S ,
Pa k ille VIC, Melbou ne, Aus alia.
Abs ac
INTRODUCTION: Plasma bioma ke s o e p omise o imp o ing he diagnosis o
Alzheime ’s disease (AD) and di e en ia ing AD and o he neu odegene a i e diso -
This is an open access a icle unde he e ms o he C ea i e Commons A ibu ion License, which pe mi s use, dis ibu ion and ep oduc ion in any medium, p o ided
he o iginal wo k is p ope ly ci ed.
© 2025 The Au ho (s). Alzheime ’s & Demen ia published by Wiley Pe iodicals LLC on behal o Alzheime ’s Associa ion.
Alzheime ’s Demen . 2025;21:e70717. wileyonlinelib a y.com/jou nal/alz 1o 15
h ps://doi.o g/10.1002/alz.70717
2o 15 ERATNE ET AL.
Email: [email p o ec ed]
Funding in o ma ion
Swedish Resea ch, G an /Awa d Numbe s:
2017-00915, 2022-00732; Swedish Alzheime
Founda ion, G an /Awa d Numbe s:
AF-930351, AF-939721, AF-968270,
AF-994551; Hjä n onden, Sweden,
G an /Awa d Numbe s: ALZ2022-0006,
FO2024-0048-TK-130, FO2024-0048-HK-24;
ALF ag eemen , G an /Awa d Numbe s:
ALFGBG-965240, ALFGBG-1006418;
Eu opean Union Join P og amme o
Neu odegene a i e Diso de s, G an /Awa d
Numbe : JPND2019-466-236; Alzheime ’s
Associa ion 2021 Zeni h Awa d, G an /Awa d
Numbe : ZEN-21-848495; Alzheime ’s
Associa ion 2022-2025, G an /Awa d
Numbe : SG-23-1038904 QC; he Swedish
Resea ch Council, G an /Awa d Numbe s:
2023-00356, 2022-01018, 2019-02397;
Eu opean Union’s Ho izon Eu ope,
G an /Awa d Numbe : 101053962; Swedish
S a e Suppo o Clinical Resea ch,
G an /Awa d Numbe : ALFGBG-71320;
Alzheime D ug Disco e y Founda ion
(ADDF), G an /Awa d Numbe :
201809-2016862; AD S a egic Fund and he
Alzheime ’s Associa ion, G an /Awa d
Numbe s: ADSF-21-831376-C,
ADSF-21-831381-C, ADSF-21-831377-C,
ADSF-24-1284328-C; Eu opean Union’s
Ho izon Eu ope Resea ch and Inno a ion
P og amme; NEu oBioS and, G an /Awa d
Numbe : 22HLT07; E ling–Pe sson Family
Founda ion, G an /Awa d Numbe :
FO2022-0270; Eu opean Union’s Ho izon
2020, G an /Awa d Numbe : 860197;
Eu opean Union Join
P og amme–Neu odegene a i e Disease
Resea ch, G an /Awa d Numbe :
JPND2021-00694; UK Demen ia Resea ch
Ins i u e, G an /Awa d Numbe : UKDRI-1003;
AFS, G an /Awa d Numbe : ALF 2022 YF 0017
de s (NDs) like on o empo al demen ia (FTD) om p ima y psychia ic diso de s
(PPDs), pa icula ly in younge pa ien s.
METHODS: In his p ospec i e s udy, we in es iga ed plasma phospho yla ed au 217
(p- au217), neu o ilamen ligh chain (N L), and glial ib illa y acidic p o ein (GFAP)
in 341 unselec ed pa icipan s om a neu opsychia y memo y clinic, including AD
(n=40),beha io al a ian FTD(b FTD)(n=15),PPD(n=69),o he NDs,andcon ols.
RESULTS: Plasma p- au217 showed s ong diagnos ic pe o mance o dis inguishing
AD om b FTD (96% accu acy) and PPD (93% accu acy). N L bes dis inguished all
NDs om PPD, while GFAP did no b ing addi ional alue. Bioma ke p o iles using
p ede ined cu -o s and age-adjus ed z-sco es u he cla i ied g oup di e ences.
DISCUSSION: Plasma p- au217 and N L ha e s ong diagnos ic u ili y in eal-wo ld,
diagnos ically complex coho s. These indings suppo implemen a ion o scalable
blood-based bioma ke s o imp o e ea ly and accu a e diagnosis in memo y clinical
se ings.
KEYWORDS
age-adjus ed e e ence anges, Alzheime ’s disease, blood-based bioma ke s, diagnos ic accu-
acy, on o empo al demen ia, GFAP, memo y clinic, neu odegene a ion, neu o ilamen ligh
chain, plasma bioma ke s, p ima y psychia ic diso de s, p ospec i e coho , psychia ic misdiag-
nosis, p- au217
Highligh s
∙Plasma p- au217 was signi ican ly ele a ed in AD compa ed o o he diso de s.
∙P- au217 dis inguished AD om b FTD wi h high accu acy.
∙P- au217 dis inguished AD om PPDs wi h high accu acy.
∙N L/p- au217 a io and GFAP added limi ed diagnos ic alue compa ed o p- au217
and N L.
∙Findings suppo blood bioma ke s in younge , eal-wo ld clinical coho s.
1BACKGROUND
Blood-based bioma ke s show g ea po en ial o ans o m diagno-
sis and clinical ca e o pa ien s p esen ing wi h cogni i e, psychia ic,
and neu ological symp oms. Thei po en ial includes imp o ing ea ly
and accu a e diagnosis o Alzheime ’s disease (AD) and esol ing eal-
wo ld diagnos ic challenges, such as dis inguishing AD om o he
neu odegene a i e diso de s (NDs) like beha io al a ian on o em-
po al demen ia (b FTD) and om common mimics such as p ima y
psychia ic diso de s (PPDs).1–6 Ea ly, accu a e diagnosis is inc eas-
ingly impo an wi h he eme gence o disease-speci ic ea men s,
such as monoclonal an ibodies a ge ing amyloid be a.
Two o he mos p omising bioma ke s a e plasma phospho y-
la ed au 217 (p- au217) and neu o ilamen ligh chain (N L) p o ein.
P- au217 shows s ong diagnos ic pe o mance and speci ici y o dis-
inguish AD om non-AD diso de s (including non-AD demen ias and
PPD).7–10 Wi hin AD and con ol coho s, plasma le els o p- au217
co ela e wi h au posi on emission omog aphy (PET) pa hology11
and angleloada au opsy
12 and a e ele a ed in diso de s wi h an-
gles bu wi hou amyloid plaques,13 suppo ing i s e lec ion o au
phospho yla ion and pa hology. N L, a non-speci ic ma ke o neu ode-
gene a ion and acu e neu onal inju y, pa icula ly in long myelina ed
axons, dis inguishes neu odegene a i ediso de s(NDs) om ND mim-
ics, including PPDs and non-NDs,3,4,14,15 and has u ili y in clinical
se ings.14,16–19 Glial ib illa y acidic p o ein (GFAP), a ma ke o as o-
cy ic ac i a ion and neu oin lamma ion, is an addi ional bioma ke o
in e es in a ange o diso de s.14,20–26 Howe e , he u ili y o GFAP
o he clinical di e en ia ion be ween AD, non-AD NDs, and PPDs,
especially when compa ed o p- au217 and N L is no ye clea .
We p e iously published diagnos ic pe o mance, cu -o s, and e -
e ence anges o plasma p- au217 and N L,10,14,18,27–29 bu we ha e
no in es iga ed hei combined use. Fu he mo e, while we and o h-
ERATNE ET AL.3o 15
e s ha e published da a showing he supe io i y o N L compa ed o
GFAP o dis inguish b FTD om PPDs,30,31 we ha e no ye in es-
iga ed all h ee, p- au217, N L, and GFAP, in a younge eal-wo ld
clinical neu opsychia ic coho wi h di e se NDs and PPDs. Some
s udies ha e in es iga ed combina ions o hese ma ke s. Benussi e al.
in es iga ed N L, p- au217, N L/p- au217 a ios, and o he bioma k-
e s in 374 pa icipan s (97 AD, 278 FTD).32 They ound s ong
diagnos ic pe o mance o p- au217 and sligh ly supe io pe o -
mance o N L/p- au217 o dis inguish be ween AD and FTD. Howe e ,
hey lacked a PPD g oup, which is c i ical o app ecia e he use o
bioma ke s o b FTD diagnosis. Rousse e al. assessed p- au217
and N L in an unselec ed memo y clinic coho , showing s ong u il-
i y o p- au217 o AD diagnosis and desc ibing bioma ke p o iles
(high/low p- au217 and N L) ac oss clinical g oups.17 To ou knowl-
edge, no s udy o da e has join ly in es iga ed p- au217, N L, GFAP,
and bioma ke p o iles, wi h a speci ic ocus on AD, b FTD, PPD,
in younge people, whe e di e en ial diagnoses, diagnos ic unce -
ain y, misdiagnosis, and delay a e g ea e .1,33–35 Fu he esea ch
is needed o p ope ly es ablish he oles o single and combina ion
bioma ke s in eal-wo ld clinicalse ings wi h di e seneu opsychia ic
coho s.
The p ima y aim o his s udy was o compa e le els and diagnos-
ic pe o mances o plasma p- au217, N L, N L/p- au217 a io, and
GFAP o dis inguishing be ween AD, b FTD, and PPDs. We hypo h-
esized ha combining p- au217 and N L would imp o e diagnos ic
accu acy compa ed o ei he alone. We also aimed o desc ibe p-
au217/N L bioma ke p o iles o AD pa hology and neu onal inju y
(e.g., low AD/low neu onal inju y, high AD/high neu onal inju y) in
an unselec ed popula ion o pa ien s om a neu opsychia y memo y
clinic. We ocused on p- au217 and N L using ou p e iously desc ibed
cu -o o plasma p- au21710 and age-adjus ed z-sco e e e ence
ange models o plasma N L.27,30 We wan ed o explo e whe he
GFAP could imp o e any dis inc ions, o example, be ween b FTD and
PPDs. Al hough some s udies epo ed age-binned e e ence anges
o GFAP in se um36 and plasma,37 we a e unawa e o any s udies ha
ha e desc ibed GFAP e e ence anges using con inuous modeling o
de i e p ecise age-adjus ed pe cen iles and z-sco es, as we did o N L.
As an explo a o y aim, we applied a no el z-sco e model o plasma
GFAP o de ine bioma ke p o iles e lec ing AD pa hology, neu onal
inju y, and neu oin lamma ion (p- au217/N L/GFAP).
2METHODS
The cu en s udy included pa icipan s p ospec i ely ec ui ed
be ween June 2019 and Ap il 2023 who had p o ided a blood sample
o bioma ke analysis. This is a ollow-up o ou p e ious s udy coho
in which we in es iga ed plasma and ce eb ospinal luid (CSF) N L.14 In
he cu en s udy, we included pa icipan s who had plasma p- au217,
N L, and GFAP bioma ke esul s a ailable. In he cu en s udy he
ocus was on AD, b FTD, and PPD diagnos ic g oups, wi h compa a-
o g oups being con ol pa icipan s and o he NDs (bo h desc ibed in
he p e ious s udy). Da a we e a ailable in he cu en s udy o addi-
RESEARCH IN CONTEXT
1. Sys ema ic e iew: We e iewed he PubMed li e a u e
on plasma p- au217, N L, and GFAP in AD, FTD, and
PPDs. Al hough se e al s udies suppo he indi idual
diagnos ic alue o p- au217 and N L, ew examined all
h ee bioma ke s oge he in eal-wo ld, diagnos ically
he e ogeneous clinical coho s, including PPDs.
2. In e p e a ion: This p ospec i e s udy demons a es he
s ong diagnos ic pe o mance o plasma p- au217 in dis-
inguishing AD om FTD and PPDs and o he non-AD
diso de sandcon i ms N Las ause ul ma ke o neu ode-
gene a ion. GFAP added limi ed alue. These indings add
o g owing e idence on he clinical applica ion o blood-
based bioma ke s in younge and diagnos ically complex
popula ions.
3. Fu u e di ec ions: Fu u e esea ch should alida e hese
indings in la ge , mo e di e se coho s, de elop mul-
imodal diagnos ic algo i hms, and e alua e he clini-
cal u ili y and cos -e ec i eness o implemen ing blood
bioma ke es ing in psychia ic, p ima y ca e, and mem-
o y clinic se ings.
ional compa a o g oups (no p e iously desc ibed): people wi h mild
cogni i e impai men (MCI) and people wi h p esymp oma ic gene ic
NDs. The b FTD g oup included pa ien s mee ing diagnos ic c i e ia
o possible b FTD. The pa icipan s in his s udy did no o e lap wi h
ou p e ious s udy on N L and GFAP in b FTD, mood, and psycho ic
diso de s.30
Pa icipan s we e ec ui ed om he Neu opsychia y Cen e a
The Royal Melbou ne Hospi al, a qua e na y se ice ecei ing e e -
als o diagnos ically complex cases om p ima y ca e and o he
specialis se ices wi hin Aus alia. Pa ien s, as pa o ou ine clini-
cal ca e h ough he Neu opsychia y Cen e, ecei ed comp ehensi e
mul idisciplina y assessmen s and mul imodal in es iga ions, including
CSF AD bioma ke analysis, wi h gold s anda d consensus diagnosis
based on es ablished diagnos ic c i e ia, as p e iously desc ibed in
de ail.14,18,28 Diagnos ic g oup ca ego iza ion was de e mined based
on he mos ecen diagnosis, a longi udinal ollow-up, based on es ab-
lished diagnos ic c i e ia, blinded o plasma bioma ke le els, as p e-
iously desc ibed.14,18,28 Con ol pa icipan s we e people ec ui ed
om he communi y. Cogni i e sc eening da a a e s ill being collec ed
and no ye a ailable o his s udy; howe e , no con ol pa icipan s
had symp oms o diagnoses o neu ological o NDs and no ac i e
psychia ic symp oms o condi ions.
EDTA plasma samples we e collec ed du ing pa ien s’ diagnos ic
wo k-up and a i s isi o communi y con ols. Samples we e s o ed
a −80◦C. Plasma N L and GFAP we e measu ed using N2PB ki s
on a Quan e ix Single molecule a ay (Simoa) HD-X analyze , acco d-
ing o he manu ac u e ’s ins uc ions (Quan e ix Co p., Bille ica, MA,
4o 15 ERATNE ET AL.
USA). Plasma p- au217 was measu ed using an in-house Uni e si y
o Go henbu g (UGOT) p- au217 assay, as p e iously desc ibed in
de ail.38 The measu emen s we e pe o med in one ound o expe i-
men s using one ba ch o eagen s by analys s blinded o clinical da a
and diagnoses, he eby educing po en ial ba ch e ec s.
This s udy, pa o he Ma ke s in Neu opsychia ic Diso de s S udy
(The MiND S udy, h ps:// heminds udy.o g), was app o ed by he
human esea ch e hics commi ee a Melbou ne Heal h (2016.038,
2017.090, 2018.371, 2020.142).
2.1 S a is ical analyses
S a is ical analyses we e pe o med using R e sion 4.5.0 (2025-
04-11). Bioma ke le els in di e en g oups we e compa ed using
s anda dized boo s apped gene al linea models (GLMs), wi h age
a blood sample and sex as addi ional co a ia es. Recei e ope a ing
cha ac e is ic (ROC) cu e analyses we e hen pe o med o in es-
iga e diagnos ic u ili y be ween di e en combina ions o g oups.
Boo s apped di e ences in a ea unde he cu e (AUC) we e used
o compa e ROC cu es. Op imal cu -o s we e selec ed based on
Youden’s Js a is ic. Addi ional diagnos ic es pa ame e s we e com-
pu ed: posi i e and nega i e likelihood a ios, posi i e and nega i e
p edic i e alues, o e all accu acy, and diagnos ic odds a io. Addi-
ional sensi i i y analyses we e pe o med: excluding ex eme ou lie s,
excluding pa ien s wi h C eu z eld –Jakob disease (CJD), and pe o m-
ing all GLMs wi h weigh included as a co a ia e. As esul s we e
simila , he esul s excluding weigh and including ou lie s and pa ien s
wi h CJD we e p esen ed o maximize he sample sizes o analyses
and p esen ed esul s (since no all pa icipan s had weigh da a).
We ocused on desc ibing p- au217 and N L bioma ke p o iles.
These we e low p- au217/low N L (a low AD pa hology and low neu-
onal inju y p o ile – a “no mal” p o ile) and abno mal p o iles: low
p- au217/high N L (low AD pa hology bu ele a ed neu onal inju y),
high p- au217/low N L (AD pa hology bu low neu onal inju y), and
high p- au217/high N L (high AD pa hology and high neu onal inju y).
Plasma p- au217 and N L bioma ke le els we e dicho omized in o
“high” and “low” based on ou p e iously published da a and cu -o s
ha used he same assays.10,27 Fo p- au217, we used a cu -o o
2.35, which was op imal a dis inguishing AD om non-AD, as p e-
iously desc ibed and published.10 Fo plasma N L, gi en he s ong
non-linea associa ion wi h age, we used age-based pe cen iles and
z-sco es de i ed om he gene alized addi i e models o loca ion,
scale, and shape (GAMLSS) model o a la ge e e ence con ol coho
ha we de eloped, p e iously desc ibed and published,14 de ining he
95 h pe cen ile as he cu -o be ween high and low. Fo his s udy,
we c ea ed a no el age-adjus ed GFAP e e ence ange model using
(GAMLSS) using his s udy’s con ol g oup. This allowed us o de i e
p ecise age-based pe cen iles and z-sco es and hus mo e p ecise ca e-
go iza ion,again gi en he signi ican s ong and non-linea associa ion
wi h age compa ed o coa se age-binned cu -o s. The 95 h pe cen ile
was de ined as he cu -o o high and low GFAP le els.
3RESULTS
The inal coho consis ed o 341 pa icipan s: 40 wi h AD (median
age 62 yea s, 53% emale), 15 wi h b FTD (median age 57 yea s,
27% emale), and 69 wi h PPDs (median age 55 yea s, 48% emale).
The e we e 119 con ols (median age 63, 74% emale), and in he
addi ional compa a o g oups: 67 wi h o he NDs, 13 wi h MCI,
and 18 wi h p esymp oma ic gene ic NDs (see Table 1 o ull
de ails).
PPDs consis ed o majo dep essi e diso de (MDD, n=21),
bipola diso de (n=6), unc ional neu ological/cogni i e diso de
(n=8), schizoph enia spec um diso de (n=16), b FTD “pheno-
copy” synd ome (n=2), and o he PPDs (n=16, which included
anxie y, pe sonali y, obsessi e-compulsi e, pos - auma ic s ess, and
undi e en ia ed psychia ic diso de s).
O he NDs consis ed o CJD (n=2), demen ia wi h Lewy bod-
ies (n=5), demen ia no o he wise speci ied (n=6), Hun ing-
on’s disease (HD, n=15), mixed AD/ ascula (n=3), subs ance-
ela ed cogni i e impai men /demen ia (n=2), ascula demen ia
(n=5),anda angeo o he NDs(n=29, which included au oim-
mune encephali is, ce eb al amyloid angiopa hy, co icobasal syn-
d ome, cen al ne ous sys em asculi is, Down synd ome, Fah dis-
ease, me abolic diso de s, Niemann–Pick Type C, Pa kinson’s disease,
and ce ebella degene a i e diso de ). The p esymp oma ic gene ic
ND g oup consis ed o AD (PSEN1,n=2), gene ic CJD (n=6),
HD (n=6), CADASIL (n=1), b FTD (C9o 72 n =1andGRN
n=2).
CSF bioma ke analysis, including AD p o eins CSF AB42 and
p- au181 (Suppo ing In o ma ion Table 1), was conduc ed on 80
pa ien s. B ie ly, 27/40 (68%) o AD pa ien s had CSF AD bioma ke s.
Mos (18/27, 67%) had a CSF AD bioma ke p o ile consis en wi h AD,
as de ined by an amyloid-posi i e, p- au-posi i e (A+T+) p o ile, ca e-
go izedusing es ablished cu -o s,as desc ibed in ou p e iouss udy.10
O he p o iles in he AD g oup we e 8/27 wi h A+T−and 1/27 wi h
A−T+. Two pa ien s wi h AD had amyloid PET (bu no CSF), and bo h
we e posi i e o amyloid plaques. Fu he in o ma ion, including phe-
no ype in o ma ion o AD, a e in he Suppo ing In o ma ion Tables
1and 2. Fo y-se en pe cen (7/15) o b FTD pa ien s had CSF AD
bioma ke s (none wi h an A+T+p o ile, 4/7 wi h A−T−,and3/7wi h
A+T−). None o he pa ien s in he o he g oups who had CSF AD
bioma ke s (20 in he PPD g oup, 21 in o he NDs, and i e MCI) had
A+T+p o iles ( u he de ails in Suppo ing In o ma ion Tables 1and
2).
3.1 Le els o plasma p- au217, N L, N L/p- au217
a io, GFAP, in AD, b FTD, and PPD
3.1.1 Plasma p- au 217
As demons a ed in Table 1and Figu e 1, plasma p- au217 le els we e
signi ican ly ele a ed in AD compa ed o b FTD (s anda dized boo -
ERATNE ET AL.5o 15
TABLE 1 S udy coho de ails and bioma ke le els.
Cha ac e is ic N
AD
N=40a
b FTD
N=15a
PPD
N=69a
Con ol
N=119a
O he ND
N=67a
MCI
N=13a
P esymp oma ic
gene ic ND
N=18a
Age 341 62 (58, 65) 57 (56, 62) 55 (45, 62) 63 (55, 70) 61 (45, 67) 65 (56, 67) 51 (43, 62)
Sex 341
Female 21/40 (53%) 4/15 (27%) 33/69 (48%) 88/119 (74%) 29/67 (43%) 3/13 (23%) 13/18 (72%)
Male 19/40 (48%) 11/15 (73%) 36/69 (52%) 31/119 (26%) 38/67 (57%) 10/13 (77%) 5/18 (28%)
Weigh 270 73 (59, 83) 84 (63, 101) 84 (73, 99) 75 (65, 85) 75 (65, 88) 84 (78, 90) 79 (70, 98)
Unknown 15 2 14 16 12 4 8
p- au217 341 3.63 (2.90, 4.41) 1.07 (0.72, 1.50) 0.92 (0.58, 1.32) 0.91 (0.65, 1.31) 1.11 (0.69, 1.68) 1.04 (0.78, 1.40) 0.83 (0.65, 1.25)
N l 341 24 (18, 28) 21 (12, 55) 11 (8, 13) 12 (9, 17) 29 (16, 42) 15 (13, 20) 12 (10, 19)
N L/p- au217 a io 341 7 (5, 9) 25 (10, 48) 12 (9, 20) 13 (9, 22) 24 (11, 47) 16 (12, 27) 14 (11, 27)
GFAP 341 212 (151, 305) 79 (53, 192) 86 (56, 117) 115 (89, 177) 132 (77, 216) 144 (86, 174) 97 (61, 125)
log p- au217 341 0.56 (0.46, 0.64) 0.03 (−0.14, 0.18) −0.04 (−0.24, 0.12) −0.04 (−0.19, 0.12) 0.05 (−0.16, 0.23) 0.02 (−0.11, 0.15) −0.08 (−0.19, 0.10)
log N L 341 1.38 (1.25, 1.45) 1.31 (1.08, 1.74) 1.02 (0.90, 1.11) 1.09 (0.94, 1.24) 1.46 (1.20, 1.62) 1.19 (1.11, 1.30) 1.06 (0.99, 1.27)
log GFAP 341 2.33 (2.18, 2.48) 1.90 (1.72, 2.28) 1.93 (1.75, 2.07) 2.06 (1.95, 2.25) 2.12 (1.88, 2.33) 2.16 (1.93, 2.24) 1.99 (1.78, 2.10)
Abb e ia ions:AD, Alzheime ’s disease; b FTD, beha io al a ian on o empo al demen ia; GFAP, glial ib illa y acidic p o ein; MCI, mild cogni i e impai men ; N L, neu o ilamen ligh chain; ND, neu odegene a i e
diso de ; PPD, p ima y psychia ic diso de ; p- au217, phospho yla ed au 217
aMedian (Q1, Q3); n/N(%).
6o 15 ERATNE ET AL.
(A) (B)
(C) (D)
FIGURE 1 Plasma p- au217, N L, N L/p- au217 a io, and GFAP le els, in AD, b FTD, PPDs, con ols, and o he compa a o g oups. Plasma
p- au217 le els we e signi ican ly ele a ed in AD compa ed o all o he diso de s and con ols. To imp o e eadabili y, se en ou lie s we e no
displayed in plo B o N L ( wo o AD [141 and 294 pg/mL], wo b FTD [101 and 214 pg/mL], h ee o he NDs [106, 352, and 1154 pg/mL]), and
six in plo C o N L/p- au217 a io (one b FTD [143], wo con ols [699, 772], h ee o he NDs [190, 198, and 604]). Dashed ed lines =p ede ined
op imal cu -o s om ou p e ious s udies, o p- au217 (2.35pg/mL). AD, Alzheime ’s disease; b FTD, beha io al a ian on o empo al
demen ia; GFAP, glial ib illa y acidic p o ein; ND, neu odegene a i e diso de ; N L, neu o ilamen ligh chain; PPD, p ima y psychia ic diso de ;
p- au217, phospho yla ed au 217.
s apped GLM wi h age and sex as addi ional co a ia es; β=1.30, 95%
CI: [0.71, 1.73], p<0.001), and AD compa ed o PPDs (β=1.59 [1.34,
1.80], p<0.001). Le els we e also ele a ed in AD compa ed o he
o he g oups (con ols, O he NDs, MCI, and P esymp oma ic ND, and
all p<0.001).
3.1.2 Plasma N L
Plasma N L was ele a ed in AD compa ed o PPD (β=0.87 [0.59, 1.13],
p<0.001), bu no o AD compa ed o b FTD (β=−0.10 [−0.82, 0.58],
p=0.788). The e we e no signi ican di e ences in plasma N L le -
els ac oss he ND g oups (AD, b FTD, O he NDs, all p>0.05). Le els
we ealsosimila be weenPPDs,con ol,MCI,andp esymp oma icND
g oups.
3.1.3 Plasma N L/p- au217 a io
The plasma N L/p- au217 a io was educed in AD compa ed o b FTD
(β=−1.57 [−2.16, −0.94], p<0.001) and PPD (β=−0.99 [−0.58,
−1.34], p<0.001). The a io was also lowe in AD compa ed o all he
o he g oups (con ols, O he NDs, MCI, P esymp oma ic NDs, and all
p<0.002).
3.1.4 Plasma GFAP
GFAP le els we e ele a ed in AD compa ed o b FTD (β=0.80 [0.23,
1.39], p=0.010), and PPD (β=0.82 [0.36, 1.21], p<0.001). GFAP le -
els we e also ele a ed in AD compa ed o some o he g oups (con ols,
MCI,P esymp oma ic ND,allp<0.038), bu no be ween ADand o he
NDs (p=0.108).
ERATNE ET AL.7o 15
FIGURE 2 ROC analyses o diagnos ic pe o mance o plasma p- au217, N L, N L/p- au217 a io, and GFAP. Plasma p- au217 had he e y
s ong diagnos ic pe o mance o AD e sus PPD, AD e sus b FTD, AD e sus non-AD, AD e sus con ols, ou pe o ming o he bioma ke s and
he N L/p- au217 a io. N L’s diagnos ic pe o mance was o b FTD e sus PPD and all NDs e sus PPDs. AD, Alzheime ’s disease; b FTD,
beha io al a ian on o empo al demen ia; GFAP, glial ib illa y acidic p o ein; N L, neu o ilamen ligh chain; PPD, p ima y psychia ic diso de ;
p- au217, phospho yla ed au 217; ROC, ecei e ope a ing cha ac e is ic.
3.2 Bioma ke diagnos ic pe o mance o
dis inguish AD om PPDs, AD om b FTD, and
b FTD om PPD
3.2.1 Dis inguishing AD om PPDs
P- au217 demons a ed he s onges diagnos ic pe o mance o dis-
inguish AD om PPDs, as demons a ed in Figu e 2and Suppo ing
In o ma ion Table 3. P- au217 had an AUC o 0.97, ou pe o ming N L,
N L/p- au217 a io, and GFAP (AUCs 0.89, 0.77, 0.86 espec i ely;
AUC di e ences all p<0.016). P- au217 had he highes accu acy
(93%), speci ici y (91%), and sensi i i y (95%), a an op imal cu -o o
1.84pg/mL. Fu he de ails o all diagnos ic pe o mance me ics a e
a ailable in Suppo ing In o ma ion Table 3.
3.2.2 Dis inguishing AD om b FTD
To dis inguish AD om b FTD, p- au217 once again demons a ed
he highes AUC (0.93) and he s onges diagnos ic pe o mance
(93% speci ici y, 98% sensi i i y, 96% accu acy, and cu -o 1.64) com-
pa ed o he o he bioma ke s. N L did no ha e signi ican diagnos ic
pe o mance o dis inguish AD om b FTD (AUC 0.51 [0.30, 0.74]).
The AUC di e ence was no s a is ically di e en be ween p- au217
and he N L/p- au217 a io (AUC 0.88, AUC di e ence p=0.509) o
GFAP (AUC 0.77, AUC di e ence p=0.10). Howe e , compa ed o
p- au217, bo h he a io and GFAP demons a ed poo e speci ici y
(87% and 67%, espec i ely), sensi i i y (80% and 85%), and accu acy
(82% and 80%) and much lowe diagnos ic odds a ios (26 and 11.33,
s p- au217’s 546).
3.2.3 Dis inguishing AD om non-AD diso de s
P- au217demons a edhighdiagnos icpe o mance odis inguishAD
om o he non-AD diso de s (consis ing o b FTD, PPDs, O he NDs),
wi hanAUCo 0.94.I was supe io oN L, N L/p- au217(all p≤0.001)
and had 88% speci ici y, 93% sensi i i y, 90% accu acy, and a cu -o o
2.19.
8o 15 ERATNE ET AL.
P- au217 demons a ed e y s ong diagnos ic pe o mance o
dis inguish AD om con ols (AUC 0.98, 96% speci ici y, and 93%
sensi i i y).
3.2.4 Dis inguishing b FTD om PPD
To dis inguish b FTD om PPD, N L had he highes AUC (0.78) and
s onges diagnos ic pe o mance (81% speci ici y, 67% sensi i i y,
79% accu acy, and cu -o 15.15pg/mL), al hough he AUC di e ence
be ween N L and N L/p- au217 a io was no signi ican (p=0.14). P-
au217 and GFAP had no diagnos ic u ili y (AUC con idence in e als
bo h c ossed 0.50).
3.3 Bioma ke diagnos ic pe o mance o
dis inguish all NDs om PPDs
As a ollow-on om ou p e ious s udies ha ocused on N L in dis-
inguishing NDs om PPDs, we in es iga ed he abili y o p- au217
and GFAP o dis inguish NDs as a g oup om PPDs. We c ea ed an
“all NDs” g oup by combining AD, b FTD, and O he NDs. N L demon-
s a ed he highes AUC (0.87) and s onges diagnos ic pe o mance
(81% speci ici y, 85% sensi i i y, 84% accu acy, and cu -o 14.35) o
dis inguishAll NDs omPPDs, signi ican ly ou pe o mingGFAP(AUC
0.73, AUC di e ence p<0.0001) and p- au217 (AUC 0.72, AUC di e -
ence p<0.001). The N L/p- au217 a io did no ha e diagnos ic u ili y
o All NDs e sus PPDs (AUC con idence in e al c ossed 0.50).xxx
3.4 Bioma ke p o iles in diagnos ic g oups
3.4.1 AD pa hology/neu onal inju y (P- au217/N L)
bioma ke p o iles
We desc ibed p- au217 and N L bioma ke p o iles in he diagnos ic
g oups, using ou p e iously desc ibed cu o o plasma p- au217, and
age-adjus ed pe cen iles model o N L.10,27
As demons a ed in Figu e 3, he high p- au217 p o iles we e mos
commonly seen in AD ( o al 88% o AD, 60% high p- au217/high N L,
and 28% high p- au217/low N L bioma ke p o iles). By compa ison
less han 10% o each o he o he g oups exhibi ed high p- au217
p o ile. b FTD only had 7% wi h high p- au217 p o iles (all high p-
au217/high N L), PPD only 6% (3% high p- au217/high N L, 3% high
p- au217/low N L), con ols only 3% (all high p- au217/low N L), and
MCI and p esymp oma ic ND 8% and 6% espec i ely. O he NDs had
13% wi h high p- au217 p o iles (10% high p- au217/high N L, 3% high
p- au217/low N L).
68% o AD had high N L p o iles (high p- au217/high N L, low p-
au217/high N L), and AD had he lowes p opo ion o “no mal” (low
p- au217/lowN L) p o iles.Fo b FTDon he o he hand, only53% had
high N L p o iles, and he emaining 47% had low p- au217/low N L
p o iles. Mos PPD had a no mal p o ile (72%), wi h 25% showing high
N L p o iles. Looking a indi idual PPD (de ails in Suppo ing In o ma-
ion Figu es 1and 2), he e was a g ea e p opo ion o pa ien s wi h
high N L p o iles in schizoph enia (37%) and O he PPD (33%), com-
pa ed o o he g oups (MDD 19%, FND 12.5%, BPAD 0%), al hough
heseexplo a o y indingsshouldbe in e p e edwi h cau ion gi en he
small sub-g oup numbe s. 83% o con ols had a no mal p o ile, and
only 14% ha ing high N L p o iles. O he ND had he highes p opo -
ionso high N Lle els(74%,64%lowp- au217/high N L,and 10%high
p- au217/highN L).Mos pa icipan sin MCI,andP esymp oma ic ND
had low p- au217/low N L p o iles (72%, 77%, and 67%, espec i ely),
simila o PPD and con ols.
These indings a e u he demons a ed in Figu e 4, a plo o
di e en diagnos ic g oups o log p- au217 e sus N L age-adjus ed
z-sco e, and dis ibu ions. The sca e plo and dis ibu ions demon-
s a e ha almos allADpa ien shadele a edp- au217le elsandhigh
p- au217/high N L p o iles (mos in uppe igh in sca e plo ), as com-
pa ed o non-AD diso de s, whe e almos all had low p- au217 le els.
Mos NDs had high N L le els (mos in lowe igh in sca e plo ), com-
pa ed o a small p opo ion o PPDs and con ols (mos in lowe le in
sca e plo ).
3.4.2 AD pa hology/neu onal
inju y/neu oin lamma ion (p- au217/N L/GFAP)
bioma ke p o iles
As an explo a o y aim, we de eloped a no el GAMLSS age-based
model o plasma GFAP o de i e pe cen iles and z-sco es, based on
ou con ol g oup. As demons a ed in Figu e 5, his shows a simila
non-linea associa ion wi h age and a U-shaped pa e n simila o he
ound in o he s udies.36,37 This enabled complex p- au217/N L/GFAP
(AD pa hology/neu onal inju y/neu oin lamma ion) bioma ke p o il-
ing, which is demons a ed in Figu e 6. Inclusion o GFAP o c ea e his
h ee-bioma ke p o ile esul ed in some di e ences compa ed o p-
au217/N L only p o iling. The p opo ion o “no mal” p o iles using
h ee bioma ke s (i.e., low p- au217/low N L/low GFAP) was sligh ly
lowe in b FTD (40%, compa ed o 47% low p- au217/low N L p o-
iles). The p opo ion o no mal p- au217/N L/GFAP p o iles in PPDs
and con ols we e simila o p opo ions o no mal p- au217/N L p o-
iles. Looking a indi idual PPDs (de ails in Suppo ing In o ma ion
Figu es 1and 2), high GFAP p o iles did no appea o be di e -
en be ween g oups, al hough hese indings should once again be
in e p e ed wi h cau ion gi en he small subg oup numbe s.
We pe o med some addi ional explo a o y analyses. To in es iga e
whe he he e we e di e ences in le els o se e i y and bioma ke s
be ween AD and b FTD, we i s compa ed MMSE sco es be ween
hese wo g oups (calcula ed om o al sco es on he Neu opsychia-
y Uni Cogni i e Assessmen Tool (NUCOG), he cogni i e sc eening
ins umen used in he Neu opsychia y Cen e).39 MMSE sco es we e
no s a is ically di e en (median [in e qua ile ange, IQR]: 21.9 [16.5
o 24.9] o AD s 23.4 [20.9 o 25.0] o b FTD, p=0.093), Sup-
po ing In o ma ion Table 1. Nex , we an GLMs explo ing whe he
diagnosis o AD o b FTD, MMSE, and diagnosis ×MMSE in e ac ion
ERATNE ET AL.9o 15
FIGURE 3 P- au217 and N L bioma ke p o iles, classi ied based on p e iously desc ibed cu -o s, in di e en diagnos ic g oups. N L,
neu o ilamen ligh chain; p- au217, phospho yla ed au 217.
in luenced N L le els, adjus ing o age and sex. Diagnosis (p=0.394),
MMSE (p=0.700), and he in e ac ion e m (p=0.634) we e all
non-signi ican . Taken oge he , hese indings indica e ha he highe
p opo ion o “high N L” p o iles in AD was unlikely o be explained
by mo e ad anced cogni i e impai men compa ed o b FTD. Finally,
we in es iga ed o any associa ions be ween bioma ke s and cogni-
i e impai men in AD. In AD cases only, highe plasma N L showed a
end owa d associa ion wi h lowe MMSE, in GLMs also adjus ing o
age and sex, bu i was no s a is ically signi ican (p=0.076). No sig-
ni ican associa ions we e obse ed in AD be ween MMSE sco es and
plasma p- au217 (p=0.928) and GFAP (p=0.614). These sugges ha
wi hin ou AD g oup, bioma ke le els did no signi ican ly a y wi h
he deg ee o cogni i e impai men .
4DISCUSSION
We in es iga ed mul iple bioma ke s in an unselec ed eal-wo ld clini-
cal coho o younge pa icipan s seen in a neu opsychia y memo y
clinic, ocusing on AD, b FTD, and PPDs. The main indings we e as
ollows: (1) e y s ong diagnos ic pe o mance o plasma p- au217
o dis inguish AD om b FTD, and AD om PPDs; (2) supe io pe -
o mance o p- au217 o AD diagnosis on i s own, compa ed o
N L/p- au217 a io, N L, and GFAP; (3) s onges pe o mance o N L
o dis inguishing b FTD om PPDs, and all NDs om PPDs, consis-
en wi h p e ious s udies.30,31 This adds impo an u he e idence
o he limi ed li e a u e so a on oles o di e en bioma ke s in a
eal-wo ldclinicalcoho , wi hp- au217 beinga e ysensi i eandspe-
ci ic es o AD and N L ha ing he s onges diagnos ic pe o mance
o dis inguish neu odegene a ion om PPDs/non-NDs, wi h li le ole
o GFAP in any o hese dis inc ions. The key s eng hs o his s udy
included a eal-wo ld, unselec ed di e se, clinical, and younge coho
– a ela i ely unde in es iga ed g oup. Es ablishing he ole o hese
bioma ke s in younge people, who ace g ea e a es o misdiagno-
sis and diagnos ic unce ain y, is impo an o inc easing clinical and
esea ch access, whe e a guably he clinical, ea men , and b oade
psychosocial bene i s could be g ea e .
This s udy builds on ou p e ious wo k on p- au217,10 p o id-
ing u he e idence o e y s ong diagnos ic pe o mance o some
o he mos common diagnos ic dis inc ions, dis inguishing AD om
b FTD, and AD om PPDs. We saw e y high speci ici y and sensi i -
i y (91% and 95%), es ablishing a ole o such an accu a e es (93%)
in younge people. We ound no bene i o N L/p- au217 a io o e p-
au217 alone in e ms o imp o ing he dis inc ion be ween AD and
