Vol.:(0123456789)
Molecula Neu obiology (2025) 62:1467–1477
h ps://doi.o g/10.1007/s12035-024-04347-6
ORIGINAL ARTICLE
Neu oaxonal Inju y May Media e heAssocia ion Be ween
Hype glycemia andP ognosis inSpon aneous Suba achnoid
Hemo hage
DanielSan ana1 · Lau aLlull1 · Alejand aMos ei o2 · Lei ePed osa3 · Gab ielPujol4 · LuigiZa e a5 ·
Ma ianoWe ne 6 · Ab ahamMa ín7,8 · Ca lesJus icia9 · ÁngelChamo o1 · RamónTo né2 ·
Se gioAma o1
Recei ed: 4 Ap il 2023 / Accep ed: 2 July 2024 / Published online: 12 July 2024
© The Au ho (s) 2024
Abs ac
Hype glycemia du ing ea ly b ain inju y (EBI) pe iod a e spon aneous suba achnoid hemo hage (SAH) is associa ed wi h
poo ou come, bu he unde lying physiopa hology is unknown. This s udy assessed i hype glycemia du ing EBI is associa ed
wi h ma ke s o neu oaxonal inju y and whe he hese bioma ke s pa ially accoun o he associa ion be ween hype glycemia
and poo clinical ou come. Nine y- wo SAH pa ien s admi ed wi hin 24h o bleeding onse we e p ospec i ely included.
Glucose le els we e measu ed a a i al and e e y 6h o 72h. Se um neu o ilamen ligh chain (NFL) le els we e meas-
u ed a 72h. Func ional ou come was assessed wi h he modi ied Rankin Scale (mRS) a 90days (poo ou come, mRS > 2).
The associa ion be ween glucose me ics, NFL le els, and clinical ou come was assessed wi h uni a ia e and mul i a ia e
analyses. Media ion analysis was pe o med o examine he po en ial chain in which NFL may media e he ela ionship
be ween glucose and unc ional ou come. Highe glucose and NFL le els du ing EBI we e associa ed wi h poo clinical
ou come in adjus ed analysis. NFL le els we e associa ed wi h olde age, highe ini ial se e i y, and highe glucose le els
du ing EBI pe iod. In adjus ed media ion analyses, he associa ion be ween glucose and clinical ou come was signi ican ly
media ed by NFL le els. The media o NFL explained 25% o he associa ion be ween glucose du ing EBI pe iod and poo
unc ional ou come a 90days. In SAH, he associa ion be ween glucose le els du ing EBI and poo clinical ou come migh
be signi ican ly media ed by NFL le els. The link be ween hype glycemia and poo clinical ou come migh be explained in
pa h ough seconda y neu oaxonal inju y.
Keywo ds Suba achnoid hemo hage· Ea ly b ain inju y· Neu o ilamen · Glucose· Neu oaxonal damage· Media ion
analysis
* Ramón To né
[email p o ec ed]
* Se gio Ama o
[email p o ec ed]
1 Ins i u e o Neu oscience, Comp ehensi e S oke Cen e ,
Hospi al Clinic o Ba celona; Uni e si y o Ba celona,
Ba celona, Spain
2 Ins i u e o Neu oscience, Neu osu ge y Depa men ,
Hospi al Clinic o Ba celona; Uni e si y o Ba celona,
Ba celona, Spain
3 Ins i u d’In es igacions Biomèdiques Agus í Pi I Sunye
(IDIBAPS), Ba celona, Spain
4 Neu oanes hesia Di ision, Anes hesiology Depa men ,
Hospi al Clinic o Ba celona, Ba celona, Spain
5 Neu oc i ical Ca e Di ision, An hesiology Depa men ,
Hospi al Clinic o Ba celona, Ba celona, Spain
6 Ins i u e o Diagnos ic Imaging, In e en ional
Neu o adiology Depa men , Hospi al Clinic o Ba celona,
Ba celona, Spain
7 Achuca o Basque Cen e o Neu oscience, Leioa, Spain
8 Ike basque Basque Founda ion o Science, Bilbao, Spain
9 Ins i u o de In es igaciones Biomédicas de Ba celona (IIBB),
Consejo Supe io de In es igaciones Cien í icas (CSIC),
Ba celona, Spain
1468 Molecula Neu obiology (2025) 62:1467–1477
In oduc ion
Spon aneous suba achnoid hemo hage (SAH) is a de as-
a ing ce eb o ascula disease caused by he up u e o an
in ac anial aneu ysm in 85% o cases [1, 2]. The disabili y
bu den caused by SAH is pa ly d i en by se e al ha m ul
mechanisms igge ed wi hin he i s 72h a e bleeding
onse esul ing in he so-called ea ly b ain inju y (EBI),
which has been consis en ly associa ed wi h an inc eased
isk o delayed sys emic complica ions as well as long-
e m poo ou come [3, 4].
The mechanisms implica ed in EBI a e s ill pa ly
known and include mic o h ombosis, oxida i e s ess,
blood–b ain ba ie dis up ion, and mul i ocal axonal
inju y, among o he s [5–10]. In acu e cen al ne ous sys-
em diseases including b ain hemo hage, mos o hese
dele e ious mechanisms a e enhanced a e he exposi ion
o high glucose le els [11–14]. Hype glycemia occu s
in abou h ee ou o ou SAH pa ien s wi hin he EBI
pe iod, and i has been associa ed wi h poo clinical ou -
come [15–20]. Hype glycemia is in pa a consequence
o an acu e s ess esponse o b ain inju y and enhances a
my iad o dele e ious physiopa hological p ocesses includ-
ing neu oin lamma o y mechanisms ha e en ually may
p omo e di use neu oaxonal damage (NAD) [21–23].
Howe e , he e a e no epo s on he link be ween abno -
mal glucose p o iles du ing he EBI pe iod and NAD
bioma ke s.
Neu o ilamen ligh chains (NFL) a e s uc u al p o eins
mainly p esen in he cy oskele on o myelina ed neu ons.
Following axonal inju y, NFLs a e eleased o he ex a-
cellula compa men hence being conside ed su oga e
bioma ke s o NAD [24]. In he se ing o SAH, NFL
le els measu ed in blood and ce eb ospinal luid samples
collec ed du ing he EBI pe iod as bioma ke s o ongoing
b ain inju y a e s ongly co ela ed wi h poo clinical ou -
come me ics a sho and long e m [25–33]. These obse -
a ions sugges ha NAD could play a majo ole in he
EBI phase a e SAH, al hough he mechanisms implica ed
in he se e i y o NAD emain o be ully unde s ood.
He ein, we hypo hesized ha in SAH su e e s, he
p ognos ic ele ance o hype glycemia du ing he EBI
pe iod could be a leas in pa media ed h ough i s ela-
ionship wi h NAD. Thus, ou speci ic objec i e was o
e alua e in a p ospec i e coho o SAH subjec s whe he
he associa ion be ween hype glycemia and poo clinical
ou come was media ed by di use NAD as measu ed wi h
ci cula ing NFL le els a he end o he EBI pe iod. E en-
ually, de e mining he link be ween glucose dys egula-
ions du ing he EBI pe iod, NAD and long- e m clinical
ou come could be o in e es o designing and moni o ing
he e ec s o new in e en ions in clinical ials aimed
o modula e glucose le els o he mechanisms associa ed
wi h glucose-d i en seconda y b ain inju y.
Me hods
S udy Design andPopula ion
We conduc ed an obse a ional, single-cen e , p ospec i e
coho s udy in a e ia y le el e e al hospi al (> 50 SAH
admissions pe yea ) p o ided wi h a mul idisciplina y Com-
p ehensi e S oke Cen e , a S oke Uni and an In ensi e
Ca e Uni specialized in neu oc i ical ca e. Pa icipan s we e
included in he s udy om Sep embe 2018 o June 2021.
The s udy inclusion c i e ia we e age olde han 18yea s,
good p emo bid unc ional s a us [modi ied Rankin Scale
(mRS) 0–2], and spon aneous SAH admi ed a ou cen e
in he i s 24h om symp om onse . Exclusion c i e ia
included he p esence o a non-aneu ysmal exclusi ely pe -
imesencephalic blood pa e n in he admission CT, p e ious
his o y o diabe es melli us, SAH seconda y o o he causes
(i.e., auma ic, enous h ombosis, myco ic aneu ysms), o
e y sho li e expec ancy ha could comp omise ollow-
up. A schema ic lowcha o he s udy is shown in Supple-
men a y Fig.1. E hics app o al was ob ained om he local
Clinical Resea ch E hics Commi ee om Hospi al Clinic o
Ba celona (HCB/2017/027). Pa ien s o a legal p oxy ga e
w i en in o med consen p io o inclusion in he s udy.
Clinical Assessmen
Baseline demog aphic da a, pas medical his o y, clini-
cal ea u es a SAH onse , and neu oimaging esul s we e
p ospec i ely eco ded. Neu ological condi ion on admis-
sion was assessed acco ding o he Wo ld Fede a ion o
Neu osu gical Socie ies (WFNS) g ading sys em a e ini-
ial esusci a ion. Pa ien s we e g ouped in o good g ade
(WFNS g ades 1 and 2) o poo g ade (WFNS g ades 3 o
5) SAH. Pa ien s wi h SAH we e admi ed o he c i ical
ca e uni (WFNS 3–5) o he S oke Uni (WFNS 1–2) as
clinically app op ia e. Neu ological s a us [Glasgow Coma
Scale (GCS) and Na ional Ins i u es Heal h S oke Scale
(NIHSS)] we e assessed daily by neu ologis s in he S oke
Uni o anes hesiologis s in he c i ical ca e uni . Sys emic
and neu ological complica ions we e moni o ed and ea ed
acco ding o in-hospi al p o ocols.
GCS is a widely used assessmen ool o pa ien s’ le el
o consciousness, based on h ee i ems: eye opening, e -
bal esponse, and mo o esponse. I anges om 15 ( ull
consciousness) o 3 (deep comma) [34]. WFNS is a SAH-
speci ic g ading sys em ha classi ies pa ien s in o i e
g oups, acco ding o hei GCS and p esence o absence o
mo o de ici [35]. Thus, WFNS I ep esen s a SAH pa ien
1469Molecula Neu obiology (2025) 62:1467–1477
wi h unpai ed consciousness (GCS 15) and no mo o de ici ,
whe eas WFNS V ep esen s a coma ose pa ien (GCS 3 o
6). Las ly, NIHSS is a s anda dized assessmen ool mainly
used in he con ex o ce eb o ascula diseases o e alua e
ocal de ici s in se e al neu ological ields such as le el
o consciousness, speech, mo o , and senso y unc ions.
I anges om 0 (no de ici ) o 42 (maximum de ici ) and
allows a p ecise and ep oducible moni o ing o he ex en
o b ain damage [36]. All h ee scales ha e been widely
epo ed in medical li e a u e in he con ex o SAH, wi h
signi ican associa ions be ween wo se sco es and poo e
p ognosis.
Imaging Me hods
Non-con as b ain CT was pe o med o all pa ien s a
a i al. The ex en o suba achnoid bleeding was ca ego ized
acco ding o he modi ied Fishe Scale (mF) classi ica ion.
This scale ca ego izes he amoun o blood obse ed in non-
con as b ain CT in o ou g oups, acco ding o bleeding
hickness in suba achnoid space and he p esence o absence
o in a en icula blood clo s. I anges om 1 ( hin suba-
achnoid bleeding, no in a en icula blood) o 4 ( hick
suba achnoid bleeding, in a en icula blood) and is posi-
i ely co ela ed wi h he isk and se e i y o complica ions
associa ed wi h he bleeding, mainly a e ial asospasm
[37]. Addi ional ele an adiological ea u es including he
p esence o in a en icula hemo hage, hyd ocephalus,
o in apa enchymal hema oma we e also collec ed. Addi-
ional imaging acquisi ions we e pe o med a disc e ion o
he a ending physician acco ding o clinical e olu ion. An
Angio-CT was pe o med o all pa ien s a a i al, as well
as a ce eb al angiog aphy wi hin he i s 24h a e hospi al
admission, as pa o ascula s udy. Acco ding o he p es-
ence o absence o a b ain aneu ysm, he pa ien s we e u -
he classi ied as non-aneu ysmal SAH o aneu ysmal SAH.
In he la e g oup, da a on he numbe , loca ion, size, and
ana omical ea u es o aneu ysms we e eco ded, and he
index aneu ysm was e alua ed o endo ascula o su gical
exclusion ea men acco ding o in-hospi al p o ocols.
Sampling andLabo a o y Analysis
Blood glucose was measu ed h ough inge p ick es ing a
a i al and e e y 6h o a leas he ollowing 72h, as pa
o usual clinical p ac ice. Glucose managemen consis ed on
a sliding scale in as ing pa ien s and/o pa ien s ecei ing
con inuous nu i ion h ough a nasogas ic ube and a basal-
bolus insulin he apy in pa ien s ecei ing egula meals.
Maximum, minimum, and mean glucose le els du ing he
i s 72h we e eco ded.
To e alua e se um NFL le els, an addi ional blood sam-
ple (15mL) was d awn a 72h a e hospi al admission.
The samples we e cen i uged a 4000 pm (1523 ela i e
cen i ugal o ce) o 15min a 20–22°C, and supe na an
was pipe ed and ozen a − 80°C un il u he analysis.
NFL le els we e measu ed h ough he Simoa echnique
(Quan e ix, Lexing on, MA, USA) in se um samples. In
b ie , Simoa is a ou h-gene a ion ELISA based on p o eic
an igen cap u ing by pa amagne ic mic oscopic beads ha
ca y sandwich an ibody complexes. Then, hose beads a e
p ecipi a ed in o single-molecule a ays ha a e indi idu-
ally ead o check o hei bond o he index an igen. Ba ch
se um NFL measu emen was pe o med acco ding o he
manu ac u e ’s guide.
Clinical Ou come E alua ion
The p ede ined p ima y ou come measu e was unc ional
disabili y, e alua ed wi h he modi ied Rankin Scale (mRS).
Func ional ou come was assessed by ce i ied neu ologis s
on in-pe son isi s a he ou pa ien clinic 90days a e
ini ial hospi al admission, h ough a semi-s uc u ed clini-
cal in e iew. Poo unc ional ou come was p ede ined as
mRS > 2.
S a is ical Analysis
Con inuous a iables a e shown as mean (s anda d de ia-
ion) o median (in e qua ile in e als), and di e ences
be ween g oups we e assessed using S uden ’s - es ,
ANOVA, o Mann–Whi ney as app op ia e. Ca ego ical a -
iables a e shown as absolu e alues (pe cen age) and com-
pa ed using chi-squa ed o Fishe ’s exac es s. A uni a ia e
analysis was un o assess he associa ion o demog aphic,
clinical, and ea men - ela ed a iables wi h clinical ou -
come a 3mon hs and wi h NFL le els. The accu acy o
he explo ed glucose me ics o p edic ing clinical ou come
was assessed wi h ecei e ope a ing cha ac e is ic cu es
(ROC), and he glucose me ic wi h highe a ea unde he
cu e (AUC) was selec ed o i s inclusion in mul i a ia e
models. A mul i a ia e logis ic eg ession model analyzed
he p edic o s o poo clinical ou come, and a mul i a ia e
linea eg ession assessed he a iables associa ed wi h NFL
le els. Spea man’s co ela ion analyses we e pe o med o
assess he associa ions be ween glucose me ics, NFL le -
els, and mRS sco es, and he nume ical esul s we e illus-
a ed h ough a co ela ion ma ix. Fo linea eg ession
and media ion analyses, we used log- ans o med NFL le els
o app oach no mali y. Age and WFNS a admission we e
o ced o emain in he inal model o hei p ognos ic el-
e ance based on p io knowledge. To a oid model o e i -
ing, a backwa d elimina ion me hod (likelihood a io) was
used o selec he inal bina y logis ic eg ession model, and
he Hosme –Lemeshow goodness-o - i s a is ic was used
o assess he inal model i . Media ion analyses we e also
1470 Molecula Neu obiology (2025) 62:1467–1477
implemen ed o assess whe he glucose in luenced clinical
ou come indi ec ly ia NFL by using he PROCESS mac o
(Hayes, 2013) o he s a is ical so wa e IBM SPSS S a is-
ics o Windows (IBM, A monk, NY, USA); 10,000 boo -
s aps p oduced 95% con idence in e als (CI) on he media-
ion coe icien pa ame e es ima es, and media ed (indi ec )
e ec s we e conside ed signi ican i he 95% CI o hei
pa ame e es ima es did no c oss he ze o alue. The analy-
ses we e pe o med using SPSS e sion 25.0, and he le el
o signi icance was es ablished a a 0.05 le el ( wo-sided).
Da a A ailabili y
The da ase s gene a ed and analyzed a e a ailable om he
co esponding au ho s on easonable eques .
Resul s
Cha ac e is ics o  heS udy Popula ion
A o al o 92 pa ien s we e included du ing he s udy pe iod.
The main ai s o he included pa ien s a e shown in Table1.
O e all, 61 (66%) pa ien s we e women wi h a median (IQR)
age o 55 (49–65) yea s. On admission, 36 (39%) pa ien s
p esen ed wi h poo clinical g ade SAH (WFNS 4–5), and
67 (73%) had a modi ied Fishe Scale sco e o 4. A b ain
aneu ysm was ound in 84 (91%) pa ien s, om whom 57
(62%) we e ea ed endo ascula ly and 23 (25%) ecei ed
su gical clipping. Ini ial neu oimaging assessmen e ealed
he p esence o hyd ocephalus in 58 (63%) pa ien s, pa en-
chymal hemo hage in 22 (24%), and in a en icula bleed-
ing in 68 (74%).
Table 1 S udy popula ion
cha ac e is ics assessed du ing
he ea ly b ain inju y pe iod
acco ding o clinical ou come a
3mon hs (whole sample)
Quali a i e da a a e exp essed as n (%) and quan i a i e da a as median (IQR)
mRS modi ied Rankin Scale, SD s anda d de ia ion
mRS 0–2 (n = 60) mRS 3–6 (n = 32) p
Demog aphics
Age (yea s) 54 (49–62) 61 (52–74) 0.054
Female sex 39 (65) 22 (69) 0.717
P e-mo bid mRS 0 (0–0) 0 (0–1) 0.003
Medical his o y
Hype ension 21 (35) 17 (53) 0.093
Dyslipidemia 9 (15) 8 (25) 0.239
Baseline ai s
WFNS g ade (4–5) 13 (22) 23 (72) < 0.001
Aneu ysm (yes) 54 (90) 30 (94) 0.543
Aneu ysm size (mm) 5 (4–6) 6 (5–9) 0.010
Modi ied Fishe Scale 4 (3–4) 4 (4–4) 0.033
In a en icula bleeding 41 (68) 27 (84) 0.095
Hyd ocephalus 33 (55) 25 (68) 0.029
Pa enchymal hemo hage 13 (22) 9 (28) 0.489
Global ce eb al edema 33 (55) 20 (63) 0.488
Aneu ysm ea men modali y 0.596
No 8 (13) 4 (13)
Endo ascula 39 (65) 18 (56)
Su gical 13 (22) 10 (31)
Glucose me ics
Glucose a admission (mg/dL) 125 (113–142) 138 (126–164) 0.005
Mean glucose wi hin 72h (mg/dL) 114 (103–128) 137 (123–150) < 0.001
Max glucose wi hin 72h (mg/dL) 150 (126–164) 165 (150–188) 0.001
Min glucose wi hin 72h (mg/dL) 87 (81–100) 103 (97–121) < 0.001
SD glucose wi hin 72h (mg/dL) 18 (14–24) 23 (16–28) 0.106
1471Molecula Neu obiology (2025) 62:1467–1477
Glucose Me ics Du ing heEBI Pe iod
Glucose me ics du ing he EBI pe iod including glucose
a admission, maximum, minimum, mean, and s anda d
de ia ion o glucose le els a e shown in Table1. In uni-
a ia e analyses, all glucose me ics excep he s anda d
de ia ion (SD) we e associa ed wi h clinical ou come a
90days. In bina y logis ic eg ession mul i a ia e analy-
ses adjus ed by baseline clinical se e i y (WFNS), only
mean glucose le els and minimum glucose le els wi hin
he i s 72h a e bleeding onse emained associa ed wi h
poo clinical ou come a 90days [mean glucose le els, OR
(95% CI) pe each mg/dL o inc ease 1.04 (1.010–1.077),
p = 0.010; minimum glucose le els, OR (95% CI) pe
each mg/dL o inc ease 1.04 (1.007–1.079), p = 0.020]. In
ag eemen wi h hese obse a ions, in ROC analysis, mean
glucose le els du ing he EBI pe iod disclosed he high-
es AUC alue, ollowed by minimum glucose du ing he
EBI pe iod, maximum glucose du ing he EBI pe iod, and
glucose le els a hospi al admission, as shown in Table2.
Acco ding o hese indings, mean glucose le els du ing
he i s 72h a e hospi al admission we e used o u he
analyses.
Va iables Associa ed wi hPoo Func ional Ou come
In his coho , a o al o 32 (35%) pa ien s showed poo
unc ional ou come a 90days. In uni a ia e analyses,
poo ou come was associa ed wi h wo se p emo bid
mRS, poo e ini ial WFNS g ade, la ge aneu ysm size,
highe mF sco e, and hyd ocephalus a ini ial neu oimag-
ing, as well as wi h highe glucose le els du ing he i s
72h (Table1). In mul i a ia e models, he a iables ha
emained independen ly associa ed wi h poo clinical ou -
come we e poo e WFNS g ade a onse and highe mean
glucose le els du ing he EBI pe iod (Table3).
Ci cula ing NFL Le els a  heEnd o  heEBI Pe iod:
Con ibu o s andAssocia ion wi hClinical Ou come
The median (IQR) le els o NFL a he end o he EBI pe iod
we e 25 (13–56) pg/mL. The a iables associa ed wi h
highe NFL le els in uni a ia e analysis included olde age,
poo e ini ial WFNS g ade, and highe mean glucose le els
du ing he EBI pe iod, as shown in Table4. In mul i a i-
a e analyses, only glucose le els emained associa ed wi h
NFL [adjus ed OR 1.01 (95% CI 1.001–1.024, p = 0.03)].
The le els o NFL we e signi ican ly highe in pa ien s
wi h poo clinical ou come [median (IQR) 56 (26–126)
e sus 19 (11–32) in pa ien s wi h poo and good clinical
ou come, espec i ely; p < 0.001], as shown in Fig.1A. A
ma ix o co ela ions be ween glucose me ics assessed du -
ing he EBI pe iod, NFL le els measu ed a 72h, and mRS
a 3mon hs is shown in Fig.1B. O no e, he associa ion
be ween NFL le els and poo clinical ou come emained
signi ican in bina y eg ession models adjus ed by age,
p e-mo bid mRS, WFNS g ade a hospi al admission, and
mean glucose le els du ing he EBI pe iod [adjus ed OR
2.61 (95% CI 1.364–5.010, p = 0.004)].
Glucose Le els Du ing heEBI Pe iod, Ci cula ing
NFL Le els a  heEnd o  heEBI Pe iod, andClinical
Ou come a 90 Days: AMedia ion Analysis
A media ion analysis was implemen ed o e alua e whe he
he associa ion be ween highe glucose le els du ing he EBI
pe iod and clinical ou come was media ed h ough i s asso-
cia ion wi h inc eased NFL le els, as schema ically shown
in Fig.2A. As shown in Fig.2B, in unadjus ed analyses, glu-
cose le els du ing EBI we e associa ed wi h highe ci cula -
ing NFL le els a 72h (media o a iable), and bo h glucose
and NFL le els we e also signi ican ly associa ed wi h poo
clinical ou come. Media ion analysis be ween glucose le els
Table 2 Recei e ope a ing cha ac e is ic cu e analyses o he accu-
acy o glucose me ics o p edic ing clinical ou come a 90days
AUC a ea unde he cu e, CI con idence in e al, SD s anda d de ia-
ion
Glucose me ics AUC (95% CI), p p alue
Glucose a admission 0.680 (0.569–0.792) 0.005
Mean glucose wi hin 72h 0.797 (0.706–0.889) < 0.001
Max glucose wi hin 72h 0.708 (0.600–0.817) 0.001
Min glucose wi hin 72h 0.764 (0.664–0.863) < 0.001
SD glucose wi hin 72h 0.605 (0.481–0.729) 0.098
Table 3 Va iables assessed du ing EBI pe iod associa ed wi h poo
clinical ou come (mRS > 2): mul i a ia e analysis
The a iables ha emained in he inal model we e selec ed h ough
bina y logis ic eg ession wi h backwa d p ocedu e (likelihood a io).
The ini ial model included all he a iables showing a p alue o < 0.1
in uni a ia e analysis (shown in Table1). Age and WFNS a admis-
sion we e o ced o emain in he inal model
OR odds a io, mRS modi ied Rankin Scale, IQR in e qua ile ange,
WFNS Wo ld Fede a ion o Neu osu gical Socie ies Scale
Va iables OR (95% CI) p alue
Age (pe yea ) 1.02 (0.980–1.069) 0.303
P e-mo bid mRS (pe IQR) 1.90 (0.932–3.865) 0.077
WFNS g ade (4–5) 4.05 (1.276–12.877) 0.018
Mean glucose wi hin 72h (pe mg/
dL) 1.04 (1.009–1.081) 0.014

1472 Molecula Neu obiology (2025) 62:1467–1477
and poo clinical ou come a 90days e ealed a signi ican
indi ec e ec (coe icien = 0.02, 95% CI = 0.007–0.041,
p < 0.005) indica ing ha he associa ion be ween glucose
and clinical ou come was in pa media ed by NFL le els.
A e adjus men o he media o NFL, glucose s ill had a
subs an ial associa ion wi h unc ional ou come. O e all, in
unadjus ed analyses, he media o NFL explained 29% o he
associa ion o highe glucose wi h poo unc ional ou come.
In analyses adjus ed by age, WFNS a hospi al admission
and p e-mo bid mRS, he associa ions be ween glucose,
Table 4 Uni a ia e and
mul i a ia e analyses o he
associa ions be ween se um
NFL le els (log- ans o med)
and baseline a iables
OR odds a io, CI con idence in e al, mRS modi ied Rankin Scale, IQR in e qua ile ange, WFNS Wo ld
Fede a ion o Neu osu gical Socie ies Scale
Va iables Uni a ia e analysis
OR (95% CI), p alue Mul i a ia e analysis
OR (95% CI), p alue
Age (pe yea ) 1.02 (1.000–1.034), p = 0.045 1.01 (0.996–1.027), p = 0.162
Female sex ( s male) 1.22 (0.765–1.928), p = 0.409 -
P e-mo bid mRS (pe IQR) 1.17 (0.874–1.575), p = 0.288 -
Hype ension (yes) 1.27 (0.814–1.972), p = 0.294 -
Dyslipidemia (yes) 1.52 (0.867–2.646), p = 0.145 -
WFNS g ade (4–5) 2.11 (1.385–3.222), p = 0.001 1.51 (0.926–2.463), p = 0.098
Aneu ysm (yes) 1.52 (0.701–3.287), p = 0.290 -
Aneu ysm size (pe IQR) 1.13 (0.925–1.379), p = 0.233 -
Modi ied Fishe Scale (pe IQR) 1.10 (0.862–1.399), p = 0.448 -
In a en icula bleeding (yes) 1.09 (0.659–1.786), p = 0.750 -
Hyd ocephalus (yes) 1.21 (0.770–1.902), p = 0.408 -
Pa enchymal hemo hage (yes) 1.61 (0.970–2.661), p = 0.065 1.34 (0.834–2.151), p = 0.227
Global ce eb al edema (yes) 1.08 (0.696–1.688), p = 0.722 -
Aneu ysm ea men modali y (pe ype) 1.03 (0.757–1.397), p = 0.858 -
Mean glucose wi hin 72h (pe mg/dL) 1.02 (1.011–1.030), p < 0.001 1.01 (1.001–1.024), p = 0.039
Fig. 1 A NFL le els a 72h a e admission in pa ien s wi h good
(mRS 0–2) and poo (mRS 3–6) ou come a 3mon hs. B Co ela ion
ma ix be ween mean glucose in 72h, NFL le els a 72h, and poo
clinical ou come a 3mon hs. Nume ical da a wi hin he ma ix ep-
esen Spea man’s ank co ela ion coe icien s; single as e isk means
p alue lowe han 0.05; double as e isks mean p alue lowe han
0.01. NFL, neu o ilamen ligh chain; mRS, modi ied Rankin Scale;
SD, s anda d de ia ion
1473Molecula Neu obiology (2025) 62:1467–1477
NFL, and clinical ou come emained signi ican , as shown in
Fig.2C. Adjus ed media ion analyses e ealed no signi ican
di ec e ec o glucose on unc ional ou come, while he e
was a signi ican indi ec e ec ha was media ed by NFL
le els (coe icien = 0.01, 95% CI = 0.001–0.040, p < 0.005).
O e all, in adjus ed analyses, he media o NFL explained
25% o he associa ion o highe glucose le els wi h poo
unc ional ou come.
Discussion
In ag eemen wi h p e ious epo s, in his coho o SAH
pa ien s, highe glucose le els du ing he EBI pe iod we e
associa ed wi h poo clinical ou come a long e m. Acu e
hype glycemia du ing he EBI pe iod also p edic ed highe
ci cula ing NFL le els measu ed a he end o he EBI
pe iod. Acco ding o media ion analyses, he associa ion
be ween glucose and poo clinical ou come was signi ican ly
media ed h ough he link be ween glucose and NFL le els.
These obse a ions sugges ha he associa ion be ween
hype glycemia and poo clinical ou come in SAH migh be
explained in pa h ough glucose-d i en seconda y NAD.
In line wi h cu en knowledge, in his coho , glucose
le els measu ed du ing he EBI pe iod a e SAH onse
eme ged as one o he mos obus p edic o s o poo ou -
come, along wi h inc eased clinical se e i y a hospi al
admission [15–20]. In ou sample, all he explo ed glu-
cose me ics excep he s anda d de ia ion we e associ-
a ed wi h clinical ou come a 90days. Acco ding o ROC
analysis, mean glucose le els du ing he EBI pe iod
disclosed he highes accu acy o p edic ing poo clini-
cal ou come a 3mon hs, especially in compa ison wi h
admission glucose le els. This obse a ion is in line wi h
p e ious epo s. None heless, we ound no associa ion
be ween s anda d de ia ion and p ognosis in ou coho ,
whe eas he e is inc easing e idence ha dynamic glucose
pa ame e s du ing he i s days a e s oke could be mo e
in o ma i e o he p edic ion o in hospi al complica ions
o long- e m clinical ou come han isola ed ini ial meas-
u emen s [16, 17, 38–42]. This could be explained due o
he limi ed empo al esolu ion o glucose me ics de i ed
om cu en s anda ds o ca e, hus limi ing he powe
o assessmen s ega ding glucose dynamics and p ogno-
sis. In his con ex , he use o comp ehensi e dynamic
e alua ions such as hose de i ed om con inuous glu-
cose moni o ing de ices could aid o ob ain a mo e p ecise
Fig. 2 Model o he media ion pa hways be ween glucose le els
wi hin he i s 72h a e bleeding onse (glucose du ing EBI) and
clinical ou come a day 90 acco ding o ci cula ing NFL le els (log-
ans o med) measu ed a 72h a e bleeding (NFL a 72h). A Theo-
e ical amewo k o a pa hway model designed o explain he media-
ion e ec o NFL on he associa ion be ween glucose and clinical
ou come a day 90 (dicho omized mRS sco e; poo ou come de ined
as mRS > 2). B, C Unadjus ed and adjus ed pa h media ion analyses,
espec i ely. Coe icien s in he pa h models a e lis ed as a, b, and c′,
whe e a and b a e pa o he indi ec pa h and c′ is he di ec pa h
adjus ed o he indi ec pa h; he o al e ec is he composi ion o he
di ec e ec (c′) and he indi ec e ec (ab). Da a a e coe icien es i-
ma es wi h 95% con idence in e als. EBI, ea ly b ain inju y; NFL,
neu o ilamen ligh chain; mRS, modi ied Rankin Scale; M, media o
1474 Molecula Neu obiology (2025) 62:1467–1477
knowledge on he p ognos ic and physiopa hological el-
e ance o glycemic pa e ns du ing he acu e phase o SAH
[42].
In his s udy, we e alua ed he p esence o NAD non-
in asi ely h ough he measu emen o ci cula ing NFL
le els a he end o he EBI pe iod. The e o e, ci cula ing
NFL le els we e used as a su oga e ma ke o consequence
o ongoing b ain inju y du ing he EBI pe iod and no as
a d i e (o di ec cause) o poo clinical ou come a long
e m. Up o less han a decade ago, NFL could only be eli-
ably measu able in CSF due o echnical limi a ions. The
de elopmen o he Simoa echnique allows NFL measu e-
men in pe iphe al biological samples, hen making possible
an in-dep h e alua ion o NFL as a bioma ke o disease
se e i y and p ognosis. Acco ding o p e ious epo s, in
he se ing o SAH, highe NFL le els co ela e consis en ly
wi h wo se clinical and adiological se e i y o he bleed-
ing a admission as well as wi h poo clinical ou come a
long e m [25–33]. In ou coho , we con i med he associa-
ion be ween se um NFL le els and poo clinical ou come
a 3mon hs a e bleeding. As a no el inding, we ound
ha highe se um NFL le els measu ed a he end o he
EBI pe iod we e p edic ed by highe glucose le els meas-
u ed du ing he acu e phase a e bleeding in adjus ed linea
eg ession analysis, independen ly o age o clinical s a us a
hospi al admission. Gi en he signi ican mul ila e al asso-
cia ions ound be ween glucose le els du ing he EBI pe iod,
NFL le els measu ed a he end o he EBI pe iod, and clini-
cal ou come a 90days, we implemen ed a media ion analy-
sis o explo e whe he he associa ion be ween glucose and
poo clinical ou come was media ed h ough NFL le els as a
media o a iable. Acco ding o hese analyses, he associa-
ion o glucose wi h NFL le els was s onge han i s di ec
associa ion wi h clinical ou come, hus sugges ing ha he
e ec o glucose on long- e m clinical ou come was a leas
in pa media ed by he magni ude o neu oaxonal inju y.
Indeed, he indi ec media ion pa hway explained be ween
29 and 25% o he associa ion be ween glucose and clinical
ou come a day 90 in unadjus ed and adjus ed media ion
analyses, espec i ely. O e all, hese da a sugges ha ele-
a ed glucose le els du ing he EBI pe iod migh hampe he
clinical eco e y a long e m by p omo ing glucose-d i en
neu oaxonal inju y. This suppo s he concep ualiza ion o
glucose as an ac i e agen in neu onal damage a he han
a passi e epiphenomenon o disease se e i y. Howe e , as
media ion analyses do no imply causali y, he alidi y and
di ec ion o hese ela ionships dese e u he in es iga ion
in addi ional la ge p ospec i e obse a ional coho s udies
and in in e en ional p eclinical and clinical ials. Acco d-
ing o ou indings, he use o su oga e ma ke s o NAD
such as NFL le els could be aluable o he assessmen o
he biological e ec o in e en ional ials aimed o imp o e
glycemic con ol o speci ic pha macological agen s.
The main s eng h o his s udy was he use o a well-cha -
ac e ized p ospec i e coho o consecu i e SAH subjec s
admi ed in a Comp ehensi e S oke Cen e . Fu he mo e,
NFL le els we e assessed h ough a well- alida ed ou h
gene a ion Simoa echnique. Also, he media ion analysis
app oach, e en i no causa ion is implied, sheds ligh on he
po en ial glucose-d i en ha m ul pa hways igge ed du ing
EBI pe iod. Ne e heless, his s udy has se e al limi a ions.
Fi s , we used unadjus ed and adjus ed media ion analyses o
cha ac e ize o he associa ion be ween hype glycemia, NAD,
and clinical ou come in ou coho o SAH pa ien s. Gi en he
obse a ional na u e o his s udy, he es ima ed di ec and
indi ec e ec s o he associa ion be ween hype glycemia and
clinical ou come may ha e been a ec ed by unmeasu ed con-
ounding. The e o e, he media ion e ec ound in his s udy
should be in e p e ed wi h cau ion and equi es u he con i -
ma ion in ex e nal obse a ional coho s and in in e en ional
s udies [43]. Second, NFL le els we e assessed a he end o
he EBI pe iod as a su oga e bioma ke o b ain inju y o
e lec he amoun o NAD occu ing du ing he whole EBI
pe iod. Consequen ly, ou esul s migh be only applicable o
he subse o SAH pa ien s who su i e beyond 72h a e
bleeding onse . Thi d, his was an obse a ional s udy whe e
hype glycemia was managed acco ding o local in-hospi al
p o ocols and he e o e a cause-and-e ec ela ionship canno
be in e ed om he ob ained da a. Fou h, NAD was exclu-
si ely e alua ed using one single blood bioma ke in a single
ime poin . The assessmen o longi udinal se um NFL le els
and/o addi ional NAD su oga e bioma ke s such as hose
de i ed om ad anced quan i a i e MRI including mic o as-
cula , me abolic, and mic os uc u al in eg i y measu emen s
could be highly in o ma i e o iden i ying glucose p o iles
associa ed wi h highe b ain damage a e SAH. A be e
knowledge on he link be ween ad anced quan i a i e neu o-
imaging me ics o NFL le els and longi udinal glucose p o-
iles du ing he EBI pe iod is needed o unde s and he mos
app op ia e and in o ma i e iming and how o combine he
di e en su oga e ma ke s o acu e b ain inju y. E en ually,
his in o ma ion may lead o imp o e glucose managemen
p o ocols and o op imize he design o clinical ials aimed o
modula e glucose le els in he acu e phase a e SAH. Finally,
his epo lacks addi ional mechanis ic s udies on molecula
le els o explain how high glucose le els lead o inc eased
NFL le els. Fu he p eclinical and clinical s udies a e needed
o con i m his ela ionship and o unde s and he mechanis ic
link be ween glucose dis u bances and neu oaxonal damage
in his disease.
1475Molecula Neu obiology (2025) 62:1467–1477
Conclusions
Bo h highe mean glucose measu ed along he EBI pe iod
and highe NFL le els measu ed a he end o he EBI pe iod
we e independen ly associa ed wi h poo clinical ou come
a 90days. The media ion analyses sugges ed ha app oxi-
ma ely 25% o he a iabili y o he associa ion be ween
glucose and poo p ognosis could be media ed h ough neu-
oaxonal damage as measu ed ia ci cula ing NFL le els.
These esul s sugges ha he link be ween hype glycemia
and poo clinical ou come may be pa ly explained by sec-
onda y neu oaxonal inju y. Addi ional esea ch is needed o
con i m hese associa ions and o explo e he mechanis ic
link be ween hype glycemia and seconda y b ain damage.
Supplemen a y In o ma ion The online e sion con ains supplemen-
a y ma e ial a ailable a h ps:// doi. o g/ 10. 1007/ s12035- 024- 04347-6.
Acknowledgemen s We hank he Spanish Minis y o Economy
and Compe i i eness, Ins i u o de Salud Ca los III, and he Eu opean
Regional De elopmen Fund (ERDF) o he inancial suppo o he
co esponding au ho s. We also hank Fundació La Ma a ó de TV3 o
he inancial suppo o Se gio Ama o. We hank Hospi al Clinic de
Ba celona as well o he inancial suppo o he i s au ho .
Au ho Con ibu ion Se gio Ama o, Ramón To né, and Ángel Cham-
o o con ibu ed o he s udy concep ualiza ion. Daniel San ana, Ale-
jand a Mos ei o, Gab iel Pujol, Lau a Llull, Ma iano We ne , and Luigi
Za e a we e in cha ge o ma e ial p epa a ion and da a collec ion.
Ab aham Ma ín and Ca les Jus icia ad ised abou neu o ilamen meas-
u emen echniques. Da a analysis was pe o med by Se gio Ama o.
Lei e Ped osa gene a ed he igu es o he manusc ip . The i s d a
o he manusc ip was w i en by Daniel San ana, and all au ho s com-
men ed on p e ious e sions o he manusc ip . All au ho s ead and
app o ed he inal manusc ip .
Funding Open Access unding p o ided hanks o he CRUE-CSIC
ag eemen wi h Sp inge Na u e. Se gio Ama o and Ramón To né
ecei ed inancial suppo om a g an gi en by he Spanish Minis-
y o Economy and Compe i i eness [p ojec PI19/00936 unded by
Ins i u o de Salud Ca los III and co- unded by he Eu opean Regional
De elopmen Fund (ERDF)]. Se gio Ama o also ecei ed inancial
suppo om a g an gi en by Fundació la Ma a ó de TV3 (g an num-
be 17/C/2017). Daniel San ana is suppo ed by a g an om Hospi al
Clinic de Ba celona (Con ac e Clínic de Rece ca Emili Le ang-Josep
Fon ).
Da a A ailabili y The da ase s gene a ed and analyzed a e a ailable
om he co esponding au ho s on easonable eques .
Decla a ions
E hics App o al This s udy was pe o med in line wi h he p inciples
o he Decla a ion o Helsinki. E hics app o al was ob ained om he
local Clinical Resea ch E hics Commi ee om Hospi al Clinic o Ba -
celona (HCB/2017/027).
Consen o Pa icipa e Pa ien s o a legal p oxy ga e w i en in o med
consen p io o inclusion in he s udy.
Consen o Publica ion No applicable. No indi idual pe son’s da a
in any o m is shown.
Compe ing In e es s The au ho s decla e no compe ing in e es s.
Open Access This a icle is licensed unde a C ea i e Commons A i-
bu ion 4.0 In e na ional License, which pe mi s use, sha ing, adap a-
ion, dis ibu ion and ep oduc ion in any medium o o ma , as long
as you gi e app op ia e c edi o he o iginal au ho (s) and he sou ce,
p o ide a link o he C ea i e Commons licence, and indica e i changes
we e made. The images o o he hi d pa y ma e ial in his a icle a e
included in he a icle’s C ea i e Commons licence, unless indica ed
o he wise in a c edi line o he ma e ial. I ma e ial is no included in
he a icle’s C ea i e Commons licence and you in ended use is no
pe mi ed by s a u o y egula ion o exceeds he pe mi ed use, you will
need o ob ain pe mission di ec ly om he copy igh holde . To iew a
copy o his licence, isi h p://c ea i ecommons.o g/licenses/by/4.0/.
Re e ences
1. Macdonald RL, Schweize TA (2017) Spon aneous suba ach-
noid haemo hage. Lance 389:655–666. h ps:// doi. o g/ 10.
1016/ S0140- 6736(16) 30668-7
2. Claassen J, Pa k S (2022) Spon aneous suba achnoid haemo -
hage. Lance 400:846–862. h ps:// doi. o g/ 10. 1016/ S0140-
6736(22) 00938-2
3. Cahill J, Cal e JW, Zhang JH (2006) Mechanisms o ea ly
b ain inju y a e suba achnoid hemo hage. J Ce eb Blood
Flow Me ab 26:1341–1353. h ps:// doi. o g/ 10. 1038/ sj. jcb m.
96002 83
4. Rass V, Helbok R (2019) Ea ly b ain inju y a e poo -g ade
suba achnoid hemo hage. Cu Neu ol Neu osci Rep 19:78.
h ps:// doi. o g/ 10. 1007/ s11910- 019- 0990-3
5. Gae ani P, Pasqualin A, Rod iguez y Baena R, Bo asio E, Ma -
za ico F (1998) Oxida i e s ess in he human b ain a e suba-
achnoid hemo hage. J Neu osu g 89:748–54. h ps:// doi. o g/
10. 3171/ jns. 1998. 89.5. 0748
6. Pe zold A, Rejdak K, Belli A, Sen J, Kei G, Ki chen N, Smi h
M, Thompson EJ (2005) Axonal pa hology in suba achnoid and
in ace eb al hemo hage. J Neu o auma 22:407–414. h ps://
doi. o g/ 10. 1089/ neu. 2005. 22. 407
7. Kumme TT, Magnoni S, MacDonald CL, Dik anian K, Milne
E, So ell J, Con e V, Bene a os JJ e al (2015) Expe imen al
suba achnoid haemo hage esul s in mul i ocal axonal inju y.
B ain 138:2608–2618. h ps:// doi. o g/ 10. 1093/ b ain/ aw 180
8. Te polilli NA, B em C, Bühle D, Plesnila N (2015) A e we
ba king up he w ong essels? Ce eb al mic oci cula ion a e
suba achnoid hemo hage. S oke 46:3014–3019. h ps:// doi.
o g/ 10. 1161/ STROK EAHA. 115. 006353
9. Egashi a Y, Zhao H, Hua Y, Keep RF, Xi G (2015) Whi e ma e
inju y a e suba achnoid hemo hage: ole o blood-b ain ba -
ie dis up ion and ma ix me allop o einase-9. S oke 46:2909–
2915. h ps:// doi. o g/ 10. 1161/ STROK EAHA. 115. 010351
10. Ge agh y JR, Da is JL, Tes ai FD (2019) Neu oin lamma ion
and mic o ascula dys unc ion a e expe imen al suba ach-
noid hemo hage: eme ging componen s o ea ly b ain inju y
ela ed o ou come. Neu oc i Ca e 31:373–389. h ps:// doi. o g/
10. 1007/ s12028- 019- 00710-x
11. Pa sons MW, Ba be PA, Desmond PM, Bai d TA, Da by DG,
By nes G, T ess BM, Da is SM (2002) Acu e hype glycemia
ad e sely a ec s s oke ou come: a magne ic esonance imaging
and spec oscopy s udy. Ann Neu ol 52:20–28. h ps:// doi. o g/
10. 1002/ ana. 10241
12. Yu T, Robo ham JL, Yoon Y (2006) Inc eased p oduc ion o
eac i e oxygen species in hype glycemic condi ions equi es
dynamic change o mi ochond ial mo phology. P oc Na l Acad