Academic Edi o s: B en Siesky and
Yoshihi o Takamu a
Recei ed: 19 No embe 2024
Re ised: 12 Decembe 2024
Accep ed: 17 Janua y 2025
Published: 19 Janua y 2025
Ci a ion: Sabucedo-Villama in, B.;
Ga cia-Quei uga, J.; Pena-Ve deal, H.;
Ga cia-Resua, C.; Yeb a-Pimen el, E.;
Gi aldez, M.J. Diagnos ic Cu -O
Values Based on Lipid Laye Pa e n
o D y Eye Disease Sub ypes
Assessmen . J. Clin. Med. 2025,14, 623.
h ps://doi.o g/10.3390/
jcm14020623
Copy igh : © 2025 by he au ho s.
Licensee MDPI, Basel, Swi ze land.
This a icle is an open access a icle
dis ibu ed unde he e ms and
condi ions o he C ea i e Commons
A ibu ion (CC BY) license
(h ps://c ea i ecommons.o g/
licenses/by/4.0/).
A icle
Diagnos ic Cu -O Values Based on Lipid Laye Pa e n o D y
Eye Disease Sub ypes Assessmen
Belen Sabucedo-Villama in 1,* , Jacobo Ga cia-Quei uga 1,2 , Hugo Pena-Ve deal 1,2,* , Ca los Ga cia-Resua 1,2,
E a Yeb a-Pimen el 1,2 and Ma ia J. Gi aldez 1,2
1GI-2092-Op ome y, Depa amen o de Física Aplicada (Á ea de Op ome ía), Uni e sidade de San iago de
Compos ela, Campus Vida s/n, 15701 San iago de Compos ela, Spain; jacoboga [email p o ec ed] (J.G.-Q.);
ca los.ga cia. [email p o ec ed] (C.G.-R.); [email p o ec ed] (E.Y.-P.); [email p o ec ed] (M.J.G.)
2
Ins i u o de In es igación Sani a ia (IDIS), T a esía da Choupana S/N, 15701 San iago de Compos ela, Spain
*Co espondence: [email p o ec ed] (B.S.-V.); [email p o ec ed] (H.P.-V.)
Abs ac : Backg ound: The aim o he p esen s udy was o es ablish a cu -o alue o
he Lipid Laye Pa e n (LLP) be ween pa icipan s wi h di e en sub ypes o D y Eye
Disease (DED) including De icien D y Eye (ADDE), E apo a i e D y Eye (EDE), and
Mixed D y Eye (MDE). Me hods: 240 pa icipan s diagnosed wi h DED acco ding o
he Tea Film and Ocula Su ace Socie y in he D y Eye Wo kshop II guidelines we e
included in he s udy. Tea Meniscus Heigh (TMH) using he Tea scope illumina ion
and Meibomian Gland Loss A ea (MGLA) using he Ke a og aph 5M we e assessed o
ca ego ize he pa icipan s in o an ADDE g oup, EDE g oup, o MDE g oup. Then, he
LLP was assessed using he Tea scope ollowing he Guillon (LLP-G) and Colou (LLP-C)
schemes. Resul s: Recei e Ope a ing Cha ac e is ics (ROC) showed ha bo h LLP-G and
LLP-C ha e no diagnos ic po en ial in dis inguishing be ween ADDE and EDE pa icipan s
(bo h p
≥
0.724). Howe e , o di e en ia e he ADDE pa icipan s om he MDE, ROC
p ocedu es showed a good diagnos ic po en ial wi h cu -o alues o Closed Meshwo k-
Wa e (
AUC ±SD = 0.609 ±0.049
,p= 0.038, sensi i i y: 23.9%; speci ici y: 76.1%) and
G ey-Whi e (AUC
±
SD = 0.611
±
0.050, p= 0.034, sensi i i y: 40.7%; speci ici y: 73.9%)
o LLP-G and LLP-C, espec i ely. Also, a signi ican po en ial o dis inguish be ween
he EDE om MDE pa icipan s was ound, wi h cu -o alues o Closed Meshwo k
(
AUC ±SD = 0.604 ±0.049
,p= 0.043, sensi i i y: 40.8%; speci ici y: 76.1%) and G ey-
Whi e (AUC
±
SD = 0.604
±
0.051, p= 0.038, sensi i i y: 44.7%; speci ici y: 73.9%) o
LLP-G and LLP-C, espec i ely. Conclusions: Using he Tea scope, bo h LLP-G and LLP-C
has diagnos ic po en ial o dis inguish MDE pa icipan s om he o he sub ypes o DED.
Keywo ds: Lipid Laye Pa e n (LLP); Aqueous De icien D y Eye (ADDE); E apo a i e
D y Eye (EDE); Mixed D y Eye (MDE); cu -o alue
1. In oduc ion
D y Eye Disease (DED) is a p e alen condi ion, a ec ing 5% o 50% o he global
popula ion, posing a challenge o clinicians in hei daily ou ines [
1
,
2
]. This condi ion
is cha ac e ized by a hype osmola i y en i onmen , which comp omises he homeos asis
o he ea ilm. Consequen ly, he ea ilm becomes uns able, esul ing in a loss o ea
olume, a low b eak-up ime o he ea ilm and an inc eased a e o e apo a ion o he ea
om he ocula su ace [
3
]. This is called he “Vicious Ci cle” o DED [
4
]. Hence, a co ec
ea ilm s abili y is indispensable in he main enance o homeos asis and consequen ly
he ocula su ace in eg i y [
3
,
4
]. One o he componen s ha plays an essen ial ole in he
J. Clin. Med. 2025,14, 623 h ps://doi.o g/10.3390/jcm14020623
J. Clin. Med. 2025,14, 623 2 o 13
in eg i y and upkeep o he ea ilm is he lipid oil. This lipid oil composes he lipid laye
and p o ides bo h s abiliza ion and p e en ion om e apo a ion [3,5].
On he Tea Film and Ocula Su ace Socie y’s D y Eye Wo kshop II (TFOS DEWS II),
DED was classi ied in o wo main ca ego ies: Aqueous De icien D y Eye (ADDE), caused
by lac imal gland dys unc ion, and E apo a i e D y Eye (EDE), associa ed wi h eyelid
o meibomian gland abno mali ies [
6
]. Howe e , ADDE and EDE a e no wo sepa a e
en i ies; hey coexis and commonly o e lap, con ibu ing o a hi d ype, Mixed D y Eye
(MDE) [
4
]. MDE occu s when pa ien s ha e bo h aqueous ea de iciency and eyelids
and/o meibomian glands a ec ed, and i is es ima ed ha abou 30% o pa ien s wi h
DED may su e om his condi ion [
7
]. Howe e , i depends on which componen is he
mos a ec ed as o whe he i will end owa ds a p edominan ly e apo a i e o aqueous
de iciency [
8
]. The common me hods used o classi y DED pa ien s in o di e en sub ypes
imply he pe o ming o se e al es s: he assessmen o he ea ilm olume and he s a us
and mo phology o meibomian glands, which gene a es a ime bu den [8–10].
The lipid laye is an essen ial componen o he ea ilm ega dless o DED ype [
3
].
The lipid laye can be assessed by e alua ing he Lipid Laye Pa e n (LLP) using ocula
su ace in e e ome e s [
5
,
11
]. LLP can es ima e he hickness o he ea ilm lipid laye ,
which could be a po en ial ool o di e en ia e be ween DED ypes in a simple way. A
hinne LLP is a ibu ed o an EDE and a hicke one o an ADDE [
3
,
11
,
12
]. Cu en ly,
co ec ly iden i ying a pa ien ’s DED sub ype when conside ing he MDE sub ype, in
addi ion o he wo main sub ypes, can be ime-consuming o clinicians because i equi es
a leas wo diagnos ic es s. This highligh s he need o ind an easie and single diagnos ic
es , which simpli ies he e alua ion, as well as es ablish a cu -o c i e ion o dis inguish
be ween EDE and MDE o ADDE and MDE, which could es ablish a ca ego iza ion a he
han a endency o be one ype o he o he [
4
,
7
]. The e o e, he aim o he p esen s udy
was o p o ide a cu -o c i e ion ha s ongly disc imina es be ween he DED sub ypes
including he MDE h ough he LLP assessmen .
2. Ma e ials and Me hods
2.1. Sample
A o al o 240 Caucasian pa icipan s om he no hwes egion o Spain we e ec ui ed
om pa ien s a ending he Op ome y Se ice o ou ine eye examina ions. O hese,
185 we e women and 55 men, wi h a mean age o 48.3
±
16.5 yea s. The pa icipan s
we e selec ed based on hei compa ibili y wi h a DED diagnos ic based on TFOS DEWS
II c i e ia. [
6
]. No one had a p io his o y o ocula su ge y, sys emic, o au oimmune
diseases, we e p egnan o b eas - eeding, wo e con ac lenses o we e unde medical
ea men . W i en consen was ob ained om all pa icipan s, and he s udy p o ocol
ecei ed app o al om he ins i u ion’s Bioe hics Commi ee (USC-08/2021), ensu ing
adhe ence o he p inciples o he Decla a ion o Helsinki.
2.2. S udy Design and Diagnos ic C i e ia
As ou lined in he TFOS DEWS II Diagnos ic Me hodology epo , a se ies o clinical
es s we e conduc ed and documen ed by a single examine du ing one session o minimize
in e obse e and in e session a iabili y, wi h measu emen s subsequen ly aken by a
second blinded obse e [
11
]. The en i e s udy p o ocol was conduc ed unde con olled
en i onmen al condi ions, main aining consis en ligh , a empe a u e ange o 20–23
◦
C,
and humidi y le els be ween 50–60%.
P ocedu es we e pe o med om leas o mos in asi e and in he same o de o all
he pa icipan s: Ocula Su ace Disease Index (OSDI), ea ilm osmola i y, Tea Meniscus
J. Clin. Med. 2025,14, 623 3 o 13
Heigh (TMH) wi h Tea scope illumina ion LLP, Meibomian Gland Loss A ea (MGLA),
Fluo escein B eak Up Time (FBUT), and co neal s aining [6].
To diagnose DED, he ollowing c i e ia we e used: an OSDI sco e
≥
13 combined wi h
a leas one o he ollowing signs: ea ilm osmola i y
≥
308 mOsm/L, FBUT < 10 s, and/o
a co neal s aining sco e
≥
2 acco ding o he Ox o d Scheme (Figu e 1) [
4
,
6
,
13
,
14
]. Once he
pa icipan s we e diagnosed wi h DED, he sample was di ided in o h ee g oups ollowing
DED sub ypes desc ibed in he TFOS DEWS II Diagnos ic Me hodology epo [6,9]:
- ADDE sub ype: TMH ≤0.16 mm and MGLA < 50%.
- EDE sub ype: TMH > 0.16 mm and MGLA ≥50%.
- MDE sub ype: TMH ≤0.16 mm and MGLA ≥50%.
J. Clin. Med. 2025, 14, x FOR PEER REVIEW 3 o 14
P ocedu es we e pe o med om leas o mos in asi e and in he same o de o all
he pa icipan s: Ocula Su ace Disease Index (OSDI), ea ilm osmola i y, Tea Meniscus
Heigh (TMH) wi h Tea scope illumina ion LLP, Meibomian Gland Loss A ea (MGLA),
Fluo escein B eak Up Time (FBUT), and co neal s aining [6].
To diagnose DED, he ollowing c i e ia we e used: an OSDI sco e ≥ 13 combined
wi h a leas one o he ollowing signs: ea ilm osmola i y ≥ 308 mOsm/L, FBUT < 10 s,
and/o a co neal s aining sco e ≥ 2 acco ding o he Ox o d Scheme (Figu e 1) [4,6,13,14].
Once he pa icipan s we e diagnosed wi h DED, he sample was di ided in o h ee
g oups ollowing DED sub ypes desc ibed in he TFOS DEWS II Diagnos ic Me hodology
epo [6,9]:
- ADDE sub ype: TMH ≤ 0.16 mm and MGLA < 50%.
- EDE sub ype: TMH > 0.16 mm and MGLA ≥ 50%.
- MDE sub ype: TMH ≤ 0.16 mm and MGLA ≥ 50%.
Figu e 1. S udy design and diagnos ic c i e ia lowcha . DED = D y Eye Disease; OSDI = Ocula
Su ace Disease Index; FBUT= Fluo escein B eak Up Time; ADDE= Aqueous De icien D y Eye;
EDE= E apo a i e D y Eye; MDE= Mixed D y Eye. TMH = Tea Meniscus Heigh . MGLA=
Meibomian Gland Loss A ea.
2.3. E alua ion P ocedu es
2.3.1. Symp oma ology Assessmen
To quan i y he DED symp oma ology, he OSDI ques ionnai e was used, which
includes 12 ques ions o a one-week ecall, and was sel -adminis e ed ia a QR code
scanned on mobile de ices [6,15,16]. Sco es, anging om 0 o 100 poin s, we e assessed
by he examine ollowing s anda dized guidelines, wi h highe alues indica ing g ea e
disabili y [6,15,16].
2.3.2. Tea Film Osmola i y
Tea ilm osmola i y was measu ed using he Tea Lab osmome e (Tea Lab Co p,
San Diego, CA, USA) [17]. Pa icipan s we e sea ed and ins uc ed o look upwa ds while
Figu e 1. S udy design and diagnos ic c i e ia lowcha . DED = D y Eye Disease;
OSDI = Ocula
Su ace Disease Index; FBUT= Fluo escein B eak Up Time; ADDE= Aqueous De icien D y
Eye;
EDE= E apo a i e
D y Eye; MDE= Mixed D y Eye. TMH = Tea Meniscus Heigh .
MGLA= Meibomian Gland Loss A ea.
2.3. E alua ion P ocedu es
2.3.1. Symp oma ology Assessmen
To quan i y he DED symp oma ology, he OSDI ques ionnai e was used, which
includes 12 ques ions o a one-week ecall, and was sel -adminis e ed ia a QR code
J. Clin. Med. 2025,14, 623 4 o 13
scanned on mobile de ices [
6
,
15
,
16
]. Sco es, anging om 0 o 100 poin s, we e assessed
by he examine ollowing s anda dized guidelines, wi h highe alues indica ing g ea e
disabili y [6,15,16].
2.3.2. Tea Film Osmola i y
Tea ilm osmola i y was measu ed using he Tea Lab osmome e (Tea Lab Co p, San
Diego, CA, USA) [
17
]. Pa icipan s we e sea ed and ins uc ed o look upwa ds while
he de ice’s p obe was ca e ully posi ioned on he lowe ea meniscus. The examine
allowed he de ice o emi a beep, signalling ha he sample had been success ully col-
lec ed. [
17
]. The de ice ansla es he elec ical impedance o he sample in o osmola i y
alues (mOsm/L) wi hin a ange o 275 o 400 mOsm/L, displaying he esul s on i s
sc een [
17
]. All measu emen s we e conduc ed using es ca ds om he same lo o
ensu e consis ency.
2.3.3. Fluo escein B eak-Up Time
FBUT was assessed wi h he Ke a og aph 5M (Oculus Op ikge a e GmbH, We zla ,
Ge many) and he luo escein unc ion p o ided by he de ice [
18
,
19
]. The pa icipan s
we e p ope ly posi ioned and ins uc ed o look up o he ceiling. Then, a luo escein s ip
hyd a ed wi h saline was applied o he lowe bulba conjunc i a and pa icipan s we e
ins uc ed o blink se e al imes o ensu e an adequa e mixing o he dye [
20
]. Immedia ely
a e , hey we e asked o look s aigh a a ed do in he de ice and blink h ee imes
o eco d he FBUT ideos. This p ocedu e was epea ed h ee imes [
9
,
20
]. FBUT was
de ined as he ime in e al be ween he las blink and he appea ance o he i s da k
spo [
6
,
9
]. Once he ideos we e ex ac ed o he compu e , he FBUT was assessed using
Vi ualDub64 1.10.4, an open so wa e which con e s he ideo eco ded in o ames
(1 s = 8 ames) [9].
2.3.4. Co neal S aining
Ocula su ace damage was e alua ed h ough co neal s aining measu ed using he
Ke a og aph 5M, immedia ely a e eco ding FBUT ideos and using he same illumina-
ion [
18
,
21
]. Pa icipan s we e ins uc ed o look a a cen al ed do and pe o m he ou
gaze posi ions while being ideo eco ded [
20
,
22
]. A e eco ding and ex ac ing images,
co neal s aining was assessed using he Ox o d Scheme, which g ades damage se e i y
om 0 o 5: 0–1 (mild), 2–3 (mode a e), and 4–5 (se e e).
2.3.5. Tea Meniscus Heigh
TMH was e alua ed using a Tea scope in e e ome e (Tea scope, Keele , Windso ,
UK) a ached o a Topcon SL-D4 sli lamp (Topcon Co po a ion, Tokyo, Japan) [
12
]. To
s anda dize he obse a ion a ea ac oss all ideos, he Tea scope was ixed o he sli
lamp, main aining a consis en dis ance be ween he chin es and he de ice h oughou
he imaging p ocess. Pa icipan s we e posi ioned a he sli lamp, main aining hei
p ima y gaze while blinking na u ally o allow obse a ion o he lowe ea meniscus.
Videos o he meniscus we e eco ded using a Topcon DC4 came a (Topcon Co po a ion,
Japan) a ached o he sli lamp [
23
,
24
]. Images we e ex ac ed om he eco ded ideos
and analyzed wi h ImageJ 1.53 so wa e (Na ional Ins i u es o Heal h, Be hesda, MD;
(h p://imagej.nih.go /ij/ (accessed on 10 Oc obe 2024)) [
9
]. The ImageJ da a, ini ially in
pixels, we e con e ed o millime es o s a is ical analysis. Acco ding o a p e ious s udy,
300 pixels equa ed o 1 mm [9].
J. Clin. Med. 2025,14, 623 5 o 13
2.3.6. Meibomian Gland Loss A ea
The isualiza ion o he meibomian glands was pe o med wi h he Ke a og aph
5M. The in a ed illumina ion p o ided by he de ice acili a es he obse a ion o he
meibomian glands while he lids a e e e ed [
10
]. The pa icipan s we e p ope ly posi ioned
on o he de ice and eques ed o look up o he ceiling o e e he lowe eyelid. Se e al
meibog aphy images we e aken and expo ed om de Ke a og aph 5M o he compu e .
Images we e ex ac ed om he eco ded ideos and analyzed wi h ImageJ 1.53 so wa e
(Na ional Ins i u es o Heal h, Be hesda, MD; h p://imagej.nih.go /ij/ (accessed on
10 Oc obe 2024)) [
9
]. MGLA ca ego iza ion ollowed he Pul e al. [
25
] scale, ea u ing
ou g ades: G ade 1 (<25% MGLA), G ade 2 (25–50% MGLA), G ade 3 (50–75% MGLA),
and G ade 4 (>75% MGLA).
2.3.7. Lipid Laye Pa e n
LLP was also assessed using he Tea scope in e e ome e [
5
]. The Tea scope is an
in e e ome e ha p o ides isualiza ion o he LLP o he lipid laye o hickness es ima-
ion. Bo h he de ice and pa icipan s we e posi ioned in he same posi ion as he TMH
measu emen wi h hei sigh s aigh o he cen e o he de ice, and we e ins uc ed o
blink h ee imes wi hou squeezing. This p ocess was epea ed h ee imes [
5
,
11
,
26
]. The
en i e p ocess was ideo eco ded. Immedia ely, LLP images we e ex ac ed, a he p ecise
momen when he LLP was s abilized and o ally expanded o one second a e blinking.
Then, LLP images we e classi ied ollowing wo di e en scales.
-
Fi s , ollowing he basic Lipid Laye Pa e n Guillon’s (LLP-G) scheme in i e s eps
(Open Meshwo k, Closed Meshwo k, Wa e, Amo phous, o Colou ) wi h he in e -
media e o each as in e -ca ego ies [
26
]. A g ade om 1 o 5, wi h middle s eps, was
assigned o analyze hickness om hinnes o hickes .
-
Secondly, LLP images we e classi ied in ou s eps ollowing he Lipid Laye Pa e n
Colou (LLP-C) cha ac e is ics scheme (G ey, Whi e, Yellow, B own o highe ) and he
in e media e mix u es o colou s as in e -ca ego ies [
27
]. A g ade om 1 o 4, wi h
middle s eps, was assigned o analyze hickness om hinnes o hickes .
In he LLP image classi ica ion o bo h scales, he p edominan g ade p esen in he
images was chosen. In he absence o his g ade, he in e media e g ade was chosen.
2.3.8. S a is ical Analysis
The da a we e analyzed using SPSS s a is ical so wa e e sion 25.0 o Windows (SPSS
Inc., Chicago, IL, USA). The signi icance le el was es ablished a p
≤
0.05 o all s a is ical
es s. P io o conduc ing he analysis, an assessmen o da a no mali y was conduc ed
using he Kolmogo o –Smi no es [
28
,
29
]. Resul s indica ed ha osmola i y and TMH
da a ollowed a no mal dis ibu ion (Kolmogo o –Smi no , all p> 0.05), whe eas OSDI,
MGLA, co neal s aining, FBUT, and bo h LLP-G and LLP-C did no (Kolmogo o –Smi no ,
all p< 0.05). Desc ip i e s a is ics we e calcula ed using he mean and SD o pa ame ic
pa ame e s, and he median wi h in e qua ile ange (IQR) o non-pa ame ic pa ame e s.
The ange o minimum and maximum alues was epo ed o bo h ypes o da a. To
assess di e ences in pa ame e alues be ween DED sub ypes, an ANOVA analysis along
wi h Bon e oni pos hoc o pai ed analyses was used on pa ame ic pa ame e s, whe eas
K uskal–Wallis along wi h he Wilcoxon es o he pai ed measu emen we e applied on
non-pa ame ic pa ame e s [
30
]; Bon e oni co ec ion was applied on he Wilcoxon es by
adjus ing he signi icance alue by he numbe o compa isons [31].
The s udy de e mined he bes , bo h LLP-G and LLP-C, h eshold by bo h classi ica ion
me hods using he Recei e Ope a ing Cha ac e is ics (ROCs) analysis o di e en ia e be-
ween pa icipan s wi h di e en eye condi ions [
32
–
34
]. This p ocess in ol ed e alua ing
J. Clin. Med. 2025,14, 623 6 o 13
a ious h eshold alues and plo ing sensi i i y agains (1-speci ici y) o de e mine he
op imal h eshold. The model’s abili y o di e en ia e condi ions was assessed using he
A ea Unde he Cu e (AUC)
±
SD, wi h alues anging om 0 (no p edic ion) o 1 (pe ec
p edic ion). Addi ionally, he 95% Con idence In e als (CI) o he AUC we e calcula ed
(Mean
±
1.96
×
SD), and he op imal h eshold o each ROC cu e was selec ed using
Youden’s J s a is ic (J = sensi i i y + speci ici y −1).
To alida e he h eshold alue ob ained, a c oss- alida ion analysis was conduc ed.
By using SPSS commands, a sample o 80% o he da a was andomly selec ed, and he
LLPs a iables we e con e ed in o a bina y pa ame e . The associa ion wi h he ini ial
diagnosis was assessed using C ame ’s V, anging om 0 (no p edic ion) o 1 (pe ec
p edic ion). The associa ion be ween his new h eshold and he ini ial diagnosis is based
on he TFOS DEWS II Diagnos ic Me hodology epo using C ame ’s V, which anges
om 0 (no p edic i e abili y) o 1 (pe ec p edic i e abili y).
3. Resul s
Desc ip i e s a is ics o all he measu emen s o he sample a e p o ided in Table 1,
while desc ip i e s a is ics o all he measu emen s on each subg oup a e p o ided in
Table 2
. The analysis showed ha he e was no gene al s a is ical di e ence in he osmo-
la i y, FBUT, co neal s aining, LLP-G o LLP-C dis ibu ion be ween DED sub ype (all
p≥0.059
), whe eas a s a is ical di e ence was ound in he age, OSDI, TMH, and MGLA
alues (all p≤0.001) (Table 2).
Table 1. Desc ip i e s a is ics o he en i e sample. n = 240.
Age
(Yea s) *
OSDI
(Sco e) **
Osmola i y
(mOsm/L) *
FBUT (s) **
Co neal
S aining
(Ox o d
Scheme) **
TMH
(mm) * MGLA ** LLP-G ** LLP-C **
To al
Sample
(n = 240)
Mean/
Median 48.3 26.04 322.38 5.27 1.00 0.186 53.92 Closed
Meshwo k G ey/Whi e
SD/IQR 16.52 20.12–
37.50 18.55 3.43–8.15 0.00–2.00 0.098 41.48–59.67
Open Mesh-
wo k/Closed
Meshwo k—
Wa e
G ey/Whi e—
Whi e
Minimum
19.0 13.36 282.00 1.29 0.00 0.060 10.24 Open
Meshwo k G ey
Maximum
81.0 83.33 400.00 65.13 4.00 0.640 82.22 Colou
B own o highe
SD = S anda d De ia ion. IQR = In e qua ile Range. OSDI = Ocula Su ace Disease Index. FBUT = Fluo escein
B eak-Up Time. TMH = Tea Meniscus Heigh . MGLA = Meibomian Gland Loss A ea. LLP-G = Lipid Laye
Pa e n Guillon Scheme. LLP-C = Lipid Laye Pa e n Colou Scheme * Mean and SD displayed on pa ame ic
pa ame e s. ** The median and in e qua ile ange (IQR) we e used o ep esen non-pa ame ic pa ame e s.
Table 2. Desc ip i e s a is ics o he g oups.
Age
(Yea s) *
OSDI
(Sco e) **
Osmola i y
(mOsm/l) *
FBUT (s) **
Co neal
S aining
(Ox o d
Scheme) **
TMH
(mm) * MGLA ** LLP-G ** LLP-C **
ADDE
(n = 91)
Mean/
Median 44.54 25.00 321.47 5.29 1.00 0.128 38.47 Closed
Meshwo k G ey/Whi e
SD/IRQ 17.3 18.75–
34.09 17.60 3.42–8.85 0.00–2.00 0.021 28.22–43.51
Open Mesh-
wo k/Closed
Meshwo k—
Wa e/
Amo phous
G ey/Whi e—
Whi e/Yellow
Minimum
19.0 13.36 292.00 1.33 0.00 0.080 10.24 Open
Meshwo k G ey
Maximum
71.0 81.25 400.00 21.83 4.00 0.160 48.78 Colou
B own o highe
J. Clin. Med. 2025,14, 623 7 o 13
Table 2. Con .
Age
(Yea s) *
OSDI
(Sco e) **
Osmola i y
(mOsm/l) *
FBUT (s) **
Co neal
S aining
(Ox o d
Scheme) **
TMH
(mm) * MGLA ** LLP-G ** LLP-C **
EDE
(n = 103)
Mean/
Median 52.7 25.0 322.56 4.79 1.00 0.260 57.02 Closed
Meshwo k G ey/Whi e
SD/IRQ 15.5 20.0–36.36 19.65 3.38–7.38 0.00–2.00 0.110 54.18–62.90 Closed
Meshwo k—
Wa e
G ey/Whi e—
Whi e
Minimum
20.0 13.36 282.00 1.29 0.00 0.160 50.38 Open
Meshwo k G ey
Maximum
81.0 83.33 400.00 65.13 4.00 0.640 79.56 Colou
B own o highe
MDE
(n = 46)
Mean/
Median 46.0 35.42 323.78 5.77 1.00 0.129 59.16 Closed
Meshwo k G ey/Whi e
SD/IRQ 15.1 24.43–
43.23 18.15 4.02–9.67 0.00–2.00 0.022 55.70–65.01
Open
Meshwo k—
Closed
Meshwo k
G ey-Whi e
Minimum
20.0 13.50 284.00 1.75 0.00 0.060 50.38 Open
Meshwo k G ey
Maximum
70.0 75.00 373.00 23.13 4.00 0.160 82.22 Amo phous Yellow/B own
o highe
p0.001 ‡0.001 †0.784 ‡0.190 †0.655 †<0.001 ‡<0.001 †0.066 †0.059 †
SD = S anda d De ia ion. IQR = In e qua ile Range. OSDI = Ocula Su ace Disease Index. FBUT = Fluo escein
B eak-Up Time. TMH-Tc = Tea Meniscus Heigh . MGLA = Meibomian Gland Loss A ea. LLP-G = Lipid Laye
Pa e n Guillon Scheme. LLP-C= Lipid Laye Pa e n Colou Scheme, ADDE = Aqueous De iciency D y Eye,
EDE = E apo a i e D y Eye, MDE = Mixed D y Eye. * Mean and SD displayed on pa ame ic pa ame e s. ** The
median and in e qua ile ange (IQR) we e used o ep esen non-pa ame ic pa ame e s.
‡
ANOVA o epea ed
measu emen s. †K uskal–Wallis es .
3.1. Analysis o LLPs Cu -O Th eshold Values o Di e en ia e ADDE om EDE Pa icipan s
The pai wise analysis showed ha he e was no s a is ical di e ence in he OSDI,
osmola i y, FBUT, co neal s aining, LLP-G, o LLP-C dis ibu ion be ween g oups (all
p≥0.376
), whe eas a s a is ical di e ence was ound in he age, TMH, and MGLA and
alues (all p
≤
0.002) (Table 2). The ROC analysis indica ed ha bo h LLP-G and LLP-C
possesses no diagnos ic po en ial in dis inguishing be ween pa icipan sub ypes wi h an
AUC
±
SD = 0.515
±
0.042 (p= 0.724, 95% CI = 0.433–0.597) and AUC = 0.509
±
0.042
(p= 0.832, 95% CI = 0.427–0.591), espec i ely (Figu e 2).
J. Clin. Med. 2025, 14, x FOR PEER REVIEW 8 o 14
Figu e 2. The ROC cu e was gene a ed o assess he sensi i i y and speci ici y o he LLP in dis in-
guishing be ween ADDE and EDE based on heo e ical h esholds. The op imal cu -off alue was
selec ed a he in lexion poin o he cu e. n = 194. LLP = Lipid Laye Pa e n; ROC = Recei e
Ope a ing Cha ac e is ic; ADDE = Aqueous De iciency D y Eye; EDE = E apo a i e D y Eye.
3.2. Analysis o LLPs Cu -Off Th eshold Values o Diffe en ia e ADDE om MDE Pa icipan s
The pai wise analysis showed ha he e was no s a is ical diffe ence in age, osmola -
i y, FBUT, co neal s aining, o TMH dis ibu ion be ween g oups (all p ≥ 0.062), whe eas
a s a is ical diffe ence was ound in he OSDI, MGLA, LLP-G, and LLP-C alues (all p ≤
0.032) (Table 2).
The ROC analysis indica ed ha bo h LLP-G and LLP-C possess diagnos ic po en ial
in dis inguishing be ween pa icipan ypes wi h an AUC ± SD = 0.609 ± 0.049 (p = 0.038,
95% CI = 0.513–0.705) and AUC ± SD = 0.611 ± 0.050 (p = 0.034, 95% CI = 0.513–0.709),
espec i ely (Figu e 3). By compu ing he Youdens index o LLP-G (Youdens J s a is ic
= 0.179) o LLP-C (Youdens J s a is ic = 0.146), a cu -off alue o Closed Meshwo k–Wa e
(sensi i i y: 23.9%; speci ici y: 76.1%) and G ey-Whi e (sensi i i y: 40.7%; speci ici y:
73.9%) we e iden i ied o disc imina ing be ween ADDE and MDE pa icipan s, espec-
i ely. In he c oss- alida ion analysis using an 80% andom sample, an associa ion was
ound be ween bo h calcula ed LLP cu -off alues and he p e iously p oposed diagnos ic
c i e ia o TFOS DEWS II o dis inguishing be ween ADDE and EDE pa icipan s (bo h,
C amé s V ≥ 0.175, p ≤ 0.041).
Figu e 2. The ROC cu e was gene a ed o assess he sensi i i y and speci ici y o he LLP in
dis inguishing be ween ADDE and EDE based on heo e ical h esholds. The op imal cu -o alue
J. Clin. Med. 2025,14, 623 8 o 13
was selec ed a he in lexion poin o he cu e. n = 194. LLP = Lipid Laye Pa e n; ROC = Recei e
Ope a ing Cha ac e is ic; ADDE = Aqueous De iciency D y Eye; EDE = E apo a i e D y Eye.
3.2. Analysis o LLPs Cu -O Th eshold Values o Di e en ia e ADDE om MDE Pa icipan s
The pai wise analysis showed ha he e was no s a is ical di e ence in age, osmola i y,
FBUT, co neal s aining, o TMH dis ibu ion be ween g oups (all p
≥
0.062), whe eas a
s a is ical di e ence was ound in he OSDI, MGLA, LLP-G, and LLP-C alues (all
p≤0.032
)
(Table 2).
The ROC analysis indica ed ha bo h LLP-G and LLP-C possess diagnos ic po en ial in
dis inguishing be ween pa icipan ypes wi h an AUC
±
SD = 0.609
±
0.049 (p= 0.038, 95%
CI = 0.513–0.705) and AUC
±
SD = 0.611
±
0.050 (p= 0.034, 95% CI = 0.513–0.709), espec-
i ely (Figu e 3). By compu ing he Youden’s index o LLP-G (Youden’s J
s a is ic = 0.179
)
o LLP-C (Youden’s J s a is ic = 0.146), a cu -o alue o Closed Meshwo k–Wa e (sen-
si i i y: 23.9%; speci ici y: 76.1%) and G ey-Whi e (sensi i i y: 40.7%; speci ici y: 73.9%)
we e iden i ied o disc imina ing be ween ADDE and MDE pa icipan s, espec i ely.
In he c oss- alida ion analysis using an 80% andom sample, an associa ion was ound
be ween bo h calcula ed LLP cu -o alues and he p e iously p oposed diagnos ic c i e ia
o TFOS DEWS II o dis inguishing be ween ADDE and EDE pa icipan s (bo h, C amé ’s
V≥0.175, p≤0.041).
J. Clin. Med. 2025, 14, x FOR PEER REVIEW 9 o 14
Figu e 3. The ROC cu e was gene a ed o assess he sensi i i y and speci ici y o he LLP in dis in-
guishing be ween ADDE and MDE based on heo e ical h esholds. The op imal cu -off alue was
selec ed a he in lexion poin o he cu e. n = 137. LLP = Lipid Laye Pa e n; ROC = Recei e
Ope a ing Cha ac e is ic; ADDE = Aqueous De iciency D y Eye; MDE = Mixed D y Eye.
3.3. Analysis o LLPs Cu -Off Th eshold Values o Diffe en ia e EDE om MDE Pa icipan s
The pai wise analysis showed ha he e was no s a is ical diffe ence in age, osmola -
i y, FBUT, co neal s aining o MGLA dis ibu ion be ween g oups (all p ≥ 0.757), whe eas
a s a is ical diffe ence was ound in he OSDI, TMH, LLP-G, and LLP-C alues (all p ≤
0.036) (Table 2).
The ROC analysis indica ed ha bo h LLP-G and LLP-C possesses diagnos ic po en-
ial in dis inguishing be ween pa icipan ypes wi h an AUC ± SD = 0.604 ± 0.049 (p =
0.043, 95% CI = 0.508–0.700) and AUC ± SD = 0.604 ± 0.051 (p = 0.038, 95% CI = 0.504–0.704),
espec i ely (Figu e 3). By compu ing he Youdens index o LLP-G (Youdens J s a is ic
= 0.169) o LLP-C (Youdens J s a is ic = 0.617), a cu -off alue o Closed Meshwo k (sen-
si i i y: 40.8%; speci ici y: 76.1%) and G ey-Whi e (sensi i i y: 44.7%; speci ici y: 73.9%)
we e iden i ied o disc imina ing be ween EDE and MDE pa icipan s, espec i ely (see
Figu e 4). In he c oss- alida ion analysis using an 80% andom sample, a s ong associa-
ion was ound be ween bo h calcula ed LLP cu -off alues and he p e iously p oposed
diagnos ic c i e ia o TFOS DEWS II o dis inguishing be ween ADDE and EDE pa ici-
pan s (bo h, C amé s V ≥ 0.202, p ≤ 0.021).
Figu e 3. The ROC cu e was gene a ed o assess he sensi i i y and speci ici y o he LLP in
dis inguishing be ween ADDE and MDE based on heo e ical h esholds. The op imal cu -o alue
was selec ed a he in lexion poin o he cu e. n = 137. LLP = Lipid Laye Pa e n; ROC = Recei e
Ope a ing Cha ac e is ic; ADDE = Aqueous De iciency D y Eye; MDE = Mixed D y Eye.
3.3. Analysis o LLPs Cu -O Th eshold Values o Di e en ia e EDE om MDE Pa icipan s
The pai wise analysis showed ha he e was no s a is ical di e ence in age, osmola i y,
FBUT, co neal s aining o MGLA dis ibu ion be ween g oups (all p
≥
0.757), whe eas a
s a is ical di e ence was ound in he OSDI, TMH, LLP-G, and LLP-C alues (all p
≤
0.036)
(Table 2).
The ROC analysis indica ed ha bo h LLP-G and LLP-C possesses diagnos ic po-
en ial in dis inguishing be ween pa icipan ypes wi h an AUC
±
SD = 0.604
±
0.049
J. Clin. Med. 2025,14, 623 9 o 13
(
p= 0.043
, 95% CI = 0.508–0.700) and AUC
±
SD = 0.604
±
0.051 (p= 0.038, 95%
CI = 0.504–0.704
), espec i ely (Figu e 3). By compu ing he Youden’s index o LLP-G
(Youden’s J
s a is ic = 0.169
) o LLP-C (Youden’s J s a is ic = 0.617), a cu -o alue o Closed
Meshwo k (sensi i i y: 40.8%; speci ici y: 76.1%) and G ey-Whi e (sensi i i y: 44.7%;
speci ici y: 73.9%) we e iden i ied o disc imina ing be ween EDE and MDE pa icipan s,
espec i ely (see Figu e 4). In he c oss- alida ion analysis using an 80% andom sample, a
s ong associa ion was ound be ween bo h calcula ed LLP cu -o alues and he p e iously
p oposed diagnos ic c i e ia o TFOS DEWS II o dis inguishing be ween ADDE and EDE
pa icipan s (bo h, C amé ’s V ≥0.202, p≤0.021).
J. Clin. Med. 2025, 14, x FOR PEER REVIEW 10 o 14
Figu e 4. The ROC cu e was gene a ed o assess he sensi i i y and speci ici y o he LLP in dis in-
guishing be ween EDE and MDE based on heo e ical h esholds. The op imal cu -off alue was
selec ed a he in lexion poin o he cu e. n = 149. LLP = Lipid Laye Pa e n; ROC = Recei e
Ope a ing Cha ac e is ic; ADDE = Aqueous De iciency D y Eye; MDE = Mixed D y Eye.
4. Discussion
DED is a global condi ion ha ep esen s a challenge in bo h i s managemen and
diagnosis. In his con ex , he iden i ica ion o he sub ypes has a undamen al ele ance
[2]. DED has been mainly subdi ided in o ADDE and EDE subg oups, whe eas in he
daily p ac ice, his diffe en ia ion is no so s ic . To dis inguish be ween ADDE om EDE
pa icipan s, he e is a consensus among au ho s in he use o he TMH o Schi me as
po en ial diagnosis es s, whe e cu -off c i e ions ha e been s a ed [6,8,9,35,36]. Also,
MGLA has been used o g ade he se e i y o EDE sub ype [6,36]. Howe e , he e a e a
signi ican numbe o pa ien s who show a combina ion o signs o bo h ypes ha a e
o en difficul o diffe en ia e, hose known and classi ied as MDE [1,36,37]. The diagnosis
o MDE pa ien s lies in pe o ming he ba e y o speci ic es s om bo h ADDE and EDE
sub ypes [4,6]. This can be ime-consuming and c ea es he need o simpli ica ion o a
single diagnos ic es . The e o e, he use o he LLP assessmen could be use ul o es ab-
lishing cu -off alues o diffe en ia e be ween DED sub ypes.
P e ious esea che s ha e used he LipiView (J&J Su gical Vision Inc., I ine, CA,
USA) in e e ome e o measu e he lipid laye hickness, and ound hinne lipid laye s
in pa icipan s wi h obs uc i e meibomian gland dis unc ion and highe OSDI and
SPEED es s alues [38,39]. A i a e al. [40] measu ed he lipid laye hickness wi h he
LipiView in e e ome e and he LLP wi h he Kowa DR-1α, epo ing ha hicke lipid
laye s we e ela ed o he LLP o mul icolou ed in e e ome ic inges. Also, Remesei o
e al. [41] s a ed ha lowe lipid laye hickness was associa ed wi h lowe g ades on bo h
Guillons and Colou schemes. These indings a e consis en wi h hose o he p esen
s udy, speci ically because he mean LLP alues ob ained we e wi hin LLP-G om Open
Meshwo k o Closed Meshwo k and wi hin LLP-C om G ey o Whi e, co esponding
wi h he hinnes lipid laye s. Also, i should be no ed ha in he p esen s udy, he sample
included was en i ely composed o DED pa icipan s, he EDE sub ype being he mos
p e alen , which explains his end in he lowe LLP g ades [4,42]. Addi ionally, i is im-
po an o no e ha all g oups exhibi ed a simila dis ibu ion in e ms o he diagnos ic
Figu e 4. The ROC cu e was gene a ed o assess he sensi i i y and speci ici y o he LLP in
dis inguishing be ween EDE and MDE based on heo e ical h esholds. The op imal cu -o alue
was selec ed a he in lexion poin o he cu e. n = 149. LLP = Lipid Laye Pa e n; ROC = Recei e
Ope a ing Cha ac e is ic; ADDE = Aqueous De iciency D y Eye; MDE = Mixed D y Eye.
4. Discussion
DED is a global condi ion ha ep esen s a challenge in bo h i s managemen and
diagnosis. In his con ex , he iden i ica ion o he sub ypes has a undamen al ele ance [
2
].
DED has been mainly subdi ided in o ADDE and EDE subg oups, whe eas in he daily
p ac ice, his di e en ia ion is no so s ic . To dis inguish be ween ADDE om EDE
pa icipan s, he e is a consensus among au ho s in he use o he TMH o Schi me as
po en ial diagnosis es s, whe e cu -o c i e ions ha e been s a ed [
6
,
8
,
9
,
35
,
36
]. Also, MGLA
has been used o g ade he se e i y o EDE sub ype [
6
,
36
]. Howe e , he e a e a signi ican
numbe o pa ien s who show a combina ion o signs o bo h ypes ha a e o en di icul o
di e en ia e, hose known and classi ied as MDE [
1
,
36
,
37
]. The diagnosis o MDE pa ien s
lies in pe o ming he ba e y o speci ic es s om bo h ADDE and EDE sub ypes [
4
,
6
].
This can be ime-consuming and c ea es he need o simpli ica ion o a single diagnos ic
es . The e o e, he use o he LLP assessmen could be use ul o es ablishing cu -o alues
o di e en ia e be ween DED sub ypes.
P e ious esea che s ha e used he LipiView (J&J Su gical Vision Inc., I ine, CA,
USA) in e e ome e o measu e he lipid laye hickness, and ound hinne lipid laye s in
