Diagnostic accuracy of non-invasive tear film break-up time assessed by the simple manual interferometric device

Abstract

Aim
To determine the diagnostic accuracy of non-invasive tear film break-up time (NIBUT) measured by the handheld lipid layer examination instrument.

Methods
108 patients were enrolled in this cross-sectional study and divided into two groups: patients with dry eye (n = 57) categorized by the presence of dry eye symptoms obtained by Schein Questionnaire and minimally-one objective dry eye sign (tear film break-up time <10 s or corneal, conjunctival and lid margin fluorescein staining), and healthy subjects (n = 51).

Results
Dry eye subjects had significantly shorter NIBUT than healthy subjects (6 s vs 20 s, p < 0.001). Logistic regression analysis showed that shorter NIBUT values were excellent indicators of dry eye disease (p < 0.001), with consistency and no significant difference between measurements, even after standardizing the results for age and sex. NIBUT cut-off point to distinguish dry eye from healthy subjects was 12 s (sensitivity 90.2 %, specificity 88.5 %, PPV 92.5 %, NPV 85.2 %, LR +7.82, LR− 0.11, DOR 70.92, DE 89.6 %). Good, but lower accuracy was observed at cut-off value of 10 s (sensitivity 87.8 %, specificity 88.5 %, PPV 92.3 %, NPV 82.1 %, LR+ 7.61, LR− 0.14, DOR 55.2, DE 88.1 %). The area under the ROC curve (AUC) of 0.944 classified NIBUT as a diagnostic test with very high accuracy.

Conclusion
This study showed a high diagnostic accuracy of NIBUT measured by the handheld lipid layer examination instrument. This simple, reliable, objective and available instrument might regularly take place in routine, standard dry eye diagnostic and can be used by almost every eye specialist.

Full text

Contact Lens and Anterior Eye xxx (xxxx) xxx
Please cite this article as: Sania Vidas Pauk, Contact Lens and Anterior Eye, https://doi.org/10.1016/j.clae.2022.101776
1367-0484/© 2022 British Contact Lens Association. Published by Elsevier Ltd. All rights reserved.
Diagnostic accuracy of non-invasive tear lm break-up time assessed by the
simple manual interferometric device
Sania Vidas Pauk
a
, Igor Petriˇ
cek
a
,
b
, Martina Tomi´
c
c
, Tomislav Bulum
b
,
c
,
*
,
Sonja Jandrokovi´
c
a
,
b
, Maja Pauk Guli´
c
d
, Miro Kalauz
a
,
b
, Dina Leˇ
sin Ga´
cina
a
a
Department of Ophthalmology, Zagreb University Hospital Center, Zagreb, Croatia
b
School of Medicine, University of Zagreb, Zagreb, Croatia
c
Vuk Vrhovac University Clinic for Diabetes, Endocrinology and Metabolic Diseases, Merkur University Hospital, Zagreb, Croatia
d
Eye Polyclinic Opto Centar, Zagreb, Croatia
ARTICLE INFO
Keywords:
Tears
Dry eye disease
Non-invasive tear break-up time
Diagnostic tests
Specicity
Sensitivity
Area under the curve
ABSTRACT
Aim: To determine the diagnostic accuracy of non-invasive tear lm break-up time (NIBUT) measured by the
handheld lipid layer examination instrument.
Methods: 108 patients were enrolled in this cross-sectional study and divided into two groups: patients with dry
eye (n =57) categorized by the presence of dry eye symptoms obtained by Schein Questionnaire and minimally-
one objective dry eye sign (tear lm break-up time <10 s or corneal, conjunctival and lid margin uorescein
staining), and healthy subjects (n =51).
Results: Dry eye subjects had signicantly shorter NIBUT than healthy subjects (6 s vs 20 s, p <0.001). Logistic
regression analysis showed that shorter NIBUT values were excellent indicators of dry eye disease (p <0.001),
with consistency and no signicant difference between measurements, even after standardizing the results for
age and sex. NIBUT cut-off point to distinguish dry eye from healthy subjects was 12 s (sensitivity 90.2 %,
specicity 88.5 %, PPV 92.5 %, NPV 85.2 %, LR +7.82, LR−0.11, DOR 70.92, DE 89.6 %). Good, but lower
accuracy was observed at cut-off value of 10 s (sensitivity 87.8 %, specicity 88.5 %, PPV 92.3 %, NPV 82.1 %,
LR+7.61, LR−0.14, DOR 55.2, DE 88.1 %). The area under the ROC curve (AUC) of 0.944 classied NIBUT as a
diagnostic test with very high accuracy.
Conclusion: This study showed a high diagnostic accuracy of NIBUT measured by the handheld lipid layer ex-
amination instrument. This simple, reliable, objective and available instrument might regularly take place in
routine, standard dry eye diagnostic and can be used by almost every eye specialist.
1. Introduction
Evidence-based denition and a new classication of dry eye disease
(DED) are dened by Tear Film and Ocular Surface Society (TFOS) Dry
Eye WorkShop (DEWS) II [1]. Tear lm homeostasis disruption is the
unifying characteristic that describes the fundamental process in the
development of DED in which, in the rst place, tear lm instability,
together with other biomarkers (hyperosmolarity, ocular surface
inammation and damage, and neurosensory abnormalities), plays a
leading etiological role, accompanied by dry eye symptoms. The diag-
nostic test of the rst choice for the tear lm stability assessment, pro-
posed by TFOS DEWS II, is non-invasive tear break-up time (NIBUT)
since the instillation of uorescein during uorescein tear lm break-up
time (FBUT) measurement alters the ocular environment [2]. In
contrast, NIBUT does not need uorescein installation; it is non-invasive
and allows tear lm stability assessment in its unaltered, physiological
condition, providing more objective and reliable results [2,3].
TFOS DEWS II recommended performing NIBUT before any other
invasive tests. The patient should be instructed to blink naturally-three
times and then to cease blinking until instructions are given to blink
again, then to blink freely between measurements. NIBUT value of less
than, or equal to, 10 s should be taken as a cut-off point in Caucasians for
subjective methods [2]. However, guidelines did not clearly state by
which instrument standardized NIBUT measurement can be performed
* Corresponding author at: School of Medicine, University of Zagreb, Vuk Vrhovac University Clinic for Diabetes, Endocrinology and Metabolic Diseases, Merkur
University Hospital, Dugi dol 4a, Zagreb, Croatia.
E-mail address: tomislav.bulum@kb-merkur.hr (T. Bulum).
Contents lists available at ScienceDirect
Contact Lens and Anterior Eye
journal homepage: www.elsevier.com/locate/clae
https://doi.org/10.1016/j.clae.2022.101776
Received 17 June 2022; Received in revised form 1 October 2022; Accepted 19 October 2022

Contact Lens and Anterior Eye xxx (xxxx) xxx
2
[2]. They propose an automated measurement system where possible
since subjective measurements have been shown to vary between mea-
surement sessions and observers.
Until recently, the subjective Keeler Tearscope and Tearscope Plus
(Keeler, Windsor, UK) were the most frequently used to measure NIBUT
in clinical practice. However, they are not commercially available
anymore [4–6]. In present literature, there are descriptions of several
newly developed interferometers, Keeler Tearscope prototypes, Polaris
(bon Optic, Lübeck, Germany) [7,8], EasyTear Viewplus (Easytear,
Rovereto, Italy) [7,8], LacryDiag (Quantel Medical, Rockwall, US)
[9–11], and DR-II Tearscope Plus (Kowa, Japan). These devices analyze
the whole cornea and the tear lm at once. Thus, the rst break-up of the
tear lm might be detected accurately, especially when the method is
automated (LacryDiag). However, the lack of objective results and the
disadvantages of being expensive, time-consuming, complicated to use,
and inaccessible to most practitioners make them unsuitable for daily
routine work. Studies investigating their applicability and feasibility are
scarce and primarily based on the data from a healthy population on a
small number of patients [7–12], evidencing that currently available
devices and methods for assessing NIBUT are not routinely used and
NIBUT as a method itself.
The authors of this article presented a simple handheld instrument,
designed and built by Igor Petriˇ
cek, that works on principles of the
interferometry of tear lipid layer, primarily developed for lipid layer
thickness assessment and named by its author as “Handheld lipid layer
examination instrument” [13–16]. However, the rst author and her
colleagues observed during their clinical use of this instrument that it
can also be used for the assessment of non-invasive tear break-up
(NIBUT), which they performed, and investigated in their previous
studies [14–16]. The main advantage of this instrument is its simplicity,
availability, and time-sparing use that does not signicantly prolong
standard ophthalmologic examination, which is of great importance for
everyday routine work. It can be available to almost every eye specialist
since it is simple, uses a simple but proven methodology, and is not
expensive.
In the present study, the authors wanted to investigate and verify the
diagnostic accuracy of NIBUT measured by the handheld lipid layer
examination instrument and obtain the cut-off point with the most sig-
nicant sensitivity and specicity to distinguish between patients with
dry eye and healthy subjects.
2. Methods
This cross-sectional study was performed in the Department of
Ophthalmology, Zagreb University Hospital Center, following the
Declaration of Helsinki and approved by the Hospital’s Ethics Com-
mittee. The patients included in the study received both written and oral
information about the study and signed written informed consent.
A total of 108 subjects consecutively attending the Ophthalmology
Department over ve months were randomly selected and included in
the study by the rst author during her routine clinical work. At the
screening visit, subjects signed informed consent, and the rst author
obtained a case history using a standardized Schein questionnaire,
modied for issues relevant to this research to determine the extent of
dry eye symptoms [17–20]. Afterward, all the subjects underwent
standard ophthalmology examination, including NIBUT, FBUT, and
uorescein surface staining assessment. All measurements were per-
formed by a single examiner who was the rst author and double-
checked by the seventh author of this work. Measurements that
differed between the two observers were discarded. Inclusion criteria
required subjects to be 18 or older, not to be contact lens wearers, have
normal other anterior ocular surface ndings, and not use topical
ophthalmic medication. Exclusion criteria were previous ocular trauma,
acute infection, glaucoma, ocular surgery in the past year, other ocular
surface diseases and irregularities, systemic diseases or medications that
could alter the ocular surface, and poor cooperation.
2.1. Schein questionnaire
Clinical parameters screened were the severity of dry eye symptoms
using the Schein questionnaire [17–20]. Schein questionnaire is chosen
since it is the one commonly used in authors’ routine clinical work, and
therefore they have the most experience using it. This test is validated
for use in Croatian translation. The rst author performed the ques-
tionnaire. It is a disease-specic six-item questionnaire used to measure
patient-reported symptoms of dry eye that patients grade on a scale from
0 to 4 (0 – none, 1 – rarely, 2 – sometimes, 3 – often, and 4 – all the time).
Subscale scores can range from 0 to 24, with higher scores indicating
more problems or symptoms. It is simple, practical, short, and under-
standable to all patients, including the elderly, and it is not protected by
copyright. Therefore, it’s widely available without restriction. This
questionnaire’s disadvantage is that it does not evaluate the impact of
dry eye symptoms on patients’ vision, everyday activities, and quality of
life. It has no validated cut-off value for DED. In contrast, some other
questionnaires do, e.g., the Ocular Surface Disease Index (OSDI) and 5-
Item Dry Eye Questionnaire (DEQ-5).
2.2. Handheld lipid layer examination instrument
It is a simple, handheld lamp that can be used with any slit lamp and
not attached to it. It ts in a pocket and is cheap. It consists of 9 standard
single light-emitting diodes (LED) covered with uniformly opaque wide,
white plastic lter. A light source satises the Class 1 light source
criteria [15,16]. It projects diffuse, cold, and white light onto the corneal
surface to visualize the lipid layer with a slit lamp, on which the illu-
mination is turned off, and is estimated using a semiquantitative inter-
ferometric scale [21]. The area of the visualized cornea is large, the light
source is wide, and the light reection at the tear lm surface is bright
and uniform. When measuring NIBUT, the tear lm surface breaks are
observed in light reection (Fig. 1A and B). The instrument can be
moved freely, and therefore the light can be projected tangentially and
near the ocular surface. By doing so, breaks in the tear lm surface
(NIBUT) can be observed over the entire cornea. Despite not being
automated, its repeatability and reproducibility were excellent [13–16].
It might be debated whether semiquantitative interferometry per-
formed by specular illumination with a diffusor on the slit lamp does the
same [22]. However, it uses a very narrow slit light source that analyzes
only tiny parts of the corneal surface and cannot be placed in various
positions, especially tangentially. Its light source is far from the cornea,
not bright enough, and the tear lipid layer cannot be clearly visualized,
which also disables accurate NIBUT measurement.
2.3. Non-invasive tear break-up time test (NIBUT)
NIBUT was measured by the handheld lipid layer examination in-
strument using slit lamp magnication [14–16]. It was performed rst,
before any other tests. The procedure was repeated three times on both
eyes, right and left.
The instrument was rst placed temporally, tangentially, and close to
Fig. 1. A) and B). Tear lm break-up is visible in the uniform, white, and wide
light reection created on the tear lm surface by the handheld instrument.
S. Vidas Pauk et al.

Contact Lens and Anterior Eye xxx (xxxx) xxx
3
the eye’s surface, about 2–3 cm from the cornea, in a way that does not
touch the eye or eyelashes, to obtain the best visualization of NIBUT
(Fig. 2A). The examiner used slit lamp magnication (X 16) to contin-
uously observe breaks in the tear lm surface (Fig. 2A). Light projected
onto the cornea covered around one-fth of the corneal surface; thus,
the examiner moved the instrument to a different position, allowing the
visualization of the entire corneal surface. The subjects were asked to
blink three times, then refrain from blinking and look straight. At that
moment, the examiner started the stopwatch. The time between the
third blink and the rst irregularity that appeared on the surface of the
lipid layer was measured in seconds as NIBUT value for that measure-
ment. The rst irregularity was the rst observable discontinuation/
break in the tear lm surface (Fig. 2B, C, and D). Between measure-
ments, the patient was allowed to blink several times to reduce
discomfort. The authors provided a short video to better understand the
methodology (Video 1).
2.4. Tear break-up time test (TBUT)
TBUT was measured by standardized uorescein strips (Biotech,
Fluorescein Sodium Ophthalmic Strip USP) [2]. The procedure was
repeated three times on both eyes, right and left.
2.5. Fluorescein surface staining
Fluorescein corneal and conjunctival staining was assessed using
National Eye Institute/Industry Workshop (NEI) scale [18].
The tear lm osmolarity was not measured in this study since the
osmometer was unavailable to the authors.
Statistical analyses were performed by Statistica software package
version 14.0 (TIBCO Inc., USA) and SPSS software package version 23.0.
(IBM, USA). The normality of data distribution was tested by the
Kolmogorov-Smirnov test and the homogeneity of variance by the Leven
test. Results of descriptive analyses were expressed as median (min-
–max) for continuous data and numbers for categorical data. Differences
in distributions of continuous data were evaluated by nonparametric
tests since the assumption of homogeneity of variance for tested vari-
ables was not met. The Mann-Whitney test was used to test the differ-
ences between independent variables, and the Friedman ANOVA test
was used to test the differences between dependent variables. Differ-
ences in distributions of categorical data were assessed by the Chi-
square test. The Spearman rank correlation test was used to evaluate
the direction of associations, and binary logistic regression analysis was
used to assess their strength and independence. A receiver operating
characteristic (ROC) curve, the area under the ROC curve (AUC),
sensitivity and specicity, predictive values, likelihood ratios, diag-
nostic odds ratio, and diagnostic effectiveness (accuracy) were used to
assess the ability of the NIBUT test to discriminate between the dry eye
and healthy non-dry eye and to determine the NIBUT optimal cut-off
value (optimal decision threshold). p <0.05 was considered statisti-
cally signicant.
3. Results
This study included 108 subjects (31 males/77 females) with a me-
dian age of 61 years (min 18 – max 87). According to the dry eye
symptoms and signs, subjects were divided into two groups: subjects
with DED (n =57) and healthy subjects (n =51). Subjects with DED had
dry eye symptoms (Schein questionnaire score 1–24, additional ques-
tions score 1–8) and minimally-one dry eye sign [TBUT <10 s, uores-
cein ocular surface staining >5 corneal spots, >9 conjunctival spots, or
lid margin (≥2mm length and ≥25 % width)], whereas healthy or
subjects without DED were those who fell out of these criteria. They had
the Schein questionnaire score 0, additional questions score 0, and TBUT
≥10 s, uorescein ocular surface staining ≤5 corneal spots, ≤9
conjunctival spots, or lid margin (<2 mm length and <25 % width).
In subjects with DED, the median Schein questionnaire score was 6
(min 1 – max 12), and the median additional questions score 4 (min 1 –
max 6), whereas, in healthy subjects, both scores were 0. The median
value of three measurements of TBUT of both eyes in subjects with DED
was 4 s (min 2 – max 9) and in healthy subjects 18 s (min 10 – max 30).
Baseline characteristics of subjects included in the study are pre-
sented in Table 1. Subjects with DED were signicantly older than
healthy subjects (p =0.007), and there were signicantly more women
among subjects with DED than among healthy subjects (p =0.044).
The groups differed signicantly in all three measurements of NIBUT
of both eyes (p <0.001) (Table 2). Subjects with DED had signicantly
shorter NIBUT than healthy subjects in all measurements. No signicant
difference between the three NIBUT measurements of both eyes was
observed in subjects with DED (p =0.115) nor healthy subjects (p =
0.081) (Table 2). Since there was no signicant difference within both
study groups between the three measurements of NIBUT of the right and
left eye, in the further statistical analyses, the median value of all six
measurements of NIBUT of both eyes was used.
NIBUT was signicantly negatively correlated with age (p =0.033),
female gender (p =0.005), dry eye symptoms evaluated by Schein (p <
0.001) and additional questionnaire (p <0.001), and signicantly
positively with TBUT (p <0.001) (Table 3).
Logistic regression analysis showed the age (OR =1.05, p =0.041)
and female gender before (OR =3.02, p =0.048), and especially after
adjustment for age (AOR =7.21, p =0.007) as the main predictors of
DED. Shorter NIBUT values indicated and increased the prevalence of
DED to a signicant extent (p <0.001), with consistency and no sig-
nicant difference between the measurements, even after adjustment of
the results for age and gender (Table 4).
The ROC curve passing near the upper left angle and the area under
the ROC curve (AUC) of 0.944 represent NIBUT as a diagnostic test with
very high accuracy (Table 5, Fig. 3). Using the Index of Union (IU)
Fig. 2. A) Non-invasive tear break-up time measurement by the handheld in-
strument. Diffuse, white, and cold light is projected onto the cornea, where
uniform reection is created on the surface of the tear lm. B), C) and D) The
appearance of the rst irregularity in the uniform reection at the moment of
tear lm rupture after the blink.
Table 1
Baseline characteristics of subjects included in the study (n =108).
Subjects with DED
(n =57)
Healthy subjects
(n =51)
Z
a
Chi
b
p
Age (years)* 65 (32–87) 51 (18–84) 2.721
a
0.007
a
Gender (m/f)** 11/46 20/31 4.069
b
0.044
b
* median (min–max) ** numbers
a
Mann-Whitney test
b
Chi-square test df =1.
Abbreviations: DED – dry eye disease. bold p<0.05.
S. Vidas Pauk et al.

Contact Lens and Anterior Eye xxx (xxxx) xxx
4
method, the NIBUT optimal cut-off value, the value whose sensitivity
and specicity are the closest to the value of the AUC and the absolute
value of the difference between the sensitivity and specicity values is
minimum, was 12 s.
Other diagnostic accuracy measures of NIBUT are shown in Table 6.
At the cut-off value of 12 s, NIBUT showed 90.2 % sensitivity and 88.5 %
specicity for detecting DED. As the sensitivity and specicity for a cut-
off value of 12 s were high, positive and negative predictive values for
detection of DED were 92.5 % and 85.2 %, and positive and negative
likelihood ratios, diagnostic odds ratio, and diagnostic effectiveness
(accuracy) were 7.82, 0.11, 70.92 and 89.6 %, respectively. By contrast,
at a cut-off value of 10 s, NIBUT showed lower accuracy in all used
measures (Table 6).
4. Discussion
The main goal of this study was to investigate the diagnostic accu-
racy of NIBUT measurement performed by handheld instrument and to
determine the cut-off value with the best sensitivity and specicity for
detecting DED. The subjects in this study were divided according to
DED, whereas in previous studies [14–16], participants were classied
according to the presence of dry eye symptoms only. Therefore, the
present study provided a more objective classication of the patients
Table 2
Three NIBUT measurements of both eyes in subjects divided into two groups
according to the presence of dry eye disease.
Subjects with
DED
(n =57)
Healthy
subjects
(n =51)
Z
a
p
a
1. NIBUT of the right eye
(seconds)
6 (3–20) 20 (7–62) −6.086
a
<0.001
a
2. NIBUT of the right eye
(seconds)
6 (2–16) 21.5 (7–51) −6.420
a
<0.001
a
3. NIBUT of the right eye
(seconds)
7 (2–17) 20 (8–52) −6.453
a
<0.001
a
1. NIBUT of the left eye
(seconds)
6 (2–24) 20 (8–60) −6.157
a
<0.001
a
2. NIBUT of the left eye
(seconds)
6 (3–18) 22.5 (7–50) −6.453
a
<0.001
a
3. NIBUT of the left eye
(seconds)
5 (2–17) 22.5 (8–45) −6.362
a
<0.001
a
Chi
b
8.861
b
10.381
b
p
b
0.115
b
0.081
b
Med (Min-Max)
a
Mann-Whitney test
b
Friedman ANOVA test df =5.
Abbreviations: NIBUT – non-invasive break-up time; DED – dry eye disease. bold
p<0.05
Table 3
Correlations between NIBUT, age, gender, dry eye symptoms, and TBUT.
NIBUT
Spearman R t(N-2) p
Age −0.261 −2.182 0.033
Gender (female) −0.349 −4.709 0.005
Schein questionnaire −0.749 −9.124 <0.001
Additional questionnaire −0.695 −7.798 <0.001
TBUT 0.761 9.451 <0.001
Abbreviations: NIBUT – non-invasive break-up time; TBUT – tear break-up time.
bold p<0.05.
Table 4
Predictors and indicators of dry eye disease by means of logistic regression
analysis.
OR (95 %CI) p AOR (95 %CI)* p*
Age (years) 1.05
(1.02–1.09)
0.041 / /
Gender (female) 3.02
(0.98–9.25)
0.048 7.21
(1.66–31.34)
0.007
1. NIBUT of the right
eye
0.69
(0.57–0.82)
<0.001 0.51 (0.33–0.78) 0.002
2. NIBUT of the right
eye
0.60
(0.46–0.79)
<0.001 0.55 (0.38–0.81) 0.002
3. NIBUT of the right
eye
0.61
(0.47–0.79)
<0.001 0.56 (0.39–0.79) 0.001
1. NIBUT of the left
eye
0.70
(0.60–0.84)
<0.001 0.65 (0.51–0.82) 0.001
2. NIBUT of the left
eye
0.66
(0.53–0.81)
<0.001 0.59 (0.39–0.82) 0.002
3. NIBUT of the left
eye
0.68
(0.56–0.84)
<0.001 0.63 (0.48–0.84) 0.001
* OR standardized for age or age and gender.
Abbreviations: NIBUT - non-invasive break-up time. bold p<0.05.
Table 5
Area under the ROC curve and 95% condence interval of NIBUT in subjects
included in the study.
AUC Std. Error 95 % CI p
NIBUT 0.944 0.027 0.89–0.99 0.000
Abbreviations: NIBUT - non-invasive break-up time; AUC - area under the ROC
curve.
Fig. 3. ROC curve of NIBUT in subjects included in the study.
Table 6
Diagnostic accuracy measures of NIBUT in subjects included in the study.
NIBUT ≤10 s NIBUT ≤12 s
Estimate 95 % CI Estimate 95 % CI
Sensitivity 87.8 % 73.8–95.9 % 90.2 % 76.9–97.3 %
Specicity 88.5 % 69.9–97.6 % 88.5 % 69.9–96.6 %
PPV 92.3 % 80.5–97.2 % 92.5 % 80.9–97.3 %
NPV 82.1 % 66.7–91.4 % 85.2 % 69.2–93.7 %
LR+7.61 2.61–22.19 7.82 2.69–22.78
LR−0.14 0.06–0.32 0.11 0.04–0.28
DOR 55.20 12.02–253-39 70.92 14.54–345.98
DE 88.1 % 77.8–94.7 % 89.6 % 79.7 %-95.7 %
Abbreviations: NIBUT – non-invasive break-up time; PPV – positive predictive
value; NPV – negative predictive value; LR+– positive likelihood ratio; LR−–
negative likelihood ratio; DOR – diagnostic odds ratio; DE – diagnostic effec-
tiveness (accuracy).
S. Vidas Pauk et al.

Contact Lens and Anterior Eye xxx (xxxx) xxx
5
and, consequently, more accurate results.
Subjects with DED in the present study had a median NIBUT of 6 s,
and healthy subjects had a median NIBUT of 20 s. The values between
the two groups differed signicantly, with p <0.001 in all measure-
ments on both eyes. Measured in this manner, NIBUT showed excellent
repeatability and reproducibility since the values did not differ signi-
cantly in all three measurements in both eyes. The method showed ac-
curate results from the rst measurement; therefore, even the rst
measurement can be considered objective.
The authors conrmed that NIBUT measurements obtained using the
handheld instrument were within the results published in recent liter-
ature [16]. In the meantime, several more papers on NIBUT measure-
ment with modern instruments have been published [8–11,23]. The
results of those papers are variable, and the methodology and study
protocols signicantly differ. However, there was substantial agreement
between NIBUT values in the present study and the studies published by
Remongin et al. and Vigo et al. in dry eye groups [11,23]. In the former
research, NIBUT was assessed by LacryDiag (Quantel Medical, France)
and IDRA Plus (SBM Sistemi, Turin, Italy); an all-in-one device was used
in the latter, both allowing the automated measurement of NIBUT. The
present study results were also in agreement with T´
oth et al. They pre-
sented the rst data on NIBUT measurement using LacryDiag (Quantel
Medical, France) in a group of healthy subjects [8]. They found that the
NIBUT values distributed in the healthy population were more likely
>12 s.
The cut-off value of 10 s was recently proposed by TFOS DEWS II
diagnostic methodology report for NIBUT assessment using subjective
methods [2]. In this study, the NIBUT cut–off value that distinguished
the subjects with DED and healthy subjects with the most signicant
sensitivity (90.2 %), specicity (88.5 %), AUC, positive and negative
predictive values (92.5 % and 85.2 %), and positive and negative like-
lihood ratios, diagnostic odds ratio, and diagnostic effectiveness (accu-
racy) (7.82, 0.11, 70.92 and 89.6 %), was 12 s. On the other hand,
sensitivity (87.8 %), specicity (88.5 %), and diagnostic accuracy were
lower even though exquisite when a cut-off value of 10 s was chosen.
The exact cut-off value of 12 s for detecting dry eye symptoms was
found in a preliminary study conducted by the same authors on a smaller
sample [15]. Moreover, T´
oth and the authors also found substantial
agreement with NIBUT cut-off point of 12 s and only slight agreement
with a cut-off point of 10 s when using automated LacryDiag [9].
The shape of the receiver operating characteristic curve (ROC)
together with the high AUC 0.944 obtained in the present study also
showed excellent diagnostic efcacy and accuracy of NIBUT measured
by a handheld instrument in detecting dry eye with a 12 s cut-off point
and slightly worse diagnostic abilities with a 10 s cut-off point.
In recent decades, numerous NIBUT assessment methods have been
published [7–12,15,16,23–30]. However, only a few reports showed
diagnostic accuracies and cut-off values of the method investigated
therein [12,23–29]. The lack of validated NIBUT cut-off value for
reaching the diagnosis of DED hampered the broad adoption of many
mentioned instruments in clinical practice. Furthermore, the reported
cut-off values of different subjective and automated methods for NIBUT
assessment range widely from 2,63 s to 12,1 s, with sensitivity and
specicity ranging between 70 % −95.5 % and 60 % −95 %
[12,23–29]. This discrepancy may result from different denitions of
dry eye and the methodologies and protocols used, making the studies
hardly comparable. So, the question is how adequate it is to apply the
same cut-off value of 10 s to all the methods assessed by using different
instruments, i.e., whether the values are comparable or are instrument
and method related. However, the results of our study were in good
agreement with ndings reported by Downie and coworkers [12]. In
their research, the NIBUT cut-off value with the best discriminative
capacity was 12, 1 s. Our results also agreed with Pult et al., where
NIBUT was assessed by Tearscope (Keeler, Windsor, UK) [29]. At a cut-
off point of 10 s, both methods had similar sensitivities and specicities.
Therefore, diagnostic parameters for the method described in the
present study are within the limits of published data, proving its
objectivity.
Furthermore, the present study’s logistic regression analysis pre-
sented NIBUT with the age and female gender signicantly connected to
DED. That highlighted the ability of short NIBUT to be an excellent
predictor, indicator, and discriminator of DED.
In conclusion, NIBUT measured using a handheld lipid layer exam-
ination instrument showed excellent repeatability, objectivity, consis-
tency, and accuracy in dry eye diagnostics.
Authors must declare potential limitations and biases regarding this
study. An important limitation of this study was the use of the Schein
questionnaire. Even though this questionnaire has scoring responses, it
has no validated cut-off value for DED, which might potentially impact
the classication of the patients into the dry eye and no-dry eye groups.
That can affect the evaluation of diagnostic metrics of a test compared to
results of other studies that used the OSDI or the DEQ-5 and TFOS DEWS
II recommended cut-off value. However, despite using this question-
naire, the present study results were in good agreement with the results
of other studies and DEWS II recommendations [2,9,11–12,16,23,29].
Additionally, the Schein questionnaire does not evaluate the impact
of dry eye on subjects’ vision, everyday activities, and quality of life,
while some other questionnaires do. However, the authors have chosen
it since it is routinely used in their clinical work, they have the most
experience using it in routine practice, and it is not protected by copy-
right in contrast to OSDI and DEQ-5.
Another potential limitation is that the handheld instrument cannot
analyze the whole corneal surface at once; thus, the examiner must
continuously move the instrument to a different position in front of the
cornea to analyze the entire cornea. Therefore, there is a slight possi-
bility of missing the rst break-up of the tear lm. LacryDiag is an
automated device where NIBUT software automatically detects minimal
and rst tear lm irregularities, blinks, records the interblink interval
and calculates the NIBUT.
In the end, TFOS DEWS II denes tear lm osmolarity as an essential
biomarker of dry eye. Unfortunately, it was not used in this study since it
is unavailable to the authors in their clinical work [2]. Using the TearLab
Osmolarity System will probably result in more objective patient cate-
gorization [31].
5. Conclusion
NIBUT measured using the handheld lipid layer examination in-
strument detected DED with high sensitivity and specicity. It showed
great consistency with dry eye signs and symptoms and was a condent
indicator, predictor, and discriminator of DED. The cut-off point with
the most signicant sensitivity of 90.2 %, specicity of 88.5 %, and AUC
of 0.944 for dry eye was 12 s, though the method also showed very good
but lower accuracy with the cut-off point of 10 s. Thus, NIBUT measured
this way was demonstrated as an accurate, reliable, objective, and, most
importantly, simple, accessible, and cheap method that can nd a place
in routine, daily dry eye diagnostics and might be used by almost every
eye clinician. Also, since instilling standard Fluorescein to perform
TBUT hinders the consequent application of soft contact lenses, this
diagnostic procedure might be particularly useful in contact lens
practice.
Funding
This research did not receive any specic grant from funding
agencies in the public, commercial, or not-for-prot sectors.
Declaration of Competing Interest
The authors declare that they have no known competing nancial
interests or personal relationships that could have appeared to inuence
the work reported in this paper.
S. Vidas Pauk et al.

Contact Lens and Anterior Eye xxx (xxxx) xxx
6
Acknowledgment
Authorship: All authors mentioned fulll the criteria for authorship as
per ICMJE guidelines. Anyone who does not meet the criteria has not
been included.
Financial disclosure: No author has a nancial or proprietary interest
in any material or method mentioned.
Informed consent: Informed consent was obtained from all individual
participants included in the study.
The persons/participants/demonstrators shown in the Figures are
not the subjects included in the study. They are the authors of the article.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi.
org/10.1016/j.clae.2022.101776.
References
[1] Craig JP, Nichols KK, Akpek EK, Caffery B, Dua HS, Joo C-K, et al. TFOS DEWS II
denition and classication report. Ocul Surf 2017;15(3):276–83.
[2] Wolffsohn JS, Arita R, Chalmers R, Djalilian A, Dogru M, Dumbleton K, et al. TFOS
DEWS II diagnostic methodology report. Ocul Surf 2017;15(3):539–74.
[3] Sweeney DF, Millar TJ, Raju SR. Tear lm stability: a review. Exp Eye Res 2013;
117:27–38.
[4] Tearscope. Tearscope Plus Clinical Handbook and Tearscope Plus Instructions.
Windsor, Keeler Ltd, Windsor, Berkshire; Keeler Insts Inc., Broomall, PA, 1997.
[5] Guillon JP. Use of the Tearscope Plus and attachments in the routine examination
of the marginal dry eye contact lens patient. Adv Exp Med Biol 1998;438:859–67.
https://doi.org/10.1007/978-1-4615-5359-5_121.
[6] Tong L, Teng LS. Review of literature on measurements of non-invasive break up
times, lipid morphology and tear meniscal height using commercially available
handheld instruments. Curr Eye Res 2018;43(5):567–75. https://doi.org/10.1080/
02713683.2018.1437454.
[7] Bandlitz S, Pult H. Advances in tear lm assessment. Optom Pract 2016;17:81–90.
[8] Bandlitz S, Peter B, Pugi T, Jaeger K, Anwar A, Bikhu P, et al. agreement and
repeatability of four different devices to measure non-invasive tear breakup time
(NIBUT). Cont Lens Anterior Eye 2020;43(5):507–11.
[9] T´
oth N, Szalai E, R´
ak T, Lillik V, Nagy A, Csutak A. Reliability and clinical
applicability of a novel tear lm imaging tool. Graefes Arch Clin Exp Ophthalmol
2021;259(7):1935–43.
[10] Ward CD, Murchison CE, Petroll WM, Robertson DM. Evaluation of the
repeatability of the LacryDiag ocular surface analyzer for assessment of the
meibomian glands and tear lm. Transl Vis Sci Technol 2021;10(9):1. https://doi.
org/10.1167/tvst.10.9.1.
[11] Remongin P-E, Rousseau A, Best A-L, Ben Hadj Salah W, Legrand M, Benichou J,
et al. Place du NIBUT automatis´
e dans l’´
evaluation multimodale de la surface
oculaire dans le syndrome sec avec le Lacrydiag. J Français d’Ophtalmol 2021;44
(3):313–20.
[12] Downie LE. Automated tear lm surface quality breakup time as a novel clinical
marker for tear hyperosmolarity in dry eye disease. Invest Ophthalmol Vis Sci
2015;56(12):7260–8. https://doi.org/10.1167/iovs.15-17772.
[13] Petriˇ
cek I. Handheld tear lm lipid layer thickness and non-invasive tear break up
time assessment tool. J Clin Ophthalmol Eye Disord 2017;1(2):1009.
[14] Vidas PS. Non-invasive tear break-up time measurement using handheld lipid layer
thickness assessment tool. University of Zagreb School of Medicine; 2019. Doctoral
thesis: Zagreb.
[15] Vidas Pauk S, Petriˇ
cek I, Juki´
c T, Popovi´
c-Sui´
c S, Tomi´
c M, Kalauz M, et al. Non-
invasive tear lm break-up time assessment using handheld lipid layer examination
instrument. Acta Clin Croat 2019;58(1):63–71. https://doi.org/10.20471/
acc.2019.58.01.09.
[16] Pauk SV, Petriˇ
cek I, Tomi´
c M, Bulum T, Jandrokovi´
c S, Kalauz M, et al. Manual
interferometric device for routine non-invasive tear lm break-up time assessment.
Semin Ophthalmol 2021;36(3):94–102.
[17] Schein OD, Tielsch JM, Mun˜
oz B, Bandeen-Roche K, Westet S, al.. Relation between
signs and symptoms of dry eye in the elderly. A population-based perspective.
Ophthalmology 1997;104(9):1395–401.
[18] Lemp MA, Foulks GN. The denition and classication of dry eye disease: report of
denition and classication subcommittee of international dry eye workshop. Ocul
Surf 2007;5:75–92.
[19] Bakija I, Filipˇ
ci´
c I, Bogadi M, ˇ
Simunovi´
c Filipˇ
ci´
c I, Gotovac M, Kaˇ
stelan S.
Comparison of the Schein and Osdi Questionnaire as Indicator of Tear Film
Stability in Patients with Schizophrenia. Psychiatr Danub 2021;33:596–603.
[20] Bandeen-Roche K, Mu˜
noz B, Tielsch JM, West SK, Schein OD. Self-reported
assessment of dry eye in a population-based setting. Invest Ophthalmol Vis Sci
1997;38:2469–75.
[21] McDonald JE. Surface phenomena of tear lms. Trans Am Ophthalmol Soc 1968;
66:905–39.
[22] Norn MS. Semiquantitative interference study of fatty layer of precorneal lm. Acta
Ophthalmol (Copenh) 1979;57(5):766–74.
[23] Vigo L, Pellegrini M, Bernabei F, Carones F, Scorcia V, Giannaccare G, et al.
Diagnostic performance of a novel non-invasive workup in the setting of dry eye
disease. J Ophthalmol 2020;2020:1–6.
[24] Hong J, Sun X, Wei A, Cui X, Li Y, Qian T, et al. Assessment of tear lm stability in
dry eye with a newly developed keratograph. Cornea 2013;32(5):716–21.
[25] Mengher LS, Pandher KS, Bron AJ. Non-invasive tear lm break-up time: sensitivity
and specicity. Acta Ophthalmol (Copenh) 1986;64(4):441–4. https://doi.org/
10.1111/j.1755-3768.1986.tb06950.x.
[26] Wang HF, Fukuda M, Shimomura Y. Diagnosis of dry eye. Semin Ophthalmol 2005;
20(2):53–62. https://doi.org/10.1080/08820530590931115.
[27] Gumus K, Crockett CH, Rao K, Yeu E, Weikert MP, Shirayama M, et al. Non-
invasive assessment of tear stability with the tear stability analysis system in tear
dysfunction patients. Invest Ophthalmol Vis Sci 2011;52(1):456.
[28] Bhandari V, Reddy JK, Relekar K, Ingawale A, Shah N. Non-invasive assessment of
tear lm stability with a novel corneal topographer in Indian subjects. Int
Ophthalmol 2016;36(6):781–90. https://doi.org/10.1007/s10792-016-0186-7.
[29] Pult H, Purslow C, Murphy PJ. The relationship between clinical signs and dry eye
symptoms. Eye (Lond) 2011;25(4):502–10. https://doi.org/10.1038/
eye.2010.228.
[30] Guillon M, Styles E, Guillon JP, Maïssa C, Guillon M. Preocular tear lm
characteristics of nonwearers and soft contact lens wearers. Optom Vis Sci 1997;74
(5):273–9.
[31] Versura P, Campos EC. TearLab® Osmolarity System for diagnosing dry eye. Expert
Rev Mol Diagn 2013;13(2):119–29. https://doi.org/10.1586/erm.12.142.
S. Vidas Pauk et al.