Download PDF - touchOPHTHALMOLOGY

Lemp_editing_A4_2011 09/08/2012 17:48 Page 157
Anterior Segment Ocular Surface
Biomarkers In Dry Eye Disease
Michael A Lemp,1 Benjamin D Sullivan2 and Leslie A Crews2
1. Chief Medical Officer, TearLab Corporation, California and Clinical Professor of Ophthalmology, Department of Ophthalmology,
Georgetown University and Department of Ophthalmology, George Washington University, Washington DC;
2. Chief Scientific Officer, TearLab Corporation, California; 3. Consultant, TearLab Corporation, California, US
Abstract
Dry eye disease is a multifactorial chronic disorder of the ocular surface that affects up to 100 million people worldwide. During the
pathogenesis of dry eye, impaired function of the lacrimal and meibomian glands results in hyposecretion of aqueous tear fluid, coupled
with increased evaporation and instability of the tear film, which becomes increasingly hyperosmolar in character at more severe stages
of disease. One critical issue in the field is that the commonly used clinical signs and symptoms for the diagnosis or classification of dry
eye often do not correlate with one another. This underscores the need to systematically evaluate current tests and highlights the
importance of developing new markers of disease progression for use as endpoints in clinical trials of diagnostic devices and potential
therapeutics. In this review, we examine the utility and limitations of commonly used signs and symptoms of dry eye disease and
comment on newer biomarkers and analytical devices that show promise for future diagnostic and therapeutic use.
Keywords
Dry eye disease, biomarkers, tear film, hyperosmolarity, symptoms, objective, subjective
Disclosure: This study was supported by TearLab Corporation (San Diego, CA). The authors wish to indicate the following financial interests in TearLab Corporation:
Consultant (MAL, LAC), Equity (MAL, BDS), Patents (BDS), Employee (BDS).
Received: 3 March 2012 Accepted: 16 March 2012 Citation: European Ophthalmic Review, 2012;6(3):157–63 DOI: 10.17925/EOR.2012.06.03.157
Correspondence: Michael A Lemp, TearLab Corporation, 7360 Carroll Road Suite 200, San Diego CA, 92121, US. E: [email protected]
Dry eye disease (DED) is thought to affect upwards of 30–40 million
people within the US.1 Officially, the definition and classification of
dry eye was updated in 2007 during the Tear Film and Ocular Surface
(TFOS) Dry Eye WorkShop (DEWS).2
‘Dry eye is a multifactorial disease of the tears and ocular surface that
results in symptoms of discomfort, visual disturbance and tear film
instability with potential damage to the ocular surface. It is accompanied
by increased osmolarity of the tear film and inflammation of the
ocular surface.’
Whether initiated by ageing, androgen deficiency, contact lens wear,
refractive surgery or an autoimmune disease,1,3 dysfunctions of the
lacrimal and meibomian glands result in hyposecretion and increased
evaporation of tear fluid, which promote instability of the tear film.2,4
This leads to significant fluctuations in vision, loss of lubrication,
inflammation, an increase in wear (epitheliopathy) and varying levels
of neuropathy and corneal sensitisation. Thus, while diagnosis of
DED may be thought to be straightforward in more severe examples
of disease, determining the severity of the disease – especially at early
stages, requires a deeper understanding of the potential aetiologies and
clinical presentation of the disease to reconcile both the statistical
and subjective aspects of severity.
Biomarkers – surrogates or substitutes for variables of clinical
disease – are often used in the assessment of disease states,
particularly in clinical trials of therapeutics and diagnostic devices.5
The terms (bio)markers, surrogates, endpoints, outcomes and others
157
have been used to describe a metric that is either objective or
subjective, which accurately reflects the characteristics of disease. In
medicine, common examples of this are the use of serum cholesterol
or blood pressure for cardiovascular disease,6 or in ophthalmology
intraocular pressure for glaucoma.7 A surrogate outcome has been
defined as a “laboratory measurement or a physical sign used as a
substitute for a clinically meaningful endpoint that measures directly
how a patient feels, functions or survives”.8 The value of such a metric
lies in how accurately it can capture important aspects of the disease.
Such measures are typically used clinically in diagnosis, assessment
of disease severity and response to therapy.
A hallmark of DED is the highly variable symptom profile at different
stages of dry eye, with signs frequently conflicting with the objective
markers used to quantify ocular surface status.2,9 In addition, although
most patients with DED complain of symptoms of irritation or visual
disturbance, not all patients experience overt symptoms. In a recent
study, approximately 30 % of subjects with clear objective evidence of
DED based on a composite severity scale derived from seven commonly
used clinical tests for dry eye10 were asymptomatic according to their
Ocular Surface Disease Index (OSDI) questionnaire scores.11 As a
consequence, symptoms alone are insufficient to facilitate proper
management of the disease. Rather, it is necessary for the clinician to
take a global view of the clinical signs and symptoms without relying
too heavily on any single measure of discomfort.
Over the last decade, concepts concerning DED have shifted and
are continuing to be developed and refined as understanding of the
© TOUCH BRIEFINGS 2012
Lemp_editing_A4_2011 09/08/2012 17:49 Page 158
Biomarkers In Dry Eye Disease
mechanisms of the pathogenesis of the disease increases. The DEWS
report advanced the concept that there are various risk factors for
the development of the disease.2 These risk/initiating factors are
summarised in Figure 1. Regardless of the initiating factor underlying
an individual patient’s development and progression of disease, the
final common clinical expression of dry eye is a dysfunctional tear film
that is unstable and hyperosmolar in nature.12–15 These two defining
characteristics of the disease contribute to inflammation of the ocular
surface, ultimately leading to evidence of cellular damage.16 Several
pro-inflammatory cascades are activated by tear hyperosmolarity
and there is an increase in apoptotic cell death in the ocular surface
of DED patients.17–19 Reflex stimulation of the lacrimal gland is a
compensatory mechanism designed to lower tear osmolarity and can
be operative transiently and asymmetrically between eyes. However,
chronic reflex tearing can lead to an overexpression of autoantigens
in the lacrimal glands and neurogenic inflammation within the gland,
ultimately compromising aqueous tear production.20 These factors
cultivate a vicious cycle of mutually re-enforcing events contributing
to increased inflammation, tear film hyperosmolarity and instability,
and exacerbation of disease severity. As noted in the DEWS Report,
increased tear osmolarity is thought to be the central pathogenic
mechanism leading to damage of the ocular surface in DED.2
There are two main mechanistic subtypes of DED: aqueous tear
deficient dry eye (ADDE), in which there is a decrease in lacrimal gland
secretion that produces aqueous tear fluid and evaporative dry
eye (EDE), which is most commonly caused by meibomian gland
dysfunction (MGD). This latter subtype has been documented in great
detail in the recently published Meibomian Gland Dysfunction
Workshop report.21 The more common of the two subtypes is
MGD, but as the disease progresses and increases in severity,
compensatory mechanisms (reflex tearing, increased blinking and
thickening of the lipid layer) tend to involve the other subtype (ADDE).
This ultimately leads to the manifestation of a mixed or hybrid type of
DED, the principal type seen in moderate to severe cases.22
Diagnosis and management of DED has been a source of frustration to
clinicians for a number of reasons; one of which is a lack of correlation
between signs and symptoms. 11,23–25 While this is undoubtedly
attributable, in part, to psychological factors and personality types,
recent evidence of sensory changes on the ocular surface have
pointed to a hyperalgesia observed in early disease and a decrease in
corneal sensation in more severe disease.26 Additionally, the most
commonly used tests in clinical practice frequently do not give clear
information as to the status of dry eye, particularly in early or mild
stages of disease.9 A common clinical presentation is that of a patient
with significant symptoms of irritation without obvious evidence of
disease as measured by current commonly used clinical tests.
This lack of correlation between signs and symptoms has led to
problems in several areas such as the accurate diagnosis of early or mild
disease and a conspicuous lack of a single metric to assess disease
severity and response to treatment. This has been made apparent by
the difficulties in identifying a single test as a primary endpoint for
clinical trials. This condition is necessary to meet the US Food and Drug
Administration’s requirement for an improvement in both a sign and
symptom for clearance of new therapeutic agents for dry eye. This is a
very high standard that presents a barrier in the process of gaining
approval of new therapeutics in dry eye and is a major reason that there
have been no new therapeutic drugs approved in the US since 2003,
EUROPEAN OPHTHALMIC REVIEW
Figure 1: Schematic Diagram Depicting the
Multifactorial Origin of Dry Eye Disease
Ageing androgen
deficiency
Systemic drug effects
Inflammatory cytokines
SSDE; NSDE
Meibomian gland
High evaporation
rate
Secretory
Downregulation
Reduced lacrimal
flow
Lacrimal gland
Neurosecretory
block
Environment
Compromised lipid layer
Meibomian gland disease
Blepharitis
Lid flora
Lipases, esterases, detergents
Neurogenic
inflammation
Xerophthalmia
Ocular allergy
Preservatives
Contact lens wear
Tear hyperosmolarity
Tear film
instability
MAPK+, NFkB+
IL-1+, TNF-a+
MMPs
Cornea
Reflex block
Refractive surgery
Contact lens wear
Topical anaesthesia
Nerve injury
Goblet cell
Glycocalyx mucin loss
Epithelial damage
Apoptosis
Shear stress
Systemic disorders (e.g. Sjögren’s syndrome dry eye (SSDE) and non-Sjögren’s dry eye
(NSDE) and other conditions can result in dysfunction of one of the three main components
of the ocular surface unit – the lacrimal gland, meibomian gland and cornea. Impairment of
any one of these components perpetuates the dry eye condition via inflammatory,
hyperosmolar and stress conditions that damage the other components of the ocular
surface, driving disease progression.
despite many Phase II and III trials. Since the lack of correlation between
conventional tests and symptoms is well-known,11,23,24,27 there is
increasing interest within both the FDA and pharmaceutical companies
in the development of biomarkers of DED.
New research has begun to elucidate the factors contributing to
this problem and may lead the way to improved diagnosis, severity
assessment and disease management for the clinician and better
endpoints for more effective drug development in clinical trials.
The central characteristics of DED, namely tear instability and
hyperosmolarity, may well hold the keys to unlock this mystery.
Objective and Subjective Clinical Tests
An example of an ideal biomarker is the measurement of human
chorionic gonadotrophin to indicate the presence of an implanted
embryo during pregnancy. This protein is a small, highly charged
molecule that is essentially absent in the circulation of non-pregnant
females and can rise in concentration by almost four orders of
magnitude over a sustained period of time. It is as close to a binary
indicator as biologically possible, making it very clear to evaluate
whether the subject is pregnant or not. Currently there is no single
biomarker of such quality for DED. Most importantly, unlike binary
conditions, such as pregnancy or presence of an infection, dry eye
exists along a continuum of severity.10 Patients transition from
intermittent, mild disease to chronic severe disease, with subtypes
defined by the involvement of various pathological mechanisms to
differing degrees at each stage. Further, while the traditional markers of
corneal staining and Schirmer test may be sufficient for diagnosing
severe subjects, they have been shown to be inadequate at quantifying
the impact of therapeutic intervention,9 as evidenced by the numerous
failed clinical trials that used these markers as primary endpoints.28
New data also suggest that punctate staining may have little to do with
apparent damage to the cornea and more to do with relative levels of
mucin expression.29 Therefore, both researchers and clinicians must be
158
Lemp_editing_A4_2011 09/08/2012 17:49 Page 159
Anterior Segment Ocular Surface
Table 1: Various Clinical Biomarkers of Dry Eye
Disease Severity
Clinical Test of Severity
space over time, and other factors. Recent developments have
incorporated the use of optical interference or thermography34–36 to
estimate evaporation in realtime. These more recent data suggest that
the dry eye features a significantly increased evaporation rate compared
with normal subjects.
Units
Parameter Reported
Evaporimetry
g/cm2/sec
Evaporation rate
Schirmer Test
mm
Tear production
Schirmer Test
Optical scattering
OSI
Visual stability
TBUT
seconds
Tear stability
Interferometry
μm
Lipid layer thickness
Blink Rate
blinks/minute
Tear stability (indirect)
Tear Osmolarity
mOsm/L
Tear concentration
MMP-9
μg/mL
Inflammation
Cytokine profiles (IL-1, etc.)
pg/mL
Inflammation
Lactoferrin/Lysozyme
mg/mL
Tear production
Corneal staining
grade 0–4
Damage (mucin
Conjunctival staining
grade 0–4
MGD grading
grade 0–28
This is a test that aims to measure aqueous tear production, wherein
a strip of filter paper is inserted over the inferior lid margin and the
length of filter paper that is wetted is measured in millimetres after
five minutes. Frequently a topical anaesthetic is instilled prior to
measurement to reduce the irritating effects of the strip on the
conjunctiva, which can trigger reflex tearing. It has been thought that
this might represent ‘basal’ tearing, however, it has been shown that
while tearing is lessened by decreasing some of the sensory input,
stimuli from the lid margin and lashes are unaffected. Therefore, most
clinicians perform this test without anaesthetic. There are several
variations on this methodology such as the phenol red test, in which
a colour-impregnated thread is substituted for the strip. One limitation
of this test is that it is a measure only of aqueous tear production
and does not capture metrics pertaining to EDE. Additionally, it is
susceptible to wide intra-subject, day-to-day and visit-to-visit
variation.37,38 More consistent readings are attainable with increasing
severity of disease, as the test scoring has a lower boundary limited
by a measurement of zero mm. A recent study has demonstrated that
in mild/moderate DED, Schirmer test results lack sensitivity and are
essentially a ‘random number generator’.9 This test tends to improve
Functional Tests
Molecular Tests
and stability
Subjective Tests
deficiency)
Tissue damage
Gland obstruction
and inflammation
Symptoms (OSDI)
score 0–100
Discomfort
TBUT = tear breakup time; MGD = meibomian gland dysfunction; OSDI = ocular surface
disease index.
able to assimilate a large number of data points, many of which are only
indirectly associated with the actual disease state, in order to come to
a conclusion about the overall disease severity.
Tests of DED severity can be broadly categorised as either objective or
subjective, with the objective markers broken down as either functional
tests or those that report the molecular constituents of the tear film
(Table 1). In an ideal world, the quality of a biomarker would be neatly
characterised by a broad dynamic range, high sensitivity to therapeutic
intervention, linearity (i.e. probability distribution at different severities),
low biological variance over time, and minimal interoperator variance
during readout. However, very little referent data exist for the majority
of biomarkers associated with disease severity.
Functional Tests
Evaporimetry
Given the increased understanding of dry eye as a consequence of
MGD, tear evaporation rate should be directly related to disease
severity. However, technical challenges to actually performing this
test in a clinical setting, coupled with very large variation in referent
values, have precluded its practical implementation. Most of the devices
developed to measure tear evaporative rates are of custom design and
incorporate a sealed goggle apparatus with inflow and outflow channels.
The differences in water content of the inflow air with the outflow are
measured. Of the seminal studies in this regard, some investigators
found that evaporation rate increased in DED, while others found that
the evaporation rate decreased.30–33 Specifically, measurement of water
evaporation is not the same as the evaporation rate from the tear
film – more water may evaporate from a normal subject in a given time
than a dry eye subject because in a normal subject there is more fluid
available to evaporate, while the overall rate of evaporation is lower than
in dry eye subjects.31,32 When measuring fluid in a closed chamber, to
understand the rate at which fluid is evaporating, one must compensate
for the ambient temperature, humidity, ocular physiology, interpalpebral
159
in accuracy in moderate-to-severe disease in which most subjects
have a mixed or hybrid form of the disease.22
Tear Breakup Time
Tear instability is a hallmark of the dry eye state. It affects vision
and contributes to the loss of homeostatic control of the lacrimal
functional unit, giving rise to the variability of many clinical tests used
to assess this disease. Tear instability is, therefore, of particular interest
and accurate measurement of this factor could provide information that
reflects a principal axis of the disease. The tear breakup test has been
used as a measure of tear instability for decades. Most commonly, it
is measured after the instillation of a defined volume of fluorescein,
followed by assessment of the rapidity with which the tear film begins
to break up after a blink. Values of less than 10 seconds have been
recommended as diagnostic thresholds, however, more recently the
use of smaller volumes (5 μl) of instilled 2 % fluorescein with a referent
value of 5 seconds has been recommended.39 Unfortunately this test
suffers from limitations such as poor reproducibility, low performance
accuracy and the fact that other events such as the presence of surface
irregularities will influence results.
Attempts to standardise the assessment of tear instability have
centred on the development of automated visual capture of serial
images using video keratography to eliminate the subjective aspects
of tear instability measurement. Several different algorithm-driven
programmes have been described.40 While more rapid inter-blink tear
breakup times (TBUTs) have clearly been demonstrated in patients
with DED, the range of normative values and relationships over the
spectrum of DED severity are yet to be elucidated.9
Interferometry
The use of interference ring patterns has been of interest for some
years in assessing dry eye. Various devices have been developed to
EUROPEAN OPHTHALMIC REVIEW
Lemp_editing_A4_2011 11/08/2012 09:31 Page 160
Biomarkers In Dry Eye Disease
measure the thickness of the lipid layer of the tear film, and
correlations between a thinned lipid layer and dry eye disease
have been described.41 Most of these devices depend on observer
judgment of colour patterns to estimate thickness of the lipid layer.
Recently an automated device has been introduced in which pattern
recognition is determined by software and correlations between a
thin lipid layer and patient symptomatology have been reported.42
However, no published studies are available as yet that establish
the reproducibility of these devices, correlations between these tests
and other measures of DED severity, or response to treatment. This
technology, in a more developed form, might prove to be useful in
the diagnosis of MGD.
Optical Quality Tests
Since the tear layer plays an important role in the optical quality of
the eye’s visual pathway, alterations in the tear film should be
relevant to the pathogenesis of DED. Variations in the optical
qualities of the tear film have been reported. These include changes
is visual acuity correlated to central corneal surface disruption and
degradation of the visual image during the interblink interval in
patients with DED. This latter characteristic – a consequence of an
unstable tear film – has been dubbed ‘functional visual acuity’, and
instruments aiming to capture the degradation of the visual image
over time, after a blink, are in development. Other visual parameters
affected by an unstable tear film include contrast sensitivity,
Hartmann–Shack wavefront sensing and double-pass optical
analysis.43–45 These measurements may contribute sensitive clinically
relevant markers in the coming years.
Tear Constituent Tests
Changes in the composition of the tear film have been identified in the
development of DED. Decreases in certain tear proteins, e.g. tear
lysozyme and tear lactoferrin, have long been known to occur and
have been utilised as diagnostic tests. Although these have been
in use in previous years, there are currently no widely utilised
commercial assay systems for these markers. It is not known why
these tests are no longer commercially available. Most probably this
is due to the technical challenges of performing these tests, the lack
of referent values, and a lack of widespread adoption. It should be
noted that as their putative relationship to DED is to the aqueous tear
deficient dry eye (ADDE), they do not serve well as a global marker
since ADDE is less common than evaporative dry eye. There is an
additional complication in interpreting these results in that reductions
in protein production may be offset by increases in tear osmolarity.
Since the publication of papers in the 1990s detailing the evidence
for inflammation as an important feature of DED, investigators have
searched for molecular markers present in the tears of DED patients.
Almost 500 individual proteins have been identified in human tears
and of these, a number of pro-inflammatory markers have been
found.17,46–49 These include interleukin-1 (IL-1), IL-6, IL-8, tumour
necrosis factor-alpha (TNF-α), transforming growth factor-beta
(TGF-β), matrix metalloproteinase-9 (MMP-9) and macrophage
inflammatory protein-1alpha (MIP-1α), among others.17,50,51
Both cytokines and chemokines have been reported to increase in
DED, particularly in moderate to severe forms of the disease such
as seen in patients with Sjögren’s syndrome. Minimal distinction
between normal subjects and patients with mild DED has been noted
using measurements of cytokines and chemokines.52 Several of these
EUROPEAN OPHTHALMIC REVIEW
molecular markers have been shown to increase in DED53,54 and to
respond to treatment with topically applied cyclosporine A.55 The
advent of microarray technology, particularly multiplex bead arrays
and highly-sensitive enzyme-linked immunosorbent assay (ELISA) tests
for proteins has spurred investigation aimed at identifying patterns
of protein elevation, although some obstacles remain in interpreting
results of microwell arrays that produce interfering effects that can
greatly reduce signal-to-noise and limit the ability to obtain meaningful
results.56 In spite of these potential concerns, a number of panels
of these proteins have shown promise as diagnostic biomarkers and
have been advanced as suitable markers for DED.51 Other promising
techniques have also recently been described for collecting and
measuring cytokines present in tear fluid, including extraction of
cytokines and matrix metalloproteinases (MMPs) from Schirmer strips
followed by Luminex analysis.57 Numerous cytokines and MMPs were
detected in the tear samples of healthy volunteers using this
methodology,57 however future studies will be required to determine if
tear quantity and quality can be similarly measured in DED subjects.
Other recent studies have presented normative values for the presence
of cytokines and chemokines in tears.58 As further investigations refine
the relationships between these patterns, their association with
severity of disease and response to treatment and the emergence of
simplified assays for clinical use, their utility may become clearer.
One particular pro-inflammatory marker that has received much
recent attention is MMP-9.17 This protein is widespread in the
body and is associated with tissue repair and re-modelling. MMP-9 is
present in very small amounts in normal tears and rises considerably
in DED. While not specific for DED,17,47 MMP-9 activation is associated
with tissue damage from inflammation, which is a prominent feature
of dry eye in moderate to severe cases. Several studies have
demonstrated the rise of MMP-9 in DED. In one study measuring
tear levels of MMP-9, the protein was elevated across four levels of
increasing disease severity.50 However, the actual expression levels
for grades 1 and 2 were significantly lower than in grades 3 and 4.
Of interest, the categorisation of subjects in the study was primarily
based on corneal staining. Since other studies have shown that less
than 50 % of patients with mild-to-moderate disease have corneal
staining (vide infra), it may be that MMP-9 correlates very well with
staining, but not necessarily DED. MMP-9 levels in the tears have been
shown to be elevated in ocular surface diseases other than DED.17 The
recent commercialisation of an office test for MMP-9 (Rapid Pathogen
Screening [RPS] InflammaDry Detector™) which is qualitative may
prove useful in identifying patients with a significant inflammatory
component and is likely to respond to anti-inflammatory therapy,
however, it should also be noted that RPS excludes the difficult to
diagnose mild-to-moderate subjects (with OSDI between 1–13) in
order to achieve the performance reported in its labelling. Given that
MMP-9 correlates so well with staining, it is not clear that the marker
gives any additional information than ocular surface staining alone.
In this era of advancing technology, broad studies at the proteomics
level are becoming more affordable and feasible to carry out in patient
populations. Thus, further research in these directions may identify
other protein candidates that represent clinically useful markers.
Quantitative Tests
Ocular Surface Staining
The definition of DED expressed in the DEWS report emphasises
damage to the ocular surface.12 For many years, damage has been
160
Lemp_editing_A4_2011 11/08/2012 09:38 Page 161
Anterior Segment Ocular Surface
assessed by the use of vital dyes which increase visibility of areas of
discontinuity on the corneal and conjunctival surfaces.59 Fluorescein
solutions are common in current clinical use to assess corneal
disruptions, and preceding this was rose bengal staining and more
recently lissamine green has been implemented to visualise these
areas of the conjunctiva.60 Ocular surface staining is used to diagnose
DED, assess its severity and as a clinical endpoint in clinical trials
for pharmaceutical effect.28
It should be noted, however, that precisely what damage is revealed
by staining methods and whether distinct clinical patterns of staining
reflect different changes to the ocular surface, remain undefined.
Traditionally, positive rose bengal staining was thought to be due to
a breakdown in the structure of the cell wall of conjunctival surface
cells, however, experimental studies revealed that discontinuities
in the mucin coating of the cells corresponded to the patterns of
staining observed.61 Fluorescein staining of the corneal surface is
thought to reflect a breakdown in the cell barrier to aqueous solutions,
but the import of differences in staining patterns (punctate versus
confluent) is unclear.
Several grading systems for assessing severity of ocular surface
staining have been developed. These include an older van Bijsterveld
system, the National Eye Institute/Industry Report system and the
Oxford system.62 The latter two are the most commonly used in
clinical trials. They differ in the subdivisions of the corneal and
conjunctival surfaces and the progression of severity (e.g. the Oxford
progression is geometric).
Recent evaluation of corneal and conjunctival staining in DED,
as measured using a composite index of severity over the entire
range of disease, demonstrated that in mild-to-moderate disease
less than 50 % of subjects demonstrated any corneal staining.10 A
slightly higher fraction of subjects displayed evidence of conjunctival
staining, however, much higher percentages would be expected
if staining were a precise indicator of the presence of disease.
This exemplifies the problems that can arise if one is using the
presence of ocular surface staining for the diagnosis of disease.
This demonstrates the danger in using a single one of the currently
used metrics for diagnosis. While staining provides an important
piece of information, it is not necessarily informative for overall
disease assessment. In summary, ocular surface staining remains a
useful data point for assessing DED, but alone can be misleading.
As with most of the tests, staining measurements become more
informative in more advanced stages of disease but lack specificity
in mild/moderate disease.
Meibomian Gland Dysfunction Grading
MGD is the most common form of EDE and the most common subtype
of DED. Its prevalence increases with more severe disease when it
is part of a mixed form of the disease. Recognising its presence and
assessing its severity is crucial not only for accurately diagnosing
disease and assessing severity but in formulation of a treatment plan,
which will differ from that for pure ADDE. There is general agreement
that the quality of the expressed secretion from the meibomian glands
is the most important component of this test. A commonly used
scale is that reported by Foulks and Bron.63 This area is discussed
in much greater detail in the recently published Report of the Tear
Film and Ocular Surface Society (TFOS) Workshop of Meibomian
Gland Dysfunction.64
161
Tear Osmolarity
Hyperosmolar tears are regarded as the central pathogenetic
mechanism of dry eye.2 Hyperosmolarity, regardless of symptomatic
status, has been shown to induce inflammation, including an
upregulation of IL-1β, TNF-α, MMP-9 and activation of c-Jun
N-terminal kinase (JNK), extracellular signal-regulated kinase (ERK)
and mitogen-activated protein kinase (MAPK) pathways. 15,18,19
Additionally, hyperosmolarity has been shown to activate caspase-3
and cytochrome c pathways that result in spontaneous epithelial
cell death. 65 Hyperosmolarity causes decreased intracellular
connection, loss of microplicae, cell membrane disruption and
increased desquamation.66 An increase in the conπcentration of
tears also has functional consequences, wherein a hyperosmolar
tear film has been directly associated with fluctuating vision.67 Most
importantly, data suggest that an increase in osmolarity directly
reduces the ability of the cornea and lid to maintain proper
boundary lubrication, thereby increasing friction during each blink
and likely manifesting in long term wear and lid epitheliopathies
evident in DED.68-70
In the recently published guideline, 2011 American Academy of
Ophthalmology Preferred Practice Patterns on Dry Eye Syndrome, it is
noted that “tear osmolarity has been shown to be a more sensitive
method of diagnosing and grading the severity of dry eye compared
to corneal and conjunctival staining, TBUT, Schirmer test, and
meibomian gland grading”.71 Tear osmolarity was also found to be a
superior correlate to a composite disease severity index as compared
with the other most common tests for DED.10 In addition, tear osmolarity
showed the lowest variability over time and the highest sensitivity
to therapeutic intervention as compared with corneal staining,
TBUT, Schirmer tests and meibomian gland grading.9 In effect, tear
osmolarity gives a direct insight to the health of the ocular surface
while predicting response to therapy before symptoms or the other
signs have resolved.9
Despite the extensive literature detailing the effects of hyperosmolarity
on the ocular surface, many clinicians have not had first hand
experience interpreting results, which can lead to confusion and
incorrect conclusions surrounding its clinical utility.72 For diagnosis,
the most important difference between osmolarity and other tests is
that the unstable tear film and its effects on compensatory mechanisms
require that both eyes be tested to make a proper determination. Once
that is completed, a positive determination is made if either of the eyes
are >308 mOsm/L or the difference between the eyes is ≥8 mOsm/L.
Since a normal homeostatic tear film is isotonic with plasma osmolarity
(290 mOsm/L),73 either the elevation of osmolarity or significant
differences between eyes are indicators of tear film instability
and an unhealthy ocular surface. For instance, a mildly symptomatic
subject may exhibit a two-eye reading of 304/295 mOsm/L
(OD/OS). While the two-eye maximum is clearly elevated from
290 mOsm/L it is below the 308 mOsm/L threshold. However, given
that the low analytical variance (±4 mOsm/L74) makes the number
statistically equivalent to 308 mOsm/L, coupled with presence of
symptoms and a ≥8 mOsm/L difference between the eyes, these
findings suggest a positive diagnosis.
Subjective Tests
Symptoms (Ocular Surface Disease Index)
It is frequently stated that DED is a symptomatic disease.3 This implies
that without assessing symptoms the diagnosis of DED is inaccurate.
EUROPEAN OPHTHALMIC REVIEW
Lemp_editing_A4_2011 11/08/2012 09:48 Page 162
Biomarkers In Dry Eye Disease
While most patients with DED do complain of irritative symptoms,
as noted above, not all do. A recent analysis of 299 subjects both
with and without DED revealed that only about 70 % of patients with
evidence of DED by other clinical tests complained of symptoms
as measured by the OSDI scoring. Symptoms alone, although they
have been thought to be more representative of the disease as
a whole than any individual signs,3 do not necessarily lead to an
accurate diagnosis.
Perhaps the most vexing aspect of the use of symptoms in DED is their
lack of correlation with signs.11,24 A newer way of looking at disease
may lead to a better understanding of their relationship.
Relationship Between Individual Tests of
Dry Eye Disease
Data presented at the 2012 meeting of the Association for Research
in Vision and Ophthalmology meeting revealed that across a general
patient population, there was no correlation between subjective and
objective tests.75 Each of the common tests was compared to each of
the other signs and symptoms and the average r2 for each test was
consistently low, as follows: TBUT (0.12), Schirmer test (0.08), corneal
(0.15) and conjunctival staining (0.16), meibomian grading (0.10),
osmolarity (0.04) and OSDI (0.11). As such, each type of measurement
was independent of the other tests, such that there was no
predictability based on the presence of any single result. For instance,
moderate staining levels did not provide any information as to
whether the subject should have a low breakup time or low tear
production. This result carries significant implications for clinical
practice, but more importantly, for clinical trials. It was not
determined whether the lack of correlation was statistically driven (i.e.
the variation in each marker over time was great enough to preclude
correlation) or whether the information given by each marker was
independent because each reveals evidence about distinctly different
patho-physiological mechanisms of the disease.
Composite Index of Disease Severity
Given the statistical complexity inherent to the signs and symptoms of
DED, there has been little guidance as to how to synthesise multiple
data points into a single coherent index of overall disease severity.
The traditional Delphi approach reprinted in the DEWS report outlines
a 1–4 grading scheme, with one being mild or episodic dry eye and
four being severe constant and disabling disease.2,76 Unfortunately,
the Delphi schema does not supply a quantitative method for dealing
with conflicts within individual patients, such as when a patient
exhibits low breakup time but no apparent staining or symptoms.
Conflicting data from a battery of clinical signs is more the norm than
the exception. When coupled with a lack of consensus on any single
gold standard test for dry eye, there is a need to quantitatively and
objectively determine disease severity for clinical trials.
In creating a composite scale, clinical signs are combined by converting
each data point into a common severity basis such as a scale of 0–4, or
0–1 scheme, adding them together to generate a final severity score,
then parsing the total into normal, mild/moderate and severe
categories. A similar but attenuated method was used in a seminal OSDI
paper, as a means to check the performance of the symptom
questionnaire.77 To create their index, the authors mixed Schirmer test,
lissamine green staining and a subjective facial expression scale, with
cutoffs for scoring determined by a literature review (e.g. Schirmer
normal values were >10 mm, mild/moderate values between
EUROPEAN OPHTHALMIC REVIEW
6 to 10 mm and severe values <5 mm in length). For each sign,
normal results were given one point, mild/moderate two points and
severe three points. The summed index was graded as ≤3 normal,
4–7 mild/moderate and then 8–9 as severe. While straightforward, there
is a series of problems with this approach. Most importantly, when
comparing signs against the total score, there is an inherent risk
of ordinate bias against the composite abscissa. Little data is available
about the facial expression scale used in the paper, but unsurprisingly,
two measures of symptoms (OSDI and facial expressions) were shown
to be very well correlated (r=0.669). Especially when a composite
is formed from a small number of variables, these biases can be
significant. As a result, the OSDI showed a sensitivity of 60 % against the
physician rating, but jumped up to 80 % sensitivity against the composite
score with the facial expression as part of the abscissa.
A second limitation of this approach is that to facilitate simplicity,
the categories are discrete. A person with a 6/9 in lissamine green
staining is given the same input to the total severity as someone
with a 2/9 in staining or a 10 mm Schirmer result. This resulting
lack of resolution typically creates spurious false negatives or false
positives in sensitivity and specificity analysis, which is unacceptable in
regulatory environments.
A more recent attempt to create an objective, unbiased composite
index outlined a continuous method for determining DED severity.10
In this approach, the standard procedure of relying upon clinical
expertise to assign thresholds to each of seven quantitative measures
was followed by fitting either a linear or exponential function to the
clinical breakpoints, depending on strength of fit. These functions
were inverted and allowed raw clinical data to be mapped on a common
0–1 scale without discretisation (converting continuous differential
equations into discrete ones). The next step, adding the individual
dimensions together to form a final severity score, was accomplished
by first rotating each dimension (equivalent to weighting each data
point) by the amount of independent information the measure
contributed to the composite. This independent components analysis,
unlike the direct linear addition approach used in previous studies,77
attempted to equalise correlation risk against the composite abscissa
(point on the x-axis). Control data sets (random data with no clinical
meaning) showed that the expected correlation was on the order
r2=0.1, which indicated the successful removal of bias for each
clinical measure, that is, a sign that may provide the same and thus
redundant, information as another sign. This was in part due to the
fact that there were many more inputs to the composite scale than in
earlier attempts, so that any one measurement could not dominate
the x-axis. Finally, the bases were geometrically combined to form
a single normalised composite index from 0–1, with 0 having no
evidence of disease and one the greatest. The resulting continuous
index revealed the gross failure of threshold-based classification
schemes for clinical trials (e.g. a patient must have N out of M tests
above a threshold to qualify as a dry eye patient in the study), wherein
basic thresholds failed to properly classify 63 % of the mild/moderate
subjects recruited for analysis.
The results of measuring the relationship between the individual tests
and the composite severity scale revealed that only one test – tear
osmolarity – demonstrated a linear relationship to the composite
severity score. This should not be too surprising since tear osmolarity
is viewed as the central pathogenic mechanism resulting in damage
to the ocular surface in DED. Thus, although it has been shown that
osmolarity may be the ideal global biomarker correlating to disease
162
Lemp_editing_A4_2011 11/08/2012 09:48 Page 163
Anterior Segment Ocular Surface
severity – whether the dry eye is evaporative, aqueous deficient, or a
hybrid – it does not define the aetiology of the disease and the other
biomarkers remain useful in that role.
Conclusion
The authors have reviewed the major candidates for biomarkers which
will accurately reflect disease effects over the entire range of DED
severity. There are a number of candidates that demonstrate either
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
Tomlinson A, Epidemiology of dry eye disease. In: Asbell
PA, Lemp MA, Dry Eye Disease: The Clinician's Guide to Diagnosis
and Treatment, New York, NY: Thieme Medical Publishers,
Inc., 2006;1–15.
International Dry Eye WorkShop, Ocul Surf, 2007;5:65–204.
The epidemiology of dry eye disease: report of the
Epidemiology Subcommittee of the International Dry Eye
WorkShop, Ocul Surf, 2007;5:93–107.
Bron AJ, Diagnosis of dry eye, Surv Ophthalmol,
2001;45 (Suppl. 2):S221–6.
Ellenberg S, Hamilton JM, Surrogate endpoints in clinical
trials: cancer, Stat Med, 1989;8:405–13.
Tatasciore A, Renda G, Zimarino M, et al., Awake systolic
blood pressure variability correlates with target-organ damage
in hypertensive subjects, Hypertension, 2007;50:325–32.
Fogagnolo P, Sangermani C, Oddone F, et al., Long-term
perimetric fluctuation in patients with different stages of
glaucoma, Br J Ophthalmol, 2011;95:189–93.
Lassere MN, Johnson KR, Boers M, et al., Definitions and
validation criteria for biomarkers and surrogate endpoints:
development and testing of a quantitative hierarchical
levels of evidence schema, J Rheumatol, 2007;34:607–15.
Sullivan BD, Crews LA, Sönmez B, et al., Clinical utility of
objective tests for dry eye disease: variability over time and
implications for clinical trials and disease management,
Cornea, 2012; [Epub ahead of print].
Sullivan BD, Whitmer D, Nichols KK, et al., An objective
approach to dry eye disease severity, Invest Ophthalmol Vis Sci,
2010;51:6125–30.
Sullivan BD, Crews LA, Messmer EM, et al., Evaluation of
the relationship among common clinical tests for the
management of dry eye disease, Arch Ophthalmol,
2011 [Submitted].
The definition and classification of dry eye disease: report
of the Definition and Classification Subcommittee of the
International Dry Eye WorkShop, Ocul Surf, 2007;5:75–92.
Gilbard JP, Rossi SR, Changes in tear ion concentrations in
dry-eye disorders, Adv Exp Med Biol, 1994;350:529–33.
Lemp MA, Bron AJ, Baudouin C, et al., Tear osmolarity in
the diagnosis and management of dry eye disease,
Am J Ophthalmol, 2011;151:792–8 e1.
Liu H, Begley C, Chen M, et al., A link between tear
instability and hyperosmolarity in dry eye,
Invest Ophthalmol Vis Sci, 2009;50:3671–9.
Johnson ME, Murphy PJ, Changes in the tear film and ocular
surface from dry eye syndrome, Prog Retin Eye Res,
2004;23:449–74.
Acera A, Rocha G, Vecino E, et al., Inflammatory markers in
the tears of patients with ocular surface disease,
Ophthalmic Res, 2008;40:315–21.
Luo L, Li DQ, Corrales RM, Pflugfelder SC, Hyperosmolar
saline is a proinflammatory stress on the mouse ocular
surface, Eye Contact Lens, 2005;31:186–93.
Luo L, Li DQ, Doshi A, et al., Experimental dry eye
stimulates production of inflammatory cytokines and
MMP-9 and activates MAPK signaling pathways on the
ocular surface, Invest Ophthalmol Vis Sci, 2004;45:4293–301.
Sullivan DA, Immunology of the lacrimal gland and tear film,
Dev Ophthalmol, 1999;30:39–53.
Nichols KK, Foulks GN, Bron AJ, et al., The international
workshop on meibomian gland dysfunction: executive
summary, Invest Ophthalmol Vis Sci, 2011;52:1922–9.
Lemp MA, Crews LA, Bron AJ, et al., Distribution of
aqueous-deficient and evaporative dry eye in a clinic-based
patient cohort: a retrospective study, Cornea,
2012;31(5):472–8.
Kallarackal GU, Ansari EA, Amos N, et al., A comparative
study to assess the clinical use of Fluorescein Meniscus
Time (FMT) with Tear Break up Time (TBUT) and Schirmer’s
tests (ST) in the diagnosis of dry eyes, Eye (Lond),
2002;16:594–600.
Moore JE, Graham JE, Goodall EA, et al., Concordance
between common dry eye diagnostic tests,
Br J Ophthalmol, 2009;93:66–72.
Fuentes-Páez G, Herreras JM, Cordero Y, et al., [Lack of
concordance between dry eye syndrome questionnaires
and diagnostic tests], Arch Soc Esp Oftalmol, 2011;86:3–7.
163
good utility for subtypes of DED (i.e. ADDE or EDE) and others are still
in early stages of testing and might prove to be of value with further
development. Currently, however, only tear osmolarity has been found
to serve as a clinically useful biomarker over a wide range of disease
severity. The increasing use of biomarkers may enable clinicians
to more effectively recognise disease in patients, particularly at
mild/moderate or early stages and facilitate the evaluation of new
treatment modalities for improved disease management. n
26. Tuisku IS, Konttinen YT, Konttinen LM, Tervo TM, Alterations
in corneal sensitivity and nerve morphology in patients with
primary Sjogren's syndrome, Exp Eye Res, 2008;86:879–85.
27. Nichols KK, Nichols JJ, Mitchell GL, The lack of association
between signs and symptoms in patients with dry eye
disease, Cornea, 2004;23:762–70.
28. Foulks GN, Challenges and pitfalls in clinical trials of
treatments for dry eye, Ocul Surf, 2003;1:20–30.
29. Gipson IK, Spurr-Michaud SJ, Senchyna M, et al.,
Comparison of mucin levels at the ocular surface of
postmenopausal women with and without a history
of dry eye, Cornea, 2011;30:1346–52.
30. Mathers WD, Daley TE, Tear flow and evaporation in
patients with and without dry eye, Ophthalmology
1996;103:664–9.
31. Tsubota K, Yamada M, Tear evaporation from the ocular
surface, Invest Ophthalmol Vis Sci, 1992;33:2942–50.
32. Yamada M, Tsubota K, [Measurement of tear evaporation
from ocular surface], Nihon Ganka Gakkai Zasshi,
1990;94:1061–70.
33. Rolando M, Refojo MF, Kenyon KR, Increased tear
evaporation in eyes with keratoconjunctivitis sicca,
Arch Ophthalmol, 1983;101:557–8.
34. Craig JP, Tomlinson A, Importance of the lipid layer in
human tear film stability and evaporation, Optom Vis Sci,
1997;74:8–13.
35. Goto E, Ishida R, Kaido M, et al., Optical aberrations and
visual disturbances associated with dry eye, Ocul Surf,
2006;4:207–13.
36. Craig JP, Singh I, Tomlinson A, et al., The role of tear
physiology in ocular surface temperature, Eye (Lond),
2000;14 (Pt 4):635–41.
37. Walker PM, Lane KJ, Ousler GW 3rd, et al., Diurnal variation
of visual function and the signs and symptoms of dry eye,
Cornea, 2010;29:607–12.
38. Cho P, Yap M, Schirmer test. II. A clinical study of its
repeatability, Optom Vis Sci, 1993;70:157–9.
39. Johnson ME, Murphy PJ, The Effect of instilled fluorescein
solution volume on the values and repeatability of TBUT
measurements, Cornea, 2005;24:811–7.
40. Alonso-Caneiro D, Turuwhenua J, Iskander DR, Collins MJ,
Diagnosing dry eye with dynamic-area high-speed
videokeratoscopy, J Biomed Opt, 2011;16:076012.
41. Yokoi N, Komuro A, Non-invasive methods of assessing the
tear film, Exp Eye Res, 2004;78:399–407.
42. Blackie CA, Solomon JD, Scaffidi RC, et al., The relationship
between dry eye symptoms and lipid layer thickness,
Cornea, 2009;28:789–94.
43. Wang Y, Xu J, Sun X, et al., Dynamic wavefront aberrations
and visual acuity in normal and dry eyes, Clin Exp Optom,
2009;92:267–73.
44. Montés-Micó R, Cerviño A, Ferrer-Blasco T, et al., The tear
film and the optical quality of the eye, Ocul Surf, 2010;8:185–92.
45. Liu H, Thibos L, Begley CG, Bradley A, Measurement of the
time course of optical quality and visual deterioration during
tear break-up, Invest Ophthalmol Vis Sci, 2010;51:3318–26.
46. Pflugfelder SC, de Paiva CS, Tong L, et al., Stress-activated
protein kinase signaling pathways in dry eye and ocular
surface disease, Ocul Surf, 2005;3:S154–7.
47. Smith VA, Rishmawi H, Hussein H, Easty DL, Tear film MMP
accumulation and corneal disease, Br J Ophthalmol,
2001;85:147–53.
48. Tong L, Zhou L, Beuerman RW, et al., Association of tear
proteins with Meibomian gland disease and dry eye
symptoms, Br J Ophthalmol, 2011;95(6):848–52.
49. Tsai PS, Evans JE, Green KM, et al., Proteomic analysis of
human meibomian gland secretions, Br J Ophthalmol,
2006;90:372–7.
50. Chotikavanich S, de Paiva CS, Li de Q, et al., Production
and activity of matrix metalloproteinase-9 on the ocular
surface increase in dysfunctional tear syndrome,
Invest Ophthalmol Vis Sci, 2009;50:3203–9.
51. Lam H, Bleiden L, de Paiva CS, et al., Tear cytokine profiles
in dysfunctional tear syndrome, Am J Ophthalmol,
2009;147:198–205 e1.
52. Yoon KC, Park CS, You IC, et al., Expression of CXCL9, -10,
-11, and CXCR3 in the tear film and ocular surface of
patients with dry eye syndrome, Invest Ophthalmol Vis Sci,
2010;51:643–50.
53. Massingale ML, Li X, Vallabhajosyula M, et al., Analysis of
inflammatory cytokines in the tears of dry eye patients,
Cornea, 2009;28:1023–7.
54. Enriquez-de-Salamanca A, Castellanos E, Stern ME, et al.,
Tear cytokine and chemokine analysis and clinical
correlations in evaporative-type dry eye disease,
Mol Vis 2010;16:862–73.
55. Narayanan S, Miller WL, McDermott AM, Conjunctival
cytokine expression in symptomatic moderate dry eye
subjects, Invest Ophthalmol Vis Sci, 2006;47:2445–50.
56. Li S, Sack R, Vijmasi T, et al., Antibody protein array analysis
of the tear film cytokines, Optom Vis Sci, 2008;85:653–60.
57. VanDerMeid KR, Su SP, Krenzer KL, et al., A method to
extract cytokines and matrix metalloproteinases from
Schirmer strips and analyze using Luminex, Mol Vis
2011;17:1056–63.
58. Carreno E, Enriquez-de-Salamanca A, Teson M, et al.,
Cytokine and chemokine levels in tears from healthy
subjects, Acta Ophthalmol, 2010;88:e250–8.
59. Eliason JA, Maurice DM, Staining of the conjunctiva and
conjunctival tear film, Br J Ophthalmol, 1990;74:519–22.
60. Korb DR, Survey of preferred tests for diagnosis of the tear
film and dry eye, Cornea, 2000;19:483–6.
61. Argüeso P, Tisdale A, Spurr-Michaud S, et al., Mucin
characteristics of human corneal-limbal epithelial cells that
exclude the rose bengal anionic dye, Invest Ophthalmol Vis Sci,
2006;47:113–9.
62. van Bijsterveld OP, Diagnostic tests in the Sicca syndrome,
Arch Ophthalmol, 1969;82:10–4.
63. Foulks GN, Bron AJ, Meibomian gland dysfunction: a clinical
scheme for description, diagnosis, classification, and
grading, Ocul Surf, 2003;1:107–26.
64. The Report of the TFOS Workshop on Meibomian Gland
Dysfunction, Invest Ophthalmol Vis Sci, 2011;52(4):1917–2085.
65. Luo L, Li DQ, Pflugfelder SC, Hyperosmolarity-induced
apoptosis in human corneal epithelial cells is mediated by
cytochrome c and MAPK pathways, Cornea, 2007;26:452–60.
66. Gilbard JP, Carter JB, Sang DN, et al., Morphologic effect
of hyperosmolarity on rabbit corneal epithelium,
Ophthalmology, 1984;91:1205–12.
67. Pisella PJ, Habay T, Nochez Y, Evaluation of tear film quality
with a double-pass scattering index, Presented at: 6th
International Conference on the Tear Film and Ocular
Surface, Florence, Italy, 22–25 September, 2010.
68. Arita R, Itoh K, Inoue K, et al., Contact lens wear is
associated with decrease of meibomian glands,
Ophthalmology, 2009;116:379–84.
69. Korb DR, Herman JP, Blackie CA, et al., Prevalence of lid
wiper epitheliopathy in subjects with dry eye signs and
symptoms, Cornea, 2010;29:377–83.
70. Samsom ML, Sullivan BD, Schmidt TA, Effect of
hyperosmolarity on PRG4’s ocular surface boundary
lubricating ability at a human cornea-eyelid biointerface,
Presented at: Association for Research in Vision and
Ophthalmology (ARVO) meeting, Florida, US, 6 May 2012.
71. Preferred Practice Pattern Guidelines: Dry Eye Syndrome –
Limited Revision, American Academy of Ophthalmology,
San Francisco, CA, US, 2011.
72. Khanal S, Millar TJ, Barriers to clinical uptake of tear
osmolarity measurements, Br J Ophthalmol, 2012;96(3):341–4.
73. Fortes MB, Diment BC, Di Felice U, et al., Tear fluid
osmolarity as a potential marker of hydration status,
Med Sci Sports Exerc, 2011;43(8):1590–7.
74. TearLab, OcuSense, Inc., TearLab Osmolarity System, 510(k).
Available at: www.accessdata.fda.gov/cdrh_docs/pdf8/
K083184.pdf (accessed 10 August 2012).
75. Sullivan BD, Crews LA, Messmer EM, et al., Correlation
between commonly used clinical tests in the management
of dry eye disease, IOVS, 2012;53:ARVO E-Abstract 550.
76. Behrens A, Doyle JJ, Stern L, et al., Dysfunctional tear
syndrome: a Delphi approach to treatment
recommendations, Cornea, 2006;25:900–7.
77. Schiffman RM, Christianson MD, Jacobsen G, et al.,
Reliability and validity of the Ocular Surface Disease
Index, Arch Ophthalmol, 2000;118:615–21.
EUROPEAN OPHTHALMIC REVIEW