Content uploaded by Milton M Hom
Author content
All content in this area was uploaded by Milton M Hom on Dec 13, 2017
Content may be subject to copyright.
CLINICAL SCIENCE
Dysfunctional Tear Syndrome
A Delphi Approach to Treatment Recommendations
Ashley Behrens, MD,* John J. Doyle, MPH,† Lee Stern, MS,† Roy S. Chuck, MD, PhD,*
Peter J. McDonnell, MD* and the Dysfunctional Tear Syndrome Study Group: Dimitri T. Azar, MD,
Harminder S. Dua, MD, PhD, Milton Hom, OD, Paul M. Karpecki, OD, Peter R. Laibson, MD,
Michael A. Lemp, MD, David M. Meisler, MD, Juan Murube del Castillo, MD, PhD,
Terrence P. O’Brien, MD, Stephen C. Pflugfelder, MD, Maurizio Rolando, MD,
Oliver D. Schein, MD, MPH, Berthold Seitz, MD, Scheffer C. Tseng, MD, PhD,
Gysbert van Setten, MD, PhD, Steven E. Wilson, MD, and Samuel C. Yiu, MD, PhD
Purpose: To develop current treatment recommendations for dry
eye disease from consensus of expert advice.
Methods: Of 25 preselected international specialists on dry eye, 17
agreed to participate in a modified, 2-round Delphi panel approach.
Based on available literature and standards of care, a survey was
presented to each panelist. A two-thirds majority was used for
consensus building from responses obtained. Treatment algorithms
were created. Treatment recommendations for different types and
severity levels of dry eye disease were the main outcome.
Results: A new term for dry eye disease was proposed: dysfunctional
tear syndrome (DTS). Treatment recommendations were based
primarily on patient symptoms and signs. Available diagnostic tests
were considered of secondary importance in guiding therapy.
Development of algorithms was based on the presence or absence
of lid margin disease and disturbances of tear distribution and
clearance. Disease severity was considered the most important factor
for treatment decision-making and was categorized into 4 levels.
Severity was assessed on the basis of tear substitute requirements,
symptoms of ocular discomfort, and visual disturbance. Clinical signs
present in lids, tear film, conjunctiva, and cornea were also used for
categorization of severity. Consensus was reached on treatment al-
gorithms for DTS with and without concurrent lid disease.
Conclusion: Panelist opinion relied on symptoms and signs (not
tests) for selection of treatment strategies. Therapy is chosen to match
disease severity and presence versus absence of lid margin disease or
tear distribution and clearance disturbances.
Key Words: Delphi panel, dry eye, dysfunctional tear syndrome, eye
lubricants, cyclosporine A, punctal plugs, steroids, dry eye therapy,
concensus, algorithm
(Cornea 2006;25:900–907)
The syndrome known as ‘‘dry eye’’ is highly prevalent,
affecting 14% to 33% of the population worldwide,
1–4
depending on the study and definition used. Symptoms related
to dry eye are among the leading causes of patient visits to
ophthalmologists and optometrists in the United States.
5
However, a stepwise approach to diagnosis and treatment is
not well established.
Treatment algorithms are often complicated, especially
when multiple therapeutic agents and strategies are available
for one single disease and for different stages of the same
disease. Dry eye syndrome is particularly challenging, because
the diagnostic criteria used vary among studies, there is poor
correlation between signs and symptoms, and efficacy criteria
are often not uniform. As a result, there is no clear current
approach to assign therapeutic recommendations as ‘‘first,’’
‘‘second,’’ or ‘‘third’’ line.
Clinical research is usually oriented to assess the efficacy
of medications in the treatment of dry eye disease. Reports are
based on either comparisons of one medication relative to
untreated placebo controls or comparisons between different
therapies.
6,7
Categorization of treatment alternatives is usually
not implicit in these studies. Strategies combining medications
or medications and surgery are usually not clearly discussed in
the literature. A panel of experts may be a good method to
develop such strategies based on current knowledge, because
publication of research may not precede practice. Furthermore,
clinical trials are typically performed on highly selected
populations with specific interventions that may not reflect
the spectrum of disease encountered in usual practice.
Where unanimity of opinion does not exist because of a
paucity of scientific evidence and where there is contradictory
evidence, consensus methods can be useful. Such methods
have been used in developing therapeutic algorithms in other
ophthalmic (glaucoma) and nonophthalmic disease states.
8,9
Received for publication June 21, 2005; revision received January 3, 2006;
accepted January 10, 2006.
From the *Wilmer Ophthalmological Institute, Johns Hopkins University
School of Medicine, Baltimore, MD; and the †Analytica Group,
New York, NY.
Supported by unrestricted educational grants from Allergan Inc. (Irvine, CA)
and Research to Prevent Blindness, Inc. (New York, NY).
Disclaimer: Some authors have commercial or proprietary interests in
products described in this study (please refer to individual disclosure).
Reprints: Ashley Behrens, MD, The Wilmer Ophthalmological Institute, 255
Woods Building, 600 North Wolfe Street, Baltimore, MD 21287-9278
(e-mail: abehrens@jhmi.edu).
Copyright !2006 by Lippincott Williams & Wilkins
900 Cornea !Volume 25, Number 8, September 2006
JOBNAME: corn 25#8 2006 PAGE: 1 OUTPUT: Friday November 3 17:15:33 2006
lww/corn/127758/ICO200300
The Delphi panel technique was first proposed in 1946
by the RAND Corporation as a resource to collect information
from different experts and to prepare a forecast of future
technological capabilities. This tool has been expanded to
technological,
10
health,
11
and social sciences research.
12
De-
spite some reasonable criticisms of this technique,
13
the Delphi
approach has been used to provide reproducible consensus to
create algorithms of treatment.
14,15
In this study, we proposed to establish expert consensus
by using the Delphi approach with an international panel to
obtain current treatment recommendations for dry eye syndrome.
MATERIALS AND METHODS
Panelist Selection
The ideal number of panelists expected with this
technique is not well defined, with reported ranges from 10
to 1685.
16
No specific inclusion criteria are established, other
than the qualification of panelists in the topic of interest. Some
authors stress the importance of the diversity of panelists’
opinion to obtain a wide base of knowledge.
17
The following criteria were considered for inclusion of
panelists:
1. Active clinicians (ophthalmologists and optometrists)
2. Scientific contributions to clinical research on dry eye
syndrome, as reflected by at least 2 of the following: peer-
reviewed publications, other forms of written scientific com-
munication, specialty meeting presentations, and member-
ship in special-interest groups focused on dry eye syndrome
3. International representation
4. Proficiency in English language to facilitate interaction
5. Able to respond to sets of questionnaires and available to
attend a final meeting at the Wilmer Ophthalmological
Institute in Baltimore, MD
The search for panelists’ scientific contributions was
conducted over available medical databases (Medline, EM-
BASE) and other major Internet-based search engines
(Scirus.com, Google.com, Alltheweb.com). Twenty-five can-
didates from 3 continents that met the selection criteria were
initially contacted.
A contract research organization (Analytica Group, New
York, NY) was selected to act as moderator/facilitator for the
questionnaire and panel meeting exercise. A 2-round modified
Delphi approach was used.
18
A set of dry eye therapy literature
was provided to each panel member along with the first-round
questionnaire. These studies were selected in part from an
ongoing systematic review of the literature on dry eye disease
therapy. Three of the panelists suggested additions of some
references that they considered valuable. Those citations were
also disseminated to the rest of the panelists.
Preparation of Surveys
Questionnaires were based on collected literature, current
practice patterns, and clinical experience in dry eye. Topics in
the survey were related to pathophysiology, diagnostic tests,
criteria used to guide treatment, and therapeutic alternatives.
Nominal variables were assigned binary values to
tabulate responses in a spreadsheet (Excel 2002; Microsoft
Corp., Redmond, WA) for analysis. Ordinal variables were
originated from 5-point Likert scales to categorize the strength
of agreement and facilitate the statistical analysis.
Survey questions were based on the use of the current
classification of dry eye disease and the available guidelines
for the treatment. Diagnostic methods and severity assessment
were also surveyed. Panelists were asked to support their multi-
level treatment recommendation with a categorical, nominal
score of 1 to 3, depending on the level of evidence to sustain
their decision:
1. Supported by a clinical trial
2. Supported by published literature of some type
3. Supported by my professional opinion
Finally, determinant factors influencing the treatment
decision-making process were stratified semiquantitatively to
evaluate the most representative for the selection of therapy.
Survey Deployment
The forms were deployed by electronic mail to the
panelists. The information obtained from the surveys was
tabulated and organized for presentation at the face-to-face
meeting of the Delphi process.
Data Analysis
Descriptive statistics were calculated for the question-
naire data by using StatsDirect 2.3.7 for Windows (StatsDirect,
Cheshire, UK).
Consensus
There exists controversy regarding the numbers neces-
sary to obtain consensus. Some authors agree that a simple
majority (.50%) is enough to constitute consensus,
19
whereas
others propose that more than 80% of panelists should be in
agreement to have the recommendation considered as con-
sensual.
20
Degree of consensus has also been quantified
statistically using the Cronbach amethod, a method for
measuring internal agreement.
21
For the purposes of this study,
consensus was defined as a two-thirds majority.
Personal Interaction
The meeting was conducted by a facilitator (J.J.D.) with
previous experience in consensus-building strategies.
8
Panel-
ists reacted and discussed the data collected from the surveys
over an intensive 1-day, 12-hour-long, face-to-face meeting.
According to the tabulated initial responses, iterative discus-
sions were conducted toward majority agreement.
RESULTS
Panelists’ Response
From the initial selection of 25 candidates who met the
inclusion criteria, 17 were able to participate in all stages of the
study and therefore were included in the panel. The candidates
who refused to join the panel did not have substantive reasons
precluding their participation. Most of them declined to
participate because of scheduling conflicts. The list of par-
ticipants is shown in Table 1. All surveys deployed were re-
turned with responses from all of the panelists.
q2006 Lippincott Williams & Wilkins 901
Cornea !Volume 25, Number 8, September 2006 Dysfunctional Tear Syndrome
JOBNAME: corn 25#8 2006 PAGE: 2 OUTPUT: Friday November 3 17:15:34 2006
lww/corn/127758/ICO200300
Conflicts of Interest
Travel expenses of panelists were covered by the
contracted company (Analytica Group), which is an in-
dependent firm. The Wilmer Eye Institute originated the
invitation, and panelists were unaware of any indirect support
from pharmaceutical industry to avoid bias in the treatment
selection.
Use of Existing Disease/Treatment Guidelines
The majority of panelists (11 of 17) responded that they
did not follow any of the available guidelines for the treatment
of dry eye syndrome. Three of 17 followed the National Eye
Institute guidelines,
22
1 of 17 followed the American Academy
of Ophthalmology Preferred Practice Patterns,
23
1 of 17 fol-
lowed the Madrid classification,
24
and 1 of 17 followed a com-
bination of the first 2 guidelines.
When panel members were asked about their opinions
regarding the adherence of the ophthalmic community to new,
simplified guidelines for the treatment of dry eye, the majority
(13 of 17) agreed that they would use them if most recent
findings on the disease were included. Those who responded
that they would not use them (4 of 17), based their response on
the low sensitivity and specificity of the available tests for the
diagnosis of dry eye and the variability of the clinical
presentation in different patients.
Diagnostic Tests for Dry Eye
When panelists were surveyed before the meeting on
diagnostic measures used to detect dry eye, the most fre-
quently cited tests were slit-lamp examination and fluorescein
staining (100% of panelists). Tear breakup time and medical
history were also frequently used (both in 94%). Schirmer test
with anesthesia (71%) and without anesthesia (65%) were less
frequently used, as well as rose bengal staining (65%). A
combination of different tests was typically preferred in an
effort to improve the specificity and sensitivity (Table 2).
Classification of Dry Eye Disease
More than one half of the respondents felt that the
current classification of aqueous-deficient versus evaporative
dry eye failed to incorporate inflammatory mechanisms and
drew a sharp distinction between disorders where there is
significant overlap.
25,26
Furthermore, the historical distinction
between Sjo¨gren keratoconjunctivitis sicca (KCS) as repre-
senting an autoimmune disorder as opposed to non-Sjo¨ gren
KCS failed to reflect the evidence that both conditions may
share an underlying immune-mediated inflammation. The
majority of experts did not consider this useful for establishing
a treatment scheme for the ocular disease (12 of 17). The
panelists considered the disease severity and the effect of
medications on symptoms and signs as the 2 most relevant
factors to consider when selecting the adequate therapy for dry
eye (Table 3).
Face-to-Face Meeting
At the face-to-face meeting, panel members made
comments on the term ‘‘dry eye’’ classically used to name the
disease. On the basis of the known pathophysiology, symp-
toms, and clinical presentation, all panelists agreed that this
term did not necessarily reflect the events occurring in the eye.
Specifically, all patients with this condition do not necessarily
TABLE 1. Experts Who Participated in the Delphi Approach
(DTS Study Group)
Panelist Name City Country
Dimitri T. Azar, M.D. Boston, MA United States
Harminder S. Dua, M.D., Ph.D Nottingham England
Milton Hom, O.D. Azusa, CA United States
Paul M. Karpecki, O.D. Overland Park, KS United States
Peter R. Laibson, M.D. Philadelphia, PA United States
Michael A. Lemp, M.D. Washington, DC United States
David M. Meisler, M.D. Cleveland, OH United States
Juan Murube del Castillo, M.D., Ph.D. Madrid Spain
Terrence P. O’Brien, M.D. Baltimore, MD United States
Stephen C. Pflugfelder, M.D. Houston, TX United States
Maurizio Rolando, M.D. Genoa Italy
Oliver D. Schein, M.D., M.P.H. Baltimore, MD United States
Berthold Seitz, M.D. Erlangen Germany
Scheffer C. Tseng, M.D., Ph.D. Miami, FL United States
Gysbert B. van Setten, M.D., Ph.D. Stockholm Sweden
Steven E. Wilson, M.D. Cleveland, OH United States
Samuel C. Yiu, M.D, Ph.D. Los Angeles, CA United States
TABLE 2. Most Commonly Used Diagnostic Tests Reported
by Panelists for Evaluating a Patient With Probable Dry Eye
Diagnostic Tests
Respondents Regularly
Using Them (%)
Fluorescein staining 100
Tear breakup time 94
Schirmer test 71
Rose bengal staining 65
Corneal topography 41
Impression cytology 24
Tear fluorescein clearance 24
Ocular Surface Disease Index Questionnaire 18
NEIVFQ-25* 6
Tear osmolarity 6
Conjunctival biopsy 6
*NEIVFQ-25: National Eye Institute Vision Function Questionnaire-25.
TABLE 3. Most Relevant Factors Influencing Treatment
Decision Making
Factor Considered Mean Score (Standard Deviation)
Severity of the disease 1.47 (0.72)
Effect of the treatment 1.79 (0.77)
Etiology of the disease 2.08 (1.07)
Diagnosis of Sjo¨gren’s syndrome 2.20 (1.05)
Use of artificial tears 3.07 (1.53)
Costs of treatment 3.80 (1.17)
Access to reimbursement 3.92 (1.10)
0 = most relevant; 5 = least relevant.
902 q2006 Lippincott Williams & Wilkins
Behrens et al Cornea !Volume 25, Number 8, September 2006
JOBNAME: corn 25#8 2006 PAGE: 3 OUTPUT: Friday November 3 17:15:36 2006
lww/corn/127758/ICO200300
suffer from reduced tear volume but rather may have abnor-
malities of tear film composition that include the presence of
proinflammatory cytokines.
25–27
The panelists unanimously
recommended dysfunctional tear syndrome (DTS) as a more
appropriate term for this disease in future references. This term
has been incorporated in the rest of this report in lieu of dry eye
disease.
Underlying Pathophysiology and
Diagnostic Testing
There was consensus that most cases of DTS have an
inflammatory basis that either triggers or maintains the
condition. However, panelists also agreed on the difficulty
in clearly identifying inflammation in most patients. The panel
therefore agreed to subclassify the disease as either DTS with
clinically apparent inflammation or DTS without clinically
evident inflammation.
After discussion at the meeting, the panelists were in
agreement that commonly available clinical diagnostic tests
did not correlate with symptoms, should not be used in
isolation to establish the diagnosis of DTS, and were of
minimal value in the assessment of disease severity.
Creation of Therapeutic Algorithms for DTS
First, the panel recommended that patients with DTS
should be classified into 1 of 3 major clinical categories at the
time of the initial examination: patients with lid margin
disease, patients without lid margin disease, and patients with
altered tear distribution and clearance.
The panel agreed that the second group, patients who do
not have coexistent lid margin disease, is the most common
form of presentation of DTS. Within each of these 3 cat-
egories, the panel listed the main subsets or specific disease
entities or, in the case of DTS without lid margin disease, the
patients were divided by severity (Fig. 1). Second, the panel
agreed that the assessment of DTS severity is important to
guiding therapy, especially in that subset of DTS patients
without lid margin disease. The panel reached consensus that
the level of severity should be based primarily on symptoms
and clinical signs.
The panel members agreed that diagnostic tests are
secondary considerations in determining disease severity. The
value of diagnostic tests was considered to be in confirming
clinical assessment. Again, many of the available tests were
deemed not useful for the diagnosis, staging, or evaluating
response to therapy in DTS.
Panelists agreed on 3 particularly relevant symptoms and
historical elements to be considered in DTS: ocular discomfort,
tear substitute requirements, and visual disturbances. In ocular
discomfort, a variety of symptoms including itch, scratch, burn,
foreign body sensation, and/or photophobia may be present.
Depending on the frequency and impact on the quality of life
of these elements, symptoms could be categorized as either
mild to moderate or severe. The relevant clinical signs to be
considered in the evaluation of DTS patients are summarized in
Table 4. The panel suggested evaluating the presence of these
clinical features to assign a severity level fluctuating from mild
to severe.
To create a categorization of the severity of the disease,
a scoring system was proposed. Basically, patients were ag-
gregated into 1 of 4 levels of severity according to the signs
and symptoms involved (Table 5). The severity of disease
indicated the appropriate range of therapeutic options available
for the patient, because the panelists agreed that certain
therapies were most appropriately reserved for patients with
more severe DTS.
Treatment Algorithm for Patients With Lid
Margin Disease
The proposed treatment algorithm for these individuals
began with division of patients according to the site (anterior
vs. posterior) of the lid pathology (Fig. 2). Anterior lid margin
disease is treated with lid hygiene and antibacterial therapy,
whereas posterior lid margin disease is treated initially with
FIGURE 1. Algorithm of the 3 major
subsets found in DTS. Each subset
should be treated separately, be-
cause treatment modality varies ac-
cording to this separation.
q2006 Lippincott Williams & Wilkins 903
Cornea !Volume 25, Number 8, September 2006 Dysfunctional Tear Syndrome
JOBNAME: corn 25#8 2006 PAGE: 4 OUTPUT: Friday November 3 17:15:37 2006
lww/corn/127758/ICO200300
warm massage, with addition of oral tetracyclines and topical
corticosteroids, if necessary.
Treatment Algorithm for DTS Patients With
Primary Tear Distribution and
Clearance Abnormalities
The panel considered that there were patients in whom
the even distribution of tears across the ocular surface is
impaired, typically related to an anatomic abnormality or to
abnormal lid function (Fig. 3). The recommended therapeutic
approach to these patients varied in accordance with the
specific underlying problem, which is summarized in Figure 3.
Treatment Algorithm for DTS Patients Without
Lid Margin Disease
Patients with mild disease are best managed with patient
education about the disease and strategies for minimizing its
impact, preserved artificial tears, modification as appropriate
of systemic medications that might contribute to the condition,
and perhaps changes in the home or work environment to
alleviate the symptoms (Fig. 4).
In patients in whom the disease state is moderate or
severe, the panelists agreed that the more frequent use of tears
mandated a switch to unpreserved lubricants, with tears during
the day, ointment at night, and consideration of progression to
a gel formulation during the day if relief was not adequate with
tears. In the absence of signs, the panel recommended lubri-
cation, with frequency determined by the clinical response.
In the presence of signs (eg, moderate corneal staining,
filaments), the panel agreed on a stepwise introduction of
additional therapies. The panelists noted that patients with DTS
may have an inflammatory component, which may or may not
be clinically evident. In addition to the use of unpreserved tears,
the panel recommended a course of topical corticosteroids
and/or cyclosporine A to suppress inflammation.
In patients who fail to respond adequately to lubricants
and topical immunomodulators, a course of oral tetracycline
therapy was recommended, as well as punctal occlusion with
TABLE 4. Clinical Signs in DTS to Consider in Severity Assessment
Lids Tear Film Conjunctiva Cornea Vision
Telangiectasia Meniscus Luster Punctate changes Blur
Hyperemia Foam Hyperemia Erosions (micro, macro) Fluctuations
Scales, crusts Mucus Wrinkles Filaments
Lash loss or Debris Staining Ulceration
abnormalities Oil excess Symblepharon Vascularization
Inspissation Cicatrization Scarring
Meibomian gland disease Keratinization
Anatomical abnormalities
TABLE 5. Levels of Severity of DTS Without Lid Margin
Disease According to Symptoms and Signs
Severity* Patient Profiles
Level 1 !Mild to moderate symptoms and no signs
!Mild to moderate conjunctival signs
Level 2 !Moderate to severe symptoms
!Tear film signs
!Mild corneal punctate staining
!Conjunctival staining
!Visual signs
Level 3 !Severe symptoms
!Marked corneal punctate staining
!Central corneal staining
!Filamentary keratitis
Level 4 !Severe symptoms
!Severe corneal staining, erosions
!Conjunctival scarring
*At least one sign and one symptom of each category should be present to qualify for
the corresponding level assignment. FIGURE 2. Algorithm on treatment recommendations for DTS
with lid margin disease.
904 q2006 Lippincott Williams & Wilkins
Behrens et al Cornea !Volume 25, Number 8, September 2006
JOBNAME: corn 25#8 2006 PAGE: 5 OUTPUT: Friday November 3 17:15:42 2006
lww/corn/127758/ICO200300
plugs. Because of the possible presence of non–clinically
apparent inflammation, punctal plugs could result in retention
of proinflammatory tear components on the ocular surface and
may enhance damage to the ocular surface, accelerate the
disease process, and produce greater patient discomfort. There-
fore, the panel agreed that it is important to treat the inflam-
matory condition before blockage of tear drainage with
punctal plugs.
Patients with severe disease who are not adequately con-
trolled after the above therapeutic interventions may benefit
from more advanced interventions. These would include sys-
temic immunomodulators for the control of severe inflamma-
tion, topical acetylcysteine for filament formation caused by
mucin accumulation, moisture goggles to reduce tear evap-
oration, and surgery (including punctal cautery) to reduce tear
drainage. Patients with Sjo¨ gren syndrome would fit within this
category.
DISCUSSION
Some researchers have stressed the use of Delphi panels
in clinical research, despite some flaws in terms of
reproducibility and other confounding factors that may
adversely influence the results.
28,29
Delphi approach is not
necessarily ‘‘evidence-based’’: Good evidence may exist
contradicting a particular consensus; or conversely, evidence
for a particular consensus may be absent, because it has not
been adequately studied. Especially for areas where there is little
or no good evidence in the literature, the process relies on the
opinion of the participating panelists, potentially tapping into
collective error.
30
Moreover, consensus is subject to particular
interpretation of evidence and personal experience, which may
affect reproducibility.
14
Nonetheless, this process has lately
become popular to delineate guidelines of treatment of various
disorders.
30–33
Bias of panelists’ selection may inevitably occur as
a result of the inclusion criteria chosen. It is a common
observation that highly published authors tend to have some
FIGURE 3. Algorithm on treatment recommendations for DTS
with abnormal tear distribution.
FIGURE 4. Algorithm on treatment recommendations for DTS
without lid margin disease according to severity.
q2006 Lippincott Williams & Wilkins 905
Cornea !Volume 25, Number 8, September 2006 Dysfunctional Tear Syndrome
JOBNAME: corn 25#8 2006 PAGE: 6 OUTPUT: Friday November 3 17:15:45 2006
lww/corn/127758/ICO200300
form of commercial support from pharmaceutical industry.
Nine of 17 panelists disclosed a past or present relationship as
a speaker/consultant/research funds recipient from companies
having products for the treatment of DTS.
The success of a Delphi panel is based largely on the
ability of the facilitator to maintain balanced participation of
panelists.
32
One of the major challenges in such panels is to
avoid the inadvertent control of one or more leaders over the
discussion.
30
The facilitator in our study was a person with
previous experience in consensus panels. He had the ability to
encourage homogeneous participation of panel members. The
facilitator focused on the varied responses previously given by
panelists in the survey to avoid discussions over a single
topic/therapeutic approach raised by individual participants
during the meeting. Inevitable discrepancies were observed
during the DTS panel meeting; however, consensual agree-
ment among panelists was finally achieved.
We believe that one significant consequence of the panel
meeting was the recommendation for a change from the term
dry eye, frequently used to describe the condition, to the term
dysfunctional tear syndrome. Panelists unanimously agreed that
the label dry eye reflects neither patient symptoms nor neces-
sarily the pathogenic mechanism of the disease. Panel members
also agreed that diagnosing patients with dry eye may be
misleading to both colleagues and patients. Patients may be
confused when excess tearing is their primary complaint and
are diagnosed as having dry eye. Even more confusing for
patients is their subsequent treatment with anti-inflammatory
agents or antibiotics. For these reasons, the term DTS was
coined, because the panel felt that this term was sufficiently
broad to encompass the myriad of etiologies while still
representing a common denominator among them.
There was consensus that severity of disease should be
the primary determinant for the therapeutic strategy chosen. In
addition, observation of the patient response to initial therapy
was deemed as an important indicator of disease severity and
further treatment selection. The failure on improvement using
medications in one level assigns the patient to additional
therapy in the immediate superior severity level. The available
diagnostic tests were not considered important in the
assessment of disease severity and therefore were not included
in the classification. However, this should not underestimate
the value of these tests in the diagnosis of DTS, because they
were regularly used by panelists to confirm the presence of the
disease.
The task of creating guidelines for DTS is complex,
because practitioners encountering DTS are faced with a mul-
tifactorial disorder with several pathophysiological events that
may require a variety of customized therapeutic schemes.
Moreover, significant overlapping between the categories
selected by the panel is also likely. The summary treatment
recommendations (Table 6) relating severity of disease with
clinical symptoms and signs created by the panel may serve as
a useful guide. It is recognized that individual patient
characteristics may require deviation from recommended
treatment, but panelists were clear that the ideal therapy for
DTS is often achieved with a combination of interventions.
Assignment of levels of severity may work only as a stepwise
guide to approaching the best combination of medications to
avoid symptoms. It is important to stress that patients may
present with signs belonging to different categories of DTS (ie,
a patient may have DTS with lid margin disease and exhibit
tear distribution problems).
Those particular patients should be treated according to
recommendations for both categories to succeed in controlling
their symptoms and signs. Published guidelines in other dis-
ease areas have proven useful to general practitioners to ap-
proach a complex disease like DTS.
14,15,17
Some examples
using the Delphi technique have been reported in esophageal
cancer management,
11
systemic hypertension treatment algo-
rithms,
15
and acute diarrhea management in children.
30
In this
study, the Delphi approach was used to gain a practical
approach to the diagnosis and treatment of DTS, as opposed to
an extensive evaluation of available diagnostic methods or
pathophysiology mechanisms, already well documented in the
literature
34–38
(Table 7).
TABLE 6. Treatment Recommendations for DTS on the Basis
of Level of Severity
DTS Severity
Treatment
Recommendations
Level 1 !No treatment !Use of hypoallergenic
products
!Preserved tears !Water intake
!Environmental
management
!Psychological support
!Allergy drops !Avoidance of drugs
contributing to
dry eye
Level 2 !Unpreserved tears !Secretagogues
!Gels !Topical steroids
!Ointments !Topical cyclosporine A
!Nutritional support
(flaxseed/fatty acids)
Level 3 !Tetracyclines
!Punctal plugs
Level 4 !Surgery !Punctal cautery
!Systemic
anti-inflammatory
!Acetylcysteine
therapy !Contact lenses
!Oral cyclosporine
!Moisture goggles
TABLE 7. Advantages of the Proposed Recommendations by
the Delphi Panel
!Proposes a new terminology for dry eye disease (dysfunctional tear
syndrome) from recent pathophysiologic findings
!Includes novel therapeutic options in the market
!Provides simplified therapeutic recommendations in a stepwise approach
!Patients without lid margin disease/tear distribution problems are assigned to
4 severity levels
!Severity levels are categorized according to patient’s signs and symptoms,
not tests
!Therapeutic options are oriented by severity levels
!Easier approach for general eye care practitioners
906 q2006 Lippincott Williams & Wilkins
Behrens et al Cornea !Volume 25, Number 8, September 2006
JOBNAME: corn 25#8 2006 PAGE: 7 OUTPUT: Friday November 3 17:15:52 2006
lww/corn/127758/ICO200300
All guidelines are limited by the future development of
new treatments and by new insights that future research will
bring. We therefore regard these guidelines as a platform onto
which future updates may be added.
REFERENCES
1. Schein OD, Munoz B, Tielsch JM, et al. Prevalence of dry eye among the
elderly. Am J Ophthalmol. 1997;124:723–728.
2. Schaumberg DA, Sullivan DA, Buring JE, et al. Prevalence of dry eye
syndrome among US women. Am J Ophthalmol. 2003;136:318–326.
3. Lin PY, Tsai SY, Cheng CY, et al. Prevalence of dry eye among an elderly
Chinese population in Taiwan: the Shihpai Eye Study. Ophthalmology.
2003;110:1096–1101.
4. Brewitt H, Sistani F. Dry eye disease: the scale of the problem. Surv
Ophthalmol. 2001;45:S199–S202.
5. Lemp MA. Epidemiology and classification of dry eye. Adv Exp Med Biol.
1998;438:791–803.
6. Matsuo T, Tsuchida Y, Morimoto N. Trehalose eye drops in the treatment
of dye eye syndrome. Ophthalmology. 2002;109:2024–2029.
7. McDonald CC, Kaye SB, Figueiredo FC, et al. A randomised, crossover,
multicentre study to compare the performance of 0.1% (w/v) sodium
hyaluronate with 1.4% (w/v) polyvinyl alcohol in the alleviation of
symptoms associated with dry eye syndrome. Eye. 2002;16:601–607.
8. Serle J, Cantor L, Gross R, et al. Best practice treatment algorithm for
primary open-angle glaucoma: implications for U.S. ophthalmology
practice. Manag Care Interface. 2002;15:37–48.
9. Coia LR, Minsky BD, John MJ, et al. Patterns of care study decision tree
and management guidelines for esophageal cancer. American College of
Radiology. Radiat Med. 1998;16:321–327.
10. Pelton JN. The future of telecommunications: a Delphi survey. J Commun.
1981;31:177–189.
11. Bellamy N, Anastassiades TP, Buchanan WW, et al. Rheumatoid arthritis
anti-rheumatic trials. III. Setting the delta for clinical trials of anti-
rheumatic drugs—results of a consensus development (Delphi) exercise.
J Rheumatol. 1991;18:1908–1915.
12. Holmes ER, Tipton DJ, Desselle SP. The impact of the internet on
community pharmacy practice: a comparison of a Delphi panel’s forecast
with emerging trends. Health Mark Q. 2002;20:3–29.
13. Goodman CM. The Delphi technique: a critique. J Adv Nurs. 1987;12:
729–734.
14. Pearson SD, Margolis CZ, Davis S, et al. Is consensus reproducible? A
study of an algorithmic guidelines development process. Med Care. 1995;
33:643–660.
15. Richter A, Ostrowski C, Dombeck MP, et al. Delphi panel study of current
hypertension treatment patterns. Clin Ther. 2001;23:160–167.
16. Powell C. The Delphi technique: myths and realities. J Adv Nurs. 2003;41:
376–382.
17. Murphy MK, Black NA, Lamping DL, et al. Consensus development
methods, and their use in clinical guideline development. Health Technol
Assess. 1998;2:i–iv, 1-88.
18. Young LJ, George J. Do guidelines improve the process and outcomes of
care in delirium? Age Ageing. 2003;32:525–528.
19. Evans C. The use of consensus methods and expert panels in
pharmacoeconomic studies. Practical applications and methodological
shortcomings. Pharmacoeconomics. 1997;12:121–129.
20. Morris CJ, Cantrill JA. Preventing drug-related morbidity—the develop-
ment of quality indicators. J Clin Pharm Ther. 2003;28:295–305.
21. Hughes R. Definitions for public health nutrition: a developing consensus.
Public Health Nutr. 2003;6:615–620.
22. Lemp MA. Report of the National Eye Institute/Industry workshop on
Clinical Trials in Dry Eyes. CLAO J. 1995;21:221–232.
23. Matoba AY, Harris DJ, Meisler DM, et al. Preferred Practice Patterns:
Dry Eye Syndrome. San Francisco, CA: American Academy of
Ophthalmology; 2003.
24. Murube J, Benitez Del Castillo JM, Chenzhuo L, et al. The Madrid triple
classification of dry eye. Arch Soc Esp Oftalmol. 2003;78:587–594.
25. Pflugfelder SC. Anti-inflammatory therapy for dry eye. Am J Ophthalmol.
2004;137:337–342.
26. Baudouin C. The pathology of dry eye. Surv Ophthalmol. 2001;45 (Suppl
2):S211–S220.
27. Lemp MA. Evaluation and differential diagnosis of keratoconjunctivitis
sicca. J Rheumatol Suppl. 2000;61:11–14.
28. Burnand B, Vader JP, Froehlich F, et al. Reliability of panel-based
guidelines for colonoscopy: an international comparison. Gastrointest
Endosc. 1998;47:162–166.
29. Jones J, Hunter D. Consensus methods for medical and health services
research. BMJ. 1995;311:376–380.
30. Armon K, Stephenson T, MacFaul R, et al. An evidence and consensus
based guideline for acute diarrhea management. Arch Dis Child. 2001;85:
132–142.
31. Campbell SM, Hann M, Roland MO, et al. The effect of panel
membership and feedback on ratings in a two-round Delphi survey: results
of a randomized controlled trial. Med Care. 1999;37:964–968.
32. Washington DL, Bernstein SJ, Kahan JP, et al. Reliability of clinical
guideline development using mail-only versus in-person expert panels.
Med Care. 2003;41:1374–1381.
33. Mathis R, Doyle S. A quality mix: using evidence and experience to
evaluate new technologies. J Healthc Qual. 2003;25:4–6.
34. Smith JA, Vitale S, Reed GF, et al. Dry eye signs and symptoms in
women with premature ovarian failure. Arch Ophthalmol. 2004;122:
151–156.
35. Horwath-Winter J, Berghold A, Schmut O, et al. Evaluation of the clinical
course of dry eye syndrome. Arch Ophthalmol. 2003;121:1364–1368.
36. Bron AJ, Evans VE, Smith JA. Grading of corneal and conjunctival
staining in the context of other dry eye tests. Cornea. 2003;22:640–650.
37. Sade de Paiva C, Lindsey JL, Pflugfelder SC. Assessing the severity of
keratitis sicca with videokeratoscopic indices. Ophthalmology. 2003;110:
1102–1109.
38. Begley CG, Chalmers RL, Abetz L, et al. The relationship between
habitual patient-reported symptoms and clinical signs among patients with
dry eye of varying severity. Invest Ophthalmol Vis Sci. 2003;44:4753–
4761.
q2006 Lippincott Williams & Wilkins 907
Cornea !Volume 25, Number 8, September 2006 Dysfunctional Tear Syndrome
JOBNAME: corn 25#8 2006 PAGE: 8 OUTPUT: Friday November 3 17:15:54 2006
lww/corn/127758/ICO200300