The Comparison of Efficacies of Topical
Corticosteroids and Nonsteroidal Anti-
inflammatory Drops on Dry Eye Patients: A
Clinical and Immunocytochemical Study
AVNI MURAT AVUNDUK, MD, MUSTAFA CIHAT AVUNDUK, MD,
EMILY D. VARNELL, BS, AND HERBERT E. KAUFMAN, MD
● PURPOSE: To investigate whether conjunctival inflam-
mation represents a primary event in the pathogenesis of
keratoconjunctivitis sicca or whether it is a secondary
inflammatory reaction caused by enhanced mechanical
irritation as a result of surface dryness and whether
anti-inflammatory drops (corticosteroids and nonsteroi-
dal anti-inflammatory) have therapeutic effects and are
similar.
● DESIGN: Single-masked, randomized, prospective clin-
ical trial.
● METHODS: Thirty-two keratoconjuctivitis patients
with or without Sjögren syndrome were included in the
study. The patients were randomized to three groups.
Group 1 patients received a topical artificial tear substi-
tute (ATS); group 2 received ATS plus nonsteroidal
anti-inflammatory drops (NSAID); and group 3 received
ATS plus topical corticosteroidal drops. The eye symp-
tom severity scores, Schirmer test values, rose bengal and
fluorescein staining scores were evaluated before treat-
ment and 15 and 30 days after start of treatment.
Impression cytology specimens were stained using immu-
nohistochemical methods to detect the percentages of
human leukocyte antigen II (HLA-DR) positive, Apo
2.7 positive, and periodic acid–Schiff positive cells. Sta-
tistical analyses were performed within and between
groups.
Accepted for publication March 14, 2003.
InternetAdvance publication at ajo.com April 24, 2003.
From the Louisiana State University, School of Medicine, LSU Eye
Center, New Orleans, Louisiana (A.M.A., E.D.V., H.E.K.); and Selçuk
University, School of Medicine, Department of Pathology, Konya,
Turkey (M.C.A.).
This study was supported in part by US Public Health Service Grant
EY02377 (H.E.K.) from the National Eye Institute, National Institutes of
Health, Bethesda, Maryland, and an unrestricted departmental grant from
Research to Prevent Blindness, Inc., New York, New York.
Inquiries to Avni Murat Avunduk, MD, KTU Lojmanlari No: 31/17
61080, Trabzon, Turkey; fax: (⫹90) 462-3252270; e-mail: avunduk@
ttnet.net.tr
0002-9394/03/$30.00 © 2003 BY ELSEVIER INC. ALL
doi:10.1016/S0002-9394(03)00326-X
● RESULTS: Group 3 patients had significantly lower
symptom severity scores, fluorescein and rose bengal
staining, and HLA-DR positive cells on days 15 and 30
compared with patients in other groups. They also had a
significantly higher number of periodic acid–Schiff posi-
tive (goblet) cells in their impression cytology specimens
on days 15 and 30 compared with the other patients. On
day 30, group 3 patients had significant differences
compared with their baseline measurements in terms of
above-mentioned parameters. However, we did not detect
a significant effect of any treatment schedule on the
Shirmer test value and the numbers of Apo 2.7 cells in
impression cytology specimens.
● CONCLUSIONS: Topical corticosteroids had a clearly
beneficial effect both on the subjective and objective
clinical parameters of moderate-to-severe dry eye pa-
tients. These effects were associated with the reduction
of inflammation markers of conjunctival epithelial cells.
(Am J Ophthalmol 2003;136:593– 602. © 2003 by
Elsevier Inc. All rights reserved.)
A
LTHOUGH THE IMMUNOPATHOLOGIC ANALYSIS OF
the lacrimal gland has received considerable atten-
tion, less work has been done on pathologic
changes occurring in the ocular surface of patients with
keratoconjunctivitis sicca (KCS). However, a strong ex-
pression of human leukocyte antigen II (HLA-DR) anti-
gens has been found in conjunctival epithelial cells of dry
eye patients.1 A strong relationship also has been proposed
between inflammatory pathways and apoptosis, which
directly affects epithelial turnover.2 Overexpression of
apoptotic markers was shown in impression cytology spec-
imens from patients with KCS.3,4 The above findings
support the immunopathogenesis of KCS. However, it is
not clear that this inflammatory reaction represents a
primary phenomenon. Conversely, it may result from
chronic surface dryness and an increased friction between
RIGHTS RESERVED. 593
palpebral and bulbar conjunctiva because of tear defi-
ciency.
Evidence of inflammatory processes in the pathogenesis
of KCS led to the development of cyclosporine A (CsA) as
a first attempt to treat this condition therapeutically. The
proposed mechanisms of CsA are twofold: immunomodu-
lation and anti-inflammation. Immunomodulatory activity
of CsA is achieved by its selective inhibition of the signal
transduction cascade of inflammatory cytokines such as
interleukin 2 and an eventual prevention of the autoim-
mune response.5 The anti-inflammatory effect of CsA is
through its inhibition of phosphatases.6 Topical CsA
treatment of dry eye patients has been reported to be
clinically effective7,8 and to reduce the numbers of acti-
vated lymphocytes and immune-related markers within the
conjunctiva.9,10 The efficacy of topical CsA in dry eye
patients suggested that other immunosuppressive or anti-
inflammatory drops would have a similar or greater bene-
ficial effect in KCS patients.
The aims of this study are twofold: (1) to investigate
whether conjunctival inflammation represents a primary
event in the pathogenesis of KCS or whether it is a
secondary inflammatory reaction caused by enhanced me-
chanical irritation as a result of surface dryness; and (2) to
investigate whether other anti-inflammatory drops (corti-
costeroids and nonsteroidal anti-inflammatory) have sim-
ilar therapeutic effects.
METHODS
THE STUDY WAS COMMENCED AS A SINGLE-SITE, PROSPEC-
tive, randomized, and single-masked clinical trial. In-
formed consent was obtained from all patients, and the
research was begun after obtaining approval from the
Institutional Review Board of the Louisiana State Univer-
sity Health Sciences Center. The research was carried out
according to the tenets of the Declaration of Helsinki. All
study medications were dispensed in coded bottles. The
examiner (A.M.A.) was masked as to the medication used
by the patients. Thirty-two KCS patients with or without
Sjögren syndrome were included in the study. All patients
were at least 21 years of age. Inclusion criteria for patients
included Schirmer test (without anesthesia) of 7 mm in 5
minutes or less in at least one eye; mild superficial punctate
keratitis defined as a corneal punctate fluorescein score of
⫹1 in either eye (scale 0 [none]–3 [severe]); and one or
more moderate dry eye related symptom including itching,
burning, blurred vision, foreign body sensation, dryness,
photophobia, and soreness or pain.
Patients were excluded from the study if they had eye
injury, infection, nondry eye ocular inflammation, trauma,
or surgery within the previous 6 months; received concur-
rent treatment that could interfere with the interpretation
of the study results (systemic corticosteroids, immunosup-
pressive therapy, and so on); had an uncontrolled disease
594 AMERICAN JOURNAL
or significant illness; or were pregnant or lactating. Post-
menopausal patients who were on hormonal replacement
therapy were also excluded. Before initialization of the
study, the patients were instructed not to use any topical or
systemic medication for at least 1 week.
On day 0, Schirmer test, tear breakup time, fluorescein
and rose bengal staining examinations were performed on
both eyes of all patients. The Schirmer test was performed
after corneal staining, because it may affect the staining
pattern of the cornea with either fluorescein or rose bengal.
We used the previously described scoring system for rose
bengal and fluorescein staining.11 We also obtained symp-
tom severity scores from patients12 who were instructed to
grade their symptoms averaging the symptom severities for
both eyes. The results of the Schirmer test and rose bengal
and fluorescein staining scores were evaluated separately
from right and left eyes.
Impression cytology specimens were obtained from the
right eyes of all patients on day 0, day 15, and day 30 to
avoid statistical comparison between different eyes in
different periods. For this purpose, Whatman nitrocellulose
filter paper (cat no: 7195004, Whatman Int. Ltd., Maid-
stone, UK) was used. The paper was cut into strips of
approximately 6 ⫻ 15 mm, which were pressed against the
temporal bulbar conjunctiva for 5 seconds and removed.
Patients were randomized to three groups according to a
computerized list generated by the LSU Eye Center bio-
statistician. Group 1 patients were treated with a preser-
vative-free topical artificial tear substitute (ATS; Refresh,
Allergan Inc., Irvine, California, USA) four times a day in
both eyes. Group 1 patients did not receive any medication
besides Refresh. Groups 2 patients were treated with
topical nonsteroidal anti-inflammatory drug (NSAID) eye
drops, flurbiprofen (Ocufen, Allergan Inc. Irvine, Califor-
nia, USA) four times a day plus ATS four to eight times a
day in both eyes. Group 3 patients were treated with
topical corticosteroidal drops (TSD) FML (Allergan Inc.,
Irvine, California, USA) four times a day plus ATS four to
eight times a day in both eyes. The patients were in-
structed to discuss their medications only with the study
coordinator and not with the examiner.
The impression cytology sample strips were cut into
three pieces: one piece was used for periodic acid–Schiff
(PAS) staining; one was used for HLA-DR (Monoclonal
Mouse Anti Human HLA-DR, Alpha-Chain Clone TAL
1B5 Code No. M0746 Lot068 DAKO); and the other for
Apo 2.7 (Monoclonal Antibody Apo 2.7 Cat. No: 2087
Immunotech) staining. The strip samples were placed
separately cell side down on gelatin-coated slides. The
slides, with adherent nitrocellulose strips, were air dried
thoroughly at room temperature, placed in 100% metha-
nol, and washed repeatedly with methanol until the
nitrocellulose totally dissolved. The slides were then
wrapped in waxed papers and mailed to the laboratory in
suitable containers. The immunohistochemistry tech-
niques reported by Bales and Durfeen13 were used.
OF OPHTHALMOLOGY OCTOBER 2003

FIGURE 1. An impression cytology specimen from a patient
with pretreatment period. The monoclonal antibody (anti hu-
man HLA-DR) stains cell membranes in a granular manner, as
evidenced by this photomicrograph. Arrows indicate the cells
stained positively (bar ⴝ 25 ␮).
Negative and positive control slides were run with each
sample. The staining pattern of goblet cells of colon
mucous epithelium was used as control to PAS staining.
Similarly, the staining pattern of human tonsil specimen
was used as control to HLA-DR and the staining pattern of
the basal cell layers of squamous epithelium was used as
control to Apo 2.7 staining. The slides were viewed and
photographed using a Zeiss Photomicroscope III (Zeiss,
Oberkochen, Germany). To standardize data collection,
four fields on each sample were viewed using a 25⫻
objective and photographed. Cell counts were carried out
on coded slides to avoid bias. In masked fashion, two
examiners quantified the pattern of staining photographi-
cally (Figures 1 and 2). Percentages of reactive cells were
determined by counting at least 200 cells in each speci-
men. Damaged cells were not considered.
Statistical analyses were performed using SPSS for
Windows (Version 6.0 and 10.0, SPSS Inc., Chicago,
Illinois, USA). The first comparison between and within
groups on different examination days was analyzed using a
two-way analysis of variance (ANOVA) test. Examination
days and groups were chosen as independent variables or
factors, and the other parameters were chosen as depen-
dent variables. Within groups, changes from baseline were
evaluated with a one-way ANOVA test (day was the
independent variable) using the Tukey post-hoc test. The
comparisons between groups on different examination
points were also analyzed with a one-way ANOVA test
using the Tukey post-hoc test (group was an independent
variable). P ⫽ .05 was considered significant for all effects.
The null hypothesis was that there was no difference
among the treatment groups with regard to changes from
baseline values. The alternate hypothesis was that there
was a change.
VOL. 136, NO. 4 DRY EYE TREATMENT WITH CORTICOSTEROIDS
FIGURE 2. An impression cytology specimen stained with
antihuman Apo 2.7 after treatment with nonsteroidal anti-
inflammatory drops plus artificial tear substitute. Apo 2.7 is a
specific apoptosis marker, and anti-Apo 2.7 antibody stains cell
membranes of reactive cells. Arrows indicate positive cells (bar
ⴝ 25 ␮).
Power was calculated to detect an among-group differ-
ence in change from baseline in symptom severity scores
on day 30. The power of the study was calculated accord-
ing to the given formula.14 For a sample size of 11 patients,
the power to calculate 1 grade difference was 0.64.
RESULTS
A TOTAL OF 32 PATIENTS WERE ENROLLED IN THE STUDY.
Four patients were discontinued for administrative reasons.
Eight patients in group 1, nine patients in group 2, and 11
patients in group 3 concluded the whole study period.
None of the above discontinued patients reported adverse
effects that could be related to the medications used in this
study.
Group 1 contained five female and three male patients
(mean age, 51.2 ⫾ 12.4 SD). Five female and four male
patients constituted group 2 (mean age, 46.67 ⫾ 8.66 SD).
Group 3 comprised seven female and four male patients
(mean age, 57.6 ⫾ 12.4 SD).
At the beginning of the study (day 0), no significant
difference was detected between groups in terms of any
parameters studied.
We did not observe any complication that could be
linked to the study medications during treatment period.
Significant differences between groups, days, and mean
effect were observed in terms of symptom severity scores
(mean effect, P ⫽ .01; groups, P ⫽ .01; days, P ⫽ .09;
two-way ANOVA). Group 3 patients had significantly less
symptom severity scores both on days 15 and 30 compared
with groups 1 and 2 patients (P ⫽ .02 and P ⫽ .03 for day
15 comparisons, and P ⫽ .03 and P ⫽ .03 for day 30
comparisons). When comparisons were made between the
AND NSAID 595
FIGURE 3. Change from baseline in symptom severity scores. *Significantly different from baseline value (all P ⴝ .000 both for
days 15 and 30; P ⴝ .003 on day 15 for ATSⴙTSD treated group, and P ⴝ .002 on day 30 of ATSⴙTSD treated group). ATS ⴝ
artificial tear substitute; NSAIDs ⴝ nonsteroidal anti-inflammatory drops; TSD ⴝ topical corticosteroidal drops.

treatment days in particular groups, no significant differ-
ence was found in groups 1 and 2 patients. However, a
significantly less symptom severity score was detected on
day 30 compared with day 0 in group 3 patients (P ⫽ .02).
Comparison between days 0 and 15 also gave a significant
result (P ⫽ .03). Other comparisons within and between
groups did not differ significantly (Figure 3).
Similarly, when rose bengal staining of right eyes was
compared between groups, significant differences were
obtained in terms of main effect (P ⫽ .03) and group
comparison (P ⫽ .04), but no significant difference was
observed in terms of day comparison (P ⫽ .07) (two-way
ANOVA). Although there was no significant difference
observed between groups on day 15, group 3 patients had
significantly lower rose bengal staining scores compared
with group 2 patients on day 30 (P ⫽ .046), but compar-
ison between groups 2 and 3 did not show a significant
difference. In group 3 patients, the mean rose bengal
staining score on day 30 was significantly lower than that
of day 0 (P ⫽ .01), and the mean score on day 15 was also
lower than that on day 0 (P ⫽ .02). However, comparison
between days 15 and 30 was not significantly different.
There was a significant difference between days 0 and 15 in
group 2 patients (P ⫽ .007). However, other comparisons
within and between groups did not give any significant
difference (Figure 4). Left eye comparisons between groups
gave similar results.
Comparison between groups in terms of fluorescein
staining patterns of right eyes was significantly different in
terms of groups (P ⫽ .019), but no significant difference
was observed in terms of days (P ⫽ .074) or main effect (P
⫽ .092). Group 3 patients had a significantly lower
fluorescein staining score compared with group 2 patients
on day 30 (P ⫽ .017). In group 2 patients, the fluorescein
staining score was significantly lower on day 15 compared
with day 0 (P ⫽ .017). Similarly, comparisons between
days 0 to 15 and 0 to 30 gave significant results in group 3
patients (P ⫽ .018 and 0.016, respectively). Other com-
parisons within and between groups did not differ signifi-
cantly (Figure 5). Comparison between left eyes among
fluorescein staining pattern gave similar results.
In impression cytology examinations, comparison be-
tween groups in terms of PAS staining gave a significant
result among group comparison (P ⫽ .003) and main effect

596 AMERICAN JOURNAL OF OPHTHALMOLOGY OCTOBER 2003

FIGURE 4. Change from baseline in rose bengal staining of right eye. *Significantly different from baseline value (P ⴝ .007 on day
15 of ATSⴙNSAIDs group; P ⴝ .02 on day 15 of ATSⴙTSD group; P ⴝ .01 on day 30 of ATSⴙTSD group). ATS ⴝ artificial
tear substitute; NSAIDs ⴝ nonsteroidal anti-inflammatory drops; TSD ⴝ topical corticosteroidal drops.

(P ⫽ .012), but no significant difference could be observed
between day comparisons. On day 15, group 3 patients had
a significantly greater number of PAS⫹ cells compared
with both groups 1 and 2 patients (P ⫽ .034 and P ⫽ .028,
respectively). Similar results were obtained on day 30
comparisons. On day 30, the mean percentage of PAS⫹
cells in group 3 patients was significantly higher than in
both groups 1 and 2 patients (P ⫽ .000 and P ⫽ .001,
respectively). In group 3 patients, the mean percentage of
PAS⫹ cell was significantly higher on day 30 than both on
days 0 and 15 patients (P ⫽ .01, P ⫽ .02). Other
comparisons within and between groups were not signifi-
cant (Figure 6).
Human leukocyte antigen (HLA) staining comparison
between groups also revealed significant differences in terms
of group comparison (P ⫽ .046), but no significant difference
could be detected in terms of either days or main effect. On
day 15, group 3 patients had significantly lower HLA⫹ cells
compared with both groups 1 and 2 patients (P ⫽ .032 and P
⫽ .042, respectively). On day 30, group 3 patients had a
significantly less HLA-DR⫹ cells compared with groups 1
and 2 (P ⫽ .024 and P ⫽ .033, respectively). In group 3
patients, mean HLA-DR⫹ cells on day 30 were significantly
lower than that in both groups 1 and 2 on days 0 and 15 (P
⫽ .042 and P ⫽ .038, respectively). Other comparisons
within and between groups did not reveal any significant
difference (Figure 7).
No significant differences at any examination point (Figure
8) were found with Schirmer test values and the percentage
of Apo 2.7⫹ cells in impression cytology specimens.
DISCUSSION
THE MOST IMPORTANT RESULTS OF THIS STUDY WERE
that treatment with TSD significantly improved the ocular

VOL. 136, NO. 4 DRY EYE TREATMENT WITH CORTICOSTEROIDS AND NSAID 597
FIGURE 5. Change from baseline in fluorescein staining of right eye. *Significantly different from baseline value (P ⴝ .017 on day
15 of ATSⴙNSAIDs group; P ⴝ .018 on day 15 of ATSⴙTSD group; P ⴝ .016 on day 30 of ATSⴙTSD group). ATS ⴝ artificial
tear substitute; NSAIDs ⴝ nonsteroidal anti-inflammatory drops; TSD ⴝ topical corticosteroidal drops.

signs and symptoms of moderate-to-severe dry eye patients,
and these improvements were associated with reduction of
HLA-DR⫹ cells and an increase of PAS⫹ cells in con-
junctival impression cytology specimens. However, ATS
alone or ATS plus NSAID did not change these parame-
ters. Keratoconjunctivitis sicca is an autoimmune disease
that involves not only the lacrimal gland but also the
whole ocular surface. Increased conjunctival inflammation
has been reported before in KCS patients.1,15 Brignole and
coworkers found that conjunctival cells from patients with
dry eye with moderate-to-severe KCS, with or without
Sjögren Syndrome, overexpressed inflammatory markers.16
However, whether inflammation is a primary phenomenon
in KCS or is the consequence of repetitive abrasion of the
corneal surface after tear film deficiency is not clear. It
causes chronic inflammation, lymphocytic infiltrates, and
apoptosis of ocular epithelial cells that could hypotheti-
cally result from chronic ocular surface dryness and epi-
thelial cell degeneration caused by increased friction
between ocular surfaces.10 Our results presented here are
somewhat contradictory to this hypothesis. This study
provides evidence that conjunctival inflammation plays a
primary role in the pathogenesis of KCS, because treat-
ment with ATS and ATS plus NSAID drops did not
change HLA-DR expression in conjunctival cells, but
treatment with corticosteroids significantly reduced these
markers in conjunctival epithelial cells.
Our study provides evidence that TSD treatment im-
proved clinical signs and symptoms, but topical ATS plus
NSAID and ATS alone had no effect on these subjective
and objective clinical parameters. The beneficial effect of
corticosteroidal drops in the treatment of KCS has been
reported before. Marsh and Pflugfelder17 treated 21 severe
KCS patients with Sjögren syndrome with nonpreserved

598 AMERICAN JOURNAL OF OPHTHALMOLOGY OCTOBER 2003

FIGURE 6. Change from baseline in PASⴙ cells. *Significantly different from baseline value (P ⴝ .01). ATS ⴝ artificial tear
substitute; NSAID ⴝ nonsteroidal anti-inflammatory drops; PAS ⴝ periodic acid–Schiff; TSD ⴝ topical corticosteroidal drops.

topical methylprednisolone and observed impressive im-
provement in clinical parameters. However, they advo-
cated using this treatment as “pulse therapy” because they
observed significant side effects of corticosteroids after
long-term therapy. We did not observe any TSD-related
complication during the study period, but using corticoste-
roids was relatively weak compared with methypred-
nisolone, and the duration of the therapy was short. Thus,
it is possible to see corticosteroidal side effects in the
longer therapy with TSD. Sainz de la Maza Serra and his
coworkers treated 15 severe KCS patients with unpre-
served TSD before punctual occlusion and compared the
clinical results with the patients who did not receive any
pretreatment before punctual occlusion.18 They reported
that 2 weeks of therapy with unpreserved TSD significantly
improved clinical parameters in severe KCS. Treating the
patients with ocular surface inflammation with nonpre-
served methylprednisolone drops was reported to avert
delayed tear clearance.19 In our study, the clinical im-
provement observed after TSD plus ATS treatment was
associated with reduction of the number of HLA-DR⫹
cells and an increase of goblet cell numbers in the
impression cytology specimens. HLA-DR is a major im-
mune-related marker normally expressed by immunocom-
ponent cells, that has been shown to be upregulated in
epithelial cells in cases of autoimmune and inflammatory
disorders, and in KCS, conjunctival cells overexpress this
marker.15 Part of the beneficial effect of TSD on dry eye
patients might be a result of the reduction in HLA-DR⫹
conjunctival cells. The decreasing effect of corticosteroids
on HLA-DR expression of human epithelial cells has been
reported before.16 The mechanism of this reduction may
involve several possibilities. As HLA-DR expression in
epithelial cells is stimulated by various cytokines, such as
interferon-␥ (I-␥) and tumor necrosis factor-␣ (TNF-␣),
which could be synthesized in the ocular surface and by
infiltrating lymphocytes, TSD might downregulate
HLA-DR expression by decreasing these cytokine produc-
tions. The reductive effects of corticosteroids on TNF-␣
and I-␥ are well known.17 Conversely, because corticoste-
roids decrease intercellular adhesion molecule-1
(ICAM-1) activation,18,19 TSD may display its beneficial
effect on the ICAM-1 system. ICAM-1 plays an important
role in cell-to-cell interactions and cell migration of

VOL. 136, NO. 4 DRY EYE TREATMENT WITH CORTICOSTEROIDS AND NSAID 599
FIGURE 7. Change from baseline in HLA-DRⴙ cells. *Significantly different from baseline value (P ⴝ .042). ATS ⴝ artificial
tear substitute; NSAID ⴝ nonsteroidal anti-inflammatory drops; TSD ⴝ topical corticosteroidal drops.
lymphocytes into the surrounding tissues such as the
conujunctival epithelium and substantia propria.20 We
failed to find any effect of NSAID on HLA-DR expression
of the conjunctival cells. However, this finding did not
seem to be surprising, especially considering the above
possible mechanisms of effect, because NSAID have no
effect on cytokine production.21 The decrease of HLA-
DR⫹ cells in conjunctival epithelial cells of dry eye
patients (Sjögren and non-Sjögren) had been reported
before after topical CsA treatment.9,10 Fukagawa and
associates speculated that chronic stimulation from envi-
ronmental challenges (contact lens, low humidity, wind,
and so on) on the ocular surface interferes with the
neuronal transmission to the lacrimal glands, resulting in
the activation of vigilant trafficking lymphocytes and sub-
sequent production of proinflammatory cytokines, promoting
the autoimmune response.22 Therefore, CsA and TSD ther-
apy are rational approaches for treating immune-based in-
flammation in KCS.
Another important finding of the current study is an
increase of goblet cell numbers in impression cytology
600 AMERICAN JOURNAL
specimens of dry eye patients after treatment with TSD, as
evidenced by a significant increase of PAS⫹ cells. Such an
effect was demonstrated previously following treatment of
non-Sjögren syndrome patients with CsA ophthalmic
emulsion.23 The authors speculated that reducing inflam-
mation with topical CsA treatment might have a prolifer-
ating effect on goblet cells. It is logical to think the same
mechanism may be valid for TSD treatment.
An interesting result was obtained in terms of apoptotic
cell numbers that were investigated by Apo 2.7 expression
in impression cytology specimens. Topical TSD treatment
had no effect on the numbers of apoptotic cells. The same
result has been reported before with topical CsA treat-
ment.10 The authors speculated that increased apoptosis
after CsA treatment might represent a very early normal-
ization of epithelial cell differentiation from a severely
affected state. Although this mechanism may operate for
TSD treatment, another explanation is also present. Both
corticosteroids and CsA have an inducing effect on apo-
ptosis.24,25 Thus, an expected decrease in apoptosis after
treatment with either CsA or TSD (because of reduction
OF OPHTHALMOLOGY OCTOBER 2003
FIGURE 8. Change from baseline in Apo 2.7ⴙ cells. Any significant difference is not present in any group. ATS ⴝ artificial tear
substitute; NSAIDs ⴝ nonsteroidal anti-inflammatory drops; TSD ⴝ topical corticosteroidal drops.
of immune-based inflammation in conjunctiva of dry eye
patients) and direct stimulation of apoptosis by the drugs
may negate each other.
Commercial TSD and NSAID used in the study (Oc-
ufen and FML) contain mercurial preservatives that could
have confounded surface inflammation; however, this fac-
tor was unlikely to affect the comparison between two
groups because both drops contain preservatives.
The results of the study implied that TSDs were more
effective than topical NSAIDs or ATS in reducing the
ocular surface inflammation in KCS patients. Topical
steroids had a clear beneficial effect both on the subjective
and objective clinical parameters of moderate-to-severe
dry eye patients. These effects were associated with the
reduction of inflammation markers of conjunctival epithe-
lial cells. Based on the data provided, it may be speculated
that conjunctival inflammation is a primary event in the
pathogenesis of KCS rather than a secondary finding,
because the decreasing surface friction by ATS did not
provide any beneficial effect. We think that TSD might be
a good alternative to topical CsA therapy in the treatment
of dry eye disease. A comparison of the two drugs would
clarify this hypothesis.
VOL. 136, NO. 4 DRY EYE TREATMENT WITH CORTICOSTEROIDS
REFERENCES
1. Baudouin C, Brignole F, Becquet F, Pisella PJ, Goguel A.
Flow cytometry in impression cytology specimens: a new
method for evaluation of conjunctival inflammation. Invest
Ophthalmol Vis Sci 1997;8:1458 –1464.
2. Sayama K, Yonehara S, Watanabe Y, Miki Y. Expression of
Fas antigen on keratocytes in vivo and induction of apoptosis
in cultured keratinocytes. J Dermatol 1994;103:330 –334.
3. Brignole F, de Saint-Jean M, Goldschild M, Becquet F,
Goguel A, Baudouin C. Expression of Fas-Fas ligand antigens
and apoptotic marker APO 2.7 by the human conjunctival
epithelium: positive correlation with class II HLA DR
expression in inflammatory ocular surface disorders. Exp Eye
Res 1998;67:679 –687.
4. Boucier T, de Saint-Jean M, Brignole F, Goguel A, Baodouin
C. Expression of CD 40 and CD 40 ligand in the human
conjunctival epithelium. Invest Ophthalmol Vis Sci 2000;
41:120 –126.
5. Ryffel B. Current aspects in organ transplantations. Schweiz
Rundsch Med Prax 1989;78:997–1002.
6. Florio T, Perrino PA, Stock PJ. Cyclic 3.5 adenosine mono-
phosphate and cyclosporine A inhibit cellular proliferation
and serine/threonine protein phosphatase activity in putu-
itary cells. Endocrinology 1996;137:4409 –4418.
7. Stevenson D, Tauber J, Reis BL. Efficacy and safety of
cyclosporine A ophthalmic emulsion in the treatment of
AND NSAID 601
 moderate-to-severe dry eye disease. Ophthalmology 2000;
107:967–974.
8. Sall K, Stevenson OD, Mundorf TK, Reis BL. Two multi-
center, randomized studies of the efficacy and safety of
cyclosporine ophthalmic emulsion in moderate to severe dry
eye disease. CsA phase 3 study group. Ophthalmology
2000;107:631–639.
9. Kunert KS, Tisdale AS, Stern ME, Smith JA, Gipson IK.
Analysis of topical cyclosporine treatment of patients with
dry eye syndrome: effect on conjunctival lymphocytes. Arch
Ophthalmol 2000;118:1489 –1496.
10. Brignole F, Pisella P-J, de Saint Jean M, Goldschild M,
Baudouin C. Flow cytometric analysis of inflammatory mark-
ers in KCS: 6-month treatment with topical cyclosporin A.
Invest Ophthalmol Vis Sci 2001;42:90 –95.
11. Lemp MA. Report of NEI/Industry workshop on clinical
trials in dry eyes. CLAO J 1995;21:221–232.
12. Oden NL, Lilienfeld DE, Lemp MA, Nelson JD, Ederer F.
Sensitivity and specificity of a screening questionnaire for dry
eye. Adv Exp Med Biol 1998;438:807–820.
13. Bales CE, Durfeen GR. Cytologic techniques. In: Koos LG,
editor. Diagnostic cytology and its histopathologic bases,
Vol. 2, 4th ed. Philadelphia: Lippincott, 1992:1–10.
14. Dawson-Saunders B, Trapp RG. Basic and clinical biostatis-
tics. Upper Saddle River, New Jersey: Prentice Hall, 1994:
119 –120.
15. Baudouin C, Haoutat N, Brignole F, Bayle J, Gastaut P.
Imunopathological findings in conjunctival cells using im-
munofluoresence staining of impression cytology specimens.
Br J Ophthalmol 1992;76:545–549.
16. Brignole F, Pisella PJ, Goldschild M, De Saint Jean M,
Goguel A, Baudouin C. Flow cytometric analysis of inflam-
matory markers in conjunctival epithelial cells of patients
602 AMERICAN JOURNAL
with dry eyes. Invest Ophthalmol Vis Sci 2000;41:1356 –
1363.
17. Marsh P, Pflugfelder SC. Topical non-preserved methylpred-
nisolone therapy of keratoconjunctivitis sicca in Sjogren’s
syndrome. Ophthalmology 1999;106:811–816.
18. Sainz de la Maza Serra SM, Simon Castellvi C, Kabbani O.
Nonpreserved topical steroids and punctual occlusion for
severe keratokonjunctivitis sicca. Arch Soc Esp Ophthalmol
2000;75:751–756.
19. Prabhasawat P, Tseng SC. Frequent association of delayed
tear clearance in ocular irritation. Br J Ophthalmol 1998;82:
666 –675.
20. Tsubota K, Fujihara T, Saito K, Takeuchi T. Conjunctival
epithelium expression of HLA-DR in dry eye patients.
Ophthalmologica 1999;213:16 –19.
21. Schwiebert LM, Schleimer RP, Radka SF, Ono SJ. Modula-
tion of MHC class II expression in human cells by dexameth-
asone. Cell Immunol 1995;165:12–19.
22. Fukagawa K, Saito H, Tsubota K, et al. RANTES production
in a conjunctival epithelial cell line. Cornea 1997;16:564 –
570.
23. Penfold PL, Wen L, Madigan MC, King NJ, Provis JM.
Modulation of permeability and adhesion molecule expres-
sion by human choroidal endothelial cells. Invest Ophthal-
mol Vis Sci 2002;43:3125–3130.
24. Tsubota K, Inoue H, Ando K, Ono M, Yoshino K, Saito I.
Adenovirus-mediated gene transfer to the ocular surface
epithelium. Exp Eye Res 1998;67:531–538.
25. Wakefield D, McCluskey P, Palladinetti P. Distribution of
lymphocytes and cell adhesion molecules in iris biopsy
specimens from patients with uveitis. Arch Ophthalmol
1992;110:121–125.
OF OPHTHALMOLOGY OCTOBER 2003