Clinical Ophthalmology
`
`Open Access Full Text Article
`
`Dovepress
`open access to scientific and medical research
`
`O r i g i nA L re s eArCh
`
`Ocular surface disease in patients with glaucoma
`or ocular hypertension treated with either
`BAK-preserved latanoprost or BAK-free travoprost
`
`gregory Katz 1
`Clark L springs 2
`e randy Craven 3
`Michela Montecchi-Palmer4
`1huron Ophthalmology, Ypsilanti, Mi,
`UsA; 2indiana University eye Care,
`indianapolis, in, UsA; 3specialty
`eye Care, Denver, CO, UsA; 4 Alcon
`research Ltd., Fort Worth, TX, UsA
`
`Correspondence: gregory Katz
`5477 W. Clark rd, Ypsilanti,
`Mi 48197, UsA
`Tel +1 734 434 6000
`Fax +1 734 434 7005
`email glbjkatz@hotmail.com
`
`Purpose: The preservative benzalkonium chloride (BAK) may adversely affect ocular surface
`health. This study evaluated symptoms of ocular surface disease (OSD) in patients previously
`treated with a BAK-preserved therapy to lower their intraocular pressure, who either continued
`that therapy or switched to a BAK-free therapy.
`Methods: Eligible adult patients with ocular hypertension or open-angle glaucoma that had
`been controlled with BAK-preserved latanoprost 0.005% monotherapy (Xalatan®) for at least
`one month and had a score of $ 13 (0 = none, 100 = most severe) on the Ocular Surface Disease
`Index (OSDI) questionnaire were entered into this prospective, double-masked, randomized,
`active-controlled, multicenter trial. By random assignment, patients either continued with
`BAK-preserved latanoprost 0.005% or transitioned to BAK-free travoprost 0.004% (Travatan
`Z® ophthalmic solution). OSDI scores were assessed again after six and 12 weeks.
`Results: For the 678 evaluable patients, mean change in OSDI score from baseline to week
`12 favored the travoprost 0.004% BAK-free group, but was not statistically different between
`groups (P = 0.10). When patients with mild OSD at baseline were assessed after 12 weeks,
`the mean OSDI score was significantly lower (P = 0.04) in the BAK-free travoprost 0.004%
`group (score = 11.6 ± 10.8 units) than in the BAK-preserved latanoprost 0.005% group
`(score = 14.4 ± 11.9 units), and a significantly larger percentage (P , 0.01) improved to normal
`OSDI scores in the BAK-free travoprost 0.004% group (62.9% of group) than in the BAK-
`preserved latanoprost 0.005% group (47.0% of group). Patients pretreated with BAK-preserved
`latanoprost 0.005% for .24 months were significantly more likely (P = 0.03) to improve to a
`normal OSDI score after 12 weeks if they were switched to BAK-free travoprost 0.004% (47.9%
`of group) than if they remained on BAK-preserved latanoprost 0.005% (33.9% of group).
`Conclusions: Switching from BAK-preserved latanoprost 0.005% to BAK-free travoprost
`0.004% yielded significant improvements in symptoms of OSD in patients with glaucoma or
`ocular hypertension.
`Keywords: ocular surface, glaucoma, benzalkonium chloride, prostaglandin analog,
`preservative
`
`Introduction
`Most of the currently available topical treatments for elevated intraocular pressure
`(IOP), including latanoprost 0.005%, are preserved with benzalkonium chloride
`(BAK).1 Chronic exposure to BAK-preserved IOP-lowering medications has been
`associated with increased frequency of patient-reported symptoms of ocular dis-
`comfort, including burning, stinging, foreign body sensation, and dry eye sensa-
`tion.2 In vitro, BAK-preserved latanoprost 0.005% and BAK-preserved travoprost
`0.004% are both toxic to ocular cells, whereas BAK-free travoprost 0.004% is not.3
`
`submit your manuscript | www.dovepress.com
`Dovepress
`DOI: 10.2147/OPTH.S14113
`
`Clinical Ophthalmology 2010:4 1253–1261
`© 2010 Katz et al, publisher and licensee Dove Medical Press Ltd. This is an Open Access article
`which permits unrestricted noncommercial use, provided the original work is properly cited.
`
`1253
`
`ALCON 2014
`Apotex Corp. v. Alcon Research, Ltd.
`Case IPR2013-00428
`
`

`

`Katz et al
`
`Dovepress
`
`In animal models, BAK-free travoprost 0.004% did not
`affect goblet cell numbers4 or corneal epithelial cells,5,6
`whereas BAK-preserved latanoprost 0.005% was shown to
`cause losses of goblet cells4 as well as pathologic changes
`in the corneal epithelium.5,6 In humans, chronic exposure
`to BAK-preserved topical IOP-lowering medications was
`associated with signs of adverse effects on the ocular
`surface, including instability of the tear film,7–9 reduced
`density of superficial epithelial cells,7 disruption of cor-
`neal epithelial barrier function,8 and conjunctival inflam-
`mation.9 Adverse reactions induced by BAK-preserved
`medications may be reversible in glaucoma patients who
`are switched to BAK-free medications.2 For these reasons,
`many researchers and clinicians have recommended BAK-
`free IOP-lowering medications.1–9
`BAK-induced changes may manifest as symptomatic
`ocular surface disease (OSD) in medically treated glaucoma
`patients.10,11 OSD is an umbrella term that includes dry
`eye, lid disease, conjunctivitis, and keratitis.12 Although
`OSD is seen in approximately 15% of the general elderly
`population,13 it has been reported to occur in 48% to 59%
`of patients with medically treated glaucoma.10,11 A higher
`incidence10 and severity11 of OSD has been reported in
`patients who received multiple BAK-preserved treatments
`concomitantly than in patients who were treated with only
`one BAK-preserved treatment. Antihypertensive medica-
`tions with alternative preservative systems (other than
`BAK) could help to maintain the long-term ocular surface
`health of patients with glaucoma, and could avoid inducing
`or aggravating OSD.
`In a previous large multicenter clinical trial of patients
`with glaucoma who had been previously treated with either
`latanoprost 0.005% or bimatoprost 0.03%, and who needed
`alternative therapy due to tolerability issues, a switch to
`BAK-free travoprost 0.004% resulted in improvement
`in OSD symptoms that were both clinically and statisti-
`cally significant, and maintained equal or better control of
`IOP.14 However, that study was not conducted in a parallel,
`randomized, masked fashion. The objective of this current
`multicenter, double-masked, randomized, controlled study
`was to quantify changes in symptoms of OSD after random-
`izing patients with open-angle glaucoma or ocular hyperten-
`sion who were previously treated with latanoprost 0.005%
`preserved by 0.02% BAK (Xalatan® ophthalmic solution;
`Pfizer Inc., NY) either to remain on BAK-preserved latano-
`prost 0.005% or to change to BAK-free travoprost 0.004%
`(Travatan Z® ophthalmic solution; Alcon Laboratories, Inc.,
`Fort Worth, TX).
`
`Methods
`This was a prospective, double-masked, randomized, active-
`controlled clinical trial, conducted at 66 clinics in the US.
`The protocol was approved by the appropriate review
`boards at all participating institutions, and the trial was con-
`ducted in accordance with the tenets of the Declaration of
` Helsinki. All participating patients signed a written informed
`consent form.
`
`general entry criteria
`Eligible patients were at least 18 years of age, had ocular
`hypertension or primary open-angle glaucoma (with or
`without pigment dispersion or a pseudoexfoliation com-
`ponent), and had IOP that was adequately controlled on
`latanoprost 0.005% monotherapy for at least one month
`prior to enrollment. Adequate IOP control was deter-
`mined for each patient by the enrolling investigator, and
`was defined as being both stable and safe for that patient.
`General ocular health exclusion criteria included the fol-
`lowing: corneal abnormalities that could prevent accurate
`applanation tonometry; any intraocular surgery or ocular
`trauma within the previous six months; any ocular laser
`surgery within the previous three months; progressive
`retinal or optic nerve disease; severe central visual field
`loss; or visual acuity worse than 0.6 logMAR in either eye.
`Patients were also excluded if they had used any ocular
`medications (other than latanoprost 0.005% or artificial
`tears) within seven days of the screening visit, or if they
`had taken any systemic medication for less than 30 days
`of stable dosing before the screening visit. Women of
`childbearing potential were allowed to participate in the
`trial only if they were not breastfeeding, were not pregnant
`or planning to become pregnant, and were using adequate
`birth control during the study. All patients were required
`to be willing and able to abstain from the use of any other
`topical ophthalmic eye drops, other than assigned study
`medication, for the duration of the study.
`
`Ocular surface entry criteria
`Exclusion criteria related to ocular surface health were
`as follows: OSD that had previously been treated with
`punctal plugs, punctal cautery, topical cyclosporine A, or
`topical corticosteroids; suspected or diagnosed Sjögren’s
`syndrome; prior corneal surgery (including keratorefractive
`surgery) within the previous one year; presence or history
`of clinically significant blepharitis within the previous two
`years; any history of other ocular inflammatory disease
`(eg, rosacea that affected the ocular adnexa or herpes
`
`1254
`
`submit your manuscript | www.dovepress.com
`Dovepress
`
`Clinical Ophthalmology 2010:4
`
`

`

`Dovepress
`
`BAK-preserved latanoprost versus BAK-free travoprost
`
`simplex virus keratitis); seasonal ocular allergies expected
`within the study period; and any contact lens wear or cor-
`ticosteroid use within the 30 days before the screening
`visit.
`All potential patients were screened using the Ocular
`Surface Disease Index (OSDI) questionnaire. The OSDI is a
`validated, self-administered instrument for assessing the pres-
`ence and severity of OSD symptoms.15 The OSDI question-
`naire includes 12 questions about the respondent’s past-week
`experience with ocular symptoms, vision-related functioning,
`and environmental triggers.15,16 Questions assessed whether
`respondents had eyes that felt gritty, painful, sore, or sensitive
`to light; whether they had blurred or poor vision; whether
`they experienced limitations with reading, driving at night,
`watching television, or working with a computer or bank
`machine; and whether their eyes felt uncomfortable in windy
`conditions, in areas with low humidity, or in air-conditioned
`places.16 Response options for each question were “all of the
`time” (score = 4), “most of the time” (score = 3), “half of the
`time” (score = 2), “some of the time” (score = 1), and “none
`of the time” (score = 0).15 Questions about vision-related
`functioning or environmental triggers could also be answered
`with “not applicable”, in which case that question was not
`factored into the final score calculation. The total OSDI score
`was calculated for each patient using the methods described
`by the OSDI originators,15 as follows:
`
`OSDI score =
`
`×
`(sum of scores for allquestions answered) 25
`Total numb
`eer of questions answered
`
`The final total OSDI score could range from 0 to 100, with
`the OSDI scores classified as #12 = normal, 13–22 = mild
`OSD, 23–32 = moderate OSD, and $33 = severe OSD.17 To
`be eligible for inclusion in the study, patients were required
`to have an OSDI score of 13 or higher.
`All potential patients were screened with corneal fluo-
`rescein staining. Corneal fluorescein staining was conducted
`according to each investigator’s standard procedure, using
`their standard staining agent. Each cornea was scored on
`the following scale, which was designed to assess stain-
`ing over the entire corneal surface with no specification of
`corneal regions: 0 = absent (no staining), 1 = mild (a few
`punctate regions of staining, but less than 10% coverage of
`the corneal surface), 2 = moderate (10%–50% coverage of
`the corneal surface), or 3 = severe (more than 50% coverage
`of the corneal surface). To be eligible for the study, patients
`were required to have a corneal fluorescein staining score
`of 1 or higher.
`
`enrollment and masked randomization
`Patients who met all entry criteria and who reported
`using latanoprost 0.005% on the evening prior to the
` screening/enrollment visit were invited to participate. At
`the enrollment visit, the informed consent form was signed,
`comprehensive medical and ophthalmic histories were
`obtained, the OSDI questionnaire was completed, and an
`ocular examination was performed (including visual acuity
`determination, slit-lamp inspection of the anterior segment,
`corneal fluorescein staining, Goldmann tonometry, and
`dilated fundus examination). Women of childbearing poten-
`tial provided urine samples for pregnancy tests.
`The targeted enrollment was approximately 700 patients
`(350 per group) in order to obtain approximately 650 evalu-
`able patients (325 per group). Power calculations (using a
`two-sample t-test with two-sided alpha = 0.05) indicated
`that, with 325 patients per treatment group, the study would
`have at least 90% power to detect a difference between treat-
`ment groups that was .5 units on the OSDI questionnaire,
`assuming a common standard deviation of 19 units in the
`mean change from baseline OSDI score.
`At the completion of the first visit, enrolled patients were
`randomized to an intervention whereby they either continued
`on therapy with BAK-preserved latanoprost 0.005% or were
`transitioned to therapy with BAK-free travoprost 0.004%
`ophthalmic solution. BAK-preserved latanoprost 0.005% was
`the commercially available Xalatan® ophthalmic solution,
`which is preserved using 0.02% BAK. BAK-free travoprost
`0.004% was the commercially available Travatan Z® ophthal-
`mic solution containing the proprietary sofZia® preservative,
`which is an ionic buffer system containing borate, propylene
`glycol, sorbitol, and zinc chloride.
`If both eyes met all of the eligibility criteria, both eyes
`were treated with the same test medication; otherwise, only
`the eligible eye was treated. At the study site, the enrolling
`clinician assigned a number to the patient, and then called
`an interactive voice response system that was hosted by the
`study sponsor in order to receive a kit number. These kit
`numbers had been randomized by the study sponsor using
`statistical software (SAS Institute, Cary, NC). The patient
`received the assigned kit of study medication. Within the
`kits, all medications (whether BAK-preserved latanoprost
`0.005% or BAK-free travoprost 0.004%) were packaged
`in identical oval 4 mL polypropylene dropper bottles. Each
`patient received two bottles of the assigned study medication
`and was instructed to instill one drop of study medication
`in the study eye(s) once daily in the evening. In case of a
`medical emergency that required information about the study
`
`Clinical Ophthalmology 2010:4
`
`submit your manuscript | www.dovepress.com
`Dovepress
`
`1255
`
`

`

`Katz et al
`
`Dovepress
`
`medication, the investigator or a designee could request
`unmasking of the test medication by calling the interactive
`voice response system.
`
`Efficacy and safety assessments
`Patients returned six weeks (42 ± 7 days) after enrollment
`for the second study visit, and 12 weeks (90 ± 7 days) after
`enrollment for the third study visit. Both of these visits
`were scheduled at approximately the same time of day as
`the entry visit for each patient. At both follow-up visits,
`an interval medical history was obtained and any adverse
`events were assessed, the OSDI questionnaire was com-
`pleted, and an ocular examination was conducted, consist-
`ing of visual acuity, slit-lamp anterior segment inspection,
`corneal fluorescein staining, and Goldmann tonometry.
`At the 12-week visit, a dilated fundus examination was
`conducted, and women of childbearing potential provided
`urine samples for pregnancy tests.
`The primary efficacy variable was the mean change
`in OSDI scores between the entry visit and the 12-week
`follow-up visit. A secondary efficacy variable was the per-
`centage of patients with a corneal fluorescein staining score
`of 0. Exploratory efficacy variables that were assessed at
`
`Table 1 Baseline values and demographics for the intent-to-treat
`population
`
`Travoprost
`0.004% BAK-free,
`n = 343
`
`BAK-preserved
`latanopros
`0.005%, n = 335
`
`Overall
`population
`n = 678
`
`131 (38.2%)
`212 (61.8%)
`
`114 (34.0%)
`221 (66.0%)
`
`245 (36.1%)
`433 (63.9%)
`
`105 (31.3%)
`230 (68.5%)
`
`237 (35.0%)
`441 (65.0%)
`
`32 (9.3%)
`
`35 (10.4%)
`
`67 (9.9%)
`
`Age, n (%)
`18–64 years
`$65 years
`Gender, n (%)
`132 (38.5%)
`Male
`211 (61.5%)
`Female
`OSDI category, n (%)
`4 (0.6%)
`2 (0.6%)
`normal*
`2 (0.6%)
`277 (40.9%)
`136 (40.6%)
`Mild
`141 (41.1%)
`176 (26.0%)
`85 (25.4%)
`Moderate
`91 (26.5%)
`221 (32.6%)
`112 (33.4%)
`severe
`109 (31.8%)
`Duration of BAK-preserved latanoprost pretreatment, n (%)
`Total with
`311 (90.7%)
`300 (89.6%)
`611 (90.1%)
`data available†
`Without
`data available
`209 (30.8%)
`100 (29.9%)
`109 (31.8%)
`1–6 months
`165 (24.3%)
`87 (26.0%)
`78 (22.7%)
`6–24 months
`237 (35.0%)
`113 (33.7%)
`124 (36.2%)
`.24 months
`Notes: *These patients were enrolled due to an error in calculating the baseline
`OsDi score; †Of those patients who could recall their start date with BAK-preserved
`latanoprost 0.005%.
`Abbreviations: OsDi, Ocular surface Disease index; BAK, benzalkonium chloride.
`
`the 12-week follow-up visit were the proportion of patients
`who had a normal OSDI score (#12 units), percentage
`of patients with a $10-point improvement from baseline
`OSDI score, OSDI outcomes stratified by duration of pre-
`treatment with BAK-preserved latanoprost 0.005% before
`entering the study, and OSDI outcomes stratified by sever-
`ity of baseline OSDI (mild, moderate, or severe). Safety
`variable assessments included best-corrected visual acuity,
`slit lamp evaluations, IOP, dilated fundus examinations,
`and adverse events.
`
`statistical analysis
`Continuous variables were assessed using a two-sample t-test
`with two-sided α = 0.05, and categoric variables were assessed
`by Chi-square test with α = 0.05. The null hypothesis stated
`that no relationship existed between BAK-preserved treatment
`and change in OSDI score. The alternative hypotheses stated
`that BAK-preserved latanoprost 0.005% had an adverse impact
`on OSD, which might accrue over a longer duration of BAK-
`preserved latanoprost 0.005% treatment, but might be reversible
`to some degree by transition to BAK-free travoprost 0.004%,
`especially in patients with mild OSDI. Unless otherwise speci-
`fied, outcome values are presented as mean ± standard deviation
`in text, and as mean ± standard error in figures.
`
`Results
`Baseline clinical and demographic data
`A total of 724 patients were enrolled, 678 of whom were
`evaluable for the intent-to-treat (ITT) analysis. Four of the
`ITT patients had normal OSDI scores at baseline, which was
`an exclusion criterion, but they received study medication and
`thus were evaluated with the rest of the population. As shown
`in Table 1, the two treatment groups were statistically similar
`(all P . 0.05) in the baseline parameters of gender, age,
`OSDI category, and duration of exposure to BAK-preserved
`latanoprost 0.005% before entry into the study. The first visit
`of the first patient was in July 2008, and the final analysis
`date was in June 2009. Participant flow through the study is
`shown in Figure 1.
`
`Mean change in OsDi scores
`For the patients who had mild OSD at baseline, the mean
`OSDI score at the 12-week time point was significantly lower
`(P = 0.04) in patients randomized to BAK-free travoprost
`0.004% (11.6 ± 10.8 units) than in patients who continued
`on BAK-preserved latanoprost 0.005% (14.4 ± 11.9 units),
`as shown in Figure 2. For the overall cohort of patients with
`all baseline OSDI scores, mean OSDI scores at the 12-week
`
`1256
`
`submit your manuscript | www.dovepress.com
`Dovepress
`
`Clinical Ophthalmology 2010:4
`
`

`

`Dovepress
`
`BAK-preserved latanoprost versus BAK-free travoprost
`
`>12 to 22
`Mild OSD,
`
`*
`
`0 to 12
`Normal,
`
`travoprost 0.004%
`BAK-free
`BAK-preserved
`latanoprost 0.005%
`
`20
`
`15
`
`10
`
`5
`
`0
`
`Mean OSDI score
`
`N = 678 enrolled
`
`Randomization
`
`travoprost
`0.004%
`BAK-free
`group
`
`BAK-preserved
`latanoprost
`0.005%
`group
`
`Total
`n = 343
`allocated
`
`Total
`n = 335
`allocated
`
`10
`
`12
`
`4 T
`
`6
`8
`ime, weeks
`
`0
`
`2
`
`n = 21 discontinued, for
`reasons that included
`adverse events (n = 12),
`protocol violation (n = 6),
`lost to follow-up (n = 1),
`other (n = 1), or
`noncompliance (n = 1)
`
`n = 35 discontinued, for
`reasons that included
`adverse events (n = 10),
`protocol violation (n = 9),
`lost to follow-up (n = 7),
`other (n = 3), or
`noncompliance (n = 2)
`
`Figure 2 Mean scores on the OsDi questionnaire for the patients who had mild
`OsD at baseline. error bars represent standard error of the mean. *P , 0.05. in the
`BAK-free travoprost group, patient numbers were 141 at baseline, 135 at week 6,
`and 140 at week 12. in the BAK-preserved latanoprost group, patient numbers were
`136 at baseline, 134 at week 6, and 132 at week 12.
`Abbreviations: BAK, benzalkonium chloride: OsDi, Ocular surface Disease index;
`OsD, ocular surface disease.
`
`0.004% group and −11.4 ± 17.4 for the 328 patients in the
`BAK-preserved latanoprost 0.005% group).
`
`Patients improving
`to normal OsDi scores
`The percentage of patients who had mild OSDI scores at base-
`line and who improved to normal OSDI scores after 12 weeks
`was significantly larger (P , 0.01) in the BAK-free travoprost
`
`Week 6
`
`Week 12
`
`n = 134 n = 135
`
`n = 132 n = 140
`
`P = 0.92
`
`P = 0.10
`
`BAK-preserved
`latanoprost 0.005%
`
`travoprost 0.004%
`BAK-free
`
`0
`
`–1
`
`–2
`
`–3
`
`–4
`
`–5
`
`–6
`
`–7
`
`Mean change from baseline mild OSDI score
`
`Figure 3 Mean change from baseline scores on the OsDi questionnaire for the
`patients who had mild ocular surface disease at baseline. error bars represent
`standard error of the mean.
`Abbreviations: BAK, benzalkonium chloride; OsDi, Ocular surface Disease index.
`
`n = 333 analyzed
`after 6 weeks
`
`n = 328 analyzed
`after 6 weeks
`
`n = 339 analyzed
`after 12 weeks
`
`n = 328 analyzed
`after 12 weeks
`
`Figure 1 Participant flow through the study. When possible, patients who
`discontinued treatment were analyzed before exiting the study, so discontinuation
`and analysis numbers are not mutually exclusive.
`Abbreviation: BAK, benzalkonium chloride.
`
`time point were not statistically different between the groups,
`ie, 18.4 ± 16.0 for 339 patients in the BAK-free travoprost
`0.004% group, and 19.4 ± 15.3 for 328 patients in the BAK-
`preserved latanoprost 0.005% group.
`When normalized to baseline values, the mean change
`in OSDI score from the entry visit to the 12-week follow-up
`visit was not significantly larger (P = 0.10) for the patients
`with mild OSD at baseline who were randomized to BAK-free
` travoprost 0.004% (−5.0 ± 10.8 units, n = 140) than for patients
`with mild OSD at baseline who continued on BAK-preserved
`latanoprost 0.005% (−2.7 ± 12.1 units, n = 132), as shown in
`Figure 3. The mean change from baseline mild OSDI score to
`score at week 12 was statistically different, from zero change
`in both treatment groups (P = 0.01 in the BAK-preserved
`latanoprost 0.005% group and P , 0.0001 in the BAK-free
`travoprost 0.004% group). Mean change from baseline OSDI
`scores in the “baseline–moderate” and “baseline–severe”
`groups were not statistically different between the treatment
`groups. For the overall cohort of patients with all baseline
`OSDI scores, mean changes in OSDI scores at the 12-week
`time point were not statistically different between groups
`(−11.3 ± 17.2 for the 339 patients in the BAK-free travoprost
`
`Clinical Ophthalmology 2010:4
`
`submit your manuscript | www.dovepress.com
`Dovepress
`
`1257
`
`

`

`Katz et al
`
`Dovepress
`
`0.004% group (62.9% of group, 88 of 140) than in the BAK-
`preserved latanoprost 0.005% group (47.0% of group, 62
`of 132), as shown in Figure 4. Percentages of those patients
`who had started at “baseline–moderate” or “baseline–severe”
`OSDI scores and who improved to normal OSDI scores at
`12 weeks were not significantly different between the treat-
`ment groups. For the overall cohort of patients with all base-
`line OSDI scores, the percentage of patients who improved
`to normal OSDI scores after 12 weeks was not significantly
`different between groups, ie, 41.6% of patients (141 of 339)
`in the BAK-free travoprost 0.004% group, and 38.4% of
`patients (126 of 328) in the BAK-preserved latanoprost
`0.005% group.
`
`Outcomes stratified by duration
`of pretreatment
`Of the evaluable population, 611 patients could recall
`how long they had been pretreated with BAK-preserved
`latanoprost before entering the study. Prior exposure to
`BAK-preserved latanoprost was stratified as .24 months in
`35.0% of patients (237 of 678), 6–24 months in 24.3% of
`patients (165 of 678), 1–6 months in 30.8% of patients (209
`of 678), and of unknown duration in 9.9% of patients (67
`of 678), as shown in Table 1. Regardless of baseline OSDI
`score, patients who were pretreated with BAK-preserved
`latanoprost 0.005% for .24 months before entering the
`
`study were significantly more likely (P = 0.03) to improve to
`a normal OSDI score after 12 weeks if they were switched to
`BAK-free travoprost 0.004% (47.9% of patients, 58 of 121)
`than if they remained on BAK-preserved latanoprost 0.005%
`(33.9% of patients, 37 of 109), as shown in Figure 5. The
`percentage of patients improving to a normal OSDI score
`was not significantly different between the treatment groups
`for patients who were exposed to BAK-preserved latanoprost
`0.005% for 1–6 months or for 6–24 months prior to entry
`into the study.
`
`Patients with $10 point
`improvement in OsDi scores
`The percentage of patients who improved $10 points in
`OSDI scores from baseline to week 12 was not statistically
`different between the treatment groups for the overall cohort,
`or for the subgroups of patients with mild, moderate, or
`severe baseline OSDI scores. For the overall population,
`an improvement of $10 points was observed in 53.4% of
`patients (181 of 339) in the BAK-free travoprost 0.004%
`group and in 51.8% of patients (170 of 328) in the BAK-
`preserved latanoprost 0.005% group.
`
`Absence of corneal staining
`The percentage of patients without corneal staining at week
`12 was not statistically different between treatment groups
`
`BAK-preserved
`latanoprost 0.005%
`
`travoprost 0.004%
`BAK-free
`
`P = 0.86
`
`36.4% 37.5%
`
`*P = 0.03
`
`47.9%
`
`33.9%
`
`40 of
`110
`
`45 of
`120
`
`6 weeks
`
`37 of
`109
`
`58 of
`121
`
`12 weeks
`
`60
`
`50
`
`40
`
`30
`
`20
`
`10
`
`0
`
`Percentage with normal OSDI scores, %
`
`Figure 5 Percentage of patients who had been exposed to BAK-preserved
`latanoprost for .24 months prior to entry in the study, and who improved to
`normal scores on the OsDi 12 weeks after switching to BAK-free travoprost.
`Abbreviations: BAK, benzalkonium chloride; OsDi, Ocular surface Disease index.
`
`BAK-preserved
`latanoprost 0.005%
`
`travoprost 0.004%
`BAK-free
`
`*P < 0.01
`
`62.9%
`
`P = 0.50
`
`50.7%
`
`46.7%
`
`47.0%
`
`68 of
`134
`
`63 of
`135
`
`6 weeks
`
`62 of
`132
`
`88 of
`140
`
`12 weeks
`
`80
`
`70
`
`60
`
`50
`
`40
`
`30
`
`20
`
`10
`
`0
`
`Percentage with normal OSDI scores, %
`
`Figure 4 Percentage of patients who had baseline mild-severity scores on the OsDi
`and who improved to normal OsDi scores.
`Abbreviations: BAK, benzalkonium chloride OsDi, Ocular surface Disease index.
`
`1258
`
`submit your manuscript | www.dovepress.com
`Dovepress
`
`Clinical Ophthalmology 2010:4
`
`

`

`Dovepress
`
`BAK-preserved latanoprost versus BAK-free travoprost
`
`for the overall cohort, or for the subgroups of patients with
`mild, moderate, or severe baseline OSDI scores. For the
`overall population, absence of corneal staining at week 12 was
`observed in 37.1% of patients (127 of 342) in the BAK-free
`travoprost 0.004% group and 40.0% of patients (132 of 330)
`in the BAK-preserved latanoprost 0.005% group.
`
`safety assessments
`No statistical differences in safety parameters were observed
`between treatment groups. The entire enrolled population, not
`just the ITT population, was assessed for safety parameters.
`The most frequently reported treatment-emergent ocular
`adverse events in both treatment groups were eye irritation,
`hyperemia of the eye, eye pain, eye pruritus, and foreign body
`sensation in the eye, as shown in Table 2.
`
`Discussion
`In this randomized, controlled, multicenter, 12-week study
`of 678 glaucoma patients who had been treated with BAK-
`preserved latanoprost 0.005% for at least one month prior
`to entry to the study, transitioning to BAK-free travoprost
`0.004% produced significant improvements in symptoms
`of OSD for patients who had mild OSDI scores at baseline
`and for patients who had been exposed to BAK-preserved
`latanoprost 0.005% for more than 24 months prior to entry
`into the study.
`For patients with mild OSDI scores at baseline, the mean
`improvement was −5.0 units on the OSDI questionnaire
`for patients who were transitioned to BAK-free travoprost
`0.004%. This value (5.0 units) is clinically relevant, because
`it has been established that the minimum clinically impor-
`tant difference in OSDI score is 4.5 units in patients with
`mild or moderate OSD.17 The patients in this study who
`continued on BAK-preserved latanoprost 0.005% had OSDI
`scores that improved by only 2.7 points, which did not meet
`
`Table 2 Most frequently reported treatment-emergent ocular
`adverse events
`
`BAK-free
`travoprost 0.004%
`n = 362
`9 (2.5%)
`11 (3.0%)
`
`BAK-preserved
`latanoprost 0.005%
`n = 362
`4 (1.1%)
`4 (1.1%)
`
`5 (1.4%)
`3 (0.8%)
`2 (0.6%)
`
`2 (0.6%)
`4 (1.1%)
`4 (1.1%)
`
`eye irritation, n (%)
`hyperemia of
`the eye, n (%)
`eye pain, n (%)
`eye pruritus, n (%)
`Foreign body
`sensation in
`eyes, n (%)
`
`that established minimum clinically important difference
`requirement.
`Of patients with mild OSDI scores at baseline, those
`who were transitioned to BAK-free travoprost 0.004% were
`significantly more likely to return to normal OSDI scores at
`week 12 and had significantly lower mean scores on the OSDI
`at week 12, when compared with the patients who remained
`on BAK-preserved latanoprost 0.005%. These outcomes
`confirm the validity of the recommendation to switch to
`BAK-free therapy that has been advocated by researchers
`who have investigated BAK-preserved IOP-lowering medica-
`tions in vitro,3,18 in animals,4–6 and in glaucoma patients.7–9,19
`Because reversal of the corneal epithelial cell layer damage in
`glaucoma patients who were treated chronically with BAK-
`preserved medications would occur gradually over time,
`patients with moderate or severe OSD may need more than
`the 12 weeks allotted in this study before significant OSD
`improvements could be observed after switching to BAK-
`free travoprost 0.004%. Patients with moderate or severe
`OSD may have had more factors influencing their ocular
`surface health than the BAK insult alone; ie, these patients
`may have had underlying dry eye pathology or other factors
`that compounded the effect of BAK. In such cases, removing
`BAK would be helpful, but not sufficient, in ameliorating
`the condition.
`Another important finding of this study was that patients
`who had been exposed to BAK-preserved latanoprost 0.005%
`for more than 24 months prior to entry into the study were
`significantly more likely to improve to normal OSDI scores
`if they were transitioned to BAK-free travoprost 0.004% than
`if they remained on BAK-preserved latanoprost 0.005%. The
`same was not true of patients pretreated with BAK-preserved
`latanoprost 0.005% for 24 months or less. These results may
`indicate some cumulative, long-term adverse effect of BAK
`on ocular surface health, as has been suggested by in vitro
`studies demonstrating that BAK has dose-dependent toxic
`effects,20–22 and by animal studies demonstrating that BAK is
`persistent in ocular tissues over time.23 Together, these pre-
`clinical studies indicate that long-term buildup of daily doses
`of BAK in glaucoma patients could have dose-dependent
`toxic effects on ocular surface health. Such long-term expo-
`sure may be common among patients with glaucoma or ocular
`hypertension: one large-scale epidemiologic study found that
`patients had been medically treated for a median of 3.9 years.2
`For the patients who were exposed to BAK-preserved latano-
`prost 0.005% for more than 24 months before entering this
`study, the effects of long-term BAK exposure on OSD were
`reversible within 12 weeks for some patients after switching
`
`Clinical Ophthalmology 2010:4
`
`submit your manuscript | www.dovepress.com
`Dovepress
`
`1259
`
`

`

`Katz et al
`
`Dovepress
`
`to BAK-free travoprost 0.004%. Similar reversibility of
`adverse ocular signs and symptoms has been demonstrated
`with therapeutic switches to BAK-free IOP-lowering treat-
`ments elsewhere in the literature.2,19
`Some of the improvements that were observed in this
`study may be attributable to regression towards the mean,
`to increased familiarity with the