To demonstrate equivalence of polyquaternium-1–preserved travoprost 0.003% with benzalkonium chloride–preserved travoprost 0.004% in patients with open-angle glaucoma or ocular hypertension.
Double-masked, randomized, 2-treatment, equivalence clinical trial.
setting : Multicenter clinical trial conducted in 60 centers in the United States and Europe. patient population : Adult patients with open-angle glaucoma or ocular hypertension. One eye per patient was analyzed. intervention : Patients were randomized 1:1 to receive polyquaternium-1–preserved travoprost 0.003% (n = 442) or benzalkonium chloride–preserved travoprost 0.004% (n = 422) once daily for 3 months. main outcome measures : Mean intraocular pressure (IOP) was assessed at 8 AM, 10 AM, and 4 PM at week 2, week 6, and month 3. Supportive outcomes were mean and percent IOP change, percentage of patients achieving IOP <18 mm Hg or ≥30% IOP reduction, and adverse events.
Mean IOP was similar between groups at all study visits (travoprost 0.003% range, 17.5–18.9 mm Hg; travoprost 0.004% range, 17.4–19.0 mm Hg). Mean change (least squares mean differences, −0.1 to 0.3 mm Hg; 95% confidence interval, −0.5 to 0.7 mm Hg) and percentage change (travoprost 0.003%, 28.4%–30.7%; travoprost 0.004%, 28.5%–31.0%) from baseline were comparable. The percentages of patients with IOP <18 mm Hg and ≥30% reduction of IOP were also similar. Hyperemia was the most frequent treatment-related adverse event with both formulations (travoprost 0.003%, 11.8%; travoprost 0.004%, 14.5%).
In patients with open-angle glaucoma or ocular hypertension, polyquaternium-1–preserved travoprost 0.003% solution provided equivalent IOP-lowering efficacy to that of benzalkonium chloride–preserved travoprost 0.004%.
Glaucoma is characterized by optic neuropathy and gradual visual field loss, the latter being a result of retinal ganglion cell atrophy. The second-leading cause of blindness worldwide, glaucoma is projected to affect approximately 79 million people by 2020.
Elevated intraocular pressure (IOP) is a leading risk factor for the development of primary open-angle glaucoma, and IOP reduction has been found to slow disease progression. Treatments to lower IOP include topical ocular hypotensive medications. Prostaglandin analogues are recommended as first-line topical agents because of their IOP-lowering ability, safety profile, and once-daily dosing.
Although not associated with systemic adverse events, prostaglandin analogues may cause local adverse events that can influence adherence and therefore compromise expected treatment outcomes. Hyperemia is the most commonly reported adverse event in patients receiving prostaglandin analogues; adverse effects on appearance (eg, bloodshot eyes) or discomfort associated with hyperemia may diminish compliance with treatment.
TRAVATAN (travoprost 0.004% ophthalmic solution; Alcon Laboratories, Inc, Fort Worth, Texas, USA) is a prostaglandin F 2α analogue indicated for the reduction of IOP in patients with open-angle glaucoma or ocular hypertension. Like many ophthalmic medications, the initial formulation contained the commonly used preservative benzalkonium chloride, which has been associated with conjunctival inflammation, tear film disruption, and symptoms of ocular surface disease or decreased ocular surface health following chronic exposure. Since its introduction in 2001, efforts have been made to improve the safety profile of travoprost 0.004%, and 2 formulations preserved without benzalkonium chloride are currently marketed. Travoprost 0.004% preserved with sofZia (Travatan Z; Alcon Laboratories, Inc) is available in the United States, Canada, and Japan; travoprost 0.004% preserved with polyquaternium-1 (POLYQUAD; Alcon Laboratories, Inc) is marketed throughout most of the world, including Europe, South America, and Asia. Noninferiority of the IOP-lowering efficacy of polyquaternium-1–preserved travoprost 0.004% compared with benzalkonium chloride–preserved travoprost 0.004% was demonstrated in a randomized, double-masked study of 371 patients with open-angle glaucoma or ocular hypertension. Safety profiles of the 2 formulations were generally similar; however, a lower incidence of hyperemia was observed with the polyquaternium-1–preserved formulation.
Further improvements in the safety profile of travoprost might be achieved by decreasing the concentration of the active drug and preserving the formulation with an alternative to benzalkonium chloride. To that end, a polyquaternium-1–preserved travoprost formulation with a reduced active drug concentration (0.003% [30 μg/mL]) is being developed for worldwide use; this formulation may confer improvements in overall drug safety while maintaining optimal IOP-lowering efficacy. The aim of the current clinical study was to demonstrate equivalence of travoprost 0.003% preserved with polyquaternium-1 to travoprost 0.004% preserved with benzalkonium chloride in patients with open-angle glaucoma or ocular hypertension.
Study Design and Medications
This was a double-masked, randomized, 2-treatment clinical trial designed to demonstrate equivalence between travoprost 0.003% solution and travoprost 0.004% in adult patients with open-angle glaucoma or ocular hypertension (registered at ClinicalTrials.gov ; trial identification number, NCT01453855 ; trial registry date, October 13, 2011). The study was conducted at clinical sites in the United States, Sweden, Germany, Austria, Spain, and Finland between November 29, 2011 and August 3, 2012. The study protocol received prospective institutional review board (IRB) approval in the United States from Sterling IRB, the Committee for the Protection of Human Subjects, and Western IRB. The protocol also received prospective approval from independent ethics committees in Europe: Regionala Etikprövningsnämnden i Uppsala (Sweden); Ceic Capio Hospital General de Catalunya (Spain); Ethikkommission für das Bundesland Salzburg and Ethikkommission der Medizinische Universität Wien (Austria); Ethik-Kommission Landesärztekammer Rheinland-Pfalz and Ethik-Kommission der Landesärztekammer Brandenburg (Germany); and Pirkanmaan sairaanhoitopiirin eettinen toimikunta (Finland). The study was performed in compliance with the Declaration of Helsinki and Good Clinical Practice. Before entering the study, all patients provided written informed consent.
The study consisted of 6 visits conducted during 2 sequential phases: the screening/eligibility phase, which included a screening visit and 2 eligibility visits, and the treatment phase, which included 3 on-therapy follow-up visits conducted at week 2, week 6, and month 3. At screening, patients discontinued use of all prestudy ocular hypotensive medications, and the first eligibility visit was scheduled after a predetermined washout period according to patients’ prestudy medication: miotics and oral/topical carbonic anhydrase inhibitors, ≥4 days; α and α/β agonists, ≥13 days; β antagonists, prostaglandin analogues, and combination drugs, ≥27 days. For combination drugs comprising ocular hypotensive medications from more than 1 class, the longest washout period for the individual components was used.
Investigators, subinvestigators, all designated IOP operators and readers, patients, the study sponsor, and monitors involved in obtaining, reporting, or reviewing clinical evaluations throughout conduct of the study were masked to treatment. Patient randomization was blocked to ensure a balance of study treatment allocations within investigational sites. The randomization was also stratified by 8 AM baseline IOP (low, 24–27 mm Hg; and high, 28–36 mm Hg) to ensure a balance of treatment groups within each IOP stratum. Upon study entry, patients were assigned screening numbers of 001 to 099 in the appropriate numerical sequence by designated site personnel. The list of patient numbers was generated by statistical personnel not involved in conduct of the study. At the end of the second eligibility visit, eligible patients were randomized in a 1:1 ratio by assigned number and the criteria described above to either polyquaternium-1–preserved travoprost 0.003% or benzalkonium chloride–preserved travoprost 0.004%. An Interactive Web Response System then instructed unmasked study staff which treatment to dispense to patients. Patients were instructed to instill 1 drop of their assigned drug in both eyes once daily at 8 PM (±30 minutes) for 3 months, unless a safety issue prevented instillation in the nonstudy eye. Individual patient treatments were masked until all study data were verified, validated, and locked.
Safety and efficacy variables were assessed at selected time points (8 AM, 10 AM, and 4 PM) during week 2, week 6, and month 3 study visits. One eye from each patient was chosen as the study eye, and only the study eye was used in the efficacy analysis. If only 1 eye of a patient was treated, that eye was selected as the study eye. If both eyes were treated, the worse evaluable eye was selected as the study eye. The worse eye was defined as the eye with the higher IOP at 8 AM averaged across the 2 eligibility visits. If IOP values were equal at 8 AM, the worse eye was defined as the eye with the higher IOP at 10 AM averaged across the 2 eligibility visits. If values were equal at 10 AM, the worse eye was defined as the eye with the higher IOP at 4 PM averaged across the 2 eligibility visits. Finally, if both eyes were equal at 4 PM, the right eye was selected for analysis.
Patients included men and women aged ≥18 years with a diagnosis of open-angle glaucoma or ocular hypertension. Patients had to have a mean IOP (after washout of previous treatments) in at least 1 eye of ≥24 mm Hg at the 8 AM (±30 minutes) time point and ≥21 mm Hg at the 10 AM (±30 minutes) and 4 PM (±30 minutes) time points (the same eye). Mean IOP could not be >36 mm Hg in either eye at any time point.
Patients were excluded if they had a modified Shaffer angle grade <2 in either eye; cup-to-disc ratio >0.8; severe central visual field loss; chronic, recurrent, or severe inflammatory eye disease; intraocular surgery or ocular trauma within the previous 6 months; ocular infection or inflammation or ocular laser surgery within the previous 3 months; central corneal thickness >620 μm; best-corrected visual acuity score worse than 55 Early Treatment Diabetic Retinopathy Study letters; clinically significant or progressive retinal disease or other severe ocular pathology; hypersensitivity to prostaglandin analogues; or any abnormality preventing applanation tonometry in either eye. Patients were also excluded if they were unable to discontinue all IOP-lowering ocular medications before the study.
The primary efficacy variable was mean IOP at the week 2, week 6, and month 3 visits, measured at 8 AM, 10 AM, and 4 PM. Supportive efficacy variables assessed at each visit and time point included mean change from baseline in IOP, percentage change from baseline in IOP, proportion of patients with IOP <18 mm Hg, and proportion of patients who achieved ≥30% IOP reduction from baseline. IOP was measured by Goldmann applanation tonometry; 2 consecutive IOP measurements were taken for each eye at all time points. Baseline IOP was determined by averaging the time-matched measurements from the 2 eligibility visits; if IOP data were missing for 1 visit, the nonmissing IOP value was used.
Safety variables assessed included solicited and unsolicited adverse events, which were coded using the Medical Dictionary for Regulatory Activities, version 13.0, and recorded throughout the study and at each visit. Adverse events were presented for each treatment group categorized by severity (mild, moderate, or severe) and relationship to the study drug.
Ocular hyperemia was evaluated at the second eligibility visit and the week 2, week 6, and month 3 visits at 8 AM, 10 AM, and 4 PM. Hyperemia assessment was conducted before IOP measurement or instillation of a tonometry-disclosing agent. Ocular hyperemia assessments were made by visual inspection, performed by the same observer throughout the study, and scored from 0 to 3 in 0.5-unit increments by comparison with a standard set of photographs.
Best-corrected visual acuity (assessed with an Early Treatment Diabetic Retinopathy Study chart) and ocular signs (eyelids/conjunctiva, cornea, lens, and iris/anterior chamber) were evaluated with slit-lamp microscopy at screening and at the 8 AM time point for every postscreening study visit. Visual field function testing (standard automated perimetry) was performed at screening and at 8 AM at the month 3 visit. Central corneal thickness (measured with pachymetry) was assessed, and dilated fundus examination (vitreous, retina, macula, choroid, optic nerve, and cup-to-disc ratio) was performed at screening and at 4 PM at the month 3 visit.
The primary efficacy analysis was conducted in the intent-to-treat analysis set, defined as all patients who received study drug and completed at least 1 scheduled on-therapy study visit. Missing data were not imputed. The per-protocol analysis set consisted of all patients who satisfied prerandomization inclusion and exclusion criteria, received study drug, and completed at least 1 scheduled on-therapy visit; the per-protocol set provided supportive data for the primary efficacy endpoint. Treatment-group differences in mean IOP (the primary efficacy variable) and mean IOP change from baseline were examined using a pairwise t test at each time point of each scheduled on-therapy study visit. Pairwise t tests and confidence intervals were based on the least squares means derived from a statistical model that accounted for correlated IOP measurements over time in individual patients and included baseline IOP stratum and investigational center as covariates.
Descriptive statistics were summarized for patient baseline demographics and IOP at each on-therapy visit (week 2, week 6, and month 3) and assessment time point (8 AM, 10 AM, and 4 PM). To conclude equivalence, the 2-sided 95% confidence interval for the difference in IOP between treatment groups (ie, the mean IOP in the travoprost 0.003% solution group minus the mean IOP in the travoprost 0.004% group) had to be within ±1.5 mm Hg at each of the 3 assessment time points for each on-therapy visit. These data were also assessed using a more stringent criterion of ±1.0 mm Hg. Safety variables were also summarized using descriptive statistics (eg, frequency and percentage, or mean change from baseline) as appropriate.
Based on an IOP standard deviation of 3.5 mm Hg, a 5% chance of type I error, and an assumption that the population means are identical between groups, a sample size of 320 patients per treatment group was determined to have ≥99% power that the 95% 2-sided confidence interval of the difference in IOP between groups at any scheduled on-therapy assessment would fall within ±1.5 mm Hg, and ≥90% power that the confidence interval would fall within ±1.0 mm Hg. Target enrollment of 720 patients was determined to ensure that ≥640 patients (320 per group) would be followed for 3 months.
Overall, 864 patients were randomized (travoprost 0.003%, n = 442; travoprost 0.004%, n = 422). Of these, 860 were included in the intent-to-treat population (travoprost 0.003%, n = 442; travoprost 0.004%, n = 418; Table 1 ); 851 patients were included in the per-protocol population (travoprost 0.003%, n = 436; travoprost 0.004%, n = 415). The study was completed by 432 of 442 patients (98%) in the travoprost 0.003% group and by 408 of 422 patients (97%) in the travoprost 0.004% group. Overall, 24 of 864 patients (3%) discontinued early from the study, including 10 (2%) in the travoprost 0.003% group and 14 (3%) in the travoprost 0.004% group ( Figure 1 ). The most common reasons for discontinuation were adverse events (n = 7), inadequate control of IOP (n = 6), and patient decision unrelated to an adverse event (n = 6). Additionally, 3 patients were lost to follow-up. Other reasons for discontinuation were noncompliance (n = 1) and “other” (n = 1).
|Polyquaternium-1–Preserved Travoprost 0.003% (n = 442)||Benzalkonium Chloride–Preserved Travoprost 0.004% (n = 418)||Total (n = 860)|
|Mean ± standard deviation||65.4 ± 10.5||65.0 ± 10.9||65.2 ± 10.7|
|<65, n (%)||189 (43)||191 (46)||380 (44)|
|≥65, n (%)||253 (57)||227 (54)||480 (56)|
|Race, n (%)|
|White||316 (72)||307 (73)||623 (72)|
|Black||112 (25)||106 (25)||218 (25)|
|Asian||11 (3)||4 (1)||15 (2)|
|Other||3 (1)||1 (0.2)||4 (0.5)|
|Sex, n (%)|
|Male||173 (39)||174 (42)||347 (40)|
|Female||269 (61)||244 (58)||513 (60)|
|Diagnosis, n (%)|
|Ocular hypertension||130 (29)||121 (29)||251 (29)|
|Open-angle glaucoma||304 (69)||290 (69)||594 (69)|
|Open-angle glaucoma with pigment dispersion||7 (2)||7 (2)||14 (2)|
|Open-angle glaucoma with pseudoexfoliation||1 (0.2)||0||1 (0.1)|
|Baseline intraocular pressure, n (%)|
|24–27 mm Hg||303 (69)||291 (70)||594 (69)|
|28–36 mm Hg||139 (31)||127 (30)||266 (31)|
Within the intent-to-treat population, patients had a mean age of 65 years; most patients were women (n = 513 of 860, 60%) and white (n = 623 of 860, 72%) ( Table 1 ). A majority of patients had a diagnosis of open-angle glaucoma (n = 594 of 860, 69%). Mean baseline IOP measurements for patients receiving travoprost 0.003% or travoprost 0.004% were comparable at 8 AM (26.9 mm Hg and 27.1 mm Hg, respectively), 10 AM (25.4 mm Hg and 25.6 mm Hg), and 4 PM (24.6 mm Hg and 24.8 mm Hg). Baseline IOP measurements across all time points and study drug groups ranged from 21 to 36 mm Hg. Mean corneal thickness at baseline was similar between groups (travoprost 0.003%, 552.9 μm; travoprost 0.004%, 551.8 μm). No substantial differences were observed between groups regarding any demographic parameter or baseline characteristic.
Primary Efficacy Analysis
On-treatment IOP values between the travoprost 0.003% solution group (range, 17.5–18.9 mm Hg) and travoprost 0.004% group (range, 17.4–19.0 mm Hg) were similar ( Figure 2 ). The least squares mean differences in IOP values between the travoprost 0.003% and travoprost 0.004% groups at each time point at each study visit ranged from −0.3 to 0.0 mm Hg, with confidence intervals ranging from –0.7 to 0.4 mm Hg ( Table 2 ). Thus, at all 9 assessments, the 95% confidence intervals for the mean differences in IOP between treatment groups were within the prespecified margin of ±1.5 mm Hg, indicating statistical equivalence. Furthermore, 95% confidence intervals for mean between-group IOP difference at each time point of all study visits were within an equivalence margin of ±1.0 mm Hg. Mean IOP and 95% confidence intervals in the per-protocol data set supported the intent-to-treat data; all 95% confidence intervals were <1.0 mm Hg.
|Polyquaternium-1–Preserved Travoprost 0.003%||Benzalkonium Chloride–Preserved Travoprost 0.004%|
|N||Least Squares Mean ± Standard Error||N||Least Squares Mean ± Standard Error|
|8 AM||442||19.4 ± 0.16||416||19.5 ± 0.17|
|10 AM||442||18.6 ± 0.16||416||18.6 ± 0.17|
|4 PM||442||18.0 ± 0.16||416||18.3 ± 0.17|
|8 AM||439||19.3 ± 0.16||413||19.3 ± 0.17|
|10 AM||440||18.5 ± 0.16||413||18.6 ± 0.17|
|4 PM||440||18.0 ± 0.16||413||18.1 ± 0.17|
|8 AM||432||19.2 ± 0.17||408||19.3 ± 0.17|
|10 AM||432||18.3 ± 0.17||408||18.6 ± 0.17|
|4 PM||431||18.0 ± 0.16||408||18.0 ± 0.17|