Comparison of Preservative-Free and Preserved Latanoprost in Primary Open-Angle Glaucoma: A Randomized Cross-Over Study Authors: Ji Heon Lee, Kwanghyun Lee, Jong Woon Park Published: 2026 ABSTRACT PurposeTo compare the effects of preservative-free (PF) and benzalkonium chloride (BAK)-preserved latanoprost on intraocular pressure (IOP), objective ocular surface signs, and subjective symptoms in patients with primary open-angle glaucoma (POAG).MethodsA prospective, randomized, cross-over study was conducted with 29 POAG patients. Participants were randomized to receive either PF or BAK-preserved latanoprost for 2 months, followed by a cross-over to the alternative formulation for another 2 months. IOP, tear break-up time (TBUT), conjunctival staining, and patient-reported symptoms were assessed at baseline and after each treatment period.ResultsIOP was significantly and comparably reduced by both formulations (PF: 16.49 ± 3.89 to 15.05 ± 2.55 mmHg, p = 0.007; preserved: 15.60 ± 3.79 to 13.79 ± 3.53 mmHg, p < 0.001). However, BAK-preserved latanoprost significantly shortened TBUT (7.23 ± 2.37 to 6.07 ± 2.40 seconds, p = 0.007) and worsened conjunctival staining (0.31 ± 0.60 to 0.69 ± 0.71, p = 0.013), whereas PF latanoprost maintained tear film stability (TBUT, 5.61 ± 2.19 to 5.22 ± 2.26 seconds; p = 0.337) and improved conjunctival staining (0.76 ± 0.83 to 0.31 ± 0.60, p = 0.003). Subjective symptoms including burning (0.34 ± 0.81 [PF] vs. 1.21 ± 1.17 [preserved], p = 0.001), stinging (0.14 ± 0.35 vs. 0.79 ± 0.79, p < 0.001), and pruritus (0.21 ± 0.41 vs. 0.50 ± 0.64, p = 0.030) were markedly lower with the PF formulation. After cross-over, patients switching from preserved to PF latanoprost showed improvements in burning (p = 0.012) and photophobia (p < 0.001).ConclusionsPF latanoprost provides equivalent IOP-lowering efficacy to BAK-preserved latanoprost while offering significant advantages in ocular surface health and symptom tolerability. These findings support PF formulations as a preferred first-line option for glaucoma management to enhance treatment tolerability and patients’ quality of life. Prostaglandin analogs are established as first-line therapy for primary open-angle glaucoma (POAG) and ocular hypertension due to their potent intraocular pressure (IOP)-lowering efficacy and convenience [1,2]. Most commercially available glaucoma medications, including prostaglandin analogs, are supplied in multi-dose bottles, which require antimicrobial preservatives to prevent microbial contamination during repeated use. As a result, the preservative benzalkonium chloride (BAK) is commonly included in these formulations. However, BAK has been shown to cause ocular surface disease (OSD), which significantly impairs patient quality of life and treatment compliance, posing a significant challenge for glaucoma specialists in their day-to-day work [3,4]. Symptoms such as burning, dryness, and foreign body sensation are not mere discomforts; they are primary drivers of treatment discontinuation. Indeed, the impact of these symptoms is significant, which can lead to poor IOP control and subsequent visual field (VF) progression [5]. Preservative-free (PF) latanoprost formulations were created to deal with this well-established preservative-induced toxicity; these formulations were intended to preserve pharmacodynamic efficacy while removing BAK [5]. Although previous studies have reported that PF formulations offer noninferior IOP-lowering efficacy and benefits in ocular surface health [6–9], many of these parallel- group comparisons were not free from the inherent limitations of ocular surface assessment, such as subjectivity, inter-observer variability, and high inter-individual variability in baseline symptoms [10–12]. Moreover, despite the growing clinical use of PF formulations, evidence remains limited on how patients’ ocular surface symptoms and signs change when they are switched from preserved to PF medications in real-world practice. To address these limitations, we designed a prospective, randomized, cross-over study in which each patient served as their own control. This self-controlled approach minimizes inter-patient variability and enhances the evaluation of treatment-related changes [10]. By directly comparing PF latanoprost with BAK-preserved latanoprost within the same individuals, we aimed to provide clinically meaningful evidence regarding their comparative effects on IOP, ocular surface health, and subjective tolerability. MATERIALS AND METHODS Ethics statement This randomized prospective study adhered to the principles outlined in the Declaration of Helsinki and received approval from the Institutional Review Board of the National Health Insurance Service Ilsan Hospital (No. NHIMC 2021-03-057). Written informed consent was obtained from all participants prior to their enrollment in the study. Study design and participants From June 2021 to March 2023, patients aged between 19 and 80 years who were diagnosed with POAG and visited the glaucoma clinic at the National Health Insurance Service Ilsan Hospital were enrolled in this study. POAG is defined as glaucoma with an open anterior chamber angle, characterized by the following features: typical glaucomatous changes in the optic disc, defects in the retinal nerve fiber layer, and associated glaucomatous VF defects without secondary causes of optic neuropathy. Glaucomatous VF defect was defined as follows: (1) a cluster of three or more points with a probability <5% on the pattern deviation map in at least one hemifield, including at least one point with a probability of <1%; or (2) a pattern standard deviation with a probability of <5%; or (3) glaucoma hemifield test values outside the normal limits. The VF test was conducted using the C24-2 Swedish Interactive Thresholding Algorithm standard program on the Humphrey Field Analyzer (Carl Zeiss Meditec). The VF test results were considered reliable if the fixation loss was less than 20%, and the false positive and false negative rates were both less than 33%. The inclusion criteria for the study were as follows: (1) participants should have no prior use of glaucoma medications or have used latanoprost for less than 6 months; (2) central corneal thickness must be between 500 and 580 μm; (3) best-corrected visual acuity should be 20 / 80 or better in both eyes; and (4) participants must demonstrate high compliance, defined as 80% or greater adherence to the prescribed eye drop regimen. Exclusion criteria included a history of angle-closure glaucoma (either primary or secondary), congenital or traumatic glaucoma, intraocular inflammation occurring within the past 3 months, corneal conditions that could affect IOP measurements, ocular surgeries that could influence dryness (such as cataract or refractive surgery within the last 3 months), or hypersensitivity to prostaglandin analogs. Patients were randomized into two groups: one group was prescribed PF latanoprost (Monoprost, Samil) for 2 months and then changed to preserved latanoprost (Xalatan, Pfizer) for the following 2 months. The other group used preserved latanoprost first and used PF latanoprost for the following 2 months. All subjects were instructed to administer one drop of PF latanoprost to the glaucoma eye once every evening. All patients instilled eye drops for a total of 120 days and visited our clinic on the 60th and 120th day after participation (D60 and D120). Randomization was performed using a computer-generated random number sequence. Both eyes of eligible patients were enrolled. Clinical assessments During each visit (D0, D60, and D120), best-corrected visual acuity (converted to logMAR for statistical analysis), IOP, limbal and bulbar conjunctival hyperemia, corneal and conjunctival staining, and tear break-up time (TBUT) were evaluated. All ophthalmic examinations were performed by two examiners (KL and JHL). In the case of disagreement between the two examiners, JWP made the final decision. Due to the distinct packaging of the two formulations (single- dose units vs. multi-dose bottles), masking of the IOP and TBUT examiners was not feasible. Participants filled out a questionnaire about the OSD index (OSDI) and ocular tolerability at each visit. IOP was measured twice at each visit using a Goldmann applanation tonometer, and the average value was used for analysis. To minimize diurnal variation in IOP, patients were instructed to visit the hospital at the same time as their first visit. Limbal and bulbar conjunctival hyperemia were evaluated on a scale of 0 to 4 based on the Efron grading scale. For corneal and conjunctival staining, the Oxford schema grading system was used and scored from grade 0 to V [13]. TBUT was measured by placing a fluorescein paper strip (FLUO 900, Haag-Streit) in the conjunctival sac, instructing the patient to blink for several seconds, and then recording the time until the first defect of the fluorescent dye appeared in the dyed tear film layer under cobalt blue light [14]. The OSDI consisted of 12 questions, and the score given was directly proportional to the severity of the symptoms. OSDI was evaluated on a scale from 0 to 100 and classified as normal (0–12), mild (13–22), moderate (23–32), and severe (33–100) [15]. To evaluate ocular tolerability, the severity of 11 symptoms was scored on a scale of 0 to 3. Six of these symptoms were assessed immediately after the instillation of the study eye drop: burning, itching, stickiness, blurred vision, a foreign body sensation, and dryness. The remaining five symptoms were evaluated as daily-life symptoms: burning, itching, tearing, a foreign body sensation, and photophobia during the daytime. Statistical analysis Data were analyzed using the R ver. 4.5.0 (R Foundation for Statistical Computing). Results are presented as mean ± standard deviation, numbers, and percentages. For comparison of the baseline demographics between the two groups, we used the t-test and the chi-square test. We evaluated the carryover, period (session), and treatment effect by means of a generalized linear mixed model with patient ID as a linear mixed-effects model random intercept to account for inter-eye correlation. A p-value of less than 0.05 was considered statistically significant. The sample size was determined based on a priori power analysis using OSDI data from our previous study [9]; a minimum of 16 participants were required, and we enrolled 30 patients (15 per sequence group) to account for potential dropout. As a sensitivity analysis, we repeated the primary analyses including only treatment-naïve patients (those with no prior use of any glaucoma medication) to evaluate the robustness of findings after excluding patients with prior BAK exposure. The results of this sensitivity analysis are presented in the Supplementary Materials. Results Demographics A total of 30 participants were enrolled in the study. Fifteen participants began treatment with PF latanoprost, while the other fifteen started with preserved latanoprost. One participant who did not want to switch to a different eye drop withdrew from the study (Fig. 1). Baseline characteristics of the 29 patients (57 eyes) who completed the study are summarized in Table 1. Among these, 24 patients (47 eyes) were treatment-naïve, and five patients (10 eyes) had prior BAK-preserved latanoprost exposure for less than 6 months. No significant differences in baseline characteristics were observed between the two subgroups, except for TBUT, which was paradoxically higher in the prior BAK group (p = 0.025). The mean age was 60.9 ± 12.9 years in the PF latanoprost group and 59.4 ± 11.5 years in the preserved latanoprost group (p = 0.744). The sex distribution was comparable between groups (female/male, 7 / 8 vs. 6 / 8, p > 0.999). The baseline mean deviation was −5.7 ± 5.6 dB in the PF group and −3.6 ± 4.3 dB in the preserved group (p = 0.119). No carryover effect was observed for any of the variables. IOP and visual acuity Both formulations achieved significant and comparable IOP reductions from baseline. In the PF latanoprost group, mean IOP decreased from 16.49 ± 3.89 mmHg at baseline to 15.05 ± 2.55 mmHg after treatment (mean reduction, 1.44 mmHg; p = 0.007). In the preserved latanoprost group, IOP decreased from 15.60 ± 3.79 to 13.79 ± 3.53 mmHg (mean reduction, 1.81 mmHg; p < 0.001). The difference in IOP-lowering efficacy between formulations was not statistically significant (p = 0.666). Visual acuity remained stable throughout the study for both groups (Table 2). Ocular surface parameters The TBUT significantly decreased when using preserved latanoprost (p = 0.007). However, no significant decrease was observed in the PF latanoprost group (p = 0.337). Conjunctival staining improved significantly during the use of PF latanoprost (p = 0.003) but worsened after switching to preserved latanoprost (p = 0.013). Subjective symptoms In the PF latanoprost group, the mean OSDI score decreased from 20.83 ± 16.34 at baseline to 16.31 ± 19.85 after the first treatment period, and slightly increased to 21.41 ± 27.22 after the cross-over; these changes were not statistically significant (p = 0.150 and p = 0.105, respectively). In the preserved latanoprost group, the mean OSDI score was 27.08 ± 19.01 at baseline, 26.04 ± 16.31 after the first period, and 27.08 ± 21.46 after cross-over, with no significant differences observed (p = 0.154 and p = 0.679, respectively). The use of PF latanoprost was associated with fewer subjective symptoms, including burning (p = 0.001), stinging (p < 0.001), and pruritus (p = 0.030) (Table 3). After switching medications, several symptom changes were observed (Table 4). Patients who changed from preserved to PF latanoprost reported significant improvements in burning (1.31 ± 1.16 to 0.77 ± 1.14, p = 0.012), as well as a significant change in photophobia (0.54 ± 0.65 to 0.92 ± 0.84, p < 0.001). Conversely, switching from PF to preserved latanoprost led to a significant increase in burning (0.34 ± 0.81 to 1.00 ± 0.93, p = 0.001), pruritus (0.21 ± 0.41 to 0.55 ± 0.63, p < 0.001), stinging (0.14 ± 0.35 to 0.48 ± 0.91, p = 0.010), and itching (0.14 ± 0.35 to 0.34 ± 0.61, p = 0.013). Sensitivity analysis A sensitivity analysis restricted to treatment-naïve patients (24 patients, 47 eyes) demonstrated overall consistency with the primary analysis. Among 65 comparisons across ocular surface parameters and subjective symptoms, 58 (89.2%) yielded concordant results, with all effect directions preserved. Discordances were confined to borderline p-values ( p = 0.019–0.056 [primary analysis] vs. p = 0.066–0.308 [naïve-only]), attributable to reduced statistical power after excluding five patients with prior BAK exposure (Supplementary Tables 1–3). Notably, two subjective symptoms that did not reach statistical significance in the primary analysis became significant in the naïve-only analysis: foreign body sensation immediately after instillation (0.37 ± 0.63 [PF] vs. 0.90 ± 0.97 [preserved], p = 0.017) and photophobia during the daytime (0.37 ± 0.63 [PF] vs. 0.80 ± 0.62 [preserved], p = 0.014). Both symptoms were more severe with the preserved formulation. Discussion Using a cross-over design, this study directly compared the effects of PF latanoprost and BAK-preserved latanoprost on IOP and the ocular surface. Both formulations showed a significant and comparable decrease in IOP from baseline, but their profiles for patient tolerability and ocular surface health differed substantially. This study has the advantage of analyzing both objective and subjective outcomes, and in particular, demonstrated the consistent superiority of the PF formulation across multiple symptom domains. In just 2 months, PF latanoprost maintained stable tear film quality, improved conjunctival staining, and minimized immediate discomfort such as burning and stinging. These findings indicate that BAK-related toxicity can manifest early—even with short-term use—and that switching to PF formulations can rapidly reverse these adverse effects. Methodologically, the prospective, randomized, cross-over design effectively eliminates inter-patient variability by allowing each patient to serve as their own control. This approach is particularly advantageous in ocular surface research, where baseline variability is high and symptom perception is subjective. Consequently, the design enhanced the statistical sensitivity to detect the treatment-induced changes. Changes in subjective symptoms were statistically significant. Considering that each treatment period lasted only 2 months—a relatively short time compared to the long-term use of glaucoma medications—these results suggest that ocular surface alterations induced by the medication occur rapidly. The incidence of post-instillation symptoms, particularly burning, stinging, and itching, was markedly lower with the PF formulation. These are the precise symptoms that drive patient complaints and nonadherence. Although there was no statistical difference in the total OSDI score, this composite score can often dilute the impact of the most bothersome acute symptoms. The marked difference in these specific, irritating daily symptoms likely has a greater and more direct impact on real- world treatment adherence than a generalized OSDI score. In the sensitivity analysis restricted to treatment-naïve patients, foreign body sensation and photophobia additionally reached statistical significance. These symptoms were more severe with BAK-preserved latanoprost, suggesting that patients initiating glaucoma therapy for the first time may be more susceptible to BAK-induced subjective discomfort than those with prior exposure. This finding supports the use of PF formulations as a first-line treatment in drug-naïve glaucoma patients, particularly to minimize foreign body sensation and photophobia that may otherwise impair early treatment tolerability and adherence. Furthermore, our objective findings align with numerous studies demonstrating that switching from preserved to PF eye drops significantly improves ocular surface signs. A multicenter study found that switching from preserved latanoprost to PF latanoprost significantly improved TBUT and reduced ocular surface staining after 12 weeks [16]. Our results, captured in an even shorter (2-month) timeframe, confirm the rapidity of these changes. One pooled analysis also found that switching to PF latanoprost was twice as effective at reducing ocular hyperemia and led to a 32.2% improvement in a composite OSDI score [7]. One of the strengths of this study is its focus on patient- reported tolerability, a key factor in treatment adherence. Up to 60% of glaucoma patients fail to use their medications as prescribed, and this rate may double for patients experiencing adverse effects [17,18]. Discomfort upon instillation is one of the most common and powerful barriers to adherence [19]. This nonadherence is not merely a quality of life concern; it leads to wider IOP fluctuations, which are a significant risk factor for glaucoma progression [20,21]. Therefore, minimizing the discomfort associated with eye drop administration is a key therapeutic strategy for achieving stable long-term IOP control and, ultimately, preserving the patient’s vision. Our findings provide robust evidence that the PF formulation is a superior option for minimizing this primary barrier to adherence. Nevertheless, some limitations exist. The relatively small sample size may have limited the statistical power to detect subtler differences in some secondary parameters, such as conjunctival hyperemia; however, the analysis of 57 eyes was considered sufficient for the central limit theorem to apply, supporting the use of parametric-based inference. Furthermore, the 2-month treatment period for each arm, while adequate to assess acute epithelial changes and tolerability, was insufficient to evaluate long-term, chronic histological changes. Chronic BAK exposure has been linked to conjunctival inflammation and therefore affects the outcome of glaucoma surgery negatively [22]. A longer-term study would be needed to assess whether switching to PF formulations can mitigate this significant long-term risk. Additionally, the absence of a washout period between treatment phases raises the possibility that residual effects of either formulation—including delayed effects of short-term BAK exposure on baseline corneal sensitivity or sub-clinical ocular surface inflammation—may have carried over to the subsequent period, potentially influencing the outcomes observed in either phase. In our cohort, however, the only baseline ocular surface parameter differing significantly between prior-exposure and treatment-naïve subgroups was TBUT, which was paradoxically higher in the prior-exposure subgroup, opposite to the direction expected from BAK-induced toxicity, suggesting that the actual impact of such preexisting changes on our observed outcomes was likely limited. Additionally, the inclusion of five patients (10 eyes) with prior BAK-preserved latanoprost exposure of less than 6 months may have introduced heterogeneity; however, a sensitivity analysis restricted to treatment-naïve patients demonstrated consistent results across all primary and secondary endpoints (Supplementary Tables 1–3). Finally, due to visible differences in medication packaging (single-dose units vs. multi-dose bottles), blinding of participants and investigators was not feasible, which may have introduced detection bias for subjective symptom reporting. For the same reason, residual observer bias in investigator-assessed outcomes cannot be entirely excluded. In conclusion, PF latanoprost demonstrated comparable IOP-lowering efficacy to that of BAK-preserved latanoprost, while providing significant benefits in ocular surface stability and patient comfort. Given that these improvements were observed within only 2 months, PF formulations should be considered a preferred option for glaucoma management, particularly for patients experiencing ocular discomfort or poor adherence to preserved medications.