Efficacy and Safety of an Ophthalmic DMPC-Based Nanoemulsion in Patients with Dry Eye Disease: A Phase I/II Randomized Clinical Trial

Purpose . Te goals of this study were to evaluate the safety and efcacy of an ophthalmic 1,2-dimyristoyl-sn-glycero-3-phos-phocholine (DMPC)-based nanoemulsion (Nanodrop ® ) in patients with dry eye disease (DED). Methods . Tis was a randomized phase I/II multicentric, prospective, double-blind clinical trial. Patients (phase I: n � 25 and phase II: n � 101) were assigned to receive either PRO-176 (Nanodrop ® ) or Systane Balance ® (control) for 29days. Once the visits of the frst 25 subjects were completed, if there were less than 20% of unexpected adverse events (AEs), related to PRO-176, recruitment was continued until the sample was completed for noninferiority (efcacy) analysis (phase II, n � 126). Efcacy endpoints were the ocular surface disease index (OSDI), tear break-up time (TBUT), epithelial defects, best corrected visual acuity (BCVA), and the incidence of expected AE. Results . For the phase I portion of the study, there were no diferences between groups regarding the incidence of AE. All related-AE symptoms in both groups were mild and expected. For the phase II subset, there was a signifcant reduction in OSDI scores at day 29 and noninferiority between treatments was confrmed ( p � 0 . 650, CI 95% [ − 8.7, 5.5]). Similar improvement was observed for TBUTalthough no signifcant intergroup diferences were found ( p � 0 . 518, CI 95% [ − 0.08, 1.6]). Tere were no signifcant diferences between treatments for epithelial staining or safety parameters. Conclusions . Topical application of PRO-176 is as safe and efective as the controls. Both groups were clinically similar in terms of efcacy and safety. Te results support the hypothesis that ophthalmic DMPC-based nanoemulsion may improve clinical parameters and symptoms in patients with DED. Tis trial is registered with NCT04111965.


Introduction
Dry eye disease (DED) is a common and complex pathology of the ocular surface that afects 6-34% of the word's adult population and signifcantly reduces quality of life [1][2][3].
Tis disorder is caused by an insufcient or inefcient and unstable tear flm that causes a range of signs and symptoms, including discomfort, foreign body sensation, hyperemia, ocular pain, visual alterations, and disruption of the outermost layer of the cornea detected by a fuorescein eye stain test, reduced tear flm break-up time (TBUT), and so on [4,5].
Te described alterations of the tear flm may be caused by decreased production of the tear's elements and/or due to its elevated evaporation rate. Tis form of DED is in many cases caused by disruptions of the lipidic component of the tear [5,6]. Te lipid layer is composed of fatty acids and fatty alcohols, as well as lipids, mainly the nonpolar type. However, small amounts of phospholipids or polar lipids are also present, and their hydrophilic properties act as an interface between the lipid and aqueous layers [7,8].
Even though the biochemical profle, rheological and clinical properties of the tear flm's lipid layer are not completely understood yet, it is known that one of its main purposes is to decrease the evaporation rate of the tear flm [5][6][7]. In patients with defciency of this portion of the tear, like in meibomian gland disfunction (MGD), one of the treatment goals must be tacking the lipid insufciency to treat the associated dry eye. Emollient-containing eye drops are useful in these cases, aiding in the replenishment of lipiddefcient tear flms and therefore reducing tear evaporation and contributing to tear flm stability [5]. However, some of the adverse efects of these emollient eyedrops include temporary blurry vision and reduced contrast sensitivity due to their viscous consistency [5].
A nanoemulsion contains small particles (10-1000 nm) and both lipidic and aqueous components. Formulating a nanoemulsion eye drop allows an aqueous-like dispersion over the ocular surface, adding lipids to the tear flm and therefore avoiding the emollient-related adverse events.
Te rheological properties of propylene glycol, a synthetic lubricant used in several commercially available eyedrops (some of which include lipids in their formulations), modify the tear flm's viscosity while the lipidic phase of the nanoemulsion acts as a surfactant, aiding in the adequate dissemination of nonpolar lipids in the aqueous component of the tears. Te resulting barrier between the aqueous and lipidic phases supports the nonpolar phase, decreasing the evaporation rate of the tear flm [9].
Phase I/II clinical trials test the safety, side efects, and how well a certain type of disease responds to a new treatment [10]. In the present study, safety and efcacy of the ophthalmic nanoemulsion composed of a novel oily phase of castor oil and 1,2-dimyristoyl-sn-glycero-3-phosphocholine, consequently called DMPC-based nanoemulsion (PRO-176), were evaluated in terms of reduction of signs and symptoms, thus helping in the treatment of DED. It was compared with a commercially available and widely distributed emulsion (Systane Balance ® ) indicated for temporary relief of some of the symptoms associated to DED. Tis reference product was selected since its formulation is similar to PRO-176 and its safety and efcacy have been previously established.

Study Design.
Tis was a multicenter, comparative, prospective, parallel group, phase I/II clinical trial. Te study was registered in clinicaltrials.gov as NCT04111965. It was conducted in fve centers in Mexico. An ethics committee in each center reviewed and approved the study's protocol and informed consent (see Acknowledgments section). Te research was conducted in compliance with the Declaration of Helsinki and in accordance with Good Clinical Practices Standards. All patients who participated in this study provided written and signed informed consent. Patients in phase I were recruited between December 1, 2020 (FPFV) and March 16, 2021 (LPLV). Patients in phase II were recruited between June 7, 2021 (FPFV) and December 30, 2021 (LPLV). Tis study's design and methods are based on previously published clinical trials [11,12].
Tis study consisted of two phases: phase I and phase II. Phase I aimed to confrm the safety of PRO-176 in patients with DED. Phase II aimed to investigate efcacy with the primary endpoint of OSDI score, taking place only after it was confrmed that a less than 20% of unexpected AE related to PRO-176 were present during phase I, and therefore allowing recruitment to continue until the sample was completed for efcacy analysis.

Participants.
Inclusion criteria comprehended patients either men or women (aged ≥18 years), with DED diagnosis (OSDI score ≥13, plus one of the following: corneal staining with more than 5 sites, conjunctival staining with more than 9 sites, or TBUT <10 seconds). Exclusion criteria were the use of topical ocular drops and systemic medication that may have afected the study's results (systemic steroids, immunomodulators, and tetracyclines), patients with conjunctivitis, anterior blepharitis, demodex, eye parasitic infections, unresolved eye trauma, healing disorders of the ocular surface, and presence of any illness that could interfere with study parameters (e.g., glaucoma or retinal diseases), history of penetrating keratoplasty, present a distance best corrected visual acuity (BCVA) of 20/200 or worse in one of the eyes, use of contact lenses, previous history of any ophthalmic surgical procedure within 3 months before baseline, subjects with a single functional eye, and patients who were pregnant or at risk for pregnancy (having no birth control treatment) or breastfeeding.

Treatment and Evaluations.
A total of one hundred and twenty-six subjects (63 subjects per group) were randomized 1 : 1 to receive a control ophthalmic emulsion of propylene glycol 0.6% (Systane Balance ® , Alcon Laboratories, Inc., Fort Worth, TX, USA (phase I, n � 12 and phase II, n � 51)) or PRO-176 ophthalmic nanoemulsion (Nanodrop ® , Laboratorios Sophia, SA de CV, Zapopan, Jalisco, Mexico; (phase I, n � 13 and phase II, n � 50)). Subjects' selection and evaluation bias was minimized by random assignment to treatment groups and by treatments-assignment masking. Randomization numbers were generated using computer software (SAS Institute, Inc., Cary, NC, USA). Subjects instilled a drop of study drug topically in the inferior conjunctival sac of both eyes four times per day (QID) for 29 days. If patients required administering additional drops, they could do it as needed (Pro re nata schedule, PRN) [12]. All the researchers and other sponsoring team members were blind to treatment assignment throughout the study. Follow-up visits took place on days 8 and 29 after randomization (day 1). A safety call was carried out one week after the fnal visit (35 th ± 1 day).
Te study drug was discontinued if either the principal investigator or patient judged that it was not in the latter's best interest to continue or if a female patient became pregnant.

Phase I: Safety Assessment.
Because this is the frst trial of PRO-176 in the target population, a safety analysis (20% of the total sample, 25/126) was conducted as the phase I subset of the study. Safety was measured by the incidence of unexpected AE related to PRO-176. Te expected AE (listed versus unexpected), their causality and relatedness assessment (related, possibly related, and not related), and their severity (mild, moderate, and severe) were analyzed [12][13][14][15][16]. Te following information about each AE was collected: severity, event duration, frequency, seriousness, relationship to investigational product (PRO-176 and control), action taken, and outcomes. For AE evaluation, any method used in the trial to elicit subject/reported AE, such as a diary, checklist, and memory aid, whether applied face-to-face or otherwise was considered. All AE occurring during the clinical protocol were recorded and included in the analysis. When the visits of the frst 25 subjects were completed (safety cohort), if there were less than 20% of unexpected AE related to PRO-176, recruitment was continued until the sample was completed for efcacy analysis (phase II, efcacy cohort) (see Figure 1). After obtaining clearance through the PRO-176 safety data analysis, research centers continued the trial, and the phase II section was conducted.

Phase II: Efcacy Assessment.
Phase II was an expanded efcacy cohort. Te primary efcacy endpoint was demonstrating the noninferiority of PRO-176 compared to the control emulsion regarding efcacy as treatment of DED by means of the OSDI score, over the 29-day treatment period [17]. Noninferiority was determined as a diference inferior to 5 points in the OSDI test score between PRO-176 and control. Secondary outcomes were TBUT (number of seconds) performed at each follow-up visit, presence of the epithelial defects evidenced by fuorescein corneal and lissamine green conjunctival staining tests (FCS and LGCS, respectively), the BCVA evaluated through the Snellen chart (to provide standardized and well-controlled assessments of visual acuity during the study, the same lighting conditions were used during the entire study visits), and the incidence of expected AE. Surface dye staining was classifed with a scale from 0 to 6, in accordance with the percentage of the afected area (using the SICCA scale, where 0 � no staining) [18].
Finally, since measurements obtained from the right and left eyes are usually correlated [19], data from the right eye were used to give a single data point per subject [20,21].

Statistical Analysis.
Statistical analyses were carried out using the R statistical software package (Te R Foundation for Statistical Computing; https://www.R-project.org). All data are expressed as mean ± SD unless indicated otherwise. Sample size calculation was performed to test the reduction in OSDI score after one month of treatment, with an expected diference of at least 5 points with an alpha of 0.05 and a power of 80% [22,23]. Terefore, 101 subjects were considered per group, allowing as much as 25% of excluded cases in the event of major protocol deviations. A total of 126 subjects were randomized. Statistical evaluations for differences were performed with the Student's t-test for continuous data. Te 95% confdence interval (CI) of these diferences was computed. Te ordinal variables were analyzed using p × q contingency tables, and the diferences were calculated with the Pearson chi-square test. All statistical analyses performed in this study were with a p < 0.05.
Demographic and baseline characteristics were similar between treatment groups, showing no signifcant diferences (see Table 1). Te mean age ± standard deviation (SD) was 46.2 ± 16.5 years (range 62 years), and 69% of subjects were female. All subjects in each group were diagnosed with DED (see Materials and Methods section). Drug administration was similar between groups, and the mean number of applications during the 8 frst days of the study was 4.26 ± 1.2 drops per day for control group vs. 4.11 ± 0.7 drops per day for PRO-176 (p � 0.414). At day 29, the QID scheme continued in both groups (4.20 ± 0.9 drops per day vs. 4.14 ± 0.8 drops per day, respectively, p � 0.722). Only 4 patients in each group followed a PRN schedule (≥5 drops per day).

Phase I: Safety by Unexpected-Related Adverse Events (AEs).
Safety data were analyzed for the frst 25 subjects who completed their follow-up visits up to day 35 (safety call). Te mean age ± SD for the control group was 38.7 ± 11.3 years vs. 45.1 ± 16.2 years for PRO-176 (p � 0.369). For the control group, 58.3% of subjects were female vs. 69.2% for PRO-176 (p � 0.688). A total of 34 AEs were reported by 64% (16/25) of the subjects randomized during the phase I period of the study. Tere were no signifcant diferences between treatments for the incidence of AE (p � 0.688). In each group, 17 AE were reported, all of them classifed as mild. Regarding the expectedness of AE, 32.4% were not treatment-related (11/34), and 67.6% were related (control 84.4% vs. 52.9% for PRO-176, p � 0.141). All related-AE in both groups were expected. During the phase I period, the most common class of reported AE was burning (29.4%), followed by itching (11.8%), and ocular pain, as well as foreign body sensations in the same proportion (8.8%).
Finally, causality assessment means fnding a causal association or relationship between a treatment and reaction. Tere are multiple criteria or algorithms available for stablishing a causal relationship in case of adverse drug reaction. Based on the World Health Organization-Uppsala Monitoring Center scale, 35.3% of AE were possible, 47.1% probable or likely, and 17.6% unlikely. Tere were no differences between groups (p � 0.246) (see Table 2).

Fluorescein Corneal Staining and Lissamine Green
Tere was a narrowed decrease in LGCS score since it was similar from baseline, at day 8 (−0.32 ± 0.86 vs.  Table 3).  Table 3).

Expected Adverse Events.
In the ITT population (n � 126), of the total AE reported in both phases, 20.4% related-AEs were unexpected for PRO-176 vs. 17.5% for the  Figure 1: Clinical study design. Phase I was a safety cohort and phase II was an expanded cohort for efcacy assessment. A total of 79.6% of the reported AEs were expected (e.g., Investigator's Brochure for PRO-176 and/or package insert/ summary of product characteristics for the approved/control product) and 77% related-AE for PRO-176 vs. 82.5% for control (p � 0.198). 40.8% were classifed as possible (PRO-176: 45.7% vs. 35.3% for control), 51.4% as probable or likely (50% vs. 52.9%, respectively), and 7.8% as unlikely (4.3% vs. 11.8%). No signifcant diferences between groups were observed (p � 0.106). Te treatment-related AE, based on their causality assessment and expectedness, is shown in Figure 4. Tere was a total of 95% mild AE (98.9% in PRO-176 and 90.6% in control group), 3.9% moderate AE (only in control group), and 1.1% serious AE (one in each group). Te most common class of reported AE was burning (26.6%), followed by itching (11.4%) and blurry vision (6.6%), without diferences between groups (p � 0.089).

Discussion
Dry eye disease (DED) is a common and complex disorder caused by an insufcient or inefcient and unstable tear flm. It causes a range of signs and symptoms which alter the quality of life of those who from sufer it [5].
One of the elements of the tear flm is the lipid layer, which is comprised by elements such as fatty acids, fatty alcohols, and polar and nonpolar lipids. It is well known that this layer has the ability to decrease the tear flm's evaporation rate [5][6][7][8].
For patients who sufer conditions that alter the lipid layer, such as MGD, emollient-containing eye drops are believed to replenish their lipid-defcient tear flms, and therefore reduce tear evaporation and contribute to tear flm stability [5].
PRO-176 is a nanoemulsion that supplements lipids to the tear flm while maintaining aqueous-like dispersion over  LGCS, lissamine green conjunctival staining; OSDI, ocular surface disease index; PP, per-protocol; SD, standard deviation; TBUT, tear break-up time.  the ocular surface, both because of the small particles it contains and through the incorporation of propylene glycol in its formulation [9].
Tis study evaluated the safety (phase I) and efcacy (phase II) of PRO-176, in comparison to the commercially available Systane Balance ® . Following a complete pipeline of development, PRO-176 had already been subjected to a cytotoxicity preclinical evaluation in NHEK-Neo cells with satisfactory results [9]. On the other hand, during clinical evaluations, safety was measured during the phase I portion of this trial through the incidence of unexpected AE-related to the product. During this period, a total of 34 adverse events were reported (17 for each group), all of them classifed as mild. A total of 67.6% were considered related to the product, and all of them were expected. Te most common AEs were burning, itching, ocular pain, and foreign body sensation. Tere were no diferences between groups neither for the incidence of AEs, nor for the causality assessment.
After safety evaluation during the phase I segment of the study, after an incidence of less than 20% nonexpected AE was confrmed, recruitment continued for the phase II stage of the study. For efcacy, during the phase II portion of the trial, the primary endpoint was the OSDI score obtained over the treatment period, compared to that of the previously described reference product. Te mean change in the OSDI score from baseline to day 29 was −31.5 ± 19.3 for the reference product and −33.1 ± 18.6 for PRO-176, with no signifcant diferences between groups. Reduction in the OSDI score results ≥30% was observed in 90.3% of the subjects, and ≥50% in 74.3%. Once again, no signifcant diferences were found between groups asserting the noninferiority of PRO-176 compared to the reference product.
Secondary variables included TBUT, corneal fuorescein staining, conjunctival lissamine green staining, BCVA, and incidence of expected AEs. Tere were no signifcant differences between groups for any of these evaluations, demonstrating the similar efcacy and safety profles of both products. Furthermore, QID dosing of PRO-176 and the reference product was not diferent to PRN dosing between treatments and in terms of the endpoints evaluated. Tis data correlates to previous fndings, which proposed that regular use of artifcial tears might provide better symptomatic relief that PRN use [12]. DED is a condition that afects millions of lives around the world, and even though the understanding of the complex interactions of the tear flm, its components, and the ocular surface grows continuously, there is still continuously growing research regarding its diagnosis, classifcation, and treatment.
Lipid insufciency in the tear flm has been identifed as one of the causes of its increased rate of evaporation, and it is now clear that most cases of DED are not limited to one type of etiology but rather to a mixture of both aqueous-defcient and evaporative DED [24]. Terefore, it is necessary to develop an ocular lubricant with the ability to restore the lipid requirements of the tear flm and prevent evaporation, while sparing its users from the adverse events related to emulsions when applied on the ocular surface. PRO-176, as a nanoemulsion, fts this profle.
Tis study's limitations included the lack of classifcation of the type of DED diagnosed in the subjects. It is possible to assume that if only subjects with a certain diagnosis of evaporative or mixed DED patients were recluted, the results could have been more conclusive [24,25]. Te fact that no specialized tools were used to measure the lipid layer could also be considered a limitation [6,26,27].

Conclusion
In this study, the safety and efcacy of PRO-176 were proven by signifcant reduction in symptoms related to DED reported by the users and exposed by the OSDI score evaluation. Te results support that ophthalmic DMPC-based nanoemulsion (PRO-176) may improve clinical parameters and symptom in patients with DED. Furthermore, it presented no signifcant diferences with an already known and commercially available eye drop, thus ofering an alternative for the treatment of patients with evaporative or mixed DED.

AE:
Adverse events DED: Dry eye disease ITT: Intent-to-treat OSDI: Ocular surface disease index PP: Per-protocol TBUT: Tear break-up time BCVA: Best corrected visual acuity.

Data Availability
In addition to summary statistics, the data points behind means, frequencies, and dispersion measures are openly available in Open Science Framework (https://osf.io) as DOI 10.17605/OSF.IO/HRPQA.

Consent
Te study protocol and informed consent form were approved by their respective Institutional Review Boards, as