A Randomized Double-Masked Study Evaluating Safety and Efficacy of a Novel Topical Cooling Device for Chronic Ocular Surface Pain Authors: Anat Galor, Christopher Rapuano, Alice Epitropoulos, Jayne Weiss, Stephen Pflugfelder, Preeya Gupta, Deepinder Dhaliwal, Laura Periman, Bonnie An Henderson, Thanh T Nguyen, Chameen Samarawickrama, Helen Wu, Tomasz Stryjewski, James A Stefater, Mark Hinds, Pooja Bhindi, Cristos Ifantides Published: 2026 ABSTRACT AbstractIntroductionNo therapy is currently FDA approved to treat ocular pain. This study evaluated the safety and efficacy of a novel topical ocular cooling device for chronic ocular surface pain (COSP).MethodsThis was a prospective, randomized, double-masked, multicenter clinical feasibility study in adults aged 22–90 years with COSP. The intervention was a handheld device that applied cold temperatures to the conjunctival surface above the long ciliary nerves via two metallic contact surfaces at the 3 and 9 o’clock positions. Subjects were randomized 2:1 to a one-time, 4-min treatment with the active (frozen) or sham (refrigerated) device and followed for 8 weeks. Subjective pain was assessed via the visual analog scale (VAS), ocular pain assessment survey (OPAS), and chronic ocular pain questionnaire (COP-Q). Corneal sensitivity was evaluated with Cochet–Bonnet esthesiometry. Main outcomes were change from baseline in VAS, OPAS, and COP-Q scores. Post hoc analyses were conducted in eyes with peripheral-dominant pain based on proparacaine challenge. Adverse events (AEs) were recorded.ResultsUnilateral treatments were administered to 31 subjects (20 active, 11 sham). All three patient-reported outcomes (PROs) showed greater reductions in eye pain at week 8 in the active versus sham arm. In eyes with peripheral-dominant pain, COP-Q eye pain severity improved significantly versus sham at week 8 (nominal p = 0.042), and OPAS overall eye pain trended towards statistical significance (nominal p = 0.088). All three PROs exceeded thresholds for clinically meaningful improvement. Baseline corneal sensitivity was below normal in both groups and improved significantly only in the active group (nominal p = 0.041). No serious AEs occurred; the most common AE was transient conjunctival hyperemia.ConclusionA novel ocular surface cooling device is well tolerated and appears to provide durable relief of COSP in peripheral-dominant pain types for at least 8 weeks.Trial RegistrationClinicalTrials.gov identifier: NCT06479382.Supplementary InformationThe online version contains supplementary material available at 10.1007/s40123-026-01392-3.Keywords: Chronic ocular surface pain, Ocular surface disease, Eye pain, Cryotherapy, Dry eye disease Key Summary Points Why carry out this study? Chronic ocular surface pain (COSP) is a common condition that has no FDA-approved treatment and is frequently recalcitrant to treatments for other ocular comorbidities such as dry eye therapiesThis study investigated whether a novel ocular cooling device could alleviate pain safely using a 4-min in-clinic treatment. What was learned from this study? The study outcomes investigated three different patient-reported outcomes (PROs) instruments, as well as safety measures such as corneal sensitivity testing and visual acuity.This novel ocular surface cooling treatment is well tolerated and appears to provide relief from COSP in peripheral-dominant pain types with durable effects at least through 8 weeks post treatment.Baseline corneal sensitivity testing was below normal in both groups and statistically significantly improved in the interventional treatment, but not in the sham. Introduction Chronic ocular surface pain (COSP) was first defined in 2022 as a chronic medical condition, perceived as originating from the ocular surface, that persists for at least 3 months, and may be described as symptoms of pain, discomfort, burning, dryness, irritation, foreign body sensation, fatigue, or grittiness [1]. The pain may be spontaneous or evoked by stimuli such as wind and light and is often debilitating [2]. COSP can occur in the presence of underlying ocular surface disease, such as meibomian gland dysfunction, aqueous tear deficiency, ocular surface inflammation, prior trauma or surgery (ex: refractive corneal surgery), and systemic disease such as Sjögren’s and other autoimmune conditions. COSP, however, can also exist with minimal signs of ocular surface disease. Recognition of COSP as a unique clinical entity is an emerging concept, as the pain symptoms defining it have been often incorrectly attributed to psychiatric disease and/or dry eye disease (DED), an umbrella term which encompasses a number of ocular surface disorders, and which affects approximately 9–22% of all Americans [3, 4]. Dissatisfaction with current treatment options that target ocular surface health is common, with up to 25–37% of patients reporting persistent and inadequate relief of their ocular pain symptoms, even when there is an improvement in tear parameters [5, 6]. As a result of the difficulty in achieving adequate relief after potentially multiple medical treatment attempts, COSP frequently negatively affects quality of life and can be debilitating. A study by Baudouin et al. [7] reported that 2.4% of patients with ocular surface pain felt that “life is not worth living.” Current COSP management focuses on therapies for dry eye disease or meibomian gland disease, heat and cold therapy, and off-label therapies such as topical morphine or naloxone and systemic gabapentin. There is currently no US Food and Drug Administration (FDA)-approved treatment for COSP; however, the treatment landscape is evolving. Urcosimod (OKYO Pharma), a ChemR23 chemerin peptide agonist, received FDA Fast Track designation for neuropathic corneal pain in 2025 and single-patient compassionate use authorization in January 2026, with a phase 2b/3 trial planned. COSP is conceptually related to, but distinct from, neuropathic corneal pain (NCP), which emphasizes a neuropathic pathophysiology involving peripheral and central sensitization of corneal nociceptive pathways. COSP is defined clinically by chronic ocular surface pain symptoms with or without obvious ocular surface signs, whereas NCP represents a subset of patients in whom neuropathic mechanisms predominate. In practice, many patients with COSP exhibit features of NCP, such as allodynia and incomplete response to topical anesthetics. There is a significant unmet medical need for a safe and effective treatment that provides sustained improvement of symptoms for patients suffering from COSP. While COSP remains understudied [1], recent research has identified a link between patients with COSP and hypersensitization of the peripheral nociceptors, which creates a state of corneal somatosensory dysfunction [8]. Topical cooling has a long, well-established association with temporarily reducing pain and discomfort. This mechanism of action may be related to activation of the TRPM8 receptor, which is responsible for the relief associated with menthol and cooling. More recently, it has been clinically observed that patients undergoing topical freezing treatments to eliminate unwanted subcutaneous fat develop a transient cutaneous hypesthesia that lasts weeks to months [9]. Further work in rodent pain models has shown that an injection of an ice slurry with a sustained temperature of − 5 °C for a period of several minutes results in diminished pain impulses along the sciatic nerve for almost 3 months after procedure, with no damage to surrounding tissue [10]. In ophthalmology, the application of cryotherapy to the ciliary nerves in eyes with chronic glaucoma is known to reduce pain, even when intraocular pressure (IOP) control was not achieved [11–13]. The ETX-4143 device (EyeCool Therapeutics, Inc.) is a novel handheld device that contains an internal frozen (− 20 °C) glycerol solution. The device weighs approximately 98.5 g with dimensions of 8.2 × 3.7 × 4.5 cm. It is stored in a medical-grade freezer and removed prior to use. The cold temperature generated by the mixture can be transferred topically to the conjunctival surface overlaying the myelinated long ciliary nerves via two metallic contact surfaces with the goal of normalizing a patient’s corneal pain signaling, thereby decreasing symptoms of chronic ocular pain (Fig. 1a–d). Both the ETX-4143 device and the silicone speculum accessory are single-use, disposable components. The primary objective of this exploratory study was to evaluate the early safety and efficacy of the study device in adult subjects with COSP. ETX-4143 device (a). A flexible silicone speculum with two apertures is used to maintain the palpebral fissure (b). The device has two metal contact surfaces that pass through the speculum’s two apertures, contacting the conjunctiva overlying the myelinated long ciliary nerve fibers at the 3 o’clock and 9 o’clock position of the eye (c). Both the speculum and the cooling device have a central aperture in order to maintain visualization of the cornea. The cornea can be visualized after the device is applied (d) Methods This randomized, double-masked, multicenter feasibility study adhered to the tenets of the Declaration of Helsinki and was approved by each study site’s ethics committee. Study participants provided written informed consent prior to initiating any study procedure. The study was approved by St. Vincent’s Hospital (Melbourne, Australia) Human Research Ethics Committee (HREC), which served as the central ethics committee covering all four study sites and registered with ClinicalTrials.gov (NCT06479382). The study was conducted across four clinical sites in Australia. Author affiliations listed on this manuscript reflect institutional appointments per ICMJE authorship criteria and are not limited to study site locations. In Australia, regulatory oversight for this device study was provided through the Therapeutic Goods Administration’s Clinical Trial Notification (CTN) scheme. Inclusion Criteria The study enrolled adult subjects between 22 and 90 years of age with a diagnosis of COSP, defined as symptoms of pain, discomfort, burning, dryness, irritation, foreign body sensation, fatigue, light sensitivity, or grittiness lasting for at least 3 months, with a further inclusion criterion of a score of greater than or equal to 40 out of 100 on at least one visual analog scale (VAS) ocular-related symptom. Exclusion Criteria Key exclusion criteria included a known allergy or hypersensitivity to copper or aluminum; subjects whose COSP symptoms were directly related to contact lens wear; active ocular infection or inflammation; any history of corneal transplant; significant corneal endothelial disease; history of ophthalmic condition or circumstance which could compromise subject safety or study results; any neuropsychiatric disorder or illness that required hospitalization within the previous 6 months. COSP could be present in one eye or both eyes. If both eyes met the inclusion/exclusion criteria, the worse eye was enrolled on the basis of the subject’s subjective reporting and the clinical screening assessments. Importantly, as a result of the severity of their ocular surface disease, enrolled patients were not asked to discontinue existing ocular surface disease therapies (e.g., artificial tears or immunomodulators) and instead were allowed to remain on these therapies. This rigorous study design approach would highlight the pain reduction that was specifically attributable to the ETX-4143 device, and that which could not be achieved with standard of care dry eye therapy. Enrolled eyes were randomized 2:1 to a one-time treatment with the study device (intervention cohort) or to a sham version of the device (sham cohort). Randomization followed a static central randomization process with subjects randomized by means of an interactive web response system, integrated into the electronic data capture platform. An unmasked site member used the allocated randomization number to assign and prepare treatment from the master randomization schedule. Treatment was performed by a masked clinician and was unilateral. Subjects were followed for 8 weeks post-treatment. Follow-up consisted of two in-person visits which were also conducted by masked personnel and six virtual visits for remote data collection. Visit 1 typically consisted of both screening and procedure, though the procedure could be conducted up to 14 days after screening, if necessary. The remaining in-person visits occurred at week 2 and week 8. In-person visits consisted of completing questionnaires, central corneal sensitivity measured with Cochet–Bonnet esthesiometry, best corrected visual acuity (BCVA), IOP, slit lamp biomicroscopy, and evaluation of adverse events. Virtual visits occurred at weeks 1, 3, 4, 5, 6, and 7, and consisted of subjects completing the three questionnaires, and reporting any adverse events. Device Description and Treatment The ETX-4143 is a novel handheld medical device that contains an internal frozen (− 20 °C) mixture of a glycerol solution which does not contain an active pharmaceutical ingredient and can be safely stored in a freezer (Fig. 1). Prior to device application, topical anesthetic is applied, followed by placement of a custom-designed flexible silicone speculum accessory used to open and insulate the eyelids. The device is then guided through the speculum’s protective apertures, allowing the contact surfaces of the device to be positioned at least 1.5 mm posterior to the limbus at 3 o’clock and 9 o’clock, directed to the position of the myelinated fibers of the long ciliary nerves. After correct device positioning, a corneal shield is applied through the central aperture to hydrate the cornea during the cooling treatment. For treatment, all subjects were placed in a reclined position in the examination chair and the eye anesthetized with topical anesthetic drops and anesthetic lidocaine 2% sterile gel applied to the ocular surface. For subjects randomized to the intervention cohort, the device was removed from the freezer (− 18 to − 22 °C) by the unmasked site member just prior to the treatment; for subjects randomized to the sham cohort, the device was removed from the refrigerator (approximately 3 to 5 °C) by the unmasked site member just prior to the treatment. Application time for both intervention and sham cohorts was 4 min. The 4-min application time was chosen on the basis of preclinical cooling-profile testing and early human feasibility work indicating that this duration achieved a sustained conjunctival temperature reduction over the targeted long ciliary nerve region without tissue injury. Shorter durations produced less consistent cooling, whereas longer applications did not further reduce temperature but increased patient discomfort. All treatments were performed during daytime clinic hours, but were not restricted to a specific time of day. Following conclusion of the procedure, the physician rinsed the ocular surface thoroughly with balanced salt solution or an equivalent saline rinse to allow warming of the contact points. The study device and speculum were then removed, and artificial tear gel/ointment was applied to the ocular surface. Follow-up care for both groups consisted of a topical ophthalmic corticosteroid drop (dexamethasone or prednisolone) taken four times a day for 3 days and over-the-counter oral analgesics as needed for any residual discomfort. Efficacy and Safety Assessments As the study was exploratory in nature, the primary outcome was safety and tolerability and there were no pre-specified efficacy endpoints. Although this feasibility trial did not specify a single primary efficacy endpoint, subjects were informed during consent that the study would assess changes in eye pain symptoms and vision-related functioning using clinical measures (corneal sensitivity, BCVA) and validated questionnaires. To guide the design of future clinical trials, three patient-reported outcome (PRO) measures were evaluated to explore changes in patients’ reported subjective pain: a VAS symptom index, the OPAS, and the COP-Q. All three instruments were administered bilaterally. The VAS symptom index evaluated seven ocular symptoms on a 0–100 scale. For each symptom, there was a horizontal 10-cm line marked with verbal descriptors on each end: 0% corresponds to “no pain” and 100% corresponds to “pain as bad as it could possibly be.” Subjects were instructed to place a vertical mark on the line to illustrate the pain magnitude for each symptom. The OPAS, based on a 0–10 scale, was specifically designed to measure ocular pain intensity in patients with eye pain of any origin, has a recall period of 24 h, and includes a “gold standard” question for pain that uses the Wong-Baker FACES® Pain Rating Scale (gold standard pain intensity scale). It has an added dimensionality containing questions regarding quality of life (QoL), aggravating factors, associated factors, and non-eye-related pain, to yield a score for each dimension that may be tracked over the course of a treatment. The COP-Q includes modules designed to assess the severity and frequency of eye pain symptoms at their worst over the previous 4-h and 24-h periods of time. The COP-Q also contains modules to address vision-dependent activities of daily living (ADL) and distal health-related quality of life (HRQoL), similar to the OPAS. The COP-Q assessments are also based on a 0–10 scale. Central corneal sensitivity was measured at each in-person visit with a Cochet–Bonnet esthesiometer. Central corneal sensitivity was measured with the Cochet–Bonnet esthesiometer because it is widely available in clinical practice, has established normative ranges across age groups, and allows direct comparison with prior ocular surface pain and dry-eye studies that used the same instrument. Non-contact esthesiometers can provide more detailed stimulus control but were not available at all participating sites. Safety evaluations included adverse events, BCVA (measured with the site’s illuminated ETDRS chart and converted to logMAR), IOP (applanation or rebound method), and slit lamp ocular examinations. At the screening visit, the three questionnaires and evaluation of corneal sensitivity were completed prior to the instillation of any anesthetic or dilating drops. Following this, subjects were given a topical anesthetic, and the VAS symptom index was re-administered. This comprised the proparacaine anesthetic challenge, which is a clinically validated method for distinguishing between central-dominant and peripheral-dominant pain types [14]. Proparacaine transiently blocks peripheral nociceptive signaling from the cornea and conjunctiva without affecting centrally mediated pain processing [15, 16]. Subjects who experienced a clinically significant reduction in pain (≥ 50%) following proparacaine application were classified as having peripheral-dominant pain, indicating that their symptoms originate primarily from nociceptor activation at the ocular surface. Conversely, those who report minimal or no pain relief (< 50%) likely have centrally mediated pain, suggesting that central sensitization within the trigeminal system is a predominant driver of their symptoms, or a non-ocular surface source of pain [17]. Although both centrally dominant and peripheral-dominant patients were included, we hypothesized that peripheral-dominant patients would be more likely to benefit from intervention targeting the ocular surface nerves, such as cryo-inhibition of the long ciliary nerves, as compared to those with centrally dominant or non-ocular surface sources of pain. Statistical Analysis This was a feasibility study designed to assess initial safety and tolerability of the investigational device. The FDA states that feasibility studies typically include fewer than 50 subjects and do not require formal statistical justification for sample size [18]. In line with these FDA guidelines, no formal hypothesis testing or formal sample size calculation was planned. Similarly, methodological literature and the CONSORT extension for pilot trials support the use of small sample sizes (commonly 20–50 subjects) in early-phase investigations [19]. The assessment of risk in this small sample size feasibility study relative to the beneficial outcomes is evaluated to justify performing larger clinical studies with statistical power. Quantitative variables were summarized descriptively using number of subjects (n), mean, standard deviation, median, minimum, and maximum. Qualitative variables were summarized using counts and percentages. All analyses were two-sided at a significance level of 0.05. Post hoc analyses were performed on the subset of subjects determined to have peripheral-dominant pain. Descriptive analyses were conducted on peripheral-dominant cases to explore patterns of change across specific VAS, OPAS, and COP-Q questions/modules, further refining the understanding of treatment effects within this distinct population. All analyses were performed using SAS Version 9.4 (SAS Institute, Cary, North Carolina, USA). Results A total of 36 subjects were enrolled across four sites in Australia between August and October of 2024. Three enrolled subjects were screen failures, and two subjects were randomized but did not undergo treatment. The screen failures were due to one case of advanced surface disease and two cases of failing to meet the ≥ 40 threshold on any one VAS symptom at baseline. Two subjects were not treated as a result of one subject failing to attend their appointment and another case of not having enough investigational product on hand. Thus, 31 subjects comprised the analysis population; 20 subjects were randomized to the intervention cohort and 11 to the sham cohort. One subject in the sham cohort was lost to follow-up, and 30 subjects completed the study. Table 1 details the baseline demographics for randomized subjects. The intervention and sham cohorts were similar with regards to age, race, laterality of treated eye, and ratio of male to female subjects. All subjects were diagnosed with COSP. Most subjects (70% in the intervention cohort and 55% in the sham cohort) also had an ongoing diagnosis of dry eye disease lasting 5 years or longer with signs of surface disease at the time of screening. There were no contact lens wearers in either group. All patients who were enrolled in this study and had been diagnosed with dry eye disease had undergone standard of care treatment for dry eye disease. Symptoms persisted despite adequate treatment. Other relevant medical history included autoimmune disease and prior corneal refractive surgery. Baseline ocular surface and dry eye parameters are presented in Supplementary Table 1. The intervention group (n = 20) had a higher proportion of participants with severe COSP compared to the sham group (n = 11) (70.0% vs. 36.4%), though this difference was not statistically significant (p = 0.067). Most participants in both groups demonstrated at least grade 1 corneal and conjunctival staining at baseline. Mean corneal staining grades were 1.40 ± 0.63 in the intervention group and 1.75 ± 1.16 in the sham group (p = 0.646). Temporal and nasal conjunctival staining grades were similar between groups (temporal: 1.74 ± 0.87 vs. 1.88 ± 0.99, p = 0.774; nasal: 1.84 ± 0.76 vs. 2.20 ± 0.92, p = 0.284). Conjunctival redness was also comparable between groups (2.14 ± 0.78 vs. 2.40 ± 0.89, p = 0.546). The intervention group had a numerically longer mean duration of COSP symptoms compared to the sham group (12.00 ± 15.76 years vs. 6.18 ± 3.84 years, p = 0.077), though this difference did not reach statistical significance. Table 1Baseline demographicsStatisticIntervention cohort(N = 20)Sham cohort(N = 11)Age (years)n2011Mean (SD)62.3 (17.5)55.1 (14.2)Minimum27.729.7Maximum88.678.7Study eyen2011 Right eye (OD)n (%)11 (55.0%)5 (45.5%) Left eye (OS)n (%)9 (45.0%)6 (54.5%)Sexn2011 Malen (%)4 (20.0%)4 (36.4%) Femalen (%)16 (80.0%)7 (63.6%)Childbearing potentialn (%)2 (12.5%)1 (14.3%)Ethnicityn2011 Not Hispanic or Latinon (%)20 (100%)11 (100%)Racen2011 Whiten (%)18 (90.0%)10 (90.9%) Black/Aboriginaln (%)1 (5.0%)0 Asiann (%)1 (5.0%)1 (9.1%)Relevant ocular & systemic comorbidities Dry eye diseasen (%)14 (70.0%)6 (54.5%) Autoimmune diseasen (%)0 (0%)1a (9.1%)Prior ocular procedures LASIKn (%)2 (10.0%)0 (0%)SD standard deviation, LASIK laser-assisted in situ keratomileusisaA single sham subject had diagnoses of Sjögren’s disease and type I diabetes Baseline demographics SD standard deviation, LASIK laser-assisted in situ keratomileusis aA single sham subject had diagnoses of Sjögren’s disease and type I diabetes Efficacy Although the primary objective of this small pilot clinical trial was safety and tolerability, exploratory efficacy evaluations were evaluated using the PRO measures (VAS symptoms index, OPAS, COP-Q). Treatment effects were estimated using a baseline-adjusted mixed model for repeated measures (MMRM, least-squares mean change-from-baseline). A negative change-from-baseline (CfB) value indicates symptom improvement. Figure 2a shows patients in both intervention and sham cohorts experiencing improvement following treatment, as compared to baseline, for the key query related to eye pain in VAS, OPAS, and COP-Q formats. Initially, minimal change is noted in the early study visits, with differences generally beginning to favor the intervention cohort at the week 2 or later visit. Table 2 presents data for the other evaluated parameters with directionally similar results. a Patient-reported outcomes, in the baseline-adjusted mixed model, for the “eye pain” query for VAS, OPAS, and COP-Q formats of the intent-to-treat population, demonstrating both intervention and sham cohorts experiencing improvement as compared to baseline, with larger improvement experienced by patients randomized to the intervention cohort. b Subgroup analysis of the peripheral-dominant pain population, showing more pronounced treatment effect experienced by patients randomized to the intervention cohort. Statistically significant reduction in the eye pain severity (COP-Q) score is noted at week 8, as compared to sham, (adjusted CfB for intervention − 5.24 vs. sham CfB − 2.86, nominal p = 0.04) Table 2Summary of least-squares mean difference (± standard error) in change-from-baseline (CfB) scores between the intervention and sham cohorts across PRO measuresPROParameterWeek 1Week 2Week 3Week 4Week 5Week 6Week 7Week 8VASBurning/stinging− 12.1 ± 10.2− 13.0 ± 10.2− 10.4 ± 10.2− 6.1 ± 10.2− 8.3 ± 10.4− 2.1 ± 10.4− 9.4 ± 10.6− 7.8 ± 10.5Itching− 1.4 ± 9.51.8 ± 9.57.5 ± 9.54.1 ± 9.53.3 ± 9.61.9 ± 9.69.9 ± 9.89.5 ± 9.8Foreign body sensation− 1.3 ± 10.9− 6.3 ± 10.8− 0.9 ± 10.81.5 ± 10.8− 0.7 ± 11.0− 5.4 ± 11.1− 4.7 ± 11.30.4 ± 11.2Eye discomfort− 0.1 ± 9.82.2 ± 9.80.0 ± 9.82.4 ± 9.8− 0.3 ± 10.04.0 ± 10.0− 1.1 ± 10.25.4 ± 10.1Eye dryness− 2.3 ± 10.1− 4.6 ± 10.00.6 ± 10.03.2 ± 10.0− 3.1 ± 10.26.1 ± 10.2− 0.8 ± 10.43.1 ± 10.4Photophobia3.2 ± 9.112.6 ± 9.17.9 ± 9.11.7 ± 9.1− 0.2 ± 9.24.8 ± 9.32.7 ± 9.43.1 ± 9.4Eye pain− 1.1 ± 10.0− 5.1 ± 10.0− 3.1 ± 10.0− 2.4 ± 10.0− 14.8 ± 10.1− 12.5 ± 10.2− 11.4 ± 10.4− 10.1 ± 10.3Average VAS score− 2.1 ± 6.32− 1.7 ± 6.280.3 ± 6.280.7 ± 6.27− 3.0 ± 6.440.2 ± 6.44− 1.4 ± 6.581.2 ± 6.51OPASOverall eye pain1.9 ± 1.1− 0.7 ± 1.1− 0.5 ± 1.1− 0.6 ± 1.1− 1.4 ± 1.1− 1.6 ± 1.1− 0.9 ± 1.1− 1.0 ± 1.1Average eye pain intensity (24 h)1.9 ± 1.00.4 ± 1.0− 0.4 ± 1.00.1 ± 1.0− 0.4 ± 1.0− 0.6 ± 1.00.1 ± 1.0− 0.6 ± 1.0Average eye pain intensity (2 weeks)1.5 ± 1.01.1 ± 0.9− 0.3 ± 0.9− 0.1 ± 0.9− 0.2 ± 1.0− 0.5 ± 1.0− 0.0 ± 1.0− 0.6 ± 1.0Average non-eye pain1.4 ± 1.00.2 ± 1.00.7 ± 1.0− 0.5 ± 1.0− 0.9 ± 1.00.1 ± 1.00.0 ± 1.00.5 ± 1.0Average QoL1.3 ± 0.90.4 ± 0.90.4 ± 0.90.2 ± 0.9− 0.4 ± 0.9− 0.7 ± 0.9− 0.4 ± 0.9− 0.3 ± 0.9Average of aggravating factors0.1 ± 1.1− 0.9 ± 1.0− 0.5 ± 1.00.8 ± 1.0− 0.7 ± 1.1− 0.7 ± 1.1− 1.4 ± 1.1− 1.3 ± 1.1Average of associated factors1.5 ± 0.81.0 ± 0.81.5 ± 0.81.1 ± 0.80.9 ± 0.80.6 ± 0.80.6 ± 0.91.0 ± 0.8COP-QEye pain severity1.4 ± 1.1− 1.1 ± 1.1− 1.2 ± 1.1− 1.3 ± 1.1− 2.2 ± 1.1− 1.2 ± 1.1− 1.5 ± 1.1− 1.6 ± 1.1Eye pain frequency0.8 ± 0.50.2 ± 0.5− 0.1 ± 0.5− 0.3 ± 0.5− 0.3 ± 0.5− 0.4 ± 0.5− 0.1 ± 0.5− 0.3 ± 0.5Symptom module0.8 ± 0.8− 0.6 ± 0.8− 0.3 ± 0.8− 0.0 ± 0.8− 0.8 ± 0.9− 0.3 ± 0.9− 0.5 ± 0.9− 0.1 ± 0.9ASM0.6 ± 0.8− 0.6 ± 0.8− 0.3 ± 0.8− 0.2 ± 0.8− 0.8 ± 0.8− 1.3 ± 0.8− 0.8 ± 0.9− 0.3 ± 0.9Visual tasking module1.1 ± 0.40.7 ± 0.40.8 ± 0.40.4 ± 0.40.4 ± 0.40.5 ± 0.40.6 ± 0.40.5 ± 0.4Health-related QoL module0.7 ± 0.30.2 ± 0.30.2 ± 0.3− 0.1 ± 0.30.1 ± 0.30.1 ± 0.30.3 ± 0.30.0 ± 0.3ASM Q1: eye pain1.7 ± 1.1− 0.0 ± 1.1− 0.6 ± 1.1− 0.6 ± 1.1− 1.4 ± 1.2− 1.9 ± 1.2− 1.5 ± 1.2− 1.7 ± 1.2Results derived from a within-subject repeated measures model using change-from-baseline as the response variable with week, treatment, and week × treatment interaction as factorsQoL quality of life, ASM alternative symptom module Summary of least-squares mean difference (± standard error) in change-from-baseline (CfB) scores between the intervention and sham cohorts across PRO measures Results derived from a within-subject repeated measures model using change-from-baseline as the response variable with week, treatment, and week × treatment interaction as factors QoL quality of life, ASM alternative symptom module Post hoc analyses were conducted on the subgroup of patients who were peripheral-dominant pain responders, as determined by proparacaine challenge at the procedure visit. This subgroup yielded a cohort of 18 interventional subjects and 7 sham subjects with peripheral-dominant pain. Figure 2b presents results for this post hoc subgroup, with both intervention and sham cohorts again experiencing improvement following treatment, as compared to baseline, for the key query related to pain in VAS, OPAS, and COP-Q formats. Of note, the interventional group showed statistically significant reduction in eye pain severity (COP-Q) scores as compared to sham at week 8 (adjusted CfB for intervention − 5.24 vs. sham CfB − 2.86, nominal p = 0.04). Table 3 presents data for the other evaluated parameters from the PROs with directionally similar results. Table 3Change-from-baseline (CfB) scores between the intervention and sham cohorts across PRO measures in peripheral-dominant proparacaine respondersTreatmentStatisticWeek 1Week 2Week 3Week 4Week 5Week 6Week 7Week 8VAS eye painInterventionn1718181817181517Mean ± SD− 27.06 ± 33.38− 33.89 ± 27.46− 39.39 ± 27.57− 40.17 ± 24.00− 44.82 ± 25.95− 43.50 ± 31.32− 48.53 ± 25.23− 50.76 ± 24.51Shamn77777777Mean ± SD− 25.14 ± 19.91− 24.86 ± 25.18− 36.14 ± 20.84− 38.29 ± 17.43− 35.00 ± 21.31− 31.86 ± 22.40− 36.71 ± 26.63− 41.00 ± 22.15Intervention − shamDiff in means ± SE− 1.92 ± 11.05− 9.03 ± 11.51− 3.25 ± 10.21− 1.88 ± 8.68− 9.82 ± 10.22− 11.64 ± 11.23− 11.82 ± 11.99− 9.76 ± 10.27Pooled SD30.3026.8825.9922.4724.7729.2625.6623.89Effect size− 0.06− 0.34− 0.12− 0.08− 0.40− 0.40− 0.46− 0.41t value− 0.17− 0.78− 0.32− 0.22− 0.96− 1.04− 0.99− 0.95p value0.86430.44780.75510.83140.35320.31590.34500.3598OPAS overall eye painInterventionn1718181817181517Mean ± SD− 1.59 ± 3.14− 3.50 ± 2.81− 4.33 ± 2.93− 4.28 ± 2.74− 4.18 ± 2.30− 4.28 ± 2.93− 5.00 ± 2.30− 4.94 ± 2.79Shamn77777777Mean ± SD− 2.29 ± 2.50− 1.86 ± 2.97− 2.71 ± 2.56− 2.71 ± 2.50− 2.43 ± 2.57− 2.00 ± 2.24− 3.14 ± 2.04− 3.00 ± 2.16Intervention − shamDiff in means ± SE0.70 ± 1.21− 1.64 ± 1.30− 1.62 ± 1.19− 1.56 ± 1.14− 1.75 ± 1.12− 2.28 ± 1.09− 1.86 ± 0.97− 1.94 ± 1.06Pooled SD2.982.852.842.682.382.762.222.64Effect size0.23− 0.58− 0.57− 0.58− 0.74− 0.82− 0.84− 0.74t value0.57− 1.26− 1.36− 1.37− 1.56− 2.09− 1.91− 1.83p value0.57450.23480.19760.19670.14940.05500.07780.0879COP-Q eye pain severity moduleInterventionn1718181817181517Mean ± SD− 1.29 ± 2.80− 3.61 ± 2.85− 4.22 ± 3.12− 4.33 ± 2.52− 4.35 ± 2.69− 4.22 ± 2.82− 5.00 ± 2.39− 5.24 ± 2.46Shamn77777777Mean ± SD− 1.86 ± 2.54− 1.57 ± 2.51− 2.71 ± 2.56− 2.57 ± 2.44− 1.71 ± 3.45− 2.29 ± 1.89− 2.71 ± 2.43− 2.86 ± 2.27Intervention − shamDiff in means ± SE0.56 ± 1.18− 2.04 ± 1.16− 1.51 ± 1.22− 1.76 ± 1.10− 2.64 ± 1.46− 1.94 ± 0.98− 2.29 ± 1.11− 2.38 ± 1.05Pooled SD2.732.772.982.502.922.612.402.41Effect size0.21− 0.74− 0.51− 0.70− 0.90− 0.74− 0.95− 0.99t value0.48− 1.76− 1.24− 1.61− 1.81− 1.98− 2.07− 2.28p value0.64090.10370.23630.13570.10320.06400.06190.0417 Change-from-baseline (CfB) scores between the intervention and sham cohorts across PRO measures in peripheral-dominant proparacaine responders Corneal Sensitivity In the intervention cohort, mean baseline corneal sensitivity measurement was 45.6 (± 12.8) mm (Table 4, with mm converted to g/mm2). Mean sensitivity increased by the week 8 visit to 50.0 (± 7.1) mm. This represented a 4.3-mm improvement from baseline and was found to be statistically significant (nominal p = 0.041). The increasing sensitivity indicates that there was no loss of corneal sensation due to treatment with the device. Instead, corneal sensitivity improved toward normal levels seen in younger, healthy eyes (60 mm) [20]. The sham cohort mean baseline corneal sensitivity was slightly higher with a mean of 48.6 (± 8.8) mm. Corneal sensitivity was also increased at the week 2 and week 8 visits to means of 51.5 (± 8.9) mm and 52.0 (± 5.6) mm, respectively, though neither of these changes was statistically significant. Table 4Change in corneal sensitivity converted from mm to g/mm2CohortVisitnMean(g/mm2)SDDifferencep valueIntervention cohort (N = 20)Baselinea200.8280.5920––Week 2200.6420.3290− 0.1850.017Week 8200.5900.1590− 0.2380.041Sham cohort (N = 11)Baselinea110.6600.2970––Week 2110.5600.1920− 0.0890.294Week 8100.5350.1100− 0.1250.202SD standard deviationaThe baseline value is defined as the last available valid assessment collected prior to first study device administration Change in corneal sensitivity converted from mm to g/mm2 SD standard deviation aThe baseline value is defined as the last available valid assessment collected prior to first study device administration Visual Acuity BCVA was measured at each in-person visit using the site’s illuminated ETDRS chart. Mean (SD) baseline BCVA was 0.073 (± 0.16) logMAR in the intervention cohort. This improved post-treatment to a mean of 0.004 (± 0.12) logMAR at the week 8 visit. Baseline BCVA in the sham cohort was similar, with a mean (SD) of 0.053 (± 0.11) logMAR. Visual acuity also improved after treatment in the sham cohort. At the week 8 visit, BCVA had improved to 0.024 (± 0.09) logMAR in the sham cohort. The improvements for both groups were nominally statistically significant (nominal p < 0.05 for both). IOP was also assessed at each in-person visit. As anticipated, no statistically or clinically significant change in IOP was observed in either group at any point. Safety Treatment with ETX-4143 was associated with a higher incidence of transient ocular AEs compared with sham, predominantly conjunctival hyperemia, edema, and surface staining, all of which resolved without sequelae. Additionally, there were no reports of serious AEs attributable to treatment with the study device, as determined by the principal investigators of the study. Two serious AEs were reported; both were non-ocular in nature and consisted of a hospitalization for a breast lumpectomy and a hospitalization for a rhinovirus infection. No AEs led to treatment reduction or study withdrawal. Increased itching was seen in the intervention group on the VAS, which was driven by two patients with underlying conditions unrelated to the treatment (blepharitis and epiphora). More subjects in the intervention cohort (14; 70%) experienced ocular AEs than in sham cohort (4; 36.4%). All events were mild or moderate, expected, and resolved without intervention or sequelae (Table 5). Table 5Ocular adverse eventsIntervention cohort(N = 20)Sham cohort(N = 11)Number of eventsn (%) of subjectsSeverity n (%)Number of eventsn (%) of subjectsSeverity n (%)MildModerateMildModerateConjunctival hyperemia1410 (50.0%)4 (28.6%)10 (71.4%)11 (9.1%)–1 (100%)Eye pain immediately following treatment116 (30.0%)8 (72.7%)3 (27.3%)11 (9.1%)–1 (100%)Foreign body sensation in eyes43 (15.0%)3 (75.0%)1 (25.0%)00 (0%)––Periorbital pain43 (15.0%)1 (25.0%)3 (75.0%)00 (0%)––Conjunctival edema33 (15.0%)1 (33.3%)2 (66.7%)00 (0%)––Lacrimation increased32 (10.0%)3 (100%)–00 (0%)––Periorbital discomfort22 (10.0%)1 (50.0%)1 (50.0%)00 (0%)––Photophobia22 (10.0%)–2 (100%)00 (0%)––Vision blurred22 (10.0%)1 (50.0%)1 (50.0%)11 (9.1%)1 (100%)–Asthenopia21 (5.0%)2 (100%)–00 (0%)––Blepharitis11 (5.0%)1 (100%)–00 (0%)––Conjunctival hemorrhage11 (5.0%)1 (100%)–00 (0%)––Dry eye11 (5.0%)1 (100%)–00 (0%)––Eye discharge00 (0%)––11 (9.1%)1 (100%)–Eye irritation00 (0%)––11 (9.1%)1 (100%)–Eyelid irritation11 (5.0%)–100 (0%)––Glare11 (5.0%)–1 (100%)00 (0%)––Ocular discomfort11 (5.0%)1 (100%)–00 (0%)––Conjunctivitis11 (5.0%)–1 (100%)00 (0%)––Hordeolum11 (5.0%)1 (100%)–00 (0%)–– Ocular adverse events Slit lamp biomicroscopy was performed at all in-person study visits. Lissamine green and/or fluorescein staining was done to assess any abnormalities of the corneal and conjunctival epithelial surface, using standard Oxford grading criteria. There were no serious findings at any visit in either study group. The most common slit lamp findings at the post-treatment and week 2 examinations were conjunctival hyperemia, conjunctival edema, conjunctival staining, corneal edema, and corneal staining. All findings were recorded as adverse events regardless of whether they were considered expected. With the exception of a single patient, all types of adverse events were seen in both the intervention and sham cohorts. All adverse event-associated signs or symptoms that were present at the week 2 exam had resolved by the next in person exam at week 8. All subjects had baseline findings of ocular surface staining before treatment. Any transient increase in ocular surface staining post-treatment returned to baseline values by the week 8 examination in all cases. One patient was found to have mild blepharitis on both sides (treated and untreated) at the week 8 visit which was not present at the week 2 visit. The finding was classified as unrelated to the device treatment. Discussion In this exploratory pilot trial, a single unilateral ETX-4143 treatment was generally well tolerated and was associated with trends toward greater pain reduction in the frozen-device cohort, particularly among peripheral-dominant pain phenotypes. This study used an active sham condition that included topical anesthetic, a cooled but non-frozen device, and a short course of topical corticosteroids. Early reductions in pain and symptoms in the sham cohort were therefore anticipated, which may have reduced between-group differences during the first 2 weeks. This study used dexamethasone or prednisolone for 3 days post-treatment. Prior peer-reviewed study results have shown a milder corticosteroid (loteprednol etabonate 0.5%) can significantly reduce ocular discomfort within 2 days, with effects lasting up to 2 weeks [21]. Despite the active sham, the intervention cohort showed larger reductions in pain and symptom scores from week 2 through week 8, particularly in participants with peripheral-dominant pain. Across VAS, OPAS, and COP-Q, pain scores decreased more in the intervention cohort than in the sham cohort. Corneal sensitivity in both groups was abnormal and moved towards the normal value (60 mm) after treatment [22]. At week 8, a statistically significant improvement in corneal sensitivity was found for the intervention cohort, but not the sham cohort. This increase in corneal sensitivity after treatment represented a change towards normal corneal sensitivity levels typically seen in younger, healthy eyes [23, 24]. Further work on the mechanism of improved corneal sensitivity is underway; we hypothesize this may reflect repair of chronically injured myelin on ocular surface nerve fibers or a transition from nociceptive nerve fibers back to non-nociceptive nerve fibers. The durability of effects observed after a single 4-min treatment aligns with reports from non-ophthalmic cryotherapy models, in which brief cooling of peripheral nerves is associated with prolonged reductions in nociceptive signaling without permanent structural damage. The sustained reduction in pain and gradual normalization of corneal sensitivity over 8 weeks are consistent with transient modulation of long ciliary nerve function rather than a short-lived anesthetic or corticosteroid effect. BCVA improved in both cohorts, with a nominally statistically significant change observed only in the intervention group. These changes may be related to improved tear-film stability or reduced ocular surface inflammation, but this requires confirmation in future studies. Corneal and ocular surface pain are difficult to quantify, and optimal outcome measures have not been established for COSP [1]. Accordingly, this pilot included multiple efficacy assessments to inform selection of primary endpoints and eligibility criteria for subsequent trials. In particular, the topical anesthetic challenge and “eye pain”-focused items appeared informative for identifying patients who may respond to this intervention. The VAS also served to identify “eye pain”-specific questions as the most informative items for other PRO measures. On the COP-Q, the intervention cohort showed a statistically significant reduction in peripheral-dominant pain severity over the last 4 h at week 8 compared with sham. Longer recall periods (e.g., 24 h) may better capture pain fluctuations in this population. Since the conclusion of this study, the psychometric validation of the COP-Q has been published [25]. The OPAS was determined to be the most appropriate PRO because it has been validated and used in studies of corneal and ocular surface pain [26]. In a literature review of PROs for COSP clinical trials, OPAS ranked highest among 16 candidates meeting inclusion criteria [27]. OPAS also uses a 24-h recall period and includes a widely used “gold standard” pain question (question 1). Across VAS, OPAS, and COP-Q, results consistently indicated reductions in pain exceeding thresholds for meaningful improvement, suggesting that the ETX-4143 system may provide therapeutic benefit for patients with COSP. No serious adverse events, complications, or device-related safety concerns were observed, and all device-related events were mild or moderate and resolved without sequelae. Because treatment involves manipulation of the ocular surface after anesthetic gel application (known to temporarily weaken epithelial tight junctions and increase vulnerability to post-treatment staining), transient epithelial staining was expected; however, slit lamp findings improved by week 2 and week 8 without evidence of persistent epithelial injury [28]. Limitations This study has several limitations, including the small sample size and limited follow-up duration. The enrollment of both centrally and peripheral-dominant pain phenotypes may also have diluted treatment effects, as suggested by the larger improvements observed in the peripheral-dominant subgroup. Larger, longer-term studies that focus on peripheral-dominant pain and include sham controls will be needed to confirm these findings. Conclusions In this exploratory pilot trial, treatment with the ETX-4143 device was associated with reductions in ocular surface pain and improvements in corneal sensitivity and BCVA in adults with COSP, with an acceptable short-term safety profile. These preliminary findings support further evaluation of ETX-4143 in adequately powered, phenotype-stratified trials. Supplementary Information Below is the link to the electronic supplementary material. Supplementary file1 (PDF 380 KB) Acknowledgements The authors thank Saadiah Goolam, MBBS (Westmead Hospital); Luke Chehade, MBBS (Sunshine Eye Surgeons); Brendan Cronin, MBBS (Ophthalmic Trial Australia); and Ben LaHood, MBBS, and Millie Meegan, OD (Innovative Eye Care) for their contributions to study treatments and clinical assessments at the participating sites. Medical Writing/ Editorial Assistance Editorial assistance in the preparation of this article was provided by Laura Fletcher, PhD of Regulatory Pathways Group Inc. Support for this assistance was funded by EyeCool Therapeutics. Author Contributions All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by Anat Galor, Mark Hinds, Pooja Bhindi, Cristos Ifantides, Tomasz Stryjewski, James A. Stefater, and Bonnie Henderson. The first draft of the manuscript was written by Anat Galor and Cristos Ifantides, Tomasz Stryjewski, and James A. Stefater. Christopher Rapuano, Alice Epitropoulos, Jayne Weiss, Stephen Pflugfelder, Preeya Gupta, Deepinder Dhaliwal, Laura Periman, Bonnie An Henderson, Thanh T. Nguyen, Chameen Samarawickrama, and Helen Wu commented on previous versions of the manuscript. All authors read and approved the final manuscript. Funding Sponsorship for this study was provided by EyeCool Therapeutics. EyeCool Therapeutics also provided funding for professional writing support and Rapid Service Fee for publication in Ophthalmology and Therapy. Data Availability The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request. Declarations Conflict of Interest Anat Galor—Alcon Bausch + Lomb, EyeCool (advisor), OKYO. Christopher Rapuano has nothing to disclose. Alice Epitropoulos has nothing to disclose. Jayne Weiss- Regeneron. Stephen Pflugfelder—Alcon, Dompe, Kala, Kowa, Unfold. Preeya Gupta—Azura, Alcon, Aldeyra, Abbvie, Bausch + Lomb, Dompe, HanAll Biopharma, Johnson & Johnson Vision, Kala, Mazado Inc, Nordic Pharma, Ocular Science, Oculis, Orasis, Sight Sciences, Science Based Health, Spyglass, Surface Ophthalmics, Tarsus, Tear Clear, Thea, Tissue Tech, Inc, Trukera, Viatris, Visionology, Vital Tears, Zeiss, EyeCool Therapeutics (advisor). Deepinder Dhaliwal—Staar Surgical, Scope Ophthalmics, Thea, Tarsus, Bausch and Lomb, Aurion, Epion, Lenz, CSI Dry Eye. Laura Periman—Alcon, Alderyra, Allergan/Abbvie, Amgen, Azura, Bausch + Lomb, Bruder, Dompe, Eyedetec, Kala, Lumenis, Mallinckrodt, Myze, Nordic Pharma, Novartis, Nusight Medical, Olympic Ophthalmics, Quench Method, Science Based Health, Scope, Sun, Tarrsus, Thea, Versea, Viatris, Visant. Bonnie Henderson—Eyecool Therapeutics (advisor). Thanh T. Nguyen has nothing to disclose. Chameen Samarawickrama has nothing to disclose. Helen Wu has nothing to disclose. Tomasz Stryjewski—EyeCool Therapeutics (advisor). James A. Stefater—EyeCool Therapeutics (advisor). Mark Hinds—Financial Support: Azura Ophthalmics, Alcon, Alexion Pharmaceuticals, Pelthos Therapeutics, Eli Lilly and Company, EyeCool Technologies, Iolyx Therapeutics, Kiora Pharmaceuticals, Lubris Biopharma, Novo Nordisk Pharmaceuticals, Stuart. Therapeutics, Tarvo Biotherapeutics, TheiaNova. Consultant/Contractor: Azura Ophthalmics, CooperVision, Eyecool Therapeutics, Kiora Pharmaceuticals, SynergEyes USA., Iolyx Therapeutics, Avant Insurance, Wotton Kearney, Remedy Medico-Legal. Pooja Bhindi has nothing to disclose. Cristos Ifantides—Abbvie, Alcon, Bausch + Lomb, EyeCool Therapeutics (advisor), JnJ, Tarsus Pharmaceuticals, Trefoil. Ethical Approval This study adhered to the tenets of the Declaration of Helsinki and was approved by each study site’s ethics committee. Study participants provided written informed consent prior to initiating any study procedure. The study was approved by St. Vincent’s Hospital (Melbourne Australia) Human Research Ethics Committee (HREC) and registered with Clinical Trials.gov (NCT06479382). References Associated Data Supplementary Materials Supplementary file1 (PDF 380 KB) Data Availability Statement The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.