Effects and Safety of Grape Skin Enzyme Fermentation Extract for Eye Fatigue and Dry Eye Symptoms: An Exploratory Randomized Controlled Trial

Abstract

Introduction

Dry eye disease impairs visual function and quality of life. This trial evaluated the efficacy and safety of grape skin enzyme fermentation extract (KL-GEFE; Nunaezone, Kitto Life Co., Ltd., Pyeongtaek, Republic of Korea), a grape skin enzyme fermentation extract, in adults with ocular fatigue and dry-eye-related symptoms.

Methods

In this 12-week, randomized, double-blind, placebo-controlled trial at Dankook University Hospital, 68 adults were allocated 1:1 to KL-GEFE (400 mg/day) or placebo. The study protocol specified the Schirmer I test, adapted Eye Strain Questionnaire (ESQ), noninvasive break-up time (NIBUT), fluorescein tear break-up time (TBUT), ocular-surface staining, critical flicker fusion frequency (CFF), and Ocular Surface Disease Index (OSDI) as efficacy outcomes but did not assign them a formal hierarchy. The protocol-specified efficacy analysis used baseline-adjusted analysis of covariance (ANCOVA) in the per-protocol set (PPS; n = 65), with supportive analysis in the full analysis set (FAS; n = 68) using last observation carried forward (LOCF).

Results

At week 12, the adjusted KL-GEFE-minus-placebo difference was 4.46 mm for the Schirmer I test (95% CI 2.96–5.95; P < 0.001) and −4.29 points for the adapted ESQ total score (95% CI −8.39 to −0.19; P = 0.041). The corresponding FAS–LOCF estimates were 4.36 mm (95% CI 2.92–5.80; P < 0.001) and −4.56 points (95% CI −8.57 to −0.56; P = 0.026), respectively. The week 12 OSDI difference was not statistically significant, and no statistically significant week 12 between-group differences were observed for NIBUT, fluorescein TBUT, CFF, or ocular-surface staining. Two mild adverse events unrelated to the study product occurred; no serious adverse events occurred.

Conclusion

KL-GEFE supplementation was associated with week 12 between-group differences in Schirmer I test values and adapted ESQ total scores, whereas no statistically significant week 12 between-group differences were observed for the other efficacy outcomes. Given the absence of a prospectively defined outcome hierarchy and multiplicity-control framework, these findings should be considered exploratory and require confirmation in prospectively designed trials.

Trial registration: Clinical Research Information Service (CRIS; https://cris.nih.go.kr), KCT0012279, registered on 14 July 2026; retrospectively registered.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s40123-026-01509-8.

Key Summary Points

Why carry out this study?

Dry eye disease and digital eye strain are common, but evidence for convenient oral interventions remains limited.

This study evaluated the effects of 12-week supplementation with grape skin enzyme fermentation extract (KL-GEFE) on ocular-surface measures and symptom scores in adults with ocular fatigue and dry-eye-related symptoms.

What was learned from the study?

KL-GEFE was associated with higher Schirmer I test values than placebo at weeks 4, 8, and 12, with an adjusted mean difference of 4.46 mm at week 12 (95% confidence interval 2.96–5.95; P < 0.001).

At week 12, the adapted Eye Strain Questionnaire score also differed between groups, favoring KL-GEFE, whereas no statistically significant between-group differences were observed for the other efficacy outcomes.

Because the study protocol did not define a formal hierarchy among efficacy outcomes or a multiplicity-control framework, the efficacy findings should be interpreted as exploratory and require confirmation in prospectively designed trials.

Introduction

Dry eye disease (DED) is one of the most prevalent ocular surface disorders worldwide. Epidemiological data indicate that approximately 5–50% of adults are affected, depending on the diagnostic criteria and population studied [1–3]. In South Korea, the prevalence of DED is estimated at approximately 33% in adults aged 65 years and older, with rates rising in younger populations as digital device use and environmental exposures increase [4, 5]. Beyond its high prevalence, DED imposes a considerable economic burden through direct medical costs—including increasingly expensive prescription therapies—and indirect costs attributable to impaired visual function, reduced work productivity, and diminished health-related quality of life [6, 7].

Current pharmacological management of DED relies predominantly on topical agents, including artificial tear substitutes, mucin secretagogues (e.g., diquafosol), and immunomodulatory agents (e.g., cyclosporine A) [8, 9]. While effective, these approaches share common limitations: the inconvenience and low compliance inherent to ocular instillation, the ocular surface toxicity associated with preservatives, and the cumulative cost of chronic use [10, 11]. Many patients, particularly those with mild-to-moderate DED, require multiple concurrent ophthalmic agents to achieve adequate symptom control, further compromising adherence [12, 13].

The increasing prevalence of digital device use has contributed to a growing population with computer-related dry eye symptoms and digital eye strain [14, 15]. Digital eye strain, characterized by reduced blink rate and incomplete blinks during near-vision tasks, promotes tear film destabilization and aqueous tear deficiency, creating an expanding demographic with mild-to-moderate DED and a need for convenient oral interventions [14].

Oral nutraceuticals represent an attractive complementary strategy, offering ease of administration and the potential to address systemic factors contributing to DED, such as oxidative stress and chronic low-grade inflammation [16–19]. Among bioactive compounds with ocular relevance, anthocyanins—polyphenolic pigments abundant in berry fruits including grape skin—have attracted considerable scientific interest. Anthocyanins have shown antioxidant and antiinflammatory activities in experimental systems, including effects on reactive oxygen species generation and inflammatory signaling pathways [20–22]. These experimental findings provide biological hypotheses but do not establish the mechanism of orally administered KL-GEFE in humans. Anthocyanin absorption and metabolism following oral administration have been investigated in pharmacokinetic studies [23–25]; however, the ocular distribution and biological activity of KL-GEFE were not evaluated in the present trial.

Early evidence for the ocular benefits of anthocyanin oligomers derived from grape skin (Vitis vinifera) was provided by Hwang et al. [26], who demonstrated significant antioxidant activity and protection against H₂O₂-induced G₂/M arrest in retinal cells in vitro. Clinical evidence followed, with Lee et al. [27] reporting improvements in nocturnal visual function in myopic subjects following supplementation with purified anthocyanin oligomers (800 mg/day). In the context of DED specifically, a randomized, double-blind, placebo-controlled trial by Riva et al. [28] using a standardized bilberry anthocyanin extract (Mirtoselect®) demonstrated significant improvements in DED symptoms over a 3-month treatment period. More recent clinical evidence for grape-skin-derived anthocyanin oligomers was provided by Fan et al. [29], who reported in a randomized, double-blind, placebo-controlled study of 108 patients with mild-to-moderate DED that oral anthocyanin oligomers (800 mg/day) improved tear break-up time, ocular surface disease index (OSDI), symptom measures, and selected clinical indices without treatment-related safety concerns.

KL-GEFE (Nunaezone; Kitto Life Co., Ltd.) is a grape skin enzyme fermentation extract standardized for anthocyanin oligomer content. The fermentation process enhances bioavailability relative to conventional grape skin extracts, and KL-GEFE has completed individual recognition as a functional food ingredient by the Korean Ministry of Food and Drug Safety (MFDS) for the health claim of improving dry eye (ingredient recognition no. 2019-27), under the MFDS eye-health functional evaluation framework [30]. Preclinical studies have reported antioxidant and antiinflammatory activity of KL-GEFE in experimental ocular models. The present trial aimed to compare the effects of once-daily KL-GEFE (400 mg/day) and placebo on prespecified ocular-surface measures and symptom scores over 12 weeks in adults with ocular fatigue and dry-eye-related symptoms.

Methods

Study Design and Setting

This was a 12-week, randomized, double-blind, placebo-controlled clinical trial conducted at the Department of Ophthalmology, Dankook University Hospital, Cheonan, Republic of Korea. Participants were evaluated at four scheduled visits: V1 (baseline/week 0), V2 (week 4 ± 7 days), V3 (week 8 ± 7 days), and V4 (week 12 ± 7 days). This study was approved by the Institutional Review Board of Dankook University Hospital (DKUH IRB 2025-02-007) and was conducted in accordance with the Declaration of Helsinki and the Korean Good Clinical Practice Guidelines. Written informed consent was obtained from all participants before enrollment. The trial was retrospectively registered with the Clinical Research Information Service (CRIS; KCT0012279) on 14 July 2026, after participant enrollment and study completion, because the investigators had misunderstood the timing requirement for clinical trial registration and believed that registration before manuscript submission was sufficient after institutional ethics approval. To mitigate concerns arising from retrospective registration, the manuscript clearly distinguishes protocol-specified analyses from post hoc analyses, explicitly states that the study protocol did not define a formal hierarchy among efficacy outcomes or a multiplicity-control framework, and consistently interprets the efficacy findings as exploratory rather than confirmatory. The registered study title was “A placebo-controlled, blinded study to evaluate the functionality and safety of grape skin enzyme fermentation extract (KL-GEFE, Nunaezone) on improving eye fatigue and dryness.”

Participants

Eligible participants were adults aged 18 years or older who reported ocular fatigue and dry-eye-related symptoms, including foreign-body sensation, ocular pain, itching, visual disturbance, or photophobia. Participants were also required to have no clinically evident organic ocular-surface abnormalities, including conjunctival hyperemia, clinically evident corneal epithelial erosion assessed by slit-lamp examination with fluorescein staining, or clinically apparent meibomian-gland abnormalities, together with a Schirmer I test value ≥ 5 mm and a fluorescein tear break-up time (TBUT) ≤ 10 s at screening. Clinically evident corneal epithelial erosion was assessed by slit-lamp examination with fluorescein staining. Although a numerical NEI cutoff was not explicitly specified in the written study protocol, the screening assessment was guided by the NEI/Industry Workshop grading framework and established dry-eye clinical-trial severity conventions. Corneal staining with a total NEI score ≥ 10 (of 15) or a score of 3 in any individual corneal region was considered clinically substantial epithelial involvement and was excluded. Lower-grade punctate fluorescein staining that did not meet these criteria was permitted and was quantified using the NEI scale as an efficacy outcome. These criteria were intended to identify participants with symptomatic tear-film instability without marked aqueous tear deficiency or clinically apparent meibomian-gland or ocular-surface disease. This eligibility framework was intended to reduce confounding from overt ocular-surface pathology or clinically apparent meibomian-gland disease while allowing for evaluation of changes in tear-film and symptom-related outcomes in a relatively homogeneous symptomatic population.

Eligibility

Key exclusion criteria included a history of allergic conjunctivitis or rhinitis; ophthalmic surgery within 6 months before V1; contact-lens use within 1 month before V1; intraocular pressure > 21 mmHg; and use within 1 month before V1 of artificial tears or other dry-eye treatments, diuretics, medications for autoimmune or thyroid disease, antihistamines, beta blockers, tricyclic antidepressants, or estrogen-containing hormonal therapies. Participants who had used eye-health supplements potentially affecting the study outcomes—including eicosapentaenoic acid/docosahexaenoic acid (EPA/DHA)-containing products, marigold extract, bilberry extract, lutein- or zeaxanthin-containing preparations, or Perilla frutescens extract—within 1 week before V1 were also excluded. Additional exclusion criteria included uncontrolled hypertension (systolic blood pressure [SBP] ≥ 160 or diastolic blood pressure [DBP] ≥ 100 mmHg), pregnancy or lactation, and participation in another interventional clinical trial within 3 months.

Concomitant Treatment

Throughout the 12-week intervention period, participants were instructed not to use artificial tears or initiate any other topical, systemic, or procedural treatment specifically intended to improve dry-eye-related signs or symptoms. They were also instructed not to initiate or resume the protocol-prohibited medications, including diuretics, medications for autoimmune or thyroid disease, antihistamines, beta blockers, tricyclic antidepressants, and estrogen-containing hormonal therapies. Use of potentially relevant eye-health supplements—including EPA/DHA-containing products, marigold extract, bilberry extract, lutein- or zeaxanthin-containing preparations, and Perilla frutescens extract—was prohibited. Contact-lens use and ophthalmic surgical procedures or related ocular treatments were also prohibited during the intervention period.

Intervention and Comparator

Participants randomized to the test group received one KL-GEFE capsule (400 mg/day; Nunaezone, Kitto Life Co., Ltd., Pyeongtaek, Republic of Korea) once daily with water for 12 weeks. The active capsule was a medium-brown hard gelatin capsule containing 400 mg of grape skin enzyme fermentation extract. The KL-GEFE used in this trial was from clinical batch AO202504. Cyanidin was used as the quantitative batch-release marker for the finished product. The labeled cyanidin content was 34.2 mg/g, with a predefined acceptance criterion of 80–120% of the labeled amount (27.36–41.04 mg/g). The measured cyanidin content of clinical batch AO202504 was 35.51 mg/g, corresponding to 103.8% of the labeled amount. Cyanidin was quantified by reverse-phase high-performance liquid chromatography using an Agilent 1260 Infinity system (Agilent Technologies, Santa Clara, CA, USA) equipped with a Capcell Pak UG120 C18 column (4.6 × 250 mm, 5 μm; Shiseido, Tokyo, Japan) and ultraviolet-visible (UV/Vis) detection at 530 nm. Batch-release quality control was performed according to the manufacturer’s Certificate of Analysis, including compositional, applicable physicochemical, heavy-metal, and microbiological specifications, and the clinical batch met the predefined release criteria. Participants randomized to the control group received an identical-appearing placebo capsule containing maltodextrin (400 mg/day). Test and placebo products were indistinguishable in appearance, color, and weight. Blinding was maintained throughout the trial for participants, investigators, and outcome assessors.

Efficacy Outcomes

Prospectively specified efficacy outcomes included the Schirmer I test, Eye Strain Questionnaire (ESQ) total score, noninvasive keratograph break-up time (NIBUT), fluorescein tear break-up time (TBUT), critical flicker fusion frequency (CFF), corneal staining score, conjunctival staining score, and Ocular Surface Disease Index (OSDI) total score. The original study protocol did not specify a formal hierarchy among these efficacy outcomes or designate primary and secondary efficacy endpoints. Other efficacy outcomes included noninvasive keratograph break-up time (NIBUT; first break-up time [NIKf-BUT] and average break-up time [NIKav-BUT], participant-level mean of both eyes) [31, 32], fluorescein tear break-up time (TBUT; participant-level mean of both eyes), critical flicker fusion frequency (CFF; not-seeing-to-seeing [NSTS] and seeing-to-not-seeing [STNS] directions), corneal staining score (NEI grading scale; participant-level mean of both eyes), conjunctival staining score (participant-level mean of both eyes), and Ocular Surface Disease Index (OSDI) total score [33]. Individual ESQ item scores and OSDI category- and item-level analyses were exploratory.

The ESQ was a 10-item Korean-language questionnaire adapted from the symptom questionnaire developed by Ames et al. [34]. Each item was rated from 0 to 6 (0, none; 1–2, slight; 3–4, moderate; 5–6, severe); yielding a total score from 0 to 60; higher scores indicate greater symptom burden. No ESQ item-level responses were missing. The OSDI is a validated 12-item instrument yielding scores from 0 to 100; higher scores indicate greater severity [33]. OSDI severity categories were defined as normal (0–12), mild (13–22), moderate (23–32), and severe (33–100) [33]. The Schirmer I test was performed without topical anesthesia using a standardized Schirmer strip placed in the lateral one third of the lower conjunctival sac. After 5 min, the length of the blue-wetted portion of the strip was measured in millimeters as an estimate of tear secretion [31]. Fluorescein TBUT was measured after fluorescein application under cobalt-blue slit-lamp illumination. The interval between the last complete blink and the first observed break in the tear film was recorded in seconds. Three consecutive measurements were obtained for each eye and averaged. NIBUT was assessed separately without fluorescein using the Keratograph 5M (Oculus Optikgeräte GmbH, Wetzlar, Germany). NIKf-BUT represented the time from the last complete blink to the first detected tear-film break, whereas NIKav-BUT represented the average of the detected tear-film break-up times generated by the device. Measurements were obtained separately for each eye. For bilateral ophthalmic endpoints, right- and left-eye measurements were averaged within each participant at each visit, and the participant-level bilateral mean was used in the efficacy analyses; no study eye was designated.

Safety Assessments

Safety was evaluated throughout the study via adverse event (AE) monitoring, vital signs (systolic blood pressure [SBP], diastolic blood pressure [DBP], pulse, and body weight), and ophthalmic safety examinations including intraocular pressure (IOP), best-corrected visual acuity (BCVA), funduscopy, and slit-lamp examination. All safety data were analyzed in the safety analysis set (SAS, N = 68).

Randomization and Blinding

Participants were randomly assigned in a 1:1 ratio to the KL-GEFE or placebo group using block randomization based on a pregenerated randomization sequence created with IBM® SPSS® Statistics before trial initiation. Study products were labeled and dispensed according to the assigned randomization codes, and allocation information was concealed from participants, investigators, clinical coordinators, and outcome assessors throughout the trial.

Sample Size Calculation

The study was designed as an exploratory randomized trial evaluating KL-GEFE at a reduced dose of 400 mg/day. The target sample size of 68 participants was prospectively determined before enrollment with reference to the previous randomized trial by Fan et al. [29], as documented in the written pre-enrollment protocol. The protocol identified Fan et al. [29] as the basis for the sample-size determination but did not reproduce the detailed numerical assumptions or their derivation. During study planning, the sample-size framework used in that trial and related clinical studies [35, 36] informed the numerical assumptions used for the calculation, including an anticipated between-group difference of approximately 5.0 mm and SD of 6.7 mm, corresponding to a standardized effect size of approximately 0.74. With a two-sided significance level of 0.05, 80% power, and a 1:1 allocation ratio, approximately 30 evaluable participants were required per group. Allowing for an anticipated dropout rate of 13%, the target sample size was set at 34 participants per group, for a total of 68 participants. Because the study protocol did not define a formal hierarchy among efficacy outcomes, this sample-size framework was not intended to establish a confirmatory primary efficacy endpoint. The planned sample size was also intended to support the exploratory estimation of treatment effects and variability to inform the design of a future confirmatory trial [37, 38].

Statistical Analysis

The protocol-specified primary analysis population was the per-protocol set (PPS; test n = 33, control n = 32), defined using prespecified exclusion criteria as participants who completed the study without major protocol deviations. Supportive analyses were conducted in the full analysis set (FAS; n = 68), which included all randomized participants analyzed according to their assigned treatment group. Three participants withdrew consent after completing at least one post-baseline assessment; no participant had baseline data only. For the FAS analyses, the last available post-baseline value was carried forward to subsequent missing visits using the protocol-specified last-observation-carried-forward (LOCF) method. Between-group differences were assessed using analysis of covariance (ANCOVA), with change from baseline as the dependent variable, treatment group as a fixed effect, and the corresponding baseline value as a covariate. For bilateral ophthalmic endpoints, participant-level means of the right- and left-eye values were used. Adjusted between-group mean differences and 95% confidence intervals were reported. Because the original protocol did not explicitly define a formal outcome hierarchy or a multiplicity-control framework, all efficacy analyses were interpreted as exploratory rather than confirmatory.

Other efficacy analyses included OSDI total score, NIBUT, fluorescein TBUT, CFF, corneal staining, and conjunctival staining. Week 4 and week 8 analyses, within-group changes, ESQ item-level analyses, and OSDI category and item-level analyses were supportive or exploratory. Normality was assessed using the Shapiro–Wilk test; within-group changes were evaluated using paired t-tests or Wilcoxon signed-rank tests, as appropriate. Categorical data were analyzed using chi-squared or Fisher exact tests. No multiplicity adjustment was applied across the efficacy outcomes; nominal P-values were therefore interpreted descriptively. Statistical analyses were performed using SPSS software (version 24; IBM Corp., Armonk, NY, USA).

Post Hoc Robustness Analyses

Additional post hoc robustness analyses were performed in the FAS without LOCF imputation. Separate mixed models for repeated measures (MMRMs) were fitted for Schirmer OU and ESQ total score using observed week 4, 8, and 12 measurements, with baseline included as a covariate. Fixed effects included treatment, visit, treatment-by-visit interaction, baseline, and baseline-by-visit interaction. Within-participant residual covariance was modeled using an unstructured covariance matrix, with REML estimation and Satterthwaite degrees of freedom. For bilateral ophthalmic outcomes, supplementary eye-level linear mixed-effects models included random intercepts for participant and eye nested within participant.

Results

Participant Disposition and Baseline Characteristics

A total of 70 adults were screened, of whom 2 did not meet eligibility criteria (Fig. 1; CONSORT flow diagram). The remaining 68 participants were randomized 1:1 to the test (n = 34) or control group (n = 34). During the study period, three participants withdrew consent (control: 2; test: 1), all before week 8, yielding a PPS of 65 (test n = 33, control n = 32; Table 1).

Fig. 1.

CONSORT flow diagram. A total of 70 participants were assessed for eligibility, of whom 68 were randomized 1:1 to the KL-GEFE group (n = 34) or placebo group (n = 34). CONSORT, Consolidated Standards of Reporting Trials; SAS, safety analysis set; FAS, full analysis set; PPS, per-protocol set

Table 1.

Subject disposition summary

Category	Control (n = 34)	Test (n = 34)	Total
Screened	35	35	70
Screen failure	1	1	2
Randomized	34	34	68
Withdrew consent	2	1	3
Completed study	32	33	65
Safety analysis set (SAS)	34	34	68
Full analysis set (FAS)	34	34	68
Per-protocol set (PPS)—primary analysis	32	33	65

Values are n. SAS, safety analysis set; FAS, full analysis set; PPS, per-protocol set

Baseline demographic and clinical characteristics were generally comparable between groups (Table 2). Mean age was 44.7 ± 10.5 years in the control group and 42.8 ± 10.1 years in the test group. Of the 68 randomized participants, 58 (85.3%) were female and 10 (14.7%) were male, with identical sex distributions in the two treatment groups. Race and ethnicity were not collected as part of the prespecified demographic assessments and therefore could not be reported. Mean daily digital device use was 4.8 ± 3.0 h in the control group and 5.2 ± 3.0 h in the test group.

Table 2.

Baseline demographic and clinical characteristics

Variable	Control (n = 34)	Test (n = 34)	P-value
Age, years (mean ± SD)	44.74 ± 10.53	42.79 ± 10.11	0.441
Sex, female, n (%)	29 (85.3%)	29 (85.3%)	1.000
Sex, male, n (%)	5 (14.7%)	5 (14.7%)	–
Height, cm (mean ± SD)	163.10 ± 6.15	164.34 ± 6.02	0.512
Body weight, kg (mean ± SD)	61.34 ± 12.50	61.92 ± 10.00	0.411
SBP, mmHg (mean ± SD)	118.00 ± 12.29	120.03 ± 8.79	0.436
DBP, mmHg (mean ± SD)	71.21 ± 9.71	72.65 ± 6.74	0.480
Pulse, bpm (mean ± SD)	74.44 ± 6.13	75.18 ± 6.13	0.575
Sleep, h/day (mean ± SD)	6.38 ± 1.10	6.71 ± 1.09	0.173
Digital device use, h/day (mean ± SD)	4.76 ± 3.04	5.24 ± 2.99	0.338
Medical history present, n (%)	10 (29.4%)	12 (35.3%)	–
Concomitant medication, n (%)	5 (14.7%)	8 (23.5%)	–

Values are mean ± standard deviation or n (%). Between-group P-values were calculated using the independent t-test for continuous variables or chi-squared test for categorical variables. SD, standard deviation; SBP, systolic blood pressure; DBP, diastolic blood pressure

Compliance

Mean compliance rates exceeded 93% in both groups at all visits (Table 3), with no significant between-group differences. These rates confirm adequate adherence to the assigned capsule regimen throughout the 12-week intervention.

Table 3.

Compliance rates by visit and group

Visit	Control (n)	Control mean ± SD (%)	Test (n)	Test mean ± SD (%)
Week 4 (V2)	34	95.2 ± 7.1	34	93.4 ± 7.3
Week 8 (V3)	32	94.7 ± 7.6	33	95.8 ± 6.6
Week 12 (V4)	32	96.2 ± 4.9	33	97.6 ± 3.4

Compliance (%) = number of capsules taken / number of capsules dispensed × 100. SD, standard deviation

Efficacy Results

Schirmer I Test

Between-group differences in Schirmer I test values favored KL-GEFE at weeks 4, 8, and 12. In the KL-GEFE group, mean Schirmer values increased from 8.74 ± 4.14 mm at baseline to 10.48 ± 5.00 mm at week 4, 10.83 ± 4.66 mm at week 8, and 10.97 ± 4.25 mm at week 12. In the placebo group, mean Schirmer values decreased from 8.36 ± 3.71 mm at baseline to 6.22 ± 2.64 mm at week 8 and 6.31 ± 2.84 mm at week 12. Baseline-adjusted between-group differences favored KL-GEFE at week 4 (2.31 mm; 95% CI 1.05–3.56), week 8 (4.36 mm; 95% CI 2.98–5.74), and week 12 (4.46 mm; 95% CI 2.96–5.95; P < 0.001) (Table 4).

Table 4.

Schirmer I test results by visit and treatment group based on participant-level bilateral means

Visit	Group	Mean ± SD (mm)	Change from BL (mean ± SD)	Within-group P	ANCOVA P (between)
Baseline (V1)	Control	8.36 ± 3.71	–	–	–
	Test	8.74 ± 4.14	–	–	–
Week 4 (V2)	Control	7.86 ± 2.94	− 0.50 ± 2.73	0.309	< 0.001***
	Test	10.48 ± 5.00	+1.74 ± 2.45	< 0.001***
Week 8 (V3)	Control	6.22 ± 2.64	−2.14 ± 2.39	< 0.001***	< 0.001***
	Test	10.83 ± 4.66	+2.09 ± 3.58	< 0.001***
Week 12 (V4)	Control	6.31 ± 2.84	−2.05 ± 3.43	< 0.001***	< 0.001***
	Test	10.97 ± 4.25	+2.23 ± 3.63	0.001**

BL, baseline (V1). Values are participant-level bilateral means of the right- and left-eye measurements. Within-group comparisons used the paired t-test or Wilcoxon signed-rank test, as appropriate. Between-group comparisons used ANCOVA with change from baseline as the dependent variable and baseline value as a covariate. **P < 0.01; ***P < 0.001. SD, standard deviation; ANCOVA, analysis of covariance

Adapted Eye Strain Questionnaire

The test group showed a progressive reduction in ESQ total score from baseline (17.42 ± 12.89) through week 4 (11.67 ± 10.98; Δ = −5.76 ± 7.60; P < 0.001), week 8 (11.00 ± 9.66; Δ = −6.42 ± 7.88; P < 0.001), and week 12 (10.03 ± 9.97; Δ = −7.39 ± 9.27; P < 0.001). The control group also improved, but to a lesser degree (baseline, 18.47 ± 12.53; week 12, 15.03 ± 13.65; Δ = −3.44 ± 9.01; P = 0.023). At week 12, ANCOVA demonstrated a between-group difference favoring KL-GEFE in ESQ total score (adjusted mean difference, −4.29 points; 95% CI −8.39 to −0.19; P = 0.041; Table 5).

Table 5.

Adapted eye strain questionnaire (ESQ) total score by visit and treatment group

Visit	Group	Mean ± SD	Change from BL (mean ± SD)	Within-group P	ANCOVA P (between)
Baseline (V1)	Control	18.47 ± 12.53	–	–	–
	Test	17.42 ± 12.89	–	–	–
Week 4 (V2)	Control	14.16 ± 12.52	−4.31 ± 6.73	0.001**	0.307
	Test	11.67 ± 10.98	−5.76 ± 7.60	< 0.001***
Week 8 (V3)	Control	13.47 ± 12.62	−5.00 ± 9.52	0.002**	0.347
	Test	11.00 ± 9.66	−6.42 ± 7.88	< 0.001***
Week 12 (V4)	Control	15.03 ± 13.65	−3.44 ± 9.01	0.023*	0.041*
	Test	10.03 ± 9.97	−7.39 ± 9.27	< 0.001***

BL, baseline (V1). Within-group comparisons used the paired t-test or Wilcoxon signed-rank test, as appropriate. Between-group comparisons used ANCOVA with change from baseline as the dependent variable and baseline value as a covariate. *P < 0.05; **P < 0.01; ***P < 0.001. SD, standard deviation; ANCOVA, analysis of covariance

The protocol-specified supportive FAS–LOCF analysis yielded estimates similar in direction and magnitude to those from the protocol-specified PPS analysis. At week 12, the adjusted between-group difference in change from baseline was 4.36 mm for the Schirmer I test (95% CI 2.92–5.80; P < 0.001) and −4.56 points for the ESQ total score (95% CI −8.57 to −0.56; P = 0.026). Both estimates were similar in direction and magnitude to the corresponding PPS estimates.

Robustness Analyses

In the post hoc FAS–MMRM using observed data without LOCF, the adjusted week 12 treatment difference remained significant for Schirmer OU (+4.48 mm; 95% CI 2.98–5.97; P < 0.0001) and ESQ total score (−4.65 points; 95% CI −8.74 to −0.56; P = 0.0266) (Supplementary Table S1). In the supplementary eye-level LMM accounting for within-participant inter-eye correlation, the week 12 Schirmer difference was +4.51 mm (95% CI 3.21–5.81; P < 0.0001). No significant week 12 eye-level treatment differences were observed for NIKf-BUT, NIKav-BUT, fluorescein TBUT, corneal staining, or conjunctival staining (Supplementary Table S2).

Ocular Surface Disease Index

At week 12, the adjusted between-group difference in OSDI total score was −7.22 points (95% CI −14.76 to 0.32; nominal P = 0.060), and the confidence interval included the null value (Table 6). Mean OSDI scores decreased from 31.85 ± 18.44 at baseline to 20.36 ± 16.42 at week 12 in the KL-GEFE group and from 35.84 ± 18.79 to 30.38 ± 22.87 in the placebo group. At week 4, the adjusted between-group difference was −7.72 points (95% CI −14.17 to −1.27; nominal P = 0.020). This intermediate-visit analysis was exploratory and was not adjusted for multiplicity. At week 12, the adjusted difference was −7.22 points (95% CI −14.76 to 0.32; nominal P = 0.060).

Table 6.

Ocular surface disease index (OSDI) total score by visit and treatment group

Visit	Group	Mean ± SD	Change from BL (mean ± SD)	Within-group P	ANCOVA P (between)
Baseline (V1)	Control	35.84 ± 18.79	–	–	–
	Test	31.85 ± 18.44	–	–	–
Week 4 (V2)	Control	31.16 ± 21.71	−4.69 ± 13.98	0.067	0.020*
	Test	20.55 ± 15.00	−11.30 ± 13.67	< 0.001***
Week 8 (V3)	Control	27.78 ± 19.24	−8.06 ± 13.15	0.002 **	0.235
	Test	21.55 ± 15.22	−10.30 ± 12.28	< 0.001***
Week 12 (V4)	Control	30.38 ± 22.87	−5.47 ± 17.09	0.080	0.060
	Test	20.36 ± 16.42	−11.48 ± 14.89	< 0.001***

Within-group comparisons used the paired t-test or Wilcoxon signed-rank test. Between-group comparisons used ANCOVA with change from baseline as the dependent variable and baseline as a covariate. P-values were not adjusted for multiplicity and are descriptive. SD, standard deviation; ANCOVA, analysis of covariance

Other Efficacy Outcomes

Week 12 results for all efficacy outcomes are summarized in Table 7. No statistically significant week 12 between-group differences were detected for NIBUT (NIKf-BUT: adjusted mean difference, 0.14 s; P = 0.797; NIKav-BUT: 0.16 s; P = 0.759), fluorescein TBUT (−0.18 s; P = 0.546), CFF (NSTS: −0.38 Hz; P = 0.259; STNS: 1.01 Hz; P = 0.161), corneal staining score (NEI: 0.02; P = 0.898), or conjunctival staining score (0.10; P = 0.203). Because no multiplicity adjustment was applied across the efficacy outcomes, the nominal P-values were interpreted descriptively. Both groups showed within-group reductions in corneal staining scores from baseline at all post-baseline visits (all P ≤ 0.001). CFF STNS increased within the test group at all visits, but the week 12 between-group difference was not statistically significant.

Table 7.

Summary of efficacy outcomes at week 12

Endpoint	Control Δ (mean ± SD)	Test Δ (mean ± SD)	Adj. mean diff. (95% CI)	ANCOVA P
Schirmer I test, bilateral mean (mm)	−2.05 ± 3.43	+2.23 ± 3.63	+4.46 (2.96, 5.95)	< 0.001***
Adapted ESQ total score	−3.44 ± 9.01	−7.39 ± 9.27	−4.29 (−8.39, −0.19)	0.041*
OSDI total score	−5.47 ± 17.09	−11.48 ± 14.89	−7.22 (−14.76, 0.32)	0.060
NIBUT NIKf-BUT, bilateral mean (sec)	−0.19 ± 2.57	−0.01 ± 3.00	0.14 (−0.92, 1.20)	0.797
NIBUT NIKav-BUT, bilateral mean (sec)	−0.27 ± 2.33	−0.25 ± 3.69	0.16 (−0.90, 1.22)	0.759
TBUT, bilateral mean (sec)	+0.58 ± 1.49	+0.42 ± 1.25	−0.18 (−0.75, 0.40)	0.546
Corneal staining, bilateral mean	−1.06 ± 1.33	−1.14 ± 1.22	0.02 (−0.24, 0.28)	0.898
Conjunctival staining, bilateral mean	−0.44 ± 0.49	−0.27 ± 0.38	0.10 (−0.06, 0.26)	0.203
CFF NSTS (Hz)	+0.48 ± 1.61	+0.23 ± 1.59	−0.38 (−1.04, 0.29)	0.259
CFF STNS (Hz)	+1.24 ± 3.19	+2.41 ± 3.47	+1.01 (−0.41, 2.43)	0.161

Adjusted mean differences and 95% confidence intervals were estimated using ANCOVA with change from baseline as the dependent variable and baseline value as a covariate in the PPS. Bilateral ophthalmic values are participant-level means of right- and left-eye measurements. No multiplicity adjustment was applied across the efficacy outcomes; nominal P-values should therefore be interpreted descriptively. SD, standard deviation; CI, confidence interval; ANCOVA, analysis of covariance; PPS, per-protocol set; ESQ, Eye Strain Questionnaire; OSDI, Ocular Surface Disease Index; NIBUT, noninvasive break-up time; NIKf-BUT, noninvasive keratograph first break-up time; NIKav-BUT, noninvasive keratograph average break-up time; TBUT, tear break-up time; CFF, critical flicker fusion frequency; NSTS, not seeing to seeing; STNS, seeing to not seeing

Safety

Two adverse events were reported during the 12-week study period, both mild and unrelated to the study product (Table 8): one case of acute upper respiratory infection in the test group (recovered in 3 days) and one case of plantar fasciitis in the control group (recovered in 8 days). No serious adverse events (SAEs) occurred in either group throughout the study.

Table 8.

Adverse events

Group	Subject	Adverse event	Start	End	Severity	Causality
Test	KL-S034	Acute upper respiratory infection	2025-09-19	2025-09-21	Mild	Not related
Control	KL-S038	Plantar fasciitis	2025-11-24	2025-12-01	Mild	Not related

No adverse event was considered related to the study product. AE, adverse event

Vital signs (SBP, DBP, pulse, and body weight) remained within normal ranges throughout the study in both groups, with no clinically meaningful changes from baseline. Ophthalmic safety examinations at weeks 0 and 12 (IOP, BCVA, funduscopy, and slit-lamp examination) revealed no clinically significant abnormalities in either group. No relevant safety signals were observed during the 12-week study.

Discussion

This randomized, double-blind, placebo-controlled trial evaluated the efficacy and safety of 12-week oral supplementation with KL-GEFE in adults with ocular fatigue and dry eye symptoms. At week 12, between-group differences favored KL-GEFE for both Schirmer-measured tear secretion (adjusted mean difference, 4.46 mm) and ESQ total score (adjusted mean difference, −4.29 points) in the PPS analysis. Supportive FAS analyses using LOCF yielded similar estimates for Schirmer (4.36 mm) and ESQ (−4.56 points). Other efficacy outcomes, including tear-film stability, CFF, ocular surface staining, and week 12 OSDI, did not demonstrate statistically significant between-group differences. No relevant safety signals were observed during the 12-week study.

The Schirmer I test showed between-group differences across follow-up visits, and the adapted ESQ showed a between-group difference at week 12. However, the clinical and biological significance of these findings remains uncertain because improvements were not consistently observed across the other efficacy outcomes [39]. Schirmer I values increased from baseline in the KL-GEFE group but decreased in the placebo group over follow-up. At week 12, the mean change was +2.23 mm in the KL-GEFE group and −2.05 mm in the placebo group. The reason for the decline in the placebo group is uncertain; therefore, the observed week 12 between-group difference should be interpreted as reflecting changes in both groups rather than an increase in the KL-GEFE group alone. Although the Schirmer I test is known to have measurement variability, particularly when performed without anesthesia, between-group differences were observed across follow-up visits [39]. However, these findings were not accompanied by corresponding between-group differences in tear-film stability, ocular-surface staining, or week 12 OSDI. Accordingly, the Schirmer finding should be interpreted cautiously and does not establish a specific aqueous-tear or tear-film mechanism.

Prior experimental and clinical evidence provides biological context for the observed Schirmer finding, particularly through reported links between oxidative stress, inflammation, and ocular surface or lacrimal dysfunction. Experimental dry eye models have shown that oxidative stress can contribute to corneal epithelial alterations and ocular surface injury [40], while inflammatory pathways are recognized as clinically important drivers of dry eye disease severity and persistence [41]. Anthocyanins have antioxidant and antiinflammatory properties, and grape-skin-derived anthocyanin oligomers have demonstrated cytoprotective activity in retinal cell models [20–26]. Clinical studies of anthocyanin-containing preparations have also suggested ocular benefits, including improvements in visual symptoms, dry eye symptoms, and tear-related parameters [27–29]. Within this context, the current trial extends the clinical evidence base by evaluating a lower daily dose of KL-GEFE (400 mg/day) and by identifying Schirmer-measured tear secretion as the outcome with the clearest treatment-placebo difference. However, the present trial did not assess oxidative stress, inflammatory biomarkers, or lacrimal-gland mechanisms; therefore, no antioxidant or antiinflammatory mechanism can be inferred from these clinical findings.

The adapted ESQ total score showed a between-group difference favoring KL-GEFE at week 12 in both the PPS and supportive FAS analyses. Exploratory item-level findings for eye fatigue, ocular pain, irritation, and eyelid twitching followed the same direction but were not adjusted for multiplicity. Because the Korean-language 10-item ESQ was adapted from a previously reported symptom questionnaire and has not undergone formal psychometric validation in the present target population, these findings should be interpreted cautiously. Symptom instruments in dry eye disease may be influenced by baseline heterogeneity, environmental exposure, visual task burden, and the recognized discordance between symptoms and objective signs [1–3, 34]. The validated OSDI did not show a statistically significant between-group difference at week 12 (adjusted mean difference, −7.22 points; 95% CI −14.76 to 0.32; nominal P = 0.060). The magnitude of the point estimate was similar to the lower end of previously reported OSDI minimal clinically important difference ranges, which vary according to baseline severity. However, those thresholds were developed primarily to interpret within-patient changes, whereas the present estimate represents an adjusted between-group mean difference. Moreover, the confidence interval included the null value. The comparison with published MCID values should therefore be considered contextual and does not establish a clinically meaningful treatment effect [33, 42].

The findings for NIBUT and fluorescein TBUT are consistent with the possibility that KL-GEFE did not substantially modify tear-film stability during the 12-week intervention period. NIBUT and TBUT are strongly influenced by tear-film lipid-layer integrity, blink dynamics, and meibomian gland function [31, 32, 43]. Therefore, a response that is most apparent in Schirmer testing would not necessarily be accompanied by proportional changes in tear break-up metrics. Similarly, corneal and conjunctival staining scores improved within both groups but did not differ significantly between groups, a pattern that may reflect regression to the mean, study participation effects, improved ocular hygiene during follow-up, or the relatively mild-to-moderate staining burden of the enrolled population. Tear osmolarity and other integrated biomarkers were not measured, and their inclusion could have helped clarify whether the Schirmer response was accompanied by broader changes in tear-film homeostasis [44].

CFF results showed within-group improvement in the STNS direction in the KL-GEFE group, but the between-group difference at week 12 was not statistically significant. Because CFF is related to visual fatigue but is not a standard diagnostic or therapeutic endpoint for dry eye disease, the present data do not provide a sufficient basis for attributing an independent effect of KL-GEFE to neural or visual fatigue-related function. Studies specifically designed for digital eye strain or visual fatigue phenotypes may better define whether changes in CFF are clinically meaningful in this setting.

The current findings should be considered alongside available therapeutic and nutraceutical evidence for dry eye disease. Conventional treatment guidelines emphasize artificial tears, antiinflammatory therapy, secretagogues, and individualized management according to disease subtype and severity [8, 9, 11]. Pharmacologic trials have demonstrated that symptom improvement and objective sign improvement do not always occur in parallel, which is a recurring challenge in dry eye disease research [45]. Oral nutraceutical interventions, particularly omega-3 fatty acids, have shown heterogeneous effects across trials and systematic reviews [16, 18, 19]. Compared with these mixed findings, the present findings support further evaluation of grape-skin-derived anthocyanin oligomers in studies designed to clarify whether the observed Schirmer signal is reproducible and clinically meaningful. This observation is consistent with prior studies of anthocyanin preparations, while larger trials remain necessary before definitive clinical recommendations can be made [28, 29].

This study has several limitations. First, it was conducted at a single center in Korea, and the generalizability of the findings to other populations remains uncertain. Second, the study population was predominantly female, limiting sex-specific inference [1, 4, 5]. Third, the 12-week follow-up did not establish durability after discontinuation or long-term safety. Fourth, the trial was retrospectively registered after study completion, which limits independent verification of the prospectively specified outcome hierarchy and analysis framework. The pre-enrollment study protocol included the Schirmer I test and ESQ among the efficacy outcomes but did not define a formal hierarchy among them. Fifth, the planned sample size was based on assumptions informed by prior clinical studies, while the protocol did not define a formal hierarchy among efficacy outcomes or a multiplicity-control framework. Accordingly, the study was not designed for multiplicity-controlled confirmatory inference across all efficacy outcomes. Sixth, the PPS was the protocol-specified primary analysis set, with supportive FAS analyses using protocol-specified LOCF. Because LOCF assumes stability of outcomes after withdrawal, the FAS results should be interpreted with this limitation in mind. Seventh, the Korean-language ESQ was adapted from a previously reported symptom questionnaire, and further psychometric validation in the target population would be valuable. Finally, mechanistic biomarkers, tear osmolarity, inflammatory cytokines, meibography, and lacrimal gland imaging were not included, limiting mechanistic inference [31, 44].

Future research should evaluate KL-GEFE in larger, multicenter trials with longer follow-up, a prospectively defined outcome hierarchy, validated patient-reported outcome measures, and sample-size calculations appropriate to the intended confirmatory framework. Such studies should stratify or enrich participants according to dry eye subtype, baseline Schirmer values, meibomian gland status, digital device exposure, and inflammatory phenotype. Additional mechanistic assessments, including tear osmolarity, tear cytokine profiling, meibography, and lacrimal gland imaging, would help determine whether the observed effect is specific to aqueous tear secretion or is accompanied by broader improvement in ocular surface homeostasis. Further validation of the Korean-language ESQ would also strengthen interpretation of symptom outcomes.

Conclusions

In this 12-week randomized, double-blind, placebo-controlled trial, oral KL-GEFE supplementation (400 mg/day) was associated with favorable week 12 between-group differences in Schirmer I test values and adapted ESQ total score in the PPS analysis, with similar estimates in supportive FAS analyses. Other efficacy outcomes, including OSDI, tear-film stability, CFF, and ocular surface staining, did not show statistically significant week 12 between-group differences. KL-GEFE was well tolerated, with no treatment-related adverse events or ophthalmic safety concerns. These exploratory findings support further evaluation in larger, multicenter trials with a prospectively defined outcome hierarchy and sample-size and multiplicity-control procedures appropriate to the intended confirmatory framework.

Supplementary Information

Below is the link to the electronic supplementary material.

Author Contributions

Conceptualization: Kyongjin Cho; methodology: Kyongjin Cho and Hoon Kim; investigation: Yuli Park, Kyongjin Cho, and Hoon Kim; formal analysis: Hoon Kim; data curation: Hoon Kim; writing—original draft preparation: Hoon Kim; writing—review and editing: Hyun Pil Yang and Kyongjin Cho; supervision: Hyun Pil Yang and Kyongjin Cho; and project administration: Kyongjin Cho. All authors read and approved the final manuscript.

Funding

This research was funded by Kitto Life Co., Ltd. (Pyeongtaek, Republic of Korea), which also supplied the investigational product and matching placebo and provided logistical support related to product manufacturing and quality-control documentation. The sponsor did not participate in participant recruitment, clinical outcome assessment, source data collection, database handling, statistical analysis, or the decision to submit the manuscript for publication. The academic investigators retained responsibility for data interpretation and the integrity of the study findings. The Rapid Service Fee was funded by Kitto Life Co., Ltd.

Data Availability

The datasets generated and/or analyzed during the current study are not publicly available because they contain information that could compromise participant privacy but are available from the corresponding author on reasonable request, subject to institutional and ethical approval.

Declarations

Conflict of Interest

Hyun Pil Yang is an employee of Kitto Life Co., Ltd., which funded the study and supplied the investigational product and matching placebo. Hoon Kim, Yuli Park, and Kyongjin Cho declare that they have no competing interests.

Ethical Approval

This study was approved by the Institutional Review Board of Dankook University Hospital (DKUH IRB 2025-02-007) and was conducted in accordance with the Declaration of Helsinki and the Korean Good Clinical Practice Guidelines. Written informed consent was obtained from all participants before enrollment. The trial was retrospectively registered with the Clinical Research Information Service (CRIS; KCT0012279) on 14 July 2026, after participant enrollment and study completion, because the investigators had misunderstood the timing requirement for clinical trial registration and believed that registration before manuscript submission was sufficient after institutional ethics approval. To mitigate concerns arising from retrospective registration, the manuscript clearly distinguishes protocol-specified analyses from post hoc analyses, explicitly states that the study protocol did not define a formal hierarchy among efficacy outcomes or a multiplicity-control framework, and consistently interprets the efficacy findings as exploratory rather than confirmatory.