Analysis of Total Antioxidant Capacity and Ascorbic Acid Levels in the Aqueous Humor of Patients With Diabetic Retinopathy

Abstract

Purpose

The purpose of this study was to investigate aqueous humor total antioxidant capacity (TAC) and ascorbic acid (AA) levels of patients with diabetic retinopathy (DR), and their correlation with serum glycated hemoglobin (HbA1c) levels.

Methods

This case-control study enrolled 177 eyes from 177 patients that underwent cataract surgery between April 2019 and November 2022 in Chang Gung Memorial Hospital. Group 1 comprised 103 eyes from patients without diabetes; group 2 comprised 40 eyes from patients with diabetes but without DR; and group 3 comprised 34 eyes from patients with DR. Aqueous humor samples were collected and analyzed for TAC and AA levels using the ferric reducing/antioxidant AA assay.

Results

Aqueous humor TAC and AA levels were decreased in group 3 compared with group 2 and group 1 (P < 0.001 and P < 0.001). Aqueous humor TAC and AA levels were independent protective factors against DR in the diabetes population, with adjusted odds ratios of 0.109 (95% confidence interval [CI] = 0.017–0.719, P = 0.021) and 0.095 (95% CI = 0.012–0.726, P = 0.023), respectively. Aqueous humor TAC and AA concentrations were weakly negatively correlated with serum HbA1c levels, with Pearson's correlation coefficients of −0.339 (P = 0.013) and −0.341 (P = 0.012), respectively.

Conclusions

TAC and AA had protective effects against DR. Decreased aqueous humor TAC and AA levels may serve as potential biomarkers for DR among the diabetes population.

Translational Relevance

This study highlights the role of antioxidants in DR and provides foundation for future research on AA supplementation's potential for the prophylaxis and treatment of DR.

Introduction

Diabetic retinopathy (DR) is a common microvascular complication of diabetes mellitus (DM) with an estimated global prevalence of 22.27% within the DM population in 2020, affecting more than 103 million people.¹ According to the Global Burden of Disease Study in 2020, DR is the fifth leading cause of blindness, with approximately 0.86 million cases in the population aged 50 years and older worldwide.² DR is usually asymptomatic until irreversible damage to the eyes occurs and leads to visual impairment. Therefore, understanding the pathogenesis and identifying potential biomarkers are the top priorities for the early detection and prevention of DR development in patients with DM.

Multiple and complex pathogenic pathways have been implicated in the development of DR. However, the primary mechanism of DR pathophysiology remains controversial. One of the crucial metabolic pathways that take part in the pathogenesis of DR is the hyperglycemia-induced imbalance between oxidative stress and antioxidant defenses. Overproduction and impaired removal of reactive oxygen species (ROS) lead to overactivation of four alternative pathways of glucose metabolism, that is, the polyol pathway, advanced glycation end products (AGEs) formation, the protein kinase C (PKC) pathway, and the hexosamine pathway. All four pathways result in stimulation of oxidative stress, leading to vascular endothelial damage, loss of pericytes, retinal neuron cell apoptosis, and inflammatory breakdown of the blood‒retinal barrier. As retinal hypoperfusion progresses, neovascularization develops in response to vascular endothelial growth factor (VEGF) produced by ischemic retinal tissue.^3,4

Enzymatic and nonenzymatic antioxidants in the serum,^5–16 aqueous humor,^6,17–19 and vitreous body^6,19–23 can serve as biomarkers for DR. Antioxidants in the aqueous humor are more appropriate biomarkers for DR than antioxidants both in the serum and vitreous body because of the blood‒ocular barrier separating the serum from the retina and the less invasive nature of aqueous humor collection versus vitreous humor collection. In addition, there is a significant correlation between the levels of antioxidants in the aqueous humor and the vitreous body.^19,24 Due to the possible synergistic or antagonistic effects among the antioxidants in ocular tissues, analysis of total antioxidant capacity (TAC) would be more relevant than measurement of individual antioxidant biomarkers. Previous studies of TAC levels in the aqueous humor of patients with DR demonstrated varied results.^{15,17–19,25}

Levels of ascorbic acid (AA), also known as vitamin C, were proven to be the most abundant antioxidant in the aqueous humor, accounting for up to 73.2% of TAC in the aqueous humor.²⁶ In vitro studies have demonstrated that AA protects human retinal pigment epithelial cells (RPECs) and blood‒retinal barrier from high-glucose-induced ROS stress and cellular damage.^27–29 Decreased AA levels in the aqueous humor were related to insufficiency of corneal endothelial cell density,²⁶ glaucoma,³⁰ and age-related nuclear cataracts.³¹ However, whether the AA levels in the aqueous humor were decreased in the DR population compared with the DM population has not been investigated. Although serum prooxidant-antioxidant balance and serum AA levels are significantly associated with poor glycemic control,^32,33 the relationship between the antioxidant levels of ocular tissues and proper glycemic control has not yet been elucidated.

In this study, we aimed to investigate whether TAC and AA levels decrease in the aqueous humor of patients with DR. We further evaluated the correlation between antioxidant levels in the aqueous humor and serum glycated hemoglobin (HbA1c) levels.

Materials and Methods

This case-control study was approved by the Institutional Review Board (IRB) of Chang Gung Memorial Hospital, Linkou, Taiwan in 2019 (IRB number: 201900017B0). The study conformed to the provisions and tenets of the Declaration of Helsinki. All patients participating in this study provided written informed consent after receiving an explanation of the nature and possible consequences of the study.

Patient Recruitment and Grouping

Patients over 20 years of age who underwent cataract surgery between April 2019 and November 2022 in Chang Gung Memorial Hospital, Linkou, were enrolled in this study. Patients with a history of ocular infection, uveitis, prior ocular surgery, or AA supplementation were excluded.

Patients included in this study were classified into three groups according to the presence of type 2 DM and/or DR. Group 1 (the control group) comprised patients without diabetes, group 2 comprised patients with diabetes without DR, and group 3 comprised patients with DR. The diagnosis of DM was based on the World Health Organization criteria.³⁴ DR was diagnosed with fundus photography and indirect ophthalmoscopy according to the Early Treatment Diabetic Retinopathy Study criteria.³⁵

For sample size estimation, a power analysis was conducted by G*Power 3.1 (Franz Faul, University of Kiel, Kiel, Germany) based on a pilot study with 45 subjects. A total sample size of at least 66 samples was considered sufficient in this study, to reach the power of 0.8 with an α value of 0.05 by 1-way independent samples analysis of variance (ANOVA).

Sample Collection and Clinical Data Acquisition

Aqueous humor samples (50–100 µL) from each eye were aspirated using limbal paracentesis with a 27-gauge needle at the beginning of the cataract surgery. The aqueous humor samples were immediately stored at −80°C and protected from light until laboratory analysis within 2 weeks. All patients underwent thorough ophthalmologic examinations by autorefractometer (KR-7000; Topcon Corp., Tokyo, Japan), pneumatic tonometry (Topcon c60; Topcon Corp., Tokyo, Japan), slit lamp (BQ 900; Haag-Streit, Bern, Switzerland), noncontact specular microscope (CEM-530; Nidek, Gamagori, Japan), and optical biometry device (IOL Master 700; Carl Zeiss Meditec AG, Jena, Germany) before cataract surgery. Cataract type and grading were assessed using the Lens Opacities Classification System III (LOCS III) based on slit lamp examination.³⁶ Cataract severity was further quantified by measuring cumulative dissipated energy (CDE; mJ) with the Infiniti Vision System (Alcon Laboratories, Inc., Fort Worth, TX, USA) following the phacoemulsification procedure. Demographic and clinical characteristics of patients were acquired from the electronic medical records. Only serum HbA1c levels measured within 3 months before or after cataract surgery were included in this study.

Laboratory Analysis

TAC and AA levels of the aqueous humor were measured simultaneously using the colorimetric Oxiselect Ascorbic Acid Assay Kit (FRASC; Cell Biolabs, Inc., San Diego, CA, USA). The FRASC assay is based on the ferric reduction reaction and the employment of ascorbic oxidase. Aqueous humor samples were thawed at 4°C, and 35 µL of each sample was diluted 20 times in assay buffer. All standards and samples were analyzed in triplicate, with the results averaged for statistical analysis. Colorimetric results were measured at 540 to 600 nm using an absorbance microplate reader (Sunrise Tecan, Switzerland). The concentrations of TAC and AA were expressed in mmol/L (millimolar [mM]).

All serum and urinary biochemical data, including HbA1c, fasting blood glucose (FBG), uric acid, lipid profile, and renal profile, were measured by the Department of Clinical Laboratory Medicine at Chang Gung Memorial Hospital within 3 months before or after cataract surgery. The lipid profiles were determined by the enzymatic colorimetric method and accelerator selective detergent method using 7600-210 Clinical Analyzers (Hitachi, Tokyo, Japan).

Statistical Analysis

Statistical analyses of the data were performed using IBM SPSS Statistics version 25.0 for Windows (IBM, Armonk, NY, USA). All descriptive statistics were expressed as the mean ± standard deviation (SD), median with interquartile range or proportions as appropriate. For categorical variables, groups were analyzed using the chi-square test. A comparison of continuous variables among the three groups was performed using 1-way ANOVA with Bonferroni post hoc multiple comparisons. We further calculated crude and adjusted odds ratios to determine the association of aqueous humor TAC and AA levels with DR using univariate and multivariate logistic regression analyses. The correlation between TAC and AA with HbA1c was analyzed with Pearson's correlation coefficients. A P value of < 0.05 was considered statistically significant.

Results

Study Population

Overall, 177 eyes from 177 patients were included in this study. Patients were classified into three groups according to the presence of DM and/or DR. Group 1 (the control group) comprised 103 eyes from 103 patients without diabetes, group 2 comprised 40 eyes from 40 patients with diabetes without DR, and group 3 comprised 34 eyes from 34 patients with DR.

The baseline characteristics of each group of patients are summarized in Table 1. There were no significant differences in age, sex, eye site, body mass index (BMI), the proportions of smokers, the presence of dyslipidemia, hypertension, or other underlying systemic diseases among the three groups. Regarding cataract type and severity, the LOCS III grading of posterior subcapsular cataract was significantly higher in group 3 compared with group 1 and group 2 (P = 0.009). Axial length was significantly lower in group 3 compared with group 1 and group 2 (P = 0.018). There were no significant differences in the LOCS III grading of nuclear and cortical cataract, CDE during phacoemulsification, anterior chamber depth, or corneal endothelial cell density among the three groups.

Table 1.

Demographic and Clinical Characteristics of the 177 Patients/Eyes

Characteristics	Group 1 Control (n = 103)	Group 2 DM Without DR (n = 40)	Group 3 DR (n = 34)	P Value
Age, y	65.93 ± 12.07	68.30 ± 8.19	62.15 ± 11.34	0.062
Male, n (%)	47 (45.6)	25 (62.5)	20 (58.8)	0.130
OD, n (%)	43 (41.7)	21 (52.5)	18 (52.9)	0.353
BMI, kg/m2	25.04 ± 3.60	25.39 ± 3.40	25.56 ± 4.12	0.733
Underlying systemic diseases
Hypertension, n (%)	41 (39.8)	21 (52.5)	21 (61.8)	0.061
Dyslipidemia, n (%)	30 (29.1)	13 (32.5)	15 (50.0)	0.081
Smoking, n (%)	9 (8.7)	7 (17.5)	2 (5.9)	0.195
Other systemic diseases, n (%)a	39 (37.9)	15 (37.5)	14 (41.2)	0.934
Cataract type and grading
LOCS III (Nuclear)	2.54 ± 0.91	2.45 ± 0.76	2.14 ± 0.96	0.116
LOCS III (Cortical)	2.38 ± 1.06	2.34 ± 0.94	1.80 ± 1.25	0.070
LOCS III (Posterior subcapsular)	0.22 ± 0.75b	0.08 ± 0.36b	0.63 ± 1.01c	0.009
CDE during phacoemulsification, mj	26.84 ± 19.67	38.43 ± 31.05	25.93 ± 19.92	0.062
Biometry
Axial length, mm	24.86 ± 2.29b	24.41 ± 1.79b	23.55 ± 1.22c	0.018
Anterior chamber depth, mm	3.14 ± 0.45	2.98 ± 0.40	3.07 ± 0.52	0.194
Corneal endothelial cell density, cells/mm2	2582.94 ± 330.25	2662.04 ± 313.19	2738.47 ± 301.76	0.142

BMI, body max index; CDE, cumulative dissipated energy; DM, diabetes mellitus; DR, diabetic retinopathy; LOCS III, Lens Opacities Classification System III; OD, oculus dexter.

^aOther systemic diseases include chronic heart diseases, lung diseases, liver diseases, kidney diseases, cerebrovascular diseases, immunocompromised status, and autoimmune diseases.

^b,cIntergroup comparison using Bonferroni post hoc test, identical symbols indicate no significant difference among groups, whereas different letters indicate a significant difference among groups.

The P values in bold represent statistical significance.

TAC Levels and AA Levels in the Aqueous Humor

There was a highly positive correlation between TAC and AA concentrations in the aqueous humor, with a Pearson's correlation coefficient of 0.943 (P < 0.001; Fig. 1). The laboratory findings of the patients are shown in Table 2. TAC levels in the aqueous humor were decreased in the DR group (1.08 ± 0.53 mM) compared with the patients with DM without DR group (1.55 ± 0.58 mM, P < 0.001) and the control group (1.59 ± 0.42 mM, P < 0.001) by 1-way ANOVA with Bonferroni post hoc test. The aqueous humor TAC levels were similar between the patients with DM without DR group and the control group (P = 1.000; Fig. 2A). Similarly, AA levels in the aqueous humor were the lowest in the DR group (0.79 ± 0.47 mM) compared with the patients with DM without DR group (1.22 ± 0.52 mM, P < 0.001) and the control group (1.19 ± 0.36 mM, P < 0.001). The aqueous humor AA levels were not significantly different between the patients with DM without DR group and the control group (P = 1.000; Fig. 2B). AA/TAC ratios in the aqueous humor were lower in the DR group (68.43 ± 17.37%) compared with the patients with DM without DR group (76.65 ± 16.20%, P = 0.023). Aqueous humor AA/TAC ratios were similar in the control group (73.79 ± 9.70%) compared to the patients with DM without DR group (P = 0.728) and the DR group (P = 0.120; Fig. 2C).

Figure 1.

Scatter plot illustrating a positive correlation between total antioxidant capacity (TAC) and ascorbic acid (AA) concentrations in the aqueous humor with Pearson's correlation coefficient.

Table 2.

Comparison of Aqueous Humor Antioxidant Levels, Serum Glycated Hemoglobin (HbA1c) Levels, Lipid Profiles, and Renal Profiles Among the Three Groups

Characteristics	Group 1 Control (n = 103)	Group 2 DM Without DR (n = 40)	Group 3 DR (n = 34)	P Value
Aqueous humor
TAC, mM	1.59 ± 0.42a	1.55 ± 0.58a	1.08 ± 0.53b	<0.001
AA, mM	1.19 ± 0.36a	1.22 ± 0.52a	0.79 ± 0.47b	<0.001
AA/TAC ratio (%)	73.79 ± 9.70a,b	76.65 ± 15.99a	68.43 ± 17.37b	0.026
Serum
HbA1c (%)	5.78 ± 0.33a	7.24 ± 1.32b	8.09 ± 1.49b	<0.001
FBG, mg/dL	95.00 ± 12.05a	124.93 ± 29.75b	140.54 ± 44.50b	0.001
Uric acid, mg/dL	6.13 ± 1.83	5.91 ± 1.55	6.91 ± 1.10	0.421
Lipid profile
TG, mg/dL	141.29 ± 83.10	112.80 ± 50.27	156.00 ± 111.32	0.376
Chol, mg/dL	173.05 ± 36.40	154.19 ± 26.76	169.47 ± 48.97	0.323
LDL, mg/dL	103.63 ± 39.14	83.94 ± 25.14	90.67 ± 32.14	0.223
HDL, mg/dL	45.64 ± 14.07	50.77 ± 8.45	46.60 ± 11.53	0.487
Renal profile
BUN, mg/dL	24.03 ± 20.86	23.70 ± 8.45	28.31 ± 22.60	0.857
Creatinine, mg/dL	1.44 ± 2.19	1.29 ± 1.80	2.36 ± 3.24	0.326
eGFR, mL/min/1.73m2	80.77 ± 33.80	72.39 ± 25.59	76.00 ± 45.71	0.782
ACR, mg/g	15.44 ± 9.59	32.40 ± 46.69	1221.37 ± 1772.48	0.181

AA, ascorbic acid; ACR, urinary albumin to creatinine ratio; BUN, blood urea nitrogen; Chol, total cholesterol; eGFR, estimated glomerular filtration rate; FBG, fasting blood glucose; HbA1c, glycated hemoglobin; HDL, high-density lipoprotein cholesterol; LDL, low-density lipoprotein cholesterol; TAC, total antioxidant capacity; TG, triglycerides.

^a,bIntergroup comparison using Bonferroni post hoc test, identical symbols indicate no significant difference among groups, while different letters indicate a significant difference among groups.

The P values in bold represent statistical significance.

Figure 2.

Comparison of total antioxidant capacity (TAC) levels (A), ascorbic acid (AA) levels (B), and AA/TAC ratios (C) in the aqueous humor between eyes from patients with diabetic retinopathy (DR), patients with diabetes without DR, and nondiabetic controls by 1-way ANOVA with Bonferroni post hoc test. The data for each group are presented as individual data points with the means and standard deviations depicted by error bars. *P < 0.05; **P < 0.01; ***P < 0.001.

For serum and urinary biochemical measurements of the patients, HbA1c (P < 0.001), and FBG (P = 0.001) were significantly higher in group 2 and group 3 compared with group 1. There were no significant differences regarding uric acid levels, lipid profiles, blood urea nitrogen (BUN) levels, creatinine levels, estimated glomerular filtration rate (eGFR), or urinary albumin to creatinine ratio (ACR) among the three groups (see Table 2).

The Association of Aqueous Humor TAC and AA Levels With DR

We then evaluated the association of aqueous humor TAC and AA levels with DR using univariate and multivariate logistic regression analyses. In univariate logistic regression models, the crude odds ratios for the association of aqueous humor TAC and AA levels with DR in the DM population were 0.223 (95% CI = 0.089–0.558, P = 0.001) and 0.183 (95% CI = 0.066–0.504, P = 0.001), respectively. In multivariate logistic regression models, the adjusted odds ratios were 0.109 (95% CI = 0.017–0.719, P = 0.021) and 0.095 (95% CI = 0.012–0.726, P = 0.023) after adjusting for age, sex, eye site, BMI, hypertension, dyslipidemia, smokers, underlying systemic diseases, cataract type with grading based on the LOCS III, and axial length as potential confounding factors. The results indicated that TAC and AA levels in the aqueous humor were independent protective factors against DR in the DM population (Table 3).

Table 3.

Univariate and Multivariate Logistic Regression Analyses of the Association of Aqueous Humor Total Antioxidant Capacity (TAC) and Ascorbic Acid (AA) Levels With Diabetic Retinopathy (DR)

	Univariate Analysis		Multivariate Analysis
Factors	Crude OR (95% CI)	P Value	Adjusted OR (95% CI)a	P Value
TAC
DM without DR vs. control	0.831 (0.382–1.807)	0.640	0.799 (0.277–2.186)	0.854
DR vs. control	0.136 (0.058–0.318)	<0.001	0.193 (0.053–0.705)	0.013
DR vs. DM without DR	0.223 (0.089–0.558)	0.001	0.109 (0.017–0.719)	0.021
Ascorbic acid
DM without DR vs. control	1.234 (0.505–3.012)	0.645	1.443 (0.448–4.645)	0.538
DR vs. control	0.128 (0.049–0.335)	<0.001	0.170 (0.039–0.748)	0.019
DR vs. DM without DR	0.183 (0.066–0.504)	0.001	0.095 (0.012–0.726)	0.023

CI, confidence interval; OR, odds ratio.

^aAdjusted for age, sex, eye site, BMI, hypertension, dyslipidemia, smokers, underlying systemic diseases, cataract type with grading based on the Lens Opacities Classification System III, and axial length as confounding factors.

The P values in bold represent statistical significance.

The Correlation Between Antioxidant Levels in the Aqueous Humor and HbA1c Levels

We further investigated the relationship between the antioxidant levels in the aqueous humor and long-term glycemic control. There was a weak negative correlation between aqueous humor TAC concentrations and serum HbA1c levels, with a Pearson's correlation coefficient of −0.339 (P = 0.013; Fig. 3A). Similarly, aqueous humor AA concentrations were weakly negatively associated with serum HbA1c levels, with a Pearson's correlation coefficient of −0.341 (P = 0.012; Fig. 3B). There was no significant correlation between aqueous humor AA/TAC ratios and serum HbA1c levels, with a Pearson's correlation coefficient of −0.239 (P = 0.085; Fig. 3C).

Figure 3.

Scatter plots illustrating the correlation between serum glycated hemoglobin (HbA1c) levels and aqueous humor total antioxidant capacity (TAC) concentrations (A), serum HbA1c levels and aqueous humor ascorbic acid (AA) concentrations (B), and serum HbA1c levels and aqueous humor AA/TAC ratios (C) with Pearson's correlation coefficients.

Discussion

The prevalence of DR is increasing as the global epidemic of DM grows. This leads to irreversible vision loss, particularly in working-age adults. Currently, there are no widely adopted effective biomarkers to identify those who are at high risk of developing DR in the DM population, nor is there prophylactic treatment to prevent DR progression before irreversible damage to the eyes occurs, except for glycemic control. In this study, we demonstrated that TAC and AA levels decreased in the aqueous humor of patients with DR compared to those in patients with DM without DR and to those in patients without diabetes (see Table 2). We further verified that TAC and AA levels in the aqueous humor were independent protective factors against the development of DR. This is the first study to indicate that decreased AA levels in the aqueous humor may serve as a potential biomarker for DR development among the diabetes population and to suggest the potential of AA supplementation for prophylaxis of DR.

Depletion of enzymatic and nonenzymatic antioxidants in the serum, aqueous humor, and vitreous body can serve as biomarkers for DR (Table 4). Although DM is a systemic disease, ocular factors are more relevant in the pathogenesis of DR than systemic factors because of the blood‒ocular barrier. AA, one of the most critical elements of the antioxidant defense system in ocular tissues, is actively transported from serum to aqueous humor and vitreous body by a sodium-dependent vitamin C transporter 2 (SVCT2) located in the pigmented ciliary epithelium, resulting in approximately a 20-fold increase in AA levels both in the aqueous humor (1.30 ± 0.62 mM) and the vitreous body (1.28 ± 0.37 mM) in comparison with those in the serum (0.06 ± 0.03 mM).^37–40 Existing literature indicates that AA is also present in RPECs, retinal endothelial cells, and retinal capillary pericytes, supported by the presence of SVCT2-dependent and glucose transporter 1 (GLUT 1)-dependent AA transport mechanisms in these cells.^41–43 AA is one of the major contributors to TAC both in the aqueous humor²⁶ and the vitreous body.⁴⁴ Moreover, there is a significantly positive correlation between the levels of AA in the aqueous humor and the vitreous body.⁶ Because aqueous humor sampling is less invasive than vitreous body collection, analysis of antioxidants in the aqueous humor would be more practical for predicting DR and guiding further antioxidant-based therapy.

Table 4.

Human Endogenous Antioxidant Biomarkers in Diabetic Retinopathy

		Antioxidant Biomarkersa
Sample Source	TAC Decrease (↓) No Significant Difference (–)	Enzymatic	Nonenzymatic	References
Serum	(↓) 10, 15, 16, 54	Superoxide	AA	5–14, 16, 55
	(–) 19, 25	dismutase	Tocopherol
		Catalase	Thiol
		Glutathione	Glutathione
		peroxidase	α-carotene
		Paraoxonase 1	Copper
			Zinc
Aqueous humor	(↓) 17–19	–	AA	6
	(–) 15, 24
Vitreous body	(↓) 19, 22, 56	Superoxide dismutase	AA	6, 20, 21, 23, 57, 58
		Catalase	Glutathione

^aThis table only includes antioxidants that are proven to be significantly different in patients with diabetic retinopathy and healthy subjects.

Previous studies of TAC levels in the aqueous humor of patients with DR have produced inconsistent results. Beyazyıldız et al., Bozkurt et al., and Mancino et al. reported that TAC levels in the aqueous humor of patients with DR were significantly impaired, which is consistent with the results from our study.^17–19 However, Kulaksızoglu et al. and Kirboga et al. revealed no significant difference in the aqueous humor TAC levels between patients with DR and the control groups without diabetes.^15,25 The discrepancy in the results of Kulaksızoglu et al. and Kirboga et al. might be due to relatively small sample sizes and relatively inappropriate measuring methods for TAC. Due to the complexity of enzymatic and nonenzymatic antioxidants contributing to TAC, none of the existing TAC assays can thoroughly detect the antioxidant capacity of all antioxidants in the aqueous humor. TAC assays using colorimetric changes based on the redox reaction of metal ions are a relatively appropriate method to evaluate aqueous humor TAC levels because small molecular antioxidants are the major components of antioxidant defensive systems in the aqueous humor.⁴⁵ Our study provided rather convincing evidence that TAC decreased in the aqueous humor of DR with the largest sample size to date using the FRASC assay based on ferric reduction reaction analysis.

AA is the most noteworthy antioxidant among the enzymatic and nonenzymatic antioxidants in the aqueous humor. One reason is that AA contributes up to 73.2% of TAC in the aqueous humor.²⁶ Another reason is that AA supplements are more accessible for further therapeutic and prophylactical applications than other enzymatic and nonenzymatic antioxidants with minimal side effects. Park et al. reported that aqueous humor AA levels were lower in patients with proliferative DR than in healthy controls without diabetes.⁶ Our study extended previous research and found that aqueous humor AA levels were lower in patients with DR than in patients with DM without DR and nondiabetic healthy controls. In addition, the aqueous humor AA levels were similar between the patients with DM without DR and the nondiabetic healthy controls. The results of our study implied that depletion of AA may play an important role in the development of DR in the DM populations aside from long-term poor glycemic control. Moreover, we found that AA was a protective factor against DR in the DM population with an adjusted odds ratio of 0.095 (95% CI = 0.012–0.726, P = 0.023). The ratio of AA to TAC in the aqueous humor was also reduced in the DR group compared with the patients with DM without DR group. The potential pathogenic mechanisms underlying ocular depletion of AA in patients with DR may include excessive consumption by ROS and impaired secretion due to dysfunction of the AA transporters.^29,43,46 Further investigation is merited to identify which antioxidants compensate for the depletion of AA.

In vivo studies have revealed that oxidative stress contributes not only to the onset of DR but also to the persistence of DR even after reinstitution of glycemic control. This metabolic memory is attributed to the accumulation of ROS, which are not easily eliminated after the restoration of normoglycemia.⁴⁷ Consequently, adequate levels of antioxidants in ocular tissues are essential for scavenging ROS and preventing DR progression. Our study is the first to investigate the relationship between the antioxidant levels of ocular tissues and proper glycemic control. We have demonstrated that the aqueous humor TAC and AA concentrations were weakly negatively correlated with serum HbA1c levels. Together with the studies mentioned above, the combination of AA supplementation and glycemic control might be more efficacious than glycemic control alone for prophylaxis and treatment of DR. Previously, dietary intake of AA has been studied for the treatment of DR. Tanaka et al. reported a protective effect of dietary AA intake against the incidence of DR with a hazard ratio of 0.66 (95% CI = 0.39–0.96, P = 0.03) in a cohort study.⁴⁸ Milan et al. observed that there was no significant overall association between the risk of retinopathy and the intake of dietary AA in a cohort study.⁴⁹ In contrast, Mayer-Davis et al. reported that increased intake of AA was a risk factor for increased severity of DR, with an odds ratio of 2.21 (P = 0.01) in a cross‒sectional study.⁵⁰ Experimentally, AA was found to suppress VEGF expression in RPECs⁵¹ and diminish the oxidative damage within human RPECs in oxidative stress models.^27,28,52 AA together with valproic acid may alleviate ROS-induced damage to the retina by promoting human fetal RPECs proliferation and regeneration.⁵³ In addition, AA may maintain the integrity of the blood‒retinal barrier by mitigating the loss and dysfunction of retinal endothelial cells and capillary pericytes.²⁹ (Walker K, et al. IOVS 2023;64:ARVO E-Abstract 995) The potential role of AA for prophylaxis and treatment of DR remains open to debate and is worthy of further research.

There are several limitations to our study. First, we did not classify patients with DR into subgroups according to the severity of DR because of the relatively small number of patients with DR enrolled. Second, we did not assess the duration of DM reported by the patients due to potential recall bias. Third, examining aqueous humor TAC and AA levels repeatedly in a longitudinal study is not very appropriate, as obtaining aqueous humor samples is an invasive procedure. Thus, we could only measure TAC and AA levels once for each subject in a retrospective study design. This restricted our capacity to observe the fluctuation in TAC and AA levels as DR status progressed over time. Further large-scale clinical trials are warranted to verify the prophylactic and therapeutic applicability of AA supplements through systemic, topical, intravitreal, and periocular administration routes for preventing the development of DR in the diabetes population and decelerating the deterioration of existing DR.

In conclusion, our study demonstrated that TAC and AA levels decreased in the aqueous humor of patients with DR compared with those of patients with DM without DR and to those of patients without diabetes. We further verified the protective effect of TAC and AA against DR. In addition, aqueous humor TAC and AA concentrations were weakly negatively correlated with serum HbA1c levels. Decreased TAC and AA levels in the aqueous humor may serve as potential biomarkers for DR among the diabetes population. These results also suggest the need for further studies to evaluate the efficacy of AA supplementation for prophylaxis and treatment of DR.