Abstract
Background
Previous studies have suggested oxidative stress may play a key role in the pathogenesis of retinopathy, while evidence from observational studies directly linking oxidative biomarkers to clinically relevant outcomes has been limited. This study aims to investigate the association between oxidative balance score (OBS) and prevalence of retinopathy in a nationally representative sample of U.S. adults, including both those with and without diabetes.
Methods
Data were obtained from the NHANES 2005–2008, including 3,287 participants. OBS was calculated from 16 dietary and 4 lifestyle components and categorized into tertiles. Weighted logistic regression was used to assess the association between OBS and retinopathy in the overall, diabetic, and non-diabetic populations, and Cox proportional hazards models evaluated the link between OBS and all-cause mortality in those with retinopathy. Restricted cubic spline (RCS) analysis explored the dose-response relationship between OBS and retinopathy. Subgroup analyses and interaction tests were conducted based on age, sex, race, BMI, education level, and diabetes status.
Results
Participants in the highest OBS tertile had a 28% lower risk of retinopathy compared to those in the lowest tertile (OR 0.72, 95% CI 0.54–0.95, P = 0.023). RCS analysis showed a significant overall association between higher OBS and reduced retinopathy risk, without a nonlinear pattern. In participants with retinopathy, higher OBS was linked to a 60% reduction in all-cause mortality (HR 0.40, 95% CI 0.24–0.66, P < 0.001). Subgroup analysis revealed stronger inverse associations between OBS and retinopathy in younger individuals and those with higher education, with a significant interaction between OBS and age (P for interaction < 0.05).
Conclusion
Our study provides evidence that higher cumulative antioxidant exposure assessed by OBS is associated with a reduced risk and severity of retinopathy and lower all-cause mortality in U.S. adults with retinopathy, highlighting the importance of maintaining a favorable oxidative balance in retinal health.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12967-025-06110-4.
Kewords: Retinopathy, Oxidative balance score, Diabetes, Mortality, NHANES
Introduction
Retinopathy, a broad term referring to damage to the retina’s blood vessels, is a significant cause of vision loss and blindness worldwide [1], affecting various populations including but not limited to those with diabetes. While diabetic retinopathy is common, affecting approximately 35% of people with diabetes [2], retinopathy can also result from other conditions such as hypertension, autoimmune diseases, and age-related macular degeneration. The condition develops when small blood vessels in the retina are damaged due to factors like chronic hyperglycemia, sustained high blood pressure, or other systemic issues. It can progress from mild non-proliferative abnormalities to more severe forms, such as proliferative retinopathy and macular edema, leading to visual impairment. The progression in diabetic individuals is often termed diabetic retinopathy, while in non-diabetic individuals, it may be called hypertensive retinopathy, central serous retinopathy, or other specific types depending on the underlying cause. With the increasing prevalence of diabetes, hypertension, and an aging population, the overall burden of retinopathy is expected to rise substantially [3], highlighting the importance of regular eye examinations and management of underlying health conditions for all individuals, regardless of diabetic status.
Oxidative stress resulting from overproduction of reactive oxygen species and diminished antioxidant defenses is implicated in the development of diabetic retinopathy [4]. Oxidative stress can induce endothelial dysfunction, inflammation, and neuronal and microvascular damage in the retina [5]. Antioxidant enzymes and nutrients can counterbalance oxidative stress and may play a protective role against diabetic retinopathy. Higher intake of antioxidants such as vitamin C, vitamin E, and carotenoids has been associated with reduced risk of diabetic retinopathy [6–8]. The oxidative balance score (OBS) provides an integrated measure of the balance between exposure to antioxidants and pro-oxidants from both dietary and lifestyle factors [9]. Studies have shown that higher OBS is inversely associated with risk of type 2 diabetes, cardiovascular disease, and all-cause mortality [10–12]. However, few studies have specifically examined the relationship between OBS and retinopathy in both diabetic and non-diabetic populations.
In this study, we aim to investigate the association between OBS and prevalence of retinopathy in a nationally representative sample of U.S. adults, including both those with and without diabetes. We will also assess the relationship between OBS and all-cause mortality among those with retinopathy. Our findings will provide novel evidence on the role of oxidative balance in retinopathy and its prognostic value in people with retinopathy.
Methods
Study population
The data for this study were obtained from the National Health and Nutrition Examination Survey (NHANES), a large cross-sectional study conducted by the National Institutes of Health (NIH) and the Centers for Disease Control and Prevention (CDC) to assess the health and nutritional status of the U.S. population, including both adults and children. For this particular study, a total of 20,497 participants were recruited during two 2-year survey cycles in NHANES (2005–2008). All participants involved in this survey provided written informed consent prior to their participation. Among them, 14,793 participants were excluded due to either the absence of diabetic retinopathy grading or missing fundus imaging data. 1,967 participants were excluded because the OBS could not be calculated. Additionally, 450 participants were excluded due to missing data on other covariates. Ultimately, a total of 3,287 participants were included in this study, including 1,728 males and 1,559 females (Fig. 1).
Fig. 1.
The flow diagram of the study participants, from NHANE 2005–2008
Data collection
Assessment of OBS
The OBS was calculated as the sum of the scores for 20 individual components, including 16 dietary and 4 lifestyle factors. The dietary OBS components comprised dietary fiber, vitamin C, vitamin B6, vitamin B12, total folate, carotenoids, riboflavin, nicotine, vitamin E, total fat, and iron, calcium, magnesium, zinc, copper, and selenium. The lifestyle OBS components included smoking, alcohol consumption, body mass index (BMI), and physical activity. These components were further categorized into antioxidants (e.g., vitamins and minerals) and prooxidants (e.g., total fat, iron, BMI, smoking, and alcohol consumption). We selected these substances as antioxidants based on their documented association with reducing oxidative stress and related ocular diseases in previous studies [13]. Each component was scored based on its oxidative or antioxidative effect. Antioxidant components were scored from 0 to 2, with higher scores indicating greater antioxidant potential. Prooxidant components were scored inversely, from 2 to 0, with higher scores indicating lower prooxidant exposure. The scores for each component were summed to generate the total OBS, with higher OBS values reflecting a more favorable oxidative balance [14].
For dietary components, OBS scores were derived using the average of two 24-hour dietary recalls, with data collected both in-person and via telephone interviews. The Food Surveys Research Group (FSRG) of the United States Department of Agriculture compiled and reviewed these data. For lifestyle components, smoking status was assessed using plasma cotinine concentration, alcohol consumption was categorized by intake levels (abstainers, non-heavy drinkers, heavy drinkers), BMI was calculated as weight divided by height squared (kg/m^2), and physical activity was based on questionnaire responses regarding exercise frequency and duration [13, 15]. The final OBS was calculated by summing the individual scores for all 20 components, with scores ranging from 0 to 40. Continuous OBS scores were categorized into tertiles: T1 (< 18), T2 (18–24), and T3 (> 24). For detailed information, please refer to eTable 1.
Assessment of retinopathy
Digital imaging systems using a retinal camera (Canon CR4-45NM) were employed by trained examiners to acquire non-stereoscopic retinal color photographs of individuals aged 40 and above. These photographs were sent to the ophthalmic research center laboratory at the University of Wisconsin for further processing and quality assessment. The classification and evaluation of retinal status were performed using the Early Treatment Diabetic Retinopathy Study grading system. The presence of any of the following conditions was defined as retinopathy: (i) hard exudates/soft exudates; (ii) intraretinal microvascular abnormalities; (iii) microaneurysms; (iv) hemorrhages; (v) non-proliferative or proliferative diabetic retinopathy, fibrous proliferation; or any evidence of diabetic retinopathy [16]. Based on these criteria, participants were categorized into either the retinopathy group or the non-retinopathy group.
Covariates
The covariates included in the analysis were age, sex, race, education level, smoking status, drinking status, hypertension, diabetes and BMI. Age was categorized into two groups: 18–65 years and ≥ 65 years. Race was divided into five groups: Mexican American, Non-Hispanic Black, Non-Hispanic White, Other Hispanic, and Other Race - Including MultiRacial. Education level was classified as follows: less than high school (including 9th grade and below, 9th-11th grade, and no diploma), high school (high school diploma or equivalent), and higher than high school (college graduate or above). Smoking status was classified as follows: former smoker (having smoked more than 100 cigarettes in lifetime but currently not smoking), never smoker (having smoked fewer than 100 cigarettes in lifetime and not currently smoking), and current smoker (having smoked more than 100 cigarettes and currently smoking either every day or some days)). Drinking status is categorized into two groups: drinkers and non-drinkers. Drinking is defined as having consumed alcohol at the time of the survey or having consumed more than 12 alcoholic beverages in their lifetime. Hypertension was characterized by a self-reported diagnosis, an elevated average blood pressure (systolic ≥ 130 mm Hg and/or diastolic ≥ 85 mm Hg), or the current use of antihypertensive medications. Diabetes classification followed the World Health Organization (WHO) criteria and was divided into four groups: (i) Diabetes Mellitus (DM) group: fasting plasma glucose (FPG) ≥ 7.0 mmol/l or (≥ 126 mg/dl), oral glucose tolerance test (OGTT) ≥ 11.1 mmol/l (200 mg/dl), glycosylated hemoglobin (HbA1c) ≥ 6.5%, or current use of oral medication or insulin treatment; (ii) Impaired Fasting Glucose (IFG): FPG between 6.1 and 6.9 mmol/l; (iii) Impaired Glucose Tolerance (IGT): OGTT between 7.8 and 11.0 mmol/L; (iv) No diabetes group. BMI was categorized into three groups: <25, [25, 30), and ≥ 30.
Mortality
To further explore the impact of OBS on the prognosis of patients with retinal lesions, we linked the National Death Index database (NDI) with the Nhanes database to obtain survival follow-up data for participants surveyed in Nhanes, including survival status, survival time, and causes of death. The follow-up data was collected until December 31, 2019. Causes of death were coded according to the International Classification of Diseases, 10th Revision (ICD-10).
Statistical analysis
Individual sample weights were calculated based on the recommended sample weights (wtint2 year) from the Nhanes database. In the statistical analysis, continuous variables were described using weighted means (standard deviations), and differences in descriptive statistics were assessed using t-tests and rank-sum tests. Categorical variables were described using sample counts (weighted percentages), and differences in descriptive statistics were assessed using Rao-Scott χ2 tests. To better evaluate the relationship between OBS and retinopathy in the general population, as well as in diabetic and non-diabetic subgroups, we transformed the continuous variable OBS into categorical variables based on tertiles and employed two models: a crude model without adjustment for any confounding factors, and the multivariable model, which adjusted for confounding factors such as age, sex, race/ethnicity, education level, smoking status, drinking status, hypertension, diabetes, and BMI. Univariate and multivariate Cox proportional hazards models were constructed to further evaluate the association between OBS and all-cause mortality in patients with retinopathy. The univariate model did not adjust for any confounding factors, while the multivariate model adjusted for the same variables as in the aforementioned multivariate logistic regression model. Restricted cubic spline (RCS) curves were developed to assess the dose-response relationship between OBS and retinopathy in the overall population, as well as in diabetic and non-diabetic subgroups. The RCS curves were also used to evaluate the association between OBS and all-cause mortality in participants with retinopathy. Kaplan-Meier survival curves were generated to study the relationship between OBS in retinal lesion patients and all-cause mortality. In the sensitivity analysis, to ensure the robustness of the findings, we excluded retinopathy patients with a follow-up death time of less than 90 days. The same Cox proportional hazards model was then applied to reanalyze the data in both the crude and adjusted models. Kaplan-Meier survival curves were also plotted to compare the survival outcomes among retinopathy patients across different OBS tertiles.
All statistical analyses and plotting were performed using R software version 4.2.2, and a two-sided P-value < 0.05 was considered statistically significant.
Results
Baseline characteristics
We categorized the participants based on the presence or absence of retinal lesions, with 2,916 individuals without retinal lesions and 371 individuals with retinal lesions. Among them, there was a higher proportion of female participants in the group without retinal lesions, accounting for approximately 48.77%, while the group with retinal lesions had a higher proportion of male participants, accounting for approximately 63.07%. In both groups, there were more participants under the age of 65, with 70.88% in the group without retinal lesions and 67.12% in the group with retinal lesions. Participants with retinal lesions were more likely to be male, non-Hispanic black, with an educational level of high school or below, have hypertension, be non-drinkers, be diabetic, with the BMI of 25 or above, and have lower OBS percentile levels. Significant differences were observed in these baseline characteristics. More specific details can be seen in Table 1.
Table 1.
Weighted characteristics of the study population by retinopathy
| Level | No Retinopathy | Retinopathy | P-value | |
|---|---|---|---|---|
| N | 2916 | 371 | ||
| OBS (%) | 0.0001 | |||
| T1 | 1066 (36.56) | 173 (46.63) | ||
| T2 | 848 (29.08) | 107 (28.84) | ||
| T3 | 1002 (34.36) | 91 (24.53) | ||
|
Race/ethnicity (%) |
61 (16.44) | < 0.0001 | ||
| Mexican American | 400 (13.72) | 61 (16.44) | ||
| Non-Hispanic Black | 468 (16.05) | 102 (27.49) | ||
| Non-Hispanic White | 1791 (61.42) | 172 (46.36) | ||
| Other Hispanic | 168 (5.76) | 28 (7.55) | ||
| Other Race - Including Multi-Racial | 89 (3.05) | 8 (2.16) | ||
| Age (%) | 0.1504 | |||
| 18–65 | 2067 (70.88) | 249 (67.12) | ||
| ≥ 65 | 849 (29.12) | 122 (32.88) | ||
| Sex (%) | < 0.0001 | |||
| Female | 1422 (48.77) | 137 (36.93) | ||
| Male | 1494 (51.23) | 234 (63.07) | ||
| Diabetes (%) | < 0.0001 | |||
| DM | 464 (15.91) | 175 (47.17) | ||
| IFG | 171 (5.86) | 19 (5.12) | ||
| IGT | 191 (6.55) | 15 (4.04) | ||
| No | 2090 (71.67) | 162 (43.67) | ||
| BMI (%) | 0.0023 | |||
| < 25 | 791 (27.13) | 70 (18.87) | ||
| 25–30 | 1068 (36.63) | 158 (42.59) | ||
| ≥ 30 | 1057 (36.25) | 143 (38.54) | ||
| Education Level (%) | 0.0002 | |||
| Low high school | 653 (22.39) | 111 (29.92) | ||
| High school | 698 (23.94) | 101 (27.22) | ||
| College or above | 1565 (53.67) | 159 (42.86) | ||
| Hypertension (%) | No | 1460 (50.07) | 127 (34.23) | < 0.0001 |
| Yes | 1456 (49.93) | 244 (65.77) | ||
| Smoking status (%) | Former | 988 (33.88) | 115 (31.00) | 0.5277 |
| Never | 1373 (47.09) | 184 (49.60) | ||
| Now | 555 (19.03) | 72 (19.41) | ||
| Drinking status (%) | No | 1018 (34.91) | 180 (48.52) | < 0.0001 |
| Yes | 1898 (65.09) | 191 (51.48) |
Notes: OBS, oxidative balance scores; IFG, impaired fasting glucose; IGT, impaired glucose tolerance; DM, diabetes mellitus; No, normal
Table 2 presents the weighted characteristics of the study population categorized by mortality status. The median follow-up time for the entire cohort was 145 months. Individuals in the deceased group (N = 119) were more likely to be non-Hispanic White, older, diabetic, former smokers, and drinkers compared to those in the surviving group (N = 252).
Table 2.
Weighted characteristics of the study population by mortality status
| Level | Alive | Deceased | P-value | |
|---|---|---|---|---|
| N | 252 | 119 | ||
| Race/ethnicity (%) | 0.0127 | |||
| Mexican American | 48 (19.05) | 13 (10.92) | ||
| Non-Hispanic Black | 71 (28.17) | 31 (26.05) | ||
| Non-Hispanic White | 103 (40.87) | 69 (57.98) | ||
| Other Hispanic | 24 (9.52) | 4 (3.36) | ||
| Other Race - Including Multi-Racial | 6 (2.38) | 2 (1.68) | ||
| Age (%) | ||||
| 18–65 | 201 (79.76) | 48 (40.34) | < 0.0001 | |
| ≥ 65 | 51 (20.24) | 71 (59.66) | ||
| Sex (%) | 0.3058 | |||
| Female | 98 (38.89) | 39 (32.77) | ||
| Male | 154 (61.11) | 80 (67.23) | ||
| Diabetes (%) | 0.0002 | |||
| DM | 104 (41.27) | 71 (59.66) | ||
| IFG | 12 (4.76) | 7 (5.88) | ||
| IGT | 7 (2.78) | 8 (6.72) | ||
| No | 129 (51.19) | 33 (27.73) | ||
| BMI (%) | ||||
| < 25 | 47 (18.65) | 23 (19.33) | 0.6636 | |
| 25–30 | 104 (41.27) | 54 (45.38) | ||
| ≥ 30 | 101 (40.08) | 42 (35.29) | ||
| Education Level (%) | 0.1024 | |||
| Low high school | 73 (28.97) | 38 (31.93) | ||
| High school | 62 (24.60) | 39 (32.77) | ||
| College or above | 117 (46.43) | 42 (35.29) | ||
| Hypertension (%) | No | 106 (36.55) | 21 (25.93) | 0.099 |
| Yes | 184 (63.45) | 60 (74.07) | ||
| Smoking status (%) | Former | 75 (25.86) | 40 (49.38) | 0.0002 |
| Never | 153 (52.76) | 31 (38.27) | ||
| Now | 62 (21.38) | 10 (12.35) | ||
| Drinking status (%) | No | 137 (47.24) | 37 (45.68) | 0.0424 |
| Yes | 153 (52.76) | 44 (54.32) |
Notes: IFG, impaired fasting glucose; IGT, impaired glucose tolerance; DM, diabetes mellitus; No, normal
Association between OBS and retinopathy
The weighted logistic regression analysis revealed a significant correlation between OBS and retinal disease occurrence. In the crude model, compared to the lowest tertile (T1 group) of OBS and the third tertile (T3 group) showed a significantly negative association with the risk of retinal disease. The risk of retinal disease in the T3 group decreased by 44% (OR 0.56, 95%CI = 0.43–0.73, P < 0.001). In the multivariable model, the risk of retinal disease in the T3 group decreased by 28% (OR 0.72, 95%CI = 0.54–0.95, P = 0.023). Please refer to Fig. 2 for detailed information. In the RCS analysis, without adjustment for variables, a significant association was observed between OBS and retinal disease, but the nonlinear relationship was not significant (Overall P < 0.001, Non-linear P = 0.040) (Fig. 3A). In multivariable model, there is also a significant association was observed between OBS and retinal disease (Overall P = 0.030, Non-linear P = 0.567) (Fig. 3B).
Fig. 2.
The associations between oxidative balance score and retinopathy. Crudel Model (no covariates were adjusted); Multivariable Model (age, sex, race/ethnicity, education level, diabetes, and BMI were adjusted)
Fig. 3.
The non-linear relationship between oxidative balance score and retinopathy. (A) No covariates were adjusted, (B) age, sex, race/ethnicity, education level, diabetes, and BMI were adjusted, C), The non-linear relationship between Oxidative Balance Score and Retinopathy in Diabetes Participants; D), in Non-Diabetes Participants; E) The non-linear relationship between Oxidative Balance Score and Death in Retinopathy Patients, no covariates were adjusted, F) age, sex, race/ethnicity, education level, diabetes, and BMI were adjusted
Among diabetic populations, the univariate model results showed that compared to the T1 group of OBS, the risk of developing retinopathy in the T3 group was reduced by 41% (OR 0.59, 95% CI = 0.40–0.87, P = 0.008). The RCS analysis indicated a significant association between OBS and retinopathy, with a decreasing trend (Overall P = 0.016, Non-linear P = 0.106) (Fig. 3C). Among non-diabetic populations, compared to the T1 group of OBS, the risk of developing retinopathy in the T3 group was reduced by 38% (OR 0.62, 95% CI = 0.42–0.90, P = 0.014) (Table 3). RCS analysis showed a significant association between OBS and retinopathy, with a decreasing trend (Overall P = 0.048, Non-linear P = 0.500) (Fig. 3D).
Table 3.
Association between weighted OBS and retinopathy in diabetes and non-diabetes populations
| Diabetes Population | Non-Diabetes Population | |||||||
|---|---|---|---|---|---|---|---|---|
| Univariable model | Multivariable model | Univariable model | Multivariable model | |||||
| OR (95% CI) | P-value | OR (95% CI) | P-value | OR (95% CI) | P-value | OR (95% CI) | P-value | |
| OBS | ||||||||
| T1 | Reference | Reference | Reference | Reference | ||||
| T2 | 0.88(0.62, 1.25) | 0.500 | 1.00(0.65, 1.55) | 0.900 | 0.71(0.48, 1.05) | 0.090 | 0.85(0.59,1.21) | 0.400 |
| T3 | 0.59(0.40, 0.87) | 0.008 | 0.66(0.40, 1.06) | 0.089 | 0.62(0.42, 0.90) | 0.014 | 0.72(0.54,1.13) | 0.200 |
Notes: OR, Odds Ratio; CI, confidence intervals; Univariable model: No covariates were adjusted; Multivariable model: Adjusted for age, sex, race/ethnicity, education attainment, BMI, smoking status, drinking status and hypertension
Association between OBS and risk of death in patients with retinopathy
A multivariable Cox regression analysis revealed a negative correlation between OBS in patients with retinal lesions and the risk of all-cause mortality. In the crude model, compared to the T1 group, the overall risk of death in the T3 group decreased by 44% (HR 0.56, 95%CI 0.35–0.89, P = 0.015) (eFigure 1). After adjusting for confounding factors, the overall risk of death in the T3 group decreased by 60% (HR 0.40, 95%CI = 0.24–0.66, P < 0.001). More specific details can be seen in eFigure 2 .
Additionally, the RCS analysis shows a significant relationship between OBS of retinal disease and the risk of death in both the crude model (Overall P = 0.014, Non-linear P = 0.356) (Fig. 3E) and the adjusted model (Overall P < 0.001, Non-linear P = 0.018) (Fig. 3F), but the nonlinear relationship was not significant. Kaplan-Meier survival curves further demonstrated significant statistical differences in all-cause mortality rates among the three quartiles of OBS in patients with retinal lesions (P = 0.04) (Fig. 4A).
Fig. 4.
Kaplan-Meier survival curves on oxidative balance score and all-cause mortality in patients with retinopathy. (A) No death time were deleted, (B) Death time < 90 days were deleted
Sensitivity analysis
In the sensitivity analysis, we excluded retinopathy patients with a time of death less than 90 days. In the crude model, compared to the T1 group, the overall risk of death in the T3 group decreased by 43% (HR 0.57, 95%CI 0.35–0.91, P = 0.019) (eFigure 3). In the adjusted model, compared to T1 group), the overall risk of death in the T3 group decreased by 60% (HR 0.40, 95%CI 0.24–0.67, P < 0.001). More details can be obtained in eFigure 4. Additionally, the Kaplan-Meier survival curves plotted showed significant statistical differences among the three quartiles of OBS for overall mortality rates in retinopathy patients (P = 0.047) (Fig. 4B). Therefore, the results of our sensitivity analysis are consistent with the above findings.
Subgroup analysis and interaction
This study conducted a stratified analysis by race, age, gender, diabetes status, BMI, and education level. The findings revealed a negative association between OBS and retinopathy across most subgroups. This inverse relationship was particularly evident among individuals aged 18–65 years (OR 0.82, 95% CI = 0.72–0.95, P = 0.007) and those with a university education or higher (OR 0.83, 95% CI = 0.69–0.99, P = 0.040). Additionally, a significant interaction between OBS and age was observed (P for interaction < 0.05). Further details are provided in Fig. 5.
Fig. 5.
Subgroup analysis of the association of the oxidative balance score and retinopathy. Each stratifcation was adjusted for race, age, sex, diabetes, BMI and education level
Discussion
In recent years, a growing body of epidemiological evidence has suggested oxidative stress may play a key role in the pathogenesis of diabetic retinopathy. While previous experimental studies have implicated oxidative damage in the development and progression of diabetic retinopathy [4], evidence from human observational studies directly linking oxidative biomarkers to clinically relevant outcomes has been limited [8, 17]. To address this knowledge gap, we conducted a population-based longitudinal cohort study of 3,287 U.S. adults, investigating the association between an integrated OBS and risk as well as severity of retinopathy.
In our study, we found significant associations between OBS and retinopathy across various population groups. For the overall population, logistic regression analysis revealed statistically significant relationships between OBS and retinopathy in both Model 1 and Model 2. This suggests that OBS may be an important predictor of retinopathy risk in the general population.
Among individuals with abnormal blood glucose levels, we observed a significant association between OBS and retinopathy in crude model, while the adjusted model showed a trend towards significance. RCS analysis further supported these findings, with univariate analysis showing significant associations and multivariate analysis demonstrating a trend. For individuals without abnormal blood glucose levels, our analysis similarly highlighted the importance of OBS. Both crude model and adjusted model showed significant associations between OBS and retinopathy, with adjusted model demonstrating a trend. RCS analysis results were comparable to those in the abnormal blood glucose group, with significant associations in univariate analysis and a trend in multivariate analysis.
The results of the subgroup analysis and interaction tests indicated that the negative association between OBS and retinopathy was more pronounced in younger individuals and those with higher educational level. There was also a significant interaction between OBS and age. This suggests that the protective effects of OBS may vary across different age groups, with younger individuals potentially benefiting more from higher OBS levels. This could be related to their higher efficiency in antioxidant intake and metabolic status. Additionally, higher educational attainment may be associated with healthier lifestyles and dietary habits, further enhancing the protective effects of OBS.
Recent studies have increasingly found that oxidative stress is not only a core pathological mechanism of retinopathy but also closely linked to the severity of retinal damage [24, 25]. For instance, research by Madsen et al. demonstrated that higher levels of oxidative stress are directly associated with the accelerated progression of retinopathy [26]. These studies align with our findings, further suggesting that optimizing dietary and lifestyle-based antioxidant factors may help prevent or slow the progression of retinopathy. Additionally, research has revealed that the interaction between oxidative stress and inflammation plays a crucial role in the onset and progression of retinopathy [27], which further supports the use of an integrated oxidative balance score in our study to assess overall risk. While previous studies have provided insights into oxidative stress and diabetic retinopathy, there are key differences between our study and prior literature. Firstly, most earlier observational studies focused exclusively on participants with existing diabetes or diabetic retinopathy [17–19]. In contrast, our study included both diabetic and non-diabetic individuals from the general population. The nationally representative sampling enhances generalizability of the findings. Secondly, the majority of preceding studies utilized only one or several circulating biomarkers such as 8-OHdG and isoprostanes to reflect oxidative stress status [20, 21]. However, our study employed an integrated oxidative balance score incorporating both pro- and anti-oxidants from dietary, lifestyle and metabolic factors. This provides a more comprehensive assessment of redox homeostasis. Thirdly, few prior studies examined hard clinical outcomes such as incident diabetic retinopathy or progression to vision loss [18]. We investigated the relationship between oxidative balance score and risk as well as severity of diabetic retinopathy over long-term follow-up. The prospective design and standardized outcome ascertainment are major strengths.
Oxidative stress exerts multifaceted effects in the complex pathogenesis of diabetic retinopathy. First, the aldose reductase pathway converts glucose into sorbitol and subsequently into fructose via an NADPH-dependent mechanism, but in diabetic patients, chronic hyperglycemia excessively activates this pathway, leading to NADPH depletion, overproduction of reactive oxygen species (ROS) through NADH oxidase, and mitochondrial overproduction of superoxide and downstream ROS, which together trigger oxidative injury to retinal capillary cells [4, 22, 23]. Excessive ROS leads to activation of major biochemical pathways implicated in diabetic retinopathy, including increased polyol flux through aldose reductase, augmented advanced glycation end-product (AGE) formation, protein kinase C activation, and hexosamine pathway flux [24–26]. ROS also stimulate inflammation by upregulating release of adhesion molecules, cytokines and vascular endothelial growth factor (VEGF). This results in breakdown of the blood-retinal barrier, leukostasis and pathological angiogenesis [27–29]. Additionally, oxidative stress causes apoptotic death of retinal endothelial cells and pericytes through DNA damage and altered gene expression [30, 31].
Antioxidant nutrients and compounds may mitigate the multifaceted effects of oxidative stress through various mechanisms. Vitamin C, vitamin E, carotenoids and polyphenols can directly neutralize ROS and prevent downstream oxidative reactions [32, 33]. Other antioxidants like taurine and lipoic acid help regenerate depleted antioxidants such as glutathione. Certain antioxidants also activate the Nrf2-KEAP1 signaling pathway to induce endogenous antioxidant enzymes [34, 35]. These combined effects on redox homeostasis likely underlie the observed protective associations between higher antioxidant exposure and risk as well as severity of retinopathy.
Among patients with existing retinopathy, we found those with higher OBS had substantially lower long-term mortality risk. Several mechanistic pathways may underlie this observed association between oxidative balance and risk of death. Firstly, oxidative stress leads to retinal capillary cell apoptosis and breakdown of the blood-retinal barrier, resulting in progression of diabetic retinopathy and severe vision loss [36, 37]. Visual impairment is an established predictor of higher mortality, which may partly mediate the link between lower antioxidant status and increased mortality [38]. Secondly, optimal oxidative balance protects against micro- and macrovascular complications of diabetes through shared endpoints of endothelial dysfunction, low-grade inflammation and accelerated atherosclerosis [4, 24]. Diabetic nephropathy, cardiovascular disease, stroke and related complications that are more prevalent in those with poorer antioxidant status could contribute to higher mortality [39, 40]. Finally, beyond diabetes, higher overall antioxidant exposure has been associated with lower risks of cancer, lung disease, infection, dementia and frailty – all of which can directly impact survival [41–43]. The cumulative effects of enhanced antioxidant capacity on this multitude of age-related conditions likely contributes to improved prognosis.
The application prospects of antioxidant interventions may encompass the following aspects: Lifestyle factors such as diet, smoking and exercise strongly influence oxidative balance, suggesting potential for lifestyle interventions to enhance antioxidant status and mitigate development and progression of retinopathy [8, 44, 45].
Epidemiological studies indicate higher intake of antioxidant-rich foods like fruits, vegetables, whole grains, nuts and fish are associated with lower incidence of diabetic retinopathy, whereas cigarette smoking increases risk [6, 7, 46]. Clinical trials also demonstrate antioxidant supplementation improves redox status in diabetic patients [44, 47]. However, direct evidence on the effects of lifestyle changes and antioxidant supplements on hard clinical outcomes like diabetic retinopathy remains limited [48, 49]. Further rigorously designed lifestyle and supplementation intervention studies are warranted to determine whether enhancing antioxidant exposure can effectively help prevent onset and progression of diabetic retinopathy.
This prospective cohort study provides initial evidence linking higher cumulative antioxidant exposure assessed through OBS to lower risk and severity of retinopathy in U.S. adults. Our study highlights the potential role of oxidative stress in contributing to human retinopathy based on a composite measure of oxidative balance, rather than individual biomarkers. This underscores the importance of overall redox homeostasis beyond isolated antioxidant levels. The nationwide sampling and long-term follow-up for clinical outcomes are also major strengths. However, this study has several limitations. First, although it is based on a nationally representative sample of U.S. adults, caution should be exercised when generalizing the findings to populations with different demographic or cultural characteristics. Second, the assessment of dietary components relies on self-reported data, which may introduce recall bias and compromise the accuracy of antioxidant exposure estimates. Additionally, the evaluation of diabetic retinopathy through retinal imaging may be subject to diagnostic variability due to differences in imaging techniques or the quality of the images obtained [50]. Lastly, the study did not fully account for potential comorbidities or the effects of drug therapies that could influence oxidative stress, even though existing evidence suggests that multifactorial interventions can improve clinical outcomes.
Future researches should focus on exploring the effects of individual antioxidants or specific antioxidant combinations to identify the most effective interventions for retinopathy. Randomized controlled trials are especially needed to test whether targeted antioxidant therapies could reduce the risk or slow the progression of retinopathy. Further studies should also evaluate how different doses, durations, or formulations of antioxidants may influence outcomes, particularly in populations at high risk, such as individuals with poorly controlled diabetes or those with pre-existing microvascular complications. Moreover, investigating the role of oxidative stress modulation in diverse populations, including ethnic minorities and those with varying genetic predispositions, could provide insights into potential personalized treatment approaches. Finally, exploring the interaction between antioxidants and other metabolic pathways involved in diabetic retinopathy, such as inflammation and glucose metabolism, could help to clarify the underlying mechanisms and optimize therapeutic strategies.
Conclusion
In summary, this study demonstrates that higher antioxidant exposure, as measured by the oxidative balance score, is associated with a reduced risk and severity of retinopathy. Oxidative balance may serve as both a useful risk predictor and a target for intervention. Further research is needed to clarify the underlying mechanisms and explore optimizing antioxidant intake through diet, supplements, and lifestyle changes to prevent vision loss and improve outcomes.
Electronic supplementary material
Below is the link to the electronic supplementary material.
Acknowledgements
The authors sincerely expressed their gratitude to all members who took part in this study.
Author contributions
XL, YC and YL analyzed and interpreted the patient data regarding the diabetic retinopathy and the oxidative balance score. YL, WS and JL participated in study design, coordination and data collection. NC and JC conceptualized and reviewed this study. All authors read and approved the final manuscript.
Funding
The authors declare that they did not receive any funding from any source.
Data availability
The data can be downloaded for free from the website: https://www.cdc.gov/nchs/nhanes/index.htm.
Declarations
Ethics approval and consent to participate
The NHANES agreement has been reviewed and approved by the NCHS Research Ethics Committee. All participants provided written informed consent before participating.
Consent for publication
All participants provided written informed consent before participation.
Competing interests
The authors declare that they have no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Xiangliang Liu, Yu Chang and Yuguang Li contributed equally to this work and shared first authorship.
Contributor Information
Naifei Chen, Email: chennaifei@jlu.edu.cn.
Jiuwei Cui, Email: cuijw@jlu.edu.cn.
References
- 1.Causes of blindness and vision impairment. In 2020 and trends over 30 years, and prevalence of avoidable blindness in relation to VISION 2020: the right to sight: an analysis for the global burden of disease study. Lancet Global Health. 2021;9(2):e144–60. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Yau JW, Rogers SL, Kawasaki R, Lamoureux EL, Kowalski JW, Bek T, Chen SJ, Dekker JM, Fletcher A, Grauslund J, et al. Global prevalence and major risk factors of diabetic retinopathy. Diabetes Care. 2012;35(3):556–64. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Wat N, Wong RL, Wong IY. Associations between diabetic retinopathy and systemic risk factors. Hong Kong Med J. 2016;22(6):589–99. [DOI] [PubMed] [Google Scholar]
- 4.Madsen-Bouterse SA, Kowluru RA. Oxidative stress and diabetic retinopathy: pathophysiological mechanisms and treatment perspectives. Reviews Endocr Metabolic Disorders. 2008;9(4):315–27. [DOI] [PubMed] [Google Scholar]
- 5.Abcouwer SF. Angiogenic factors and cytokines in diabetic retinopathy. J Clin Cell Immunol. 2013;Suppl 111:1–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Millen AE, Klein R, Folsom AR, Stevens J, Palta M, Mares JA. Relation between intake of vitamins C and E and risk of diabetic retinopathy in the atherosclerosis risk in communities study. Am J Clin Nutr. 2004;79(5):865–73. [DOI] [PubMed] [Google Scholar]
- 7.Ganesan S, Raman R, Kulothungan V, Sharma T. Influence of dietary-fibre intake on diabetes and diabetic retinopathy: Sankara Nethralaya-diabetic retinopathy epidemiology and molecular genetic study (report 26). Clin Exp Ophthalmol. 2012;40(3):288–94. [DOI] [PubMed] [Google Scholar]
- 8.Sasaki M, Kawasaki R, Rogers S, Man RE, Itakura K, Xie J, Flood V, Tsubota K, Lamoureux E, Wang JJ. The associations of dietary intake of polyunsaturated fatty acids with diabetic retinopathy in well-controlled diabetes. Investig Ophthalmol Vis Sci. 2015;56(12):7473–9. [DOI] [PubMed] [Google Scholar]
- 9.Lee JH, Son DH, Kwon YJ. Association between oxidative balance score and new-onset hypertension in adults: a community-based prospective cohort study. Front Nutr. 2022;9:1066159. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Cheng S, Han Y, Jiang L, Lan Z, Liao H, Guo J. Associations of oxidative balance score and visceral adiposity index with risk of ischaemic heart disease: a cross-sectional study of NHANES, 2005–2018. BMJ open. 2023;13(7):e072334. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Annor FB, Goodman M, Okosun IS, Wilmot DW, Il’yasova D, Ndirangu M, Lakkur S. Oxidative stress, oxidative balance score, and hypertension among a racially diverse population. J Am Soc Hypertension: JASH. 2015;9(8):592–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Talavera-Rodriguez I, Fernandez-Lazaro CI, Hernández-Ruiz Á, Hershey MS, Galarregui C, Sotos-Prieto M, de la Fuente-Arrillaga C, Martínez-González M, Ruiz-Canela M. Association between an oxidative balance score and mortality: a prospective analysis in the SUN cohort. Eur J Nutr. 2023;62(4):1667–80. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Bryl A, Mrugacz M, Falkowski M, Zorena K. A Mediterranean diet may be protective in the development of diabetic retinopathy. Int J Mol Sci. 2023;24(13). [DOI] [PMC free article] [PubMed]
- 14.Zhang W, Peng SF, Chen L, Chen HM, Cheng XE, Tang YH. Association between the oxidative balance score and telomere length from the National Health and Nutrition Examination Survey 1999–2002. Oxidative medicine and cellular longevity. 2022;2022:1345071. [DOI] [PMC free article] [PubMed]
- 15.Alfonso-Muñoz EA, Burggraaf-Sánchez de Las Matas R, Mataix Boronat J, Molina Martín JC, Desco C. Role of oral antioxidant supplementation in the current management of diabetic retinopathy. Int J Mol Sci. 2021;22(8). [DOI] [PMC free article] [PubMed]
- 16.Grading diabetic retinopathy. From stereoscopic color fundus photographs–an extension of the modified Airlie House classification. ETDRS report number 10. Early Treatment Diabetic Retinopathy Study Research Group. Ophthalmology. 1991;98(5 Suppl):786–806. [PubMed] [Google Scholar]
- 17.Song P, Yu J, Chan KY, Theodoratou E, Rudan I. Prevalence, risk factors and burden of diabetic retinopathy in China: a systematic review and meta-analysis. J Global Health. 2018;8(1):010803. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Ting DS, Cheung GC, Wong TY. Diabetic retinopathy: global prevalence, major risk factors, screening practices and public health challenges: a review. Clin Exp Ophthalmol. 2016;44(4):260–77. [DOI] [PubMed] [Google Scholar]
- 19.Sasongko MB, Wong TY, Nguyen TT, Kawasaki R, Jenkins A, Shaw J, Wang JJ. Serum apolipoprotein AI and B are stronger biomarkers of diabetic retinopathy than traditional lipids. Diabetes Care. 2011;34(2):474–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Graille M, Wild P, Sauvain JJ, Hemmendinger M, Guseva Canu I, Hopf NB. Urinary 8-OHdG as a biomarker for oxidative stress: a systematic literature review and meta-analysis. Int J Mol Sci. 2020;21(11). [DOI] [PMC free article] [PubMed]
- 21.Montuschi P, Barnes PJ, Roberts LJ 2. nd: Isoprostanes: markers and mediators of oxidative stress. FASEB journal: official publication of the Federation of American Societies for Experimental Biology. 2004;18(15):1791–800. [DOI] [PubMed]
- 22.Sasaki M, Ozawa Y, Kurihara T, Kubota S, Yuki K, Noda K, Kobayashi S, Ishida S, Tsubota K. Neurodegenerative influence of oxidative stress in the retina of a murine model of diabetes. Diabetologia. 2010;53(5):971–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Caturano A, D’Angelo M, Mormone A, Russo V, Mollica MP, Salvatore T, Galiero R, Rinaldi L, Vetrano E, Marfella R, et al. Oxidative stress in type 2 diabetes: impacts from pathogenesis to lifestyle modifications. Curr Issues Mol Biol. 2023;45(8):6651–66. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Giacco F, Brownlee M. Oxidative stress and diabetic complications. Circul Res. 2010;107(9):1058–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Madsen-Bouterse SA, Mohammad G, Kanwar M, Kowluru RA. Role of mitochondrial DNA damage in the development of diabetic retinopathy, and the metabolic memory phenomenon associated with its progression. Antioxid Redox Signal. 2010;13(6):797–805. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Du Y, Miller CM, Kern TS. Hyperglycemia increases mitochondrial superoxide in retina and retinal cells. Free Radic Biol Med. 2003;35(11):1491–9. [DOI] [PubMed] [Google Scholar]
- 27.Mamputu JC, Renier G. Advanced glycation end products increase, through a protein kinase C-dependent pathway, vascular endothelial growth factor expression in retinal endothelial cells. Inhibitory effect of gliclazide. J Diabetes Complicat. 2002;16(4):284–93. [DOI] [PubMed] [Google Scholar]
- 28.Cheng HM, González RG. The effect of high glucose and oxidative stress on lens metabolism, aldose reductase, and senile cataractogenesis. Metab Clin Exp. 1986;35(4 Suppl 1):10–4. [DOI] [PubMed] [Google Scholar]
- 29.Kowluru RA. Diabetic retinopathy: mitochondrial dysfunction and retinal capillary cell death. Antioxid Redox Signal. 2005;7(11–12):1581–7. [DOI] [PubMed] [Google Scholar]
- 30.Kowluru RA, Koppolu P, Chakrabarti S, Chen S. Diabetes-induced activation of nuclear transcriptional factor in the retina, and its inhibition by antioxidants. Free Radic Res. 2003;37(11):1169–80. [DOI] [PubMed] [Google Scholar]
- 31.Du Y, Sarthy VP, Kern TS. Interaction between NO and COX pathways in retinal cells exposed to elevated glucose and retina of diabetic rats. Am J Physiol Regul Integr Comp Physiol. 2004;287(4):R735–741. [DOI] [PubMed] [Google Scholar]
- 32.Li C, Miao X, Li F, Wang S, Liu Q, Wang Y, Sun J. Oxidative stress-related mechanisms and antioxidant therapy in diabetic retinopathy. Oxidative medicine and cellular longevity. 2017;2017:9702820. [DOI] [PMC free article] [PubMed]
- 33.Xiaojing L, Bijun, Zhu, Haidong, Zou, Daode H, Qing C, GJGsAf et al. Carbamylated erythropoietin mediates retinal neuroprotection in streptozotocin-induced early-stage diabetic rats. 2015;253(8):1263–72. [DOI] [PubMed]
- 34.Zhu X, Liu Q, Wang M, Liang M, Yang X, Xu X, Zou H, Qiu J. Activation of Sirt1 by resveratrol inhibits TNF-α induced inflammation in fibroblasts. PLoS ONE. 2011;6(11):e27081. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Ungvari Z, Bagi Z, Feher A, Recchia FA, Sonntag WE, Pearson K, de Cabo R, Csiszar A. Resveratrol confers endothelial protection via activation of the antioxidant transcription factor Nrf2. Am J Physiol Heart Circ Physiol. 2010;299(1):H18–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Kowluru RA, Abbas SN. Diabetes-induced mitochondrial dysfunction in the retina. Investig Ophthalmol Vis Sci. 2003;44(12):5327–34. [DOI] [PubMed] [Google Scholar]
- 37.Silva KC, Rosales MA, Biswas SK, Lopes de Faria JB, Lopes de Faria JM. Diabetic retinal neurodegeneration is associated with mitochondrial oxidative stress and is improved by an angiotensin receptor blocker in a model combining hypertension and diabetes. Diabetes. 2009;58(6):1382–90. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Zhang X, Saaddine JB, Lee PP, Grabowski DC, Kanjilal S, Duenas MR, Narayan KM. Eye care in the United States: do we deliver to high-risk people who can benefit most from it? Archives of ophthalmology (Chicago, Ill: 1960). 2007;125(3):411–418. [DOI] [PubMed]
- 39.Boyce G, Button E, Soo S, Wellington C. The pleiotropic vasoprotective functions of high density lipoproteins (HDL). J Biomedical Res. 2017;32(3):164–82. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Liu SX, Hou FF, Guo ZJ, Nagai R, Zhang WR, Liu ZQ, Zhou ZM, Zhou M, Xie D, Wang GB, et al. Advanced oxidation protein products accelerate atherosclerosis through promoting oxidative stress and inflammation. Arterioscler Thromb Vasc Biol. 2006;26(5):1156–62. [DOI] [PubMed] [Google Scholar]
- 41.Mühlhöfer A, Mrosek S, Schlegel B, Trommer W, Rozario F, Böhles H, Schremmer D, Zoller WG, Biesalski HK. High-dose intravenous vitamin C is not associated with an increase of pro-oxidative biomarkers. Eur J Clin Nutr. 2004;58(8):1151–8. [DOI] [PubMed] [Google Scholar]
- 42.Raphalle Varraso WCW, Carlos A. Camargo JJAJoE: prospective study of dietary fiber and risk of chronic obstructive pulmonary disease among US women and men. 2010;171(7). [DOI] [PMC free article] [PubMed]
- 43.Gusti AMT, Qusti SY, Alshammari EM, Toraih EA, Fawzy MS. Antioxidants-related Superoxide Dismutase (SOD), Catalase (CAT), Glutathione Peroxidase (GPX), Glutathione-S-Transferase (GST), and Nitric Oxide Synthase (NOS) gene variants analysis in an obese population: a preliminary case-control study. Antioxidants (Basel, Switzerland). 2021;10(4). [DOI] [PMC free article] [PubMed]
- 44.Ho JI, Ng EY, Chiew Y, Koay YY, Chuar PF, Phang SCW, Ahmad B, Kadir KA. The effects of vitamin E on non-proliferative diabetic retinopathy in type 2 diabetes mellitus: are they sustainable with 12 months of therapy. SAGE open Med. 2022;10:20503121221095324. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Mohammad G, Kowluru RA. Novel role of mitochondrial matrix metalloproteinase-2 in the development of diabetic retinopathy. Investig Ophthalmol Vis Sci. 2011;52(6):3832–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Tanaka S, Yoshimura Y, Kawasaki R, Kamada C, Tanaka S, Horikawa C, Ohashi Y, Araki A, Ito H, Akanuma Y, et al. Fruit intake and incident diabetic retinopathy with type 2 diabetes. Epidemiol (Cambridge Mass). 2013;24(2):204–11. [DOI] [PubMed] [Google Scholar]
- 47.Ceriello A. New insights on oxidative stress and diabetic complications may lead to a causal antioxidant therapy. Diabetes Care. 2003;26(5):1589–96. [DOI] [PubMed] [Google Scholar]
- 48.Raman R, Ganesan S, Pal SS, Kulothungan V, Sharma T. Prevalence and risk factors for diabetic retinopathy in rural India. Sankara Nethralaya Diabetic Retinopathy Epidemiology and Molecular Genetic Study III (SN-DREAMS III), report 2. BMJ open Diabetes Res care. 2014;2(1):e000005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49.Mansour SE, Browning DJ, Wong K, Flynn HW Jr., Bhavsar AR. The evolving treatment of diabetic retinopathy. Clinical ophthalmology (Auckland, NZ). 2020;14:653–78. [DOI] [PMC free article] [PubMed]
- 50.Sasso FC, Simeon V, Galiero R, Caturano A, De Nicola L, Chiodini P, Rinaldi L, Salvatore T, Lettieri M, Nevola R, et al. The number of risk factors not at target is associated with cardiovascular risk in a type 2 diabetic population with albuminuria in primary cardiovascular prevention. Post-hoc analysis of the NID-2 trial. Cardiovasc Diabetol. 2022;21(1):235. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data can be downloaded for free from the website: https://www.cdc.gov/nchs/nhanes/index.htm.