Association between dietary antioxidant quality score and cognitive function in older adults: A cross-sectional study based on NHANES 2011–2014

Abstract

The aim of this study was to examine the connection between the dietary antioxidant quality score (DAQS) and cognitive impairment in older adults. This cross-sectional study of 2713 older adults aged ≥ 60 years from the 2011 to 2014 National Health and Nutrition Examination Survey cohort evaluated cognitive performance at quartile thresholds using the digit symbol substitution test, the animal fluency test, and the consortium to establish a registry for Alzheimer disease word learning test. Vitamins A, C, E, zinc, selenium, and magnesium consumption were used to calculate the DAQS score. Multiple linear regression and logistic regression analyses were used to evaluate the association between DAQS and the cognitive capacities of senior citizens. In the fully adjusted model, higher DAQS was linked to less cognitive impairment (odds ratio = 0.34, 95% confidence interval: 0.22–0.53). According to our research, older adults’ cognitive decline is adversely correlated with higher DAQS, and improving their antioxidant intake may help them improve cognitive function.

1. Introduction

Cognitive function pertains to the brain’s capacity to perceive and comprehend objective reality, encompassing various faculties such as language, cognition, focus, and memory. The cognitive abilities of older individuals experience substantial age-related changes, which can have a profound impact on their quality of life and hinder their daily functioning.^[1] Consequently, the preservation of regular cognitive function in elderly individuals has emerged as a primary area of interest in public health research and a significant worry for the aging demographic. Recent projections estimate that by 2050, over 150 million people globally will be living with dementia. Therefore, identifying modifiable risk factors is urgent.^[2]

An imbalance between excessive oxidation (free radicals) and insufficient antioxidant system destruction of free radicals as an internal defensive mechanism is known as oxidative stress.^[3] Oxidative stress has been linked to the pathophysiology of numerous illnesses, including kidney disease, rheumatoid arthritis, and cardiovascular disease.^[3–7] It has also been demonstrated to be a significant contributor to cognitive impairment. On the one hand, atherosclerosis and thrombosis can cause cognitive impairment through cerebral infarction or white matter damage, and oxidative stress is now thought to be the primary cause of atherosclerosis.^[8] On the other hand, because the brain needs a lot of oxygen and is high in fats, it has a low antioxidant capacity, which renders neuronal cells more vulnerable to oxidative damage.^[9] Increasing antioxidant consumption in the diet may help avoid oxidative stress, according to several research studies.^[3,4] Systematic reviews have further highlighted that adopting sustainable, nutrient-rich dietary patterns is a viable strategy for promoting health in the United States (US) population.^[10] Mineral supplements are essential for maintaining brain and cognitive function in addition to antioxidant vitamins.^[11] Mineral shortages have been linked to compromised antioxidant enzyme synthesis and function, according to numerous studies.^[12] Given that individuals eat complex food components rather than isolated nutrients on a regular basis and that multiple dietary antioxidants from related sources may have intricate synergistic effects in the body,^[13] a “single nutrient” approach may not be sufficient, given the complex relationships between nutrients. The effectiveness of the dietary antioxidant quality score (DAQS) has been demonstrated in numerous studies, and this approach focuses on 6 antioxidant nutrients: zinc, selenium, magnesium, and vitamins A, C, and E.^[14,15] However, the relationship between DAQS and cognitive impairment remains unclear. In summary, this study aims to investigate the correlation between DAQS and cognitive impairment in older adults using nationally representative data from the US population.

2. Methods

2.1. Study population

The National Health and Nutrition Examination Survey (NHANES) database provided the data for this investigation, which evaluates the health and nutritional condition of individuals in the US, including both adults and children. The Centers for Disease Control and Prevention conducts NHANES, a survey that involves selecting 5000 participants each year using a sophisticated method called complex probability sampling. The sampling design for each 2-year cycle employs a stratified multistage probability approach, which involves conducting standardized interviews and physical examinations. To obtain additional information regarding NHANES, please visit the Centers for Disease Control and Prevention website at (https://wwwn.cdc.gov/nchs/nhanes/). The study protocol was approved by the National Center for Health Statistics Research Ethics Review Board, and all participants provided written informed consent. This study is a secondary analysis of publicly available, de-identified data from NHANES; therefore, additional ethical approval was not required. The study specifically examined 2713 individuals who were 60 years of age or older. These participants were selected from the NHANES database, and the data used was collected between 2011 and 2014 (Fig. 1). It was not appropriate or possible to involve patients or the public in the design, conduct, reporting, or dissemination plans of our research.

Figure 1.

Flowchart showing the selection of the study population. N = number of participants, NHANES = National Health and Nutrition Examination Survey.

2.2. DAQS

The DAQS was determined by contrasting daily dietary consumption with the daily recommended intake (DRI) specified in the 2015 to 2020 Dietary Guidelines (https://health.gov/sites/default/files/2019-09/2015–2020_Dietary_Guidelines.pdf). Six micronutrients have been identified: vitamins A, C, E, selenium, magnesium, and zinc. The DAQS assigns a value of “0” to each micronutrient if the intake is < two-thirds of the DRI and a value of “1” if the intake is at or above two-thirds of the DRI. The individual scores for each of the 6 micronutrients were added together to calculate a total DAQS sum score, ranging from 0 to 6. Greater scores indicated greater levels of dietary antioxidants. The research classified DAQS into 3 specific categories according to their quality: low (0–2 points), moderate (3–4 points), and high (5–6 points). Each micronutrient and total calorie intake were calculated using the database’s cumulative sum of the data acquired during the preceding 24 hours.

2.3. Assessment of cognitive function

The cognitive assessments were administered to participants aged 60 and above during the 2011 to 2014 NHANES survey, specifically by the mobile examination center. The assessments comprised the consortium to establish a registry for Alzheimer disease (CERAD) word learning subtest, the animal fluency test (AFT), and the digit symbol substitution test (DSST).

Subjects agreed to audio recordings of the administration of these tests for quality control purposes. The CERAD test, the AFT, and the DSST were transcribed and scored by 2 interviewers from audio recordings of the interviews conducted in both English and Spanish. This was typically done on the same day. Consultants who are fluent in other languages transcribed and scored tests in those languages exactly as they were spoken. An impartial entity resolved any conflicting scores. As part of the data collection process, around 10% of the recorded interviews were subjected to independent review to ensure accuracy.

The CERAD test comprises 3 consecutive learning trials and 1 delayed recall trial to evaluate the immediate and delayed acquisition of new linguistic information.^[16] Participants verbally recited groups of 10 unrelated words and promptly attempted to recall as many words as they could. The delayed recall test was conducted subsequent to the administration of the AFT and DSST tests. The CERAD score was determined by adding up the scores of each learning trial, which ranged from 0 to 40, and has been used in a variety of studies.^[16]

The AFT is a component of the executive function assessment that evaluates categorical verbal fluency. It involves asking the participant to list as many animals as they can within a 1-minute time frame. The total number of accurate responses provided determines the score.^[17] The DSST is a component of the Wechsler adult intelligence scale that evaluates the speed at which information is processed, the ability to maintain focus over time, and the capacity to hold and manipulate information in the mind. Participants must match symbols to numbers within a time limit of 2 minutes. They can use the keys provided at the top of a sheet of paper that contains 133 numbered boxes. The score spans from 0 to 133, representing the total count of accurate matches.^[18]

There is no universally accepted threshold for determining low cognitive performance on the CERAD, AFT, and DSST tests. Hence, we employed the 25th percentile, also known as the lowest quartile, as the threshold, which aligns with the approach utilized in existing research.^[19,20] The study employed participants who met the criteria of having a total CERAD score of 21 or below, an AFT score of 13 or below, and a DSST score of 34 or below in order to distinguish individuals with cognitive impairment.

2.4. Covariates

We analyzed multiple factors that may affect the outcomes, including age, gender, race, level of education, body mass index (BMI, kg/m²), poverty-to-income ratio (PIR), alcohol use, smoking, hypertension, diabetes mellitus, and consumption of certain vitamins (A, C, and E) and minerals (zinc, magnesium, and selenium). The PIR assesses household income relative to the poverty threshold, acting as a standard for income level evaluation; it is employed to establish income thresholds. Hypertension was included as a covariate because of its established association with cognitive decline.^[21] Hypertension and diabetes were identified based on a physician’s self-reported diagnosis. The smoking status was determined based on the following criteria: “never smoked” refers to individuals who have never participated in smoking or have smoked fewer than 100 cigarettes in their lifetime; “ever smoked” pertains to those who have consumed 100 or more cigarettes but are not currently smoking; “current smoker” describes individuals who have smoked 100 or more cigarettes and are presently active smokers.^[22] Alcohol consumption is defined as the ingestion of at least 12 alcoholic beverages of any type within a year. An individual alcoholic beverage is characterized as 12 ounces of beer, 1.5 ounces of distilled spirits, or 5 ounces of wine.^[20]

2.5. Statistical analysis

The analyses were conducted using R version 3.4.3 (http://www.R-project.org, The R Foundation) and Empower software (www.empowerstats.com; X&Y Solutions, Inc.). The statistical significance level was established at a d value of <0.05. We presented baseline data in the DAQS for patients participating in this study. Continuous variables are represented by the mean value along with the standard error, while categorical variables are represented by the combined ratio n (%). We employed chi-square tests to analyze the P values of the distributions. Next, we employed 3 logistic regression models: Model I without any variable adjustments; Model II with adjustments for specific general variables such as age, sex, and race; and Model III with adjustments for a comprehensive set of variables including age, sex, race, education level, BMI, PIR, smoking status, drinking status, hypertension, and diabetes.

3. Results

3.1. Description of the study population

Table 1 presents the characteristics of 2713 participants, highlighting the differences between those with low cognitive performance (n = 201) and those without (n = 2512). The mean age of all participants was 69.42 years, with 50.76% being female. Additionally, 62.51% had hypertension, and 23.33% had diabetes. The results also indicated that the non-low cognitive performance group had higher mean DAQS scores and higher intakes of vitamin E, vitamin A, vitamin C, magnesium, zinc, and selenium compared to the low cognitive performance group.

Table 1.

Demographic characteristics stratified by cognitive performance (n = 2713).

Variables	Total (n = 2713)	Non- cognitive impairment (n = 2512)	Cognitive impairment (n = 201)	P
DAQS score, Mean (SE)	3.45 ± 1.65	3.50 ± 1.64	2.79 ± 1.59	< .001
DAQS score, n (%)				< .001
0–2	839 (30.93)	744 (29.62)	95 (47.26)
3–4	1040 (38.33)	968 (38.54)	72 (35.82)
5–6	834 (30.74)	800 (31.85)	34 (16.92)
Age, yrs, Mean (SE)	69.42 ± 6.76	69.36 ± 6.74	70.12 ± 7.01	.154
Gender, n (%)				.998
Male	1336 (49.24)	1237 (49.24)	99 (49.25)
Female	1377 (50.76)	1275 (50.76)	102 (50.75)
Race, n (%)				.044
Mexican American	233 (8.59)	217 (8.64)	16 (7.96)
Other Hispanic	274 (10.10)	253 (10.07)	21 (10.45)
Non-Hispanic White	1332 (49.10)	1249 (49.72)	83 (41.29)
Non-Hispanic Black	642 (23.66)	577 (22.97)	65 (32.34)
Other Race - Including Multi-Racial	232 (8.55)	216 (8.60)	16 (7.96)
Education levels, n (%)				.066
< high school	675 (24.88)	618 (24.60)	57 (28.36)
High school	637 (23.48)	581 (23.13)	56 (27.86)
> high school	1401 (51.64)	1313 (52.27)	88 (43.78)
BMI, kg/m2, n (%)				.614
< 25	719 (26.50)	665 (26.47)	54 (26.87)
25–30	973 (35.86)	907 (36.11)	66 (32.84)
> 30	1021 (37.63)	940 (37.42)	81 (40.30)
Drinking, n (%)				.955
Yes	1854 (68.34)	1717 (68.35)	137 (68.16)
No	859 (31.66)	795 (31.65)	64 (31.84)
Smoking, n (%)				.998
Never	1326 (48.88)	1228 (48.89)	98 (48.76)
Former	1047 (38.59)	969 (38.57)	78 (38.81)
Current	340 (12.53)	315 (12.54)	25 (12.44)
High blood pressure, n (%)				.688
Yes	1696 (62.51)	1573 (62.62)	123 (61.19)
No	1017 (37.49)	939 (37.38)	78 (38.81)
Diabetes, n (%)				.096
Yes	633 (23.33)	574 (22.85)	59 (29.35)
No	1956 (72.10)	1824 (72.61)	132 (65.67)
Borderline	124 (4.57)	114 (4.54)	10 (4.98)
PIR	2.62 ± 1.60	2.64 ± 1.60	2.47 ± 1.59	.169
Vitamin A intake, mcg, Mean (SE)	657.93 ± 907.99	668.22 ± 920.38	529.38 ± 725.49	< .001
Vitamin E intake, mg, Mean (SE)	8.11 ± 6.50	8.27 ± 6.59	6.13 ± 4.73	< .001
Vitamin C intake, mg, Mean (SE)	83.26 ± 90.49	84.61 ± 92.11	66.32 ± 64.97	.006
Zinc intake, mg, Mean (SE)	9.88 ± 5.81	9.97 ± 5.84	8.67 ± 5.34	.002
Magnesium intake, mg, Mean (SE)	281.58 ± 139.96	285.11 ± 141.30	237.48 ± 113.23	< .001
Selenium intake, mcg, Mean (SE)	102.04 ± 59.97	103.09 ± 60.87	88.84 ± 45.34	.001

BMI = body mass index, DAQS = dietary antioxidant quality score, n = number of participants, PIR = poverty-to-income ratio, SE = standard error.

3.2. Relationship between DAQS and cognitive impairment

Table 2 depicts the outcomes of a correlation analysis between DAQS and cognitive impairment. The initial model demonstrated a significant inverse correlation between DAQS and cognitive impairment when DAQS was treated as a continuous variable (odds ratio [OR] = 0.77, 95% confidence interval [CI]: 0.70, 0.84, P < .0001). After controlling for age, gender, and ethnicity, a significant association was identified between DAQS and cognitive impairment (OR = 0.78, 95% CI: 0.71, 0.86, P < .0001). After accounting for all potential factors that could influence the results, the link between DAQS and cognitive impairment remained significant (OR = 0.77, 95% CI: 0.70–0.85, P < .0001). This research categorized DAQS into 3 distinct quality levels: low, moderate, and high. The fully adjusted model revealed a significant association between the high-quality DAQS group and a reduced risk of cognitive impairment compared to the low-quality DAQS group (0.34 (95% CI: 0.22–0.53), P < .0001).

Table 2.

Relationship between dietary antioxidant quality score and cognitive impairment.

	Crude Model	P	Model 1	P	Model 2	P
	OR (95% CI)		OR (95% CI)		OR (95% CI)
DAQS score	0.77 (0.70–0.84)	< .0001	0.78 (0.71–0.86)	< .0001	0.77 (0.70–0.85)	< .0001
DAQS group
0–2	1.0		1.0		1.0
3–4	0.58 (0.42–0.80)	.0010	0.61 (0.44–0.85)	.0031	0.60 (0.43–0.84)	.0033
5–6	0.33 (0.22–0.50)	< .0001	0.36 (0.24–0.54)	< .0001	0.34 (0.22–0.53)	< .0001

The Crude Model was unadjusted. Model 1 took race, gender, and age into account. Gender, age, race, education level, PIR, BMI, diabetes, high blood pressure, alcohol use, and smoking were all controlled for in Model 2.

BMI = body mass index, CI = confidence interval, DAQS = dietary antioxidant quality score, OR = odds ratio, PIR = poverty-to-income ratio.

3.3. Subgroup analyses

As illustrated in the forest plot (Fig. 2), the protective effect of a higher DAQS against cognitive impairment was generally consistent across various clinical and demographic subgroups (indicated by ORs consistently below 1.0). This visually demonstrates that the cognitive benefits of dietary antioxidants remain evident regardless of the participants’ baseline characteristics or chronic conditions. Furthermore, interaction tests showed no statistically significant differences across subgroups defined by smoking status, race/ethnicity, drinking status, BMI, and hypertension (all interaction P > .05). However, we did observe significant interactions for gender (interaction P = .0280), education level (interaction P = .0094), and diabetes status (interaction P = .0059). This suggests that while a higher DAQS is generally beneficial, the strength of this protective association is influenced by these factors and appears to be more pronounced in women, nondiabetic individuals, and those with a high school education or above.

Figure 2.

Subgroup analysis of the relationship between the DAQS and the risk of cognitive impairment. The ORs and 95% CIs were calculated by multiple logistic regression analyses adjusted for the factors mentioned in the text. BMI = body mass index, CI = confidence interval, DAQS = dietary antioxidant quality score, OR = odds ratio.

3.4. Nonlinear association between the DAQS and cognitive function

The adjusted smoothing curve in Figure 3 shows a nonlinear relationship characterized by a saturating effect between DAQS and cognitive function. We used a linear regression model to identify this saturating effect between DAQS and cognitive impairment. As shown in Figure 3, we calculated the turning point of DAQS to be 5. When DAQS is below 5, cognitive impairment decreases linearly with increasing DAQS. However, cognitive impairment eventually stabilizes when DAQS reaches the saturation point of 5, indicating that increasing DAQS beyond this point does not further reduce cognitive impairment.

Figure 3.

Smooth curve fitting of the association between the DAQS and the risk of cognitive impairment. The black dashed line represents the adjusted mean for the risk of cognitive impairment, and the red vertical bars represent the 95% CI. DAQS = dietary antioxidant quality score, CI = confidence interval.

4. Discussion

This study investigated the correlation between the DAQS and cognitive performance. The study used a large, randomized national sample from the US. The findings suggest a noteworthy correlation between DAQS and cognitive function in older adults, even after accounting for potential confounders. This provides a rationale for reducing cognitive decline in older adults by adopting an antioxidant-rich diet. This study is notable as it is the first to specifically investigate the relationship between DAQS and cognitive decline in older adults, addressing a gap in previous research that often focused on single nutrients or overall antioxidant capacity.

In addition, smoothing curve and threshold effect analyses revealed a nonlinear relationship characterized by a saturating effect between DAQS and cognitive function. That is, when DAQS is below 5, cognitive impairment decreases linearly with increasing DAQS. However, cognitive impairment eventually stabilizes when DAQS reaches the saturation point of 5, indicating that increasing DAQS beyond this point does not further reduce cognitive impairment. This nuanced finding suggests that while increasing antioxidant intake is beneficial up to a certain level, there may be a plateau effect.

Although direct evidence was previously limited, a growing body of recent literature utilizing the NHANES database has begun to explore this relationship.^[23–25] Our findings align with these recent studies,^[25–29] which consistently report a positive association between composite antioxidant indices and cognitive performance. Notably, our observation of a saturation effect (turning point at 5) is corroborated by similar nonlinear or threshold effects reported in these cohorts,^[26,27] suggesting a biological plateau for antioxidant benefits where moderate but adequate intake may be sufficient for maximal cognitive protection. However, a study of a group of older women failed to confirm a significant correlation between overall dietary antioxidant intake and cognitive performance.^[30] Similarly, 1 study found that older adults’ cognitive function was unaffected by vitamin E administration,^[31] while another found that combined use of vitamin E and vitamin C supplements was associated with better cognitive performance.^[28] The reason for the inconsistency in the results of the above studies may be related to the differences in calculation methods and dietary patterns of the various populations. To summarize, these factors need to be taken into account in future studies in order to make a comprehensive assessment.

The DAQS consists of 6 essential micronutrients: zinc, magnesium, selenium, and vitamins A, C, and E. The higher the score, the higher the antioxidant content. The DAQS has been recognized as a valid indicator of dietary quality.^[29] Similar composite indices have also been successfully applied to predict risks of cardiovascular disease^[32] and colorectal cancer,^[33] supporting the rationale of using such composite scores to assess cognitive health. Our findings suggest that higher levels of DAQS are associated with a lower likelihood of cognitive impairment in older adults, after controlling for variables such as age, gender, race, education, BMI, PIR, smoking status, alcohol consumption, hypertension, and diabetes. The OR was 0.77, with a 95% CI of 0.70 to 0.85. Notably, the ORs remained highly stable across the unadjusted crude model and the fully adjusted models (Models I and II). This consistency indicates that the negative association between DAQS and cognitive impairment is robust and independent, rather than being heavily confounded by variables such as age, gender, race, or other lifestyle factors. Some animal studies have shown the cognitive performance of aged rats and young rats treated with hyperoxia, and the nerve endings of the former were found to be more susceptible to the effects of oxidative stress. Experimental results showed that hyperoxia treatment increased thiobarbituric acid-responsive substances in the hippocampus, cerebral cortex, and synaptic membranes of aged rats and decreased the release of acetylcholine from nerve endings, suggesting that cognitive functions of rats may be affected by oxidative stress and aging.^[34] A distinctive feature of our study, compared to other composite indices like the composite dietary antioxidant index used in recent literature,^[26,27] is the inclusion of magnesium as a key component of the DAQS. Beyond its role as a cofactor in antioxidant enzymes, magnesium plays a critical role in regulating N-methyl-D-aspartate receptors. By blocking N-methyl-D-aspartate receptor channels, magnesium prevents excitotoxicity caused by excessive calcium influx, a process vital for synaptic plasticity and memory preservation. This suggests that the cognitive benefits observed in our study may stem from the synergistic effects of antioxidant defense and neuroprotection against excitotoxicity. Reactive oxygen species are metabolic byproducts that, while playing roles in physiological signaling, can induce severe tissue damage when homeostasis is disrupted.^[35] Cumulative oxidative stress from these sources impedes mitochondrial function and causes damage to different parts of the body, especially the central nervous system.^[36] However, the exact mechanisms by which these dietary antioxidants exert their protective effects on human cognitive function remain unclear and require further verification through more relevant basic research. Cognitive decline is an important neurodegenerative disease, and age is a key factor in its development. Studies have shown that the effectiveness of the body’s natural antioxidant system decreases with age.^[37] Previous studies have shown that dietary intake of antioxidants prevents reactive oxygen species generation^[38]; therefore, exogenous antioxidants in the diet are essential for combating cognitive impairment. The negative correlation between DAQS and cognitive impairment in older adults also explains this phenomenon.

This study emphasizes the broad nature of the DAQS, which covers a wide range of antioxidants, including vitamin A, vitamin E, vitamin C, zinc, selenium, and magnesium, providing a comprehensive view of dietary antioxidant levels. Consistent with our dietary findings, urinary metal concentrations (including zinc and selenium) have also been linked to cognitive performance in the NHANES elderly population,^[39] reinforcing the importance of maintaining adequate mineral status. This research does have several drawbacks, however. The study’s cross-sectional approach could only identify association, not cause. In addition, potential causative factors, such as the presence of a history of coronary heart disease, had to be taken into account when analyzing the database. In addition, dietary intake data were obtained by self-report and may be subject to recall bias.

In conclusion, our research suggests that DAQS and cognitive decline in elderly persons are related. This implies that increasing antioxidant intake through the diet could have a beneficial effect on cognitive performance in older adults. However, more extensive prospective studies and interventional trials are needed in the future to validate these findings, clarify the exact causal relationship, and elucidate the precise mechanisms involved.

5. Conclusions

In older persons, lower DAQS is linked to a higher risk of cognitive impairment. However, to determine the exact causation of this association, more extensive prospective studies are required.

Acknowledgments

The authors would like to thank Zengying from Jiangxi Medical College, Nanchang University, for valuable statistical consultation and feedback on the analysis and interpretation of the results. The authors also thank all members involved in the preparation and revision of this manuscript.

Author contributions

Conceptualization: Dehong Xu, Wan Zhou, Heng Qiu.

Methodology: Heng Qiu.

Project administration: Dehong Xu, Wan Zhou.

Software: Wan Zhou.

Supervision: Heng Qiu.

Writing – original draft: Dehong Xu, Wan Zhou.

Writing – review & editing: Dehong Xu, Wan Zhou.