Association of antioxidants intake in diet and supplements with risk of Alzheimer’s disease: a systematic review and dose-response meta-analysis of prospective cohort studies

Abstract

Background & aims

Previous studies have shown that antioxidants may be associated with risk of Alzheimer’s disease (AD). However, some findings have failed to demonstrate a significant correlation. To rigorously evaluate this relationship, a comprehensive review and meta-analysis were conducted.

Methods

All relevant cohort studies reporting association between antioxidants intake (diet and/or supplement use) and AD risk were searched in 9 electronic databases and 4 registration platforms from their inception up to March 15, 2023. Pooled hazard ratios (HRs) and 95% confidence intervals (CIs) were calculated using either a fixed-effects or random-effects model. Heterogeneity was assessed using I² statistics. Furthermore, a dose-response meta-analysis was conducted to explore potential dose-response relationships.

Results

Eleven cohort studies were included. The pooled HRs of AD were 0.90 (95% CI = 0.60–1.34) and 0.94 (95% CI = 0.75–1.17) for the dietary intake of vitamin E, 0.90 (95% CI = 0.76–1.07) for the vitamin E supplement use. The pooled HRs of AD were 0.84 (95% CI = 0.76–0.93) and 0.60 (95% CI = 0.35–1.02) for the dietary intake of vitamin C, 0.85 (95% CI = 0.72-1.00) for the vitamin C supplement use. The pooled HRs of AD were 1.02 (95% CI = 0.85–1.22) and 0.86 (95% CI = 0.68–1.07) for the dietary intake of beta-carotene. Notably, no significant dose-response relationship was observed.

Conclusions

A high dietary intake of vitamin C (≥ 75 mg/d) was found to have a statistically significant impact on reducing the risk of AD. However, no significant association was observed between dietary intake of vitamin E or beta-carotene, or the use of vitamin E or vitamin C supplement use, and the risk of AD.

Supplementary Information

The online version contains supplementary material available at 10.1007/s40520-024-02893-6.

Introduction

Dementia is currently the 7th leading cause of mortality globally, as stated in the World Alzheimer’s Disease Report 2021, and it affects more than 55 million individuals worldwide [1]. The most prevalent form of dementia is Alzheimer’s disease (AD), which accounts for 60–70% of the total [2]. AD is a group of primary brain degenerative diseases of unknown etiology, which can impair the thinking, memory, and independence of older adults, affect their quality of life, and even lead to death [3]. Moreover, AD imposes a substantial economic burden on both the society and the families of patients. The World Alzheimer’s report predicts that the social-economic cost of dementia worldwide will reach $2.11 trillion in 2030 and $7.45 trillion in 2050 [1]. Studies have confirmed that the potential pathophysiology mechanism of AD may be mainly related to the imbalance between production and clearance of amyloid-beta (Aβ). The neurotoxic plaques formed by abnormal levels of Aβ between neurons in the brain result in neuronal degeneration [4, 5]. However, the systemic therapies available for AD remain inadequate. A previous study revealed that one in three AD cases worldwide can be completely prevented [6]. It is imperative to explore controllable risk factors and develop feasible preventive measures. A combination of genetics, lifestyle and environmental factors seem to influence the onset and progression of AD [7]. Notably, the incidence of AD may be intimately related to some modifiable lifestyle factors, such as poor diet, lack of exercise, and smoking, among others [8].

Antioxidants often exists in natural or synthetic substances that inhibit or retard oxidation reactions, counteracting the damaging effects of oxidants in animal tissues [9]. In terms of antioxidants, vitamin C, vitamin E, and beta-carotene are regarded as the most prevalent types [10]. Robust evidence has demonstrated that increased oxidative stress is a sign of neurodegeneration and is highly linked to the development of AD [11]. Antioxidants have been proven to effectively counteract and mitigate the oxidative stress, protecting the brain and cognitive function [12]. For instance, vitamin E, the most potent antioxidant among nutrients, can dissolve within the lipid structure of the central nervous system, acting synergistically with other intracellular antioxidants to inhibit the lipid peroxidation induced by oxygen free radicals [13]. Additionally, vitamin C plays a crucial role in brain development and function, significantly increasing cellular protein synthesis and enhancing cognitive and memory functions [14]. Beta-carotene can block the chain reaction of free radicals, thereby protecting the body from damages caused by free radicals and lipid peroxidation [15]. Therefore, dietary intake and/ or antioxidant supplementation may alleviate the oxidative stress and forestall the occurrence of AD.

Several meta-analyses have investigated the relationship between antioxidants and AD [16–19]. The inconsistent results of these studies may be attributed to the discrepancy in literature retrieval methodologies, inclusion criteria, statistical analyses, and adjustments for confounding variables. A number of studies have suggested that dietary and supplementary intake of antioxidants may diminish the risk of AD [16, 18, 19]. However, in contrast to other reports, dietary and supplemental antioxidants were not found to be associated with the incidence of AD [16, 17]. Furthermore, the dietary plans of nearly healthy individuals are devised and evaluated using Dietary Reference Intakes (DRIs), which represent quantitative estimates of nutritional intakes [20]. For instance, the current recommended dietary allowance (RDA) for vitamin E for adults aged 51 and over is 15 mg per day. Nevertheless, when formulating recommendations for antioxidants (vitamin E, vitamin C, and beta-carotene), the connection between dietary antioxidant consumption and health-related outcomes, such as AD, has not been taken into consideration.

To provide a foundation and support for the future formulation of antioxidant guidelines, systematic reviews and meta-analyses are necessary to synthesise the available evidence. Given that prospective cohort studies constitute a powerful method for examining the causal relationship between antioxidants and health outcomes, the present study was based on a comprehensive meta-analysis of cohort studies, aiming to draw attention to research gaps pertaining to the establishment of guidelines for the use of antioxidants. Therefore, the following two research questions were addressed in the present study: (1) Is there an association between antioxidants intake from diet and supplements and the risk of AD? (2) Is there a dose-response relationship between antioxidants intake and AD?

Methods

This meta-analysis was carried out rigorously in accordance with the referred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) 2020 guideline [21] and the protocol was registered in the International Prospective Register of Systematic Reviews (CRD42023409306).

Searches strategy

A systematic inspection was conducted of electronic databases (such as PubMed, Cochrane Library, Embase, Web of Science, CINAHL, CNKI, WANFANG, VIP, and CBM) and registration platforms (such as Clinical Trials, WHO-ICTRP, NHS Trusts, and ACTR), the initial search time range from inception to March 15, 2023. The primary terms used in the initial query included: “antioxidant*” OR “vitamin E” OR “vitamin C” OR “beta-carotene” OR “tocopherols” OR “ascorbic acid” AND “dementia” OR “Alzheimer*” AND “prospective” OR “cohort” OR “follow-up” OR “longitudinal”. We also examined reference lists and review articles for relevant studies (Supplementary Table 1).

All studies were meticulously screened and exported to the Endnote X7 reference manager, a tool utilised for managing searched citations. Upon the removal of duplicate studies, the remaining studies were reviewed by two authors, who independently examined the titles, abstracts, and full-texts. The inconsistencies included in the study were discussed and resolved by the two authors.

Eligibility criteria

The eligibility criteria for this study were specified using the PECOS framework. Studies were eligible if they met the following inclusion criteria: (1) Population: individuals aged 60 years old or above; (2) Exposure: high levels of dietary antioxidant intake or supplementation (such as vitamin E, vitamin C, and beta-carotene); (3) Comparison: lower levels of dietary antioxidant intake or absence of supplementation; (4) Outcome: exclusively studies focusing on AD, excluding other forms of dementia, such as vascular dementia (VaD). AD was diagnosed by the National Institute of Neurological and Communicative Disorders and Stroke and the Alzheimer’s Disease and Related Disorders Association (NINCDS-ADRDA); (5) Study design: prospective cohort studies. Animal studies, case reports, reviews, letters, meta-analyses, commentaries, or consensuses were excluded. Furthermore, studies not published in English or Chinese were also omitted.

Data extraction

Two authors independently extracted the following characteristics of included studies, namely: first author, publication year, study location, study design, sample source, the ratio of cases and samples, baseline age, gender composition, duration of follow-up, diagnosis of AD, and study findings. Unless divergences could not solved through discussion, the third author made the final determination.

Risk of bias and certainty of evidence

The Newcastle-Ottawa Quality Assessment Scale (NOS) was employed to evaluate the quality of the overall eligible cohort studies [22]. The NOS scale comprises 3 dimensions, encompassing 8 items: 4 items pertaining to study selection, 1 item concerning group comparability, and 3 items related to outcome ascertainment. Apart from group comparability, the maximum score for other items is 2, yielding a total score range of 0–9. A higher total score indicates a lower risk of bias in the study.

The level of evidence for the outcomes was assessed using the Grading of Recommendations, Assessment, Development, and Evaluations (GRADE) framework, a transparent and structured approach to quality assessment. Starting from an initial high degree of evidence rating, five factors are considered for downgrading the evidence rating among the tools suitable for observational research. These factors include risk of bias, inconsistency, indirectness, imprecision, and publication bias. The quality of evidence is categorised into four levels: no downgrade indicating “high certainty”, one downgrade for “moderate certainty”, two downgrades “low certainty” and three downgrades for “very low certainty” [23]. Two authors independently appraised the risk of bias and quality of evidence. Discussion will carried out if any disagreement arose, and if necessary, the third author was consulted to assess the quality of the evidence.

Data synthesis and statistical analysis

The incorporated study data underwent preprocessing and synthesis, subsequently undergoing meta-analysis using Stata 15.0 statistical software. The study included relevant statistical indexs, namely the hazard ratio (HR), relative risk (RR), odds ration (OR), and 95% confidence interval (95% CI). In scenarios where outcomes are infrequent across all populations and subgroups, or when the study sample size is substantial, distinctions among various measures of relative risk, including HR, RR, and OR, can generally be disregarded [24]. Consequently, the summary HRs and 95% CIs for measuring effectiveness were calculated using either the fixed-effects or random-effects models. Regarding dietary intake of antioxidants, we compared each of the higher levels of antioxidants (above and below DRIs) and the lower levels of antioxidants (reference category) to obtain the HR and 95%CI for different levels of antioxidants. For antioxidant supplements, participants who used antioxidant supplements were compared with non-users to derive meta-analysis results. The Q-test and I² estimates were employed to investigate heterogeneity of various studies, categorised distinctly as strong heterogeneity (I² > 50%), moderate heterogeneity (25% < I² < 50%), and low heterogeneity (I² < 25%). Random-effects models were applied when statistical significance was attained (P < 0.05) [25]. For each meta-analysis encompassing ten or more studies, a funnel plot was used to evaluate publication bias. A visually symmetrical funnel plot was deemed indicative of the absence of publication bias.

In addition, if dietary antioxidant intake is significantly associated with an increased risk of AD, a dose-response analysis is performed on a selection of studies that provide sufficient information. In accordance with the theory proposed by Greenland and Longnecker [26], a few original studies with reference to other groups were converted to the lowest dose group as a reference. The dose-response nonlinear model is formulated through the application of the three-node restricted cubic spline function (at the 10th, 50th, and 90th percentiles of the distribution) in conjunction with the generalized least-squares trend (GLST) model. The nonlinear test is carried out by Wald test. If P < 0.05, it is deemed that a nonlinear dose-response relationship exists. Otherwise, a linear dose-response meta-analysis was conducted, and the fitting effect of model is evaluated [27]. In instances where the median or mean intake is not reported by the study, we use the midpoint between the upper and lower bounds of each category to represent the average intake. Furthermore, if the lower and upper boundaries for the lowest and highest categories, respectively, are unspecified, we infer that the amplitude of the boundary aligns with the nearest category.

Results

Study selection

The study selection process is shown below (Fig. 1). Eleven cohort studies were included in this systematic review. A literature search identified a pool of 4309 articles, initially identified by a comprehensive literature search. After the elimination of 982 duplicate records, a thorough screening of 3327 titles and abstracts was conducted to identify potentially pertinent studies, and a comprehensive full-text review was conducted on 81 studies. Seventy more articles were later discarded because they did not fit the inclusion criteria. Ultimately, this meta-analysis contained 11 studies [28–38].

Fig. 1.

PRISMA flow diagram of search

Study characteristics and quality assessment

Table 1 summarizes the characteristics of the included studies. Eleven cohort studies that were published between 2000 and 2019 were included in this review. The included studies’ sample sizes varied from 579 to 5395, and their follow-up periods varied from 3 to 30.2 years. The included studies had a total sample size of 35,954 participants. There were eight studies with both male and female subjects, two with exclusively male participants, and one did not specify gender. Additionally, eight studies were conducted in the United States, two in the Netherlands, and one in Canada. The results of the quality assessment of selected studies is shown in Supplementary Table 2. All but two studies had a quality score of greater than or equal to 7. Overall, one, four, four, and remaining two three studies had scores of 9, 8, 7, and 6, respectively.

Table 1.

Characteristics of the included studies for vitamin E, vitamin C, and beta-carotene

First Author, Year, country	Study design	Sample source	Cases/Sample	Sex, Female, n (%)	Baseline age, y, Mean ± SD or range	Follow-up time, y	Types of antioxidant	Statistical indexs	Adjustment for confounding factors	Diagnosis criteria of AD	Study findings
Masaki et al. 2000, USA [28]	HAAS/HHP cohort	The island of Oahu, HI	132/3385	0	71 to 93	3 to 5	VE, VC	OR	Age at mailed survey, years of education, years of childhood spent in Japan, apoE4 category, and history of stroke	NINCDS-ADRDA	No significant associations were found for vitamin E and C supplement use with AD.
Engelhart et al. 2002, Netherlands [29]	Rotterdam cohort	A suburb in Rotterdam, the Netherlands	146/5395	3183 (59.0)	67.7 ± 7.8	6	VE, VC, beta-carotene	RR	Age, sex, baseline Mini-mental State Examination score, alcohol intake, education, smoking habits, pack-years of smoking, body mass index, total energy intake, presence of carotid plaques, and use of antioxidantive supplements	NINCDS-ADRDA	High dietary intake of vitamin C and vitamin E may lower the risk of AD.
Morris et al. 2002, USA [30]	CHAP cohort	A stratified random sample of community-dwelling residents.	131/815	313 (62.0)	73.3 ± 9.7	3.9 ± 1.7	VE, VC, beta-carotene	RR	Age (years), sex, education (years), APOE ε4 status (any allele vs. none), race (black or white), an interaction term between race and APOE ε4, and period of observation	NINCDS-ADRDA	Increasing vitamin E intake from foods was associated with decreased risk of developing AD. Intake of vitamin C, β caritene, and vitamin E from supplements was not significantly associated with risk of AD.
Luchsinger et al. 2003, USA [31]	WHICAP cohort	Healthy Medicare beneficiaries from WHICAP	242/4023	2695 (67.0)	75.3 ± 5.8	4.0 ± 1.5	VE, VC	RR	Age, sex, APOE ε4 allele presence, smoking status, and years of education	NINCDS-ADRDA	Neither dietary and supplemental of carotenes, vitamin C, vitamin E was associated with a decreased risk of AD.
Laurin et al. 2004, USA [32]	HAAS/HHP cohort	The island of Oahu, HI	235/2459	0	76.5	30.2(25.7–33.0)	VE, VC, beta-carotene	HR	Age, education, smoking status, alcohol intake, body mass index, physical activity, systolic and diastolic blood pressures, year of birth, total energy intake, cholesterol concentration, history of cardiovascular disease, supplemental vitamin intake, and apolipoprotein E ε4	NINCDS-ADRDA	Intakes of β carotene, vitamins E and C were not associated with the risk of dementia or its subtypes.
Zandi et al. 2004, USA [33]	Cache County Study	A large, population-based investigation	104/4740	2712 (57.2)	≥ 65	3	VE, VC	HR	Age, the squared deviation of age from the population median, sex, education, dummy-coded terms for the presence of 1 and 2 apolipoprotein E ε4 alleles, interactions between age and the dummy-coded apolipoprotein E ε4 terms, and an indicator term for general health status.	NINCDS-ADRDA	Use of vitamin E and vitamin C supplements were not associated with reduced prevalence and incidence of AD.
Corrada et al. 2005, USA [34]	BLSA cohort	The cohort comprises well-educated, predominantly white, community-dwelling volunteers	57/579	220 (38.0)	69.6	9.3 (0.4–14.6)	VE, VC	RR	Age, gender, education, and caloric intake	NINCDS-ADRDA	No association was found between total intake of vitamin C, carotenoids and risk of AD.
Gray et al. 2008, USA [35]	ACT cohort	Study participants were randomly sampled from Seattle-area members of Group Health Cooperative (GHC)	405/2969	1768 (59.5)	≥ 65	5.5 ± 2.7	VE, VC	HR	Age, sex, education, exercise, smoking status, self-reported health, and coronary heart disease	NINCDS-ADRDA	The use of supplemental vitamin E and C did not reduce risk of AD or overall dementia.
Devore et al. 2010, Netherlands [36]	Rotterdam cohort	Omoord residents aged 55 and above	465/5395	/	67.4–68.1	9.6	VE, VC, beta-carotene	HR	Age, education, APOE ε4 genotype, total energy intake, alcohol intake, smoking habits, BMI, and supplement use	NINCDS-ADRDA	High intake of foods rich in vitamin E may modestly reduce long-term risk of dementia and AD.
Basambombo et al. 2016, Canada [37]	CSHA cohort	Community-dwelling and institutionalized participants	821/5269	3223 (61.2)	≥ 65	5.2 ± 1.7	VE, VC	HR	Ever regular smoking, alcohol drinking, regular physical activity, NSAID use, history of diabetes, and vascular risk factors	NINCDS-ADRDA	The use of vitamin E and C supplements was associated with a reduced risk of AD and all-cause dementia.
Agarwal et al. 2019, USA [38]	the cohort of Rush Memory and Aging Project	2152 residents of retirement communities and senior public housing	245/925	694 (75.0)	81 ± 7.2 (58–98)	6.7 ± 3.6	VC	HR	Age, sex, education, physical activity, participation in cognitive activities, Apo-ε status, dietary intake of other fruits, and total calorie intake	NINCDS-ADRDA	Consumption of foods rich in vitamin C may reduce the risk of AD.

Abbreviations: AD, Alzheimer’s disease; HAAS/HHP: Honolulu–Asia Aging Study/Honolulu Heart Program; VE: Vitamin E; VC: Vitamin C; OR: Odds Ratio;

NINCDS-ADRDA: National Institute of Neurological and Communicative Disorders and Stroke and Alzheimer Disease and Related Disorders Association; RR: Relative

Risk; CHAP: Chicago Health and Aging Project; WHICAP: Washington Heights-Inwood Columbia Aging Project; HR: hazard ratio; BLSA: Baltimore Longitudinal Study

of Aging; ACT: Adult Changes in Thought; CSHA: Canadian Study of Health and Aging; NSAID: Nonsteroidal Anti-inflammatory Drug

The meta-analysis of association between antioxidants status and risk of AD

Vitamin E status and risk of AD

Five cohort studies investigated the association between dietary intake of vitamin E and risk of AD [29–32, 36]. A statistically significant correlation between vitamin E intake and the risk of AD was not observed for intakes of ≥ 15 mg/d (HR = 0.90, 95% CI = 0.661 − 0.34). Vitamin E intake below 15 mg/d did not show a statistically significant association with the risk of AD (HR = 0.94, 95% CI = 0.75–1.17) (Fig. 2A). We used a random-effects model to combine the effect sizes since there was heterogeneity (I² = 66.0%, P = 0.032; I² = 39.2%, P = 0.106). Very low certainty was found in the evidence for each of the two effect estimates (Supplementary Table 3).

Fig. 2.

Forest plots of dietary intake (A) or supplement use (B) of vitamin E and AD risk

The relationship between the use of vitamin E supplements and the risk of AD was examined in six cohort studies [28, 30, 31, 33, 35, 37]. Compared with those not taking vitamin E supplements, the HR of AD was 0.90 (95% CI = 0.76–1.07) for those taking vitamin E supplements, and the heterogeneity among studies was low (I² = 8.1%, P = 0.365) (Fig. 2B). As shown in Supplementary Table 3, the evidence supporting the effect estimation was a very low certainty.

Vitamin C status and risk of AD

The final analysis comprised six studies that examined the relationship between dietary intake of vitamin C intake and the risk of AD [29–32, 36, 38]. We discovered a statistically significant correlation between vitamin C intake ≥ 75 mg/d and the risk of AD (HR = 0.84, 95% CI = 0.76–0.93). Additionally, the I² statistics exhibited no evidence of heterogeneity among the studies (I² = 0.0%, P = 0.472) (Fig. 3A). Furthermore, Supplementary Fig. 1 shows that a funnel plot demonstrates that there was no publishing bias for the 15 included studies, and that the scattering distribution was almost symmetrical. The association between dietary intake and AD risk was not statistically significant for vitamin C intakes less than 75 mg/d (HR = 0.60, 95% CI = 0.35–1.02), and the I² statistics showed no evidence of study heterogeneity (I² = 0.0%, P = 0.810) (Fig. 3A). Supplementary Table 3 shows that there was low certainty evidence linking vitamin C intake (≥ 75 mg/d and < 75 mg/d) to the risk of AD.

Fig. 3.

Forest plots of dietary intake (A) or supplement use (B) of vitamin C and AD risk

The correlation between vitamin C supplement use and risk of AD was revealed in six studies [28, 30, 31, 33, 35, 37]. With very low certainty of evidence (HR = 0.85, 95% CI = 0.72-1.00), there was no statistically significant connection between the use of vitamin C supplements and an elevated risk of AD. The I² statistics, however, indicated a considerable level of heterogeneity among the studies (I² = 38.8%, P = 0.147) (Fig. 3B).

Beta-carotene status and risk of AD

The relationship between dietary intake of beta-carotene and the risk of AD was examined in four studies [29, 30, 32, 36]. The dietary intake of beta-carotene greater than 2.6 mg/d was not statistically significantly associated with an increased risk of AD (HR = 1.02, 95% CI = 0.85–1.22; low certainty). Figure 4 shows that there was no heterogeneity among the studies using the I² statistics (I² = 0.0%, P = 0.384). Between low dietary beta-carotene intake (< 2.6 mg/d) and the risk of AD, no statistically significant relationship was found (HR = 0.86, 95% CI = 0.68–1.07; low certainty). According to the I² statistics, there was no heterogeneity amongst the studies (I² = 0.0%, P = 0.482) (Fig. 4).

Fig. 4.

Forest plot of dietary intake of beta-carotene and AD risk

Dose-response analyses

We performed a dose-response relationship analysis since the current result indicated that a higher dietary intake of vitamin C (≥ 75 mg/d) decreased the risk of AD. A total of 4 studies [29, 30, 32, 36] were included for meta-analysis, and the results of chi-square goodness-of-fit test showed that there was no need to consider the heterogeneity among different studies (P = 0.303), fixed effect model was used to merge. Using the testparm command to test whether there is a curve relationship, the results showed that the difference is not statistically significant (χ²=3.28, P = 0.194), indicating that the nonlinear model is not suitable. Furthermore, the fixed effect linear model was fitted by glst command, and the results showed that the difference was not statistically significant (χ²=3.04, P = 0.081), and the linear model could not be constructed. Therefore, there was no significant linear or non-linear association between dietary intake of vitamin C and the risk of AD (Fig. 5).

Fig. 5.

Dose-response analyses between dietary intake of vitamin C and AD risk

Discussion

In this meta-analysis, a comprehensive examination was conducted of the relationship between dietary or supplemental intake of antioxidants and the risk of AD. A total of eleven cohort studies were included, which investigated the dietary consumption and/or supplementary use of antioxidants, such as vitamin E, vitamin C, and beta-carotene, in relation to the AD risk. Our findings reveal that dietary intake of vitamin C at a level of ≥ 75 mg/d demonstrates a statistically significant correlation with the risk of AD. Conversely, the pooled results indicate no statistically significant association between the dietary intake of vitamin E or beta-carotene, or the supplementary use of vitamin E or vitamin C, and the risk of AD.

As an pivotal nutrient in the diet, the relationship between vitamin E and the risk of AD has been extensively investigated, yielding varied results [39, 40]. In our meta-analysis, neither dietary intake nor supplemental use of vitamin E exhibited a statistically significant correlation with the risk of AD. This is consistent with the previous studies [18, 39]. For example, Wang et al. [18] evaluated the link between vitamin E supplementation and AD, concluding that there was insufficient evidence to suggest an association between vitamin E supplementation and the risk of AD based on the synthesised data. However, Li et al. [41] observed that dietary intake of vitamin E demonstrated a notable protective effect in reducing the risk of AD. The potential mechanism underlying this effect could be its lipid-soluble nature, which facilitates its antioxidant actions against lipid peroxidation in the pathogenesis of AD. Conversely, some researchers have postulated that this protective effect may be disappointing and negligible, particularly in the context of supplement use [16].

Our results confirm that there is a statistically significant correlation between the risk of AD and a high dietary consumption of vitamin C (≥ 75 mg/d). However, this correlation diminishes as vitamin C intake falls below < 75 mg/d. Moreover, no statistically significant association was observed between the supplement use of vitamin C and risk of AD. These results are in concordance with a prior study [19]. Considering the underlying pathomechanisms, many potential approaches of functioning are possible. For instance, one possible explanation for the apparent outcome could be that vitamin C protect SH-SY5Y human neuroblastoma cells from Aβ-mediated apoptosis. This could slow down the rate of endogenous amyloid synthesis and lessen the risk of AD [42]. Zhou et al.’s study [19] examined the relationship between AD risk and vitamin C intake patterns, as well as the potential benefits of vitamin C supplementation and dietary intake, by employing linear dose-response analysis. Further research is required to elucidate the precise mechanism underlying this association and to determine the optimal dose for reducing AD risk.

The relationship between beta-carotene and the risk of AD has been a subject of considerable debate for an extended period. Regardless of dose, our investigation shows no statistically significant correlation between beta-carotene intake and AD risk. Conversely, a meta-analysis has lent support to the notion that beta-carotene exerts a beneficial influence on AD risk and may potentially contribute to the prevention of cognitive decline [17]. This is attributed to its antioxidant properties, which may modulate telomerase activity in older individuals, and its anti-inflammatory functions, which interact with inflammatory signalling cascades [43]. However, a preponderance of recent studies have focused primarily on serum carotenoid levels, rather than on dietary intake patterns [44, 45]. Consequently, there is a pressing need for more robust and reliable studies to elucidate the underlying mechanisms and to ascertain the precise effect of beta-carotene on AD risk.

The strengths of this study are as follows. Firstly, cohort studies represent the observational research paradigm offering the highest level of evidence in evidence-based medicine. Compared with randomised controlled trials (RCTs), cohort studies exhibit superior feasibility in terms of implementation ease and mitigation of ethical risks. Additionally, we conducted a dose-response analysis, meticulously considering the statistical significance of the association between AD risk and vitamin C intake from dietary sources. Besides, as the GRADE framework estimates the degree of evidence certainty, our findings are rendered more trustworthy and impartial. However, it is noteworthy that GRADE inherently assigns a low initial rating to observational studies, resulting in an overall certainty level of the evidence body for all research outcomes being low to very low. Lastly, our groups were categorised according to the RDA standard, providing tentative guidelines for dietary nutrient intake.

Limitations inherent in this study merit consideration. Firstly, the elimination of confounding effects from other variables is unattainable, as potential influencing factors such as genetic disparities, economic statuses, and educational backgrounds continue to impact the observed associations. Given that the sources of evidence comprise solely observational research data, the conclusions drawn may be overly definitive. Furthermore, the total patient count included in the studies was relatively modest, which could potentially bias the analytical outcomes. Moreover, despite our comprehensive search across major databases, limitations in the studies due to language and time constraints may still persist, potentially leading to deviations in the results to some degree.

Conclusions

This meta-analysis concluded that dietary intake at least 75 mg of vitamin C per day considerably reduces the risk of AD. Nonetheless, the evidence available remains insufficient to conclusively establish a relationship between dietary intake of vitamin E, beta-carotene, or the use of vitamin E and vitamin C supplements, and the risk of AD. We strongly recommend the conduct of additional high-quality RCTs and research with larger sample sizes to further investigate the correlation between antioxidants and the risk of AD.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Author contributions

Study conception and design: JZ, XH and ZW; Analyses and interpretation of data: JZ, XH and YS; Drafting of the manuscript: JZ and XH; Revised the manuscript: ZW; Project administration and supervision: ZW; Approved the manuscript: all authors.

Funding

This study was supported by the National Natural Science Foundation of China (grant number: 72274007); the Incubation Fund for Scientific Research and Innovation of Leading Talents of Unnamed Nursing of Peking University (grant number: LJRC22YB03).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Competing interests

The authors declare no competing interests.