ABSTRACT
Background
This study examined the association between dietary intakes of total and specific carotenoids and cognitive decline among Chinese middle‐aged and older adults.
Methods
We included 4043 participants aged 55 years and older from the China Health and Nutrition Survey (1997–2006). Average daily dietary intakes of total carotenoids, α‐carotene, β‐carotene, β‐cryptoxanthin, lycopene, and lutein + zeaxanthin were calculated from 3‐day 24‐h dietary recalls. Cognitive function was measured repeatedly using the Telephone Interview for Cognitive Status‐Modified. Linear mixed‐effects models were utilized to estimate beta coefficients (β) and the 95% confidence intervals (CIs) for the association of energy‐adjusted carotenoid intake with changes in cognitive z‐scores.
Results
Higher intake of total dietary carotenoids was associated with slower cognitive decline. Comparing the top to bottom quintiles (median intake: 37.0 vs. 3.1 mg/day) of total carotenoids, the adjusted difference in annual decline rates (95% CI) was 0.018 (0.001–0.035, p‐trend = 0.037). The strongest association was observed for moderate intake at the fourth quintile (median intake: 23.5 mg/day, β = 0.028, 95% CI: 0.011–0.046, p‐quadratic = 0.030). Protective associations were also found for specific carotenoids, including α‐carotene (β Q5 vs. Q1 = 0.037, 95% CI: 0.020–0.055), β‐cryptoxanthin (β = 0.032, 95% CI: 0.014–0.049), lutein + zeaxanthin (β = 0.020, 95% CI: 0.002–0.037), lycopene (β = 0.018, 95% CI: 0.001–0.035) and β‐carotene (β = 0.019, 95% CI: 0.001–0.036).
Conclusions
A potential nonlinear association between dietary carotenoids and cognitive decline was observed in Chinese middle‐aged and older adults. Further research is warranted to confirm optimal intake level for dietary carotenoids and to investigate the underlying biological mechanisms.
Keywords: Chinese middle‐aged and older adult, cognitive decline, dietary carotenoid, prospective study
Higher intake of dietary carotenoids was associated with slower cognitive decline. The strongest protective association was observed for moderate carotenoid intake. Total carotenoid intake exhibited a non‐linear association with cognitive decline.
Abbreviations
- BMI
body mass index
- CFCT
Chinese Food Composition Table
- CHNS
China Health and Nutrition Survey
- CI
confidence interval
- IQR
interquartile range
- JFCT
Japanese Food Composition Table
- SD
standard deviation
- USDA
United States Department of Agriculture
1. Background
As the global population ages, the prevalence of age‐related cognitive decline and the dementia burden are increasing dramatically [1]. Currently, over 50 million individuals worldwide are living with dementia, and the number is projected to triple by 2050 [2]. Given the difficulty of reversing the progression of dementia, identifying modifiable lifestyle factors that can slow cognitive decline and prevent dementia would have substantial public health benefits [3].
Carotenoids are a large class [4] of dietary antioxidants abundant in deeply pigmented vegetables and fruits, including those with yellow, red, and dark green color. Among over 700 identified carotenoids, six carotenoids, lutein, zeaxanthin, β‐carotene, β‐cryptoxanthin, α‐carotene, and lycopene, are the most commonly detected in both the human diet and serum [5]. Epidemiological evidence has increasingly suggested the neuroprotective potential of dietary antioxidants [6]; the role of carotenoids in preventing cognitive decline and dementia has thus received growing research interest.
However, evidence from observational studies and randomized controlled trials on carotenoids and cognitive function has been inconsistent. A recent meta‐analysis reported lower brain carotenoid concentrations in individuals with Alzheimer's disease [7], whereas another meta‐analysis of nine intervention studies found that lutein + zeaxanthin and β‐carotene interventions were associated with better cognitive performance [5]. Besides, potential protective associations were demonstrated for total dietary carotenoids and cognitive function in several large‐scale cohort studies [3, 8, 9, 10, 11], while no significant associations were found in another three cohort studies [11, 12, 13]. In addition, concerns have also been raised about potential detrimental effects from high‐dose specific carotenoids [13, 14, 15]. For example, daily supplementation with high‐dose β‐carotene (20–30 mg/day) increased lung cancer risk and mortality among smokers in two previous studies [14, 15]. The postulated mechanism for this adverse effect is that carotenoids may exhibit prooxidant activity under conditions of high concentration, elevated oxygen tension, and imbalanced intracellular redox status, and toxicity is hypothesized to occur when supplemental β‐carotene exceeds the typical dietary intake levels, which were found to be protective and beneficial in observational studies [15, 16]. It is, thus, of great importance to identify the optimal intake level of carotenoids to inform safe and effective dietary intervention studies.
The current study examined the associations of total and specific dietary carotenoids with cognitive decline among a nationally representative sample of middle‐aged and older Chinese adults. Given that existing evidence is mainly limited to Western populations, whose dietary habits differ from those of Asian populations, we aimed to extend knowledge on the role of carotenoids in slowing cognitive decline, specifically within the Chinese population, which has a long‐established tradition of consuming plant‐based, carotenoid‐rich diet.
2. Methods
2.1. Study Population
The China Health and Nutrition Survey (CHNS), established in 1989, is an ongoing prospective cohort study that recruits new participants every 2–4 years in each wave. Using a multistage random cluster sampling design, households were sampled from nine provinces in mainland Chinese. Between 1997 and 2006, cognitive function was assessed in person among participants aged 55 years and older during face‐to‐face household interviews. This study was conducted in accordance with the Declaration of Helsinki and its subsequent revised editions. The original study protocol was approved by the Institutional Review Boards of the University of North Carolina at Chapel Hill and the National Institute of Nutrition and Food Safety at the Chinese Center for Disease Control and Prevention. All participants provided written informed consent. In addition to the approvals obtained by the original cohorts, the secondary analysis has been reviewed and approved by the Ethics Committee of Zhejiang University School of Public Health (Approval Code: ZGL202403‐2[TS1.1], Approval Date: March 21, 2024).
Among 5038 eligible participants aged 55 years or older, 4594 (91.2%) completed at least one cognitive assessment. We further excluded 26 individuals without dietary information, 34 with extreme energy intake (< 1st or > 99th percentile) [17], and 491 with severe cognitive impairment (global cognitive score < 7) at baseline [18]. Finally, a total of 4043 (80.3%) participants were included, in which 1339, 804, 1201, and 699 participants were enrolled in 1997, 2000, 2004, and 2006, respectively (Figure S1).
2.2. Dietary Assessment
During each wave, individual‐level dietary intake was assessed through 3‐day 24‐h dietary recalls, while household‐level consumption of condiments and cooking oil was measured through a 3‐day weighing method [19]. The dietary assessment method has been validated for energy intake [20]. Intakes of total and specific carotenoids (lutein, zeaxanthin, β‐carotene, β‐cryptoxanthin, α‐carotene, and lycopene) were calculated based on the United States Department of Agriculture (USDA) Nutrient Database, which provides values for total carotenoids and each of the five individual carotenoid components in foods. Major food sources include green leafy vegetables (such as spinach, collard, and kale), red and yellow vegetables (such as tomato, carrot, and pepper), other vegetables (such as broccoli, cabbage, and onion), berry fruits, citrus fruits, other fruits (such as red and yellow fruits, e.g., watermelon), and others. During follow‑up, the top contributors in our study were green leafy vegetables, red and yellow vegetables, other vegetables, berry fruits, citrus fruits, and other fruits. Green leafy vegetables provided most β‐carotene and lutein + zeaxanthin, whereas red and yellow vegetables were the main sources of α‐carotene, β‐cryptoxanthin, and lycopene (Figure S2). To account for cross‑population dietary differences, we additionally estimated carotenoid intake using the Chinese Food Composition Tables (CFCT; 1991, 2002, and 2004 editions) and the Japanese Food Composition Tables (JFCT; 2015 edition). The CFCT provides total carotene information, and the JFCT gives detailed values for β‐carotene, α‐carotene, and β‐cryptoxanthin. All intakes were energy‑adjusted via the residual method [21].
2.3. Cognitive Assessment
In the CHNS study, cognitive tests were derived from the Telephone Interview for Cognitive Status‐Modified [22], which evaluated multiple cognitive domains, including memory (immediate and delayed recall of a 10‐word list, 10 points for each task), calculation (serial 7 subtraction from 100, 5 points), and attention (backward counting from 20 to 1, 2 points) [23]. The global cognitive score was calculated as the sum of all domain scores, ranging from 0 to 27. To enhance comparability with other studies, a composite score was constructed by averaging the z‐scores of all cognitive domains [24].
2.4. Assessment of Other Covariates
Sociodemographic characteristics, including age, sex, household income, education, region, and residence, were considered in this study. Lifestyle factors encompassed smoking status, drinking status, physical activity, body mass index (BMI), and total energy intake. Physical activity was estimated from self‐reported hours spent in occupational, transportation, household, and leisure‐time activities, and quantified as metabolic equivalent‐hours per week. Additionally, self‐reported diagnosis of diabetes, hypertension, and myocardial infarction were included as history of chronic diseases. Missing values were handled using single imputation, with the mode for categorical variables and the mean for continuous variables.
2.5. Statistical Analysis
Baseline characteristics of participants were summarized by quintiles of total carotenoid intake, with categorical variables presented as numbers (percentages) and continuous variables as means (standard deviations [SDs]). Group differences were assessed using the Chi‐square test for categorical variables and one‐way analysis of variance (ANOVA) for continuous variables. Linear mixed‑effects models were used to evaluate the association of total carotenoid intake with changes in annual global cognitive z‐score, incorporating the repeated diet and cognitive measures collected across waves from 1997 to 2006. A positive beta coefficient (β) indicated slower decline with higher intake. To further clarify, the key fixed effect was the interaction between carotenoid intake and time, which shows how the rate varies by intake level. Random intercepts for individuals accounted for baseline differences and the within‐person correlation from repeated measures. We included 699 participants newly enrolled in 2006 as they could contribute to covariate estimation with a time indicator set to 0 in the linear mixed model. Three adjustment models were fitted. Model 1 was adjusted for age, age squared, sex, household income (tertiles, as low, median and high), education (illiteracy, primary school, middle school and above), region (northern, southern), residence (rural, urban), smoking status (never, ever), drinking status (never, ever), BMI (normal weight [BMI < 24 kg/m2], overweight [24 ≤ BMI < 28 kg/m2], obesity [BMI ≥ 28 kg/m2]), and physical activity (tertiles, as low, median and high). Model 2 was additionally adjusted for total energy intake (continuous), intake of whole grains, nuts, fish and shellfish, red meat, poultry, pastries and sweets, tea and coffee (continuous). Model 3 was additionally adjusted for hypertension, diabetes, and myocardial infarction (all categorized as yes or no). Potential linear and quadratic trends were assessed by including the exposure as a continuous variable and its quadratic term. Stratified analyses were conducted across subgroups among age (≥ 65, < 65 years old), sex, income (low and medium, high), education (illiteracy or primary school, middle school and above), smoking status (never, ever), drinking status (never, ever), physical activity (low and medium, high), and BMI (normal, overweight and obesity). To examine potential effect modification, we tested for the interaction term between total carotenoid intake and the above covariates in the fully‐adjusted models.
In the secondary analysis, we evaluated the associations between cognitive decline and the intake of: specific carotenoids (α‐carotene, β‐carotene, β‐cryptoxanthin, lycopene, and lutein + zeaxanthin) calculated by the USDA database; carotene from the CFCT; and α‐carotene, β‐carotene, and β‐cryptoxanthin from the JFCT. Several sensitivity analyses were performed to test the robustness of the associations. First, we restricted to participants who completed all follow‐up surveys. Second, we repeated the primary analysis among participants with at least two cognitive assessments. Third, to address the potential clustering effects, we randomly selected one participant in each household. All statistical analyses were conducted using R 4.1.2 (R Foundation for Statistical Computing, Vienna, Austria), and two‐sided p < 0.05 was considered statistically significant.
3. Results
Baseline characteristics of study participants by quintiles of total carotenoid intake were presented in Table 1. Among 4043 participants (mean age 62.2 years, 49.5% female), median daily intake and interquartile range (IQR) of carotenoids was 14.9 (6.3–25.4) mg overall. The corresponding median intakes across increasing quintiles were 3.1 (1.7–4.2) mg, 8.3 (6.9–10.0) mg, 15.4 (13.5–17.5) mg, 23.5 (21.5–26.3) mg, and 37.0 (32.2–44.5) mg, respectively. The primary dietary carotenoids were lutein + zeaxanthin and β‐carotene, with the median intake being 8.1 and 5.4 mg/day overall. Intakes of α‐carotene, β‐cryptoxanthin, and lycopene were positively correlated with each other (Spearman correlation coefficients [r] were 0.86 between α‐carotene and β‐cryptoxanthin, 0.77 between α‐carotene and lycopene, and 0.76 for β‐cryptoxanthin and lycopene, respectively) (Table S1). The highest correlation was observed between the intakes of β‐carotene and lutein + zeaxanthin (r = 0.93). Participants with higher carotenoid intake were more likely to be educated, more physically active, and more likely to reside in urban areas and southern regions.
Table 1.
Population characteristics by quintile of total energy‐adjusted carotenoid intake in the China Health and Nutrition Survey.
| Variable | Overall | Quintile 1 | Quintile 3 | Quintile 5 | p |
|---|---|---|---|---|---|
| N | 4043 | 886 | 817 | 745 | |
| Age (years), mean (SD) | 62.2 (7.0) | 62.6 (7.1) | 62.1 (6.7) | 61.8 (7.0) | 0.065 |
| Female, n (%) | 2002 (49.5) | 424 (47.9) | 407 (49.8) | 393 (52.8) | 0.091 |
| Resident in urban China, n (%) | 1662 (41.1) | 266 (30.0) | 359 (43.9) | 309 (41.5) | < 0.001 |
| Resident in Southern China, n (%) | 2359 (58.3) | 419 (47.3) | 484 (59.2) | 576 (77.3) | < 0.001 |
| Education, n (%) | < 0.001 | ||||
| Illiteracy | 1773 (43.9) | 444 (50.1) | 355 (43.5) | 325 (43.6) | |
| Primary school | 1040 (25.7) | 225 (25.4) | 208 (25.5) | 190 (25.5) | |
| Middle school and above | 1230 (30.4) | 217 (24.5) | 254 (31.1) | 230 (30.9) | |
| Income, n (%) | < 0.001 | ||||
| Low | 1445 (35.7) | 409 (46.2) | 278 (34.0) | 226 (30.3) | |
| Median | 1400 (34.6) | 293 (33.1) | 272 (33.3) | 266 (35.7) | |
| High | 1198 (29.6) | 184 (20.8) | 267 (32.7) | 253 (34.0) | |
| Smoking status, n (%) | 0.270 | ||||
| Never | 2695 (66.7) | 586 (66.1) | 533 (65.2) | 512 (68.7) | |
| Ever | 1348 (33.3) | 300 (33.9) | 284 (34.8) | 233 (31.3) | |
| Drinking status, n (%) | 0.087 | ||||
| Never | 2717 (67.2) | 598 (67.5) | 537 (65.7) | 522 (70.1) | |
| Ever | 1326 (32.8) | 288 (32.5) | 280 (34.3) | 223 (29.9) | |
| Physical activity, n (%) | < 0.001 | ||||
| Low | 1570 (38.8) | 411 (46.4) | 304 (37.2) | 245 (32.9) | |
| Median | 1195 (29.6) | 249 (28.1) | 243 (29.7) | 221 (29.7) | |
| High | 1278 (31.6) | 226 (25.5) | 270 (33.0) | 279 (37.4) | |
| BMI, n (%) | 0.371 | ||||
| Normal | 2561 (63.3) | 553 (62.4) | 501 (61.3) | 497 (66.7) | |
| Overweight | 369 (9.1) | 81 (9.1) | 90 (11.0) | 62 (8.3) | |
| Obesity | 1113 (27.5) | 252 (28.4) | 226 (27.7) | 186 (25.0) | |
| Total calorie intake (kcal/day), mean (SD) | 2103.7 (589.0) | 2120.8 (610.4) | 2124.1 (581.9) | 2074.4 (593.4) | 0.352 |
| Total carotenoids (mg/day), median (IQR) | 14.9 (6.3–25.4) | 3.1 (1.7–4.2) | 15.4 (13.5–17.5) | 37.0 (32.2–44.5) | < 0.001 |
| α‐Carotene (mg/day), median (IQR) | 0.16 (0.10–0.49) | 0.09 (0.06–0.13) | 0.19 (0.11–0.55) | 0.25 (0.16–0.74) | < 0.001 |
| β‐Carotene (mg/day), median (IQR) | 5.4 (2.6–8.8) | 1.2 (0.5–1.7) | 5.6 (4.9–6.3) | 12.9 (11.2–15.1) | < 0.001 |
| β‐Cryptoxanthin (mg/day), median (IQR) | 0.08 (0.04–0.20) | 0.03 (0.02–0.07) | 0.10 (0.04–0.24) | 0.13 (0.06–0.32) | < 0.001 |
| Lycopene (mg/day), median (IQR) | 0.08 (0.03–0.90) | 0.05 (0.02–0.08) | 0.12 (0.04–1.11) | 0.10 (0.03–1.46) | < 0.001 |
| Lutein + zeaxanthin (mg/day), median (IQR) | 8.1 (2.7–15.4) | 1.3 (0.6–2.1) | 8.9 (6.9–10.5) | 22.6 (19.9–27.5) | < 0.001 |
| Whole grains (g/day), mean (SD) | 29.8 (76.8) | 36.5 (67.3) | 24.7 (62.1) | 22.4 (71.4) | < 0.001 |
| Nuts (g/day), mean (SD) | 6.9 (27.6) | 5.8 (25.5) | 8.3 (28.2) | 8.9 (36.6) | 0.044 |
| Fish and shellfish (g/day), mean (SD) | 68.0 (122.5) | 46.0 (90.6) | 67.8 (110.9) | 89.7 (171.3) | < 0.001 |
| Red meat (g/day), mean (SD) | 96.9 (105.2) | 74.5 (90.0) | 104.6 (110.2) | 109.0 (104.2) | < 0.001 |
| Pastries and sweets (g/day), mean (SD) | 36.6 (102.0) | 24.4 (59.8) | 42.3 (119.0) | 40.4 (154.3) | < 0.001 |
| Tea and coffee (g/day), mean (SD) | 0.4 (15.5) | 0.0 (0.8) | 0.0 (0.0) | 1.8 (35.6) | 0.103 |
| Hypertension, yes, n (%) | 642 (15.9) | 121 (13.7) | 149 (18.2) | 110 (14.8) | 0.010 |
| Diabetes, yes, n (%) | 133 (3.3) | 29 (3.3) | 34 (4.2) | 27 (3.6) | 0.407 |
| Myocardial infarction, yes, n (%) | 42 (1.0) | 12 (1.4) | 9 (1.1) | 5 (0.7) | 0.088 |
Note: Data are presented as number (percentage), mean (SD) or median (IQR).
Abbreviations: BMI, body mass index; IQR, interquartile range; SD, standard deviation.
Participants were followed for a median of 3 years. In the primary analyses, higher total carotenoid intake was related to slower global cognitive decline after adjustment for sociodemographic characteristics, lifestyle factors, other dietary factors, total energy intake, and chronic diseases (Table 2, Figure 1, Table S2). Compared to participants in the bottom quintile, the estimated annual decline rates (95% confidence interval [CI]) were 0.002 (−0.016 to 0.020), 0.013 (−0.005 to 0.031), 0.028 (0.011–0.046), and 0.018 (0.001–0.035) for the increasing quintiles, respectively. The strongest association was observed in the fourth quintile group (p‐quadratic = 0.030). Additionally, the associations were similar for composite cognitive decline (β Q5 vs. Q1 = 0.016, 95% CI: 0.003–0.030, p‐trend < 0.001, Table S3).
Table 2.
Associations of total energy‐adjusted carotenoid intake with global cognitive decline.
| Total carotenoids intake (energy‐adjusted) | Quintile 1 | Quintile 2 | Quintile 3 | Quintile 4 | Quintile 5 | p‐trend |
|---|---|---|---|---|---|---|
| N | 886 | 793 | 817 | 802 | 745 | |
| Intake (mg/day), median (IQR) | 3.1 (1.7–4.2) | 8.3 (6.9–10.0) | 15.4 (13.5–17.5) | 23.5 (21.5–26.3) | 37.0 (32.2–44.5) | |
| Model 1 | Ref | 0.002 (−0.016 to 0.020) | 0.015 (−0.003 to 0.032) | 0.030 (0.013–0.047) | 0.020 (0.002–0.037) | 0.022 |
| Model 2 | Ref | 0.002 (−0.016 to 0.020) | 0.014 (−0.004 to 0.032) | 0.029 (0.011–0.046) | 0.018 (0.001–0.036) | 0.036 |
| Model 3 | Ref | 0.002 (−0.016 to 0.020) | 0.013 (−0.005 to 0.031) | 0.028 (0.011–0.046) | 0.018 (0.001–0.035) | 0.037 |
Note: Model 1 adjusted for age, age squared, sex, education (illiteracy/primary school/middle school and above), residence (urban/rural), region (northern/southern), income (low/medium/high), smoking status (never/ever), drinking status (never/ever), BMI (normal weight/overweight/obesity), and physical activities (low/medium/high). Model 2 additionally adjusted for total intake of energy, whole grains, nuts, fish and shellfish, red meat, poultry, pastries and sweets, tea and coffee (continuous). Model 3 additionally adjusted for chronic diseases, including hypertension, diabetes, and myocardial infarction (yes/no). Data are presented as numbers or median (IQR).
Abbreviations: BMI, body mass index; IQR, interquartile range; Ref, reference.
Figure 1.
Rates of change in global cognitive score over four waves for CHNS participants across quintiles of total energy‐adjusted carotenoid intake. The rates of change were based on the linear mixed‐effects model and adjusted for age, age squared, sex, education (illiteracy/primary school/middle school and above), residence (urban/rural), region (northern/southern), income (low/medium/high), smoking status (never/ever), drinking status (never/ever), BMI (normal weight/overweight/obesity), total intake of energy (continuous), physical activities (low/medium/high), intake of whole grains, nuts, fish and shellfish, red meat, poultry, pastries and sweets, tea and coffee (continuous), and chronic diseases, including hypertension, diabetes, and myocardial infarction (yes/no). CHNS, China Health and Nutrition Survey; BMI, body mass index.
For specific carotenoids, higher intakes of β‐cryptoxanthin, α‐carotene, β‐carotene, lycopene, and lutein + zeaxanthin were each associated with slower global cognitive decline (Figure 2). Comparing participants in the top quintile to those in the bottom quintile, the effect estimates (95% CI) were 0.037 (0.020–0.055) for α‐carotene, 0.032 (0.014–0.049) for β‐cryptoxanthin, 0.020 (0.002–0.037) for lutein + zeaxanthin, 0.018 (0.001–0.035) for lycopene, and 0.019 (0.001–0.036) for β‐carotene, respectively. Stronger associations were also observed at the fourth quintile for lutein + zeaxanthin and β‐carotene, with β Q4 vs. Q1 being 0.030 (95% CI: 0.012–0.047) and 0.034 (95% CI: 0.017–0.052), whereas a significant association was seen at the highest intake level of lycopene. In addition, significant associations persisted for dietary carotene calculated by CFCT (β Q5 vs. Q1 = 0.032, 95% CI: 0.015–0.049), and β‐cryptoxanthin (β Q5 vs. Q1 = 0.042, 95% CI: 0.025–0.059), β‐carotene (β Q5 vs. Q1 = 0.033, 95% CI: 0.016–0.051), and α‐carotene (β Q5 vs. Q1 = 0.019, 95% CI: 0.002–0.036) from JFCT (Table S4).
Figure 2.
Associations of total and specific energy‐adjusted carotenoid intake with global cognitive decline. (a) Total carotenoid; (b) α‐carotene; (c) β‐carotene; (d) β‐cryptoxanthin; (e) lycopene; (f) lutein + zeaxanthin. Models were adjusted for age, age squared, sex, education (illiteracy/primary school/middle school and above), residence (urban/rural), region (northern/southern), income (low/medium/high), smoking status (never/ever), drinking status (never/ever), BMI (normal weight/overweight/obesity), total intake of energy (continuous), physical activities (low/medium/high), intake of whole grains, nuts, fish and shellfish, red meat, poultry, pastries and sweets, tea and coffee (continuous), and chronic diseases, including hypertension, diabetes, and myocardial infarction (yes/no). BMI, body mass index.
Stratified analyses showed no evidence of effect modification by age, sex, income, education, smoking status, drinking status, or BMI. However, we observed statistically significant interactions for physical activity (median/high vs. low), indicating a stronger protective association among physically active participants (p‐interaction < 0.001; Figure S3). As for sensitivity analyses (Table S5), the results were unchanged when restricted to participants with complete follow‐up data or to those with at least two cognitive assessments, and remained similar after randomly selecting one participant per household to account for household clustering.
4. Discussion
In this cohort of 4043 Chinese middle‐aged and older adults followed for a median of 3 years, higher dietary carotenoid intake was associated with slower subsequent cognitive decline, with evidence of potential nonlinearity. Compared to participants with the lowest intake level (median 3.1 mg/day), the strongest association was observed at a median intake of 23.5 mg/day. Similar trends were found for specific carotenoids, including α‐carotene, β‐cryptoxanthin, lutein + zeaxanthin, and β‐carotene.
Our findings align with previous cohort studies conducted among the US population. In the Nurses' Health Study cohort of 16,010 participants, higher total dietary carotenoids (median intake 21.3 vs. 8.63 mg/day; median follow‑up 6.4 years) was linked to better objective cognitive function, with similar benefits for lycopene and lutein + zeaxanthin [11]. Another large‐scale prospective investigation within the same cohort (n = 49,493) reported that long‐term higher intake of total carotenoids was associated with 33% lower odds of poor subjective cognitive function in older age (median intake 21.7 vs. 9.40 mg/day; follow‐up of 30 years). Specific carotenoids, including β‐cryptoxanthin, α‐carotene, β‐carotene, lycopene, and lutein + zeaxanthin, each showed significant reverse associations [9]. In concurrence, the Rush Memory and Aging Project (n = 927) found that higher intake of total carotenoids (median intake 24.8 vs. 6.7 mg/day) was associated with a 48% lower risk of Alzheimer's disease [8].
Furthermore, prior cohort studies also support neuroprotective roles for specific carotenoids. In one study of 960 participants from the Rush Memory and Aging Project, higher intakes of lutein + zeaxanthin (median intake 7.89 vs. 2.03 mg/day) and β‐carotene (median intake 7.86 vs. 2.68 mg/day, but not significant) were each associated with 0.04 and 0.02 units slower cognitive decline [8]. In addition, a recent study from the Age‐Related Eye Disease Study (AREDS) and AREDS2 (n = 6334) found protective associations of higher intake of lutein, zeaxanthin, β‐carotene, and lycopene with lower cognitive impairment risk and higher cognitive function scores [3]. Taken together, these findings provide substantial evidence that dietary carotenoids may help preserve cognitive health, and this effect likely extends beyond any single carotenoid.
While acknowledging potential variations, our study reinforces the beneficial role of both total and specific dietary carotenoids in slowing cognitive decline. Based on previous findings from Western populations, our study extends this evidence to the Chinese population, which represents one of the largest aging societies worldwide. Although the effect sizes appear modest, this magnitude is expected given that cognitive outcomes were standardized as z‐scores and is consistent with estimates from prior epidemiologic studies (range: 0.001–0.05/year) [10, 25, 26, 27, 28, 29]. Such a modest yet cumulative effect may hold potential public health implications. Notably, in the current study, the most pronounced protective association for total carotenoid intake was observed at the fourth quintile (median intake: 23.5 mg/day), rather than the fifth quintile (median intake: 37.0 mg/day). The median carotenoid intake among the highest quintile in Western populations ranged from 21.7 to 24.8 mg/day [8, 9], which approximately corresponds to the fourth quintile in our study [8, 9]. This nonlinear pattern may suggest that cognitive benefits of higher carotenoid intake could reach a plateau beyond a certain level, rather than continue to increase proportionally [16, 30]. Indeed, previous trials have linked high‐dose β‐carotene supplementation (20–30 mg/day) to increased lung cancer risk and mortality in high‐risk groups, such as smokers or those with asbestos exposure [14, 15]. Therefore, further evidence is still required to determine the optimal ranges for dietary carotenoid intake. Nevertheless, substantial evidence continues to support the cognitive benefits of carotenoid‐rich foods, including green‐leafy, red, and yellow vegetables and fruits, suggesting that naturally sourced carotenoids may confer protective effects [10, 31].
Despite an incomplete understanding of the underlying mechanisms, it has been suggested that the neuroprotective effect is mainly attributable to the antioxidant properties of carotenoids [32]. Previous evidence showed that carotenoids may suppress proinflammatory cytokines, reduce oxidative stress, and limit lipid peroxidation [33, 34], thus limiting neuronal damage from free radicals and neuroinflammatory processes, which potentially serve as a modifiable risk factor for cognitive decline [4]. Moreover, carotenoids were reported to exert neuroprotective activity against pathology, including limiting amyloid‐beta deposition and tau protein aggregation [8, 35, 36]. Nonetheless, more research is required to confirm and extend our understanding of these neuroprotective roles.
To our knowledge, this is one of the few prospective studies to evaluate dietary carotenoids in relation to cognitive decline among the Chinese population, which further refined and expanded the evidence base for dietary prevention of cognitive decline. Strengths of our study include its longitudinal design, relatively large sample size, and representativeness of the middle‐aged and older Chinese population. In addition, the prudent design of sensitivity analyses and rigorous control of potential confounding factors further contributed to the robustness of our results. However, we should take several limitations into account when interpreting results. First, while repeated 24‐h recalls across waves helped improve the estimation of habitual intake, they still fail to capture long‑term dietary patterns or seasonal variation within each wave. Second, carotenoid intakes were primarily calculated using the USDA Nutrient Database, which may not fully reflect the food composition of the Chinese diet. To address this, we additionally applied the CFCT and JFCT. Despite their incomplete carotenoid coverage, the consistent results across three data sources strengthen the robustness of our primary findings, though residual measurement error cannot be ruled out. Third, cognitive function was assessed using the Telephone Interview for Cognitive Status‐Modified, which covers selected domains and may not capture the full spectrum of cognition, although the instrument has been validated as an effective screening tool. Fourth, despite careful adjustment for multiple confounders, residual confounding and potential reverse causality could still exist. Fifth, the loss to follow‐up could introduce selection bias, but we have conducted several sensitivity analyses, and the associations were consistent. Sixth, the data analyzed were collected between 1997 and 2006, which limits the temporal generalizability of our findings. We therefore emphasize that future studies with more recent data are essential to understand the evolving dietary carotenoids‐cognitive health relationship in China. Seventh, due to methodological restrictions and potential measurement error, we cannot derive specific recommendations for optimal intake levels based on our current data. Eighth, high correlation and potential multicollinearity (tested by variance inflation factors, range: 10.6–77.0) between carotenoid subgroups prevented mutual adjustment in the models, complicating the interpretation of their independent effects. Finally, our sample was drawn exclusively from the Chinese population; the generalizability of the findings to other populations may be limited.
5. Conclusions
Our findings suggested that higher dietary carotenoid intake may slow cognitive decline in Chinese middle‐aged and older adults, with the strongest protective association observed at a moderate intake level. Further study is needed to validate these results and to explore the optimal intake level and underlying mechanisms.
Author Contributions
Liyan Huang: writing – original draft, formal analysis. Ting Shen: writing – review and editing, software, validation. Minyu Wu: writing – review and editing, software. Yiying Gong: writing – review and editing, data curation. Gulisiya Hailili: writing – review and editing, software. Ziping Wang: writing – review and editing, data curation. Caifeng Zhao: writing – review and editing, data curation. Wei Chen: writing – review and editing, methodology. Shuang Rong: methodology, writing – review and editing. Changzheng Yuan: methodology, supervision, validation, writing – review and editing.
Ethics Statement
This study was conducted in accordance with the Declaration of Helsinki and its subsequent revised editions. The original study protocol was approved by the Institutional Review Boards of the University of North Carolina at Chapel Hill and the National Institute of Nutrition and Food Safety at the Chinese Center for Disease Control and Prevention. In addition to the approvals obtained by the original cohorts, the secondary analysis has been reviewed and approved by the Ethics Committee of Zhejiang University School of Public Health (Approval Code: ZGL202403‐2[TS1.1], Approval Date: March 21, 2024).
Consent
All participants provided written informed consent.
Conflicts of Interest
Changzheng Yuan is a member of the Health Care Science Editorial Board. To minimize bias, she was excluded from all editorial decision‐making related to the acceptance of this article for publication. The other authors declare no conflicts of interest.
Supporting information
Figure S1: Flowchart of participants inclusion.
Figure S2: Major food sources of carotenoids by subgroups.
Figure S3: Multivariable‐adjusted associations of total energy‐adjusted carotenoid intake (Q5 vs. Q1) with cognitive decline in subgroups.
Table S1: Spearman correlation coefficients among individual carotenoids.
Table S2: Fixed and random effects from linear mixed‐effects models of global cognitive decline on total energy‐adjusted carotenoids intake.
Table S3: Associations of total energy‐adjusted carotenoids intake with composite cognitive decline.
Table S4: Associations of individual energy‐adjusted carotenoid intake calculated by CFCT and JFCT with global cognitive decline.
Table S5: Sensitivity analysis for associations of total energy‐adjusted carotenoid intake with global cognitive decline.
Acknowledgments
The authors thank the staff and all the participants of the China Health and Nutrition Survey. The authors acknowledge the use of DeepSeek (V4‐Flash‐0731) for language polishing and grammatical checking in the preparation of this manuscript. The authors have carefully reviewed all AI‐assisted content and take full responsibility for the published work. This abstract was accepted by the Nutrition 2024 conference; it was presented orally on June 30, 2024, in Chicago, United States, and the conference abstract has been published by Current Developments in Nutrition.
Huang L., Shen T., Wu M., et al., “Dietary Carotenoids and Cognitive Decline: A Longitudinal Study Among Chinese Middle‐Aged and Older Adults,” Health Care Science 5 (2026): 415‐423. 10.1002/hcs2.70096.
Liyan Huang and Ting Shen contributed equally as co‐first authors.
Contributor Information
Shuang Rong, Email: rongshuang@ustc.edu.cn.
Changzheng Yuan, Email: chy478@zju.edu.cn.
Data Availability Statement
The data sets and study materials that support the findings of our study can be found on the China Health and Nutrition Survey official website.
References
- 1. Rajaram S., Jones J., and Lee G. J., “Plant‐Based Dietary Patterns, Plant Foods, and Age‐Related Cognitive Decline,” Advances in Nutrition 10 (2019): S422–S436, 10.1093/advances/nmz081. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Livingston G., Huntley J., Sommerlad A., et al., “Dementia Prevention, Intervention, and Care: 2020 Report of the Lancet Commission,” Lancet 396, no. 10248 (2020): 413–446, 10.1016/S0140-6736(20)30367-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Keenan T., Agrón E., and Chew E. Y., “Dietary Nutrient Intake and Cognitive Function in the Age‐Related Eye Disease Studies 1 and 2,” Alzheimer's & Dementia 19, no. 10 (2023): 4311–4324, 10.1002/alz.13033. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Tan B. L. and Norhaizan M. E., “Carotenoids: How Effective Are They to Prevent Age‐Related Diseases?” Molecules 24, no. 9 (2019): 1801, 10.3390/molecules24091801. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Davinelli S., Ali S., Solfrizzi V., Scapagnini G., and Corbi G., “Carotenoids and Cognitive Outcomes: A Meta‐Analysis of Randomized Intervention Trials,” Antioxidants 10, no. 2 (2021): 223, 10.3390/antiox10020223. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Zhou F., Xie X., Zhang H., and Liu T., “Effect of Antioxidant Intake Patterns on Risks of Dementia and Cognitive Decline,” European Geriatric Medicine 14, no. 1 (2023): 9–17, 10.1007/s41999-022-00720-7. [DOI] [PubMed] [Google Scholar]
- 7. Dorey C. K., Gierhart D., Fitch K. A., Crandell I., and Craft N. E., “Low Xanthophylls, Retinol, Lycopene, and Tocopherols in Grey and White Matter of Brains With Alzheimer's Disease,” Journal of Alzheimer's Disease 94, no. 1 (2023): 1–17, 10.3233/JAD-220460. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Yuan C., Chen H., Wang Y., Schneider J. A., Willett W. C., and Morris M. C., “Dietary Carotenoids Related to Risk of Incident Alzheimer Dementia (AD) and Brain AD Neuropathology: A Community‐Based Cohort of Older Adults,” American Journal of Clinical Nutrition 113, no. 1 (2021): 200–208, 10.1093/ajcn/nqaa303. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Yuan C., Fondell E., Ascherio A., et al., “Long‐Term Intake of Dietary Carotenoids Is Positively Associated With Late‐Life Subjective Cognitive Function in a Prospective Study in US Women,” Journal of Nutrition 150, no. 7 (2020): 1871–1879, 10.1093/jn/nxaa087. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Morris M. C., Wang Y., Barnes L. L., Bennett D. A., Dawson‐Hughes B., and Booth S. L., “Nutrients and Bioactives in Green Leafy Vegetables and Cognitive Decline: Prospective Study,” Neurology 90, no. 3 (2018): e214–e222, 10.1212/wnl.0000000000004815. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Devore E. E., Kang J. H., Stampfer M. J., and Grodstein F., “The Association of Antioxidants and Cognition in the Nurses' Health Study,” American Journal of Epidemiology 177, no. 1 (2013): 33–41, 10.1093/aje/kws202. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Laurin D., “Midlife Dietary Intake of Antioxidants and Risk of Late‐Life Incident Dementia: The Honolulu‐Asia Aging Study,” American Journal of Epidemiology 159, no. 10 (2004): 959–967, 10.1093/aje/kwh124. [DOI] [PubMed] [Google Scholar]
- 13. Devore E. E., Grodstein F., van Rooij F. J. A., et al., “Dietary Antioxidants and Long‐Term Risk of Dementia,” Archives of Neurology 67, no. 7 (2010): 819–825, 10.1001/archneurol.2010.144. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Omenn G. S., Goodman G. E., Thornquist M. D., et al., “Risk Factors for Lung Cancer and for Intervention Effects in CARET, the Beta‐Carotene and Retinol Efficacy Trial,” Journal of the National Cancer Institute 88, no. 21 (1996): 1550–1559, 10.1093/jnci/88.21.1550. [DOI] [PubMed] [Google Scholar]
- 15. The Alpha‐Tocopherol Beta Carotene Cancer Prevention Study Group , “The Effect of Vitamin E and Beta Carotene on the Incidence of Lung Cancer and Other Cancers in Male Smokers,” New England Journal of Medicine 330, no. 15 (1994): 1029–1035, 10.1056/nejm199404143301501. [DOI] [PubMed] [Google Scholar]
- 16. Palozza P., Serini S., Di Nicuolo F., Piccioni E., and Calviello G., “Prooxidant Effects of Β‐Carotene in Cultured Cells,” Molecular Aspects of Medicine 24, no. 6 (2003): 353–362, 10.1016/s0098-2997(03)00031-1. [DOI] [PubMed] [Google Scholar]
- 17. Shang X., Hill E., Li Y., and He M., “Energy and Macronutrient Intakes at Breakfast and Cognitive Declines in Community‐Dwelling Older Adults: A 9‐Year Follow‐up Cohort Study,” American Journal of Clinical Nutrition 113, no. 5 (2021): 1093–1103, 10.1093/ajcn/nqaa403. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Crimmins E. M., Kim J. K., Langa K. M., and Weir D. R., “Assessment of Cognition Using Surveys and Neuropsychological Assessment: The Health and Retirement Study and the Aging, Demographics, and Memory Study,” Journals of Gerontology Series B: Psychological Sciences and Social Sciences 66B, no. S1 (2011): i162–i171, 10.1093/geronb/gbr048. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Batis C., Gordon‐Larsen P., Cole S. R., Du S., Zhang B., and Popkin B., “Sodium Intake From Various Time Frames and Incident Hypertension Among Chinese Adults,” Epidemiology 24, no. 3 (2013): 410–418, 10.1097/EDE.0b013e318289e047. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Yao M., McCrory M. A., Ma G., et al., “Relative Influence of Diet and Physical Activity on Body Composition in Urban Chinese Adults,” American Journal of Clinical Nutrition 77, no. 6 (2003): 1409–1416, 10.1093/ajcn/77.6.1409. [DOI] [PubMed] [Google Scholar]
- 21. Willett W., Howe G., and Kushi L., “Adjustment for Total Energy Intake in Epidemiologic Studies,” American Journal of Clinical Nutrition 65, no. 4 (1997): 1220S–1228S, 10.1093/ajcn/65.4.1220s. [DOI] [PubMed] [Google Scholar]
- 22. Plassman B. L., Welsh K. A., Helms M., Brandt J., Page W. F., and Breitner J. C. S., “Intelligence and Education as Predictors of Cognitive State in Late Life: A 50‐year Follow‐up,” Neurology 45, no. 8 (1995): 1446–1450, 10.1212/wnl.45.8.1446. [DOI] [PubMed] [Google Scholar]
- 23. Shi Z., Li M., Wang Y., Liu J., and El‐Obeid T., “High Iron Intake Is Associated With Poor Cognition Among Chinese Old Adults and Varied by Weight Status: A 15‐y Longitudinal Study in 4852 Adults,” American Journal of Clinical Nutrition 109, no. 1 (2019): 109–116, 10.1093/ajcn/nqy254. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Qin B., Adair L. S., Plassman B. L., et al., “Dietary Patterns and Cognitive Decline Among Chinese Older Adults,” Epidemiology 26, no. 5 (2015): 758–768. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Morris M. C., Tangney C. C., Wang Y., et al., “MIND Diet Slows Cognitive Decline With Aging,” Alzheimer's & Dementia 11, no. 9 (2015): 1015–1022, 10.1016/j.jalz.2015.04.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Liu X., Dhana K., Barnes L. L., et al., “A Healthy Plant‐Based Diet Was Associated With Slower Cognitive Decline in African American Older Adults: A Biracial Community‐Based Cohort,” American Journal of Clinical Nutrition 116, no. 4 (2022): 875–886, 10.1093/ajcn/nqac204. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Huang L., Tao Y., Chen H., et al., “Mediterranean‐Dietary Approaches to Stop Hypertension Intervention for Neurodegenerative Delay (MIND) Diet and Cognitive Function and Its Decline: A Prospective Study and Meta‐Analysis of Cohort Studies,” American Journal of Clinical Nutrition 118, no. 1 (2023): 174–182, 10.1016/j.ajcnut.2023.04.025. [DOI] [PubMed] [Google Scholar]
- 28. Huang L., Zhao C., Gao M., et al., “Associations of Vegetable and Fruit Intake With Cognitive Function and Its Decline: Two Longitudinal Studies,” Journal of Nutrition, Health and Aging 28, no. 6 (2024): 100223, 10.1016/j.jnha.2024.100223. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Cochar‐Soares N., da Silva T. B. P., Luiz M. M., et al., “Vitamin D Deficiency as a Risk Factor for Cognitive Decline in Individuals Aged 50 or Older,” GeroScience 48, no. 2 (2026): 2491–2503, 10.1007/s11357-025-01800-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Shin J., Song M.‐H., Oh J.‐W., Keum Y.‐S., and Saini R. K., “Pro‐Oxidant Actions of Carotenoids in Triggering Apoptosis of Cancer Cells: A Review of Emerging Evidence,” Antioxidants 9, no. 6 (2020): 532, 10.3390/antiox9060532. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Yuan C., Fondell E., Bhushan A., et al., “Long‐Term Intake of Vegetables and Fruits and Subjective Cognitive Function in US Men,” Neurology 92, no. 1 (2019): e63–e75, 10.1212/WNL.0000000000006684. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Cho K. S., Shin M., Kim S., and Lee S. B., “Recent Advances in Studies on the Therapeutic Potential of Dietary Carotenoids in Neurodegenerative Diseases,” Oxidative Medicine and Cellular Longevity 2018, no. 1 (2018): 4120458, 10.1155/2018/4120458. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Decandia D., Gelfo F., Landolfo E., Balsamo F., Petrosini L., and Cutuli D., “Dietary Protection Against Cognitive Impairment, Neuroinflammation and Oxidative Stress in Alzheimer's Disease Animal Models of Lipopolysaccharide‐Induced Inflammation,” International Journal of Molecular Sciences 24, no. 6 (2023): 5921, 10.3390/ijms24065921. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Hira S., Saleem U., Anwar F., Sohail M. F., Raza Z., and Ahmad B., “Β‐Carotene: A Natural Compound Improves Cognitive Impairment and Oxidative Stress in a Mouse Model of Streptozotocin‐Induced Alzheimer's Disease,” Biomolecules 9, no. 9 (2019): 441, 10.3390/biom9090441. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Lakey‐Beitia J., Kumar D J., Hegde M., and Rao K. S., “Carotenoids as Novel Therapeutic Molecules Against Neurodegenerative Disorders: Chemistry and Molecular Docking Analysis,” International Journal of Molecular Sciences 20, no. 22 (2019): 5553, 10.3390/ijms20225553. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Yu L., Wang W., Pang W., Xiao Z., Jiang Y., and Hong Y., “Dietary Lycopene Supplementation Improves Cognitive Performances in Tau Transgenic Mice Expressing P301L Mutation via Inhibiting Oxidative Stress and Tau Hyperphosphorylation,” Journal of Alzheimer's Disease 57, no. 2 (2017): 475–482, 10.3233/JAD-161216. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Figure S1: Flowchart of participants inclusion.
Figure S2: Major food sources of carotenoids by subgroups.
Figure S3: Multivariable‐adjusted associations of total energy‐adjusted carotenoid intake (Q5 vs. Q1) with cognitive decline in subgroups.
Table S1: Spearman correlation coefficients among individual carotenoids.
Table S2: Fixed and random effects from linear mixed‐effects models of global cognitive decline on total energy‐adjusted carotenoids intake.
Table S3: Associations of total energy‐adjusted carotenoids intake with composite cognitive decline.
Table S4: Associations of individual energy‐adjusted carotenoid intake calculated by CFCT and JFCT with global cognitive decline.
Table S5: Sensitivity analysis for associations of total energy‐adjusted carotenoid intake with global cognitive decline.
Data Availability Statement
The data sets and study materials that support the findings of our study can be found on the China Health and Nutrition Survey official website.