ABSTRACT
Background
Previous studies have shown inconsistent results for the relation between dietary fat intake and cognitive function in the elderly. Furthermore, prospective studies on this topic among the Chinese population are scarce.
Objectives
We aimed to examine the association between midlife dietary fat intake and risk of cognitive impairment in the elderly.
Methods
Prospective cohort analysis was conducted among 16,736 participants from the Singapore Chinese Health Study. Dietary information was assessed by a validated FFQ at baseline (1993–1998) when participants aged 45–74 y (mean: 53.5; SD: 6.22). Cognitive impairment was identified using the Singapore modified Mini-Mental State Examination at the third follow-up visit (2014–2016) when participants aged 61–96 y (mean: 73.2; SD: 6.41). Multivariable logistic regression models were used to calculate ORs and 95% CIs.
Results
Cognitive impairment was presented in 2397 participants. When substituted for total carbohydrate, dietary fat intake was inversely related to cognitive impairment (OR comparing extreme quartiles: 0.80; 95% CI: 0.67, 0.94; P-trend = 0.003). The OR (95% CI) comparing extreme quartiles of specific dietary fats was 1.08 (0.89, 1.31; P-trend = 0.51) for SFAs, 0.80 (0.64, 0.99; P-trend = 0.02) for MUFAs, 0.84 (0.72, 0.99; P-trend = 0.02) for PUFAs, 0.92 (0.77, 1.09; P-trend = 0.49) for n–3 PUFAs, and 0.83 (0.70, 0.98; P-trend = 0.01) for n–6 PUFAs. An inverse association was found for plant-based fat (OR: 0.84; 95% CI: 0.72, 0.98; P-trend = 0.02), but not for animal-based fat (OR: 0.96; 95% CI: 0.81, 1.15; P-trend = 0.76). When substituted for SFAs, the OR (95% CI) was 0.77 (0.61, 0.97; P-trend = 0.02) for MUFAs and 0.82 (0.70, 0.95; P-trend = 0.003) for PUFAs.
Conclusions
We found that substitution of total carbohydrate or SFAs with MUFAs and PUFAs, particularly n–6 PUFAs, was related to a lower risk of cognitive impairment in elderly Chinese participants. In addition, an inverse association with cognitive impairment was found for plant-based fat.
Keywords: dietary fat, fatty acids, animal-based fat, plant-based fat, cognitive impairment, cohort study
Introduction
Population aging has become a public health challenge globally, and age-related neurodegenerative disorders such as dementia are expected to increase dramatically along with a huge economic burden. It is estimated that the number of people with dementia will increase to 74.7 million by 2030 with an economic burden of US$2 trillion (1). No effective therapeutic strategy is currently available to cure dementia, and thus identifying modifiable risk factors to prevent the onset of dementia and its clinical phenotype, such as cognitive impairment, is imperative.
Mechanistic studies have indicated that fatty acids (FAs) play important roles in brain structure and function, and the association between dietary fat intake and cognitive function has long been investigated (2). However, the current evidence from epidemiological studies is inconsistent (3). A number of studies have reported no significant relation of total fat intake with risk of cognitive decline or dementia (4–6), whereas 2 studies reported a positive association (7, 8). As for different types of FAs, both positive (5–7) and nonsignificant positive and negative (4, 9–14) associations have been reported for SFAs. Although some studies showed better cognition and less cognitive decline with higher intakes of MUFAs (6, 11, 14, 15) and PUFAs (14, 16), other studies showed no significant associations for MUFAs (4, 5, 7, 9, 10, 12, 13) or PUFAs (6, 9, 10). Although intake of n–3 PUFAs was found to be associated with lower risks of cognitive decline and dementia in some cohort studies (4, 16, 17), other studies did not find significant associations (18–20), and a meta-analysis of 12 randomized clinical trials (RCTs) failed to show a beneficial effect of n–3 PUFAs on cognitive function among elderly people (21). Studies focusing on n–6 PUFAs are limited and report inconsistent results (4, 9, 17, 22). The inconsistency of the literature could be due to differences in the study populations (age, sex, socioeconomic status, lifestyle, health status, etc.), follow-up years, exposures and outcome measures, as well as differences in statistical models.
Most of the aforementioned studies were conducted in Western populations, and there is very limited data from Asians, particularly Chinese adults. Unlike the Western dietary pattern with fat mainly derived from meat and dairy products, a major dietary source of fat in Chinese population is plant-based cooking oil (23). Fat from different food sources may have different biological effects on human health (24). In addition, very few long-term follow-up studies have evaluated the association between dietary fat intake in midlife and risk of cognitive function in later life. Therefore, to fill the literature gap, we aimed to comprehensively explore the relations of dietary intakes of fat and FAs with the risk of cognitive impairment among Chinese individuals living in Singapore.
Methods
Study population
The Singapore Chinese Health Study was a population-based cohort study among 63,257 participants aged 45–74 y at recruitment (April 1993 to December 1998). Participants were Chinese Singaporean citizens or permanent residents living in government-built housing estates, where 86% of Singapore residents lived during the time of recruitment. They came from 2 major Chinese dialect groups, the Hokkiens and the Cantonese, who were originally from Fujian and Guangdong provinces in Southern China, respectively. The participants were followed approximately every 5 y, and only participants who finished cognition testing during the third follow-up visit (2014–2016) were included in the present analysis (Supplemental Figure 1). All enrolled participants provided informed consent and the Institute Review Board of the National University of Singapore approved the study.
Assessment of diet and covariates
At recruitment, all participants were interviewed face-to-face in their homes using structured questionnaires to collect information on demographics, height, weight, smoking status, alcohol consumption, physical activity, sleep duration, history of physician-diagnosed medical conditions, including cancer, coronary heart disease, stroke, hypertension, and diabetes. History of cancer was additionally ascertained by linking with the database from the Singapore Cancer Registry.
Information on habitual dietary intake was gathered using a validated semiquantitative FFQ at baseline. In brief, participants were questioned about their usual dietary intake of 165 commonly consumed foods in Singapore in the previous year, in terms of intake frequencies (ranging from “never or hardly ever” to “2 or more times a day”) and portion sizes (small, medium, large). Food categories in the FFQ mainly included rice and noodles, meat, seafood, vegetables and fruits, cereals, legumes, dairy products, and cooking fats and oils. The nutrients and energy intakes were subsequently calculated using FFQ data and the Singapore Food Composition Database, which was specifically developed for this cohort (25). The FFQ validation study was conducted by comparing data from repeated FFQ and 2 24-h recalls among a subset of 810 participants (25). For the macronutrients (carbohydrate, fat, and protein), the energy-adjusted correlation coefficients comparing estimated intakes from 24-h recalls with those from FFQ ranged from 0.36 to 0.61 (25). In our study, SFAs intake included FA 4:0, FA 6:0, FA 8:0, FA 10:0, FA 12:0, FA 14:0, FA 15:0, FA 16:0, FA 17:0, FA 18:0, FA 20:0, FA 22:0, and FA 24:0. MUFAs intake included FA 14:1, FA 16:1, FA 18:1, FA 20:1, and FA 22:1. PUFAs intake included FA 18:2, FA 18:3 (including α-linolenic acid and γ-linolenic acid), FA 18:4, FA 20:4, FA 20:5 (EPA), FA 22:5 (docosapentenoic acid; DPA), and FA 22:6 (DHA). In our study, n–3 PUFAs mainly included α-linolenic acid, EPA, DPA, and DHA.
Assessment of cognitive function
Cognitive testing was only conducted during the third follow-up visit (July 2014 to February 2016) for the consenting participants at their homes using a Singapore modified version of the Mini-Mental State Examination (SM-MMSE) (26). The test consists of 30 items and the score is from 0 to 30, with higher MMSE scores representing better cognitive function. Western studies often used <24 as the cut-off point for the definition of cognitive impairment (27); however, MMSE scores could be heavily influenced by education level. Unlike many studies in Western populations, education level among our study population was relatively low: 63.6% individuals had ≤6 y of education. Hence, we used education-specific cut-off points for the definition of cognitive impairment: <18 for those with no formal education, <21 for those with 1–6 y of education, and <25 for those with ≥7 y of education. These cut-off points originated from a Shanghai dementia screening survey among 5055 Chinese participants with a comparable education level to our study population (28).
The SM-MMSE was completed by in-person interviews. The interviewers were trained by a geriatric and neurocognitive assessment expert in the team (LF). All the interviews were audio recorded and 20% of these recordings were randomly selected for quality control. Interviewers who did not follow the stringent protocol were retrained and asked to pass an examination before they could resume work.
Among the 63,257 participants enrolled in the Singapore Chinese Health Study, 18,148 participants died before the third follow-up visit, and 17,107 participants were contacted again and the SM-MMSE test performed at the third follow-up visit. The third follow-up visit was stopped because of funding restraint, and we did not intentionally select participants for visits. After further exclusion of 55 participants who had missing items in the SM-MMSE, 104 participants who were incapable of completing the SM-MMSE questionnaire (mute, blind, or deaf), and 212 participants who reported extreme energy intake (<700 kcal/d or >3700 kcal/d for men; <600 or >3000 kcal/d for women), the final analysis included 16,736 participants. The selection of participants is shown in Supplemental Figure 1.
Statistical analysis
Energy-adjusted intakes of fat were calculated using the residual method (29) and categorized into quartiles. Spearman's rank correlation coefficients were used to compare the correlations between different FAs. The characteristics of the participants were compared by chi-square test (for categorical variables) or ANOVA (for continuous variables) across quartiles of fat intake.
Multivariate-adjusted logistic regression models were used to estimate ORs and 95% CIs of cognitive impairment comparing extreme quartiles of fat intake. To quantify the linear trend, we assigned the median intake value to each respective quartile and treated these values as a continuous variable in the models. In model 1, we adjusted for age at cognition assessment (years), sex, education level (no formal education, primary school, secondary school, diploma/college, or higher), marital status (married, widowed, separated/divorced, never married), dialect groups (Hokkien, Cantonese), and total energy intake (kcal/d). In model 2, we further included smoking status (never, former, current), alcohol consumption (never/monthly, weekly, daily), moderate or vigorous physical activity level (<0.5, 0.5–3.9, ≥4.0 h/wk), BMI (<18.5, 18.5–22.9, 23.0–27.4, ≥27.5 kg/m2), sleep duration (≤5, 6, 7, 8, ≥9 h/d), physician-diagnosed history of diabetes, hypertension, heart disease, stroke, and cancer (yes, no), and dietary intakes of protein (g/d), fiber (g/d), and total cholesterol (mg/d) (all in quartiles). When examining the different types of FAs (SFAs, MUFAs, and PUFAs), they were added in the model simultaneously; when examining the n–3 and n–6 PUFAs, these 2 variables were added in the model together with SFAs and MUFAs. The OR in model 2 can be interpreted as the estimated effect size of replacing a certain amount of energy intake from dietary carbohydrate with equivalent energy from fat, which is common practice in the nutritional epidemiology field (29). We further tested the substitution association of other FAs with SFAs by including dietary carbohydrate in model 3 but excluding SFAs. We also classified total fat intake into plant-based fat and animal-based fat based on the food sources, and conducted similar analyses to investigate their individual association with cognitive impairment. Animal-based fat was FAs mainly from red meat, poultry, fish/shellfish, eggs, and dairy products. Plant-based fat mainly included FAs from vegetable oils, legumes, all vegetables and fruits, grain products, nuts and seeds.
We performed a set of sensitivity analyses. First, we used the nutrient-density method to compute dietary intake instead of the residual method in the main analysis, and dietary macronutrient intake was presented as percentage of energy. In this analysis, we also treat exposure as a continuous variable and computed the OR of substituting 5% energy from carbohydrate by fat and from SFAs by other FAs. Second, we repeated analysis restricted to participants without cancer or cardiovascular disease at baseline. Third, we used <24 as the cut-off point to define cognitive impairment to explore the robustness of our results. Fourth, we treated the MMSE score as a continuous variable and evaluated the association between fat intake and MMSE score using a generalized linear regression model. Finally, in order to decrease the heteroscedasticity in the estimated intake of energy and nutrients, the nutrients were first log-transformed and then used to calculate the residuals (29), which were further used as exposures (quartiles) to explore the relations with cognitive impairment.
We also conducted exploratory analyses of testing multiplicative interactions by including fat intake (quartiles), the potential interaction factor, and their interaction term simultaneously in model 2. The potential effect modifiers included age at cognitive assessment (aged <75, ≥75 y), sex, education level (no formal education, primary school, and higher education), BMI level (<23, ≥23), smoking status (never, former, current), alcohol consumption (never/monthly, weekly, daily), and history of chronic diseases. For those with significant interactions, stratified analyses were subsequently conducted.
All analyses were performed using SAS software (version 9.4; SAS Institute, Inc.), with 2-tailed P values <0.05 considered as statistically significant.
Results
We documented 2397 cases of cognitive impairment in the third follow-up visit. The mean age of the participants was 53.5 y (SD: 6.22) at baseline and 73.2 y (SD: 6.41) at the third follow-up interview. Characteristics of participants in the first and fourth quartiles of dietary intakes of total and 3 main types of FAs are shown in Table 1. Participants with a higher intake of fat (including all types of FAs) were more likely to be younger, women, have a higher intake of protein and lower intake of carbohydrate, but were less likely to be current smokers and alcohol drinkers. Participants with higher intakes of SFAs and MUFAs were more likely to have a higher BMI and lower physical activity, whereas participants with a higher intake of PUFAs were more likely to be physically active. Participants with higher intakes of MUFAs and PUFAs had higher education levels. In this population, 66.5% of dietary fat came from plant-based foods. Different types of FAs were highly correlated with total fat intake and the Spearman coefficients ranged from 0.54 (n–6 PUFAs) to 0.92 (MUFAs), whereas the Spearman coefficients between major FAs ranged from 0.02 between SFAs and n–6 PUFAs to 0.99 between n–6 PUFAs and PUFAs (Table 2).
TABLE 1.
Characteristics of participants according to quartiles of fat intake:1 the Singapore Chinese Health Study
| Total fat | SFAs | MUFAs | PUFAs | |||||
|---|---|---|---|---|---|---|---|---|
| Characteristics | Q1 | Q4 | Q1 | Q4 | Q1 | Q4 | Q1 | Q4 |
| Median intake,2 g/d | 33.9 (26.7–37.1) | 55.3 (52.7–59.4) | 10.7 (8.64–12.0) | 20.7 (19.5–22.7) | 11.1 (8.83–12.1) | 19.0 (18.0–20.6) | 5.95 (4.63–6.53) | 12.9 (11.6–14.9) |
| Age at baseline, y | 54.2 ± 6.27 | 52.5 ± 5.94* | 54.0 ± 6.22 | 52.7 ± 5.97* | 54.1 ± 6.25 | 52.7 ± 5.99* | 54.0 ± 6.25 | 52.8 ± 6.00* |
| Age at interview, y | 74.0 ± 6.34 | 71.8 ± 6.17* | 74.0 ± 6.33 | 71.9 ± 6.21* | 73.9 ± 6.33 | 72.0 ± 6.23* | 73.5 ± 6.35 | 72.8 ± 6.26* |
| Sex, female | 1722 (41.2) | 2763 (66.0)* | 1845 (44.1) | 2674 (63.9)* | 1739 (41.6) | 2700 (64.5)* | 1725 (41.2) | 2670 (63.8)* |
| Dialect group | ||||||||
| Cantonese | 2188 (52.3) | 2132 (51.0)* | 2305 (55.1) | 2043 (48.8)* | 2179 (52.1) | 2149 (51.4)* | 2075 (49.6) | 2224 (53.2)* |
| Hokkien | 1996 (47.7) | 2052 (49.0) | 1879 (44.9) | 2141 (51.2) | 2005 (47.9) | 2035 (48.6) | 2109 (50.4) | 1960 (46.9) |
| Education level | ||||||||
| No formal | 751 (18.0) | 631 (15.1)* | 682 (16.3) | 729 (17.4)* | 701 (16.8) | 673 (16.1)* | 763 (18.2) | 568 (13.6)* |
| Primary | 2061 (49.3) | 1675 (40.0) | 1953 (46.7) | 1787 (42.7) | 2011 (48.1) | 1705 (40.8) | 2136 (51.1) | 1646 (39.3) |
| Secondary | 1108 (26.5) | 1488 (35.6) | 1209 (28.9) | 1366 (32.7) | 1184 (28.3) | 1442 (34.5) | 1035 (24.7) | 1502 (35.9) |
| Higher | 264 (6.3) | 390 (9.3) | 340 (8.1) | 302 (7.2) | 288 (6.9) | 364 (8.7) | 250 (6.0) | 468 (11.2) |
| Marital status | ||||||||
| Married | 3735 (89.3) | 3740 (89.4) | 3782 (90.4) | 3707 (88.6)* | 3739 (89.4) | 3708 (88.6) | 3698 (88.4) | 3782 (90.4)* |
| Others | 449 (10.7) | 444 (10.6) | 402 (9.6) | 477 (11.4) | 445 (10.6) | 476 (11.4) | 486 (11.6) | 402 (9.6) |
| Tobacco smoking | ||||||||
| Never | 2818 (67.4) | 3414 (81.6)* | 2969 (71.0) | 3308 (79.1)* | 2860 (68.4) | 3362 (80.4)* | 2780 (66.4) | 3487 (83.3)* |
| Ever | 543 (13.0) | 316 (7.6) | 516 (12.3) | 331 (7.9) | 533 (12.7) | 323 (7.7) | 516 (12.3) | 341 (8.2) |
| Current | 823 (19.7) | 454 (10.9) | 699 (16.7) | 545 (13.0) | 791 (18.9) | 499 (11.9) | 888 (21.2) | 356 (8.5) |
| Alcohol drinking | ||||||||
| Never/monthly | 3457 (82.6) | 3813 (91.1)* | 3522 (84.2) | 3752 (89.7)* | 3482 (83.2) | 3783 (90.4)* | 3407 (81.4) | 3823 (91.4)* |
| Weekly | 483 (11.5) | 322 (7.7) | 454 (10.9) | 366 (8.8) | 477 (11.4) | 340 (8.1) | 530 (12.7) | 303 (7.2) |
| Daily | 244 (5.8) | 49 (1.2) | 208 (5.0) | 66 (1.6) | 225 (5.4) | 61 (1.5) | 247 (5.9) | 58 (1.4) |
| Physical activity | ||||||||
| <0.5 h/wk | 2505 (59.9) | 2671 (63.8)* | 2442 (58.4) | 2738 (65.4)* | 2432 (58.1) | 2722 (65.1)* | 2559 (61.2) | 2484 (59.4)* |
| 0.5–3.9 h/wk | 1003 (24.0) | 1012 (24.2) | 1014 (24.2) | 965 (23.1) | 1041 (24.9) | 961 (23.0) | 1004 (24.0) | 1063 (25.4) |
| ≥4.0 h/wk | 676 (16.2) | 501 (12.0) | 728 (17.4) | 481 (11.5) | 711 (17.0) | 501 (12.0) | 621 (14.8) | 637 (15.2) |
| BMI, kg/m2 | ||||||||
| <18.5 | 271 (6.5) | 205 (4.9)* | 257 (6.1) | 223 (5.3)* | 268 (6.4) | 216 (5.2)* | 260 (6.2) | 213 (5.1)* |
| 18.5–22.9 | 1870 (44.7) | 1827 (43.7) | 1899 (45.4) | 1780 (42.5) | 1863 (44.5) | 1827 (43.7) | 1830 (43.7) | 1862 (44.5) |
| 23–24.9 | 1741 (41.6) | 1752 (41.9) | 1718 (41.1) | 1763 (42.1) | 1723 (41.2) | 1747 (41.8) | 1775 (42.4) | 1772 (42.4) |
| ≥25 | 302 (7.2) | 400 (9.5) | 310 (7.4) | 418 (10.0) | 330 (7.9) | 394 (9.4) | 319 (7.6) | 337 (8.1) |
| Sleep duration | ||||||||
| ≤5 | 314 (7.5) | 382 (9.1)* | 294 (7.0) | 385 (9.2)* | 310 (7.4) | 372 (8.9) | 320 (7.7) | 332 (7.9) |
| 6–8 h | 3645 (87.1) | 3546 (84.8) | 3670 (87.7) | 3536 (84.5) | 3660 (87.5) | 3556 (85.0) | 3628 (86.7) | 3627 (86.7) |
| ≥9 h | 225 (5.4) | 256 (6.1) | 220 (5.3) | 263 (6.3) | 214 (5.1) | 256 (6.1) | 236 (5.6) | 225 (5.4) |
| Hypertension | 783 (18.7) | 790 (18.9) | 820 (19.6) | 772 (18.5) | 814 (19.5) | 760 (18.2) | 741 (17.7) | 836 (20.0)* |
| Heart disease | 91 (2.2) | 84 (2.0) | 96 (2.3) | 68 (1.6) | 93 (2.2) | 74 (1.8)* | 76 (1.8) | 102 (2.4) |
| Stroke | 24 (0.6) | 21 (0.5) | 21 (0.5) | 15 (0.4) | 24 (0.6) | 19 (0.5)* | 19 (0.5) | 27 (0.7) |
| Diabetes mellitus | 145 (3.5) | 280 (6.7)* | 174 (4.2) | 237 (5.7)* | 149 (3.6) | 277 (6.6)* | 142 (3.4) | 231 (5.5)* |
| Cancer | 67 (1.6) | 86 (2.1) | 67 (1.6) | 85 (2.0)* | 66 (1.6) | 84 (2.0) | 67 (1.6) | 90 (2.2) |
| Total energy, kcal/d | 1760 ± 539 | 1720 ± 541* | 1780 ± 507 | 1690 ± 537* | 1770 ± 524 | 1700 ± 542* | 1770 ± 529 | 1650 ± 527* |
| Protein, g/d | 50.6 ± 7.98 | 67.4 ± 9.52* | 53.0 ± 9.98 | 65.1 ± 9.63* | 51.0 ± 8.36 | 66.8 ± 9.42* | 52.9 ± 9.21 | 63.6 ± 9.89* |
| Carbohydrate, g/d | 261 ± 23.9 | 194 ± 19.7* | 255 ± 28.7 | 201 ± 22.9* | 260 ± 24.7 | 196 ± 21.2* | 251 ± 29.1 | 213 ± 27.5* |
| Cholesterol, mg/d | 124 ± 60.6 | 217 ± 83.6* | 128 ± 64.0 | 213 ± 80.7* | 121 ± 58.3 | 220 ± 82.9* | 146 ± 72.2 | 181± 76.3* |
| Total fat, g/d | 31.5 ± 6.70 | 57.0 ± 6.07* | 33.9 ± 9.30 | 54.6 ± 7.27* | 32.3 ± 7.55 | 56.1 ± 6.80* | 35.3 ± 9.36 | 50.4 ± 9.45* |
| SFAs, g/d | 10.9 ± 3.27 | 20.2 ± 4.10* | 9.99 ± 2.51 | 21.5 ± 2.81* | 11.1 ± 3.57 | 19.9 ± 4.08* | 14.1 ± 4.69 | 14.8 ± 4.40* |
| MUFAs, g/d | 10.6 ± 2.51 | 19.3 ± 2.69* | 11.3 ± 3.27 | 18.5 ± 3.01* | 10.3 ± 2.27 | 19.7 ± 2.32* | 12.0 ± 3.25 | 16.3 ± 3.97* |
| PUFAs, g/d | 6.61 ± 2.66 | 11.5 ± 4.12* | 8.91 ± 4.45 | 9.01 ± 2.60 | 7.31 ± 3.55 | 10.7 ± 3.51* | 5.47 ± 1.36 | 13.8 ± 3.18* |
| Animal-based fat, g/d | 10.2 ± 4.92 | 19.2 ± 7.01* | 10.1 ± 4.93 | 19.4 ± 6.71* | 9.96 ± 4.82 | 19.6 ± 6.84* | 12.8 ± 6.34 | 15.4 ± 6.61* |
| Plant-based fat, g/d | 21.3 ± 5.85 | 37.8 ± 7.82* | 23.9 ± 8.27 | 35.2 ± 7.87* | 22.3 ± 6.82 | 36.5 ± 8.16* | 22.5 ± 6.78 | 35.0 ± 8.35* |
All data are shown as n (%) or mean ± SD except for median intake. *P <0.05.
Median intakes are shown as median value (IQR).
TABLE 2.
Spearman correlation coefficients between dietary fat intake (energy-adjusted)1 in the Singapore Chinese Health Study
| Dietary fat | Total fat | SFAs | MUFAs | PUFAs | n–3 PUFAs | n–6 PUFAs |
|---|---|---|---|---|---|---|
| Total fat | 1.00 | 0.78 | 0.92 | 0.56 | 0.56 | 0.54 |
| SFAs | — | 1.00 | 0.74 | 0.05 | 0.33 | 0.02 |
| MUFAs | — | — | 1.00 | 0.44 | 0.44 | 0.42 |
| PUFAs | — | — | — | 1.00 | 0.57 | 0.99 |
| n–3 PUFAs | — | — | — | — | 1.00 | 0.52 |
| n–6 PUFAs | — | — | — | — | — | 1.00 |
All P values <0.05.
The relations of dietary fat intake with risk of cognitive impairment are shown in Table 3. When replacing for total carbohydrate (model 2), a higher dietary intake of total fat was associated with a lower risk of cognitive impairment and the OR (95% CI) was 0.80 (0.67, 0.94) comparing extreme quartiles (P-trend = 0.003). A significant association was found for plant-based fat (OR: 0.84; 95% CI: 0.72, 0.98; P-trend = 0.02) but not for animal-based fat (OR: 0.96; 95% CI: 0.81, 1.15; P-trend = 0.76). As for specific FAs, both MUFAs and PUFAs showed a significantly inverse association and the OR (95% CI) comparing extreme quartiles was 0.80 (0.64, 0.99; P-trend = 0.02) and 0.84 (0.72, 0.99; P-trend = 0.02), respectively, whereas no significant association was found for SFAs (OR comparing extreme quartiles: 1.08; 95% CI: 0.89, 1.31; P-trend = 0.51). When substituting for SFAs (model 3), higher intakes of MUFAs and PUFAs were both significantly associated with a lower risk of cognitive impairment (OR comparing extreme quartiles: 0.77; 95% CI: 0.61, 0.97; P-trend = 0.02 for MUFAs; OR: 0.82; 95% CI: 0.70, 0.95; P-trend = 0.003 for PUFAs).
TABLE 3.
Multivariable-adjusted OR and 95% CI for cognitive impairment according to quartiles of fat intake: the Singapore Chinese Health Study
| Quartiles of fat intake | |||||
|---|---|---|---|---|---|
| Q1 | Q2 | Q3 | Q4 | P-trend1 | |
| Total fat | |||||
| Cases/n | 674/4184 | 676/4184 | 558/4184 | 489/4184 | |
| Median,2 g/d | 33.9 (26.7–37.1) | 42.2 (40.6–43.6) | 47.7 (46.3–49.2) | 55.3 (52.7–59.4) | |
| Model 13 | 1.00 | 0.97 (0.86–1.10) | 0.83 (0.73–0.94) | 0.78 (0.68–0.89) | <0.001 |
| Model 24 | 1.00 | 1.01 (0.88–1.16) | 0.86 (0.74–1.00) | 0.80 (0.67–0.94) | 0.003 |
| SFAs | |||||
| Cases/n | 646/4184 | 645/4184 | 588/4184 | 518/4184 | |
| Median,2 g/d | 10.7 (8.64–12.0) | 14.4 (13.7–15.1) | 17.0 (16.3–17.7) | 20.7 (19.5–22.7) | |
| Model 13 | 1.00 | 1.03 (0.91–1.17) | 0.96 (0.84–1.09) | 0.93 (0.81–1.06) | 0.16 |
| Model 24 | 1.00 | 1.08 (0.94–1.25) | 1.05 (0.89–1.24) | 1.08 (0.89–1.31) | 0.51 |
| MUFAs | |||||
| Cases/n | 666/4184 | 681/4184 | 559/4184 | 491/4184 | |
| Median,2 g/d | 11.1 (8.83–12.1) | 14.1 (13.5–14.6) | 16.1 (15.6–16.7) | 19.0 (18.0–20.6) | |
| Model 13 | 1.00 | 1.01 (0.89–1.15) | 0.84 (0.73–0.95) | 0.78 (0.68–0.89) | <0.001 |
| Model 24 | 1.00 | 1.03 (0.88–1.19) | 0.85 (0.71–1.02) | 0.80 (0.64–0.99) | 0.02 |
| Model 35 | 1.00 | 1.08 (0.92–1.26) | 0.87 (0.71–1.05) | 0.77 (0.61–0.97) | 0.02 |
| PUFAs | |||||
| Cases/n | 663/4184 | 642/4184 | 553/4184 | 539/4184 | |
| Median,2 g/d | 5.95 (4.63–6.53) | 7.75 (7.37–8.10) | 9.35 (8.90–9.92) | 12.9 (11.6–14.9) | |
| Model 13 | 1.00 | 0.91 (0.80–1.04) | 0.76 (0.66–0.86) | 0.74 (0.65–0.85) | <0.001 |
| Model 24 | 1.00 | 0.99 (0.86–1.14) | 0.86 (0.74–1.01) | 0.84 (0.72–0.99) | 0.02 |
| Model 35 | 1.00 | 0.99 (0.86–1.14) | 0.85 (0.73–0.99) | 0.82 (0.70–0.95) | 0.003 |
| Animal-based fat | |||||
| Cases/n | 639/4184 | 612/4184 | 611/4184 | 535/4184 | |
| Median,2 g/d | 8.66 (6.03–10.2) | 13.2 (12.3–14.0) | 16.5 (15.6–17.4) | 21.3 (19.7–24.2) | |
| Model 13 | 1.00 | 0.93 (0.82–1.06) | 0.98 (0.86–1.12) | 0.92 (0.81–1.05) | 0.31 |
| Model 24 | 1.00 | 0.98 (0.85–1.13) | 1.04 (0.89–1.22) | 0.96 (0.81–1.15) | 0.76 |
| Plant-based fat | |||||
| Cases/n | 662/4184 | 634/4184 | 583/4184 | 518/4184 | |
| Median,2 g/d | 21.0 (17.2–23.2) | 27.2 (26.0–28.3) | 31.5 (30.3–32.6) | 38.3 (35.9–42.3) | |
| Model 13 | 1.00 | 0.91 (0.80–1.04) | 0.86 (0.75–0.98) | 0.78 (0.68–0.89) | <0.001 |
| Model 24 | 1.00 | 0.97 (0.84–1.10) | 0.91 (0.79–1.05) | 0.84 (0.72–0.98) | 0.02 |
P-trend was estimated using median values of quartiles as a continuous variable in logistic models.
Median intakes are shown as median value (IQR).
Model 1 was adjusted for age at cognition assessment, sex, educational level, marital status, dialect groups, and total energy intake.
Model 2 was further adjusted for BMI, physical activity, smoking status, alcohol consumption, sleep duration, physician-diagnosed history of diabetes, hypertension, heart disease, stroke, and cancer, and dietary intake of protein, cholesterol, and fiber (quartiles). Different types of fat were included in the model simultaneously.
Model 3 further included carbohydrate intake but excluded SFA from the model.
As for subtypes of PUFAs (Table 4), a higher dietary intake of n–6 PUFAs was significantly related to a lower risk of cognitive impairment when replacing carbohydrate (model 2) and SFAs (model 3), and the corresponding OR (95% CI) comparing extreme quartiles was 0.83 (0.70, 0.98) and 0.80 (0.69, 0.94), respectively. As for n–3 PUFAs, neither marine nor plant-based n–3 PUFAs showed a significant association with cognitive impairment after mutual adjustment for other types of FAs (model 2 and model 3, all P-trend >0.05).
TABLE 4.
Multivariable-adjusted OR and 95% CI for cognitive impairment with PUFAs: the Singapore Chinese Health Study
| Quartiles of PUFA intakes | |||||
|---|---|---|---|---|---|
| Q1 | Q2 | Q3 | Q4 | P-trend1 | |
| n–3 PUFAs | |||||
| Cases/n | 661/4184 | 588/4184 | 603/4184 | 545/4184 | |
| Median,2 g/d | 0.637 (0.539–0.695) | 0.802 (0.771–0.832) | 0.931 (0.897–0.968) | 1.17 (1.07–1.38) | |
| Model 13 | 1.00 | 0.84 (0.74–0.95) | 0.87 (0.76–0.99) | 0.78 (0.69–0.89) | <0.001 |
| Model 24 | 1.00 | 0.90 (0.78–1.03) | 0.97 (0.83–1.14) | 0.92 (0.77–1.09) | 0.49 |
| Model 35 | 1.00 | 0.91 (0.79–1.05) | 0.98 (0.84–1.15) | 0.91 (0.76–1.09) | 0.41 |
| Marine n–3 PUFAs | |||||
| Cases/n | 644/4184 | 593/4184 | 568/4184 | 592/4184 | |
| Median,2 g/d | 0.166 (0.118–0.197) | 0.268 (0.245–0.289) | 0.354 (0.331–0.378) | 0.491 (0.442–0.574) | |
| Model 13 | 1.00 | 0.90 (0.79–1.02) | 0.87 (0.76–0.99) | 0.94 (0.83–1.06) | 0.33 |
| Model 24 | 1.00 | 0.91 (0.79–1.04) | 0.89 (0.76–1.03) | 0.97 (0.82–1.14) | 0.77 |
| Model 35 | 1.00 | 0.91 (0.79–1.04) | 0.88 (0.76–1.03) | 0.96 (0.81–1.14) | 0.70 |
| α-linolenic acid | |||||
| Cases/n | 653/4184 | 623/4184 | 571/4184 | 550/4184 | |
| Median,2 g/d | 0.390 (0.319–0.425) | 0.497 (0.476–0.516) | 0.579 (0.557–0.604) | 0.746 (0.676–0.964) | |
| Model 13 | 1.00 | 0.90 (0.80–1.03) | 0.81 (0.71–0.92) | 0.78 (0.68–0.89) | <0.001 |
| Model 24 | 1.00 | 0.98 (0.85–1.12) | 0.92 (0.79–1.07) | 0.95 (0.80–1.12) | 0.52 |
| Model 35 | 1.00 | 0.99 (0.86–1.13) | 0.93 (0.80–1.08) | 0.94 (0.80–1.11) | 0.45 |
| n–6 PUFAs | |||||
| Cases/n | 666/4184 | 636/4184 | 566/4184 | 529/4184 | |
| Median,2 g/d | 5.19 (4.00–5.71) | 6.84 (6.48–7.16) | 8.36 (7.91–8.90) | 11.7 (10.5–13.6) | |
| Model 13 | 1.00 | 0.92 (0.81–1.04) | 0.77 (0.68–0.88) | 0.72 (0.63–0.82) | <0.001 |
| Model 24 | 1.00 | 1.01 (0.88–1.16) | 0.89 (0.77–1.04) | 0.83 (0.70–0.98) | 0.01 |
| Model 35 | 1.00 | 1.00 (0.87–1.15) | 0.88 (0.76–1.03) | 0.80 (0.69–0.94) | 0.001 |
P-trend was estimated using median values of quartiles as a continuous variable in logistic models.
Median intakes are shown as median value (IQR).
Model 1 was adjusted for age at cognition assessment, sex, educational level, marital status, dialect groups, and total energy intake.
Model 2 was further adjusted for BMI, physical activity, smoking status, alcohol consumption, sleep duration, physician-diagnosed history of diabetes, hypertension, heart disease, stroke and cancer, and dietary intake of protein, cholesterol, and fiber (quartiles). Different types of fat were included in the model simultaneously.
Model 3 further included carbohydrate intake but excluded SFA intake from the model.
The results remained robust in the sensitivity analyses when using percentages of energy from dietary fat intake as both categorical and continuous variables (Supplemental Table 1) and restricting participants to those without baseline cancer or cardiovascular disease (Supplemental Table 2). When we used <24 as the cut-off point for SM-MMSE to define cognitive impairment, the results remained unchanged except for MUFAs (Supplemental Table 3). In this sensitivity analysis, an inverse but nonsignificant association was observed for MUFAs (OR comparing extreme quartiles: 0.86; 95% CI: 0.71, 1.04; P-trend = 0.11). When we analyzed the MMSE score as a continuous variable, higher intakes of total fat, MUFAs, n–6 PUFAs, or plant-based fat were related to higher MMSE scores (all P-trend <0.05). Conversely, a higher intake of SFAs was associated with lower MMSE scores (P-trend = 0.01; Supplemental Table 4). Dietary fat intakes at baseline were not substantially different comparing individuals who participated or not in the third follow-up visit (Supplemental Table 5). When the log-transformed nutrients were used to calculate residuals, the results were robust except for MUFAs and n–3 PUFAs. In this sensitivity analysis, the inverse association with MUFAs was attenuated and became nonsignificant (OR comparing extreme quartiles in model 2: 0.83; 95% CI: 0.66, 1.04; P-trend = 0.09); whereas a significant inverse relation was found for n–3 PUFAs comparing extreme quartiles (OR: 0.82; 95% CI: 0.69–0.99; P-trend = 0.07; Supplemental Table 6).
In the exploratory analyses, we did not find any significant interactions between fat intake (including total fat, SFAs, MUFAs, and PUFAs) and potential effect modifiers including age, sex, education level, BMI level, smoking status, alcohol consumption, and history of chronic diseases (all P-interaction values >0.05; data not shown).
Discussion
In this cohort study of Chinese adults in Singapore with a mean age of 53.5 y (range 45–74) at baseline, we found that higher dietary intakes of total fat during midlife, in particular, fat from plant-based foods, were associated with a lower risk of cognitive impairment after 20 y of follow-up. In addition, dietary intakes of MUFAs and PUFAs, specifically n–6 PUFAs, were related to a lower risk of cognitive impairment when replacing total carbohydrate or SFAs. No significant relations were found for SFAs and n–3 PUFAs.
Emerging epidemiologic studies have investigated the association between dietary fat and risk of cognitive impairment, but the results remain inconsistent. Most cohort studies in Western populations indicated a nonsignificant association between total fat intake and the risk of cognitive decline (4, 5) and dementia (9, 10, 30). In the Rotterdam study, fat intake provided an estimated 36.7% of energy among the participants, and the study reported a positive association between total fat intake and dementia risk in a short-term follow-up (2.1 y) (8), but not in a long-term follow-up (6 y) study (9). The Cardiovascular Risk Factors, Aging and Dementia study followed 1341 middle-aged participants for ∼21 y and found a positive association between dietary total fat and mild cognitive impairment (7), but not for dementia or AD (30). In contrast to these studies, our study among Chinese adults indicated a lower risk of cognitive impairment with higher dietary fat intake. One of the plausible reasons is that the consumption level was relatively moderate in our study participants compared with those reported by Western populations. Total fat accounted for 25.7% of energy intake on average in our study, which is lower than the percentage reported in Western countries (31). The moderate amount of fat intake in this Chinese cohort may be in the optimum range, comparable with the range in the Mediterranean and Dietary Approaches to Stop Hypertension (DASH) diets, leading to better physiological functioning (32). Second, food sources of fat varied across different populations. Unlike Western populations, the main food sources of fat among our participants was plant-based fat, which comprised 66.5% of total fat. In particular, cooking oils were mostly plant-based oils. This is consistent with the study by Morris et al. (22), in which 815 community residents aged ≥65 y in the USA were followed for a mean of 3.9 y, and an inverse relation between vegetable fat intake and risk of AD was observed.
Different types of FAs may have variable associations with cognitive function, and results on this topic are also inconclusive. We found that MUFAs and PUFAs were inversely associated with cognitive impairment whereas SFAs showed no significant association. Previous studies investigating the association between SFAs and cognitive function have reported both positive (5–7) and nonsignificant (4, 9–14) associations. As for unsaturated FAs, some studies showed better cognition and less cognitive decline with higher intakes of MUFAs (6, 11, 14, 15) and PUFAs (14, 16), whereas others showed no significant associations for MUFAs (4, 5, 7, 9, 10, 12, 13, 22) or PUFAs (6, 9, 10). The inconsistency could be due to several reasons. First, many previous studies did not adjust for other FAs in the models (6, 8–14, 16), thus residual confounding was possible, particularly when different types of FAs correlated with each other because of overlap in food sources (2). Second, the follow-up duration varied substantially in those studies ranging from 2.1 to 5 y (4, 6, 8, 10, 11, 13, 16, 30), and since cognitive impairment requires many years to develop, the effect of dietary fat intake on cognition might not be detected in studies of a relatively short period. Third, some studies focused on the elderly when the process of cognitive decline cannot be easily reversed by dietary factors (5, 10, 13, 33), and very few studies have examined the association between dietary fat intake in midlife and risk of cognitive impairment in old age (4, 7). The Cardiovascular Risk Factors, Aging and Dementia study assessed dietary fat intake in midlife and cognition in old age with an average of 21 y of follow-up, and the authors reported a positive association between SFAs intake and mild cognitive impairment but no significant associations for MUFAs and PUFAs (7). However, this study only collected dietary fat intakes from milk products and spreads; therefore, the results cannot be compared with ours. The Doetinchem Cohort Study followed 2612 Dutch adults aged 43–70 y at baseline for 5 y and did not find significant relations of SFAs, MUFAs, and PUFAs with cognitive decline. However, this study only had a relatively short follow-up of 5 y (4).
Mechanistic studies have suggested that long-chain n–3 PUFAs like EPA and DHA may provide neuroprotective effects via anti-inflammation, antioxidation, and playing a role in neuron membrane (mainly DHA) function (34–37). Although some cohort studies have reported that n–3 PUFAs intake was associated with lower risks of cognitive decline and dementia (4, 16, 17, 38), null results were also reported in other studies (9, 18, 19). A meta-analysis of 3 cohort studies in the USA and Europe found no significant association between long-chain n–3 PUFAs (DHA and/or EPA) and the risk of dementia or AD (39). Another meta-analysis of 4 cohort studies among Western populations reported a significant inverse relation of DHA intake with AD (OR: 0.63; 95% CI: 0.51, 0.76; n = 3) and dementia (OR: 0.86; 95% CI: 0.76, 0.96; n = 2), but no significant association for the intake of EPA or total n–3 PUFAs (40). However, both meta-analyses included very limited studies. A meta-analysis of 12 RCTs failed to show a significant beneficial effect of 0.9–1.8 g/d EPA/DHA supplementation on cognition function (21). Some studies reported that the association between n–3 PUFAs and cognitive health may depend on the apoE genotype (33, 41, 42), however, we could not test this hypothesis in our study and further investigations are still needed.
We observed that a higher dietary n–6 PUFAs intake was associated with a lower risk of cognitive impairment. To date, only 4 cohort studies have reported the relation of dietary n–6 PUFAs intake and cognitive function (9, 17, 20, 22). Three short follow-up studies (3.9–6 y) showed nonsignificant associations between n–6 PUFAs and the risk of AD or dementia (9, 17, 22). Consistent with our study, the SU.VI.MAX study among 3362 French adults (mean age 65 y at baseline) found better cognitive function with a higher n–6 PUFAs intake after 13 y of follow-up (20). Previous studies have demonstrated the role of n–6 PUFAs in regulating lipid profiles and prostaglandin metabolism (43), and their subsequent protective effect on vascular disease and hypertension (44), which are both risk factors for cognitive impairment and dementia (45, 46). Further studies are still needed to explore the potential mechanisms.
Our study has the following strengths: a large sample size, long follow-up period, comprehensive measurements of dietary intake (165-item FFQ that included commonly consumed food in Singapore), and high-quality assessment of cognitive function with SM-MMSE which was designed specifically for this population. To our knowledge, this is the first large-scale prospective study investigating the association between dietary fat intake in midlife and the risk of cognitive impairment in late-life in the Chinese population. We simultaneously adjusted for different FAs in the models to reduce the residual confounding given that different FAs correlated with each other, and also applied the isocaloric substitution method to provide more realistic estimates for the impact of fat on cognitive health.
Our study also has several limitations. First, reverse causation is possible. We only evaluated participants’ cognitive function once at the third follow-up visit so we could not exclude people with cognitive impairment at recruitment. Second, we only assessed dietary intake at 1 time point at baseline, and subsequent changes in diet during follow-up may lead to nondifferential misclassification that could potentially bias our results towards null. Third, SM-MMSE is only a screening test and further studies using clinical diagnosis of mild cognitive impairment and dementia as outcomes are still needed. Fourth, although we have adjusted for various nutrients in the models, which are standardized in the studies on the relations of fat and FAs with health outcomes (29), the models may be overadjusted because the nutrients may come from the same food sources. Finally, selection bias is possible because a large proportion of participants were not included in the third follow-up visit. As shown in our previous publication (47), those who attended the third follow-up visit were younger and generally had healthier lifestyles and less comorbidities at baseline compared with those who did not attend the visit. However, dietary fat intakes were not substantially different across groups, and the selection bias might underestimate our results because of the nondifferential loss-to-follow-up.
In summary, we found that a higher intake of total fat, particularly the plant-based fat, in midlife, was associated with a lower risk of cognitive impairment in later life in Chinese adults. Dietary intakes of MUFAs and PUFAs, specifically n–6 PUFAs, were related to a lower risk of cognitive impairment.
Supplementary Material
Acknowledgments
We thank Siew-Hong Low of the National University of Singapore for supervising the fieldwork of the Singapore Chinese Health Study and Renwei Wang for the maintenance of the cohort study database. Finally, we acknowledge the founding Principal Investigator of the Singapore Chinese Health Study, Mimi C Yu.
The authors’ contributions were as follows—Y-WJ: conceived the study, analyzed data, drafted the initial manuscript, critically revised and edited the final manuscript; L-TS: checked the statistical code and the final data in the manuscript, reviewed and contributed to discussions about the results; X-FP: reviewed and edited the final manuscript; LF: contributed to data collection of MMSE interviews, reviewed and edited the final manuscript; J-MY: supervised data collection, reviewed and edited the final manuscript; W-PK and AP: supervised data collection, conceived the study, contributed to discussions about the results, critically revised and edited the final manuscript, were the guarantors of the work, had full access to all of the data in the study, and took primary responsibility for final content; and all authors read and approved the final manuscript.
Notes
This work was supported by the National Medical Research Council, Singapore (NMRC/CSA/0055/2013), National Institutes of Health, USA (grant numbers UM1 CA182876 and R01 CA144034), and the Saw Swee Hock School of Public Health, National University of Singapore. AP is supported by the National Key Research and Development Program of China (2017YFC0907504) and Hubei Province Science Fund for Distinguished Young Scholars (2018CFA033).
Author disclosures: The authors report no conflicts of interest.
Supplemental Tables 1–6 and Supplemental Figure 1 are available from the “Supplementary data” link in the online posting of the article and from the same link in the online table of contents at https://academic.oup.com/jn.
Abbreviations used: AD, Alzheimer's disease; FA, fatty acid; MMSE, Mini-Mental State Examination; RCT, randomized clinical trial; SM-MMSE, Singapore modified version of the Mini-Mental State Examination.
References
- 1. Prince M. World Alzheimer Report 2015: The Global Impact of Dementia: an Analysis of Prevalence, Incidence, Cost and Trends. London: Alzheimer's Disease International (ADI); 2015. [Google Scholar]
- 2. Morris MC, Tangney CC. Dietary fat composition and dementia risk. Neurobiol Aging. 2014;35:S59–64. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Solfrizzi V, Frisardi V, Capurso C, D'Introno A, Colacicco AM, Vendemiale G, Capurso A, Panza F. Dietary fatty acids in dementia and predementia syndromes: epidemiological evidence and possible underlying mechanisms. Ageing Res Rev. 2010;9(2):184–99. [DOI] [PubMed] [Google Scholar]
- 4. Nooyens ACJ, van Gelder BM, Bueno-de-Mesquita HB, van Boxtel MPJ, Verschuren WMM. Fish consumption, intake of fats and cognitive decline at middle and older age: the Doetinchem Cohort Study. Eur J Nutr. 2018;57(4):1667–75. [DOI] [PubMed] [Google Scholar]
- 5. Morris MC, Evans DA, Bienias JL, Tangney CC, Wilson RS. Dietary fat intake and 6-year cognitive change in an older biracial community population. Neurology. 2004;62(9):1573–9. [DOI] [PubMed] [Google Scholar]
- 6. Okereke OI, Rosner BA, Kim DH, Kang JH, Cook NR, Manson JE, Buring JE, Willett WC, Grodstein F. Dietary fat types and 4-year cognitive change in community-dwelling older women. Ann Neurol. 2012;72(1):124–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Eskelinen MH, Ngandu T, Helkala EL, Tuomilehto J, Nissinen A, Soininen H, Kivipelto M. Fat intake at midlife and cognitive impairment later in life: a population-based CAIDE study. Int J Geriatr Psychiatry. 2008;23(7):741–7. [DOI] [PubMed] [Google Scholar]
- 8. Kalmijn S, Launer LJ, Ott A, Witteman JC, Hofman A, Breteler MM. Dietary fat intake and the risk of incident dementia in the Rotterdam Study. Ann Neurol. 1997;42(5):776–82. [DOI] [PubMed] [Google Scholar]
- 9. Engelhart MJ, Geerlings MI, Ruitenberg A, Van Swieten JC, Hofman A, Witteman JC, Breteler MM. Diet and risk of dementia: does fat matter?: The Rotterdam Study. Neurology. 2002;59(12):1915–21. [DOI] [PubMed] [Google Scholar]
- 10. Roberts RO, Roberts LA, Geda YE, Cha RH, Pankratz VS, O'Connor HM, Knopman DS, Petersen RC. Relative intake of macronutrients impacts risk of mild cognitive impairment or dementia. J Alzheimers Dis. 2012;32(2):329–39. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Naqvi AZ, Harty B, Mukamal KJ, Stoddard AM, Vitolins M, Dunn JE. Monounsaturated, trans, and saturated fatty acids and cognitive decline in women. J Am Geriatr Soc. 2011;59(5):837–43. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Psaltopoulou T, Kyrozis A, Stathopoulos P, Trichopoulos D, Vassilopoulos D, Trichopoulou A. Diet, physical activity and cognitive impairment among elders: the EPIC-Greece cohort (European Prospective Investigation into Cancer and Nutrition). Public Health Nutr. 2008;11(10):1054–62. [DOI] [PubMed] [Google Scholar]
- 13. Solfrizzi V, Colacicco AM, D'Introno A, Capurso C, Del Parigi A, Capurso SA, Argentieri G, Capurso A, Panza F. Dietary fatty acids intakes and rate of mild cognitive impairment. The Italian Longitudinal Study on Aging. Exp Gerontol. 2006;41(6):619–27. [DOI] [PubMed] [Google Scholar]
- 14. Solfrizzi V, Colacicco AM, D'Introno A, Capurso C, Torres F, Rizzo C, Capurso A, Panza F. Dietary intake of unsaturated fatty acids and age-related cognitive decline: a 8.5-year follow-up of the Italian Longitudinal Study on Aging. Neurobiol Aging. 2006;27(11):1694–704. [DOI] [PubMed] [Google Scholar]
- 15. Solfrizzi V, Panza F, Torres F, Mastroianni F, Del Parigi A, Venezia A, Capurso A. High monounsaturated fatty acids intake protects against age-related cognitive decline. Neurology. 1999;52(8):1563–9. [DOI] [PubMed] [Google Scholar]
- 16. van Gelder BM, Tijhuis M, Kalmijn S, Kromhout D. Fish consumption, n-3 fatty acids, and subsequent 5-y cognitive decline in elderly men: the Zutphen Elderly Study. Am J Clin Nutr. 2007;85(4):1142–7. [DOI] [PubMed] [Google Scholar]
- 17. Laurin D, Verreault R, Lindsay J, Dewailly E, Holub BJ. Omega-3 fatty acids and risk of cognitive impairment and dementia. J Alzheimers Dis. 2003;5(4):315–22. [DOI] [PubMed] [Google Scholar]
- 18. Devore EE, Grodstein F, van Rooij FJ, Hofman A, Rosner B, Stampfer MJ, Witteman JC, Breteler MM. Dietary intake of fish and omega-3 fatty acids in relation to long-term dementia risk. Am J Clin Nutr. 2009;90(1):170–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Morris MC, Evans DA, Tangney CC, Bienias JL, Wilson RS. Fish consumption and cognitive decline with age in a large community study. Arch Neurol. 2005;62(12):1849–53. [DOI] [PubMed] [Google Scholar]
- 20. Assmann KE, Adjibade M, Hercberg S, Galan P, Kesse-Guyot E. Unsaturated fatty acid intakes during midlife are positively associated with later cognitive function in older adults with modulating effects of antioxidant supplementation. J Nutr. 2018;148(12):1938–45. [DOI] [PubMed] [Google Scholar]
- 21. Jiao J, Li Q, Chu J, Zeng W, Yang M, Zhu S. Effect of n-3 PUFA supplementation on cognitive function throughout the life span from infancy to old age: a systematic review and meta-analysis of randomized controlled trials. Am J Clin Nutr. 2014;100(6):1422–36. [DOI] [PubMed] [Google Scholar]
- 22. Morris MC, Evans DA, Bienias JL, Tangney CC, Bennett DA, Aggarwal N, Schneider J, Wilson RS. Dietary fats and the risk of incident Alzheimer disease. Arch Neurol. 2003;60(2):194–200. [DOI] [PubMed] [Google Scholar]
- 23. Shen X, Fang A, He J, Liu Z, Guo M, Gao R, Li K. Trends in dietary fat and fatty acid intakes and related food sources among Chinese adults: a longitudinal study from the China Health and Nutrition Survey (1997–2011). Public Health Nutr. 2017;20(16):2927–36. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Liu AG, Ford NA, Hu FB, Zelman KM, Mozaffarian D, Kris-Etherton PM. A healthy approach to dietary fats: understanding the science and taking action to reduce consumer confusion. Nutr J. 2017;16(1):53. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Hankin JH, Stram DO, Arakawa K, Park S, Low SH, Lee HP, Yu MC. Singapore Chinese Health Study: development, validation, and calibration of the quantitative food frequency questionnaire. Nutr Cancer. 2001;39(2):187–95. [DOI] [PubMed] [Google Scholar]
- 26. Feng L, Chong MS, Lim WS, Ng TP. The Modified Mini-Mental State Examination test: normative data for Singapore Chinese older adults and its performance in detecting early cognitive impairment. Singapore Med J. 2012;53(7):458–62. [PubMed] [Google Scholar]
- 27. Folstein MF, Folstein SE, McHugh PR. "Mini-mental state". A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res. 1975;12(3):189–98. [DOI] [PubMed] [Google Scholar]
- 28. Katzman R, Zhang MY, Ouang Ya Q, Wang ZY, Liu WT, Yu E, Wong SC, Salmon DP, Grant I. A Chinese version of the Mini-Mental State Examination; impact of illiteracy in a Shanghai dementia survey. J Clin Epidemiol. 1988;41(10):971–8. [DOI] [PubMed] [Google Scholar]
- 29. Willett W, Stampfer MJ. Nutritional Epidemiology. 3rd ed New York, NY: Oxford University Press; 2013. [Google Scholar]
- 30. Laitinen MH, Ngandu T, Rovio S, Helkala EL, Uusitalo U, Viitanen M, Nissinen A, Tuomilehto J, Soininen H, Kivipelto M. Fat intake at midlife and risk of dementia and Alzheimer's disease: a population-based study. Dement Geriatr Cogn Disord. 2006;22(1):99–107. [DOI] [PubMed] [Google Scholar]
- 31. Dehghan M, Mente A, Zhang X, Swaminathan S, Li W, Mohan V, Iqbal R, Kumar R, Wentzel-Viljoen E, Rosengren A et al.. Associations of fats and carbohydrate intake with cardiovascular disease and mortality in 18 countries from five continents (PURE): a prospective cohort study. Lancet. 2017;390(10107):2050–62. [DOI] [PubMed] [Google Scholar]
- 32. Morris MC. Nutrition and risk of dementia: overview and methodological issues. Ann N Y Acad Sci. 2016;1367(1):31–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. van de Rest O, Wang Y, Barnes LL, Tangney C, Bennett DA, Morris MC. APOE epsilon4 and the associations of seafood and long-chain omega-3 fatty acids with cognitive decline. Neurology. 2016;86(22):2063–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Dyall SC. Long-chain omega-3 fatty acids and the brain: a review of the independent and shared effects of EPA, DPA and DHA. Front Aging Neurosci. 2015;7:52. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Wu YQ, Dang RL, Tang MM, Cai HL, Li HD, Liao DH, He X, Cao LJ, Xue Y, Jiang P. Long chain omega-3 polyunsaturated fatty acid supplementation alleviates doxorubicin-induced depressive-like behaviors and neurotoxicity in rats: involvement of oxidative stress and neuroinflammation. Nutrients. 2016;8(4):243. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Calviello G, Su HM, Weylandt KH, Fasano E, Serini S, Cittadini A. Experimental evidence of omega-3 polyunsaturated fatty acid modulation of inflammatory cytokines and bioactive lipid mediators: their potential role in inflammatory, neurodegenerative, and neoplastic diseases. BioMed Res Int. 2013;2013:743171. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Yehuda S, Rabinovitz S, Carasso RL, Mostofsky DI. The role of polyunsaturated fatty acids in restoring the aging neuronal membrane. Neurobiol Aging. 2002;23(5):843–53. [DOI] [PubMed] [Google Scholar]
- 38. Morris MC, Evans DA, Bienias JL, Tangney CC, Bennett DA, Wilson RS, Aggarwal N, Schneider J. Consumption of fish and n-3 fatty acids and risk of incident Alzheimer disease. Arch Neurol. 2003;60(7):940–6. [DOI] [PubMed] [Google Scholar]
- 39. Wu S, Ding Y, Wu F, Li R, Hou J, Mao P. Omega-3 fatty acids intake and risks of dementia and Alzheimer's disease: a meta-analysis. Neurosci Biobehav Rev. 2015;48:1–9. [DOI] [PubMed] [Google Scholar]
- 40. Zhang Y, Chen J, Qiu J, Li Y, Wang J, Jiao J. Intakes of fish and polyunsaturated fatty acids and mild-to-severe cognitive impairment risks: a dose-response meta-analysis of 21 cohort studies. Am J Clin Nutr. 2016;103(2):330–40. [DOI] [PubMed] [Google Scholar]
- 41. Yassine HN, Braskie MN, Mack WJ, Castor KJ, Fonteh AN, Schneider LS, Harrington MG, Chui HC. Association of docosahexaenoic acid supplementation with Alzheimer disease stage in apolipoprotein E epsilon4 carriers: a review. JAMA Neurol. 2017;74(3):339–47. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Barberger-Gateau P, Samieri C, Feart C, Plourde M. Dietary omega-3 polyunsaturated fatty acids and Alzheimer's disease: interaction with apolipoprotein E genotype. Curr Alzheimer Res. 2011;8(5):479–91. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43. Wang DD. Dietary n-6 polyunsaturated fatty acids and cardiovascular disease: epidemiologic evidence. Prostaglandins Leukot Essent Fatty Acids. 2018;135:5–9. [DOI] [PubMed] [Google Scholar]
- 44. Nakamura H, Hara A, Tsujiguchi H, Thi Thu Nguyen T. Relationship between dietary n-6 fatty acid intake and hypertension: effect of glycated hemoglobin levels. Nutrients. 2018;10(12):1825. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45. Deckers K, Schievink SHJ, Rodriquez MMF, van Oostenbrugge RJ, van Boxtel MPJ, Verhey FRJ, Kohler S. Coronary heart disease and risk for cognitive impairment or dementia: systematic review and meta-analysis. PLoS One. 2017;12(9):e0184244. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46. Perrotta M, Lembo G, Carnevale D. Hypertension and dementia: epidemiological and experimental evidence revealing a detrimental relationship. Inter J Mol Sci. 2016;17(3):347. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47. Wu J, Song XY, Chen GC, Neelakantan N, van Dam RM, Feng L, Yuan JM, Pan A, Koh WP. Dietary pattern in midlife and cognitive impairment in late life: a prospective study in Chinese adults. Am J Clin Nutr. 2019;110(4):912–20. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.