Abstract
Objectives
There is increasing attention for dietary patterns as a potential strategy to prevent cognitive decline. We examined the association between adherence to a recently developed Mediterranean-DASH Intervention for Neurodegenerative Delay (MIND) diet with cognitive function and cognitive decline, taking into account the interaction between the apolipoprotein E ε4 genotype and the MIND diet.
Design
Population-based prospective cohort study.
Participants
A total of 16,058 older women aged 70 and over from the Nurses' Health Study.
Measurements
Dietary intake was assessed five times between 1984 and 1998 with a 116-item Food Frequency Questionnaire. The MIND score includes ten brain-healthy foods and five unhealthy foods. Cognition was assessed four times by telephone from 1995 to 2001 (baseline) with the Telephone Interview for Cognitive Status (TICS) and by calculating composite scores of verbal memory and global cognition. Linear regression modelling and linear mixed modelling were used to examine the associations of adherence to the MIND diet with average cognitive function and cognitive change over six years, respectively.
Results
Greater long-term adherence to the MIND diet was associated with a better verbal memory score (multivariable-adjusted mean differences between extreme MIND quintiles=0.04 (95%CI 0.01-0.07), p-trend=0.006), but not with cognitive decline over 6 years in global cognition, verbal memory or TICS.
Conclusion
Long-term adherence to the MIND diet was moderately associated with better verbal memory in later life. Future studies should address this association within populations at greater risk of cognitive decline.
Key words: Cognition, dietary pattern, MIND, Mediterranean, DASH
Introduction
The proportion of the world's population over 60 years is expected to double from 11% in 2010 to 22% by 2050 (1). With population ageing, there will be an increase in the prevalence of age-related diseases and disabilities, such as dementia (2). As, to date, there is no effective treatment for dementia (3), it is important to identify disease-modifying risk factors to prevent cognitive decline. One of these factors could be nutrition. Especially studying dietary patterns as a potential strategy to prevent cognitive decline has increasingly received attention in recent years. The Mediterranean diet (MeDi) is a frequently studied dietary pattern which has been shown to be associated with a slower rate of cognitive decline (4). A less studied dietary pattern in relation to cognition is the Dietary Approaches to Stop Hypertension (DASH) diet. So far, three observational studies and one intervention study have demonstrated that higher adherence to the DASH diet was associated with less cognitive decline (5, 6, 7, 8). However, the MeDi and DASH diets do not include specific dietary components that optimize brain health.
Recently, a hybrid of these two dietary patterns has been developed by Morris et al. (9), namely the Mediterranean-DASH Intervention for Neurodegenerative Delay (MIND) diet. Similar to the MeDi and DASH diets, the MIND diet emphasizes natural plant-based foods and limited intakes of animal and high saturated fat foods. Uniquely, it also specifies the consumption of berries and green leafy vegetables (9), which have been demonstrated to have antioxidant and antiinflammatory properties and inhibit β-amyloid deposition and neurotoxic death (10). Greater adherence to the MIND diet was associated with slower rates of cognitive decline in the only study conducted to date (10).
Besides lifestyle factors, the apolipoprotein E ε4 (ApoE ε4) allele has been shown to be a major genetic risk factor for more rapid cognitive decline and earlier onset of Alzheimer's Disease (AD) (11, 12). The few human studies that have investigated an interaction between dietary patterns and cognitive function showed inconsistent findings (13, 14, 15, 16, 17). Overall, the contribution of the ApoE ε4 allele in the association between dietary patterns and cognitive function is unclear.
We evaluated the association between long-term adherence to the MIND diet with cognitive function and cognitive decline during 6 years in 16,058 women from the Nurses' Health Study. Additionally, we examined the interaction between the ApoE ε4 genotype and the MIND diet in relation to cognitive function.
Methods
Study population
The Nurses' Health Study (NHS) began in 1976, with 121,700 female registered nurses aged 30-55 years and resided in 11 U.S. states. Participants completed a mailed questionnaire about their health and lifestyle. Follow-up questionnaires are sent every two years. To date, follow-up of the cohort remains ~90%. In 1984, participants completed a 116-item semi- quantitative food frequency questionnaire (FFQ) and similar FFQs were sent in 1986 and every 4 years thereafter (18). During 1995-2001, women who were >70 years old and free of stroke were invited to participate in the first telephone-based study of cognitive function. A total of 19,415 eligible women completed the first interview (93%). Follow-up assessments were performed three times at two-year intervals until 2008, with a participation rate of > 90% among those who remained alive and could be contacted at each follow-up point. The study was approved by the Institutional Review Board of Brigham and Women's Hospital (Boston, MA) and all participants consented.
In the present study, we excluded those who had not completed at least one of the two initial FFQs in 1984 and 1986 (n=2,990), without at least one complete cognition battery at follow-up (n=331), or had no data for physical activity (n=36), resulting in a total population for analysis of 16,058 participants.
Dietary assessment and MIND diet score
For each item of the FFQ, possible responses ranged from ‘never or <1 time/month' to ‘>6 times/day', and specific standardized portion sizes. Food intake estimations were converted into nutrient intakes by multiplying the frequency of consumption of each food by its nutrient content by using the U.S. Department of Agriculture database and other sources of updated supplementary information.
In this study, we used the 1984, 1986, 1990, 1994 and 1998 FFQs to estimate daily energy intake and to construct a MIND diet score at each of these time points using the methodology described by Morris et al. (9). The MIND score is based on 10 brain-healthy foods groups: namely, higher intake of green leafy vegetables (spinach, lettuce, kale), other vegetables, berries (blueberries, strawberries), nuts, whole grains, fish, beans, poultry, limited intake of wine (red and white wine), and use of olive oil as primary source of fat; and 5 unhealthy foods: namely, lower intakes of butter and margarine, cheese, red meat and products, fast fried foods, and pastries and sweets (9) (Table 1). For healthy components, we assigned 0, 0.5 or 1 point for higher intakes, and for unhealthy components the scoring was reversed (total score 0 to 15 points). As long- term dietary habits are likely most relevant because cognitive decline develops over many years (19), long-term MIND score was computed as the mean of up to five MIND scores from all dietary assessments from 1984 (or 1986, if 1984 FFQ was missing) to the questionnaire immediately preceding the first cognitive interview. Total MIND scores were divided into quintiles based on the study population distribution.
Table 1.
Components and scoring of the MIND diet within the Nurses’ Health Study
| MIND component | 0 | Score0.5 | 1 |
|---|---|---|---|
| Green leafy vegetables | ≤2 servings/week | >2 to <6/week | ≥6 servings/week |
| Other vegetables | <5 servings/week | 5 to <7/week | ≥1 serving/day |
| Berries | <1 serving/week | 1-2/week | ≥2 servings/week |
| Nuts | <1/month | 1/month to <5/week | ≥5 servings/week |
| Olive Oil | Not primary oil | Primary oil used | |
| Butter, margarine | >2 table spoon/day | 1-2 table spoon/day | <1 table spoon/day |
| Cheese | ≥7 servings/week | ≥1-<7/week | <1 serving/week |
| Whole grains | <1 serving/day | ≥1-<3/day | ≥3 servings/day |
| Fish (not fried) | Rarely | 1-3/month | ≥1 meals/week |
| Beans | <1 meal/week | 1-3/week | >3 meals/week |
| Poultry (not fried) | <1 meal/week | ≥1-<2/week | ≥2 meals/week |
| Red meat and products | >6 meals/week | ≥4-≤6/week | <4 meals/week |
| Fast fried foods | ≥4 times/week | 1-<4/week | <1 time/week |
| Pastries and sweets | ≥7 servings/week | ≥5-<7/week | <5 servings/week |
| Wine | >1 glass/day or never | 1/month to 6/week | 1 glass/day |
| Total score | 0 | 7.5 | 15 |
Cognitive assessment
Cognitive testing was performed by trained interviewers using validated telephone interviews. The cognitive battery included: 1) the Telephone Interview for Cognitive Status (TICS) (20); 2) immediate; and 3) delayed recalls of the East Boston Memory test (EBMT) (21); 4) delayed recall of the TICS 10-word list; 5) category fluency; and 6) digit spanbackward test. The sample size slightly differs across tests as in the initial stage of the first cognitive interview only the TICS was administered and gradually five other tests were added. The participation rate remained identical for all tests.
The TICS (20) (0–41 points) is a telephone adaptation of the Mini-Mental State Examination (22) and assesses overall cognitive performance. The EBMT (21) (0–12 points) and delayed recall of the TICS 10-word list (0–10 points) assesses verbal (episodic) memory. The category fluency test assesses language and executive function (23); performance is based on naming as many animals as possible in one minute. Finally, the digit-span backward test (0–12 points) assesses working memory and attention, where participants repeat backwards an increasingly long series of digits.
The primary outcomes were the TICS and composite scores of global cognition and verbal memory. A global cognitive score was computed as the mean of z-scores of all six cognitive tests. A verbal memory score was calculated as the mean of z-scores of four tests assessing verbal memory (i.e., immediate and delayed recalls of both the EBMT and the TICS 10-word list). We calculated z-scores at each follow-up using means and standard deviations of scores at the first cognitive assessment.
Other variables
Socio-demographic, lifestyle and health-related information were obtained from questionnaires. Using cumulative and updated information from 1976, covariates were determined at the time of the first cognitive exam. For physical activity and energy intake values were averaged across multiple assessments over time, similar to diet. Physical activity was assessed in 1986, 1988 and 1992, and every 2 years thereafter by estimating mean energy expended per week (in metabolic equivalenthours, Met-hrs). In a randomly selected subsample (n=5,822), data on ApoE polymorphisms was available from cheek cell specimens (n=3,469) and genome wide association studies (n=2,353).
Statistical analyses
Long-term MIND adherence, cognitive function and decline
Statistical analyses consisted of two complementary approaches. As primary analyses, we averaged the four repeated measures of cognitive function to create an outcome representing overall cognitive status at older ages. Averaging repeated measures of cognition was relevant to our data, as it attenuates variability in each single cognitive assessment, which may be helpful when cognition is measured over a relatively short follow-up period in largely healthy, educated participants, as described in previous studies (24, 25). The association of long-term MIND score with cognitive status was modelled using linear regression models.
As secondary approach, we modelled trajectories of the four repeated cognitive scores using linear mixed models (26). The linear mixed models included an intercept representing the level of cognitive score at baseline and a slope representing the annual change in scores over time, as well as a random intercept and random slope to account for inter-individual variability.
Adjustments were made for confounding factors age and education (registered nurse, bachelor, master or doctorate) (Model 1), and additionally for long-term energy intake (quintiles) and physical activity (Met-hrs/week, quintiles), body mass index (BMI, ≤21, 22-24, 25-29, ≥30kg/m2), smoking status (never, former, current), alcohol intake (<1, 1-15, ≥15g/d), history of depression (yes/no), multivitamin use (yes/no), and cardiovascular risk factors (history of diabetes, hypertension, hypercholesterolemia, and/or myocardial infarction (yes/no) (Model 2). Because for BMI and multivitamin use data was missing for >4% of the sample (4.4 and 7.1%, respectively); a missing category was created. For other covariates, participants with missing data were <1% of the sample and were assigned to the reference group. Effect modification was tested by adding an interaction term for MIND score quintiles with age (median split), BMI (>25/≥25kg/m2), high blood pressure (yes/no), hypercholesterolemia (yes/no), myocardial infarction (yes/no). In a subset of 5,822 participants, a variable indicating the product of the number of ApoE ε4 alleles (0,1,2) was used to test for effect modification by ApoE ε4 genotype.
MIND-components and cognitive function
To examine the relative importance of the individual MIND components, we recalculated separate MIND scores by excluding one MIND component at a time while including this component as covariate (in quintiles of intake).
All statistical analyses were carried out using SAS software version 9.2 (SAS Institute Inc., Cary, NC, USA). A two-sided p-value of <0.05 was considered significant. We examined linear trends across quintiles of the MIND score using a continuous variable in which participants in a given category were assigned the median value.
Results
General characteristics individual MIND components across quintiles of the average MIND score of 4.4 dietary assessments over 12.9 years (Table 2). Women with greater long-term MIND adherence were more likely to have a lower BMI, to be more physically active, to be higher educated, to use a multivitamin supplement, to report a history of hypercholesterolemia and were less likely to be p ]At the first cognitive assessment, the mean age of participants was 74.3 ± 2.3 years (mean ± SD), and the mean TICS score was 33.8 ± 2.7 points. The median MIND score was 6.4 points (range 2.6–11.0) with a large variation in intake of current smokers and to report a history of depression.
Table 2.
Baseline characteristics of the participants in the Nurses’ Health Study by quintiles of long-term adherence to the MIND score (n=16,058) a
| Quintile 1 | Quintile 2 | Quintile 3 | Quintile 4 | Quintile 5 | |
|---|---|---|---|---|---|
| 4.9 (2.6-5.4) | 5.8 (5.5-6.1) | 6.4 (6.1-6.7) | 7.0 (6.7-7.3) | 7.9 (7.4-11.0) | |
| n=3,339 | n=3,259 | n=2,962 | n=3,261 | n=3,237 | |
| Age at first cognitive exam | 74.3±2.3 | 74.2±2.3 | 74.3±2.3 | 74.3±2.4 | 74.3±2.3 |
| BMI at first cognitive exam ≤21, kg/m2 | 709 (21.2%) | 612 (18.8%) | 561 (18.9%) | 602 (18.5%) | 686 (21.2%) |
| 22-24, kg/m2 | 780 (23.4%) | 771 (23.7%) | 829 (28.0%) | 866 (26.6%) | 860 (26.6%) |
| 25-29, kg/m2 | 1,138 (34.1%) | 1,090 (33.4%) | 955 (32.2%) | 1,127 (34.6%) | 1,041 (32.2%) |
| ≥30, kg/m2 | 592 (17.7%) | 647 (19.9%) | 488 (16.5%) | 526 (16.1%) | 476 (14.7%) |
| Long-term physical activity (mets) Educational level | 12.5±12.6 | 15.0±13.8 | 17.2±14.7 | 18.7±15.9 | 23.1±19.8 |
| Registered nurse | 2,785 (83.4%) | 2,589 (79.4%) | 2,325 (78.5%) | 2,445 (75.0%) | 2,304 (71.2%) |
| Bachelor | 403 (12.1%) | 505 (15.5%) | 459 (15.5%) | 588 (18.0%) | 657 (20.3%) |
| Master or doctorate | 151 (4.5%) | 165 (5.1%) | 178 (6.0%) | 228 (7.0%) | 276 (8.5%) |
| Smoking Never | 1,582 (47.4%) | 1,554 (47.7%) | 1,360 (45.9%) | 1,512 (46.4%) | 1,466 (45.3%) |
| Former | 1,376 (41.2%) | 1,408 (43.2%) | 1,377 (46.5%) | 1,552 (47.6%) | 1,627 (50.3%) |
| Current | 381 (11.4%) | 297 (9.1%) | 225 (7.6%) | 197 (6.0%) | 144 (4.4%) |
| History of: Myocardial infarction | 214 (6.4%) | 197 (6.0%) | 180 (6.1%) | 196 (6.0%) | 182 (5.6%) |
| Hypertension | 1,860 (55.7%) | 1,810 (55.5%) | 1,634 (55.2%) | 1,829 (56.1%) | 1,757 (54.3%) |
| Hypercholesterolemia | 2,107 (63.1%) | 2,122 (65.1%) | 1,942 (65.6%) | 2,179 (66.8%) | 2,152 (66.5%) |
| Diabetes | 331 (9.9%) | 336 (10.3%) | 329 (11.1%) | 326 (10.0%) | 289 (8.9%) |
| Depression | 346 (10.4%) | 315 (9.7%) | 273 (9.2%) | 305 (9.4%) | 271 (8.4%) |
| Multivitamin use | 1,736 (54.5%) | 1,888 (58.8%) | 1,925 (58.7%) | 1,983 (61.6%) | 2,026 (64.1%) |
| Long-term dietary variables (servings per day unless otherwise noted) | |||||
| Daily energy intake, kcal/day | 1,646±440 | 1,686±428 | 1,694±421 | 1,712±409 | 1,783±420 |
| Alcohol, g/day | 4.6±8.8 | 5.2±8.9 | 5.8±9.0 | 6.1±8.8 | 6.3±8.4 |
| MIND components Green leafy vegetablesb | 3.4±5.2 | 5.0±6.7 | 6.5±8.6 | 7.4±8.6 | 9.6±10.1 |
| Other vegetables | 2.0±1.8 | 2.6±2.3 | 3.1±2.9 | 3.4±2.9 | 4.2±3.7 |
| Berriesb | 0.8±2.7 | 1.3±3.5 | 1.6±4.1 | 1.8±4.0 | 2.4±4.7 |
| Nutsb | 1.5±2.6 | 2.1±3.9 | 2.6±4.7 | 2.6±4.8 | 3.4±6.1 |
| Primarily vegetable oil | 0.2±0.3 | 0.3±0.4 | 0.3±0.4 | 0.4±0.5 | 0.6±0.5 |
| Butter, margarine | 1.4±1.0 | 1.2±0.9 | 1.1±0.9 | 1.0±0.8 | 0.8±0.8 |
| Cheeseb | 5.9±7.9 | 5.4±7.2 | 5.5±7.6 | 5.3±7.3 | 5.2±7.6 |
| Whole grains | 1.0±1.2 | 1.4±1.6 | 1.7±2.1 | 1.8±2.1 | 2.4±2.6 |
| Fish (not fried)b | 1.1±2.8 | 1.8±3.8 | 2.3±4.8 | 2.6±5.0 | 3.3±5.2 |
| Beansb | 0.6±2.2 | 1.2±3.4 | 1.5±4.1 | 1.8±4.6 | 2.3±5.4 |
| Poultry (not fried)b | 2.3±3.8 | 3.0±4.7 | 3.5±4.8 | 3.7±4.8 | 4.3±5.2 |
| Red meat and productsb | 6.7±6.0 | 6.3±6.9 | 6.1±7.6 | 5.8±8.8 | 5.4±9.6 |
| Fast fried foodsb | 8.0±1.6 | 8.0±1.7 | 8.2±1.7 | 8.3±1.7 | 8.3±1.9 |
| Pastries and sweetsb | 15.1±15.0 | 14.3±17.8 | 14.2±21.5 | 13.0±22.2 | 13.0±27.4 |
| Wine | 0.1±0.4 | 0.2±0.6 | 0.3±0.6 | 0.3±0.6 | 0.3±0.7 |
Abbreviation: MIND=Mediterranean-DASH Intervention for Neurodegenerative Delay; a. Values are means ± SD or number (%); b. Servings per week
Average cognitive function
Greater long-term adherence to the MIND score was linearly related to significantly higher verbal memory scores, comprising the mean of the 4 repeated measures of cognitive function, in both the age and education adjusted model and the multivariable-adjusted model (p-trend < 0.0001 for model 1, 0.006 for model 2), but not with a better global cognition or TICS score (Table 3). The mean difference in mean verbal memory between the top and bottom quintiles of the MIND score in the fully adjusted model was 0.04 standard units.
Table 3.
Multivariable-adjusted mean differencesa in cognition averaged over the follow-up period across quintiles of the MIND score in 16,058 older women participating in the Nurses’ Health Study
| Model 1b | Model 2c | |
|---|---|---|
| Global cognitive score | ||
| Quintile 1 | Ref | Ref |
| Quintile 2 | 0.02 (-0.01,0.05) | 0.00 (-0.02,0.03) |
| Quintile 3 | 0.02 (-0.01,0.05) | 0.00 (-0.03,0.03) |
| Quintile 4 | 0.02 (-0.01,0.05) | 0.00 (-0.03,0.03) |
| Quintile 5 | 0.03 (0.00,0.06) | 0.00 (-0.03,0.03) |
| p-trend | 0.03 | 0.88 |
| Verbal memory score | ||
| Quintile 1 | Ref | Ref |
| Quintile 2 | 0.02 (-0.01,0.06) | 0.01 (-0.02,0.05) |
| Quintile 3 | 0.03 (0.00,0.07) | 0.02 (-0.01,0.05) |
| Quintile 4 | 0.04 (0.01,0.08) | 0.02 (-0.01,0.06) |
| Quintile 5 | 0.06 (0.03,0.09) | 0.04 (0.01,0.07) |
| p-trend | <0.0001 | 0.02 |
| TICS score | ||
| Quintile 1 | Ref | Ref |
| Quintile 2 | -0.01 (-0.13,0.11) | -0.05 (-0.18,0.07) |
| Quintile 3 | 0.06 (-0.07,0.19) | 0.00 (-0.13,0.12) |
| Quintile 4 | 0.07 (-0.05,0.19) | -0.02 (-0.14,0.11) |
| Quintile 5 | 0.02 (-0.10,0.14) | -0.09 (-0.21,0.04) |
| p-trend | 0.44 | 0.31 |
Abbreviations: MIND=Mediterranean-DASH Intervention for Neurodegenerative Delay; TICS=Telephone interview of cognitive status; a. Values are mean difference in score (95% confidence interval); b. Model 1: adjusted for age and education; c. Model 2: additionally adjusted physical activity, calorie intake, alcohol intake, smoking status, multivitamin use, BMI, depression, and history of high blood pressure, hypercholesterolemia, myocardial infarction, and diabetes mellitus.
The associations did not change when additionally adjusting for family history of dementia and socio- economic variables such as income, job, and marital status (data not shown). There was no interaction between adherence to the MIND diet and age, BMI, high blood pressure or ApoE e4 status for global cognition, verbal memory or TICS (p-interaction age: 0.45, 0.70, 0.82, BMI: 0.63, 0.61, 0.69, high blood pressure: 0.90, 0.70, 0.85, ApoE e4 status: 0.38, 0.40, 0.27 for global cognition, verbal memory and TICS, respectively). The association between the MIND diet and cognitive function was not mediated by high blood pressure as the effect estimates remained exactly the same when taking out high blood pressure from the multivariable adjusted model (data not shown).
Cognitive decline over time
Long-term adherence to the MIND score was not significantly associated with change over time in the global cognitive score, verbal memory score, or the TICS score, in either the age and education adjusted models, or in the multivariable-adjusted model (p-trend 0.95, 0.98, and 0.73 for global cognition, verbal memory and TICS, respectively) (Table 4).
Table 4.
Multivariable-adjusted mean differencesa in slopes of cognitive change by quintiles of long-term MIND-score in 16,058 older women
| Model 1b | Model 2c | |
|---|---|---|
| Global cognitive score | ||
| Quintile 1 | Ref | Ref |
| Quintile 2 | 0.002 (-0.005,0.010) | 0.001 (-0.007,0.009) |
| Quintile 3 | -0.002 (-0.010,0.006) | -0.004 (-0.011,0.004) |
| Quintile 4 | 0.000 (-0.009,0.008) | -0.002 (-0.010,0.006) |
| Quintile 5 | 0.004 (-0.003,0.012) | 0.001 (-0.007,0.009) |
| p-trend | 0.44 | 0.95 |
| Verbal memory score | ||
| Quintile 1 | Ref | Ref |
| Quintile 2 | 0.002 (-0.007,0.011) | 0.000 (-0.009,0.009) |
| Quintile 3 | -0.005 (-0.015,0.004) | -0.007 (-0.017,0.002) |
| Quintile 4 | -0.001 (-0.010,0.008) | -0.003 (-0.013,0.006) |
| Quintile 5 | 0.005 (-0.004,0.015) | 0.002 (-0.008,0.011) |
| p-trend | 0.44 | 0.98 |
| TICS score | ||
| Quintile 1 | Ref | Ref |
| Quintile 2 | 0.017 (-0.015,0.048) | 0.014 (-0.018,0.045) |
| Quintile 3 | 0.007 (-0.025,0.039) | 0.003 (-0.030,0.035) |
| Quintile 4 | -0.007 (-0.039,0.024) | -0.011 (-0.043,0.020) |
| Quintile 5 | 0.011 (-0.021,0.042) | 0.004 (-0.028,0.036) |
| p-trend | 0.95 | 0.73 |
Abbreviations: MIND=Mediterranean-DASH Intervention for Neurodegenerative Delay; TICS=Telephone interview of cognitive status; a. Values are mean difference in slopes (95% confidence interval); b. Model 1: adjusted for age and education; c. Model 2: additionally adjusted for physical activity, calorie intake, alcohol intake, smoking status, multivitamin use, BMI, depression and history of high blood pressure, hypercholesterolemia, myocardial infarction, and diabetes mellitus.
MIND-components and cognitive function
Excluding the component ‘Butter and margarine' and ‘Pastries and sweets' attenuated the association between the MIND score and average verbal memory (adjusted mean differences between extreme quintiles of intake were 0.02 (95%CI -0.02,0.05) for ‘Butter and margarine' and 0.03 (95%CI -0.01,0.06) for ‘Pastries and sweets, p-trends across categories of intake = 0.14 for both components), demonstrating that these components were the most important contributors to the association, as part of the MIND diet.
Discussion
In this large observational study in 16,058 older women, we observed that greater long-term adherence to the MIND diet was associated with a better verbal memory score, irrespective of ApoE ε4 status. We did not find evidence of an association between MIND diet adherence with global cognition or the TICS score, nor with cognitive change over time in any of the domains.
To put our results into perspective, we observed that each increase of one year of age of participants in this study population was associated with a mean difference of -0.05 standard units for verbal memory score. To provide a more clinical interpretation of the findings, this means that the effect estimate we found for MIND adherence was cognitively equivalent to being approximately one year younger. The same estimate was observed with a higher adherence to the Mediterranean diet (25).
The Memory and Aging Project (MAP), an open prospective cohort study in 960 older persons, demonstrated that the MIND diet was associated with slower decline in the global cognitive score and also in each of the five cognitive domains studied. The difference in rate of decline for highest adherence to the MIND diet compared to the lowest adherence was equivalent to that seen in women who were 7.5 years apart in age (10). To date, there have been no other studies on the MIND diet and cognitive function. A possible explanation for the difference in the strength of associations as reported in the MAP study and the present study, could be related to different population characteristics of the two cohorts. Compared to NHS participants, MAP study participants were on average ten years older than NHS participants, had a lower education level, mainly lived in retirement houses, were more likely to have ever smoked and more likely to have reported high blood pressure, diabetes and myocardial infarction. This would suggest that MAP participants were at higher risk of poor cognitive function and faster cognitive decline compared to NHS participants, resulting in a stronger association between the MIND diet and cognitive decline over time. Another explanation may relate to a difference in the cohort design, specifically the number and timing of dietary and cognitive assessments. First, the MAP study used 19 cognitive tests ten times over a period of nine years to compute a global measure of cognitive function and five cognitive domains, whereas the NHS used six cognitive tests four times during six years to construct a global measure of cognitive function and three cognitive domains. It could be that the MAP study was better able to detect differences in cognitive change with the more extensive cognitive assessment battery over a longer period of follow-up. Second, in the NHS study, we used all dietary assessments prior to the first cognitive assessment to create a measure of long-term dietary intake capturing changes in dietary intake over time, whereas the MAP study used the first dietary assessment to study cognitive change from that point forward. Hence, it could be that MAP study participants with good cognitive function at baseline may have reported healthier dietary intakes, which could have led to less decline in cognitive function at follow-up. On the other hand, those with poor cognitive function at baseline probably reported worse dietary intakes, resulting in a stronger association between the MIND diet and cognitive decline. These two issues could also explain why we were not able to find any association with cognitive change over time.
We additionally examined the relative contribution of individual components when excluding one component from the MIND score at a time, while including it as a covariate into the regression models. Using this approach, we observed that butter plus margarine and pastries plus sweets were the main components as part of the total MIND diet driving the association between the MIND diet and verbal memory within the NHS population. So far, no other studies have taken into account the relative contribution of these components to cognitive function. Future research should further explore the relative contribution of components of dietary patterns to better understand underlying mechanisms.
There are several biological mechanisms by which the MIND diet could impact cognitive function. First, by keeping the intake low for items such as butter and margarine or pastries and sweets, the consumption of saturated fats and trans fatty acids would be kept low, resulting in a better fat composition of the diet that can improve the blood-brain barrier function and decrease Aβ aggregation (27). Second, in an Alzheimer's disease mouse model, higher intakes of long-chain n-3 fatty acids from fish has been shown to reduce Aβ formation and oxidative damage and increase synaptic proteins and dendritic spine density (28, 29). Finally, the components of berries and green leafy vegetables provide high amounts of carotenoids, flavonoids, folate and vitamin E, which, in rodents, have been shown to have antioxidant and anti-inflammatory properties (30, 31) and to inhibit Aβ deposition (30, 32, 33, 34, 35). All of these mechanisms directly impact brain health and may be indirectly beneficial for cognitive maintenance via possible cardioprotective effects (36). Thus, there is a solid biological basis that the MIND diet could impact brain health.
Genetically, carrying one or more copies of the ApoE ε4 is associated with an increased risk of Alzheimer's disease (AD) and cognitive function decline (11, 37). Observational studies have demonstrated the largest cognitive decline among homozygous and an intermediate cognitive decline in heterozygous ApoE ε4 carriers after three years (38) and a more rapid decline in all cognitive domains in those carrying at least one ApoE ε4 allele (12). This evidence supports the importance of finding strategies to prevent cognitive decline especially among ApoE ε4 carriers. However, we were not able to demonstrate an interaction between the MIND diet and ApoE ε4 status. Additionally, adding the ApoE ε4 genotype to our fully adjusted model did not result in different or stronger associations. Other studies investigating an interaction between different dietary patterns and cognitive function have shown inconsistent findings, with studies reporting better cognitive function among non-carriers (13, 14), one study reporting an association among carriers (15), and three studies presenting no interaction (14, 16, 17). Overall, no consistent conclusion on the role of the ApoE4 ε4 allele between diet and cognitive function can be drawn and this requires further research.
This study has several limitations to consider. First, the FFQ did not specifically ask if olive oil was the primary source of fat, as this is not commonly consumed in this US population. As a best estimate, we chose to study the use of primarily vegetable oil instead of butter or margarine. Second, cognitive function was assessed by telephone assessment, possibly leading to some misclassification of cognitive function. A validation study, however, has shown that telephone-based cognitive battery performed well compared with detailed, in-person interviews (Q = 0.81 comparing the two modes of assessment) (39). Third, participants were relatively healthy and highly educated, limiting the generalizability of the results to other populations. It could be that a healthy diet is not so beneficial in the NHS population, whereas it may be more beneficial in the general population with more risk factors for poorer health and a lower educational level. Finally, we cannot exclude the possibility of residual confounding by unmeasured variables within this cohort, although we have adjusted for many potential confounders.
The relatively long follow-up allowed us to assess longterm associations between the MIND score and cognitive function. By averaging repeated measures of diet to obtain an assessment of long-term overall diet, we accounted for within-person change in intake and reduced measurement error (as previously documented by our group (40)). The same approach has been taken to limit measurement error in cognitive function assessments, by averaging all repeated measures into a more robust measure of cognitive function at older ages. The prospective nature of this analysis reduces the probability of recall bias and selection bias. In addition, a high follow-up rate reduced potential bias due to loss of follow-up. Nevertheless, our results need to be replicated in other populations at risk of poorer health to clarify the relevance of this association within populations at risk of cognitive decline.
To summarize, long-term adherence to the MIND diet, with high intakes of berries, green leafy vegetables and low intakes of pastries and sweets, was associated with a better verbal memory in later life, but not with global cognitive function, TICS score or cognitive decline.
Acknowledgements
Participants, staff and investigators of the NHS are gratefully acknowledged. This work was supported by the National Institutes of Health who provided funding for the Nurses' Health Study (NHS) (grant UM1 CA186107). The NHS cognitive assessment was supported by: R01 AG015424 and R01 AG015424. The NHS and the genetic data collection were supported by the following: the National Cancer Institute (NCI, UM1 CA186107, P01CA087969, P01CA049449, R01CA137178), the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK, R01DK058845), and the National Heart, Lung, and Blood Institute (NHLBI, R01HL35464), with additional support for the collection and management of cognitive and genetic data. The NHS assessment of cognitive function was supported by R01 AG015424 from the National Institute of Aging. The NHS Breast Cancer GWAS (dbGaP:phs000147.v1.p1) was performed as part of the Cancer Genetic Markers of Susceptibility initiative of the NCI (R01CA40356, U01CA98233). The NHS type 2 diabetes (T2D, dbGaP:phs000091.v2.p1) and openangle glaucoma (GA, dbGaP:phs000308.v1.p1) GWAS were funded as part of the Gene Environment-Association Studies (GENEVA) project under the NIH Genes, Environment, and Health Initiative (T2D: U01HG004399, GA: U01HG004728). Genotyping for the NHS coronary heart disease GWAS was supported by Merck/Rosetta Research Laboratories, North Wales, PA. The NHS kidney stone GWAS (dbGaP:phs000460.v1.p1) was supported by NIDDK (5P01DK070756). The NHS colon cancer GWAS (dbGAP: in progress) was funded as part of the Colorectal Cancer GWAS Consortium funded by the NCI (U01 CA137088, R01CA059045).
Conflicts of interest
No conflicts of interest are reported by Berendsen, Kang, van de Rest, de Groot and Feskens. Grodstein reports grants from International Nut Council, other from California Walnut Council, outside the submitted work.
References
- 1.United Nations., Department of Economic and Social Affairs, Population Division. World Population Ageing 2013., 2013
- 2.World Health Organization. 10 facts on ageing and the life course Geneva, Switzerland, 2012 [cited 2015 15-06-2015]. Available from: http://www.who.int/features/factfiles/ageing/ageing_facts/en/index3.html.
- 3.Nelson L, Tabet N. Slowing the progression of Alzheimer’s disease; what works. Ageing Res Rev. 2015;23:193–209. doi: 10.1016/j.arr.2015.07.002. 10.1016/j.arr.2015.07.002 PubMed PMID: 26219494. [DOI] [PubMed] [Google Scholar]
- 4.van de Rest O, Berendsen AM, Haveman-Nies A, de Groot CPGM. Dietary patterns, cognitive decline, and dementia: a systematic review. Adv Nutr. 2015;6(2):154–168. doi: 10.3945/an.114.007617. 10.3945/an.114.007617 PubMed PMID: 25770254, PMCID 4352174. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Smith PJ, Blumenthal JA, Babyak MA, Craighead L, Welsh-Bohmer KA, Browndyke JN, et al. Effects of the dietary approaches to stop hypertension diet, exercise, and caloric restriction on neurocognition in overweight adults with high blood pressure. Hypertension. 2010;55(6):1331–1338. doi: 10.1161/HYPERTENSIONAHA.109.146795. 10.1161/HYPERTENSIONAHA.109.146795 PubMed PMID: 20305128, PMCID 2974436. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Tangney CC, Li H, Wang Y, Barnes L, Schneider JA, Bennett DA, et al. Relation of DASH-and Mediterranean-like dietary patterns to cognitive decline in older persons. Neurology. 2014;83(16):1410–1416. doi: 10.1212/WNL.0000000000000884. 10.1212/WNL.0000000000000884 PubMed PMID: 25230996, PMCID 4206157. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Wengreen H, Munger RG, Cutler A, Quach A, Bowles A, Corcoran C, et al. Prospective study of Dietary Approaches to Stop Hypertension-and Mediterraneanstyle dietary patterns and age-related cognitive change: the Cache County Study on Memory, Health and Aging. Am J Clin Nutr. 2013;98(5):1263–1271. doi: 10.3945/ajcn.112.051276. 10.3945/ajcn.112.051276 PubMed PMID: 24047922, PMCID 3798079. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Berendsen AM, Kang JH, Van de Rest O, Feskens E, De Groot CPGM, Grodstein F. Adherence to the Dietary Approaches to Stop Hypertension diet, cognitive function and cognitive decline in American older women. unpublished results. [DOI] [PubMed]
- 9.Morris MC, Tangney CC, Wang Y, Sacks FM, Bennett DA, Aggarwal NT. MIND diet associated with reduced incidence of Alzheimer’s disease. Alzheimers Dement. 2015;11(9):1007–1014. doi: 10.1016/j.jalz.2014.11.009. 10.1016/j.jalz.2014.11.009 PubMed PMID: 25681666, PMCID 4532650. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Morris MC, Tangney CC, Wang Y, Sacks FM, Barnes LL, Bennett DA, et al. MIND diet slows cognitive decline with aging. Alzheimers Dement. 2015;11(9):1015–1022. doi: 10.1016/j.jalz.2015.04.011. 10.1016/j.jalz.2015.04.011 PubMed PMID: 26086182, PMCID 4581900. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Corder EH, Saunders AM, Strittmatter WJ, Schmechel DE, Gaskell PC, Small GW, et al. Gene dose of apolipoprotein E type 4 allele and the risk of Alzheimer’s disease in late onset families. Science. 1993;261(5123):921–923. doi: 10.1126/science.8346443. 10.1126/science.8346443 PubMed PMID: 8346443. [DOI] [PubMed] [Google Scholar]
- 12.Wilson RS, Schneider JA, Barnes LL, Beckett LA, Aggarwal NT, Cochran EJ, et al. The apolipoprotein E epsilon 4 allele and decline in different cognitive systems during a 6-year period. Arch Neurol. 2002;59(7):1154–1160. doi: 10.1001/archneur.59.7.1154. 10.1001/archneur.59.7.1154 PubMed PMID: 12117364. [DOI] [PubMed] [Google Scholar]
- 13.Barberger-Gateau P, Raffaitin C, Letenneur L, Berr C, Tzourio C, Dartigues JF, et al. Dietary patterns and risk of dementia: the Three-City cohort study. Neurology. 2007;69(20):1921–1930. doi: 10.1212/01.wnl.0000278116.37320.52. 10.1212/01.wnl.0000278116.37320.52 PubMed PMID: 17998483. [DOI] [PubMed] [Google Scholar]
- 14.Martinez-Lapiscina EH, Galbete C, Corella D, Toledo E, Buil-Cosiales P, Salas-Salvado J, et al. Genotype patterns at CLU, CR1, PICALM and APOE, cognition and Mediterranean diet: the PREDIMED-NAVARRA trial. Genes Nutr. 2014;9(3):393–397. doi: 10.1007/s12263-014-0393-7. 10.1007/s12263-014-0393-7 PubMed PMID: 24643340, PMCID 4026432. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Gardener SL, Rainey-Smith SR, Barnes MB, Sohrabi HR, Weinborn M, Lim YY, et al. Dietary patterns and cognitive decline in an Australian study of ageing. Mol Psychiatry. 2015;20(7):860–866. doi: 10.1038/mp.2014.79. 10.1038/mp.2014.79 PubMed PMID: 25070537. [DOI] [PubMed] [Google Scholar]
- 16.Cherbuin N, Anstey KJ. The Mediterranean diet is not related to cognitive change in a large prospective investigation: the PATH Through Life study. Am J Geriatr Psychiatry. 2012;20(7):635–639. doi: 10.1097/JGP.0b013e31823032a9. 10.1097/JGP.0b013e31823032a9 PubMed PMID: 21937919. [DOI] [PubMed] [Google Scholar]
- 17.Roberts RO, Geda YE, Cerhan JR, Knopman DS, Cha RH, Christianson TJ, et al. Vegetables, unsaturated fats, moderate alcohol intake, and mild cognitive impairment. Dementia and geriatric cognitive disorders. 2010;29(5):413–423. doi: 10.1159/000305099. 10.1159/000305099 PubMed PMID: 20502015, PMCID 2889256. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Willett WC, Sampson L, Stampfer MJ, Rosner B, Bain C, Witschi J, et al. Reproducibility and validity of a semiquantitative food frequency questionnaire. Am J Epidemiol. 1985;122(1):51–65. doi: 10.1093/oxfordjournals.aje.a114086. 10.1093/oxfordjournals.aje.a114086 PubMed PMID: 4014201. [DOI] [PubMed] [Google Scholar]
- 19.Launer LJ. The epidemiologic study of dementia: a life-long quest. Neurobiol Aging. 2005;26(3):335–340. doi: 10.1016/j.neurobiolaging.2004.03.016. 10.1016/j.neurobiolaging.2004.03.016 PubMed PMID: 15639311. [DOI] [PubMed] [Google Scholar]
- 20.Brandt J, Spencer M, Folstein M. The telephone interview for cognitive status. Neuropsychiatry Neuropsychol Behav Neurol. 1988;1:111–117. [Google Scholar]
- 21.Albert M, Smith LA, Scherr PA, Taylor JO, Evans DA, Funkenstein HH. Use of brief cognitive tests to identify individuals in the community with clinically diagnosed Alzheimer’s disease. Int J Neurosci. 1991;57(3-4):167–178. doi: 10.3109/00207459109150691. 10.3109/00207459109150691 PubMed PMID: 1938160. [DOI] [PubMed] [Google Scholar]
- 22.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):189–198. doi: 10.1016/0022-3956(75)90026-6. PubMed PMID: 1202204. [DOI] [PubMed] [Google Scholar]
- 23.Morris JC, Heyman A, Mohs RC, Hughes JP, van Belle G, Fillenbaum G, et al. The Consortium to Establish a Registry for Alzheimer’s Disease (CERAD) Part I. Clinical and neuropsychological assessment of Alzheimer’s disease. Neurology. 1989;39(9):1159–1165. doi: 10.1212/wnl.39.9.1159. PubMed PMID: 2771064. [DOI] [PubMed] [Google Scholar]
- 24.Devore EE, Kang JH, Stampfer MJ, Grodstein F. The association of antioxidants and cognition in the Nurses’ Health Study. Am J Epidemiol. 2013;177(1):33–41. doi: 10.1093/aje/kws202. 10.1093/aje/kws202 PubMed PMID: 23221724. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Samieri C, Okereke OI, Devore EE, Grodstein F. Long-term adherence to the Mediterranean diet is associated with overall cognitive status, but not cognitive decline, in women. J Nutr. 2013;143(4):493–499. doi: 10.3945/jn.112.169896. 10.3945/jn.112.169896 PubMed PMID: 23365105, PMCID 3738244. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Laird NM, Ware JH. Random-effects models for longitudinal data. Biometrics. 1982;38(4):963–974. 10.2307/2529876 PubMed PMID: 7168798. [PubMed] [Google Scholar]
- 27.Morris MC, Tangney CC. Dietary fat composition and dementia risk. Neurobiol Aging. 2014;35(2):S59–64. doi: 10.1016/j.neurobiolaging.2014.03.038. 10.1016/j.neurobiolaging.2014.03.038 PubMed PMID: 24970568. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Calon F, Lim GP, Yang F, Morihara T, Teter B, Ubeda O, et al. Docosahexaenoic acid protects from dendritic pathology in an Alzheimer’s disease mouse model. Neuron. 2004;43(5):633–645. doi: 10.1016/j.neuron.2004.08.013. 10.1016/j.neuron.2004.08.013 PubMed PMID: 15339646, PMCID 2442162. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Lim GP, Calon F, Morihara T, Yang F, Teter B, Ubeda O, et al. A diet enriched with the omega-3 fatty acid docosahexaenoic acid reduces amyloid burden in an aged Alzheimer mouse model. J Neurosci. 2005;25(12):3032–3040. doi: 10.1523/JNEUROSCI.4225-04.2005. 10.1523/JNEUROSCI.4225-04.2005 PubMed PMID: 15788759. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Jiang Q, Ames BN. Gamma-tocopherol, but not alpha-tocopherol, decreases proinflammatory eicosanoids and inflammation damage in rats. FASEB J. 2003;17(8):816–822. doi: 10.1096/fj.02-0877com. 10.1096/fj.02-0877com PubMed PMID: 12724340. [DOI] [PubMed] [Google Scholar]
- 31.Yamada K, Tanaka T, Han D, Senzaki K, Kameyama T, Nabeshima T. Protective effects of idebenone and alpha-tocopherol on beta-amyloid-(1-42)-induced learning and memory deficits in rats: implication of oxidative stress in beta-amyloid-induced neurotoxicity in vivo. Eur J Neurosci. 1999;11(1):83–90. doi: 10.1046/j.1460-9568.1999.00408.x. 10.1046/j.1460-9568.1999.00408.x PubMed PMID: 9987013. [DOI] [PubMed] [Google Scholar]
- 32.Chan A, Shea TB. Folate deprivation increases presenilin expression, gammasecretase activity, and Abeta levels in murine brain: potentiation by ApoE deficiency and alleviation by dietary S-adenosyl methionine. J Neurochem. 2007;102(3):753–760. doi: 10.1111/j.1471-4159.2007.04589.x. 10.1111/j.1471-4159.2007.04589.x PubMed PMID: 17504266. [DOI] [PubMed] [Google Scholar]
- 33.Katayama S, Ogawa H, Nakamura S. Apricot carotenoids possess potent antiamyloidogenic activity in vitro. J Agric Food Chem. 2011;59(23):12691–12696. doi: 10.1021/jf203654c. 10.1021/jf203654c PubMed PMID: 22043804. [DOI] [PubMed] [Google Scholar]
- 34.Nishida Y, Ito S, Ohtsuki S, Yamamoto N, Takahashi T, Iwata N, et al. Depletion of vitamin E increases amyloid beta accumulation by decreasing its clearances from brain and blood in a mouse model of Alzheimer disease. J Biol Chem. 2009;284(48):33400–33408. doi: 10.1074/jbc.M109.054056. 10.1074/jbc.M109.054056 PubMed PMID: 19679659, PMCID 2785184. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Obulesu M, Dowlathabad MR, Bramhachari PV. Carotenoids and Alzheimer’s disease: an insight into therapeutic role of retinoids in animal models. Neurochem Int. 2011;59(5):535–541. doi: 10.1016/j.neuint.2011.04.004. 10.1016/j.neuint.2011.04.004 PubMed PMID: 21672580. [DOI] [PubMed] [Google Scholar]
- 36.Duron E, Hanon O. Hypertension, cognitive decline and dementia. Arch Cardiovasc Dis. 2008;101(3):181–183. doi: 10.1016/s1875-2136(08)71801-1. 10.1016/S1875-2136(08)71801-1 PubMed PMID: 18477946. [DOI] [PubMed] [Google Scholar]
- 37.Saunders AM, Strittmatter WJ, Schmechel D, George-Hyslop PH, Pericak-Vance MA, Joo SH, et al. Association of apolipoprotein E allele epsilon 4 with late-onset familial and sporadic Alzheimer’s disease. Neurology. 1993;43(8):1467–1472. doi: 10.1212/wnl.43.8.1467. 10.1212/WNL.43.8.1467 PubMed PMID: 8350998. [DOI] [PubMed] [Google Scholar]
- 38.Feskens EJ, Havekes LM, Kalmijn S, de Knijff P, Launer LJ, Kromhout D. Apolipoprotein e4 allele and cognitive decline in elderly men. BMJ. 1994;309(6963):1202–1206. doi: 10.1136/bmj.309.6963.1202. 10.1136/bmj.309.6963.1202 PubMed PMID: 7987151, PMCID 2541698. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Devore EE, Kang JH, Breteler MM, Grodstein F. Dietary intakes of berries and flavonoids in relation to cognitive decline. Annals of neurology. 2012;72(1):135–143. doi: 10.1002/ana.23594. 10.1002/ana.23594 PubMed PMID: 22535616, PMCID 3582325. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Hu FB, Stampfer MJ, Rimm E, Ascherio A, Rosner BA, Spiegelman D, et al. Dietary fat and coronary heart disease: a comparison of approaches for adjusting for total energy intake and modeling repeated dietary measurements. Am J Epidemiol. 1999;149(6):531–540. doi: 10.1093/oxfordjournals.aje.a009849. 10.1093/oxfordjournals.aje.a009849 PubMed PMID: 10084242. [DOI] [PubMed] [Google Scholar]