PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
 
J Alzheimers Dis. Author manuscript; available in PMC 2011 January 1.
Published in final edited form as:
PMCID: PMC2940991
NIHMSID: NIHMS198327

Polyunsaturated Fatty Acids and Reduced Odds of MCI: The Mayo Clinic Study of Aging

Abstract

Mono- and polyunsaturated fatty acids (MUFA, PUFA) have been associated with a reduced risk of dementia. The association of these fatty acids with mild cognitive impairment (MCI) is not fully established. The objective of the study was to investigate the cross-sectional association of dietary fatty acids with MCI in a population-based sample. Participants aged ≥ 70 years on October 1, 2004, were evaluated using the Clinical Dementia Rating Scale (participant and informant), a neurological evaluation, and neuropsychological testing. A panel of nurses, physicians, and neuropsychologists reviewed the data for each participant in order to establish a diagnosis of MCI, normal cognition, or dementia by consensus. Participants also completed a 128-item food-frequency questionnaire. Among 1,233 non-demented subjects, 163 (13.2%) had MCI. The odds ratio (OR) of MCI decreased with increasing PUFA and MUFA intake. Compared to the lowest tertile, the OR (95% confidence interval) for the upper tertiles were 0.44 (0.29–0.66; p for trend = 0.0004) for total PUFA; 0.44 (0.30–0.67; p for trend = 0.0004) for omega-6 fatty acids; 0.62 (0.42–0.91; p for trend = 0.012) for omega-3 fatty acids; and 0.56 (0.38–0.83; p for trend = 0.01) for (MUFA+PUFA):saturated fatty acid ratio after adjustment for age, sex, number of years of education, and caloric intake. In this study, higher intake of PUFA and MUFA was associated with a reduced likelihood of MCI among elderly persons in the population-based setting.

Keywords: Cross-sectional studies, dietary fats, polyunsaturated fatty acids, monounsaturated fatty acids, population-based, mild cognitive impairment

INTRODUCTION

Dietary fat intake represents a potential modifiable risk factor for mild cognitive impairment (MCI). Several studies have reported associations of high intake of dietary saturated fat with Alzheimer’s disease (AD), vascular dementia, and cognitive impairment [17]. In contrast, monounsaturated fatty acids (MUFA) and n-3 polyunsaturated fatty acids (PUFA) have been reported to have beneficial effects on cognitive function [4, 8, 9]. MUFAs are derived primarily from olive oil. PUFAs include omega-3 PUFAs and linolenic acids (primarily from fish and plant sources), and omega-6 PUFAs and linoleic acids are derived primarily from seed oils. A few other studies have not found an association of any dietary fats with incident dementia [10], or have reported that high intake of n-6 PUFA is associated with cognitive impairment assessed from the Mini-Mental State Exam (MMSE) [11, 12].

The association of the components of dietary fat intake with MCI has not been fully established in a population-based setting. Our objective is to investigate these associations in a community-based sample of subjects in the Mayo Clinic Study of Aging who have been well-characterized for MCI using previously specified criteria for MCI at the time of the evaluation, and in whom dietary fat intake has been ascertained using a comprehensive food frequency questionnaire.

MATERIALS AND METHODS

Study participants

The details of the study design have been published previously [13]. Briefly, we identified all Olmsted County, MN, residents aged 70–89 years on October 1, 2004, using the medical records-linkage system of the Rochester Epidemiology Project [14]. From among 9,953 subjects enumerated, we identified 4,398 eligible subjects: 2,719 agreed to participate (61.8% response) by telephone (n = 669) or via a face-to-face evaluation (n = 2,050). Information from the medical records-linkage system showed that non-participants were more likely to be older, men, less educated, to have diabetes, and greater comorbidity [13]. Only subjects who participated in the face-to-face evaluation were considered for eligibility for the dietary assessments.

Of the 2,050 subjects that had an in-person evaluation at baseline, 67 had dementia at baseline and 14 did not complete the evaluation and therefore could not be assigned a diagnosis; these 81 subjects were not eligible for the food frequency questionnaire study. Of the remaining 1,969 non-demented subjects who completed the full evaluation at baseline, subjects who had died (n = 22), refused further participation since the baseline evaluation (n = 22), or were non-English speaking (n = 1) were considered ineligible. Thus, the food frequency questionnaires were mailed to 1,924 eligible subjects. A total of 1,567 (81.4%) subjects returned the food frequency questionnaire and 357 did not. We also excluded data for 334 subjects who had 1) missing responses on > 10 of the 128 ‘How often’ questions on frequency of consumption of foods (n = 268); 2) persons who reported extreme caloric intake (low intake = < 800 [in men] or < 600 [in women] kcal/day; high intake = > 6,000 [in men] or > 5,000 [in women] kcal/day); and 3) persons who were demented at the time of the evaluation (n = 10; at baseline, 9 had MCI and 1 was cognitively normal; Fig.). Thus, of those who received the questionnaire, 691 subjects were not included in the present analyses.

Fig
Flow-chart of subjects included in the analyses. Subjects who were not eligible to receive the food frequency questionnaire included 45 subjects who died (n = 22) after the baseline assessment, refused further participation (n = 22), or were non-English ...

Measurements

Assessment of cognitive status

Each study participant was evaluated by a nurse or study coordinator, and underwent a neurological evaluation by a physician and neuropsychological testing. A nurse or study coordinator administered the Clinical Dementia Rating (CDR) Scale [15] to the participant and an informant by interview. We performed neuropsychological testing using nine tests to assess performance in four cognitive domains: memory, executive function, language, and visuospatial skills, as previously described [13]. The raw scores from the neuropsychological test battery were age-adjusted using normative data from the Mayo’s Older American Normative Studies and scaled to a mean of 10 and standard deviation (SD) of 3 [16], and were also summed and scaled to obtain domain scores [13]. An expert panel of physicians, neuropsychologists, and nurses reviewed all the information collected for each participant to reach a consensus diagnosis of normal cognition, MCI, or dementia.

Definition of cases and controls

We defined MCI (cases) according to previously published criteria: cognitive concern by participant, physician, nurse, or informant (from CDR); impairment in one or more of the four cognitive domains from the cognitive testing battery; essentially normal functional activities (from CDR); and absence of dementia [17]. A diagnosis of dementia was made according to the Diagnostic and Statistical Manual of Mental Disorders, 4th edition criteria [18]. Subjects were characterized as cognitively normal according to published criteria [16]. Subjects with dementia were not included in this study.

Assessment of dietary lipid intake from a food frequency questionnaire

We mailed a self-administered food frequency questionnaire that was a modification of the Block 1995 Revision of the Health Habits and History Questionnaire [19] to the homes of participants, with a self-addressed, stamped envelope for its return. The questionnaire included 103 food items, 25 beverages, cooking doneness for 7 meats, and use of 15 types of supplements; it was focused on usual eating habits during the year prior to enrollment, including food eaten in a restaurant. It also included questions about 1) type of fat most often used in cooking; 2) type of fat added at the table to bread, potatoes, and vegetables; 3) types of margarine used; 4) other forms of fat intake; and 5) fat preferences for milk and meat products. Dietary sources of MUFAs include olive oil, canola oil, peanut oils, peanuts and almonds, dairy, and meat. Omega-3 PUFA is present primarily in fish and seafood but is also found in animal products such as meat, poultry, eggs, and dairy; sources of linolenic acid include some nuts and vegetable oils (flaxseed, canola oil, soybean oils, walnuts) as well as from green leafy vegetables. Omega-6 PUFA and linoleic acids are present in vegetable oils (safflower, canola, and sunflower oils), soybeans, butter, and cheese. For each food, participants 1) indicated their usual portion size consumed (small, medium, or large), with the medium size specified (e.g., medium serving = 1 banana, 1 cup); and 2) how often they had consumed each food (never or < 1/month, 1–3/month, 1/week, 2–4/week, 5–6/week, 1/day, 2–3/day, 4–5/day, 6+/day). We analyzed the data using the Food Processor SQL nutrition analysis software (version 10.0.0., ESHA Research, Salem, OR), under the direction of a registered dietician (HMO). Using this database, we multiplied daily food consumption by 30 and weekly food consumption by 4 to determine monthly intake, and calculated the total nutrient intake (grams) and total caloric intake (kilocalories/day).

Assessment of demographic variables and other covariates

We ascertained date of birth, number of years of education, marital status, and potential confounders including diabetes, hypertension, coronary heart disease, and stroke by interview and from medication bottles provided by subjects. We validated comorbidities using the medical records-linkage system [14] with the medical record as the gold standard when there was a discrepancy. We assessed a history of depressive symptoms from an informant using the Neuropsychiatric Inventory Questionnaire [20]. We assessed frequency of moderate physical exercise in the year prior to the evaluation as ≤ 1/month, 2–3/month, 1–2/week, 3–4/week, 5–6/week, and daily [21]. Each participant underwent a blood draw at baseline to assess Apolipoprotein (ApoE) ε4 allele status. The Mayo Clinic and Olmsted Medical Center Institutional Review Boards approved all study protocols.

Statistical analyses

To calculate the energy-adjusted values of dietary fats, we used the residual from a regression model with the log(energy) as the independent variable and the log(nutrient) as the dependent variable, plus the mean log(nutrient) value derived from the regression model [22]. This method yields a measure of nutrient intake that is not (or minimally) correlated with the total energy intake.

We included linoleic acid as a marker of omega-6 (n-6) PUFA because it is the most frequently consumed dietary n-6 PUFA in Western populations [23], and we also evaluated total omega-6 PUFA intake. Similarly, we assessed linolenic acid as a measure of plant-derived omega-3 (n-3) PUFA intake, and total omega-3 PUFA that includes linolenic acid, docosahexaenoic acid (DHA), and eicosapentaecoic acid (EPA). We computed odds ratios (OR) and 95% confidence intervals (95% CI) for the association between tertiles of fatty acids and MCI using multiple logistic regression models and tested for linear trends across tertiles. The tertiles were determined from sex-specific distributions for cases and controls combined. We performed the same analyses by MCI subtype (amnestic [a-MCI] and non-amnestic [na-MCI]). In the first model, we adjusted for age, sex, number of years of education, and total energy intake (as continuous variables). In the fully-adjusted model, we also included diabetes, stroke, ApoE ε4 allele status (ApoE ε4+ vs. ApoE ε4-), coronary heart disease, body mass index, and depressive symptoms. Total energy intake was included as a covariate to avoid spurious associations [24]. We investigated potential effect modification by age, sex, number of years of education, and ApoE genotype by including a product term of these variables with dietary fats in the model. We used the information from the face-to-face evaluation that was closest to the date of completion of the food frequency questionnaire. We also assessed the effects of additional potential confounders: marital status, dyslipidemia, alcohol intake (moderate [> 0 - < 30 gm/day] vs. low or high [0 or ≥30 gm/day]), cigarette smoking (yes vs. no), and moderate exercise (yes vs. no), with each variable added separately to the base model and with all variables simultaneously included in the fully-adjusted models.

Potential non-participation

To adjust for potential non-participation bias, we used a propensity score approach [25, 26]. We developed a logistic regression model that predicted the probability of participation. We included demographic (e.g., age, sex, number of years of education) and clinical (e.g., diabetes, stroke, depressive symptoms, MCI status), characteristics that could affect participation. We then used the reciprocal of the predicted probabilities as covariates in the multiple regression analyses. The results for the models without the adjustment for non-participation were essentially the same and are not presented, but are available online (Appendix 1 and Appendix 2).

Appendix 1
Association of tertiles of PUFA, MUFA, and SFA with MCI*
Appendix 2
Association of tertiles of omega-6, omega-3, linoleic acid, and linolenic acid with MCI*

Sensitivity analyses

To determine whether recall bias from subjects with memory problems could influence our findings, we performed separate analyses after excluding 1) subjects with a memory domain score at the 5th percentile or lower (n = 110); 2) subjects with more than 12 months between the diagnosis and completion of the food frequency questionnaire (n = 19); and 3) subjects who had a medical record notation of cognitive impairment prior to enrollment in the study based on a review of their medical record (n = 17).

Preliminary longitudinal analyses

In subjects who were cognitively normal at baseline, we used proportional hazards models to investigate the association of baseline levels of PUFA, MUFA, omega-3 and omega-6 fatty acids, linoleic and linolenic acids (key variables) with incident MCI or dementia during follow-up; subjects were censored at the time of death or loss to follow-up. All analyses were performed using SAS® (SAS Institute, Cary, NC).

RESULTS

A total of 1,233 subjects were included in the present study. Subjects who were not included (n = 691) were older (median = 82.2 vs. 80.1 years; p < 0.0001); had a higher frequency of stroke (14.2% vs. 10.1%; p = 0.007), diabetes (16.1% vs. 12.1%; p = 0.014), depressive symptoms (16.3% vs. 12.3%; p = 0.018), and MCI (23.9% vs. 13.2%; p < 0.0001) compared to participants; a lower frequency was married (52.5% vs. 64.9%; p < 0.0001); and a lower frequency had > 12 years of education (49.2% vs. 56.0%; p < 0.0001) but they were similar in regard to history of hypertension, coronary heart disease, sex, and ApoE ε4 genotype. There were no differences between MCI cases who were excluded and MCI cases who were included in regard to age (median = 83.6 vs. 82.9 years; p = 0.21), percentage of men (51.5% vs. 59.5%; p = 0.15), and any ApoE ε4 allele (31.4% vs. 33.3%; p = 0.72) respectively.

Our present findings are based on the 1,233 subjects who provided useable data. Questionnaires for these subjects were completed at a median of 3.3 months (interquartile range = 1.5–6.1) from the most recent diagnostic assessment. The median memory domain score was 0.28 (interquartile range = −0.46–1.08). MCI cases were more likely to be older (p < 0.0001), male (p = 0.04), have fewer years of education (p = 0.001), have a higher frequency of prior stroke (p < 0.0001) and depressive symptoms (p < 0.0001), and an ApoE ε4 allele (p = 0.008) than controls (Table 1).

Table 1
Demographic and clinical characteristics of study participants by MCI status, Mayo Clinic Study of Aging, 2004–2006*

MCI cases had lower daily intake of several dietary fats compared to controls (Table 2). There were significant differences between cases and controls regarding intake of PUFA, omega-6 PUFA, omega-3 PUFA, fatty acids, linoleic acid, and linolenic acid. The (MUFA+PUFA):SFA ratio was also lower in MCI cases (p = 0.049), and trans-fatty acids intake was marginally higher in MCI cases.

Table 2
Dietary fatty acids of participants (median [interquartile range]) at baseline by MCI status*

The OR of MCI decreased significantly with increasing intake of total PUFA and (MUFA+PUFA):SFA ratio (Table 3). A similar trend was observed for MUFA intake, MUFA:SFA ratio, and total fat, but these trends were not statistically significant. Table 4 demonstrates a dose-response association of the components of PUFA with MCI. Increasing intake of omega-6 fatty acid and linoleic acid (both n-6 PUFA), and of omega-3 fatty acid and linolenic acid (both n-3 PUFA), was associated with a reduced OR of MCI. There was no potential confounding by dyslipidemia, hypertension, marital status, current smoking status, moderate alcohol intake, or moderate exercise when these variables were included in the base models separately (data not presented). When they were simultaneously included in the fully-adjusted models, there was little change in the estimates. For example, the OR (95% CI) for the highest vs. lowest tertile (reference) were as follows: 0.42 (0.24–0.73) for PUFA; 0.52 (0.30–0.91) for (MUFA+PUFA):SFA ratio; 0.43 (0.25–0.75) for omega-6 fatty acid; 0.46 (0.27–0.79) for omega-3 fatty acid; 0.42 (0.24–0.74) for linoleic acid; and 0.35 (0.19–0.62) for linolenic acid.

Table 3
Association of tertiles of PUFA, MUFA, and SFA with MCI*
Table 4
Association of tertiles of omega-6, omega-3, linoleic acid, and linolenic acid with MCI*

There was no interaction of fatty acids with age, sex, or number of years of education. However, there was a significant interaction of MUFA with sex, and of cholesterol with ApoE ε4 allele. High intake of MUFA was associated with a reduced OR of MCI in men (OR = 0.71; 95% CI = 0.53–0.94; ordinal p value across tertiles = 0.018) but not in women (OR = 1.08; 95% CI = 0.77–1.51; ordinal p value across tertiles = 0.67; p for interaction = 0.04). High intake of cholesterol was associated with a reduced OR of MCI in persons who were ApoE ε4+ (OR = 0.62; 95% CI = 0.43–0.92; ordinal p value = 0.016) but not in persons who were ApoE ε4- (OR = 1.06; 95% CI = 0.83–1.35; ordinal p value = 0.64; p for interaction = 0.03). We examined the associations further across specific ApoE ε4 genotypes and age group (Table 5). The OR (95% CI) decreased with increasing numbers of ε4 alleles: ApoE ε3ε3–1.05 (0.80, 1.36; p = 0.73); ApoE ε3ε4–0.78 (0.51, 1.20; p = 0.26); and ApoE ε4ε4–0.20 (0.03, 1.26; p = 0.09); however, the sample sizes were much reduced and none of the associations reached statistical significance. The same pattern was observed when stratified by age.

Table 5
Association of cholesterol intake with MCI across ApoE genotypes

Table 6 presents the associations of PUFA, MUFA and their components with MCI subtypes. Intake in the upper tertile was significantly associated with a reduced OR of a-MCI for all the variables considered in this table. The OR of na-MCI was significantly reduced for intake in the upper tertile of omega-6 fatty acid and for linoleic acid; estimates for PUFA, (MUFA+PUFA):SFA ratio, omega-3 fatty acid, and linolenic acid were also reduced, but the CIs included 1.

Table 6
Association of tertiles of PUFA, MUFA, omega-6, omega-3, linoleic acid, and linolenic acid with MCI subtypes*

Sensitivity analyses

After exclusion of subjects with a memory score in the 5th percentile or lower (n = 110), our results remained essentially the same as for the total sample. Compared to the lowest tertile, there was a dose-response association of MCI with daily intake of PUFA (middle–tertile OR = 0.60, 95% CI = 0.38–0.95; upper tertile–OR = 0.42, 95% CI = 0.25–0.68; p for trend = 0.002); omega-6 fatty acids (middle tertile–OR = 0.64, 95% CI = 0.41–1.01; upper tertile–OR = 0.43, 95% CI = 0.26–0.71; p for trend = 0.004); linoleic acid (middle tertile–OR = 0.64, 95% CI = 0.41–1.01; upper tertile–OR = 0.43, 95% CI = 0.26–0.70; p for trend = 0.004), and linolenic acid (middle tertile–OR = 0.98, 95% CI = 0.63–1.53; upper tertile–OR = 0.55, 95% CI = 0.33–0.91; p for trend = 0.038). The dose-response association persisted for omega-3 fatty acids (middle tertile–OR = 0.79, 95% CI = 0.50–1.26; upper tertile–OR = 0.76, 95% CI = 0.47–1.22; p for trend = 0.46), but the test for trend was no longer significant due to the reduced power. When we excluded subjects with more than 12 months lag between diagnosis and food frequency questionnaire (n = 19), and subjects who had a medical record notation of cognitive impairment prior to enrollment in the study (n = 17), the results were essentially unchanged, and the data are not presented.

Preliminary longitudinal analyses

Among subjects who were cognitively normal at baseline, 889 completed ≥ 1 longitudinal clinical evaluations after assessment of dietary intake, and there were 93 incident MCI events over a median follow-up of 2.7 years (interquartile range = 2.5, 3.7). Due to the relatively short follow-up time and number of incident events, we did not have adequate power to detect significant associations of dietary intakes with incident MCI. Nonetheless, the HR of MCI was reduced between 7% and 21% for persons with intake in the upper tertile at baseline: 0.79 (95% CI = 0.46–1.35) for PUFA; 0.90 (95% CI = 0.53–1.54) for omega-3 fatty acid, 0.92 (95% CI = 0.57–1.48) for (MUFA+PUFA):SFA ratio; and 0.93 (95% CI = 0.53–1.64) for omega-6 fatty acids compared to the lowest tertile, but the CIs included 1.

DISCUSSION

We observed significant dose-response associations between increasing dietary intake of polyunsaturated fatty acids, increasing (MUFA+PUFA):SFA ratio, and prevalent MCI. A similar but non-significant trend was observed for MUFA. In contrast, higher intake of trans-fatty acids was associated with an elevated OR of MCI. Although other studies have reported similar associations with dementia and cognitive impairment (using different diagnostic criteria), our findings provide support for a potential beneficial role of PUFA and MUFA in MCI. We observed similar associations with the MCI subtypes. Since MCI is an intermediate stage between normal cognitive aging and dementia, these findings may have potential implications for intervention if established in longitudinal studies assessing incident MCI.

Our study has several strengths. Given the population-based design, our findings are generalizable to the community. We used a previously validated questionnaire to assess dietary intakes [19]; this previously validated questionnaire had an 81% response rate of which 79% provided useable data. Our study endpoint, MCI, was rigorously assessed using a clinical protocol that involved three independent evaluators. In this epidemiological setting, the ascertainment of information on cognitive concern includes a subjective concern expressed by the study participant, a concern about the participant raised by a study partner, or concern about cognition raised by the examining clinician, thus enhancing the capture of all potential MCI cases. The ultimate diagnosis of MCI was made by consensus using previously specified criteria. However, the diagnosis of MCI was made at the time of the evaluation; this is in contrast to other studies that have either retrospectively applied MCI criteria to previously collected data or have assigned diagnoses based solely on neuropsychological test results.

Our findings are consistent with previous studies. High MUFA and PUFA intake were significantly associated with better cognition in the Italian Longitudinal Study of Aging [27]. In longitudinal studies, high PUFA intake was associated with a reduced risk of cognitive impairment in two Italian cohorts [3, 28]; olive oil and MUFA had beneficial effects on MMSE scores [12], and linoleic acid was associated with a reduced risk of dementia [7]. In the Chicago Health and Aging Project, high MUFA and PUFA:SFA ratios were associated with less cognitive decline [2, 4], total MUFA intake was associated with a reduced risk of AD [29], and high intake of trans-fatty acids was associated with greater cognitive decline [2]. Beneficial effects of PUFA and PUFA:SFA ratios were reported in the Cardiovascular Risk Factors, Aging, and Dementia Study after 21 years of follow-up [30], and in elderly French women [31]. In the Nurses Health Study, higher intake of trans-fatty acids and a lower PUFA:SFA ratio were associated with worse cognitive decline in women with type 2 diabetes mellitus [32]. In clinical trials, mild AD subjects treated with omega-3 fatty acids had a slower decline in MMSE scores [33], and short-term supplementation (90 days) with arachidonic acid and DHA was associated with improvement in memory and attention in MCI patients [34].

In other studies, high PUFA intake was adversely associated with MMSE scores [12], linoleic acid was associated with prevalent and incident cognitive impairment, and n-3 PUFA was not associated with cognitive impairment [11]. Others have found no association of low MUFA intake and cognitive function [10, 35].

The implications of the interactions of MUFA with sex, and cholesterol with ApoE ε4 allele in the present study are not clear. However, the latter is consistent with the suggestion that high serum levels of cholesterol may have beneficial effects in the elderly [36], in contrast to the adverse effects of high levels in mid-life [37, 38]. Other studies have observed no beneficial association [39, 40] or an adverse effect [41]. Yet, this finding may be of interest since ApoE is the principal cholesterol carrier protein in the brain [42]. Alternately, it is highly probable that the significant interaction observed when grouped as ε4 carrier versus non-carrier was due to a spurious association caused by the small number of ε2ε4 MCI cases and, in particular, to the five ε4ε4 MCI cases who survived to a late age. In the analyses stratified by all six ApoE genotypes, we observed no significant associations even among the ε3ε4 genotype, the ε4 carrier group with the largest number of MCI cases. Further longitudinal studies will help determine whether the beneficial association of high cholesterol intake in ApoE ε4 carriers is real, due to bias, or simply a spurious association as is possible in our study.

The beneficial effects of PUFA and MUFA on cognitive function may occur through several mechanisms. PUFA may reduce the risk of thrombosis, cardiovascular risk [43], and stroke [44], and may also inhibit inflammation [45]. They may improve the lipid profile of the brain [46], maintain the structural integrity of neuronal membranes [47], exert antioxidant effects [48], or decrease β-amyloid levels [49]. Alternately, others suggest that n-6 PUFA constitutes a high component of low-density lipoprotein, and consequently may increase the susceptibility of low-density lipoproteins to oxidation [50]. This would, in turn, promote atherosclerosis, thrombosis, strokes, infarctions, and white matter changes that may lead to cognitive impairment [11].

Despite the small number of subjects with na-MCI, the magnitude of the associations for na-MCI was similar to a-MCI, albeit with fewer significant associations. This would suggest that the potential beneficial effects of PUFA and MUFA are on both neurodegenerative and vascular mechanisms, or through an alternate mechanism.

The inconsistencies across studies could be due to several reasons. Methods used to assess cognitive function differ; some investigators use a single test such as the MMSE, whereas others use a neuropsychometric testing battery with or without a clinical assessment. Differences in nutritional databases, volunteer versus random sample, cross-sectional or longitudinal design, duration of longitudinal studies, and cultural differences in diet across populations may also be important. For example, in contrast to a Mediterranean setting, use of olive oil (MUFA) and fish intake in a Midwestern community may be well below the threshold at which significant beneficial effects can be detected in the research setting. Daily MUFA intake in the present study was about half that reported in an Italian community, whereas PUFA intake was higher [27].

Potential limitations include recall bias in reporting of dietary nutrients. In particular, it is possible that cases would report dietary intakes in a biased manner that would affect the validity of the study findings. However, the MCI cases were relatively mild (the median CDR sum of boxes was only 1.0, and the median memory domain score was −1.1). Cases were not informed that they received an MCI diagnosis at the evaluation, and the results were essentially the same after exclusion of subjects with a medical record notation of cognitive impairment prior to enrollment. Furthermore, the results remained essentially the same after exclusion of subjects in the lowest 5% of memory domain or subjects with more than 1 year between evaluation and completion of the food frequency questionnaire. However, these efforts to assess bias do not provide information on the direction or magnitude of any potential bias. A possible mechanism to validate dietary intakes and to determine the magnitude of bias would have been to ascertain dietary intakes in cases and controls by interview of the participant informants (i.e., proxy interview) using a matched case-control design. Nevertheless, this method would also yield valid data only where the proxy had detailed knowledge about the dietary habits of the participant over the period of time evaluated in the dietary questionnaire.

Another potential limitation that cannot be evaluated is that of reverse causality. The consistency of the findings across several longitudinal studies, including population-based studies, suggests that the associations between MUFA and PUFA and cognition may not be due entirely to reverse causality [24, 7, 21, 28]. Nonetheless, given that some studies have not confirmed the relationships, further evaluation of the associations in large population-based studies are needed to clarify this issue. Despite the potential for non-participation bias, the differential participation by MCI cases did not influence our results; the results with and without adjustment for propensity to participate were essentially unchanged. Due to multiple testing, a Bonferroni correction would require a p value of 0.0036 for statistical significance; yet, even at this conservative p value, the associations of PUFA with MCI remain statistically significant. Finally, the study was conducted among a predominantly Caucasian population; thus, the findings are generalizable to similar settings.

Acknowledgments

This work was supported in part by the National Institutes of Health (P50 AG016574 to VSP and RCP, U01 AG006786 to RCP, K01 AG028573 to ROR, R01 AR030582 to WAR, and K01 MH068351 to YEG) and by the Robert H. and Clarice Smith and Abigail van Buren Alzheimer’s Disease Research Program, MN (to RCP).

References

1. Solfrizzi V, Capurso C, D’Introno A, Colacicco AM, Santamato A, Ranieri M, Fiore P, Capurso A, Panza F. Lifestyle-related factors in predementia and dementia syndromes. Expert Rev Neurother. 2008;8:133–158. [PubMed]
2. 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;62:1573–1579. [PubMed]
3. Kalmijn S, van Boxtel MP, Ocke M, Verschuren WM, Kromhout D, Launer LJ. Dietary intake of fatty acids and fish in relation to cognitive performance at middle age. Neurology. 2004;62:275–280. [PubMed]
4. 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;60:194–200. [PubMed]
5. Luchsinger JA, Tang MX, Shea S, Mayeux R. Caloric intake and the risk of Alzheimer disease. Arch Neurol. 2002;59:1258–1263. [PubMed]
6. Kivipelto M, Helkala EL, Laakso MP, Hanninen T, Hallikainen M, Alhainen K, Soininen H, Tuomilehto J, Nissinen A. Midlife vascular risk factors and Alzheimer’s disease in later life: longitudinal, population based study. BMJ. 2001;322:1447–1451. [PMC free article] [PubMed]
7. 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;42:776–782. [PubMed]
8. Solfrizzi V, Panza F, Capurso A. The role of diet in cognitive decline. J Neural Transm. 2003;110:95–110. [PubMed]
9. Barberger-Gateau P, Letenneur L, Deschamps V, Peres K, Dartigues JF, Renaud S. Fish, meat, and risk of dementia: cohort study. BMJ. 2002;325:932–933. [PMC free article] [PubMed]
10. 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;59:1915–1921. [PubMed]
11. Kalmijn S, Feskens EJ, Launer LJ, Kromhout D. Polyunsaturated fatty acids, antioxidants, and cognitive function in very old men. Am J Epidemiol. 1997;145:33–41. [PubMed]
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;11:1054–1062. [PubMed]
13. Roberts RO, Geda YE, Knopman DS, Cha RH, Pankratz VS, Boeve BF, Ivnik RJ, Tangalos EG, Petersen RC, Rocca WA. The Mayo Clinic Study of Aging: design and sampling, participation, baseline measures and sample characteristics. Neuroepidemiology. 2008;30:58–69. [PMC free article] [PubMed]
14. Melton LJ., 3rd History of the Rochester Epidemiology Project. Mayo Clin Proc. 1996;71:266–274. [PubMed]
15. Morris JC. The Clinical Dementia Rating (CDR): current version and scoring rules. Neurology. 1993;43:2412–2414. [PubMed]
16. Ivnik RJ, Malec JF, Smith GE, Tangalos EG, Petersen RC, Kokmen E, Kurland LT. Mayo’s Older Americans Normative Studies: WAIS-R, WMS-R and AVLT norms for ages 56 through 97. Clin Neuropsychol. 1992;6:1–104.
17. Petersen RC. Mild cognitive impairment as a diagnostic entity. J Intern Med. 2004;256:183–194. [PubMed]
18. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) 4. American Psychiatric Association; Washington, D.C: 1994.
19. Block G, Coyle LM, Hartman AM, Scoppa SM. Revision of dietary analysis software for the Health Habits and History Questionnaire. Am J Epidemiol. 1994;139:1190–1196. [PubMed]
20. Kaufer DI, Cummings JL, Ketchel P, Smith V, MacMillan A, Shelley T, Lopez OL, DeKosky ST. Validation of the NPI-Q, a brief clinical form of the Neuropsychiatric Inventory. J Neuropsychiatry Clin Neurosci. 2000;12:233–239. [PubMed]
21. Geda YE, Roberts RO, Knopman DS, Christianson TJ, Pankratz VS, Ivnik RJ, Boeve BF, Tangalos EG, Petersen RC, Rocca WA. Physical exercise, aging, and mild cognitive impairment: a population-based study. Arch Neurol. 2010;67:80–86. [PMC free article] [PubMed]
22. Willett W, Stampfer MJ. Total energy intake: implications for epidemiologic analyses. Am J Epidemiol. 1986;124:17–27. [PubMed]
23. Wood DA, Riemersma RA, Butler S, Thomson M, Macintyre C, Elton RA, Oliver MF. Linoleic and eicosapentaenoic acids in adipose tissue and platelets and risk of coronary heart disease. Lancet. 1987;1:177–183. [PubMed]
24. Michels KB, Bingham SA, Luben R, Welch AA, Day NE. The effect of correlated measurement error in multivariate models of diet. Am J Epidemiol. 2004;160:59–67. [PubMed]
25. Kessler RC, Little RJ, Groves RM. Advances in strategies for minimizing and adjusting for survey nonresponse. Epidemiol Rev. 1995;17:192–204. [PubMed]
26. D’Agostino RB, Jr, Rubin DB. Estimating and using propensity scores with partially missing data. J Am Stat Assoc. 2000;95:749–759.
27. 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;27:1694–1704. [PubMed]
28. 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;41:619–627. [PubMed]
29. 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;60:940–946. [PubMed]
30. 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;23:741–747. [PubMed]
31. Vercambre MN, Boutron-Ruault MC, Ritchie K, Clavel-Chapelon F, Berr C. Long-term association of food and nutrient intakes with cognitive and functional decline: a 13-year follow-up study of elderly French women. Br J Nutr. 2009;102:419–427. [PMC free article] [PubMed]
32. Devore EE, Stampfer MJ, Breteler MM, Rosner B, Hee Kang J, Okereke O, Hu FB, Grodstein F. Dietary fat intake and cognitive decline in women with type 2 diabetes. Diabetes Care. 2009;32:635–640. [PMC free article] [PubMed]
33. Freund-Levi Y, Eriksdotter-Jonhagen M, Cederholm T, Basun H, Faxen-Irving G, Garlind A, Vedin I, Vessby B, Wahlund LO, Palmblad J. Omega-3 fatty acid treatment in 174 patients with mild to moderate Alzheimer disease: OmegAD study: a randomized double-blind trial. Arch Neurol. 2006;63:1402–1408. [PubMed]
34. Kotani S, Sakaguchi E, Warashina S, Matsukawa N, Ishikura Y, Kiso Y, Sakakibara M, Yoshimoto T, Guo J, Yamashima T. Dietary supplementation of arachidonic and docosahexaenoic acids improves cognitive dysfunction. Neurosci Res. 2006;56:159–164. [PubMed]
35. Ortega RM, Requejo AM, Andres P, Lopez-Sobaler AM, Quintas ME, Redondo MR, Navia B, Rivas T. Dietary intake and cognitive function in a group of elderly people. Am J Clin Nutr. 1997;66:803–809. [PubMed]
36. Solomon A, Kareholt I, Ngandu T, Winblad B, Nissinen A, Tuomilehto J, Soininen H, Kivipelto M. Serum cholesterol changes after midlife and late-life cognition: twenty-one-year follow-up study. Neurology. 2007;68:751–756. [PubMed]
37. Kalmijn S, Foley D, White L, Burchfiel CM, Curb JD, Petrovitch H, Ross GW, Havlik RJ, Launer LJ. Metabolic cardiovascular syndrome and risk of dementia in Japanese-American elderly men. The Honolulu-Asia Aging Study. Arterioscler Thromb Vasc Biol. 2000;20:2255–2260. [PubMed]
38. Kivipelto M, Helkala EL, Hanninen T, Laakso MP, Hallikainen M, Alhainen K, Soininen H, Tuomilehto J, Nissinen A. Midlife vascular risk factors and late-life mild cognitive impairment: A population-based study. Neurology. 2001;56:1683–1689. [PubMed]
39. Stewart R, White LR, Xue QL, Launer LJ. Twenty-six-year change in total cholesterol levels and incident dementia: the Honolulu-Asia Aging Study. Arch Neurol. 2007;64:103–107. [PubMed]
40. Li G, Shofer JB, Kukull WA, Peskind ER, Tsuang DW, Breitner JC, McCormick W, Bowen JD, Teri L, Schellenberg GD, Larson EB. Serum cholesterol and risk of Alzheimer disease: a community-based cohort study. Neurology. 2005;65:1045–1050. [PubMed]
41. Mielke MM, Zandi PP, Sjogren M, Gustafson D, Ostling S, Steen B, Skoog I. High total cholesterol levels in late life associated with a reduced risk of dementia. Neurology. 2005;64:1689–1695. [PubMed]
42. Puglielli L, Tanzi RE, Kovacs DM. Alzheimer’s disease: the cholesterol connection. Nat Neurosci. 2003;6:345–351. [PubMed]
43. Marchioli R, Barzi F, Bomba E, Chieffo C, Di Gregorio D, Di Mascio R, Franzosi MG, Geraci E, Levantesi G, Maggioni AP, Mantini L, Marfisi RM, Mastrogiuseppe G, Mininni N, Nicolosi GL, Santini M, Schweiger C, Tavazzi L, Tognoni G, Tucci C, Valagussa FG. ISSI-Prevenzione Investigators. Early protection against sudden death by n-3 polyunsaturated fatty acids after myocardial infarction: time-course analysis of the results of the Gruppo Italiano per lo Studio della Sopravvivenza nell’Infarto Miocardico (GISSI)-Prevenzione. Circulation. 2002;105:1897–1903. [PubMed]
44. Keli SO, Feskens EJ, Kromhout D. Fish consumption and risk of stroke. The Zutphen Study. Stroke. 1994;25:328–332. [PubMed]
45. Blok WL, Katan MB, van der Meer JW. Modulation of inflammation and cytokine production by dietary (n-3) fatty acids. J Nutr. 1996;126:1515–1533. [PubMed]
46. Mensink RP, Katan MB. Effect of dietary fatty acids on serum lipids and lipoproteins. A meta-analysis of 27 trials. Arterioscler Thromb. 1992;12:911–919. [PubMed]
47. Lopez GH, Ilincheta de Boschero MG, Castagnet PI, Giusto NM. Age-associated changes in the content and fatty acid composition of brain glycerophospholipids. Comp Biochem Physiol B Biochem Mol Biol. 1995;112:331–343. [PubMed]
48. Solfrizzi V, Capurso C, D’Introno A, Colacicco AM, Frisardi V, Santamato A, Ranieri M, Fiore P, Vendemiale G, Seripa D, Pilotto A, Capurso A, Panza F. Dietary fatty acids, age-related cognitive decline, and mild cognitive impairment. J Nutr Health Aging. 2008;12:382–386. [PubMed]
49. Calon F, Lim GP, Yang F, Morihara T, Teter B, Ubeda O, Rostaing P, Triller A, Salem N, Jr, Ashe KH, Frautschy SA, Cole GM. Docosahexaenoic acid protects from dendritic pathology in an Alzheimer’s disease mouse model. Neuron. 2004;43:633–645. [PMC free article] [PubMed]
50. Reaven PD, Grasse BJ, Tribble DL. Effects of linoleate-enriched and oleate-enriched diets in combination with alpha-tocopherol on the susceptibility of LDL and LDL subfractions to oxidative modification in humans. Arterioscler Thromb. 1994;14:557–566. [PubMed]