Total and Specific Polyphenol Intakes in Midlife Are Associated with Cognitive Function Measured 13 Years Later 1 3

Transcription

1 The Journal of Nutrition Nutritional Epidemiology Total and Specific Polyphenol Intakes in Midlife Are Associated with Cognitive Function Measured 13 Years Later 1 3 Emmanuelle Kesse-Guyot, 4 *Léopold Fezeu, 4 Valentina A. Andreeva, 4 Mathilde Touvier, 4 Augustin Scalbert, 5 Serge Hercberg, 4,6 and Pilar Galan 4 4 Unité de Recherche en Epidémiologie Nutritionnelle, U557 Inserm; U1125 Inra; Cnam; Université Paris 13, CRNH IdF, Bobigny, France; 5 International Agency for Research on Cancer, Nutrition and Metabolism Section, Biomarkers Group, Lyon, France; and 6 Département de Santé Publique, HôpitalAvicenne,Bobigny,France Abstract Polyphenols, and in particular flavonoids, are omnipresent plant-food components displaying biochemical properties possibly beneficial to brain health. We sought to evaluate the long-term association between total and class-specific polyphenol intake and cognitive performance. Polyphenol intake was estimated using the Phenol-Explorer database applied to at least six 24-h dietary records collected in as part of the SU.VI.MAX (Supplémentation en Vitamines et Minéraux Antioxydants) study. The cognitive performance of 2574 middle-aged adults participating in the cohort was assessed in using the following four neuropsychological tests: phonemic and semantic fluency, the RI-48 Cued Recall test, the Trail Making test, and Forward and Backward Digit Span. Inter-correlations among the test scores were estimated with principal component analysis. Associations between polyphenol intake and cognition were assessed by multivariate linear regression and ANCOVA. In multivariate models, high total polyphenol intake was associated with better language and verbal memory (P = 0.01) but not with executive functioning (P = 0.09). More specifically, intake of catechins (P = 0.001), theaflavins (P = 0.002), flavonols (P = 0.01), and hydroxybenzoic acids (P = ) was positively associated with language and verbal memory, especially with episodic memory assessed by the RI-48 test. In contrast, negative associations between scores on executive functioning and intake of dihydrochalcones (P = 0.01), catechins (P = 0.01), proanthocyanidins (P = 0.01), and flavonols (P = 0.01) were detected. High intake of specific polyphenols, including flavonoids and phenolic acids, may help to preserve verbal memory, which is a salient vulnerable domain in pathological brain aging. Further investigations are needed to clarify the observed negative associations regarding executive functioning. J. Nutr. 142: 76 83, Introduction As the elderly population is rapidly increasing, age-associated neurodegenerative disorders, including dementia, represent a major public health concern. The lack of curative treatment for cognitive decline and dementia argues for improvement of the prevention strategies via modifiable risk factors and behaviors, 1 Supported by the French National Research Agency (no. ANR-05-PNRA-010), the French Ministry of Health, Médéric, Sodexo,Ipsen,MGEN, andpierre Fabre. Mederic and Mutuelle Générale de l Education Nationale are French health insurance organizations, which are complementary to the National Health Insurance System. Ipsen and Pierre Fabre are private pharmaceutical companies. Sodexo, a food catering company, supported the study by organizing meetings between researchers and study participants. Sponsors were not involved in the analysis or interpretation of findings. 2 Author disclosures: E. Kesse-Guyot, L. Fezeu, V. Andreeva, M. Touvier, A. Scalbert, S. Hercberg, and P. Galan, no conflicts of interest. 3 Supplemental Tables 1 3 are available from the Online Supporting Material link in the online posting of the article and from the same link in the online table of contents at jn.nutrition.org. * To whom correspondence should be addressed. e.kesse@uren.smbh. univ-paris13.fr. such as dietary intake (1). Polyphenols represent a ubiquitous group of microcomponents contained in various plant-derived foods. Flavonoids and phenolic acids are the two most frequently consumed classes of polyphenols (2). Over the past few years, the potential of polyphenols to preserve brain health has become the subject of substantial research interest. Experimental studies suggest that these compounds, especially flavonoids, display neuroprotective effects, including enhancement of the neuronal function, stimulating brain flow, and inducing neurogenesis, and might prevent age-related damage to the central nervous system through their antioxidant properties and anti-inflammatory activities (3 6). However, such findings are based on specific phenolic compounds regardless of their bioavailability, level of consumption, or ability to cross the blood-brain barrier (7). Such deficiencies argue for the need for well-designed epidemiological studies carried out in human populations. Research on the relationship between the consumption of polyphenol-rich foods (8 12) and various cognitive outcomes has hypothesized a positive impact of these micronutrients on ã 2012 American Society for Nutrition. 76 Manuscript received May 11, Initial review completed July 26, Revision accepted October 3, First published online November 16, 2011; doi: /jn

2 cognition. A Dutch study showed a decreased risk of AD 7 among current smokers with high concentrations of flavonoid intake (13). Two other studies carried out in the same population reported a lower risk of dementia and cognitive decline among those with high flavonoid intake compared to their low-intake counterparts (14,15). The physiological properties, bioavailability, and thus potential effects on brain health may vary widely from one polyphenol compound to another, yet currently there are to our knowledge no specific research data regarding the different subtypes of flavonoids. Recently, the Phenol-Explorer database (16) has become available to the scientific community, providing data on a total of 502 different polyphenols (in aglycones, glycosides, and ester forms) from 452 different foods (17). In addition to total flavonoid composition, the database also includes detailed composition data for the various subtypes of flavonoids, i.e., flavonols, flavanols, flavones, flavonones, and catechines as well as phenolic acids, lignans, and stilbenes. Recent developments of the Phenol-Explorer system offer the opportunity for improved precision in estimating polyphenol intakes in the SU. VI.MAX cohort (2). The aim of the present study was to evaluate the hypothesized positive associations between the midlife level of intake of different types of polyphenols and cognitive function assessed 13 y later. Materials and Methods Study population. The SU.VI.MAX study ( ) was initially designed as a randomized, double-blind, placebo-controlled, primary prevention trial that included a total of 12,741 (7713 women aged y, 5028 men aged y) individuals for a planned follow-up of 8 y. The aim of the study was to test the potential efficacy of daily supplementation with antioxidant vitamins and minerals at nutritional doses [ascorbic acid (120 mg), vitamin E (30 mg), b-carotene (6 mg), selenium (100 mg), and zinc (20 mg)] regarding the incidence of cancer, CVD, and all-cause mortality (18,19). From the initial SU.VI.MAX cohort, a total of 6850 participants were included in the SU.VI.MAX 2 study ( ) aimed at evaluating quality of aging (20). The SU.VI. MAX and SU.VI.MAX 2 studies were conducted according to the guidelines laid down in the Declaration of Helsinki and were approved by the Ethics Committee for Studies with Human Subjects of Paris- Cochin Hospital (CCPPRB nos. 706 and 2364, respectively) and the Comité National Informatique et Liberté (CNIL nos and , respectively). Written informed consent was obtained from each participant. Dietary data collection. During the course of the SU.VI.MAX study, participants were invited to complete a 24-h dietary record every 2 mo for a total of 6 records/y. These dietary records were randomly distributed across 2 weekend days and 4 weekdays/y, so that each day of the week and all seasons were covered, thus accounting for individual variability in nutrient intake. Data were collected via computerized questionnaires using the Minitel, a small terminal used in France as an adjunct to the telephone. For the codification of foods, participants were assisted by using an instruction manual that included validated photographs of.250 generic foods represented in three main portion sizes. Participants could also choose from two intermediate or two extreme portions for a total of seven different portion sizes (21). French recipes validated by food and nutrition professionals were used to assess the amounts of specific nutrients consumed from composite dishes. The Phenol-Explorer database (17) and a published validated composition 7 Abbreviations used: AD, Alzheimer s disease; BP, blood pressure; CVD, cardiovascular disease; PCA, principal component analysis; SU.VI.MAX, Supplémentation en Vitamines et Minéraux Antioxydant; TMT, Trail Making test. table (22) were used to compute polyphenol and other nutrient intakes. Only polyphenol classes or subclasses with a mean daily intake.5 mg/d were considered in the present analysis (2). Cognitive assessment. Self-reported memory troubles (yes/no) were recorded at baseline (1994). In , all participants in SU.VI. MAX 2 were invited to undergo a medical check-up that included a clinical examination and a neuropsychological evaluation carried out by trained neuropsychologists. Episodic memory was evaluated using the RI-48, a delayed, cued recall test based on a list of 48 words belonging to 12 different categories. This test was designed to minimize ceiling effects encountered in some list-learning instruments. The score was the total number of words retrieved correctly (maximum score of 48) (23). Lexical-semantic memory was assessed via two verbal fluency tasks: a semantic fluency task, which consisted of naming as many animals as possible, and a phonemic fluency task consisting of citing words beginning with the letter P. The score was the total number of correct words produced over a 2-min period for each task (24). Working memory was assessed with the Forward and Backward Digit Span (25). Participants were asked to repeat two sequences of digits, forward or backward. The number of digits increased by one until the participant failed two consecutive trials of the same digit span. One point was scored for each correct sequence repeated, with a maximum score of 14 points for digit span forward as well as backward (25). Mental flexibility was assessed via the Delis-Kaplan TMT consisting of connecting numbers and letters, alternating between the two series. The score was the time in seconds needed to complete the task (26), implying that a lower value indicated better performance. For our analyses, we used the inverse of the TMT score; thus, a higher score corresponded to a better result. Covariates. At baseline, information on gender, date of birth, smoking status (never smoked, former or current smoker), physical activity (irregular, equivalent to,1 h walking/d, equivalent to at least 1 h walking/d), and education (primary, secondary, or university level) was collected. Anthropometric measurements were obtained at the first clinical examination ( ). Weight was measured using an electronic scale, with participants wearing indoor clothing and no shoes. Height was measured under the same conditions with a wall-mounted stadiometer. BP was measured with a mercury sphygmomanometer usi ng a standardized procedure. It was taken once from each arm in partici pants who had been lying down for at least 10 min. The mean of these two measurements was used in the analyses. Medication use was self-reported at baseline and the end of the follow-up. Fasting blood samples were obtained at baseline and the end of the follow-up and all bi ochemical measurements were centralized. Fasting blood glucose was measured using an enzymatic method (Advia 1650; Bayer Diagnostics). In case of suspected CVD during the follow-up period, relevant medi cal data (clinical, biochemical, histological, and radiological reports) were requested from participants, physicians, and/or hospitals. All reported CVD events were reviewed and validated by an independent expert committee. In SU.VI.MAX 2, depressive symptoms were assessed using the French version of the Center for Epidemiologic Studies Depression Scale (27). Statistical analyses. Among participants with available neuropsychological evaluation data aged between 45 and 60 y at enrollment, those with at least three 24-h dietary records from May to October and at least three records from November to April (thus accounting for seasonality and intra-individual variability in nutrient intake) provided during the first 2 y of follow-up were selected for inclusion in the present study. Among the 6850 adults included in the SU.VI.MAX 2 study, all cognitive tests were completed by a total of 4447 individuals aged y at baseline. Among them, 2819 participants had complete dietary intake data (i.e., at least six complete 24-h dietary records provided during the first 2 y of follow-up). From that subsample, we selected a total of 2574 participants without missing values on the covariables for inclusion in the present analysis. We modeled food and nutrient intakes as the respective mean values across all available 24-h records. Energy adjustment was performed using the residual method (28). s of energy-adjusted polyphenol Polyphenols and cognition 77

3 intake were calculated. BMI was calculated as the ratio of weight: squared height (kg/m 2 ). Hypertension at baseline was defined as systolic BP $140 mm Hg, diastolic BP $90 mm Hg, or self-reported antihypertensive drug use. Presence of diabetes was defined as blood glucose $7 mmol/l or self-reported antidiabetic drug use. To limit exclusion due to missing values, we excluded only participants with missing data for covariates with,5% missing. Thus, in case of missing clinical measurements for BMI, self-reported data were used when available. A missing category was used in case of missing values for hypertension status at baseline. The inverse of the TMT score was log-transformed to improve normality. PCA was performed to compute the inter-correlations among the various test scores, thereby maximizing the explained variance. Factors were rotated via an orthogonal transformation. A Cattel s scree plot (a plot of the total variance related to each pattern) was used to determine the number of patterns to be retained in the final model. For purposes of facilitating interpretability, the cognitive scores were converted into T scores ( , mean 6 SD). Thus, a 1-point difference in the test score corresponded to one-tenth of a SD difference. Mean polyphenol intakes are reported by gender. Descriptive baseline characteristics are reported as mean 6 SD or percentage across total polyphenol intake quartiles. The provided P values refer to the nonparametric Wilcoxon s rank test, the linear contrast test, or the trend chi-square test, as appropriate. ANCOVA were used for assessing the association between quartiles of total polyphenol and polyphenol subtype intakes and the cognitive scores. The initial model was adjusted for gender and age at the time of the cognitive evaluation. The final multivariate model was further adjusted for educational level, energy intake, number of 24-h dietary records, Western and healthy dietary patterns extracted using PCA (Supplemental Table 1), depressive symptoms concomitant with the cognitive performance assessment, and diabetes and CVD status during the follow-up as well as the following baseline covariates: memory troubles, alcohol intake, physical activity, intervention group (active vs. placebo), BMI, tobacco use, and hypertension. All tests of significance were 2-sided and the type I error was set at 5%. Statistical analyses were performed using SAS software (version 9.1, SAS Institute). Results Intake of polyphenols. Total crude polyphenol intakes were and g/d among men and women, respectively (Table 1). Flavonoid and hydroxycinnamic acids were the two main classes of polyphenols consumed, with intakes of and g/d, respectively, among men and and g/d, respectively, among women. The main food and beverage sources of polyphenols were coffee (36.8%), fruits (18.7%), wine (11.3%), and tea (9.4%) (data not tabulated). Participant characteristics. The mean age of the population at the time of the cognitive evaluation was 66 y. The baseline characteristics of the participants are shown across quartiles of total polyphenol intake (Table 2). Compared to participants in the bottom quartile, those in the highest quartile of polyphenol intake were more often men and smokers. Increasing polyphenol intake was significantly associated with a higher level of education, physical activity, and total energy intake as well as alcohol, protein, and micronutrient intakes and with a decreased likelihood of self-reported memory troubles. The percentage of men (P = 0.15) and women (P = 0.74) with memory troubles was similar across polyphenol intake quartiles (data not tabulated). Cognitive factors. Two main cognitive factors were extracted with PCA, accounting for 61% of the total initial variance in TABLE 1 Baseline intakes of total and subtypes of polyphenols by gender (SU.VI.MAX study, ) 1 Men (n = 1413) Women (n = 1161) P mg/d Total polyphenols , Flavonoids , Anthocyanins , Dihydrochalcones Dihydroflavonols , Catechins Theaflavins , Flavanones Flavones , Flavonols , Proanthocyanidins , Stilbenes , Lignans , Hydroxybenzoic acids Hydroxycinnamic acids , Sources of polyphenols Polyphenols from coffee , Polyphenols from fruit Polyphenols from wine , Polyphenols from chocolate Polyphenols from tea , Polyphenols from vegetables Polyphenols from bread , Polyphenols from potatoes , Data are mean 6 SD. SU.VI.MAX, Supplémentation en Vitamines et Minéraux Antioxydant. cognitive performance (Supplemental Table 2). The first factor accounted for 42% of the variance and reflected language and verbal memory abilities (e.g., memory for words and other abstractions involving language) (24). The test scores with the highest loadings on factor 1 pertained to the semantic (0.80) and phonemic fluency (0.66) tests and the RI-48 cued recall task (0.74). The second factor accounted for 19% of variance and reflected executive functioning. Test scores with the highest loadings were the Forward (0.85) and Backward (0.84) Digit Span tasks and, to a lesser extent, the TMT (0.50). Association between polyphenol intake and language and verbal memory (Factor 1). Results of the analyses of the association between energy-adjusted total and class-specific polyphenol intake and language and verbal memory are presented in Table 3. In the initial, age- and gender-adjusted model, intakes of total polyphenol, total flavonoids, anthocyanins, catechins, theaflavins, flavanones, flavonols, lignans, and hydroxybenzoic acids were positively associated with the language and verbal memory score. In the final multivariate model, the associations were somewhat attenuated but remained significant for total polyphenols, total flavonoids, catechins, theaflavins, flavonols, and hydroxybenzoic acids. Performance on the RI-48 test, measuring episodic memory, largely accounted for the observed relationships (Supplemental Table 3). Association between polyphenols intake and executive functioning (Factor 2). Results of the analyses of the association between energy-adjusted total and class-specific polyphenol intake and executive functioning are presented in Table 4.In 78 Kesse-Guyot et al.

4 TABLE 2 Baseline characteristics of the population across total polyphenol intake quartiles (SU.VI.MAX Study; ) 1 1 (#863 mg/d) 2 (.863 to #1145 mg/d) 3 (.1145 to #1487 mg/d) 4 (.1487 mg/d) P 2 Male, % , General characteristics Age at cognitive evaluation, y BMI, kg/m Physical activity, % Irregular ,1 h/d $1 h/d Education, % Primary Secondary University Smoking status, % Never smoker , Former smoker Current smoker Intervention group, % Dietary information 24-h records, n Energy intake, 3 kcal/d , Alcohol intake, g/d , Lipids, 4 % Proteins, 4 % , Carbohydrates, 4 % Fiber, g/d , Folic acid, mg/d , Vitamin C, mg/d , b-carotene, mg/d , Values are mean 6 SD or percent as appropriate, n = SU.VI.MAX, Supplémentation en Vitamines et Minéraux Antioxydant. 2 P values based on linear contrast or chi-squared trend test. 3 Energy intake excluding energy from alcohol. 4 Values are percentage of total daily energy intake (without alcohol). the initial, age- and gender-adjusted model, intakes of dihydrochalcones and flavanones were negatively associated with executive functioning scores. In the final multivariate model, the association with dihydrochalcones remained significant. In addition, increased intakes of catechines, proanthocyanidins, and flavanols were related to lower scores on the executive functioning factor. There was no association with total polyphenol intake. In-depth analyses indicated that intakes of dihydrochalcones, catechins, and flavanols were negatively associated with Forward Digit Span scores, whereas proanthocyanidin intake was negatively associated with Backward Digit Span scores (Supplemental Table 3). Discussion In this large prospective study, a positive association was observed between total polyphenol intake and cognitive factors reflecting language and verbal memory assessed after a follow-up of 13 y, even after accounting for various confounding factors, including vascular parameters, alcohol intake, and overall dietary patterns. More specifically, increasing intake of total flavonoids, catechins, theaflavines, flavonols, and phenolic acids was positively associated with long-term language and verbal memory capacities. In contrast, no beneficial association was detected between total polyphenol intake and a cognitive score reflecting executive functioning (e.g., working memory and mental flexibility). Unexpectedly, an adverse effect was found for some specific classes of polyphenols (catechins, proanthocyanidins, and flavonols). Overall, our findings suggest that the various polyphenol compounds may be differentially involved in the preservation of distinct cognitive domains. The clinical importance of our estimates may be evaluated according to educational level, which has consistently demonstrated an inverse association with cognitive function (1). In this study, the mean values on the verbal memory and executive functioning factors were 7.42 and 6.82 among those with high and low levels of education, respectively. Thus, a difference of 1 point on the language and verbal memory factor between the top and bottom quartiles of total polyphenol intake may be considered relatively low at the individual level, but this difference is probably important at the population level. Animal models support the positive effect of specific flavonoids on some aspects of cognition (6), in agreement with our findings regarding language and verbal memory. Some epidemiological studies are also in favor of a beneficial effect of consumption of polyphenol-rich foods on various cognitive outcomes (8 12). In a recent cross-sectional study carried out in Norway (12), better cognitive performance was observed among participants with increased consumption of chocolate, wine, and tea. However, that study did not exclude the possibility that other components of these foods, or the overall dietary patterns conducive to such intakes, may be responsible for the observed relationships. As underlined by these authors, Polyphenols and cognition 79

5 TABLE 3 Energy-adjusted associations between quartiles of total and class-specific polyphenol intakes and language and verbal memory (cognitive factor 1) (SU.VI.MAX and SU.VI.MAX 2 studies, ) 1 TABLE 3 Continued P P 2 Total Polyphenols Model Model Flavonoids Model Model Anthocyanins Model Model Dihydrochalcones Model Model Dihydroflavonols Model Model Catechins Model Model Theaflavins Model Model Proanthocyanidins Model Model Flavonones Model Model Flavones Model Model Flavonols Model Model Stilbenes Model Model Lignans Model Model Hydroxybenzoic acids Model (Continued) Model Hydroxycinnamic acids Model Model Values are adjusted mean 6 SEM of the language and verbal memory factor across quartiles of polyphenols intake. CVD, cardiovascular disease; PCA, principal component analysis; SU.VI.MAX, Supplémentation en Vitamines et Minéraux Antioxydant. 2 P-linear contrast across quartiles of polyphenols intake. 3 Model 1: adjusted for age (y) and gender. 4 Model 2: model 1 + energy intake (kcal), number of 24-h dietary records, education (primary/secondary/university) and intervention group (active vs. placebo), BMI (kg/m 2 ), tobacco use status (former/never/current), physical activity (irregular, equivalent to,1h walking/d, equivalent to $1 h walking/d), diabetes mellitus during follow-up (yes/no), baseline hypertension status (yes/no), CVD during follow-up (yes/no), CES D (Center for Epidemiologic Studies Depression Scale) score, self-reported memory troubles (yes/no) at baseline, and Western and healthy dietary patterns extracted via PCA. investigating foods rich in certain nutrients does not permit a full understanding of the specific effects of these nutrients. Three other studies have reported links between flavonoids and risk of dementia (13,15), cognitive performance, and cognitive decline (14). Two of these studies were carried out in the same population and supported a protective effect of flavonoid intake on cognitive outcomes (14,15). The third study observed a decreased risk of AD among participants with high intake of flavonoids, but these associations were significant only in smokers (13). These findings support the hypothesis of a potentially favorable effect of high intake of flavonoids on global cognitive function and prevention of AD. However, no studies to our knowledge have focused specifically on individual cognitive functions; indeed, cognitive domains have been suggested to be differentially affected during pathologic aging (29). Moreover, it has been suggested that nutritional factors exert site-specific effects within brain regions (30). For example, a cross-sectional study reported a positive association between tea consumption and cognitive functions other than attention (9). That finding is in agreement with those in the present study regarding the differential effects of nutritional factors according to specific cognitive function domains. Most existing mechanistic hypotheses concerning the link between polyphenol intake and brain health pertain to specific flavonoid properties. These compounds, especially when derived from fruit, may enhance memory and learning through their effects on synaptic plasticity, neuronal communication, and blood flow (31 33). For example, long-term memory and synaptic plasticity are modulated by the activity of genes regulated by the transcription factor CREB, which has been shown to be activated by flavonoids. Such mechanisms could help explain our findings regarding better language and verbal memory performance with increasing flavonoid intake (31). In addition, one in vitro study recently suggested that phenolic acids, especially caffeic acid, may have similar or greater neuroprotective effects than flavonoids against neurotoxic molecules involved in Parkinson s disease progression (34), but further investigations regarding hydroxybenzoic acids and memory are needed. To our knowledge, no molecular mechanisms accounting for the negative association between dihydrochalcones, catechins, proanthocyanidins, and flavonols and executive function have been advanced. However, it has been suggested that, under 80 Kesse-Guyot et al.

6 TABLE 4 Energy-adjusted associations between quartiles of total and class-specific polyphenol intakes and executive functioning (cognitive factor 2) (SU.VI. MAX and SU.VI.MAX 2 studies, ) 1 TABLE 4 Continued P P 2 Total polyphenols Model Model Flavonoids Model Model Anthocyanins Model Model Dihydrochalcones Model Model Dihydroflavonols Model Model Catechins Model Model Theaflavins Model Model Proanthocyanidins Model Model Flavanones Model Model Flavones Model Model Flavonols Model Model Stilbenes Model Model Lignans Model Model Hydroxybenzoic acids Model (Continued) Model Hydroxycinnamic acids Model Model Values are adjusted mean 6 SEM of executive functioning score across quartiles of polyphenol intake. CVD, cardiovascular disease; PCA, principal component analysis; SU.VI.MAX, Supplémentation en Vitamines et Minéraux Antioxydant. 2 P-linear contrast across quartiles of polyphenols intake. 3 Model 1: adjusted for age (y) and gender. 4 Model 2: model 1 + energy intake (kcal), number of 24-h dietary records, education (primary/secondary/university) and intervention group (active vs. placebo), BMI (kg/m 2 ), tobacco use status (former/never/current), physical activity (irregular, equivalent to,1 h walking/d, equivalent to $1 h walking/d), diabetes mellitus during follow-up (yes/no), baseline hypertension status (yes/no), CVD during follow-up (yes/no), CES D (Center for Epidemiologic Studies Depression Scale) score, self-reported memory troubles (yes/no) at baseline, and Western and healthy dietary patterns extracted via PCA. certain conditions, some catechins (derived from tea, fruits, wine, and to a lesser extent cocoa) may have dual effects on the brain through prooxidant action (35). Finally, experimental (30) and epidemiological (36) studies have reported that different foods or nutrients from various sources might exert region-specific actions within brain structures. Overall, further investigations aimed at elucidating the mechanisms of action of these components regarding different cognitive domains are warranted. Most published studies in this area have focused on total flavonoid intake, estimated with respect to the main flavonoid compounds rather than on class-specific flavonoids. For example, in the PAQUID study, flavonoid intakes were estimated via quercetin, kaempferol, myricetin, luteolin, and apigenin intake, thus possibly biasing the estimates, because other flavonoids were not taken into consideration. The Phenol-Explorer database is the first database of its kind that provides detailed and accurate information about a wide variety of polyphenols. Previously developed databases mainly focused on specific flavonoids, preventing a direct comparison of polyphenol intakes across studies. However, levels of such intakes have been shown to dramatically differ from one population to another. For example, flavonoid intake was mg/d in the PAQUID study (14), mg/d in the Rotterdam study (13), and mg/d among women and mg/d among men in our study. Such variability may stem from differences in the dietary patterns, under- or overestimation of food consumption, or methodological discrepancies regarding the assessment of nutrient composition with different food composition tables (2). Making food composition tables consistent with one another is crucial for accurately comparing intake levels among studies (2). Some limitations of our study should be mentioned. Because cognitive evaluation data at baseline were not available, we were unable to control for baseline levels of cognitive performance according to level of polyphenol intake. However, considering the baseline age range and general health status of the participants, it is unlikely that cognitive impairment was present at baseline. Moreover, self-reported memory troubles did not differ across level of polyphenol intake. Another limitation pertains to the measurement of dietary practices only at baseline without accounting for dietary changes over the years. If indeed dietary change had occurred, such participants nutrient intake might Polyphenols and cognition 81

7 have been misclassified, resulting in statistical power reduction. In an effort to minimize this potential source of confounding, we accounted for the presence of some comorbidities (CVD, diabetes) that often necessitate dietary modification. Furthermore, some inaccuracies regarding polyphenol intake may have occurred as a result of disparities in terms of food composition data quality when computing intakes from self-reports. As previously reported (2,17), some food components (e.g., proan thocyanidin) are insufficiently addressed in existing food composition tables. In addition, due to the restricted number of sources of some classes of polyphenols (e.g., dihydrochalcone from apples, dihydroflavols from wine, theaflavins from tea, flavanones from oranges, hydroxycinnamic acids from coffee), it is possible that the observed associations between these particular compounds and cognitive function may be due to another component of the vector foods. For more ubiquitous polyphenols such as catechins (mostly provided by fruits, wine, tea, and chocolate) and flavonols (mostly provided by fruits, wine, tea, and vegetables), such confounding and differential effects according to the sources need future in-depth investigations. Additionally, the analyzed sample was likely overselected, because participants with available dietary, covariate, and cognitive evaluation data may have been particularly compliant or health conscious. Thus, caution is needed when generalizing our findings to other populations. Finally, because many associations were evaluated, we cannot rule out the possibility of observing significant findings by chance. In turn, the absence of other associations may be due to lack of statistical power, especially for phenolic compounds with a very narrow range of intake. Further studies are needed to confirm our findings. Original aspects of our study include its longitudinal design and dietary data reflecting midlife exposure. Indeed, our findings are of major interest from a public health viewpoint, because the prevention of cognitive decline is a cost-effective strategy and dementia prevention should be initiated in middle age when potential cognitive disorders are presymptomatic (37,38). Additional strengths of the present study include cognitive evaluation in a relatively young population and the use of a neuropsychological battery of sensitive tests to avoid ceiling and floor effects. Furthermore, diet was assessed via at least six repeated 24-h records, providing comprehensive and relatively accurate dietary information and accounting for intra-individual variability. Moreover, to our knowledge, this is the first study in which polyphenol intake was evaluated in a general population with a relatively high level of precision and the first to investigate a substantial number of polyphenol compounds. Preservation of cognitive function remains a major public health concern, because no curative treatment for dementia is currently available and even small delays in the onset of cognitive decline may produce a sizable effect at the population level. Thus, polyphenols may be of major public health interest. To our knowledge, this study is the first to report that increased midlife consumption of some subtypes of polyphenols is associated with better language and verbal memory capacities, which constitute cognitive domains particularly vulnerable to dementia. Associations with executive functioning are less clear-cut and require further investigation. Because the relevant literature is scarce, further research is needed to confirm these findings in other populations and settings. Acknowledgments The authors especially thank Stéphane Raffard, neuropsychologist, who was responsible for standardization of the cognitive evaluation; Nathalie Arnault (statistician), who coordinated data management; Frédérique Ferrat, who coordinated the logistic aspects of the neuropsychological evaluation; and Gwenaël Monot (computer scientist), who coordinated the computing aspects. E.K.G. carried out data checking and analyses and was responsible for drafting the manuscript; V.A.A., L.F., A.S., S.H., M.T., and P.G. were involved in interpreting results and editing the manuscript; and E.K.G., P.G., and S.H. were responsible for developing the design and protocol of the study. All authors read and approved the final manuscript. Literature Cited 1. Middleton LE, Yaffe K. Promising strategies for the prevention of dementia. Arch Neurol. 2009;66: Pérez-Jiménez J, Fezeu L, Touvier M, Arnault N, Manach C, Hercberg S, Galan P, Scalbert A. Dietary intake of 337 polyphenols in French adults. Am J Clin Nutr. 2011;93: Darvesh AS, Carroll RT, Bishayee A, Geldenhuys WJ, Van der Schyf CJ. Oxidative stress and Alzheimer s disease: dietary polyphenols as potential therapeutic agents. Expert Rev Neurother. 2010;10: Rossi L, Mazzitelli S, Arciello M, Capo CR, Rotilio G. Benefits from dietary polyphenols for brain aging and Alzheimer s disease. Neurochem Res. 2008;33: Vauzour D, Vafeiadou K, Rodriguez-Mateos A, Rendeiro C, Spencer JP. The neuroprotective potential of flavonoids: a multiplicity of effects. Genes Nutr. 2008;3: Spencer JP, Vauzour D, Rendeiro C. Flavonoids and cognition: the molecular mechanisms underlying their behavioural effects. Arch Biochem Biophys. 2009;492: Manach C, Williamson G, Morand C, Scalbert A, Remesy C. Bioavailability and bioefficacy of polyphenols in humans. I. Review of 97 bioavailability studies. Am J Clin Nutr. 2005;81:S Dai Q, Borenstein AR, Wu Y, Jackson JC, Larson EB. Fruit and vegetable juices and Alzheimer s disease: the Kame Project. Am J Med. 2006;119: Feng L, Gwee X, Kua EH, Ng TP. Cognitive function and tea consumption in community dwelling older Chinese in Singapore. J Nutr Health Aging. 2010;14: Ng TP, Feng L, Niti M, Kua EH, Yap KB. Tea consumption and cognitive impairment and decline in older Chinese adults. Am J Clin Nutr. 2008;88: Kuriyama S, Hozawa A, Ohmori K, Shimazu T, Matsui T, Ebihara S, Awata S, Nagatomi R, Arai H, Tsuji I. Green tea consumption and cognitive function: a cross-sectional study from the Tsurugaya Project 1. Am J Clin Nutr. 2006;83: Nurk E, Refsum H, Drevon CA, Tell GS, Nygaard HA, Engedal K, Smith AD. Intake of flavonoid-rich wine, tea, and chocolate by elderly men and women is associated with better cognitive test performance. J Nutr. 2009;139: Engelhart MJ, Geerlings MI, Ruitenberg A, van Swieten JC, Hofman A, Witteman JC, Breteler MM. Dietary intake of antioxidants and risk of Alzheimer disease. JAMA. 2002;287: Letenneur L, Proust-Lima C, Le Gouge A, Dartigues JF, Barberger- Gateau P. Flavonoid intake and cognitive decline over a 10-year period. Am J Epidemiol. 2007;165: Commenges D, Scotet V, Renaud S, Jacqmin-Gadda H, Barberger- Gateau P, Dartigues JF. Intake of flavonoids and risk of dementia. Eur J Epidemiol. 2000;16: Phenol-Explorer database. Internet; 2011 [cited 2011 Jan 15]. Available from: Neveu V, Perez-Jimenez J, Vos F, Crespy V, du Chaffaut L, Mennen L, Knox C, Eisner R, Cruz J, et al. Phenol-Explorer: an online comprehensive database on polyphenol contents in foods. Database (Oxford); 2010 [cited 16 December 2010]. Available from: Hercberg S, Galan P, Preziosi P, Roussel AM, Arnaud J, Richard MJ, Malvy D, Paul-Dauphin A, Briancon S, Favier A. Background and rationale behind the SU.VI.MAX Study, a prevention trial using nutritional doses of a combination of antioxidant vitamins and minerals to reduce cardiovascular diseases and cancers. SUpplementation en VItamines et Mineraux AntioXydants Study. Int J Vitam Nutr Res. 1998;68: Kesse-Guyot et al.

8 19. \Hercberg S, Galan P, Preziosi P, Bertrais S, Mennen L, Malvy D, Roussel AM, Favier A, Briancon S. The SU.VI.MAX Study: a randomized, placebo-controlled trial of the health effects of antioxidant vitamins and minerals. Arch Intern Med. 2004;164: Kesse-Guyot E, Amieva H, Castetbon K, Henegar A, Ferry M, Jeandel C, Hercberg S, Galan P. Adherence to nutritional recommendations and subsequent cognitive performance: findings from the prospective Supplementation with Antioxidant Vitamins and Minerals 2 (SU.VI.MAX 2) study. Am J Clin Nutr. 2011;93: Le Moullec N, Deheeger M, Preziosi P, Montero P, Valeix P, Rolland- Cachera MF, Potier de Courcy G, Christides JP, Galan P, Hercberg S. Validation du manuel photos utilisé pour l enquête alimentaire de l étude SU.VI.MAX. Cahier de Nutrition et de Diététique. 1996;31: Hercberg S. Table de composition SU.VI.MAX des aliments. Paris: Les éditions INSERM/Economica; Ivanoiu A, Adam S, Van der Linden M, Salmon E, Juillerat AC, Mulligan R, Seron X. Memory evaluation with a new cued recall test in patients with mild cognitive impairment and Alzheimer s disease. J Neurol. 2005;252: Lezak MD, Howieson DB, Loring DW. Neuropsychological assessment. 4th ed. New York: Oxford University Press; Wechsler D. Wechsler Adult Intelligence Scale-revised. New York: Psychological Corporation; Delis DC, Kaplan E, Kramer JH. Delis-Kaplan Executive Function System (D-KEFS) examiner s manual. San Antonio (TX): The Psychological Corporation; Radloff L. The CES-D Scale: a self-report depression scale for research in the general population. Appl Psychol Meas. 1977;1: Willett WC. Nutritional epidemiology. 2nd ed. New York: Oxford University Press; Amieva H, Le Goff M, Millet X, Orgogozo JM, Peres K, Barberger- Gateau P, Jacqmin-Gadda H, Dartigues JF. Prodromal Alzheimer s disease: successive emergence of the clinical symptoms. Ann Neurol. 2008;64: Shukitt-Hale B, Carey AN, Jenkins D, Rabin BM, Joseph JA. Beneficial effects of fruit extracts on neuronal function and behavior in a rodent model of accelerated aging. Neurobiol Aging. 2007;28: Spencer JP. Flavonoids and brain health: multiple effects underpinned by common mechanisms. Genes Nutr. 2009;4: Spencer JP. Flavonoids: modulators of brain function? Br J Nutr. 2008; 99 E Suppl 1:ES Spencer JP. Beyond antioxidants: the cellular and molecular interactions of flavonoids and how these underpin their actions on the brain. Proc Nutr Soc. 2010;69: Vauzour D, Corona G, Spencer JP. Caffeic acid, tyrosol and p-coumaric acid are potent inhibitors of 5-S-cysteinyl-dopamine induced neurotoxicity. Arch Biochem Biophys. 2010;501: Mandel SA, Avramovich-Tirosh Y, Reznichenko L, Zheng H, Weinreb O, Amit T, Youdim MB. Multifunctional activities of green tea catechins in neuroprotection. Modulation of cell survival genes, irondependent oxidative stress and PKC signaling pathway. Neurosignals. 2005;14: Sabia S, Nabi H, Kivimaki M, Shipley MJ, Marmot MG, Singh- Manoux A. Health behaviors from early to late midlife as predictors of cognitive function: The Whitehall II study. Am J Epidemiol. 2009;170: de la Torre JC. Alzheimer s disease is incurable but preventable. J Alzheimers Dis. 2010;20: Mortimer JA, Borenstein AR, Gosche KM, Snowdon DA. Very early detection of Alzheimer neuropathology and the role of brain reserve in modifying its clinical expression. J Geriatr Psychiatry Neurol. 2005; 18: Polyphenols and cognition 83