| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
ORIGINAL RESEARCH COMMUNICATION |
1 From the Department of Pharmacology (EN) and the Department of Physiology, Anatomy and Genetics and the Oxford Project to Investigate Memory and Ageing (HR and ADS), University of Oxford, Oxford, United Kingdom; the Institute of Basic Medical Science, Department of Nutrition, University of Oslo, Oslo, Norway (EN, CAD, HR and KS); the Department of Public Health and Primary Health Care (SEV, HAN, and GST) and the Institute of Medicine (ON), University of Bergen, Bergen, Norway; and the Department of Geriatric Medicine, Norwegian Centre for Dementia Research, Ullevål University Hospital, Oslo, Norway (KE)
See corresponding editorial on page 1259.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Objective: The objective was to examine the cross-sectional relation between intake of different amounts of various seafood (fish and fish products) and cognitive performance.
Design: The subjects (n = 2031 subjects; 55% women), aged 70–74 y, were recruited from the general population in Western Norway and underwent cognitive testing. A cognitive test battery included the Kendrick Object Learning Test, Trail Making Test (part A), modified versions of the Digit Symbol Test, Block Design, Mini-Mental State Examination, and Controlled Oral Word Association Test. Poor cognitive performance was defined as a score in the highest decile for the Trail Making Test and in the lowest decile for all other tests.
Results: Subjects whose mean daily intake of fish and fish products was 
10 g/d (n = 1951) had significantly better mean test scores and a lower prevalence of poor cognitive performance than did those whose intake was <10 g/d (n = 80). The associations between total intake of seafood and cognition were strongly dose-dependent; the maximum effect was observed at an intake of 
75 g/d. Most cognitive functions were influenced by fish intake. The effect was more pronounced for nonprocessed lean fish and fatty fish.
Conclusions: In the elderly, a diet high in fish and fish products is associated with better cognitive performance in a dose-dependent manner.
Key Words: Cognitive deficit • cognition • elderly • fish • fish oils • processed fish • seafood
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The purpose of this cross-sectional study on community-dwelling elderly people was to seek answers to the following questions: 1) Is the intake of fish and fish products associated with better cognitive performance? 2) Is dietary intake of seafood associated with the performance of all, or only some, cognitive abilities? 3) Is there any relation between the type of seafood product consumed and cognitive test results? and 4) Is there any relation between the amount of fish consumed and cognitive test results? We had the opportunity to examine these issues among >2000 elderly men and women, who represented a subset of subjects within the Hordaland Health Study (HUSK).
|
SUBJECTS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Data collection
Cognitive testing was performed at the study location by trained nurses after the standard cardiovascular examinations of the National Health Screening Service (17) were completed. The cognitive test battery included 6 tests.
Kendrick Object Learning Test
The Kendrick Object Learning Test (KOLT; episodic memory) (18) is designed to assess dementia status and memory performance among noninstitutionalized elderly. The maximum possible KOLT score is 70.
Trail Making Test
The Trail Making Test, part A (TMT-A; executive function) (19) is a test of visual conceptual and visuomotor tracking, involving motor speed and attention functions. The score is the total time in seconds to complete the items. The shortest and longest times in our population were 16 and 331 s.
Modified version of the Digit Symbol Test
The modified version of the Digit Symbol Test (m-DST; perceptual speed) (20) is regarded as a measure of focused attention, visuomotor coordination, and psychomotor speed. In the present version, the number of correct matches between digits and symbols in 30 s was recorded.
Block Design
The Block Design (m-BD; visuospatial skills) (20) tests visuospatial and motor skills. The short form, used in the present study, included 4 of the 10 patterns (patterns 1, 2, 5, and 6) in the full test. Every correct matching gives 4 points; thus, a possible maximum score on the m-BD short form is 16.
Modified version of the Mini-Mental State Examination
A modified version of the Mini-Mental State Examination (m-MMSE; global cognition) (21) covers various aspects of cognitive function, including orientation to time and place, naming, repeating, writing, copying, instantaneous recall, short-term memory, backward spelling, and performing a 3-stage command (22). The m-MMSE consists of 12 of the 20 items in the full version of the MMSE and has been shown to be just as effective as the full version when the purpose is to identify elderly subjects with cognitive impairment (22).
Abridged version of the Controlled Oral Word Association Test
The abridged version (S-task) of the Controlled Oral Word Association Test (access to semantic memory) (23) is a test of verbal fluency and psychomotor speed. The subjects were required to generate as many words as possible beginning with the letter S within 60 s. The maximum possible score is theoretically infinite; in our subjects it was 39.
Dietary habits
To assess habitual food consumption, a modified version of a comprehensive food-frequency questionnaire created at the Department of Nutrition, University of Oslo (24, 25), was handed out on the day of the examination and filled out later at home by the participants and then mailed to the HUSK Project Centre in Bergen. The questionnaire included 169 food items that were grouped according to Norwegian meal patterns. It was designed to obtain information on usual food intake during the past year. The frequency of consumption was given per day, week, or month. The portion sizes were given as household measures or units such as slices or pieces.
Fish intake included sandwich spread and fish intake as a part of a main meal. Questions related to sandwich spread were as follows: How many slices of bread with the following spreads do you eat per week: tinned mackerel in tomato paste, smoked mackerel, pilchard, pickled herring, anchovies or similar fish, salmon, or trout? All these questions had 11 categories (0, 0.5, 1, 2–3, 4–5, 6–7, 8–14, 15–21, 22–28, 29–35, and 36 slices/wk). Questions related to fish intake as a main meal were 2-fold, related to frequency of intake and to the typical amount of intake per meal. The frequency of fish intake included 9 categories (0, <1, 1, 2, 3, 4, 5–6, 7–8, and 9 times/mo). The amount of fish intake was categorized into 5 categories that differed according to type of product: fishcakes, fish pudding, and fish balls (from 1 to 5 units); fish fingers (from 1–2 to 10 pieces); boiled cod, coalfish, and haddock (from 1 to 5 pieces); fried cod, coalfish, or haddock (from 1 to 5 pieces); fresh, salt-cured, or smoked herring (from 1 to 5 fillets); fresh or smoked mackerel (from 0.5 to 3 fillets); salmon or trout (both wild and farmed) (from 1 to 5 steaks); and fish stew, fish soup, and fish au gratin (from 1–2 to 9 dl). In the data analyses, the main meal was divided into 3 different categories: fatty fish (herring, mackerel, salmon, and trout), lean fish (cod, coalfish, and haddock), and processed fish (fish cakes, fish pudding, fish balls, fish fingers, fish stew, fish soup, and fish au gratin). The last category, "processed fish," is imprecise in terms of nutritional value.
The questionnaire also included questions about dietary supplement intake, in which the product names of the most used supplements in Norway were considered. Use of cod liver oil and fish oil was reported as "seasonal use" (during the whole year or only winter half of the year), frequency per week, and amount per time.
Dichotomous variables were created for different types of fish and fish products. Individuals who reported use of certain products in the abovementioned categories were considered as users, whereas all those who reported that they never ate that product were considered nonusers. To identify the individuals who never ate any fish and fish products (including fish either as sandwich spread or as part of main meals, cod liver oil, and other fish oils), all dichotomous variables were combined. The amount of fish and fish products in grams per day was calculated by using a food database and software system developed at the Department of Nutrition, University of Oslo (Kostberegningssystem, version 3.2; University of Oslo, Oslo, Norway). Calculations of total intake of fish and fish products also included intake of cod liver oil and other fish oils. For convenience, we use the term seafood to include all the fish and fish products described above.
Other variables
Self-reported information on history of myocardial infarction, angina pectoris, stroke, thrombosis, phlebitis, and hypertension was recorded in 1992–1993 and in 1997–1999. On the basis of information from both surveys, the subjects were categorized as with or without a history of cardiovascular disease (CVD), including the diseases and conditions mentioned above. About four-fifths (79.3%) of self-reported CVD cases and noncases were validated on the basis of hospitalizations records used in our earlier study (26), whereas the remaining 20% of CVD cases were presumably less severe and did not require hospitalization or occurred before 1992.
Educational level was self-reported and recorded in 5 categories: primary school (
9 y), vocational secondary school (10–12 y), theoretical secondary school (10–12 y), college or university <4 y, and university of 4 y.
Nonfasting EDTA blood samples were collected for the measurement of total homocysteine (tHcy) and gene polymorphisms. The EDTA sample was kept cool until centrifuged. The blood samples were stored at –80 °C. The duration of storage ranged from a few days to 18 mo. Plasma tHcy was measured by using a fully automated HPLC assay (27, 28). Apolipoprotein E genotypes were determined in the packed cell fraction of blood samples by using a one-stage polymerase chain reaction method (29).
Statistical analysis
Cutoff points for poor cognitive test scores were set at about the 10th percentile of the cognitive test score, except for TMT-A, for which the 90th percentile was used. Preliminary analyses showed that cognitive test performance was most frequently associated with education (all test scores were significantly associated). Three of 6 cognitive test scores were significantly associated with tHcy, the KOLT and m-DST were significantly associated with sex and ApoE 
4 variant allele, and the KOLT score was the only test that was significantly associated with some important CVD variables referred to above. Intake of fish or fish products was significantly associated with sex, education, and tHcy. Body mass index, total cholesterol, and HDL cholesterol were not associated with cognitive test scores or with fish and fish product intakes, and inclusion of these variables in the statistical models did not change the results and were therefore excluded from final models. We also adjusted for depression score using a 7-item self-administered subscale for depression from the Hospital Anxiety and Depression Scale (30), which focuses mainly on the reduced pleasure response aspect (anhedonia) of depression, but also psychomotor retardation and impaired mood. Although the depression score was significantly associated with intake of seafood and with most cognitive test scores, it made little difference when introduced in the statistical models. Because inclusion of the depression score significantly reduced the numbers of subjects because of missing data, the results are not shown. The participants of the cognitive substudy were homogeneous in age (born 1925–1927); therefore, age was not included as a cofactor in the statistical models. Throughout the article, we analyzed the results using 2 different models: a simple model, in which we only adjusted for sex, and a multiple adjusted model, in which sex, ApoE 4 variant allele, education, history of CVD (limited to some important CVD variables), and tHcy were cofactors.
For comparison between the groups of fish and fish product intake, the chi-square test or the univariate analysis of variance was used. Adjusted mean cognitive scores by intake of various types of seafood were obtained from the univariate analysis of variance. Multinominal logistic regression analysis (adjusted for the same variables) with fish and fish products intake as the independent variable, and the number of tests with poor results as the dependent variable, was applied to find risk ratios for poor cognitive performance by using subjects without poor test performance as the reference group. Gaussian generalized additive regression models, as implemented in S-PLUS 6.2 for WINDOWS (Insightful Corporation, Seattle, WA), were used to generate graphic representations of the dose-response relations, with the use of both models. On the y axis, the model generated a reference value of zero that approximately corresponds to the value of cognitive test score associated with the mean of the average total fish intake in grams per day for all subjects. Multiple linear regression analyses were used to examine significant associations between the cognitive test scores and average of total fish intake. Except for generalized additive models, all statistical analyses were performed by using the STATISTICAL PACKAGE FOR THE SOCIAL SCIENCES 12.0 for WINDOWS (SPSS Inc, Chicago, IL). P values <0.05 were considered significant.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
). The mean total intake of fish and fish products among consumers was 85 g/d. About four-fifths of the population ate fatty fish as part of a main course, whereas even more reported having lean fish and processed fish for a main course. Nearly 2 of 3 subjects consumed fish as a sandwich spread, and >2 of 5 reported intake of cod liver oil or other fish oils.
|
. The proportions of women and subjects with low education (9 y) were significantly higher among subjects who consumed <10 g/d of any type of fish or fish product. The mean scores on all 6 cognitive tests were significantly better for those who ate any type of fish or fish product. The prevalence of poor cognitive performance was 2–3 times lower in those who reported eating seafood than in those who did not report eating seafood, independent of the cognitive test used. The proportion of subjects with poor scores on more than one cognitive test was greater in subjects who did not eat fish and fish products (27%) than in those who did (11%; P < 0.001).
|
4 variant allele, education, history of some important CVD variables, and tHcy, the overall results changed moderately (data not shown), but because of missing data the significance was reduced for mean test scores (Table 2
). Compared with users of fish and fish products, the nonusers had a somewhat poorer health status. The prevalence of several diseases was significantly higher among those who did not eat fish and fish products than among those who did (epilepsy: 5.5% compared with 0.8%, P = 0.015; asthma: 19.5% compared with 9.1%, P = 0.005; chronic bronchitis: 14.5% compared with 6.7%, P = 0.036; and osteoporosis: 20.3% compared with 11.0%, P = 0.034). We found no significant differences in the prevalence of history of some important CVD variables, diabetes, renal diseases, liver diseases, arthritis, thyroiditis, depression, psoriasis, or some other skin diseases between those who reported eating and those not eating fish and fish products (data not shown).
Cognitive scores in relation to type of fish and fish products
We studied the associations between cognitive performance and the main types of consumed seafood, ie, fatty fish, lean fish, processed fish, fish sandwich, and fish or cod liver oil. The adjusted mean cognitive test scores by different types of fish and fish products intake are presented in Table 3. Subjects who ate fatty fish or lean fish as their main meal performed significantly better in 5 of the 6 tests than did those who did not eat fatty fish or lean fish. Intake of processed fish as part of a main meal was associated with significantly better mean scores in 3 cognitive tests: KOLT, m-BD, and m-MMSE. Similarly, those who ate fish sandwiches performed better on 3 tests (m-DST, m-BD, and S-task). In contrast, those who ate fish oils performed better only in the S-task.
|
). Linear regression analyses adjusted only for sex indicated that dose-response associations were significant for all cognitive tests. In the multiple-adjusted analyses, most of the relations remained significant (KOLT, m-DST, m-BD, and m-MMSE).
|
. A diet including fish or fish products was associated with a low risk of having 4 tests with poor scores. The risk decreased by as much as 80% when all types of fish and fish products were combined.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Prevalence of participants who reported eating seafood
The proportion of participants who reported eating fish and fish products and the mean daily fish intake were higher in our population than in populations in previous studies (5, 11, 12). Despite the relatively low fish intake in previous studies, all suggested that fish eaters exhibit better cognitive performance than do noneaters (5, 11, 12). Our study, in a population who reported eating a lot of fish, allowed us to investigate the associations between fish intake and cognition variables in more detail.
Intake of seafood and cognitive test performance
We found that fish eaters had significantly better results on all cognitive tests than did nonconsumers. Most of these associations remained significant after adjustment for several nonnutritional factors (sex, ApoE 4 variant allele, education, history of some important CVD variables, and tHcy) with a proven association with cognition. Moreover, the prevalence of poor scores, independent of the cognitive test used, was significantly lower among those who reported eating than in those who reported not eating fish and fish products.
The population-based Rotterdam Study was one of the first large prospective studies to show that fish consumption is associated with a reduced risk of dementia, particularly of AD (4). The favorable effect of fish consumption was observed at a relatively low intake (from 20 g/d) (4, 5). Other prospective studies (6-8) showed that subjects who consumed seafood at least once a week had a significantly lower risk of developing dementia or AD in the following 4–7 y. In contrast, an evidence-based study (31) focusing on the role of n–3 fatty acids reported that fish consumption was only weakly associated with a reduced risk of cognitive impairment and had no association with cognitive decline.
The associations between different types of fish and impaired cognitive outcome have been little studied. Huang et al (9) observed that the consumption of fatty fish was associated with a greater reduction in risk of dementia and AD than was eating lean fish. In the present study, the effects of fatty and lean fish intake on cognitive test performance were comparable. However, when the significance level was restricted to P < 0.001 in the multiple-adjusted model, only the association between lean fish intake and the KOLT remained.
In contrast with another study (32), we observed that intake of fish oils had a beneficial effect only on the S-task. In the groups who reported consuming processed fish and fish sandwiches, the associations were inconsistent, possibly because cooking and preparation methods influence the nutritive value of the meal (9). Thus, our most consistent finding was related to total intake of fish and fish products. In line with this, we found that those who reported consuming 10 g seafood/d were protected against having multiple tests with poor scores. A reduction of up to 80% in the risk of having poor scores on 4 tests was found in those who reported eating 10 g seafood/d compared with those who reported eating less or no seafood.
Dose-response relation
Although few subjects never ate seafood, there was a wide span in intake. Using dose-response curves, we observed that the performance on all 6 of the cognitive tests, adjusted for sex, improved with increasing total fish intake up to 70–80 g/d and reached a plateau thereafter. The dose-response relations remained significant in all except the TMT-A and S-task after multiple adjustments, which indicated the robustness of these associations. A dose-response relation was suggested in an earlier study (9), in which the hazard ratio for dementia decreased with the number of servings per week of fatty fish.
Cognitive test abilities
The protective effect of eating fish and fish products extended to almost all of the tested cognitive abilities. Notably, the test of global cognitive ability, the m-MMSE, was frequently not affected by individual types of seafood. However, as with the other tests, a strong dose-response relation was found with the combined intake of any type of fish or fish oil. Few other studies have examined fish intake in relation to specific cognitive abilities in the elderly, but one report described a positive effect of eating fatty fish on processing speed, but not on tests of memory or flexibility (11). Another study found that the BD and the DST scores were higher in consumers of fish oil (32).
Strengths and limitations of the study
The strengths of this study included a large population-based sample with a relatively high consumption of fish and fish products and inclusion of 6 different tests to study cognitive performance. The food-frequency questionnaire used was validated in several studies, including the correlation between self-reported dietary intake of fish and essential n–3 fatty acids in plasma phospholipids among 579 men and women (25) and 14-d weighed diet records with the intakes calculated from the food-frequency questionnaire in a group of 38 elderly women (24).
One limitation of dietary studies is errors in estimates of nutrients (11, 14, 31). Thus, it is possible to over- or underestimate true associations with outcomes. In addition, lack of a uniform quantification of fish intake makes comparisons between different studies difficult.
Because 77.3% of the study participant volunteered for cognitive testing, recruitment bias may have been an issue. Several differences between those who underwent and who did not undergo cognitive testing were reported earlier (33). However, the dietary habits were similar, and the difference in total fish intake was only 0.3 g/d (P = 0.91).
Cognition in the elderly is shaped by long-term exposures (34, 35). Thus, an important limitation of this study was its cross-sectional design. Furthermore, subjects with impaired cognition may have altered their diet as a consequence of a change in their cognitive status, although the direction of such a change is not predictable. In addition, self-reported dietary data collected from subjects who are cognitively impaired or demented may be less reliable. However, participants in the present study were not seriously impaired, and we do not believe this to have a major effect on our findings. Last but not least, foods are not consumed individually but as part of a diet; therefore, confounding by other food items is always an issue in studies using dietary assessments.
Conclusion
In a population-based study, we showed that intake of fish and fish products is associated with better performance across several cognitive abilities and that the associations are strongly dose-dependent. We also observed that the effect depends on the type of fish consumed. Our data indicate the need for additional studies in which more details about the type or species of fish and methods of preparation should be taken into account. Nevertheless, because this and earlier studies have shown that fish intake is associated with better cognition, the next question is what component of fish makes it good for the brain? Studies of n–3 fatty acids (36), niacin (37), and any other factor known to be enriched in fish are needed to answer this question.
|
ACKNOWLEDGMENTS |
|---|
The authors' responsibilities were as follows—HR, SEV, and GST: participated in the study design and the organization of data collection; KE, HAN, and ADS: assisted with the design and organization of the cognitive substudy; KS: assessed food intakes and helped develop the food-frequency questionnaire; CAD: helped develop the food-frequency questionnaire; EN: conducted the statistical analysis and wrote the first draft of the manuscript; and EN, CAD, HR, KS, SEV, ON, HAN, KE, GST, and ADS: interpreted the results and contributed to the study design and the writing of the paper. None of the authors had any financial conflicts of interest.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
4. Neurology 2005;65:1409–14.Related articles in AJCN:
This article has been cited by other articles:
|
|
 |
|
  E. Albanese, A. D Dangour, R. Uauy, D. Acosta, M. Guerra, S. S G. Guerra, Y. Huang, K. Jacob, J. Llibre de Rodriguez, L. H. Noriega, et al. Dietary fish and meat intake and dementia in Latin America, China, and India: a 10/66 Dementia Research Group population-based study Am. J. Clinical Nutrition, August 1, 2009; 90(2): 392 - 400. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Vogiatzoglou, A D. Smith, E. Nurk, P. Berstad, C. A Drevon, P. M Ueland, S. E Vollset, G. S Tell, and H. Refsum Dietary sources of vitamin B-12 and their association with plasma vitamin B-12 concentrations in the general population: the Hordaland Homocysteine Study Am. J. Clinical Nutrition, April 1, 2009; 89(4): 1078 - 1087. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Nurk, H. Refsum, C. A. Drevon, G. S. Tell, H. A. Nygaard, K. Engedal, and A. D. Smith Intake of Flavonoid-Rich Wine, Tea, and Chocolate by Elderly Men and Women Is Associated with Better Cognitive Test Performance J. Nutr., January 1, 2009; 139(1): 120 - 127. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Whelan (n-6) and (n-3) Polyunsaturated Fatty Acids and the Aging Brain: Food for Thought J. Nutr., December 1, 2008; 138(12): 2521 - 2522. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Samieri, C. Feart, L. Letenneur, J.-F. Dartigues, K. Peres, S. Auriacombe, E. Peuchant, C. Delcourt, and P. Barberger-Gateau Low plasma eicosapentaenoic acid and depressive symptomatology are independent predictors of dementia risk Am. J. Clinical Nutrition, September 1, 2008; 88(3): 714 - 721. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
I. H Rosenberg Rethinking brain food Am. J. Clinical Nutrition, November 1, 2007; 86(5): 1259 - 1260. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
