This study was conducted in compliance with the guidelines laid down in the Declaration of Helsinki (ethical principles for research involving human subjects). All procedures involving human subjects were approved by the Regional Ethical Review Board in Lund, Sweden (protocol 2008/5). Written informed consent was obtained from all subjects.
Volunteers, 30 women and 14 men aged 51–72 years (mean ± SD: 63 ± 5 years) with BMI between 20–30 kg/m2 (mean ± SD: 25 ± 3) from the south of Sweden, were recruited to the study through advertisement in local newspapers. Exclusion criteria were blood glucose > 6.1 mmol/L, BMI > 30 kg/m2 and known metabolic diseases, gastro-intestinal disorders or known cognitive decline. Medication for high blood pressure was allowed, but had to be kept constant during the study period. The subjects had to be fluent in the Swedish language. All but one subject (born in UK, but fluent in the Swedish language) were native Swedish.
Recruitment began October 2008. Trials took place between January 2009 and June 2009. Out of the forty-four volunteers enrolled in the study, four dropped out after the first intervention period (two after the placebo period and two after omega-3); three due to personal reasons and one was excluded due to reporting suffering from a disorder affected wakefulness. Forty subjects completed both intervention periods, but two subjects were excluded due to detection of abnormal fasting blood glucose concentrations (7.0 and 6.7 mmol/L, respectively). In total, results from 38 completers (28 women and 10 men) aged 51–72 years (mean 63.3 ± 5.3) and with BMI between 20–31 kg/m2 (mean 25.0 ± 2.8) were evaluated. Twenty out of the 38 subjects started with n-3 PUFA and 18 with placebo. Twenty-two subjects were senior citizens. One of the participants was an occasional smoker, but didn’t smoke the day before or during the test days. The participants were interviewed regarding eating habits and health status prior to the onset of the study. In addition, at each visit, i.e. prior to and after each intervention period, a questionnaire regarding current diet, physical activity, health status and use of medicine, were filled in. All participants consumed an ordinary Swedish diet, including meat, and fish every week. The subjects were told to continue with their habitual diet throughout the study period. When comparing diet journals within each subject, it could be stated that there were no major changes in diet or physical activity during the whole study period. No physical examination was included. Based on information from interviews and forms it was revealed that seven participants were medicated for hypertonia, and one subject with cholesterol lowering medication (Simvastatin). One subject was medicated for depression since several years, without any symptoms and changes in medicine for at least twelve month. Three participants suffered from mild osteoarthritis (one was medicated with glucosamine). Post-intervention evaluation of base-line data revealed that five subjects had high levels of triacylglycerides (2.3, 2.3, 2.7, 2.8, and 3.3 mmol/L) and three subjects showed borderline concentrations (1.8, 2.0, and 2.2 mmol/L).
No major side effects or problem to consume the supplements were reported.
Test product and placebo
The dietary supplement consisted of capsules containing 1000 mg fish oil, whereof 600 mg was n-3 PUFA (EPA 300 mg, DHA 210 mg and 90 mg unspecified) (Pikasol fish oil capsules, Axellus VS, ORKLA, Oslo, Norway). Five capsules per day were consumed; resulting in a total daily intake of 3 g n-3 PUFA (EPA 1500 mg, DHA 1050 mg and 450 unspecified).
The placebo supplement was provided as two tablets per day containing in total 366 mg dicalcium phosphate (E 341), 150 mg microcrystalline cellulose (E 460) and 4 mg magnesium salts of fatty acids (E 470b). The placebo was provided by Axellus VS, ORKLA, Oslo, Norway.
The study had a cross-over, randomised but balanced design. Out of the forty-four subjects described above, forty completed both intervention periods. Twenty subjects (14 women and 6 men) started with five weeks daily consumption of omega-3 PUFA, and consumed placebo in a second five weeks intervention period, and 20 subjects (14 women and 6 men) were enrolled to start with five weeks consumption of the placebo. The subjects visited the experimental department at four occasions; in the mornings prior to start of each intervention periods, and in the mornings after finishing the PUFA- and placebo period, respectively.
The evening prior to attendance, at 9.00 pm. the test subjects ate a standardized meal, consisting of white wheat bread with optional spread, and had coffee, tea or water to drink. Thereafter they were fasting until the arrival at the research department. At 07.45 am, the test subjects were weighed and seated to rest for a minimum of 10 minutes before the blood pressure was registered and fasting blood samples collected. A standardized breakfast was served consisting of white wheat bread (Dollar Storfranska, Lockarps bakery, Malmö, Sweden) and apricot marmalade (Ica, Sweden) corresponding to in total 55 g available carbohydrates. Water, 250 ml, or a plain cup of decaffeinated coffee or tea was served with the bread. The breakfast was consumed within 15 min. Thereafter, cognitive tests were performed and capillary blood tests were collected repeatedly up to 180 min post commencing the breakfast. The occupation during in-between the cognitive tests were standardized such that the subjects performed Sudoku.
The study was performed at the division of Applied Nutrition and Food Chemistry, Department of Food Technology, Engineering and Nutrition, Lund University, Sweden.
Prior to the intervention periods, i.e. at visit no. 1 and 3, the subjects performed pilot versions of the cognitive tests to reduce learning effects and stress at the cognitive test days. Measurements of cognitive performance were performed after completion of each intervention period, i.e. at visit no. 2 and 4.
Working memory (WM) test
The tests employed in the present study represent an extension of the methodology developed by Radeborg et al. . There are several reasons for choosing WM as a measure of cognitive performance in this study. WM can be defined as a system responsible for simultaneous temporary short term storing and processing of information, and is involved in many everyday activities; such as mathematical problem solving where one often has to remember part of the result in a calculation while performing further mathematical operations. WM represent a fundamental ability for higher-level cognitive processes. Thus, measures of WM capacity have been shown to correlate significantly with activities as diverse as e.g. reading comprehension , note taking , the following of directions , reasoning , and complex learning . Some authors [30, 31], even claim that WM and general problem solving ability or intelligence, as measured by e.g. Raven´s Matrices, reflect nearly identical constructs. However, whereas intelligence tests generally only can be administered once due to risk of considerable learning effects, WM can be measured repeatedly. In total, three oral WM-tests were included at each experimental day (performed at 60, 110, and 160 min). Two of the tests (at 60 and 160 min) were executed principally as described previously , modified by including 12 sets of 3–5 short declarative sentences (four of each number) instead of 4 sentences in all sets. As previously, the sentences could be either semantically meaningful of the type ‘the boy brushed his teeth’, or nonsensical, such as ‘the rabbit struck the idea’. The test leader was blind to the product provided to the test subjects. The sentences were read one by one to the subjects. Immediately after a sentence, he/she had to indicate if it was a semantically meaningful sentence or not. After each set of sentences the subjects had to repeat, in any order, the first noun in each of the sentences. One test could at maximum generate 48 credits. The tests consisted of equal number of sentences that were semantically meaningful (24 credits) and nonsensical (24 credits). It has been described that remembering of a noun in a semantically nonsensical sentence is more demanding . The WM-test could therefore be divided into two parts differing in degree of difficulty. The third WM-test was performed at 110 min and was similar to the tests just described, with the exception of that instead of short sentences, the test was composed of simple additions of two single digit numbers. The test leader presented orally the two figures to be added, and the test subjects were supposed to immediately give an oral answer to the addition. After a set of 3–5 additions, the subjects had to repeat the first figure in each addition. The test could at maximum generate 48 credits. One WM-test took approximately 8–10 min to perform. Four different but comparable WM-tests composed of sentences and two different but comparable WM-tests composed of figures were included in the study.
Selective attention (SA) test
The test was based on spatial perception and primarily measured the ability to sustain a prolonged attention, and to control and split the attention to the entire picture on a computer screen. Like the WM-test, the test also dealt with simultaneous temporary storing and processing of information (WM capacity). The storing time required was however shorter compared with the WM-test, whereas the time pressure was higher. The SA was measured using a computerized test made up of 96 pictures, each shown for two seconds on the computer screen. The pictures consisted of a square on a white background, divided into four equally sized smaller squares. One of the smaller squares was red, one square was green, and two squares were uncolored (white), resulting in a total of 12 unique picture combinations. The subjects had to remember the positions of the colored squares, and to compare each new picture that emerged on the screen with the preceding one. Each time a new picture emerged, either the green, the red, or none of the colored squares were in the same position compared with the previous picture. Within the two seconds each picture was shown, the subjects were supposed to indicate by pressing one of three different keys on the keyboard, which of the three possible alternatives that occurred for each new picture. The test began with a short training session, and took approximately 10 min to perform. The test was scored with the number of correct responses (CR, total 95 credits) and for the reaction time (RT) needed to give the answer (i.e. press one of the keys).
Metabolic risk markers
Physiological test variables were determined prior to and after completing each intervention period. Blood pressure was determined with an automatic blood pressure cuff (Digital Automatic Blood Pressure Monitor, Model M3 Intelligence, OMRON HEALTHCARE CO., LTD, Kyoto, Japan). Finger-prick capillary blood was withdrawn at fasting and at 15, 30, 45, 60, 90, 125, 160 and 180 min after the start of the standardized breakfast for determination of glucose concentrations and glucose tolerance (HemoCue®B-glucose, HemoCue AB, Ängelholm, Sweden). Venous blood was withdrawn for determination of fasting levels of serum (s) insulin, s-TNF-α, s-adiponectin, s-free fatty acids (s-FFA), s-triacylglycerol, and plasma (p) malonedialdehyd (MDA). The venous blood samples were centrifuged and plasma and serum separated and stored in a freezer (−40°C) until analyzed.
Methods for analyses of insulin, FFA, adiponectin and triacylglycerols are described elsewhere . S-TNF-α was determined with a sandwich enzyme immunoassay kit (TNF-α ELISA Kit, Immunodiagnostik AG, Germany). Plasma MDA was determined by measure of lipid peroxidation as TBARS as is described in , modified by excluding the n-butanol.
Calculations and statistical methods
Primary outcome measure was results in the WM-test. The sample size was calculated based on a study in healthy middle aged subject, including a similar WM-test as was used in the current study . A significant effect (p < 0.05) was detected on WM, with an effect size (Cohen´s d) of d = 0.75. In the present study we assumed a smaller effect (d = 0.50), resulting in a power of 0.86 in a one tailed statistical hypothesis test, and a power of 0.77 in a two tailed test. In a power calculation, based on table 9–9 and 9–10 in Aron and Aron 82003: Statistics for Psychology, 33 subjects would be enough to get a 80% power. However, to have the possibillity to have a balanced design (balansed with respect to order of products and order of test sequenses), we decided to involve 40 subjects.
The results are expressed as means ± SEM. The influences of the test- and placebo products on the cognitive tests were analyzed by repeated measures ANOVA at the test points, with order of test meals and test meals as independent variables and performance on cognitive tests as dependent variables. Statistical calculations were performed in Stat View 5.0 and SuperAnova 1.11. Treatment effects on physiological test parameters (based on changes from baseline in the intervention -and placebo periods, respectively) were assessed with analysis of variance (ANOVA general linear model) in MINITAB Statistical Software (release 13.32; Minitab inc., State College. PA. USA). Time effects on cognitive tests were assessed with analysis of variance (ANOVA general linear model) followed by Tukey’s pairwise multiple comparison method for means (adjusted means were reported) in MINITAB. Participants acted as their own control. GraphPad Prism (version 4.03; GraphPad Software, San Diego, CA, USA) was applied for calculation of blood glucose incremental areas under the curves (IAUC). Blood glucose IAUC (0–90 min) was used as an estimate of glucose tolerance. Pearson correlations were applied to study relations between physiological test parameters and results in the cognitive tests. Values of P ≤ 0.05 were considered statistically significant. Cohen’s d is presented to report effect size for significant results . n= 38.