Flavonoid-rich orange juice is associated with acute improvements in cognitive function in healthy middle-aged males

Twenty-four healthy males were recruited from Reading, UK, and surrounding areas. Two participants withdrew following the completion of arm one due to work commitments (data not included). Inclusion criteria were males aged 30–65 (mean 51, SD 6.6), native English speaking, BMI 25–32 kg/m² (mean 28.3, SD 3.1), non-smokers. Exclusion criteria were history of stroke/myocardial infarction, alcohol misuse, renal or bowel disease, pancreatitis, diabetes or any other endocrine disorder, dementia or mild cognitive impairment (mini–mental state examination [18] (MMSE) <26), current weight loss regime, vegetarian diet, regular intake of supplements including fish oil, fatty acid, vitamins and minerals, any medication for hypertension, hyperlipidaemia, inflammation, hypercholesterolemia and hypercoagulation, and history or evidence of depression (as indicated by the Brief Symptom Inventory, BSI). Pilot data by our group with a sample size of 24 healthy adults using the Digit Symbol Substitution Test (DSST) indicated an effect size (eta squared) of 0.32 following consumption of a fruit-based juice. Therefore, for a crossover design, assuming two-sided test criteria with alpha level 0.05 and a correlation of 0.5 between repeated-measures test scores, for two treatment conditions and three time-point measurements, a sample size of 24 would give power of at least 86 %. Recruitment commenced in June 2012 and terminated in December 2012.

A randomized, double-blind, placebo-controlled, crossover design was used, and good clinical practice and ethical principles were followed according to ICH and the Declaration of Helsinki (1996). There were two drink conditions: placebo and FR orange juice, and measurements were taken at three time points: baseline (prior drink consumption), 2 h post consumption and 6 h post consumption of the drink. All procedures were approved by the University of Reading Research Ethics committee (No. 12/06).

The 240-ml FR drink consisted of Tropicana Pure Premium Orange Juice with 5.5 g of added orange pomace fiber (240 ml) which contained a total of 272 mg flavonoids (see Table 1 for drink composition). Analysis of the flavonoid content was performed by PepsiCo (Barrington, Illinois, USA). The orange pomace comprised of the micronized, edible part of the whole orange which is leftover during the production of Tropicana Pure Premium Orange Juice. The pomace is rich in fiber (40:60 ratio of soluble to insoluble) and contains small amounts of micronutrients, and a high proportion of the flavonoids found in whole orange. The control drink was prepared by dissolving a placebo mix containing glucose, fructose, sucrose and 0.67 % citric acid (for flavor purposes) in 240 ml of water. The drinks were closely matched for volume, taste, appearance, energy, glucose, fructose and sucrose (see Table 1). The ingredients for both drinks (except water) were provided by PepsiCo Inc. and were kept frozen (−20 °C) until preparation. The drinks and randomization sequence were prepared at the University of Reading by an independent researcher (AW) who did not come into contact with the participants or administer any of the test-day procedures. Both drinks were stored frozen (−20 °C) as individual portions (240 ml) in aluminum canisters and labeled with a three-digit code known only by AW. Blinded experimenters (MA, GD) performed the test-day procedures and provided participants with the drinks. Participants consumed the drinks from the canisters with black straws. The key for the codes remained sealed until data analysis was complete.

Participants initially attended a screening visit during which inclusion and exclusion criteria were checked with the MMSE [18] (mean 28.2, SD 1.5), BSI (mean 1.05, SD 1.89) and a Health and Lifestyle Questionnaire. A practice version of the cognitive battery was completed for familiarization (data not collected). In addition, participants also completed the following tests which provide a proxy for IQ, National Adult Reading Test (NART; mean score 30.5, SD 8.5) and Block Design (from the Wechsler Adults Intelligence Scale-R [19]; mean score 38.2, SD 9.2). Height (mean 1.8 m, SD 7 cm) and weight (92.8 kg, SD 3.1) were also measured. During screening, participants were provided with oral and typed instructions to consume a low-polyphenol diet 24 h prior to each test day which included a description of foods and drinks to avoid (berries, fruits, fruit juices, jams and preserves, red wine, black, green and fruit teas, coffee, cocoa, soy products, caffeinated energy drinks and vegetables except potatoes). The screening visit and each test day were separated by a 2-week washout. The evening prior to each test-day participants consumed (at home) a low-fat standardized chicken and rice meal provided by the research team (350 kcal, 6.9 g fat of which 3 g saturates, 52.1 g carbohydrate of which 9.7 g sugars, 19 g protein, 1.4 g fiber, 0.9 g salt) to avoid second meal cognitive effects [20]. For each test day, participants arrived at 8:30 a.m. following a 12-h fast. During the fast, participants were permitted to consume only water provided at screening. A standardized breakfast of croissants and low-fat cream cheese was consumed upon arrival (24.5 g fat, 10.1 g protein, 38.4 g carbohydrates, 415 kcal). The baseline cognitive battery commenced at 9:10 a.m. followed by the consumption of the FR or control drink at 10:00 a.m. Instructions were to consume the whole drink within 5 min. Repeated testing of the cognitive battery commenced 2 and 6 h post-drink consumption (approximately 12:00 and 16:00 h, respectively). Following completion of the second cognitive battery, a lunch was provided (identical to breakfast at approximately 13:00). The rationale for cognitive testing at 2 and 6 h was based on the predicted bioavailability of hesperidin and narirutin metabolites in the plasma which peak between 4 and 7 h postingestion [21, 22]. All visits took place at the University of Reading Hugh Sinclair Nutrition Unit. Participants remained within the Nutrition Unit for the entire test day during which only water consumption was permitted (notwithstanding the test-day foods and drinks). Participants received a £100 honorarium upon completion of both arms.

The 45-min cognitive battery consisted of the following tests administered in the respective order: Immediate Word Recall; Simple and Complex Finger Tapping; DSST; Continuous Performance Task (CPT); Serial Sevens; Positive and Negative Affect Scale (PANAS—a measure of subjective mood); Contrast Sensitivity; Delayed Word Recall. Immediate Word Recall involved computerized serial presentation of fifteen words in a random order at a rate of one word every 2 s. Following the presentation, participants were required to orally recall the words in any order. Presentation and recall occurred three times. The dependent variable was the mean number of words recalled per trial (averaged of three trials). False positives were not documented (e.g., saying a new word). Delayed Word Recall (45 min postinitial presentation) involved one attempt to orally recall the words presented during Immediate Word Recall. Simple Finger Tapping involved pressing the “2” key as many times as possible with the index finger over 10 s. The forearm was fixed, and two trials were performed with each hand. The first trial was considered a practice, and data were only collected for the second trial. Complex Finger Tapping followed the same procedure expect the required response was 1–2–3 (consecutively) using the forefinger for 1, the second finger for 2 and the third finger for 3. The dependent variable was the total number of responses, averaged across the dominant and non-dominant hand (errors were not recorded). The DSST [19] is a pen and paper test which contains a key of nine digit-symbol pairs and an accompanying list of digits. Under each listed digit, a space is provided for the participant to enter the corresponding symbol. The dependent variable was the number of symbols correctly entered over 90 s. The CPT involved random presentation of letters at a rate of one every 250 ms for a duration of 6 min. Participants were required to press the space bar when a letter other than “X” was presented. The dependent variable was the number of false positives (i.e., pressing the space bar when an “X” appeared). Serial Sevens required participants to orally count down in sevens from a randomly determined starting number ranging from 700 to 1000. The dependent variable was the number of correct responses over 2 min. Correct answers were determined relative to the previous number given, regardless of whether the previous response was correct or incorrect. The PANAS [23] requires a response to twenty adjective items on a scale ranging from 1 to 5 according to how the respondent is feeling at that moment with “not at all” and “extremely” as the anchor points. Scoring is collated into two factors reflecting Positive Affect and Negative Affect. In addition, analysis of each PANAS mood adjective was also performed. The Contrast Sensitivity test required identification of a series of letters on a computer screen. The letters become increasingly difficult to distinguish as the contrast between the letters and the background is reduced. The dependent variable was the point at which the letters could not be identified (Michelson Contrast).

Given the importance of considering baseline cognitive performance prior to drink consumption, change from baseline was calculated at each time point (2 and 6 h) for each cognitive test, and 2 × 2 repeated-measures, Bonferroni-corrected ANOVAs (Drink × Time) were performed using the change from baseline data with planned contrasts (t tests) to compare each drink (FR vs placebo) at 2 and 6 h. To examine equality at baseline, t tests were performed comparing FR and placebo at baseline for each cognitive test. In addition, for consistency, interpretability and agreement with other studies [24], the cognitive tests were combined to create a measure of global cognitive function. To do this, initially for each cognitive test, the change from baseline data was combined across all drink conditions (placebo and FR) and time points (2 and 6 h), thus forming one variable for each cognitive test which was subsequently converted into z-scores [24]. Subsequently, the mean z-score for each drink condition (placebo and FR) and each time point (2 and 6 h) was calculated such that global cognitive function = (Z _DSST + Z _{Serial Sevens} + Z _{Immediate recall} + Z _{Delayed recall} + Z _CPT + Z _{simple finger tapping} + Z _{complex finger tapping} + Z _{Contrast Sensitivity})/8. The z-score data for Contrast Sensitivity and the CPT were multiplied by −1 prior to calculating global cognitive function since a higher score on these two tests indicates worse performance. Repeated-measures 2 × 2 ANOVAs (Drink × Time) were then performed with change from baseline z-score data for global cognitive function. Change from baseline was calculated for subjective mood data which were subsequently analyzed with Bonferroni-corrected repeated-measures 2 × 2 (Drink × Time) ANOVAs. Data were analyzed using SPSS Inc. PASW Statistics 18.