Effects of Breakfast Glycaemic Index on Cognition in Earlier or Later Chronotypes.

Background Studies in young adults indicate beneficial effects of a low glycemic index (GI) breakfast on cognitive performance. A meta-analysis of 17 studies revealed benefits of a low-glycemic load (GL) breakfast on immediate memory only in the late post-prandial period. The meta-analysis reported no overall effect on attention in adults, yet suggests advantages for attention in children, adolescents, and young adults. A limitation of previous studies was that dairy was often allowed despite its strong insulinotropic effect, dampening the postprandial glycaemic response. To ensure an unbiased evaluation, the GlyCoBrain intervention study 1 will use pure water-carbohydrate mixtures, avoiding nutrient interactions, to examine the isolated effect of a high-GI breakfast.

Mechanisms explaining the time-dependent benefits of a low-GL breakfast are not fully understood. The provision of glucose to the brain is currently discussed as a central mechanism linking meal GI/GL to cognition. Generally, glucose is required for cognitive efforts as evidenced by decreases in local extracellular glucose concentrations in the activated brain area. The synthesis of key neurotransmitters in the brain requires a glucose supply. Glucose metabolism may be linked to memory via effects on tryptophan utilization and serotonin concentrations. Overall,the memory enhancing effect of glucose may require relatively constant blood glucose levels in the brain rather than high glucose amounts per se. Hence, whilst high GI foods elicit a transient increase in blood glucose level, low GI foods will result in a more sustained and longer-lasting provision of glucose. Since recent work suggests that there may be an optimal blood glucose range for cognitive performance and that this optimal range may differ by cognitive domain, it appears prudent to use continuous glucose monitoring in addition to the conventional measurement points directly before the cognitive test. However, benefits of low GI/GL breakfast for cognitive function have also been observed without differences in blood glucose levels. This may be attributable to an acutely improved postprandial insulin sensitivity and/or lower cortisol levels elicited by low GI meals. Accordingly, concomitant examinations of all three factors, i.e. glucose, insulin and cortisol levels, seem necessary to disentangle potential mechanisms.

The above mentioned meta-analysis also found greater benefits for individuals with better glucose tolerance, although this may be confounded by age. Studies suggest that young adults with poorer glucose tolerance, even within the healthy range, may be more vulnerable to these effects. While glucose tolerance follows a circadian rhythm, being lowest in the evening due to reduced pancreatic β-cell function and insulin response, circadian misalignment-common among young adults with later chronotypes-can also impair glucose tolerance, even at breakfast, due to decreased insulin sensitivity. This misalignment, often caused by social jetlag, has been linked to obesity, type 2 diabetes, and higher blood glucose and HbA1c levels, and an increased risk of depression. Considering results from our previous study showing high glucose responses among young adults with a late chronotype in the morning to a high GI breakfast, the GlyCoBrain intervention study 1 examines the hypothesis that consuming an early breakfast "against the inner clock" affects cognitive performance.

Hypothesis

(1) A low GI beverage breakfast (isomaltulose® beverage) results in a better memory and attention until 180 minutes compared to a high GI beverage breakfast (glucose beverage). (2) a low GI beverage breakfast results in fewer fluctuations in memory and attention throughout the morning. (3) the benefits of a low GI breakfast are particularly observed among people with later chronotype.

Aim The main objectives of this controlled nutrition trial are to analyze the effect of a high or low GI beverage at breakfast on memory and attention over the course of the morning including the late postprandial phase among young healthy university students. As young students are at high risk for a later chronotype, we aim to explore these effects in two samples of students with an earlier and later chronotype. In addition, our study will allow us to explore the relevance of chronotype for the postprandial course of glucose, insulin and cortisol levels.

Methodology Participants Healthy, German speaking students at Paderborn University with early or late chronotype.

Sample size calculation:

We base our sample size calculation on the difference in immediate memory and the study rated best in terms of quality in the meta-analysis. Given a standard deviation of 3.6 and 4.3 in both groups, and a standardized mean difference of 0.31 for people with a better glucose tolerance (corresponding to mean difference of 1.12) we require 44 participants in the smallest subgroup with a given power of 80% and alpha of 0.05 considering a correlation of 0.8 between both measurements (calculated with STATA, vers. 17.0). We assume, that 66% of invited students will participate and that 75% of those will complete the study. Hence, to have the full data set of 88 persons (44 with an early chronotype and 44 with a late chronotype) completing the trial, 117 need to start the trial and 177 people need to be invited from the GlyCoBrain Observational study which recruited 356 students from Paderborn University and screened for chronotype via MCTQ. In both chronotype samples equal sex distribution is envisaged.

Both samples will be invited to participate in the crossover nutrition trial, excluding smokers, people working on shift or having travelled for >2 time zone in the past 3 months, persons taking methylphenidate or using melatonin. Students will participate in 2 breakfast-cognition tests, each after an overnight fast at 8 a.m. to be taken over a course of 2 weeks including 1 week wash-out. Participants will randomly be assigned to one of two sequence groups, which differ only regarding the sequence of the intervention (Glucose beverage, isomaltulose® beverage or isomaltulose® beverage, glucose beverage). Randomization lists are provided by an external statistical advisor from the University of Esbjerg stratified by sex and chronotype. A randomized list was generated for 4 groups, each with 30 slots, using a block randomization approach. The randomization was performed using STATA (Vs 18.5; randomize) with 5 blocks, ensuring balance across the groups

Participants will receive to two different simple beverage breakfasts:

1. high GI beverage (glucose beverage) consisting of 75 g glucose dissolved in 500 ml tap water,
2. low GI beverage inducing no reactive hypoglycaemia (isomaltulose® beverage) consisting of 75 g isomaltulose dissolved in 500 ml tap water.

The intervention is structured into a 2-week schedule:

Week 1. Participants are invited for preparation day 1, during which a venous fasted blood sample is taken to determine fasting lipids, high-sensitivity-C-reactive protein, alanine aminotransferase and gamma-glutamyltransferase to metabolically characterize the recruited participants. To validate the consistency of the chronotype that had been determined in the GlyCoBrain Observational study previously, the participants' chronotype will again be determined via the Munich chronotype questionnaire. Body composition will be reassessed by Bioimpedance Analysis to estimate individual percentages of body fat, muscle mass, total body water and extracellular water. Additionally, body weight will be measured directly by the medical Body Composition Analyzer (mBCA, seca, Germany). Body height will be measured using an ultrasound measuring station (SECA287 db). Waist circumference will be assessed on the exposed upper body at the midpoint between the lower ribcage and the hip bone.

Continuous glucose monitoring (CGM, G7, Dexcom, Inc., San Diego,CA) is activated, accelerometry will be used to measure movement and sleep/wake conditions. A standardized evening meal will be provided to be consumed in the evening before the intervention to avoid secondary meal effects e.g. by pulse consumption. Participants will be asked to perform a training session on the cognition battery (see below).

The actual intervention starts three to five days later at 8 a.m. with each day running as follows:

• minimal invasive capillary blood samples collected at -40, 30, 80 and 140 minutes (in relation to the time point of drinking the beverage) will be used to determine glucose and insulin levels
• non-invasively cortisol samples from saliva at -40, 80 and 140 minutes
• continuous glucose measurements
• Heart rate belt to measure the heart rate variability in combination with the Unite fitness watch (Polar Electro, Finland).
• Cognition test at -30, 90 and 150 minutes
• test beverage breakfast including either Glucose or Isomaltulose to be drunken at time point 0, 9 a.m.
• appetite, thirst and mood will be assessed at -40, 30, 80 and 140 minutes each using a validated visual analogue scale
• removal of all devices and provision of a voluntary breakfast at 180 minutes Glucose levels will be directly determined using HemoCue Glucose 201 RT Analyzer (HemoCue AB, Schweden). Plasma insulin concentration and cortisol concentration in saliva will be determined by an enzyme-linked immunosorbent assay (ELISA) using a kit from IBL/Tecan.

Week 2. Again, participants repeat the preparation day as before (day 2), except for BIA, MCTQ and venous blood sample measurements. The intervention then continues 3 to 5 days after the preparation day following the same time scale as on intervention day 1.

Cognition tests Assessment using the test battery is expected to last approximately 25 minutes in total each time. Tests will include a set of tests developed by the ALA Institute Bochum, that was applied in our previous studies.

• verbal memory test (ability to correctly recall a total of 30 words immediately (recall) and after 20minutes (delayed recall))
• Visual-spatial short-term working memory test (Corsi-block tapping test equivalent, i.e., tapping a sequence of up to nine identical spatially separated blocks, sequentially increasing the difficulty by the number of presented blocks).
• Test of selective and sustained attention and visual scanning speed (modified version of the d2 Test of attention, during which the reaction specific signs within a series of presented signs will be assessed)
• Test of spatial attention and switching abilities between two different tasks (alternative version of the Trail Making Task)
• Test of tonic alertness (simple reaction time test).
• Test of inhibition, i.e. ability to inhibit a prepotent response, using a flanker task. In each trial three superposed triangles will be presented to the participants. The upper and lower triangles (flankers) will be pointing in the same direction but independent from the middle triangle (target). During the no-go trials a circle replaces the target. The participants will be supposed to press the buttons left or right according to the direction of the target or not to react in the case of a no-go condition. Each trial will be categorized as compatible-go, incompatible-go or no-go.
• A dummy test to ensure that delayed recall starts exactly 20 minutes after the immediate recall Statistical analysis Multilevel regression models will be used to analyze the effect of the interventions on verbal memory at 150-180 minutes. Identical models will be used for the other cognitive domains as outcome. Tests for interactions of the intervention effect with chronotype will be performed and stratified analyses by chronotype will be run. Note that consideration of interactions may also be relevant if p<0.1. Additional analyses will control for insulin sensitivity, cortisol, appetite, thirst and/or mood to explore whether observed effects may be partly attributable to these factors.