The Omnibus Satiety Metric: Predicting Satiety in Humans Through Brain, Blood, and Subjective Data

Overview

Designing food and drink that maximizes satiety has long been an ambition of industry and public health. For obvious reasons, foods that fill faster and for longer are desirable to consumers for controlling their weight, and to public health programs that are designed to prevent obesity. Current methods for measuring satiety have weak predictive value, commonly fail to replicate, and are yet to be validated with respect to energy consumption in everyday life. The investigators propose to overcome this deficiency by developing the Omnibus Satiety Metric (OmniSaM). OmniSaM is proposed as a multi-modal metric that targets the full spectrum of processes underlying the satiety cascade, composing brain, blood, mind and behavior of consumers. As a proof-of-concept, subjects (normal BMI) will undergo a preload ad libitum paradigm, with a 2-parameter factorial design comparing milk based products differing in levels of caloric load and protein-to-carbohydrate ratio. The investigators will acquire 1. high-resolution neuroimaging data, whilst 2. simultaneously recording appetitive hormones, 3. blood metabolite composition, 4. subjective sensory indices of appetite and 4. behavioral metrics of consumption.

The overarching strategy is to give subjects preload foods of variable combinations of protein-to-carbohydrate ratios (two levels) and caloric load (two levels), testing the effect this has on ad libitum consumption, using neuroimaging, sensory and endocrine metrics as predictor variables. The investigators will conduct a repeated measures cross-over study, meaning that all subjects will be exposed to all four combinations of protein-to-carbohydrate ratio and caloric loads. A total of 25 subjects will complete the study. Subjets will consume a shake preload drink outside the Magnetic Resonance-scanner (MR-scanner) and after a wait time of 4 hours during which scanning, endocrine and sensory measures are obtained, the subjects will eat an ad libitum meal consisting of standardized prepared spaghetti bolognese. Endocrine and sensory measures will continue until one hour after intake of the ad libitum meal. All subjects will undergo four test days of \~5 hours in the Danish Research Centre for Magnetic Resonance (DRCMR), Hvidovre Hospital, a maximum of one test day per week because of blood loss considerations. On study days the participants will arrive at 11:30am having fasted since 22:00 the previous evening with a fixed consumption of 200 ml of water in the morning 2 hours before the test. A Intravenous cannula will be placed in participants arm to take blood samples, and baseline blood samples and baseline measures of appetite sensations will be collected. At time 0 participant will be instructed to consume a preload shake of Queal (Queal, Queal Smooth Vanilla, Queal Quick Meal, The Netherlands) added whey protein isolate (LACPRODAN SP-9225 INSTANT, Arla Foods Ingredients, Denmark) or lactose monohydrate (VARIOLAC 992, Arla Foods Ingredients, Denmark) and water. This candidate comparison is designed to exploit known differences in satiating capacity for both caloric load, and for protein to carbohydrate.The logic behind the selection of these factors is to use factors for which we have the strongest a priori evidence for their satiety effects, such that the method can be developed as a proof-of-principle. Along with the preloads, participants will drink 200 ml water containing 1.5g of paracetamol as a tracer for gastric emptying. Afterwards the participants will be placed on a bed connected to the MRI-scanner, and fMRI scanning will start. The scanning will be stopped shortly when presenting questionnaires and taking blood samples. In the following two and a half hours participants will alternate between being scanned, resting, and completing questionnaires, repeating at \~15 minutes intervals. At \~t=180 min an ad libitum meal of Spaghetti Bolognese will be served. Subjects will be instructed to eat until pleasantly satiated, and intake will be calculated via weight of food consumed. Blood sampling and sensory measures will continue until one hour after intake of the ad libitum meal. Blood samples: Samples will be drawn through intravenous access by a catheter in the cubital fossa, taken at regular intervals of between 5 and 30 minutes from t=-15 to t=240 which includes testing 1 hour after the ad libitum meal. The amount of blood drawn from participants will be a maximum of 180ml (for each test day), which will not impose any risks for the participants who are all healthy (typically, voluntary blood donation involves donating 500 ml). Participants will be tested no more than once per week such that there is sufficient time for recovery from the experimental blood loss. Glucose, insulin, C-peptide and CCK, PYY, Ghrelin, GIP, GLP-1, FFA, Triglyceride and glucagon will be measured. Blood samples will be centrifuged, plasma and serum will be pipetted, frozen, and labeled with non-sensible and non-identifiable personal data for later analyses. Serum/Plasma samples will be analyzed at Clinical Biochemistry, Hvidovre Hospital where automatic biochemist analyses are performed. Other biochemical analyses that cannot be performed locally will be sent to our collaborating laboratory either at the A) Institute of Biomedicine (Professor Jens Juul Holst), which has specific equipment for analysis of blood samples for intestinal hormones or B) Department of Endocrinology and Internal Medicine, Aarhus University Hospital (Professor Kjeld Hermansen) which will perform another subset of biochemical analyses. Subjective measures of appetite: Subjective sensations will be measured by presenting questions about; liking, hunger, fullness, satiety, desires, prospective consumption and thirst, one question at the time in randomized order. Appetitive measures will be taken regularly throughout the test day, from prior to intake of pre-load to post ad libitum meal consumption. When lying in the scanner, subjects will rate the magnitude of their responses on a visual analogue scale (VAS), using button presses. And outside the scanner these responses will be recorded via an equivalent pen and paper VAS. MR-Scanning: During MRI scanning, participants are inserted into a long tube, surrounded by a coil, which yields a strong magnetic field. By changing the slope of the magnetic field, we can detect oxygenation of the blood. Subjects will be scanned through different imaging sequences. The participant will not be able tell the difference between these, but will hear the scanner makes a variety of sounds, depending on the sequence running. During a scan, the various anatomical images of the brain are taken. Additionally, the investigators measure changes in the magnetic signal generated by an increased blood oxygenation to certain parts of the brain. This is the BOLD signal, and is the purpose of the fMRI. Subjects will be inserted into the scanner at timepoint \~t=15, to around t=180 with a 10 minute break outside the scanner at timepoint \~t=80. From t=0, the subjects will drink the queal preload prior to entering the scanner. Subjects will perform the in-scanner subjective ratings (specified above) at the time points following the blood samples. Scanning will stop shortly to allow for blood samples to be taken. To take blood samples, a researcher will approach the subject who will remain in the bore of the magnet throughout, and draw blood from the catheter. Once blood sampling and appetite ratings have been collected, scanning will resume. In the gaps between functional sequences, standard structural will be run, lasting approximately 15 minutes. Subjects will be functionally scanned up to 6 more times. Once complete, subjects will be removed from the scanner and complete the rest of the experiment in a nearby testing room. Ad libitum meal: At \~t=180 subjects will begin the ad libitum (all you can eat) meal of spaghetti bolognaise, in a dedicated private room, where they will be on their own and unobserved. They can sit and eat for up to 30mins. Subjects will be instructed to eat until "pleasantly satiated". After consumption the amount eaten will be calculated by the weight differential. Subjects will be provided with 200 ml water during the meal (forced intake). Additional measures: As physical activity can affect appetite and ultimately energy intake in the ad libitum meal, energy expenditure will be restricted by advising participants to refrain from physical activity of high intensity three days in advance of each test day. Further, on test days participants must travel by automated vehicle to the hospital to ensure a low level of physical activity immediately prior to testing. To measure participants physical activity levels an activity tracker will be used. Participants will receive an activity tracker minimum four days before the test day and receive instruction in its use. Blood will be sampled during each study days (for a total of 720 mL over 4 days with minimum 1 week between each test day). On each test day, blood samples will be centrifuged, plasma or serum will be pipetted, frozen and labeled with non-sensitive and non-identifiable personal data for batch analysis after completion of the whole trial. Serum/Plasma samples will be analyzed at Clinical Biochemistry, Hvidovre Hospital where automatic biochemist analyses are performed. Other biochemical analyses that cannot be performed locally will be sent to a collaborating laboratory either at the A) Endocrinology Research Section, Panum Institute (Professor Jens Juul Holst), which has specific equipment for analysis of blood samples for intestinal hormones or B) Department of Endocrinology and Internal Medicine, Aarhus University Hospital (Professor Kjeld Hermansen) which will perform another subset of biochemical analyses. All samples will be kept until analysis in a research biobank at Hvidovre hospital. Unanalyzed samples will be kept until 01.01.2027 after which samples will be destroyed according to routine procedures. The biobank is only accessible for research staff involved in the project. Participants provide informed written content on the amount and storage of blood samples and can withdraw consent at any time resulting in destruction of the biological material. Power calculations: 25 x 4 repeated measures = 100 data points. For a two-tailed test, for a medium effect size with f-square value of 0.15, for an alpha of 0.05. With 35 predictor variables (25 subject factors, and then 10 aggregated predictor variables) we have 64 degrees of freedom, the calculated power for this design \> 0.9. If we calculate the power for inferring the effect of protein to carbohydrate ratio on consumption, the power is also very high. If we suppose an effect size of 10% difference in consumption from the lowest to the highest protein to carbohydrate ratio, then that is a slope of 3.85 (with respect to log PC ratio). If the null hypothesis is a slope of 0, with sd of 4, then the power also \> 0.9.