Effects of pre-meal whey protein consumption on acute food intake and energy balance over a 48-hour period

The effects of pre-meal whey protein consumption on acute food intake and subsequent energy balance measured over 48-h was investigated in males of healthy-weight (HW) or living with overweight and obesity (OV/OB). On two separate trial days, following a controlled breakfast (09:00) and lunch (13:00), 12 HW and 12 OV/OB males consumed either whey protein (20 g) or flavoured water beverages (16:40), and ad libitum test meal (17:00). A controlled 48-h assessment of energy intake and expenditure was used to determine any compensatory behaviour. Test meal energy intake reduced 15.9 % in HW (P = 0.003), and 17.8 % in OV/OB (P = 0.005) following whey protein, compared to placebo. We report no between-group differences and no changes in compensatory behaviour. A small dose of whey protein reduces energy intake at the next meal, without upregulating compensatory behaviours in both HW and OV/OB males. However, chronic effects on body composition and weight loss remain to be elucidated.


Introduction
Dietary protein intake induces satiety and reduces food intake to a greater extent than other macronutrients (Poppitt et al., 1998;Speakman, 2022).Milk proteins are of specific interest in the management of overweight and obesity due to strong associations between high dairy consumption and low body mass (Phillips et al., 2003;Lu et al., 2014).Indeed, whey protein has been reported to suppress appetite and subsequent food intake in comparison to placebo or isocaloric doses of casein, egg or soy protein in healthy individuals and people with obesity (Akhavan et al., 2010;Pal & Ellis, 2010;Poppitt et al., 2011;Zafar et al., 2013).This may be due to the abundance of branched chain amino acids (BCAAs) and bioactive peptides in whey protein, providing a wide range of physiologic functions including delayed gastric emptying and stimulation of appetite satiating hormones (King et al., 2018;Madureira et al., 2010;Stanstrup et al., 2014).Physiologically, 20 min appears to be the minimum interval for postabsorptive effects of the preload to influence energy intake (Booth et al., 1976), with inter-meal intervals (IMI) of between 20 and 120 min reported to reduce food intake in adults (Almiron-Roig et al., 2013).The smallest efficacious dose of whey protein required to suppress food intake was 20 g when IMI was 30 min (Akhavan et al., 2010).
Despite the acute appetite supressing effects of whey protein, when no concomitant dietary or exercise intervention is undertaken, the effects of chronic supplementation of whey protein on body composition are equivocal (Baer et al., 2011;Pal & Ellis, 2010).Baer et al. (2011) reported reductions in body mass and waist circumference following 23 weeks of whey protein supplementation (56 g/d) in people with obesity when compared to isoenergetic (1670 kJ/d) carbohydrate.In contrast, Pal and Ellis (2010) reported no change in body composition after 12week supplementation (54 g/d) when compared to isoenergetic glucose supplementation.This may, in part, be attributed to compensatory adaptations in other components contributing to energy balance.Previous short-term feeding studies have failed to identify whether the energy deficits reported were sustained in the longer-term, with no quantification of energy expended through physical activity and diet induced thermogenesis (DIT).Furthermore, basal metabolic rate (BMR) and daily energy requirements are typically estimated using physical activity questionnaires for which validity is questionable (Neilson et al., 2008).
The objectives of this study were to determine the effects of whey protein consumption (20 g) 20 min prior to an ad libitum test meal on acute food intake and energy balance over the subsequent 48-hours in males of healthy-weight or with overweight and obesity.Therefore, this study is designed to advance our understanding of the appetite suppressing effects of whey protein and identify compensatory behaviour following whey protein ingestion, which could optimise body mass management strategies.We hypothesized that a significant reduction in food intake would be observed following whey ingestion at the ad libitum meal, but that compensatory adjustments in behaviour would negate the energy deficit over the subsequent 2 days.Secondly, we hypothesized that a greater degree of compensation would be observed in the people living with overweight and obesity.

Participants
The study population consisted of 12 healthy-weight males (HW, 20.0-24.99 kg/m 2 ) and 12 males living with overweight or obesity (OV/ OB, >25 kg/m 2 ) aged 18-65 years, without diagnosed metabolic/ autoimmune disease, common food allergens, intolerances, or dietary restrictions.Breakfast skippers, smokers, dieters and those currently taking prescribed or over-the-counter medications that influence appetite or gastric motility were excluded.Restrained eaters were identified for exclusion by a score of > 12 on the cognitive restraint scale of the 51 Item Three Factor Eating Questionnaire (Stunkard & Messick, 1985).Female participants were excluded due to the time constraints of not being able to control for the effect of the menstrual cycle on dietary and physical activity behaviour (Buffenstein et al., 1995).Participants were recruited from Northumbria University and external local organisations through poster advertisement, and local areas of Newcastle-upon-Tyne through radio broadcast advertisements.Participants were financially compensated (£20 shopping voucher).The present study was conducted in accordance with the Declaration of Helsinki and the procedures were approved by the Northumbria University Ethics Committee (REF: HLSDK090916).Written informed consent was obtained from all participants prior to enrolment and they had the right to withdraw at any time.

Pre-trial procedures
For pre-trial procedures participants arrived at the Northumbria University laboratory following an overnight fast (10-12 h) but allowed water ad libitum.Participants removed footwear and excessive clothing to permit accurate stature, body mass and BMI measurements (Seca, Hamburg, Germany).Resting metabolic rate (RMR) was assessed by indirect calorimetry using an online gas analyser (Oxycon Pro, Care-Fusion, USA).Participants lay supine for 30 min while metabolic gas exchange parameters of oxygen consumption (O 2 ) and expired carbon dioxide (CO 2 ) were collected for 20 min.Alcohol and caffeine use and strenuous exercise were prohibited for 24 h and 48 h, respectively as Rocha et al. (2006) has shown that even a light bout of intensity activity can influence energy balance in the days following exercise.RMR was calculated using the abbreviated Weir equation (((3.94* O 2 ) + (1.106 * CO 2 )) * 1.44) (Weir, 1949).
On the same visit, participants were familiarised to the ad libitum test meal procedures and university food laboratory.Participants were separated into individual feeding booths whilst personal possessions such as smart phones and computers were prohibited to minimise distractions, and the potential viewing of time or food cues.A uniform, homogenous meal of pasta (Tesco fusili pasta, UK) and tomato sauce (Lloyd Grossman, UK) (100 g; 614 kJ, 145 kcal, 4 % fat, 82 % carbohydrate, 14 % protein) was served until the participant signalled that they were sufficiently full and satisfied.Interactions between the principal investigator were minimised but bowls were removed from the participant before the contents were fully consumed and replenished.Food intake was recorded by weighing the bowls before and after consumption, meal-time duration was also recorded.
Two days preceding experimental trials, Actiheart accelerometers (Actiheart, CamNtech, United Kingdom) were fitted to participants, providing valid assessments of energy expenditure via a combination of heart rate monitoring and movement registration (Lof et al., 2013).The accelerometer attached onto two ECG electrodes (3 M Health Care, Canada) placed at 12-lead positions V1 and V4 and worn for 4 days, collecting data in 1-minute intervals.Participants also received weighing scales to assist accurate reporting of daily food intake for diary reporting and were guided in their use.Weighing measurements of all ingredients was carried out twice in repetition to ensure accurate reporting.Participants were asked to replicate dietary choices in the h prior to each main trial.To assist this standardisation, pre-packaged meals (Tesco cottage pie, UK) and snacks (Nature Valley bar, UK) were provided to be consumed in the evening prior to each trial with the aim of normalising appetite perceptions, glucose metabolism and gut hormone parameters (Chandarana et al., 2009).

Protocol
Participants were studied on three separate occasions with 7 d between each study visit.Trials were conducted in a randomized, singleblind, crossover design.Trial sequences were randomly assigned with the use of a computerised random-number generator (https://www.randomization.com).Participants were instructed to avoid consuming any alcoholic beverages and conducting any strenuous physical exercise 24 h prior to the study day.A schematic of the main trial protocol is presented in Fig. 1.Participants arrived at 08:45 following an overnight fast and completed visual analogue scales (VAS) to rate subjective hunger, fullness, prospective fullness, desire to eat, thirst, mood and nausea on a 100 mm continuum (Flint et al., 2000).At 09:00, a fixednutrient breakfast meal containing cereal (Cheerios, Nestle, UK) and whole milk (Tesco, UK), equivalent to 15 % of RMR, was consumed within 15 min.At 13:00 a fixed-nutrient lunch meal of chicken soup (Heinze, UK), crisps (Kettle Foods, UK) and oat bar (Nature Valley, UK), equal to 35 % of RMR, was consumed within 15 min.VAS were recorded every 30 min throughout the study day.
At 16:40, participants received either whey protein (20 g, Lacpro-dan©, Arla Foods Ingredients Group, Denmark) or flavoured water placebo to be consumed as quickly as possible and within 1 min.The mixture of whey protein was achieved using protein shakers (Smart Shake, UK) with a standardised vigorous mixing time of 15 s.Drinks were served in a ready-made 150 ml opaque bottle with calorie-free citrus flavoured sweetener (20 ml, Fun One, Germany) added to both beverages to standardise taste and palatability which was tested previously (King et al., 2018).VAS were recorded to measure subjective ratings for how pleasant, salty, bitter, sweet, creamy, thick, sticky, fruity, and refreshing the preload tasted.Nutritional composition of the Lacprodan© whey protein powder is shown in Table 1.
At 17:00, participants commenced consumption of the ad libitum test meal, as described above.VAS were recorded once fullness was signalled.Upon departure from the laboratory, participants were required to complete diet diaries for 48 h using the provided weighing scales.Nutrition analysis was performed on software (Microdiet, Downlee systems Ltd, UK) along with item packaging, to determine the composition and nutritional content of foods consumed.

Statistical approach, analysis and power
Sample size calculation was determined to identify the minimal clinically important difference set to 180 kcal.Previous research suggests a within-subjects standard deviation of 120 kcal for ad libitum food intake in healthy-weight individuals and people living with obesity (Seimon et al., 2013).Therefore, using a 2-tailed p-value of 0.05, D.G. King et al. healthy-weight, and 10 people with overweight or obesity were needed to reject the null hypothesis that the population means are equal with a power of 90 %.The sample size calculation was processed using the software Power and Sample Size Calculations (PS) (Dupont & Plummer, 1998).To account for potential drop-out, a sample size of 12 healthyweight males and 12 males with overweight or obesity were targeted.
All hypotheses were specified a priori.All data were analysed using the Statistical Package for the Social Sciences (SPSS 24, IBM, United States) and reported as means and their standard deviation (mean ± SD).Tests of normality and sphericity were performed using the Shapiro-Wilk test, and Mauchly's test, respectively.Composite appetite score (mm) was calculated as [desire to eat + hunger + (100 − fullness) + prospective food consumption]/4 (Anderson et al., 2002).Paired ttests were used to examine differences between trials for HW and OV/ OB groups for outcomes including food intake and physical activity energy expenditure.To identify the effect of the intervention between groups, a mixed-model ANOVA was conducted (within-subjects' variables [Treatment]; between-subjects' variables [Group]).Area under the curve was calculated using the trapezoid method.Relationships between variables were assessed using Pearson's linear correlations assuming normality of data.A p-value ≤ 0.05 was regarded as being statistically significant.The analytic plan was pre-specified, and any data-driven analyses are clearly identified and discussed appropriately.

Participants
Twenty-four males completed the study (Table 2).With the exception of body mass and BMI, there were no significant differences in baseline characteristics.All preloads were tolerated by participants.

Laboratory standardisation
Figs. 2 and 3 show time-course changes in self-reported ratings of hunger, fullness, desire to eat, and prospective food intake after breakfast and lunch, respectively, in HW and OV/OB participants.All fasting self-reported ratings of appetite were similar upon arrival at the laboratory prior to breakfast during whey and control trials for HW and OV/ OB participants (p > 0.05).Standardisation of all appetite sensations was achieved with no differences in total postprandial AUC following breakfast and lunch during whey and control trials in both HW and OV/ OB males (p > 0.05), as presented in Table 3. Composite appetite scores were also similar following breakfast and lunch meals in HW and OV/OB males during both trials (p > 0.05; Fig. 2).

Post-laboratory phase
In the evening of the test meal, there were no compensatory behaviours in food intake in response to whey and control trials, with similar caloric intake reported for HW (Whey 310.9 ± 268.

Discussion
We report a statistically significant and meaningful reduction in food intake following whey protein ingestion in both HW (193.4 kcal, 15.9 %) and OV/OB (215.81 kcal, 17.8 %), when compared to placebo.Considering that 68 kcal was consumed as part of the whey protein preload, the net energy deficits achieved during the test meal were 125.4 kcal and 147.81 kcal for HW and OV/OB, equating to 86 g and 102 g of pasta respectively.Interestingly, our analysis on self-reported Fig. 2. Time-course changes in composite appetite scores ([desire to eat + hunger + (100 -fullness) + prospective consumption]/4) following breakfast (A) and lunch (B) during whey and control trials in HW and OV/OB individuals (n = 12 for HW and OV/OB).Data are presented as mean ± SD.No between group differences were observed (p > 0.05).

Table 3
Postprandial areas under the curve (AUCs) for self-reported ratings of appetite following control and whey trials post-breakfast and post-lunch in HW and OV/ OB males.Data are presented as ± SD, percentages represent change as a percentage of the control trial.DTE, desire to eat; PI, prospective food intake.
diet diaries and physical activity accelerometery detected no significant differences in compensatory behaviour over the following 2 days.There is strong evidence from earlier studies that a whey protein preload reduces food intake at a subsequent test meal in healthy, lean populations (Akhavan et al., 2010;Anderson & Moore, 2004;Astbury et al., 2010;Chungchunlam et al., 2017;Hall et al., 2003;Zafar et al., 2013) with few conflicting studies (Chungchunlam et al., 2009).However, in studies including people living with overweight and obesity, the satiating effect of whey protein is more unclear, reporting reduced food intake (Bowen, Noakes, & Clifton, 2006;Zafar et al., 2013), or no effect (Bowen, Noakes, & Clifton, 2007;Poppitt et al., 2011).When compared to a glucose preload, two studies reported significant reductions in food intake in overweight cohorts (BMI; 30.7 ± 2.5 kg/m 2 ; 30.1 ± 1.1 kg/m 2 ) by 15.6 % and 10 % following the ingestion of 25 g and 50 g whey protein 180 min prior to a test meal, respectively (Bowen, Noakes, & Clifton, 2006;Zafar et al., 2013).Conversely, Bowen, Noakes, and Clifton (2007) observed increased VAS-rated fullness following 50 g whey when compared to fructose, but no significant difference in energy intake when ingested 240 min prior to a buffet meal in people with obesity (32.5 ± 0.6 kg/m 2 ).Similarly, Poppitt et al. (2011) reported that 20 g whey ingestion suppressed immediate postprandial measures of satiety, however the effects were short-term and not sufficient to significantly impact on subsequent food intake when measured 2 h later.
Therefore, to the authors knowledge this is the first study to investigate the effects of a low dose of whey protein on food intake when ingested only 20 min before an ad libitum meal in males with overweight and obesity.The anorexigenic effect we report may be due to the release of several gut peptides including cholecystokinin (CCK), glucagonlike peptide 1 (GLP-1), peptide YY (PYY) and insulin, along with the suppression of acylated ghrelin (Batterham et al., 2006;Blom et al., 2006;Burton-Freeman, 2008;El Khoury et al., 2006;Foster-Schubert et al., 2008;Hall et al., 2003), although the evidence for the latter is inconsistent (Cummings, 2006;Lejeune et al., 2006).The present experimental protocol did not allow for the investigation of these possible mechanisms which is a limitation, although it is possible that the small inter-meal interval may have aligned the test meal alongside peak GLP-1 and insulin concentrations, resulting in reduced energy intake (Bowen, Noakes, & Clifton, 2006;Ma et al., 2009).However, postprandial hormone responses do not always translate into a more satiating effect of a given protein (Juvonen et al., 2011;Veldhorst et al., 2009).Furthermore, ingestion of whey protein preloads < 90 min prior to feeding have been shown to slow gastric emptying, as assessed by the plasma concentrations of oral paracetamol consumed with the meal (Akhavan et al., 2014;Ma et al., 2009).In the present investigation, the reduction in food intake following whey ingestion may also be attributed to the reduced eating rate in both HW (19 %) and OV/OB (21.0 %).Indeed, evidence suggests that a slowed eating rate reduces energy intake (Bolhuis et al., 2014) and similar responses have been observed when whey protein is ingested after resistance exercise in males (Monteyne et al., 2018) but not females (Martin et al., 2007).
Our results suggest that a low dose of whey protein, consumed 20 min before a meal, elicits a meaningful energy deficit in both HW and people living with OV/OB, without compensatory changes in food intake and physical activity over the following 2 days.Therefore, in theory, weight loss strategies could incorporate whey protein as a tool to achieve long-term energy deficits.However, the feasibility of such a strategy remains unclear since little is known about the effects of chronic whey protein ingestion on food intake and body composition in people with overweight or obesity.Future research should focus on identifying whether the effect of whey protein preloading on energy intake persists with regular exposure, similar to its effects on reducing postprandial glycaemia over 4 weeks (Ma et al., 2015).Furthermore, unlike previous studies administering carbohydrate control groups (Baer et al., 2011;Pal & Ellis, 2010), comparisons between whey protein and no intervention would capture the effects of the caloric burden ingested within the preload within free-living conditions.
The current investigation was robust in its design to ensure standardisation of appetite perceptions and gut hormone parameters between trials.This design allows direct observation of energy intake at the breakfast, lunch and evening test meal in controlled environments, overcoming the issue of misreporting of intake.Where this wasn't possible for the evening meal prior to study days, participants received take away meals and snacks for ease of replication.Furthermore, the homogenous, uniform tomato-based pasta meal prevented confounding factors often observed in buffet-style test meals.Large variation in energy intake has been reported despite two separate buffet test meals with identical feeding conditions in the same individuals (Stensel, 2010), which may be due to preferences for expensive, palatable foods that may not be readily available in everyday life.However, our introduction of dietary reporting in the subsequent 2 days under free-living conditions relies on the compliance of the individual who may alter dietary habits, such as under-reporting by up to 30 % in overweight cohorts (Lichtman et al., 1992).
In conclusion, our findings show a meaningful reduction in food intake following a small whey protein preload in both healthy-weight males or those with overweight and obesity.There were no compensatory dietary and physical activity behaviours identified over the 2 days of post-intervention monitoring, suggesting whey protein is an effective tool for inducing an energy deficit in males.This could be effective in the long term, but future research is required to assess the effect of chronic whey protein supplementation on food intake and changes in body composition.

Fig. 1 .
Fig. 1.Schematic of experimental trial protocol.Document symbol denotes VAS completion, food images denote breakfast/lunch/evening meal, drinks bottle denotes beverage consumption.

Fig. 3 .
Fig. 3. Caloric intake at the ad libitum test meal for HW and OV/OB individuals following whey protein and placebo trials (n = 12 for HW and OV/OB).Data presented mean bar and individual before and after plot.* denotes whey and control trials different (HW, p = 0.003; OV/OB, p = 0.005).

Table 1
Nutritional composition of Lacprodan© whey protein concentrate.

Table 2
Participant characteristics categorised by body mass status.
Data presented as ± SD.BMI, Body mass index; RMR, Resting metabolic rate; PAL, Physical activity level.Physical Activity Level (PAL) was calculated by dividing participants' total daily energy expenditure by RMR.* denotes significant difference between groups (p < 0.05).