Acute Effects of Substitution, and Addition, of Carbohydrates and Fat to Protein on Gastric Emptying, Blood Glucose, Gut Hormones, Appetite, and Energy Intake

Whey protein, when ingested on its own, load-dependently slows gastric emptying and stimulates gut hormone concentrations in healthy young men. The aim of this study was to determine the effects of substitution, and addition, of carbohydrate (dextrose) and fat (olive oil) to whey protein. In randomized, double-blind order, 13 healthy young men (age: 23 ± 1 years, body mass index: 24 ± 1 kg/m2) ingested a control drink (450 mL; ~2 kcal/‘control’) or iso-volumetric drinks containing protein/carbohydrate/fat: (i) 14 g/28 g/12.4 g (280 kcal/‘M280′), (ii) 70 g/28 g/12.4 g (504kcal/‘M504′), and (iii) 70 g/0 g/0 g (280 kcal/‘P280′), on 4 separate study days. Gastric emptying (n = 11, 3D-ultrasonography), blood glucose, plasma insulin, ghrelin, cholecystokinin (CCK) and glucagon-like peptide-1 (GLP-1) concentrations (0–180 min), appetite (visual analogue scales), and ad-libitum buffet-meal energy intake (180–210 min) were determined. Substitution of protein with carbohydrate and fat was associated with faster gastric emptying (lower 50% emptying time (T50)), reduced suppression of ghrelin, and stimulation of GLP-1 (all P < 0.001); while the addition of carbohydrate and fat to protein did not affect gastric emptying or gut hormone responses significantly. Total energy intake (i.e., drink plus meal) was greater after all caloric drinks than control (P < 0.001). In conclusion, substitution of whey protein with dextrose and olive oil accelerated gastric emptying. Higher protein content of a mixed macronutrient drink increased gut hormone and insulin responses.


Introduction
The gastrointestinal tract plays a central role in determining appetite, energy intake, and postprandial glycemic excursions [1]. The stomach regulates the rate of delivery of ingested nutrients to the small intestine to optimize their digestion and absorption. When nutrients enter the small intestine they generate feedback signals that slow gastric emptying and suppress appetite, via both neural and hormonal mechanisms, which includes stimulation of cholecystokinin (CCK) and glucagon-like peptide (GLP-1) and suppression of ghrelin [2]. The lowering of postprandial blood glucose concentrations by GLP-1 is attributed in part to slowing of gastric emptying [3].

Subjects
Thirteen healthy men aged between 18 and 30 years (Mean ± standard error of mean (SEM): Age: 23 ± 1 years; body weight: 78 ± 2 kg; height: 1.79 ± 0.02 m; body mass index (BMI): 24 ± 1 kg/m 2 ) were recruited by advertisement. Exclusion criteria were alcohol abuse, use of illicit substances, smoking, use of medications known to potentially affect energy intake, appetite, or gastrointestinal motor function), significant gastrointestinal symptoms (abdominal pain, heartburn, diarrhea, or constipation), gastrointestinal surgery (apart from uncomplicated appendectomy), diabetes, gallbladder or pancreatic disease, known lactose intolerance or food allergies, low plasma ferritin levels or blood donation in the 12 weeks prior to the study, and failing to comprehend the study protocol. Subjects provided written informed consent. The study was approved by the Royal Adelaide Hospital Human Research Ethics Committee (140407), conducted in accordance with the Declaration of Helsinki, and registered as a clinical trial with the Australian New Zealand Clinical Trial Registry (www.anzctr.org.au; ACTRN12614000846628).

Protocol
This study had a randomized (using the method of randomly permuted blocks; www.randomization.com), double-blind, cross-over design. Subjects were studied on 4 occasions, each separated by 3-14 days, to determine the effects of substitution, and addition, of carbohydrate and fat to whey protein liquid supplements (Table 1) compared to control on gastric emptying, gut hormones, perceptions of appetite and gastrointestinal symptoms, and energy intake. In 'M 280 , 56 g, of the 70 g pure protein load ('P 280 ) was substituted for fat and carbohydrates, and in M 504 , the same amount of fat and carbohydrates was added. Drinks were matched for taste, served in a covered cup, and prepared by dissolving whey-protein isolate (Fonterra Co-Operative Group Ltd., Palmerston North, New Zealand) and dextrose (glucose), and homogenizing olive oil (Bertolli Australia Pty Ltd., Unilever Australasia, Sydney, NSW, Australia) in varying volumes of demineralized water and diet lime cordial (Bickford's Australia Pty Ltd., Salisbury South, SA, Australia).
Subjects consumed a standardized evening meal (beef lasagna (McCain Foods Pty Ltd., Wendouree, VIC, Australia),~591 kcal) on the night before each study day at 19.00 h. They were furthermore instructed to fast overnight from solids and liquids and to refrain from strenuous physical activity.
In each subject, blood sampling from an intravenous cannula, ultrasound measurements of gastric volume, and appetite ratings were obtained immediately before (during fasting), and at regular time intervals after ingestion of each drink until 180 min. Subjects were seated in an upright position on a wooden chair after removal of all metal objects, instructed to consume the drink within 2 min, and presented with a standard, cold, buffet-style meal in excess of what they were expected to consume (180 min). The meal had a total energy content of 2457 kcal, with 19% protein, 50% carbohydrates, and 31% fat. The subjects were allowed to eat for 30 min (180-210 min) until comfortably full. They were in a room by themselves whilst eating to limit external distractions.

Gastric Emptying by 3D Ultrasonography
Gastric volume was measured by a Logiq™ 9 ultrasound system (GE Healthcare Technologies, Sydney, NSW, Australia) with TruScan Architecture including a 3D sensor, equipped with a 3.5C broad spectrum 2.5-4 MHz convex transducer [12]. The stomach was scanned using a continuous translational movement along its long axis (~10 s) at 0, 5,15,30,45,60,75,90,105,120,135,150,165, and 180 min. The raw data (original scan planes) were transferred for 3D reconstruction and volume estimation using EchoPAC-3D software (GE Vingmed Sound, Horten, Norway). When ultrasound images lacked sufficient clarity to determine the volume of the stomach at some time points, data were imputed by linear interpolation [13]. In two subjects, the quality of ultrasound stomach images was insufficient to determine gastric emptying during one or more entire study days, and all data related to gastric emptying in these subjects were, therefore, excluded from analysis.
Intragastric retentions from 0-60 min 'early' and 60-180 min 'late' were calculated as total gastric volume minus baseline 'fasting' gastric volume at each time point, expressed as percentage of the maximal gastric volume (100%), i.e., volume of the ingested drink. The time at which 50% of the drink had emptied from the stomach (50% gastric emptying time; T50; min) and 'complete' gastric emptying time (100% gastric emptying time; T100; min) were calculated for all conditions. 'Complete emptying' was defined as the time when the residual volume of the drink in the stomach was ≤5%, and set to 180 min when the residual volume at 180 min was ≥5%. Rate of gastric emptying was calculated as the mean of rates of emptying (kcal/min) during each 15-min interval, respectively, of the 'early' (0-60 min) and 'late' (60-180 min) phases [13]. Glucose and Plasma Insulin, Ghrelin, CCK, and GLP-1 Concentrations   Blood samples were collected into ice-chilled, EDTA-coated tubes at 0, 5, 15, 30, 45, 60, 90, 120, 150, and 180 min. No inhibitors were added [14]. Blood glucose concentrations (millimoles per liter) were determined immediately after collection by the glucose oxidase method using a portable glucometer (Optium Xceed, Abbott Laboratories, Doncaster, VIC, Australia).
Peak/nadir and time to peak/nadir concentrations for blood glucose, plasma insulin, ghrelin, CCK, and GLP-1 were calculated for the caloric drink conditions.

Energy Intake
Energy and protein, carbohydrate, and fat intakes were determined at a buffet-style meal. The meal consisted of palatable food items, including sliced bread, margarine, ham, chicken, cheese, mayonnaise, fruit, fruit salad, yoghurt, custard, orange juice, iced coffee, and water [13]. Absolute energy intake (kcal), both as the intake at the buffet meal and as the cumulative energy intake (buffet meal plus preload drink), and the percentage suppression/change energy intake at the buffet meal (expressed as % of energy intake of the control day) by a given protein load compared to control, were calculated.

Data and Statistical Analyses
Statistical analyses were performed using SPSS software (version 22; IBM, Armonk, NY, USA). On the basis of our previous work [13], with an observed within-subjects SD of 31 min for gastric emptying half time, we calculated that 13 subjects would allow detection of a within-groups difference between treatments for T50 of 35 min, with power equal to 0.8 and alpha equal to 0.05. Differences between study conditions for gastric emptying, glucose, hormones, appetite, and energy intake were determined using one-way repeated-measures ANOVA, with the treatment as the within-subject factor, and Bonferroni post hoc comparisons. Interaction effects of time by treatment for glucose, hormones, and appetite were determined using a two-way repeated measures ANOVA, with treatment and time as the within-subject factors. Within-subject correlations were determined by using a general linear model with a fixed slope and random intercept [17]. Area under the curve (AUC) for gastric emptying, glucose, hormones, and appetite were calculated from baseline to 60 min (early phase), and 60 to 180 min (late phase), using the trapezoidal rule. Peak/nadir and time to peak/nadir hunger, fullness, prospective food consumption, desire to eat, nausea, and bloating were calculated for all conditions. Assumptions of normality were verified for all outcomes before statistical analysis.
Statistical significance was accepted at P < 0.05. Data in the text and tables are presented as mean values ± SD, and data in the figures are presented as mean values ± SEMs.

Gastric Emptying
The M 280 and control drink emptied in a non-linear pattern, whereas the pattern of emptying of M 504 and P 280 was more linear (Figure 1). Gastric emptying of all caloric drinks was slower compared to control (P < 0.05). Gastric emptying of P 280 and M 504 were slower (P < 0.001), compared to M 280 , with greater 50% and 100% gastric emptying times after both P 280 and M 504 compared to M 280 (P < 0.001, Table 2). Gastric emptying of M 504 was not significantly slower than P 280 , with the stomach taking 21% more time to empty 50%, (70 vs. 58 min), but only 4% more time to empty 100% (176 vs. 169 min) after the addition of fat and carbohydrate.

Insulin
Insulin increased after all caloric drinks (P 280 , M 280 , and M 504 ; interaction effect of time by drink-condition: P < 0.001, Figure 2); AUC 0-60min (P = 0.005) and peak concentrations (P = 0.001) were higher compared to control ( Table 2). M 280 and M 504 evoked a rapid increase in insulin concentrations, with higher AUC 60-180min concentrations after M 504 than M 280 (P = 0.004), while insulin remained elevated after M 504 and P 280 , with 180-min concentrations being higher after P 280 compared to M 280 (P < 0.009).

Ghrelin
Plasma ghrelin concentrations decreased after all caloric drinks; concentrations returned to baseline at~120 min after M 280 , and remained below baseline until the buffet meal at 180 min after P 280 and M 504 (interaction effect of time by drink-condition: P < 0.001, Figure 2). Nadir ghrelin concentrations following P 280 and M 504 were lower than control (P = 0.003, Table 2) and remained suppressed, with AUC 60-180min and 180-min concentrations being lower compared to M 280 and control (P < 0.001).

CCK
CCK increased promptly after all caloric drinks (interaction effect of time by drink-condition: P < 0.001, Figure 2). Peak and AUC 0-60min CCK concentrations after P 280 , M 280 , and M 504 were higher than control (P < 0.001, Table 2) and remained elevated, with AUC 60-180min and 180-min concentrations being higher than control, as well as after M 504 compared to M 280 (P < 0.001).

Correlations between Gastric Retention, Hormones, and Energy Intake
AUC 60-180min concentrations of GLP-1 were, within subjects, negatively (r = −0.33, P = 0.038), and perceptions of hunger and desire to eat and concentrations of ghrelin positively (range of r = 0.33 to 0.66, P < 0.05), correlated with energy intake. AUC 0-60min perceptions of hunger, desire to eat and prospective food consumption, and plasma gut hormones, but not blood glucose concentrations were, within subjects, associated with gastric retention; i.e., lower plasma ghrelin, perceptions of hunger, desire to eat, and prospective food consumption (range of r = −0.39 to −0.42, P < 0.05) and higher insulin, CCK, and GLP-1 concentrations (range of r = 0.52 to 0.65, P < 0.01) correlated with higher gastric retention. AUC 60-180min concentrations of ghrelin were, within subjects, negatively (r = −0.74, P < 0.001) correlated, and insulin, CCK, and GLP-1 positively (range of r = 0.73 to 0.76, P < 0.001), correlated with gastric retention.

Discussion
This study determined the acute effects of substitution, and adding, of carbohydrate and fat to protein on gastric emptying, glucose, gut hormones, appetite, and energy intake in healthy young men. Equi-energetic substitution of 224 kcal fat (112 kcal) and carbohydrate (112 kcal) for whey protein in the 280-kcal drinks (M 280 vs. P 280 ) resulted in faster gastric emptying, particularly during the early phase, greater increases in glucose concentrations, less prolonged suppression of ghrelin, a tendency for a smaller increase in CCK, and a marked abbreviation of the duration of the GLP-1 response. Thus, whey protein slowed gastric emptying more than an equi-energetic combination of protein, fat and carbohydrate under these study conditions. As both CCK [18,19] and GLP-1 [20,21] slow gastric emptying, the lesser increases in their blood concentrations after fat and carbohydrate than protein may, in part, account for the faster gastric emptying after the mixed macronutrient drink. The lower increases in CCK and GLP-1 after M 280 may be a result of the emptying of fat and protein/carbohydrates in different phases. The drinks in this study were mixed for~45 min until immediately before their ingestion and the macronutrients were evenly mixed at the time of ingestion. As the subjects were seated during the study, however, it is possible that the fat (olive oil; 9 kcal/g) separated from the protein/carbohydrate solution (4 kcal/g) after ingestion of the M 280 and M 504 drinks, and emptied from the stomach slower than the aqueous phase by 'layering' on the denser aqueous components [22].
The addition of 224 kcal fat (112 kcal) and carbohydrate (112 kcal) to 280 kcal whey protein (M 504 vs. P 280 ) produced a slight, but not significant, further slowing of gastric emptying and had no additional stimulatory effect on CCK and GLP-1, or suppressive effect on ghrelin, but predictably increased glucose and insulin concentrations above those after whey protein alone [23,24]. The lack of additional stimulation of CCK and GLP-1 may be an explanation for the lack of further slowing of gastric emptying, consistent with a greater slowing effect of protein than fat and carbohydrate, and a role for CCK and GLP-1 in mediating this effect. Our results are in line with the report that gastric emptying of a protein-enriched yoghurt preload is slower than that of a carbohydrate-enriched preload, associated with higher plasma CCK and GLP-1 concentrations after the high protein than high carbohydrate preload [9]. Administration of a fat and carbohydrate drink on its own was not part of this study and would be needed to test the possibility that 224 kcal combined fat and carbohydrate, unlike 280 kcal whey protein, has little slowing effect on gastric emptying, which is possible given the reduced slowing effect of fat and carbohydrate compared to equi-caloric whey protein demonstrated in the substitution part of the study. Alternately, the 280 kcal whey-protein drink may have produced near maximal nutrient-induced slowing, with fat and carbohydrate unable to slow emptying much more, even if they would do so if administered without this protein.
Adding protein to a glucose meal increases the insulin response [25][26][27], as with the addition of extra protein to a mixed macronutrient drink in this study (M 504 vs. M 280 ). This effect is particularly strong after ingestion of whey, which has a high content of insulinotropic amino acids [28], when compared to casein [29]. Despite the higher energy content of M 504 than M 280 , early phase blood glucose concentrations were lower after M 504 than M 280 . Factors contributing to this difference could include the higher blood insulin concentrations and an increased ratio of whey protein to dextrose, resulting in lower amounts of carbohydrates entering the small intestine, where they are absorbed. A daily meal pattern of three meals, with snacks in between, results in a limited number of hours of true fasting. Postprandial glycemia predominates over fasting blood glucose in contributing to glycated hemoglobin (HbA1c), a risk indicator of developing diabetes-related complications, and a deterioration in postprandial glycemic control precedes any substantial elevation of fasting blood glucose [30]. As such, our finding indicates that increasing the whey protein content of a drink may have postprandial glucose lowering effects and may be a valid nutritional strategy to manage glycaemia in patients with type 2 diabetes.
The postprandial plasma incretin hormone GLP-1 concentrations seemed to be particularly related to the protein content of the drinks; concentrations were higher after M 504 and P 280 , which had comparable responses from 30 min onwards, than after M 280 . Initial peak GLP-1 concentrations at 15 min after M 280 and M 504 were comparable and slightly higher than P 280 . GLP-1 after M 280 increased from 120 min onwards until before meal intake (180 min) after the transient initial peak, which may be related to the 'layering' of the fat component of the drink in the proximal stomach. Plasma CCK and ghrelin concentrations showed a similar pattern, although slightly less pronounced; when compared to M 280 , M 504 and P 280 had a more sustained decrease in ghrelin and slightly higher increases in CCK. Consistent with our results, preloads rich in whey [9,31], whey and egg white [32], whey and soy [33], casein [31], and dairy protein [34] resulted in suppressed plasma ghrelin concentrations [9,[31][32][33][34]. Protein vs. glucose [9,31], but not fat [35], resulted in increased plasma CCK concentrations, and protein vs. carbohydrate [9,[36][37][38] and fat [36,38,39] resulted in larger and more sustained plasma on GLP-1 concentrations in most, but not all [40,41], studies. Protein increased postprandial insulin concentrations more than fat [40].
The use of high protein supplements in young people is widespread, and increasing, in response to greater awareness that ingestion of protein has a muscle sparing effect and greater satiating effects than dietary carbohydrate and fat [4,5], making it attractive when weight control is desired. Before the test meals, more than 90% of all drinks had emptied from the stomach. Although hunger and desire to eat were lower after P 280 compared to control, there was a substantial effect on ghrelin, CCK, and GLP-1 at the time of meal ingestion, and energy intake was, within subjects, associated with plasma insulin, ghrelin, CCK, and GLP-1 concentrations, and perceptions of hunger in this study, ad libitum energy intake at the buffet meal was not different after the drinks, and, consequently, there was an increase in total energy intake during the caloric-drink study days compared to control. Energy intake at the buffet meal was assessed three hours after drink ingestion to allow for complete emptying of the drinks from the stomach and thus detailed assessment of gastric emptying and gut hormone responses. It has been reported in a systemic review that the optimal timing of a preload before a meal, if suppression of ad libitum food intake at that meal is intended, is much closer to the meal, with a time interval of 30 min or less associated with the greatest subsequent suppression of energy intake. Furthermore, suppression of energy intake is less after a liquid than semi-solid and solid preload [42].
Limitations included the relatively small subject numbers and blood glucose being measured by a glucometer, which is less than optimal. Nevertheless, any errors in glucose measurement would be expected to be uniform between study conditions and the glucose results were clear-cut. Olive oil and dextrose were used as a substitute for, and addition to, whey protein. Therefore, the results of this study are related to these single types of fat and carbohydrate. Furthermore, the nutrients were ingested in liquid form, and, therefore, the results may not translate to the administration of nutrients in solid form. While the drinks were palatable and matched for taste, we did not assess the subjects' perceptions of taste and/or pleasantness of the drinks. Energy intake at the buffet meal was assessed three hours after drink ingestion at a buffet meal, but not during the remainder of the day.

Conclusions
In conclusion, substitution of whey protein with dextrose and olive oil resulted in faster gastric emptying, less suppression of ghrelin, and less stimulation of CCK and GLP-1 concentrations; while adding the dextrose and olive oil to whey protein did not further slow gastric emptying, suppress ghrelin, or increase CCK and GLP-1 responses of the drink. A higher protein content of the mixed macronutrient drinks slowed gastric emptying and evoked stronger ghrelin, CCK, GLP-1, and insulin responses and a lower increase in blood glucose concentrations.