| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Human Nutrition (J.B., M.N., P.M.C.), Commonwealth Scientific and Industrial Research Organisation, Adelaide 5000, Australia; and Department of Physiology (J.B.), University of Adelaide, Adelaide SA 5000, Australia
Address all correspondence and requests for reprints to: J. Bowen, Commonwealth Scientific and Industrial Research Organisation, Human Nutrition, P.O. Box 10041 BC, Adelaide SA 5000, Australia. E-mail: jane.bowen{at}csiro.au.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Objective: The objective of the study was to assess postprandial responses to different protein sources, compared with glucose, in males with normal and high BMI.
Design: This was a randomized, crossover study of four preloads followed by blood sampling (+15, 30, 45, 60, 90, 120, 180 min) and buffet meal.
Setting: The study was conducted at an outpatient clinic.
Participants: The study population included 72 men, with a BMI range 20.6–39.9 kg/m2.
Interventions: Interventions consisted of liquid preloads (1.1 MJ, 450ml) containing 50 g whey, soy, gluten, or glucose.
Main Outcome Measures: Fasting and postprandial plasma glucose, insulin, ghrelin, glucagon-like peptide-1 (GLP-1) and cholecystokinin (n = 38), ad libitum energy intake, and appetite ratings were measured.
Results: Energy intake was 10% lower after all protein preloads, compared with the glucose treatment (P < 0.05), independent of BMI status and protein type. All protein loads prolonged the postprandial suppression of ghrelin (P < 0.01) and elevation of GLP-1 (P < 0.01) and cholecystokinin (P < 0.05). Fasting GLP-1 concentrations [overweight, 17.5 ± 1.3; lean, 14.7 ± 0.1 pg/ml (5.2 ± 0.4 and 4.4 ± 0.1 pmol/liter, respectively); P < 0.001] and postprandial responses (P = 0.038) were higher in overweight subjects.
Conclusions: Whey, soy, and gluten similarly tend to reduce ad libitum food intake 3 h later in lean and overweight males relative to glucose. Postprandial ghrelin, GLP-1, insulin, and cholecystokinin may contribute to this higher satiety after protein consumption. GLP-1 concentrations are increased in overweight subjects, which may affect satiety responses in this group.
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The type of protein ingested may also affect postprandial responses. EI was lower 1–1.5 h after consuming whey-based preloads, compared with casein (15) and egg albumin (16) treatments. However, in a similar study, we found no difference in EI 3 h after the two dairy proteins (13). Consumption of mixed meals (5 MJ) high in plant (soy, pea, or gluten) or animal (egg albumin, casein, gelatin) proteins did not affect EI, although this was assessed 8 h after the test meal (17). Gastrointestinal responses to plant- and animal-derived proteins may give further insight into the mechanism by which proteins differentially affect appetite, compared with carbohydrates.
Dietary manipulations that maximize satiety have obvious applications for the overweight population as a means of improving compliance with energy-restricted diets. However, appetite studies are frequently performed in lean samples (1, 2, 3, 4, 5, 9, 11, 14, 15, 16, 17, 18). Outcomes derived from lean subjects may not be applicable to overweight subjects due to differences in glucose metabolism, ghrelin regulation (19), and eating behavior (20). Macronutrient-specific effects on appetite regulation should also be compared across weight groups.
The aims of this study were to compare the effect of soy, whey, and gluten proteins in liquid preloads on subjective (appetite ratings) and objective (ad libitum EI) appetite markers and postprandial changes in plasma ghrelin, GLP-1, and cholecystokinin relative to a glucose control. We also investigated the effect of body weight on these outcomes. This study was performed in males because the menstrual cycle is known to influence EI (21).
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Healthy men with a body mass index (BMI) greater than 20 kg/m2, aged 20–65 yr and a stable body weight were recruited by public advertisement. Exclusion criteria were hypersensitivity to foods used in the study, a score of greater than 10 on the eating restraint section of the validated Three Factor Eating questionnaire (22), and illnesses or use of medications that affect glucose metabolism or appetite. Eighty-two participants were eligible for participation. Ten subjects withdrew before study commencement (n = 3 illness, n = 3 personal reasons, n = 4 lost to contact). The study was approved by the Commonwealth Scientific and Industrial Research Organisation’s Human Nutrition Ethics Committee, and all participants gave informed, written consent to participate.
Dietary protocol
Standardized evening meals (3.5 MJ, 26% of total energy from protein, 31% from fat, and 43% from carbohydrate), consisting of frozen meals, cheese (20 g), chocolate (15 g), and fruit (140 g) were consumed by participants before each visit to control intake.
The preloads were beef-flavored soups, which contained water, beef flavoring, vegetable oil (2 g; whey and glucose treatments only), and protein (whey, soy, gluten) or glucose (Table 1
). Energy, energy density (Table 1), palatability, and consistency were matched and preloads did not contain dietary fiber.
|
Experimental protocol
Subjects attended the outpatient clinic on four occasions with a 7-d interval between treatments. Subjects refrained from alcohol and strenuous exercise for 24 h before visits, consumed the standardized meal on the evening before, and fasted thereafter (water permitted).
Subjects’ weight and height were measured (Mettler scales, model AMZ14; A&D Mercury, Kinomoto, Japan) in light clothing. BMI was calculated as weight (kilograms) divided by height (square meters). Total fat mass [coefficient of variation (CV) 2.3 ± 8.7%] and total fat-free mass (CV 2.1 ± 0.4%) were assessed by tetrapolar, single-frequency (50 Hz) bioelectrical impedance analysis (BIA; IMP5TM; Impedimed Pty. Ltd., Brisbane, Australia) using the offline general algorithm. Duplicate measurements were made while subjects were supine (legs apart, arms not touching the body) and after emptying the bladder. BIA shows good correlation with the gold standard method for assessing body composition, dual-energy x-ray absorptiometry, although in absolute values BIA tends to underestimate percentage fat mass in obese males with a percentage fat mass greater than 25% (25).
A randomly selected subgroup of 18 lean (BMI < 25.0 kg/m2) and 20 overweight (BMI > 25.1 kg/m2) participants had blood samples collected during the 3-h test period at all visits. An indwelling cannula was inserted into a lower arm vein of these participants upon arrival at the unit. The remaining subjects (n = 34) had single fasting blood samples collected by venipuncture at each visit for analysis of insulin and glucose.
All subjects then completed a visual analog scale questionnaire asking "how hungry do you feel" and "how much food would you like to eat now?" Opposing extremes of each feeling were described at either end of a 100-mm horizontal line, and subjects marked the line to indicate how they felt at that moment (26).
The preloads were served in a randomized order and consumed within 7 min. All subjects completed the appetite questionnaire at 15, 30, 45, 60, 90, 120, and 180 min after commencing the preload. Blood samples were also collected from the subgroup of cannulated subjects at these times for analysis of plasma cholecystokinin, ghrelin, GLP-1, glucose, and insulin. Cannulae were removed after the final blood sample. All subjects were then offered a buffet-style lunch, at which each subject was provided with large servings of all four foods. Instructions were given to eat until comfortably satisfied, and the food was removed after 30 min.
Biochemistry
Blood was collected into prechilled sodium fluoride/EDTA (1 g/liter) tubes for plasma insulin and glucose analysis. Aprotinin (500 KIU/ml of blood; Trasylol; Bayer, Leverkusen, Germany) was added to tubes for plasma cholecystokinin and ghrelin analysis and dipeptidyl peptidase-IV inhibitor (10 µl/ml blood; Linco, St. Charles, MO) was added to tubes for plasma GLP-1 analysis. Blood samples were stored on ice, and the plasma was isolated within 30 min of collection by centrifugation (10 min, 2000 x g, 5 C) (Beckman GS-6R Centrifuge; Fullerton, CA). Aliquots were stored at –80 C.
Commercially available RIA kits were used to measure total ghrelin (Phoenix Pharmaceuticals, Belmont, CA; CV 5.5%) and cholecystokinin-8 (Euria-Diagnostica, Malmo, Sweden; CV 14%). Ethanol extraction was performed on plasma for cholecystokinin analysis according to the manufacturer’s instructions. Active GLP-1(7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37) was measured by fluorescence immunoassay (Linco; CV 8.0%). Plasma insulin was measured using an ELISA immunoassay kit (Mercodia, Uppsala, Sweden). Plasma glucose was determined using an enzymatic kit (Hoffmann-La Roche Diagnostics, Basel, Switzerland) and control sera on a Hitachi 902 automatic analyzer (Roche Diagnostics, Basel, Switzerland).
Statistics
Results are expressed as means ± SEM and are presented for 72 subjects for all baseline characteristics, appetite, and ad libitum EI. BIA data are presented for 70 subjects; two subjects with internal metal pins were not measured. Postprandial blood parameters are presented for 38 subjects. For analysis of the appetite questionnaire, the baseline value was subtracted from postprandial responses to normalize between-subject differences, and total area under the curve (AUC) was calculated using a trapezoidal equation.
ANOVA with repeated measures was used to determine the effect of time (minutes) and treatment. BMI status (lean; BMI < 25 kg/m2, overweight; BMI > 25.1 kg/m2) was included as a between-subject factor, age was a covariate, and Bonferroni adjustments were used for multiple comparisons. Where ANOVA showed a significant main effect, Tukey’s post hoc tests were performed to compare group differences. Appetite AUC was compared using two-way ANOVA with treatment and BMI as factors. Relationships between EI and glucose, insulin, ghrelin, GLP-1, and cholecystokinin at all time points were examined using multiple linear regressions, which adjusted for the repeated nature of the data. Percentage variance is derived from the r2 value. The satiety hormone that might be most closely related to the impact of meal type on EI was examined in an analysis of covariance (ANCOVA) with EI as the dependent variable using general linear model. Differences between treatments at the 3-h time point were examined first. All statistical analysis was performed using SPSS 11.5 for WINDOWS (SPSS Inc., Chicago, IL) except the ANCOVA (SAS-STAT; SAS Institute, Cary, NC). Differences are considered significant if P < 0.05.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Seventy-two men completed the study (Table 2). There were no differences in baseline characteristics or responses to preloads between subjects allocated to the group in which fasting blood samples were drawn compared with multiple blood samples (data not shown; P > 0.05). All preloads were well tolerated by participants.
|
Ad libitum EI was significantly higher after the glucose treatment than gluten (+560 ± 136 kJ; P < 0.05; Table 3). A similar, nonsignificant trend was observed with lower intake after the soy and whey treatments.
|
Appetite ratings (P > 0.05, data not shown) and change (
) in desired amount of food AUC or hunger AUC (P > 0.05; Table 3
) were independent of treatment.
Glucose and insulin
Postprandial plasma glucose increased after the glucose treatment (time-by-treatment effect; P < 0.0005; Fig. 1A). Peak postprandial insulin occurred at 30–45 min and was highest after the glucose treatment (time-by-treatment interaction; P < 0.01; Fig. 1B).
|
There was a time-by-treatment interaction such that the ghrelin nadir after all protein treatments occurred later (90–120 min) and remained below baseline at 180 min, compared with the glucose treatment (time by treatment interaction, P < 0.01; Fig. 1C).
The highest GLP-1 concentration occurred later (30–90 min) and remained elevated for longer after all protein treatments, compared with the glucose preload (time by treatment interaction; P < 0.01; Fig. 1D).
Cholecystokinin remained elevated until 120 min after the protein preloads, whereas concentration declined after the initial peak at 15 min for the glucose treatment (time by treatment interaction; P < 0.05; Fig. 1E).
Body weight
Age was significantly higher in overweight subjects, compared with lean subjects (P < 0.01; Table 2). Inclusion of age as a covariate did not influence outcomes. Postprandial plasma glucose and appetite hormones were assessed in 18 lean (mean BMI 23.2 ± 0.3, range 21.3–24.9 kg/m2) and 20 overweight (mean BMI 31.4 ± 0.8, range 25.2–39.9 kg/m2) subjects.
There was no effect of BMI status on mean overall EI (lean, 3371 ± 141 kJ; overweight, 3310 ± 105 kJ; P > 0.05). Baseline and AUC for both appetite questions were also independent of body weight (P > 0.05).
Fasting (P < 0.001; Table 2) and postprandial response (P < 0.001; Fig. 2) for glucose and insulin were higher in overweight subjects. There was a time by treatment by BMI status interaction such that the postprandial change in plasma glucose after the glucose preload was smaller in lean subjects, compared with overweight subjects (P < 0.05; Fig. 2A).
|
), postprandial response (P = 0.003; Fig. 2C), and change from fasting to nadir concentrations were lower in overweight compared with lean subjects [–97.7 ± 7.5 vs. –151.8 ± 16.2 pg/ml, respectively (–28.9 ± 2.2 vs. –44.9 ± 4.8 pmol/liter); P < 0.01]. Fasting GLP-1 concentrations (P < 0.001; Table 2), postprandial responses (P = 0.038; Fig. 2D), and change from fasting to peak [21.5 ± 2.8 vs. 13.1 ± 1.8 pg/ml, respectively (6.4 ± 0.8 vs. 3.9 ± 0.5 pmol/liter); P < 0.01], were higher in overweight subjects. The macronutrient-specific changes in ghrelin and GLP-1 described above were observed in both lean and overweight subjects, despite overall differences in concentration. There was no significant effect of body weight on cholecystokinin. Regression analysis
The relationship among satiety hormones, glucose, and insulin was explored in a regression model using all time points. Glucose (P = 0.0005), insulin (P = 0.002), and GLP-1 (P = 0.0005) predicted 9.7% of the variance of ghrelin. Cholecystokinin was the only predictor of GLP-1 and explained 25.2% of the variance (P = 0.0005). Insulin and GLP-1 explained 15.3% of the variation in cholecystokinin (P = 0.0005).
Power calculations
With 72 subjects completing the study, we had 80% power (P < 0.05) to detect a difference between treatments of 200 kJ in EI (
7%). Thus, the differences observed between glucose and gluten treatment were at the threshold of detection.
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Differences in EI previously observed 1–1.5 h after consumption of various dietary proteins in liquid preloads (15, 16) were not replicated after 3 h in the present study, which reflects a more typical intermeal interval. However, an effect of protein type on appetite may arise when the proteins are consumed in solid form due to variation in the rate of digestion and nutrient absorption. This may be important because it is the main form of protein consumed in the free living setting. Indeed a hunger-fullness rating was higher 2–3 h after consuming fish, compared with chicken and beef; however, palatability and EI were not assessed (27). EI was not affected after mixed meals (1.8–5.0 MJ) in which the protein source in test foods (mousse and soup) was covertly manipulated (17, 28), although EI was recorded after 8 h and the consistency of the test foods was similar (17, 28). The use of liquid preloads in the present study enabled a comparison of gastrointestinal responses with plant- and animal-derived proteins and controlling for potential differences in gastric emptying rate. The influence of varying gastric emptying and digestion rates of solid proteins on hormonal responses remains unclear.
Our findings suggest that the higher satiety associated with consumption of protein, compared with carbohydrate, may be at least partially mediated by prolonged ghrelin suppression. Such a reduction in the orexigenic signal may delay the initiation of a subsequent feeding episode or lower hunger and EI. Ghrelin remained below baseline for longer after high protein mixed meals with yogurt as the primary protein source (29, 30). However, appetite and EI either were not assessed (29) or did not correlate with ghrelin (30) in these studies. Two studies reported that plasma ghrelin concentration remain unchanged 3 h after consumption of cooked turkey (31) and 4 h after pork (32). These conflicting findings may be due to the test food form; the influence of solid forms of protein on postprandial ghrelin may require assessment over a period longer than 3–4 h. This is because slow gastric emptying delays the postprandial ghrelin nadir (33) and changes in ghrelin require postgastric feedback (34). Indeed, treatment effects for ghrelin are not observed until 120–180 min after liquid (13, 33, 35) and semisolid (yogurt-rich meal) (29) preloads.
We report that postprandial GLP-1 secretion was prolonged after dietary proteins, compared with the glucose treatment. Earlier studies are suggestive of a pattern of longer GLP-1 elevation after protein-rich mixed meals, relative to carbohydrate and fat (14, 36). Our findings show a clear distinction in the pattern of secretion between proteins and glucose treatments, and higher GLP-1 at 180 min was related to lower EI. It remains unclear whether GLP-1 reduces appetite centrally and/or as a consequence of slowing gastric emptying (37). If the latter is important, the liquid preloads used in this study (which bypass gastric distension-induced satiety signals) may have compromised the influence of GLP-1 on appetite regulation. Nevertheless, our observation of prolonged GLP-1 stimulation after protein consumption suggests that it may contribute to the satiety associated with dietary proteins.
The postprandial cholecystokinin response was temporally similar to GLP-1; cholecystokinin remained elevated for 1.5 h longer after the protein preloads, compared with glucose. This confirms previously reported macronutrient- related differences after duodenal infusion (38) and oral consumption (12, 13) of protein, carbohydrate, and fats. Whereas cholecystokinin is typically associated with meal termination (i.e. satiation), our findings suggest it may also contribute to greater satiety (i.e. delay the return of hunger), supported by the inverse association between cholecystokinin and EI that almost reached significance (P = 0.056).
The macronutrient-specific effects on gastrointestinal hormones reported in this study are small and relationships are relatively weak; however, we show that a range of hormones respond similarly after consumption of different macronutrients and in a coordinated way that is likely to influence acute appetite regulation. The ANCOVA results indicate that the low glucose and low GLP-1 at the end of the sampling period had an influence on EI. The present study also shows that ghrelin has an inverse secretory pattern to GLP-1 and cholecystokinin, confirming two similar recent observations (13, 33). Interestingly cholecystokinin inhibits the central effects of ghrelin in animals (39).
We believe this is the first demonstration that fasting and postprandial GLP-1 concentrations are significantly increased in overweight subjects, compared with lean counterparts, independent of the macronutrient consumed. It has previously been reported that GLP-1 is reduced in overweight and obese subjects (8), although the evidence to support this is limited; incremental GLP-1 AUC (but not total AUC or fasting concentration) was lower in overweight compared with lean subjects (40). Fasting and postprandial responses to a breakfast plus guar gum solution were independent of body weight (41); however, lean subjects had slightly higher GLP-1 concentration at 30 min, compared with obese subjects, when the breakfast was consumed with water (41). Duodenal infusion of fat and carbohydrate produced a similar GLP-1 response in lean and obese subjects (42). The higher GLP-1 we observed in overweight subjects appears to be an appropriate response to a positive energy balance and would presumably lower appetite. Yet there was no effect of weight status on EI or appetite ratings. This implies that overweight subjects may have diminished sensitivity to this hormone, which is analogous to the established reduced sensitivity to insulin and leptin in this population. A metaanalysis found that GLP-1 infusion had a weaker effect on reducing ad libitum EI in obese compared with lean subjects, although this did not remain significant after controlling for the overall lower intake in obese participants (8). The similarity in EI in our study may also reflect a behavioral effect of randomly mixing lean and overweight subjects in groups and eating the buffet lunch together.
In summary, we show that dietary protein consumed in liquid preloads prolongs the postprandial suppression of ghrelin, elevation of cholecystokinin and GLP-1, and maintenance of glucose levels, compared with glucose ingestion. These responses are not affected by the type of protein consumed (soy, whey, or gluten) and are similarly observed in lean and overweight subjects, regardless of the overall differences in hormone levels. Preload type influenced EI through ghrelin, cholecystokinin, and glucose, whereas insulin was a predictor of EI that was independent of treatment, although overall these relationships were relatively weak. We have also observed increased fasting and postprandial GLP-1 concentration in overweight males, although this does not appear to affect appetite or EI.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Disclosure statement: The authors have nothing to disclose.
First Published Online May 30, 2006
Abbreviations: , Change; ANCOVA, analysis of covariance; AUC, area under the curve; BIA, bioelectrical impedance analysis; BMI, body mass index; CV, coefficient of variation; EI, energy intake; GLP, glucagon-like peptide.
Received March 20, 2006.
Accepted May 18, 2006.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  E. R Dove, J. M Hodgson, I. B Puddey, L. J Beilin, Y. P Lee, and T. A Mori Skim milk compared with a fruit drink acutely reduces appetite and energy intake in overweight men and women Am. J. Clinical Nutrition, July 1, 2009; 90(1): 70 - 75. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 X. Yin, Y. Li, G. Xu, W. An, and W. Zhang Ghrelin fluctuation, what determines its production? Acta Biochim Biophys Sin, March 1, 2009; 41(3): 188 - 197. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. E. Foster-Schubert, J. Overduin, C. E. Prudom, J. Liu, H. S. Callahan, B. D. Gaylinn, M. O. Thorner, and D. E. Cummings Acyl and Total Ghrelin Are Suppressed Strongly by Ingested Proteins, Weakly by Lipids, and Biphasically by Carbohydrates J. Clin. Endocrinol. Metab., May 1, 2008; 93(5): 1971 - 1979. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. J Moran, M. Noakes, P. M Clifton, G. A Wittert, C. W Le Roux, M. A Ghatei, S. R Bloom, and R. J Norman Postprandial ghrelin, cholecystokinin, peptide YY, and appetite before and after weight loss in overweight women with and without polycystic ovary syndrome Am. J. Clinical Nutrition, December 1, 2007; 86(6): 1603 - 1610. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. L. Luhovyy, T. Akhavan, and G. H. Anderson Whey Proteins in the Regulation of Food Intake and Satiety J. Am. Coll. Nutr., December 1, 2007; 26(6): 704S - 712S. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
), whey (
), gluten (
), or glucose (•) on postprandial plasma glucose (A), insulin (B), ghrelin (C), GLP-1 (D), and cholecystokinin (E) in males (n = 38). Data are expressed as mean ± SEM. For conversion of milligrams per deciliter to millimoles per liter for glucose, multiply by 0.056. For conversion of milliunits per liter to picomoles per liter for insulin, multiply by 6.95. For conversion of picograms per milliliter to picomoles per liter for glucose, multiply by 0.296. For conversion of picograms per milliliter to picomoles per liter for GLP-1, multiply by 0.298. There are significant time-by-treatment effects for glucose (P < 0.0005), insulin (P < 0.01), ghrelin (P < 0.01), GLP-1 (P < 0.01), and cholecystokinin (P < 0.05) (repeated measures ANOVA with Tukey’s post hoc test).
–; protein, –
-) males for plasma glucose (A), insulin (B), ghrelin (C), GLP-1 (D), and cholecystokinin (E). Data are expressed as mean ± SEM. For conversion of milligrams per deciliter to millimoles per liter for glucose, multiply by 0.056. For conversion of milliunits per liter to picomoles per liter for insulin, multiply by 6.95. For conversion of picograms per milliliter to picomoles per liter for glucose, multiply by 0.296. For conversion of picograms per milliliter to picomoles per liter for GLP-1, multiply by 0.298. There are significant time by BMI status effects for glucose (P < 0.0005), insulin (P = 0.001), ghrelin (P = 0.003), and GLP-1 (P = 0.038) (repeated measures ANOVA with Tukey’s post hoc test). There was a time by treatment by BMI status effect for plasma glucose (P < 0.05).