Mechanism of action of whole milk and its components on glycemic control in healthy young men

Available online at www.sciencedirect.com

ScienceDirect Journal of Nutritional Biochemistry 25 (2014) 1124 – 1131

Mechanism of action of whole milk and its components on glycemic control in healthy young men☆,☆☆, ★ Shirin Panahi a , Dalia El Khoury a , Ruslan Kubant a , Tina Akhavan a , Bohdan L. Luhovyy b , H. Douglas Goff c , G. Harvey Anderson a,⁎ b

a Department of Nutritional Sciences, Faculty of Medicine, University of Toronto, Toronto, ON, Canada M5S 3E2 Department of Applied Human Nutrition, Faculty of Professional Studies, Mount Saint Vincent University, Halifax, Nova Scotia, Canada B3M 2J6 c Department of Food Science, University of Guelph, Guelph, ON, Canada N1G 2W1

Received 13 April 2014; accepted 21 July 2014

Abstract Milk reduces post-meal glycemia when consumed either before or within an ad libitum meal. The objective of this study was to compare the effect of each of the macronutrient components and their combination with whole milk on postprandial glycemia, glucoregulatory and gastrointestinal hormones and gastric emptying in healthy young men. In a randomized, crossover study, 12 males consumed beverages (500ml) of whole milk (3.25% M.F.) (control), a simulated milk beverage based on milk macronutrients, complete milk protein (16g), lactose (24g) or milk fat (16g). Whole and simulated milk was similar in lowering postprandial glycemia and slowing gastric emptying while increasing insulin, C-peptide, peptide tyrosine tyrosine (PYY) and cholecystokinin (CCK), but simulated milk resulted in higher (41%) glucagon-like peptide-1 (GLP-1) and lower (43%) ghrelin areas under the curve (AUC) than whole milk (P=.01 and P=.04, respectively). Whole and simulated milk lowered glucose (P=.0005) more than predicted by the sum of AUCs for their components. Adjusted for energy content, milks produced lower glucose and hormone responses than predicted from the sum of their components. The effect of protein/kcal on the AUCs was higher than fat/kcal for insulin, C-peptide, insulin secretion rate, GLP-1, CCK and paracetamol (Pb.0001), but similar to lactose except for CCK and paracetamol, which were lower. The response in PYY and ghrelin was similar per unit of energy for each macronutrient. In conclusion, milk lowers postprandial glycemia by both insulin and insulin-independent mechanisms arising from interactions among its macronutrient components and energy content. © 2014 Elsevier Inc. All rights reserved. Keywords: Whole milk; Milk macronutrients; Glycemic control; Gastric emptying; Glucoregulatory and gastrointestinal hormones

1. Introduction Epidemiological studies have linked frequent dairy consumption with healthier body weights [1] and lower risk of type 2 diabetes (T2D) [2]. This possible link between milk consumption and obesity and T2D is of growing interest because milk and its components contribute to metabolic control, including postprandial glycemia [3,4].

Abbreviations: T2D, type 2 diabetes; M.F, milk fat; GLP-1, glucagon-like peptide-1; CCK, cholecystokinin; PYY, peptide tyrosine tyrosine; BMI, body mass index; ANOVA, analysis of variance; ISEC, Insulin SECretion; DPP-IV, dipeptidyl peptidase IV. ☆ This work was supported by the National Science and Engineering Research Council of Canada (NSERC-CRDJP 385597-09), Dairy Farmers of Ontario, and Kraft Canada Inc. ☆☆ Author disclosures: No conflicts of interests. ★ This trial was registered at clinicaltrials.gov as NCT01812967. ⁎ Corresponding author. Department of Nutritional Sciences, Faculty of Medicine, University of Toronto, 150 College Street, Toronto, ON, Canada M5S 3E2. Tel.: +1 416 978 1832; fax: +1 416 978 5882. E-mail address: [email protected] (G.H. Anderson). http://dx.doi.org/10.1016/j.jnutbio.2014.07.002 0955-2863/© 2014 Elsevier Inc. All rights reserved.

Fluid milk products reduce post-meal glycemia when consumed either before or within an ad libitum meal by healthy young adults [3,4] or with carbohydrate foods [5]. In a comparison of familiar beverages consumed before a meal, including isovolumetric (500ml) servings of 2% milk, 1% chocolate milk, orange juice, soy beverage, infant formula and water, postprandial glycemia following a pizza meal was lowest after milk [4]. Similarly, 1% milk, in contrast to other caloric and noncaloric beverages, when consumed at a pizza meal, reduced postprandial glycemia and produced the lowest post-meal appetite [3]. Milk consumed with carbohydrate foods, such as bread and pasta, also reduces postprandial glycemia than observed after the carbohydrate food alone [5]. The effect of milk on postprandial glycemia has been attributed to its stimulatory effect on insulin [6] because milk proteins, when consumed in beverage form or with carbohydrate, reduce glycemia [7], consistent with a rise in blood insulin concentrations [6,8]. Although milk proteins stimulate insulin, attributed to the rapid digestion and absorption of their branched-chain amino acids [9], post-meal reduction of glycemia after milk consumption may not be only due to its protein content and insulin release [7]. Milk proteins release gut hormones including glucagon-like peptide-1 (GLP-1), peptide tyrosine tyrosine (PYY) and cholecystokinin (CCK) which also

S. Panahi et al. / Journal of Nutritional Biochemistry 25 (2014) 1124–1131

affect blood glucose, delay stomach emptying [7,10,11] and suppress food intake [10,12]. In addition, milk is a complex mixture of proteins (whey protein and casein in the ratio of 20:80), fats (saturated, monounsaturated and polyunsaturated and trans-fatty acids), carbohydrate (lactose), and micronutrients with a wide range of bioactivities [13]. Milk fat is an important regulator of metabolic responses via the ileal brake, slowing stomach emptying and stimulating the release of gastrointestinal peptides such as CCK and PYY [14]; however, its relevance to glycemic control by milk has not been reported. Furthermore, lactose contributes to lower glycemia than other sugars or starch [5], but also simulated glucoregulatory and gastrointestinal hormones [15]. Thus, each of the macronutrients of milk may contribute to reducing glycemia and food intake [15–17]; however, there are no reports of the individual and collective contribution of milk's macronutrient components to the effect of whole milk. Therefore, we hypothesized that regulation of postprandial glycemia after milk consumption occurs through both insulin and insulin-independent actions due to interactions among its macronutrient components and energy content. The objective was to compare the effects of isovolumetric (500ml) beverages of whole milk (3.25% M.F.), each of its macronutrient components (protein, lactose and fat) and their combination (a simulated milk beverage) on postprandial glycemia, glucoregulatory and gastrointestinal hormones and gastric emptying in healthy young men. 2. Methods and materials 2.1. Participants Participants were recruited through advertisements posted at the University of Toronto campus. Healthy men between 20 and 30 years of age with a body mass index (BMI) of 20–24.9kg/m2 were eligible to participate. Exclusion criteria included smoking, dieting, skipping breakfast, lactose intolerance or allergies to milk, taking medications that may affect glucose metabolism, diabetes (fasting blood glucose ≥7.0mmol/L) or other metabolic diseases that could interfere with study outcomes. Based on previous clinical studies on gastrointestinal hormones with the sample size required for blood glucose response, 12 participants were recruited and completed the sessions [18]. Participants were financially compensated for completing the study. The procedures of the study were approved by the Human Subject Review Committee, Ethics Review Office at the University of Toronto. 2.2. Beverages Beverages included isovolumetric amounts (500ml) of (1) whole milk (3.25% M.F.; Neilson Dairy, Toronto, ON, Canada) (control); (2) complete milk protein (16g, whey protein/casein ratio of 20:80; American Casein Company, Burlington, NJ, USA); (3) lactose (24g; Davisco Foods International Inc., Eden Prairie, MN, USA); (4) milk fat (16g, from unsalted butter, 80% M.F.; Lactantia, Parmalat Canada Inc., Toronto, ON, Canada); and (5) a simulated milk beverage consisting of complete milk protein (16g), lactose (24g) and milk fat (16g). Each of the macronutrients and that in the simulated milk beverage were formulated at the same concentration as in whole milk. Whole milk (3.25% M.F.), which contains the highest fat content compared to other commercially available types of milk, was used as the control for the following reason. In our previous study [4], despite infant formula and chocolate milk having similar carbohydrate contents, infant formula, containing a higher fat content, resulted in lower pre-meal blood glucose concentrations compared to chocolate milk. Thus, we hypothesized that the fat component may also contribute to glycemic control. A simulated milk beverage was provided to assess the effects of the combination of macronutrients without the possible interference of other components of whole milk on glycemic control. Complete milk protein and lactose beverages were prepared at the Department of Nutritional Sciences at the University of Toronto by adding each of the powders to 500ml of water and stirred at room temperature for 20min until mixed. Milk fat beverages were prepared by the Department of Food Science at the University of Guelph from butter (80% M.F.; Lactantia, Parmalat Canada Inc.). Butter was added to water (at 4.35%), heated to 75°C and mixed using an industrial mixer. During the mixing process, a saturated monodiglyceride (0.2%, Danisco, Toronto, ON, Canada) was added as an emulsifying agent. The fat mixture was run through a two-stage homogenizer (Model 31MR [APV Gaulin Inc., Wimington, MA, USA], at 17.5/3.5MPa) to reduce the size and size distribution of milk fat globules. The milk fat beverage was pasteurized at 75°C for 15min, then heated to 90°C and poured into 500-ml sterilized

1125

bottles. Simulated milk beverages were prepared at the Department of Nutritional Sciences at the University of Toronto by adding lactose (24g) and protein (16g) (both in powder form) to 460ml of the ready-made milk fat beverages to achieve a volume of 500ml with 16g of fat and stirred at room temperature for 20min. Paracetamol (1.5g, Panadol; GlaxoSmithKline) was dissolved in each of the five beverages so the rate of its appearance in the blood can be used as a proxy to measure rate of gastric emptying [19]. Vanilla extract (1.2ml; Flavorganics, Newark, NJ, USA) and sucralose (0.02g; McNeil Specialty Products Company, New Brunswick, NJ, USA) were added to all beverages to equalize palatability and sweetness and blind the participants to the beverages. All beverages were isovolumetric (500ml) based on the commercially available serving size of milk beverages and were served chilled. The nutritional composition of the beverages is provided in Table 1. 2.3. Protocol This study was a randomized, crossover, single-blind design consisting of five sessions separated by a 1-week washout period to minimize any carryover effects. Individuals who fulfilled eligibility requirements were invited to participate in the study. Participants attended the Department of Nutritional Sciences at the University of Toronto following a 12-h overnight fast, except for water, which was permitted until 1h before each session. To minimize within subject variability, all participants were scheduled to arrive at the same time and on the same day of the week for each treatment, instructed to refrain from alcohol consumption and to maintain the same dietary and exercise patterns the evening before each test. To ensure that these instructions were followed, participants completed a questionnaire detailing presession information about their diet and lifestyle patterns. The order of beverages was randomized using a randomization block design, which was generated with a random generator script in SAS version 9.2 (SAS Institute Inc, Cary, NC, USA). On arrival, participants completed visual analog scale questionnaires assessing their “sleep habits,” “stress factors,” “food intake and activity level,” “feelings of fatigue” and “motivation to eat” [20,21]. Before the beginning of each test, each subject provided a baseline finger-prick capillary blood sample using a Monoejector Lancet device (Sherwood Medical, St. Louis, MO, USA) to ensure compliance with fasting instructions. Plasma concentration of glucose was measured with a glucose meter (Accu-Chek Compact; Roche Diagnostics Canada, Laval, Quebec, Canada). A baseline measurement ofN5.5mmol/L indicated noncompliance with the fasting instructions, and participants were rescheduled accordingly. Following the finger-prick blood glucose measurement, an indwelling intravenous catheter was inserted in the antecubital vein by a registered nurse and a baseline blood sample was obtained. Immediately thereafter, each person was instructed to consume one of the five beverages within 5min at a constant pace. Blood samples were collected at 0min (baseline) and at 30, 45, 60, 90, 120, 150 and 180min. Participants were asked to remain seated for the duration of the experimental session and were permitted to read, do homework or listen to music. 2.4. Blood parameters Blood was collected in 8.5ml BD P800 tubes (BD Diagnostics, Franklin Lakes, NJ, USA) containing spray-dried K2EDTA anticoagulant and a proprietary cocktail of additives which includes DPP-IV, esterase and other protease inhibitors to prevent the proteolytic breakdown of hormones. The tubes were centrifuged at 1300 RCF for 20min at 4°C. Collected plasma samples were aliquoted in Eppendorf tubes and stored at −70°C for analyses. Plasma concentrations of glucose, insulin, C-peptide, GLP-1, PYY, CCK, ghrelin and paracetamol were measured. Plasma glucose was measured using the enzymatic hexokinase method (intra-CV (coefficient of variation): b5%; inter-CV: b8%; Roche Diagnostic). Insulin was assessed with an electrochemiluminescence assay “ECLIA” (intra-CV: b3%; inter-CV: b7%; Roche Diagnostic). These analyses were performed by the Pathology and Laboratory Medicine

Table 1 Nutritional composition of beverages Composition a

Energy (kcal) Fat (total) (g) Carbohydrate (g) Lactose (g) Protein (g) Whey (g) Casein (g) a

Beverages b Whole milk (3.25% M.F.)

Simulated milk beverage

Protein

Lactose

Fat

300 16

300 16

64 0

96 0

144 16

0 16 3.2 12.8

24 0 0 0

0 0 0 0

24 16 3.2 12.8

24 16 3.2 12.8

Composition of each beverage as provided by the manufacturer. Amounts given are per 500ml serving. Paracetamol (1.5g), vanilla extract (1.2ml) and sucralose (0.02g) were added to all beverages. b

1126

S. Panahi et al. / Journal of Nutritional Biochemistry 25 (2014) 1124–1131

Department at Mount Sinai Hospital (Toronto, ON, Canada). The remaining biomarkers were measured at the Department of Nutritional Sciences, University of Toronto. To assess gastric-emptying rate, free paracetamol (acetaminophen) was measured with a commercially-available paracetamol enzymatic assay (intra-CV: b2.4%; interCV: b2.6%; Cambridge Life Sciences, Ely, Cambridge, UK). Human active GLP-1 (intraCV: b8%; inter-CV: b5%; CAT#EGLP-35K), total ghrelin (intra-CV: b2%; inter-CV: b8%; CAT#EZGRT-89K), total PYY (intra-CV: b6%; inter-CV: b8%; CAT#EZHPYYT66K) and Cpeptide (intra-CV: b4%; inter-CV: b8%; CAT#80-CPTHU-EO1.1) were measured with ELISA kits (Millipore, Billerica, MA, USA). Human CCK-8 [CCK-(26–33)] was measured with an enzyme immunoassay kit (intra-CV: b10%; inter-CV: b15%; CAT#EK-069-04, Phoenix Pharmaceuticals, Inc, Burlingame, CA, USA). Plasma concentrations of glucose, insulin, C-peptide and ghrelin were measured at all sampling times. However, due to the high cost of the kits and measurements, plasma concentrations of paracetamol, GLP-1 and PYY were measured at 0 (baseline), 30, 45, 60, 90, 150 and 180min and CCK measured at 0 (baseline), 30, 45, 60, 90 and 150min.

content [23], and food intake is primarily influenced by energy content [4,24], AUCs were also calculated per kcal of energy in the treatments. The insulin secretion rate (Table 2) was calculated from the deconvolution of plasma concentrations of C-peptide by using the ISEC (Insulin SECretion) computer program [25]. Pearson's correlation coefficients were used to detect associations between dependent measures. All results are presented as mean±standard error of the mean (S.E.M.). Statistical significance was concluded with a P-value less than .05.

3. Results 3.1. Subject characteristics Twelve healthy males (age: 22.4±0.4years; BMI: 21.9±0.3kg/m2) completed the study. An additional three participants withdrew from the study after the first session (n=2 personal time constraints; n=1 physical discomfort after the lactose beverage).

2.5. Statistical analyses Statistical analyses were conducted using SAS version 9.2 (SAS Institute Inc, Cary, NC, USA). Two-factor repeated-measures analysis of variance (ANOVA; via PROC MIXED procedure) were performed to analyze the effects of time, beverage and their interaction on dependent variables including plasma glucose, insulin, C-peptide, paracetamol and gastrointestinal hormones. When an interaction was found, onefactor ANOVA was performed followed by Tukey's post hoc test to compare the effect of beverages at each time of measurement. For analysis of all parameters, the baseline value was subtracted from postprandial responses to normalize between-subject differences. Incremental areas under the curve (AUCs) using the trapezoidal rule [22] were calculated for all measures. For each parameter, AUC values represented the positive or negative areas enclosed by the pre-ingestion baseline and the postingestion response curve until the return to baseline. Incremental AUC for whole milk and the simulated milk beverage were compared to the sum of the components by adding AUCs for protein, lactose and fat [AUC(sum)] for glucose, insulin, C-peptide, insulin secretion and gastrointestinal hormones. In addition, because the response of many hormones is related not only to macronutrient composition but also energy

3.2. Postprandial responses unadjusted for energy content On an energy-unadjusted basis, the milk components (16g of protein or fat, 24g lactose) evoked similar responses (AUCs) in insulin, insulin secretion rate, PYY, GLP-1, CCK and ghrelin. Plasma glucose was highest after lactose. The sum of the AUCs for the components predicted the response to whole milk and the simulated milk beverage for most measures, with the exception that plasma glucose AUC was only 44% of that predicted and the simulated milk beverage raised CCK more than that predicted from the sum of the components (Table 2).

Table 2 Incremental AUCs for glucose, insulin, C-peptide, insulin secretion rate, gastrointestinal hormones and gastric-emptying rate (plasma paracetamol) after consumption of 500ml of beverages 1 Biomarkers2,3

Whole milk

Simulated milk beverage

Protein

Lactose

P value 4

Fat

−1

Glucose (mmol*min*L AUC

)

AUC(sum) Insulin (pmol*min*L−1) AUC AUC(sum) C-peptide (pmol*min*L−1) AUC AUC(sum) Insulin secretion rate (pmol*kg−1*min−1) AUC AUC(sum) GLP-1 (pg*min*L−1) AUC AUC(sum) PYY (pg*min*L−1) AUC AUC(sum) CCK (ng*min*L−1) AUC AUC(sum) Ghrelin (pg*min*L−1) AUC AUC(sum) Paracetamol (mmol*min*L−1) AUC 1

18.3±3.1ab

18.3±3.3ab

18.3±3.1b

18.3±3.3b

4745±912a

5248±820a

4745±912

5248±820 a

10.4±3.2b

1930±340b

a

18,643±2066

18,643±2066

42±7a

53±9a

42±7

53±9 153±18a 153±18a

684±209b

7266±1866

17±4b

c

2117±474

8±3b

b.0001 NS

35±9c

1768±310

2056±357 1.6±0.2a

470±67b

50±10bc

697±138b

0.9±0.2b

−1440±313

−2515±533

−1440±313

−2515±533 7.7±0.5abc

701±169b

0.8±0.1b

∑2.4±0.3 b

−925±329

.004 .0005

b

−896±240

b

−828±180

∑−2649±449 6.8±0.4c

b.0001 NS

a

1.6±0.2

a

0.7±0.1b

b.0001 .01

∑1868±265

ab

b.0001 NS

∑128±17ab

2056±357a

b.0001 NS

b

∑15,321±2222

42±7c

1768±310a

7.0±0.6bc

1979±618b

∑39±6

91±18b

ab

5938±623

14±2b

91±18b

1.0±0.1

bc

.005 .0005

∑ 4592±564

15,793±1659

b

7.4±1.4b

∑ 41.7±6.8a

15,793±1659

1.0±0.1ab

23.9±5.1a

8.6±0.5ab

.002 .04

9.1±0.6a

b.0001

All values are±S.E.M. n=12. Values in the same row with different superscript letters are significantly different, Pb.05 (beverage effect using one-factor ANOVA with proc mixed procedure, Tukey's post hoc). 2 AUCs are calculated as change from baseline from 0 to 150min for CCK and 0 to 180min for glucose, insulin, C-peptide, insulin secretion, GLP-1, PYY, ghrelin and paracetamol. 3 AUCs for whole milk and the simulated milk beverage were compared to the sum (∑) of the components by adding AUCs for protein, lactose and fat [AUC(sum)] for glucose, insulin, C-peptide, insulin secretion, GLP-1, PYY, CCK and ghrelin. 4 NS=sonsignificant.

S. Panahi et al. / Journal of Nutritional Biochemistry 25 (2014) 1124–1131

1127

Whole milk and the simulated milk beverage resulted in a 56% lower AUC than the sum of protein, lactose and fat over 180min (P=.005). Plasma insulin concentrations (0–180min) were affected by time (Pb.0001), beverage (Pb.0001) and an interaction between time and beverage (Pb.0001). Plasma insulin concentrations were higher after whole milk and simulated milk beverage than after fat and protein beverages at the peak time of 30min (Pb.0001) and after lactose, protein and fat beverages at 45min (Pb.0001), 60min (Pb.0001) and 120min (Pb.0001) (Fig. 1B). Plasma insulin AUCs were higher after whole milk and the simulated milk beverage than after any one of the AUCs for protein, lactose and fat (Pb.0001) (Table 2). However their AUCs were similar to the AUC reflecting the sum of protein, lactose and fat. Plasma C-peptide concentrations (0–180min) were affected by time (Pb.0001), beverage (Pb.0001) and an interaction between time and beverage (Pb.0001). Plasma C-peptide concentrations were higher after whole milk and simulated milk beverage than after fat beverages at the peak time of 30min (Pb.0001) and after lactose, protein and fat beverages at 45min (Pb.0001), 60min (Pb.0001) and 120min (Pb.0001) (Fig. 1C). Plasma C-peptide AUCs were higher after whole milk and the simulated milk beverage than after any one of protein, lactose and fat (Pb.0001) (Table 2). Whole milk and the simulated milk beverage resulted in a C-peptide AUC similar to the sum of AUCs for protein, lactose and fat. 3.2.2. Insulin SECretion (ISEC) Insulin secretion AUCs were higher after whole milk and the simulated milk beverage than any one of protein, lactose and fat, but similar to their sum (Pb.0001) (Table 2). 3.2.3. Plasma GLP-1 concentrations Plasma GLP-1 concentrations (0–180min) were affected by time (Pb.0001), beverage (Pb.0001) and an interaction between time and beverage (Pb.0001). The interaction is explained by the higher plasma GLP-1 concentrations after whole milk compared to protein, lactose and fat at 30min (P=.007) and 45min (Pb.0001) (Fig. 2A) and the higher GLP-1 concentrations at 60 and 90min (Pb.0001) after the simulated milk beverage compared to whole milk and its components (Fig. 2A). Plasma GLP-1 AUC was higher after whole milk compared to lactose and protein, but not different from fat (Pb.0001) (Table 2). In addition, the simulated milk beverage resulted in higher (41% AUC) GLP-1 than whole milk, but was not different from the sum of protein, lactose and fat.

Fig. 1. Effect of whole milk and its separate components on plasma (A) glucose, (B) insulin and (C) C-peptide concentrations. Results are shown as change from baseline. Means with different superscripts are significantly different at each measured time (one-way ANOVA, Tukey–Kramer post hoc test, Pb.05). All values are means±S.E.M.s; n=12.

3.2.1. Plasma glucose, insulin and C-peptide concentrations Plasma glucose concentrations were affected by time (Pb.0001), beverage (P=.002) and an interaction between time and beverage (Pb.0001). Plasma glucose concentrations were higher after lactose, whole milk and simulated milk beverage compared to protein and fat beverages at 30min (Pb.0001) and 45min (P=.03) (Fig. 1A). Glucose concentrations increased by 20% after lactose alone, but increased only 12% after whole milk and the simulated milk beverage, reaching peak values at 30min (Fig. 1A). Plasma glucose AUC was higher after lactose compared to fat and protein beverages, but not different from whole milk or the simulated milk beverage (P=.005) (Table 2).

3.2.4. Plasma PYY concentrations Plasma PYY concentrations were affected by time (P=.02) and beverage (Pb.0001), but not a time and beverage interaction. Whole milk and the simulated milk beverage resulted in the highest PYY concentrations at 45min (P=.0004), 60min (P=.0007), 90min (Pb.0001), 150min (P=.004) and 180min (P=.003) compared to its components (Fig. 2B). Plasma PYY AUC was higher after whole milk and the simulated milk beverage compared to each of the components (Pb.0001) (Table 2). Whole milk and the simulated milk beverage resulted in a similar PYY AUC compared to the sum of protein, lactose and fat. 3.2.5. Plasma CCK concentrations Plasma CCK concentrations were affected by time (Pb.0001) and beverage (P=.002), but not a time and beverage interaction. Plasma CCK concentrations were higher after whole milk compared to the fat beverage alone at 30min (P=.02) (Fig. 2C), but were not different from the other beverages. The simulated milk beverage resulted in the highest plasma CCK concentrations at 45min (P=.005) and 60min (P=.02), and its AUC was higher than any one of protein, lactose and fat (Pb.0001), but not different from whole milk (Table 2). Whole milk

1128

S. Panahi et al. / Journal of Nutritional Biochemistry 25 (2014) 1124–1131

Fig. 2. Effect of whole milk and its separate components on gastrointestinal hormones (A) GLP-1, (B) PYY, (C) CCK and (D) ghrelin. Results are shown as change from baseline. Means with different superscripts are significantly different at each measured time (one-way ANOVA, Tukey–Kramer post hoc test, Pb.05). All values are means±S.E.M.s; n=12.

and the simulated milk beverage resulted in a 200% and 60% lower CCK AUCs than the sum of AUCs for protein, lactose and fat (P=.0005).

3.2.6. Plasma ghrelin concentrations Plasma ghrelin concentrations were affected by time (P=.004) and beverage (Pb.0001), but not a time and beverage interaction. All beverages reduced plasma ghrelin concentrations; however, the simulated milk beverage resulted in the lowest ghrelin concentrations at 30min (P=.0005), 45min (P=.0008), 90min (P=.007), 150min (P= .0004) and 180min (P=.02) (Fig. 2D) and remained below baseline for the duration of the study. Plasma ghrelin concentrations after protein and lactose beverages returned to above baseline levels after 120min (Fig. 2D). Ghrelin AUC was more negative after the simulated milk beverage compared to any one of protein, lactose and fat (P=.0003), but not different from whole milk (Table 2). In addition, the simulated milk beverage reduced ghrelin AUC by 160 % more than whole milk, and at all times of measurement produced the lowest below baseline blood concentrations (Fig. 2D). Whole milk and the simulated milk beverage resulted in a similar ghrelin AUC compared to the sum of AUCs for protein, lactose and fat.

3.2.7. Gastric-emptying rate (plasma paracetamol concentrations) Plasma concentrations of paracetamol were affected by time (Pb.0001), beverage (Pb.0001) and an interaction between time and beverage (Pb.0001). Plasma paracetamol concentrations were lower after whole milk and protein beverages at 45min (Pb.0001) and after whole milk, the simulated milk beverage and protein at 60min compared to lactose and fat beverages (Pb.0001) (Fig. 3). Whole milk and protein reduced plasma paracetamol AUC compared to fat (Pb.0001), but were not different from the simulated milk beverage (Table 2).

3.3. Postprandial responses adjusted for energy content The AUCs for glucose and all hormones adjusted for energy content were similar for whole milk and the simulated milk beverage (Table 3). However, all AUCs after the whole and simulated milk beverages were lower than predicted by summing their component effects. The effect of protein per kcal on the AUCs was higher than fat per kcal for insulin, C-peptide, insulin secretion rate, GLP-1, CCK and paracetamol (Pb.0001), but similar to lactose except for CCK and paracetamol, which were lower. The response in PYY and ghrelin was similar per unit of energy for each macronutrient.

Fig. 3. Effect of whole milk and its separate components on gastric-emptying rate (plasma paracetamol concentrations). Results are shown as change from baseline. Means with different superscripts are significantly different at each measured time (one-way ANOVA, Tukey–Kramer post hoc test, Pb.05). All values are means±S.E.M.s; n=12.

S. Panahi et al. / Journal of Nutritional Biochemistry 25 (2014) 1124–1131

1129

Table 3 Energy-adjusted incremental AUCs per 50kcal for glucose, insulin, C-peptide, insulin secretion rate, gastrointestinal hormones and gastric-emptying rate (plasma paracetamol) after consumption of 500ml of beverages 1 Biomarkers2,3

Whole milk

Simulated milk beverage

Protein

Lactose

P value 4

Fat

−1

Glucose (mmol*min*L AUC

)

AUC(sum) Insulin (pmol*min*L−1) AUC

3±0.5b

3±0.5b

b

b

3±0.5

790±150bc b

790±150

AUC(sum) C-peptide (pmol*min*L−1) AUC

bc

AUC(sum) Insulin secretion rate (pmol*kg−1*min−1) AUC

4650±500

7.5±1ab

11±2a

3800±950

750±150

16.5±3

28±3.5

16.5±3b

28±3.5b

295±50

340±60

AUC(sum) CCK (pg*min*L−1) AUC

295±50b

340±60b

160±25b

260±40b

AUC(sum) Ghrelin (pg*min*L−1) AUC

160±25b

260±40b

33±5.5

3±1b

.0006

−240±50

−415±90

−240±50b

−415±90b

b.0001 ab

18.5±4.5

b

17.5±3.5

∑69±9a 365±55

365±70

245±60

NS

205±40b

b.0001 200±35b

∑960±135a −725±255

−465±125

a

1.5±0.1

6±0.5

b

4.5±0.5

.0006 b.0001

−315±55

NS

∑−1505±285a

d

.01 b.0001

∑975±110a 555±125a

b.0001 b.0001

8.5±4a

a

b.0001 b.0001

c

∑23±4

AUC(sum) PYY (pg*min*L−1) AUC

1.5±0.1

ab

a

ab

d

250±150c

∑9150±1250a

7.5±1

b

AUC(sum) Paracetamol (mmol*min*L−1) AUC

a

b

7±1

AUC(sum) GLP-1 (pg*min*L−1) AUC

1050±300ab ∑2850±600

ab

.0005 b.0001

a

865±150

3500±400b

b

1500±250a

b

3500±400

2.5±0.5b

∑22.5±4

865±150bc

2650±275

12±2.5a a

3±0.5

2650±275b 7±1ab

8±2.5ab

b.0001 c

3.5±0.5

b.0001

1

All values are±S.E.M. n=12. Values in the same row with different superscript letters are significantly different, Pb.05 (beverage effect using proc mixed procedure, Tukey's post hoc). 2 AUCs are calculated as change from baseline from 0 to 150min for CCK and 0 to 180min for glucose, insulin, C-peptide, insulin secretion, GLP-1, PYY, ghrelin and paracetamol. 3 AUCs for whole milk and the simulated milk beverage were compared to the sum (∑) of the components by adding AUCs for protein, lactose and fat (AUCsum) for glucose, insulin, C-peptide, insulin secretion, GLP-1, PYY and ghrelin. 4 NS=nonsignificant.

3.4. Relations between dependent measures Blood glucose AUC was associated with AUC for insulin (r=0.30, P=.02), but not GLP-1, PYY, CCK or paracetamol. Insulin AUC was positively associated with AUC for PYY (r=0.31, P=.02), but not GLP-1 or CCK and negatively associated with ghrelin (r=−0.38, P=.003) and paracetamol (r=−0.53, Pb.0001). C-peptide AUC was associated with glucose (r=0.1, P=.04), insulin (r=0.88, Pb.0001), PYY (r=0.47, P=.0001), ghrelin (r=−0.88, P=.0002) and paracetamol (r=−0.56, Pb.0001), but not GLP-1 or CCK. GLP-1 AUC was negatively associated with paracetamol (r=−0.30, P=.02), but was not associated with PYY, CCK or ghrelin. PYY AUC was negatively correlated with AUC for ghrelin (r=−0.31, P=.02) and paracetamol (r=−0.52, Pb.0001).

CCK AUC was positively correlated with paracetamol (r=0.29, P= .02). Ghrelin AUC was also positively correlated with paracetamol (r= 0.31, P=.02). Caloric content of beverages were positively correlated with insulin (r=0.56, Pb.0001), C-peptide (r=0.68, Pb.0001) and PYY (r=0.35, P=.005), but not GLP-1 or CCK and negatively associated with ghrelin (r=−0.29, P=.03) and paracetamol (r=−0.28, P=.03) (Table 4). 4. Discussion The hypothesis that regulation of postprandial glycemia after milk consumption occurs through both insulin and insulin-independent actions due to interactions among its macronutrient components

Table 4 Relations between dependent measures a Dependent variables

Glucose AUC (mmol*min/L)

Insulin AUC (pmol*min/L)

C-peptide AUC (pmol*min/L)

GLP-1 AUC (pg*min/L)

PYY AUC (pg*min/L)

CCK AUC (ng*min/L)

Ghrelin AUC (pg*min/L)

Paracetamol AUC (mmol*min/L)

Glucose AUC (mmol*min/L) Insulin AUC (pmol*min/L) C-peptide AUC (pmol*min/L) GLP-1 AUC (pg*min/L) PYY AUC (pg*min/L) CCK AUC (ng*min/L) Ghrelin AUC (pg*min/L) Paracetamol AUC (mmol*min/L) Caloric content (kcal)

– r=0.30 r=0.10 NS NS NS NS NS NS

r=0.30 – r=0.88 NS r=0.31 NS r=−0.31 NS r=0.56

r=0.10 r=0.88 – NS r=0.47 NS r=−0.88 r=−0.56 r=0.68

NS NS NS – NS NS NS NS NS

NS r=0.31 r=0.47 NS – NS r=−0.31 NS r=0.35

NS NS NS NS NS – NS NS NS

NS r=−0.38 r=−0.88 NS r=−0.31 NS – NS r=−0.29

NS r=−0.53 r=−0.56 r=−0.30 r=−0.52 r=0.29 r=0.31 – r=−0.28

NS=nonsignificant. a All values are±S.E.M. n=12. Pearson's correlation coefficients (r) indicate significant differences between dependent measures, Pb.05.

1130

S. Panahi et al. / Journal of Nutritional Biochemistry 25 (2014) 1124–1131

and energy content is supported. The postprandial glycemia after whole and simulated milks was much lower than predicted from the sum of the glycemic effects of their protein, lactose and fat contents, even though the glucoregulatory hormones, insulin, C-peptide and GLP-1 reflected the sums of the components. AUCs expressed per kcal show that each of the macronutrients differs in impact per kcal on glucose and hormone responses and that the stimulatory effects of whole and simulated milks per kcal were markedly attenuated compared with that predicted by the sum of the AUCs for protein, lactose and fat. Thus, the approach taken herein to assess the contribution of macronutrient components in foods with intrinsic health properties shows that the functionality of the whole is more than a sum of the parts. The glycemic response of the milks was 56% lower than that calculated as the sum of the AUCs for their protein, lactose and fat contents (Table 2). However, plasma insulin, C-peptide and insulin secretion rate AUCs were similar after whole milk and the simulated milk beverage compared to the sum of the individual responses to protein, lactose or fat (Table 2). Altogether, these results suggest that whole milk affects postprandial glycemia through mechanisms in addition to that which can be assumed from blood concentrations of insulin. One possibility is enhanced insulin function [26] as the insulin secretion rate was not increased. This observation is consistent with previous studies showing that whey protein ingestion reduces postmeal glycemia in a dose-dependent manner without increased Cpeptide release or insulin concentrations indicating that whey also stimulated insulin-independent mechanisms controlling glucose metabolism [7,18]. The most probable explanation for the insulin-independent actions of whole milk on blood glucose control resides in the rate of gastric emptying [27]. GLP-1, PYY and CCK slow gastric-emptying rate, and their role is supported in our study by the negative correlations observed between these hormones and paracetamol concentrations (Table 4). GLP-1 [21], PYY [22] and CCK [28] rapidly cross the blood–brain barrier to directly transmit signals that inhibit gut motility and delay gastric emptying, thus lowering postprandial glycemia without concurrent increases in insulin [18]. In addition, both GLP-1 [29] and PYY [30] augment portal-mediated glucose clearance. Furthermore, paracetamol AUCs were similar after whole milk, the simulated milk beverage and protein but lower after protein than fat (Table 2), and its concentration in blood after fat and lactose at 45 and 60min was highest, indicating a more rapid stomach emptying (Fig. 3). Thus, these results are consistent with other studies reporting reduced gastric-emptying rate after protein consumed either alone [18,31] or with carbohydrate [27]. A novel finding was that the simulated milk beverage resulted in a higher AUC and more sustained concentrations of GLP-1 (Fig. 2A) and greater suppression of ghrelin, as indicated by a larger AUC below baseline for ghrelin (Fig. 2D) compared to whole milk. The ingestion of fats, carbohydrates and protein results in a rise in GLP-1 [32,33] (Table 2), which enhances glucose-stimulated insulin secretion [33]. However, the higher and greater duration of GLP-1 after the simulated milk beverage suggests that source and processing of the macronutrients was a factor. The milk protein concentrate used in this study contained micellar casein in a powder form after spray drying which results in a large particle size that is less soluble than intrinsic proteins in milk. The peak size for milk was approximately 0.2μm compared to 100μm for the milk protein concentrate, as assessed by laser light scattering (Malvern Mastersizer; data not shown). Thus, while similar in chemical composition to milk, it was not in structure. In addition, homogenization of whole milk rearranges physical properties of casein micelles such that they adsorb to fat which affects milk structure and bioactivities [34] adding a further possible difference in the functionality of these milks. Furthermore, the absence of micronutrients in the simulated milk may have been a factor. Milk minerals attenuate the rise in total cholesterol and low-density

lipoprotein cholesterol concentrations due to a high-fat diet in rats and humans [35]. A better understanding of the cause of these differences may have been achieved if the milk macronutrients were obtained from the same sample of homogenized whole milk. The results of this study show that functionality of foods is not readily predicted from the composition of their macronutrient components alone. Because a primary objective of this study was to examine the contribution of individual macronutrients in amounts within milk to the effect of whole milk on glycemic control and gut hormones, energy content of the treatments varied from 64 to 304kcal. Consistent with the literature, caloric content of the treatments was positively correlated with insulin, C-peptide and PYY and negatively correlated with ghrelin, but not GLP-1 or CCK (Table 4) [23]. Furthermore, expressing AUC responses per unit of energy clearly shows that energy content was a strong determinant of the metabolic effects of the protein, lactose and fat that it contains. The sum of the individual macronutrient effects expressed per unit of energy resulted in large overestimates of their combined effect in the milks on postprandial glucose and hormone responses (Table 3). The effect of protein per kcal on the AUCs was higher than fat per kcal for insulin, C-peptide, insulin secretion rate, GLP-1, CCK and paracetamol (Pb.0001), but similar to lactose except for CCK and paracetamol, which were lower. In contrast, the response in PYY and ghrelin was similar per unit of energy for each macronutrient (Table 3). Although previous studies report that postprandial inhibition of ghrelin is influenced by a meal's macronutrient content [36,37], including carbohydrates [38] and milk proteins [32], the present results support that energy content is the primary signal for ghrelin release, as has been shown [18]. This approach of adjusting for caloric content to compare macronutrient treatments is supported by a recent study comparing whey protein and glucose at equicaloric doses [18]. Pre-meal consumption of whey protein (10 and 20g) compared with glucose (10 and 20g) led to similar postprandial insulin concentrations, but lower plasma glucose, higher GLP-1 and PYY concentrations and reduced gastric-emptying rate [18]. However, there are some limitations in this study. First, the paracetamol absorption test is an indirect measure of gastric-emptying rate of liquids in humans [31]. Although it provides a reasonably accurate and inexpensive estimate [39,40], it does not have the precision of scintigraphy and other noninvasive methods such as the C13-acetate breath test, magnetic resonance imaging and ultrasound, which are technically challenging and require specialized equipment [40]. Nevertheless, while the quantitative rates of stomach emptying may not be as accurate, the primary purpose of the measure was to compare the response to beverages, and differences were found. Second, total ghrelin accounting for both acyl and des-acyl ghrelin was measured in this study. Acylated ghrelin is now recognized as the preferred measure as it relates more clearly to functionality [41]; however, total ghrelin is reflected by acyl ghrelin as previously reported [42]. Finally, because this study assessed the short-term effects of whole milk on glycemia and hormonal responses in healthy young men, its effects in women and on long-term glycemic control are unclear. Nonetheless, the results of our study add to the accumulating evidence that milk has potential in the dietary management T2D and provides encouragement to conduct both shortand longer-term studies in this population. In conclusion, regulation of postprandial glycemia after milk consumption occurs through both insulin and insulin-independent actions due to interactions among its macronutrient components and energy content to achieve lower postprandial glycemia than predicted from the sum of its components. Statement of authors' contributions to manuscript S.P. contributed to conception and design of the study, carried out the study, performed the hormone assays, data analysis and interpretation, and drafted the manuscript. D.E.K. contributed to conception and design

S. Panahi et al. / Journal of Nutritional Biochemistry 25 (2014) 1124–1131

of the study, assisted in data interpretation, and participated in the critical revision of the manuscript. R.K., T.A., B.L.L. and H.D.G. contributed to data interpretation and participated in the critical revision of the manuscript. R.K. assisted in sample and data analysis. G.H.A. contributed to conception and design of the study, data interpretation and writing of the manuscript. All authors read and approved the final manuscript. Acknowledgments We thank Drs. Paul Pencharz and Vladimir Vuksan for their intellectual contribution and constructive guidance during the research. We also thank Chesarahmia Dojo Soeandy and Urshila Sriram (work-study students) and our nurses for their assistance during the study. References [1] Pereira MA, Jacobs Jr DR, Van Horn L, Slattery ML, Kartashov AI, Ludwig DS. Dairy consumption, obesity, and the insulin resistance syndrome in young adults: the CARDIA Study. JAMA 2002;287(16):2081–9. [2] Elwood PC, Pickering JE, Givens DI, Gallacher JE. The consumption of milk and dairy foods and the incidence of vascular disease and diabetes: an overview of the evidence. Lipids 2010;45(10):925–39. [3] Panahi S, El Khoury D, Luhovyy BL, Goff HD, Anderson GH. Caloric beverages consumed freely at meal-time add calories to an ad libitum meal. Appetite 2013;65:75–82. [4] Panahi S, Luhovyy BL, Liu TT, Akhavan T, El Khoury D, Goff HD, et al. Energy and macronutrient content of familiar beverages interact with pre-meal intervals to determine later food intake, appetite and glycemic response in young adults. Appetite 2013;60(1):154–61. [5] Ostman EM, Liljeberg Elmstahl HG, Bjorck IM. Inconsistency between glycemic and insulinemic responses to regular and fermented milk products. Am J Clin Nutr 2001;74(1):96–100. [6] Nilsson M, Stenberg M, Frid AH, Holst JJ, Bjorck IM. Glycemia and insulinemia in healthy subjects after lactose-equivalent meals of milk and other food proteins: the role of plasma amino acids and incretins. Am J Clin Nutr 2004;80(5):1246–53. [7] Akhavan T, Luhovyy BL, Brown PH, Cho CE, Anderson GH. Effect of premeal consumption of whey protein and its hydrolysate on food intake and postmeal glycemia and insulin responses in young adults. Am J Clin Nutr 2010;91(4):966–75. [8] Pal S, Ellis V. The acute effects of four protein meals on insulin, glucose, appetite and energy intake in lean men. Br J Nutr 2010;11:1–8. [9] Nilsson M, Holst JJ, Bjorck IM. Metabolic effects of amino acid mixtures and whey protein in healthy subjects: studies using glucose-equivalent drinks. Am J Clin Nutr 2007;85:996–1004. [10] Luhovyy BL, Akhavan T, Anderson GH. Whey proteins in the regulation of food intake and satiety. J Am Coll Nutr 2007;26(6):704S–12S. [11] Boirie Y, Dangin M, Gachon P, Vasson MP, Maubois JL, Beaufrere B. Slow and fast dietary proteins differently modulate postprandial protein accretion. Proc Natl Acad Sci U S A 1997;94(26):14930–5. [12] Anderson GH, Moore SE. Dietary proteins in the regulation of food intake and body weight in humans. J Nutr 2004;134(4):974S–9S. [13] Ebringer L, Ferencik M, Krajcovic J. Beneficial health effects of milk and fermented dairy products — review. Folia Microbiol (Praha) 2008;53(5):378–94. [14] Maljaars P, Romeyn E, Haddeman E, Peters H, Masclee A. Effect of fat saturation on satiety, hormone release and food intake. Am J Clin Nutr 2009;89:1019–24. [15] Nuttall FQ, Gannon MC. Plasma glucose and insulin response to macronutrients in nondiabetic and NIDDM subjects. Diabetes Care 1991;14(9):824–38. [16] Anderson GH, Tecimer SN, Shah D, Zafar TA. Protein source, quantity, and time of consumption determine the effect of proteins on short-term food intake in young men. J Nutr 2004;134(11):3011–5. [17] Anderson GH, Catherine NL, Woodend DM, Wolever TM. Inverse association between the effect of carbohydrates on blood glucose and subsequent short-term food intake in young men. Am J Clin Nutr 2002;76(5):1023–30. [18] Akhavan T, Luhovyy BL, Panahi S, Kubant R, Brown PH, Anderson GH. Mechanism of action of pre-meal consumption of they protein on glycemic control in young adults. J Nutr Biochem 2014;25(1):36–43.

1131

[19] Heading RC, Nimmo J, Prescott LF, Tothill P. The dependence of paracetamol absorption on the rate of gastric emptying. Br J Pharmacol 1973;47(2): 415–21. [20] Akhavan T, Luhovyy BL, Anderson GH. Effect of drinking compared with eating sugars or whey protein on short-term appetite and food intake. Int J Obes 2011; 35:562–9. [21] Hamedani A, Akhavan T, Samra RA, Anderson GH. Reduced energy intake at breakfast is not compensated for at lunch if a high-insoluble-fiber cereal replaces a low-fiber cereal. Am J Clin Nutr 2009;89(5):1343–9. [22] Wolever TM. Effect of blood sampling schedule and method of calculating the area under the curve on validity and precision of glycaemic index values. Br J Nutr 2004;91(2):295–301. [23] Murphy KG, Dhillo WS, Bloom SR. Gut peptides in the regulation of food intake and energy homeostasis. Endocr Rev 2006;27(7):719–27. [24] Almiron-Roig E, Drewnowski A. Hunger, thirst, and energy intakes following consumption of caloric beverages. Physiol Behav 2003;79(4–5): 767–73. [25] Hovorka R, Soons PA, Ma Y. ISEC: a program to calculate insulin secretion. Comput Methods Programs Biomed 1996;50(3):253–64. [26] Ahlkvist L, Vikman J, Pacini G, Ahren B. Synergism by individual macronutrients explains the marked early GLP-1 and islet hormone responses to mixed meal challenge in mice. Regul Pept 2012;178(1–3):29–35. [27] Karamanlis A, Chaikomin R, Doran S, Bellon M, Bartholomeusz FD, Wishart JM, et al. Effects of protein on glycemic and incretin responses and gastric emptying after oral glucose in healthy subjects. Am J Clin Nutr 2007;86(5): 1364–8. [28] Reidelberger RD, Hernandez J, Fritzsch B, Hulce M. Abdominal vagal mediation of the satiety effects of CCK in rats. Am J Physiol Regul Integr Comp Physiol 2004;286 (6):R1005–12. [29] Drucker DJ. The biology of incretin hormones. Cell Metab 2006;3(3):153–65. [30] Vrang N, Madsen AN, Tang-Christensen M, Hansen G, Larsen PJ. PYY(3–36) reduces food intake and body weight and improves insulin sensitivity in rodent models of diet-induced obesity. Am J Physiol Regul Integr Comp Physiol 2006;291 (2):R367–75. [31] Hall WL, Millward DJ, Long SJ, Morgan LM. Casein and whey exert different effects on plasma amino acid profiles, gastrointestinal hormone secretion and appetite. Br J Nutr 2003;89(2):239–48. [32] Aziz A, Anderson GH. The effect of dairy components on food intake and satiety: mechanisms of actions and implications for the development of functional foods. In: M S, editor. Functional Dairy Products. Cambridge, UK: Woodhead Publishing Limited; 2007. [33] Brubaker PL, Anini Y. Direct and indirect mechanisms regulating secretion of glucagon-like peptide-1 and glucagon-like peptide-2. Can J Physiol Pharmacol 2003;81(11):1005–12. [34] Michalski MC, Januel C. Does homogenization affect the human health properties of cow's milk? Trends Food Sci Technol 2006;17:423–37. [35] Lorenzen JK, Jensen SK, Astrup A. Milk minerals modify the effect of fat intake on serum lipid profile: results from an animal and a human short-term study. Br J Nutr 2014;111(8):1412–20. [36] Sanchez J, Oliver P, Palou A, Pico C. The inhibition of gastric ghrelin production by food intake in rats is dependent on the type of macronutrient. Endocrinology 2004;145(11):5049–55. [37] Koliaki C, Kokkinos A, Tentolouris N, Katsilambros N. The effect of ingested macronutrients on postprandial ghrelin response: a critical review of existing literature data. Int J Pept 2010;2010. [38] Monteleone P, Bencivenga R, Longobardi N, Serritella C, Maj M. Differential responses of circulating ghrelin to high-fat or high-carbohydrate meal in healthy women. J Clin Endocrinol Metab 2003;88:5510–4. [39] Naslund E, Bogefors J, Gryback P, Jacobsson H, Hellstrom PM. Gastric emptying: comparison of scintigraphic, polyethylene glycol dilution, and paracetamol tracer assessment techniques. Scand J Gastroenterol 2000;35(4):375–9. [40] Willems M, Quartero AO, Numans ME. How useful is paracetamol absorption as a marker of gastric emptying? A systematic literature study. Dig Dis Sci 2001;46 (10):2256–62. [41] Asakawa A, Inui A, Fujimiya M, Sakamaki R, Shinfuku N, Ueta Y, et al. Stomach regulates energy balance via acylated ghrelin and desacyl ghrelin. Gut 2005;54 (1):18–24. [42] Foster-Schubert KE, Overduin J, Prudom CE, Liu J, Callahan HS, Gaylinn BD, et al. Acyl and total ghrelin are suppressed strongly by ingested proteins, weakly by lipids, and biphasically by carbohydrates. J Clin Endocrinol Metab 2008;93(5): 1971–9.