| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
ORIGINAL RESEARCH COMMUNICATION |
1 From the Department of Human Nutrition, Faculty of Life Sciences, University of Copenhagen, Frederiksberg, Denmark (JKL, SN, and AA); the Department of Medical Physiology, Panum Institute, University of Copenhagen, Copenhagen, Denmark (JJH); the Department of Nutrition, Danish Institute for Food and Veterinary Research, Søborg, Denmark (IT); and the Department of Clinical Biochemistry, Rigshospitalet, University Hospital of Copenhagen, Copenhagen, Denmark (JFR)
2 Supported by the Danish Dairy Research Foundation, FOOD Graduate School, Arla Foods Ingredients amba, Pharma Vinci A/S, The Dutch Dairy Association, Programme Commission on Food and Health (Danish Research Agency), and the Danish Medical Research Council. The cacao butter was a gift from Aarhus United Denmark. 3 Reprints not available. Address correspondence to A Astrup, Department of Human Nutrition, LMC, The Royal Veterinary and Agricultural University Rolighedsvej 30, DK-1958 Frederiksberg C, Denmark. E-mail: ast{at}life.ku.dk.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Objective: Our aim was to examine whether a high calcium intake from dairy products or from supplements affects postprandial fat metabolism and appetite through fat malabsorption.
Design: Four different isocaloric meals were tested in 18 subjects according to a randomized crossover design. The test meals contained high (HC meal: 172 mg/MJ), medium (MC meal: 84 mg/MJ), or low (LC meal: 15 mg/MJ) amounts of calcium from dairy products or a high amount of calcium given as a calcium carbonate supplement (Suppl meal: 183 mg/MJ). Concentrations of plasma total triacylglycerol, chylomicron triacylglycerol, serum total cholesterol, HDL cholesterol, cholecystokinin, glucagon-like peptide 1, ghrelin, peptide YY, glucose, and insulin and appetite sensation were measured before and at regular intervals until 420 min postprandially.
Results: Dairy calcium significantly diminished the postprandial lipid response. The baseline adjusted area under the curve for chylomicron triacylglycerol was 
17% lower after the MC meal (P = 0.02) and 19% lower after the HC meal (P = 0.007) than after the LC meal and 15% lower after the MC meal (P = 0.0495) and 17% lower after the HC meal (P = 0.02) than after the Suppl meal. No consistent effects of calcium on appetite sensation, or on energy intake at the subsequent meal, or on the postprandial responses of cholecystokinin, glucagon-like peptide 1, ghrelin, peptide YY, insulin, or glucose were observed.
Conclusions: Increased calcium intakes from dairy products attenuate postprandial lipidemia, most probably because of reduced fat absorption, whereas supplementary calcium carbonate does not exert such an effect. This may be due to differences in the chemical form of calcium or to cofactors in dairy products. Calcium did not affect appetite sensation, glucose metabolism, or gut hormone secretion.
Key Words: Dietary calcium • calcium supplementation • postprandial lipidemia • appetite • men • cholecystokinin • glucagon-like peptide 1 • ghrelin • peptide YY
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
3% of the variation in body weight. However, other observational studies have failed to find this association (7-9). Relatively few intervention studies have been reported, and those that have been reported are ambiguous with respect to the effect of calcium intakes on body weight and body composition (1, 10-16). In addition, in both animal models and in human studies it has been found that calcium from dairy products has a more profound effect on body weight and composition than does supplementary calcium (6, 14). Many different mechanisms have been suggested to be responsible for the positive effect of a high calcium intake on energy balance. Zemel et al (6) suggested that calcium affects adipocyte metabolism and fat oxidation. Animal studies have found that dairy calcium increases fecal fat excretion (17). In a short-tem intervention study, we recently showed that increases in daily calcium intake increase fecal fat and energy excretion, whereas we failed to find any effect on energy expenditure (18). An increase in calcium intake of 1300 mg/d, mainly from low-fat dairy products, caused a daily increase in fecal fat excretion of 8.2 g in subjects consuming a diet with 30% of energy derived from fat. Others have shown a similar effect, though quantitatively smaller (19-21). The difference in effects might have been due to the source of calcium or to the protein source (18). The subjects in our study were given calcium from dairy products, whereas the subjects in the other studies were given calcium supplements or fortified products. The increased fat excretion was presumably due to formation of insoluble calcium fatty acid soaps or to the binding of bile acids, which impairs the formation of micelles (22-25). When fat is absorbed, it enters blood circulation in the form of intestinally derived triacylglycerol-rich lipoproteins, ie, chylomicrons. Thus, if calcium partly inhibits fat absorption, a decrease in the postprandial increase in chylomicron triacylglycerol would be expected. As far as we know, the effect of calcium on postprandial fat absorption has not been examined previously. Dairy calcium may produce weight loss by a mechanism other than fat malabsorption. Although we, and others, have failed to find any effect of calcium on 24-h energy expenditure during energy balance (18, 26), the effects of calcium on appetite regulation have been less well studied. It is also possible that the increased fecal energy loss produced by dairy calcium might be recognized by the body and result in a compensatory increase in appetite and calorie intake. The aim of the present study was to examine whether a high calcium intake from dairy products or from supplements affects postprandial fat metabolism and appetite. |
SUBJECTS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
US$500) on completion of all tests.
Experimental design
Four different isocaloric meals were tested in 18 subjects according to a randomized crossover design The subjects were randomly assigned to the sequence of the test meals. There was a washout period of 
3 wk between the test meals. Each test meal lasted 8.5 h. The subjects were instructed to not eat or drink dairy products or food items rich in fat, such as potato crisps or chocolate, for 2 d immediately before each test meal or to drink alcohol or perform hard physical activities on the day before each test meal. The subjects consumed a standardized meal supplied by the study, which consisted of a pasta dish, orange juice, and biscuits (4 MJ; 17.6% of energy from protein, 22.1% of energy from fat, and 60.3% of energy from carbohydrate) on the evening before the test. The subjects fasted after consuming this meal at no later than 2000. They were supplied with 0.5 L water and were instructed to consume half of it before they went to bed and the remaining half in the morning before leaving the house. On the morning of the tests, the subjects arrived at the Department of Human Nutrition by car, bus, train, or slowly walking (ie, the least strenuous means of transportation). On arrival they were weighed while wearing only underwear and after emptying their bladder. Weight was measured in kilograms with one decimal by a Lindetronic 8000 scale (Copenhagen, Denmark). A Venflon catheter (BD Infusion Therapy AB, Helsingborg, Sweden) was then inserted in an antecubital arm vein. After 20 min of rest, a fasting blood sample was taken and the subject's appetite sensation was assessed with a visual analogue scale (VAS). The test meal was then served. The subjects were instructed to consume the meal within 15 min, after which they completed a questionnaire on the palatability of the meal. Blood samples were collected 15, 30, 60, 90, 120, 180, 300, and 420 min after the start of the meal, and appetite sensation was assessed every 30 min throughout the 7 h after the meal. During the day the subjects were allowed to read, walk around at the Department of Human Nutrition, listen to the radio, or watch television or videos. One to 3 subjects were tested on each test day. The subjects were allowed to talk with each other as long as the conversation did not involve food, appetite, or related issues. The subjects were not allowed to eat or drink during the day except for 2 glasses of water (250 mL each), which were served 180 and 300 min after the meal (after the blood samples were drawn). At the end of day, after the last blood sample was taken, an ad libitum meal consisting of a pasta salad (15% of energy from protein, 30% of energy from fat, and 55% of energy from carbohydrate) and water was served, food intake was registered, and energy intake was calculated. The subjects were instructed to eat until "comfortable satisfaction." Afterward, they completed a questionnaire on the palatability of the meal, and appetite sensation was assessed. The subjects' water intake was registered on the first test day, and this intake was repeated on the following test days.
Test meals
Four isocaloric test meals were provided to the subjects in a randomized order on 4 different days (Table 1
). The test meals contained dairy proteins as the main protein source and a high amount of calcium from dairy products (HC meal), a medium amount of calcium from dairy products (MC meal), a low amount of calcium from dairy products (LC meal), or a low amount of calcium from dairy products and a calcium supplement served as a drink containing calcium carbonate (Suppl meal).
|
). All of the test meals consisted of the same servings of bread rolls, yogurt, and liquid. The content of macronutrients and micronutrients in each meal was estimated by using Dankost 3000 dietary assessment software (Danish Catering Center, Herlev, Denmark) (27) and information from Arla Foods Ingredients amba (Viby, Denmark). The calcium and energy contents were analyzed. The test meals had the same macronutrient composition: 15% of energy from protein, 39% of energy from fat, and 46% of energy from carbohydrate. In addition, all test meals contained the same amounts of lactose, the content of which was adjusted by adding lactose powder to the bread roll or the drink, and the same amounts of whey and casein, which were achieved with the addition of whey powder (Lacprodan DI-9213; Arla Foods Ingredients amba) and casein powder (Miprodan 30; Arla Foods Ingredients amba).
The test meals were prepared at the Department of Human Nutrition from normal Danish food items. Portions were matched to each subject's individual energy requirement and adjusted to the nearest 1 MJ. Each test meal contained 50% of the subject's daily energy requirement. The subjects' energy requirements were determined by using the following formula, which is based on a meta-analysis of our 24-h calorimetry studies (B Buemann, unpublished observations, 2003):
 | (1) |
The subjects' habitual diet was assessed with the use of a 3-d weighed food record (2 weekdays and 1 Saturday). The subjects were given both verbal and written instructions on how to complete the food record, and digital kitchen scales were supplied. Dankost 3000 dietary assessment software (Danish Catering Center) (27) was used to calculate the energy and nutrient composition of the diets.
Visual analogue scale
Visual analogue scales, 100 mm in length with words expressing the most positive and the most negative rating anchored at each end, were used to measure subjective appetite sensation (hunger, satiation, fullness, prospective food consumption, thirst, well-being, and desire to eat meat or fish or something sweet, salty, or fatty) and palatability of the test and ad libitum meals (Table 2). The questions were provided in small booklets, one question at a time. The subjects were instructed to not compare their ratings with each other and could not refer to their previous ratings when filling in the visual analogue scale.
|
Triacylglycerol
Triacylglycerol was measured in blood samples drawn at baseline and 60, 120, 300, and 420 min after the start of the test meal. The chylomicrons were isolated within 24 h by ultracentrifugation as described elsewhere (29). Plasma total triacylglycerol, chylomicron triacylglycerol, and triacylglycerol concentrations in the bottom fraction were measured with an enzymatic endpoint method (Test-Combination Triacylglycerol GPO-PAP kit; Roche, Basel, Switzerland) by using a Cobras Mira Plus (Roche). The intraassay and interassay CVs were 0.5% and 1.4%, respectively. The percentage of recovery was calculated, and the chylomicron triacylglycerol concentration was adjusted to the concentration corresponding to 100% recovery.
Plasma concentrations of glucose, cholecystokinin, ghrelin, glucagon-like peptide 1, and peptide YY
Plasma concentrations of glucose, cholecystokinin (CCK), ghrelin, glucagon-like peptide 1 (GLP-1), and peptide YY (PYY) were measured in blood samples drawn at baseline and 15, 30, 60, 90 (only glucose), 120, 180, 300 (except glucose), and 420 min after the start of the test meal. Glucose was measured with the use of an enzymatic endpoint method (Hexokinase) (Gluco-quant Glucose/HK, Roche Diagnostics, Basel, Switzerland) with a Cobras Mira Plus (Roche). The intraassay and interassay CVs were 1.1% and 1.3%. CCK concentrations in plasma were measured by radioimmunoassay with an antiserum (code no. 92128) that binds the circulating CCK with equimolar potency without cross-reactivity with any gastrin. Details of the assay and its reliability parameters are described elsewhere (30).
Plasma ghrelin concentrations were determined by using a radioimmunoassay kit (catalog no. GHRT-89k; Linco Research, St Charles, MO) that measures total ghrelin (intact and desoctanoylated ghrelin). The sensitivity was 100 pg/mL, and the intraassay CV was <10%. Quality controls were always within acceptable limits.
Radioimmunoassay of PYY in plasma was performed by using antiserum code no. 8412-2II (31), which reacts equally with PYY1–36 and PYY3–36. Synthetic human PYY 1–36 or PYY 3–36 (Peninsula Laboratories, St Helens, United Kingdom) were used for standards. [125I]-PYY1–36 (code no. IM259) was from Amersham Biosciences, Buckinghamshire, United Kingdom. The assay buffer consisted of 0.05 mol sodium phosphate/L (pH 7.5), 400 KIE Trasylol-aprotinin/mL, 0.1 mol NaCl/L, 10 mmol EDTA/L, and 0.6 mmol merthiolate/L. Unknown plasma samples (150 µL) + 150 µL assay buffer or 150 µL charcoal-treated plasma + 150 µL standards were preincubated with antiserum (100 µL) and diluted 1:20.000 (final concentration) for 48 h at 4 °C. Then, 100 µL tracer (5 fmol; specific activity: 70 MBq/nmol) was added, and the mixture was incubated for 24 h before bound and free peptide moieties were separated by plasma-coated charcoal (32). The detection limit of the assay was <2.5 pmol/L, and 50% inhibition was obtained with 23 pmol PYY/L. Recovery of PYY added to plasma in concentrations between 5 and 50 pmol/L deviated <15% from expected values. The intraassay CV was <5%. The antiserum showed no cross-reaction with human neuropeptide Y or human pancreatic peptide in concentrations up to 500 pmol/L.
GLP-1 concentrations in plasma were measured by radioimmunoassay after extraction of plasma with 70% ethanol (by vol, final concentration). Carboxy-terminal GLP-1 immunoreactivity was determined by using antiserum code no. 89390, which has an absolute requirement for the intact amidated carboxy-terminus of GLP-1 7–36 amide and cross-reacts <0.01% with carboxy-terminally truncated fragments, and 89% with GLP-1 9–36 amide, the primary metabolite of dipeptidyl-peptidase IV–mediated degradation (33). The sum of the 2 components (total GLP-1 concentration) reflects the rate of secretion of the L cell. Sensitivity was <5 pmol/L, and the intraassay CV was <10%.
Serum concentrations of insulin, total cholesterol, and HDL cholesterol
Serum concentrations of insulin, total cholesterol, and HDL cholesterol were measured in blood samples drawn at baseline and 30, 60, 90, 120, 180, 300 (except insulin), and 420 min after the start of the test meal. Insulin was measured by solid-phase, 2-site chemiluminescent immunometric assay (Immulite/immuliter 1000 insulin; Diagnostic Products Corporation, Los Angeles, CA) with the use of an Immulite 1000 analyzer (Diagnostic Products Corporation). The intraassay and interassay CVs were 2.5% and 4.9%, respectively. Total cholesterol was measured with an enzymatic endpoint method (CHOD-PAP) (cholesterol kit; Roche) by using a Cobras Mira plus (Roche, Basel, Switzerland). The intraassay CV was 0.9%, and the interassay CV was 1.5%. HDL cholesterol was measured with an enzymatic colorimetric test (HDL cholesterol plus second generation; Roche) by using a Cobras Mira Plus (Roche). The intraassay and interassay CVs were 1.9% and 3.5%, respectively.
Calcium and energy contents of the test meals
A portion of each meal was freeze-dried and homogenized. Gross energy content was obtained by using a bomb calorimeter (Ika-calorimeter system C4000; Ika, Heitersheim, Germany). The analyses were performed as duplicates. Before the calcium content was measured, the samples were lyophilized and microwave digested (MES-1000; CEM Corporation, Matthews, NC) with 65% HNO3 (suprapur; Merck, Darmstadt, Germany) and 30% H2O2 (suprapur; Merck). Calcium was measured by atomic absorption spectroscopy (SpectraAA-200 VARIAN; Varian Techtron Pty, Limited, Victoria, Australia) after dilution with a lanthanium oxide solution (Merck). Standards were prepared from a 1000-mg Ca/L standard (Tritisol; Merck) by dilution with lanthanium oxide solution. A reference diet (Standard Reference Material 1548a, Typical Diet; National Institute of Standards and Technology, Gaithersburg, MD) was analyzed in the same run. The percentage of recovery of calcium in the reference diet was 89%. The calcium content in the samples was adjusted to concentrations corresponding to 100% recovery.
Calculations and statistical methods
Data were analyzed with the use of SAS software (version 8; SAS Institute, Cary, NC). Data are reported as means ± 95% confidence limits (CLs) unless otherwise indicated. A P value <0.05 is considered significant.
Repeated-measures analysis of variance (ANOVA) was used to assess the effect of time, meal, and interaction of meal and time. To adjust for differences in baseline values, these factors were included as covariate in all analyses. The analysis was performed in PROC MIXED. Based on the structure of the covariance matrix, the Gaussian model of spatial correlation was chosen for covariance structure. The study had a crossover design, so "subject" was included as a random effect. We tested for a significant effect of sequence of the test meals, and sequence of the test meals was included as a covariate when this was the case.
Area under the curve (AUC) was calculated as the total increase above zero. Analysis of covariance (ANCOVA) was used to examine the effect of meal on AUC. ANCOVA was performed in PROC MIXED. "Subject" was included as a random effect and baseline values as a covariate in all analyses. The sequence of the test meals was included as a covariate when it was found to have a significant effect. ANOVA was used to examine the effect of meal on the palatability of the meals and ad libitum intake. ANOVA was performed in PROC MIXED, with subject included as a random effect. Model control for all analyses was performed in PROC MIXED, and data were transformed before analysis if necessary. Least-squares means were used to estimate the adjusted means. Pairwise comparisons of the least-squares means were performed by paired t test.
The recording of appetite sensation measured by VAS after intake of the ad libitum meal was not included in the calculation of AUC or in any of the analyses.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Blood variables from one subject were excluded from the data analyses because of abnormal triacylglycerol values. The record of appetite sensation after the HC meal was missing for one subject.
|
. Plasma total and chylomicron triacylglycerol concentrations increased after intake of each of the 4 test meals. The concentrations peaked 180 min after intake of the MC and the HC meals and 300 min after intake of the LC meal and the Suppl meal. At the end of the study, the concentrations had decreased to values similar to the baseline concentration after intake of the LC, MC, and HC meals but not after intake of the Suppl meal. Repeated-measures ANOVA showed a significant meal-by-time interaction on the plasma total triacylglycerol concentration (P = 0.04) when adjusted for baseline concentration. However, pairwise comparisons of the adjusted means showed no persistent effects. There was a significant effect of time on chylomicron triacylglycerol concentrations (P < 0.0001) when adjusted for baseline concentration. There was no significant meal-by-time interaction and no significant effect of meal.
|
. ANCOVA showed that meal had a significant effect on AUC for chylomicron triacylglycerol (P = 0.01) when adjusted for baseline concentration. The adjusted AUC for chylomicron triacylglycerol was lower after intake of the MC (17%; P = 0.02) and HC (19%; P = 0.007) meals than after the LC meals. Similarly, the AUC was lower after intake of the MC (15%; P = 0.0495) and the HC (17%; P = 0.02) meals than after the Suppl meal. There were no significant differences between the other meals. Meal had no significant effect on AUC for plasma total triacylglycerol adjusted for baseline concentration.
Serum cholesterol
Repeated-measures ANOVA of serum total cholesterol and of HDL cholesterol, both adjusted for baseline concentrations, showed no significant meal-by-time interaction and no significant effect of meal (data not shown). However, a significant effect of time on both serum total cholesterol (P = 0.01) and HDL cholesterol (P < 0.0001) was observed.
ANCOVA showed that meal had no significant effect on the AUC for total cholesterol when adjusted for the baseline concentration of total cholesterol or on the AUC for HDL cholesterol when adjusted for the baseline concentration of HDL cholesterol and the sequence of the test meals (data not shown).
Plasma concentrations of glucose, cholecystokinin, ghrelin, glucagon-like peptide 1, and peptide YY
The postprandial responses of glucose, CCK, ghrelin, GLP-1, PYY, and insulin are shown in Figure 2. Repeated-measures ANOVA showed a significant meal-by-time interaction on GLP-1 (P = 0.007) when adjusted for baseline concentration. Comparisons of adjusted least-squares means showed that the GLP-1 concentration was higher 60 min after intake of the Suppl meal than after the LC (P = 0.03), MC (P = 0.01), and HC (P = 0.005) meals. No significant differences were observed at the other time points, and no significant meal-by-time interaction or significant effect of meal was observed on CCK, ghrelin, PYY, glucose, or insulin when adjusted for baseline concentrations. However, a strong significant effect of time on CCK, ghrelin, PYY, and insulin (P < 0.0001 for all) was observed when adjusted for baseline concentration and on glucose (P < 0.0001) when adjusted for baseline concentration and the sequence of the test meals.
|
. ANCOVA showed that meal had no significant effect on the AUC for CCK, GLP-1, ghrelin, PYY, or insulin when adjusted for baseline concentrations or on the AUC for glucose when adjusted for the baseline concentration and the sequence of the test meals.
Palatability of the test meal
ANOVA showed that meal had a significant effect on the palatability of the test meal with regard to taste (P = 0.02), visual appeal (P < 0.0001), and overall palatability (P < 0.0001) (data not shown). The subjects found that the taste of the HC meal was significantly better than the taste of the LC (P = 0.0008) and the supplement (P = 0.007) meals when adjusted for the sequence of the test meals. The subjects found that the HC meal was significantly more visually appealing than were the LC (P < 0.0001) and Suppl (P = 0.02) meals, and the LC meal was significantly less appealing than were the MC (P = 0.0003) and Suppl (P = 0.02) meals. The subjects found that the overall palatability of the HC meal was significantly better than that of the LC (P < 0.0001) and the Suppl (P = 0.0002) meals and that of the MC meal was significantly better than that of the LC (P = 0.0003) and the Suppl (P = 0.01) meals.
Appetite sensation
The postprandial responses in subjective appetite sensation measured by VAS are shown in Figure 3 (data on well-being and desire to eat meat or fish or something salty or sweet are not shown). Repeated-measures ANOVA showed no significant meal-by-time interaction and no significant effect of meal on hunger, satiation, fullness, prospective food consumption, thirst, well-being, or desire to eat meat or fish or something sweet, salty, or fatty after adjustment for sensation measured at baseline. However, a strong significant effect of time on all measures, except well-being, was observed (P < 0.0001 for all). ANCOVA showed that meal had no significant effect on AUC for any of the measures after adjustment for sensation measured at baseline (Figure 3).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Postprandial fat metabolism
Consumption of the high-fat meals resulted, as expected, in pronounced postprandial lipidemia. Compared with the LC meal, the lipid response in chylomicron triacylglycerol was reduced by 17% by the MC meal and by 19% by the HC meal, which indicated that a high calcium intake from dairy products reduces the lipid response but also that a plateau value exists above which an increased calcium intake does not seem to have any additional effect. The total lipid response in plasma total triacylglycerol was not affected by calcium intake.
The decrease in the lipid response in chylomicron triacylglycerols may reflect either a decrease in fat absorption or an increase in chylomicron clearance. However, no evidence in the literature suggests that calcium intake interferes with chylomicron clearance, and the unaffected concentrations of total cholesterol and HDL cholesterol suggest that this is unlikely. It is therefore most likely that the decreased postprandial lipid response was due to decreases in fat absorption. It was previously shown that the lipid response in chylomicron triacylglycerols increases with the amount of fat ingested (0-50 g per meal) (34). As far as we know, the present study is the first to show that an increased calcium intake from dairy products exerts a lowering effect on postprandial fat absorption. However, several studies, in both animals and humans, have shown that calcium intake increases the fecal excretion of fat, presumably via the formation of insoluble calcium fatty acid soaps in the gut or by binding of bile acids, which impairs the formation of micelles (17-21, 24, 25, 35, 36).
Our main aim in studying postprandial fat metabolism was to examine whether the previously reported increased fecal fat excretion induced by a diet high in dairy calcium could be reflected in a diminished postprandial plasma excursion of triacylglycerol and chylomicron triacylglycerol, but we are aware that the findings might also have implications for the role of a diet high in dairy products in the risk of cardiovascular diseases (CVD). Observational studies have previously shown an inverse association between dairy and calcium intakes and CVD and dyslipidemia (2, 37-39). Orlistat, a gastrointestinal lipase inhibitor that reduces dietary fat absorption by 30%, has been shown to have a positive effect on fasting lipid profile by decreasing fasting total cholesterol and LDL (40, 41) and reducing postprandial lipids (42). Although the effect of calcium in the present study is not as pronounced as the effect of Orlistat, it is likely that a long-term high intake of calcium from dairy products may have beneficial effects on the lipid profile and, thereby, on the risk of CVD. A beneficial effect on the lipid profile was previously shown in intervention studies using calcium from supplements or fortified foods (43, 44).
It was previously shown that calcium from dairy products has a more profound effect on body weight than does calcium from supplements (6, 14). The mechanism behind this difference is unknown. We previously suggested that this may be due in part to the time at which the supplement is consumed. If calcium is to inhibit the absorption of fat it, is a condition that fat and calcium are present in the gut at the same time (5). We therefore gave the supplement as a part of the meal and in a solution that mimicked the milk given with the other meals. However, the lipid response in chylomicron triacylglycerols was not significantly different after the intake of the LC and Suppl meals, which indicated that there must be another property of milk calcium that we did not take into account, eg, the chemical form of calcium, other bioactive components in dairy products, differences in the solubility of calcium from milk and calcium carbonate, or differences in pH. In dairy products, calcium is largely found as calcium phosphate and it is possible that phosphate contributes to the effect of calcium. It was previously shown that supplemental calcium increases the fecal excretion of phosphate and inhibits its absorption, probably because of the formation of insoluble calcium phosphate, which has been suggested to bind bile acids and thereby partly impairs the formation of micelles (25).
Appetite sensation and regulation
One of our considerations was that a partial inhibition of fat absorption by increased calcium intake could have a compensatory stimulatory effect on food intake, which might lead to increased food intake. However, we found no effect of a high calcium intake, from either dairy products or the supplement, on the sensation of appetite, on the secretion of the appetite-regulating hormones, or on energy intake at the subsequent ad libitum meal. A significant interaction was found between meal and time for GLP-1 secretion, but this was due to an inexplicably high concentration of GLP-1 60 min after intake of the supplement meal and it was not a persistent difference between the meals. In addition, there was no significant difference in the total GLP-1 response; therefore, we do not attach much importance to the significant interaction between meal and time. Our results indicate that the apparent inhibition of fat absorption does not have a stimulating effect on energy intake and that the previously observed effect of calcium on body weight and composition is not due to a reduced energy intake. However, we cannot exclude a subchronic effect of calcium on appetite.
We used a VAS to measure appetite sensation. The reproducibility and validity of the VAS was previously examined and found to be reliable (45). Flint et al found that, to detect a difference of 10% in the mean value, a study with a paired design should include 18 subjects (20 subjects for the variable "desire to eat something salty") to obtain a power of 0.9 (45). In the present study we used the AUC and not mean values in the data analyses. However, because the study included 18 subjects, it seems unlikely that the negative outcome found in the present study was due to type 2 errors. If possible, a gap should be entered because the study limitations apply to the entire study and not just the part dealing with appetite sensation and regulation.
The study had some limitations. Although all efforts were exerted to make the 4 meals as similar as possible, except for the calcium intakes, there were some differences in the compositions of the meals, and we could not rule out that these differences may have confounded the outcome. Moreover, we cannot exclude the possibility that an adaptation to a long-term high calcium intake may occur, which could reduce the effect. However, because our subjects had a relatively high habitual calcium intake, it seems unlikely that a total adaptation takes place. Further studies are necessary to examine this aspect further.
Consistent with our previous finding that a diet high in dairy calcium induces fat malabsorption in humans, the present findings indicate that the effect is also reflected in an attenuated postprandial lipid response after intake of dairy calcium from milk and low-fat yogurt, but not after intake of supplementary calcium. This may have been due to differences in the chemical form or to cofactors in the dairy products. The findings of the present study do not suggest any effect on appetite regulation as assessed by appetite sensations, appetite hormones, or ad libitum intake of a subsequent meal.
|
ACKNOWLEDGMENTS |
|---|
JKL, IT, and AA designed the study. JKL and SN collected the data. JJH and JFR analyzed most of the hormones. JKL analyzed the data. All authors participated in the discussion of the results and commented on the manuscript. AA is a member of the Arla Nutrition Advisory Board and receives an honorarium for attending each board meeting.
|
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]
|
|
|
|
 |
|
 J. A. Yanovski, S. J. Parikh, L. B. Yanoff, B. I. Denkinger, K. A. Calis, J. C. Reynolds, N. G. Sebring, and T. McHugh Effects of Calcium Supplementation on Body Weight and Adiposity in Overweight and Obese Adults: A Randomized Trial Ann Intern Med, June 16, 2009; 150(12): 821 - 829. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Tremblay and J.-A. Gilbert Milk Products, Insulin Resistance Syndrome and Type 2 Diabetes J. Am. Coll. Nutr., February 1, 2009; 28(Supplement_1): 91S - 102S. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Van Loan The Role of Dairy Foods and Dietary Calcium in Weight Management J. Am. Coll. Nutr., February 1, 2009; 28(Supplement_1): 120S - 129S. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Astrup The role of calcium in energy balance and obesity: the search for mechanisms Am. J. Clinical Nutrition, October 1, 2008; 88(4): 873 - 874. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
), high (HC; 
), and medium (MC; 
) amounts of calcium from dairy products or high amounts of calcium as supplementary calcium carbonate (Suppl; 
). n = 17. The mean AUCs are adjusted for the baseline concentration, and the bars represent 95% confidence limits. Repeated-measures ANOVA showed a significant interaction of meal and time for plasma total triacylglycerol concentration (P = 0.04) and a significant effect of time on chylomicron triacylglycerol concentration (P < 0.0001) when adjusted for the baseline concentration. ANCOVA showed a significant effect of meal on the AUC for chylomicron triacylglycerol (P = 0.01) when adjusted for the baseline concentration. Values with different lowercase letters are significantly different, P < 0.05. 1n = 16 because of missing values due to analytical problems.
