Objective: To examine the effect of overfeeding isoenergetic diets enriched in 1-monoglyceride or triglyceride on nutrient oxidation and appetite throughout the ...
European Journal of Clinical Nutrition (1998) 52, 610±618 ß 1998 Stockton Press. All rights reserved 0954±3007/98 $12.00 http://www.stockton-press.co.uk/ejcn
Overfeeding fat as monoglyceride or triglyceride: effect on appetite, nutrient balance and the subsequent day's energy intake AM Johnstone1, LM Ryan1, CA Reid2 and RJ Stubbs1 1
The Rowett Research Institute, Greenburn Road, Bucksburn, Aberdeen AB21 9SB, Scotland, UK; and 2Biomathematics and Statistics Scotland, The Rowett Research Institute, Greenburn Road, Bucksburn, Aberdeen AB21 9SB, Scotland, UK
Objective: To examine the effect of overfeeding isoenergetic diets enriched in 1-monoglyceride or triglyceride on nutrient oxidation and appetite throughout the day that it was given and the subsequent day's food and energy intake. Design: Six men [mean (s.d.) weight; 76.89 (7.00) kg, height; 1.77 (0.05) m, age; 26.4 (6.0) y], were each studied twice in a 3 d protocol. On day 1 (maintenance day) they were fed a medium fat (MF) maintenance diet (MF: 40% fat, 47% carbohydrate and 13% protein by energy) calculated at 1.66resting metabolic rate (RMR). Subjects entered the calorimeter at 06.30 on day 2 for 49.5 h. On day 2 (manipulation day) subjects consumed a MF diet at 1.66RMR with an additional 0.456RMR as either 1-monoglyceride or triglyceride. On day 3 (outcome day), subjects had ad libitum access to isoenergetic, isoenergetically dense MF (40 : 47 : 13, 550 kJ/ 100 g) foods. Subjective hunger and satiety were tracked hourly, during waking hours throughout days 1±3. Results: There was no signi®cant effect of diet on nutrient oxidation or balance either during day 2 (manipulation day) or day 3 (outcome day), fat oxidation was similar on both diets. Subjective hunger was not affected by diet on either day with mean values of 34.3 and 35.0 mm (SED 5.2) on manipulation day (day 2) and outcome day (day 3), 35.3 and 40.8 mm (SED 5.2) on the 1-monoglyceride or triglyceride diets respectively. Day 3 food and energy intake were unaffected by the previous day's dietary treatment, with mean intakes of 15.9 and 15.6 MJ (SED 1.07) on the 1-monoglyceride or triglyceride treatments, respectively. Conclusions: This study suggests that when 1-monoglyceride is covertly incorporated into a diet at unusually high levels, it behaves in a manner that is very similar to triglyceride, in its effects on appetite, feeding behaviour and net nutrient balance. Sponsorship: This work was supported by the Scottish Of®ce Agriculture, Environment and Fisheries Department. Descriptors: monoglyceride; triglyceride; fat; energy; macronutrients; food intake; appetite; humans; calorimetry
Introduction The role of high-fat, energy-dense diets in bringing about excess energy intakes (Lissner et al, 1987; Thomas et al, 1992) and their contribution to the current secular trends in overweight and obesity is now well documented (Department of Health, 1992). Increased intake of dietary fat is believed to facilitate overconsumption because it increases the palatability and energy density of the diet (Blundell et al, 1995), both of which appear to facilitate higher energy intakes. Once ingested, dietary fat appears to exert poor postabsorptive feedback onto subsequent intake (Stubbs, 1995). It has been suggested that nutrient oxidation is an important component of satiety, and since dietary fat is preferentially stored, rather than oxidised, it may well contribute less to postabsorptive satiety than the macronutrients protein and carbohydrate. The intake of these macronutrients stimulate autoregulatory increases in their own oxidation (Abbot et
Correspondence: Dr RJ Stubbs. Received 16 October 1997; revised 20 April 1998; accepted 4 May 1998
al, 1988). It has been suggested that the different metabolic fate of these macronutrients may partially underlie the differences that they exert on satiety (Stubbs, 1995). Thus fat may exert relatively weak postabsorptive feedback to suppress subsequent feeding because it is preferentially stored rather than oxidised. The majority of studies that have examined the effects of dietary fat on appetite and energy intake in humans do not discriminate between the types of dietary fat used, and these studies tend to use mixed fats in their dietary preparations. Fats can vary in structure in terms of (i) chain length, (ii) degree of saturation and (iii) degree of esteri®cation of the glycerol backbone. Storlien (1990) has pointed out that not all fats may be as readily stored and poorly oxidised as has generally been presumed, and that this may have implications for the role of different fats in the development of obesity. He points to a number of indications in the literature which suggest that the metabolic fate of a fat may partially depend on its degree of saturation (Mead et al, 1956; Leyton et al, 1987). In this context it is intriguing that Blundell's group have recently shown that when 20 subjects were each fed isoenergetic lunches rich in monounsaturates, polyunsaturates and saturates, monounsaturates signi®cantly suppressed short-term
Overfeeding fat as monoglyceride or triglyceride AM Johnstone et al
intake to the greatest extent, and saturates were the least satiating (Lawton et al, 1997). There is evidence that the chain length of fat may in¯uence energy intake. In general the shorter the chain length of a fat the more immediate its oxidation. In this context it is of note that there is evidence that fat metabolites (including ketones) may suppress food intake in rodents (Carpenter & Grossman, 1982; Rich et al, 1988). There is also data in rats (Furuse et al, 1992) and humans (Stubbs & Harbron, 1996) which shows isoenergetic substitution of medium-chain triglycerides (MCT) for long-chain triglycerides (LCT) at high levels, in highfat, energy-dense diets, limits the excess energy intakes and weight gains that are frequently apparent when subjects feed ad libitum on such diets. Incorporation of more moderate amounts of MCT into the diet is unlikely however, to be of use in promoting spontaneous weight loss. There is little data in the literature on whether the degree of esteri®cation of dietary fat affects feeding behaviour and energy intake. Gregory et al (1989) have infused different fats into the gastrointestinal tract of pigs. They found 1monoglyceride to be especially effective at suppressing subsequent intake (in excess of its energy content) when infused into the duodenum. Gastric infusions suppressed subsequent intake by the amount of energy infused. Gregory & Rayner (1987) have also experimentally examined the relationship between duodenal infusion of fat, CCK secretion and feeding in the pig and concluded that `monoglyceride-induced CCK secretion is mainly responsible for the satiety to duodenal fat in the pig, but that there is also a CCK-independent effect via the fatty acid'. These relatively fragmentary ®ndings in relation to the potential effects of fat structure on appetite and energy intake suggest that fats which are more readily absorbed and/or metabolised may suppress energy intake to a greater extent than mixed long chain triglyceride. Few if any studies in humans have examined whether monoglyceride differentially affects appetite and energy intake relative to triglyceride. A recent short-term study suggested that 1monoglyceride behaved in a similar manner to triglyceride as regards short-term appetite and energy intake (Ryan et al, 1997). The purpose of the present study was to extend this initial study, by giving a larger dose of monoglyceride or triglyceride, over the course of one day, under controlled conditions of ®xed mandatory intakes, so that subjective motivation to eat could be tracked hourly. By housing subjects in a whole-body indirect calorimeter, it was possible to determine over this time period whether overfeeding 1-monoglyceride, relative to triglyceride, differentially affected fat oxidation (which is believed to be linked to the satiating ef®ciency of fat). The effect of this manipulation on the subsequent day's energy intake was then assessed. Materials and methods Subjects and their characteristics Six healthy, lean (BMI 20±25) non-smoking men were recruited by advertisement. Their mean (s.d.) weight was 76.8 (7.0) kg, height was 1.77 (0.05) m and age was 26.4 (6.0) y. They were each studied twice in a 3 d experiment, with at least one week in between each dietary treatment. Subjects were resident in the Human Nutrition
Unit for the duration of each study period. Height, weight and resting metabolic rate were measured as described by Johnstone et al (1998). Design Subjects were each studied for two 3 d periods, 48 h of which were spent in a whole-body indirect calorimeter. On day 1, subjects were fed a medium fat (MF: 40% fat, 47% carbohydrate and 13% protein by energy) ®xed, maintenance diet estimated at 1.66RMR, served as three isoenergetic, isoenergetically dense meals. On day 2 (manipulation day) subjects entered the calorimeter at 06.30. In the calorimeter, subjects received a diet calculated at 1.66RMR, with an additional 0.456RMR as either monoglyceride (Dimodan PV, Danisco ingredients (UK) Ltd., Suffolk) or triglyceride (largely as Soya Dream, Vandemoortele (UK) Ltd., Hounslow, UK). Total energy intake on day 2 therefore amounted to 2.056RMR. The position of the fatty acid will affect the absorption of the fat, as it is more ef®ciently absorbed in the 2 position than in the 1 or 3 position (Bistrian, 1997). Dimodan is a 1 position monoester, made from edible vegetable oil, which has a high polyunsaturate content. For this reason the corresponding LCT was derived largely from polyunsaturated sources. The mean food, energy and nutrient intakes during day 2 (manipulation day) are given in Table 2, the recipes are given in Appendix 1. Meal times were as follows: 08.30 (breakfast) 13.00 (lunch) and 19.00 (dinner). On day 3 (outcome day), subjects had ad libitum access to a MF diet, which enabled them to alter the amount but not the composition of foods ingested. The menu is given in Appendix 2. The composition of these foods is given in Appendix 3. During day 3 in the calorimeter, subjects requested food by telephone, with the food then being heated up in the microwave by a member of staff. Access to food was as follows: items on the menu called `breakfast' were served on request from 08.30±11.30, lunch from 11.30±17.30, and supper from 17.30±22.30. Snacks were available at any time between 08.30±22.30. From 22.30± 8.30 no food was given. In the calorimeter, volunteers had two cycling periods of 40 min at 50 W to increase energy expenditure to 1.56RMR, which is fairly typical for a free-living sedentary routine. Cycle times were 11.10±11.50 h and 16.00± 16.40 h. The volunteers were requested not to undertake any other strenuous physical activity during the 3 d study. Visual analogue scales were completed on every waking hour throughout each study day to assess changes in subjective motivation to eat and completed pleasantness and satisfaction ratings 15 min after each meal as described by Hill & Blundell (1982). In the calorimeter urine was collected every 4 h and overnight (8 h). Nitrogen content was analysed using the Foss Heralus Nitrogen Analyser (macro N) (Foss Electric UK Ltd., Wheldrake, York, UK). This study was approved by the Joint Ethical Committee of Grampian Health Board and the University of Aberdeen. Formulation and preparation of the diets The composition of each dish in terms of energy, fat, carbohydrate, protein and non-starch polysaccharide was calculated from McCance & Widdowson's The Composition of Foods, 5th Edition and supplements (Holland et al, 1991). The diets for day 2 were formulated to be of a very similar energy content and density so that differences in
611
Overfeeding fat as monoglyceride or triglyceride AM Johnstone et al
612
amount of food ingested did not in¯uence ad libitum food intake on the subsequent day. The ad libitum diet was formulated so that every item on the menu comprised 47% fat, 40% carbohydrate and 13% protein as a proportion of energy and contained 550 kJ/100 g wet weight of food. This was done so that food intake directly paralleled energy intake. The food was prepared by the dietetic research assistant in the metabolic kitchen. Therefore, the ad libitum diets were formulated to contain, as far as possible, normal every-day ingredients and to contain recognisable food which varied little in composition (for example pre-prepared foods were not used). They had to be palatable and appealing to the volunteers. Diets were pilot tested and altered accordingly to resemble familiar food items, yet have the appropriate composition. Presentation of the diets and measurement of food intake If the subjects were working during day 1 (maintenance day), food was packed in cool bags to be taken with them. On the calorimeter days, food was placed through the chamber hatch. On ad libitum days in the chamber, food and drink was requested from the given menu by telephone. Food was presented to the subjects in the following amounts: breakfast, 600 g; main courses 800 g; sweets 150 g; milkshakes 300 g and hot drinks 350 g. Extra portions were readily available on request. Subjects could therefore determine the time, size (weight) and frequency but not composition of each meal. When a subject requested food, the weight was recorded by the researcher before the subject had eaten. Feeding behaviour in terms of meal size and frequency could thus be determined. Calorimetry The study was conducted in the Rowett's two whole body, indirect calorimeters, which are identical in design and lay out. A previous paper (Johnstone et al, 1996) describes the chambers, their initial calibration, ongoing system checks and statistical simulation of the variation of results for recovery of oxygen and carbon dioxide. The gas analysers were calibrated prior to every run, using atmospheric gas, N2 and a span scaling gas. The span gases were checked by comparison with alpha standard gases, corrected to STP (British Oxygen Company, Guildford, Surrey, UK). During the run, the analysers were corrected for drift every 3 h using the atmosphere as a reference. Precision estimates for the chambers suggest that the standard deviation in the estimates of fat and carbohydrate oxidation are 21.7 kJ/h and 18.9 kJ/h, respectively, giving a coef®cient of variation (CV) about the calculated hourly substrate oxidation of 10.5% and 8.9%, respectively. Over the 24 h these values would be expected to decrease. This calculation excludes all other possible instrumental errors and all errors relating to coef®cients and constants used in the calculation of substrate oxidation. There may be a further 1±2% error in fat oxidation calculation on the monoglyceride diet, because the ratio of fatty acid to glycerol is lower than conventional triglycerides (Livesey & Elia, 1988). The major source of calorimetric errors originate from calibration of the ventilation rate, linearity of the carbon dioxide analysers and the composition of the carbon dioxide span gas. These remained unchanged throughout the course of the study and the same calorimeter was used for both runs for each subject. Thus, while errors in the
calculation of substrate oxidation may not be insigni®cant, they would have been primarily systematic and relatively constant over the course of the study, and would have had little impact on the relative comparison across diets. Oxygen consumption and carbon dioxide production were estimated using rapid-response calculations of Brown et al (1984). Energy expenditure was calculated from Elia & Livesey's equation (Elia and Livesey, 1992). Substrate oxidation rates were calculated from oxygen and carbon dioxide exchanges and urinary nitrogen excretion using the values of Livesey & Elia (1988) for volumes of oxygen consumed per oxidised gram of protein, fat and carbohydrate and the associated respiratory quotients. Statistical analysis The visual analogue ratings (VAR) were analysed using analysis of variance (ANOVA) by calculating a mean rating for each 24 h period with diet and day as a factor and subject as a blocking factor. Under conditions where data were not normally distributed a square root transformation was used for the analysis. Real mean values were used for reporting appetite plus hunger ratings. Additionally, day 2 (manipulation day) was analysed by splitting the day into three inter-meal periods and for each subject on each diet, calculating a mean rating for each period. The three periods were, pre-lunch 09.00±12.00 h inclusive, postlunch 14.00±18.00 h inclusive and post-supper 20.00± 23.00 h inclusive. These values were analysed by ANOVA with diet, day and time as factors and subject as blocking factor. Mean hourly values were also analysed by ANOVA day, diet and time as factors and subject as blocking factor. Subjectively rated pleasantness and satisfaction was analysed by ANOVA with diet and meal as factors and subject as a blocking factor. Twenty four hour intake, oxidation and balance on days 2 and 3 were analysed by ANOVA, with diet and day as factors and subject as blocking factor. In addition, day 2 nutrient oxidation patterns were analysed by splitting the day into three separate periods (09.00±13.00, 14.00±19.00 and 20.00±08.00). Nutrient oxidations were then analysed by ANOVA with diet, day, and period as factors and subject as blocking factor. Daily food and energy intakes were also analysed with the inclusion of mean daily subjectively rated pleasantness of the food as a covariate. All analysis was performed using the Genstat 5 statistical program (Genstat 5 Rothampstead Experimental Station, Harpenden, UK). Results Subjective motivation to eat Subjective hunger was very similar throughout both maintenance days, over the 24 h prior to the consumption to the monoglyceride and triglyceride diets (mean values of 35.8 and 44.2 mm (SED 5.2) respectively. Table 1 shows mean daily hunger and fullness on days 1±3. Figure 1 shows the mean subjective hunger rated by the six men on the two dietary treatments throughout day 2 (manipulation day) of the study. There was no signi®cant diet effect throughout this day. Mean values were, 34.3 and 35.0 mm (SED 5.2) on the monoglyceride and triglyceride diet respectively. Neither was there any signi®cant effect when mean subjective hunger was analysed in the three inter-meal periods. Mean values for pre-lunch (09.00±12.00) were 29.1 and
Overfeeding fat as monoglyceride or triglyceride AM Johnstone et al
Table 1 Days 1±3: average (s.e.m.) subjective feelings of hunger and fullness between 07.00 and 22.00 h in 6 men. There was no day, diet or day/diet interactions for hunger or fullness at the 5% level of probability
Hunger Day 1 Day 2 Day 3 Fullness Day 1 Day 2 Day 3
Mono
Tri
36 8 34 5 35 5
44 11 35 9 41 6
40 6 50 4 47 3
39 9 51 7 41 5
Figure 1 Mean subjectively rated hunger on the monoglyceride and triglyceride diets throughout manipulation day 2 and outcome day 3. ANOVA showed that there were no diet effects or diet-time interactions at the 5% level of signi®cance.
33.3 mm, post-lunch (14.00±18.00) were 38.2 and 36. mm and post-supper (20.00±22.00) 31.0 and 32.8 mm (SED 9.2) for the monoglyceride and triglyceride diets respectively. There was no signi®cant difference in hunger between diets on day 3 (ad libitum outcome day) with values of 35.3 and 40.8 mm (SED 5.2) on the monoglyceride and triglyceride diets respectively. Subjectively rated fullness exhibited a reciprocal pattern to hunger. The average 24 h values on day 2 were 49.7 and 51.0 mm (SED 5.2) for the monoglyceride and triglyceride diets respectively. Subjects felt signi®cantly more full on day 2 than on day 1, which is indicative of their overfed state (F(3, 30) 3.53; P 0.027). The same patterns as for hunger were apparent for subjectively rated `desire to eat',
`urge to eat', `prospective consumption' and `thoughts of food'. Analysis of the individual hourly values showed that there were no diet effects for any parameter. Figure 1 shows that time had a major effect on hunger (F(15, 431) 8.52; P < 0.001). The effect for time was apparent for all other variables and will not be discussed further, as this is a common ®nding in studies such as this. There were no diet-time interactions for any parameter. Day-time interactions were apparent for all responses indicating the difference between ®xed meal times for breakfast and lunch on day 2 as compared to subject-determined meal times for day 3. The fasting values for hunger were similar on the monoglyceride and triglyceride diets on day 2 [7 and 7 mm, respectively diet (F(1, 5) 0.07; P 0.806)]. The same effect was apparent for day 3. These patterns prevailed for all appetite responses. Surprisingly, subjects preferred the monoglyceride diet over the triglyceride (F(1, 5) 8.49; P 0.03). On average, subjectively rated pleasantness expressed by the subjects 15 min after the completion of each meal, were 68.2 and 60.6 mm (SED 2.6) for the monoglyceride and triglyceride diet respectively. Ratings for satisfaction, also expressed by the subjects 15 min after the completion of each meal, were not signi®cantly different between diets, with values of 74.2 and 71.7 mm (SED 5.1) for the monoglyceride and triglyceride diets respectively. Nutrient intake, oxidation and balance Table 2 gives the average 24 h energy and nutrient intake, net oxidation and balance of protein, carbohydrate and fat for day 2 (manipulation day) together with F-ratios and probability values for the main effects for the six men on each of the two dietary treatments. Day 2 energy intakes were slightly lower on the monoglyceride dietÐalthough only by a maximum of 2%. Analysis of day 2 intakes with mean daily subjective pleasantness of the food as a covariate demonstrated that the day and diet effect remained unaltered by the covariate. This suggests that differences in the hedonic qualities of the diets did not exert quantitatively signi®cant in¯uences on intake. By the end of day 2, there was no signi®cant difference between treatments in energy expenditure, nutrient oxidation or energy and nutrient balance. Fat oxidation was very similar between diets, with 4.9 and 5.0 MJ/d oxidised on the monoglyceride and triglyceride diets respectively. Similarly, there was no difference between oxidation of carbohydrate or protein between dietary treatments. By the end of day 2, average energy balance was notably positive and very similar on each treatment at 4.2 and 4.1 MJ/d on the monoglyceride and triglyceride diets respectively. This is mostly attributable to an accumulation of fat balance, by the end of day 2, of 2.9 and 2.7 MJ/d on the monoglyceride and triglyceride diets respectively. There was no signi®cant diet effect on energy or nutrient balance (Table 2). There was no treatment effect on ad libitum energy intake throughout day 3 as a consequence of ingesting the monoglyceride or triglyceride diets. Subjects consumed 15.9 and 15.6 MJ/d on the monoglyceride and triglyceride diets respectively. Table 3 gives the average net 24 h energy and nutrient intake, oxidation and balance of protein, carbohydrate and fat for day 3 (outcome day) together with F-ratios and probability values for the main effects for
613
Overfeeding fat as monoglyceride or triglyceride AM Johnstone et al
614
Table 2 Day 2 (Manipulation day), average (s.e.m.) food, energy and nutrient intake, oxidation and balance for the two dietary treatments, together with the variance ratios and statistical probability for the main treatment effects
Intake Wet weight (g) Energy (MJ) Protein (MJ) Fat (MJ) CHO (MJ) Oxidation Energy (MJ) Protein (MJ) Fat (MJ) CHO (MJ) Balance Energy (MJ) Protein (MJ) Fat (MJ) CHO (MJ)
Mono
Tri
Variance ratio (F1, 5)
Probability
SED
2188 167 15.51 1.10 1.84 0.13 7.76 0.55 5.93 0.42
2592 186 15.78 1.13 1.92 0.14 7.69 0.55 6.16 0.44
11.30 0.85 1.19 0.13 4.88 0.98 5.23 0.59
11.68 0.79 1.46 0.17 4.96 0.97 5.26 0.73
0.34 2.33 0.02 0.00
0.587 0.187 0.905 0.964
646.8 174.2 641.1 592.7
4.21 0.84 0.64 0.07 2.88 0.99 0.70 0.60
4.10 0.56 0.46 0.19 2.73 0.97 0.91 0.63
0.03 1.16 0.06 0.13
0.875 0.331 0.823 0.736
650.1 174.3 656.2 580.2
Table 3 Day 3 (Outcome day): average (s.e.m.) food, energy and nutrient intake, oxidation and balance for the two dietary treatments, together with the variance ratios and statistical probability for the main treatment effects
Intake Wet weight (g) Energy (MJ) Protein (MJ) Fat (MJ) CHO (MJ) Oxidation Energy (MJ) Protein (MJ) Fat (MJ) CHO (MJ) Balance Energy (MJ) Protein (MJ) Fat (MJ) CHO (MJ)
Mono
Tri
Variance ratio (F1, 5)
Probability
SED
3169 305 15.94 1.61 2.14 0.21 6.44 0.68 7.36 0.73
3098 304 15.59 1.59 2.10 0.21 6.22 0.64 7.23 0.74
0.15 0.11 0.12 0.22 0.04
0.713 0.753 0.746 0.656 0.845
183.6 1067.1 123.0 463.1 498.4
11.38 0.68 1.37 0.18 4.31 0.74 5.70 0.28
11.65 0.77 1.50 0.20 3.79 0.83 6.36 0.86
0.08 5.60 0.55 0.88
0.784 0.064 0.493 0.392
907.9 52.4 697.0 699.1
4.56 0.86 0.76 0.07 2.13 0.99 1.65 0.60
3.94 1.51 0.60 0.12 2.43 1.07 0.90 0.56
0.49 1.66 0.34 0.82
0.513 0.253 0.583 0.406
878.8 128.5 504.9 835.1
the six men on each of the two dietary treatments. Because the composition of the diet was constant on this day, the intakes of the macronutrients exhibited a similar trend. Energy expenditure was also similar across treatments during day 3, with 11.4 and 11.7 MJ/d expended on the monoglyceride and triglyceride diets respectively. The actual values can be obtained from Table 2. Since energy and therefore nutrient intakes were similar on the dietary treatments throughout day 3, any differences in substrate oxidation, or speci®cally fat oxidation, could be largely attributed to different types of fat (but not energy) ingested on day 2. There were no signi®cant differences in nutrient oxidation on that day. Interestingly, after being overfed by around about 45% in excess of energy requirements during day 2, in the form of either triglyceride or monoglyceride, subjects still overate on the outcome day leading to a similar positive energy balance by the end of day 3 of 4.6 and 3.9 MJ on the monoglyceride and triglyceride diets respectively. Splitting day 2 nutrient oxidation patterns into the three time periods (09.00±13.00, 14.00±19.00 and 20.00±08.00) showed that there were no diet-time interactions for fat [F(2, 554) 0.59; P 0.553] or CHO [F(2, 554) 2.34; P 0.097] oxidation. There was a diet-time interaction for protein oxidation [F(2, 554) 8.00; P < 0.001]. Speci®-
cally, protein oxidation was higher on the triglyceride treatment, between 09.00±13.00 and 14.00±19.00 than on the monoglyceride diet. Mean values were 0.05, 0.06 and 0.05 MJ/h and 0.07, 0.07 and 0.05 MJ/h on the monoglyceride and triglyceride rich diets, during the morning, afternoon and evening periods respectively. On both diets, carbohydrate oxidation increased [F(2, 554) 69.24; P < 0.001] between morning and evening and fat oxidation decreased [F(2, 554) 39.81; P < 0.001] during the same periods, indicating the overfed state of subjects during day 2. Discussion The effect of the composition of the overfed diets on subjective hunger, fullness and appetite The dose of monoglyceride and triglyceride was, on average 3.45 MJ per day (50% of total fat) or 25% of daily energy intake. This very large dose of monoglyceride did not exert any differential effect on subjective hunger or motivation to eat, at any point during day 2. It is also of interest that despite the large extra load of fat consumed on day 2, there was no change in hunger relative to the maintenance day. Clearly this effect is not due to a lack of sensitivity of the subjective ratings since they have been shown to be sensitive to nutritional and other manipulations
Overfeeding fat as monoglyceride or triglyceride AM Johnstone et al
in a number of laboratories (Hill et al, 1984; DeCastro, 1987; de Graff, 1993). In our laboratory, under experimental conditions such as these, changes in subjective hunger have been found to be sensitive to differences in the energy density of diets (Stubbs et al, 1997) to the macronutrient composition of isoenergetically-dense diets (Johnstone et al, 1996). Furthermore, in the present study, differences due to time-course effects were detectable in the study, with hunger rising and falling in accordance with meal times. Presumably the excess energy ingested as either monoglyceride or triglyceride was able to be assimilated without producing strong physiological changes that are associated with satiety and that would have decreased hunger relative to day 1. In other words, overfeeding fats of differing degree of esteri®cation did not invoke a physiological response that was as readily detected as a treatment or day effect, through sensory experience and expressed as motivation to eat. Subjects did however feel signi®cantly more full on day 2 relative to day 1 (F(3, 30) 3.53; P 0.027). This probably re¯ected the higher levels of food and energy intake on day 2. The dose of monoglyceride administered to these subjects was greatly in excess of the levels ingested in normal life. Since the large loads of monoglyceride used in this and our previous experiment did not differentially in¯uence subjective motivation to eat it is reasonable to conclude that normal levels of 1-monoglyceride ingestion will not exert a differential effect on subjective motivation to eat in freeliving humans. The effect of the composition of the overfed diets on energy intake on day 3 (outcome day) By the end of day 2 subjects has attained a mean imbalance of around 3.0 MJ of the overfed macronutrient, fat, and a positive energy balance of around 4.0 MJ. This had no differential effect on the subjects' energy intake during day 3. In the present study, subject's average subjective hunger and appetite was the same regardless of the diet they consumed on the previous day. Therefore, there was no apparent unconditioned stimulus apparent by the beginning of day 3 upon which to base any behavioural response. Subjects did not compensate for the excess energy intake of day 2 on the following day. Studies in other laboratories have also shown that when nutrient balance is manipulated over one (Stubbs et al, 1993), two (Shetty et al, 1994) or 3 days (Snitker et al, 1997) there is no response in feeding behaviour on the subsequent day. Day-to-day effects in the relationship between nutrient balance/metabolism and energy intake have been found when subjects feed ad libitum on covertly manipulated diets for several consecutive days, where changes in energy and nutrient balance can build up over time (Stubbs et al, 1997). Heavey et al (1995) have found more rapid responses to day-to-day energy de®cits, and it may well be that humans compensate more accurately for day-to-day energy de®cits than surfeits (Johnstone et al, 1996). In the study of Johnstone et al (1996) we argued that the different physiological effects macronutrient loading on one day had decayed by the next day, and therefore failed to elicit a behavioural response. In the present study we argue that the dietary manipulation of fat-type did not induce any detectable difference in nutrient processing, and therefore in¯uenced neither subjective motivation to eat during the day of the manipulation, nor feeding behaviour on the subsequent day.
The effect of energy and nutrient intakes on nutrient oxidation and balance During day 2, subjects were given approximately 0.456RMR as fat in addition to a MF maintenance diet. As can be seen from Table 2 this large load of 1-monoglyceride did not affect nutrient oxidation patterns throughout day 2. This suggests that the 1-monoglyceride was absorbed and physiologically processed in the same way as triglyceride, from the perspective of whole-body net nutrient balance. Despite being in a positive fat balance of around 3.0 MJ by the end of day 2 total fat oxidation was very similar on both days on both treatments. It appears that the monoglyceride acted in a similar way to the triglyceride and was preferentially stored, as oxidation did not increase speci®cally in response to this physiological load. It has been previously demonstrated that fat oxidation does not show a large increase in response to a high-fat overfeeding diet (Flatt et al, 1988). Protein and carbohydrate balance tend to be regulated (by changes in their oxidation) at the expense of fat balance. This appeared to be the case in the present study. Over the 2 days subjects were in the chambers they ingested 31.9 and 31.0 MJ in total, comprising fat, 14.1 and 14.0 MJ; carbohydrate 13.7 and 13 MJ; protein 4.1 and 3.9 MJ, on the triglyceride and monoglyceride treatments, respectively. By the end of day 3 (after 48 h in the calorimeter), on average (across dietary treatments) 62% of the fat was oxidised, 85% of the carbohydrate was oxidised and 70% of the ingested protein was utilised (not accounting for extraneous nitrogen loss). It should be noted that protein balance was somewhat positive due to the marked positive energy balance of 8.4 MJ by the morning of day 4. Thus taking into account the nitrogen that was retained in association with a positive energy balance (Young et al, 1992) these data con®rm those of previous studies which have shown that fat balance is at the bottom of a hierarchy in the immediacy with which macronutrient stores are regulated by increases in oxidative disposal (Abbot et al, 1988). In this context there did not appear to be any differences in the net whole body metabolic fate of 1-monoglyceride compared to triglyceride. Limitations of the present results As in most studies of feeding behaviour, there were limitations to the design of this experiment. It is important to indicate where the design (and hence conclusions arising from it) was subject to the following constraints: (i) The experimental environment of the calorimeter, while offering great precision and accuracy of measurement is highly arti®cial. Equal attention should be given to studies conducted in a more naturalistic environment; (ii) It should be remembered that the subject's response in terms of food intake could only vary quantitatively. Selection of different foods, which vary in composition and/or energy density, was precluded; (iii) Compensatory feeding responses to dietary manipulations are likely to have a large learned component. Studies of the nature and duration as that described in this paper do not allow the connection between unconditioned stimuli and the conditioning process to be reinforced; (iv) The type of fat used in this study (Dimodan-PV Distilled monoglyceride) is mainly a 1-mono-ester. In terms of chemical structure, the fatty acid is mainly found on the 1-position of the glycerol backbone. Positions 1 and 3 are hypothesised to be least readily absorbed, in
615
Overfeeding fat as monoglyceride or triglyceride AM Johnstone et al
616
comparison to position 2 which is more readily absorbed (Bistrian, 1997). It is therefore possible that the 2-monoglyceride may exert differential effects on appetite and feeding behaviour; (v) The present study is of a relatively short-duration. Statements relating to energy balance therefore pertain to short-term energy balance and caution should be exercised when extrapolating to longer term conclusions regarding energy balance; and (vi) This experiment used six lean, young men as the study population. It may be inappropriate to extrapolate these ®ndings to other groups in the population such as women, older subjects or overweight subjects. Conclusions This study has demonstrated that overfeeding large isoenergetic doses of dietary monoglyceride or triglyceride over the course of one day did not exert differential effects on subjective motivation to eat, or nutrient oxidation on the
day it was fed, nor did it affect feeding behaviour or nutrient oxidation pro®les on the subsequent day. This lack of effect is consistent with the hypothesis that dietary fats which tend to be more readily absorbed and utilised tend to be more satiating than long chain triglyceride. The dose of monoglyceride administered to these subjects was greatly in excess of the levels ingested in normal life. Since the large loads of monoglyceride used in this and our previous experiment did not differentially in¯uence nutrient ¯ux, subjective motivation to eat or feeding behaviour it is reasonable to conclude that normal levels of 1-monoglyceride ingestion are unlikely to exert a differential effect on appetite or energy balance in normal free-living subjects. AcknowledgementsÐThis work was supported by funding from the Scottish Of®ce Agriculture, Environment and Fisheries Department. We are grateful to Danisco Ingredients (UK) Ltd., Suffolk, England, for the donation of the 1-monoglyceride.
Appendix 1 Recipes for the meals consumed during day 2 (manipulation day) of the study: mean intake for the six men Monoglyceride diet Food Breakfast
Ice-cream
Lunch
Ice-cream
Supper
Ice-cream
a
Milk Corn¯akesa Orange juice Sugar Dimodan (mono)b Maple & pecan cerealc Dried skimmed milk Protifard Low-calorie ice-creame Milk Lettuce Tomato Green pepper Wholemeal bread Cheddar cheese Dimodan (mono)b Maple & pecan cerealc Dried skimmed milk Protifard Low-calorie ice-creame Beef Carrots Potatoes Gravy Milk Dimodan (mono)b Maple & pecan cerealc Dried skimmed milk Protifard Low-calorie ice-creame
Triglyceride diet Wt (g) 265 30 115 12 32 75 6 6 219 38 35 58 58 81 30 32 75 6 6 219 61 115 115 69 76 32 75 6 6 219
Kellogg's corn¯akes, Kelloggs, Manchester, UK. Dimodan, Danisco Ingredients (UK) Ltd., Suffolk, UK. c Maple & pecan cereal, Tesco Stores, Cheshunt, UK. d Protifar, Cow & Gate Nutricia Ltd., Wiltshire, UK. e Low-calorie ice-cream, Heinz Co. Ltd., Hayes, UK. f Soya dream, Vandemoortele (UK) Ltd., Hounslow, UK. b
Food Milk Corn¯akesa Orange juice Sugar Soya dream (Tri)f Maple & pecan cerealc Dried skimmed milk Protifard Low-calorie ice-creame Milk Lettuce Tomato Green pepper Wholemeal bread Cheddar cheese Soya dream (tri)f Maple & pecan cerealc Dried skimmed milk Protifard Low-calorie ice-creame Beef Carrots Potatoes Gravy Milk Soya dream (Tri)f Maple & pecan cerealc Dried skimmed milk Protifard Low-calorie ice-creame
Wt (g) 265 30 115 12 177 75 6 3 219 38 35 58 58 80 30 177 75 6 3 219 61 115 115 69 76 117 75 6 6 219
Overfeeding fat as monoglyceride or triglyceride AM Johnstone et al
617
Appendix 2 Ad libitum menu for day 3 (Outcome day) Day 3 menu with portion sizes Breakfast Weetabix (600 g) Lunch Chicken stew (26400 g) (with a salad garnish of lettuce 40 g, cress 10 g, celery 20 g and green pepper 30 g) Supper Chicken curry (26400 g) (with a salad garnish of lettuce 40 g, tomato 30 g, cucumber 30 g) Snack choice Milkshake (300 g chocolate, strawberry or banana) Cocoa (350 g) Raspberry fool (26150 g) Sweetcorn soup (400 g) Milk allowance (200 g)
Appendix 3 Energy and nutrient content (MJ/100 g) for the ad libitum diets available throughout day 3a Food Weetabix Chicken stew Chicken curry Milkshake Cocoa Raspberry fool Sweetcorn soup
Weight (g)
Energy (MJ)
Fat (MJ)
Carbohydrate (MJ)
Protein (MJ)
100 100 100 100 100 100 100
0.68 0.55 0.55 0.56 0.55 0.46 0.51
0.34 0.21 0.21 0.22 0.22 0.22 0.22
0.27 0.26 0.26 0.27 0.26 0.17 0.22
0.07 0.08 0.08 0.07 0.07 0.07 0.07
a
Because the subjects themselves determined how much or how little of each item they ate, it is not possible to give a representative day's intake of food on this diet.
References Abbot WGH, Howard BV, Christin L, Freymond D, Lillioja S, Boyce VL, Anderson TE, Bogardus C & Ravussin E (1988): Short-term energy balance: relationship with protein, carbohydrate and fat balances. Am. J. Physiol. 255, E332±E337. Bistrian BR (1997): Novel lipid sources in parenteral and enteral nutrition. Proc. Nutr. Soc. 56, 471±477. Blundell J, Cotton J, Delargy H, Green S, Greenough A, King N & Lawton C (1995): The fat paradox: fat-induced satiety signals versus high fat overconsumption. Int. J. Obes. 19, 832±835. Brown D, Cole TJ, Dauncey MJ, Marrs RW & Murgatroyd PR (1984): Analysis of gaseous exchange in open-circuit indirect calorimetry. Med. Biol. Eng. Comput. 21, 333±338. Carpenter R & Grossman S (1982): Plasma fat metabolites and hunger. Physiol. Behav. 30, 57±63. De Castro JM (1987): Macronutrient relationships with meal patterns and mood in the spontaneous feeding behaviour of humans. Physiol. Behav. 39, 561±569. de Graff C (1993): The validity of appetite ratings. Appetite 21, 156±160. Department of Health (1992): The Health of the Nation: A Strategy for Health in England. London: HMSO. Elia M & Livesey G (1992): Energy expenditure and fuel selection in biological systems. World Rev. Nutr. Diet 70, 68±131. Flatt JP, Ravussin E, Acheson HJ & Jequier E (1988): Effects of dietary fat on postprandial substrate oxidation and on carbohydrate and fat balances. J. Clin. Invest. 7, 1019±1024. Furuse M, Choi Y-H, Mabayo RT & Okumura J-I (1992): Feeding behaviour in rats fed diets containing medium chain triglyceride. Physiol. Behav. 52, 815±817. Gregory PC & Rayner DV (1987): The in¯uence of gastrointestinal infusion of fats on regulation of food intake in pigs. J. Physiol. 385, 471±481. Gregory PC, McFadyen M & Rayner DV (1989): Duodenal infusion of fat, cholecystokinin secretion and satiety in the pig. Physiol. Behav. 45, 1021±1024.
Heavey PM, McKenna APM, Goldberg GR, Murgatroyd PR & Prentice AM (1995): Underfeeding by reduction in fat or carbohydrate intake: effects on energy expenditure, macronutrient oxidation and subsequent food intake in lean men. Proc. Nutr. Soc. 54, 160A (Abstract). Hill AJ & Blundell JE (1982): Nutrients and behaviour: research strategies for the investigation of taste characteristics food preferences hunger sensations and eating patterns in man. J. Psychol. Res. 17, 203±212. Hill AJ, Magson LD & Blundell JE (1984): Hunger and palatability: tracking ratings of subjective experience before, during and after the consumption of preferred and less preferred food. Appetite 5, 361±371. Holland B, Welch AA, Unwin I, Buss DH, Paul AA & Southgate DAT (1991): Fifth Revised Edition of McCance & Widdowson's The Composition of Foods. London: The Royal Society of Chemistry and Ministry of Agriculture Fisheries and Food. Johnstone AM, Stubbs RJ & Harbron CG (1996): Effect of overfeeding macronutrients on day-to-day food intake in man. Eur. J. Clin. Nutr. 50, 418±430. Johnston AM, Ryan LM, Reid CA & Stubbs RJ (1998): Breakfasts high in monoglyceride or triglyceride: no differential effect on appetite or energy intake. Eur. J. Clin. Nutr. 52, 603±609. Lawton C, Delargy H, Smith F & Blundell J (1997): Does the degree of saturation of fatty acids affect post-ingestive satiety? Int. J. Obes. 21 (Suppl 2), S35 (abstract). Leyton J, Drury PJ & Crawford MA (1987): Differential oxidation of saturated and unsaturated fatty acids in vivo in the rat. Br. J. Nutr. 57, 383±393. Lissner L, Levitsky DA, Strupp BJ, Kalkwarf HJ & Roe DA (1987): Dietary fat and the regulation of energy intake in human subjects. Am. J. Clin. Nutr. 46, 886±892. Livesey G & Elia M (1988): Estimations of energy expenditure, net carbohydrate utilization: evaluation of errors with special reference to the detailed composition of fuels. Am. J. Clin. Nutr. 47, 608±628. Mead JF, Staton WH & Decker AB (1956): Metabolism of the essential fatty acids II. The metabolism of stearate, oleate and linoleate by fat de®cient and normal mice. J. Biol. Chem. 218, 401±407.
Overfeeding fat as monoglyceride or triglyceride AM Johnstone et al
618
Rich AJ, Chambers P & Johnston IDA (1988): Are ketones an appetite suppressant?. J. Parenter Enteral Nutr. 13, 7S (Abstract). Ryan LM, Stubbs RJ, Johnstone AM, Lyons HE & Robertson K (1997): Breakfasts high in mono- or triglyceride: no differential effect on within-day appetite and energy intake. Proc. Nutr. Soc. 56, 129A. Shetty PS, Prentice AM, Goldberg GR, Murgatroyd PR, McKenna APM, Stubbs RJ & Volschenk PA (1994): Alterations in fuel selection and voluntary food intake in response to isoenergetic manipulation of glycogen stores in man. Am. J. Clin. Nutr. 60, 534±543. Snitker S, Larson DE, Tataranni PA & Ravussin E (1997): ad libitum food intake in humans after manipulation of glycogen stores. Am. J. Clin. Nutr. 65, 941±946. Storlien L (1990): Not all dietary fats may lead to obesity. Am. J. Clin. Nutr. 51, 1114±1115. Stubbs RJ & Harbron CG (1996): Covert manipulation of the ratio of medium to long-chain triglycerides in isoenergetically dense diets: effect on food intake in ad libitum feeding men. Int. J. Obes. 20, 435±444. Stubbs RJ (1995): Macronutrient effects on appetite. Int. J. Obes. 19, Suppl 5, S11±S19.
Stubbs RJ, Goldberg GR, Murgatroyd PR & Prentice AM (1993): Carbohydrate balance and day-to-day food intake in man. Am. J. Clin. Nutr. 57, 897±903. Stubbs RJ, Harbron CG & Johnstone AM (1997): The effect of covertly manipulating the energy density of high-carbohydrate diets on ad libitum food intake in ``pseudo free living'' humans. Proc. Nutr. Soc. 56, 133A. Thomas CD, Peters JC, Reed GW, Abumrad NN, Sun Ming & Hill JO (1992): Nutrient balance and energy expenditure during ad libitum feeding of high-fat and high-carbohydrate diets in humans. Am. J. Clin. Nutr. 55, 934±942. Young VR, Yu Y-M & Fukagawa NK (1992): Energy and nitrogen (protein) relationships. In Energy Metabolism Tissue Determinants and Cellular Corollaries, eds. HM Kinney and HN Tucker, pp 139± 162. New York: Raven Press.