The obese population is growing worldwide as energy intake exceeds energy expenditure and consequently more individuals are at risk of developing obesity-related diseases such as type 2 diabetes mellitus and CVD(Reference Seidell1). High-protein diets have been shown to induce weight loss and weight maintenance after a weight-loss period, through an appetite-suppressing effect and increased diet-induced thermogenesis (DIT), as proteins are not energy efficient(Reference Lejeune, Westerterp and Adam2–Reference Weigle, Breen and Matthys4). High-protein diets may be helpful in preventing overweight or decreasing body weight in obese individuals, as they are effective for the regulation of food intake. Absorption of different proteins can vary, as proven by Boirie et al. (Reference Boirie, Dangin and Gachon5), who showed that whey protein passes quicker through the stomach than casein. Whey protein reaches the jejunum intact in contrast to casein that coagulates under the acidic environment of the stomach. Whey protein releases peptides in the intestine after peptic hydrolysis. This process occurs slowly, which ensures that the digestion and absorption take place over a larger length of the small intestine while descending(Reference Boirie, Dangin and Gachon5–Reference Hall, Millward and Long7). As a result, whey protein has been shown to reduce short-term energy intake and affects appetite suppression through an increase in plasma amino acid levels, mainly branched-chain amino acids such as leucine(Reference Luhovyy, Akhavan and Anderson6, Reference Anderson, Tecimer and Shah8). Other mechanisms such as the utilisation of amino acids for gluconeogenesis and the thermic effects of proteins may contribute as well to the differences seen in appetite suppression(Reference Veldhorst, Smeets and Soenen9).
Whey protein itself is a constituent of milk protein (20 %), and consists of 20–25 % of α-lactalbumin. In a study by Bourthegourd et al., rats that received a meal with α-lac increased their energy expenditure more than rats that were fasting or rats that received a meal containing either glucose or whole milk protein(Reference Bouthegourd, Roseau and Makarios-Lahham10). This increase in energy expenditure could in its entirety be attributed to the increase in protein oxidation, as lipid oxidation was equal with the other groups that ate a meal. Glucose oxidation was increased in rats that consumed a meal enriched with glucose, compared with the other groups. The oxidation of glucose had a protein-sparing effect, through which protein oxidation remained unchanged. The elevated protein oxidation after α-lac-enriched whey protein was the result of a rapid rise in plasma amino acid levels(Reference Bouthegourd, Roseau and Makarios-Lahham10), related to the so-called ‘slow’ and ‘fast’ proteins. It remains unknown whether the increased energy expenditure after α-lac-enriched whey protein could be attributed to the total whey protein or to the enrichment with α-lac.
Therefore, the aim of the present study was to examine whether the addition of total whey protein (whey) or caseinomacropeptide-depleted α-lac-enriched whey protein (α-lac) to a normal milk protein yoghurt drink in order to achieve a specific high-protein yoghurt drink, which was served as breakfast, had an effect on energy expenditure and appetite suppression in human subjects, and whether α-lac and whey differed in these effects.
Subjects and methods
Subjects
Thirty-five healthy subjects (eighteen females, seventeen males) with a mean age of 20·9 (sd 1·9) years and with a mean BMI of 23·0 (sd 2·1) kg/m2 participated in the present study. The subjects were recruited by advertisements in local newspapers and on notice boards at the university. All volunteers (n 94) participated in an initial screening that involved measurements of body weight and height and included the completion of a questionnaire related to eating behaviour (the Three Factor Eating Questionnaire(Reference Stunkard and Messick11)) and the completion of a questionnaire related to health, use of medication, physical activity, alcohol consumption, food allergies, smoking behaviour and daily caffeine consumption. All subjects were in good health, non-smokers, not using medication (except for contraception), at most moderate alcohol consumers and unrestrained eaters (as assessed by factor 1 of the Three Factor Eating Questionnaire). Baseline characteristics of the subjects are presented in Table 1.
(Mean values and standard deviations)
FFM, fat-free mass; FM, fat mass; TFEQ, Three Factor Eating Questionnaire; F1, cognitive restraint; F2, disinhibition; F3, hunger.
* Data collected before the last test day.
† The TFEQ measures three different factors of human eating behaviour.
Experimental design
The study had a randomised, three-arm, single-blind, cross-over design. Subjects attended the university laboratory once per week, during three consecutive weeks. They travelled by public transport or car, in order to avoid physical activity that would have increased resting energy expenditure (REE). Subjects arrived in the fasted state at 08.15 hours and were kept in time-blinded surroundings. They emptied their bladder before the test. After resting on a bed for 30 min, the REE and the substrate oxidation of the subjects were measured for 30 min by means of an open-circuit, ventilated-hood system. Gas analysis was performed by a paramagnetic oxygen analyser (OmniCal type 1155B; Servomex, Crowborough, East Sussex, UK) and an infrared carbon dioxide analyser (Omnical type 1520/1507; Servomex). Energy expenditure was calculated using Weir's formula(Reference Weir12). Respiratory quotient was calculated as CO2 produced/O2 consumed. The subjects were lying in the supine position. After measuring the REE, subjects received one of the three breakfasts, consisting of either the control which was a normal-protein (NP) yoghurt drink with whole milk protein (Friesland Foods, Leeuwarden; 15, 47 and 38 % energy from protein, carbohydrate and fat, respectively), a high-protein (HP) yoghurt with added total whey protein (Hiprotal® 60 MP; Friesland Foods Domo®, Zwolle, The Netherlands; 41, 47 and 12 % energy from protein, carbohydrate and fat, respectively) or a HP yoghurt that contained caseinomacropeptide-depleted α-lactalbumin-enriched whey protein (Vivinal®Alpha; Friesland Foods Domo®; 41, 47 and 12 % energy from protein, carbohydrate and fat, respectively). The 41 % energy as protein in the HP breakfasts consisted of 14 % energy from casein protein and 27 % energy from total whey protein or caseinomacropeptide-depleted α-lac-enriched whey protein. It was unknown how much extra α-lactalbumin was included in the α-lac yoghurt compared with the whey yoghurt or the NP yoghurt. Breakfasts were isoenergetic, had the same flavour (vanilla) and contained the same volume for each condition; also the HP breakfasts had the same energy density (3·2 kJ/g). Before the study started, ratings for palatability and acceptability were assessed with visual analogue scales. With an average rating of 64 mm, subjects experienced no significant differences between conditions concerning colour, taste and viscosity. The energy content of the breakfast meal was tailored to the energy requirements of each subject by calculating BMR with the Harris–Benedict equation, which uses sex, age, height and weight(Reference Harris and Benedict13). To estimate the total energy requirement, the BMR was multiplied with a physical activity index of 1·6, estimated by means of a computer simulation program(Reference Westerterp, Donkers and Fredrix14). The breakfast contained 15 % of the daily energy need. Energy content varied from 1·42 to 2·28 MJ, with an average of 1·74 MJ. During the consumption of the breakfast the hood was removed temporarily.
After the subjects had finished their breakfast the hood was placed back and the measurements continued for another 4 h, during which DIT was determined. During the measurements questionnaires concerning appetite sensations were completed every 30 min as well as immediately before and after breakfast. Subjects were not allowed to talk, laugh, move or sleep while lying under the hood(Reference Schoffelen, Westerterp and Saris15). The urine collected during and after the experiment was used to determine protein turnover. Body composition was measured previous to their last test day.
Anthropometry
Body composition was measured using the 2H2O dilution technique. The dilution of the 2H isotope is a measure for total body water(Reference Westerterp, Wouters and van Marken Lichtenbelt16). In the evening before the last test day, the subjects ingested a dose of 2H-enriched water (2H2O; Cambridge Isotope Laboratories, Inc., Andover, MA, USA) after collecting a background urine sample. After ingestion of the 2H solution no further fluid or food consumption was permitted. The following morning, another urine sample (second voiding) was collected. 2H concentration in the urine samples was measured using an isotope ratio mass spectrometer (Micromass Optima, Manchester, UK). Total body water was obtained by dividing the measured 2H dilution space by 1·04 to correct for exchange of the 2H label with non-aqueous hydrogen of body solids(Reference Schoeller, van Santen and Peterson17). Fat-free mass was calculated by dividing the total body water by a hydrating factor (0·73). By subtracting fat-free mass from body weight, fat mass was obtained. Fat mass expressed as a percentage of body weight gives percentage of body fat.
Questionnaires
The appetite questionnaire was composed of visual analogue scale (in mm) questions on subjective feelings of hunger, fullness, appetite suppression, thirst and desire to eat. Opposing extremes of each feeling were described at either end of a 100 mm horizontal line, and subjects marked the line to indicate how they felt at that moment.
Statistical analysis
Data are presented as mean values and standard deviations or standard errors. Data were analysed using SPSS 11 (SPSS, Inc., Chicago, IL, USA). A three-factor repeated-measures ANOVA including a Scheffé F post hoc test was used to determine a possible difference between breakfast effects on DIT and appetite ratings. Factor DIT, which expresses DIT after correction for body size, was calculated as follows: (REE+DIT)/REE. The level for establishing significant differences was taken at P < 0·05.
The present study was conducted according to the guidelines laid down in the Declaration of Helsinki and all procedures involving human subjects were approved by the Medical Ethical Committee of Maastricht University Medical Centre. Written informed consent was obtained from all subjects.
Results
No adverse events occurred, as no subjects reported any feelings of discomfort after consuming the yoghurt drinks. No different effects for men or women were observed, so these data were analysed together. Average energy intake was equal in all three conditions (1·74 (sd 0·25) MJ). There were significant treatment effects (P < 0·05) for energy expenditure after consumption of the yoghurt drinks. REE did not differ between the three conditions before consuming the whey, α-lac or NP yoghurt drinks (5·08 (sd 0·89), 4·99 (sd 0·83) and 5·02 (sd 0·80) kJ/min, respectively) (Fig. 1). On average, across the 4 h post-breakfast period, total energy expenditure increased significantly compared with baseline measurements in all three conditions (5·99 (sd 0·93), 5·97 (sd 1·02) and 5·77 (sd 0·96) kJ/min; P < 0·001). Areas under the curve for energy expenditure increased significantly after the HP yoghurt drink with whey and after the HP yoghurt drink with α-lac compared with after the NP yoghurt drink (217·05 (se 10·00) kJ × 4 h (P < 0·05); 234·30 (se 11·61) kJ × 4 h (P < 0·01); 179·71 (se 10·87) kJ × 4 h (P < 0·05), respectively) (Fig. 2). No sex differences were seen after comparing the increase of DIT in percentages between all conditions, between males and females. This was also supported after correction for body size by means of calculating factor DIT, where the increase in energy expenditure remained significantly different between whey and placebo (P < 0·05) as well as α-lac and the NP yoghurt drink (P < 0·001). Total energy expenditure was positively related to fat-free mass (whey, r 2 0·81, y = 0·08x+1·54; α-lac, r 2 0·76, y = 0·08x+1·24; NP, r 2 0·78, y = 0·08x+1·26; P < 0·001), as was REE (whey, r 2 0·75, y = 0·07x+0·96; α-lac, r 2 0·78, y = 0·07x+1·12; NP, r 2 0·78, y = 0·07x+1·24; P < 0·001), respectively, for all conditions.
Energy expenditure (EE; kJ/min) during the entire measurement for the placebo (–♦–), total whey protein (–△–) and caseinomacropeptide-depleted α-lactalbumin-enriched whey protein (– × –) conditions in thirty-five subjects (eighteen women and seventeen men). Resting EE was measured during the first 30 min, followed by a breakfast and another 4 h of measuring. Values are means. Data were analysed by repeated-measures ANOVA.
Areas under the curve (AUC) for energy expenditure after the placebo, total whey protein (Whey) and caseinomacropeptide-depleted α-lactalbumin-enriched whey protein (α-Lac) conditions in thirty-five subjects (eighteen women and seventeen men). Values are means, with standard errors represented by vertical bars. Mean value was significantly different from that for the placebo condition: * P < 0·05, ** P < 0·01 (repeated-measures ANOVA).
Subjects were in positive energy balance during the 4 h post-treatment, and energy balance was significantly different between the HP conditions and the NP condition (whey, 0·67 (sd 0·14) MJ/4 h; α-lac, 0·68 (sd 0·16) MJ/4 h; NP, 1·25 (sd 0·21) MJ/4 h; P < 0·001). No significant differences in respiratory quotient were seen between the three conditions at baseline (whey, 0·87 (sd 0·08); α-lac, 0·86 (sd 0·06); NP, 0·87 (sd 0·06); NS) and post-treatment (whey, 0·88 (sd 0·07); α-lac, 0·87 (sd 0·05); NP, 0·86 (sd 0·04); NS). Figure 3 shows the separate macronutrient balances. The protein and fat balances were significantly different between the NP and the HP conditions. Fat balance was lower after the HP breakfasts compared with the NP breakfast (α-lac, 0·11 (sd 0·26) MJ/4 h; whey, 0·15 (sd 0·29) MJ/4 h v. placebo, 1·03 (sd 0·29) MJ/4 h; P < 0·001) and protein balance was higher in the HP conditions compared with the NP condition (α-lac, 0·35 (sd 0·18) MJ/4 h; whey, 0·37 (sd 0·20) MJ/4 h v. placebo, − 0·04 (sd 0·16) MJ/4 h; P < 0·001).
Macronutrient balances for the placebo (■), total whey protein (□) and caseinomacropeptide-depleted α-lactalbumin-enriched whey protein (
) conditions in thirty-five subjects (eighteen women and seventeen men). Values are means, with standard deviations represented by vertical bars. *** Mean value was significantly different from that for the placebo condition (P < 0·001; repeated-measures ANOVA).
Baseline values of the measured variables from the appetite questionnaires (hunger, fullness, satiety and desire to eat) were not significantly different between the conditions. The results and the areas under/above the curve of the visual analogue scales are presented in Fig. 4 (desire to eat) and Table 2. The area above the curve was significantly higher after the α-lac condition compared with the whey condition, for hunger and desire to eat (P < 0·01). No differences were seen between the placebo and α-lac conditions or between the placebo and whey conditions concerning areas above the curve.
Changes in desire to eat (mm visual analogue scale; VAS) after the placebo (–○–), total whey protein (–▲–) and caseinomacropeptide-depleted α-lactalbumin-enriched whey protein (α-lac) (– × –) conditions in thirty-five subjects (eighteen women and seventeen men). Values are means, with standard errors represented by vertical bars. * Mean value for the α-lac condition was significantly lower than that for the total whey protein condition (P < 0·05; repeated-measures ANOVA). † Mean value for the α-lac condition was significantly lower than that for the placebo condition (P < 0·05; repeated-measures ANOVA).
(Mean values with their standard errors)
NP, normal-protein breakfast; Whey, total whey protein yoghurt; HP, high-protein breakfast; α-Lac, caseinomacropetide-depleted α-lactalbumin-enriched whey yoghurt.
Mean value was significantly lower than that after the Whey (HP) breakfast: * P < 0·05, ** P < 0·01 (repeated-measures ANOVA).
Discussion
Consumption of a high-protein breakfast with either added whey or α-lac is more thermogenic than consumption of a normal-protein breakfast containing milk protein. The results of the present study confirmed that high-protein breakfasts enhance thermogenesis more compared with a normal-protein breakfast(Reference Veldhorst, Nieuwenhuizen and Hochstenbach-Waelen18). The NP breakfast contained milk protein, which consists of 80 % casein and 20 % whey. Milk protein has a higher energy efficiency due to casein, which has the lowest cost for ATP production (8·2 kJ/ATP) compared with α-lac and whey (8·6 and 8·4 kJ/ATP)(Reference van Milgen19). Much research has already been conducted concerning the effect of high-protein diets on DIT in comparison with normal-protein diets, in the long-term as well as in acute studies(Reference Lejeune, Westerterp and Adam2, Reference Westerterp-Plantenga3, Reference Luscombe, Clifton and Noakes20–Reference Smeets, Soenen and Luscombe-Marsh25). Moreover, effects of high-protein diets on weight loss and weight maintenance have been shown, through the thermogenic effect and preservation of fat-free mass, following a relatively high-protein diet, which is the main determinant of REE(Reference Westerterp-Plantenga3). Together, this guarantees sustained thermogenesis on a high-protein diet. The positive protein balance after the HP conditions indicate that anabolism was taking place, which may support the preservation of fat free mass(Reference Soenen and Westerterp-Plantenga26). However, more amino acids are available after the HP breakfasts compared with the NP breakfast, which will cause an increase in protein oxidation in the postprandial and post-absorptive state. As a consequence, the daily protein balance will probably be less positive. The low fat content of the HP conditions compared with the NP condition resulted in a fat balance that was reduced, which means that less fat was stored after the HP conditions. The present study was performed under acute conditions; therefore subjects were not adapted to the HP breakfast. This might also explain some of the effects observed with the HP meals compared with the NP meal and particularly the reduction in fat balance.
An explanation for the thermogenic effect of α-lac may be that it is a complete protein. The effect of proteins on DIT is dependent on the quantity of protein, the rate of absorption and thereby increasing the level of amino acids in the bloodstream and also the concentration of essential amino acids, which determines whether a protein is ‘complete’ or ‘incomplete’. α-Lac is a well-balanced protein, which contains high concentrations of indispensable amino acids such as cystine, lysine and tryptophan(Reference Heine, Klein and Reeds27). Not all amino acids are used for protein synthesis and those that remain are oxidised, causing an increase in energy expenditure. Costs for ATP production are higher after α-lac compared with whey and milk protein (8·6 v. 8·4 v. 8·3 kJ/ATP), which means that α-lac is not as energy efficient as whey and milk protein, i.e. the higher the cost for ATP production, the more thermogenic the protein(Reference van Milgen19). There are hardly any studies that have investigated the thermogenic properties of whey and α-lac. The study mentioned before from Bouthegourd et al. showed that whey enriched with α-lac caused an increase in DIT in rats, but in the present study whole milk protein had a comparable effect size(Reference Bouthegourd, Roseau and Makarios-Lahham10).
Appetite scores showed that the HP breakfast with α-lac suppressed hunger significantly more than the HP breakfast with whey and the NP breakfast. No suppressive effect on hunger of the HP breakfast with whey compared with the NP breakfast was seen. The suppressive effect of hunger by α-lac may be attributed to the fast rise in amino acid levels in plasma, as Mellinkoff stated in his aminostatic theory in 1956 that elevated concentrations of plasma amino acids create a appetite-suppressing signal(Reference Mellinkoff, Frankland and Boyle28). Furthermore, α-lac contains high concentrations of leucine and lysine, which are ketogenic amino acids(Reference Heine, Klein and Reeds27). It is known that hunger is suppressed during a ketogenic state(Reference Johnstone, Horgan and Murison29). However, compared with total whey protein there is not much difference between concentrations of ketogenic amino acids(Reference Heine, Klein and Reeds27). Nieuwenhuizen et al. found the same hunger-suppressive effect in a study where they compared α-lac with gelatin, and with gelatin with added tryptophan(Reference Nieuwenhuizen, Hochstenbach-Waelen and Veldhorst30). They concluded that tryptophan, which is present in high concentrations in α-lac and often seen as a modulator of appetite regulation via serotonin, was not involved in the suppressive effect. Neither involved were the ratio between tryptophan and large neutral amino acids or total amino acids, glucagon-like peptide-1 or ghrelin responses(Reference Nieuwenhuizen, Hochstenbach-Waelen and Veldhorst30). In a study by Veldhorst et al. the lowering effect of α-lac on energy intake was also demonstrated. Energy intake at lunch, after a breakfast with α-lac, was 19 % lower than after a breakfast with whey(Reference Veldhorst, Nieuwenhuizen and Hochstenbach-Waelen18).
Furthermore in a rodent study, Pichon et al. showed that energy intake was reduced after a high-protein diet with whey compared with a normal-protein diet with milk protein(Reference Pichon, Potier and Tome31). Diepvens et al. found that after a breakfast with whey protein subjects were more satiated than after a breakfast with milk protein(Reference Diepvens, Haberer and Westerterp-Plantenga32). Bowen et al. (Reference Bowen, Noakes and Trenerry33) and Hall et al. (Reference Hall, Millward and Long7) have studied the effects between casein and whey, the constituents of milk protein. Both studies found different results, as only Hall et al. (Reference Hall, Millward and Long7) could demonstrate that energy intake was lower after a whey preload compared with a casein preload(Reference Hall, Millward and Long7, Reference Bowen, Noakes and Trenerry33).
Summarising, HP breakfasts increased thermogenesis more than a NP breakfast, which can be attributed to the amount of protein but also to the higher bioactivity of α-lac and whey compared with milk protein. Furthermore, α-lac suppressed hunger more than whey and placebo. In conclusion, the addition of α-lac, i.e. whey, to a NP milk protein yoghurt drink increases diet-induced thermogenesis; additional α-lac also suppresses hunger and the desire to eat.
Acknowledgements
We would like to acknowledge Friesland Foods Domo for sponsoring the study and providing the yoghurt drinks. R. H., L. Z. and M. S. W.-P. designed the study. R. H. collected and analysed the data. R. H. wrote the manuscript and M. S. W.-P. contributed to the interpretation of the data and reviewed the manuscript. The study was executed under M. S. W.-P.'s supervision. R. H. and M. S. W.-P. had no personal or financial conflict of interest. L. Z. is nutritional research and development manager at Friesland Foods, which funded the study.
