Added thermogenic and satiety effects of a mixed nutrient vs a sugar-only beverage

Article metrics


OBJECTIVE: To examine the effects of a sugar-only (SO) beverage vs one containing a mixed-nutrient (MN) composition on energy expenditure and feelings of hunger and satiety.

HYPOTHESIS: A beverage containing a mixed macronutrient composition will lead to greater thermic effect of food and feelings of fullness than an isocaloric beverage containing only sugar.

RESEARCH METHODS AND PROCEDURES: Adults were randomly assigned to receive a 2510 kJ (600 kcal) SO liquid formula followed by an isovolumic, isoenergetic, MN liquid formula with an energy distribution of 17% protein, 67% carbohydrates as sucrose and corn syrup solids, and 16% fat, or vice versa, in a crossover design. The carbohydrate source in the two beverages was identical: 1:1 ratio of sucrose and corn syrup solids (25 dextrose equivalents). The thermic response was calculated as the 7 h deviation from resting metabolic rate (RMR). Subjects provided hunger/satiety ratings and other related information by visual analog scales at regular intervals throughout the study period.

RESULTS: In all, 20 subjects completed the protocol; one was removed from the thermic effect analysis due to discrepant RMRs. Following beverage ingestion, SO and MN liquid meals produced 7 h thermic effects of (X±s.e.m.) 274.1±27.6 kJ (65.5±6.6 kcal) and 372.0±33.9 kJ (88.9±8.1 kcal), respectively, resulting in a significant (P<0.01) difference between meals (Δ=97.9±35.1 kJ [23.4±8.4 kcal]). Analysis of satiety ratings using area under the curve analysis showed greater feelings of satiety (P<0.05) with MN compared to SO consumption. Also, subjects felt that they could eat less (P<0.05) after consumption of the MN vs SO beverage.

DISCUSSION: In comparison to MN beverages, SO beverages are associated with a relatively high-energy retention without accompanying subjective hunger/fullness compensations, suggesting a basis for their role in long-term unintentional weight gain in healthy adults.


Consumption of beverages containing only simple sugars, such as soft drinks, has increased in all age groups in the past few decades.1 The additional energy provided by beverages containing a large amount of simple sugars, such as sucrose and corn syrup solids, may in part be responsible for the inordinate weight gain observed in some populations, particularly children and adolescents.2 A recent study by DiMeglio and Mattes3 showed that individuals do not compensate for energy from simple sugars provided in a beverage format but do so when the energy is provided in the form of a liquid. Two mechanisms leading to weight gain with high simple sugars-containing beverage intake can be hypothesized: the first being that the net energy yield (energy intake–energy output) from a beverage containing only simple sugars is greater than that of an equivalent energy intake from a mixed-nutrient (MN) beverage, such as milk, which includes protein and fat;4 and the second being that ingested energy, in the form of simple sugars, is not fully compensated for by later meal ingestion through effects on hunger and fullness.5 The net effect of these two purported mechanisms, alone or in combination, is long-term positive energy balance and weight gain.

Although experimental data are available to support both of these mechanisms,6,7 there exists at present no study in which the two have been quantitatively evaluated following a single meal. Moreover, most of the earlier studies were not of sufficient duration to capture the entire meal response, and the quantitative energy retention difference between meals varying in macronutrient content is thus largely unknown. In addition, earlier studies have mostly examined the effects of large amounts of protein (70–100% of energy intake) on energy expenditure (EE) and satiety, with few trials testing the effects of lower protein content meals (<20% of energy intake), representative of current dietary recommendations.8 The present study tests the hypothesis that a sugar-only (SO) beverage, high in sucrose and corn syrup solids, leads to lower EE but less compensatory postmeal hunger and fullness sensations than a corresponding MN meal.

Subjects and methods


Subjects were recruited through local advertisements and enrolled in the study by telephonic interview. Subjects were required to be healthy, as assessed by the use of a medical history and self-report, and were excluded if they took medications known to influence EE. Additionally, subjects were excluded on the basis of the presence of any of the following: pregnant or lactating females, use of tobacco products, excessive caffeine use (>6 caffeinated beverages/days), severe chronic health conditions, prior gastrointestinal surgery with the exception of distal appendicectomy, acute infections or current use of antibiotic therapy, regular use of medications that impact gastrointestinal function and mental acuity (ie, H2-receptor antagonists, analgesics, and sedatives), inability to consume the two test meals, consumption of alcohol within 24 h of the test days, any type of mental incompetence that would preclude compliance with the protocol, and persons directly involved with the research protocol. The study was approved by the Institutional Review Board of St Luke's/Roosevelt Hospital (New York, NY, USA). In addition, all subjects signed an informed consent prior to participation in the investigation.

Study design

Initially, subjects were randomized to one of two isovolumic, isoenergetic liquid meal sequences, either corn syrup solids (25 dextrose equivalents) and sucrose only (SO) followed by MN or MN followed by SO.

The 2510 kJ (600 kcal) test meals were provided as 590 ml beverages and served chilled. The MN test meal contained 17% protein (milk protein concentrate), 16% fat (canola oil, high oleic sunflower oil, corn oil), and 67% carbohydrate (1:1 corn syrup solids, 25 dextrose equivalents, sucrose) (Boost, Mead-Johnson Nutritional Division, Evansville, IN, USA) (Table 1). The SO test meal consisted of the same carbohydrate source as the MN beverage. Water was then added up to a level of 590 ml to make the beverages isovolumic. The mixture was shaken vigorously and allowed to stand for several minutes for complete dissolution of the powdered mix.

Table 1 Nutrient composition of the mixed--nutrient and sugar-only beverages

Subjects arrived in the laboratory by 0800 after an overnight fast. Body weight was measured to the nearest 0.1 kg via a Weight-Tronix Scale (Scale Electronics Development, New York, NY, USA). Height was measured to the nearest 0.1 cm via a wall-mounted stadiometer (Holtain Limited, Cromwell, Wales, UK). Subjects then rested quietly while lying recumbent on a bed in a thermoneutral, dimly lit room for 30 min.

Energy-expenditure measurements

After the initial 30 min resting period, and while in a supine position, a ventilated canopy was placed over the subject's head and readings from the metabolic cart were taken for approximately 20 min. The baseline resting metabolic rate (RMR) was taken as the average of the last 10 min of the resting period. Subjects then ingested the liquid meal in a maximum time of 15 min and were again placed under the ventilated canopy for the 7 h measurement period.

Gas-exchange indirect calorimetry measurements were made continuously for the entire duration of the measurement period (ie, 420 min) using a ventilated hood (MMC Horizon™ Systems—Nutritional Evaluation Cart, Sensormedics Inc, Yorba Linda, CA, USA). The system was calibrated as specified by the manufacturer using standard gases and the modified Weir equation was used to derive EE from oxygen (VO2) consumption and carbon dioxide (VCO2) production with specific assumptions for nitrogen excretion.9 Readings from the metabolic cart were collected at 3 min intervals.

During each measurement period, subjects were permitted to watch taped documentary programs on a television monitor but were not allowed to sleep. The taped programs were purposely selected such that they would not incite emotions or reactions that would confound the results. Although subjects were allocated a maximum of two 15 min breaks, they were allowed to use the bathroom at their request.

Satiety/hunger questionnaires

At baseline and at 60 min intervals, each subject answered a set of six questions describing their satiety and other meal-related impressions using visual analog scales.10 Subjects were asked to make a mark across a 100 mm line where 0 meant ‘not at all’ and 100 meant ‘very much’. Questions asked included (1) how hungry do you feel; (2) how satisfied do you feel; (3) how full do you feel; (4) how much do you think you can eat; (5) how energetic do you feel; and (6) how sluggish do you feel?

Statistical methods

Following liquid meal ingestion, subjects requested the allowed bathroom breaks and these were occasionally accompanied by aberrant metabolic rate measurements immediately before or after they changed body position. These data points, about 1% of the total, were removed from the analysis. Therefore, of the total 420 min measurement period, approximately 10% of the time points were not included in the analysis; those corresponding to time points without data, such as during the breaks, and the aberrant data immediately preceding or following the break.

RMR was then subtracted from each measured postmeal EE data point to obtain a thermic effect of food value. The thermic effect points were then fitted with a fourth-order polynomial curve. The area under this fitted curve, 420 min in duration, was then taken as the total thermic effect of the meal. This value is expressed in kJ (kcal) and as a percentage of the liquid meal energy intake. A trapezoidal analysis approach was also applied in estimating the total thermic effect and the acquired results were in agreement with the polynomial analysis. Accordingly, we present only the polynomial analysis thermic effect results. A paired t-test was used to determine the differences between the areas under the curve (AUC) obtained after each meal.

Thermic effect values were also averaged over each 1 h period after meal consumption. Hourly thermic effect values were calculated by subtracting RMR from the average hourly EE. Repeated measures mixed model analysis of variance was conducted including effects for sequence, subject within sequence, test period, period-by-subject within sequence, meal, hour, and meal-by-hour interaction as main effects. Individual paired Student's t-tests were then used to test differences between meals at each of the hour periods. In addition, a general linear model analysis was carried out to test the effects of meal, hour, and meal-by-hour interaction after controlling for BMI.

The satiety ratings were analyzed with the same repeated measures mixed model as was used for the hourly thermic effect results. In addition, for each of the six questions, the AUC was computed by the trapezoidal method to determine the total 7 h ratings. For this analysis, sequence, subject within sequence, meal, and period were included in the statistical model.

All statistical analyses were carried out with a combination of Excel (Microsoft, 2002) and SAS 8.1 software (SAS Online Doc, version 8). Data are presented as means±s.e.m. A P-value of <0.05 was taken to establish statistical significance.


There were a total of 20 subjects, 12 men and eight women with a mean age and BMI of 35.1±2.8 y and 26.9±0.96 kg/m2, respectively. The RMR data collected in one subject were erratic and the reasons for this discrepancy in data were unclear. This male subject's data were removed from the thermic effect analysis. From the 19 remaining subjects, six had a BMI that fell within the normal range, while nine and three subjects were overweight and obese, respectively.

Energy expenditure

The RMR was stable over the 10 min measurement period, and RMR on the SO and MN meal days were significantly correlated with each other (RMR-SO (kJ/days)=3.56 × (RMR-MN)+882.8 [kcal/day: 0.85 × (RMR-MN)+211], r=0.88, P<0.01) and had similar mean values of 6518.7±288.7 kJ/days (1558±69 kcal/day) and 6435.0±280.3 kJ/day (1538±67 kcal/day) for the periods preceding SO and MN meal consumption, respectively.

The total thermic effect of the SO meal (274.1±27.6 kJ [65.5±6.6 kcal] or 10.9±1.1% of the energy content of the meal) was significantly lower (P<0.01) than that of the MN meal (372.0±33.9 kJ [88.9±8.1 kcal] or 14.8±1.4%). The SO liquid meal led to a greater net energy retention of 97.9±35.1 kJ (23.4±8.4 kcal) or 3.9±1.4% of the ingested meal compared to the MN meal (Figure 1).

Figure 1

Thermic effects of the two study meals. Data are averages±s.e.m., n=19.

Using the general linear model procedure, there was a significant effect of meal (P=0.01) and hour (P<0.0001) on the thermic effect. The effects of meal (P=0.047) and hour (P<0.0001) remained significant after controlling for BMI. Using regression model analyses with meal, hour, and meal-by-hour interaction, there tended to be an effect of meal (Table 2). The average hourly thermic effects following meal ingestion were significantly lower on the SO meal compared to the MN meal at 1 h (immediately following the meal, P<0.05), 2 h (P<0.05), 6 h (P<0.05), and 7 h (P<0.05). There was a trend for lower thermic effect with SO vs MN at 4 h (P=0.10) and 5 h (P=0.08) after consumption of the meal.

Table 2 Regression model analysis of the effects of meal and hour on the thermic effect

Satiety/hunger data

The satiety and hunger scores are shown in Figure 2, respectively, at baseline and at each time point after MN and SO consumption. Baseline responses to all questions on the satiety questionnaire were not significantly different between MN and SO testing periods. MN beverage consumption resulted in greater feelings of satiety vs SO at 3 h (50.7±4.6 vs 37.5±4.3 mm, P<0.05) and of fullness at 3 h (50.8±4.6 vs 37.3±4.6 mm, P<0.05). AUC values for satiety perceptions were greater (P<0.05) with MN compared to SO consumption. In addition, subjects recorded lower (P<0.05) AUC scores on the question ‘How much do you think you can eat?’ after MN compared to SO consumption. Subjects scored 51.4±2.9 cm h with MN compared to 59.0±2.9 cm h with SO consumption on their perceptions of how much they could eat (P<0.05 for difference between test beverages). Ratings for this question were lower (P<0.05) at 1 and 5 h after MN vs SO consumption and trends, in the same direction, were observed at 2 h (P=0.08) and 3 h (P=0.06) postmeal. There was no difference between test beverages on perceptions of energy and sluggishness.

Figure 2

Basal (h 0) and postprandial scores on the questions ‘How hungry do you feel?’ (a); ‘How satisfied do you feel?’ (b); ‘How full do you feel?’ (c); ‘How much do you think you can eat?’ (d); ‘How energetic do you feel?’ (e); ‘How sluggish do you feel?’ (f) with SO and MN meal consumption. Data are averages±s.e.m., n=20. SO meal=closed squares; MN meal=open squares. *Significantly different between meals, P<0.05, trend for difference between meals, P<0.1.


The present study results support the hypothesis that energy provided by a beverage containing only sugar is retained by the body to a greater extent than a beverage of equal volume and energy content that includes protein and fat. Despite greater energy retention, subjects experienced more feelings of hunger and less feelings of fullness following the SO meal. This combination of decreased EE with diminished postmeal satiety perceptions would, if sustained, create a framework for long-term positive energy balance and weight gain. Alternatively, consumption of an MN beverage in place of an SO beverage may result in improved weight maintenance, due to greater thermogenesis and improved feelings of satiety.

Thermic effect

A classic observation is that the thermic effect of ingested protein is greater than that of an isoenergetic carbohydrate load.8,11,12,13 However, many of the earlier studies were not of sufficient duration nor did investigators employ appropriate measurement systems and protocols for quantitatively evaluating the magnitude of this difference following intake of carefully designed liquid meals.11,12 Our results show that over the 7 h postmeal evaluation period, the between-meal difference for the two study formulas was only approximately 98 kJ (23 kcal) or 4% of the ingested 2510 kJ (600 kcal) meal.

Although relatively small in magnitude, if replicated every day for a high-sugar beverage consuming subject and left uncompensated for by reduced food intake, this would result in a total energy storage of 35.1 MJ (8395 kcal) over 1 y, equivalent to a weight gain of approximately 1 kg.14 This simple example assumes that subjects replace MN beverages with an isoenergetic sugar-containing drink, although actual food practices are more complex. Additional studies are needed to evaluate the nutritional and metabolic effects of SO beverages in differing dietary contexts.

In a recent compilation of studies, Eisenstein et al8 developed a thermic effect prediction formula based upon the amount of protein, fat, and carbohydrate provided in a meal. The Eisenstein formula predicts an 8.1 and 9.3% thermic effect for the SO and MN meals, respectively. The corresponding results in the present study were 10.9 and 14.8%, yielding a between-meal difference of 3.9%. Our proportionally larger thermic effects and greater between-meal difference may be due to longer postmeal measurement periods, 7 h, than most of the earlier studies used in developing the prediction model. However, our results are similar to those of Schutz et al,15 who reported a difference of 3% points in 4 h thermic effect measurements between meals containing 5 and 14% of energy as protein.

Satiety ratings

A reasonable prediction is that greater net energy retention with the SO meal would be accompanied by corresponding reductions in postmeal hunger and increased fullness. In contrast, our results showed the opposite effect, less self-reported feelings of fullness for the SO meal compared to its MN counterpart. In fact, the results from visual analog scale ratings show, for the first time, that adding a small amount of protein and fat to a carbohydrate-rich drink improves feelings of satiety compared to a SO beverage. In fact, subjects had higher scores with the SO compared to the MN beverage when asked how much they thought they could eat. Only one previous study has compared high sugar to MN drinks on subjective feelings of hunger and satiety.16 However, the protein content of the MN beverage was more than twice that fed in the present trial and corresponds to an extremely high protein intake.8,17 The protein content of the beverage tested in the present study is more representative of typical dietary protein intakes. Nevertheless, the results of the present trial agree with those of trials comparing drinks of varying macronutrient content in which high carbohydrate preloads led to increased hunger and lessened satiety compared to high protein preloads.18

High-sugar drinks have been compared to sugar-free beverages for their effects on satiety ratings and food intake. In the trial of Holt et al,19 although the sugar-free drink provided no energy, vs 629 kJ (150 kcal) for the sugar-rich drink, there was no difference in fullness and hunger AUC ratings over the 110 min period after consumption of the different beverages. There is thus some reason to believe that high-sugar, energy-providing beverages may lead to weight gain. In fact, Ludwig et al2 followed children over a period of 19 months and found that the odds of becoming obese increased by a factor of 1.6 for each additional serving of sugar-sweetened drink consumed per day. More recently, Raben et al20 found that subjects supplemented with high-sugar beverages and foods, from which approximately 70% of energy was in a beverage format, gained 1.6 kg of body weight over the 10-week period. In contrast, subjects supplemented with the same amount of artificially sweetened foods and drinks lost 1.0 kg of body weight.

Study limitations

Although our results suggest that consuming high-sugar beverages instead of MN beverages decrease EE and feelings of satiety, suggestive of potential weight-promoting effects of high-sugar beverages, our study did not fully test this hypothesis. Food intake following the two test meals was not measured, and therefore it is unknown whether subjects would have consumed more food after the SO meal compared to the MN meal, as their feelings of hunger and satiety imply. Future, studies are thus needed to evaluate whether the lower EE following an SO beverage compared to an MN beverage is compensated for by lower food intake at a subsequent meal.

The objectives of the present study were focused on the quantitative aspects of a single meal's thermic effect and postmeal satiety ratings. Beverages were closely matched in energy content, carbohydrate source, temperature, and volume. Follow-up studies that systematically varied beverage macronutrient content and sources of carbohydrate, protein, and fat would be useful. Lastly, whether or not the beverages produce actual short- or long-term differences in food intake and thus weight change was not examined in this trial, but would provide useful information in future studies.


In comparison with an MN beverage, results from the present study show that an SO beverage is associated with greater energy retention without subjective hunger/fullness compensation. MN beverages may thus represent a better nutritional alternative as they provide, in addition to metabolic fuel, essential nutrients rather than ‘empty calories’, improved subjective feelings of satiety, and less net energy retention, all of which would favor weight maintenance relative to SO beverages. These observations have fundamental implications and provide a foundation for future long-term clinical and mechanistic studies.


  1. 1

    Nielsen SJ, Siega-Riz AM, Popkin BM . Trends in energy intake in US between 1977 and 1996: similar shifts seen across age groups. Obes Res 2002; 10: 370–378.

  2. 2

    Ludwig DS, Peterson KE, Gortmaker SL . Relation between consumption of sugar-sweetened drinks and childhood obesity: a prospective, observational analysis. Lancet 2001; 357: 505–508.

  3. 3

    DiMeglio DP, Mattes RD . Liquid versus solid carbohydrate: effects on food intake and body weight. Int J Obes Relat Metab Disord 2000; 24: 794–800.

  4. 4

    Dubois EF . Carbohydrate, fat and protein. In: Dubois EF (ed) Basal metabolism in health and disease. Lea & Febiger: Philadelphia; 1936. pp 14–50.

  5. 5

    Bellisle F, Rolland-Cachera MF . How sugar-containing drinks might increase adiposity in children [commentary]. Lancet 2001; 357: 490–491.

  6. 6

    Birch LL, McPhee L, Sullivan S . Children's food intake following drinks sweetened with sucrose or aspartame: time course effects. Physiol Behav 1989; 45: 387–395.

  7. 7

    Belko AZ, Barbieri TF, Wong EC . Effect of energy and protein intake and exercise intensity on the thermic effect of food. Am J Clin Nutr 1986; 43: 863–869.

  8. 8

    Eisenstein J, Roberts SB, Dallal G, Saltzman E . High-protein weight-loss diets: are they safe and do they work? A review of the experimental and epidemiologic data. Nutr Rev 2002; 60: 189–200.

  9. 9

    Weir JB . New methods for calculating metabolic rate with special reference to protein metabolism. Nutrition 1990; 6: 213–221.

  10. 10

    Kissileff HR, Wentzlaff TH, Guss JL, Walsh BT, Devlin MJ, Thornton JC . A direct measure of satiety disturbance in patients with bulimia nervosa. Physiol Behav 1996; 60: 1077–1085.

  11. 11

    Johnston CS, Day CS, Swan PD . Postprandial thermogenesis is increased 100% on a high-protein, low-fat diet versus a high-carbohydrate, low-fat diet in healthy, young women. J Am Coll Nutr 2002; 21: 55–61.

  12. 12

    Luscombe ND, Clifton PM, Noakes M, Parker B, Wittert G . Effects of energy-restricted diets containing increased protein on weight loss, resting energy expenditure, and the thermic effect of feeding in type 2 diabetes. Diabetes Care 2002; 25: 652–657.

  13. 13

    Mikkelsen PB, Toubro S, Astrup A . Effect of fat-reduced diets on 24-h energy expenditure: comparisons between animal protein, vegetable protein, and carbohydrate. Am J Clin Nutr 2000; 72: 1135–1141.

  14. 14

    Pietrobelli A, Allison DB, Heshka S, Heo M, Wang ZM, Bertkau A, Laferrere B, Rosenbaum M, Aloia JF, Pi-Sunyer FX, Heymsfield SB . Sexual dimorphism in the energy content of weight change. Int J Obes Relat Metab Disord 2002; 26: 1339–1348.

  15. 15

    Schutz Y, Bray G, Margen S . Postprandial thermogenesis at rest and during exercise in elderly men ingesting two levels of protein. J Am Coll Nutr 1987; 6: 497–506.

  16. 16

    Latner JD, Schwartz M . The effects of a high-carbohydrate, high-protein or balanced lunch upon later food intake and hunger ratings. Appetite 1999; 33: 119–128.

  17. 17

    St Jeor ST, Howard BV, Prewitt TE, Bovee V, Bazzarre T, Eckel RH . Dietary protein and weight reduction: a statement for healthcare professionals from the nutrition committee of the council on nutrition, physical activity, and metabolism of the American Heart Association. Circulation 2001; 104: 1869–1874.

  18. 18

    De Graaf C, Hulshof T, Weststrate JA, Jas P . Short-term effects of different amounts of protein, fats, and carbohydrates on satiety. Am J Clin Nutr 1992; 55: 33–38.

  19. 19

    Holt SHA, Sandona N, Brand-Miller JC . The effects of sugar-free vs sugar-rich beverages on feelings of fullness and subsequent food intake. Int J Food Sci Nutr 2000; 51: 59–71.

  20. 20

    Raben A, Vasilaras TH, Moller AC, Astrup A . Sucrose compared with artificial sweeteners: different effects on ad libitum food intake and body weight after 10 wk supplementation in overweight subjects. Am J Clin Nutr 2002; 76: 721–729.

Download references


Supported by NIH grant # DK PO1-42618, a grant-in-aid from Mead Johnson Nutritional Division, and fellowship funding for M-P St-O by Bristol Myers Squibb–Mead Johnson.

Author information

Correspondence to S B Heymsfield.

Rights and permissions

Reprints and Permissions

About this article


  • glucose
  • sucrose
  • protein
  • energy expenditure
  • satiety
  • body weight

Further reading