Antoine Lavosier (1743–1794) may have been the first to directly connect digestion with an increased heat production and oxygen consumption 1. Later in the 20^th century it was discovered that increases in energy expenditure after eating, called the thermic effect of feeding (TEF), varied depending on the composition of the meal 2. Metabolic rate not only rises to a greater extent with an increased caloric intake 34 but there is a 6–8% increase in energy expenditure with carbohydrate meals, 3% increase with fat, and 25–40% with protein based meals 2. Research tends to focus on the extent and timing of the peak TEF where measurements can go on for hours 5. Measurement intervals too can be lengthy often occurring in 15 or 30-minute periods. Lengthy measurement intervals reduce measurement variability 6 but they also make it difficult to pinpoint precise time periods; for example, determining when TEF rises.

Sports nutritionists have brought timing issues to attention where it has been recommended that nutrient ingestion post-exercise (i.e., recovery) be completed sooner rather then later to optimize nutrient uptake by muscle 7. Likewise, it is apparent that both colder drinks 8 and lower drink volumes (meal size) 9 tend to hasten the gastric emptying time of a liquid meal. Dependent on the meal consumed blood glucose levels often rise within 10 or 15-minutes but again, measurement periods are often conducted in 10 to 15-minute intervals 1011. In this regard it is not known if the rise in blood glucose above fasting levels is meaningful or not within the first measurement period.

In this investigation we asked the following 'timing' questions: 1) at a given 5% and 10% increase in RMR, when exactly does energy expenditure begin to rise after feeding and, 2) is the onset of TEF different between equi-caloric carbohydrate and protein liquid meals (1255 kJ; 300 kcal)? Comparisons between sexes also were made.

All data are reported as mean ± standard deviation. Subject characteristics for RMR are reported in Table 1: Men had a higher RMR as compared to women for both trials (p < 0.02). No differences in RMR were found between trials within males (p = 0.64), females (p = 0.24) or all subjects (p = 0.77) (Table 1).

For all subjects there were no significant differences in the onset of TEF between the CHO and PRO liquid meals when examined as either a 5% or 10% increase above RMR: 5% CHO, 8.4 ± 6.2 min; 5% PRO, 8.6 ± 5.2 min (p = 0.77); 10% CHO, 14.1 ± 7.5 min; 10% PRO, 16.7 ± 6.7 min (p = 0.36) (Table 2). Several subjects did not show any increase in TEF over the 30-min measurement period: 2 men never had a 5% increase in TEF with PRO; 1 man, 4 women had no 10% increase in TEF with CHO; 3 men, 1 woman had no 10% increase in TEF with PRO (Table 2).

Between sex differences were evident for the initial 5% increase after CHO consumption in men (4.7 ± 2.9 min) as compared to women (10.8 ± 6.6 min) (p = 0.04). No other comparisons revealed significance between the sexes (Table 2).

Previous research has demonstrated the timing of gastric emptying after ingestion of a liquid meal, variability is evident 89. Our study was designed to determine the onset of the TEF. These results indicate that for a compilation of male and female subjects and regardless of whether the (room temperature, 1255 kJ) liquid meal is predominantly carbohydrate or protein, a 5% rise in energy expenditure after feeding occurs in about 8-minutes; a 10% increase in energy expenditure occurs in almost double that time at about 14–16 minutes. It is apparent that subject variability was evident, perhaps in accordance with the considerable between-subject variability that has been found with gastric emptying 89 (see standard deviations in Figures 1 and 2).

This investigation has limitations. Given between-subject variability, subjects could have been classified into specific groups for analyses, for example: comparing age, activity level, or body composition. A more direct investigation may better correlate gastric emptying with TEF. Our methodology also is subject to interpretation. For example, we considered the origin of the TEF as the first of 2 consecutive data points rising 5% and 10% above RMR. Perhaps 3 consecutive data points should have been chosen. Or perhaps the onset of TEF could better be depicted by a different statistical analysis (dependent on a measure of variance or being significantly different from a RMR baseline). We chose a 5% increase in energy expenditure based on the statistical alpha level where significance is found at p < 0.05. Some subjects never achieved 2 consecutive 1-minute rises in energy expenditure: 2 men for 5% increase with PRO, 3 men and 1 women for 10% increase with PRO, 1 man and 4 women for 10% increase with CHO (Table 2). Post-feeding measurements could have been extended to ensure a TEF onset with all participants (a pilot study of 3 subjects provided the rationale for our protocol revealing an observable increase in energy expenditure in less than 10-minutes). These data could also be affected by a limited 3-hour fast, the size (kcal and grams), temperature and type (solid versus liquid) of meal consumed and perhaps the amount of time spent consuming and the palatability of the meal [see 8914]. Female subjects in the current study were not tested according to their menstruation pattern. Research has shown that the phases of menses do 15 and do not 16 affect RMR and the TEF. In the carefully designed study of Tai et al. where TEF decreased during the luteal phase of menstruation, measurement times were conducted in 30-min intervals 15; in our investigation onset time was recorded in 1-minute intervals throughout a 30-min period. Another factor that may have affected our results was the subject's anticipation and added movement after removing the hood for consumption of the drink. This resulted in considerable variation (standard deviation) for the first minute of TEF measurements (the first 2-min of TEF data were not considered as consecutive time points) (see Figures 1 and 2). Extended measurement times ensure less variability but our 1-minute collection periods were designed to help pinpoint the rise in the TEF.

Tai et al. postulated that glucose uptake was slowed during the luteal phase of menstruation suggesting the likelihood of a gender effect. Gender differences were only evident for the initial 5% increase in TEF with the CHO drink where women had a significantly lengthier onset of TEF as compared to men. A larger sample size will be needed to verify this. When using a 10% increase in TEF as a marker of significance, no gender or drink related (CHO vs PRO) differences were found.

Previous research has revealed a rather high degree of individual variability in gastric emptying 89 and the time to peak increase in TEF 5. We conclude that the onset of TEF is likewise variable among subjects but is initiated within about 5 to 20-min for most subjects after ingestion of a 1255 kJ (300 kcal) liquid meal. The onset of TEF does not appear to differ between carbohydrate or protein liquid meals.

The author(s) declare that they have no competing interests.

C.B. Scott conceived and designed the study, performed statistical analysis and wrote the final manuscript. J. Fernandes and M. Lehman conceived and designed the study and collected all data. The authors have read and approve of this manuscript.