Chewing increases postprandial diet-induced thermogenesis

Slow eating, which involves chewing food slowly and thoroughly, is an effective strategy for controlling appetite in order to avoid being overweight or obese. Slow eating also has the effect of increasing postprandial energy expenditure (diet-induced thermogenesis). It is still unclear whether this is due to oral stimuli; that is, the duration of tasting food in the mouth and the duration of chewing. To investigate the effects of oral stimuli on diet-induced thermogenesis in 11 healthy normal weight males, we conducted a randomized crossover study comprising three trials: (1) drinking liquid food normally, (2) drinking liquid food after tasting, and (3) adding chewing while tasting. Oral stimuli (i.e., the duration of tasting liquid food in the mouth and the duration of chewing) significantly increased diet-induced thermogenesis after drinking liquid food. This result demonstrates that the increase in diet-induced thermogenesis is due to oral stimuli rather than the influence of the food bolus. Increased diet-induced thermogenesis induced by chewing and taste stimuli may help to prevent overweight and obesity.

different days, with consecutive trial separated by more than 3 days. The subjects arrived at the laboratory at the same time between 8:30 a.m. and 10:00 a.m. after having abstained from eating, consuming caffeinated or alcoholic beverages, and intense exercise since dinner on the previous night, i.e., they had fasted for more than 10 h. Each subject was seated on a chair in a semisupine position in a quiet room in which the temperature and humidity were controlled to within 25.4 ± 0.4 °C and 47 ± 7%, respectively. After allowing the subjects to adjust to the experimental setup for 20 min, baseline data of gas-exchange variables and the splanchnic circulation were recorded while resting for 20 min. The subjects completed a 100-mm visual analogue scale (VAS) questionnaire to assess their hunger and fullness before the test drink. A 200-mL cocoa-flavored drink was divided into ten 20-mL cups. After performing baseline measurements, the subjects swallowed the ten 20-mL test drinks over a 5-min experimental period in three ways. In the control trial, subjects swallowed one 20-mL test drink every 30 s. In the long-duration taste stimulation trial (taste trial), subjects kept the 20-mL test drink in their mouth for 30 s without chewing, and then swallowed it; stimulation was taste only. In the chewing stimulation trial (chewing trial), subjects chewed the 20-mL test drink for 30 s at a frequency of once per second, and then swallowed it; stimulation was both taste and chewing. Gas-exchange variables and the splanchnic circulation were measured until 90 min after swallowing the test drink.
Test drink. The subjects consumed the same 200-mL cocoa-flavored drink (Calorie Mate Can, Otsuka, Japan; 200 kcal; protein 7.6 g, fat 4.4 g, and carbohydrate 31 g) as the test drink at the three trials. The temperature of the test drink was measured using an infrared thermometer (A&D, Japan), and the test drink was provided at a controlled temperature (7.4 ± 0.5 °C).
Hunger and fullness. Subjects scored their hunger and fullness on a 100-mm VAS before the test drink.
The subjects marked a tick on the line to indicate their feeling, with the score corresponding to the distance in millimeters from the left starting point of the line to the tick. The left and right ends of the scale were labelled "not at all" and "extremely", respectively.

Gas exchanged variables and DIT.
Oxygen uptake (·VO 2 ), carbon dioxide output (·VCO 2 ), and respiratory exchange ratio (RER) were measured using a gas analyzer (AE-310S, Minato Medical Science, Japan) on a breath-by-breath basis before and after the test drink. The average data every 15 min and the last 15 min of the resting baseline were used for the analysis. Energy expenditure at resting baseline (REE) and after the test drink was calculated using the abbreviated Weir equation 28 : DIT was calculated from the postprandial increments in energy expenditure above the resting baseline. Accumulation for DIT over 90 min was calculated area under the curve (AUC) using the trapezoidal rule.
Protein oxidation g/min = resting energy expenditure (REE, kcal/min) × 0.152 / 4 kcal  Splanchnic circulation. The heart rate (HR) and the mean blood velocities (MBVs) and vessel diameters of the celiac artery (CA) and superior mesenteric artery (SMA) were measured. HR was determined using the electrocardiograph (MEG-2000, Nihon Kohden, Japan). Simultaneous pulsed and echo Doppler ultrasound flowmetry was used to measure the MBVs and vessel diameters of the CA and SMA, as in previous studies of our research group 21,22,30,31 . A curved-array Doppler-scan probe was operated at a pulsed Doppler frequency of 3.3 MHz (LOGIQ P6, GE Healthcare, USA), and the Doppler beam insonation angle relative to the blood vessel was maintained at ≤ 60°. After obtaining these signals for measuring MBV for 1 min, a cross-sectional image of the vessel was recorded for 30 s. This was repeated every 5 min. The B-mode images sent from the Doppler monitor were recorded to enable later measurement of the vessel diameters using image-editing software (ImageJ 1.47, Wayne Rasband, National Institute of Mental Health, USA). The HR, MBVs and EMG signals of the chewing muscles were sampled at 20 kHz using an A/D converter (PowerLab 8/30, ADInstruments, Australia). The spectra of the MBV signals were analyzed offline with our own Doppler signal processing software, and beat-bybeat MBV values were calculated. MBV was determined by averaging the ten largest values in each minute (for 1 min every 5 min) in order to eliminate data variations originating from the abdominal movement associated with respiration 30,31 . The blood flows (BFs) in the CA and SMA were calculated using the following formula: Accumulation for splanchnic BF in the CA and SMA over 90 min was calculated from incremental AUC using the trapezoidal rule.
Statistical analysis. The sample size was estimated using G*Power 3.1 32 , using the data from a previous study that investigated the effects of the number of chews and meal duration on DIT 22 . To detect changing of DIT with a power of 80% and an alpha level of 5%, a sample size of more than six subjects was required. All statistical analyses were performed with SPSS (IBM SPSS Statistics 21.0, IBM, Japan). A P value of less than 0.05 was considered statistically significant, and the data were presented as mean ± SEM (range) values. One-way analysis of variance (ANOVA) was used to compare hunger and fullness scores before the test drink, baseline data of gasexchange variables, substrate oxidation, and splanchnic circulation, and accumulation for DIT, substrate oxidation, and splanchnic BF over 90 min after consuming the test drink among all trials. Two-way repeated ANOVA was used to examine effects of trials and time on time course data. When a significant F value was detected, this was further examined by using Bonferroni's post-hoc test.
Gas-exchange variables, DIT, and substrate oxidation. The time courses of the gas-exchange variables (·VO 2 , ·VCO 2 , and RER) are shown in Fig. 2. The gas-exchange variables at the resting baseline did not differ significantly among the trials. Significant interaction of trial and time was found for ·VO 2 . ·VO 2 was significantly greater in the chewing trial than in the control trial at 45-60 min after the test drink. The duration over which ·VO 2 was significantly greater than the resting baseline was longer in the chewing trial than in the control and taste trials. Significant interaction of trial and time was found for ·VCO 2 . ·VCO 2 was significantly greater in the chewing trial than in the control trial at 45-60 min after the test drink. RER did not differ significantly among the trials.
A significant interaction of trial and time was found for DIT (Fig. 2). DIT was significantly greater in the taste and chewing trials than in the control trial; these differences continued until 90 min after the test drink between chewing and control trials, and was evident at 45-75 min between taste and control trials. The duration over which DIT was significantly greater than the resting baseline was longer in the taste and chewing trials than in the control trial.
The time courses of the substrate oxidations (protein, fat, and carbohydrate) are shown in Fig. 3. Significant interaction of trial and time was found for protein oxidation. Protein oxidation was significantly greater in the chewing trial than in the control trial at 45-60 min after the test drink. Fat and carbohydrate oxidation did not differ significantly among the trials. Splanchnic circulation. The time courses of the splanchnic circulation (MBV, diameter, BF for the CA and SMA) are shown in Fig. 4. The splanchnic circulation at the resting baseline did not differ significantly among the trials. Significant interaction of trial and time was found for CA MBV. CA MBV was significantly greater in the chewing trial than in the taste trial at 30-45 min after the test drink. Significant interaction of trial and time Fat oxidation g/min = 1.67×VO 2 (L/min) −1.67×VCO 2 (L/min) − 0.307× protein oxidation g/min Carbohydrate oxidation g/min = 4.55×VCO 2 (L/min) −3.21×VO 2 (L/min) − 0.459× protein oxidation g/min .
Blood flow (mL/min) = π × r radius of artery (mm) 2 × MBV(m/s) × 60.   Circles, triangles, and squares denote data from the control, taste, and chewing trials, respectively. ·VO 2 oxygen uptake, DIT diet-induced thermogenesis, ·VCO 2 carbon dioxide output, RER respiratory exchange ratio. * P < 0.05, vs. resting baseline in each trial. † P < 0.05, difference between control and taste trials. ‡ P < 0.05, difference between control and chewing trials. # P < 0.05, difference between taste and chewing trials. . Time courses of changes in substrate oxidation (protein, fat, and carbohydrate) in the control, taste, and chewing trials. Circles, triangles, and squares denote data from the control, taste, and chewing trials, respectively. * P < 0.05, vs. resting baseline in each trial. ‡ P < 0.05, difference between control and chewing trials.

Discussion
DIT refers to the increased energy production after consuming a meal, and it was found to increase with the duration of each taste stimulation and the duration of chewing. This result, supporting our hypothesis, demonstrates that oral stimuli (i.e., the duration of tasting food in the mouth and the duration of chewing) increase DIT, rather than representing the influence of the food bolus. CA BF increased with the duration of each of taste stimulation and with the duration of chewing. CA supplies blood to the liver, stomach, abdominal esophagus, spleen, and the superior halves of the duodenum and the pancreas, and so the motility of the upper gastrointestinal tract was increased by taste and chewing. This notion is consistent with our previous report of an increase in CA BF associated solely with taste and chewing food 30 .
In this study we used a beverage as the stimulus in order to avoid the influence of the food bolus, which meant that we were only examining the influence of oral stimuli on DIT. We have previously shown using blocky foods and regular foods that DIT increases with the chewing duration 21,22 , but did not control for the effect of the food bolus. By using beverages, we succeeded in eliminating this effect in the present study.
The present study provides novel insight into the mechanism underlying the increase in DIT induced by taste and chewing. We have also provided evidence that the oral stimuli provided by the combination of taste and chewing are important to increasing DIT. Thus, slow eating, which involves chewing food slowly and thoroughly, increases DIT and may be an effective strategy for preventing overweight and obesity.
The increase in DIT induced by chewing is smaller after consuming solely liquid food than normal food. We have previously shown that chewing increases DIT: by 6 kcal and 10 kcal over 90 min after eating 100-kcal and 300-kcal blocky foods, respectively, and by 15 kcal over 180 min after eating 621 kcal of regular food including pasta 21,22 . Comparison with these previous results reveals that we obtained a smaller DIT by using a 200-kcal . Time courses of changes in splanchnic circulation in the control, taste, and chewing trials. Circles, triangles, and squares denote data from the control, taste, and chewing trials, respectively. CA celiac artery, SMA superior mesenteric artery, MBV mean blood velocity, BF blood flow. * P < 0.05, vs. resting baseline in each trial. † P < 0.05, difference between control and taste trials. ‡ P < 0.05, difference between control and chewing trials. # P < 0.05, difference between taste and chewing trials. www.nature.com/scientificreports/ beverage: tasting increased DIT from 3.4 kcal to 5.6 kcal over 90 min, and adding chewing increased it to 7.4 kcal over 90 min. Thus, the increase in DIT by chewing was 1.8-4.0 kcal over 90 min. The use of liquid reduces the chewing stimulation, resulting in a smaller DIT, which is consistent with our hypothesis. This reveals that not only oral stimuli (i.e., the duration of tasting food in the mouth and the duration of chewing) but also the size of the food bolus may contribute to DIT.

Conclusion
Oral stimuli (i.e., the duration of tasting food in the mouth and the duration of chewing) increase DIT. We speculate that overweight and obesity may be avoided by chewing and tasting via increases in DIT.