J Nutr Sci Vitaminol, 60, 246–254, 2014

Evaluation of the Relative Available Energy of Several Dietary Fiber Preparations Using Breath Hydrogen Evolution in Healthy Humans

Tsuneyuki Oku and Sadako Nakamura

Graduate School of Human Health Science, University of Nagasaki, Siebold, Nagasaki 851–2195, (Received December 9, 2013)

Summary A standardized simple, indirect method for assessing the relative energy of dietary fiber carbohydrates is not yet established. There is a need for a standardized in vivo assay. The objective of the present study was to evaluate the relative available energy (RAE) for 9 major dietary fiber materials (DFMs) based on fermentability from breath hydrogen excretion (BHE) in subjects. Fructooligosaccharide (FOS) was used as a reference. The study was conducted using a within-subject, repeated measures design and approved by the Ethi- cal Committee of University of Nagasaki. After DFM ingestion, end-expiratory gas (750-mL) was collected at 1-h intervals for 8 h, as well as at 2-h intervals between 8 h and 14 h, and 30 min after waking up and 24 h after DFM ingestion. Breath hydrogen concentration was assessed with a gas chromatograph. The RAE of DFMs tested was evaluated based on the area under the curve (AUC) of BHE of FOS. Based on the ratio of AUC for 8 h, the RAE of polydextrose, partially hydrolysed guar gum, resistant maltodextrin and partially hydrolysed alginate was 1 kcal/g, and that of glucomannan, heat-moisture treatment and high-amy- lose cornstarch and cellulose was 0 kcal/g, while the RAE of all tested DEMs including cel- lulose and glucomannan was 1 kcal/g in the calculation based on AUCs for 14 h and 24 h in subjects. We suggest that a breath hydrogen collection period of 14 h or more could be used to measure RAE for a range of fiber preparations in vivo. Key Words relative available energy, dietary fiber, fermentability, breath hydrogen

Dietary fiber (i.e., largely indigestible carbohydrates) However, this is not practical, because balance studies has beneficial effects for health and is actively used in are time-consuming, expensive, and very stressful for health foods and functional foods (1, 2). Dietary fiber subjects. Although an indirect and simple method to taken orally reaches the large intestine without being evaluate the available energy of nondigestible carbohy- digested by a-amylase or disaccharidases present in the drates is desirable, it has not been established yet. mucosa of the small intestine. In the large intestine, The amount of available energy of nondigestible dietary fiber is fermented to varying degrees by intes- and fermentable carbohydrates is dependent upon the tinal microbes and metabolized into short chain fatty amount of short-chain fatty acids produced in the fer- acids, carbon dioxide, hydrogen, methane, and bacterial mentation by intestinal microbes. Thus, energy produc- cell components (3–6). Among them, short-chain fatty tion is dependent on the fermentability of nondigestible acids are absorbed from the large intestine and used carbohydrates which reach the colon. We estimated the solely as the energy source of the host. That is, even relative available energy value (RAE) of dietary fiber carbohydrates that are not digested and not absorbed materials (DFMs) based on the breath hydrogen excre- provide available energy to a living body by being fer- tion (BHE) and calculated the relative ratio versus the mented and absorbed in the large intestine (5). amount of hydrogen excretion from the ingestion of Energy values for fermentable fibers are an important fructooligosaccharide (FOS), which is fermented com- matter for regulatory bodies as they are incorporated pletely by intestinal microbes. Therefore, the profile of into on-pack labels, and for weight control products, BHE seems to be different from the fiber sources. But, as a help to direct consumer choice (2). The available FOS is a convenient reference because it is fermented energy of nondigestible carbohydrate in humans is eval- completely by intestinal microbes. Besides, FOS has been uated using balance studies and other methods (7, 8). classified as a nondigestible carbohydrate of which the energy value is 2 kcal/g by the Japanese Health Promo- E-mail: [email protected] tion Law (9). In this study we expressed the energy val- Abbreviations: AOAC, Association of Official Analytical Chemists; AUC, area under the curve; BHE, breath hydrogen ues derived from the indirect method as the relative ratio excretion; DFMs, dietary fiber materials; FOS, fructooligosac- to FOS. charide; GM, glucomannan; HMT-HACS, heat-moisture treat- In the nutrition labelling for processed foods, the ed, and high amylose cornstarch; PD, polydextrose; PHAA, indication of energy is an essential item. The available partially hydrolysed alginic acid-Na; PHGG, partially hydro- energy derived from digestible carbohydrates is 4 kcal/g. lysed guar gum; RAE, relative available energy; RMD, resistant However, that of nondigestible carbohydrates is not maltodextrin. known because the available energy derived from nondi-

246 Evaluating Relative Energy of Dietary Fiber 247

Table 1. Preparation of test materials.

Soluble fibers Polydextrose (Lytes; Danisco Japan Ltd.) Resistant maltodextrin (Fibersol-2; Matsutani Chemical Industry Co., Ltd.) 5 g of test substance was dissolved in 120 mL of soup. Partially hydrolyzed guar gum (Sunfiber-PG; Taiyo Kagaku Co., Ltd.) Partially hydrolyzed sodium alginate (Sorgin granule; Kaigen Co., Ltd.) Glucomannan 5 g of glucomannan was gelated in 120 mL of soft drink (Propol-A; Shimizu Chemical Corp.) over 24 h. Insoluble fibers HMT-HACS (Loadstar; Nihon Shokuhin-Kako Co., Ltd.) 5 g of test substance was suspended in 120 mL of potage. Cellulose (Avicel; Asahi Kasei Chemicals Corp.) Possitive control Fructooligosaccharide 5 g of FOS was dissolved in 120 mL of water. (Meiji Seika Kaisha Ltd.)

HMT-HACS: heat-moisture treated, high amylose cornstarch. gestible and fermentable carbohydrates such as oligosac- to participate in the study. All experiments were carried charides cannot be evaluated for common foods listed in out in the Public Health Nutrition Laboratory of the standard food tables. To estimate the available energy Graduate School of Human Health Science, Siebold Uni- from nondigestible and/or nonabsorbable carbohydrates versity of Nagasaki. such as oligosaccharides and sugar alcohols, we pro- Subjects. Nine healthy females, aged 21.860.6 y posed a unique and simple method which enables calcu- and with a BMI of 20.162.4 kg/m2, voluntary partici- lation from the fermentation equations with the ratio of pated in this study. The exclusion criteria were a history short-chain fatty acids produced by intestinal microbes of gastrointestinal diseases, carbohydrate malabsorp- (10–13), and demonstrated that a carbohydrate which tion, obesity, or pulmonary disease. The average and is fermented completely has an energy available value standard deviation (SD) values of fasting blood glucose of 2 kcal/g. This value is used for nutrition labelling in levels of the subjects was 82.764.1 mg/100 mL. Prior The Japanese Health Promotion Law (9). The estimated to the experiments, we ensured that all of the subjects available energy based on BHE is classified into three were hydrogen producers and not methane producers. energy categories, 0 kcal/g, 1 kcal/g, and 2 kcal/g, The subjects had not taken antibiotics or laxatives for which are expressed as integers for the necessity of prac- 2 wk prior to the experiments. tical use in the calculation of nutritional intake. Materials. To evaluate the available energy, DFMs We have proposed an indirect and simple method and which are frequently used in foods for specified health tried to evaluate the energy available from some DFMs uses and functional foods were chosen. The water solu- based on the fermentability from BHE for 8 h (13). The ble dietary fiber materials were polydextrose (PD), resis- results obtained were different from the energy coef- tant maltodextrin (RMD), partially hydrolysed guar gum ficients of DFMs based on the fermentability obtained (PHGG), partially hydrolysed alginic acid-Na (PHAA) from animal and human experiments in the Japanese and glucomannan (GM); water insoluble dietary fiber Health Promotion Law. Furthermore, the values of materials were heat-moisture treated high amylose corn- available energy obtained were not correct because the starch (HMT-HACS) and cellulose. period of breath-gas collection for 8 h to calculate the PD (Lytes, Danisco Japan Ltd., Tokyo, Japan) was a fermentability was too short. Therefore, in the present mixture of polymers in which glucose, sorbitol and study, we collected breath gas for 24 h after the inges- citrate (90 : 9 : 1) were condensed under high pressure tion of DFMs, and tried to evaluate again the RAE of and high temperature, and had an average molecular DFMs in healthy human subjects. weight of 1,500. It contained ~25% of monosaccha- rides and oligosaccharides. The available energy of PD MATERIALS AND METHODS has been estimated as 0.8–1.2 kcal/g by other meth- All of the experiments complied with the code of eth- ods, administered orally as 14C-polydextrose to rats and ics of the World Medical Association (Declaration of humans (14, 15). RMD (Fibersol-2, Matsutani Chemical Helsinki, Oct. 2008). The study protocol was approved Industry Co., Ltd., Osaka, Japan) is a mixture of mono- by the Ethical Committee of Siebold University of Naga- saccharides, oligosaccharides and glucose polymers, saki (received No. 41–1, approval No. 39, Nagasaki, and contains 93.73% dietary fiber (by the Association of Japan). All subjects provided written informed consent Official Analytical Chemists (AOAC) method). The aver- 248 Oku T and Nakamura S age molecular weight was ~2,000. The available energy refrigerator overnight (13). of RDM has been estimated as 0.98 kcal/g by indirect Special meals for experiments. The period over which calorimetric measurements in healthy adults (7). PHGG the subjects were restricted with regard to food intake (Sunfiber-PG, Taiyo Kagaku Co., Ltd., Mie, Japan) had was very long. Hence, special meals and snacks from an average molecular weight of .20,000 and con- which hydrogen is not produced when ingested were tained .80% dietary fiber (by the AOAC method). The given to participants during the experiment. All food- ratio of galactose to mannose was 1 : 2. PHAA (Sorgin stuffs used in the meals had been confirmed not to pro- granule, Kaigen Co., Ltd., Osaka, Japan) had a molecu- duce breath hydrogen. The combinations of a cookie lar weight of 40,000–70,000 (average, 55,000). GM (Butter Cookie Sable 60 g, University Coop., Tokyo), (Propol-A, Shimizu Chemical Corp., Hiroshima, Japan) canned tuna fish (Sea-chicken Mild, 80 g, Hagoromo had an average molecular weight of .1,000,000 and Food Co., Ltd., Shizuoka, Japan), boiled egg (50 g), and the ratio of d-glucose to d-mannose was 1 : 1.6. HMT- sport drinks (Pocari Sweat 250 mL) were used as food HACS (Loadstar, Nihon Shokuhin-Kako Co., Ltd., Tokyo, materials for special meals (breakfast, lunch, supper, Japan) comprised 64% resistant starch (by the Prosky snacks). The amount of energy and protein intake is method). Cellulose (Avicel, Asahi Kasei Chemicals Corp., shown in Table 2. Tokyo, Japan) was a polymer in which glucose bonded On the day of the experiment, all subjects took the with a b-1,4 linkage, and contained .99% of dietary special meals and snacks given at the indicated time in fiber. a day (Fig. 1). If needed, they could drink or These DFMs were used to evaluate the available en- water. If the subject had an empty stomach, a sports ergy. FOS (purity, .98%; Meiji Seika Kaisha Ltd., Tokyo, drink containing sugars was given. In addition, each Japan) containing 1-kestose (36.8%), nystose (51.9%), subject was given a multivitamin tablet (Nature Made, 1F-b-fructofranosyl-nystose (9.3%), and monosaccha- Otsuka Pharm. Co., Ltd.) to supply some vitamins every rides (2%) was used as a reference for BHE which was d ay. completely fermented and not evacuated to feces (16, Experimental protocol. The present study had a 17). within-subject, repeated-measures design. All test sub- Preparation of test substances. We used water soluble stances were given in a random order with intervals of and insoluble dietary fibers and FOS as a reference in 1 wk. Experiments were carried out under the direc- the present study. The practical dose for dietary fiber tion of a physician (Fig. 1). The experimental proto- was less than 5 g. In addition, DFMs are tasteless, fla- col was carried out in accordance with the methods vourless and odourless, so test materials were prepared employed in our previous study (13, 18). to be ingested readily (Table 1). That is, PD, RMD and PHGG were dissolved in 120 mL of boiled soy flavoured soup (dried powder 6 g, Riken Vitamin Co., Ltd., Tokyo, Table 2. Energy and protein of experimental meals. Japan). PHAA, HMT-HACS and cellulose were dissolved or suspended in 120 mL of boiled cream potage (dried Energy (kcal) Protein (g) powder 17 g, Ajinomoto Co., Ltd., Tokyo, Japan). We used just frozen dried powder in soy flavoured soup and Meal-1 437 9.7 improvised cream potage, dissolved in boiling water Meal-2 587 22.2 and filtered after being well stirred. The soups had pre- Snack-1 366 3.5 viously been confirmed not to excrete breath hydrogen Meal-3 412 9.4 using the subjects participating in this study. GM (5 g) Snack-2 336 3.2 was suspended in 120 mL of (Pocari Sweat, Otsuka Pharm. Co. Ltd., Tokyo, Japan); 8 g of sucrose Total 2,138 48.0 was added, and GM formed a hard gel and was stored in Per person, per day. a refrigerator overnight. The tolerance level of FOS due Cookie 60 g, canned tuna fish 80 g, boiled egg 50 g, sport to osmotic effects in the large bowel is 0.3 g/kg of body drink 250 mL were used as food materials for special weight in Japanese (4). So, in the present study FOS (5 g) meals. The energy value and protein content were used was dissolved in 120 mL of tap water and stored in a from the food labelling.

Times of gas collection 1 2 3 4 5 6 7 8 9 10 11 12 13 14

- 2 0 1 4 8 10 12 14 22 24 （h） Ingestion of in bed wake up test material Sleeping (23:00) (6：30) Meal-1 Meal-2 Snack-1 Meal-3 Snack-2 Meal-1 (7:00） (12:00) (17:00) (20：00) （7：30)

Fig. 1. Experimental protocol of breath hydrogen gas test. Evaluating Relative Energy of Dietary Fiber 249

40 a,b,c - - FOS a,b,c a,b,c - - RMD a,b,c - - PHGG 30 a,b,c - - PD

a,b,c a,c 20

a a a a a b b a 10 b b b c c a Hydrogen gas excretion (ppm) c c b c c c 0 0 2 4 6 8 10 12 14 16 18 20 22 24 Sleeping period Period after ingestion of test materials (h)

Fig. 2. Profiles of breath hydrogen excretion by oral ingestion of polydextrose (PD), resistant maltodexitrin (RMD), and partially hydrolyzed guar gum (PHGG) in healthy subjects. Data are expressed as the mean and SD (n59). a–c: signifi- cantly different between the same letters at each time point, at p,0.05 by Dunnett’s post hoc test. FOS: fructooligosac- charide used as a reference.

After an overnight fast, the subject’s health status The result of the ratio was classified into three energy was examined, the food intake for the previous week categories based on the Japanese Health Promotion Law. was reviewed, and blood pressure and pulse rate were It was evaluated to the category of 2 kcal/g if the ratio measured. Then 750-mL samples of end-expiratory gas of test material versus FOS was 0.75; the category of were collected. After ingestion of the test substance, the 0 kcal/g if the ratio was #0.25; the category of 1 kcal/g end-expiratory gas was collected at 1-h intervals for 8 h, if the ratio was between 0.25 and 0.75. The AUC was and then at 2-h intervals between 8 h and 12 h after the considered to be significant if p,0.05 was obtained by ingestion, with a special collection bag after the removal two-sided analysis using ANOVA and Dunnett’s post hoc of dead-space gas. The sleeping period was between test using SPSS for Windows (Japan version) 12.0 (SPSS 14 h and 20 h after ingestion of the test substance. Inc., Tokyo, Japan). Breath gas was collected 30 min after waking up and RESULTS 24 h after ingestion. Participants consumed their usual diet but were pro- Subject participation hibited from ingesting foods containing nondigestible No subject dropped out of the study and none of the carbohydrates starting 3 d before each experimental subjects experienced side effects. Their health status day. We provided a supper, which consisted of cooked through the entire study period was good. The average paddy white rice, fried chicken, a small amount of cab- intake of energy by a subject was 2,0486220 kcal/d bage, and soup, on the day before each experiment. Pre- and that of protein was 45.266.0 g/d. viously, we evaluated that breath hydrogen was detected Breath hydrogen excretion (BHE) following consumption of negligibly by the ingestion of these foodstuffs given for dietary fiber test foods supper. During the experiment, they were also prohib- The profiles of BHE values with mean values and SDs ited from ingesting foods or beverages except for water, of 9 subjects are shown in Figs. 2–4. as well as from sleeping or smoking. Subjects were a) FOS. When each subject was given FOS (5 g), all required to sit on a chair and were prohibited from exer- subjects excreted breath hydrogen gas and did not pro- cising with hyperventilation until the final collection of duce methane gas. Breath hydrogen gas (which is pro- expiratory gas (13, 18). Participants ingested an experi- duced through only the fermentation caused by intesti- mental meal from which breath hydrogen was not pro- nal microbes) started to be excreted ~3 h after the intake duced, so that they would not feel hungry. of FOS. The amount of hydrogen gas excreted reached Analyses of breath hydrogen. Five millilitres of end- the peak at 5–6 h and decreased gradually until 14 h expiratory gas was sucked into a plastic syringe and after ingestion. BHE was maintained 24 h after ingestion loaded into a simple gas chromatograph (Breath Gas and did not recover to basal levels. The concentration of Analyzer BGA1000D, Laboratory for Expiration Bio- hydrogen in the first collection of breath gas after wak- chemistry Nourishment Metabolism Co., Ltd.) to mea- ing up was slightly higher than that taken before sleep- sure the concentrations of hydrogen and methane gas. ing, but that of hydrogen after 24 h was small (Fig. 2). Calculations and statistical analyses. BHE was calcu- The amount of breath hydrogen produced from DFMs lated as mean values and SDs; normal distribution was was markedly lower than that from FOS. tested. Evaluation of the available energy of DFMs was b) PD, RMD and PHGG. When PD (5 g) was ingested calculated based on a ratio. The ratio was calculated by subjects, BHE started to increase 2–3 h after ingestion based on the areas under the curve of hydrogen excre- and reached a peak 5 h after ingestion (Fig. 2). Thereaf- tion by the ingestion of test materials versus that of FOS. ter, BHE decreased gradually and was excreted little by 250 Oku T and Nakamura S

40 a - - FOS a,b a,b - - GM a,b - - PHAA 30 a,b

a,b b 20

10 b a a a,b b Hydrogen gas excretion (ppm) b a a b a b 0 0 2 4 6 8 10 12 14 16 18 20 22 24 Sleeping period Period after ingestion of test materials (h)

Fig. 3. Profiles of breath hydrogen excretion by oral ingestion of glucomannan (GM) and partially hydrolyzed alginic acid- Na (PHAA) in healthy subjects. Data are expressed as the mean and SD (n59). a, b: significantly different between the same letters at each time point, at p,0.05 by Dunnett’s post hoc test. FOS: fructooligosaccharide used as a reference.

- - FOS 40 a a,b - - HMT-HACS a,b a,b - -Cellulose 30 a,b

a,b b 20

10 b b b Hydrogen gas excretion (ppm) b b b a a a a 0 0 2 4 6 8 10 12 14 16 18 20 22 24 Sleeping period Period after ingestion of test materials (h)

Fig. 4. Profiles of breath hydrogen excretion by oral ingestion of heat-moisture treated, high amylose cornstarch (HMT- HACS) and cellulose in healthy subjects. Data are expressed as the mean and SD (n59). a, b: significantly different between the same letters at each time point, at p,0.05 by Dunnett’s post hoc test. FOS: fructooligosaccharide used as a reference.

little until the end of experiment. PD contains ~20% of ~3 h after ingestion and reached a peak at 5 h (Fig. 3). low-molecular-weight saccharides which are ferment- PHAA was very resistant to fermentation by intestinal able. The total amount of breath hydrogen excreted was microbes even though it has a low molecular weight and markedly lower than that for FOS. is water-soluble. GM was very resistant to fermentation When RMD (5 g) was digested by subjects, breath by intestinal microbes, and the start of BHE was very hydrogen started to be excreted 2–3 h after ingestion late (7–8 h after ingestion). However, the total amount and reached a maximum 5–6 h after ingestion (Fig. 2). of breath hydrogen excreted for 14 h was similar to that Thereafter, BHE decreased gradually until 14 h after of PHAA (Fig. 3). ingestion. A part of RMD was readily fermented by d) HMT-HACS and cellulose. When HMT-HACS (5 g), intestinal microbes, but was less than that of FOS. The which is very poorly soluble in water, was ingested by total amount of breath hydrogen excreted for 14 h was subjects, the start of BHE was late and very low. BHE between that of PD and that of FOS. For the ingestion started to increase 8 h after ingestion and showed a of PHGG (5 g), breath hydrogen started to be excreted peak at 10 h. The fermentability was lower than that of 2–3 h after ingestion and reached a maximum 6–8 h PHAA and GM (Fig. 4). When cellulose (5 g), which is after ingestion (Fig. 3). Thereafter, BHE decreased grad- water-insoluble, was ingested, the excretion profile of ually until 14 h after ingestion as in the case of RMD. breath hydrogen was similar to that of HMT-HACS, and The fermentation of PHGG by intestinal microbes was the total amount of breath hydrogen excreted for 14 h poorer than that of RMD, but was much more than that was similar to that of HMT-HACS (Fig. 4). of PD. Area under the curve (AUC) of BHE 24 h after ingestion c) PHAA and GM. When PHAA (5 g) was ingested of DFMs by subjects, breath hydrogen started to be excreted The rate of fermentation by intestinal microbes was Evaluating Relative Energy of Dietary Fiber 251

Table 3. Areas under the curve versus time of breath hydrogen gas excretion.

AUC AUC AUC (ppm/8 h) (ppm/14 h) (ppm/24 h)

Fructooligosaccharide 6,00062,100 11,90064,300 13,00064,300

Glucomannan 1,30061,100 3,10062,500 4,10063,000 Guar gum 2,70061,900 5.50063,700 6,60064,300 Sodium alginate 1,6006 1,000 3,60063,100 4,10063,300 Polydextrose 2,0006 600 3,90061,400 5,00062,200 Resistant maltdextrin 3,50061,600 7,10063,900 8,30064,300 Cellulose 1,4006 800 4,00062,100 4,90062,700 HMT-HACS 1,0006 600 4,40062,400 4,90062,700

Data are expressed mean and SD (n58). AUC: areas under the curve versus time. HMT-HACS: heat-moisture treated, high amylose corn starch.

Table 4. Energy estimation from ratio of AUC versus FOS.

8 h-collection 14 h-collection 24 h-collection Estimation by Health Promotion Ratio Estimated Ratio Estimated Ratio Estimated Law vs FOS energy vs FOS energy vs FOS energy (kcal/g) (%) (kcal) (%) (kcal) (%) (kcal)

FOS 100.0 2 100.0 2 100.0 2 2

Glucomannan 21.7 0 26.1 1 31.5 1 2 Guar gum 45.0 1 46.2 1 50.8 1 2 Sodium alginate 26.7 1 30.3 1 31.5 1 0 Polydextrose 33.3 1 32.8 1 38.5 1 0 Resiatant maltdextrin 58.3 1 59.7 1 63.8 1 1 Cellulose 23.3 0 33.6 1 37.7 1 0 HMT-HACS 16.7 0 37.0 1 37.7 1 2

Data are expressed mean and SD (n58). AUC: areas under the curve versus time. HMT-HACS: heat-moisture treated, high amylose cornstarch. different according to the type of DFM, so the AUC of tine) is ingested, it is completely fermented by intestinal BHE 8 h, 14 h and 24 h after ingestion was calculated microbes and not excreted to feces (16, 17). The avail- using the results in Fig. 2 to Fig. 4, and the results of able energy of FOS is 2 kcal/g. The RAE of DFMs based AUC are shown in Table 3 and the relative ratios versus on the AUC of FOS is shown in Table 4. Thus, in the cal- FOS are shown in Table 4. culation based on AUC for 8 h, the RAE value for PD, The AUC of the reference of FOS (the available PHGG, RMD, and PHAA was 1 kcal/g and that for GM, energy of which has been evaluated to be 2 kcal/g) HMT-HACS and cellulose was 0 kcal/g. However, in the was 6,000 ppm for 8 h, 12,000 ppm for 14 h, and calculation based on the AUC for 14 h and 24 h, the 13,000 ppm for 24 h. The increment in the AUC RAE of all the DFMs tested became 1 kcal/g. Thus, the between 14 h and 24 h after ingestion was slight. To RAE of all DFMs tested was similar. However, all RAEs calculate the relative ratio, these AUC values of FOS of the DFMs obtained from a short period after ingestion were set at 100% at 8 h, 14 h and 24 h, respectively. were lower than those of the long period after ingestion, The AUC of PD was 2,000 ppm for 8 h, 3,900 ppm and were different according to the type of dietary fiber. for 14 h, and 5,000 ppm for 24 h (Table 3). The rela- In terms of appropriate experimental periods, these tive ratios versus FOS were 33.3 for 8 h, 32.8 for 14 h results demonstrate that even GM (which is a water- and 38.5% for 24 h (Table 4). These values were signifi- soluble dietary fiber with a high molecular weight) and cantly smaller than those of FOS (p,0.05). Using the cellulose (which is a water-insoluble dietary fiber with a same method, the AUC and the relative ratios of each high molecular weight) are slowly fermented by intesti- test material were calculated. nal microbes and produce short chain fatty acids. There- Evaluation of relative available energy (RAE) of each dietary fore, we must use the fermentability value obtained from fiber material (DFM) based on the available energy of FOS a long period (.14 h) after the ingestion of test mate- When FOS (which is not digested in the small intes- rials, when the available energy of the dietary fiber is 252 Oku T and Nakamura S evaluated based on BHE. used in energy evaluation of the Japanese Health Pro- motion Law (9). DISCUSSION To calculate the RAE from the AUC for each DFM, we Dietary fiber reaches the large intestine where it is pre-supposed that the AUC of BHE for FOS was 100%, so fermented to varying degrees by the resident microbial the available energy was 2 kcal/g. The estimated RAEs population. Short-chain fatty acids produced by fer- of PHGG, PHAA, PD, and RMD by 14-h collection were mentation are utilized solely as the energy source of the 1 kcal/g and they were in the same category after 8-h host. However, an indirect and simple method used to collection. However, although the RAEs of GM, cellu- evaluate the available energy of nondigestible and fer- lose, and HMT-HACS by 8-h collection were 0 kcal/g, mentable carbohydrates has not been established. In the they increased to 1 kcal/g after 14-h collection. The present study, we tried to evaluate the available energy value of RAE was different according to the collection of several DFMs using a method that calculates from period of breath gas, because the velocity of fermenta- the fermentation equations which consist of the ratio of tion by intestinal microbes is different in the materials short-chain fatty acids produced by intestinal microbes tested. (10–12). The fermentability of DFMs was estimated by The available energy of RMD has been already dis- measuring the excretion of breath hydrogen in healthy cussed and evaluated as 1 kcal/g using the CO2 and O2 subjects. It was an indirect but very simple method and ratio in a balance study (7). In the present study using had potential to estimate the available energy of DFMs. an indirect and simple method based on BHE, the RAE Fermentability reflects the degree of breakdown of of RMD was evaluated as 1 kcal/g, and the value con- fiber carbohydrates and the production of short chain formed to that of the balance study (21). This supports fatty acids. It is affected by the quantity as well as the the validity of the indirect method using BHE. Further- type of DFM consumed, because intestinal microbes more, Livesey et al. reported that the fermentation of have the fermentation capacity, and the amount of avail- PD was low and that there were interactions between able energy is changed by fermentability. In the pres- in vitro and in vivo study (21). Moreover, Figdor and ent study, 5 g of each test material was ingested by the Rennhard and Anchour et al. evaluated the available subjects, because 10 g of tested materials could not energy of PD as 1 kcal/g in human and rat experiments be tolerated by the subjects. Furthermore, the amount using [14C]-PD (14, 15). These results are not contradic- of DFM which is added to processed foods with health tory to the value obtained from the present study. benefits is ,5 g. We consider 5 g of test material to be Cellulose is not hydrolyzed by a-amylase or small an appropriate dose to evaluate the available energy of intestinal enzymes. Although cellulose has very low fer- DFMs. mentability and yields few short chain fatty acids in in It has been stated in the detailed enforcement regu- vitro incubation using human feces (22, 23), cellulolytic lations for the application of the Japanese Health Pro- bacterium has been isolated in the human gut microbial motion Law (9) that carbohydrate that is not digested in community (24–26) and it was different between pres- the small intestine and which is completely fermented ence and absence of methanogen (26). These reports in the large intestine has 2 kcal/g of available energy. support our result that BHE was detected in cellulose The value is derived from some proposed fermentation ingestion and the RAE was not 0 kcal/g in this study. equations (11–13). FOS is a typical nondigestible oli- BHE by the ingestion of GM was clearly lower than gosaccharide with such properties as well as lactulose that of FOS in this study, and the RAE of GM was evalu- (16–18). Hence, we used FOS (2 kcal/g) as a refer- ated as 1 kcal/g. Matsuura reported that glucoman- ence to calculate the RAE of several DFMs in the pres- nan was degraded almost 100% by soluble enzymes in ent study. Livesey reported that the available energy human feces in an in vitro study (27). But, GM increases of digestive carbohydrate is 4 kcal/g and that of full- the fecal volume and accelerates the defecation (28). fermentable carbohydrate is a half of that of available These results suggest that the fermentation or degrada- carbohydrate, 2 kcal/g (19). In addition, the EU, Canada tion by intestinal microbes is different between in vivo and Australia/New Zealand established 2 kcal/g of the study using humans and in vitro study. energy value for fermentable fiber (20). Therefore, it is HMT-HACS, which is similar to high amylose maize reasonable that the energy value of FOS as a reference starch, showed incremental changes between 8 h and is 2 kcal/g, as used in the present study. 14 h with regard to the AUC. It was reported that the A DFM that is not fermented by intestinal microbes excretion of breath hydrogen by high amylose maize does not produce short-chain fatty acids, so the avail- starch in healthy humans increased markedly (29), able energy is 0 kcal/g. Therefore, the available energy and high amylose maize starch is partly digested in the of DFMs which are fermentable becomes between human intestine by ileostomy (30) and in pigs (31). 0 kcal/g and 2 kcal/g. However, in the present study, we However, the fermentability is affected by source, and used integers after rounding off because complicated particle size (29, 32). HMT-HACS employed in this study numerical values are not practical in nutrition educa- comprised 64% resistant starch, and the remaining tion or nutrition labelling. Nevertheless, the value of carbohydrates were difficult to be digested by intestinal available energy for each DFM is calculated based on the enzymes. Thus, 1 kcal/g is valid for HMT-HACS. fermentability from BHE and is not an integer. In addi- The fermentability of nondigestible carbohydrate is tion, the category of 0 kcal/g, 1 kcal/g and 2 kcal/g was dependent on the preparations, its molecular weight Evaluating Relative Energy of Dietary Fiber 253 and the particle size of ingested test materials. Fur- Shokuhin-Kako Co., Ltd. (Tokyo, Japan) for HMT-HACS, thermore, using a within-subject, repeated-measures Asahi Kasei Chemicals Corp. (Tokyo, Japan) for cellulose, design is important, because the population of intes- and Meiji Seika Kaisha Ltd. (Tokyo, Japan) for FOS. This tinal microbes is different according to individual and study was supported in part by Health Science Research also ethnicity. Therefore, the present study in which the Grants from the Ministry of Health and Welfare in Japan available energy of dietary fibers is expressed as RAE and university funding for research to T.O. The authors appears to be suitable. declare that there is no conflict. T.O. designed research Thus, the evaluated available energy of some DFMs and had primary responsibility for final content, and was different according to the collection period of BHE. T.O. and S.N. conducted research, analysed data and However, the available energy which was based on the performed statistical analysis. Both authors read and 24-h AUC was identical to that of 14-h AUC. In terms approved the final manuscript. of the detection of increasing hydrogen at 24 h after REFERENCES ingestion as shown in the figures, our examination was based on a 36 h continuous experiment, and we con- 1) Oku T, Nakamura S. 2008. Analysis for “Food for Speci- cluded that the detection of hydrogen in the morning fied Health uses” currently approved and short history was artifact. These results demonstrated that the collec- on FOSHU regulatory system. Japan J Dietary Fiber Res 12: 51–58 (in Japanese). tion period of breath hydrogen must be 14 h because 2) Yamada K, Sato-Mito N, Nagata J, Umegaki K. 2008. the velocity of fermentation by intestinal microbes is dif- Health claim evidence requirements in Japan. J Nutr ferent according to the type of DFM. Hence, we recom- 138 (Suppl): s1192–s1198. mend a 15-h collection period to evaluate the available 3) Miller TL, Wolin MJ. 1981. Fermentation by the human energy and to alleviate the stress of subjects. However, intestine microbial community in an in vitro semi- the available energy of GM, PHAA, PD and RMD was continuous culture system. Appl Environ Microbiol 42: not different among values based on 8-h AUC, 14-h 400–407. AUC and 24 AUC because these materials are readily 4) Oku T, Nakamura S. 2002. Digestion, absorption, fer- fermented by intestinal microbes, as is FOS. mentation and metabolism of functional sugar sub- In the detailed enforcement regulations for the appli- stitutes and their available energy. Pure Appl Chem 74: cation of The Japanese Health Promotion Law (9), the 1253–1261. 5) Livesey G, Elia M. 1988. Estimation of energy expendi- values of available energy (energy coefficient) used are ture, net carbohydrate utilization, and net fat oxidation 2 kcal/g for GM, PHGG and HMT-HACS, 1 kcal/g for and synthesis by indirect calorimetry: evaluation of RMD, and 0 kcal/g for PHAA, PD and cellulose, respec- errors with specific reference to the detailed composition tively. Apart from RMD, the value for DFMs obtained of fuels. Am J Clin Nutr 47: 608–628. from the present study were not in accordance with the 6) Cummings JH, Macfarlane GT. 1991. The control and values noted in the detailed enforcement regulations for consequences of bacterial fermentation in the human the application of The Japanese Health Promotion Law. colon. J Appl Bacteriol 70: 443–459. The reason is that fermentability values were obtained 7) Goda T, Kajima Y, Suruga K, Tagami H, Liversey G. from animal experiments and in vitro experiments (not 2006. Availability, fermentability, and energy value of human experiments) and then used to evaluate the resistant maltodextrin: modelling of short-term indirect available energy. In the present study, the evaluation of calorimetric measurements in healthy adults. Am J Clin Nutr 83: 1321–1330. available energy was carried out based on the results 8) Poppitt DS, Livesey G, Elia M. 1998. Energy expenditure from human experiments alone. Therefore, the energy and net substrate utilization in men ingesting usual and coefficients of DFMs in the present study are recom- high amounts of nonstarch polysaccharide. Am J Clin mended for the nutrition labelling. Nutr 68: 820–826. We tried to evaluate the RAE of nondigestible and fer- 9) The Ministry of Health and Labour and Welfare. 2012. mentable carbohydrates using the fermentability based Detailed enforcement regulations for the application on BHE for .14 h, and found that the collection period of Japanese Health Promotion Law. [Online]. Available: 8 h after ingestion of the test material is too short to http://www.mhlw.go.jp/[accessed December 25, 2012]. evaluate the available energy. The method which we 10) Oku T. 1988. Available energy of nondigestible carbohy- propose in the present study is indirect but very simple drate and intestinal microbes. Shoku no Kagaku No. 121: and valid. The energy coefficient of several DFMs should 79–82 (in Japanese). 11) Oku T. 1994. Special physiological functions of newly be used for the nutrition labelling of processed foods. developed mono- and oligo-saccharides. In: Functional Foods: Designer Foods, Pharmafoods, Nutraceuticals Acknowledgments (Goldberg I, ed), Chapter 10: p 202–218. Chapman & The authors would like to thank the subjects who Hall, New York & London. participated in this study. We also wish to thank the fol- 12) Oku T. 1996. Oligosaccharides with beneficial health lowing commercial organisations based in Japan for pro- effects: A Japanese perspective. Nutr Rev 54: s59–s66. viding test materials: Danisco Japan Ltd. (Tokyo, Japan) 13) Nakamura S, Oku T. 2005. Evaluation of available for PD, Matsutani Chemical Industry Co., Ltd. (Osaka, energy of several dietary fiber materials based on the fer- Japan) for RMD, Taiyo Kagaku Co., Ltd. (Mie, Japan) for mentability from breath hydrogen excretion in healthy PHGG, Kaigen Co., Ltd. (Osaka, Japan) for PHAA, Shi- human subjects. Japan J Dietary Fiber Res 9: 34–46 (in mizu Chemical Corp. (Hiroshima, Japan) for GM, Nihon Japanese). 254 Oku T and Nakamura S

14) Figdor SK, Rennhard HH. 1981. Caloric utilization and 74: 303–322. disposition of [14C]polydextrose in the rat. Agric Food 24) Kopecny J, Hajer J, Mrazec J. 2004. Detection of cellu- Chem 29: 1181–1189. lolytic bacteria from the human colon. Folia Microbiol 15) Anchour L, Flourie B, Briet F, Pellier P, Marteau P, (Praha) 49: 175–177. Rambaud JC. 1994. Gastrointestinal effects and energy 25) Chassard C, Delmas E, Robert C, Lawson PA, Bernalier- value of polydextrose in healthy nonobese men. Am J Donadille A. 2012. Ruminococcus champanellensis sp. Clin Nutr 59: 1362–1368. Nov., a cellulose-degrading bacterium from human gut 16) Oku T, Tokunaga T, Hosoya N. 1984. Nondigestibility microbiota. Int J Syst Evol Microbiol 62: 138–143. of a new sweetener, “Neosugar,” in the rat. J Nutr 114: 26) Chassard C, Delmas E, Robert C, Bernalier-Donadille 1574–1581. A. 2010. The cellulose-degrading microbial commu- 17) Tokunaga T, Oku T, Hosoya N. 1989. Utilization and nity of the human gut varies according to the presence excretion of a new sweetener, fructooligosaccharide or absence of methanogens. FEMS Microbiol Ecol 74: (Neosugar), in rats. J Nutr 119: 553–559. 205–213. 18) Oku T, Nakamura M, Hashiguchi-Ishiguro M, Tanabe K, 27) Matsuura Y. 1998. Degradation of glucoman- Nakamura S. 2009. Bioavailability and laxative thresh- nan by enzymes in human feces and formation of short- old of 1-kestose in human adults. Dyn Biochem Process chain fatty acids by intestinal anaerobic bacteria. J Nutr Biotechnol Mol Biol 3: 90–95. Sci Vitaminol 44: 423–436. 19) Livesey G. 2001. A perspective on food energy standards 28) Marzio L, Del Bianco R, Donne MD, Pieramico O, Cuccu- for nutrition labeling. Br J Nutr 85: 271–287. rullo F. 1989. Mouth-to-cecum transit time in patients 20) Bureau of Nutritional Sciences, Food Directorate, Health affected by chronic constipation: Effect of glucoman- Products and Food Branch, Health Canada. 2010. Pro- nan. Am J Gastoenterol 84: 888–891. posed policy: definition and energy for dietary fiber. 29) Van Munster IP, Trangerman A, Nagengast FM. 1994. 21) Livesey G, Johnson IT, Gee JM, Smith T, Lee WE, Hillan Effect of resistant starch on colonic fermentation, bile KA, Meyer J, Turner SC. 1993. “Determination” of sugar acids metabolism, and mucosal proliferation. Dig Dis Sci alcohol and polydextrose absorption in humans by the 39: 834–842. breath hydrogen (H2) technique: the stoichiometry of 30) Zhou Z, Topping DL, Morell MK, Bird AR. 2010. Changes hydrogen production and interaction between carbohy- in starch physical characteristics following digestion drates assessed in vivo and in vitro. Eur J Clin Nutr 47: of foods in the human small intestine. Br J Nutr 104: 419–430. 573–581. 22) Vince AJ, McNeil NI, Wager JD, Wrong OM. 1990. The 31) Bird AR, Vuaran M, Brown I, Topping DJ. 2007. Two effect of lactulose, pectin, arabinogalactan and cellulose high-amylose maize starches with different amounts of on the production of organic acids and metabolism of resistant starch vary in their effects on fermentation, ammonia by intestinal microbes in a faecal incubation tissue and digesta mass accretion, and bacterial popula- system. Br J Nutr 63: 17–26. tions in the large bowel of pigs. Br J Nutr 97: 134–144. 23) Barry JL, Hoebler C, MacFarlane GT, Mathers JC, Reed 32) Ferguson LR, Tasman-Jones C, Englyst H, Harris PJ. KA, Mortensen PB, Nordgaard I, Rowland IR, Rumney 2000. Comparative effects of three resistant starch CJ. 1995. Estimation of the fermentability of dietary preparations on transit time and short-chain fatty acids fiber in vitro: a European interlaboratory study. Br J Nutr production in rats. Nutr Cancer 36: 230–237.