点击显示 收起
1 From the Department of Clinical Nutrition, Göteborg University, Göteborg, Sweden, and the Department of Animal Nutrition and Physiology, Danish Institute of Agricultural Sciences, Research Center, Foulum, Denmark.
2 Supported by grants from the Swedish Cancer Foundation and MD Food Ingredients Amba, Videbæk, Denmark. 3 Reprints not available. Address correspondence to L Normén, Department of Clinical Nutrition, Annedalsklinikerna, Göteborg University, S-413 45 Göteborg, Sweden. E-mail: nutrition{at}nutrition.gu.se.
ABSTRACT
Background: The ketohexose D-tagatose is a new sweetener with a low energy content. This low energy content may be due to either low absorption of the D-tagatose or decreased absorption of other nutrients.
Objective: The aims of this study were to measure the excretion of D-tagatose from the human small bowel, to calculate the apparent absorption of D-tagatose, and to study the effects of D-tagatose on the small-bowel excretion of other carbohydrates.
Design: A controlled diet was served for 2 periods of 2 d during 3 consecutive weeks to 6 ileostomy subjects. In one of the periods, 15 g D-tagatose was added to the diet daily. Duplicate portions of the diet and ileostomy effluents were freeze-dried and analyzed to calculate the apparent net absorption of D-tagatose and carbohydrates.
Results: Median D-tagatose excretion was 19% (range: 12–31%), which corresponded to a calculated apparent absorption of 81% (69–88%). Of the total amount of D-tagatose excreted [2.8 g (1.7–4.4 g)], 60% (8–88%) was excreted within 3 h. Between 3 and 5 h, 32% (11–82%) was excreted. Excretion of wet matter increased by 41% (24–52%) with D-tagatose ingestion. Sucrose and D-glucose excretion increased to a small extent, whereas no significant changes were found in the excretion of dry matter, energy, starch, or D-fructose.
Conclusions: The apparent absorption of 15 g D-tagatose/d was 81%. D-Tagatose had only a minor influence on the apparent absorption of other nutrients.
Key Words: D-Tagatose • sweetener • ketohexose • carbohydrates • ileostomy • small bowel • digestion • absorption • excretion
INTRODUCTION
The ketohexose D-tagatose is structurally similar to D-fructose, except for an inverted optically active center. Because of its excellent taste and bulk properties, combined with a possibly very low energy value, D-tagatose has potential for use as a sweetener (1, 2). Different explanations are possible for the low energy value of D-tagatose. A pronounced thermogenic effect of D-tagatose was suggested in growth studies of rats supplemented with the sweetener (1, 3); however, this finding was not supported by a human study in which indirect calorimetry was used (4). D-Tagatose may also suppress the absorption of other carbohydrates, either by inhibiting the intestinal enzymes connected to the brush border membranes or by increasing transit time. Kinetic studies of intestinal enzymes showed that D-tagatose inhibits sucrase-isomaltase (oligo-1,6-glucosidase or sucrose -glucosidase), which may partly explain the lower energy value (5). Another explanation may be a malabsorption of the sweetener, a theory supported by animal studies showing limited D-tagatose absorption in the range of 20–25% (6, 7). In 64% of unadapted human subjects, a single dose of 29 g D-tagatose caused flatulence, which is a typical adverse effect of malabsorbed carbohydrates (8). Increased hydrogen production, as measured by a breath-hydrogen test, was also reported, indicating that part of the 30-g dose was not absorbed but was excreted into the large bowel, where it would be fermented (4).
To calculate the energy value of D-tagatose with the factorial model (9), absorption in the human small bowel must be quantified. Because D-tagatose is metabolized by the liver, urinary excretion cannot be used to study intestinal absorption. The ileostomy model provides a feasible technique for evaluating small-bowel absorption and was used extensively in many studies of carbohydrate absorption (10–15). The ileostomy model provides reliable estimates of complex carbohydrates that have escaped small-bowel absorption (16). For noncomplex carbohydrates such as monosaccharides, disaccharides, and fructooligosacchararides, however, the reliability of this model has not been evaluated.
The objectives of this study were to calculate the apparent absorption of a single 15-g D-tagatose dose by measuring net excretion and transit time. The effects of D-tagatose on the excretion of other carbohydrates, eg, starch, sucrose, D-glucose, and D-fructose, were also examined. Finally, fermentation was evaluated to test the validity of the ileostomy model for D-tagatose excretion.
SUBJECTS AND METHODS
Subjects
Six subjects, 4 men and 2 women, all of whom had been proctocolectomized for ulcerative colitis, participated in the study. The subjects' median age was 64 y (range: 38–74 y) and their median body mass index (in kg/m2) was 25.9 (range: 22.5–30.6). The subjects had well-functioning ileostomies with <10 cm of terminal ileum removed. The median time since the operation was 14 y (range: 5–27 y). Three subjects used no medications. Two subjects used medications for hypertension, 1 of whom also used medication for hypothyroidism, and 1 subject used medication for sacroiliitis. The subjects were otherwise healthy and showed no signs of anemia, inflammation, hepatic disease, or thyroid disease as confirmed by means of history and standard laboratory tests. The study was approved by the Ethics Committee at the Medical Faculty of Göteborg University. Subjects gave their informed consent after being provided with oral and written information about the aim and methods of the study. No conflict of interest existed for any of the authors.
The ileostomy model
The subjects were served a controlled diet during two 2-d periods. D-Tagatose, provided by MD Food Ingredients Amba (Videbæk, Denmark), was added to the diet during one of the periods and the other period served as a control period. The order of the D-tagatose period was assigned randomly and there was a minimum washout time of 5 d between diet periods. The first day of each diet period was used as an adaptation day. On the second day, ileostomy effluents were collected every second hour to minimize bacterial degradation. The bags were sealed immediately and frozen in containers filled with dry ice. The ileostomy effluents were then weighed, freeze-dried, weighed again, homogenized, and pooled individually. The contents of the ileostomy bags were analyzed for gross energy, starch, sucrose, D-glucose, D-fructose, and D-tagatose. To analyze the transit time of D-tagatose, bags were also pooled and analyzed in 5 time intervals at the following hours of the study day: 0800–1200, 1200–1400, 1400–1600, 1600–1800, and 1800–0800 (the next day). Absolute dry mass was measured before the analyses by an extra drying of the freeze-dried samples for 20 h at 104°C.
Diet
The diet was prepared by a trained cook in a controlled manner. Meals were eaten at specific hours, ie, breakfast at 0900, lunch at 1200, coffee break at 1400, dinner at 1700, and an evening snack at 2000. For breakfast, the subjects received orange juice, coffee, sandwiches, and sweetened natural yogurt. Lunch consisted of a vegetarian lasagna and a green salad. During the afternoon the subjects had coffee and a chocolate muffin. Dinner consisted of a ground meat pie, fresh peppers, and a low-alcohol beer. The evening snack was sandwiches and tea. Addition of sugar to coffee or tea was not allowed. The nutrient content of the diet was calculated with the software DIETIST (Näringsdata, Bromma, Sweden; Table 1) by using the Swedish food-composition tables (17). Dietary starch, sucrose, D-glucose, and D-fructose were analyzed.
View this table:
TABLE 1.. Nutrient content of the daily experimental diet
For the D-tagatose period, 15 g D-tagatose was mixed into the breakfast yogurt together with 15 g sucrose. The amount of D-tagatose was restricted to limit gastrointestinal symptoms. The D-tagatose was not a substitute for sucrose but was given in addition to sucrose. This was necessary to evaluate the possible inhibition of sucrose absorption by D-tagatose. A higher dose of sucrose was not used because the saccharide content of the breakfast would then not have represented a normal Swedish diet. During the control period, only sucrose was mixed with the yogurt. To avoid a negative energy balance, subjects were offered extra bread and butter during the study. Intake of water was not limited, but subjects were asked to drink the same amounts at the same hours during the control and intervention periods.
Evaluation of the method
Bacterial degradation of D-tagatose in the morning yogurt by the mix of Lactobacillus acidophilus and Bifidobacterium bifidus was tested. According to the study protocol, 15 g D-tagatose was mixed with 100 g natural yogurt (unsweetened) 15 h before the experimental breakfast. The yogurt was stored in a refrigerator overnight and thereafter freeze-dried and analyzed for D-tagatose.
To evaluate the possibility of nutrient and D-tagatose fermentation in the distal part of the small bowel or in the ileostomy bags, the pooled samples were analyzed for short-chain fatty acids (SCFAs). In a further evaluation after the main study, 2 subjects (subjects 3 and 5) were asked to eat prepared doses of 3 x 5 g D-tagatose/d in addition to their normal diet for 2 consecutive days. On the third day, subjects were asked to change bags before breakfast and then to consume a large breakfast ad libitum. Both subjects changed their morning ileostomy bag 3 h after the meal at the study center. Bags were treated under aerobic conditions and 10 g D-tagatose was added directly to each bag. The bags were sealed and put into a furnace in a water bath with a temperature of 37°C. Every 30 min a 10-mL sample was drawn from the ileostomy bag and frozen immediately. The samples were analyzed for D-tagatose and SCFAs. The density of the wet matter of ileostomy subjects 3 and 5 was measured for calculations of recovery of D-tagatose.
Analytic procedures
Gross energy was analyzed by using bomb calorimetry (Automatic Adiabatic Bomb Calorimetry; Gallenkamp, Loughborough, United Kingdom). Total starch was analyzed by a modified method of Englyst et al (18). D-Glucose, D-fructose, and sucrose were measured according to an enzymatic-colorimetric assay (19). D-Tagatose was analyzed by HPLC essentially as described by Johansen et al (20), with 50% ethanol as the extraction medium and eluent and sorbitol as an internal standard. Extraction took place in a heating block at 50°C for 60 min. Samples were cleaned with a Sep-Pak Alumina-B cartridge (Waters Corp, Milford, MA) before evaporation. D-Tagatose was separated from the other components by using a calcium-based resin column (Aminex HPX-87C; BioRad Laboratories, Hercules, CA). The content of D-tagatose in the in vitro incubates was determined by capillary gas chromatography after silanization of hydroxy groups with sorbitol (1 g/L) as an internal standard (21). Concentrations of SCFAs were measured according to the method of Jensen et al (22).
Calculations and statistics
The apparent absorption of nutrients and D-tagatose was calculated as intake minus excretion as a percentage of intake. After the tested dose of D-tagatose was dried in the same way as the duplicate portion of the diet and the ileostomy effluents, on a dry matter basis, the dose was found to be 14.2 g, which was later used in the calculations. The transit of D-tagatose was estimated as the recovery of D-tagatose in the 5 separate time intervals, calculated as a percentage of the total D-tagatose excretion. Net wet and dry weight were estimated as the total weight minus the amount of D-tagatose excreted. The total amount of SCFAs was calculated as dry matter (mmol/kg) multiplied by the total excreted amount of dry matter (kg/24 h). Because the density of the ileostomy effluents of subjects 3 and 5 was 1.0 kg/L, the recovery of the incubated D-tagatose was calculated by multiplying the tagatose concentration by the total wet weight.
Statistics were performed with the statistical package SYSTAT for WINDOWS (version 7.0; SPSS Inc, Chicago). Results are presented as medians and ranges because of the lack of a normal distribution. Pairwise comparisons were made for variables of interest (excretion of energy, wet matter, dry matter, starch, sucrose, glucose, D-fructose, and SCFAs) with the nonparametric Wilcoxon's signed-rank test at a probability of 95%.
RESULTS
The amount of D-tagatose excreted was low, corresponding to 19% (12–31%) of intake, indicating a high apparent absorption (Table 2). The median calculated apparent absorption was 81%, with individual variations from 69% to 88%.
View this table:
TABLE 2.. Individual excretion and apparent absorption of D-tagatose
The transit time of D-tagatose was short. Within 3 h, two-thirds of the total amount excreted had passed the small bowel; between 3 and 5 h, the final one-third was excreted (Table 3). Four subjects excreted most of the D-tagatose within 3 h of ingestion; the 2 remaining subjects excreted the most between 3 and 5 h of ingestion. After 5 h, most of the given dose of D-tagatose had been excreted.
View this table:
TABLE 3.. Excretion of D-tagatose and wet and dry weights during different time intervals
After the D-tagatose excreted was subtracted from the median wet weight of the ileostomy excreta, the wet matter increased by 258 g (124–394 g), corresponding to a 41% (24–52%) increase in basal content (P = 0.028; Table 4). No significant difference in the net excretion of dry matter was found between the control diet period and the D-tagatose period. When we examined the excretion of D-tagatose over the same 5 intervals used to study transit time, it became clear that D-tagatose increased the excretion of wet matter during the first interval (0800–1200). The excretion of dry matter also increased during the first interval (0800–1200) during the D-tagatose period.
View this table:
TABLE 4.. Ileal excretion of energy, carbohydrates, and short-chain fatty acids (SCFAs) in 6 ileostomy subjects during a control period and an intervention with 14.2 g D-tagatose added to the diet daily1
The effect of D-tagatose on carbohydrate digestibility is expressed here as the change in ileostomy excretion of starch, sucrose, D-glucose, and D-fructose. The increase in sucrose excretion was small but significant, corresponding to a sucrose absorption of 98.9% (range: 98.2–99.6%) with the control diet alone and of 96.7% (range: 95.5–98.8%) with the addition of D-tagatose (P = 0.046). The same pattern was seen for D-glucose absorption, which decreased from 99.6% with the basal diet alone (range: 97.8–100%) to 98.3% with the addition of D-tagatose (range: 96.4–99.3%; P = 0.046). There were no significant effects of D-tagatose on the excretion of total energy, starch, or D-fructose. The analysis of D-tagatose in the yogurt revealed a 2–6% breakdown of D-tagatose during the 15 h of refrigeration.
The median total amount of SCFAs (mmol) in the dry matter of the pooled bags (24 h) increased with the addition of D-tagatose (P = 0.046), but it is important to note that 2 of 6 subjects (subjects 2 and 5) had lower SCFA amounts during the D-tagatose period than during the control period. Concentrations of D-tagatose and SCFAs during the 4 h of incubation of the ileostomy bags of subjects 3 and 5 remained essentially the same. The wet weights of the ileostomy bags were 58 g for subject 3 and 348 g for subject 5. This large difference in wet weight content induced a large difference in D-tagatose concentration even though concentrations did not decrease over time in either of the 2 incubated bags. During the first 2.5 h, D-tagatose concentrations in the bag of subject 3 were lower than the final concentration (probably because of an incomplete mix of D-tagatose in the bag). Thereafter, the D-tagatose concentration remained 163–194 g/L for the last 1.5 h. The bag of subject 5 had less variation in D-tagatose concentration and remained in the range of 24.8–28.5 g/L during the whole 4-h period. The recoveries of the added 10 g D-tagatose after 4 h corresponded to 101% for subject 3 and 89% for subject 5. Fermentation of D-tagatose, as indicated by elevated SCFA concentrations, did not seem to occur because the SCFA concentrations remained stable during all 4 h: in the range of 39–50 mmol/L for subject 3 and 36–42 mmol/L for subject 5.
None of the subjects chose to eat extra bread during the study, which simplified the comparison of excretion patterns. Subject 1–5 reported no gastrointestinal complaints during the study. Subject 6 reported an increased excretion of wet mass during the D-tagatose period, but otherwise experienced no discomfort during the study.
DISCUSSION
To our knowledge this is the first study of the small-bowel excretion of D-tagatose in humans. The calculated apparent absorption in the present study was high compared with the 20% D-tagatose absorption described in animals (6, 7). This difference may be partly explained by the relatively high doses of D-tagatose used in the animal studies. A dose-response relation has never been evaluated in humans, but the importance of such a relation depends on the mechanism of D-tagatose absorption. Incubation of rat intestinal epithelial brush border membranes with D-fructose and D-tagatose showed that despite a ratio of 1:100 or 1:150, there was no inhibition of D-fructose transport over the epithelia (23, 24). Thus, despite the structural similarity of D-tagatose and D-fructose, D-tagatose could have an absorption mechanism in humans other than the active saturated process shown for D-fructose. If passive diffusion is the only mechanism available for D-tagatose absorption, absorption would be expected to be lower than that found in the present study. Because a low absorption rate is typical for monosaccharides and polyols of similar molecular size, which are absorbed by passive diffusion (9), the result of the present study raises doubts as to whether D-tagatose is only passively absorbed in humans. Thus, another factor, eg, a higher ileal fermentation in persons with ileostomies than in pigs, may explain the discrepancy in results between the animal studies and the current human study.
To calculate the effect of D-tagatose on energy metabolism, it is important to evaluate the effects on the absorption of other carbohydrates. There are different ways in which D-tagatose may affect the absorption of other nutrients. The first is by inducing a more rapid transit time. This possibility is supported by the results of a study by Buemann et al (4), who found an increased production of hydrogen as early as 0.5–1.5 h after ingestion of D-tagatose. The second possibility is a reduction in the transmucosal gradient as a result of a change in osmosis. A lower gradient, due to an osmotic dilution of gut content, is likely to reduce absorption of nutrients. The increase in wet weight in the present study was probably due to an osmotic overload induced by the unabsorbed D-tagatose. Considering the 53% increase in wet weight for ileostomy subjects who ingested 5 g lactulose (25), the unabsorbed amount of D-tagatose, 2.8 (1.7–4.4) g, might create a similar response. Inhibited sucrase (sucrose -glucosidase) activity is a third possibility that could further diminish absorption (5). D-Tagatose substitution in pigs, corresponding to 10% by wt of the daily diet, decreased the digestibility of sucrose to only 90% compared with 96% in the controls and thereby inhibited sucrose absorption (7). In the present study, sucrose absorption was high with both diets and malabsorption after the addition of D-tagatose corresponded to <8 kJ. The small but significant malabsorption in response to a 15-g dose of D-tagatose seemed to be of minor importance for the absorption of other nutrients and thereby also for energy reduction. The quantity of 33 g sucrose in the present study corresponded to only 6% of energy, lower than the average intake in an American population, which can be as high as 12% (26). It is possible that if a higher dose of sucrose had been used in the experimental design, relatively higher malabsorption would have resulted, which would have led to greater energy reduction.
There are 2 direct methods for studying carbohydrate excretion from the small bowel (the ileostomy model and the intubation technique) and 1 indirect method (the breath-hydrogen technique). The breath-hydrogen technique, however, is not quantitative and can be used only to qualitatively confirm the results of other methods (16). The validity of the ileostomy model has been questioned because of an assumption that ileostomy subjects have an abnormally rapid transit time (27). However, several studies showed that the physiology of the small intestine is similar in ileostomy subjects and in subjects with an intact gastrointestinal tract and that the transit time through the small bowel is also similar (13, 28–30). In previous ileostomy studies, transit times of 4–6 h were observed (10–12, 15).
It has been hypothesized that the distal part of the small bowel in ileostomy subjects has an active microbial population (9). Bacterial fermentation is minimized with the frequent changing of ileostomy bags and the rapid freezing of the bags by the subjects. As shown by the incubation study of ileostomy bags, D-tagatose did not seem to be fermented for even up to 4 h at 37°C, supporting the idea that D-tagatose is not easily fermented in the bags. Measurement of the total amount of SCFAs excreted into the ileostomy bags has the disadvantage of not taking into account the SCFAs that might have been absorbed in the human small bowel. In the present study, the total amount of SCFAs produced in both periods was low compared with the 300–400 mmol estimated to be produced by the flora of a normal colon every day (31), even though this does not rule out the possibility of a more active bacterial population in the distal part of the small bowel in ileostomy subjects than in healthy subjects. Bacterial counts were estimated to be 1 x 106/g in ileostomy effluents, compared with 1 x 1010/g produced by a normal fecal flora (32), with ratios of aerobic to anaerobic bacteria in ileostomy effluents of 1000:1 compared with 1:1000 in the flora of a normal colon.
The small but significant increase in the total amount of SCFAs with the addition of D-tagatose to the diet was probably a consequence of fermentation of the increased amounts of sucrose and glucose not absorbed, and probably also to some extent the unabsorbed D-tagatose. The question is then, To what extent would this increase in SCFAs account for the 60% difference in D-tagatose excretion found between animals and humans? A comparison of SCFA concentrations in several ileostomy studies, which were in the same range as in this study, concluded that the SCFAs excreted correspond to an average fermentation of 2–4 g carbohydrate (KE Bach Knudsen, Å Lia, H Andersson, P Åman, and G Hallmans, unpublished observations, 1997). It is reasonable to assume that a similar figure is valid also for this study and that the carbohydrate fermented consisted of dietary fiber, starch, other carbohydrates, and, in part, D-tagatose. Considering the limited aerobic flora, which has a short time to exert its action on D-tagatose because of its rapid transit time, it seems unlikely that fermentation in the distal ileum could explain the large difference in the figures for apparent absorption of D-tagatose between the animal studies and the present study. Assuming that one-half of the mentioned maximum value corresponding to fermentation of carbohydrates, ie, 2 g, would be D-tagatose, this would still mean a median apparent absorption of 66% (55–74%). This figure is still considerably higher than the one reported in animals.
Also important to note is an ileostomy study of the fructooligosaccharide inulin, which gave estimates of absorption of 80%, even though inulin is used as a prebiotic compound that stimulates the growth of anaerobic bacteria (33). In that study, inulin was not fermented, which provides additional support to the validity of the ileostomy model for studies of other saccharides. The hypothesis of a highly active microbial population in the distal part of the small bowel in ileostomy subjects is therefore not likely to be correct. However, apparent absorption of sorbitol, another sweetener with a low molecular size, was shown to be higher with lower doses in ileostomy studies (34). It thereby seems reasonable that a saturable fermentation by the aerobic ileostomy flora of small monosaccharides and polyols cannot be completely excluded.
In summary, the present study showed that 15 g D-tagatose/d had a high apparent absorption in the small intestine of humans, even though some bacterial degradation of this highly fermentable hexose cannot be ruled out. A minor overestimation of D-tagatose absorption could be the result of the fermentation that takes place inside the gut, which is not easily estimated. Moreover, ingestion of D-tagatose had only a minor influence on the apparent absorption of other nutrients. Despite its limitations, the ileostomy model seems to still be the most accurate way of quantitatively studying the small-bowel absorption of carbohydrates because other techniques for human studies of carbohydrate excretion and absorption have more serious methodologic problems.
ACKNOWLEDGMENTS
We thank Vibeke Malmros, Anette Almgren, Maj-Lis Montheli, and Jacob Jensen for excellent technical assistance and Lars Ellegård for scientific advice.
REFERENCES