| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
ORIGINAL RESEARCH COMMUNICATION |
1 From the Clinical Nutrition & Risk Factor Modification Center, St Michael's Hospital, Toronto, Canada (VV, ALJ, DJAJ, ALR, JLS, and EJ), and the Department of Nutritional Sciences, University of Toronto, Toronto, Canada (VV and DJAJ)
2 Supported by Kellogg Company. 3 Reprints not available. Address correspondence to V Vuksan, Departments of Nutritional Sciences and Medicine, Faculty of Medicine, University of Toronto, Clinical Nutrition & Risk Factor Modification Center, St Michael's Hospital, 70 Richmond Street East, Toronto, ON M5C1N8, Canada. E-mail: v.vuksan{at}utoronto.ca.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Objective: We aimed to assess the effects of increasing fiber intake on bowel habits and gastrointestinal tolerance in healthy persons consuming a typical Canadian or US diet.
Design: Under a randomized crossover design, 23 free-living participants consumed a typical Canadian or US diet (35% fat, 12 g fiber/d) and received 25.0–28.7 g fiber/d from each of 5 breakfast cereals: All-Bran (AB), Bran Buds with Corn (BBC), Bran Buds with Psyllium (BBP), BBC with viscous fiber blend (VFB), or a low-fiber control for 3 wk, with each study arm separated by a washout of 
1 wk. Seven-day stool collections and a symptom diary were obtained during the last week of each study arm.
Results: All study cereals induced significant (P < 0.05) increases in fecal bulk from the control diet at 128 ± 38 g to 199 ± 56, 199 ± 57, 247 ± 87, and 197 ± 63 g with consumption of AB, BBC, BBP, and VFB, respectively; less intestinal transit time; and significantly (P < 0.05) greater bowel movement frequency. Despite the increased activity of the bowel, a positive level of comfort was maintained. BBP was more effective than other cereals in terms of increasing fecal wet weight (P < 0.05).
Conclusion: Water-insoluble dietary fibers (ie, AB and BBC) and their mixtures with water-soluble fibers (ie, BBP and VFB) in the form of breakfast cereals (2.5 servings/d) proved to be a practical way of increasing fiber intake to recommended levels, while maintaining a good level of tolerance.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Despite the broad range of potential health applications attributed to dietary fiber, including the treatment of colonic disorders (5) and the association of dietary fiber with a lower risk of heart disease (6, 7), diabetes (8), and colon cancer (9), the ideal types of fiber and the proportions in which they should be fed are still not clearly defined. The water-insoluble forms of fiber, such as wheat bran, have been advocated for their effects in promoting colonic mucosal health by causing various changes in bowel function, such as increasing fecal bulking or reducing transit time through the gut. Conversely, the water-soluble or viscous dietary fiber is commonly believed to be more influential at a metabolic level, by decreasing the blood cholesterol or glucose concentration. Thus far, in terms of bowel function and side effects, abundant information has been gathered for water-insoluble fibers, mainly wheat bran, but little information exists for the water-soluble forms. It was shown in the past few years that water-soluble fibers (glucomannan and psyllium) also function extensively in the large colon to promote laxative effects and consequently to promote metabolic benefits (10, 11). Most previous studies focused primarily on one or the other form of fiber (12, 13, 14), and thus we felt that different types of fiber in a form in which most fiber is consumed warranted further assessment. Moreover, an increase in dietary fiber intake to the recommended level, which averages more than double the amount currently consumed, is of concern to most people because of negative gastrointestinal effects, such as flatulence, abdominal discomfort, and diarrhea.
The present study investigated a combination of water-insoluble (wheat and corn) and -soluble [psyllium and a viscous fiber blend (VFB) containing glucomannan and xanthan fibers consumed in breakfast cereal with respect to the fibers' ability to induce changes in bowel habits, including fecal bulk, defecation frequency, and transit time and to the gastrointestinal side effects measured by a symptom diary in healthy persons consuming a typical Canadian or US diet. The test fibers in this study were provided in the recommended amounts for the general population and combined in the form of different breakfast cereals; breakfast cereals are a source of dietary fiber whose health properties and practicality have been well established for many decades.
|
SUBJECTS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The inclusion criteria were self-reported 3 bowel movements per week with wet fecal output >50 g/d. Subjects were advised to eat a typical Canadian or US diet. Persons following a special diet, such as a vegetarian diet, and those with gastrointestinal disorders, bowel abnormalities, or a major illness were excluded, as were subjects using laxatives, diuretics, antidepressants, codeine, or antibiotics. Subjects were advised to consume >50% of any of the study supplements; changes of body weight of no more than ±2 kg/mo during the study were requested.
Assuming that the average wet fecal output among healthy subjects consuming the low-fiber cereals is on the order of 120 g/d (15, 16) and that the addition of 1 g wheat bran/d to the control diet increases fecal output by an average of 2.5 to 3.0 g/d (16), we calculated that the supplementation of 20 g of a test fiber would increase fecal output from 120 g/d to 170 g/d. Assuming an SD of 70 g/d (16), a sample size of 16 subjects is needed to ensure an 80% chance of correctly identifying a difference between the test and control means at an 0.05 level of significance. We further revised this sample size upwardly to improve the power of the study.
Written informed consent was obtained from all participants. The study was approved by the University of Toronto Committee for the Use of Human Subjects in Research and was carried out at the University of Toronto and St Michael's Hospital.
Experimental design
A 5-phase randomized crossover design was used, in which all subjects took each treatment and acted as his or her own control. Each treatment phase was 21 d in duration and was followed by a washout phase of 7 d. The first 10 d of each treatment phase, designated as the habitual period, consisted of the consumption of a test cereal in conjunction with the participant's typical diet; during the latter 11 d, or metabolic period, each participant continued to consume the test cereal, but a metabolic diet was supplemented to replace his or her habitual food.
Dietary record and symptom diary
Before the experimental portion of the study, a screening period of 7 d was designed to establish the subjects' baseline to ensure that candidates met all inclusion criteria. During this period, dietary records and symptom diaries of all participants were monitored. These records involved descriptions of the types and quantities of foods consumed by the participants. A weighing scale was provided to the subjects to aid in the measurement of food quantity. Participants were not asked to discontinue including any foods in their diet. During each of the experimental phases, 3 assigned days of food records were collected during the habitual period. Dietary intake for both the 7-d baseline and 3-d habitual periods were analyzed by using the University of Toronto Nutriput Diet Analysis Program based on the US Department of Agriculture Handbook # 8 Series for energy and nutrient information (Internet: http://www.nal.usda.gov/fnic/foodcomp/search/). Missing fiber values were added by using data from the Minnesota Database (Internet: http://www.ncc.umn.edu/products/database.html). When a difference in carbohydrate was observed between the 2 databases, a ratio was used to adjust the fiber value. The form and duration of habitual physical activity were recorded by the dietitian. If a participant regularly performed exercise, he or she was instructed to maintain the pattern for the entire study. Furthermore, any symptoms related to bowel habits, unusual sensations, or both were also noted. Symptom diaries used a 5-point scale and included flatus (1, no gas; 5, extreme flatulence), bloating (1, no bloating; 5, extreme bloating), stool consistency (1, watery; 5, hard), ease of movement (1, easy; 5, hard), and abdominal pain (1, no pain; 5, extreme pain). The scale was designed to describe the qualitative observations of the participants regarding their gastrointestinal and bowel sensations. Any unusual stresses were noted, because psychological and emotional factors may affect defecation. Body weight was measured at each visit to the research center.
Diet and test cereals
The study involved the investigation of 4 different types of fibers given in the form of breakfast cereals. The test fibers included 4 high-fiber cereals—All-Bran (AB), Bran Buds with Corn (BBC), and Bran Buds with Psyllium (BBP) (all: Kellogg Company, Battle Creek, MI), and BBC plus a VFB—and a low-fiber control. VFB is a proprietary fiber blend (patented combination, US patent 7 326 404; Internet:http://www.uspto.gov/web/patents/patog/week06/OG/html/1327-1/US07326404-20080205.html) with a characteristic, highly viscous polysaccharide. Viscosity of the VFB was estimated to be 700 poise, as measured with a viscometer (RVT; DW Brookfield Ltd, Cooksville, Canada) at a concentration of 1% and a shear rate of 1/30 s with a spindle type F. The total fiber content in VFB is 95% (70% glucomannan and 30% xanthan), and 98% of the fiber is water soluble.
The content (%) of raw wheat bran, corn bran, psyllium, and VFB in the cereals is shown in Table 1
. All cereals were prepared with the use of the same milling process, and they did not differ significantly in particle size. The cereals were provided in portions of 58–87 g (
2.5 servings/d) depending on the fiber content. The aim was to provide 25.0–28.7 g total dietary fiber (TDF)/d from cereals in addition to 12 g fiber/d from the metabolic diet, proportional to caloric intake.
|
The metabolic diet was typical of a Canadian or US diet in terms of both nutrient profile and types of food provided; it was similar in energy level and included a 3-d rotation of foods. The target nutrient profile for the metabolic diet was 15% protein, 35% fat, and 50% carbohydrates. Foods containing seed or thick skins were minimized to avoid obscuring the transit time measurements. Energy levels were individualized according to caloric intake (2000, 2500, 3000, or 3500 kcal) to promote weight maintenance and, subsequently, adherence to the diet. Because subjects varied in ethnic background, individual food preferences were accommodated where possible. Beverages such as coffee, tea, and alcohol were allowed in moderation, but participants were asked to set patterns for the entire study period and were told of the importance of not deviating from this pattern.
During all but one metabolic period, each subject consumed a constant level of TDF that was proportional to his or her caloric intake; in the control phase, a low-fiber cereal was provided. TDF was provided on an energy-density basis, at a level of 1.4 g TDF/100 kcal, as recommended by the major health agencies.
Sample collection and analysis
A few days before each stool collection period (the last 7 consecutive days of the metabolic period), waterproof plastic bags and personalized labels with space for marking the time and date were provided to each subject along with a stool collection kit. Subjects were also asked to note in their symptom diaries the time of each bowel movement. From the recording and the physical collection of the fecal samples, the bowel habits of the participants including stool frequency and the interval between defecations, were determined. Each fecal specimen was collected separately and immediately sealed and frozen by placing it in dry ice provided in insulated containers. Stool samples were transported to the laboratory on the same day and stored at –20 °C. Frozen stool samples were weighed within 48 h after transportation to the laboratory. Each sample was individually weighed on a digital electronic scale (PJ6000; Mettiler-Toledo Inc, Columbus, OH) that was accurate to 0.1 g.
Fecal dry matter content was established after the stools were freeze-dried with the use of the Labconco system (Labconco Corp, Kansas City, MO) for 48 h. The Mettiler digital scale was used to determine the weight before and after freeze-drying.
Determination of breath hydrogen (H2) and methane (CH4) gases was conducted on day 21 of each experimental phase. The breath sample collection was to begin before breakfast and then was conducted hourly for 12 h. Breath samples were collected in Haldane-Priestley tubes and stored at –20 °C immediately after collection. Within 2 d, the breath samples were analyzed by gas chromatography (model DP; Quintron Microlyzer, Milwaukee, WI). A standard of 105:75 ppm (hydrogen/methane) (Linde Canada, Guelph, Canada) was used before and after the analysis of each breath sample.
To determine intestinal transit time, radiopaque barium–impregnated polyethylene pellets (Sitzmarks; Konsyl Pharmaceuticals Inc, Fort Worth, TX) encased in gelatin capsules (24 pellets/capsule) were assigned to the subjects. Participants consumed the capsules at breakfast on the first, second, and third days of the 7-d stool collection period. All fecal samples were X-rayed to ascertain the passing of the pellets. Transit time was calculated as the mean time needed for the first 80% of the pellets to be passed in the stool (17).
Statistical analysis
The results are presented as means ± SDs for parametric data and as medians (interquartile ranges) for non-Gaussian distributions. The primary outcome measured was the average daily fecal wet weight expressed as a total fecal wet weight divided by the number of days of collection. Daily fecal dry weight, intestinal transit time, and frequency of defecation were reported and evaluated as secondary variables, which acted as validity checks. Breath gas analyses also were conducted. Parametric analyses were conducted after a comparison of the sampling distribution to a normal distribution (Shapiro-Wilk and Kolmogorov-Smirnov tests). The tests did not reject the null hypothesis at 
= 0.05. Statistical analysis was undertaken with SAS (version 8.02; SAS Institute Inc, Cary, NC) and SPSS (version 15.0; SPSS Inc, Chicago, IL) software. The mixed model adjusted for age, sex, and caloric intake, and the post hoc Tukey-Kramer comparison test was used to compare different diets. All comparisons were paired, and thus each subject served as his or her own control. For nonnormally distributed data, a global comparison was performed by using the nonparametric Friedman test, taking repetition into account, and with, if the test was significant, subsequent paired comparisons with the control diet by using Wilcoxon's signed-ranks test. Data were considered significant at P < 0.05.
Comparisons between the baseline and control diets were performed by using the paired t test for parametric distributions and Wilcoxon's signed-ranks test for non-Gaussian distributions.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
97% of food and 89–97% of cereals delivered were consumed. The composition of the experimental cereals and the amounts of cereals consumed are shown in Table 1
. Because the cereals were matched for quantity of TDF, significant (P < 0.05) between-cereal differences were observed in the amount of cereal consumed and the energy and macronutrient intakes. However, these differences were compensated for by the metabolic diet, and daily energy and macronutrient intakes did not differ significantly, except for a lower amount of TDF from the control diet (Table 2).
|
35% less VFB than psyllium fiber was VFB's very high level of viscosity, which would cause gelled clumps that make cereals less palatable. However, despite the fact that a smaller quantity of VFB was given, subjects still reported that some of the VFB, as much as 10%, remained on the side of the bowl.
The control diet did not differ significantly from each participant's habitual diet in terms of fecal bulking and other variables (Table 3). This finding indicated that the addition of a control cereal to the diet was not a confounding factor and that any changes observed among the high-fiber cereal phases are most likely the effect of the fiber content in AB, BBC, BBP and VFB.
|
Fecal weight and intestinal transit time
Based on a 7-d fecal collection period, the mean fecal output of the control diet was 128 ± 38 g/d (Table 3). Test cereals resulted in a fecal weight increase of 71 g for AB and BBC, 69 g for VFB, and 119 g for BBP (Table 3). When fecal output was expressed as 1 g stool per 1 g dietary fiber added (increment above control), tested fibers gave an increase of 2.8 ± 1.8, 3.0 ± 2.1, 4.8 ± 2.7, and 2.9 ± 2.1 g in wet stool for AB, BBC, BBP, and VFB, respectively. The stool per each 1-g increase in fiber for BBP was significantly larger, by 60–71%, than that for other cereals. All test cereals produced a significantly higher fecal dry weight than did the low-fiber control; there were no significant differences among the 4 treatments. The percentage of dry matter did not differ significantly among the treatments—exception BBP, which was significantly lower than BBC (Table 3). Furthermore, BBC had the slowest transit time of all of the experimental cereals, although the difference was not statistically significant. In each test cereal phase, there were 4 energy intake levels. When fecal weights were adjusted for energy intake, the between-diet differences remained similar to those with unadjusted data (Table 3). When male-female differences in fecal output were examined, the output for males was consistently larger than that for females, because of their different caloric intakes. There were no sex differences in daily fecal weights after adjustment for caloric intake (P > 0.05).
AB and VFB cereals were able to shorten transit time by 12 h below control, and BBP was almost as effective, shortening transit time by 10 h (Table 4). Furthermore, the number of stools per day was significantly higher with all test fibers than with the control cereal (Table 4).
|
6 ppm) increased the median breath H2 above control concentrations, whereas BBP did not cause an elevation of breath H2 (Table 3).
Nonparametric analysis of methane gas data included all subjects, whether they were methane producers or not. The median breath CH4 concentration was significantly lower with the AB and VFB test cereals than with the control cereal (Table 3).
Bowel habits and symptoms
Despite the increased defecation frequency, larger stool size, and greater hydrogen production, a positive level of comfort was maintained in terms of the bowel movement symptoms, including ease of movement, stool consistency, and abdominal pain, which were similar to the control diet for all tested cereals (Table 4). Whereas stool consistency and ease of defecation remained primarily unchanged, a trend of softer stools and more rapid defecation was reported by some subjects. Although the symptoms of flatulence and bloating were somewhat increased, all of these variables were recorded at a level indicative of comfort and were in the lower range of the given scale.
Medical records and symptom diaries showed that 6 subjects had a feeling of urgency to defecate during the BBP study arm. This symptom may be the result of the significantly higher fecal weight and greater moisture concentration during that study arm (Table 3).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The average fecal output among healthy persons consuming the low-fiber cereals was 128 g/d, which is an average fecal weight in Canada or the United States (15, 16, 18). Undigested fiber, retention of water (13, 19), and the mechanical action of fiber (20) are responsible for much of the increase in fecal weight with high-fiber cereal consumption, as is illustrated by significantly greater increases in both wet and dry fecal weights with consumption of all test cereals than with consumption of the control cereal. BBC cereal, in which 22% of AB was replaced with corn bran, did not result in a change of fecal weight from AB, which suggests that the fecal bulking ability of corn bran matches that of AB.
The increase of 2.8 g stool/g additional dietary fiber in the control diet, as compared with AB, is consistent with findings in some other studies using wheat bran (12, 15, 16, 18). Cummings (13) included a table of the average increase in fecal output per 1 g fiber fed; for wheat, this was 5.4 g; for corn, it was 3.3 g; and for psyllium, it was 4.0 g. The lower value in the present study for AB (2.8 g/g fiber) could indicate an effect of processing on the bran, because many of the studies summarized by Cummings used unprocessed wheat bran. At the same time, the laxative effect of BBP was surprisingly high in the present study, with almost 5 g feces/g added dietary fiber. Because the fecal weight was significantly higher with psyllium than with other cereals and because the fecal dry weight was similar in all cereals, the increase in fecal wet mass with BBP is likely to be due to moisture. Psyllium has the ability to form a gel matrix that binds water on the external surfaces of fiber particles, thus increasing the moisture content of the stool (10).
Both insoluble and soluble types of fiber may be fermentable by intestinal microflora, but soluble fiber usually has a higher fermentability (13). An increase in fermentation was shown in our study by an increase in flatulence and bloating, indicating a possibly higher level of microflora activity that was illustrated by increases in breath hydrogen and decreases in breath methane with AB, BBC, and VFB. Whereas hydrogen adds to the colonic bulk (13), methane appears to slow intestinal transit by augmenting small-bowel contractile activity (21), and thus it provides an additional mechanism for the bulking effect. In addition, because a fiber-rich diet may lower the risk of colon cancer (4, 14, 22-24), lower CH4 concentrations could be associated with the prevention of bowel malignancy (25, 26).
With the low hydrogen and high methane values recorded in the BBP phase, it would appear that fermentation of psyllium is either minimal (if any occurs) or appreciably slow. It has been shown that a gel-forming fraction of psyllium escapes microbial fermentation (10). Wolever et al (27) also reported that psyllium had no effect on breath hydrogen, methane, or serum acetate concentrations. A slow production of H2 introduces a stable environment for methanogenic bacteria that may use hydrogen for the generation of methane as quickly as it appears (27).
As expected, we found a reduction in intestinal transit time with consumption of the cereals, although all of the transit measurements were well within the normal range (repeated transit measurements in a person under rigid diet control vary by 20% to 30%). All cereals produced a transit time that was 30% shorter, except for BBC, which reduced transit time by 20% compared with the control. It is interesting that, despite greater fecal weight after consumption of BBC, the transit time was minimally improved. The explanation could be the percentage of dry matter with BBC that was higher than that with other experimental cereals and the positive correlation between these 2 variables.
The greater cereal fiber content induced significant increases in daily bowel movement. It was previously reported that the frequency of defecation is correlated with the size of the stool (28), and the improvements observed in the ease of passage and stool consistency were most likely the result of a higher moisture content in the feces.
Viscous fibers are therapeutically useful in modifying postprandial hyperglycemia (29) and have been associated with alterations in serum cholesterol concentrations (30), prolonged gastric emptying, and slower transit time through the small intestine (31). It has been shown that physiologic efficacy of dietary fiber is the same whether fiber is incorporated in breakfast cereals or sprinkled on the cereal (32). Although many dietary fibers alter viscosity in the gastrointestinal tract (31), few data exist for a link between viscosity and laxation. Findings summarized by Cummings (13) and subsequently published research indicated that psyllium is a unique viscous fiber, and even its analysis as a soluble fiber requires special steps. Whereas the study of the effects of glucomannan on the large intestine (11) provided evidence that some fibers analyzed as a soluble may increase stool weight, many studies of other soluble fibers (eg, pectin or guar gum) have not supported a large-bowel effect. Therefore, the influence of fiber viscosity on laxation requires further study.
In conclusion, water-insoluble (AB and BBC) dietary fibers and their mixtures with water-soluble (BBP and VFB) fibers in the form of breakfast cereals (2.5 servings/d) proved to be a practical way of increasing fiber intake to recommended levels, and they induced an increase in fecal weight, decreased transit time, and improved bowel movement frequency with sustained gastrointestinal tolerance. Fiber can safely be consumed at twice the quantity of current intake with no change or little change in gastrointestinal symptoms. Combinations of insoluble fibers with psyllium have stronger bulking effects than do insoluble fibers alone.
|
ACKNOWLEDGMENTS |
|---|
The authors' contributions were as follows—VV: conceived of the project, wrote the protocol, obtained the funding, and contributed significantly to the manuscript; ALR: contributed to data analysis and manuscript preparation; DJAJ: contributed to the study design, data interpretation, and manuscript preparation; ALJ: contributed to preparation of the diet, diet analysis, and manuscript preparation; JLS: contributed to the study interpretation and manuscript preparation; and EJ: contributed to data sorting and analysis and manuscript preparation. None of the authors had a personal or financial conflict of interest.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
