点击显示 收起
1 From the Department of Gastroenterology, Monash University, Box Hill Hospital, Victoria, Australia (JGM); the Andrew Love Cancer Centre, The Geelong Hospital, Geelong, Victoria, Australia (EGWY); CSIRO Health Sciences and Nutrition, Adelaide, Australia (JK and ARB); the Department of Colorectal Medicine and Genetics, The Royal Melbourne Hospital, Victoria, Australia (CP and FAM); the Department of Mathematics & Statistics, University of Melbourne, Victoria, Australia (KS); and the Menzies School of Health Research, Casuarina, Northern Territory, Australia (KO).
2 Supported by grants from the National Health and Medical Research Council of Australia and Meat and Livestock Australia. Starch Australasia (Lane Cove, Australia) supplied the Hi-maize, waxy maize, and wheat bran and Meat and Livestock Australia donated meat. 3 Reprints not available. Address correspondence to J Muir, Department of Gastroenterology, Monash University, Box Hill Hospital, Level 8, Clive Ward Centre, Arnold St, Box Hill, Victoria, Australia, 3128. E-mail: jane.muir{at}med.monash.edu.au.
ABSTRACT
Background: Wheat bran (WB) increases fecal bulk and hastens colonic transit, whereas resistant starch (RS) has effects on colonic fermentation, including increasing concentrations of butyrate.
Objective: We hypothesized that a diet combining WB with RS would produce more favorable changes in fecal variables (eg, fecal bulk, rapid transit time, lower pH, and higher butyrate) than would WB alone.
Design: This was a randomized crossover block-design study for which 20 volunteers with a family history of colorectal cancer were recruited. The study included 3 diets: control, WB (12 g fiber/d), and WBRS (12 g WB fiber/d plus 22 g RS/d), each continued for 3 wk. In each diet, the major source of protein was lean red meat. During 5 consecutive days (days 15–19) of each dietary period, the subjects collected their total fecal output for analysis.
Results: The WB diet resulted in greater fecal output (by 23% and 21% for wet and dry weights, respectively) and a lesser transit time (–11 h) than did the control diet but did not have major effects on fermentation variables. Compared with the control diet, the WBRS diet resulted in greater fecal output (by 56%) and a shorter transit time (–10 h), lower fecal pH (–0.15 units), higher fecal concentration (by 14%) and daily excretion (by 101%) of acetate, higher fecal concentration (by 79%) and daily excretion (by 162%) of butyrate, a higher fecal ratio of butyrate to total short-chain fatty acids (by 45%), and lower concentrations of total phenols (–34%) and ammonia (–27%).
Conclusions: Combining WB with RS had more benefits than did WB alone. This finding may have important implications for the dietary modulation of luminal contents, especially in the distal colon (the most common site of tumor formation).
Key Words: Wheat bran • resistant starch • fecal bulk • colonic transit • luminal butyrate • colonic fermentation
INTRODUCTION
The effects of wheat bran (WB) on reducing transit time, fecal bulking, and dilution of potentially carcinogenic compounds are well established (1, 2). WB has been shown to be effective at protecting against tumor formation in animal models (3, 4). Intervention trials monitoring the occurrence of new colorectal adenomas in human volunteers, however, have been less conclusive (5), with the Australian Polyp Prevention Project being the most convincing positive trial to date (6). Fermentation of WB by colonic microflora produces short-chain fatty acids (SCFAs), including acetate, propionate, and butyrate (7), which are rapidly absorbed by the colonic mucosa (7). Butyrate, the preferred energy source of the colonocyte (8), has a range of effects relevant to reducing colorectal cancer risk (7, 9–12). In contrast, when protein is fermented in the colon, the byproducts ammonia and phenols (p-cresol and phenol) may be deleterious (13). These tend to accumulate when protein is fermented in a carbohydrate-deficient environment. WB can dilute the concentration of these byproducts (14).
Resistant starch (RS) is another carbohydrate to reach the colon undigested (15). The 3 main types of RS are RS1 (physically inaccessible starch), RS2 (ungelatinized starch granules), and RS3 (retrograded starch polymers) (15). The major physiologic effects of RS on the colon of humans include increased concentrations of SCFAs (16–19), lowered pH (17, 19), and lowered concentrations of ammonia, phenols, and secondary bile acids (20–22). RS, however, has little (19, 23) or only modest effects on fecal bulking (16–18, 21) and tends to delay rather than hasten intestinal transit time (18, 23). Particular interest has been focused on the observation that fermentation of RS produces more butyrate than does fermentation of nonstarch polysaccharides, including cereal brans (19, 24). Yet not all reported effects of RS are positive. For example, RS (from raw potato starch) was found to cause colonocyte hyperproliferation and increased tumor formation in animal models (25). Moreover, a recent human study found that a high-RS diet [based on the high amylose maize starch Hylon V11 (National Starch Chemical Company, Bridgewater, NJ)] increased the formation of DNA adducts (26). These adverse outcomes may be prevented by using a mix of dietary fiber types. We have shown that combining WB with RS from potato starch attenuates the RS-induced increase in tumor formation in rats (25). We suggested that this effect may relate to the extent and location of fermentation events in the colon (25, 27, 28).
In humans, most tumors occur in the distal portion of the colon. To reduce the incidence of colorectal cancer, it may be important to focus on fermentation-dependent events, including butyrate production, at this site. Earlier work in pigs (29) found that the addition of WB to the diet can shift the site of RS fermentation more distally, thereby improving luminal conditions, including higher butyrate and lower ammonia concentrations. In the present study, we studied a WB plus RS diet in human volunteers with a family history of colorectal cancer. Red meat provided the major source of dietary protein. We hypothesized that combining WB with RS would produce more favorable changes in the fecal variables (eg, bulking, rapid transit, lower pH, higher butyrate, lower phenols, and lower ammonia) than would WB alone.
SUBJECTS AND METHODS
Subjects
Twenty healthy volunteers (9 women, 11 men) aged 42 ± 2 y (
This study was conducted over an 11-wk period (including 3 x 3 wk of test diets and 2 x 1-wk washout periods between each 3-wk dietary period). The 3 test diets were as follows: control (C: low RS, no added WB), WB (low RS, added WB), and WB plus RS (WBRS: WB plus moderate RS). The subjects were randomly assigned to 6 possible diet orders, with as close to a balanced design as possible. The subjects consumed each test diet for 3 wk, followed by a 1-wk break before crossing over to the next diet, until all 3 test diets had been consumed. At the end of each dietary period, the subjects were weighed and completed a questionnaire to assess gastrointestinal symptoms relating to each diet. During the study, no more than 4 standard drinks of alcohol per week were allowed.
The subjects completed food and drink diaries throughout the entire study. Dietary records were analyzed by using the FOODWORKS database (Xyris Software, Highgate Hill, Queensland, Australia), which is based on the Australian food-composition tables.
Experimental diets
During the test dietary periods, macronutrient intake was kept constant while the amounts of WB and RS were varied by the use of specially prepared foods. Background intakes of dietary fiber and RS were kept low throughout the study. Intake of WB and RS was based on energy intake, with the men consuming more WB- and RS-containing foods than did the women.
Three sets of foods were prepared, and the composition is given in Table 1. During the C diet, the subjects were supplied with cornbread, muffins, cakes, cereals, and desserts containing finely ground (<1 mm) low-amylose maize and no added WB. For the C diet, RS intakes ranged from 2 to 5 g/d. During the WB diet, the same foods were supplemented with unprocessed WB, which provided 10-15 g fiber/d (particle size: 1-3 mm); RS intakes ranged from 2 to 5 g/d. During the WBRS diet, the same foods were prepared with a high-amylose (85% amylose) maize [Hi-maize; Starch Australasia Ltd, Lane Cove, Australia (currently National Starch and Chemical Co, Bridgewater, NJ)]) that had been ground coarsely (particles ranging from <1 to 3 mm). The WBRS diet provided 20-30 g RS/d in addition to the 10-15 g fiber/d from WB. Preparation of similar low- and high-RS dietary foods have been described in detail previously (17). The amounts of RS in cooked ready-to-eat foods were measured with an in vitro assay developed in this laboratory and validated previously in an ileostomy model (30). In addition to these dietary foods, the subjects were supplied with 250 g raw lean red meat (5-7% fat) and requisite cooking instructions or were given frozen preprepared dishes containing this meat (150-250 g/d).
View this table:
TABLE 1. Composition of foods supplied to the volunteers during the control (C), wheat bran (WB), and WB plus resistant starch (WBRS) dietary periods1
The meat dishes included meat loaf, lasagna (made with bolognaise sauce), and lamb stew. A sample menu, which also includes the quantities of RS and WB in the supplied dietary foods, is shown in Table 2.
View this table:
TABLE 2. Sample meal plan for one day of the control (C), wheat bran (WB), and WB plus resistant starch (WBRS) diets
Fecal collections and analysis
During the third week of each dietary period, total feces were collected for 5 consecutive days and placed immediately onto dry ice (23). Mouth to anus transit time was determined as described previously (23). Briefly, early on day 14 of the study, each subject (the exact time and date was recorded) took a gelatin capsule containing 24 radiopaque rings (Sitzmarks; Konsyl Pharmecuiticals Inc, Fort Worth, TX). The collected frozen fecal samples were X-rayed and radiopaque rings were counted. For 2 subjects, <10 rings were recovered, and the transit time results for these subjects were not included in the final data set. For the remaining subjects, recovery of the rings was 98%.
Feces were analyzed for pH, SCFA, ammonia, and starch by established procedures (23). Fecal samples for each subject were thawed overnight at 4°C and were then weighed, pooled, and homogenized on ice before further storage in aliquots at –45°C. Samples of pooled fecal homogenate in sealed tubes were brought to 37°C before fecal pH was determined. An average of 3 readings was taken. Samples were processed quickly to minimize exposure to air. SCFA content was analyzed by gas chromatography. Fecal ammonia was estimated colorimetrically. Aliquots of fecal homogenate were freeze-dried to a constant weight for estimation of fecal dry weight. Fecal starch content was determined by using a commercially available kit (Megazyme Australia, Warriewood, Sydney, Australia) and involved first solubilizing the starch by boiling in dimethyl sulfoxide followed by enzymatic hydrolysis to quantitatively release glucose. This assay accurately quantifies starch in both fecal and food samples. The content of fecal phenols (ie, phenol and p-cresol) was determined by HPLC by the method described previously (31), except that a vacuum distillation procedure and ultraviolet light detection (270 nm) were used. Fecal and not urinary phenols were measured here, because an earlier study (20) showed that fecal phenols are more responsive to RS-induced change.
Statistical analyses
Each subject undertook each of the 3 diets at different time periods. Comparisons made between diets are therefore within subjects with allowance made for possible period effects. Two analyses were carried out for each outcome variable: one using diet as a nominal factor with 3 levels (an intention-to-treat analysis) and one using the actual amounts of WB and RS consumed, as derived from each subject’s food diary. The intention-to-treat analyses also considered possible carryover effects between study periods and sex-by-diet interactions; neither of these effects was found to be significant for any of the outcome variables, and they are excluded from the analyses reported below. Some of the outcome variables were transformed by using either a square-root or log transformation to allow for heterogeneity in the variability of the raw data. Repeated-measures analysis of variance followed by Tukey’s multiple comparisons was used throughout, with diet as a factor for some analyses and the actual amounts of WB and RS as covariates in others. In most cases, the analyses that used the actual amounts of WB and RS consumed merely confirmed the results of the intention-to-treat analyses, and thus only limited details are reported. For many of the outcome variables, usually one, but up to 3, outliers were identified. However, the conclusions were largely unaffected by the outliers and the analyses reported here are those without any of the outliers omitted. Data were analyzed with the SPLUS (version 6) statistical package [Lucent Technology (formerly Bell Laboratories), Murray Hill, NJ].
RESULTS
Dietary intake
The subjects consumed similar amounts of energy and macronutrients during all 3 dietary periods (Table 3). Throughout the study, red meat contributed >50% of dietary protein; the men consumed
View this table:
TABLE 3. Daily composition of the control (C), wheat bran (WB), and WB plus resistant starch (WBRS) diets1
Gastrointestinal symptoms and weight maintenance
During the C diet, some subjects experienced difficulties with constipation and were advised to ensure that they consumed all of their fruit, vegetable, and water allocation. Otherwise, the only gastrointestinal symptom reported during the study was moderate flatulence (median value of 6 on a symptom scale from 0 to 10), which was reported during the WBRS diet (data not shown). The mean body weight of the subjects at the end of each dietary period remained stable.
Changes in fecal variables
Less than 7% of the RS supplied in the diet could be recovered in the feces. Eighteen of the 20 subjects excreted more starch in feces during the WBRS diet than during the C and WB periods, suggesting good adherence to the dietary prescription (Table 4).
View this table:
TABLE 4. Fecal variables during the control (C), wheat bran (WB), and WB plus resistant starch (WBRS) diets1
Fecal output, both wet and dry, was significantly higher during both the WB and the WBRS diets than during the C diet (Table 4). Transit time was significantly shorter, by 11 and 10 h, respectively, with the WB and WBRS diets than with the C diet. After allowance for the amount of added WB in the diet, RS did not further reduce transit time (P = 0.79). Fecal pH was significantly reduced by RS (by 0.15 units/10 g RS), but not by the amount of WB (P = 0.51).
The median fecal concentration of total phenols (phenol, p-cresol) and ammonia was significantly lower after the WBRS diet than after the C or WB diet (Table 5). The effect of diet on the concentration of phenols and ammonia may be accounted for by the increase in fecal output during the WBRS diet.
View this table:
TABLE 5. Fecal concentrations and daily excretion of phenol and ammonia during the control (C), wheat bran (WB), and WB plus resistant starch (WBRS) diets1
Fecal short-chain fatty acids
Only the WBRS diet resulted in significant increases in the daily excretion of acetate (101%) and total SCFAs (92%) and also the fecal concentration of acetate (14%) and total SCFAs (22%) (Table 6). Compared with the C diet, daily excretions of propionate were significantly higher by 53% with the WBRS diet. These increases were offset by the increased fecal output so that, compared with the C diet, concentrations decreased (–17%) significantly with the WBRS diet. The WBRS diet also increased the daily excretion of isobutyrate (67%) and decreased the fecal concentration of valerate (–6%) and isovalerate (–32%).
View this table:
TABLE 6. Fecal concentrations and daily excretion of short-chain fatty acids during the control (C), wheat bran (WB), and WB plus resistant starch (WBRS) diets1
The daily excretion of butyrate and its concentration increased significantly with the WBRS diet (Table 6). They also increased with the WB diet, but the effect was not significant. Compared with the C diet, the amount of butyrate excreted during the WBRS diet was 162% higher and the concentration of butyrate excreted in feces was 79% higher.
Changes in SCFA excretion in feces were also expressed as a percentage of total SCFAs (Table 7). The relative proportions of the SCFAs changed significantly during the WBRS diet. The WBRS diet produced a marked and significant increase in the percentage of butyrate (45%) and a decrease in the percentage of propionate (–25%) produced. There was also a significant decrease in the proportion of iso-SCFAs, iso-butyrate (–18%), and iso-valerate (–33%) excreted during the WBRS diet.
View this table:
TABLE 7. Percentages of short-chain fatty acids excreted in feces during the control (C), wheat bran (WB), and WB plus resistant starch (WBRS) diets1
DISCUSSION
The results of the present study show that combining unprocessed WB (12 g dietary fiber/d) with RS (RS1 and RS2, 22 g RS/d) in the diet of human volunteers with a family history of colorectal cancer produced marked changes in all fecal variables measured. The combination of WB with a moderate dose of RS was well tolerated by the participants and did not cause excessive flatus. These changes were achieved within the context of a background diet that was high in lean red meat. In contrast, the WB diet increased fecal bulking and hastened intestinal transit time but did not produce major changes in fermentation-dependent variables.
We are confident that the observed effects of the WBRS diet were due to the combination of WB with RS and not to the RS component alone. A group consuming RS only was not included in the present study because earlier work (16–23) showed that RS alone does not produce the wide range of luminal changes considered essential for improving colonic health and reducing colorectal cancer risk. Even at high doses (30-48 g/d), RS has only modest bulking ability and tends to delay rather than hasten intestinal transit time (17, 18, 21). RS does, however, have a major effect on fermentation-dependent events such as increasing the fecal excretion of the SCFAs acetate and butyrate and lowering pH. Furthermore, a study by Noakes et al (19) that was conducted in human volunteers used a similar quantity and source of RS as in the current study. Those authors found that this moderate amount of RS had no effect on fecal bulking but did lower fecal pH. The RS also increased the concentration of fecal butyrate without significantly changing the excretion of acetate or the molar ratio of butyrate to total SCFAs excreted. The molar ratio is a more reliable indicator of dietary induced change because >95% of the SCFAs produced are rapidly absorbed by the colon (16, 19). In the present study, the WBRS diet significantly increased the percentage of butyrate to total SCFAs, whereas the percentage of propionate to total SCFAs decreased, a finding that suggests a major shift in the products of fermentation. This may have implications for colorectal cancer risk, because a higher ratio of acetate to total SCFAs and a lower ratio of butyrate to total SCFAs has been reported in enema samples from patients with adenomatous polyps and colon cancer than in samples from healthy control subjects (32).
The present study extends earlier studies conducted by our group in pigs (29) and rats (25). In pigs, 4 diets (control, RS, WB, and WBRS) were examined. The RS (Hi-maize) was shown to be fermented in the proximal region of the colon, whereas addition of WB moved the site of fermentation further down the colon, improving conditions in the distal lumen by increasing concentrations of butyrate and lowering concentrations of ammonia (29). Tumor formation was studied in rats after a 10-wk course of dimethylhydrazine (25). The RS diet (raw potato starch) increased the number and size of tumors (25). Importantly, this negative effect of RS was reversed when the RS was combined with WB (25). Similar results were found for the highly fermentable fibers guar gum and pectin (33). The finding that diets high in RS without other fiber may not be protective against tumor formation is also borne out by a recent study by Wacker et al (26). Human volunteers were fed a diet containing RS from high-amylose maize (Hylon VII) and effects on lipid-peroxidation-induced DNA damage and proliferation in the colon were assessed. A significant increase in DNA adduct formation was evident during the high-RS period compared with the low-RS period (26). Formation of DNA adducts is thought to be important in the process of carcinogenesis. Precisely how intake of RS in the context of a fiber-poor diet relates to risk of colorectal cancer in humans remains unknown. Combining different fiber types (particularly more slowly fermented fibers with the rapidly fermented fibers) may be required for maximum benefits.
Red meat has been implicated as a risk factor for colorectal cancer in some (34) but not all (35) population studies. In the present study, we chose a background diet in which lean red meat contributed >50% of daily protein intake. When protein reaches the colon undigested, it can be processed by the microflora to potentially deleterious compounds, including ammonia and phenols (p-cresol, phenol) (13, 36-38). These products tend to accumulate in a carbohydrate-deficient environment. We showed previously that the fermentation of RS is effective at reducing fecal phenols and ammonia in humans (20). In the present study, the combination of WB and RS was effective at lowering concentrations of phenols and ammonia, principally through dilution effects from the increased fecal bulk. Fecal iso-SCFAs can be used as markers for amino acid fermentation in the colon (7, 14). In the present study, there was a significant decrease in the proportion of iso-SCFAs excreted during the WBRS diet, which is consistent with a decrease in colonic protein fermentation (7, 14).
The effects of WB on fecal variables are well reported (1, 2). The quantity of WB used here was similar (12 g fiber/d) to that used in the Australian Polyp Prevention Project (6). Effects on fecal bulking in the current study were less than anticipated, which may have been because the higher background amount of dietary fiber in the C diet masked some of the WB effects. Nevertheless, the WB diet did have a marked effect on hastening intestinal transit time (by 11 h), which was not enhanced further by the addition of RS. Bulking and faster transit have been suggested as 2 major protective characteristics of fiber against colorectal cancer (39). In contrast with the results of many animal studies, WB in the present study did not have major effects on the SCFA profile, especially butyrate (3, 27, 28). Other studies using higher doses of WB (18 g nonstarch polysaccharides/d) in humans, however, produced results similar to those observed here (18).
The increase in SCFAs during the WBRS diet resulted in more acidic feces (7). Lower luminal pH affects several processes relevant to reducing colorectal cancer risk (40, 41). Production of butyrate was also enhanced. Butyrate has been reported to have several important effects relevant to normal cellular function and protection against malignant transformation. For example, butyrate stabilizes DNA, induces differentiation (9, 10), reduces the growth rate of mammalian colorectal cell lines (10, 11), and induces apoptosis (12). RS has been shown, both in vitro and in animal studies, to be a better substrate for the production of butyrate than other nonstarch polysaccharide sources (7, 19, 24). Not all published studies of the effects of RS in humans, however, have shown this effect on butyrate production (18, 21). This variability may be due to the type and quantity of RS used or the quantity and nature of the background dietary fiber. These factors in turn may affect the location of RS fermentation along the bowel and whether it persists to the distal end.
In humans, the distal colon is the most common site of tumor formation (42); consequently, the site of fermentation may have important implications for colorectal cancer risk. Stable production of butyrate along the entire length of the colon may be important for protection against malignant change (43). Studies in rats have shown that differences in regional patterns of fermentation have different effects on outcomes of tumor formation (27). Perrin et al (43) showed that fibers that produce a stable production of butyrate along the length of the colon were more effective at decreasing the rate of aberrant crypt foci formation in azoxymethane-treated rats. Furthermore, RS3 was more effective than starch-free WB (43).
The strategy of combining WB with RS was effective at producing beneficial changes in fecal bulking, transit time, and fermentation-dependent indexes that persisted to the distal regions of the human colon. The WBRS combination clearly produced more benefits than did the WB diet alone (as shown here) or the RS diet alone (as shown in previous studies). Moreover, combining WB with RS may attenuate any potentially negative effects of RS. This work suggests that to maximize the health benefits of dietary fiber in the gastrointestinal tract, a combination of different fiber types may be required.
ACKNOWLEDGMENTS
We thank the Kingston Centre for providing the cooking facilities, the volunteers who took part in this study, radiographers Cassie Cooper and Dana Jackson of the Radiology Department at Monash Medical Centre for radiography of fecal samples, and Debbie Davies and Corinna Flory for their assistance in performing phenol analyses. We also thank Karen Walker for reading the manuscript.
JGM, KO, and FAM designed the study and contributed to the interpretation of the results and preparation of the manuscript. JGM, EGWY, JK, and CP carried out the study. ARB carried out the analysis of the phenols and contributed to the interpretation of the results and preparation of the manuscript. KS conducted the statistical analysis of the data. JGM, EGWY, JK, CP, ARB, and KS had no financial or personal interests in the National Health and Medical Research Council or Meat and Livestock Australia. KO is a member of the Council of the National Health and Medical Research Council but had no role in funding decisions and had no affiliations with Meat and Livestock Australia. FAM participates as a member of the Scientific Advisory Board of Meat and Livestock Australia but had no role in funding decisions and had no affiliations with the National Health and Medical Research Council.
REFERENCES