Effect of dietary fiber on intestinal gas production and small bowel transit time in man13

Transcription

1 ffect of dietary fiber on intestinal gas production and small bowel transit time in man13 John H. Bond,4 M.D. and Michael D. Levitt,5 M.D. ABSTRACT The influence of dietary fiber on intestinal gas production and on small bowel transit time was studied in eight healthy subjects using breath H2 excretion as an indicator of colonic gas production. Hydrogen excretion following ingestion of bran was substantially less than that following ingestion of lactulose, a nonabsorbable fermentable sugar. Likewise, human fecal homogenates produced only about 1% as much H2 and CO2 during incubation with bran as with glucose or lactulose. Thus, the polysaccharides in bran appear to be relatively poor substrate for colonic bacterial gas production, and reported gas-related symptoms after bran ingestion may be due to some other mechanism. The small bowel transit time of bran was greater than that of lactulose; however, addition of bran to lactulose did not slow lactulose transit. Am. J. Clin. Nutr. 31: S 174, The symptomatic and physiological response of an individual to dietary fiber may be influenced by the bacterial metabolism of fiber in the colon. Most attempts to study the interaction of fiber with colomc bacteria have used in vitro culture techniques that may not accurately simulate the complex environmental conditions present in the colon. In the present investigation we studied the in situ degradation of ingested fiber indirectly by measuring the excretion of one product of this reaction, intestinal gas. Gas production by colomc bacteria The three major gases formed in the colon are hydrogen (H2), carbon dioxide (C2), and methane (CH4). All H2 produced in man is derived from bacterial metabolism (1). Although a variety of aerobic and anaerobic species of bacteria are capable of liberating H2, the organisms responsible for its production are unknown. We previously studied the site and rate of H2 production in the human intestine by perfusing the gut with nitrogen until a steady state was achieved, and then by analyzing gas samples obtained from the proximal ileum, terminal ileum, and rectum for H2 content (2). These studies showed that H2 is normally produced almost solely in the colon (small bowel production being negligible) and that production in the colon depends on the presence of exogenously supplied fermentable substrate, primarily carbohydrate. xcessive H2 production in the colon may occur in patients with malabsorption caused by failure to absorb normally the ingested, fermentable carbohydrate, which then passes down to the colonic bacteria. In addition, some vegetables and grains contain carbohydrate that cannot be digested and absorbed by the normal small intestine; hence these substances also serve as substrate for H2 production in the colon. Methane is also produced in man solely by colomc bacteria; however, its production does not appear to be affected by the presence or absence of unabsorbed, fermentable carbohydrate. Only about one-third of the adult population excretes appreciable CH4, and this tendency appears to be a familial trait (3). Carbon dioxide is derived in the upper bowel from the interaction of bicarbonate 1 From the Department of Medicine, Minneapolis Veterans Administration Hospital and the University of Minnesota Hospital, Minneapolis, Minnesota. 2 Supported by grants from the Veterans Administration and the United States Public Health Service Grant ROl AM Address reprint requests to: Dr. John H. Bond, Veterans Administration Hospital, 54th Street and 48th Avenue South, Minneapolis, Minnesota Chief, Gastroenterology Section, Veterans Administration Hospital. Staff physician in Gastroenterology Section, University of Minnesota Hospital. The American Journal of Clinical Nutrition 31: OCTOBR 1978, pp Printed in U.S.A. Sl69 Downloaded from

2 S17 BOND AND LVITT with hydrochloric acid and fatty acids released during digestion of fats. This CO2 is rapidly absorbed, and little reaches the colon. Like H2, CO2 appears to result in the colon from bacterial metabolism of unabsorbed substrates. Colonic CO2 may result from the bicarbonate neutralization of bacterially produced acids or as a direct bacterial metabolic product. Thus, either in malabsorptive disease or as the result of ingestion of foods containing unabsorbable plant polysaccharides, substrate is supplied to the colonic bacteria resuiting in increased production of two gases formed in the colon, H2 and CO2. A wellknown example is the effect of beans on intestinal gas. Steggerda observed that a diet comprising 56% baked beans increased the mean flatus excretion of five subjects from 15 to 176 ml/hour, and that this increase consisted primarily of H2 and CO2 (4). Likewise, lactase-deficient individuals who malabsorb the readily fermentable sugar, lactose, commonly complain of increased flatus, abdomma! cramping, and bloating when they drink milk (5). Since bran consists largely of nonabsorbable plant polysaccharides, one might also expect a high-bran diet to increase colonic gas production. Indeed, there are several reports of individuals developing what they believed to be gas-related abdominal symptoms shortly after they added bran to their diets (6). - ffect of bran on H2 production The effect of bran on intestinal H2 production was determined in eight healthy subjects by measuring their pulmonary excretion of H2 after ingestion of 1, 2, and 3 g of unprocessed bran (General Mills, Minneapohs, Minn.). This technique was based on the observation that the quantity of H2 produced in the colon as determined by intestinal perfusion is accurately reflected, minute to minute, by the rate of pulmonary H2 excretion (2). After ingestion of each test meal, all expired air was collected for several hours by having the subject breathe into a closed collection system from which gas samples were serially removed and analyzed for H2 concentration by gas chromatography. The collection system consists of a polyvinyl cylindrical hood that is placed over the subject s head and sealed at the neck with a soft rubber diaphragm. A pump circulates the gas from the hood through a spirometer, a CO2-absorbing canister, a cooling chamber, and back to the hood. Oxygen is added at the rate used by the subject. Figure 1 compares H2 excretion by one subject after ingesting 1 g of lactulose with H2 excretion after 1, 2, and 3 g of bran. While ingestion of lactulose was followed by a brisk rise in breath H2, no increase above base-line was observed after an equal quantity of bran. When the bran meal was increased to 2 and 3 g, an increase in H2 excretion above base-line did occur; however, even after the largest dose, the increase was appreciably less than that following only 1 g of lactulose. Similar results were obtained in all eight subjects studied, and the mean H2 excretion for this group is shown in Figure 2. Over a 2- hr period following the initial rise in breath H2 above base-line, less than half as much H2 was excreted after 3 g of bran as occurred after 1 g of lactulose. After 2 g of bran, H2 excretion was slightly lower than with 3 g in all subjects; no H2 above base-line was detected after 1 g. We considered two explanations for the low H2 excretion following bran ingestion. First, the bran may move so slowly through the stomach and small bowel that colonic H2 production was underestimated because the studies were not continued for a sufficient time. To rule out this possibility, we monitored breath H2 excretion of four subjects for 8 hr after ingestion of 3 g of bran, and no appreciable later H2 production was detected. Second, the colonic bacteria may be less efficient in digesting the polysaccharides in bran, and less gas is therefore produced. To test more directly the ability of colonic bacteria to form gas from bran, we collected freshly passed feces from four subjects and incubated homogenized samples in phosphate buffer at 37 C with equal quantities of lactulose, glucose, and bran. As shown in Figure 3, similar volumes of H2 resulted from addition of glucose and lactulose to the fecal homogenates; however, the volume of H2 resulting from incubation with bran was less Downloaded from

3 GAS PRODUCTION AND TRANSIT TIM log.1.2 2g -.1 C.Q 8) U C ).2 3g (sj I a > 2.4 log Lactulose.2 I Time (hours) FIG. 1. Pulmonary H2 excretion by a normal subject after ingestion of 1 g of lactulose, and 1, 2, and 3 g of bran. than 1% of that produced from the two sugars. Fecal samples collected from four subjects who regularly consumed bran were similarly tested to determine if the ability of colonic bacteria to produce H2 from bran polysaccharides increases as the result of chronic exposure. H2 production after bran again remained low (Fig. 3), indicating that no bacterial adaptation with regard to gas formation had occurred. Figure 4 shows the effect of lactulose, glucose, and bran on production of C2, the other major gas produced from dietary carbohydrate by the colonic bacteria. As was the case with H2, only about 1% as much CO2 was produced during incubation with bran as with glucose and lactulose. In all incubation studies, H2 comprised about one-third and CO2 two-thirds of the total gas evolved. All subjects participating in these studies were carefully questioned about symptoms resulting from ingestion of the different test mixtures. Supporting the fmdings of low gas 2 I LoctutOSe FIG. 2. Pulmonary H2 excretion (mean ± 1 SM) by eight normal subjects after ingestion of 1 g of lactulose, and 1, 2. and 3 g of bran. production from bran was the observation that symptoms of abdominal distention, cramping, and increased flatus were far more frequent and severe following lactulose ingestion. Downloaded from

4 S172 BOND AND LVITT I FIG. 3. Hydrogen production by fecal homogenates collected from four subjects who do not ingest bran and four subjects who ingest bran daily, following incubation with glucose, lactulose, and bran. Data are expressed as the percent H2 produced from glucose and bran compared to that produced from lactulose (mean ± 1 SM). S. (.2 Loctuose Glucose LOCtUIOSe Glucose FIG. 4. Carbon dioxide production by fecal homogenates collected from four subjects who do not ingest bran and four subjects who ingest bran daily, following incubation with glucose, lactulose, and bran. Data are expressed as in Figure 3. ffect of bran on small intestinal transit time Several studies have shown that bran decreases both colonic and mouth-to-anus transit time (7, 8). Measurement of breath H2 excretion appears to provide a unique method to assess the influence of bran on transit time through the small bowel. Since a rise in breath H2 occurs within a few minutes after an ingested, unabsorbed carbohydrate reaches the colon, the time interval between ingestion and detection of the increase represents the transit time of that material to the cecum. This technique for measuring small bowel transit time was recently validated in studies comparing breath H2 measurements with direct measurements of transit time employing aspiration from the terminal ileum (9). Table 1 compares the small bowel transit time of bran with that of lactulose. In eight subjects, lactulose reached the colon in a mean of 48.2 ± 6 (SM) miii compared with 72.8 ± 8 nun for the 2-g dose ofbran and 64 ± 6 mm for the 3-g dose. Since no H2 was detected after ingestion of 1 g of bran, its transit time could not be measured with this method. Downloaded from

5 TABL 1 Small bowel transit time of lactulose and bran determined by pulmonary H2 excretion in eight normal subjects (mean ± 1 SM) Thus, this relatively large dose of bran moves somewhat more slowly through the small bowel than does a smaller quantity of lactulose. To determine if bran, when added to a test meal such as lactulose, would retard the rate of transit of the lactulose through the small bowel, we repeated studies using a mixture of 1 g of lactulose plus 2 g of bran. Figure 5 shows that adding bran to lactulose did not significantly alter the rate the lactulose moved through the small bowel. Transit time of lactulose with and without bran av- GAS PRODUCTION AND TRANSIT TIM S173 eraged 52 ± 8 (SM) and 58 ± 14 min, respectively. Conclusions C I-. I Test meal Small transit bowel - - time Lactulose, 1 g 48.2 ± 6, log, 2g 72.8±8, 3 g 64. ± 6 mis It appears that bran serves as a relatively poor substrate for gas production by the colonic bacteria, both in vitro and in vivo. The abdominal symptoms reported by some subjects who supplement their diets with bran are therefore unlikely to be caused solely by excessive gas. In a recent study of patients with chronic functional abdominal symptoms attributed to excessive gas, we found perfectly normal volumes of intestinal gas both in the fasting state (1) and following meals (1 1). These patients did appear, however, to have disordered intestinal motility that interfered with the passage of gas through the bowel, and they developed abdominal pain following distention of the bowel with volumes of gas well tolerated by normal subjects. Since patients with functional bowel disease are frequently treated with bran, it is possible that reports of excessive gas following bran ingestion are actually a result of the tendency of dietary fiber to stimulate motility in these patients rather than its tendency to gasify in the gut. While the small bowel transit time of bran is longer than that of lactulose, addition of bran to the lactulose meal did not alter its rate of passage to the cecum. Cl log Loctulose log Loctulose + 2g FIG. 5. Small bowel transit time of 1 g lactulose with and without 2 g bran in four normal subjects (mean ± I SM). References 1. Lavrrr, M. D., P. FRNCH AND R. M. DONALDSON. Use of hydrogen and methane excretion in the study of the intestinal flora. J. Lab. Chin. Med. 72: 988, LVITT, M. D. Production and excretion of hydrogen gas in man. New ngl. J. Med. 281: 122, BOND, J. H., R. R. NGL AND M. D. LVITT. Factors influencing pulmonary methane excretion in man. J. xptl. Med. 133: 572, STGGARDA, F. R. Gastrointestinal gas following food consumption. Ann. N.Y. Acad. Sci. 15: 57, LvITT, M. D., R. B. LASSR, J. S. SCHWARTZ AND J. H. BOND. Studies ofa flatulent patient. New ngl. J. Med. 295: 26, PAINTR, N. S. Colonic diverticular disease. Diseases Colon Rectum 18: 549, KIRWAN, W.., A. N. SMITH, A. A. MCCONNLL, W. D. MITCHLL AND M. A. ASTWOOD. Action of different bran preparations on colonic function. Brit. Med. J. 4: 187, Downloaded from

6 S174 BOND AND LVITT 8. FINDLAY, J. M., A. N. SMITH, W. D. MITCHLL, A. J. B. ANDRSON AND M. A. ASTWOOD. ffects of unprocessed bran on colon function in normal subjects and in diverticular disease. Lancet 1: 146, BOND, J. H., AND M. D. LVITT. Investigation of small bowel transit time in man utilizing pulmonary hydrogen (H2) measurements. J. Lab. Chin. Med. 85: 546, LASSR, R. B., J. H. BOND AND M. D. LVITT. The role of intestinal gas in functional abdominal pain. New ngl. J. Med. 293: 524, LASSR, R. B., M. D. LVITT AND J. H. BOND. Studies of intestinal gas after ingestion of a standard meal. Gastroenterology 7: 96, Downloaded from