Recent studies have demonstrated that colonic bacterial fermentation of malabsorbed carbohydrate (CHO) produces short-chain fatty acids (SCFA), which could contribute to dietary energy salvage. In animals, there appears to be a symbiotic relationship with intestinal flora. The bacteria obtain metabolic energy from fermentation, the end-products of which (SCFA) are absorbed and used by the host. Animal physiology studies have shown that SCFA production and absorption provides up to 85% (ruminants), 30%-40% (rabbits), and 10% (pigs) of metabolized energy.1-3 The overall contribution of fermentation to daily energy requirements in man, however, remains unknown.
Calculations derived from fermentation equations4 and also based on the heat of combustion of individual SCFA indicate that 70%-75% of the energy originally available in CHO is liberated as SCFA.5' Because up to 98% of SCFA produced in the colon is absorbed and only 2% is excreted in feces, colonic metabolism is likely to contribute a significant proportion of daily energy requirements.6,7 It is estimated that 5%-10% of daily energy requirements is obtained from colonic fermentation in healthy adults.5,8
Patients with pancreatic enzyme deficiency are likely to malabsorb an increased amount of dietary CHO. Nutrient malabsorption in these patients may lead to negative energy balance; however, an increased CHO load entering the colon would be expected to lead to colonic adaptation to increase fermentation capacity. This could possibly contribute significantly to total body energy requirements through SCFA absorption and metabolism. We have previously reported9 increased breath hydrogen production in patients with chronic pancreatic insufficiency after ingestion of malabsorbed CHO, suggesting colonic bacteria adaptation with hyperfermentation. An understanding of the extent to which this adaptation contributes to the total energy acquisition would further increase our knowledge of nutrient assimilation in these patients and possibly lead to better clinical management through dietary manipulations. In this study, we attempted to quantify CHO malabsorption and to assess the contribution of colonic bacterial fermentation of malabsorbed CHO to the total energy requirements in these patients.
Aims
This study aimed to measure CHO malabsorption in patients with chronic pancreatic enzyme deficiency and to calculate total energy salvaged by colonic bacterial metabolism of malabsorbed CHO.
MATERIALS AND METHODS
Ten male adult patients (mean age, 49.1 ±5.1 years; weight, 54 ± 5 kg) previously diagnosed with chronic pancreatic exocrine insufficiency due to chronic pancreatitis were enrolled. Establishment of the diagnosis was by standard clinical, radiologic (endoscopic retrograde cholangiopancreatography), and biochemical evidence (Table I). Results were evaluated in comparison to 15 healthy volunteers (mean age, 38 ± 4 years; weight, 58 ± 7 kg). Patients were all receiving pancreatic enzyme replacement therapy, which was withdrawn 14 days before the study, and no further medication was taken before or during the test. Exclusion criteria included severe metabolic disturbances such as uncontrolled diabetes, bowel resection, significant respiratory or cardiovascular disease, and malabsorption/maldigestion disorders caused by factors other than pancreatic insufficiency, bacterial overgrowth, breath methane production, or recent use of antibiotics.
The study was approved by the Research Ethics Committee of Faculty of Medicine, University of Cape Town. Written, informed consent was obtained from all subjects, and the research was carried out in accordance with the declaration of Helsinki.
The patients were admitted to the Gastro-Intestinal Ward of Groote Schuur Hospital for 3 days and commenced receiving a standard diet containing 100 g fat, 329 g CHO, and 154 g protein (24.6 gN) per day, constituting a total energy value of 2800 kcal. A 72-hour stool collection (for analysis of SCFA, fat, CHO, and nitrogen) was started 24 hours after the patients began receiving the diet. Stool collection was stopped 24 hours after patients finished receiving standard diet. After the stool collection, a hydrogen breath test was performed after ingestion of 200 g (dry weight) cooked maize meal. The patients were then recalled for breath hydrogen test in response to 10 g oral inulin (as standard) in order to calculate malabsorption of CHO. Breath hydrogen tests were done after an overnight fast, in which both patients and controls were advised to eat low-CHO meals the night before. A 2-week washout period was allowed between the tests. Sleeping or smoking during the tests was not allowed.
Test Meals
Maize Meal
Commercially prepared maize meal with declared nutrient composition of 370 kcal/100 g derived from CHO 81.6 g, protein 8.9 g, fat 2.5 g, and fiber 3.4 g was prepared by adding 200 g maize meal into a preweighed cooking pot containing 200 mL of boiling water. The mixture was stirred gently for 10-15 minutes to form a semisolid paste. The meal was then removed from the cooking stove and allowed to cool before being served. The patients and controls were requested to eat the meal as quickly as possible. In this way, 165 g of CHO in the maize meal was expected to provide 10% (16.5 g) resistant starch.10,11
Inulin
Inulin at 10.6 g (fructooligosaccharide) powder extracted from Dahlia tubers (Sigma Chemical Company, St. Louis, MO), containing 94% (10 g) inulin, 5% fructose, <1% glucose, and <1% sucrose, was dissolved in 100 mL of tap water. The patients and controls were asked to drink the solution in 2-3 swallows.
Testing
Stool Studies
Samples were collected in preweighed tins and stored at -20°C for the analysis of:
1. SCFA by gas chromatography after vacuum distillation, as described by Zijlstra et al12;
2. fecal fat by the method of Kamer et al13;
3. nitrogen by automated chemiluminescence nitrogen analyzer (Anatek Instruments, Inc, Houston, TX) according to the method of Grimble et al14; and
4. CHO using Dreywood's Anthrone Reagent,15 according to the method of Ameen and Powell.1
Breath Hydrogen
Following an overnight fast, end-expiratory samples were taken from patients and controls before ingestion of the test substance. Breath samples were then taken at 30-minute intervals for 6-10 hours, until there was a return to baseline levels. The samples were analyzed by a Quintron gas analyzer (Quintron Instrument Co, Milwaukee, WI). Breath hydrogen concentrations were plotted against time, and the area under the curve for each triangulated area under the maximal rise of hydrogen concentration was calculated by the trapezoidal rule as proposed by Rumessen et al. The baseline determined by calculating the areas demarcated by the lowest previous hydrogen concentration to orocecal transit time (OCTT) was subtracted from the total area under the curve. Breath hydrogen curves that did not return to baseline at the end of the test period were extrapolated as recommended.17
Analysis
Assessment of CHO Malabsorption
Conversion of SCFA to Monosaccharide Equivalents
Daily SCFA excretion in stool was converted to monosaccharide equivalents (CHO g/d) according to the formula of Hammer et al.18 SCFA monosaccharide equivalents (starch, fructose, or galactose) all have molecular weights of 180. Consequently, starch, which is a polymer of glucose units, would have a molecular weight of (C^sub 6^H^sub 10^O^sub 5^)^sub n^, or 180^sub n^, where "n" is the number of glucose molecules. One mol of starch would therefore have the same molecular weight as "n" molecules of glucose, and hence there was no need to correct for molecular weights derived from SCFA monosaccharide equivalents when converted to malabsorbed starch.
Colonic salvage of CHO for energy was reflected by the difference between the colonic bacteria fermented CHO in the standard diet, assessed by breath hydrogen test, and the monosaccharide equivalents of SCFA measured in the stool.
Calculation of Colonic Salvage of CHO for Energy
Metabolizable energy content of food was calculated using Atwater conversion factors for available energy in a mixed diet19:
* 1 g protein = 4 kcal
* 1 g fat = 9 kcal
* 1 g CHO = 4 kcal
* Energy equivalent of SCFA derived from malabsorbed CHO was calculated as 1 g CHO = 10 mmol SCFA.20
Calculation of Energy Values
Energy requirement was defined according to the World Health Organization (WHO) guideline as "the level of energy intake that will balance energy expenditure when the individual has a body size and composition, and level of activity, consistent with long-term good health."21 Thus, the equation of energy balance is: energy intake = energy expenditure + changes in stored energy. The energy requirement was calculated from energy expenditure, which can be predicted from basal metabolic rate (BMR),21 as per the Food and Agriculture organization (FAO)/World Health Organization (WHO)/United Nations University (UNU) consultation group's recommendations.22'23 Thus, for a 30to 60-year-old man, energy requirement in kcal/d is BMR × relative activity coefficient or (11.6 × W + 879) × 1.7, where W is the body weight and 1.7 is the relative activity coefficient for an active patient.22
The BMR in the patients (W = 54 kg) in this study was therefore calculated as {(11.6 x 54) -I- 879} = 1505.4 kcal/d; hence, the energy requirement was 1505.47 × 1.7 = 2559.2 kcal/d.
A flow diagram of the experimental procedure is summarized in Figure 1.
RESULTS
A representative chromatogram of SCFA profile obtained from pooled faecal aliquots from patients is shown in Figure 2. Acetate, propionate, and butyrate formed the major peaks while branched chain SCFA, isobutyrate and isovalerate were minor. Calculated molar ratios of principal SCFA was 48:11:55 for acetate:propionate:butyrate respectively (Table II).
Balance Studies
The results of the 72-hour stool analysis are illustrated in Table III. Mean daily fecal weight was 376 ± 263 g/d, with a fecal nitrogen output of 2.4 ± 1.5 gN/d (normal, <1.5 g/d). Linear regression analysis demonstrated a significant positive correlation (r = 0.97; p < .05) between fecal nitrogen output and fecal wet weight (Figure 3). Mean fecal fat output was 38.1 ± 25 g/d (normal <5 g/d). There was significant positive correlation between fecal fat output and fecal wet weight (r = 0.83; p < .05; Figure 3). Fecal SCFA production positively correlated with wet weight (r = 0.95; p < .05; Figure 3), whereas no significant correlation was found between stool CHO excretion and wet weight (Figure 3). Reducing sugar constituted 36% of total CHO excreted in stool.
Malabsorbed CHO
The results of the breath hydrogen responses to maize meal and inulin in patients and controls are shown in Figure 4. Breath methane excretion was excluded because not everyone harbors methane-producing bacteria; hence, methane excretion is inversely proportional to hydrogen excretion. In addition, previous studies have shown that quantitative estimates of CHO malabsorption using methane excretion do not seem to offer any advantage over breath hydrogen measurement.24,25
OCTT, defined as the interval between ingestion and initial sustained rise in breath hydrogen concentration of 10 parts per million (ppm) or more,26 was calculated as 4.4 ± 0.6 hours vs 4.2 ± 0.6 hours after inulin ingestion and 3.8 ± 0.4 hours vs 4.0 ± 0.9 hours after maize meal ingestion in patients and controls, respectively. There was no significant difference in OCTT between patients and controls.
There was also no significant difference in the amount of CHO ingested between patients and controls (55 ± 25 g vs 52.3 ± 17 g, respectively). However, patients produced significantly more breath hydrogen than controls after both maize meal and inulin ingestion, respectively (p < .05; Table IV). Subsequently, calculated malabsorbed CHO was significantly greater in patients (11.16 ± 10 g or 21.4% ± 17% of ingested CHO) than in controls (5.21 ± 1.6 g or 10.2 ± 1.4% of ingested CHO; p < .05; Table IV).
Calculations based on CHO malabsorption from the breath hydrogen test therefore suggested that patients receiving the test meal malabsorbed 70.4 g CHO (281.6 kcal/d; Tables IV and V). Total CHO loss in stool was measured as 8.1 g/d (4.8 g as monosaccharide equivalents of SCFA and 3.3 g as stool CHO; Tables III and V). CHO loss in stool therefore accounted for 32 kcal/d, or 2.4% the of total daily dietary CHO intake (Table V).
CHO absorbed as SCFA equivalents for energy metabolism was therefore 70.4-8.1 g = 62.3 g, or 19% of total daily dietary CHO intake (Table V). Thus, colonic bacterial fermentation liberated 88.5% (62.3/70.4) of total energy originally present in malabsorbed CHO as SCFA. This is equivalent to 649 mmol SCFA,18 and because 50 mmol/d SCFA excretion was measured in stool, 92.8% (649/699) of SCFA produced in the colon was therefore absorbed, and hence 7.2% was excreted in feces.
The amount of hydrogen produced from fermentation of malabsorbed CHO in the test meal and lost as expired air was calculated as 0.04 g/d, assuming an average ventilation rate of 7 L/min and that 1 mol of hydrogen occupies 22.4 L at standard room temperature and pressure.
Energy Salvage
From schematic representation of colonic CHO energy budget in Figure 5, CHO intake (A) is therefore 329 g; malabsorbed CHO (B) = 70.4 g; loss in stool as simple sugars (C) = 3.3 g; loss in stool as SCFA equivalents (E) = 4.8 g; available for energy salvage (G) = 62.3 g (260 kcal/d, or 649 mmol SCFA). Colonic energy salvation, defined here as the ability of the colon as a metabolic organ to contribute to the total energy requirement expressed as a fraction of energy expenditure, was (G)/energy expenditure. Thus, 62.3 g CHO (260 kcal) divided by calculated energy expenditure (BMR × 1.7) as explained in "Materials and Methods,"22,23 is equivalent to 260/2559 = 10.2%.
DISCUSSION
Patients with pancreatic enzyme deficiency are characterized by steatorrhea (high stool fat), creatorrhea (high levels of nitrogenous compounds in stool), and poor nutrition status resulting from negative energy balance.27 In this study, for the test meal, mean fecal weight was 375 g/d, and fat excretion was 38.1 g/d. Normal healthy adults have fecal weights of <250 g/d18'28'29 and fecal fat excretions of <7 g/d.27'30 Our results show positive correlation between stool fat and wet weight suggesting that faecal fat was contributing to the increased faecal wet weight (Figure 3; Table III). Stool nitrogen excretion was measured as 2.4 gN/d, suggesting that the patients had parallel protein malabsorption. Stool nitrogen excretion in normal adults has been recorded as less than 1.5 gN/d.27
CHO malabsorption/maldigestion has been shown to cause diarrhea by increasing osmotic load in the colonic lumen,18'29 contrary to earlier hypotheses that SFCA derived from colonic fermentation of malabsorbed CHO caused diarrhea by lowering colonic pH.31 Increased SCFA production in the colon reduces the osmotic effects of CHO overload and also enhances fluid, chloride, and sodium absorption. Parallel stimulation of bicarbonate excretion raises the colonic pH, thus correcting the pathologic effects of diarrhea.3 Stool CHO excretion was measured as 3.3 g/d, and there was no significant correlation between stool CHO output and wet weight (Figure 3), suggesting that the stool CHO was not contributing to the total fecal wet weight. Noorgaard et al28 previously reported that up to 28 g CHO/d may be lost in stool in patients with malabsorptive disorders. That our results are rather similar to the findings of Hammer et al,18 who measured stool CHO loss in healthy adults as 3.0 g/d, suggests a possibly increased role of colonic bacteria in converting malabsorbed CHO to other metabolites in these patients. We measured stool CHO loss using Dreywood's anthrone reagent,15 which binds CHO irrespective of molecular size as opposed to the commonly used reducing sugar assay, which reacts only to the reducing-end moiety of the CHO molecule and would consequently give a negative result with some disaccharides like sucrose. Our results are therefore unlikely to have underestimated CHO loss in stool. It has previously been shown that CHO-induced diarrhea occurs only when the fermentation capacity of the anaerobic colonic flora is exceeded, and not by limitation in the colonic capacity to absorb SCFA.16,32,33 Osmotic diarrhea due to CHO malabsorption has been reported to occur only in patients with severe pancreatic insufficiency (fecal fat 40-120 g/d).18 The patients we studied had mild to moderate pancreatic enzyme insufficiency (fecal fat output of 38 g/d) and did not have severe diarrhea (Table I). It is therefore likely that malabsorbed CHO in the standard diet was largely metabolized and absorbed as SCFA by these patients. Our calculations here suggest that 88.5% of energy originally available in malabsorbed CHO was converted to SCFA, and that 92.8% of SCFA produced was absorbed and 7.2% excreted in feces. These results are comparable to previous estimates, based on heat of combustion, indicating that 70%-75% of energy originally available in CHO is liberated as SCFA5'6 and that up to 98% of SCFA produced in the colon is absorbed.6,7 Stool SCFA excretion in healthy subjects has previously been estimated as <2%6,7 as opposed to our measurement of 7.2%. Stool SCFA excretion is an indication of the difference between production and absorption,34 and in the absence of diarrhea, or any other pathologic condition that may affect SCFA metabolism and absorption, high stool SCFA measurement may imply increased production. Cummings et al35 have reported stool SCFA output in British subjects as 10-20 mmol/d. Stool SCFA output in our patients was 50 mmol/d (Table III), indicating a greater production of SCFA in patients with maldigestive states.
Like dietary CHO, protein malabsorption may also provide substrates for colonic bacterial fermentation. Colonic bacterial proteolytic enzymes yield amino acids, which may serve as further substrates for fermentation. However, it has been shown that amino acid fermentation yields not only SCFA but also disproportionately high amounts of branched-chain fatty acids such as isobutyrate, isovalerate, and 2-methylbutyrate, which arise from branched-chain amino acids (valine, leucine, and isoleucine, respectively).36 However, the amount of branched-chain SCFA is quantitatively much smaller than the 3 major SCFAs of acetate, propionate, and butyrate.36 Our results here suggest that protein fermentation did not contribute significantly to overall SCFA production, because there was no significant disproportionate increase in branched-chain fatty acids to indicate increased protein catabolism in the fermentation process (Figure 2; Table II). The contribution of fermentation of malabsorbed proteins to the total energy requirement by the host was therefore likely to be insignificant compared with that of CHO.
Despite the widely recognized importance of malabsorbed CHO in colonic physiology, accurate measurement of malabsorbed CHO remains elusive. Breath hydrogen measurement, though widely used, has been criticized on account of lack of appropriate standard and wide intra- and interindividual variations in breath hydrogen excretion.25,37 However, unlike tolerance tests that measure absorbed rather than malabsorbed CHO, the breath hydrogen test seems to be desirable in screening malabsorption because it is noninvasive and does not disturb the normal physiologic functions of the colon.38 It has recently been suggested that breath hydrogen measurement of CHO malabsorption be factored into the interpretation of indirect colorimetric estimation of energy expenditure in patients with short bowel syndrome.39
Previous studies based on breath hydrogen measurements estimated CHO malabsorption as 30%-40% (Ladas et al30), 48% (Royall et al40) in patients with pancreatic enzyme deficiency, 40% (Woolf et al41) in patients with short bowel syndrome, and 5%-10% (Anderson et al42) in normal subjects. In this study, malabsorption of CHO in maize meal was significantly greater in patients (21.4%) than in normal controls (10.2%; p < .05; Table IV). There was no significant difference in the amount of CHO ingested by patients and controls, yet breath hydrogen production was significantly greater in patients than in controls after ingestion of both maize meal and inulin, respectively (Table IV). We could have expected breath hydrogen production in response to inulin to be similar in both patients and controls, but our findings here are consistent with a previous one9 suggesting that the patients we studied are experiencing an adaptation of the colonic bacteria to hyperfementation in response to increased CHO load.
We used inulin as a standard in breath hydrogen test assessment of malabsorbed CHO, rather than the commonly used lactulose, because previous studies have shown that lactulose fermentation by colonic bacteria leads to overproduction of breath hydrogen compared with complex CHO.9,43 Inulin, being a complex CHO like starch, ferments less rapidly compared with lactulose and may better mimic starch fermentation9 because it has previously been shown that colonic hydrogen absorption is highly effective at low but not higher colonic hydrogen accumulation rates.44 It may be argued that the rate of hydrogen production per g of inulin, maize meal, or mixed diet may not necessarily be the same, but the similarity we observed in breath hydrogen profiles, as well as transit times, between inulin and maize meal (Figure 4) may suggest that the rate at which bacterial enzymes depolymerize inulin and maize meal is similar. We have no evidence that colonic bacterial hydrolysis of malabsorbed CHO in a mixed diet would be different.
Even though our results seem to be somewhat lower than those previously reported, it must be considered that breath hydrogen test assessment of malabsorbed CHO, despite its advantages, is only a semiquantitative method25 and is dependent on the chosen reference standard. Nordgaard et al45 could not establish a linear relationship between breath hydrogen excretion and malabsorbed CHO in patients with short bowel syndrome and cited different routes of hydrogen disposal as a possible reason. Despite these limitations, breath hydrogen test remains the method of choice, and the significance of our findings using this approach is that 19% of the total energy available in dietary CHO is salvaged as SCFA for energy provision in these patients (Table V). Our calculations suggested that 0.04 g/d of hydrogen produced from colonic fermentation of malabsorbed CHO was lost in exhaled air. This represents a negligible amount of CHO energy loss (Figure 5) because hydrogen is neither produced nor used by human tissues,39 and the amount absorbed in systemic circulation is a net product of acetogenesis, methanogenesis, or sulfate reduction in the fermentation process. Energy lost in breath as carbon dioxide may similarly be insignificant because the bulk of it is a by-product of the respiratory processes of the body tissues, whereas the proportion generated by the bacterial fermentation is largely not absorbed but rather used in acetogenesis in the colon.
Our calculations here indicate that colonic anaerobic bacterial metabolism of malabsorbed CHO in these patients contributes 10.2% to the overall energy expenditure/requirement. To our knowledge, our study has provided the first noninvasive direct measurement/calculation of the actual contribution of colonic bacterial CHO metabolism to the total energy expenditure in patients with exocrine pancreatic insufficiency. Our observations therefore suggest that giving patients with chronic exocrine pancreatic enzyme insufficiency a low-fat/high-CHO diet may be beneficial in correcting the negative energy balance commonly experienced by these patients. Further studies are suggested.
CONCLUSIONS
Colonic bacterial metabolism could represent a significant source of energy salvage in patients with chronic pancreatic exocrine insufficiency.
ACKNOWLEDGMENTS
The authors acknowledge G. O. Young, PhD, for her laboratory assistance with this work.
The study was supported by a grant from the Medical Research Council of South Africa.
