progress report - gut · galactose), pentose (xylose and arabinose), and uronic acid monomers which...

Gut, 1981 22, 763-779

Progress report

Short chain fatty acids in the human colon

The human colon contains a luxuriant mixed culture of bacteria, probablyat least 175 g. The special property of these bacteria is their metabolism,which is in the main strictly anaerobic, the end-products being primarilythe short chain, or volatile, fatty acids, acetic, propionic, and butyric acid.These are the principal known anions in the colon but, although this hasoften becni rcported,' little thought has been given to the possibleimplications.

PhysiologyOCCU I11 N C' I O 1` S H 0 R C H A I N 1: ATTY A C I 1) S I N C O 1, 0 N

Acetate, propionate, acln butyrate occur in the large intestine of all mam-malian species so far stutdied. They are the pre(donminalnt anions in herbi-vorous animals like the horse,- kangaroo, rabbit, and guinea-pig,,omnivores such as the pig," and even the dlog, a carnivorous species.! Theyare the major anions in humnan faeces' but no values have yet been re-ported of their concentrations in the rest of the colon. In addition to thelarge intestine short chain fatty acids are also found in the rumen wherethe complex metabolic pathways leadling to their formation have beenextensiv elv studlied (see re iew by Prins, 1977)."'

Fermentation in the rumen is very similar to that in the colon and,given appropriate substrates, experimental evidence indicates that thispattern of metabolisnm takes place in the humani colon.'' ` The microfloraof human faeces is similar to that of the runmen' and reflects that of therest of the colon.'' 1lunman colonic microflora are capable of fermentationreactions necessary for the production of short chain fatty acids.'-' "f; Thereis a close similarity in the molar ratios of short chaini fatty acids in thecolon amongst various species (Table). Concentrations of these acids inhuman faeces are, however, lower than those seen in the colon of the rator pig and in the rumen. This niay be a true species (liffer-ence but is morelikely to reflect differences betweeii faecal antl colonic contents.

Production and absorption rates are the primary factors controllingshort chain fatty acid concentrations and nmolar ratios, although the rateof passage of digesta through the gut (transit) is also important.i2- Theeffect of diet is less significant than one would expect. For example, chang-ing a dog from a cereal- to meat-based dlict hadl virtually no effect on faecaland colonic short chalin fatty aciel levels,!' altlhough in pigs a cereal-baseddiet led to higlher acetate andJ lower propionate levels than a formula-typediet.22 Fasting redluces prodluction andl concentrations.i -' In man, ' achange from an ad libitlwni diet to one containing only carbohydlrate ledlto an overall fall in total slhort chain fatty acidls concentration in faecaldialysate from 85 to 46 mmol/l antdl a small change in the molar ratios ofacetate: proprionate: butyrate from 59 :22: 19 to 68: 17: 13. When the

763

on October 14, 2020 by guest. P

rotected by copyright.http://gut.bm

j.com/

Gut: first published as 10.1136/gut.22.9.763 on 1 S

eptember 1981. D

ownloaded from

http://gut.bmj.com/

Table Shiort chain fatty acid concentrations in gut

Species Site Acetate Propionate Butyrate References(mmol/kg)

Rat Caecum 101 57 32 Remsey and Demigne (1976)16Colon 70 29 7Faeces 75 27 16

Pig Caecum 118 68 25 Argenzio and Southworth(1974)8

Sheep Rumen 65 21 18 Hungate (1966)18Cow Rumen 66 23 20 Hungate (1966)18Man Faecal

dialysate 41 12 13 Cummings et al. (1979)1946 17 15 Rubinstein et a!. (1969)439 19 14 Bjork et al. (1976)20

Faecalwater 93 46 24

Faeces 48 11 5 Zilstra et al. (1977)21

Molar ratios (0

Pig Faeces 66 22 7 Sambrook (1979)22Caecum 56 32 12 Argenzio and Southworth

(1974)8Rat Caecum 61 25 14 Remsey and Demigne (1976)17Sheep Rumen 62 21 16 Hungate (1966)18and cow

Man Faeces 60 24 16 *

*Average value from Rubenstein et al.Cummings et al. (1979)19.

(1969)4; Bjork etal. (1976)20; Zijlstra etal. (1977)21;

same subjects took only methyl cellulose an additional fall in concentrationwas seen (to 23 mmol/l) and the molar ratios were 60 : 24: 15. Calorieintake was considerably reduced, however, in the experimental diets. In adifferent study1" no significant change in short chain fatty acid concentra-tions was seen when four healthy subjects went from a low protein (66+4.6 (SEM) mmol/l) to an equicaloric high protein diet (58±+4.7 mmol/l)or when 39 g of dietary fibre were added to the high protein diet (66+4-2 mmol/l).

SOURCE S AND MET ABOL ISM Ol SHOR I CHAINFAI TY ACIDS

Short chain fatty acids are produced by fermentation from carbohydrate.The major source of this in the human colon is thought to be plant cell-wall polysaccharides such as cellulose, pectins, and hemicelluloses, cur-rently referred to in human nutrition as dietary fibre. Starch would alsobe a suitable substrate if it were to reach the colon in significant quantities.Plant cell-wall polysaccharides are composed of hexose (glucose andgalactose), pentose (xylose and arabinose), and uronic acid monomerswhich are fermented by gut micro-organisms along a variety of pathways.The important feature to remember of this metabolism is that it isanaerobic.Hexose breakdown is mainly via the Embden-Myerhoff-Parnas glyco-

lytic pathway to pyruvate. Alternatively, hexose is converted to 6-phospho-

764 Cummings



j.com/


eptember 1981. D

ownloaded from

http://gut.bmj.com/

Slhort chain fatty acids in thc Ihuiman colon

gluconate and then metabolisednmainly by the pentose phosphiate cycle.'This latter route, a niajor one in some aerobic bacteria, is also used forthe conversion of pentoses to glycolytic intermediates and intermediatescssential for the synthesis of nucleic acidls, cofactors., and amino acidis.Pyruvate is the main metabolite of these fermentation reactions but verylittle is found in the gut because it is converted to a series of end prodticts

namely, acetate, propionate, butyrate, carbon dioxile, hydirogen,niethane, and water. Acetate is usually fornied by the oxidative dlecarboxy-lation of pyruvate and butyrate by reluction of acetoacetate formed fromacetate. The production of propionate is by two main, but circuitous,routes, firstly, involving fixation of CO., to form succiniate which is subse-cquently decarboxylated (the 'dicarboxylic acid pathway') or, secondlly,froni lactate and acrylate (the 'acrylate pathway')."'

Lactate (both D andl L) may be formed froni pyruvate by niaiiy gutanaerobes but is not a key interniediate in fernientationi and significantamounts are rarely found in the human or animal gut, except in the pig'sstomach.< When, howexer, flux through the glycolytic pathway is! high, aswhen large quantities of reatily fermentable soluble carbohydrate areavailable, lactate prodiuction is favoured. Increased fermentation ratesalso result in a lowering of pH which in turn iiilibits the nietabolisnm oforganisms which uitilise lactate. Thus lactate levels may become signifi-cant.') "i Up to 50 mmniol/l latctate levels in faeces have beell reported ininfants suffering froni acute infectious diarrhoea when fed on milk27 andin children with deficiency of sugar-splitting enzymes in the small gut.2 29In these circumstances mialabsorption of sugar favours colonic lactate pro-duction, althoughi acetate usually remains the principal anion. In manD-lactic-acidosis has been reported in a patient with the short bowelsyndlrome' where cir-cumstances would dictate similar fermentationpatterns in the colon.Other interniaeliates in the anaerobic breakdown of coniplex carbo-

hydrates in the rumen and colon are hydrogen, ethanol, and formate.Formaite is not found in significant amounts because it is rapidly con-verted to hydrogen and carbon dioxide.l' In human faeces formateconcentrationis are very low (1-2 mmol/l).' Ethanol (and methanol) areproduced in relatively small quantities and are absorbed without additionalmicrobial metabolisni. Highi levels (90-130 mmol/l) have been reported inthe caecum of the ptarmigan'" and, as with lactate, alcohol production issaid to be favoured by acid pH.'9Hydrogen is a key intermediate in any fermentation system but con-

centrations in the rumen are low because it is rapidly converted tomethane by clirect reduction of CO. by methanogenic bacteria. Methanemay also be produced from a number of other substrates includiing acetatebut this is seen only when retention of material in the system is four tofive days or longer. Both hydrogen andnmethane are produced in man.:,In clear dlistinction from the rumen, however, only about 30-60% ofhunians prodluce niethane, at least of those exainiiietd in the USA ' andUK.'9 These differences niay possibly be (lue to the smaller aniounts ofCO., available for fernientation in the 'Westernised' colon or to differencesin transit tinie. About 75-80°/O of Nigerians prodluce nietliane (Tompkins,Drasar, Wk,igginis, Bradley, andl Gyselynck; personal coninilication) which

765



j.com/


eptember 1981. D

ownloaded from

http://gut.bmj.com/

may be due to their much higher intake of complex carbohydrate and thusmore typically rumen-like fermentation in the caecum and colon.

TOTAL, DAI LY PRODUCTION 01: SHORTr CHAI N

I A T I Y A C I 1) S

There can be little doubt therefore that a fermentative system similar tothat of the rumen occurs in the colon of many animals including man. Itsimportance, in terms of overall metabolism, will depend on the amountand type of substrate degraded and on the fate of the short chain fattyacids. Fermentation has been quantified in the rumen and an overallequation for the anaerobic breakdown of complex carbohydrate derived:

58(Ct;Hi20,;)-#62 CH1COOH+22 CH3CHL2COOH+ 16 CH,(CH,)2COOH +605 C02+33.5 CH,+27 HR0 (1)

Miller and Wolin (1979)1' have derived a similar equation for fermentationin the human colon based on the molar ratios of short chain fatty acidsin faeces and known proluction of carbon dioxide and methane:

34-5 C,;H 64 SCFA + 23.75 CH, + 34-23 CO, + 10.5 HWO (2)

This equation assumes that short chain fatty acids are the sole non-gaseousend-product, but takes no account of non-methane producers.How can colonic fermentation be quantified in man? Two possible

approaches can be made, both assuming the rumen data to be approxi-mately correct for the colon. Firstly, if the amount of substrate availablefor fernmentation is known then short chain fatty acid proluction can becalculatedl. The anmount of dietary carbohydrate, presumably the mainsubstrate for the bacteria, whiclh is fermented in the colon each (lay isarounl 20 g. About 15 g from dietary fibre and 5-6 g of soluble carbo-hydrate.."' From equiation (2) this would yield about 200 mmol shortchain fatty acids. This figure will be an underestimate in people consumingmore (lietary fibre than the UK population' ' or if other carbohyiratesubstrates are available.An alternative approach, used by Smith antl Bryant, 2 is based on the

amount of bacteria excretedl in human faeces each (lay. In i?itro fermenta-tion studies have demonstratedl the amount of carbohydrate needlel to pro-vide energy for the growth of anaerobic bacteria. One mole of hexose,metabolised anaerobically, yields 5 mol ATP.21 2` The theoretical yield ofbacterial cells (dry) is 26 g/mol ATP.'"-' This falls to about 20 g/molwhen the cost of maintenance energy for bacteria is taken ilito aiccount ata turnover time of about 48 hours. On average for eaich mole of hexosefer-nmented, about 62 g of bacterial cells are pro(lucedl or 34O' of theweight of carbohydrate. Smitlh antl Bryant-- suggest, on theoreticalgrounds, that for man, bacteriai produced are equivalenit to 'at least 1300of the weight of carbohy(drate utilised atnaerobically', which is a( muchlower figure. In fact, the actual yield of bacterial cells per- gram ofhexose fermentedl varies widely dependling on the availability of pre-formed monomners (such as amino acids) for biosynthesis, cell wall coni-position, and maintenance enlergy requirements of the baicteria. M9uclh isknown about this in the rumen'- but little for man, which makes it

766 Cummings



j.com/


eptember 1981. D

ownloaded from

http://gut.bmj.com/

Sliort chain fatty acids in thehl'uman colon

difficult to predict accurately the substrate requirenments for bacterialgrowth in the coloIn.

However, the amount of bacterial cells produced in the colon in man isknown, and can be used as a basis for calculating short chain fatty acidyields. Stephen and Cummings7 '` have shown that bacterial cell yields(dry weight) from fermentation of dietary fibre from cabbage are 28%' ofthe weight of fibre fermented and for wheat bran 36o%.l In these studies itwas assumed that the only carbohydrate available in the colon, as a resultof the experimental dietary changes, was non-starch polysaccharidle fromthe cell wall material. Were other carbohydrate to be carried into thecolon as a result of the dietary changes, then these appairent yields wouldbe overestimates. Nevertheless, they are remarkably close to rumen values.Daily excretion of bacterial cells in man is about 15 g while on an averageUnited Kingdom diet, although it can be considerably more than this."'If a conversion factor of 25%, is taken, this is equivalent to 60 g carbo-hydrate fermented daily and a yield of 600 mmol short chaini fatty acids.This is an intriguing approach. It is not currently thought that so muchcarbohydrate is metabolised in the colon. However, anaer obically, bac-teria get much less energy for growth from protein andl one would have toassume the whole of normal protein intake to pass to the coloni to sustainmicrobial growth, an unlikely occurrence. Long chain fatty acids are notused by anaerobes as an energy source."' To sustain known bacterialgrowth either more carbohydrate passes into the human coloin each daythan is currently believed, or the yield of bacterial cells per g of carbo-hydrate is higher than that in the rumen. Alternative possible sources ofcarbohydrate other than dietary fibre available for breakdown to provideenergy for bacterial growth include unabsorbed star-ch, which recentstudies reported by Anderson et al. " suggest could be an important con-tributor, or mucous secretion.Whatever the possible source, if the energy requirements for the growth

of 15 g of bacteria in the gut are met each day by anaerobic fermentationthen a considerable amount of short chain fatty acids are produced. Whlatis their fate? These acids are the end products of carbohydrate fermenta-tion. With the exception of small amounts of acetate being converted tomethane in subjects who have slow transit times, they must either beexcreted in the faeces or absorbed. In the adult human with a faecal out-put of 80-230 g/day only 7-20 mmol/day are excreted.l' so most shortchain fatty acids must be absorbed.

ABSORP 10 N

There can be little doubt that short chain fatty acids are absorlSed from thehuman colon. Were they not, man would be unique in the animal king-dom. All mammalian species studied absorb these acids including therat,'7 pig,' horse, and goat. ' Absorption of all three slhort chainfatty acids from the human colon has been shown. McNeil ` meacsuredabsorption of these acids in the human rectum and delscending antd trans-verse colon and showed that absorption rates are comparable with thoseobserved in animals-for example, human 6.1-12-6 [Lmol/cm2/h; horse8 ,umol/cm2/h;`" pig 8-10 Amol/cm2/h,. and from the rumen 10.5 £mol/

767



j.com/


eptember 1981. D

ownloaded from

http://gut.bmj.com/

cm2/h. '(W Net movement out of the colonic lumen of short chain fattyacids is more rapid than net sodium transport. ';The exact mechanism whereby these acids are absorbed still requires

additional study in all species, but a number of general characteristicsemerge from published reports. Transport from the lumen is invariablyassociated with the appearance of HCO. , with the stimulation of sodiumabsorption (except in the horse)" and is independent of bulk waterflow.; " Important unanswere£1 questions in man include whether theyare absorbed primarily as acids Dr in the ionised form, whether there is aspecific membrane carrier for short chain fatty acids, how the transport ofthese acids affects that of other ions and, more particularly, the influenceof mucosal short chain fatty acid metabolism on their transport.

Role of bicar-bonateAn understanding of HCO:, metabolism must hold the key to the mech-anism of short chain fatty acid transport, as it appears consistently in thelumen during absorption of these acids. HCO, accumulation is indepen-dent of the chloride/HCO;, exchange known to occur in the human ileumand colon, 2 G3 as its appearance continues when chloride is replaced in thelumen by sulphate.' In both the rat5 " and pig colon-' HCO2- appear-ance is independent of chloride/HCO:- exchange. On a molar basis HCO,-appearance is equal to about half total short chain fatty acid absorption inthe rumen';'i' and colon. " "

The source of HCO: is either plasma, or hydration of CO, with car-bonic anhydrase (CA):

H20+C02 CAH2 CO±, H` HCO:1- (3)

In those parts of the gut where fermentation occurs that is, rumen,caecum, and colon the epithelia contain high concentrations of carbonicanhydrase,';"- ' although levels in the human colon have yet to be reportedl.Intracellular HCO: could be exchanged for the ionised form of shortchain fatty acids across the luminal border of the cell in a way analogouswith chloride exchange. With an average pKa of 4.8 short chain fatty acidlswill be about 95%o ionised at the pH of human caecum (6.0) and leftcolon (7.0).2 The rise in luminal pH seen when these acids are absorbedfrom all parts of the human gut i' -' and in the horse"' andl pig.'would fit with HCO.- exchange because the accumulation in the lumenof HCO,- would shift the reactions shown in equation (3) to the left. Thiswould result in a rise of pCO. which has been observed in the rat2 andhorse-! but in the pig, the rumen, and man7 pCO, falls.

In fact, gut mucosa is relatively impermeable to the ionised form ofthese acids, so an alternative and more likely mechanism of absorption ofshort chain fatty acids has been suggested. '!7 In this mechanismluminal hydration of CO. occurs and allows protonation of the acid andits rapid absorption by diffusion in the unionised form. This would lead toHCO3: accumulation and a fall in pCO, as observed in man, and in the pigwhose colon is similar to the human. Absorption in the unionised formwould be additionally favoured by the acid microclimate at the colonicmucosal surface,-'; although this is currently disputed.>9

768 Cummings



j.com/


eptember 1981. D

ownloaded from

http://gut.bmj.com/

Sl70rt chain fatty acids in tdeIehuman colon 769

Sodium absorptionAnother clue to the mechanism of short chain fatty acid absorption liesin the stimulation of sodium absorption seen during acetate transport inseveral species. 2 !. In the rat colon it is notable that neither succinatenor lactate, which are poorly absorbed, stimulate sodium absorption,whereas acetate dloes. The most likely interaction between short chainfatty acids and sodlium would be through a recycling of hydrogen ion (H+).Transport of the unionised acid into the cell drives Na H' exchange andstimulates sodium absorption. In the studies of Argenzio and Whipp t ofthe perfused pig colon marked differences in luminal pCO2 and pH wereobserved when either Na2SO, or CH:,COONa was perfused. Perfusion withNa2SO, was associatel with a fall in HCO. concentration, rise in pCO2,andl acidification of the luminal solution; this suggests that net Na absorp-tion is associated with I-' secretion. With sodium acetate the reverse isseen. The secreted H' ion facilitates absorption of the unionised acid. How-ever, in man an Nai-H exchange is not thought to occur in the colon" andlperfusion of Na2SO, solutions through the isloated human colon does notlead to luminal acidification.S'l

Several other possibilities for the control of short chain fatty acid ab-sorption have been postulatedl. One proposed by Jackson9' depends on thedevelopment of an area of high pH within the mucosa, probably in theintracellular space. This would facilitate the absorption of weak acid. Onthe other hand Lanmers'` has shown evidence for the existence of a carrierfor monocarboxylate ions across the gut mucosa and suggested that thiswas likely to be situated in the basolateral border of the epithelial cell.These disparate observations indicate that a great deal more needs to be

learnt about the tralnsport of short chain fatty acids across the humancoloIlic mucosa. Both species differences and those in transport at variouslevels in the gut make the dlevelopment of a unifying hypothesis difficult.It is remarkable, however, that the study of electrolyte transport in thelhuman colon should have proceeded for more than 20 years, yet onlyrecently has attentioni been given to the role of short chain fatty acids,the major anioni in the colon, and one which clearly affects both sodiumand bicarbonate transport.

NI U CO S A 1. N[l I A 13 0 1,1 S NI

Study of the transport of short chain fatty acids across colonic mucosa iscomplicated by the fact that substantial amounts are metabolised withinthe epithelial cell itself. In the rumen absorption rates of short chainfatty acids increase with chain length (acetate<propionate<butyrate),but the amount transported to the blood dlecreases,';" differences attribut-able to cellular metabolism. Ketone body production (acetoacetate andfl-hydroxy butyrate) accounts for 30°O of acetate metabolisnm, 75%/, ofbutyrate, and some conversion of butyrate to acetate occurs.'' In thesheep Bergman"' has shown that 500o of absorbed propioniate atnd 90%of butyrate are metabolised in the mucosa, the rest being almost entirelycleared by the liver. Very little propionate and butyrate are found in thearterial blood of ruminants. In addition 24-350/ of acetate is removedfrom the blood during passage through portal viscera (180o/ of acetateturnover).



j.com/


eptember 1981. D

ownloaded from

http://gut.bmj.com/

Much less information is available about short chain fatty acid metab-olism by caecal and colonic mucosa. In the pig substantial amounts (over50%0) of total short chain fatty acids appear to be used by the mucosaSbut in the caecum of rabbits7 and rats'7 only about 12% of butyrate is con-verted to ketone bodies and virtually no conversion occurs in the mucosaof the colon. In man, using isolated human colonic epithelial cells,Roediger92 has shown that butyrate is an important energy source forcolonic mucosa, accounting for the major part of energy needs even in thepresence of glucose.

Measurement of short chain fatty acid absorption from colonAs adequate methods are available for chemical estimation of these acids":most of the problems in studying their metabolism in the colon arise fromdifficulties in studying this organ. A variety of techniques have beenapplied to measuring short chain fatty acid moxemetnt in the colon,including in vitro tissue studies, colonic perfusion, and the dialysis method.

Colonic mucosal electrolyte transport has been measured in vitro usingmodifications of the Ussing technique.';:'., 5 This method has the advantageof allowing bidirectional fluxes of ions across the mucosa to be measureddirectly and to isolate the effect on electrolyte transport on the highspontaneous electrical potential difference found in vivo. Problems in inter-pretation arise from difficulties with the viability of tissue and the fact thatit is removed from its normal neural and hormonal controls. It is useful,however, for the study of mechanisms of electrolyte transport and, whileno data have yet been published of its use for human short chain fattyacid absorption, the techniques have been applied to animal studies.5!' Analternative in vitro method using isolated colonic epithelial cells hasrecently been reported by Roediger and Truelove.-i The metabolism ofindividual coloncytes can be studied and already the technique has yieldedvaluable information on short chain fatty acid metabolism by these cells.>'

In vivo perfusion of the whole colon has been established over a numberof years7 "! and has been applied to the study of short chain fatty acidabsorption. The advantages of being able to study colonic functiondirectly in vivo are obvious and the technique has provided useful in-formation on overall colonic absorptive capacity. There are, however,problems with contamination of perfusing solution by ileal effluent, ofreflux of perfusate into the ileum, in validating steady state conditions, inthe large amounts of isotope needed for unidirectional flux studies and,particularly for short chain fatty acid metabolism, of contamination ofsolutions with retained faecal material. The technique has not been suc-cessfully applied to individual segments of the colon, and, as colonicfunction varies from region to region,` !10 this information is lost. Someof these problems can be overcome by perfusion of the isolated wholecolon, either defunctioned` or surgically isolated," but the precisionof the technique can never be great.

Edmonds'' has described a method for studying colonic electrolytetransport in which a small amount of test solution, in a sac of dialysistubing, is placed in the rectum or colon and changes in ion concentrationwith time observed. The technique has been well validatedl` !' andl has theadvantage of allowing clearly defined areas of the colon to be studliedl witlh

770 Cummings



j.com/


eptember 1981. D

ownloaded from

http://gut.bmj.com/

Slwort chain fatty acids in the human colon

minimal disturbance to the subject. Because of the much smaller amountsof fluid used, it is a highly sensitive technique and very low doses of radio-isotopes are needed for detailed studies. At the present time it seems tooffer the best prospect of obtaining accurate information on short chainfatty acid transport.

Short chain fatty acid metabolism in the colon has been studied in avariety of other ways. These include faecal analysis,'"' 1¶2 the dialysis bagtechnique of Wrong,` " direct installation of substrates into the caecum,"'and in vitro fermentation studies, which are used widely in animal physi-ology. -All of these should be considered when approaching a particularproblem of colonic short chain fatty acid metabolism.

Clinical implicationsELE.1 C T R O L Y TI I R A N S P O R T A N D D I A R R H O E AIt is said that changes in human stool output are due to the effect of shortchain fatty acids generated in the colon. These acids are thought of aspoorly absorbed anions in man and to cause fluid retention in the colon byan osmotic effect. This view has been strengthened by the finding that, inchildren with diarrhoea due to carbohydrate malabsorption, stool outputcorrelates with short chain fatty acid output27 2" and similarly in adultswith diarrhoea due to malabsorption,"' lactose intolerance,' and evencatharsis with magnesium sulphate.} In some circumstances short chainfatty acids may even induce fluid secretion in the colon."; Other evidencefor the laxative role of short chain fatty acid comes from a series of studiesof the effect of dietary fibre on stool composition in man by Williams andOlmsted."'' "1 They showed that stool weight correlated with short chainfatty acid output, and concluded that dietary fibre breakdown producedshort chain fatty acids which, being unabsorbed, increased faecal weight.

However, it is now clear that short chain fatty acids are absorbed fromthe colon and that the increase in faecal weight with dietary fibre is dueto physical effects of undigested fibre and bacterial mass. As colonic shortchain fatty acid levels remain more or less constant despite dietarychanges, etc, any factor increasing stool weight will increase output ofthese acids, even laxatives. The change in stool output in diarrheoa causedby carbohydrate malabsorption is now thought to be partly due to theosmotic effect of malabsorbed sugars."9 "I' Furthermore, Argenzio1ll' haspointed out that short chain fatty acid production from glucose (equation2) results in only a small increase in the theoretical osmotic pressure, ascarbon dioxide is absorbed and bicarbonate reacts with short chain fattyacids as follows:

NaHCO, + CH:COOH ±CH ,COO Na + CO2 + H2O (4)

He also points out that these acids are the most rapidlly absorbedl ions inthe colon andl that their likely contribution to an increase in osmoticpressure will be small. In addition, acetate stimulates absorption of sodiumand water from the colon at pH 614 andl a case couldl be madle for sug-gesting that absorption from the colon would be impaired in the absenceof short chain fatty acids. "The case against these acids being involxved in dliarr-hoea is strong. A

possible effect could be through changes in p1I. Where short chain fatty

771



j.com/


eptember 1981. D

ownloaded from

http://gut.bmj.com/

acids are being generatedl rapidly the colonic buffering systemn may notbe able to dleal with prodluctioni of acid and pH will fall. This may impairabsorption of salt and w ater by the colonic mucosal' an,l in acdldition, asalready dlescribedl, proniote lactate formnation (a less xxell-absorbed anion)by the bacteria. Lactate may be important in the diarrhoea of chlildhoodmalabsorption, as significant quaitities are foundl in the stools of thesechildren. In general, however, it is difficult to implicate short chain fattyaci'ds directly in either the control of faecal output or the genesis ofdiarrhoea in man.

1 N E R G Y N l I A B1 0 1 1 S NIThe contribution of colonic fermentation in mnan to overall energy balanceis unknown, but presents an intriguing problenm. If, as was discussedearlier, the only substrate for fermenitationi in the colon is undligesteddietary carbohydrate then about 20 g/lday is broken dow\n in man eatinga United Kingdoom type of dliet. If we assume that about 70°' of the poteni-tial energy is available for absorptioni as short chain fatty acidls the netenergy yield would be about 224 kJ (56 kcal), an insignificaint anmount. If,however, the energy requirement for growth and maintenlance of bacteriain the colon are considered, then the potential energy available for ab-sorption (as short chain fatty aci(ds) would be 600-700 kJ or 70' of normalenergy intake (assuming a 250/ vield of bacterial cells fronm carbohydratemetabolised anaerobically). This represents considerably more carbo-hydrate breakdown than is currently believed to occur in the colon insubjects eating typical Western diets atind needs to be furthler substantilted.The colon has not previously been thouglht of as a major- factor in humanenergy metabolism. Much larger aniounts of fer-mcntable carbohydrateenter the colon of people living on high fibre diets, those w ithi Ialabsorp-tion, and in subjects who have had a jejunoileal bypass for obesity. Hereshort chain fatty acidls may be mnaking a significant contribution to energybalance."'

N 1 T RO G 1 N A N D ANI NI 0 N I A NI1 T A 13 0 1, I S NICarbohydrate breakdown in the colon has important implications fornitrogen metabolism. When dietary fibre intakes increase in man facecalnitrogen excretion increases,` '' an effect not seen with changes inprotein intake.'! '" Faecal N was at one time thought to be either en-dlogenous or 'metabolic' in origin (from gastrointestinal secretions andsloughed epitlhelial cells, etc) or from undigestedl dlietary proteini. It is clearfrom animal andl human studlies, however, that much of faecal nitrogen ispresent in bacteria' " andl that when 'undligestible polysaccharides' arefedl the increase in N excr-etion is dlue to an increase in faecal bacteria.""Similalr findlings have now been reported in main.]"t'The mnain non-protein source of nitrogeni in the colon is anmmoniacle-

rived from the hydrolysis of urea. In the albsence of active colonic fer-mentation muchi of this amnionia is reabsorbed to be reconiverted to ureain the liver, ac metabolic cycle which would seem to serve no useful pur-pose. Given a suitable ener-gy source such lias undigestedl carbohydrate,'bacteria utilise anmmonia nitrogen for protein synthesis2 aind bacterial

772 Cumm,ng



j.com/


eptember 1981. D

ownloaded from

http://gut.bmj.com/

Slhort chlain fatty acids in the hiuman colon

growth is stimulated. In the horse"" the amount of protein nitrogen in thecolon increases when the diet is supplemented with cellulose and ammoniaconcentration falls. Vince et al."1' have shown, using human colonicmicroflora incubated anaerobically in vitro, that ammonia concentrationsfall in the incubation system when an energy source such as lactulose isadded, suggesting ammonia use for bacterial protein synthesis. In therat,"'-' blood urea is an important determinant of ammonia levels in thecaecum and when plant cell-wall carbohydrate or lactose is added to caecalcontents in vivo the production of short chain fatty acid increases andconcurrently there is an increase in blood urea uptake, a fall in ammoniaconcentration in the caecum, and an increase in the portal arteriovenousdifference in ammonia concentration. In patients with hepatic cirrhosisWeber" observed urea production to fall and faecal N excretion to risewhen his subjects were given lactulose, suggesting a reduction in ammoniarecycling through the portal system and an increase in protein synthesisby the colonic microflora. All this evidence suggests therefore that micro-bial fermentation of carbohydrate reroutes nitrogen into bacterial proteinsynthesis and in so doing lowers colonic and portal venous blood ammonialevels.

Production of short chain fatty acids in the colon may also affectdirectly ammonia absorption. Carbohydrate fermentation lowers intra-luminal pH,72 which would tend to trap ammonium ion in the lumen, asit is thought to cross the mucosa in the unionised (NH:1) form.""' Con-versely, the effect of these acids on bicarbonate secretion would favourammonia absorption by coupled non-ionic diffusion as suggested byWrong."`1 7 A direct study of the effect of short chain fatty acids onammonia absorption would be worthwhile to try to clarify which ofthese possible interactions between fermentation and ammonia metabolismpredominates in man.

Visek and his colleagues have pointed out"' -'- that there could beother important implications of the relationship between the production ofshort chain fatty acids and ammonia metabolism in the colon. They haveshown that ammonia can induce changes in intestinal cells which favourtumour growth. The generation of these acids in the colon by directingnitrogen metabolism towards bacterial protein synthesis, by loweringluminal ammonia levels, and by diminishing ammonia absorption wouldall reduce this risk. Such a possible protective effect for dietary fibre in

the development of colon, and possibly other, malignancies, would fitcurrent epidemiological findings.

NI U C O S A I G R 0 W I H A N D I S 1 A S E-

There is little doubt that both the rumen epithelium and colonic mucosametabolise short chain fatty acids, especially butyrate and propionate. Inthe rumen these acids have a stimulatory effect on epithelial growth.'22 Itis quite possible that they may exert a stimulatory effect on colonicmucosa as well. Recent studies'2 of diet on colonic mucosa in rats haveshown that mucosal cell turnover declines during intravenous feeding andalso when the same liquid diet is fed orally. A chow diet, however, whichis rich in complex carbohydrate, stimulated cell turnover and increasedcolonic mucosal weight. This was attributed by the authors to the trophic

773



j.com/


eptember 1981. D

ownloaded from

http://gut.bmj.com/

effect of gastrin, serum levels of which were higher with chow than withthe liquid diet. When, however, the liquid diet was mixed with celluloseand fed, colonic mucosal growth was stimulated in the absence of a rise inserum gastrin, an effect which might well be ascribed to the presence offermentation products.

In man the colonic mucosa is known to be dependent on butyrate as anenergy source,"2 particularly the distal colon. When Roediger'2' measuredbutyrate uptake in the colonic mucosal cells of patients with ulcerativecolitis, he showed that it was reduced when compared with normal con-trols and suggested that failure to utilise butyrate could be a primary fac-tor in the aetiology of ulcerative colitis. If this is the case, then it is equallypossible that low butyrate production in the colon could precipitate colitisin susceptible individuals.

OT HI: R I1MPI1 CAT IONSShort chain fatty acids are known to have an antibacterial effect.'2 '127Production in the colon of these acids is one mechanism which preventsthe establishment of pathogenic bacteria, such as salmonella species, al-though the low redox potential maintained by the predominantly anaerobicflora which fermentation supports may also contribute to this effect.Other consequences of a lowering of pH include a favourable effect on

vitamin K absorption,'2 the stimulation of mucus production,!", and ofmagnesium absorption. 12

ConclusionThere are close parallels between the fermentative processes which go onin the rumen, caecum, and colon of herbivorous animals and colonicmetabolisnm in man. Short chain fatty acids, which are the main end-product of carbohydrate breakdlown in these organs, exert a controllinginfluentce on intraluminal events, absorption, mucosal metabolism, and areacceptedl as such by animnal phvsiologists.Such recognition has yet to be given to this aspect of colonic function

in man, and many studies of colonic metabolism have failed to take ac-count of the possible effect of short chain fatty acids. A great deal stillneeds to be learnt about these acids in the human colon-in particular, theoverall amount produced each day. the main substrates for fermentation,the effect of diet, the molecular form in which they are absorbed, theircontribution to energy metabolism, and their interaction with a wide rangeof other colonic events. Such knowledge should yield important infor-mation which may be relevant to the aetiology of colonic disor(ders whichare so prevalent in the human species.My thanks are due especially to Professor Budl Tennant of Cornell Veter-inary Medical School and to Dr David Saunders of the Medical Schoolin Seattle for helpful comments during the preparation of this progressreport.

Dunn Clinical Nutrition Centre .I H CUMNMINGSA ddenbrooke.s HospitalTrumpingwon Stu-eetCarnbt-idgc Received for publication 16 February 1981

774 Cummings



j.com/


eptember 1981. D

ownloaded from

http://gut.bmj.com/

Short chain fatty acids in the human colon 775

References

'Brieger L. Ueber die fluchtigen Bestandtheile der Menschlicher Excremente. JPrakt Chlemie (Neue Folge) 1878; 17:124-38.2Grove EW, Olmsted WH, Keonig K. The effect of diet and catharsis on the lowervolatile fatty acids in the stools of normal men. J Biol Chlem 1929; 85:127-36;:'Goiffon R, Goiffon B, Fron G. Contributions a l'etude des electrolytes des selles.IV. Essai d'ionogramme fecal. Gastroenterologia (Basel) 1961; 69:326-32.

4Rubinstein R, Howard AV, Wrong OM. In vivo dialysis of faeces as a method ofstool analysis. IV. The organic anion component. Clin Sci 1969; 37:549-64.5Argenzio RA, Southworth M, Stevens CE. Sites of organic acid production andabsorption in the equine gastrointestinal tract. Am J Phyysiol 1974; 226:1043-50."IHenning SJ, Hird FJR. Concentration and metabolism of volatile fatty acids in thefermentative organs of 2 species of kangaroo and the guinea-pig. Br J Nutr 1970;24:146-55.7Henning SJ, Hird FJ. Transport of acetate and butyrate in the hind-gut of rabbits.Biochem J 1972; 130:791-6.`Argenzio RA, Southworth M. Sites of organic acid production and absorption ingastrointestinal tract of the pig. Am J Physiol 1974; 228:454-60."Banta CA, Clemens ET, Krinsky MM, Sheffy BE. Sites of organic acid productionand patterns of digesta movement in the gastrointestinal tract of dogs. J Nutr 1979;109:1592-1600.

"Prins RA. In Clarke RTJ, Bauchop T, eds. Microbial ecology of the gut. London:Academic Press, 1977; 73.

"McBee RH. In: Clarke RTJ, Bauchop T, eds. Microbial ecology of the gut. London:Academic Press, 1977; 185.

'2Miller TL, Wolin MJ. Fermentations by saccharolytic intestinal bacteria. Am JClin Nutr 1979; 32:164-72.

' 'Bryant MP. Nutritional features and ecology of predominant anerobic bacteriaof the intestinal tract. Arn J Clin Nutr 1974; 27:1313-20.

'1"Moore WEC, Cato EP, Holdeman LV. Some current concepts in intestinal bac-teriology. Am J Clin Nutri 1978; 31:533-42.

'"Salvers AA, Vercellotti JR, West SEH, Wilkins TD. Fermentation of mucin andplant polysaccharides bv strains of Bacteroides from the human colon. Appl En-viron Microbiol 1977; 33:319-22.

";Salvers AA. Energy sources of major fermentative anaerobes. Am J Clin Nutr1979; 32:158-63.

'7Remesy C, Demigne C. Partition and absorption of volatile fatty acids in thealimentary canal of the rat. Ann Rech Vet 1976; 7:39-55.

I vHungate RE. The rumen and its microbes. New York: Academic Press, 1966."Cummings JH, Hill MJ, Bone ES, Branch WJ, Jenkins DJA. The effect of meatprotein and dietary fiber on colonic function and metabolism. Part 11. Bacterialmetabolites in feces and urine. Am J Clin Nutr 1979; 32:20)94-101.

20Bjork JT, Soergel KH, Wood CM. The composition of 'free' stool water. Gastro-enterology 1976; 70:A6/864.

'2Zijlstra JB, Beukema J, Wolthers BG, Byrne BM, Groen A, Donkert L. Pretreat-ment methods prior to gas chromatographic analysis of volatile fatty acids fromfaecal samples. Clin Chlim Acta 1977; 78:243-50.'2Sambrook IE. Studies on digestion and absorption in the intestines of growingpigs. 8. Measurements of the flow of total lipid, acid-detergent fibre and volatilefatty acids. Br J Nutr 1979; 42:279-87.

2:'Owens FN, Isaacson HR. Ruminal microbial yields: factors influencing synthesisand bypass. Fed Proc 1977; 36:198-202.

2'Bergman EN. In: McDonald IW. Warner ACI, eds. Digestion and metabolism intlie ruminant. Australia: University of New England Publishing Unit, 1975; 292.

25Hungate RE. In: Code CF, ed. Handbook of ply.'siology, Section 6, Vol V, 1968;2725.

26Dunlop RH. In: Brandley CA, Cornelius CE, eds. Advances in veterinarY scienceand comparative medicine. New York: Academic Press, Vol 16, 1972; 259.

27Torres-Pinedo R, Lavastida M, Rivera CL, Rodriguez H, Ortiz A. Studies on infantdiarrhoea. I. A comparison of the effects of milk feeding and intravenous therapy



j.com/


eptember 1981. D

ownloaded from

http://gut.bmj.com/

776 Cummings

upon the composition and volume of the stool and urine. J Clin Invest 1966; 45:469-80.

2sWeijers HA, Van De Kamer JH. Aetiologv and diagnosis of fermentative diarrhoea.A(ta Paediatr 1963; 52:329 37.

"!)Weijers HA, Van De Kamer JH, Dicke WK, jisseling J. Diarrhoea caused bydeficiency of sugar splitting enzymes. I. Ad(ta Pediatr (Stockli) 1961; 50:55 -71.

:"'Oh MS, Phelps KR, Traube M, Barbosa-Salvidar JL, Boxhill C, Carroll HJ. D-Lactic acidosis in a man with the short bowel syndrome. N Engl J Med 1979; 301:249-52.

:"lMcBee RH. Metabolic contribution of the cecal flora. Am J Clin Nutr 1970; 23:1514-8.

32Smith CJ, Bryant MP. Introduction to metabolic activities of intestinal bacteria.Am J Clin Nutr 1979; 32:149-57.

:i:ILevitt MD. Production and excretion of hydrogen gas in man. N Engl J Med1969; 281:122-7.

34Bond JH, Engel RR, Levitt MD. Factors influencing pulmonary methane excretionin man. J Exp Med 1971; 133:572-88.

:'"Levitt MD. Volume and composition of human intestinal gas determined by meansof an intestinal washout technique. N Engl J Med 1971; 184:1394-8.

e"'Hickey GA, Murphy EL, Calloway DH. Intestinal gas-production following in-gestion of commercial wheat cereals and milling fractions. Cereal Cl/em 1972; 49:276-83.

:37Pitt P, Bruijn KM De, Beeching MF, Goldberg E, Blendis LM. Studies on breathmethane: the effect of ethnic origins'and lactulose. Gut 1980; 21:951-9.

3SHaines A, Metz G, Dilawari J, Blendis L, Wiggins HS. Breath methane in patientswith cancer of the large bowel. Lancet 1977; 2:481-3.

:3¶'Cummings JH. Dietary fibre. Br Med Bull 1981; 37:65-70.4"Bond JH, Currier BE, Buchwald H, Levitt MD. Cclonic conservation of mal-absorbed carbohydrate. Gastroenterololgy 1980; 78:444- 7.

41McNeil NI, Bingham SA, Grant AM, Cummings JH. Ileostomy function at home.Gut 1977; 18:A958.

42Bingham S, Cummings JH. In: Spiller GA, Kay RM, eds. Medi(cal aspects ofdietary fibre. New York: Plenuim Press, 1980; 261- 84.

4:Southgate DAT, Bingham S, Robertson J. Dietary fibre in the lBritish diet. Nature1978; 274:51-2.

44Hespell RB. Efficiency of growth by ruminal bacteria. Fed Proc 1979; 38:27)7 -12.45Hespell RB, Bryant MP. Efficiency of rumen microbial growth: influence of some

theoretical and experimental factors on YATP. J Anim S(i 1979; 49:1640 59.4GIsaacson HR, Hinds FC, Bryant MP, Owens FN. Efficiency of energy utilization bymixed rumen bacteria in continuous culture. J Dairy Sci 1975; 58:1645 9.

4 7Stephen AM, Cummings JH. The microbial contribution to human faecal mass.J Med Microbiol 1980; 13:45--56.

4SStephen AM, Cummings JH. Mechanism of action of dietary fibre in the humancolon. Nature 1980; 284:283-4.

49Stephen AM. Dietary fibre and lhuman (colonic function. PhD Thesis, Universityof Cambridge 1980.

51'Anderson I, Levine A, Levitt MD. Use of breath H excretion to study absorp-tion of wheat flour. Ga.stroenterology 1980; 78:1131.

51'Cummings JH, Hill MJ, Jenkins DJA, Pearson JR, Wiggins HS. Changes in fecalcomposition and colonic function, due to cereal fiber. Amn J Ckn Nutr 1976; 29:1468-73.

rX2Umesaki Y, Yajima T, Yokokura T, Mutai M. Effect of organic acid abosrptioll onbicarbonate transport in rat colon. Pfiugcrs Arc/i 1979; 379:43 7.

*,:'Argenzio RA, Whipp SC. Interrelationship of soodiunm, chlori(de, bicarbonate anidacetate transport by the colon of the pig. J P/ivsiol 1979; 295:315 Xl.

54Argenzio RA, Miller N, Von Engelhardt W. Effect of volatile fatty acids on waterand ion absorption from the goat colon. Am J P/i.siol 1975; 229:997 1(M)2.

55Daw.son AM, Holdsworth CD, Webb J. Absorption of short chain fatty acids inman. Proc Soc Exp Biol Med 1964; 177:97-1 00.

5'"McNeil NI, Cummings JH, James WPT. Short chain fatty aci(l absorption by thehuman large intestine. Gut 1978; 19:819- 22.



j.com/


eptember 1981. D

ownloaded from

http://gut.bmj.com/

Short chain fatty acids in the human colon 777

,7Ruppin H, Bar-Meir S, Soergel KH, Wood CM, Schmitt MG. Absorptioni of shortchain fatty acids by the colon. (Gastroenterology 1980; 78:15(X)-7.

55McNeil NI. Short chain fatty acid absorption in the human large intestine. ThesisUniversity of Cambridge 1980.

'"Argenzio RA, Southworth M, Lowe JE, Stevens CE. Interrelationiship of Na, HCO3and volatile fatty acid transport by equine large intestine. Am J Physiol 1977; 233:E469-78.

';"Stevens CE, Stettler BK. Factors affecting the transport of volatile fatty acidsacross rumen epithelium. Am J Physiol 1966a; 210:365-72.

"'lStevens CE, Stettler BK. Transport of fatty acid mixtures across rumen epitheliunm.Am J Physiol 1966b; 211:264-71.

6;2Turnberg LA, Bieberdorf FS, Morawski SG, Fordtran JS. Interrelationship ofchloride, bicarbonate, sodium and hydrogen transport in the human ileum. J ClinInvest 1970); 49:557--67.

G':1Hawker P7, Mashiter KE, Turnberg LA. Mechanisms of transport of Na, Cl andK in the human colon. Gastroenterology 1978; 74:1241-7.

64McNeil NI, Cummings JH, James WPT. Rectal absorption of short chain fattyacids in the absence of chloride. Gut 1979; 20:400-3.

6"Lamers JMJ. Some characteristics of monocarboxylic acid transfer across the cellmembrane of epithelial cells from rat small intestine. Biochfim Biophys Acta 1975;413:265-76.

'1';Stevens CE. In: Phillipson AT, ed. IPhy.siology of dige.stion ancd mtietabolismtz in thleruminlant. Newvcastle: Oriel Press, 1970; 101.

67Stevens CE. In: Ussing IH, Thorn NA, eds. Iransport mnechanismns in epithelia.Copenhagen: Munksgaard, 1973; 404.

0;&Carter MJ, Parsons DS. Carbonic anhydrase activity of mucosa of small intestinleand colonl. Nature (Lond) 1968; 219:176-7.

";'Carter MJ, Parsoni DS. The is.oelnzymes of carboiic anhydrase: tissue, subcellulardistribution and functional significance swith particular referenice to the intestinaltract. J Phy.siol 1971; 215:71-94.

7'Carter MJ, Parsons DS. The isoenzymes of carbonic anhydrase: kinetic propertieswith particular reference to the functions of the G-I tract. J Physiol 1972; 220:465-78.

71Parsons DS. In: Bolis L, Schmidt-Nielson K, Maddress SHP, eds. Comzparativephysiology. Amsterdam: North Holland Publishing Co, 1973; 417.

7TBown RL, Gibson JA, Sladeni GE, Hicks B, Dasson AM. Effects of lactulose andother laxatives on ileal and colonic pH as measured by a radiotelemetry device.Gut 1974; 15:999-1004.

7:Schmitt MG, Soergel KH, Wood CM. Absorption of short chain fatty acids fromthe human jejunum. Gaastroenterology 1976; 70:211-5.

71Schmitt MG, Soergel KH, Wood CM, Steff JJ. Absorption of short-chain fattyacids from the human ileum. Am J Dige.st Dis 1977; 22:34(-7.

7-Sallee VL, Dietschy JM. Determinants of intestinal mucosal uptake of short andmedium chain fatty acids and alcohols. J Lipid Res 1973; 14:475-84.

7'McNeil NI, Ling KLF. The mucosal surface pH of the large intestine. Gavtro-enterology 1980; 78:1220.

T7Hogben CAM, Tocco DJ, Brodie BB et al. On the mechaniism of intestinal ab-sorption of drugs. J Pharmnacol Exp Thierap 1959; 125:275-282.

>Jackson MJ, Williamson AM, Dombro".ski WA et al. Intestinal transport of weakelectrolytes determination of influx at the luminal surface. J Gen Physiol 1978;71:301-27.

7"'Crump MH, Argenzio RA, Whipp SC. Effects of acetate on absorptionl of soluteand water from the pig colon. Ai J Vet Re.s 1980; 41:1565-8.

"Bown RL, Sladen G.E, Rousseau B, Gibson JA, Clark ML, Dawson AM. A studyof water and electrolyte transport by the excluded human colon. Clin Sci 1972;43:891-90)2.

81Jackson MJ. In: Smyth DH, ed. Biomembranes, vol 43: Intestinal absorption. NewYork: Plenum Press, 1974; 673.

S2Roediger WEW. Role of anaerobic bacteria in the metabolic welfare of the colonicmucosa in man. Cut 1980; 21:793-8.



j.com/


eptember 1981. D

ownloaded from

http://gut.bmj.com/

778 Cummings

s:'Cochrane GC. A review of the analysis of free fatty acids (C2-C,). J ChromatogrSci 1975; 13:440-7.'Grady GF, Duhamel RC, Moore EW. Active transport of sodium by human colonin vitro'. Gastroenterology 1970; 59:583-8.Archampong EQ, Harris J, Clark CG. The absorption and secretion of water andelectrolyte.s across the healthy and the diseased human colonic mucosa measuredin vitro. Gut 1972; 13:880-6.

`'ERoediger WEW, Truelove SC. Method of preparing isolated colonic epithelialcells (colonocytes) for metabolic studies. Gut 1979; 20:484-8.7Levitan R, Fordtran JS, Burrowss BA, Ingelhinger FJ. Water and salt absorptioll inthe human colon. J Clin Invest 41:1754-9.

,'Devroede GJ, Philips SF. Studies of the perfusion technique for colonic ab.iorption.Gastroenterology 1969; 56:92-100.

9!Harris J Shields R. Absorption and secretion of water and electrolytes by theintact human colon in diffuse untreated proctocolitis. Gut 1970; 11:27-33.

"(Devroede GJ. Regional differences in rates of insorption of sodium and water fromthe human large intestine. Can J Plh.iul Plarinacol 1971; 49:1023-9.1Edmonds CJ. Absorption of sodium and water by human rectum measured by adialysis method. Gut 1971; 12:356-62.

"2Williams RD, Olmsted WH. The effect of cellulose, hemicellulose and lignin onthe weight of stool. A contribution to the study of laxation in man. J Nutr 1936;11:433-49.

93Bond JH, Levitt MD. Fate of soluble carbohvdrate in the colon of rats and man.J Clin invest 1976; 57:1158-64.

"4Bustos-Fernandez LB, Gonzalez E, Marzi A, Paolo MIL. Faecal acidorrhoea.N Engl J Med 1971; 284:295-8.

'`McMichael HB, Webb J, Dawson AM. Lactose deficiency in adults. Lancet 1965;1:717-20.

"';Bustos-Fernandez LB, Gonzalez E, Paolo MIL, Celemer D, de Furuya KO. Organicanions induce colonic secretion. Am J Dige.st Di.s 1976; 21:329-32.

"7Williams RD, Olmsted WH. The manner in which food controls the bulk of faeces.Ann Intern Med 1936; 10:717-27.

"9Launiala K. The mechanisms of diarrhoea in congenital disaccharide malabsorption.Acta Paediatr Scand 1968; 57:425-32.

99Launiala K. The effect of malabsorbed sucrose or mannitol-induced acceleratedtransit on absorption in the human small intestinle. Scand J Ga.stroenterol 1969;4:25-32.

100Christopher NL, Bayless TM. Role of the small bowel and colon in lactose induceddiarrhoea. Gastroenterology 1971; 60:845-52.

"11Argenzio RA. Physiology of diarrhoea-large intestine. J Antm Vet Med A.ssoc1978; 173:667-72.

'1'Rousseau B, Sladen GC. Effect of luminal pH on the absorption of water, sodiumand chloride by the rat intestine in vivo. Biocheem BiophYs Acta 1971; 233:591--3.

"':iCummings JH, Southgate DAT, Branch W, Wiggins HS, Houston H, JenkinsDJA, Jivraj T, Hill MJ. The digestion of pectini in the human gut and its effect oncalcium absorption and large bowel function. Br J Nutr 1979; 41:477-185.

"04Southgate DAT, Durnin JVGA. Calorie conversion factors. An experimental re-assessment of the factors used in the calculation of the energy! value of humandiets. Br J Nutr 1970f; 24:517-35.

"0.5Gibson JA, Park NJ, Sladen GE, Dasson AN4. The role of the coloni in ureametabolism in man. Clin Sci Mol Med 1976; 50:51 -9.

"'mMacNeal WJ, Latzer LL, Kerr JE. The faecal bacteria of heaclthy menl. Part 1.Introduction aand direct quantitatixe observations. J Inject Di.s 1909; 6:123-69.

"07Mason VC. Some observations on the distribution aind origin of nitrogenl in sheepfaeces. J Agri S(ci (Canmbridge) 1969; 73:99-1 1 1.

'0VMason VC, Palmer R. The influenlce of bacterial activity in the alimentary canalof rats on faeccal nitrogeni excretioni. Act A gric .Scand 1973; 23:141- 5).

"'Steplhcn AM, Cummings JH. The influenice of dietary fibre on faecal nitrogenexcretion in man. Proc Nutri Soc 1979; 38:141A.

''Wooten JF, Argenzio RA. Nitrogen utilization within equitle large intestine. A tiJ Physiol 1975; 229:10)62-7.



j.com/


eptember 1981. D

ownloaded from

http://gut.bmj.com/

Slhort chain fatty acids in the hu1uman colon 779

"'Vince A. Killingley NI, Wvrong OM. Effect of lactulose on ammonia productionin a fecal incubatioll system. (,astro(eiterolog,y 1978; 74:544-9.l)emiul 'C. Remesx C. Urea recvcline! and ammonia absorption in vivo in theligestive tract of the rat. Annl lliol Aniin Biochfin IJiopllys 1979; 19:929-35.

"3Weber FL. The effect of lactulose on urea metabolism and nitrogen excretion incirrhotic patients. Ga.stroenterology 1979; 78:518-23.

'I'Down PF. Aostini L. Murison J, Wrong OM. The interrelations of faecal am-moniia pH and bicarbonate: evidence of colonic absorption of ammonia by non-ioniic diffuslion. C('iti Si 1972; 43:101- 14.1'Poxn RL. Gibsoni JA. Fentoni JC3. Sneddenl \. Clark ML. Sladen GE. Ammoniaandl urea transport by the excluded humani colon. C(in Sci Mol Med 1975; 48:279--87.\rong O)1M. The role of humani colon in homeostasis. IBriti.sl l'o%stgraduate Feder-atiori '.(ihtatifi(ha.sis of inedicine ainnual rev,iewt.s 1971; 192-215.

1'1-ront 0MN. Nitrogen metabolism in the gut. A in J C(lin NuVtr 1978; 31:1587-93.''Vise;k Wi. Effects of urea hlJdrolysis on cell life-span and metabolism. Fed Proc

1972; 31:1178 93.1''Visek Wi. Diet and cell growth modulation by ammonia. A/I1 J C lin Nutr 1978;

31:S216 S20.12Visek WJ. Clinton SK, Truex CR. Nutrition and experimental carcinogeniesis.

Cornell Vet 1978; 68:3- -39.2'ITopping DC, Visek WJ. Nitrogen intake and tumorigenesis in rats injected with

I.2-(liniethlhvdrazine. J Nutr 1976; 106:1 583-90.I22SadEr EG, Warner RG, Harrison HN. Loosli JK. The stimulatorv effect ofsodiumn butvrate and sodium propionate on the development of rumen mucosa inthe vounc calf. J Dairv S(ri 1959; 42:160)0 5.R\an (GP, Dudrick SJ, Copelantd FM. Johnson LR. Effect of various diets oncolomic crowth in rats. (Gastroeznterology 1979; 77:658-63.

'I2Roediger WEW. The colonic epithelium in ulcerative colitis: an energy-deficiencydisease'? l.anct 98X); 2:71 2 5.

12-Hentges DJ. Inhibition of shigella flexneri by the normal intestinal flora. I. Mech-anisms of inhibition by klebsiella. J Bacteriol 1967; 93:1369-73.I2;M'vnlll GG. Antibacterial mnechanilismns of the mouse gut. 11. The role of EH andvolatile fatty acids in the normal gut. /r J Exp Pat/z!o 1963; 44:209-19.

2713)othoff M. Miller CP. Malrtini WR. Resistance of the mouse's intestinal tract toexperimental salmonella infectiotn. 1. Factors which interfere with the initiation ofinfectioni bv oral innoculation. J Exp Med 1964; 120:805 16.

L>Ho,llatlder D, Rim E Ruble PE. Vitamin K-2 colonic and ileal in vivo absorption.lile. fattv aci(is and pH effects on transport. (;astroenzterology' 1977; 72:A48/1071.Ra,vssiuier Y. Reniesv C. Magnesium absorption in the caecum of rats related tovolatile fatty aci(ds prodluctioni. Ann Reci Vet 1977; 8:1(05-1).



j.com/


eptember 1981. D

ownloaded from

http://gut.bmj.com/

progress report - gut · galactose), pentose (xylose and arabinose), and uronic acid monomers which...

Documents