Lactulose, Disaccharides and Colonic Flora Clinical Consequences

Drugs 1997 Ju n: 53 (6): 930-942 REVIEW ARTICLE 00 12-6667/97/rJ:XJO-O

Mette Rye Clausen and PerBrebech Mortensen Department of Medicine CA, Division of Gastroenterology, Rigshospitalet, University of Copenhagen, Copenhagen, Denmark

Contents Summary . 930 1. Dietary Carbohydrate . 931 1.1 Lactulose . . . . . 931 2. Colonic Fermentation and Producti on of Short-Chain Fatty Acids (SCFA) 931 2.1 Fermentation of Lactulose . 932 3. Lactulose in the Management of Constipation ...... 933 3.1 Colonic Compensation in Car bohydrate Malabsorption 933 3.2 Fermentation Capac ity Threshold: Individual Variability . 934 3.3 Does Lactulose Exert Its Laxative Effect Through the Osmotic Activity of Unferme nted Carbohydrate? . 935 3.4 Capac ity of the Colon to Absorb Fluid . 936 3.5 Possible Me chanisms of Action of Lactulose in the Treatment of Constipation. 936 4. Lac tulose in the Management of Hepatic Encephalopathy ...... 937 4.1 Effects of Lac tulose on Ammonia and Nitrogen Metabolism ...... 937 4.2 Role of Short- and Medium-Chain Fatty Acids in Hepatic Encephalopathy . 938 4.3 Effect of Lactulose on Colon ic SCFA Production in Hepatic Encephalopathy 939 4.4 Possible Mechanisms of Action of Lactulose in the Treatment of Hepatic Encephalopathy 940 5. Conclusions...... 940

Summary Lactulose is one of the mos t freq uently utilised age nts in the treatment of constipation and hepatic ence pha lopathy because of its efficacy and good safe ty profile. The key to understanding the possible modes ofaction by which lactulose ac hieves its therapeutic effects in these diso rders lies in certain pharmacological phe nomena: (a) lactulose is a sy nthetic disaccharide that does not occur naturally; (b) there is no disaccharidase on the microvillus membrane of enterocytes in the human small intesti ne that hydrolyses lactulose; and (c) lactulose is not absorbed from the small intestine. Thus, the primary site of action is the colon in which lactulose is readi ly fermented by the co lonic bacterial tlora with the produ ction ofshort-chain fatty acid s and various gases. T he purpose ofthis review is to focu s on some pertinent basic aspects of the clinical pharmacology of lactulose and to discuss the possible mec hanisms by which lactulose benefits patients with co n stipation and hepatic encephalopat hy. Disaccharides and Colonic Flora 931

1. Dietary Carbohydrate human small intestine.Pv' Once in the caecum, lactulose is a fully fermentable carbohydrate. Dietary carbohydrates may be divided into Lactulose is widely considered to be an effective monosaccharides, disaccharides, oligosaccharides agent in the management of constipation and he and polysaccharides. Their physiological proper patic encephalopathy. It is the purpose of this re ties and health benefits depend on the site, rate and view to discuss the interaction of lactulose with the extent of their digestion or fermentation in the hu bacterial flora of the colon and the possible modes man gastrointestinal tract. The polysaccharide ofaction by which it achieves its therapeutic effects fraction of plant cell walls (dietary fibre) and some in patients with constipation and hepatic encepha forms of starch (resistant starch) are resistant to lopathy. The review deals almost exclusively with hydrolysis in the small intestine. These carbohy lactulose. The principles expressed, however, are drate polymers as well as indigestible oligosaccha applicable to any malabsorbed and fermentable di rides, e.g. stacchyose and raffinose, enter the colon saccharide, e.g. lactitol, a disaccharide analogue of through the ileocaecal valve to be fermented by the lactulose for which there is also no disaccharidase colonic bacterial flora. Monosaccharide and disac on the microvillus membrane of enterocytes, or charide sugars are absorbed as they pass through lactose when it is administered to lactase-deficient the healthy small intestine and only become avail individuals. able to the colonic flora as a consequence of defi ciencies in the transport system, or, more usually, specific disaccharidase deficiencies, e.g. lactase 2. Colonic Fermentation and Production deficiency. Physiological carbohydrate malabsorp ofShort-Chain Fatty Acids (SCFA) tion, that is, the fermentable carbohydrate entering Fermentation, the process whereby the micro the colon in healthy individuals, is in the range of bial population that inhabits the human colon 30-60 g/day in Western communities.l!l This load breaks down carbohydrates in order to obtain en is normally fermented by the vast numbers of co ergy for maintenance of cellular function and lonic bacteria. growth, is an important component of normal colonic activity (fig. I). Short-chain fatty acids 1.1 Lactulose (SCFA) and the gases H2, CO 2 and, in some indi Lactulose is a synthetic disaccharide composed viduals, CH4, are the principal end products of car of the monosaccharides galactose and fructose [p bohydrate fermentation.l!' The major SCFA found ( 1-4)-galactosido-fructose]. The disaccharidases in the human colon are the C2 (acetate), C3 (pro that split naturally occurring disaccharides into pionate) and C4 (butyrate) members of the ali their hexose moieties do not include a lactulase, phatic monocarboxylic acid series. Formate (C I), and lactulose is not metabolised or absorbed in the valerate (C5), hexanoate (C6), and the branched-

Dietary mono·, di- and oligosaccharides

Bacterial enzymes IPolysaccharides 1------ICarbohydrates Ir---- SCFA,H2,C02,CH.and increased bacterial mass

Fig. 1. Bacterial fermentation in the human colon . Abbrevialion: SCFA = short-chain fatty acids .

<0Adis International Limited. Ail rights reserved. Drugs 1997 Jun; 53 (6) 932 Clausell & Mor!('//s('//

chain fatty acids isobutyrate (iC4) and isovalerate 3-fo ld by the additio n of 100 mmol/L lactulose or (iC5) , whi ch are produced during amino acid fer its mon osacch ari de co nstituents, ga lac tose and ment ation,14's1are also present in the colon, though fructose, during inc ubation for 6 hours, secondary in smaller quantities. Oth er orga nic ac ids such as to an isolated 4- to 6-fo ld increase in the synthes is lactate and succ inate are ferment ation intermedi of acetate. In patient s undergoing elective chole ates. Th ey can be further metaboli sed to SC FA, and cystectomy, caecal insti llatio n of lactulose ca used do not usually accumulate to any extent, except an immediate increa se in portal plasma SCFA during rapid carbohydrate breakd own. P"! co nce ntrations, principally acetate; peak co ncen Th e SCFA are weak acid s and their pKa values trations were achieve d in 15 to 45 minutes, co n are very similar, rangin g from 4.76 to 4.87. Thu s, firming efficient fermen tatio n of lactulose and ab more than 95% of the SCFA are present in their sorption of the resultant SCFAP ol di ssociated form at the phy siological pH of the co Th e various gases ge nerated in the co lon are ei lon (pH 6-8) . Total co ncentrations of SCFA are low ther absorbed and ex pired in breath or expelled as in the small intestine, but in the co lon and faeces flatu s, and excess gas production may ca use bor SCFAco nstitute the predominant ani ons at concen borygm i and flatul ence. Sin ce H2 produced in the tration s ranging from about 80 to 130 mmol/Lp ·IOI human bod y deri ves entirely from bacter ial fer Concentrati ons are highest in the caecum, where ment ation of nonabsorbed carb oh ydrate, breath H2 substra te avai lability is grea tes t, and fall progres exc retion in the abse nce of small intestinal bacte sive ly towards the distal co lon. By co ntras t, pH is rial ove rgrowth ca n be used as a marker of carbo lowest in the right colon and rises in the distal hydrate rnalabsorpt ion .F! ' Th e extent of malab bowel.l?' SC FA are avidly abso rbed by the co lonic sorption may be quantified by co mparing breath H2 epithelium,II I.1 21 and used as a source of energy, excretion following a test dose of lactulose with either locall y (co lonic mucosa)113.151 or syste mi tha t after ingestion of a different carbohydrate.l->' ca lly (liver, peripheral tissues).' 161 In the colon, ab Fermentation of a readily fermen tab le carbohy sorpt ion ofSCFA is accompanied by luminal bicar drate result s in acidification of co lonic co ntent s. bon ate acc umulatio n and increased sodium and Using a pH-sensitive rad iotelem etry de vice, Bown wate r absorption.t'I -P!Thu s, the micro bia l process et aL/231 showed that lactul ose 30 to 40g daily de co nve rts unabsorb able materi al to rapidly absorbed creased the pH of the right co lon fro m control lev SCFA, thereby redu cing the effective osmo tic pres els of 6.0 to 4.8 5. As the intes tinal co ntent reached sure of the colonic co ntents, enhanci ng water and the left co lon and rectum, the median pH rose, electrolyte abso rption, and sa lvaging calories whi ch wo uld otherwise be lost in the faeca l strea m. probably as a res ult of SCFA absorpt ion and bicar bonate sec retion. Th e effect of lactulose on faeca l pH depend s on the dose employed, and the drop in 2.1 Fermentation of Lactu lose pH is usually maint ain ed throughout the co lon and Lactul ose, after arriving in the colon, is digested the stool when larger doses of lactul ose are ad min by bacteri al disaccharidases and metabolised to or istered.124.261 ga nic aci ds, mainl y acetate and lactate, H2 and Alth ough co nsistent changes in the relative pro C02.1 61 A variety of studies have show n that lactu portion s or in the abso lute numbers of differe nt lose is usuall y vigorously fermented . In an ill vitro bacterial species have been difficul t to ide ntify in ex perime nt using human faecal homogen ates, patient s treated with lactul ose ,1 27-291this co mpou nd Vince et aLI171showed that co mplete fermen tatio n does appear to have a significant effect on the me of 10 giL (30 mmol/L ) of lactul ose occurred within tabolic activities of the microb es. In healthy volun 6 hours. By use of a similar tech nique, Mortensen teers, faeca l levels of ~ - g al ac t o s id a s e activity in et aLII X.191 showed that the total product ion of creased significantly after an 8-da y period of a SCFA in faecal homoge nates was increased 2- to non diarrhoeogeni c dose of lactulose (20g twice

dailyj.!'" Co mpared with day I, caecal co ntents edly enhances net sodi um and water abso rptio n sampled on day 8 afte r lactul ose administ ration ex and bicarb onate secretion into the lumen. Thu s, the hibited a mar ked fall in pH, a faster disappearance intracolonic fer mentation of nonabsorbed suga rs of lactulose and its co nstitutive hexoses, and in into readily absorbable SCFA reduces the osmotic creased co ncentrations of SCFA and lactate.l'" The load , and suppresses or mit igates the diarrhoea co lonic microflora therefore are able to adapt to which would have resulted from unm odified ileal ferment lactul ose, and the more efficie nt bacteri al output. fermentation of lactul ose modifies the diarrhoea indu ced by a larger load of this sugar.P'!' 3.1 Colonic Compensation in Th e colonic bacteria use ferme ntable carbohy Carbohydrate Ma labsorption drate, e.g . lactul ose, as a so urce of ene rgy, thu s prom oting their growth and multiplication. P!' It is Th e protective role of coloni c fermentation in bel ieved that bacteri al growth ass imilates the am reducing the severity of carbohydrate-induced diar moni a produ ced by degradation and putrefaction rhoea has been demonstrated in studies co mparing of ingested protein s as a source of nitrogen , and the output of faecal water in response to a non faecal nitrogen exc retion increas es 2- to 4-fold af ferme ntable os mo tic load with a load of ferm ent ter lactul ose administration.tv-P! able carbohydrate. Hammer et al.1391co mpared the diarrh oea res ulting fro m increasin g iso -os mo lar 3. Lactulose in the Management loads of the nonabsorbable, nonferment able and of Constipation electrically neut ral co mpound polyeth ylene glyco l (PE G) and lactulose in healthy individuals. In Lactul ose is frequently used as an effective lax creasing osmotic loads of PEG caused a near-li near ative in the management ofconstipation.P"! Based increase in stool wa ter output. At any load, on the absence of its respecti ve disaccharidase, the lactulose ind uced less diarrhoea than did PEG. rationale for the use of lactulose as a laxative was With low (45 g/day) or moderate (95 g/day ) doses to indu ce disacch aride malab sorption, a we ll of lact ulose, stool water losses were redu ced by as known cause of diarrh oea. In spite ofthe freq uency much as 600 g/day (co mpared with equimolar os and importance of ca rbo hydrate-i nduce d diar motic loads of PEG). As lactulose doses of >45 rhoea, its pathophysiology is not full y understood . g/day were ingested, faecal ca rbohydrate exc retion Until relative ly rece ntly, SC FA were thought to be rose progressively, thereb y contributing more and poorly abso rbed and to functi on as an irritating and more to the osmotic dr iving force for diarrh oea. os motic force stimulating intes tinal motilit y and With the largest dose of lactul ose (125 g/day) the imp eding co lonic water and elec tro lyte absorp difference between lactulose- and PEG-induced tion .135.3xl Early work on carbohydrate malabsorp diarrh oea markedly decreased. In lactulose-induced tion therefo re impli cated bacter ial fermentation as diarrhoea, daily stoo l collections co nta ined an the prim ary mechani sm for diarrhoea in disaccha average of I, 12 and 45g of carbohyd rate when the ride malabsorption (' ferme ntative diarrhoea' ). Th e participant s were ingesting 45 , 95 and 125g of stro ng co rrelation between faecal water output and lactul ose/d ay, respectively, indi catin g a maximum output of SCFA135.371 was thought to incriminate ca pacity of the co lonic bacteria to met aboli se SCFA in the path ogenesis of diarrhoea. As co lonic lactulose of approximately 80 g/day.' 39I SCFA levels remain more or less co nstant despite Saunders and Wiggi nsl'! obtained dose-response dietary changes, any fac tor increasing stool we ight curves for mann itol (a poorly absorbed sugar alco will, however, increase their output. Later in vivo hol), raffi nose, lactulose and magnesium sulfate in studies of intestinal SCFA abso rption have consis relation to faecal water output over 48 hours afte r tent ly de mo nstrated efficient absorption of SCFA the dose of test substance. With magnesium sulfate from the human colon.' I 1.1 21 Th is absorption mark- there was an immediate increase in faecal output,

whereas a lag period was observed for the non flora in these older piglets were capable offerment absorbed sugars. When an increase in faecal water ing the osmotic load to SCFA, which could then be occurred after ingestion of mannitol, lactulose or rapidly absorbed from the colonic contents. Once raffinose, the sugar or its constitutive hexoses ap the microbial process is well established, pro peared in the stools. The dose-response experi longed ingestion of carbohydrate results in an im ments indicated that the colonic flora can normally proved colonic fermentation capacity.l'v'" convert 40 to 60g of single doses of carbohydrate The higher capacity to ferment carbohydrates to SCFA without the individual experiencing diar may be a result of enzymatic induction in existing rhoea.I" rnicro-organisms!''! and/or an alteration in micro bial populations.lv-'!' Flourie et al.1301 studied 3.2 Fermentation Capacity Threshold: healthy volunteers for 2 test periods, at the begin Individual Variability ning of which they ingested a diarrhoeogenic load (60g) of lactulose; the 2 periods were separated by The human colon is obviously capable of re a lactulose feeding period of 8 days, during which moving appreciable amounts offermentable carbo a nondiarrhoeogenic load (20g) of lactulose was hydrates from colonic contents, even in patients taken twice daily. Stool weight and frequency, and with malabsorbed carbohydrate.Ue'" If, however, faecal outputs of carbohydrates after the ingestion the fermentative capacity of the colonic microflora of lactulose 60g, dropped significantly after the is exceeded, unmodified and osmotically active lactulose maintenance period (20g twice daily for sugars may produce an osmotic diarrhoea. The re 8 days), indicating that the colonic flora can adapt sponse threshold when a given amount ofcarbohy to a nondiarrhoeogenic load of lactulose, and that drate is malabsorbed differs among individuals, In this adaptation has a beneficial effect on the diar the study by Saunders and Wiggins.U' one partici rhoea induced by a larger load of this sugar. pant tolerated a 73 mmol (25g) dose of lactulose Similarly, Launiala!"!' reported on a 3-week whereas another tolerated 176 mmol (60g) before faecal water output rose above 400 ml/48 hours and period during which an infant with congenital lac test carbohydrate appeared in the stool. Similarly, tose malabsorption was maintained on a lactose doses of 120 mmol (22g) and 350 mmol (64g) of containing diet. In comparison with the first day of mannitol were tolerated by different participants. treatment, continued treatment resulted in a de The individual variability may, at least in part, creased stool volume and disaccharide output, and be explained by differences in the fermentation ca a sharp rise in faecal lactate concentration. No ad pacity of the bacterial flora. Individuals with a lim aptation occurred in the small intestine.H!' ited fermentation capacity or a high dependence on Hypothetically, regular consumption of carbo the existing capacity for colonic carbohydrate deg hydrates malabsorbed in the small bowel may re radation might be especially susceptible to an in sult in a more efficient colonic carbohydrate creased carbohydrate load. A similar situation may fermentation and an increased response threshold apply to the development of microbial fermenta when a given amount of carbohydrate is mal tion in newborn animals. Argenzio et al.1401 pro absorbed. In addition to differences in colonic bac duced carbohydrate malabsorption by administra terial fermentation capacity, variations among in tion ofcorona virus to newborn pigs. This resulted dividuals in the buffering capability of the colon in diarrhoea in 3-day-old pigs, with high concen may contribute to individual differences in the re trations of carbohydrate and low concentrations of sponse to identical amounts of nonabsorbable car SCFA in colonic contents. Pigs 3 weeks of age bohydrates. It is known that caecal pH decreases showed no diarrhoea, with negligible amounts of even with the amounts of unabsorbed carbohydrate carbohydrate and normal levels ofSCFA in colonic normally entering the colon .F" Furthermore, pa contents, indicating that the more mature micro- tients with severe carbohydrate malabsorption, ei-

150 ::J::, 0 E .s 100 2 l!! "0 >- JC::. 50 0 of? III U 0

J: 0- 6

4 )2280 )2100 )2300 1500 )1678

i 1000 ~ .E OJ . ~ ~ 8: 500

o 20 40 80 160 320 0 20 40 80 160 320 Lactulose (g/day) Fig. 2. Faecal weight,pH and carbohydrateconcentrationin faeces of 12 individualstaking increasingdoses of lactulose.The results have been divided in accordancewith the responsein faecal output to low doses of lactulose(from Holtug et al.,126] with permission). ther due to disaccharidase deficiency or due to in 3.3 Does Lactulose Exert ItsLaxative Effect gestion of excessive amounts of lactulose, have de Through the Osmotic Activity of creased faecal pH in the range of 4 to 5.124,26,35] Unfermented Carbohydrate? This decrease in pH is sufficient to inhibit bac Many physicians believe that the change in terial fermentation and SCFA production as shown stool output in the diarrhoea of carbohydrate mal in vivo and in vitro.126,42,43j In subjects with a re absorption is due to the osmotic effect of malab stricted colonic buffering system, the tolerance of sorbed and nonfermented sugars. This assumption carbohydrate malabsorption may be decreased as seems to be supported by the studies mentioned the low pH decreases fermentation leading to ac above (sections 3.1 and 3.2) in which diarrhoea cumulation of undigested sugars and osmotic diar ensued when carbohydrate appeared in the stools. rhoea early on in the process . The study by Holtug et aI.,126 1however, queries this

assumption. Healthy individuals were given lactu major importance. Rapid infusion (8.3 ml/min) of lose twice daily for 3 consecutive days in each ex a single bolus of 250ml of t1uid did not influence perimental period. Doses were doubled (20, 40,80, faecal output whereas 500ml delivered at the same 160 and 320 g/day) from each period to the next, rate overwhelmed the absorptive capacity of the with the end-point being diarrhoea in excess of colon and produced liquid stools.U''!In healthy hu 1000 g/day. Faecal output responded differently man volunteers, Chauve et al.1471investigated pres among individuals, and 2 types of behaviour were sure recordings in the right , transverse and left co identified (fig . 2). lon, while isotonic saline was continuously infused In accordance with the interpretation that carbo in the caecum. The data obtained indicated that hydrate-induced diarrhoea is due to the osmotic ef once a certain volume of content is reached in the fect of nonfermented mal absorbed sugars, diar proximal part of the colon, it contracts and propels rhoea occurred suddenly in six of 12 individuals in its content.l''?' which may explain the different co association with the appearance ofcarbohydrate in lonic responses comparing slow continuous and faeces . In the 6 other subjects, however, faecal out rapid bolus infusions of t1uid into the caccum .lv' put gradually increased with increasing doses of The extra volume load to the colon due to lactulose and diarrhoea occurred before the appear lactulose can be estimated to be 1.65L at a lactulose ance of faecal carbohydrates.P''! How does unab intake of 60g,130,41,45 1and the sudden arrival in the sorbed carbohydrate cause diarrhoea in those caecum of this considerable amount of ileal dis cases? charge might lead to colonic 'decompensation' and prompt colonic evacuation. 3.4 Capacity of the Colon to Absorb Fluid 3.5 Possible Mechanisms of Action of The mechanism of action of lactulose as a laxa Lactulose in the Treatment of Constipation tive is probably multifactorial, involving effects on both the colon and the small bowel. Under normal A single primary mechanism for the laxative ef circumstances, about 1500ml of fluid is estimated fect of lactulose cannot be determined because of to enter the colon each day.14411ndisaccharide mal multiple possible effects. Most likely, the laxative absorption an increased volume offluid is retained effect of lactulose results from the combination of: in the lumen of the small intestine by the osmotic • water retention in the small intestine through the effect of disaccharide.141,451 Analysis of distal ileal osmotic activity of the unabsorbed disaccha fluid collected during the passage ofa lactose meal ride; and in lactase-deficient people and a lactulose load in • interference with net t1uid absorption in the co healthy people indicates that about two-thirds of lon due to the osmotic effect of malabsorbed and the osmotic load entering the colon consists of intact sugars when the bacterial fermentation endogenous electrolytes.130,41,45] Thus, the water capacity is exceeded. load delivered to the colon is about 3 times that Whether or not the laxative effect of lactulose is calculated to be osmotically held by the non accomplished is probably just a matter of dose . It absorbed sugar. is, however, important to realise that the response A logical extension of the above observations is threshold when lactulose is prescribed differs to ask what the colon can accomplish under stress. among individuals (fig . I) . A gradual increase in In fact, the colon has a surprisingly large capacity faecal output with increasing amounts of lactulose to adapt to fluid overload. When the healthy human occurs in some individuals, while others suddenly colon was subjected to a slow, continuous infusion respond with severe diarrhoea.P'" Unfortunately, it of isotonic fluid (1.4 to 2.8 ml/min), net absorption is not possible to predict the response of a given of water increased markedly, up to 5 to 6 L/day.1461 individual, and firm guidelines for dosage recom The rate and pattern of t1uid entry are, however, of mendations are impossible.

4. Lactulose in the Management of bacterial metabolism of amino acids , peptides and Hepatic Encephalopathy proteins.P?' and as the simplest nitrogenous com pound which can be used by bacteria as a starting Lactulose has been successfully used in the material in the synthesis of their own nitrogenous treatment of hepatic encephalopathy ever since constituents. In the absence of active colonic fer the pioneering experiments of Bircher et al. in mentation most of the ammonia formed is ab 241 1966.' The exact mechanisms whereby lactulose sorbed to be converted to urea in the liver. exerts its beneficial effect on hepatic encephalop When lactulose was first used in the treatment athy are not fully understood, but several hypoth of hepatic encephalopathy, it was presumed that eses have been suggested. the colonic acidification favoured the growth of Hepatic encephalopathy is a complex neuropsy acid-tolerant lactobacilli and otheracidophilic, fer chiatric syndrome that appears to be characterised mentative bacteria and reduced the growth of predominantly by augmented neuronal inhibition. acidophobic, proteolytic bacteria responsible for The syndrome is associated with hepatocellular ammonia formation.P'! However, quantitative failure, increased portal-systemic shunting of studies failed to confirm this rearrangement of gut blood and multiple metabolic changes, and is gen f1ora.127,281 Alternatively, luminal acidification erally considered to be a potentially reversible me would reduce the proportion of ammonia available tabolic encephalopathy. The precise pathogenetic for reabsorption by passive nonionic diffusion by mechanism responsible for the CNS dysfunction increasing the ratio of ionised to unionised ammo remains to be determined. nia, thereby enhancing the faecal excretion of am Products of bacterial metabolism in the colon, monia.P'!' Attempts to demonstrate increases in which are normally absorbed and extracted by the stool ammonia caused by lactulose administration liver, tend to accumulate in the peripheral blood of have failed,151 ,521 except when excessive amounts patients with liver disease either because the dis of lactulose (>80 g/day) have been given.F'l ln that eased liver is unable to remove these compounds case, stool ammonia increases approximately 3 from the portal blood perfusing it, or because por 133 1 tal hypertension has led to the development of ve fold. The main point is, however, that ammonia nous collaterals which transmit portal blood di constitutes only a small percentage (3 to 5%) of rectly into the systemic circulation. If one or more total faecal nitrogen. The debate on whether or not of these metabolites can cross the blood-brain bar lactulose increases faecal excretion of ammonia is rier and promote neural inhibition, they may con therefore of minor importance in the context of to tribute to hepatic encephalopathy. Despite decades tal nitrogen excretion. Of more importance for fae of research endeavour, the nature of the toxic, gut cal nitrogen clearance are other sources of nitro derived, substances defies identification. Three gen, constituting about 95% of total stool nitrogen. candidate toxins seem likely pathogenetic factors By use of an anaerobic faecal incubation sys I171 because they are present to excess in patients with tem, Vince et al. demonstrated that net genera liver failure and cause coma experimentally in ani tion of ammonia by faecal bacteria can be con mals. These are ammonia, SCFA and medium-chain verted to net utilisation of ammonia by the fatty acids (MCFA), and mercaptans; of these, am provision of a fermentable source of energy, e.g. monia probably plays the most central role.'48I lactulose, ammonia disappearing from the incuba tion system. The results may be explained by a 4.1 Effects of Lactulose on Ammonia and combination of factors: (a) the availability of a Nitrogen Metabolism readily fermentable substrate may encourage bac terial proliferation and assimilation of ammonia as Ammonia occupies a central position in nitrogen a nitrogen source;' 17,31 I (b) preferential use of metabolism in the colon, both as an end-product of lactulose as a carbon and energy source may exert

a sparing effect on the metabolism of both exoge lism;117,19,26,421 or (c) both.A direct effect of lac nous and endogenous aminated compounds with a tulose in reducing the colonic transit time may also subsequent decrease in the generation of ammonia be a factor contributing to an increase in faecal sol by colonic bacteria;117, 191 and, (c) low pH values uble nitrogen. such as are induced in the colon by lactulose may bring about a general reduction in bacterial meta 4,2 Role of Short- and Medium-Chain Fatty bolism, including that of ammonia-producing bac Acids in Hepatic Encephalopathy teria.l 17,19,26,421 If, in the presence of lactulose, in corporation of ammonia into bacterial protein is The suggestion that SCFA and MCFA may play stimulated, faecal nitrogen excretion should in a role in the pathogenesis of hepatic encephalopa crease and hepatic urea production and total body thy is mainly based on animal experiments. Sam 54 urea pool should decrease during administration of son et aI.i 1 investigated the narcotic action of SCFA (acetate, propionate, butyrate, valerate and lactulose. hexanoate) and MCFA (heptanoate and octanoate) This is exactly what Weber l321and Mortensen'Vl in experimental animals, Intravenous or intra have shown. Administration of lactulose caused a peritoneal injection of the sodium salt s of these 2- to 4-fold increase in faecal nitrogen content in acids produced a reversible coma with a definite patients with cirrhosisl321and in healthy individu relationship between the amount of the fatty acid als,1 331 and the increase in nitrogen excretion was which would produce unconsciousness and the accompanied by a significant reduction in the urea carbon length of the compound; i.e. the longer the production rate leading to a reduction in the total carbon chain, the more potent the encephalopathic body urea pool.132 1Using recently developed tech effect, acetate being almost nontoxic.P'l niques that permit quantitative separation of faecal In a subsequent experiment, White and Sam solids into bacterial, soluble and fibre fractions, son l551demonstrated that intravenous injection of Weber et al.1531analysed the effects of lactulose on propionate, butyrate, valerate and hexanoate im the compartmentalisation offaecal nitrogen. In pa mediately induced electroencephalographic (EEG) tients with cirrhosis, administration of lactulose changes from a resting type to one of sleep in un (56 ± 6 g/day) caused a marked increase in both the anaesthetised rabbits. During these EEG alter solid weight and nitrogen content of the bacterial ations, overt signs of drowsiness or sleep always and soluble fractions ofstool ; lactulose administra occurred and the typical 'alert' EEG responses to tion increased nitrogen excretion in the bacterial nociceptive and auditory stimuli were diminished and soluble fractions by 165 and 135%, respec or absent. Again, the longer the carbon chain, the tively. The major increment in nitrogen content of more potent the effect. In rats , the rate of blood the bacterial fraction indicated that lactulose stim brain barrier penetration ofstraight-chain saturated ulated bacterial growth and multiplication signifi monocarboxylic acids has also been shown to in cantly, and thereby increased the incorporation of crease with chain length and is virtually complete nitrogenous compounds such as ammonia into bac at lengths greater than that of hexanoate (C6).1561 terial protein. In humans, octanoate has been shown to cross the The increase in faecal soluble nitrogen may be blood-brain barrier in both normal individuals and explained by reduced bacterial catabolism of ni patients with cirrhosis.lV' trogenous compounds to absorbable metabolites, The mechanisms by which SCFA and MCFA in e.g. ammonia, due to: (a) the introduction of a car terfere with CNS functions are not entirely clear. bohydrate which is preferentially used as a carbon In vitro and in vivo stud ies have shown that SCFA and energy source;117,191 (b) an increase in the and MCFA inhibit cerebral Na+,K+-ATPase activ hydrogen ion concentration which may bring ity,15X,59 1and, as for the narcotic potency, the inhib about a general reduction in bacterial metabo- itory capacity increases with increasing chain

<0Adi s International Limited . All right s reserved. Drugs 1997 Jun; 53 (6) Disaccharides and Colonic Flora 939

length.P'" The con sequences of inhibiting this en or rectal instill ation without any influence on the zyme would include an increase in intracellular so mental or neurological state.16X.691 It therefore dium and possibl y impaired neurotransmi ssion. seems unlikely that the elevated levels of SCFA in Muto and Takahashi t'v ' were the first to sugges t patients with hepatic encephalopathy are the pri a role of SCFA in the pathogenesis of hepatic en mary cause of the neurological deterioration, cephalopathy. In patients with hep atic coma, plas A secondary role for SCFA, however, cann ot be ma levels of SCFA of 4- to 6-carbon length were excluded as the coma-producing potential of sepa increased 3- to 4-fold compared with controls and rate toxic substances may be multiplied several patients with other kind s of coma. Significantly fold when they are pre sent together. In experi elevated levels ofSCFA in peripheral venous blood mental anima ls the coma-producing effects of of patients with fulminant hepatic failure and por ammonia, fatty acid s (octanoate) and mercaptans tal systemic encephalopathy have been confirmed have been shown to be strikingly interdepen 170, recently. Lai et al.1611investigated 6 patients with dent. 7!I In healthy animals, the dose of ammo fulminant hepatic failure due to paracetamol (aceta nium salt required to produce coma is sharply re minophen) intoxication and hepatitis. All patients duced by the presence of a subcoma dose of either were in grade IV coma. Plasma levels of SCFA a fatt y acid (octanoate) or mercaptan.Fv-" Iand the were significantly elevated in patients with en incidence ofcoma can be raised from zero to 100% cephalopathy (2034 ± 11341lmo1/L; mean ±2 SO ) by injection of subcoma doses of any of these sub compared with 10 healthy controls ( 1113 ± 662 stances simultaneously. In rats with acute experi um ol/L), but a correlation between SCFA con cen mental hepatic co ma resulting from massive liver necrosis, the average blood levels of ammoni a, free trations and the clinical course could not be dem fatty acids (oc tanoa te) and methanethi ol are much onstrated. lower than blood levels required to produce co ma Clausen et al.1621inve stigated 32 patients with in health y rats with each of these substances indi cirrhosis, 15 patients with and 17 patients without vidually.!" I When , however, health y rats are give n hepatic ence phalopathy, and II health y indi vidu a combination of doses of ammonia, octan oat e and als. Plasma concentrations of SCFA were signifi mercaptan which produce blood level s in the range cantly elevated in patients with hepatic encepha of those observed in rats with experiment al hepatic lopathy (362 ±831lmo1/L; mean ± SEM) compared coma, they becom e cornato se.P!' with patients without encephalopathy ( 178 ± 57 umol/L) and healthy controls (60 ± 8IlmoI/L). Re 4,3 Effect of Lactulose on Colonic SCFA petitive sampling from the encephalopathic pa Production in Hepatic Encephalopathy tients showed no relationship between SCFA con centrations and the grade of encephalopathy.I'<' Although SCFA may not be the major ca use of In comparison with the encephalopathy asso encephalopathy in patients with cirrhosis, the evi ciated with inborn errors of organic acid metabo dence of synergism in animalsl70.71 I suggests that lism, el evated levels of SCFA in patients with SCFA may playa role as an augmenter or intensi hepatic encephalopathy are modest. Massive ele fier of other putativ e neurotoxins, e.g. amm oni a vations of propionate are found in severe and fatal and mercaptans. In incubated stool samples, fer ca ses of propionic acidaemias (5400 to 38 000 mentation of lactulose results in the form ation of IlmoI/L ),1 63-651 and levels of isovalerate in the acetate prim aril y, whereas bacterial degradation of rang e of 340 to 2990 umol/L are associ ated with blood, albumin and am ino acids significa ntly in stupor and unconsciousne ss in children with iso creases the production of all C2-6-SCFA, es pe valeric ac idaemia.166,671 Moreover, an 8- to IO-fold cially the potentially toxic C4-6-fatt y ac ids.119,4 31 rise of venous propionate and oct anoate (C8) is When lactul ose is added to incubations of amino seen in patients with liver cirrhos is after duodenal ac ids, protein and blood, the SCFA produ ction is

shifted almos t co mpletely toward s the nontox ic tate at the expense of the potent ially toxic C4-6 C2-fatty acid, acetate. At lactul ose concen trations SCFA. of 10 to 25 mmol/L, corresponding to a pH of ap proximately 5, this effec t is largely mediated by the 5. Conclusions acidifying effect of lactulose.F'J pH-Independ ent inhibition of blood and amino acid degradation to A single primary mech an ism of actio n of SCFA requires conce ntrations of lactulose exceed lactul ose in the treatment of co nsti patio n and he ing 50 to 100 mmol/L.1431The co nce ntratio n of patic ence phalopathy cannot be determined due to lactul ose after a single dose of 20g is 50 to 100 several possible effects the dru g may have. In pa mmol/L in the caecum.l''l with a corres ponding de tient s with consti pation, the laxative effect of lac crease in pH to values in the range of 3.5 to 6.1 in tulose possibly results from the combination of: (a) the prox imal colon.'6.231 flu id retenti on in the small intestine through the In co ntras t, redu ction of faecal pH by 20 to 40g osmotic activity of the unabsorbed disaccharide; of lactul ose is small in magnitudelV' or eve n neg and (b) interference with net flu id absorpti on in the ligiblel'" due to the rapid fermentation of lactul ose colon due to the osmotic effect of malabso rbed and and the neutralisation ofluminal co ntents by bicar intact sugars when the bacteri al fermentatio n ca bonate secretion. The result s obtained in vitrol431 pacit y is exceeded. However, indivi dual sensitivi indicate that lactul ose should be given in doses that ties to the drug should be recogni sed, i.e. the ther not only acidify the caec um but also exceed the apeutic dose sho uld start at a relatively low level buffering capacity in the left co lon. with a gradual increase according to the frequency Enemas of lactul ose or glucose may be benefi and consistency of the stoo ls. cia l supplemen ts to aid in obtaining this effect. The possibl e mec hanisms by whic h lactulose benefits patient s with hepatic encephalopathy in cl ude: (a) increased faecal nitrogen excretion; (b) 4,4 Possible Mec hanisms of Ac tion of colonic acidification; (c) decreased tra nsit time; Lactulose in the Treatm ent of and (d) redu ced production of potentially toxic Hepatic Encephalopathy SCFA . In order to obtain these effects it is recom Several mechanisms of action of lactulose in the mend ed that the faeca l pH be measured in patients treatm ent of hepatic ence pha lopathy apparently with hepat ic ence phalopathy treated with lactu - ex ist (ta ble I). They inc lude: (a) provision of a readi ly fermentable substrate, whic h enco urages Table I. Possible mechanisms of action of lactulose in the treatment bacteri al proli feration and converts the norm al bal of hepatic encephalopathy ance betw een ammonia generation and utilisat ion Effect Result from a net increase to a net decrease (assimil ation) ; Utilisation Stimulates bacterial growth and increases (b) a preferenti al use of lactulose as a carbon and incorporation of ammonia into bacterial protein energy source, which exerts a sparing effect on the (assimilation), and exerts a sparing effect on the metabolism of aminated compounds with bacterial metabolism of both exogenous and en subsequent decrease in ammonia generation by dogenou s aminated co mpo unds with subsequent colonic bacteria decrease in the generation of ammonia; (c) an acid Acidifying Brings about a general reduction in bacterial metabolism, including that of ammonia-producing ifyi ng effect, which brings abo ut a general reduc bacteria tio n in bacterial metabolism, including that of am Laxative Reduces the colonic transit time. thereby moni a-p rodu cin g bact eria; (d) a laxative action, decreasing the time available both for production which reduces the time for the production and ab and absorption of ammonia and other potentially toxic substances sor ption of ammonia and increases the soluble ni SCFA Enhances the production of the nontoxic acetate trogen content of faeces; and, (e) an ability to alter modification at the expense of the potentially toxic C4-6-SCFA the productio n of SCFA towards the nontoxic ace- Abbreviation: SCFA - short-chain fatty acids.

'" AdisInternational Umited. All rightsreserved. Drugs 1997 Jun: 53 (6) Disaccharides and Colonic Flora 941

lose, and that the lactulo se dose should be titrated 20. Peters SG, Poma re EW, Fisher CA. Portal and peripheral blood short chain fatty acid concentrations after caecal lactulose to pH levels <6.0. Acidifying enemas, e.g. of instillati on at surgery. Gut 1992; 33: 1249-52 lactulo se, lactitol or glucose, may be a beneficial 21. Bond JH, Levitt MD. Use of pulmonary hydrogen measure supplement in obtaining this effect. ments to quantitate carboh ydrate absorption: study of par tially gas trectomized patient s. J Clin Invest 1972; 51: 1219-25 References 22. Strocchi A, Cora zza G, Ellis CJ, et al. Detection of malabsorp tion of low doses of carbohydra tes: accu racy of various brea th I. Cummings JH, Macfarlan e GT. The control and consequences H2 criteria. Gastroenterology 1993; 105: 1404-10 of bacterial fermentation in the human colon . J Appl Bacteriol 199 1; 70: 443-59 23. Bown RL, Gibson JA, Siaden GE, et al. Effects of lactu lose and 2. Carulli N, Salviolo GF, Manent i F. Ahsorpti on of lactulose in other laxatives on ileal and colonic pH as measured by a man. Digestion 1972; 6: 139-45 radiotelemetry device. Gut 1974; 15: 999- 1004 3. Saunders DR, Wiggins HS. Conservation of mannitol , lactu 24. Bircher J, Muller J, Guggenheim P, et al. Treatment of chronic lose, and raffinose by the human co lon. Am J Physiol 1981; portal-system ic encephalopathy with lactulose. Lancet 1966; 241 : G397 -402 I: 890-3 4. Zarling EJ, Ruchim MA. Protein origin of the volatile fatty 25. Conn HO, Leev y CM , Vlahcev ic ZR, et al. Comparison of acids isobutyrate and isovalerate in human stool. J Lab Clin lactulose and neomycin in the treatment of chronic portal-sys Med 1987; 109: 566- 70 temic encephalopathy: a doubl e blind controlled trial. Gastro 5. Rasmussen HS, Holtug K, Morten sen PB. Degradat ion of entero logy 1977; 72: 573-83 amino acids to short chain fatty acids in humans: an in vitro 26. Holtug K, Clausen MR, Hove H, et al. The colon in carbohy study. Scand J Gas troenterol 1988; 23: 178-82 drate malabsorption: short-chain fatty acids, pH, and osmotic 6. Florent C, Flourie B, Leblond A, et al. Influence of chronic diarrhoea. Scand J Gastroenterol 1992; 27: 545-52 lactulo se ingestion on the colonic metabolism of lactulo se in 27. Conn HO, Floch MH. Effects of lactulose and lactobacil man (an in vivo study). J Clin Invest 1985; 75: 608-13 lus acidophilus on the fecal flora. Am J Clin Nutr 1970; 23: 7. Hove H, Morten sen PB. Colonic lactate metabolism and D-Iac 1588-94 tate acidosis. Dig Dis Sci 1995; 40: 320-30 28. Vince A, Zeegen R, Drinkw ater JE, et al. The effect of lactulose 8. Bustos D, Pons S, Pernas JC, et al. Fecal lactate and short bowel on the faecal nora of patients with hepatic encephalopathy. J syndrome. Dig Dis Sci 1994; 39: 23 15-9 Med Microbiol 1974; 7: 163-8 9. Cu mmings JH, Pomare EW, Branch WJ, et al. Short chain fatty 29. Riggio 0 , Varriale M, Testore GP, et al. Effect of lactitol and acids in human large intestine, portal, hepatic and venous lactul ose administration on the fecal flora in cirrhotic pa blood . Gut 1987; 28: 1221-7 tients. J Clin Gastroe nterol 1990; 12: 433-6 10. Mortensen PB, Hove H, Clausen MR, et al. Fermentation to 30. Flourie B, Briet F, Florent C, et al. Can diarrhea induced by short-chain fatty acids and lactate in human faecal batch cul lactulose be reduced by prolonged ingestion of lactulose'? Am tures. lntra- and inter-individual variations versus variations J Clin Nutr 1993; 58: 369-75 caused by changes in ferment ed saccharides. Scand J Gas 3 1. Stephen AM, Cu mmings JH. Mech anism of action of dietary troenterol 199 1; 26: 1285-94 fibre in the huma n colon. Nature 1980; 284: 283-4 II. McNeil NI, Cummings JH, James WPT. Short chai n fatty acid 32. Weber FL. The effect of lactulose on urea metabolism and ni absorption by the human large intestine. Gut 1978; 19: 819-22 trogen excre tion in cirrhotic patie nts. Gastroe nterology 1979; 12. Ruppin H, Bar-Meir S, Soergel KH, et al. Absorption of short 77: 5 18-23 chain fatty acids by the colon. Gastroenterol ogy 1980; 78: 33. Mortensen PB. The effe ct of ora l-administered lactulose on co 1500-7 lonic nitrogen metaboli sm and excretion. Hepatolo gy 1992: 13. Roediger WEW. Role of anaerobic bacteria in the metabolic 16: 1350-6 welfare of the coloni c mucosa in man. Gut 1980; 21: 793-8 34. Kot T V, Pettit-Young NA. Lactulose in the management ofcon 14. Clausen MR, Mortensen PB. Kinetic studies on the metabolism stipation: a cu rrent review. Ann Pharm acother 1992; 26: of short-chain fatty acids and gluco se by isolated rat co l 1277-82 onocytes. Gastroenterolo gy 1994; 106: 423-32 35. Weijers HA, van de Kamer JH, Dicke WK, et al. Diarrhoea 15. Claus en MR, Mortensen PB. Kinetic studies on colonocyte me cau sed by deficiency of suga r splitting enzymes. Acta tabolism of short -chain fatty acids and glucose in ulcerat ive Paediatr 1961; 50: 55-71 co litis. Gut 1995; 37: 684-9 36. Haemmerli UP, Kistler H, Ammann R, et al. Acquired milk 16. McNeil NI. The contribution of the large intestine to energy supplies in man. Am J Clin Nutr 1984; 39: 338-42 intolerance in the adult caused by lactose malabsorption due 17. Vince A, Killingley M, Wrong OM. Effect of lactulose on am to a selective deficiency of intestinal lactase activity. Am .I monia production in a fecal incubat ion system. Gastroenter Med 1965; 38: 7-30 ology 1978; 74: 544- 9 37. Bustos Fernandez L, Gon zale z E, Marzi A, et al. Feca l 18. Mortensen PB, Holtug K, Rasmus sen HS. Short-chain fatty acidorrhea. N Engl J Med 1971; 284: 295-8 acid producti on from mono- and disaccharides in a fecal in 38. Christopher NL, Bayless TM. Role of the small bowel and co cubation system: implication s for colon ic fermentati on of di lon in lactose-induced diarrh ea. Gastroenterology 1971; 60: etary fiber in human s. J Nutr 1988; 118: 321-5 845-52 19. Mortensen PB, Rasmus sen HS, Holtug K. Lactulose detoxifi es 39. Hammer HF, Santa Ana CA, Schill er LR, et "I. Studies of os in vitro short- chain fatty acid production in colonic contents motic diarrhea induced in normal subjects by ingestion of induced by blood: implications for hepatic coma. Gastroen polyethylene glycol and lactulose. J Clin Invest 1989; 84: terology 1988; 94: 750-4 1056-62

40 . Argenzio RA, Moon HW, Kemeny LJ, et al. Colonic compen 58. Dahl DR. Short chain fatty acid inhibition of rat brain Na-K sation in transmissible gastroenteritis of swine. Gastroenter adenosine triphosphatase. J Neurochem 1968; 15: 815 -20 ology 1984; 86: 1501-9 59 . Trauner DA. Regional cerebral Na+K+ ATPase activity follow 4 J. Launiala K. The mechanism of diarrhoea in cong enital disac ing octanoate administration. Pediatr Res 1980; 14: 844-5 charide malabsorption. Acta Paediat Scand 1968; 57: 425-32 60 . Muto Y, Takahashi Y. Gas chromatographic determination of 42 . Perman JA , Modler S, Olson AC. Role of pH in production of plasma short chain fatty acids in diseases of the liver. J Jpn hydrogen from carbohydrates by colonic bacterial flora: stud Soc Int Med 1964; 53 : 828-39. Cited in Postgrad Med 1965; ies in vivo and in vitro. J Clin Invest 1981; 67: 643-50 37: A 158 43 . Mortensen PB, Holtug K, Bonnen H, et al. The degradation of 61. Lai JCK , Silk DBA , Williams R. Plasma short-chain fatty acids amino acid s, protein s, and blood to short-chain fatty acids in in fulminant hepatic failure . Clin Ch im Acta 1977; 78: 305- 10 colon is prevented by lactulose. Gastroenterology 1990; 98 : 62 . Clausen MR, Mortensen PB, Bendt sen F. Serum levels of short 353-60 44 . Phillips SF, Giller J. The contribution of the colon to electrolyte chain fatty acids in cirrhosis and hepat ic com a. Hepatology and water conservation in man . J Lab Clin Med 1973; 81: 1991; 14: W40-5 733-46 63. Hommes FA, Kuipers JRG , ElemaJD, et al. Propionicacidemia, 45 . Bond JH , Levitt MD. Quantitative measurement of lactose ab a new inborn error of metabolism . Pediatr Res 1968; 2: 519-24 sorption. Gastroenterology 1976; 70 : 1058-62 64 . Bran ski D, Gale R, Gross-Kieselstein E, et al. Propionic acide 46. Debongnie JC, Phillips SF. Capacity of the human colon to mia and anorectal anomalies in three siblin gs. Am J Dis Child absorb fluid . Gastroenterology 1978; 74: 698-703 1977; 131: 1379-81 47 . Chauve A, Devroede G, Bastin E. Intraluminal pressures during 65 . Hillman RE. Simple, rapid method for determination of propi perfusion of the human colon in situ. Gastroenterology 1976; onic acid and other short-chain fatty acids in serum. Clin 70 : 336-40 Chern 1978; 24: 800-3 48 . Zieve L. The mechanism of hepatic coma. Hepatology 1981; I: 66. Budd MA, Tanaka K, Holmes LB, et al. Isovaleric acide mia: 360-5 clinical features of a new genetic defect of leucine metabo 49 . Vince A, Down PF, Murison J, et al. Generation of ammonia lism . N Engl J Med 1967; 277 : 321-7 from non-urea sources in a faecal incubation system. Clin Sci 67. Duran M, Sprang FJ van, Drewes JG , et al. Two sisters with Mol Med 1976; 51 : 313-22 isovaleric acidaemia, multiple attacks of ketoacidosis and 50 . Haemmerli UP, Bircher J. Wrong idea , good results. (The normal development. Eur J Pediatr 1979; 131: 205-11 lactulose story.) N Engl J Med 1969; 281 : 441-2 68. Scheppach W, Richter F, Joeres R, et al. Systemic avail ability 51 . Zeegen R, Drinkwater JE, Fenton JCB , et al. Some observations on the effects of treatment with lactu lose on patients with of propionate and acetate in liver cirrhosis. Am J Gas chronic hepat ic encephalopathy. Q J Med 1970; 39: 245-63 troenterol 1988; 83 : 850-3 52. Agostini L, Down PF, Murison J, et al. Faecal ammonia and pH 69. Linscheer W, Castell DO, Platt RR. A new method for evalua during lactulose administration in man : comparison with tion of portosystemic shunting.Gastroenterology 1969; 57: other cathartics. Gut 1972; 13: 859-66 415-23 53 . Weber FL, Banwell JG , Fresard KM, et al. Nitrogen in fecal 70. Zieve FJ, Zie ve L, Doi zaki WM, et al. Synergism between am bacterial, fiber, and soluble fractions of patients with cirrho mon ia and fatty acids in the production of coma: implications sis : effects of lactulose and lactulose plus neomycin. J Lab for hepatic coma.J Pharmacol Exp Ther 1974; 191: 10-6 Clin Med 1987; 110: 259-63 71. Zieve L, Doizaki WM , Zieve FJ. Synergism between rnercap 54 . Samson FE, Dahl N, Dahl DR. A study on the narcotic action tans and ammonia or fatty acids in the production of com a: a of the short chain fatty acids. J Clin Inves t 1956; 35 : 1291-8 possible role for mercaptans in the pathogenesis of hepatic 55. White RP, Samson FE. Effects of fatty acid anions on the elec coma. J Lab Clin Med 1974; 83: 16-28 troencephalogram of unanesthetized rabbits. Am J Physiol 1956; 186: 271-4 56. Oldendorf WH oCarrier-mediated blood-brain barrier transport of short-chain monocarboxylic organic acids . Am J Physiol Correspondenceand reprints: Dr MetteRyeClausell, Depart 1973; 224: 1450-3 57 . Linscheer WG, Blum AL, Platt RR. Transfer of medium chain ment of Medicine CA 2101, Division of Gastro enterology, fatty acids from blood to spinal fluid in patients with cirrhosis. Rigshospitalet, Blegdamsvej 9, DK-2100 Copenhagen 0, Gastroenterology 1970; 58 : 509- J5 Denmark.