key: cord-0006357-rub02my3 authors: Clausen, Mette Rye; Mortensen, Per Brøbech title: Lactulose, Disaccharides and Colonic Flora: Clinical Consequences date: 2012-10-12 journal: Drugs DOI: 10.2165/00003495-199753060-00003 sha: dffecbe1b5408e91638af6158e04f4006513578c doc_id: 6357 cord_uid: rub02my3 Lactulose is one of the most frequently utilised agents in the treatment of constipation and hepatic encephalopathy because of its efficacy and good safety profile. The key to understanding the possible modes of action by which lactulose achieves its therapeutic effects in these disorders lies in certain pharmacological phenomena: (a) lactulose is a synthetic disaccharide that does not occur naturally; (b) there is no disaccharidase on the microvillus membrane of enterocytes in the human small intestine 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 readily fermented by the colonic bacterial flora with the production of short-chain fatty acids and various gases. The purpose of this review is to focus on some pertinent basic aspects of the clinical pharmacology of lactulose and to discuss the possible mechanisms by which lactulose benefits patients with constipation and hepatic encephalopathy. Lactu lose is one of the mos t freq uen tly utilised age nts in the treatment of constipation and hepatic ence pha lopat hy because of its efficacy and good safe ty profile. T he key to understand ing the possible modes of action by which lactulose ac hieves its therapeutic effects in these diso rders lies in certain pharmaco logical phe nomena : (a) lactu lose is a sy nthetic disaccharide that does not occur naturally; (b) there is no disacchari dase on the microvi llus membrane of enterocytes in the human sma ll inte sti ne that hydrol yses lact ulose; and (c) lactulose is not absor bed 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 of short-chain fatty acid s and various gases. T he pu rpose of this review is to focu s on so me pertinent basic aspects of the clinica l pharmacology of lactulose and to di scu ss the possible mec hanisms by which lactulose benefits patients with co nstipation and hepatic encephalopat hy. Dietary carbohydrates may be divided into monosaccharides, disaccharides, oligosaccharides and polysaccharides. Their physiological properties and health benefits depend on the site, rate and extent of their digestion or fermentation in the human gastrointestinal tract. The polysaccharide fraction of plant cell walls (dietary fibre) and some forms of starch (resistant starch) are resistant to hydrolysis in the small intestine. These carbohydrate polymers as well as indigestible oligosaccharides, e.g. stacchyose and raffinose, enter the colon through the ileocaecal valve to be fermented by the colonic bacterial flora. Monosaccharide and disaccharide sugars are absorbed as they pass through the healthy small intestine and only become available to the colonic flora as a consequence of deficiencies in the transport system, or, more usually, specific disaccharidase deficiencies, e.g. lactase deficiency. Physiological carbohydrate malabsorption, that is, the fermentable carbohydrate entering the colon in healthy individuals, is in the range of 30-60 g/day in Western communities .l!l This load is normally fermented by the vast numbers of colonic bacteria. Lactulose is a synthetic disaccharide composed of the monosaccharides galactose and fructose [p- ( 1-4)-galactosido-fructose] . The disaccharidases that split naturally occurring disaccharides into their hexose moieties do not include a lactulase, and lactulose is not metabolised or absorbed in the human small intestine.Pv' Once in the caecum, lactulose is a fully fermentable carbohydrate. Lactulose is widely considered to be an effective agent in the management of constipation and hepatic encephalopathy. It is the purpose of this review to discuss the interaction of lactulose with the bacterial flora of the colon and the possible modes of action by which it achieves its therapeutic effects in patients with constipation and hepatic encephalopathy. The review deals almost exclusively with lactulose. The principles expressed, however, are applicable to any mal absorbed and fermentable disaccharide, e.g. lactitol, a disaccharide analogue of lactulose for which there is also no disaccharidase on the microvillus membrane of enterocytes, or lactose when it is administered to lactase-deficient individuals. Fermentation, the process whereby the microbial population that inhabits the human colon breaks down carbohydrates in order to obtain energy for maintenance of cellular function and growth, is an important component of normal colonic activity (fig. I) . Short-chain fatty acids (SCFA) and the gases H2, CO 2 and, in some individuals, CH 4 , are the principal end products of carbohydrate fermentation.l!' The major SCFA found in the human colon are the C2 (acetate), C3 (propionate) and C4 (butyrate) members of the aliphatic monocarboxylic acid series. Formate (C I) , valerate (C5), hexanoate (C6), and the branched- chain fatty ac ids isobutyrate (iC4) and iso valerate (iC5) , whi ch are produced during amino acid ferment ation ,14's1are also presen t in the co lon, thou gh in smaller qu antities. Oth er orga nic ac ids such as lactate and succ inate are ferment ati on intermediates. Th ey ca n be further metaboli sed to SC FA, and do not usu all y accumulate to any extent, except during rapid carbo hydrate breakd o wn. P"! Th e SCFA are weak acid s and their pKa values are ve ry similar, ran gin g from 4.76 to 4 .87 . Thu s, more than 95% of the SCFA are present in their di ssoci ated form at the phy siological pH of the colon (pH [6] [7] [8] . Total co ncentrations of SCFA are low in the sma ll intestine, but in the co lon and faeces SC FA co nstitute the predominant ani on s at con centrati on s ran ging from about 80 to 130 mm ol/Lp ·IOI Concentrati ons are high est in the caecum, where substra te avai lability is grea tes t, and fa ll progressive ly towards the dista l co lon. By co ntras t, pH is lo west in the right co lo n and rises in the d istal bowel .l?' SC FA are av idly abso rbed by the co lonic epit he lium,II I.1 2 1 and used as a so urce of energy, e ither locall y (co lonic mucosa)113. 15 1 or sy ste mica lly (l iver, peripheral tissues).' 161 In the colon, abso rpt ion of SCFA is accompanied by luminal bicarbon at e acc umulatio n and inc reased sod ium and wate r absorption.t'I -P!Thu s, the micro bia l pro cess co nve rts unabsorb abl e materi al to rapidly abso rbed SCFA, thereby redu cing the effective os mo tic pressure of the co lonic co ntents, enhanci ng water and el ectrol yt e abso rptio n, and sa lvag ing calo ries whi ch wo uld otherwise be lost in the faeca l strea m. Lactul ose, afte r arriving in the co lon, is digested by bacteri al disaccharidases and metaboli sed to orga nic aci ds, mainl y acetate and lactate, H 2 and C02.1 61 A variety of studies have show n that lactulose is usuall y vigorously fermented . In an ill vitro ex pe rime nt using human faecal ho mogen ates, Vince et aLI 17 1showed that co mplete fermen tatio n of 10 gi L (30 mm ol/L ) of lactul ose occurred within 6 hours. By use of a sim ilar tech niqu e, Mort en se n et a LII X . 191 showed tha t the total produ ct ion of SCFA in faeca l ho moge nates was increased 2-to e Ad is Interna Nona l Umite d . All rights reserved . Clausell & Mor!('//s('// 3-fo ld by the additio n of 100 mm ol/L lactulose or its mon osacch ari de co ns titue nts, ga lac tose an d fru ctose, during inc ubation fo r 6 ho urs , seco ndary to an iso late d 4-to 6-fo ld inc rease in the synthes is of acetate. In patient s undergoing elective cholecystectomy, caecal insti lla tio n of lactulose ca used an immedi at e inc rea se in por tal plasm a SCFA co nce ntratio ns, princ ipall y acetate; pea k co ncentration s we re ac hieve d in 15 to 45 minutes, co nfirmin g efficient fer men tatio n of lactul ose an d abso rption of the resultant SCFA P ol Th e vario us gases ge nerated in the co lon are e ither absorbed and ex pired in breath or expe lled as flatu s, and excess gas production may ca use borbor ygm i and flatul en ce. Sin ce H 2 pro duce d in the human bod y deri ves entire ly fro m bacter ial ferment ation of non abso rbed carb oh yd rate, brea th H 2 exc retion in the abse nce of sma ll intes tinal bacterial ove rgrowth ca n be used as a mar ker of car bohydrate rnalab sorpt ion .F! ' Th e exte nt of ma labsorption may be quantified by co mparing breath H 2 excretion followi ng a test dose of lactulose with tha t after ingestion of a d ifferent carbohydrate.l->' Fer mentation of a readily fer men tab le carbohydrate result s in aci d ificatio n of co lonic co ntent s. Us ing a pH-sensitive rad iotelem etry de vice, Bown et aL/ 231 showed that lactul ose 30 to 40g da ily decreased the pH of the right co lon fro m contro l le vels of 6.0 to 4.8 5. As the intes tinal co ntent reached th e left co lon and rectum, the med ian pH rose, prob ab ly as a res ult of SCFA ab sorpt ion and bicarbon ate sec retion. Th e effect of lactulose o n faeca l pH depend s on the dose employed, and the drop in pH is usu ally maint ain ed throughout the co lon and the stoo l when larger doses of lactul ose are ad mi nistered.124.261 Alth ou gh co ns istent cha nges in the re lative propor tion s or in the abso lute numbers of d iffere nt bacterial spec ies have been difficul t to ide ntify in pat ient s treated with lactul ose,1 27-29 1this co mpou nd does appea r to have a significant effect on the metab oli c ac tivities of the microb es. In healthy vo lunteer s, faeca l level s of~-gal acto s id ase activity increased significantly after an 8-da y period o f a non di ar rhoeogeni c dose of lactu lose (20g twice daily j.!'" Co mpa red with day I , caecal co ntents sa mpled on day 8 afte r lactul ose adm inist ration exhibit ed a mar ked fall in pH , a faster disappearance of lactul ose and its co ns titutive hexoses, and increased co ncentrations of SC FA and lactate.l'" The co lonic microflora therefore are able to ada pt to ferme nt lactul ose, and the more efficie nt bacteri al fermentatio n of lactul ose mod ifies the d iarrhoea indu ced by a larger load of th is sugar.P'!' Th e co lonic bacter ia use ferme ntable ca rbo hydrate, e.g . lactul ose, as a so urce of ene rgy, thu s prom oting their gro wth and multipl ication. P!' It is bel ieved that bacteri al growth ass im ilates the ammoni a produ ced by degradation and putrefaction of ingested prot ein s as a so urce of nitrogen , and faeca l nitrogen exc retion increas es 2-to 4-fo ld after lactul ose admi nistration.tv-P! Lactul ose is frequently used as an effective laxative in the management of constipation.P"! Based o n the ab sence of its respecti ve disaccharidase, the ration ale for the use of lactulose as a laxative was to indu ce d isacch aride malab sorption, a we ll known ca use of diarrh oea. In spite of the freq uency and impo rtan ce of ca rbo hy drate-i nd uce d d iarrhoea, its patho physiology is not full y understood . Until re lative ly rece ntly, SC FA were thou ght to be poorl y abso rbed and to functi on as an irrit ating and os mo tic fo rce stimula ting intes tina l motilit y and imp edin g co lo nic water and e lec tro lyte absorption .135.3xl Early work on carbohydrate malab sorption therefo re impli cated bacter ial ferme ntation as the prim ary mechani sm for diarrhoea in di saccharid e malabsorption (' ferme ntative di arrh oea ' ). Th e stro ng co rre lation bet ween faecal water o utput and output of SCFA135. 371 was thou ght to incriminate SCFA in the path ogenesis of diarrhoea. As co lonic SCFA level s remain mor e or less co nstant desp ite dietary cha nges, any fac tor increasin g stool we ight will, ho we ver, increase their o utput. Later in vivo studies of intes tinal SCFA abso rption have consistent ly de mo nstrated efficient absorptio n of SCFA from the human co lon.' I 1.1 21 Th is absorption mar k-edly enha nces net sodi um and wa ter abso rptio n and bicarb onate secretion int o the lume n. Thu s, the intracolonic fer me ntation of non absorbed suga rs into rea dily absorbable SCFA reduces the osmotic load , and suppresses or mit igates the d iarrhoea wh ich wo uld have res ulted from unm odified ileal output. Th e prot ect ive ro le of col oni c ferme ntation in redu cing the severity of carbohydrate-induced diarrho ea has been dem on str ated in studies co mparing the o utput of faecal water in respon se to a nonfe rme ntable os mo tic load with a load of ferm entable ca rbohydrate. Hammer et al. 1391 co mpared the d iarrh oea res ulting fro m incr easin g iso -os mo lar loads of the nonabsorbabl e, non ferment abl e and elec trically neut ral co mpo und polyeth ylen e g lyco l (PE G) and lactul ose in healthy ind ividual s. Increasing os mo tic loads of PEG ca use d a near-li near increase in stool wa ter o utput. At any load, lac tu lose ind uced less diarrhoea than d id PEG . Wit h low (45 g/day) or mode rate (95 g/day ) do ses of lact ulose, stool wa ter losses were redu ced by as much as 600 g/day (co mpared with equimo lar osmotic loads of PEG ). As lactul ose doses of >45 g/day we re inge sted, faecal ca rbo hydrate exc retion rose progressively, thereb y co ntributing more and more to the osmo tic dr iving force for di arrh oea. With the largest dose of lactul ose (125 g/day) the differ en ce between lactulos e-and PEG-induced diarrh oea markedly decreased . In lactulose-induced d iarrhoea, daily stoo l co llections co nta ined an average of I, 12 and 45 g of ca rbohyd rate whe n the participant s wer e ingesting 45 , 95 and 125g of lactul ose/d ay, respecti vely, indi catin g a maximum ca pac ity of th e co lo nic bact eria to met aboli se lactulose of approxi ma tely 80 g/day.' 39I Saund ers and Wiggi nsl' ! obtai ned dose-response curves fo r mann itol (a poorly absorbed sugar alcohol ), raffi nose, lactulose an d magnesium sulfate in relation to faecal wa ter ou tp ut over 48 hours afte r the dose of test substance. With magnesium sulfate there was an im med iate inc rease in faecal output, whereas a lag period was observed for the nonabsorbed sugars. When an increase in faecal water occurred after ingestion of mannitol, lactulose or raffinose, the sugar or its constitutive hexoses appeared in the stools. The dose-response experiments indicated that the colonic flora can normally convert 40 to 60g of single doses of carbohydrate to SCFA without the individual experiencing diarrhoea.I" The human colon is obviously capable of removing appreciable amounts offermentable carbohydrates from colonic contents, even in patients with malabsorbed carbohydrate.Ue'" If, however, the fermentative capacity of the colonic microflora is exceeded, unmodified and osmotically active sugars may produce an osmotic diarrhoea. The response threshold when a given amount of carbohydrate is malabsorbed differs among individuals, In the study by Saunders and Wiggins.U' one participant tolerated a 73 mmol (25g) dose of lactulose whereas another tolerated 176 mmol (60g) before faecal water output rose above 400 ml/48 hours and test carbohydrate appeared in the stool. Similarly, doses of 120 mmol (22g) and 350 mmol (64g) of mannitol were tolerated by different participants. The individual variability may, at least in part, be explained by differences in the fermentation capacity of the bacterial flora. Individuals with a limited fermentation capacity or a high dependence on the existing capacity for colonic carbohydrate degradation might be especially susceptible to an increased carbohydrate load. A similar situation may apply to the development of microbial fermentation in newborn animals. Argenzio et al. 1401 produced carbohydrate malabsorption by administration of corona virus to newborn pigs. This resulted in diarrhoea in 3-day-old pigs, with high concentrations of carbohydrate and low concentrations of SCFA in colonic contents. Pigs 3 weeks of age showed no diarrhoea, with negligible amounts of carbohydrate and normal levels of SCFA in colonic contents, indicating that the more mature micro-'£JAdis International limited. All rights reserved. Clausell & Mortellsell flora in these older piglets were capable offermenting the osmotic load to SCFA, which could then be rapidly absorbed from the colonic contents. Once the microbial process is well established, prolonged ingestion of carbohydrate results in an improved colonic fermentation capacity.l'v'" The higher capacity to ferment carbohydrates may be a result of enzymatic induction in existing rnicro-organisms!''! and/or an alteration in microbial populations.lv-'!' Flourie et al. 1301 studied healthy volunteers for 2 test periods, at the beginning of which they ingested a diarrhoeogenic load (60g) of lactulose; the 2 periods were separated by a lactulose feeding period of 8 days, during which a nondiarrhoeogenic load (20g) of lactulose was taken twice daily. Stool weight and frequency, and faecal outputs of carbohydrates after the ingestion of lactulose 60g, dropped significantly after the lactulose maintenance period (20g twice daily for 8 days), indicating that the colonic flora can adapt to a nondiarrhoeogenic load of lactulose, and that this adaptation has a beneficial effect on the diarrhoea induced by a larger load of this sugar. Similarly, Launiala!"!' reported on a 3-week period during which an infant with congenital lactose malabsorption was maintained on a lactosecontaining diet. In comparison with the first day of treatment, continued treatment resulted in a decreased stool volume and disaccharide output, and a sharp rise in faecal lactate concentration. No adaptation occurred in the small intestine.H!' Hypothetically, regular consumption of carbohydrates mal absorbed in the small bowel may result in a more efficient colonic carbohydrate fermentation and an increased response threshold when a given amount of carbohydrate is malabsorbed. In addition to differences in colonic bacterial fermentation capacity, variations among individuals in the buffering capability of the colon may contribute to individual differences in the response to identical amounts of nonabsorbable carbohydrates. It is known that caecal pH decreases even with the amounts of unabsorbed carbohydrate normally entering the colon .F" Furthermore, patients with severe carbohydrate malabsorption, ei- ther due to disaccharidase deficiency or due to ingestion of excessive amounts of lactulose, have decreased faecal pH in the range of 4 to 5. 124 , 26, 35] This decrease in pH is sufficient to inhibit bacterial fermentation and SCFA production as shown in vivo and in vitro. 12 6,4 2,43 j In subjects with a restricted colonic buffering system, the tolerance of carbohydrate malabsorption may be decreased as the low pH decreases fermentation leading to accumulation of undigested sugars and osmotic diarrhoea early on in the process . Through the Osmotic Activity of Unfermented Carbohydrate? Many physicians believe that the change in stool output in the diarrhoea of carbohydrate malabsorption is due to the osmotic effect of malabsorbed and nonfermented sugars. This assumption seems to be supported by the studies mentioned above (sections 3.1 and 3.2) in which diarrhoea ensued when carbohydrate appeared in the stools. The study by Holtug et aI.,1 26 1however, queries this assumption. Healthy individuals were given lactulose twice daily for 3 consecutive days in each experimental period. Doses were doubled (20, 40, 80, 160 and 320 g/day) from each period to the next, with the end-point being diarrhoea in excess of 1000 g/day. Faecal output responded differently among individuals, and 2 types of behaviour were identified ( fig . 2 ). In accordance with the interpretation that carbohydrate-induced diarrhoea is due to the osmotic effect of nonfermented mal absorbed sugars, diarrhoea occurred suddenly in six of 12 individuals in association with the appearance of carbohydrate in faeces . In the 6 other subjects, however, faecal output gradually increased with increasing doses of lactulose and diarrhoea occurred before the appearance of faecal carbohydrates.P''! How does unabsorbed carbohydrate cause diarrhoea in those cases? The mechanism of action of lactulose as a laxative is probably multifactorial, involving effects on both the colon and the small bowel. Under normal circumstances, about 1500ml of fluid is estimated to enter the colon each day. 14411n disaccharide malabsorption an increased volume of fluid is retained in the lumen of the small intestine by the osmotic effect of disaccharide. 141,451 Analysis of distal ileal fluid collected during the passage of a lactose meal in lactase-deficient people and a lactulose load in healthy people indicates that about two-thirds of the osmotic load entering the colon consists of endogenous electrolytes. 130,41,45] Thus, the water load delivered to the colon is about 3 times that calculated to be osmotically held by the nonabsorbed sugar. A logical extension of the above observations is to ask what the colon can accomplish under stress. In fact, the colon has a surprisingly large capacity to adapt to fluid overload. When the healthy human colon was subjected to a slow, continuous infusion of isotonic fluid (1.4 to 2.8 ml/min), net absorption of water increased markedly, up to 5 to 6 L/day.146 1 The rate and pattern of t1uid entry are , however, of © Adi s International Limited . All rights reserved. Clausen & Morten sen major importance. Rapid infusion (8.3 ml/min) of a single bolus of 250ml of t1uid did not influence faecal output whereas 500ml delivered at the same rate overwhelmed the absorptive capacity of the colon and produced liquid stools.U''!In healthy human volunteers, Chauve et al. 1471 investigated pressure recordings in the right , transverse and left colon, while isotonic saline was continuously infused in the caecum. The data obtained indicated that once a certain volume of content is reached in the proximal part of the colon, it contracts and propels its content.l''?' which may explain the different colonic responses comparing slow continuous and rapid bolus infusions of t1uid into the caccum .lv ' The extra volume load to the colon due to lactulose can be estimated to be 1.65L at a lactulose intake of 60g,130,41,45 1 and the sudden arrival in the caecum of this considerable amount of ileal discharge might lead to colonic 'decompensation' and prompt colonic evacuation. A single primary mechanism for the laxative effect of lactulose cannot be determined because of multiple possible effects. Most likely , the laxative effect of lactulose results from the combination of: • water retention in the small intestine through the osmotic activity of the unabsorbed disaccharide; and • interference with net t1uid absorption in the colon due to the osmotic effect of malabsorbed and intact sugars when the bacterial fermentation capacity is exceeded. Whether or not the laxative effect of lactulose is accomplished is probably just a matter of dose . It is, however, important to realise that the response threshold when lactulose is prescribed differs among individuals ( fig . I) . A gradual increase in faecal output with increasing amounts of lactulose occurs in some individuals, while others suddenly respond with severe diarrhoea.P'" Unfortunately, it is not possible to predict the response of a given individual, and firm guidelines for dosage recommendations are impossible. Lactulose has been successfully used in the treatment of hepatic encephalopathy ever since the pioneering experiments of Bircher et al. in 1966 .' 241 The exact mechanisms whereby lactulose exerts its beneficial effect on hepatic encephalopathy are not fully understood, but several hypotheses have been suggested. Hepatic encephalopathy is a complex neuropsychiatric syndrome that appears to be characterised predominantly by augmented neuronal inhibition. The syndrome is associated with hepatocellular failure, increased portal-systemic shunting of blood and multiple metabolic changes, and is generally considered to be a potentially reversible metabolic encephalopathy. The precise pathogenetic mechanism responsible for the CNS dysfunction remains to be determined. Products of bacterial metabolism in the colon, which are normally absorbed and extracted by the liver, tend to accumulate in the peripheral blood of patients with liver disease either because the diseased liver is unable to remove these compounds from the portal blood perfusing it, or because portal hypertension has led to the development of venous collaterals which transmit portal blood directly into the systemic circulation. If one or more of these metabolites can cross the blood-brain barrier and promote neural inhibition, they may contribute to hepatic encephalopathy. Despite decades of research endeavour, the nature of the toxic, gutderived, substances defies identification. Three candidate toxins seem likely pathogenetic factors because they are present to excess in patients with liver failure and cause coma experimentally in animals. These are ammonia, SCFA and medium-chain fatty acids (MCFA), and mercaptans; of these, ammonia probably plays the most central role.'48I Ammonia occupies a central position in nitrogen metaboli sm in the colon, both as an end-product of bacterial metabolism of amino acids , peptides and proteins.P?' and as the simplest nitrogenous compound which can be used by bacteria as a starting material in the synthesis of their own nitrogenous constituents. In the absence of active colonic fermentation most of the ammonia formed is absorbed to be converted to urea in the liver. When lactulose was first used in the treatment of hepatic encephalopathy, it was presumed that the colonic acidification favoured the growth of acid-tolerant lactobacilli and other acidophilic, fermentative bacteria and reduced the growth of acidophobic, proteolytic bacteria responsible for ammonia formation .P'! However, quantitative studies failed to confirm this rearrangement of gut f1ora. 127 , 281 Alternatively, luminal acidification would reduce the proportion of ammonia available for reabsorption by passive nonionic diffusion by increasing the ratio of ionised to unionised ammonia, thereby enhancing the faecal excretion of ammonia.P'!' Attempts to demonstrate increases in stool ammonia caused by lactulose administration have failed,151 , 521 except when excessive amounts of lactulose (>80 g/day) have been given.F'l ln that case, stool ammonia increases approximately 3fold . 1331 The main point is, however, that ammonia constitutes only a small percentage (3 to 5%) of total faecal nitrogen. The debate on whether or not lactulose increases faecal excretion of ammonia is therefore of minor importance in the context of total nitrogen excretion. Of more importance for faecal nitrogen clearance are other sources of nitrogen, constituting about 95% of total stool nitrogen. By use of an anaerobic faecal incubation system , Vince et al. I171 demonstrated that net generation of ammonia by faecal bacteria can be converted to net utilisation of ammonia by the provision of a fermentable source of energy, e.g. lactulose, ammonia disappearing from the incubation system. The results may be explained by a combination of factors: (a) the availability of a readily fermentable substrate may encourage bacterial proliferation and assimilation of ammonia as a nitrogen source;' 17,31 I (b) preferential use of lactulose as a carbon and energy source may exert a sparing effect on the metabolism of both exogenous and endogenous aminated compounds with a subsequent decrease in the generation of ammonia by colonic bacteria;117, 191 and, (c) low pH values such as are induced in the colon by lactulose may bring about a general reduction in bacterial metabolism, including that of ammonia-producing bacteria.l 17 ,19,26,421 If, in the presence of lactulose, incorporation of ammonia into bacterial protein is stimulated, faecal nitrogen excretion should increase and hepatic urea production and total body urea pool should decrease during administration of lactulose. This is exactly what Weber l321 and Mortensen'Vl have shown. Administration of lactulose caused a 2-to 4-fold increase in faecal nitrogen content in patients with cirrhosisl 321 and in healthy individuals,1 331 and the increase in nitrogen excretion was accompanied by a significant reduction in the urea production rate leading to a reduction in the total body urea pool.132 1 Using recently developed techniques that permit quantitative separation of faecal solids into bacterial, soluble and fibre fractions, Weber et al. 1531 analysed the effects of lactulose on the compartmentalisation of faecal nitrogen. In patients with cirrhosis, administration of lactulose (56 ± 6 g/day) caused a marked increase in both the solid weight and nitrogen content of the bacterial and soluble fractions of stool ; lactulose administration increased nitrogen excretion in the bacterial and soluble fractions by 165 and 135%, respectively. The major increment in nitrogen content of the bacterial fraction indicated that lactulose stimulated bacterial growth and multiplication significantly, and thereby increased the incorporation of nitrogenous compounds such as ammonia into bacterial protein. The increase in faecal soluble nitrogen may be explained by reduced bacterial catabolism of nitrogenous compounds to absorbable metabolites, e.g. ammonia, due to: (a) the introduction of a carbohydrate which is preferentially used as a carbon and energy source;117, 191 (b) an increase in the hydrogen ion concentration which may bring about a general reduction in bacterial metabo-<0 Adi s Internation al Limited . All right s reserved. lism;117,19,26,421 or (c) both. A direct effect of lactulose in reducing the colonic transit time may also be a factor contributing to an increase in faecal soluble nitrogen . The suggestion that SCFA and MCFA may play a role in the pathogenesis of hepatic encephalopathy is mainly based on animal experiments. Samson et aI.i 54 1 investigated the narcotic action of SCFA (acetate, propionate, butyrate, valerate and hexanoate) and MCFA (heptanoate and octanoate) in experimental animals , Intravenous or intraperitoneal injection of the sodium salt s of the se acids produced a reversible coma with a definite relationship between the amount of the fatty acid which would produce unconsciousness and the carbon length of the compound; i.e. the longer the carbon chain, the more potent the encephalopathic effect, acetate being almost nontoxic.P'l In a subsequent experiment, White and Samson l551 demonstrated that intravenous injection of propionate, butyrate, valerate and hexanoate immediately induced electroencephalographic (EEG) changes from a resting type to one of sleep in unanaesthetised rabbits . During these EEG alterations, overt signs of drowsiness or sleep always occurred and the typical 'alert' EEG responses to nociceptive and auditory stimuli were diminished or absent. Again , the longer the carbon chain, the more potent the effect. In rats , the rate of bloodbrain barrier penetration of straight-chain saturated monocarboxylic acids has also been shown to increase with chain length and is virtually complete at lengths greater than that of hexanoate (C6).1 561 In humans, octanoate has been shown to cross the blood-brain barrier in both normal individuals and patients with cirrhosis.lV' The mechanisms by which SCFA and MCFA interfere with CNS functions are not entirely clear. In vitro and in vivo stud ies have shown that SCFA and MCFA inhibit cerebral Na+,K+-ATPase activity,15X,59 1 and, as for the narcotic potency, the inhibitory capacity increases with increasing chain length.P'" The con sequences of inhibiting thi s enzyme would include an increase in intracellular sodium and possibl y impaired neurotransmi ssion. Muto and Takahashi t'v ' were the first to sugges t a role of SCFA in the pathogenesis of hepatic encephalopathy. In patients with hep ati c coma, pla sma level s of SCFA of 4-to 6-carbon length were increased 3-to 4-fold compared with controls and patients with other kind s of coma. Significantly elevated levels of SCFA in peripheral ven ous blood of patients with fulminant hepatic failure and portal sys temic encephalopathy have been confirmed recently. Lai et al. 1611 investigated 6 patients with fulminant hepatic failure due to paracetamol (acetaminophen) intoxication and hepatitis. All patients were in grade IV coma. Pla sma level s of SCFA were significantly elevated in patients with encephalopathy (2034 ± 11341lmo1/L; mean ±2 SO ) co mpared with 10 healthy controls ( 1113 ± 662 um ol/L), but a correlation between SCFA con centrations and the clinical course could not be demon strated. Clau sen et al. 1621 inve stigated 32 patients with cirrhosis, 15 patients with and 17 patients without hepatic ence phalopathy, and II health y indi viduals. Plasm a concentrations of SCFA were significantly elevated in patients with hepatic en cephalopathy (362 ± 831lmo1/L; mean ± SEM ) co mpared with patients without encephalop athy ( 178 ± 57 umol/L) and healthy controls (60 ± 8IlmoI/L). Repetitive sa mpling from the encephalopathic patients showed no relationship between SCFA concentrations and the grade of encephalopathy.I '<' In comparison with the encephalopathy associated with inborn errors of organic acid metabolism, el e vated levels of SCFA in patients with hepatic encephalopathy are modest. Massive elevations of propionate are found in severe and fatal ca se s of propionic acidaemias (5400 to 38 000 Ilm oI/L ),1 63-651 and level s of isovalerate in the rang e of 340 to 2990 umol/L are associ ated with stupo r and uncon sciousne ss in children with isovaleric ac idaemia.166,671 Moreover, an 8-to IO-fold rise of veno us propionate and oct an oate (C8) is seen in pat ients with liver cirrhos is after duodenal