TheBritish Society of Gastroenterology - BMJ

Gut, 1974, 15, 822-841

The British Society of Gastroenterology

The following are abstracts of papers given at the annual meeting held at the University of Birmingham on19-21 September. On 19 September there was a discussion organized by the Liver Group (Convener, ProfessorSheila Sherlock) on 'The carrier of hepatitis B antigen'. In the afternoon there was a symposium shared by theBritish Society for Digestive Endoscopy on 'Diagnosis of pancreatic disease' (Moderator, Dr Sheila Waller),and a clinicopathological conference on 'Progressive intestinal disease in an adolescent' (Moderator, Dr W. T.Cooke). The Sir Arthur Hurst Memorial Lecture was given by Dr A. F. Hofmann on 20 September. Thescientific sessions followed on 20 and 21 September. -A fuller account of the meeting appears in 'Notes andactivities' on page 844, together with the names of the officers of the Society and.those of the newly electedCouncil.

P1

The extent of rectal premalignant changeIn ulcerative colitisR. H. RIDDELL (introduced by B. C.MORSON) (St Mark's Hospital, London) Ithas been suggested that in patients withulcerative colitis regular rectal biopsy canbe used for detecting premalignant changeand that this reflects similar changes inthe proximal colonic mucosa1. To assessthe accuracy of this policy the frequencyand extent of premalignant change in therectum was measured using an eyepiecemicrometer on slides of rectal mucosafrom 56 proctocolectomy specimens.Forty of these specimens had carcinoma(25 rectal and 15 above the rectosigmoidjunction), and a further 16 had premalig-nant change anywhere in the large intes-tine but no carcinoma.

Five out of the 25 patients with rectalcarcinoma had premalignant changearound the tumour only and this involvedless than half of the mucosa examined. Intwo out of the 15 patients with cancer inthe colon there was no evidence of pre-malignant disease in the rectum and in afurther two patients it involved less thanhalf of the rectal mucosa. Among the 16specimens without carcinoma premalig-nant change in one patient did not extendto the rectum, and in a further five halfor less of the rectal mucosa examined wasinvolved. These changes were least markedin the rectum in four patients and mostmarked in three.

In view of these findings it is suggestedthat (a) ideally the site of rectal biopsy todetect premalignant change should be

recorded and multiple biopsies takenfrom different parts of the rectum at eachexamination, (b) consideration should begiven to a policy of colonoscopy withmultiple biopsy in all patients with ex-tensive colitis who are statistically at in-creased risk of malignant change. Thefrequency with which colonoscopic ex-aminations should be performed has yetto be determined.

Reference

'Morson, B. C., and Pang, L. S. (1967), Gut, 8, 423.

P2

Diet and growth after resection of ileum inchildhood

H. B. VALMAN (Northwick Park Hospitaland Clinical Research Centre, Harrow,Middlesex) Twelve children who havesurvived resection of more than 45 cm ofileum (eight during the neonatal periodand four later in childhood) have been re-assessed at periods between three and 16years. Two children received a low-fat,high-protein diet for nine and 13 yearsafter resection and showed no advantagein growth compared with 10 children whohad a normal diet two years after resec-tion. No child had persistent diarrhoea forlonger than two years after resection. Theweights of individual children are appro-priate for their heights. Their heights tendto be less than those of their siblings butonly one is less than the third centile. Twochildren are below the third centile forthe height expected from their parents'heights. Despite the lack of diarrhoea theolder children still have marked steator-

rhoea, although serum levels of vitaminA, D, and E are normal. In contrast tothe adult, a normal diet usually is indicatedtwo years after extensive resection of ileumin childhood.

P3

Perianal Crohn's disease

J. ALEXANDER-WILLIAMS, D. M. STEINBERG,J. S. FIELDING, H. THOMPSON, AND W. T.COOKE (Nutritional and Intestinal Unit, TheGeneral Hospital, Birmingham) The nat-ural history and results of surgical treat-ment of perianal disease were assessed in418 patients with Crohn's disease.A prospective review was made of 109

patients initially diagnosed as havingperianal Crohn's disease in 1968. All butfive have been followed and examinedafter five years, following a policy of con-servative management. In the majority(78Y) the disease status was unchanged;in only 16% had the disease deterioratedin the five years. Continued or intermit-tent symptoms occurred in 35 %.A separate review of 96 local anal

operations in 68 patients with Crohn'sdisease has shown that the anal lesions arenot necessarily serious or progressive andsimple establishment of drainage gavegood results in 70 %. Incontinence islikely to be the result of aggressive surgerynot of progressive disease.Even in the presence of underlying

active bowel disease good results occur in54% and the eradication of proximalactive disease is not a necessary prerequi-site, contrary to the views of Garlock(1967). Some patients may remain in good

822

on January 3, 2022 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.15.10.822 on 1 October 1974. D

ownloaded from

http://gut.bmj.com/

The British Society of Gastroenterology 823

health with minimal symptoms even in thepresence of gross anal stenosis withfissures and fistulae.

Although an occasional indication forradical surgery, perianal Crohn's diseasecan be managed conservatively in mostpatients.Reference

Garlock, J. H. (1967). 'Surgery of the small intes-tine' in surgery of the alimentary tract.London: Butterworths.

P4

The HL-A system and the immunologicalresponse to dietary antigens

B. B. SCOTT, L. M. SWINBURNE, S. M. RAJAH,AND M. S. LOSOWSKY (Department ofMedicine, St James's (Universitv) Hospital,Leeds) There is ample evidence of relation-ships between histocompatibility antigensand immune responses in animals but notin man. The close association betweenHL-A8 and coeliac disease (CD) mightsuggest that immunological abnormalitiesin CD are related to the HL-A system.High levels of serum gluten antibodies

are common in CD and in other condi-tions with damaged gut mucosa and maymerely reflect absorption of dietary pro-teins. However, gluten antibodies are notfound in all such patients and are foundin some patients with chronic liver disease.The relationship between these immuneresponses to gluten and histocompatibilityantigens is not known.Serum gluten and milk antibodies and

HL-A-status were studied 'blind' in 124subjects with various conditions. Glutenantibody titres were significantly higher inHL-A8-positive subjects (p < 0O0005;limit 1/128). If the 23 CD patients areexcluded gluten antibody titres are stillsignificantly greater (p < 0-005). Further-more, milk antibody titres tend to behigher in HL-A8-negative subjects. Thusfrom both antibodies a discriminantfunction can be calculated which separatesHL-A8-positive and -negative subjectswith a high degree of certainty.These results suggest that in man as well

as in animals histocompatibility antigensare related to immune responses. Therelevance of these findings to the aetiologyof CD will be discussed.

P5

Enteric loss of lymphocytes in coeliacdisease and in Crohn's disease

A. P. WEETMAN, J. HAGGITH, AND ADRIAN P.

DOUGLAS (Gastroenterology Group,Department of Medicine, Royal VictoriaInfirmary, and Department of MedicalPhysics, General Hospital, Newcastle uponTyne) Normal small intestinal epithe-lium contains interepithelial lymphocytesbut their fate is unknown'. The observa-tion that in both coeliac disease' andCrohn's disease2 there is an increase inthe number of interepithelial lymphocytesin the small intestine suggested that inthese conditions there may be an increasedloss of lymphocytes into the bowel lumen.Such a situation is known to occur inlymphangiectasia3. We have thereforestudied the distribution and faecal loss ofperipherally harvested lymphocytes label-led with 51-Cr and reinjected into humansubjects. Localization of these labelledlymphocytes has been made by externalcounting and by counting of faecal, urine,and blood samples daily for five days. Thedistribution of the lymphocytes withinthe body was different from normal inpatients with coeliac disease. In fivenormal subjects enteric loss of lympho-cytes over five days was 0-24% (SD ±0-16). In three cases of lymphangiectasiathe five-day faecal loss was 0 47 %, 0 55 %,and 0 95% respectively. In two untreatedcoeliac patients faecal loss was 0-66% and1-85% and in two patients with Crohn'sdisease faecal loss was 1-45% and 2-11 %.The results have been correlated withinterepithelial lymphocyte counts inmucosal biopsies and with measurementof 51-Cr in resected intestine.Our observations suggest that there is

normally a pathway of lymphocyte re-moval into the gastrointestinal tract ofprobable importance in lymphocyte mi-gration streams4. The finding of an in-crease in this lymphocyte loss may explainsome of the immunological abnormalitiesobserved in the conditions studied, andmay yield a method of screening for in-flammatory bowel disease and coeliacdisease.

References

'Ferguson, A., and Murray, D. (1971). Gut, 12,988.'Otto, H. F. (1973). Current topics in pathology,

57, 81.3Weiden, P. L., et al (1972). J. clin. Invest., 51, 1319.'Yoffey, J. M., and Courtice, F. C. (1970). Lymph,

lymphatics and the lymphomyeloid complex.Academic Press, London.

P6

Tropical sprue in Rhodesia

G. E. THOMAS, AND D. J. CLAIN (Department

of Medicine, University of Rhodesia,Salisbury, Rhodesia) The objects of thisstudy was to demonstrate the existence oftbopical sprue said not to occur in Africa.We investigated 38 megaloblastic anae-mias and three patients clinically thoughtto have malabsorption syndrome. Fullhaematological investigation and compre-hensive small intestinal work up was done.Twenty-four patients had tropical sprue.The following details concern only these24. Clinically they presented typically withanaemia, weight loss, anorexia, abdominalpain, and diarrhoea. Of the 21 with mega-loblastic anaemia, the mean haemoglobinwas 5-4 g; six had a low serum B12 andfolate, two a low B12 only, and nineeither a low serum or red cell folateTwenty-one had severely abnormal jejunalbiopsies. Twenty-two had abnormal D-xylose absorption. All 24 had impairedCo57 vitamin B12 absorption. Steator-rhoea occurred in only eight, perhaps dueto inadequate fat-loaded diets, but thetropical sprue group excreted markedlymore faecal fat than a control group.Eighteen had malabsorption patterns onsmall intestinal radiology. Six had atro-phic gastritis and achlorhydria, fivesuperficial gastritis, and 12 hypochlorhy-dria. There was no statistical differencebetween the diets of the patients andmatched controls. Patients were rein-vestigated in three groups. After sixmonths tetracycline and haematinics, thecombined regimen; after six months tetra-cycline alone; and some, that subsequentlyjoined one or other of these groups,after two to three weeks' tetracyclinealone. To date 15 abnormal biopsies havebeen repeated. Eight were treated with thecombined regimen and seven with tetra-cycline alone; all returned to normal.D-xylose absorption improved in all nineafter the combined regimen, six to morethen 5 g; in all eight after six months'tetracycline, five to more than 5 g and inall seven on two to three weeks' tetra-cycline, four to more than 5 g. VitaminB12 absorption became normal in sevenof eight on the combined regimen, theother improved; after six months' tetra-cycline all nine returned to normal; andfour were normal and three improved ontwo to three weeks' tetracycline. Faecalfat levels improved in all four on the com-bined regimen, in all six on six months'tetracycline, and in four of five on theshort course of tetracycline. The meanweight gain was 9 kg on the combinedregimen, 10 kg on six months' tetracycline,and 2-9 kg after two to three weeks'


http://gut.bmj.com

/G


ownloaded from

http://gut.bmj.com/

824 The British Society of Gastroenterology

tetracycline. Six megaloblastic anaemiasachieved a haematological response ontetracycline alone, all six haemoglobins re-turning to normal. Of four bone marrowsrepeated, to date, two have becomenormoblastic and two showed minimalmegaloblastic change, having been severe-ly megaloblastic. Our patients had amalabsorption syndrome which respondedto conventional therapy for tropical sprue.We have excluded coeliac disease, abdo-minal tuberculosis, lymphomas, parasites,and diet as alternative causes.We thereforeconsider tropical sprue to occur in centralAfrica and to cause most of our megalo-blastic anaemias.

P7

Observer variation in the assessment ofresults of surgery for peptic ulceration

F. T. DE DOMBAL, JANE C. HORROCKS, ANDSUSAN E. CLAMP (The General Infirmary,Leeds, and the County Hospital, York)Results of surgery for peptic ulcerationare usually graded according to (a) pre-sence or absence of various postoperativesymptoms; (b) severity of these symptoms;and (c) an overall grading according tothe classification of Visick (1948). Thepresent paper examines the observer varia-tion inherent in these three assessmentmethods.Some 232 patients were studied (fol-

lowing peptic ulcer surgery) presenting tothe clinics of the Leeds/York gastricfollow-up group. Each patient was re-viewed by the usual panel of observers;concomitantly a separate (blind) assess-ment was made by a further panel ofobservers, the results being subsequentlycompared.A very high degree of observer agree-

ment was recorded between observers andbetween panels both in York and in Leeds(95 %) concerning the presence or absenceof specific postoperative complaints. Lessagreement (under 90 Y.) was recorded con-cerning the severity of symptoms. As re-gards Visick classification, initial agree-ment was low (64% in York, 68% inLeeds) but rose with experience to over80% in both centres.These data suggest that the current

degree of observer agreement in the clinicsstudied is high. However, 'inter-series'comparisons based on overall Visickclassifications should be viewed withextreme caution. Future comparisonsshould concentrate upon the presence orabsence of postoperative symptoms.

Reference

Visick, A. H. (1948). Ann. roy. Coll. Surg. Engl., 3,266.

P8

Antibiotics in biliary disease

M. R. B. KEIGHLEY, R. DRYSDALE, D. W.BURDON, AND J. ALEXANDER-WILLIAMS(The General Hospital, Steelhouse Lane,Birmingham) In patients undergoing bi-liary operations, and particvlarly thosewho are jaundiced, there is a close corre-lation between the presence of micro-organisms in the bile and the develop-ment of wound sepsis and septicaemia.To investigate the value of chemotherapyin preventing these sequelae an antibioticexcreted almost entirely in bile (rifamide)and one with low bile but adequate serumconcentrations (gentamicin) were com-pared with controls.A randomized trial was conducted in

150 patients who were allocated to re-ceive rifamide, gentamicin, or no anti-biotic. Infective complications were re-corded by an independent observer. Ade-quate antibiotic levels were assumed if theconcentration was twice the minimuminhibitory concentration for more than80% of biliary organisms.

In non-jaundiced patients adequatebile concantrations occurred in 97% onrifamide and 15% on gentamicin butamongst jaundiced patients levels wereonly 10 and 0% respectively. Adequateserum concentrations were found only inpatients on gentamicin (87 %). Bacter-bilia was significantly reduced only in thenon-jaundiced rifamide patients (p <0-05). Only in the gentamicin group waswound infection and bacteraemia signifi-cantly less than controls (p = 0-05 and< 0-02 respectively). Also bacteraemiawas eliminated in the jaundiced patientsreceiving gentamicin (p = 0-05) but not inthose on rifamide.We conclude that adequate serum anti-

biotic levels are more important than bilelevels in preventing septic complicationsof biliary operations, particularly inpatients with jaundice.

P9

Percutaneous needle biopsy of the liver andlarge bile duct obstruction

J. S. MORRIS, G. A. GALLO, P. J. SCHEUER,AND S. SHERLOCK (The Departments ofMedicine and Histopathology, Royal Free

Hospital, London) The value and safetyof percutaneous needle biopsy of the liverhas been studied in 125 patients withmechanical obstruction confirmed at sur-gery or necropsy. Biopsies were done onlyon those patients with normal bloodcoagulation and adequate platelet counts.Jaundice, of varying degree (in somepatients > 20 mg//), had been presentfor from two weeks to two years.

There was no mortality followingbiopsy. One patient developed biliaryperitonitis and underwent surgery fromwhich he made an uneventful recovery.Minor complications occurred in sevenpatients and these all responded to medi-cal treatment. Percutaneous cholangi-ography with the finding of dilated bileducts, was performed successfully in 38patients only one of whom had a compli-cation following needle biopsy.

Histological features of large bile ductobstruction were seen in 83% of thepatients and in a further 13 % the biopsyshowed cholestasis. The initial biopsy re-port was misleading in only six patients.The histological findings were particularlyvaluable in 17 patients in whom previousdrug ingestion or the clinical history madea diagnosis of biliary obstruction lesslikely.

In patients with large bile duct obstruc-tion, therefore, percutaneous needle bi-opsy of the liver is useful in confirmingthe diagnosis and is of value in the inves-tigation of the patient with cholestasis ofuncertain cause. The incidence of signifi-cant side effects was low.

Plo

Effect of wheat bran on bile salt metab-olism and bile composition

E. W. POMARE, K. W. HEATON, T. S. LOW-BEER, AND C. WHITE (University Depart-ment of Medicine, Royal Infirmary,Bristol) Feeding bran reduces the en-trance into the bile of deoxycholate (DC),a colonic bacterial metabolite, and in-creases the proportion of chenodeoxy-cholate (CDC) in the bile'. To elucidatethe mechanism of these changes we haveperformed kinetic studies, using the iso-tope dilution technique after simultaneousiv administration of 14C-labelled cholate(C) and CDC, in six healthy subjects withcholecystographically normal gallblad-ders (two containing translucent stones).The effects of bran (mean 57 g/day takenfor four to six weeks) were as follows:DC pool shrank from 1-20 to 0-80 mmoles


http://gut.bmj.com

/G


ownloaded from

http://gut.bmj.com/


(p < 0025), CDC pool expanded from1-72 to 2-19 mmoles (p < 0 05), while Cpool and total bile salt (BS) pool were notsignificantly altered. Synthesis of CDCincreased from 0-46 to 057 mmoles/day(p < 0 025) but that of C was unchanged.Total BS synthesis was 1V24 and 1-46mmoles/day before and after bran (0 05< P < 010). The half-lives of C and CDCwere unchanged, indicating that bran haslittle if any bile salt-binding action.

In six subjects with radiolucent gall-stones, bran feeding (mean 48 g/day forfour to six weeks) consistently reducedthe cholesterol saturation of bile. Molarpercentage of cholesterol fell from 14-1 ±3-2 to 10-6 ± 2-7 (p < 0-0125).These findings suggest that bran, like

fed chenodeoxycholic acid, reduces choles-terol secretion into bile. A high fibre dietmay have a place in the prevention andeven treatment of gallstones.

Reference

'Pomare, E. W., and Heaton, K. W. (1973). Brit.med. J., 4, 262.

P11

Inhibition of external pancreatic secretionby extracts of ileal and colonic mucosa

A. A. HARPER, A. J. C. HOOD, J. MUSHENS,AND J. R. SMY (School of Pharmacy,Sunderland Polytechnic, Chester Road,Sunderland) Instilling oleic acid into thecolon or lower part of the ileum of anaes-thetized cats inhibits pancreatic secretion,stimulated by secretin and pancreozymin.Hage, Palasciano, Tiscornia, and Sarles(Biomedecine: in the press) have alsoobserved inhibition with oleic acid in thecolon of chronic dogs, but found that oleicacid in the lower ileum stimulated thepancreas.As the inhibitory effects in cats can be

elicited after cutting the vagal and splan-chnic nerves, it seemed likely that the in-hibition was humorally mediated. Wehave therefore extracted the terminalileal and colonic mucous membrane ofpigs and cats by a modification of themethod for preparing secretin and pan-creozymin (Crick, Harper, and Raper,1950). Intravenous infusion of the extractsinhibits the volume and enzyme outputof pancreatic secretion. Gallbladder con-traction in response to pancreozymin pre-parations is also inhibited. The inhibitoryeffect is not secondary to a reduction inblood flow. The behaviour of the extractson millipore filters and the abolition of

their activity by tryptic digestion suggestthat the active material is polypeptide innature.

Pancreotone is suggested as the appro-priate name for this pancreatic inhibitor.Its relation to GLI will be discussed.

Reference

Crick, J., Harper, A. A., and Raper, H. S. (1950).J. Physiol., 110, 367-376.

Al

Cholestyramine therapy in cholestatic liverdisease of children

R. NELSON, G. M. MURPHY, SUSAN EDKINS,SHEILA NUTrER, AND CHARLOTTE M. ANDER-SON (Institute of Child Health, Universityof Birmingham) Six children aged from 4months to 8 years with persistent chole-static liver disease from early infancy havebeen treated for up to 10 months withcholestyramine 6-12 g daily. In five ofthese, the response to treatment wasassessed in terms of (i) circulating liverfunction tests, (ii) liver histology, (iii) bileacids in duodenal aspirate.

In four patients there was improvementin the following liver function tests,whose mean values before and aftercholestyramine were: serum bile acids135 ± 130 and 29 ± 32 ,umols/litre(p < 0 05); serum bilirubin 4 2 ± 3-3 and1-2 ± 1-4 mg/I 00 ml (ns); serum alkalinephosphatase 79 ± 22 and 37 ± 18 KAunits/litre (p < 0-05); SGOT 201 ± 56and 95 ± 43 iu/litre (p < 0-05) andserum cholesterol 318 ± 72 and 202 ± 77mg/100 ml (p < 0-05). Histological evi-dence of cholestasis was considerably re-duced.Duodenal bile acid output (jpmols/10

min/kg body weight) measured followingintravenous pancreozymin was signifi-cantly higher in six controls (mean 19 ±11) than in the patients with cholestasis(4 ± 3) (p < 0-05). Following cholestyr-amine therapy, output in all four children(mean 34 ± 39) increased and was notthen significantly different from that ofthe controls. With the increase in bileacid output the proportion of glycocholicacid increased and that of monohydroxybile acids fell. In one, all liver functiontests have remained normal since stoppingcholestyramine two months ago.Of the other two patients, one showed

no improvement in any of the featuresstudied and, in the other, studies are notyet complete.

A2

Cell-mediated immunity to hepatitis Bantigen in blood donors with persistentantigenaemia or high titre antibody

W. M. LEE, W. D. REED, CHRISTINE G.MITCHELL, A. L. W. F. EDDLESTON, IANDYMOCK, AND ROGER WILLIAMS (Thie LiverUnit, King's College Hospital, London)The continued presence of either hepatitisB antigen (HBAg) or antibody to HBAgin the serum of asymptomatic blooddonors implies a faulty immune mecha-nism and a specific defect in cellularimmunity to HBAg has been proposed.In the present study 43 antigen-positiveblood donors and 28 with persistent anti-body were investigated for cellular im-munity to HBAg using the leucocytemigration test.An abnormal migration index was

found in 12 (28 %) of the antigen-positivegroup. In an additional 13 subjects,titres of HBAg in serum were higher thanthat used in the migration test chamberand sensitization could also have beenpresent in some of these. However, in 18subjects with low titres of HBAg themigration indices were normal. Serumtransaminase elevations and abnormalappearances on liver biopsy were presentin the majority of those with evidence ofcellular immunity to HBAg but were in-frequent in those with normal migrationindices. In the antibody-positive group,only five (18%) showed sensitization toHBAg and one of these had chronicpersistent hepatitis.The increased frequency of liver

disease in those carriers with evidence ofcellular immunity to HBAg is consistentwith the hypothesis that liver cell injuryis mediated by sensitized lymphocytes.

A3

The effects of portacaval shunting andportacaval transposition on serum gamma-globulin levels in the rat

I. S. BENJAMIN, C. J. RYAN, A. L. C. MCLAY,C. H. W. HORNE, AND L. H. BLUMGART(University Departments of Surgery andPathology, Glasgow Royal Infirmary, andUniversity Department of Pathology,Foresterhill, Aberdeen) The aetiology ofhypergammaglobulinaemia in liver dis-ease is uncertain (Lancet, 1972). Sincecirculating antibody in the normal animalis thought to be determined largely byintestinal flora, its elevation may reflect


http://gut.bmj.com

/G


ownloaded from

http://gut.bmj.com/


increased stimulation by intestinal antigenwhich has failed to be sequestered byKupffer cells (Inchley and Howard, 1969)due either to incompetence of these cellsor to intra- or extrahepatic shunting ofblood past the cells (Bjorneboa et al, 1972).Male rats (250-350 g) were subjected tooperation as follows:

Group I: 11 animals Portacaval anastomosisGroup II: 11 animals Portacaval transpositionGroup III: 19 animals Laparotomy with dissection

of portal vein and vena cava

Serum levels of y2-globulin were esti-mated weekly by radial immunodiffusionassay up to five weeks, when the animalswere sacrificed.

After the first week, y2-globulin roseprogressively in all groups. Group IIIshowed an increase by five weeks of 40 %,consistent with normal age changes(Horne and Ferguson, 1972). Groups Iand II rose together, increasing by 98%and 124% respectively by five weeks.Groups I and II were never significantlydifferent, but differed from group III attwo weeks (p < 0-05) and at three, four,and five weeks (p < 0 005 - p < 0 001).

Evidence is presented that portacavaltransposition achieves portal-systemicshunting with little change in hepaticfunction. It is proposed therefore thathypergammaglobulinaemia in this situa-tion may be a result of shunting of portalblood past the reticuloendothelial cells ofthe liver, and that hepatocellular damagemay not play a major role in this process.References

Bjorneboa, M., Prytz, H., and Orskov, F. (1979).Lancet, 1, 58

Horne, C. H. W., and Ferguson, J. (1972). J.Endocr., 54, 47

Inchley, C. J., and Howard, J. G. (1969). Clin. exp.Immunol., 5, 189

Leading aticle (1972). Lancet. 1, 80

A4

Viral and bacterial antibodies in primarybiliary cirrhosis and other chronic liverdisease

H. C. THOMAS, R. HOLDEN, J. IRONSIDE, ANDR. G. SOMERVILLE (University Departmentof Bacteriology and Immunology, GlasgowUniversity, and Virology Department,Belvidere Hospital, Glasgow) Increasedantibody titres to measles and rubellaviruses have been described in chronicactive hepatitis and increased antibodytitres to the Gram-negative intestinalbacteria have been demonstrated in

alcoholic liver disease and chronic activehepatitis. It is suggested that the elevationof antibody titres directed against theintestinal bacteria is secondary to a failureof the liver to sequestrate these bacteria,but since elevated titres to measlesand rubella have only been described inchronic active hepatitis, the possibility ofan aetiological association of these virusesto this disease has arisen. We have inves-tigated the incidence of high titres ofantibody to Gram-negative intestinalbacteria and viruses in primary biliarycirrhosis (40 cases) and other forms ofchronic liver disease.The patients with primary biliary

cirrhosis had significantly elevated titresto Gram-negative intestinal bacteria. Thetitres were similar to those found inpatients with other forms of chronic liverdisease. This is further evidence of afunctional reduction in the phagocyticcapacity of the liver in all forms of chronicliver disease.Antibody titres to measles were elevated

in chronic active hepatitis (p < 0.001) andprimary biliary cirrhosis (p < 0 001) butwere normal in alcoholic liver disease.Titres to rubella were elevated in chronicactive hepatitis but not in the primarybiliary cirrhosis or alcoholic liver diseasegroups. This suggests that these viruseshave a special association with the auto-immune liver diseases.

A5

Cell-mediated immunity to human Tamm-Horsfall glycoprotein in autoimmune liverdisease associated with renal tubularacidosis

D. C. TSANTOULAS, I. F. MCFARLANE, B.PORTMANN, A. L. W. F. EDDLESTON, ANDROGER WILLIAMS (The Liver Unit, King'sCollege Hospital, London) Renal tubularacidosis (RTA) occurs in about 30% ofpatients with active chronic hepatitis orprimary biliary cirrhosis, and immunolo-gical abnormalities have been implicatedin this association.

In this study, we have used the leucocytemigration test to investigate cell-mediatedimmunity to a protein known as theTamm-Horsfall glycoprotein, which ispresent in normal human urine and isproduced within the tubules of the kidneyat the principal sites of acidification.Abnormal responses were found in 91 %of patients with active chronic hepatitis orprimary biliary cirrhosis associated withRTA, but in only 19% of those without

RTA. In nearly all of a group of patientswith other liver diseases and in a controlgroup of normal subjects, results werewithin normal limits. In addition, using animmunofluorescent technique with rabbitantibody to human Tamm-Horsfall glyco-protein, it was possible to demonstratethe presence in human liver cell membraneof material reacting immunologically asTamm-Horsfall.

These findings suggest that the develop-ment of an immune response to this glyco-protein, initiated by release of cross-reacting antigens from damaged hepato-cytes, could be the mechanism underlyingthe occurrence of RTA in some patientswith autoimmune liver disease.

B1

The production of lysozyme by the Panethcell

G. R. VANTRAPPEN, AND T. L. PETERS (Uni-versity of Leuven, Leuven, Belgium) Thefunction of the Paneth cell in normal andpathological conditions is still largely un-known. In previous cell fractionation andhistochemical studies we demonstratedthat the Paneth cell granules containlysozyme. Based upon this observation wedeveloped a method to study Paneth cellsecretion, using immunofluorescence andintestinal perfusion techniques. Lysozymewas isolated from the small intestine ofmice and antibodies prepared against it inrabbits. The antiserum was used to staincryostat-cut, alcohol-fixed tissue sliceswith the indirect antibody technique. Infasting mice fluorescence was largely con-fined to the secretory pole of the Panethcells. After the intravenous administra-tion of 4 mg of pilocarpine the fluor-escence in the crypts was abolished, butstreaks of fluorescent material granular inappearance could be seen between thevilli. Ten minutes after the stimulation thefluorescent material had reached the topof the villi. Perfusion studies confirmedthese findings. A 15 cm segment of theileum was perfused with physiologicalsaline at a constant rate of 4 0 ml/hr. Theconcentration of lysozyme during thecontrol period was 0 1-0-2 jig/ml, and wassharply increased after the intravenousadministration of 4 0 mg of pilocarpine.Much smaller doses were also effective: astatistically significant rise was stillobtained with a dose of 0-01 mg (peakrate 0-8 ,ug/ml). Amylase and acid phos-phatase were determined concomitantly;their initial concentration was 5-6 Somog-


http://gut.bmj.com

/G


ownloaded from

http://gut.bmj.com/


yi units per ml for amylase and 2-6mU/ml for acid phosphatase. The con-centration of these two enzymes decreas-ed during the perfusion and was un-affected by the injection of pilocarpine.Our results demonstrate the presence oflysozyme in the Paneth cell granules, andthe ability of lysozyme to serve as amarker of Paneth cell secretion.

B2

Small intestinal mucosal morphology in agroup of infants and young children withdelayed recovery after acute diarrhoea andvomiting

M. J. BOYCE, N. E. FRANCE, AND J. A.WALKER-SMITH (Queen Elizabeth Hospitalfor Children) Acute gastroenteritis (GE)is usually a self-limiting illness with a lowmortality, although the morbidity may beconsiderable (Gribbin et al, 1974). Duo-denal mucosal damage, sometimes severe,is common during the acute illness al-though on the basis of a few serial biop-sies it is suggested that mucosal recoveryis rapid (Barnes and Townley, 1973). Anecropsy study of 10 children who diedat variable intervals after acute GE re-vealed an abnormal mucosa in seven(Walker-Smith, 1972).Twenty-two infants and young children

with acute antecedent diarrhoea who sub-sequently either failed to gain weightsatisfactorily or continued to lose weightdespite adequate caloric intake weresubjected to small intestinal biopsy. Someinfants showed signs of severe malnu-trition. All biopsies were expected toassist diagnosis and management.

Biopsies were performed not less thanthree weeks after the acute antecedent ill-ness and in some cases as late as threemonths. Seven biopsies were consideredhistologically normal. The remainder (15)showed variabled egrees of abnormality,which, on the basis of villous architecture,epithelial cell changes, and infiltration ofthe lamina propria with inflammatorycells, were graded into mild (7), moderate(5), and severe (3). The severe changeswere similar to those seen in coeliacdisease.

Generalized disaccharidase depressionoccurred in association with abnormalmorphology. In two instances disacchari-dase depression also occurred despitenormal morphology. Only two childrendeveloped sugar malabsorption on rein-troduction of cow's milk formula afterrehydration.

The management of this group ofinfants and young children, together withthe hypothesis that acute antecedent GEmay lead to prolonged small intestinaldamage, will be discussed.

B3

Continuing gluten ingestion and its detec-tion in treated coeliac patients

P. G. BAKER AND A. E. READ (UniversityDepartment of Medicine, Bristol RoyalInfirmary) In order to assess the inci-dence and effects of continuing gluten in-gestion in patients with coeliac disease, wehave studied 51 patients after a period offour to 132 (mean 63) months on aprescribed gluten-free diet. Each patientcompleted a dietary questionnaire, under-went a repeat jejunal biopsy and hadserum taken for gluten antibody estima-tion (tanned erythrocyte method).The questionnaire showed that 30 (60 %)

had failed to eliminate dietary glutencompletely. Of these nine were takinglarge amounts, and 21 were taking smallamounts.

All nine of the patients taking largeamounts of gluten showed subtotal villousatrophy on repeat jejunal biopsy com-pared with only four of 20 patients takingno gluten (P <0-01).

Similarly, all nine patients taking largeamounts of gluten had antibodies togluten in the serum compared with onlyfour of the 20 patients apparently takingno gluten (P < 0-01).Nine of 10 patients with both subtotal

villous atrophy and gluten antibodieswere still ingesting large amounts ofgluten.These findings suggest that many

patients with coeliac disease continueto ingest gluten despite initial dietaryinstruction, and that patients with bothsubtotal villous atrophy and circulatinggluten antibodies are almost certain to bestill taking large amounts of gluten.

B4

Family study of coeliac disease

C. J. ROLLES, T. 0. KYAW MYINT, WAI-KEESIN, AND CHARLOTTE M. ANDERSON (Insti-tute of Child Health, University ofBirmingham) The reported incidence ofcoeliac disease amongst the patient'sfirst-degree relatives varies from under1 % to over 15 %. However, in none of thereports has the small intestinal mucosa

been examined in all the first-degreerelatives of all the propositi.We have investigated all first-degree

relatives of 15 newly diagnosed coeliacchildren. Families known to containanother coeliac patient were excluded.

Investigations included detailed clinicalhistory, anthropometry, and jejunal muco-sal biopsy; measurement of haemoglobin,serum iron, red cell and serum folate; theHLA status and, in children under 30 kg,a one-hour blood xylose test.Mucosal appearances typical of coeliac

disease were found in two of the 30parents (6-6%) and two of the 42 siblings(4 8%)-an overall incidence of 5 5% of72 first-degree relatives. Both childrenwere in one family and thus three of 15families (20%) contained another coeliacmember.

In none of the four was the historyfrankly suggestive nor were any of thetests other than biopsy uniformly reliablefor screening.The incidence of 'asymptomatic' af-

fected first-degree relatives was 5 5% withone of five families containing more thanone affected member. Studies where allpropositi are children may give resultsdifferent from those where adult casesare used. No test other than intes-tinal biopsy was shown to select withcertainty the individuals previously un-suspected.

B5

Ileal function in patients with coeliacdisease

D. B. A. SILK, PARVEEN J. KUMAR, M. L.CLARK, AND A. M. DAWSON (Medical Unit,Department of Gastroenterology, StBartholomew's Hospital, London) In ani-mals after jejunal resection the remainingileal mucosa becomes hypertrophied andileal function enhanced (Dowling andBooth, 1967). Untreated coeliac diseasemight be compared to this experimentalmodel because the mucosal lesion is moremarked in the jejunum than the ileum, andsome patients absorb vitamin B12 to agreater extent than normal subjects (Eliaset al, 1973).

In the present study the possibility ofenhancement of other aspects of ilealfunction in patients with untreated coeliacdisease has been investigated by studyingabsorption during ileal perfusion of anisoosmotic electrolyte solution containing50 mM glucose and 25 mM bicarbonate.The differential handling of bicarbonate


http://gut.bmj.com

/G


ownloaded from

http://gut.bmj.com/


by the jejunum and ileum was used toconfirm the correct positioning of theperfusion tube in normal controls, andapplying the same criteria to the coeliacpatients ileal positioning was satisfactoryin five out of seven subjects studied. Threeof these patients had enhanced ileal func-tion as shown by increased absorption ofeither glucose, sodium, chloride or watercompared with any of the respectiveabsorption values observed in eight con-trol subjects.These findings indicate that enhanced

ileal function, in addition to the recentlydescribed colonic adaptation (Phillipsand Giller, 1973), may explain why somecoeliac patients with known jejunalsecretory states do not have diarrhoea.

References

'Dowling, R. H., and Booth, C. C. (1967). Structu-ral and functional changes following smallintestinal resection in the rat. Clin. Sci., 32,139-149.

'Elias, E., MacKinnon, A. M., Short, M. D., andDowling, R. H. (1973). Factors controllingileal adaptation after proximal small bowelresection and in coeliac disease. Eur. J. clin.Invest., 3, 226.

'Phillips, S. F., and Giller, J. (1973). The contribu-tion of the colon to electrolyte and waterconservation in man. J. lab. clin. Med., 81,733-746.

C1

Actions of bran on colonic motility relatedto its physical properties

W. 0. KIRWAN, A. N. SMITH, A. A.MCCONNELL, W. D. MITCHELL, AND M. A.EASTWOOD (Department ofClinical Surgeryand Wolfson Gastrointestinal Laboratories,University of Edinburgh, Western GeneralHospital, Edinburgh) The effectivenessof two different bran preparations inlowering intraluminal colonic pressureand decreasing intestinal transit time hasbeen assessed in 14 subjects given fine andcoarse bran over one month. Coarse brangiven to nine patients lowered intraluminalpressure motility indices significantly:basal 892-8 + 196 to 648-7 + 339; afterfood 1513-2 ± 456 to 446-1 ± 130 (P <0-01); after prostigmine 2120 + 405 to1216'8 + 398 (P < 0 01), and transit time93.4 hours + 13-8 to 57*9 hours ± 8-0(P < 0-01): fine bran had no such effects.Patients first given fine bran and laterchanged to coarse bran also showed signi-ficant change. The water-holding capacityof the two brans differed markedly:6'15 g and 2 36 g of water per g bran forthe coarse and fine preparations respect-ively. The lignin content of each was 4-1 %

and 2-6% and the cationic exchangecapacities 1-2 m-equiv and 0-7 m-equivper gram. After milling both brans to thesame particle size, the advantageouscharacteristics of the coarse bran werelost.

It is concluded that the water-holdingcapacity of coarse bran is a function ofparticle size, and that these two factorsare related to motility changes.

C2

Effects of bile acids on the motility of thecolon

W. 0. KIRWAN, A. N. SMITH, W. D.MITCHELL, AND M. A. EASTWOOD (WolfsonGastrointestinal Laboratories, WesternGeneral Hospital and Department ofClinical Surgery, University of Edinburgh)Bile was reported by Horral (1938) andHarvey et al (1939) to stimulate intestinalmotor activity, but specific effects on themotor function of the colon have not yetbeen described. The effects of bile acidson the colon have therefore been examinedin acute and chronic experiments in therabbit and in the human.

In acute experiments in 14 rabbits bileacids were infused into the caecum in con-centrations 2 mM-16 mM. Concentrationin the range 2 mM-4 mM produced thegreatest motor activity (p < 0-0125)starting at 10 minutes. In 12 chronic ex-periments bile in the caecum producedmaximal motor stimulation (p < 0-0125)after a delay of 30 to 45 minutes. Whenbile was introduced into the sigmoid, theeffect took place in 10 minutes. Trhere wasa significant correlation between the moti-lity response and the concentration used(r = 0 98). Bile introduced into the humansigmoid had a similar effect.

In patients with cholerrhoeic entero-pathy the motility responses correlatedwell with daily bile acid excretions (r =0-58). It is suggested that bile can act asstimulant of colonic motility.References

Haney, H. F., Roley, W. C., and Cole, P. A. (1939).The effect of bile on the propulsive motilityof Thiry-Vella loops in dogs. Amer. J.Physiol., 126, 82-88.

Horral, 0. H. (1938). Bile: its toxicity and relationto disease. University of Chicago Press.

C3

CEA estimation in ulcerative colitis withand without malignant change

J. B. DILAWARI, J. E. LENNARD-JONES,

A. M. MACKAY, JEAN K. RITCHIE, AND H. G.STURZAKER (St Mark's and the CentralMiddlesex Hospitals, and the ChesterBeatty Research Institute, London) Theaim of this study was to investigate levelsof plasma CEA in patients with ulcerativecolitis, particularly in thosewith carcinomaor especially liable to malignant change.CEA was measured by a double

antibody radioimmunoassay method(Laurence et al, 1972), the upper limit ofnormal being taken as 12-5 ng/ml based onmeasurements in 60 healthy control sub-jects.CEA levels above 12-5 ng/ml were

found in 11 of 25 patients with non-inflammatory bowel disease or minoranal conditions, originally intended to actas a hospital control group.One hundred and fifty-two patients with

ulcerative colitis attending St Mark's orthe Central Middlesex Hospitals betweenNovember 1972 and May 1974 werestudied, including seven with severedysplastic changes in the rectal mucosaand six with established carcinoma.CEA levels above 12-5 ng/ml were found

in 53 of the 152 patients with colitis; norelationship between actual levels andactivity, duration, or extent of the diseasecould be found. Eight of the 13 patientswith either dysplastic or carcinomatouschanges showed levels above normal ( fourin each group). Only two of the 152patients (one in the dysplastic and one inthe carcinoma group) had levels above40 ng/ml, a level usually associated withcarcinoma.

This study highlights the problem of theselection of controls for CEA estimationand suggests that CEA levels do not con-tribute to the management of patientswith colitis.

Reference

Laurence, D. J. R., Stevens, U., Bettelheim, R.,Darcy, D., Leese, C., Tuberville, C.,Alexander, P., Johns, E. W., and MunroNeville, A. (1972). Role of plasma carcino-embryonic antigen in diagnosis of gastro-intestinal, mammary, and bronchial carci-noma. Brit. med. J., 2, 605-609.

C4

Immunosuppressive consequences of thetreatment of ulcerative colitis with azathio-prine

J. M. SKINNER, A. C. CAMPBELL, CATHERINEWALLER, JANET WOOD, AND I. C. N.MaCLENNAN (Nuffield Department of Clini-cal Medicine and Nuffield Department of


http://gut.bmj.com

/G


ownloaded from

http://gut.bmj.com/

Thne British Society of Gastroenterology 829

Pathology, Radcliffe Infirmary, Oxford,introduced by Dr S. C. Truelove) Stud-ies were made of lymphoid subpopula-tions in the blood and rectal mucosa of 40patients who were taking part in a double-blind trial of continuous treatment withazathioprine, the clinical results of whichare reported by D. P. Jewell and S. C.Truelove. In addition to blood leucocyteand differential counts, the followingassays of lymphocyte function or sub-populations were applied: mitotic re-sponse to phytohaemagglutinin (T cellfunction), staining for surface immuno-globulin (B cells), and cytotoxicity againstantibody-sensitized target cells (K cellactivity). In some cases T cells were alsoidentified by their capacity to form rosetteswith sheep red blood cells. Plasma cells ofspecific class were identified in the laminapropria of rectal biopsies by an immumo-peroxidase technique and were quanti-tated. From the blood leucocyte countsthe only change attributable to azathio-prine was a slight and gradual fall inlymphocyte count. The lymphocyte assaysshowed that azathioprine had little effecton the major T and B cell subpopula-tions. Patients on azathioprine did, how-ever, show a highly significant depressionof K cell activity. After 12 months' treat-ment, this lymphocyte cytoxicity wasalmost immeasurable in several patients.The assays of plasma cells in the rectal

lamina propria revealed that these cellswere markedly depleted by azathioprinetreatment.

C5

Treatment of chronic proctitis with di-sodium cromoglycate

R. V. HEATLEY, B. J. CALCRAFT, J. RHODES,E. OWEN, AND B. EVANS (Department ofGastroenterology, University Hospital ofWales, Cardiff) Mast cells and eosino-phils are commonly found in the rectalmucosa of patients with ulcerative colitisand proctitis. The mast cells are oftendegranulated during disease activity andmay play a role in the pathogenesis. Sincedisodium cromoglycate (DSCG) stabilizesmast cells in bronchial asthma, it may beof therapeutic value in proctitis.We have examined DSCG in 24 patients

with proctitis. All had active symptomsand inflammatory changes on sigmoido-scopic examination; the barium enemawas normal in all patients. DSCG wasgiven by enema (200 mg twice daily, andorally 100 mg three times daily). The

trial was of a double-blind crossoverdesign in which patients received active orplacebo treatment in a randomized order,each for a period of four weeks. Patientsrecorded symptoms on a diary card, and asigmoidoscopy with rectal biopsies wasperformed initially and at the end of eachtreatment period.On the basis ofchange in symptoms and

and sigmoidoscopic appearance, 11 pati-ents improved during the DSCG period;this improvement was marked in nine.Five patients improved during the placeboperiod and the improvement was markedin two. Four patients showed no changeduring either period and a further fourwere withdrawn because they failed tocomplete the trial satisfactorily. Thenumber of eosinophils in biopsy speci-mens varied from 0 to 115 per five high-power field. Five patients with less than 10eosinophils/5 hpf did not respond to intal;eight of 10 with more than 10 eosinophils/5 hpf responded to the drug. The resultssuggest that DSCG may be of value in thetreatment of proctitis. The findings mayalso be important in relation to thepathogenesis of this disease.

Dl

Prolyl hydroxylase activity in regeneratingrat liver

1. S. BENJAMIN, J. O'D. MCGEE, THAN THAN,MARION C. RODGER, AND L. H. BLUMGART(University Departments of Surgery andPathology, Glasgow Royal Infirmary)Most investigations in the field of liverregeneration have focused on the kineticsand mechanisms underlying the cellularproliferative response. This paper is con-cerned with the collagen productive re-sponse in regenerating liver, since collagenis the main extracellular product whichprovides structural support for a pro-liferating cell population in vivo. Activityof prolyl hydroxylase, a key enzyme incollagen biosynthesis, was used as anindex of collagen production. The tissuelevel of this enzyme in general parallelsthe rate of collagen synthesis in tissues(McGee et al, 1974) and preliminary datausing labelled precursors indicate that therelationship is valid in the present model.

Sixty-eight per cent hepatectomy(Higgins and Anderson, 1931) or shamoperation was performed in male rats. Atintervals after surgery, hepatic concentra-tion of prolyl hydroxylase in the regen-erating posterior lobes was measured.Activity of the enzyme increased to a

maximum of two to three times normal at36 hours. Elevated levels were main-tained for at least four days, but returnedalmost to normal at seven days, when theliver approaches its estimated preopera-tive weight.The present data suggest an appropriate

functional relationship between collagensynthesis and hepatocyte and mesenchy-mal cell proliferation in regeneratingliver. Further studies of this relationshipmay be of value in studying the mecha-nisms not only of liver regeneration butalso of liver pathology.

References

Higgins, G. M., and Anderson, R. M. (1931). Arch.Path., 12, 186.

McGee, J. O'D., Patrick, R. S., Rodger, Marion C.,and Luty, Caroline L. (1974). Gut, in press.

D2

Adenocarcinoma of the bile ducts: Rela-tionship of anatomical location to clinicalfeatures

DAVID A. INGIS AND RICHARD G. FARMER(Department of Gastroenterology, TheCleveland Clinic Foundation, Cleveland,Ohio, USA) An unusual, but not rare,cause of obstructive jaundice is carcinomaof the bile ducts. It was postulated thatlesions at the hepatic duct bifurcationrepresent a clinical entity separate fromthose arising elsewhere in the bile ducts.This review of 23 cases ofcarcinoma of thebile ducts (1960-1974) compared carci-noma of the hepatic ducts and commonbile duct regarding clinical presentation,treatment, and prognosis. There were 14patients with hepatic duct lesions and ninewith common bile duct carcinoma. Age,sex, presenting symptoms, physical ex-amination, and laboratory findings weresimilar regardless of location of carci-noma. Location did not influence whetherthe correct diagnosis was made at theinitial operation, only 25 %. Percutaneouscholangiograpby, operative cholangio-graphy, and endoscopic retrograde cho-langiopancreatography were equally suc-cessful in making the diagnosis in bothlocations. Pathologically, the hepaticduct carcinomas were scirrhous and distallesions were adenocarcinomas. Locationinfluenced operative feasibility; resectionwas possible with common bile ductlesions. Survival was less than one yearfor both locations, and unaffected by 5-fluoracil or radiation. Since prognosisis so poor, attention must focus on means


http://gut.bmj.com

/G


ownloaded from

http://gut.bmj.com/


of earlier diagnosis, but separation intodistinct clinical entities based on locationof the lesion is artificial.

D3

Cholelithiasis: A metabolic consequence ofileostomy

WILLIAM S. J. MAIR, GRAHAM L. HILL, ANDJOHN C. GOLIGHER (Department ofSurgery,The General Infirmary, Leeds) All ileost-omists attending a large clinic weresubmitted to cholecystography to de-termine the incidence of gallstones andto determine the factors governing this.Of 67 patients studied to date, four

have had cholecystectomy performed anda further 14 have developed stones sinceileostomy. The overall incidence of stonesis thus 28% and for the 28 patients over50 years of age, 13 (46%) have gallstones.

Detailed operative and pathologicaldata are available for each patient, and,of the many risk factors studied, only age,sex, and ileal resection were significant.

Eleven out of 18 over 53 years of agedeveloped stones compared with 12 outof 49 under 53 (p = < 0-0025). Thirteenout of 18 women developed stones com-pared with 25 out of 49 men (P = <0005).The mean length of ileal resection in

those who developed stones was 18-3 in.compared with 6-3 in. in those who did not(p = < 001).Of 23 patients with a resection < 4 in.

nine had stones. None of those withoutstones were over 53 years of age.

It is concluded that there is an in-creased incidence of gallstones in ileos-tomists, particularly in women, thosewith an ileal resection, and those who areover 53 years of age.

D4

Fulninant hepatic failure: Histologicalquantitation of surviving liver parenchymain the evaluation of the cause of death

B. PORTMANN, B. G. GAZZARD, I. M. MURRAY-LYON, AND ROGER WILLIAMS (Liver Unit,King's College Hospital, London) Since1968, 130 adult patients with grade III orIV encephalopathy due to fulminanthepatic failure (FHF) have been seen.Only 12 of the first 99 patients survived,although a more encouraging figure (45%survival) has been obtained in a recentseries of 31 patients treated with charcoalcolumn haemoperfusion.

Analysis of the necropsy findings in 95patients showed cerebral oedema to be astriking feature in 36, and in 20 it wasassociated with brain herniation. Majorgastrointestinal bleeding was responsiblefor death in 26 patients, 10 died of variousinfections, and in five there was acutepancreatitis. In the remaining 21 cases,death appeared to be due to severe hepaticfailure without these complicating factors.When the area of surviving liver paren-chyma was assessed by a morphometrictechnique on necropsy liver biopsies, thisgave a value of less than 12% (normalabout 85%) in all the patients in whomhepatic failure was the cause of death,whereas in those cases with sepsis as amajor cause of death and in half the caseswith cerebral oedema or bleeding, thehepatocyte volume fraction was between25% and 55%.

Thus, better understanding and manage-ment of these complications is necessaryif the survival rate is to be improved evenfurther.

D5

Effect of administration of a colonicmetabolite of cholic acid on cholesterollevels in bile and blood

E. W. POMARE AND T. S. LOW-BEER (Univer-sity Department of Medicine, BristolRoyal Infirmary, Bristol) Cholic acid ismetabolized by colonic bacteria to deoxy-cholic acid (DC)'. This metaboliteadministered at doses up to 150 mg daily,selectively suppresses hepatic synthesisand pool size of chenodeoxycholate(CDC), without altering total bile saltpool size2. CDC, when used for dissolvinggallstones, decreases the cholesterol con-tent of biliary lipids by decreasing choles-terol secretion into bile3. We anticipatedthat DC administration would increasethe cholesterol content of bile lipids.Sixteen healthy volunteers received 100-150 mg DC dailv for two weeks. Thecholesterol content of biliary lipids in-creased significantly (p < 0-0125) andreversibly (p < 0-025), and this wasaccompanied by a significant fall in serumcholesterol to 88 % of baseline levels.

These results suggest that populationsin whom there is a high colonic absorp-tion of bacterially metabolized cholate,ie, DC, have an increased predispositionto cholesterol gallstones.References

"Morris, J. S., Low-Beer, T. S., and Heaton, K. W.

(1973). Bile salt metabolism and the colon.Scand. J. Gastroent., 8, 425-431.

'Pomare, E. W., and Low-Beer, T. S. (1974). Theselective inhibition of chenodeoxycholatesynthesis by cholate metabolites in man.Gut, in press.

3Northfield, T. C., and Hofmann, A. F. (1973).Biliary lipid secretion in gallstone patients.Lancet, 1, 747-748.

D6

The effect of florantyrone (Zanchol) on thecomposition of T-tube bile

J. R. KIRKPATRICK, M. A. EASTWOOD, ANDW. D. MITCHELL (The Department ofClinical Surgery and the Wolfson Gastro-intestinal Laboratories, Western GeneralHospital, University of Edinburgh, Edin-burgh) In 1956 McGowan describedchanges in common duct bile resultingfrom the use of a new choleretic floran-tyrone. The results in the physical proper-ties of the bile suggested that this drugshould be useful in the treatment ofbiliary disease. Recent emphasis on thevalue of altering bile composition promp-ted a new study of this drug.Four patients were studied following

cholecystectomy and choledochotomy.T-tube bile was collected for three daysbefore and three days during therapy with250 mg of florantyrone (Searle) adminis-tered thrice daily orally. Within 24 hoursthere were changes in the bile, both incolour and chemical composition. Theprincipal change was in the bile acid con-centration from 4-8 ± 1-8 mM to 8-9 +3-2 mM, particularly in the cholic acidfraction (3-2 ± 1-3 to 5 9 + 1 9 mM) andin phospholipids (570 ± 77T6 mg% to142 i 134-0 mg%).There were also significant but less

striking increases in the concentration ofcholesterol and in dry weight and osmolal-ity of the bile. The changes in the deter-gent content of the bile might well beexploited in studies and management ofbiliary problems.

Reference

McGowan, J. M. (1956). Surg. Gynec. Obst.. 103,163.

El

Ion transport characteristics of humanjejunal and ileal mucosa in vitro

P. ISAACS AND L. A. TURNBERG (Division ofGastroenterology, Manchester Royal Infir-mary, and Department of Medicine, HopeHospital, Salford) In-vivo studies of ion


http://gut.bmj.com

/G


ownloaded from

http://gut.bmj.com/


transport demonstrate differences betweenhuman jejunum and ileum. To investigatethese differences at the mucosal level wehave studied ion transport in human in-testinal mucosa in vitro. Pieces of strippedmucosa from fresh surgical specimenswere mounted in Ussing-type flux cham-bers and bathed in oxygenated glucose-Krebs-Ringer solution at 37°. There wereclear differences between the behaviourof short circuited jejunal (n = 8) andileal (n = 9) mucosa. Net Na- absorption(netNa+) was greater in ileal than jejunalmucosa (12-8±1.2 against 4-0±1.1 ,uequiv/cm2/hr) due to a greater mucosa to serosaunidirectional flux (Jm); serosa to mucosaflux (Jm) was similar in the two tissues.JnetCl- was near zero in jejunum butileal mucosa absorbed 5 4 ,uequiv/cm2/hrof Cl-. Electrical characteristics are shownin the table.

PD (m V) Resistance SC Current(Q/cm') (Aequiv/cm2/

hr)

Jejunum 9 0 t 05 17-0 2-3 6-3 1 0Ileum 9-1 03 11-6 1-6 93 09

Theophylline (10 mM) had no effect onnet jejunal ion transport while in the ileumit increased, Jsmcl-producing a net secre-tion of 2 0 uequiv/cm2/hr and JnetNa+was slightly reduced to 9-6 /Aequiv/cm2/hr.These results demonstrate that ilealmucosa absorbs more Na+ per cm2 thanjejunal mucosa and that while the ileumabsorbs Cl- actively the jejunum does not.Differences in response to theophyllinealso suggest cyclic AMP stimulatedo intransport processes are probably differentin these two regions.

E2

The influence of ethacrynic acid and aspirinon water and electrolyte secretion inducedby prostaglandin E1 in the human jejunum

C. MATUCHANSKY, J. Y. MARY, J. C.RAMBAUD, AND J. J. BERNIER (Groupe deRecherches sur la Physiopathologie de laDigestion, INSERM, Hopital St-Lazare,Paris, France) Intraluminal prostaglan-din E1 (PGE1) promotes a copious jejunalsecretion ofwater and electrolytes in man.1To investigate further this secretory effect,the action of ethacrynic acid (EA) oraspirin upon basal jejunal absorption andPGE1-induced secretion of water andelectrolytes was studied in 25 healthyvolunteers using an intestinal perfusiontechnique (proximal occlusive balloonand glucose-saline isoosmotic solution).

EA (2 5 mg/kg), administered in jejunallumen prior to PGE1 (0 9 ,sg/kg/min),significantly reduced (p < 0 001) the netsecretory effect of intraluminal PGE1 byan average of 4-8 ml/min/25 cm for waterand 632 pequiv/min/25 cm for sodium;EA suppressed the PGE1-induced increasein plasma to lumen unidirectional flux ofsodium. Intravenous or intraluminalaspirin (25 to 45 mg/kg) did not modifyPGE,-induced secretion nor control ab-sorption rates. No significant changes insystemic plasma level of PGE occurredduring the different periods of study.

These results suggest that aspirin, whichinhibits prostaglandin synthesis in vitro,and reduces cholera secretion in animals2,has no effect on the intestinal fluid move-ments induced by preformed PGE1.Conversely, they indicate that EA inhibitsin man PGE.-elicited jejunal secretion, aneffect of EA similar to that observed, inanimals, on cholera and cyclic AMP-mediated secretion3.

References'Matuchansky, C., and Bernier, J. J. (1973).

Gastroenterology, 64, 1111-1118.'Finck, A. D., and Katz, R. L. (1972). Nature

(Lond.), 238, 273-274.3AI-Awqati, Q., Field, M., and Greenough, W. B.

(1974). J. clin. Invest., 53, 687-692.

E3

The effect of frusemide on small-intestinalabsorption of water and electrolytes

J. F. MaCKENZIE, K. M. COCHRAN, AND R. I.RUSSELL (Department of Gastroenterology,Royal Infirmary, Glasgow) Although diu-retics are widely used, little is known oftheir effects on the movements of waterand electrolytes in the small intestine.We have studied the effect of frusemide

on small intestinal absorption of waterand electrolytes using a jejunal perfusionsystem in man. An isomotic solution con-taining sodium, potassium, chloride, andbicarbonate in glucose and using phenol-sulphonphthalein as a water-solublemarker, was perfused using a double-lumen tube with proximal occludingballoon". After a steady state had beenachieved and baseline collections ob-tained, frusemide was given intravenouslyin doses of either 20 mg or 40 mg and theeffect on intestinal absorption determined.After 20 mg of frusemide the mean waterabsorption (± SEM) in five subjects fellfrom a baseline of 3-15 ± 1 01 ml/minuteto 1-39 ± 1-97 ml/minute. Similar fallswere obtained with electrolyte absorption.After 40mgof frusemidewasadministered

intravenously, the mean water absorptionin five subjects (± SEM) fell from 1 99 ±0-87 ml/minute to -0O81 ± 1-49 ml/minute. Similar results were obtained withelectrolyte absorption. In three subjects,net secretion of water and electrolytesoccurred after frusemide.These findings may account for the

diarrhoea which has been reported insome patients when high doses of fruse-mide are used, especially in patients withchronic renal failure. In addition, thisinhibition of absorption in the small intes-tine may affect the absorption of otherdrugs which may be given together withfrusemide. Modification of the dosageregimes of such drugs may be necessarywhen frusemide is administered in largedoses.

Reference1Phillips, S. F., Summerskill, W. H. J. (1966).

Occlusion of the jejunum for intestinal per-fusion in man. Proc. Mayo Clin., 41, 224-231.

E4

Effects of ethanol on small intestinalabsorption

N. KRASNER, K. M. COCHRAN, C. G.THOMPSON, H. A. CARMICHAEL, AND R. I.RUSSELL (Gastroenterology Unit, StobhillGeneral Hospital, Glasgow, and Depart-ment ofGastroenterology, Royal Infirmary,Glasgow) It is well established thatchronic alcoholic subjects may exhibitmalabsorption1. In the present study, theabsorption of water, electrolytes, andglucose by the jejunum was measured in10 chronic alcoholic patients using atriple-lumen tube perfusion system2. Amalabsorption screen was also performedon each patient.

Jejunal histology was normal in everypatient and only isolated biochemicalabnormalities were found in individualpatients. However, the mean absorptionof water in the alcoholic subjects (50 0 ±21-3 ml/hr) was significantly lower (p <0 001) than the mean value in 14 healthycontrol subjects (205.0 ± 15-9 ml/hr).Alcoholic patients also showed a signifi-cant reduction (p < 0 001) in the absorp-tion of sodium (7-5 ± 2-9 m-equiv/hr)when compared to control subjects (23.1± 1-3 m-equiv/hr); chloride absorptionshowed a similar trend.

Since the availability of adenosine tri-phosphate (ATP) plays a significant rolein the active transport of Na+ and K+across the intestinal cell membrane, anin-vitro model was set up to study ATP


http://gut.bmj.com

/G


ownloaded from

http://gut.bmj.com/


following the exposure of segments ofguinea pig jejunum to 2% ethanol forperiods of one hour. ATP levels in seg-ments perfused with Krebs' solution weresignificantly higher (p < 0-02) than levelsin ethanol-perfused portions.

These findings suggest that ethanol mayinduce changes in jejunal ATP andATPase content which may explain theinhibition of intestinal absorption of waterand electrolytes found in the chronicalcoholic patients.ReferencesIlber, F. L. (1971). Alcohol and the gastrointestinal

tract. Gastroenterology, 61, 120-123.'Cooper, H., Levitan, R., Fordtran, J. S., and

Ingelfinger, M. D. (1966). A method forstudying absorption of water and solutefrom the human small intestine. Gastroenter-ology, 50, 1-7.

E5Intestinal absorption of tetrahydrobiop-terin and pteroic acid in man: A comparisonof models for the transport of folates acrossthe small intestine

J. A. BLAIR, K. RATANASTHIEN, AND R. J.LEEMING (Department of Chemistry, Uni-versity ofAston in Birmingham and GeneralHospital, Birmingham) (introduced byW. T. Cooke) Biopterin derivatives arepresent in normal human serum (1-3 ng/ml) and urine (1-2 ,tg/ml). When largeoral doses of tetrahydrobiopterin aregiven to man only slight rises in the serumand urinary levels are seen in contrast tothe large rises with other tetrahydrop-terins, eg, 5-methyltetrahydrofolic acid.Similarly large oral doses of pteroic acidgave no increase in serum and urinarylevels, again in contrast to the substantialrises obtained with other pterins, eg, folicacid. Thus both compounds are at bestpoorly absorbed through the intestine.

It has been previously suggested' thathuman small intestine contains a permeasefor 5-methyltetrahydrofolic acid which isaffected by the stereochemistry at C6 ofthe tetrahydropterin ring. These observa-tions with tetrahydrobiopterin and pteroicacid suggest that there are no specificcarrier systems in the human intestine foreither the tetrahydropterin or pterin ring.

Another model2,3 for the intestinaltransport of folates suggests that they aretransported as the neutral species throughthe lipoid membrane after conversion tothese forms in an acid microclimate atabout pH 3.5. In such a microclimate theconcentration of the neutral forms oftetrahydrobiopterin and pteroic acidwould be almost zero and therefore theyeither would not be transported or only

slowly transported as is observed experi-mentally.References'Weir, D. G., Brown, J. D., Freedman, D. S., and

Scott, J. M. (1973). Clinical Science andMolecular Medicine, 43, 625-631.

'Blair, J. A., Johnson, 1. T., and Matty, A. J. (1974).J. Physiol., 236, 653-661.

'Blair, J. A., and Matty, A. J. (1974). Clinics inGastroenterology, 3, 183-197.

E5The electrical measurement of glucoseabsorption

N. W. READ, R. J. LEVIN, AND C. D.HOLDSWORTH (Departments of Medicineand Physiology, University of Sheffield)The active absorption of glucose from thesmall intestine is linked to sodium irontransfer and is therefore electrogenic. Inthe rat the electrical potential difference(PD) across the jejunal wall increases withincreasing concentrations of luminal glu-cose. Plotting PD against glucose con-centration results in a saturation curvefrom which the kinetic parameter of'apparent Km' can be obtained. In therat this is identical with that obtained bychemical methods of measuring activeabsorption.

In man the potential difference betweenan intraluminal and a subcutaneouselectrode was recorded. Infusion of solu-tions of the same osmolarity and ioniccomposition but containing serially in-creasing concentrations of glucose causedthe jejunal lumen to become progressivelymore negative and a saturation curve wasfound. The apparent Km in 99 normalvolunteers (61-1 mM ± 3-6 SE) wassignificantly higher (p < 0 001) than thatfound in three patients with malabsorp-tion (23-7 ± 8-0 SE).

Kinetic parameters estimated fromchemical absorption curves are the mostsensitive method of assessing changes injejunal function, but are technically diffi-cult in man. This new technique of deter-mining electrically the kinetic parametersof active glucose absorption uses a singlelumen tube, is rapid and more clinicallyacceptable. Because of these features itshould find wide application in studyingjejunal absorption function in health anddisease.

FlImpairment of upper alimentary vagalfunction in diabetes

I. M. STEWART, D. J. HOSKING, B. J. PRESTON,AND M. ATKINSON (Nottingham General

Hospital) Proof that autonomic neuro-pathy causes gastrointestinal symptomsin diabetes is lacking and the site of anysuch changes is conjectural. A cine radio-graphic and manometric assessment ofoesophageal motor function and thegastric secretory response to insulin havebeen used to study autonomic function ina group of 20 diabetics with diarrhoea,vomiting, impotence, and bladder symp-toms.

Oesophago-gastric sphincteric tone waslow and oesophageal peristalsis althoughcoordinated was of low amplitude result-ing in delayed oesophageal emptying inthe 150 Trendelenburg position.

Bethanecol increased sphincteric toneto normal but not to the high levels itdoes in Chagas disease or achalasia,suggesting that the ganglion cells of themyenteric plexus were intact in diabetics.The Hollander test revealed evidence ofcomplete vagotomy in six of eight diabe-tics with diarrhoea but was usuallynormal in those without.

These findings indicate a lesion in thepreganglionic vagal fibres rather than inthe myenteric plexus which may be animportant factor in the causation ofdiabetic diarrhoea and could account forthe reportedly low incidence of duodenalulcer in diabetics.

F2

Can symptoms or radiographic reflux bepredicted by measuring cardiac sphincterpressure?

J. B. DILAWARI, D. A. W. EDWARDS, ANDD. H. GIRMES (MRC GastroenterologyUnit, Central Middlesex Hospital, andDepartment of Statistics and ComputerScience, University College, London)There is a widespread supposition thatlow cardiac sphincter pressure is asso-ciated with reflux and that measurementof the pressure indicates whether thepatient is suffering from reflux. Two hun-dred and forty-six subjects with dys-peptic symptoms were grouped into138 with and 108 without symptomsattributed to reflux. The pressures re-corded by four-channel perfusion mano-metry of the cardiac sphincter failed toseparate the groups. Odds in favour of thepresence of symptoms were calculated;odds better than 19:1 (5% level of signifi-cance) existed if the pressure was 5 mmHg or less, but only 5% of all subjects and9% of those with symptoms had thispressure. Two hundred and twenty-oneof the subjects had a standardized


http://gut.bmj.com

/G


ownloaded from

http://gut.bmj.com/


radiographic examination for the cap-acity to herniate the sphincter and forradiographic reflux. Reflux of bariumwas seen in 58. The distribution ofsphincter pressures was the same in thegroups with and without radiographicreflux indicating that low sphincter pres-sure is not the sole cause of radiographicreflux. Simultaneous herniation of thesphincter occurred in those with radio-graphic reflux and this seems the likelyexplanation for the failure of the anti-reflux mechanism. We conclude that arandom measurement of cardiac sphincterpressure is unhelpful in deciding whethera patient has reflux or is suffering from it.The use of probability calculations indecision making of this sort will bediscussed.

F3

A double blind trial of carbenoxolone andgeranyl farnesylacetate in gastric ulcer

P. M. SMITH, G. E. SLADEN, M. J. S. LANGMAN.AND J. E. LENNARD-JONES (University Hos-pital of Wales, St Bartholomew's Hospital,City Hospital, Nottingham, and UniversityCollege Hospital, London) Carbenoxo-lone, a liquorice derivative, has been usedwidely in the treatment of peptic ulcersince Doll's controlled studies (1964)showed that it would significantly in-crease the rate of healing of gastric ulcerin the ambulant patient. Unfortunately,side effects are common, and we thereforedecided to compare carbenoxolone withgeranyl farnesylacetate, a synthetic iso-.prenoid extracted from white headedcabbage previously shown to aid ulcerhealing (Newcombe, Stone, andRichardson, 1970).

Fifty-seven ambulant patients were ran-domly allocated to the two groups, onereceiving carbenoxolone 100 mg tds andthe other geranyl farnesylacetate 100 mgtds. At entry into the trial the groups weresimilar with respect to age and sex, eachpatient having had a barium meal notmore than five days beforehand. Thesurface area of the ulcer profile wasdeterminedindependentlybyoneof us, andagain four weeks later at the end of treat-ment. Weekly recordings were made ofblood pressure, weight, and ulcer symp-toms. Plasma electrolytes were measuredat the beginning and end of the trial.

Fifty-two patients completed thetrial. Of the 25 who received carbenoxo-lone, mean reduction in ulcer profile area

was 58 %. In contrast the mean reductionin ulcer profile area in the 27 geranylfarnesylacetate-treated patients was 47 %.This difference is not statistically signifi-cant (p > 01). Side effects occurred inhalf of the carbenoxolone-treated patients,whereas none of the patients receivinggeranyl farnesylacetate developed oedemaor electrolyte disturbance.

References

Doll, R. (1964). Scot. med. J., 9, 183.Newcombe, P. B., Stone, W. D., and Richardson,

D. C. (1970). Adv. Abstracts, 4th WorldCongress of Gastroenterology, Copenhagen,313.

F4

Effective treatment of gastric ulcerationwith a Bismuth preparation (De-Nol)

B. E. BOYES, I. L. WOOLF, R. Y. WILSON, D. J.COWLEY, AND I. W. DYMOCK (The Depart-ments of Medicine and Surgery Univer-sity Hospital of South Manchester) Ofdrugs known to affect the rate of healingof gastric ulcers in ambulant patients theliquorice derivatives are most widely usedbut may produce adverse side effects.Recently a bismuth preparation, tri-potassium di-citrate bismuthate (De-Nol)has been advocated for use in patientswith gastric ulceration. In order to assessthe efficacy of this preparation we haveconducted a randomized double-blindtrial of De-Nol in patients with radiologi-cally and endoscopically proven benigngastric ulcers.Of the 20 patients admitted to the trial

10 received De-Nol and 10 an identicalplacebo. These two groups were com-parable with reference to age, sex, andulcer type at the start of the trial. Sevenof the 10 patients receiving the activepreparation had complete radiologicalhealing of the ulcer but only two of thosereceiving the inert preparation. Endo-scopically nine of the 10 treated by De-Nol healed compared with three in theplacebo group. The mean percentagereduction in the cross-sectional area of theulcer asjudged by comparable radiographswas 91 % in the treated group and 36% inthe control group. None of the patientsdeveloped side effects during therapy, andthere was no evidence of drug toxicity.

This degree of healing with De-Nol iscomparable to or better than that achievedby the most extensively used ulcer-healingagent, carbenoxolone sodium.

F5

The patient's assessment of the result ofoperation for peptic ulcer

ELIZABETH L. CAY, A. E. PHILIP, W. P.SMALL, M. A. HENDERSON, AND J. NEILSON(Gastric Follow Up Clinic, Western GeneralHospital, Department of PsychologicalMedicine, University of Edinburgh, Edin-burgh, and Duimfries and Galloway RoyalInfirmary, Dumfries) When we seek toevaluate a new operation for ulcer, it iscustomary to set up a controlled trial.Yet for 15 years or more, controlledtrials have failed to distinguish betweenthe various vagotomies, the variousdrainage procedures, and the various re-sections and anastomoses.Are there no differences, or have we

become obsessed with the method andneglected the indices on which the judge-ment of success ultimately depends?We believe that doctor-determined criteriaare outdated and too insensitive and that anew look at what is expected of an opera-tion is required. The opinion of the patientwho judges the outcome ofsurgery in termsof the quality of his life has been largelyneglected because it has been thought toodifficult to handle 'soft' data of this kind.But other professions confronted with asimilar problem have been forced to lookat attitudes, and by the use of question-naires have successfully hardened 'soft'data.From a list of over 100 statements such

as 'everything seems to go straight throughme', 'I wake as fresh as a daisy', used bypatients to describe the result of operation,a preliminary inventory of 53 items wasmade and tested on 66 postoperativepatients. These patients had been treatedin two different centres and had an opera-tive outcome ranging from success tofailure.

Statistical analysis of their replies showsthat the response to the 53 items can beaccounted for by three underlying factorsor dimensions. The first of these drawstogether statements reflecting psycho-social state after surgery, the second linksitems concerning specific gastric stateafter surgery, while the third is a dimen-sion of general physical well being.;These preliminary results indicate that

it is possible to assess the quality of apatient's life after surgery in a way whichis simple, brief, and open to publicscrutiny. Since these measures obtained bymeans of a self-administered inventory arequantified, their value as indices can be


http://gut.bmj.com

/G


ownloaded from

http://gut.bmj.com/


put to the test by relating them to clinicalstate and other relevant criteria.

F6

Long-term effects of truncal, selective, andhighly selective vagotomy on gastric secre-ition in man

M. J. GREENALL, P. J. LYNDON, J. C.GOLIGHER, AND D. JOHNSTON (UniversityDepartment of Surgery, The GeneralInfirmary, Leeds) Preservation of thevagal nerve supply to the antrum of thestomach as in HSV allows the antral milland the pyloric sphincter to be kept intact.Side effects of gastric surgery are therebydiminished. However, reports of signifi-cant increases in acid output during thefirst year after HSV and of an increasingincidence of positive Hollander responseshave led to fears that the incidence of re-current ulceration might be prohibitiveafter HSV.We compared basal (BAO), penta-

gastrin-'maximal' (PAOPS) and insulin-stimulated (PAOI) acid and pepsin out-puts in patients who had undergone TV +P (n = 20), SV + P (n = 15), or HSV(n = 25) more than four years previously.The three groups were w-ll matched andeach patient had had a negative insulintest one week after operation.Mean BAO (± ISE) were: TV, 1I1

(+ 0 4); SV 1416 (± 04); HSV 1P50 (±04.Mean PAOP" (iSE) were: TV 16-7

(± 641); SV 13-1 (± 6 0); HSV 23-80 (±3 42).Mean PAOI (+ ISE) were: TV 3-5

1-7); SV 1P72 (+ 0 6); HSV 5-21 (± 1-05).None of the differences between groups

was statistically significant. Pepsin outputsafter HSV were also no higher than afterTV or SV. Eighty-five per cent of patientshad Hollander-positive insulin tests afterHSV but only 60% were positive after TVor SV.These findings lend little support to the

idea that the incidence of recurrent ulcera-tion will be high after HSV. To date, inLeeds, in 250 patients, the incidence ofrecurrence is nil in a follow-up period of0 to five and a half years.

Gl

Agglutinating antibodies in the duodenalsecretions of infants with enteropathic E.coli gastroenteritis

A. S. MCNEISH, H. GAZE, AND N. EVANS (In-

stitute of Child Health, University ofBirmingham) The appearance of agglu-tinating antibodies in the duodenal secre-tions was measured serially in a group of15 infants (aged 1 month to 2 years)with enteropathic E. coli (EPEC) gastro-enteritis. The specificity of these anti-bodies was measured against a panel ofEPEC, before and after absorptionwith a suspension of the specific infectingorganisms. The immunoglobulin (Ig)classes of the antibodies were assessed byabsorption of the intestinal secretions withspecific anti-Ig antisera.The antibody response in the duodenal

secretions was measurable within threedays of infection, and reached a peak at10 to 14 days. IgM antibodies were foundearly in the response, and were the onlydetected antibodies in the youngest in-fants. IgA antibodies were found pre-dominantly in later specimens, obtainedtowards the peak of the agglutinin re-sponse.

In three infants, a poor secretory anti-body response was associated with severeclinical illness.

It is concluded that the antibody re-sponse of the small intestine in EPECgastroenteritis may be an importantfactor in determining the clinical outcome.

G2

Tissue damage mediated by lymphocytes inlocal small intestinal immune reactions

ANNE FERGUSON AND T. T. MCDONALD (Uni-versity ofGlasgow, Department ofBacteri-ology and Immunology, Western Infirmary,Glasgow) The villi of mammalian smallintestine contain lymphoid cells, of whichonly a small proportion are thymus-dependent lymphocytes'. No physiologicalrole for intestinal T lymphocytes has yetbeen defined. However, experiments inrodents suggest that a local cell-mediatedimmune (CMI) reaction may itself con-tribute to tissue damage in small intestinaldisease.

(a) In rejection of allografts of mousesmall intestine , T lymphocyte-mediatedreaction causes crypt hyperplasia andvillous atrophy; however, the morphologydiffers from that of human jejunum incoeliac disease in that in rejection theenterocytes appear normal; there aremany lymphocytes but few plasma cells inthe lamina propria, and relatively fewintraepithelial lymphocytes.(b) Villous atrophy and crypt hyperplasia

are also found in mice in the course oflocal CMI to tuberculoprotein, and in ratsinfected with the parasite Nippostrongylusbrasiliensis (associated with a thymus-dependent immune reaction in the smallintestine).

Local CMI may cause the crypt hyper-plasia of coeliac disease, giardiasis, etc,but is unlikely to be the cause of entero-cyte damage in these conditions. It shouldbe possible to test this theory by measuringsecretion of lymphokines by jejunal biop-sies in short-term organ culture.

References

'Ferguson, A., and Parrott, D. M. V. (1972).Clin. exp. Immunol., 12, 477.

'Ferguson, A., and Parrott, D. M. V. (1973).Transplantation, 15, 546.

G3

Histocompatibility antigens in coeliacdisease

R. FERGUSON, P. ASQU1TH, AND W. T.COOKE (The Nutritional and IntestinalUnit, The General Hospital, Birmingham)Since approximately 80% of coeliacpatients possess HL-A 81 and histocom-patibility antigens are closely linked withimmunological reactions2 the questionarises as to whether coeliacs withoutHL-A 8 are a distinct subpopulation fromthose with HL-A 8 and show a differentimmunological reaction to gluten.To study this possibility the presenting

clinical features, haematological and bio-chemical status, and morphological param-eters of immunological reactivity havebeen compared between 19 HL-A 8coeliacs and 13 'non-8' coeliacs. Theeffect on these factors of gluten with-drawal of at least one year's duration wasalso examined. Results show a significantincrease in serum globulin levels and adecreased lymphocyte population in thelamina propria of the jejunum in theHL-A 8 group when receiving gluten. Nosuch differences are found on glutenwithdrawal, nor in the other parametersdetermined suggesting there may indeedbe a difference in immunological reacti-vity to gluten between coeliacs with andwithout HL-A 8.

References

'Stokes, P. L., Asquith, P., Holmes, G. K. T.Mackintosh, P., and Cooke, W. T. (1972).Lancet, 2, 162-164.

'McDevitt, H. O., and Bodmer, W. F. (1972). Amer.J. Med., 52, 1-8.


http://gut.bmj.com

/G


ownloaded from

http://gut.bmj.com/


G4

Evidence of immunodeficiency in patientswith coeliac disease

P. G. BAKER, J. VERRIER JONES, D. B.PEACOCK, AND A. E. READ (Departments ofMedicine and Bacteriology, University ofBristol) #X 174 is a potent antigen whichhas been used to study antibody produc-tion in animals and in normal subjects(Peacock et al, Clin. exp. Immunol., 43,1973). There is evidence for hyposple-nism in a proportion of patients withcoeliac disease, and impairment of anti-body production has been reported(Pettingale, K. W., Clin. Sci., 38, 1970,16p).

Eight patients (two untreated) withadult coeliac disease were given primaryand secondary injections ofOX 174. Threepatients showed gross impairment ofprimary antibody production, and thesecondary response for the whole groupwas significantly lower than that of 40normal subjects. There was a markedassociation between hyposplenism and apoor secondary response.

This study provides strong evidence forimpaired antibody production in coeliacdisease.

G5

Serum complement components C, and C4in inflammatory bowel disease

M. WARD AND M. A. EASTWOOD (WolfsonGastrointestinal Laboratories, WesternGeneral Hospital, University ofEdinburgh,Edinburgh) The finding of an increasedincidence of serum precipitin reactions tocomplement component C,q in ulcerativecolitis and Crohn's disease' and of C,deposition in the rectal mucosa of patientswith ulcerative colitis2, has raised thepossibility of a role for complement-binding immune complexes in the patho-genesis of inflammatory bowel disease.

In the present study serum C3 and C4component levels were estimated by radialimmunodiffusion in 30 patients withulcerative colitis; 13 patients who hadundergone total colectomy for ulcerativecolitis from two to 14 years previously;32 patients with Crohn's disease; and 14healthy controls.

Significantly elevated C4 levels werefound in the Crohn's disease group as awhole (P < 0 01), including those patientsin apparent remission (P < 0 01). In theulcerative colitis group, C4 levels were

significantly raised only in those cases insevere relapse (p < 001): those in re-mission did not differ from the controlgroup. C, levels in the postcolectomygroup were also significantly elevated(p < 0-01). A similar though less exag-gerated disturbance was seen in C3 levels.These findings suggest: (1) additional

support for the role of complement in thepathogenesis of Crohn's disease andulcerative colitis; (2) differences in comple-ment metabolism between the two diseaseswhen in apparent remission; (3) a persist-ing disturbance of complement metab-olism in some cases of ulcerative colitistreated several years previously by totalcolectomy.

References

"Doe, W. F., Booth, C. C., and Brown, D. L. (1973).Lancet, 1, 402-405.

'Ballard, J., and Shiner, M. (1974). Lancet, 1, 402-403.

Hl

Provoking gastrin release-which stimulus?

L. P. FIELDING, R. C. G. RUSSELL, AND S. R.BLOOM (Surgical Unit, St Mary's Hospital,London) (Introduced by J. G. Walker) Inour experience there is an overlap in thelevel of fasting plasma gastrin between theZollinger-Ellison syndrome and the upperlimit of normal. The response of theplasma gastrin to a protein stimulus helpsto discriminate between these two groups.We have therefore sought to standardizea 'test meal' of this kind.

Seven fasting subjects were studiedwith:(1) Calcium caseinatelOOg(2) Composite packagedfood 100 g(3) Calcium carbonate Made up to2 g 300 ml with(4) Beef extract 25 g tap water(5) Milk 300 ml(6) Arachis oil emulsion100 ml(7) 'English' breakfastThree preliminary blood samples were

taken followed by specimens at 15-minuteintervals for two hours. Plasma gastrinwas measured by radioimmunoassay;statistical significance was assessed bypaired Student t test.The integrated gastrin response (IGR

= area under curve, pg/2 hr) was thelargest after the breakfast and was signifi-cantly different from the next largest,

arachis oil (breakfast 3551; SEM 200;arachis oil 2579; SEM 93; t = 2 94, p <0 02). The breakfast stimulus was re-producible in these subjects on two furthertests.

Seven patients who had had a truncalvagotomy and drainage were then studied.The results showed a highly significantdifference from normals (normals; IGR =3076 ± 460, vagotomy and drainage:IGR = 6150 ± 804, t = 3-63; P < 0 01).These results suggest that a standard

'English' breakfast is the most satisfactorystimulus for the initial investigation ofpatient with equivocal fasting plasmagastrin levels.

H2

Experimental hypergastrinaemia and an-tral-gastrin-cell hyperplasia in dogs

H. M. JENNEWEIN, P. C. GANGULI, F.WALDECK, R. SIEWERT, J. M. POLAK, ANDA. G. E. PEARSE (Pharmaforschung-Biologie,C. H. Boehringer Sohn, Ingelheim-Rhein,University Department of Surgery,Royal Infirmary, Manchester, Klinik undPoliklinik fur Allgemeinchirurgie der Uni-versitdt, Gottintgen, Department of Histo-chemistry, Royal Postgraduate MedicalSchool, London) In order to investigatechronic hypergastrinaemia and G-cellhyperplasia in dogs with Heidenhainpouches, the antrum was separated fromthe rest of the stomach by a mucosalseptum or by complete division. Thegastroduodenal junction was either leftintact or a pyloroplasty was performed.

Gastric secretion was collected from theHeidenhain pouches and the acid outputwas determined by titrating to pH7.Gastrin levels were measured pre- andpostoperatively by radioimmunoassay. Insome samples the gastrins were separatedaccording to their molecular size, employ-ing a 1 x 200 cm Sephadex G-50 SFcolumn. Between two and four monthsafter operation the antra were examinedfor the number of G-ells present (immu-nofluorescence histological technique).

Distinct hypergastrinaemia and acidhypersecretion had developed in dogswhere the antrum was separated from thestomach by a mucosal septum leaving thegastroduodenal junction intact. Gastrinlevels rose to 500 pg/ml (normal 50-80pg/ml). Analysis of the gastrin patternshowed a predominance of little gastrincomponents. The elevated gastrin levelscould be suppressed by secretin. Althoughthere was a marked hypersecretion in most


http://gut.bmj.com

/G


ownloaded from

http://gut.bmj.com/


of the dogs with hypergastrinaemia, no-close correlation between these param-eters was evideWt. In two of three of thedogs with hypcastrinaemia there was ademonstrable G-cell hyperplasia.

H3

A comparison of basal and stimulatedgastric and pancreatic secretion in patientswith and without duodenal ulcer disease

L. V. GUTIERREZ AND J. H. BARON (Depart-ment of Surgery, Royal PostgraduateMedical School, London, England) Duo-denal hyperacidity in patients with duode-nal ulcer disease may be due to gastrichypersecretion and/or decreased duodenalneutralization. Basal gastric secretion andthe maximum response to pentagastrin(6 ug/kg im) and basal secretion into theduodenum and its maximum response tointravenous secretin were measured in 20(12 male) control subjects (aged 24-69)and 10 patients (nine male) with duodenalulcer disease (aged 27-68).

In control subjects the volumes (but notconcentrations or outputs) of stimulatedgastric and duodenal secretions weresignificantly correlated (r = 0-73, p <01001). In patients with duodenal ulcer,there were no significant correlationsbetween basal gastric and duodenal secre-tion, but after stimulation both volumes (r= 0 85, p < 0 01) and outputs (r = 0'86,p x 0101) (but not concentrations) ofgastric and duodenal secretions weresignificantly correlated. The bicarbonatesecretory capacity in patients with duode-nal ulcer (mean 32-2 mmol/h) was com-parable with their gastric acid secretorycapacity (34 5 mmol/h) and similar to thebicarbonate secretory capacity in controlsubjects (31V8 mmol/h).

While in control subjects basal acidoutput (1P2 mmol/h) was only one and onehalf times the basal bicarbonate output,in patients with duodenal ulcer basal acidoutput (3 3 mmol/h) was almost seventimes basal bicarbonate output. Thus, thebasal bicarbonate output in patients withduodenal ulcer (05 mmol/h) was onlyhalf that in control subjects (0-8 mmol/h),perhaps another example of the failure ofthe duodenum in patients with duodenalulcer to respond to a gastric acid loadin spite of the normal bicarbonate secre-tory capacity.

H4

The importance of an innervated intact

antrum in preventing symptomatic duodeno-gastric reflux

M. R. B. KEIGHLEY, P. ASQUITH, A. HOARE,AND J. ALEXANDER-WILLIAMS (The GeneralHospital, Steelhouse Lane, Birmingham)After gastric operations, dyspeptic symp-toms have been shown to be related toduodenogastric reflux. A prospectiverandomized trial was therefore conductedto compare the incidence of duodenalreflux after an operation designed to pre-serve an innervated intact antro-pyloricmechanism: proximal gastric vagotomy(PGV) and one in which the antrum andpylorus was removed, vagotomy andantrectomy (TV + A).

Assessment of 24 patients before andafter operation was by a symptomaticscore based on a questionaire, measure-ment of the concentration of bilirubin infour samples of fasting gastric juice,endoscopic appearances, gastric biopsy,and a radiological study of reflux.There was a significant correlation

between the concentration of bile in thefasting gastric juice and radiological re-flux (P < 0 01), also between dyspepsiaandduodeno-gastric regurgitation (P < 0-05).The relationship between reflux and histo-logical gastritis, however, did not reachstatistical significance (P < 01). Radio-graphic reflux was observed in fivepatients before operation (21 %), in twopatients after PGV (17%), and in sevenfollowing TV + A (58 %).We confirm that after ulcer curative

surgery there are a number of patientswith moderately severe symptoms asso-ciated with duodenal reflux. The re-sults of this trial also suggest that anoperation designed to preserve an intactand innervated antro-pyloric mechanismwill protect against duodenal reflux andsymptoms.

H5

Is there an antral-body portal system in thestomach?

T. V. TAYLOR AND H. BRUCE TORRANCE(Department of Gastroenterology, RoyalInfirmary, Manchester) The direction ofblood flow within the stomach has beenstudied in the rat and in man. In four basicexperiments carried out on groups of sixrats the transport of the isotope Rb86CI,which is extracted from the blood streamto the tissues, has been studied (Sapistein,1958).The first experiment showed that acti-

vity is directly transported from the antralmucosa to the body of the stomach. In thesecond experiment occlusion of all bloodvessels feeding the stomach arrested thetransfer of activity from the antrum to thebody, in the third experiment division ofthe stomach at the antral-body junctionobviated the transfer of activity from theantrum to the body but activity was stilltransferred to the liver. Finally in the in-tact stomach activity was seen to betransmitted from the body to the antrumof the stomach.

In man a similar study has been em-ployed which illustrates the direct transferof radioactivity from the antral mucosato the body of the stomach using a y-camera.A 'portal system' has been demon-

strated between the mucosa of the an-trum and the body of the stomach throughwhich gastrin could be directly trans-ported from G cells in the antrum to actdirectly on the parietal cell mass beforebeing diluted by the systemic circulationand being excreted in the liver and othersites.ReferenceSapirstein, L. A. (1958). Regional blood flow by the

fractional distribution of indicators. Amer.J. Physiol., 193, 161-166.

H6

Protein synthesis and secretion in culturedrabbit gastric mucosal biopsies

D. R. SUTTON AND R. M. DONALDSON(Nottingham General Hospital) Investiga-tion of gastric metabolic function inanimal models is limited by inability toisolate the stomach from all extraneousinfluences while culture of gastric tissuehas only previously been achievedusing free cell suspensions which are onlyviable for a few hours.We report the successful culture of

rabbit gastric mucosal biopsy tissue inwhich a steady state of protein synthesisand secretion was maintained for 24hours. Characterization of the secretionsuggested that most protein was pepsino-gen. Protein synthesis was stimulated bypentagastrin while secretion was enhancedby acetylcholine, pentagastrin, secretin,and cholecystokinin. These gastro-intestinal hormones only effected secretionin the presence of a background of acetyl-choline. This type of organ culture allowssynthesis and secretion of various macro-molecules to be studied in detail for pro-longed periods.


http://gut.bmj.com

/G


ownloaded from

http://gut.bmj.com/


JA

HypercataboHsm of third component ofcomplement in pimary biliary cirrhosis

B. J. POITER, E. ELIAS, AND E. A. JONES(Departments of Medicine, Royal FreeHospital, London, and University ofLiverpool) Complement-mediated im-mune injury is one of the possible mecha-nisms by which small bile ducts may bedamaged in primary biliary cirrhosis(PBC). Evidence consistent with activa-tion of the complement system was soughtby (1) analysing the plasma radioactivitycurve and urinary excretion of radio-activity" after the intravenous administra-tion of pure undenatured radioiodinatedC3' and (2) screening sera for conversionproducts of C3 using Laurell crossedimmunoelectrophoresis3 and specific anti-C3 antiserum.The metabolism of 1251-C3 was studied

in seven patients with PBC and five normalcontrol subjects. The fractional catabolicrate of C3 was higher in PBC (430%intravascular pool/hr ± 0 94; mean ±1 SD) than in normals (2-04 ± 020)(p < 0O0005). There was no overlapbetween the plasma 1512-C3 disappearancecurves in PBC and normals. The syntheticrate of C3 also was higher in PBC (3-10mg/kg/hr + 1P59) than in normals (105 +031) (p < 001). Three of the plasmadisappearance curves of 12SI.-C3 failed toreveal an early terminal exponentialdecline, suggesting that some C3 may beconverted to a metabolic product, pos-sibly C3d, which is catabolized moreslowly than C3. In the serum of onepatient conversion products of C3 weredemonstrated.The results suggest that the complement

system may be activated in PBC.

Referenices

'Nosslin, B. (1973). Ciba Foundation Symposium,9, 113.

'Nilson, U. R., and Muller-Eberhard, H. J. (1965).J. exp. Med., 122, 277.

'Minchin Clarke, H. G., and Freeman, T. (1968).Clin. Sci., 35, 403.

J2

The effect of portocaval shunt on colon&Waer of ammonia in cirrhosis of theliver

MAUNG MAUNG SEIN, ANNE HARDY-SMITH,AND. R. SHIELDS (Department of Surgery,University ofLiverpool) Faecal ammoniaand electrolytes were measured using an

'in-vivo' dialysis method (Wrong et al,1961) in six control subjects, and 12patients with liver disease including sixwho had had portocaval shunts.

In the patients with shunts, the con-centration of ammonia in faecal dialysatewas similar to that of controls; in patientswithout shunts, the faecal concentrationof ammonia was raised. In the patientswith shunts the concentration of ammoniain the peripheral blood was signifi-cantly higher than those without shunts;both were higher than in control subjects.Sodium concentrations in the faecal

dialysates of patients with shunts weresignificantly decreased compared withthose from patients without shunts, orcontrols, while the potassium concentra-tions were increased. These findingssuggest a state of hyperaldosteronism.

It is suggested that following a shuntthere is an increased absorption of ammo-nia from the colon into the portal circu-lation. This increased absorption ofammonia is one factor contributing tothe hyperammonaemia, and possiblyhepatoportal encephalopathy, found aftershunt.

Reference

Wrong, 0. M., Morrison, R. B. I., and Hurst, P. E.(1961). A method of obtaining faecal fluidby in vivo dialysis. Lancet, 1, 1208-1209.

J3

An approach to the diagnosis of liverdisorders using 1231-BSP

A. REUBEN, K. NARASIMAH, M. MYERS, T.WOOD, AND K. BRrrON (Institute ofNuclear Medicine, Middlesex Hospital,London, England) The differentiation of'surgical' jaundice from intrahepatic cho-lestasis, the detection of liver diseasebefore biochemical changes in the blood,and the demonstration of space-occupyinglesions and their pathology are areas ofdiagnostic importance. A method of in-vestigation is described which is appli-cable to each of these problems, using123I-BSP. Blood clearance and hepatic up-take curves, and in addition liver and gall-bladder pictures, are obtained using ex-ternal gamma camera monitoring follow-ing a small intravenous injection of1"3I-BSP (< 100 mg) after O9mTc-albuminis injected as a background marker.Jaundice does not interfere with themeasurement unlike chemical BSP. Theradiation dose is approximately one tenth

of a single standard chest radiograph. Thehigh count rate obtained with 1as3I-BSPsafely allows good gamma camera defini-tion of space-occupying lesions and 99mTc-albumin enables blood pool scanning inthese cases. Liver, gallbladder, and gutactivity may be seen showing the level ofsurgical obstruction. A more accurateestimate of bile output is obtained than bycompartmental analysis of plasma clear-ance curves of chemical BSP. In non-obstructive liver disease deconvolutionanalysis of blood clearance and bile out-put curves gives a distribution of BSPtransit times. In seven normal subjectsthe first peak of this distribution is narrowand unimodal with a mean transit time ofless than 10 minutes. In cases of intra-hepatic cholestasis, eg, primary biliary cir-rhosis, this distribution becomes widenedand the mean transit of the first peaktime greatly prolonged.The BSP transit time distribution is a

promising way of characterizing differentliver diseases and we present results in avariety of disorders.

J4

The influence of chronic liver disease onthe elimination of d-propranolol, anti-pyrine, and indocyanine green

R. A. BRANCH, J. A. JAMES, AND A. E.READ (University Department of Medicine,Royal Infirmary, Bristol) Many factorscan influence drug pharmacokinetics inpatients with chronic liver disease. Thisstudy assesses the relative importance ofdrug-metabolizing activity and liver bloodflow for a rapidly metabolized drug, d-propranololl, and a slowly metabolizedone, antipyrine2. Twenty patients withstable chronic liver disease and sixhealthy controls received 40 mg iv d-pro-pranolol, 1200 mg oral antipyrine, and0 5 mg/kg indocyanine green, to measureliver blood flow (LBF), on separateoccasions.The half-life of each drug increases in

chronic liver disease, the increase beingmore marked in patients with severeliver disease. The serum albumin andbilirubin correlated significantly with thehalf-lives of each drug. There were alsosignificant correlations between the half-lives of individual drugs. A comparison ofthe intrinsic metabolic clearance, an indi-cator of drug metabolizing efficiencyindependent of LBP, was made withLBF and actual drug clearance; thedecreases in actual drug clearance of anti-


http://gut.bmj.com

/G


ownloaded from

http://gut.bmj.com/


pyrine and d-propranolol are mainlydue to a reduction in intrinsic metabolicclearance. However, patients with anormal serum albumin but a reducedLBF had a reduced clearance of d-pro-pranolol but not of antipyrine. In conclu-sion, the rate of elimination of a drugdepends more on the drug metabolizingactivity of a damaged liver than on itsblood supply. However if the metabolicactivity is high, a fall in LBF will decreasethe rate of actual drug elimination.

References

'Shand, D. G., and Ragno, R. E. (1972). Disposi.tion of propranolol. Pharmacology, 7, 159.

'Branch, R. A., Herbert, C. M., and Read, A. E.(1973). Determinants of antipyrine half-lives in patients with liver disease. Gut, 14,569.

'Rowland, M., Benet, L. Z., and Graham, G. G,(1973). Clearance concepts in pharmaco.kinetics. J. Pharmacokin. Biopharmacol., 1,123.

J5

Lysosomal changes in human and experi-mental iron overload

C. A. SEYMOUR, G BUDILLON, AND T. J.PETERS (Department of Medicine, RoyalPostgraduate Medical School, London)Little is known of the mechanism oftissue damage in iron storage diseases.Lysosomal disruption has been implicatedas a possible mechanism of cell damage.Therefore the lysosomal changes in liverbiopsies from patients with iron overloaddue to haemochromatosis or secondaryhaemosiderosis were studied. The specificactivities (milliUnits per milligram pro-tein) of six acid hydrolases (acid phos-phatase, /-glucuronidase, N-acetyl-fl-glu-cosaminidase, ox-glucosidase, /3-galacto-sidase, acid diesterase) were two to 10times higher in biopsies from patients withiron overload than in biopsies from con-trol subjects or patients with other chronicliver diseases. This suggests an accumula-tion of undegradable material within thelysosomes.

Determination of lysosomal integrityin these biopsies (latent N-acetyl-3-glu-cosaminidase and sedimentable acid hy-drolases) showed that these organelleswere particularly fragile. No such changeswere found in biopsies from patients withcryptogenic cirrhosis or other chronicliver diseases. After iron was removed byvenesection these parameters returnedto normal.

Hepatic lysosomes from iron over-loaded rats were also shown to be morefragile than normal. Using sucrose gra-dient centrifugation these lysosomes werefound to be very dense because of accu-mulated insoluble iron compounds withinthem.

It is suggested that accumulation ofexcess iron, probably as haemosiderin,within hepatocyte lysosomes leads todisruption of these organelles and thisinitiates cell damage.

J6

Prolonged survival in the Crigler-Najjarsyndrome: Ultrastructural and dietarystudies

J. L. GOLLAN, S. N. HUANG, B. H. BILLING,AND S. SHERLOCK (Department of Medi-cine, Royal Free Hospital, Gray's InnRoad, London) Three brothers, who hadbeen deeply jaundiced for over 50 years,have been studied. Although their plasmaunconjugated bilirubin concentrations(UCB) were greater than 20 mg per 100ml they showed no abnormal neurologicalsigns. Prolonged exposure to severe Un-conjugated hyperbilirubinaemia does nottherefore necessarily increase morbidity.

Conventional tests of liver function werenormal and on light microscopy the onlyabnormality of liver histology wasoccasional bile plugs in undilated bilecanaliculi. Electronmicroscopy showedhypertrophy and hyperplasia of smoothendoplasmic reticulum with unusualprominence of the Golgi apparatus. Focalmodification of the cell surface mem-branes was also evident. These changesmay be implicated in the transport ofunconjugated bilirubin from the hepato-cyte.

Dietary restriction for 72 hours to 400calories per day produced a dramaticincrease in UCB. The addition of 2400calories by the intravenous administra-tion of 50% dextrose did not reduce theelevated UCB. In contrast, when 2400calories was fed as a normal diet the UCBreturned to basal levels. While on pheno-barbitone therapy (which caused a reduc-tion of UCB to 4 mg per 100 ml withintwo weeks) the response to caloric with-drawal and parenteral feeding was similar.These results indicate that the hyper-bilirubinaemia of fasting is not justattributable to caloric deprivation andsuggests that the type of diet may in-fluence the degree of hyperbilirubinaemia.

Kl

Biochemical studies of pure pancreaticjuice obtained by duodenoscopic cannu-lation of the pancreatic duct in consciouspatients

P. B. COTTON, M. CREMER, P. ROBBERECHT,AND J. CHRISTOPHE (St Thomas's Hospital,and Middlesex Hospital, London H6pitalUniversitaire Brugmann, and Laboratoirede Biochemie, Universite libre de Bruxelles,Belgium) In 27 patients undergoingendoscopy and retrograde cholangio-pancreatography (ERCP), it has beenpossible to obtain pure secretions fromwithin the pancreatic duct. Intravenousinjection of secretin (Boots, 1 unit/kg)produced a brisk flow, varying in appa-rently normal patients from 2-7 to 8-0ml/minute. In patients with definite pan-creatic pathology, the response range was0-94 ml/minute. Further injections ofpancreozymin (Boots 1 unit/kg) or ceru-lein (20 ng/kg) were given, and collectionscontinued each minute for up to 40 min-utes. Pancreatic juice sodium and potas-sium concentrations remained constantin each patient. Despite immediate freez-ing of samples, marked loss of bicarbonateoccurred unless paraffin was added. Insamples so protected, measured bicarbo-nate concentration closely resembled thedifference between chloride concentra-tion and the sum of sodium and potas-sium concentrations. The minimal chlor-ide concentration and maximum bicar-bonate concentration were achieved withinfive to 10 minutes of secretin injection.Calculated bicarbonate concentrationsvaried little between normal and ab-normal patients (normals 115-143 m-equivll; definite abnormals 86-133 m-equiv/l). Maximum bicarbonate outputranged from 90 to 1200 4-equiv/minute.The concentrations of protein and hydro-lases (lipase, amylase, chymotrypsinogen)reached a sharp peak within two minutesof a stimulus. The responses varied con-siderably between patients, and in rela-tion to bicarbonate secretion.

This method of collecting pure pancre-atic juice in conscious man provides a newavenue for research into pancreaticphysiology and pathophysiology.

K2

Studies on the nature of cholecystokinin-pancreozymin in small-intestinal mucosalextracts

R. F. HARVEY, LYNDA DOWSETT, AND A. E.


http://gut.bmj.com

/G


ownloaded from

http://gut.bmj.com/

T7he British Society of Gastroenterology 839

READ (University Department of Medicine,Bristol Royal Infirmary, Bristol) Thecholecystokinin-pancreozymin (CCK) iso-lated from small-intestinal mucosa byJorpes and Mutt is a 33-amino acid pep-tide". In order to determine whether otherCCK-like peptides are present in thesmall intestine, we have used gel filtrationto study the properties of the CCK insmall-intestinal mucosal extracts. CCKlike imrnmunoreactivity was measured by asensitive radioimmunoassay, using anantiserum (MS6/72) reacting at the bio-logically active (C-terminal) end of themolecule'.During passage through Sephadex

G-50 columns (30 x 500 mm), CCK-like immunoreactivity separated intothree distinct components. Onecomponent'corresponded in elution position toJorpes-Mutt CCK, one had apparentlya greater molecular weight than this('big' CCK), and the third componenteluted in a position similar to that of theC-terminal octapeptide of CCK. A'little' CCK component, also similar tothe octapeptide in its elution character-istics, could be produced by incubatingeither 'big' or Jorpes-Mutt CCK withtrypsin at 35°C for two hours.As the biological activity of the C-

terminal octapeptide is considerablygreater than that of Jorpes-Mutt CCK',these findings suggest that a large part ofof the CCK-like biological activity oftisssbe extracts may result from the actionof a relatively small C-terninal fragment,which is possibly identical with the octa-peptide. Such a fragment may be producedin vivo by enzymes acting on the largercholecystokinins.

'Jorpes, J. E., and Mutt, V. (1973). Secretin, chole-cystokinin, pancreozymin and gastrin.Handbook of Experimental Pharmacology.vol. 34. Springer, Berlin, pp. 1-179.

'Harvey, R. F., Lynda Dowsett, Hartog, M., andRead, A. E. (1974). Radioimmunoassay ofcholecystokinin-pancreozymin. Gut (inpress).

'Rubin, B., and Engel, S. L. (1973). Some biologicalcharacteristics of cholecystokinin (CCK-PZ)and synthetic analogues. Frontiers in gastro-intestinal hormone research. Nobel Sympo-'sium 16. Andersson, S. Ed. Almqvist andWiksell, Stockholm, pp. 41-56.

K3

Panreatic changes induced by chronic(two years) ethanol treatment in the4Dg

0.- M. TISCORNIA, G. PALASCJANO, AND H.

SARLES(Unit6de Recherches de PathologieDigestive de l'INSERM 46, chemin de laGaye, 13009, Marseille, France) Inprevious studies' , the action of ethanolon the dog pancreas was studied in non-alcoholic animals and in animals sub-mitted to short duration consumption ofethanol.

Twelve dogs were provided with chronicpancreatic and gastric fistulae (Thomascannula); six served as control, and sixwere alcohol fed (2-0 g/kg) for two years.In chronic alcoholic animals, the follow-ing changes were observed:

1 After a test meal, either alone orassociated with 50% intragastric (ig)ethanol (1 0 g/kg) gastrin blood levels arehigher and better sustained,

2 Accordingly, when pancreatic secre-tion is stimulated by a continuous secretin(GIH, 10 CU/kg/h) and CCK-PZ (GIH,3 0 CHR U/kg/h) perfusion and gastriccontent is prevented from entering theduodenum, an acute ig 15% ethanoladministration (1 3 g/kg) elicits a higherpancreatic protein secretion,

3 Xylocaine spraying of the duodenalpapilla every 10 min for two h evokes anincrease mainly on the protein pancreaticsecretion plateau values. This effect ismore marked in alcoholic animals,

4 The increase of the different CCK-secretin-induced pancreatic secretoryvalues elicited by an acute iv ethanol infu-sion (1 3 g/kg) on the alcohol-fed dogs isabolished by atropine (I10 mg/kg/h) butnot by a pentolinium perfusion (0-75 mg/kg/h): the iv ethanol-evoked stimulationofpancreatic secretion is exerted through acholinergic mechanism triggered at thehypothalamic bulbar centres and/or onnon-nicotinic receptors of the intrapan-creatic ganglia,

5 Dose response curves to GIH secre-tin perfusion (0-5 to 0-8 CU/kg/h) showsa higher maximal bicarbonate output thanin the control. This ethanol-inducedincrease in the mass and/or number of'pancreon' units must be related, at leastin part, to the trophic effects of raisedlevels of gastrin and CCK-PZ.

Refwenc

'Tiscornia, O., Gullo, L., Sarles, H., Devaux, M. A.,Michel, G., and Grimaud, R. (1973). Diges-tion, 9, 231-240. Gastroenterology (abstract)62, no. 4, 866 (1972).

'Sarles, H., Tiscornia, O., Palasciano, G., Brasca,A., Hage, G., and Devaux, M. A. (1972).Scand. J. Gastroent., 8, 85-96.

K4

The proximal jejunal mucosa in colonicCrohn's disease and ulcerative colitis

R. FERGUSON, R. N. ALLAN, AND W. T.cooKE (The Nutritional and IntestinalUnit, The General Hospital, Birmingham)Morphological abnormalities of the jeju-num have been reported in inflammatorybowel disease1 2. We have studied themucosal cell population of the proximaljejunum in 26 patients with Crohn'sdisease of the colon and for comparisonthat in 20 randomly selected patients withulcerative colitis and in 20 normal con-trols. Our results show that a significantincrease (p < 0-005) in the plasma cellpopulation of the lamina propria occurswith both ulcerative colitis and Crohn'sdisease in comparison to normal controls,even in those cases with previous colec-tomy. In addition, patients with Crohn'sdisease have significantly higher (p <0O05) plasma cell populations than subjectswith ulcerative colitis. This increase inplasma cells is not related to the severityof illness, extent of disease, nutritionalstate, or therapy in Crohn's disease, incontrast to patients with ulcerative colitis,where exacerbation of disease was asso-ciated with a greatly elevated plasmacell infiltrate. These findings are alsosignificantly different from the resultswe have observed in coeliac disease'.

References

'Salem, S. M., and Truelove, S. C. (1965). Brit. med.J., 1, 827-831.

'Hermos, J. A., Cooper, H. L., Kramer, P., andTrier, J. S. (1970). Gastroenterology, 59,868-873.

'Ferguson, R., Asquith, P., and Cooke, W. T.(1974). Gut (in press).

K5

Bidirectional sodium flux across theintestinal mucosa in Crohn's disease

R. N. ALLAN, D. M. STEINBERG, K. DIXON,AND W. T. COOKE (Nutritional and IntestinalUnit, The General Hospital, Birmingham)We have observed that patients treatedby panproctocolectomy and ileostomy forCrohn's colitis are more prone to epi-sodes of flux than a similarly treated groupof patients with ulcerative colitis, eventhough they are otherwise in good healthand there is no radiological evidence ofrecurrent disease in the small bowel. Themagnitude of these fluxes resembles thatof the fluid loss from the intestinal mucosa


http://gut.bmj.com

/G


ownloaded from

http://gut.bmj.com/


in cholera where a disturbance of bidirec-tional sodium flux has been demonstrated(Love et al, 1972).We have measured the bidirectional

sodium flux in patients with Crohn'sdisease treated by panproctocolectomywith ileostomy and compared them with asimilarly treated group of patients withulcerative colitis using a radioisotopetechnique with Na22 as a marker, modifiedfrom the method described by Love et al(1973).

Patients with Crohn's disease have asignificant reduction in bidirectionalsodium flux which is one reason they areprone to episodes of flux but also addsweight to the concept that Crohn's diseaseis a diffuse disease of the gastrointestinaltract and may interfere with normal intes-tinal function even in the absence of mac-roscopic or radiological evidence of re-current disease.

References

Love, A. H. G., Phillips, R. A., Rohde, J. E., andVeall, N. (1972). Sodium iron movementacross intestinal mucosa in cholera patients.Lancet, 2, 151.

Love, A. H. G., Rohde, J. E., Abrams, A. E., andVeall, N. (1973). Measurement of bidirec-tional sodium flux across the intestinal wallin man using whole gut perfusion. Clin. Sci.,44, 267.

K6

Physical characteristics of vegetablefoodstuffs that could influence bowelfunction

A. A. MCCONNELL, M. A. EASTWOOD, ANDW. D. MITCHELL (Wolfson GastrointestinalLaboratories, Western General Hospital,University ofEdinburgh, Edinburgh) Thereis a need to explore physical characteris-tics of dietary vegetable fibre that can betested in nutritional studies. This paperdescribes two properties of the fibre ofcommonly eaten vegetables and fruit,namely water-holding capacity (WHC)and ion exchange capacity.

Twenty-six different fruit and vege-tables were dried to a powder. The capa-city of the powder to hold water was esti-mated. Lettuce, carrot, cucumber, celery,and aubergine had the greatest WHCwhereas maize, oatmeal, potato, banana,wheat, and bran had the least. The wateradsorptive capacity of any dietary plantis determined by its fibre content and thewater-holding capacity of the fibre.Estimations of the fibre content of theoriginal plant and the WHC of that fibre

suggest that in their natural state, bran(447 g water holding/100 g originalmaterial), mango (312 g), carrot (208 g),apple (177 g), brussel sprouts (168 g) arebetter as hydrophilic laxatives thanbananas (68 g), cauliflower (68 g), pota-toes (49 g), and turnip (37 g).Most of the fibres act as monofunctional

weak cation exchange resins; lettuce (3-1m-equiv/g), cabbage, carrot, orange,turnip (2-4 m-equiv/g) are in the range ofcommercially available weak cation ex-change resin.When a diet is changed to one with

high fibre content then these physicalcharacteristics may influence the choice ofvegetables and fruit which are included inthe diet.

Ll

Effect of urogastrone in the Zollinger-Ellison syndrome

J. B. ELDER, P. C. GANGULI, I. E. GILLESPIE,I. DELAMORE, AND H. GREGORY (UniversityDepartments ofSurgery and Haematology,Royal Infirmary, Manchester, and Re-search Division, ICI Laboratories.Alderley Park, Cheshire) In normalvolunteers urogastrone (0-25 ,ig/kg/hriv) inhibited exogenously stimulatedgastric secretion but had no effect onserum gastrin concentration and producedno significant clinical side effects (Gillespieet al, 1974).Four male patients with proven

Zollinger-Ellison syndrome have beenstudied: one had previously undergonefour operations including two gastricresections, two hadhad a truncal vagotomyand pyloroplasty, and one remained un-operated at the time of study. Informedconsent was obtained.

Gastric secretion and venous bloodsamples were collected at 15-minute inter-vals after an overnight fast. Following a90-minute plateau period of basal secre-tion, urogastrone in the above dose wasgiven intravenously for one hour. Sampl-ing continued for a further minimumperiod of one hour after the end of theinfusion. Measurements of acid, pepsin,intrinsic factor, and plasma gastrin con-centrations were made in the appropriatesamples.

After urogastrone both acid volumeand concentration decreased and basalacid output was reduced by 60 to 82%.The concentrations of intrinsic factorand pepsin in gastric juice increased by60 to 300%. Peak plasma gastrin con-

centration after urogastrone infusion in-creased by 127 to 164% of basal concen-tration. A significant negative correlationbetween increase in plasma gastrin con-centration and decrease in acid outputwas observed (r = 0-72, p < 0-01).Ulcer pain was relieved 30-60 minutesafter the beginning of urogastrone infu-sion.

These results suggest that urogastronecan inhibit the endogenously stimulatedacid hypersecretion in Zollinger-Ellisonpatients. The pattern of inhibition iscompatible with a competitive actionagainst gastrin for receptor sites on theoxyntic cell.

Reference

Gillespie, I. E., Elder, J. B., Ganguli, P. C.,Gregory, H., and Gerring, L. (1974). Gut,15, 337.

L2

The G cell population of the gastricantrum, plasma gastrin, and gastric acidsecretion

C. M. S. ROYSTON, JULIA M. POLAK, S. BLOOM,W. M. COOKE, C. R. C. RUSSELL, A. G. E.PEARSE, J. H. BARON, J. SPENCER, ANDR. B. WELBOURN (Departments of Surgeryand Histochemistry, Royal PostgraduateMedical School and Institute of ClinicalResearch, Middlesex Hospital) The Gcell populations of gastric antra removedat operation in 20 patients with duodenalulcers were estimated by a modificationof the method of Card and Marks (1960).The G cells were stained by indirectimmunofluorescence (Coons et al, 1955)with an antibody to synthetic human gas-trin 1. BAO, PAOPG, basal gastrin andgastrin release by insulin and bicarbo-nate (Hansky et al, 1971) were measuredpreoperatively. The mean G cell popula-tion was 19 6 x 106 (SE 3 8), mean BAO10-3 m-equiv/h (SE 4'4) and mean PAOpG46-9 m-equiv/h (SE 7 7). The populationshowed no correlation with BAO (r =0-165), PAOpG (r = 0 253), basal gastrin(r = 0-164) or gastrin release (r = 0 039).There was highly significant correlationbetween basal gastrin and BAO-(r =0-673, 0-01 > P > 0-001) but no correla-tion between basal gastrin and PAO(r = 0 283) or gastrin release and PAO(r = 0-051).The G cells were graded subjectively

as normal in number or hyperplastic inthese 20 patients and in four others.Seventeen were 'normal' and had G cell


http://gut.bmj.com

/G


ownloaded from

http://gut.bmj.com/


populations of 2-3 to 2617 x 106; sevenwere 'hyperplastic' with populations of18-3 to 45-8 x 10.In 28 antra the distribution of G cells

was assessed by examination of about 15blocks from each and the average numbersof G cells per 1 mm of mucosa werecalculated. In 24 cases the distributionwas even and in four cases, showingdiffuse gastritis, it was uneven.

Refeeces

Card, W. I., and Marks, S. (1960). Clin. Sci., 19,147.

Coons, A. H., Leduc, E. H., and Connolly, J. M.(1955). J. exp. Med., 102, 49.

Hansky, J., Korman, M. G., Cowley, D. J., andBaron, J. H. (1971). Gut, 12, 959.

I3

The secretin provocation test in thediagnosis of Zollinger-Ellison syndrome

S. BONFILS, M. MIGNON, AND J. P. ACCARY(Units de Recherches de Gastroenterologie,INSERM U-10, HOpital Bichat, Paris)(Introduced by Dr J. H. Baron) Themeasurement of blood gastrin in fastingpatients is a valuable tool for the diagnosisof Zollinger-Ellison Syndrome (ZES).However, increased levels range fromtwice to more than 20 times normal, sothat in some patients provocative testsare needed.The claim that an injection of secretin

is a paradoxical stimulant for both bloodgastrin and acid secretion in patients withZES has often been made, but the failuresof the test are seldom, if ever, mentioned.

Gastric acid and blood gastrin weremeasured- before, during, and after theinfusion of GIH secretin 3 Unit/kg/hourfor one hour in 10 patients with ZES,

five patients with gastric hypersecretionwithout pancreatic tumour, and in fourpatients with achlorhydria.

In the four unoperated patients withZES blood gastrin and acid output in-creased simultaneously, but in threepatients with a total gastrectomy and aremaining endocrine tumour, only oneshowed an increase in blood gastrinduring the infusion of secretin. In fourother patients with a total gastrectomyand whose tumours had apparently beencompletely removed (fasting blood gastrin14-41 pg/ml), a secretin-induced increasein blood gastrin was observed in two.

In the five patients with gastic hyper-secretion without pancreatic tumour(fasting blood gastrin 10-103 pg/ml)neither blood gastrin nor acid outputincreased. Blood gastrin decreased aftersecretin in three of the four patients withachlorhydria.

These results suggest that the secretinprovocation test is of considerable valuein the diagnosis of ZES patients with anon-operated stomach, and (on the basisof one additional case), for familialscreening in the Wermer syndrome. Inother situations, especially in postopera-tive follow up, it seems doubtful whetherthe secretin provocation test can reliablyassess the presence or absence of a persis-tent gastrin-secreting tumour.

1A

Gastric emptying rate measurement inman: A method for simultaneous study ofsolid and liquid phases

R. C. HEADING, P. TOTHILL, G. P.MCLOUGHLIN, AND D. J. C. SHEARMAN(Gastrointestinal Section, Department of

Therapeutics and Department of MedicalPhysics, Royal Infirmary of Edinburgh)Gastric emptying rates for the liquid andsolid components of a standard meal havebeen assessed using a new method bywhich both are measured simultaneously.The meal contained cornflakes and milktogether with 200 tci lismln DTPAchelate as a marker of the aqueous phaseand, as a marker of the solid phase,approximately 30 pieces of filter paper, 3mm square, impregnated with 200 ,uci*9mTc sulphur colloid and coated with athin film of perspex. Sequential scinti-scanning of the upper abdomen was per-formed at 25-minute intervals beginning30 minutes after ingestion of the meal(Heading et al, 1971).

In 15 normal subjects emptying of theaqueous phase of the meal approximatedto a simple exponential process but thesolid phase marker apparently emptiedat a constant rate. In almost all patientsthis was substantially slower than empty-ing of the liquid phase and the two ratescorrelated poorly, indicating that onecannot be inferred from the other. Somepatients who had undergone gastric sur-gery exhibited particularly fast emptyingof the solid relative to the liquid phase.The results indicate that measurements

relating to both the solid and liquidcomponents of a natural meal may berelevant in defining the specific abnormal-ities of emptying which cause symptomsafter gastric surgery.

Reference

Heading, R. C., Tothill, P., Laidlaw, A. J., andShearman, D. J. C. (1971). An evaluation ofi'lmindium DTPA chelate in the measure-ment of gastric emptying by scintiscanning.Gut, 12, 611-615.


http://gut.bmj.com

/G


ownloaded from

http://gut.bmj.com/

Documents

TheBritish Society of Gastroenterology - BMJ