the british society of gastroenterology · study day on the wednesday a full scientific programme...

Gut, 1978, 19, A955-A998

The British Society of Gastroenterology

The 39th Annual Meeting of the British Society of Gastroenterology and the 7th Annual Meeting of theBritish Society for Digestive Endoscopy were held at Edinburgh on 20-23 September 1978 under the Presi-dencies of Dr W. Sircus, President of the BSG, and Dr M. Atkinson, President of the BSDE. Following astudy day on the Wednesday a full scientific programme of 167 papers was presented to the two Societies.The Plenary Session included the Sir Arthur Hurst Lecture given by Professor Dame Sheila Sherlock entitled'Untoward hepatic drug reactions'. The abstracts of the 167 papers follow.

ENDOSCOPY

Open access GP endoscopy

G. HOLDSTOCK, M. WISEMAN, AND C. LOEHRY(Royal Victoria Hospital, Bournemouth)We report on three years' experience of aGP direct referral gastroscopy service.In all, there were 1077 GP referrals,compared to 728 hospital ones. Therewas no significant difference in either thetype or duration of symptoms betweenthe two groups, the vast majority havingdyspepsia. Fewer of the GP patientshad had a trial of antacids than thehospital referrals. Comparing the firstyear of the service with the third, 265additional endoscopies were performedbut no extra ulcers or cancers were found.The pick-up rate for these two conditionscombined fell from 25 to 13 %. There wasno significant difference between thepathology found in either group. Despitethe increase in numbers of GP referrals,there was no reduction in the number ofpatients referred either from hospitalclinics for endoscopy or for bariumstudies.

All the GPs in the areas were sentquestionnaires and all found the serviceuseful. Seventy-five per cent felt thatclinic referral was reduced. It is concludedthat, although feasible and useful to theGPs, the introduction of the serviceresults in too many patients being referred,and, consequently, there are few objectiveadvantages. Possible ways of limiting thenumber of patients referred are discussed.

Preventing transfer of malignant cells onnon-disposable brushes used for cytology

M. R. B. KEIGHLEY, T. MAKURIA, JANETMOORE, AND H. THOMPSON (The GeneralHospital, Birmingham) The accuracy ofdiagnosing gastric carcinoma is increasedby the use of brush cytology.1 In ourexperience, even with non-disposablebrushes, false positive cytology is rare.2Four patients in this unit were endoscopedwith the same brush on one day, none hadvisual or biopsy evidence of cancer, butmalignant cells were seen on brushcytology in all patients. The neoplasticcells were traced to a patient examinedon the previous day. The incidence andprevention of malignant cell transfer wasinvestigated.

Cells were recovered from brushes usedon a previous endoscopy list in six of11 occasions. Vigorous cleaning of thebrush removed cells from only sevenof 12 contaminated brushes. Washing withdetergent, hibitane, pyroneg, and pan-creatin was without effect but autoclavingremoved cells in all of eight brushes.Repeated autoclaving (x 50) did notdamage the bristles on disposable brushes.The mean number of cells recovered fromdisposable brushes (6 x 104) was lessthan from non-disposable brushes(1 X 106).We conclude that, because of an in-

creased yield of cells and reduced costof non-disposable brushes, this practiceshould be retained provided that brushesare autoclaved after each examination.

References1Witzel, L., Halter, F., Gretillat, P. A., Scheurer, U.,

and Keller, M. (1976). Evaluation of specificvalue of endoscopic biopsies and brush cytologyfor malignancies of the oesophagus and stomach.Gut, 17, 375-377.

2Thompson, H., Hoare, A. M., Dykes, P. W.,Allan, R. N., and Keighley, M. R. B. (1977).A prospective randomised trial to comparebrush cytology before or after punch biopsyfor endoscopic diagnosis of gastric cancer.Gut, 18, A398-399.

Importance of repeated endoscopy indetermining the source of haemorrhage inportal hypertension

K. J. MITCHELL, B. R. D. MACDOUGALL,D. B. A. SILK, AND ROGER WILLIAMS(Liver Unit, King's College Hospital andMedical School, Denmark Hill, London)Although it is known that patients withportal hypertension and haematemesiscommonly develop gastric and duodenalerosions, there is uncertainty as to whetherthe haemorrhage originates from theseor from ruptured varices. To investigatethis, two experienced observers performedemergency endoscopy in 76 patients withportal hypertension and oesophagealvarices referred for management ofhaematemesis.

All patients were endoscoped within24 hours (median four hours) of haemate-mesis. By this time no evidence of anyvisible bleeding was found in 54 patients(71 %). Spurting or oozing varices wereobserved in 18 (23-7%). Although super-ficial erosions were seen in a further18 patients (23 7%), visible bleeding wasobserved in only four (5 3 %).

Despite cessation of bleeding on initialendoscopy, 33 patients re-bled during the

A955

on February 9, 2021 by guest. P

rotected by copyright.http://gut.bm

j.com/

Gut: first published as 10.1136/gut.19.10.A

955 on 1 October 1978. D

ownloaded from

http://gut.bmj.com/

A956 The British Society of Gastroenterology

same admission. Twenty-one were re-endoscoped within six hours (median09 hours). Bleeding from rupturedvarices was observed more frequently inthis group (71P4%) than at initial endo-scopy (23-7%, p > 0 05) and none wasfound to be bleeding from erosions.Furthermore, two patients previouslynoted to be bleeding from erosions werebleeding from varices when re-endoscoped.

In two respects these findings differfrom those of previous endoscopy series.Firstly, no visible haemorrhage could beidentified in most cases on initial endo-scopy. Secondly, recurrent haemorrhagewas common, and originated from varicesand not other sites.

Peptic ulceration in patients with chronicliver disease

A. P. KIRK, R. H. HUNT, J. S. DOOLEY,AND SHEILA SHERLOCK (Department ofMedicine, Royal Free Hospital, London)Peptic ulceration is said to be morecommon in chronic liver disease,' andoften asymptomatic. To assess this,163 patients with chronic liver diseaseunderwent upper gastrointestinal endo-scopy, 106 having dyspeptic symptoms.

Twenty-four peptic ulcers were found(14 7y%), 12 were duodenal, eight gastric,and four pre-pyloric. Twenty-two of these24 patients were symptomatic. Ulcerswere found in five of 15 patients withhepatitis B antigen positive chronic liverdisease (33 %), 10 of 46 with chronicalcoholic liver disease (21.7%), five of 35with primary biliary cirrhosis (14-3 %),two of 19 with miscellaneous liver diseases(105 yo), two of 25 with cryptogeniccirrhosis (8 %). Ulcers were not demon-strated in any of 23 patients with hepatitisB negative chronic active hepatitis.Thirty patients were receiving cortico-steroid therapy, five had peptic ulceration,compared with 19 of the remaining133 patients. The difference was notsignificant.

Fifty-nine patients presented withgastrointestinal bleeding, but in only fourwas this due to peptic ulceration.

In conclusion, peptic ulcerationoccurred in 14-7% of patients withchronic liver disease. It was common inalcoholics and those with hepatitis Bsurface antigen, but rare in hepatitis Bantigen negative chronic active hepatitis.Ulcers were not more frequent in thosereceiving corticosteroid therapy. The

majority of patients with ulcers weresymptomatic, and ulcers were rarelythe cause of gastrointestinal bleeding.

Reference1Tabaqchali, S., and Dawson, A. M. (1964).

Peptic ulcer and gastric secretion in patients withliver disease. Gut, 5, 417-421.

Laser photocoagulation for upper gastro-intestinal haemorrhage

S. G. BOWN, P. R. SALMON, D. KELLY,B. M. CALDER, AND A. E. READ (Depart-ments of Medicine and Pathology,University of Bristol) Much interest cur-rently centres on therapeutic endoscopyfor upper gastrointestinal haemorrhage.The most promising method is laserphotocoagulation, but the safety andefficacy have not yet been fully established.We are currently comparing an Argonwith a Neodymium YAG laser in ananimal model, and here present theArgon results. Our Spectra-Physics 171argon ion laser produces a continuouswave beam of up to 15W. This is trans-mitted to the bleeding site by a 200 micronsingle quartz fibre which is enclosed in a2 mm catheter enabling a coaxial streamof CO2 gas to blow blood away fromthe lesion. This catheter may be passeddown the biopsy channel of a standardendoscope. Our work was performed atopen gastrotomy and endoscopically in10 beagles. Artificial ulcers were createdin the stomach of heparinised animals,and the resulting haemorrhage treatedwith argon laser irradiation of varyingtotal power, power density and duration.The optimum power was 7-9 W, whicharrested haemorrhage in 20 out of 21ulcers. Histological examination showedno thrombosis in acute lesions, whereaschronic experiments showed fibrinoidnecrosis in irradiated vessels with delayedhealing. Most of the energy was absorbedin the mucosa and submucosa.

Direct magnification during ERCP

T. OGURI, N. KUNO, AND T. KASUGAI(introduced by P. R. SALMON) (AichiCancer Center Hospital, Ist department ofInternal Medicine, Tashirocho, Chikusaku,Nagoya, 464 Japan) The branches andfine pancreatic ducts when opacified withcontrast material can be studied moreprecisely after direct fourfold magnifica-tion during ERCP than by routine ERCP.

Routine post-mortem pancreatograms

(PMP) and direct fourfold magnificationduring PMP (PMPM) were comparedin 17 cases. Greater accuracy was ob-tained by the latter method, resulting inan increase of abnormal areas and atwofold increase in the number of branchesof the pancreatic ducts that could beanalysed.ERCP and direct fourfold magnification

during ERCP (ERCPM) were performedin 30 subjects. Twenty-six pairs of ERCPand ERCPM were obtained. These26 cases included four cases of pancreaticcarcinoma, 20 of chronic pancreatitis,and two normal subjects. Good fillingduring ERCPM gave more informationin 22 cases than during ERCP. In foursubjects ERCPMwas of particular value inthediagnosis ofpancreatic disease.ERCPMrevealed focal pancreatitis in two patientsin whom there was a suspicion of pan-creatic carcinoma at ERCP. There wasno advantage of ERCPM in three casesbecause of a large carcinoma of thepancreas, but particular advantages werefound in one case with a small carcinoma.

New method for endoscopic retrogradecholangiopancreatography (ERCP) in casesof unsuccessful cannulation in jaundicedpatients

G. C. CALErrl, L. BOLONDI, G. VERUCCHI,G. LABO (introduced by P. B. COrTON)Istituto di Clinica Medica e Gastro-enterologia dell'Universitai di Bologna.Policlinico S. Orsola Bologna, Italy)In five patients with obstructive jaundiceof unknown origin traditional ERC failed.Using an Olympus JFB3 duodenoscopesmall artificial endoscopic choledocho-duodenal fistulas by means of a speciallydesigned diathermic cutter (needle type)were performed at the lower end of theintramural portion of the common bileduct in order to achieve ERC.ERC through the endoscopic fistula

succeeded with no complications afterthis procedure in four patients.

In all four a precise diagnosis was made.In one patient the depth of the fistulawas not sufficient to reach the CBD lumenand the contrast medium was injectedin the duodenal submucosa with thedevelopment of a mild fever (38 5°C),with recovery after five days'-antibiotictherapy. An attentive exploration of theduodenum during surgery (two patients)or at necropsy (two patients) showed nosign of the endoscopic fistula. In onepatient with normal bile ducts at ERC, an



j.com/



ownloaded from

http://gut.bmj.com/

The British Society of Gastroenterology A957

5

endoscopic control four weeks latershowed a normal mucosa at the same siteof the fistula.

It is concluded that this method is lessinvasive and probably safer than PTCin order to carry out cholangiographywhen ERC fails. Moreover, it may permitan endoscopic choledochoduodenostomyfor choledocholithiasis in cases ofunsuccessful endoscopic papillosphinct-erotomy.

Statistical evaluation of the correlationbetween endoscopic retrograde cholangio-pancreatography and pancreozyminsecretin test

T. OGURI, T. KASUGAI, AND N. KUNO(introduced by P. R. SALMON) (AichiCancer Center Hospital, 1st department ofInternal Medicine, Tashirocho, Chikusaku,Nagova, 464 Japan) Failure of agreementbetween ERCP and a pancreozymin-secretin test (PS test) may be ascribed tovariations in histopathology, differencesin the method of ERCP and functiontesting, accuracy in the performance ofthe procedures, and the interpretationof the obtained data. As a result, correla-tions between ERCP and function testsshould be analysed statistically. In 1972the authors proposed criteria for thegrading of ERCP in the diagnosis ofchronic pancreatitis-that is, minimal,moderate, and advanced stages.

Regressions of total volume, totalamylase output, and maximum bicarbon-ate concentration by PS test were cal-culated against the grade of ERCP andfound to be highly significant (P < 0 001)for all three factors. Regressions of fre-quency of abnormality below mean value(m)-standard deviation (s) and m-2s inthese three factors were highly significant(P < 0-001). The ERCP in those cases inwhich the ERCP grade did not correlatewith PS test showed marked local differ-ences in the grades of pancreatogram.This finding could explain the discrepancyon the assumption that PS test is afunctional mean value, whereas ERCP is astructural maximum value.

'Lassoo' polypectomy

C. B. WILLIAMS AND P. E. GILLESPIE (St.Mark's Hospital, City Road, London)Removal of large and broad-basedcolonic polyps poses a special problemfor the colonoscopist as there is a risk of

bleeding from the site of resection andperforation of the bowel wall. The endo-scopist has had the choice of removing thepolyps 'piecemeal' in several sessions orof referring patients for abdominalsurgery.

Recently we have adapted the methodof sigmoid-rectal intussusception, pre-viously described with the rigid procto-sigmoidoscope.' In our cases, havingstarted conventional snare polypectomyit was considered dangerous to continue;the polyps were 'lassooed' using thehandleless snare wire2 which was left in situtightened onto the base of the polyp,and the colonoscope withdrawn. Undergeneral anaesthesia the polyps wereintussuscepted to the anus and locallyexcised and sutured without difficulty.This intussusception technique should beconsidered in those with broad-basedsigmoid colon polyps judged to behazardous or impossible to remove byconventional polypectomy.

References'Parks, A. G. (1976). Ergebnisse Der Angologie,

p. 338. New York, F. K. Schattauer, Verlag-Stuttgart.

'Shinya, H., and Wolff, W. I. (1975). HospitalPractice, 10, 71.

PAEDIATRICS

Duodenal mucosal antibody against bac-teria grown from duodenal luminal fluidand mucosa in children with diarrhoea

P. D. MANUEL, V. BAMPOE, S. AVIGAD,M. SHINER, AND J. A. WALKER-SMITH.(CRC, Northwick Park, Harrow, Middlesexand the Queen Elizabeth Hospital forChildren, Hackney Road, London) De-layed recovery after acute gastroenteritisof infancy remains a major clinical prob-lem. Its cause is unclear, although somecases are due to intolerance to lactose orcow's milk protein. The role of bacteria,particularly the 'normal' gut flora,remains unknown.We have investigated 29 children under

the age of 2 years, 18 with diarrhoea forat least two weeks before study, and 11without diarrhoea during this period.Bacteria were cultured aerobically andanaerobically from duodenal juice andmucosa taken from these children as partof routine clinical investigation. Usingan indirect immunofluorescence tech-nique1 we have looked for antibody in the

duodenal mucosa to bacteria culturedfrom each patient.

Total bacterial counts were not stat-istically significantly different betweenthe two groups. Antibody was found in17 of 18 patients with diarrhoea, but inonly two of 10 patients without. Thepresence of antibody did not depend onan abnormal mucosa, nor on the conditioncausing the diarrhoea. In a number ofpatients the antibody has been classed asIgA or IgG or both.The role of this antibody is unclear;

it may mark a changed relationshipbetween these 'non-pathogenic' bacteriaand the host's immune system.

Reference'Manteiro, E., Fossey, J., Shiner, M., Drasar, B. S.,and Allison, A. C. (1971). Lancet, 249.

Abnormalities of intestinal transportsystems in the post-enteritis syndrome(PES) and 'toddler' (non-specific) diar-rhoea

J. H. TRIPP, J. A. MANNING, D. P. R. MULLER,A. KILBY, J. A. WALKER-SMITH, AND J. T.HARRIES (Institute of Child Health, TheHospital for Sick Children, Great OrmondStreet, and Queen Elizabeth Hospital forChildren, London) There is now goodevidence that some small intestinalsecretory states are mediated by altera-tions in the activities of (Na+-K+)-ATPaseand adenylate cyclase (AC) in the mucosalenterocyte. Examples include increasedAC activity induced by cholera and certainother toxins and reduced (Na+-K+)-ATP-ase in viral diarrhoea, and we have recentlydemonstrated increased AC and reduced(Na+-K+)-ATPase in active coeliac dis-ease.'We have assayed both enzymes in

biopsies obtained from 19 children withPES during the active phase of the diseaseand from eight during recovery; in fourpatients sequential studies were performedduring and after recovery from PES.Biopsies were also performed in a furthergroup of 23 children with 'toddler'diarrhoea (TD).

In the PES patients both enzymeactivities were reduced (P < 0-05),and returned to normal or supranormallevels on recovery. In contrast the meanactivity of both enzymes was increasedin patients with TD (P < 0-05); six ofthe seven biopsies with raised (Na+-K+)-ATPase activity also had raised ACactivity, suggesting the possibility of alinked compensatory process.



j.com/



ownloaded from

http://gut.bmj.com/


Our results suggest that the reductionin both enzyme activities in PES is notsimply due to enterocyte immaturity.The similarities between enzyme activitiesin TD and the recovery phase of PES mayindicate that the increased (Na+-K+)-ATPase activity is a response of normalvillous cells to crypt cell secretion.

Reference'Tripp, J. H., Manning, J. A., Muller, D. P. R.,Walker-Smith, J. A., O'Donoghue, D. P.,Kumar, P. J., and Harries, J. T. (1978). Mucosaladenylate cyclase and sodium-potassium stimu-lated adenosine triphosphatase in jejunalbiopsies of adults and children with coeliacdisease. In Perspectives in Coeliac Disease,Proceedings of Third Symposium on CoeliacDisease, Galway 1977. Edited by B. McNicholl,C. F. McCarthy, and P. F. Fottrell. MTP Press:Lancaster.

Liver histology in obstructive jaundice ofinfancy

R. NELSON, D. SCOTT, AND A. J. WATSON(Departments of Child Health, and ClinicalPathology, Royal Victoria Infirmary,Newcastle upon Tyne) Liver biopsiesobtained during the first few months oflife, from 58 children with persistentcholestatic jaundice, were studied retro-spectively. Twenty-four histological fea-tures were graded 0 to 3, according tothe severity of the abnormality, by twopathologists, without knowledge of tbeclinical details.

Twenty-six patients had extrahepaticbile duct atresia and 32 had 'neonatalhepatitis'. The patients with 'hepatitis',were also grouped according to theirclinical progress; 10 died of liver failurein infancy, 15 had recovered normalliver function, and seven had chronicliver disease.

Portal tract expansion, bile duct pro-liferation, and cirrhosis were features ofthe liver biopsy in bile duct obstruction,but also in fatal hepatitis. There was nosignificant difference between obstructionand hepatitis in the incidence of giantcell transformation, pseudo-acini forma-tion, the severity and type of inflammatorycell infiltration, or the presence of piece-meal necrosis, and these features were nothelpful in predicting the pi ognosis ofinfants with hepatitis; 92% of patientswith obstruction, but only 61 % with'hepatitis' were correctly diagnosed byexamination of the liver biopsy. 'Hepatitis'patients with poor prognosis were morefrequently misdiagnosed as obstruction.

Percutaneous cholangiography in prolongedjaundice of childhood

EDWARD R. HOWARD AND HEATHER B.NUNNERLEY (King's College Hospital,Denmark Hill, London) The radiologicaldemonstration of undilated intrahepaticbile ducts was made easier by the introduc-tion of the Chiba fine needle (Okudaet al., 1974). The needle has now beenmodified for the investigation of paediatricpatients with prolonged jaundice by re-ducing the external diameter to 0 5 mmand by reducing its length.

Percutaneous transhepatic cholangio-graphy was performed on 22 jaundicedchildren in whom screening tests suggestedsurgically correctable lesions in the biliarytract.

In the preoperative investigation of12 infants with extrahepatic biliaryatresia isolated segments of bile ducts,dilated extrahepatic lymphatics and nor-mal hepatic veins were seen. Postoperativecholangiograms in two patients withatresia were successful in demonstratinga patent bile duct anastomosis in one caseand portal hypertension in the other.

Patent biliary tracts were visualised incases of hepatitis syndrome and hepaticfibrosis and in consequence surgery wasavoided in two infants. Three childrenwith intrahepatic biliary hypoplasia werealso investigated.The anatomy of the extrahepatic ducts

was clearly demonstrated in three olderchildren who had lesions of the commonbile duct.

Percutaneous cholangiography withthe very fine needle has been free ofcomplications in these young patientsand the investigation has contributed toboth the diagnosis and surgical manage-ment of their jaundice.

Effect of serum from patients with pan-creatic insufficiency (PI) on jejunal trans-port in the rat jejunum in vivo.

G. BANCHINI, J. H. TRIPP, E. ROMA, P. J.MILLA, D. P. R. MULLER, AND J. T. HARRIES(The Hospital for Sick Children, GreatOrmond Street, and Institute of ChildHealth, Guilford Street, London) Arakiet al.1 demonstrated that cystic fibrosis(CF) serum inhibited glucose-stimulatedshort circuit current in rat jejunum invitro, and suggested that this was a specificphenomenon reflecting a CF 'factor'.Subsequently we showed that CF seruminhibited glucose transport in rat jejunum

in vivo.2In this study we compare the effects of

serum from children with CF, Shwach-man's syndrome (SS), and protracteddiarrhoea (PD) without PI on jejunaltransport in the rat jejunum in vit'o usinga closed-loop technique. Compared withcontrol serum, CF serum inhibited netabsorption of glucose (p < 0 001),sodium (p < 0 01), and water (p < 0 05).Glucose absorption was not inhibitedby heterozygote serum. SS serum behavedin a similar way to CF serum, inhibitingglucose (p < 0-05) and water (p < 0-05)absorption. In contrast, PD serum had noeffect on glucose absorption.

These results indicate that the inhibitoryeffects of CF serum are not specific, andoccur in other types of PI. We suggest thateither (1) a 'factor' is present in normalserum, and is presumably secreted by thepancreas, which enhances absorption ofglucose, water, and sodium, or (2) aninhibitory 'factor' is present in PI serum.

References'Araki, H., Field, M., and Shwachman, H. (1975).A new assay for cystic fibrosis factor: effects ofsera from patients with cystic fibrosis in thein vitro electrical properties of rat jejunum.Pediatric Research, 9, 932.

'Roma, E., Milla, P. J., Tripp, J. H., and Harries,J. T. (1978). Inhibition of jejunal transport inrat jejunum by cystic fibrosis (CF) serum.Archives of Disease in Childhood, 53, 263.

Diagnosis of childhood Crohn's disease:value of colonoscopy

C. A. CAMPBELL, C. B. WILLIAMS, ANDJ. A. WALKER-SMITH (Department ofGastroenterology and Child Health, St.Bartholomew's Hospital) Crohn's dis-ease is being increasingly diagnosed inthe paediatric age group; we have seen13 cases between the ages of 6 and 15years in this department in the last twoyears. The diagnosis was ultimately madeon radiological or pathological grounds.There was often a prolonged delay

between onset of symptoms and diagnosisbecause of lack of awareness of the diseasein childhood, the vague symptom pattern,and inconclusive initial investigations.1We now perform early total colono-

scopy in suspected cases under sedation,without general anaesthesia, using theadult colonoscope.

Small aphthoid ulcers can easily beseen in the colon and terminal ileum evenat a stage when they are radiologicallyundetectable. Such appearances are re-corded by conventional and polaroidphotographs and videotape; they are



j.com/



ownloaded from

http://gut.bmj.com/


characteristic of Crohn's disease, althoughthe small biopsy specimens infrequentlyshow diagnostic granulomata.

Colonoscopy is a simple, practical,and rewarding procedure as an earlyinvestigation in any child in whom there isa possibility of Crohn's disease.Reference'O'Donoghue, D. P., and Dawson, A. M. (1977)

Crohn's disease in childhood. Archives ofDisease in Childhood, 52, 627-632.

GI HORMONES

Cellular localisation of gastrin in thehuman small intestine

A. M. J. BUCHAN AND J. M. POLAK (Depart-ment of Histochemistry, RPMS, Hammer-smith Hospital, London) Two majorforms of gastrin, little G17 and big G34,are extracted from antral and upper smallintestinal mucosa. The origin of bothpeptides in antral mucosa was establishedas the ultrastructurally defined G cell.G cells contain secretory granules, rangingfrom small (200 nm) and electrondenseto large (400 nm) and electronlucent.Although ultrastructurally identifiableG cells are present in the Brunner'sglands, they are absent in intestinalmucosa. Therefore the cellular originof intestinal gastrin remained undeter-mined.We present here data gathered from the

investigation of 30 samples of duodenumand jejunum on the identification of theintestinal gastrin-producing cell. We usedthree region specific antisera to immuno-cytochemically identify the cells by thesemithin/thin method.

Antisera 1295 (specific for G17) andL33 (specific for G34) stain a totallydistinct endocrine cell population fromthe CCK specific antiserum (Z58). Thegastrin-containing cells have small(190 nm) electrondense secretory granulescompared to the large I cell granules(250 nm) containing CCK. These cellswere probably included in the D1 cellcategory in previous classifications. Wesuggest the name SIG for these smallintestinal gastrin-containing cells in orderto distinguish them from other smallgranulated cells found in the intestine.

Raised motilin in diarrrhoea

S. R. BLOOM, N. D. CHRISTOFIDES, ANDH. S. BESTERMAN (Department of Medi-cine, Royal Postgraduate Medical School,London) Motilin, a newly discoveredhormone from the upper intestine, haspowerful pharmacological actions on gutmotor functions. It greatly acceleratespostprandial gastric emptying' and caninduce the formation of interdigestivemyoelectric complexes.2 In order to deter-mine if it might play a role in the abnormalmotility associated with diarrhoeal states,fasting motilin concentrations were mea-sured in several different groups of patientswith active diarrhoea.Plasma motilin was determined by a

radioimmunoassay which was capableof detecting changes of 3 pmol/l andshowed no cross-reaction with other guthormones. In 16 healthy controls themean fasting level was 56 ± 8 (SEM)pmol/l. In eight patients with diarrhoeadue to severe tropical sprue, the motilinconcentration was 142 + 16 pmol/l, in12 patients with acute infective diarrhoeait was 140 ± 26 pmol/l, in 13 patientswith Crohn's disease it was 202 +55 pmol/l, and in 12 ulcerative colitispatients it was 150 ± 41 pmol/l (allp < 0 01 against controls using Wilcoxonsum of ranks test). In contrast, 11 patientswith constipation had normal levels of55 ± 9 pmol/l. Thus, motilin is sig-nificantly raised in diarrhoea states andmay well play a significant role in theupper gastrointestinal motor changesassociated with diarrhoea.

References'Christofides, N. D., Modlin, I., Fitzpatrick, M. L.,and Bloom, S. R. (1978). Effect of mctilin ongastric emptying of solid meals in man. Gut,19, 436-437.

'Itoh, Z. (1977). Motilin-a hormone to controlinterdigestive gut motility. Journal ofthe JapaneseBiochemical Society, 49 (4).

Vipergic innervation of the human pancreas

A. E. BISHOP, J. M. POLAK, M. G. BRYANT,AND S. R. BLOOM (Departments of Histo-chemistry and Medicine, RPMS, Hammer-smith Hospital, London) Vasoactive in-testinal polypeptide (VIP) has a varietyof effects on the exocrine and endocrinepancreas. These include the stimulationof bicarbonate secretion and alkalinejuice flow, and the release of glucagon,pancreatic polypeptide, and insulin.As VIP is rapidly cleared from the

circulation, it is unlikely that it normallyacts as a circulating hormone. VIP prob-ably functions as a local tissue hormone

or neurotransmitter. This suggests thatthe powerful effects of VIP on thepancreas may be mediated by a sourceof VIP within the pancreas. This possibilitywas investigated by a morphological andquantitative study carried out by thecombined techniques of immunocyto-chemistry and radioimmunoassay.Radioimmunoassay showed that VIP

can be extracted from the human pancreasin amounts averaging 33 ± 9 pmol/g.Immunocytochemistry localised this VIPcontent in fine, varicose nerve fibres.The fibres ran through the connectivetissue and the exocrine pancreas. Theycould also be seen in close contact withthe pancreatic islets.The finding of the presence of VIP can,

therefore, help in the interpretation ofmany physiological and pathologicalresponses of the exocrine and endocrinepancreas. Further functional studies areneeded to demonstrate the exact role of theVIPergic fibres in the pancreas.

'Big' cholecystokinin?: evidence for exist-ence in human serum

C. E. MARSHALL, S. L. HOWELL, AND A. G.JOHNSON (Professorial Department ofSurgery, Charing Cross Hospital MedicalSchool, London) Experiments were per-formed to investigate the possible exist-ence of multiple molecular weight formsof biologically active cholecystokininin human serum. Serum samples werefractionated on Sephadex columns usingKrebs' solution as the eluting buffer.Eluates from the columns were fed directlyinto a modification of the superfusionsystem for bioassay of cholecystokininactivity with rabbit gall-bladder stripsalready described." 2'3 By splitting thestream the analysis was performed induplicate. Columns were calibrated withBoots pancreozymin, octapeptide of chole-cystokinin, and 99% pure cholecysto-kinin. Under these conditions, on applica-tion of sera, two distinct peaks of chole-cystokinin activity were eluted fromG50 columns; the first corresponded toa large molecular weight molecule elutedwith the void volume, the second to33 amino acid cholecystokinin, althoughthese columns did not separate the fullcholecystokinin molecule from its octa-peptide. Analysis of the sera from subjectsfasting and after a standard fatty mealindicates that it is the low molecularweight component that is released aftera stimulus, the high molecular weight

5*



j.com/



ownloaded from

http://gut.bmj.com/


molecule predominating in the fastingstate. These experiments provide evidencefor the existence of a 'big' moleculewith cholecystokinin-like activity inhuman serum in certain conditions.

References'Johnson, A. G., and McDermott, Susan J. (1973).

Sensitive bioassay of cholecystokinin in humanserum. Lancet, 2, 589-591.

'Marshall, C. E. (1976). Automated biological assayof cholecystokinin. Medical and BiologicalEngineering, 14, 327-329.

3Marshall, C. E., Egberts, E. H., and Johnson, A. G.(1978). An improved method for estimatingcholecystokinin in human serum. Journal ofEndocrinology. (In press).

PANCREAS/INFLAMMATORY BOWEL

Clearance of I125 labelled human trypsinfrom the serum: evidence for recirculationin man

G. LAKE-BAKAAR, C. E. RUBIO,S. MCKAVANAGH, AND J. A. SUMMERFIELD.(introduced by PROFESSOR DAME SHEILASHERLOCK) (Department of Medicine,Royal Free Hospital, Pond Street,London) We have recently provided evi-dence that circulating serum trypsin inman is derived mainly from gut reabsorp-tion and suggested that this might repre-sent one axis of an entero-pancreaticcirculation.1 Using 1125 labelled humantrypsin, the possibility that the circulat-ing enzyme may be actively takenup by the pancreas and resecreted hasbeen investigated.Three healthy volunteers received intra-

venous injections of 1125 trypsin and F131albumin. 1125 trypsin showed a biexponen-tial decay in the blood. The rapid initaldisappearance (mean Ti 11-7 min, range8-16 min) was followed by a slower decay(mean Ti 303 min; range 279-350).Duodenal juice was aspirated con-

tinuously and pooled in 15 minute aliquotsduring pancreatic stimulation with sec-retin (1 unit/kg/h). 1125 activity, but not1131, appeared in duodenal juice withinthe first 15 minules and increased to amaximum of up to 14 times the cor-responding concentration in blood. Themean recovery of 1125 counts in duodenalcontents during the four hours of aspira-tion was 26-8% (range 12-3-45-5 %) of thetotal dose administered. The remainderwas recovered mainly in the urine. Lessthan 17% of the F131 albumin administeredwas detected in the duodenal aspirate.

These experiments suggest the presenceof an active uptake process for circulat-ing serum trypsin which is then resecretedinto the duodenum.

Reference'Lake-Bakaar, G., SmithLaing, G., and Summer-

field, J. A., (1978). Gut, 19, A445.

Urine trypsin concentration in the diagnosisof pancreatic disease

G. LAKE-BAKAAR AND J. A. SUMMERFIELD(introduced by PROFESSOR DAME SHEILASHERLOCK) (Department of Medicine,Royal Free Hospital, Pond Street, London)A radioimmunoassay (RIA) for the mea-surement of serum trypsin concentrationhas recently been described.1 In general,high fasting serum levels are found incancer of the pancreas and low levels inchronic pancreatitis, but there is con-siderable overlap between the groups.

In this study the fasting early morningurine trypsin concentration in 38 subjectswere estimated by RIA. Each specimenwas concentrated up to 25 times using aMinicon B 15 concentrator. Urinecreatinine concentration (pmol/ml) wasmeasured by the Jaffe reaction.

In 13 healthy volunteers, a mean urinetrypsin concentration of 1-02 + 3-4 ng/ml(mean ± SEM) was obtained. In threepatients with chronic renal failure andsix with acute pancreatitis, mean urinetrypsin concentrations of 28 + 23 ng/mland 7-5 ± 4-3 ng/ml respectively werefound.

In nine patients with chronic pan-creatitis a mean urine trypsin concentra-tion of 0-3 ± 0-2 ng/ml was obtained.However, in seven patients with cancerof the pancreas, the mean concentrationwas 10-3 ± 4.5 ng/ml. This was signific-antly different from the chronic pan-creatitis group (t = 2-74, p < 0-01).The ratio of the trypsin concentration

to the creatinine concentration in theurine yielded a similar pattern of results.We conclude that in the differential

diagnosis between chronic pancreatitisand cancer of the pancreas, a single urinetrypsin determination in a fasting earlymorning specimen may prove useful.

Reference'Elias, E., et al. (1977). Lancet, 66.

Hypotrypsinaemia in diabetes mellitus

T. E. ADRIAN, A. J. BARNES, AND S. R. BLOOM

(Department of Medicine, Royal Post-graduate Medical School, London) It haspreviously been reported that plasmaconcentration of trypsin-like immuno-reactivity is extremely low in patients withpancreatic exocrine deficiency.1 Sub-clinical impairment of exocrine pancreaticfunction has been reported in severediabetes mellitus.2 However, the pre-valence and degree of such dysfunctionis unknown. We have therefore measuredplasma trypsin by radioimmunoassay(Hoechst UK Ltd) in 106 non-diabetics,163 non-insulin-dependent diabetics, and236 insulin-dependent diabetics. In 10patients who had undergone total pan-createctomy trypsin was undetectable.In 106 non-diabetic subjects plasmatrypsin was 12-3 ± 0 9 (mean ± SEM),whereas levels were markedly reducedin both insulin-dependent diabetics 6-0 +0-3 (p < 0 001) and in the tablet and diet-treated patients 7-7 ± 0 4 (p < 0-005).

Within the insulin-dependent group,plasma trypsin was significantly lowerwith increasing insulin requirements(<40 U/day 7 9 + 0-6, >80 U/day4-5 ± 04, p < 0-001). Plasma trypsinlevels were lower in younger diabetics(<25 years, 2-9 + 0 4; 25-50 years.5-8 ± 0-4; and >50 years, 7-8 ± 0-7,p < 0 001). No differences were observedin plasma concentrations of this enzymebetween patients with and withoutdiabetic complications and no correlationwas seen with time on treatment withinsulin. The gross reduction in plasmatrypsin fits well with the previouslyreported acinar cell failure which appearsto be associated with mild as well as severediabetes.

References'Adrian, T. E., Besterman, H. S., Mallinson, C. N.,

Galarotis, C., and Bloom, S. R. (1978). Plasmatrypsin in patients with steatorrhoea due tochronic pancreatitis. Clinical Science andMolecular Medicine, 54, 24.

2Frier, B. M., Saunders, J. H., and Wormsley,K. G. (1976). Exocrine pancreatic function injjvenile-onset diahetes mellitus. Gut, 17, 685-691.

Parathyroid hormone (PTH) response inpatients with very severe acute pancreatitis

C. W. IMRIE, G. H. BEASTALL, B. F. ALLAM,J. O'NEILL, I. S. BENJAMIN, AND A. J.MCKAY (Departments of Surgery, Bio-chemistry and Regional RadioimmunoassayCentre, Royal Infirmary, Glasgow) It hasbeen suggested that an inadequate PTHresponse characterises severe acute pan-creatitis and is especially associated with



j.com/



ownloaded from

http://gut.bmj.com/


hypocalcaemia.1'2'3We have investigated 21 patients with

very severe acute pancreatitis (at leastfour objective prognostic factors present).Sequential monitoring of plasma PTHlevels was performed during the firstweek of hospitalisation and correlatedwith hypocalcaemia and outcome ofdisease. All eight patients with severehypocalcaemia (corrected calcium <2-0mmol/l) had associated rises of plasmaPTH levels (normal up to 600 ng/l).The mean of the maximal PTH in thesepatients was 1183 ng/l (range 660-1700)and these high levels were usually recordedwithin the initial 36 hours. Thereafter agradual decrease in plasma PTH occurred,accompanied by a restoration of serumcorrected calcium within the normal range(2.20-2-60 mmol/l).The mortality in this group of patients

was 33 %, and six of the seven who diedexhibited significant rises of PTH, theexception being a death associated withrecurrent myocardial infarction and nor-mocalcaemia. This study indicates anadequate PTH response to be present inthe most severe pancreatitis.

References'Condon, J. R., et al. (1975). British Journal of

Surgery, 62, 115-118.'McMahon, M. J., et al. (1978). British Journal of

Surgery, 65, 216-218.'Robertson, G. M. Jr., et al. (1976). New EnglandJournal of Medicine, 294, 512-516.

Screening for pancreatic disease: com-parison of grey-scale ultrasonography andisotope scanning

J. M. RHODES, J. A. AGNEW, J. A. SUMMER-FIELD, E. ELIAS, R. A. HORROCKS, P. M.

CHUDLEIGH, AND L. A. BERGER (RoyalFree Hospital, Pond Street, London)There is still no clearly established screen-ing test for pancreatic disease. 75Seleno-methionine isotope scanning generallygives few false-negatives but many false-positives.' Grey-scale ultrasonography,while promising,2 has not yet been ade-quately compared with isotope scanning.Both isotope and ultrasound scan

wereperformed prospectively in 40 patientswith suspected pancreatic disease. Thediagnosis was established by laparotomy(23), endoscopic pancreatography (nine)or clinical follow-up for at least ninemonths (eight). All 13 patients withpancreatic carcinoma had abnormal iso-tope scans and only one had a normalultrasound scan. The eight patients withchronic pancreatitis all had abnormal

ultrasound and isotope scans. However,out of 19 normal patients, 12 had abnormalisotope scans compared with only threefalse-positive ultrasound scans.A review of isotope scanning in the

Royal Free Hospital since 1973 (n = 314)and ultrasound scanning during 1977(n = 115) yielded similar results. In pan-creatic carcinoma ultrasound was falsely-normal in 1/22 (5%), isotope scanningfalsely-normal in 2/92 (2 %). In chronicpancreatitis the false-negative rate forultrasound was 1/15 (7%), for isotope10/88 (11 %). However, in normal patients,ultrasound yielded only 12/78 (16 %) false-positive scans in contrast to isotope scan-ning with 64/134 (48 %) false-positives.

In conclusion, ultrasound scanning ofthe pancreas has a false-negative ratesimilar to isotope scanning but a muchlower false-positive rate. It should now bethe screening test of choice for pancreaticdisease.

References'Agnew, J. E., and Maze, M. (1976). Pancreatic

scanning. British Journal of Radiology, 49,979-995.

"Lees, W. R., Vallon, A. G., Denyer, M. G.,Vehl, S. P., and Cotton, P. B. (1978). Prospectivestudy of pancreatic ultrasonography. Gut, 19,A499.

Postprandial secretin stimulates pancreaticbicarbonate secretion in man

G. R. GREENBERG, S. DOMSCHKE, M. G.BRYANT, W. DOMSCHKE, W. ROSCH, ANDS. R. BLOOM (Royal Postgraduate MedicalSchool, London, and University ofErlangen,Niirnberg, Erlangen, Germany) We haverecently reported that in man a rise inplasma secretin of 3-5 pmol/l is observedduring both endogenous acid stimulation(pentagastrin) and after a meal. Whethersuch low levels are sufficient to stimulatepancreatic bicarbonate has not, however,been determined. Pure pancreatic juicewas collected after direct cannulation ofthe main pancreatic duct (n = 6) duringintravenous GIH secretin (003 CUkg-' h-1). Because potentiation may occurwith cholecystokinin, the addition ofcaerulein (15 ng kg-' h-') was also studied.During secretin, bicarbonate output

increased from basal levels of 27 ± 12,uEq/5 min (mean ± SEM) to 182 ± 24(P < 0 001). Addition of caerulein resultedin a further significant increase to 396 ± 50.Plasma secretin increased to 2 1 pmol/lduring the infusion and was not affectedby caerulein. The similarity of porcineto human secretin was suggested by gel

chromatography studies of postprandialhuman plasma, where one peak elutedin an identical position to both porcinesecretin and to human duodenal tissue.

These studies demonstrate that plasmalevels of secretin after a meal, thoughsmall, are sufficient to stimulate pancreaticbicarbonate and support the hypothesisthat secretin has a physiological role inman.

0-14D-O breath test of the fat digestivecapacity of the pancreas

Y. GHOOS, G. VANTRAPPEN, P. RUTGEERTS,AND P. SCHURMANS (Department ofMedical Research, University of Leuven,B-3000 Leuven, Belgium) The rationaleof this new test is the fact that hydrolysisof the triglyceride O-14D-O (1,3 dioleyl,2-14C decanoyl glycerol, synthesised byG. Koch, Unilever) results in a labelledmedium chain monoglyceride or fattyacid which is readily absorbed andmetabolised to 14CO2. 5 ,uCi 0-14D-Ois taken with lunch and breath CO2 istrapped in 2 mmol hyamine hydroxidebefore and 1, 2, 3, 4, 5, 6, 8, 10, 20 and 24 hafter lunch. The amount of 14CO2 isexpressed as the 10 h cumulative per-centage of the administered dose. Twenty-five normal subjects, 12 proven chronicpancreatitis, eight pancreatectomy, 12gluten enteropathy, and eight bile aciddeficient patients were studied. Results:(1) All pancreatic patients, even thosewithout steatorrhoea, had 14CO2 excre-tions (6 %-50%) well below the normalrange (544 %-90 5 %); (2) pancreaticenzyme treatment increased 14CO2 excre-tion in eight of 10 pancreatic patientsstudied (mean increase: 23%); (3) the14C02 excretion in patients with steator-rhoea due to bile acid deficiency was 25%higher than in patients with pancreaticsteatorrhoea; (4) 14CO2 excretion waslow in seven of 12 gluten enteropathies.Submaximal CCK stimulation normalisedthe test in the two patients studied. Thedata suggest that the O-14D-O breath testis a sensitive and specific test of fatdigestive capacity of the pancreas.

Physiological plasma levels of pancreaticpolypeptide (PP) inhibit stimulated pan-creatic and biliary secretion in man

G. R. GREENBERG, R. F. MCCLOY, T. E.ADRIAN, V. S. CHADWICK, J. H. BARON,AND S. R. BLOOM (Royal Postgraduate



j.com/



ownloaded from

http://gut.bmj.com/


Medical School, London) A physiologicalrole for PP in man is not established,yet mean concentrations in plasma of250 pmol/l are observed after a meal.Although pharmacological doses of PPinhibit pancreatic secretion in the dog, itis not known in man whether a similareffect occurs at these post-prandial plasmaconcentrations. Therefore, using aduodenal perfusion technique to measurepancreatic and biliary outputs, we infusedbovine PP (60 pmol kg-1 h-1) in sevenhealthy subjects during stimulation withGIH secretin (Od CU kg-1 h-1) andcaerulein (10 ng kg-1 h-1) closely toapproximate physiological conditions.Mean stimulated trypsin output of

20-4 + SEM 013 KU h-1 was reduced to8-5 + 0-2 (p < 0-001) during PP. Asignificant 80% reduction in bilirubinoutput was also observed. After PPboth parameters recovered to preinfusionlevels. In contrast, mean bicarbonateoutput of 20 + 4 mmol h-1 was unaffectedby PP. Incremental plasma PP levels of250 ± 23 pmol/l during the infusion areequivalent to those after a meal.We conclude that, in man, PP inhibits

stimulated trypsin and bilirubin outptitsat plasma concentrations which areobserved physiologically. These effects,which are in direct opposition to actionsof cholecystokinin, suggest PP may havea role in the regulation of post-prandialpancreatic enzyme secretion and gallbladder storage.

Peroperative transduodenal pancreaticbiopsy

D. E. E. TWEEDLE (University Hospital ofSouth Manchester) Peroperative pan-creatic biopsy can provide valuable infor-mation but has been associated withmajor complications (particularly fistulae)and a mortality of 3-8%1. Peroperativetransduodenal biopsy has been attemptedin 65 patients since 1972 using the dis-posable Tru-cut needle (Baxter-Travenol)and pancreatic tissue was obtained in 62.

In 28 cases the preoperative diagnosisof malignancy was confirmed by biopsyin 23 and, of the remaining five patients,three had malignancy. In six cases thepreoperative diagnosis of chronic pan-creatitis was confirmed by biopsy in fiveand there was one unsuspected malignancyrevealed by biopsy. In 28 cases with apreoperative diagnosis of acute or acuterelapsing pancreatitis, inflammation wasobserved in five, normal tissue in 22,

and unsuspected malignancy in one case.The procedure produced a subserosal

haematoma in three patients but thisdid not complicate the postoperativecourse. Acute pancreatitis, fistulae, pseudo-cyst, and subphrenic abscess did notoccur.

Peroperative transduodenal biopsy is arelatively safe method of providing ahistological diagnosis in disease of thehead of the pancreas, and may revealunsuspected malignancy.Reference'Shulz and Saunders (1963). Annals of Surgery,

158, 1053.

Cancer of the pancreas and extrahepaticbiliary apparatus-a study of its behaviourin 150 patients

R. M. RAINSBURY, LORD SMITH, AND J-CGAZET (Gastrointestinal Unit, St. George'sHospital, London) A consecutive seriesof 150 patients with carcinoma of thepancreas or extrahepatic biliary apparatushave been reviewed and reclassifiedaccording to a proposed (S)TNM(P)classification based on surgical explora-tion. Results of surgery have been relatedto multiple parameters, including present-ing symptoms and time, sex, age, delay indiagnosis, type of surgery, and subsequentcare with (S)TNM(P). Whereas there wasa close correlation in all tumours betweentumour size and survival, there was noconsistent relationship between tumourdifferentiation, or nodal involvementand survival in any group. But the idealtumour-that is TNM(P) 1001 or 1002--did not necessarily do best, as four outof five survived less than 300 days(182, 203, 206, 281, and 1234).But radical resection in carefully

selected patients was the most effectivetreatment and biliary bypass alone wasassociated with longer survival than whencombined with gut bypass or when gutbypass alone was performed. However, inall other groups and in all tumour sites,those treated subsequently with chemo-therapy (59) survived longer with a meanof 359 days, than those not (29) whosurvived an average of 286 days.

Aspiration cytology of the pancreas atERCP

R. A. MOUNTFORD, P. R. SALMON, J. LEVER,P. BROWN, S. G. BOWN, AND A. E. READ(University Departments of Medicine and

Pathology, Bristol Royal Infirmary, Bristol)Patients with suspected pancreatic car-cinoma, supported by ERCP, weresubjected to percutaneous aspirationcytology using a Chiba needle. Lesionsof the head were approached using aposterior approach, lesions of the bodyfrom the front. Three types of pancreato-graphic appearance were selected assuitable:

1. Minor narrowing of the mainpancreatic duct. In this circumstance thecatheter was maintained in position,with the patient prone, so that contrastwas retained in the duct system. Fluoro-scopy was employed to guide the needlealong the axis of the x-ray beam.Depth of insertion was assessed by:

(a) preliminary CT scanning; (b) simul-taneous real time ultrasound.

2. Major narrowing of the pancreaticduct. In these cases contrast remainedwithin the duct, allowing the duodeno-scope to be removed, so that the needletip could be positioned by parallax.

3. Complete occlusion of the juxta-papillary pancreatic duct. The biliarytree was opacified and aspiration per-formed medial to the common bile duct,again rotating the patient.

In 15 attempted aspirations, pancreaticcells were identified in 11 cases (73%),and carcinoma confirmed in four cases.No complications were encountered.The technique provides a tissue diag-

nosis, and illustrates the importance of amulti-disciplinary approach to the diag-nosis of pancreatic cancer.

Gastroduodenal lesions in Crohn's disease

G. WYNDHAM STEVENSON (Department ofPathology, McMaster University MedicalCentre, Ontario, Canada) Results ofdouble contrast barium meal from 87patients with ileal and/or colonic Crohn'sdisease, and 88 patients without Crohn'sdisease have been reviewed. Superficialgastric and duodenal erosions, thickenedand abnormal antral and duodenal folds,and duodenal strictures were found morefrequently in patients with Crohn'sdisease but occurred in those without.Certain combinations of these abnor-malities, however, were strongly suggestiveof Crohn's disease, particularly in theyounger age group. Endoscopy confirmedthe validity of the radiological diagnosesof superficial gastric and duodenalerosions, and showed that, where thick-ened abnormal antral and duodenal



j.com/



ownloaded from

http://gut.bmj.com/


folds were present on radiographs,additional superficial erosions werefrequently seen at endoscopy. Whereerosions were detected endoscopically,they were demonstrated on radiographsin 50% of the patients. Endoscopicbiopsies were taken in five patients withboth gastric and duodenal erosions, andhistological lesions of Crohn's diseasefound in four. Double contrast bariummeals showed combinations of abnor-malities strongly suggestive of gastro-duodenal Crohn's disease in 18 patients(21 %). The radiological and endoscopicfeatures of gastroduodenal Crohn's dis-ease will be illustrated.

Lymphocytotoxic antibodies in Crohn'sdisease

A. S. PENA. I. T. WETERMAN, G. KUIPER,M. C. CASTELLI, A. VAN LEEUWEN, ANDJ. J. VAN ROOD (Department of Gastro-enterology and Department of Immuno-haematology, Leiden University MedicalCenter, The Netherlands) We have in-vestigated the presence of lymphocyto-toxic antibodies in the sera of 119 un-related patients with Crohn's diseasewith two different techniques, the micro-cytotoxicity assay of Ting et al. (1973)1and the two colour fluorescence tech-nique (TCF) of van Rood et al. 19762.As controls we have screened with thesame techniques 119 sera from healthydonors matched for sex and age. Patientsand controls were screened against apanel of at least 29 cells with differentand well defined HLA specificities.We have found that the TCF technique

is much more sensitive to detect lympho-cytotoxic antibodies than Ting et al.'stechnique.

In the TCF 84% of patients withCrohn's disease had antibodies directedagainst B cells as opposed to 6% of thehealthy controls. Sixty-six per cent ofCrohn's patients had antibodies againstT cells as opposed to 3% of the healthycontrols. Several of the positive patientshad not received previous blood trans-fusions and/or had been pregnant. Mostantibodies were not directed against well-defined HLA specificities.These findings confirm and extend the

findings of Korsmeyer et al.3 using adifferent technique and a different patientpopulation.

References'Ting, A., Hasegava, T., Ferrone, S., and Reisfeld.

R. A. (1973). Presensitization detected by sensitive

crossmatch tests. Transplantation Proceedings,5, 813-817.

'Van Rood, J. J., van Leeuwen, A., and Ploem, J. S.(1976). Simultaneous detection of two cellpopulations by two colour fluorescence andapplication to the recognition of B-cell deter-minants. Nature, 262, 195-197.

'Korsmeyer, S., Strickland, R. G., Wilson, I. D.,and Williams, R. C. (1974). Serum lympho-cytotoxic and lymphocytophilic antibody activityin inflammatory bowel disease. Gastroenterology,67, 578-583.

Lysosomal enzymes in monocytes and insera of patients with inflammatorybowel disease

A. S. MEE AND D. P. JEWELL (AcademicDepartment of Medicine, Royal FreeHospital, London) Granulomata arelargely composed of macrophages derivedfrom circulating monocytes. The activityof peripheral blood monocytes has there-fore been assessed in 21 patients withCrohn's disease (CD), 19 patients withulcerative colitis (UC), and 21 healthycontrols by assaying the activity of thelysosomal enzyme N-acetyl-P-D-glucos-aminidase. Monocytes were isolated fromheparinised blood by density gradientcentrifugation followed by glass adher-ence, resulting in 80-90% viable mono-cytes. Monocyte enzyme activity (± SEM)in CD patients was 4-54 ± 0-28 n mol/104cells/h, 4 99 + 0-46 in UC patients and3 09 ± 0-09 in the normal controls(CD vN,P < 0O01;UCvN,P < 0-01).There was a positive correlation betweenenzyme activity and disease severity forpatients with UC (P < 0 01) and a similartrend was seen for patients with CD,although this did not reach statisticalsignificance.Enzyme activity was also measured in

sera from 34 patients with CD, 27 patientswith UC, and 39 controls. Serum enzymeactivity was significantly increased inpatients with both diseases (P < 0-01)but there was no correlation with diseaseseverity.No difference was observed in either

monocyte or serum enzyme activitybetween patients with CD and those withUC.The results show that monocytes are

activated in inflammatory bowel disease.Raised serum levels suggest that therelease of lysosomal enzymes from cells,including monocytes, may contribute totissue damage in both diseases.

Controlled trial of disodium cromoglycateas maintenance therapy for ulcerativecolitis

C. P. WILLOUGHBY, M. F. HEYWORTH,J. PIRIS, AND S. C. TRUELOVE (NuffieldDepartment of Clinical Medicine, Rad-cliffe Infirmary, and Department ofMorbidAnatomy, Radcliffe Infirmary, Oxford)There is evidence that disodium cromo-glycate (DSCG) is of some value in thetreatment of ulcerative colitis.1,2 Thepresent study was designed to compareits efficacy in the prevention of relapsewith that of sulphasalazine (SASP) andwith that of the two agents used in com-bination. The trial period lasted for sixmonths and doses used were 800 mgDSCG daily (the manufacturers' recom-mended dose) and 20 g SASP daily.One hundred and twenty patients with

ulcerative colitis were admitted to thetrial and assigned at random to the threetypes of treatment. A relapse was definedas the recurrence of colitic symptomsaccompanied by sigmoidoscopic andbiopsy evidence of inflammation. Theinterim results show that the patients onDSCG were four times as likely to suffera relapse as those on SASP or on thetwo treatments combined.We conclude that, In the dose used,

DSCG is an ineffective maintenancetreatment for ulcerative colitis and thatno benefit is obtained by combining itwith a standard dose of SASP.

References'Heatley, R. V., Calcraft, B. J., Rhodes, J., Owen,

E., and Evans, B. K. (1975). Disodium cromo-glycate in the treatment of chronic proctitis.Gut, 16, 559-563.

"Mani, V., Green, F. H. Y., Lloyd, G., Fox, H.,and Turnberg, L. A. (1976). Lancet, 1, 439-441.

Doubie-blind controlled trial of disodiumcromoglycate (DSCG) in the treatment ofCrohn's disease

M. J. GRUNDMAN, S. E. WILLIAMS, ANDL. A. TURNBERG (Department ofMedicine,Hope Hospital (University of ManchesterSchool of Medicine), Salford) Disodiumcromoglycate has been shown to be ofsome value in the treatment of ulcerativecolitis' and proctitis2 in limited numbersof cases. The present trial was performedto assess the value of this treatment inCrohn's disease.

Twenty-three ambulant outpatientswere randomly allocated for six monthstreatment with DSCG (1200 mg/day)or placebo in addition to their regulartherapy. Patients were crossed over to



j.com/



ownloaded from

http://gut.bmj.com/


the alternative treatment after six months.Monthly assessments were made of symp-toms, signs, haematological and bio-chemical tests. Sigmoidoscopy and rectalbiopsies were performed three-monthly.All data were analysed by computerfor comparison of individual indicators ofactivity and of a calculated Crohn'sDisease Activity Index.3 Six patientsfailed to complete the trial for a varietyof reasons. No significant side-effectswere noted. No difference was foundbetween active treatment and placeboperiods for the Activity Index or for anysymptom, sign, or laboratory result.Neither the site nor the extent of thedisease had any influence on thesefindings.We conclude that DSCG, given for

six months, is of no benefit in the treat-ment of patients with Crohn's disease.

References'Mani, V., Lloyd, G., Green, F. H. Y., Fox, H.,and Turnberg, L. A. (1976). Lancet, 1, 439.

2Heatley, R. V., Calcroft, B. J., Rhodes, J., Owen,E., and Evans, B. K. (1975). Gut, 16, 559.

3Best, W. R., Becktel, J. M., Singleton, J. W., andKern, F. (1976). Gastroenterology, 70, 439.

LIVER AND BILIARY/CLINICAL DIAGNOSTIC

Is raised plasma oestrone in cirrhotic menjust an incidental finding?

J. R. B. GREEN, H. L. GOBLE, C. R. W.EDWARDS, A. M. DAWSON (Departmentof Gastroenterology, St. Bartholomew'sHospital, London) Failure by the dis-eased liver to inactivate endogenouslyproduced oestrogens is the traditionalexplanation for the feminising changesseen in some cirrhotic men. Recentstudies in these men, however, have shownincreased plasma concentrations of oestr-one rather than of the more biologicallypotent oestradiol. As oestrone is con-sidered a weak oestrogen with littlebiological activity, the significance, if any,of increased plasma oestrone in feminisedcirrhotic men is not clear.To assess possible pathogenetic sig-

nificance of increased plasma oestrone incirrhotic men, oestrone was infused intoeight men without hepatic disease toproduce plasma oestrone concentrationssimilar to those found in cirrhotic menwithout altering other plasma oestrogens.In controlled experiments, gonadotrophic-releasing hormone (Gn-RH) was given

both by bolus injection and separatelyby slow infusion. The effect of high plasmaoestrone concentrations on the stimulatedpituitary release of gonadotrophins wasmeasured.The results of the study show that high

plasma oestrone concentrations signific-antly impair pituitary responsiveness to aslow infusion of Gn-RH, although notto a single intravenous bolus.

This study clearly shows for the firsttime that raised plasma oestrone incirrhotic men could well exert importantpathogenetic effects and it suggests thatoestrone may not be just an inert metabol-ite of little significance in these patients.

Liver lipid and liver function after smallintestinal resection and two types ofjejunoileal bypass in rats

R. MCGOURAN, K. R. P. RUTTER, L. ANG,H. CLEEVE, AND J. D. MAXWELL (St.George's Hospital Medical School, London)Fatty liver and hepatocellular failure,the most serious complications of in-testinal bypass surgery, have been attri-buted to intestinal bacterial overgrowthand consequent bacterial metabolismof ingested nutrients to hepatotoxicmetabolites.We therefore compared liver function

and lipid concentrations in rats with:(1) 90% small intestinal resection (groupR); (2) 90% jejunoileal bypass (group B);(3) 90% jejunoileal bypass, with theexcluded segment disconnected fromfunctioning gut and exteriorised (group E).

All groups lost 20% of body weightafter two weeks. Thereafter (R) increasedto 90% of initial weight. However B and Efailed to regain, and by 10 weeks were65-70% initial weight (P < 0 01).

Liver function in R did not differfrom controls, but B and E had threefoldrises in plasma alanine-aminotransferase(P < 0-02) and aspartate-aminotransferase(P < 0-05), and lower plasma albumin.Hepatic triglyceride increased in all threegroups (control: 7 0 + SD 1-6 mg/g;R: 14-3 ± SD 10-3 mg/g, P < 0 01;B: 15-0 + SD 8-8 mg/g, P < 0-01;E: 34 9 ± SD 16-0, P < 0 001). Hepaticcholesterol was unchanged.Thus hepatic dysfunction associated

with a long bypassed segment of smallbowel is (1) independent of hepatic lipidaccumulation, and (2) occurs whetheror not the segment is in continuity withfunctioning gut.

These findings do not support the view

that intestinal bacterial metabolism ofingested nutrients contributes to hepaticdisease after jejunoileal bypass surgery.

Cell-mediated immunity to a liver specificantigen in patients with cystic fibrosis andliver disease

G. MIELI, H. T. PSACHARAPOULOS,A. NICHOLSON, A. L. W. F. EDDLESTON,A. P. MOWAT, AND R. WILLIAMS (LiverUnit and Department of Child Health,King's College Hospital and MedicalSchool, Denmark Hill, London) Im-proved management of pulmonary andpancreatic complications of cystic fibrosisallows 80% of children to survive to adultlife, but in 10% significant liver diseasethen develops. We have studied therelationship between cell-mediatedimmune responses to liver antigens andthe development of liver damage.

Inhibition of leucocyte migration bypurified liver specific lipoprotein (LSP),derived from hepatocyte plasma mem-brane, was shown in nine of 11 childrenwith cystic fibrosis and liver diseasebut in only five of 14 with cystic fibrosisand no overt liver disease (P > 0-025).Of 12 normal children only one had aslightly reduced migration index.Lymphocyte cytotoxicity to isolated

rabbit hepatocytes was significantly in-creased in 10 of 13 children with cysticfibrosis and liver disease, but in onlysix of 29 cases of cystic fibrosis withoutliver disease (P > 0 001). Experimentsusing lymphocyte subpopulations showedthat the cytotoxicity was mediated by anon-T cell population and could beblocked with LSP in seven of 10 cases,suggesting that the reaction in thesepatients was specifically directed againstLSP.The study demonstrates an association

between potentially damaging immuneresponses to liver antigens and the pre-sence of clinically evident liver disease inpatients with cystic fibrosis.

Functional activity of the alternativecomplement pathway in chronic liverdisease

L. E. MUNOZ, K. TSITSIKAS, H. C. THOMAS,AND S. SHERLOCK (Department ofMedic-ine, The Royal Free Hospital, London)Immune complexes (IC) are associatedwith increased catabolism in patientswith HBsAg positive chronic active



j.com/



ownloaded from

http://gut.bmj.com/


liver disease (CALD) and primarybiliary cirrhosis (PBC), but not in HBsAgnegative CALD and alcohol-inducedliver disease (ALD).To determine whether failure of IC

to activate C3 is due to diminished alter-nate pathway amplification function, weexamined the percentage conversion of C3to C3b by a standardised activant of thispathway (saccharomyces cerevisae yeast)and the ability of these sera to supportphagocytosis of this yeast by normalpolymorphs (alternative pathway depen-dent function).'The percentage conversion of C3 to

C3b in 12 normal subjects was 50 5% ±11 8 (M ± SD). Normal values wereobtained in PBC (48 % ± 9-4) and HBsAgpositive CALD (45-1% ± 17-8) butsignificantly reduced levels occurred inHBsAg negative CALD (36% ± 7)(p < 0-001) and ALD (31% ± 96)(P < 0-01).The opsonic capacity of these sera was

increased or normal in PBC, HBsAgpositive and negative CALD and de-creased in ALD.To summarise: the functional activity

of the alternate pathway of complementis reduced in ALD and to a lesser extentin HBsAg negative CALD. It is normalin HBsAg positive CALD and PBC.This reduced activity may explain why inHBsAg -ve CALD and ALD IC do notresult in increased activities of C3, thebiologically important component ofcomplement.

Reference'Yamamura, M. and Valdimarsson, H. (1978).Immunology, 34, 689-694.

Chronic active hepatitis in the UnitedKingdom and Iraq

G. HOLDSTOCK, S. RASSAM, J. G. C. KING-HAM, II. M. SADLER, AND R. WRIGHT

(Southampton University Hospitals andDepartment of Medicine, Medical City,Iraq) The clinical, immunological, andbiochemical features of biopsy-documented chronic active hepatitis havebeen compared in 26 patients from theUnited Kingdom and 40 patients fromIraq. Immunological studies were under-taken in one laboratory (Southampton)and included HBV markers, autoanti-bodies and DNA-binding antibodies.United Kingdom patients had a lowincidence of HBsAg (8 %) compared withIraquis (83 %). They were more likely tobe female and have an onset with jaundice.

They had a high incidence of other putat-ive autoimmune diseases and autoanti-bodies, with the exception of DNA-binding antibodies, and higher globulinlevels. Anticore antibodies were onlyrarely detected in HBsAg-negative patientsfrom both countries (less than 4%).

Iraqui patients were more likely topresent with fluid retention and abdominalpain and had lower albumin levels andless pronounced rises in the serum trans-aminases. Only one patient from Iraqhad clinical and immunological featuresof a 'lupdid' hepatitis.

This suggests that HBsAg positive andnegative chronic active hepatitis areaetiologically distinct and that there is astriking geographical variation in theirprevalence.

Hepatitis Bs antibody in chronic liverdisease

P. R. MILLS, T. H. PENNINGTON, R. N. M.MACSWEEN, AND G. WATKINSON (Depart-ments ofMedicine, Virology andPathology,Western Infirmary, Glasgow) The hepatitisB surface antibody (anti-HBs) is foundcommonly in sera from healthy com-munities as an indicator of previous sub-clinical hepatitis infection. We haveconducted a prospective survey of anti-body carrier rate among 160 patientswith histologically-proven chronic liverdisease and 28 chronic alcoholics withoutliver disease. Anti HBs was assayed byAbbot Ausab radioimmunoassay kit onicoded sera and recorded as positive ornegative. All patients sera were HBsAgnegative.

There was a striking increase in theantibody carrier rate in patients withalcoholic cirrhosis (7/28-25 %) (P < 0-05)and alcoholic cirrhosis with portal hyper-tension (12/23-52%) (P < 0-001) whencompared with a control population of161 hospital patients without liver disease(14/161-8-7%) who were age and sex-matched with the patients with alcoholiccirrhosis. Chronic alcoholics withoutliver disease fell within the normal range(3/28-11 %). Patients with cryptogeniccirrhosis, PBC, and CAH, did not differfrom the control population. Liverbiopsies examined by orcein stain andelectron-microscopy showed no evidenceof intrahepatic viral particles.

These findings are suggestive of anassociation between previous hepatitisB infection and the development ofchronic alcoholic liver disease. This

association requires confirmation andfurther investigation.

Orcein positive copper associated proteinin childhood liver disease

J. EVANS, P. J. SCHEUER, B. ARCHER,S. P. NEWMAN, AND S. SHERLOCK (Depart-ments of Medicine, Histopathology, andMedical Physics, Royal Free Hospital,London) An orcein positive copper asso-ciated protein (CAP) has been shown inadult hepatocytes and is believed to be anabnormal copper protein'. CAP has notbeen described in children.

Liver sections from six neonates(including two foetuses) without liverdisease and four groups of age-matchedchildren, n = 38 (normal subjects, non-cirrhotic liver disease, cirrhosis, intra-hepatic cholestasis of childhood (IHCC))were semi-quantitatively assessed forCAP and copper by staining with orceinand rhodanine respectively. Liver copperconcentration was measured by neutronactivation analysis.

In eight normal children with normalliver copper concentration, CAP wasabsent. Liver copper levels were increasedin six of six neonates (physiological rise)and 12 of 14 cases of IHCC. Orceinpositive material was found in liver cellsof all six neonates, in the 12 patients withIHCC and raised hepatic copperconcentrations, and one cirrhotic patientwith normal hepatic copper concentration.When hepatic copper concentration

exceeded 3-9 ,amol/g in neonates andIHCC, rhodanine and orcein stains wereusually moderately positive. The resultslend further support to a relationshipbetween CAP and copper and the positiveresults in normal neonates support theview that CAP is a physiological substanceand not a consequence of biliary obstruc-tion.

Reference'Salaspuro, M., and Sipponen, P. (1976). Demon-

stration of an intracellular copper-bindingprotein by orcein staining in long standingcholestatic liver disease. Gut 17, 787-790.

Serum CEA-another test of liver function?

K. R. HINE, S. N. BOOTH, AND P. W. DYKES

(Department ofImmunology, University ofBirmingham) Serum carcinoembryonicantigen (CEA) is of little value in theearly diagnosis of alimentary cancer.However, none of the extensive clinical



j.com/



ownloaded from

http://gut.bmj.com/

A966 The British Society of Gastroenterologyreports has tried to assess its value in aprospective series of undiagnosed patients.Three-hundred-and-eighty-one patients infour diagnostic categories have beeninvestigated by standard techniques andby estimation of CEA. In patients withupper gastrointestinal symptoms, lowergastrointestinal symptoms, and iron de-ficiency anaemia, CEA was of limitedprospective value in the diagnosis ofmalignancy. However, in the group of 74patients with abnormal liver functiontests and hepatomegaly, 38 had a finaldiagnosis of malignant disease and innearly all cases (90%) the CEA level wasraised, compared with 53% of those withbenign conditions. All values in patientswithout neoplasia were below 45 ng/ml,whereas, in 43% of those with malignancy,this value was exceeded. This information,in a number of instances, was not obtain-able in any other way. CEA was markedlysuperior to other laboratory tests(alkaline phosphatase, ESR, 5-nucleo-tidase) in discriminating between the twogroups.

It would appear that CEA might be ofmore use in evaluating this type of clinicalproblem that many of the indices currentlyemployed.

Use of alpha-l-antitrypsin (al-AT) as anendogenous marker to detect protein-losing enteropathy

C. L HIRONDEL, C. FLORENT, C. DESMAZURES,C. AYMES, AND J. J. BERNIER (HopitalSaint-Lazare, Paris) a1-antitrypsin isnot destroyed by pancreatic enzymes andis found in the faeces; it was thereforeused to measure the plasma protein lossinto the gastrointestinal tract.We examined 28 patients without

gastrointestinal disease and 19 withprotein-losing enteropathy diagnosedeither by isotopic technique or by evidenceof active chronic inflammatory boweldisease.

Stools were collected during a four to10 day period. Blood samples were takenat the beginning and the end of the study.a,-AT was measured through an immuno-diffusion technique using commerciallyavailable plates.The following measures were made

daily: al-AT faecal concentration, a,-ATfaecal loss (F), al-AT faecal clearance (C).

C =F where P is the mean a,-AT serum

concentration.

Our results show that a,-AT faecalconcentration and a,-AT faecal loss werehigher in patients than in control subjects(p < 0 01); (2) the determination ofa1l-AT clearance was more sensitive toseparate the two groups (p < 0 001); (3)the mean clearance calculated over a threeday period allowed a diagnosis of protein-losing enteropathy in all patients exceptone; (4) no false positive results wereobserved in control subjects.

This simple and reliable test willprovide new facilities in patients' diagnosisand in management of the treatment.

Is faecal alpha-i antitrypsin measurementa reliable diagnostic test of protein-losingenteropathy?

M. R. HAENEY, J. FIELDS, R. A. CARTER,R. A. THOMPSON, AND P. ASQUITH(Regional Immunology Laboratory,Department of Nuclear Medicine andMetabolic Research Unit, East BirminghamHospital, Birmingham) It has beensuggested that estimation of al-antitrypsin(alAT) in random faecal samples pro-vides a reliable index of intestinalprotein loss'.We have measured faecal alAT

concentrations in 20 adults (11 men,nine women), average age 39 years (range18-60 years), investigated for suspectedprotein-losing enteropathy in a metabolicunit where accurate faecal collectionswere possible. The diagnoses were Crohn'sdisease (six), ulcerative colitis (three),contaminated bowel syndrome (three),coeliac disease (two), chronic activehepatitis (two), sigmoid diverticulosis(one), small intestinal ischaemia (one),chronic pancreatitis (one), and intestinallymphangiectasia (one). Alpha-i-anti-trypsin concentrations in stool homo-genates were measured by single radialimmunodiffusion using a monospecificrabbit anti-human alAT and the resultsexpressed in three ways: mg alAT/g dryweight of stool/day, mg alAT excreted/day, or as a ratio of faecal/serum alAT.Mean (five day) faecal alAT results werecompared with the simultaneous faecalloss of 51Cr-albumin and with serumlevels of albumin, immunoglobulins andorosomucoid, and with faecal nitrogenexcretion.No significant correlation was found

between faecal alAT results and 5"Cr-albumin loss or other parameters, thusseriously questioning the validity offaecal alAT estimation as a screening testfor protein-losing enteropathy.

Reference'Crossley, J. R., and Elliott, R. B. (1977). Simplemethod for diagnosing protein-losing entero-pathies. British Medical Journal, 1, 428-9.

Radio-opaque markers for faecal fat

F. G. SIMPSON, G. P. HALL, J. KELLEHER,AND M. S. LOSOWSKY (University ofLeeds Department ofMedicine, St. James'sHospital, Leeds) Inert markers enableaccurate faecal fat estimations with onlyshort collection periods1 but chemicalestimation of marker is needed. Radio-opaque beads have been tried as inertmarkers in normal subjects2. Estimation issimple by radiography. We report theiruse in routine faecal fat studies in patients,and comparison with established markers,polyethylene glycol 4000 (PEG) andchromium sesquioxide (Cr2O3).

Thirty-seven patients with a variety ofdiagnoses were maintained on constantfat diets and given PEG 500 mg tds,Cr203 500 mg tds, and beads 8 tds,After five days' equilibration, faeces werecollected into polyethylene bags for two-day periods. A total of 84 collections wasmade. Water was added before homo-genising with an external mechanicalpaddle. The homogenised specimens werefrozen, radiographed, thawed, and analiquot removed for analysis.Over a wide range of faecal fat (2-0-

62-0 g/day), recovery of markers, asexpected, varied widely (28-282%, mean94%, for PEG; 33-244%, mean 97%, forpellets; 28-250 %, mean 99-7 %, for Cr2O3).There were very close correlations betweenexcretions of all markers (correlationcoefficients between 0-92 and 0 95).Corrected faecal fats also correlatedextremely closely (correlation coefficientsbetween 0-92-0-97).Thus radio-opaque pellets are reliable

and offer advantages over traditionalmarkers for routine faecal fat balances.Estimation is precise and simple. Faecescan be collected, homogenised, and themarker determined in the polyethylenebag used for collection.References'Walker, B. E., Kelleher, J., Davies, T., and Losow-

sky, M. S. (1971). Chemical faecal fat usingsingle stools. Scandinavian Journal of Gastro-enterology, 6, 277-280.

'Branch, W. J., and Cummings, J. H. (1978).Comparison of radio-opaque pellets and chrom-ium sesquioxide as inert markers in studiesrequiring accurate faecal collection. Gut, 19,371-376.



j.com/



ownloaded from

http://gut.bmj.com/


Use of 99mTc-HIDA in hepatobiliaryscintigraphyT. V. TAYLOR, D. C. CARTER, G. P.MCLOUGHLIN, A. MILLER, AND M. DSUMERLING (Departments of ClinicalSurgery, Medical Physics, and Radiology,Edinburgh Royal Infirmary) Scintigra-phic imaging of the hepatobiliary systemhas been recently improved by theavailability of ssmTc-pyridoxylidene-glutamatel. s'mTc HIDA may haveadvantages over pyridoxylidenegluta-mate2, but has not been fully evaluatedin man.

Thirty patients with a variety of hepato-biliary pathology were investigated. Afteran overnight fast, 2 mCi of 99mTc wasgiven intravenously and the patientscanned continuously for one hour. Whenfunctioning normally the liver wasvisualised within five minutes, excretioninto the common bile duct, gall bladder,and duodenum occurred at an early stageand the Ti of the tracer in the bloodstream was 20 minutes. In completeobstructive jaundice, isotope was excretedthrough the renal tract without visualisa-tion of the liver.The technique was combined with

oral cholecystography and ultrasono-graphy in 20 patients presenting asemergencies with abdominal pain andsuspected biliary disease. Agreementbetween 99"mTc-HIDA and cholecysto-graphy in terms of gall bladder fillingoccurred in 18 patients but in additionstones were visualised on three cholecysto-grams. Ultrasonography also agreed withHIDA scintigraphy in 18 patients withabnormal gall bladders. Scintigraphy wasparticularly useful in assessing thejaundiced patient. We have not witnessedany adverse side-effects from the use ofthis radiopharmaceutical.

References'Matalo, N. M., Stadalnik, R. C., and Wolfman,

E. F. (1977). Hepatobiliary scanning using99mTc-pyridoxylideneglutamate. AmericanJournalof Surgery, 133, 116-120.

2Wistow, B. W., Subramanian, G., Van Heartum,R. L., Henderson, R. W., Hall, R. C., andMcafee, J. G. (1977). An evaluation of 99mTc-labelled hepatobiliary agents. Journal of NuclearMedicine, 18, 455-461.

Ultrasonography in the differential diag-nosis of jaundice

C. RUBIO, L. BERGER, J. DOOLEY, AND S.SHERLOCK (Departments of Medicine andDiagnostic Radiology, Royal Free Hospital,London) This investigation assessed the

accuracy of ultrasound in the differentialdiagnosis of jaundice. One-hundred-and-sixteen jaundiced patients referred to theRoyal Free Hospital for ultrasonographywere studied from 1 September 1977 to30 April 1978. The diagnosis of 'surgical'or 'medical' jaundice was based on thepresence or absence of dilated ducts.Fifty-one were found to have dilated, and65 non-dilated ducts. Forty-nine of the51 (96%) had obstructed ducts confirmedby percutaneous transhepatic cholangio-graphy (PTC), endoscopic retrogradecholangiopancreatography (ERCP),surgery, or necropsy. In 39 of these 49(79 5 %) the site of obstruction wascorrectly identified. However, the causewas diagnosed in only 14 (28 5 %).

All 65 patients with non-dilated ductshad a definite diagnosis established byliver biopsy, PTC, ERCP, surgery, ornecropsy. In six of the 65 (9-2 %) theultrasound findings were incorrect. Fourof these six had moderately dilated ductsand stones in the common bile duct, onehad sclerosing cholangitis, and the otherone had clots in the common bile ductvisualised by PTC.

Ultrasound differentiated between'medical' and 'surgical' jaundice in 93%of the patients. As a non-invasive tech-nique, it should be used routinely in theinvestigation of the jaundiced patients.

Non-invasive techniques in the diagnosis ofjaundice-ultrasound and computer

A. THEODOSSI, P. J. WHEELER, R. PICKFORD,J. LAWS, AND R. WILLIAMS (Liver Unit,King's College Hospital and MedicalSchool, Denmark Hill, London) Thisstudy aimed to document the accuracy indifferentiating between medical and surgi-cal types of jaundice using ultrasound in84 patients and computer-aided diagnosis(CAD) in 169. Fifty of these patients wereevaluated by both techniques.The correct final diagnosis was estab-

lished in all patients by prolonged clinicalfollow-up and by histology in those with-out extrahepatic biliary obstruction, andby laparotomy or necropsy in those with.Two of 84 patients were technical

failures for ultrasound. Seventy-five ofthe remaining 82 (91 5%) were correctlyseparated into medical and surgicalcategories. One-hundred-and-fifty-two ofthe 169 patients (90%) were similarlycorrectly classified with CAD.

In the 50 patients evaluated by bothtechniques correct diagnostic separation

was achieved in 43 (86%) by ultrasoundand in 42 (84 %) by computer. This groupof patients included the two ultrasoundtechnical failures, both being correctlyidentified as surgical by the computer.Six of the computer diagnoses were falsenegatives and therefore missed a surgicaldiagnosis and two were false positive. Allthe errors for ultrasound representedfalse negatives.

These data therefore indicate that ultra-sound and CAD are similarly effective indistinguishing between medical and surgi-cal causes of jaundice.

Clinical value of the dumping provocationtest: an independent assessment

0. LAWAETZ, Y. ARITAS, D. N. L. RALPHS,AND M. HOBSLEY (Department of SurgicalStudies, The Middlesex Hospital, London)The dumping provocation test (150 ml50% glucose orally) may be used to evalu-ate patients with symptoms after gastricsurgery. Patients with a fall in plasmavolume, derived from haematocrit mea-surements, of at least 9% have beenfound to experience dumping symptomsduring the test.' Using this objectivecriterioni alone 81 patients were tested(on average four years) after a varietyof operations for duodenal ulceration;a fall in plasma volume of less than 9%Obeing deemed a negative test and a 9% orgreater fall a positive one. Before the testan independent observer had placed thepatients into a dumping (39) or non-dumping (42) group after assessmentof their clinical symptoms.2A positive test showed an 83% correla-

tion with the group assessed as 'dumpers'symptomatically. The predictive value of anegative test was 85%. Twelve patients,who experienced pain after ordinary mealsand subsequently were shown to haverecurrent ulceration, had negative tests.The clinical value of the test lies

particularly in its ability to sort patientswith any postprandial symptoms aftergastric surgery. We conclude, that apositive dumping provocation test con-firms the diagnosis of dumping withnearly 100% accuracy. A negative testindicates the need for further investiga-tions.

References'Lawson-Smith, C., and Thomson, J. P. S. (1975).A dumping provocation test. British Journal ofSurgery, 62, 1 53.

'Meurling, S. (1953). Post cibal symptoms after



j.com/



ownloaded from

http://gut.bmj.com/


partial gastrectomy for peptic ulcer. ActaSocietatis Medicorum Upsaliensis, Supp. 3.

Clinical relevance of gastrin assay

T. P. CORBISHLEY, R. C. G. RUSSELL, ANDS. J. MITCHELL (Departments of SurgicalStudies and Nuclear Medicine, TheMiddlesex Hospital, London) The esti-mation of the level of plasma gastrin is ofestablished value in the diagnosis of theZollinger Ellison syndrome (ZES). Theclinical impact of gastrin radioimmuno-assay has been assessed by a review offive years' experience. In 1972 gastrinlevels were estimated in three patients;subsequent years have shown a progressiverise (1973: 50; 1974: 273; 1975: 275;1976: 775; 1977: 992) The proportionof patients with raised gastrin levels(50 pmol/l) has increased (1974: 14%;1977: 34%), but the percentage whoseraised gastrin level was confirmed asbeing due to the ZES has remained atapproximately 4% (n = 84).A survey of all patients from whom

samples were received in 1977 wasconducted to determine the reasons forthe request. The indications for requestingthe gastrin level in 423 replies were:gastrointestinal haemorrhage 25% (75%sample levels 50 pmol/l); perforation10% (70% 50 pmol/l); aggressive pepticulcers 35% (66% 50 pmol/l); multiplepeptic ulcers 17%; jejunal ulcer 6%;diarrhoea 21 %; failed surgery 28 %. Ofthose patients whose diagnosis of ZESwas confirmed the commonest presentingsymptom was aggressive peptic ulceration(37 %). The clinician considered that thegastrin estimation was of clinical relevancein 80% patients and 35% replies con-sidered that the gastrin result had in-fluenced clinical management.

This survey suggests that the radio-immunoassay of gastrin is of value inclinical gastroenterology and that pre-sentation alone will not accurately denotethe ZES.

ABSORPTION AND MALABSORPTION/LIVERAND BILIARY

Different rates of fat absorption fromhomogenised and solid mealsA. CORTOT, S. F. PHILLIPS, AND J.-R.MALAGELADA (Gastroenterology Unit,Mayo Clinic, Rochester, Minnesota, USA)

Fats leave the stomach faster wheningested as homogenised meals thanwhen eaten as solids.' We measured fatabsorption after ingestion of identicalmeals in different physical forms, toachieve different rates of gastric empty-ing. Meals consisted of hamburger, bread,butter, ice cream, and water (total fat15 g). Four subjects received homogenisedmeals, nine ate the solid meal; mealscontained 3H-glycerol triether (3H-GTE)and/or sucrose octaoleate ('4C-SPO) asnon-absorbable lipid-phase markers.2 In-tubation allowed gastric samplingand total recovery of contents proximalto a jejunal balloon, 30 cm beyond theduodenojejunal junction.

1. Ratios of lipid markers to totalfatty acids (TFA) in the stomach remainedconstant. Homogenised meals emptied infour hours; % of marker emptied/h was53, 27, 11, 9. Solid meals required sixhours to empty; (%/h) 6, 17, 18, 21,23, 15. Total amounts of fat emptiedwere not different: homogenised 8-7 +1-4 g, solid 7-7 ± 0 7 g.

2. Ratios of lipid marker/TFA injejunal samples were higher after the solidmeal, reflecting greater absorption of fatafter this meal.

3. Total fat absorbed was greaterafter solid meals (81-1 ± 3 4%) thanafter homogenised meals (47 5 ± 10-7 %;p < 0-005). In conclusion, fat ingestedwith a solid meal emptied from thestomach much more slowly but wasabsorbed from the proximal intestinemore efficiently.

References'Cortot, A., Phillips, S. F., and Malagelada,

J.-R. (1978). Gastric emptying of lipids aftermixed meals in man. Gastroenterology, (abstract),74, 1021.

2Cortot, A., Phillips, S. F., and Malagelada,J.-R. (1978). Sucrose polyester ("'C-SPO) asoil phase marker in man. Gastroenterology.(Abstract). 74, 1021.

Bile acid induced colonic secretion:increased mucosal CAMP or altered perme-ability?

M. CAMILLERI, B. T. COOPER, AND V. S.CHADWICK (Department of Medicine,Royal Postgraduate Medical School, Lon-don) Di-hydroxy bile acids have beenreported to increase cyclic AMP pro-duction' and mucosal permeability2 inthe mammalian colon. Either or bethmechanisms could account for bile acidinduced colonic secretion. To test therelative roles of cAMP and permeability

changes, in vivo colonic perfusions wereperformed in rabbits, using isotonic salinecontaining PEG4000 as non-absorbablemarker. Intestinal permeability was asses-sed by the absorption of different sizedPEGs (mol. wt. 242-638), cAMP wasassayed by Gilman's method.3 Controlsolution was perfused for two hours andtest solution for three hours (sodiumchenodeoxycholate 5 mmol/l). Colonicmucosal biopsies were taken in the controlperiod and after one and three hours ofbile acid perfusion.

Fluid secretion increased progressivelyduring the bile acid perfusion to a peaklevel of 5 2 + 0 9 ml/10 min. during thethird hour. Mucosal permeability simi-larly showed a stepwise increase duringthis period. Mucosal cAMP did notchange significantly being 5 0 + 1P4pmol/mg protein during the controlperiod and 6 5 + 1 8 pmol/mg proteinduring bile acid perfusion.

These results suggest that changes inmucosal permeability and not increasedcAMP production are responsible forbile acid induced colonic secretion.

References'Binder, H. J., Filburn, C., and Volpe, B. T.

(1975). Bile salt alteration of colonic electrolytetransport: role of cyclic adenosine mono-phosphate. Gastroenterology, 68, 503-508.

2Chadwick, V. S., Gaginella, T. S., Debongie,J-C, Carlson, G. L., Philips, S. F., and Hofmann,A. F. (1976). Mucosal epitheliolysis: a mechan-ism for the increased colonic permeabilityinduced by dihydroxy bile acids. Gut, 17, 10.

'Gilman, A. G. (1970). A protein binding assay forAdenosine 3' : 5'-cyclic monophosphate.Proceedings of the National Academy of Sciences,67, 305-312.

Sodium dependence of glucose and maltoseabsorption in the jejunum in man

G. I. SANDLE, R. W. LOBLEY, AND R. HOLMES(University Department of Gastro-enterology, The Royal Infirmary, Man-chester) In vitro work suggests Na+-independent disaccharidase-linked glucosetransport from maltose in addition to themonosaccharide transport system.'To investigate this, 20 cm segments of

proximal jejunum were perfused (flowrate 18 ml/min) by double-lumen tubein four normal subjects using solutionsin sequence containing 56 mM glucose +122 mM Na+ (I), 56 mM glucose withoutNa+ (II), 28 mM maltose + 122 mMNa+ (III), and 28 mM maltose withoutNa+ (IV). Solutions were made isotonicwith mannitol as appropriate and con-tained PEG 5 g/l. Glucose absorption



j.com/



ownloaded from

http://gut.bmj.com/


6

(mmol/10 min ± SEM) from II (3-80 +04) was 235% less (0005 < p < 001)than from I (4 97 ± 0 19). Similarly,absorption from IV (4-91 ± 0 52) was22% less (0-025 < p < 0-05) than from III(6-29 ± 0-37).

Glucose absorption from III (6-29 +0 37) was significantly higher than fromI (4 97 ± 0 19, 0 005 < 2P < 0 01) whilenet water absorption (ml/10 min ± SEM)from Ill (25-49 ± 3 45) was significantlylower than from I (36 58 ± 3-38, 0 025 <2P < 0 05). Similarly, glucose absorptionfrom IV (4-91 ± 0 52) was significantlyhigher than from II (3-8 ± 0 4, 0*02 <2P < 0 025), while net water secretion wasno different (50-36 ± 17-49 and 45-31 +8-07 respectively).

Maltose hydrolysis rates (mmol/10 min± SEM) with II (4-21 ± 0-13) andIV (4-19 ± 0-24) were not significantlydifferent (0 45 < p < 0-475).

It is concluded that (1) Na+-independentdisaccharidase-linked glucose transportwas not confirmed in man, (2) glucosewas absorbed more rapidly from 28 mMmaltose than 56 mM glucose during bothnet absorption and secretion of water,(3) maltose hydrolysis was unaffected bylow intraluminal sodium concentrations.

Reference1Ramaswamy, K., Malathi, P., Caspary, W. F.,and Crane, R. K. (1974). Studies on the transportof glucose from disaccharides by hamster smallintestine in vitro. II. Characteristics of thedisaccharidase-related transport system. Bio-chimica et Biophysica Acta, 345, 39-48.

Movement of water and sodium during theabsorption of glucose and maltose in thejejunum in man

G. I. SANDLE, R. W. LOBLEY, AND R. HOLMES(University Department of Gastro-enterology, The Royal Infirmary, OxfordRoad, Manchester) Twenty centimetresegments of proximal jejunum wereperfused in three normal subjects (flowrate 21 ml/min) with isotonic glucose-saline solutions containing PEG 5 g/l:(1) 55.5 mM glucose, 122 mM NaCI;(2) 166-5 mM glucose, 66 mM NaCl;(3) 222 mM glucose, 39 mM NaCI;(4) 277-5 mM glucose, 11 mM NaCI.Glucose absorption (mmol/10 min +SEM) increased from 5-34 + 0-61 with(1) to 9-73 + 0-77 with (4), while netwater movement (ml/10 min ± SEM)changed from 35-3 ± 5-27 absorbed(1) to 15-89 + 4-91 secreted (4). Netsodium movement (mmol/10 min +SEM) changed from 3-79 ± 0-67 absorbed(1) to 5-5 ± 0-11 secreted (4).

Four normal subjects were perfusedwith isotonic solutions containing PEG5 g/l, 21 mM NaCl and appropriateamounts of mannitol: 35 mM maltose;222 mM glucose; 35 mM maltose +222 mM glucose; 111 mM maltose.Absorption rates of glucose were 6-23 +099, 1016 ± 034, 11-58 ± 034, and11-33 ± 0 53 respectively. Net water andsodium movements (secretion in all)were 42-08 ± 5-84, 24-62 + 3-76, 32-42 +10-63, 42-89 ± 618 and 6-97 + 0-38,5'86 + 0-58, 6-17 ± 0-80, 6'18 ± 077respectively.

Five normal subjects were perfusedwith isotonic solutions containing PEG5 g/l, 21 mM NaCl and mannitol: 111 mMglucose; 111 mM maltose; and 111 mMglucose + 111 mM maltose. Absorptionrates of glucose were 6-39 ± 0 43,8-51 + I1 1, and 1098 ± 1-5 respectively.Net water and sodium movements (secre-tion in all) were 30-25 + 3 34, 51-16 +6-07, 35 04 + 3-16 and 8-34 + 1 10,8-24 + 1V52, 7.47 ± 1-13 respectively.Thus (1) increasing secretion of water

and sodium may accompany increasingglucose absorption, (2) water and sodiummove together, the direction dependingon sodium concentration perfused ratherthan rate of glucose absorption, (3) secre-tion of water is greater with maltose thanwith equimolar glucose.

Glucose-stimulated intestinal fluid secre-tion: an acquired intestinal transport defect

J. DAWSON, H. J. F. HODGSON, T. J. PETERS,AND V. S. CHADWICK (Department ofMedicine, Royal Postgraduate MedicalSchool, London) Diarrhoea in commonvariable immunodeficiency is often associ-ated with jejunal mucosal abnormalities,bacterial overgrowth, or giardiasis.1 Westudied a patient with profound secretorydiarrhoea in whom extensive investiga-tions failed to show any of these abnor-malities.To elucidate the pathophysiology of the

diarrhoea we perfused the jejunum, ileum,and colon with various isosmotic solutionsby triple lumen perfusion techniques.2In the jejunum, fluid and electrolytetransport were normal using saline andbicarbonate-saline solutions. However,addition of 30 mM glucose to the per-fusate stimulated profound net secretionof water (125 ml/30 cm segment/h),sodium (16-4 mmol/h), and chloride(26-6 mMol/h). In contrast, glucosestimulated absorption in normal subjects.

Glucose itself was absorbed normally.An exactly similar phenomenon wasdemonstrated in the ileum. Perfusion ofthe whole colon with saline demonstratednet secretion of water and electrolytes.

Failure of glucose-stimulated absorp-tion was not due to failure of glucoseabsorption per se nor was there any brushborder abnormality demonstrable byelectron microscopy or analytical sub-cellular fractionation.3 A hormonalmechanism for this glucose stimulatedsecretion is unlikely since it was notabolished when the perfusion was re-peated during intravenous somatostatininfusion. The phenomenon represents ahitherto undescribed acquired intestinaltransport defect.

References1Hermans, P. E., Diaz-Buxo, J. A., and Stobo,

J. D. (1976). Idiopathic late onset. Immuno-globulin deficiency: clinical observations in 50patients. American Journal of Medicine, 61,221-237.

"Cooper, H., Levitan, R., Fordtran, J. S., andIngelfinger, F. J. (1966). A method for studyingabsorption of water and solute from the humansmall intestine. Gastroenterology, 50, 1-7.

3T. J. Peters (1976). Analytical subcellar fractiona-tion of jejunal biopsy specimens: methodologyand characterisation of the organelles in normaltissue. Clinical Science and Molecular Medicine,51, 557-574.

Drug and dietary modification of carbo-hydrate absorption

R. H. TAYLOR, D. J. A. JENKINS, ANDR. NINEHAM (introduced by J. J. MISIEWICZ)(Department of Gastroenterology, CentralMiddlesex Hospital, London; DepartmentofRegius Professor of Medicine, RadcliffeInfirmary, and University PhysiologyLaboratory, Oxford) Small intestinallesions are known to alter the pattem ofcarbohydrate absorption. The develop-ment of a glycosidehydrolase-inhibitor(BAY g 5421) makes possible controlledcarbohydrate malabsorption withoutaffecting transit. These studies aim toshow that gastrointestinal absorption ofcarbohydrate can be modified to flattenthe post-prandial glycaemia curve withoutproducing nonabsorption.

Sucrose with and without BAY g 5421was given to healthy male volunteers.Serial measurements were made of bloodglucose and breath hydrogen.BAY g 5421 200 mg given with 50 g

sucrose produced a reduction in thecontrol blood sugar rise of 76 + 7%(P < 0001, n = 6). Calibration of thesubjects for hydrogen production fromunabsorbed carbohydrate with lactulose



j.com/



ownloaded from

http://gut.bmj.com/


25 g indicated that 63 ± 15% (p < 0 002,n = 6) of the original sucrose was notabsorbed. Fifty milligrams of the drugwith 50 g sucrose produced no significanthydrogen or flattening of the curve.

However, when half the test mealcarbohydrate was given as starch, 50 mgof BAY g 5421 reduced the 30 minutemean peak blood sugar rise by 22%(p < 0 01, n = 4) and by a further 42 %-that is, total 64% (p < 0-05, n = 4)-whenfibre (guar gum) was added, with nobreath hydrogen evidence of carbo-hydrate malabsorption.We conclude that modification of

glucose absorption may allow flatteningof the post-prandial glycaemia curvewithout producing malabsorption andits symptoms. These results may explainthe absence of post-prandial glycaemiawithout symptoms of carbohydrate mal-absorption in gastrointestinal enzymedeficiency states.

Ursodeoxycholic acid reduces fatty diar-rhoea after ileal resection

T. M. COX, G. VAN BERGE HENEGOUWEN,AND V. S. CHADWICK (Department ofMedicine, Royal Postgraduate School,London, and Universiteitskliniek VoorInwendige Ziekten, Nijmegen) Profusesteatorrhoea is a major problem afterileal resection. It results largely fromdecreased micellar solubilisation of fatdue to intraluminal bile salt deficiency.Attempts to correct the fat maldigestionby feeding sodium taurocholate1 reducedsteatorrhoea but aggravated diarrhoea,because of bacterial conversion to deoxy-cholate in the colon.

Activity studies2 showed that ursode-oxycholic acid (UDCA), unlike deoxy-cholic acid and chenodeoxycholic acid,did not produce colonic secretion.Accordingly, a trial of bile acid replace-ment with UDCA was undertaken in sixpatients with ileal resections and steator-rhoea. After a four day control period0 25 g UDCA was given before meals forfour days; during this period analysis ofbile showed that UDCA accounted for51 % of bile acids. At this low dose,daily faecal fat fell by a mean of 15%(p < 0-05) with 11% decrease in faecalwet weight (p < 0 01). In addition,UDCA rich-bile was shown in vivo to bean effective mixed micelle-former withdietary fat. When 4g UDCA was givendaily in a further patient, there was a 40%reduction in faecal weight and a 50%reduction in faecal fat.

These studies show that bile acid replace-ment with UDCA reduces steatorrhoeaand diarrhoea after ileal resection.

References"Hardison, W. G. M., and Rosenburg, I. H.

(1967). Bile salt deficiency in the steatorrhoeafollowing resection of the ileum and proximalcolon. New England Journal of Medicine, 277,337-343.

2Chadwick, V. S., Carlson, G. L., Gaginella, T. S.,Debongnie, J-C., Phillips, S. F., and Hofmann,A. F. (1977). Structure activity relationships ofbile acids in the rabbit colon. European Journalof Clinical Investigation, 7, 241-242.

Composition and enzymes of crypt andvilius cells in the small intestine of folatedeficient rats

JACQUI BADCOCK AND A. M. TOMKINS(Department of Human Nutrition, LondonSchool of Hygiene and Tropical Medicine,London) Mucosal folate deficiency ofthe small intestine is a feature of infectivemalabsorption syndromes, but its contri-bution to diarrhoea is uncertain. Experi-mental folate deficiency is associatedwith the development of diarrhoea,morphological lesions, and a reduction intotal mucosal DNA, RNA, and protein,to 44%, 59%, 63%, of control animals,respectively. Fractions of epithelial cellsfrom villi and crypts of control animals,showed higher ratios of RNA/DNA andprotein/DNA in villus than crypt cells(259 ± 13 vs. 165 + 6 and 3-77 ± 0 23 vs.2-06 ± 0-18, P < 0 01). In folate deficiencyRNA/DNA and protein/DNA weregreater, especially at the villus tips, com-pared with controls (395 ± 34 and5-35 + 032, p < 0 01). Radioautographyshowed a delay in cell migration time infolate deficiency (4-7 vs. 10-2 ,tM/h incontrols, p < 0 001) and might suggestthat the compositional changes of theepithelial cells represent 'hypermaturity'.In villus tip fractions, sucrase activityper cell was sufficiently increased fortotal gut sucrase not to be decreased byfolate deficiency. However, Na-K-ATPaseactivity per villus tip cell did not increasein folate deficiency and total mucosalNa-K-ATPase was lower. This suggeststhat, despite the longer migration timepermitting acquisition of enzymes forcarbohydrate absorption, the cells did notacquire increases of enzymes associatedwith sodium and water transport, andthis may contribute to the diarrhoea ofmucosal folate deficiency.

Absorption of dietary cholesterol andhepatic cholesterol synthesis in cirrhosis

M. PONZ DE LEON, F. ZIRONI, P. LORIA,A. SMERIERI, AND N. CARULLI (Istitutodi Clinica Medica, Policlinico, Modena,Italy) Dietary cholesterol absorptionand neosynthesis, mainly in the liver,are among the factors that regulate bloodcholesterol levels. The low cholesterollevels frequently observed in cirrhosis,"2therefore, could either be due to animpaired cholesterol absorption or to areduced synthesis. We aimed to assesshow these two variables could contributeto the genesis of hypocholesterolaemiain cirrhosis.We evaluated hepatic cholesterol syn-

thesis by determining the activity of thehydroxymethylglutaryl-CoA reductase(HMG-CoA-R), the rate limiting enzymein cholesterol synthesis, in surgical liverbiopsies of six patients undergoingportocaval shunt. Cholesterol absorptionwas estimated in 12 cirrhotic patients byfeeding a standard dose of C-14 labelledcholesterol, plus sitosterol-H-3 as a non-absorbable marker, and by measuringthe recovered radioactivity in the faecalneutral sterol fraction for the next sixdays. All patients had marked hypochole-sterolaemia (102 ± 14 mg/dl SDM).

Results showed that HMG-CoA-Ractivity was 59 7 ± 6-6 pmol/m/mg prot.in a control group of eight patientsoperated on for duodenal ulcer and30-2 + 5-6 (P < 001, 50 5% decrement)in the cirrhotics. Cholesterol absorptionwas 38-4 ± 4% of the administered dosen normal controls and 27-9 + 10%(P < 0-05, 27-4% decrement) in thecirhrotics.We conclude that both a decreased

cholesterol synthesis and a low cholesterolabsorption could explain the low chole-sterol levels observed in cirrhosis; syn-thesis, however, seems to be more affectedthan absorption.

References'Heder, H. A., Russ, A. M., Rees Pritchett, R. A.,

et al. (1955). Journal of Clinical Investigation,34, 1147-1162.

2Saudek, C. D. (1977). Anmerican Journal ofMedicine, 63, 453-458.

Dynamics of gall-bladder contraction inresponse to different cholecystokinins

JANE M. OLIVER, J. F. REY, AND R. F.HARVEY (Department ofMedicine, BristolRoyal Infirmary and Frenchay Hospital,



j.com/



ownloaded from

http://gut.bmj.com/


Bristol) Gall-bladder contraction beginsvery soon after food is eaten and toniccontraction continues for at least one totwo hours after a normal meal. Theseeffects are mediated mainly by the peptidehormone cholecystokinin-pancreozymin(CCK).Radioimmunoassay studies indicate

that several different molecular forms ofCCK are present in the body. The molecu-lar sizes of the three known forms,cholecystokinin(33 aminoacids, CCK-33)CCK-variant (39 aminoacids, CCK-39),and octapeptide (eight aminoacids,CCK-8) are very different.The dynamics of action of different

CCK preparations on rabbit gall-bladdermuscle strips were studied in vitro using asuperfusion technique. Gall-bladder con-traction begins within a few seconds aftercontact with each CCK preparation.Peak contraction was reached after amean of 8X16 minutes (range 6X8 to 9 0minutes). Initial speed of contraction wasgreater with CCK-8 than CCK-33 (50% ofpeak contraction in 1X75 + 0 35 vs.2X55 ± 0-64 min, P < 0 05). Gall-bladderrelaxation was significantly slower aftercessation of CCK-33 superfusion thanafter CCK-8 (Ti = 5.39 ± 0-75 vs.4.67 + 0-63 min, P < 0-05). The BootsCCK preparation (Boots Company Ltd.,Nottingham) produced both faster con-traction (50% contraction in 2-03 ± 0*29vs. 2-65 ± 0-62 min, P < 0 05) and fasterrelaxation after ending the superfusion(Ti = 3.53 ± 0 37 vs. 5 39 ± 0-75,p < 0 001) than GIH cholecystokinin(Karolinska Institute, Stockholm).We conclude that (1) different molecu-

lar forms of CCK have qualitativelydifferent effects on the gall-bladder, and(2) the effects of the Boots preparation ofCCK are consistent with the finding thatit contains mainly smaller molecularforms of the hormone.

Investigation of hormonal control of gallbladder storage in man: studies usingindocyanine green (ICG)

0. G. BJORNSSON, J. DAWSON, G. R. GREEN-BERG, R. F. MCCLOY, T. PANAYIOTIDIS,T. E. ADRIAN, S. R. BLOOM, AND V. S.CHADWICK (Department of Medicine,Royal Postgraduate Medical School,London) Cholecystokinin (CCK) isknown to promote gall bladder contrac-tion. Recent investigations suggest thatpancreatic polypeptide (PP) may haveopposite actions leading to gall bladder

storage. To investigate this hypothesiswe have measured gall bladder storageusing intravenously administered ICGas a marker of biliary output by aduodenal perfusion technique, in healthycontrols, and after cholecystectomy. Inthe steady state a gall bladder storagefraction was calculated from any decreasein duodenal ICG output.

In normal fasting subjects the gallbladder storage fraction was 0-8 to 10.Fasting storage patterns were abolishedby simulating the interprandial statewith intravenous secretin (0-1 CU/kg/h)and caerulein (10 ng/kg/h). As pre-dicted, the storage fraction in fastingcholecystectomy patients was zero. Duringsecretin-caerulein infusion in normalsubjects, bovine PP at physiologicalplasma concentrations significantly (p <0-001) decreased ICG output. In contrast,BPP did not change ICG output incholecystectomised patients.We conclude that low circulating levels

of secretin and caerulein abolished fastinggall bladder storage patterns. Under theseconditions PP promoted gall bladderstorage.

Failure of PP to decrease ICG output incholecystectomised patients suggests adirect action on the gall bladder.

Real-time ultrasound-a new method forinvestigating gall-bladder dynamics

M. H. ORNSTEIN, A. PALFRAMAN, ANDH. MEIRE (Northwick Park HospitalHarrow) (introduced by I. M. BAIRD)Little is known of gall-bladder dynamics.Research methods have been indirector invasive, depending on the metabolismof oral radiographic agents or the risksof intravenous agents. Radiation dosageis appreciable and magnification is amajor problem, compounded by move-ment of the gall bladder during contrac-tion. Radio-isotopic scanning suffersfrom similar disadvantages and intubationis non-physiological.

These problems can be overcome bythe use of real-time, grey-scale, ultra-sound B-scanning,1 and in this preliminarystudy we have investigated gall-bladderemptying in 13 normal male fastingvolunteers. The gall bladder is visualisedusing a hand held probe which is rotateduntil the largest possible area of gallbladder is obtained on the video-screenand this is measured on a photographiccopy using planimetry. After an initialscan a standard fatty meal is taken and

scans repeated at frequent regular intervalsfor 30-45 minutes.

In keeping with previous work,2 despiteconsiderable initial variation the gallbladders showed a remarkably similar,almost linear, reduction in size reaching76 + 25% (mean = SEM) at 15 minutes,50 ± 26% at 30 minutes, but only41-3 ± 4X3% at 40 minutes-that is,with this stimulus complete emptyingdoes not occur.We conclude that this is a straight-

forward non-invasive and physiologicalmethod of studying dynamic gall-bladderfunction, superior to previous methods.

References"Gonzalez, A. C., and Johnson, J. A. (1978).

Ultrasonic examination of the gall bladder:a review. Clinical Radiology, 29, 171-176.

"Boyden, E. A. (1928). An analysis of the reactionof the human gall bladder to food. AnatomicalRecord, 40, 147-191.

Direct measurement of the first passextraction of bile acids by the liver in man

I. T. GILMORE AND R. P. H. THOMPSON(Gastrointestinal Laboratory, Rayne Insti-tute, St. Thomas' Hospital, London)The hepatic extraction of bile acids isassumed to be almost complete but directmeasurements have not been previouslyreported in man. We report such measure-ments for cholic and glycocholic acidin control subjects and patients withchronic liver disease (CLD) and comparethem with indirect estimates of extractionratio (ER) from the ratio of the oral andintravenous clearances.

After intravenous injection or constantinfusion of 14C-glycocholic acid bloodsamples were taken simultaneously fromperipheral and hepatic veins. The ER inthree control subjects was 90 + 2 %,84 ± 3 %, and 88 + 1 % (mean + SD).Extraction was greater than for simul-taneously measured indocyanine green.In five patients with CLD, ER wasreduced to 24-65 %. The ER of 14C-cholicacid in two controls was 72 and 70% andin eight patients with CLD was from17-74%; indirect measurement of oraland intravenous clearances overestimatedER by a mean of 64% of the directlymeasured value.Thus a high hepatic extraction of bile

acids in man is confirmed by directmeasurement and is greater for conjugatedthan unconjugated cholic acid. This isimpaired in CLD where indirect estimatesfrom oral and intravenous clearance mayoverestimate liver function.



j.com/



ownloaded from

http://gut.bmj.com/


Effect of chenodeoxycholic acid (CDCA)on cholesterol absorption in man

M. PONZ DE LEON, P. LORIA, R. IORI, ANDN. CARULLI (Istituto di Clinica Medica,Policlinico, Modena, Italy) The effects ofCDCA administration in reducing biliarycholesterol secretion and desaturating bilehave been related to the inhibition ofCDCA on hepatic cholesterogenesis1.However, only part of the biliary choles-terol is neosynthesised in the liver2 andcholesterol from other sources, includingdietary cholesterol, can be secreted in bile.Accordingly, biliary saturation may bereduced by inhibiting cholesterol absorp-tion3. We decided, therefore, to investigatethe effect of CDCA feeding on cholesterolabsorption.

Eight volunteers were given a standardmeal containing 500 mg cholesterol, 5,LCi cholesterol-C-14, and 10 ,uCisitosterol-H-3 to correct losses fromsources other than absorption. The radio-activity of the faecal neutral sterolfraction was determined for the nextsix days to provide an estimate of un-absorbed cholesterol. The same subjectswere restudied after a 20 day treatmentwith CDCA (15 mg/kg/day). Changes inbiliary lipid and bile acid compositionwere also taken into account.During treatment, the proportion of

CDCA in bile increased from 35 0 +12.4% (SD) to 75 3 + 5-7% of totalbile acids and percent saturation of bilefell from 6-4 ± 2-1 to 4*9 + 1 0%.Cholesterol absorption was 33-7 ± 14%of the given dose in basal condition anddecreased to 15-5 ± 9-8 % (P < 0 01) aftertreatment.We conclude that,given in doseseffective

to desaturate bile, CDCA seems todecrease dietary cholesterol absorption.This effect could contribute to the de-saturation of bile duringCDCA treatment.

References1Salen, G., Nicolau, G., Shefer, S., et al. (1975).

Gastroenterology, 69, 676-684.'Schwartz, C. C., Berman, M., Vlahcevic, Z. R.,

et al. (1978). Journal of Clinical Investigation,61, 408-423.

'Begemann, F., Bandomer, G., and Herget, H. G.(1978). Scandinavian Journal of Gastroenterology,13, 57-63.

Rowachol, a proprietary essential oil pre-preparation, lowers hepatic microsomalHMGCoA reductase and dissolvescholesterol gallstones

G. D. BELL, J. DORAN, ALIYA MIDDLETON,

B. MIDDLETON, C. R. RICHMOND, AND D. A.WHITE (Department of Therapeutics, CityHospital, Nottingham, Department ofSurgery, General Hospital, Nottingham,and Department of Biochemistry, MedicalSchool, Queen's Medical Centre, Notting-ham) Rowachol, a proprietary prepara-tion of essential oil, is freely available inthe UK. Like chenodeoxycholic acid(CDCA) and ursodeoxycholic acid(UDCA), R lowers the lithogenic index ofhuman bile'. We have studied, in the rat,the effect of R on HMGCoA reductase-the rate limiting enzyme for cholesterolbiosynthesis. R significantly (p < 0.01)inhibited hepatic HMGCoA reductase.This is of interest, as R has also beenreported to lower serum cholesterol levelin man.

Twenty-seven patients with radiolucentgallstones were treated with R (twocapsules tds) for periods of six months ormore. There was evidence of gallstonedissolution and/or disappearance in sevenof the 27 patients. The drug was welltolerated and no evidence ofhepatotoxicityor other serious side-effects have emerged.We conclude that prolonged oral

administration of R appears to be safe.This compound and possibly otherterpene substances merit more detailedconsideration as cholelitholytic agentseither alone or in combination with CDCAor UDCA.

References'Doran, J., Keighley, M. R. B., and Bell, G. D.

(1977). Rowachol-a possible treatment forcholesterol gallstones? Gut, 18, A977.

2Beiglbock, W. (1960). As to the effect of essentialoils on the secretory function of the liver and theserum cholesterol. Wiener Medizinische Wochen-schrift, 110, 11-16.

Effects of ursodeoxycholic acid treatment(UDCA) on biliary lipids in man

M. C. BATESON, P. E. ROSS, J. MURISON, ANDI. A. D. BOUCHIER (Department ofMedicine,Ninewells Hospital and Medical School,Dundee) Twelve gallstone patients havebeen reviewed after at least two months'treatment with UDCA. Fasting biliarylipids have been measured in duodenalbile before and after one month's treat-ment with UDCA 250 and 500 mg dailyrespectively in a crossover study. Meanpatient weight was 71-9 ± 15-6 kg andheight 162 ± 8 cm.

After both doses there was a fall inbiliary cholesterol to the normal range.However, there was no consistent fall incholesterol saturation indices. The pro-

portion of UDCA in bile rose, but therewere no changes in the proportions ofbiliary lithocholic or deoxycholic acid.Both cholic and chenodeoxycholic acidfell.

Total serum bile acids were slightlyincreased after treatment, though UDCAwas not always present. The only otherchange detected in blood tests was a 39%fall in y-glutamyl transpeptidase.

There was no diarrhoea after treatmentand most patients reported an improve-ment in abdominal symptoms.We conclude that doses ofUDCA in the

range shown in Japan to dissolve radio-lucent gallstones' 2 reduced biliarycholesterol content of fasting bile tonormal but did not consistently improvethe cholesterol saturation.

References"Nakagawa, S., Makino, I., Ishizaki, T., and Dohi,

1. (1977). Dissolution of cholesterol gallstonesby ursodeoxycholic acid. Lancet, 2, 367-369.

2Okumura, M., Tanikawa, K., Chuman, Y.,Koji, I., Nakagawa, S., Nakamura, Y., lino, H.,Yamasaki, S., and Hisatsugu, T. (1977). Clinicastudies on dissolution of gallstones using ursode-oxycholic acid. Gastroenterologio Japonica, 12,469-475.

PLENARY

Gut hormone profile after gut resection

H. S. BESTERMAN, S. R. BLOOM, T. E. ADRIAN,N. D. CHRISTOFIDES, D. L. SARSON, C. N.MALLINSON, A. PERO, AND R. MODIGLIANI(Royal Postgraduate Medical School,London, Lewisham Hospital, London,Ospedale Mauriziano, Turin, Italy, andHopital St Lazare, Paris, France) Wehave previously demonstrated a greatlyexaggerated postprandial rise in plasmaenteroglucagon in coeliac disease and havepostulated that it might be exerting acompensatory trophic effect on theintestinal mucosal. We therefore wenton to study the effect of intestinal resectionon enteroglucagon and other gut hor-mones, where similar mechanisms mightbe operating, Gut hormone release wasinvestigated after ingestion of a standardtest breakfast (530 Kcal) in 18 patientswith partial ileal resection, in ninepatients with partial colonic resection,and in 11 normal subjects. Gut resectionhad been undertaken for inflammatorybowel disease, neoplasia, radiation, fibro-sis or trauma, a minimum of one yearbefore study. Basal and peak levels of



j.com/



ownloaded from

http://gut.bmj.com/


pancreatic polypeptide, gastrin, motilin,and enteroglucagon were all signifi-cantly raised in the group with ilealresection. In marked contrast, however,only pancreatic polypeptide levels weresignificantly increased in patients withcolonic resection.Thus we find striking abnormalities

in the levels of enteroglucagon and anumber of other gut hormones in patientswith partial ileal resection, which mayreflect compensatory mechanisms. Thegut hormone profile in ileal resectiondiffers markedly from that seen in colonicresection and also from that in coeliacdisease, in ulcerative colitis and in Crohn'sdisease.

Reference'Besterman, H. S., Bloom, S. R., Sarson, D. L.,

Blackburn, A. M., Johnston, D. I., Patel, H. R.,Stewart, J. S., Modigliani, R., Guerin, S., andMallinson, C. N. (1978). Gut hormone profilein coeliac disease. Lancet, 1, 785-788.

Improved diagnostic accuracy of a modifiedoral pancreatic function test (PFT)

C. J. MITCHELL, C. S. HUMPHREY, A. W.BULLEN, J. KELLEHER, AND M. S. LOSOWSKY(University Departments of Medicine andSurgeryv, St. James's Hospital, Leeds) Theoral PFT depends upon urinary recoveryof p-aminobenzoic acid (PABA) releasedby chymotrypsin hydrolysis of orallygiven N-Benzoyl-L-Tyrosyl-p-aminoben-zoic acid (Bz-Ty-PABA). However, falselyabnormal results are frequently foundin bowel or liver disease, as PABArecovery is also affected by abnormalabsorption or hepatic conjugation.Furthermore, pancreatic dysfunction maycoexist with liver or bowel disease.To improve diagnostic accuracy, we

have compared urinary recovery fromBz-Ty-PABA with that from an oral doseof pure PABA, giving a PABA excretionindex:

% 8h urine PABA recovery fromPEI _

oral Bz-Ty-PABA% 8h urine PABA recovery from

oral pure PABAWe have assessed the modified test in

six normal subjects, seven patients withchronic pancreatic insufficiency, and 11patients with small bowel disease (six)or chronic liver disease (five) in whomthe conventional oral PFT was falselyabnormal.The mean PEI for normal subjects was

1.1 (range 0-82-1c18) but was markedlyreduced in the pancreatic patients at 051

(range 0 29-0 67). The PEI was similarto normal for patients with bowel disease(mean 0 9, range 0-72-1-30) and liverdisease (mean 0-99, range 0 73-1-9).Thus the PEI distinguished abnormal

results due to pancreatic disease fromfalsely abnormal results in liver and boweldisease. We conclude that the modifiedoral PFT greatly improves diagnosticaccuracy.

Breath hydrogen in pneumatosis cystoidesintestinalis

J. GILLON, K. TADESSE, R. F. A. OGAN,S. HOLT, AND W. SIRCUS (Gastrointest-inal Unit and Wolfson Laboratory, WesternGeneral Hospital, Edinburgh, Gastro-intestinal and Liver Service, Royal In-firmary, Edinburgh) Pneumatosis cys-toides intestinalis (PCI) is a rare conditionof unknown aetiology. Previous research'suggested that bacterial gas-productionmay be an important factor, and thehydrogen content of the cysts lendssupport to this view, since H2 is notproduced by human cells.We have studied breath H2 excretion in

six patients known to have PCI. Subjects1 and 2 had greatly raised fasting breathH21evels(normal 0-20ppm) and bothshow-ed a coasid-rable rise after 50 g glucoseorally. Both had gross PCI at the timeof the test. Subject 6 had a high-normalbasal breath H2 and a moderate rise inresponse to 50 g glucose. She had under-gone resection of localised sigmoid PCItwo months earlier. Three subjects showeda negative response, only one ofwhom hadPCI demonstrable on plain x-ray.The high basal breath H2 in subjects

1 and 2 implies constant bacterial H2production, and the response to glucosein subjects 1, 2, and 6 suggests smallintestinal colonisation. Subject 2 achievedclinical and radiological remission afterthree weeks' oral tetracycline, basalbreath H2 returning to normal.These results lend support to the

hypothesis that gas production by bacteriamay be an important aetiological factorin PCI.

Reference'Yale, C. E., Balish, E., and Wu, J. D. (1974).

Archives ofSurgery, 109, 89.

Duodenal iron uptake in vitro: studies iniron overloaded subjects

T. M. COX AND T. J. PETERS (introduced by

DR V. S. CHADWICK) (Department ofMedi-cine, Royal Postgraduate Medical School,London) A method for determininginitial rates of iron uptake into humanduodenal biopsies has been developed'.Uptake of iron in normal subjects wassaturable over the concentration range18-450,umol/l and was shown to have thefeatures of active transport.

Because there is evidence that idio-pathic haemochromatosis is caused byinappropriate intestinal absorption ofdietary iron, the kinetics of uptake werestudied in duodenal biopsies taken frompatients with both primary and secondaryiron overloaded. Five patients withidiopathic haemochromatosis (two un-treated; three partially treated-meantime from venesection 34 ± 6 months)and five patients with iron overloadsecondary to thalassaemia major(two), alcoholic cirrhosis (two), andhereditary spherocytosis (one), werestudied.Compared with 11-15 controls, uptake

at 18, 90, and 180 ,umol/l medium ironconcentrations was increased up to 85%in haemochromatosis (p < 0-005,p < 0 025, P < 0 05, respectively), in-dicating an increased tissue affinity foriron. In contrast, no significant abnorma-lity in the kinetic behaviour of ironuptake by biopsies from patients withsecondary iron overload was found.These data indicate that there is a

specific abnormality of the mucosal ironcarrier in idiopathic haemochromatosis,which is not a consequence of iron over-load.

Reference'Cox, T. M.. and Peters, T. J. (1977). Gut, 18,A961.

Transnasal bile duct catheterisation afterendoscopic sphincterotomy: a method forbiliary drainage, perfusion, and sequentialcholangiography

P. B. COTTON, P. G. J. BURNEY, AND R. R.MASON (Gastrointestinal Unit, TheMiddlesex Hospital, London) Most duc-tal stones pass spontaneously afteradequate sphincterotomy, but repeatendoscopic cholangiography is necessaryto prove this. We have evolved a methodwhich reduces the need for repeat endo-scopy, and has other advantages. Im-mediately after sphincterotomy a fineTeflon tube is passed deep into the biliarysystem and preferably looped in the com-mon hepatic duct. The endoscopeisremov-



j.com/



ownloaded from

http://gut.bmj.com/


ed and the proximal end of the tube is re-routed via the nose. Tubes have beenmaintained in 26 out of 35 attempts, andhave been well tolerated for periods of upto two weeks. Immediately after sphincter-otomy the tube allows free bile drainageand cultures. A transnasal cholangiogramis taken after two days, and repeated ifnecessary after three days' saline perfusionof the duct. If stones are still presentat this time, attempts are made to removethem endoscopically, or to dissolve ordislodge them with further perfusions.

Will ursodeoxycholic acid (UDCA) replacechenodeoxycholic acid (CDCA) as themedical treatment of choice for gallstonedissolution ?

G. WILLIAMS, P. N. MATON, G. M. MURPHY,AND R. H. DOWLING (GastroenterologyUnit, Guy's Hospital and Medical School,London) Both CDCA and UDCA de-saturate bile and dissolve cholesterol-rich gallstones but, despite more difficultsynthesis and greater cost, UDCA stillmay have advantages over CDCA. Tostudy this, in 120 patients with radiolucentgallstones in functioning gall bladderstreated with CDCA and 22 similar patientsgiven UDCA, we compared dose, efficacy,symptoms, and side-effects.

Biliary cholesterol saturation indicescorrelated with bile acid dose (mg kg-l-day-' for CDCA (r = 043; P < 001)and UDCA (r = 0 855; P < 0 001). Mini-mum doses required to desaturate bilewere 4 0 mg kg-' (CDCA) and 2-5 mg kg-'(UDCA) and to desaturate bile con-sistently-14-5 mg CDCA (excludingobese' and resistant' patients) and9-5 mg UDCA. Mean doses required todissolve gallstones were 14-2 mg CDCAand 8-0 mg UDCA.

After six months' treatment with mixedUDCA doses (5-15 mg kg-' day-') sevenof the eight patients (88 %) showedpartial or complete gallstone dissolutioncompared with 27 of 35 (77 %) given.,r 13 mg CDCA kg-' day-'.Dose-related diarrhoea3 occurred in

36 of 84 (43%) treated with CDCA butin only one of 22 (4%) given UDCA.

Dose-related hypertransaminaseaemiawas seen in 33 % ofCDCA-treated patientsbut never occurred with UDCA.We conclude that UDCA has com-

parable efficacy to CDCA, is effectivein two-thirds of the dose, and does notcause diarrhoea. If long-term absence ofdiarrhoea, hypertransaminasaemia, and

toxicity are confirmed, UDCA seemslikely to replace CDCA.

References'Iser, J. H., Dowling, R. H., Mok, H. I. Y., and

Bell, G. D. (1975). Chenodeoxycholic acidtreatment of gallstones. A follow-up reportand analysis of factors influencing response totherapy. New England Journal of Medicine,293, 378-383.

2Iser, J. H., Murphy, G. M., and Dowling, R. H.(1978). Resistance to chenodeoxycholic acidtreatment in obese patients with gallstones.British Medical Journal, 1, 1509-1512.

3Maton, P. N., and Dowling, R. H. (1978). Hepaticcholesterol synthesis in cholelithiasis: role ofHMGCoA reductase in response to, and resist-ance to medical treatment. European Journal ofCliniccl Investigation, 8, (Abstr.) In press.

Serum methionine: a biochemical measureof the severity of liver failure

A. J. KNELL, D. J. COLLINS, AND R. ROBINSON(Warwick Hospital, Warwick) Lack of asimple test of the severity of liver failure(like the blood urea in kidney failure)hinders the selection of patients forartificial liver support. Coma is theindication for treatment, although allow-ing the disease to progress so far reducesthe chances of success. We have developeda simple automated assay of serummethionine to assist in patient evaluation'.Our reference range is 20-60 unmol/l. Wehave used the test routinely for over ayear.Serum methionine concentrations in-

crease rapidly as liver failure develops.Values above 300 umol/l are associatedwith encephalopathy. The onset anddisappearance of encephalopathy cor-related with the rise and fall of serummethionine during acute liver failure intwo patients who recovered. Valuesare low (<20,umol/1) during the first 24hours after paracetamol poisoning, thenincrease as liver damage progresses. Noincreases occur in simple hepatitis or ob-structive jaundice. Serum methionine ismoderately increased (100-200,umol/l) insevere kidney failure. We attribute thisto product inhibition of methioninemetabolism by creatinine.Measurement of serum methionine adds

precision to assessing patients in liverfailure, and permits early identificationof cases who may need artificial liversupport.

Reference'Collins, D. J., Robinson, R., and Knell, A. J.

(1978). Clinica Chimica Acta. In press.

The case for high dose antacid therapyin duodenal ulcer

R. A. KEENAN, R. H. HUNT, DIANA VINCENT,B. WRIGHT, AND G. J. MILTON-THOMPSON(Royal Naval Hospital, Haslar, Gosport,Hampshire) High dose antacid therapyaccelerates the healing of duodenal ulcer'.Using a technique previously described2we have studied such a regimen over 24hours in 11 patients with duodenal ulcersin remission. Patients received 30 mlmagnesium and aluminium hydroxideliquid (Wyeth) with 140 mmol bufferingcapacity to pH 3-0, or placebo, one andthree hours after main meals and onretiring. Gastric juice was sampled hourlythroughout the study, measured for pHand returned to the stomach.

Antacid caused a highly significantreduction in intragastric acidity betweenmeals but the effect was less after mid-night. Mean 24 hour intragastric hydrogenion activity was 57-8 + SEM 2-9 mmolon placebo and antacid reduced this by50% to 29-1 ± 2-6 mmol (p<0001).From midnight to 0700 activity wasreduced by 31% from 81-9 ± 4-5 mmol(placebo) to 56 5 + 4-8 mmol (antacid)(p < 0-001).

Cimetidine 1 g a day was studied in aseparate group of 15 duodenal ulcerpatients. Mean 24 hour intragastrichydrogen ion activity was reduced by 50%from placebo and nocturnal activity by66%.High dose antacid therapy appears as

effective as histamine H2-receptor antag-onists in reducing intragastric acid over24 hours but has less effect on nightsecretion.

References"Peterson, W. L., Sturdevant, R. A. L., Frankl,H. D. et al. (1977). Healing of duodenal ulcerwith an antacid regimen. New England JournalofMedicine, 297, 341-345.

'Pounder, R. E., Williams, J. G., Milton-Thompson,G. J., and Misiewicz, J. J. (1976). Effect ofcimetidine on 24-hour intragastric acidity innormal subjects. Gut, 17, 133-138.

Trial of plasmapheresis in patients withCrohn's disease

G. HOLDSTOCK, A. FISHER, C. LOEHRY, ANDT. HAMBLIN (Royal Victoria Hospital,Bournemouth) Immune complexes havebeen found in patients with chronicinflammatory bowel disease, particu-larly during relapse. It has been suggestedthat these complexes may be an importantfactor in the pathogenesis of the disease,especially the extragastrointestinal com-



j.com/



ownloaded from

http://gut.bmj.com/


plications. Plasmapheresis has been shownto be effective at removing immunecomplexes, with clinical improvementin patients with SLE-a disease knownto be associated with high levels ofcirculating immune complexes.With these facts in mind, we have

performed a pilot study of plasmapheresisin patients with Crohn's disease. Onenewly diagnosed patient and five withlong-standing disease were studied.Patients were selected because of theirunacceptable steroid requirements.

Patients were studied for up to six-months. The mean steroid requirementsfor the five long-standing cases wasreduced from 17-5 to 2-5 mg prednisolonedaily. Thirteen clinical relapses werecontrolled by plasmapheresis alone andall patients remained well. All had highlevels of circulating immune complexes,and these levels were reduced afterplasmapheresis. Levels of these complexescorrelated with the severity of patients'symptoms.

Plasmapheresis is worthy of futureinvestigation as a steroid-sparing agentin Crohn's disease and may give furtherclues to the pathogenesis of the disease.

CLINICAL MEDICAL/CLINICAL SURGICAL

Aetiological aspects of finger clubbing(FC) in inflammatory bowel disease (IBD)

G. KrTIS, H. THOMPSON, AND R. N. ALLAN(Nutritional and Intestinal Unit, andDepartment of Pathology, The GeneralHospital, Birmingham) The aetiology ofFC, commonly associated with IBD, isunknown. Suggested aetiological factorsinclude the extent of disease activity' thedegree of fibrosis in tumours2, and reflexmechanisms of which the vagus nerve maybe the afferent pathway.We assessed FC by objective measure-

ment of nail profile and hyponychialangles3. The mean profile angle was 166-50(±4 5) and the mean hyponychial angle1780 (±5 0) in 118 healthy volunteers, 51of whom had duplicate measurementsmade.

Finger clubbing (defined as one or bothangles > 1-65 SD from the control mean)was present in 40% of patients withCrohn's disease (75/189 patients), 15%with ulcerative colitis (15/101), and 8%with proctitis (2/25).

Finger clubbing was significantly

associated with disease activity assessedby laboratory indices (p < 0001), thedegree of fibrosis in resected specimens(p < 0.001), and macroscopic disease inareas of the gut innervated by the vagusnerve (p < 0-001). FC was significantlyassociated with the severity assessed bythe need for surgical intervention (p <0-001) and corticosteroid therapy (p <0 01). There was no correlation betweenFC and duration of disease.These data have identified that disasee

activity, severity of disease, and degree offibrosis and vagal innervation are im-portant in the pathogenesis of FCassociated with IBD.

ReferencesFielding, J. F., and Cooke, W. T. (1971). Fingerclubbing and regional enteritis. Gut, 12, 442-444.

'Yacoub, M. H., Simon, G., and Ohnsorge, J.(1967). Hypertrophic pulmonary osteoarthro-pathy in association with pulmonary metastasesfrom extrathoracic tumours. Thorax, 22,226-231.

3Bentley, D., and Cline, J. (1970). Estimation ofclubbing by analysis of shadowgraph. BritishMedical Journal, 111, 43.

Functional bowel disorder: a new perspec-tive

W. GRANT THOMPSON AND K. W. HEATON(University Department of Medicine,Bristol Royal Infirmary, Bristol) Although30-50% of patients referred to gastro-enterologists suffer from functional dis-orders, their prevalence in the communityis unknown.A questionnaire about functional

gastrointestinal symptoms was admini-stered to 301 apparently healthy Britishsubjects in young, middle-aged, andelderly categories. Abdominal pain, afeeling of incomplete evacuation afterdefaecation, urgency, scybala, runnystools, straining at stool, borborygmi,distension, heartbum, and laxative usewere very common.The spastic colon syndrome occurred in

13-6% and was identifiable by clusteranalysis. These subjects had abdominalpain relieved by defaecation and signifi-cantly associated with mucus, feeling ofincomplete evacuation, urgency, disten-sion, and more frequent, looser stoolsfollowing pain onset. Dyspepsia con-sisting of pain, usually in the upperabdomen, not relieved by defaecation,and associated with heartburn, occurredin 7%. A further 3-7% had painlessdiarrhoea characterised by frequent runnystools and a strong association withurgency. Six per cent suffered painless

constipation defined as frequent strainingat stool and significantly associated withscybala, less frequent bowel movements,mucus, feeling of incomplete evacuation,distension, and increased laxative use.

Thus, four distinct functional bowelsyndromes existed in at least 30% of sub-jects who were not patients. Studies offunctional gastrointestinal disorder musttake into account the large number ofuncomplaining sufferers in the community.

Therapeutic trial of Ativan, Buscopan, andFybogel in the irritable bowel syndrome

J. A. RITCHIE AND S. C. TRUELOVE (Nuf-field Department of Clinical Medicine,Radcliffe Infirmary, Oxford) A double-blind controlled therapeutic trial offactorial design has been used to test thetherapeutic effects of lorazepam (Ativan),hyoscine N-butyl bromide (Buscopan),and ispaghula husk (Fybogel) in 12randomised blocks of eight patients withthe irritable bowel syndrome. Therapeuticsuccess implied a definite symptomaticimprovement within one month, main-tained over the following two months.Failure implied either a lack of improve-ment or relapse within the three months'duration of the trial.Each of the three agents conferred some

benefit but the only one to give results thatwere significantly better than placebo wasFybogel. There was no evidence of inter-action between them. When the eightpossible combinations of treatment wereanalysed, no successes occurred among the12 patients who received only dummy pre-parations of the three agents. Sevensuccesses occurred among the 12 whoreceived potent preparations of all three.Groups receiving one or two of the potentpreparations obtained intermediate scores.These results show that the effects of

drugs used in treating the irritable bowelsyndrome can be additive. Other com-binations deserve to be similarly studiedif the efficacy of medical treatment is to beimproved.

Colonic bacterial activity, bile cholesterolsaturation, and the pathogenesis of gall-stones

T. S. LOW-BEER AND SHEILA NUTrER (SellyOak Hospital, Birmingham) This studyinvestigates how colonic metabolism ofbile acids affects biliary bile acid com-position and cholesterol saturation.



j.com/



ownloaded from

http://gut.bmj.com/

A976

Eleven healthy men ingested metroni-dazole 2 g daily for 10 days to inhibitanaerobic bacteria. Fasting gall-bladderbile was aspirated from the duodenumbefore, at the end of 10 days, and onemonth after stopping metronidazole. Therelative concentrations of cholate, cheno-deoxycholate, and deoxycholate weremeasured, and also those of phospholipidand cholesterol. Bile cholesterol satura-tion was then calculated, and the two-tailed paired t test applied.

Metronidazole decreased deoxycholatefrom a mean of 24% to 7% returning to22% a month later. Chenodeoxycholatechanged from 33% to 46% and back to33% (p < 0-001 throughout), cholatefrom 44% to 48% to 45% (p < 0-005and not significant respectively). Bilecholesterol saturation fell in 10 of the 11subjects, from a mean of 100 to 0-83 andthen increased in all to a mean of 111

after discontinuing metronidazole (p <0.01 and < O-COI respectively).

This study demonstrates that deoxy-cholate absorption selectively reduceschenodeoxycholate concentrations in bile,previously shown to result from sup-pression of chenodeoxycholate synthesisin the liver. With decreasing deoxycholateabsorption bile cholesterol saturation falls,with increasing absorption cholesterolsaturation rises, also shown previously.Diets known to decrease the absorptionand recirculation of deoxycholate (suchas high-fibrediets)should reduce the risk ofgallstones.

Observations on the treatment of primarypneumatosis coli

S. HOLT AND R. C. HEADING (UniversityDepartment of Therapeutics and ClinicalPharmacology, The Royal Infirmary,Edinburgh) Successful treatment ofpneumatosis coli by high flow oxygentherapy has been reported but there is noagreement on the optimum procedure. Wehave treated five patients with primarypneumatosis coli, using high flow oxygentherapy in a standardised manner withsymptomatic and radiological resolutionof disease. In each case, the extent ofdisease and its response to therapy weredefined endoscopically and radiographi-cally. Bacteriological examination of thepatients' stool and colonic mucosa re-vealed no qualitative or quantitativeabnormalities and the flora was notaltered to any great extent by oxygentherapy. Clinical and radiological re-

mission has been maintained in all patientswith follow-up periods ranging from sixmonths to three years.

Intermittent high flow oxygen therapyprovides a safe and comfortable form oftreatment for patients with symptomaticpneumatosis coli. Its success is apparentlynot dependent on eradication of anaerobicbacteria from faeces or colonic mucosa.

Adenocarcinoma in Barrett's oesophagus:a report of 16 cases

J. DEES, M. VAN BLANKENSTEIN, AND M.FRENKEL (introduced by J. R. Bennett)(Erasmus University Rotterdam, Depart-ment of Internal Medicine II, Rotterdam)Barrett's oesophagus (BO) may be a pre-malignant lesion. Only about 40 cases ofadenocarcinoma in BO have been pub-lished. Since 1974 we have diagnosed 16cases at endoscopy. Endoscopic criteriawere (1) histologically confirmed normalgastric mucosa above the tumour; (2) notumour in the cardia region of thestomach.

In a group of 5300 patients undergoingupper gastrointestinal endoscopy between1974 and 1978, 71 cases of primaryoesophageal carcinoma were seen, 44(62%) squamous cell- and 27 (38%)adenocarcinomas. The relatively lowpercentage of squamous cell carcinoma isthe result of patient selection. The 27adenocarcinomas form an unselectedgroup, 16 of these (59%) were found inBO. The 71 cases of primary oesophagealcarcinoma are part of a group of 161patients referred to a regional cancercentre. Therefore, in this unselectedgroup, at least 16 (10 %) of the oesophagealcarcinomas were found in BO. This per-centage should be contrasted with a 0-6 %prevalence of BO in the 5300 uppergastrointestinal endoscopies. These find-ings support the hypothesis that BOentails an increased risk of oesophagealcarcinoma.

Review of the results of dilatation in themanagement of benign oesophageal stric-tures in a surgical clinic

F. H. YOUNAN AND C. W. VENABLES (New-castle University Hospitals, Newcastleupon Tyne) The management of benignpeptic oesophageal strictures remains con-troversial. Since 1972 we have adopted aninitial conservative approach to thisproblem using endoscopic dilatation. This


paper presents a review of our materialand the results of management.Between 1972-77, 47 patients (32 male)

have had endoscopically confirmed be-nign oesophageal strictures under ourcare. A total of 133 dilatations have beenperformed using the following techniques:(1) forcible dilatation with endoscope(38), (2) Eder Peustow (42), (3) blindbougie (53), with only two perforations(1.3 %). One dilatation alone wassuccessful in relieving symptoms in 23patients but 15 needed three or moredilatations. Sixteen patients have hadsurgical management after dilatationeither for coexistent duodenal ulceration,hiatus hernia, or failure of conservativemanagement.

In conclusion, this study demonstratedthat dilatation is a safe method of man-agement. The Eder Peustow dilatorappears to be the best if the endoscopealone is unsuccessful. With this conserva-tive approach surgical intervention can beavoided in many patients.

Perforation of duodenal ulcers in theelderly

M. R. THOMPSON (Royal Devon andExeterHospitals (Wonford and Heavitree),Exeter) In a six-year review of duodenalperforation in Exeter it was found that40% (36 out of 89) of duodenal perfora-tions occurred in patients over the age of70 years and approximately 50% of thesewere women. Perforation time (PT) waslonger in the over 70s (over 70, PT =25-1 h + 4-3; N = 26; cf under 70, PT =12-7 h + 1-7; N = 47 p < 0 005). Themortality rate in patients over the age of70 years was 35% compared with 4% forthe under 70s. The major factors in thecause of death in the over 70s was PT andchronicity of the ulcer. In patients over 70years with a PT of > 18 h there was a 54%mortality rate compared with 8% inpatients with a PT of < 18 h. Seven of thenine elderly patients who died developedsevere complications of a persistingchronic ulcer-namely, bleeding, pyloricstenosis, and perforation-before dyingthree to 43 days after admission.

It is suggested that, although simplesuture is adequate in elderly patients whohave perforation for less than 18 hours,paradoxically those that have perforatedfor more than 18 hours and who have achronic ulcer require more extensivesurgical procedures.These data show that there is room for



j.com/



ownloaded from

http://gut.bmj.com/


improvement in the management of thiscondition and that better results can beachieved only by earlier diagnosis and achange in the surgical management.

Intraoperative tests for completeness ofvagotomy

M. H. THOMPSON, R. C. N. WILLIAMSON,P. W. DAVIES, W. K. ELTRINGHAM, ANDD. JOHNSTON (Department of Surgery,Bristol Royal Infirmary, Bristol) Theresults of peptic ulcer surgery would beimproved if all vagotomies were com-plete, but this is likely to be achieved onlyby using an intraoperative test. Two testshave been evaluated: the Grassi, usingpentagastrin stimulation and an intra-luminal pH electrode, and the Burge,using direct nerve stimulation and intra-gastric pressure measurement.

In phase 1, 25 patients had highlyselective vagotomy for duodenal ulcer.After completion of vagotomy both testswere performed and no further dissectionwas done. One week post-operativelygastric secretion was measured. Hol-lander's criteria, peak acid output (insulin)and peak acid output (pentagastrin) wereused to measure completeness of vago-tomy. The Burge test gave misleadingresults. (All Hollander positive patientswere in the Burge 'negative' group).Patients with negative Grassi test hadsignificantly lower postoperative secretionthan those with positive Grassi test (P =0.048).

In phase 2, if the Grassi test waspositive, more vagal nerve fibres were cut,and postoperative secretion measured asin phase 1. All patients were Hollandernegative, and measured postoperativesecretion was not significantly greater thanin the Grassi negative patients in phase 1.We conclude that the Grassi test is help-

ful in producing complete vagotomy, butthat the Burge test is not.

Evaluation of a non-invasive insulin test

T. V. TAYLOR, H. SHARMA, B. R. PULLAN,AND H. B. TORRANCE (Department ofClinical Surgery, University of Edinburgh,andManchester Royal Infirmary) A directrelationship between acid output andintragastric accumulation of B9mTcl,2 hasled to the development of a non-invasivemethod of assessing gastric function3.The possibility of performing an insulintest non-invasively has been investigated

in 39 patients. After an overnight fast99mTc (1 mCi) was given intravenouslyand 40 minutes later insulin 0-2 u/kg IV.Five minute y-camera computer frameswere taken for 120 minutes. Under basalconditions a slow and steady accumula-tion of intragastric activity occurs, but ifa secretory response to insulin is obtaineda sharp rise in activity above the predictedcourse of the curve is seen, as the stomachis the only organ in the field of studycapable of concentrating the isotope.A quantitative sharp rise in the intra-

gastric activity was obtained in the firsthour after insulin in 18 of 21 patients withan intact stomach. Of 18 patients who hadundergone recent vagotomy and whowere insulin negative by routine testing, 17failed to show a sharp rise in intra-gastric activity after insulin. Of the 18insulin positive results some gastricemptying was seen on the scan in 15,whereas of the 17 insulin negative patientsonly four had observed gastric emptyingof 99mTc

References'Taylor, T. V., Pullman, B. R., Elder, J. B., and

Torrance, B. (1975). Observations of gastricmucosal blood flow using S9mTc in rat and man.British Journal of Surgery, 62, 788-791.

2Wine, C. R., Nahrwold, D. I., Rose, R. C., andMiller, K. S. (1976). Effect of histamine onS9mTc excretion by gastric mucosa. Surgery, 80,591-594.

3Taylor, T. V., Bone, D., and Torrance, B. (1977).A non-invasive test of gastric function. BritishJournal of Surgery, 64, 702-708.

CLINICAL SURGICAL/G I HORMONES

Preliminary results of a modified Hill-Larrain procedure for gastroesophagealreflux

E. AMDRUP, M. 0STER, P. FUNCH-JENSEN, ANDA. CSENDES (introduced by Rory McCloy)(Surgical-gastroenterological Dpt., Kom-munehospitalet, Aarhus University, Den-mark) Thirty-six patients suffering fromfailure of at least two years' medicaltreatment were operated on for gastro-oesophageal reflux (heartburn+regurgita-tion). Four had additional duodenal ulcer,three were previously operated on forduodenal ulcer, and two for refluxoesophagitis. All had a positive acid refluxtest, eight also had radiological reflux.All had histological and 12 also macro-scopic signs of oesophagitis. Four of thepatients had fibrotic stenosis with dys-

phagia.The operative steps were: (1) parietal

cell vagotomy; (2) cardioplasty with onenon-absorbable suture; (3) calibrationwith a 12 mm bougie; (4) gastropexy tothe median arcuate ligament.One patient died postoperatively from

oesophageal tear. One had a postopera-tive subphrenic abscess drained. Theother operations were uncomplicated.

Observation at two to 20 months:initial dysphagia subsided in one to twomonths. All were relieved from symptomsof gastroesophageal reflux. In one apositive acid reflux test could be inducedand in another radiological reflux wasstill observed. In the rest exam;nationswere negative.Adding of parietal cell vagotomy

reduces gastric acid secretion, providesbetter technical access, and securesagainst unintended vagotomy.

Importance of an intact, innervated pyloricantrum in the control of gastric emptyingof liquids in man

C. M. WHITE, VALERIE POXON, M. R. B.KEIGHLEY, AND J. ALEXANDER-WILLIAMS(The General Hospital, Birmingham)Studies in dogs' suggest that gastricemptying of liquids is controlled by thetone of the body and fundus of thestomach; the antrum and pylorus appearunimportant. In man the gastric outletdoes seem important2. The faster empty-ing and the diarrhoea and dumping aftervagotomy and pyloroplasty may be causedby the pyloroplasty or the denervationof the antrum as well as the body andfundus.We have examined the separate effects

of pyloroplasty and vagal denervation.Thirty-three patients with uncomplicatedduodenal ulcer were randomly allocatedto one offour operations: proximal gastricvagotomy (PGV) N = 10; proximalgastric vagotomy and pyloroplasty(PGV + P) N = 8; total gastric vagotomy(TGV) N = 7; total gastric vagotomyand pyloroplasty (TGV + P) N = 8. Theemptying of 10% dextrose was studiedbefore and after operation3.

Early passive emptying was significantlyfaster than preoperatively after alloperations (PGV P < 0-01; PGV + Pp < 0-025; TGV P < 0-01; TGV + PP < 0-025-paired t test). Later emptyingwas unchanged after PGV but wassignificantly faster after all the otheroperations (PGV + P P < 0.025: TGV



j.com/



ownloaded from

http://gut.bmj.com/

A978 The British Society of Gastroenterologyp < 0001; TGV + P P < 0005-pairedt test).Our results indicate that, although

proximal gastric denervation causes fasterearly emptying, the innervated intactterminal antrum and pylorus are import-ant during the later emptying of liquids.

References'Wilbur, B. G., and Kelly, K. A. (1973). Effect

of proximal gastric, complete gastric and truncalvagotomy on canine gastric electrical activity,motility and emptying. Annals of Surgery,178, 295-303.

'McKelvey, S. T. D. (1970). Gastric incontinenceand post-vagotomy diarrhoea. British Journal ofSurgery, 57, 741-747.

'Donovan, I. A. D. (1976). The different com-ponents of gastric emptying after gastric surgery.Annals of The Roydrl College of Surgeons ofEngland, 5, 368-373.

Why do patients die after surgical treat-ment for peptic ulcer haemorrhage evenwhen diagnosis is established?

C. W. VENABLES (Newcastle UniversityHospitals (NGH and RVI), Newcastleupon Tyne) The value of emergencyendoscopy in the management of upperGI bleeding is being questioned becauseof a failure to demonstrate changes inmortality. Clearly endoscopy cannotaffect the outcome unless it modifiesnianagement. Since 1972 emergency endo-scopy has been attempted in all cases ofupper GI bleeding coming to surgery inthe two hospitals in which I work. Thispaper presents an analysis of 113 casestreated surgically between 1972-78 withan analysis of various factors that mightbe related to mortality.

In Hospital A, 73 patients were operatedupon with an overall mortality of 8-2%,while in hospital B, the mortality was25 %. Various factors have been analysed,including age; sex; diagnosis; delaysbefore admission, endoscopy, and opera-tion; surgical expertise, length, and tech-nique; blood transfused; other diseasespresent. The analysis demonstrates thatthe most important factors are delays inadmission and the presence of coexistentdiseases.

It is concluded that comparisonsbetween mortality figures should be madeonly if all other factors are comparable.

Failure

W. P. SMALL (Gastric Follow-Up Clinic,Western General Hospital, Edinburgh) Astrangely neglected index of failure of

surgery for peptic ulcer is that the patientkeeps coming back'. On whom does thisclinical load fall? What is the cost? Canfailure be avoided?The records of 25 'good' results and 25

'bad' results2 randomly selected from theGastric Follow-Up Clinic have been com-pared in an attempt to answer thesequestions.The consequences of surgical failure

weigh most heavily on the laboratories, theradiologists and the endoscopists. Re-operation is not so large a burden.

Specialised units attract failures andwith them disproportionate demands oninvestigative techniques. The need torecognise this work load in terms offacilities and staffing is obvious.The investigation of failure costs at

least £500, to which must be added thecosts of inpatient management, work loss,and social security.

Failure is potentially avoidable. It stemsfrom misjudged operation in patients withno clear indications for surgery and fromthe complications of radical surgery. Toinvest in the avoidance of failure by betterinvestigation, better selection, lesseroperations, and, indeed, fewer operationsmakes sense.The trend towards proximal vagotomy

and the use of cimetidine as an alternativeto surgery is to be encouraged.

References'Johnstone, F. R. C., and Holubitsky, I. B. (1967).

Post-gastrectomy problems in patients withpersonality defects: The 'albatross' syndrome.Canadian Medical Association Journal, 96,1559-1564.

2Small, W. P. et al. (1978). A questionnaire forassessment of the result of peptic ulcer surgery.Digestion. (In press).

Is it necessary to add neomycin to metro-nidazole for oral prophylaxis in colonicoperations?

S. VALLANCE, B. JONES, M. R. B. KEIGHLEY,J. ALEXANDER-WILLIAMS, AND Y. ARABI(The General Hospital, Birmingham) Pro-phylactic neomycin (N) and metroni-dazole (M) will reduce postoperativesepsis' but it has been suggested thatprophylactic M alone may abolish aerobicand anaerobic sepsis2. However, oral Mhas no influence on colonic microflora.3The aim of this study has been to com-

pare the influence ofM alone with that ofM and N on the incidence of sepsis afterelective colonic resections. So far 52patients have entered a double blindplacebo controlled study: 26 patients

received M alone (200 mg tds) and anequal number received M (200 mg tds)and N (1 g tds) for two days preopera-tively. The groups were similar withrespect to age, sex, and method of bowelpreparation.Wound sepsis occurred in 50% of

patients receiving M alone compared withonly 23% after M and N. The incidenceof perineal sepsis, septicaemia, andanastomotic dehiscence was similar in thetwo groups. The overall incidence ofseptic complications was 57 6% after Malone and only 38-4% after M and N.The preliminary results of this study

indicate that neomycin should be added tometronidazole for effective antimicrobialprophylaxis in bowel surgery.

ReferencesMatheson, D., Arabi, Y., Baxter-Smith, D., and

Keighley, M. R. B. (1977). Multicentre studyof bowel preparation for colorectal cancer.British Journal of Surgery, 64, 839.

'Ingham, H. R., Sisson, Penelope, Tharagonnet,Danka, Selkon, J. B., and Codd, A. A. (1977).Inhibition of phagocytosis in vitro by obligateanaerobes. Lancet, 12, 1252-1254.

3Arabi, Y., Dimock, Frances, Budon, D. W.,Alexander-Williams, J., and Keighley, M. R. B.(1977). Influence of neomycin and metro-nidazole on faecal flora in healthy volunteers.Gut, 18, A969.

Variation in antral gastrin concentrationand the effect of gastritis

R. J. MCFARLAND, J. S. SLOAN, AND K. D.BUCHANAN (The Queen's University ofBelfast, Northern Ireland) Antral gastrinconcentration has been studied by endo-scopic biopsy in 34 patients with non-ulcer dyspepsia, 19 with acute gastritis, 19with duodenal ulcer, and six with per-nicious anaemia. Three adjacent antralbiopsies were taken from each patient.and buffer extract assayed for gastrin.Mucosal gastrin concentration wasmarkedly higher in pernicious anaemia(159 ± 65 ng/g, M ± SE), compared withnon-ulcer dyspepsia, (35 + 6 ng/g), acutegastritis (43 ± 14 ng/g), and duodenalulcer (18 ± 4 ng/g). A striking feature wasthe marked variation within biopsies fromthe same patient confirming previousreports of uneven gastric mucosal con-centration. To determine the effect of localmucosal pathology on gastrin cell popula-tion, a separate study on 35 gastrectomyspecimens was carried out using im-munofluorescent and immunoperoxidasetechniques. The numbers of detectableimmunoreactive gastrin cells diminishedwith increasing severity of gastritis and



j.com/



ownloaded from

http://gut.bmj.com/


very few were detected in the presence ofintestinal metaplasia.lThus it appears that the presence and

severity of antral gastritis is one importantfactor in the marked variation in gastrincell numbers and mucosal concentrationseen in small biopsies and should be takeninto account in their assessment.

Simple technique for the demonstration ofpolypeptide-secreting endocrine cells inhuman small intestinal biopsies

C. SHAW, J. S. SLOAN, AND P. TITTERINGTON(introduced by K. D. Buchanan)(Departments of Medicine and Pathology,Queen's University of Belfast) Presentmethods for the demonstration of poly-peptide-secreting endocrine cells occurringin the mucosa of human small intestinerequire elaborate and tedious fixationtechniques which are not applicable to theroutine laboratory.We have developed a simple technique

using a liquid phase fixative which hasresulted in the easy identification of poly-peptide-secreting endocrine cells in humansmall intestinal mucosa obtained atbiopsy. These include cells containinggastric inhibitory polypeptide, vasoactiveintestinal polypeptide, gastrin, somato-statin, secretin, and glucagon immuno-reactive materials. Demonstration may beachieved using either immunofluorescentor immunoperoxidase techniques. Thetissue morphology is well preserved, thuspermitting routine histological assessment-a factor of considerable importancewhen small intestinal biopsies are underanalysis for clinical purposes.The fixative procedure will allow the

assessment of the role of small intestinalendocrine cells in a variety of disorders,and has been successfully used to demon-strate polypeptide-secreting endocrine cellsin other tissues including antrum, pancreasand large bowel.

Gastrointestinal hormone patterns asso-ciated with interdigestive and postprandialmotor activity in the human proximalbowel

W. D. W. REES, J.-R. MALAGELADA, ANDV. L. W. GO (introduced by S. F. Phillips)(Gastroenterology Unit, Mayo Clinic,Rochester, Minnesota, USA) We havecharacterised interdigestive and post-prandial motor pattems in the humanproximal bowel'. In the present study,relationships among these motor pattems

and blood levels of gastrointestinal hor-mones were examined. Intraluminalpressure was recorded from the antrum,duodenum and jejunum of six healthysubjects using intraluminal transducers.On the first day, fasting motility and bloodlevels of gastrin, motilin, glucagon, andgastric inhibitory polypeptide weremeasured at 15 minute intervals for eighthours. On the second and third days, fourdifferent liquid meals (carbohydrate; pro-tein; lipid; combined) were randomlyadministered (two each day) during phaseII interdigestive activity and levels ofgastrin and motilin were measured.

Fasting gastrin, glucagon, and gastricinhibitory polypeptide remained low andconstant, while motilin showed cyclicvariation (range: 138 ± 21 to 523 ± 96pg/ml). Motilin peaks were associatedwith antroduodenal phase III activity intwo subjects, but such a relationship wasabsent in the remaining subjects. Proteinand combined meals produced significantgastrin release (3200 ± 500 and 1400 +400 pg/60 min) and lipid caused significantmotilin release (1100 ± 400 pg/60 min).All meals inhibited interdigestive activity,the duration being significantly longer forlipid and combined meals'. Conclusionswere that (1) fluctuation in fasting motilincorrelates poorly with interdigestivemotility while other hormones show littlevariation; (2) feeding disrupts motorcycles, protein and combined mealsreleasing gastrin and lipid releasingmotilin.

Reference'Rees, W. D. W., et al. (1978). Gastroenterology,

74. 1083.

Bombesin evokes calcium-dependent amy-lase secretion in the isolated mousepancreas

0. H. PETERSEN (Department ofPhysiology, University of Dundee MedicalSchool) The tetradecapeptide bombesinevokes pancreatic secretion by a directaction on the plasma membrane of theacinar cells". The calcium dependence ofthis process and the possible interactionswith other secretagogues have been investi-gated using a sensitive on-line fluoro-metric method for determining amylaseoutput2 from isolated superfused segmentsof mouse pancreas.Bombesin (10-8 M) (maximally stimu-

lating concentration1) caused a marked,but only transient increase in amylase out-put during exposure of the tissue to

calcium-free solution containing EDTA(calcium chelator). Admission of calcium(50 ,uM to 2-6 mM) during sustainedmaximal stimulation caused an immediatedose (calcium) dependent increase in amy-lase secretion. In the presence of calciumthe secretory response to bombesin wassustained. Removal of calcium duringsustained bombesin stimulation caused animmediate abolition of the stimulant-evoked secretion. In the presence ofmaximal acetylcholine or caerulein stimu-lation, bombesin had no effect. In thecase of submaximal acetylcholine orcaerulein stimulation, bombesin (10-8 M)caused an increase in secretion up to themaximal level.Bombesin evokes enzyme secretion

through exactly the same calcium depen-dent mechanism3 as acetylcholine andcholecystokinin.

ReferencesIlwatsuki, N., and Petersen, 0. H., (1978)..Journalof Clinical Investigation, 61, 41-46.

'Matthews, E. K., Petersen, 0. H., and Williams,J. A. (1974). Anolytical Biochemistry, 58, 155-160.

3Petersen, 0. H., and Iwatsuki, N., (1978). Annalsof the New York Academy of Sciences, 307,599-618.

Measurement of trypsin concentration induodenal juice using a radioimmunoassay

G. LAKE-BAKAAR, S. MCKAVANAGH, C. E.RUBIO, AND J. A. SUMMERFIELD (introducedby Professor Dame Sheila Sherlock)(Department of Medicine, Royal FreeHospital, London) Trypsin output induodenal aspirate following pancreaticstimulation with either a Lundh meal orcholecystokinin-pancreozymin (CCK-PZ)and secretin remains a standard test ofpancreatic function. However the usualmethod of trypsin estimation, based onenzymatic hydrolysis of the substrate p-tosyl-L-arginine methyl ester-HCl(TAME) has several disadvantages, in-cluding autodegradation of the enzymewhich precludes storage before estimation.A radioimmunoassay (RIA) has been

described for the estimation of trypsineven in the presence of inhibitors such asTrasylol. The trypsin concentration induodenal aspirate has been determined byRIA in a total of 25 estimations and com-pared with the enzymatic method. Theeffect of Trasylol on the stability of trypsinin duodenal aspirate stored at - 70°C wasalso studied.

Enzymatic activity correlated well withthe mass of trypsin obtained by RIA (p <0 005, Student's t test.) No cross-reactivity



j.com/



ownloaded from

http://gut.bmj.com/


was shown with Trasylol. Storage ofduodenal aspirate led to rapid loss oftrypsin (mean 99% range 6-3-19%) atfour weeks. This was completely preventedby the addition of Trasylol (p < 0-05,Student's paired t test).We conclude that RIA determination of

duodenal juice trypsin concentration is auseful alternative method to enzymaticanalysis and should be performed when-ever prior storage of the enzyme isnecessary, since autodegradation can becompletely prevented with Trasylol.

LIVER AND BILIARY/OESOPHAGUS/STOMACH/DUODENUM

Effect of age on hepatic drug metabolismand on the response to enzyme induction

N. CARULLI, M. PONZ DE LEON, G. ROMANO,AND R. IORI (Istituto di Clinica Medica,Policlinico, Modena, Italy) Althoughage-dependence of drug metabolism is awidely held concept, only few reportsfocus attention on the impairment ofhepatic drug metabolism in human aging.In this study we evaluated the metabolismof three model drugs in the attempt todefine the role of the various determinantsof hepatic drug metabolism, such as drugmetabolising capacity, liver blood flow,and plasma protein binding of the drug.Two groups of subjects were studied:

(1) 21 elderly subjects (mean age 79 ± 7 9years) and (2) 21 young adults (meanage 27-2 ± 6-9 years). Liver function andother appropriate tests ruled out hepatic,renal, or metabolic dysfunctions. The sub-jects had not been medicated for at leastfour weeks before testing. In subgroups ofseven subjects for each age group, amino-pyrine, tolbutamide and propranololpharmacokinetics were evaluated. Tol-butamide and propranolol metabolismwas studied in the same subjects also afterphenobarbital treatment (100 mg/day forseven days).

In the elderly group the metabolic clear-ance of aminopyrine (67 5 ± 12 ml/m)and tolbutamide (9 3 ± 1-6 ml/m) wassignificantly lower (p < 0-01) than in theyoung group (106-7 ± 15-2 and 18-9 ± 3-2respectively). Propranolol clearance didnot differ in the two groups (108 ± 0-31/m versus 1*12 ± 0-41). In the elderlysubjects phenobarbital treatment pro-duced an increase of tolbutamide clear-ance (+70% of the basal value) greater

than in young subjects (+ 21 %). Thetreatment did not change the propranololclearance.We conclude that decreased hepatic

drug metabolism in elderly subjects seemsto be due mainly to a reduced activity ofthe hepatic drug metabolising system.

Alcohol reduces liver acetaldehyde dehydro-genase

W. J. JENKINS AND T. J. PETERS (introducedby v. s. CHADWICK) (Department ofMedicine, Hammersmith Hospital, London)Acetaldehyde, the first product of ethanolmetabolism, is more toxic than alcoholitself. Alcoholics display higher bloodlevels of acetaldehyde than non-alcoholicsgiven the same dose of ethanol', but thecause of this difference is unknown.Normally acetaldehyde is rapidly removedby oxidation to acetate in the liver.Acetaldehyde dehydrodenase, which cata-lyses acetaldehyde oxidation, has beenlittle studied in man, and there has beenno previous work on its activity or intra-cellular localisation in fresh human liver.We have developed an assay for acetal-

dehyde dehydrogenase, and determinedits specific activity in liver biopsies fromcontrols and non-cirrhotic alcoholics. Thespecific activity in alcoholics (43 4 ± 6-6mI/mg protein) was significantly less thanin controls (72-6 ± 8-7 m,u/mg protein)(p < 0-05). Intracellular localisationstudies showed that the decrease wasspecific to the cytosolic component of theenzyme, whose activity in the mito-chondria was normal.

Thus, chronic ethanol abuse isassociated with a reduced hepatic activityof the enzyme catalysing the removal ofits more toxic oxidation product-acetal-dehyde. This may explain the higher bloodlevels of acetaldehyde in alcoholics, andhas important implications in understand-ing the pathogenesis of alcoholic liverinjury.

Reference'Korsten, M. A., Matsazuki, S., Feinman, L.,and Lieber, C. S. (1975). High blood acetalde-hyde levels after ethanol administration. NewEngland Journal of Medicine, 292, 386-389.

Vitamin D metabolism in primary biliarycirrhosis and extrahepatic obstructivejaundice

R. T. JUNG, M. DAVIE, P. SIKLOS, T. M.CHALMERS, J. 0. HUNTER, AND D. E. M.

LAWSON (Addenbrooke's Hospital andMRC Dunn Nutrition Unit, Cambridge)Osteomalacia is frequently observed incholestatic liver disease. To study theimportance of bile duct obstruction invitamin D metabolism, we investigatedsix cases of acute extrahepatic obstructivejaundice (OJ) and eight cases of primarybiliary cirrhosis (PBC) (three supple-mented with vitamin D).

Plasma 25-hydroxy vitamin D (25-OHD) was low in PBC (unsupplemented)(5-2 + 4 5 ng/ml) (p < 0 05 of normal)but normal in OJ (8-4 ± 4 9). No case ofPBC showed histological osteomalacia.Dietary intake of vitamin D in PBC waslow (40 ± 24 IU), but response to ultra-violet light was normal. Plasma 25-OHD-binding protein (Gc) was low in PBC(223 ± 54 ug/ml) and OJ (235 ± 17 nr300-550 ug/ml). This reflects low serumalbumin, which correlates with Gc (r =0i78, p < 0-001).3HD3 half-life in OJ was normal but

longer in PBC in contrast to the short half-life in alcoholic cirrhosis (p < 0-05).

All OJ patients and five with PBC hadnormal 3H-25-OHD production. Threewith PBC (two supplemented, one afterUV light) produced less 3H-25-OHD thanexpected.

Urinary 3H after 3HD3 correlated withserum bilirubin (r = 0-84, p < 0 02). Pre-liminary work suggests that this is notderived from 25-OHD but from othervitamin D metabolites.Low 25-OHD results from poor intake,

inadequate UV exposure, and possiblyurinary vitamin D wastage. 25-hydroxyla-tion is normal in OJ and most PBC.

Impromidine (SK & F 92676)-a highlyspecific agonist for histamine H2 receptors

R. H. HUNT, J. A. BILLINGS, JANE G. MILLS, W.L. BURLAND, AND G. J. MILTON-THOMPSON(Royal Naval Hospital, Haslar, and TheResearch Institute, Smith Kline & FrenchLaboratories Ltd, Welwyn Garden City)Impromidine is a potent and specifichistamine H2-receptor agonist in animalsand in vitro.'We have studied the gastric secretory

response to impromidine in healthy malevolunteers using the phenol red correctiontechnique.2 A 60 minute intravenousinfusion of impromidine 10 jig kg-1 h-1was compared with pentagastrin 6 ugkg-1 h-1 and histamine acid phosphate 40yg kg-1 h-1 intravenously in 11 subjects.Comparisons were also made between



j.com/



ownloaded from

http://gut.bmj.com/


intramuscular pentagastrin 6 jug kg-1 andimpromidine 10 ug kg-' in 12 subjects.Analysis by Wilcoxon's paired rank sumtest showed no significant differences inpeak or maximal acid output betweensecretagogues nor in the time course of theresponse following either intravenous orintramuscular administration.

Cimetidine 2 mg kg-' h-1 intravenouslyover 30 minutes inhibited impromidinestimulated acid output by a mean of 65%in five subjects.The log-dose response curve to impro-

midine in five subjects was linear from 2.5to 20 ,ug kg-' h-1. Cimetidine 0 5 mg kg-lh-1 caused a highly significant parallelshift in the dose response curve, con-sistent with direct competitive antagonism.No untoward clinical, haematological, orbiochemical effects were observed.The evidence suggests that impromidine

is a specific H2 agonist in man and will beof value in the investigation of gastricsecretion.

References'Durant, G. J., Duncan, W. A. M., Ganellin,

C. R., Leslie, G. B., Parsons, M. E., Blakemore,R. C., and Rasmussen, A. C. (1978). Impromid-ine-(N- [3-(imidazol-4-yl) propyl ]-N'-{2- [(5-methyl imidazol-4-yl)-methyl thio ] ethyl}guanidine-a very potent and highly specificagonist for H, receptors (Submitted for publica-tion).

'Hobsley, M., and Silen, N. (1969). Use of an inertmarker (phenol red) to improve accuracy ingastric secretion studies. Gut, 10, 787-795.

Braking of stomach emptying: 'receptors'in duodenum or not?

0. LAWAETZ AND RINETTE ANDREASEN(introduced by M. HOBSLEY) (Institute forExperimental Research in Surgery, Uni-versity of Copenhagen, Denmark) Itseems to be well established that sugars,fat, and acids instilled into the duodenallumen inhibit gastric emptying'. However,the location of this braking mechanism tothe duodenum itself has never beenproved.To identify the site of the braking

'receptors', 62 gastric emptying studieswere performed on four healthy consciousdogs during continuous instillation ofdifferent fluid substances (distilled water,20% soya bean oil, and 25% glucose)separately into the duodenum and into thesmall bowel. All the dogs were preparedwith gastric, duodenal, and jejunal fistulae.A simple isotope technique was used forcontinuous evaluation of the gastricemptying pattern and for determination ofthe half time of emptying (Ti min).

Glucose and fat introduced into thestomach or into the duodenum did notinhibit gastric emptying (distilled water,Ti < 3 min; glucose and fat, Ti < 3 min).However, both fat and glucose instilledseparately into the jejunum did inhibitgastric emptying (distilled water, T* < 3min; fat, Ti > 60 min; glucose, Ti > 200min).The results show that in dogs the

braking 'receptors' lie in the smallintestine, distal to the duodenojejunaljunction, and that there is no brakingmechanism in the duodenum.

Reference'Hunt, J. N. (1975). Interactions of the duodenal

receptors which control gastric emptying.Scandinavian Journal of Gastroenterology, 10,Suppl. 35, 9-21.

Gastric cancer detection in gastric ulcerdisease

R. A. MOUNTFORD, P. BROWN, P. R. SALMON,C. S. NEUMANN, AND A. E. READ (Univer-sity Department ofMedicine, Bristol RoyalInfirmary, Bristol) In an attempt todetermine which clinical and diagnosticfeatures might increase the suspicion ofmalignancy in patients with gastriculceration, a retrospective survey of allpatients with gastric ulcer diagnosedendoscopically in our unit over a threeyear period has been performed. Two-hundred-and-sixty-five patients withgastric ulcer have been studied with anaverage length of follow-up of twoyears, and, of these, 37 (14%) provedmalignant. Presenting complaints ofanorexia, weight loss, nausea andvomiting, and multiple (>3) symptomswere significantly more common in themalignant group, but ulcer site and co-existing duodenal ulceration were un-helpful in determining the status of anulcer. Larger (> lcm) ulcers had a signifi-cantly higher risk of malignancy thansmall ulcers. If a definitive statement onthe nature of an ulcer was made, thenradiology was unreliable in distinguishingbenign from malignant ulcers (14% falsepositives; 40% false negatives). Anequivocal opinion on the nature of anulcer was common at both endoscopyand radiology, expecially if the ulcer wasbenign (endoscopy 56%, radiology 49 %).Three cases of superficial gastric car-cinoma were detected and are the majorreason for advocating repeated endoscopyand biopsy of all ulcers until completehealing is achieved.

Effect of short- and long-term cimetidine onhistological duodenitis and gastritis

H. M. GILMOUR, J. A. H. FORREST, MARY R.FETTES, R. F. A. LOGAN, AND R. C. HEADING(Department of Pathology and UniversityDepartment of Therapeutics and ClinicalPharmacology, The Royal Infirmary,Edinburgh) Patients with endoscopicduodenal ulceration were treated withcimetidine (1 or 2 g/day) for four or eightweeks until there was endoscopic evidenceof ulcer healing and then with 600 mg bdfor six months. Before and after eachperiod of treatment, multiple endoscopicbiopsies were taken from the duodenum,gastric body, and antrum. Histologicalgrading of duodenitis and gastritis wasbased on the criteria of Whitehead'.Over the initial four- or eight-week

period, the most severe histological duo-denitis in each patient was unchanged in16, improved in seven and increased inthree. Antral gastritis increased in allpatients (n = 6), no change occurring inbody mucosa. Over the six-month main-tenance period, six patients showed nochange in duodenitis, 15 improved, and 13deteriorated. During this period, nostatistically significant changes occurredin body or antral gastritis (n = 15). Thedegree of histological duodenitis beforefull-dose treatment or at entry to the 600mg bd period was of no value in predictingthose patients who would take more thanfour weeks of full-dose treatment to healtheir ulcers or who would relapse duringmaintenance therapy.The results demonstrate that gastritis

and duodenitis associated with chronicduodenal ulceration are unaffected bycimetidine therapy.

Reference"Whitehead, R. (1973). Mucosal Biopsy of the

Gastrointestinal Tract. W. B. Saunders.

Cimetidine treatment in chronic and acuteforms of Zollinger-Ellison syndrome (ZES)

S. BONFILS, M. MIGNON, AND J. GRATTON(Unite de Recherches de Gastroente'rologieHopital Bichat, Paris, France) Thirteencases of ZES were treated with cimetidine.This population could be divided into: (1)chronic forms, mostly presenting as acommon duodenal ulcer, (2) acute formsresulting in critical problems requiringintensive medical care.Among the seven chronic ZES cases,

cimetidine treatment was unsuccessful intwo; satisfactory clinical control was



j.com/



ownloaded from

http://gut.bmj.com/


obtained in three others, but gastrinomaexcision was the final treatment; cime-tidine treatment has been prolonged formore than 15 months in the last two cases.If in this condition acute pharmacologicalsecretory inhibition was constantlyobtained, therapeutic efficiency criteriaare not sensitive enough to feel secure inthe patient's long-term follow-up. Totalgastroectomy is still a valuable altemativeif gastrinoma excision is not possible.Out of the six acute ZES cases, four

were treated by addition of pirenzepin(0 5 mg kg-' tid, im) to cimetidine infusion(2.4 mg/day) resulting in an increased anti-secretory activity. However, total gastrec-tomy was the final outcome of every case,with one immediate postoperative death.

In conclusion, cimetidine in ZES treat-ment, although capable of inducing ulcerhealing, disappearance of diarrhoea anddramatic secretory inhibition, is stillchallenged by surgery which allows eithercomplete cure of the gastrinoma, ordefinitive suppression of the secretoryvirulence.

Reduction by selective gastric vagotomyand by proximal gastric vagotomy of thepancreatic-polypeptide response to foodand insulin hypoglycaemia

T. W. SCHWARTZ, N. A. L0VGREN, J.POULSON, AND E. AMDRUP (introduced byJ. F. REHFELD) (Institute of Medical Bio-chemistry and Department of SurgicalGastroenterology L, University of Aarhus,Denmark) The initial pancreatic-poly-peptide (PP) response to food is abolishedby truncal vagotomy and the secondaryprolonged response is significantlyreduced'. We have studied the PP secre-tion before and three months after twotypes of selective gastric vagotomy.The PP response to food was reduced

by proximal gastric vagotomy PGV;integrated initial response 0-30 min 77(47-103) % of preoperative value, secon-dary response 30-120 min: 66 (47-98) %,median and interquartile range, N = 15.Both the initial and the secondary PP

response to food were greatly reduced byselective gastric vagotomy (SGV); in-tegratedinitialresponse: 16 (8-46) % of pre-operative value, secondary response: 24(8-42) %, N = 23. No rapid increase (5-10 min) in PP concentrations were foundin 10 out of 23 patients after SGV, but inonly one out of 15 patients after PGV.The PP response to insulin hypogly-

caemia is abolished by truncal vago-

tomyt'3. After SGV or PGV a positivecorrelation was found between thereduced, initial PP response to food andthe integrated PP response to insulinhypoglycaemia, r, = 0 93, P 0 01, N = 9.A weaker correlation was found betweenthe secondary PP response to food and tohypoglycaemia, r. = 0-63, P 0 05.Assuming that insulin hypoglycaemia

is a central stimulus, this study indicatesthat the vagal innervation to the pancreasis severely injured during SGV butgenerally spared during PGV.

References"Schwartz, T. W., Rehfeld, J. F., Stadil, F., Larsson,

L.I., Chance, R. E., and Moon, N. (1976).Pancreatic polypeptide response to food induodenal ulcer patients before and after vagot-omy. Lancet, 1102-1105.

2Adrian, T. E., Bloom, S. R., Besterman, H. S.,Barnes, A. J., Cooke, T. J. C., Russell, R. C. G.,and Faber, R. G. (1977). Mechanism of pan-creatic polypeptide secretion. Lancet, 1, 161-163.

'Schwartz, T. W., Holst, J. J., Fahrenkrug, J.,Lindkaer Jensen, S., Nielsen, 0. V., Rehfeld,J. F., Schaffalitzky de Muckadell, 0. B., andStadil, F. (1978). Vagal cholinergic regulationof pancreatic polypeptide secretion. Journal ofClinical Investigation, 61, 781-789.

Does abnormal gastric mucosal permea-bility contribute to poor results afterpeptic ulcer surgery?

M. J. GOUGH, LINDA WOODHOUSE, C. S.HUMPHREY, AND G. R. GILES (UniversityDepartment of Surgery, St James'sUniversity Hospital, Leeds) Somepatients do badly after peptic ulcersurgery, possibly due to bile reflux,abnormal gastric emptying, or gastricmucosal damage. Gastric mucosal per-meability assessed by back diffusion oflithium chloride' may reflect mucosaldamage. This study measured lithiumfluxes in 18 patients with peptic ulcer, 12patients with recurrent ulcer post-vagotomy, and 12 patients with dumping,bile vomiting, or non-specific pain afterpeptic ulcer surgery. Simultaneously bilereflux under basal conditions, peak acidoutput to pentagastrin, and liquid testmeal emptying times were measured.

Pre-operative lithium fluxes were signi-ficantly lower (0.121 mmol Li ± 0-017/15min) than in patients with recurrent ulcer(0177 mmol Li ± 0 022) and patients witha poor surgical result (0 194 mmol Li +0-023) (p < 0-05). Bile reflux was absentin preoperative patients but present inpatients with recurrent ulcer (0.22 mmol/h ± 0.16) and symptomatic patients(0-36 mmol/h ± 0.13). However, therewas no significant correlation between the

degree of bile reflux and back diffusion oflithium in any group.

Test meal emptying times in recurrentulcer patients and symptomatic patientswas significantly faster than in pre-operative patients (p < 0-025). Theseresults did not correlate with lithiumfluxes, though postoperative bile refluxwas associated with more rapid gastricemptying.

Gastric mucosal permeability measuredby lithium back diffusion is increased inpatients with poor surgical results. Thesedifferences are not directly related to thedegree of bile reflux, nor changes ingastric emptying or acid secretion.

Reference'Chung, R. S. K., Field, M., and Silen, W. (1971).

Gastric permeability to hydrogen and lithium:A new method for quantitation of the gastricmucosal barrier. Surgical Forum, 21, 297-299.

SMALL AND LARGE BOWEL/LIVER ANDBILIARY

Functional classification of intestinal endo-crine cells

A. M. J. BUCHAN AND J. M. POLAK (Depart-ment of Histochemistry, Royal Post-graduate Medical School, HammersmithHospital, London) Many biologicallyactive gut peptides are produced bymucosal endocrine cells. Accurate identifi-cation of the cell types is essential asrecent results indicate.We have two groups of motilin anti-

sera, one directed to the C-terminal, theother to the whole molecule. The anti-sera to the whole molecule stain two celltypes: enterochromaffin with large (350nm), polymorphic, argentaffin granulesand cells with small (180 nm), round,electrondense, non-argentaffin granules.This indicates either a cycle of cellularmotilin production or that there are atleast two motilin-like peptides.The availability of neurotensin anti-

bodies allowed the identification of a pre-viously unknown ileal cell type with large(300 nm), round, electrondense granules.Using antibodies to glicentin (specific forgut glucagon-like immunoreactivity), wehave shown that (in man) glicentin isproduced by cells with small (190 nm)round, electrondense granules.Use of highly specific antibodies com-

bined with the semithin/thin technique hasshown the inaccuracy of previous cell



j.com/



ownloaded from

http://gut.bmj.com/


classifications by morphological criteria.For example, the D1 cell type has provedto be a heterogeneous group of cells pro-ducing intestinal gastrin, motilin-likeimmunoreactivity, and glicentin.A functional classification of peptide-

producing cells, stressing the productrather than morphological appearance, istherefore advisable in view of the in-creasing complexity of gut endocrinology.

Comparative study of long-term systemicand topical glucocorticoids on the rat smallintestine

J. SCOTT, R. M. BAIT, AND T. J. PETERS (intro-duced by V. S. Chadwick) (RoyalPostgraduate Medical School, London)Short-term prednisolone selectively en-hances the digestive-absorptive capacity ofthe rat jejunum, without altering mor-phology or enterocyte kinetics.1 In thepresent study the effect of long-termsoluble prednisolone-21-phosphate andtopical betamethasone-17-valerate hasbeen examined. Adult male Wistar ratswere fed either prednisolone (0 75 mg kg-1day-' orbetamethasone(O06mgkg-lday-1for 28 days or were pair-fed as controls.Jejunal enterocytes were isolated, brushborder marker enzymes assayed, andactivity expressed per enterocyte and percm of intestine. D-galactose absorptionwas measured using a recirculation per-fusion technique and a non-absorbable(14C)-polyethylene glycol marker)2. Enter-ocyte kinetics were studied using (3H)-methylthymidine 1.Both prednisolone and betamethasone

increased brush border enzyme activitiesand D-galactose absorption per entero-cyte. With prednisolone this resulted insignificantly enhanced brush borderenzyme activities and D-galactose absorp-tion per cm jejunum. In contrast beta-methasone produced marked mucosalatrophy and consequently no net altera-tion in either brush border enzymeactivity or D-galactose absorption perlength of intestine.Thus prednisolone produces a sustained

increase in the digestive-absorptivecapacity of the jejunum, which may be ofclinical value. Betamethasone-17-valerate,however, causes marked atrophy and-thus is unlikely to be useful in intestinaldisease; it may be contraindicated wherethere is pre-existing mucosal injury.

References"Batt, R. M., and Peters, T. J. (1976). Effects ofprednisolone on the small intestinal mucosa of

the rat. Clinical Science and Molecular Medicine,50, 511-523.

2Batt, R. M., and Peters, T. J. (1976). Absorptionof galactose by the rat intestine in vivo: proximal-distal kinetic gradients and a new method toexpress absorption per enterocyte. ClinicalScience and Molecular Medicine, 50, 499-509.

Adaptation of the shortened gut: differen-tial response in functioning and isolatedbowel

R. C. N. WILLIAMSON, F. L. R. BAUER, ANDR. A. MALT (Department of Surgery,University ofBristol, United Kingdom, andSurgical Services, Massachusetts GeneralHospital, Boston, USA) Evidence support-inghormonal control of intestinal adaptat-ion includes the transmission ofan entero-tropic factor between parabiotic rats' andthe greater intensity of distal hyperplasiafollowing jejunal resection as opposed tojejunal bypass.2 Humoral regulation ofintestinal growth was further tested inrats (N = 80) submitted to 50% proximalsmall bowel resection, with exclusion ofthe upper half of the remaining ileum as aThiry-Vella fistula (TVF). Other groupshad creation of an ileal TVF alone (shamresection) or simple laparotomy (control).Mucosal scrapings from 5-cm intestinalsegments were analysed for nucleic acidcontent and 3HTdR-labelled radioactivity.Compared with controls at 48 hours,

jejunal resection increased RNA and DNAby 15% (p < 0-05) in duodenum, by 18-21 % (p < 0-01) in defunctioned (upper)ileum and by 27-35% (p < 0-001) infunctioning (lower) ileum. Compared withsham resection, radioactivity in the TVFwas 36% higher (p < 0-02). One weekafter resection, nucleic acids remained 20-38% higher than either controls or shamsin the duodenum and 16-33 % higher in theileal TVF (p < 0-05 to 0'001); but thegreatest increases (27-86 %) again occurredin the functioning lower ileal segment.Although luminal factors contribute to

mucosal adaptation in bowel exposed tothe nutrient stream, they cannot explaincompensatory hyperplasia found either inisolated bowel or in bowel proximal to thesite of resection.

References'Williamson, R. C. N., Buchholtz, T. W., and

Malt, R. A. (1977). Adaptation of the shortenedgut: transfer of a humoral agent by cross-circulation. Gut, 18, 431.

'Williamson, R. C. N., Bauer, F. L. R., Ross,J. S., and Malt, R. A. (1978). Proximal enterect-omy stimulates distal hyperplasia more thanbypass or pancreaticobiliary diversion. Gastro-enterology, 74, 16-23.

Colon adenomas-a colonoscopic appraisal

C. B. WILLIAMS, P. E. GILLESPIE, B. C.MORSON, T. CHAMBERS, AND J. C. CHUAN(St Mark's Hospital, London) A colono-scopy survey of 620 patients with 1049colon adenomas shows a predominantlyleft-sided distribution (77 %). Of theselesions 97% were amenable to endoscopicremoval or ablation. Sixty per cent ofpatients presented with rectal bleeding astheir major symptom. Forty-eight per centof adenomas in our series were less than1-0 cm in diameter. Of the larger adeno-mas (>2-0 cm in diameter), 66% weresituated in the sigmoid colon, and of thosewith invasive malignancy (4-8% of thetotal) an even higher percentage (94%)were in the sigmoid and low descendingcolon. With increasing polyp size, therewas a greater predominance of villouselements and this was associated with ahigher risk of malignant change than themore frequent and generally smallertubular adenoma. Local colonoscopicexcision alone is sufficient treatment foradenomas with malignant change, unlessthey are poorly differentiated histologically,providing adequate resection is demon-strated. Twenty-eight patients treated inthis way are alive without recurrence atperiods from six to 62 months. Although65% of patients had only one adenoma,and 90% three or less, there is a risk ofdeveloping further benign and malignantcolon neoplasms and careful follow-up isrequired.

Dietary fibre, Vivonex, cholesterol and ex-perimental colon cancer

J. P. CRUSE, M. R. LEWIN, G. P. FERULANO,AND C. G. CLARK (Surgical Unit, Univer-sity College Hospital Medical School,London) Epidemiological studies im-plicate low-fibre, high-fat diets in theaetiology of colon cancer. We have testedaspects of this hypothesis in the dimethyl-hydrazine (DMH)-induced ratcolon cancermodel.

Five groups of 20 rats were allocateddietary regimens as follows: high fibre(20% w/w); standard fibre (3 % w/w); lowfibre (< 0 1 w/w); Vivonex (elemental, no-residue, liquid diet), and Vivonex pluscholesterol (100 mg/l). Within each group,half the animals received carcinogen(DMH, 40 mg kg b.wtl- wk-1, subcu-taneously, for 13 weeks) and half receivedsaline subcutaneously as controls. Animalswere weighed weekly, inspected daily for



j.com/



ownloaded from

http://gut.bmj.com/

A984 The British Society of Gastroenterologytumour presentation, isolated when mori-bund, killed when deemed terminal, andsubjected to necropsy.At 44 weeks, all controls were alive,

while mortality in the fibre groups was60 %, 80 %, and 60% respectively with nodifference in latent period, incidence oftumours or metastases. However, in theVivonex group, mortality was significantlyreduced to 10% (p < 0 01) with greatlyincreased latent period. The 'protective'effect of Vivonex was abolished by theaddition of cholesterol, this grouphaving a 60% mortality and similartumour parameters to the fibre groups.

Manipulation of dietary fibre in thismodel did not influence carcinogenesis.Vivonex however exerted a protectiveeffect which was nullified by cholesterol,implicating dietary fat as a potent co-carcinogen.

Vitamin D prophylaxis and osteomalacia inchronic cholestasis

JULIET E. COMPSTON, J. CROWE, I. WELLS,L. W. L. HORTON, D. HIRST, A. L. MERRETT,J. S. WOODHEAD, AND R. WILLIAMS(Gastrointestinal Research Unit, RayneInstitute, and Department of SurgicalPathology, St Thomas' Hospital, London;the Liver Unit and Departments ofChemical Pathology and Radiology, King'sCollege Hospital, London; and the Depart-ment of Medical Biochemistry, The WelshNational School of Medicine, Cardiff,Wales) The prevalence of histologicalosteomalacia was investigated in 32patients with chronic cholestatic liverdisease, 17 of whom were receiving highdoses of parenteral vitamin D. Undecalci-fied sections of transiliac biopsy speci-mens were quantitatively assessed. Plasma25-hydroxyvitamin D was measured by acompetitive protein-binding assay, andplasma parathyroid hormone by radio-immunometric assay.Four patients had histological osteo-

malacia. In two, who were receivingprophylactic vitamin D, the bone diseasewas mild and may have been healing,while in the remaining two, who were notreceiving vitamin D, more severe osteo-malacia was present. Clinical symptoms,radiology, and serum calcium and phos-phate concentrations were not reliableindicators of bone disease, and thedevelopment of osteomalacia wasunrelated to the duration or severity ofliver disease.Plasma 25-hydroxyvitamin D levels

were normal in all patients receivingvitamin D (mean ± SD, 45 9 ± 16-8nmol/l) but were low in many untreatedpatients (22-9 ± 15-2 nmol/l) beinglowest in the two with severe bonedisease. Plasma parathyroid hormoneconcentration was increased in only onepatient.

It is concluded that vitamin D defi-ciency is common in untreated patientswith chronic cholestasis, and some maydevelop severe osteomalacia. High-doseparenteral vitamin D therapy is usuallyeffective in preventing osteomalacia.

Reversible focal seizures as a manifestationof chronic portosystemic encephalopathy

N. E. F. CARTLIDGE, D. BATES, A. ANDERSON,AND 0. JAMES (Department of Neuro-logy, Victoria Infirmary, and DepartmentofMedicine (Geriatrics), Freeman Hospital,Newcastle upon Tyne) The classicalclinical features of chronic hepatic porto-systemic encephalopathy (PSE), knownfor 2300 years, were definitively describedby Adams and Foley'. In current texts onliver disease and neurology, little or nomention is made of focal (Jacksonian)seizures, myoclonus, or even generalisedepilepsy as features of chronic PSE.Epilepsy is well recognised in acutePSE .

In a current neurological study ofhepatic coma, 50 patients with coma orsevere pre-coma associated with chronicliver disease have been studied. In four,generalised epilepsy was a prominentfeature of deteriorating encephalopathy.All four, together with a fifth patient, alsohad well defined, frequently repeated,Jacksonian fits; in two, myoclonus wasalso a mark of worsening encephalo-pathy. All five subjects had biopsy-provenalcoholic cirrhosis (one with portocavalshunt) but no recent alcohol ingestion.Subsequent post mortem brain histologyrevealed no focal abnormality in thecerebral cortex of three patients, a fourthdied of a stroke, the fifth patient is stillalive.We suggest that (1) chronic PSE may

produce focal neurological signs with nounderlying anatomical lesion, and (2)careful neurological assessment andobservation of patients with relapsingPSE may reveal these features morefrequently than has hitherto been realised.

References"Adams, R. D., and Foley, J. M. (1953). The neuro-

logical disorders associated with liver disease,

in metabolic and toxic disease of the nervoussystem. Vol 32, Proceedings of the Associationfor Research-Nervous and Mental Disorders.Williams and Wilkins: Baltimore.

"Schiff, L. (1975). Disease of the Liver. 4th Edition.J. B. Lippincott.

3Steigman, F., and Clowdus, B. F. (1971). HepaticEncephalopathy. C. C. Thomas: Springfield,Illinois.

Impaired bacterial opsonisation due tocomplement deficiency in patients withfulminant hepatic failure

R. J. WYKE, I. A. RAJKOVIC, D. B. A. SILK,AND R. WILLIAMS (Liver Unit, King'sCollege Hospital Medical School, London)Patients in fulminant hepatic failure(FHF) are known to show an increasedsusceptibility to bacterial infection1'2. Wehave studied opsonisation, an importanthumoral factor in most defence to in-fection, by incubation of E. coli (1 x 106/ml) and normal polymorphonuclear leuco-cytes (2.5 x 106/ml) with sera diluted 1:25from 12 patients with FHF in grade IVcoma and matched controls3. Patient serakilled only 10-5 + SD 17-9% of theE. coli compared with 93-3 ± 7-6% forcontrol sera (p < 0 001). Cross-overdilution studies showed that defectiveopsonisation was not related to the pre-sence of toxins but to a deficiency inpatients' sera that was corrected byadding 1:160 dilutions of the heat labilecomponent of normal serum. This findingled us to measure the serum levels of com-plement factors in these patients. Func-tionally active total haemolytic comple-ment, C3, alternative pathway factors Band D were all highly significantly (p <0'001) reduced (irrespective of aetiologyor outcome).

Serial studies in all four survivorsshowed that E. coli opsonisation returnedto normal within 16-6 + SE 6-7 days inparallel with a return to normal ofcomplement factors.We conclude therefore that patients

with FHF have impaired bacterial op-sonisation due to complement deficiency,which may explain their increased suscepti-bility to infection.

References'Mummery, R. V., Bradley, J. M., and Jefferies,D. J., (1971). Microbiological monitoring ofpatients in hepatic failure with particularreference to extra-corporeal porcine liver per-fusion. Lancet, 2, 60-4.

'Gazzard, B. G., Portmann, B., Murray-Lyon,I. M., and Williams, R. (1975). Causes of deathin fulminant hepatic failure and relationshipto quantitative histological assessment ofparenchymal damage. Quarterly Journal ofMedicine, 176, 615-26.



j.com/



ownloaded from

http://gut.bmj.com/


'Quie, P., White, J., Holmes, B., and Good, R.(1967). In vitro bactericidal capacity of humanpolymorphonuclear leucocytes: diminished activ-ity in chronic granulomatous disease of child-hood. Journal of Clinical Investigation. 46, 668.

Intracranial pressure monitoring in fulmi-nant hepatic failureM. A. HANID, M. H. DAVIES, P. J. MELLON,J. J. MACCABE, L. STRUNIN, D. B. A. SILK,AND R. WILLIAMS (Liver Unit, King'sCollege Hospital Medical School. London)Cerebral oedema remains one of the com-monest causes of death in fulminanthepatic failure (FHF)' 2. The aims of thepresent study were to monitor intra-cranial pressure (ICP) in patients withFHF and to investigate the effects on ICPof prophylactic dexamethasone therapy(32 mg intravenously followed by 8 mgfour hourly), mannitol therapy (40-100mg intravenously following a sustainedpressure rise > 20 mmHg/h), and poly-acrylonitrile-membrane haemodialysis.

Subdural pressure transducers wereinserted in eight patients with FHFshortly after admission in grade IV coma.Pressure on insertion was 13-0 ± SD 11 4mmHg. Two patients survived in whomthe highest ICP recordings were 47 and 35mmHg. In these mannitol reversed orarrested ICP rises.

Precipitous rises in ICP (>110 mmHgover six to nine hours) were observed infive of the six patients who died. How-ever, mannitol was effective only if givenwhen ICP < 50 mmHg.

Haemodialysis (six periods in eightpatients) was associated with a significantincrease in ICP (29-3 ± SD 29-1 mmHgover four hours) compared with thatobserved in the preceding four hours (7 0+ 6-8 P < 0-05).We conclude, therefore, that, despite

high dose dexamethasone prophylaxis,precipitous rises in ICP still occur in FHFand that mannitol should be given beforethe pressure exceeds 50 mmHg.

References"Ware, A. J., D'Agnostino, A. N., and Coombes, B.

(1971). Cerebral oedema: a major complicationof massive hepatic necrosis. Gastroenterology,61, 877-884.

'Silk, D. B. A., Hanid, M. A., Trewby, P. N.,et al. (1977). Treatment of fulminant hepaticfailure by haemodialysis using a polyacrylo-nitrile membrane. Lancet, 2, 1-3.

Liver disease in the male homosexual

P. W. N. KEELING, J. BULL, J. E. BANATVALA,J. C. COLEMAN, I. M. MURRAY-LYON, AND

R. P. H. THOMPSON (Departments ofGastroenterology, Microbiology, andVirology, St Thomas' and Charing CrossHospitals, London) We have ascertainedthe prevalence of HBsAg and liver diseaseamong male homosexuals attending twoLondon venereal disease clinics and cor-related the histological changes with viralmarkers.Routine liver function tests were

assessed in HBsAg + ve patients and ifpersistently abnormal, liver biopsy wasrecommended. Over three years 1221homosexuals were surveyed: 91 wereHBsAg + ve. Of these 91, 15 had normaland 39 abnormal liver function tests, and37 declined investigation. Of the 39patients with abnormal tests, 10 had acutetype B hepatitis, nine refused further in-vestigation, and 20, all of whom wereasymptomatic, underwent liver biopsy.The histological appearances of the speci-mens showed five with minimal changes,four persistent hepatitis, two chroniclobular hepatitis, seven active chronichepatitis, and two with active cirrhosis.se' antigen was present in all patients withactive chronic hepatitis or cirrhosis, butin only half of those with milder liverdisease. To date, in two patients 'e'antigen has spontaneously disappearedfrom the blood.These results indicate a high prevalence

of liver disease among HBsAg + ve malehomosexuals, particularly those carryingthe 'e' antigen, and suggest that liverdisease is more frequent and severe thanamong HBsAg + ve blood donors.

CLINICAL MEDICAL AND CLINICAL SURGICAL

Controlled trial of a carbenoxolone/alginate antacid combination (Pyro-gastrone) in reflux oesophagitis

P. I. REED AND W. A. DAVIES (Departmentof Medicine, Wexham Park Hospital,Slough, Berkshire, and Royal PostgraduateMedical School, Hammersmith Hospital,London) Thirty-seven patients with en-doscopically proved reflux oesophagitiswere treated with Pyrogastrone (carben-oxolone/alginate antacid combination) orwith an active control drug (alginateantacid) under double blind conditions.Whereas after four weeks there was com-parable symptom relief with both com-pounds, after eight weeks the result with

Pyrogastrone was significantly better (p <0-025), complete remission or persistenceof only minimal symptoms being obtainedin 89% of patients. Oesophageal ulcerhealing was achieved in 100% and com-plete endoscopic healing or persistence ofonly minimal inflammation was noted in95% on Pyrogastrone, compared with33% (p = 0-02) and 67% (p < 0-05)respectively of controls. Side-effects inboth groups were mild and required notreatment. There was no demonstrablerelationship between healing, side-effects,and serum carbenoxolone levels. Thepossible mechanisms of action of carben-oxolone in oesophagitis are discussed.Compared with data from trials of otherdrugs, the results of this study wouldsuggest that Pyrogastrone is now the mosteffective agent available for the treatmentof reflux oesophagitis.

Double-blind trial of cimetidine in themanagement of resistant peptic oeso-phageal stricture

R. FERGUSON, M. W. DRONFIELD, AND M.ATKINSON (General Hospital, Notting-ham) Despite conventional medicalmeasures to control gastro-oesophagealreflux, regular dilatation at frequentintervals is required in about a third ofpatients with peptic oesophageal stricture'.In a randomised double blind trial we havetreated 15 such patients with cimetidine1-6 daily and placebo, each given for aperiod of six months. The diameter of thestricture and severity of oesophagitis wasassessed by endoscopy at 0, two, andsix months in each treatment period, anddilatation of the stricture was performedwhen necessary to relieve dysphagia.Symptoms of dysphagia and gastro-oesophageal reflux were recorded daily ondiary cards.There was significant improvement in

oesophagitis as graded endoscopicallyduring cimetidine treatment when com-pared with placebo. Restricturing, asassessed by the degree of dysphagia, thediameter of the stricture measured atendoscopy, and the need for furtherdilatations was, however, the same in thecimetidine and placebo treatment periods.We conclude that, when given for six

months, cimetidine does diminish oeso-phagitis but that this is not reflected in areduction in the need for dilatation.

Reference"Ferguson, R., and Atkinson, M. (1978). Outlook



j.com/



ownloaded from

http://gut.bmj.com/


withconservative treatment ofbenign oesophagealstricture. Gut, 19, 432.

Studies on the effects of cimetidine onanterior pituitary hormones; are theyclinically relevant?

S. LA BROOY, J. J. MISIEWICZ, G. DELITALA,A. JONES, C. R. W. EDWARDS, G. M. BESSER,W. A. STUBBS, K. G. M. M. ALBERTI (CentralMiddlesex Hospital, London; St Bartholo-mew's Hospital, London; University CollegeHospital, London; Southampton GeneralHospital, Southampton) Raised serumprolactins and gynaecomastia have beenreported on chronic treatment withcimetidine. We have previously shown thatcirculating prolactin levels remain un-changed in subjects given intravenouscimetidine 100 mg h-1 over four hours.Intravenous bolus doses of up to 400 mg,however, raised serum prolactin and thisresponse was dose dependent.' However,there are no data relating to prolongedmeasurements of serum prolactin levelsafter oral cimetidine.

Fourteen patients with peptic ulcer havebeen studied. In 10, serum prolactin wasmeasured at 30 minute intervals for 2-5hours at the end of a month's treatmentwith cimetidine Ig/day. Four other patientswere studied during two 12 hour periods,before and after four weeks of cimetidineIg/day. Two-hourly blood samples weretaken for anterior pituitary hormones,cortisol, insulin, and blood metabolites.Identical meals were eaten on both studydays. Analysis of serum prolactin levels inthe 10 patients studied after treatmentshowed no increase above the normalrange, nor was there any significant differ-ence between values in the four patientsstudied before and after cimetidinetherapy.These results suggest that the clinical

relevance of the reported hyperprolactin-aemia in patients taking cimetidine isdoubtful, particularly as prolactin is astress-related hormone and may beraised nonspecifically.

Reference1Delitala, G., Stubbs, W. A., Jones, A., and Besser,G. M. Clinical Endocrinology.

Increasing gallstone disease in Gibraltar-aclue to aetiology?

R. H. HUNT (Royal Naval Hospital, Haslar,Gosport, Hampshire) Gallstone diseaseis common in Europe and the incidence isincreasing.1'2

A marked increase has occurred inGibraltar where cholecystectomy is nowfive times more common than in Bristol.Between 1963 and 1976 the mean adultpopulation remained 14060 ± 190 (SE)and gall-bladder operations increasedfrom means of 27 for the first quadrenniumto 86 for the last quadrennium. This three-fold increase is linear and highly signifi-cant over the time period (x2 P < 0.001).These findings are supported by analysisof gall-bladder operations as a proportionof general surgical procedures.

Analysis by age and sex shows a shift toa younger age, although this is not signifi-cant and contrasts with other reportedstudies.Between 1969 and 1974, 1243 individual

patients (8-8% of the adult population)underwent 1812 radiological investiga-tions of the biliary tract. The films of 1024patients were rescrutinised and gallstonesseen in 308 (30%) or a non-functioninggall bladder in 326 (32 %). Yearly analysisshowed a highly significant linear increasein total and positive examinations (x2p < 0-001).

Analysis of food imports and consump-tion showed no significant changes and thecause of this increase is not clear.

ReferencesVan der Linden, W., and Rentzhog, U. (1960). Thechanging character of gallstone disease asobserved in a hospital population. Acta Chir-urgica Scandinavica, 119, 489-501.

'Holland, C., and Heaton, K. W. (1972). Increasingfrequency of gallbladder operations in theBristol clinical area. British Medical Journal,3, 672-675.

Gut hormone profile in morbid obesity andafter jejunoileal bypass

H. S. BESTERMAN, D. L. SARSON, A. M.BLACKBURN, J. CLEARY, T. R. E.PILKINGTON, J. C. GAZET, AND S. R.BLOOM (Royal Postgraduate MedicalSchool, London, St James' Hospital,London, and St George's Hospital,London) The role of gut hormones in themetabolic disorders of obesity and afterbypass is poorly understood. We havetherefore investigated the effect of astandard test breakfast (18 g protein, 22 gfat, 66 g carbohydrate, 530 k calories) ongut hormone release in 19 patients withmorbid obesity (225 ± 7% ideal weight)and compared it with 16 age and sexmatched normal controls (106 ± 3%ideal weight). In addition, 21 patients werestudied between two and 12 months afterjejunoileal bypass (181 ± 8% ideal

weight). All patients had lost weightpost-operatively (28 + 2 kg).

In obesity we find no gut hormoneabnormality corresponding to the aug-mented blood glucose and insulinresponses. After bypass, there is adramatic five-fold diminution in GIPrelease, an eight-fold increase in neuro-tensin release, and a massive 16-foldincrease in enteroglucagon release. Thestriking alteration in the pattern of guthormone release after jejunoileal bypassprovides important new insight into themetabolic changes following this opera-tion.

Team approach to long-term intravenousfeeding in gastroenterological patients

J. POWELL-TUCK, THALIA NIELSON, J.FARWELL, AND J. E. LENNARD-JONES (StMark's Hospital, London) A system hasbeen developed for administration ofintravenous feeding for prolonged periodsin general wards. The service is co-ordinated by a clinician, specialised nurse,and pharmacist. A silicone rubber catheteris introduced under local anaesthesia usingthe infraclavicular route and creating askin tunnel.' With this technique 22 of 27treatment periods (mean duration 44 days)have needed only one catheter. Ward careis simplified by the use of three-litre bagscontaining all nutritional requirementsexcept fat. Each bag is prepared underlaminar flow conditions by a pharmacistin a specially designed room. Studies of thesolutions we use have shown that theamino acids, glucose, electrolytes, andtrace elements are compatible. Vitaminsare added to the mixture and intralipid isinfused separately as needed.

In 38 treatment periods over a total of1551 days there was satisfactory weightgain which correlated with gain in fat freemass, assessed from skin-fold measure-ments (r = 086), and arm muscle cross-sectional area (r = 072). Some patientswho have learnt to spigot and care for thecatheter have left the hospital for periodsof up to three days, and two patients havegiven themselves supplemental infusionsregularly by night at home.

Reference"Powell-Tuck, J. (1978). British Medical Journal,

625.

Effect of a dietary fibre on gastric emptyingin dumpers

D. N. L. RALPHS, 0. LAWAETZ, N. J. G.



j.com/



ownloaded from

http://gut.bmj.com/


BROWN, AND A. R. LEEDS (introduced byM. Hobsley) (Departments of SurgicalStudies, The Middlesex Hospital andGastroenterology, The Central MiddlesexHospital, London) Pectin decreases post-prandial glycaemia in normal subjectswhen added to a carbohydrate meal' andalso improves symptoms after a glucosedrink in patients with the dumpingsyndrome.2

In this study six patients, who dumpedafter meals following gastric surgery forduodenal ulceration, were given 150 ml50% glucose on one occasion and the samemeal with an added 10-5 g pectin onanother. Both meals were labelled with1-5 mCi Indiumli3m. The symptoms ex-perienced were recorded and serial haema-tocrit measurements made on venousblood samples. Gastric emptying wasmonitored using a gamma camera system.The addition of pectin alleviated the

symptoms. The maximum falls in plasmavolume, derived from the haematocritreadings, were significantly less in thepectin studies (p = 0-05) and the Tiemptying time significantly longer (p =0O05). The distribution of the mealswithin the stomach and the rate of empty-ing showed different patterns in the twosets of studies.The results confirm the symptomatic

improvement with pectin in dumpingsubjects in these circumstances. Theysuggest that the modification in the wayin which the stomach handles the pectinmeal may be the prime cause of this effect.

References'Jenkins, D. J. A., et al. (1976). Annals of Internal

Medicine, 86, 22-23.'Abstracts of6th World Congress ofGastroenterology

(1978). Madrid.

Barium infusion examination in acute smallbowel obstruction

C. G. MARKS AND D. J. NOLAN (Depart-ments of Surgery and Radiology, RadcliffeInfirmary, Oxford) Twenty-three patientswho were admitted to hospital over a 21month period with suspected small bowelobstruction were examined with a bariuminfusion. The examination was performedby infusing barium through a tubedirectly into the small intestine and takingradiographs when indicated under screen-ing control.'

Mechanical small bowel obstructionwas confirmed in 22 patients and the smallintestine was shown to be normal in onepatient. The level of obstruction was

shown in all except one patient, who had acarcinoma of the caecum.

Seven patients were shown to haveCrohn's disease of the small intestine butonly two required emergency surgery. Sixpatients had metastatic disease involvingthe small intestine and they all came tooperation. Small bowel obstruction wascaused by adhesions in three patients andtwo were operated on. In the third patientthe barium appeared to release adhesionsadjacent to a drainage tube. The remain-ing five patients had miscellaneous lesions:a radiation stricture, femoral hernia,ileostomy twist, ischaemic stricture withperforation, and a malrotation. They allcame to operation.

There were no complications associatedwith the examination or with the sub-sequent operation.

Reference'Sellink, J. L. (1976). Radiological Atlas ofCommon

Diseases of the Small Bowel. Stenfert Kroese:Leiden.

Pathogenesis of sepsis after appendicec-tomy

I. A. DONOVAN, D. ELLIS, D. GATEHOUSE,G. LITTLE, R. GRIMLEY, S. ARMISTEAD,M. R. B. KEIGHLEY, AND C. J. L. STRACHAN(General Hospital, Birmingham; QueenElizabeth Hospital, Birmingham; DudleyRoad Hospital, Birmingham; and RoyalInfirmary, Huddersfield) The relative im-portance of aerobic and anaerobicorganisms in the genesis of infection afterappendicitis remains uncertain. We there-fore designed a prospective randomisedtrial of single dose prophylaxis to com-pare an agent effective against anaerobes(clindamycin phosphate) with one effectiveagainst aerobes (cefazolin sodium).

Two-hundred-and-fifty patients agedover 12 years entered the trial, there werel 2withdrawals. Seventy-two patients receivedplacebo, 81 received 600 mg clindamycin,85 received 1 g cefazolin given intra-muscularly immediately before operation.

Overall wound sepsis rates were placebo33 %, clindamycin 17 %, cefazolin 35%(p < 0 05 clindamycin versus others. Chi-squared Yates correction). In 52 patientswith gangrenous or perforated appendi-citis overall wound sepsis rates wereplacebo 78%, clinamycin 44%, cefazolin80% (p < 0-05 clindamycin versus others).

Microbiology was obtained in 50 woundinfections. Anaerobes were isolated in60% of placebo, 38% of clindamycin,and 60% of cefazolin infections; aerobes

from 90% placebo, 100% clindamycin,and 64% cefazolin infection. Cefazolinsignificantly reduced the number ofaerobicisolates (p < 0 05).We conclude that an agent active against

anaerobes (clindamycin) significantly re-duced wound infection, whereas oneactive against aerobes (cefazolin) did not,and that anaerobic organisms are thereforeof much greater importance than aerobicorganisms in the pathogenesis of sepsisafter appendicectomy.

Trial of high fibre diet or local treatment forpatients with haemorrhoids

Y. ARABI, T. MAKURIA, P. BUCHMANN, J.ALEXANDER-WILLLAMS, AND M. R. B.KEIGHLEY (The General Hospital Birming-ham) It has been suggested that onlypatients with piles who have high analpressures should be treated by procedureswhich reduce anal pressure because of therisk of prolapse in the remaining patients1.We have therefore undertaken two

randomised clinical trials of treatment forpatients with piles. Patients with high analpressure (n = 110) were allocated totreatment by sphincterotomy (LSS), analdilatation (MDA), or high fibre diet(diet). Patients with low pressure (n =111) were allocated to rubber bandligation (RBL), cryosurgery (C), or diet. Asymptomatic review and measurement ofanal pressure2 were repeated one and fourmonths after treatment.

In the high pressure group: LSS, MDA,and diet were equally effective and relievedsymptoms in 85-92% of patients at fourmonths. There was a 9% reduction in analpressure by all forms of therapy. In thelow pressure group: diet improved only50% at four months compared with 83%after C and 87% by RBL. There were morecomplications and an increase in time offwork after C than RBL.These data indicate that diet is effective

and has economic advantages over LSSand MDA for patients with high analpressure, but that RBL is preferable forpatients with low anal pressure.

References"Arabi, Y., Gatehouse, D., Alexander-Williams, J.,and Keighley, M. R. B. (1977). Rubber bandligation or lateral subcutaneous sphincterotomyfor treatment of haemorrhoids. British JournalofSurgery, 64, 737-740.

'Hancock, B. D., and Smith, K. (1975). The internalsphincter and Lord's procedure for haemor-rhoids. British Journal of Surgery, 62, 833-836.



j.com/



ownloaded from

http://gut.bmj.com/


Excision or resection for carcinoma of themiddle third of the rectum?

R. J. NICHOLLS, JEAN K. RITCHIE, JANEWADSWORTH, AND A. G. PARKS (StMark's Hospital, London) Resection withanastomosis is being used increasingly forcarcinoma of the middle third of therectum, although it is not known if thisis as curative as total rectal excision fortumours in this site. More advanced andanaplastic tumours tend to be treated byexcision: overall survival rates are notcomparable therefore.Between 1963 and 1972 at St Mark's

Hospital, 199 patients underwent poten-tially curative operations (excision 112,resection 87) for a single adenocarcinomaat > 8 cm < 12 cm from the anal verge.

Using single pathological variables, thenumber of deaths due to carcinoma orunknown cause within five years was notsignificantly different for patients withDukes' B or average grade tumours aftereither type of operation. However,patients with Dukes' C tumours faredsignificantly better after resection.With two variables, there was no signifi-

cant difference in five year survival forDukes' B or C tumours of average grademalignancy after excision or resection.When the extent of local spread was

added as a third variable, a significantlyimproved survival after resection wasfound in two subgroups.

These results suggest that resection isas curative as excision for middle-thirdrectal carcinoma.

Patterns of change in carcinoembryonicantigen (CEA) levels in patients developingrecurrent colorectal cancer

C. B. WOOD, R. W. BURT, J. G. RATCLIFFE,A. J. H. MALCOLM, AND L. H. BLUMGART(University Department of Surgery, RoyalInfirmary, Glasgow) Serial CEA estima-tions were performed in 199 patients afterapparently curative surgery for colorectalcancer. In a follow-up period of 15-51months, 41 patients had two consecutiveraised CEA levels (> 25 jAg/l). No tumourrecurrence was detected in five (12 %), andclinical detection preceded raised CEAlevels in one (2-4 %). CEA levels rose con-currently with, or before, detection ofrecurrence in the remaining 35 patients(85 %). Raised CEA levels preceded de-tection of recurrence in 32 patients (78 Y.)by two to 28 months (median = fourmonths).

Two patterns of rises in CEA wereobserved. A slow rise, with levels notexceeding 75 itg/l within 12 months of thefirst CEA rise, were seen in 15 patients. Ofthese five have died (33 %). In 21 patientsCEA levels rose to > 100 ,ug/l within sixmonths of the first raised level beingrecorded. Of these 17 (81 %) have died.Thus serial CEA assays will allow earlydetection of recurrent colorectal cancer.The rate of CEA rise may have prog-nostic significance.

Clinical trial to compare manual dilatationof the anus and lateral subcutaneoussphincterotomy for anal fissure

Y. ARABI, T. MAKURIA, P. BUCHMANN,J. ALEXANDER-WILLIAMS, AND M. R. B.KEIGHLEY (The General Hospital, Bir-mingham) Anal fissures are associatedwith increased activity of the internalanal sphincter'. Manual dilatation (MDA)gives good symptomatic relief but re-quires a general anaesthetic. Sphinc-terotomy (LSS) is an alternative techniquehaving the advantage that it may be per-formed under local anaesthesia2.A prospective randomised trial has

compared MDA with LSS in 152 con-secutive patients with anal fissures (LSS:n = 75, MDA: n = 77). Anal pressureshave been measured3 before, at four and12 months after treatment. The incidenceof complications, time off work, and asymptomatic assessment were alsorecorded.

Complications of pain and bleedingoccurred in two patients after LSS com-pared with nine after MDA. Within threedays of LSS 87% returned to work com-pared with only 56% after MDA (P <0.01). At four months there were threepatients after LSS and four patients afterMDA who developed recurrent fissures.At 12 months four more patientsdeveloped recurrent fissure after LSS.Compared with the preoperative pressuresthere was a sustained reduction at four and12 months but the percentage reductionwas greater after MDA (19%) than afterLSS (8%/).These results indicate that LSS has less

morbidity and potential economic advan-tages over MDA but the long-term resultsof MDA are slightly better than LSS.

References1Eisenhammer, S. (1953). The internal anal

sphincter. Its surgical importance. South AfricanMedical Journal, 27, 266-269.

'Notaras, M. J. (1971). The treatment of analfissure by lateral subcutaneous sphincterotomy.

A technique and results. British Journal ofSurgery, 58, 96-100.

3Hancock, B. D. (1976). Measurement of analpressure and motility. Gut, 17, 645-651.

INFLAMMATORY BOWEL/CLINICAL MEDICAL

Gut hormone profile in inflammatory boweldisease

H. S. BESTERMAN, S. R. BLOOM, N. D.CHRISTOFIDES, C. N. MALLISON, A. PERA,AND R. MODIGLIANI (Royal PostgraduateMedical School, London; LewishamHospital, London; Ospedale Mauriziano,Turin, Italy; and H6pital St Lazare, Paris,France) We have previously reportedcharacteristic abnormalities in gut hor-mone release in certain well-defined con-ditions affecting the alimentary tract-forexample, coeliac disease1 and acutetropical sprue2. No previous informationhas been reported on gut hormones ininflammatory bowel disease. We have,therefore, investigated the gut hormoneresponse to standard test breakfast (530 Kcal) in 14 patients with Crohn's disease, in24 patients with ulcerative colitis and com-pared this with 14 age and sex-matchednormal subjects. Patients with ulcerativecolitis had significantly raised basal andpeak gastrin levels (11-8 ± 1-7, 50 ± 8pmol/l) compared to normal subjects (5-4+ 0-6 25-0 + 7-1, p < 0-001, P < 0-02)and also to patients with Crohn's disease(7-3 ± 1, 21-4 + 5.9 ip < 0-05, p < 0-01).In contrast, basal levels of motilin wereraised in both the ulcerative colitis group(141 ± 23 pmol/l) and the Crohn'sdisease group (157 ± 35) compared tonormal subjects (74 ± 15, P < 0-05 forboth). The three hour incrementalresponse of enteroglucagon was greater inthe Crohn's disease patients (6.2 ± 1-3nmol/1) compared both to normal subjects(1-7 ± 0-5 nmol/l, P < 0-005) and topatients with ulcerative colitis (3-0 ± 0-5,p < 0-05). The release of pancreatic poly-peptide and insulin was, however, normalin both groups.We therefore find that there is a signifi-

cant difference in the patterns of guthormone release between Crohn's diseaseand ulcerative colitis and that these aremarkedly different from those previouslyreported for other gut diseases.

References'Besterman, H. S., Bloom, S. R., Sarson, D. L.,

Blackburn, A. M., Johnston, D. I., Patel,



j.com/



ownloaded from

http://gut.bmj.com/


H. R., Stewart, J. S., Modigliani, R., Guerin, S.,and Mallinson, C. N. (1978). Lancet, 1, 785.

'Besterman, H. S., Cook, G. C., Sarson, D. L.,Christofides, N. D., and Bloom, S. R. (1978)Gut, 19, A438.

Prostaglandin synthetase activity in acuteulcerative colitis: the effects of treatmentwith sulphasalazine, codeine phosphate andprednisolone

D. J. DAWSON, P. R. SMITH, AND C. H. J.SWAN (Department of Gastroenterology,North Staffordshire Hospital Centre,Stoke-on-Trent) Prostaglandins (PGs)have been implicated in the pathogenesisof diarrhoea in disease states. In activeulcerative colitis, faeces contain high levelsof PG-like material, and metabolites areexcreted in the urine. We have previouslyshown raised PG synthetase activities inrectal biopsies of ulcerative colitis.

Seven patients with active ulcerativecolitis were treated with conventionaltherapy and rectal biopsies taken at 0,two and four weeks treatment. Rectal PGsynthetase activities were assayed by aradiometric method and were shown to fallsignificantly with treatment (P < 0-001 fordifference at 0 and four weeks).

In vitro studies using affected colon froman acute colitic undergoing panprocto-colectomy and uninvolved colon from apatient with polyposis coli have shownthat sulphasalazine is a potent inhibitorof PG biosynthesis in mucosal homo-genates, whereas prednisolone and codeinephosphate are ineffective. Indomethacin,was even more potent than sulphasalazine.Percentage inhibition was similar fornormal and colitic mucosae, althoughabsolute values of PG biosynthesis wereincreased in colitics.

It is shown that prostaglandin syn-thetase activity is increased in the mucosaof ulcerative colitis during acute attacksand falls during conventional therapy.One of the effects of sulphasalazine maybe to inhibit PG biosynthesis, and atherapeutic role for other PG synthesisinhibitors is suggested.

Bacteria, prostaglandins, and ulcerativecolitis

M. J. HUDSON, M. J. HILL, S. R. GOULD, ANDJ. LENNARD-JONES (Bacterial MetabolismResearch Laboratory, Colindale, London,and St Mark's Hospital, London) Thefaeces of patients with acute colitis havebeen shown to contain high levels of

prostaglandins (PGs) and treatment withsulphasalazine (SASP) is paralleled by afall in PGs to normal low levels'. Anincreased in vitro synthesis of PGs bycolorectal mucosal biopsies from acutecolitics compared to controls and SASP-treated colitics has also been reported2.We have examined the ability of isolates

of the intestinal microflora to synthesisePGs from arachidonic acid (AA), a pre-cursor of PGs present in faeces. Mixedcultures of the predominant flora fromcolitics and controls were grown separatelyin a complex medium with AA (100ag/ml). Bacterial pellets, disrupted by

ultrasonication or EDTA-toluene treat-ment were found to contain between 25and 4500 ng/g PG when assayed by bothbioassay and radioimmunoassay. Theratio of PGE2 to PGF2a was approxi-mately 3 to 1, a similar ratio to that foundfor faecal PGs in colitics. Preliminaryqualitative studies using thin-layer chro-matography of bacterial PGs producedfrom 14C-AA suggest that Bacteroides spp.are particularly active in PGE2 production,and that SASP is able to inhibit synthesis.Studies are continuing to investigate thefactors controlling bacterial PG synthesisin vitro, the ecology of these bacteria andtheirroleinthe pathogenesis of the disease.

References'Gould, S. R., and Lennard-Jones, J. (1976).

Production of prostaglandins in ulcerativecolitis and their inhibition by sulphasalazine.Gut, 17, 828.

'Harris, D. W., and Swan, C. H. J. (1977). Increasedsynthesis of prostaglandins in ulcerative colitis.Lancet, 2, 196.

Immunohisto- and ultracytochemical obser-vations on the early lesion in ulcerativecolitis

JAN-OLAF GEBBERS AND HERWART F. OITO(introduced by D. P. Jewell) (The RoyalFree Hospital, London Institute of Path-ology, University of Hamburg, Hamburg,W. Germany) Biopsies from 21 youngpatients with ulcerative colitis in an activestate, who were under no treatment withcorticosteroids, have been studied by theindirect immunoperoxidase methoddemonstrating IgA, IgG, IgM and Clq,C3. Parallel electron microscopic observa-tions were carried out in 12 of these cases.A striking deviation from the normal

local Ig-cell class pattern was found,resulting in a 30-fold increase of IgG-cells.Abundant extracellular IgG occurred,which was released by a large number ofnecrobiotic IgG-cells. IgG aggregates

were bound to the basement membrane ofthe surface epithelium, where activatedcomplement (Clq, C3) could also bedemonstrated. Simultaneously granulo-cytes infiltrated the surface epitheliumforming microabscesses where IgG, Clq,and C3 are detectable as well.

Electronmicroscopically the infiltratinggranulocytes were found to be in a stateof degranulation and disintegration andultracytochemically a release of theirlysosomal peroxidase into the intersticeswas detectable near the basement mem-brane. In the vicinity of these granulocytesmicro-ulcerations were seen.The association of IgG and activated

complement in the active state suggests alocal antigen-antibody reaction. Theearly lesion is probably not only causedby activated complement but also granulo-cytic mediated as a consequence of a'frustrated phagocytosis' of membrane-associated immune complexes.

New mast cell stabilising compound in thetreatment of ulcerative colitis

P. S. DAVIES AND J. RHODES (Departmentof Gastroenterology, University Hospitalof Wales, Cardiff) Sulphasalazine is thestandard treatment for maintaining clini-cal remission in ulcerative colitis. Somepatients are unable to tolerate the com-pound and attempts have been made tofind altematives. A mast cell stabiliser,disodium cromoglycate, is of someclinical value in this situation. We haveexamined a new mast cell stabiliserPRD-92 (Boehringer), which is wellabsorbed, and have compared it withsulphasalazine in 44 patients duringremission with ulcerative colitis using adouble-blind study. Twenty-two patientstook active PRD and placebo sulphasa-lazine; two of these could not toleratethe new drug and were withdrawn afterless than a week's treatment. Twelveof the 20 patients continued in remissionfor the six-month period of the trial(relapse rate 40%/); assessments werebased on clinical symptoms, sigmoido-scopy, and rectal histology. Twenty-twopatients were given active sulphasalazineand placebo PRD; 20 continued inremission for six months (9% relapse). Weconclude that sulphasalazine maintainsclinical remission in ulcerative colitis,while PRD-like other mast cell stabi-lisers-has a relatively weak effect in thisclinical situation.



j.com/



ownloaded from

http://gut.bmj.com/


Sphincter on the ileum: mucosal proc-tectomy, colectomy, and ileo-anal ana-stomosis for ulcerative colitis

D. JOHNSTON AND N. S. WILLIAMS (Uni-versity Department of Surgery, TheGeneral Infirmary at Leeds) Removalof the anal canal and anal sphinctersin patients with ulcerative colitis (UC)is wasteful, because these structures areusually normal. Also, as the disease isprimarily mucosal, the rectal muscleis also worth preserving. Provided thatthe sphincteric apparatus (including pubo-rectalis and levators) is preserved, patientsremain continent even after completeexcision of the rectum1. Complete ano-proctocolectomy, in contrast, may pro-duce impotence2, bladder dysfunction,and psychological upset due to thepermanent ileostomy.An alternative approach is to remove

the disease by means of subtotal colectomyand mucosal proctectomy, ileum beingdrawn through the tube of rectal muscleto be anastomosed just above the anus3.In a pilot study, we did this in threeyoung men. Since dissection is endo-rectal, there is no danger to the pelvicparasympathetic nerves.

All three patients are continent duringthe day. One is incontinent at night. Thefrequency of bowel action ranges fromseven to 12 per day.

This procedure might be a usefulalternative to the reservoir ileostomy.

References'Lane, R. H. S., and Parks, A. G. (1977). Functionof the anal sphincters following colo-analanastomosis. British Journal of Surgery, 64,596-599.

'Burnham, W. R., Lennard-Jones, J. E., and Brooke,B. N. (1977). Sexual problems among marriedileostomists. Gut, 18, 673-677.

'Ravitch, M. M., and Sabiston, D. C. (1948).Anal ileostomy with preservation of the sphinc-ter. Surgery, Gynecology and Obstetrics, 84.1095-1105.

VIP nerves in Crohn's disease

A. E. BISHOP, J. M. POLAK, M. G. BRYANT, ANDS. R. BLOOM(DepartmentsoffHistochemistryand Medicine, Royal PostgraduateMedical School, London) Abnormalitiesof autonomic nerves are known to existin Crohn's disease. As VIP nerves forman important part of the autonomicnervous system a combined immunocyto-chemistry and radioimmunoassay studyof their involvement in Crohn's diseasewas undertaken.

Surgical specimens were removed from

nine patients with Crohn's disease.Samples were taken from regions ofmaximal pathology and from proximaland distal areas. Control tissue wasobtained from 16 patients with carcinomasand sampled from the following areas:(1) macroscopically normal; (2) distendedwith no hypertrophy; (3) grossly hyper-trophied.The controls had the normal VIP nerve

pattem of fine fibres running through themuscle layers and forming meshes in thesubmucosa. In groups 1 and2the myentericplexuses were normal in appearance andVIP innervation. The plexuses in group 3were distorted but contained normal VIPinnervation. The nerves in the diseasedsamples were increased in number anddegree of immunostaining with distendedfibres and disorganised, densely packedmeshes and plexuses. Radioimmunoassayof gut extracts showed increased VIPconcentration in the Crohn's tissues(243 ± 51 pmol/g) compared with controltissues (121 ± 16 pmol/g) (p < 005).

These results suggest that there is aremarkable involvement of VIP inner-vation in the pathology of Crohn'sdisease and that further investigation isimportant to ascertain whether thesechanges play any role in aetiology.

Crohn's disease activity index-is ituseful ?

A. S. MEE, D. J. BROWN, AND D. P. JEWELL(Academic Department ofMedicine, RoyalFree Hospital, London) Assessment of act-ivity in Crohn's disease is notoriouslydifficult. TheCrohn'sdisease activity index(CDAI)1 attempts to provide a measure ofseveritybasedon objectivecriteria, allowinga comparison of activity scores betweendifferent centres. We have found theCDAI to be cumbersome to use and havetherefore compared it with a simpleclinical assessment at the time of seeingthe patient, the ESR, and the serum levelof C-reactive protein.The CDAI was obtained in 42 con-

secutive patients. Simultaneously a clinicalassessment was made which includedgeneral well-being, presence or absenceof pain, diarrhoea, and complications.The disease was graded as inactive, mild,moderate, or severe. Blood was takenfor ESR and C-reactive protein (measuredby Laurell immunoelectrophoresis).Twenty-six per cent of patients had aCDAI of more than 150 and wouldtherefore be regarded as having active

disease.The four methods of assessment were

compared using Spearman's rank cor-relation coefficients (with mathematicaladjustment for tied ranks).There was a highly significant cor-

relation (p < 0 01) between each methodof assessment. No one method appearedto be a more sensitive indicator of diseaseactivity.For multicentre studies of Crohn's

disease, the CDAI provides an objectivemethod of disease assessment but doesnot appear to be superior to an ESR orC-reactive protein estimation.

Reference"Best, W. R., Becktel, J. M., Singleton, J. W., and

Kern, F. Jr. (1976). Development of a Crohn'sdisease activity index. National CooperativeCrohn's Disease Study. Gastroenterology, 70,439-444.

Effects of antacids on the absorption ofcimetidine

G. BODEMAR, JANE G. MILLS, B. NORLANDER,P. OSBORNE, W. L. BURLAND AND A.WALAN (University Hospital, Linkoping,Sweden, and The Research Institute,Smith Kline & French Laboratories Ltd,Welwyn Garden City) The absorptionof a single oral dose of cimetidine 400 mgtaken with and without antacids has beenstudied after a meal in 11 healthy subjects,and in 18 fasting patients with pepticulcer disease.

In healthy man 20 ml (or two tablets)antacid was given before and 30, 90, and150 minutes after cimetidine. Concurrentadministration of Aludrox, magnesiumtrisilicate BPC, Maalox, and Rennies hadno effect on the bioavailability of an oraldose of cimetidine. Neutralising capacitywas 10, 11, and 25 mmol HCl/10 mlantacid and 15 mmol HCI/tablet, re-spectively.

In fasting patients concomitant admini-stration of 30 ml of an antacid containingaluminium and magnesium hydroxide(Novalucol, neutralising capacity 35 mmolHCl/10 ml) caused a significant decreasein mean peak blood concentration(P < 0 001) and the mean area under theblood concentration curve was reducedby 22% (P < 0{001). However, no effectwas seen when 30 ml Aludrox was givenunder the same experimental conditions.

Concurrent administration of cimeti-dine and antacid suspension with a highbuffering capacity should be avoided whenthe patient is fasting, but this is unlikelyto occur in practice, as dosage instructions



j.com/



ownloaded from

http://gut.bmj.com/


for cimetidine recommend that it is takenwith or after meals and at bedtime.

Effect of maintenance cimetidine treatmentin duodenal ulcer and the eventual outcomeafter cessation of treatment

M. W. DRONFIELD, A. J. BATCHELOR, W.LARKWORTHY, AND M. J. S. LANGMAN(University Department of Therapeutics,City Hospital, Nottingham: and RAFHospital, Nocton Hall, Lincoln) Forty-two patients with endoscopically diag-nosed duodenal ulcer (DU) were studiedin a double-blind trial after their ulcershad been healed by cimetidine. Cimetidinewas effective in preventing relapse, onlyfive of the 20 patients allocated to cime-tidine 400 mg twice daily relapsingduring the six months' treatment (cumu-lative relapse rate, CRR, 25 %.), comparedwith 16 of the 22 on placebo treatment(CRR 76%, P < 0 01). The 21 patientswho were asymptomatic at the end ofthe trial had their treatment discontinuedand have been followed-up for at leasta further eight months in all but five cases.Relapse in these patients was common,particularly in those previously treatedwith maintenance cimetidine, in whomthe CRR rose from 25% to 68% inthe eight months after discontinuing thetrial, compared with a rise from 76%to 84% in those previously treated withplacebo. Thus, the eventual outcomein patients treated with maintenancecimetidine was only slightly better thanthat in those who had received placebo,and it is likely that, if cimetidine is to beused for DU, the majority of patients willrequire prolonged if not indefinite main-tenance treatment. Until such treatmentis shown to be safe and effective, surgeryremains the treatment of choice for mostpatients with severely symptomatic DU.

Adenine arabinoside therapy in hepatitis Bantigen positive chronic liver disease

M. F. BASSENDINE, R. G. CHADWICK, E. M.CRAWFORD, H. C. THOMAS, AND S. SHERLOCK(Department of Medicine, Royal FreeHospital, London) Adenine arabinoside(ARA-A) is a synthetic purine nucleosidewith activity against DNA viruses. Wereport the effect of ARA-A on thehepatitis B virus (HBV) in chronic liverdisease (CLD).Four male patients, aged 23-48 years,

with biopsy proven HBSAg + ve CLD

(duration > six months) were treated.These patients had Dane particles (HBV)in their serum by immune electronmicroscopy and raised HBV specificDNA polymerase activity (DNAp).ARA-A was initially given in a dose

of 10 mg/day intravenously for 10 days.DNAp fell immediately in all patients.In three patients activity returned onstopping treatment and HBsAg, HBeAg,and Dane particle positivity did notchange. In the fourth patient DNApremained absent after treatment andthis was associated with falling levelsof HBsAg, absent HBeAg and Daneparticles.

Repeat courses of ARA-A in twopatients showed that a higher dose of20 mg kg-' day-' and more prolongedtreatment for 21 days again produced arapid fall in DNAp but this reappearedon stopping treatment.ARA-A has activity against HBV

in HBsAg + ve CLD but short coursesmay be insufficient to eradicate the virus.Further studies are under way to deter-mine optimum treatment regimes.

Evidential value of the hospital recordin clinical decision-making

W. I. CARD, W. SIRCUS, AND A. N. SMITH(Gastro-intestinal Unit, Western GeneralHospital, Edinburgh; Diagnostic Metho-dology Research Unit, Southern GeneralHospital, Glasgow) In a formalisedstructure of medicine the activity of thedoctor could be analysed as a sequenceof decisions each based on the evidencehe elicits". Part of this evidence may bein the medical record of a past illness.The problem is: what facts need to berecorded about an illness which wouldhave evidential value should the patientbe readmitted?A retrospective study was undertaken of

53 re-admissions to a gastrointestinalunit. Evidence about the present state ofthe patient and, if indicated, from therecord of the past illness, was soughtfrom a 'referee', by a physician and asurgeon. In a simulated exercise they usedthis information to arrive at the diagnosisand management. A numerical estimatewas made of the evidential weight of eachitem requested, its degree of irrecovera-bility, and its benefit to the patient.Each piece ofevidence was given a score,

and the total score derived from eachsource of evidence for all the patientscompared. The sources of evidence in

order of importance were operation find-ings, pathology, laboratory, radiology,and endoscopy. Other sources contributedlittle or none. It is concluded that, giventhe diagnosis at discharge and the detailsof any operation performed, the evidenceworth recording could be specified; thisevidence would be brief and could benumerically coded.

ReferenceGood, I. J., and Card, W. I. (1971). The applicationof rationality to medical records. MathematicalBiosciences, 10, 157-176.

Faecal bile acid (BA) excretion in patientswith cholelithiasis before and duringchenodeoxycholic (CDCA) and ursode-oxycholic (UDCA) acid therapy

M. T. PODESTA, C. A. MCGUFFIE, G. M.MURPHY, AND R. H. DOWLING (Gastro-enterology Unit, Guy's Hospital andMedical School, London) During CDCAand UDCA therapy of gallstones, in-increased amounts of potentially hepa-totoxic lithocholate form, but do notaccumulate appreciably, allegedly becausehepatic sulphation prevents their intestinalreabsorption1 but resultant changes infaecal excretion have never been studied.Furthermore, why some patients treatedwith CDCA get diarrhoea, while others,and UDCA-treated patients, do not, isunknown.

Therefore in 14 controls, five untreatedgallstone patients, 12 taking CDCA (sixwith diarrhoea and six without) and seventaking UDCA, we measured total andindividual faecal BAs before and aftersolvolysis2 (which removes sulphates) andcorrelated findings with faecal wetweight.

Results showed that total faecal bileacid output (mg/day) was comparable incontrols (680 + SEM 106), untreatedpatients (845 ± 64), and in patients with-out diarrhoea taking CDCA (864 ± 186)or UDCA (650 ± 165), but was signifi-cantly increased in patients with diarrhoea(1333 ± 178; p < 0-005). Stool weightalso correlated with total faecal BAexcretion (r = 0-54; p < 0-001) and withexcretion of non-sulphated cathartic(CDCA ± DCA) BAs (r = 0-56; p <0-001). Contrary to expectation', litho-cholate excretion (mg/day) was notincreased during CDCA (86 ± 49) orUDCA (94 ± 34) when compared withuntreated patients (192 ± 82).We conclude that CDCA-induced

diarrhoea occurs mainly with high faecal



j.com/



ownloaded from

http://gut.bmj.com/


BA excretion; lithocholate sulphate ex-cretion is not increased during BA therapy.

References'Cowen, A. E., Korman, M. G., Hofmann, A. F.,

Cass, 0. W., and Coffin, S. B., (1975). Metabol-ism of lithocholate in healthy man II Entero-hepatic circulation. Gastroenterology, 69, 67-76.

WVan Berge Henegouwen, G. P., Brandt, K. N.,Eyssen, H., and Parmentier, G. (1976). Sul-phated and unsulphated bile acids in serum,bile and urine of patients with cholestasis.Gut, 17, 861-869.

COELIAC/LIVER AND BILIARY

Coeliac disease presenting as malabsorp-tion from the tropics

R. M. MENDELSON, S. G. WRIGHT, AND A. M.TOMKINS (Hospital for Tropical Diseases,London, and Department of HumanNutrition, London School of Hygiene andTropical Medicine) In a series of 274patients fully investigated over a five yearperiod for persisting symptoms of malab-sorption after return from the tropics,85 had giardiasis and 100 had tropicalmalabsorption (including 47 with tropicalsprue).

Thirteen cases had subtotal villousatrophy on initial jejunal biopsy. Six ofthese had persistent severe mucosallesions despite elimination of giardia fromthree and treatment for tropical sprue inone.

Gluten withdrawal in these six patientsresulted in improvement in symptoms andvillous morphology. Reintroduction ofgluten in three thus far has resulted inrelapse.

Before tropical exposure these patientswere asymptomatic and well-nourished.We suggest that acute gastrointestinalinfection precipitated symptoms in thesepatients with coeliac disease.We conclude that (1) a severe mucosal

lesion and (2) the failure to improve aftertreatment for giardiasis or tropical sprueshould suggest the diagnosis of coeliacdisease.

Splenic size and function in coeliacdisease

P. J. ROBINSON, A. W. BULLEN, R. HALL,R. C. BROWN, AND M. S. LOSOWSKY (De-partments ofRadiology, Haematology, andMedicine, Leeds (St. James's) UniversityHospital, Leeds) There is increasing

evidence that hyposplenism has adverseclinical consequences, some potentiallyavoidable.' A substantial proportion ofpatients with coeliac disease have impairedsplenic function as shown by measurementof heat-damaged red cell clearance,2but this method is time-consuming. andcomplex. The technically simpler pro-cedure of scintiscanning has shown smallspleens in some immunologicallyabnormal coeliac patients,3 but cor-relation between splenic size and functionhas not been fully assessed.

Splenic scintiscans were performedafter measurement of S9mTc labelled heat-damaged red cell clearance in 28 coeliacsand six controls. 99mTc sulphur colloidscans were performed in 24 coeliacs and10 controls. Splenic volumes computedfrom colloid scans correlated well withthose from red cell scans (rs = 0-889,p < 0-001). Correlation between splenicvolume computed from red cell scansand heat-damaged red cell clearance inthe coeliacs was good (rs = -0-715,p < 0-001). Eleven out of 28 coeliacshad both prolonged red cell clearance andcomputed splenic volume below thenormal range; three patients were ab-normal on one or other test.

Splenic function is impaired in about50% of patients with coeliac disease.99mTc sulphur colloid scintiscanning pro-vides a simple quantitative method ofassessing splenic function which can beused in exploring the possibilities ofavoidance of the adverse effects ofhyposplenism.

References

'Ammann, A. J., Addiego, J., Wara, D. W., Lubin,B., Smith, W. B., and Mentzer, W. C. (1977).Polyvalent pneumococcal-polysaccharide im-munization of patients with sickle-cell anaemiaand patients with splenectomy. New EnglandJournal of Medicine, 297, 897-900.

2Marsh, G. W., and Stewart, J. S. (1970). Splenicfunction in adult coeliac disease. British Journalof Haematology, 19, 445-457.

3Baker, P. G., Jones, J. V., Peacock, D. B., andRead, A. E. (1975). The immune response to4kX 174 in man. III. Evidence for an associationbetween hyposplenism and immunodeficiency inpatients with coeliac disease. Gut, 16, 538-542.

Is prolactin trophic to the intestine incoeliac disease?

FIONA M. STEVENS, C. F. MCCARTHY, ANDA. CRAIG (Department of Medicine,Regional Hospital, Galway, Ireland; SearleDiagnostic, High Wycombe, Bucking-hamshire) Prolactin has been measuredby radioimmunoassay in the blood of

13 untreated and 42 treated coeliacpatients. Raised values were found infour of the five untreated male patients(558 + 229 mU/l) (mean + 1 SD) (normalrange < 350 mU/I). The levels in sevenof the eight untreated female patients(325 + 160 mU/l) (mean ± 1 SD) fellwithin the normal range (<650 mU/1).

Raised prolactin values (799 ± 1089mU/1) (mean + 1 SD) were found in 35patients responding favourably to thegluten free diet. Significantly lower levels(p < 0-001) were found in seven patientsnot responding to the diet (131'4 ± 45 6mU/I) (mean + 1 SD). The finding ofpersistently raised prolactin values onserial sampling suggests that stress is nota contributing factor. No other knowncause of hyperprolactinaemia could bedefined.

In spite of the evidence in rats thatperphenazine or pituitary transplantinduced hyperprolactinaemia does notaffect the growth of intestinal villi1, thepresent results, together with the villoushyperplasia of lactation in rats, indicatethat a possible enterotrophic role forprolactin, or an associated hormone,needs further investigation.

Reference'Muller, E., Bates, T., Samson, D., and Dowling,R. H. (1977). Is prolactin trophic to the intestine?Gut, 18, A965.

Subcellular distribution of organelle-associated enzymes in untreated coeliacdisease

R. W. LOBLEY, P. W. PEMBERTON, RITAWARWICK, G. I. SANDLE, AND R. HOLMES(University Department of Gastroenter-ology, The Royal Infirmary, Manchester)To investigate the subcellular distributionof organelle-associated enzymes in coeliacdisease, jejunal biopsies from controlsand patients with untreated coeliacdisease were homogenised in isotonicmannitol, and four subcellular fractionsisolated by differential centrifugation.

In coeliac homogenates, the specificactivity of sucrase (brush-border enzyme)was 11 % of control (2P < 0 002); that ofmonoamine oxidase (mitochondria) was22% (2P < 0-002); non-specific esterase(endoplasmic reticulum) was 30%(2P < 0-002); and P-N-acetylglucosami-nidase, ,B-galactosidase and acid phos-phatase (lysosomes) were 99% (NS), 91 %(NS) and 147% (2P < 0'02) respectively.Compared to controls, the proportionof sucrase present in the brush-border



j.com/



ownloaded from

http://gut.bmj.com/


fraction of coeliacs was significantlyreduced and that in the microsomal andsoluble fractions increased. fi-N-Acetyl-glucosaminidase, P-galactosidase, andacid phosphatase were each reduced in thebrush-border and increased in the solublefraction, but acid phosphatase was uniquein showing increased microsomal andsoluble specific activity.Thus in untreated coeliac disease (1)

the fragility of brush-borders and lyso-somes is increased; (2) lysosomal enzymeactivity may be generally unchanged, but(3) acid phosphatase is increased, sug-gesting either a different origin for thisenzyme or a specific derangement; (4)mitochondrial and endoplasmic reticulumfractions may be impaired.

Serum lysozyme levels in malignanthistiocytosis of the intestine

J. R. HODGES, P. ISAACSON, 0. E. EADE, ANDRALPH WRIGHT (Southampton UniversityHospitals, Southampton) A group oflymphomas associated with villous atrophyand malabsorption have recently beencharacterised as malignant histiocytosisof the intestine'. Our finding of a markedlyraised serum lysozyme level in two suchcases suggested that this estimation mayhelp in distinguishing malignant histio-cytosis from uncomplicated adult coeliacdisease.Serum lysozyme levels were significantly

raised in a group of eight patients withmalignant histiocytosis of the intestine(P < 0O001); four were markedly raised. Incontrast, there was no significant dif-ference between groups of patients withuncomplicated adult coeliac disease andhealthy controls. Three lymphomapatients' sera had been stored for threeyears and in only one of these was thelevel raised. Comparison of lysozymelevels in control sera stored for threeyears or more with those obtainedrecently showed significant loss of lyso-zyme activity of long-term storage, sothat the levels of stored sera from lym-phoma patients may originally have beenmuch higher.The estimation of serum lysozyme is a

simple test to perform and may bevaluable in the diagnosis of malignanthistiocytosis of the intestine and, inparticular, in differentiating it fromuncomplicated adult coeliac disease.

Reference'Isaacson, P., and Wright, D. H. (1978). Intestinallymphoma associated with malabsorption.Lancet, 1, 67.

Ulcerative jejunitis: its relationship tomalignant lymphoma

P. ISAACSON AND J. R. HODGES (Depart-ments of Pathology and Medicine, Uni-versity of Southampton Medical School,Southampton) Ulcerative jejunitis (non-granulomatous ileojejunitis) is a conditioncharacterised by multiple ulcers of thesmall intestine which is frequently asso-ciated with long-standing malabsorptionbut which may be preceded by a very shorthistory. The condition carries a highmortality caused by perforation of oneor more ulcers. The histology of theseulcers is usually described as non-specificpolymorphic chronic inflammation andjejunal villous atrophy with crypt hyper-plasia has been a feature of all reportedcases. The finding of small foci of malig-nant lymphoma in three cases presentingas ulcerative jejunitis, the development ofsmall intestinal lymphoma six months afterresection of three jejunal ulcers in anothercase, and the presence of multiple'benign' ulcers in two cases presenting asprimary small intestinal lymphoma, hasled us to conclude that multiple smallintestinal ulcers occurring in associationwith jejunal villous atrophy is a mani-festation of intestinal lymphoma. Themalignant lymphoma is of the typerecently characterised as malignant histio-cytosis of the intestine. Based on theseconclusions, a series of investigations issuggested in patients with unexplainedjejunal ulceration to rule out malignantlymphoma.

Effects of prostaglandin E1 (PGE1) onwater and solute movements in the jejunumof patients with adult coeliac disease (ACD)

R. MODIGLIANI, P. COUZIGOU, E. RENE,M. HAUTEFEUILLE, P. BORIES AND J. J.BERNIER (Unite de Recherches sur laPhysiopathologie de la Digestion U54,H6pital Saint Lazare, Paris, France)In order to test in man the concept thatintestinal fluid secretion originates fromthe crypts of Lieberkiihn1, we assessed thejejunal secretory effect of intraluminalPGE1 (09 ,ug kg-1 min-') in active ACD(n = 11), treated ACD (n = 6) andnormal subjects (n = 48), using theintestinal perfusion technique. A 130mM Na-30 mM mannitol and a 130mM Na-30 mM glucose solutions wereperfused basally and during PGE1administration. In active ACD (1) waterand solutes were malabsorbed (or secreted)

in the basal period; (2) fluid and ionsecretion during PGE, infusion and PGE,net secretory effect (difference in transportbetween basal and PGE, periods) weredepressed (p < 0-01); (3) the normalPGEI-induced inhibition of sodium in-sorption and stimulation of sodiumexsorption were respectively decreased andabolished; (4) normal subjects in whomthe patients' basal secretion was mimickedby perfusing a mannitol-rich solutionstill showed PGE1-induced secretoryrates higher than in ACD (p < 0-001).In treated patients PGE1-induced secre-tion returned towards normal values.Our finding of a depressed secretory

response to PGE1 in the face of markedcrypt hypertrophy does not support theconcept that crypt cells are the majorsite of intestinal fluid secretion.

Reference"Hendrix, T. R., and Bayless, T. M. (1970).

Digestion: intestinal secretion. Annual Review ofPhysiology, 32, 139-164.

Serum conjugated cholic acid in liverdisease

M. VAN BLANKENSTEIN, J. W. 0. VAN DENBERG, M. FRENKEL AND F. J. W. TEN KATE(introduced by J. R. BENNET) (Depart-ment of Internal Medicine II and Depart-ment of Pathology I, Erasmus University,Rotterdam) The hypothesis that fastingand especially post-prandial serum bileacid levels are a sensitive test of liverfunction has been investigated by deter-mining serum conjugated cholic acid(CCA) values fasting and one and twohours after a liquid test meal in over 250patients undergoing liver biopsy.The fasting and highest post-prandial

(stimulated) CCA-values in various diag-nostic categories were compared withvalues found in 23 normal healthyvolunteers (controls).No significant difference in fasting

and stimulated CCA-values was foundbetween controls and patients with:normal biopsies, minimal lesions andchronic persistent hepatitis.

In hepatitis with SGPT < 10 x normal,stimulated but not fasting values weresignificantly raised. In chronic activehepatitis, alcoholic steatosis with fibrosis,cirrhosis, primary biliary cirrhosis, acutehepatitis with SGPT 10-20 x normal,portocaval shunts and acute hepatitis withSGPT > 20 x normal, mean fasting andstimulated values were significantly higherthan controls, with, however, overlap in



j.com/



ownloaded from

http://gut.bmj.com/


all but the last two categories.Stimulated values improved discrimi-

nation only in acute hepatitis withSGPT < 10 x normal and alcoholic stea-tosis. These results indicate that (1)mean CCA-values are significantly raisedin moderate to severe liver disease only;and (2) minimal improvement in sensi-tivity is obtained by post-prandial deter-minations.

Demonstration of a circulating antibodyto halothane altered hepatocytes in patientswith hepatitis

D. VERGANI, G. MIELI-VERGANI, A. ALBERTI,A. L. W. F. EDDLESTON, M. DAVIS ANDR. WILLIAMS (Liver Unit, King'sCollege Hospital and Medical School,London Because many of the clinicalfeatures of halothane hepatitis suggest animmune mediated pathogenesis, we havelooked in patients with this conditionfor circulating antibodies directed againsthalothane altered hepatocyte membranecomponents.

Isolated hepatocytes were preparedfrom rabbits 12 hours after a 60 minuteperiod of halothane anaesthesia, andincubated with sera from two patientswith presumed halothane hepatitis. Thesehad previously been adsorbed with normalhepatocytes to remove any antibodiesreacting with normal cell surface constit-uents. Using indirect immunofluorescence,an antibody coating the surface of halo-thane pre-treated hepatocytes was visibleas a distinct pattem of membranefluorescence. This could be abolished by asecond adsorption with halothane pre-treated hepatocytes.To investigate the possible relationship

between this antibody and cell-mediatedcytotoxicity to hepatocytes, the twoadsorbed sera were incubated in tissueculture with isolated liver cells obtainedfrom halothane pre-treated rabbits. Whenlymphocytes from normal individualswere added to this system, strong cytotoxi-city was induced, but only when usingsera from patients with halothane hepa-titis.Thus patients with halothane hepatitis

have a circulating antibody directedspecifically against halothane alteredhepatocyte membrane components. Thesecan induce normal lymphocytes tobecome cytotoxic to liver cells obtainedfrom halothane pre-treated rabbits.

Anti-LSP antibody in acute viral hepatitis

R. M. MELICONI, A. ALBERTI, D. JENSEN, I. G.MCFARLANE, A. L. W. F. EDDLESTON ANDROGER WILLIAMS (Liver Unit, King'sCollege Hospital and Medical School,London) Immune responses to a livermembrane lipoprotein (LSP) have beenimplicated in the pathogenesis of chronicactive hepatitis. In this study we have useda sensitive radioimmunoassay to lookfor antibody to this antigen in serialserum specimens from 26 patients withacute viral hepatitis; this conditionsometimes progresses to chronic liverdisease. Of 21 patients with an un-complicated course, 20 (95 %) had detec-table anti-LSP antibody in the first twoweeks of the illness with titres of 1:40 to1:3400. Antibody titres were similar inthe 14 type B and the seven non-B cases.In three of the HBsAg positive grouppresenting very early in the course of thedisease, highest anti-LSP values werefound when first tested, before the peakin serum aminotransferases, when viralDNA polymerase levels were high,indicating active viral replication. Anti-LSP and DNA polymerase levels then fellin parallel. In all the patients with anuncomplicated course anti-LSP titres fellrapidly, but in five patients with a pro-tracted clinical illness high levels persistedbeyond three months. These results showthat anti-LSP production in acute typeB hepatitis is not simply secondary toliver damage but may be linked to viralreplication.

Effect of penicillamine on immune com-plexes and immunoglobulins in primarybiliary cirrhosis (PBC)

0. EPSTEIN, D. DE VILLIERS, S. JAIN, B. J.POTTER, H. C. THOMAS AND S. SHERLOCK(Department of Medicine, Royal FreeHospital, London) Twenty eight patientswith PBC were randomly allocated eitherinto a treatment group receiving 600-900mg pencillamine daily, or a control group.Patients were followed for up to 24months. Fifty per cent of patients hadraised levels of circulating immunecomplexes. After 12 months' therapy,serum immune complexes (measured by1211-Clq binding) had fallen significantlyin treated patients compared to controls(p < 0-05) and this effect was sustainedafter 24 months (p < 0-01). InitiallyIgM concentrations were increased in themajority of patients, while IgA and IgG

concentrations were normal. Treatmentresulted in a fall in all three classes ofimmunoglobulins, with IgM concentra-tions being significantly different fromcontrols after six, 12, and 24 months(p < 0-01). There was no evidence thatpencillamine caused deaggregation ofimmune complexes. Aspartate trans-aminase levels fell in response to treat-ment, while a sustained increase occurredin controls (p < 0-01 at six, 12, and 24months), Bilirubin concentrations in-creased at a slower rate in patientsreceiving penicillamine, but the significantdifference between groups at six months(p < 0-01) and 12 months (p < 0 05)was not apparent after 24 months. Inaddition to its copper chelating effect inPBC, pencillamine may favourably in-fluence the course of the disease by itsimmunological action.

Comparative effects of short-term feedingof chenodeoxycholic and ursodeoxycholicacid on hepatic metabolism in man

N. CARULLI, F. ZIRONI, M. PONZ DE LEON,A. PINETTI, AND P. FERRARI (Istituti diClinica Medica, Clinica Chirurgica eChimica organica Universita di Modena,41100 Modena, Italia) Both chenodeoxy-cholic acid (CDCA) and ursodeoxycholicacid (UDCA) administration reducebiliary cholesterol secretion and promotegallstones dissolution. It has been sug-gested that a possible mechanismcould be the reduction of cholesterolsynthesis by the liver. We decidedto compare the effects of short termfeeding of both CDCA and UDCA inequal dose (10-15 mg/kg/day) on bilelipid and biliary bile acid composition,cholesterol content of the liver and thetwo rate-limiting enzymes of cholesteroland bile acid synthesis, HMG-CoA-reductase (HMG-CoA-R) and 7a-hydro-xylase (7a-OH).

Twenty-four gallstones patients wererandomly divided in three groups: (A)eight untreated; (B) eight treated withCDCA and (C) eight treated with UDCA.The treatment was started seven to eightdays before cholecystectomy. In groupsB and C bile samples were obtainedbefore and after the treatment. Theliver samples (100-200 mg) obtained atthe operation were used for biochemicalassays and morphological examinations.The results are shown below by groups.HMG-CoA-R 7a-OHpmol/mg prot.

A994



j.com/



ownloaded from

http://gut.bmj.com/


(A) 64-7 29-0 27-6 + 7-5(B) 27-9 + 165++ 25-6 ± 90(C) 123-5 + 239++ 24 5 ± 34++p < 0-01Micr. Chol Administerednmol/mg prot. BA in bile

(% of total)(A) 78-7 22*1(B) 64*8 ± 116 72-6 + 3.7(C) 59-1 ± 105 53-4 ± 58Neither CDCA nor UDCA treatmentproducedchanges in bilelipidcomposition.We conclude that, in short-term

administration, sufficient to produce apredominance of administered bileacid in the bile acid pool, CDCA andUDCA affect differently the hepaticcholesterol metabolism. Our data suggestthat the mechanism of bile cholesteroldesaturation could be different for thetwo bile acids and may not be relatedonly to reduced cholesterol synthesis.

Hepatic ultrastructure in Gilbert's syn-drome: evidence for two populations

J. DAWSON, D. L. CARR-LOCKE, I. C. TALBOT,AND F. D. ROSENTHAL (Departments ofMedicine and Pathology, Leicester GeneralHospital, Leicester, and Department ofMedicine, Royal Postgraduate MedicalSchool, London) Gross hypertrophy ofthe hepatocyte smooth endoplasmic reti-culum on electron microscopy has beensuggested as a diagnostic feature ofGilbert's syndrome'. Characteristic bili-rubin responses after 48 hours on a 400calorie diet2 or 180 minutes after intra-venous injection of 50 mg nicotinic acid3have also been reported. We studiedhepatic ultrastructure in 24 patients andcorrelated the findings with response tocaloric restriction and nicotinic acid.The smooth endoplasmic reticulum was

hypertrophied in only 12 patients (EMpositive). The remainder (EM negative)did not differ from normal controls. TheEM positive group showed a significantlygreater mean percentage rise in bilirubinaftercaloricrestriction(92-8 + 16-8SEM)compared with the EM negative group(46-1 ± 10-1, p < 0-05), who were similarto normal controls (39-5 + 10 9). Themean percentage bilirubin response tonicotinic acid was higher, but not sig-nificantly so in the EM positive (610 +17-8) compared with the EM negativegroup (32 5 ± 7 4), who were similarto normal subjects (43 5 ± 12 8). The twogroups were otherwise similar in respect

of mean bilirubin (33 7 ± 35 ,umol/l;40-2 + 77 ,umol/l), age, clinical presen-tation and mean red cell survival.We suggest that smooth endoplasmic

reticulum hypertrophy is not a constantfeature of Gilbert's syndrome but is acharacteristic of a distinct subpopulation.

References'McGee, J. O'D., Allan, J. G., Russell, R. I.,and Patrick, R. S. (1975). Liver ultrastructurein Gilbert's syndrome. Gut, 16, 220-224.

"Owens, D., and Sherlock, S. (1973). Diagnosis ofGilbert's syndrome: role of reduced caloricintake test. British Medical Journal, 3, 559-563.

3Davidson, A. R., Rojas-Bueno, A., Thompson,R. P. H., and Roger Williams (1975). Reducedcaloric intake and nicotinic acid provocationtests in the diagnosis of Gilbert's syndrome.British Medical Journal, 2, 480.

SMALL AND LARGE BOWEL/OESOPHAGUS/STOMACH/DUODENUM

Comparison of colonic motor function indiverticular disease and the irritable colonsyndrome

J. HYLAND, C. DARBY, P. HAMMOND, ANDI. TAYLOR (Departments of Surgery andBioengineering, Liverpool) It is oftensuggested that the irritable colon syndrome(ICS) is a pre-diverticular condition andit may be that patients with diverticulardisease and predominant diarrhoea arereally suffering from ICS.

In this study colonic motor function wascompared in 20 patients with diverticulardisease (seven with predominant diar-rhoea), in 20 patients with ICS (sevenwith predominant diarrhoea) and in eightcontrols. Patients with diarrhoea werematched for symptom score, stool weight,and transit time. An intraluminal suctionelectrode was used to record sigmoidmyoelectrical activity and the percentageof 2-4 c/m and 6-9 c/m activity was cal-culated using an automated frequencyanalyser.No statistically significant difference in

the incidence of these two frequencybands was found between patients withdiverticular disease and normal controls.The ratio of the percentage incidence ofeach was 0 85 and 1 1. However, in ICSthe mean incidence of the 2-4 c/mfrequency was statistically significantlygreater than the 6-9 c/m (ratio 2:2).

Treatment with bran (12 g per day)resulted in a normalisation of stool weightand transit time in patients with diver-

ticular disease and diarrhoea but notinvariably so in ICS. Thus, on the basis ofcolonic motor function, there appears tobe no evidence to link these two condi-tions.

Energetics of the colonic epithelial cells(colonocytes) in man: the concept ofenergy deficiency diseases of the colonicmucosa

W. E. W. ROEDIGER AND S. C. TRUELOVE(Nuffield Department of Clinical Surgeryand Clinical Medicine, Radcliffe Infirmary,Oxford) Colonocytes require energy forsodium pumping and for cell growth. Wewished to establish the types of metabolicfuel utilised by normal colonocytes of theascending (AC) and descending colon(DC) in man.Mucosae were obtained from colectomy

specimens (seven AC, seven DC). Colono-cytes were prepared by incubating mucosalstrips in calcium-free Krebs-Henseleitsaline with 10 mM EDTA, 4 mg/mlhyaluronidase or 20 mM DL-dithio-threitol and 0-25% w/v bovine serumalbumin.Aerobic glycolysis was observed in both

AC and DC. Glucose oxidation accountedfor 85-4% of the oxygen consumption inthe AC and 29-6% in the DC. In thepresence of 10 mM n-butyrate these pro-portions decreased to 40-5 % in the ACand 15-8 % in the DC. The percentageoxygen consumption attributable to n-butyrate, when this was the only substrate,was 72-6% in the AC and 75 1 % in theDC. In the presence of 10 mM glucosethese proportions changed to 59'2% in theAC and 72-0% in the DC. We concludethat n-butyrate, under the conditions ofstudy, is the predominant fuel of colono-cytes, particularly those of the DC.

In vivo, n-butyrate produced by bacteriais freely available to colonocytes of man.In the light of the above findings and thework of others, there is the possibility thatantibiotic diarrhoea is due to a mildtemporary deficiency of butyric acid' andthat ulcerative colitis is due to a chronicpersistent failure to utilise n-butyrate.AReferences'Asatoor, A. M., Chamberlain, M. J., Emmerson,

B. T., Johnson, J. R., Levi, A. J., and Milne,M. D. (1967). Metabolic effects of oral neomycin,Clinical Science, 33, 111-124.2Ellestad-Sayed, J. J., Nelson, R. A., Adson, M. A.,Palmer, W. M., and Soule, E. H. (1976). Panto-thenic acid, coenzyme A and human chroniculcerative and granulomatous colitis. AmercanJournal of Clinical Nutrition, 29, 1333-1338.



j.com/



ownloaded from

http://gut.bmj.com/

A996 The British Society of GastroenterologyCholinergic innervation of human smallintestine

A. N. CHATTERJI (introduced by ProfessorL. A. Turnberg) (Department ofMedicine, Hope Hospital (University ofManchester School of Medicine), Salford)It has been suggested that control of iontransport by intestinal epithelial cells maybe influenced by acetyl choline (ACh),released from nerve fibres near themucosal.To explore this hypothesis, the distri-

bution of cholinergic nerve fibres in 15specimens of human small intestinalmucosa was studied by light and electronmicroscopic (EM) cytochemical techni-ques2.

Light microscopy revealed a densenetwork of cholinergic fibres extendingthroughout the intestinal wall. At the EMlevel, abundant nerve fibrils were identi-fied in the lamina propria of the villi andadjacent crypts. However, cholinergicfibres, recognised by specific reaction pro-duct localised to axon membranes3, wereshown not to be in contact with villousepithelial cell membranes, nor in closeopposition to the basement membrane(BM). The minimum separation betweenBM and nerve fibres was at least 1micron. Nerve fibres were in closeapposition to crypt epithelium.

It is concluded that the intestinal wallhas a rich cholinergic innervation, speciallyaround the crypts, although the distancefor ACh diffusion into villous epitheliumthrough basement membrane is consider-able.

These observations would suggest thatif cholinergic fibres influence epithelial iontransport they do so via the crypt ratherthan villous epithelium.

References"Isaacs, P. E. T., Corbett, C. L., Riley, A. K.,Hawker, P. C. and Turnberg, L. A. (1976).In vitro behaviour of human intestinal mucosa.The influence of acetyl choline on ion transport.Journal of Clinical Investigation, 55, 535-542.

2Karnovsky, Morris J. (1964). The localisation ofcholinesterase activity in rat cardiac muscle byelectron microscopy. Journal of Cell Biology, 23,217-232.

'Burnstock, G., and Robinson, P. M. (1967).Localisation of catecholamines and acetylcholinesterase in autonomic nerves. CirculationResearch, Suppl. III, XX, XXI, 43-55.

Demonstration of opiate receptors inintestinal mucosal epithelium

J. S. MCKAY, P. C. HAWKER, B. D.LINAKER, AND L. A. TURNBERG (University

Department of Medicine, Hope Hospital(University of Manchester School ofMedicine), Salford) The demonstrationof opiate receptors in brain and intestinalmuscle raises the possibility that endo-genous opiates (endorphins) play aphysiological role in these tissues. Becauseof the known effects of opiates in diar-rhoea therapy, we searched for evidencethat such receptors exist in intestinalmucosa by which means opiates mayinfluence intestinal transport.

Using an in vitro technique', segmentsof rabbit ileal mucosa, stripped of musclelayers, were clamped between Perspex halfchambers and bathed in oxygenatedbicarbonate/saline buffer. Morphine (2 x10-5 M) caused a fall in potential difference(0-36 mV, P < 0-001) and short circuitcurrent (8-2 uAcm-2, p < 0 001) butresistance was unchanged (n = 20).Morphine increased net chloride absorp-tion (from -019 to + 198 ,uEq cm-2h-1, P < 0-02, n = 10) due to a decrease inthe serosa to mucosa unidirectional flux.Sodium transport was unaffected by thecalculated residual ion flux, probablyindicating that bicarbonate secretion wasincreased. Naloxone (10-6 M) enhancedrather than antagonised this response.Dextromoramide acted similarly butlaevomoramide was inactive, suggestingthat this was not a non-specific effect. Ilealsecretory responses to prostaglandin E2,theophylline or acetyl choline were notblocked by morphine.

These results indicate that opiatesenhance ileal chloride absorption. Theresponse to naloxone suggests the existenceof a novel type of opiate receptor on themucosa.

Reference'Corbett, C. L., Isaacs, P. E. T., Riley, A. K.,and Turnberg, L. A. (1977). Gut, 18, 136.

Migrating motor complex in man: itsassociation with transjejunal potentialdifference

N. W. READ (Department of Physiology,University ofSheffield, Sheffield) We havepreviously decribed a temporal relation-ship between fluctuations in intraluminalpressure and transmural potential differ-ence (PD) in the human small intestine'and given evidence which suggests thatthe latter may represent electrogenicsecretion of anions. We have now investi-gated the association of the PD with therecurrent phases of fasting motor activityknown as quiescence, intermittent activity,

and the migrating motor complex (MMC).Intraluminal pressure and transmural PDwere recorded over seven hours from threejejunal sites (20 cm apart) in eight healthyfasted volunteers, using catheters per-fused with saline. MMCs were observed inall subjects (duration = 10-8 ± 1-4 min;rate of propagation = 3-7 ± 0 4 cm/min;period between complexes = 105 + 12min; mean + SEM). The average PDduring the MMC (-5-3 ± 0 5 mV) wassignificantly higher than the PD duringintermittent activity (-2-7 ± 05 mV),which was significantly higher than thatobserved during quiescence (-1 3 ± 0-3mV). Moreover, the migrating complexwas accompanied by a single largemigrating PD hump, while intermittentactivity was associated with irregular PDfluctuations, some of which accompaniedperistalsis and progressed at a much fasterrate than the MMC (approximately 100cm/min). These results suggest thatdifferent phases of intestinal motor activitymay be associated with different rates ofintestinal secretion.

Reference'Read, N. W., Smallwood, R. H., Levin, R. J.,

Holdsworth, C. D., and Brown, B. H. (1977).Relationship between changes in intraluminalpressure and transmural potential difference inthe human and canine jejunum in vivo. Gut,18, 141-151.

Effects on pyloric reflux of antral andduodenal pacing in the cat

J. MUNK, M. HOARE, C. J. C. KIRK, AND A. G.JOHNSON (Professorial Department ofSurgery, Charing Cross Hospital MedicalSchool, London) Pyloric reflux probablydepends far more on the relative con-traction patterns either side of the pylorusthan on the diameter of the pyloric ringitself.

Bipolar silver recording electrodes werepermanently implanted in the antrum,pylorus, proximal, middle, and distalduodenum. In addition, stimulating elec-trodes were inserted in the antrum,proximal and distal duodenum, and anytwo of these could be stimulated at thesame time. Synchronous radiology moni-tored the movement of 5 ml isotonicbarium sulphate suspension introducedinto the duodenum through an indwellingcannula. Recordings were started at leastone week after operation while theanimals were lightly sedated with Saffan.

Distal duodenal pacing consistentlyproduced pyloric reflux. Synchronousproximal duodenal pacing reduced the



j.com/



ownloaded from

http://gut.bmj.com/


reflux but did not abolish it. Strong antralactivity, produced spontaneously, bypacing or the administration of meto-clopramide, did not completely preventreflux but rapidly emptied the refluxedduodenal juice from the antrum so that itdid not reach the fundus. Prolongedduodenal pacing produced vomiting insome animals.

It is suggested from these experimentsthat the cause of duodenogastric refluxshould be sought in abnormal or inde-pendent duodenal activity rather than thediameter or response of the pylorus itself.

Colonic inhibition of gastric secretion inman

R. JIAN, J. HOSTEIN, C. AYMES, H. BESTERMAN,S. BLOOM, AND J. C. RAMBAUD (ResearchUnit on Pathophysiology of Digestion,Hopital Saint Lazare, Paris, France) Thecolon has been shown to contain endo-crine cells, whose role is unknown. Wetherefore studied in 12 healthy volunteersthe influence of colonic perfusion ofvarious solutes on gastric secretion andcirculating levels of gastrointestinal hor-mones. Gastric secretion was stimulatedby pentagastrin infusion at a constantrate (0-1 ,tg kg-' h-1) and continuouslyaspired through a Salem's tube. Constantinfusion of PEG 4000 allowed assessmentof gastric juice recovery. Four hundredmillilitres of each solute were perfusedinto the colon during 20 minutes after atwo-hour basal period.

Hypertonic (823 mOsm kg-') glucose,mannitol, and saline solutions induced aprompt and sustained inhibition of gastricH+ and pepsin outputs: the mean per-centages of H+ output inhibition were 74(p < 0-01), 66 (p < 0-01), and 68 (p <0-01)% respectively. Isotonic glucose andtriglyceride had a much lower inhibitoryeffect.

Hypertonic glucose and mannitolcolonic infusions had no effect on plasmalevels of GIP, VIP, serotonin, somatos-tatin, and neurotensin. Both solutesincreased circulating enteroglucagon;however, the effect of mannitol was muchweaker than that of glucose.These results suggest that gastric secre-

tion can be inhibited by a presently un-known humoral substance of colonicorigin.

Single scan technique for estimating maxi-mal acid output

T. V. TAYLOR, S. HOLT, G. P. MCLOUGHLIN,

AND R. C. HEADING (Departments ofClinical Surgery, Medical Physics andTherapeutics, Edinburgh Royal Infirmary)Gastric accumulation of 19mTc estimatedby continuous aspiration of secretions hasbeen shown accurately to correlate withacid secretion in the dog' and man2. Usinga non-invasive method of estimating 99mTcuptake by the stomach, it has been shownthat duodenal ulcer patients have a higheraccumulation of isotope than those withgastric ulcer and gastric cancer3. Thisnon-invasive technique has now been cor-related with conventional measurements ofmaximal acid output and the optimaltiming for the scan has been investigated.

Thirty-six patients with mixed gastro-duodenal pathology have had theirmaximal acid output estimated con-ventionally and by O9mTc scanning inresponse to pentagastrin (6 ug/kg).Fifteen minutes after pentagastrin 19mTc(1 m Ci) was given intravenously andscans were performed at 15, 30, and 45minutes. The intragastric activity at 15minutes, expressed as a percentage of thetotal injected dose of the isotope, cor-related with that at 30 and 45 minutes (r009). Of the patients studied four hadpernicious anaemia, five gastric cancerand 27 duodenal ulceration. The correla-tion coefficient between maximal acid out-put and intragastric activity was 0-88.

This non-invasive test of gastric func-tion gives an accurate estimate of maximalacid output using only one scan taken 15minutes after injection of 9BmTc.

References"Wine, C. R., Nahrwold, D. I., Rose, R. C., and

Miller, K. S. (1976). Effect of histamine on*9mTc excretion by gastric mucosa. Surgery,80, 591-594.

'Taylor, T. V., Pullan, B. R., Elder, J. B., andTorrance, B. (1975). Observations of gastricmucosal blood flow using 99mTc in rat and man.British Journal of Surgery, 62, 788-791.

'Taylor, T. V., Bone, D., and Torrance, B. (1977).A non-invasive test of gastric function. BritishJournal of Surgery, 64, 702-708.

Role of gastric acid secretion in the genera-tion of human interdigestive motor activity

W. D. W. REES, L. J. MILLER, J.-R.MALAGELADA, V. L. W. GO (introduced byS. F. Phillips) (Gastroenterology Unit,Mayo Clinic, Rochester, Minnesota, USA)The release of motilin by gastric acidentering the duodenum has been suggestedas a mechanism for generating inter-digestive motor cycles. This study testedthis hypothesis by examining fasting motor

activity and plasma motilin during cime-tidine-induced achlorhydria.

Fasting motor activity was recordedfrom the antrum, duodenum, and proxi-mal jejunum of five healthy volunteersusing intraluminal pressure transducersduring nine hours on two days. Intra-venous saline was given on one day andcimetidine (300 mg/4 h) on the other, inrandomised sequence. During the lastthree hours of each day the duodenum wasperfused with 041 N HCl using a 50 mlbolus and continuous infusion (5 ml/min).Intragastric pH and plasma motilin weremeasured at 15 minute intervals. Cime-tidine produced achlorhydria (meangastric pH: cimetidine = 7-5 ± 0 3; saline= 2-0 ± 04) but failed to alter inter-digestive motor activity (mean cycleduration: cimetidine = 108 ± 16, saline= 90 ± 11 min; mean phase duration:cimetidine I = 64 + 8 II = 39 + 9 III =5 ± 1 min;salineI = 54 ± 6, II = 30 + 7,III = 5 ± 1 min). Cyclic variation inplasma motilin was observed on bothdays (range: cimetidine = 129-349,saline = 119-370 pg/ml). Intraduodenalacid inhibited cyclic motor activity in allsubjects but produced a significant andsustained motilin release.We conclude that (1) interdigestive

motor cycles and cyclic release of motilinoccur in achlorhydria; (2) intraduodenalacid stimulates release of motilin butinhibits interdigestive motor activity; (3)entry of gastric acid into the duodenum isnot essential for generation of inter-digestive cycles of motor activity.

Is the chemical half life of cimetidinerelevant to its duration of action?

V. P. GERSKOWITCH, H. K. ADAM, E. J.DOUGLAS, AND C. L. HUGHES (ResearchDepartment, ICIPharmaceuticals Division,Alderley Edge, Cheshire) The effect ofcimetidine on stimulated acid secretion inrelation to blood levels has been deter-mined in Heidenhain and Pavlov pouchdogs.

Intravenous infusion of cimetidine at0 7 and 10 mg kg-' h-' gave steady stateplasma concentrations of between 800 to1550 ng/ml after two hours. By subsequentchallenge with histamine (30 and 90 bugkg-' h-' s.c.) or pentagastrin (0-5 ytg kg-'h-1 i.v.) this plasma concentration wasfound tocorrespond to 80-95% inhibition ofacid secretion. Observing plasma levels andthe rate of return of acid secretion to pre-determinedcontrol values after discontinu-



j.com/



ownloaded from

http://gut.bmj.com/


ation of cimetidine infusion enabled boththe biological and chemical half life tobe calculated.The effective concentration (EC50) of

cimetidine in these studies (. 600 ng/ml)agrees with published values. The chemicalhalf life was 72 ± 7 (mean ± SE) in con-trast to the biological half life of 37 + 7minutes. The 2:1 ratio in half lives wasobserved in all our studies.The unexpected difference between

chemical and biological half life may bedue to (1) biotransformation or (2)selective elimination from the site ofaction. The latter appears more probab!ein view of our findings of high concentra-tions of cimetidine in gastric juice.

This abbreviated biological half lifemay be of clinical relevance in that certainpatients may have an unexpectedly shorterduration of biological effect.

Effect of cimetidine on gastric acid secre-tory mechanismsPHYLLIS A. MOONEY, JACQUELINE M.WALTERS, J. M. O'DONNELL, AND C. F.MCCARTHY (Department of Medicine,Regional Hospital, University College,Galway and Department ofPharmacology,University College, Galway) Little isknown about the effect of cimetidine onbiochemical mechanisms within gastricmucosal cells, which lead to gastric acidsecretion and transport. The objectives ofthis project were to study the effects ofcimetidine on levels of cyclic nucleotidesand histamine-activated ATPase in gastricmucosa of patients with duodenal ulcer(DU).

Biopsies were taken for diagnostic pur-poses from the antrum and body of 12patients, before and after one month ofcimetidine treatment. Biopsies were alsotaken from patients with dyspepsia whohad normal endoscopies. ATPase wasassayed in biopsy homogenates1 andcyclicnucleotides were measured by radio-immunoassay2.

Results showed that histamine-activatedATPase was present only in biopsies fromthe body of the stomach. This enzymeactivity was higher (p < 0-02) in biopsiesfrom patients with untreated DU, than inbiopsies from patients who had normalendoscopies. There was no differencefound in the histological characteristics ofbiopsies from these two groups of patients.

The levels of histamine-activated ATPasewere reduced (p < 0005) while cyclicAMP levels were increased (p < 0001)after one month of cimetidine treatment.These results suggest that cimetidine mayexact a marked influence at a cellular levelon several biochemical mechanisms in-volved in gastric acid secretion transport.

References'Mooney, P., and McCarthy, C. F. (1974). Studieson the ATPase enzyme systems in human in-testinal epithelial cells. Clinical Chimica Acta,56, 37-47.

'Toney, K. C., Oldham, K. G., and Whelan,J. A. M. (1974). A simple direct assay for cyclicAMP in plasma and other biological samplesusing an improved competitive protein bindingtechnique. Clinica Chimica Acta, 56, 221-234.

Gastric mucosal levels of pepsinogen I andII and their relationship to serum pepsino-gens I and II

L. J. LIBMAN AND I. M. SAMLOFF (Divisionof Gastroenterology, Harbor GeneralHospital, Torrance, California, USA) Thepeptic cell mass for pepsinogen II (PG II)in pyloric and oxyntic gland mucosa, isgreater than for pepsinogen I (PG I)present only in oxyntic mucosa.112 How-ever, serum PG I levels are normallyhigher than serum PG II levels.3 In thisstudy, PG I and PG II levels were deter-mined by rocket immunoelectrophoresis inendoscopic gastric biopsies from 12 con-trols, six duodenal ulcer patients, and sixwith a gastric resection. In controls oxynticmucosal PG I (lug/mg tissue protein), 55*9± 8-4 (mean ± SE) was significantlyhigher (P < 001) than oxyntic mucosalPG II, 21-7 + 4-6. PG II in pyloricmucosa 87 + 1 8, was significantly lower(P < 0-02) than in oxyntic mucosa. SerumPG I (ng/ml), 73-6 ± 11-8 was signifi-cantly higher (P < 001) than serum PG II,32.6 ± 7.4. The mucosal PG I: PG IIratio 2-6 :1, was similar to the serum PGI:PG II ratio 2-3 :1. In duodenal ulcerpatients PG I levels in oxyntic mucosa87'8 ± 14-4, and serum 110O8 ± 29-3,were higher than controls. PG II levels inoxyntic mucosa 22-1 + 5-0, and serum46-2 + 11-9, were similar to controls. Inpatients with a gastric resection both PG Iand PG II levels in mucosa and serumwere lower than in controls (mucosal PG Iand PG II 28-6 ± 9-4 and 11-2 ± 2-5,respectively; serum PGI and PG II 25-5 +

6-9 and 23-6 ± 5-9, respectively). Thus, asin serum, PG I is the predominant pep-sinogen in mucosa. Mucosa pepsinogenlevels are raised in duodenal ulcer patientsand decreased in patients with gastricresection.

References'Samloff, I. M. (1971). Cellular localisation ofgroup I pepsinogens in human gastric mucosaby immunofluorescence. Gastroenterology, 61,185-188.

'Samloff, I. M., and Liebman, W. M. (1973).Cellular localisation of the group II pepsinogensin human stomach and duodenum by immuno-fluorescence. Gastroenterology, 65, 36-42.

3Samloff, I. M. (1977). Radioimmunoassay ofgroup II pepsinogens in serum. Gastroenterology,72, A102/1125.

Diffuse oesophageal spasm (D) and loweroesophageal sphincter spasm (C) aredistinct entities

W. C. WATSON, S. N. SULLIVAN, M. BELSHEIM,AND W. BERRY (GI Unit, VictoriaHospital, University of Western Ontario,London, Ontario, Canada) Thirty-threepatients were allocated to diagnosis D (20)or C (13) on the basis of symptoms andmanometric findings. Another five did notclearly fit either group. The manometricdata were then analysed and compared.Reference points are S-lower oesopha-geal sphinter, F-5 cm proximal, T-10cm proximal. In the basal or unstimulatedstate there was a significant differencebetween the mean pressures (mm Hg) at F(in D 58, in C 43, p < 0-05) but not at Tor S. After methacholine, 12-5 mg sub-cutaneously, significant changes occurred.In group D the mean pressure at T rose55-4 (p < 0001), at F 62-9 (p < 0-00l)butat S only 7-9 (NS). In group C the meanpressure at T rose 14-7 (p < 0-05), at F16-7 (p < 0-05), and atS 49-6 (p < 0-001).The differences between the degree ofchange at each point comparing thegroups are highly significant-at T and F,p < 0-005, and at S, P < 0-001. Re-examination of the indeterminate group offive allowed a diagnosis of D in two. Theremaining three were 'atypical', in thatunusually high basal sphincter pressures(mean 113) fell after methacholine to mean50. This paradoxical response may indicatea group of oesophageal motility disordersdistinct from the entities of diffuseoesophageal spasm and lower oesophagealsphincter spasm.



j.com/



ownloaded from

http://gut.bmj.com/

the british society of gastroenterology · study day on the wednesday a full scientific programme...

Documents