thebritish society of gastroenterology · hydrochloric acid for one month at 40°c, the breaking...

Gut, 1981, 22, A414-A444

The British Society of Gastroenterology

The Spring meeting of the British Society of Gastroenterology took place in Bristol on 9 and 10 April 1981under the Presidencies of Professor A E Read and Dr P B Cotton (Endoscopy) and Professor C G Clark(President Elect). Following the Teaching Day and splinter groups, 104 papers were read to the Societyand 12 posters were on display. Dr N W Read delivered the Research Medallist lecture 'Waves, wind anda tricky passage'. The abstracts of the papers follow.

ENDOSCOPYT1-T5

(T1)Controlled trial of argon laser photo-coagulation for haemorrhage from pepticulcers

S G BOWN, D W STOREY, P SWAIN, J SKIRKHAM, T C NORTHFIELD, AND P RSALMON (The Rayne Institute, UniversityCollege Hospital and The Norman TannerGastroenterology Unit, St. James' Hos-pital, London) We present the results ofa controlled clinical trial from twocentres in London of the efficacy of argonlaser photocoagulation in the endoscopiccontrol of haemorrhage from pepticulcers. From an unselected series of 231consecutive admissions for upper gastro-intestinal haemorrhage, all patients seenat emergency endoscopy to have a pepticulcer were included in this trial if theyshowed stigma of recent haemorrhageand if it was possible to aim the laser atthe source of haemorrhage. One hundredand eight ulcers were seen, 79 had stigmaof haemorrhage and 53 were included inthe trial. Identification of the exactbleeding points was achieved usingstandard endoscopes with meticulouswashing of the ulcer bases. The laserpower, exposure time, and aiming wereoptimised in accordance with our previousanimal and clinical experience.

Stratification was in three groups:ulcers with a visible vessel (39 total, six

actively spurting), those without a vessel(11), and those with an overlying clotpersisting after washing (four). Withvisible vessel 9/19 treated and 9/20untreated rebled. Without a vessel orwith clot, 2/7 treated and 0/8 untreatedrebled. No significant benefit from lasertreatment has been shown. In its presentform, there is little prospect of argonlaser photocoagulation playing a usefulrole in the endoscope treatment of majorupper gastrointestinal haemorrhage.

(T2)Outcome of endoscopic intubation in 100patients with oesophagogastric carcinoma

A L OGILVIE, M W DRONFIELD, R FERGUSON,AND M ATKINSON (University Hospital,Nottingham) Inoperable carcinoma ofthe oesophagus or cardia causes progres-sive relentless dysphagia and an effectivemeans of palliation is important in man-agement. We describe the results ofsuccessful insertion of prosthetic tubes atfibreoptic endoscopy in 100 patients withoesophagogastric neoplasms. Fifty pa-tients had adenocarcinomas and 41 hadsquamous carcinomas. Their ages rangedfrom 50 to 91 years. Five fatal, and ninenon-fatal oesophageal perforations weresustained in 118 intubations. Anotherseven patients died within a week ofintubation, giving an overall mortality of12%. Eighty-five patients survived toleave hospital, with marked improvementin dysphagia. Tube function was good,but occasionally problems arose fromblockage with food or overgrowth of the

A414

tumour, tube migration, or disintegration.In three patients, after radiotherapy tosquamous lesions, it was possible toremove the tube without recurrence ofdysphagia. Median survival time afterintubation was three months, and 28 havesurvived for six months. It would appearthat this is a safe and effective method ofpalliation of oesophagogastric neoplasms.

(T3)Prosthetic tube function after palliativeendoscopic intubation of oesophagogastricneoplasms

F J BRANICKI, A L OGILVIE, M WILLIS, ANDM ATKINSON (University Hospital andDepartment of Physical Chenmistry, Uni-versity of Nottingham, Nottingham) Onehundred and eighteen prosthetic tubeswere inserted at fibreoptic endoscopyacross oesophagogastric neoplasms. Thesewere followed up for a total of 428 tubemonths. Food blockage occurred on 24occasions and this was related to thelength of the tube. Thirteen tubes dis-placed proximally, eight of which had nodistal shoulder. Distal displacement oc-curred once, requiring laparotomy andtwo tubes disappeared without trace.Structural deterioration resulted in dis-integration and fragmentation in twotubes, three and eight months afterinsertion. Using the Instron method,tensile strengths of virgin latex and sili-cone rubber tubes were measured in vitroafter accelerated ageing by intubation invarying concentrations of hydrochloricacid and bile. After incubation in 01 M

on Decem

ber 24, 2020 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.22.5.A414 on 1 M

ay 1981. Dow

nloaded from

http://gut.bmj.com/


hydrochloric acid for one month at40°C, the breaking force of latex fell by31 % (48 Newtons), while there was nosignificant change for silicone rubbertubes. Neither the addition of bile norprior irradiation of the tubes had asignificant effect. There was no change inthe breaking strength of silicone rubbertubes after three months in vivo. It issuggested that prosthetic tubes shouldhave a distal shoulder to minimise dis-placement and be fabricated of siliconerubber for improved durability.

(T4)New disinfecting apparatus for gastro-intestinal fibre endoscopes

H J O'CONNOR, J ROTHWELL, C LINCOLN,S MAXWELL, AND A T R AXON (TheGastroenterology Unit and Department ofMicrobiology, The General Infirmary atLeeds, Leeds) Bacterial contaminationof gastrointestinal endoscopes is a poten-tial source of clinically significant infec-tion. Activated 2% alkaline glutaralde-hyde effectively disinfects fibre endoscopesin the short time available between patientprocedures. Unfortunately, up to 37% ofBritish endoscopy units using glutaralde-hyde are experiencing serious staff hyper-sensitivity reactions, especially contactdermatitis.

In response to this problem we havedesigned a 'closed system' disinfectingapparatus that ensures minimal staffexposure to glutaraldehyde. The appara-tus consists of two Perspex chambersconnected to suction apparatus and mainswater. The endoscope is disinfected for30 minutes before the endoscopy session,for two minutes between patient pro-cedures, and for 10 minutes beforestorage. Tap water flushes the endoscopefree of glutaraldehyde.

Samples were taken for bacteriologicalstudy from the channels and exterior ofthe endoscope before and after eachdisinfection cycle. Gross bacterial con-tamination of the endoscope was foundafter each patient procedure. The wash-ings and swabs from the endoscope werevirtually sterile after two, 10, or 30minutes' disinfection by our method.

Therefore, this relatively simple dis-infecting apparatus offers rapid andadequate disinfection of fibre endoscopes,while protecting endoscopy staff from thepotentially toxic effects of contact withglutaraldehyde.

(T5)An overtube concept in upper gastro-intestinal endoscopy

P B COTTON (Gastrointestinal Unit, TheMiddlesex Hospital, London) Overtubeshave previously been advocated for pro-tection of the oesophagus during endo-scopic removal of foreign bodies. Wehave designed a new tube and greatly~extended the range of applications. Aplastic sleeve 40 cm long is placed overthe proximal shaft of the fibrescope; oncethe fibrescope is in the oesophagus, thesleeve is advanced until its tip lies beyondthe cricopharyngeus. The sleeve allowsrepeated removal and reintroduction offibrescopes without patient discomfortand therefore greatly facilitates retrievalof polyps, gastric suction and lavage (byreplacing the scope temporarily with alarge bore rubber tube), changing thetype of fibrescope, and the endoscopiccarriage of jejunal biopsy capsules. Thesleeve permits the use of muzzle-loaded'jumbo' biopsy forceps, and can providea gas-venting system during laser photo-coagulation. It can also be used in anentirely new method for introducingfibrescopes, which may be easier andsafer for the inexperienced. The patientswallows the tip of a 36 French gaugerubber lavage tube, with the sleeve at itshilt; the sleeve is advanced over the tube,which is then replaced by the fibrescope.Patient toleration is excellent and therehave been no adverse effects.

CLINICAL MEDICAL

T6-T20

(T6)Intra-oesophageal acidmonitoring in asthma

studies and pH

D N COOPER, A PIERRY, M MARPLE, ABERNSTEIN, AND J BANCEWICZ (M H IRVING)(Departments of Thoracic Medicine andSurgery, Hope Hospital, Salford) Anumber of reports have suggested a rela-tionship between gastro-oesophageal re-flux and asthmatic symptoms. This mightbe due either to pulmonary aspiration ofrefluxed acid or to reflux bronchocon-striction induced by stimulation of theoesophagus. We have studied 12 patientswith asthma and gastro-oesophageal re-flux and nine patients with gastro-

oesophageal reflux alone (controls). Pul-monary function tests were performedduring the course of an acid infusion(Bernstein) test. Respiratory symptomswere also recorded during overnightmonitoring of oesophageal pH.

Significant bronchoconstriction wasobserved after oesophageal intubation innine asthmatic patients and four controls.Infusion of N/10 hydrochloric acid intothe lower oesophagus caused broncho-constriction in five asthmatic patients andfour controls. Those with asthma wereaffected to a greater degree. A secondacid infusion was done in the asthmaticpatients after instillation of 4% ligno-caine into the lower oesophagus. Bron-choconstriction was absent or reduced onthis occasion.

Overnight pH monitoring revealedfrequent episodes of reflux with relatedepisodes of cough and wheeze.These results are consistent with the

hypothesis that acid reflux into the loweroesophagus induces reflex bronchocon-striction. In asthmatic subjects this maybe associated with episodes of wheeze.

(T7)Comparison of surgical and medical man-agement of bleeding peptic ulcers

K D VELLACOTT, M W DRONFIELD, MATKINSON, AND M J S LANGMAN (Depart-ments of Surgery and Therapeutics,University and City Hospitals, Notting-ham) There are conflicting opinions asto what the indications for surgery shouldbe in patients with bleeding peptic ulcersas illustrated by answers to question-naires sent to gastroenterologists both inthe United Kingdom and the UnitedStates. Of 1373 patients admitted toNottingham hospitals with upper gastro-intestinal haemorrhage in a 4j yearperiod from mid-1975 to the end of 1979,699 (51 %) were found to have pepticulcers. Two hundred and ninety-eight hadgastric ulcers and 401 had duodenalulcers. The ulcer mortality was 74(10-5%), 32 with gastric ulcers (10.7%)and 42 with duodenal ulcers (10-4%).There has been a decline in the annualmortality rate for gastric ulcers from18-7% in 1976 to 9.4% in 1978, and alsofor duodenal ulcers from 11-9% in 1976to 4-6% in 1978. In addition, the percent-age of patients being operated on hassteadily fallen from 38-8 % in 1975 to16-5 % in 1979. Of the total 699, 196(28 %) had operations, 85 for gastric

A415

on Decem


http://gut.bmj.com

/G


ay 1981. Dow

nloaded from

http://gut.bmj.com/


ulcers and 111 for duodenal ulcers, withmortality of 41 (209 %). The surgicalmortality each year has remained rela-tively constant.We conclude that a more conservative

attitude towards emergency surgery hasresulted in a reduction in the mortalityfor both bleeding gastric and duodenalulcers.

(T8)Experience of an upper gastrointestinalbleeding management team

G G BIRNIE, F KENNEDY, C MACKAY, ANDG WATKINSON (Departments of Surgeryand Gastroenterology, The Western In-firmary aiid Gartnavel General Hospital,Glasgow) As an agreed hospital policyall patients admitted with upper gastro-intestinal bleeding are managed by a dutyteam consisting of a gastroenterologistand a surgeon.

Ninety-four per cent of 769 patientsadmitted in 1978-79 and here reportedwere endoscoped, 96% within 24 hours.Emergency endoscopy within six hourswas necessary in only 14%.

Bleeding chronic peptic ulcers accoun-ted for 36-2% of the series, gastritis for15.10%, and varices for 3 9%. The man-agement team ensured a much lowerrecourse to surgery than in other series-namely, 21 % for gastric ulcer and 91 %for duodenal ulcer-a selection which didnot increase the risk of exanguination,which accounted for the death of onlyone patient of 279 bleeding chronic ulcerstreated. The surgical mortality was cor-respondingly increased in the highlyselected group of 36 patients requiringemergency surgery, reaching 42% forgastric ulcers and 21 % for duodenalulcers. Such patients had rebled, wereusually elderly, suffering from associateddisease, and had ulcers presenting difficulttechnical problems.The overall mortality for the series was

81 %, half the deaths being attributableto malignancy or hepatic failure. Con-servative approach to surgery ensured anacceptably low overall mortality at theexpense of a high operative mortality inpoor risk patients.

(T9)Accuracy of urgent endoscopy in acuteupper gastrointestinal bleeding

J A H FORREST, C ZAMBARTAS, R J CREGEEN,AND N D C FINLAYSON (The Gastro-

intestinal and Liver Service, Royal In-firmary, Edinbutrgh) The accuracy ofurgent endoscopy in acute upper gastro-intestinal bleeding was investigated bycomparing the findings at endoscopy withthose at surgery and/or necropsy per-formed on the same admission. Of 308consecutive endoscopies (performed by anumber of endoscopists with varyingdegrees of experience) 99 (32%) under-went surgery alone and 23 (23 %) of thesedied. Five patients had necropsy alone.The mean age of the 65 males and 39females was 58 years (range 16-88 years).Endoscopy correctly identified the bleed-ing lesion in 79 (76 %) cases, and identifiedthe site of bleeding correctly in another 11(11 %) cases. Excess blood prevented anadequate endoscopy in three (3 %) cases.Endoscopy was incorrect in 11 (1 1 %)cases; four of 24 (l7 %) gastric ulcers,five of 43 (12 %) duodenal ulcers, oneoesophageal ulcer, and gastric erosionsin one patient being missed. Three of thefour gastric ulcers were misdiagnosed aserosions.

In 31 patients the endoscopy andbarium meal findings were correlatedwith those of surgery/necropsy. The twoinvestigations agreed in 32% of cases,endoscopy only was correct in 580%, thebarium meal only was correct in 7 %, andboth were incorrect in 3%.

This study shows that in acute uppergastrointestinal bleeding urgent endo-scopy may miss or misdiagnose 10-150%of gastric and duodenal ulcers.

(T1O)Cimetidine interaction with phenytoin

D J HETZEL, F BOCHNER, J HALLPIKE, ANDD J C SHEARMAN (Royal Adelaide Hos-pital, Adelaide, South Australia) Cime-tidine has been shown to inhibit thehepatic microsomal enzyme oxidationsystem, and drug interaction with cime-tidine resulting in an increase in the effectof warfarin and diazepam has beenreported. Phenytoin, a commonly usedanticonvulsant with a narrow therapeuticindex, also undergoes hepatic microsomalmetabolism. We therefore investigatedthe effect of cimetidine on steady stateplasma phenytoin concentrations in fourpatients with long-standing epilepsy.Serum phenytoin was measured daily

in the week preceding cimetidine adminis-tration, then during six days' treatment

with cimetidine 200 mg three times a daywith meals and 400 mg at bedtime, andup to six days afterwards. All othermedication remained unchanged. In eachpatient cimetidine treatment resulted instatistically significant rises in plasmaphenytoin (P< 002), which fell towardspretreatment levels when cimetidine waswithdrawn. The mean increase in eachpatient was 13%, 20%, 20%, and 33%over pretreatment phenytoin levels. Dur-ing cimetidine treatment one patientdeveloped symptoms of mild phenytoinintoxication which resolved when cime-tidine was withdrawn. Urinary output ofphenytoin and its metabolite pHPPHincreased during cimetidine administra-tion. In another patient receiving pheny-toin intravenously, plasma levels alsoincreased during cimetidine treatment.Thus cimetidine may reduce hepatic

metabolism of phenytoin and caution isnecessary when adding cimetidine tophenytoin therapy.

(T11)Retrograde cholangiogram after cholecys-tectomy

I HAMILTON, W S J RUDDELL, D J LINTOTT,C J MITCHELL, AND A T R AXON (Gastro-enterology Unit and Department ofRadiology, Gener.al Infirmary at Leeds,Leeds) Intrahepatic and extrahepaticbile duct diameter was measured at ERCPin 50 patients without biliary, hepatic, orpancreatic disease (normal subjects), andin 109 patients after cholecystectomy, ofwhom 39 had biliary stricture or retainedstone and 19 were jaundiced.Common bile duct diameter in normal

subjects was 6.5 ±20 mm (mean ±SD)and common hepatic duct diameter 6-02-0 mm. In post-cholecystectomy patientswithout biliary stones or strictures com-mon bile duct diameter (10-0±3.0 mm)and common hepatic duct (9.0 ±30 mm)were greater than in normal subjects(P=< 0-001), and in post-cholecystectomypatients with stones or stricture commonbile duct diameter (14-5 ±4 mm) andcommon hepatic duct diameter (15.55 I4.5mm) were greater than in post-cholecys-tectomy patients with no stones (P=< 0.001), but all groups overlapped.

Ducts of jaundiced patients with stoneswere not more dilated than those of non-jaundiced patients with stones.

Intrahepatic bile ducts were moredilated in post-cholecystectomy patients

A416

on Decem


http://gut.bmj.com

/G


ay 1981. Dow

nloaded from

http://gut.bmj.com/


without stones than in normal subjects(P=< 0001), and, in post-cholecystec-tomy patients with stones, intrahepaticducts were more dilated than in thosewithout stones (P=<0-05), with con-siderable overlap between groups.The bile ducts after cholecystectomy

are dilated compared with normal andduct dilatation does not imply the pres-ence of retained stone. Diagnosticmethods relying on determination of ductcalibre-for example, ultrasound-maybe misleading in the investigation ofpatients after cholecystectomy.

(T12)Pancreatic incorporation of '1C-L-methio-nine evaluated by external detection andduodenal aspirate in normal subjects andpatients with chronic pancreatitis

M CERF, M DOP-NGASSA, AND A SYROTA(Service de Gastroenterologie, HopitalBichat 75018 Paris, SHFJ Departementde Biologie, Commissariat a l'EnergieAtomique 91400 Orsay, France) Theuptake of a natural L-amino acid by thepancreas has never been evaluated inman. L-methionine labelled with "1C, ashort half life isotope (20.4 minutes)produced in a cyclotron was injected in25 normal subjects and in 16 patients withchronic pancreatitis. Data were recordedfor 90 minutes with a gamma camera.Duodenal intubation was simultaneouslyperformed in 14 patients using an infusionof secretin (1 CU/kg/h) and coerulein(70 ng/kg/h) for 30 minutes.

In normal patients time-activity-curvesobtained in areas of interest selected overthe pancreas showed a plateau or a slightincrease of activity with time. In contrast,in patients with chronic pancreatitis theplateau was preceded by a peak lasting10 to 20 minutes.A similar peak of radioactivity was not

detected in the duodenal aspirate suggest-ing that the peak seen by external detec-tion was not due to a passive leakage ofL-methionine from the blood into thepancreatic ducts. The fraction of theamino acid incorporated into proteinswas low during the first 20 minutes andthen increased with time in normalsubjects (51 ±23 % at 70 minutes). It wassignificantly lower in pancreatitis (6±10,P< 0.001). Concentration and output ofbicarbonate and amylase were alsosignificantly lower.

11C-L amino acids could be used for

evaluating protein synthesis by externaldetection.

(T13)Glucose tolerance after acute pancreatitis-short-term and long-term studies

P BRADLEY, M J MCMAHON, AND J K WALES(University Departments of Surgery andMedicine, The General Infirmary, Leeds)Diabetic states are occasionally encoun-tered in patients with acute pancreatitis;yet glucose tolerance after acute pancrea-titis has not been the subject of extensivestudy.

Glucose tolerance tests were carriedout on 88 patients who were convalescentfrom their first attack of acute pancrea-titis, at a mean time of 19 days afteradmission to hospital. By WHO criteria,36 (41%) were diabetic, 45 (51%) werenormal, and seven (8 %) showed inter-mediate values. None had a past orfamily history of diabetes.

In a prospective study of 38 of thesepatients, none of whom had receivedtreatment for diabetes, a second glucosetolerance test was carried out betweenfour and 60 months later (mean 37months). Of the 19 patients whoseconvalescent test showed a diabeticresponse, five (26 %) remained diabetic,10 (53 %) were normal (including the onlypatient to have had a second attack ofacute pancreatitis), and four (21 %) wereintermediate. None of the 19 patientswhose convalescent test was normal (16)or intermediate (three) showed a diabeticresponse, but two (12.5 %) previouslynormal now showed intermediate values.

After the first attack of acute pan-creatitis islet cell damage appears to becommon, and may be of long duration ina few patients.

(T14)Simple screening tests for fat malabsorp-tion: comparison of plasma triglyceride,optical density and 14CO2 breath excretionafter fat meal containing 14C-triolein

G E GRIFFIN, P S WEST, G E LEVIN, ANDJ D MAXWELL (Departments of Medicineand Chemical Pathology, St. George'sHospital, London) Three simple screen-ing tests-measurement of changes inplasma triglyceride, and optical densityafter a standard fat meal, and 14C02breath excretion after ingestion of 14Ctriglyceride-are claimed to discriminatebetween healthy subjects and those with

malabsorption. However, they have notbeen compared together, nor againstmeasurements of faecal fat excretion.

These tests were evaluated againstadjusted faecal fat using non-absorbableradio-opaque markers in 52 patients. Ineight healthy staff faecal collections wereomitted. After a 60 g fat meal containing10 ,uCi glycerol tri I-14C oleate followedby a slice of toast, blood was collectedhourly for six hours and plasma trigly-ceride and optical density measured.Breath 14CO2 was measured hourly fornine hours. After 48 hours' equilibration,stools were collected for three days andfat content adjusted using markerrecovery.

There was no discrimination betweensubjects with normal and raised faecalfat (adjusted) using maximal increases inoptical density or plasma triglyceride.However, the seven and eight hour cumu-lative % 14C02 breath excretion providedexcellent separation, with four apparentfalse positives and no false negativeresults. The simplicity, convenience, andreliability of breath analysis make it anattractive alternative to faecal collectionin screening for fat malabsorption.

(TI 5)Small bowel transit after gastric surgery

C L CORBETT, S THOMAS, N HOBSON, ANDC D HOLDSWORTH (Gastroenterology Unit,Royal Hallamshire Hospital, Sleffield)Small bowel transit time (SBTT) measure-ment can contribute to an understandingof the pathophysiology of diarrhoealstates. This transit time is decreased indiarrhoea due to the irritable bowelsyndrome and can be increased byconstipating drugs.

Small bowel transit time was deter-mined by serial measurement of end-expiratory breath hydrogen after inges-tion of 10 g lactulose in 100 ml wateruntil a rise in hydrogen excretion oc-curred. In seven post-gastrectomy patientswith no diarrhoea, small bowel transittime was 85.0 1i0 min (M ± SE), simi-lar to that of 20 normal subjects (930 ±L6-6 min). In eight post-gastrectomy pa-tients with diarrhoea small bowel transittime was significantly faster (44.3 ± 5.7min, P< 0.01), but in 10 patients withpost-vagotomy diarrhoea it was slightly,though not significantly, slower than innormal subjects (1 12.5 ±130 min).

Small bowel transit time is clearly notreduced in post-vagotomy diarrhoea, in

A417

on Decem


http://gut.bmj.com

/G


ay 1981. Dow

nloaded from

http://gut.bmj.com/


contrast with both the irritable bowelsyndrome and post-gastrectomy diarrhoea.The demonstration that there is prematurepost-prandial return of jejunal fastingmotor complexes after vagotomy has ledto the inference that small bowel peri-stalsis may be increased by vagotomy.Our data do not support this and addi-tional studies of jejunal motor functionafter vagotomy are clearly needed.

(T16)Single dose gluten challenge in treatedcoeliac disease: the first 24 hours

M G BRAMBLE, C 0 RECORD, AND N A

WRIGHT (Department of Gastroentero-logy, Royal Victoria Infirmary, Newcastleupon Tyne) Changes in brush borderenzymes and morphology in the earlyperiod after gluten challenge are poorlydocumented, although a rapid fall inenzyme concentrations has been observed.Using a Quinton biopsy tube we wereable to biopsy the distal duodenum of 10patients with treated adult coeliac diseaseevery hour for the first four hours, afterinstillation of gluten fraction III (=25 ggluten) into the proximal duodenum.Additional biopsies were obtained ateight and 24 hours.

Lactase, maltase, and alkaline phos-phatase concentrations fell rapidly duringthe first three to four hours. Lactaseappeared to be the most sensitive, fallingfrom (mean ±SEM) 23±6 ,umol/min/g X 103 to 14 ± 3 at three hours (P< 0.02).Alkaline phosphatase fell from 531 66,umol/min/mg to 337 i47 during thesame period (P< 0.05). Villous cell popu-lation counts fell from 3310 ± 145 to2633 ±135 cells per villus at four hours(P< 002) and to 2525 ± 182 cells pervillus at 24 hours.

This change was not reflected by directmeasurement of villous height, whichremained unchanged. At 24 hours asignificant increase in the crypt populationcount was observed (P< 0.05).The rapid fall in brush border enzymes

paralleled a similar fall in the villouspopulation count. Toxic effects of glutenappear earlier than is generallyappreciated.

(T17)Crohn's disease of the mouth as an earlyindicator of intestinal involvement

P B MACINTYRE, M A K GHOURI, C SCULLY,

D G MACDONALD, F D LEE, K M COCHRAN,M M FERGUSON, AND R I RUSSELL (Gastro-enterology Unit, Victoria Infirmary, Glas-gow, Department of Oral Medicine andPathology, Glasgow Dental Hospital, andGastroenterology Unit, Royal Infirmary,Glasgow) Crohn's disease can affect anypart of the gastrointestinal tract. A fewpatients have granulomatous lesions inthe mouth, suggestive of Crohn's disease.The prevalence of intestinal involvementin these patients is unknown.We have studied 20 patients (age range

3 to 64 years; mean 25.5 years) whopresented with swelling of the lips and/orface, thickening of the buccal mucosae,mucosal tags, and oral ulcerations. Therewere no gastrointestinal symptoms at thetime of presentation. Oral mucosal biopsyin all patients showed granulomatouschanges suggestive of Crohn's disease.

Gastrointestinal assessment includedtests of haematological, biochemical, andnutritional status, sigmoidoscopy andrectal biopsy, jejunal biopsy (in mostpatients), small and large bowel radiology.

Seven of the 20 patients (35%) werefound to have evidence of intestinalinvolvement. Rectal histology was ab-normal in five patients. Six had lowalbumin, five low haemoglobin, three lowserum folate, and three low serum zinc.Low serum iron, calcium, magnesiumwere also found, and five had a raisedESR.

Patients presenting with these oralfeatures may have intestinal Crohn'sdisease and should undergo a full gastro-intestinal assessment, even in the absenceof symptomatic or clinical evidence ofintestinal disease.

(T18)IndiumII labelled autologous leucocytes inCrohn's disease

S H SAVERYMUTTU, A M PETERS, J P LAVEN-

DER, H J HODGSON, AND V S CHADWICK(Royal Postgraduate Medical School,London) The distribution and fate ofautologous IndiumIII labelled leucocytes,after intravenous injection, was investi-gated in 11 patients with active sympto-matic Crohn's disease and in 11 patientswith the irritable bowel syndrome. Afterintravenous injection of the labelled cells,scans were obtained at two to four hoursand 18 to 28 hours. Stools were collectedfor four days and Indium content mea-sured to determine the percentage ofadministered dose excreted. The time

course of faecal excretion of labelledwhite cells was compared with that oflabelled plasma transferrin in two patientsto compare the kinetics of white cell andprotein loss.

All patients showed normal uptake oflabelled cells in liver, spleen, and bonemarrow. Abnormal intra-abdominal local-isation of radioactivity (positive scan)was found in all the patients with Crohn'sdisease. Faecal counting showed meanrecovery of 9-21 ±SD 9-8% (range 1.9-34%) of the dose. All patients with theirritable bowel syndrome had negativescans and mean faecal isotope excretionwas 04±SD 0.4% (range 0-1-1.0%).The rate of faecal excretion of labelledwhite cells was much more rapid thanthat of labelled transferrin.Indium labelled white cell scanning

with measurement of faecal recovery mayallow regional localisation of Crohn'sdisease and provides an objective mea-surement of white cell loss into theintestine.

(T19)A new look at the cancer risk in ulcerativecolitis

S GYDE, P PRIOR, J A H WATERHOUSE, ANDR N ALLAN (Gastroenterology Unit, TheGeneral Hospital, Birmingham, and CancerEpidemiology Research Unit, Universityof Birmingham, Birmingham) The mag-nitude and pattern of risk of colorectalcancer in ulcerative colitis is uncertain,the apparent risk varying greatly fromone reported series to another. In thisanalysis several models have been used toshow how selection factors, operating inmost reported series, may alter theapparent cancer risk and the pattern ofrisk over time. These models have beenapplied to a hospital series of 676 patientsunder long-term review with ulcerativecolitis. The number of colorectal andhepatobiliary cancers observed in theseries was significantly in excess of thatexpected in a comparable general popula-tion. The colorectal cancer risk variedfrom six to 17 times that in the generalpopulation according to which model wasused for analysis. This variation helps toexplain the widely differing estimates ofcolorectal cancers in published series.The model which seems to most closelyresemble the natural history of thedisease suggests that the risk of develop-ing colorectal cancer increases in the first15 years after onset of ulcerative colitis

A418

on Decem


http://gut.bmj.com

/G


ay 1981. Dow

nloaded from

http://gut.bmj.com/


and thereafter becomes a constant in-creased risk. This pattern suggests thatcancer may be initiated and colitisinduced by a common aetiological agentat or near the onset of colitic symptoms.

(T20)Reticuloendothelial function in coeliacdisease and ulcerative colitis

K R PALMER, D C BARBER, S SHERRIFF, ANDC D HOLDSWORTH (GastroenterologyUnit and Department of Medical Physics,Royal Hallamshire Hospital, Sheffield)It is still not known whether the hypo-splenism which may occur in coeliacdisease and ulcerative colitis is an isolatedphenomenon or is part of a more general-ised reticuloendothelial atrophy. We havetherefore assessed circulatory clearanceof 1251-micro-aggregated albumen (MAA),which is almost exclusively cleared byKupffer cells, in 13 patients with coeliacdisease and 17 with ulcerative colitis.Four and six of these respectively werehyposplenic as assessed by clearance ofheat damaged red cells. Control groupswere six normal subjects and five patientsafter surgical splenectomy.The splenectomised patients had slower

clearance of 1251-MAA than any othergroup (P< 005). As splenic uptake ofMAA is negligible, this implies that afactor produced by the spleen enhancesMAA phagocytosis by Kupffer cells.Hyposplenic patients with ulcerativecolitis or coeliac disease had normal oronly slightly impaired clearance of MAA.The data suggest that hyposplenism in

both coeliac disease and ulcerativecolitis is not associated with generalisedreticuloendothelial atrophy, but thatKupffer cell function can itself be de-pressed as a secondary effect of hypo-splenism. This could account for theincreased susceptibility to perioperativeGram negative endotoxaemia, whichappears to be present in patients withulcerative colitis complicated by hypo-splenism.

LIVER/IMMUNOLOGYT21 T35

(T21)Impaired acetaldehyde oxidation in alco-holism

K R PALMER, W J JENKINS, AND S SHERLOCK(Academic Department ofMedicine, Royal

Free Hospital, London) Many of thetoxic effects of ethanol may be mediatedby its metabolite, acetaldehyde. Alco-holics given ethanol develop high bloodacetaldehyde levels. This could be secon-dary to liver damage, increased formationrelated to induction of ethanol oxidation,or a primary defect of acetaldehyderemoval. Using a sensitive and repro-ducible assay, we have studied ethanolclearance and blood acetaldehyde in ninenormal subjects and 17 recently drinkingalcoholic patients.

Ethanol clearance did not differ be-tween the two groups (199 ± 3-5 (SD) and21.4±5l1 mg/100 ml/h), but the meanacetaldehyde was higher in the alcoholics(17-8 ±6 FM and 10.4±4 FM, P< 001).Thus, the defect is one of acetaldehyderemoval rather than increased formation.Aldehyde dehydrogenase activity, mea-sured on liver biopsies from nine of thealcoholic subjects, inversely related toblood acetaldehyde levels. Those with theleast damage (fatty liver-nine) showedhigher levels than those with more severeliver disease (alcoholic hepatitis andcirrhosis).Although raised blood acetaldehyde

and reduced aldehyde dehydrogenaseactivity may be a consequence of liverdamage, the lack of correlation withhistological severity suggests that thismay not be the whole explanation.

(T22)Granulomas in primary biliary cirrhosis:a prognostic feature

0 EPSTEIN, R G LEE, H 0 JAUREGI, S SHER-

LOCK, AND P J SCHEUER (Department ofMedicine and Pathology, Royal FreeHospital, London) In Crohn's disease,Hodgkin's disease, and leprosy the pre-sence of granulomas is associated withimproved prognosis. To evaluate thesignificance of granulomas in primarybiliary cirrhosis, 295 liver biopsies from100 patients with primary biliary cir-rhosis were reviewed. Granulomas werefound in biopsies of 54 patients and wereless frequently noted as histological stageincreased. They often persisted in mul-tiple biopsies from individual patients.Granulomas did not correlate with clini-cal presentation or biochemical values,but were associated with better survival,as, of the 18 patients who died, only twoshowed granulomas on biopsy. To testthis finding in patients at a uniform stageof progression patients with late primary

biliary cirrhosis were evaluated separ-ately. Patients with and without granu-lomas closely matched clinical, labora-tory, and histological features other thangranulomas. Although 23 of these patientsshowed granulomas, only one of 14 whodied did so (P< 0-01). Thus the presenceof granulomas in primary biliary cir-rhosis seems to be of prognostic signific-ance independently of previously re-ported prognostic indicators.

(T23)Aryl hydrocarbon hydroxylase in alcoholiccirrhosis and primary biliary cirrhosis

K W WOODHOUSE, P H CHAPMAN, 0 F WJAMES, A F MACKLON, AND M D RAWLINS(Departments of Medicine (Geriatrics)and Clinical Pharmacology, The Univer-sity, Newcastle upon Tyne) Aryl hydro-carbon hydroxylase (AHH) is involved inthe tratsformation of polycyclic hydro-carbons to carcinogenic metabolites. Inanimals, hepatic AHH activity is a majordeterminant of the carcinogenicity ofpolycyclic hydrocarbons. Because theoccurrence of primary liver cancer inpatients with primary biliary cirrhosis isstrikingly lower than in alcoholic cir-rhosis, we have compared hepatic micro-somal AHH activity in these two groups,and in histologically normal liver tissueobtained during diagnostic workup. Bothbasal and induced (after 18 hours' incu-bation with benzanthracene) AHH activi-ties were measured.

Basal and induced AHH activity (pmol3-OH-BP/mg protein/h) was 95.2 ±10 3and 152±22 in normal biopsies (n= 8)compared with 54.7±4 and 122±11 inalcoholic cirrhotics (n=10) and 28.2 ±3and 55±7 in primary biliary cirrhosis(n=12). AHH activity in both cirrhoticgroups was significantly lower (p<0-01)than in normal patients, and the activityin primary biliary cirrhosis was signifi-cantly less (p< 0001) than in alcoholiccirrhosis. These differences remainedsignificant when smokers and womenwere considered as separate groups.These results suggest that (1) although

AHH may play a role in the pathogenesisof primary liver cancer in alcoholic cir-rhosis, other factors must be involved;(2) this potential impairment in produc-tion of carcinogenic metabolites inprimary biliary cirrhosis may offer someexplanation for the low incidence ofprimary liver cancer in this group.

A419

on Decem


http://gut.bmj.com

/G


ay 1981. Dow

nloaded from

http://gut.bmj.com/


(T24)Possible recurrence of primary biliarycirrhosis after orthotopic liver trans-plantation

J NEUBERGER, B PORTMANN, B R D MAC-DOUGALL, R WILLIAMS, AND R Y CALNE(The Liver Unit, King's College Hospitaland Medical School, London, and TheDepartment of Surgery, Addenbrooke'sHospital, Cambridge) Three patientssurviving more than two years afterorthotopic liver transplantation for pri-mary biliary cirrhosis were investigated2.6, 3.3, and 3-6 years after transplanta-tion for possible recurrence. Two have arecurrence of jaundice, not due to biliaryobstruction, and one has developedpruritis. Postoperatively, antimitochon-drial antibody levels fell but, despitesome fluctuation, have shown an overallrise since; only one has raised immuno-globulin M. Liver biopsy in one showedonly fibrosis, but in others there werefeatures of primary biliary cirrhosis, withreduced numbers of bile ducts, bile ductdestruction, lymphoid aggregates, granu-loma formation, and slight cholestasiswith increased deposition of copper-binding protein.As a control group, eight other patients

surviving more than two years aftertransplantation were investigated. Nonehad antimitochondrial antibody. Only inone patient did liver biopsy show smallbile duct damage, during an episode ofNocardial infection, but this was notseen on subsequent biopsies. Cholestasis,present in five, was not associated withincreased deposition of copper-bindingprotein.

In chronic rejection, liver histologymay be that of predominant bile ductdamage, and in graft-versus-host diseasethe picture may resemble primary biliarycirrhosis. However, recurrence of primarybiliary cirrhosis is suggested by a rise inantimitochondrial antibody and com-patible biopsies only in those withprimary biliary cirrhosis initially, althoughthe disease is still at a mild and earlystage.

(T25)Dietary protein manipulation in mild (andsubelinical) portal-systemic encephalopathy

KAREL M DE BRUIJN, MADELEINE F BEECHING,CINDY LI, AND LAWRENCE M BLENDIS(Toronto General Hospital and AddictionResearch Foundation, Toronto, Canada)

In order to investigate possible beneficialeffects of vegetable protein in chronicportal systemic encephalopathy, eightpatients with stable cirrhosis and ahistory of spontaneous portal systemicencephalopathy after shunt-surgery werestudied while they were off all antience-phalopathic therapy. Under metabolicconditions for five consecutive one weekperiods, equal amounts of mixed proteinswere alternated with animal or vegetableprotein in a crossover protocol.On admission only three patients had

signs of clinical portal systemic ence-phalopathy, while all had slowing ofEEG, raised ammonia levels and abnor-mal Trailmaking Test performance.During the vegetable diet the apparent

nitrogen (N-) balance (2.1 ±0.66 g/day)tended to be more positive than duringthe animal diet (1.1±063, n=017),because a decrease in the urinary N-excretion (6.1 ±74 vs 7.4±0.87 respec-tively P=007) but no change in faecalN-output. The mean peak frequency ofcomputer analysed EEGs was lowerduring the animal diet (6.58 ± 0.43 Hz)than during the vegetable diet (7.10±0.44Hz P=001). A high arginine intakeduring the vegetable and a high methio-nine intake during the animal diet mayexplain part of this effect. However,plasma levels of amino acids did notchange.

Vegetable protein diets, rather thansimply protein restriction should beconsidered in the management of chronicportal systemic encephalopathy, especi-ally if the patient is undemourished.

(T26)Are platelet antibodies the cause ofthrombocytopenia in patients with chronicliver disease?

I G BARRISON, I D KNIGHT, L VIOLA, MBOOTS, T MITCHELL, AND I M MURRAY-LYON (Gastrointestinal Unit and Depart-ment of Haematology, Charing CrossHospital and Medical School, London)Multiple platelet abnormalities, includingthrombocytopenia, occur in patients withchronic liver disease. Increased autoanti-body activity is also well recognised inthis group of patients. The aim of thisstudy was to establish whether thrombo-cytopenia in patients with chronic liverdisease was due to platelet antibodies.

Twenty-five patients were studied, 14with alcoholic cirrhosis, five with non-alcoholic cirrhosis and six patients with

alcoholic hepatitis and/or steatosis. Plate-let associated immunoglobulins weremeasured by a modification of the anti-globulin consumption technique, and, inorder to establish whether platelet asso-ciated immunoglobulins were specificallydirected against the platelets, immunecomplexes were measured with the Clqbinding assay.Nine of 14 patients with alcoholic

cirrhosis and all five patients with non-alcoholic cirrhosis had raised levels ofplatelet associated immunoglobulins. Im-mune complexes were measured in 18cases (14 cirrhotics). Nine of this sub-group of 14 cirrhotics had raised plateletassociated immunoglobulin levels andfive of these had normal levels of circu-lating immune complexes suggesting thatin the majority platelet associated im-munoglobulins were specifically directedagainst the platelets.We conclude that platelet associated

immunoglobulins may be important inthe pathogenesis of thrombocytopenia inpatients with chronic liver disease.

(T27)Jejunal biopsy and lymphocyte co-culturein coeliac disease

F G SIMPSON, P D HOWDLE, D A F ROBERT-SON, AND M S LOSOWSKY (Department ofMedicine, St. James's University Hospital,Leeds) It has been suggested that animmunological reaction to gluten causesthe jejunal mucosal damage in coeliacdisease.Organ culture of jejunal biopsies has

demonstrated gluten toxicity in vitro inuntreated coeliac disease by quantitativehistology and biochemical techniques.This contrasts with results of culture oftreated coeliac mucosa where glutenproduces only minor histological changes.One possible explanation is that treatedcoeliac mucosa contains fewer effectorlymphoid cells than untreated mucosa.We have therefore cultured jejunalmucosa from 10 treated coeliacs with andwithout gluten and with and withoutadded autologous peripheral blood lym-phocytes.

Biopsies cultured with gluten aloneshowed no reduction in mean enterocyteheight (mean 25.1 ,u, SD 2.2) comparedwith controls (mean 25.4 ,, SD 1.3) butthose cultured with lymphocytes didshow a reduction (mean 24.3 g, SD 0-8,P< 0.05), as did those cultured withgluten and lymphocytes (mean 23.1 gs,

A420

on Decem


http://gut.bmj.com

/G


ay 1981. Dow

nloaded from

http://gut.bmj.com/


SD 1.5, P< 002). Enterocyte height inbiopsies cultured with gluten and lym-phocytes was significantly lower thanthose cultured with gluten alone (P=0-02)or lymphocytes alone (P< 0 05).These results demonstrate that the

combination of gluten and lymphocytesis more toxic to treated coeliac mucosa

than either alone and support the sugges-tion that immune reactions to gluten mayproduce jejunal mucosal damage incoeliac disease.

(T28)Improved technique for the detection ofalpha gliadin sensitised cells in the peri-pheral blood of untreated coeliac patients

CLIONA O'FARRELLY, C. FEIGHERY, J F

GREALLY, AND D G WEIR (Immunology De-partment, School of Pathology, TrinityCollege, Dublin) It has been shown thatleucocytes obtained from coeliac patientswill react with alpha gliadin, a purifiedwheat protein shown to be toxic to coeliacmucosa. However, using the direct or

one-step leucocyte migration inhibitionfactor (LMIF) assay, 100% of treated,but only 500% of untreated, coeliacpatients (and 13 % of controls) respondedto the antigen, which confirms theobservations of Haeney and Asquith. Inorder to reduce the incidence of falsenegative results in untreated coeliacpatients, a more sensitive indirect or two-step LMIF assay has been developed.Peripheral blood lymphocytes are exposedto the antigen to be tested for 72 hours.The cells are separated from the tissueculture medium and the resultant super-natant is assayed for the presence oflymphokines using normal leucocytes as

indicator cells. By this method, 100%0 ofuntreated and 930% of treated coe!iacpatients reacted to alpha gliadin as com-

pared with 150% of control groupscomprising normal subjects and patientswith other gastrointestinal diseases. Thetechnique is highly sensitive, requiringone thousandth the amount of antigen as

the direct method. This study indicatesthere are lymphocytes sensitised to alphagliadin in the peripheral blood of allcoeliac patients. This may be of aetio-logical as well as of diagnostic importance.

(T29)Specific antibody production to milkproteins in the jejunal mucosa of childrenwith cows' milk protein intolerance

J R PEARSON, D KINGSTON, AND M SHINER

(Intestinal Studies Group, Division ofClinical Sciences, Clinical ResearchCentre, Harrow, Middlesex) In cows'milk protein intolerance the local immunereaction in the intestine includes a raisedplasma cell count. To obtain more

specific information we have looked forcells in the jejunal mucosa producingantibody to milk proteins. Jejunal biop-sies from children aged under 2 years,with and without such intolerance, were

examined for plasma cells containingantibodies to the bovine milk proteins3-lactoglobulin and bovine serum albu-

min. A direct immunofluorescence tech-nique with fluorescein-labelled antigenswas used. Antibody-containing cells were

present in 10 of 14 patients intolerant ofcows' milk protein but only one of sevencontrols, all on a diet containing cows'milk, and in two of seven patients withcows' milk protein intolerance on a cows'milk-free diet at time of biopsy. Theantibody-containing cells comprised a

small, varying proportion of the totalplasma cells, equivalent to up to 7% ofthe IgA count. In two patients, doubleimmunofluorescence with rhodamine-labelled antigens and fluorescein-labelledclass-specific anti-human immunoglobu-lins showed the antibody-containing cellsto be 64-80% IgA cells and 20 400% IgEcells. No IgM antibody-containing cellshave yet been identified. Our resultsprovide direct evidence of local produc-tion of specific antibody to milk proteinsin the jejunal mucosa in cows' milkprotein intolerance.

(T30)Further studies in cell-mediated immunity(CMI) to gluten fraction III (GFIII) inthe small intestinal mucosa in coeliacdisease

P D HOWDLE, A W BULLEN, F G SIMPSON,

AND M S LOSOWSKY (Department ofMedicine, St. James's University Hospital,Leeds) The demonstration of cell-mediated immunity to gluten in the smallintestinal mucosa of untreated coeliacsprovided further evidence of the import-ance of immune mechanisms in coeliacdisease. This suggested that sensitisedlymphocytes produce a substance inresponse to gluten, but other explanationsare possible. To provide additionalinformation about such a substance we

have extended our previous work usingpuromycin, a protein synthesis inhibitor.

Small intestinal biopsies were cultured

from 0-5 and 5-25 hours in the presenceand absence of 1 mg/ml GFIII, with andwithout the addition of 15 ,ug/ml puro-mycin. Leucocyte migration inhibitiontests, using normal leucocytes, wereperformed on the culture media, appro-priately reconstituted with GFIII orpuromycin.

Migration inhibition in the presence ofGFIII was demonstrated in the fiveuntreated coeliacs but not in the seventreated coeliacs or three controls, con-firming previous findings. Puromycin didnot prevent this inhibition and was alsoshown to have no non-specific effect onleucocyte migration.These results confirm that a substance

which inhibits leucocyte migration isproduced by untreated coeliac mucosa inthe presence of gluten but suggest thatactive protein synthesis is not requiredfor its release. This casts some doubt uponthe role of cell-mediated immunity in thepathogenesis of coeliac disease.

(T31)Changes in the populations of mucosalimmunoglobulin-containing cells precedinga relapse of ulcerative colitis

E LOPES PONTES, J PIRIS, D P JEWELL, AND

S C TRUELOVE (Gastroenterology Unit,The John Radcliffe Hospital, Oxford)Previous studies have shown an increasein immunoglobulin-containing cells inthe colonic lamina propria of patientswith active ulcerative colitis. In thepresent study, serial observations havebeen made of the immunoglobulin-containing cell populations in patients inremission in order to determine whetherchanges in these populations precede arelapse.Twenty patients with ulcerative colitis

in clinical and sigmoidoscopic remission,who were on no treatment, have hadsigmoidoscopy and rectal biopsy per-formed at monthly intervals for a periodof six months. Normal rectal tissue wasobtained from 20 patients with the irrit-able colon syndrome. Biopsies were fixedin formol sublimate and stained withantisera to lgA, lgG, and IgM by animmunoperoxidase technique. Pointcounting was used to quantify theimmunoglobulin-containing cells.

IgA-, IgG-, and lgM-containing cellswere significantly increased in patients inremission compared with the controlgroup. Nine patients relapsed and inmost of these all three populations ofcells increased three to five weeks before

A42 1

on Decem


http://gut.bmj.com

/G


ay 1981. Dow

nloaded from

http://gut.bmj.com/


clinical relapse was apparent. The greatestproportional rise was seen in the IgG-containing cells which increased six-fold.

These results support the hypothesisthat mucosal antibody production has arole in the pathogenesis of ulcerativecolitis.

(T32)Specific antibody-dependent cell mediatedcytotoxicity (ADCC) against isolated por-phyric hepatocytes in porphyria cutaneatarda

E BARAVALLE AND J PRIETO (introduced byR E Pounder) (Department of Medicine,University Clinic, University of Navarra,Pamplona, Spain) The mechanism ofhepatic damage in porphyria cutaneatarda remains unknown. Rats were madeporphyric by administering hexachloro-benzene. An ADCC test was carried outas follows: hepatocytes from normal andporphyric rats were isolated and used astargets; serum from normal subjects orpatients was added, and normal peri-pheral blood mononuclear cells wereused as effectors. LDH release wasmeasured as an indicator of cytotoxicity.The cytotoxicity index was obtained bythe expression:

LDH released LDH releasedby patient serum - by normal serum

x 100Total LDH content of hepatocytesIn six of eight porphyria cutanea tarda

patients a high cytotoxicity index wasfound with porphyric hepatocytes andvery low or null with normal hepatocytes.The six sera with positive ADCC test be-longed to patients with chronic activehepatitis and/or cirrhosis. Sera from twoporphyria cutanea tarda patients withoutliver disease did not show positive ADCCeither with porphyric hepatocytes or withnormal hepatocytes. Sera from patientswith non-porphyric chronic active hepa-titis showed very low LDH release withporphyric hepatocytes and a positiveADCC towards normal hepatocytes.Alcoholic liver disease serum did notexhibit ADCC with porphyric hepato-cytes nor with normal hepatocytes.

In immunofluorescence studies, por-phyria cutanea tarda serum from patientswith active liver disease showed anintense immunofluorescence with isolatedporphyric hepatocytes, and weak im-munofluorescence with normal hepato-cytes.We conclude that porphyria cutanea

tarda patients have antibodies reactingspecifically with porphyric hepatocytes.

(T33)Different mechanisms of lymphocyte-mediated liver cell damage in 'auto-immune' and HBsAg positive chronicliver disease

G MIELI-VERGANI, D VERGANI, B PORTMANN,I MURRAY-LYON, R THOMPSON, 1 WOOLF,A L W F EDDLESTON, AND R WILLIAMS(The Liver Unit, King's College Hospitaland Medical School, London) Mech-anisms of lymphocyte-mediated damageto hepatocytes have been investigated in57 patients with chronic liver disease byincubating subpopulations of peripheralblood lymphocytes with the patients' ownhepatocytes. In 10 of 16 patients withHBsAg negative chronic active hepatitisnon-T cells showed increased cytotoxicitybut only one, with a previous history ofnon-A non-B hepatitis and no auto-antibodies, had in addition cytotoxicT-cells. In contrast, T-cell cytotoxicitywas found in seven of 12 patients withHBsAg positive chronic active hepatitis.Five also had non-T cytotoxicity. Aggre-gated IgG and liver membrane lipopro-tein blocked non-T cytotoxicity but didnot affect T-cell killing, results consistentwith the proposal that non-T killing isantibody mediated and directed at anormal membrane component. T-cellcytotoxicity was decreased but not com-pletely blocked by purified HBsAg andin 41 patients with a wide range ofchronic liver disease due to HBV infectionthe frequency of T-cell cytotoxicity washigher in HBeAg positive cases, findingsconsistent with a relationship betweenT-cell killing and active viral replication.The results show a clear difference in

immune-effector mechanisms between'autoimmune' chronic active hepatitis andchronic active hepatitis associated withpersistent viral infection. While non-Tlymphocytes reacting with normal mem-brane components may be involved inboth sub-groups, T-cells, probably react-ing with viral determinants, are exclusiveto those with viral infection.

(T34)Drug influences on the enhanced produc-tion of acid hydrolases from macrophagesrecruited into rat livers during liver injury

A R TANNER, A H KEYHANI, AND RALPH

WRIGHT (Professorial Medical Unit,Southampton General Hospital, South-ampton) Earlier studies have demon-strated that macrophages recruited intorat livers during an inflammatory reactionproduce increased amounts of tissue-damaging lysosomal enzymes. In order tostudy the influence of various drugs onthis enhanced production, macrophageshave been isolated from rat livers usingpronase digestion; these cells have beenexposed to the appropriate drugs in vitrofor two hours and the production of anacid hydrolase (n-acetyl-B-glucosamini-dase, NAG) measured at the end of anadditional 24 hours in culture. Bothnormal liver macrophages and macro-phages recruited into the liver after anintravenous injection of Corynebacteriumparvum have been studied. Drugs usedwere prednisolone (200 ng/ml), levami-sole (10 -3 M), azathioprine (100 ng/ml),or 6 mercapto-purine (6 MP, 100 ng/ml).These drugs did not influence normalmacrophage NAG production. Signifi-cant decreases in the enhanced produc-tion by recruited cells were observed inthe presence of azathioprine (P< 0-05)and 6 MP (P< 001). Total NAG (nmolactivity/kg cell protein) for the respectivegroups at the end of the culture periodwere: normal macrophages 0-25 ±004;recruited macrophages, RM 0-63 ±0.08;RM after prednisolone exposure 0.55±0.06; RM, levamisole 0.53 ±004; RM,azathioprine 0.44±0-08; RM, 6 MP0.36±0-06. These changes may representone mechanism of action for the bene-ficial influence of azathioprine on chronicliver disease.

(T35)Does proximity to colorectal carcinomasuppress the immunological competence oflymphocytes?

G H HUTCHINSON, D HEINEMANN, M 0SYMES, AND R C N WILLIAMSON (Depart-ment of Surgery, Bristol Royal Infirmary,Bristol) The finding that lymphoreticu-lar infiltration of solid tumours mayimpart a favourable prognosis prompteda study of lymphocyte immunoreactivityin human colorectal cancer. Tumourdigests (collagenase/DNAse) from 16large-bowel carcinomas were passedthrough a nylon-wool column to separateneoplastic cells and tumour infiltratinglymphocytes. Lymphocytes were alsoobtained from the peripheral blood bycentrifugation on a Ficoll-Isopaque gra-

A422

on Decem


http://gut.bmj.com

/G


ay 1981. Dow

nloaded from

http://gut.bmj.com/


dient. The cytotoxicity of each populationof lymphocytes was assayed againstautoplastic tumour cells, using a two-hour 51Cr release assay.

In 11 of 16 patients peripheral bloodlymphocytes showed moderate cytotoxi-city, whereas tumour infiltrating lympho-cytes showed similar cytotoxicity in onlyfive cases. The cytotoxicity of tumour in-filtrating lymphocytes from 11 patientswas therefore compared before and afterwashing six times in medium 199. Reac-tivity against tumour cells was shown byunwashed tumour infiltrating lympho-cytes in three of 1 1 and by washed tumourinfiltrating lymphocytes in nine (P< 0.02).At effector: target ratios of 10:1 and 20:1the level of reactivity was significantlygreater for washed tumour infiltratinglymphocytes (P=005-0 001).The cytotoxicity of lymphocytes derived

directly from the tumour may be impairedby a blocking factor (perhaps tumour-associated antigen), adherent to thelymphocyte membrane but removed bywashing. Proximity to the carcinoma maysuppress the immunological competenceof lymphocytes.

PHYSIOLOGY IT36-T50

(20 mg/kg enteric). All rats were killed atsix hours.

Raised serum amylase levels occurredin all except the control group.Lung compliance was reduced in the

pancreatitis group (0-61 _±008 ml/cmH20/g) compared with the control group(0.73 -0 1 ml/cm H20/g, P< 005) andthis change was reversed by dexametha-sone (072iO08 ml/cm H20/g, P< 002),heparin (0.74-40.09 ml/cm H20/g, P<0.01), and aspirin (0.77-0 07 ml/cmH20/g, P< 0-001).

Percentage lung weight was raised inthe pancreatitis group (0.45 ±005 %)compared with the sham operation group(037+±0.02O%, P< 001) and this changewas reversed by heparin (0.36±0 04%,P<0 01) and aspirin (0.39+0.03o%, P<0.05) but not by dexamethesone (0.45 ±0.07%, P>0 05).

'Stiff' and heavy lungs occurred inexperimental acute pancreatitis. Thesechanges were completely reversed byheparin and aspirin and partially reversedby dexamethasone, suggesting that pul-monary intravascular coagulation con-tributes to the pulmonary changes ofacute pancreatitis.

(T37)Leupeptin, a low molecular weight trypsininhibitor, prolongs survival in acuteexperimental pancreatitis

(T36)Effect of various drugs on the pulmonarychanges occurring in experimental acutepancreatitis

A R BERRY AND T V TAYLOR (UniversityDepartment of Clinical Surgery, RoyalInfirmary of Edinburgh, Edinburgh) Therespiratory complications of acute pan-creatitis are closely related to prognosisand an undertanding of the mechanisminvolved is vital to improve management.With the use of a rat model of acute

pancreatitis lung compliance and lungweight were studied and the effects ofdexamethasone, heparin, and aspirin onthese parameters in acute pancreatitiswere investigated. Compliance was cal-culated from lung inflation studies afterthe rats had been killed.

Five groups of eight rats were studied:a control sham operation group, a pan-creatitis group, and pancreatitis groupsgiven dexamethasone (3 mg/kg intra-venously), heparin (150 IU/l subcutane-ously and intravenously) and aspirin

PETER JONES, D A W GRANT, J HERMON-

TAYLOR (Department of Surgery, St.George's Hospital Medical School, Lon-don) Acute pancreatitis has a mortalityof about 1I1 % and results in the liberationof pancreatic enzymes into the circula-tion. The mortality is higher in those withthe haemorrhagic form. The kallikreinand trypsin inhibitor Trasylol (aprotinin)failed to reduce mortality or morbidityin recent multicentre trials. This failurehas been attributed to the inability ofTrasylol to abolish catalytic activity oftrypsin in the circulation, where it isbound predominantly to ac2 macroglobu-lin. Inhibition of the a2 macroglobulin-trypsin complex is complete in thepresence of low concentrations of leu-peptin, a low molecular weight peptidealdehyde proteinase inhibitor of micro-bial origin. Its use in pancreatitis has notbeen reported.We studied the effect of leupeptin on

the survival of rats with experimentalacute haemorrhagic pancreatitis. Thiswas induced by controlled, standardised

intraductal infusion of a mixture of thebile salt sodium taurocholate (3.5%) andtrypsin (6000 BAEE units per ml).The survival times of 15 rats given

continuous intravenous leupeptin insaline I ml per hour was 3444 --4.4 hourscompared with 28 0 4 2 hours in 15simultaneous control animals receivingsaline alone (P< 0.001). We conclude thatthe proteolytic enzyme inhibitor leupeptinsignificantly prolongs survival in acuteexperimental pancreatitis.

(T38)Physiological actions of pancreatic poly-peptide on pancreatic exocrine secretion

P G DEVITT, J LONOVICS, S GUZMAN, P LRAYFORD, AND J C THOMPSON (introducedby R Shields) (Department of Surgery,University of Texas, Medical Branich,Galveston, Texas, USA) Pancreaticpolypeptide in pharmacological doses hasan inhibitory effect on pancreatic exo-crine secretion. We have studied the actionof exogenous pancreatic polypeptide, indoses which could be considered physio-logical, on endogenously stimulated pan-creatic secretion.

Five dogs with pancreatic fistulae wereinfused intraduodenally with either hydro-chloric acid or an amino acid mixture.Pancreatic secretion of water bicarbonateand total protein, and plasma concentra-tions of secretin, cholecystokinin andpancreatic polypeptide were measured.Pancreatic polypeptide (0.8 and 2-1 sg/kg/h) was infused for 45 minutes duringthe period of stimulated pancreaticsecretion.

Hydrochloric acid stimulated secretionof water (59.7 - 15-8 ml/45 min), bicar-bonate (764-1-7 mEq/45 min), and pro-tein (1837731-8 mg/45 min), while theamino acid mixture increased secretion ofwater (8.0±1.7 ml/45 min) and protein(183-3+14-3 mg/45 min). Hydrochloricacid increased circulating concentrationsof secretin only, and the amino acidmixture increased plasma concentrationsof cholecystokinin. These increases inpancreatic exocrine secretion were inhi-bited by pancreatic polypeptide: lowdose figures for hydrochloric acid stimu-lation: water (24.4±6.4ml/45 min), bi-carbonate (3.3 4-11 mEq/45 min), pro-tein (50.8 128.8 mg/45 min) (P< 0 05).Pancreatic polypeptide did not affect therelease of secretin or cholecystokinin.We conclude that pancreatic polypep-

tide inhibits pancreatic exocrine secretion,

A423

on Decem


http://gut.bmj.com

/G


ay 1981. Dow

nloaded from

http://gut.bmj.com/


and this action may be of physiologicalsignificance. This inhibition does notappear to be through interference withthe release of secretin or cholecystokinin,and may be mediated by direct action onthe pancreatic acinar cell.

(T39)Duodenal reflux oesophagitis in the rat

H J MUD, S E KRANENDONK, D L WESTBROEK,H OBERTOP, H VAN HOUTEN, AND M VANBLANKENSTEIN (Departments of GeneralSurgery and of Internal Medicine II,University Hospital Dijkzigt-Rotterdam)We planned to study (1) the effect ofreflux of gastric juice, bile, or pancreaticjuice separately or in combination on therat oesophagus; (2) the relation betweenthe presence of active trypsin and/or bileacids in oesophageal washout samplesand the occurrence and degree of oeso-phagitis.Male Wistar rats (n=88) of equal age

and weight were randomly subjected toeither one of seven different reflux-inducing operations. An oesophagealwashout sample was obtained before theselected operation and seven days afterthe operation when the rats were killedfor trypsin and bile acid determinations.In all rats the oesophagus was examinedboth macroscopically and microscopicallyfor signs of oesophagitis.

Oesophagitis followed all operationsthat allowed pancreatic juice to refluxinto the oesophagus, by itself or in com-bination with either gastric juice or bile orboth.

In the case of reflux of pancreatic juiceactive trypsin was detected in theoesophageal washout samples. Thepresence of active trypsin correlated wellwith the occurrence, but not with thedegree, of oesophagitis.The presence of bile acids in the oeso-

phageal washout samples was not associ-ated with the occurrence or the degree ofoesophagitis.

(T40)Ambulatory monitoring of oesophageal pHin reflux oesophagitis using a portableradiotelemetry system

F J BRANICKI, D F EVANS. A L OGILVIE, MATKINSON, AND J D HARDCASTLE (De-partment of Surgery, University Hospital,Queen's Medical Centre, Nottingham)

Long-term oesophageal pH monitoringprovides a measure of the duration ofhigh oesophageal acidity and may thushelp to identify patients at greatest riskof stricture formation.We have monitored gastro-oesopha-

geal reflux occurring in hospitalisedpatients and in the same group at workand in the home using a newly developedpH sensitive radiotelemetry capsule (RTC7006 Rigel) and a portable receivingsystem. A three-aerial chestband detectedfrequency modulated signals emitted bythe tethered RTC positioned 5 cm proxi-mal to the lower oesophageal high pres-sure zone. pH was continuously moni-tored by the portable receiver and aNedilog 24 hour cassette recorder wornon a waist belt, allowing the patientcomplete freedom of movement.

Studies were performed in 10 patientswith graded symptoms of gastro-oeso-phageal reflux accompanied by oesopha-gitis graded at endoscopy and biopsy.The mean number of reflux episodes

per two hours' recording time for hos-pitalised patients was 2-32 (+0 49 SEM)compared with 7-3 (±1-53 SEM) foroutpatients-a significant difference (t=3-08, P< 0.02). Similar analysis for thetotal duration of episodes for bothhospitalised and outpatients gave meandurations of 15.88 minutes (±7.23 SEM)and 39-62 minutes (±4 94 SEM) respec-tively again significant (t= 2.7, P< 0.05).Ambulatory outpatient oesophageal pH

monitoring is useful in assessing patientswith atypical symptoms and may demon-strate significant gastro-oesophageal re-flux when inpatient studies and endo-scopy findings show minimal abnormality.

(T41)Radionuclide transit studies of the oeso-phagus

J N BLACKWELL, W J HANNAN, S HOLT,P TOTHILL, AND R C HEADING (Depart-ments of Therapeutics and of MedicalPhysics, The Royal Infirmary, Edinburgh)Current methods of investigating patientswith motility disorders of the oesophagusare unsatisfactory. Contrast radiology isinsensitive, and manometry is both time-consuming and uncomfortable for thepatient. Our aim was to assess whetheroesophageal radionuclide transit studiesmight prove a satisfactorily sensitive,objective, and non-invasive method foridentifying oesophageal dysfunction.Ten normal volunteers (aged 24-74

years) were studied to ascertain the normalrange and reproducibility of transit times.A bolus of 500 ,1Ci/99mTc sulphur col-loid in 10 ml of water was swallowed, withthe patient positioned supine beneath agamma camera linked to video-recorder.Computer processing allowed analysis oftime activity curves for the upper, middle,and lower thirds of the oesophagus. Theaverage mean transit time in the normalsubjects was 2-5 s (i SD 0.6), and theaverage total transit time was 6.23(l SD 1.1).The procedure was then undertaken on

10 patients with the manometric featuresof diffuse oesophageal spasm and whosuffered from chest pains and dysphagia.The total transit time, which proved themore discriminating measurement, wasprolonged in all cases (range 90-90 s)and accorded with the degree of mano-metric abnormality.

Radionuclide transit measurement is aquick, comfortable, and reproducible newprocedure which should prove useful inidentifying oesophageal motility disorders.

(T42)Effect of natural prostaglandin E2 (PGE2)and derivatives on aspirin-induced fall ingastric mucosal electropotential differencein man

H A CARMICHAEL, L M NELSON, AND R IRUSSELL (Gastroenterology Unit, RoyalInfirmary, Glasgow) We have investi-gated the effect of pretreatment by oraladministration of several PGE2 com-pounds on the fall in electropotentialdifference induced in six normal humanvolunteers by 1200 mg aspirin in 50 mlsaline given orally. The dose of prostag-landin that prevented the electropotentialdifference falling by more than 50% ofthe fall induced by aspirin alone wasdetermined for each prostaglandin in eachsubject.Using the Kendall coefficient of con-

cordance, there was a highly significantagreement between (a) the subjects withregard to ranking of potencies of thecompounds and (b) the compounds onranking of the sensitivities of the subjects(p <001 in both cases). (a) The rankingof the compounds in order of potencywas: 15 (S) 15 Me PGE2 >15 (R) 15 MePGE2 > 16, 16 di Me PGE2 NaturalPGE2. (b) There was marked variation inthe sensitivity of individuals to theprotective effect of the prostaglandins,

A424

on Decem


http://gut.bmj.com

/G


ay 1981. Dow

nloaded from

http://gut.bmj.com/


with > 100-fold difference between cer-tain individuals. The PGE2 compoundsalone did not reduce electropotentialdifference.

Prostaglandin E2 compounds, includingnatural PGE2, prevent breaking of thegastric mucosal electropotential differencebarrier and may be useful cytoprotectiveagents in man. The marked interindivi-dual sensitivity to the protective effect ofPGE2 compounds could be relevant inthe context of susceptibility to pepticdisease.

(T43)Motility of the denervated human stomachand duodenum

D L WINGATE AND A GORCHEIN (LondonHospital, London, and St. Mary's Hos-pital, London) A 34 year old man,subsequently found to have acute inter-mittent porphyria, developed quadri-plegia, respiratory paralysis, and signs ofautonomic denervation after a laparo-tomy. After six months on a ventilatorwith total parenteral nutrition, somerecovery of neck and shoulder movementprompted a study of gastroduodenalmotility in the hope of instituting enteralfeeding. Motility was recorded for 72hours through three perfused lumens ofa nasoduodenal tube with openingslocated in the distal, and mid-duodenum,and in the antrum.During the first 61 hours of study, the

patient received intravenous dopamine,but, as the dose was diminished, antralcontractions at 3/min appeared andpersisted, although varying in amplitude,for most of the rest of the study. Duringthe final 18 hours, some periods of low-amplitude contractions at 10/min werenoted (? antral tachygastria). There wasno evidence of duodenal contractileactivity at any time during the study.Neither the antral motility nor the duo-denal atony were in any way altered byoral feeding, intraduodenal nutrition(Isocal), or intraduodenal neostigmine.The findings suggest that both the

normal periodic fasting motor activityand the normal motor response to food isneurally dependent, but that the neuralcontrol of the stomach and duodenumdiffer; in the antrum it is inhibitory,while the duodenum requires neuralactivation.

(T44)Motility patterns of the human antrumand jejunum and their association withsleep: studies using a radiotelemetrysystem

D F EVANS, G E FOSTER, AND J D HARD-CASTLE (Department of Surgery, Uni-versity Hospital, Queen's Medical Centre,Nottingham) Gastrointestinal pressurescan be conveniently recorded using atethered radiotelemetry capsule. However,this does not permit the recognition ofpropagation of activity, essential if themigrating motor complex and its phasesare to be studied. We have used twolinked radiotelemetry capsules, each tunedto a different frequency and receiver, torecord activity in the antrum and jejunumof 20 fasting volunteers, during the dayand during a night's sleep; during sleepheart rate and electro-oculogram wererecorded to identify REM periods. Acorrelation is seen between REM periodsand a phase 111 activity front in theantrum (x2=19.9, P= <0 001), antralfronts always propagated to the jejunum,the interval between successive jejunalfronts being 136-4 minutes (±6-01 SEM).In 14 subjects fronts were recorded in thejejunum only without preceding antralactivity often occurring shortly before orafter a propagated front; their inclusionin migrating motor complex intervalanalysis reduced the mean interval to107-5 minutes (45.9 SEM). Isolatedjejunal fronts were never associated withpreceding REM sleep periods and com-prised 300% of all jejunal fronts recorded.Two forms of activity front occur in thehuman jejunum: a propagated frontassociated with preceding REM sleepand therefore central influence, andisolated fronts apparently free of suchcentral control.

(T45)Effect of parenteral prostaglandin and aprostaglandin synthetase inhibitor on antralmucosal permeability

H J E LEWI AND D C CARTER (UniversityDepartment of Surgery, Royal Infirmary,Glasgow) Exogenous prostaglandins re-duce monovalent ion flux across fundicmucosa exposed to topical damagingagents, while prostaglandin synthetaseinhibitors alter monovalent ion flux bothin the presence and absence of topicaldamaging agents. The present studieswere undertaken to explore the effect of

prostaglandins and prostaglandin synthe-tase inhibitors on antral mucosal barrierdamage induced by topical bile salts.Dogs were prepared with denervatedantral pouches (n=4). Control studiesdemonstrated that topical 20 mM sodiumtaurocholate (NaT) caused a transientbut significant increase in net H+ lossfrom -91 ± 15 ,umol/30 min to -372+ 51,umol/30 min (p < 0.05) and a correspond-ing increase in net Na+ gain from 145±46,mol/30 min to 407±115 gimol/30 min(p < 0.05).

Intravenous 15 (S) 15 methyl prostag-landin E2 (PG; intravenous bolus dose12,g/kg) reduced net H+ loss induced byNaT from -372± 51 ,umol/30 min to -67±36,mol/30 min (p < 0.05) but caused asignificant increase in net Na+ gain from407±115 F±mol/30 min to 751 ± 130,mol/30 min (p < 0.05).

Intravenous indomethacin (IND; in-travenous bolus dose 5 mg/kg) similarlyreduced net H+ loss induced by NaTfrom -372±51 ,umol/30 min to -51±18t,mol/30min (P<0-02) but had no signi-ficant effect on net Na+ gain.We conclude that both exogenous

prostaglandin E2 and indomethacin re-duce taurocholate-induced net H+ flux,but may do so by differing mechanisms.Prostaglandin E2 may stimulate Na+ richsecretion while decreasing the severity ofNaT induced damage, while indo-methacin directly affects monovalentionic permeability of the antral mucosa.

(T46)Histamine is not the final common pathwayfor gastric acid secretion

W K MAN, C J H INGOLDBY, AND J SPENCER(Department of Surgery, Royal Post-graduiate Medical School, HammersmithHospital, London) The relationship be-tween gastric acid secretion and gastricand plasma histamine release, after penta-gastrin infusion, is so close as to suggestthat histamine is acting as a transmitterto induce acid release.We have attempted to discover whether

a similar pattern of histamine releasefollows insulin-induced gastric acid secre-tion.Nine patients with proven duodenal

ulcer were studied by a combined insulin/pentagastrin gastric secretion test. Acidoutput, gastric histamine output, andplasma histamine levels were measured.There were no significant changes in

gastric or plasma histamine levels during

A425

on Decem


http://gut.bmj.com

/G


ay 1981. Dow

nloaded from

http://gut.bmj.com/


insulin-induced acid secretion (medianbasal 0-98; median highest fraction 0.93;median basal plasma 4.06; medianhighest plasma 6.68).

After pentagastrin infusion there wasan immediate increase in gastric hista-mine to a peak (median 2.66, P<0.05)followed by a fall and then a second peak(median 4.89, p< 0.001). The biphasicnature of the gastric histamine release isunexplained. The occurrence of acidsecretion without histamine release duringinsulin infusion suggests that the changesinduced by pentagastrin are not simply apassive 'washout' of histamine duringacid secretion.The lack of histamine release during

insulin infusion is not compatible withthe theory that histamine is the commonfinal transmitter to the parietal cell.

(T47)SK&F 93479 is a long-acting inhibitor ofgastric acid secretion in man

W L BURLAND, A C CLANCY R H HUNT,JANE G MILLS, DIANA VINCENT, AND G J

MILTON-THOMPSON (Smith Kline andFrench Research Ltd, Welwyn GardenCity, Herts; Royal Naval Hospital, Haslar,Gosport, Hants) The H2-receptor anta-gonists cimetidine, oxmetidine, and rani-tidine differ in potency but not in durationof action. SK&F 93479, a non-imidazoleH2-antagonist, is 16 times more potent ona molar basis and longer acting thancimetidine.We studied the effect of oral cimetidine

3 mg/kg and SK&F 93479 0-25 mg/kg onthe gastric secretory response to mealsgiven 30 minutes or one hour and sixhours after dosing using a modifiedintragastric titration technique.Mean two hour acid output in six

healthy men during the first meal wasreduced from 64.6 mmol (placebo) to21-2mmol (67%) by cimetidine and to23.7 mmol (63%) by SK&F 93479. Asustained response to SK&F 93479 oc-curred in all subjects but in four this wasachieved only after the first hour. Meanacid output after the second meal wasreduced from 59-2 mmol to 46-7 mmol(21 %) by cimetidine and to 34-0 mmol(43 %) by SK&F 93479. The differencebetween drugs was significant (p < 0.05).Our re-sults suggest that the antisecre-

tory effect of SK&F 93479 is prolongedand that on a molar basis it is approxi-mately 20 times more potent than cime-tidine. The maximum antisecretory effect

may be maintained beyond three hoursand a significant effect for longer thaneight hours. Additional studies are re-quired to examine these possibilities.

(T48)Gut-glucagon and GIP may cause sufficienthypoglycaemia in duodenal ulcer patientsto activate the vagus nerves

N F KNIGHT, S ZAHEDI-ASL, D J SANDERS,A P MARR, C W VENABLES, AND A MOODY(Departments of Phlysiology and Surgery,The University, Newcastle utpon Tyne)We have previously reported to thissociety that after eating a standard break-fast, a group of patients with endoscopic-ally proven active duodenal ulcer hadsignificantly higher plasma insulin andsignificantly lower glucose concentrationsthan age and sex-matched controls.The hypoglycaemia in duodenal ulcer

patients may be very important becauseit was recently shown that hypoglycaemiaover the range 5-3 mM produces gastricacid secretion. Our duodenal ulcerpatients went below 3 mM glucose fromabout 75 to 180 minutes after feeding.We have now measured the plasma

glucose-dependent insulinotrophic pep-tide (GIP) and shown that control andduodenal ulcer basal are not significantlydifferent. There is a significantly greaterGIP response in duodenal ulcer patientsfrom 15 to 120 minutes after feeding andthis probably accounts for the highinsulin response in these patients.

In addition, the significantly higherplasma gut-type glucagon concentrationsseen in duodenal ulcer patients from 15to 120 minutes after feeding may alsohelp to account for the hypoglycaemia inthese patients. It is known that gut-typeglucagon can bind to liver cell glucagonreceptors and this may prevent the actionof pancreatic glucagon in stimulatingglucose release from the liver.The increased release of gut-type

glucagon and GIP in duodenal ulcerpatients may be secondary to the ulcera-tion or it may be a primary change. Itwould appear to produce sufficient hypo-glycaemia to activate the vagus.

(T49)Indomethacin inhibits duodenal HCO3secretion: protection by prostaglandin

J N L SIMSON AND W SILEN (introduced byProfessor I McColl) (Department of

Suirgery, Harvard Medical School, andGuy's Hospital, London) The effects ofindomethacin and 16-16 dimethyl PGE2(16 dm PGE2) on duodenal HCO3'secretion were studied in vitro. Strippedamphibian duodenum was mounted inUssing chambers and HCO3' secretionmeasured by pH stat. Indomethacininhibited active HCO3' secretion by 36±3% at 10-5 M and 58±6% at 5 x 104M.16 dm PGE2 (10 -5M) stimulated activeHCO3' secretion by 151±8% of controlvalves, and completely reversed theinhibitory effect of low-dose indomethacin(10- M). The effect of high-dose indo-methacin (5 x 10 4M) was significantlymitigated by 90 minutes' pretreatmentwith 16 dm PGE2 and was completelyprevented by simultaneous addition of16 dm PGE2, but was not reversed if16 dm PGE2 was added 90 minutes afterindomethacin. 16 dm PGE2 demonstratedprotection only when given on the serosalside of the tissue; mucosal administrationwas ineffective.We conclude that 40% of duodenal

HCO3' secretion is driven by endogenousprostaglandins and this component isabolished by low-dose indomethacin.Exogenous prostaglandins on the serosalside stimulate HCO3' secretion, reversethe inhibitory effects of low-dose indo-methacin, and protect HCO3' secretionagainst high-dose indomethacin. Suchobservations provide a possible explana-tion for the observed but unexplainedulcerogenic effects of indomethacin andprostaglandin 'cytoprotection' in animaland human duodenum.

(T50)Nicotine-induced relaxation of the ampullaof Vater

HELEN L LEATHARD AND T G ALLEN-MERSH(Departments of Pharmacology and ofSurgery, Charing Cross Hospital MedicalSchool, London) Nicotine, at levelscomparable with those caused by cigarettesmoking, stimulates duodenal motilityin vivo. To investigate the possibleinfluence of nicotine on bile and pancrea-tic juice propulsion we have recorded itseffects on strips of bile duct, pancreaticduct, and longitudinal and circularduodenal muscle from the dog.The ducts were dissected free of duo-

denal muscle and cut spirally to recordsphincter muscle responses. Thin strips ofduodenum were cut parallel to the longi-tudinal or circular muscle fibres. The

A426

on Decem


http://gut.bmj.com

/G


ay 1981. Dow

nloaded from

http://gut.bmj.com/


strips were suspended in Krebs' solutionat 37°C and responses recorded isotonic-ally (1 g load).

Preparations from five dogs yielded aconsistent pattern of responses to nicotine(100 ng/ml, 10 t±g/ml): the bile ductsphincter was predominantly relaxed;duodenal circular muscle gave a biphasicresponse of weak relaxation and thencontraction; duodenal longitudinal musclecontracted strongly; pancreatic ductfailed to respond. These responses re-flected those caused by stimulating theintramural nerves electrically at 4 Hz(1 ms pulses, supramaximal voltage).

These results suggest that the bile duct,but not the pancreatic duct, can respondindependently of the surrounding duo-denal muscle. By relaxing the bile ductsphincter, nicotine could cause prematureflow of bile into, or promote reflex fromthe duodenum, depending on prevailingpressures.

(73 %, and 860% respectively). Parenteraldrug abuse was more common in thenon-A, non-B group than in the hepatitisB virus group (55 % vs 36%). None of thenon-A, non-B group was homosexual incontrast with the hepatitis B virus groupwhere 18% were homosexuals. None ineither the non-A, non-B or hepatitis Bvirus groups worked in health-relatedoccupations. Twenty-eight per cent of thenon-A, non-B hepatitis patients hadevidence of past infection with hepatitisB virus.

In London, acute non-A, non-B hepa-titis accounts for 16% of all acute viralhepatitis. An important risk factor isparenteral drug abuse. Sexual transmis-sion was not noted. There was no associa-tion with health-care occupations. Hepa-titis A virus is commonly transmittedamong male homosexuals.

(F2)Primary biliary cirrhosis-a frequentlyasymptomatic disease

PLENARYF1-F9

(F1)Prospective study of acute non-A, non-Bhepatitis in a London population

M BAMBER, A MURRAY, B BANNISTER, H CTHOMAS, AND S SHERLOCK (Departmentof Medicine, Royal Free Hospital, London,and Coppetts Wood Hospital, MuswellHill, London) The epidemiology ofnon-A, non-B hepatitis unrelated totransfusions, was studied in 70 consecu-tive cases of acute viral hepatitis admittedto a London infectious diseases hospital.Non-A, non-B hepatitis was diagnosed ifserological markers of recent hepatitis A(IgM anti-HAV) and B (HBsAg, HBcAb),Epstein-Barr virus, and cytomegaloviruswere absent. Twenty-nine (41 %) of thecases were due to hepatitis A virus, 28(40%) due to hepatitis B virus, two (3 %)due to Epstein-Barr virus, and 11 (16%)due to non-A, non-B hepatitis.

In the hepatitis A virus group nine(31%) of the patients were homosexualmales.The non-A, non-B and hepatitis B

virus groups were similar in their averageage (31 years, and 28 years respectively)and predominant occurrence in males

O F W JAMES, A F MACKLON, AND A J

WATSON (Departments of Medicine(Geriatrics) and Pathology, University ofNewcastle upon Tyne) Primary biliarycirrhosis is thought to be a disease inwhich virtually all patients have markedrises in serum alkaline phosphatase. Onlya small proportion of patients are thoughtto be asymptomatic. We present 84patients with primary biliary cirrhosis(seven males), all histologically confirmed,83/84 had antimitochondrial antibody(AMA) titre > 1/40. There were fourgroups. Group 1: 12 patients (14%)asymptomatic of liver disease havingnormal alkaline phosphatase, bilirubin,aspartate aminotransferase (LFTs). Allwere detected on review of +ve anti-mitochondrial antibody found duringinvestigation of an unrelated problem(usually thyroid or joint disease). All hadabnormal liver biopsies, seven were diag-nostic of primary biliary cirrhosis, fivehad raised serum gamma GT, five raisedserum IgM. After mean follow-up 4.0years two had developed abnormal LFTs,none had symptoms of liver disease.Group 2: 28 patients (33 %) asymptomaticof liver disease but with abnormal LFTs.After mean follow-up of 2-5 years fivepatients had developed symptoms ofliver disease, one died of liver failure.Group 3: 36 patients (43 %) presentedwith symptoms suggesting liver disease.

After mean follow-up of 3-7 years six haddeveloped complications of portal hyper-tension, three died. Group 4: eight patients(10%) presented with complications ofportal hypertension. After mean follow-upof 2.5 years three died.We conclude that many patients with

primary biliary cirrhosis (46% here) areasymptomatic, an unknown proportion(14% here) have normal LFTs but signi-ficant liver pathology. Asymptomaticpatients have an excellent prognosis.

(F3)Advantages and disadvantages of percu-taneous transhepatic biliary drainage aspart of a staged approach to obstructivejaundice

G A D MCPHERSON, I S BENJAMIN, B NATH-ANSON, I B BLENKHARN, N B BOWLEY ANDL H BLUMGART (Departments of Sur-gery, Bacteriology, and Radiology, RoyalPostgraduate Medical School, London)A two-stage approach has long beenadvocated for high-risk patients withobstructive jaundice. The recent introduc-tion of percutaneous transhepatic biliarydrainage as a non-surgical first stage hasbeen reported as a significant advantagein the preoperative management of thejaundiced patient. No reports, however,have provided a critical assessment ofpercutaneous transhepatic biliary drain-age based on defined criteria for selection.We have treated 26 patients by percu-

taneous transhepatic biliary drainage (19with high bile duct tumours, and sevenwith low bile duct or pancreatic tumours),of whom 23 came to second-stage surgery.Satisfactory changes in plasma bilirubin(T1/2= 10-4 days) and renal functionwere obtained, but the incidence of com-plications rose with increasing durationof drainage. Infection was the mainproblem. Seven (27%) had infected bileat the time of drainage. This high inci-dence for malignant obstruction mayhave been related to previous surgery.Serial bile sampling showed that theincidence of colonisation of the drainagesystem increased progressively after thefirst four days. This colonisation signifi-cantly affected the operative outcome.There were two deaths in the post-operative period which were secondary toinfection arising in the drainage system.Exogenous infection, common in theearly patients, was later excluded byadding povidone iodine to the system.

A427

on Decem


http://gut.bmj.com

/G


ay 1981. Dow

nloaded from

http://gut.bmj.com/


(F4)Duodenoscopic sphincterotomy in patientswith gallbladders

A G VALLON AND P B COTTON (Gastro-intestinal Unit, The Middlesex Hospital,London) Duodenoscopic sphincterotomyis widely used in patients with duct stonesafter cholecystectomy; increasingly pa-tients are referred who still have gall-bladders in place. We have attemptedsphincterotomy in 56 patients (mean age73 years) presenting with jaundice andcholangitis. Sphincterotomy was possiblein 55 patients, and duct clearance was

achieved in 48 (87 %). Stones weregreater than 15 mm in diameter in five ofthe seven patients whose stones could notbe extracted. There were no fatal com-plications. One patient required a bloodtransfusion, and two needed urgentcholecystectomy for acute cholecystitiswithin seven days of sphincterotomy.Clinical follow-up (mean 14 months,range six to 48 months) has been possiblein 44 of the 46 remaining patients withgallbladders and successful bile ductclearance. None has suffered cholangitisor jaundice; nine have undergone electivecholecystectomy because of relative youthand fitness, and four others because ofbiliary pain. from six weeks to one yearafter sphincterotomy. The remaining 31patients are well. Duodenoscopic sphinc-terotomy is recommended for acutely illpatients with duct stones even in thepresence of a gallbladder; longer follow-up is required to judge the indications forsubsequent cholecystectomy, but it seemsreasonable to postpone this indefinitelyin many elderly and frail patients.

(F5)Pancreatic synthetic rates: a new test ofpancreatic function

E J S BOYD AND K G WORMSLEY (Depart-ment of Therapeutics, University of Dun-dee, Dundee) Ten patients with noevidence of pancreatic disease and sevenpatients with unequivocal chronic pan-creatitis (two of whom had normalenzyme outputs in response to secretin/CCK) had synthetic rate estimationsusing the incorporation of Se75-methio-nine into pancreatic proteins. After thepatient had been fasted overnight a tubewith vented gastric and duodenal aspira-tion channels was positioned under radio-logical control. Pancreatic secretion wasstimulated for two hours by constant

intravenous infusion of secretin (1 clinicalunit/kg/h) and cholecystokinin (1 Ivy Dogunit/kg/h). Se75-methionine (3.0 micro-curies/kg) was added to the infusion.Pancreatic juice was collected in 15minute batches and assayed for pH,bicarbonate, and enzymic activities. Eachsample was immediately mixed with 0-6 NTCA and the activity of the precipitatedprotein counted in a gamma-counter.After the infusion, tin colloid was givenby bolus intravenous injection and apancreatic subtraction scan performed.The rate of incorporation of Se75-methio-nine into pancreatic proteins was signifi-cantly greater in all the patients withchronic pancreatitis, suggesting that inchronic pancreatitis the surviving acinarcells undergo compensatory increase intheir capacity to synthesise and secreteproteins. Synthetic rate studies can detectchronic pancreatitis before secretoryinsufficiency occurs.

(F6)Ergometrine provocation in the diagnosisof oesophageal spasm

H ALBAN DAVIES, M KAYE, A M DART,A H HENDERSON, AND J RHODES (Depart-ment of Gastroenterology and Cardiology,University Hospital of Wales, Cardiff)Oesophageal disease frequently mimicscardiac pain; we have investigated 42patients with apparent angina but nocardiovascular abnormality (all had exer-cise tests and 35 had coronary angio-graphy), by giving them ergometrine toprovoke oesophageal spasm.

Oesophageal motility was studied witha triple lumen assembly perfused withwater (05 ml/min) by syringe pump.Ergometrine (500-1000,ug intravenously)was given to the patients as well as to 10volunteers after an injection of saline, andfull resuscitation facilities were available.The tracings were masked and read'blind' by an experienced observer.Barium swallow examinations were alsoperformed.Asymptomatic oesophageal spasm was

present in nine patients at rest and therewere minor abnormalities in another 13.After ergometrine 22 of the 42 patientshad oesophageal spasm and 19 of themexperienced their typical pain. Anothereight patients developed pain withoutdefinite spasm; seven of these showed anincrease in abnormal oesophageal mo-tility, but no coronary vasospasm, after

ergometrine. Twelve patients had neitherpain nor oesophageal spasm.

In 10 of the 29 patients with eitherproven or suspected oesophageal spasm,this problem was not considered untilergometrine was given. Ergometrine pro-vocation alone saved many patients fromunwarranted invalidism as cardiaccripples.

(F7)Effect of vagotomy on gastric nitrosamineproduction

P I REED, P L R SMITH, F R HOUSE, ANDC L WALTERS (Gastrointestinal Unit,Wexham Park Hospital, Slough, Berk-shire; British Food Manufacturing Indus-tries Research Association, Leatherhead,Surrey, and Department of Pharmacology,Guy's Hospital Medical School, London)Gastric cancer develops more frequentlyin achlorhydric conditions, includingpartial gastrectomy and vagotomy withdrainage, a recent report noting a muchshorter induction period after the latterprocedure. There is strong evidenceindicating that N-nitroso compounds,formed by gastric flora converting dietarynitrate (via nitrite) and dietary or endo-genous nitrosatable amines, may beimportant in gastric carcinogenesis inachlorhydria. Animal studies using N-Methyl-Nl-nitro-N-nitrosoguanidine havedemonstrated accelerated gastric cancerformation after vagotomy. Recently wedemonstrated a highly significant rela-tionship between a rise in gastric pH,increase in total N-nitrosamine concentra-tion and growth of nitrate reducing gutflora in upper gastrointestinal disease,including cancer, and also during cime-tidine treatment.

Forty-eight fasting gastric juice speci-mens from 29 vagotomised patients, 26when not treated and 22 during cimeti-dine treatment, were analysed bacterio-logically, for pH, nitrite, and total N-nitroso compounds. Vagotomy resultedin significantly higher bacterial growth(P 0-014) and mean N-nitrosamine con-centration (P< 002) compared with nor-mal controls, especially when performedmore than six years previously. Levelsalso rose during cimetidine treatment,significantly so in six patients studiedserially, N-nitrosamines (P < 0-002) andpH (P < 0-05). These data confirm thesignificant relationship between pH andN-nitrosamine production especially inachlorhydria. More detailed studies are

A428

on Decem


http://gut.bmj.com

/G


ay 1981. Dow

nloaded from

http://gut.bmj.com/


required to establish which vagotomyoperation enhances this risk.

(F8)Late mortality after curative surgery forpeptic ulcer

A H MCLEAN ROSS, J R ANDERSON, M ASMITH, AND W P SMALL (Gastric Follow-up Clinic, Department of Surgery, WesternGeneral Hospital, Edinburgh) For manyyears it has been suggested that surgicaltreatment of benign peptic ulcerationresults in a shortened lifespan. This studyexamines this suggestion and its possiblecauses.The prospectively gathered records of

856 men aged 30-59 years who underwentcurative surgery (gastrectomy 86 %, drain-age±vagotomy 13 %; other 1 %) forbenign gastric or duodenal ulcerationduring the period 1947-65 were examined.Patients with subsequent recurrent pepticulceration were excluded from this group,as were patients dying less than one yearafter surgery. Nine per cent of the groupwere lost to follow-up. By 1980 46.2%had died.

Actuarial life table analysis demon-strated highly significant increases indeath rate between each age groupstudied (30-39, 40-49, and 50-59 years)and predicted death rates for a compar-able population (P < 0 001, < 0.001,< 0-001, respectively). The operatedgroup died on average 9.1 years prema-turely.

Eighty-four per cent of the study groupwere smokers and of all the deathsrecorded 675 % were from diseasesrelated to smoking. Death from causestraditionally related to peptic ulcersurgery (gastric carcinoma, pancreaticcarcinoma, TB, suicide, alcoholism)amounted to 11.9%. of the deaths (2.2%,4.7 %, 0-6 %, 2-5 %, 1.9% respectively).

This study confirms that patients whoundergo surgery for peptic ulcers arelikely to die prematurely. The contribu-tion to mortality of diseases related tosmoking or peptic ulcer surgery is dis-cussed and compared with figures for thegeneral population.

(F9)Results of resection and colo-anal anasto-mosis for carcinoma of the rectum

J P PERCY AND A G PARKS (St. Mark'sHospital, London) Restorative resection

is used increasingly for rectal carcinomabecause it obviates the need for a per-manent colostomy and is statistically ascurative as total excision. However,where an anastomosis is technically diffi-cult. alternatives to total excision aredescribed but have not gained widespreadacceptance.

Since 1973, 75 patients with rectalcarcinoma in a situation unsuitable foranterior resection have had sphincterfunction preserved by an overlappingsutured anastomosis performed withinthe anal canal from the perineal aspect.The results have been analysed.The pathological characteristics of

these tumours were comparable withthose of all rectal carcinomas treated atSt. Mark's Hospital between 1948 and1972. Two patients developed pelvicsepsis after colonic necrosis and anasto-motic breakdown. Eight developed pelvicsepsis without major anastomotic break-down. Four patients have developedrecurrent pelvic tumour. Twenty-one of32 patients are alive without sign ofrecurrence at three years and 12 of 19 arealive five years after a curative operation.Sixty-nine of the 70 patients whosecolostomies had been closed at the timeof assessment had normal or near normalbowel function.The operation is straightforward, has a

morbidity and mortality comparable withalternative procedures, and producesgood functional results.

THERAPYF10-F24

(FIO)Alginate: microanatomy of drug action?

M C BATESON, D HOPWOOD, W S HISLOP,AND I A D BOUCHIER (Departments ofMedicine and Pathology, Ninewells Hos-pital and Medical School, Dundee, Scot-land) Forty-seven patients with a nor-mal oesophagus were given 20 ml alginate-sodium bicarbonate mixture (Gaviscon)up to five hours before upper digestiveendoscopy. Biopsies were taken on entry30 cm from the incisors. Sections werecut, stained, and examined with theelectron microscope. In five patientsbiopsies were incubated in autologous

gastric juice before ultrastructuralexamination.

There was no macroscopic evidence ofalginate at endoscopy in these patients,but on electron microscopy dark-stainingglobules could be identified in the inter-cellular spaces of the deeper functionaland prickle cell layers. Globules wereseen in seven of 18 patients examined upto one hour after alginate treatment, butin only two patients examined thereafter.These appearances could be reproducedby incubation in alginate of oesophagealbiopsies from untreated patients. No suchglobules were found in more than 100other oesophageal biopsies from patientsnot receiving alginate treatment.

Alginate pretreatment partially pre-vented the severe damage expected afterincubation in autologous gastric juicepH 1 in each of three patients. In twoother patients with gastric juice pH 7 or8 no protection was observed.The protective action of alginate may

resaLlt from plugging of the oesophagealintercellular spaces and prevention of thecytotoxic effects of retrograde reflux.

(Fl 1)Elemental diet composition: effect on bileacid metabolism and hepatic lipids

LESLEY M NELSON, I WILLIAMSON, ANDR I RUSSELL (Gastroenterology Unit,Royal Infirmary, Glasgow) We havepreviously demonstrated that treatmentwith the low fat elemental diet Vivonexreduces faecal bile acid excretion andimproves cholerheic diarrhoea in man.

In this study the mechanism of theseeffects and the influence of elemental dietcomposition have been investigated byfeeding rats Vivonex (n= 6), Flexical(n=6), or control diet (n=6) for threemonths.Median (range) faecal bile acid excre-

tion was reduced from 1094 (855-1115)[Lmol/kg body weight/wk for control dietto 541 (407-695) for Flexical and 207(126-280) for Vivonex (P <0002 in allcases). Cholic acid half-life increasedfrom 2-8 (2.5-3.7) days to 4-2 (3.3-7.5)for Flexical and 10(5 (8-8-11-6) forVivonex (P<0003). Liver cholesterol was198 (135-220),umol/kg body weight forcontrol diet, 241 (211-358) for Flexical,and 725 (397-1054) for Vivonex (P<0.005).

Thus, reduced faceal bile acid excretionduring elemental diet feeding does notresult from a reduced bile acid pool but

A429

on Decem


http://gut.bmj.com

/G


ay 1981. Dow

nloaded from

http://gut.bmj.com/


reduced turnover of a normal sized pooland a rise in liver cholesterol reflectsreduction of necessary bile acid synthesis.

Liver histology showed marked fattychange in 5/6 Vivonex and slight changesin 516 Flexical. Total liver lipid wasincreased from 5 9 (3.6-7.3) mmol/kgbody weight to 8.3 (7-2-12.6) for Flexicaland 25.7 (14'1-29.8) for Vivonex (P<0 005).

This study indicates that, although thecomposition of Vivonex has a moremarked effect on bile acid turnover thanFlexical, the latter may be preferable ifthe development of fatty liver during long-term elemental diet therapy is to beavoided.

(F12)Effect of concentration on amino acidabsorption from a protein hydrolysate andan equivalent amino acid mixture

J E HEIGARTY, K J MORIARTY, P D FAIR-CLOUGH, M L CLARK, AND A M DAWSON(Department of Gastroenterology, St.Bartholomew's Hospital, London) Jeju-nal perfusion studies in man demonstrat-ing greater absorption of amino acidsfrom complex mixtures of peptidescompared with their equivalent aminoacid mixtures have suggested that pep-tides might confer nutritional advantages.However, the relative rates of absorptionof amino acids from single peptides andequivalent amino acid mixtures is critic-ally dependent on concentration. Thepresent studies were designed to examine,using a jejunal perfusion technique inman, the effect of concentration on therelative rates of absorption of aminoacids from a protein hydrolysate contain-ing a complex mixture of peptides andfrom an equivalent mixture of free aminoacids.

Studies were performed using solutionsof a hydrolysate of lactalbumin and anequivalent amino acid mixture containing40 and 100 mmol aNH2/l. The resultsshow that the five amino acids (isoleucine,leucine, methionine, proline, arginine),which were absorbed to a greater extent(p< 0-05) from the amino acid mixturethan for the hydrolysate at the low con-centration (40 mmol aNH2/l), were ab-sorbed to a similar extent from the highconcentration (100 mmol aNH2/l). Simi-larly, those amino acids (phenylalanine,tyrosine, serine, histidine, threonine,glutamine), which were absorbed to asimilar extent from the peptide and amino

acid mixtures at the low concentration(40 mmol ocNH2/l), were absorbed to agreater extent (P < 005) from the peptidemixture at the high concentration (100mmol aNH2/l).We conclude that the greater absorp-

tion of amino acids from protein hydroly-sates compared with their equivalentamino acid mixtures is a concentrationdependent phenomenon and that intes-tinal perfusion studies do not allow anystatement regarding the relative nutri-tional merits of different nitrogen sources.

(F1 3)Dietary nitrogen formulation: does itreally matter?

K J MORIARTY, J E HEGARTY, P D FAIR-CLOUGH, M L CLARK, AND A M DAWSON(Department of Gastroenterology, St.Bartholomew's Hospital, London) Avariety of enteral diets containing eitherwhole protein, protein hydrolysates, orfree amino acids as the nitrogen sourceare available for the treatment of protein-calorie malnutrition. However, the opti-mal composition of the nitrogen sourceused in these diets has not been estab-lished, although intestinal perfusionstudies have suggested that diets contain-ing small peptides as the nitrogen sourcemay confer nutritional advantages. Thepurpose of the present study was toevaluate the relative merits of differentnitrogen sources in the maintenance ofnitrogen balance in normal man.

Nitrogen conservation was induced byfeeding four healthy volunteers for 38days a low nitrogen intake (47 mgN/kg/24 h), which produced a state of negativenitrogen balance. All diets were isonitro-genous and isocaloric, energy (151 kJ/kg)being provided by a glucose polymer anda fat emulsion. After an 18 day adapta-tion period, there was equal conservationof nitrogen by the three forms of dietarynitrogen used, as shown by comparableurinary excretion of nitrogen over 10 daystudy periods in the subjects fed lactal-bumin whole protein, its hydrolysate,and an equivalent amino acid mixture.We conclude that whole protein is as

effective as amino acids and proteinhydrolysates in maintaining nitrogenbalance and that there is therefore nonutritional logic in the current practice ofprescribing liquid feeds containing expen-sive hydrolysates and amino acids tosubjects with normal gastrointestinalfunction.

(F14)Old ladies, drugs, and gastric ulceration

P COOKE AND M R THOMPSON (Depart-ment of Surgery, Bristol Royal Infirmary,Bristol) The controversy over the extentto which non-steroidal anti-inflammatorydrugs (NAD) cause peptic ulcerationcontinues. In view of this, the medicationtaken by 202 patients with gastric ulceradmitted to the Bristol Royal Infirmarybetween 1974-78 was reviewed. Thirty-seven patients (10/108 men=9%; 27/94women=29%) were on a non-aspirinNAD (NANAD). Thirty-four (91 %)were aged over 60 years and 25 (68%)were women over 60 years. Thus, 25 outof 70 women (36%) aged over 60 yearswere on NANADS (indomethacin, 14;phenylbutazone, seven; other NANAD,four). This was significantly greater(P<0.01, x2 test) than an age-matchedgroup of 220 women admitted to thesame hospitals over the same time periodwith either a colonic carcinoma or gall-stones (5/220 indomethacin; 3/220 phenyl-butazone; 2/220 other NANAD).

Seven gastric ulcer patients were takingNANADS for rheumatoid arthritis, sixfor osteoarthritis and three for non-specific arthritis. In nine the reason wasindefinite.

In comparison with the 35 women agedover 60 years not on NANADS the 25on NANADS were more likely to have agiant ulcer (11/25 vs 4/35; p< 0.05), needa blood transfusion (19/25 vs 15/35;(p< 0-05), be on a diuretic (12/25 vs 4/35;p< 0.01), less likely to have an operation(5125 vs 18/35; p<0.05), and had asignificantly shorter history of dyspepsia(p < 0'001).Thus the majority of NANAD gastric

ulcers occur in elderly women in whomit is a significant problem. These ulcershave several features which distinguishthem from idiopathic ulcers. Diureticsmay be important cofactors.

(FI 5)Pirenzepine in the treatment of duodenalulcer: a multicentre controlled trial versuscimetidine

M GUSLANDI, A TITTOBELLO, M GALEONE,G MOISE, E BIANCHINI, P BODINI, A VOLTA,G VERCELLINO, E BAFFELLI, AND M GIORGI-CONCIATO (introduced by P I Reed)(Third Medical Clinic, University ofMilan; Surgery Department, Hospital ofCastel Goffredo, 2nd Medical Department,

A430

on Decem


http://gut.bmj.com

/G


ay 1981. Dow

nloaded from

http://gut.bmj.com/


Hospital of Cremona, 2nd Medical De-partment, Hospital of Sesto S. Giovanni;Medical Department, Hospital of Breno;Clinical Department, Istituto De Angeli,Milano, Italy) Pirenzepine, a selectiveanti-muscarinic agent, was found to beeffective in promoting healing of pepticulcer. In a controlled clinical trial involv-ing five hospitals we have comparedpirenzepine and cimetidine in the treat-ment of duodenal ulcer.One hundred and two outpatients with

endoscopically proved duodenal ulcerswere randomly allocated to a six weektreatment with either pirenzepine (100 mgdaily) or cimetidine (1 g daily). The twogroups were comparable as regards sex,age, disease history, smoking, and alcoholand coffee consumption. Ninety-sixpatients completed the study, but only 84agreed to undergo endoscopical controlat six weeks. Endoscopy was performedby endoscopists unaware of the adminis-tered treatment and healing was definedas complete re-epithelisation of ulcerarea. Statistical analyses were carried outby an independent statistician.The healing rate was 71.5% with piren-

zepine and 59.5% with cimetidine. Thedifference is not statistically significant,showing that both drugs are equallyeffective. No differences in the two groupsas regards pain relief and antacid con-

sumption were observed.Side-effects were mild and infrequent

with both drugs. Only one patient in thepirenzepine group and one in the cimeti-dine group discontinued treatment be-cause of severe side-effects (dry mouthand diarrhoea respectively).

(F16)Effect of cimetidine on intraduodenal bileacid precipitation, pancreatin inactivation,and lipid solubilisation in pancreaticsteatorrhoea

P L ZENTLER-MUNRO, D R FINE, M GANNON,AND T C NORTHFIELD (Department ofMedicine, St. George's Hospital MedicalSchool, London) In pancreatic steator-rhoea cimetidine enhances the effect ofpancreatin on fat absorption. Our hypo-thesis is that this enhancement is due, notonly to prevention of pancreatin inactiva-tion by gastric acid, but also to preventionof pH-dependent bile acid precipitation.To test this, we have aspirated jejunalcontents after a Lundh meal in sevenpatients with pancreatic steatorrhoea onthree regimens in random order: (1)

pancreatin alone; (2) cimetidine alone;(3) pancreatin plus cimetidine. We havecompared these results with those on notreatment. On pancreatin alone, therewas an increase in mean lipase (15-1 vs37 U/i, P<0.01) and lipolysis (12.5 vs8.8%, P<0 05), but no increase inaqueous-phase lipid (6.4 vs 79 mM/I,NS) because bile acid precipitation re-mained unchanged (313 vs 27.5% NS).Fifty-nine per cent of the meal enteredthe jejunum below the critical pH of 5;on the cimetidine regimens all the mealwas at pH>6. On cimetidine alone, bileacid precipitation was reduced comparedwith pancreatin alone (178±3.8 % vs31*3±10*1 %, NS) and aqueous lipid wasincreased (9.0 vs 6-4 mM/i, P < 0.05)despite minimal lipase and lipolysis. Thecombined regimen was significantly betterthan either pancreatin or cimetidine alonefor lipase (40.5 U/l P< 0.01), lipolysis(19-7%, p< 0-05), and bile acid precipita-tion (4.3 %, P < 0.05). This led to a markedimprovement in aqueous phase lipid(18.8 mM/I, P<0.05). We conclude thatfor optimum therapy of pancreatic stea-torrhoea, bile acid precipitation, as wellas pancreatin inactivation, must beprevented.

(F17)Disposition in normal volunteers of sodiumazodisalicylate, a potential therapeuticagent in inflammatory bowel disease

C P WILLOUGHBY, J K ARONSON, H AGBACK,N 0 BODIN, E ANDERSSON, AND S C TRUELOVE(Gastroenterology Unit and the Depart-ment of Clinical Pharmacology, JohnRadcliffe Hospital, Oxford, and theAnalytical Research Department, Phar-macia AB, Uppsala, Sweden) Colonicbacteria split sulphasalazine to liberatesulphapyridine and 5-aminosalicylic acid(5-ASA). 5-ASA is the active constituent;many adverse effects of the compoundare attributable to its sulphapyridinecontent. An oral agent which releasedonly 5-ASA in the colon would, in theory,be a preferable drug.Sodium azodisalicylate (di-5-ASA) is

two 5-ASA molecules linked by a diazobond. We have studied its disposition ineight normal volunteers. Serum, urinary,and faecal concentrations of di-5-ASAand total 5-ASA were measured after asingle oral dose (0.5 g di-5-ASA) andduring long-term treatment (0-25 g or1.0 g daily).

During 72 hours after a single dose ofdi-5-ASA, less than 5% of the amountingested appeared intact in the urine.Nineteen to 41% (median 26.7%) of thedose was recovered in the urine, and7-36% (median 17.3%) in the faeces, as5-ASA and acetyl-5-ASA. Serum 5-ASAlevels were negligible.During long-term oral ingestion (0.25 g

daily) the serum half-time of di-5-ASAwas 74-10 days. After 1.0 g di-5-ASAhad been taken daily for eight days, the24 hour faecal recovery of total 5-ASAwas 19-70% (median 35.8 %) of the dailydose.The quantity of 5-ASA reaching the

colon after ingestion of di-5-ASA issimilar to that found after ingestion ofsulphasalazine.

(F18)Sulphasalazine retention enemas in ulcera-tive colitis: a double-blind trial

K R PALMER, J R GOEPEL, AND C D HOLDS-WORTH (Department of Gastroentero-logy, Royal Hallamshire Hospital, Shef-field) Oral sulphasalazine is of provenvalue in the treatment of ulcerativecolitis, but in up to 20% of cases side-effects preclude its use. Sulphasalazineretention enemas represent an alternativeapproach in these subjects, though theirefficacy has not been adequately demon-strated.

Thirty-four patients with active ulcera-tive colitis therefore completed a double-blind assessment of the efficacy of twoweeks of treatment by nightly retentionenemas containing 3 g sulphasalazine,compared with placebo. Clinical, sig-moidoscopic and rectal biopsy gradingand diary records were used to assesschange in disease activity and side-effects.Significant advantage of active drug wasshown by change in clinical and sig-moidoscopic scores (p< 0.05), stoolconsistency, and percentage bloody stool(p < 0.05) and by histological improvementin seven subjects, contrasting withdeterioration in four of the controls.Improvement occurred only in thosesubjects not already taking oral sulpha-salazine.

Overall assessment showed improve-ment in 70% of patients on active treat-ment, but only 11 % on placebo (p< 0.01).No significant side-effects were observedeven in patients (two) previously in-tolerant of oral sulphasalazine.

A431

on Decem


http://gut.bmj.com

/G


ay 1981. Dow

nloaded from

http://gut.bmj.com/


(F19)5 aminosalicylic acid, a weak inhibitor ofhuman colonic prostaglandin synthetase

C J HAWKEY (Nuffield Department ofClinical Medicine, John Radcliffe Hospital,Oxford) 5 aminosalicylic acid (5ASA) isprobably the active therapeutic moiety ofsulphasalazine. Although it is assumed toact through inhibition of prostaglandinsynthetase there is no direct evidence thatit does so in human colonic muocsa. Amethod allowing construction of dose-response curves using single rectal biop-sies was therefore developed, and used toinvestigate the effects of 5ASA, sulpha-pyridine, prednisolone, indomethacin, andflurbiprofen.

Rectal biopsies taken from untreatedcolitics in remission were homogenised inice cold Tris HC1 pH 7.4. Aliquots induplicate or triplicate were preincubatedon ice with serial dilutions of inhibitorbefore incubation for 30 minutes at37°C with arachidonic acid 6.1 x 10-5 M,adrenalin 3.0x 10-3 M, and reduced glu-tathione 1.3 x 10-3 M (final concentra-tions). The reaction was stopped byaddition of indomethacin (final concen-tration 10-3 M) in ice-cold ethanol. Afterextraction, the prostaglandin E2 synthe-sised was measured by radioimmunoassayMean IC50s -50% inhibitory concen-

trations (with 95% confidence limits)were: for 5ASA, 2.68 (194-3A42) x 10 -2 M(n=4); for indomethacin, 3.55 (0.59-6.51)x10-7M (n=4); for flurbiprofen,5-17 (2-89-7.45)x 10-7 M (n=3). Neithersulphapyridine nor prednisolone waseffective.

These results provide the first directdemonstration that 5ASA can inhibithuman colonic prostaglandin synthetase.However, its low potency makes it ques-tionable whether this is its therapeuticmode of action in ulcerative colitis.

(F20)Novel mechanism of action for chenodeoxy-cholic acid in gallstone dissolution

J GRAHAM, R BIRD, AND T C NORTHFIELD(Departments of Biochemistry and Medi-cine, St. George's Hospital MedicalSchool, London) Chenodeoxycholic acid,but not cholic acid, makes fastinggallbladder bile unsaturated in cholesteroland causes dissolution of cholesterolgallstones. It also reduces biliarycholesterol secretion and hepaticHMGCOA reductase levels. It is there-

fore widely believed that the underlyingmechanism of action is a reduction inhepatic cholesterol synthesis. Our hypo-thesis is that this is secondary to aneffect on bile canalicular membranebecause bile acids are known tohave marked membrane effects. We havetherefore isolated bile canalicular mem-brane from hamster by the method ofEvans. The hamster was chosen becauseit provides a good model for formationof cholesterol gallstones, which can beprevented by chenodeoxycholic acid butnot by cholic acid. Purity of bile canali-cular membrane vesicles was checkedenzymically and by electron microscopy.They were incubated (37°C, pH 7.4) forfive minutes with conjugated bile acids1-4 mM. Chenodeoxycholic acid solu-bilised less cholesterol than cholic acid(mean±SD: 18-0±1.3% vs 24.2±3-5%at 1mM, 38.5_--4-1% vs 504±4.9% at2mM, 48-8±5.9% vs 85.4±5.5% at4 mM) but more phospholipid (14.3±1.2% vs 7.2±0.9% at 1 mM, 62.0±5.8%vs 11.9±1.9% at 2mM, 77.5±1.2% vs31.0±2.0% at 4 mM).

Dehydrocholate, a non-micelle-formingbile acid, solubilised no lipid. We con-clude that these findings are consistentwith the micellar theory of biliary lipidsecretion; and that they provide a novelexplanation for the finding that chenode-oxycholic acid, but not cholic acid,reduces biliary cholesterol saturation.

(F21)Gallstone calcification caused by ursode-oxycholic acid (UDCA)

M C BATESON, D P MAUDGAL, D B TRASH,T C NORTHFIELD, AND I A D BOUCHIER(Department of Medicine, Ninewells Hos-pital, Dundee; Norman Tanner Institute ofGastroenterology, St. George's Hospital,London; Manor Hospital, Walsall) Gall-stone calcification has been observed infive patients after therapy with UDCA500-1000 mg daily for initially radio-lucent gallbladder stones. This occurredafter six to 12 months' therapy.

In three patients there was a con-current reduction in stone size of at least50%, but after calcification no furtherresponse in gallstone sizes has occurred.Biliary lipid analysis in four of thepatients showed consistent falls in choles-terol content (14.3 to 6-7, 9.3 to 5.3, 6-8to 3.1, and 6.1 to 3.2mol%), and litho-genic index (1-82 to 1.00, 1.90 to 1.06,0-76 to 0.68, and 0.78 to 0.53). There was

an increase in biliary UDCA from 0-3.8%to 323-858% of the biliary bile acidsafter treatment. UDCA was 90-99%glycine conjugated after therapy. Aftercorrection for the increased biliary UDCAcontent the lithogenic index did notchange.

Gallstone calcification was an un-common event on UDCA therapy (fiveof 54 cases) but was not seen in 122patients we have viewed after chenodeoxy-cholic acid therapy. This difference isunlikely to have arisen by chance (p<0.006, Fisher's exact test), and must beconsidered when comparing the relativetherapeutic influences of these bile acids.

(F22)Is recurrence inevitable on withdrawingtreatment after complete gallstone dissolu-tion with chenodeoxycholic (CDCA) andursodeoxycholic (UDCA) acids?

D C RUPPIN AND R H DOWLING (Gastro-enterology Unit, Department of Medicine,Guy's Hospital and Medical School,London) Recurrence rates of 10-30%have been reported after discontinuingtreatment but surprisingly little is knownabout the frequency, timing, and riskfactors of gallstone recurrence. Since 1972we have seen complete gallstone dissolu-tion on 61 occasions in 56 patientsfollowed with annual radiographs for upto five years after discontinuing CDCA(n=47) or UDCA (n=14).

Excluding 11 patients lost to follow-upand seven who have not yet had post-dissolution cholecystograms, to datethere have been 23 recurrences in 22patients (21 of 30 (70%) after CDCA andtwo of eight (25%) after UDCA), while16 patients have remained gallstone free.In five, recurrent stones were dissolved byfurther therapy only to recur again in one.Mean time from discontinuing treat-

ment until detection of recurrence was20±SD 14-9 months (range 3-63 months),20 of the 23 recurrences (87 %) appearingwithin two years. Time-related recurrencerates were nine of 46 (20%) at one year,11 of 31 (37%) at two, one of 12 (8%) atthree, and two of seven (29 %) at four tofive years.There were no significant differences

between 22 patients with gallstone recur-rence and the 16 who remained gallstonefree, in age (57±6 vs 53±13 years), bodyweight (64±10 vs 67±12kg), pretreat-ment stone size (6.8±4 vs 78±t5 mmdiam) or number (9410 vs 19 ±6), dose

A432

on Decem


http://gut.bmj.com

/G


ay 1981. Dow

nloaded from

http://gut.bmj.com/


(14.5±1-8 vs 15-3-LI-9mg CDCA/kg/day) and (10.0 vs 89- -2-00 mg UDCA/kg/day) or duration of treatment to dis-solution (16±8 vs 13 --5 months), norbetween pretreatment and on-treatmentSIs (1-45-_03 vs 1-46 -0 3 and 0.74±0-2vs 0-76 -0 2 respectively). However, thefemale:male ratio was significantly higher(p<0-05) in the recurrence group (22:1)than in the non-recurrence patients (11:5).With up to 70%0 gallstone recurrence,

post-dissolution treatment seems essential.

(F23)Diarrhoea and bile acid malabsorption inmild radiation enteropathy

M VAN BLANKENSTEIN (Department ofInternal Medicine II, Rotterdam Univer-sity Hospital, Dijkzigt) From a groupof women being investigated for diarrhoeaafter pelvic radiotherapy 15 were selectedwho had painless, urgent diarrhoea withincontinence. All had regained theiroriginal weight, they had no anaemia,vitamin deficiency, steatorrhoea, or signi-ficant ileal stricture on radiographicexamination of the small intestine.

Bile acid metabolism was studied bythe cholyl-14C-glycine breath test, 24 hourfaecal bile acid excretion, and the per-centage faecal excretion of a tracer doseof tauro-14C-cholic acid (TCA) in fourdays.The breath test was abnormal in 14 of

15 patients, the 24 hour faecal bile acidexcretion in 12 of 15 (mean 2-05 mmol/24 h, range 0.6-5.9, normal < 1-25), theTCA in all 15 (mean 78-3% range 62-91,normal < 56 %). Mean 24 hour faecalproduction was 200 g (range 92-297 g).

All patients reported alleviation ofsymptoms on cholestyramine 1-8 g daily.

Diarrhoea in mild radiation entero-pathy is associated with impaired ilealbile acid absorption. This may be respon-sible for urgency. However, the low stoolvolumes suggest abnormal continencemechanism as an additive factor. Patientsresponded well to low dose cholestyra-mine therapy.

(F24)D-penicillamine treatment improves sur-vival in primary biliary cirrhosis

0 EPSTEIN, S JAIN, R G LEE, D G COOK,A M BOSS, P J SCHEUER, AND S SHERLOCK(Academic Department of Medicine andDepartment of Pathology, Royal Free

Hospital, London) In primary biliarycirrhosis, immunological damage to intra-hepatic bile ducts causes chronic chole-stasis, liver fibrosis on liver copper reten-tion. The immunological, cupriuretic, andantifibrotic effects of d-penicillaminemake it potentially suitable for the treat-ment of primary biliary cirrhosis. Wehave evaluated d-penicillamine (Hom-burg) in a radomised trial. Fifty-fivepatients received penicillamine (600 mg/day), and 32 were given an identicalplacebo. The groups were matched forage, duration of follow-up, liver tests,immunoglobulins and liver copper con-centrations. There were similar propor-tions of symptomatic and asymptomaticpatients. Stage 4 histology was morecommon in the penicillamine group, andstage 3 in the controls. Eighteen patientson penicillamine developed major side-effects. Over a mean follow-up of 33months (range six to 71 months), four(7%) penicillamine treated patients, and10 (31 %) controls died (p<0-02). Alldeaths occurred in patients with stage 3or 4 histology on entry to the study. Theimproved survival is only apparent inpatients surviving 24 months in the study.Changes in liver tests, immunoglobulinsand liver copper concentrations in sur-viving patients are similar to those pre-viously reported. There was evidence ofhistological improvement in treatedpatients, but not controls.

CLINICAL DIAGNOSTIC/SURGICALF25-F39

(F25)Effect of cimetidine on the surgical man-agement of peptic ulcer disease

M R THOMPSON (Department of Surgery,Bristol Royal Infirmary, Bristol) Whencimetidine was introduced many clini-cians, with varying degrees of delight anddismay, felt that it heralded the rapiddisappearance of surgery for peptic ulcerdisease. It is now possible to measure inreal terms what effect cimetidine hasactually had on the number of operationsperformed for this disorder. This wasdone in the Bristol area for the three yearperiod before and after the introductionof cimetidine, using data from the SouthWestern Health Authority StatisticsBureau.

There has been a significant reductionin the number of elective operations forboth gastric and duodenal ulcers since1977 (61-38 duodenal ulcers, 13-10gastric ulcers; per six month period;P<0-01 Mann-Whitney U test). As therewas no similar reduction in the numberof cholecystectomies or admissions forperforated duodenal ulcer, it would seemthat this decrease is not due either to areduction in hospital facilities or to areduction in the incidence of the disease.

Further, there was a reduction in thenumber of emergency duodenal ulceroperations, excluding perforation (41-37per six month period; P>0-05) and anoverall reduction (16-12; P<0-05) in thenumber of emergency gastric ulceroperations.These data suggest that cimetidine has

reduced the referral rate for electiveoperations and is also possibly affectingthe natural history of the disease.

(F26)Does insulin-stimulated gastric secretionchange with time after vagotomy?

D J BUTTERFIELD, J V PARKIN, P F WHIT-FIELD, AND M HOBSLEY (Department ofSurgical Studies, The Middlesex Hospitaland Medical School, London) Insulin-stimulated gastric secretion in the post-vagotomy patient has been reported tobe stable with time, and to be an accuratepredictor of the liability to clear recur-rence when corrected for pyloric loss andduodenogastric reflux and standardisedfor height.We measured the half to two hour VG

in 47 patients who had had various formsof vagotomy, twice (40 patients) or threetimes (seven patients) at various timesafter the operation.The mean VG did not differ between

the recurrent and non-recurrent ulcergroup at the time of the first test, and didnot alter in the non-recurrent group atthe time of the second test. However, themean VG of the recurrent group at thetime of the second test had risen signi-ficantly over the VG of the same groupwhen first tested (P <00025) and wassignificantly higher than the VG of thenon-recurrent group when tested for thesecond time (p < 00005).

It would seem that within the first yearafter operation, the VG cannot differen-tiate between the recurrent ulcer and non-recurrence groups. After this, patientswho develop recurrent ulcer have a

A433

on Decem


http://gut.bmj.com

/G


ay 1981. Dow

nloaded from

http://gut.bmj.com/


significant rise in VG which does notoccur in patients without recurrence.

(F27)Cimetidine and duodenal ulcer: can surgerycure the failures?

R L BLACKETT AND D JOHNSTON (Univer-sity Department of Surgery, The GeneralInfirmary, Leeds) If cimetidine is con-sidered to be a 'medical' form of vago-tomy, then cimetidine failures might farebadly after vagotomy also.

Since 1977, 53 patients with duodenalulcer (M 38, F 15, mean age 40 years) whohad failed to respond to cimetidine infull dosage, have undergone highly selec-tive vagotomy (HSV). Fifty-three consecu-tive patients (M 39, F 14, mean age 41years) who were treated by HSV between1969 and 1971 served as controls. Meanlength of history was 10 years for bothgroups.

Preoperative BAO, 6.2±1 mM/h andPAOP G, 47±2 mM/h, in the cimetidinetreated patients were similar to controls,7±1 mM/h and 44±1 mM/h respec-tively. After operation, mean BAO was1 mM/h in both groups and PAOP G,20 mM/h in both groups. Mean PAO(minus BAO) was 0.7±0-2 mM/h incimetidine failures and 0-04±0-02 mM/hin the controls (p< 0.01).

After a follow-up of one to three years,84% of the cimetidine failures and 86%of the control patients had obtained goodresults (Visick grades I and II). In thecimetidine-treated patients there werethree proven recurrences and in thecontrols two suspected recurrences.Thus so far the results of surgery in

cimetidine failures have been as good asthe results obtained in the pre-cimetidineera.

(F28)Gastric juice enzymes-an aid in thediagnosis of gastric cancer

K ROGERS AND G T WILLLAMS (Depart-ments of Surgery and Pathology, WelshNational School of Medicine) Raisedlevels of lactic dehydrogenase and B-glucuronidase are found in gastric juicein disease states.We have studied fasting gastric juice in

113 patients in whom double contrastbarium examination and endoscopicbiopsy have been carried out and corre-lated the pathological diagnosis with the

values of lactic dehydrogenase and B-glucuronidase.The diagnostic subgroups studied com-

prise 28 patients with normal findings;42 with gastric adenocarcinomas; 14 withgastritis and intestinal metaplasia; 11with gastric ulcers; nine with duodenalulcers; and nine post-gastrectomy patients.The distribution of lactic dehydrogen-

ase and B-glucuronidase were bothpositively skewed; accordingly, subse-quent analysis was based on their cuberoots. In all patients there was goodcorrelation between lactic dehydrogenaseand B-glucu ronidase values, 0-412(p < 0-001).Both variables discriminated substan-

tially between the diagnostic subgroupsand, in particular, the presence or absenceof malignancy could be predicted with ahigh degree of accuracy (p < 0O001).

In this study both cases of early gastriccancer were correctly diagnosed by gastricjuice analysis.The test is easy to perform and readily

reproducible. In addition to being usefulin the investigation of symptomaticpatients it may be of value in the identi-fication and screening of patients atincreased risk of gastric malignancy.

(F29)Effect of biliopancreatic bypass surgeryfor morbid obesity on the hormonalresponse to food

D L SARSON, N SCOPINARO, D CIVALLERI,E GIANETTA, AND S R BLOOM (Depart-ment of Medicine, Royal PostgraduateMedical School, London, and II PatologiaChirurgica Universita, Ospedale S Mar-tino, Genoa, Italy) Biliopancreatic by-pass entails a roux-en-y reconstructionwith the small bowel transected at itsmid-point and an enteroenterostomyfashioned 50 cm proximal to the ileo-caecal valve. We have therefore investi-gated the gut hormone responses to anormal meal in 13 lean controls, 16 obesesubjects, 21 patients within six monthsand 17 patients six to 12 months aftersurgery. After biliopancreatic bypassglucose-dependent insulinotropic poly-peptide (GIP), pancreatic polypeptide(PP), and insulin release were muchreduced, while the distally producedenteroglucagon and neurotensin wereconsiderably raised. Gastrin responseswere little changed. The poor insulinresponse may be the result of the reduc-tion of GIP, although the mechanism of

the entero-PP axis is still uncertain. Arise in the neurotensin response may beresponsible for the reduction of uppergastrointestinal motility and enteroglu-cagon for the mucosal hypertrophy whichensues; both effects may be viewed as&normalising strategies'. If anything, thesedifferences become more marked withtime.

Biliopancreatic bypass results in pro-found weight reduction and the hormonalchanges observed after surgery relatelargely to their cellular distribution. Thesechanges provide unique clues to thephysiology of gastrointestinal weight andappetite control.

(F30)Non-invasive measurement of duodeno-gastric reflux and its association withflatulent dyspepsia in patients with gall-stones

I A EYRE-BROOK, A M HOLROYD, AND A GJOHNSON (University Departments of Sur-gery and Medical Physics, Royal Hallam-shire Hospital, Sheffield) A relationshipbetween flatulent dyspepsia and duodeno-gastric reflux has been demonstrated inpatients with cholelithiasis. In thosestudies, however, nasogastric intubationwas used, which may well affect pyloricfunction. We have studied post-prandialduodenogastric bile reflux in gallstonepatients using a non-invasive technique.

Severity of each of nine symptoms andtheir frequency were scored separately ona three point scale (0-1-2). The scoreswere summated to provide a numericalvalue for flatulent dyspepsia (range 0-36).

Duodenogastric bile reflux was moni-tored scintigraphically; the position ofthe stomach being first established withintravenous technetium 99 sodium per-technecate (100 ,uCi). Biliary secretionafter a Lundh meal was then studieddynamically (one minute frames) usingtechnetium p butyl iminodiacetic acid(3 mCi). Reflux was recorded as the per-centage of bile secreted by the liver whichappeared in the stomach area. Five percent reflux was considered significant.

Fifteen patients were studied beforecholecystectomy and eight were reinvesti-gated one month after operation. Threepatients with high preoperative symptomscores (>10), but none of four patientswith low scores (< 5) had significantreflux ( >5 %). Two of eight patients withintermediate scores (5-10) had significantreflux.

A434

on Decem


http://gut.bmj.com

/G


ay 1981. Dow

nloaded from

http://gut.bmj.com/


Postoperatively, two asymptomaticpatients (score=O) had no significantreflux (<5 %), while six symptomaticpatients (scores 2-11) had significantreflux (> 5 %).

This study confirms the associationbetween postprandial duodenogastric re-flux of bile and flatulent dyspepsia andestablishes that reflux is not a feature ofasymptomatic patients.

(F31)Diagnosis of acute pancreatitis-have we

improved on the plasma amylase?

M J MCMAHON AND JAHAN HODGSON(Department of Surgery, The GeneralInfirmary, Leeds) The occasional un-

reliability of plasma (or serum) amylaseestimation has led to enthusiasm fornumerous alternative methods of measur-ing amylase in order to detect acutepancreatitis. Measurement of urinaryamylase was developed into the amylase-creatinine clearance ratio, and advant-ages have been claimed for selectivemeasurement of pancreatic isoamylase andthe heat-labile fraction of the plasmaamylase. There has been no comparativestudy of these alternatives.Twenty patients with acute pancreatitis

were studied on the first, fifth, and 10thdays after admission. Plasma was takenfor total amylase (Phadebas) pancreaticisoamylase (separated using polyacryla-mide gel electrophoresis) and heat-labilefraction (60°C for 15 minutes). Simul-taneous urine samples were collected foramylase measurement and calculation ofamylase-creatinine clearance. Results wereconsidered abnormal if they were >2 SDabove the mean for normal subjects.

Amylase-creatinine clearance and theheat-labile fractions were of little value,being 35% and 36% respectively ofresults being abnormal on day 1. Totalamylase, pancreatic isoamylase, andurinary amylase were all grossly abnormalon day 1, 40-50% abnormal on day 5,and 30-47% abnormal on day 10.Neither pancreatic isoamylase nor urinaryamylase appeared to have any advantageover total plasma amylase, and the clinicalrelevance of these alternatives must beopen to doubt.

(F32)Familial incidence of colorectal cancer

J L DUNCAN AND J KYLE (Gastrointestinal

Research Unit, Royal Infirmary, Aberdeen)A prospective study of the familial inci-dence of cancer has been performed on50 consecutive patients treated for adeno-carcinoma of the colon and rectum. Alldiagnoses were confirmed histologicallyand the presence or absence of adjacentpolyps was noted. Each patient wasquestioned about all first-degree relatives,particularly in regard to intestinal cancer.Confirmation was obtained from patho-logy reports or close relatives. The con-trol group consisted of 50 patients,matched for age and sex, and who did nothave cancer.Nine out of the 50 patients had a

parent, sibling, or child with colorectalcancer (18 %). In five out of the ninefamilies the mother was the other affectedmember. In the control group only twopatients had an affected relative (4 %).The difference is highly significant(x2=8.13; 0-01 > p>0.001).The familial incidence was comparable

with the 16% incidence noted inNebraska, but the 'cancer families' withmore than two affected members werenot encountered in the present series.There was no difference in the frequencyof polyps in the group of patients whohad affected relatives (22%) comparedwith the group whose close relatives todate are free from colorectal cancer(24%). The implications for screeningprogrammes are discussed.

(F33)A practical solution to the diagnosis andtreatment of colorectal cancer?

P A FARRANDS, R L GRIFFITHS, AND D CBRITTON (Royal United Hospital, Bath)If early colonic tumours bleed, faecaloccult blood estimation of the asympto-matic populace may prevent 20 000deaths annually, and eliminate the mor-bidity of colonic cancer surgery.A well-publicised prospective trial was

undertaken in the single-practice markettown of Frome, Somerset. Each of the9050 townspeople aged over 40 yearsreceived an envelope containing a per-sonal letter of explanation, three Haemoc-cult test slides, and an instruction sheet.Completed slides were returned by 1817

people, 59 with positive tests were ex-amined and endoscoped in a specialclinic. Barium enema was arranged whenindicated and haemorrhoids treated withsclerotherapy. One patient had colitis,

six had severe diverticulosis, and 25 hadperianal conditions which can bleed.Twenty-one had no clinical or radio-logical abnormality and are still underreview. Six had colorectal neoplasms,four benign and two malignant. Threemore patients in Frome presented withcolonic tumours during the study. Twohad refused to comply with the protocolbecause of cancerphobia; the third hadbeen negative in the trial.Mass occult blood screening does

detect asymptomatic colonic neoplasms,but a sympathetic national publicitycampaign is required to increase thecompliance if this method is to achieveits optimum potential.

(F34)Ultrasound monitoring of liver metastasesfrom colorectal cancers

R H TAYLOR, J M GILBERT, MAGGIE G EVANS,HEATHER G LANE, AND P G CASSELL (De-partment of Gastroenterology, CentralMiddlesex Hospital, London; UltrasoundDepartment, University College Hospital,London; Imperial Cancer Research FundLaboratories, London, and Wexham ParkHospital, Slough) Patients with livermetastases from colorectal cancer oftenreceive chemotherapy. It is valuable inselecting patients and supervising theirclinical condition to monitor the develop-ment and size of metastases. We reporta prospective study using static grey-scale ultrasound scanning to detect andmeasure liver metastases.

Fifty patients with resected colorectalcancers had ultrasound scans as part oftheir routine follow-up at three monthlyintervals for between three months andfour years (median 15 months). Scanswere performed by one operator using astandard method on the same machine(Diasonograph NE4102, 3.5 MHz trans-ducer). Half were receiving chemo-therapy.

Scans on nine patients were abnormal.Four had metastases shown on their firstscan and five developed metastases duringfollow-up. Ultrasound detected thesethree to six months before they weredetectable clinically. Serial ultrasonicmeasurements of tumour size increasedduring follow-up. The remaining scans on41 patients were normal and these patientscontinue to be well clinically. Tumourswere graded at surgery and pathologicallyon a modified Dukes's classification.None of the four Dukes's A patients, 1/1S

A435

on Decem


http://gut.bmj.com

/G


ay 1981. Dow

nloaded from

http://gut.bmj.com/


Dukes's B, 4/22 Dukes's C, and 4/'6Dukes's D have developed hepaticmetastases.

Ultrasound appears to be a simple,non-invasive method of early detectionand monitoring of hepatic metastasesfrom colorectal tumours.

(F35)Portal vein liver scanning for early detec-tion of liver metastases in colorectal cancer

B MOONEY, M CRITCHLEY, S GRIME, AND ITAYLOR' (Departments of Slurgery and Nui-clear Medicine, Royal Liverpool Hospital,Liverpool) The methods currently avail-able for the preoperative detection ofcolorectal liver metastases are limited intheir ability to detect small lesions. Thishas both prognostic and therapeuticimplications.

Access to the portal circulation via theobliterated umbilical vein is availableand is used for the cytotoxic treatment ofestablished liver metastases as well asdelivery of adjuvant therapy.

In this study the cannulated umbilicalvein has been used to perform (1) portalvenography; (2) liver blood flow measure-ments using Xenon133 clearance; and (3)liver perfusion studies using TC99mmicrospheres (average 10 ,u diameter). Aportal venogram and hepatogram phaseare obtained. Liver blood flow andperfusion studies were performed using agamma camera with online computer anddata storage facilities. This allowedqualitative and quantitative measure-ments of blood flow to different areas ofthe normal and metastatic liver.The results of these studies have been

compared with preoperative Tc99m sul-phur colloid and ultrasound scans of theliver as well as liver function tests.The initial results of these studies

suggest that as many as 30%0 of patientswith macroscopically normal liver atinitial surgery for colorectal cancer haveevidence of metastatic disease.

These techniques may allow prognosisto be refined and treatment to be plannedon a more rational basis.

(F36)Infrared coagulation in the treatment ofhaemorrhoids

R J LEICESTER, R J NICHOLLS, AND C V

MANN (St. Mark's Hospital, London,The London Hospital, Whitechapel, and

the Royal Naval Hospital, Haslar, Gos-port, Hants) Many methods have beendescribed for the outpatient treatment ofhaemorrhoids to reduce the number ofpatients requiring admission. These maybe painful and complications such asnecrosis and incontinence have beenreported. A new technique using infraredcoagulation was described in 1977.A prospective trial comparing infrared

coagulation with conventional treatment(injection or rubber band ligation) hasbeen conducted. One hundred patientswith non-prolapsing and 100 with pro-lapsing haemorrhoids were allocatedrandomly to infrared coagulation orconventional treatment groups.

Age, sex, length of history and previoustreatment were similar in each group. Atsix weeks, sympatomatic improvementwas achieved in 90% of all patients withno difference between the groups. Atthree months 80% of patients with non-prolapsing haemorrhoids were asympto-matic after infrared therapy, comparedwith 50% after injection (x2 =4-84, DF= 1,P= <005). There was no significantdifference between infrared coagulationand conventional treatment in the pro-lapsing haemorrhoids. Seventy per cent ofpatients complained of pain after conven-tional treatment compared with 85 0%after infrared coagulation (x2= 35-03,DE=1, P= <0 001). No complicationoccurred in either group.

Infrared appears to be superior toinjection in non-prolapsing haemorrhoidsand as effective as rubber band ligationin most patients with prolapsing haemor-rhoids. It is a simple technique and is lesspainful than conventional methods.

(F37)Management of fistulas in Crohn's disease

J C GIVEL, P C HAWKER, M R B KEIGHLEY,R N ALLAN, AND J ALEXANDER-WILLIAMS(Gastroenterology Unit, The GeneralHospital, Birmingham) Our manage-ment of 68 patients with Crohn's diseasecomplicated by fistula, excluding perianalfistula, treated between 1970 and 1980 isreviewed.

Although many were complex, theprincipal fistula connection was cutaneousin 29. Four developed soon after opera-tion as the result of an anastomotic leak.Twelve occurred later from the site of anoperation, six after appendicectomy, sixafter resection. Thirteen arose spon-taneously from recrudescent disease, of

which three were associated with extra-luminal abscess. Only the early uncom-plicated postoperative fistulas closedspontaneously. All others required sur-gical resection before permanent closure.Of the 12 enterovaginal fistulas none

healed with conservative management,although some have minimal symptoms.Five have undergone proctectomy andone a successful operative closure. Onepatient developed an adenocarcinoma inthe fistula tract.One early postoperative enterovesical

fistula closed spontaneously; three othersrequired bowel resection.The 23 enteroenteric fistulas did not

cause specific symptoms related to thefistula but 21 of them have needed sur-gical treatment for indications such asstenosis. Two remain symptom free.

Except for uncomplicated early post-operative fistulas, conservative treatmenthas no place in the management ofCrohn's fistulas. A chronic symptomaticfistula requires excision of the diseasedbowel from which it originates. Short-term total parenteral nutrition, if needed,is only to make the patient fit foroperation.

(F38)Long-term effects of proctocolectomy ongenitourinary function in patients withinflammatory bowel disease

D E NEAL, A J PARKER, N S WILLIAMS, ANDD JOHNSTON (University Department OfSurgery, The General Infirmary, Leeds)Bladder dysfunction and sexual problemsafter abdominoperineal excision of therectum for carcinoma are not uncommon.Sexual dysfunction and bladder dysfunc-tion after proctocolectomy for inflamma-tory bowel disease are also not uncom-mon, despite the opinion of many sur-geons. This study was carried out todetermine the frequency and type ofbladder dysfunction in the long-term afterproctocolectomy for inflammatory boweldisease. Thirty-five patients (18 F, 17 M:age=47±14 years) were studied a meanof eight years after proctocolectomy.With the use of voiding pressure cysto-metry, they were compared with a seriesof controls (18 M, 16 F: age=48±17years) who had undergone bowel resec-tion for carcinoma of inflammatorybowel disease with no pelvic dissection,a mean of five years previously.Abnormal symptoms (impotence, ejacu-

latory failure, or dyspareunia, nocturia

A436

on Decem


http://gut.bmj.com

/G


ay 1981. Dow

nloaded from

http://gut.bmj.com/


> 2, incontinence and a poor stream)were more common after proctocolec-tomy (x2 18, on 4 DOF; P< 0-001).There was no difference in mean peakflow, maximum detrusor pressure, orcompliance between the two groups.Seven proctocolectomy and two controlpatients had abnormal cystometry; in theproctocolectomy group (1 F, 6 M) threehad bladder neck obstruction and fourhad denervated bladder; two controlpatients had bladder neck problems.Thus both symptoms and objective

signs of genitourinary dysfunction werefound to be common in the long-termafter proctocolectomy for inflammatorybowel disease.

(F39)Prolonged antibiotic cover is required toprevent sepsis after operations for inflam-matory bowel disease

M M HARES, F GRECA, E NEVAH, R N ALLANAND M R B KEIGHLEY (Departmentof Surgery, The General Hospital, Birm-ingham) Operations for inflammatorybowel disease are frequently complicatedby sepsis, probably because many patientsare receiving steroids, and there is a highincidence of pre-existing sepsis particu-larly in Crohn's disease.We have undertaken a trial of anti-

biotic cover in 74 patients undergoingelective operation for inflammatorybowel disease. Excluded were patientsrequiring emergency operations, thosewith obvious intra-abdominal sepsis orenterocutaneous fistulae. Patients wereallocated to one of three groups: threedoses of metronidazole and gentamicin(group A: n=21), identical placebosolutions (group B: n=27), or five daycover with metronidazole and gentamicin(group C: n=26). The overall incidenceof sepsis was 42% in group A, 44% ingroup B, and 120% in group C. Henceonly the group receiving five day anti-biotic cover received clinical benefit.Failure of antimicrobial prophylaxis didnot relate to steroid therapy but wasassociated with established infection atthe time of operation. With five day anti-biotic cover the incidence of abdominalwound sepsis was 8 %, abscess 8 %, andsepticaemia did not occur.These results indicate that (1) steroids

do not increase the risk of sepsis in in-flammatory bowel disease; (2) patients

who develop sepsis after surgery for thisdisorder usually have pre-existing infec-tion; and (3) that prolonged antimicrobialcover is required to prevent postoperativecomplications.

PHYSIOLOGY II

F40-F54

(F40)Prostacyclin-like activity in experimentalportal hypertension in dogs

J J CORZO, J M ZOZAYA, F GUARNER, AND

J PRIETO (introduced by R E Pounder)(Department of Medicine, UniversitaryClinic, University of Navarra, Pamplona,Spain) A possible relationship betweenportal hypertension and platelet aggrega-tion was investigated in dogs. Animalswere anaesthetised with pentobarbitone.A catheter was placed in the mesentericvein for pressure determinations andblood collections. In five (group A),pressure was increased to twice the basallevel during 30 minutes by partial ligationof the portal vein. In four (group B),300 mg/kg of aspirin were administeredintravenously 30 minutes before theincrease of pressure. Platelet aggregationby thrombin was measured in platelet-rich plasma obtained from basal and30 minute post-portal hypertensionsamples in group A and from basal,30 minute post-aspirin and 30 minutepost-portal hypertension in group B.

In group A, a significant decrease ofthe platelet aggregation was observedwhen basal and 30 minute post-portalhypertension were compared (P < 0 0005).In group B there were no significantdifferences in platelet aggregation amongbasal, 30 minute post-aspirin and post-portal hypertension samples. It is con-cluded that (1) portal hypertension in-duces a decrease in mesenteric venousblood platelet aggregation; (2) thisaction is prevented by a large dose ofaspirin; (3) as this dose of aspirin blocksboth the platelets and vessels cyclo-oxygenase, it is reasonable to think thatthe hypoaggregation observed after por-tal hypertension is mediated by PGI2liberation by portal vascular bed.

(F41)Effect of catechin (cyanidanol-3) onalcoholic fatty liver

P R RYLE, J CHAKRABORTY, G K SHAW,AND A D THOMSON (Department ofGastroenterology/Liver Unit, GreenwichDistrict Hospital, Vanbrugh Hill, London,and Department of Biochemistry, Univer-sity of Surrey, Guildford, Surrey) Thebioflavanoid, catechin, acting as a freeradical scavenger, protects the liveragainst carbon tetrachloride toxicity. It isalso reported to inhibit collagen synthesisand to reverse the raised NADH1:NADratio that arises from ethanol oxidation.We have studied the effect of catechin onthe development of fatty liver.Two groups of 10 Sprague-Dawley rats

(males, 200-260 g) were fed a nutrition-ally complete liquid diet containing 28%0of the total calories as ethanol, replacingan isocaloric amount of carbohydrate,and compared with control groups.Concomitant administration of catechin(intragastrically, 200 mg/kg, once daily)significantly reduced the raised hepaticlevels of total lipid (by 29% comparedwith controls), triglycerides (by 32 %),free fatty acids (by 53 %), and microsomalprotein (by 51 %) observed after ingestionof ethanol for three weeks. No toxiceffects attributable to the drug wereobserved.These results suggest that catechin

increases utilisation and lowers synthesisof fatty acids, rather than facilitating lipidremoval from the liver. The detailed bio-chemistry will be presented, but catechinis clearly effective at preventing fattyliver. These studies confirm the drug'simportance as a powerful hepatoprotec-tive agent which may be valuable in thetreatment of all stages of alcoholic liverdisease.

(F42)Motilin-induced gallbladder contraction: anew mechanism

T E ADRIAN, P MITCHENERE, G R SAGOR,N D CHRISTOFIDES, AND S R BLOOM (De-partment of Medicine, Royal PostgraduateMedical School, London) Gallbladderfunction has not been fully investigatedfor the influence of the newer gut hor-mones-for example, motilin, which con-tracts smooth muscle and is released by afatty meal.

In the conscious healthy pig we havecontinuously monitored intraluminal gall-

A437

on Decem


http://gut.bmj.com

/G


ay 1981. Dow

nloaded from

http://gut.bmj.com/


bladder pressure using a fundally attachedmicroradio pressure capsule. Pressureresponses to motilin, pancreatic polypep-tide (PP), vasoactive intestinal peptide(VIP), and somatostatin were comparedwith changes induced by feeding andcholecystokinin octapeptide.

After a meal pressure rose by 108+1.6 mmHg at 30 minutes with a peak at65 minutes (increment 14.8±2.3 mHg,meaniSEM, P<0 005, n=5). CCK-OP(doses 0.4, 1-5, 6, and 24 pmol/kg/h)caused a dose-related increase in pressureof 2.5±05, 6.2±1'4, 11.9±3.0, and16-5±2-3 mmHg respectively. Motilin(doses 1.8, 7, 24, and 96 pmol/kg/h) alsosignificantly increased gallbladder pres-sure by 2.7±1.3, 4-0+P14, 7.5±1O0, and12.0±2.0 mmHg respectively. Somato-statin, VIP and PP all caused a dose-related fall in gallbladder pressure-forexample, PP (doses 6, 25, 100, and 400pmol/kg/h) reduced pressure by 2.2+0-9,8.2±0.4, 11.6±7, and 14.8±0.8 mmHgrespectively.Thus CCK caused the expected rise and

PP the expected fall in gallbladderpressure. Low doses of motilin produceda rise in gallbladder pressure as large asCCK. This is of great interest, as motilincirculates in higher concentrations thanCCK and rises greatly after fat ingestion.

(F43)New method for the direct recording ofintestinal contractions

JOHN CUMMING, J M KELLY, AND C LSMITH (Department of Surgery, QueenAlexandra Hospital, Portsmouth, Hamp-shire) Previous methods for recordingmotor activity of the colon, have reliedon intraluminal pressure recordings usingballoons of various sizes and perfusedopen-ended tubes. However, intraluminalpressure is the result of a number offactors, among them being the contrac-tion of the intestinal smooth muscle. It isthese contractions in which we areprimarily interested.We describe a new method for record-

ing intestinal contractions directly, bytheir action on the strain gauge pairs ofan intraluminal strain gauge probe.A comparison of pressure (open-ended

tube) and contraction (strain gaugeprobe) recordings shows there is a highcorrelation between the two methods(r=0-8; P < 0-01) when comparing motilityindices before and after stimulation withneostigmine (0-6 mg subcutaneously).

Three contractile patterns have beenidentified:A contractions have a low amplitude

and a frequency of 10-17 cycles perminute (cpm) and are not associated withpressure wave changes.B contractions are A contractions on

a raised baseline and occur in conjunctionwith type III pressure waves which theyresemble closely in shape and character.C contractions have a high amplitude

and low frequency (2.2-42 cpm), andhave been associated with pressure wavechanges with similar shape and character.C contractions represent segmentation.Three patients with the irritable bowelsyndrome who were symptomatic at thetime of the recording, all had C con-tractions.The strain gauges respond to changes

in radius of lumen and measure smoothmuscle contraction directly, unlike pres-sure recordings which respond to smoothmuscle contractions indirectly.

Changes in smooth muscle tone may bedifferentiated from distension or deflationby combining intraluminal pressure mea-surement with the strain gauge probe atthe same level.

(F44)Stress and jejunal motor activity

D L WINGATE, SHEENA MCRAE, KIRSTENYOUNGER, AND D G THOMPSON (LondonHospital Medical College, London) Whileit is often assumed that stress influencesintestinal motility, previous studies ofthis effect have usually been brief andhave not allowed for the normal periodi-city of fasting motor activity. In thisstudy, 11 healthy fasted volunteers weresubjected to controlled mental stress forfour hours using a modified dichotomouslistening test; jejunal motor activity wasmonitored using a radiotelemetry capsuletethered at the duodenojejunal flexure.Each subject was also studied for a fourhour unstressed control period on thesame day. Pulse rate and blood pressurewere measured at regular intervals, and astressful response was arbitrarily definedas an increase greater than 5 % in meanaggregate pulse and blood pressurebetween control and stress periods.Thus defined, 7/11 subjects showed a

stressful response, and they also showeda significant (p=0-016) reduction in thenumber of activity fronts in the first twohours of stress compared with the controlperiod, but not thereafter. This method

of inducing controlled stress may proveuseful in further studies of gastrointes-tinal function, but the results emphasisethat individual variation in the responseto stress is an important variable. Fromour data, we conclude that stress maydelay motor complexes, but that adapta-tion occurs as stress is sustained.

(F45)Effect of porcine gastric fundic factor(GFF), bombesin, neurotensin, glucagonand VIP on jejunal glucose absorption

N J ANDREWS, S RINNO-BARMADA, KBURDETT, AND J B ELDER (UniversityDepartments of Surgery and MedicalBiochemistry, Manchester) Recently apeptide fraction isolated from gastricfundic mucosa was shown to decreasejejunal glucose absorption. Comparisonof gastric fundic factor (GFF) withcandidate fundic hormones has beenmade. 10 mM glucose in 0 15 M NaCIcontaining 0-1 % PEG was continuouslyperfused for up to one hour through a10 cm jejunal segment in a rat bioassay(n=80). Bolus caval injections of albuminin NaCl as control (n=17), GFF (n=8),bombesin, neurotensin, and VIP (1, 0-1,0.01 ±g/rat, n=12 for each peptide), andglucagon (10, 1, 0-1 gg/rat, n=12) weregiven. Samples of perfusate were taken at15 minute intervals for measurement ofglucose and PEG concentrations. Absorp-tion of glucose was calculated.At 15 minutes bombesin (1 pg/rat) and

neurotensin (0-1 ,ug/rat) gave increases of28% and 46% respectively in glucoseabsorption, whereas GFF decreased ab-sorption by 25 %. No significant effect onglucose absorption was demonstratedwith any of the doses of VIP or glucagonor with the other doses of bombesin orneurotensin throughout the experiment.Of the known peptides present in thegastric fundic mucosa, the gastric fundicextract used contained negligible amountsof somatostatin and substance P. Thisobservation, together with the presentdata using other candidate fundic hor-mones, strengthens the evidence that anovel fundic peptide may be involved inthe control of jejunal glucose absorption.

(F46)Simultaneous recording of myoelectricalactivity and motility from the human colon

I MORRIS, C DARBY, P HAMMOND, ANDI TAYLOR (Departments of Surgery and

A438

on Decem


http://gut.bmj.com

/G


ay 1981. Dow

nloaded from

http://gut.bmj.com/


Bioengineering, University of Liverpool,Liverpool) A new method of studyinghuman colonic motility has been assessed.Variations in electrical resistance at anintramuscular electrode correspond tolocal movement of the muscle. Thus,simultaneous recordings of myoelectricalactivity and motility can be obtained froma single intraluminal suction electrode.These were compared with motilitymeasured by intraluminal pressurechanges.

In 15 normal subjects at rest, motilitywas detected by resistance changes for agreater proportion of the time (mean748 ±6.7 %) compared with percentagemotility measured by intraluminal pres-sure (mean 32.9±7.4%). This increasedobservation of motility was also seen in15 patients with diverticular disease(resistance motility-mean 88-5±4.6 %,pressure motility 170O±4.2%) and in fivepatients with the irritable colon syndrome(resistance motility 92-6±3.8%, pressuremotility 28.2±16.5 %).

In five normal subjects and in fivepatients with the irritable colon syndrome,intravenous glucagon (0025 mg/min in-travenously) resulted in complete inhibi-tion of motility. This is additional evi-dence that the resistance method accur-ately detects colonic muscle motility andnot other movement.The method appears to be a sensitive

technique for studying motility and, bysimultaneously recording myoelectricalactivity, should prove a valuable tech-nique for examining the relationshipbetween the two.

(F47)Enteroglucagon and intestinal adaptation

G R SAGOR, M Y T AL-MUKHTUR, M AGHATEI, N A WRIGHT, AND S R BLOOM(Royal Postgraduate Medical School,London) Luminal nutrition and intes-tinal secretions are known to be importantin intestinal adaptation; however,humoral agents may also play a part. Toinvestigate the relationship further, wemeasured plasma enteroglucagon in twoexperimental rat models, and relatedthese findings to the most reliable indica-tor of proliferative status, the crypt cellproduction rate (CCPR). One group hadeither subtotal colectomy or ileal transec-tion. A second group had 75 % proximalsmall bowel fashioned into an isolated

Thiry-Vella fistula. Half of this groupwere allowed food ad libitum, while therest were fed intravenously.

In the colectomy group, ileal CCPR/hincreased from 18 ±09 after ileal transec-tion to 31 2 ±1.1 after subtotal colectomy(P< 0001). Similarly, enteroglucagon in-creased from 108.8±32 pmol/l in ilealtransection, to 234 5±20 8 pmol/l aftercolectomy (P<0-005). In the Thiry-Vellagroup, in the ileum in continuity, CCPR/hwas greater in orally fed rats (52±8),compared with intravenously fed animals(18±5) (p<0-01). In the excluded fistula,CCPR/h was reduced to 23.8±2 in orallyfed animals but this was greater than inthe intravenously fed group (16±1.5)(P < 0.01). Plasma enteroglucagon wasgreater with oral (566±59 pmol/l) thanwith intravenous feeding (250±41.9pml/l) (P < 0-01).These studies indicate that luminal

nutrition and intestinal secretions cannotexplain all the changes in crypt cellproduction site and suggest that a humoralagent plays a part. Enteroglucagon is stillthe most favoured candidate.

(F48)Search for an appropriate measurement ofproliferative status in experimental workin intestinal adaptation

N A WRIGHT, M AL-MUKHTAR, G SAGOR,AND S BLOOM (Departments of Histo-pathology and Medicine, Royal Post-graduate Medical School, HammersmithHospital, London) Workers in the fieldof intestinal adaptation need a precisemeasurement of proliferative status, whichis economical in terms of time and effort;serious doubt has been cast on the use oftritiated thymidine incorporation intoDNA as an estimator of proliferativerate and consequently we have sought formore appropriate measurements. Thethree factors which control crypt cellproliferation are the cell cycle time, cryptgrowth fraction, and crypt populationsize; an ideal measurement would besensitive to all three parameters.Such a measurement is the crypt cell

production rate; we have combined acrypt microdissection method with vin-cristine metaphase arrest to provide anestimate of the crypt cell production ratewhich gives acceptable interval estimatesand excellent statistical discrimination inexperimental work; the method is econo-

mical in terms of animals and timeexpenditure, and, because the experi-mental period is short, is suitable formonitoring ephemeral proliferative events.The method has been compared with

the more elaborate and time-consumingautoradiographic methods; in both con-trol and experimental situations there wasexcellent agreement.We therefore propose that the crypt

cell production rate is the most appropri-ate proliferative estimator in work onintestinal adaptation.

(F49)Faecal excretion of intravenous 3H-25-hydroxyvitamin D3 in normal subjects andin patients with intestinal malabsorption

J E COMPSTON, A L MERRETT, J E LEDGER,AND B CREAMER (Gastrointestinal Re-search Unit, Rayne Institute, St. Thomas'Hospital, London) Interruption of anenterohepatic circulation of 25-hydroxy-vitamin D (250HD) has been postulatedas a mechanism for the development ofvitamin D deficiency in intestinal malab-sorption. We have determined faecalexcretion of intravenously administered3H-250HD3 in three normal subjects andsix patients with malabsorption (threejejunoileal bypass, three small intestinalresection).

10 ,uCi of 3H-250HD3 in 20 ml of 10%ethanol was injected intravenously andstools were collected for five to six days.Faecal radioactivity was determined inhomogenates by liquid scintillation count-ing after oxidative combustion and silicicacid chromatography was performed onfaecal extracts.The mean daily excretion of radio-

activity in the normal subjects was 0-8-1. 6% of the injected dose. In four patientswith malabsorption faecal radioactivitywas increased; in two of these, the meandaily excretion was over 6% of theadministered dose. There was a significantcorrelation between faecal fat excretionand faecal radioactivity excretion (r=+0-89; P<0 005). Faecal chromato-graphy revealed unchanged 3H-250HD3in patients with malabsorption and innormal subjects.

These results provide support for theexistence of an enterohepatic circulationof 250HD in man and demonstrate thatfaecal loss of circulating 250HD may beincreased in intestinal disease.

A439

on Decem


http://gut.bmj.com

/G


ay 1981. Dow

nloaded from

http://gut.bmj.com/


(F50)Prolonged organ culture of human colonicmucosa using a supplemented medium:observations on the degeneration andregeneration of epithelial crypts

P V SENIOR, C J PRITCHETT, J P SUNTER,D R APPLETON, AND A J WATSON (Depart-ments of Pathology, Surgery, and MedicalStatistics, University of Newcastle uponTyne) Using a system which provedsuccessful in preserving adult mouse

colonic mucosa we have cultured adulthuman colonic mucosa and report thesequential morphological changes ob-served.During the initial 48 hours of culture

the normal microarchitecture of thetissue is well preserved although there issome progressive depletion of cytoplas-mic mucus from crypt epithelial cells.Subsequently accelerated degenerativechanges are seen and cells are lost fromthe upper third of the crypt. By between72 and 96 hours the upper part of thecrypts are denuded, leaving small acini or

solid clusters of epithelial cells in thelamina propria discontinuous with theintact surface epithelium.

These remnants then display intenseproliferative activity until by 186 hoursthe crypts are repopulated with a mono-layer of columnar epithelial cells, againcontinuous with the surface, possiblyutilising the fibrocellular sheath of thecrypt to guide their migration. From 264hours goblet cells are again seen amongthe crypt epithelial cells.These findings suggest that the primary

control of crypt cell production, migra-tion, and differentiation is inherent in thetissue. This will be a useful in vitrosystem for studying the response ofmucosa to various insults.

(F51)VIP nerves in endoscopic rectal biopsiesfrom patients with inflammatory boweldisease: a double-blind study

C O'MORAIN, ANNE E BISHOP, JULIA M

POLAK, S R BLOOM, A J LEVI, AND T JPETERS (Divisions ofClinical Cell Biologyand Clinical Sciences, Clinical ResearchCentre, Harrow, Middlesex, and Depart-ments of Histochemistry and Medicine,Hammersmith Hospital, London) Pre-vious reports, using surgically resectedspecimens, have suggested that there is aselective increase in the mural nervous

elements containing vasoactive intestinalpolypeptide (VIP), as determined byimmunocytochemistry, and in VIP con-tent, as estimated by radioimmunoassay,in patients with Crohn's disease.

In the present study, endoscopic recialbiopsies from control subjects (n 13),patients with ulcerative colitis (n= 13),and Crohn's disease, affecting (n=9) ornot involving (n =5) the rectum, havebeen examined for possible changes inthe levels of VIP nerves. The assessmentwas carried out in a double-blind fashion.

In patients with Crohn's disease affect-ing the rectum all biopsies had increasedlevels of VIP innervation, with six showingvery marked changes including distortionof individual fibres. In patients withoutrectal involvement some minor increaseswere found. In 10 of the patients withulcerative colitis VIP nerves were thesame as in the normal controls. However,in the three cases with the most severe

inflammatory reactions VIP nerves were

found to be increased, although themorphology of individual fibres was

essentially normal.This study confirms the finding of

hyperplasia of VIP nerves in Crohn'sdisease and suggests that this is relatedto the level of transmural inflammation.

(F52)Partial colectomy promotes experimentalcarcinogenesis only at the site of intestinalanastomosis

R C N WILLIAMSON, P W DAVIES, J B

BRISTOL, AND M WELLS (University De-partment of Surgery, Royal Infirmary,Bristol) The possibility that metachro-nous colorectal cancers arise in mucosa

rendered hyperplastic by partial intestinalresection was tested in male Sprague-Dawley rats (N=200) weighing 159±14 g

(SD). Rats received mid-colonic transec-tion, right or left hemicolectomy, caecalresection, or no operation (controls).Each operation was performed eitherbefore or one week after a course of fiveweekly subcutaneous injections of azoxy-methane (10 mg/kg/wk) or water. Thetiming of operation did not affect ultimatetumour yields.Although caecal resection and right

hemicolectomy both increased ileal wetweight by 22-40% (P<0005), there wasno evidence of adaptive growth in thecolon after partial colectomy. A total of217 large-bowel tumours were distributed

as follows: proximal colon 16-5 %, colonicanastomosis 23-5 %, distal colon 40 %,rectum 200%. Consistent with this left-sided predominance, left hemicolectomyreduced the number of tumours by 52-58 % compared with transection orcontrols (P=005-0.005). Caecal resectionhad no effect. A twofold increase in distaltumours after transection and right hemi-colectomy (p < 0.02) reflected the manyanastomotic tumours found after eachoperation. Moreover, one rat receivingvehicle (not carcinogen) developed aninvasive mucinous adenocarcinoma at thecolorectal anastomosis after left hemi-colectomy.

Ileocolic and colorectal anastomosesare favoured sites for tumour develop-ment after partial colectomy. The lack ofadaptive hyperplasia may explain un-altered carcinogenesis in the remainingcolon.

(F53)Increased colonic carcinogenesis afterjejunoileal bypass in rats is prevented byextreme reduction in body weight

J B BRISTOL AND R C N WILLIAMSON (Univer-sity Department of Surgery, Bristol RoyalInfirmary, Bristol) Small bowel resectionconsistently promotes experimental intes-tinal carcinogenesis, whereas subtotalbypass may have a protective effect. Aschanges in body weight might be critical,colorectal carcinogenesis was studied aftervarious types of enteric bypass. MaleSprague-Dawley rats weighing 117±0-8 g(SEM) were given six subcutaneousinjections of azoxymethane (15 mg/kg) atweekly intervals. One week later, each ratwas subjected to 85-90%. jejunoileal by-pass (three groups) or sham bypass,which comprised jejunal transaction,ileotomy, and resuture. Bypass was (1)end-to-side; (2) end-to-side with a Thiry-Vella fistula; or (3) end-to-end withdrainage of the bypassed loop into thedescending colon.At 30 weeks the 46 surviving rats were

killed. Sham bypass rats weighed 586±23 g and had 3.9±1-0 colorectal tumoursper rat. End-to-side bypass more thandoubled the number of colorectal tumoursto 10-3±1i4 (P<0.01), despite reducingbody weight to 73% of that of shambypass rats (p < 0-001). Bypass withThiry-Vella fistula or loop into descendingcolon caused even greater weight loss to

A440

on Decem


http://gut.bmj.com

/G


ay 1981. Dow

nloaded from

http://gut.bmj.com/


55% of sham bypass. Colorectal tumouryields of 5.3 ±08 (Thiry-Vella fistula) and5.8 - 1.2 (loop into descending colon)were not significantly different from shambypass rats (p>005), but were less thanafter end-to-side bypass (e < 0-02; p < 0.05respectively.Enhanced neoplasia after jejunoileal

bypass may reflect compensatory colonichyperplasia, as after enterectomy, butthis effect is abolished by profoundreduction of body weight.

(F54)Bile acid binding to dietary protein

A LANZINI, R BIRD, W FITZPATRICK, PZENTLER-MUNRO, AND T C NORTHFIELD(Department of Medicine, St. George'sHospital Medical School, London) Bileacid precipitation occurs in jejunalsamples after Lundh meal in patients withpancreatic or ileal resection steatorrhoeaand healthy controls-50% at pH < 5 and20% at pH > 6. In pure solutions, bileacids do not precipitate unless pH < 5 forglycine and <2 for taurine conjugates.We have therefore investigated in vitrowhether precipitation at pH>5 is due tobile acid binding to Lundh meal consti-tuents. We incubated these with tauro-cholate (3706, pH 6.5) for one hour.Adsorption was calculated from differencebetween total and aqueous phase bileacid concentration after ultracentrifuga-tion at 100 000 g for eight hours, andexpressed as mM/g protein. At 5 mM/1,4.7±0t9x 10-smM/g (mean±SEM) wasadsorbed using Lundh meal, 5.3±04x10-5mM/g using milk protein (Casilan),and none using corn oil, sucrose, saline,or calcium. Very little binding occurredusing hydrolysate of Casilan. Adsorptionisotherms for Casilan showed that bindingreached a maximum of 101.2±12 x 10-5mM/g at bile acid concentration of 24mM, suggesting a saturable process. AtpH 4.5, adsorption was markedly in-creased at all concentrations. Binding wasgreater for dihydroxy than tridroxy bileacid, but was uninfluenced by type ofamino conjugation at pH 6.5. We con-clude that marked bile acid bindingoccurs to dietary protein-for example,milk; that this may contribute to steator-rhoea in patients with low total bile acidconcentrations-for example, ileal resec-tion; and that it can be prevented in vitroby using the constituent amino acids.

POSTERSFP1-FP12

(FP1)Faecal microbial flora in the irritablebowel syndrome

M GUSLANDI, A TITTOBELLO, A BALSARIAND E FESCE (introduced by S Gottfried)(Institutes of 3rd Medical Clinic and ofVeterinary Microbiology and Immunology,University of Milan, Milan, Italy) Theirritable bowel syndrome is a functionalgut disorder associated with an abnormalintestinal motility. In clinical situationswith altered peristalsis changes in intes-tinal microflora can be detected. We haveinvestigated the behaviour of intestinalflora in patients with the irritable bowelsyndrome.

Stool samples were collected from 20outpatients with the irritable bowelsyndrome and 20 healthy controls. Noneof them had received antibiotics duringthe previous month. Stool specimenswere immediately placed in sterile plasticcontainers with Aranki diluting fluid toprevent damage to anerobic bacteria,transferred in a 'glove box', homogenised,and diluted up to 10-9. Bacteriologicalexaminations were performed within twohours and bacterial populations of eachspecies were compared by the method ofKruskal and Wallis.Compared with control subjects, the

irritable bowel syndrome patients showeda significant decrease of coliforms(P < 0001), lactobacilli (P <002), andbifidobacteria (P < 0 05). No significantchanges in the other bacterial populationswere observed. Disturbances of peristalticmovements may account for removal ofsome of the autochthonous microflorafrom the intestinal tract.The significance of this altered micro-

bial pattern in the maintenance and inthe severity of the irritable bowel syn-drome remains unknown. However, ad-ministration of autochtonous lactobacilliand bifidobacteria can sometimes inducesymptomatic improvement in thesepatients.

(FP2)Diagnostic accuracy of the PABA excre-tion index (using 14C-PABA)

V A TETLOW, G KAY, H HERMAN, AND J MBRAGANZA (University Department of

Gastroenterology and the Departmentof Medical Physics, Manchester RoyalInfirmary, Manchester) In a six monthperiod 140 consecutive patients referredbecause pancreatic disease was considereda serious possibility, or after recrudes-cence of symptoms from chronic pan-creatitis, were investigated by a modifiedPABA test. Drugs were withdrawn andafter an overnight fast each subject drank500 ml of a flavoured cocktail containing25 g casein, 0.5 g N-Benzoyl-L-Tyrosyl-p-aminobenzoic acid and 5 pCi 14C-labelled PABA. Urine was collected forsix hours, PABA and 14C content mea-sured (expressed as percentage of ingesteddoses) and their ratio calculated to givea PABA excretion index which in healthis 0.76-112 (-3SD to +3SD).The diagnosis arrived at after full

investigation of each patient was com-pared with the PABA excretion index.

1. Drug or isotopic interference invali-dated 16 tests.

2. PABA excretion index was repro-ducible (r=0-97 in 10 studies).

3. PABA excretion index was normal in67 of 70 patients with non-pancreaticabdominal disorders-that is, specificity,95.7 %. When gastrointestinal anatomywas intact and renal function normal (64patients) 76.6±1.9% of the administered14C was recovered in urine in six hours.

4. PABA excretion index was reducedin 32 of 43 patients with chronic pan-creatitis or pancreatic cancer-that is,sensitivity, 74.40%.The predictive value and efficiency of

the test has been analysed.

(FP3)Premature bone ageing in primary biliarycirrhosis

O EPSTEIN, Y KATO, R DICK, AND S SHERLOCK(Academic Department of Medicine, andDepartment of Radiology, Royal FreeHospital, London) The study of bonedisease in primary biliary cirrhosis hasbeen restricted to patients with an indica-tion for bone biopsy, and, therefore, theprevalence and pattern of bone diseasehas not been fully established. An alterna-tive radiological approach is to measurethe shaft (S) and medullary cavity width(M) of the second right metacarpal, at itsmid-point. From these values, corticalthickness can be measured (S-M), as wellas the percentage cortical area (100S2-M2/S2), which reflects the proportionof bone in the bone envelope.

A441

on Decem


http://gut.bmj.com

/G


ay 1981. Dow

nloaded from

http://gut.bmj.com/


We have measured these parameters in83 patients with primary biliary cirrhosis(age range 37-78 years), and comparedthem with normal age-matched controls.Cortical thickness was significantly de-creased in all age groups, especially in the5th to 7th decades (p < 0001). The bonethinning was due to marked endostealresorption, reflected by a significantincrease in medullary width (p < 0001)and a significant reduction in percentcortical area (p < 0-025). There was nocorrelation between the reduction incortical thickness and serum bilirubin,nor was there a correlation with thehistological stage of the disease.

In primary biliary cirrhosis, there ismetacarpal cortical thinning, predomin-antly due to endosteal resorption, ratherthan failure of accretion. This resemblesthe normal pattern of ageing, occurringprematurely.

(FP4)Unusually enlarged liver due to heavydrinking for one year in a girl of 20: acase report

S K MAJUMDAR, A D THOMSON, G K SHAW, PO GORMAN, E J APS, AND J BUGLER (Gastro-enterology and Liver Unit, GreenwichDistrict Hospital, London, Elmdene Alco-holic Treatment Unit, Bexley Hospital,Bexley, Kent, Queen Mary's Hospital,Sidcup, Kent, and Brook General Hospital,London) (introduced by A D Thomson)Hepatomegaly due to chronic ethanolingestion is well known. We report belowa rare case of ethanol-induced hepato-megaly in a young unmarried English girl(accounts clerk) aged 20 years (date ofbirth 19 January 1960). She was admittedfor conventional detoxification treatmentfor ethanol withdrawal syndrome. Shestarted social drinking at the age of 14years but admitted drinking heavily forthe last one year (about three-quarters ofa bottle of vodka daily). Her parents areboth social drinkers. She was depressedclinically. There was no history of drugaddiction.The results of investigations (routine

haematological profile, prothrombin time,serum folate, alpha fetoprotein, bloodgroup (A-RhD positive), autoantibodies,plasma urea, creatinine, electrolytes, andradiograph of the chest (P-A view)) wereall within normal limits. Serum B2= 1250(N=160-900ng/l); Red cell folate=158(N= 160-640 tg/l); Australia antigen-negative.

EMI brain scan (CAT scan) showednormal examination but both temporallobes were faintly visualised. Straightradiograph of the abdomen showed liverenlarged up to the iliac crest. Liver biopsydisclosed a histologically normal liver,while liver scintigram showed an enlargedliver with normal spleen. There wasincreased activity in the spleen relative tothe liver but the features were thosesuggestive of hepatitis rather than earlycirrhosis. Liver ultrasound disclosed nospecific diagnostic lesion. The biliarysystem was not dilated.Known causes of hepatomegaly-for

example, malignant deposits, leukaemia,Hodgkin's disease, amyloidosis, conges-tive cardiac failure, gross steatosis-wereexcluded clinically, radiologically, bio-chemically, and histologically. This rarecase is unique and interesting because ofthe unusual hepatomegaly at such ayoung age with such a short history ofheavy drinking.

(FP5)Propranolol in the treatment of bile acidinduced diarrhoea

M J HALL, L M NELSON, R G MURRAY, ANDR I RUSSELL (Gastroenterology Unit,Royal Infirmary, Glasgow) Propranololhas been shown to inhibit bile acid-induced adenylate cyclase activity inhuman colon in vitro, and deoxycholicacid induced sodium and water secretionin rat colon in vivo.To determine if propranolol is effective

in man we studied seven patients withbile acid-induced diarrhoea. Propranololor placebo was given for four weeks andthe alternative preparation after a twoweek washout period. The design allowedfor an increase in propranolol dosagedepending on the degree of beta-blockadeas assessed by exercise testing up to amaximum of 360 mg daily. The patientskept diary cards, were interviewed blindly,and a five day stool was collected beforeand after each treatment period.

Six patients were beta-blocked on thedosage used. Three patients reported asubjective improvement in diarrhoea onpropranolol, but, of these, two reporteda similar improvement on placebo. Therewas no significant difference (Wilcoxon'ssigned ranks test) between the dailynumber of bowel motions on placebo(median 3-6; range 2 0-8.0) and pro-pranolol (3-2; 0.9-6.6) and none in dailyfaecal wet weights between placebo

(358 g; 88-440) and propranolol (274 g;87-450).We conclude that doses of propranolol

which produced significant cardiovascularbeta-blockade were ineffective in thetreatment of bile acid-induced diarrhoea.

(FP6)14C-triolein breath test as an outpatientscreening test for detecting steatorrhoca

A AVGERINOS, A K BEAVIS, J J MISIEWICZ,AND D B A SILK (Department of Gastro-enterology, Central Middlesex Hospital,London) The poor acceptability of thefaecal fat excretion test by patients,physicians, and chemical pathologists hasled to the development of tests in whichbreath 14CO2 is measured after the inges-tion of triglycerides labelled with 14C. Inthe present study we have compared thespecificity of a triolein breath test withmeasurement of faecal fat excretion in 35patients, 16 of whom had normal faecalfat excretion and 19 steatorrhoea (>18mmol fat/24 h). Hourly collections ofbreath 14C02 were performed for eighthours after ingestion of 5 IuCi '4C-triolein dissolved in a test meal containing25 g fat.

Breath 14CO2 increased slowly, peakingat six to eight hours. Mean hourly 14C02was higher at five, six, seven, and eighthours (p >0 01 or less) in the steatorrhoeacompared with the control group. Themaximum specific activity of 14C02 in thebreath at five to eight hours provided thebest discrimination between the twogroups, and at eight hours only four(21 %) of the 19 steatorrhoeic patientshad normal '4C02-breath excretion values(within normal mean+t2 SD).These findings confirm that the 14C-

triolein breath test is a sensitive non-invasive, and simple outpatient screeningtest for the detection of steatorrhoea,although in our hands its specificity isnot as precise as previously reported.

(FP7)Bacterial colonisation of the jejunum: anevaluation of five diagnostic tests

R H TAYLOR, A AVGERINOS, A J TAYLOR,M J HILL, AND J J MISIEWICZ (Departmentof Gastroenterology, Central MiddlesexHospital, London, and Bacterial Meta-bolism Research Laboratory, Central Pub-lic Health Laboratory, London) Whenbacterial colonisation of the jejunum is

A442

on Decem


http://gut.bmj.com

/G


ay 1981. Dow

nloaded from

http://gut.bmj.com/


suspected clinically, investigations toconfirm the diagnosis can be difficult,time-consuming, and unreliable, requiringthe use of expensive equipment. In thisstudy we have compared the results offive tests with quantitative analysis ofaerobic and anaerobic bacteria, includingidentification of all isolates, in samples ofjejunal fluid from 95 symptomaticpatients.

Gas/liquid chromatography of volatilefatty acids on 95 samples of jejunal juicegave positive results in 6/23 patients withsignificant colonisation (>104 faecalorganisms/ml) and 11/72 withoutcolonisation.

Analysis of urine for phenols asevidence of bacterial overgrowth in 23patients gave positive results in 0/6 withcolonisation and 3/17 without.

Glucose/hydrogen, lactulose/hydrogenand 14C-glycocholate breath tests weredone on 11 patients with confirmedcolonisation and 18 patients without.One or more breath tests were positivein 8/11 colonised and 9/18 withoutcolonisation. Positives in colonised pa-tients occurred in 6/11 on 14C-glycocho-late, 4/11 on glucose/hydrogen, and 1/11on lactulose/hydrogen. In those withoutcolonisation, positives occurred in 7/18on 14C-glycocholate, 3/18 on glucose/hydrogen and 2/18 on lactulose/hydrogen.After treatment of the colonised patientsall breath tests became negative exceptfor lactulose/hydrogen in one patient.None of these tests individually, or in

combination, appears to give a reliablemeasure of the degree and nature ofjejunal bacterial colonisation.

(FP8)Clinical course of solitary ulcer of therectum

M J FORD, J ANDERSON, S HOLT, H MGILMOUR, R C HEADING, AND W SIRCUS(Departments of Therapeutics and Patho-logy, Royal Infirmary, and the Gastro'intestinal Unit, Western General Hospital,Edinburgh) A review of 39 patients (18males, mean age 35 years, 21 females,mean age 42 years) with solitary ulcer ofthe rectum has provided additionalinformation about the natural history ofthis usunusal condition. Symptoms usu-ally comprised the passage of blood andmucus from the rectum, alteration inbowel habit, and anorectal or lowerabdominal pain; two patients wereasymptomatic. Difficulty with defaeca-tion was reported by 25 patients, of whom

12 admitted to self-digitation of therectum. At presentation solitary ormultiple ulcers were found within 13 cmof the anal margin in 25 patients and wereoften sited anteriorly. In 14 patientsmacroscopic ulceration was absent. Rec-tal prolapse was found in 20 patients.The diagnosis was confirmed histologic-ally in all patients; delay in diagnosisafter onset of symptoms was common, onaverage being five years.

Thirty patients were followed-up for amean of 4-6 years. Patients withoutmacroscopic ulceration at presentationdid not develop an ulcer during follow-up;thus two types of solitary ulcer wereapparent. Covert rectal prolapse wasfrequently associated with both but in somepatients prolapse was not demonstrable.The clinical course was unpredictable andneither medical nor local surgical treat-ment consistently achieved symptomrelief or ulcer healing.

(FP9)Prediction of the course of acute pancrea-titis using acute phase reactant proteins

P BRADLEY, MARGARET BOWDEN, E HCOOPER, AND M J MCMAHON (Unit forCancer Research, University of Leeds, andUniversity Department of Surgery, TheGeneral Infirmary, Leeds) It is importantto identify patients who develop severecomplications of acute pancreatitis suchas abscess or pseudocyst, but oftendifficult to do so, especially in the earlystages of an attack. We have investigatedacute phase proteins for this purpose asrapid nephelometric methods of assaymay soon be available.

Daily blood samples were collectedfrom 17 patients with acute pancreatitisfor measurement of C-reactive protein,al-antitrypsin, acid glycoprotein, anti-chymotrypsin, white cell count, and ESR.Three patients had a severe attack ofpancreatitis complicated by a pseudocyst(two) or abscess (one), 11 were classifiedas moderate, and were without complica-tions, and three were mild.A rise in the level of each parameter

was seen in all patients, but with theexception of C-reactive protein was notof value in differentiation of the groups.The level of C-reactive protein (g/l -3;mean± SD) two days after admissionwas 189±47 for severe, 117±55 formoderate, and 22±15 for mild pancrea-titis (P < 0-05 for all differences;Wilcoxon). Patients who developed com-

plications showed a sustained rise inC-reactive protein but, in other patients,levels fell from a peak at about 48 hours.

C-reactive protein may have potentialfor the early detection and investigationof complications of acute pancreatitis.

(FP10)Antral gastrin (G) cells in man after oneyear's treatment with cimetidine

J F MORRIS, N J MCC MORTENSEN, R WSPENCE, L R CELESTIN (Department ofHuman Anatomy, Oxford; Department ofSurgery, Bristol University; and Depart-ment of Surgery, Frenchay Hospital,Bristol) G cell hyperplasia occurs inachlorhydria and might complicate cime-tidine treatment. Integrated gastrin res-ponses (IGR) were raised three-fold overpretreatment levels in 24 patients withduodenal ulcer treated with cimetidine(16 g daily) for one year. Antral biopsieswere taken from these patients both pre-treatment (nine suitable) and after oneyear of treatment. The one year treatedgroup were divided into low integratedgastrin responses (< 14 000 ng/min/l;seven patients) and high integrated gas-trin responses (> 14 000 ng/min/l; ninepatients) subgroups. Gastrin cells wereanalysed immunocytochemically and ul-trastructurally. The numerical density ofG cells identified either immunocyto-chemically or ultrastructurally was notsignificantly different among the threegroups and did not correlate well withintegrated gastrin responses. This doesnot preclude hyperplasia due to increasedantral mass. The size of individual Gcells was not significantly changed but, inthe high-integrated gastrin responsegroup, cells contained significantly moreimmature granules (26±2/100lm2 Cyto-plasm; pretreatment 17± 3; low-inte-grated gastrin response 18±2 (mean±SEM), and increased rough endoplasmicreticulum. This suggests that the raisedintegrated gastrin response associatedwith one year of cimetidine treatmentresults from hyperactivity of individualG cells rather than G cell hyperplasia.

(FP1 1)Biological diagnosis of the hepatocellularcarcinoma associated with liver cirrhosis:clinical and experimental study

M MUNOZ, J A AMIGUET, F CONCHILLO,J PRIETO, AND P LISO (introduced by R E

A443

on Decem


http://gut.bmj.com

/G


ay 1981. Dow

nloaded from

http://gut.bmj.com/


Pounder) (Department of Medicine,University Clinic, University of Navarra,Pamplona, Spain) Hepatocellular car-

cinoma develops frequently from livercirrhosis and it is often very difficult todiagnose clinically. With this in mind we

studied two oncofetal antigens (carcino-embryonic antigen and alpha-fetoprotein)and five acute-phase reactants (alpha-i-acid glycoprotein, alpha-1-antitrypsin,haptoglobin, ceruloplasmin and alpha-2-macroglobulin) in 10 patients withhepatocellular carcinoma associated withliver cirrhosis.As controls we used 56 healthy subjects

and 58 patients who had liver cirrhosiswithout hepatocellular carcinoma.We found that the hepatocellular carci-

noma associated with liver cirrhosis can

be biologically differentiated from livercirrhosis by the positivity of the alpha-fetoprotein and by a significant increaseof haptoglobin and the other four acutephase reactants. Moreover, regular deter-minations of haptoglobin in patients with

liver cirrhosis showed us that, in fourcases, a rise in the haptoglobin levelcorrelated with malignity.

Furthermore, an experimental study ofhaptoglobin in rats with liver cirrhosisand with liver cirrhosis plus hepato-cellular carcinoma confirmed these results.

In conclusion, acute-phase reactantsand alpha-fetoprotein are good biologicalindicators for the diagnosis of hepato-cellular carcinoma which has developedfrom liver cirrhosis.

(FP12)Experimental carcinoma of the colon:serum haptoglobin in early diagnosis

J A AMIGUET, M MUNOZ-NAVAS, F CON-CHILLO, E ORTIZ DE LANDAZURI, AND

P LISO (introduced by R E Pounder)(Department of Medicine, UniversityClinic, University of Navarra, Pamplona,Spain) Haptoglobin has shown its use-fulness in the biological diagnosis of

colon cancer. The possible interest ofthis protein indicator for purposes ofearly diagnosis motivated this experi-ment.The weekly administration of 40 mg/kg

weight of 1,2 dimethylhydrazine to Wistarrats produced early neoplastic lesionsfrom the 15th week. Haptoglobin wasquantified by simple radial immuno-diffusion against a specific immunoserumanti-human haptoglobin 1-1, obtained byus through an original technique. Thisantiserum had been previously used inmeasuring haptoglobin in experimentalhepatoma of the rat. This subtype hapto-globin 1-1 possesses crossed antigenicitywith the rat variety.

Early histological lesions coincidedwith a statistically significant rise ofhaptoglobin in serum.The cause of haptoglobin serum in-

crease in neoplasms is still unknown.In conclusion, we have found that

haptoglobin is a useful biological markerin early carcinoma of the colon.

A444

on Decem


http://gut.bmj.com

/G


ay 1981. Dow

nloaded from

http://gut.bmj.com/

thebritish society of gastroenterology · hydrochloric acid for one month at 40°c, the breaking...

Documents