The British Society of Gastroenterology - Gut

Gut, 1985, 26, A1100-A1158

The British Society of Gastroenterology

The 46th Annual Meeting of the British Society of Gastroenterology was held at the University of Newcastleupon Tyne from 18-20 September 1985, under the presidency of Professor E L Blair. Out of the paperssubmitted, the Programme Committee of the Society selected 136 for presentation as posters, and 52 for oralpresentation; the abstracts* are printed below. Further details of the meeting appear in the News column onp. 1098.

BSG/BASL LIVERW1-13

WIErythrocyte aldehyde dehydrogenase acti-vity in health and disease and its value as amarker for hepatic aldehyde dehydrogenaseactivity

K MATrHEWSON AND C 0 RECORD (Gastro-enterology Unit, Royal Victoria Infirmaryand University of Newcastle upon Tyne)Considerable controversy surrounds therole of hepatic aldehyde dehydrogenase inalcoholic liver disease. Its activity is re-duced and this could be a pre-existingabnormality or simply a non-specific con-sequence of alcohol consumption. Erythro-cytes contain an enzyme identical to thehepatic cytosolic enzyme which couldprove to be an accurate marker of thehepatic enzyme with the added advantageof ready availability. Using a spectro-photometric technique, erythrocytealdehyde dehydrogenase was assayed in 15healthy controls, seven actively drinkingalcoholics (ADA), five subjects with alco-holic liver disease currently abstaining(AALD) and seven with non-alcoholicliver disorders (NALD). In addition thehepatic cytosolic enzyme was also assayedin 18 of the subjects. The ADA group hadsignificantly reduced erythrocyte enzymeactivity (p<0-01) but the AALD andNALD groups did not. The correlationsbetween erythrocyte and hepatic cytosolicactivity for all subjects consideredtogether, for the NALD group alone, andfor the ADA and AALD groups con-sidered together were all insignificant. Inthe control group alone, however, there

* Because of the time lag between acceptance of abstractsand the meeting, the data presented at the scientificsessions may not correspond exactly to the informationcontained in the abstracts.

was a significant positive correlation(r=0.943, p<005). We conclude thaterythrocyte aldehyde dehydrogenase acti-vity is reduced in actively drinking alco-holics but that red cell activity gives a poorindication of hepatic cytosolic activity insubjects with liver disease. In contrast, itappears a good marker of hepatic activityin normal subjects, and might therefore beused to study enzyme activity in 'pre-alcoholic' subjects.

W2Decreased red cell aldehyde dehydrogenaseactivity in patients with alcoholism

N R TURNER, S THOMAS, M J P ARTHUR, AND RWRIGHT (Professorial Medical Unit, South-ampton General Hospital, Southampton)Acetaldehyde has been implicated as amediator of hepatic injury in alcoholic liverdisease. Alcoholics develop higher bloodacetaldehyde concentrations than normalsubjects after drinking alcohol. Reducedactivities of aldehyde dehydrogenase(AldDH) have been reported in liver sam-ples from alcoholic subjects. An enzymesimilar to the cytoplasmic AIdDH of livercells is present in red blood cells and showsa similar reduction of activity in alcoholicpatients. If this reduction is specific forexcess alcohol consumption, it may beuseful in the detection and follow up ofpatients with early alcoholism.The red cell AldDH activities of patients

with a history of alcoholism (ALC, n=35)were compared with those activities foundin patients with non-alcoholic liver disease(NALD, n=21) and a group of normalcontrols (NC, n=32). AldDH activities inthe ALC group were significantly lowerthan those in both the NALD (p<0O05)and NC (p<001) groups. Moreover,patients with clinical evidence of continueddrinking had the lowest activities. Activi-

A1100

ties in the NALD group, however, werethemselves lower than those in the NCgroup (p<0-05). These results suggest thatdecreased red cell AldDH activity is ofsome value in the detection of alcoholismbut its major clinical potential is as amarker of continued drinking in alcoholics.

W3Immunoregulatory defects and aberrantcellular reactivity to liver cell antigens infamilies of patients with autoimmunechronic active hepatitis

C J O'BRIEN, S VENTO, A L W F EDDLESTON,AND R WILLIAMS (Liver Unit, King'sCollege Hospital and School of Medicine &Dentistry, Denmark Hill, London) Thefinding of a T suppressor cell defect specificfor liver membrane antigens in auto-immune but not HBsAg positive chronicactive hepatitis (CAH) raises the possi-bility that a discrete defect in immuno-regulation may be fundamental to thepathogenesis of this condition. As non-specific disturbances in immune functionhave been observed in relatives of patientswith autoimmune CAH, we have assessedT lymphocyte responses to liver derivedantigens and their suppressor T cell (Ts)control in healthy relatives and spouses ofpatients with autoimmune CAH. Using anindirect T lymphocyte migration inhibitoryfactor assay it was found that 18/18patients, 8/41 relatives, 3/11 spouses and0/23 unrelated controls exhibited sensitisa-tion to a liver membrane lipoprotein (LSP)complex. In contrast to all nine patientsstudied, none of the LSP sensitised rela-tives exhibited sensitisation to the asialo-glycoprotein receptor - a purified consti-tuent of LSP. In T cell co-culture experi-ments, using patient and non-patient cellsin a 9:1 ratio, T lymphocytes from 20/41relatives and 1/11 spouses (p<005) were

on July 14, 2022 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.26.10.A1100 on 1 O

ctober 1985. Dow

nloaded from

http://gut.bmj.com/


unable to suppress T cells from patientssensitised to LSP. These studies suggestthat a Ts cell defect for liver membraneantigen(s) is a genetically determinedmarker of susceptibility for autoimmuneCAH. Clinical disease only arises when inaddition to the defect, T cell sensitisationto the asialoglycoprotein receptor appears.

W4Natural killer (NK) cell activity in acuteviral hepatitis

L CHEMELLO, M MONDELLI, F BORTOLOTlTI, ESCHIAVON, A ALBERTI, E G RONDANELLI, AND

G REALDI (Istituto di Medicina Clinica,Cattedra di Clinica Medica 11, University ofPadua, and Department of Infectious Di-seases, IRCCS Policlinico S Matteo, Uni-versity of Pavia, Italy) Recent experi-ments in animal models suggest thatnatural killer (NK) cells play an importantrole in resistance to viral infections. In thisstudy, we have investigated natural cyto-toxicity for the NK-susceptible K562 targetcells in 24 patients with acute hepatitis B(AHB) and in 11 with acute non-A, non-Bhepatitis (NANBH), using a standard 51Cr-release assay. Simultaneously, serum levelsof ax-interferon (Q-IFN) were determinedby radioimmunoassay. Peripheral bloodmononuclear cells showed significantly in-creased cytotoxicity in patients with AHBat all effector to target cell ratios rangingfrom 10:1 to 80:1 (p<0.01). Percent speci-fic lysis was significantly higher in patientstested within 20 days of clinical onsetcompared with those tested during con-valescence, in whom cytotoxicity was notstatistically different from healthy controls(mean % cytotoxicity±SEM: 20-7±2 5,34 3±3*5, 50-5±3*7, 65±3-2, and 12±1,21-3±2-2, 32-3±3-4, 46-2±4-6, respec-tively, p<0-01). In contrast, patients withNANBH, all studied in the early phase,showed normal cytotoxicity. Serum con-centrations of (A-IFN were normal in bothgroups of patients. These functional studiessupport recent data showing increasedproportions of lymphoid cells with NKphenotype in the liver infiltrate of patientswith AHB. The finding of enhanced cyto-toxicity in the early phase of AHB suggeststhat NK cells may be important in controll-ing HBV infection before virus-specificcytolytic T cells become fully operative.Non-A and non-B agents do not generatesignificant NK activity and this may be oneof the factors contributing to frequentprogression to chronicity. Serum concen-

tration of ot-IFN was unrelated to NKcytotoxicity, at least in this setting.

W5Clinical and experimental studies of theinfluence of ethanol on paracetamol hepato-toxicity

J M TREDGER, HEATHER M SMITH, R B READ, BPORTMANN, R WILLIAMS (Liver Unit, King'sCollege Hospital and School of Medicine &Dentistry, Denmark Hill, London) Re-peated ethanol ingestion is thought topotentiate paracetamol hepatotoxicitywhile ethanol taken concomitantly withparacetamol may reduce liver damage. Of247 patients seen at King's College Hos-pital after paracetamol overdose, alcoholwas used in 60 (24%). There was nosignificant difference in clinical course oroutcome between patients consuming alco-hol either chronically (>80 g/week), con-comitantly with paracetamol or bothchronically and concomitantly. In mice,however, ethanol had pronounced effectson paracetamol metabolism and hepato-toxicity evaluated using semi-quantitativehistology (scale 0-5). Concomitant ethanoladministration prevented liver damage(score: 0±0 vs 3-2±0-5 after paracetamolalone) and decreased the production oftoxic paracetamol metabolites. Ethanolhad no effect on paracetamol covalentbinding in vitro. Chronic ethanol consump-tion potentiated paracetamol hepato-toxicity (score: 3-9±0-8; mortality 27%(0% in control); p<0.05) but did notinduce the enzymes catalysing paracetamolintoxication. In mice given ethanol bothchronically and concomitantly, paraceta-mol was mildly hepatotoxic (score:1-1±0i7; 0% mortality) and the acutemetabolic effects of ethanol were reduced.Interindividual variations in dosages ofethanol and paracetamol, in the use ofantidotes and in metabolising activity, mayexplain the differences between our humanand animal data.

W6Partial hepatectomy enhances proliferationof ectopically sited liver cells in experi-mental animals

S GUPTA, R JOHNSTONE, Y PRICE, H DARBY, A C

SELDEN, AND H J HODGSON (Department ofMedicine, Royal Postgraduate MedicalSchool, Hammersmith Hospital, London)

Experimentally, isolated hepatocytestransplanted into the spleen in syngeneicanimals eventually proliferate, formingconfluent hepatic cord structures. Thisprocedure offers a model for exploringcontrol of hepatocyte growth in a vascularbed not receiving portal blood, and alsooffers a potential therapy for metabolicdisorders. We have explored the effect ofpartial hepatectomy, the strongest knownstimulus to hepatic regeneration, on estab-lished ectopic grafts of liver cells in Augustrats. Ten months after implantation, twogroups of animals (n=6) underwent 70%partial hepatectomy or sham laparotomy.Cell proliferation was assessed 24 hourslater by autoradiography, after administra-tion of colchicine and 3H-thymidine. Innormal liver, <2 cells/1000 were labelled,compared with 300 cells/1000 in regenerat-ing liver. Ectopic hepatocytes in spleenswithout the stimulus of partial hepatec-tomy, were in a state of proliferationshown by a greater labelling index incomparison with normal liver (9 cells/1000), which moderately increased afterpartial hepatectomy (20 cells/1000). Theseresults show that the regenerative stimulusof partial hepatectomy reaches the sys-temic circulation, and ectopic hepatocytesrespond to this stimulus albeit not to thesame degree as the liver itself. Either localfactors, or the presence of nutrients inportal blood, may account for this differ-ence.

W7Can dynamic liver scanning replace liverbiopsy in methotrexate patients?

E ANN BINGHAM, M D O HARA, R FERGUSON, JD LAIRD, D BURROWS, AND M E CALLENDER(Royal Victoria Hospital, Belfast, NorthernIreland) Because methotrexate therapyfor psoriasis can produce hepatic fibrosisand cirrhosis without disturbance of serumliver function tests these patients requireserial liver biopsy, bridging fibrosis orcirrhosis being a contraindication to theinitiation or continuation of methotrexatetherapy.The 'dynamic liver scan' indicates hepa-

tic dysfunction by a reduction in the portalvein contribution to total hepatic bloodflow assessed by computer analysis ofhepatic and splenic uptake of 9Tc-colloidinjected intravenously.

Sixty paired dynamic liver scans and liverbiopsies were obtained in 46 patients withpsoriasis and a single pathologist assessed

A1101


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


the degree of fibrosis 'blind'. All patientshad normal serum liver function tests andnormal static isotope scans.

In 54 cases where the biopsy showed nofibrosis or only periportal or pericentralfibrosis the dynamic scan was normal in 44.In six cases where the biopsy showedbridging fibrosis or cirrhosis the dynamicliver scan was normal in one. Although thepredictive value of an abnormal dynamicscan is low at 5/15 (30%), the predictivevalue of a normal dynamic scan is high at44/45 (97-8%).

It may thus be possible to reduce liverbiopsies by 75% in psoriasis patients bybiopsying only those with an abnormaldynamic liver scan.

W8Effects of portasystemic shunting on hepatichaemodynamics in the cirrhotic rat

S A JENKINS, J N BAXTER, P DEVIrr, AND RSHIELDS (Department of Surgery, Univer-sity of Liverpool, Liverpool) The aim ofselective portasystemic shunts is to reduceportal pressure while maintaining a satis-factory liver blood flow. As there is littleevidence to support the theoretical advan-tages of selective shunts we compared theeffects of different types of portasystemicshunts on hepatic haemodynamics in thecirrhotic rat. Rats with dimethylnitro-samine-induced cirrhosis underwent one ofthe following shunts: end-to-end portacav-al (n=10); side-to-side mesocaval (n=10);mesocaval H-graft (n=10); or splenopan-creaticocaval (n= 10). Liver blood flow(LBF) and wedged hepatic venous pressure(WHVP) were measured before, immedi-ately after and four weeks following shunt-ing. Immediately following end-to-endportacaval shunting there were significantreductions (p<005 Student's t test) inWHVP (13.9±0-4 to 6 5±0-5 mmHg) andLBF (18.6±1.3 to 10-1±1-7 ml/min/100 g).After side-to-side mesocaval shunting LBFand WHVP fell by approximately the samemagnitude. Mesocaval H-grafting pre-served liver blood flow to a greater extent(19.3±1.2 to 14 7±0-9 ml/min/100 g) andreduced WHVP from 12-8±0 7 to 8-9±1 1mmHg. However, 60% of the grafts be-came thrombosed. Most marked preserva-tion of LBF followed splenopancreatico-caval shunting, but the reduction in WHVPwas the least, falling from 13-8±0-6 to10-3±0-9 mmHg. No further significantchanges occurred one month after opera-tion. These results suggest that shuntoperations for cirrhosis with portal

hypertension which maintain liver bloodflow may not decompress the splanchniccirculation as effectively as portacavalshunts.

W9Effects of a long acting analogue of somato-statin (SMS 201-995) on hepatic haemo-dynamics in the pig and on intravaricealpressure in cirrhotic patients

S A JENKINS, J N BAXTER, W A CORBETr, AND R

SHIELDS (Department of Surgery, Univer-sity of Liverpool, Liverpool) SMS 201-995is a new analogue of somatostatin with alonger biological half life. Furthermore,preliminary studies have shown that SMS201-995 is more potent than somatostatinin inhibiting glucagon and insulin releaseand gastric secretion. In view of the greaterpotency of SMS 201-995 we studied theeffects of the analogue on hepatichaemodynamics in the pig and on intra-variceal pressure in patients with cirrhosis.

Pigs (n=6), received iv infusions of 250,ug/h of SMS 201-995. Portal venous flow,hepatic artery flow, portal pressure (PP),systemic blood pressure and cardiac outputwere measured before and after the infu-sion of the SMS 201-995. In six patientswith cirrhosis, intravariceal pressure wasmeasured before and after iv administra-tion of 50 gg SMS 201-995.

In pigs, systemic administration of 250,ug SMS 201-995 significantly reduced PP(15-6±2-4 to 12-3±2-5 mmHg), portalvenous flow and hepatic artery flow. Innine patients with cirrhosis, a bolus injec-tion of 50 ,ug SMS 201-995 significantly(p<005, Student's paired t test) reducedthe intravariceal pressure (27-4±2-5 to15-8±2-1 mmHg). Furthermore, in onepatient undergoing an elective portacavalshunt, the PP was reduced from 29 to 22mmHg following iv administration of 50 j,gSMS 201-995.These results suggest that SMS 201-995 is

effective in lowering PP and may be ofvalue in the control of acute varicealhaemorrhage.

w1oEffect of acute and chronic propranololadministration on antipyrine and paraceta-mol clearance in chronic liver disease

P C HAYES AND I A D BOUCHIER (UniversityDepartment of Medicine, Ninewells Hospi-tal and Medical School, Dundee, Scotland)The effect of propranolol on hepatic drug

metabolism was examined in 12 patientswith chronic liver disease in a placebocontrolled double blind study. Clearance ofantipyrine and paracetamol which are oxi-dised and conjugated respectively by theliver was determined before, after 120 mgof propranolol over 24 hours and after sixand 12 months of a long acting proporano-lol preparation (Inderal LA 160 mg).

Propranolol therapy caused an acutereduction in antipyrine clearance(1586±1177 to 1349±1198 ml/h) comparedwith the placebo treated patients(1202±534 to 1444±229 ml/h, p=0-01).This reduced clearance, however, did notpersist with chronic administration.The clearance of free paracetamol re-

mained unchanged throughout the treat-ment period but clearance of total (freeand conjugated) paracetamol fell on pro-pranolol from 5824±2249 to 4101 ± 1898ml/h compared with the placebo group(4551±1202 to 4421±1148 ml/h, p<0-05).The reduced clearance of total paracetamolwas probably because of impaired renalexcretion of the conjugated metabolite andwas accompanied by a rise in serum urea inthe propranolol group which became signi-ficant after 12 months therapy.

Propranolol therapy therefore has anacute effect on hepatic drug metabolismwhich is not maintained with chronicadministration although renal excretion ofdrug metabolites may be modified.

WilProphylactic sclerotherapy of oesophagealvarices: a preliminary report

D R TRIGER AND A G JOHNSON (UniversityDepartments of Medicine & Surgery, RoyalHallamshire Hospital, Sheffield) Fiftythree patients (age 21-70) with cirrhosisand oesophageal varices which had notbled were subjected to wedged hepatic veinpressure (WHVP) measurement. Thosewith WHVP ,12 mmHg were randomisedto either sclerotherapy (n= 19) or observa-tion (n= 18). Groups were comparable withregard to age, aetiology, presence ofascites and Child's grading. Cirrhosis wasdue to alcohol (21), PBC (16), CAH (10)and others (6). Patients with WHVP <12mmHg were not randomised (n= 16).Patients in any group who subsequentlypresented with variceal bleeding weretreated by sclerotherapy. During a meanfollow up of 28 months survival in thosereceiving prophylactic sclerotherapy wasslightly better than in the control group(one year 81% vs 62%; two years 67% vs

Al1102


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


45%). This was entirely due to improvedsurvival in the alcoholics receiving sclero-therapy. Survival in the non-randomisedgroup was 89% at one year and 77% at twoyears. Variceal haemorrhage occurred in6/19 undergoing sclerotherapy, 7/18 in thecontrol group and 3/16 in the non-randomised group. Size of varices corre-lated poorly with WHVP and was a poorpredictor of haemorrhage. Only 6/16deaths to date have been due to varicealbleeding. No patient in the non-randomised group has died of bleedingvarices. Preliminary results do not suggestthat prophylactic sclerotherapy in non-alcoholic cirrhosis prolongs survival; thebeneficial effect in alcoholic cirrhosis mightbe related to alteration in alcohol con-sumption. Longer follow up with largernumbers of patients is required.

W12Prophylactic propranolol therapy inchronic liver disease

P C HAYES AND I A D BOUCHIER (UniversityDepartment of Medicine, Ninewells Hospi-tal and Medical School, Dundee, Scotland)The effect of long term propranolol (Inde-ral LA-160 mg) therapy in patients withchronic liver disease was examined in adouble blind placebo controlled study.Ninety five patients (41 alcoholic cirrhosis;18 primary biliary cirrhosis; nine chronicactive hepatitis; six cryptogenic cirrhosisand 21 'others'), predominantly Child'sgroup A & B, took part in the study, 47 inthe propranolol and 48 in the placebogroup. The groups were well matched forage, sex and severity of liver disease.Twenty five patients were withdrawn be-cause of adverse reactions (13 placebo; 12propranolol) and 64 patients completed 12months in the study. Fifteen patients werechanged to half strength trial medication.Eleven patients died during the 12 monthstudy period, eight in the placebo group(17%) and three in the propranolol group

(6.5%).Four of the deaths in the placebo group

were related to upper gastrointestinalhaemorrhage and/or hepatocellular failurecompared with none in the propranololgroup. Hepatoma and cardiovasculardeaths appeared with similar frequency.There was no consistent change in Child'scategory in either group and the onlysignificant change between the two groupswas a fall in serum testosterone on pro-pranolol (p=002).Long term propranolol therapy in

patients with chronic liver disease is freefrom serious side effects, does not impairliver function and may improve survival.

W13Enhanced respiratory burst activity of Cparvum elicited hepatic macrophages com-pared with normal Kupffer cells in the rat

M J P ARTHUR, P KOWALSKI-SAUNDERS, AND RWRIGHT (Department of Medicine, South-ampton University Medical School, South-ampton) We have previously shown thatoxygen derived free radicals promotehepatocellular necrosis in the C parvumlendotoxin rat model of liver injury. Theaim of this study was to compare therespiratory burst activity of normal Kupffercells (KC) with C parvum (28 mg/kg, giveniv, six days previously) elicited hepaticmacrophages (CPHM). These were iso-lated by collagenase perfusion of the portalvein and purified by centrifugal elutriation.Respiratory burst activity was studied inthe 'resting' state and after stimulation withphorbal myristate acetate (PMA) by 1-14C-glucose oxidation via the hexosemonophosphate shunt (HMPS). This re-flects the production of oxygen derivedfree radicals.There was a significantly increased yield

of CPHM (meanxl106+SD, 376±145)compared with normal KC (47±24,p<O-001). CPHM (n=8) demonstrated en-hanced HMPS activity compared with nor-mal KC (n=8) both in the resting state andafter PMA stimulation (mean±SD, CPHMvs KC: (a) resting, 3 41±1-65 vs1-12±0-33, p<0-01: (b) PMA stimulated5-35±2-99 vs 1-18±0-31 CPMx 103/mgprotein/60 min, p<0-001).The increased number of CPHM and

their enhanced respiratory burst activitydemonstrate a substantially increasedpotential for release of cytotoxic oxygenderived free radicals in this model of liverinjury.

OESOPHAGUSP1-17

P1Abnormal response to distension inoesophageal clearance disorders

G P N KENDALL AND D G THOMPSON (StMark's and The London Hospital, Lon-

don) Motor responses to oesophagealdistension were studied in 13 normal sub-jects and five patients with pooroesophageal clearance without structuralabnormality. Oesophageal manometry wasperformed, both proximal and distal to aballoon sited 10-15 cm from the gastro-oesophageal junction, during five minuteperiods of balloon inflation (I) and defla-tion (D). During distension, primary peri-stalsis was unaltered proximally, but wasreduced distally ((D) 7-5±1-8 (mean con-tractions/5 min±SEM) vs (I), 3-9+1-4,p<005). Secondary peristalsis increasedproximal to the distension ((I) 22-8±3-9 vs(D) 2-5±1-3, p<0.001) with a small distalreduction ((D) 5-2±1-5 vs (I) 3.5±1.3).Powerful aboral propulsion of the balloonwas induced.

Supine radionucleotide transit was pro-longed in all patients. One patient withcorkscrew oesophagus, exhibited no distalinhibition. In two patients with hypo-motility and two with normal manometry,distension induced neither a proximalincrease in contractions nor aboralpropulsion of the balloon.These results show a propulsive response

to balloon distension in the normaloesophagus, resembling the in vitro 'peri-staltic reflex'. Abnormalities of thisresponse may be a cause of impairedoesophageal clearance. Addition of bal-loon distension to standard manometrythus appears useful for identifying specificpropulsive abnormalities in such patients,particularly when other manometric para-meters are normal.

P2Barium radiology in gastro-oesophageal re-flux disease - a reappraisal

R J SELLAR, J S DE CAESTECKER, AND R CHEADING (Department of Radiology andDepartment of Medicine, Royal Infirmary,Edinburgh) Forty six patients (24 withheartburn, 22 with non-cardiac chest pain)were studied using radiology, endoscopyand 24 hour ambulatory intraoesophagealpH monitoring. Barium radiology was car-ried out using a 'physiological' method(rolling the patient into the right decubitusposition) and a 'compression' method(abdominal binder inflated to 100 mmHg)to induce reflux. Double contrast oesopho-grams were also obtained.

Thirty one patients had significant acidreflux by pH probe. Using this as astandard, the 'physiological' method de-tected 17, the 'compression' method 23 and

A1103


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


endoscopy 14 refluxers, with three, fiveand one false positives respectively. 'Com-pression' barium was the most sensitive(74%) and accurate (71%), while endo-scopy was the most specific (93%) but leastaccurate (58%) in detecting acid refluxers.Using double contrast barium technique,measurement of the internal diameter ofthe cardiac oesophagus [(IDCO); -25 mmabnormal] had a sensitivity of 67%, speci-ficity 91% and accuracy 74% in detectingacid refluxers. When the results of 'com-pression' method and measurement of theIDCO were combined, sensitivity was87%, specificity 69% and accuracy 81%.These results suggest that, by combining atest to provoke reflux with a double con-trast technique to measure IDCO, bariumradiology becomes a sensitive screeningtest for reflux disease.

P3Endoscopic ultrasonography in the investi-gation of oesophageal strictures

P J SHORVON, R A FROST, W R LEES, AND P B

COTFON (Departments of Radiology andGastroenterology, The Middlesex Hospital,London) Diagnosis of oesophageal stric-tures is made by barium swallow andendoscopy, but neither method suppliesinformation beyond the mucosa. Endo-scopic ultrasonography (EUS) enablesassessment of the depth of mucosal lesionsand the detection of associated enlargedlymph nodes.We have carried out EUS on five

patients with oesophageal strictures, usingthe prototype Olympus ultrasonic endo-scope which has a 7-5 MHz transducer.Three patients had known oesophagealcarcinoma. In one EUS demonstrated asmall resectable tumour with no enlargedlymph nodes, confirmed at operation. CThad suggested more advanced disease. Inthe other two patients the ultrasonic endo-scope could not be passed beyond thetumour despite prior endoscopic dilatationin one. Inoperability was demonstrated byvisualisation of an extensive tumourextending into the mediastinum and anenlarged subaortic node in one, however.In the other no associated lymph nodeswere seen above the tumour, and this wasconfirmed at operation.One patient had an oesophageal stricture

proximal to the anastomosis of an oeso-phagogastrectomy two years before. Nomucosal abnormality was visible and EUSclearly showed that the stricture was be-

cause of compression by surrounding en-larged lymph nodes.The last patient had a recurring benign

stricture despite dilatation. It was decidedto diathermy the stricture endoscopicallyand this was adjudged safe after demon-stration of the depth of fibrosis by EUS.Endoscopic ultrasonography gives

unique information about oesophagealstrictures and a study is in progress toassess its accuracy in staging ofoesophageal cancer.

P4Gastro-oesophageal reflux in the irritablebowel syndrome

H L SMART, D NICHOLSON, AND M ATKINSON(University Hospital, Queen's MedicalCentre, Nottingham) The irritable bowelsyndrome (IBS) and symptomatic gastro-oesophageal reflux (GOR) are commongastrointestinal disorders which may berelated. To investigate this possibility weexamined oesophageal symptoms, endo-scopic appearances and undertookoesophageal manometry and pH recordingin 25 consecutive patients with IBS. Week-ly symptoms of GOR were significantlymore common (p=0-0003) in IBS patients(52%) than in age and sex matched con-trols (17%). Similar significant differenceswere observed for globus and dysphagia.Endoscopy revealed macroscopic evidenceof oesophagitis in eight cases and micro-scopic oesophagitis in a further 11 subjects.Patients with IBS had a significantly lower(p<0-001) lower oesophageal sphincterpressure (14.0±5.3 cmH2O (mean±stan-dard deviation)) than 25 age and sexmatched controls (19.4±5.1 cmH2O) with-out symptoms of GOR. No other signifi-cant manometric abnormalities werefound. Ambulatory radiotelemetricoesophageal pH monitoring revealedabnormal reflux in 16 of 21 IBS patients.Two of 13 patients with frequent GORsymptoms had a normal pH study, whereasfive of eight with infrequent symptoms hadan abnormal study.We conclude that GOR is significantly

commoner in IBS and that the loweroesophageal sphincter pressure is signifi-cantly lower in IBS. These two findingsmay be causally related.

P5Comparison between three methods of

oesophageal pH recording in the diagnosisof gastro-oesophageal acid reflux

G BIANCHI PORRO, F PACE, S BARONI, F

PARENTE, AND M LAZZARONI (Gastro-intestinal Unit, L Sacco Hospital, Milano,Italy) Oesophageal pH-monitoring is themost sensitive method of detecting andquantitating gastro-oesophageal reflux(GOR). Prolonged pH testing is usuallydone as a 24 hour investigation, in order totake into account postural or other circa-dian changes in oesophageal pH. Thesuperiority of a 24 hour period of studyover shorter tests, however, such as over-night or postprandial test, has not exten-sively been evaluated in terms of specificityand sensitivity. The purpose of this study istherefore to assess if these two shorter testsretain the diagnostic validity of a 24 hourtest. Oesophageal pH testing was under-taken for 24 hours (with subjects in uprightposition during the day and supine whenretiring) in 20 patients (14 men, six women,mean age 46 years, range 18-64 years) withendoscopic and/or histologic diagnosis ofoesophagitis, and in 20 healthy volunteers(nine men, 11 women, mean age 25 years,range 19-42 years) free of any symptoms ofGOR. The 24 hour test, as well as the 12hour and the postprandial ones, yielded nofalse positive results, with a 100% diagnos-tic specificity and a 100% positive predic-tive value. Twelve hour and postprandialtests showed a lower diagnostic sensitivitythan the 24 hour test (50% and 70% vs81%), respectively).We conclude that a 24 hour oesophageal

pH monitoring is the most accurate anddesirable clinical test of GOR.

P6Oesophageal dysmotility after myocardialinfarction

J K RAMAGE, M DEAKIN, R EDGE, L JENKINS,AND J G WILLIAMS (Departments of Gastro-enterology and Nuclear Medicine, RN Hos-pital, Haslar, Gosport, Hants) The rela-tionship between oesophageal chest painand the pain of cardiac ischaemia has notyet been fully established.We have measured the time to clear 90%

of a liquid bolus of "mTechnetium sulphurcolloid from the oesophagus in 13 patientswith ischaemic heart disease (proven onexercise testing or angiography) and 37patients within 10 days of myocardial in-farction (MI). Mean ages were 60*3 years(infarcts) and 54*3 years (IHD).

A1104


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


Of the patients with recent MI, 70% hadabnormal transit times (greater than 17seconds) and 40% had transit times greaterthan 120 seconds. In the ischaemic heartdisease group, only one was abnormal at 47seconds, the rest being less than 17seconds. There was no correlation of tran-sit time with age, sex, drug therapy withthe beta blockers, maximum enzyme riseor site of infarct. Six of the infarct patientswere studied again more than two monthslater. Five of these improved (four re-turned to normal) and one remained thesame at 20 seconds. Four patients had diedbefore they could be restudied - all hadacute transit times greater than 120seconds.Oesophageal transit time is abnormal

after myocardial infarction, indicating dys-motility which may be a cause of episodesof chest pain in the postinfarct patient.

P7Does dilatation of benign oesophagealstrictures (BOS) affect gastro-oesophagealreflux?

R PENAGINI, M AL DABBAGH, P F EVANS, I FTROTMAN, AND J J MISIEWICZ (Departmentof Gastroenterology and Nutrition, CentralMiddlesex Hospital, London) The role ofoesophageal dilatation in the managementof peptic oesophageal strictures is wellestablished. Dilatation improves dysphagiabut it is not known if it has any effect ongastro-oesophageal reflux.We studied nine consecutive patients

(median 61 years; range 29-77 years, fivemen), with a BOS, admitted to hospital foroesophageal dilatation. Before and fourdays after the procedure each patient had astandard meal composed of foods withdifferent consistency to assess dysphagia(score 0-20), a standard swallow with ahigh density barium suspension to measurethe diameter of stricture and 22 hourintraoesophageal pH monitoring (Synecticssystem). All dilatations were undertakenusing the Celestin dilator to 18 mm. Analy-sis of results was done with the pairedWilcoxon test.

After dilatation the dysphagia score im-proved from (mean±SEM) 10.2±2.0 to18-9±0-7 (p<0.01), the diameter (mm) ofthe stricture increased from 7-7±0-7 to9-4±0-6 (p<005), while % of time intra-oesophageal pH was <4 in the upright,supine and upright+supine position did notchange significantly, being respectively13-7±2-7 vs 20-1±6-2, 18-2±6-2 vs19-7±7-0 and 15-6±3-2 vs 19-3±4-7.

A1105

We conclude that oesophageal dilatationhas a minor effect on radiologically deter-mined bore of BOS, does not make gastro-oesophageal reflux worse and has a strikingeffect on dysphagia.

P8Computed tomography evaluation of oeso-phageal carcinoma - intubate or operate?

G P MCENTEE, J P DUIGNAN, D A O'CONNELL, EO'MALLEY, AND D BOUCHIER-HAYES (MaterMisericordiae Hospital, Dublin, Eire)Recent studies have recommended placingperoral endoprostheses in patients withextensive oesophageal carcinoma, thusavoiding unnecessary surgery. This studyanalysed the role of computed tomography(CT) in evaluating oesophageal cancer withrespect to other organ involvement andtumour resectability. Fifty patients withhistologically proven disease (17 middlethird, 33 lower third) were assessed usingCT and subsequently re-assessed atsurgery. The findings were correlated by anindependent observer and the accuracy ofCT determined separately for middle thirdand lower third lesions.The accuracy of CT in evaluating organ

involvement in middle third tumours was84% (false positives five, false negativesfour, total observations 51) compared with94% for lower third tumours (false posi-tives four, false negatives six, total obser-vations 165). Regarding tumour resect-ability, the accuracy of CT for middle thirdtumours was 64-7% (false positives 0, falsenegatives six) compared with .90% forlower third lesions (false positives 0, falsenegatives three). Tumours deemed resect-able on CT were always resectable atsurgery, but nine tumours deemed unre-sectable on CT were in fact resectable atoperation.

In conclusion CT provided useful infor-mation pre-operatively regardingextra-luminal tumour spread but was notsufficiently accurate to replace surgery asthe final arbiter of tumour resectability.

P9Oesophageal ulceration after extravasationof sodium tetradecyl sulphate and ethanol-amine oleate during endoscopic sclero-therapy

J D R ROSE AND P M SMITH (Department ofGastroenterology, Llandough Hospital,Penarth, S Glam) It is said that 3%sodirm tetradecyl sulphate (STD) pro-

duces more ulceration during oesophagealsclerotherapy than 5% ethanolamineoleate (EO). Twenty cirrhotic patients withoesophageal varices were randomly treatedwith either EO or STD, and at weeklyinjection sessions the volume of sclerosantused, the number of radiologically demon-strable extravasations of a contrast-sclerosant mixture and the number ofulcers were recorded. The 10 patients ineach group were similar, except for theirinitial variceal score, being higher for thosereceiving EO (mean 19-9, range 9-32) thanthose receiving STD (mean 14-2, range7-20; p<005). Two patients from eachgroup did not complete the trial; tworequiring oesophageal transection and twodying of hepatic failure.Four oesophageal ulcers developed after

56 EO extravasations and four after 44STD extravasations (NS). The mean num-ber of treatments required to obliterate thevarices was 6-4 (range 4-10) for EO and 4-5(range 2-6) for STD but allowing for initialsize of the varices, the two agents wereequally effective.We conclude that EO and STD and

equally effective for oesophageal sclero-therapy, and, in small quantities, areequally safe.

PloEndoscopic sclerotherapy for bleedinggastric varices

YASSIN M YASSIN, MOHAMED S EITA, ANDABDEL MONEIM T HUSSEIN (Gastroenter-ology Unit, Medical Academy GeneralHospital, Kobri-El-Kobba, Cairo, Egypt)Gastric varices are often associated withoesophageal varices in portal hypertension,but bleed less frequently and seem to bemore difficult to control. The results ofendoscopic sclerotherapy to control theirbleeding are reported. Total obliteration ofall gastric variceal channels was possible inonly six of the 35 cases reported (17.1%),otherwise it was limited to bleeding anddistended columns. The main complicationwas large deep sclerotherapy ulcers in eightcases (25.7%) with four deaths; two ofuncontrollable haemorrhage and two ofrupture. Three more patients died in hos-pital of intra-procedural cardiac arrest,failure to stop bleeding, and hepatic fail-ure. Bleeding renewed early in five morecases, two only from their gastric varices,and one died eventually of a bleedingantral ulcer. Hospital deaths totalled eight(22.9%). During a one year follow up,bleeding recurred in five patients (only one


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


from gastric varices); two died. Two morepatients died of hepatic failure. The oneyear survival rate was 65-7%. These resultsindicate that the serious complications ofendoscopic sclerotherapy for bleeding gas-tric varices including recurrent bleeding arenumerous and frequently fatal. We suggestit should be limited to selected cases.Necrotic complications are greater with oldage and poor hepatic function.

P11Flow cytometric analysis of the DNAcontent of gastric cancer

K C BALLANTYNE, P D JAMES, R A ROBINS, R WBALDWIN, AND J D HARDCASTLE (Depart-ments of Surgery & Histopathology, Uni-versity Hospital, Nottingham, and CancerResearch Campaign Laboratories, Notting-ham) Abnormal tumour cell DNA con-tent (aneuploidy) is associated with worseprognosis in a variety of cancers and in arecent Japanese study only 17/54 (32%)gastric cancers were aneuploid.Seventy two consecutive patients,

median age 67 years (43-88 years) whounderwent gastrectomy between 1979-1982 were studied. DNA content wasmeasured by flow cytometry after disaggre-gating representative paraffin embeddedsections (2x20,u) with diaminidophenyl-indole hydrochloride.

Forty four (61%) had a significant popu-lation of cells (<10%) with an abnormalDNA content (aneuploid). Two separatetumour blocks were examined in 22 casesand concordance found in 17 (77%). Nocorrelation was found between DNA con-tent of primary tumours and histologicaltype, histological grade or pathologicalstage. Curative resection was carried out in40 cases. Twelve patients survived over twoyears and nine remain disease free. Themedian survival was 17 months (n=13) fordiploid tumours and 18 months (n=27) foraneuploid tumours.We conclude that factors other than

tumour cell DNA are responsible for theaggressive nature of gastric cancer. Only39% of cancers studied were diploid com-pared with 68% of tumours in Japan; thismay reflect a difference in the geographicalpattern of this disease.

P12Therapy in symptomatic advanced gastriccarcinoma

G RUSTIN, E S NEWLANDS, R H J BEGENT, K D

BAGSHAWE, J L MAiTHEWS, AND T COOKE(INTRODUCED BY A PARKINS) (CharingCross Hospital, Fulham Palace Road, Lon-don) Effective therapy is required forsymptomatic unresectable or recurrent gas-tric carcinoma. Following the response of a

gastric cancer producing HCG to a drugregimen used in choriocarcinoma, we haveassessed its efficiency in further patients.Twelve patients with advanced gastric

carcinoma received a weekly schedule ofetoposide, methotrexate and actimomycinD (EMA) alternating with cyclophophas-mide and vincristine (CO). Treatment con-tinued for 12 weeks unless there wasevidence of progression. In 10 tumoursDNA ploidy values were determined bymicrodensitometry after Feulgen stainingand DNA histograms plotted and classifiedas aneuploid or diploid. Response tochemotherapy was determined sympto-matically and by WHO criteria.

Five of the seven patients (71%) withobjectively assessable disease responded.Only two patients in the total group hadprogressive disease during the threemonths of chemotherapy. Of the fivepatients without assessable disease, symp-toms improved with complete pain relief,loss of ascites or relief of obstruction and intwo, marked improvement at repeatgastroscopy. By manipulating dosageappropriately, chemotherapy was welltolerated.

Six patients had both measurable diseaseand DNA ploidy estimations. Four hadaneuploid tumours and responded, twopatients with diploid profiles had progres-sive disease.The initial results using this novel

chemotherapy regime suggests it may be ofvalue to patients with gastric carcinoma.

P13Local gastric antibodies to Campylobacterpyloridis

B J RATHBONE, J I WYATT, B WORSLEY, S

SHIRES, L K TREJDOSIEWICZ, R V HEATLEY,AND M S LOSOWSKY (Departments ofMedicine and Pathology, St James Univer-sity Hospital, Leeds, and Department ofMicrobiology, University of Leeds, Leeds)The association between Campylobacterpyloridis (CP) and active chronic gastritis isnow well established. The colonisation ofnormal gastric mucosa and the subsequentdevelopment of gastritis has also beendemonstrated. We have previously identi-fied raised circulating IgG and IgA anti-bodies in CP +ve patients, but this sys-

temic immune response probably has littlerelevance to events at the mucosal level.

Using immunohistochemical techniquesand an enzyme linked immunosorbentassay (ELISA), local antibodies to CPwere studied both in gastric biopsies andgastric juice. Twenty two dyspepticpatients were studied: seven out of 13 CP+ve patients had demonstrable IgA to CPin their gastric juice, IgM was shown inthree +ve patients. No IgG antibody wasdetected in any patient.

IgG, IgA and IgM antibodies were con-sistently shown coating the surface oforganisms on inflamed mucosae. No anti-body labelling could be detected on organ-isms deep in the gastric pits, however.The local antibody response to CP does

not appear to inhibit bacterial colonisation.One possible reason is that the organismssituated deep in the gastric pits are, bytheir position, protected from the secretedantibody.

P14Ranitidine for stress ulceration: effect ofbolus or infusion administration

D L MORRIS, S MARKHAM, A BEACHEY, FIONAHICKS, K SUMMERS, P LEWIS, AND A BYRNE(Department of Surgery, University Hospi-tal Nottingham) Stress ulcers are a com-mon problem in critically ill patients andmay largely be prevented by antacid or H2antagonist administration. The optimalmode of ranitidine administration is un-known. Forty patients who all requiredrespiratory support on our intensive careunit underwent an untreated control periodof 12 hours and were then randomlyallocated to (1) ranitidine 50 mg six hourlyby iv bolus, (2) ranitidine infusions 0-125mg/kg/hr, or (3) ranitidine infusion 0-24mg/kg/hr. Gastric juice was aspirated hour-ly for pH measurement. Serum concentra-tions of ranitidine were assayed by HPLC.pH Data is currently available in 20

patients and good pH control was achieved(>pH 4) in all but three patients. Sixteenof 140 samples were <4 in group 1 com-pared with two of 100 in group 2, and 19 of101 in group 3.High peak serum concentrations (mean

2359 ,ug/l± 1593 SD) were seen immediate-ly after bolus administration with a meantrough concentration of 243±49 at sixhours. In the infusion groups a muchsteadier serum level was achieved. Meanserum concentrations at four and 12 hourswere 280 and 461 ,ug/l for group 2 and 429

A1106


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


and 740 ,ug/l for group 3.Ranitidine infusion produces adequate

pH control and has possible pharmaco-kinetic advantages over bolus admin-istration.

P15SK&F 93574 - Preliminary evaluation of a

potent and long-acting parenteral H2-receptor antagonist in man

W L BURLAND, JANE G MILLS, LINDA

RICHARDSON, AND KATHIE WAREHAM (SmithKline & French Research Ltd, The Frythe,Welwyn, Hertfordshire) SK&F 93574 is a

potent and specific histamine H2-receptorantagonist in animals with a long durationof action. The parenteral administration ofan antisecretory compound with such a

profile of activity could offer therapeuticadvantage in several indications, including(1) the management of patients at risk fromdeveloping stress-related lesions of theupper gastrointestinal tract; (2) the reduc-tion of gastric acidity and volume, whichare associated with morbidity and mortalityin the event of pulmonary aspiration.Two studies have been conducted in

healthy male subjects. Inhibition of theheart rate response to the intravenousinjection of impromidine 300 mcg was usedto establish an effective dose-range forSK&F 93574 in man and to examine thetime course of the response. Eleven sub-jects aged 24-41 years were studied onthree or four separate occasions when theyreceived doses of 10 to 60 mcg/kg by 15minute intravenous infusion: SK&F 93574:25 mcg/kg inhibited the response toimpromidine, the effect was dose relatedand appeared to be maximal after two tothree hours. At doses ¢50 mcg/kg theincrease in heart rate was reduced by atleast 50% 30 minutes after the start of theinfusion; 20-24 hours after SK&F 93574 50or 60 mcg/kg the heart rate response toimpromidine was reduced by 48% (n=4).SK&F 93574 2-5-15 mcg/kg had no

significant effect on impromidine stimu-lated gastric acid secretion; three hoursafter the start of the infusion of SK&F93574 25 mcg/kg stimulated gastric acidoutput was reduced by 69-78% (n=3);within the first hour of the infusion ofSK&F 93574 50 mcg/kg acid output was

reduced by 36 to 52% (n=4) increasing to96 to 100% in the second hour with 59 to80% inhibition still present during the sixthhour. The iv administration of SK&F 93574was well tolerated.

P16A new role for polyacrylates in gastricmucosal protection

P W DETTMAR, A G LYNN, E C LEACH, AND J G

LLOYD-JONES (INTRODUCED BY A ALLEN)(Departments of Pharmacology and Clini-cal Sciences, Reckitt & Colman, Hull) Therole of polyacrylates in preventing ethanol-induced gastric necrosis and their ability tobind to gastric mucus adherent to thegastro mucosal surface of the rat has beeninvestigated. These studies showed thatthere was a synergistic effect when thesodium polyacrylate carbomer 934P, whichis a high molecular weight non-absorbedpolymer, was administered together withthe anti-ulcer agent carbenoxolone sodiumwith a resultant increase in gastric mucosalprotection.The minimum dose of carbenoxolone to

significantly protect against the ethanol-induced gastric necrosis was 60 mg/kg(69.1% protection, p<0.01), carbomerpossessed only weak mucosal protectionactivity maximal at 50 mg/kg (39-1%,p<002). When 50 mg/kg carbomer wascombined with an inactive dose of carben-oxolone (5 mg/kg) significant protection(92.5%, p<0.001) was produced - that is,synergism was exhibited.

Alcian blue was used to detect thepresence of carbomer binding to mucus

adherent on the surface on the rat gastricmucosa. At the above dose carbomerbinding increased -by 118-9% (p<0-001)compared with control treatment and per-sisted for up to five hours.The gastric mucosal protection afforded

by a combination of carbomer and carben-oxolone is currently being clinically evalu-ated for the treatment of gastritis.

P17Early and median term results of verticalbanded gastroplasty in the management ofmorbid obesity

E R T C OWEN, F D BEGGS, AND A E KARK

(INTRODUCED BY A G COX) (Department ofSurgery, Northwick Park Hospital, andClinical Research Centre, Harrow, Middle-sex) We report the results of 21 cases ofvertical banded gastroplasty undertakenwith modification of the Mason technique.Essential technical features are exposure,pouch size construction and accurate outletbanding which will be illustrated. At pre-sentation patients were between 30% and154% above their ideal weight (mean96.3%).

One death occurred six weeks afterleaving hospital from pulmonary embol-ism, seven patients had minor woundinfections, but nine had a cholecystectomycarried out it the same operation. Onepatient had a wound dehiscence. Theaverage hospital stay was nine days.The mean follow up period was 12

months. Mean weight loss for all patientsduring this period was 27 kg (range 8-5-47kg). At 26 weeks patients were a mean of57% above their ideal weight. Thesefigures compare favourably with otherforms of gastroplasty.

After weight loss many patients have hadcosmetic procedures - a total of 12 opera-tions have been carried out.We conclude that vertical banded gastro-

plasty is a relatively safe, simple andreliable procedure. Surgeons performingbariatric surgery should appreciate theinevitable further multiple cosmetic pro-cedures which are necessary followingweight loss.

BASIC SCIENCEP18-34

P18Gastric mucosal protection by a throm-boxane synthesis inhibitor (TXSI)

C J HAWKEY, R P WALT, R T KEMP, B FILIPO-WICZ, J DAVIES, AND N K BASKAR (Depart-ment of Therapeutics, University Hospital,Nottingham) Prostaglandins (PDs) areprotective to gastric mucosa but throm-boxane synthesis enhances damage. Wehave investigated whether the TxSIdazmegrel affords protection. Male Wistarrats (200-220 g) were dosed orally withvehicle or dazmegrel 1-5 or 25 mg/rat (n=8all experiments). After two hours theywere challenged with sodium taurocholate100 mM in HCl 0-2 N and mucosal damagesubsequently graded 'blind'. Other ratswere killed two hours after dosing and exvivo release of TxB2 and PGE2 fromgastric mucosal fragments measured byradioimmunoassay.Dazmegrel 1 mg/rat inhibited TxB2

synthesis by 23±8% (mean±SEM,p<0.05), did not significantly affect PGE2synthesis (122±24% control) and reducedmucosal damage (median grade 2-5 to 1-5,p<0.05). Dazmegrel 5 mg/rat inhibitedTxB2 synthesis (by 34±5%, p<0-0001),did not affect PGE2 synthesis (102±16%control) and reduced mucosal damage

A1107


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


(grade 2 5 to 1, p<0.01). Dazmegrel 25mg/rat inhibited synthesis of both TxB2(by 46±7%, p<0-001) and PGE2 (by27±9%, p<005), without significant effecton mucosal damage (grade 2 5 to 2).

Thus, lower doses of dazmegrel selec-tively inhibit thromboxane synthesis andprotect gastric mucosa. At higher dosesboth selectivity and protection may be lost.Assessment of Tx SI's in man is worth-while.

P19Antigenic determinants of wheat protein incoeliac disease

C O'MAHONEY, C O'FARRELLY, M MANSFIELD,D G WEIR, A WHELAN, AND C F FEIGHERY(Departments of Immunology and ClinicalMedicine, St James's Hospital, Dublin,Eire) Wheat protein antibodies are foundin patients with untreated coeliac disease(CD). The major antigenic determinants ofthis antibody response are found in crudegliadin - a complex mixture of alpha, beta,gamma, and omega gliadins. Interest hasrecently focused on a-gliadin, due to itsremarkable amino acid sequence similarityto adenovirus 12- a virus which may beimplicated in the aetiology of coeliacdisease.

Using two assay systems we tested frac-tions of crude gliadin (obtained by ionexchange chromatography) with sera frompatients with coeliac disease (23) andnormal controls (5).

(1) In the enzyme linked immunosorbentassay sera from patients with untreated CDreacted with most of the gliadin fractions.The highest antibody levels were, however,found using the ox-gliadin fraction.

(2) Crude gliadin and its fractions werefurther separated on agarose isoelectricfocusing and then immunoblotted ontonitro cellulose paper strips. The strips werethen incubated with sera from patients andcontrols. The areas of reactivity werestained by immunoperoxidase. Again, allmajor antigens were contained in the x-gliadin fraction. Finally, a monoclonal anti-body to (x-gliadin reacted to several bandsof fractionated a-gliadin suggesting thepresence of a recurring identical antigenicmoiety.

This work confirms that (x-gliadin is themajor antigen of crude gliadin. This maybe due to a repeating antigenic site on themolecule. Thus is a susceptible individual,a damaging immune response, first elicitedby a viral infection, could be perpetuatedby immune activity against cross-reactive

cx-gliadin and cause the enteropathy ofcoeliac disease.

P20Does gastric fundic pH or distension effectileal absorption?

A BILGE AND J B ELDER (University Depart-ments of Surgery, Manchester Royal Infir-mary and University of Keele, Keele) Nodata are available on the influence ofgastric fundic pH or gastric distension onileal absorption. Male Sprague-Dawleyrats were prepared under general anaes-thetic with a 4 cm ileal loop and the gastricfundus isolated by ligation of the oeso-phagogastric junction and cross clampingof the antrofundic junction. Five minutesafter the introduction of tracer (100 gCiTc99 m) to the ileal loop, 0 1 ml portal bloodsamples from an indwelling cannula wereobtained at five minute intervals for 50minutes. The IVC was cannulated andinfused with 0-154 M NaCl containing01I% albumin at 3 ml/h. Thirty four controlrats (fasting gastric fundic pH (GF pH)1 5-5, no gastric distension), and fourgroups each of six rats with GF pHbuffered at pH 1.2, 4, 7, 9 and a group withGF Ph in the fasting resting range but withthe gastric fundus distended by injection of3 ml air were studied. When GF pH washeld at four ileal tracer absorption in-creased by 27% at 10 minutes and 32% at15 minutes (p<0001); when the gastricfundus was distended by air ileal absorp-tion after 15 minutes was 38% abovecontrol (p<0001). Alkaline GF pH re-duced ileal absorption (p<0-01). Fourweeks after truncal vagotomy and pyloro-plasty the increased ileal tracer absorptionnoted after gastric fundic distension wasabolished (n=5), but the effect of intra-luminal acid (GF pH, n=5) enhancing ilealabsorption remained. These data suggest ahumoral effect from the gastric fundus onileal absorption in the rat.

P21Correlations between the acute effects ofthioacetamide on hepatic morphology andon hexokinase activity and isoenzymic com-position in the rat

M A JEPSON, G M LAWRENCE, AND D GWALKER (INTRODUCED BY R COLEMAN)(Department of Biochemistry, University ofBirmingham, Birmingham) In normaladult rat liver, four hexokinase isoenzymescontribute to total glucose-phosphorylating

activity. The major, high-Km form, gluco-kinase, constitutes 88% of the total activ-ity, is entirely cytoplasmic and occurs onlyin hepatocytes where it has a predomin-antly perivenous zonal distribution. Theremaining, low-Km, activity is largely con-fined to non-parenchymal cells. Ten percent are tightly associated with the outermitochondrial membrane and up to 70%are due to type I hexokinase.During the two to three day period after

a single 200 mg/kg body weight injection ofthioacetamide, there is a 300-500% in-crease in low-Km hexokinase activitywhereas high-Km activity decreases to 10-20% of control values. Up to 20% of theraised low-Km activity is mitochondriallybound and the type II and type III iso-enzymes predominate in both the solubleand the particulate fractions.Recovery begins three to four days after

treatment and low- and high-Km activitiesreturn to control levels six to seven dayslater.The early changes in hexokinase activity

correlate with rapidly developing peri-venous hepatocytic necrosis and with theproliferation of non-parenchymal, ovalcells. The return to normality coincideswith the disappearance of the proliferatingzones, the reappearance of the originalparenchymal cell morphology and the re-emergence of normal metabolic zonationpatterns.

P22Permissive role for the vagus nerves in thegenesis of antro-antral reflexes in the anaes-thetised ferret

D GRUNDY, D HUTSON, AND T SCRATCHERD(Department of Physiology, The Universityof Sheffield, Western Bank, Sheffield) Anincrease in antral motility after distensionof the stomach depends, in part, on avago-vagal reflex activated by mechano-receptors in the corpus. In the presentstudy we have considered the possiblereflex effects of antral distension.The experiments were carried out on

urethane (1.5 g/kg) anaesthetised, splanch-nectomised ferrets. Antral motility wasrecorded manometrically from a catheterinserted through the pylorus. Antral dis-tension was achieved by passing salinethrough a second catheter inserted throughthe mouth and secured in the antrum by aligature across the incisura. Antral disten-sion with 5-10 ml of saline increased theamplitude of antral contractions. Vagalblockade, achieved by cooling the cervical

A1108


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


vagi to <4°C, caused a fall in antral toneand attenuated the contractions evoked byantral distension. This might implicate avagal reflex in the genesis of the evokedcontractions. In vagotomised animals,however, close arterial infusions of acetyl-choline sufficient to return antral motilityto basal levels allowed the reflex to againbecome manifest. We conclude, therefore,that the increase in antral motility follow-ing antral distension is mediated by localreflexes which require a tonic vagal input.

P23Secretin stimulates gastric mucus barrierthickness without increasing luminal mucus

N J H CARROLL, A ALLEN, AND B H HIRST(Department of Physiological Sciences,Medical School, University of Newcastleupon Tyne) The anaesthetised rat pro-vides a suitable model for studying simul-taneous changes in the protective adherentmucus barrier and the output of luminalmucus. Ligated stomachs were perfused viaa double lumen orogastric tube, and gastriccontents recovered and analysed for mucusglycoprotein. The thickness of the ad-herent mucus barrier was measured onunfixed gastric mucosal sections.

Secretin, 4 U/kg/h iv, caused a significantprogressive rise in adherent mucus thick-ness reaching a maximum median value of175 um (quartile values 165-230) (n=6),compared with controls 100 am (60-130)(n=6). Significant increases in adherentmucus thickness were also observed withtopical 16,16-dimethyl prostaglandin E2(dm-PGE2 5 ,ug/ml) (44% increase) andcarbachol (100 ug/kg ip 55% increase).Luminal mucus glycoprotein output wasunchanged following secretin infusion;0*17±0-02 (n=6) mg/30 min (mean±1SEM) compared with control values of0-14±0-02 (n=6) mg/30 min. In contrastdm-PGE2 resulted in a seven-fold rise inluminal mucus glycoprotein output.These experiments show a novel stimula-

tory affect of secretin on gastric mucusbarrier thickness. Further, increases inadherent mucus thickness are not neces-sarily associated with increased luminalmucus output.

P24Mechanism for the mucosal protectiveaction of polyacrylate on the gastric mucusbarrier

S N E FOSTER, A ALLEN, AND J P PEARSON

(Department of Physiological Sciences,University Medical School, FramlingtonPlace, Newcastle upon Tyne) Carbomer934P, a polyacrylate (3x106) acts syner-gistically with carbenoxolone, protectingagainst ulceration.Carbomer (1-10 mg/ml), when added to

gastric mucus glycoprotein (range 0*5-10mg/ml) in isotonic pH 2 buffer, produced alarge synergistic increase in mucus viscosityrising with mucus and carbomer concentra-tion to over 1000% of the theoreticaladditive viscosity (mucus/carbomer both at10 mg/ml). Carbomer also caused a smallersynergistic increase in the viscosity of thepepsin degraded glycoprotein (95%increase).

Pepsin activity at pH 2-2 was inhibited bycarbomer (45% inhibition of albumin hyd-rolysis at carbomer 4 mg/ml, pepsin 1-0,ug/ml). Two methods (centrifugation andultrafiltration) which measured bound pep-sin compared with free showed this inhibi-tion can be explained primarily by revers-ible binding of pepsin to carbomer. At lowpepsin concentrations (0.01-0.4 mg/ml) thepercentage bound to carbomer (4 mg/ml)rose linearly plateauing at higher pepsinconcentrations (1-5 mg/ml). Pepsin bind-ing was also dependent on carbomer con-centration - for example, for pepsin 2mg/ml, carbomer 4 or 0 4 mg/n}l, percen-tage bound = 73% and 31% respectively.The above results show that carbomer

may act in mucosal protection by physicallystrengthening the mucus barrier and byinhibiting degradation by pepsin.

P25Mechanism of acetate absorption in normalrat jejunum

A J M WATSON, M J KELLY, M WILKS, M J GFARTHING, AND P D FAIRCLOUGH (Depart-ment of Gastroenterology and Departmentof Bacteriology, St Bartholomew's Hospi-tal, London) Acetate has been proposedas an alternative to bicarbonate in oralrehydration solutions for correction of aci-dosis due to diarrhoeal disease. Our pre-vious experiments show that in the normalrat jejunum acetate, like glucose, stimu-lates sodium and water absorption.Although acetate is absorbed at a similarrate to glucose, its transport mechanism isstill unclear. Studies of the absorptionkinetics, competition with propionate, andion dependency have therefore been per-formed using steady state perfusion of thenormal rat jejunum in situ. Acetate absorp-tion from isotonic solutions containing

5-150 mmol/l of acetate apparently fol-lowed saturation kinetics (Vmax = 13Atmol/min/g, Km = 47 mmol/1). Competi-tion experiments showed that acetateabsorption from a 30 mmol/l solution wasinhibited by 70 mmol/l propionate(5-2±0-3 ,umol/min/g vs 2-9±0-3 ,umol/min/g; p<0-01). Absorption of propionatefrom a 70 mmol/l solution was also inhi-bited by 80 mmol/l acetate (9.9±1.5 Amol/min/g vs 5-6±0-4 ,umol/min/g; p<001).Ion substitution experiments (sulphate re-placing chloride, and choline and lithiumreplacing sodium) failed to show chlorideor sodium dependence in this model.

Demonstration of saturation kinetics andinhibition by propionate are compatiblewith carrier-mediated transport of acetateby the rat small intestine.

P26Secretion of adherent mucus gel by amphi-bian gastric mucosa in vitro

J P KEOGH, S MCQUEEN, A ALLEN, AND AGARNER (Department of PhysiologicalSciences, Medical School, Newcastle, andICI Pharmaceuticals Division, AlderleyPark, Macclesfield) The secretion ofadherent mucus gel has been studied invitro using gastric mucosal sheets (mountedin Ussing chambers) and stomach sacs(oesophageal intubated, pylorus ligated)from Rana temporaria.

In control mucosal sheet preparationsmucus thickness increased over one hourby 2-7 fold (from mean 72±11 ,um to92±16 ,um), while in the stomach sacs noincrease was observed (from 83±6 ,m to71±8 ,um). This may be related to mucosalstretching when mounted in the Ussingchambers since distension of stomach sacsby applying a luminal hydrostatic pressure(3 cm) caused a 2-2 fold increase in mucusthickness after one hour.

16,16-dimethyl prostaglandin E2 (10-5M) significantly stimulated mucus thick-ness after one hour on mucosal sheets andstomach sacs (dose dependent) by 1-4 foldand 2-7 fold respectively over the corres-ponding one hour control values. Carba-chol (10-3 M, nutrient side) also signifi-cantly increased adherent mucus thicknessby 1-4 fold in mucosal sheet preparations, aresponse inhibited by atropine. Addition ofcimetidine (10A M, nutrient side) did notaffect mucus thickness on mucosal sheetsafter one hour although acid secretion wasinhibited.These results are the first demonstration

in vitro of stimulated secretion of adherent

A1109


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/

AlllO

mucus by prostaglandins, cholinergicmechanisms and mucosal distension.

P27Effects of eicosanoids on amphibianduodenal luminal alkaline secretion in vitro

J P KEOGH, A M STANIER, J R HEYLINGS, A

ALLEN, AND A GARNER (Department ofPhysiological Sciences, Medical School,Newcastle, and ICI Pharmaceuticals Divi-sion, Alderley Park, Macclesfield) Epi-thelial bicarbonate secretion is believed tobe an important mechanism in duodenalmucosal protection against luminal acid.We have compared the stimulatory effectsof various eicosanoids on bicarbonatesecretion in the duodenum in vitro.Segments of Rana catesbeiana proximal

duodenum were mounted as cylinders inglass chambers and rates of luminal alkalin-isation measured by continuous back titra-tion to pH 7-4. Cumulative dose responseswere determined to serosal administrationof eicosanoids. Prostaglandin E2 (1 Am)was administered at the end of each experi-ment to elicit a maximal response in orderto standardise responsiveness between in-dividual mucosae. The following agents (10Am) produced significant increases in alka-line secretion, expressed as percentage ofPGE2 maximal response: PGE1 (100%),PGI2 (68%), PGF2a, (64%), PGD2 (62%),arachidonic acid (60%) 6 keto-PGF1a:(55%), leukotrieneD4 (48%), PGA2(48%), thromboxaneB2 (42%). Linoleicacid and leukotrieneB4 (lum) were inactive.Indomethacin (lum) attenuated arachi-donic acid response and enhanced that toPGE2.While prostaglandins of the E series are

the most potent stimulants of HCO3 secre-tion a variety of other eicosanoids are alsoactive.

P28Effects of protein kinase C activation on

intestinal fluid transport and blood flow

J D FONDACARO, J S STEFANKIEWICZ, L S

HENDERSON, AND A SJOQVIST (Smith Klineand French Laboratories, Philadelphia,PA, USA) Studies were designed toexamine the effects of protein kinase C(PKC) stimulation on fluid transport andblood flow in the small intestine of theanaesthetised cat. The phorbol ester 4p-phorbol 12,13-dibutyrate (PDB) was usedto activate PKC. Intraluminal administra-tion of PDB into a segment of isolated

small bowel produced a copious intestinalsecretion and a concomitant mesenterichyperaemia. Net fluid movement in theintestine was converted from absorption inthe control state to secretion followingPDB. Intravenous atropine reduced PDB-induced secretion by 56%; clonidine abol-ished the remaining secretory response.Intra-arterial administration of PDB pro-duced intense vasoconstriction. Further-more, intraluminal PDB caused increasedsegmental contractions in the bowel seg-ment, which were totally inhibited byatropine. In Ussing chamber studies of therat ileum, PDB increased short-circuit cur-

rent, inhibited net Na+ and Cl- absorptionand increased serosal-to-mucosal Cl- flux.These studies demonstrate that stimulationof PKC produces a full secretory responsein the small intestine. Secretion is likely tobe a result of inhibition of Na+ and Cl-absorption and stimulation of electrogenicCl- secretion. Evidence suggests that thissecretion is accompanied by a metabolichyperaemia. We propose that PKC playsan important role in the regulation ofintestinal fluid transport.

P29Duodenal brush border membrane HCO3-ATPase, implicated in HCO3 secretion is anexpression of alkaline phosphatase

J M WILKES, A GARNER, AND T J PETERS

(Clinical Research Centre, Watford Road,Harrow, Middx, and ICI Pharmaceuticalsplc, Mereside, Alderley Park, Macclesfield)HCOj-activated ATPase has been impli-cated in duodenal HCO3-secretion, an

important process in mucosal protectionagainst acid. Rat duodenal brush bordermembrane (BBM) was isolated by Ca2"precipitation, forming tightly sealed right-side-out vesicles capable of Na+-dependentglucose transport. The BBM was enriched16-fold in (x-glucosidase, with a 30% re-

covery. HCO -ATPase was enrichedseven-fold, confirming an association withthe BBM. The activity was strongly inhi-bited by 10 mM L-phenyl-alanine, an

inhibitor of alkaline phosphatase.BBM HCO3-ATPase, solubilised in Tri-

ton X-100, was subjected to anion ex-

change, gel exclusion and phenyl boronatecolumn chromatography. In all cases

HCO--ATPase co-eluted with alkalinephosphatase. A number of detergents usedover a range of concentrations failed toshow significant latent HCO3-ATPase inintact BBM vesicles. Brush border mem-

brane alkaline phosphatase shows similar


activation by HCO3 as Mg2+-dependentATPase activity. Detergent solubilisedduodenal BBM HCO3-ATPase thereforeappears to be an expression of alkalinephosphatase activity.

P30Site of kinin action in the intestine

G WARHURST, M LEES, N HIGGS, AND L A

TURNBERG (Department ofMedicine, HopeHospital, University of Manchester Schoolof Medicine, Salford) Kinins as mediatorsof the inflammatory response are potentsecretagogues and may, therefore, be im-portant in secretory diarrhoea associatedwith inflammatory diseases. Their mecha-nism of action involves an increase in theproduction of prostaglandins. We haveexamined the cellular site of these actionsby studying the influence of kallidin onisolated rat enterocytes and subepithelialtissues. Kallidin increased cyclic AMP con-centrations in a dose-dependent manner inwhole mucosa (epithelium + subepithelialtissue) (4-11±0-52 in control tissues to9-85±0-81 at 10-5 M and 12-5±1-56 at 104M pm/mg protein, n=5). Kallidin did not,however, influence cyclic AMP concentra-tions in isolated epithelial cells alone.Prostaglandin E2 and forskolin, however,did stimulate cyclic AMP concentrations inwhole mucosa and epithelial cells. Kallidinstimulated a 30-fold increase in prosta-glandin E2 production in whole mucosawithin one minute (0-036±0.02 in controltissues and 1-18±0-12 ng/min after lx10M kallidin). Prostaglandin release was par-tially inhibited by the removal of Ca2+from the bathing medium. Kallidin failedto stimulate prostaglandin E2 production insuspensions of epithelial cells alone. Weconclude that kinins raise prostaglandinproduction by subepithelial tissues and thatthe prostaglandins then activate epithelialcell adenylate cyclase and cyclic AMPproduction so leading to ion secretion.

P31Effect of electrical field stimulation onbicarbonate secretion by isolated amphi-bian duodenum

J R CRAMPTON, L C GIBBONS, AND W D W REES(Department of Medicine, Hope Hospital,University of Manchester School of Medi-cine, Salford) Electrical field stimulation(EFS) is a technique used to provokerelease of neurotransmitters from endo-genous neurones in both gut and exocrine


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


glands. Using an in vitro chamber prepara-

tion, the effect of EFS on bicarbonatesecretion by a 2 cm segment of bullfrog(Rana catesbeiana) proximal duodenumhas been examined. Repetitive trains ofimpulses were passed longituding throughthe mucosa by a pair of plantinum elec-trodes in contact with the epithelium.Preliminary experiments establishedthresholds of pulse wavelength (0 5 ms),voltage (50 V) and frequency (5 Hz) belowwhich no change in secretion occurred. Astimulus of one train per second with trainlength of 0*5 s containing square waves of 2ms duration and 100 V amplitude at a

frequency of 10 Hz was found to give an

increase in the rate of alkali secretion of50±28% (mean±SE, n=5, p<0 005).With cessation of stimulation secretionreturned to basal levels. Repeat stimula-tion produced a similar response for as longas the tissue remained viable. Addition ofthe specific neurotoxin tetrodotoxin (10'M) reduced the secretory response to EFSby 68±18% (mean±SE, n=5, p<0005)indicating that the mode of action is pri-marily related to excitation of endogenousneurones and may be due to release ofneurotransmitter. This study shows thatEFS is likely to prove a useful method ofdetermining the role of enteric neurones inthe control of bicarbonate secretion.

P32Cisapride inhibits the secretory action ofserotonin (5-HT) in mammalian small intes-tine

K J MORIARTY, N B HIGGS, M WOODFORD, G

WARHURST, AND L A TURNBERG (Depart-ment of Medicine, Hope Hospital, Univer-sity of Manchester School of Medicine,Salford) Cisapride is a synthetic drugwhich stimulates gastrointestinal motoractivity in animals and man. This effectmay be mediated in part by inhibition ofserotonin (5-HT). Furthermore, cisapridehas been shown to bind in vitro to type 2serotonin receptors. We examined the in-fluence of cisapride on ion transport across

intestinal mucosa in vitro and studied itseffect on the action of 5-HT. Segments ofileum of male Sprague-Dawley rats were

stripped of muscle layers and mounted influx chambers. The addition of cisapride(5 x 10-5 M) to the mucosal and serosalaspects of the mucosa had no effect on theshort-circuit current, transmural potentialdifference, resistance or sodium and chlo-ride fluxes across the mucosa. The applica-tion of serotonin (10-5 M) to the serosal

aspect of the mucosa caused a rapid in-crease in short-circuit current and potentialdifference. Cisapride, however, inhibitedthis response in a dose-dependent manner

and blocked it completely at a concentra-tion of 5x10-5 M. Serotonin (5x 10-5 Mserosally) increased serosal to mucosal fluxof chloride from 12*6±0-8 to 15-2±0*6,umol/cm2/h (p<0 025), thus reducing netchloride absorption from 4.65±0.81 to1.49±1-04 IAmol/cm2/h (p<0.05). Thiseffect was completely blocked by cisapride(5x 10-5 M). In summary, cisapride inhibitsthe secretory action of 5-HT in the ratsmall intestine, possibly by blocking type 2serotonin receptors.

P33Is cholera toxin-induced intestinal secretionmediated via a neurogenic mechanism?

We conclucintestinal secrvia a neurolol

P34Electrophysiolcrypt-villus ax

C P STEWART Ament of Medisity of MancSalford) Incells at differaxis were im

under visualpreparation othe middle awere punctureopenings andcrypt. In glucopotential diff

K J MORIARTY, N B HIGGS, M WOODFORD, AND rime to C1 nl

L A TURNBERG (Department of Medicine, brane potentHope Hospital, University of Manchester (n=7) in mid-School of Medicine, Salford) Cholera base-villus celtoxin is thought to stimulate intestinal ing and withsecretion by direct activation of mucosal 58-2±2 0 mVadenylate cyclase. Lundgren and co- (n=5) respectworkers, however, provide evidence that was 100±11-2cholera toxin stimulates secretion in vivo al resistanceindirectly via enteric neural reflexes. We (n= 15), indicexamined this hypothesis further by study- RT resides ining the influence of neuronal blockade on The response

cholera toxin-induced changes in fluid assessed durirtransport across rabbit ileum in vitro. glucose had lMucosa, stripped of muscle layers, was meters, butmounted in flux chambers and crude partial depolhcholera toxin (1 ,ug/ml mucosally) caused a cells but notdelayed but sustained rise in the short- and RT did r

circuit current and electrical potential dif- results show tference (PD). The nerve blocking drug, ise transportitetrodotoxin (10-7 M and 5 x 10' M sero- crypt-villus a

sally), failed to influence the subsequent means of elecresponse to cholera toxin and addition oftetrodotoxin at the peak response tocholera toxin also had no effect. Thattetrodotoxin could block neurally-mediated secretagogues was confirmed bythe demonstration that the short-circuit SMALL BOWEL

current and PD responses to neurotensin P35-51(10-7 M) were blocked by tetrodotoxin(5 x 10' M). Furthermore, the response to P35cholera toxin of segments of ileum, which Osteoporosisincluded circular and longitudinal muscle bowel diseaselayers as well as enteric neurones, was notinfluenced by tetrodotoxin. An analysis of D JUDD, W EV

sodium and chloride flux responses to RHODES, AND

cholera toxin in the presence and absence of Gastroenteof tetrodotoxin suggested that the effects of cal Physics,cholera toxin on ion secretion were not Heath Park,inhibited by neuronal blockade. oesteoporosis

de that cholera toxin-inducedretion in vitro is not mediatedgical reflex arc.

ilogical recording along thexis of rat ileum in vitro

AND L A TURNBERG (Depart-icine, Hope Hospital, Univer-chester School of Medicine,this study, surface epithelialrent sites on the crypt villuspaled with micro-electrodescontrol. Using an in vitrof stripped rat ileum at 31°C,nd basal third of the villused, as well as cells at the cryptat a short distance within theose-free medium, transmuralference (Vms) declined withnV. The brush-border mem-tial (Va) was 459±4*4 mV- and 52*8±17 mV (n=25) inIls, whereas at the crypt open-lin the crypt itself, Va was

i (n=23) and 59X2±2X8 mVtively. Tissue resistance (RT)2 Q cm2 (n= 10), and fraction-(AVa/AVms) was 0-61±0*05cating that more than half ofthe brush-border membrane.to 10 mmol/l D-glucose was

ng some impalements. Serosalno effect on the above para-mucosal addition caused a

arisation of Va in some villusin crypt openings. AVa/AVmsnot change significantly. Thethat it is possible to character-ing epithelial cells along thelxis in rat small intestine byctrophysiological recordings.

in patients with inflammatory

IANS, E 0 CRAWLEY, C EVANS, JIJ E COMPSTON (Departmentserology, Radiology and Medi-University Hospital of Wales,Cardiff) The prevalence ofs in patients with inflamma-

Allll


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


tory bowel disease (IBD) has not beenaccurately established. Using single photonabsorptiometry of the radius and vertebralquantitative computerised tomography tomeasure bone mineral content (BMC) wehave determined the prevalence of corticaland trabecular osteoporosis in 58 patients(25 men) with large bowel (n= 17) or smallbowel (n=14) IBD.

Osteoporosis defined as >2 SD belowmean BMC was shown in 13 patients (eightwomen) with mean age 43 years (range21-77 years). Four had both cortical andtrabecular osteoporosis, six cortical only,and three trabecular only. Of these, twowomen aged 33 years and 38 years hadsevere clinical osteoporosis with loss ofheight and multiple vertebral fractures:three other patients had one or morevertebral fractures. All 13 patients with lowBMC had small bowel IBD with previousresections and 10 had received large dosesof steroids (>10 g total). Three of theseven premenopausal females with lowBMC were amenorrhoeic.

In this group of patients with IBD,cortical and/or trabecular osteoporosis waspresent in 22%, with severe clinical diseasein two young women. Patients with smallbowel IBD and previous resections whohave received large doses of steroidsappear to be mainly at risk. Within thiscategory, premenopausal women withamenorrhoea may develop severe clinicalosteoporosis and oestrogen replacementtherapy should be considered in suchpatients.

P36Effect of codeine and loperamide on carbo-hydrate malabsorption in postvagotomydiarrhoea (PVD)

J D O BRIEN, D G THOMPSON, H IBBOTSON, W R

BURNHAM, AND E WALKER (Departments ofGastroenterology, The London and Old-church Hospitals, London) Codeine phos-phate (60 mg) is more effective thanloperomide (4 mg) in reducing the speed ofupper gut transit in normal subjects (Br JClin Pharmacol 1985; 19: in press). Ouraim was to examine the effects of thesedrugs in patients with diarrhoea after trun-cal vagotomy, a disorder in which symp-toms are associated with rapid upper guttransit and nutrient malabsorption.

Orocaecal transit was studied in sevenpatients and in seven matched controls byexhaled breath hydrogen (H2) samplingafter a 20 g lactulose containing meal.Control transit was 63-3±2-9 min

(mean±SEM) vs PVD, 28-1±1-5 min,p<0-01).The functional absorptive capacity for

carbohydrate was then studied using a mealcomprising 50 g glucose in 250 ml water.No control showed a breath H2 rise,indicating complete absorption. In con-trast, all patients showed a breath H2 rise(>20 ppm) with subsequent diarrhoea,indicating incomplete absorption. Repeatstudies with barium added to the meal andcaecal screening, confirmed that this H2rise was due to rapid transit, not bacterialovergrowth. Prior administration ofcodeine (60 mg) abolished the H2 rise anddiarrhoea in all patients, indicating im-proved glucose absorption. Loperamide (4mg) had no effect. After one monthcodeine therapy, all patients reportedsymptomatic relief. Codeine thus seems tobe a more rational therapy than loper-amide for reducing carbohydrate malab-sorption in PVD.

P37Galanin in gut peptide secreting tumoursand its diagnostic value in phaeochromo-cytomas

F E BAUER, G W HACKER, T E ADRIAN, J M

POLAK, AND S R BLOOM (Departments ofMedicine and Histochemistry, RPMS,Hammersmith, London) The new intes-tinal peptide galanin, originally isolatedfrom porcine gut with potent biologicalactions on smooth muscle contractility andinhibition of insulin secretion, was found insignificant quantities in the gastrointestinaltract of different species including man.Therefore, we investigated galanin-IR ingut peptide secreting tumours andphaeochromocytomas with a newly de-veloped radioimmunoassay with both N-and C-terminal directed antibodies and byimmunocytochemistry. In none of the pan-creatic endocrine tumours was galanin sig-nificantly raised compared with normalpancreatic tissue. In phaeochromo-cytomas, however, the galanin content wassignificantly higher (tissue: 21±2-3 pmol/g,n=11, x±SEM, plasma: 161±21-5 pmol/l,n=6) than those of normal adrenals(2.6±0 9 pmol/g, n=4) and plasma ofvolunteers (<50 pmol/l, n=6). Gelchromatography and HPLC of adrenal andphaeochromocytoma extracts revealed twomolecular forms compared with one formin the porcinie standard. The C-terminalantibody did not detect human galaninsuggesting C-terminal molecular differ-ences. Immunocytochemistry localised

galanin-IR to cells in phaeochromo-cytomas. There is no evidence that galanin-IR is produced by peptide secreting gastro-intestinal tumours. The raised galaninlevels in phaeochromocytomas, however,may be responsible for some symptoms -that is, abdominal pain, constipation andhyperglycaemia. Plasma galanin could be amarker in the diagnosis of these tumours.

P38Comparison of modular elemental andpolymeric liquid diets on growth, nitrogen(N) balance, N wastage, faecal residue andhepatic lipid in rats

R H R PARK, A DUNCAN, G MITCHELL, W EAST,AND R I RUSSELL (Gastroenterology Unit,Royal Infirmary, Glasgow) New modularenteral liquid diets offer greater flexibilityof nutritional therapy. A controlled meta-bolic study was carried out to investigatetheir metabolic and nutritional effects.Vivonex HN (VHN), enteral 400 (E),elemental 028 (ELE), pepdite (P), MCTpepdite (MCT), and control rat chow(Oxoid 41B) (0), differing in composition,were fed to 36 rats (six rats in each group)for 28 days in isocaloric amounts (62 kcal(260 kJ) per rat per day). Mean weight gain(mean±SEM % of initial weight) was lessfor P (41±3-5) and MCT (40±3.5) than 0(59±4.3) (p<0.01), VHN (52±3.3)(p<0-05 and E (68±3.4) (p<0-001). Nbalance (mmol/24 h) was significantly in-creased with VHN (26.7±1.37) comparedwith E (17.4±1.5) (p<0.001), ELE(14.7±0.56) (p<0-001), P (15.5±0.81)(p<0-001) and MCT (16-9±0-99)(p<0.001). No significant differences wereobserved for mean N wastage (N excretionas % intake). Faecal residue (mg/24 h dryfaecal weight) for VHN (203±5) was re-duced significantly compared with ELE(304±8) (p<0-001), P (386±22) (p<0-001)and MCT (267±5) (p<0.001). Hepaticlipid (mg/g liver) was increased significant-ly with the elemental diets VHN and ELE(57.8±4-6 and 56-5±5-1 respectively) com-pared with the polymeric diets E(41-4±1.8) (p<0-01) and P (42-7±4.5)(p<0-05). New modular enteral liquid dietsdo not offer extra nutritional advantagesand VHN remains the diet of choice forvery low faecal output.

P39cx-1-antitrypsin (AT) and 51-CR-albumin(CrA) in the assessment of faecal proteinloss

A1112


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


E M M QUIGLEY, I N ROSS, M R HAENEY, I B

HOLBROOK, AND M N MARSH (UniversityDepartments of Medicine and ChemicalPathology and Department of Immun-ology, Hope Hospital, Salford) Althoughfaecal AT has been proposed as a reliablemeasure of enteric protein loss the accu-

racy and reproducibility of this methodremains uncertain. Our aim was to com-

pare AT excretion with the standard 5"Cr-labelled albumin technique. ct-1-antitrypsinand CrA excretion were compared in: (i)single 24 hour stool samples from 20normal subjects, (ii) five day stool collec-tions pooled from each of 35 patients withvarious GI diseases, and (iii) eight serial 24hour collections from each of seven

patients with small intestinal disease.Faecal CrA was expressed as % injecteddose excreted during the collection period.oa-1-antitrypsin was measured by singleradial immunodiffusion and expressed as:

(i) faecal excretion (mg/g dry weight or

mg/ml faeces), and (ii) intestinal clearance(ml AT/24 h). Mean values for excretionand 24 hour intestinal clearance of AT innormal subjects were 0*7 mg/g dry weight(range 0.1-6), 0-3 mg/ml faeces (0-03-2.3)and 5*2 ml/24 h (0.4-73.4) respectively. Inthe GI patients sensitivity and specificity ofAT excretion in defining excess proteinloss were 44% and 88% respectively, com-

pared with CrA. Measuring AT excretionin serial 24 hour collections did not im-prove its accuracy indicating that samplingerror alone cannot explain the poor sensi-tivity. We conclude that while excretion ofAT is significantly increased in patientswith various gastrointestinal diseases, thereis considerable overlap with the controlrange so that correlation with CrA loss isextremely weak (r2: 0.02-0-34). In indi-vidual patients AT excretion underesti-mates true protein loss and is therefore an

unreliable test.

P40How does dietary lipid lower blood alcohollevels?

I MCL WELCH, A MCFARLANE, L POOLEY, AND

N W READ (Clinical Research Unit, RoyalHallamshire Hospital, Glossop Road, Shef-field) To determine the mechanism,whereby food lowers blood alcohol levels,gastric emptying and blood alcohol profileswere measured in six healthy malevolunteers after ingestion of a 200 mlsolution of vodka and orange juice contain-ing 0-5 g/kg alcohol. Subjects were studiedon two separate occasions during infusion

of isosmotic solutions of either intralipid orsaline into the ileum. Gastric emptying wassignificantly delayed by ileal infusion ofintralipid (t': 149+17 vs 46±6 min,p<O-005) and the peak blood alcohol levelswere significantly depressed (24±4 vs 37±3mg/100 ml, p<0.01). Similar effects wereobserved in three subjects when the solu-tions were infused into the duodenum (t0:126±28 vs 34±6 min) (peak alcohol 19±4vs 36±3 mg/100 ml). The results suggestthat the reduction in alcohol absorption bylipid in food does not depend on thephysical relationship between the alcoholand lipid or between the lipid and absorb-ing epithelium, but is probably caused by adelay in the delivery of alcohol to the uppersmall intestine where it is rapidly absorbed.

P41IgG subclass antibodies to wheat gliadin inpatients with coeliac disease

P J CICLITIRA, H J ELLIS, AND M J KEMENY

(Gastrointestinal Unit, Department ofMedicine, The Rayne Institute, St Thomas'Hospital, and Department of Medicine,Guy's Hospital, London) Circulating anti-bodies to dietary antigens including wheatgliadin are present in coeliac patients.Untreated coeliac jejunal mucosa secretesmore IgG, M and A antibodies to gliadinthan casein. IgG can be divided into foursubclasses of which IgGI and 3 can bindand therefore activate complement whileIgG2 can to a lesser extent and IgG4cannot.

Titres of circulating IgGi, 2, 3 and 4subclass antibodies to gliadin and casein, adietary control protein, were measured byELISA in normal subjects (n=12), treated(n= 12) and untreated coeliac patients(n=12). Untreated coeliac patients hadgreater IgGl titres to gliadin and casein(p<0.01) than controls with results fromthe treated patients falling in between.IgG2, 3 and 4 antibodies to gliadin andcasein could only be detected in a minorityof subjects; the results for these subclasseswere not significantly different for thethree subject groups.The majority of circulating IgG gliadin

and casein antibody in untreated coeliacpatients is IgGl. This suggests that circulat-ing complement fixing gliadin antibodiesare not involved in the disease mechanismbut does not exclude a pathogenetic rolefor locally produced specific antibody. Thepresence of raised circulating IgGl anti-bodies to both gliadin and casein in un-treated coeliac patients implies that these

antibodies are because of dietary antigenabsorption.

P42Postprandial gut hormone profile afterintestinal glycosidase inhibition

R H TAYLOR, H M BARKER, E A BOWEY, J E

CANFIELD, AND K D BUCHANAN (Depart-ment of Gastroenterology and Nutrition,Central Middlesex Hospital, London, andDepartment of Metabolic Medicine,Queen's University, Belfast) Postprandialgut hormone profile is determined by mealcomposition, volume and other factors.The purpose of this study was to measurechanges in this response to a standard mealgiven with intestinal glucosidase inhibitors.

Six healthy subjects took a standard testbreakfast three times with either placebo,BAY m1099 50 mg or BAY o1248 20 mg inrandom order. These are potent, reversibleglycosidase inhibitors of different substratespecificity. Blood samples were taken for 3hours for measurement of glucose, insulin,GIP, N-terminal glucagon-like immuno-reactivity (N-GLI), C-GLI and gastrin.Breath hydrogen was measured as anindirect index of carbohydrate malabsorp-tion compared with lactulose 25 g alone.Both inhibitors reduced the postprandial

glucose peak (placebo 7-2±0-5 mmol/l;m1099 5-3±0-1 (p<001; o1248 5-9±0 2(p<0.05)) and the serum insulin peak(placebo 89±13 mU/l; m1O99 39±3(p<0-01); o1248 37±3 (p<0.01). Insulinrelease was reduced significantly from 30-120 min (p<005). GIP release was re-duced significantly between 30 and 150 minand peaks .fell from 995±395 ng/l to260±50 (m1099) and 305±60 (ol248). N-GLI rose from 60 min (m1099) and 90 min(o1248) to 134±7 ng/l (placebo, 222±31(m1099, p<O-05) and 190± 19 (ol248,p<O0.5) at 180 min. Gastrin levels rose inall groups and C-GLI did not change.Breath hydrogen indicated carbohydratemalabsorption of 10±4 g (m1099) and21±9 g (o1248).We conclude that glycosidase inhibition

slows digestion, reducing insulin and GIPrelease, but unabsorbed nutrients stimulateN-GLI distally. Slow absorption does notaffect gastrin or C-GLI release. Theseexperimental changes simulate those foundin digestive impairment due to enzymic ormucosal abnormality.

P43Regulation of ileal Na'-dependent bile acidtransport in man

A1113


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/

The British Society of Gastroenterologv

F W M DE ROOIJ, J W 0 VAN DEN BERG, A J P

VAN TILBURG, AND M VAN BLANKENSTEIN

(Department of Internal Medicine II, Eras-mus University, Rotterdam, Rotterdam,The Netherlands) We are interested in theregulation of ileal Na+-dependent bile acidtransport (INBAT) in gastrointestinaldiseases. In 177 patients undergoingcolonoscopy, ileal biopsies were takenfrom which brush border membrane ves-ciles (BBMV) were prepared. INBAT wasmeasured in vitro as the uptake of 3H-taurocholate (4 uM) into BBMV in thepresence of a 100 mM NaCl gradient.Uptake was followed during one minute.INBAT was quantified in pmoles tauro-cholate/20 sec/mg membrane protein.

Patients were placed in 10 diagnosticcategories. Mean INBAT values in sixcategories did not differ significantly, threecategories had a significantly decreasedINBAT, ileal disease (n=11, 19.4±7.4),ileal resection (n=5, 17-4±9.4), and consti-pation (n=8, 37.3±6.0). Increased INBATwas found in patients (n=8) with bile acidlosing diarrhoea with an endoscopicallyand histologically normal ileum(123.5±17.3).Our results show that in patients with

increased faecal bile acid loss (FBAL)INBAT is high, while in constipatedpatients, presumed to have a low FBAL,INBAT is low. This suggests that INBAT isregulated by the availability of bile acids tothe ileal enterocytes.

P44Improvement of abnormal lactulose/rham-nose permeability in active Crohn's diseaseof the small bowel by an elemental diet

I R SANDERSON, P BOULTON, I MENZIES, ANDJ A WALKER-SMITH (Department of ChildHealth, St Bartholomew's Hospital, Insti-tute of Child Health, St Thomas' Hospital,London) Elemental diet is as effective assteroids in the induction of remission ofsmall bowel Crohn's disease in children, as

judged by disease activity, ESR, CRP andalbumin. Intestinal sugar permeability is anobjective marker of small bowel functionwhich can assess the efficacy of elementaldiet during treatment. Fourteen childrenaged 11-17 years with active small boweldisease (proven by radiology and ilealhistology) were given an elemental diet forsix weeks. All had the following sugarpermeability study undertaken before andafter treatment. An 80 ml solution contain-ing 5-25 g lactulose (L) and 0-75 g rham-nose (R) was ingested after an overnight

fast. L/R ratio was calculated from thepercentage recovery of each of the sugarsin a 5 hour urine collection. Seven childrenwith no disease of the small intestine actedas controls. They had L/R ratios below0*08, the upper limits of normal found byBeach et al. All 14 children with Crohn'sdisease had an abnormally raised permea-bility ratio (0.256±0.037) before treatmentwhich fell significantly (p<0.001, pairedStudent's t test) after elemental diet(0.93±0-012). This coincided with markedclinical improvement assessment by diseaseactivity index. An elemental diet producesmarked reduction in the severity of smallbowel Crohn's disease as measured objec-tively by sugar permeability.

P45Does glycine content affect the efficacy ofamino acid solutions in TPN?

R G REES, G K GRIMBLE, P FROST, F AL-UBAIDI,AND D B A SILK (Departments of Gastro-enterology and Nutrition and ChemicalPathology, Central Middlesex Hospital,London) Despite the fact that the value ofglycine as a source of nitrogen for proteinsynthesis is disputed, it often represents asignificant proportion of non-essentialnitrogen in amino acid solutions for TPN.If glycine N is not retained, but channelledto urea, lower levels of N balance andplasma proteins might be expected whenglycine-rich solutions are used.We have prospectively compared two

solutions of differing glycine content in aheterogeneous group of seven patients(24-63 years) who required TPN. All weremetabolically stable without significantorgan failure and received two consecutivefive-day feeding regimes in random order.Amino acids were supplied as Vamin N or'old' formula Synthamin providing 141 gand 14-3 g N with 2.7% and 23% respec-tively, contributed by glycine. The regimeswere isocaloric and provided 2200 Kcal/dayfrom a 45%/55% energy lipid/glucosemixture.Twenty four hour total urinary N and

urea and serum urea were measured daily.Plasma albumin, prealbumin and transfer-rin were measured on day five of eachregime.There were no differences between

matched pairs for any parameter. Valuesfor Vamin and Synthamin (mean±SEM)were N balance (cumulative) -4-0±9 vs-14-8±13 g; urinary urea N/total N80±2% vs 82±2%; alb 26±2 vs 26±2 g/l;prealb 122±17 vs 107±19 mg/I; transferrin

2-2±0-1 vs 2-1±0 2 g/l; BUN 12 2+1 6 iss14-6±2-8 mmol/l.We conclude, from these data, that there

is no nutritional disadvantage for parenter-ally fed patients when significant amounts(up to 23%) of total amino acid N isprovided as glycine.

P46Intestinal absorption and laxative thresholdof lactitol - a new hydrogenated derivativeof lactose

D H PATIL, G K GRIMBLEL AND 1) B A Sll K

(Department of Gastroenterology atnd Nu-trition, Central Middlesex Hospital. Loti-don) Lactitol is a high soluble disaccha-ride with excellent taste properties. Asanimal studies suggest that it is poorlx,absorbed, it has potential as at bulksweetener of a low calorific value in thefood industry. This potential has beeninvestigated by characterising its assimila-tion by the human small intestine and hbdetermining its laxative threshold.

In vivo jejunal perfusion experimentswere carried out in normal human subjects(n=6). Intestinal uptake from isotonicsolutions containing 10. 30, 60 and 10()mmol lactitol/l was not significantly differ-ent from zero.To determine laxative thresholds, 21

healthy volunteers entered a single blindrandomised cross over trial, taking. individed doses, increasing amounts (10 gday) of lactitol, sorbital or placebo. I'helaxative thresholds of lactitol (74-7±SEM.6-3 g/day) and sorbitol (71 9+4-9) wercsimilar and the incidence of gastrointestinalside effects were not significantlv differenton similar doses of sorbital and lactitol.We conclude that as lactitol is notabsorbed in the small intestine and has alaxative threshold of >40 g/day, its poten-tial as a low calorie bulk sweetener in thefood industry is confirmed. At high doses(>70 g/day) Lactitol could find an impor-tant place as an osmotic cathartic agent.

P47Low phytate wheat bran inhibits zincabsorption less than standard bran

M J HALL, D ENE, D FARAH, AN[) 1. I)OWNS(University Department of Medicine, Bris-tol Royal Infirmary, Bristol) Binding otfzinc in the intestine by the phytate and/orfibre content of cereal products can lead tozinc deficiency which has been Increasingly

A1114


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


reported in gastrointestinal disorders. Wehave compared the effect of a new wheatbran, Testa Triticum Tricum, containing80% fibre but low in phytate, with phytate-rich standard bran on zinc absorption inhealthy fasting volunteers.Twenty subjects in two equal groups

each took zinc sulphate 220 mg. Plasmazinc was measured beforehand and up tosix hours afterwards by atomic absorptionspectrophotometry. The experiment wasrepeated with group 1 taking, in addition,17 g standard wheat bran and group 2 asimilar quantity of low phytate bran. Acontrol group of six subjects took RiceKrispies instead of bran.

Zinc absorption expressed as area underthe plasma zinc time curve (AUC) fell from(mean±SEM) 99-6±8-8 to -4*6±2-1(p<0.001) after standard bran, from74-6±7-4 to 18-5±4-2 (p<0.001) after low-phytate bran and from 105-9±10-7 to47-3±10-1 (p<0-05) after Rice Krispies.Mean percentage reduction in AUC afterstandard wheat bran was significantlyhigher than after low-phytate bran(104.9±1*9 vs 75*9±4 8%, p<0-001).We conclude that wheat bran is a potent

inhibitor of zinc absorption which can beimproved by reduction of the phytatecontent.

P48Effect of somatostatin analogue SMS 201-995 on fluid and electrolyte transport in apatient with secretory diarrhoea

C A EDWARDS, P A CANN, N W READ, AND C DHOLDSWORTH (Gastrointestinal Unit,Royal Hallamshire Hospital, Sheffield)SMS 201-995, a long acting analogue ofsomatostatin can be administered sub-cutaneously and has been shown to be ofbenefit in the treatment of diarrhoeasecondary to tumours secreting vasoactiveintestinal polypeptide (VIP). We usedsteady state perfusion techniques to inves-tigate its effect on fluid and electrolytetransport in the jejunum and ileum of apatient with a VIPoma. Fifty microgramsof the analogue was given subcutaneouslyfollowing a 60 minute control period andmeasurements repeated after a 40 minutesequilibration. In the jejunum, endogenousflow was abolished (110.4-0 ml/h), netfluid absorption increased (4.2-73-8 ml/25cm/h), sodium secretion reversed (-1-2-+10-1 mmol/25 cm/h), potassium absorp-tion increased (0.14-0.49 mmol/25 cm/h)and chloride absorption increased (0.5-6-56 mmol/25 cm/h). Ileum: endogenous

flow (488*4-283 8), net fluid absorption(24.5-86.3), sodium (1.6-12.2), potassium(0-1-0.3) and chloride (3-3-19-3). PlasmaVIP fell from 168 pmol/l (control) to 99pmol/l (end of test period). Clinically thepatient was able to return to a normal dietand leave hospital on a regime of 50 ,ug tds.He maintained a stool output of less than 1litre daily during treatment, enabling himto live normally until the eventual resectionof the tumour four months later.

P49Intestinal transit times and stool outputduring intake of a hypocholesterolaemicdose of guar in man

R PENAGINI, P VELIO, AND P A BIANCHI(INTRODUCED BY J J MISIEWICZ) (Cattedra diPatologia Medica III, Istituto di ClinicaMedica I, University of Milan, Italy) Guargum is generally believed to prolongmouth-to-caecum transit time (MCTT) andthis has been suggested to play a role in itsmetabolic effects. Previous studies ofMCTT used liquid meals and large doses ofguar. Whole gut transit time (WGTT) andstool output are reported to be unaffectedby guar, but data are scanty.

Six healthy male volunteers, aged 21-28years, ate a controlled diet (2721 Kcal and22 g dietary fibre daily) for two periods oftwo weeks (CD1 and 2) with a two-weekinterval on unrestricted diet; during CD2guar gum 5-7 gbd was given daily incrispbread during the two main meals.Daily stool weight and WGTT (measuredwith radio-opaque markers given on threeconsecutive days) were determined in thesecond week of CD1 and CD2 and totalserum cholesterol at the beginning and theend of each CD period. After both CDperiods, MCTT of a solid test meal (554Kcal) was measured using the hydrogenbreath test; guar gum 5 7 g in crispbreadwas added to the second test meal. Resultswere analysed using the paired Wilcoxon'stest and expressed as mean±SD. Serumcholesterol changes were -5 5 mg/dl±14-6and -30±17*6 (p=0-05) in CD1 and CD2respectively. MCTT (276-7 min±112-7 vs266.7±128), WGTT (46.2 h±8-5 vs53.8±17-6) and daily stool output(67-9±15-4 vs 76-2±35-2) showed no signi-ficant variations.We conclude that guar exerts its

hypocholesterolaemic effect without anysignificant change in MCTT and does notaffect WGTT and stool output in healthyhumans.

P50Bile salt uptake by Giardia lamblia: possiblerole in fat malabsorption

C E W HALLIDAY, P M G INGE, J WEBB, ANDM J G FARTHING (Department of Gastro-enterology, St Bartholomew's Hospital,London) Our preliminary observationsindicate that Giardia lamblia (GL) tropho-zoites take up bile salts (BS) in vitro.Although the biological implications forthe parasite are as yet unknown, we haveattempted to characterise this uptake pro-cess and estimate its potential impact onintraluminal BS concentration and on theBS pool. Giardia lamblia trophozoiteswere cultivated axenically and BS uptakeinvestigated with respect to time (0-4 h),temperature (4 and 37°C), trophozoitefixation (1% glutaraldehyde), parasitegrowth phase (active or stationary) and BSconcentration (glycocholic (GC) acid, 0-1-10 mmol/l). Glycocholic uptake was deter-mined using l4[C]-GC and expressed perunit number of organisms. Glycocholicuptake plateaued at one hour, was inhi-bited -50% at 4°C and almost completelyabolished by glutaraldehyde fixation. Up-take was greater during stationary growthphase (50-6 nmol/108 trophs/h from 2mmol/l GC) than during active multiplica-tion (2 mmol/l uptake - 29-4±nmol/108trophs/h). Uptake was concentrationdependent and appeared to follow satura-tion kinetics (Km = 0 45 mmol/l; Vmax =

0-55 mmol/min/108 trophs) compatible withan active transport process. This Km forGC is similar to published data for intacthuman and rat ileum. Infection with 1012organisms would result in the consumptionof 1 mmol GC per day which represents-20% of the normal adult BS pool.Chronic infection may lead to reduction ofintraluminal BS concentration which mightcontribute to fat malabsorption ingiardiasis.

P51Trypsin-sensitive surface ligand mediatesattachment of Giardia lamblia (GL) to ratenterocytes

P M G INGE, A D PHILLIPS, AND M J G FARTHING(Department of Gastroenterology, StBartholomew's Hospital, London) Wehave shown previously with an erythrocytemodel of attachment that GL has surfacelectin-like activity which may be importantin mediating parasite-enterocyte inter-action. We have now further characterisedthis ligand and its intestinal receptor using

A1115


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/

The British Society of Gastroe tacrologYisolated rat enterocytes. Cultivated GLtrophozoites and citrate-EDTA eluted ratenterocytes (-.105 cells of each) wereincubated 1:1 in microplates and parasiteattachment was quantified by phase micro-scopy. Attachment was maximal after onehour at 37°C, was markedly inhibited at4°C but remained unaltered in the absenceof divalent cations. Relatively high concen-trations of D-mannose (0-1 M) and themannose-rich glycoprotein mannan (10 mg/ml) inhibited attachment by 49+±10% and48±10% respectively, whereas othersugars were without effect. Subagglutinat-ing concentration (1:256) of heat inacti-vated, non-cytotoxic rabbit polyclonal anti-giardia antiserum with surface specificityalso inhibited GL attachment (42±9-0%inhibition), confirming the importance ofsurface determinants for attachment. Tryp-sin pretreatment (1 mg/ml, 30 min) of GLsignificantly reduced attachment(10-6±1.8%) trophozoites attached vs19-4±1 8% control, p<0-001), althoughviability was unchanged. However, shortterm (4 h) recultivation of trypsin-treatedGL normalised their ability to attach toenterocytes. We conclude that attachmentof GL to rat enterocytes is mediated by atrypsin sensitive, lectin-like surface ligandwhose intestinal receptor may include D-mannosyl residues.

GASTRODUODENALP52-68

P52Effects of single nocturnal doses of pirenze-pine on overnight gastric secretion

C W HOWDEN, D W BURGET, C SILLETrI, MDICKEY, AND R H HUNT (Department ofGastroenterology, McMaster University,Hamilton, Ontario, Canada) Treatmentof DU has focused on single nocturnaldosing with H2 antagonists but pirenzepine(Pir) has not previously been evaluated inthis manner.

In a double-blind randomised study, sixhealthy men were given a three day courseof placebo (Pla), Pir 100 mg nocte and Pir150 mg nocte. On day three, nocturnalhydrogen ion activity [H+], and acid output(AO) were measured. Inhibition by Pir wasassessed by percentage reduction of mean[H+], and the area under the H+ activity/time curve (AUC) (linear trapezoidalrule).

Mean [H+] was 36-1±7-9 mmol/l on Pla,16-7±8-6 on Pir 100 mg (-54%, p<0.01)and 16-6±6-0 on Pir 150 mg nocte (-53%,p<0-01). Acid output was reduced by 67%by both doses (p<0-001). Volume wasreduced by 47% and 52% by Pir 100 and150 mg respectively. Pepsin output was notaltered. AUC was 251-9±55-2 mmol H+h/l on Pla, 115-8±59-7 on Pir 100 mg(-54%, p<0.05) and 110-7±38-6 on Pir150 mg (-56%, p<0.05). Side effects ofdry mouth occurred only with Pir 150 mg.

Pirenzepine given as a single night timedose significantly reduced nocturnal acidsecretion. There was no significant differ-ence between the doses studied. As sideeffects were frequent with Pir 150 mg, theoptimal night time dose is 100 mg. Thisdose should be evaluated in a controlledclinical trial.

P53Efficacy of H2 receptor antagonists (H2RA)in gastric ulcer (GU) is independent ofantisecretory effect

C W HOWDEN, D B JONES, AND R H HUNT(Department of Gastroenterology, McMas-ter University, Hamilton, Ontario, Canada)Although patients with GU usually havenormal or reduced gastric acidity, they areoften treated with antisecretory drugs suchas H2RA. This study examines whether arelationship exists between GU healingand suppression of 24 h or nocturnalacidity. We analysed the results of over 40controlled clinical trials of H2RA in GU.The highest healing rates were obtainedwith 10 to 12 weeks' treatment with stan-dard doses of H2RA (88.9 to 93.8%).Placebo healing rates after four, six andeight weeks were 36-1% (n=208), 49 2%(n=65) and 61 5% (n=65), suggestingnatural healing of GU with time. From ananalysis of 11 studies of 24 hour andnocturnal intragastric acidity with variousdose schedules, therer was found to be nocorrelation between healing after four, sixor eight weeks and suppression of 24 houror nocturnal acidity. At six weeks, healingrates versus suppression of 24 hour acidityyielded r=-0-3908; p=0-553. With noctur-nal acidity, r=-0-3397: p=0514.There is no relationship between GU

healing and suppression of 24 hour ornocturnal acidity. Prolonging treatmentresults in maximal healing rates. Placebohealing also improves with time; the re-sponse to H2RA may simply reflect anaccelerated natural healing rate. As effica-cy in healing is not related to acid suppres-

sion, H2RA may heal (Gt thirouh somcother mechanism.

P54Are there immunological forms of duodenalulcer (DU) as a consequence of gastricparietal cell stimulating antibodies (PU'S-Ab)?

F DE LAZZARI, R MIRAKIAN, ( ViN1iRi 1BORTOLAMI, R NACCARA-RIo0, D D)Ni1V Hi. ANI)G F BOTrAZZO (Gastrc('litcrolo(4 L)Dcp rt-ment, Padova Universitu Itillod, ltmmuno-logy. Department, MVliddlcs t - oH /pitaiMedical School, Loncdoni) PreCs ios -Cports indicated that Ig(J obtailned tIromn1some patients with DIl stirmiulate alcidisecretion when injected into rats hv inaction on the histamine rcceptots ( I-R)on gastric parietal cells (PC ec in esti-gated this stimulatory actls its dircct1s h.measuring cAMP responses in PC enricleidcultures from guinea pig gastric mucosaAmmonium sulphate precipitIted I gs ftrom30 DU patients and 2() healthy controlShave been studied. The DU! cases wercselected on the basis of pcntagast in-stimulated (MAO) hyperacidity or hiohserum pepsinogen-I levels. alnd resplonse totreatment (cimetidine or ranitidine). 'IlhcPC suspensions were ohtained K% atcpcollagenase digestioll (.I) mc mill,Worthington type IV) of selectise hodsmucosa to avoid contamincation w ith antralendocrine cells. After ses cirl ws ashes andcentrifugations. the enriched P( prepai-ations (45-55%) were cultured os crnithito allow their full recovers and incuatcedfor four hours with 2 and 4 mrre m-l of li withadded 2 mM 3-isobuthxl\--methsl -xaintinic(IBMX). cAMP was meascired hv RIA\(Amersham kits). Control cultcires con-tained IBMX only, positiseV negatisc andcontrols Ig were included in esers hatch oftests and all samples were tested in clupli-cate. 10-3 M histamine producedimaximumcAMP stimulation at 3() min. Pelntagastrinand G1-17 had no effect. Igs trom 13 ot the30 DU cases produced a significant closerelated increase in total cAMP indiciatimntthe presence of PC stimulating antibodiels(PCS-Ab) similar to the well known th-roid stimulating antibodies of (Oras es clis-ease (TS-Ab). 8/13 stimulating-antihocl-positive DU patients were non responidersto anti-H2-R drugs as compared swith 3 17of the negative cases. Pepsinogen-I serumlevels were over 100 ng,'ml in I() 13 positisCcases and in 6/17 of negatise cases. Fhescresults suggest that in some paticnts, parti-cularly those resistant to cimctidinc

A1116


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


hypersecretors, duodenal ulcer diseasemay result from an autoimmune stimula-tory receptor antibody, acting selectivelyon the histamine-2-receptor.

P55Peptic ulcer in Bangladesh - an endoscopicsurvey

M HASSAN, SHAH MD K ALI AND A K AZADKHAN (Department of Gastroenterology,Institute of Postgraduate Medicine and Re-search, and Department of Epidemiology,Bangladesh Institute of Research andRehabilitation in Diabetes, Endocrine andMetabolic Disorders, Dhaka) A survey onpeptic ulcer was carried out in a ruralcommunity in Bangladesh. All subjects of15 years or above (n=2675) were surveyedby a questionnaire and those with ulcerdyspepsia (n=1106) were selected on thebasis of predetermined criteria. A randomsample (n=313) of these subjects werechosen for endoscopy and 283 of them(89 5%) have been endoscoped.Duodenal ulcer was found in 48,

duodenal erosions in nine, deformed bulbin 23 and gastroenterostomy in two sub-jects. The prevalence of duodenal ulcerdisease was 11 9%. Gastric ulcer was foundin 24 (prepyloric 14, antral eight, bodytwo) subjects, the prevalence rate being3-5%. Both duodenal ulcer and gastriculcer subjects were predominantly males.No case of carcinoma of the upper GI tractwas detected.Duodenal ulcer is common but gastric

ulcer is uncommon in this population.Reasons for the high prevalence ofduodenal ulcer remain unknown. Thosewith and without ulcer have been identifiedin a defined population and further studieson gene'tic and environmental factors cannow be undertaken.

P56Pathogenic role of campylobacter-likeorganisms in duodenal ulcer

W M HUI, S K LAM, P Y CHAU, J HO, M T NG, C L

LAI, AND A LOK (Departments of Medicine,Pathology and Microbiology, University ofHong Kong, Queen Mary Hospital, HongKong) Campylobacter like organismshave been associated with antral gastritisand duodenal ulcer. To study the pathoge-nic role of these organisms in duodenalulcer, endoscopic biopsies, two from thefirst part of duodenum, four from antrum,and four from body and fundus were taken

in random order before and after fourweeks of cimetidine treatment (1-2 g/day)from 49 patients with active duodenal ulcerthat healed at the end of the treatmentperiod. The biopsies were examined for thepresence and severity of any inflammationhistologically, and for campylobacter-likeorganisms by smear and culture. Beforetreatment, inflammation was present in71%, 98%, and 25%, while the organismswere present in 37%, 89%, and 82% of theduodenal, antral and fundal biopsies re-spectively. With complete healing ofduodenal ulcer, inflammation was presentin 64%, 100%, and 27%, while the organ-isms were positive in 39%, 83%, and 81%of the respective mucosae. With ulcerhealing, duodenitis and antral gastritis be-came milder and significantly so for antralgastritis. These findings indicate that heal-ing of duodenal ulcer is not influenced bythe presence of campylobacter-like organ-isms, which are frequently found in thegastroduodenal mucosa of patients withduodenal ulcer, but do not appear to beassociated with mucosal inflammation ex-cept in the antrum.

P57Neutral micro-climate lines human gastro-duodenal mucosa in vivo

E M M QUIGLEY AND L A TURNBERG(Department of Medicine, Hope Hospital(University of Manchester School of Medi-cine, Eccles Old Road, Salford) Gastro-duodenal mucosa may protect itself bymaintaining a neutral zone in the mucouslayer. The demonstration of a neutral pHon the mucosa in vitro and in vivo inanimals lends credence to this hypothesisand we now present evidence for a similarphenomenon in vivo in man. In 21 normalsubjects (aged 22-86 years) pH was meas-ured using a flexible pH electrode (OD1*35 mm) passed through the biopsy chan-nel of an endoscope. At six sites (loweroesophagus, fundus, body, antrum,duodenal cap and loop) pH was recordedwith the electrode (a) in the lumen and (b)touching the mucosa. In nine subjectsduodenal recordings were repeated duringluminal acidification. In five subjects elec-trical potential difference (PD) readingswere taken at the same sites. Lumen (L) tomucosal (M) pH gradients were identifiedin oesophagus (L 3-29±0-32 vs M4-18±0-25; p<0006), gastric fundus (L2-01±0-17 vs M 4-84±0-37; p<0003),body (L 1*82±0*12 vs 5*50±0*15; p<0001)and antrum (L 3-52±0*34 vs M 5-42±0-29;

p<0004). In duodenum, luminal andmucosal pH were near neutral but onluminal acidification a gradient was de-tected in both cap (L 2-57±0-15 vs M6-74±0-13; p<0.005) and loop (L2-44±0- 14 vs 6-39±0*20; p<0.0001).Potential difference (mV) in the lumen andat the mucosa were similar in oesophagus(L -15±3 vs M -15±4), body (L -23±3vs M -24±4), antrum (L -19±5 vs M-22±6) and duodenum (L -5±2 vs M-5±2) [NB: 1 pH unit=61 mV]. Weconclude that lumen-to-mucosal pH gra-dient is present in the human upper gut andsupports the importance of surface factorsin mucosal protection.

P58Luminal and mucosal pH in patients withduodenal ulcer, reflux oesophagitis andantral gastritis

E M M QUIGLEY AND L A TURNBERG(Department of Medicine, Hope Hospital(University of Manchester School of Medi-cine, Salford). We sought evidence for adefect in the 'mucus-bicarbonate' barrier inupper GI diseases by recording juxta-mucosal pH with a flexible pH electrode(ED 1-35 mm) passed through the biopsychannel of a standard endoscope. Luminaland mucosal pH were measured in loweroesophagus, fundus, body, antrum,duodenal cap and loop and repeated induodenum and oesophagus during intra-luminal acidification in patients with un-treated duodenal ulceration (DU) (n=5),reflux oesophagitis (RO) (n=7) and antralgastritis (AG) (n=7) and compared withvalues in 21 normal subjects. Duodenalulcer patients exhibited lower luminal pHin antrum (mean±SE, DU vs normals:1-88±0-37 vs 3-52±0-34, p<0-03), duoden-al cap (5-28±1.43 vs 6-89+0-20, p<0-05)and loop (4-67±1.81 vs 6*84±0*19,p<O004). Mucosal pH was lower in gastricbody (3*52±0-76 vs 5-50+0-15, p<0.005)in DU, but was similar to normals at allother sites. In RO luminal pH in antrumwas lower (2.27±0-29 vs 3-52+0*34,p<0.04) while in Ag luminal pH in gastricbody was raised (2-54±0.49 vs 1-82±0*12,p<0-05). Luminal pH at other sites andmucosal pH at all sites for RO and AGwere similar to normals.We conclude that in DU patients,

though antrum and duodenum are exposedto more acid the juxtamucosal neutral zoneis preserved; mucosal pH in gastric body,in contrast is depressed. No deficiencies inthe ability to maintain juxtamucosal neut-

A1117


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


rality could be identified in patients withreflux oesophagitis or antral gastritis.

P59Isolated pyloric contractions (IPC) in fastedand fed human subjects

N W READ, L A HOUGHTON, R HEDDLE, G JMADDERN, J DENT, J DOWNTON, J B WYMAN,AND T TOOULI (Departments of Medicineand Surgery, Flinders Medical Centre, andRoyal Adelaide Hospital, Adelaide, Austra-lia) It is controversial whether phasicpyloric contractions occur in humans inde-pendently from antral and duodenal con-tractions. We have recorded pyloric press-ures in nine healthy subjects with a 4-5 cmlong sleeve sensor, positioned by measure-ments of transmucosal potential difference.Pressures were also measured with per-fused side holes at four sites in the duode-num and three sites in the antrum.Under fasting conditions five subjects

exhibited sequences of between 18 and 89regular pyloric contractions (frequency, 2-9to 3-3 min) which were not associated withcontraction in the antrum or duodenum 1cm from either end of the sleeve. IPCsoccurred immediately before the onset ofphase III of the MMC in four subjects andwithin 15 minutes of the end of phase III infour subjects.

Ingestion of 300 ml chocolate milk eli-cited IPCs in all subjects (19 to 153contractions occurring between 2-5 and 2-9min) postprandial IPCs occupied 25±8%(SEM) of recording time compared with7±4% during fasting. Postprandial IPCswere also more likely to be interspersedwith episodic peristaltic waves whichswept from the antrum into and along theduodenum. Isolated pyloric contractionsmay play a role in the control of transpylor-ic flow, especially in the fed state.

P60New evidence for the pathophysiology ofpostvagotomy diarrhoea

S A RAIMES, V SMIRNIOTIS, E J WHELDON, C WVENABLES, AND I D A JOHNSTON (UniversityDepartment of Surgery, Newcastle uponTyne) Diarrhoea is more common aftertruncal vagotomy than after other gastricoperations, but there is no satisfactoryexplanation for this. The hypertonic glu-cose 'dumping provocation test' repro-duces diarrhoea in those patients for whomthis is an important complaint. We haveinvestigated the malabsorption of glucose

during this test by measuring changes inbreath hydrogen. Four groups of un-selected male subjects were studied - 42after truncal vagotomy and pyloroplasty(TV+P), 14 after proximal gastric vago-tomy (PGV), 12 after Bilroth I gastrectomy(BG) and 10 healthy controls. Glucosemalabsorption was significantly more com-mon in the TV+P group (39/42 vs PGV:5/14; p<0-001, vs BG: 4/12; p<0-001 andvs controls: 0-10; p<0.001; Fisher exacttest). Small bowel colonisation by hyd-rogen producing bacteria was excluded byrepeating tests with lactulose. Twenty pa-tients experienced diarrhoea during thetest, all but one in the TV+P group. In theTV+P group those patients with diarrhoeahad a greater fall in plasma volume (me-dian 13% vs 8%; p=0-001; Mann-WhitneyTest) and a shorter mouth-caecum transittime (median 15 minutes vs 20 minutes;p<0048; Mann-Whitney Test).We conclude that patients after TV+P

do not fully absorb a hypertonic glucoseload. This is abnormal and helps to explainthe pathophysiology of postvagotomydiarrhoea.

P61Emergency peptic ulcer surgery - an associ-ation with NSAIDs

A J WALKER AND E P DEWAR (ProfessorialSurgical Unit, RN Hospital, Haslar, Gos-port, Hants) The elderly have a dispro-portionately high mortality from complica-tions of peptic ulceration. They may havemore concomitant diseases, be treated lessaggressively or be prescribed more ulcer-ogenic drugs, particularly NSAIDs.For three years all patients admitted with

an episode of upper GI tract haemorrhageor perforation of a PU who requiredemergency surgery were studied.

Sixty patients were admitted with 64ulcers. Thirty two (25 DU, seven GU)operations for haemorrhage and 32 (17DU, 15 GU) for perforation were per-formed.

Fifty per cent of the patients were takingNSAIDs, the incidence in women doublethan in men. Sixty seven per cent were over60 years old, women significantly olderthan men.

Perforation in association with NSAIDsin the over 60s was double that in the under60s.

In an aggressive surgical policy the mor-tality from haemorrhage was 3*4% andfrom perforation 9*7%. Contrary to surgic-al tradition only four partial gastrectomies

were performed. Of three deaths fromperforation, all were over 60 and all takingNSAIDs.A low incidence of serious complications

associated with NSAIDs is claimed. Thereis an associated serious morbidity, howev-er, requiring emergency surgery and amortality. Prescribers should be madeaware of the benefit/risk ratio.

P62Does cimetidine alter the prognosis afterperforated duodenal ulcer?

C J SIMPSON, G LAMONT, I MCDONALD, AND I SSMITH (Victoria Infirmary, Glasgow)Sixty consecutive patients with perforationof a duodenal ulcer undergoing emergencysimple closure were randomised to receivefull dose cimetidine for eight weeks, fol-lowed by maintenance of 400 mg nocte fora further 18 weeks, or, not to receive anyanti-ulcer therapy. Follow up on all pa-tients took place at one month, twomonths, six months and 12 months andendoscopy was undertaken on symptoma-tic patients. Both groups were comparablein age, pre and postoperative smokinghabits, prior treatment with ulcer healingagents, dyspeptic history and duration ofperforation before surgery. Althoughcimetidine did not affect immediate recov-ery, there being three (10%) deaths in thecimetidine group and five (16%) in thecontrol group, the cimetidine group didenjoy a significant benefit (p<0.001) in thelong term. Eleven (45%) patients in thecontrol group developed recurrent ulcersymptoms, three requiring emergencysurgery for bleeding or reperforation andeight requiring medical treatment. Thecimetidine group remained symptom freefor 12 months.

P63Effect of glucomannan on postprandialreactive hypoglycaemia after gastricsurgery

W P M HOPMAN, G M P HOUBEN, P A J SPETH,AND C B H W LAMERS (Departments ofGastroenterology-Hepatology, Universitiesof Nijmegen and Leiden, The Netherlands)Presently no satisfactory treatment is avail-able to patients who suffer from post-prandial reactive hypoglycaemia aftergastric surgery. A-glucoside hydrolase in-hibitors (acarbose) are poorly toleratedbecause of side effects due to carbohydratemalabsorption, while pectin is unpalatable

A1118


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


and effective only when ingested in verylarge doses. Glucomannan, a polysacchar-ide consisting of glucose and mannose, istasteless and has strong gel-forming prop-erties. In a double-blind study of eightpatients (21-61 years) with reactive hypo-glycaemia after gastric surgery we havecompared the effect of 5-2 g glucomannan,2-6 g glucomannan and 2-6 g placebo, and5-2 g placebo added to a normal carbohy-drate-rich breakfast on plasma glucose,plasma insulin and breath hydrogen excre-tion. Glucomannan had no significanteffect on the peak glucose increment(4-4±0-6 mmol/l with placebo, 4 2±0-5mmol/l with 2-6 g and 3-9±0-5 mmol/l with5-2 g of the fibre). On the other hand,glucomannan induced a dose dependentsignificant (p<0.05) inhibition of the post-prandial decrease in plasma glucose(2-2±0-3 mmol/l during placebo, 1-5±0-2mmol/l during 2-6 g and 0 7±0 9 mmol/lduring 5-2 g of the substance) and of thepeak insulin increments (154±19 mU/lduring placebo, 80±16 mU/l during 2-6 gand 75±21 mU/l during 5-2 g of glucoman-nan). Six of the patients had a breathhydrogen excretion of more than 15 ppmduring placebo, three after 2-6 g and oneafter 5-2 glucomannan.We conclude that glucomannan is highly

effective in preventing postprandial reac-tive hypoglycaemia in patients with pre-vious gastric surgery.

P64RP 40749 in the treatment of duodenal ulcerand its influence on serum gastrin, serumpepsinogen I and gastrin content of theantral mucosa

G F NELIS, C B H W LAMERS, AND G PALS(INTRODUCED BY J J MISIEWICZ) SophiaZiekenhuis, Zwolle, Radboud Ziekenhuis,Nijmegen, and Institute of Human Gene-tics, Free University, Amsterdam) Weevaluated the effectiveness of two doses ofa new H+-K+-ATPase antagonist (RP40749, pyridyl-2-tetrahydrothiophene) inthe treatment of 20 patients with duodenalulcer. Treatment was prescribed double-blind either as 100 mg or 200 mg as a singledaily dose for four weeks. Blood samplesand gastric biopsies were taken immedi-ately before treatment and at the last dayof treatment. There was a rapid clinicalimprovement and after four weeks allulcers were healed endoscopically, exceptfor one in the 100 mg group. Side effectsdid not occur, there were no changes in theroutine laboratory parameters and gastric

histology.We evaluated the influence of RP 40749

on basal and meal-stimulated serum gastrinand serum pepsinogen I and the gastrincontent of the antral mucosa. After treat-ment there was a significant rise in basalserum gastrin (55.6 vs 83.8) and pepsi-nogen I (59-0 vs 136-7), meal-stimulatedserum gastrin (96.0 vs 144.4) and antralgastrin (17-2 vs 27.1). There were no majordifferences between the 100 and 200 mggroup.

P65Effect of H2 receptor antagonists on pros-taglandin E2 and leukotriene B4 productionin duodenal ulcer disease

J P WALSH, F J BLOOMFIELD, W J MAXWELL,F P HOGAN, D KELLEHER, AND P W N KEELING(Department of Clinical Medicine, TCDMedical School, St James' Hospital, Dub-lin, Eire) Prostaglandins (Pg) may play arole in the prevention and treatment ofduodenal ulcer (DU) disease by antisecre-tory and cytoprotective mechanisms. Theeffects of pro-inflammatory leukotrienes(LT) have not yet been quantified. Thisstudy examines PgE2 and LT B4 produc-tion by peripheral blood monocytes (PBM)of DU patients, stimulated with opsonisedzymosan, before and after four weekstreatment with H2 antagonists.Twenty patients with active DU were

studied. Reduced PgE2 production wasseen in untreated DU compared withnormal control subjects (CS) (n=20,18.2±1.8 vs n=20, 40+8-0 ng/106 PBM,x±SE p<0-001, DU vs CS), with a riseafter DU treatment to 19-7±2-45 ng/106PBM. Leukotriene B4 production wasmarkedly raised in untreated DU com-pared to CS (n=20, 6-1±1 vs n=10,1-3±0-2 ng/106 PBM, p<0-001, DU vs CS)with a significant reduction to 3-94±0.60ng/106 PBM after DU treatment.Reduced PgE2 production in the un-

treated DU may result form diversion ofarachidonic acid substrate to an activatedlipoxygenase pathway with consequent en-hanced LT B4 production. LT B4 may beof pathogenic importance in the persist-ence of chronic inflammation. Reversal ofthese abnormalities by H2 antagonists sug-gests an additional mode of action for theseagents.

P66Effect of sucralfate on isolated amphibiangastroduodenal bicarbonate secretion

J R CRAMPTON, L C GIBBONS, AND W D W REES(Department of Medicine, Hope Hospital(University of Manchester School of Medi-cine), Salford) Sucralfate is a basic alumi-nium salt of sucrose sulphate which ex-hibits ulcer healing and cytoprotectiveproperties in man and experimentalanimals. There is evidence that its cyto-protective activity may be, in part, prosta-glandin mediated and gastric luminal pros-taglandin E2 release has been shown to bestimulated by the drug. As prostaglandinsof the E series have been shown to stimu-late gastroduodenal alkali secretion theeffect of sucralfate on the rate of bicarbon-ate secretion by stripped bullfrog (Ranacatesbeiana) antral, fundic and duodenalmucosa has been examined. An isolatedchamber preparation has been used ena-bling pH stat titration of the luminalsolution and recording of transmucosalpotential difference. Addition of sucralfate0 5 g/l at pH 7-4 to the mucosal sidesolution induced, within 15 minutes, anincrease in the rate of bicarbonate secre-tion by fundus (mean±SE: 183±87%,n=4, p<0-05) and antrum (mean±SE:156±58%, n=5, p<0-005). At this concen-tration there was no effect on duodenum(mean±SE: 4±15%, n=6, NS) but in ahigher dosage of 1 g/l there was an increasein alkalinisation (mean±SE: 42±15%,n=6, p<0-05). Transmucosal potential dif-ference was not altered in these studies.These results suggest that the cytoprotec-tive and anti-ulcer activity of sucralfatemay, in part, be mediated by an increase inmucosal bicarbonate secretion and enhan-cement of the mucus-bicarbonate barrier.

P67Measurement of increased pepsin degrad-ation of mucus by gastric juice in pepticulcer patients

D A HUTTON, J P PEARSON, A ALLEN, W JCUNLIFFE, C W VENABLES, AND R WARD(Departments of Physiological Sciences andSurgery, University Medical School, Fram-lingon Place, Newcastle upon Tyne) Theintegrity of the protective adherent gastricmucus gel barrier is a balance between itssecretion and its erosion by luminal pepsinand/or mechanical shear.We have developed: (1) a sensitive assay

for pepsin activity which measures newN-terminal groups from cleavage of pep-tide bonds. (2) An assay for mucolyticactivity of pepsin based on the fall inviscosity and the accompanying increase innew N-terminal groups when gastric mucus

A1119


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


is incubated with juice. (3) Quantitation ofthe different pepsin types in gastric juice byelectrophoretic separation on agar gels,elution and asssay of pepsin activity.At pH 2.0 pepsin 1 had twice the

mucolytic activity of pepsin 3. At pH 4-0pepsin 1 had six times more mucolyticactivity than pepsin 3 and caused substan-tial mucolysis up to pH 5*1. Gastric juicefrom duodenal ulcer patients exhibitedsubstantial mucolytic activity between pH 2to 5, similar to pepsin 1. Gastric juice fromnon-symptomatic volunteers exhibited lit-tle mucolytic activity above pH 4.These studies suggest increased peptic

degradation of the mucus barrier, associ-ated with higher pepsin 1 activity, could beimportant in the pathology of peptic ulcerpatients.

P68Treatment of duodenal ulcer by H2 block-ade, large single daily dose, how much andwhen?

M DEAKIN, H GLENNY, J K RAMAGE, JANE

MILLS, W BURLAND, S P GRAY, J BILLINGS,AND J G WILLIAMS (Department of Gastro-enterology, Royal Naval Hospital, Haslar.Smith Kline and French Research, Wel-wyn,. Herts.) The therapeutic efficacy ofa single night time dose of H2 antagonist isproven but treatment is usually given onretiring. During a 24-hour period thelongest period of unbuffered intragastricacidity covers the evening and night. It ispossible that more effective inhibitionwould be achieved by higher or earlierdosing.We have undertaken identical studies of

24-hour intragastric acidity, nocturnal acidand pepsin output in two groups ofvolunteers with duodenal ulcers in remis-sion, dosing at either 1800 (n=8) or 2300hours (n=10).Compared with placebo mean hourly

hydrogen ion activity during the eveningand overnight period (1800-0730) was de-creased by 51-4% (cimetidine 800 mg, 2300hours), 54% (C800 mg, 1800 hours), 54%(cimetidine 1600 mg, 2300 hours), 78%(C1600, 1800 hours) and 80% (ranitidine300 mg 1800 hours). Despite considerablereduction in overnight acid output bydosing at 1800 hours, decrease in pepsinoutput was less marked than following a2300 hours dose. After dosing at 1800hours pH readings ranged from 3-6, (mak-ing significant denaturation unlikely),whereas after 2300 hours dosing with cime-tidine 1600 mg all specimens were anacidic,

and denaturation of pepsin probably occur-red.While more effective overall control of

intragastric acidity is achieved by 1800hours dosing, we would not recommenddosing at this time because overnight acid-ity is sufficient to allow peptic activity toremain.

INFLAMMATORY BOWEL DISEASEP69-84

P69Serological studies in Crohn's disease

J P IBBOTSON, R N ALLAN, AND P PEASE(The Gastroenterology Unit, General Hos-pital, Steelhouse Lane, Birmingham) Ithas been suggested that cell-wall deficientforms of Pseudomonas maltophilia mightbe involved in the aetiology of inflamma-tory bowel disease. In addition, certainserotypes of Yersinia enterocolitica cause aselflimiting ileitis which resembles Crohn'sdisease. The aim of this study was tomeasure antibody levels to these organismsand to Klebsiella aerogenes, a commoninhabitant of the gut, in sera from inflam-matory bowel disease pEitients.Serum samples were obtained from 20

patients during exacerbation of Crohn'sdisease and from 20 age- and sex-matchedpatients with active ulcerative colitis and 20healthy controls. Antibody levels weremeasured using an enzyme linked immuno-sorbent assay.

In comparison with control groups, ahighly significant number of patients withCrohn's disease had raised antibody levelsto Y enterocolitica and K aerogenes. Levelsto P maltophilia were not significantlyraised. Patients with ileal Crohn's diseasehad significantly higher levels than those inother groups. There was no correlationbetween antibody levels and diseaseactivity.Although the raised antibody levels may

be due to leakage of normal gut floraacross a damaged mucosa, it could be thata range of organisms are primarily involvedin the aetiology of the disease.

P70Patient education in Crohn's disease

H L SMART AND J F MAYBERRY (University

Hospital, Nottingham) Three hundredand fifty patients with Crohn's disease wereoffered an information booklet about theircondition. This provided information onsymptoms, investigations and treatmenttogether with addresses of self help organ-isations. A copy was requested by 232patients and 175 of these completed aquestionnaire assessing its value. The ma-jority of patients (85%) who completed thequestionnaire found the booklet helpful.but required more information about com-plications of the disease, long term prog-nosis, cancer risk and inheritance of thecondition. Seventy eight per cent of thepatients felt that such a leaflet should begiven to all patients shortly after diagnosis.One year later a further survey was con-ducted to assess the effect this booklet hadon anxiety and consultation levels. Onehundred and sixty of the original group of175 patients who completed the question-naire were contacted and 78% responded.There was a significant reduction in anxie-ty; 30% of patients felt less anxious com-pared with 13% who were more anxious(X2=10 2 p<0.005) and this was associatedwith a reduction in consultation rates by17% of the patients compared with anincrease by 5% (X2=8 5 p<(0)05). Im-proved patient education may alter theclinical management of this disease in thefuture.

P71Expression of MHC Class I (HLA-A,B,C)and Class II (HLA-DR) antigens by colonicepithelium in human intestinal schistoso-miasis

S BADR EL-DIN, L K TREJDOSIEWICZ, J OAKES,R V HEATLEY, AND A ABOU-KHADR (Depart-ments of Medicine,, St James's UniversitVHospital, Leeds, and Alexandria Univer-sity, Egypt) The role of the expression otMHC antigens by epithelial cells has re-cently received considerable attention, andit has been shown that expression ofHLA-DR is increased in epithelial cells oforgans affected by some autoimmune dis-eases. In the bowel, HLA-DR expressionby enterocytes is known to increase inTrichinella spiralis infection and in inflam-matory bowel disease, although alterationof expression HLA-A,B,C has not beenreported.The expression of MHC antigens in

colonic epithelium was studied by im-munofluorescence with monoclonal anti-bodies on mucosal colonoscopy biopsies of13 patients with intestinal schistosomiasis

A1120


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


(six with schistosomal polyposis) and fivenormal Egyptian controls.Normal colonic epithelium was negative

for HLA-DR, but become HLA-DR+ inschistosomiasis patients, expression beingstrongest by glandular epithelium. HLA-DR+ epithelium was observed in colonicmucosae of schistosomal colitis and shisto-somal polyposis patients, as well as in thepolyps themselves. By contrast, expressionof HLA-A,B,C was high in the surfaceepithelium of controls, and only slightlyweaker in the glandular epithelium. Essen-tially, no differences in HLA-A,B,C ex-pression was observed in the patients,irrespective of whether polyposis was pre-sent.These data suggest that colonic epithe-

lium expresses Class II MHC antigens as aconsequence of parasitic egg deposition,which in turn may allow the HLA-DR+epithelial cells to act as antigen-presentingcells in the development and maintenanceof local anti-parasite immunity.

P72T-lymphocyte subsets in the colonic mucosaof patients with intestinal schistosomiasis

S BADR EL-DIN, L K TREJDOSIEWICZ, J OAKES,R V HEATLEY, G JANOSSY, AND A ABOU-KHADR(Departments of Medicine, St James's Uni-versity Hospital, Leeds, and AlexandriaUniversity, Egypt, and Academic Depart-ment of Immunology, Royal Free Hospital,London) The mechanisms of immuneresistance to schistosomiasis are not fullyunderstood, and it is not known why, inchronic intestinal schistosomiasis, about10% of patients develop multiple colonicpolyposis. Little is known of the role of Tcell subpopulations in the local immuneresponses to parasite eggs.T cell subsets were studied by double-

label immunofluorescence in cryostat sec-tions of colonoscopy mucosal biopsies us-ing a panel of monoclonal antibodies.Thirteen patients with schistosomiasis(seven with schistosomal colitis and sixwith schistosomal colonic polyposis) andfive normal Egyptian controls were stu-died, as were the actual polyps from twopatients. In schistosomiasis, there weremarked changes in the intra-epithelial Tcells: there was a significant increase(p<0-05) in all patients in the percentageof T4+ (helper/inducer) cells, the majorityof which co-expressed the T2 marker of Tlymphocyte stimulation. There was also anincreased tendency for the T8V (cytotoxic/suppressor) cells to co-express the Ti

'pan-T' marker (p<0-01 in schistosomalpolyposis). In the lamina propria, therewere no major differences between con-trols and schistosomal colitis patients,however, although in polyposis, the T4:T8ratio was significantly decreased (p<005),whereas paradoxically in the actual polyps,the T4:T8 ratio was greatly increased (7:1versus 2:1).These results show that there are consid-

erable alterations in immuno-regulatory Tlymphocyte subpopulations in the colonicmucosa in chronic intestinal schistoso-miasis. Polyp formation may be a functionof infiltrating T4+ cells, which migratefrom the lamina propria of the colonicmucosa.

P73Enhanced prostaglandin E3 production bycolonic epithelial cells and resident mac-rophages in inflammatory bowel disease

W J MAXWELL, F J BLOOMFIELD, F P HOGAN,J P WALSH, D KELLEHER AND P W N KEELING(Department of Clinical Medicine, TCDMedical School, St James's Hospital, Dub-lin, Eire) Prostaglandin (Pg) produced inthe colon may have important modulatoryfunctions in inflammatory bowel disease.PgE2 is produced by both colonic epithelialcells and resident macrophages and hasimportant cytoprotective and secretoryeffects. As recent evidence suggests thatPgs also act as immunomodulators in in-flammatory bowel disease (IBD), the aimsof this study were to assess the relativeproduction of prostaglandin E2 (PgE2) bycolonic epithelial cells and tissue fixedmacrophages, isolated from inflamed (In)and non-inflamed (NIn) tissue from pa-tients with IBD and challenged with opson-ised zymosan. As controls, cells isolatedfrom tumour free resection margins ofcolonic cancer (CC) were used.PgE2 production by In colonic epithelial

cells was enhanced relative to NIn cells(n=5, 7-8±1-8 vs n=5, 3-3±1-0 ng/106cells, x SE, In vs NIn p<002). Similarenhancement was seen by macrophagesfrom In mucosa in response to Zy (n=5,21-8±5-2 vs n=5, 8-9+1-9 ng/106 cells In vsNln, p<0.02). Control (CC) PgE2 produc-tion by epithelial cells was 9-1 ng/106 cellsand macrophages 14±1 ng/106 cells. PgE2production by stimulated macrophages inthis study were higher than those previous-ly reported for spontaneous secretion in-dicating an enhanced response to immunechallenge.

In conclusion, these data support the

concept that increased local production ofprostaglandins in areas of disease activitymay be of pathogenic significance in thechronic inflammatory response.

P74Inhibition of PgE2 secretion by Salazopyrinand prednisolone in normal and Crohn'sdisease monocytes

W J MAXWELL, F J BLOOMFIELD, F P HOGAN,J P WALSH, D KELLEHER, AND P W N KEELING(Department of Clinical Medicine, TCDMedical School, St James's Hospital, Dub-lin, Eire) Salazopyrin (SP) reduces intes-tinal prostaglandin (Pg) secretion inCrohn's disease (CD). Salazopyrin and itsactive moiety, 5 amino salicylic acid are,however, both weak cyclooxygenase in-hibitors. We studied the in vitro effects ofdirect addition and 24 hour preincubationwith SP and prednisolone (Pred), a phos-pholipase A2 (PLA2) inhibitor on PgE2secretion by peripheral blood mononuclearcells (PMBC) stimulated by opsonisedzymosan (Zy) or Zy+ 10 mMol arachidonicacid (AA) for 30 min at 37°. Secretion wascorrected for the number of esterase posi-tive PBMC.

Crohn's disease monocytes had signifi-cantly enhanced PgE2 secretion comparedwith normal subjects (NS) when stimulatedwith Zy (n=7, 70±15 vs n=7, 21±5 ng/106monocytes, x±SE, p<0-001, CD vs NS),and Zy+AA (n=7, 2343±577 vs n=7,322±123 ng/106 monocytes, p<0.001).Direct addition of SP resulted in a nonsignificant increase in PgE2 secretion byboth CD and NS monocytes using Zystimulation only. Pred added directly hadno effect on PgE2 secretion. Twenty fourhour preculture with Pred, however, re-sulted in marked reduction of PgE2 secre-tion in both CD and NS using Zy stimula-tion (78% and 69% reduction, CD and NSrespectively, partially reversible on addi-tion of AA (16% and 31%, CD and NS). Aless marked inhibition of PgE2 secretionwas seen following preculture withSP onZy stimulation (39% and 26%, CD andNS) which was reversible on addition ofAA (2% and 3%, CD and NS).Both Pred and SP reduced PG produc-

tion by 24 hour cultured monocytes. This isnot because of cyclooxygenase inhibition asit is reversible by addition of exogenousAA. Reduced PgE2 production by SP maybe because of PLA2 inhibition. Alterna-tively because mixed mononuclear cells areused, it may be due to an indirect effect onintercellular signalling.

A1121


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


P75Factors affecting suppression of cell medi-ated immunity in inflammatory bowel dis-ease

C AINLEY, J CASON, R A WOLSTENCROFT, B MSLAVIN, AND R P H THOMPSON (The Gastro-intestinal Laboratory, Department of Im-munology, and Department of ChemicalPathology, St Thomas's Hospital, London)There is abnormal suppression of cellmediated immunity (CMI) in ulcerativecolitis (UC) and Crohn's disease (CD), butthe factors involved are unknown. We haveinvestigated the effect of spontaneous sup-pressor cell activity (SSCA) and in-domethacin sensitive suppression (ISS) onlymphocyte transformation (LT) in rela-tion to lymphocyte populations (LP) andnutrition in 24 controls (CO), 14 patientswith UC and 31 patients with CD. Lym-phocyte populations were estimated usingOKT monoclonal antibodies. Lymphocytetransformation with phytohaemagglutinin(PHA) and Concanavalin A (Con A) werecarried out in parallel (1) control LT; (2) 24hour preincubation (SSCA assay); (3) with1 mcg/ml indomethacin (ISS assay).Seventeen CD patients were normallynourished (nCD) and 14 malnourished,being <90% of ideal body weight. Thepercentage OKT3 cells, and theOKT4:OKT8 ratio were normal in CD, butin UC the ratio was reduced (CO1-93±0-07 vs UC 1*12±019, p<0.001).Control LT responses were reduced inmalCD (PHA p<0-01, Con A <p 0.01),but not in nCD; in UC responses to Con Aonly were reduced (p<0-05). There weresmall increases in ISS in both UC and CD.In UC and nCD there were small reduc-tions in SSCA. In malCD SSCA wasincreased (PHA p<0-05, Con<A p 0.01),and was partially responsible for the re-duced control LT of malCD.

In UC, LT is reduced with an abnormalOKT4:OKT8 ratio, but SSCA and ISS arenormal. In nCD, LT, SSCA and ISS arenormal. In malCD, LT is reduced, in partdue to increased SSCA. Malnutritionunderlies abnormal CMI in CD.

P76Peripheral blood T cell subsets in sclerosingcholangitis (PSC) and ulcerative colitis (UC)

G K SACHDEV, R W G CHAPMAN, AND D PJEWELL (Gastroenterology Unit, RadcliffeInfirmary, Oxford) Previous studies of Tcell subsets in patients with ulcerativecolitis have produced conflicting results.

One explanation may be methodologicalbecause subsets have been evaluated usingmononuclear cell suspensions. T cell sub-sets have now been re-examined usingwhole blood smears.

Patients with UC (n=40, remission 27,active 13), and PSC with UC in remission(eight) were compared with normal healthycontrol subjects (15) and patients with theirritable bowel syndrome or peptic ulcers(17). Blood smears were air-dried, fixedwith acetone-methanol and stained withmonoclonal antibodies: Dako-TI (panT),Dako-T4 (helper), Dako-T8 (suppressor-cytotoxic, CR23/43 (Ia).

Patients with active UC showed a signifi-cantly (p<0.001) lower proportion of T8+cells (16%) than those in remission (25%),the healthy controls (27%) or diseasecontrols (24%). This difference was relatedto disease activity and not to length ofhistory or therapy. Patients with PSC,whose UC was in remission, had signifi-cantly fewer T8+ cells than healthy controls(19-2, p<0-02). There were no differencesbetween the groups for T1+, T4+ orCR23/43+ cells.We conclude that reduced proportions of

peripheral blood lymphocytes of suppres-sor-cytotoxic phenotype are associatedwith active UC and PSC regardless ofactivity of the UC.

P77Peanut lectin binding and dysplasia inmulticolonic biopsies from patients withulcerative colitis complicated by carcinoma

J B J FOZARD, S B GRIFFITHS, M F DIXON, A T RAXON, AND G R GILES (University Depart-ment of Pathology, Department of Surgery,St James's University Hospital, Leeds, andGastroenterology Unit, Leeds General In-firmary, Leeds) The demonstration ofdysplasia in the mucosa of patients withulcerative colitis (UC) is important incancer surveillance and may determine theneed for surgery. There are problems inthe interpretation of dysplasia especially inthe presence of inflammatory changes. In apreliminary study using an immunoperox-idase technique we assessed 11 lectins forthe identification of dysplasia in colectomyspecimens from patients with UC peanutlectin bound in all cases of dysplasia.We have now studied peanut lectin

binding and dysplasia in 165 colonoscopicbiopsies from patients with UC compli-cated by carcinoma (n=6) and cancer freecontrols (n=10). A significant increase inpeanut lectin binding was found in biopsies

from the cancer group 81/89 (83%) com-pared with controls 43/67 (64%) p=0-005x2. Comparable rates of dysplasia occurredin the cancer (32/98, 32%) and controlgroups (20/67, 30%. High grade dysplasiawas absent, however, in the control groupand universally present in the cancergroup.

Peanut lectin binding is a sensitive in-dicator of cancer complicating UC, butlacks specificity. High grade dysplasia indi-cates the presence of carcinoma.

P78Detecting premalignancy in the colon

J MAiTHEWS, T COOKE (INTRODUCED BY A

PARKINS) (Department of Surgery,Charing Cross and Westminster MedicalSchool, London) We have previouslyreported that in an animal carcinogenesismodel the mean DNA content per epi-thelial cell in the upper regions of thecolonic crypts increases as carcinogenesisprogresses. We have now studied the DNAcontent of colonic mucosa in patients withcolorectal carcinomas.

Using microdensitometry, DNA contentwas measured in the cells in the prolifera-tive and functional zones of histologicallynormal Feulgen stained sections takenadjacent to and distal from colonic carcino-mas, and related to stem cell DNA con-tent. DNA content was measured similarlyin cytological brushings from the sameareas and the percentage of 2N cellscalculated.There was a significant increase in the

amount of DNA in the proliferative cellsadjacent to the tumours (100%±1.3) com-pared to distally (91%±2.0, p<0.002).Although a similar increase was seen in thefunctional cells adjacent to the tumourscompared to distally the difference was notsignificant. In cytological preparationsthere was an increase in the proportion ofdividing or aneuploid cells in the transition-al mucosa (7-3%+1-2) compared with dis-tal mucosa (3-5±0-9, p<0-02) and to themucosa of patients with non-cancer relatedbowel problems (21% ±0-5, p,0-001).These techniques appear to be reliable in

detecting early malignant or pre-malignantchanges in the colonic mucosa patients.

P79" 'Indium granulocyte scanning in acutegraft versus host disease after bone marrowtransplantation

A1 122


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


S H SAVERYMU1TU, C O'BRIEN, J GOLDMAN, EGORDON-SMITH, A M PETERS, V CHADWICK,AND H J F HODGSON (Royal PostgraduateMedical School, Hammersmith Hospital,London) The successful use of bonemarrow transplantation to cure aplasticanaemia and certain haematological malig-nancies has been hindered by acute graftversus host disease (GVHD), which affectsprimarily the gastrointestinal tract in addi-tion to the liver and skin. Because intestin-al GVHD displays many of the histologicalfeatures of acute inflammatory bowel dis-ease we have investigated the value of"'Indium granulocyte scanning in diagno-sis and management.Ten patients with suspected or definite

GVHD after bone marrow transplantationwere studied and scans compared withhistological grading of rectal histology. Sixpatients with active GVHD showed exten-sive abnormal bowel activity on 'Indiumgranulocyte scan (both small and largeintestinal activity) and five subsequentlydied. Rectal histology was abnormal in allsix cases but in half the cases underesti-mated the severity of the disease. Threefurther patients were studied after treat-ment with methylprednisolone, whenGVHD was quiescent and showed justlocalised ileocaecal involvement while rec-tal histology was normal. The remainingpatient had a normal "'Indium granulo-cyte scan and subsequently GVHD wasexcluded.

"lIndium granulocyte scanning showsclear differences between active and quies-cent GVHD and appears to be a promisingnon invasive technique for assessing dis-ease severity and prognosis in this difficultgroup of patients.

P80Colonic permeability to 51Cr-EDTA in in-flammatory bowel disease

R T JENKINS, D B JONES, R L GOODACRE, R HHUNT, AND J BIENENSTOCK (Intestinal Dis-ease Research Unit, Departments of Pathol-ogy and Medicine, McMaster University,Hamilton, Ontario, Canada) The purposeof this study was specifically to measurecolonic permeability to 51Cr-EDTA inCrohn's disease (CD) and ulcerative colitis(UC) after rectal administration. Fivevolunteers (four men, one woman; aged21-43 years) served as the control group.Seven patients (four men, three women,aged 22-44 years) with colitis (one ileoco-lonic CD, one colonic CD, and five subtot-al or total UC) also were studied. No bowel

preparation was used. After normal eva-cuation of the bowel just before the test,25 uCi of 51Cr-EDTA in 30 ml of normalsaline (pH 6-2, 280 mOsm/kg) was instilledinto the rectum via a 30 cm, 8F, paediatricfeeding tube, while the patient was supinein the left lateral position. Urine wascollected for 24 hours. Volunteers wererequested to retain the enema for at leasttwo hours. Food and drink were allowed adlibitum. In the control group, urinaryexcretions of the probe ranged from 0.23-1 42%/24 h (mean 0 69%/24 h, SD 0-51%/24 h). In the patients with colitis, the 24hour urinary excretions of the proberanged from 2-95 to 21-57%/24 h (mean8 17%/24 h, SD 6-49%/24 h). The differ-ence between the control group and pa-tients was signifcant (p<003). This studyemphasises that colonic absorption of 51Cr-EDTA does occur in health and thatincreased permeability may be found inpatients with extensive colitis.

P81An audit of ulcerative colitis in a districtgeneral hospital

H W JONES, J GROGONO, AND A M HOARE(Wycombe General Hospital, HighWycombe, Bucks) Previous reports haveshown a high mortality for ulcerative colitis(UC) in district hospitals. We have carriedout an audit of all patients with UC in onehealth district (pop 270 000) between1975-1984. The incidence and prevalencewere 6-7 and 70/100 000 respectively. Nine-ty six patients required 114 admissions withacute colitis. Before admission 37 5% wereundiagnosed. Nineteen per cent requiredemergency surgery. There were no deathsfrom acute colitis suggesting an improvedprognosis for colitis in district hospitals.Three hundred and two patients werefollowed for 1151 patient years. Colono-scopy performed routinely eight to 10 yearsafter diagnosis detected two Dukes' Acarcinomas and one severe dysplasia.Seventeen per cent of patients initiallyassessed as having distal colitis were foundto have extensive disease including the twowith malignancy. Of 99 patients lost tofollow up two represented with carcino-mas, one died, being the only colitis relateddeath. Therefore, close follow up androutine colonoscopy even of patients withapparent distal disease appears worth-while. In this district the workload toexamine all patients 10 years after diagno-sis, and subsequently those with extensive

disease biannually, requires an estimated30 colonoscopies/year.

P82Clinical importance and complications ofthe early postoperative water-soluble con-trast enema

I G HAYNES, M GOLDMAN, S H SILVERMAN, J RLEE, J ALEXANDER-WILLIAMS, AND M R BKEIGHLEY (The General Hospital, Birm-ingham) The early postoperative water-soluble contrast enema (WSCE) is a wellestablished technique to assess the integrityof large bowel anastomoses. We haveassessed the safety and accuracy of aWSCE from a prospective series of 117consecutive patients undergoing colorectaloperations.Twenty four radiological leaks (24%)

were detected and 14 clinical leaks (12%)occurred. In four patients (3%), however,with clinical leaks, the anastomosis wasradiologically intact. The overall accuracyof a WSCE was 84-6% (false positive11-9%, false negative 3-4%; sensitivity71%; specificity 86%).

Septicaemia occurred in five patients(4-2%) after WSCE, one of whom died andonly one patient had a radiological leak.

Fifty two patients (44%) had a stapledanastomosis, and in 12 patients (23%) thering was disrupted on plain radiograph. Allthe clinical leaks after stapled anastomosiswere identified by this non-evasive techni-que.

This study of postoperative WSCE indi-cates that the investigation is potentiallydangerous and does not always identifyclinically important anastamotic dehis-cence.

P83An analysis of anal sphincter competence bymeasurement of anal compliance

C P GIBBONS, A TROWBRIDGE, J J BANNISTER,AND N W READ (Departments of Surgery,Physiology and Medical Physics, Universityof Sheffield, Sheffield) The circularsmooth muscle of the anal sphincter isrequired to contract sufficiently to close theanal canal in order to preserve continence,but also to stretch sufficiently to allow thepassage of a stool. These functions dependupon the elastic properties of the sphincter,which have been little investigated. Analdistensibility was assessed in 14 normalmen and 11 normal women by measuringsphincter pressures via perfused catheters,

A1123


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


set into probes of increasing diameter (0.4to 3-0 cm) and connected to pressuretransducers. Sphincter pressures duringrest, maximum voluntary contraction, andinflation of a 100 ml rectal balloon (mini-mum residual pressure) rose with increas-ing anal diameter between 0 4 and 2-0 cm(p<O.OOl). Thereafter resting and squeezepressures reached a plateau. The estimatedsphincter tension was proportional to dia-meter over the whole studied range. Theslope of this relationship (men 70 5±4-4g/cm2; women 57-5±2-3 g/cm2;mean±SEM) increased (p<0-01) duringsqueezing (men 177.1+3.6; women126±12.6) and decreased (p<0-01) duringrectal distension (men 32 8±3 5; women31-0±2.8). These results conform to amathematical model of the anal sphincteras a thin elastic tube whose elasticity andintrinsic diameter change with the state ofsphincter contraction. The model predictsthat the smooth muscle itself cannot closethe anal canal and shows the importance ofthe bulk and turgor of the vascular analcushions in the maintenance of continence.

P84Patients with pruritis ani leak liquids fromthe anal canal more readily than normalsubjects

N S AMBROSE, A ALLAN, S SILVERMAN, ANDM R B KEIGHLEY (Department of Surgery,The Genral Hospital, Birmingham) Wehave assessed the physiological abnormali-ties in the anorectum of 20 patients withpruritis ani compared with 12 age matchedcontrols.A saline infusion (1500 ml) test showed

that leakage started after infusion of 425 mlin patients with pruritis ani, compared with1500 ml in controls (p<0-001).The anal canal high pressure zone was

greater in patients (4 cm) than controls (3cm). The resting anal canal pressures werealso greater in patients (100 cm H20) thancontrols (82 cm H20). A similar trend wasfound with regard to the maximum squeezepressures (280 cm of water and 213 cm ofwater respectively). The degree of perinealdescent was 1-5 cm in patients comparedwith 14 cm in controls. The rectoanalinhibitory reflex was absent in five of thepatients (25%) compared with only onecontrol (8%), however, the percentagereduction of anal pressure after 50 mlinflation of a rectal balloon was greater inthe patients with pruritis ani (40%) com-pared with controls (23%).

Results indicate that patients with prur-

itis ani leak liquids through the anal canalmore readily than controls, despite appa-rently normal sphincter pressures.

BSG/BASL: LIVERP85-100

P85Partial purification of a high molecularweight hepatotrophic factor from humanserum

A C SELDEN, R JOHNSTONE, S GUPTA, HDARBY, AND H J HODGSON (Department ofMedicine, Royal Postgraduate MedicalSchool, Hammersmith Hospital, London)Several circulating low molecular weighthepatotrophic factors have been associatedwith liver regeneration after partialhepatectomy including insulin, glucagonand epidermal growth factor (EGF). Incontrast, we have partially purified a highmolecular weight (approx 150 000 daltons)factor from human serum taken 24 hoursafter partial hepatic resection. Hepato-trophic activity was shown on rat hepato-cytes cultured in supplemented Williams Eby 3H-thymidine incorporation into DNA.The hepatotrophic factor, prepared by gelfiltration and heparin-sepharose affinitychromatography, or EGF and insulin, wereadded 20 hours postplating of the cultures.The 'hepatotrophic factor' stimulated

DNA synthesis in a dose dependent man-ner (161 ,ug/ml, 0 42x 106 dpm 3H/mgprotein, 320 ,ug/ml, 0-54x 106 dpm 3H/mgprotein), to a greater extent than dex-amethasone alone (0*293x 106 dpm 3H/mgprotein). At concentrations studied thefactor was 20% as potent as optimalconcentrations of EGF plus insulin. Asimilar high molecular weight hepato-trophic factor from rat serum afer partialhepatectomy reached 60% the potency ofEGF+insulin, suggesting either partialspecies specificity, or differing time-relationships of production of these factorsafter hepatic resection in different species.

P86Effects of chenodeoxycholic (CDCA) andursocholic (UCA) acids on gall bladder(GB) emptying

P HOWARD, G M MURPHY, AND R H DOWLING

(Gastroenterology Unit, Division of Medi-

cine, UMDS of Guy's and St Thomas'sHospitals, London) We previouslyshowed that ursodeoxycholic acid treat-ment reduces GB emptying after a CCKinfusion of a Lundh meal, in gall stonepatients. We recently found, however, thatnet GB emptying after a liquid+solid mealis complex with early emptying, refillingand late emptying phases, superimposedon which are minute-by-minute episodes offilling and emptying. As the effect of otherbile acids on this complex pattern is un-known, we studied the GB' response to ameal of baked beans on buttered toast withmilk before, and after six weeks treatment(15 mg per/kg per/d) with either CDCA(five men and two women) or UCA (fourmen). Treatment consistently increasedpreprandial GB volumes from 11-2±SEM2-97 ml to 19 6±3 20 on CDCA and from17 6±9-30 to 32-1±5-58 on UCA(p<005), and late nadir volumes from2-09±0 86 to 5-36±0 5 ml on CDCA and4-23±0 6 to 12 4±2-2 ml on UCA(p<0-05) with corresponding changes inthe early nadir and refilling volumes. Theoverall emptying time was not altered byCDCA, though with UCA is decreasedfrom 122±5-8 to 88±9 2 minutes (p<0.05).The minute-by-minute changes in GBvolumes were also greater during therapyso that the net flux of bile thorugh the GBincreased in six of seven CDCA- and threeof four UCA-treated subjects.

In summary we conclude that treatmentwith CDCA and UCA markedly increasesGB size and flux of bile through the GB.UCA also shortens GB emptying time.These changes may relate to cholereticproperties of the bile acids and to resultantchanges in bile volume loads handled bythe gall bladder.

P87HBV infection and alcohol abuse: a synergiceffect leading to a more severe liver diseaseor a casual association?

M CHIARAMONTE, A FLOREANI, D MARTINES, MSALVAGNINI, E PORNARO, AND R NACCARATO(Department of Gastroenterology, Policli-nico Universitario, Padova, Italy) Eightysix (77 men) heavy drinkers - 29HBsAg+ve, 13 antiHBs/antiHBc+ve, 14antiHBc+ve, 30 negative for any HBVmarker - were studied to verify whether:(1) the HBV can be responsible for chronichepatitis in alcoholics; (2) the concomitantpresence of HBV infection and alcoholabuse enhances the severity of the disease.Histological features of ACH were present

A1124


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


in 12 patients (14%) (six HBsAg+ve, fiveantiHBc+ve and/or antiHBs+ve and oneHBV negative) and of CPH in 13 (15%)(six HBsAg+ve. three antiHBc/anti-HBs+ve and four HBV-ve); fourHBsAg+ve patients had signs of acutehepatitis superimposed on cirrhosis. Thirtypatients had cirrhosis (11 HBsAg+ve andeight antiHBV+ve) and 27 steatosis (twoHBsAg+ve, 10 antiHBV+ve and 15 HBVnegative). Liver tissue HBcAg was de-tected by indirect immunofluorescence in50% of HBsAg+ve patients (but withoutrelationship with serum HBeAg/antiHBepositivity), in 33% of antiHBV+ve pa-tients and in one of 30 HBV serumnegative patients. The clinical course hadbeen very severe in HBsAg+ve patients:four aged 25, 31, 43, 44 years, died of liverfailure, while none of the HBsAg negativedied during a comparable follow up period.The prevalence of 'severe' liver disease(cirrhosis or ACH) was 80% in theHBsAg+ve, 61% in antiHBs/antiHBc+veand 33% in HBV negative.We conclude that (1) the HBV has a

putative aetiopathogenic role in chronichepatitis in alcoholics; (2) alcohol abuseenhances HBV pathogenicity and - viceversa - the course of alcoholic liver diseaseis accelerated by the HBV infection.

P88Assessment of portal vein patency: value ofultrasound scanning

N RABY, J KARANI, P POWELL-JACKSON, HMEIRE, AND R WILLIAMS (Liver Unit, King'sCollege Hospital, and School of Medicineand Dentistry, Denmark Hill, London)The value of ultrasound scanning (USS) forestimating portal vein (PV) patency has notbeen assessed previously in a large series.In the present study the accuracy of USSwas estimated either by comparison withthe findings at operation (21 patients) orwith the findings after aortoportography(94 patients). In the first group, USS wasaccurate in 18 (86%). Non-visualisation ofthe PV by USS in one and misdiagnosis ofPV occlusion in two was attributed in eachinstance to distortion of hepatic anatomyby shunt surgery in two and a Kasaioperation in one. In the second group,confirmation of USS findings was obtainedby aortoportography in 73 (78%). Non-visualisation of the PV by USS could beexplained in eight of 11 patients as due toportal vein thrombosis (two), hepaticmalignancy (four), aberrant PV (one) andprevious shunt surgery (one). Misinterpre-

tation of USS findings in five could beattributed to cavernous transformation ofthe PV (two), hepatic malignancy (two),and previous Kasai operation (one). Inconclusion, the accuracy of USS for esti-mating PV patency is high but in casescomplicated by previous Kasai operationor shunt surgery, aortoportography shouldalso be carried out.

P89Comparison of lactulose and lactitol on ilealand colonic pH

D H PATIL, D WESTABY, Y R MAHIDA, K RPALMER, R REES, M L CLARK, AND D B A SILK(Departments of Gastroenterology, CentralMiddlesex Hospital, London, and St Bar-tholomew's Hospital, London) Lactitol isan unabsorbed disaccharide with a definedlaxative threshold. It has superior tasteproperties to lactulose and has been sug-gested as an alternative to lactulose fortreatment of chronic hepatic encepha-lopathy. The aim of the present study wasto compare the effects of these two sugarson luminal pH in the terminal ileum, colonand rectum of five normal subjects. Theluminal pH was recorded every one to twohours using a pH-sensitive radiotelemeter-ing device, either after ingestion of anormal diet, or this supplemented withsufficient lactulose or lactitol to produce2-4 semiformed stools daily. Neither sugarhad an effect on terminal ileal pH (basal7-48+0-40; lactulose 7-20±0-48; lactitol7-02+0-31, mean±SD). The pH of theright colon (basal 6.36±0.32) was signifi-cantly lowered during ingestion of bothlactulose (5-1±0-89) and lactitol(5-78±0-45; p<005 or less). There was nosignificant difference between the acidifica-tion properties of the two sugars. Neitherlactulose or lactitol had a significant effecton the pH of the left colon or rectum. If themode of action of lactulose in the treat-ment of hepatic encephalopathy is depen-dent upon its ability to lower right sidedcolonic pH, then our data lend support tothe suggestion that lactitol may also have arole to play in the treatment of thiscondition.

P90Effects of dietary protein on plasma andCSF amino acid levels: correlation with thedegree of encephalopathy

S A JENKINS, N B ROBERTS, J N BAXTER, GSKERRITI, AND R SHIELDS (Departments ofSurgery and Veterinary Anatomy, Univer-

sity of Liverpool, Liverpool) There islittle information on the precise relation-ship between diet and the development ofhepatic encephalopathy (HE) after porta-caval shunting (PCS). Therefore, westudied the effects of varying protein dietson plasma and CSF amino acid profiles andthe development ofHE in dogs with a PCS.Fasting blood and CSF samples were takenfor amino acid and ammonia estimationsfrom four dogs with a PCS at the end of 28days of respectively protein free (controlperiod), 15% and 40% protein diets. Thedegree of HE was assessed at the time ofsampling. Normal fasting ranges of CSFand plasma amino acids were determinedin dogs without a PCS. The degree of HEincreased with increasing protein intake.There was a marked decrease in plasmalevels of branched chain amino acids innon-shunted dogs which did not alter withincreasing protein intake. In contrast,plasma phenylalanine, glycine and ammo-nia levels increased with higher proteinintake. The marked encephalopathy at theend of the 40% protein regimen wasassociated with significant increases in CSFammonia (52 to 316 ,umol/l), glutamine(1126 to 2919 ,umol/l), phenylalanine (51 to75 umol/l and glycine (22 to 172 ,umol/l)(Students t test; p<005). However, CSF,GABA and tryptophan showed no consis-tent change. The results suggest; (1) porta-caval shunting maintains low levels ofbranched chain amino acids irrespective ofdietary protein intake; (2) the effects of ahigh protein diet on CSF amino acids andammonia may have important implicationsin the management of patients with HE.

P91Detection of CA-50 in serum from patientswith malignant liver disease

N A HABIB, M HERSHMAN, A GRAUER, M

BLOUNT, L LINDHOLM, J HOLMGREN, AND C BWOOD (Department of Surgery, RoyalPostgrauate Medical School, London, De-partment of Medical Microbiology, Univer-sity of Goteborg, Sweden) Malignanttransformation of the cell is accompaniedby changes in surface glycolipids. Amonoclonal antibody C-50 was raisedagainst a tumour associated antigen, CA-50 which has been isolated as a complexmono sialoganglioside from a high percen-tage of various carcinomas. A solid-phaseRIA-inhibition test was developed for de-tection of CA-50 in serum from tumourpatients. A total of 137 patients werestudied, 83 with liver tumours, 24 control

A1 125


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/

The British Society ofl(, stro lter-oilo

and 30 normal subjects. Among the cancerpatients studied there were 25 with hepa-tocellular carcinoma (HCC), 20 cholan-giocarcinoma and 36 with secondary liverneoplasms and two with angiosarcomas.The control patients included 10 livercirrhosis, six iatrogenic biliary stricutre,three haemangioma, two sclerosing cholan-gitis, two hydatid disease, two intrahepaticcholelithiasis, two liver cyst, two adenomaand one liver sarcoidosis.The CA-50 RIA test was positive in 15

(60%) HCC, 14 (70%) cholangiocarcino-ma and 26 (72%) of those with secondaryliver neoplasms. The test was negative inthe two patients with angiosarcoma, allcontrol patients and in all normal subjects.Therefore the CA-5O RIA test could helpin the differential diagnosis of pathologicalliver conditions.

P92Expression of oncogene related proteins inhuman malignant liver neoplasms

N A HABIB, H NIMAN, A THOMPSON, AND C B

WOOD (Department of Surgery, RoyalPostgraduate Medical School, London, andDepartment of Molecular Biology, ScrippsResearch Centre, La Jolla, USA) Protooncogenes are responsible for normal cellgrowth and their conversion into activatedcellular oncogenes are associated withcancer. We investigated the expression ofras-oncogene product in human liver withthe use of monoclonal antibodies raisedagainst c-Ki-ras oncogene normal product(Mo-EP). Using the peroxidase-anti-peroxidase technique, liver tissue wasstained with Mo-RAP to a concentration of1:200. The patients studied were five nor-mal, 10 hepatoma and 20 colorectal liversecondaries. Positive staining in the cyto-plasm was found in six of the 10 hepatomasand 16 of the 20 liver secondaries. Theremaining cancer patients and the fivecontrol patients (apart from weak non-specific staining in the connective tissue)had negative hepatocytes staining. Thesefindings suggest that liver tissue may ex-press ras-oncogene in carcinoma and thismay offer new diagnostic and therapeuticapplications.

P93Lack of osteomalacia in chronic activehepatitis on maintenance corticosteroidtherapy

A J STELLON, J COMPSTON, AND R WILLIAMS

(Liver Unit, King's College Hospital,London SE5, and Department of Histo-pathology, St Thomas's Hospital, LondonSE1) Subnormal 25-hydroxyvitamin-D(25-OHD) levels have been reported tooccur in 45-60% patients with chronicactive hepatitis (CAH), although the num-ber that developed osteomalacia is un-certain as diagnostic iliac crest histologywas performed in few patients. We reportserum 25-OHD levels and iliac crest histol-ogy to determine the incidence of osteoma-lacia in 36 patients, aged 20-66 years, 34women, with biopsy-proven CAH, inbiochemical and histological remission, ofwhom 23 had cirrhosis on biopsy. All hadreceived prednisolone therapy for one to12 years (median 3-6 years) and the main-tenance dose ranged 5-12.5 mg/day (me-dian 10mg). All patients had an iliac crestbiopsy to determine total trabecular bonevolume (TBV), mean osteoid seam width(MOSW) and mineralisation lag time(MLT). Osteomalacia was defined ifMOSW was >15 gm coupled with a MLT>100 days. 25-OHD levels were low infour (11%) patients, TBV was significantlylower in the patients when compared to ageand sex-matched controls (19.2+4.6 vs23-5±5-8 respectively; p<0001). No pa-tient was found to have an MOSW >15 ,umand the mean values were not significantlydifferent from controls (8 41±2 202 vs9-59±2 36 respectively; p=NS) but MLTwas significantly prolonged in the patients(28.4+310 vs 13 7±8-4 respectively;p<0-01). Low TBV and absent osteomala-cia suggests that osteoporosis is the meta-bolic bone disorder associated with steroid-treated CAH.

P94Prophylactic chemotherapy after opera-tions for hydatid disease - an animal study

D L MORRIS, JANET B CHINNERY, AND J DHARDCASTLE (Department of Surgery, Uni-versity Hospital, Nottingham) There isapproximately a 10% risk of recurrenceafter operation for hydatid cyst (Echino-coccus granulosus). The surgical treatmentof recurrent disease is associated withconsiderable morbidity and whilst chemo-therapy with Mebendazole, and more re-cently Albendazole has been investigatedin such patients, prevention of recurrencewould have great advantages.

In order to study the effect of Albenda-zole on the development of cysts, 25 gerbilswere given intraperitoneal infections of live

E granulosus protoscol(TC). "iNitC iocihiliremained as untre'tc(i Stn\its SLreceived Albendazolce 1imc, k( uuiatllx ltoone week before injectioiicand cikhtrceived the same dosage t( Ii0Tl \xCcck Atictcinfection. The animnals Wci c klilledi at siomonths. All control tnnimalls dcxc%lop)cperitoneal cysts. Thev haltd Ia nn ntiLticof 44 1 (SD16.5) exsts pcCcrihil xhilci:eight gerbils wxho rccixlcd Alhcndiiuilzlafter infection, txo hald nio 1Stisianmean number ot cyxstS xx a it (SD) "-((p<001). The c\sts xx hrch didi dc clop IIthe treated animals wx cic Ut siil 'Csi -

controls. There was no sinnitctllt iCdLdntion in number ot cy sts n hc mnllildl,itreated before infectionl ('I c%\ iCollclude tht even a short postopcl ix C cmIIi .of Albendazole is likelx to eniticanreduce the number of spilledi iiritoscolcccwhich are able to implant miid Mroxcysts.

P95Importance of clinical staging for prognosisin primary biliary cirrhosis

O EPSTEIN. E FRA\(\ ((iok

(Departments of Alctlii incto nti8aa ,wt.Roval Free Hospitual S chool/ X/mAlto(li(oiLondon) In PBC the. csI'.iin U

prognosis is importanrt oi (od ix dal P J-

tients, interpretation ol clinical iitrils L

timing of liver transplkinuition Batsed oni 1

group of 56 poor risk PBC patIticnts xshIdied of their liver diSetSxc, \x C ix C'vdcx 5cca staging protocol to prox ide nidldliocs fmthe rational managemenit of the dissc, (3415 patients referred xith hilinuhrin lexcl,>100 ,umol/l, 12 died within txxo \xciTrs ifiall had succumbed within thi-ce \Ca(lrsFourteen of 24 patients reterrcd x ith liilrubin levels <34 umol sLr ixveitcor tixc l13 years. Serum biliruhin lcxels ot tOlpatients were pooled 'it x arlx iMtcrix atindicated that on axera-c. sxnmptoinattmpatients with normal serum hilirLnbin cx clxreach a level-of 34 jumol within tix-c xCaitsand 100 ,umol/l within li0 xctrs Deatthusually occurs within two xcatrs of icachinelthis level. Primary bilihar\ cirrhosis catni hedivided into four stages. Staec A\ asxnip-tomatic, normal biliruhbin (no:-llial litc cx-pectancy); stage B: symptomatic. hlir-ulhill<34 ,umol/l (survixal sexcn, to xciistage C: bilirubin 34-100( /Lnf(ol snli-xixV1ltwo to seven yeairs : stiac 1) hilirinhili>100,mol/l (surxvixval lcss thalxxiixctirsi'Stage A patients do not i- clcirc slpictictreatmment. Clinical trials oimedical ti-cztment should focus on stitac B aimd ( cliscilst

A1126


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


and liver transplantation should be con-sidered in stage D.

P96Long term prognosis of Budd-Chiari syn-drome

S GUPTA AND H J HODGSON (Department ofMedicine, Royal Postgraduate MedicalSchool, London) The Budd-Chiari syn-drome is commonly considered a seriouscondition with progressive deteriorationand a high mortality. It is, however, anheterogeneous condition, and we haveanalysed the histories of 18 patients seenover 20 years to elucidate determinants ofprognosis.

In all patients diagnosis was based onliver biopsy and/or hepatic venography. Inseven men and 11 women, mean age 37-6(19-60 years range), symptoms had beenpresent on average for 6-6 months beforediagnosis. Patients were followed for 4-7years (range 0-5-19). Eight patients died(44%), two after surgery, two because ofassociated tumours, and four from compli-cations of hepatic failure. Acturial analysisshowed a one year mortality of 22*9%, andthree years of 51 1%, but no furthermortality beyond this period. A poor prog-nosis was associated with an associatedmalignancy, and bleeding fromoesophageal varices, but in the absence ofthese features the prognosis is unexpected-ly good. The 10 survivors all lead normallives with control of symptoms by medicalmeans, though in two surgical therapy(side-to-side shunt or removal of hepaticvenous webs) has been necessary. Manypatients with Budd-Chiari syndrome haveonly mild symptoms and a good prognosis.

P97Liver biopsy in patients with impairedcoagulation - which route?

M V TOBIN AND I T GILMORE (Gastroenter-ology Unit, Royal Liverpool Hospital,Liverpool) Histology is crucial in theinvestigation of hepatic disease but there isstill a signifcant morbidity and mortalityassociated with conventional liver biopsy.Its safety, however, has usually been evalu-ated in patients with normal coagulationand the complication rate in those withprothrombin times greater than 15 secondsor platelet counts below 80x 10'/l is un-known.

In our unit during the past three years, inpatients with impaired coagulation, we

have preferred the recently described'plugged' biopsy to the transvenousapproach. Sixty five patients withprothrombin times prolonged by up toeight seconds or platelet counts as low as20x109/l have been biopsied and the nee-dle track filled with absorbable gelatinsponge. The technique is straightforwardand there have been no complications.Specimens have all been satisfactory forhistological examination and superior tothose taken transvenously. Our results aresufficiently encouraging to recommend itsuse in this high risk group in preference tothe laparoscopic or transvenous approachwhich both require more expertise, equip-ment and time.

P98Long-term oral contraceptive use andserum total cholesterol and total bile acids

P R BAKER, J S BUMBRA, A D REID, P E PREECE,AND J D E KNOX (University Departmentsof Surgery and General Practice, NinewellsHospital and Medical School, Dundee)Oral contraceptive steroids (OC) increasethe cholesterol saturation index of bile andare associated with raised serum totalcholesterol concentrations. Changes inserum bile acid levels might also be ex-pected, especially after several years use ofOC comprising ethinyloestradiol (EE) anda progestogen. Serum concentrations oftotal cholesterol (CHOD-PAP method)and total bile acids (TBA; Sterognost-3oxFlu assay) were therefore determined in10 women (27-8±4-7 years (mean±SD)with well documented long-term EE/prog-estogen OC intake and compared with-levels in 10 age-matched women (26-0±4-9years) from the same general practice whohad never taken OC (NON-OC). The OCwere taken, as prescribed, continuously for3-12 years (mean 5-6 years), and totaloestrogen and progestogen intakes wereestimated as 25-156 mg and 93-2948 mgrespectively, and all were currently on 30Ag EE and 50-250 jig L-Norgestrel. Bloodwas taken between 1800 and 1845 hoursafter a six hour fast and serum obtainedwithin one hour. Total cholesterol (mM)were significantly higher in the OC group(5-83±0-57 (mean±SD) vs 4-75±0-67;p<0-05, t-test) and four of these womenhad values >5-7 (suspicion limit) comparedwith only one in the NON-OC group.Serum TBA (MuM) was lower in women onOC (3-7±1-9 vs 6-1±2-8) although thedifference was of border-line significance(Mann-Whitney U=24, z=1-968 (cor-

rected for ties); t=2.25). The TBA/cholesterol molar ratio was significantlylower in the OC group (0.7±0.4x 10-3 VS1-3±0-7x 10-3; p<0-05, Mann-Whitney)and the two constituents exhibited a nega-tive correlation (rho=0-51, p<0-025).There was, however, no correlation be-tween total cholesterol or TBA and totalduration or total amount of OC steroidintake. Oral contraceptive steroid intakeover several years appears to result in arelatively large increase (23%) in serumtotal cholesterol and a decrease in serumtotal bile acids which might reflect a changein the output of bile acids by the liver.

P99Biopsy findings in liver allograft rejection

S HUBSCHER, D CLEMENTS, E ELIAS, AND PMCMASTER (Queen Elizabeth Hospital,Edgbston, Birmingham) It is difficult todistinguish liver transplant rejection fromother causes of graft dysfunction on thebasis of histological criteria alone. Earlystudies, based mainly on necropsy findings,reflect this problem and the histologicalfeatures which characterise rejection re-main the subject of controversy.During the past 12 months we have

routinely carried out needle biopsies in theassessment of abnormal liver function aftertransplantation. In all cases a diagnosis ofrejection was based on excluding otherknown causes of graft dysfunction bymicrobiological, serological and radio-logical criteria. Changes ascribed to rejec-tion were seen in 15 biopsies from fivepatients. The features of acute rejectionwere (1) a mixed portal inflammatoryinfiltrate and (2) infiltration of bile ductepithelium by polymorphonuclear leuko-cytes. In one patient this was followed bycomplete loss of small bile ducts and theneed for retransplantation. In a secondpatient there was chronic bile duct damageresembling that seen in graft-versus-hostdisease. In the other three cases there wasclinical and histological improvement fol-lowing immunosuppressive therapy.We conclude that liver biopsy is useful in

the diagnosis and management of rejectionafter liver transplantation.

P100Increased permeability to proteins of hepa-tic tight junctions in x-naphthyl-isothiocyanate (ANIT)-treated rats

K S KAN, P J LOWE, AND R COLEMAN

A1127


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/

A1 128

(Department of Biochemistry, University ofBirmingham, Birmingham) The move-ment of proteins from blood to bile hasbeen studied using isolated perfused liveroperating under one pass conditions. Aftera one minute perfusion pulse of horserad-ish peroxidase (HRP) two peaks wereobserved in bile, at five, and at 20-25minutes (abolished by colchicine and prob-ably representing a vesicle mediated trans-cytosis). The first HRP peak was substan-tially increased in rats pretreated withANIT, known to increase paracellularpermeability, probably through an effecton tight junctions. There was no decreasein bile flow within the time-period of thepresent experiments (until 15 hours), andthus the effects represent one of the ear-liest stages in ANIT-cholestasis.

Rats were treated with ANIT 0-15 hoursbefore pulsing with HRP. Biliary concen-tration, in the five minute peak, ofcopulsed inulin (MWt 5000) rose steadilyover the dose period, but the peak of HRP(MWt 40 000) and of ovalbumin (MWt40 000) began to rise only after 10 hourexposure to ANIT, suggesting that theextent of ANIT-increased permeability isdependent upon the molecular weight. Thefive minute HRP peak was increased afteronly two hours of ANIT exposure inphenobarbitone-treated rats suggestingthat ANIT metabolites, rather than ANITalone, may be involved.

COLORECTALP101-117

P101Laparotomy: still a valuable investigation inpatients with bleeding from the intestine

S BREARLEY, P C HAWKER, N J DORRICOTr, RLEE, N S AMBROSE, P W DYKES, AND M R BKEIGHLEY (General Hospital, Birming-ham) Bleeding lesions of the small intes-tine and colon are often difficult todiagnose despite the availability of a widerange of diagnostic techniques. In recentpapers, laparotomy has not been recom-mended and particularly good results havebeen claimed for selective mesentericangiography.

Sixty three patients who had had eithercolonoscopy or mesenteric angiographywhile being investigated for intestinalbleeding were reviewed. Twenty six hadhad trivial bleeding and were not consi-

dered further. The remaining patientscould be divided into two groups, 14 withoccult bleeding and 23 who had had amajor intestinal bleed. Colonoscopy wasdiagnostic on six of 37 occasions (two of 15following occult bleeding and four of 22after major haemorrhage). Angiographywas diagnostic on three of 14 occasions(occult bleeding none of three, majorbleeding three of 14). Thirteen undi-agnosed patients had a laparotomy whichwas diagnostic in nine (occult bleedingthree of three, major bleeding six of 10).Patients with intestinal bleeding shouldhave upper and lower gastrointestinal en-doscopy. Undiagnosed patients who con-tinue to bleed should have a laparotomywith an endoscopist available to undertakeon table panendoscopy. Angiographyshould be reserved for the few patients stillundiagnosed after laparotomy and shouldbe carried out in centres with specialexpertise in the technique.

P102Modified oxygen electrode as a probe fordetecting local blood flow in canine colon

C PIASECKI AND S LAKE (INTRODUCED BYR POUNDER) (Department of Anatomy,Royal Free Hospital School of Medicine,London) There is a need for simple, fast,non-invasive probing of blood flow in areasof - for example 1/2 sq cm. Postanastomoticcolonic leakage might thus be prevented iflocal ischaemia was detectable peropera-tively. A Clark type surface oxygen elec-trode was modified to perform this task. Itdoes not measure P02 because its oxygenconsumption is high. Instead, it depletesthe tissue of oxygen, the maintenance ofwhich is dependent on inflow of blood.Thus the instrument's reading varies withblood flow. A 95% response is obtained in20 seconds of application, thus sites can beprobed at 15 second intervals.The instrument was tested in six anaes-

thetised dogs. Electromagnetic blood flowin the inferior mesenteric artery was com-pared with probe reading on colon, at fivelevels of graded arterial constriction. Asemi-linear relationship emerged (r=0-91;p<l%), showing dependence on blood-flow. The electrode was more sensitive toflow changes at low flows that at highflows.Being cheap, sterilisable, and simple to

use, the instrument may prove valuable indetecting and grading areas with reducedflow.


P103Comparison of low residue diet and purga-tion on different faecal occult blood tests

C Y YIU, L BAKER, P B BOULOS, AND C G CLARK(Department of Surgery, University Col-lege, London) It is generally accepted thata high fibre diet improves the yield offaecal occult blood tests in colonic lesions.It has been suggested that laxatives have asimilar effect. We have examined thisconcept by testing stool specimens using allavailable test procedures. One hundredand thirty nine patients were included inthe study and faecal occult blood testingwas performed after a three day lowresidue diet and again after purgation withPicola'x. Comparison was made betweenHemo-Fec, Hema-Check, Haemoccult andFecaTwin tests. Patients were admitted forcolonoscopy or surgery so that diagnosiswas known.The patients were divided into normal or

pathological groups (cancer, polyp, diverti-cular, and inflammatory bowel diseases).The positive results in 74 patients withpathology before and 70 after purgationare compared with 42 normal subjects.Statistical analysis by the x2 test showed nodifference in the results before and afterpurgation. Hemo-Fec and FecaTwin wereapparently statistically similar but betterthan Hema-Chek and Haemoccult whichwere not different. These results questionthe need for special bowel preparation forfaecal occult blood testing and indicate thata positive occult blood test is not related tolaxative intake. The choice of test willinfluence the diagnostic yield and theproportion of positive results in eachpathological group.

P104Is increased colonic prostaglandin E2 a riskfactor for colorectal cancer?

M R LEWIN, S PUGH, SIAN WILLIAMS, TINABARTON, C G CLARK, AND P B BOULOS(Department of Surgery, University Col-lege, London) Several colorectal condi-tions are associated with an increased riskof colorectal malignancy. The commonestof these are ulcerative colitis and colorectalpolyps. Previous studies have demon-strated that the synthesis of PGE2 bycolonic mucosa is increased in the presenceof UC. Other studies have shown it in-creased in colorectal cancer and postulatedthat PGE2 leads to a loss of normal growthrestraint and the increased cyclooxygenaseactivity necessary to produce PGE2 pro-


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


motes the formation of carcinogens frombenzo(a)pyrene. We are interested in theadenoma-carcinoma sequence and havemeasured PGE2 synthesis in biopsies takenfrom polyps and cancers at colonoscopy.Normal mucosal samples were obtained atnegative procedures. Biopsies were vortex-ed to stimulate PGE2 synthesis and re-leased PGE2 measured by RIA. Resultswere (mean±SD in pg PGE2/mg wetweight): normals 102-8±24-3 n=11, polyps162-1±29-6 n=6 (p<0-001 vs normal),cancers 177-1±13-1 n=7 (p<0-001 vs nor-mals, NS vs polyps) and uninvolved mu-cosa of tumour bearing colons 117-3±32-6n=12 (p<0-002 vs polyps and cancers, NSvs normals). We conclude that like otherpremalignant conditions of the colon,polyps synthesise increased quantities ofPGE2 and that this may be a factor in thetransformation of benign polyp to coloniccancer.

P105In vitro binding and in vivo localisaiton of ananti CEA monoclonal antibody (C46)

N C ARMITAGE, K C BALLANTYNE, L DURRANT,R C HARRISON, A M L RILEY, I 0 ELLIS, A CPERKINS, AND J D HARDCASTLE (Depart-ments of Surgery, Pathology and MedicalPhysics, University Hospital, Nottinghamand Amersham International, Bucks)Antitumour monoclonal antibodies havebeen shown to localise in colorectal cancerby several workers. The directly measuredtumour:non-tumour (T:NT) uptake in re-sected specimens after pre-operative injec-tion of radiolabelled antibody has beenonly 2-5-3-3:1. A monoclonal antibodyC46, raised against carcinoembryonic anti-gen was studied for in vitro binding and invivo localisation to colorectal cancer.

Binding was measured by flow cytometryof disaggregated tumour cells from 15primary and three metastatic/recurrenttumours and by immunoperoxidase stain-ing in 18 cancers. Eight patients (sevenprimary, one recurrent) were injectedpreoperatively with 111In-labelled C46 andtumour and normal tissue from the re-sected specimens counted.

Binding of C46 was increased 10-fold,median fluorescence 463F1U (33-2203)compared with normal immunoglobulin43F1U (24-86) (U=20.5, p<0.0001). Im-munohistologically intense staining wasseen in 15/18 sections, moderate staining in2/18 and only one section did not stain. TheT:NT uptake in resected specimens wasmean 5*8±1-7:1 for primary cancers

(seven) and 4-6:1 for the recurrence. Allpatients gave positive preoperative gammacamera images.The high affinity and high T:NT ratio of

5-8:1, considerably greater than previouslyreported, increases the prospects for effec-tive targeting of antitumour agents withmonoclonal antibodies.

P106Duodenal bile acid profiles in patients withcolorectal polyps

R J MOORHEAD, JOAN D DONALDSON, AND S TD MCKELVEY (Department of Surgery, TheQueens University of Belfast, Belfast) Ithas been reported that patients with col-orectal adenomas have increased concen-trations of secondary bile acids in duodenalbile. To investigate this finding we havecarried out an analysis of duodenal bileacids using gas liquid chromatography, in38 patients with histologically proven ade-nomas and compared them with a group ofcontrols matched for age and sex.We have shown that the levels of che-

nodeoxycholic acid were significantly high-er in those with polyps compared withcontrols (mean percentage±SD,30 5%±10 6: 26-5±9-0: p<0-04). Thelevels of cholic acid and deoxycholic aciddid not differ significantly between the twogroups. (Cholic acid: 47-5%±13-0:50%±14-1: deoxycholic acid 18-7%±7-9:22.2%±12-7). We have also shown anunidentified bile acid which was significant-ly more prominent in the polyp groupcompared with the controls (2-9% ±4-2:0.9%±1*6: p<002). The levels of cheno-deoxycholic acid and of this unidentifiedbile acid correlated with the increasingmalignant potential of the adenomas withrespect to size, type and degree of dyspla-sia.We have confirmed that the duodenal

bile acids of patients with colorectal adeno-mas are abnormal when compared withcontrols. Contrary to the findings of theprevious study we have shown the levels ofchenodeoxycholic acid to be abnormallyelevated in those with polyps, and havebeen unable to detect any significant differ-ences in levels of deoxycholic acid. Furtherwork is being done to identify the unknownbile acid.

P107ABSTRACT WITHDRAWN

P108Influence of sialomucin at the resectionmargin on survival of patients with colorec-tal cancer

P M DAWSON, N A HABIB, R C N WILLIAMSON,AND C B WOOD (Department of Surgery,Royal Postgraduate Medical School, Lon-don, and Bristol Royal Infirmary, Bristol)There is strong evidence to suggest that apredominant sialomucin production in thecolonic mucosa represents a preneoplasticstage in the carcinogenic process. Wetherefore studied the relation betweensialomucin at the resection margin andsurvival in patients with colorectal cancer.In a multicentre prospective trial, 204patients have been followed for a mean of14-8 months (SD 7-3 months). The pre-sence or absence of sialomucins at theresection margin was studied histochemi-cally using the High iron diamine-alcianblue (HID-AB) stain. There were 32deaths relating to tumour recurrence: 15deaths in the sialomucin positive group(n=50) and 17 deaths in the negative group(n=137) (p<0-01).

Percentage survival was correlated withtime and with the presence or absence ofsialomucin in the resection margin. Regres-sion analysis predicts 63-2% five yearsurvival for patients with no sialomucin and28-4% five year survival with sialomucinpresent. There was no significant statisticalcorrelation between the sialomucin stain-ing at the resection margin and Dukes'stage, tumour site or differentiation. Sialo-mucin production in either resection mar-gin appears to be of poor prognosticsignificance.

P109Incidence of colorectal cancer and polyps ina health care district: experience at one yearof registration

M PONZ DE LEON, A ANTONIOLI, S BONILAURI,K ARDUINI, A MERIGHI, G P RIGO, M PULVI-RENTI, G GIBERTINI, P DI DONATO, AND FMANENTI (Istituto di Patologia Medica,Cattedra di Gastroenterologia, Divisione diChirurgia generale, Universitd di Modena,Italy) From January 1984 a 'tumourregistry' for colorectal cancer and polypshas been instituted in a predominantlyurban population (263 546 inhabitants) ofnorthern Italy. Based on the experience ofthe first year of registration, the purposesof this report were three-fold: (1) to deter-mine the incidence of colorectal tumours in

A1129


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


a rather homogeneous population; (2) tofind out possible associated risk factors; (3)to ascertain if the data from registrationcould fit into the 'polyp-cancer theory'.The observed incidence of colorectal

cancer was 52-8 new cases/100 000/1984(53X4 in men, 52-2 in women). The inci-dence of polyps was 59-6. cases (83.4 inmen, 37-3 in women). The incidence in-creased with age either for cancer or forpolyps; however, polyps were more fre-quent than cancer until the sixties and thepeak of incidence of polyps anticipated thatof cancer by a five years period. Bothcancer and polyps had a similar distributionin the large bowel, more than 60% beinglocated in the left distal portion. No dietary(including alcohol) or occupational factorwas clearly associated with colorectalcancer or polyps. Smoking was more fre-quent in patients with polyps than in thosewith cancer (46-6% vs 27-6%, X2=5-61,p<001), presumably because of the malepreponderance in the former group.We conclude that the incidence of colo-

rectal cancer in northern Italy is higherthan that of many other european countriesand comparable to that observed, on aver-age, in the United States. No obvious riskfactor could be detected during the firstyear of registration. The earlier rise ofincidence of polyps - as compared tocancer - and the similar distribution in thevarious tracts of the large bowel lendfurther support to the 'polyp-cancer sequ-ence'.

P1 10Prospective comparison of ultrasonic scan-ning and static and dynamic isotope imag-ing in the pre-operative diagnosis of livermetastases from colorectal cancer

P J FINAN, S H LEVESON, G R GILES, P AWIGGINS, P J ROBINSON, AND H IRVING(Departments of Surgery and DiagnosticImaging, St James's University Hospital,Leeds) The pre-operative detection ofhepatic metastases from colorectal canceris of importance for accurate staging andplanning of adjuvant therapy. Assessmenthas relied previously on a variety of imag-ing techniques including static isotopescanning ultrasound and computerisedtomography. Isotopically measured liverblood flow ratios have previously beendescribed for the detection of hepaticmetastases. We wish to present a prospec-tive comparison of ultrasonic imagingtogether with static and dynamic scintigra-phy in patients undergoing surgery for

colorectal cancer; the results being corre-lated with the findings at subsequent lapar-otomy.

Forty five patients have been fully evalu-ated of whom 10 had metastatic disease atlaparotomy. The sensitivities for ultra-sound, static and dynamic scintigraphywere 50%, 50%, and 70% respectively.The specificity for these three methods was91%, 100%, and 66%. When either staticscanning or ultrasound was used in com-bination with dynamic scintigraphy, thesensitivity was increased from the original50% to 80%.

It is known that at least 30% of patientswith no evidence of liver metastases atsurgery will subsequently develop overtdisease and it has been shown that dynamicscintigraphy will identify patients harbour-ing occult metastases. Based on this studyit is apparent that a combination of ultra-sound or static isotopic imaging used incombination with flow scintigraphy in-creases the accuracy of either of thesediagnostic procedures in the detection ofmetastatic colorectal disease.

PillDiagnosis and surgical management of in-tractable constipation

A M ROE, D C C BARTOLO, AND N J MCCMORTENSEN (University Department ofSurgery, Bristol Royal Infirmary, Bristol)Intractable constipation presents a clinicalproblem in diagnosis and management. Wehave investigated 52 patients using a pro-tocol involving transit studies, manometry(anal sphincter pressures, rectal com-pliance and rectosigmoid motility), EMGmeasurements of volitional pelvic flooractivity, and proctographic techniques inorder to identify those patients who may behelped by surgery. Twenty eight patients(27 women, one man) had slow transit andseven patients have had nine procedures:two had internal sphincterotomy withoutobjective improvement and subsequentlyhad a colectomy. A further four had acolectomy and ileorectal anastomosis(IRA) with satisfactory results. One had apuborectalis division. Twenty four patients(17 women, seven men) had normal transitbut symptoms of obstructed defaecation.Dynamic proctograms showed rectal intus-susception in five, treated by rectopexywith good results in four. One patient had arectopexy and IRA and one an excision ofanterior mucosal prolapse. A combinationof transit studies and a dynamic procto-gram were the most helpful investigations.

Manometry and EMG studies did not vieldany additional information, althoughsphincter manometry was important whensurgery was considered. We have sug-gested an algorithm for the investigationand management of these difficult patients.

P112Anorectal myectomy; a valuable treatmentfor chronic constipation

N D HEATON AND E R HOWARI) (Kin'g .sCollege Hospital, Denmark Hill, London)Partial excision of smooth musclc from theinternal anal sphincter and rectum (anorec-tal myectomy) has been suggested as atechnique for the management ot seerechronic constipation.We used a modified technique in the

treatment of 53 new cases who had failed torespond to treatment with laxatives andanal stretch, (age range two months to 66years, mean 8-5 years; men 31, women 23).There were no clinical features of congenit-al aganglionosis (Hirschsprung's disease) inthese patients.The resected strips of smooth muscle

were assessed with neurohistochemicaltechniques, using acid phosphatase stainingfor ganglia, and assessing nerves for acetvl-cholinesterase activity and catecholaminefluorescence. Significant ahnormalitieswere detected in 34 cases - hypogangliono-sis (32), hyperganglionosis (one), andanganglionosis (one). There were minorabnormalities in five cases. Fourteenshowed a normal pattern of innervation.

Postoperative assessment in 48 patients(follow up two months to five years; mean14 months) showed an excellent result in 26(54%) and a significant improvement ineight (17%). A poor result was found in 14patients (29%) and eight of these haxesubsequently undergone a large howel re-section.

Anorectal myectomy is valuahle in thediagnosis of neuronal disorders of thehindgut, and is a useful therapeutic man-oeuvre in a significant number of patients.

P113Impaired recruitment of the pelvic floormusculature by intra-abdominal pressurein faecal incontinence

N R WOMACK, J F B MORRISON, AND N SWILLIAMS (University Departments ofSurgery and Physiology, The General Infir-mary, Leeds) Patients with idiopathic

A1 130


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


faecal incontinence (IFI) often report ex-acerbation of symptoms with activity thatraises intra-abdominal pressure. Normallyneural reflexes generate a compensatoryincrease in the activity of the pelvic di-aphragm when intra-abdominal pressure(IAP) rises. To assess the integrity of thesereflexes in incontinence 15 patients withIFI (12 women, three men, age 62±14years) were compared with 15 age and sexmatched normal subjects (12 women,three men, mean age 59+17 years; p=NS).

Intra-abdominal pressure, measured viaan intrarectal balloon, was raised by aseries of forced expirations to graded pre-set levels. Activity in the puborectalismuscle was recorded using a concentricEMG needle, and the signal rectified andintegrated. In each subject there was apositive linear correlation between muscleactivity and IAP (R=0.67-0-98, median0-88). Extrapolation of the regression lineallowed measurement of the rise in IAPthat was necessary to increase muscleactivity. Since these muscles are tonicallyactive this increase in activity depends onthe sensory side of the reflex mechanism.In the control group recruitment occurredat 0-2±2-0 cm water, whereas in the IFIgroup recruitment required a significantlyhigher rise in IAP of 7-6±5-0 cm water(p<0.01 Mann Whitney U test). Thisdeficient response of the pelvic floor mus-cles to raised IAP in IFI explains theincrease of symptoms with activity. It alsosuggests that IFI may not be entirely motorin origin, as has been postulated, there alsobeing a deficient sensory input.

P1 14Experimental support for a simple model ofdefecation

J J BANNISTER, C P GIBBONS, E A TROWBRIDGE,AND N W READ (Departments of Surgeryand Medical Physics, Royal HallamshireHospital, Glossop Road, Sheffield) Asimple model of the forces involved in thepassage of a stool, through the anal canal,predicts that small stools will be moredifficult to pass than large stools and that alinear relationship will be found betweenthe intrarectal pressure required to pass astool and the reciprocal of the radius of thestool. To test this model, in sevenvolunteers (four men, three women) theintrarectal pressure was measured whilstthe subjects passed five incompressiblespheres of known diameter. The smallestsphere (diameter 0-5 cm) was passed by

only two subjects, the largest (diameter 2-0cm) by all seven. Analysis of the pressuretraces for the spheres, that were passed,shows that the intrarectal pressure and thetime required to pass the largest spherewere significantly less than those requiredto pass the smallest (diameter 0-5 cm or 1-0cm) (p<0.01). Plotting the peak intrarectalpressure, to pass each sphere, against thereciprocal of the radius gave linear correla-tion coefficients of 0-75-0-98 (median0.82).These findings support our simple analy-

sis of defecation and confirm the clinicalimpression that the size of the stool is animportant contributory factor in difficultiesin defecation.

P1 15Urological abnormalities in patients withslow transit constipation

J J BANNISTER, W T LAWRENCE, D G THOMAS,AND N W READ (Department of Surgery,Royal Hallamshire Hospital, Sheffield, andDepartment of Urology, Lodge Moor Hos-pital, Sheffield) In a study of 24 youngwomen with severe slow transit constipa-tion, it was noted that a surprisingly highnumber (75%) had symptoms of urinarydysfunction, including hesitancy, frequen-cy, urgency, stress incontinence and symp-toms of urinary infection. In 10 patients acombined radiological and manometric,urodynamic and anorectal evaluation wasperformed. The urodynamic assessmentshowed that all the subjects required blad-der volumes larger than the normal range(150-300 ml) to stimulate the first desire tomicturate (p<001) and their bladdercapacity exceeded the normal range (300-500 ml) in all except one subject (p<0-01).Detrusor contraction was normal with noevidence of instability, the rate of micturi-tion was normal with no radiological evi-dence of impaired bladder neck opening.The anorectal studies revealed that theconstipated subjects required larger rectalvolumes than normal to stimulate thedesire to defecate (200±50 ml vs 110±10ml (Mean±SEM); p<0.05) and that theirmaximum tolerated volume was also larger(380±30 ml vs 290±20 ml (Mean±SEM);p<0-05), other modalities of rectal sensa-tion and compliance were normal. Thesimilar findings in the bladder and anorec-tum suggest an analogous sensory defect inboth organs, raising the possibility of acommon abnormality in their extrinsic in-nervation.

P116Investigation of colonic motility patterns inthe irritable bowel syndrome using radio-telemetry

J R REYNOLDS, A G CLARK, D F EVANS, AND J DHARDCASTLE (Department ofSurgery, Uni-versity Hospital, Nottingham) An un-tethered pressure sensitive radiotelemetrycapsule and portable receiving apparatushave been used to measure colonic motilityfor 24 hours in ambulatory subjects. Twen-ty one patients with irritable bowel syn-drome (mean age±SD 34-7±9-8 years)were compared with 10 healthy volunteers(24-8±5.67 years). All patients complainedof abdominal pain with constipation in nineand diarrhoea in 12. Data were analysedusing a microcomputer and a motility indexderived for three periods: interfood (IF),postprandial (PP) and night (N). Signifi-cant differences between IF, PP and Nwere found in controls (7-520.91,13-14±1-47 and 1-86±0-66 p<O.01) andpatient group (10-65±1-35, 18-16±2-04,and 6-01± 108 p<002). Night time activityin the patient group (6-01±1-08) was signi-ficantly higher than control (p<0-05) andthere was a similar trend, though notsignificant, in the interfood periods. Post-prandial activity was significantly greater inpatients with diarrhoea (20-52±3.01p<O-05) whereas no difference was foundin the PP response in the constipationgroup (15-54±2-59). This non-invasivemethod has identified a sub-group of 30%of patients who have periods of relativecolonic hypermotility studied in their nor-mal environment and subject to usual dailystress.

P117Management and differential diagnosis ofperianal hidradenitis

B J HARRISON AND L E HUGHES (Departmentof Surgery, University of Wales College ofMedicine, Cardiff) Perianal skin is proneto involvement with hidradenitis suppur-ativa and is characterised by the develop-ment of recurrent skin nodules whichprogress to multiple chronically discharg-ing sinuses. Natal cleft and buttock skinmay be affected, superficial fistulae involv-ing the distal anal canal have been de-scribed.

Since 1979, 15 patients with perianalhidradenitis in whom conservative mea-sures had failed required wide excision ofperianal skin. In all cases active diseaseinvolving the axillae and/or puboinguinal

A1131


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


region was present. The wounds wereallowed to heal by granulation with the aidof silastic foam dressings and satisfactoryresults were obtained in all cases. Followup of 12 patients (mean 2-5 years) hasrevealed no evidence of recurrent disease.Careful examination usually differenti-

ates hidradenitis from other perianaldisease. Anorectal Crohn's and perianalfistulae are conditions commonly causingdiagnostic difficulty. In the former, cavitat-ing ulcers are found within the anal canal,in the latter, inflammation rarely spreadsto involve the buttocks. Less commonproblems of diagnosis include primary pe-rianal pilonidal sinus, postnatal pilonidalsinus, Paget's disease, Bowen's disease andcolloid carcinoma.A normal anal canal and hidradenitis at

other sites are the most important clinicalfeatures in diagnosis.

PANCREATICO-BILIARYP118-133

P118Are bile acids involved in the regulation ofmouth-to-caecum transit time (MCTT) inman?

R PENAGINI, R C SPILLER, D B A SILK, AND J JMISIEWICZ (Department of Gastroenterol-ogy and Nutrition, Central Middlesex Hos-pital, London) Bile acids can affect waterand electrolyte movements in the humanjejunum and ileum. After cholecystec-tomy, the postprandial peak luminal andblood concentration of bile acids ismarkedly lowered.To test if this could delay intestinal

transit we studied 10 patients two to fourmonths after cholecystectomy (C) andcompared them with 10 healthy controls(N). MC1T was measured in all subjectsafter a standard liquid 440 kcal (1842 kJ)meal containing lactulose 15 g, using thehydrogen breath test. Blood samples werecollected at 0, 30, 60, 120, 180, 240 min (8C and 8 N) for measurements of total bileacids (enzymatic method) and cholylgly-cine (RIA); peak incremental response(PIR) and area under the curve (AUC)were calculated. Analysis of results wasdone with the Mann-Whitney U test.MCiTl (mean±SEM) was 65-0±7-6 min inN vs 41*0±5-3 in C (p<005). Cholylgly-cine PIR was 2-17±0-48 ,umol/l in N vs1-11±0-23 in C (p<005) and its AUC

306-15±53-7 in N vs 204-12±36 1 in C(p<002), while total bile acids PIR was17-2±1-91 ,umol/l in N vs 10-75±1-1 in C(p=0025) and the AUC 2328-86±328-26in N vs 2893-39±208-77 in C (p=ns).

Contrary to our hypothesis patients aftercholecystectomy appear to have a fasterMCTI than healthy controls. As AUC oftotal bile acids was the same in C and Nwith lower AUC for cholylglycine in C,higher concentrations of dihydroxy bileacids may be the responsible factor. RapidMCTT may be relevant to the pathogenesisof post-cholecystectomy diarrhoea.

P1 19Morphological and biochemical studies onrat pancreatic ducts maintained in tissueculture

S ARKLE AND B E ARGENT (INTRODUCED BY AALLEN) (Department of Physiological Sci-ences, University Medical School, New-castle Upon Tyne) Recently, we havedeveloped a technique for the isolation ofviable small interlobular ducts from thepancreas of copper deficient rats. Copperdeficiency causes a non-inflammatory atro-phy of pancreatic acinar cells while the ductcells remain structurally and functionallyintact. When isolated ducts were main-tained in tissue culture their cut ends sealedwithin eight hours. This process wasaccompanied by an overall swelling of theduct, a marked dilatation of the lumen anda flattening of the epithelium against thesurrounding connective tissue layer. Secre-tin (0.1 nmol to 1 ,Lmol) caused a doserelated increase in cyclic AMP contentand, usually, further swelling of the ducts.Puncture of the cultured ducts caused a fallin duct size, largely accounted for by areduction in lumen volume, and also in-creased the height of the epithelium.Taken together these observations sug-

gest that the duct swelling which occursduring maintenance in culture is because ofan increased luminal pressure, resultingfrom fluid secretion into the closed luminalspace.

P120Cholesterol absorption by the human gallbladder

M R JACYNA, P E ROSS, D HOPWOOD, AND I A DBOUCHIER (Department of Medicine,Ninewells Hospital and Medical School,Dundee, Scotland) Although guinea piggall bladder is known to absorb significant

amounts of luminal cholesterol, there areno data currently available relating to thehuman gall bladder. Consequently a modi-fied Ussing Chamber was used to investi-gate cholesterol absorption and transport inhuman gall bladder mucosa using artificialbiles of varying cholesterol saturation. Up-take and transport of cholesterol was deter-mined by tissue and serosal fluid content of1-('4C)-cholesterol used as tracer in theartificial bile solutions while (3H)-Dextranwas used to correct for adherent bile on themucosal surface. Tissue viability duringexperiments was established by using elec-tron microscopy to confirm ultrastructuralintegrity and also by serial measurementsof transmural potential difference and dif-fusion potentials.

In 33 human gall bladders studied so far,cholesterol was absorbed from the lumenand transported into the serosal fluid.During this process, almost 15% of theabsorbed cholesterol was esterified.Absorption and transport rates increasedwith increasing cholesterol saturation,reaching a maximum value (of approxi-mately 3-5 nmol/cm2/min) when bile be-came supersaturated. These results showabsorption and esterificiation of biliarycholesterol which may be important in thepathogenesis of human gall bladder dis-ease.

P121Molecular forms of cholecystokinin inhuman plasma after ingestion of fat

J B M J JANSEN AND C B H W LAMERS(Departments of Gastroenterology-Hepatology, Universities of Nijmegen andLeiden, The Netherlands) There is sub-stantial controversy about the nature ofcirculating CCK in response to meal stimu-lation. We therefore studied the molecularforms of CCK in basal plasma and 20minutes after ingestion of 250 ml 20%Intralipid by region-specific radioimmuno-assays. CCK-like immunoreactivity (CCK-LI) was concentrated from 10 ml plasmasamples from three healthy volunteers byC18 SEP-PAK cartridges eluted with amixture of 80% acetonitril-20% 1%-trifluoroacetic acid. In addition, 5 ml plas-ma samples from three gastrectomisedpatients with high plasma CCK responsesto oral fat were extracted with ethanol. Allsamples were evaporated to dryness, dis-solved in 1 ml elution buffer and applied toSephadex G-50 columns. The eluates weremeasured by radioimmunoassay using anti-

A1132


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


body T204 (directed to the sulphatedtyrosyl region of CCK) and antibody 1703(binding to C-terminal CCK-peptides con-taining at least 14 amino acid residues). NoCCK-LI was demonstrable in fractionatedbasal plasma. Chromatography revealedfour peaks in fat stimulated plasma. Peak Ieluted in the void volume and comprised1-10% of CCK-LI, peak II eluted at 35%and comprised 3-35%, peak III eluted at50% and comprised 24-62%, and peak IVeluted at 75% and comprised 15-45% ofCCK-LI. No additional peaks were de-tected by antibody 5135, directed to theC-terminus of CCK and gastrin.We conclude that CCK in plasma after

fat stimulation is heterogeneous. From thecross reactivity pattern it is also concludedthat the small form of CCK is differentfrom CCK-8.

P122Cholecystokinin octapeptide-like material isproduced from larger forms during circula-tion in man

C J SPRINGER AND J CALAM (HammersmithHospital, London) Cholecystokinin inplasma exists in forms containing 58, 39(CCK39), 33 (CCK33), and eight (CCK8)amino acid residues but the relationshipsbetween different forms is ill understood.In this study we have shown that CCK8-like material is formed from CCK33/39during circulation.

Six fasted normal volunteers receivedintravenous natural porcine CCK (KabiDiagnostics) at a measured rate of 0-7-2-5pmol/kg/min for 15 minutes. CCK formswere separated on Sephadex-G50 and mea-sured by C-terminal specific radioimmu-noassay. CCK in infusates eluted chiefly inthe characteristic position of CCK33 andCCK39 (CCK33/39). A second peak, elut-ing between CCK33 and CCK8, accountedfor 0-15% and CCK8-like materialaccounted for less than 3% of immunoreac-tivity. During infusions CCK33/39-like im-munoreactivity appeared in plasma androse to a mean concentration of 49 pmol/lat the end of the infusion. In addition,CCK8-like material was detected during allinfusions, rising about 3 minutes later thanCCK33/39 to a mean concentration of 48pmol/l at the end of the infusion. A steadystate was not achieved but both formsbecame undetectable (<10 pmol/l) 11minutes after the infusion. The CCK-8-likematerial seen in normal human plasma maybe the product of post secretory cleavage oflarger forms.

P123Methods to assess 'enzyme induction' inpatients with idiopathic pancreatitis

L N SANDLE, D W K ACHESON, L P HUNT, A HGOWENLOCK, AND J M BRAGANZA (Depart-ments of Biochemistry and Gastroenter-ology, and Faculty of Medicine Computa-tion Group, Royal Infirmary, Manchester)Recent studies suggest that pancreatitismay be facilitated by an imbalance be-tween enzyme induction and available anti-oxidants. There is a need to assess thisinduction/antioxidant axis. We have com-pared the information on induction pro-vided by various tests done in the sameweek on 11 patients with idiopathic recur-rent pancreatitis, without overt liver dis-ease. The last attack of pain in thesepatients occurred between one week and12 months before the tests: the time sincetheir first symptoms varied widely, be-tween two months and 13 years.The group displayed: (1) rapid theophyl-

line clearance, indicating induction of cyto-chromes P450 (median 179, vs 74 ml/kg/hin controls, p<0.005); (2) acceleratedearly-phase disappearance of sulphobro-mophthalein, suggesting induction of ligan-din (mean k, 23-0, vs 14-3%/min in con-trols, p<0-025); (3) increased D-glucaricacid excretion in 'spot' samples of urine,reflecting heightened 'phase II' reactions(median 3.7, vs 2-9 mmol/mol creatinine incontrols, p<0.025), (4) increased post-secretin bilirubin output, suggesting induc-tion of haem oxygenase as well as 'phase II'reactions (median 8363, vs 4000 u/30 min incontrols, p<0.025).Theophylline clearancewas increased in nine of the i1 patients:none of the other induction 'markers'achieved this sensitivity.The theophylline test is thus the obvious

choice to assess the induction componentof the induction/antioxidant axis in pan-creatic disease.

P124Modified sham feeding induces gall bladdercontraction by an atropine-sensitive, CCK-independent mechanism

W P M HOPMAN, G M P HOUBEN, M C A

VERMEULEN, J B J M JANSEN, G ROSENBUSCH,AND C B H W LAMERS (Departments ofGastroenterology-Hepatology, and Depart-ment of Radiology, Universities of Nij-megen and Leiden, The Netherlands) Thisstudy was undertaken to determinewhether modified sham feeding inducesgall bladder contraction, to assess the

relative contribution of the cephalic phaseto postprandial gall bladder contraction,and to elucidate the mechanism of cephalicstimulation of gall bladder contraction. Onseparate mornings eight fasting healthyvolunteers (four men, four women, 20-65years) underwent the following studies:sham feeding during 30 minutes, shamfeeding after atropine (0-015 mg/kg asbolus followed by infusion of 0*005 mg/kg/h), and ingestion of the same meal in 30minutes. Gall bladder volumes were mea-sured by ultrasonography and plasma CCKby a sensitive and specific radioimmunoas-say. Sham feeding induced a significant gallbladder contraction of 33±4% of theoriginal volume (p<0.0005). After inges-tion of the meal gall bladder contractionwas 67±47% (p=0-00001). Gall bladdercontraction after modified sham feedingwas abolished by atropine. Plasma CCKconcentrations were not affected by shamfeeding, whereas ingestion of the mealincreased plasma CCK from 2-3±0-6 to4-6±1-1 pM (p<0-001).We conclude that modified sham feeding

induces gall bladder contraction by anatropine-sensitive, CCK-independentmechanism. The extent of gall bladdercontraction after modified sham feedingwas about half of that after ingestion of themeal.

P125Diagnosis and management of pancreaticduct injuries in children: a report on fourcases

R I HALL, M I LAVELLE, AND C W VENABLES(Departments of Surgery and Radiology,Freeman Hospital, Newcastle upon Tyne)Major pancreatic injuries in children aredifficult to diagnose and may go unrecog-nised for very long periods. Of crucialimportance is the identification of thosepatients with damaged pancreatic ducts,because they require urgent surgical treat-ment. We have encountered four children,aged three to 13 years, with major pancrea-tic injuries. Although three had sustainedthe type of bicycle and sledging accidentsassociated with pancreatic injury, the di-agnosis was delayed in all children forperiods of up to one year. All childrencomplained of persistent abdominal painand vomiting. Traumatic pancreatitis wassuggested by raised serum amylase concen-trations in each case. Ultrasound scansidentified pseudocysts in three, althoughthe location was incorrect in one. Com-puted tomography scanning suggested a

A1133


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


lacerated pancreas in one, but was unhelp-ful in another. ERCP was performed undergeneral anaesthetic in all patients. Thepresence and location of the duct injurywas correctly identified in every case.Distal injuries (two) were treated by distalpancreatectomy, proximal lacerations(two) were internally drained into Roux-en-Y loops. All patients recovered. ERCPshould be performed in all children withtraumatic pancreatitis and any duct injuryidentified must be treated by urgentsurgery.

P126Evidence for an inhibitory effect of bombe-sin on pancreatic polypeptide secretion inman

A J L DE JONG, J P M BLAAUWHOF, M C AVERMEULEN, AND C B H W LAMERS (Depart-ments of Gastroenterology-Hepatology,Universities of Nijmegen and Leiden, TheNetherlands) Bombesin containing nerveshave been demonstrated in the pancreas.Because infusion of bombesin stimulatesthe secretion of pancreatic polypeptide(PP), it is possible that bombesin is in-volved in postprandial PP release. Todetermine the interactions betweenbombesin and food in the regulation ofpostprandial PP secretion, we have studiedthe effect of increasing doses of bombesin(1, 2-5 and 5 ng/kg/min for two hours) onpostprandial PP secretion in six healthyvolunteers (four men, two women, 20-26years). Thirty minutes after the start of thebombesin infusion the subjects ingested aliquid test meal. As expected the mealinduced significant increases in plasma PPfrom 17±3 to 47±8 pM (p<0-01). Infusionof bombesin did not stimulate, but ratherinhibited and even abolished postprandialPP secretion in a dose-related manner.Integrated postprandial plasma PP secre-tion during saline infusion was 1850±362pM/90 min, 1245±542 pM/90 min (ns)during 1 ng/kg/min, 668±218 pM/90 min(p<005) during 2-5 ng/kg/min, and-426±478 pM/90 min (p<0-01) during 5ng/kg/min bombesin. Infusion of the threedoses of bombesin without the meal in-duced small, non-significant increases inplasma PP. After stopping the bombesininfusion, however, a steep, dose-dependent increase in plasma PP wasfound (13±5 pM after 1 ng/kg/min, 29±7pM after 2-5 ng/kg/min bombesin;p<O.05).The findings of the present study suggest

that bombesin stimulates the secretion of

an inhibitor of PP release. The nature ofthis inhibitor is at the present time un-known.

P127Partial purification of pancreotone

A A HARPER, A J C HOOD, J MUSHENS, J R SMY,C SNELL, P SNELL, AND R K VEITCH(Department of Pharmacology, SunderlandPolytechnic, Medical School, Newcstle Uni-versity, and MRC NeuroendocrinologyUnit, Newcastle General Hospital, New-castle) Pancreotone is an aqueous alcoholpeptide extract of distal but mucosa pre-pared by precipitation onto bile salts at pH3-9. Pancreotone iv in chloralose-anaesthetised cats inhibits secretion of pan-creatic juice (antisecretin effect) and gas-tric pepsin secretion. In cats and guineapigs gall bladder contractility is also inhi-bited (anticholecystokinin effect).

Reprecipitation of crude pancreotonefrom aqueous alcohol at pH 4-1 followedby extraction with acid alcohol and precipi-tation with acetone, increased the specificactivity six-fold and reduced the bile saltcontent by 85%. When these partly puri-fied preparations were subjected to isoelec-tric focusing with ampholines in Ultrodexgel, two peaks of antisecretin activity wererevealed. One at pl 8-1-8-3 containedapproximately 2/3 of the recovered activity,and the other at pl 48-5-1 contained inaddition all the recovered anti-cholecystokinin activity. Both fractions in-hibited pepsin secretion. The material at pI8-1-8-3 cross reacted strongly with anti-bodies to PYY (S R Bloom, personalcommunication). Furthermore, when in-jected iv into cats it produced a smallpressor response, whereas the pI 48-5-1material produced a small depressor re-sponse.The activities at p14-.85-1 do not co-

incide with those of characterised peptidesof the distal gut and include all the effectsof pancreotone.

P128Preliminary evaluation of a modified PABAtest

W H BRADBURY, A R W FORREST, C DHOLDSWORTH, A ROB, AND J R WORTERS(Department of Clinical Chemistry andGastrointestinal Unit, Royal HallamshireHospital, Sheffield) The combined BTPABA/14C test is accepted as a specific andsensitive test of pancreatic function. The

usual analytic method for urinary PABAis, however, subject to drug interference.We have investigated anthranilic acid, theorthoisomer of PABA as a substitute for'4C PABA in the BT PABA/'4C test.Anthranilic acid is absorbed and metabol-ised independently of pancreatic chymot-rypsin.The test has been evaluated in normal

volunteers, patients with known pancreaticsteatorrhoea and patients with suspectedpancreatic disease.

After fasting, BT PABA 1 g and anthra-nilic acid 340 mg was taken with water.Urine was collected for six hours and bothisomers were measured by high perform-ance liquid chromatography (HPLC) usinga previously unpublished method. Anexcretion ratio (% PABA recovery/%anthranilic acid recovery) was calculated.The test clearly distinguished between

normal subjects (excretion ratio 0-63-1-14)and patients with known pancreaticsteatorrhoea (excretion ratio 0.13-0-43).Patients found to have no evidence ofpancreatic disease had excretion ratios inthe range of the normal subjects. Thosewith equivocal Lundh tests had excretionratios of 0-467-0-837. This modified testmay become a useful alternative to the BTPABA/'4C test, and eliminates problemswith both drug interference and the ad-ministration and counting of '4C.

P129Absorption of omeprazole in Zollinger-Ellison syndrome is accelerated by alkali

C B H W LAMERS, L TEUNISSEN, AND J B M JJANSEN (Departments of Gastroenterologyand Hepatology, Universities of Nijmege-nand Leiden, The Netherlands) Omepra-zole is a potent inhibitor of gastric acidsecretion in patients with Zollinger-Ellisonsyndrome. As omeprazole is inactivated byacid, an enteric-coated preparation hasbeen developed from which the drug isreleased only when the pH is greater thansix. About half of patients with Zollinger-Ellison syndrome, however, show low in-hibition of gastric acid secretion in the firsthours after ingestion of the drug. We havemeasured the absorption of 80 mg omepra-zole ingested either as enteric-coated pre-paration together with 250 ml saline or asenteric-coated preparation together with250 ml (40 mmol) sodium bicarbonate or asuncoated preparation with sodium bicar-bonate in six patients with Zollinger-Ellison syndrome (three men, 35 women;35-63 years). Plasma samples for omepra-

A1 134


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


zole were obtained at regular intervals forthree hours. There was a wide variation inomeprazole absorption with the enteric-coated preparation and saline (AUC 00-18-8, median 0-6 ,.mol/l/3h). Ingestion ofenteric-coated omeprazole with sodiumbicarbonate resulted in significantly greaterabsorption (AUC 2-1-23-1, median 14-3,umol/l/3h; p<0.05). Omeprazole absorp-tion after uncoated omeprazole withsodium bicarbonate (1.4-31-2, median 4-3,umol/l/3h) was slightly greater than thatwith the enteric-coated preparation andsaline (p<0.10), but not significantly diffe-rent from that with enteric-coated omepra-zole and sodium bicarbonate. The acidinhibitory effect of omeprazole was relatedto the absorption of the drug. In tests withan AUC of more than 0.2 ,umol/l/h acidinhibition was greater than 80%, whereasin tests with an AUC of less than 0-2mmol/l/h inhibition ws smaller than 30%.We therefore conclude that addition of

alkali accelerates absorption of omeprazolein patients with Zollinger-Ellison syn-drome resulting in early inhibition of acidsecretion.

P130Ultrasonography in choledocholithiasis - anew look

M TOBIN, R M MENDELSON, G LAMB, AND I T

GILMORE (Gastroenterology Unit and De-partment of Radiology, Royal LiverpoolHospital, Liverpool) Ultrasound (US)diagnosis of common duct stones has beendisappointing. To assess its accuracy in aspecialist centre and its role as a screeningtest in selecting patients requiring directcholangiography, US was done by one oftwo experienced operators immediatelybefore endoscopic cholangiography (ERC)in 104 consecutive patients referred forsuspected biliary disease.Of 36 patients with choledocholithiasis

on ERC, US showed stones in 20, commonduct dilatation in a further 10, and failed tovisualise the common duct in five. Thesensitivity was 64*5% and 55-5% for calcu-lus visualisation and 97% and 83% for anabnormal common duct, excluding andincluding technical failures respectively. Of39 normal cholangiograms, US agreed in37 (95%), was technically unsatisfactory inone, and incorrectly diagnosed a calculus inone. One further false positive US forcalculus occurred in 16 patients showingERC dilatation only. Using US visualisa-tion of common duct calculus, dilatation ortechnical failure as criteria for proceeding

to ERC would have resulted in the omis-sion of only one patient with calculus and14 'unnecessary' cholangiograms. Ultra-sound is a sensitive screening test insuspected choledocholithiasis. A normalUS virtually excluded choledocholithiasisin this series.

P131Electrohydraulic lithotripsy of gall stones,and in vitro study

J D HARRISON, D L MORRIS, JULIE HAYNES,AND D C WHERRY (Department of Surgery,University Hospital, Nottingham) Elec-trohydraulic lithotripsy is widely used inthe endoscopic management of urinarytract calculi. Its role in the management ofgall stones has received little attention.An electrohydraulic generator and probe

(ACM) with variable voltage, plus lengthand number of pulses was used to fragment12 mixed pigment/cholesterol gall stonesrecently removed from five patients in0-9% saline. Two stones (6x7, 5x4 mmdiameter) were fragmented at 60 volts (V),two required 80 V (9x7, 7x5 mm), twowere fragmented at 100 V (13x 12, 12x13mm) while two were not affected until 120V was used (12x8, 15x13 mm). In onlyone stone did we fail to achieve fragmenta-tion (28x 21 mm). Three stones (7x 5, 5x 6,and 7x6 mm) were then fragmented withinhuman and ovine bile ducts at 60, 60 and120 V without obvious macroscopic evi-dence of damage to the ducts.

Direct contact of the end of the probewith bile duct or gall bladder wall produceda perforation even at 60 V, but both wereresilient to shock waves (rather than sparkeffect).Most gall stones may be fragmented by

the electrohydraulic lithotripter. This in-strument may be useful to both endoscopistand surgeon for impacted bile duct calculi.

P132Results of endoscopic stenting in malignantstricture of the biliary tract

J STOKER, J DEES, M VAN BLANKENSTEIN, ANDG A J J NIX (Departments of InternalMedicine and Radiology, University Hos-pital Rotterdam, Rotterdam, The Nether-lands) In 1983-1984 endoscopic place-ment of a biliary stent (EPS) by theHuibregtse-technique was attempted in 123patients suffering from malignant stricture.The results were assessed in relation to thelocalisaiton of the stricture, a comparison

between common duct and perihilar stric-tures being found to be relevant. Stentingwas successful in 97 cases (79%). In com-mon duct strictures the failure rate was11*8%. After successful stenting early com-plications were seen in 10% and there wasa satisfactory drainage in 94%. In hilartumours these percentages were 42, 29 and73 respectively. Early complications in-cluded haemorrhage, requiring transfusionin five patients and cholangitis in 11. Meansurvival in 79 patients who were onlytreated by stent drainage was 18-6 weeks.Late complications were recurrent jaun-dice in 33 patients, and fever in fourpatients. Only eight patients were referredfor stent replacement.We conclude that (1) EPS is an effective

and safe treatment for malignant stricturesof the common bile duct. It is less success-ful in hilar strictures which may be moreamenable to percutanous drainage. (2)Patients and their doctors should beallerted to the necessity of prompt stentreplacement if fever or jaundice recur.

P133Randomised trial comparing endoscopicand percutaneous prostheses in poor riskpatients with malignant obstructive jaun-dice

A G SPEER, P B COTrON, A HATFIELD, R R

MASON, R C G RUSSELL, J LEUNG, T P YIN, J

LENNARD-JONES, J BAILLIE, AND K MCCRAE(Departments of Gastroenterology, Radiol-ogy and Surgery, The Middlesex and Lon-don Hospitals; Cancer Research CampaignCentre, King's College Hospital, London)Bypass surgery for palliation of biliaryobstruction due to malignancy has a highmortality in poor risk patients. We havecompared the insertion of a prosthesis viathe percutaneous transhepatic method(PTE) or endoscopically (EP) in a prospec-tive randomised trial. Seventy five patientswith primary tumours (pancreatic and cho-langio-carcinoma) who were judged to beunsuitable for surgery were entered.Analysis was according to a sequentialblock design. The results are presented ona strict intention to treat basis. The twogroups of patients were well matched apartfrom the incidence of hilar stricture - EP -

49%; PTE - 28%.Criterion for significance (p.0-016) was

achieved for 30 day mortality using logrank analysis stratified according to site oflesion. The significantly increased earlymortality and risk of complications of PTEwere related to puncturing the liver -

A1135


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


haemorrhage, biliary peritonitis and abs-cess formation. The endoscopic method isnow the preferred route for insertion ofbiliary prostheses. The relative merits ofprostheses and bypass surgery in fitterpatients are being compared in anothertrial.

ENDOSCOPYP134-135

P134Balloons against bougies for dilatation ofbenign oesophageal stricture - a random-ised prospective trial

J C COX, R K WINTER, R JONES, J F DYET, D RSUTTON, AND J R BENNETT (Hull RoyalInfirmary, Hull) Balloons have beenadvocated as safer and more effective thanbougies for dilating oesophageal strictures,but there is no evidence to support this.Twenty six patients, aged 30-90 years,

were randomly allocated to dilatation bybougie or balloon to a predetermineddiameter. Stricture size was measuredbefore, one week and one month after-wards using the barium sphere technique.Patients' acceptance of each technique wasassessed by a simple scoring system.

All dilatations were achieved withoutcomplications. There was no significantdifference in the mean change from theinitial measurement between the twogroups at one week or one month. In 12patients measured at three months, fourmonths, and five months after dilatationthe diameter diminished more rapidly inthe balloon group than in the bougie group(p<0-1). Two patients in each group re-quired repeat dilatation because of dys-phagia. There was no significant differencein the patients' acceptance score betweenthe two groups.

Balloon dilatation appears to be a safetreatment for benign oesophageal stric-tures, but is neither more effective normore acceptable to patients thanbougienage.

P135Endoscopic injection of adrenaline in bleed-ing peptic ulcers

J W C LEUNG AND S C S CHUNG (CombinedEndoscopy Unit, Departments of Medicine& Surgery, The Chinese University ofHongKong, Hong Kong) Early endoscopy

frequently reveals active bleeding lesions inpatients with upper gastrointestinalhaemorrhage. We report a simple methodof endoscopic haemostasis in bleedingpeptic ulcers using local injection of adren-aline. Between November 1984 and April1985, 206 patients were admitted withupper gastrointestinal haemorrhage.Twenty one patients (19 men, two women,mean age of 48 years) had an activelybleeding ulcer (17 DU, four GU) onendoscopy. Mean haemoglobin on admis-sion was 9-5 g/dl, and 15 patients requiredblood transfusion (mean=5 units, range1-12). Using a needle injector through anendoscope, alliquots (0-5 ml) of 1:10000adrenaline (volume 1-6 ml, mean=3-2 ml)were injected submucosally around thebleeding site. No complication wasobserved. Initial haemostasis was achievedin all. Three patients rebled at four hours,three days and four days later, and two(one GU, one DU) required emergencysurgery and the third (DU) improved withconservative treatment. Thus definitivehaemostasis was achieved in 19/21 patients(90-5%). Repeat endoscopy at eight weeksshowed healed ulcer in 13 patients, whilesix patients are awaiting reassessment.Local injection of bleeding ulcers withadrenaline is a simple, effective and econo-mical method for haemostasis in patientswith bleeding peptic ulcers.

PLENARYT1-6

TiDangers of surgical treatment for perianalCrohn's disease

M R B KEIGHLEY AND R N ALLAN (Depart-ment of Surgery, General Hospital, Birm-ingham) We have examined 202 consecu-tive patients with Crohn's disease to deter-mine the current status of perianal diseasetogether with the influence of surgicaltreatment on outcome. The mean durationof follow up was 7-6 years (range: 1-36years) and all except 35 patients (17%) hadundergone some form of intestinal re-section.One hundred and ten patients had evi-

dence of perianal disease at some time intheir illness (54%). The principal lesionswere skin tags (n=75), anorectal abscess(n=53), fistula (n=52), fissure (n=35) andstricture (n= 19). Seven fissures were tre-ated by dilatation but only four were

improved and one became incontinent.Twelve fistulae were laid open hut onix oneresolved and six became incontinentFifteen abscesses were drained but onl1five resolved and four developed a fistulalNone of the six strictures treated by dilat-tion were improved. Proctectomr wxas per-formed in 36 patients: 19 tf 27 withperinanal disease have a persistent perinealsinus (70%), compared with none in thenine having proctectomrn and no peritanaldisease (10%: p<001). Seventeen ot the19 patients with an unhealed perineal sinushad a rectal stricture.These results imply that local surger%

should be avoided for perilanal Crohn'sdisease and that proctectomrn fot rectaldstricture is usually associated with a pci-sistant perineal sinus.

T2Segmental colonic function in experimentalsteatorrhea-decreased capitance of theproximal colon

R C SPILLER, M I. BROWN AND) S F PItt lIlS(Mayo Clinic, Rochester, X1Iotics ohl, l rS A)In steatorrhea, long chaiti latty acids im-pair colonic absorption. howexver the im -

portance of altered colonic transit in theassociated diarrhoea is unknoxx n. Atteipassing an orocaecal manometrs tube. wehave infused the normal unprepared colon(n=6) at 2 5 ml/min with a solutioni simllLl-lating postprandial colonic inflows Insteatorrhea [oleic acid emulsion (0A). 5

g/100 ml]. Seven control subjects receixed0-9% NaCl; both solutions also contained2-5 g/l lecithin, 10 mmoll1Na taurocholitc.and 20 mmol/l glycerol (all ad"justed to pIT6-4, 290 mosmol/kg) and were labelled writh12 mCi/l of "'In DTPA. Passage ot' Infrom ascending (AC) to transverse (I(and descending (DC) colon and rccto-sigmoid (RS) were quantified hx serild1-min gamma scans.

Oleic acid emulsion induced cpisodic.prolonged (>10 sec), propagated pressuicwaves (>60 mmHg) in thc A('(mean±SEM, 75+3 mmHg. 30j+4 sec)which were associated with cramps anidmass movements of isotope caudalls. Suchwaves occurred 4.1+2.4 times/h during OAinfusion, but only once in 33 h WNith saline,p<0-001. Concomitantly. isotope overflowed from the AC earlier during OA(19±3 vs 39±6 min) and accumulated in]the DC and RS morc rapidly. reachigLl>50% infused dose bx 106r+ 14.8 inill(range 34-135), p<t)()()l is saline (all>183 min). Urgent defecation occurredl

A1136


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


after 312±20 ml of OA compared with1049±71 ml of saline (p<0-001). Thisnovel research technique, the 'isotope col-ometrogram' which quantifies regionalcolonic transit, promises new insights intothe pathophysiology of diarrhoea.

T3Randomised trial of vasopressin andvasopressin plus nitroglycerin in the controlof acute vericeal haemorrhage

A E S GIMSON, D WESTABY, J HEGARTY, A

WATSON, AND R WILLIAMS Liver Unit,King's College Hospital and School ofMedicine & Dentistry, Denmark Hill, Lon-don) The systemic haemodynamic com-plications of vasopressin (VP) have beenan important limitation to its use in man-agement of variceal haemorrhage. Recentstudies suggest that nitroglycerin (NG)may reverse these changes and augmentthe fall in portal pressure. A randomisedtrial was undertaken to compare the effi-cacy and complication rate of VP alone(0-4 units/min constant infusion (and VP +intravenous NG (40-400 ,g/min to main-tain systolic blood pressure > 100 mmHg)for a period of 12 hours. Seventy twobleeding episodes in 57 patients were in-cluded: VP alone on 34 and VP + NG on38 occasions. At the end of the 12 hourperiod haemorrhage had stopped signifi-cantly more frequently in the VP + NGgroup (26 of 38: 68%) compared with theVP group (15 of 34: 44%, p<0-05). Majorcomplications requiring cessation oftherapy were significantly less common inthe VP + NG group than in the VP alonegroup (one and seven respectively,p<002). Hospital mortality was similar inboth groups. In conclusion, the addition ofNG to a VP infusion significantly reducesthe complication rate and has been shownto be more effective in the management ofactive variceal haemorrhage.

T4Non-steroidal anti-inflammatory drugs andbleeding peptic ulcer

K W SOMERVILLE, G FAULKNER, AND M J S

LANGMAN (Department of Therapeutics,University Hospital, Nottingham) Non-aspirin non-steroidal anti-inflammatorydrugs (NSAID) are widely believed tocause peptic ulceration or its complicationsbut good supportive evidence is lackingbecause no planned study using appro-priate controls has ever been done. Since

1983 we have carried out such a study inwhich patients over 60 years of age admit-ted to the Nottingham City and UniversityHospitals with a bleeding peptic ulcer havebeen questioned about immediate antece-dent drug intake. For each patient acommunity control matched for age andsex was selected from the same generalpractice list. The same questions were alsoposed to an age and sex matched hospitalcontrol by a trained interviewer (GF) usingthe same structured questionnaire. Ninetyper cent of the 193 community controls sofar approached agreed to participate as didall the hospital controls.Two hundred and thirty of the 289

eligible cases admitted during the two yearstudy period, were directly questioned.Thirty five per cent were users of NSAIDscompared with 14% of 230 hospital con-trols and 15% of 173 community controls.Patients with bleeding peptic ulcers werenearly four times as likely to be NSAIDtakers as hospital or community controls(relative risk 3-8, X2=26.1 and 3 7, X2=21 7respectively, p<0-001 in both cases byMcNemar's test). By extrapolation frromthese results, of about 1000 deaths annuallyof patients over the age of 60 admitted withbleeding peptic ulcers approximately 200would be associated with NSAID use.

T5Gluten sensitive oral ulceration in theabsence of coeliac disease

C O'MAHONY, C O'FARRELLY, D G WEIR, T

FINCH, AND C F FEIGHERY (Departments ofImmunology and Clinical Medicine, StJames's Hospital, Dublin, Eire) Theclinical management of severe recurrentoral ulceration is often unsatisfactory andsuch ulceration may continue for life.Symptomatic treatment alone is used un-less, as is the case in a small proportion ofpatients, an underlying systemic illness,such as coeliac disease, is found.The purpose of this study was to deter-

mine if gluten sensitive oral ulcerationoccurred in the absence of coeliac disease.Nine patients, presenting primarily withsevere recurrent oral ulceration, were in-vestigated. All had normal small intestinalbiopsies. Three patients had raised alpha-gliadin antibodies, however (raised levelsare found in up to 85% of patients withcoeliac disease). All three were HLA DR3- the antigen found in up to 90% of ourcoeliac patients. A trial of gluten free dietwas instituted in these three patients. Twohad an excellent response with complete

remission of the ulceration and alpha-gliadin antibody levels returned to normalA subsequent gluten challenge brought anacute relapse in oral ulceration althoughthe small intestinal biopsy remainednormal. Alpha-gliadin antibody remainedelevated in the non-responding patientand dietary assessment showed noncompliance.Two of the remaining six patients with

oral ulceration but negative alpha-gliadinantibodies also went on a gluten free dietbut no clinical improvement was noted.

This study shows that in a subpopulationof patients with severe recurrent oral ulcer-ation the lesion responds to gluten exclu-sion. Thus the clinical spectrum of glutensensitive disease is wider than is oftensuspected.

T6Influence of neuropeptide Y on rabbit ilealmucosa

K J MORIARTY, N B HIGGS, M WOODFORD, J MALLEN, S R BLOOM, AND L A TURNBERG(Department of Medicine, Hope Hospital,Salford, and Department of Medicine,Hammersmith Hospital, London) Neuro-peptide Y (NPY) has been shown in highconcentration in the enteric nervous systemand in particular in the submucosalplexuses of the small intestine of a varietyof mammalian species including man. Thisdistribution suggests that NPY may play arole in the control of mucosal fluid trans-port in the intestine and we thereforeexamined its effect on segments of rabbitileal mucosa, stripped of muscle coats, andmounted in flux chambers. NPY, added tothe serosal aspect of the mucosa in concen-trations of 1-2x10-9 M to 1 2x10-7 M,caused a rapid dose-dependent fall inshort-circuit current and electrical poten-tial difference across the mucosa, a signifi-cant response being observed at 1-2x10-9M. The maximal decrease in short-circuitcurrent was 1-12±0-19 ,umol/cm2/h(p<O00l) and in potential difference was1-62±0i29 mV (p<0-001) at 1*2x10-7 M.A fall in short-circuit current and potentialdifference is usually associated with en-hanced absorption and this response issimilar to that induced by opiates, whichinfluence the mucosa indirectly via a pre-sumed neural intermediary, such as NPY.Application of morphine (10' M and 10'M) to the serosal aspect of the mucosacaused a decrease in short-circuit currentand potential difference. The response ofthe ileum to morphine was not however

A1137


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


influenced by pretreatment with NPY(1-2x10-7 M).We conclude that these findings support

a role for NPY in the control of ileal ionabsorption and suggest that NPY is not theneural mediator of opiate responses in themucosa.

BSG/BASL LIVERT7-17

T7Liver copper content correlates poorly withseverity of histological abnormality inIndian childhood cirrhosis (ICC)

I C TALBOT, M S TANNER, AND A M PRADHAN

(Departments of Pathology and ChildHealth, University of Leicester and KingEdward Memorial Hospital, Pune, India)Among 327 0-14-year-old children withliver disease in Pune, 139 (aged 7-66months) had liver biopsies diagnostic ofICC. Characteristic features were: peri-cellular fibrosis (96%); hepatocyte bal-looning (92%); Mallory's hyaline (91%);foci of polymorphonuclear leucocytes(96%); spotty necrosis (64%); orcein andrhodanine staining (all). There was disrup-tion of the limiting plate in 99%. Mild bileduct proliferation was seen in only 33%,cholestasis in only 18% (25 cases, includingsix necropsies), and fat was absent.

Histological features were graded 0(absent) to 3 (severe). Liver copper, mea-sured by atomic absorption spectro-photometry, was grossly raised (1610±997gg/g dry weight, normal <100 ,ug/g), andcorrelated with rhodanine (p=0-009) andweakly with orcein (p=0-16), but did notcorrelate with any other histologicalfeature. Raised hepatic copper in otherdisorders also did not correlate with histo-logical severity.Although the histological changes in ICC

may result from copper overload, theseverity of damage bears no relationship tocopper content at the time of examination.

T8Effects of a long acting somatostatinanalogue (SMS 201-995) on the activity ofthe reticuloendothelial system (RES) in rats

S A JENKINS, J N BAXTER, AND R SHIELDS(Department of Surgery, University ofLiverpool, Liverpool) Because the activ-ity of the RES is markedly reduced in

cirrhosis, the inability of the liver to de-toxify enterically derived endotoxin mayaccentuate existing hepatic damage.Somatostatin (SRIF) has been shown to beeffective in controlling acute varicealhaemorrhage and preliminary reports sug-gest that it is cytoprotective with respect tothe liver. The purpose of this study was toinvestigate the effects of a new long actinganalogue of somatostatin (SMS 201-995)on RES function in rats. Male Wistar ratsreceived 2 ,ug SMS 201-995 bd sub-cutaneously for seven days. A controlgroup of rats received similar injections ofthe same volume of isotonic saline.Reticuloendothelial system function wasassessed by the uptake of technetium sul-phur colloid, Cr51 red cells damaged byacetylphenylhydrazine and colloidalcarbon by the liver and spleen. A furtherstudy was undertaken to establish whetherSMS 201-995 would protect against E coliendotoxaemia. Administration of SMS201-995 significantly increased the liver toblood ratio (6-5±0.9 to 27-1±5-1) (p<0-01Student's t test) and spleen to blood ratio(4-66±1-1 to 29-3±4.4) of sulphur colloid.Similarly, SMS 201-995 stimulated theuptake of colloidal carbon and Cr51 redcells by the liver and spleen. Furthermore,SMS 201-995 significantly improved sur-vival after intraperitoneal administration ofE coli endotoxin (p=0-004, Log RankTest). These results suggest that SMS201-995 may stimulate the RES system andbe of value in protecting against endo-toxaemia in patients with cirrhosis.

T9Hepatic infarction after orthotopic livertransplantation

P POWELL-JACKSON, R J POLSON, R Y CALNE,AND R WILLIAMS (Liver Unit, King'sCollege Hospital and School of Medicine &Dentistry, Denmark Hill, London)Although the clinical features of hepaticinfarction due to hepatic artery thrombosis(HAT) after orthotopic liver transplanta-tion are well described, the picture that canoccur with thrombosis of the portal vein orhepatic veins is less well documented. In45% (5/11) of cases of hepatic infarctionproven by necropsy or reoperation thatoccurred in 199 patients in our series,thrombosis of vessels other than the hepa-tic arteries was the cause - portal veinthrombosis in three and hepatic veinthrombosis in two cases. In all cases theclinical picture was characterised by a rapiddeterioration of consciousness progressing

to coma within 24 hours, hypotension andhypoxia with levels of aspartate amino-transferase >600 IU/I and prothrombintime over 25'seconds prolonged. Deathwithin 12 days of onset of symptomsoccurred in 91% (10/11) of these cases andwas responsible for 13% (10/75) of alldeaths occurring within six weeks of livergrafting. In only four patients (two withHAT) could technical difficulties be identi-fied while possible aetiological factors inthe remainder included prolongedhypotension in three and a presumedcoagulopathy in one patient with BuddChiari. In the one survivor of this series, asuitable donor became available within 12hours of hepatic infarction and retrans-plantation was successfully carried out.

T10Increased iron absorption after chronicethanol feeding in rats

R MAZZANTI, S K SRAI, M A BOSS, E S DEBNAM,AND P GENTILINI (INTRODUCED BY 0 EPSTEIN)(Clinica Medica IV, Universita di Firenze,Italy, and Royal Free Hospital School ofMedicine, London) Alcoholics often havean increased amount of iron in the liver,which may contribute to the developmentof alcoholic liver disease, however themechanism is unknown. It has been shownthat chronic alcoholism reduces the entero-cyte turnover and increases galactoseabsorption. Whether chronic alcoholintake affects iron absorption is still con-troversial. The aim of this study was toinvestigate the effect of chronic alcoholismon whole body iron absorption in rats.Twenty eight adult male Sprague-Dawleyrats were pair fed a liquid diet containingeither ethanol as 36% of total calories or anisocaloric diet where fat was substituted forethanol. On the 28th day, four hour fastedrats were given an oral dose of 59Fe (0-5,uCi) and then immediately counted by awhole body counting technique. 59Fe levelswere then monitored in the following ninedays, using the same technique. Althoughethanol fed and control rats had a similarhepatic iron levels (59.5±5.8 vs 60 2±7 41,ug/100 mg dry liver weight (mean±SEM),the 59Fe total body iron content (% of theadministered dose) was significantlygreater in the ethanol group (75±3) com-pared with control group (45+4; p 0-001).These results show that chronic ethanolintake increases iron absorption in rats.This phenomenon may be one explanationof the abnormally high liver iron levels inalcoholics.

A1138


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


T1lThe Birmingham liver transplantationprogramme: our early experience

D CLEMENTS, R M KIRBY, S HUBSCHER, W AJUREWICZ, M SEALY, E ELIAS, AND PMCMASTER (Queen Elizabeth Hospital,Edgbaston, Birmingham) Our indicationsfor liver transplantation are primary livertumours without evidence of dissemina-tion, chronic liver disease at end stage(projected life expectancy less than sixmonths), and acute liver failure. Of thepatients referred for assessment approxi-mately one quarter undergo transplanta-tion; the remainder being too early in thecourse of their disease, receive othermedical or surgical therapy, or are unsuit-able for other reasons.Twenty five grafts have been carried out

in 24 patients. The diagnosis was primarybiliary cirrhosis in eight, and primary livertumour in eight. The other eight patientswere grafted for ca1-antitrypsin deficiency(two), cryptogenic cirrhosis (two), Budd-Chiari syndrome (two), chronic activehepatitis (one), and fulminant hepaticfailure (one).

Fourteen of the 24 patients are alive(median seven months, longest survivorover three years). Seven patients have diedas inpatients after the operation (mean 15days postop) and the mean hospital stay forprimary discharge was 32 days followingtransplantation. There have been three latedeaths, two from recurrent hepatomas.

Five patients alive beyond 12 months arein normal health with minimal immuno-suppression.

T12Does male sex predispose to the HBsAgcarrier state or to chronic liver disease?

M CHIARAMONTE, A FLOREANI, M ZAGOLIN, E

PORNARO, AND R NACCARATO (Departmentof Gastroenterology, Policlinico Universi-tario, Padova, Italy) The male/femaledifferences in response to HBV is a stilldiscussed problem. In the attempt to clarifythis point we studied 319 adult HBsAgcarriers (aged 15 to 70 years) presentedduring epidemiological screening or bloodtesting for suspected liver disease. Twohundred and twenty two were men (M/F=2.3). Out of the men 55% had CALD,23% had CPH and 22% were healthycarriers, while among women 32% hadCALD, 16% had CPH and 52% werehealthy carriers. M/F ratio was 0-98 inhealthy carriers, 3-4 in CPH and 3-9 in

CALD. Age at presentation, prevalence ofHBsAg and seroconversion rate to anti-HBe were similar in men and women ineach group, while cofactors such as drugaddiction, alcohol abuse, Delta infection,infection in adulthood were significantlyassociated with liver disease and were morefrequent in males. Evidence of intrafami-lial infection was present in 16% of menand in 29% of women and was morefrequent in the healthy carrier group(34%), suggesting that, when the HBVinfection occurs in the household and/or atearly ages the risk of chronicity is the samefor males and females. We tested thishypothesis in 86 silblings from high riskfamilies: the HBV infection prevalence was46-6% in men and 43-5% in women andthe HBsAg carrier rate was 34% in menand 38% in women.

In conclusion, the difference in theHBsAg carrier rate between sexes is in-creasing with the severity of the disease.Other harmful factors (more frequent inmen?) and/or the age at infection canpartially explain this phenomenon.

T13The enigma of asymptomatic primarybiliary cirrhosis

M R LUCEY, J M NEUBERGER, AND RWILLIAMS (Liver Unit, King's CollegeHospital and School of Medicine & Den-tistry, London) Primary biliary cirrhosisin patients without symptoms of hepaticdisease ('asymptomatic PBC') is well de-scribed. Although progression occurs insome cases, these patients are said to havenormal life expectancy. In order to con-sider whether asymptomatic PBC is abenign condition we reviewed the clinicalcourse of 231 patients with PBC followedfrom 1970 to 1984 of whom 31 (22 women)had no hepatic symptoms, as previouslydefined, up to the time of diagnosis. Atdiagnosis the mean age of symptomatic andasymptomatic patients was similar (55 vs 56years) and disease severity was less inasymptomatic patients; serum bilirubin(median, range) 13,umol/l, 3-65 (asympto-matic), 35 ,umol/l, 4-390 (symptomatic).Life table analysis showed longer survivalin asymptomatic patients; actuarial fiveyear survival: asymptomatic 70%, sympto-matic 50% (p<0.05). In a median followup period of 45 months, however, 18 (12women) patients developed hepatic symp-toms and their survival from the onset ofsymptoms was not different from that inpatients with hepatic symptoms prior to

diagnosis. These data suggest that manypatients with asymptomatic PCB do be-come symptomatic and they should beconsidered a presymptomatic group inwhom the apparent benign course is aconsequence of lead time bias.

T14Autonomic neuropathy and alcoholic liverdisease

F BARTER AND A R TANNER (Department ofMedicine II, Southampton General Hospi-tal, North Tees General Hospital, Cleve-land) Autonomic neuropathy (AN) hasbeen assessed in 16 patients with provenalcoholic liver disease (ALD), 14 alcoho-lics from the community without evidenceof liver disease and 30 sex and age matchedcontrols attending for endoscopy. Tests ofparasympathetic function were heart rateresponses to Vasalva, to deep breathingand to standing. Sympathetic function wasassessed by BP response to standing andsustained handgrip. The presence of two ormore abnormal tests was taken to indicateAN. Fifty six per cent of subjects withALD, 14% of community alcoholics andnone of the controls had AN. Femalealcoholics were more likely to develop thiscomplication (80% vs 28%; p=0.004).There was no significant correlationbetween symptoms and objective signs ofAN. Peripheral neuropathy was docu-mented in 64% of both groups of alco-holics, but not in the controls. Alcoholicswith AN were older, had been drinkinglonger and 91% had an associatedperipheral neuropathy. Prospective studiesin alcoholics with AN would be of interestsince studies in diabetics have shown a highmortality and morbidity associated withobjective signs of AN.

T15Screening for haemochromatosis in the UK:preliminary results

A R TANNER, S DESAI, W LU, AND R WRIGHT(Department of Medicine II, SouthamptonGeneral Hospital, North Tees General Hos-pital, Cleveland) Recent studies inEurope and North America have indicatedthat the gene frequency for haemochroma-tosis in caucasian populations is muchhigher than previously suspected. In thesepopulations the prevalence of heterozy-gotes (HO) has varied between 9% and14% with a homozygote (HH) frequency of0*3-0-5%. In the present study, serum

A1139


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


ferritin, serum iron and iron binding satur-ation have been measured in 1800 consecu-tive blood donors in order to assesspossible iron overload. The frequency ofblood donation was such that little differ-ence would be made to iron stores in HHsubjects. Those subjects with raised serum

ferritin (>200 ,ug/l), raised serum iron(>35 ,umol/l) and increased iron bindingsaturation (>48%) were classified as prob-able HH and those with elevated saturationand/or serum iron classified as possible HOsubjects. Approximately 30% of knownHO subjects have minor biochemicalabnormalities. Five subjects fulfilled thecriteria for HH giving a homozygote preva-lence of 0-3% (95% confidence limits01-0.6%) and the heterozygote frequencywas approximately 12%. Further studiesare continuing in order to independentlyassess iron stores in these individuals andtheir families. This study indicates thatbiochemical screening of an asymptomaticpopulation may be of value.

T16Increased incidence of menstrual abnormal-ities and hysterectomy preceding primarybiliary cirrhosis

A J STELLON AND R WILLIAMS (Liver Unit,King's College Hospital and School ofMedicine & Dentistry, London) Men-strual disturbances have been reported inpatients with chronic active hepatitis(CAH) and alcoholic liver disease (ALD)but no studies have been described inwomen with primary biliary cirrhosis(PBC). In this prospective study men-

strual, gynaecological and obstetric his-tories were obtained from PBC patientsand compared both with age-matched con-trol subjects and patients with other typesof chronic liver disorders. Histories were

obtained from 87 patients, aged 35-70years, with PBC and compared with acontrol group of 100 age matched (35 71years) hospital personnel and 80 agematched (34-73 years) female patients witheither CAH (45) or ALD (35). A signifi-cantly higher rate of hysterectomy and D &C was found in PBC when compared withcontrol subjects (p<0-025 and p<005respectively) and other patients withchronic liver disease (p<0005 and p<005respectively). Both hysterectomy and D &C were carried out a mean of 10-7 yearsand 13-2 years respectively, in the majorityof patients, before the diagnosis of PBC.The main clinical indications for hysterec-tomy in PBC was menorrhagia while

fibroids (51 %), endometrosis (21%) and the higher incendometrial polyps (7%) were the main tation. Developerative findings. Endometrial hyperpla- depend moresia was detected macroscopically in 24% rather than thPBC patients at hysterectomy and 33%patients in which histological material was

obtained at D & C. The findings would beconsistent with an underlying hormonaldisturbance favouring oestrogen excess SMALL BOWEL

during the presymptomatic stage of PBC. T18-28

T17Prognostic features in chronic activehepatitis

J J KEATING, A J STELLON, P J JOHNSON, C JO'BRIEN, R D JOHNSON, B PORTMANN, J EHEGARTY, AND R WILLIAMS (Liver Unit,King's College Hospital and School ofMedicine & Dentistry, London) Resultsrelating to survival in chronic active hepati-tis (CAH) vary between centres and are

more likely to depend upon the nature andstage of the disease than the treatmentused. We have, therefore, carried out ananalysis of presenting features, histology,frequency of relapse, survival and relationto sex in 106 patients (21 men) with'autoimmune' CAH, 69 patients (28 men)with 'cryptogenic' CAH and 29 patients (26men) with HBsAg seropositive (HBsAg+)CAH. Variceal bleeding and encephalo-pathy were more common presenting fea-tures in the cryptogenic group (p<0-01)while the presence of oedema, ascites andjaundice did not differ between the groups(NS). Cirrhosis was more frequent at pre-sentation in the cryptogenic group (53%)than in the autoimmune (30%) or theHBsAg+ group (35%) (p<0002). Deve-lopment of hepatocellular carcinoma(HCC) occurred in five men with CAH andin five men with HBsAg+ CAH. Therelapse rate per year in autoimmune CAHwas 15% for those treated with predniso-lone, 4-4% for those treated with predniso-lone and azathioprine and 77% for those inwhom treatment was withdrawn. This was

similar to the relapse rate in the crypto-genic CAH of 15%, 6-2%, and 72%respectively. The overall five year survivalwas 87% in the autoimmune CAH groupcompared with 65% in the cryptogenicgroup and 80% in the HBsAg+ group(p<0-001) although no significant differ-ence in survival was observed in thosepresenting without cirrhosis. The responseto immunosuppressive therapy is com-

parable in patients with and without auto-immune markers and the worse prognosisin the cryptogenic group appears related to

cidence of cirrhosis at presen-lopment of HCC appears toe on the sex of the patienthe HBV status.

T18Bacterial contamination in chronic renalfailure

M HEALY, CATRIONA LiTYLE, P O'CONNOR,MONA O'MOORE, A SPEEKENBRINK, B KEOGH, CT KEANE, D G WEIR, AND R R O'MOORE(Departments of Medicine, Clinical Bioche-mistry and Mircobiology, Trinity College,Dublin, Eire) Bacterial contamination ofthe upper small intestine (USI) in chronicrenal failure (CRF) has been implicated inthe causation of uraemic diarrhoea anddecreased mental alertness possibly due tobacterial production of dimethylamine andtrimethylamine (DMA, TMA).The incidence and extent of contamina-

tion in CRF had been estimated as approxi-mately 40% by the '4C glycocholate breathtest (14CBT). Intubation and microbio-logical culture was carried out in 28patients with positive 14CBT. The controlswere 18 patients with contaminated smallbowel syndrome and 20 normal subjects.Production of DMA, TMA was monitoredserially during digestion of a meal (threehours). After receiving antimicrobialtherapy, reassessment of mental alertness(Wechsler), was made in 11, and microbialculture and plasma DMA, TMA levels insix CRF patients.Twenty four of all 28 CRF patients were

contaminated. The 14CBT and anaerobicbacterial counts became normal in all CRFpatients after antimicrobial treatment.There was also a significant increase inboth weight and mental alertness(p<001). Serial plasma and intestinalDMA levels increased significantly duringthe meal (p<0.001). Plasma DMA concen-trations also decreased significantly follow-ing treatment.Our findings suggest USI bacterial con-

tamination may be clinically important andrelatively common in CRF. Antimicrobialtherapy lead to improvement of symptoms.

T19Macrophages in the microenvironment ofcoeliac disease

A1140


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


L K TREJDOSIEWICZ, G MALIZIA, J OAKES, L W

POULTER, G JANOSSY, AND M S LOSOWSKY

(Department of Medicine, St James'sUniversity Hospital, Leeds, and AcademicDepartment of Immunology, Royal FreeHospital, London) Involvement of theimmune system in coeliac disease is welldocumented; attention has hitherto beenfocused on antibody responses and cellmediated immune responses. Althoughthere are many macrophages in the laminapropria of the small intestine in coeliacdisease, these have not hitherto beenstudied. We have used simultaneousdouble labelling of jejunal biopsy cryostatsections with monoclonal antibodies todemonstrate macrophage subsets and theirrelationship to HLA-DR expression inindividual cells in 19 patients with coeliacdisease (11 treated, Grades 1 to 3; eightuntreated, Grades 3 and 4) and 11 controls.Antigen presenting macrophages, as

identified by RFD1 antibody were a minorsubpopulation. In the controls, all suchcells expressed HLA-DR whereas up to70% of antigen presenting cells in coeliacpatients had lost their expression of HLA-DR. HLA-DR expression is believedessential for functioning of these cells.Scavenger macrophages, as indicated byRFD2 antibody, were increased aboutthree-fold in coeliac disease as comparedwith controls, although theproportion ofscavenger cells which appeared to be in theactivated state, as judged by histochemicalmarkers and HLA-DR expression, wasunchanged. Regulatory macrophages havepreviously been shown to exist in func-tional assays and are now believed to beidentified by the RFD7 marker. SomeRFD7 cells in coeliac -disease are of a

distinctive type, not hitherto described.They are crescentic in shape and surroundaggregates of lymphocytes. All thesechanges were proportional to the degree ofpathological damage.

It is clear that macrophages are involvedin the immunological disturbances in theintestinal mucosa in coeliac disease. Thepresence of regulatory macrophages sur-

rounding lymphoid aggregates might sug-gest that they were in some way controllingthe cells within, perhaps being responsiblefor reducing the expression of HLA-DRand thus reducing antigen presentation andminimising T cell stimulation. The increasein scavenger cells may represent a responseto the tissue damage.

T20Inflammatory cell subpopulations in nor-

mal and coeliac small intestinal mucosa

JACINTA KELLY, C O'MAHONEY, C O'FARRELLY,C FEIGHERY, AND D G WEIR (Departments ofImmunology and Clinical Medicine, TrinityCollege, and St James's Hospital, Dublin,Eire) Coeliac disease is characterised byvillous atrophy and chronic inflammatorycell infiltrate. Using an immunoperoxidasestaining technique on frozen biopsies, thecell subpopulations defined by monoclonalantibodies were studied in untreated (10)and treated coeliac patients (four) andcompared with normal controls (eight). Apanel of monoclonal antibodies was used toidentify Pan T cells, T helper cells, Tsuppressor/cytotoxic (S/C) cells, (using Leu2a and Leu 8) macrophages, and HLA-DRantigens.

Intraepithelial lymphocytes (IELs) were

almost exclusively of the S/C phenotype inboth coeliac and normal small intestine. Inthe coeliac mucosa, however, the numberof these IELs was greatly increased, oftenoutnumbering the number of epithelialcells present. In some cases these IELsexpressed HLA-DR antigens, suggestingthat they are in an activated state. S/C IELswere also found in the crypt epithelium ofthe coeliac mucosa, but not in the normalcrypts. The lamina propria contained Thelper cells and macrophages in closeassociation and both cell tvpes were foundin increased numbers in the coeliac mucosa.HLA-DR expression was found along

the surface epithelial cells of-the normalsmall bowel, but was absent from theepithelial cells of the crypts. In coeliacmucosa however, expression of this antigenwas not confined to the surface epithelium,but was found extensively on the cryptepithelium. HLA-DR antigens are in-volved in antigen presentation and im-munological reactions; their increased ex-

pression in the coeliac mucosa demons-trates their involvement in maintaining an

activated immunological state which mayculminate in tissue destruction.

T21Small round viruses in acute diarrhoealdisease in children

V LARCHER, A OLIVER, D LEWIS, AND A

PHILLIPS (Queen Elizabeth Hospital forChildren, London) Electron microscopyof stools from children with acute di-arrhoeal disease (ADD) is of major im-portance in the specific identification ofviral pathogens, for example, rotavirus.Small round viruses (SRV) which lackcharacteristic morphology are found in

1-2% of stool samples from children with-ADD but their specific clinical associationsare ill defined. We therefore used Freonenhancement and immuno-electron mic-roscopy (IEM) to further categorise stoolSRVS in ADD and retrospectively ana-lysed their clinical sssociation.

Sixty one stool samples (1-6% of total)contained SRV. Stools were available foranalysis in 50, and a specific diagnosispossible in 45. These comprised previouslyunrecognised structured viruses (astrovirus13, calici two, morwalk 1), adeno associ-ated virus (AVV) 15, parvovirus six, en-terovirus six, parvo and entero one, andhepatitis A one.

Astrovirus was associated with mildADD, had an equal sex distribution, me-dian age six months, a winter peak andhigh rate of hospital acquisition (six). AVVwas associated with a history of failure tothrive (eight), previous gastroenteritis(eight), and atopy (six). There was a malepreponderance, median age 12-5 months,and no clear winter peak. Parvo and enteroviruses each produced mild ADD, butwithout specific features.

Morphological enhancement with Freonand IEM permits a specific viral diagnosisin 90% cases of ADD associated withSRVs. Prospective studies may definemore specific clinical associations thanthose detailed above.

T22Toddler's diarrhoea - intestinal hurry?

R GUERRERO, G A BROWN, AND A S MCNEISH(Institute of Child Health, Birmingham)Toddler's diarrhoea, one of the forms ofirritable bowel syndrome (IBS), is a com-monly diagnosed condition with an un-defined aetiology. It has previously beensuggested that children with this syndromehave decreased mouth to caecum transittime, but data are few. Using the breathhydrogen test (BHT) we have measuredthe mouth to caecum transit time (MCTT)of 10 g of lactulose in a group of 22 childrenaged 1-4 years (mean 2-5 years) presentingwith IBS-toddler's diarrhoea. The resultswere compared with a group of 13 healthychildren age matched (mean 3-5 years).The mean MCTT in the IBS group was 42minutes (SD 8-81), significantly longerthan that observed in the control group(32.7, SD 8.80), (p<0.01). Eight of thechildren in the IBS group had a secondBHT after 3-13 months (mean 8-6months). Seven of them were said by theirparents to have improved, and their MCTT

A1141


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


decreased towards control values. Nodrugs or basic dietary modification hadbeen used. One child had relapsed onemonth before the test, after being well overthe previous year, and her MCTT in-creased.The results of this study show, contrary

to what has been accepted, that childrenwith IBS-toddler's diarrhoea have pro-longed small bowel transit which decreasesas the condition resolves.

T23Study of postoperative gastrointestinalmotility in man using radiotelemetry

D L MORRIS, A G CLARK, D F EVANS, AND J D

HARDCASTLE (Department of Surgery, Uni-versity Hospital, Nottingham) Abnormalmotility has been shown in the stomach andduodenum after operation but has not beenmeasured simultaneously in the stomach,small bowel and colon in man.

In 10 patients undergoing cholecystec-tomy intraluminal pressure was measuredby three radiotelemetry capsules. One un-tethered capsule was swallowed 12 hoursbefore operation to allow it to reach thecaecum followed by two capsules tetheredin the gastric antrum and proximal je-junum. After a period of recording whenfasting migrating motor complexes (MCCs)and colonic motility were confirmed,patients underwent surgery. Motility wascontinuously monitored for up to 72 hoursafter the operation. In two patients record-ings were of insufficient quality foranalysis. Fasting jejunal activity returnedquickly after operation (mean 100-4min±47SD) but with shorter interval(mean 36-6 min±32SD) compared withnormal (approximately 120 min). Normalfasting patterns denoted by a gastro-duodenal MMC were not seen in anypatient during the recording period, but atypical gastric activity was seen in threepatients between 12-24 hours. Colonicmotility returned variably from 1-50 hoursin six patients, but was absent in two. Wehave confirmed that gastrointestinal motil-ity is disrupted by anaesthesia and surgeryand returned at a variable rate in thestomach and colon and although recoverywas more rapid in the jejunum, it wasabnormal in type.

T24Clinical efficacy of peripheral intravenousnutritional support

C M ROYCE, M MULLEE, AND S I KARRAN

(University Surgical Unit, SouthamptonGeneral Hospital, Southampton) Totalparenteral nutrition is of proven efficacy inthe perioperative period but remains un-

popular. Blackburn pioneered the use ofistotonic amino acids for nutritional sup-port but their value remains controversial.

Forty six consecutive patients whounderwent oesophagectomy for carcinomawere randomly allocated on the day aftersurgery to receive one of three peripheralintravenous support regimens: (a) an isoto-nic amino acid solution (Perifusin) at a rateof 1 gaox/kg/day. (b) an isotonic amino acidsolution as above plus 500 ml 20% intra-lipid/day. (c) 4th dextrose and 0 18%n/saline in equivalent volumes.Groups (a) and (b) were designated as

being 'fed' and group (c) formed the'unfed' clinical control group. The regimenwas continued until adequate oral fluidintake was reestablished. Patients in thethree groups were well matched for age,nutritional status, stage and grade oftumour, operative procedure, surgeon,duration and blood loss.

Postoperative complications were re-

duced in 'fed' patients (a) and (b), com-

pared with (unfed) group (c) (p=0O004).The reduction was due to a much lowerincidence of nutritionally associated com-

plications (NAC), mainly sepsis. Technical(T) and non-nutritionally associated com-

plications (NNAC), - for example,myocardial infarction, were similar in allgroups.

T25Influence of human and murine giardiasison intestinal permeability

F ANDRE, C ANDRE, J GUZMAN, AND S

CAVAGNA (INTRODUCED BY R N ALLAN)(Groupe d'Immunopathologie DigestiveINSERM, Centre Hospitalier LYON SUD,Pierre Benite, France) Reports of an

association of giardiasis with urticaria andbronchial asthma suggest that troubles ofintestinal permeability may be provoked bythis parasite. Using gas chromatography,we have measured five hour urine clear-ance of 5 g mannitol and of 5 g lactuloseingested as markers of intestinal permea-bility, respectively to small and to largemolecules. This study was undertaken inseven patients with untreated giardiasis,four of these patients after effective treat-ment with metronidazole, and in 90 con-trols.Mean mannitol excretion was 14 1l%

and mean lactulose excretion was llifeltz inIcontrols. In patients w ith geilirdlasis theresults were 12 91', aindc '>%' beftoictreatment. The results wcrc 14 ii iiict0 63% after treatment.

During the spontaneous hlimination olgiardiasis by BALB/c mice the numnber ofintestinal mast cells, the histarmine niuCosallcontent and the intestinal perrneabilitx tomannitol and lactulose wsere nicasured tItweekly intervals during eight sweks.The experimental disease was associated

with an increase of the nuimhcr of mistcells and of mucosal histamilne content andwith the same abnormalitics of eut piermelaibility as detected in humrn.ns Ahsorption oflactulose was increased thrce-told. MNiccreceiving 6 mg/kg histamine orallv alsodeveloped gut hyperpermeahilitM to lkit L-

lose: absorption of lactulose \ as also increased threefold.These results suggest thit the tspe oneC

hypersensitivity reaction to giardia intc-tion induces an increase of gult pernieahilitMto macromolecules. Therefore this inlec-tive episode may be the occatsion ot foomdsensitisation.

T26Effect of bicarbonate on efficacN(of oralrehydration therapy in a rat model ofsecretory diarrhoea

E J FLLIOTT, M J KELIIY, sx1i \ISi".WALKER-SMITH, ANt) N1 YARiHiIN\(Department oJ Gastrooeutc o/ogyx nAcademic Departmemt of C hil/d Ilaltli, SiBartholomew's Hospital, ILondoun) C on-troversy exists regarding the nccessits! foiinclusion of bicarbonate (Bic) and otherbase precursors in oral rehydration solu-tions (ORS). Our prcxvions sxork in t-altproximal small intestine indicates that BiCenhances cholera toxin (C 1) -induLCed W,itCrIand sodium secretion. Recent clinicil stnL-dies suggest that the abhsencn 1ol3i fioenot alter clinical efficacy of ()RSR -

investigate this furthcr ss. hixe fioss lierfused the entire rat small intestire (Si )1 x*iththe widely used UK orall rchydraltion for-mulation 'NaCl and glucose oral J)oW dcicompound' (Na 35. K 2ff. C 37, isib I8.Glucose 200 mmol/l; British National For-mulary) and an equivalent solution inIwhich Bic was replaced hb (l PertLusionSwere done before and after induCtioll Otf ,secretory state by 2h pre-exposuirc to 75 ,ueCT. In the normal SI water absorlption 'Nssignificantly greater t roia the+ 13i-containing ORS (+l18--X til min g dirweight) than the Bic-free R(Ri 72 4-

A1142


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


p<0.01). In the secreting SI, however, netwater absorption was two-fold greater withBic-free ORS (+25.1±14) than Bic-containing ORS (+11 1±16; p<005).Both solutions failed to produce net Naabsorption in normal and secreting SI, Nasecretion being significantly greater withBic-free ORS (-16-8±4 vs -9-6±5;p<002) in the secreting SI.These findings indicate that Bic does not

contribute to the promotion of waterabsorption during secretory diarrhoea. Failure to achieve net Na absorption fromORS in the SI suggests an important rolefor the colon during oral rehydrationtherapy.

T27Effect of ileal and intravenous lipid infu-sions on feeding and satiety in humans

I MCL WELCH AND N W READ (ClinicalResearch Unit, Royal Hallamshire Hospital,Glossop Road, Sheffield) The effect ofileal infusion of a lipid emulsion, contain-ing 50% corn oil and 3% albumen, on foodintake and satiety was measured in pairedexperiments in six healthy volunteers. Sub-jects ate for shorter periods of time duringileal lipid infusions compared with infu-sions of albumen and saline (25±1 vs 32±3min, p<0025) and consumed less food(670±23 vs 884±89 g or 1016±79 vs1591±228 Kcal, p<0-05). The quantity ofliquid drunk and the rates of eating anddrinking were not affected by the lipidinfusion. In five further experiments, ilealinfusion of lipid emulsion delayed gastricemptying compared with infusion of albu-men and saline (tl/2: 203±48 vs 68±12 min,p<0.02). Food intake was not affected byintravenous infusions of intralipid com-pared with saline in six volunteers suggest-ing that the above observations were notrelated to the effect of absorbed fat.Intestinal lipid may interact with ileal recep-tors to induce early satiety; this result isprobably explained by early gastric disten-sion caused by delayed gastric emptyi'ng,although the release of an ileal mechanismhaving a direct action on the satiety centrecannot be excluded.

T28Abnormal c-myc oncogene product expres-sion in coeliac small intestinal epithelium

J STEWART, G EVAN, K SIKORA, AND P CICLI-TIRA (Ludwig Institute for Cancer Re-search, MRC Centre, Cambridge, and Gas-

troenterology Research Unit, Rayne Insti-tute, St Thomas's Hospital, London)Oncogenes are sections of DNA that areinvolved in normal cellular growth control.There is evidence that they may be associ-ated with the rapid cellular transformationresulting in neoplastic change. Molecularcloning has allowed characterisation of theamino acid sequences of several oncogenesincluding c-myc. Production of a monoc-lonal antibody to a solid-phase synthesisedpeptide, permitted investigation of thedistribution of the oncogene product inhuman tissue.The pathogenesis of coeliac disease is

not understood. We have investigated thedistribution of the c-myc oncogene productin normal (n=5), treated (n=5) and untre-ated (n=5) coeliac small intestinal biop-sies. The method used peroxidase stainingof the c-myc gene product in paraffinembedded sections. Normal subjects andtreated coeliac patients express very lowlevels of the oncogene product while un-treated coeliac patients or those on a glutenfree diet subjected to a gluten challengeexhibited raised levels in the enterocytes ofthe upper third of the villi.

OESOPHAGO-GASTRO-DUODENALT29-39

T29Late failures of the Angelchik prosthesis

R L WOLVERSON AND J G TEMPLE (QueenElizabeth Hospital, Edgbaston, Birming-ham) Since its introduction in 1979 theAngelchik prosthesis has gained wide-spread popularity as a simple and safeprocedure for the treatment of gastro-oesophageal reflux. We reported our earlyfavourable results using this device, butrecently increasing numbers of significantproblems have been encountered. Between1981 and September 1983 we inserted 17prostheses in 25 patients, six of these(24%) have now had to be removed. In twoof the six this was because of disruption ofthe securing tapes on the prosthesis. Oneof the patients had a second prosthesisinserted and this subsequently had to beremoved as it migrated into the posteriormediastinum and caused dysphagia byangulating. Another prosthesis was re-moved again for a similar reason. Twofurther prostheses were removed for dys-phagia thought to be due to fibrous capsule

formation around the Angelchik and there-fore around the gastro-oesophageal junc-tion.The technical problem of the tape dis-

ruption has been solved by the manufactur-ers. A failure rate of 3/25 (16%), however,is still unacceptably high. Detailed followup of our remaining 19 patients indicatesthat 3/19 (15-8%) are developing symp-toms related to swallowing and may wellrequire removal of their prosthesis andsubsequent alternative antireflux surgerycarried out.

T30Patterns of gastroesophageal reflux associ-ated with oesophagitis

S SADEK, W CHEADLE, G VITALE, C CRANFORD,N W CARTER, AND A CUSCHIERI (Departmentof Surgery, Ninewells Hospital and MedicalSchool, Dundee) Prolonged ambulatorypH monitoring in normal subjects andpatients with varying degrees of refluxinjury as assessed by endoscopy can pro-vide information on the evolution ofoesophageal disease. A comparative studywas therefore carried out between thefollowing groups: Group 1: asymptomaticnormal volunteers (n=50); group 2: symto-matic patients with normal endoscopy(n=26); group 3: symptomatic patientswith oesophagitis (n=46); group 4: symp-tomatic patients with ulcerative oesophagi-tis and/or strictures (n=31).Comparison between groups 1 and 2 -

that is, the development of symptoms -

showed an increase in acid exposure exclu-sively in the erect posture due to increasein the number of short events. The transi-tion from group 2 to 3 (development ofoesophagitis) involved changes in thesupine posture and decrease in oeso-phageal clearance. The transition fromgroup 3 to 4 (development of ulcerativeoesophagitis and/or stricture) was associ-ated with increased acid exposure and areduction in the oesophageal clearance ofacid in both the erect and supine positions.

T31Hydrostatic balloon dilatation of oeso-phageal strictures

B J M JONES, G F MASKELL, AND A R W

HATFIELD (Department of Gastroenter-ology, The London Hospital, London)Hydrostatic balloons are now available fordilatation of oesophageal strictures otherthan achalasia. We report here our pre-

A1143


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


liminary experience with polyvinyl ballooncatheters (W Cook) of 10, 15, and 22 mmdiameter on full inflation.

Seventeen consecutive patients with be-nign (13), malignant (two) or postradiationstricture (two) underwent 25 dilatations.The balloons were passed over an endosco-pically placed guide wire in succession andpositioned across the stricture underradiological cbntrol. Each balloon wasinflated with dilute contrast medium forthree minutes and the expansion of the'waist' in the balloon at the site of thestricture observed radiologically.

In nine procedures (36%) full dilatationto 22 mm was achieved and in the remain-ing 16 (64%) some degree of waistingpersisted. To assess the significance of theresidual 'waist' on the 22 mm balloon,Eder-Puestow olives were passed in ninepatients and the effective stricture dia-meter was found to vary from 11 to 15 mm.In these patients a greater degree ofdilatation was then obtained with E-Polives or Celestin bougies.The balloons were very easy to pass over

the oropharynx with minimal discomfortbut some patients experienced markedretrosternal pain on full inflation. In onepatient with a post radiotherapy stricture asmall localised perforation occurred afterfull dilatation which was successfully man-aged conservatively.

In conclusion balloon dilatation was sim-ple and particularly suitable for elderlypatients with cervical spine deformity butvery narrow strictures were more difficultto dilate completely and additional conven-tional bouginage was often necessary.

T32Does Valsalva's manoeuvre cause varicealbleeding?

S W HOSKING AND A G JOHNSON (UniversitySurgical Unit, Royal Hallarshire Hospital,Sheffield) Large increases in varicealpressure recorded during Valsalva's man-oeuvre may cause variceal bleeding but nostudies have measured intraoesophagealpressure simultaneously. This preventsmeasurement of the pressure differenceacross the varix wall which is a more likelypredictor of variceal rupture. We havemeasured intraoesophageal and intravar-iceal pressure simultaneously during Val-salva's manoeuvre to determine this press-ure difference. Thirteen patients withoesophageal varices were studied beforesclerotherapy. A perfused manometer linewas passed into the oesophagus followed

by an Olympus GIF-Q1O endoscope toapproximately 4 cm above the oesophago-gastric junction. A perfused varicealinjector was used to obtain intravaricealpressure readings. At rest, corrected intra-variceal pressure (intravariceal minusintraoesophageal pressure) varied betweenpatients from 6-22 mm Hg. During voluntary Valsalva's manoeuvre, uncorrectedintravariceal pressure rose to between 16and 74 mm Hg (p<0.001 Paired t-test).After correcting for intraoesophagealpressure, the change in pressures werefrom -9 to + 18 mm Hg (not significant),and showed no correlation with readingsobtained at rest. Repeated Valsalva's man-oeuvre within each patient showed a simi-lar change on each occasion. These resultssuggest that Valsalva's manoeuvre causesonly a small rise or even a fall in thepressure across the varix wall, and seemsunlikely to initiate variceal haemorrhage.

T33Follow up of laser palliation for malignantdysphagia

S G BROWN, K MATrHEWSON, C P SWAIN, ANDC G CLARK (Department of GastroenterologyUniversity College Hospital, London) Wetreated 18 patients aged 57 to 88 withendoscopic Nd YAG laser therapy topalliate dysphagia from advanced malig-nancy of the oesophagus and gastric cardiaand were able to restore satisfactory swal-lowing in 14. Failure was caused by massiveextrinsic tumour in three and a laserperforation in one. The other 14 werefollowed up. Thirteen have died. Threedied without further dysphagia after eight,11 and 42 weeks (the first of these was theonly patient also to receive radiotherapy,which he tolerated poorly). Four had poor-ly defined difficulty with swallowing nearto the time of death from disseminateddisease after four, five, 14 and 14 weeks.Three had recurrent dysphagia attributableto exophytic tumour at four, five and 11weeks. Two of these had excellent resultsfrom further laser therapy, one survived afurther 19 weeks without dysphagia and theother is still swallowing well after a further28 weeks. The third had a Celestin tubeinserted but its introduction caused atracheobronchial fistula from which shedied after three weeks. The other four hadrecurrent stenosis due to extrinsic tumourat 10, 12, 14 and 16 weeks. Intubation wasattempted in three with bad results; oneperforated, one aspirated and one died 24hours postoperatively.

Necropsy histology in cases with goodresults showed extensive fibrosis in thelaser treated areas, even in the absence ofprevious radiotherapy. This treatmentappears effective in cases with obstructiondue primarily to exophytic tumour. Recur-rent exophytic tumour responds well tofurther laser therapy, but other treatmentsfor recurrent obstruction have poor results.

T34Ambulatory oesophageal pH monitoring inachalasia

H L SMART, P N FOSTER, D F EVANS, B SLEVIN,AND M ATKINSON (UniWersits Hospital,Queen's Medical Centre, Nottingham)Radiotelemetric ambulatory oesophagealpH monitoring was performed in 12 pa-tients with symptomatic achalasia, seven ofwhom had a dilated oesophagus with re-tained food residue. pH Monitoring re-vealed that classical episodes of gastro-oesophageal reflux occurred in only onepatient. Overall an abnormally high percen-tage acid exposure time (AET), was foundfor pH<5 (44.4%) and pH<4 (20()3c)prior to treatment. In patients with foodresidue these values were significantlyhigher than those found in patients withoutretained food. Repeat studies one weekafter pneumatic dilatation showed a fall inAET in patients with initial food residue(pH<5 from 65.1% to 51-7%) but anincrease was seen in those without initialfood residue (pH<5 from 7.8Clc to 46-2%/C ).Gastro-oesophageal reflux occurs infre-quently in untreated achalasia and theabnormally high AET is presumably be-cause of fermentation of retained foodmaterial. The fall in AET after pneumaticdilatation is explained by improvement inoesophageal emptying in those patientswith initial food residues whereas the risein AET in those without initial residue isprobably attributable to gastro-oeso-phageal reflux after dilatation.

T35Altered drug pharmacokinetics in smokers- an effect of smoking on gastric emptying

D A JOHNSON, E J S BOYD, AND K G WORMSI FY(Ninewells Hospital, Dundee, Scotland)Gastric inhibition by antisecretory drugs isimpaired by cigarette smoking. We under-took studies to determine whether thiseffect is attributable to an alteration ofdrug pharmacokinetics. Eight habitualsmokers underwent two studies on sepa-

A1144


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


rate days. The stomach was intubated,resting contents aspirated, and 300 ml of ameal containing Oxo meat broth and PEG4000 (2-5 g/l) was introduced into thestomach. Gastric contents were sampledafter 10 minutes and aspirated completelyafter 20 minutes. Gastric secretory re-sponses and the volume of gastric contentsdischarged into the duodenum were calcu-lated by the method of Hunt. On one of thestudy days, in random order, subjectssmoked at a rate they found comfortable.On the other day cigarettes were prohi-bited. Ten habitual smokers underwenttwo studies on separate days. On each dayeither. ranitidine 150 mg (five subjects) orcimetidine 200 mg (five subjects) was takenorally with a meal. Blood levels of raniti-dine or cimetidine were measured at 0, 15,30, 45, 60, 90, 120, 150, 180, 240, and 360minutes after dosing. Cigarettes weresmoked on only one study day. Ten furthersmokers underwent similar studies in whichthey received either ranitidine 50 mg (n=5)or cimetidine 200 mg n=5) intravenously.When cigarettes were smoked the

volume of gastric contents discharged intothe duodenum during the meal increasedby 23%. Drug absorption after oral dosingwas more rapid, and peak plasma levelsachieved sooner, when cigarettes weresmoked. For any given time after peakplasma levels had been achieved, however,plasma drug levels were lower whencigarettes were smoked. Smoking did notalter the pharmacokinetics of intravenouslyadministered drugs.We conclude that cigarette smoking al-

ters the pharmacokinetics of antisecretorydrugs in a manner which may contribute tothe impaired therapeutic response. Theeffect is attributable solely to an increase inthe rate of gastric emptying.

T36Natural history of silent duodenal ulcer

G BIANCHI PORRO, M LAZZARONI, M PETRILLO,F PARENTE, AND F PACE (GastrointestinalUnit, L Sacco Hospital, Milano, Italy)Endoscopic follow up of duodenal ulcerpatients during maintenance treatmentwith H2-blockers have raised the problemof asymptomatic recurrence of duodenalulcer. In order to define the natural historyand the clinical significance of this subset ofulcer, we followed up during nine months.62 patients with a silent recurrence ofduodenal ulcer, detected in the course of aroutine endoscopic examination of patientsundergoing long-term ulcer treatment.

Among these patients, 25 were receivingcimetidine (400 mg nocte), 21 placebo (2tablets nocte), and 16 no therapy except forantacids when needed. Each patient con-tinued the treatment unchanged during thenine month follow up period. Endoscopicreexaminations took place after three andnine months, and at symptoms occurrence,whereas clinical examinations were per-formed at bimonthly intervals. The cumu-lative rate of spontaneous ulcer healingafter nine months was 39%: 48% in H2-blockers group, 37% in the group withoutany therapy, and 29% in placebo group,respectively (p>0-05). The cumulative fre-quency of symptoms occurrence was 50%:40% in patients receiving cimetidine, 44%in those receiving no therapy, and 67% inthose receiving placebo (p>0-05). Thepercentages of unhealed asymptomaticulcer after nine months were 12% in theH2-blockers treated group, 19% in theno-treatment group, and 49% in the place-bo treated group (p>0-05). One patientreceiving no treatment bled from his ulcerduring the follow up period.

T37Immediate effects of vagotomy on parietaland oxyntic cell function

P D SCOTT AND R F MCCLOY (UniversityDepartment of Surgery, Manchester RoyalInfirmary, Oxford Road, Manchester)The time course and pathophysiology ofchanges in gastric secretion by vagotomyremain unclear. Grassi suggested that gas-tric acidity falls to >pH 5-5 immediatelybut gastric function tests in the earlypostoperative period reveal the pH is often<pH 2.

Gastric acidity and pepsin concentrationwere measured in samples of gastric juiceaspirated hourly in 11 patients undergoingvagotomy (seven proximal gastric vago-tomy, four truncal vagotomy and pyloro-plasty) for duodenal ulcer disease. Stan-dard pre- and postoperative basal/sham/pentagastrin tests were performed to assessthe changes in basal and stimulated gastricacid outputs.

In nine of 11 patients there was a rise inpH above 5-5 within two hours of comple-tion of vagotomy. The pH remained >4 fora mean duration of 20-8 hours (range4-41). There was no correlation betweenthis time period and postoperative reduc-tion in BAO. The pH returned to pre-operative levels in all patients within 14-54hours (mean 31-4). Pepsin concentrationsdid not mirror this pattern and showed a

variable response.These findings suggest that the operation

of vagotomy leads to a near total inhibitionof parietal and oxyntic cell function whichthen recovers to expected levels by thethird day.

T38Human gastric enterochromaffin-like(ECL) cells - demonstration of histaminecontent and its cholinergic nerve supply

W M HUI, H C LIU, AND S K LAM (Depart-ments of Medicine and Anatomy, Univer-sity of Hong Kong, Queen Mary Hospital,Hong Kong) The source of histamine inthe human stomach has not been estab-lished. To examine for possible relation-ship between the nerve supply and theendocrine ECL cells of the stomach andwhether the latter contain histamine, en-doscopic biopsies of the gastric fundus offour healthy subjects and eight patientswith duodenal ulcer were meticulouslystudied. Modified silver impregnationmethod was used to stain up simultaneous-ly the ECL cells and nerve fibres, andcholinesterase activity of nerve fibres wereexamined histochemically. Modified o-phthalaldehyde fluorescence method wasused to examine for possible histaminecontent of the ECL cells, which were thencounter-stained for ECL cells by Grimeliussilver method. The results were (i) o-phthalaldehyde staining was taken up bythe mast cells and the ECL cells, whichcorresponded to the argyrophil stainingcells. (ii) Nerve fibres possessing acetylcho-linesterase activity ramified to the base-ment membrane of the ECL and parietalcells, which were in close proximity to eachother. This intimate anatomical relation-ship and the demonstration for the firsttime that human ECL cells contain hista-mine strongly suggest that the human ECLcells play an important role in the controlof acid secretion, possibly under vagalinfluence.

T39Thromboxanes and gastric mucosal damage

C A PRICE, G PIPKIN, A CURRINGTON, L DAVIES,L DARLING, AND M E PARSONS (Smith Kline& French Research Limited, The Frythe,Welwyn, Hertfordshire) Thromboxane(TX) synthetase inhibitors reduce ex-perimental gastric damage in the rat. Ifthey act by preventing the formation ofpotentially damaging TX then a TX anta-

A1 145


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


gonist should also be effective. Dazoxiben(TX synthetase inhibitor, 30 and 60 mg/kgpo) reduced ethanol induced damage in theconscious rat. At 60 mg/kg% area ofmucosa damaged was reduced from16 3±2-3 (n=8) to 5 9±2 0% (n=8,p<0-01). The TX antagonists SK&F 88046(10 and 50 mg/kg po) and BM13177 (50mg/kg po) also reduced this form of dam-age. SK&F 88046 (50 mg/kg) reduced %area of mucosa damaged from 9-8±2-7(n=7) to 0-1±0-1% (n=6, p<0.01). Theduration of action of SK&F 88046 againstU46619 (TX mimetic) induced hyperten-sion was short (40 min) in the rat, thereforein the rat ex vivo gastric chamber prepara-tion SK&F 88046 was given as an iv bolus(20 mg/kg) followed by an infusion (40mg/kg/h). This dose reduced blood andalbumin loss into the gastric chamber aftertwo hours exposure to acidified Na tauro-cholate (50 mM). Blood (5'Cr red bloodcells) loss, was reduced from 11-9±3-4 to2-8±0 6 mlx 10-3 (p<0-05). SK&F 88046antagonised the effect of U46619 on ratplatelet rich plasma (IC50 4-2 ,uM) and onthe rat fundus strip (log KB 7.53) confirm-ing that SK&F 88046 acted as a TXantagonist in the rat, as in other species.These results suggest that TX may beinvolved in the pathogenesis of ethanol andbile induced damage in the rat.

BASIC SCIENCEF1-12

FlIdentification of human progastrin

H P DESMOND, S PAUWELS, R DIMALINE, AND

G J DOCKRAY (Department of Physiology,University of Liverpool, Liverpool, andCentre de Medecine, Nucleaire, Universityof Louvain, Brussels, Belgium) The genesequence encoding the human gastrin pre-cursor, progastrin, has been elucidated,but little else is known of the biosynthesisof this hormone. We have raised antibodiesto a synthetic analogue of the C-terminalhexapeptide of progastrin; unlike othergastrin antibodies, these are expected toreact with the intact precursor as well asfragments of it. They should therefore helpelucidate biosynthetic pathways. In humanantral mucosal extracts fractionated onSephadex G50 we found two peaks ofimmunoreactivity: a major one (84±2% oftotal; n= 10) corresponded to the C-

terminal tryptic fragment of progastrin(Kav 0.65), and a minor one was of higherapparent molecular weight (Kav 0.20). Incontrast, in three of eight gastrinomas, thelatter material accounted for over 50% oftotal immunoreactivity (mean 39+6%;n=8). Digestion of the high molecularweight form with trypsin liberated peptidesidentified by antibodies to the N-terminusof G17 and of G34, indicating a structureextending from the C-terminus of pro-gastrin to beyond the N-terminus of G34.We conclude that: (1) material with the

properties of intact progastrin is found inantral mucosa and gastrinomas. (2) Insome gastrinomas it is the predominantform of gastrin. (3) There are differencesbetween gastrinomas and antral mucosa inthe biosynthetic processing of progastrin.

F2Recombinant urogastrone-EGF can providea proliferative signal after small resection inparenterally fed rats

R A GOODLAD, W LENTON, H GREGORY, A PSAVAGE, K G MCCULLAGH, AND N A WRIGHT(Department of Histopathology, RoyalPostgraduate Medical School, Hammer-smith Hospital, London, ICI, AlderleyPark, Macclesfield, and G D Searle, HighWycombe, Bucks) It is now known thatanimals maintained on total parenteralnutrition (TPN) do not show the dramaticproliferative response to partial smallbowel resection seen in orally fed animals,suggesting that the presence of food in thelumen is an important signal. If this pro-liferative signal is hormonal it might beexpected that hormone(s) themselvescould initiate this process. The response ofthe gastrointestinal tract to urogastrone-EGF was investigated in rats maintainedon TPN with or without 75% small bowelresection. Orally fed rats (±resection)were also studied. Two way analysis ofvariance showed that resection caused asignificant increase (p<0-01) in prolifera-tion below the anastomosis and in theileum of TPN rats; however, the responseof the ileum was much less than thatobserved in orally fed rats, which confirmsthe importance of 'luminal nutrition' in theresponse to resection. There was no evi-dence for a significant positive interaction(synergistic effect) between the effects ofurogastrone-EGF and resection. Con-tinuous infusion of 60 gg/rat/day of recom-binant beta urogastrone significantly in-creased proliferation (measured by theaccumulation of vincristine arrested

metaphases) in the stomach (p<005),small intestine and colon (p<0-001) inresected and unresected rats maintained byTPN. Thus urogastrone-EGF has the novelproperty of being capable of stimulating aproliferative response in the resected intes-tine of parenterally fed rats.

F3Serotonin, a physiological role in gastro-intestinal motility?

G B HOPKINSON, J HINSDALE, AND B JAFFE(INTRODUCED BY J B ELDER) (Departmentof Surgery, Downstate Medical School,Brooklyn, New York, USA) Freyburgerin 1951 showed a spasmogenic action forSHT in the canine jejunum, but its physio-logical significance was unknown. Fivechronic conscious dogs fitted with gastricand jejunal strain gauges were infusedintravenously with SHT in doses between0 25 ,ug/kg/min and 120 Ag/kg/min over 20minute periods during the quiescent phase(Ph I) of the interdigestive myoelectricalcomplex (IDMEC). Over 200 such infu-sions were carried out. Coeliac artery SHTlevels were monitored by radioimmuno-assay. Below 1 Ag/kg/min no contractileresponse occurred. Between 1-4 Atg/kg/minthree dogs showed a response. At 8 Ag/kg/min all dogs showed responses to intra-venous 5HT. The gastric contraction rateincreased from 2 1±0 50 (contractions perminute (CPM)±SEM) before infusion, to4-6±0 21 CPM during infusion of 5HT at 8,ug/kg/min (p<0-001) (n=20). The gastriccontraction force increased from16 1±6 20 mV/mins before infusion, to145-8±14-1 mV/mins during infusion(p<0001). The jejunal contraction rateincreased from 1-2+0-57 CPM before infu-sion, to 16 4±0 50 CPM (p<0001)(n=20), with an increase in contractionforce from 6 1±3 10 mV/mins before infu-sion to 125 2±15 2 mV/mins. Arteriallevels before infusion of 5HT were658-3±187-6 ng/ml of whole blood andincreased to 719 5±68 7 ng/ml during infu-sion. The contraction force during 5HTinfusion was similar to that found duringthe spontaneous maximum contraction ofthe bowel. During infusion of 5HT at 8,ug/kg/min serotonin appears a majormediator of intestinal motility.

F4Glucocorticoid-induced Na and K channelsin the apical membrane of rat distal colon

A1 146


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


G I SANDLE (Department of Medicine.University of Manchester School of Medi-cine, Hope Hospital, Salford) Na and Kchannel blockers (amiloride and tetraethyl-ammonium chloride (TEA), respectively)have been used with microelectrodes tostudy stimulation of Na absorptive and Ksecretory processes in rat distal colon bydexamethasone (600 ,ug/100 BW/day forthree days). In dexamethasone treatedanimals (n= 15), transepithelial voltage(V,: -31±4 mV) and total conductance(Gt:8-5±0-5 mS/cm2) were higher than incontrols (-6±1 mV, p<0-001 and 5-3±0-2mS/cm2, p<0-001 respectively, n= 12);these changes reflected an increase inbasolateral membrane voltage, and de-creases in apical membrane voltage and thebasolateral/apical conductance ratio.Mucosal amiloride (10' M) and then TEA(30 mM) had no effects in control animalsIn dexamethasone treated animals, amilo-ride reduced Vt to the control value,decreased Gt from 8-5±0 5 to 6-9±0-3mS/cm2 (p<0-001), hyperpolarised theapical membrane by 13 mV (p<0-005), andincreased the basolateral/apical conduc-tance ratio by 63%, indicating inhibition ofapical Na channels and electrogenic Natransport. Subsequent addition of TEAdecreased Gt from 6-9±0-3 to 6*4±0 4mS/cm2 (p<O-001), depolarised the apicalmembrane by 8 mV (p<0025), and in-creased the basolateral/apical conductanceratio by 66%, indicating inhibition of apicalK channels. Thus, stimulation of Naabsorptive and K secretory processes in ratdistal colon by dexamethasone involves an

increase in the conductance of the apicalmembrane to both Na add K.

F5Is gall bladder emptying really exponential?

P HOWARD, G M MURPHY, AND R H DOWLING

(Gastroenterology Unit, Division of Medi-cine, UMDS of Guy's and St Thomas'Hospitals, London) Previous ultrasoundstudies of gall bladder (GB) emptying haveused standard Lundh meal or CCK stimuliand long intervals (>15 min) betweenimaging. The pattern of response to a

normal, dual phase (liquid plus solid) meal,is unknown. Therefore, in 10 control sub-jects, we measured GB size with frequentultrasound recordings, after a meal ofbaked beans on buttered toast with milk.

In nine of 10 subjects, the GB emptiedbefore the meal, the mean reduction in GBvol (A vol) recorded during this 'cephalicphase' being 5-28±SEM 1-76 ml. After the

meal, all 10 subjects showed the sametriphasic pattern of response with: (i) anearly phase of GB emptying from a meanpre-ingestion vol of 18-9±3-48 ml to11-21±2-35 (p<0-05) at 17*1±5-4 min; (ii)a net refilling phase to 16-6±3-24 ml at36-4±6-5 min; and (iii) a late emptyingphase to a final nadir vol of 5-6±1-51 ml at1-38±9-9 min. In four subjects, US imag-ing at more frequent intervals (1/min)showed that this overall triphasic responsewas punctuated by min to min fluctuationsin vol with mean filling and emptying ratesof 0-82 and 0-95 ml/min respectively.We conclude that, surprisingly, the pat-

tern of GB emptying after a dual phasemeal is not a simple exponential but isinterrupted by an episode of net refillingbefore the final nadir is reached, withsuperimposed minute by minute fluctua-tions in volume throughout.

F6Intestinal tissue oxygen tension measure-ment using a surface electrode

M B HALLETT, A SHANDALL, R H LOWNDES,H L YOUNG (INTRODUCED BY J V PSAILA)(University Department of Surgery, Univer-sity of Wales College of Medicine, HeathPark, Cardiff) If a surface electrode trulymeasures tissue oxygen tension (TO2), thiscould be used to assess intestinal perfusion/viability. To determine this we havestudied the relationship between TO2measured with the Clark electrode, bloodflow (F) measured by Xei33 clearance andarterial 02 tension (PaO2) measured byblood gas analysis, in a rabbit experimentalmodel.Using serial devascularisation in nine

rabbits F & TO2 were measured in mid-ileum and mid-sigmoid colon. In six rabbitsinspired 02 (FiO2) was varied between0-100% with PaO2 measured on rabbit eararterial blood, and TO2 on ileum and colonas above.

Using known physiological parametersof oxygen capacity, oxygen supply, oxygenconsumption, oxygen diffusion andhaemoglobin saturation a formula relatingF, PaO2 and TO2 was derived. Curvesbased on this showed the expected sigmoidrelationship between F and TO2, PaO2 andTO2, which fitted the experimental datawith a correlation of r=0-93, p<0-001(TO2 vs F), r=0-96, p<0-001 (PaO2 vsTO2) on linearising the equation. Thisrelationship would not be expected if theelectrode was merely detecting capillary02, or if ambient air was significantly

influencing the measurement.We conclude that the Clark electrode is a

true measure of tissue oxygen tension andcan be used clinically for this purpose.

F7Dissociated effect of atropine on hyper-gastrinaemia induced by single dose andrepeated omeprazole treatment

F HALTER, F EIGENMANN, AND H R KOELZ

(Gastrointestinal Unit, University Hospital,Inselspital Berne, Switzerland) Omepra-zole has been shown to increase serumgastrin and antral G-cell density in the rat.Acute hypergastrinaemia following acidinhibition induced by histamine H2-antagonists can be blocked by high doses ofatropine, both in fasting and fed rats. Westudied whether hypergastrinaemia in-duced by single or repeated dose treatmentwith omeprazole can be prevented byadditional atropine.

Gastric pH and serum gastrin weremeasured in 10 chronic gastric fistula ratsbefore and four hours after sc omeprazolee(0, 40 ,umol/kg); omeprazole combinedwith atropine (O+A, 40 ,mol/kg and 3mg/kg, respectively); atropine (A, 3mg/kg); or the omeprazole solvent (S, 40%PEG). Gastric pH of all animals treatedwith 0 or O+A became neutral within fourhours. After 0, fasting serum gastrin levelsrose from 34±4 to 110±9 pmol/l (x±SEM,p<O-00l), but were not influenced in theother groups. Four groups of six intact, fedanimals received similar treatment for 10days. 0 was given once daily, A twicedaily. In contrast to single dose treatment,both 0 and O+A increased mean serumgastrin levels by approximately seven-foldas measured four and 12 hours after the lastdose of 0 or O+A. Treatments A and Sdid not influence serum gastrin.We conclude that atropine blocks the

acute rise of serium gastrin after a singledose omeprazole administration. The morepronounced hypergastrinaemia followingprolonged omeprazole treatment, how-ever, is atropine resistant, suggesting achange in the regulatory mechanisms ofG-cells. It appears thus unlikely that G-cellhyperplasia following prolonged ome-prazole treatment represents a simple workhyperplasia directly related to acid inhibi-tion.

F8Prostaglandin E2 stimulates chloride secre-tion in guinea-pig isolated gastric mucosa

A1 147


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


C F SPRAGGS AND K T BUNCE (Department ofNeuropharmacology, Glaxo Group Re-search Ltd, Ware, Herts) The effect ofprostaglandins on gastric mucosal iontransport has been investigated in vitromainly in amphibian mucosa and littleinformation is available on the effect ofprostaglandins in mammalian tissue. Wehave investigated the effects of prosta-glandin E2 (PGE2) on sodium and chloridetransport and short circuit current (SCC) inguinea-pig isolated gastric mucosa in whichacid secretion had been inhibited by ome-prazole (100 ,uM).PGE2 (1 ,.M) produced increases in SCC

of 73±5 ,uA Cm-2. The ionic basis of thisresponse was determined by 36Cl and 22Naflux studies in mucosae pretreated withomeprazole (n=7). Under control condi-tions the net secretion of 36Cl (2-90±0-69,.uEq Cm-2/30 min) was not significantlydifferent from SCC (4-30±0.39 ,uEq cm /30 min, p>005), while the net absorptionof 22Na (-0.02±0-08 gEq Cm-2/30 min)was significantly different (p<0001).PGE2 (1 ,uM) stimulated a significantincrease in both SCC (5.36±0.31 uEqcm-2/30 min, p<0.001) and 36CI secretion(5 53±0 63 ,uEq Cm-2/30 min, p<0001),but had no significant effect on net 22Naabsorption (-0.27±0 12 ,Eq cm-2/30 min,p>O0O5).These results show that PGE2 stimulates

electrogenic chloride secretion in guinea-pig isolated gastric mucosa, and provide anionic basis for the stimulation of a NaCl-rich secretion by E-prostaglandins in mam-malian gastric mucosa in vivo.

F9Adaptive cytoprotection: evidence againstmediation by prostaglandins

C J HAWKEY, R T KEMP, R P WALT, N K BASKAR,J DAVIES, AND B FILIPOWICZ (Department ofTherapeutics, University Hospital, Notting-ham) Mild irritants - for example, 20%ethanol - increase gastric mucosal resist-ance to subsequent necrotising stimuli - forexample, 100% ethanol. We have investi-gated the proposition that this phenome-non (adaptive cytoprotection) occursbecause mild irritants stimulate prosta-glandin (PG) synthesis.Male Wistar rats were dosed orally with

vehicle or indomethacin 2-5 mg/kg or 10mg/kg. One hour later 20% ethanol (1 ml)was introduced into the stomach followed15 minutes later by 100% ethanol (1 ml)and mucosal necrosis subsequently quanti-tated macroscopically. Other rats were

killed 15 minutes after 20% ethanol and exvivo release of PGE2 from mucosal frag-ments measured by radioimmunoassay.

In control rats release of PGE2 was 28±3pg/mg/minute (mean±SEM, n=5):36±10% of the mucosa was necrosed by100% ethanol (n=7). After pretreatmentwith 20% ethanol PGE2 release was99±12% of control values but mucosalnecrosis induced by 100% ethanol wasreduced to 6±2% (p<0.05). After pre-treatment with indomethacin and 20%ethanol PGE2 release was reduced by62±15% (indomethacin 2*5 mg/kg,p<0-01) and 80±8% (indomethacin 10mg/kg, p<0-001) but there was only 8±1%(p<0-01) and 16±4% (p<0.05) mucosalnecrosis respectively.These observations show that adaptive

'cytoprotection' can occur in the face ofreduced prostaglandin synthesis and mayinvolve other mechanisms.

F10Arachidonic acid metabolism and leukocyteinfiltration, as determined by myeloperoxi-dase activity in a model of IBD

N K BOUGHTON-SMITH, J L WALLACE, ANDB J R WHITrLE (Department of MediatorPharmacology, Wellcome ResearchLaboratories, Beckenham, Kent) Achronic model of IBD in the rat, which hasthe pathological features of Crohn'sdisease can be induced by colonic adminis-tration of trinitrobenzene sulphonic acid(TNB). We have investigated colonic "'C-arachidonic acid (AA) metabolism and, asa measure of leukocyte infiltration,myeloperoxidase (MPO) activity, duringthe development of colonic inflammationin this model.The distal colons of rats were removed at

various times (one hour to one week) aftera single intracolonic application of TNB(20 mg in 0-25 ml of 30% EtOH). Seg-ments of colon (200 mg) were homogen-ised, incubated (30 min, 37°C) with14C-AA (0.25 ,uCi/ml) and the 14C-AAmetabolites separated by TLC. Cell freesupernatants of colon were also used forspectrophotometric (at 460 nm) determina-tion of MPO activity.The relative formation of eicosanoids

from 14C-AA by control colon wasPGE2>6-keto-PGF ,>HETE>PGE2>PGF2,, TXB2. After TNB administrationthere was a marked increase in metaboliteformation, at 24 hours HETE formationwas 317% control; 6-keto-PGF,:, 157%control and TXB2 133% control (all

p<0.05). MPO activity was also markedlyincreased (16-fold over control; p<005).Formation of HETE returned to controllevels by 72 hours, but formation of 6-keto-PGFI,,, and PGF2, was increased (160%and 133% control; p<0-05). MPO activityincreased further (35 fold, p<005) peak-ing between 54 and 72 hours. The MPOactivity remained raised at one week, atwhich time "4C-AA metabolism had re-turned to control levels.A significant increase in 14C-AA meta-

bolite formation was apparent during theacute phase of TNB induced colonic in-flammation which coincided with anincreased leukocyte infiltration (MPOactivity). In the chronic phase of theinflammation, however, there was littlechange in "4C-AA metabolism despite thepersistance of leukocyte infiltration. Theseresults suggest a greater importance ofthese AA metabolites in acute comparedwith chronic inflammation.

FliSegmental variability of ionic conductancesin rat colonic epithelium

G I SANDLE (Department of Medicine,University of Manchester School of Medi-cine, Hope Hospital, Salford) Segmentaldifferences exist in colonic ion transport.Microelectrodes and ion substitutions havebeen used to study passive (paracellular)transport in rat proximal and distal colon.In proximal colon (n=6) bathed in NaCl-Ringer, transepithelial voltage (V,:-5±1mV) was lower, and total conductance(G,: 11-4+0-8 mS/cm2) was higher, than indistal colon (-9±1 mV, p<0-05 and5-9±0-5 mS/cm2, p<0-001 respectively,n=9), and there was no amiloride-sensitiveNa conductance in either segment. Para-cellular shunt and apical membraneconductances were higher in proximalcolon (5*8±0.9 mS/cm2 and 5*0±1+1mS/cm2 respectively) than in distal colon(2.8±0-4 mS/cm2, p<0-01 and 2-2±0-3mS/cm2, p<0 02 respectively); basolateralmembrane conductance was similar in bothsegments. With Cl-free Ringer, total con-ductance in proximal colon decreased(AG,:6-1+1-2 mS/cm2, n=7) to a greaterextent than in distal colon (AG,:2-6±0-4mS/cm2, p<002, n=7), without a changein the basolateral/apical conductance ratio.With Na-free Ringer, there were smallerdecreases in proximal and distal colonicconductances (AG,:2-5±0-5 mS/Cm2 and1-3±0-5 mS/cm2 respectively, n=6). Thus:(i) proximal colon has a greater paracel-

A1148


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


lular conductance than distal colon, whichmainly reflects its permeability to Cl; (ii)proximal colon has a greater apical con-ductance than distal colon, which may

reflect K channels, as there appear to be no

Na or Cl channels in this membrane.

F12Activation of intestinal epithelial proteinkinases by calcium

G WARHURST, G S SMITH, A TONGE, AND L A

TURNBERG (Department of Medicine, Uni-versity of Manchester School of Medicine,Hope Hospital, Salford) Intestinalsecretagogues such as acetylcholine and5-HT are believed to mediate their actionsby raising cytosolic Ca concentrations. Thecellular events which follow and which leadto ion secretion are largely unexplored butmay involve the phosphorylation of keyproteins by specific protein kinases. Wetherefore investigated Ca activated proteinkinase activity in cytosolic and particulatefractions from rat enterocytes. Endo-genous protein phosphorylations deter-mined in cell homogenates incubated with32P ATP revealed a Ca dependent phos-phorylation of several proteins, the mostapparent being of Mr=50 000. Labelling ofthis protein was stimulated at 3 x 10'8 M Caand was maximal at 10- M Ca. Inclusion ofcalmodulin (CDR) had no effect on thelevel of phosphorylation, although theCDR antagonist TFP inhibited labellingsuggesting the involvement of CDR-dependent protein kinase activity.

Using a specific histone protein as exo-

genous substrate a second Ca activatedkinase could be demonstrated. Thisenzyme was located in the 100 000 g

cytosolic fraction and was dependent on

the presence of phosphatidylserine (PS) forfull activity. PS produced a dose-dependent stimulation of activity (4 ,.tg/ml,142%; 18 jig/ml, 235% of control). Inconclusion this study shows the presence inrat enterocytes of two distinct Ca activatedprotein kinase activities - both of whichmay serve to mediate the secretory eventsinitiated by increases in intracellular Ca.

GASTRODUODENALF13-24

F13Comparison between enprostil and raniti-dine in the treatment of gastric ulcerationand subsequent follow up (interim report)

A G MORGAN, W A F MCADAM, C PACSOO

(Endoscopy Unit, Airedale General Hospi-tal, Keighley, West Yorkshire) So far 48patients with benign gastric ulceration havebeen enrolled in a double blind endo-scopically controlled study, and randomlyallocated to treatment with either enpros-til, a synthetic dihydro-prostaglandin E2(70 i.g twice daily) or ranitidine. Endo-scopic examination was repeated at month-ly intervals for three months or untilhealing. Dyspeptic symptoms were re-

corded on a diary card and a return drugcount and safety screening performed ateach endoscopic visit.The results of the treatment were similar

for the two drug regimes. With enprostil63% had healed ulcers at one month, 91%at two months and three months. Thehealing rates for ranitidine were 50%,83%, and 96% respectively. Diary cardanalysis showed that both drugs rapidlyrelieved dyspeptic symptoms. No clinicallyimportant side effects were encountered. Ayears follow up after ulcer healing andwithout maintenance therapy is planned.So far there have been 20 recurrences outof the 32 patients who have completed thefirst six months of follow up.

Enprostil appears to be an effective andsafe drug in the treatment of gastriculceration.

F14Enprostil (E) versus ranitidine (R) induodenal ulcer (DU)

K D BARDHAN, K BOSE, R F C HINCHLIFFE,LESLEY WHITrAKER, PAMELA MORRIS, HELEN

MASSEY, AND MOIRA THOMSON (DistrictGeneral Hospital, Rotherham, and SyntexResearch, Maidenhead, Berks) We haveinvestigated enprostil (E), a new syntheticprostaglandin E2 derivative with antisecre-tory and mucosal-protective effects, in DUhealing. In a double blind double dummystudy, 85 patients with DU were randomlyallocated to receive either E 35 mcg (n=44)or R 150 mg (n=41) both taken twice daily.The patients were interviewed at two, four,and six weeks and endoscoped at fourweeks and again at six weeks if unhealedearlier.

Patients in both treatment groups were

well matched for age, sex, length of his-tory, smoking, and ulcer size. Healing inthe two groups was: at four weeks, E 46%,R 93% (p<0.01); at six weeks E 82%, R

97% (ns). Smoking retarded healing butulcer size and length of history had no

effect. Reduction of daytime and night-

time pain was equally quick with bothdrugs.The main adverse events noted were:

abdominal pain (E 16%, R 0%), diarrhoea(E 5%, R 2%), nausea (E 7%, R 2%),dizziness (E 2%, R 5%), and depression (E0%, R 5%); and three patients (El, R2)were withdrawn because of continuing orrecurrent abdominal pain. There was nomajor treatment related haematological orbiochemical abnormality.

In conclusion, at this dose E heals DUless rapidly than R but is equally effectivein relieving ulcer pain.

F15Natural history of chronic antral gastritisin duodenal ulcer (DU) and its reponse totreatment with prostaglandin El (miso-prostol)

W H HUI, J HO, S K LAM, I LUI, M T NG, C L LAI,

AND A LOK (Departments of Medicine &Pathology, University of Hong Kong,Queen Mary Hospital, Hong Kong) Thenatural history of chronic antral gastritis inrelation to healing of DU and its responseto treatment, if any, are unknown. Wecarried out a double blind controlled trialusing an oral prostaglandin El, miso-prostol (Searle), in 213 patients with activeDU randomised to receive placebo(n=69), misoprostol 200 ,ug (n=73), ormisoprostol 300 ug (n=71) qid respective-ly. Healing of DU was assessed bi-weeklyup to 12 weeks by endoscopy at which atleast two antral and two fundal biopsieswere taken. The activity and chronicity ofgastritis as assessed histologically by theinfiltration of, respectively, polymorphsand chronic inflammatory cells weregraded blind by the pathologist as nil, mild,moderate or severe. Before treatment,80% of patients had moderate to severeantral gastritis and 1.5% had fundal gastri-tis. In the placebo group, healed DU wasassociated with significantly (p<0.01, lifetable analysis) higher rates of regression ofantral gastritis (nil or mild as end point)than unhealed DU (30% vs 4% at week 8).Irrespective of whether DU was healed orunhealed, significantly (p<0.01) morepatients on misoprostol (50% at week 8)showed regression of antral gastritis thanthe placebo group. The chronicity of antralgastritis showed similar changes.

In conclusion, healing of DU was associ-ated with improvement of chronic antralgastritis, which, as shown for the first time,could be further enhanced by a therapeuticagent - prostaglandin El.

A1149


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


F16Will overnight response to cimetidine pre-dict healing of duodenal ulcers?

M DEAKIN, J RAMAGE, ANGIE PAUL, JO

SHOULER, S P GRAY, J BILLINGS, D G COLIN-

JONFS, AND J G WILLIAMS (Department ofGastroenterology and Biochemistry, RoyalNaval Hospital, Haslar, DepartmentGastroenterology, Queen Alexandra Hos-pital, Portsmouth, Devon) We have stu-died prospectively the evening and noctur-nal pH profiles, nocturnal acid and pepsinoutputs of 33 patients with endoscopicallydiagnosed acute duodenal ulcers. Thestudy was carried out before treatment andwhile taking cimetidine 400 mg bd (0800and 2300 hours). All patients were re-endoscoped after six weeks treatment withcimetidine: in 13/33 the ulcer had nothealed, in 7/33 an erosive duodenitis re-

mained and in 13/33 complete healing hadoccurred.

All 13 patients whose ulcers had healedhad a profound pharmacological responseto cimetidine 400 mg with a mean over-

night pH of 6*06±0-46 SEM comparedwith 1-92±0-14 before treatment and a

97% fall in nocturnal acid output,22-46±3-6 to 0-75±0*35 mmol between0030-0730 hours. While taking cimetidinefew overnight specimens contained detect-able peptic activity. Mean fall in pepsinoutput was 28-1±5-39 to 3-45±1-55 IU.The 13 patients with unhealed ulcers had

smaller pH changes overnight (1-47±0-05to 3.74±0.62) (p<001), a smaller reduc-tion in mean acid output, 57-29±18-07 to13-28±5-12 mmol (p<0.01) and pepsinoutput 45*96±11*71 to 22-16±5-6 IU(p<O0O1).A poor response to cimetidine overnight

will predict slow healing of duodenal ulcer-ation. This may be because of inadequateinactivation of pepsin as well as poorinhibition of acid.

F17Is persistent duodenal prostaglandin E2deficiency the cause of relapse in DU?

S PUGH, SIAN WILLIAMS, M ISHAQUE, M R

LEWIN, TINA BARTON, K BOSE, K BARDHAN,AND C G CLARK (Department of Surgery,University College London, and Depart-ment of Gastroenterology, Rotherham Dis-trict General Hospital, Rotherham) It hasbeen shown that the ability of the duodenalmucosa to synthesise PGE2 is deficient inassociation with DU. As PGE2 may be an

important mediator of duodenal defences

such a deficiency may be causally related toDU. We report a two centre study on theeffects of treatment with H2 receptor anta-gonists and the healing of the DU onduodenal PGE2 synthesis. At endoscopy,biopsies were obtained from the duodenumin normals (22), untreated DU rim (18),patients with healed DU but still on H2receptor antagonist treatment (22),patients with healed DU off treatment (6)and also patients being treated with H2receptor antagonists for GU or oesophagi-tis (the patients act as treatment controls(20)). All biopsies were treated similarly byinducing synthesis of PGE2 by vortexingand measuring released PGE2 by RIA.Results were (mean±SD in pg PGE2/mgwet wt), normals 110-2±32-2, untreatedDU rim 60-3±22-5 (p<0.001 vs normal),healed DU on treatment 84-7±38-2(p<0-025 vs normal), healed DU off treat-ment 75-7±26-1 (p<0-02 vs normal, NS vshealed DU on treatment) and patients ontreatment for conditions other thanduodenal disease 121-6±40-2 (NS vsnormal, p<002 vs healed DU on treat-ment).We conclude that the initial deficiency of

PGE2 in association with DU is confirmedand that this deficiency persists despitetreatment with H2 receptor antagonists.This persistent deficiency may be one ofthe causes of the rapid recurrence of DU inmost patients off treatment.

F18Prostaglandin E2 in the prevention ofgastric stress bleeding

H A VAN ESSEN, M VAN BLANKENSTEIN, J H PWILSON, B VAN DEN BERG, AND H A BRUINING(Departments of Internal Medicine andSurgery, University Hospital, Rotterdam,The Netherlands) The effect of prosta-glandin E2 (PGE2) in the prevention ofacute gastric stress bleeding in intensivecare patients was investigated in a prospec-tive, double blind, placebo controlledstudy. Ninety patients with two or morerisk factors (major surgery, multipletrauma, respiratory insufficiency, renal in-sufficiency, jaundice, hypotension, peri-tonitis, sepsis) were randomised for treat-ment with either PGE2 0-5 mg every fourhours via a nasogastric tube, or placebo.Blood loss in 24 hour gastric aspirates wasmeasured by a peroxidase test (orthotoli-dine) and by 51Cr-labelled autologouserythrocytes, a loss >15 ml/day being thecritierion of bleeding.

Fifty seven patients could be evaluated

after at least three days: 29 had receivedPGE2 and 28 placebo. Bleeding occurredin nine (31%) of PGE2 treated patients andin 13 (46%) placebo treated patients (NS).No correlation was found between theblood content of gastric aspirates as mea-sured by the 5"Cr-method and the peroxi-dase test.

It is concluded that (1) PGE2 0.5 mgadministered intragastrically four hourlydid not provide adequate protectionagainst stress bleeding. (2) Peroxidase testscannot be used to quantify blood loss ingastric aspirates. The results of previoustrials on the prevention of stress bleedingusing peroxidase tests are thereforequestionable.

F19Effect of aluminium on bicarbonate secre-tion by isolated amphibian gastroduodenalmucosa

J R CRAMPTON, L C GIBBONS, AND W D W REES(Department of Medicine, Hope Hospital,University of Manchester School of Medi-cine, Salford) Aluminium containingantacids provide symptomatic relief ofdyspepsia despite the negligible bufferingcapacity of doses commonly used. Recentevidence suggests that aluminium antacidsmay possess cytoprotective propertiesalthough the possible mechanism of suchaction is not clear. The effect of neutralaluminium salts on bicarbonate secretionby bullfrog (Rana catesbeiana) fundic,antral, and duodenal stripped mucosa hastherefore been examined. An isolatedchamber preparation has been used enabl-ing measurement of bicarbonate secretionby pH stat titration of the luminal solutionwith recording of transmucosal potentialdifference. Addition of neutral aluminiumsulphate 3 x 10-3 M (equivalent to onetablet of aluminium hydroxide in 2 1)caused a marked increase in bicarbonatesecretion by antrum (mean±SE:214±63%, n=4, p<0-05), fundus(mean±SE: 144±48%, n=5%, p<0-05)and duodenum (mean±SE: 133+44%,n=6, p<0-005). Transmucosal potentialdifference was not altered during theseexperiments. These results demonstratethat aluminium is a potent stimulant ofmucosal bicarbonate secretion in concen-trations which may be achieved in vivo.The mechanism of this stimulation de-serves further evaluation since it mayprovide a clue to the therapeutic effect ofaluminium containing antacids in pepticdisorders.

A1150


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


F20Experimental studies of new mechanicalmethods of endoscopic haemostasis: stitch-ing, banding, clamping, and ulcer removal

C P SWAIN, T N MILLS, AND T C NORTHFIELD(The Norman Tanner GastroenterologyUnit, St James' Hospital and Department ofMedical Physics, University College Hospi-tal, London) Mechanical occlusion ofbleeding vessels in the gastrointestinal tractat endoscopy might offer greater securityof haemostasis than thermal methods. Wehave designed and tested new mechanicalmethods for occlusion of bleeding vesselsat endoscopy including: (1) an improvedendoscopic sewing machine (SM); (2) anulcer clamp (C); (3) a method for deliver-ing rubber bands (B) or self-retainingplastic ties over a suction polyp at endo-scopy; (4) banding or tying of a suctionpolyp over an acute ulcer followed bymonopolar snare ulcer removal (UR).Forty five standard bleeding canine ulcerswere randomised to endoscopic treatmentwith SM, B, C, UR or control. Bleedingwas terminated by SM in 9/10, C 10/10, B9/10, UR 10/10. Controls did not stopspontaneously. No secondary rebleeding orperforation occurred in these survivalexperiments. Mechanical methods werecompared with thermal methods; hotsqueeze bipolar forceps (HS), heater probe(HP), monopolar probe (MP), Nd-YAGlaser (YAG) and argon laser (A) in bleed-ing mesenteric vessels and isolated arteriesof 1-4 mm, measuring bursting pressure ofvessel occlusion. Mechanical occlusionachieved higher (p<0:01) bursting press-ures (mean±SEM 1200±230 mmHg) thanthermal methods (320±120 mmHg) (SM+C>HS+B+UR>HP+YAG with coapta-tion >MP+YAG>A).We conclude that permanent mechanical

occlusion of acute experimental ulcers bymeans of stitching, banding, clamping andulcer removal is feasible and can occludelarge vessels more securely than thermalmethods.

F21Trial of the 7 FG bipolar probe in bleedingpeptic ulcers

J D O'BRIEN AND W R BURNHAM (Depart-ment of Gastroenterology, Oldchurch Hos-pital, Romford, Essex) Four hundred andsixty patients with upper gastrointestinalbleeding over a 27 month period weregastroscoped by two endoscopists. Twohundred and four patients examined within

24 hours who had peptic ulcers with activebleeding, visible vessel or adherent clotwere allocated randomly to electrocoagula-tion with the probe (101 patients) or not(103 patients). Otherwise all patients weretreated identically. Groups were stratifiedby ulcer site to give similar numbers ineach. Management decisions and assess-ment of rebleeding were made by a clini-cian unaware of the randomisation.The treated patients were older than

controls (mean age 68-7 years vs 64 6years; p=0-054) had a lower initial meanhaemoglobin (9.8 g vs 10-2 g) and lowermean transfusion requirement after endo-scopy (4.6 units vs 7-3 units; p=013).Seventeen treated patients continuedbleeding or rebled compared to 34 controls(X2=628, p<0.015); benefit was mostmarked in those actively bleeding; sevenhad surgery (10 controls) and nine diedafter rebleeding (12 controls). One third ofthe mortality and nearly one quarter of therebleeds in the treated group were in thefirst seven patients treated suggesting thatinexperienced application of the probe maybe hazardous.

F22Risk of gastric cancer after benign ulcersurgery

G D CORCORAN, J WARE, D W DAY, R F A

LOGAN, AND S GRAY (Departments ofSurgery and Pathology, University of Liver-pool, and Department of CommunityHealth, Nottingham University, Notting-ham) The gastric adenocarcinoma (GC)registrations between 1970-79 for theMersey Region have been reviewed and acase controlled study from five inner cityhospitals done to estimate the relative risk(RR) of late malignancy (OSC) after ulcersurgery.Of the total number of GC, 1*8% were

OSC cases (116/6613), GC mean age 68*6years and OSC, 66.4 years. The meaninterval between operation and diagnosisof OSC was 21-8 years while the threeoperative categories, numbers and age atoperation were: gastric resection (GR) 80,44.9 yr; gastrojejunostomy (GJ) 27, 42*2yr; and vagotomy and drainage, nine, 42-4yr. There were 1610 registrations including38 OSCs from five hospitals. These caseswere matched for age, sex, time of deathand hospital with necropsy controls. Theoverall RR was 1-4 (95% conf limits0.8-2-3) and the RR for GR (n=28) was1.3 (0.7-2-2) and for GJ (n=10) was 2-0(0.7-5.8).

It is concluded that OSC represents onlya small number of all gastric cancers.Moreover, it is suggested that the largerRR for OSC previously reported mayreflect the character of the control materialrather than a real risk.

F23Pathologists problems in the recognition ofearly gastric cancer and gastric dysplasia

A B PRICE, G WILLIAMS, H THOMPSON, ANDBIBA UNWIN (Departments of Pathologies:Northwick Park Hospital, London, Univer-sity Hospital of Wales, Cardiff, Birming-ham General Hospital, and Department ofSurgery, Leeds) This is preliminary patho-logical data from a multicentre investiga-tion of early gastric cancer (EGC), and thenatural history of gastric dysplasia. Thisreport concerns interobserver variation of41 cases of EGC, 22 with pregastrectomybiopsies, and a second biopsy group with aconsensus opinion of dysplasia from 13non-operated patients (follow up threeweeks - two years). Assessment was bythree pathologists using a proforma of 20attributes.For 41 cases of EGC there was full

agreement in 40 surgical specimens. Therewere six disagreements amongst the 22pregastrectomy biopsies from this group,with never more than one dissenting opin-ion. Only once was there dissention overthe actual presence of dysplasia, five.of thesix disagreements involved grades of dys-plasia.Of the 13 non-operated cases with at

least one biopsy classified dysplasia, theinitial 'blind' assessment produced totalagreement in only four. In six at least oneopinion was against dysplasia regardless ofgrade. This reflects the pathologist's prob-lem with the lesser grades of dysplasiapredominant in this group in contrast to theEGC group and borne out by analysis ofthe proforma attributes by Kappa statisticsand positive predictive values. Clearly theinterpretation of graded gastric dysplasia,like dysplasia in colitis, suffers interpreta-tive difficulties, while follow up from thisstudy will determine any clinical signifi-cance.

F24BSG early gastric cancer/dysplasia survey:the first 104 cases

F T DE DOMBAL, B J UNWIN, P COTrON, G R

GILES, A G MORGAN, A B PRICE, H THOMPSON,

A1151


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


AND G T WILLIAMS (University of Leeds,Leeds) At the BSG/SKF InternationalWorkshop on early gastric cancer (EGC) in1982, it was decided to set up a register ofpatients diagnosed as having 'EGC' or

'gastric dysplasia' in Britain, and create a

databank of information about these pa-

tients. This interim report outlines some

initial findings of the survey.

By May 1985, some 59 BSG membersfrom 22 hospitals had contributed 104 fullyevaluated patients (a further 40 patientsawait review). After evaluation by theBSG panel (three independent patholo-gists) only 41 cases were confirmed as

having EGC; 21 cases were classified as

'advanced gastric cancer', 25 patients as

'dysplasia' and 17 as other conditions.All but five EGC cases presented with

pain, usually epigastric. The commonestmode of presentation (15 cases) was find-ing cancer in association with an existinggastric ulcer. Eleven patients were on

long term H2 blocking drugs when theircancer was discovered. Only two patientshad undergone previous gastrectomy.These interim results emphasise (i) dif-

ferences between UK and Japanese pre-sentation, a high proportion of UK cancers

occurring in 'benign' gastric ulcers; and (ii)dangers in prescribing H2 blockers for'benign' gastric ulcer without regularbiopsy.

COLORECTALF25-38

F25Collagenous colitis: a report of five cases

C H MASON AND D P JEWELL (Departmentsof Histopathology and Gastroenterology,John Radcliffe Hospital, Headington,Oxford) We have studied five cases ofcollagenous colitis occurring in four womenand one man, varying from 62 to 79 yearsof age. Four patients had a history ofwatery diarrhoea. The fifth case had a 20year history of intermittent diarrhoea withoccasional bleeding suggestive of ulcerativecolitis. Diagnosis was made on colorectalbiopsy specimens by the presence of asubepithelial collagen band greater than1O,u in thickness.Two patients had borderline collagen in

the rectum (5-17,) but had a welldeveloped band (12-50,u) throughout therest of the colon similar to that seen in therectum of the other three. The collagen

band showed marked variability in thick-ness. All cases showed a mild, non-specificinflammation. One case had had multiplerectal biopsies over several years. Initially,these showed non-specific chronic inflam-mation but subsequently a collagenousband appeared. Immunohistochemistryshowed the collagen bands to be composedmainly of Type III collagen with somefibronectin. One case responded dramatic-ally to steroid enemas and two showedpartial response to oral sulphasalazine.

Collagenous colitis may exhibit minimalrectal involvement and the diagnosis couldbe missed on rectal biopsy. Clinical obser-vation and histochemical staining are com-patible with collagen developing inresponse to an inflammatory stimulus.

F26Does super-efficient starch absorptionpromote diverticular disease?

J R THORNTON, A DRYDEN, J KELLEHER, AND

M S LOSOWSKY (Department of Medicine,St James's University Hospital, Leeds)Populations eating a relatively low fibrediet have an increased prevalence of diver-ticular disease, but the factors determiningindividual susceptibility to this disease re-main unclear. We, and others, have shownpreviously that dietary starch is incom-pletely absorbed and that unabsorbedstarch is a quantitively important source ofcolonic carbohydrate additional to thatprovided by fibre. The degree of starchmalabsorption shows considerable varia-tion between individuals.We tested the hypothesis that super-

efficient starch absorption, by reducing theprovision of colonic carbohydrate, maypromote diverticular disease. Eight pa-tients with extensive, symptomatic diver-ticular disease were compared with eightage- and sex-matched healthy controls. Onseparate days, all subjects consumed, inrandom order, a standardised potato mealproviding 60 g starch or 6-5 g lactulose.Breath H2 was measured every 15 minutesfor up to 12 hours. The relative quantitiesof H2 generated enabled calculation of theamount of malabsorbed potato starch. Theamount of unabsorbed potato starch pro-vides a good approximation of unabsorbedstarch from all food sources.

Percentage unabsorbed starch was low inall patients and was only about one quarterof that found in the controls (mean±SEM:3-3%±0 5% vs 12-4%±1*8%, p<0-01).For the average Briton consuming 150 gstarch and 21 g fibre daily, this difference

represents nearly 14 g of unabsorbedstarch, equivalent to almost two-thirds ofthe colonic carbohydrate provided by con-sumption of fibre. Mouth-to-caecum tran-sit time of unabsorbed potato starch wassimilar (patients 321 min vs controls 272min, NS).

Super-efficient starch absorption, byreducing the provision of colonic carbo-hydrate, may promote the development ofdiverticular disease.

F27Effect of ampicillin on the colonic salvage ofcarbohydrate

S S C RAO, C A EDWARDS, C J AUSTEN, V A

BEATTIE, N W READ, AND C D HOLDSWORTH(Clinical Research Unit, Royal HallamshireHospital, Sheffield) The colon normallysalvages unabsorbed carbohydrate bybacterial conversion to volatile fatty acids,which are then rapidly absorbed togetherwith water. Impairment of bacterialfermentation may therefore result in anosmotic diarrhoea in people ingesting un-absorbable carbohydrate. We have testedthis hypothesis by studying the effect of 500mg ampicillin tid on stool weight andfrequency and breath hydrogen productionin 13 normal volunteers before and afteradministration of 20 g lactulose. Studieswere separated by two weeks and dietaryintake was similar during each study.Administration of lactulose did not in-crease stool weight and frequency undercontrol conditions (190-1±31-4 g/day vs195-4±28-6 g, mean±SEM and 1-38±0-28motions/day vs 1.46±0.13) but afteradministration of ampicillin stool weight(273-3±62-2 g vs 390-7±43 g, p<0.02) andstool frequency were significantly in-creased (1-38±0-26 vs 2*30±0 35, p<0-01).Ampicillin did not significantly increasebasal stool weight and frequency underbasal conditions. There was no significantchange in the breath hydrogen response tolactulose after ampicillin, although in twosubjects the response was nearly abolished.Mouth-to-caecum transit of the lactulosemeal was prolonged during antibiotic in-gestion (49±5.9 min vs 75±7-4 min,p<0-05) but there was no significantchange in the whole gut transit time. Theseresults suggest that antibiotic associateddiarrhoea could result from impaired sal-vage of carbohydrate.

F28Emergence of Clostridium difficile and

A1 152


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


disturbance of faecal flora after single doseintravenous antibiotic administration

R G TUDOR, N S AMBROSE, M JOHNSON, D

YOUNGS, D W BURDON, AND M R B KEIGHLEY

(Department of Surgery, The GeneralHospital, Birmingham) We have studiedthe effect on faecal flora and the emer-gence of Cl difficile (CD) over a two weekperiod after a single dose of eight cephalo-sporins and seven penicillins. Each anti-biotic was administered to six volunteers(total 90 volunteers) and a further group ofsix volunteers served as controls. Stoolsamples were obtained prior to antibioticadministration and at 1, 4, 7, and 14 dayssubsequently, and at similar time intervalsin the controls. Total viable counts ofindividual bacterial and fungal species wereobtained and the presence of Cl difficileand its toxin specifically sought.

Alteration in faecal flora by antibioticswas estimated by expressing the number ofindividual bacterial species present in allfive specimens as a percentage of the totalnumber of species identified.These results show that the cephalo-

sporins produce a more marked disturb-ance in faecal flora and this is oftenassociated with emergence of Cl difficile.

F29Use of monoclonal antibodies to ras-oncogene product in the differentiationbetween benign and malignant diseases ofthe colon

N A HABIB, H NIMAN, A THOMPSON, R C N

WILLIAMSON, AND C B WOOD (DepartmentofSurgery, RPMS, London, Department ofSurgery, Bristol Royal Infirmary, and De-partment of Molecular Biology, Scripps ResCenter, La Jolla, California, USA) Protooncogenes are responsible for normal cellgrowth and their conversion to a cellularoncogene leads to abnormal cell prolifera-tion.The expression of c-ras oncogene on

chromosome 12 has been demonstrated incarcinoma of the colon and rectum. Usingthe peroxidase-anti-peroxidase techniquecolonic tissue (obtained from mountedparaffin blocks) was stained with a mono-clonal antibody against ras-oncogene pro-duct (Mo-RAP) to a concentration of1:200. The patients studied were fivenormal subjects and 25 patients withvarious pathological conditions (fiveCrohn's disease, five ulcerative colitis, fivediverticular disease, 10 colorectal carci-noma). A sharp contrast was found

between the strong positive staining in thecytoplasms of all 10 carcinomas and theweak cryoplasmic staining mainly at thecrypt base in normal and inflamed largeintestine. Therefore, staining of colorectaltumour tissue with Mo-RAP may open newavenues for diagnostic and therapeuticmodalities. These findings suggest a muchstronger expression of ras-oncogene incancer tissue compared with normal orinflammatory tissue, and that Mo-RAPstaining offers new diagnostic and thera-peutic opportunities in colorectal malig-nancy.

F30Histological grading of rectal cancer: amultivariate analysis

J R JASS AND W S ATKIN (St Mark's and StBartholomew's Hospitals, London) Thepathological grading of rectal carcinoma isa subjective exercise associated with con-siderable interobserver variation. The aimof this study was to identify and rankprognostic factors in order to refine currentmethods of grading.

Tissue sections through the primarytumours from a consecutive series of 447patients were examined. All patients hadsurvived radical surgery for rectal cancerfor at least 28 days and had been followedfor up to 20 years. Histological factorsgraded subjectively were: architecture(regular, irregular, no glands), nuclearpolarity (easily discerned, just, lost), pat-tern of growth (expanding, infiltrating),lymphocyte infiltration at growing edge(marked, moderate, little) and fibrosis(little, moderate, marked). Lymph nodestatus was not known. Multivariate analysisusing the Cox regression model was em-ployed.

Factors most strongly related to survivalwere ranked: (1) lymphocyte infiltration,(2) pattern of growth, (3) architecture, (4)nuclear polarity, (5) fibrosis. A prognosticmodel was constructed by adding eachfactor in turn to the first. Prediction ofsurvival was not improved beyond architec-ture. When lymph node status and extentof spread were added to the model, onlylymphocyte infiltration was found to besignificant (p<0-001). The overriding in-fluence of lymphocyte infiltration uponsurvival has not been demonstrated pre-viously.

F31Frequency of colorectal cancer among the

flrst-degree relatives of patients with canceror polyps of the large bowel

M PONZ DE LEON, A ASCARI, A ANTONIOLI, F

MANENTI, G MELOTI, C PEZCOLLER, I PIC-CAGLI, A GRISENDI, D MAZZEO, AND A

MISELLI (Istituto di Patologia Medica,Cattedra di Gastroenterologia, Divisione diChirurgia Generale, Universitd di Modena,Italy) Despite the importance of en-vironmental factors there is evidence thatcancer of the large bowel may involve agenetic component. The institution of aregistry for colorectal tumours in ourhealth care district gave us the opportunityto test if close relatives of patients withcolorectal cancer or polyps are more likelyto develop large bowel cancer than mem-bers of the general population. For eachregistered patient a careful clinical historywas taken and the genealogic tree, limitedto the first-degree relatives, was drawn.Particular attention was given to relativesaffected or deceased of neoplastic diseases.Each patient was matched to a control -that is, a patient of the same age (±5 years)and sex hospitalised but not for neoplasticor colonic diseases. At one year, a total of139 cases of cancer and 157 of polyps wereregistered; there were 2202 first-degreerelatives in the diseased group (1325 alive)and 2203 in the controls (1328 alive).Among the relatives of patients withtumours we found 71 cases of colorectalcancer as compared to 16 in the controls(Relative Risk=RR=4-55, X2=34 1,p<O.00l). As for the parents there were 23cases vs 10 (RR=2-35, p<0.02); among thesiblings 44 vs 4 (RR= 12.2, p<0001); as forsons 4 vs 2 (ns). When considered separate-ly, an increased frequency of cases ofcolorectal cancer among first-degree rela-tives was found either in the cancer or inthe polyps group.

It is concluded that large bowel canceroccurs four times more often in relatives ofpatients with colorectal cancer or polypsthan in controls. Nearly 20% of the regis-tered tumours may be defined as familial.Since a familial aggregation was observedboth in the cancer and in the polypsgroups, our findings provide further evi-dence to the 'polyp-cancer theory'.

F32Use of monoclonal antibodies against ras-oncogene products to inhibit the metab-olism of human colorectal cancer cells invitro

N A HABIB, C B WOOD, H L NIMAN, B FERMOR,

A1153


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/

A1154

M 0 SYMES, AND R C N WILLIAMSON Depart-ment of Surgery, Royal Bristol Infirmary,Bristol, Department of Surgery, RoyalPostgraduate Medical School, London, andDepartment of Molecular Biology, ScrippsInstitute, La Jolla, USA) Proto-oncogenesare responsible for normal cell growth.During malignant cell transformation theybecome activated oncogenes and producegrowth factors. c-ras oncogene is specifi-cally expressed in colorectal carcinomas.To test their potential therapeutic role,monoclonal antibodies against ras-oncogene products (Mo-RAP) were ap-plied to disaggregated human colorectalcancer cells obtained from nine primarytumours. After 24 hours exposure in tissueculture for Mo-RAP, the tumour cells werecultured to a further 24 hours with 1 ,uCi/mlof radioactive selenomethionine (75Se).Protein synthesis was then measured interms of incorporated radioactivity.Mo-RAP caused marked inhibition of

protein synthesis by colorectal cancer cellsobtained from four out of the nine patients.The isotope uptake relative to untreatedcells was significantly reduced in fourtumours (p<005, <0-02, <0-001 and<0.005). In all these four patients a cytos-pin preparation of the tumour cells stainedspecifically with Mo-RAP. The widespreadexpression of oncogenes and their encodedproteins in human malignancies could thusprovide possible therapeutic targets.

F33Monoclonal antibody binding to primaryand metastatic colorectal cancer

K C BALLANTYNE, L G DURRANT, N C ARMI-TAGE, R A ROBINS, R W BALDWIN, AND J D

HARDCASTLE (Departments of Surgery andCancer Research, University of Notting-ham, Nottingham) The pattern of antigendistribution in primary colorectal tumourshas been defined by monoclonal anti-bodies. The distribution and degree ofantibody binding to metastatic, however,compared with primary colorectal cancerhas not been fully evaluated. Using flowcytometry and immunohistology weassessed the binding of the monoclonalantibodies 791T/36 - antiosteosarcoma,C14/1/46 - anticolonic adenoma andC161/25 - anticarcinoembryonic antigen,to 35 primary colorectal cancers, 13 lymphnode metastases and nine hepatic/omentalmetastases.

After disaggregation tumour cell bindingwas measured by flow cytometry usingindirect immunofluorescence. Fluoresc-

ence values were corrected for non-specificbinding using normal mouse immuno-globulin. Immunohistology was carried outusing the indirect peroxidase technique.The median linear fluorescence

(fluorescence units FlU) for each antibodywas: 791T/36 162 FIU (primary), 180 FIU(node), 239 FIU (metastasis), C14/1/46 407FIU (primary), 623FlU (node), 833 FIU(metastasis), C161/25 899 FlU (primary),380 FIU (node), 1312 FIU (metastasis).Immunohistology demonstrated that theantigen distribution of primary colorectalcancer is retained in metastatic tumourdeposits.Confirmation that metastases retain

similar pattern with increased antigenexpression to primary colorectal cancerprovides further evidence that targetedimmunotherapy may effectively treat meta-static colorectal cancer.

F34T lymphocyte subpopulations in the blood,primary tumour and normal colonicmucosa of patients with colorectalcarcinoma

T W J LENNARD, A WARFORD, B K SHENTON, GPROUD, R M R TAYLOR (INTRODUCED BY I D AJOHNSTON) (Departments of Surgery andPathology, University of Newcastle uponTyne, Newcastle upon Tyne) The degreeof 'host response' within tumours has beenrelated to prognosis. The immunoregula-tory T lymphocyte subpopulations in 36primary colorectal carcinomas have beencharacterised using cryostat immuno-histochemistry and a panel of monoclonalantibodies to pan T (OKT3), helper(OKT4) and suppressor/cytotoxic (OKT8)lymphocytes. Normal colonic mucosa fromthe proximal resection line of 26 of thetumours was also analysed. The mean ratioof helper to suppressor/cytotoxic lympho-cytes ±SD in the stroma of the tumourswas 0-55±0-21, and in the lamina propriaof normal mucosa 1-68±0-34 (p=<0-001)(Student's paired t test). The mean pheri-pheral blood ratio ±SD in 34 patients withcolorectal carcinomas was 1-5±0-28. Thefinding of a significant increase in suppres-sor/cytotoxic lymphocytes within colorectalcarcinomas compared to normal mucosaand peripheral blood may contribute todown regulation of the host immuneresponse in situ. These findings wereindependent of histological grade or stageand the presence or absence of venousinvasion by tumour.

The British Society of GastroenterologyF35Cell cycle kinetics during experimentalcolonic carcinogenesisT COOKE, J L MAlTHEWS (INTRODUCED BY APARKINS) (Charing Cross Hospital,Fulham Palace Road, London) Increasesin crypt cell production rate, maximal inthe descending colon, occurs during in-duced carcinogenesis. To determine itsmechanism, cell cycle kinetics in coloniccrypts were investigated.At week 25, rats treated with azoxy-

methane or saline sc for 12 weeks weregiven vincristine 1 mg/kg ip, to arrest cellsin metaphase, and killed at 15 minuteintervals. Metaphases were counted in 100descending colon crypts per animal ex-pressed as a percentage of total crypt countand plotted against time. Entry into mitosiswas calculated from the slope of the result-ing regression line and cell cycle timederived as its reciprocal.There was a significant hyperplastic re-

sponse in treated rats compared with con-trols increasing from 36-69±2-71 to48-84±4-64 cells crypt (p<0-001). Theproliferation zone increased from 23 to 29cells although the proportion of dividingcells (growth fraction) did not alterbetween controls (0.62) and carcinogengroups (0-59). Rate of entry into mitosisdecreased significantly from control valuesof 6-48± to 0 87% cells per hour (p<0-001)and cell cycle time increased from 30-86 to59 09 hours.

Absolute increases in the numbers ofproliferating crypt cells, despite a slowingdowli of cell division, explains the pre-viously observed elevation in crypt cellproduction rate (1), advancing our under-standing of the mechanism of coloniccarcinoma development.

F36Specialist investigation of obscuregastrointestinal haemorrhage

R SALEM, J N THOMPSON, A P HEMINGWAY, H CREES, H J F HODGSON, C B WOOD, D J ALLISON,AND J SPENCER (Departments of Surgery,Radiology and Medicine, Royal Post-graduate Medical School, HammersmithHospital, London) The cause of gas-trointestinal haemorrhage is not identifiedby routine investigations in about 5% ofcases. Over six years 131 patients with'obscure' gastrointestinal bleeding havebeen investigated. The median age was 62years (range 10-95) and 70 patients were


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


men. Only 25 presented as emergencies,

while 87% were referred for specialistinvestigation. The mean presenting featurewas melaena (55 cases), anaemia (35),rectal bleeding (34), haematemesis (six),and ileostomy bleeding (one). The medianduration of symptoms was 50 weeks (iangeI day-36 years) with a median two previoushospital admissions (range 0-20). Twentysix had undergone previous surgery.

The lesions identified were colonicangiodysplasia (52 patients), small bowelvascular anomalies (16), Meckel's diver-ticula (nine), leiomyomas (seven), gastricvascular anomalies (four), chronic pan-

creatitis (three), colonic diverticulardisease (three), and 16 single other causes.

Sixty per cent of lesions were first shownon visceral angiography, 22% at surgery,10% on endoscopy, 5% on barium exami-nations and 3% at ERCP. Seventy eightpatients underwent surgery. No lesion was

found in 21 cases (16%).Specialist investigation, especially expert

angiography, identified the cause ofobscure gastrointestinal haemorrhage inthe majority of cases. In a significantproportion, however, the cause was notfound until exploratory laparotomy.

F37Functional changes after mucosal proctec-tomy with colo-anal anastomosis (CSA) forchronic radiation rectal injury

J S VARMA, A N SMITH, AND A BUSUUTIL

(University Department of Surgery! Urologyand Department of Pathology, WesternGeneral Hospital, Edinburgh) Despite itssuccess in dealing with the complications ofchronic radiation rectal injury, markedurgency and frequency of defecation andanorectal incontinence occasionally followCSA. To elucidate the cause, anorectalphysiology was compared in eight patientsand matched control subjects. Conven-tional manometric techniques were used tomeasure rectal compliance (RC, ml/cmH2O), maximal tolerable volume (MTV,ml H2O), resting (MRP) and squeeze (SP)anal pressures (cm H.O), sphincter length(HPZ, cm) and the rectosphincteric reflex(RSR). Concentric needle EMG deter-mined pelvic floor activity during contrac-tion, straining and balloon rectal disten-sion. Resected rectal specimens were

examined using conventional techniquesfor histopathology.There is significant reduction in RC,

MTV, MRP and SL but not in SP afterCSA (RC: control 8 7±1 1, mean+SEM,

CSA 18 ±1, p<0(0l; MTV: control504±29, CSA 120+25, p<0.l; MRP:control 104±5, CSA 54+8, p<0-0l; SL:control 35+±0)25, CSA 2 4±0+2, p<0(02;Wilcoxon's signed rank sum test). Four CSApatients had an absent RSR, four expelledthe balloon at the MTV and four showedincreased pelvic floor EMG activity on

straining and rectal distension (none of thecontrols). Myenteric plexus abnormalities,including paucity and vacuolation of gang-

lion cells and nerve fibre proliferation, was

a prominent feature in all the excisedspecimens and may be responsible formany of these abnormalities.

F38Outcome of surgery in colonic angio-dysplasia

R R SALEM, J N THOMPSON, H C REES, A P

HEMINGWAY, D J ALLISON, J SPENCER, AND

C B WOOD (Departments of Surgery,Histopathology and Diagnostic Radiology,Royal Postgraduate Medical School, Lon-don) Bleeding from colonic angiodyspla-sia may be treated by endoscopic coagula-tion or by resection. Twenty nine patientswho were diagnosed as having angiodyspla-sia by selective visceral angiography or

colonoscopy underwent resection for per-sistent gastrointestinal bleeding.

Peroperative colonoscopy with trans-illumination was undertaken in 10 patientsand in eight this was valuable in visualisingthe lesions and determining their extent.Histology on 28 of the 29 resected speci-mens (with barium injections in 17) con-

firmed the diagnosis in 20 (71%). Therewere three postoperative deaths. Of theremaining 26 patients, five bled subse-quently (median follow up = 24 months)one from a definite source and two fromless clearly defined sources. In two of thesepatients additional lesions were resected atthe same operation while in seven otherpatients resection of the angiodysplasia was

not undertaken because alternative lesionsmore likely to have bled were identified atlaparotomy.

Endoscopic coagulation is being usedincreasingly to arrest haemorrhage inangiodysplasia, however resection may berequired if this is unsuccessful. In 19% ofpatients, however, bleeding recurred de-spite successful resection of an area ofproven angiodysplasia. An extensivesearch for concurrent pathology appearsindicated to identify alternative sources ofhaemorrhage.

INFLAMMATORY BOWEL DISEASE/PANCREATICO-BILIARYF39-52

F39Smoking and inflammatory bowel disease

M V TOBIN, R F A LOGAN, M J S LANGMAN, R BMCCONNELL, AND I T GILMORE Gastro-enterology Unit, Royal Liverpool Hospital,Liverpool, Departments of Therapeuticsand Community Health, Queen's MedicalCentre, University of Nottingham, Notting-ham) While the intriguing association ofulcerative colitis (UC) with non-smoking isnow established, the association of smok-ing and Crohn's disease (CD) has yet to beconfirmed and it is uncertain whether theseassociations are not secondary to develop-ing inflammatory bowel disease (IBD). Wehave, therefore, investigated smokinghabits in 280 patients (157 UC, 123 CD)and 280 general practice controls matchedfor age, sex and locality, with over 95% ofsubjects returning a postal questionnaire.Compared with the controls CD patients

were more likely to report being currentsmokers (48% relative risk (RR) 1-9p<0.25) and UC patients less likely (14%RR 0 17 p<0001). Crohn's disease pa-tients were also more likely to have eversmoked (76% RR 3 2 p<0-001) and UCpatients less likely (50% RR 0-53p<0025). Before the onset of disease CDpatients were again more likely to besmokers than their controls (63% RR 4-1p<OOOl) and UC patients less likely (17%RR 0-24 p<0-001). Compared to neversmokers ex-smokers had small but non-significant increased risks of both UC (RR1.4) and CD (RR 1.8). Smokers with IBDare almost three times as likely to have CDas UC.Both associations are strong, consistent

and antedate disease onset and use ofidentical methods minimises the possibilityof systematic biases. By confirming thecontrasting associations of smoking habit inUC and CD this study strengthens theclaim for an aetiological role of cigarettesmoking in IBD.

F40Indirect evidence against the viral hypo-thesis in Crohn's disease in the Lyons area

C ANDRt, L DESCOS, AND S DANItRE (INTRO-DUCED BY R N ALLAN) (Croupe d'Immuno-pathologie Digestive INSERM, Centre Hos-pitalier, LYON SUD, Pierre Benite,

A1 155


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


France) Many discrepancies exist withregard to the direct demonstration of thepresence of an infectious agent in Crohn'sdisease. A general agreement does existwith regard to two indirect elements, (1)the high prevalence of lymphocytotoxicantibodies suggest that a viral agent may beinvolved in the pathogenesis, (2) Crohn'sdisease ultrafiltrates produce lymphoma inathymic mice.

Crohn's disease is infrequent in Franceand particularly in the Lyons area. Thusthe same studies were repeated on patientsfrom this region. Lymphocytotoxic anti-bodies were screened by an objectivetechnique measuring the lymphocyte via-bility of a panel of eight donors by thedetermination of the residual ATP content(average killing of 20% or more of targetcells from at least four donors).Serum lymphocytotoxic antibodies were

detected in three out of 56 healthy controls(5%), in two out of 56 patients withCrohn's disease (3.5%), in 16 out of 57patients with ulcerative colitis (28%) and in11 out of 13 patients with infectious mono-nucleosis (85%).Twenty patients with Crohn's disease

and 23 patients with ulcerative colitis werestudied both when the disease was activeand quiescent. No evident relationshipcould be established between the presenceof lymphocytotoxic antibodies and clinicalactivity.Mucosal and mesenteric lymph node

homogenates from 10 patients withCrohn's disease were injected into 100nude mice. Mice were killed after seven to14 months. None developed a lymphoma.

If an infective agent is responsible forCrohn's disease it does not seem to bepresent in the Lyons area or possibly itdiffers in its characteristics from thoseagents demonstrated in other countries.

F41How can the excess mortality from Crohn'sdisease be reduced?

H A ANDREWS AND R N ALLAN (TheGastroenterology Unit, General Hospital,Steelhouse Lane, Birmingham) A recentstatistical study has shown a two-fold ex-cess mortality among patients with Crohn'sdisease. The present study analyses howthis excess mortality might be reduced.Of 792 patients under review between

1944-1984, 142 have died. The pattern ofunrelated deaths (52) was similar to thatexperienced in the general population.There were 90 Crohn's disease related

deaths. Twenty nine died postoperativelymostly (20) from sepsis. Sepsis was themain cause of death in a further 13 pa-tients. Cancer of the gastrointestinal tractwas the primary cause of death in 22patients. Other less common problemsincluded pulmonary embolism (eight),steroid related deaths (seven), amyloidosis(four), others (six) and early in the serieselectrolyte imbalance (10).We conclude that the use of peroperative

antibiotic prophylaxis and the identifica-tion and treatment of sepsis at other timesshould improve mortality rates. Changes inperoperative management should reducethe incidence of thromboembolic disease.Finally, screening for colorectal cancershould be considered in patients with long-standing extensive colitis.

F42Adhesive E coli in ulcerative colitis

D A BURKE AND A T R AXON (Gastroenterol-ogy Unit, The General Infirmary, Leeds)E coli from patients with ulcerative colitishave been studied for adhesive propertiesusing a buccal epithelial cell (BEC) adhe-sion assay. This technique has severaladvantages over HeLa cells as it allowsquantitation of adhesive properties andassessment of host factors.

Suspensions of BEC were incubated withE coli isolates in the presence of D-mannose for 30 minutes. An 'adhesionindex' was obtained for each assay bycounting the number of BEC (out of 100)with greater than 50 adherent organisms. Ecoli isolates from colitics (n=18) adheredsignificantly more than those from controls(n=12). Colitic isolates; median adhesionindex=43-5 (18-68): control isolates; me-dian=2-0 (0-15) p<0.0001. Using a singleadhesive strain there was no significantdifference between the adhesion to BEC'sobtained from colitics (n=12 median 67(45-84)) and those from controls (n=12median 69 (59-78)). Similarly for non-adhesive E coli.

This study shows different populations ofE coli in the two groups. The BEC adhe-sion assay clearly separates colitic E colifrom controls by their adhesive propertiesraising the possibility of a pathogenic rolefor adhesive organisms in ulcerative colitis.

F43Abnormal presence of lactosylceramide de-tected in Crohn's disease by thin layerchromatography

C STEVENS, V OBERHOLZER, J WALKER-SMITH,AND A PHILLIPS (Queen Elizabeth Hospitalfor Children, London) We are currentlyinvestigating the involvement of phospho-lipids and related compounds in Crohn'sdisease. Lipids were extracted from mucos-al biopsy homogenates with chloroform,methanol and water. Two dimensional thinlayer chromatography was used to separatethe phospholipid fraction. Small and largebowel samples (from adults and children)were studied from 10 patients with Crohn'sdisease, 11 patients with active ulcerativecolitis (UC), and 11 patients with normalcolonic histology. Samples from involvedsites in nine patients with Crohn's diseasestrongly displayed two abnormal com-pounds of chromatography. These werefaintly present in eight of the 11 patientswith UC and the remaining patients withCrohn's disease, but did not appear on anyother chromatography. Orcinol anddiphenylamine tests showed the com-pounds to be glycolipids. They were alsoisolated from neutrophil-enrichedperipheral leucocytes of healthy adults andco-migrated with purified lactosylceramidefrom human neutrophils (donated by Dr BA Macher, California). We therefore havea biochemical marker which appears todiscriminate between Crohn's disease andUC. The degree of mucosal neutrophilinfiltration was similar in the samples fromCrohn's disease and UC and thus cannotfully explain the strong presence of lacto-sylceramide in Crohn's disease. Additionalmucosal sources of lactosylceramide arebeing sought, as is its presence in otherinflammatory diseases.

F44Rapid bowel preparation for outpatientflexible sigmoidoscopy

S H SILVERMAN AND M R B KEIGHLEY(Department of Surgery, General Hospital,Birmingham) Fibreoptic flexible sig-moidoscopy (FFS) can easily be done at theinitial outpatient attendance provided thatefficient bowel preparation is possible. In aprospective randomised study the efficien-cy of two preparations, phosphate enema(Fletchers phosphate PHARMAX) andMicralax microenema (SK&F) were com-pared in unstarved and unpurged outpa-tients at a Rectal Clinic.

Faecal loading was assessed during rigidsigmoidoscopy and patients then receivedrandomly either a phosphate enema(n=32) or Micralax microenema (n=24).The patient having relieved himself, FFS

A1156


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


was performed and faecal loading reasses-sed. Faecal loading was graded as I=nofaeces, II=minor residue, III=gross re-sidue.

Before preparation faecal loading wasequivalent in the two groups Phosphateenema produced a highly significant reduc-tion in faecal loading (x2=9-43, p<0-001)but Micralax did not (X2=09 p=NS). Afterpreparation faecal loading was significantlylower in the phosphate group than theMicralax group (X2=11-25 p<0-001). Phos-phate enema is superior to Micralax inpreparing unpurged unstarved patients forFFS; furthermore, purging prior to FFS isunnecessary as a single phosphate enemagave adequate preparation in 29/32 (91%)of cases.

F45Azodisalicylate for the treatment of activeulcerative colitis (UC)

A IRELAND, W S SELBY, G D BARR, C H MASON,AND D P JEWELL (Gastroenterology Unit,Radcliffe Infirmary, Oxford) Sodium azo-disalicylate (ADS) consists of a dimer of5-aminosalicylic acid (5-ASA), the activeingredient of sulphasalazine. Azodisalicy-late has been assessed, double blind, inboth enema and oral form in patients withmildly active distal UC. All patients wereassessed clinically, sigmoidoscopically andhistologically. For the enema trial, 60patients were randomised to receive eitherADS enema 1 g nightly or a dummy enemafor two weeks. Nineteen of 29 on ADSimproved, nine remained unchanged andone deteriorated. Twelve of 28 on dummyimproved, 14 remained unchanged and twodeteriorated. These differences were notsignificant. Sigmoidoscopic responsesshowed a similar pattern. Response wasunrelated to previous sulphasalazine ther-apy or length of history. For the oral trial,40 patients were randomised to receiveeither ADS capsules 2 g/day or dummycapsules for two weeks. Sulphasalazine wasstopped on entry. Thirteen of 20 on ADSimproved, six remained unchanged andone deteriorated. On dummy, eight of 20improved, three remained unchanged, andnine deteriorated. These differences werestatistically significant (p<0-02). Sigmoido-scopic responses were also better in theADS group (p<005). Responses wereunrelated to length of history or previoussulphasalazine. The drug was well toler-ated in both trials, and adverse reactionswere minimal. Azodisalicylate in oral formis therefore effective in mildly active distal

UC, and is a useful alternative to sulphasa-lazine.

F46High performance liquid chromatography(HPLC) of plasma PABA - a sensitive assayfor pancreatic exocrine dysfunction

I M CHESNER, J BERG, N LAWSON, AND P

ASQUITH (The Metabolic Unit and Depart-ment of Clinical Chemistry, East Birming-ham Hospital, Birmingham) The NBT-PABA test has a recognised role in theinvestigation of pancreatic dysfunction.Conventionally PABA excretion is mea-sured in the urine by chemical methods.The assay is subject to interference by awide range of drugs - for example, aspirinand paracetamol, and the method dependsupon normal hepatic and renal function,plus a complete six hour urine collection.Using an HPLC assay for the detection ofPABA in plasma we have studied ninepatients with proven pancreatic steator-rhoea, seven patients with chronic pancrea-tic disease without steatorrhoea, and 11healthy controls. The test was carried outin the standard way using lg NBT-PABA,25 g casein and 5 ml C14 PABA takenorally. Serial blood specimens were takenat one, two, three and four hours. Separa-tion between the three groups was evidentat two, three and four hours and wasmaximal at three hours. The three hourmean (SD) for healthy controls was 47-9(±10.4) that for the pancreatic diseasegroup without steatorrhoea 17-6 (±5-4)(p<0-01). In the group with proven pan-creatic steatorrhoea eight of the nine hadan undetectable three hour level, the othera level of 9.3 ,umol/l (p<0-001).We conclude that an HPLC assay of

PABA in plasma is a rapid and accuratemethod of assessing pancreatic exocrinefunction. A single three hour plasma levelwill discriminate between normals, pan-creatic damage and pancreatic malabsorp-tion.

F47Use of fresh frozen plasma in the treatmentof acute pancreatitis

T LEESE, K P WEST, AND A W HALL (TheDepartment of Surgery and Pathology,University of Leicester, Leicester) Admi-nistration of fresh frozen plasma (FFP)may reduce mortality in acute pancreatitis,possibly by replenishing the plasma anti-proteinase system. An animal study has

been undertaken to assess further the valueof FFP in acute pancreatitis. Acutehaemorrhagic pancreatitis was induced inmale AS rats 250-300 g by transduodenalinfusion of 0-2% enterokinase and 3.5%sodium taurocholate into the pancreaticduct system. Animals underwent right in-ternal jugular vein cannulation and, onrecovery from anaesthetic, 30 animals wererandomised to each of three intravenousfluid regimes. Group A received 8 ml ofnormal saline and 16 ml of dextrose salineper 24 hour. Group B received 8 ml of'Haemacel' and 16 ml of dextrose salineand group C received 8 ml of FFP and 16ml of dextrose saline. Infusions were re-peated each 24 hour for 72 hours.There were nine survivors (30%) at 72

hours in group A, 13 (43%) in group B and22 (73%) in group C. Using a X2 test theimproved survival in the FFP groupreached significance when compared withthe crystalloid controls (p<0-001) and thecolloid controls (p<0-05).A randomised prospective trial is now

underway comparing FFP (2 units per dayfor three days) to plasma protein fraction(400 ml per day for three days) as part ofthe intravenous fluid administered to pa-tients admitted with acute pancreatitis.

F48Ultrasound assessment of bile duct obstruc-tion: level, cause, and tumour resectability

R N GIBSON, E YEUNG, H A BRADPIECE, J NTHOMPSON, D H CARR, A P HEMINGWAY, I SBENJAMIN, D J ALLISON, AND L H BLUMGART(Departments of Radiology and Surgery,Royal Postgraduate Medical School, Ham-mersmith Hospital, London) In a prospec-tive study of 65 patients with bile ductobstruction the role of ultrasound wasevaluated in relation to other radiologicalmodalities. Computed tomography (CT)was carried out in 51 patients, directcholangiography in 57, and angiography in32. Ultrasound determined the correctlevel of obstruction in 61 patients and inpredicting cause was correct in 56, incor-rect in three and indeterminate in five. Thecorresponding results for CT were 45, 32, 3and 16.

Thirty seven of the 58 patients withmalignant obstruction had hilar tumours.Of this group irresectability was estab-lished radiologically in 27 and operativelyin a further two. Ultrasound alone pre-dicted irresectability in 21 and in two ofthese it was the only modality to do so.One patient judged by ultrasound to be

A1157


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/


probably irresectable was in fact resect-able. Computerised tomography was ableto predict irresectability in only 11 out of 25cases. In 21 patients with non-hilar malig-nant obstruction ultrasound assessment ofresectability disagreed with the combinedprediction of the other modalities in onlyfive cases and in three of these the ultra-sound prediction was correct on operativefindings.We conclude that good quality ultra-

sound is the most useful and frequently theonly necessary imaging modality forpatients with bile duct obstruction.

F49Acid stable lipase in the treatment ofpancreatic steatorrhoea

S M GRIFFIN, D ALDERSON, AND J R FARNDON(INTRODUCED BY M B CLAGUE) (Departmentof Surgery, New Medical School, Framling-ton Place, Newcastle upon Tyne) Pancrea-tic exocrine insufficiency in cystic fibrosisand chronic pancreatitis results in steator-rhoea with concomitant poor nutrition andsocial embarrassment. Long term conven-tional treatment has been unsatisfactorybecause up to 90% of the lipase content ofsuch therapy is inactivated by gastric acid.This results in a need for large volumes ofmedication to provide adequate sup-plementation. The aim of this study was toinvestigate the efficacy of a new acidresistant agent - fungal lipase in the treat-ment of pancreatic exocrine deficiencysteatorrhoea.

Eleven dogs with pancreatic insuffici-ency were studied while taking a fixeddietary intake of fat. All animals wererandomly allocated to fungal lipase, pan-creatin and placebo for two week coursesof treatment regimens. Ten grams (60 000U lipase) of pancreatin was compared with400 mg (24 000 U lipase) of fungal lipasegiven with each meal against a placebo.Mean faecal fat concentrations over a threeday period were performed. Data wereanalysed using the Wilcoxon's rank sum testfor paired data. There was no significantdifference between pancreatin and lipasetreated animals. Both groups had signifi-cantly reduced steatorrhoea when com-pared with placebo (p<001). This studyshows that a markedly reduced volume oftreatment in the form of fungal lipasecontrols steatorrhoea as effectively as pan-creatin, and represents a potentially valu-able new therapy for pancreatic exocrineinsufficiency.

F50Is cholecystectomy needed after endoscopicremoval of common bile duct stones in theelderly?

D F MARTIN AND D E F TWEEDLE (UniversityHospital of South Manchester, Manchester)Between July 1981 and December 1983, 35patients aged between 50 and 94 years(mean 79) underwent successful endosco-pic clearance of common bile duct (CBD)stones. They presented with a combinationof obstructive jaundice (25 patients), acutecholangitis (10), acute pancreatitis (two)and biliary colic (11). Ten patients had ahistory of myocardial infarction or strokeand three others had been treated forbronchial carcinoma. Twenty five patientsremain alive and symptom free 19 to 46months (mean 28 months) after treatment.Seven patients have died, none as a resultof gall bladder disease. One patient, aged50, underwent elective cholecystectomybut had been symptom free for fourmonths since endoscopy. Two patientshave required cholecystectomy for acutecholecystitis five and six months aftercommon duct clearance. In both the gallbladder had filled at ERCP. Exploration ofthe CBD was not necessary. These resultssuggest that endoscopic treatment alonefor CBD stones is indicated in elderly highrisk patients with gall bladders, subsequentcholecystectomy proving necessary in lessthan 10%. Lack of filling of the gallbladder at ERCP does not appear to be auseful indicator of the future need forcholecystectomy.

F51Does drainage increase the risk of sub-hepatic collection following cholecystec-tomy? - a prospective study

J R T MONSON, H IRVING, A W TANNER, JMACFIE, AND T G BRENNAN (St James'sUniversity Hospital, Leeds, West York-shire) Intraperitoneal drainage aftercholecystectomy is traditional but unsup-ported by scientific data. In an uncontrol-led study in 1983, Elboim demonstratedwith ultrasound a significant incidence ofsubhepatic collections following cholecy-stectomy, with the figure much higher indrained patients, suggesting a causitivelink. We have evaluated the influence ofintraperitoneal.drains on the incidence ofthese collections in a randomised prospec-tive trial. Ninety three patients undergoingcholecystectomy were randomised at the

end of the operation into a drainage, (4"suction drain for 48 hours (n=44)), ornon-drainage group (n=49). Patients withperforated gall bladders or undergoingexploration of the common bile duct wereexcluded. All patients had abdominalultrasound carried out 72 hours postopera-tively. The number of patients with acutecholecystitis was similar in the drainageand non-drainage group (11% vs 12%).

Ultrasound detected six collections inthe drainage group and one collection inthe non-drainage group (p<0O05 X2). Theaverage volume of the collections was 300ml. None was clinically significant. Theseresults suggest that drains may actuallyincrease the risk of developing a subhepaticcollection after cholecystectomy. Furtherprospective studies are required.

F52Cholecystokinin provocation test insuspected acalculous biliary pain - threeand a half years of clinical use

T W J LENNARD, J R FARNDON, AND R M RTAYLOR (INTRODUCED BY I D A JOHNSTON)(Department ofSurgery, University ofNew-castle, Newcastle upon Tyne) A doubleblind placebo controlled infusion ofcholecystokinin (CCK) has been used toattempt to reproduce the presenting symp-tom in 123 patients with unexplained rightupper quadrant abdominal pain. Sixty twopatients (53 women, nine men, mean age44 years) developed pain with CCK andnot with saline infusions (CCK positive).Fifty nine patients (46 women and 13 men,mean age 43 years) developed no painduring either infusion (CCK negative) andtwo patients developed pain during bothinfusions. Fifty four patients who wereCCK positive ve have undergone cholecy-stectomy to date, and of these 51 (94%)have been relieved of their pain and havehad no response to postoperative CCK/placebo infusions (mean follow-up 11-9months, range two months to 3-5 years).Gall bladder histology was abnormal in 48cases (chronic cholecystitis and/orcholesterolosis) and a small gall stone wasfound in two cases. Cholecystokinin nega-tive patients have been followed up andnine have spontaneously improved. Fifteenremain undiagnosed and in 35 an alteran-tive cause for their pain has been found.The cholecystokinin provocation test willidentify those patients who are likely tohave a good resonse to cholecystectomy.

A1158


http://gut.bmj.com

/G


ctober 1985. Dow

nloaded from

http://gut.bmj.com/

Date post:	24-Mar-2023
Category:	Documents
Upload:	khangminh22
View:	0 times
Download:	0 times

The British Society of Gastroenterology - Gut

Documents