British Society of Gastroenterology · Gut: first published as 10.1136/gut.23.10.A880 on 1 October...

Gut, 1982, 23: A880-926

The British Society of GastroenterologyThe Autumn Meeting of the Society was held from 22 to 24 September 1982 at the University of Leeds underthe presidency of Professor C G Clark. One hundred and fifty-three communications (including 29 posters)were selected for presentation in the scientific programme on September 23 and 24; the abstracts are printedbelow.

POSTERSWP1-WP29

WP1Peptide-containing nervous system of thegut in three forms of megacolon

A E BISHOP, G P MCGREGOR, B D LAKE, D MPRESTON, S R BLOOM, AND J MPOLAK (Departments of Histochemistryand Medicine, Hammersmith Hospital,Department of Pathology, The Hospital forSick Children, Department of Medicine, StMark's Hosptial, London) It has beenestablished that the peptide-containingnerves of the gut show distinct abnor-malities in certain diseases. In the presentstudy, the techniques of immunocyto-chemistry and radioimmunoassay havebeen used to investigate these nerves insurgical specimens from patients with threeforms of megacolon: congenital (n=8),acquired (n=10), and idiopathic (n=6).A marked reduction in peptide-

containing nerves was observed in bothcongenital (Hirschsprung's disease) andacquired (Chagas' disease) megacolon. InHirschsprung's disease the loss of peptide-containing nerves was largely in proportionto the length of the aganglionic segment. InChagas' disease, however, where theneuronal degeneration in the gut is diffuse,the extent of the reduction in peptide-containing nerves was variable. Radio-immunoassay demonstrated a decrease inthe content of neuropeptides in bothdiseases. For example, the content of VIPwas reduced from 142±18 pmol/g wetweight of tissue in normal neonatal bowelto 57±8 pmol/g in aganglionic bowel. Incontrast, no abnormality of peptide-containing nerves could be detected inidiopathic megacolon.These results provide new means for the

investigation of the pathogenesis ofmegacolon and indicate that the underlyingabnormality in idiopathic megacolon doesnot involve the peptide-containingcomponent of the enteric nervous system.

WP2Localisation of distinct sub-populations ofgastrin-containing secretory granules in themammalian gastrointestinal tract using anew double immunogold staining method

I M VARNDELL, F J TAPIA, J DE MEY, NYANAIHARA, S R BLOOM, AND J MPOLAK (Departments of Histochemistryand Medicine, Royal Postgraduate MedicalSchool, Hammersmith Hospital, London,Department of Life Sciences, JanssenPharmaceutica, Belgium, Department ofBio-Organic Chemistry, Shizuoka Schoolof Pharmacy, Shizuoka, Japan) Two bio-active molecular forms of gastrin areknown to occur in endocrine cells in themammalian gut. Gastrin 17 (G17), thecarboxy terminal heptadecapeptide of biggastrin (G34), is the predominantmolecular form in the human gastricantrum and some authors have speculatedthat G17 is contained within the electron-lucent vesicles which are characteristic ofantral G cells. In addition to the electron-lucent vesicles there is a subpopulation ofsmaller granules with an electron-densecore and a closely apposed limitingmembrane. We have attempted to localisegastrin immunoreactivity within these twodistinct subpopulations in the gastro-intestinal tract of cat and man using a newdouble immunogold staining method. The

'double immunogold staining procedure isbased on the conjugation of colloidal gold(20 or 40 nm) to immunoglobulins whichrecognise the primary antisera (rabbit (R)anti-N-terminal gastrin 34 and guinea-pig(GP) anti-N-terminal gastrin 17). Theprimary antisera were visualised using goatanti-R40 nm and goat anti-GP20 nm. UsingN-terminal G17 antisera we have been ableto demonstrate specific immunoreactivityin the electron-lucent vesicles of mam-malian antral G cells and also ingastrinomas which are known to secretepredominantly G17. In addition, usingN-terminally directed antisera to G34 wehave visualised immunoreactivity in theelectron-dense secretory granules of

A880

intestinal G cells in man (160 nm) and cat(280 nm). Both G17 and G34 immuno-reactivities have been demonstratedsimultaneously in the electron-densegranules which coexist with the electron-lucent vesicles in antral G cells.

WP3Pharmacological properties of two newsulphasalazine analogues

J E LENNARD-JONES, J H BARON, ROSALIND PCHAN, D J POPE, A P GILBERT, AND P JSACRA (St Mark's Hospital, London, StCharles Hospital, London, BiorexLaboratories Ltd., London) Circulatingsulphapyridine is the main cause ofunwanted side-effects with sulphasalazine,whereas the 5-aminosalicylic acid (5ASA)component may be the active moiety. Twoanalogues have been synthesised in whichsulphapyridine has been replaced by either4-aminobenzoylglycine (ABG) or 4-aminobenzoyl-f-alanine (ABA) to yieldhipsalazine and balsalazine respectively.Both new compounds are crystalline, non-hygroscopic stable dihydrates, highlysoluble in water (up to 22% W/V). Acutetoxicity of hipsalazine and balsalazine wasnot demonstrable in rats and mice withsingle doses up to 2-4 g/kg body weight.Chronic toxicity of balsalazine was notdemonstrable in dose levels up to 1 g/kg inrats and 0-5 g/kg in ferrets over 90 days.Metabolic studies in rats showed thatrecoveries of 5-ASA in urine and faeceswere similar for sulphasalazine, hip-salazine, and balsalazine. In normal humanvolunteers only trace amounts ofunchanged hipsalazine and balsalazinewere excreted in urine and faeces after asingle oral dose of 2 g. Two-thirds of the5-ASA in sulphasalazine and hipsalazine,and half of that in balsalazine appeared inthe faeces. Total recoveries of 5-ASA infaeces and urine were over 80% with eachdrug. There was a marked difference in theexcretion pattern of the carrier moleculessulphapyridine (faeces 6%, urine 85%) and

on June 4, 2020 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.23.10.A880 on 1 O

ctober 1982. Dow

nloaded from

http://gut.bmj.com/

The British Society of Gastroenterology

ABG (faeces 8%, urine 45%) comparedwith ABA (faeces 74%, urine 19%).Serum concentrations of hipsalazine,balsalazine, ABA and 5-ASA were belowthe limits of detection (-2 ,ug/ml), meanlevels of ABG did not exceed 3 ,ug/mlcompared with free sulphapyridine whichreached a mean level of 24 ,ug/ml. Thesetwo analogues appear to be suitable fortherapeutic trial.

WP4Imbalance of prostacyclin and throm-boxane synthesis in Crohn's disease

C J HAWKEY, F KARMELI, AND D

RACHMILEWITZ (Nuffield Department ofClinical Medicine, John Radcliffe Hospital,Oxford, and Hadassah University Hospital,Jerusalem, Israel) Rectal mucosa frompatients with ulcerative colitis in relapse(but not remission) synthesises excessquantities of prostanoids in organ culture,but data are lacking for patients withCrohn's disease.

Thirty rectal biopsy specimens fromBritish patients (23 Crohn's disease, sevennormal) were placed in organ culture for23 hours. The accumulation of prosta-glandin (PG)E2, 6 keto PGFI,, and throm-boxane (TX)B2 was measured byradioimmunoassay.

Synthesis of TXB2 by uninflamedmucosa from patients with Crohn's disease(0.90 ng/mg, 0-11-3-52 ng/mg, median andrange, n=11) was greater than by normalmucosa (0.18 ng/mg, 0-15-0-74 ng/mg,n=7, p<005) and was similar to that seenwith inflamed mucosa (0.66 ng/mg,0-21-3-76 ng/mg). The groups did notdiffer in the synthesis of 6 keto PGF1, butthe ratio of 6 keto PGF1, to TXB2 wasreduced in Crohn's disease (0.60,0.08-2.33) compared with normal (1.48,0-64-2-78, p<0-02). Inflamed mucosasynthesised more PGE2 (4.67 ng/mg,1-48-20-44 ng/mg, n=12) than uninflamedmucosa (2.22 ng/mg, 0-88-12-75 ng/mg,n=11, p<002).We conclude that there is an imbalance

of prostacyclin and TXB2 synthesis inCrohn's disease even in the absence ofinflammation. The consequences mayinclude vasoconstriction, platelet aggrega-tion, reduced suppressor cell activity, andpossibly diminished 'cytoprotection'.

WP5Prediction and treatment of severe acutepancreatitis by peritoneal lavage

A P CORFIELD, M J COOPER, R C N WILLIAMSON,AND A V POLLOCK (Department of Surgery,Bristol Royal Infirmary, Bristol, andScarborough Hospital, Scarborough) Onehundred and fifty-one patients with raisedserum amylase (>1200 IU/l Phadebas)were entered into a prospective controlledtrial to evaluate reports that peritoneallavage may predict, and improve survivalin severe acute pancreatitis. There were 78males and 73 females with a mean age of60-8 years, the majority (64%) havinggallstone-related pancreatitis. Two predic-tive criteria of severity on admission wereassessed; hypoxaemia (PaO2 <7-98 kPa)and toxic peritoneal broth. One or bothcriteria were fulfilled by 39 patients ofwhom 12 died (31%) and eight developedother major complictions (21%). Two ofthe 112 with predicted mild disease died(2%) and eight developed major com-plications (8%). The overall accuracy ofthese predictive criteria is 81%. Patientswith predicted severe disease (n=39) wererandomised to receive therapeutic lavagewith hourly cycles of 21 dialysate for 72hours (n= 16) or to act as controls (n=23).The following results show figures forcontrol first and lavage second: benigncourse, 11, 8; major complications, 4, 4;death, 8, 4.An accurate guide to the likely severity

of an attack of acute pancreatitis is pro-vided by the predictive criteria, but thevalue of therapeutic peritoneal lavageremains unproven.

WP6Adaptation of the shortened gut: increasednumbers of sialomucin-containing gobletcells

I 0 OLUBUYIDE, R C N WILLIAMSON, J B

BRISTOL, AND A E READ (UniversityDepartments of Medicine and Surgery,Royal Infirmary, Bristol) Both clinicaland experimental studies have shownhyperplasia of the small bowel remaining incontinuity after jejunoileal bypass, and thecolon may contribute to the compensatoryresponse. Numerical changes in thedifferent types of epithelial cell have notbeen identified, however. Adaptation ofthe goblet cell population was thereforestudied in male Sprague-Dawley rats(n=40) weighing 75-0±4-6 g (SEM) andrandomised to receive 85% end-to-side

jejunoileal bypass or sham bypass (jejunaltransection, ileotomy, and resuture). At 36weeks jejunoileal bypass increased thelength and wet weight of the duodenum by38-55% and of the large intestine by13-25% (p 0.05-0.001). In the duodenumand functioning jejunoileum, villi were43-63% taller and crypts 41-69% deeper(p 0.02-0.001). Jejunoileal bypass alsoincreased crypt depth in the distal colon by24% (p<0-001). Goblet cells containingspecific acid mucins were differentiatedhistochemically, using high iron diamine-alcian blue. Irrespective of operation,sulphomucins predominated throughoutthe gut. Jejunoileal bypass increased thenumber of sulphomucin-containing gobletcells only in ileal villi (by 16%) anddistal-colon crypts (by 27%). By contrast,sialomucin cells were two to three timescommoner both in the distal colon(p<002) and in the villi and crypts of thefunctioning jejunoileum (p 0.04-0-002).Adaptation to jejunoileal bypass involvesall the residual functioning intestine andincludes goblet-cell hyperplasia.Sialomucin-containing cells are increaseddisproportionately.

WP7Effect of sodium and acetyl-salicylates onbile flow in man

M J COOPER AND R C N WILLIAMSON(University Department of Surgery, BristolRoyal Infirmary, Bristol) We havepreviously shown that sodium salicylate is apotent choleretic in the Rhesus monkey,but were unable to confirm the alterationof lithogenicity observed in dogs afteradministration of lysine acetyl-salicylate.To investigate the possible value ofsalicylates in the treatment of humanbiliary disease, five patients with indwellingT-tubes were studied. T-tubes wereclamped for 48 hours to restore the entero-hepatic circulation before study. Bile wasthen collected over a one-hour baseline,and 1-8 g sodium salicylate and 1-8 g acetylsalicylate were administered p.o. onconsecutive days; collection was continuedfor a further two hours. Baseline bile flowwas 7-07±0-16 ml/kg/24 h (mean±SEM)and increased to 11-46±0-43 ml/kg/24 h(p<0-01) with sodium salicylate and to9-67±0-44 ml/kg/24 h (p<0.001) withacetyl-salicylate. Baseline bile acidsecretion (0.26±0.02 mmol/kg/24 h) wasunchanged after either drug, confirmingthat the choleresis is bile acid-independent.Baseline secretion of phospholipid

A881


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/


(0.07±0.01 mmol/kg/24 h) and cholesterol(0.03±0.003 mmol/kg/24 h) was alsounchanged after both drugs, indicating nochange in lithogenicity; the lithogenic ratioremained at 11-81±0-4. These resultsdemonstrate that both sodium and acetylsalicylates increase human bile flow, but ata lesser order of magnitude than thecholeresis seen in dogs or monkeys.Neither agent affects the lithogenicity ofprimate bile.

WP8Determinants of cholesterol saturation offasting gall bladder bile in health

R P JAZRAWI, R M KUPFER, C BRIDGES, A

JOSEPH, AND T C NORTHFIELD (Depart-ments of Medicine and Nuclear Medicine,St George's Hospital Medical School,London) Since saturation index is higherfor fasting than for postprandial hepaticbile, percentage gall bladder filling mayinfluence saturation index of fasting gallbladder bile. We have therefore developeda simple, rapid scintiscanning technique forassessing this. The conventional measure-ment of saturation index does not indicatewhether a rise is due to excess ofcholesterol, or deficiency of bile acid andphospholipid within the gall bladder. Wehave therefore measured the mass of allthree biliary lipids by combining our scinti-scanning technique with direct analysis ofgall bladder bile. We also measured bileacid pool size by isotope dilution. Westudied 12 non-obese healthy malevolunteers. After intravenous 99mTc-HIDA, gall bladder and gut activity weremeasured using two isosensitive scanningheads (anterior and posterior). Gallbladder bile was obtained by nasoduodenalintubation. The concentration of biliarylipids was determined, and Tc-HIDAactivity was measured in vitro in exactly thesame way as for the in vivo studies. Bileacid pool size was determined by isotopedilution. Percentage gall bladder filling(mean ± SEM) was 54±8%, and wassignificantly correlated with saturationindex (r=0.63, p<005). Saturation indexwas also significantly correlated withcholesterol mass in the gall bladder(r=0.64, p<005), but not with bile acid(r=0.40) or phospholipid mass (r=0.03),or with bile acid pool size (r=0.36). Weconclude that saturation index of fastinggall bladder bile in health is influenced bythe degree of gall bladder filling, and bycholesterol mass in the gall bladder.

WP9Natural history of chronic active hepatitis

F CARUBBI, M PONZ DE LEON, P DI DONATO, P

LORIA, I IORI, AND N CARULLI (IstitutoClinica Medica II, Universitd di Modena,Italy) A recent long-term trial suggestedthat the natural history of untreatedchronic active hepatitis seems to proceedfrom an active hepatitis to inactivecirrhosis, with a 27% survival at 10 years,and that prednisolone improves survival.This study, however, dealt with severely illpatients probably not representative of thevast majority of patients with chronicactive hepatitis, as pointed out by morerecent studies. A detailed review wastherefore undertaken of the patients withchronic active hepatitis without cirrhosisadmitted in our Institute between 1971 and74 and subsequently followed up to now.Twenty patients were included in the study(M18, F2) with the following criteria: (1)none was treated with corticosteroids orimmunosuppressive drugs; (2) all had aliver biopsy at the first admission.and atleast a second biopsy during the follow-up;(3) minimum follow-up was 7-5 years(mean 9-6, range 7.5-11 years).

Results showed that HBsAg was positivein 12 (60%) and remained positivethroughout the study except for one patientwho seroconverted to the negative state.At the time of diagnosis 14 patients werevirtually asymptomatic; two patients hadascites that disappeared with low salt dietand did not reappear, in contrast onepatient developed ascites, with other signsof cirrhosis, after six years. Jaundice waspresent in four patients at the diagnosis butdisappeared rapidly during the hospi-talisation without reappearing. Oeso-phageal varices were not present in anypatient at the diagnosis and subsequentlyappeared in three. At diagnosis thebiochemical profile showed a modestimpairment of the liver function; this, onaverage, tended to improve slightly duringthe follow-up. Histology was consideredimproved in two patients (whose diagnosisbecame chronic persistent hepatitis),virtually unchanged in 13 (65%), andworsened in five patients (25%), whoshowed histological features of cirrhosis.None of the patients died and most of themcontinued their usual work.We conclude that this long-term follow-

up showed that chronic active hepatitiswithout cirrhosis is a less severe diseasethan previously reported; chronic activehepatitis shows a protracted clinical coursewithout impairment of the general state in

most of the patients. Cirrhosis, however,eventually develops in a definite pro-portion of these patients. In our opinion,the real benefit of drugs on the naturalevolution of the disease has yet to beproved.

WP1OTerpene treatment for gallstones: five yearsof experience with rowachol 1977-2

W R ELLIS, K W SOMERVILLE, D H ROSE, A T R

AXON, AND G D BELL (University Depart-ment of Therapeutics and X-ray Depart-ment, Nottingham City Hospital andGastroenterology Unit, General Infirmary,Leeds) We treated 48 cases of radiolucentgall bladder stones with rowachol alone(1-6 capsules daily): nine responded,including 3/42 (7%) complete dissolution(excluding six planned withdrawals), or 8.3per 100 patient-years. Rowachol did notalter duodenal bile cholesterol saturation(1.36±0.21 SD before, 1-31±0-28 after,n=9), Rowachol (2-3 capsules daily) wascombined with suboptimal chenic acid(CDC) dosage - 375 mg/day (A) or 7-5-10mg/kg/day (B). On dosage A, completedissolutions were 6/31 (19%) at 12 monthsand 8/30 (27%) at 24 months. Threepatients' biliary symptoms requiredsurgical intervention and one haddiarrhoea, nine changing treatment after12 months. On dosage B, five patientsshowed progressive dissolution duringtreatment; stone disappearance occurred in9/30 (30%) at 12 months and 12/25 (48%)at 24 months. No patient had severesymptoms, one had possible side-effects,and nine elected to discontinue treatmentafter six or 12 months.

Using Rowachol alone, 6/18 (33%) ofductal stones dissolved in 12 months and7/17 (41%) in 24 months; no radio-opaquestones disappeared but 6/21 (29%)responded partially within two to threeyears.We concluded that Rowachol alone was

active, comfortably exceeding placebodissolution rates (about 0-4 per 100 patientyears) for gall bladder stones andcomparing favourably with bile acidtherapy for duct stones. Tolerance ofRowachol and Rowachol-CDC combina-tions was excellent; continued biliary painand adverse effects were rare. Efficacy ofcombination treatments using suboptimalCDC doses considerably exceeds that ofeither drug alone and compared well withfull-dose bile acid treatment. Sinceduodenal bile remains saturated, alter-

A882


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/


native mechanisms, as recently suggested,must be proposed for Rowachol's adjuvantcholelitholytic properties.

WP11Demonstration of immunoreactive sites ofinsulin biosynthesis using immunogoldprocedures

F J TAPIA, I M VARNDELL, J DE MEY, L HEDING,N YANAIHARA, S R BLOOM, AND J MPOLAK (Departments of Histochemistryand Medicine, RPMS, HammersmithHospital, London, Department of LifeSciences, Janssen Pharmaceutica, Belgium,Novo Research Insitute, Denmark, andShizuoka College of Pharmacy, Japan)The enzymatic cleavage of proinsulin in theGolgi region and secretory granules of theB-cell yields insulin and C-peptide. Thesecretory granules of the B-cell are of threetypes, spherical electron-opaque (pale),spherical electron-dense, and crystal-linecore granules. We have identifiedbiosynthetic sites of insulin in normalhuman pancreas (n=4) and insulinomas(n=15) using antisera to porcine insulinand to synthetic human C-peptide (knownto cross-react extensively with proinsulin).Immunoreactive sites were detected at theelectron microscopical level using singleand double immunogold procedures. C-peptide (proinsulin) immunoreactivity wasdemonstrated in a subpopulation ofspherical pale and electron-dense coregranules. Very little immunostaining wasobserved in the crystalline-cored granules.In contrast, insulin was immunostained inall the granules. Some atypical (non-diagnostic) tumour cells were immuno-reactive to insulin and C-peptide (proin-sulin), other atypical cells were found tocontain pancreatic polypeptide (n=4).Insulin and C-peptide (proinsulin) werealso demonstrated around the Golgiregion. These results show for the first timethe ultrastructural localistion of proinsulinto a subpopulation of the B-cell granules.

WP12Differential distribution of the peptidergicinnervation in the human intestinal mucosa

G-L FERRI, S R BLOOM, AND J MPOLAK (Departments of Histochemistryand Medicine, Royal Postgraduate MedicalSchool, Hammersmith Hospital,London) Several peptides are present inthe innervation of the mammalian gut.Their differential distribution in the

various compartments of the mucosa,however, has been little investigated inman. By immunocytochemistry we studiedthe three-dimensional pattern of nervefibres containing VIP, substance P, andmet-enkephalin in the ileal and colonicmucosa (n=20 samples, fixed in 0.4%p-benzoquinone). Mucosal strips (1 x3mm) were dehydrated, cleared in xylene,rehydrated and immunostained whole byimmunofluorescence, using prolongedincubations (24-36 h) and washings (6-8h). After immunostaining, thin mucosalslices (1 villus or 2-3 crypts wide) and themuscularis mucosae layer were micro-dissected and mounted on slides. An exten-sive network of nerve fibres containing VIPwas consistently demonstrated across thewhole thickness of the mucosa, but wasdistinctly more abundant underneath theluminal epithelium at all levels investi-gated. Substance P-immunoreactive fibresformed a sparser network of fibres tendingto run vertically in the villus core and inbetween the epithelial crypts. In themusculatris mucosae, in addition to VIP-and substance P- containing fibres,scattered met-enkephalin-immunoreactivenerve bundles were also seen, reachingonly the very bottom of the epithelialcrypts. The differential distribution of VIP-immunoreactive fibres is in keeping with apossible role for this peptide in the controlof intestinal secretion, while all threepeptides investigated are likely to influencethe motility of the muscularis mucosae.

WP13Dual localisation of PHI in the gut andbrain

N D CHRISTOFIDES, Y YIANGOU, G P

MCGREGOR, M A GHATEI, K TATEMOTO, J M

POLAK, AND S R BLOOM (Department ofMedicine, Royal Postgraduate MedicalSchool, London) PHI, the most recentlydiscovered gut peptide, has recently beenlocalised immunochemically in the humancolon. The aim of this study was todocument its precise distribution in thegastrointestinal tracts of three species andalso to investigate its localisation in thebrain. The full gastrointestinal tracts of fiverats, six guinea-pigs, and five cats wereexamined. These were dissected intofundus, antrum, duodenum, jejunum, andcolon. In addition, the brains of six ratswere dissected into the cortex,hippocampus, mid-brain, brain stem, andcerebellum. Tissues were extracted in 05Macetic acid and assayed using a previously

described PHI radioimmunoassay. In allthree species the lowest PHI concentrationoccurred in the fundus (<20 pmol/g), levelsincreasing caudally reaching a peak in thecolon (rat 140+13 pmol/g and guinea-pigand cat 360±40 pmol/g). In the rat brainPHI was found in the cortex (35+4) andhippocampus (35±8 pmol/g) with lowerconcentrations in the mid-brain (10 pmol/g) and brain stem (1±0.6 pmol/g). PHI wasundetectable (<0.3 pmol/g) in thecerebellum.Thus PHI is the latest member of the

group of peptides with a dual (gut andbrain) localisation, and which are thoughtto have a role as neurotransmitters orneuromodulators.

WP14Outpatient preparation for colonoscopy bysodium picosulphate

D G KARAMANOLIS, P B BOULOS, AND P RSALMON (Departments of Surgery andGastroenterology, School of Medicine,University College London, UniversityStreet, London) Preparation of patientsfor colonoscopy involves colonic washoutsand requires hospital admission if dayfacilities are not available. Bowelpreparation with oral sodium picosulphatewithout colonic washout has been success-fully used in barium studies of the largebowel but its use in colonoscopy has notbeen examined. In this study 46 con-secutive patients undergoing colonoscopywere randomised to either a test groupprepared with sodium picosulphate andlow residue diet (without washout enema)the day before colonoscopy or to a controlgroup prepared by a standard regimeconsisting of three day low residue diet,purgatives, and washout enemas. Thepatient's acceptability was measured by ascoring scale and the endoscopist'sassessment of the bowel cleanliness wasdescribed as good, adequate, or moderate.There were 26 patients in the test group

and 20 in the control group. Theacceptability score (0-11 median 4.25) inthe test group was significantly better(p<0-01) than in the controls (2-13 median6.0). In the test group, the bowelcleanliness was good in 14, adequate infour, moderate in eight; in the controls 12,six, two respectively and were not differentstatistically by x2 test. The haematocrit,urea, and elctrolytes before and afterbowel preparation in both groups didnot show any change demonstrablestatistically. These results show the efficacy

A883


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/


of sodium picosulphate in bowelpreparation for colonoscopy and can berecommended for outpatient use.

WP15Collagenase inhibition in colonic mucosa byproteinase inhibitors

A P JAYARAJ, P R HAWLEY, P B BOULOS, AND C

G CLARK (Department of Surgery, Schoolof Medicine, University College London,University Street, London) It has beenshown that the colonic mucosa producescollagenase after injury which may beresponsible for the breakdown of colonicanastamosis and this is demonstrable bylysis of collagen substrate. This studyinvestigates the possibility of inhibitingcollagenase by proteinase inhibitors.

Forty-two male New Zealand whiterabbits were randomly allocated into seven

equal groups; one group acted as a controland six groups were treated with a singledose of soy or lima bean trypsin inhibitor -

SIGMA (1 mg in 2 ml water) administeredeither intramuscularly, or by instillationinto the stomach or into the rectum. Afterthree days, the animals were killed and 2mm colonic mucosa was removed by tissuepunch from each animal. These were

explanted on collagen gel plates andincubated for 72 hours. The lysed area on

the collagen gel was drawn on tracingpaper, cut out and weighed to express lysisin milligrams. The mean areas of lysis inanimals treated with soy bean were

20-0±3-2 when given intramuscularly,20*0±6.0 when instilled into the stomach,0.1±0.38 when instilled into the rectumand the respective areas in animals treatedwith lima bean were 4-0±1-8, 15-0±3-2with no lysis in the group treated rectally.The areas of lysis in the treated animalswere significantly less (p<0.0001) than inthe control group (160-0±12-8).These results indicate that collagenolysis

is inhibited by proteinase inhibitors andraises the possibility of using a protinin(Trasylol) to reduce the risk of dehiscenceof colonic anastomosis.

WP16Alpha gliadin sensitive suppression incoelliac disease

CLIONA O'FARRELLY, U MCKEEVER, C

FEIGHERY, C A WHELAN, AND D G

WEIR (Departments of Immunology andClinical Medicine, Trinity College, Dublin,and St. James's Hospital, Dublin,

Ireland) It has been proposed that animmunoregulatory abnormality isresponsible for the depressed cellularimmunity in sarcoidosis, Crohn's disease,and coeliac disease and thus may beinvolved in their pathogenesis. It has beenshown that the peripheral blood mono-nuclear cells (PBMCs) of Crohn's andsarcoidosis patients show increasedsuppressive ability when compared withnormal individuals. In this study, thegeneration of suppression was examined in26 untreated and 17 treated coeliac patientsusing a functional assay. A significantlyincreased level of suppression was found inboth groups of coeliacs when comparedwith the control group.Alpha gliadin sensitised PBMCs have

been demonstrated in treated anduntreated coeliac patients. The hypothesisthat these cells were involved in theimmunoregulatory abnormality describedabove was examined by testing the abilityof alpha gliadin to generate suppression.Alpha gliadin activated cells from treatedand untreated coeliacs demonstrated ahighly significant ability to suppress Con Astimulated autologous cells. Alpha gliadindid not generate suppression in threeCrohn's disease patients. Beta lacto-globulin and casein, which were alsotested, showed no significant differencebetween the level of suppression in treatedor untreated coeliac patients whencompared with controls.

It is possible that alpha gliadinsuppressor cells have developed to dampenthe cellular and humoral response to wheatprotein. Alternatively, perhaps sensitisedhelper cells have been immobilised in thegut in order to augment the reaction toalpha gliadin, thus depleting the peripheralblood of specific helper cells.

WP17Effects of oral vs subcutaneous EGF onage-induced changes in rat gut epithelia

A F WREN, S R. CROSBY, AND J B

ELDER (Department of Surgery, Man-chester University Medical School, Man-chester) Seventy per cent of all fatalBritish cancers arise from the epithelia ofthe digestive tract, skin, and secretoryorgans. As the incidence of gastric andcolonic adenocarcinoma increases withage, we have investigated the maturation-induced changes in DNA and proteincontent of rat gut epithelial cells. Theeffects of the mitogen EGF have also beenexamined. Male albino rats of either 2

months or 9+ months of age wererandomly allocated into four groups andgiven saline or EGF (10 ug/kg) by gavageor subcutaneous injection six times over 48hours. They were killed 16 hours after finalinjection and epithelial cells from discreteareas of the gastrointestinal tractharvested. The cell protein and DNAcontent were measured using the Lowryand ethidium bromide techniquesrespectively. There was a significantdecrease in both DNA and protein contentof cells from the small intestine (30% and24% respectively) and large intestine (45%and 34% respectively) of mature rats.Subcutaneous EGF reversed all thesematuration-induced changes. In contrast,oral EGF reversed only the changesoccurring in mature colonic epithelial cells.In addition, oral EGF significantlyincreased the protein content (from 0-512to 0-766 mg/106 cells) and DNA content(from 8-42 to 13-3 1£g/106 cells) of maturegastric epithelial cells. It is possible that amaturation-induced fall in DNA andprotein synthesis may result in smallerepithelial cells and that EGF can reversethis. EGF may also have a direct effectupon the gastric epithelium after oraladministration.

WP18Protein composition of ascitic fluid

G E GRIFFIN AND P KNOX (Departments ofMedicine II and Biochemistry, St George'sHospital Medical School, London)Ascitic fluid and plasma were collectedsimultaneously from patients (n=16) withcirrhosis, nephrotic syndrome, abdominalmalignancy, and tuberculous ascites. Theprotein composition of each ascitic fluidand plasma were compared using gelfiltration chromatography. Results areexpressed in terms of the level of protein inascitic fluid compared with the level ofprotein in homologous plasma (CA/CP).This ratio for total protein is variable (0.12to 0.78). When the ratio is calculated forproteins of differing molecular weights,however, two groups of ascitic fluidbecome apparent irrespective of totalprotein concentration. In one group theratio is the same for all proteins. In thesecond group CA/CP varies inversely withmolecular weight - that is, in comparisonwith plasma, proteins with lower molecularweight are more abundant in ascites.

Ascites results from an imbalance in theformation and removal of interstitial fluidsformed within the peritoneal cavity. From

A884


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/


current knowledge of the protein com-position of lymph from liver and otherabdominal organs we suggest that thepathophysiology of the two groups ofascites is distinct. In the first group(cirrhosis) the protein compositionsuggests a hepatic origin, while the proteincomposition of the second group suggeststhat it is an exudate from splanchniccapillaries. Thus the origin of ascitic fluidcannot be deduced from measurement oftotal protein but is, however, indicated bysimple chromatographic analysis.

WP19Aqueous solubilisation of triglyceride anddiglyceride in pancreatic lipase deficiency

D FINE, P ZENTLER-MUNRO, J C BATrEN, AND TC NORTHFIELD (Department of Medicine,St George's Hospital Medical School andBrompton Hospital, London) Pancreaticlipolysis is said to be a necessaryprerequisite for aqueous solubilisation oflipid. This assumption is based on thefinding that jejunal chyme from healthysubjects contains fatty acid and mono-glyceride, but virtually no triglyceride ordiglyceride. No similar studies have beenreported in pancreatic steatorrhoea. Wehave therefore studied eight patients withsteatorrhoea due to adult cystic fibrosis,and eight healthy controls. Jejunal samplesobtained after a Lundh meal were ultra-centrifuged overnight to separate theaqueous phase. Total lipids and fatty acidswere measured by titration with TEBAHwith and without saponification. Glycerideclasses were separated on thin layerchromatography and quantified by ahydroxamate-perchlorate method.Aqueous solubilisation of lipid was notmarkedly reduced in cystic fibrosis; meanfatty acid and non-fatty acid lipid (mM/1) inaqueous phase were 3.4 and 5.8 re-spectively, compared with 7-2 and 6.5 incontrols (NS). In preliminary studies,cystic fibrosis aqueous phase showed nodetectable glyceride in low pH samples(<5), but 1.0 to 1.5 mM/i glyceride at highpH (>6), of which 10% was mono-glyceride, 55% diglyceride, and 35%triglyceride. This pattern is consistent withlingual lipolysis; pancreatic lipolysisnormally proceeds to monoglycerideformation, and there was no detectablepancreatic lipase. We conclude that inpancreatic lipase deficiency lingual lipolysisprobably occurs; that diglyceride andtriglyceride can enter the aqueous phase;and that pancreatic lipolysis is not a

necessary prerequisite for aqueoussolubilisation of fat.

WP20FMRF-NH2-immunoreactive peptide in gutand pituitary of man and mammals

A ALI-RACHEDI, G L FERRI, I VARNDELL, JYATES, S VAN NOORDEN, L P C SCHOT, S RBLOOM, AND J M POLAK (Departments ofHistochemistry and Medicine, RoyalPostgraduate Medical School,Hammersmith Hospital, London) Themolluscan cardio-excitatory peptideFMRF-NH2 has recently been reported inendocrine cells of the gut and pancreas ofseveral mammals. The identity of theimmunoreactive cells has not yet beenclearly defined. By immunocytochemistry,we demonstrated numerous FMRF-NH2immunoreactive cells in the upper gut andpancreas of man, pig, cat, and guinea-pig.All immunostaining was abolished bypreincubation with FMRF-NH2 at 1 nmol/ml while a range of other peptides(including gastrin, CCK, met-enkephalin,PP hexapeptide, ACTH) had a partialeffect at 10 nmol/ml. The identity of thecells was established by comparing seriallystained 2 gm sections. The FMRF-NH2immunoreactive cells were characterised asG cells in duodenal and antral mucosa andPP cells in pancreas. By immunoelectron-microscopy, the FMRF-NH2 immuno-reactivity was localised to the smallelectron dense granules in G cells. Threegastrinomas and one PPoma were alsoinvestigated and found to containnumerous FMRF-NH2 immunoreactivecells. Both G- and PP- cells have beenrepeatedly reported to contain severalACTH-related peptides. We thereforestudied the adenohypophysis (man, ox,dog, pig, guinea-pig) and found FMRF-NH2 immunoreactivity to be localised incorticotrophs. In seven ectopic ACTH-producing tumours numerous FMRF-NH2immunoreactive- cells were found. Ourfindings demonstrate that a peptide similarto FMRF-NH2 is present in mammalsincluding man. It is of interest that it islocalised in cells producing ACTH orACTH-related peptides.

WP21Intestinal permeability to probe molecules

I HAMILTON, J ROTHWELL, D A ARCHER, AND AT R AXON (Gastroenterology Unit, GeneralInfirmary, Leeds and Department of

Organic Chemistry, University of Leeds,Leeds) The passive permeability of thesmall intestine has been studied using aseries of carbohydrate probe molecules inan animal model. The 10 molecules usedhad a molecular volume between 130x 103and 850x 103 nm3, and none is transportedby active mechanisms. An aqueoussolution of each molecule is injected into aligated loop of rat small bowel in vivo, thebladder is cannulated, and urine collected.The urinary recovery of each of the probemolecules after intravenous injection wassimilar, and urinary recovery thereforereflected the absorption of the probemolecule from the intestine. Moleculardimensions were calculated by a computerprogramme using x-ray crystallographicdata.

Urinary recovery of the probe moleculeswas related to molecular volume but not tomolecular radius or cross-sectional area. Itwas not possible to define a relationshipbetween recovery and molecular volumefor the entire series but molecules weredivided into two distinct populations.Those with a molecular volume greaterthan 250x 103 nm3 were absorbed lesscompletely than those with a smallervolume; there was only a slight change inabsorption with increasing molecularvolume for the large molecules, while amarked reduction in absorption of thesmaller molecules occurred with even veryminor increases in volume.The results demonstrate that the

absorption of hydrophilic molecules with avolume in excess of 250x 103 nm3 differsqualitatively and quantitatively from thatof smaller molecules supporting thehypothesis that the passage of watersoluble molecules across the intestinalmucosa depends on 'aqueous pores' whichare of limiting size. An alternative routeexists for molecules excluded from thepores because of their large volume.

WP22Hepatic and skeletal distribution of 65Zn incirrhosis

R W CROFTON, S GVOZDANOVIC, DGVOZDANOVIC, P J AGGETr, N A G MOWAT,AND P W BRUNT (Departments ofMedicine, Physiology and BiochemicalPhysics and Bio-Engineering, University ofAberdeen, Aberdeen) We have in-vestigated five healthy volunteers (HV)and six patients with hepatic cirrhosis(HC). Subjects were first injected with 18-5kBq + 0-25 mg elemental Zn; two months

A885


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/


later they were given an oral dose of 92.5kBq 65Zn incorporated into a standardmeal. An ultra-high sensitive imagingwhole body counter was used to determineabsorption, distribution and retention of65Zn.

65Zn absorption in the cirrhotic patientswas no different from that in the healthyvolunteers (HC 33-0±12-3%; HV31.6±1.2%). Whole body retention half-life after intravenous injection in thesepatients was significantly prolonged (HC299-8±74-8 days; HV 219-0±21-8 days,p<O0O5).

After intravenous injection of 65Zn therewas no difference in liver uptake (HC18-6±7-7% HV 23.5±3.6%). After oraladministration, however, there was asignificant difference in uptake (HC6-5±2.5%, HV 20-7±6-4%, p<0.05). Theturnover of 65Zn in the liver can beaccurately described by a two-compartment kinetics model. Afterintravenous injection there was the sameproportion of 65Zn in the first compartmentin the two groups (HC 8-8±7-7%; HV7.1±3.3%); after ingestion there was asignificant difference (HC 2-5±1-3%; HV17*3±8.5%, p<0.05). A ratio of countsover the sacroiliac and liver region wasused to quantify the transfer of 6Zn fromliver to bone, which was significantlyincreased in cirrhosis (HC 1-33±0-60; HV0*51±0*05, p<0.05).Abnormal Zn metabolism is well

recognised in cirrhosis. This study showsincreased skeletal depositions as well asimpaired hepatic uptake and retention ofZn in cirrhosis.

WP23Adhesins in enteroinvasive strains ofEscherichia coli

S KNUlTON, D C A CANDY, D R LLOYD, P H

WILLIAMS, AND A S MCNEISH (Institute ofChild Health, University of Birmingham,Departments of Genetics, University ofLeicester, Leicester) Enteroinvasivestrains of Escherichia coli (EIEC) producea dysentery-like illness. Little is knownabout the mechanisms of mucosal attach-ment and epithelial cell penetration inthese strains. EIEC strains 444-3 and 469-3isolated from children with dysentery-likediarrhoea cause a mannose-resistanthaemagglutination (MRHA) of humanerythrocytes, adhere strongly to HEp-2and HeLa cells but only weakly to adulthuman enterocytes. The adhesins havebeen isolated and partially purified from

both strains. They appear to be proteinswith subunit molecular weights of -14,500(444-3) and --14,000 (469-3), showhaemagglutination activity, and inhibit theadhesion of 444-3 and 469-3 to HEp-2 andHeLa cells. Mutants of 444-3 and 469-3which lack the adhesins also lackhaemagglutination activity and do notadhere to the HEp-2 or HeLa cells.Electron microscopy has revealed that theadhesins are not fimbrial antigens but areassociated with the bacterial cell wall, asclose apposition between cell membraneand bacterial cell wall is seen when thesestrains agglutinate human erythrocytes orbind to HeLa cells; adhesion to HeLa cellsis predominantly to cell surface microvilli.Invasion of HeLa cells by 444-3 and 469-3has been demonstrated by killing surface-associated bacteria with antibiotics andrecovering 'internalised' bacteria fromdisrupted cells. The extent of invasion was-1 bacterium/cell.

In conclusion, we have identifiedadhesins in two EIEC strains anddemonstrated invasiveness in an in vitrosystem. Our observations also suggest thatthe formation of specific molecular contactbetween bacterium and cell plasmamembrane is an important primary event inthe process of bacterial invasion.

WP24Hepatic, mesenteric, and splenic blood flowafter digestion

B H WALMSLEY, J S FLEMING, AND S JKARRAN (University Surgical Unit andDepartment of Nuclear Medicine,Southampton General Hospital,Southampton) Most of our knowledge ofthe blood flow to abdominal organs hasbeen derived from animal experimentationbecause of the invasive nature of themethods used. We have developed a non-invasive technique in the rat, which wasused in the present study to determinehepatic, mesenteric, and splenic blood flowafter digestion in man.Twelve healthy volunteers underwent

dynamic abdominal imaging after a bolusinjection of 99mTc sulphur colloid, whichis extr4cted by the R-E system. Initialstudies were performed after a 12 hour fastand a second study after a standardisedmeal of 1800 kcal. Computer analysis ofthe stored camera images enabled effectiveblood flow to the various organs to becalculated.

Total hepatic blood flow rose from amean of 1271±88-7 (SEM) ml/min to

1632±107 ml/min after the meal (p<0-01).Total mesenteric flow rose from 674±61-5ml/min to 1280±95 ml/min (p<0001).Splenic arterial flow fell from 257± 16mi/min to 187±16 ml/min (p<0.01). Meanhepatic arterial flow also fell from 577±40ml/min to 352±28 ml/min (p<0.01). Pulserate and blood pressure in the individualsat each imaging session showed nosignificant change.The reduction in hepatic and splenic

arterial flow described in this study lendssupport to the hypothesis that the increasein mesenteric blood flow brought about bydigestion is due to a redistribution of bloodrather than an increase in cardiac output.

WP25Problematical boundary conditions inunstirred layer modelling

M L LUCAS AND J C TAYLOR (Institute ofPhysiology and Institute of NaturalPhilosophy, Glasgow University,Glasgow) Calculations of unstirred layer(UWL) thickness b depend on a solution ofthe diffusion equation (DE). The solutionusually quoted is based on the unlikelyboundary condition that the membrane isimpermeable to solute. In fact, only if themembrane is highly permeable is the rateof transport across the UWL determinedby the latter's characterisitics. Thus theusual solution corresponds to an un-physiological membrane.A more reasonable procedure is to solve

the DE with varying degrees of mucosalpermeability at the UWL: membraneinterface. This was done by usingnumerical methods to solve thefundamental equation dC/dt = D.d2C/dx2with the boundary condition, flux throughmembrane = k(C, - C.) which is non-zero, unlike the currently-used solution.Here Cj is the interfacial solute concentra-tion, C. is an arbitrarily low concentrationacross the membrane. Substitution ofappropriate values of k for membranepermeability demonstrated unexpectedlythat the half steady state concentration atthe UWL: membrane interface wasreached earlier than the simple formulat(1/2) = 0-38 62/D allows for. This simplestcase solution is often used to calculate 6.With moderate to high membranepermeability, the proportionality factorcan change from 0 38 to 0-18 and values aslow as 0-10. The practical implications arethat the simple half-time formula may bean inaccurate approximation, especially if apermeant solute such as NaCI is used to

A886


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/


induce diffusion potentials from which (b)is estimated. One likelihood is that alteredmucosal permeability may be erroneouslyinterpreted as changes in UWL thickness indisease states.

WP26Sites of production and inactivation ofprostaglandin E2 in the rat small intestine:implications for the role of prostaglandinsin the gut

G S SMITH, G WARHURST, AND L ATURNBERG (Department of Medicine,Hope Hospital, Salford) Prostaglandins(PGs) can provoke secretion in the smallintestine and may play an important role inmodulating electrolyte and water transportin health and diarrhoeal disease.Endogenous PG levels will depend upontheir relative rates of synthesis anddegradation and the distribution of theseactivities in the intestinal wall will beimportant in determining their effects. Wehave therefore investigated these further inrat small intestine. Epithelial cells wereseparated from the subepithelium byvibration in EDTA buffer and the syntheticand degradative capacity of these fractionsdetermined, by RIA for PGE2 and radio-chemical detection of the PGE2metabolites formed respectively. Synthesiswas located almost exclusively in the sub-epithelial fraction (subepithelium76-5±15-6; epithelium 1-45±0*47p.mol.min- (mg.protein)- 1), whereasdegradation was found predominantly inthe epithelial cell fraction (subepithelium36-3±7*9; epithelium 274±22-3p.mol.min- ' (mg.protein)- '; n=6).Separation of the epithelial cells into avillous to crypt gradient demonstrated thatthe capacity to degrade PGE2 decreasedfrom villous to crypt, whereas the respon-siveness of epithelial adenylate cyclase toexongenous PGE2 increased and wasmaximal in the crypts. Subepithelialadenylate cyclase activity was poorlystimulated by PGE2.These data support the hypothesis that

PGE2 is synthesised in the subepithelialtissues and degraded in the epitheliumwhere it activates secretion. On the basis ofadenylate cyclase responsiveness secretionis likely to be maximally located in thecrypt epithelium.

WP27Influence of a new selective 5-HT2 receptorantagonist (ketanserin) on jejunal PGE2

release and ion secretion due to malignantcarcinoid syndrome

S ANTONSEN, M G J HANSEN, K BUKHAVE, ANDJ RASK-MADSEN (Department of MedicalGastroenterology, Odense UniversityHospital, and Department of Medicine C,Herlev Hospital, University ofCopenhagen, Denmark) Recent work hasimplicated prostaglandins (PGs) releasedby excessive 5-hydroxytryptamine (5-HT)in the pathophysiology of carcinoidsyndrome. A PG mediated 5-HT effect incarcinoid syndrome has been questionedbecause anti-inflammatory agents as well asclassical 5-HT antagonists appear ineffec-tive in ameliorating diarrhoea and flushing.Two distinct 5-HT receptors exist, but onlybinding to the 5-HT2 receptor correlateswith both in vitro and in vivo effects of5-HT. This has prompted us to investigatewhether the new selective 5-HT2 receptorantagonist, ketanserin, would affectjejunal release of PGE2 and - likesomatostatin - inhibit intestinal ionsecretion, diarrhoea, and flushing incarcinoid syndrome.

In a 68 year old female with metas-tasising midgut carcinoid oral treatmentwith ketanserin for eight days (20 mg tid)reduced immunoreactive PGE2 levels injejunal fluids from 1255 to 385 pg/ml(5-205 pg/ml-99% confidence limits),decreased stool volumes from 0-75 to 0-391/day (p<0.02), and relieved flushingattacks. Jejunal perfusion studies beforemedication revealed net secretion of fluid,Na+, Cl-, and K'. Ketanserin infusedintravenously for 60 minutes (10 mg + 2mg/h) decreased net secretion of fluid,Na4, and K' and increased lumen toplasma fluxes of Cl- (p<0.05). Additionalinfusion of somatostatin the following 60minutes (8 ,ug/kg/h) reversed net secretionto absorption (p<0.001). Spontaneoussecretion of Na+ and Cl- occurred mainlyas a consequence of increased plasma tolumen fluxes, which decreased (p<0.001)in response to combined ketanserin/somatostatin infusion. These resultssuggest that ketanserin and somatostatininhibited a PG-mediated 5-HT sensitivesecretory component of fluid and ionmovement in the small intestine of thepatient with carcinoid syndrome.

WP28Cholesterol esterase activity of humanintestinal mucosa: properties of the enzymeand effect of different bile acids

F CARUBBI, M PONZ DE LEON, P DI DONATO,AND N CARULLI (Isttituto di Clinica MedicaII, Universitd di Modena, Modena, Italy)Ithas been suggested that cholesterol absorp-tion is increased by cholic acid feeding,whereas it tends to decrease afteradministration of dihydroxy bile acids. Inthis context it has been shown, in rats, thatcholic acid and its conjugates stimulate theactivity of cholesterol esterase - an enzymewhich accounts for most of the esterifyingcapacity of the rat intestine - and in thisway may promote cholesterol absorption;in contrast, dihydroxy bile acids are muchless effective. The aim of this study was todetermine some general properties of thehuman cholesterol esterase and, inparticular, to investigate the effect ofdifferent bile acids on cholesterol esteraseactivity. Eighteen samples (2-5 cm) ofnormal terminal ileum were obtained atsurgery for total or hemi colectomy. Themucosa was scraped off and homogenisedin phosphate buffer containing theinvestigated bile acid. Cholesterol (23.2,umol + 1 ,uCi `4C-cholesterol) and oleicacid (139 ,mol) were added and themixture (total volume 6 ml) was incubatedat 37°C for one to 24 hours. Experimentswere carried out at 0-0, 0-1, 1-0, and 20mmol/l of bile acids. Cholesterol esterswere separated by TLC and the enzymaticactivity expressed as nmol of esters formed/h/g of mucosa.The time-activity relationship was linear

for two hours; by increasing the amount ofproteins (1.5, 3, and 6 ml of homogenate) alinear increase of esterification wasobserved. Subcellular distribution (byultracentrifugation) of cholesterol esteraseactivity was a follows: 78±12% in thesupernatant, 20±2% in the microsomes,and trace in the nuclei and mitochondria.Optimal pH ranged between 5.4 and 6-2.In absence of bile acids cholesterol esteraseactivity was 188±25 nmol/h/g. Bile acids,either individually or in a mixture ofphysiological proportion, did not changecholesterol esterase activity when added at0-1 and 1-0 mmol/l (range of activity154-206 nmol/h/g), whereas the activitywas significantly (p<0.01) reduced (rangeof activity 57-106) when bile acids wereadded at 20 mmol/l.We conclude that cholesterol esterase

activity is detectable in human ileal mucosaand shows similarities with the rat enzyme.At variance with this species, however,human cholesterol esterase does notappear to be affected by bile acids and,indeed, high concentrations may actuallyinhibit cholesterol esterase activity.

A887


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/


WP29Changes in the structure of the humansurface gastric mucus gel in disease

R WARD, A ALLEN, J P PEARSON, C WVENABLES, AND F YOUNAN (Department of.Physiological Sciences, Department ofSurgery, Medical School, University ofNewcastle upon Tyne, Newcastle uponTyne) The mucus gel covering the gastricmucosal surface consists of a highmolecular weight glycoprotein which, onproteolysis, results in a lower molecularweight form lacking the viscous and gel-forming properties of the nativeglycoprotein. We have shown that gastricmucus gel from mucosa resected for gastriculcer or duodenal ulcer contained higherproportions of lower molecular weightglycoprotein than that from normal mucosaobtained by pancreatoduodenectomy forperiampullary carcinoma.

In the present studies surface mucusfrom resected antral mucosa of patientswith gastric carcinoma containedsignificantly more lower molecular weightglycoprotein (76.8±2.4%) compared withpatients with gastric ulcer (65.1±2-8%),duodenal ulcer (50.2±3.3%), or normalmucosa (33.4±5.1%). The results wereunaffected by the use of isolation buffer atpH 2-7 or pH 7-0 or addition of a cocktailof proteolytic enzyme inhibitors; evidencethat the results reflect the state of themucus in vivo rather than degradationduring the isolation procedure.Our results show there is a higher

proportion of lower molecular weightglycoprotein in mucus gel from patientswith gastric pathology than in normalstomachs, implying weakening of thegastric mucus barrier.

PLENARYP1-P7

P1Epidemic hypochlorhydria

T GLEDHILL, MAXINE BUCK, ANGELA PAUL, B JADDIS, AND R H HUNT (Departments ofGastroenterology, Royal Naval Hospital,Haslar, Gosport, Hants) During studiesmeasuring the effect of antacid on the pHresponse to a meal, four cases of hypo-chlorhydria occurred. Six previouslyhealthy volunteers were to be studied,receiving six separate treatments. After400 ml Clinifeed 400 (Roussel), 10 mmsamples of gastric juice were aspirated and

pH recorded before return to the stomach.On the first study day, every subject had

a rise in pH to greater than 5-3 pH unitsafter Clinifeed. After six hours, the intra-gastric pH had fallen to less than 1-5 pHunits in every subject. Several study dayslater (range from two to four) four of thesubjects experienced a transient illnessconsisting of nausea, vomiting, and mildabdominal pain which usually lasted twodays. All subjects who experienced thisillness showed no pH response to Clinifeedon their next study day. Basal acid outputafter the illness was 0 mmol/h in everysubject and mean peak acid output was2-32 mmol/h (range 0-8) in response toimpromidine and, in one subject studiedwith pentagastrin, peak acid output was0-32 mmol/h. Gastroscopy showed anormal gastric mucosa but gastric biopsiesshowed an active superficial gastritis, mostmarked in the body of the stomach. Thesubjects have remained anacidic to date(two months). We presume the cause to bea virus, although serological screening hasnot revealed a causative agent.One previous outbreak of epidemic

gastritis with hypochlorhydria has beenreported from the United States but this isthe first in the United Kingdom. Trans-mission of illness may have been via acontaminated electrode and we thereforequestion the safety of returning gastricjuice to the stomach after pH estimation.

P2Optimal timing of surgery for peptic ulcerhaemorrhage

D L MORRIS, P C HAWKER, M H SIMMS, P W

DYKES, AND M R B KEIGHLEY (The GeneralHospital, Birmingham) Widely differingviews are held as to the correct indicationsfor surgery in peptic ulcer haemorrhage.No previous prospective randomised studyhas been done. Endoscopy is performed assoon as resuscitation allows. If the patienthas a duodenal or gastric ulcer, is over 30years of age, and has not had previousgastric surgery, he is randomised. Surgeryis performed if any of these criteria aremet: 'Early' (1) 4 units of blood, (2)rebleed, (3) endoscopic stigmata, (4)previous bleed. The corresponding criteriain the 'Delayed' group are: (1) 8 units ofblood, (2) two rebleeds, (3) persistenthaemorrhage. One hundred and eightpatients have entered the study: 80duodenal and 28 gastric ulcer. No patientunder 60 years of age has died; in the moreelderly there have been six deaths. Five of

these deaths have been managed by thedelayed policy and only one by the earlypolicy. Rebleeding occurred in only two of18 duodenal ulcer patients under the age of60 years managed by the delayed policy,whereas 11 of 21 of the elderly grouprebled (0.025>p>0.01). Early and delayedgroups are well matched.The excess of deaths in the over 60 years

delayed management group (15%) com-pared with the early group (3%), while notyet statistically significant, would lead us tofavour an aggressive diagnostic andsurgical policy, while in the young theapproach should be more conservative.

P3Nature of the bleeding point in massivelybleeding gastric ulcers

C P SWAIN, S G BOWN, P R SALMON, T CNORTHFIELD, J S KIRKHAM, AND JO'SULLIVAN (The Norman Tanner Gastro-enterology Unit and Department of Histo-pathology, St James's Hospital, and theDepartment of Gastroenterology,University College Hospital,London) This study relates endoscopicfindings to postoperative angiographic andhistological observations on the nature ofthe bleeding point when gastric ulcersbleed massively. Emergency endoscopy in580 unselected patients admitted for uppergastrointestinal haemorrhage revealedpeptic ulcers in 281; full examination of thecrater was possible in 231. Visible vesselswere identified in 110, other stigmata ofrecent haemorrhage in 47, and no stigmataof recent haemorrhage in 74. Ninety-fourpatients with stigmata of recenthaemorrhage, randomly allocated to noendoscopic treatment, were observed forevidence of rebleeding. Thirty-eight of 68patients with visible vessels rebled (56%)compared with two of 26 with otherstigmata of recent haemorrhage (8%)(p<0.001). Nineteen patients in whom avisible vessel was identified endoscopicallywere subsequently submitted to emergencysurgery for rebleeding and the resectedspecimen was available for study. In 18 ofthese (95%) the vessel identified endo-scopically was found in the specimen. Thevessel protruded above the ulcer crater infour, and clot in continuity with the breechin the vessel wall protruded in a further 12.The bleeding artery (mean externaldiameter 0-6 mm in the fixed specimen -

range 03-1.5 mm) was submucosal in 11and serosal in seven. Angiography, whentechnically possible, showed the breeched

A888


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/


artery to cross the ulcer in all. Thedominant pathological changes at thebleeding point were arteritis with fibrinoidnecrosis (14), loose fibrous intimalthickening (10), and recanalised thrombus(four). The ulcer penetrated to serosa in10. These observations are of relevance inplanning endoscopic therapy. They alsovalidate endoscopic identification of avisible vessel and confirm that suchidentification is helpful in predictingrecurrent haemorrhage.

P4Soya-induced pancreatic hypertrophy andrise of circulating cholecystokinin

T E ADRIAN, C PASQUALI, F PESCOSTA, A JBACARESE-HAMILTON, AND S R

BLOOM (Department of Medicine, RoyalPostgraduate Medical School,London) Soybean flour which is usedincreasingly for dietary protein supple-mentation in man contains a heat labiletrypsin inhibitor which causes pancreatichypertrophy in experimental animals. Thisincrease in pancreatic growth can be pre-vented by proximal intestinal resection,suggesting that a gut hormone mediates theeffect. Thirty male Wistar rats were fed onraw soybean flour for 21 days and 30controls were pair-fed with flour in whichthe enzyme inhibitor had been heat-inactivated (130°C, 30 min). Plasma CCKconcentrations were measured using asubtraction system with two antibodies,one raised to sulphated CCK-8 measuringall gastrins and CCKs and the other specificfor the gastrins. Both assays can detectchanges of 0-2 fmol/tube with 95%confidence. After 21 days the pancreaticweight of the group receiving active trypsininhibitor had increased by 76% comparedwith controls (2-71±0-08 g vs 1-53±0-04 g,p<OO005). Similarly, plasma CCK levelswere grossly raised in this group (24.5±3.1pmol/l vs 8.2±0*8 pmol/l, p<0.001). Therewas also a small increase in plasma gastrinconcentrations from 31±3 pmol/l incontrols to 46±4 with enzyme inhibitor(p<O.OO5).Thus pancreatic hypertrophy resulting

from a dietary protease inhibitor isaccompanied by an increase in CCK andalso gastrin. In the light of the currentepidemic of pancreatic cancer in thewestern world (and also soya consumption)the study of factors which cause hyper-trophy of this organ is of great importance.

P5Periampullary diverticula and commonduct calculi: a combined transhepatic andendoscopic technique for difficult cases

A R W HATFIELD, R S MURRAY, AND J ELENNARD-JONES (Academic Unit ofGastroenterology and Department ofRadiology, The London Hospital, White-chapel, London) The relationshipbetween periampullary diverticula andcommon bile calculi is not fully under-stood. The distal route of the bile ductaround the diverticulum may encouragestasis and hence stone formation. Thepresence of the diverticulum and site of thepapilla may make endoscopicsphincterotomy difficult if not impossible.

In a series of 263 consecutive patientswithout gall stones undergoing ERCP (agerange 20-90 years, mean 59 years) peri-ampullary diverticula were found in 20(8%). In 119 consecutive patients withcommon bile duct calculi (age range 32-92years, mean 67 years), however,diverticula were present in 42 (35%)(p<O.00l). If older patients (>60 years)without gall stones were examined,diverticula were seen in only 10 of 82patients (12%) (p<0-001).Endoscopic sphincterotomy was

attempted in 68 patients. Diverticula werepresent in 29 patients, with the papilla onthe edge in 21 and inside the diverticulumin eight. In three patients sphincterotomywas impossible as the papilla was inside thediverticulum. Using a transhepatic catheterand guide-wire, the papilla can be broughttowards the duodenal lumen. The orificecan even be dilated with a Grundzigballoon catheter. It is then easy to followthe direction of the guide wire and performa sphincterotomy.There appears to be a close association

between diverticula and common bile ductcalculi. This combined approach facilitatesendoscopic sphincterotomy even when thepapilla is hidden inside a diverticulum.

P6Selective and non-selective B blockade inthe reduction of portal hypertension

D WESTABY, A E S GIMSON, D BIHARI, I RCROSSLEY, AND R WILLIAMS (The LiverUnit, King's College Hospital and MedicalSchool, London) It has recently beensuggested that the reduction in portalpressure after propranolol administrationis due to a fall in hepatic blood flowresulting from the drug-induced reduction

in cardiac output. To test this hypothesiswe have studied the effect of a selective(metoprolol) and a non-selective B blocker(propranolol) on cardiac output (CO:thermodilution technique), estimatedhepatic blood flow (EHBF: ICGclearance), and portal pressure (wedged-free hepatic venous pressure) in 18 patientswith cirrhosis and recent variceal bleeding.Patients were randomly allocated toreceive a single intravenous dose of eitherpropranolol or metoprolol to reduce pulserate or cardiac output by 15%. There was acomparable reduction in pulse rate and COin both groups. Portal pressure fell signifi-cantly in both groups (p<0.005) but wasmore marked with propranolol (mean fall6-8 mmHg) than metoprolol (mean fall 3-6mmHg). In the propranolol group therewas a reduction in EHBF of 20.4% (pre1-4±SD 0-15 1/min; post 1-13±SD 0-161/min, p<0.005), while there was no changewith metoprolol (pre 1-15±SD 0-21 1/min;post 1-17±SD 0-23 1/min). In neither groupwas there any correlation between fall incardiac output and change in portalpressure or hepatic blood flow. There was asignificant relationship, however, betweenthe decrease in EHBF and the fall in portalpressure with propranolol (r=0.80,p<O-Ol).

Thus, the fall in portal pressure after Bblockade cannot be entirely explained by afall in cardiac output or EHBF. Thegreater fall in portal pressure and thesignificant changes in EHBF seen withpropranolol suggest that B2 receptorblockade in the splanchnic bed maycontribute to the reduction in portalpressure.

P7Alpha gliadin and ELISA (enzyme-linkedimmunosorbent assay): a serological test forcoeliac disease.

CLIONA O'FARRELLY, J KELLY, B BRADLEY, A

THOMPSON, G MACDONALD, W HEKKENS, AND

D G WEIR (Departments of Immunologyand Clinical Medicine, Trinity College,Dublin, and St James's Hospital,Dublin) The ELISA system was adaptedto measure antibody levels to wheatprotein with a view to developing a sero-logical test for coeliac disease. Using thisassay, there was no significant differencebetween the ELISA indices (EI) of 26untreated coeliac patients and 26 controlswhen the following antigens were used:digested gluten, gliadin, a and X gliadins,casein, or P lactoglobulin. In contrast,

A889


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/


using a gliadin, significantly raised EIswere found in coeliac disease but not in thenormal control group nor in 13 patientswith Crohn's disease (p<O0.01 in bothcases). These results indicated that aGAlevels are specifically raised in coeliacdisease.To test the validity of this assay for the

diagnosis of coeliac disease, oGA levelswere estimated in 75 patients being investi-gated for the disorder. The levels wereraised in 31/35 patients whose jejunalbiopsies were diagnostic of coeliac disease,whereas in 40 individuals with normaljejunal mucosa, four had raised otGAlevels (p<0.001).

Eleven patients were started on a gluten-free diet and serial serum samples wereobtained to determine the effect of the dieton the oGA levels. A further jejunalbiopsy was obtained six months afterstarting treatment. The jejunal biopsies ofeight patients were consistent with that oftreated coeliac disease and in 7/8 cases theaGA levels decreased with time. In threepatients, the jejunal biopsy was consistentwith untreated coeliac disease indicatingnon-dietary compliance; in each case thecxGA levels remained raised.The results of this study indicate that

estimation of (xGA levels by ELISA is auseful screening test for coeliac disease andmay be used to monitor compliance withgluten-free diet.

GENERAL IT1-T12

TiVitamin D status and bone histomorph-ometry after biliopancreatic bypass forobesity

S VEDI, J E COMPSTON, G J WATSON, L Pl?TON,AND N SCOPINARO (GastrointestinalResearch Unit, Rayne Institute, St Thomas'sHospital, London, and the Department ofSurgery, Ospedale S Martino, Genoa,Italy) In the operation of biliopancreaticbypass for obesity the biliopancreatic andalimentary tracts are separated withoutinterruption of the enterohepatic circula-tion of bile acids or formation of a blindloop. We have measured plasma 25-hydroxyvitamin D (250HD) levels andexamined bone histology in 41 patientsstudied 12-61 months (mean 32) afterpartial (n=21) or total (n=20) bilio-pancreatic bypass.

Metabolic bone disease was present in 30patients (73%), consisting of osteomalaciaand secondary hyperparathyroidism.Patients with bone disease weresignificantly older than those without(38.5±10.9 years; mean + SD, vs29-2±6-9, p<0-02), but did not differsignificantly with respect to type ofoperation, time since operation, or post-operative weight loss. Plasma calcium andalkaline phosphatase did not differ signifi-cantly between the two groups, but theplasma phosphate was significantly lower inpatients with bone disease (3.34±0.6mmol/l vs 3-86±0-5; p<0-02). Plasma250HD levels were normal in all 41patients.We conclude that metabolic bone

disease is common after either partial ortotal biliopancreatic bypass and isassociated with normal plasma 250HDlevels. The pathogenesis of the bonedisease is unclear, but simple vitamin Ddeficiency appears unlikely to beimportant.

T2Gastrointestinal manifestations of systemicvasculitis

M CAMILLERI, C D PUSEY, V S CHADWICK, ANDA J REES (Department of Medicine, RoyalPostgraduate Medical School, Hammer-smith Hospital, London) Involvement ofthe gastrointestinal tract is a recognised butpoorly characterised complication ofsystemic vasculitis. In 65 patients withsystemic vasculitis (and clinically pre-dominant renal disease), there wasevidence of gut involvement in 18 patients(27%): 4/8 with polyarteritis nodosa; 9/17with microscopic polyarteritis; 4/36 withWegener's granulomatosis, and 1/4 withChurg-Strauss syndrome. Histologicalevidence suggestive of systemic vasculitiswas detected in skin (40%) and renal(94%) biopsies. Clinical features includedabdominal pain (85%), diarrhoea (50%),gut haemorrhage (44%), and abnormalliver function tests (50%). These occurredinvariably with clinical evidence of activesystemic vasculitis; were evident at pre-sentation in 50% and constituted a fatalcomplication in 27% (two polyarteritisnodosa, two microscopic polyarteritis, oneWegener's granulomatosis). Ileus,perforation, mucosal abnormalities, andslow transit were detected on plain andcontrast radiology, while selective angio-graphy showed aneurysms (four poly-arteritis nodosa) and organ infarction (one

Churg-Strauss syndrome). Histology of gutbiopsies (chiefly rectal) revealed non-specific inflammation or ulceration in ninepatients and intramucosal haemorrhage intwo; intestinal arteritis was found only infull thickness surgical biopsies. Circulatingimmune complexes were detected in 40%cases. In the 73% patients who survived,gut features resolved with improvement ofsystemic illness with conventional treat-ment. We conclude that gastrointestinalinvolvement occurs frequently in patientswith vasculitis and may be fatal. Intestinalinvestigations and angiography may behelpful but are frequently normal.Vasculitis of the gut must be consideredwhenever abdominal pain, diarrhoea, orgut haemorrhage occur in a patient withactive systemic vasculitis.

T3Dark red bleeding as a marker comparablewith faecal occult blood testing in screeningfor large bowel neoplasms

R J NICHOLLS, A J SILMAN, P MITCHELL, FMACRAE, R J LEICESTER, C I BARTRAM, AND MSIMMONS (St Mark's Hospital, London,and The London Hospital, London) Theprevalence of large bowel symptoms andtheir association with large bowelneoplasms were investigated amongemployees of two industrial organisations.Results were compared with faecal occultblood testing (Haemoccult, Eaton Labora-tories).Symptom questionnaires and

Haemoccult kits were returned by 916 outof 1805 employees (>40 years) (com-pliance 50.7%). Twenty-eight (3.1%) wereHaemoccult positive and 114 (12.4%) hadone or more symptoms including bleedingin 108 (11.8%). All positives (129 persons)were examined by flexible endoscopy andbarium enema and polyps were removedfor histological examination. No cancerwas found but seven patients with anadenoma >10 mm in diameter werediscovered. Haemoccult was positive in sixbut all seven had at least one symptom:dark bleeding (four), bright bleeding(two), diarrhoea (one). Predictive valuesfor an adenoma >10 mm for Haemoccult(21.4%), dark bleeding (16.0%), anddiarrhoea (16.6%) were not significantlydifferent, but these values were geater thanthose for bright bleeding (2%), mucus(0%), and constipation (0%) (p<0.05).The predictive value of a positiveHaemoccult test plus at least one symptomincreased significantly to 46% (p<005).

A890


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/


Faecal occult blood testing in screeningmay not be identifying 'asymptomatic'cases as often as has been previouslypresumed. A symptom questionnaire mayreduce the false positive rate ofHaemoccult and dark bleeding might be auseful marker for neoplasms in screening.

T4When can antibiotics be used safely inpatients with active or recently treatedClostridium difficile colitis?

R P BOLTON AND M S LOSOWSKY (UniversityDepartment of Medicine, St James'sHospital, Leeds) The need for continuedantibiotic therapy in some patients withantibiotic-induced Clostridium difficilecolitis poses a real clinical problem and theeffect of concurrent or subsequentantibiotic therapy given for otherconditions on the course of the disease hasnot been established. Of 11 patients withC. difficile colitis, six received additionalantibiotics during therapy; all respondedclinically with clearance of organism andtoxin and none relapsed. Antibiotics weresubsequently used on 20 occasions in ninepatients after therapy with C. difficile hadceased. C. difficile colitis recurred in twopatients; both had responded well tovancomycin with clearance of faecal toxin,but both had residual faecal organisms.Seven patients remained well during 18subsequent episodes of antibiotic usage; allhad cleared C. difficile after initial therapyand none reacquired the organism or faecaltoxin. Patients receiving specific anti-microbial therapy for C. difficile colitis areunlikely to be affected by antibiotics givenfor intercurrent infection, and prophylaxisduring subsequent antibiotic therapy is notindicated in those patients who havecleared C. difficile from their stool.Relapse may occur during subsequent anti-biotic therapy in those patients withpersistent faecal C. difficile despite sympto-matic recovery, and prophylaxis isprobably indicated in this group ofpatients.

T5Does failure of bisacodyl-induced colonicperistalsis indicate intrinsic nerve damage?

D M PRESTON AND J E LENNARD-JONES (StMark's Hospital, London) Previous workhas suggested that patients with con-stipation show excessive colonicsegmenting activity. Patients complaining

of constipation can be divided into twogroups. Some (presumed irritable bowelsyndrome, IBS) have normal whole guttransit rate and pain is prominent; others(termed slow-transit constipation, STC)have a prolonged transit rate anddecreased bowel frequency is the majorsymptom. The motility of the pelvic colonhas been studied in 28 patients over onehour under resting conditions after a 12hour fast. The mean motility index for eachgroup was: controls 1803+297 SEM, IBS11,934+3066, STC 397±91. This findingsuggests that hypersegmentation occursonly in some patients complaining ofconstipation; others tend to have an inertcolon.

Eighteen patients with STC were studiedafter the introduction of a surface actinglaxative at 25 cm from the anus. In 11patients a normal response was seen withperistaltic waves passing caudally. Themotility index in this group over 30 minutesrose from 201±54 to 7201±1874(p<0.001). In seven patients there was noresponse. Motility index 153±68 to262±146 (NS). The patients whoresponded tended to be younger and havea shorter history. Six of seven non-responders reported being given regularlaxatives before the age of 10, comparedwith two of 11 who responded.The topical effect of bisacodyl may be a

test of intrinsic nerve function.

T6Elemental diets in the treatment of acuteCrohn's disease: a controlled study

C A O'MORAIN, A W SEGAL, AND A JLEVI (North wick Park Hospital andClinical Research Centre, Harrow,Middlesex) A randomised controlledsingle centre study was conducted tocompare the efficacy of an elemental diet(Vivonex) for one month with that ofprednisolone 0-5 to 0-75 mg/kg bodyweight. Patients were entered infto studywho had acute severe documented Crohn'sdisease and who had received no previousspecific therapy. The patients wereassessed by a prearranged clinical storingsystem for symptoms and signs, bodyweight, haemoglobin, erythrocytesedimentation rate, and serum albuminestimated at entry and weekly intervals forthe duration of the trial and at threemonths' follow-up. Twenty-one consecu-tive patients were entered into the trial,aged 15-68 years; 14 males, seven females.Ten received steroids and 11 patients an

elemental diet. The two groups werecomparable for age, site, and activity ofdisease. Two patients in each group werewithdrawn from the trial. Eight of the 10steroid-treated patients and nine out of the11 elemental diet treated group wereconsidered to be in clinical remission at theend of the trial period. There was asignificant fall in clinical score, ESR, andan increase in haemoglobin and albumin inboth treatment groups at four weekscompared with base line values whenanalysed by one-way analysis of variance.There was no difference in theseparameters between both groups. At threemonths one patient in each group hadrelapsed. The patients in the steroidtreatment group were receivingprednisolone 10-15 mg/day, whereas theelemental diet group were on no specifictherapy. We conclude from this study thatan elemental diet is as effective as steroidsin inducing a remission in acute Crohn'sdisease and is simple and safe toadminister.

T7Granulomatous crypt destruction: a subtlelesion of Crohn's disease?

G HOLDSTOCK, CLAIR DU BOULAY, C L SMITH,AND P ISAACSON (Department ofMedicine and Pathology, SouthamptonGeneral Hospital, Southampton) Wedescribe a lesion, which we have namedgranulomatous crypt destruction, whichcan be used to differentiate betweenulcerative colitis and Crohn's disease. Thelesion is characterised by a collection ofmacrophages, some almost epithelioid intype, and associated eosinophils, which aredestroying crypts. The lesion has beenfound in rectal biopsy material from 14patients with inflammatory bowel disease(seven males and seven females, age range24-72 years). The duration of diseasebefore recognition of this lesion rangedfrom six months to 12 years. All patientswere previously clinically considered tohave ulcerative colitis with diarrhoea,blood, and mucus. None had significantpain. Only one patient had an anal fissureand none had rectal sparing. Barium mealand follow-through was normal in allpatients and barium enema showed distalchanges in eight patients but was normal insix patients. Colonoscopy, with biopsies,revealed more extensive disease in all andshowed two patients to have 'skip lesions'and another five to have true granulomas inbiopsies from other sites. Furthermore,

A891


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/


review of histological material availableshowed preserved architecture, a paucityof crypt abscesses, and focal changes more

in favour of Crohn's disease than ulcerativecolitis.We consider granulomatous crypt

destruction to be an important finding in a

subgroup of patients with Crohn's disease.Furthermore, if granulomatous cryptdestruction represents a partial granulo-matous response it is suggested that inflam-matory bowel disease is a spectrum ofdisease possibly dependent on theindividual's ability to form granulomas.

T8Sulphasalazine-induced reversible maleinfertility in man and rat

C A O'MORAIN, P SMETHURST, AND A J

LEVI (Northwick Park Hospital andClinical Research Centre, Harrow, Middle-sex) In clinical studies semen wasanalysed from 26 men (age 19-30 years)taking 2-4 g sulphasalazine (SASP) dailyand compared with semen from 25 men(age 22-42 years) with inflammatory boweldisease not on SASP for the precedingthree months. There was a significantdecrease in sperm counts (31.9±6.4 vs

62.6+7*9x 106 ml) and motility 24-2±+16 vs49.3±3.1) progressively motile sperm, anda deterioration in morphology in thetreated patients. The semen qualityimproved three months after withdrawal.Fourteen pregnancies resulted after with-drawal (median 2-5 months). Threepatients fathered children while on thedrug. Semen acid phosphatase, fructose,and PGE2 were measured in 10 treatedpatients and showed no difference fromnormal levels. Gonadotrophin hormoneprofile was identical in eight patients whileon and off SASP for three months. Malerats fed SASP had a dose dependentreversible reduction in litter size whenmated (13.7 to 3.0 live foetus). Libido was

unaffected. The effect was reversed within14 days of drug withdrawal. The uteri offemale rats mated with treated malesshowed a significant decrease in fertilisedeggs at the two cell stage compared withcontrols (2.4 vsl2). Sulphapyridine but not5-amino-salicylic acid nor a 5ASA polymerinduced a similar reduction in litter size.We conclude that SASP causes a reversibleinfertility in man and rat. This is notrelated to ill health or inflammatory boweldisease. The time course, morphology, andrecovery experiments suggest a quanti-tative effect on later stages of sperm

maturation. The SP part of the SASPmolecule is responsible for infertility as

well as other side-effects and this shouldstimulate the formulation of othercompounds such as SASA polymers, foruse in inflammatory bowel disease.

T9Oral 5-amino salicylic acid for themaintenance of remission in ulcerativecolitis: a controlled trial

M J DEW, P J HUGHES, A D HARRIES, GWILLIAMS, B K EVANS, AND J

RHODES (Departments of Gastro-enterology, Pharmacy and Pathology,University Hospital of Wales, Cardiff) 5-amino salicylic acid (SASA) is the activeconstituent of sulphasalazine (SLP salazo-pyrine), which is released in the colon,healing and probably preventing relapse inulcerative colitis. The effect of 5ASA on

the colon is likely to be related to localrelease and a topical effect which could notbe achieved after simple oral adminis-tration because of absorption by the smallintestine. We have recently described a

'colonic delivery system' whereby oralpreparations coated with an acrylic-basedresin (Eudragit S) 120 ,u in thickness are

released in the colon. This double-blindrandomised study compares the effective-ness of SASA coated in this way with SLPfor the maintenance of remission inulcerative colitis. Seventy-two patients (36male and 36 female) with ulcerative colitisin remission on at least 2 g SLP/day (lessthan three stools per day without blood or

slime for one month) were given eitheractive SASA and placebo SLP or activeSLP and placebo SASA for 16 weeks. Theminimum dose of SASA was 1-2 g per dayand of SLP 2 g/day. Sigmoidoscopy was

performed at entry into the trial and at thetime of relapse or on completion of the trialperiod. Sixty-seven patients completed thetrial with 15 relapses, nine on SASA, and 6on SLP (p not significant). SASA coatedwith Eudragit S and SLP are equallyeffective for prevention of relapse inulcerative colitis.

T10Impact of Crohn's disease in childhood

J PUNTIS, A S MCNEISH, AND R N

ALLAN (Gastroenterology Units, Generaland Children's Hospital, Birming-ham) The impact of Crohn's disease inchildhood has been evaluated in a series of

67 children (M=37, F=30) whosesymptoms started before 16 years of age.The mean age at onset of symptoms was13-0 years and at diagnosis 14-5 years. Thedistribution of disease at diagnosis wasdistal ileum 37%, ileum + right colon19%, colon alone 27%, diffuse small boweldisease 13%, other 3%. During review21% developed diffuse small boweldisease. At presentation abdominalsymptoms were nearly always present(87%) often combined with weight orheight retardation (39%). Weight loss orheight retardation alone was rare (6%).

Retardation was usually corrected bysurgical intervention and was a permanentfeature in only five patients.

Cumulative recurrence rates after ileo-colonic resection were initially higher thanin the adult group: at one year 7% vs 5%,five years 33% vs 20%, but similar at 10years (54% vs 49%).

Nine patients have died, six with diffusesmall bowel disease. The causes of deathwere disease related (three), associateddisorders (three), drug induced (two),unrelated (one).Of the 58 survivors 38 are currently fit

and well with no evidence of recurrentdisease, 14 have evidence of recurrent orresidual disease but are symptom free,while six have asymptomatic recurrent orresidual disease.The initial high morbidity poses

challenges in management, but the long-term prognosis is good in most patients.

T1lWhat length of terminal ileum is requiredfor bile salt absorption?

G COOPER, B J ABEL, A G HUTCHISON, AND CMACKAY (Department of Surgery, WesternInfirmary, Glasgow, and Department ofUrology, Victoria Infirmary,Glasgow) The critical length of ilealresection necessary to cause malabsorptionof bile salts and vitamin B12 has not beenclearly defined by previous work on thesubject. With regard to bile salts, however,40 cm has been suggested.The use of terminal ileum in urinary

conduit construction results in an interest-ing model since a standard length, approxi-mately 25 cm, is isolated for the technique.Bowel habit, faecal bile salt excretion, andvitamin B12 were investigated in 12 patientswho had undergone this operation. Sevenhad developed looseness or frequency ofstool. Bile salt losses were determined bygas-liquid chromatography using three-day

A892


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/


stool collections. The mean faecal bile saltconcentration was 56-00 ,umol/g and theoutput 6-22 mmol/day in comparison withcontrol values of 24.97 ,umol/g and 0-53mmol/day respectively; both these differ-ences were statistically significant (p<0025and p<0001). Four patients hadsubnormal Schilling test results.Terminal ileal resections of the order of

25 cm, therefore, are likely to cause

abnormal faecal bile salt losses and may

result in vitamin B12 malabsorption.Cholegenic diarrhoea, cholesterol gall-stone formation, and megaloblasticanaemia are possible consequences ofusing ileum to make a urinary conduit.

T12Quantitative faecal excretion of "'In-labelled autologous granulocytes in assess-

ment of disease activity in Crohn's disease

S H SAVERYMUTrU, A M PETERS, M B PEPYS, J P

LAVENDER, H J HODGSON, AND V S

CHADWICK (Departments of Medicine andRadiology, Royal Postgraduate MedicalSchool, Hammersmith Hospital,London) Assessment of disease activityin Crohn's disease involves determinationof either clinical indices - for exampleCDAI- or laboratory (ESR, CRP, oroso-

mucoids) measurements. These are

indirect and non-specific correlates of gutinflammation. Neutrophil granulocyteinfiltration of the intestine is the character-istic histological feature of disease activitybut quantitative histology of biopsies fromdifferent and sometimes inaccessibleregions of the gut to assess activity is oftenimpractical. In preliminary studies with"`In labelled mixed leucocyte preparationswe reported that faecal "'Indium excretionafter intravenous administration of labelledcells correlated with disease activity ininflammatory bowel disease. We have now

developed a method of efficiently labellingpure granulocyte preparations in plasmausing a discontinuous plasma densitygradient technique and I 'In tropolonatechelation. With this technique we havecompared quantitative faecal granulocyteexcretion with conventional parameters ofdisease activity in 32 studies in patientswith Crohn's disease.

Results showed that faecal granulocyteexcretion ranged from 1-5 to 52% of theinjected dose and was significantlycorrelated with CDAI (r=0.731, p<0.001),ESR (r=0.676, p<0.001), and CRP(r=0.716, p<0-001). Quantitative faecalexcretion of "'In labelled granulocytes is a

reliable method of assessing diseaseactivity in Crohn's disease, is specific forgut inflammation, and is suitable forobjective assessment of disease activity intherapeutic trials.

LIVER/GUT IMMUNOLOGYT13-T24

T13Platelet associated immunoglobulins inprimary biliary cirrhosis (PBC)

M F BASSENDINE, J COLLINS, J STEVENSON, P

SAUNDERS, AND 0 F W JAMES (Departmentof Medicine, Freeman Hospital and Depart-ment of Haematology, General Hospital,Newcastle upon Tyne) Thrombocyto-penia in cirrhotic patients is classicallyattributed to splenic pooling, whereas inidiopathic thrombocytopenia purpura(ITP) it is related to surface boundimmunoglobulin (IgG). Recently plateletassociated IgG (PAIg) has been reportedin patients with chronic liver disease,mainly alcoholic cirrhotics. We have nowlooked for PAIg in primary biliarycirrhosis, as this is an 'autoimmune'disorder, in order to assess the frequencyand possible relationship of PAIg tothrombocytopenia in this condition. PAIgwas estimated by a method using radio-labelled staphylococeal protein A; in ourlaboratory normal subjects have a meanPAIg of 0.89±0i38 (n=50), whereas in ITP87.5% patients (n=48) have raised PAIg(>1.6). Spleen size was graded 0-3 oncolloid scan by an independent Qbserver.

Sixty-two primary biliary cirrhosispatients were studied (28 stage I-II, 34stage III-IV). Raised PAIg was found in 25patients (40%) of whom 18 had cirrhosis.Thrombocytopenia (<100,000) was foundin nine (15%), all were cirrhotic. Whereasthere was no correlation between plateletcount and amount of PAIg (r=0.14) orspleen size (r=0.03) in the total group, inthe nine thrombocytopenic patients all hadraised PAIg (p<0.0005) and eight hadspleen size 2 or 3 (p<0-005). Two of thesewith bleeding and platelets <30,000 weretreated with steroids with great benefit.We conclude that platelet bound

immunoglobulins, present in 40% patientswith primary biliary cirrhosis, may con-tribute to thrombocytopenia, particularlyin the presence of splenic enlargement.Steroid treatment may benefit suchpatients.

T14Different immunoregulatory mechanismsfor IgM production and B cell proliferation

K NOURI-ARIA, J NEUBERGER, AND A L W F

EDDLESTON (The Liver Unit, King'sCollege Hospital and Medical School,London) To investigate suppressor cellfunction in patients with primary biliarycirrhosis and its relationship to increasedIgM production, the ability of concanavalinA (Con A) stimulated suppressor cells toinhibit differentiation and secretion of IgMproducing cells was measured byhaemolytic plaque and immunofluorimetricassays in 20 normal subjects and 23 patientswith primary biliary cirrhosis. Monomericand pentameric IgM were also determinedby immunodiffusion in polyacryl-amide/agarose. The median number ofIgM producing cells in primary biliarycirrhosis was comparable with normalsubjects (1811 and 1798/106 lymphocytes,p>005), whereas the amount of IgMsecreted into culture media was signifi-cantly greater in primary biliary cirrhosisthan in normal subjects (1887 and 792ng/106 lymphocytes, p<0.01). Thus, IgMsynthesis per B cell was significantlyincreased in primary biliary cirrhosis (2.23compared with 0-83 ng, p<0-05).Autologous Con A activated suppressorcells inhibited IgM production significantlyless than controls (% suppression = 52-1 +SD 25-2 and 69-1±16-7, p<005), butinhibition of proliferation was comparablewith normal subjects. Monomeric IgM wasfound in 37% of primary biliary cirrhosissera and was related to the serum IgM levelmeasured by nephelometry. Theseobservations suggest that control ofproliferation of B cells and secretion ofimmunoglobulin may be under separateimmunoregulation and that in primarybiliary cirrhosis suppression of prolifera-tion of IgM producing cells isnormal butcontrol of secretion is abnormal. Whetherthis defect is due to an intrinsicabnormality of B cell function or asuppressor cell abnormality remains to bedetermined.

T15Significance of delta agent infection inchronic hepatitis B viral infection in GreatBritain

I V D WELLER, P KARAYIANNIS, A S F LOK, M

BAMBER, H C THOMAS, AND SSHERLOCK (Department of Medicine,Royal Free Hospital and School of

A893


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/


Medicine (University of London)London) Delta antigen (o) is a trans-missible agent requiring hepatitis B virus(HBV) for its replication. It is prevalent inItalian HBV carriers and more common inpatients with chronic liver disease. HBVcarriers with current o infection have aserum antibody (anti-b), in high titre,detectable by radioimmunoassay. Theprevalence of anti-o in British HBVcarriers is unknown. Seventy-one HBVcarriers born and resident in Great Britainwere studied. Fifty-nine had abnormalLFTs and 49 histologically proven chronicliver disease. Stored sera together withsamples from HBsAg negative controlswere coded and analysed 'blind' for anti-oby Dr M Rizzetto (Turin, Italy). Anti-6was positive in nine out of 71 (13%) BritishHBV carriers: six were intravenous drugabusers and two were haemophiliacs. Anti-o was negative in 30 homosexual HBVcarriers. Cirrhosis was more common inpatients with anti-b and those with anti-6positive cirrhosis were significantlyyounger than those with anti-b negativecirrhosis. Anti-b was not found in HBsAgnegative blood donors (n=100) orhaemophiliacs (n=79). Two out of 58HBsAg negative intravenous drug abusershad anti-6 in low titre and both hadanti-HBc and anti-HBs indicating recentinfection with HBV and b. In British HBVcarriers b infection is associated with intra-venous drug abuse and haemophilia andperhaps a more rapid progression ofchronic liver disease.

T16T-cell cytotoxicity and circulating T-cellsubsets in chronic HBV infection

M MONDELLI, G J M ALEXANDER, K NOURI-ARIA, A L W F EDDLESTON, AND RWILLIAMS (The Liver Unit, King's CollegeHospital and Medical School, DenmarkHill, London) The proportion ofperipheral blood lymphocytes (PBL)reacting with the monoclonal anti-T-cellantibody OKT8 is increased in untreatedpatients with chronic hepatitis B virus(HBV) infection and, as this antibodyreacts with a suppressor cell population,this has been interpreted as indicatingaltered immunoregulation in these cases.However, OKT8 also reacts with cytotoxicT cells and the possibility that these cellsare responsible for the increased propor-tion of OKT8+ PBL in this virus-related

disease has largely been ignored. We havetherefore investigated the relationshipbetween the cytotoxicity of T enrichedPBL for autologous hepatocytes and theproportion of OKT8+ cells measured byindirect immunofluorescence in 13 patientswith chronic HBV infection, eight of whomhad HBeAg in serum. Five were oncorticosteroids and one of these was alsoon cyclosporin. Percentage T-cell cyto-toxicity was higher in HBeAg positive thanHBeAg negative patients (40.1+25.4 and24.4±18.6). There was a positive correla-tion (r=0-81, p<002) between T cellcytotoxicity and the proportion of OKT8+cells which was independent of therapy butconfined to HBeAg positive patients.

In earlier studies with similar patients wehave shown that there is a negativecorrelation between the proportion ofOKT8+ cells and suppressor cell functionin HBeAg positive cases and this, inconjunction with the present results,suggests that the increased proportion ofOKT8+ cells reflects increased numbers ofcytotoxic T cells in the peripheral blood,and may be unrelated to any changes insuppressor cell function.

T17Reactivity of anti-hepatocyte antibodyRL23/36 with normal and abnormal humanliver tissue

C J HAWKEY, C H HOLMES, E B AUSTIN, B

GUNN, M J EMBLETON, R W BALDWIN, P GSMITH, AND P J TOGHILL (UniversityHospital, Queen's Medical Centre, Notting-ham, and Cancer Research Laboratory,University of Nottingham, Notting-ham) We have developed a monoclonalantibody RL23/36 which reacts with anantigenic determinant expressed on thesurface and in the cytoplasm of rat andhuman hepatocytes. We have studied itsreactivity with histologically normal andabnormal human liver tissue.Needle liver biopsies were obtained from

58 patients at diagnostic liver biopsy or atcholecystectomy. The reactivity of RL23/36 was examined on frozen sections usingan indirect immunoperoxidase technique.No reactivity was seen with cells other

than hepatocytes. In 24 out of 25 histo-logically normal liver biopsies diffuse cyto-plasmic binding of RL23/36 was observedin all hepatocytes. In histologically normalliver tissue from an insulin-dependentdiabetic no binding occurred. Failure of

binding also occurred in six of 33 histo-logically abnormal liver biopsies with thefeatures of autoimmune liver disease(three), haemochromatosis (one),alcoholic liver disease (one), and livertissue adjacent to a metastasis (one).RL23/36 did not bind to hepatoma tissuepresent in one liver biopsy.

Expression of the antigenic determinantwith which RL23/36 reacts is lost duringexperimental hepatocarcinogenesis and inthe only human hepatoma examined. Itsloss in liver diseases associated with anincreased risk of malignancy may lead to adeeper understanding of the pathogenesisof chronic liver disease and the develop-ment of malignancy.

T18Liver graft rejection in the immuno-competent subhuman primate

O EPSTEIN, P J SCHEUER, M C KEW, AND J AMYBURGH (Royal Free Hospital MedicalSchool and University of WitwatersrandMedical School, Johannesburg) The liveris considered 'immunologically privileged',and HLA matching is not considered acritical factor when considering suitabilityfor liver transplantation. To test thishypothesis in primates we have retro-spectively studied 120 liver-transplantedchacma baboons, none of which hadreceived postoperative immuno-suppression. Median survival was 28 days(range 0-390), and all animals surviving 14days had biochemical evidence of markedcholestasis. At necropsy, only 11% hadnormal-looking livers. Necropsy liverbiopsy specimens from 108 baboons weresuitable for histological evaluation. Liverhistology was normal in 12, and in 29 therewere features consistent with septiccholangitis or shock. In 67 animals, therewere features suggestive of liver rejection.These included mononuclear cell portalinflammation, damage to small calibre bileducts with or without vascular damage,portal sclerosis, fibrous septum formation,and varying degrees of sinusoidal infiltra-tion and cholestasis. Features of rejectionusually occurred seven days after trans-plantation, and cirrhosis was present in62% of 21 animals surviving 60 days. Weconclude that, although liver cells are not atarget for rejection, portal tract damage iscommon and results in cholestasis andrapidly developing cirrhosis.

A894


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/


T19HLA- A and B antigens in alcoholic andnon-alcoholic chronic pancreatitis: is non-alcoholic chronic pancreatitis an

autoimmune disease?

R A ANDERSON, D DONNAI, P DYER, AND J MBRAGANZA (Manchester Royal Infirmary,Manchester) HLA- A and B antigenfrequencies were determined in 88Caucasoids with chronic pancreatitis. Thepatients, who had been fully investigated,were divided into four groups dependingon whether or not alcohol was an

aetiological factor, and whether or notpancreatic calculi were detected bytomography. A detailed family history was

obtained from all patients.A and B antigen frequencies were not

significantly different between the studygroup as a whole, compared with 344controls. In the non-alcoholic group thefrequency of HLA-A1 was significantlyhigher than in controls (66.7% comparedwith 34.8%, p 0-002, Fisher's exact test);the frequency of HLA-B8 was also higherbut the difference did not reach statisticalsignificance (38 9% compared with27.9%). In the alcoholic group HLA-B21occurred more frequently than in controls(13.8% compared with 2%, p 0.001).There was a much higher incidence ofautoimmune disease in non-alcoholicpatients and their close relatives than in thealcoholic group.

In conclusion: (1) the alcoholic andnon-alcoholic varieties of chronicpancreatitis are associated with differentHLA antigens; (2) the development ofpancreatic calculi seems to be independentof tissue type; (3) HLA-A1 and B8 are

well-established markers of autoimmunedisease. The incidence of these antigensand of autoimmune diseases in the non-

alcoholic group suggests a possible auto-immune basis for non-alcoholic chronicpancreatitis.

T20Effects of neonatal antigen feed on subse-quent systemic and local CMI responses

S STROBEL AND A FERGUSON (Gastro-Intestinal Unit, Western General Hospital,and University of Edinburgh, Edin-burgh) Systemic CMI is measured byintradermal skin tests, intestinal mucosalCMI can be detected by measurement ofintestinal villi and crypts and intraepitheliallymphocyte counts. Adult mice given a

single feed of the protein antigen

ovalbumin (OVA) do not developintestinal CMI, and have systemic hypo-responsiveness (oral tolerance) whensubsequently parentally immunised andtested. We have tested the hypothesis thatthe age at first feed of a protein antigeninfluences subsequent immune responses:neonatal mice, aged 1 day, and adult micewere fed a weight-related dose of OVA orsaline, by intragastric gavage. Four weekslater all mice were immunised with OVA incomplete Freund's adjuvant and systemicimmunity measured by serum antibodiesand skin testing. Feeding of adult mice ledto oral tolerance. Neonatal feeding led topriming of systemic humoral (+46%) andCMI responses (+26%). To test theinduction of the local intestinal immuneresponse, mice immunised as above werechallenged with 2 mg% OVA in drinkingwater and mucosal morphology and IELcount measured. There was no evidence ofmucosal CMI induction in adult mice.However, mice fed OVA on the first day oflife and subsequently challenged hadsignificantly high IEL counts whencompared with saline fed ovalbuminchallenged litter-mates (9.3 vs 14-6/100,p<0.005). These experiments show thatprotein feeding on the first day of lifeprofoundly alters subsequent immuneresponses, causing priming of the immunesystem and induction of a local CMIresponse to the same antigen whenpresented later in life.

T21Local immune responses preceding clinicalrelapse in ulcerative colitis: changes inIg-containing cells, T-lymphocytes, andeosinophils

E LOPES PONTES, J PIRIS, D P JEWELL, AND S CTRUELOVE (Gastroenterology Unit,Radcliffe Infirmary, Oxford) Twentypatients with ulcerative colitis in completeclinical, sigmoidoscopic, and histologicalremission volunteered to be studied byregular monthly rectal biopsy over a periodof six months while on no treatment.During remission, every patient had anexcessive number of Ig-containing cells(especially IgG) and tissue eosinophils inthe lamina propria. Eleven patientsrelapsed and in six of these a pronouncedincrease in the number of Ig-containingcells (especially IgG) occurred at least onemonth before the clinical relapse. Tissueeosinophils showed a similar pattern. Bycontrast, tissue neutrophils continued to bescanty at this stage and only became

numerous at the stage of clinical relapse. Ina similar study of 10 other patients, T-lymphocytes were demonstrated by animmunoperoxidase method using a mono-clonal antibody (UCH TI). The T-lymphocytes resembled the Ig-containingcells in being raised even when the patientswere in remission and in showing a furtherincrease before and during clinical relapse.It is concluded that these patients withulcerative colitis in remission showevidence of a continuous antigenicstimulation and that the immunologicaldisturbance becomes more pronouncedbefore a clinical relapse supervenes.

T22Neutrophil luminescence in chronic inflam-matory bowel disease (CIBD)

D KELLEHER, C FEIGHERY, AND D GWEIR (Departments of Immunology andClinical Medicine, Trinity College, Dublin,and St James's Hospital,Dublin) Disorders of neutrophil chemo-taxis have been described in chronicinflammatory bowel disease. Superoxideproduction by polymorphonuclearleucocytes can now be examined usingbioluminescence techniques. Superoxideproduction by neutrophils stimulated withopsonised zymosan was measured in 29patients with chronic inflammatory boweldisease as compared with 20 normalcontrols.

Patients with chronic inflammatorybowel disease had decreased superoxideproduction (49±28) as compared withcontrols (106±26; p<0-001). When eightpatients with pure ileal disease wereexamined, however, their superoxideproduction was significantly higher(73±42) than patients with colonic disease(49±30), six of them lying within thenormal range. The patients with colonicCrohn's disease had higher superoxideproduction than those with ulcerativecolitis (30±20). No opsonisation defect wasevident in these studies, as determined byzymosan opsonised with patient serumcross-reacting with control cells. Toexclude the possibility that immunecomplex blockage of receptors wasresponsible a soluble stimulant (formylmethionyl leucylphenylalanine) was addedafter maximal stimulation with opsonisedzymosan. No further stimulation of activitywas observed. The depression of neutro-phil luminescence correlated with theseverity of clinical disease. It is possiblethat patients with colonic inflammatory

A895


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/


bowel disease have an intrinsic defect ofneutrophil function, while those withterminal ileal disease are able to localisethe disease effectively by a strong response

from the phagocytes. It is more likely,however, that these results represent a

secondary process analogous to neutrophillysozyme depletion.

T23Mucosal lymphocyte subpopulations ininflammatory bowel disease

W S SELBY, G JANOSSY, M BOFILL, G

GOLDSTEIN, AND D P JEWELL (Gastro-enterology Unit, John Radcliffe Hospital,Oxford, and Department of Immunology,Royal Free Hospital, London) Lympho-cyte subpopulations in the intestinalmucosa of patients with ulcerative colitis or

Crohn's disease have been studied using a

double marker immunofluorescence tech-nique. In tissue sections, over 85% ofintraepithelial lymphocytes (IEL) were Tcells (HuTLA+ UCHT1+'). The majorityof these T IEL were of suppressor-

cytotoxic phenotype (OKT8+: 83±2%),although only one-third of these reactedwith another T cell antibody, OKT1. Theremainder of the T IEL were of helperphenotype (OKT4: 17+2%). No Blymphocytes were found within theepithelium. Within the lamina propria,OKT4+ T cells predominated (OKT4+:62±1%; OKT8+: 38±1%). These findingsdid not differ from those seen in tissuesfrom controls. However, in active inflam-matory bowel disease the colonicepithelium, both surface and glandular,showed expression of HLA-DR antigens.This was not seen in inactive disease nor incontrol tissues.

In nine of the above subjects, lympho-cytes were isolated from the intestinallamina propria using an enzymatic tech-nique. Although OKT4+ and OKT8+ Tcell populations could be isolated frompatients and controls, comparison withresults from tissue sections indicated thatthe procedure tended to deplete OKT8+cells.These findings suggest that an imbalance

of mucosal immunoregulatory T cells, as

defined by monoclonal antibodies, doesnot occur in inflammatory bowel disease.The epithelial expression of HLA-DRantigens, however, may be of pathogenicimportance. The results also indicate thatfunctional studies of isolated intestinallymphocytes should be interpreted withcaution, and with reference to the results ofmorphological studies.

T24Natural killer (NK) cells and their activityin intestinal mucosa

P R GIBSON, E L DOW, W S SELBY, D P JEWELL,AND R G STRICKLAND (Radcliffe Infirmary,Oxford) Mononuclear cells (MNC)isolated from the intestine display markedreduction or absence of natural killer cellactivity when compared with autologousperipheral blood mononuclear cells. Thepresent study examines possiblemechanisms for this observation. Propor-tions of natural killer cells were determinedin isolated intestinal mononuclear cells(enzymatic separation), mucosal tissuesections, and autologous peripheral bloodmononuclear cells using the HNK-1 mono-clonal antibody. Natural killer function ofthe isolated mononuclear cells fromintestine and peripheral blood was studiedin parallel using a four-hour 5`Cr-releaseassay with K562 as the target cell. Theeffect of adjusting effector:target (E:T)ratios of intestinal mononuclear cells toapproximate those existing in peripheralblood was examined.

Proportions of HNK-1+ intestinal mono-nuclear cells ranged from 1-3% comparedwith 7-19% HNK-1+ mononuclear cells inautologous peripheral blood. The paucityof HNK-1+ mononuclear cells in theintestine was confirmed by parallel study oftissue sections which also showed thesecells to be located in lymphoid aggregatesand lamina propria. Natural killer activityof intestinal mononuclear cells at E:T of50:1 ranged from 0-5%; autologousperipheral blood natural killer activity atthis E:T ratio was 18-38%. Adjustment ofE:T ratios in intestinal mononuclear cellsresulted in the appearance of significantnatural killer activity (3-21%) in all experi-ments.

Previous observations of low or absentnatural killer activity in intestinal mono-nuclear cells are probably attributable tolow proportions of these cells in themucosa. However, natural killer activity isdemonstrable with these cells at high E:Tratios.

BASIC SCIENCE IT25-T36

T25Functional liver blood flow measured byXenon-133 clearance

A R TAYLOR, S A JENKINS, J JOHNSON, S K

SHIMIRTY, AND R SHIELDS (Department ofSurgery, University of Liverpool/RoyalLiverpool Hospital, Liverpool) Existingmethods of measuring the functionalcomponent (FHBF) of total (THBF)hepatic blood flow are highly invasive. Theclearance of Xenon-133, following its intra-parenchymal injection, may reflect FHBFand be of considerable clinical importance.Experimentally, in 10 healthy dogs, THBFwas measured by electromagneticflowmetry and by indicator dilution. FHBFwas measured by the established techniqueof dual indicator dilution and by theclearance of Xenon-133 injected intra-parenchymally into the liver.There was good correlation (r=0.83)

between THBF measured by electro-magnetic flowmetry (178.6+24.5) and byindicator dilution (200.9+18.6). Mostimportantly, a similar strong correlation(r=0.80) was confirmed between measure-ments of FHBF by dual indicator dilution(146.5+20.0) and by the intraparenchymalinjection of Xenon-133 (42.7+4.7),although the latter results were consider-ably lower. (All results expressed in ml/min/100 g liver tissue + SEM.) There wasno correlation between FHBF, measuredby Xenon-133, and THBF, measured byeither indicator dilution or electromagneticflowmetry.Thus FHBF may be simply assessed by

the intraparenchymal injection of Xenon-133, a procedure which may be readilyapplicable in the preoperative assessmentof cirrhotic patients.

T26Enhanced hepatic tumour growth afterportal venous diversion

G JACOB, A K C LI, AND K E F HOBBS

(Academic Department of Surgery, RoyalFree Hospital School of Medicine, London)Whereas suggestions that deprivation ofportal venous blood leads to regression ofestablished liver tumours in clinical andexperimental studies, the relationship ofportal perfusion to initial tumour growthhas not been fully investigated. We havestudied the growth of implanted tumour inrat liver after portal venous diversion.

Partial or complete portal venousdiversion after staged portal vein occlusionwas first confirmed in rats by corrosivecasts. Serial liver histology and hepaticenzyme profiles were recorded in theseanimals (n= 15). Immediately or at 48hours after partial or total portal diversion,further rats (n=20) had a single inoculum

A896


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/


of Walker carcinoma cells (5 x 104) injectedinto the liver. A control group (n=5)underwent sham surgery and received thesame inoculum. Ten days later, the liver,tumour, and body weights were measuredfor each animal.

Although none of the rats had detectablebiochemical or histological hepatic damageafter staged portal vein diversion, animalsinoculated 48 hours after total portaldiversion showed increases in tumour/liverand tumour/body weight ratios (p=002,t-test).We conclude that initial tumour growth

is enhanced by specific alteration of portalblood flow. Since there was no hepato-cellular damage the tumour growthstimulation must be unrelated to factorswhich cause cellular proliferation underthose conditions.

T27Synthesis of lipoxygenase and cyclo-oxygenase products by human colonicmucosa

C J HAWKEY, N K BOUGHTON-SMITH, AND B J RWHITTLE (Department of Medicine,University Hospital, Nottingham, andWellcome Research Laboratories, Becken-ham, Kent) The human colon synthesisesseveral prostanoids, which may have a rolein inflammatory bowel diseases. Aslipoxygenase products of arachidonic acidmetabolism have been implicated in theinflammatory process, we have investi-gated the formation of both lipoxygenaseand cyclo-oxygenase metabolites from '4Carachidonic acid (`'C-AA) by humancolonic tissue.Homogenates of human colonic mucosa

were incubated with `'C-AA and, afterextraction into diethyl ether, separated byTLC using two solvent systems thatallowed resolution of cyclo-oxygenase andlipoxygenase products. Human colonicmucosa converted 10±3% (n=7) of theadded 14C-AA. The predominant cyclo-oxygenase products, as identified by theirchromatographic mobility were, PGE2 >PGF2o, > PGD2 > TXB2 > 6-keto PGFI,,.The formation of these products wasinhibited both by indomethacin (1-10 ,uM)and the dual pathway inhibitor, BW755C(1-30 ,M). The predominant lipoxygenaseproducts formed, which had the chromato-graphic mobility of 11-,12-,15-HETE(which ran together) were inhibited byBW755C (30 ,uM) but not by indomethacin(3 ,iM). Using tissue from a colitic and anon-colitic patient, further resolution of

the 11-,12-,15-HETE TLC band was per-formed using normal-phase HPLC. Thenon-colitic tissue formed predominantly12-HETE, whereas both 12-HETE and15-HETE were major products formed bythe colitic tissue.The present findings indicate that human

colonic tissue can convert 14C-AA intoboth lipoxygenase and cyclo-oxygenaseproducts, as identified by chromatographicmobility and selectivity of inhibition byindomethacin and BW755C. Thelipoxygenase products formed by humancolonic tissue, including 15-HETE, mayhave a role in inflammatory bowel disease.

T28Bacteril neuraminidase: a possible cause ofrelapse in ulcerative colitis

J M RHODES, RUTH GALLIMORE, Z JMOJADDEDI, R ALLAN, E ELIAS, AND J F

KENNEDY (Departments of Medicine andMicrobiology, General and QueenElizabeth Hospitals, Birmingham, andBioactive Carbohydrate and ProteinResearch Laboratory, Department ofChemistry, University of Birmingham,Birmingham) Histochemical studies inulcerative colitis have shown abnormalsubstitution of sialic (neuraminic) acids incolonic mucus. This may cause reducedresistance to neuraminidase, so we areinvestigating the hypothesis that relapse ofulcerative colitis results from colonic over-growth by neuraminidase-producingbacteria.

Faeces were sampled from 28 patientswith ulcerative colitis, eight patients withCrohn's disease, and eight control subjects.Patients with ulcerative colitis were gradedinactive, mild, moderate, or severe. Faeceswere homogenised, centrifuged, and thesupernatant filtered through 0-22 ,um porefilters. Supernatant neuraminidase activitywas assayed fluorimetrically and expressedas me-umbelliferyl units/g pellet wt where 1unit=1 ,mol of 4 me-umbelliferonereleased/hour at 37°C, pH 4-6 caDglucosidase was assayed fluorimetrically asa marker of bacterial exo-enzyme activity.As expected there was no significant differ-ence overall in faecal neuraminidasebetween controls (56.6 u/g ± 56-1 SD),ulcerative colitis (44.9 u/g ± 52.8), andCrohn's disease (52.2 u/g ± 71.1). Therewas, however, a strong positive correlationbetween disease activity and faecalneuraminidase in ulcerative colitis: inactivemean neuraminidase 23-4 u/g, mild 22-3u/g, moderate 39-1 u/g, and severe 87-6

u/g; t for B = 0, 2-79; p<0.02. Nosignificant difference in faecal etglucosidase was found between controls(24.4 u/g ± 34-7 SD), ulcerative colitis(28.6 u/g ± 23.6), and Crohn's disease(49.7 u/g ± 60.6). a Glucosidase did notalter with disease activity in ulcerativecolitis so increased neuraminidase in activeulcerative colitis is not due to a non-specificincrease in bacterial exo-enzymes.These results support the hypothesis that

relapse in ulcerative colitis is due to colonicovergrowth by neuraminidase-producingbacteria.

T29Effect of bombesin on plasma cholecysto-kinin concentrations in normal subjects andpatients with partial gastrectomy

J B M J JANSEN AND C B H W

LAMERS (Division of Gastroenterology, StRadboud Hospital, Nijmegen, The Nether-lands) It has previously been shown thatinfusion of bombesin stimulates pancreaticenzyme secretion and gall bladdercontraction. Several mechanisms for theseactions have been suggested: direct effectof bombesin; activation of a duodeno-pancreatic reflex by bombesin; release ofcholecystokinin by bombesin. We havemeasured plasma cholecystokinin con-centrations during infusion of bombesin(100 ng/kg 20 min) in eight normalsubjects. In addition, eight patients withpartial gastrectomy were studied in orderto exclude an effect of bombesin-stimulated gastrin and gastric acid oncholecystokinin-release. Plasmacholecystokinin was measured by a specificand sensitive radioimmunoassay. Theantibody used was directed against themid-portion of cholecystokinin and did notshow any cross-reactivity with gastrin.

Infusion of bombesin induced significantincreases in plasma cholecystokinin in bothnormal subjects and gastrectomisedpatients. There was no significant differ-ences between basal plasma cholecysto-kinin (1.2±0.3 vs 1-4±0-3 fmol/l),bombesin-induced increases in plasmacholecystokinin (6.6±1.9 vs 5.7±1.0), andintegrated plasma cholecystokinin duringinfusion of bombesin (72.4±12.4 vs92-6±21-1 fmol/ml.20 min). Duringinfusion of bombesin serum gastrinincreased from 16-7±1-4 to 49-6±8-1 fmol/ml (p<0.005) in the normal subjects, whilethere was no significant change in serumgastrin in the antrectomised patients.We conclude that bombesin releases

A897


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/


cholecystokinin in man by a gastrin-independent mechanism.

T30Glucose and insulin levels in duodenal ulcerdisease before and after surgery

R J CADE, G H LESTER, G R PHILPOT, AND L R

CELESTIN (Departments of Gastro-enterology and Biochemistry, FrenchayHospital, Bristol) The aim of this studywas threefold: to investigate glucose andinsulin patterns in a group of patients withchronic duodenal ulceration (1) beforesurgery and (2) after surgery, and (3) toexamine the effect of a glucosidaseinhibitor (Acarbose) on any postoperativechanges.Ten males with proven chronic duodenal

ulcers and 10 male volunteers were eachgiven a standardised breakfast and venousblood samples collected every 15 minutesfor two hours. Seven of the former groupwere retested after selective vagotomy andpyloroplasty on two separate mornings, onone of which they were given 100 mgAcarbose with the breakfast.

Glucose concentrations in the duodenalulcer group were significantly higher thancontrols at 45, 60, and 90 minutes(p<0-05). The integrated response wasalso higher (p<0.005). There was nosignificant difference in insulin levels.

After surgery there was an earlyexaggerated peak in the glucose curvewhich did not reach statistical significance.From 45 to 120 minutes, however, glucoselevels were significantly lower than pre-operatively (p<005). Insulin levels weresignificantly higher postoperatively from 0to 45 minutes (p<0.01), but differences inthe later part of the curve were notsignificant.Acarbose attenuated both hyper- and

hypoglycaemic phases of the glucose curve,and decreased the insulin responsethroughout the study period.Duodenal ulcer patients, therefore, have

an abnormal glucose tolerance which isexacerbated by surgery and which, aftersurgery, is improved by Acarbose.

T31Cholecystokinins in human small intestine:distribution of molecular forms

P N MATON, A C SELDEN, AND V SCHADWICK (Department of Medicine,Royal Postgraduate Medical School,Hammersmith Hospital, London) Know-

ledge of the distribution of various forms ofcholecystokinins (CCKs) in humanintestine is limited and the similarity ofhuman and porcine cholecystokininsremains questionable. Accordingly weanalysed the distribution of various formsof cholecystokinin throughout the smallintestine.

Samples were washed in saline andstored at -20°C before processing.Mucosal and muscle layers were separatedand each boiled in water to extractcholecystokinins. The supernatants weresubjected to two-step reverse phase high-pressure liquid chromatography andsamples analysed using (1) a 'non-specific'radioimmunoassay detecting the carboxyl-terminal of all cholecystokinins andgastrins and (2) a gastrin-specific assay.

Total cholecystokinin immunoreactivityin mucosal samples was 137 ng CCK 8equivalents/g in jejunum, 23 ng/g in midsmall intestine and <1 ng/g in terminalileum. The majority of mucosal cholecysto-kinin (66-76%) co-chromatographed withporcine CCK 8. A second peak of immuno-reactivity (16-17% of total) co-chromatographed with porcine CCK 33/39.Assuming similar cross-reactivities ofporcine and human cholecystokinins withinthe assay, the molar ratios of CCK 8:CCK33/39 were 1-9-2-2:1. Jejunal musclecontained 13 ng CCK/g, principally co-chromatographing with CCK 8.Cholecystokinins were not detected in theremainder of small intestinal muscle (<1ng/g).We conclude that (1) human and porcine

intestinal cholecystokinins have similarhydrophobic properties; (2) most mucosalcholecystokinin is CCK 8-like and togetherCCK 8 and 33/39-like peptides constitute85% of total cholecystokinins; (3) onlysmall amounts of cholecystokinins arepresent in the muscle layers of human smallintestine.

T32Cephalic phase of pancreatic secretion inman: effects of sham feeding with andwithout atropine

A A ANAGNOSTIDES, P N MATON, AND V S

CHADWICK (Department of Medicine,Royal Postgraduate Medical School,Hammersmith Hospital, London) Therelative importance of neurogenic andhormonal influences on pancreaticsecretion in man is disputed. In order toexamine the role of vagal stimuli westudied the effect of sham feeding on

pancreatic outputs.After an overnight fast five normal

volunteers underwent a secretin-caeruleintest (secretin 1 CU/kg/h and caerulein 100ng/kg/h) using duodenal perfusion withgastric aspiration to measure both basaland 'maximal' pancreatic secretory outputsof trypsin, bicarbonate, and fluid. Onanother occasion a 40 minute period ofsham feeding (chewing and spitting baconsandwiches) was used instead of exogenousstimulation. On a third occasion the shamfeeding study was repeated after adminis-tration of atropine (0.6-12 mg intra-venously).

Basal and 'maximal' outputs of trypsinwere 743±441 (SD) and 3475±576 IU/30min respectively. Output during shamfeeding was 3192±1045 (p<0.001 vs basal)equivalent to 92% of 'maximal'. Atropinesuppressed basal output to 112±59(p<0.001) and output during sham feedingto 489±389 IU/30 min (p<0-001). Asimilar result was observed with fluidsecretion. Basal and 'maximal' outputs ofbicarbonate were 0.9±0*7 and 22±21mmol/30 min respectively. Sham feedingproduced only a modest increase inbicarbonate secretion to 3-3±2 (15% ofmaximal). Atropine produced no signifi-cant reduction in basal or sham feedingstimulated bicarbonate secretion.We conclude that (1) there is tonic vagal

stimulation of pancreatic trypsin secretionin the fasting state; (2) sham feedingproduces near 'maximal' secretion of bothtrypsin and water; (3) vagal stimuli are oflittle importance in bicarbonate secretion.

T33Trypsin secretion in response to endogenousand exogenous stimulation in chronicpancreatitis

H J KLASS, G I SANDLE, P KAY, P DAVIES, ANDJ M BRAGANZA (Manchester RoyalInfirmary, Oxford Road, Manchester)Studies in volunteers and patients withoutpancreatic disease showed that the meantryptic activity in duodenal juice after aLundh meal approximated the peak trypticactivity after intravenous injection ofpancreozymin (2 CHRu/kg). In patientswith chronic pancreatitis mean trypticactivity was disproportionately reducedcompared with peak tryptic activity - aphenomenon which could reflect impairedrelease of CCK-PZ from the intestinalmucosa, or excessive dilution of thesecreted pancreatic enzymes. Wedistinguished between these two possi-

A898


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/


bilities by comparing the output of trypsinafter endogenous and exogenouspancreatic stimulation using validateddouble-marker techniques in fivevolunteers and seven patients with chronicpancreatitis.Thje mean 10-minute trypsin output

after a test meal was similar to the peak10-minute trypsin output after pancreo-zymin in both groups. The cumulativevolume of gastric contents emptied into theduodenum two hours after the meal wassignificantly higher in the chronicpancreatitis group. In three patients withchronic pancreatitis gastric acid hyper-secretion was apparent, but there were nodifferences overall in the cumulativeduodenal acid load, or duodenal pHbetween the two groups.We conclude that the disproportionate

reduction in mean tryptic activitycompared with peak tryptic activity inchronic pancreatitis is not due to impairedrelease of CCK-PZ, but to excessivedilution of the secreted pancreaticenzymes.

T34Studies of the gastric 'mucus-bicarbonate'barrier: the influence of prostaglandin onthe pH gradient across rat gastric mucus invivo

I N ROSS AND L A TURNBERG (Departmentof Medicine, Hope Hospital,Salford) Previous studies have demon-strated a pH gradient across the mucuslayer adherent to gastric mucosa indicatingthe existence of a 'mucus-bicarbonate'barrier. We have now investigated themaximal pH in the mucus layer which themucosa can produce and the effect of a'cytoprotective' prostaglandin. Using anantimony chloride pH microelectrode, tipdiameter 50 At, the maximal pH recorded inthe mucus on anaesthetised rat gastricmucosa in vivo was determined afteradjusting luminal pH to 7 with phosphatebuffer (n=5). The maximal pH rose aboveluminal pH to 7-88±0-2 (range 7-59 to8.08), which is significantly higher thanluminal pH (6.85±0.33; p<0.001). Thisintramucus pH must be close to that ofbasal secretion of the surface mucosa. Ineight rats given 25 ,.&g of 16-16 dimethylprostaglandin E2 (DMPGE) sub-cutaneously the mean maximum pHrecorded in the mucus layer was 7-89±0-45compared with 7.43±0-56 in untreated rats

(p<0.05), when luminal pH was 2.Luminal application of aspirin (20 mmol)dropped the intramucus pH in seven ofeight untreated rats but in eight rats treatedwith DMPGE only two showed any fall onexposure to aspirin. Mean intramucus pHfell from 6-9 to 6.85 in treated rats andfrom 6.8 to 5-3 in untreated rats (p<0.005).DMPGE did not prevent the fall in intra-mucus pH induced by n-acetyl cysteine orluminal pH 1.4 HCl.These results suggest that the pH of the

fluid secreted into mucus is alkaline (meanpH 7.88) and that prostaglandins enhancethis alkaline component and preventdissipation of the gradient by aspirin.These observations support a role for the'mucus-bicarbonate' barrier in mucosalprotection.

T35Effect of a mast-cell stabilising agent ongastric acid secretion in response to penta-gastrin in conscious dogs

S P CANFIELD AND B P CURWAIN (introducedby D L Wingate) (Department ofPhysiology, St Mary's Hospital MedicalSchool, London) FPL 52694 stabilisesmast-cells and has been shown to inhibitpentagastrin stimulated secretion inanaesthetised, pylorus ligated rats anddogs. We have studied the effect of FPL52694 in conscious beagles with gastricfistulae during continuous intravenousinfusion of pentagastrin (2 ,ug/kg/h). In oneseries of experiments FPL 52694 (5 or 10mg/kg/h) was added to the intravenousinfusion after 90 minutes for the next 60minutes. Results were similar at both dosesand a maximum inhibition of acid output of53±8% mean, SEM, n=4) was seen after45 minutes. The volume of secretion wasreduced to a lesser extent (30±10%) andthe [H'] in the juice was reduced by28±10%. In control experiments therewere no significant changes in anyparameter. In a second series FPL 52694was given intragastrically (4-5 mg/ml indistilled H20 given as 4x10 ml aliquotsover 30 minutes with the fistulae closed) 90minutes after beginning intravenousinfusion of pentagastrin. The fistulae werethen opened, the stomachs allowed todrain and secretion followed for a further60 minutes of pentagastrin infusion.Maximum inhibition of acid output was90±4% (n=5) after 15 minutes and was74±6% after 60 minutes. At 15 minutes the

volume of secretion was reduced by60±12% and [H'] by 79±5%.FPL 52694 inhibits pentagastrin stimu-

lated acid secretion in conscious dogs and ismore potent when given intragastricallythan intravenously. The possibility thatmast-cell stabilising agents offer analternative method for control of gastricacid secretion merits further investigation.

T36Inhibition of gastric acid secretion byAH22216, a new long-acting histamine H2-receptor antagonist

J M HUMPHRAY, M J DALY, AND R STABLES(introduced by J J Misiewicz) (GlaxoGroup Research Limited, Ware, Hertford-shire) AH22216 is a new, potenthistamine H2-receptor antagonist with aprolonged duration of action; it differsqualitatively from cimetidine andranitidine in that it does not appear tobehave competitively in vitro.The antisecretory activities of AH22216

and ranitidine have been compared in thedog. AH22216 was four to 10 times morepotent than ranitidine as an inhibitor ofgastric acid secretion induced by histamineor pentagastrin in the Heidenhain pouchdog (n=5). Antisecretory ED50 values inmg/kg (95% confidence limits) forAH22216 were 0.019 (0.016-0.025) intra-venous and 0-029 (0.025-0.033) p.o. vshistamine, and 0.025 (0.020-0.029) intra-venously and 0.049 (0.034-0.066) p.o. vspentagastrin. AH22216 0.03-0.10 mg/kgintravenously, inhibited the secretoryresponse to a test meal in dogs (n=3) witha gastric fistula. The duration of action ofAH22216 was very prolonged comparedwith that of ranitidine. At 1 mg/kg p.o.,ranitidine inhibited histamine-inducedgastric acid secretion by 95±2% at twohours, 62±8% at four hours and 23±9% ateight hours, whereas an equipotent dose ofAH22216, 0-10 mg/kg p.o., inhibitedsecretion by 67±9% at two hours, 95±3%at four hours, 89±2% at eight hours,54±8% at 18 hours, and 32±13% at 24hours. In a separate series of experimentsin the dog, secretory dose-response curvesto histamine were displaced in a parallel,dose-related manner by AH22216,indicating competitive antagonism.

In summary, AH22216 is a potent, long-lasting inhibitor of gastric acid secretion inthe dog, and warrants investigation in man.

A899


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/


SURGERYF1-F14

FlSurgical treatment of gastro-oesophagealreflux by partial fundoplication (Lindprocedure)

C J STODDARD AND J A R SMITH (UniversitySurgical Units, Royal Liverpool Hospital,Liverpool, and Royal HallamshireHospital, Sheffield) The Belsey Mark IVand Nissen procedures are effective opera-tions for the control of gastro-oesophagealreflux (GOR) but may be followed bydysphagia, gas bloat syndrome, andrecurrence. We have evaluated a techniqueof transabdominal partial fundoplication,the Lind procedure, in 25 patients withgastro-oesophageal reflux.

All patients had symptoms of gastro-oesophageal reflux which had failed torespond to medical treatment, radiologicalevidence of reflux, and endoscopic histo-logical evidence of oesophagitis. Patientsunderwent clinical, radiological and mano-metric evaluation pre- and postoperatively.A transabdominal 2700 partial fundo-plication was performed in all patients andin three patients this was combined withsimultaneous dilatation of an oesophagealstricture and with vagotomy for duodenalulceration in three others. Five patientsrequired splenectomy for intraoperativesplenic trauma.At a six months' postoperative assess-

ment no patients were suffering fromheartburn, acid regurgitation, or the gasbloat syndrome. Five patients had experi-enced a transient dysphagia in theimmediate postoperative period. Gastro-oesophageal reflux was still present radio-logically in only one patient. Post-operatively, there was a significant rise inpressure at the lower end of theoesophagus p<0O01. We conclude that thisis a satisfactory technique for the control ofgastro-oesophageal reflux.

F2Cholecystectomy and vagotomy - anunhappy combination?

M W N WARD, C G CLARK, AND D

KARAMANOLIS (Department of Surgery,University College London, the RayneInstitute, University Street, London) Gallstones and duodenal ulcer coexist in some10% of duodenal ulcer patients requiringsurgery, and it has been shown that there isa high incidence of post-vagotomy

diarrhoea when these patients are treatedby combined cholecystectomy and truncalvagotomy and pyloroplasty. There are,however, no reports of the outcome ofcombining cholecystectomy with PGV. Wehave therefore compared the results ofsurgery in 64 patients subdivided into fourmatched groups; (1) PGV, (2)PGV + cholecystectomy, (3) TV + PY,(4) TV + PY + cholecystectomy.There was no diarrhoea after PGV alone

but when combined with cholecystectomydiarrhoea occurred in 30% of patients. ForTV + PY, diarrhoea occurred in 25% butwhen combined with cholecystectomy thisfigure increased to 60%, with a greaterproportion of more troublesomediarrhoea. The aetiology of post-vagotomydiarrhoea is speculative but some form ofbile acid malabsorption may beresponsible. This view is supported byobservations comparing the outcome ofPGV with and without cholecystectomy.Whichever combination of operation isused for treatment there is an increase inpost-vagotomy diarrhoea, but the lesser ofthe two evils appears to be the combinationof cholecystectomy with PGV.

F3Infection after cholecystectomy in aprovincial district hospital

A FLOWERDEW, D BADENOCH, AND R GFABER (Department of Surgery, BattleHospital, Reading) This prospectivestudy was made of 183 consecutive patientsadmitted under the care of one surgeon.The objects were to determine (1) theincidence of organisms in the bile and itsrelationship to infection; (2) the incidenceof systemic and wound infection; (3) theeffects of certain clinical factors oninfection and organisms in bile.For analysis the patients were divided

into four groups: group I: 162 patientsundergoing planned cholecystectomy only;group II: eight patients undergoing semi-emergency cholecystectomy; group III: 22patients undergoing unexpected explora-tion of bile ducts; group IV: 19 patientswith obstructive jaundice due to stones.

Forty-five patients (23%) had positivebile cultures; group I, 18 (13%); group II,five (63%); group III, 10 (45%); group IV,12 (63%). One patient in group IVdeveloped septicaemia despite havingprophylactic antibiotics. No patientdeveloped intra-abdominal infection.Twenty-two patients (12%) developedwound infections; group I, 11 (7%); group

II, two (25%); group III, five (23%); groupIV, four (21%). In only six patients,however (27% of wound infections, 3% ofwhole series), was the same organismcultured from wound and bile. In group IVtwo wound infections occurred in 11patients given prophylactic antibiotics,while two infections occurred in eight whowere not.Contrary to previous reports, the

incidence of systemic and wound infectionis lower, and organisms in the bile rarelycause infection.

F4Management of portal hypertension andpost-cholecystectomy bile duct strictures

C J KELLEY, G A D MCPHERSON, AND L HBLUMGART (Hepatobiliary SurgicalDepartment, Hammersmith Hospital, andRoyal Postgraduate Medical School,London) Damage to the biliary tree occursonce in every 400 cholecystectomies and ifuncorrected will result in biliaryobstruction and biliary fibrosis. Portalhypertension will then develop in a signifi-cant proportion of patients especially if theobstruction is prolonged and associatedwith frequent episodes of cholangitis. Bothof these factors may be influenced by thenumber of previous attempts at estab-lishing biliary drainage. Once portal hyper-tension is established subsequent stricturerepair is associated with a high mortality.Of 69 patients referred with such

strictures nine (13%) had portal hyper-tension. Eight patients had multiplelaparotomies before referral and all hadrecurrent cholangitis. Five patients hadclinically evident portal hypertension,three were diagnosed on either endoscopyor angiography, and one at laparotomy.Four patients had portasystemicanastomoses, one with a simultaneoushepaticojejunostomy. Four had hepatico-jejunostomy alone and one hadlaparotomy at which no procedure waspossible. Five patients have subsequentlydied of liver failure.

If biliary bypass is performed early in thecourse of the disease portal hypertensionmay regress. After prolonged obstruction,however, a portasystemic shunt may benecessary to lessen peroperative complica-tions. Portal hypertension must thereforebe considered in all patients with benignstrictures and confirmed by endoscopy andangiography when necessary. Earlydiagnosis is essential when planning theappropriate treatment, which may still beassociated with a high mortality.

A900


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/


F5Prolonged irrigation of the pancratic bedafter debridement for severe necrotisingpancreatitis

N J MCC MORTENSEN AND H JESPINER (Department of Surgery, BristolRoyal Infirmary, Bristol) The mortalityfrom severe necrotising pancreatitis mayapproach 70%, and the survivors oftenrequire repeated operations to debride thepancreas and drain abscesses. We reportthe results of prolonged irrigation of thepancreatic bed after surgical debridementin nine patients with severe necrotisingpancreatitis. Patients with pancreaticabscess or pseudocyst presenting later intheir clinical course have been excluded.

All nine patients were critically ill withhypoalbuminaemia, leucocytosis, and afalling haemoglobin (mean 3.9 g/dl). Atdefinitive surgery an average of 17 days(range eight to 25 days) after onset ofsymptoms the pancreatic slough wasthoroughly debrided and two or more largedrains were placed in the pancreatic bed.Irrigation with saline (2 I/day) was startedafter two days and continued for a mean of23 days (range 10-54 days).There were two deaths, five and 25 days

after surgery, and one of these patientsrequired packing for massive postoperativehaemorrhage. No other patient requiredfurther surgery. Four patients developedfistulae (one gastric, three biliary) which allclosed spontaneously.After thorough debridement and

drainage, prolonged irrigation of thepancreatic bed may reduce mortality andthe need for repeated drainage proceduresin patients with severe necrotisingpancreatitis.

F6Identification of patients with 'occult'complications of acute pancreatitis

A D MEYER, M J MCMAHON, MARGARETBOWEN, AND E H COOPER (UniversityDepartment of Surgery, General Infirmary,Leeds, and Unit for Cancer Research,University of Leeds, Leeds) Patients whodevelop pancreatic collections (pseudocystor abscess) after an attack of acutepancreatitis are frequently discharged fromhospital without the presence ofcomplication being suspected. In this studymarkers of continuing inflammation havebeen evaluated in order to select thosewhich might be useful to screen patients forthese complications.

Body temperature, leucocyte count(WBC), ESR and plasma levels of C-rective protein, a1-protease inhibitor andanti-chymotrypsin were measured duringthe initial 13 days in hospital in 53 patientswith acute pancreatitis. Sixteen patientshad an attack defined as severe and 11 ofthem developed a pseudocyst or abscess.The initial part of the illness was clinicallymild in seven of the patients whodeveloped complications, and themeasured parameters showed only limiteddiscrimination between severe and mildattacks during the first five days. Betweenfive and 13 days 95% confidence limits forC-reactive protein in severe attacks showedcomplete separation from those for mildones. The leucocyte count also hadpredictive power but other parameterswere less useful.

This study suggests that (1) pancreaticcollections originate early during thecourse of acute pancreatitis, (2) measure-ment of C-reactive protein in the secondweek of pancreatitis can select patients inwhom pancreatic collections aredeveloping.

F7Intestinal microcirculatory changes in thenormal and diseased bowel

N D CARR, B PULLAN, AND P FSCHOFIELD (Teaching Unit Five,Withington Hospital, University Hospital ofSouth Manchester) Several papers havereported similarities between colonicischaemia and inflammatory bowel diseasein older patients. Occlusive vascularchanges have been noted in the smallintramural intestinal vessels in inflam-matory bowel disease. We have investi-gated and quantified the intestinal micro-circulation in the normal and abnormalfrom excised specimens of bowel. Thespecimens were perfused with bariumsulphate suspension, which penetrates tothe smallest intramural vessels. Theconcentration of barium has beenestimated using an x-ray fluorescencemethod in tissue blocks of resected bowel.Twenty-one control cases, eight patients

with Crohn's disease, three patients withulcerative colitis, five patients withradiation bowel damage, and two patientswith colonic ischaemia have been studied.Our results indicate that: (1) there is a

highly significant negative correlation(p<0.002) between the concentration of

barium and age in the normal group(r=-0.73); (2) the Crohn's disease groupshows a highly significant reduction in theconcentration of barium (p<0.001); (3) asimilar reduction in the concentration ofbarium was seen in radiation enteritis andischaemic colitis (p<0-001).This reduction in microcirculatory

volume may play a role in the pathogenesisof Crohn's disease, radiation boweldamage, and colonic ischaemia, but not inulcerative colitis.

F8Prevention of peritoneal adhesions afterthermal injury using the formaldehydecarriers noxythiolin and taurolin

D J LEAPER AND A KAPLUN (introduced by RC N Williamson) (Departments ofSurgery, Westminster Hospital and BristolRoyal Infirmary) Although postoperativeperitoneal adhesions have become thecommonest cause of intestinal obstructionmany adhesions may be protective inwalling off vulnerable suture lines,damaged bowel, or inflammatorycollections.

In this study mice were submitted tolaparotomy under intravenous nembutalanaesthesia. An unpublished method ofproducing a measurable peritonealadhesion was assessed by subjecting thecaecum of each mouse to a thermal injuryof 69°C for 5 s using a water heated 2.5 mmbrass disc. Before abdominal closuregroups of mice had 0-1-0.75 ml isotonicsaline (controls) or 2-5-6 25 mgnoxythiolin or taurolin in similar amountsof saline instilled intraperitoneally. Allmice were killed seven days later andperitoneal adhesions examined withoutknowledge of the treatment groups. 61/65controls developed an adhesion to thecaecal burn compared with 41/66 in thenoxythiolin treated mice (p<0.001) and49/68 in the taurolin treated mice(p<0.002). There was no macroscopicevidence of caecal dehiscence or bacterialcontamination from the burn site.The formaldehyde carriers noxythiolin

and taurolin have both been shown toinhibit the formation of peritonealadhesions produced by thermal injury. Onthe surgical occasions where adhesionformation should be minimised, as ingeneral bacterial peritonitis, it is suggestedthat intraperitoneal instillation ofnoxythiolin or taurolin may be of clinicalvalue as an adjunct to meticulous surgicaltechnique.

A901


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/


F9Non-invasive assessment of intestinalischaemia

B H WALMSLEY, J S FLEMING, AND S J

KARRAN (University Surgical Unit andDepartment of Nuclear Medicine, South-ampton General Hospital, South-ampton) The diagnosis of mesentericischaemia is often difficult and delay insurgical intervention may prejudice theoutcome. We have developed a non-invasive technique, using dynamic hepaticscintigraphy after a bolus injection of99mTc sulphur colloid, which is able toassess the portal component of totalhepatic blood flow. The ability of thismethod to detect reduction in mesentericflow following progressive mesentericischaemia was examined.

Thirty-five Wistar rats were studied infive groups of seven animals: group 1,control; group 2, sham operation; group 3,ligation of ileocolic artery proximal to theterminal caecal branch; group.4, ligation ofsuperior mesenteric artery distal to theorigin of the middle colic artery; group 5,ligation of superior mesenteric arteryproximal to the origin of the middle colicartery.

Mesenteric blood flow in the controlanimals was 72% ± SD 3-2 of total hepaticflow and this value did not differ signifi-cantly in the sham operated group.Mesenteric flow fell to 58% ± SD 3-8 ingroup 3 (p<0.01). A further fall to52% ± SD 6-8 was seen in group 4. Thereduction in flow was greatest in group 5animals, 32% ± SD 8-7 (group 4 vs group5 p<0.01). Total liver blood flow was alsoprogressively reduced from groups 2 to 5.As this imaging technique can be used

clinically, the results of this study suggestthat it may be of value in the diagnosis andassessment of patients with mesentericischaemia.

FIOConservative treatment of fissure-in-ano

M J GOUGH AND A A M LEWIS The RoyalFree Hospital, London) Posterior fissure-in-ano (PFIA) is an exquisitely painfulcondition for which conservative treatmentwith a topical local anaesthetic and a StMark's Hospital anal dilator is widely used.The rationale for using the dilator, whichmay be both painful and distasteful to thepatient, is difficult to understand. The

value of such therapy has therefore beenassessed in a randomised clinical trial of 82patients with a PFIA. Within one monthhealing occurred in 17/39 patients (43.6%)treated with lignocaine gel alone comparedwith 18/43 patients (41.9%) who also usedthe dilator. The mean duration ofsymptoms was the same in both groups.The success of both methods of treatmentin PFIA present for less than three months(50%) was no different from that in fissuresof longer standing (40%, X2=0-6458). Asthis trend was reversed if the results werereconsidered with a six month cut-off,there is no evidence to suggest that con-servative therapy for PFIA should bereserved for fissures of short duration.On the basis of these results it is

suggested that the use of an anal dilator inthe management of PFIA is abandoned infavour of treatment with anaesthetic gelalone, and early surgical intervention ifnecessary.

In 12 of 15 patients with an anteriorfissure-in-ano there was an obvious pre-disposing factor. Resolution of underlyinganal pathology was presumably responsiblefor the greater success (73.3% healed) of asimilar period of conservative treatment inthese patients.

FllAnal manometry in neuropathic faecalincontinence and complete rectal prolapse

M PESCATORI, G G P BROWNING, AND A GPARKS (Surgical Research Department, StMark's Hospital, London) Previous worksuggests that in patients with pelvic floorneuropathy combined faecal incontinenceand complete rectal prolapse is associatedwith more severe sphincter deficiency thanincontinence alone.Seventy-two patients with electro-

physiological evidence of primary neuro-pathic faecal incontinence were studied.Group 1 consisted of 33 patients (46%)with faecal incontinence accompanied bycomplete rectal prolapse; group 2comprised the remaining 39 patients (54%)with faecal incontinence alone. Anal canalpressures and rectal capacity weremeasured and compared in these twogroups of patients.Mean anal canal length did not differ

significantly; 2.4±0-6 cm in group 1 and3 1±1i1 cm in group 2 (mean ± SD, x2-test, p NS). Anal pressures were less ingroup 1; mean resting pressure was 31±19

cm water in group 1 and 43±23 cm water ingroup 2 (p<005) and mean squeezepressure 24±17 cm water and 41±29 waterrespectively (p<002). Mean rectalcapacity in group 1 was 155±71 ml air andin group 2 200±49 ml air; this differencewas not significant (p NS).These results confirm previous reports of

greater anal sphincter deficiencies whenneuropathic faecal incontinence isaccompanied by complete rectal prolaspe.Further studies are required to quantifythis by electrophysiological means.

F12Results of delayed anal sphincter musclerepair for trauma

G G P BROWNING AND A G PARKS (SurgicalResearch Department, St Mark's Hospital,London) Ninety-two patients, 38 menand 54 women, have been treated bydelayed anal sphincter muscle repair fortraumatic sphincter division. The aetiologywas operative injury in 53 (M21, F32),trauma in 26 (M17, F9), and obstetricinjury in 13 patients. Sixty (65%) patientspresented with frank faecal incontinence of8 0±3-0 years (mean ± SEM) duration inthe men and 11-5±1-6 years in the women(p<005). The remaining 32 (35%)patients had stomas performed at the timeof injury 2 6±0-5 years before repair.

In all cases the sphincter muscle ring wasrepaired by an overlapping techniqueprotected by a temporary loop colostomy;four patients had combined sphinctermuscle repair and post anal repair. Therewas no postoperative death. Complicationsoccurred in 25 patients and included sinus,12; stricture, eight; fistula, three; andnecrosis, two. Two of these patientsrequired revision of the repair within onemonth.

Closure of colostomy has been carriedout in 79 patients; 10 await closure.Function has been assessed in 73 fromthree to 114 months (mean 30-1 months)afterwards. Sixty-two (86%) patientsregained full continence apart from minorproblems with flatus and the remaining 11(14%) have normal control provided thestool is solid. The repair failed on threeoccasions; one patient had a successfulsecond repair but two required permanentcolostomies.The operation is straightforward, has a

low morbidity and mortality, and producesgood functional results.

A902


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/


F13Results of colectomy for slow-transitconstipation

hospital orwhole. Therbetween thetumour, its

D M PRESTON, P R HAWLEY, J E LENNARD- logical graadJONES, AND I P TODD (St Mark's Hospital, the anal ver

London) From 1969-82 96 adults with higher ones

chronic constipation were treated by bowel of recurrenresection; 43 had Hirschsprung's disease indication o

and 35 idiopathic megacolon. We report The highestthe results in 18 patients with an apparently was in thenormal colon but greatly reduced whole tumours ingut transit rate. All were female, with equally fixtmean age at operation 25 years. They were positive lincapacitated by abdominal symptoms; quadrants irthree had no spontaneous bowel actions of the trialvand the interval between bowel actions in cases: yet cthe others ranged from seven to 35 days. UK were m

Operation was advised as a last resort after the tumoufailure of medical treatment. Thirteen were significancetreated by subtotal colectomy (seven curative rescaecorectal and six ileorectal anastomosis), of 401 mobitwo by sigmoid colectomy, two by left with 36% fhemicolectomy, and one by Duhamel's survival ralprocedure. There were no deaths or stages A, Eoperative complications. Over a mean high survivfollow-up of 4 5 years (range six months to cases. The11 years), six have experienced episodes of closely relatsmall bowel obstruction, three of whom three yearsneeded laparotomy (all IRA). Dukes's A,No patients treated by sigmoid eight per ce

colectomy or left hemicolectomy was low grade timproved. Of the 13 patients who had disease at tisubtotal colectomy, the result was satis- of patientsfactory in 10. Constipation was relieved inseven (bowels open 31xdaily) and theother three were improved, though are stillconstipated and need laxatives. Threepatients were not helped; two still cannot GENERAL IIdefaecate and one has diarrhoea probably F15-F28due to laxative addiction.

Subtotal colectomy, but not partial F15colectomy, with removal of an apparently Jejunal molnormal colon, can benefit women with abdominalsevere intractable constipation.

F14Prognostic factors in operable rectalcancer, as seen in the United Kingdom

A N SMITH, L S FREEDMAN, AND W

DUNCAN (Departments of Surgery andRadiation Oncology, University of Edin-burgh, and MRC Cancer Trials Office,Cambridge) The opportunity arose ofstudying 824 cases of operable rectal cancer

entered into an MRC trial of irradiation inoperable cancer of the rectum. The cases

had been contributed from 17 centreswidely scattered in the UK, thus giving a

picture, not of the disease in any one

region, but in the country as are were significant relationshipse prognosis and the fixity of the;Dukes's stage and the histo-le. Tumours less than 8 cm fromrge had a poorer prognosis than*. There was a greater frequencyice in older patients but no)f a sex difference in prognosis.proportion of mobile tumoursDukes's A group and of fixedthe Dukes's B, with Dukes's Ced or mobile. There was arelationship between thenvolved and fixity. The protocolwas designed to recruit operableonly 49% of such cases in theiobile tumours. The mobility ofJr was of great prognostic: 80% of mobile tumours had;ections. The three year survivaltile tumours was 58% comparedfor 364 tethered tumours. Thete at three years in Dukes's3, and C showed the expectedal in the more favourable Alocal disease-free rate was alsoLted to Dukes's staging being ats, 81%, 65%, and 51% forB, and C respectively. Fifty-

ent of patients with histologicaltumours were alive and free of:hree years compared with 24%with high grade lesions.

tility in patients with functionalpain

J KINGHAM, R BOWN, ELSPETH BELLHOUSE,AND A M DAWSON (GastroenterologyDepartment, St Bartholomew's Hospital,London) There is now accumulatingevidence to suggest a disturbed colonicmotility pattern in patients with functionalabdominal pain. In addition preliminaryobservations with a single radiotelemetrypressure-sensitive capsule have suggestedthat there may be a disturbance of smallgut motor pattern in the irritable bowelsyndrome. Using paired, tethered, radio-pills 20 cm apart we have studied jejunalmotor activity in six patients presentingwith abdominal pain in the irritable bowelsyndrome and 10 matched controls. Thepills were screened into the jejunum with

the proximal pill at the duodenojejunalflexure and fasting, fed, and sleep record-ings obtained for up to 30 hours.We recorded the duration of proximal

phase III complexes (5.39±2.03 min fornormal subjects, 5-33±1-83 min for thosewith the irritable bowel syndrome), theduration of distal phase III complexes(5.70±2.62 min for normal subjects,5-59±2-15 min for those with the irritablebowel syndrome), the speed of propaga-tion (3.15±3.44 min for normal subjects,4-09±5-82 min for those with the irritablebowel syndrome), the frequency ofmigrating motor complexes, (96.54±49.94min for normal subjects, 77-64±51-63 minfor those with irritable bowel syndrome).In both groups 7% of phase III activityappeared only at the proximal radio-pillwith 26% in both groups appearing only atthe distal pill. Two migrating motorcomplexes exhibited retrogradepropagation. Feeding usually, but notinvariably, delayed the first migratingmotor complexes in both groups but afterfood the return of phase III activity wasoften seen only at one locus. During sleeptype II activity was reduced but phase IIIactivity remained unchanged in bothgroups. There was no correlation betweeneither phase II or III motor activity andabdominal pain in the irritable bowelgroup.These studies demonstrated a marked

variability in the pattern of phase IIIactivity (migrating motor complexes)within and between normal subjects and asimilar variability in patients withfunctional abdominal pain. No differencecould be elicited between the two groupscontrary to previous reports.

F16Antral gastrin cell hyperplasia

N A L0VGREEN, P DELIKARIS, J POULSEN, L I

LARSSON, AND E AMDRUP (Surgical Gastro-enterological Department L, Aarhus Kom-munehospital, Aarhus, Denmark) Hyper-gastrinemia due to gastrinoma is a well-established clinical entity, but non-tumerous hypergastrinaemia of antralorigin, associated with duodenal ulcerationhas not yet been universally accepted.

This paper describes three cases of ulcerrecurrence where antral G-cell hyperplasiahas been adequately documented, byimmunocytochemical counts of gastrin cellsin the resected antrum. The patients hadpreoperatively basal gastrin levels in pg/mlA 180, B 170, C 92 respectively and after

A903


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/


food response A>800, B>800, C 390.Se-gastrin levels were similar aftervagotomy and were eliminated afterprecise antrectomy. In all patients gastrincells were found three to four times moreper field of vision than normal.As a conclusion we believe that antral

G-cell hyperplasia is a definite clinicalentity, characterised by recurrent pepticulceration, hypertrophic gastric mucosalfolds, and postprandial hypergastrinaemia,not necessarily associated with very highgastric acidity. We also believe that thesepatients can be identified at present bymeans of food-stimulated gastrin secretorystudies.

F17Smoking and peptic ulceration: anoutpatient endoscopic survey

C C AINLEY, L C FORGACS, P W N KEELING,AND R P H THOMPSON (The GastrointestinalLaboratory, The Rayne Institute, St Thoms'Hospital, London) The relationship ofsmoking to peptic ulceration is contro-versial. Epidemiological surveys suggest anassociation, but these have rarely beenbased on endoscopy. We have thereforestudied the smoking histories of patientsattending for outpatient endoscopy.The endoscopy unit at St Thomas'

Hospital provides a service open to allhospital doctors and local generalpractitioners. Between April 1981 andMarch 1982 all patients attending for out-patient endoscopy completed a question-naire on their smoking history and wereinterviewed; 1217 patients over the age of18 were studied 42 with carcinoma of theoesophagus or stomach were excluded.Three hundred and four had never smokedregularly, 241 were ex-smokers, and 594smokers. There was no difference in theincidence of benign gastric ulcer betweennever smoked regularly and ex-smokers,but there was a significantly increasedincidence in smokers compared with bothnever smoked regularly (2= 15-5,p<0001) and ex-smokers (x2=5.4,p<002). Similarly, in duodenal ulcerationthere was no difference between neversmoked regularly and ex-smokers, butthere was an increased incidence insmokers compared with never smokedregularly (x2=12-2, p<0.001) and ex-smokers (Q2=69, p<0.01). In addition,there was an increased incidence of bothgastric ulcer and duodenal ulceration withincreasing cigarette consumption.

In conclusion, in patients having out-patient endoscopy there is an associationbetween current smoking and both gastricand duodenal ulcers.

F18Effects of smoking on gastric secretion: lossof inhibition by antisecretory drugs

E J S BOYD, J A WILSON, AND K GWORMSLEY (Department of Therapeutics,Ninewells Hospital, Dundee) Smokingadversely affects healing and recurrencerates in patients with duodenal ulcers. Weinvestigated the effects of smoking onpentagastrin-stimulated and overnightbasal gastric secretion in patients receivingantisecretory drugs. Six habitual smokersunderwent three tests. After an overnightfast gastric secretion was stimulated by asubmaximal dose of pentagastrin (0.5 mg/kg/h). Gastric juice was collected for 6x 15minute periods, volumes noted, andaliquots assayed for pH, H', and pepsin.On two of the days the subjects receivedcimetidine 200 mg orally 60 minutes beforegastric intubation. On one of the days onwhich cimetidine was given subjectssmoked cigarettes at a rate they foundcomfortable. Four of the subjects also hadtwo further studies in which poldine 4 mgwas given 60 minutes before the study.Eleven patients being treated for duodenalulcer had two tests on separate days. Theytook their evening dose of antisecretorydrug (ranitidine 150 mg in five, cimetidine400 mg in three, and LM24056 200 mg inthree) at 18-00 h and gastric juice wascollected from 20-00 h until 08-00 the nextmorning. On one of the nights they wereallowed to smoke cigarettes ad libitum.Inhibition of pentagastrin-stimulatedsecretion was not affected by smoking (acid(X±SEM) mmol/90 min - control 43±7,cimetidine 20+3, cimetidine + cigarettes18±4, poldine 34±9, poldine + cigarettes35±6; pepsin mg/90 minutes - control208 ± 36, cimetidine 93 ± 1 5,cimetidine + cigarettes 91±17, poldine148±31, poldine + cigarettes 156±46).Ten of the 11 patients showed increases ingastric acid and pepsin secretion onsmoking during the overnight studies. Acidincreased by a mean of 92% (from 25±8mmol/12 h to 48±7 mmol/12 h p<0.02) andpepsin by 59% (from 446±89 mg/12 h to709±63 mg/12 h p<005).Smoking can markedly decrease the

inhibitory effects of antisecretory drugs onbasal gastric secretion. This may contribute

to the poor therapeutic response in patientswith duodenal ulcer who smoke.

F19Luminal pancreatic enzymes: aspects oftheir fate and function

E J S BOYD, J G R CUMMING, A CUSCHIERI, ANDK G WORMSLEY (Departments of Thera-peutics and Surgery, Ninewells Hospital,Dundee) It has been proposed that pan-creatic enzymes in the upper smallintestinal lumen can modulate pancreaticsecretion by inhibition of CCK release, andthat up to 92% of luminal pancreaticenzymes may be reabsorbed intact andrecirculated to the pancreas as part of anenteropancreatic conservation mechanism.We studied changes in serum immuno-reactive trypsin after intraduodenal loadsof endogenous trypsin, and, also, theeffects of oral pancreatic enzyme supple-ments on pancreatic fistula output ofenzymes.

Eleven patients had complete collectionof duodenal contents during stimulation ofpancreatic secretion by secretin (1 CU kgh) and CCK (1 IDU kg h) for 45 minutes.Five patients had a repeat test on aseparate day in which duodenal contentswere sampled (5 ml/15 minutes) instead ofbeing completely aspirated. In six patientspancreatic juice collected during the 45minute period of stimulation was rapidlyreinfused into the duodenum at the end ofthe period of aspiration serum immuno-reactive trypsin was measured in all sub-jects before stimulation with secretin-CCKand after 20 and 45 minutes of the hormoneinfusion. Serum immunoreactive trypsinwas also measured after 60, 75, and 90minutes in those subjects in whomduodenal juice was reinfused.

Daily volumes, trypsin output and meantrypsin concentration were measured inthree patients with pancreatic fistulaeduring control periods and while receivingoral pancreatic enzyme supplements(Pancrex V, 2 capsules two hourly).Serum immunoreactive trypsin did not

change during stimulation with secretin-CCK, nor did it increase when anendogenous trypsin load was provided byavoiding aspiration or reinfusion.Recirculation of large amounts of trypsin istherefore unlikely. Evocative tests ofpancreatic function based on serumimmunoreactive trypsin are not likely to beof diagnostic value. Enzyme supplementsconsistently reduced mean trypsin concen-tration and trypsin output and there was a

A904


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/


rebound on stopping. This supports thetheory that luminal enzymes can inhibitpancreatic secretion. Pancreatic enzymesupplements are useful in the managementof pancreatic fistulae.

F20Tubeless pancreatic function testing withdual labelled cobalamin

J LEUNG, R FROST, H WATERS, J BRAGANZA,AND P B COlTON (Departments of Gastro-enterology, The Middlesex Hospital,London, and Royal Infirmary,Manchester) The fact that some patientswith chronic pancreatitis have abnormalSchilling tests led to proposals for a test ofpancreatic function based on pancreaticenzyme separation of R protein fromcobalamin before transfer to intrinsicfactor. We have assessed this system in 66subjects. The fasting patient swallows acapsule containing 0.125 ,ug intrinsic factor(IF) bound Co57 cyanocobalamin, 0-125 ,ug(R) protein bound Co58 cyanocobalamin,and 125 jig cobinamide, and is given anintramuscular flushing dose of cyano-cobalamin. The patient fasts for anothertwo hours. Urine is collected for 24 hours.An aliquot is counted, and the relativeabsorption of R-Co58 cobalamin and IF-Co57 cobalamin calculated.

In 10 normal subjects, the mean ratiowas 0-75 (95% reference range 0.52-0.99),16 non-pancreatic gastrointestinal diseases0-79 (0.48-1.09), 21 chronic pancreatitis(CP) 0*56 (0-1.11), eight pancreaticcancers (CA) 0.43 (0-.93), and 11recurrent acute pancreatitis 0-84 (0.63-1.03). The sensitivity for chronic pancreaticdiseases (CP + CA) was 48% and thespecificity for non-pancreatic disease was87.5%. In recurrent acute pancreatitis, alltests were normal. There were 13 patientswith steatorrhoea; two with small bowelcauses had results within the normal range,whereas seven (three CA + four CP) ofthe 11 with pancreatic steatorrhoea hadabnormal results. Abnormal results werefound in only four of 12 patients withpancreatic calcification. The test is simpleto perform, and irradiation dose is low (0.5mR-50 mR). Further assessment isjustified.

F21Atropine suppression of pancreaticpolypeptide (PP) completely distinguishestumour PP production from other rises

T E ADRIAN, S M WOOD, AND S R

BLOOM (Department of Medicine, RoyalPostgraduate Medical School,London) To date there have been fewgenuine routine clinical applications of themeasurement of gut hormones. We haverun a weekly pancreatic endocrine tumourscreen for several years. A commonproblem is moderate rise in plasma PPconcentrations which may be due to anapudoma or, alternatively, to a variety ofother causes - for example, diabetes, renalimpairment, infections, stress, chronicalcohol abuse, or old age. As normally PPrelease is dependent on the cholinergicinnervation, it has been suggested thatatropine administration should distinguishbetween autonomous PP secretion bytumours and normal PP release, which isalways suppressible by atropine. The effectof atropine (1 mg, intramuscularly) onplasma PP concentrations was investigatedin 12 patients with producing pancreaticendocrine tumours (two insulinomas, threegastrinomas, one glucagonoma, oneVIPoma, and five PPomas). After atropinethere was no significant change incirculating PP concentrations (medianbasal PP level 2050 pmol/l, range 440-30,000). Thus at 15 minutes PP was107±4% of basal (mean±SEM) at 30minutes 99±4%, and at 60 minutes102±6%. In contrast, all subjects incontrol groups showed a large fall by 30minutes (p<0005). These findings suggestthat tumour-produced PP is unaffected bycholinergic blockade. The 'atropine test'with two key blood samples (zero and 30minutes) may greatly help to detect early,and treatable, tumours in those patientswho have an intermediate rise of plasmaPP.

F22Neodymium YAG laser photocoagulationfor major acute -upper gastrointestinalhaemorrhage

I A MACLEOD, P R MILLS, J F MACKENZIE, R I

RUSSELL, AND D C CARTER (UniversityDepartment of Surgery and Gastro-enterology Unit, Royal Infirmary,Glasgow) A prospective single blindcontrolled study was started in October1980 to evaluate the neodymium YAGlaser in the management of majorhaemorrhage from peptic ulcers and singlevessels. Entry was restricted to patientswith major blood loss as judged by clinical(shock, haemoglobin <10 g/ml, and needfor blood transfusion) and endoscopic(arteries or red and black spots in ulcer

base) criteria. Patients were assignedrandomly to laser or sham therapy and theincidence of further haemorrhage and needfor emergency surgery assessed byindependent observers.

Six hundred and fifty-seven patientswere admitted with acute non-varicealupper gastrointestinal bleeding of whom184 were bleeding from peptic ulcers orsingle vessels; 130 of these failed to fulfilthe entry criteria and all settled onconservative management. Fifty-fourpatients were eligible for the study and 45were included. Twenty patients bled fromarteries: only one of eight who receivedlaser therapy required emergency surgeryand survived, whereas all eight receivingsham therapy needed emergency surgeryand two died. Four patients allocated tolaser therapy did not receive it for technicalreasons. Twenty-five patients werebleeding from red/black spots and allsettled irrespective of therapy.A group of patients at high risk of

returrent haemorrhage can be identified.Laser therapy reduces the incidence offurther haemorrhage and need foremergency surgery in patients bleedingfrom arteries.

F23N-nitrosocompounds after operation forduodenal ulcer

M R B KEIGHLEY, D MORRIS, V POXON, DYOUNGS, D W BURDON, T J MUSCROFT, JBARNARD, P M G BAVIN, R W BRIMBLECOMBE,D W DARKIN, P J MOORE, AND N VINEY (TheGeneral Hospital, Birmingham, and SmithKline & French Research Limited, Welwyn,Herts) There is some evidence that thereis an increased risk of gastric carcinomaafter resection or vagotomy for duodenalulcer. N-nitrosocompounds are powerfulcarcinogens which may be formed in thehypochlorhydric stomach. Analysis ofnitrosamines from fasting gastric juice werereported to be high after partialgastrectomy but not after vagotomy.We have analysed half-hourly gastric

samples for pH, nitrite, total, and stableN-nitrosocompounds from four groupsover 24 hours taking a standard diet:proximal gastric vagotomy (PGV n=7),truncal vagotomy and pyloroplasty (TVPn=7), truncal vagotomy and antrectomy(TVA n=8), and controls (n=8). Statisticalcomparison of groups was performed bycalculating the area under the curve perhour (auc/h) for individuals by analysis ofvariance.

A905


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/


pH was significantly higher after TVA(auc/h=4-67) than after TVP (auc/h=2-52), PGV (auc/h=2.94), and control(auc/h=2.21) (p<0.001). Nitrite valuesincreased with pH and were significantlyhigher after TVA than control (p<0-01).By contrast, neither total nor stable N-nitrosocompounds increased with pH andthere was no significant difference in totalN-nitrosocompounds between the groups.

If gastric carcinoma is more commonafter operation for duodenal ulcer it isunlikely to be due to intragastricnitrosation.

F24New radioimmunoassay for cholecystokinin(CCK): relationship between post-prandialCCK levels and gall bladder (GB)contraction

I C FORGACS, M N MAISEY, G M MURPHY, ANDR H DOWLING (Department of Medicine,Guy's Hospital and Medical School.London) Although the pharmacologicaleffects of cholecystokinin are wellestablished, little is known about thephysiological role of circulating cholecys-tokinin in contracting the gall bladder -mainly because its RIA is notoriouslydifficult, as antisera to cholecystokininoften lack specificity and conventionalradiolabelling reduces cholecystokininimmunoreactivity. We therefore used aniodination technique, which avoidsoxidative damage to the peptide, andraised several anti-cholecystokininantisera, one of which cross-reactedequally with CCK8 and CCK33 but not withgastrin, and in six controls and sevenpatients with radiolucent gall stonesmeasured fasting and post-prandial (Lundhmeal) plasma cholecystokinin levels duringstudies of gall bladder emptying assessedboth by 99Tc-HIDA scanning and real-timeultrasound.

Results showed that fasting cholecys-tokinin levels were below the lower limit ofassay detectiocn (2 pmol/l) but 15 minutesafter the meal, rose to 33-8±SEM 5-8 incontrols and 37-2±5-3 in the patients, the0-60 minute post-prandial profiles beingcomparable in both groups. Gall bladdercontraction, however, was markedlydifferent, the ti emptying (HIDA scan)being 10-2±1-9 minutes in controls and21-7±3-1 in gall stone patients. A similarpattern of results was seen with ultrasound,the reduction in gall bladder volume afterthe meal being greater in controls(14.8±2-5 ml) than in gall stone patients

(9-1±1-8; p<0.05).To summarise: (1) a sensitive and

specific RIA for plasma cholecystokininhas been developed; (2) post-prandialcholecystokinin release was comparable ingall stone patients and in controls; while(3) gall bladder emptying, studied by twoseparate methods, was reduced in gallstone patients suggesting that gall bladderend-organ response to endogenous chole-cystokinin is altered in gall stone patients.

F25Is percutaneous functional cholecystectomyfeasible?: an experimental study in therabbit

T V TAYLOR, D FRENCH, AND ICAPPERAULD (Manchester Royal Infirmaryand Ethicon Research LaboratoriesEdinburgh) For the elderly, frail, unfitpatient with gall stones, drug therapy takesmonths to become effective, is expensive,and on stopping treatment the stonesrapidly recur. The present study was per-formed to assess the feasibility of defunc-tioning the gall bladder by occluding itslumen with Ethibloc. On hydration thissubstance undergoes rapid physical changefrom a viscous liquid to a solid similar tosponge rubber'.When injected subcutaneously into the

rat Ethibloc became encapsulated with noeffect on the bolus. Twenty-five rabbitsunderwent aspiration of their gall bladdercontents, which were replaced by Ethibloc.Groups of five animals were killed atseven, 14, 28, and 182 days. Controlanimals underwent cystic duct ligation.Up to 28 days from the time of injection,

Ethibloc remained in the gall bladderunchanged, occluding the lumen and pro-ducing minimal inflammatory change. Atsix months partial resorption of the sub-stance had occurred in some animals. Nogall stones or cholesterol depositsdeveloped in the gall bladder and none ofthe substance could be detected in thecommon bile duct. No infection occurredand liver function tests remained normalthroughout the study. The implications ofthese findings are discussed in the light ofperforming percutaneous 'functionalcholecystectomy' under ultrasoundcontrol.

F26Retained common bile duct stones: mono-octanoin or endoscopic sphincterotomy?

J DAWSON AND R COCKEL (Selly Oak

Hospital, Birmingham) The increasingrecognition of retained common bile ductstones as a cause of post-cholecystectomysymptoms necessitates constant evaluationof the optimum mode of therapy. Endo-scopic sphincterotomy is now regarded bymany as the treatment of choice but carriesa definite morbidity and mortality. Mono-octanoin has recently been reported tosuccessfully dissolve stones in selectedpatients but has not been evaluated inconjunction with sphincterotomy. In thisstudy we have evaluated the alternatives ofendoscopic sphincterotomy and mon-octanoin infusion in 41 consecutive patientswith retained stones presenting to aDistrict Hospital.Mono-octanoin infusion completely

dissolved stones in seven of 12 patientswith T-tubes in situ. In five patients withlarge stones mono-octanoin, infusedthrough a nasobiliary cannula, dissolvedstones in two and markedly reduced thesize of stone in a further patient. In 29 of 32patients with residual stones completeremoval was achieved by endoscopicsphincterotomy. There were no complica-tions of mono-octanoin therapy, but endo-scopic sphincterotomy was complicated byhaemorrhage in two patients andpancreatitis in one.Thus mono-octanoin is a safe first line

agent for dissolving retained stones inselected patients, while the remainder canbe effectively treated by endoscopicsphincterotomy without recourse tosurgery.

F27Gall stones and mortality: a post mortemstudy of gall stone-related deaths

T BATES, P J GODFREY, M HARRISON, M B KING,AND N R PADLEY (William HarveyHospital, Ashford, Kent) In a prospectivestudy of gall stones, all post mortemreports and death certificates issued in aHealth District were scanned over a 22 yearperiod. The prevalence of gall stones atnecropsy was 17% but of these subjectsonly 11% had had a cholecystectomy.

During the period of study, 391 chole-cystectomies were carried out in theDistrict of which 73% were in women.There were 10 deaths in the necropsy seriesof 1701, directly related to gall stone-disease, of which five were postoperative(1.3% operative mortality). Eleven furtherdeaths were considered to have been dueto gall stones in the 6356 subjects who didnot have a necropsy. Choledocholithiasis

A906


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/


was found at necropsy in 19 of those withgall stones (6.5%) and was responsible fortwo deaths.

Gall stones have been reported as havinga negative correlation with coronary heartdisease and a positive one with certainforms of cancer. The age-sex correctedprevalence of coronary heart disease andcancer, however, was similar in thosesubjects with gall stones and those without.Carcinoma of the gall bladder was noted inone woman with gall stones but carcinomaof the caecum was not found in any of thepost-cholecystectomy subjects.

It is concluded that, although gall stonesare common, most remain unoperated andthe mortality is low. The relationship toheart disease and cancer is still uncertain.

F28Inhibition of gastric acid secretion inZollinger-Ellison syndrome by omeprazole,a potent and long-acting antisecretory drug

C B H W LAMERS AND J B M J

JANSEN (Divison of Gastroenterology, StRadboud Hospital, Nijmegen, TheNetherlands) The efficacy of histamineH2-receptor blocking agents in patientswith Zollinger-Ellison syndrome is limitedby the short duration of acid inhibition. Ithas been shown that omeprazole (Hassle,Sweden), a substituted benzimidazole,inhibits pentagastrin stimulated gastric acidsecretion in normal subjects for at least 48hours. We therefore studied the effect of a

single oral dose of 80 mg omeprazole ongastric acid secretion in six patients withZollinger-Ellison syndrome (four men, twowomen, aged 31-51 years). All had hyper-gastrinaemia (serum gastrin 440-6475 pg/ml) with a positive secretin provocationtest and gastric acid hypersecretion (BAO13-3-90-8 mmol/h). In the second hourafter ingestion of the drug gastric acidsecretion was inhibited by 98±2%. Fivepatients were achlorhydric, while onepatient, having the lowest plasmaomeprazole concentrations, had 90%inhibition of gastric acid. At 24 hours afteringestion of the drug gastric acid was

inhibited by 83±6% (56-100%). At thattime omeprazole was undetectable inplasma from four patients, while theomeprazole concentrations were very lowin the two other patients with Zollinger-Ellison syndrome. Clinical symptoms(diarrhoea in five, gastric discomfort intwo, heartburn in two) were completelyabolished for at least 48 hours. Serumgastrin concentrations were not influenced

by omelaboratorthe drug.We co

and Ionpatients i

LIVERF29-F42

iprazole. No side-effects orry abnormalities were induced by

nclude that omeprazole is a potentg-acting antisecretory drug inwith Zollinger-Ellison syndrome.

F29Identification of alcohol dehydrogenase(ADH) and aldehyde dehydrogenase(ALDH) isoenzymes in patients withalcoholic liver disease

J B SAUNDERS, B R RICCIARDI, D A HOPKINSON,AND R WILLIAMS (The Liver Unit, King'sCollege Hospital and Medical School,London) Variations in the isoenzymecomposition of ADH and ALDH underlycertain acute effects of alcohol such as

facial flushing. The relationship betweenthe isoenzyme components and the chronictoxic effects of alcohol has not yet beenestablished. We have developed a methodfor simultaneous analysis of ADH andALDH isoenzymes by starch gel electro-phoresis in 2-5 mg liver biopsy samples,and have investigated their phenotypes andrelative activities in 65 patients withalcoholic liver disease (28 with fatty liver,12 with alcoholic hepatitis, and 25 withcirrhosis).

Isoenzymes coded by the four ADH lociwere identified in all biopsy samples. Thephenotype frequencies of the isoenzymescoded by ADH2 and ADH3 loci were

similar in the three histological groups andalso comparable with those in the generalpopulation. Overall ADH activity and therelative activities of the individual iso-enzymes were also similar in the threegroups.Up to five distinct isoenzymes of ALDH

were seen. In all patients the fast-migratingmitochondrial band was present, indicatingthe 'usual' ALDH phenotype. In patientswith alcoholic cirrhosis there was a

selective reduction in staining intensity ofthis component (mean score 2-6 comparedwith 3-7 and 4-0 for patients with fatty liverand alcoholic hepatitis respectively,p<0-025). Other components were rela-tively weaker and there were no significantdifferences in mean scores betweenpatients in the three histological groups.Selective reduction in the mitochondrialcomponent of ALDH in alcoholic

cirrhotics, by limiting oxidation of acet-aldehyde, could be an important factorpotentiating liver damage.

F30Hepatoma-specific alkaline phosphatase (H-ALP) in high and low incidence areas ofprimary liver cancer

P KAY, P J JOHNSON, T W WARNES, RWILLIAMS, M LONGSON, I LAING, A SMITH,AND N W LEE (University Department ofGastroenterology, Manchester RoyalInfirmary, Manchester; The Liver Unit,King's College Hospital, London; and theDepartment of Surgery, Queen MaryHospital, Hong Kong) We wished tocompare the diagnostic value of H-ALPand a-fetoprotein (AFP) in primary livercancer in Hong Kong, an area of highincidence with high carriage rates of thehepatitis B virus (HBV) and in the UnitedKingdom, an area of low incidence withlow carriage rates of HBV.

Forty-four Chinese and 87 Caucasianpatients with primary liver cancer werestudied. ALP isoenzymes were demon-strated by polyacrylamide gel electro-phoresis. AFP was measured by radio-immunoassay and HBsAg by passivehaemagglutination inhibition and radio-immunoassay.

Results showed that, although theincidence of raised AFP was the same inthe two groups, the serum AFP levels weremuch higher in the Chinese than in theCaucasians (1.4x106 ng/ml compared with7-8 x104 ng/ml, p<0001). The incidence ofthe HBsAg was 74% in the Chinese and18% in the Caucasians (p<0-001). Theincidence of H-ALP in the Chinese wasonly 2% compared with 18% in Caucasians(p<0-05). In the Caucasian group, patientswith cirrhosis had a higher incidence ofraised AFP than non-cirrhotics (p<0-0005)and H-ALP was found in 29% of non-cirrhotics compared with only 10% ofcirrhotics (p<0-05).We conclude that Chinese with primary

liver cancer have high levels of AFP andHBsAg but a low incidence of H-ALP(2%). In Caucasians, the levels of AFP andincidence of HBsAg are lower but H-ALPis much more common (18%). Thus, thevarying incidence of tumour markers mayreflect biological differences in hcpatomafound in high and low incidence areas ofthe world. H-ALP is of particular value inthe diagnosis of primary liver cancer inCaucasian non-cirrhotics, in whom AFPlevels are often normal.

A907


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/


F31Prevention of perinatal transmission ofhepatitis B by selective passiveimmunisation

R J STOCKS, A E FLOWER, P NUlTALL, AND M S

TANNER (introduced by Professor A SMcNeish) (Department of Child Health andPHLS, Leicester Royal Infirmary andRegional Transfusion Centre,Sheffield) Chronic carriage of HBsAgconfers a high risk of chronic liver diseaseand hepatoma, and frequently results fromperinatal HBV infection from carriermothers. The efficacy of selective passiveimmunisation of babies at risk is evaluated.Antenatal HBsAg positive sera areexamined for HBeAg and anti-HBe.Babies of HBeAg positive mothers, and ofmothers with perinatal acute hepatitis B,receive hepatitis B immune globulin 250mg at birth, then monthly for six months.Anti-HBe positive mothers' babies arefollowed without HBIG. No attempt ismade to sterilise the vagina, to wash outthe baby's stomach, or to inhibit breastfeeding.Of 36 HBsAg-positive pregnancies, eight

had HBeAg (four from SE Asia). SevenHBIG-treated infants, together with twoHBIG-treated infants of mothers withacute hepatitis, were followed; none hasdeveloped HBsAg, and three havedeveloped anti-HBc or anti-HBs in thesecond year. Twenty-eight mothers hadanti-HBe, and none of their 18 babiesfollowed has become an HBsAg carrier;one infant developed transient antigen-aemia and raised transaminases at 3months of age, subsequently developinganti-HBs and anti-HBc. Among oldersiblings of study infants, five chroniccarriers of HBsAg and HBeAg weredetected. We conclude that (1) anti-HBepositive pregnancies do not result inHBsAg carrier infants, but transientantigenaemia and subsequent activeimmunity may occur; (2) HBIG givesshort-term protection in all and, bypermitting active immunity to develop,long-term protection in some babies ofHBeAg positive mothers. These datasupport the selective use of HBV vaccinewhen available.

F32Nuclear magnetic resonance (NMR)imaging of the liver

S H SAVERYMUTTU, H J F HODGSON, G M

BYDDER, AND R E STEINER (Departments ofMedicine and Radiology, Royal Post-graduate Medical School, HammersmithHospital, London) Nuclear magneticresonance scans of the liver were obtainedin 12 normal volunteers and 32 patientswith biopsy proven liver diseases and theresults compared with x-ray computertomography. In addition to the visualimages derived from nuclear magneticresonance imaging, tissue proton longi-tudinal relaxation times (T1) werecalculated in the different conditions.Both methods demonstrated focal liver

disease, but an extensive area of focalatrophy in one patient, and of hepaticinfarction in another were detectable onlyby nuclear magnetic resonance. In diffuseliver disease, it showed a prolongation ofT1 in cirrhosis but was not of diagnosticspecificity. The images in three patientswith cirrhosis were abnormal only with thenuclear magnetic resonance technique, andnot by computer tomography. In contrast,in steatosis computer tomography imageswere virtually diagnostic, while nuclearmagnetic resonance scans showed nospecific features.

Initial results of nuclear magneticresonance scanning as a non-invasivetechnique for investigating liver disease arepromising, but more experience is requiredto evaluate its potential.

F33Single photon emission tomography (SPET)in the evaluation of liver disease

A I MORRIS, M ROBERTS, M CRITCHLEY, AND J SGRIME (Gastro-Intestinal Unit, WaltonHospital, and Departments of Medicine andNuclear Medicine, Royal LiverpoolHospital, Liverpool) Conventionalisotope scans provide limited informationabout hepatic function because of failure toaccount for liver volume. Rather thansearching for mass lesions within the liver,we have developed a technique using SPETto measure hepatic volume and uptake ofradiopharmaceutical. Thirty patients withhistologically proven liver disease werescanned with a GE 400T tomographiccamera, 15 minutes after intravenousinjection of 4mCi9Tm sulphur colloid.Transverse sections of the liver were recon-structed using SPET software, apd livervolume derived using an edge detectionprogramme.The sensitivity in measuring isotope

activity and accuracy of volume estimations

were assessed using phantoms, and highdegrees of correlation were found (sensi-tivity r=0-98, volume r=0.99). Repro-ducibility of liver volume estimations was±3%. Patients with the most severe liverdisease (decompensated cirrhosis) hadlowest uptakes per injected dose of isotope19%±6 (SD); those with inactive cirrhosisor clinically moderate chronic liver diseasehad uptakes of 41%±5, while patients withminimal hepatic dysfunction, either withnormal biopsies or chronic persistenthepatitis, had the highest uptakes(57%±5).The use of 99T' sulphur colloid allowed

sufficient time for the development ofanalytical techniques without interferencefrom secretion into the gut. Despitespecific uptake by Kupffer cells in the liverthe results correlated well with clinical andhistological assessments. Studies of morespecific hepatocellular radiopharma-ceuticals - for example, BSP - are con-tinuing, as the method permits non-invasive longitudinal assessment of liverfunction.

F34Propranolol for prevention of recurrentvariceal bleeding in cirrhotic patients

R P WALT, A K BURROUGHS, A A DUNK, W JJENKINS, AND S SHERLOCK (Royal FreeHospital School of Medicine,London) We report a controlled study ofpropranolol for the prevention ofrebleeding in patients with cirrhosis.

Since June 1981 42 consecutive patientswith cirrhosis, who had been admitted withbleeding oesophageal varices, wererandomly allocated to propranolol orplacebo treatment after their condition hadstabilised. Propranolol was given to reducethe resting pulse by 25%. The two groupsdid not differ significantly in age, sex, causeof cirrhosis, or Child's grade.

Results showed that there was nosignificant difference in rebleeding,between the groups - seven in each (p0.90), all from oesophageal varices. Therewere three deaths, all from recurrentvariceal haemorrhage - two in thepropranolol group and one in the placebogroup.These results contrast with those of a

previous study of mainly alcoholic patientsin good condition, in which propranolol didreduce recurrent gastrointestinal bleeding.Our patients were unselected, so that boththe aetiology of the cirrhosis and the

A908


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/


severity of the liver damage varied, butthey were representative of the populationat risk of rebleeding. Propranolol not onlyfailed to prevent rebleeding, but the ,-blockade seriously complicated theresuscitation of two patients who bled.

F35Controlled trial of high and low doseD-pencillamine (DP) in primary biliarycirrhosis (PBC): results at three years

M F BASSENDINE, A F MACKLON, R MULCAHY,AND 0 F W JAMES (Department ofMedicine, University of Newcastle uponTyne, Newcastle upon Tyne) Fifty-nineconsecutive patients with primary biliarycirrhosis were randomly allocated to one ofthree groups - 19 received 1 g DP daily, 21received 250 mg DP daily, 19 controlsreceived no DP. The groups werecomparable with respect to age, symptoms,disease duration, LFTs and histologicalstage. Mean period of follow-up was 37months, minimum follow-up in survivorswas 29 months. Changes in bilirubin, AST,alkaline phosphatase, albumin, IgG, lgM,and antimitochondrial antibody titre wereanalysed at six-monthly intervals.The following achieved statistical signifi-

cance (p<0.05). High dose vs control: ASTlower at six, 12, and 30 months, IgM lowerat six, 12, and 30 months, IgG lower at 12,18, 24, and 30 months. Low dose vscontrol: AST lower at six and 12 months,igG lower at 12 months, alkalinephosphatase lower at 24 months. No othermeasurement showed significant change ineither group at any stage.

Mortality: high dose 2/19 (deaths at fiveand 20 months), low dose 8/21 (six liverrelated, four of whom received <18months DP), controls 2/19. Eleven of 12patients who died had cirrhosis on entry totrial. Side effects occurred in 14 (35%) -nine high dose, five low dose; theserequired withdrawal of drug in nine (23%).We conclude that (1) in high dose, while

DP produced a consistent improvement inAST and immunoglobulins, no significantchange occurred in bilirubin nor histo-logical stage, nor was mortality influenced;(2) in low dose transient improvement onlywas seen in AST and immunoglobulins. Itis possible that these limited changesindicate that higher dose DP over manyyears could benefit those patients able totolerate the drug.

F36Species differences in copper toxicosis andtheir relevance to Indian childhoodcirrhosis (ICC)

A KANTARJIAN, A GHOSH, AND M S TANNER(introduced by Professor A SMcNeish) (Department of Child Health,University of Leicester, Leicester) Cu-supplemented lambs and rats were studiedas animal models of Indian childhoodcirrhosis. Groups of five male lambsreceived 0, 1, 2-5, and 5 mg/kg/day Cu for13 weeks. Liver Cu in percutaneousbiopsies rose from 199±96 ,ug/g to 930±60,1810±340, and 1720±100 ,ug/g at fourweeks; 1770±220, 3200±450, and2980±320 gg/g at eight weeks. Lambsthrived initially, with normal serum bGT,copper + caeruloplasmin values, but ateight to 12 weeks rapidly deteriorated withjaundice, rising 6GT, haemolysis, andmethaemoglobinaemia. Rats differed infive respects: (1) pretreatment liver Cu waslower (27±14 ,tg/g); (2) a higher Cu dosewas required; 3 mg/kg/day was ineffective,but Cu supplementation of drinking waterto supply 20 and 40 mg/kg/day for 10 weeksincreased liver Cu to 316±230 ,sg/g and562±424 ,ug/g; (3) no haemolytic crisisoccurred; AST was raised to twice normalat four weeks and 4x normal at eightweeks; (4) whereas Cu-loaded lambbiopsies showed heavy rhodanine, Timm's,and orcein staining, rat biopsies withcomparable Cu concentrations measuredbiochemically were orcein negative; (5)ultrastructurally, lamb biopsies containednumerous electron-dense lysosomes,shown by electron probe analysis tocontain Cu and S, while rat lysosomes weremuch less electron-dense and containedscanty Cu and S.

Species differences in physiologicalhepatic Cu concentration, in its lysosomalaggregation, in orcein-demonstrable Cu-binding proteins, and in susceptibility toCu-toxicosis, are important constraints onextrapolating animal data to man.

F37Is intrahepatic cholestasis in sickle haemo-globinopathies a benign condition?

S I TERRY, B HANCHARD, AND G RSERJEANT (Departments of Medicine,Pathology and MRC Laboratories,University of the West Indies, Mona,Kingston 7, Jamaica, WI) Intrahepaticcholestasis in sickle haemoglobinopathies(IHCSS) is said to carry a high mortality.

Experience in 16 patients with IHCSSadmitted to the University Hospital of theWest Indies suggests a more benignprognosis. Fifteen had SS disease and oneSD Punjab (11 males and five females) withages ranging from 8 to 42 years (mean 20years). The diagnosis of IHCSS was basedon rapidly deepening jaundice, cholestaticfeatures, absence of large duct obstruction,and a liver biopsy with typical changes asdescribed below. Clinical features includeda prodrome with or without fever (12),dark urine (16), and soft, smooth hepato-megaly (13). None had evidence of liverfailure. Previous episodes of IHCSS hadoccurred in five. Serum bilirubin and liverenzymes were raised, whereas serumalbumin and prothrombin time werenormal. Haemoglobin levels and reticulo-cyte counts did not differ from the steadystate. Liver biopsy revealed normal liverarchitecture without hepatocellularnecrosis, intrahepatic bile pigment, andcongestion of the sinusoids by sicklederythrocytes in all cases. Ultrastructuralexamination of the liver, performed in 12,showed mitochondrial elongation orgranulation in seven. Resolution occurredin all, within two weeks in 15. Although theoutcome of IHCSS in our patientsappeared benign, monitoring of suchepisodes in hospital is consideredadvisable.

F38Prospective study of jaundice after cardio-pulmonary bypass surgery

J COLLINS, M F BASSENDINE, R FERNER, AMURRAY, AND 0 F W JAMES (Departmentsof Medicine and Medical Physics, FreemanHospital, Newcastle upon Tyne) Westudied 248 patients undergoing cardio-pulmonary bypass surgery to determine theincidence and possible causes of early'postpump' jaundice. Factors, suggested byprevious retrospective studies to beassociated with this type of jaundice, wereexamined prospectively; preoperativeLFTs, peroperative minimum meanarterial blood pressure (mMAP), PaO2,bypass time, oesophageal temperature,and type of operation. Alcohol intake,drugs and anaesthetics, blood transfused,postoperative blood pressure and PaO2were also recorded. LFTs and tests forhaemolysis were measured postoperativelyon alternate days. Postpump jaundice(bilirubin :50 jumol/l) occurred in 49patients (20%); this was detectable in 48/49by postoperative day 2, when the bilirubin

A909


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/


was over 75% conjugated; in 44/49bilirubin began to fall by postoperative day6; early postoperative ALT was normal.Hepatic alkaline phosphatase did not riseuntil after day 6, the rise often beingprolonged. Bilirubin was slightly raisedpostoperatively (18-49 umol/l) in a further140 subjects (56%). Fourteen of 248patients died; of these 12 were in thepostpump jaundice group (p<0.0005). Atnecropsy liver biopsy in seven showed onlychanges of chronic venous congestion.Postpump jaundice patients received more

blood (10±7.6 units vs 5.3±3.6, p<0-001),and had longer bypass times (103±35minutes vs 89±30, p<0.01) than thosewithout postpump jaundice. Postpumpjaundice developed after mitral valvesurgery in 32/93 patients (p<0.0005).Contrary to previous hypotheses there was

no association of postpump jaundice withmMAP, PaO2, or haemolysis. No otherfactors were related.We conclude that the incidence of

postpump jaundice is high (20%) andcarries a bad prognosis (12/49 died). It is anearly conjugated hyperbilirubinaemia notassociated with other evidence of liver celldamage or cholestasis, not explained byhypoxia, hypotension, or haemolysis.

F39Endoscopic retrograde cholangiopancreato-graphy (ERCP) in primary biliary cirrhosis

I HAMILTON, D J LINTOTT, W S J RUDDELL,AND A T R AXON (Gastroenterology Unitand Department of Diagnostic Radiology,General Infirmary at Leeds, Leeds) Wehave compared the endoscopic retrogradecholangiogram and pancreatogram of 22unselected consecutively investigatedpatients with primary biliary cirrhosis withthose of 33 patients with other forms ofintrahepatic disease ('control patients') and28 controls with normal pancreatic andbiliary duct systems and no evidence ofliver disease ('normal subjects').The pancreatic duct was opacified in 17

primary biliary cirrhosis patients and was

abnormal in only one; of the 29 controlpatients in whom a pancreatogram was

obtained there were changes of chronicpancreatitis in eight. These changes were

particularly frequent in seven patients withsclerosing cholangitis, of whom three hadchronic pancreatitis.The bile duct was opacified in 19 primary

biliary cirrhosis patients and 28 controlpatients. The intrahepatic cholangiogramshowed calibre variation and irregularity in

seven primary biliary cirrhosis patients, 15control patients, and one of the normalcontrols (p<005). There was a smooth,localised indentation of the commonhepatic duct at the porta hepatis in 11patients with primary biliary cirrhosis, andin none of the control patients or normalcontrols (p<0.05). It was not possible tocorrelate the presence of this indentationwith the histological stage of the disease,but it was present in nine of the 12 patientswith a normal intrahepatic cholangiogramand in only two of the seven with radio-logical features of cirrhosis (p<0-05). Inthree such patients undergoing laparotomyenlarged hilar lymph nodes were present,and may have been compressing the bileduct.

While the presence of this radiologicalabnormality has been described in a smallproportion of patients with primary biliarycirrhosis its frequency has not been pre-viously established. The presence of asmooth 'notch' on the extrahepaticcholangiogram is highly suggestive ofprimary biliary cirrhosis as distinct fromother forms of intrahepatic disease.

F40Gall bladder abnormalities in acutehepatitis: a prospective study

D P MAUDGAL, A E A JOSEPH, AND M H

WANSBROUGH-JONES (Departments ofCommunicable Diseases and NuclearMedicine, St George's Hospital andMedical School, London) Oedema andincreased thickness of the gall bladder wallfound in cholecystitis can now be preciselymeasured by ultrasonography. Patientswith acute hepatitis do not have any clinicalsigns of cholecystitis, but increased thick-ness of gall bladder wall has been reportedin some cases. The cause and progressionof thickness of gall bladder wall is notknown.We have compared gall bladder wall

thickness in 22 healthy volunteers and 20patients presenting with acute hepatitis dueto hepatitis A in nine; B in one; EpsteinBarr virus in two; and non-A non-B ineight patients. Gall bladder wall measured2-0±0-06 mm (mean ± SEM) in controlsand 5.35±0-46 mm in patients (p<0.001).On repeat measurements after clinical andbiochemical recovery from acute hepatitisin nine patients, gall bladder wall thicknessdecreased from 6-22±0-59 mm to2-44±0-29 mm (p<0.001). In sevenpatients the gall bladder contents gave anechogenic pattern suggestive of sludge,

which cleared during recovery phase.Serum alanine transaminase was raised upto 2932 IU/I (normal <35 IU/l) and serumalbumin ranged from 22 to 44 g/l (normalrange 35-50 g/l). Gall bladder wall thick-ness significantly correlated with serumtransaminase (r=0-65; p<0-001) andalbumin levels (r=-0.49; p<0.001).We conclude that local inflammation and

hypoalbuminaemia might contribute to theincreased thickness of gall bladder wallduring acute hepatitis, and gall bladderreturns to normal after complete recoveryfrom hepatitis.

F41Detergency of the bile acid (BA) pooldetermines biliary cholesterol (XOL)secretion and saturation in man

P LORIA, M BERTOLOMTI, R IORI, G TADDIA, M

PONZ DE LEON, AND N CARULLI (Istituto diClinica Medica II, Universita di Modena -

Divisione di Chirurgia, Ospedale di Castel-franco, Modena, Italy) It has beensuggested that cholesterol secretion intobile might be dependent on detergentpower of the bile acid pool. To test thishypothesis we investigated the effect of bileacid having different detergent capacity:deoxycholic acid (DCA), chenodeoxy-cholic (CDCA), cholic (CA), and urso-deoxycholic (UDCA), in decreasing orderon bile lipid secretion in man.

Six T-tube patients operated for chole-docolithiasis were studied. Each patientunderwent two studies: three patientsreceived CA and CDCA and three UDCAand DCA at three days' interval. At thebeginning of the study the enterohepaticcirculation was interrupted via an occlusiveballoon placed in the distal arm of theT-tube and bile was drained for five hours(pretreatment period); intraduodenalinfusion of bile acid was then started at therate of 1 g/h for five hours (replacementperiod). Bile samples were collected hourlythroughout the study. Parameters investi-gated included bile flow, bile lipid output,and bile acid composition.

During the pretreatment period bile acidoutput was the lowest and only primarybile acids were present. In the replacementperiod the administered bile acidconstituted 80-90% of the pool.We conclude that at low bile acid output

(pretreatment period), XOL/BA ratio andsaturation index are the highest. In the bileacid stimulated bile lipid secretion, XOLoutput seems to be significantly related tothe detergent power of the infused bile acid

A910


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/


(DCA > CDCA > CA > UDCA). Ourresults support the hypothesis that adeterminant of cholesterol secretion andsaturation index is the detergency of bileacid pool.

F42Effects of the vasoactive substancecyclandelate on hepatic HMGCoAreductase activity in rats and on biliarylipid composition in gall stone patients

B H WHIITEN, B MIDDLETON, A MIDDLETON, AMICIAK, D A WHITE, K W SOMERVILLE, AND GD BELL (Department of Therapeutics andDepartment of Biochemistry, University ofNottingham, Nottingham) Cyclandelate isa vasoactive substance (Cyclospasmol Gist-Brocades) consisting of the mandelic acidester of 3,3,5-trimethylcyclohexanol(TMC). TMC is structurally similar tomenthol (2-isopropyl, 5-methyl-cyclohexanol) one of a group of cyclicmonoterpenes that we have shown todecrease hepatic cholesterol synthesis inrats and man by their inhibition ofHMGCoA reductase activity in vivo.Hepatic HMGCoA reductase activity wasmeasured in Wistar rats given one of thefollowing: (1) TMC, (2) cyclandelatedissolved in olive oil, (3) olive oil alone, (4)-mandelate in normal saline, or (5) normalsaline alone. The TMC, cyclandelate, andmandelate were all given as a single dose of6 mmol/kg. The rats were killed 17 hourslater and the assays performed. Ninepatients with radiotranslucent gall stonesand radiologically functioning gall bladdershad a bile sample taken by duodenalintubation after an overnight fast. Theythen took cyclandelate in a dose of 400 mgbds (3) or 800 mg bds (6) for a mean of 29days before a second sample was obtained.A single dose of cyclandelate in the ratscaused a 62% inhibition of HMGCoAreductase activity (p<0.01). The samedegree of inhibition was obtained by TMC(p<0.01). Mandelate, the acid componentof cyclandelate, gave no significantinhibition. The mean lithogenic index (Hand D) of the gall stone patientsfell significantly from 1*4±0*05(mean ± SEM) to 1-1+±007 (p<0-05) aftercyclandelate. The improvement was due toa significant fall (p<0-05) in the molar ratioof cholesterol. Cyclandelate may prove tobe a useful adjuvant to bile acid therapy,particularly in obese subjects with highhepatic HMGCoA reductase levels.

BASIC SCIENCE IIF43-F56

F43Subcellular localisation of 1251 intrinsicfactor (IF) and 57Co cyancobalamin(57CoBI2) during absorption by guinea-pigileum

H A SHEPHERD, J PRIDDLE, W J JENKINS, ANDD P JEWELL (Radcliffe Infirmary,Oxford) Previous studies have shownthat a B12 is processed by the lysosomesand cystol during absorption through theileal enterocyte. The fate of IF during B12absorption is unknown.Pure human 125I IF-57CoB12 complex

prepared from a protein A-anti IF columnwas applied to guinea-pig ileal tissue main-tained in organ culture from 15 minutes to22 hours. At selected time intervals, tissuewas counted for gamma activity of bothisotopes and homogenised. Subcellularorganelles of the post nuclear supernatant(PNS) were fractionated on Percoll densitygradients. The distribution of organelleswas identified by enzyme markers and eachfraction was counted for 57Co and 125Iactivity.Both 125I and 57Co activity was detected

in ileal tissue but only approximately 50%was removed after washing with 0-IMEDTA representing a net uptake of bothisotopes. Subcellular fractionation at alltime intervals demonstrated 125I and 57Coactivity associated with the enzymemarkers for brush border, lysosomes, andcytosol fractions. The lysosomal fractionswere solubilised and run on SephadexG150. 125I activity eluted in the range forIF and IF-B12 complex, whereas 57Coactivity eluted in the range for the complexand free B12. No free 1251 activity wasdetected. These results demonstrate thatIF enters the cell during B12 absorptionprobably as the IF-BI2 complex which isthen processed by lysosomes.

F44Effect of pectin on jejunal glucoseabsorption in man

B FLOURIE, N VIDON, C FLORENT, AND J J

BERNIER (Inserm U54 - Hopital Saint-Lazare, Paris, France) Ingestion ofpectin improves glucose tolerance in bothnormal and diabetic patients. It has beenshown that pectin delays gastric emptying,but a direct effect on intestinal glucoseabsorption has not been demonstrated so

far in man, and experiments in rats gaveconflicting results.The acute effect of pectin on glucose

absorption was studied by segmental per-fusion technique with proximal occlusiveballoon to eliminate the gastric factor. Theeffect of three different concentrations ofpectin (6, 10, 15 g/l) was studied in 22normal volunteers. Two, solutions wereinfused (10 mg/min) at the ligament ofTreitz. They contained PEG (5 g/l), NaCI(130 mM), KCI (5 mM) with either glucoseor mannitol (30 mM). The length of thestudied intestinal segment was 25 cm.

Results showed that, for each concentra-tion of pectin, there was a significantreduction of water and sodium absorption(p<1%), compared with either glucose ormannitol control solutions. This reduction,however, was not dose-dependent. Theabsorption of 30 mM glucose was reducedby addition of pectin (p<1%). Thisreduction of glucose absorption was dosedependent; mean percentages of inhibitionon 25 cm segment were 10, 13, and 27%,for 6, 10, and 15 g/l pectin, respectively(r=0.58, p<l%). Pectin had no effect onglucose-linked sodium transport.Our results show that, in healthy man,

there is an acute inhibitory effect onintestinal glucose absorption. This effect isobserved with low intraluminal pectinconcentrations, similar to those achievedwhen pectin is given in a test meal.

F45Studies of the mechanisms of action of theantidiarrhoeal agent loperamide

S HUGHES, N B HIGGS, AND L A

TURNBERG (Department of Medicine,Hope Hospital (University of ManchesterSchool of Medicine, Salford) Loperamideis an effective antidiarrhoeal agent withwell-recognised effects on motility. Weinvestigated the possibility that it mightalso influence intestinal absorption andhave anti-secretory effects. Stripped rabbit'ileal mucosa was mounted in a modifiedUssing chamber. Loperamide (10`8 to10-6M) added serosally caused a dose-related fall in potential different and short-circuit current (p<0.01) leaving resistanceunchanged but was inactive on the mucosalside. This effect was inhibited by naloxone10-6M (p<0-02). Loperamide 10-6Malone failed to influence basal ion transportbut did inhibit secretion provoked byprostaglandin E2 (PGE2) and by E. coliheat labile toxin. Electrical responses toPGE2 (n=18) and E. coli (n=12) were not

A911


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/


prevented by loperamide. The PGE2reduction in net Na absorption, however,(from 7-14±0-97 to 2.39±1-07, p<002,n=6) was partially inhibited by loperamide(7.16+0.88 to 5.47+0.82, n=6). PGE2induced net Cl secretion (from control of1.89±0.70 to -1.2±0.99, p<0.01) was alsoinhibited (2.29±0.48 to 1-08±0-92 NS).Unidirectional flux responses were alsoinhibited.Heat labile E. coli toxin (200 ,ul crude

extract) reduced Na absorption (from5*14±0X76 to 1-96±0-99, p<0 01, n=6)and induced Cl secretion (from 1-98±0-90to -1-49±0-72, p<0.05, n=6) by reducingmucosa to serosa fluxes. Loperamideinhibited these responses (net Naabsorption 4-25±1 04 to 3-46±1 43, NSand Cl, +1-74±1 06 to -0-32±0-83, NS).We conclude that loperamide influences

intestinal mucosa through opiate receptorsand part of its anti-diarrhoeal effect may bemediated by its anti-secretory activity.

F46Can rapid small bowel transit limitabsorption of a meal?

A N HOLGATE AND N W READ (Departmentof Physiology, University of Sheffield,Sheffield) The small bowel transit andabsorption of radiolabelled solid meal ofknown composition were measured in 14patients, who had had a terminal ileostomyfashioned for ulcerative colitis. Measure-ments were carried out under controlconditions and after administration ofthree agents, which speeded up smallbowel transit by different mechanisms.Small bowl transit time was taken as thedifference between the half times forgastric emptying, measured by countingover the surface of the stomach with a

crystal scintillation detector, and ilealemptying, measured by estimating thedelivery of isotope from the stoma. Thelatter provided an excellent index of thedelivery of fat, protein, and carbohydratecomponents of the meal (r=0.99).Absorption of the meal was determined bychemical analysis of the ileostomy efiluent.Administration of either metoclopramide(20 mg tds), magnesium sulphate (7 g), or

lactulose (40 g) significantly reduced smallbowel transit time (p<005) and increasedthe weight of ileal effluent (p<0.05). Allthree agents significantly reduced theabsorption of fat (p<0.05), and lactuloseand magnesium sulphate also decreasedthe absorption of protein and carbohydrate(p<O.OS).

We concluded that rapid small boweltransit may limit absorption by reducingthe contact of food with the absorptiveepithelium.

F47Adult hereditary fructose intolerance:isolation and characterisation of a mutantliver aldolase

T M COX, M W O'DONNELL, AND M

CAMILLERI (Department of Medicine,Royal Postgraduate Medical School,London) Hereditary fructose intolerance(HFI) is a metabolic disorder caused byenzymic deficiency of aldolase 13 but themolecular basis of this defect is unknown.We have studied a remarkable family withhereditary fructose intolerance in a 45 year

old man and his three children, all of whonmhad suffered typical abdominal pain,vomiting, and hypoglycaemic synmptomsafter fructose ingestion, since infancy.Assay of fructose-l-phosphate aldolaseconfirmed a profound deficiency (<2%control) of this specific enzyme activity inbiopsy samples of liver and intestine.Moreover, kinetic studies showing a 10-fold Mean increase in Michaelis constant(24-92 versus 1-6-3 mM) suggested thepresence of an aberrant enzyme.To characterise and isolate this putative

mutant enzyme, immunochemical tech-niques were used: aldolase B was purifiedfrom human liver and monospecific anti-aldolase immunoglobulins were preparedfrom antiserum raised in sheep. Immuno-diffusion gels showed a precipitin linecommon to the pure enzyme, normal liver,and intestinal extracts but reactions withother control tissues of hereditary fructoseintolerance liver were not detected. Asensitve and specific radioimmunoassaywas thus used to search for immuno-reactive aldolase (detection limit 7-5 ng;ID50 125 ng). Extracts of tissues fromhereditary fructose intolerance patientsgave 10-25% of control levels of immuno-reactive enzyme; immunoreactive aldolasein four proven heterozygotes was, as

expected, reduced (to 55%) whencompared with seven control samples(p<O-05).

Immunoaffinity chromatography on anti-aldolase Sepharose gels enabled purealdolase to be recovered from liversupernatants. Electrophoresis in SDS-polyacrylamide showed a single proteinband from hereditary fructose intoleranceliver apparently identical with normalactive aldolase (subunits 42600 daltons). In

hereditary fructose intolerance there was astriking reduction in apparent specificactivity of the aldolase and displacementimmunotitration showed a markedlydecreased affinity for antibody.We conclude that aldolase B deficiency

in hereditary fructose intolerance isassociated with the synthesis of animmunoreactive but structurally alteredenzyme protein.

F48Improved methods for the metabolicprofiling of bile acids in faeces

J M GILBERT, A M LAWSON, J WORTHINGTON,AND K D R SETCHELL (introduced by A GCox) (Department of Surgery andDivision of Clinical Chemistry, ClinicalResearch Centre, Harrow,Middlesex) Because of their probablerole in the aetiology of diseases such ascolon cancer and other gastrointestinaldisorders, there has been renewed interestin the measurement of bile acids in faeces.Improved techniques have been

developed in our laboratory for thedetermination of metabolic profiles offaecal bile acids. These employ acombination of liquid-gel and liquid-solidextraction before group separation of bileacids according to their mode ofconjugation. Metabolic profiles areobtained using high resolution capillarycolumn gas chromatography withidentification of compounds by massspectometry.Twenty-seven different bile acids have

been identified in the unconjugated state inthe faeces of normal adult female Sprague-Dawley rats. Quantitative changes in thefaecal excretion of these bile acids werestudied over 14 consecutive days in twoanimals. Deoxycholic acid was thepredominant bile acid excreted by bothanimals and ranged from 686-4545 ,ug/day.There was an abundance of 6-hydroxylatedbile acids of which hyodeoxycholic was themost prevalent (491-3193 ,g/day) and atleast four muricholate isomers wereidentified. Contrary to previous reportschenodeoxycholic and cholic acid were notpresent.These data indicate that the metabolic

profile is more complex than previouslysuggested. The methods provide increasedsensitivity and specificity for investigationof the relationship between faecal bileacids and gastrointestinal diseases inhumans and rats.

A912


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/


F49Reversibility of sucrase inhibition byacarbose in in vivo perfusion of the ratjejunum

R H TAYLOR AND HELEN M BARKER (Depart-ment of Gastroenterology, CentralMiddlesex Hospital, London) The ox-glycoside hydrolase inhibitor acarbose is apowerful competitive inhibitor of sucrasein man. An oral dose of 200 mg results inalmost complete malabsorption of asucrose load. In establishing a therapeuticdosage regime the kinetics of recovery ofsucrase activity is important. In theseexperiments we have studied this in an invivo perfusion model.At laparotomy in five anaesthetised

adult Sprague-Dawley rats a segment ofproximal jejunum 20 cm long wascannulated and perfused in vivo with anisosmolar sucrose/saline solution. '4C PEG4000 was used as an unabsorbable markerfor calculation of absorption rates. Sugarsin the aspirate were measured by highpressure liquid chromatography andelectrolytes by flame photometry. After ahalf-hour stabilisation period, aspirateswere collected at five minute intervals forthree hours. In the second half-hour of thethree hour period, acarbose 30 mg/l wasalso present in the perfusion solution.Acarbose caused a rapid onset 86%

inhibition of the sucrose absorption ratefrom 83.5±5.9 to 11*6±1*8 ,umol/h/segment. There was a similar reduction inwater and sodium absorption. Luminalfructose fell by 90% and glucosedisappeared. Over the next two hoursthere was a gradual return of sucraseactivity reaching 50% recovery after 75minutes. A similar recovery in sodium andwater absorption rates was seen.The results show that acarbose is an

effective reversible inhibitor of sucrase inthe rat in vivo with half life of the order of75 minutes.

F50Hyperenteroglucagonaemia and smallintestinal hyperplasia after colonicperfusion of glucose in the rat

B M MIAZZA, A FILALI, M A GHATEI, M Y TAL-MUKHTAR, N A WRIGHT, S R BLOOM, AND JC RAMBAUD (Inserm U54 - Hopital Saint-Lazare, Paris, France; Departments ofMedicine and Histopathology, Royal Post-graduate Medical School,London) Hormonal factors play animportant role in intestinal adaptation and

enteroglucagon is a strong candidate forthe role of enterotrophin. It is distributedthroughout the intestine, with a maximalconcentration in terminal ileum and colon.In most conditions associated with smallbowel hyperplasia, such as lactation,enterectomy and pancreaticobiliarydiversion, enteroglucagon plasma concen-trations are raised. To study further therole of enteroglucagon in promoting smallbowel adaptation, enteroglucagon releasewas stimulated by continuous colonic per-fusion of glucose in orally-fed rats. Possiblechanges in cell proliferation and mucosalmass of an otherwise intact small intestinewere assessed by measuring: crypt cellproduction rate (CCPR), mucosal wetweight, DNA and protein per cm in jejunaland ileal segments. Enteroglucagon plasmaconcentration was measured by RIA. Ratswere killed eight days after continuouscolonic perfusion of 10% glucose (1-8 ml/h)and resultant changes in the intestinecompared with those found in 10%mannitol- or isotonic saline - perfusedcontrol rats.

Colonic perfusion of glucose provoked atwo- to six-fold increase in plasmaenteroglucagon compared with controlvalues, without significant change inpancreatic glucagon. In the jejunum,CCPR per hour increased from 21.18±SE0.6 in controls to 31*87±1*93 after glucoseperfusion (p<0.001) and in the ileum from177± 1*86 to 30*41±1*79 (p<0.001).Stimulation of crypt cell poliferationresulted in a significant increase in mucosalwet weight from 47-1±SEM 2-5 mg/cm to63-1±2-1 (p<0.001) in jejunum and from33-2±1-6 to 44-3±3-0 (p<0.01) in ileum,with corresponding changes in DNA andprotein.We conclude that colonic perfusion of

glucose stimulates the release ofenteroglucagon and is associated withjejunal and ileal mucosal growth in orally-fed rats. These results suggest that entero-glucagon plays an important role inregulating small intestinal cell proliferationand mucosal mass.

F51Effect of the intestinal peptide PHI on netintestinal fluid transport in the rat

M GHIGLIONE, N D CHRISTOFIDES, L 0UTTENTHAL, K TATEMOTO, AND S RBLOOM (Royal Postgraduate MedicalSchool, London, Karolinska Insititute,Stockholm, Sweden) Many factors areinvolved in the control of the transpor-

tation of fluid across the intestinal mucosa.Regulatory peptides form components ofsuch control. PHI is a peptide recentlyisolated from porcine intestine which hasbeen included in the glucagon-secretinfamily of peptides because of homologiesin its sequence and in its actions. We havestudied the effect of intravenous infusion ofPHI on net fluid transport in the smallintestine in the rat, using a previouslydescribed method.Male Wistar rats (n=6), weighing 300-

350 g, were infused with a nominal dose of500 pmol/kg/min of synthetic PHI over 45minutes. Control rats (n=7) were infusedwith normal saline over the same period oftime. Net fluid transport was measured inone duodenal segment, three jejunalsegments and three ileal segments (2.5 cmeach), and expressed as mean change(±SEM) of luminal fluid in ml/g ofintestine. During PHI infusion netabsorption was reduced in the duodenumto -0-29±0-15 (control -0-99+0*15,p<0.01) and in jejunal segments to-0.12±0.07 (control -0-84±0-04,p<0.001), whereas net secretion was seenin the ileal segments (+0.40±0.06) insteadof the net absorption (-0.56±0.08,p<0.001) observed in controls. Thussynthetic PHI appears to be capable ofcausing intestinal secretion in the rat.

F52Site of chemoselective resistance to gastricemptying of liquids

N S WILLIAMS, J MILLER, AND J H MEYER(introduced by D Johnston) (Center forUlcer Research and Education, LosAngeles, USA, and University Departmentof Surgery, The General Infirmary,Leeds) Glucose or oleate empty moreslowly from the stomach than 0. 15M NaCi.This chemoregulation may be achieved byresistance beyond the proximal stomach.We sought to define the site of thisresistance. Expt. 1: gastric emptying ofNaCI and 600 mM G was measured in fivedogs with intestinal cannulas 45 cm distal tothe pylorus while gastric pressure wasmaintained at 10, 18, or 25 cm H20 with abarostat. The dogs were studied either withan uninterrupted intestinal stream (Intact)or with duodenal effluent diverted througha second barostat before its return down-stream (Diverted). As the latter ensured aconstant pressure at the ligament of Treitz,any observed effects on emptying could notbe due to resistance further distally. Expt.2: emptying of NaCl, glucose, and 40 mM

A913


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/


oleate emulsion was measured in five dogswithout cannulas after control antro-duodenal transection and after antrectomywhile gastric pressure was controlled at 11,17 and 23 cm H20.Under both conditions in expt. 1

emptying rose linearly with pressure, yetNaCl emptied significantly faster thanglucose. (Mean emptying rate ± SE ml/min. Intact: NaCl= 14-8±2-6;glucose=5-6±2-0, p<005. Diverted:NaCl=15*3+4*3; glucose=3-2+1.4,p<0.05). Thus a major site of chemo-selective resistance lies proximal to theligament of Treitz. After antrectomy NaCland oleate emptied significantly faster thanbefore; however, the difference betweensaline and nutrients was maintained(Before NaCl=*8.6±2.1; glucose=3-2±0-7; oleate=4- 1±0-7. After NaCl=*t22.2±3.7; glucose=5-5+7-7; oleate=t9-3±1-0; *NaCl v glucose or oleatep<005. tAfter v. before p<0-05). Thusresistance does not reside in the antrum orpylorus and clearly the duodenum isimplicated.

F53Porcine histidine isoleucine (PHI) inducessecretion in the pig small intestine

A A ANAGNOSTIDES, K MANOLAS, N D

CHRISTOFIDES, Y YIANGOS, R B WELBOURN, S R

BLOOM, AND V S CHADWICK (Departmentsof Medicine and Surgery, Royal Post-graduate Medical School, HammersmithHospital, London) Porcine histidineisoleucine (PHI) is a newly discovered gutpeptide with structural similarities tovasoactive intestinal polypeptide (VIP), anestablished intestinal secretagogue. Wetherefore investigated the effect of PHI onnet water and electrolyte movements. Insix anaesthetised pigs, five jejunal, fourileal, and three colonic loops were perfusedat 6 ml/min with a bicarbonate/electrolytesolution containing PEG 4000. After a 30minute equilibration period, net water,sodium, potassium, chloride, andbicarbonate transport were measured ineach loop in three test periods: (1) a 30minute basal period, (2) a 40 minute periodduring which PHI was infused into theaorta at 30 pmol/kg/min, and (3) a 30minute post-infusion period.

In the basal period there was netabsorption of water in the jejunum(-34.7±17.8 ,ul/cm/10 min) and ileum(-27.8±32 ,ul/cm/10 min). During infusionof PHI marked water secretion occurred(+64.8±54.9 in jejunum and +27-4±34-3

in ileum, p<0-01), with return toabsorption in the post-infusion period(-23.8±23.9 in jejunum and -12*8±13*2in ileum, ns. v. basal). PHI also causedmarked small intestinal secretion of sodiumand chloride which again returned towardsbasal levels in the post-infusion period. Netpotassium and bicarbonate fluxes did notchange significantly. No consistent patternof secretion was observed in colonic loops.We conclude that infusion of PHI causes

reversible secretion in the porcine smallintestine. In view of these finding PHI mayplay a role in certain cases of secretorydiarrhoea.

F54Prolonged and intermittent stress inhibitshuman fasting motor complexes (MCs)

R M VALORI, M P H PATRICK, A RAIMAN, A

PARNHAM, AND D L WINGATE (Gastro-intestinal Science Research Unit, LondonHospital Medical College, London)Previous studies have shown thatprolonged mental stress inhibits humanfasting upper gastrointestinal motility;however, adaptation occurred after twohours with the single stressor used. Usingan ambulant system and three differentstressors, we have assessed the effects ofprolonged but intermittent stress on thehuman motor complex. Fasting duodenaland jejunal motility in healthy volunteers(n=8) was monitored continuously for 33hours using dual pressure sensitiveingestible radiotelemetric capsules. Theincidence of fasting motor complexesduring seven hours of rest (day 1) wascompared with the incidence in a sevenhour period (day 2) during which subjectswere stressed intermittently by thecomputer game 'Asteroids' (two hours),driving in heavy traffic (two hours), and adelayed auditory feedback technique (onehour). Stress response was quantified usingblood pressure, heart rate, and self-reported visual analogue scales. Positiveresponses, being taken as a 5% increase inblood pressure and 10% in the othervariables, were 50% in systolic bloodpressure, 60% in diastolic, 70% in heartrate, and 75% in the visual analoguescores. The incidence of motor complexes/hour in day 1 was 0-53±0-09 (mean ± SD)and 0-27±0-18 on day 2, this differencebeing significant (p<0-02; Wilcoxonmatched pairs signed rank test). Thevalidity of this was further emphasised bythe similarity of motor complex incidenceduring the inter-stress rest periods on day 1.

We conclude that the intermittentapplication of different stressors cansignificantly reduce the incidence of motorcomplexes over prolonged periods of timeand postulate that these findings may beconnected with symptoms of functionalbowel disease in susceptible individuals.

F55Effect of high fat meals on the disruption offasting motility patterns in man

D F EVANS, G E FOSTER, AND J DHARDCASTLE (Department of Surgery,University Hospital, Nottingham) Feedingdisrupts the migrating motor complex inseveral species including man. It has alsobeen shown in dogs that the percentage offat in a meal delays the return of the first ofthese complexes. We have studied theeffect of different percentages of fat inmeals on the disruption of fasting motilityin man.Twenty fasted healthy volunteers were

given one of four 600 kcal test mealscontaining 0, 15, 30, or 60 g of total fat.Antral and jejunal motility was monitoredusing linked radiotelemetry capsules aspreviously described.

All four meals disrupted fasting activity.The median return to fasting activityindicated by the occurrence of a phase IIIactivity front was significantly shorter forthe 0 g and 15 g fat meals compared withthe 60 g high fat meal (Og:360 min, range184-400, 15g:390 min, range 132-576,60g:424 min, range 344-454) (p 004 in bothcases). Fifty per cent of the activity frontswere not preceded by antral activity,however, and may not migrate the fulllength of the small intestine.

In this study we have found that onlymeals with an unusually high fat contentcause any significant delay in the return tofasting activity and this may therefore havelittle clinical significance.

F56Effect of colonic motor activity on theinternal anal sphincter

G G P BROWNING AND A G PARKS (SurgicalResearch Department St Mark's Hospital,London) The myenteric plexus is thoughtto mediate intestinal peristalsis and therecto-anal inhibitory reflex. The aim of thisstudy was to determine whether peristalsisin the distal colon had any effect on theinternal anal sphincter.

Fifteen patients were studied. Group 1

A914


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/


consisted of 10 patients with a normalinternal anal sphincter and deficiency ofthe external anal sphincter; seven withpelvic floor neuropathy, and three withcauda equina lesions. Group 2 consisted offive patients with completely normal analsphincters (normal controls). Intraluminalpressures in the distal colon, rectum, andproximal anal canal were recorded inresponse to a standardised motilitystimulation test. A solution of the surfaceactive agent bisacodyl, applied to thecolonic mucosa at 27 cm from the analverge, was used to stimulate propulsivecolonic motor activity.

After stimulation all patients exhibitedincreased motor activity of a propulsivetype. Reflex inhibition of the internal analsphincter was recorded in response topropulsive sigmoid motor activity in all 10patients in group 1. The internal analsphincter inhibition was not abolished bymaintaining the lower rectum at atmos-pheric pressure to prevent distension. Incontrast, in group 2, only one patientshowed reflex inhibition of the internalanal sphincter; the remaining four patientswere observed to voluntarily contract theexternal anal sphincter resulting inincreases in intra-anal pressure.These results show that a mechanism

exists whereby propulsive motor activity inthe distal colon reflexely inhibits internalanal sphincter resting tone. Such a reflexmay be mediated by the myenteric nerveplexus.

ENDOSCOPYF57-F63

F57Duodenoscopic treatment of acutecholangitis

A G VALLON, P J SHORVON, AND P BCOTrON (Department of Gastroenterology,The Middlesex Hospital, London)Duodenoscopic sphincterotomy is nowroutine treatment for duct stones, but itsvalue in patients with acute cholangitis isnot widely appreciated.Of 186 patients referred for endoscopic

management during 1979 and 1980, 14(7.5%) were acutely ill, with pain,jaundice, fever, and leucocytosis. Allpatients in whom specimens were takenhad positive cultures (blood 7/7, bile 13/13). Ages ranged from 47 to 92 years(mean 72 years); nine had previously

undergone cholecystectomy.Nine patients had impacted stones, all

with pus; five patients had mobile stones.Sphincterotomy with immediate stoneextraction or pernasal biliary drainage wasperformed without complication, leadingto rapid clinical improvement. Stonesgreater than 15 mm diameter could not beextracted endoscopically in two patients.One was treated surgically; the second wasjudged to be totally unfit for surgical orother treatment, and died with recurrentsepsis.Duodenoscopy allows effective man-

agement of acute purulent cholangitis, andhas potential advantages over per-cutaneous drainage or urgent surgery.

F58Bacteraemia after flexible fibreopticsigmoidoscopy

P A FARRANDS, M J STOWER, M S OSMAN, ANDJ D HARDCASTLE (Departments of Surgeryand Microbiology, University Hospital,Nottingham) Bacteraemia may followcolonoscopy in up to 20% of examinations.With the increasing use of the 60 cmflexible fibreoptic sigmoidoscope in out-patient departments we felt it important todetermine the incidence of bacteraemia inpatients undergoing this examination.One hundred consecutive patients

undergoing flexible fibreoptic sigmoido-scopy were given a single phosphate enemahalf an hour before examination. Tenmillilitres of blood were taken from theantecubital fossa immediately before andafter the endoscopy and incubated foraerobic and anaerobic organisms. Culturebottles were inspected daily for seven days.Fifty-two patients were male, with a meanage of 60-2 years, range 16-80 years. Inonly 13 could the sigmoidoscope not beadvanced to its full 60 cm. The averagetime taken for the examination was 8*3(±3.7 SD) minutes. Fifty-eight patientshad significant abnormalities on examina-tion, 16 of which were bleeding at the time.Biopsies were performed in 42 patients.Only one of the blood cultures waspositive. This was taken after endoscopyfrom a woman with florid ulcerative colitisand grew Escherichia coli.Bacteraemia after flexible fibreoptic

sigmoidoscopy is rare and routineprophylaxis for the majority of patientsundergoing this examination is unneces-sary. Patients with inflammatory boweldisease and a concomitant cardiacabnormality or compromised immune

system appear to be at risk, however, andshould be given antibiotic prophylaxis.

F59Radiological control of oesophageal sciero-therapy: an improved technique

J D R ROSE, P M SMITH, AND M DCRANE (Department of Gastroenterology,Llandough Hospital, Penarth, SGlamorgan) Intravariceal rather thanparavasal sclerosant injection is moreeffective in obliterating varices but someinjections will be paravasal, because it isdifficult to be certain of the site ofinjection. This may lead to ulcer andstricture formation. We have, therefore,determined the site of intended intra-variceal injections radiologically with a 1:3mixture of 76% Urografin and 3% sodiumtetradecyl sulphate in 20 patients. Of 68injections, 51 (75%) were intravenous.Injection into large varices was intravenous35 times in 43 (81%) but into small varicesonly 16 times in 25 (64%).Fewer complications and faster variceal

obliteration might be expected if allinjections were intravariceal. Eightpatients have been treated purely underradiological control. Two millilitres of thesclerosant-contrast mixture were given, butthe injection was stopped if a paravasal sitewere demonstrated. There has been onlyone ulcer and no stricture formationcompared with an incidence of 27% forboth complications in our previouslyreported series. Variceal obliteration wasachieved in a mean of 2-8 sessions (rangetwo to four) compared with a mean of six(three to 13) in our earlier series(p<0.002).

Injection under fluoroscopic controldemonstrates that a quarter of intendedintravariceal injections are extravascular,but, if injections are kept within the lumen,complications are reduced and obliterationis achieved in significantly fewer sessions.

F60Controlled trial of Nd YAG laser photo-coagulation in bleeding peptic ulcers

C P SWAIN, S G BOWN, P R SALMON, J SKIRKHAM, AND T C NORTHFIELD (Depart-ment of Gastroenterology, UniversityCollege Hospital, and Norman TannerGastroenterology Unit, St James's Hospital,London) The efficacy of Nd YAG laserphotocoagulation in the endoscopic controlof haemorrhage from peptic ulcers was

A915


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/


tested in a controlled trial at two centres inLondon. One hundred and eighty-sevenunselected patients admitted consecutivelywith upper gastrointestinal haemorrhagewere submitted to emergency endoscopy.Peptic ulcers were seen in 83. All 47 withstigma of recent haemorrhage (SRH)accessible to laser therapy were included inthe trial (11 inaccessible, 25 had no SRH).Prospective stratification was in threegroups - those with a visible vessel, thosewith other SRH, and those in which theclot could not be washed off beforetherapy. Within each group, ulcers wererandomised to conventional managementwith or without endoscopic laser therapy.Treatment was carried out using 0-5 secondlaser pulses at 80 watts. Laser and controlgroups were well matched for other factorsknown to influence prognosis. Overall,three of 26 treated patients and 10 of 21control patients rebled (p<0-02). Con-sidering ulcers with a visible vessel, threeof 17 treated and nine of 13 controls rebled(p<0.02). Considering other SRH, none ofthe seven treated rebled compared withone of seven controls. No treated patientsbut three control patients died after an

episode of rebleeding. These preliminaryresults of this continuing trial suggest thatNd YAG laser has significantly reduced therebleeding rate in patients with a bleedingpeptic ulcer accessible to laser therapy.

F61Bipolar endoscopic electrocoagulation

D L MORRIS, P C HAWKER, M R B KEIGHLEY,AND P W DYKES (The General Hospital,Birmingham) Patients with a non-bleeding 'visible vessel' in the base of a

peptic ulcer have a 56% risk of rebleedingand those with active bleeding have aneven worse prognosis. In the caninestomach full thickness necrosis was notproduced in any of 20 single ACM bipolarprobe applications; 10 separate applica-tions to the same site produced serosalinjury in only 2/13. Bleeding from caninegastric ulcers (Protell) was controlled in13/13 cases. Arterial bleeding from caninemesenteric vessels was stopped in 16/17cases at a rate of <10 ml/min, but in onlyfour of 11 where bleeding was >10 ml/min.We repeated our study of the histologicaldepth of necrosis in man during electivegastric resection and confirmed the safetyof this device.

Endoscopic electrocoagulation has beenused in 20 patients without any related

complication or death. Nineteen of thesepresented with a major upper gastro-intestinal bleed. Active bleeding was

stopped in three out of four cases. Of the10 patients with a visible vessel at the baseof a peptic ulcer who have beencoagulated, two have rebled, both at twoweeks, but neither required operation. Incontrast with the unipolar probe, stickinghas not been a problem.The ACM bipolar probe is safe and is

capable of controlling bleeding from pepticulcers. The ability of this probe to reducethe risk of rebleeding in patients with a

visible vessel deserves further study.

F62Use of needle diathermy wire at ERCP

ANDREW CLARK (Brighton GeneralHospital, Brighton, Sussex) Successfulcannulation with papillotomy wherenecessary is not achieved in all patients.The rate of success varies according to themethod of reporting. In this series 60%were cannulated in the system of majorinterest. Of the remainder, three-quarters(31) were pre-cut with a 'needle' type ofdiathermy wire, leading to better biopsy ofthree tumours, images of the system ofmajor interest in three cases, images ofpapillotomy in 17 cases, and second-goimages with papillotomy in seven cases.This contrasts with the small number ofpatients who can be helped by con-ventional pre-cutting with the Classenloop. Complications included onehaemorrhage and one case of continuingpancreatitis, but no perforations occurred.

F63Duodenoscopic palliation of malignantobstructive jaundice with 10 Frenchprostheses

P B COTrON, L SAFRANY, B SCHOlT, AND

J LEUNG (Departments of Gastro-enterology, The Middlesex Hospital,London, and Reinhard Nieter Hospital,Wilhelmshaven, West Germany) SinceJune 1981, we have attempted to placeendoscopic biliary prostheses in 88 high-risk patients with malignant obstructivejaundice. We used an experimentalOlympus duodenoscope with a 3-7 mm

channel, and 10 French gauge prostheseswith double pigtails. Prostheses were

inserted in 73 patients (83%), with

obstruction of the lower duct (26), mid-duct (17), and hilum (30).

Jaundice was relieved in 69 patients, buteventually returned in 15 because of tubemigration (three), or obstruction (12).Four patients died within 30 days fromsepsis (three), and perforation (one).Mean survival was 104 days, and threepatients survived more than six months.None has required surgery for duodenalobstruction. Techniques are still beingimproved.

GENERAL IIIF64-F69

F64Medical therapy of hydatid disease

D L MORRIS, F BURROWS, S GOULD, fDICKSON, J BOGAN, AND P W DYKES (TheGeneral Hospital, Birmingham) Whilesurgery is usually successful in rembvingsolitary hydatid cysts, it is not withouthazard. Recurrent or disseminated hydatiddisease is a much more serious problem.Mebendazole has shown some effect onEchinococcus granulosus, but a dosage ofup to 200 mg/kg/day has been required,and improvement is seldom evident beforethree months. We have measuredmebendazole levels in serum and cyst fluidat operation by radioimmunoassay in sixpatients taking 50 mg/kg/day. Serum levelswere usually below 30 ng/ml and cyst levelsbelow 1 ng/ml.

Albendazole is a new benzimidazolecompound. We have used this drug at adose of only 10 mg/kg/ml for four weeks infour patients. The first patient had multiplecysts in the lung, mediastinum, chest wall,and retroperitoneum demonstrated by CTscan. As well as marked clinical improve-ment, repeat CT scan at one month showedmarked cyst shrinkage. Two of theremaining three have had similar cystshrinkage. The peak serum levels after one400 mg dose have usually exceeded 500ng/ml (high performance liquid chromato-graphy). A progressive fall in CFT titre wasseen in three of these cases.

Little mebendazole penetrates intohydatid cysts. Blood levels of albendazoleare much higher and our early clinicalresults are encouraging. This new drug maybe an important advance in the manage-ment of hydatid disease.

A916


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/


F65Illlndium plasma labelled autologousleucocyte scanning in the detection of intra-abdominal abscess

S H SAVERYMUTTU, A M PETERS, M E CROFTON,H J F HODGSON, J P LAVENDER, AND V SCHADWICK (Departments of Medicine andRadiology, Royal Postgraduate MedicalSchool, Hammersmith Hospital, London)"`Indium leucocyte scanning is a recentlyintroduced technique for the detection ofintra-abdominal abscesses with a diagnosticaccuracy comparable with ultrasound andcomputer tomography. With currentlyused cell-labelling techniques ("'Inacetonate or oxime in saline), however,there is a variable and significant delay inlocalisation of labelled cells at sites ofinflammation and variable sequestration ofcells in the lungs. A new method ofcell-labelling in plasma using "'Intropolonate has been developed whichabolished initial pulmonary sequestration.This improved labelling technique has beenused in 60 patients with suspected intra-abdominal abscess to study the rapidity ofabscess localisation and the diagnosticaccuracy of the technique. Fourteen out ofthe 60 patients were shown to haveabscesses, at laparotomy, at necropsy, andclinically by the spontaneous discharge ofpus. "'In tropolonate leucocyte scanningcorrectly identified 13 (93%) of abscesses.In seven patients abnormal activity corres-ponding to the site of the abscess wasvisible within one hour of return of thecells. In 12 other patients extra-abdominal(five) or intra-abdominal (seven) sites ofsepsis (other than intra-abdominalabscesses) were demonstrated whichaccounted for the clinical symptoms. Ultra-sound abdominal examination wasavailable in eight patients with provenabscesses and detected five (62.5%),including the single abscess missed on"'Indium leucocyte scanning. These resultsdemonstrate that "`In autologoustropolonate scanning is a rapid andaccurate method of detecting intra-abdominal abscesses.

F66Role of hepatic embolisation in advancedcarcinoid tumours

P N MATON, M CAMILLERI, C HASLElT, H JHODGSON, D J ALLISON, AND V S CHADWICK(Departments of Medicine and Radiology,Royal Postgraduate Medical School,Hammersmith Hospital, London) Hepatic

arterial embolisation therapy for advancedcarcinoid tumours has been evaluated in11 patients. Ten had flushing, sevendiarrhoea, three severe abdominal pain,two wheezing, three right heart failure, andthree a pellagroid rash. Beforeembolisation, patients received p-chlorphenylalanine, cypropheptadine, anaprotinin infusion, methyl prednisolone,and pethidine. Tobramycin, flucloxacillin,and metronidazole were given for 10 daysafterwards. After ensuring that the portalvein was patent, selective hepaticembolisation was performed under localanaesthesia. Successful embolisation withtumour necrosis was manifest by mildconstitutional symptoms, fever :380C andleucocytosis ¢10-7x 109 WBC/1. Meanaspartate amino transferase and alkalinephosphatase levels rose to 800 IU/l (n<45)and 300 IU/l (n<130) respectively. Serumbilirubin levels remained normal.Abdominal pain with hepatic friction ruboccurred in three patients and pleuraleffusion in two. One patient died ofsepticaemia after stopping antibiotics atseven days. Dramatic improvement insymptoms occurred in eight patients withabolition of flushing and abdominal painand reduction in diarrhoea (mean stoolfrequency from 10 to 2/day, n=6). Urinary5HIAA levels fell from 1132±746 (SD) to315±178 ,mol/day. Symptom relief lastedfrom one to 24 months (mean 10 months).Recurrent symptoms in two patients weretreated by a second embolisation withfurther remission of symptoms for up to sixmonths. Five patients have died one to 40months after their first embolisation, as aresult of systemic metastases or heartfailure.We conclude that hepatic artery

embolisation for carcinoid tumours mayproduce striking improvement insymptoms. Whether long-term survival isimproved is not yet clear.

F67Ultrasound guided percutaneous pan-creatography

W R LEES AND P B COTTON (The MiddlesexHospital, London) Ultrasound guidedpercutaneous pancreatography was

attempted on 13 patients. Standardmethods of real-time ultrasound guidancewere used to puncture the duct with a 22swg needle. Position of the needle tip in theduct was confirmed by aspiration of severalmillilitres of juice which was then replacedwith contrast (Conray 280). At the end of

the procedure the duct was decompressed.Satisfactory pancreatograms were

obtained in 10 patients. All three failureswere in patients with a normal calibre duct;in two of these puncture of the duct wasachieved but injection of contrast resultedin extravasation. All of the successfulprocedures were in patients with a dilatedduct (range 3-7 mm). In six of these onlyone pass of the needle was required, andonly one patient needed more than threepasses. No complications were recorded inany of the 13 patients.The indications for pancreatography

were a ventral pancreas with failure tocannulate the dorsal duct - four patients(includes three failures); a failed ERCP inchronic pancreatitis - one; inadequateERCP in pancreatic cancer - two; com-bined ultrasonographic and radiographicguidance for biopsy purposes - six.Percutaneous pancreatography is a

feasible, and, in most cases, easy methodof producing high quality pancreatogramsin patients with a dilated duct system.

F68Endoscopic review of patients havingpreviously undergone gastric surgery

P A FARRANDS, J R S BLAKE, J D HARDCASTLE,I D ANSELL, AND R E COTTON (Departmentsof Surgery and Pathology, University andCity Hospitals, Nottingham) After gastricsurgery for benign conditions patientsappear to be at risk of developing gastriccancer. Epithelial dysplasia is consideredto be pre-malignant and Morson hasrecently suggested a prospective study toevaluate its role in gastric screening.

Six hundred and thirty-six patients under75 years of age, having undergone gastricsurgery for benign peptic disease 15-20years previously, were identified. Sevenpatients had died of cancer of their gastricremnant seven to 20 years after operation.One hundred and eighty-nine patients stillalive and living in the area were traced andinvited to attend for interview; 85 (44.9%)accepted and 71 (37.5%) were gastro-scoped and multiple biopsies taken. Onendoscopy two patients had invasivecarcinoma of the stomach, 66 had histo-logical evidence of gastritis (56 (78.8%)chronic atrophic gastritis, 10 (14.1%)superficial gastritis). Thirty-eight patients(53.8%) had intestinal metaplasia and 15regenerative epithelial cell change. Elevenpatients (15.4%) had epithelial dysplasia(six mild, two moderate, three severe(carcinoma in situ)). There was no

A917


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/

A918

correlation between symptoms and thepresence of epithelial dysplasia, eight outof 11 patients with dysplasia being Visickgrade 1. In only one patient with milddysplasia was chronic atrophic gastritis andintestinal metaplasia not present. In thetwo patients with invasive cancer, severe

dysplasia, chronic atrophic gastric, andintestinal metaplasia was also present inthe areas of adjacent mucosa. We recom-

mend that patients after gastric surgery are

gastroscoped and patients with dysplasiareviewed regularly.

F69Natural history of the Mallory-Weisssyndrome

C P SWAIN, S G BOWN, J S KIRKHAM, P R

SALMON, AND T C NORTHFIELD (NormanTanner Gastroenterology Unit, St James'sHospital, and the Department of Gastro-enterology, University College Hospital,London) Eighty-one Mallory-Weisslacerations were found at emergency endo-scopy in 63 of 617 patients (10%) consecu-

tively admitted with upper gastrointestinalhaemorrhage to two London hospitals over

21 years. The male:female ratio was 1-5:1,mean age 42 years. Classic symptoms ofvomiting preceding haematemesis were

reported in 30 (48%). Fifty-seven (90%)presented with haematemesis, six withmelaena alone (10%). Heavy alcoholintake was admitted in 39 (62%). Pain on

vomiting was present in 16 (25%),dyspepsia in 28 (44%). Epigastric tender-ness was found in 23 (37%), nine (14%)were shocked. At endoscopy a singlelaceration was seen in 50, doublelacerations in eight, triple in five. Fifty-six(69%) were in gastric mucosa alone, 16(20%) crossed the cardio-oesophagealjunction, and nine (11%) were

oesophageal alone. The tear lay in a

quadrant posterior to the mid-point of thelesser curve in 68 (84%). Associatedmucosal abnormality was seen in 42 (67%),haemorrhagic gastritis in 27 (43%), andpeptic ulceration in eight (13%), probablythe major source of haemorrhage in six(10%). Hiatus hernia was present in 31(49%). Spurting arterial haemorrhage was

seen in four, which settled spontaneouslyin all except one, a visible vessel in two,oozing or adherent clot in 47, and no

stigma of bleeding in 28 tears. Rebleedingoccurred in nine, although the tear was

thought to be the main cause in only fivepatients (8%). Eight (13%) required more

than 6 units of blood. Three patients (5%)

came to surgery but two were bleedingpredominantly from other lesions; allsurvived. One alcoholic patient died ofpneumococcal septicaemia without furtherbleeding in hospital. We conclude that theMallory-Weiss syndrome is a commonfinding in patients presenting withhaematemesis; almost all will settle withconservative management.

OESOPHAGUSF70-F75

F70Palliative relief of malignant upper gastro-intestinal obstruction by endoscopic lasertherapy

S G BOWN, C P SWAIN, D A W EDWARDS, ANDP R SALMON (Department of Gastro-enterology, University College Hospital,London) We report the successfulpalliative relief of dysphagia in five patientswith carcinoma of the oesophagus orstomach after endoscopic therapy with anArgon or Nd YAG laser. All wereconsidered unsuitable for surgery or radio-therapy. The patients were all male, aged71-88 years and had tumours from 5-13 cmin length which occluded at least 90% ofthe lumen (two squamous and three adeno-carcinomas). Treatment was carried outunder direct vision, the laser fibre beingpassed through the biopsy channel of astandard endoscope. To optimisemechanical relief of obstruction, therapywas directed at areas of tumour projectingfrom the walls, particularly when nodulesof tumour were somewhat mobile blockingthe lumen (ball valve effect, three cases).Superficial tumour vaporised immediately,whereas deeper layers sloughed over thesucceeding two to three days using energiesof 2-10,000 Joules per session. Sympto-matic relief of dysphagia was achieved inall cases after two to six treatments. Allwere able to eat solid foods aftertreatment, and all were able to go home.Quantitative barium studies before andafter treatment showed an average four-fold increase in luminal diameter. Onepatient had slight chest discomfort aftertherapy. There were no perforations orother complications. These encouragingresults warrant further evaluation of thismethod in relation to other endoscopictechniques for palliation of dysphagia.


F71Can benign oesophagel strictures be safelyand adequately dilated in the outpatientdepartment?

J M SIMMS, C J STODDARD, J A R SMITH, ANDA G JOHNSON (University Surgical Units,Royal Hallamshire Hospital, Sheffield, andRoyal Liverpool Hospital, Liverpool)Benign oesophageal stricture dilatationusually requires hospitalisation and may befollowed by patient discomfort andiatrogenic oesophageal perforation. Thesafety and effectiveness of outpatientdepartment dilatation using Maloney'smercury-weighted dilators has beenstudied.Two hundred and fourteen dilatations

have been performed in 57 patients withdysphagia and benign oesophagealstrictures secondary to gastro-oesophagealreflux. Their mean age was 69-3 years(range 25-89 years), 27 being over 70years. Macroscopic oesophagitis was con-firmed endoscopically and malignancyexcluded by biopsy. After a six-hour fast,and with only oropharyngeal anaesthesia,dilatations were performed in anoutpatient clinic. Patients were allowedhome immediately with instructions tohave nothing by mouth for two hours andto return if complications developed.Gastro-oesophageal reflux was controlledmedically and dilatations repeated whendysphagia recurred. Between one and 24dilatations were performed in each patient(mean 3.64).Dysphagia was improved in all patients

and the procedure was well tolerated bymost. No oesophageal perforationsoccurred, the only complications being aninsignificant haematemesis in one patient.The interval between dilatations increasedwith time. This outpatient dilatationtechnique is effective, well tolerated, andmay be safer than other methods ofoesophageal dilatation.

F72Evaluation of ambulatory recordings ofgastro-oesophageal reflux: a frequencyduration index (FDI)

F J BRANICKI, D F EVANS, J A JONES, A L

OGILVIE, M ATKINSON, AND J D HARDCASTLE

(Department of Surgery, UniversityHospital, Nottingham) Scoring tables foracid reflux have been described but aredesigned with unjustifiable weighting ofsupine reflux episodes. We have describeda radiotelemetric technique to con-


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/


tinuously monitor oesophageal pH in fullyambulant subjects at work and at home.Gastro-oesophageal reflux was recorded in20 normal subjects and 20 symptomaticpatients with oesophagitis graded at endo-scopy and biopsy. Analysis of our 24 hourpH recordings using this scoring systemunderestimated the severity of reflux, asseven out of 20 symptomatic patients withknown oesophagitis were classified as

normal.We have therefore devised the Fre-

quency-Duration Index (FDI). A refluxepisode is defined as a fall of at least 2 pHunits to less than pH 5. The FDI was

obtained by multiplying the frequency andcumulative duration (minutes) of refluxepisodes per hour.

In symptomatic patients during the day(0630-2230) the median FDI was 36-21(range 0.38-158.11) compared with 0-65(range 0-42.21) in normal subjects(p<0-001). At night (2230-0630) themedian FDI was 2-01 (range 0-73.53) insymptomatic patients compared with 0.01(range 0-0.38) in normal subjects(p<0-001). The index misclassified onlyone normal subject and two patients (oneof whom was thought to have takenantacids) and is considered to be a satis-factory discriminant of abnormal gastro-oesophageal reflux.

F73Gastro-oesophageal scintiscan (GOR scan)compared with pH monitoring

J C R MARTINS, P ISAACS, S EDWARD,

M MAISEY, AND G E SLADEN (Department ofGastroenterology and Department ofNuclear Medicine, Guy's Hospital,London) The value of GOR scan in 22symptomatic patients and eight asympto-matic controls has been assessed bysimultaneous intra-oesophageal pHmonitoring.From a continuous record of pH the

mean duration of reflux (pH <4.0) over 24hours was calculated (normal <3-5 min/h).A pneumatic binder was wrapped roundthe abdomen and the subjects drank 500 ml0-1M HCl in orange juice containing 11MBq 99mTc sulphur colloid. Serial supineimages were taken at 30 second intervals,the binder pressure beCing raised by 20 mmincrements to 120 mmHg; visible isotopeabove the diaphragm or a pH <4-0 at anypressure were considered positive. Usingstored data from 21 patients, 'blind'

quantification of reflux was obtained andconsidered positive if >2% of ingesteddose refluxed.Of 18 subjects with positive GOR scan,

17 showed reflux by pH probe. Of 12subjects with negative GOR scans, threeshowed reflux by pH probe. Six subjectshad 24 hour reflux >3.5 min/h and all hadpositive GOR scans but 14 other subjectswith 24 hour reflux <3-5 min/h also hadpositive scans. This poor specificity (42%)was improved to 79% by quantification ofthe scan data.

Simultaneous pH monitoring shows thatGOR scan detects reflux with 85% sensi-tivity and 90% specificity. GOR scan, withquantification, can predict the result of 24hour intra-oesophageal pH monitoring andprovides a useful non-invasive test forreflux.

F74Pharyngo-oesophageal submucosal fibrosis:a possible cause of webs and strictures

D A W EDWARDS (Surgical Unit, Faculty ofClinical Sciences, University CollegeLondon, London) Fifty-six patients (17male) with dysphagia had one (15) or two(41) concentric ring constrictions at thepharyngo-oesophageal junction, com-

monly called 'sideropenic webs'. All had a

sensation of obstruction synchronous withthe pharyngeal phase of swallowing. Manycould sip but not drink. The obstructionwas not casued by cricopharyngealmyopathy. The upper ring was in the planeof the pharyngo-oesophageal junction andthe lower 8 to 15 mm (x 12 mm) beyond.The 'webs' appear to be rings of increasedresistance to stretch rather than an

ingrowth of epithelium and are so similar inposition from patient to patient that theyare likely to represent an anatomicalfeature. They may not be visible unless theadjacent more compliant wall is welldistended by fluid. Shallow notches are

sometimes seen in normal subjects. 'Webs'are often missed at endoscopy andnecropsy. Twenty-six also had a cylindricalor posteriorly eccentric narrowing betweenthe webs. Biopsies and one necropsy revealsubmucosal fibrosis. One male developedstenosis overnight from haemorrhage fromthe venous plexus. Evidence of pastanaemia was occasional. The concept of'webs' as epithelial sheets needs revisionand the role and origin of submucosalfibrosis should be explored.

F75'Tender oesophagus': a new syndrome

D A W EDWARDS (Surgical Unit, Faculty ofClinical Sciences, University CollegeLondon, London) More than 65 patientshave been recognised as having a syndromecharacterised by an abnormal sensitivity tostretch of part of the oesophageal wall.Sometimes there is hypersensitivity to heatand acid without evidence of oesophagitis.This suggests that the receptors are in thesubepithelial layer. They can sip but notdrink comfortably, the more viscous thefood the more it hurts. Pain simulates a

sensation of obstruction and they presentas painful (apparently) obstructivedysphagia, living on liquids, sometimesregurgitating food. Duration is weeks tomany years. Age and sex incidence seems

random. Radiological examination demon-strates normal motility and compliance ofthe wall, and no evidence of obstruction.Endoscopy shows a normal mucosa inappearance and biopsy. Originally thedefinition excluded patients with abnormalreflux or possible oesophagitis.The diagnosis is supported by observing

the descent of a radio-opaque semi-solidbolus and the site of induction of pain; andconfirmed by inflating and deflatingballoon in 2 cm steps up the oesophagus.Any part or the whole of the oesophagusmay be consistently sensitive to a balloondiameter unnoticed by the remainder ofthe patients' oesophagus, or normalsubjects or patients with oesophagitis.Treatment is empirical. A liquid diet iscomfortable. Spontaneous relief iscommon. Two patients had an oesophagealdenervation which relieved 'tenderness'but produced other problems.

COELIACF76-F81

F76Splenic function and intestinal enzymeactivity in coeliac relatives: is there a'coeliac trait'?

J G O GRADY, E M CRYAN, F M STEVENS, B

MCNICHOLL, P FOTTRELL, T A O'GORMAN, AND

C F MCCARTHY (Medical Research CouncilofIreland, Coeliac Unit, Regional Hospital,Galway and University College,Galway) Splenic function was assessed bycounting the percentage of 'pitted' erythro-cytes in venous blood in 29 healthy

A9)19


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/


controls, 29 first-degree coeliac relativeswith normal small bowel histology, and 37coeliacs (both treated and untreated). Allcoeliacs and relatives were under the age of25 years and controls under the age of 30years. The 'pitted' erythrocyte counts weresignificantly higher in both the coeliacs(7.4±3.6, p 0.0001) and the relatives(6.1±2.7, p 0.0001) than in the controlpopulation (3.15±1.7). No significantdifference was found between the counts inthe coeliacs and relatives (p 0.1).

Small intestinal brush border activity oflactase, sucrase, and alkaline phosphatasewas measured in 36 coeliac relatives withnormal histology. The levels of activitywere compared with normal ranges for age,sex, and site of biopsy and expressed inrelation to the mean and 5% confidencelevel for the appropriate control group.Enzyme activity tended to be in the lowerrange for all three enzymes and in the caseof lactase activity 17% of cases were belowthe 5% confidence level.No difference was found between HLA

B8 positive and HLA B8 negative subjectsfor either 'pitted' erythrocyte counts orenzyme activity.The finding of these abnormalities in

healthy first-degree coeliac relatives withnormal small bowel histology raises thepossibility of the existence of a 'coeliactrait'.

F77HLA antigens in coeliac disease associatedwith malignancy

CHRISTINE M SWINSON, P J HALL, PENELOPE A

BEDFORD, AND C C BOOTH (Northwick ParkHospital and Clinical Research Centre,Harrow, Middlesex) HLA A, B, C, andDR antigen typing was carried out on 45unrelated patients included in a nationalregister of coeliac patients with malig-nancy, 20 (44%) of whom had lymphomasand 25 (56%) carcinomas. The results werecompared with a group of 48 coeliacpatients without malignancy to see whetherthere are genetic factors separate fromthose predisposing to coeliac diseaseassociated with the development of malig-nancy in coeliac patients. The study hasalso provided data on HLA DRassociations with coeliac disease in Britain.No significant differences in HLA antigenfrequencies were found between the twocoeliac groups, but compared with normalcontrols, coeliac disease was associatedwith DR 3 (X2 with Yates correction = 92,p<<0.001, relative risk 23.4) and DR 7 (X2

with Yates correction = 10-5, p<0-02,relative risk 2.45). Thus this study hasshown for the first time that, in Britain,coeliac disease is associated with DR 3 andDR 7, and that coeliac patients whodevelop malignancy share a commongenetic background with those who do not,suggesting that the incresed susceptibilityof coeliac patients to malignancy is deter-mined mainly by environmental factors.The authors are greatly indebted to numerous clinicians

throughout Britain who have sent blood samples frompatients under their care, Professor J J van Rood whosupplied some of the anti-sera. and to Dr A S Penia. fortheir generous help and cooperation.

F78Malnutrition caused by occultmalabsorption in elderly patients

A MCEVOY AND 0 F W JAMES (Departmentof Medicine (Geriatrics), FreemanHospital, Newcastle upon Tyne) Westudied 55 patients who were over 65 yearsold, admitted to hospital, and found tohave malnutrition on the following criteria.Clinical, anthropometric, haematological(Hb, blood film, folates, B 12), andalbumin + bone chemistry. Thecommonest problems leading to admissionwere: 'off feet', 15 patients; diarrhoea(often spurious), 15; weight loss, 13;dementia, seven; bone pain, six. Carefuldietary assessment suggested that in 31/55the cause of malnutrition was dietarydeficiency alone. Among the remaining 24patients, with no previously documentedcause for malabsorption, the followingtests were carried out: xylose tolerance test(one hour serum sample), 14C-trioleinbreath test, barium meal and radiology ofthe upper gastrointestinal tract, endoscopywith small bowel biopsy, and duodenalaspiration for aerobic and anaerobicbacteriology + volatile fatty acids (VFAs),culture was said to be positive if > 105organisms detected; also '4C-glychocholicacid breath test. If pancreatic disease wassuspected ERCP was carried out. Thefollowing were the causes for mal-absorption: bacterial contamination ofsmall bowel, 17 patients (nine diverticula,four post-gastric surgery, four 'no sump').Only one had anaerobic organisms with+ve VFAs. Coeliac disease, two; chronicpancreatitis, two; short bowel, one;refused tests, two. Following nutritionalsupport and specific treatment 20/24malabsorption patients returned home,only 15/31 poor diet patients were able todo so.We conclude that, in old patients with

occult malabsorption, small bowelbacterial contamination is common andtreatable.

F79Gastrointestinal symptoms and specific foodintolerance

D A FARAH, I T CALDER, L BENSON, AND J FMACKENZIE (Gastroenterology Unit andDietetic Department, Royal Infirmary,Glasgow) Forty-eight patients presentingwith long-standing gastrointestinalsymptoms of unknown cause, principallyvomiting and/or diarrhoea, have beeninvestigated for specific food intolerance.Organic disease was excluded as far aspossible using routine tests. Patients took alow allergenicity diet for up to 14 days.Those patients whose symptoms remittedcompletely, during this time, were con-sidered likely to have specific foodintolerance. Offending specific foods wereidentified in subsequent weeks by single,weekly, food reintroduction. In sevensubjects foods were identified in this way asa cause of symptoms. These were con-sidered suitable for further investigation byrandomised, double-blind, specific food/placebo challenge. The challenge was givenin the form of opaque capsules as freeze-dried specific food or placebo (glucose).Three capsules were given three times perday for one week or until symptomsbecame intolerable. In two subjectsdouble-blind challenge confirmed thesuspected food as a cause of the patients'typical symptoms. In the remaining fivesubjects there were substantial placeboreactions. It is concluded that specific foodintolerance is a clinical entity, that double-blind challenge with specific foods isfeasible, but that the clinical yield in thegastrointestinal clinic is likely to be low.

F80Effect of dietary fibre on zinc absorption

D A FARAH, M J HALL, P R MILLS, AND R IRUSSELL (Gastroenterology Unit, RoyalInfirmary, Glasgow) Dietary fibre, nowcommonly used in the long-term manage-ment of diverticular disease and IBS, maycontribute to zinc deficiency by reducing itsabsorption. Most of the evidenceimplicating fibre in the aetiology of zincdeficiency is, however, indirect andperhaps inconclusive. The effect of dietaryfibre on zinc absorption was thereforestudied in normal volunteers using a zinc

A920


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/


tolerance test and whole body monitor tomeasure the seven-day retention of 65Zinc(65Zn).Twenty-three subjects (age 21-26 years)

were randomised into three groups. Allsubjects were given 5 11Ci of 65Zn orally ina 10 ml solution of zinc sulphate containing15 mg of elemental zinc. In addition, groupA were given 20 g of wheat bran; group B20 g of Rice Krispies (a low-fibre foodwhich contains one-tenth the amount offibre in an equivalent weight of wheatbran), and group C had zinc alone. Theseven-day percentage retentions(mean ± 1 SD) for the three groups wereas follows: group A, 2.43±1t28; group B,21-45±6-6; group C, 57-9±13-3. Alldifferences were significant (A vs B,p<0c05; B vs C and A vs C, p<0U01). Thezinc tolerance test response in group A wasflat; in groups B and C there was asignificant rise in serum zinc at 90-120minutes. The areas under the zinctolerance test curves (mean ± 1 SD ingmol/min/l) in the three groups were asfollows: group A, 3349+271-8; group B,4004±357-1; group C, 5189±367-4. Alldifferences were again significant (A vs B,p<0-05; B vs C and A vs C, p<0-01).We conclude that dietary fibre leads to a

significant reduction in zinc absorptionwhich could eventually induce a state ofzinc deficiency.

F81Differential absorption of D-xylose and3-O-methyl-D-glucose in coeliac disease andacute gastroenteritis

C NOONE, R C BEACH, J BULL, AND I SMENZIES (St Thomas's Hospital MedicalSchool) The rate of gastric emptying andspace of distribution after absorption arevariables that complicate interpretation ofthe D-xylose (D-xyl) test. Inclusion of asecond marker with susceptibility to thesefactors should provide correction, butresponse to the absorptive lesion mustdiffer from that of D-xyl to render a usefuldiscrimination. 3-O-methyl-D-glucose(3mGlc) has been given orally, with D-xyl,for this purpose to fasting subjects as adifferential absorption test (2.5+5.0 g,respectively, in 250 ml water: half-dose forinfants). Samples of plasma taken at 30, 60,90, and 120 minutes for adults and of wholeblood at 60 minutes for infants werepreserved for sugar analysis by thin-layerchromatography.Plasma D-xyl concentration at 60

minutes gave separation between healthy

adults and patients with untreated coeliacdisease (0.847±0.119 and 0-277±0-168mmol/l, respectively, mean ± SD: n=16for each group) but the D-xyl/3mGlc ratiosgave a better within-group reproducibilityand between-group discrimination(0.866±0.104 and 0-359±0-129,respectively). Resolution derived fromboth values was greatest at 60 minutes andbecame progressively less at 30, 90, and120 minutes. Twenty-five infants (age 3weeks to 2 years) with acute rotaviralgastroenteritis gave abnormal D-xyl/3mGlcratios (0-386±0- 123) rising to normal(0.760±0.077) after recovery two to fourweeks later.Employment of a D-xyl/3mGlc ratio

exploits the relative insensitivity of 3mGlccompared with that of D-xyl to the effectsof small intestinal pathology and gives amore reliable index of absorption than thatprovided by the use of D-xyl alone.

PAEDIATRIC GASTROENTEROLOGY GROUP

Fundoplication in children: before and afterassessment using oesophageal pHmonitoring

C A CAMPBELL, D OUTRAM, AND V M WRIGHT(introduced by J A Walker-Smith) (QueenElizabeth Hospital for Children, London)Using a small flexible pH probe, 24 hourmonitoring of oesophageal pH was per-formed in 48 infants and children either inan effort to diagnose suspected gastro-oesophageal reflux or in the assessment ofthe severity of known reflux. A usefuladjunct has been the ability to dynamicallymonitor the effects of various medicaltreatments and it has helped to clarifycriteria for selection of children foroperative correction of gastro-oesophagealreflux.

Results of monitoring in children arepresented with special emphasis on sixchildren assessed before fundoplicationand 12 assessed after surgery. The resultsshow that oesophageal pH monitoringgives additional useful informationregarding gastro-oesophageal reflux and isan easily performed method of assessingthe success of treatment including surgery.

Fibreoptic endoscopy in the early diagnosisof upper gastrointestinal tract haemorrhagein childhood

P J MILLA, J T HARRIES, AND L SPITZ (Institute

of Child Health and Hospital for SickChildren, London) Early fibreoptic endo-scopy is a routine in the manage-ment of upper gastrointestinal tracthaemorrhage in adult patients. With theavailability of more suitable paediatricinstruments we have routinely investigateda series of 28 children early in the course oftheir haemorrhage. We present the endo-scopic findings here.The site of bleeding was localised in 24

cases (86%), which compares well withboth adult and paediatric series. Thisdiagnostic score in considerably higherthan with other conventional techniquescurrently employed. Bleeding was notinfrequent in infancy, 28% of childrenpresenting were under the age of 1 year,35% aged 1 to 6 years, and 37% over theage of 6 years. Oesophagitis was thecommonest lesion encountered in infantsbut the bleeding was mild compared withthe older children. Over the age of 6 yearsgastroduodenal ulceration was thecommonest lesion and frequently resultedin significant haemorrhage causing bothhaematemesis and malaena. Aspiriningestion preceded bleeding in 50% of thechildren with gastritis, the remainder beingsick with severe systemic disease. In thisseries bleeding in patients withoesophageal varices was variceal ratherthan due to gastritis or ulcer.

Fibreoptic endoscopy gives valuableinformation regarding the site of the lesionresponsible for the bleeding and addition-ally immediate therapy may be possible.Endoscopy should be performed as a firststep investigation in upper gastrointestinaltract haemorrhage.

Rectal mucosal acetylcholinesterase inHirschsprung's disease

G DALE, D J SCOrr, AND J WAGGET (FlemingMemorial Hospital for Sick Children, AtticLaboratory, Royal Victoria Infirmary,Institute of Pathology, Newcastle GeneralHospital, Newcastle upon Tyne) The pre-operative diagnosis of Hirschsprung'sdisease is dependent on the histologicalevidence of hypertrophied nerve trunksand the absence of ganglion cells. Thebiopsy may have insufficient submucosaltissue to show that ganglion cells are trulyabsent; inadequate samples are common.The demonstration of mucosal nerve

fibre proliferation by acetylcholinesterasehistochemistry, together with direct assayof acetylcholinesterase activity in mucosalbiopsies, have usefully extended the

A921


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/


investigative repertoire. Duplicate rectalsuction biopsies were performed on 120children (newborn to 14 years). Onespecimen was processed for histochemistryand, in the other, total cholinesterase andacetylcholinesterase activities were assayedusing ethopropazine as a selective inhibitorof non-specific cholinesterase.

In patients found not to have Hirsch-sprung's disease ('normals'), the acetyl-cholinesterase was independent of age,with a mean value for the group of 6-7 U/l(SD 4.04). In the group with Hirsch-sprung's disease (16 cases) the mean was40.3 U/i (SD 28-6, range 8.5-130.5).Acetylcholinesterase expressed as apercentage of total cholinesterase activityis a useful additional measurement: normalsubjects had a mean value of 503% (SD11-2%, range 40-76%), cases with Hirsch-sprung's disease gave a mean of 76% (SD4-7%, range 66-84%).We conclude that the combination of

histochemistry and direct measurement ofacetylcholinesterase in rectal biopsies is auseful adjunct to histology, offering thecapability of same-day response with fewersamples of inadequate size - an importantconsideration in the neonatal period.

Lymphocyte cytotoxicity to autologoushepatocytes in al antitrypsin deficiency: aconsetluence of liver damage

M MONDELLI, A L W F EDDLESTON, RWILLIAMS, G MIELI VERGANI, AND A P MOWAT(Department of Child Health and LiverUnit, King's College Hospital and MedicalSchool, London) The severity of liverdisease in children with antitrypsin (odAT) deficiency is variable and themechanism responsible for hepatocytedamage is unknown. To investigatewhether cell-mediated immune reactionsare involved in the pathogenesis of liverdamage, peripheral blood lymphocytesfrom nine children with liver disease andal AT deficiency were incubated in amicrocytotoxicity assay with the patients'own hepatocytes obtained from a diag-nostic liver biopsy. All had presented withhepatitis in infancy, but were studied fromone month to eight years later. Four withhistological features of severe neonatalhepatitis were studied within six months ofthe onset and none had significantlyincreased lymphocyte cytotoxicity(mean ± SD = 13.2%±15.6). Twochildren, 12 and 15 months old, one withcirrhosis and one with minor histological

changes, had values around the upper limitof normal (32% and 34%), while theremaining three children all over a yearold, had significantly increased lymphocytecytotoxicity (53%+7) irrespective of theseverity of liver damage (one minorchanges and two cirrhosis). Overall therewas a significant exponential relationshipbetween percentage cytotoxicity andduration of disease (r=0.81, p<0.01) orage. Thus, cell-mediated immune reactionsdirected against autologous hepatocytesare present in some children with liverdisease associated with (xl AT deficiencybut not at the onset. This suggests that suchimmunological reactions are secondary tothe initial liver injury, but whether theymay contribute to the perpetuation ofhepatocyte damage remains to be estab-lished.

Does Indian childhood cirrhosis (ICC)result from viral infection in a copper-ladenliver?

M S TANNER, A J E FLOWER, S A BHAVE, ANDA N PANDIT (Public Health Laboratoryand Department of Child Health, Leicester,UK, and Department of Paediatrics, KingEdward Memorial Hospital, Pune, India)During a prospective study of paediatricliver disease in Pune, India, 115 pretrans-fusion sera were examined for evidence ofhepatitis A, hepatitis B, Epstein-Barr, andcytomegalovirus infection. Fifty-sevenchildren with Indian childhood cirrhosishad liver copper values of 1836±673 ,ug/g(controls 8-118 1£gIg). Of these, 16 hadHAVAB (28%); 19 had EBV IgG (33%),and five EBV IgM; 41 had CMV IgG(72%) and one CMV IgM. None hadmarkers of HBV. Of 35 children with otherhepatic disorders in whom liver coppervalues were 189±298 ,ug/g, 24 (69%) hadHAVAB and 11 HAIgM; 69% EBV IgG,and 57% CMV IgG; five had HBVmarkers. In controls, 46% had HAVAB,39% EBV IgG, and 85% CMV IgG.None of the viruses tested are specifically

associated with Indian childhood cirrhosis.The absence of HBV markers in thisdisorder contrasts with studies fromsouthern India, and may explaindifferences in descriptions of Indian child-hood cirrhosis. HA was common,increased with age, and correlated withEBV infection. The high frequency of viralinfections indicates the importance ofseeking non-A non-B hepatitis markerswhen available.

Wilson's disease in childhood: a plea forearly diagnosis

H NAZER, R EDE, A P MOWAT, AND ROGERWILLIAMS (Departments of Child Healthand Liver Unit, King's College Hospital,London) Although Wilson's disease is atreatable disorder, nine of 15 cases referredwith undiagnosed liver disease have diedwithin three to 53 days of admission. Toidentify features that would enable earlierdiagnosis and improvement of manage-ment we have retrospectively reviewedthese cases. The presenting symptoms werelethargy and malaise, 11 cases; jaundice,11; abdominal pain, nine; deterioratingschool performance, four. The meanduration of symptoms was 4-8 months(range one to 12 months) in fatal cases and15-4 (one to 18 months) in survivors.At diagnosis all fatal cases had jaundice

and ascites, eight had hepatomegaly, sevensplenomegaly, and six oedema. Only onesurvivor had ascites, two jaundice, twooedema, although four had hepatospleno-megaly. Kayser Fleischer rings werepresent in six fatal cases and five survivors.All cases had biochemical evidence of liverdisease and 14 had low caeruloplasmin andraised urinary copper concentration.Serum bilirubin concentrations, aspartatetransaminase, and prolongation of pro-thrombin time were significantly moreabnormal in the fatal cases (p<0.01) ascompared with the survivors. Cirrhosis waspresent in all fatal cases and in foursurvivors. Penicillamine caused gradualclinical and biochemical improvement inthe survivors, except one who requiredTETA, but in the fatal cases penicillamineand steroids (six cases), BAL (one case),TETA (one case) had no discernible effect.

Wilson's disease must be excluded notonly in children presenting with liverdisease but also in children with persistinglethargy, abdominal pain, or deterioratingschool performance. Specific therapy mustbe initiated before decompensatedcirrhosis develops.

Abnormalities of postprandial smallintestinal motor activity in childhood: theirrole in the pathogenesis of irritable bowelsyndrome

T R FENTON, J T HARRIES, AND P J MILLA(Institute of Child Health, Department ofChild Health, and the Hospital for SickChildren, Great Ormond Street, London)Toddler diarrhoea, or the irritable bowel

A922


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/


syndrome in childhood, is the commonestcause of chronic diarrhoea without failureto thrive in childhood, yet little is known ofthe pathophysiological mechanisms.

Using constantly perfused triple lumencatheters; we have previously shown thatthe fasting small intestinal motor activity isnormal in these children, but, in contrastwith controls, their migrating motorcomplexes are not disrupted by intra-duodenal dextrose.With the same technique we have

demonstrated that postprandial activity isinduced by constant duodenal perfusion ofmilk formulae and isosmotic hexosesolutions in children with a variety ofdisorders. When compared with thosesubjects (n=9) whose migrating motorcomplexes were disrupted by intra-duodenal dextrose (group 1), subjects(n=10) whose migrating motor complexeswere not disrupted by intraduodenaldextrose (group 2) demonstrateddifferences in their postprandial activity.Normal postprandial activity was

induced in only six out of 10 in group 2 vsnine out of nine in group 1. The return offasting activity (signalled by the appear-ance of an migrating motor complex) wasfaster in group 2, occurring in less than 30minutes in five out of six subjects vs oneout of five in group 1. Migrating motorcomplexes were seen during the duodenalperfusion in four out of 10 of group 2 vsnone out of nine in group 1.We suggest that these abnormalities in

postprandial activity play a role in thepathogenesis of the diarrhoea in childrenwith irritable bowel syndrome.

'M cells' in rat and childhood follicle-associated epithelium in the small intestine

D JACKSON, J A WALKER-SMITH, AND A D

PHILLIPS (Queen Elizabeth Hospital forChildren, London) Specialised M cells,found in follicle-associated epithelium arethought to be important in the uptake andpresentation of antigen to lymphocytes inthe epithelium. They have been shown toexist in adult human Peyer's patches and,in view of their importance, we haveinvestigated their incidence and cyto-chemistry in isolated lymphoid follicles inchildhood, and compared them with Mcells in rat Peyer's patches.

Isolated lymphoid follicles are frequentlyfound in jejunal biopsies from children andmorphologically resemble follicles fromPeyer's patches. They contained, however,only an occasional M cell. In contrast,

follicular epithelium from jejunal Peyer'spatches in rats contained numerous Mcells, their numbers varying betweenindividual follicles, with up to 30% of theepithelial cells overlying the follicle beingM cells. In both children and rats M cellmorphology was found to be variable.An alkaline phosphatase reaction was

performed on some of the follicles. It wasfound that M cells had a reduced amount ofalkaline phosphatase activity along theirapical membranes compared with neigh-bouring columnar epithelial cells overlyingthe follicles.Thus the incidence of M cells varies

greatly in Peyer's patches from rats and arerarely seen in isolated lymphoid folliclesfrom children. Their incidence may berelated to the size and/or functional state ofthe follicle, as the isolated lymphoidfollicles from children tended to be smallerand less often possessed a germinal centre.The reduced alkaline phosphatase activityfound along the M cell apical membranemay be a sign of immaturity or of a processof differentiation from columnar cells.

Characteritics of jejunal lntraepithelialand Peyer's patch lymphocytes of neonataland malnourished rats

M J BRUETON AND N LYSCOM (Departmentof Child Health, Westminster Children'sHospital, London) The development ofantigen handling mechanisms in neonatalgut is essential for the establishment ofprotective immunity, systemic tolerance tooral antigen, and the avoidance of hyper-sensitivity reactions. As part of an investi-gation of immunoregulatory mechanisms,T lymphocyte subsets and B cells in thesmall bowel epithelium and Peyer's patcheshave been characterised in isolated cellsuspensions from suckling rats and normaland malnourished (6% protein diet)animals. B cell markers and the mono-clonal antibodies W3/13, W3/25, and MRCOX 8 were used. The latter are consideredto be specific for T cells, T helpers (Th),and T suppressors (Ts) respectively.

Intraepithelial lymphocyte numberswere reduced in the young and in themalnourished. At birth the MRC OX 8marker was the only one expressed (40%).These unique intraepithelial lymphocytes,which expressed the Ts marker but lackedreceptors for the Pan-T marker, were alsopresent among mature IEL (50%). Cellsbearing the W3/25 and W3/13 markersappeared from the age of 3 weeks, B cellswere rarely seen in intraepithelial lympho-

cyte preparations. At three weeks Peyer'spatch lymphocytes showed a maturedistribution of subsets (18% W3/13, 11%W3/25, 10% MRC OX 8, 49% B). In themalnourished animals the number of intra-epithelial lymphocytes expressing MRCOX 8 markers was significantly lower(p<0.01) than in well-nourished controlstwo and three weeks after weaning. Noother differences in lymphocyte subtypeswere noted in intraepithelial lymphocytesor Peyer's patch lymphocytes. The functionof intraepithelial lymphocytes is poorlyunderstood but this demonstration ofdisturbances in their numbers and surfacemarker characteristics at times when theindividual is immunologically com-promised suggest the possibility of intra-epithelial lymphocyte involvement inmucosal immunity.

Astrovirus within human small intestinalmucosa

A D PHILLIPS, S J RICE, AND J A WALKER-SMITH(Queen Elizabeth Hospital for Children,Hackney Road, London) Astrovirus hasnot been detected within cells of the humangastrointestinal tract, although it is foundin stools from children with gastroenteritis,has been associated with diarrhoealoutbreaks, and can produce diarrhoea andsero conversion in adult volunteers.We now report the findings of astrovirus

particles within the small intestinalepithelium of two children admitted forinvestigation of gastrointestinal problems.One child (JF) had a history of vomitingand weight loss after cow's milk formulafeeding. The other child (RS) hadrecurrent intermittent diarrhoea. Bothdeveloped bouts of diarrhoea in hospital,that of RS following regrading onto cow'smilk. Small intestinal biopsies wereperformed to investigate the presence ofenteropathy. Stools taken for microbiologywere found to contain astrovirus on the dayof biopsy (JF and RS) and subsequently(RS). Electronmicroscopy demonstratedastrovirus within epithelium in the lowerpart of the villus (RS) and exposed surfaceepithelium of a severe enteropathy (JF).This finding implicates the small intestineas a site of astrovirus replication.The interpretation of its relevance to the

patient is complicated by the presence ofother stool pathogens before the astrovirus(E coli 086 (JF) and rotavirus (RS)) and byother confirmed diagnoses (cow's milksensitive enteropathy (JF) and sucrase-isomaltase deficiency (RS)).

A923


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/

A924

These astrovirus infections highlightintercurrent infection as a problem ofdifferential diagnosis and give insight intothe interaction between astrovirus andsmall intestinal mucosa in man.

In vitro gluten sensitivity in a patient withfamilial hypomagnesaemia andunresponsive subtotal villous atrophy(SVA)

P D HOWDLE, A J CROLLICK, M S LOSOWSKY,AND J M LiTFLEWOOD (Departments ofMedicine and Paediatrics, St James'sUniversity Hospital, Leeds) Familialhypomagnesaemia is believed to be due toa specific defect in intestinal absorption.Jejunal mucosa has been normal in the fewcases biopsied. Subtotal villous atrophy inolder British children is usually due tocoeliac disease. We report a case of familialhypomagnesaemia with subtotal villousatrophy, an association not previouslyreported.LW, a 12-year-old daughter of English

parents, presented at 2 weeks of age withconvulsions. Hypomagnesaemia wasdiagnosed and magnesium supplementswere started. Two infant brothers hadpreviously died of hypomagnesaemia. At 4years magnesium malabsorption wasconfirmed; however, malabsorption of fatand xylose and iron-deficient anaemia werealso found. Jejunal biopsy showed subtotalvillous atrophy, but there was no histo-logical improvement 10 and 23 monthsafter starting a gluten-free diet. For thepast three years she has taken a normaldiet. She has always grown normally withno gastrointestinal symptoms, and hasremained on oral magnesium supplements.Jejunal biopsies on three occasions toassess the effect of vitamin and zinc supple-mentation, before re-starting a gluten-freediet have shown subtotal villous atrophy.Each time, mucosa was maintained inorgan culture, and was sensitive in vitro toFrazer's fraction III, alpha- andunfractionated gliadin, but not casein, asassessed by changes in enterocyte height,thus behaving similarly to that fromuntreated coeliac patients.These results suggest that, despite the

lack of response to a gluten-free diet, thischild has coeliac disease, and raise thequestion whether the hypomagnesaemiaprevents restoration of the jejunal mucosain vivo.

Food allergy: the major cause of infantilecolitis?

H R JENKINS, P J MILLA, J R PINCOTr, J F

SOOTHILL, AND J T HARRIES (The Hospitalfor Sick Children, Great Ormond Street,and the Institute of Child Health, London)Ulcerative colitis is rare in childhood andthere is an apparent peak incidence in thefirst year of life, although why this shouldbe is not clear.

In this paper we report eight suchpatients seen during a three year period,and six of these presented below the age of4 months; colonoscopy showed the typicalappearances of ulcerative colitis.There was a very close relationship

between the introduction of artificialfeeding and the onset of symptoms - thatis, bloody diarrhoea. Cow's milk proteinwas the commonest allergen, but soya

protein and beef provoked symptoms inthree of the cases. Eczema, eosinophilia,and a strong family history of allergy werecommon. Endoscopic biopsies showed amarked increase of eosinophils in thelamina propria, and IgE antibodies tocow's milk protein were common. Thecolitis rapidly and completely resolved onan antigen exclusion diet.These results suggest that food allergy is

an important cause of ulcerative colitis inthe first year of life, and that an antigenexclusion diet is the treatment of choice.Moreover, the results emphasise theimportance of colonoscopy as a diagnostictool.

Intestinal permeability in Crohn's disease

E J EASTHAM, A D J PEARSON, M F LAKER, ANDR NELSON (Department of Child Health,Royal Victoria Infirmary, Newcastle)Intestinal permeability has been studied in17 children with Crohn's disease involvingthe small bowel. The probe moleculesmannitol and lactulose were given orally ina moderately hypertonic solution (580mosmol/l) and urine collected for fivehours. Analysis of sugars was performed bygas liquid chromatography. By expressingthe result as a ratio of these moleculesexcreted, patients with active Crohn'sdisease had a five-fold increase inpermeability when compared with anormal control group of 31 children(p>0.001). This was found to be duemainly to an increased lactulosepermeability, suggesting changes atjunctional complexes and an increase in


extrusion zones at the villous tips.Comparing the permeability ratio with aclinical rating of disease severity andseparate index of inflammatory activitythere was a significant correlation bothbetween patients (p>O0O5) and particularlywithin (p>0.01) patients studied on severaloccasions. This became more significantwhen correlating percentage lactuloseexcretion. We conclude that intestinalpermeability, particularly to lactulose, isincreased in active Crohn's disease of thesmall bowel. The test is of clinical value asit is non-invasive, can be performed simplyat home, and appears to correlate withdisease activity.

Intracellular particles in childhood inflam-matory bowel disease - an enteric micro-organism?

D C LEWIS, J A WALKER-SMITH, J D ALMEIDA,AND A D PHILLIPS (Queen ElizabethHospital for Children, London, andWellcome Research Laboratories, Becken-ham) It has often been postulated that atransmissable organism is involved in thepathogenesis of Crohn's disease, however,electronmicroscopy has failed to demon-strate the presence of an infective agent.Using electronmicroscopy we have

found morphological evidence of an entericmicro-organism of a previously un-described type, in (1) a colonic resectionfrom a child with Crohn's disease, (2) ajejunal biopsy from a child with inflam-matory bowel disease, and (3) a jejunalbiopsy from a child with a persistententeropathy of undetermined cause, whohas since been lost to follow up.The putative micro-organism appeared

as dense intracellular particles in the apicalcytoplasm and microvilli of surfaceepithelial cells, and as free membranebound bodies (70-110 nm diameter).Particles also occurred inside protrusionsof the microvillous membrane suggestingthat they are either budding out of the cell,or are luminal bodies which have fusedwith the plasma membrane. In our experi-ence no epithelial organelle could accountfor this appearance. Negative stainingelectronmicroscopy of an homogenate of acolonic biopsy from case 1 revealedparticles of a size consistent with those seenin sections, and when reacted withhomologous serum these showed antibodyattachment.The identity of the particle and its

relevance to inflammatory bowel disease


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/


are of considerable interest. Studies arecurrently underway to establish thefrequency of this finding and to determineif the particle has nucleic acid content.

Zinc in children with Crohn's dsease

N J MEADOWS, C A CAMPBELL, P AGGETT,S CHONG, AND J A WALKER-SMITH (Academic Unit of Child Health, StBarthblomew's Hospital, London) Zinc isan important co-factor in DNApolymerase, RNA polymerase, andthymidine kinase activity, these enzymesbeing essential for cell replication andprotein synthesis. Zinc deficiency maytherefore prove to be an important growth-limiting factor in children with Crohn'sdisease. As there have been conflictingreports of both raised and reduced plasmazinc levels in adults with Crohn's diseasewe have studied plasma zinc and copperlevels in 21 children with Crohn's disease,monitoring disease activity and growth.The mean plasma zinc level in the

children with Crohn's disease was 10.3nmol/l compared with the normal mean of17-5 nmol/l, with 57% of the values lessthan two standard deviations from thenormal mean. The differences are highlysignificant, but there was no correlationwith alkaline phosphatase (a zinc-dependent enzyme), serum albumin, ESRor disease activity index. The plasmacopper values in the children were normal.This pattern suggests that true tissue zincdepletion is present, which may beimportant in the mechanism of growthfailure in children with Crohn's disease.Work is in progress to determine the tissuezinc status in the children.

Disaccharidase activities in intestinal fluidrefect local tissue levels

L ARAMAYO, H DE SILVA, G A BROWN, AND A SMCNEISH (Institute of Child Health,University of Binningham, Birmingham)There are clinical and research circum-stances in which it would be desirable tomake (serial) measurements of brushborder disaccharidase but ethical con-siderations make repeated jejunal biopsyunjustifiable. We have therefore deter-mined the degree to which disaccharidaseactivities in jejunal fluid reflect theactivities of these enzymes in mucosalbiopsies obtained simultaneously from thesame site.Duplicate assays of disaccharidase

activities were made on fasting,homogenised fluid and conventionalmucosal homogenates. Eight children withsubtotal villous atrophy (SVA) and 15children with normal mucosal (N) were

studied.There was a highly significant

(p<0.0005) positive correlation(Spearman's rank correlation) betweenenzyme levels in tissue (T, U/g) and fluid(F, U/1) in all cases (lactase: NT 5 04±1 80(mean ± SD); NF 17.49±8-58; SVAT0-48±0.51; SVAF 2.54±1.94; sucrase: NT8.80±2.91; NF 26.41±16.9; SVAT2-15+1.94; SVAF 9-07±6-9; maltase: NT29.8±10; NF 138±64; SVAT 8-83±5-56;SVAF 66.26±22.6).The significant positive correlation

between tissue and juice disaccharidaseactivities suggests that the mucosa is thepredominant or sole source ofdisaccharidase activity in the juice.Preliminary kinetic studies comparingtissue and juice enzyme properties alsoindicate a mucosal origin for the juiceenzyme.We conclude that disaccharidase

activities in jejunal fluid reflect local tissuelevels and may be a useful alternativeprocedure when jejunal biopsy is notpossible.

Fetal secretory diarrhoea: a new congenitalintestinal transport defect

P J MILLA, T R FENTON, AND J T HARRIES(Institute of Child Health and Hospital forSick Children, London) Secretorydiarrhoea may occur from birth. To datethe only defect solely of intestinal iontransport which has been clearly defined isof anion exhange in congenitalchloridorrhoea. We present a patient inwhom intestinal secretion started in utero.

Transport of water and solutes wasstudied in the jejunum and rectum of thispatient using jejunal perfusion and non-equilibrium rectal dialysis techniques. Inthe jejunum, water (-35, control +89±10gd/min/cm) and electrolytes (Na+ -1-26controls +7.65±1.59; Cl- -1*18 controls+4.62±1.0 1mol/min/cm) were secreted,but glucose and fructose absorption was

intact. In the rectum there was impairmentof Na+ and Cl- absorption with increasedK+ and HCO3 secretion. Further studiesof the jejunum using perfusates of high(8.2), neutral (7.1), and low (6.2) pHshowed that secretion of hydrogen ion(-0.015 ,smol/min/cm) occurred only athigh luminal pH with a fall in pH from 8-2

to 7.4; in these circumstances Na+ wassecreted (-15.67 gimolImin/cm). At lowpH H' absorption occurred (+0-28 umol/min/cm, pH 6-2 to 7.1) and Na+ secretionwas reduced to -1-83 ,mol/min/cm.These data suggest that Na+/H+

exchange is defective in the proximaljejunum of this patient with only passivemovement of H' ion. We speculate thatthe secretory diarrhoea is due to apreviously undescribed defect of intestinalion transport.

Functional development of the neonatal gutas measured by electrogenic glucoseabsorption

A J MAYNE, D A DUCKER, G AUCOlT, C AHUGHES, AND A S MCNEISH (Institute ofChild Health, University of Birmingham,and Department ofChild Health, Universityof Leicester) The optimal regime forfeeding immature and low birth weightinfants is not known. This dilemma wouldbe helped by a better understanding of theabsorptive capacities and postnataldevelopment of the immature gut. Wehave therefore continued our studies ofglucose absorption from the small intestineof newborn infants using changes in trans-mucosal potential difference (PD) as ameasure of active glucose transport.The distal duodenum of infants was

perfused via a polyvinyl feeding catheter at1.5 ml/min with 0.40 mM glucose solutionsmade isomolar with mannitol andcontaining 134 mM NaCl. Fullterm (FT)and preterm (PT) infants appropriate forgestational age (AGA) and light for gesta-tional age (LGA) were studied in the firstpostnatal week. A further group ofPTAGA infants was studied for up to sixweeks postnatally.We confirmed our previous results that

PDmax was significantly reduced inPTAGA (n=6) and PTLGA (n=9)compared with FTAGA (n=8) infants(p<0-02, p<0-001). In PTAGA infantsPDmax increased from 10-1 ± 1x0 mV duringthe first postnatal week to 15.2±0.9 in thethird postnatal week (p<0.01), followed bya decrease to a new plateau in the fourth(12.0±0.7 mV), fifth (11.5± 10 mV), andsixth (12-0±0.7) weeks.We conclude that preterm and growth

retarded neonates have a reduced capacityfor glucose absorption, compared withfullterm infants. In preterm infants there ispostnatal surge in absorptive capacity forglucose during the second and third post-natal weeks.

A925


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/


The Society is grateful for the support of the following exhibitors and sponsors:

American Hospital Supply (UK) LimitedAyerst Laboratories LimitedBarr & Stroud LimitedBerk Pharmaceuticals LimitedBoots Company LimitedBrocades (GB) LimitedChurchill LivingstoneConcept Pharmaceuticals LimitedDowns Surgical LimitedDuphar Laboratories LimitedEthicon LimitedFisons LimitedGlaxo Laboratories LimitedGMI (Medical) LimitedKabivitrum LimitedKeymed (Medical & Industrial Equipment) LimitedLabaz Sanofi UK LimitedLilly IndustriesMCP Pharmaceuticals Limited, Infusion Solutions DivisionMCP Pharmaceuticals LimitedMedisco Equipment LimitedMerrell Pharmaceuticals LimitedMurray Systems LimitedNordic Pharmaceuticals LimitedNorgine LimitedOrmed LimitedPharmax LimitedPharmacia (GB) LimitedPortex LimitedPyser LimitedReckitt & Colman Pharmaceutical DivisionRimmer Bros. LimitedRoche Products LimitedRorer PharmaceuticalsW B Saunders CompanySeward SurgicalSmith Kline & French Laboratories LimitedStuart PharmaceuticalsChas. F Thackray LimitedTillotts LaboratoriesWeddel Pharamceuticals LimitedWinthrop Laboratories

A926


http://gut.bmj.com

/G


ctober 1982. Dow

nloaded from

http://gut.bmj.com/

British Society of Gastroenterology · Gut: first published as 10.1136/gut.23.10.A880 on 1 October...

Documents

Transcript of British Society of Gastroenterology · Gut: first published as 10.1136/gut.23.10.A880 on 1 October...